digital.csic.esdigital.csic.es/bitstream/10261/101768/4/tesis_boras.pdf · institut de ciències...

Institut de Ciències del Mar (CSIC) Universitat de Barcelona

Impact of viruses on bacterial communities

in marine systems

Impacto del virioplancton sobre las comunidades bacterianas

en sistemas marinos

Julia Anna Boras

Memoria presentada para optar al grado de Doctor en Ciencias del Mar por la Universidad de Barcelona, Departamento de Ecología, Programa de Ciencias del Mar, por Julia A. Boras.

Visto bueno Visto bueno Visto bueno de la directora de la tesis de la codirectora de la tesis de la tutora de la tesis

Dra. Dolors Vaqué Dra. M. Montserrat Sala Dra. Montserrat Vidal Investigadora – CSIC Investigadora – CSIC Investigadora – UB

Barcelona 2009

…a mis tres Chicas y dos Chicos –

mi Familia…

…to my Family…

…mojej Rodzinie…

Contents

AKNOWLEDGMENTS ......................................................................................................................9

LIST OF ABBREVIATIONS .............................................................................................................. 11

1. INTRODUCTION..................................................................................................................... 13

1.1. Thesis outline ............................................................................................................ 16

1.2. Introduction to virology ........................................................................................... 16

1.3. Bacteriophages as modifiers of bacterial communities ........................................ 22

1.3.1. Bacterial mortality ............................................................................................... 22

1.3.2. Bacterial diversity ................................................................................................ 24

1.4. Bacteriophages in the microbial loop .................................................................... 27

2. HYPOTHSES AND OBJECTIVES .............................................................................................. 31

3. REPORT OF THE THESIS DIRECTORS ........................................................................................ 35

4. PUBLICATIONS ...................................................................................................................... 41

4.1. Paper I. Annual changes of bacterial mortality due to viruses and protists in an oligotrophic coastal environment (NW Mediterranean) ................................. 43

4.2. Paper II. Role of viruses in shaping bacterial phylogenetic and functional diversity in oligotrophic marine coastal waters ...................................................... 61

4.3. Paper III. Changes of bacterial mortality and diversity as a response to viral and nanoflagellate communities interactions in an oligotrophic marine coastal site. ................................................................................................. 83

4.4. Paper IV. Effects of viruses and protists on bacteria in eddies of the Canary Current region (subtropical Northeast Atlantic) ................................................... 109

4.5. Paper V. Effect of ice melting on bacterial carbon fluxes channeled by viruses and protists in the Arctic Ocean ............................................................... 137

5. DISCUSSION AND CONCLUSIONS ...................................................................................... 163

5.1. Virus-mediated mortality of bacteria in the different systems studied ................ 165

5.2. Lysogeny ................................................................................................................ 172

5.3. Factors shaping viral mortality of bacteria ........................................................... 172

5.4. Viral infections in deep waters .............................................................................. 175

5.5. Effect of viral infections on bacterial phylogenetic diversity ............................... 175

5.6. Open questions ..................................................................................................... 176

5.7. Main thesis conclusions ......................................................................................... 176

6. RESUMEN DE LA TESIS (THESIS SUMMARY - Spanish version).............................................. 179

6.1. Introducción .......................................................................................................... 182

6.2. Hipótesis y objetivos de la tesis ............................................................................. 188

6.3. Discusión ................................................................................................................ 189

6.3.1. Mortalidad de bacterias debida a virus en distintos sistemas ........................ 189

6.3.2. Lisogénia ........................................................................................................... 192

6.3.3. Factores determinantes de la mortalidad bacteriana debida a los virus ..... 193

6.3.4. Infecciones víricas en aguas profundas .......................................................... 195

6.3.5. Effecto de las infecciones víricas sobre la diversidad filogenética de las bacterias .......................................................................................................... 195

6.3.6. Preguntas abiertas ............................................................................................ 196

6.4. Conclusiones de la tesis ........................................................................................ 197

7. ANNEX – METHODOLOGY .................................................................................................. 199

7.1. Virus enumeration by flow cytometry ................................................................... 202

7.2. Virus-mediated mortality of bacteria ................................................................... 205

7.3. Protist-enumeration by flow cytometry ................................................................ 208

8. REFERENCES ........................................................................................................................ 211

Acknowledgments/Agradecimientos/Podziękowania

P a g e | 9

ACKNOWLEDGMENTS /

AGRADECIMIENTOS/

PODZIĘKOWANIA

This thesis is the dream that came true thanks to a lot of people. I’m very grateful to all of

them, and especially to…..

To the director of my thesis, Dolors Vaqué, for the trust, and for giving me this great

opportunity of working with you. Thank you for a wonderful possibility of visiting the Arctic and

Antarctic, of travelling, and for your patience and help. This thesis wouldn’t be possible without

you.

I feel very lucky that I could spend this last five years with you, and learn from you not only about

the microworld, but also about how to be better person, more patient, and more positive.

To my codirector, Montse Sala. Thank you for your help with all details, your advices, and

reflections. And for the last minute-corrections, so important along this years J From the beginning

I felt more as you colleague than a student, and it was a great experience to work with you.

To my “angels”, who gave me their smiles and friendship. To Eli Sà – for listening always, for

your time for coffees, and your mails asking ‘How is it going?’. You are one of the very few real

angels that I’ve met in my life. To Sofia Loureiro – for your friendship, for motivating me and for

remembering me the real sense of our work. I hope that our energies will stay in a synchrony for a

long time.

To girls from my office: Arancha Lana, Pati Homs, and Bego Vendrell for conversations,

laugh, and for making each day in the ICM special and funny. And especially to Clara Ruiz, for

making me laughing always, even in the most difficult moments.

To Evariso Vázquez-Domínguez, for your help, conversations, and the JPM programme that saved

my life ;-). To Jordi Felipe, for hours and hours spent with the cytometer, or in the car, and for the

companion conversations.

Acknowledgments/Agradecimientos/Podziękowania

10 | P a g e

To my colleagues from the ICM – Ero, Ana Gomes, Pati Jimenez, Elena Lara, Gisela

Llaveria, Clara Cardelius, Vanessa Balagué, Irene Forn, Silvia Acinas, Bea Díaz, Albert Calbét, Rafel

Simón, Thomas Leffort, Pep Gasol, Celia Marrasé – for your company, conversations, smiles, and for

making the ICM so nice place to come to each day.

To colleagues that shares with me thee unforgettable cruises: RODA1, ATOS-Ártico y ATOS-

Antártida, for making of this cruises so extraordinary experience, and especially to a dream team:

Antonio Tovar and Txetxu Arrieta, for their friendship.

To Markus Weinbauer, for all very thoughtful comments, motivation, and beer (I owe you

one…). Thanks for the possibility of staying in your lab, and for sharing some of your ideas with me.

To Angel Navarro, for being so special person and so extraordinary friend.

Special thanks to Rodrigo Almeda – just for everything. For your friendship, for your support

in so many situations, your honesty, and for your help in all small and big details. I feel proud of

being your friend. We started this road together from the first hour at the University of Barcelona,

and I’m very happy to finish it next to you. Now it’s your turn. Go for it!

To my Family, for help and support. To my Mother, for believing in me more than myself

and for sharing her wisdom with me; my Sister for being my biggest fan; my Daughter for being my

biggest motivation and my sunshine; and Israel – for being there always, for your help in a variety

of ways, for warm mails in cold nights, and – I have to be honest – for your patience.

Mojej Rodzinie, za pomoc i wsparcie: Mojej Mamie, za to że wierzy we mnie mocniej niż ja

sama i że dzieli ze mną swoją mądrość; mojej Siostrze za to że jest moim najwirkszym fanem;

mojemu Ojcu za zaszczepienie we mnie milości do nauki; mojej Córce za to że wnosi motywację i

światlo w moje życie, i że okazuje mi swoją milość każdego dnia.

A Israel – por estar siempre, por tu ayuda en muchísimas cosas, por los mails cálidos en las

noches muy frías, y – tengo que admitirlo – por tu paciencia.

Gràcies, Gracias, Thanks, Dziękuję!!!

Frequently used abbreviations

P a g e | 11

List of abbreviations

a – specific net growth rate

AE – anticyclonic eddy stations

AW- Atlantic Waters

BA – bacterial abundance

BP – bacterial production

BS – burst size

BSS – bacterial standing stock

Chl a – chlorophyll a

CE – cyclonic eddy stations

DFM – deep fluorescence maximum

DOC – dissolved organic carbon

DOM – dissolved organic matter

FF – far field stations

FLB – fluorescently labeled bacteria

G – grazing rate

g – specific grazing rate

%HNA – percentage of high nucleic acid content bacteria

HNF – heterotrophic nanoflagellates

OC – organic carbon

OM – organic matter

PMM – protist-mediated mortality of bacteria

PNF – phototrophic nanoflagellates

POC – particulate organic carbon

POM – particulate organic matter

PSW - Polar Surface Waters

PSWm - Polar Surface Waters from melted ice mixed with Atlantic Waters

RLC – rate of lysed cells

TEM – transmission electron microscopy

VA – viral abundance

VBR – virus-bacterium ratio

VDA – virus dilution approach

VDR – viral decay rate

VMM – virus-mediated mortality of bacteria

Boras J. A., Impact of viruses on bacterial communities

12 | P a g e

VP – viral production

VPL – lytic viral production

VPLG – lisogenic viral production

VRA – virus reduction approach

Introduction

Introduction

P a g e | 15

The world was a home for me,

in all its corners I was looking for the knowledge.

Beneath diverse skies I liked to wander taking possession of each shade of thinking,

like varied fruits that supplied with the nourishment my philosophy.

Muhammad Hafez Sherazi

1. INTRODUCTION

iruses are a numerous and active element in all marine ecosystems, from coastal

waters, through open sea, and to hydrothermal vents and the dark deep ocean

layers. With concentrations of approximately 107 viruses per milliliter of seawater and 109 viruses per

gram of soil or sediment (Wommack & Colwell 2000), they gain the status of the most abundant

biological entities in the biosphere. Assuming that each virus contains roughly 10-17 grams of

carbon, and a total number of viral particles in the ocean of 1030 (Suttle 2005), it can be

calculated that they constitute the ocean second biomass (ca. 200 Mt) after bacteria (Hambly &

Suttle 2005). Viruses also represent the largest reservoir of genetic diversity in the ocean (Rohwer

2003).

Since viral abundance usually follows host abundance, viruses of bacteria and

cyanobacteria, called bacteriophages, are believed to be the most abundant kind of viruses in

marine systems. Heterotrophic bacteria, in turn, are an essential part of the food webs in aquatic

environments, being the major consumers of dissolved organic matter (DOM) in the ocean (Azam

1998), and represent a large fraction of the global genetic diversity (Hammond 1995). Changes in

the productivity and community structure of bacteria lead to modifications of carbon metabolism

and its fluxes in the ocean. Among a variety of factors that can modify these processes,

bacteriophages are one of the most important ones, and play a key role within the microbial food

webs. Regarding the great diversity of bacterial viruses and often complicated character of

interactions among phages and their hosts, as well as relationships between bacteria and other

trophic levels, there is still a significant lack of knowledge in many aspects of marine viral ecology,

in spite of intense studies on these topics.

This work aims to add new insights to the existent pool of information about marine

bacteriophage-host relationship and intents to point some consequences of these new data for

V


16 | P a g e

the marine systems, as well as try to clarify the role that bacteriophages play within the microbial

food webs in marine systems.

1.1 THESIS OUTLINE

The memory of this doctoral thesis consists of five central chapters in scientific-article

format, an introduction and a general results’ discussion. The aim of this introduction chapter is to

present the current state of art of the viral aquatic ecology with emphasis on marine environments,

essential to the understanding of further presented topics,. General discussion intents to give a

global review of the obtained results, compare with previous studies and indicate the meaning of

these results for the understanding of microbial processes in the sea.

Five papers, on which the thesis is based, will be referred to along the whole memory by their

roman numerals.

1.2. INTRODUCTION TO VIROLOGY

Viruses are one of the most powerful and influencing biological entities on earth. For some

scientists they are clearly nonliving particles (Kirchman 2008), while for others they represent the

simplest living creatures (Fields et al. 2001). It is believed that all living organisms have their viruses.

Considering, that a given organism often is susceptible to more than one type of virus it is likely that

there are more kinds of viruses than kinds of hosts (Munn 2006). Furthermore, a virus can have his

own infecting virus as well (Pearson 2008). This would give to viruses a status of the most diverse

biological component in the biosphere. However, only a few thousand viruses have been

described till now.

Beginnings of virology.

The existence of infectious agents that would be able to cause specific diseases, and too

small to be observed with the light microscope was proposed as early as in 1840 by the German

anatomist J. Henle of Gottingen (Fields et al. 2001). Yet, in absence of direct evidence the

existence of viruses was not confirmed till 1892, when the Russian scientist D. Ivanofsky first proved

and described the existence of ultramicroscopic infectious agents. He demonstrated that a

tobacco disease was caused by particles that were not stopped by the Chamberland filter

(Ivanofsky 1892). Finally, in 1898, M. Beijerinck, a Dutch soil microbiologist, independently from

Ivanofsky discovered that the agent causing a tobacco disease could reproduce itself, but only in

living tissue (Beijenrinck 1898). He called this infectious agent a “contagium vivum fluidum”.

Introduction

P a g e | 17

In 1915 bacteriophages were first described by F. W. Twort (1915). Twort observed that

some of the colonies of the bacterium Staphylococcus aureus in his lab became more

transparent, and were constituted of dead bacteria. In parallel, a Canadian bacteriologist, F.

d’Herelle, observed clear circular spots (“plaques”) on the colonies of Shigella bacterium growing

on Petri dish (d’Herelle 1917). He called the infectious agent of bacteria the “bacteriophage”

(from Greek: phage = eater). D’Herelle had a great input into the knowledge, showing the

adsorption of phages to hosts as a first step of infection and the host specificity of viruses, and

describing in modern terms the cell lysis and the release of infectious viruses (d’Herelle 1921, 1926).

One decade later a German scientist, M. Schlesinger, showed that viruses of bacteria are

composed of proteins and contain phosphorous and deoxyribonucleic acid (Schlesinger 1934).

After a number of studies, the viral protein subunits were dissociated, and a viral particle was

reconstructed from its ribonucleic acid (RNA) and protein subunits to produce again an infectious

virus by Fraenkel-Conrat and Williams (1955). Those authors showed also that the nucleic acid from

virus injected to the host has infectious properties, is more unstable, however, and much less

effective as an infection factor than the untouched virus (Fraenkel-Conrat et al. 1957).

Originally from these discoveries, bacterial viruses served as a model for understanding life

processes for generations of scientists.

Phage structure.

Bacteriophages (phages), viruses that infect Bacteria, are the largest viral group in the

nature, with over 5000 types of phages examined and described (Ackermann 2001).

Phages are tiny particles with dimensions between 30 nm and 60 nm, although smaller and

bigger phages can also be found (Weinbauer 2004). All known viruses have the same basic

structure: they are segments of nucleic acid, deoxyribonucleic (DNA) or ribonucleic (RNA),

enclosed generally in a protein coat. Nucleic acid can be linear or circular, single-stranded (ss) or

double-stranded (ds). The protein sheath is called capsid, and is composed of one to a few

different types of protein molecules (Fig. 1.1).

Fig. 1.1. Scheme of viral structure on the example of T4 bacteriophage.


18 | P a g e

Bacteriophages are diverse both structurally and functionally. According to the symmetry

of a viral particle, phages can be divided into tailed, cubic, filamentous and pleomorphic (Fig.

1.2). Tailed phages constitute ca. 96 % among all bacterial viruses (Ackermann 2006). Many of

bacteriophages are large and complex, with relatively large content of nucleic acid and proteins.

Most contain dsDNA, but there are small phage groups with ssDNA, ssRNA or dsRNA. Some virions

(extracelular phase of virus) have a lipid-containing envelope (pleomorphic phages, which do not

form capsids, and some filamentous phages), or contain lipids as part of the particle wall (some

cubic phages).

Fig. 1.2. Types of bacteriophages based on viral particle symmetry: tailed (a), cubic (b), filamentous (c) and

pleomorphic (d) phages.

Viral infection.

Viruses have no metabolism and are unable to replicate by their own. To create a new

progeny, they must enter a host cell and use the host subcellular machinery and energy supplies.

Viral infection goes according to the general scheme: (1) binding of viral attachment protein to

specific receptors on the host cell surface; (2) penetration of the cellular membrane and injection

of the viral genome into the host cell cytoplasm; (3) uncoating, i.e. disassembly of viral

components; this step begins immediately after genome injection and leads to genome

replication; (4) replication of viral genetic information; (5) virus assembly; (6) cellular lysis and new

progeny release (Fig. 1.3; Campbell 2001).

Although this general scheme is common for all viruses, viral infection can go by different

paths. In terms of the new viral progeny viral infection can be classified into productive (lytic and

chronic), reductive (pseudolysogenic and lysogenic), and destructive (restrictive and abortive)

(Abedon 2008). Table 1.1 provides a short description of all these processes.

Introduction

P a g e | 19

Table 1.1. Types of phage infection (modified from Abedon (2008)).

Category Type Description

Productive

lytic

Replication of the prophage genome starts rapidly after viral

nucleic acid injection. Release of mature phages kills the

host cell.

chronic

Infection results in release of produced virions without

immediate destruction of the host.

Reductive

lysogenic

Infection results in indefinite persistence of a temperate

phage in the host cell. Its presence does not induce cell

death or the production of phage particles.

pseudolysogenic

After infection neither lysogeny, nor productive phage

replication is initiated, although viral genome remains

present in the host cell.

Destructive

restrictive

Involves loss of phage, but not of bacterial viability. Bacteria

are “immune” to infecting phages thank to blocking of

genome infection (exclusion), destruction of phage genome

within bacterial cytoplasm (restriction), blocking of the

replication superinfecting phage genomes by prophage

(immunity) or prophage eradication (curing).

abortive

Infection results in loss of both, phage and bacterial viability.

Viral replication strategies

Bacteriophages can potentially display three different replication strategies, according to

the type of infection: productive (lytic or chronic), lysogenic or pseudolysogenic (Fig. 1.3).

Lytic replication strategy leads to the death of infected bacterium shortly after the

injection of the genome of an obligate lytic phage (a phage that is not temperate or is not

capable of displaying a lysogenic cycle in any of its mutations; Barksdale & Arden 1974). Shortly

after infection, the phage genome redirects host metabolism towards the production of new


20 | P a g e

phages. The synthesis of viral components is followed by the release of mature phages through a

burst of the bacterial cell. Some studies suggest that lytic infection could predominate in natural

bacterial communities (Wilcox & Fuhrman 1994). Moebus (1983) calculated that 65 % of the

phages he co-isolated from the Atlantic Ocean were lytic.

Fig. 1.3. Types of viral replication strategies. Model modified from Weinbauer (2004).

Introduction

P a g e | 21

Chronic replication strategy involves the release of new produced virions, yet not the

immediate destruction of the host (Russel & Model 2006). Bacterial cell remains infected and

produce phages, which liberate by budding or extrusion, without lysing the bacterium. To my best

knowledge there is only one report on a potential chronic infection in situ (Hofer & Sommaruga

2001).

In lysogeny, long-term relationship between temperate phage and host is established.

Phage genome forms part of the host genome. At this step the phage is called prophage, and the

bacterium that carries a prophage – lysogenic bacterium. Prophage replication is coordinated

with host genome replication, and progeny genomes are distributed to each daughter bacterial

cell. Prophage presence in the host cell does not cause the death of the bacterium or the

production of new viruses, and often results in the increase of host fitness. This gives to the

bacterium-phage relationship a symbiotic character. The lytic cycle in prophage can be induced

by a number of factors (e.g. pH, UV light, chemical agents), and ends with the death of the

bacterial cell.

It is believed, that because in marine environments prevail conditions that favour lysogeny,

i.e. low nutrient concentrations, slow-growing bacterial hosts at relative low concentrations,

lysogeny would be the dominant replication strategy in those systems (Miller 2006). Ackerman and

DuBow (1987) estimated that between 21 % and 60 % of environmental bacteria are lysogens.

Field studies, however, give no clear response in this topic. On the one hand, higher percentage of

lysogens was found in offshore than coastal environments (Jiang & Paul 1994, Weinbauer & Suttle

1999) and during periods with low inorganic nitrogen and phosphorous concentrations (Williamson

et al. 2002). On the other hand, Moebus (1983) found that only 10 % of the 300 marine phages that

he examined were truly temperate. Also, some authors report very few inductions of lysogenic

bacteria in oligotrophic offshore waters comparing with the coastal systems (Jiang & Paul 1996),

which introduce some interrogates in the theory of lysogeny occurrence.

Pseudolysogeny is the phage replication strategy, in which viral genome (called

preprophage) is maintained in the host cell for potentially extended periods of time. No

integration of host and phage genomes occurs, and the phage genetic information is not

replicated and segregated equally among daughter bacterial cells. Pseudolysogen can be cured

of the preprophage, e.g. through repeated subcloning (Wommack & Colwell 2000). It is believed,

that the low nutrient concentrations decide on the pseudolysogenic infection decision (Ripp &

Miller 1997). Pseudolysogeny can be a common phenomenon among bacteriophages (Ackerman

& DuBow 1987), also in marine ecosystems (Moebus 1996), although so far only few aquatic

pseudolysogenic phages have been described (e.g. Moebus 1997).


22 | P a g e

One of the goals of this thesis was to assess the relative frequency of

lysogenic and lytic infection in systems with different trophy, as coastal

waters of the Mediterranean Sea, subtropical Atlantic Ocean, and the

Arctic Ocean (Papers I, IV and V).

1.3. BACTERIOPHAGES AS MODIFIERS OF BACTERIAL COMMUNITIES.

Presence of viruses in marine ecosystems was confirmed as early as in 1955 (Spencer 1955),

however, due to the low detected concentrations they were not considered an important part of

microbial food webs. In 1989, high abundances of phages in sea water, often exceeding by one

to two orders of magnitude bacterial concentrations, were reported by Bergh and colleagues

(1989). Since then the importance of phages in marine environments, and their impact on

ecosystems have been widely studied. Today it is established, that marine viruses are an important

component of microbial food webs as they influence bacterial abundances, diversity, and nutrient

cycling within the ecosystem.

1.3.1. BACTERIAL MORTALITY

Viral lysis, along with grazing by protists, is an important source of mortality in aquatic

bacterial communities. It is believed that phages daily remove about 10 % to 40 % of the standing

stock of bacteria in marine environments (Thingstad et al. 2008). Losses from undetectable (e.g.

Guixa-Boixereu et al. 1996) to over 100 % (e.g. Guixa-Boixereu et al. 2002) of bacterial production

were found in different systems. In different ecosystems, bacteriophages have been found to be

the main factor of bacterial losses (Guixa-Boixereu et al. 1999a, Wells & Deming 2006), to have

similar significance (Fuhrman & Noble 1995, Hwang & Cho 2002), or even to be less significant than

protists (Guixa-Boixereu et al. 1996, Choi et al. 2003). However, although a high number of studies

that evaluate virus mediated mortality, the comparison of results obtained by different authors is

difficult due to the great variety of methods used. Furthermore, even studies where the same

methods were applied, show high variety of results, depending on location, season, physico-

chemical and biological factors (Table 1.2).

Introduction

P a g e | 23

Table 1.2. Ranges of virus-mediated mortality (VMM) of bacteria in various marine environments assessed using

viral dilution / reduction approach, the same as used along this thesis. BG – bacterial growth, BSS – bacterial

standing stock, BP – bacterial production.

Location VMM Reference

Ria de Aveira, Portugal (estuary)

brackish

marine

30 – 41 % BG

49 – 74 % BG

Almeida et al. 2001

Strait of Georgia, Canada

tidally mixed

non-mixed

46 – 137 % BSS h-1

19 – 27 % BSS h-1

Wilhelm et al. 2002

Tropical Atlantic Ocean

Southern North Sea

1 - 7 % BSS h-1

11 % BSS h-1

Winter et al. 2004

North Pacific Subtropical Gyre,

Stn ALOHA

5 m

25 m

45 m

75 m

3.2 – 5.9 % BSS h-1

4.7 – 8.8 % BSS h-1

3.4 – 6.4 % BSS h-1

8.8 – 16.5 % BSS h-1

Brum 2005

Chesapeake Bay, USA

Delaware Bay, USA

20 – 38 % BP d-1

1.5 – 18 % BP d-1

Helton et al. 2005

Chesapeake Bay, USA

Delaware Bay, USA

84 – 207 % BP

105 % BP

Winget et al. 2005

It is believed that the impact of phages on bacterial abundances varies with the system

trophy. While in eutrophic systems viruses would be an important cause of bacterial losses, protists

would be a main factor of bacterial mortality in oligotrophic waters. This is based on two

arguments. First, mathematical models predict that contact rates between viruses and hosts

increase at high bacterial abundances (Murray & Jackson 1992), enhancing the infection rate and

prokaryotes losses due to viral lysis. The second argument is based on the fact that phages infect

preferentially more active and thus more productive hosts (Lenski 1988). According to that, many

studies have found viruses as an important cause of prokaryotic mortality in eutrophic systems

(Weinbauer & Suttle 1999, Noble & Fuhrman 2000), and protists as a principal cause of bacterial

losses in oligotrophic waters (Guixa-Boixereu et al. 1996; Bettarel et al. 2002). Other authors,


24 | P a g e

however, reported higher proportion of lysed than grazed bacteria in poor-nutrient environments

(Wells & Deming 2006).

One of the most important effects of lysis of bacterial cells due to viral infections is an input

of organic carbon, nutrients and trace elements to the water. It was estimated that carbon

remobilization rates in the thermally stabilized water column could supply ca. 5 % to 7 % of the

bacterial carbon demand in oligotrophic offshore, 30 % in mesotrophic nearshore waters of the

Gulf of Mexico, and ca. 80 % to 95 % in the mesotrophic Strait of Georgia (Wilhelm & Suttle 2000),

and up to 180 % in the eutrophic Baltic Sea (Riemann et al. 2009, Holmfeldt 2009). In well vertically

mixed waters, these values reached from 140 % to >1000 % of the bacterial carbon demand

(Wilhelm & Suttle 2000). This indicates that viral lysis may be one of the key mechanisms providing

DOM in some marine systems.

Bacterial losses due to phages were evaluated in different marine systems,

and compared with bacterial losses due to nanoprotists. Results of these

studies are presented in Papers I, IV and V.

1.3.2. BACTERIAL DIVERSITY.

Bacteriophages can modify bacterial species and genetic diversity in four main ways (Fig.

1.4): (1) Differentially killing bacteria that win the competition for resources, the “killing the winner”

hypothesis; (2) Releasing lysis products from the cell to the environment; (3) Changing bacterial

community structure thanks to lysogenic infections; and (4) Mediating the genetic exchange

among bacterial strains and even species.

The “killing the winner” hypothesis (Thingstad & Lignell 1997, Thingstad 2000) suggests that

the most abundant bacterial taxa in the ecosystem will be the most affected by viral infections

and thus less abundant taxa will benefit from the competitors elimination and/or increase of

resource availability and will increase their concentration. This hypothesis is based on two main

assumptions: that the growth of bacteria is limited by nutrients, and that bacterial abundance is

controlled by protists predation. Furthermore, it bases on the mathematical model, according to

which the contact rate between the host and its phage increases with increasing bacterial

concentration. Thus, the probability of infection is higher for the successful, more abundant

bacterial strains (Murray & Jackson 1992). According to this hypothesis, viruses will affect neither

growth rate, nor abundance of the steady state bacterial community, but they will control

bacterial diversity. Phages will sustain species richness, preventing any single bacterial species from

Introduction

P a g e | 25

dominating the community, and allowing the existence and growth of less competitive bacterial

species.

Although the “killing the winner” hypothesis seems to be correct for a variety of

ecosystems and organisms involved (e.g. Bohannan & Lenski 1999, Bouvier & del Giorgio 2007),

other theoretical models have shown a different impact of phages on bacterial diversity. Model

presented by Miki and Yamamura (2005) suggests that because viral infections eliminate bacterial

species with a higher affinity for nutrients, the existence of viruses leads to less effective use of

nutrients throughout the bacterial community, instead of higher species richness. In another

model, unlike that in “killing the winner”, authors assume that not all bacterial groups compete for

a single resource (Miki et al. 2008). They argue that if different bacteria use different carbon

sources, no “winner” of competition for resources exists. According to this model, phages favour

the abundance of infection-resistant bacteria, which have the same grazers as infection-

susceptible bacteria. By this, viruses indirectly regulate bacterial consumption of different kinds of

organic carbon. Authors conclude, that phages “do not necessarily promote the coexistence of

bacteria, but they can facilitate the remineralization of diverse organic carbon and suppress the

accumulation of specific organic carbon fractions”. They assume, however, that viruses may

regulate bacterial diversity in systems with high diversity of organic carbon, e.g. in lakes or coastal

ecosystems.

Fig. 1.4. Potential mechanisms of phage impact on prokaryotic diversity in aquatic systems.


26 | P a g e

The “killing the winner” hypothesis was supported by a number of in situ findings.

Oscillations of different phages and their hosts’ abundances with a typical predator-prey pattern,

with relatively constant total abundances of bacteria and viruses were found in Chesapeake Bay

(Wommack et al. 1999 a, 1999b). Presence of viruses was found to change community

composition of bacteria (Weinbauer & Höfle 1998, Auguet et al. 2009) and phytoplankton (Suttle

1992). Although these and other data support the “killing the winner” hypothesis, some

experimental bacterial communities often show a negative responses on viral presence and

activity (e.g. Hewson & Fuhrman 2006), which suggests, that the impact of viruses on

bacterioplankton diversity is a very complex process, which may depend on a variety of factors.

The second way in which viruses can affect bacterial diversity is through the release of

organic material from bacteria during cell lysis (Weinbauer & Rassoulzadegan 2004). This may

change the bioavailability of organic nutrients and lead to changes in bacterial community

structure. Indeed, changes in the DOM composition were found to modify bacterial community

composition (Arrieta & Herndl 2002, Jones et al. 2009).

Through lysogenic infection by some temperate phages and expression of its genes, the

host’s phenotype is modified by the addition or loss of some characteristics. This process is termed

as lysogenic conversion (Miller & Day 2008). These new phenotypes acquired can include an

acquisition of components of metabolic systems, such as the photosynthetic complex (Mann et al.

2003), the ability to produce bacteriocins (Ivanovics et al. 1976), antibiotics (Martinez-Molina &

Olivares 1979), or pili (Karaolis et al. 1999). One of the most spectacular examples of lysogenic

conversion is the case of the bacterium Vibrio cholera. These bacteria cause cholera only when

cells are infected with a temperate phage, which carries encoded in its genome the cholera-

causing toxin (Waldor & Mekalanos 1996). Lysogenic bacteria also gain protection from

superinfection by the virions of the same immunity type as the resident prophage (Campbell 2006),

or even from infections by other kind of bacteriophages (Saye & Miller 1989). Furthermore, lysogens

have often better fitness than uninfected cells (Lin et al. 1977). All these changes can lead to the

better adaptation to environmental conditions, or even occupation of new niches (Frye et al.

2005). Lysogeny was thus suggested to shape bacterial diversity by allowing survival and / or

dominance of lysogens in the environment (Edlin et al. 1975).

Bacterial genetic diversity may be also increased by horizontal gene transfer by

transduction. Transduction consists on the transfer of genetic material from bacterium to bacterium

by phage particles. High rate of transduction was found in some bacterial strains (e.g. Escherichia

coli, Hayashi et al. 2001), and prophages were found to be an integral part of some bacterial

genomes (Ohnishi et al. 2001). These facts suggest that phages played a predominant role in the

emergence of various bacterial strains. It was demonstrated that transduction occurs also in

Introduction

P a g e | 27

freshwater and marine systems (Saye et al. 1987, Jiang & Paul 1998). Theoretical estimations

showed that there might be about 1014 transductants per year in Tampa Bay, Florida (Jiang & Paul

1998), and about 1013 transductants per year in the Mediterranean basin (Weinbauer &

Rassoulzadegan 2004). However, no empirical data are available by now.

Besides presented above, phages can affect bacterial community structure also by other

ways. Bacterial DNA released during viral lysis might stimulate DNA transfer by transformation, i.e.

by uptake of free DNA by bacteria (Weinbauer & Rassoulzadegan 2004). Moreover, phage activity

can change the clonal composition of the host by induction of resistance mechanisms

(Middelboe et al. 2001). The variety of known strategies with which phages shape host diversity

suggest that virus-mediated gene transfer may play a crucial role in biogeochemical and

ecological processes in the ocean, and that there is a need of studies which would link those fields

(Weinbauer & Rassoulzadegan 2004).

Effect of viral activity on bacterial community composition in the coastal

waters of the Mediterranean Sea was evaluated in situ and in experimental

conditions (Papers II and III).

1.4. BACTERIOPHAGES IN THE MICROBIAL LOOP

Since the early 1930s scientists were aware that a variety of bacteria exist in the ocean,

which have their importance for biogeochemical cycles in marine ecosystem (Sherr and Sherr

2008). However their basic role for pelagic ecosystems was not proved and described till 1974,

when L. Pomeroy gave a central role to heterotrophic bacteria and their protists grazers in the

material flow through food webs (Pomeroy 1974). This induced an intense investigation in this topic

which led to the “microbial loop” idea (Azam et al. 1983, Ducklow 1983), which nowadays is a well

established paradigm of aquatic food webs functioning.

The microbial loop is a fundamental part of all aquatic food webs, providing the

regeneration of macronutrients, particularly nitrogen and phosphorus, which allow further

phytoplankton growth and flow of the organic matter to higher trophic levels (Fig. 1.5). It bases on

the ingestion of dissolved products of primary production by bacteria, returning to the biomass the

organic matter that otherwise would be lost for larger grazers (Azam et al. 1983). The essential

component of the marine microbial loop are heterotrophic bacteria, as the major consumers of

DOM in the ocean (Azam 1998), and herbivorous protists as the major consumers of primary

production in marine systems. Small heterotrophic protists (mainly flagellates) shunt therefore


28 | P a g e

carbon from bacteria and small phytoplankton (<5 μm) to larger protists (mainly ciliates), and

finally to zooplankton. Bacterioplankton incorporate from 10 % to 50 % of primary production

(Azam et al. 1983). On the other hand, the carbon and nutrients contained in bacterial and

phytoplankton cells can be returned to the environmental DOM pool by their respective viruses.

Fig. 1.5. Microbial loop including the role of viruses. Blue arrows show the flux of carbon within the system, red

arrows show the viral loop. Model adapted from Wommack and Colwell (2000) with some modifications.

Viral lysis of photosynthetic organisms, algae and cyanobacteria, transfers about 6 % to 26

% of the photosyntetically fixed carbon into the DOM (Wilhelm & Suttle 1999). Through bacterial

cell lysis, phages convert organic matter contained within the cells into a dissolved form, available

again only to heterotrophic bacteria, which take up nutrients directly across their cell membranes.

Introduction

P a g e | 29

This process, called “viral loop” (Fig. 1.5; Bratbak et al. 1992), exclude the higher trophic levels from

the utilization of this carbon pool. As conversion of bacterial biomass in DOM is a very efficient

process, phage lysis represents thus a source of nutrient-rich substrate for bacterial production

(Bratbak et al. 1990, Middelboe & Jørgensen 2006). Furthermore, phage-mediated DOM release

results in an increase of respiratory loss, but also in a substantial increase of bacterial production

(Fuhrman 1999).

Direct result of the activity of viruses is the removing of the potential protists’ prey from the

environment. Models show that the presence of phages always decreases the carbon transport to

protists (Miki et al. 2008), which has a profound impact on the whole trophic web. The theoretical

assessment of the impact of phages on the trophic net showed that with 50 % of bacterial

mortality due to viral lysis, the bacterial production due to new available DOM increased by 27 %,

and an export of DOM to nanoplankton grazers decreased by 37 %, compared with the food web

without viral lysis. This results in a net loss of 25 % in nanoprotists production (Fuhrman 1992). Similar

calculations, which included also viral infection of phytoplankton (10 % of mortality) and loss of

viruses due to consumption by nanoprotists (13 % of viral production) resulted in a 33 % increase in

bacterial production and respiration, and a 20 % reduction in nanoprotists production (Fuhrman &

Suttle 1993). It is believed that bacteriophage infection have the greatest impact on

mesozooplankton production (organisms with size of 20 – 200 μm; Murray & Eldridge 1994) in

oligotrophic systems, where recycling of organic matter predominates (Murray & Eldridge 1994). In

those environments biomass consist primarily of bacteria (Fuhrman et al. 1989), and bacterial

production accounts for 15 % to 25 % of mesozooplankton nutrition. In turn in mesotrophic systems

greater proportion of primary producers is directly available to mesozooplankton. In those

environments bacterial losses due to phages were predicted to cause minor reduction (1.2 % - 7.4

%) in mesozooplankton production (Murray & Eldridge 1994).

The process of viral lysis also allows the retention of nutrients in the euphotic zone of

aquatic systems, which can be important especially in oligotrophic environments, where it could

partially avoid nutrients’ sinking to the deep sea (Bratbak et al. 1990, Fuhrman 1999). Due to their

substantial impact on natural ecosystems, viruses were included to the classic microbial loop (Fig.

1.5; Fuhrman 1992, Wommack & Colwell 2000), as an integrate and basic element.

The input of nutrients to the water column due to viral lysis of bacterial cells

was assessed for different marine systems, and compared with the flux of

carbon up in the trophic chain mediated by nanozooplankton (Papers I, IV

and V).

Hypotheses and objectives

Objectives

P a g e | 33

2. HYPOTHESES and OBJECTIVES

his thesis aimed to make a contribution to the knowledge in the relatively new area of

viral marine ecology. Its main purpose was to determine the effects of bacteriphage

activity on bacterioplancton (abundance, losses and diversity) comparing with protists activity.

Factors influencing processes of viral infection of bacteria and protistan grazing were evaluated.

Furthermore, the type of interactions among phages and protists was determined, and the effect

of the coexistence of those two predators on bacterial communities.

Hypotheses of the thesis

Based on the information presented in the “Introduction”, this thesis is aimed to test following

hypotheses:

In oligotrophic systems losses of bacteria due to protists should be higher than due to

viruses. In contrast, in eutrophic systems bacterial losses due to viruses would be similar or

more important than due to protists. Thus, viral infections could modify the carbon fluxes

channelled in marine systems.

Lysogeny might be the dominant type of viral replication strategy in oligotrophic systems,

and lytic replication strategy should predominate in eutrophic systems.

Phages should be important players in shaping bacterial community composition, and

protists, through the establishment of synergistic or antagonistic interactions with viruses,

could have an influence on bacterial diversity.

T

Boras J. A, Impact of viruses on bacterial communities

34 | P a g e

Objectives of the thesis

The main objectives of this thesis were:

1. Estimation of the range of bacterial mortality caused by viral infections in different marine

systems, and :

comparison with bacterial losses caused by protistan grazing (Papers I, IV

and V),

determination of the character of interactions between phages and

protists, and the effect of these interactions on bacterial losses due to

phages (Paper III),

evaluation of the range of lysogenic viral infections across the systems

(Papers I, IV, and V)

assessment of the input of organic carbon to the environment mediated

by viral lysis (Papers I, IV and V).

2. Investigation of the effect of the system trophic status on the rate of viral infections (Paper

IV).

3. Determination of the consequences of viral and protistan activity for the bacterial

community structure (Paper II), and:

evaluation of the effect of possible interactions between phages and

protists on bacterial diversity (Paper III).

Report of the thesis directors


P a g e | 37

3. REPORT OF THE THESIS DIRECTORS

on the impact factor of each journal for the published

articles and implication of the candidate

Dr. Dolors Vaqué, as the director, and Dr. Maria Montserrat Sala, as the co-director of the

PhD thesis entitled: “Impact of viruses on bacterial communities in marine systems” developed by

the PhD candidate Julia A. Boras,

Report the implication of the PhD candidate in each scientific article included in the

current thesis, and the impact factor of each journal for the published articles.

Article I

“Annual changes of bacterial mortality due to viruses and protists in an oligotrophic coastal

environment (NW Mediterranean)” by J.A. Boras, M.M. Sala, E. Vázquez-Domínguez, M.G.

Weinbauer, and D. Vaqué, published in Environmental Microbiology (2009, 11(5), 1181–1193), with

an impact factor (2008) of 4.707, and located in the first quartile (rank 16 out of 91 journals) in the

category “Microbiology”.

This article describes two annual cycles of bacterial mortality due to protists and viruses,

and its implication in the carbon cycle in an oligotrophic coastal area (Blanes Bay). The sampling

program and the variables needed for this study was discussed together with the PhD student and

the director and co-director. Julia Boras performed determination of viruses, bacteria and protists

(nanoflagellates and ciliates) abundance, as well as bacterivory, viral production (lysis and

lysogeny), and bacterial mortality due to viruses. Dr. E. Vázquez-Domínguez supplied data of

bacterial production by leucine incorporation, and Dr. Weinbauer supervised the viral mortality

calculations. Finally the Net of the Marine Microbial Observatory of Blanes determined chlorophyll

a concentration and physicochemical parameters as temperature, salinity and nutrient

concentration. The PhD student has analyzed the data and elaborated the manuscript under the

supervision and advices of the director, comments of the co-director and the consensus of the


38 | P a g e

other coauthors. The PhD candidate is the first author of this article. The data included in this article

were not used to the preparation of any other doctoral thesis.

Article II

“Role of viruses in shaping bacterial phylogenetic and functional diversity in oligotrophic marine

coastal waters” by J.A. Boras, D. Vaqué, F. Maynou, E.L. Sà, M.G. Weinbauer, and M.M. Sala,

submitted to The ISME Journal, with an impact factor (2008) of 5.029, and located in the first

quartile (rank 10 out of 124 journals) in the category “Ecology”.

This article describes the main factors shaping bacterial community composition and

bacterial metabolism along two annual cycles in Blanes Bay. The design of the study was planned

together with the PhD student, the director and co-director. The PhD student determined bacterial

diversity through a DGGE (denaturing gel gradient electrophoresis) finger printing technique with

help of the technician E.L. Sà. Bacterial enzymatic activity and functional diversity was determined

by Dr. M.M. Sala. To obtain which were the main factors affecting bacterial phylogenetic and

functional diversity, the PhD student used the data set of bacterial mortality due to protist and

viruses, as well as other parameters presented in Article I. She applied the CCA statistical package

under the supervision of Dr. F. Maynou. The PhD student has analyzed the results and has

elaborated the manuscript, under the supervision of the director, and having thoughtful discussions

with the co-director specially related with the bacterial functional diversity as well as the

agreement of the rest of coauthors. The PhD candidate is the first author of this article. The data

included in this article were not used to the preparation of any other doctoral thesis.

Article III

“Changes of bacterial mortality and diversity as a response to viral and nanoflagellate

communities interactions in an oligotrophic marine coastal site” by J.A. Boras, E. Vázquez-

Domínguez, E.L. Sà, M.M. Sala, and D. Vaqué, submitted to Environmental Microbiology, with an

impact factor (2008) of 4.707, and located in the first quartile (rank 16 out of 91 journals) in the

category “Microbiology”.

This study, conducted with water samples from Blanes Bay, shows mostly synergistic

interactions between viruses and protists, and describes the effect of those interactions on

bacterial losses and diversity. The microcosms experimental design was planned and performed

together with the PhD student, director, co-director, and other members of the research group.

The PhD student determined viral and microbial abundances, viral production, virus lytic and

lysogenic infection on bacteria, and bacterial diversity. Dr. E. Vázquez-Domínguez carried out

bacterial production measurements. The PhD student has analyzed all data and has elaborated

the manuscript after consultations with the coauthors and under the supervision and advices of


P a g e | 39

director and co-director. The PhD candidate is the first author of this article. The data included in

this article were not used to the preparation of any other doctoral thesis.

Article IV

“Effect of viruses and protists on bacteria in eddies of the Canary Current region (subtropical

Northeast Atlantic)” by J.A. Boras, M.M. Sala, F. Baltar, J. Arístegui, C.M. Duarte, and D. Vaqué,

accepted to be published in Limnology and Oceanography, with an impact factor (2008) of 3.663

and located in the first quartile (rank 2 out of 50 journals) in the category “Oceanography”.

This study explores how oceanic eddies interact with biological activities. This study shows

that viral lysis could be a dominant pathway for the flow of bacterial carbon in the NE Subtropical

Atlantic, particularly significant in areas dominated by eddy activity. The study was framed within

the RODA project (PIs Dr. J.Arístegui and C.M. Duarte) and accomplished on board of the R/V BIO-

Hespérides. The design of the investigation and the choice of sampling depths (from epipelagic to

bathypelagic layers) was planned and discussed together with the PhD student, the director and

co-director. Julia Boras run all experiments on board to measure bacterial mortality by viruses and

protists, as well as collection of samples for viral and microbial (nanoflagellates and ciliates)

abundance, that were lately quantified in the Institute of Marine Sciences, Barcelona (ICM - CSIC).

Other variables, as physicochemical parameters (temperature, salinity fluorescence and nutrients

concentration), or part of bacterial abundance measurements were performed by the other

coauthors and members of the Unitat de Tècnologia Marina (UTM – CSIC). The PhD student has

analyzed the data and elaborated the manuscript under the supervision and guidance of the

director, comments of the co-director, and the agreement with other coauthors. Dr. J. Arístegui

and Dr. C.M. Duarte added thoughtful discussions specially related with the hydrography of the

area and on the nutrient fluxes in the ocean. The PhD candidate is the first author of this article. The

data included in this article were not used to the preparation of any other doctoral thesis.

Article V.

“Effect of ice melting on bacterial carbon fluxes channeled by viruses and protists in the Arctic

Ocean” by J.A. Boras, M.M. Sala, J.M. Arrieta, E.L. Sá, J. Felipe, C.M. Duarte, and D. Vaqué,

submitted to Polar Biology, with an impact factor (2008) of 1.515 and located in the second

quartile (rank 12 out of 28 journals) in the category “Biodiversity conservation”.

This study investigates the effect of ice melting in the Arctic Ocean on carbon fluxes from

bacteria to protists and, consequently, to higher trophic levels, and back to the water column as

dissolved and particular organic matter due to lysis by viruses. The study was carried out in the

frame of the project ATOS (PI Dr. C.M. Duarte) and accomplished on board of the R/V BIO-

Hespérides. The investigation and choice of the sampling stations and depths were planned

together with the PhD student, the director and co-director. All experiments to determine


40 | P a g e

bacterivory and viral production, and lytic and lysogenic infections on bacteria were run on board

by the PhD student, as well as part of bacterial abundance measurements determined by Dr. J. M.

Arrieta. Microscopy and flow cytometery counts of viral, bacterial and protistan abundances were

determined in the lCM by the PhD student. The physicochemical parameters (temperature, salinity

fluorescence and nutrients concentration) were determined by the other coauthors and members

of the Unitat de Tècnologia Marina (UTM – CSIC). Finally, Julia Boras has analyzed the data and

elaborated the manuscript under the supervision and advices of the director, comments of the co-

director as well as helpful discussion of the other coauthors. The PhD candidate is the first author of

this article. The data included in this article were not used to the preparation of any other doctoral

thesis.

Publications

Paper I

Annual changes of bacterial

mortality due to viruses and protists

in an oligotrophic coastal

environment (NW Mediterranean)

Summary – Paper I

P a g e | 45

Resumen

Cambios anuales en mortalidad bacteriana debida a los virus

y protistas en un sistema oligotrófico costero (Mediterráneo

NO)

Artículo publicado en Environmental Microbiology (2009, 11: 1181-1193, Annual changes of

bacterial mortality due to viruses and protists in an oligotrophic coastal environment (NW

Mediterranean))

Durante dos años (mayo 2005 – abril 2007) se evaluó mensualmente el papel que juegan

los virus y protistas como principales causantes de la mortalidad del bacterioplancton en un

sistema oligotrófico costero (Bahía de Blanes, Mediterráneo NO). Debido a la oligotrofia del

sistema se esperaba: a) que la mortalidad bacteriana fuera debida principalmente a la actividad

de protistas, y b) que la lisogénia fuera el tipo más importante de infección vírica. Durante el

periodo de estudio, virus y protistas juntos fueron responsables de 50.6 ± 40.1 % de pérdidas de la

abundancia bacteriana (BSS) por día, y de 59.7 ± 44.0 % d-1 de pérdidas de la producción

bacteriana (BP). Durante el primer año (mayo 2005 – abril 2006), los protistas fueron la causa

principal de la mortalidad bacteriana, eliminando 19.9 ± 20.4 % d-1 de BSS y 33.9 ± 24.3 % d-1 de BP,

mientras que la lisis vírica eliminaba el 13.5 ± 17.0 % d-1 de BSS y 12.3 ± 12.3 % d-1 de BP. Durante el

segundo año (mayo 2006 - abril 2007), la mortalidad bacteriana debida a virus fue

significativamente más alta que en el primer año (VMMBSS: F1, 22 = 12.3, p < 0.005; VMMBP: F1, 22 =

23.8, p < 0.001), y comparable con la tasa de mortalidad debida a protistas (virus: 29.2 ± 14.8 % d-1

de BSS y 40.9 ± 20.7 % d-1 de BP; protistas: 28.6 ± 25.5 % d-1 de BSS y 32.4 ± 20.0 % d-1 de BP). La lisis

vírica incrementó con el aumento de la abundancia y producción bacteriana, lo que sugiere que

el estado metabólico del hospedador es importante para la infección y proliferación de los virus.

En el 37% de los casos las pérdidas de BP debidas a virus eran más altas que las debidas a

protistas. En once de los 24 muestreos fue detectada infección lisogénica, y constituía un


46 | P a g e

porcentaje variable de la producción vírica total, desde 0.6 % en noviembre de 2006, hasta 78.4

% en agosto de 2005, coincidiendo respectivamente con la más alta y más baja producción

vírica lítica. Los resultados de este estudio sugieren que la lisogénia puede estar influenciada por

el estado trófico de sistema, ya que se detectaron más casos de lisogénia cuando la

concentración de nutrientes en el sistema era más baja. Sin embargo, contrariamente a lo

esperado, se observó un predominio de la infección lítica sobre la lisogénia durante los dos años

de estudio. El hecho de que la lisis vírica era más alta durante el periodo con concentraciones de

nutrientes más bajas sugiere, que las porinas (estructuras que permiten la captación de nutrientes

por la bacteria) podrían aparecer en la pared celular bacteriana en dichas condiciones de

nutrientes favoreciendo las infecciones víricas. Los resultados de nuestro estudio muestran que,

aunque los protistas juegan un papel destacado en la canalización del flujo de carbono hacia

niveles tróficos superiores, también los virus contribuyen de forma significativa en la mortalidad

bacteriana en la Bahía de Blanes a lo largo del período de estudio.

Annual changes of bacterial mortality due to virusesand protists in an oligotrophic coastal environment(NW Mediterranean)

Julia A. Boras,1* M. Montserrat Sala,1

Evaristo Vázquez-Domínguez,1

Markus G. Weinbauer2,3 and Dolors Vaqué1

1Institut de Ciències del Mar (CSIC), Passeig Marítim de

la Barceloneta 37-49, 08003 Barcelona, Spain.2CNRS, Microbial Ecology and Biogeochemistry Group,

Laboratoire d’Océanographie de Villefranche, 06230

Villefranche-sur-Mer, France.3Université Pierre et Marie Curie-Paris 6, Laboratoire

d’Océanographie de Villefranche, 06230

Villefranche-sur-Mer, France.

Summary

The impact of viruses and protists on bacterioplank-

ton mortality was examined monthly during 2 years

(May 2005–April 2007) in an oligotrophic coastal envi-

ronment (NW Mediterranean Sea). We expected that

in such type of system, (i) bacterial losses would be

caused mainly by protists, and (ii) lysogeny would be

an important type of virus–host interaction. During

the study period, viruses and grazers together were

responsible for 50.6 6 40.1% day-1 of bacterial stand-

ing stock losses (BSS) and 59.7 6 44.0% day-1 of bac-

terial production losses (BP). Over the first year (May

2005–April 2006), protists were the principal cause of

bacterial mortality, removing 29.9 6 20.4% day-1 of

BSS and 33.9 6 24.3% day-1 of BP, whereas viral lysis

removed 13.5 6 17.0% day-1 of BSS and 12.3 6 12.3%

day-1 of BP. During the second year (May 2006–April

2007), viruses caused comparable bacterial losses

(29.2 6 14.8% day-1 of BSS and 40.9 6 20.7% day-1

of BP) to protists (28.6 6 25.5% day-1 of BSS and

32.4 6 20.0% day-1 of BP). In 37% of cases higher

losses of BP due to viruses than due to protists were

found. Lysogenic infection was detected in 11 of 24

samplings. Contrary to our expectations, lytic infec-

tions dominated over the two years, and viruses

resulted to be a significant source of bacterial mor-

tality in this oligotrophic site.

Introduction

During the last decades, marine viruses have been rec-ognized to be an important component of microbial foodwebs. It is known that viral lysis, along with grazing byprotists, can be an important source of mortality in aquaticbacterial communities (Wommack and Colwell, 2000).Dominance of those two mortality processes variesamong ecosystems. In marine habitats, bacteriophageswere found to be the main factor of bacterial losses(Guixa-Boixereu et al., 1999a; Wells and Deming, 2006),to have similar significance (Fuhrman and Noble, 1995;Hwang and Cho, 2002) or to be less significant thanprotists (Guixa-Boixereu et al., 1996; Choi et al., 2003).Viruses transform particulate carbon and nutrientsfrom prokaryotes to dissolved compounds, which can beassimilated again by bacterioplankton, allowing the reten-tion of nutrients in the euphotic zone (Bratbak et al., 1990;Fuhrman, 1999). This process can be important especiallyin oligotrophic systems, where it could partially avoid nutri-ents’ sinking to the deep sea (Fuhrman, 1999). Thus, inorder to understand the functioning of nutrient-poor habi-tats, it is important to evaluate the carbon and nutrientfluxes at the very basic level – from prokaryotes to virusesand to grazers. Moreover, as the proportion of viral infec-tions versus grazing can vary even within the hours(Winter et al., 2004), there is a need of simultaneously runexperiments for relatively realistic prokaryotic loss pro-cesses evaluation. Yet, very few studies based on suchcomparisons have been performed (e.g. Guixa-Boixereuet al., 2002; Wells and Deming, 2006), and most of themwere carried out sporadically or covering short timeperiods.Mathematical models predict that contact rates

between viruses and hosts increase at high bacterialabundances (Murray and Jackson, 1992; Thingstad,2000), increasing the infection rate and prokaryoteslosses due to viral lysis. According to that, many studieshave found viruses as an important cause of prokaryoticmortality in eutrophic systems (Weinbauer and Peduzzi,1995; Weinbauer and Suttle, 1999), and protists as aprincipal cause of bacterial losses in oligotrophic waters(Guixa-Boixereu et al., 1996; Bettarel et al., 2002). Otherauthors, however, reported higher proportion of lysed than

Received 28 August, 2008; accepted 15 November, 2008. *For cor-respondence. E-mail [email protected]; Tel. (+34) 932309500;Fax (+34) 932309555.

Environmental Microbiology (2009) 11(5), 1181–1193 doi:10.1111/j.1462-2920.2008.01849.x

© 2009 The AuthorsJournal compilation © 2009 Society for Applied Microbiology and Blackwell Publishing Ltd

grazed bacteria in cold and nutrient-poor environments(Wells and Deming, 2006). Similarly, bacterial density canalso affect the protists–bacteria interactions, and hencehigher grazing is expected when the encounter probabilityincreases (Peters, 1994; Vaqué et al., 1994). Moreover,the relative dominance of virus- or protists-mediated mor-tality of bacteria can also change in relation with predationon nanoflagellates, enhancing the mortality due toviruses.In addition, system trophy apparently affects also the

life strategy of viruses. Lysogeny seems to predominateover lytic infections in oligotrophic systems, characterizedby low bacterial abundances and low productivity (e.g.Herndl, 1991; Sorokin and Mamaeva, 1991) as a survivalstrategy of phages at low host densities. In this type ofsystems, virus–host contact rates, and thus chances forsuccessful infection and lytic replication, are low (Stewartand Levin, 1984). Moreover, lysogenic bacterial cells(lysogens) appear to be more competitive than non-lysogenic ones in low-nutrient conditions because onceinfected, they cannot be infected again by homologousphages (Lwoff, 1953) and can benefit from the products ofthe lysis of other cells. Higher percentage of lysogens wasfound in offshore than coastal environments (Jiang andPaul, 1998; Weinbauer et al., 2003) and during periodswith low inorganic nitrogen and phosphate concentrations(Williamson et al., 2002). Nevertheless, other studieshave not found a clear relationship between lysogeny andsystem trophy (Weinbauer and Suttle, 1999).Here we present results of a 2-year study on the impact

of grazers and viruses on the bacterioplankton communityin a relatively oligotrophic coastal marine environment(Blanes Bay, NW Mediterranean), characterized by lownutrient concentration and plankton biomass (Duarteet al., 1999; Pinhassi et al., 2006; Alonso-Sáez et al.,2008). Due to the trophic status of this ecosystem wehypothesized that (i) grazing by predators would be theprincipal cause of prokaryotic mortality, and (ii) lysogenywould represent an important percentage of viral infec-tion. We performed monthly, from May 2005 to April2007, experiments to evaluate simultaneously losses ofprokaryotes due to grazers and phages. To simplify, fromnow on, the period May 2005–April 2006 will be called ‘thefirst year’, and May 2006–April 2007 will be ‘the secondyear’.

Results

Physicochemical parameters and chlorophyll

a concentration

Over the sampling period the seawater temperatureranged from 12.0°C to 26.0°C (Fig. 1A), with similar meanvalues for both years (c. 18.5°C) (Table 1). Salinity ranged

between 35.8 and 39.1, and light penetration was onaverage c. 15 m (Table 1).Inorganic nutrient concentrations ( PO4

-3, NO3-) were

relatively low during the whole sampling period (Table 1,Fig. 1B). Average NO3

- concentration was higher duringthe first year than in the second (Table 1), especiallyduring September 2005–April 2006 (Fig. 1B), althoughthe difference between the two years was not statisticallysignificant (P = 0.085). The lowest value of nitrate wasreached in August 2005 (0.10 mM), and the highest in April2006 (3.2 mM). PO4

-3 concentrations were highly variablebetween months. The lowest value was observed in July2006 (0.02 mM), and the highest in July 2005 and March2007 (0.2 mM) (Fig. 1B).The average concentration of chlorophyll a (Chl a) was

very similar between the two years, c. 0.6 mg l-1 (Table 1).Two peaks of Chl a concentration were detected in Marchand May in both years (Fig. 1A). The lowest value wasrecorded in August 2005 (0.02 mg l-1), and the highest inMay 2006 (2.5 mg l-1). Chlorophyll a concentration wascorrelated with NO3

- concentration (Table 2), but not withPO4

-3 concentrations.

Abundances of microorganisms

Abundances of microorganisms did not differ significantlybetween the two years of study (Table 1). However, withinyears viral and heterotrophic nanoflagellates’ (HNF)abundances varied one order of magnitude, whereasbacterial abundance varied 2.5-fold only.During the two years viral abundance showed the

lowest value in February 2007 (8.9 ¥ 106 viruses ml-1),and the highest in May 2006 (6.0 ¥ 107 viruses ml-1)(Fig. 1C). No significant correlation between abundanceof viruses and others parameters was found.Bacterial abundance did not follow a clear seasonal

pattern. The minimum value was detected in August 2005(4.6 ¥ 105 cells ml-1), and two maxima in May 2005(1.6 ¥ 106 cells ml-1) and May 2006 (1.5 ¥ 106 cells ml-1),coinciding with peaks of Chl a and viral abundance(Fig. 1A and C). Percentage of high-nucleic-acid-contentbacteria (%HNA) reached the highest value in April 2006(88.6%), coinciding with the peak of nitrates, while thelowest value of %HNAwas detected in July 2006 (11.6%).Percentage of high-nucleic-acid-content bacteria pre-sented a significant positive correlation with NO3

-, PO4-3

and Chl a concentrations (Table 2). However, no correla-tion was detected between total bacterial abundance andthese variables. Average virus–bacteria ratios weresimilar over the two years (c. 25). The lowest value wasrecorded in August 2006 (9.7), and the highest in January2007 (68.8) (Fig. 1C).Abundance of HNF in both years showed a marked

decrease in late autumn and winter, and two peaks of

1182 J. A. Boras et al.

© 2009 The AuthorsJournal compilation © 2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1181–1193

Fig. 1. Temporal variability of (A) temperatureand chlorophyll a concentration, (B) nutrientconcentrations, (C) abundances of virusesand bacteria and (D) abundances ofheterotrophic nanoflagellates (HNF) andciliates during a 2-year study in Blanes Bay,NW Mediterranean.

Impact of grazers and viruses on bacterioplankton 1183


abundance were recorded in spring and late summer(Fig. 1D). The lowest value was observed in November2005 (1.1 ¥ 102 cells ml-1) and the highest in May 2006(2.5 ¥ 103 cells ml-1). Peaks of HNF abundance coincidedwith peaks of bacterial abundance in spring and summermonths (Fig. 1C and D). Heterotrophic nanoflagellates2–5 mm size class, considered mainly bacterivores(Wikner and Hagström, 1988), represented 53.4 6 13.7%of total HNF. Bacterial abundance showed a slightlyhigher correlation coefficient with HNF 2–5 mm than withtotal HNF abundance (Table 2).Ciliate abundance during the first year showed a

predator–prey pattern with HNF abundance, whereas this

trend was less evident over the second year (Fig. 1D).The lowest ciliate abundance was reached in July 2005(1.0 ¥ 103 cells l-1), and the highest in May 2005 (9.6 ¥ 103

cells l-1, Fig. 1D), as was observed for the other variables.The ciliate community was dominated by oligotrichs,mainly Strombidium sp. (54.5 6 14.2% of total ciliatesabundance in the first year and 67.8 6 13.5% in thesecond, Table 1). Strictly bacterivorous ciliates, such asscuticociliates, represented only 3.0 6 4.9% over the twoyears. Ciliates abundance showed significant correlationwith HNF fractions of 5–10 and 10–20 mm (Table 2). Nocorrelation was found between ciliates and total HNFabundance.

Bacterial production

Average bacterial production (BP) rates were similarbetween the two years (Table 1). The lowest value wasobserved in January 2007 (1.3 ¥ 105 cells ml-1 day-1), andthe highest in May 2005 (2.1 ¥ 106 cells ml-1 day-1)(Fig. 2). For the whole study period, BP presented ahigher correlation coefficient with %HNA than with totalbacterial abundance (Table 2). Only for the second yearBP was correlated with total HNF abundance (r = 0.686,P < 0.02).

Bacterial mortality

Lytic viral production (VPL) was significantly higher in thesecond year than in the first (F1,22 = 10.3, P < 0.005)(Table 1). VPL did not show a clear trend over theseasons. The lowest value was detected in August 2005

Table 1. Average, minimum and maximum values of the physicochemical and biological parameters in both study years.

VariableFirst yearAverage (min.–max.)

Second yearAverage (min.–max.)

Temperature (°C) 18.0 (12.5–26.0) 19.0 (13.4–25.6)Salinity (‰) 37.4 (35.8–39.0) 38.0 (36.3–39.1)NO3

- (mM) 1.4 (0.1–3.2) 0.5 (0.1–1.2)PO4

-3 (mM) 0.2 (0.2–0.2) 0.1 (0.02–0.2)Secchi disc (m) 16.1 (12.0–22.5) 15.3 (8.0–19.5)Chlorophyll a (mg l-1) 0.6 (0.02–1.5) 0.6 (0.2–2.5)Viruses (¥ 107 viruses ml-1) 1.8 (0.5–4.0) 2.0 (0.9–6.0)Bacteria (¥ 105 cells ml-1) 8.0 (4.6–15.6) 8.8 (4.7–14.6)HNF (¥ 102 cells ml-1) 7.0 (1.1–13.4) 8.6 (1.9–24.8)Ciliates (¥ 103 cells l-1) 4.4 (1.0–9.6) 5.3 (1.6–9.4)BP (105 cells ml-1 day-1) 8.7 (2.7–21.2) 7.8 (1.3–17.4)VPL (¥ 105 viruses ml-1 day-1) 6.7 (0.8–16.4) 22.2 (1.9–131.0)VPLG (¥ 105 viruses ml-1 day-1) 1.0 (nd-7.4) 1.6 (nd-6.3)VMMBSS (% day-1) 13.5 (2.5–64.7)** 29.2 (9.6–56.0)**VMMBP (% day-1) 12.2 (3.1–47.7)*** 40.9 (9.6–84.1)***PMMBSS (% day-1) 19.9 (0.0–66.6) 28.6 (0.0–91.6)PMMBP (% day-1) 33.9 (0.0–77.6) 32.4 (0.0–77.0)

Significance of differences between the two years: **P < 0.005, ***P < 0.001.HNF, heterotrophic nanoflagellates; BP, bacterial production; VPL, lytic viral production; VPLG, lysogenic viral production; VMM, virus-mediatedmortality of bacteria (BSS – % of bacterial standing stock; BP – % of bacterial production); PMM, protists-mediated mortality of bacteria; nd, nodetectable.

Table 2. Results of Pearson correlation analysis used to test forsimple correspondence among variables.

Variable n r P

Chl a – NO3- 23 0.583 < 0.01

Chl a – %HNA 20 0.579 < 0.02%HNA – NO3

- 19 0.572 < 0.05%HNA – PO4

-3 19 0.534 < 0.05BA – HNF 24 0.403 < 0.05BA – HNF 2–5 mm 24 0.428 < 0.05Ciliates – HNF 5–10 mm 24 0.597 < 0.01Ciliates – HNF 10–20 mm 24 0.524 < 0.01BP – BA 24 0.501 < 0.02BP – %HNA 20 0.677 < 0.01VPL – Chl a 24 0.604 < 0.01VPL – BA 24 0.515 < 0.01VMMBSS – BA 24 0.476 < 0.02VMMBSS – VPL 24 0.551 < 0.01

Chl a, chlorophyll a concentration; BA, bacterial abundance; %HNA,percentage of high nucleic acid content bacteria. The rest of variableabbreviations are the same as in Table 1.



(2.3 ¥ 105 viruses ml-1 day-1), and the highest in Novem-ber 2006 (2.8 ¥ 107 viruses ml-1 day-1). VPL was positivelycorrelated with Chl a concentration and bacterial abun-dance (Table 2). Lysogeny was observed in four occa-sions along the first year, and seven times during thesecond. The lowest percentage of lysogeny to total viralproduction was detected in November 2006 (0.6%), andthe highest in August 2005 (78.4%), coinciding with thehighest and the lowest VPL respectively. Lysogeny eventsoccurred mainly in summer (four times) and winter (fourtimes) months (Fig. 3).Virus-mediated mortality of bacteria (VMM) was signifi-

cantly different between the two years, either related tolosses of bacterial standing stock (VMMBSS, F1,22 = 12.3,P < 0.005) or to BP (VMMBP, F1,22 = 23.8, P < 0.001)(Table 1). Average VMMBSS and VMMBP for the firstyear were 13.5 6 17.0% day-1 and 12.2 6 12.3% day-1,respectively, while in the second, higher values wererecorded for both variables (29.2 6 14.8% day-1 and40.9 6 20.7% day-1) (Table 1). For the whole period,

VMMBSS varied 26-fold between minimum and maximumvalue (2.5% day-1 and 64.7% day-1, Fig. 4A), and 27-foldfor VMMBP (3.1% day-1 and 84.1% day-1, Fig. 4B).VMMBSS was significantly correlated with bacterial abun-dance and VPL (Table 2). Only for the second year,VMMBSS was correlated with %HNA (r = 0.677, P < 0.05)and BP (r = 0.439, P < 0.05), coinciding with higher valuesof VMMBSS than in the first year.Protists-mediated mortality of bacteria (PMM) did not

differ between the two years (Table 1). PMM was notdetectable in August 2005 and April 2007 (Fig. 4). Whenmeasurable, the lowest value of PMM related to the bac-terial standing stock (PMMBSS) was observed in October2006 (6.3% day-1, Fig. 4A), and to BP (PMMBP) in March2006 (3.9% day-1, Fig. 4B). The highest value of PMMBSS

was observed in May 2006 (91.6% day-1), and for PMMBP

in September 2005 (77.6% day-1). In two occasions,viruses and grazers together removed c. 100% day-1 ofBP (May and November 2006) (Fig. 4B). PMMBSS in thesecond year was positively correlated with total bacterial

Fig. 2. Average bacterial production (BP)over the study period (n = 3). Bars indicatestandard deviation.

Fig. 3. Temporal variability of the averagelytic (VPL) and lysogenic (VPLG) viralproduction. In bold: percentage of VPLG intotal viral production. Each value correspondsto experimental triplicates 6 standarddeviation.



abundance (r = 0.620, P < 0.05), BP (r = 0.620, P < 0.05)and HNF (r = 0.585, P < 0.05). The correlation with HNFwas mainly due to 2–5 mm (r = 0.596, P < 0.05) and5–10 mm (r = 0.598, P < 0.05) fractions.

Discussion

Evaluation of the used methodology

The methods to estimate complex ecosystem variables,such as PMM based on fluorescent labelled bacteria(FLB) disappearance and bacterial production, have theirsuite of assumptions and uncertainties. For grazing onbacteria one can refer to Strom (2000) and referencestherein, where the used method is widely discussed. Thebulk disappearance of FLB as tracers of natural bacteriaduring 24 h incubation time is in our view the best optionin studies where the dynamics of the whole bacterialassemblage is targeted, as it introduces the least watermanipulation possible which is always a potential sourceof artefact. In order to evaluate bacterial growth enhance-ment owing to confinement, we compared BP ratesmeasured as the sum of the net increase of bacterialabundance (BPN), losses of bacteria due to protists

(G) and viruses (RLCGR) in the same incubation bottlesas for grazing estimations, with in situ measurementsof BP by 3H-leucine incorporation method (BPLeu; J.M.Gasol, unpublished), as described in Kirchman (1993).We obtained a significant relationship between bothmeasurements (log BPLeu = -1.65 + 1.17 (6 0.28) log BP,n = 24, r = 0.670, P < 0.01), with a slope not significantlydifferent from 1. This indicates that our BP results arecomparable to those obtained in situ with standardmethods, although BP values in the experimental bottleswere higher than the in situ values.Viral mortality rates are subject to the calculation of

burst size. We assumed that the only cause of bacterialdecrease over short time periods (1 h) in the experimentalfalcon tubes was cell lysis, and we did not take intoaccount the viral decay and bacterial production rateduring this time interval. Burst size values reported in thisstudy, between 6 and 375, are within the range obtainedfrom different aquatic environments (Guixa-Boixereuet al., 1996; Parada et al., 2006), and are lower thanvalues found in the anoxic part of an eutrophic lake(~500, Weinbauer and Höfle, 1998). Also, bacterial lossescaused by lysis measured in this study are the potentiallosses, as we did not consider grazing on infected cells by

Fig. 4. Temporal variability of the averageprotists- (PMM) and virus-mediated mortalityof bacteria (VMM), as a percentage ofbacterial standing stock BSS (A) and bacterialproduction BP (B). Each value corresponds toexperimental triplicates 6 standard deviation.



protists in our calculations. This process could reduce thepercentage of bacteria that burst due to viral activity innatural communities.The virus reduction approach (VRA) used to evaluate

VP and VMM is based on the assumption that all viralproduction observed during the experiment is a result ofinfections previous to the incubation. It is also assumedthat no new infections occur, and the filtration and incu-bation does not induce lysogenic bacteria. This methodallows direct observation of changes in viral abundanceover time, taking into account the bacterial losses duringfiltration, and allows the distinction between production ofvirulent and temperate phages. In addition, it is relativelyeasy and inexpensive to perform (Winget et al., 2005).Detection of lysogeny is based on lysis induction by mito-mycin C. It is known that in some circumstances (i.e.nutrients availability, pH) this agent is not sufficient toinduce the lytic cycle in all prophages (Cochran et al.,1998; Weinbauer and Suttle, 1999 and referencestherein). However, this method is widely used and mito-mycin C is superior to other inducing agents, hence theobtained results are comparable to other studies. Finally,the main drawbacks of this technique are the consider-able sample manipulation and the loss of a portion ofbacterial community during filtration. Despite of these dis-advantages, VRA is considered one of the best-suitedincubation-based methods for VP and VMM estimations(Helton et al., 2005).


Abundances of viruses and bacteria found in Blanes Baywere similar to others found in the Mediterranean Sea(Guixa-Boixereu et al., 1999b; Bettarel et al., 2002). Viraland bacterial abundances (total and %HNA) were notfollowing the same pattern during the study period andthere was no significant correlation between them, neitherwith Chl a concentration. This suggests that (i) the virusesdetected could be pathogens of other organisms besidesbacteria, and/or (ii) these phages could be in differentphases of cell infection at the sampling moment. Otherauthors found a similar lack of correlation in temporalstudies in the NW Mediterranean Sea (Bettarel et al.,2002; M.G. Weinbauer, unpublished) and in the North Sea(Winter et al., 2004). Nevertheless, Jiang and Paul (1994)have found a significant positive relationship betweenviral and bacterial abundances during an annual cyclein Tampa Bay (Florida). Abundance of HNF wassimilar to values found in previous years in Blanes Bay(Vaqué et al., 1997; Unrein et al., 2007). Heterotrophicnanoflagellates were weakly correlated with bacterialabundance, and this correlation did not improve muchmore when considering the bacterivore HNF 2–5 mm frac-tion (Wikner and Hagström, 1988), suggesting that trophic

cascade effects (Calbet et al., 2001) could have blurredthe real bacteria–HNF relationship. Thus, the lack ofstrong correlations observed between most variables inthis study could be due to changes in metabolic pro-cesses and succession of communities at all trophiclevels. We are also aware that the considered microbialprocesses could be highly variable within days, or evenhours (e.g. Winter et al., 2004). Then, with the span ofsampling we probably missed small oscillations withinshort periods of time, which would be reflected in weakcorrelations among variables. It is interesting that roughlyconstant bacterial abundances (variation about 2.5-fold)can be maintained in the environment in spite of highlyoscillating abundances of predators within years.

Lysogeny

There are indications that lysogeny dominates in olig-otrophic systems as a survival strategy of viruses atlow host densities (Stewart and Levin, 1984). Weinbauerand colleagues (2003) have found that the frequencyof lysogenic cells was inversely related to bacterialabundance and production in Mediterranean and BalticSeas. Also, higher percentage of lysogens was observedin bacteria isolates from an offshore poor environment(Jiang and Paul, 1994), which could suggest that thetrophic conditions, not only host density, determine theoccurrence of lysogeny. During our study we did notfind any correlation between bacterial abundance andlysogenic infection. However, we have detected highernumber of lysogeny cases in the second year, when nutri-ent concentrations were lower than in the first year(Fig. 1B). Thus, our results seem to indicate that lysogenymight be influenced by the trophic status of the system.We are aware, however, that more data would be neededto prove this suggestion.

Bacterial losses due to protists and viruses

Average values of bacterial losses by protists per yearwere similar (Table 1), indicating constant PMM in thissystem over the two years. In the same study area, similargrazing rates were found in previous studies (Unreinet al., 2007). Within protists, HNF are considered the mainbacterivores, while ciliates feed preferentially on largercells as nanoflagellates (Stoecker and Capuzzo, 1990),and/or small phytoplanktonic cells (Sherr and Sherr,2002). In agreement with this, HNF abundances, espe-cially HNF 2–5 mm and HNF 5–10 mm size classes, werepositively correlated with bacterial abundance, grazingrates and BP, while ciliates presented correlation with theHNF 5–10 mm and HNF 10–20 mm fractions.Viruses were an important source of bacterial mortality

in this oligotrophic environment, sometimes overweight-



ing the impact of protists on the bacterial community.Comparison of bacterial losses due to protists and virusesin different marine systems shows that in eutrophic habi-tats those two predators cause similar losses of bacterialproduction (Fuhrman and Noble, 1995), in extremesystems (Arctic and Antarctic) viruses appear to be themain source of bacterial mortality (Guixa-Boixereu et al.,2002), and in oligotrophic waters impact of protists isoften more important than viruses (Bettarel et al., 2002)(Table 3). However, in the second year of our study,viruses and protists had a similar impact on bacteria, andthese results are in agreement with findings of Hwang andCho (2002) in the oligotrophic East Sea, Korea. The typeof bacterial mortality has a great importance for carbonand nutrient fluxes through the ecosystem. Grazing onbacteria by protists shifts the particular organic carbon(POC) to higher trophic levels, whereas viral infectioncauses lysis of bacterial cell and release of the dissolvedand detrital organic matter. This dissolved organic matter(DOM) can be a source of nutrients, such as N, P or Fe,

for other microorganisms (Fuhrman, 1999). We have cal-culated that during our study, the potential bacterialcarbon flux to higher trophic levels mediated by HNF wason average 3.12 6 3.35 mg C l-1 day-1 (range: no detect-able to 16.1 mg C l-1 day-1), assuming a cell-carbon factorof 12 fg C cell-1 (Simon and Azam, 1989). In contrast,the input of bacterial carbon to the DOC pool due toVMM was in average 2.5 6 2.9 mg C l-1 day-1 (range:0.2–12.1 mg C l-1 day-1). The dominance of one of thesetwo processes will, thus, shape the carbon flow patternthrough the food webs.Mortality of bacteria due to phages during our study

increased with the increase of bacterial abundance andproduction, suggesting that the metabolic status of thehost is critical for viral infection and proliferation, as it wasfound in other studies (Steward et al., 1996; Weinbaueret al., 2003). The interesting fact is that the averages ofVPL and VMM over the first year were approximatelythreefold lower than during the second year, which meansthat bacterial mortality due to viruses was a more variable

Table 3. Bacterial production losses caused by protists and viruses detected simultaneously in different systems and methods used to determinateprokaryotic mortality caused by viruses.

System/geographical location Method Losses by protists Losses by viruses Reference

Extreme systems

ArcticBering and Chukchi Seas TEM Steward et al. (1996)Integrated depthsBottom 23 (17–30) 23 (9–37)Bottom 5 (1–9) 23 (9–37)50 m 17 (6–27) 11 (3–20)Bottom 25 (9–40) 12 (4–21)50 m 3 (1–6) 11 (3–19)Bottom 4 (1–8) 9 (2–16)

Franklin Bay, Canadian Arctic VDA nd to -0.004a -0.006 to -0.015a Wells and Deming (2006)AntarcticBellingshausen Sea VDR 35.3 400 Guixa-Boixereu et al. (2002)Bransfield Strait 37.2 180Gerlache Strait 0.9 44

Eutrophic

Santa Monica, California, USA BDR 1.8–2.2b 1.2–1.4b Fuhrman and Noble (1995)2.8–3.4b 2.7–2.8b

Ria de Aveiro, Portugal VDA Almeida et al. (2001)Marine waters 69 36Brackish waters 73 59

Masan Bay, Korea TEM 41 9.4 Choi et al. (2003)Mediterranean Sea, Spain (Masnou harbour) VDR 120 100 Guixa-Boixereu et al. (1999a)

Oligotrophic

Mediterranean Sea, Spain TEM 31.0 nd Guixa-Boixereu et al. (1999b)36.5 nd D. Vaqué (unpubl. data)74.0 nd

Mediterranean Sea, Spain TEM 19 17 Guixa-Boixereu et al. (1996)East Sea, Korea TEM 9.5 (1.6–32.9) 13.1 (6.9–26.9) Hwang and Cho (2002)Mediterranean Sea, France TEM 26–80 nd-18 Bettarel et al. (2002)Mediterranean Sea, Spain VRA 33.9 (nd-77.6) 12.2 (3.1–47.7) This study

32.4 (nd-77.0) 40.9 (9.6–84.1)

a. Bacterial mortality in h-1 (for details see reference).b. Bacterial loss rate in % h-1 (for details see reference)BDR, bacterial decay rate; VDR, viral decay rate; TEM, determination of frequency of visibly infected cells by transmission electron microscopy;VDA, virus dilution approach; VRA, virus reduction approach; nd, no detected



process than mortality due to protists. On the other hand,the averages of viral and bacterial abundances remainedsimilar between the two years, as well as other measuredparameters, except nitrate, which had higher, yet notsignificantly, concentrations during the first year thanduring the second year (Fig. 1B). Thus, variation in VPLand VMM could be related with the occurrence of struc-tures named porins on the bacterial surface wall (Szmel-cman and Hofnung, 1975; Delcour, 1997), which allow thecontinuous exchange of nutrients across the bacterial cellwall with the surrounding environment. These structuresare also receptor sites for phages (Braun and Hantke,1977). Porins’ number on the cell wall can be modified inresponse to changing environmental conditions, e.g. canbe lost under higher viral pressure (Lenski, 1988), or canappear under nutrients depletion (Korteland et al., 1982;Poole and Hancock, 1986) favouring new viral infections.It is also possible that changes in VMM during our studywere caused by changes in bacterial community compo-sition induced by nutrient conditions (Øvreås et al., 2003;Alonso-Sáez et al., 2007). The bacterial mortality due toviruses could be therefore enhanced or depleted due todistinct susceptibility on viral infection of newly dominatingbacterial groups (Bouvier and del Giorgio, 2007).In summary, preservation of bacterial population abun-

dance at constant level, in spite of environmental changesand variation of predation pressure, is possible. Interest-ingly, in a microcosm experiment, Middelboe and col-leagues (2001) also observed that bacterial abundancewas not affected by phages during their experiments,since infection-resistant bacteria complemented theabundance decrease caused by viral lysis. It seems clearthen that although the bacterial standing stocks aresimilar between years, different mechanisms can controlbacterial growth and create different paths for the carbonflow in the system. However, further experimental studiesare needed to determine factors that regulate the domi-nance of grazing or bacterial lysis through the year.

Conclusions

During our study, phages were an important source ofbacterial mortality in an oligotrophic coastal area (NWMediterranean), causing in 37% of cases higher losses ofbacterial production than protists. Lytic infection domi-nated over lysogeny during the two years. We suggestthat changes in nutrient concentrations in the environmentplay an important role in regulating viral infection of bac-terial cells.

Experimental procedures

Study site and sampling strategy

Surface water samples (0.5 m) were collected from May 2005to April 2007 in Blanes Bay, Spain (The Blanes Bay Microbial

Observatory, NW Mediterranean, 41°40′N, 2°48′E, 20 mdepth), at 0.5 miles from the shore. Samples were collectedonce a month, in 10 l polyethylene carboys, and kept in thedark and refrigerated until reaching the lab (~2 h). Watertemperature and salinity were measured in situ with a CTD(conductivity, temperature, depth).

Chlorophyll a and nutrient concentrations

Chlorophyll a was extracted from 250–500 ml of watersamples and determined fluorometrically (Yentsch andMenzel, 1963). Inorganic phosphorous and nitrates wereanalysed using standard methods (Grasshoff et al., 1983).


Viral abundances were determined by flow cytometry. Sub-samples (2 ml) were fixed with glutaraldehyde (0.5% finalconcentration), refrigerated, quick frozen in liquid nitrogenand stored at -80°C as described in Marie and colleagues(1999). Counts were made using FACSCalibur flow cytom-eter (Becton and Dickinson) with a blue laser emitting at488 nm. Samples were stained with SYBR Green I, and runat a medium flow speed (Brussaard, 2004).The term ‘bacteria’ used along this article refers to all

prokaryotes, as in our study we did not distinguishbetween Bacteria and Archaea. Bacteria and HNF in situ

abundances were obtained by epifluorescence microscopy(Olimpus BX40). Subsamples 100 ml were fixed with glutaral-dehyde (1% final concentration). Aliquots of 20 ml were fil-tered through 0.2 mm (for bacteria), and 60 ml through 0.6 mm(for HNF) black polycarbonate filters, and stained with DAPI(4,6-diamidino-2-phenylindole; Porter and Feig, 1980) at afinal concentration of 5 mg ml-1 (Sieracki et al., 1985). At least200–300 bacteria and 20–100 HNF were counted persample. Heterotrophic nanoflagellates were grouped into foursize classes: # 2, 2–5, 5–10 and 10–20 mm.Ciliate abundances were determinate using the Utermöhl

method. One litre of sample was fixed with acidic lugol (2%final concentration). After sedimentation of 100 ml aliquots,ciliates were counted in an inverted microscope (Zeiss) andidentified to genera level when possible (Lynn and Small,2000).

Bacterial mortality and production

Bacterial losses due to protists were evaluated following theFLB (Spanish Type Culture Collection, http://www.cect.org/index2.html, Burjassot, València) disappearance method(Sherr et al., 1987; Vázquez-Domínguez et al., 1999). Threepolycarbonate (2 l) bottles were filled with 1 l of natural sea-water, and other two (controls) with virus-free water. Fluores-cent labelled bacteria were added to a final concentration of15–20% of bacterial in situ concentration. Bottles were incu-bated in a thermostatic chamber during 24 h, simulating in

situ temperature and light conditions. Samples were taken attime 0 and 24 h to evaluate abundances of bacteria, FLB andHNF by epifluorescence microscopy. Control bottles showedno decrease of FLB during the experiment.



Grazing rates of bacteria were obtained following the math-ematical model #3 of Salat and Marrasé (1994), based on thespecific grazing rate (g) and specific net growth rate (a). First,net bacterial production (BPN) in the incubation bottles wasobtained:

BP BA e cells ml day ,N 0= × −( )[ ]− −at 1 1 1

where BA0 is bacterial abundance at the beginning of theexperiment; t is time of experiment (1 day).Then, grazing rate (G) was calculated

G g a= ( ) × [ ]− −BP cells ml dayN1 1 .

Finally, protists-mediated mortality of bacteria (PMM) wascalculated as the percentage of BSS and BP losses:

PMM BA day ,BSS 0= ×( ) [ ]−G 100 1%

PMM BP dayBP = ×( ) [ ]−G 100 1% .

The virus reduction approach (Weinbauer et al., 2002;Wilhelm et al., 2002) was followed to determine viral produc-tion and bacterial losses due to phages. Briefly, 1 l of seawa-ter was pre-filtered by 0.8-mm-pore-size cellulose filter(Whatman), and next concentrated by a spiral-wound car-tridge (0.22 mm pore size, VIVAFlow 200), obtaining 60 ml ofbacterial concentrate. Virus-free water was collected filtering1 l of seawater using a cartridge of 100 kDa molecular masscut-off (VIVAFlow 200). A mixture of virus-free water (240 ml)and bacterial concentrate (60 ml) was prepared and distrib-uted into six sterile falcon plastic tubes (50 ml in each one).Three of the tubes were maintained without any manipula-tions as controls, while in other three, mitomycin C (Sigma)was added (1 mg ml-1 final concentration) as inducing agentof the lytic cycle in lysogenic bacteria. All falcon tubes wereincubated in a thermostatic chamber simulating in situ tem-perature and light conditions, during 12–19 h. Samples forbacterial and viral abundances were collected at time zeroand each hour during the first 6 h of the experiment, and atthe end of the experiment. The choice of the sampling periodwas made based on previous experiments that showed thatalmost 100% of the lysis occurs during the first 6 h. Samplesfor viral and bacterial counts were fixed with glutaraldehydeand stored as described before for viruses. Viruses werecounted by flow cytometry (see above). Bacteria werestained with SYTO 13 and run in a flow cytometer at a lowflow speed, using 50 ml of 0.92 mm yellow-green latex beadsas an internal standard (del Giorgio et al., 1996; Gasol anddel Giorgio, 2000). Burst size (BS) was estimated from viralproduction experiments, as in Middelboe and Lyck (2002)and Wells and Deming (2006). Briefly, an increase of viralabundance during an incubation period of viral productionexperiment was divided by a decrease of bacterial abun-dance at the same period of time. In all these experiments,growth of bacteria was observed and such calculations werepossible only for very short time intervals (1 h). Burst sizeestimated ranged from 6 to 375 viruses per cell, with meanvalues of 93 6 92 in the first year, and 91 6 94 in the secondyear.Estimation of virus-mediated mortality of bacteria (VMM)

was performed following the model presented by Weinbauerand colleagues (2002) and Winter and colleagues (2004).Briefly, virus increase in the control tubes represents lytic viral

production (VPL), and an increase in mitomycin C treatmentsrepresents total (VPT), i.e. lytic plus lysogenic, viral produc-tion. A difference between VPT and VPL represents lysogenicproduction (VPLG). As during tangential flow filtration the lossof part of bacterial in situ standing stock occurred (from 1% to78%, mean of 56.3 6 20.3%), we multiplied VPL and VPLG bythe bacterial loss factor (Winget et al., 2005) to compare thevalues between different months. The percentage of lysog-eny in total viral production was calculated:

% % . VP VP VPLG LG T= ×( ) [ ]100

Following the method used by Guixa-Boixereu (1997), therate of lysed cells (RLC) was obtained dividing VPL by BS:

RLC VP BS cells ml dayL= [ ]− −1 1 .

RLC was used to calculate VMM as a percentage of bacterialstanding stock (VMMBSS):

VMM RLC BA day ,BSS 0= ×( ) [ ]−100 1%

where BA0 is the initial bacterial abundance in the viral pro-duction experiment. Assuming that percentage of losses ofbacterial standing stock due to viruses is the same in falcontubes and in the grazing polycarbonate bottles, we usedVMMBSS to calculate the rate of lysed bacteria during thegrazing experiment (RLCGR):

RLC VMM BA cells ml day ,GR BSS GR= ×( ) [ ]− −100 1 1

where BAGR is bacterial abundance in the grazing bottles attime zero. Finally, using RLCGR, VMM as a percentage ofbacterial production (VMMBP) could be calculated:

VMM RLC BP day ,BP GR= ×( ) [ ]−100 1%

where BP is total bacterial production.Bacterial production was calculated summing the BPN, G

and RLCGR:

BP BP RLC cells ml dayN GR= + + [ ]− −G 1 1 .

Statistical analyses

Normal distribution of data was checked using the Shapiro–Wilk W-test, and data were logarithmic transformed if neces-sary. Annual variations of different parameters were analysedby one-way ANOVA for normal distributions, and by Wilcoxontest for non-normal distributions. The Pearson correlation andregression analyses were used to determine the relationshipsbetween parameters. All statistical analyses were performedusing the JMP program.

Acknowledgements

This study was supported by the PROCAVIR (CTM2004-04404-C02) and MICROVIS (CTM2007-62140) projectsfunded by the Spanish Ministry of Education (MEC). J.A.B.work was supported by PhD fellowship from the MEC (FPUgrant) and M.M.S. by I3P-CSIC postdoctoral contract fundedby the Fondo Social Europeo. The authors thank to R.Ventosa and M.I. Abad for inorganic nutrients analyses. Sam-pling, CTD data and Chl a analyses were provided by col-leagues involved in the Blanes Bay Microbial Observatorynet. We are indebted to Dr C.P.D. Brussaard for helpful



advice in both scientific and lab issues. We also thank to twoanonymous reviewers for their helpful comments that havecontributed to improve the manuscript.

References

Almeida, M.A., Cunha, M.A., and Alcântara, F. (2001) Loss ofestuarine bacteria by viral infection and predation in micro-cosm conditions. Microb Ecol 42: 562–571.

Alonso-Sáez, L., Balagué, V., Sà, E.L., Sánchez, O.,González, J.M., Pinhassi, J., et al. (2007) Seasonality inbacterial diversity in north-west Mediterranean coastalwaters: assessment through clone libraries, fingerprintingand FISH. FEMS Microbiol Ecol 60: 98–112.

Alonso-Sáez, L., Vázquez-Domínguez, E., Pinhassi, J.,Cardelús, C., Sala, M.M., Lekunberri, I., et al. (2008)Factors controlling the year-round variability in carbon fluxthrough bacteria in a coastal marine system. Ecosystems11: 397–409.

Bettarel, Y., Dolan, J.R., Hornak, K., Lemée, R., Masin, M.,Pedrotti, M.-L., et al. (2002) Strong, weak, and missinglinks in a microbial community of the NW MediterraneanSea. FEMS Microbiol Ecol 42: 451–462.

Bouvier, T., and del Giorgio, P.A. (2007) Key role of selectiveviral-induced mortality in determining marine bacterialcommunity composition. Environ Microbiol 9: 287–297.

Bratbak, G., Heldal, M., Norland, S., and Thingstad, T.F.(1990) Viruses as partners in spring bloom microbial tro-phodynamics. Appl Environ Microbiol 56: 1400–1405.

Braun, V., and Hantke, K. (1977) Bacterial receptors forphages and colicins as constituents of specific transportsystems. In Microbial Interactions. Reissig, J.L. (ed.).London, UK: Chapman & Hall, pp. 99–137.

Brussaard, C.P.D. (2004) Optimisation of procedures forcounting viruses by flow cytometry. Appl Environ Microbiol

70: 1506–1513.Calbet, A., Landry, M.R., and Nunnery, S. (2001) Bacteria–flagellate interactions in the microbial food web of the oli-gotrophic subtropical North Pacific. Aquat Microb Ecol 23:

283–292.Choi, D.H., Hwang, C.Y., and Cho, B.C. (2003) Comparisonof virus- and bacterivory-induced bacterial mortality in theeutrophic Masan Bay, Korea. Aquat Microb Ecol 30: 117–125.

Cochran, P.K., Kellogg, C.A., and Paul, J.H. (1998) Proph-age induction of indigenous marine lysogenic bacteria byenvironmental pollutants. Mar Ecol Prog Ser 164: 125–133.

Delcour, A.H. (1997) Function and modulation of bacterialporins: insights from electrophysiology. FEMS Microbiol

Lett 151: 115–123.Duarte, C.M., Agustí, S., Kennedy, H., and Vaqué, D. (1999)The Mediterranean climate as a template for Mediterra-nean marine ecosystems: the example of northeastSpanish littoral. Prog Oceanogr 44: 245–270.

Fuhrman, J.A. (1999) Marine viruses and their biogeochemi-cal and ecological effects. Nature 399: 541–548.

Fuhrman, J.A., and Noble, R.T. (1995) Viruses and protistscause similar bacterial mortality in coastal seawater.Limnol Oceanogr 40: 1236–1242.

Gasol, J.M., and del Giorgio, P.A. (2000) Using flow cytom-etry for counting natural planktonic bacteria and under-standing the structure of planktonic bacterial communities.Sci Mar 64: 197–224.

del Giorgio, P.A., Bird, D.F., Prairie, Y.T., and Planas, D.(1996) Flow cytometric determination of bacterial abun-dance in lake plankton with the green nucleid acid stainSYTO 13. Limnol Oceanogr 41: 783–789.

Grasshoff, K., Ehrhardt, M., and Kremling, K. (1983)Methodsof Seawater Analysis. Kiel, Germany: Verlag Chemie.

Guixa-Boixereu, N. (1997) Abundància i dinàmica dels virusen ecosistemes planctònics. Doctoral Thesis. Barcelona:Universitat de Barcelona.

Guixa-Boixereu, N., Calderón-Paz, J.I., Heldal, M., Bratbak,G., and Pedrós-Alió, C. (1996) Viral lysis and bacterivoryas prokaryotic loss factors along a salinity gradient. AquatMicrob Ecol 11: 215–227.

Guixa-Boixereu, N., Lysnes, K., and Pedrós-Alió, C. (1999a)Viral lysis and bacterivory during a phytoplankton bloom ina coastal water microcosm. Appl Environ Microbiol 65:

1949–1958.Guixa-Boixereu, N., Vaqué, D., Gasol, J.M., and Pedrós-Alió,C. (1999b) Distribution of viruses and their potential effecton bacterioplankton in an oligotrophic marine system.Aquat Microb Ecol 19: 205–213.

Guixa-Boixereu, N., Vaqué, D., Gasol, J.M., Sánchez-Cámara, J., and Pedrós-Alió, C. (2002) Viral distributionand activity in Antarctic waters. Deep Sea Res II 49: 827–845.

Helton, R.R., Cottrell, M.T., Kirchman, D.L., and Wommack,K.E. (2005) Evaluation of incubation-based methods forestimating virioplankton production in estuaries. Aquat

Microb Ecol 41: 209–219.Herndl, G.J. (1991) Microbial biomass dynamics along atrophic gradient at the Atlantic Barrier Reef off Belize(Central America). Mar Ecol 12: 41–51.

Hwang, C.Y., and Cho, B.C. (2002) Virus-infected bacteria inoligotrophic open waters of the East Sea, Korea. AquatMicrob Ecol 30: 1–9.

Jiang, S.C., and Paul, J.H. (1994) Seasonal and dielabundance of viruses and occurrence of lysogeny/bacteriocinogeny in the marine environment. Mar Ecol

Prog Ser 104: 163–172.Jiang, S.C., and Paul, J.H. (1998) Significance of lysogeny inthe marine environment: studies with isolates and a modelof lysogenic phage production. Microb Ecol 35: 235–243.

Kirchman, D.L. (1993) Leucine incorporation as a measure ofbiomass production by heterotrophic bacteria. In Hand-

book of Methods in Aquatic Microbial Ecology. Kemp, P.F.,Sherr B.F., Sherr E.B., and Cole, J. (eds). Boca Raton, FL:Lewis Publishers, pp. 509–512.

Korteland, J., Tommassen, J., and Lugtenberg, B. (1982)PhoE protein pore of the outer membrane of Escherichiacoli K12 is a particularly efficient channel for organic andinorganic phosphate. Biochim Biophys Acta 690: 282–289.

Lenski, R.E. (1988) Dynamics if interactions between bacte-ria and virulent bacteriophage. Adv Microb Ecol 10: 1–44.

Lwoff, A. (1953) Lysogeny. Bacteriol Rev 17: 269–337.Lynn, D.H., and Small, E.B. (2000) Phylum Ciliophora. In: An



Illustrated Guide to the Protozoa. Lee, J.J., Bradbury, P.C.,and Leedale, G.F. (eds). Lawrence, KS, USA: Society ofProtozoologists, pp. 371–640.

Marie, D., Brussaard, C.P.D., Thyrhaug, R., Bratbak, G., andVaulot, D. (1999) Enumeration of marine viruses in cultureand natural samples by flow cytometry. Appl Environ

Microbiol 65: 45–52.Middelboe, M., and Lyck, P.G. (2002) Regeneration of dis-solved organic matter by viral lysis in marine microbialcommunities. Aquat Microb Ecol 27: 187–194.

Middelboe, M., Hagström, A., Blackburn, N., Sinn, B.,Fischer, U., Borch, N.H., et al. (2001) Effects of bacterioph-ages on the population dynamics of four strains of pelagicmarine bacteria. Microb Ecol 42: 395–406.

Murray, A.G., and Jackson, G.A. (1992) Viral dynamics: amodel of the effects of size, shape, motion and abundanceof single-celled planktonic organisms and other particles.Mar Ecol Prog Ser 89: 103–116.

Øvreås, L., Bourne, D., Sandaa, R.-A., Casamayor, E.O.,Benlloch, S., Goddard, V., et al. (2003) Response of bac-terial and viral communities to nutrient manipulations inseawater mesocosms. Aquat Microb Ecol 31: 109–121.

Parada, V., Herndl, G.J., and Weinbauer, M.G. (2006) Viralburst size of heterotrophic prokaryotes in aquatic systems.J Mar Biol Ass UK 86: 613–621.

Peters, F. (1994) Prediction of planktonic protistan grazingrates. Limnol Oceanogr 39: 195–206.

Pinhassi, J., Gómez-Consarnau, L., Alonso-Sáez, L., Sala,M.M., Vidal, M., Pedrós-Alió, C., and Gasol, J.M. (2006)Seasonal changes in bacterioplankton nutrient limitationand their effects on bacterial community composition inthe NW Mediterranean Sea. Aquat Microb Ecol 44: 241–252.

Poole, K., and Hancock, R.E. (1986) Phosphate-starvation-induced outer membrane proteins of members of thefamilies Enterobacteriaceae and Pseudomonodaceae:demonstration of immunological cross-reactivity with anantiserum specific for porin protein P of Pseudomonasaeruginosa. J Bacteriol 165: 987–993.

Porter, K.G., and Feig, Y.S. (1980) The use of DAPI foridentifying and counting aquatic microflora. Limnol Ocean-ogr 25: 943–948.

Salat, J., and Marrasé, C. (1994) Exponential and linearestimations of grazing on bacteria: effects of changes inthe proportion of marked cells. Mar Ecol Prog Ser 104:

205–209.Sherr, B.F., Sherr, E.B., and Fallon, R.D. (1987) Use ofmonodispersed, fluorescently labeled bacteria to estimatein situ protozoan bacterivory. Appl Environ Microbiol 53:

958–965.Sherr, E.B., and Sherr, B.F. (2002) Significance of predationby protists in aquatic microbial food webs. Antonie Van

Leeuwenhoek 81: 293–308.Sieracki, M.E., Johnson, P.B., and Sieburth, J.M. (1985)Detection, enumeration and sizing of planctonic bacteria byimage analyzed epifluorescence microscopy. Appl EnvironMicrobiol 49: 799–810.

Simon, M., and Azam, F. (1989) Protein content and proteinsynthesis rates of planktonic marine bacteria. Mar Ecol

Prog Ser 51: 201–213.Sorokin, Y.I., and Mamaeva, T.I. (1991) Role of planktonic

bacteria in productivity and cycling of organic matter in theEastern Pacific Ocean. Hydobiologia 209: 39–50.

Steward, G.F., Smith, D.C., and Azam, F. (1996) Abundanceand production of bacteria and viruses in the Bering andChukchi Seas. Mar Ecol Prog Ser 131: 287–300.

Stewart, F.M., and Levin, B.R. (1984) The population biologyof bacterial viruses: why be temperate? Theor Popul Biol

26: 93–117.Stoecker, D.K., and Capuzzo, J.M. (1990) Predation on pro-tozoa: its importance to zooplankton. J Plankton Res 12:

891–908.Strom, S.L. (2000) Bacterivory: interactions between bacteriaand their grazers. In Microbial Ecology of the Oceans.Kirchman, D.L. (ed.). New York, USA: Wiley-Liss, pp. 351–386.

Szmelcman, S., and Hofnung, M. (1975) Maltose transport inEscherichia coli K12. Involvement of the bacteriophagelambda receptor. J Bacteriol 124: 112–118.

Thingstad, T.F. (2000) Elements of a theory for the mecha-nisms controlling abundance, diversity, and biogeochemi-cal role of lytic bacterial viruses in aquatic systems. LimnolOceanogr 45: 1320–1328.

Unrein, F., Massana, R., Alonso-Sáez, L., and Gasol, J.M.(2007) Significant year-round effect of small mixotrophicflagellates on bacterioplankton in an oligotrophic coastalsystem. Limnol Oceanogr 52: 456–469.

Vaqué, D., Gasol, J.M., and Marrasé, C. (1994) Grazing rateson bacteria: the significance of methodology and ecologicalfactors. Mar Ecol Prog Ser 109: 263–274.

Vaqué, D., Blough, H.A., and Duarte, C.M. (1997) Dynamicsof ciliate abundance, biomass and community compositionin an oligotrophic coastal environment (NW MediterraneanSea). Aquat Microb Ecol 1: 71–83.

Vázquez-Domínguez, E., Gasol, J.M., Peters, F., and Vaqué,D. (1999) Measuring the grazing losses of picoplankton:methodological improvements to the use of fluorescentlylabeled tracers combined with flow cytometry. Aquat

Microb Ecol 20: 110–128.Weinbauer, M.G., and Höfle, M.G. (1998) Size-specific mor-tality of lake bacterioplankton by natural virus communities.Aquat Microb Ecol 15: 103–113.

Weinbauer, M.G., and Peduzzi, P. (1995) Significance ofviruses versus heterotrophic nanoflagellates for controllingbacterial abundance in the northern Adriatic Sea.J Plankton Res 17: 1851–1856.

Weinbauer, M.G., and Suttle, C.A. (1999) Lysogeny andprophage induction in coastal and offshore bacterial com-munities. Aquat Microb Ecol 18: 217–225.

Weinbauer, M.G., Winter, C., and Höfle, M.G. (2002)Reconsidering transmission electron microscopy basedestimates of viral infection of bacterioplankton usingconversion factors derived from natural communities.Aquat Microb Ecol 27: 103–110.

Weinbauer, M.G., Brettar, I., and Höfle, M.G. (2003) Lysog-eny and virus-induced mortality of bacterioplankton insurface, deep, and anoxic marine waters. Limnol Oceanogr48: 1457–1465.

Wells, L.E., and Deming, J.W. (2006) Significance of bacte-rivory and viral lysis in bottom waters of Franklin Bay,Canadian Arctic, during winter. Aquat Microb Ecol 43: 209–221.



Wikner, J., and Hagström, Å. (1988) Evidence for a tightlycoupled nanoplanktonic predator–prey link regulating thebacterivores in the marine environment. Mar Ecol Prog Ser50: 137–145.

Wilhelm, S.W., Brigden, S.M., and Suttle, C.A. (2002) A dilu-tion technique for the direct measurement of viral produc-tion: a comparison in stratified and tidally mixed coastalwaters. Microb Ecol 43: 168–173.

Williamson, S.J., Houchin, L.A., McDaniel, L., and Paul, J.H.(2002) Seasonal variation in lysogeny as depicted byprophage induction in Tampa Bay, Florida. Appl EnvironMicrobiol 68: 4307–4314.

Winget, D.M., Williamson, K.E., Helton, R.R., and Wommack,

K.E. (2005) Tangential flow diafiltration: an improved tech-nique for estimation of virioplankton production. Aquat

Microb Ecol 41: 221–232.Winter, C., Herndl, G.J., and Weinbauer, M.G. (2004) Dielcycles in viral infection of bacterioplankton in the NorthSea. Aquat Microb Ecol 35: 207–216.

Wommack, K.E., and Colwell, R.R. (2000) Virioplankton:viruses in aquatic ecosystems. Microbiol Mol Biol Rev 64:

69–114.Yentsch, C.S., and Menzel, D.W. (1963) A method forthe determination of phytoplankton, chlorophyll andphaeophytin by fluorescence. Deep-Sea Res 10: 221–231.



Paper II

Role of viruses in shaping bacterial

phylogenetic and functional

diversity in marine coastal waters

Summary – Paper II

P a g e | 63

Resumen

Papel de los virus en los cambios de diversidad filogenética y

funcional bacteriana en un sistema oligotrófico marino costero

Artículo enviado para ser publicado en The ISME Journal (Role of viruses in shaping bacterial

phylogenetic and functional diversity in an oligotrophic marine coastal waters)

Para evaluar los principales factores que modifican la diversidad filogenética y funcional

bacteriana en un sistema oligotrófico marino costero, se llevó a cabo un estudio durante dos años

(mayo 2005 – abril 2007), basado en un muestreo mensual en la Bahía de Blanes (Mediterráneo

NO). Los factores que se presumía podían tener un impacto eran: (1) temperatura y nutrientes,

junto con la concentración de clorofila a como indicador de biomasa fitoplanctónica y por tanto

fuente de carbono (llamados aquí “factores abióticos + Chl a”), y (2) la presión de los

depredadores (virus y protistas) como causa de mortalidad bacteriana. La diversidad filogenética

fue evaluada usando la técnica de DGGE (denaturing gradient gel electrophoresis) a partir de la

amplificación del gen 16S rRNA como cebador universal. La diversidad funcional fue evaluada

analizando la utilización de diversas fuentes de carbono en placas Biolog Ecoplates, y

determinando las actividades ectoenzimáticas. El dendrograma basado en el análisis de DGGE

correspondientes a diversidad filogenética, dio lugar a tres ramas principales. Dos de ellas se

diferenciaban significativamente en la temperatura del agua, concentración de nitratos, y en la

concentración de la fracción de Chl a < 3 μm. Tercera rama tenía una superior mortalidad

bacteriana debida a virus. Respecto a la diversidad funcional bacteriana, la depredación por

protistas parecía no tener ningún efecto sobre ella, debido a que no se observaron correlaciones

entre la mortalidad bacteriana causada por protistas y las actividades ectoenzimáticas o

utilización de las fuentes de carbono. En cambio, la mortalidad bacteriana causada por virus

estaba correlacionada con los cambios en porcentaje de la utilización de las fuentes de carbono

(por ejemplo D,L-a-Glycerol Phosphate, Glucose-1-Phosphate), y negativamente correlacionada


64 | P a g e

con la actividad de la aminopeptidasa y β-glucosidasa. Esto sugiere que los productos de la lisis

bacteriana podían ser una fuente de carbono fácilmente utilizable por las bacterias,

probablemente compuesto, entre otras moléculas, por aminoácidos y glucosa, que podrían

inhibir la actividad proteolítica y glucosídica. Los resultados indicaban que los factores abióticos,

como temperatura y nutrientes, junto con el fitoplancton como una fuente de carbono,

contribuían a modificar la estructura y la actividad de la comunidad bacteriana. En el caso de los

factores bióticos, los virus mostraron un papel más importante que los protistas en los cambios de

la diversidad filogenética y funcional bacteriana en el sistema estudiado.

The ISME Journal, submitted

Role of viruses in shaping bacterial phylogenetic and functional

diversity in oligotrophic marine coastal waters

Julia A. Boras1*, Dolors Vaqué1, Francesc Maynou1, Elisabet L. Sà1, Markus G. Weinbauer2, and M. Montserrat Sala1

1Institut de Ciències del Mar (CSIC), Passeig Marítim de la Barceloneta 37-49, 08003

Barcelona, Spain ; 2CNRS, Microbial Ecology and Biogeochemistry Group, Laboratoire d'Océanographie de Villefranche, 06230 Villefranche-sur-Mer, France; Université Pierre et

Marie Curie-Paris 6, Laboratoire d'Océanographie de Villefranche, 06230 Villefranche-sur-

Mer, France

To evaluate the main factors shaping the bacterioplankton phylogenetic and functional diversity in marine coastal waters, we carried out a two-year study (May 2005 – April 2007) based on a monthly sampling in Blanes Bay (NW Mediterranean). We expected that these factors would be (1) temperature and nutrients, together with the chlorophyll a concentration as an indicator of phytoplankton biomass and source of dissolved carbon (called here abiotic factors plus Chl a), and (2) predators (viruses and protists) pressure as a cause of bacterial mortality. Phylogenetic diversity was analyzed by denaturing gradient gel electrophoresis (DGGE) of 16S rRNA. Functional diversity was assessed both by the use of monomeric carbon sources in Biolog Ecoplates and the determination of 6 ectoenzymatic activities. The dendrogram based on the DGGE formed three main sample clusters. The main two clusters showed differences in temperature, nitrate and chlorophyll a concentration, and the third characterized by the highest losses of bacterial production due to viral lysis. Protistan grazing did not seem to have an effect on bacterial functional diversity, since no correlations between protist-mediated mortality (PMMBP) and ectoenzymatic activities were found, and utilization of only 2 out of the 31 carbon sources was correlated with PMMBP (N-acetyl-D-glucosamine and α-cyclodextrin). In contrast, virus-mediated mortality correlated with changes in the percentage of use of several carbon sources (e.g. D-mannitol and glucose-1-phosphate), and correlated negatively with aminopeptidase and β-glucosidase activity. This suggests that viral lysate provides to bacterioplankton a pool of labile carbon sources, presumably amino acids and glucose among them, which may inhibit proteolytic and glucosidic activity. Our results indicate that abiotic factors plus Chl a are the main factors that shape bacterial community structure and activity. Furthermore, they suggest a major role of viruses than of protists in modifying phylogenetic and functional diversity of bacteria in oligotrophic marine coastal waters.

Correspondence: Julia A. Boras, E-mail: [email protected]

Introduction

Prokaryotes constitute half of the total amount of carbon enclosed in living organisms on Earth, and represent the largest pool of nitrogen and phosphorous in the biosphere (Whitman et al., 1998). In aquatic environments, heterotrophic bacteria are an essential part of the food webs, being the major consumers of dissolved organic matter (DOM) in the ocean (Azam, 1998). The carbon and nutrients contained in bacterial cells can move up into the trophic web through grazing by protists, or return as DOM and particulate organic matter (POM) pools to the water through viral infections and subsequent cell lysis. Thus, viral activity provides an additional input of organic matter and nutrients to ocean waters that could be available for bacterioplankton. Although the process of nutrients recycling from bacterial cells to the environment can be quantitatively important, little is known on the composition of the bacterial lysate. Our present knowledge on that unknown pool of DOM reveals that lysis products are labile and turn over rapidly (Noble and Fuhrman, 1999), they may constitute an important source of P for marine bacterioplankton (Weinbauer et al., 1995, Riemann et al., 2009), and that the released organic carbon could be composed in average of 51 – 86 % by dissolved combined amino acids (Middelboe and Jørgensen, 2006).

DOM in aquatic systems is composed mostly by polymeric compounds and therefore has to undergo a previous hydrolysis before it could be taken up by bacterial cells. Hydrolysis is carried out by specific ectoenzymes, and their activity is an indirect indication of the polymeric molecules available for bacteria (Hoppe, 1983). Monomeric DOM is a preferable carbon source for bacteria, since it is easily taken up as it does not require the synthesis of specific ectoenzymes. The utilization of monomers is characteristic for each community, and changes in the utilization of sole carbon sources in marine samples have shown differences among contrasting coastal environments (Sala et al., 2005a), seasons

(Sala et al., 2006a), or depth and temporal patterns (Sala et al., 2008). Products of viral lysis provide an additional input of different types of monomers and polymers to the environment. Preferential use of these monomers might suppress the activity of specific bacterial ectoenzymes and modify the pattern of sole carbon sources utilization by the microbial community.

The phylogenetic composition of natural bacterial assemblages can be shaped by a variety of parameters and processes, such as environmental factors and predators’ pressure. Several studies have shown temporal changes in the dominance of particular bacterial groups in coastal (Pinhassi and Hagström, 2000; Ghiglione et al., 2005) and oceanic waters (Morris et al., 2005). These and other studies (Schauer et al., 2003; Alonso-Sáez et al., 2007) suggest that temperature or substrate availability shape bacterial diversity, and it was demonstrated that an addition of specific substrates can induce a succession of bacterial species (Pinhassi et al., 1999). Also, the variability in phytoplankton assemblages, as blooms or variation in species composition, can influence the shape of the bacterial community (Pinhassi et al., 2004; Ghiglione et al., 2005). On the other hand, activity of predators, protists and phages, can also produce changes in bacterial diversity. Grazing by protists was shown to impact the taxonomic structure of bacterial communities (Šimek et al., 1997), for example by selective grazing (Hahn and Höfle, 1999), indirectly by providing the substrates for bacterial growth (Caron et al., 1988), or by elimination of competitive strains. Bacteriophages can modify bacterial diversity in a variety of ways, e.g. by lysogenic conversion, transduction, resistance induction or by release of the lysis products to the environment (Weinbauer and Rassoulzadegan, 2004). It is also hypothesized that through differentially killing of bacteria that win the competition for resources, viruses increase or maintain bacterial diversity in the environment (“killing the winner” hypothesis;

Impact of viruses on bacterial diversity

3

Thingstad and Lignell, 1997; Thingstad, 2000). Microcosm and field studies showed a great variability of responses of the bacterial assemblage to the presence of viruses (Hewson and Fuhrman, 2006; Bouvier and del Giorgio, 2007). This shows, that the impact of viruses on bacterioplankton diversity is a complex process, which may depend on factors like the presence of binding sites (porins) on the bacterial cell wall (Lenski, 1988), resistance on viral infection (Weinbauer et al., 2007), or the quality of available DOM (Hewson and Fuhrman, 2006).

To identify the main factors shaping phylogenetic bacterial diversity, and regulating bacterial utilization of monomeric and polymeric carbon sources, we performed a two-year study in an oligotrophic coastal marine environment (Microbial Marine Observatory of Blanes Bay, NW Mediterranean). Our aim was to (1) follow changes in bacterial community structure along the year, (2) detect the main factors shaping bacterial phylogenetic and functional diversities, the latter based on ectoenzymatic activity and utilization of sole carbon sources, and (3) evaluate the role of viruses in shaping bacterial phylogenetic and functional diversity. We expected to find (1) changes in bacterial phylogenetic diversity caused by viral lysis and/or protistan grazing, (2) lower specific ectoenzymatic activities with higher viral mortality, (3) higher variability in the utilization of monomeric carbon sources with virus-mediated mortality, and a minor effect of protist-mediated mortality.


Study site and sampling strategy Surface water samples (0.5 m depth)

were collected from May 2005 to April 2007 in Blanes Bay, Spain (The Blanes Bay Microbial Observatory, NW Mediterranean, 41º 40’ N, 2º 48’ E, 20 m depth), at 0.5 miles from the shore. Samples were collected once a month, in 10 litter polyethylene carboys, and

kept in the dark until reached the lab (~2 h). Water temperature and salinity were measured in situ with a CTD (conductivity, temperature, depth).

Physicochemical and biological parameters

Detailed description of the determination of physicochemical and biological parameters is presented in Boras and colleagues (2009). Shortly, concentrations of inorganic nutrients (PO4

-3 and NO3-),

chlorophyll a (Chl a), and Chl a fraction smaller than 3 μm (Chl a < 3 μm) were determined using standard methods (Grasshoff et al., 1983 for inorganic nutrients; Yentsch and Menzel, 1963 for Chl a). Viral abundance was determined by flow cytometry as described in Brussaard (2004). Bacterial and heterotrophic nanoflagellate (HNF) abundances were obtained by epifluorescence microscopy (Olympus BX40), after staining with DAPI (Porter and Feig, 1980; Sieracki et

al., 1985). HNF were grouped into four size classes: ≤ 2 µm, 2 - 5 µm, 5 - 10 µm and 10 - 20 µm.

Losses of bacterial production due to protists (PMMBP) were evaluated following the FLB (fluorescent labeled bacteria; Spanish Type Culture Collection, http://www.cect.org/index2.html) disappearance method (Sherr et al., 1987; Vázquez-Domínguez et al., 1999), and losses due to viruses (VMMBP) were evaluated by the virus reduction approach (Weinbauer et

al., 2002; Wilhelm et al., 2002).

DGGE and phylogenetic analysis Nucleic acids extraction - Denaturing

gradient gel electrophoresis (DGGE) of polymerase chain reaction (PCR) amplified 16S rRNA gene fragments was used to compare bacterial communities from each sampling. To collect microbial biomass, 20 litters of seawater were filtered through a 3.0

m pore size filter (Millipore) and a 0.2 m Sterivex filter (Durapore, Millipore), immediately after arriving to the lab. The

Sterivex units were filled with 1.8 ml of lysis


4

buffer (40 mM EDTA, 50 mM Tris-HCl pH = 8.3, 0.75 M sucrose) and stored at temperature of -80 C. The nucleic acids were treated with lysozyme, proteinase K and sodium dodecyl sulphate, extracted with phenol and concentrated in a Centricon-100 (Millipore) (Massana et al., 1997). The quality of the DNA was checked by agarose gel electrophoresis.

DGGE gel analysis – The DGGE was carried out as previously described in Schauer and colleagues (2003). The PCR amplification of bacterial 16S rRNA was done using the bacterial specific primer GC358f and the universal primer 907rM. PCR products were verified by agarose gel electrophoresis with a standard in the gel (Low DNA Mas Ladder, Gibco BRL).

To perform the DGGE, the PCR product from each sample was loaded on 6 % polyacrylamide gels with a DNA-denaturant gradient ranging from 40 % to 80 %. The 24 samples were run in three DGGE gels, whereas identical samples were run in the extreme side of consecutive gels as standards to allow the comparison among them (Sep-05 and Sep’-05 for gels I and II, and Jun’-06 and Jun’’-06 for gels II and III). The DGGE gels were run at 100 V for 16 h at 60 ºC in 1x TAE running buffer using DGGE-2000 system (CBS Scientific Company), as previously described for Blanes samples (Alonso-Sáez et al., 2007). The gels were stained with 3 μl of SYBR Gold in 20 ml of TAE 1x, and analyzed using the Fluor-S MultiImager (Bio-Rad) with the Multi-Analyst software (Bio-Rad). Digitized DGGE images were analyzed using the Diversity Data software (Bio-Rad), described previously (Schauer et al., 2000; Díez et al., 2001). A matrix was constructed taking into account the relative contribution of each band (as percentage) to the total intensity of the band line. Ranked matrix of similarities among samples was constructed using the Bray-Curtis similarity measure (Bray and Curtis, 1957). Based on this matrix a

hierarchical clustering was performed, resulting in a dendrogram. Functional diversity

Functional diversity was assessed both by the utilization of carbon sources in the Biolog-Ecoplates and the activity of a set of selected ectoenzymes.

Biolog-Ecoplates - Microplates were used to assess the differences in the functional diversity of the bacterioplankton assemblage (Preston-Mafham et al., 2002). Each of the 96 wells of the microplate contains a carbon source (31 carbon sources, in triplicates). Together with the carbon source, tetrazolium violet is included in each well to indicate substrate catabolism. Each sample provides a unique pattern of utilization of carbon sources. After inoculation of 150 μl in each well, samples were incubated at room temperature in the dark during 6 days and then kept at -20 ºC until the measurement of the absorbance at 590 nm wavelength in a spectrophotometric microplate reader (ELX800 BIOTEK Instruments, Inc.Winooski, Vermont, USA). The mean color development of the three replicate wells for each substrate was calculated, and the mean absorbance of the blanks (with only water) was subtracted. Absorbance of each substrate was standardized to total absorbance of the plate to avoid effect of the inoculums’ size. Further details can be found in Sala and colleagues (2005b).

Ectoenzymatic activities - We determined the activity of six ectoenzymes: α-glucosidase, β-glucosidase, xylosidase, esterase, alkaline phosphatase and aminopeptidase using fluorogenic substrates: MUF-4-α-glucosidase, MUF-4-β-glucosidase, MUF-4-β-xylosidase, MUF-4-butyrate, MUF-4-phosphate, and Leucine 7-amido-4 methylcoumarin, respectively. We followed the methodology of Sala and colleagues (2001). Briefly, substrates were added at 100 μM final concentration to 0.9 ml samples in duplicates. Fluorescence of the sample was read immediately after addition of the


5

substrate and after an incubation time between 15 and 120 min, on a Shimadzu spectrofluorometer RF-540 at 365 nm excitation and 446 nm emission wavelengths. The increase of fluorescence units during the incubation was transformed to activity with a standard curve of the end product of the reactions. Activities of each ectoenzyme were standardized as the percentage of the highest activity of the sampling and a percentage of total activity. Statistical Analyses

Normal distribution of data was checked using the Shapiro-Wilk W test, and data were logarithmic transformed when necessary. Bacterial mortalities in percentages of bacterial production (VMMBP and PMMBP) were arcsin-square root transformed to achieve normality of distributions. Differences in values of parameters among clusters were tested using 1-way ANOVA in the case of normally-distributed parameters, and using Mann-Whitney U-test for non-normally distributed parameters. The Pearson correlation coefficient was used to determine the relationships between pairs of parameters. These statistical analyses were performed using the JMP program. Samples similarity matrices and a dendrogram were constructed and analyzed using the PRIMER v.6 (Plymouth Marine Laboratory) program. Similarity between pairs of data matrices (phylogenetic, based on Biolog, enzymatic, biotic and abiotic plus Chl a data) was checked by Mantel-type test (PRIMER v.6). Canonical correspondence analysis (CCA) was performed to evaluate multivariate patterns in the data, using the XLSTAT-ADA software. For abiotic factors plus Chl a, temperature, nutrient concentrations, Chl a and Chl a < 3 μm fraction were used to build the predictor (environmental) matrix, and bacterial apparent richness, diversity (as Shannon-Weaver index), and specific ectoenzymatic activities (activity per cell) were used to build the response (biological) matrix. The permutation test based on the F-

statistic (Legendre and Legendre, 1998) was used to assess the significance of the environmental matrix in explaining the patterns observed in the biological matrix. A second CCA consisted in using the biotic factors: abundances of virus and HNF, as well as VMMBP and PMMBP as predictor matrix, vs. the same response matrix as above.

Results

Physicochemical and biological parameters

Detailed data of changes of water temperature, inorganic nutrients and Chl a concentrations, abundances of microorganisms, and bacterial mortality due to viruses and protists along the study period are presented in Boras and colleagues (2009). Water temperature followed a seasonal trend, with the highest temperatures in summer (25.0 ± 1.0 ºC), and the lowest in winter (13.4 ± 1.2 ºC; Table 1). Among inorganic nutrients, only nitrate concentration showed significant seasonal changes, with the highest values in winter (1.49 ± 0.75 μM), and the lowest in summer (0.29 ± 0.17 μM; Table 1). Concentrations of Chl a and Chl a < 3 μm followed the same trend as nitrate concentration (Table 1). Abundance of viruses reached significantly higher values in spring (3.3 ± 1.5 × 107 viruses ml-1) than in the rest of the year (Table 1). Among HNF, the bacterivorous 2 - 5 μm size fraction reached the highest abundances in summer (7.4 ± 4.3 × 102 cells ml-1), and the lowest in autumn (1.8 ± 1.4 × 102 cells ml-1; Table 1). Virus-mediated mortality of bacteria (VMMBP) and protist-mediated mortality (PMMBP) over the first year of study reached average values of 12.3 ± 12.3 % d-1 and 33.9 ± 24.3 % d-1 respectively, and 40.9 ± 20.7 % d-1 and 32.4 ± 20.0 % d-1 over the second year. Bacterial community composition

Analysis of the DGGE gel yielded a total of 58 band positions (operational taxonomic units, OTUs) detected in 24 samples (data not shown). The number of

Table 1. Average seasonal values of selected physicochemical and biological parameters, as well as ectoenzymatic activities determined along a two-year study in Blanes Bay, NW Mediterranean (n = 24; SD: standard deviation). Chl a: chlorophyll a; HNF: heterotrophic nanoflagellates; VMMBP: virus-mediated losses of bacterial production (BP); PMMBP: protists-mediated losses of BP; OTUs: operational taxonomic units.

Parameter Average value ± SD

Spring Summer Autumn Winter

Water temperature (ºC) 17.7 ± 3.2 25.0 ± 1.0 18.0 ± 2.6 13.4 ± 1.2

NO3- (μM) 1.08 ± 1.23 0.29 ± 0.17 1.18 ± 1.01 1.49 ± 0.75

Chl a total (μg l-1) 0.97 ± 0.79 0.23 ± 0.12 0.59 ± 0.11 0.77 ± 0.38

Chl a < 3 μm fraction (μg l-1) 0.13 ± 0.07 0.11 ± 0.09 0.23 ± 0.07 0.30 ± 0.15

Viruses (× 107 viruses ml-1) 3.3 ± 1.5 1.5 ± 0.4 1.6 ± 0.6 1.7 ± 0.8

Bacteria (× 105 cells ml-1) 10.6 ± 3.7 9.0 ± 3.3 7.5 ± 1.5 6.4 ± 1.6

HNF (× 102 cells ml-1)

HNF < 2 μm

HNF 2 - 5 μm

HNF 5 - 10 μm

HNF 10 - 20 μm

3.91 ± 3.17

5.48 ± 5.01

0.62 ± 0.46

0.13 ± 0.08

3.78 ± 2.17

7.36 ± 4.31

0.45 ± 0.42

0.06 ± 0.06

1.97 ± 1.49

1.82 ± 1.36

0.23 ± 0.11

0.06 ± 0.08

3.21 ± 1.47

2.93 ± 1.08

0.25 ± 0.17

0.03 ± 0.03

VMMBP (% day-1) 23.5 ± 16.4 21.8 ± 14.7 36.7 ± 32.8 24.3 ± 23.2

PMMBP (% day-1) 29.2 ± 27.3 44.6 ± 26.9 35.9 ± 10.9 22.9 ± 16.9

OTUs 11 ± 2 15 ± 4 19 ± 5 20 ± 3

α-Glucosidase (nmol l-1 h-1) 10.2 ± 9.6 10.8 ± 18.1 5.8 ± 12.3 3.7 ± 5.7

β-Glucosidase (nmol l-1 h-1) 29.4 ± 24.3 5.9 ± 4.1 6.8 ± 6.4 6.6 ± 2.3

β-Xylosidase (nmol l-1 h-1) 16.9 ± 8.5 24.1 ± 21.5 34.6 ± 57.2 35.6 ± 54.7

Esterase (nmol l-1 h-1) 3186 ± 1360 2872 ± 1368 1978 ± 628 3404 ± 4653

Alkaline phosphatase (nmol l-1 h-1) 263.7 ± 202.0 194.7 ± 106.1 229.3 ± 268.4 153.4 ± 152.8

Aminopeptidase (nmol l-1 h-1) 1234 ± 907 283 ± 220 258 ± 156 261 ± 193

OTUs ranged between 8 (Jun 05 and Jul 05) and 27 (Dec 06) per sample. Twelve of the 58 detected bacterial taxa (21 % of total OTUs) were restricted to single samples (8 samples). Only one OTU (2 % of the bacterial phylotypes) was observed in all analyzed samples, along the two years. Samples analyzed on the same DGGE gel did not show a particularly high similarity between them. Samples used as standards for comparison of gels were similar in 75 % (Sep-05 and Sep-05’) and 85 % (Jun’-06 and Jun’’-06).

The number of bands per sample was significantly different between seasons (F3,20 = 9.0, p < 0.01), increasing from spring to winter from 11 to 20 bands. Also, significant differences were detected between the spring-summer and autumn-winter periods (F1,22 = 16.9, p < 0.01), with higher number of bands in the second period (20 ± 4 bands) than in the first one (13 ± 4 bands). No significant differences in OTUs number were found between two study years, or relationship with any biotic factor.


7

Hierarchical clustering of samples (Fig. 1) revealed two main clusters, which separated the samples collected in months with high water temperature (from April to October, “warm” cluster) from samples collected in months with low water temperature (November to March, “cold” cluster; Table 2) at a 35 % similarity level. Other parameters, like nitrate concentration and Chl a < 3 μm, as well as HNF abundance were significantly different among clusters (Table 2). One sample, from Nov 06, shared with the rest of the samples only 20 % of similarities and formed the third cluster. This sample showed significantly higher VMMBP compared to the other samples (F1,22 = 9.7, p < 0.01).

Functional diversity

The pattern of ectoenzyme activities changed along the annual cycles. Mean seasonal values of several ectoenzymes, like β-glucosidase, alkaline phosphatase and aminopeptidase were higher in spring than in the rest of the year (Table 1). Seasonal differences were significant for aminopeptidase (F1, 21 = 3.5, p < 0.05), which reached the highest mean activities in spring (1234 nM -1 h-1). Other activities, in contrast, showed the highest mean values in autumn or winter, such as β-xylosidase or esterase. Among the 31 carbon sources in the Biolog-Ecoplate, none showed significant differences among seasons.

Fig. 1. Dendrogram of the hierarchical clustering of 24 samples from Blanes Bay, NW Mediterranean, using group-average linking of Bray-Curtis similarities. Samples are grouped in three main clusters: the “cold” cluster, the “warm” cluster, and the sample from November 06. The bold line means the significantly different branches at the significance level of 5 %.

Table 2. Differences among selected parameters between two main clusters of samples obtained by hierarchical clustering of DGGE results, assessed by ANOVA analysis. Chl a: chlorophyll a; HNF: heterotrophic nanoflagellates; OTUs: operational taxonomic units.

Parameter

Differences among “warm” and “cold” cluster

F n p

Water temperature (ºC) 9.8 23 <0.01

NO3- (μM) 8.2 22 <0.01

Chl a (μg l-1) - - -

Chl a < 3 μm fraction (μg l-1) 5.9 23 <0.05

Viruses (× 107 viruses ml-1) - - -

Bacteria (× 105 cells ml-1) 5.1 23 <0.05

HNF (× 102 cells ml-1)

HNF 2 - 5 μm

HNF 10 - 20 μm

5.4

4.6

23

23

<0.05

<0.05

OTUs 12.0 23 <0.01

α-Glucosidase (nmol l-1 h-1) 5.5 24 <0.05

Alkaline phosphatase (nmol l-1 h-1) 4.5 24 <0.05

Aminopeptidase (nmol l-1 h-1) 4.8 24 <0.05

D,L –α-Glycerol phosphate (%) 4.7 22 <0.05

L-Arginine (%) 4.8 22 <0.05

L-Serine (%) 3.9 22 <0.05

Phenylethylamine (%) 3.9 22 <0.05

D-Mannitol (%) 3.7 22 <0.05

α-D-Lactose (%) 4.5 22 <0.05

γ-Hydroxybutyric acid (%) 5.9 22 <0.05

Itaconic acid (%) 5.6 22 <0.05

D-Malic acid (%) 4.1 22 <0.05

2-Hydroxy benzoic acid (%) 4.0 22 <0.05

Tween 80 (%) 6.1 22 <0.01

In contrast, comparison of

ectoenzymatic activities between the ‘cold’ and ‘warm’ clusters yielded significant differences between them: α-glucosidase, alkaline phosphatase, and aminopeptidase showed higher mean activity values in the samples of the ‘warm’ than of ‘cold’ cluster

(Table 2). Also significant differences in substrate utilization in the Biolog plates for 11 carbon sources among two main clusters were observed. Four of them were higher in the ‘warm’ than the ‘cold’ periods (L-arginine, L-serine, D-mannitol and D-malic acid), and seven in the ‘cold’ than in the ‘warm’ periods


9

(D,L-α-glycerol phosphate, phenylethylamine, α-D-lactose, γ-hydroxybutyric acid, itaconic acid, 2-hydroxy benzoic acid, and tween 80).

The VMMBP showed significant correlation with several percentages of ectoenzymatic activities: β-glucosidase (r = -0.426, n = 23, p < 0.05), aminopeptidase (r = -0.417, n = 23, p < 0.05), and β-xylosidase (r = 0.569, n = 23, p < 0.01). In contrast, none of the ectoenzymatic activities were significantly correlated with PMMBP. Out of the 31 substrates in the Biolog-Ecoplates, four showed correlations with VMMBP (glucose-1-phosphate, r = -0.508, n = 22, p < 0.02; L-asparagine, r = 0.601, n = 22, p < 0.01; I-erythritol, r = 0.542, n = 22, p < 0.01; D-mannitol, r = 0.481, n = 22, p < 0.05), and only two with PMMBP (N-acetyl-D-glucosamine, r = 0.492, n = 22, p < 0.03; and α-cyclodextrin, r = -0.449, n = 22, p < 0.05). Relation among parameters

The Mantel test showed a significant positive relationship between the matrix of phylogenetic (based on the DGGE) and functional (based on Biolog) data (r = 0.345, p < 0.01).

The CCA analysis showed relation among abiotic plus Chl a data and bacterial apparent richness (S), diversity (H') and ectoenzymatic activity (p = 0.05, Fig. 2A). The first ordination axis accounted for 45 % of the variance in the biological matrix, while the 2nd axis accounted for 30 %. The first axis was mainly related to a combination of Chl a, nitrate and phosphate, while the second axis was positively defined by Chl a < 3 μm and negatively by temperature. Chl a < 3 μm fraction was the main factor shaping bacterial richness and phylogenetic diversity, as well as the esterase excretion. Other ectoenzymes (aminopeptidase, β-glucosidase) activities were mainly determined by phosphate and nitrate concentrations. The α-glucosidase

activity was strongly determined by temperature.

The biotic variables showed less clear relationships with the evaluated parameters (Fig. 2B). The ordination was mainly determined by the first canonical axis, accounting for 69.76 % of the variance in the biological data. This first axis resulted from a combination of virus- and protist-mediated mortalities and abundance of HNF. The second axis was determined mainly by viral abundances, but explained a small proportion of the variance (16.79 %) The main factor that determined bacterial richness, phylogenetic diversity, and xylanase and aminopeptidase activity was VMMBP (Fig. 2B). Virus abundance strongly determined the second ordination axis, but as it accounted for a low proportion of the variance explained (16.79 %), species richness, phylogenetic diversity and most enzymatic activities are probably not determined by virus abundance.

Discussion

Method evaluation

The DGGE is a widely used fingerprinting method to evaluate the community composition of marine picoplankton. It is highly reproducible (Schauer et al., 2000; Díez et al., 2001), although it also has some caveats, as that the use of different primers significantly changes the obtained bacterial assemblage composition (Cardinale et al., 2004). In our study we used standard primers commonly used in other works, which allow the comparison of our results with other studies. We are also aware, that the bands obtained on the gel may not reflect the whole variability of bacterial diversity, mainly due to low in situ abundance of specific strains, or the presence of different sequences in one band (Bano and Hollibaugh, 2002).

Fig. 2. Results of the canonical correspondence analysis (CCA) showing multivariate patterns among the bacterial diversity data (apparent richness S, diversity as Shannon-Weaver index H’, and specific ectoenzymatic activities as activity per cell; response matrix), and (2A) abiotic data (temperature – Temp.; phosphate, nitrate, nitrite, and ammonium concentrations) plus chlorophyll a – Chl a, and Chl a < 3 μm concentrations (predictor matrix), and (2B) biotic data (virus- VMM, and protist-mediated losses of bacterial production PMM; abundance of viruses VA, and of heterotrophic nanoflagellates HNF; predictor matrix). Ectoenzymes: Est - esterase, Apa - alkaline phosphatase, Ama - aminopeptidase, Xyl - xylosidase, Agl – α-glucosidase, Bgl - β-glucosidase.


11

Although firstly conceived as a tool for bacterial identification, Biolog plates are a useful tool to characterize metabolic properties of whole microbial communities (Garland and Mills, 1991), with both inconveniences and benefits (see review in Preston-Mafham et al., 2002). One of the possible drawbacks is the lack of representativity of the used carbon sources in natural environments. We have reduced this problem using recently created Biolog-Ecoplates (Insam, 1997), which contain carbon sources more relevant for natural ecosystems, instead of the commonly used GN-plates. Another drawback is that results obtained with this method do not reflect the actual in situ utilization of carbon sources. We are aware that the observed activities are the potential metabolic responses to substrate addition, similar to miniaturized enrichment experiments. In spite of the disadvantages, this approach has been widely applied in the study of soil ecosystems, and lately has also provided information on marine bacterioplankton. It has shown, for instance, seasonal patterns of functional diversity of bacterioplankton in coastal ecosystems (Sala et al., 2005a, 2006a), contrasting patterns among coastal ecosystems of different trophic levels (Sala et al., 2006b), among different depths in the Arctic Ocean (Sala et al., 2008), and of different water masses in the Antarctic Ocean (Sala et al., 2005b).

For the estimation of ectoenzymatic activities, we used an approach based on the addition of fluorogenic substrates (Hoppe, 1983). The main caveat of the methodology is that it does not provide information on the actual enzymatic rates but on potential activities since saturating concentrations are used. However, this method is the most widely used for marine environments and has shown, for instance, clear gradients of activities from the Equator to Antarctica (Christian and Karl, 1995), and also in transect across the Indian Ocean (Misic et al., 2006). Additionally, we have increased the

number of activities determined (six enzymes) in order to get a pattern that, similarly to Biolog plates, would provide us an additional insight into the functional bacterial diversity.

Determination of sampling frequency is a fundamental factor in temporal studies. Previous studies in Blanes Bay showed that the time needed by a single bacterial OTU to appear in significant amounts in this system ranged from weeks to months (Schauer et al., 2003; Ghiglione et al., 2005). Thus, we considered that monthly sampling let us perform a relatively detailed picture of changes in bacterial community structure. Changes in bacterial diversity

Number of bands on the DGGE gels detected along this two-year study was changing gradually along the year, with clear distinction between high and low water temperature periods. Similar seasonality in bacterioplankton assemblage composition was found in other studies run in the same study site (Schauer et al., 2003; Alonso-Sáez et al., 2007). According to Schauer and colleagues (2003) the main factors that affect bacterial diversity in Blanes Bay along the year are the changing DOM supply and the temperature. These authors also suggest that top-down factors do not impact the bacterial diversity in this particular system. Furthermore, in the study of Alonso-Sáez and colleagues (2007), a succession between SAR11 and Roseobacter groups along seasons was found, which apparently was due to changing nutrient concentrations in the water. Results of our study suggest that abiotic factors, like water temperature and nutrients concentration, plus Chl a < 3 μm, determined phylogenetic richness and diversity of bacterioplankton. Low number of OTUs was detected during months with low Chl a < 3 μm, high temperatures, and low nitrate concentrations. The Chl a < 3 μm representing picophytoplankton biomass is the potential source of carbon for bacteria, as POM and DOM is released to the environment during


12

the lysis of autotrophic cells and sloppy feeding (Bird and Kalff, 1984; Vaqué et al., 1989). Activity of predators, viruses and protists, can also modify bacterial community structure. Previous studies showed that viral activity affects differently individual members of the same marine bacterial community (Winter et al., 2004), and that possibly rare bacterial groups are more susceptible to virus-induced mortality than the abundant ones (Bouvier and del Giorgio, 2007). It was also seen that viruses typically cause a reduction of the number of detected bacterial phylotypes (Schwalbach et al., 2004; Winter et al., 2004). Protists can change the taxonomic composition of bacterial communities as well, causing the development of rare bacterial strains (Suzuki, 1999) or the replacement of strains (Hahn and Höfle, 1999). Finally, Zhang and colleagues (2007) found that viral lysis and protistan grazing can act additively, substantially increasing the apparent richness of the studied bacterial communities in eutrophic marine waters. In our study no clear relationship among biotic factors and phylogenetic and functional diversity of bacteria was found. However, our results show (Fig. 2B) that predators’ activity, mainly virus-induced mortality, is more important in determining phylogenetic and functional diversity than overall virus abundance per se. Likewise, it was observed that viral activity can occasionally modify the composition of the bacterial community. This was clearly seen in the November 06 sample that formed the third cluster on the DGGE dendrogram, with a pattern of bands completely different compared with other samples. In this month the only factor that was significantly different from the rest of the samples was the VMMBP, which was the highest along the whole study period (84 % d-1; Boras et al., 2009). This suggests that in conditions of severe viral predation, the effect of phages on bacterial community can be more pronounced that the effect of abiotic factors plus Chl a.

Previous studies of functional diversity in Blanes Bay have shown clear seasonal patterns in the utilization of carbon sources in contrast to more eutrophic coastal stations (Sala et al., 2006a). In the present study, the pattern of utilization of carbon sources in the Biolog plates was correlated with the patterns of phylogenetic diversity, although this was not found in another region of the NW Mediterranean coast (Sala et al., 2005a). Similarly to previous findings in Blanes Bay (Alonso-Sáez et al., 2008) that showed peaks of β-glucosidase and aminopeptidase in spring and summer coinciding with high bacterial production, the highest mean values of several ectoenzymatic activities were found in spring, especially aminopeptidase activity. Those peaks coincided with quite high VMMBP values (20 - 40 % d-1; Boras et al., 2009). Ectoenzymatic diversity of bacteria in our study was found to be driven by abiotic factors, like nutrient concentrations plus Chl a < 3 μm. In contrast, biotic factors, i.e. viral and protozoan abundances and mortality induced by both predators, showed no significant relationship with ectoenzymatic activity of bacterial community. Nevertheless, it could be observed that VMMBP was the main factor determining the pattern of activities of xylosidase and aminopeptidase, while protistan grazing had no relation with ectoenzymes. Viral lysate has been suggested to be a significant source of P, at least via D-DNA, for marine bacteria (Weinbauer et al., 1995, Riemann et al., 2009), and in the NW Mediterranean viral infection has been suggested as a route for phosphate recycling (Noble and Fuhrman, 1999). Middelboe and colleagues (1996) found increased alkaline phosphatase and aminopeptidase in a model system of Vibrio and viruses, suggesting a degradation of polymeric DOM containing N and P molecules. However, our data does not provide evidence for a relationship between viral infection and alkaline phosphatase increase in spite of sampling in the generally


13

P-limited bacterioplankton area (Sala et al., 2002, Pinhassi et al., 2006). Furthermore, our results contrast with those of Middelboe and colleagues (1996), since a negative correlation was observed between VMMBP and aminopeptidase activity. This suggests that highly labile DOM in the lysate might be partly composed by amino acids.

Additionally to the role of ectoenzymes in the utilization of organic matter, Noble and Fuhrman (1999) postulated ectoenzymes as a defense mechanism against viruses since proteases could destroy the capsid protein of living particles. In months when aminopeptidase activity was higher, we found a significant lower viral mediated mortality. It is unclear if proteases would be produced for the purpose of resisting viruses or is just a beneficial side effect of the ectoenzymes synthesized to hydrolyse peptides, but our results provide slight evidence in that direction and further research is needed in order to clarify that point.

Analysis of the relationship between the utilization of carbon sources in the Biolog plates and bacterial mortality also showed more cases of correlation between substrate utilization and VMMBP than with PMMBP. Viral lysis provides to bacterioplankton a highly labile DOM (Noble and Fuhrman, 1999) of still relatively unknown composition. The correlations among the utilization of 4 substrates with virus-mediated mortality might be an indication that the changes in the utilization of carbon sources, which could be a response to the new input of DOM from lysed cells. Conclusions

We conclude that abiotic factors, as water temperature, nutrients availability, plus DOM sources represented by Chl a < 3 μm, shaped the phylogenetic, enzymatic and substrate utilization structure of bacterioplankton community in the studied oligotrophic system over the seasonal cycle. Among bottom-up factors, DOM sources were

the main parameter that influenced phylogenetic and functional diversity of bacteria. Among predators, mainly viruses caused changes in the analyzed processes, although no statistically significant relationship among the overall VMMBP and diversities was found. Nevertheless, under conditions of severe viral predation viral mortality may modify significantly the phylogenetic diversity of bacterioplankton, and the effect of phages on bacterial community can be more pronounced that those of abiotic factors. Acknowledgments Funding for this research was provided by PROCAVIR and MICROVIS (CTM2004-04404-CO2-00/MAR, CTM2007-62140 to D. Vaqué), and ÍCARO (200830I120 to M.M. Sala) projects, funded by the Spanish Ministry of Education and Science (MEC). J.A.B. work was supported by a PhD fellowship from the MEC (FPU grant). The authors thank to J.M. Gasol for his thoughtful comments, and to V. Balagué for her help in the molecular and digital analysis of the samples. Sampling, CTD data and chlorophyll a analyses were provided by colleagues involved in the Blanes Bay Microbial Observatory net. This is a contribution to the European Network of Excellence EurOceans. References

Alonso-Sáez, L., Balagué, V., Sà, E.L., Sánchez, O., González, J.M., Pinhassi, J. Massana, R., Pernthaler, J., Pedrós-Alió, C., and Gasol., J.M. (2007) Seasonality in bacterial diversity in north-west Mediterranean coastal waters: assessment through clone libraries, fingerprinting and FISH. FEMS Microbiol Ecol 60: 98-112.

Alonso-Sáez L., Vázquez-Domínguez, E., Pinhassi, J., Cardelús, C., Sala, M.M., Lekunberri, I., Unrein, F., Massana, R., Simó, R., and Gasol, J.M. (2008) Factors controlling the year-round variability in carbon flux through bacteria in a coastal marine system. Ecosystems 11: 397-40.


14

Azam, F. (1998) Microbial control of oceanic carbon flux: the plot thickens. Science 280: 694-696.

Bano, N., and Hollibaugh, J.T. (2002) Phylogenetic composition of bacterioplankton assemblages from the Arctic Ocean. Appl Environ Microbiol 68: 505-518.

Bird, D.F., and Kalff, J. (1984). Empirical relationship between bacterial abundance and chlorophyll concentration in fresh and marine waters. Can J Fish Aquat Sci 41: 1015-1023.

Boras, J.A., Sala, M.M., Vázquez-Domínguez, E., Weinbauer, M.G., and Vaqué, D. (2009) Annual changes of bacterial mortality due to viruses and protists in an oligotrophic coasta environment (NW Mediterranean). Environ Microbiol 11: 1181-1193.

Bouvier, T., and del Giorgio, P.A. (2007) Key role of selective viral-induced mortality in determining marine bacterial community composition. Environ Microbiol 9: 287-297.

Bray, J.R., and Curtis, J.T. (1957) An ordination of the upland forest communities of Southern Wisconsin. Ecol Monogr 27: 325-349.

Brussaard, C.P.D. (2004) Optimisation of procedures for counting viruses by flow cytometry. Appl Environ Microbiol 70:

1506-1513. Cardinale, M., Brusseti, L., Quatrini, P., Borin, S.,

Puglia, A.M., Rizzi, A., Zanardini, E., Sorlini, C., Corselli, C., and Daffonchio, D. (2004) Comparision of different primer sets for use in Automated Ribosomal Inetrgenic Spacer Analysis of complex bacterial communities. Appl Environ Microbiol 70:

6147-6156. Caron, D.A., Goldman, J.C., and Dennett, M.R.

(1988) Experimental demonstration of the roles of bacteria and bacterivorous protozoa in plankton nutrient cycles. Hydrobiologia 159: 27-40.

Christian, J.R., and Karl, D.M. (1995) Bacterial ectoenzymes in marine waters: Activity ratios and temperature responses in three oceanographic provinces. Limnol Oceanogr 40: 1042-1049.

Díez, B., Pedrós-Alió, C., Marsh, T.L., and Massana, R. (2001) Application of

denaturing gradient gel electrophoresis (DGGE) to study the diversity of marine picoeukaryotic assemblages and comparision of DGGE with other molecular techniques. Appl Environ Microbiol 67: 2942-2951.

Garland, J.L., and Mills, A.L. (1991) Classification and characterization of heterotrophic microbial communities on the basis of patterns of community-level sole-carbon-source utilization. Appl Environ Microbiol 57: 2351-2359.

Ghiglione, J.-F., Larcher, M., and Lebaron, P. (2005) Spatial and temporal scales of variation in bacterioplankton community structure in the NW Mediterranean Sea. Aquat Microb Ecol 40: 229-240.

Grasshoff, K., Ehrhardt, M., and Kremling, K. (1983) Methods of seawater analysis. Kiel: Verlag Chemie.

Hahn, M.W., and Höfle, M.G. (1999) Flagellate predation on a bacterial model community: Interplay of size-selective grazing, specific bacterial cell size, and bacterial community composition. Appl Environ Microbiol 65: 4863-4872.

Hewson, I., and Fuhrman, J.A. (2006) Viral impacts upon marine bacterioplankton assemblage structure. J Mar Biol Ass UK 86: 577-589.

Hoppe, H.-G. (1983) Significance of exoenzymatic activities in the ecology of brackish water – measurements by means of methylumbelliferyl-substrates. Mar Ecol Prog Ser 11: 299-308.

Insam, H. (1997) A new set of substrates proposed for community characterization of environmental samples. In Microbial communities. Functional versus structural approaches. Insam, H., and Rangger, A. (eds). Berlin Heidelberg: Springer Verlag, p. 261.

Legendre, P., and Legendre, L. (1998) Numerical Ecology. 2nd English ed. Elsevier.

Lenski, R.E. (1988) Dynamics if interactions between bacteria and virulent bacteriophage. Adv Microb Ecol 10: 1-44.

Massana, R., Murray, A.E., Preston, C.M., and DeLong, E.F. (1997) Vertical distribution and phylogenetic characterization of marine planktonic Archaea in the Santa Barbara


15

Channel. Appl Environ Microbiol 63: 50-56.

Middelboe, M., Jørgensen, N.O.G., and Kroer, N. (1996) Effects of viruses on nutrient turnover and growth efficiency of noninfected marine bacterioplankton. Appl Environ Microbiol 62: 1991-1997.

Middelboe, M., and Jørgensen, N.O.G. (2006) Viral lysis of bacteria: an important source of dissolved amino acids and cell wall compounds. J Mar Biol Ass UK 86: 605-612.

Misic, C., Castellano, M., Fabiano, M., Ruggieri, N., Saggiomo, V., and Povero, P. (2006) Ectoenzymatic activity in surface waters: A transect from the Mediterranean Sea across the Indian Ocean to Australia. Deep-Sea Res I 53:1517-1532.

Morris, R.M., Vergin, K.L, Cho, J.-C., Rappé, M.S., Carlson, C.A., and Giovannoni, S.J. (2005) Temporal and spatial response of bacterioplankton lineages to annual convective overturn at the Bermuda Atlantic Time-series Study site. Limnol Oceanogr 50: 1687-1696.

Pinhassi, J., Azam, F., Hemphälä, J., Long, R.A., Martinez, J., Zweifel, U.L., and Hagström, Å. (1999) Coupling between bacterioplankton species composition, population dynamics, and organic matter degradation. Aquat Microb Ecol 17: 13-26.

Pinhassi, J., and Hagström, Å. (2000). Seasonal succession in marine bacterioplankton. Aquat Microb Ecol 21: 245-256.

Pinhassi, J., Sala, M.M., Havskum, H., Peters, F., Guadayol, Ò., Malits, A., and Marrasé, C. (2004) Changes in bacterioplankton composition under different phytoplankton regimens. Appl Environ Microbiol 70: 6753-6766.

Pinhassi, J., Gómez-Consarnau, L., Alonso-Sáez, L., Sala, M.M., Vidal, M., Pedrós-Alió, C., and Gasol, J.M. (2006) Seasonal changes in bacterioplankton nutrient limitation and its effects on bacterial diversity in the NW Mediterranean Sea. Aquat Microb Ecol 44: 241-252.

Porter, K.G., and Feig, Y.S. (1980) The use of DAPI for identifying and counting aquatic microflora. Limnol Oceanogr 25: 943-948.

Preston-Mafham, J., Boddy, L., and Randerson, P.F. (2002) Analysis of microbial community functional diversity using sole-carbon-source utilization profiles – a critique. FEMS Microbiol Ecol 42: 1-14.

Riemann,L., Holmfeldt, K., and Titelman, J. (2009) Importance of viral lysis and dissolved DNA for bacterioplankton activity in a P-limited estuary, northern Baltic Sea. Microb Ecol 57: 286-294.

Sala, M.M., Karner, M., Arin, L., and Marrasé, C. (2001) Measurement of ectoenzyme activities as an indication of inorganic nutrient imbalance in microbial communities. Aquat Microb Ecol 23: 301-311.

Sala, M.M., Peters, F., Gasol, J.M., Pedrós-Alió, C., Marrasé, C., and Vaqué, D. (2002) Seasonal and spatial variations in the nutrient limitation of bacterioplankton growth in the Northwestern Mediterranean. Aquat Microb Ecol 27: 47-56.

Sala, M.M., Balagué, V., Pedrós-Alió, C., Massana, R., Felipe, J., Arin, L., Illoul, H., and Estrada, M. (2005a) Phylogenetic and functional diversity of bacterioplankton during Alexandrium spp. blooms. FEMS Microbiol Ecol 54: 257-267.

Sala, M.M., Arin, L., Balagué, V., Felipe, J., Guadayol, O., and Vaqué, D. (2005b) Functional diversity of bacterioplankton assemblages in western Antarctic seawaters during late spring. Mar Ecol Prog Ser 292:13-21.

Sala, M.M., Estrada, M., and Gasol, J.M. (2006a) Seasonal differences in functional diversity of bacterioplankton in contrasting environments of the Catalan coast (NW Mediterranean). Aquat Microb Ecol 44:1-9.

Sala, M.M., Pinhassi, J., and Gasol, J.M. (2006b) Estimation of bacterial use of dissolved organic compounds (DON) in aquatic ecosystems using Biolog plates. Aquat Microb Ecol 42: 1-5.

Sala, M.M., Terrados, R., Lovejoy, C., Unrein, F., and Pedrós-Alió, C. (2008) Metabolic diversity of heterotrophic bacterioplankton over winter and spring in the Coastal Arctic Ocean. Environ Microbiol 10: 942-949.


16

Schauer, M., Massana, R., and Pedrós-Alió, C. (2000) Spatial differences in bacterioplankton composition along the Catalan coast (NW Mediterranean) assessed by molecular fingerprinting. FEMS Microbiol Ecol 33: 51-59.

Schauer, M., Balagué, V., Pedrós-Alió, C., and Massana, R. (2003) Seasonal changes in the taxonomic composition of bacterioplankton in a coastal oligotrophic system. Aquat Microb Ecol 31: 163-174.

Schwalbach, M.S., Hewson, I., and Fuhrman, J.A. (2004) Viral effects on bacterial community composition in marine plankton microcosm. Aquat Microb Ecol 34: 117-127.

Sherr, B.F., Sherr, E.B., and Fallon, R.D. (1987) Use of monodispersed, fluorescently labeled bacteria to estimate in situ protozoan bacterivory. Appl Environ Microbiol 53: 958-965.

Sieracki, M.E., Johnson, P.B., and Sieburth, J.McN. (1985) Detection, enumeration and sizing of planctonic bacteria by image analyzed epifluorescence microscopy. Appl Environ Microbiol 49: 799-810.

Suzuki, M.T. (1999) Effect of protistan bacterivory on coastal bacterioplankton diversity. Aquat Microb Ecol 20: 261-272.

Šimek, K., Vrba, J., Pernthaler, J., Posch, T., Hartman, P., Nedoma, J., and Psenner, R. (1997) Morphological and compositional shifts in an experimental bacterial community influenced by protists with contrasting feeding modes. Appl Environ Microbiol 63: 587-595.

Thingstad, T.F., and Lignell, R. (1997) A theoretical approach to the question of how trophic interactions control carbon demand, growth rate, abundance and diversity. Aquat Microb Ecol 13: 19-27.

Thingstad, T.F. (2000) Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnol Oceanogr 45: 1320-1328.

Vaqué, D., Marrasé, C., Iñiguez, V., Alcaraz, M. (1989). Zooplankton influence on the phytoplankton-bacterioplankton coupling. J Plank Res 11: 625-632.

Vázquez-Domínguez, E., Gasol, J.M., Peters, F., and Vaqué, D. (1999) Measuring the

grazing losses of picoplankton: methodological improvements to the use of fluorescently labeled tracers combined with flow cytometry. Aquat Microb Ecol 20:

110-128. Weinbauer, M.G., and Rassoulzadegan, F. (2004)

Are viruses driving microbial diversification and diversity? Environ Microbiol 6: 1-11.

Weinbauer, M.G, Fuks, D., Puskaric, S., and Peduzzi, P. (1995) Diel, seasonal, and depth-related variability of viruses and dissolved DNA in the Northern Adriatic Sea. Microb Ecol 30: 25–41.

Weinbauer, M.G., Winter, C., and Höfle, M.G. (2002) Reconsidering transmission electron microscopy based estimates of viral infection of bacterioplankton using conversion factors derived from natural communities. Aquat Microb Ecol 27: 103-110.

Weinbauer, M.G., Hornák, K., Jezbera, J., Nedoma, J., Dolan, J.R., and Šimek, K. (2007) Synergistic and antagonistic effects of viral lysis and protistan grazing on bacterial biomass, production and diversity. Environ Microbiol 9: 777-788.

Whitman, W.B., Coleman, D.C., and Wiebe, W.J. (1998) Prokaryotes: the unseen majority. Proc Natl Acad Sci USA 95: 6578-6583.

Wilhelm, S.W., Brigden, S.M., and Suttle, C.A. (2002) A dilution technique for the direct measurement of viral production: a comparison in stratified and tidally mixed coastal waters. Microb Ecol 43: 168-173.

Winter, C., Smit, A., Herndl, G.J., and Weinbauer, M.G. (2004) Impact of virioplankton on archeal and bacterial community richness as assessed in seawater batch cultures. Appl Environ Microbiol 70: 804-813.

Yentsch, C.S., and Menzel, D.W. (1963) A method for the determination of phytoplankton, chlorophyll and phaeophytin by fluorescence. Deep-Sea Res 10: 221-231.

Zhang, R., Weinbauer, M.G., and Qian, P.-Y. (2007) Viruses and flagellates sustain apparent rechness and reduce biomass accumulation of bacterioplankton in coastal marine waters. Environ Microbiol 9: 3008-3018.

Paper III

Changes of bacterial mortality and

diversity as a response to viral and

nanoflagellate communities

interactions in an oligotrophic

marine coastal site

Summary – Paper III

P a g e | 85

Resumen

Cambios en la mortalidad y diversidad bacteriana como

respuesta a las interacciones entre las comunidades de virus y

nanoflagelados en un sistema oligotrófico marino costero

Artículo enviado para ser publicado en Environmental Microbiology (Changes of bacterial

mortality and diversity as a response to viral and nanoflagellate communities interactions in an

oligotrophic marine coastal site)

Para el estudio del efecto de los protistas sobre la infección vírica de las bacterias en un

sistema costero, se han llevado a cabo cuatro experimentos estacionales en microcosmos con

poblaciones naturales de bacterias de la Bahía de Blanes, Mediterraneo NO, durante el año 2006-

2007. También se han estudiado los cambios en la diversidad bacteriana causados por virus y el

tipo de interacciones producidas entre protistas y fagos. El diseño experimental contemplaba 4

tratamientos, con y sin nanoflagelados heterotróficos (HNF), y con y sin virus (virus y virus-

suprimidos). Para ello se utilizó la técnica de fraccionamiento de las muestras filtrando por 5.0 µm y

0.8 µm, y añadiendo virus activos o virus inactivados por exposición al calor (80 ºC, 30 min). Al

inicio de los experimentos, la producción vírica era significativamente más alta en los tratamientos

con presencia de HNF, que en ausencia de HNF (F1, 14 = 16.78, p < 0.01). En tres de cuatro casos, la

mortalidad bacteriana debida a virus era significativamente más alta en presencia de HNF que

en su ausencia. Esto sugería que había una interacción sinérgica entre los dos tipos de

predadores bacterianos. Así los virus se beneficiaban del mejor estado fisiológico de las bacterias

y/o de los cambios en la diversidad bacteriana causada por los protistas, a partir de la materia

orgánica excretada por ellos al depredar sobre bacterias y otros microorganismos. Se observó

una correlación significativa y positiva entre el número de bandas de DNA (que representan los

taxones bacterianos) y la abundancia vírica, así como una correlación significativa y negativa

entre las dos variables anteriores y la abundancia de HNF. En tres de cuatro experimentos se


86 | P a g e

observó una separación significativa (p = 0.05) entre los tratamientos de virus + HNF y los de virus-

suprimidos + HNF. Esto sugeriría que la aparente riqueza de especies bacterianas en el sistema

estudiado estaba favorecida por la presencia/actividad de los virus. Los resultados obtenidos

indican que la presencia de HNF incrementa la mortalidad bacteriana debida a las infecciones

víricas en la mayoría de los casos debido a un efecto de sinergia entre HNF y virus. Además, la

actividad de los fagos en los tratamientos en los que se añadieron virus activos propició un

incremento de la riqueza de especies de la comunidad bacteriana.

Submitted to Environmental Microbiology

Changes of bacterial mortality and diversity as a response to viral and nanoflagellate communities interactions in an oligotrophic marine coastal site Julia A. Boras

*, Evaristo Vázquez-

Domínguez, Elisabet L. Sà, M. Montserrat Sala, and Dolors Vaqué

Institut de Ciències del Mar (CSIC), Passeig Marítim de la Barceloneta 37-49, 08003 Barcelona, Spain Summary Four seasonal microcosm experiments were carried out in a marine coastal system (Blanes Bay, NW Mediterranean Sea) in order to evaluate the effect of protists on virus-mediated bacterial mortality, and on changes of bacterial diversity caused by viruses, as well as the type of interactions established between those two main predators of bacteria. For this purpose, experimental treatments were prepared with and without nanoflagellates (HNF), and with and without viruses (virus-enhanced and virus-suppressed). This was achieved by sample size fractionation through 5.0 µm and 0.8 µm, and addition of active and heat-inactivated viruses. In three out of four cases, virus-mediated mortality of bacteria was significantly higher in the presence of HNF than without them. This suggests synergistic interactions among viruses and protists, where viruses benefited from an increase of bacterial fitness and/or changes in bacterial species structure caused by grazers. Positive correlation between viral abundance *For correspondence. E-mail [email protected]; Tel: (+34) 932309500; Fax: (+34) 932309555.

and detected number of bacterial taxa was observed, and in three out of four experiments a significant separation of virus-enhanced + HNF from virus-suppressed + HNF samples on the dendrogram was found. This suggests that bacterial apparent richness in the studied system was enhanced by viruses. We conclude that the presence of HNF results in an increase of virus-mediated mortality of bacteria in most cases as an effect of synergy with viruses. Furthermore, viral activity in virus-enhanced treatments results in an increase of richness of the bacterial community. Introduction

In aquatic environments, bacterial losses due to viral lysis are often similar to that caused by protists (Fuhrman and Noble, 1995; Boras et al., 2009), or even higher in some systems (Guixa-Boixereu et al., 1999; Wells and Deming, 2006). Dominance of grazing or viral lysis is crucial for carbon and nutrient fluxes in the ecosystem. While grazing shifts the organic carbon up through the trophic web, viral lysis releases dissolved organic matter (DOM), that can be taken up again by heterotrophic bacteria and reduce sinking of the organic matter to the deeper layers of the ocean (Fuhrman, 1999).

Predators, both viruses and protists, are able to change the composition of the bacterial community. Viral effect on bacterial diversity was described by the “killing the winner” hypothesis (Thingstad and Lignell,

J. A. Boras et al.

2

1997). This hypothesis predicts that the bacterial strain that wins the competition for nutrients will increase its abundance, and consequently the contact rate between members of this strain and their phage will also increase. This would result in higher infection rates (Murray and Jackson, 1992) decreasing abundance of the winning competitor, and allowing the development of other, less abundant taxa. Some microcosms and field studies support this theory (Bouvier and del Giorgio, 2007; Weinbauer et al., 2007), yet other works showed a great variability of responses of the bacterial assemblage to the presence of viruses (Hewson and Fuhrman, 2006). Activity of protists, mainly heterotrophic nanoflagellates, can also modify the bacterial community structure. Selective grazing (Hahn and Höfle, 1999), production of the additional substrates for bacterial growth (Caron et al., 1988), or elimination of competitive strains can change the taxonomic structure of bacterial communities (Šimek et al., 1997).

Bacterial response (i.e. changes in bacterial production, mortality rates, etc.) on the presence of grazers and viruses and their mutual relationship is an additional element describing the functioning of the microbial loop. Intuitively, an antagonistic relationship between phages and protists would reflect a competition for prey. Thus, some studies showed a substantial decrease of virally caused bacterial mortality in the presence of grazers (Maranger et al., 2002). In contrast, in other studies higher production of viruses and higher losses due to viral lysis were found in the presence of grazers than without them, indicating a synergistic interaction (Sime-Ngando and Pradeep Ram, 2005; Weinbauer et al., 2007). Finally, no relationship between viral and protistan activities, reflected in similar bacterial response on the presence of viruses and protists, was found by Horňák and colleagues (2005).

Interactions among phages and grazers can be crucial also for the structure of the bacterial community. Theoretical predictions attribute mainly to viruses the control of bacterial diversity, and to protists the control of prokaryotic abundance and biomass (Thingstad, 2000). Recent studies showed, however, that those two predators could induce more changes in bacterial diversity and richness acting synergistically, than each of them separately (Šimek et al., 2007; Zhang et al., 2007). The type of interactions in the triangle viruses-bacteria-protists is important for evaluation and prediction of the organic matter fluxes within the microbial loop in aquatic systems (Miki and Jacquet, 2008). Up to now, only two studies evaluate the effect of those interactions on bacterial diversity in marine systems (Zhang et al., 2007; Bonilla-Findji et al., 2009). Studies that assess changes in bacterial mortality due to both predators refer exclusively to freshwater environments (Šimek et al., 2001; Weinbauer et al., 2003; Weinbauer et al., 2007). In order to fill this gap, we conducted four seasonal experiments with microbial communities from an oligotrophic coastal marine system (NW Mediterranean). The main objectives of our study were: (1) to evaluate at what extent the presence/absence of protists can modify the mortality of bacteria due to viral infections; (2) to assess changes in bacterial community structure due to viral activity; and (3) to determine the character of bacteriophages-protists interactions and the effect of these interactions on bacterial mortality and richness. We hypothesize that the presence of protists could enhance viral infections due to an increase of bacterial production, and hence a synergistic interaction between viruses and grazers would exist. Based on the “killing the winner” theory, we also expected that richness of the bacterial community would increase with enhanced viral abundance.

Mortality and diversity of bacteria

3

Table 1. Pearson correlation analysis among variables for the four seasonal experiments. HNF: heterotrophic nanoflagellates, BP: bacterial production, VPL: lytic viral production, VPLG: lysogenic viral production, VMMBSS: virus-mediated mortality of bacteria as a % of bacterial standing stock, OTUs: bacterial operational taxonomic units. T0: correlation detected among variables at T0 of the experiments, n: number of data, r: correlation coefficient, p: significance level.

Variables n r p

Viruses vs. Bacteria Viruses vs. HNF Bacteria vs. HNF BP vs. HNF (T0) VPL vs. Bacteria (T0) VPL vs. VPLG (T0) VPLG vs. BP (T0) VMMBSS vs. VPLG (T0) OTUs vs. Viruses OTUs vs. HNF OTUs vs. Bacteria (T0) OTUs vs. BP (T0) OTUs vs. VPL

122 48 45 16 15 16 16 32 32 16 15 16 16

0.557 -0.471 -0.602 0.546 0.862

-0.505 0.572

-0.408 0.407

-0.681 0.566

-0.596 0.608

< 0.01 < 0.01 < 0.01 < 0.05 < 0.01 < 0.05 < 0.05 < 0.05 < 0.05 < 0.01 < 0.05 < 0.02 < 0.02

Results

Changes in microorganisms’ abundances

Values of in situ temperature, chlorophyll a (Chl a) concentration, and microorganisms abundances are shown in Table 1A (Supporting Information). Temperature varied from 13.5 ºC in spring, coinciding with the maximal values of viral abundance and Chl a, to 23.7 ºC in summer, when the maximal values of bacteria and HNF abundances were observed. Detailed information about monthly and seasonal in situ variability of physical and biological variables in the studied system can be found in Boras and colleagues (2009). The Chl a concentrations decreased during all experiments, although these changes were not statistically significant (data not shown). The preparation of the experimental setup (by size fractionation)

led to differences in the proportion of viruses and bacteria in the samples. Higher virus-bacterium ratio (VBR) in the 0.8 μm treatments at T0 of the experiments (range of 13 - 38) than in the 5 μm treatments (range of 5 – 19; F1,

13 = 6.59, p < 0.05) was observed. Abundances of microorganisms in different treatments of the four experiments are shown in Table 2A (Supporting Information). In autumn and winter, viral abundance increased after 6 - 24 h of incubation in all treatments, and declined (autumn) or did not changed (winter) at T48. In spring and summer, viral abundance augmented slightly at 24 - 48 h, mainly in virus-enhanced treatments. Abundance of bacteria increased in all treatments during autumn experiments from 6 h to 48 h of incubation, while in spring the increment occurred at 24 - 48 h. Finally, only in 0.8 μm treatments, increase of bacterial

J. A. Boras et al.

4

Fig. 1. Mean bacterial production detected at the beginning (T = 0) and at the end (T = 48h) of the experiments. B: bacteria, VB: viruses+bacteria, FB: HNF+bacteria, VFB: viruses+HNF+bacteria treatments; HNF: heterotrophic nanoflagellates. abundance was detected from 24 h to 48 h in summer, and from 6 h to 48 h in winter (Table 2A). The HNF achieved its maximal abundances in summer and winter (24 – 48 h of incubation, Table 2A), while in spring and autumn a negative growth was observed after this incubation time. Pooling all data from all experiments we detected a significant positive correlation between viral and bacterial abundance, and negative significant correlations between both of these variables and HNF abundance (Table 1).


Bacterial production (BP) ranged between 0.21 × 106 and 5.85 × 106 cells ml-1 d-1 at the beginning of the experiments, and between 0.13 × 106 and 5.59 × 106 cells ml-1 d-1 at the end of incubations (Fig. 1). The lowest BP values were observed in the spring, and the highest in the summer experiments (F3, 28 = 17.15, p < 0.01). No differences between treatments were found, except in summer, where higher BP was found in 0.8 μm than in 5 μm treatments (F3,4 = 236.30, p < 0.01). Positive relationship between BP and HNF abundance was


5

Fig. 2. Lytic and lysogenic viral production detected (A) at the beginning and (B) at the end of the four seasonal experiments. In bold: the percentage of lysogeny of the total viral production. B: bacteria, VB: viruses+bacteria, FB: HNF+bacteria, VFB: viruses+HNF+bacteria treatments; HNF: heterotrophic nanoflagellates. Bars indicate minimal and maximal values of duplicates. detected at the beginning of the experiments (Table 1).

Viral production

Lytic viral production (VPL) varied between 7.65 × 105 and 5.33 × 107 viruses ml-1 d-1 at the beginning of the experiments (Fig. 2), and decreased significantly in most of treatments at T48 (F1, 30 = 4.28, p < 0.05), being not detectable in the viruses+HNF+bacteria (VFB) treatment in summer at T48. However, for the bacteria (B) treatment in summer and the viruses+bacteria (VB) treatments in summer, autumn and winter no significant decrease of VPL was observed. At T0 of the experiments, VPL was significantly higher in 5 μm than in 0.8 μm treatments (F1, 14 = 16.78, p < 0.01). This trend was not observed at the end of the experimental time. No differences in VPL between virus-enhanced and virus-suppressed treatments were observed. Statistically significant correlation between VPL and

bacterial abundance was detected at the beginning of the experiments (Table 1).

Lysogenic viral production (VPLG) ranged between not detected and 93.46 % of total viral production (Fig. 2), and was detected in 63 % of cases at T0, and in 69 % at T48. The VPLG expressed in viruses produced ml-1 d-1 was significantly higher at the end of experiments than at the beginning (F1, 30 = 5.50, p < 0.05). No differences in VPLG between 0.8 μm and 5 μm treatments were detected, while it was significantly higher in virus-suppressed than in virus-enhanced treatments at the end of the experiments (F1, 14 = 4.95, p < 0.05). At T0 of the experiments, VPLG was negatively correlated with VPL, and positively with BP (Table 1). Bacterial mortality

Virus-mediated mortality of bacteria expressed as losses of bacterial standing stock (VMMBSS), varied between 3.14 and 34.23 % d-1 at T0, and

J. A. Boras et al.

6

Fig. 3. Virus-mediated mortality (VMM) as the percentage of losses of bacterial standing stock (BSS) at the beginning (T0) and at the end (T48) of four seasonal experiments. B: bacteria, VB: viruses+bacteria, FB: HNF+bacteria, VFB: viruses+HNF+bacteria treatments; HNF: heterotrophic nanoflagellates. Bars indicate minimal and maximal values of duplicates.

between no detected and 27.02 % d-1 at T48 (Fig. 3). The VMMBSS was significantly higher at T0 than at T48 (F1, 30 = 6.28, p < 0.02). Also, at T0 significantly higher VMMBSS was observed in 5 μm than in 0.8 μm treatments (F1, 14 = 5.67, p < 0.05), yet this trend was not observed after 48 h of incubation. No differences in VMMBSS were observed between virus-enhanced and virus-suppressed treatments, except in autumn, where significantly higher VMMBSS in virus-enhanced treatments was detected at T48 (F1, 2 = 195.66, p < 0.01). Significant negative relationship between VMMBSS and VPLG was observed at T0 of the experiments (Table 1).

Bacterial phylogenetic diversity

Analysis of denaturing gradient gel electrophoresis (DGGE) gels gave from 10 (summer) to 22 (spring) different band positions, or operational taxonomic units (OTUs; Table 2). Number of bands decreased significantly during the incubation time (F1, 30 = 5.68, p < 0.05), being this trend more pronounced in the spring and winter experiments (Table 2). At T0, the OTUs number was significantly higher in 5 μm than in 0.8 μm treatments (F1, 14 = 5.84, p < 0.05), yet this was not observed at T48. This decrease was more pronounced in virus-supressed treatments (Table 2). Moreover, in three cases an increase on total number of OTUs was observed, all in virus-


7

Figure 4. Dendrograms for hierarchical clustering of samples from four seasonal experiments, using group-average linking of Bray-Curtis similarities. The bold line means the significantly different branches at the significance level of 5 %. The dashed squares indicate separation of virus-enhanced and virus-suppressed samples at T48. In bold: T0 samples, “a” and “b” means experimental duplicates at T48. B: bacteria, VB: viruses+bacteria, FB: HNF+bacteria, VFB: viruses+HNF+bacteria treatments; HNF: heterotrophic nanoflagellates.

enhanced treatments (Table 2). Number of OTUs was significantly positively correlated with viral abundance and negatively with HNF abundance (Table 1). Additionally, at T0 significant relationships with bacterial abundance, BP, and VPL were detected (Table 1). Hierarchical clustering of samples (Fig. 4) showed significant (at p = 0.05) separation of the 0.8 μm and 5 μm treatments, at similarity levels from 55 % (winter) to 85 % (autumn) at T0, and from 55 % (winter) to 75 % (summer) at T48. The exception was the autumn experiment, where the separation was less clear. Virus-enhanced and virus-suppressed treatments did not differ significantly at T0 of the experiments. However, at T48

the VFB and the HNF+bacteria (FB) treatments formed separated clusters (p = 0.05) in all experiments except in spring. Experimental replicates were clustered together, indicating reproducibility of the treatments in each experiment. Discussion

Evaluation of used methodology

Experimental setup was based on the size fractionation technique (5.0 μm and 0.8 μm filtration) and on the addition of active vs. heat inactivated viral concentrate to pre-filtered water samples, which contained natural populations of virioplankton. Size

J. A. Boras et al.

8

Table 2. Number of bacterial operational taxonomic units (OTUs) detected in samples from different treatments at the beginning (T0) and at the end (T48) of four seasonal experiments. Values presented as mean ± standard deviation. B: bacteria, VB: viruses+bacteria, FB: HNF+bacteria, VFB: viruses+HNF+bacteria treatments; HNF: heterotrophic nanoflagellates. Δ OTUs: difference (increase or decrease) of bacterial bands number between T0 and T48. Bold numbers indicate positive differences.

Experiment Treatment T0 T48 Δ OTUs

Spring B VB FB VFB

17 ± 0.0 15 ± 0.0 20 ± 0.0 22 ± 0.0

11 ± 0.0 13 ± 0.0 17 ± 0.0 18 ± 0.0

- 6 - 2 - 3 - 4

Summer B

VB FB VFB

13 ± 0.0 10 ± 0.0 15 ± 0.0 15 ± 0.0

13 ± 0.0 12 ± 0.0 11 ± 0.7 16 ± 0.0

0 + 2 - 4 + 1

Autumn B VB FB VFB

16 ± 0.0 13 ± 0.0 18 ± 0.0 18 ± 0.0

14 ± 0.7 14 ± 0.7 11 ± 1.4 15 ± 0.7

- 2 + 1 - 7 - 3

Winter B VB FB VFB

16 ± 0.0 15 ± 0.0 15 ± 0.0 16 ± 0.0

11 ± 0.0 10 ± 1.4 11 ± 2.1 13 ± 1.4

- 5 - 5 - 4 - 3

fractionation is a broadly used technique to evaluate bacterial growth limitations (Fuhrman and Azam, 1980), and to assess the trophic cascade effects (Calbet et al., 2001). The main drawback of this method is the loss of cells during filtration, which was higher in the 0.8 μm filtrate than in the 5 μm filtrate (Table 1A and Table 2A). Consequently, virus-bacterium ratio was higher in the 0.8 μm than in the 5 μm treatments; however, our results suggest that evaluated processes were not affected by this difference. Also, due to filtration the number of observed OTUs was higher in the 5 μm than in the 0.8 μm treatments at T0, but this difference diminished at the end of the experiments and so allowed the observation of the effect of predators.

The collected seawater used in

the experiments was handled similarly, i.e. storage in carboys, filtration, virus addition and sample confinement in microcosms. Although we are aware of bottle effect, we assume that bacterial responses observed in the different treatments are comparable.

Estimations of bacterial mortality due to viruses are subjected to the calculation of the burst size. We estimated the BS taking into account the increase of viral abundance in viral production experiments, and the decrease of bacterial abundance in the same period of time (2 h). We assumed that the only cause of bacterial abundance decrease over short time periods was cell lysis, and we did not take into account the viral decay and bacterial production during this time


9

interval. Burst size values reported in this study, between 15 and 90, is within the range obtained from different aquatic environments (Parada et al., 2006).

Virus reduction approach used to evaluate VP and VMM is based on the assumption, that viral production observed during the incubation comes from infections prior to the incubation, that no new infections occur, and that the filtration and incubation do not induce lysogenic bacteria. This approach is considered one of the best-suited incubation-based methods for estimations of viral production and bacterial mortality due to viral lysis, is widely used and allows the comparison of the obtained results with other studies (Winget et al., 2005). Bacterial losses caused by lysis measured in our study are the potential losses, as we did not consider grazing on infected cells by protists in our calculations. This process could reduce the percentage of bacteria that burst due to viral activity in natural communities.

Finally, detection of lysogeny is based on lysis induction by mitomycin C. It is known, that in some circumstances (i.e. nutrients availability, pH) this agent is not sufficient to induce the lytic cycle in all prophages (Weinbauer and Suttle, 1999, and references therein). However, this method is widely used and mitomycin C is better suited than other inducing agents, hence the obtained results are comparable to other studies. Bacterial losses due to viral infections

Bacterial mortality caused by phages found in this study was within the range of in situ bacterial losses detected earlier in the same sampling station (Boras et al., 2009). Lytic VP and VMM were enhanced in the presence of protists in most cases, as summarized on Fig. 5, where values for VFB compared with VB were overall above the 1:1 line, indicating synergy among phages and protists. Similar findings were reported in other studies (Table 3).

Higher abundance of viruses (Weinbauer et al., 2003; Zhang et al., 2007) and higher rates of viral infection (Šimek et al., 2001) in the presence of protists were detected in marine and freshwater environments. This relationship among protists and viruses could be caused by DOM enrichment due to an egestion of organic matter and/or sloppy feeding by grazers, which can stimulate cell-specific production of bacteria (Posch et al., 1999). Higher bacterial production could, in turn, promote viral infection and proliferation (Weinbauer et al., 2003). Indeed, in three out of four experiments in our study, higher, although no significantly, bacterial production was observed in the presence of protists. However, it is noteworthy that in summer bacterial production achieved the highest values for all experiments in the 0.8 μm treatments, indicating that in this season bacterial losses were mainly due to protists. This is in agreement with the rate of bacterial mortality found in July 2006 at the same sampling station (Boras et al., 2009). Increase of viral production in grazers’ presence could be also caused by changes in phylogenetic composition of bacterial community due to grazing. It was proposed, that grazing enhances growth of phylotypes that are capable of rapid growth, or resistant to grazing but less resistant to viral infection (Weinbauer et al., 2003). Actually, some bands were detected in the 5 μm treatments that were not observed in the 0.8 μm mesocosms (data not shown). Production of DOM by grazers could favour an induction of the lytic cycle of prophages, as higher metabolic status of the host leads to more lytic rather than lysogenic infections (Williamson et al., 2002). Consequently, less lysogenic viral production in the presence of protists should be found (Weinbauer et al., 2003). Yet, we did not find significant differences in lysogeny among treatments with and without grazers in our study. However, we detected significantly higher lysogenic viral

J. A. Boras et al.

10

Table 3. Type of interaction among protists and bacteriophages observed in different aquatic systems. Enhancement of viral abundance and/or virus-mediated mortality of bacteria (VMM) in presence of heterotrophic nanoflagellates (HNF) reflects a synergistic interaction (+), whereas a decline of those parameters reflects an antagonistic interaction (-). – HNF: without HNF, + HNF: with HNF, VRA: virus reduction approach, VMMBSS: VMM as a % of bacterial standing stock, TEM: transmission electronic microscopy, FVIC: frequency of visibly infected cells, FIC: frequency of infected cells, nd: not determined.

System and Site Viral abundance (107 ml-1) VMM (%) VMM method Interaction Reference

- HNF + HNF - HNF + HNF

Marine

China: Clear Water Bay 6.0 7.5

7.0 7.5

nd nd

nd nd

nd + +

Zhang et al., 2007

Victoria Harbour 6.0 6.0

7.0 6.0

nd nd

nd nd

+ +

Peng Chau 8.0 75.0

12.5 55.0

nd nd

nd nd

+ +

Mediterranean Sea (Spain) 2.2 1.3 1.4 0.7

3.6 1.6 1.5 0.9

17.0 4.4 7.0

10.2

11.8 34.2 13.6 61.8

VRA (VMMBSS) - + + +

This study

Freshwater Římov Reservoir (Czech Republic) 1.7 4.7 22.0 37.0 TEM (FVIC) + Šimek et al., 2001 Furuike Pond (Japan) 20.0

10.0 5.0

0.3 0.1 0.1

nd nd nd

nd nd nd

nd - - -

Manage et al., 2002

Lac Cromwell (Canada) 12.0 8.0

20.0 16.0

11.0 9.0

19.0 14.0

nd nd nd nd

nd nd nd nd

nd - - - -

Maranger et al., 2002

Římov Reservoir (Czech Republic) 3.7 10.0

5.0 13.0

2.0 3.2

4.0 5.2

TEM (FVIC) + +

Weinbauer et al., 2003

Římov Reservoir (Czech Republic) 5.0 6.1

6.3 4.4

1.5 0.8

1.8 0.8

TEM (FVIC) No effect No effect

Horňák et al., 2005

Sep Reservoir (France) 1.3

1.6

5.0 15.0 TEM (FIC) + Sime-Ngando and Pradeep Ram, 2005

Římov Reservoir (Czech Republic) 10.0 7.0 18.0 39.0 VRA (FIC) + Weinbauer et al., 2007


11

production in virus-suppressed than in virus-enhanced treatments. This was also reflected in changes of viral abundance. Increase of viral concentrations and low lysogenic viral production was observed in virus-enhanced treatments, while a decrease of viral abundance and significantly higher lysogeny was detected in virus-suppressed treatments. This, together with the observed negative correlation between lytic and lysogenic viral production, suggests that an increase of viral abundance promoted a release of DOM from lysed cells, which in turn might have enhanced bacterial fitness and induced a lytic cycle in lysogens. However, considering all experiments, we did not observe a significantly higher VPL or bacterial mortality due to viruses in virus-enhanced treatments, as expected from the above outlook. Even so, at the end of the autumn experiment we detected significantly higher mortality in virus-enhanced treatments, and similar, although not significant, relationship at T0 of the summer and winter experiments.

Effect of phages on bacterial community structure

Bacteriophages can modify bacterial diversity in a variety of ways, e.g. by lysogenic conversion, transduction, resistance induction or by release of the lysis products to the environment (Weinbauer and Rassoulzadegan, 2004). It is also hypothesized that through differentially killing of bacteria that win the competition for resources, viruses increase or maintain bacterial diversity in the environment (“killing the winner” hypothesis; Thingstad and Lignel, 1997; Thingstad, 2000). In our study we found two arguments that support the “killing the winner”. The first argument is based on the positive correlation between viral abundance and the detected number of bacterial taxa, and on the negative correlation between protistan abundance and OTUs number. This suggests that

indeed, viral presence might enhance bacterial apparent richness. Phages, removing infection sensitive or/and abundant bacterial strains, promote growth of other, prior less abundant strains, that perhaps in their low abundances are not detected on the DGGE gel and become visible after reaching higher abundances. In fact, few bands were detected that were specific to virus-enhanced or suppressed treatments (data not shown). These bands could represent bacterial phylotypes resistant to viral infection and/or those which have higher affinity to lysis products than other phylotypes (Weinbauer et al., 2007; Zhang et al., 2007). Negative correlation of protistan abundance with the number of bacterial taxa could be an effect of grazing of protists. It is believed, that grazers’ activity reduce the abundance of all bacterial taxa in a less selective way than phages (Pernthaler, 2005), which could lead to the decrease of the apparent richness of bacterial community, if as a result of grazing some strains would reach very low, not detectable abundances. The second argument supporting the “killing the winner” hypothesis is grounded on the effect of active viruses’ addition on bacterial community structure. At the end of experiments, a separation of VFB treatments from FB treatments on the dendrogram was observed (Fig. 4). Those two groups of samples were significantly different among each other in three out of four experiments. Also, the increase of OTUs number was detected only in virus-enhanced treatments. Moreover, within experiments where decrease of bands number was observed between T0 and T48 in all treatments, the lowest reduction of number of OTUs occurred also in virus-enhanced treatments (Table 2). This evidence suggests the importance of viruses in shaping the bacterial community composition. This process could be intensified by the presence of protists, as was suggested by some authors (Sime-

J. A. Boras et al.

12

Ngando and Pradeep Ram, 2005; Zhang et al., 2007).

Interactions among viruses and protists

In our study a synergistic type of interactions among protists and viruses was observed (Fig. 5). Viral production and consequent virus-mediated mortality of bacteria were higher in the presence of protists. Also, bacterial production was enhanced in most treatments where grazers were present (Fig. 5). This suggests that an augment of bacterial production due to an increase of DOM derived from grazers’ activity could enhance viral infections and proliferation (Šimek et al., 2001, Weinbauer et al. 2007). Thus, bacteriophages benefit from the presence of protists taking advantage of increased bacterial fitness. Negative correlation among viral and protistan abundances in treatments where they were jointly present could suggest, however, that at the same time and in spite of this synergy, a competition for prey exist among those two predators. This would indicate an antagonistic (competition-like) interaction among phages and protists, since they compete for the same prey (Maranger et al., 2002). By lysing bacterial cells, phages could remove potential prey of protists, or bacterial cells after infection by viruses could become grazing resistant (Weinbauer et al., 2007). Figure 5. Comparision of bacterial production (BP), lytic viral production (VPL) and virus-mediated mortality of bacteria (VMMBSS) between VFB and VB treatments. Synergistic or antagonistic interactions among protists and phages were interpreted when values were situated above or below the 1:1 line, respectively. VB: viruses+bacteria, VFB: viruses+HNF+bacteria treatments; HNF: heterotrophic nanoflagellates.


13

In contrast, protists may control viral production by grazing on potential viral hosts, infected bacterial cells and free viral particles (González and Suttle, 1993; Manage et al., 2002). Furthermore, theoretical models suggest that viral losses due to protists grazing on both, free viruses and infected bacterial cells, could result in a decrease of viral infections, and could have a negative effect on bacterial species richness (Miki and Yamamura, 2005).

The simultaneous existence of synergy and antagonism among predators does not imply a paradox. We suggest that whereas synergy is the effect of the coexistence of phages and protists, where phages benefit from grazers’ activity, antagonism is the effect of sharing the same prey/host. In our study, however, synergistic interactions among phages and protists seem to be of major importance for bacterial production and mortality. Conclusions

This study was aimed to test the effects of the main bacterial predators on bacterial community and their interactions. Our results indicate an increase of richness of the bacterial community correlated with viral activity/presence, supporting the “killing the winner” hypothesis, and a decrease of richness related with the presence of HNF. We propose that interactions among phages and protists have two dimensions: (1) the effect of grazers’ activity on viral infections and proliferation would have a synergistic character, and will result in an increase of bacterial mortality due to viruses, (2) the effect of competition for the same prey/host will have an antagonistic character, resulting in a decrease of viral abundance with an increase of protistan concentration.


Study site and sampling strategy

One hundred fifty liters of surface water (0.5 m) was collected in four occasions (April 2006, July 2006, November 2006 and February 2007) in Blanes Bay, Spain (The Blanes Bay Microbial Observatory, NW Mediterranean, 41º 40’ N, 2º 48’ E, 20 m depth), at 0.5 miles from shore. Water was prefiltered by 50 μm filter, collected in 50 liter polyethylene carboys, and kept in the dark and in situ temperature until reaching the lab (~ 2 h). Water temperature and salinity was measured in situ with a CTD (conductivity, temperature, depth).

Experimental setup

After reaching the lab, the 150 L of seawater were split into two parts: 100 L of water was filtered through 5 μm pore size cellulose filter (Whatman) to remove microzooplakton (> 5 μm), and then, half of this water (50 L) was filtered through 0.8 m filter to remove HNF. The remaining 50 L of seawater were concentrated to 2 L of viral concentrate using a cartridge of 100 kDa molecular mass cutoff (PREP/SCALE-TFF, 0.23 m2). The 5 m and 0.8 m filtrates were poured to four plastic carboys, 25 L in each one, using a peristaltic pump. Half of the viral concentrate (1 L) was heated in a water bath at 80 ºC for 30 minutes for virus inactivation. Adding 0.5 L of respectively active or inactivated viral concentrate to each carboy, two virus-suppressed treatments: B - bacteria (0.8

m filtrate + inactivated viruses), FB - HNF and bacteria (5 m filtrate + inactivated viruses), and two virus-enhanced treatments: VB - viruses and bacteria (0.8 m filtrate + active viruses), and VFB - viruses, HNF and bacteria (5

m filtrate + active viruses) were prepared. All treatments contained in situ viral communities, as they were not removed from the original water sample.

J. A. Boras et al.

14

The final concentration of viruses in the experimental carboys was 20 % - 29 % higher for active viruses, and 27 % - 40 % higher for heat-inactivated viruses in respect to the initial viral standing stock. Twelve liter duplicates of each treatment were prepared in 20 L mesocosms and incubated at the in situ temperature and in the dark during 48 h. Samples for viral and bacterial abundances were collected at 0, 6, 24 and 48 h of incubation. Samples for abundance of HNF, Chl a concentration and bacterial production were collected at time 0, 24 and 48 h of experiment. Furthermore, at the beginning and at the end of the experiment samples for determination of viral production and bacterial diversity were collected.

Biological parameters

Chl a was extracted from 250-500 ml of water samples and determined fluorometrically (Yentsch and Menzel, 1963). Viral abundances were determined by flow cytometry. Subsamples (2 ml) were fixed with glutaraldehyde (0.5 % final concentration) at 4 ºC, quick frozen in liquid nitrogen and stored at -80 ºC (Brussaard, 2004). Counts were made using FACSCalibur flow cytometer (Becton & Dickinson) with a blue laser emitting at 488 nm. Samples were stained with SYBR Green I, and run at a medium flow speed (Brussaard, 2004). Bacteria were fixed and stored as described for viruses, stained with SYTO 13 and run in a flow cytometer at a low flow speed, using a 50 l of a 0.92 m yellow-green latex beads solution as an internal standard (del Giorgio et al., 1996). The HNF were enumerated by epifluorescence microscopy (Olimpus BX40). Subsamples were fixed with glutaraldehyde (1 % final concentration), and filtered through 0.6 m black polycarbonate filters after staining with DAPI (4,6-diamidino-2-phenylindole; Porter and Feig, 1980) at a final concentration of 5 μg ml-1 (Sieracki et al.,

1985). At least 20-100 HNF were counted per sample.


The BP was estimated from radioactive 3H-leucine incorporation (Kirchman et al., 1985), with the modifications established for the use of microcentrifuge vials (Smith and Azam, 1992). Samples of 1.2 ml were taken at time 0, 24 and 48 h from each treatment and were dispensed into four 2 ml vials plus two TCA-killed control vials. Next, 48

l of a 1 M solution of 3H-leucine was added to the tubes providing a final concentration of 40 nM (which was found to be saturating in these waters). Incubations were run for 2 h in the same thermostatic chambers as the experimental microcosms, and stopped with TCA (50 % final concentration). Following, tubes were centrifuged for 10 minutes at 16000 g. Liquid was sucked with a Pasteur pipette connected to a vacuum pump. Pellets were rinsed with 1.5 ml of 5 % TCA, vortexed and centrifuged again. Supernatant was removed again and 0.5 ml of scintillation cocktail was added. The vials were counted in a Beckman scintillation counter. For each time point, BP was calculated according to the equation:

BP = Leu × CF [ g C L-1 h-1],

where Leu is the 3H-leucine incorporation (pmol L-1 h-1), and CF is the conversion factor (3.1 kg C mol Leu-1). The BP values were converted to bacterial produced cells (ml-1 d-1), using the conversion factor of 12 fg C per bacterium (Lee and Fuhrman, 1987), and averaged between 0 and 24 h (T0), and between 24 and 48 h (T48).

Bacterial mortality

Bacterial losses due to viruses were evaluated twice during each experiment, at the beginning (T0) and at the end (T48) of the incubations. The virus reduction


15

approach (Wilhelm et al., 2002; Weinbauer et al., 2002) was followed to determine viral production and bacterial losses due to phages. Briefly, bacterial concentrate was obtained by concentrating 1 L of seawater on a spiral-wound cartridge (0.22 m pore size, VIVAFlow 200). Virus-free water was collected by filtration of the 0.22 m filtrate through a cartridge of 100 kDa molecular mass cutoff (VIVAFlow 200). A mixture of virus-free water and bacterial concentrate was prepared and distributed into 4 sterile falcon plastic tubes. Two of them were maintained without any manipulations as controls, while in other two, mitomycin C (Sigma) was added (1 µg ml-1 final concentration) as inducing agent of the lytic cycle in lysogenic bacteria. All falcon tubes were incubated in a thermostatic chamber at in situ temperature and in the dark during 24 h. Samples for bacterial and viral abundances were collected at time zero and every two hours during the first six hours, as well as at the end of the incubation, and processed as described above. The number of viruses released by bacterial cell (burst size, BS) was estimated from the viral production experiments, as in Wells and Deming (2006). Briefly, the increase of viral abundance was divided by a decrease of bacterial abundance during the same period of time in the incubation of the viral production experiment. In all experiments, an overall growth of bacteria was observed and therefore BS could be calculated only for very short time intervals, i.e. 2 h. The BS varied between 15 and 90 viruses per cell and was within the range obtained for marine environments (Wommack and Colwell, 2000). We assumed that viral lysis was the only cause of bacterial decrease over short time periods (2 h) in the experimental falcon tubes, and the viral decay and bacterial production rate during this time interval were not considered. Virus-mediated mortality of bacteria (VMM) was estimated following the

model of Weinbauer and colleagues (2002). The equations are described in detail in Boras and colleagues (2009). Briefly, lysogenic viral production (VPLG) was calculated resting lytic VP (VPL; i.e. an increase of viruses in the controls) from total VP (i.e. an increase of viruses in the samples tubes amended with the mitomycin C). Due to loss of bacterial standing stock (BSS) during tangential flow filtration, a bacterial loss factor was calculated (from 1.0 to 22.1) to allow the comparison of the VP between experiments (Winget et al., 2005). Following the method used by Guixa-Boixereu (1997), the rate of lysed cells (RLC) was obtained dividing VPL by BS. The RLC was used to calculate VMM as a percentage of BSS at T0 and T48 respectively (VMMBSS). Since grazing by protists neither on viruses nor on infected cells was considered in our calculations (González and Suttle, 1993), the VMMBSS

are potential losses, and thus their values can be overestimated.

Phylogenetic analysis

Nucleic acids extraction – The DGGE of the polymerase chain reaction (PCR)-amplified 16S rRNA gene fragments was used to compare bacterial communities from each treatment. To collect microbial biomass, 1 L of seawater was filtered through a 3.0 m pore size filter (Millipore) and a 0.2 m Sterivex filter (Durapore, Millipore). The Sterivex units were filled with 1.8 ml of lysis buffer (40 mM EDTA, 50 mM Tris-HCl pH = 8.3, 0.75 M sucrose) and stored at -80 C. The nucleic acids were treated with lysozyme, proteinase K and sodium dodecyl sulphate, extracted with phenol and concentrated in a Centricon-100 (Millipore) (Massana et al., 1997). The quality of the DNA was checked by agarose gel electrophoresis.

DGGE gel analysis – The DGGE was carried out as previously described in Schauer and colleagues (2003). The PCR amplification of bacterial 16S rRNA was performed using the bacterial

J. A. Boras et al.

16

specific primer GC358f and the universal primer 907rM. The PCR products were verified by agarose gel electrophoresis with a standard in the gel (Low DNA Mas Ladder, Gibco BRL). To perform the DGGE, the PCR product was loaded on 6 % polyacrylamide gels with a DNA-denaturant gradient ranging from 40 % to 80 %. The DGGE gels were run at 100 V for 16 h at 60 ºC in TAE 1x running buffer using DGGE-2000 system (CBS Scientific Company), as previously described for Blanes samples (Alonso-Sáez et al., 2007). The gels were stained with 3 μl of SYBR Gold in 20 ml of TAE 1x, and analyzed using the Fluor-S MultiImager (Bio-Rad) with the Multi-Analyst software (Bio-Rad). Digitized DGGE images were analyzed using the Diversity Data software (Bio-Rad), described previously (Schauer et al., 2000). Matrix was constructed for all detected bands taking into account the relative contribution of each band (as percentage) to the total intensity of the band line. Ranked matrix of similarities among samples was constructed using the Bray-Curtis similarity measure (Bray and Curtis, 1957). Based on this matrix a hierarchical clustering was performed, which resulted in a dendrogram. Statistical analyses

Normal distribution of data was checked using the Shapiro-Wilk W test, and data were logarithmically transformed if necessary. Variations of parameters among clusters were analyzed by one-way ANOVA for normal distributions, and by Wilcoxon test for non-normal distributions. Pearson correlation analysis was used to determine the relationships between parameters. These statistical analyses were performed using the JMP program. The similarity matrices of DGGE results and a dendrogram were constructed and analyzed using the PRIMER v.6 (Plymouth Marine Laboratory) program.

Acknowledgments

Funding for this research was provided by PROCAVIR and MICROVIS (CTM2004-04404-CO2-00/MAR, CTM2007-62140 to D. Vaqué), and ÍCARO (200830I120 to M. M. Sala) projects, funded by the Spanish Ministry of Education and Science (MEC). J.A.B. work was supported by a PhD fellowship from the MEC (FPU grant). We wish to thank to C. Balestra and J. Felipe for their help in the flow cytometer analysis of the samples. This is a contribution to the European Network of Excellence EurOceans.

References

Alonso-Sáez, L., Balagué, V., Sà, E.L., Sánchez, O., González, J.M., Pinhassi, J., Massana, R., Pernthaler, J., Pedrós-Alió, C., and Gasol, J.M. (2007) Seasonality in bacterial diversity in north-west Mediterranean coastal waters: assessment through clone libraries, fingerprinting and FISH. FEMS Microbiol Ecol 60: 98-112.

Bonilla-Findji, O., Herndl, G.J., Gattuso, J.-P., and Weinbauer, M.G. (2009) Viral and flagellate control of prokaryotic production and community structure in offshore Mediterranean waters. Appl Environ Microbiol 75: 4801-4812.

Boras, J.A., Sala, M.M., Vázquez-Domínguez, E., Weinbauer, M.G., and Vaqué, D. (2009) Annual changes of bacterial mortality due to viruses and protists in an oligotrophic coastal environment (NW Mediterranean). Environ Microbiol 11: 1181-1193.

Bouvier, T., and del Giorgio, P.A. (2007) Key role of selective viral-induced mortality in determining marine bacterial community composition. Environ Microbiol 9: 287-297.

Bray, J.R., and Curtis, J.T. (1957) An ordination of the upland forest


17

communities of Southern Wisconsin. Ecol Monogr 27: 325-349.

Brussaard, C.P.D. (2004) Optimization of procedures for counting viruses by flow cytometry. Appl Environ Microbiol 70: 1506-1513.

Calbet, A., Landry, M.R., and Nunnery, S. (2001) Bacteria-flagellate interactions in the microbial food web of the oligotrophic subtropical North Pacific. Aquat Microb Ecol 23: 283-292.

Caron, D.A., Goldman, J.C., and Dennett, M.R. (1988) Experimental demonstration of the roles of bacteria and bacterivorous protozoa in plankton nutrient cycles. Hydrobiologia 159: 27-40.

del Giorgio, P.A., Bird, D.F., Prairie, Y.T., and Planas, D. (1996) Flow cytometric determination of bacterial abundance in lake plankton with the green nucleid acid stain SYTO 13. Limnol Oceanogr 41: 783-789.

Fuhrman, J.A. (1999) Marine viruses and their biogeochemical and ecological effects. Nature 399: 541-548.

Fuhrman, J.A., and Azam, F. (1980) Bacterioplankton secondary production estimates for coastal waters of British Columbia, Antarctica, and California. Appl Environ Microbiol 39: 1085-1095.

Fuhrman, J.A., and Noble, R.T. (1995) Viruses and protists cause similar bacterial mortality in coastal seawater. Limnol Oceanogr 40: 1236-1242.

González, J.M., and Suttle, C.A. (1993) Grazing by marine nanoflagellates on viruses and virus-sized particles: ingestión and digestión. Mar Ecol Prog Ser 94: 1-10.

Guixa-Boixereu, N. (1997) Abundància i dinàmica dels virus en ecosistemes planctònics. PhD dissertation, Universitat de Barcelona, Barcelona, Spain.

Guixa-Boixereu, N., Lysnes, K., and Pedrós-Alió, C. (1999) Viral lysis and bacterivory during a phytoplankton bloom in a coastal water microcosm.

Appl Environ Microbiol 65: 1949-1958.

Hahn, M.W., and Höfle, M.G. (1999) Flagellate predation on a bacterial model community: Interplay of size-selective grazing, specific bacterial cell size, and bacterial community composition. Appl Environ Microbiol 65: 4863-4872.

Hewson, I., and Fuhrman, J.A. (2006) Viral impacts upon marine bacterioplankton assemblage structure. J Mar Biol Ass UK 86: 577-589.

Horňák, K., Mašín, M., Jezbera, J., Bettarel, Y., Nedoma, J., Sime-Ngando, T., and Šimek, K. (2005) Effects of decreased resource availability, protozoan grazing and viral impact on a structure of bacterioplankton assemblage in a canyon-shaped reservoir. FEMS Microbiol Ecol 52: 315-327.

Kirchman, D., K’nees, E., and Hodson, R. (1985) Leucine incorporation and its potential as a measure of protein synthesis by bacteria in natural aquatic systems. Appl Environ Microbiol 43: 599-607.

Lee, S., and Fuhrman, J.A. (1987) Relationships between biovolume and biomass of naturally derived marine bacterioplankton. Appl Environ Microbiol 53: 1298-1303.

Manage, P.M., Kawabata, Z., Nakano, S.-I., and Nishibe, Y. (2002) Effect of heterotrophic nanoflagellates on the loss of virus-like particles in pond water. Ecol Res 17: 473-479.

Maranger, R., del Giorgio, P.A., and Bird, D.F. (2002) Accumulation of damaged bacteria and viruses in lake water exposed to solar radiation. Aquat Microb Ecol 28: 213-227.

Massana, R., Murray, A.E., Preston, C.M., and DeLong, E.F. (1997) Vertical distribution and phylogenetic characterization of marine planktonic Archaea in the Santa Barbara Channel. Appl Environ Microbiol 63: 50-56.

J. A. Boras et al.

18

Miki, T., and Jacquet, S. (2008) Complex interactions in the microbial world: underexplored key links between viruses, bacteria and protozoan grazers in aquatic environments. Aquat Microb Ecol 51: 195-208.

Miki, T., and Yamamura, N. (2005) Intraguild predation reduces bacterial species richness and lossens the viral loop in aquatic systems: “kill the killing the winner” hypothesis. Aquat Microb Ecol 40: 1-12.

Murray, A.G., and Jackson, G.A. (1992) Viral dynamics: a model of the effects of size, shape, motion and abundance of single-celled planktonic organisms and other particles. Mar Ecol Prog Ser 89: 103-116.

Parada, V., Herndl, G.J., and Weinbauer, M.G. (2006) Viral burst size of heterotrophic prokaryotes in aquatic systems. J Mar Biol Ass UK 86: 613-621.

Pernthaler, J. (2005) Predation on prokaryotes in the water column and its ecological implications. Nat Rev Microbiol 3: 537-546.

Porter, K.G., and Feig, Y.S. (1980) The use of DAPI for identifying and counting aquatic microflora. Limnol Oceanogr 25: 943-948.

Posch, T., Šimek, K., Vrba, J., Pernthaler, J., Nedoma, J., Sattler, B., Sonntag, B., and Psenner, R. (1999) Predator-induced changes of bacterial size-structure and productivity studied on an experimental microbial community. Aquat Microb Ecol 18: 235-246.

Schauer, M., Massana, R., and Pedrós-Alió, C. (2000) Spatial differences in bacterioplankton composition along the Catalan coast (NW Mediterranean) assessed by molecular fingerprinting. FEMS Microbiol Ecol 33: 51-59.

Schauer, M., Balagué, V., Pedrós-Alió, C., and Massana, R. (2003) Seasonal changes in the taxonomic composition of bacterioplankton in a coastal oligotrophic system. Aquat Microb Ecol 31: 163-174.

Sieracki, M.E., Johnson, P.B., and Sieburth, J.McN. (1985) Detection, enumeration and sizing of planctonic bacteria by image analyzed epifluorescence microscopy. Appl Environ Microbiol 49: 799-810.

Sime-Ngando, T., and Pradeep Ram, A.S. (2005) Grazer effects on prokaryotes and viruses in a freshwater microcosm experiment. Aquat Microb Ecol 41: 115-124.

Smith, D.C., and Azam, F. (1992) A simple, economical method for measuring bacterial protein synthesis rates in seawater using 3H-leucine. Mar Microb Food Webs 6: 107-114.

Šimek, K., Vrba, J., Pernthaler, J., Posch, T., Hartman, P., Nedoma, J., and Psenner, R. (1997) Morphological and compositional shifts in an experimental bacterial community influenced by protists with contrasting feeding modes. Appl Environ Microbiol 63: 587-595.

Šimek, K., Pernthaler, J., Weinbauer, M.G., Hornák, K., Dolan, J.R., Nedoma, J., Mašín, M., and Amann, R. (2001) Changes in bacterial community composition and dynamics and viral mortality rates associated with enhanced flagellate grazing in a mesoeutrophic reservoir. Appl Environ Microbiol 67: 2723-2733.

Šimek, K., Weinbauer, M.G., Hornák, K., Jezbera, J., Nedoma, J., and Dolan, J.R. (2007) Grazer and virus-induced mortality of bacterioplankton accelerates development of Flectobacillus populations in a freshwater community. Environ Microbiol 9: 789-800.

Thingstad, T.F. (2000) Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnol Oceanogr 45: 1320-1328.

Thingstad, T.F., and Lignell, R. (1997) A theoretical approach to the question of how trophic interactions control carbon demand, growth rate,


19

abundance and diversity. Aquat Microb Ecol 13: 19-27.

Weinbauer, M.G., and Rassoulzadegan, F. (2004) Are viruses driving microbial diversification and diversity? Environ Microbiol 6: 1-11.

Weinbauer, M.G., and Suttle, C.A. (1999) Lysogeny and prophage induction in coastal and offshore bacterial communities. Aquat Microb Ecol 18: 217-225.

Weinbauer, M.G., Winter, C., and Höfle, M.G. (2002) Reconsidering transmission electron microscopy based estimates of viral infection of bacterioplankton using conversion factors derived from natural communities. Aquat Microb Ecol 27: 103-110.

Weinbauer, M.G., Christaki, U., Nedoma, J., and Šimek, K. (2003) Comparing the effects of resource enrichment and grazing on viral production in a meso-eutrophic reservoir. Aquat Microb Ecol 31: 137-144.

Weinbauer, M.G., Hornák, K., Jezbera, J., Nedoma, J., Dolan, J.R., and Šimek, K. (2007) Synergistic and antagonistic effects of viral lysis and protistan grazing on bacterial biomass, production and diversity. Environ Microbiol 9: 777-788.

Wells, L.E., and Deming, J.W. (2006) Significance of bacterivory and viral lysis in bottom waters of Franklin Bay, Canadian Arctic, during winter. Aquat Microb Ecol 43: 209-221.

Wilhelm, S.W., Brigden, S.M., and Suttle, C.A. (2002) A dilution technique for the direct measurement of viral production: a comparison in stratified and tidally mixed coastal waters. Microb Ecol 43: 168-173.

Williamson, S.J., Houchin, L.A., McDaniel, L., and Paul, J.H. (2002) Seasonal variation in lysogeny as depicted by prophage induction in Tampa Bay, Florida. Appl Environ Microbiol 68: 4307-4314.

Winget, D.M., Williamson, K.E., Helton, R.R., and Wommack, K.E. (2005) Tangential flow diafiltration: an

improved technique for estimation of virioplankton production. Aquat Microb Ecol 41: 221-232.

Wommack, K.E., and Colwell, R.R. (2000) Virioplankton: viruses in aquatic ecosystems. Microbiol Mol Biol Rev 64: 69-114.

Yentsch, C.S., and Menzel, D.W. (1963) A method for the determination of phytoplankton, chlorophyll and phaeophytin by fluorescence. Deep-Sea Res 10: 221-231.

Zhang, R., Weinbauer, M.G., Qian, P.-Y. (2007) Viruses and flagellates sustain apparent rechness and reduce biomass accumulation of bacterioplankton in coastal marine waters. Environ Microbiol 9: 3008-3018.

J. A. Boras et al.

20

SUPPORTING INFORMATION Table 1A. In situ values of temperature and biological parameters of water for the experiments’ performance. Temp.: temperature, Chl a: chlorophyll a, HNF: heterotrophic nanoflagellates (≤ 5 μm).

Experiment Temp.

ºC

Chl a

μg L-1

Viruses

ml-1

Bacteria

ml-1

HNF

ml-1

Spring 13.5 1.23 3.03 × 107 1.74 × 106 1.48 × 102

Summer 23.7 0.09 1.88 × 107 3.94 × 106 7.72 × 102

Autumn 17.0 0.60 9.99 × 106 4.87 × 105 4.56 × 102

Winter 14.0 1.04 1.10 × 107 4.80 × 105 7.00 × 102


21

Table 2A. Abundances of microorganisms at the beginning (T0), 6 h, 24 h and at the end (T48) of four seasonal experiments. SD: standard deviation that coincides with maximal and minimal range of duplicates; nd: not detected; HNF: heterotrophic nanoflagellates (≤5 μm); B: bacteria, VB: viruses+bacteria, FB: HNF + bacteria, VFB: viruses+HNF+bacteria treatments.

Expta

Treatb Viruses

107 ml-1 ± SD Bacteria

106 ml-1 ± SD HNF

102 ml-1 ± SD

T0 T6 T24 T48 T0 T6 T24 T48 T0 T24 T48

Spring

B VB FB VFB

2.0±0.1 2.0±0.1 3.7±0.1 3.4±0.1

2.0±0.0 2.0±0.2 3.0±0.3 3.2±0.1

1.8±0.2 2.2±0.3 3.1±0.2 3.2±0.1

1.9±0.0 2.2±0.2 3.2±0.1 3.6±0.2

0.8±0.0 1.3±0.0 1.9±0.1 2.0±0.2

0.9±0.0 1.3±0.2 1.6±0.1 1.9±0.0

1.5±0.0 1.8±0.0 2.0±0.0 2.1±0.0

1.9±0.1 1.8±0.1 2.1±0.0 2.2±0.1

nd nd

1.8±0.0 2.3±0.0

nd nd

1.4±0.2 1.3±0.1

nd nd

2.0±0.0 2.0±0.3

Summer B VB FB VFB

1.5±0.0 1.4±0.0 1.9±0.0 1.9±0.0

1.3±0.0 1.3±0.0 2.0±0.4 1.9±0.2

1.3±0.2 1.5±0.1 2.1±0.1 1.9±0.0

1.5±0.0 1.3±0.1 1.6±0.0 1.6±0.0

0.4±0.0 0.4±0.0 1.8±0.0 1.6±0.0

0.9±0.0 0.3±0.0 1.2±0.0 1.0±0.0

0.9±0.0 0.8±0.1 1.0±0.0 1.1±0.1

1.2±0.2 0.9±0.2 1.4±0.8 0.9±0.0

nd nd

14.4±0.3

21.1±2.2

nd nd

32.2±25.3 46.7±7.9

nd nd

59.2±16.4

63.8±7.8

Autumn B VB FB VFB

0.8±0.0 0.8±0.0 1.3±0.0 1.1±0.0

1.2±0.5 1.3±0.0 1.5±0.0 1.6±0.1

1.2±0.2 1.3±0.0 1.3±0.0 1.5±0.0

1.3±0.1 1.4±0.0 1.5±0.2 1.5±0.0

0.4±0.0 0.5±0.0 0.7±0.0 0.8±0.0

0.6±0.0 0.9±0.0 1.2±0.1 1.2±0.2

1.3±0.0 1.3±0.0 1.7±0.0 1.6±0.0

1.5±0.2 1.4±0.0 1.7±0.1 1.6±0.1

nd nd

6.8±0.3 4.0±0.3

nd nd

4.0±0.8 4.1±1.5

nd nd

4.2±0.8 2.8±1.1

Winter B VB FB VFB

0.8±0.0 0.5±0.0 1.0±0.0 0.7±0.0

1.3±0.2 0.8±0.0 1.3±0.0 1.0±0.1

0.9±0.0 0.7±0.0 3.2±3.3 0.9±0.0

0.5±0.0 0.7±0.0 0.8±0.1 0.9±0.0

0.4±0.0 0.4±0.0 1.4±0.0 1.4±0.0

0.5±0.0 0.5±0.0 0.8±0.0 1.2±0.5

0.9±0.0 0.9±0.2 1.2±0.0 0.6±0.8

1.1±0.1 1.0±0.0 1.0±0.1 1.4±0.3

nd nd

3.8±1.3 4.9±1.4

nd nd

8.3±5.8 13.5±0.0

nd nd

24.6±1.0 17.4±1.6

Paper IV

Effect of viruses and protists on

bacteria in eddies of the Canary

Current region (subtropical

Northeast Atlantic)

Summary – Paper IV

P a g e | 113

Resumen

Efecto de virus y protistas sobre las bacterias en los remolinos

de la región de la Corriente Canaria (Atlántico NE subtropical)

Artículo acceptado para la publicación en Limnology and Oceanography (Effect of viruses and

protists on bacteria in eddies of the Canary Current region (subtropical Northeast Atlantic))

El efecto de los remolinos oceánicos en los procesos microbianos, con énfasis en la

mortalidad bacteriana debida a fagos y protistas, fue examinado en la región de la Corriente

Canaria (NE Atlántico subtropical) a través de la columna de agua (hasta 1000 m), en Agosto de

2006. Las estaciones de muestreo estaban situadas en una zona de remolinos ciclónicos (2

estaciones) y anticiclónicos (2 estaciones), y en otra situada fuera de la influencia del campo de

los remolinos (2 estaciones “far-field”). En la zona eufótica (desde la superficie hasta 200 m de

profundidad) de los remolinos ciclónicos, la mortalidad bacteriana debida a virus y protistas fue,

respectivamente, de 25.6% hasta 69.8%, y de “no detectada” hasta 46.8% de la producción

bacteriana (BP) d-1. En los remolinos anticiclónicos estos valores variaban entre 20.6% y 90.2% BP d-

1 para la lisis vírica, y entre 8.0% BP d-1 y 79.4% BP d-1 para la depredación por protistas. En las

estaciones “far-field”, la mortalidad bacteriana oscilaba desde 48.7% BP d-1 hasta 66.9% BP d-1

para la lisis vírica, y desde “no detectada” hasta 44.8% para la depredación por protistas. La lisis

vírica detectada en el máximo profundo de fluorescencia (DFM) era significativamente más alta

en los remolinos anticiclónicos que en los ciclónicos o en las estaciones “far-field” (F2, 3 = 31.9, p <

0.01). Considerando todas las estaciones y profundidades (desde la superficie hasta 1000 m), la

mortalidad bacteriana debida a los virus era significativamente más alta que la mortalidad

debida a los protistas. Además, en los remolinos anticiclónicos, las pérdidas bacterianas fueron

significativamente más altas en la zona eufótica (desde la superficie hasta 200 m) que en aguas

más profundas (700 – 1000 m). La lisogénia fue detectada en el 46% de los casos yendo desde un

13.4% hasta un 84.6% de la producción vírica total. La lisogénia fue más frecuente en las

Boras J. A., Impact of viruses on bacterioal communities

114 | P a g e

estaciones anticiclónicas, donde fue observado un impacto mayor de los protistas sobre la

biomasa y producción bacteriana. La notable actividad vírica en esta región sugiere que,

además de los nutrientes aportados por la acción de los remolinos, los productos de la lisis de las

bacterias podrían ser una fuente adicional de nutrientes en las aguas superficiales oligotróficas

del océano.

Editor-in-Chief: Everett Fee Phone: 403-609-2456 Journals Manager: Lucille Doucette <[email protected]> Fax: 403-609-2400 <[email protected]>

LIMNOLOGY

AND

OCEANOGRAPHY L&O Editorial Office, 343 Lady MacDonald Crescent, Canmore, Alberta T1W

1H5 Canada Dr. Julia A. Boras Institut de Ciències del Mar (CSIC) P. Marítim de la Barceloneta 37-49 08003 Barcelona SPAIN 34 932309500 34 932309555 (fax) Dear Julia, Re: L&O 09-135 -- Boras, Julia A., M. Montserrat Sala, Federico Baltar, Javier Arístegui, Carlos M. Duarte, and Dolors Vaqué. Effect of viruses and protists on bacteria in eddies of the Canary Current region (subtropical Northeast Atlantic). This is formal confirmation that this paper has been accepted for publication in L&O. You are hereby granted permission to include a copy of this manuscript in your thesis submission to the University of Barcelona. Everett Fee

Editor in Chief

Limnology and Oceanography, Accepted

Effect of viruses and protists on bacteria in eddies of the Canary Current region (subtropical Northeast Atlantic)

Julia A. Boras1 and M. Montserrat Sala

Institut de Ciències del Mar (Consejo Superior de Investigaciones Científicas), Passeig Marítim de la Barceloneta 37-49, 08003 Barcelona, Spain Federico Baltar and Javier Arístegui

Facultad de Ciencias del Mar, Universidad de Las Palmas de Gran Canaria, Campus 14 Universitario de Tafira, 35017 Las Palmas de Gran Canaria, Spain

Carlos M. Duarte

Institut Mediterrani d'Estudis Avançats (Consejo Superior de Investigaciones Científicas – Universidad de las Islas Baleares), Miquel Marqués 21, 07190 Esporles (Mallorca), Spain Dolors Vaqué

Institut de Ciències del Mar (Consejo Superior de Investigaciones Científicas), Passeig Marítim de la Barceloneta 37-49, 08003 Barcelona, Spain

Abstract

The effect of oceanic eddies on microbial processes, with emphasis on bacterial losses due to protists and phages, was examined in the Canary Current region (subtropical NE Atlantic), through the water column (down to 1000 m), during August 2006. Sampling stations were located in cyclonic and anticyclonic eddies, as well as in region situated outside the influence of the eddy field (far-field stations). In the euphotic zone, in cyclonic eddies losses of bacteria due to viruses and protists were from 25.6% to 69.8%, and from not detected to 46.8% of bacterial production (BP) d-1, respectively. In anticyclonic eddies, these values ranged from 20.6% to 90.2% of BP d-

1 for viruses, and from 8.0% to 79.4% of BP d-1 for protists. At far-field stations, losses of bacteria ranged from 48.7% to 66.9% for viruses, and from not detected to 44.8% for protists. In addition, covering all stations and depths (from epipelagic to bathypelagic layer) bacterial losses due to viruses were significantly higher than losses by protists, and did not differ significantly among depths except for the stations situated in anticyclonic eddies, where they were significantly higher in the epipelagic layer. Lysogenic infection was more frequent at anticyclonic stations, where the highest pressure of protists on bacteria was observed. Due to the importance of viral activity we suggest that the lysis products from bacteria may be a source of regenerated nutrients in the surface of the oligotrophic ocean, in addition to the input of nutrients upwelled by eddies.

Corresponding author: [email protected]

Boras et al.

2

Introduction

Mesoscale eddies are common features in the ocean. Cyclonic eddies raise deep, nutrient-rich water to the surface, and anticyclonic eddies deepen the warm, nutrient-poor surface water. Eddies are thus a key mode of vertical nutrient transport in the ocean. Because of the overall nutrient limitation of primary producers in the ocean, eddies can regulate oceanic primary production. Some studies showed that nutrient input to the eutrophic waters by cyclonic eddies can significantly enhance the production of autotrophs (Falkowski et al. 1991). Due to the tight coupling between marine prokaryotes and primary producers, eddies should also affect processes within the microbial heterotrophic community. Indeed, higher bacterioplankton abundances inside cyclonic eddies relative to surrounding waters have been reported (Lochte and Pfannkuche 1987). However, the evidence of the effect of eddies on the functioning of microbial food webs is still scarce (Olaizola et al. 1993). For instance, virioplankton are recognised to be an essential component of the microbial loop, but their distribution and activity (production and lysis of prokaryotes) in oceanic eddies has not yet been described. Moreover, the effect of eddies on microbial processes should not be restricted to surface waters. Enhanced primary production at the surface of cyclonic eddies may lead to an increased flux of particulate organic matter (POM) to their meso- and bathypelagic waters. The interpretation of the effect of these mechanisms in meso- and bathypelagic layers is difficult, as there is still lack of knowledge of microbial food web processes therein (Magagnini et al. 2007).

Viral infection and subsequent cell lysis recycles organic carbon (C) and nutrients contained in bacterial cells, which can be particularly important in nutrient-limited systems (Fuhrman 1999), whereas protistan grazing on bacteria transfers the POM up into the food web. Models predict that viral infections of bacterioplankton will be more prevalent in eutrophic than oligotrophic

systems, due to higher hosts density (Murray and Jackson 1992), as supported by some empirical observations (Weinbauer and Suttle 1999). In contrast, protistan grazing is considered to be the principal cause of bacterial mortality in oligotrophic waters (Guixa-Boixereu et al. 1996). Trophy of the system may also affect the ‘life strategy’ of viruses, and thus

a higher percentage of lysogens was found in oligotrophic than in eutrophic waters (Weinbauer et al. 2003). Because bacteria are the major node in C cycling in oligotrophic systems (del Giorgio et al. 1997), the nature of bacterial mortality, cell lysis or protistan grazing, shapes the C flow in oligotrophic oceanic ecosystems.

Here we compare the bacterial mortality mediated by viruses and protists in contrasting mesoscale eddies in the Canary Current region, where a recurrent eddy field is located south of the Canaries Archipelago. We hypothesize that the relaxation of oligotrophic conditions in cyclonic eddies would lead to a higher control of viruses over bacteria along the euphotic zone, while protists would be the dominant source of bacterial mortality outside these eddies (far-field stations). In the euphotic zone of anticyclonic eddies, we expected to find no changes in microorganisms growth and activity compared with those in the far-field stations, as no modification in water trophy was assumed to result from the deepening of surface water masses. Lysogeny was expected to be the dominant type of infection in anticyclonic and far-field stations where oligotrophic conditions prevail, and should be less frequent in waters under the influence of cyclonic eddies. Also, a higher percentage of lysogeny was expected in deeper waters than in the euphotic zone, due to the lower abundance of host. Methods

Study area and sampling – The study was carried out at stations located at the subtropical NE Atlantic, in the Canary Current region (26.5 - 30º N, 15 – 23.1º W, Fig. 1), on board of the Buque de Investigación Oceanográfica (BIO) - Hespérides during the

Bacterial mortality in ocean eddies

3

Table 1. Average values of selected physico-chemical and biological parameters determined during the RODA1 cruise in the cyclonic eddie (CE), anticyclonic eddie (AE), and far-field (FF) stations. Depth layers: epi. - epipelagic, 5 to < 200 m; trans. - transition, 200 m to < 500 m; meso. - mesopelagic, 500 m to 800 m; bathy. - bathypelagic, > 800 m to 2000 m. Temperature (Temp., ºC); fluorescence (Fluor., arbitrary fluorescence units); viral abundance (VA, 106 viruses mL-1); bacterial abundance (BA, 105 cells mL-1); heterotrophic nanoflagellates (HNF, 102 cells mL-1); ciliates (102 cells L-1); lytic viral production (VPL, 106 viruses mL-1 d-1); bacterial production (BP; 105 cells mL-1 d-1); rate of lysed cells (RLC, 104 cells mL-1 d-1); grazing rate (G, 104 cells mL-1 d-1). nd – not detected.

Station and

layer

Temp.

Salinity

Fluor.

VA

BA

HNF

Ciliates

VPL

BP

RLC

G

FF1 epi. trans. meso. bathy.

21.0 15.6 11.0

7.0

37.0 36.2 35.6 35.3

0.27 0.05 0.04 0.04

2.5 0.4 0.2 0.2

1.9 0.4 0.3 0.2

2.3 0.9 0.2 0.1

16.2

3.4 0.6 0.2

1.4 0.4 0.9 1.2

2.9 1.0 0.2 0.6

15.8

4.8 2.0 5.7

12.8

2.9 0.4 0.6

FF2 epi. trans. meso. bathy. Average

20.3 15.7 11.1

7.2

13.4

36.9 36.3 35.6 35.3

36.0

0.25 0.04 0.04 0.04

0.09

3.3 0.6 0.7 0.2

1.0

2.6 0.7 0.7 0.3

0.9

1.9 0.4 0.9 0.2

1.0

8.6 2.8 0.3 0.1

4.8

1.7 0.4 0.5 0.1

0.8

1.5 0.5 0.2 0.2

0.9

9.8 3.2 0.8 1.2

5.4

nd 0.7 0.2 0.3

2.2

CE1 epi. trans. meso. bathy.

19.2 13.7

9.5 7.1

36.6 35.9 35.4 35.2

0.27 0.04 0.04 0.04

4.3 0.9 0.5 0.4

3.6 1.0 0.6 0.4

3.4 0.7 0.2 0.3

- - - -

1.1

0.04 0.2 0.2

1.8 0.3 2.0 0.7

4.3 1.3

17.2 5.9

3.3 nd 2.9

<0.1

CE2 epi. trans. meso. bathy. Average

20.0 14.4 10.5

6.7

13.6

36.8 36.0 35.5 35.2

35.9

0.39 0.03 0.04 0.04

0.13

4.1 0.6 0.3 0.2

1.8

2.2 1.2 0.6 0.2

1.5

16.5

0.6 0.4 0.1

3.6

16.8

3.8 0.6 0.3

6.5

0.3 0.7 1.0 1.4

0.6

4.3 0.4 0.2 0.2

1.2

15.9

2.6 1.6 0.6

6.2

20.1

0.6 0.4 nd

3.4

AE1 epi. trans. meso. bathy.

22.8 15.7 10.6

8.0

37.0 36.3 35.5 35.3

0.32 0.05 0.04 0.04

4.4 0.7 0.3 0.3

5.4 1.3 0.8 0.5

4.9 1.1 0.4 0.2

11.0

2.3 0.5 0.1

1.0 0.8 0.7 1.0

5.8 0.6 0.4 0.1

52.8

1.2 2.3 0.5

4.7 4.8 1.9 0.4

cruise ‘Remolinos Oceánicos y Deposición Atmosférica’ (RODA1), from 11 August to 05 September 2006. Six stations were sampled: two stations in cyclonic eddies (CE1, CE2), two stations in anticyclonic eddies (AE1, AE2), and

two presumably undisturbed far-field stations, located in regions where no eddies were detected (FF1, FF2; Fig. 1). At each station, temperature, salinity and fluorescence were recorded down to 2000 m depth using a SeaBird

Boras et al.

4

Fig. 1. Map of the area studied during the RODA1 cruise. FF – far-field stations, CE – cyclonic eddy stations, AE – anticyclonic eddy stations. Black lines correspond to XBT sections across eddies, represented in Figure 2. 911 plus CTD system, mounted on a General Oceanics rosette sampler, equipped with 12 L Niskin bottles. Samples for microbial abundances and processes were taken at intervals of 25 m up to 100 m, at 150 m, and at 200 m, including the deep fluorescence maximum (DFM) layer, and from 200 m at intervals of 100 m down to 1000 m. Furthermore, samples for viral and bacterial abundances were taken also at 2000 m. At FF2, CE2, and AE1 samples for nutrient analysis (phosphate, nitrate + nitrite, and ammonium) were collected down to 500 m. Samples for the determination of the dissolved inorganic phosphate concentrations, and the nitrate + nitrite concentrations were kept frozen until analyzed in a Bran+Luebbe AA3 autoanalyser, following standard spectrophotometric methods (Hansen and Koroleff 1999), and ammonium concentrations were measured spectrofluorometrically within 1 h of collection.

Microorganisms’ abundances – In situ viral and bacterial abundances were determined by flow cytometry. Subsamples (2 mL), taken at 12 selected depths, were fixed with glutaraldehyde for viruses (0.5% final concentration), or paraformaldehyde for bacteria (1% final concentration). Samples of viruses were fixed at 4ºC during 15-30 min, then quick frozen in liquid nitrogen and stored at -80ºC as described in Brussaard (2004).

Fig. 2. Vertical sections of potential temperature (ºC) across the different eddies studied (from west to east; see Fig. 1). Arrows on the top axis indicate XBT stations. Vertical dotted lines indicate positions at the eddy centers were CTD casts were carried out.


5

Table 2. Significant correlation coefficients among variables. Fluor. (fluorescence); VA (viral abundance); BA (bacterial abundance); VBR (virus-bacterium ratio); HNF (heterotrophic nanoflagellate abundance); %VPLG (percentage of lysogenic viral production in total VP); BP (bacterial production); VMM (virus-mediated mortality of bacteria: BSS - % of bacterial standing stock); PMM (protist-mediated mortality of bacteria).

Variables n r p Variables n r p

VA – Depth 80 -0.741 <0.01 BP – Fluor. 24 0.759 <0.01

VA – Fluor 80 0.753 <0.01 BP – VA 24 0.750 <0.01

VA – BA 75 0.897 <0.01 BP – BA 22 0.631 <0.01

BA – Depth 75 -0.776 <0.01 BP – HNF 24 0.661 <0.01

BA – Fluor 75 0.625 <0.01 %VPLG – BP 24 0.460 <0.05

VBR – Depth 75 -0.238 <0.01 %VPLG – Fluor. 24 0.462 <0.05

HNF – Depth 37 -0.818 <0.01 %VPLG – VA 24 0.407 <0.05

HNF – BA 37 0.919 <0.01 %VPLG – HNF 24 0.455 <0.05

Ciliats - BA 28 0.805 <0.01 VMMBSS – BP 24 0.490 <0.02

Ciliats – HNF 30 0.846 <0.01 PMMBSS – BP 24 0.510 <0.02

BP – Depth 24 -0.659 <0.01 VMMBP – PMMBP 24 -0.474 <0.02

Samples of bacteria were analyzed on board immediately after fixation. Counts were made on a FACSCalibur (Becton and Dickinson) flow cytometer. Samples of viruses were stained with SYBR Green I, and run at a medium flow speed (Brussaard 2004), and samples of bacteria were stained with a DMS (dimethyl sulfate)-diluted SYTO13, and run at a low speed using 50 mL of 0.92 mm yellow-green latex beads as an internal standard (del Giorgio et al. 1996).

In situ nanoflagellate abundance was determined by epifluorescence microscopy (Olympus BX40-102/E at 1000 X). Subsamples (100 mL) were taken at 6 selected depths, fixed with glutaraldehyde (1% final concentration), filtered through 0.6 mm black polycarbonate filters, and stained with DAPI (4,6-diamidino-2-phenylindole) at a final concentration of 5 μg mL-1. Heterotrophic and phototrophic nanoflagellates (HNF and PNF) were distinguished under ultraviolet (UV) and blue light (B2 filter). At least 20-100 HNF were counted per sample. The HNF were grouped into four size classes: ≤ 2 µm, 2-5 µm, 5-10 µm,

and 10-20 µm. To determine ciliate abundance and community composition, one liter of water from the same depths as for nanoflagellates counts was immediately fixed with acidic lugol (2% final concentration). The fixed samples were allowed to settle for 48 h in the same sampling bottles, and the supernatant was gently removed, leaving ~ 200 mL. A hundred milliliters of this concentrate was further sedimented in 100 mL chambers for at least 48 h before enumeration at 400 X magnification. Ciliates were counted in an inverted microscope (Zeiss AXIOVERT35), and identified to genus level when possible. No ciliate samples were taken at CE1 station.

Bacterial production and mortality -

Experiments to determine bacterial (BP) and viral (VP) production, and bacterial losses due to protists (PMM) and viruses (VMM) were run using water samples from four selected depths: DFM, 200 m, 700 m (800 m for FF2 and CE2), and 1000 m. Those depths were chosen in order to select the most active ocean layer (DFM and

Boras et al.

6

Fig. 3. Temperature (Temp.), fluorescence (Fluor.), and nutrients values along a depth profile at far-field (FF), cyclonic eddies (CE), and anticyclonic eddies (AE) stations in NE Atlantic during RODA1 cruise.


7

200 m, which is a boundary of the euphotic zone), and potentially low-productive deeper layers (700 m - 1000 m). Depths of 700 m - 800 m correspond to the minimum oxygen concentration layer, selected as an example of mesopelagic waters, and 1000 m was chosen as an example of deep sea waters.

Bacterial mortality due to protists was evaluated following the fluorescent labeled bacteria (FLB) disappearance method (Sherr et al. 1987). For each grazing experiment, duplicates (1 L of seawater each) and one control (1 L of virus-free water) were prepared in 2 L polycarbonate bottles. All bottles (control and duplicates) were inoculated with FLB at 20% of the natural bacterial concentration. The FLB were prepared with a culture of Brevundimonas diminuta (strain obtained from the Spanish Type Culture Collection http://www.cect.org/index2.html). B. diminuta was heat-killed and stained with DTAF (5-([4,6 dichlorotriazin-2yl) amino]-fluorescein). Bottles were incubated in a thermostatic chamber, simulating in situ temperature and in the dark. Samples from DFM were incubated during 48 h, and samples from the rest of depths during 72 h. Samples for evaluation of bacterial and FLB's abundances were taken at the beginning and at the end of the experiments. Abundances of bacteria and FLB were assessed by epifluorescence microscopy (Olympus BX40-102/E; a 1000X magnification). To this end, aliquots of 20 mL sample were filtered through 0.2 μm black polycarbonate filters, and stained with DAPI at a final concentration of 5 μg mL-1. Natural bacteria were identified by their blue fluorescence when excited with UV radiation, while FLB were identified by their yellow-green fluorescence when excited with blue light. Control bottles showed no decrease of FLB during the experiments.

Grazing rates of bacteria were obtained following the equations of Salat and Marrasé (1994), based on the specific grazing rate (g, FLB mL-1) and the specific net growth rate (a, cells mL-1), and calculated as follows:

g = - (1 / t) ln (Ft / F0), (1)

a = (1 / t) ln (BAt / BA0), (2) where t is the incubation time; Ft is the abundance of FLB at final time, F0 the abundance of FLB at initial time, and BAt and BA0 are bacterial abundances at the end and at the beginning of the experiment respectively.

Net bacterial production (BPN, cells mL-

1 d-1) in the incubation bottles was obtained:

BPN = BA0 × (eat – 1). (3) Then, grazing rate (G, cells mL-1 d-1)

was calculated:

G = (g / a) × BPN. (4) Finally, protist-mediated mortality of

bacteria (PMM, % d-1) was calculated as the percentage of bacterial standing stock (BSS) and BP:

PMMBSS = (G × 100) / BA0, (5)

PMMBP = (G × 100) / BP. (6) To determine viral production and

bacterial losses due to phages, we followed the virus reduction approach (Wilhelm et al. 2002). Briefly, 2 L of seawater were prefiltered through 0.8 mm pore size cellulose filter (Whatman) (except 1000 m depth samples), and then concentrated by a spiral-wound cartridge (0.22 mm pore size, VIVAFlow 200), obtaining 40 mL of bacterial concentrate. Virus-free water was collected filtering one liter of seawater using a cartridge of 100 kDa molecular mass cutoff (VIVAFlow 200). A mixture of virus-free water (160 mL) and bacterial concentrate (40 mL) was prepared and distributed into 4 sterile falcon plastic tubes. Two of the tubes were maintained without any manipulations as controls, while in other two, mitomycin C (Sigma) was added (1 µg mL-1 final concentration) as inducing agent of the lytic

Boras et al.

8

cycle in prophages. All falcon tubes were incubated in a thermostatic chamber simulating in situ temperature and in the dark, during 12 hours. Samples for viral and bacterial abundances were collected at time zero and every 4 hours of the experiment, fixed with glutaraldehyde and stored as described before for viruses. Viruses and bacteria from viral production experiments were counted by flow cytometry. The number of viruses released by bacterial cell (burst size, BS) was estimated from viral production experiments, as in Middelboe and Lyck (2002), Wells and Deming (2006) and Boras et al. (2009). Increase of viral abundance during short time intervals (4 h) in viral production experiments was divided by a decrease of bacterial abundance in the same period of time. We assumed that the bacterial production and viral decay in this time interval were negligible. The estimated BS ranged from 5 to 299 viruses per cell.

Virus-mediated mortality of bacteria (VMM) was determined as previously described in Weinbauer et al. (2002) and Winter et al. (2004). Briefly, viral increase in the control tubes represents lytic viral production (VPL), and the difference between viral increase in the mitomycin C treatments and VPL gives the lysogenic production (VPLG). To compare the values of VPL and VPLG from different experiments, we multiplied them by the bacterial loss factor (0.4 – 8.9), as the loss of the part of in situ BSS during tangential flow filtration was observed (Winget et al. 2005). Then, we calculated the rate of lysed cells (RLC, cells mL-1 d-1) dividing VPL by burst size (BS), following the method of Guixa-Boixereu (1997):

RLC = VPL / BS. (7) RLC was used to calculate VMM as a

percentage of bacterial standing stock (VMMBSS, % d-1):

VMMBSS = (RLC × 100) / BA0, (8) where BA0 is the initial bacterial abundance in the viral production experiment.

Assuming that percentage of losses of BSS due to viruses is the same in falcon tubes and in the grazing bottles, we used VMMBSS to calculate the rate of lysed bacteria in the grazing experiment (RLCGR, cells mL-1 d-1):

RLCGR = (VMMBSS × BAGR) / 100, (9)

where BAGR is bacterial abundance in the grazing bottles at time 0.

Finally, using RLCGR, VMM as a percentage of bacterial production (VMMBP, % d-1) could be calculated:

VMMBP = (RLCGR × 100) / BP, (10) where BP is bacterial production.

Bacterial production (cells mL-1 d-1) was calculated as the sum of net bacterial production (BPN), grazing rate (G) and rate of lysed cells (RLCGR):

BP = BPN + G + RLCGR. (11)

Nutrient fluxes – Nutrient fluxes from bacteria to the dissolved organic matter (DOM) pool, or to higher trophic levels as POM were determined transforming lysed or consumed bacteria to C, nitrogen (N), and phosphorous (P) units, using the following cell factors: 12 fg C cell-1 (Simon and Azam 1989), 5.6 fg N cell-1 (Simon and Azam 1989), and 0.5 fg P cell-1 (Fagerbakke et al. 1996).

Statistical analyses - The Shapiro-Wilk

W-test was used to check normal distribution of data, and data were logarithmically transformed prior to analyses if necessary. One-way ANOVA for normal distributions and Wilcoxon test for non-normal distributions were used to evaluate the differences between the three types of stations, or between water layers. Pearson correlation and regression analyses were used to determine the relationships between the various properties examined. These statistical analyses were performed using the JMP program. Canonical correspondence analysis (CCA) was performed to evaluate multivariate patterns in the data, using the XLSTAT-ADA software.


9

Fig. 4. Viral and bacterial abundances along a depth profile at far-field (FF), cyclonic eddy (CE), and anticyclonic eddy (AE) stations in NE Atlantic during RODA1 cruise. Depth, fluorescence, salinity, and temperature were used as predictor (abiotic) variables, and viral, bacterial, and HNF abundances, viral and bacterial productions, and losses of bacterial production due to phages and grazers as response variables (biotic). The permutation test confirmed the significance of both canonical correspondence axes. Results

Oceanographic conditions - Cyclonic and anticyclonic eddies were identified during the cruise by satellite sea-surface temperature (AVHRR). Once the approximate location was obtained, high-resolution expendable bathythermographs (XBT) transects were carried out to determine the thermal gradients of the mesoscale eddy field (Fig. 2). Four eddies were selected for this study: two cyclonic, CE1 and CE2, and two anticyclonic, AE1 and AE2 (see Fig. 1 for locations, and Fig. 2 for vertical potential temperature sections). Four layers of the water column were studied: epipelagic (< 200 m), that includes the euphotic zone, transition (200 m to < 500 m), mesopelagic (500 m to 800 m) and bathypelagic (> 800 m to 2000 m). Average water temperature ranged from 24.8ºC at the surface, to 3.9ºC at 2000 m (Fig. 3), and salinity varied from 37.3 to 35.1, from 5 m to 2000 m (Table 1). Nitrites + nitrates varied between 0.04 μmol L-1 at 5 m, and 11.77 μmol L-1 at 1000 m (Fig. 3), increasing with depth. Phosphate concentration was below the detection level from the surface to ca. 120 m at far-field and anticyclonic eddy stations, being detectable however at cyclonic eddy station (Fig. 3). There, concentration of PO4

-3 increased with depth, reaching a maximum of 1.63 μmol L-1 at 1000 m. Ammonium levels varied between 2.19 μmol L-1 in the upper 60 m, and below the detection level at 500 m and deeper

Boras et al.

10

(Fig. 3). The nitrites + nitrates concentrations were significantly higher in the epipelagic layer of cyclonic eddy, than at far-field and anticyclonic eddy stations (F1, 13 = 5.4, p < 0.05), and ammonium levels were significantly higher in the epipelagic layer of anticyclonic eddy than at cyclonic eddy and far-field stations (F1, 13 = 74.9, p < 0.01). Fluorescence (relative units) values, as indicator of chlorophyll a concentration, ranged from 0.03 to 1.0 (Fig. 3). The DFM changed its depth with the occurrence of eddies. The DFM in cyclonic eddies was found at 56 and 68 m (CE2 and CE1, respectively), whereas at the far-field stations it was located much deeper, at 120 m (Fig. 3). The DFM in anticyclonic eddies was observed at 59 and 75 m (AE2 and AE1, respectively; Fig. 3). No significant differences in fluorescence were found among the epipelagic layer of the three types of stations.

Distribution, abundance and production

of microorganisms – Viral abundances (VA) ranged between 1.1 × 105 viruses mL-1 at 2000 m and 8.1 × 106 viruses mL-1 at 50 m, and bacterial abundances ranged between 1.5 × 104 cells mL-1 at 2000 m and 7.8 × 105 cells mL-1 at 5 m (Fig. 4). At all stations, the highest viral and bacterial (BA) abundances were found in the epipelagic layer, above or in the DFM, and decreased significantly with depth (Table 2). No differences in VA were found among the three types of stations, whereas BA was significantly higher in cyclonic and anticyclonic eddies than at far-field stations, yet only for the transition layer (F1,15 = 11.9, p < 0.01). Both, VA and BA, showed strong positive correlation with fluorescence (Table 2). Virus to bacterium ratio (VBR) varied between 1.8 and 26.0, decreasing significantly with depth (Table 2). The lowest VBR was overall found between 800 m and 2000 m. HNF abundance decreased with depth (Table 2, Fig. 5), reaching the lowest values in the bathypelagic layer (0.1 × 102 cells mL-1 at 1000 m), and highest in the epipelagic layer (1.8 × 103 cells mL-1 at 5 m). Significant differences in HNF abundance between the three types of stations were detected only for the epipelagic layer, with highest abundances at cyclonic eddy

Fig. 5. Heterotrophic nanoflagellates (HNF) and ciliate abundances along a depth profile at far-field (FF), cyclonic eddy (CE), and anticyclonic eddy (AE) stations. No ciliate sample was taken at CE1.


11

Fig. 6. (A) Bacterial production BP, and (B) lytic (VPL) and lysogenic (VPLG) viral production, detected at far-field (FF), cyclonic eddy (CE), and anticyclonic eddy (AE) stations. Each value corresponds to experimental duplicates, bars indicate the maximum and minimum values of duplicates. The VPLG percentages of total VP are shown.

stations, and the lowest at far-field stations (F2,8 = 8.0, p < 0.02). Slightly, but not statistically significant, higher HNF abundances were detected at anticyclonic eddy stations compared to far-field or cyclonic eddy stations in transition and mesopelagic layers (Table 1). Ciliate abundance declined with depth, from a maximum of 3.0 × 103 cells L-1 in the DFM, to a minimum of 5 cells L-1 at 1000 m (Fig. 5). At all stations, except for AE1, a peak of abundance was detected in the DFM. No significant

differences between the three types of station were found. The ciliate community was composed mainly of the autotrophic Mesodinium sp. (range 23.3% – 55.6%), followed by Strombidium sp. (range 20.1% – 40.8%) in the euphotic zone, and by Strobilidium sp. (range 19.5% – 100%) and tintinnids (37.4% – 73.9%) in the dark ocean zone. Bacterivorous ciliates, as scuticociliates, were very rare in the collected samples (from 1.7% to 16.4%). Positive correlations between

Boras et al.

12

Fig. 7. Virus- (VMM) and protist- (PMM) mediated mortality of bacteria, as a percentage of bacterial production (BP) detected at far -field (FF), cyclonic eddy (CE), and anticyclonic eddy (AE) stations. Bars indicate the maximum and minimum values of duplicates. the abundance of all components of the microbial food web: VA, BA, HNF, and ciliates were found (Table 2).

Bacterial production (BP) decreased with depth (Fig. 6 A), and increased with fluorescence and microorganisms abundances

(VA, BA, HNF; Table 2). No differences in BP between stations were detected (Table 1).

Lytic viral production (VPL) varied between 4.3 × 104 viruses mL-1 d-1 at 200 m and 2.8 × 106 viruses mL-1 d-1 at 1000 m (Table 1, Fig. 6 B). At four out of six stations, the highest VPL was detected in the DFM. High VPL values


13

Fig. 8. Plot of the rate of lysed cells (RLC) against grazing (G) detected at the three types of stations:

far-field (FF; open symbols), cyclonic (CE; black symbols), and anticyclonic eddies (AE; grey symbols). Different depth layers (epipelagic, transition, mesopelagic, and bathypelagic) are marked with diamonds, triangles, circles, and squares; dashed line is the 1:1 line. Values above the line represent RLC > G; values below the line represent RLC < G. Grazing rates = 0 on the plot correspond to ‘not detectable’. were also found at 1000 m at four stations, for instance at AE2 station VPL was ca. 4-fold higher at 1000 m than in the DFM. No differences in VPL between far-field, cyclonic and anticyclonic eddy stations were found in any of the layers. In 46% of viral production measurements, prophage induction was detected, and the percentage of lysogenic viral production (%VPLG) with respect to the total VP (VPT) varied from 13.4% at 700 m to 84.6% in DFM (Fig. 6 B). In one of anticyclonic eddies (AE2), lysogeny was detected at all analyzed depths. The %VPLG increased with increasing fluorescence, VA, HNF abundance, and BP (Table 2).

Bacterial mortality - Losses of bacterial

production due to viruses (VMMBP) varied between 20.6% d-1 at 200 m, and 92.0% d-1 in DFM (Fig. 7), and did not differ significantly among depths (Table 1). Stations in anticyclonic eddies were characterized by significantly higher VMMBP in the epipelagic layer (F2, 3 = 31.9, p < 0.01) than in cyclonic eddy or far-field

stations (Table 1). Losses of BSS due to viruses (VMMBSS) increased with increasing BP (Table 2), and did not show any differences among the three types of stations. At far-field stations, VMMBSS showed a decrease with depth down to a minimum of 3.8% d-1 at 700 m, and then an increase up to 55.8% d-1 at 1000 m (data not shown). Bacterial production grazed by protists (PMMBP) ranged from 0.8% d-1 at 1000 m to 79.4% d-1 at 200 m (Fig. 6). In three occasions, grazing was not detectable (Table 1). The PMMBP did not show a clear pattern at far-field stations, whereas at cyclonic stations reached the lowest (0.8% d-1) values at 1000 m (Fig. 7). Significant differences in PMMBP between the three types of stations were detected only for the mesopelagic layer (F2,3 = 92.1, p < 0.01), with the highest values at anticyclonic stations. Losses of BSS due to protists (PMMBSS) followed the same pattern as PMMBP (data not shown), and increased significantly with BP (Table 2). Anticyclonic eddies supported the highest, and cyclonic eddies the lowest PMMBSS values (F2,3 = 10.7, p < 0.05).

Boras et al.

14

Plotted values of rates of lysed against grazed bacteria (cells mL-1 d-1) showed that in anticyclonic eddies losses of bacteria due to viral lysis were higher than due to grazing for epipelagic layer (DFM), and close to the line 1:1 for all the other depths (Fig. 8). In cyclonic eddies, losses due to phages were higher than due to protists at all depths, except for epipelagic layer (DFM). At far-field stations viral lysis dominated in the majority of samples for all depths, except for two samples: from the epipelagic (DFM) and transition (200 m) layers, where both types of mortality were similar (Fig. 8). The CCA analysis showed that data were clustered mostly within the sampling depths (data not shown), indicating that this was the main factor shaping the biological variability among samples. However, samples from the same types of stations also clustered together within each depth. The permutation test showed that the biotic data had a linear relation with abiotic data (p < 0.001) Discussion

Methodological considerations - The methods to estimate complex ecosystem variables, such as bacterial production and losses of bacteria, have their suite of assumptions and uncertainties. In most cases these methods give high reproducibility of results among duplicates, even dealing with natural communities. However, in some experiments they resulted in relatively high range of values, as we observed e.g., for PMMBP (FF1, 700 m).

The FLB disappearance technique (Sherr et al. 1987) was used to measure the grazing rates on bacteria by the whole community of protists. Taking into account that only a small part of the ciliates counted was strictly bacterivorous, such as scuticociliates, the main grazers in our system should be small HNF (≤ 5 µm) (Sherr and Sherr 2002). The use of FLB as bacterial surrogates has known limitations, such as disturbance of the sample and underestimation of bacterial losses due to prey selection (Christoffersen et al. 1997). This

technique has been widely discussed in Vaqué et al. (1994).

Bacterial production was estimated as the sum of the net increase of bacterial abundance (BPN), losses of bacteria due to protists (G) and viruses (RLCGR). This method was employed in previous studies, and results of BP were comparable, although slightly higher, with those obtained with in situ measurements of BP by 3H-leucine incorporation method (Boras et al. 2009).

Estimations of bacterial mortality due to viruses are subjected to the calculation of the burst size. We estimated the BS taking into account the increase of viral abundance in viral production (VP) experiments, and the decrease of bacterial abundance in the same period of time (4 h). We assumed that the only cause of bacterial abundance decrease over short time periods was cell lysis, and we did not take into account the viral decay and bacterial production during this time interval. Burst size values reported in this study, between 5 and 299, are within the range obtained from different aquatic environments (Guixa-Boixereu et al. 1996; Parada et al. 2006), and are lower than values found in the anoxic layer of an eutrophic lake (~500; Weinbauer and Höfle 1998). Also, bacterial losses caused by lysis measured in our study are the potential losses, as we did not consider grazing on infected cells by protists in our calculations. This process could reduce the percentage of bacteria that burst due to viral activity in natural communities.

The virus reduction approach (VRA) used to assess VP and VMM is based on the assumption that all viral production observed during the experiment is a result of infections prior to incubation. It is also assumed that no new infections occur, and that both filtration and incubation, do not induce lysogenic bacteria. This method allows direct observation of changes in viral abundance over time, and the distinction between production of virulent and temperate phages. In addition, it is relatively easy and inexpensive to perform (Winget et al. 2005). Detection of lysogeny is based on lysis induction by mitomycin C. It is known, that in some circumstances (i.e., nutrients availability,


15

pH) this agent is not sufficient to induce the lytic cycle in all prophages (Weinbauer and Suttle 1999). However, this method is widely used and mitomycin C is better suited than other inducing agents, hence the obtained results are comparable to other studies. Finally, the main drawbacks of this technique are the considerable levels of sample manipulation and the loss of a portion of bacterial community during filtration. Despite these disadvantages, VRA is considered one of the best-suited incubation-based methods for VP and VMM estimations (Winget et al. 2005).

Distribution, abundance, and

production of microorganisms - Eddies studied during the RODA1 cruise are common mesoscale features in the region of Canary Islands, NE Atlantic, and are generated by the interaction of the islands with the southward flowing Canary Current and the intrusion of coastal upwelling waters into the Canary region. Previous studies have shown that in addition to the nutrients rising from the lower water layers, rich upwelling waters are incorporated into cyclonic eddies in this region (Arístegui et al. 1997), thereby enriching the otherwise oligotrophic waters around the Canaries archipelago. Our results show a clear contrasting pattern of nutrient concentrations in the epipelagic layer of cyclonic and anticyclonic eddies, and far-field stations. Higher concentrations of nitrates + nitrites and phosphates in cyclonic eddy than in anticyclonic eddy and far-field stations were observed. Moreover, higher ammonium concentrations at anticyclonic stations suggest intense recycling processes therein.

In open-ocean waters, viral and bacterial abundances have been reported to be highest in the epipelagic layer (down to 200 m), and decline reaching relatively constant, low concentrations in deep waters (Hara et al. 1996). Similar pattern for both viral and bacterial abundances was observed in this study, with the highest abundances between 5 m and 120 m. We found peaks of viral and bacterial

abundances above and in the DFM, consistent with previous reports from the North Pacific (Hara et al. 1996), and from the Eastern Tropical Atlantic Ocean (Winter et al. 2008). Strong positive correlation between viral and bacterial abundance throughout the water column, consistent with other reports (Hara et al. 1996; Magagnini et al. 2007) suggests that bacteriophages may be an important fraction of the viral community in that region. Also, the positive correlation of both, viral and bacterial abundances with fluorescence suggests that viruses of photosynthetic organisms may also be a significant fraction of virioplankton in the epipelagic layer. The virus-bacterium ratios reported here are similar to the values reported for other marine waters (Wommack and Colwell 2000), but showed a decreasing trend with depth, in contrast with data reported by other studies (Magagnini et al. 2007). Protists (HNF and ciliates) abundance also decreased with depth, reaching very low (≤ 10 cells mL

-1) values in transition, mesopelagic and bathypelagic layers, similarly to abundances found by Gasol et al. (2009). Bacterial production declined with depth, as reported in the past (Reinthaler et al. 2006), with a difference of an order of magnitude between the values in the DFM and at 1000 m.

Stations influenced by eddies, either cyclonic or anticyclonic, showed higher viral and bacterial abundances than stations away from the eddy field in most of the layers. This could be due to the input of nutrients from deeper to shallower waters in cyclonic eddies, or due to an increase of ammonium concentration originated from the remineralisation of the organic matter (OM) accumulated in the centre of anticyclonic eddies (Nelson et al. 1995), as observed at the AE2 station. Similarly to Lochte and Pfannkuche (1987), we have found no differences in BP between eddies’ and far-field stations. However, we detected slightly higher bacterial and HNF abundances in the transition, meso- and bathypelagic layers of anticyclonic eddy stations than at cyclonic eddy and far-field

Boras et al.

16

Table 3. Average carbon and organic nutrients fluxes shunted from bacteria to higher trophic levels by protistan grazing (PMM), and to dissolved organic matter pool by viral lysis (VMM). Averaged values for all water column (± SD, n = 8), epipelagic + transition layers (epi., trans.; ± SD, n = 4), and for mesopelagic + bathypelagic layers (meso., bathy.; ± SD, n = 4). * values of C, N, and P shunted by PMM are significantly lower than by VMM (p < 0.01).

Layer and

station

Carbon μg L-1 d-1

Nitrogen μg L-1 d-1

Phosphorous μg L-1 d-1

PMM VMM PMM VMM PMM VMM

All water

column

FF 0.27 ± 0.33 0.65 ± 0.27 0.13 ± 0.15 0.31 ± 0.12 0.01 ± 0.01 0.03 ± 0.01

CE 0.41 ± 0.31 0.75 ± 0.17 0.19 ± 0.14 0.35 ± 0.08 0.02 ± 0.01 0.03 ± 0.01

AE 0.37 ± 0.04 1.38 ± 0.49 0.17 ± 0.02 0.64 ± 0.23 0.02 ± 0.00 0.06 ± 0.02

epi.

trans.

FF 0.50 ± 0.65 1.02 ± 0.33 0.23 ± 0.30 0.48 ± 0.15 0.02 ± 0.03 0.04 ± 0.01

CE 0.73 ± 0.73 0.74 ± 0.56 0.34 ± 0.34 0.34 ± 0.26 0.03 ± 0.03 0.03 ± 0.02

AE 0.53 ± 0.04 2.40 ± 1.23 0.25 ± 0.02 1.12 ± 0.57 0.02 ± 0.00 0.10 ± 0.05

meso.

bathy.

FF * 0.05 ± 0.01 0.29 ± 0.20 0.02 ± 0.00 0.13 ± 0.09 0.00 ± 0.00 0.01 ± 0.01

CE 0.10 ± 0.11 0.76 ± 0.90 0.05 ± 0.05 0.36 ± 0.42 0.00 ± 0.00 0.03 ± 0.04

AE 0.21 ± 0.12 0.35 ± 0.25 0.10 ± 0.06 0.16 ± 0.12 0.01 ± 0.01 0.01 ± 0.01

stations, suggesting that the deepening of the surface water masses by anticyclonic eddies could alter the vertical distribution of microorganisms.

The highest lytic VP was detected in the DFM, but it was also high at 1000 m, occasionally exceeding the VPL in the epipelagic layer (AE2, CE2). It is not a surprising result, considering the relatively high viral abundances and VBR found in deeper waters, which would suggest also a high viral production. Parada et al. (2007) reported similar viral abundances in bathypelagic waters (> 1000 m) for the subtropical North Atlantic, and

proposed a contribution of allochthonous viral input or lysogenic production. In our study, however, the measured VPL would be sufficient to maintain the viral abundance found in dark waters. Detected values of VPL are comparable with viral production observed in euphotic (Rowe et al. 2008) and deeper (Parada et al. 2007) waters in the same area. Rowe et al. (2008) reported the VP that ranged from ca. 0.5 × 106 to 7.2 × 106 viruses mL-1 d-1, and in our study we observed VPL from 0.1 × 106 to 1.7 × 106 viruses mL-1 d-1 (Table 2). Parada et al. (2007) reported VP values at the level of ca. 0.1 × 106 viruses mL-1 d-1 in deeper waters (900 –


17

1100 m), which are similar to our observations for stations FF2 and CE1 at 1000 m. However, at the other stations we detected higher VPL, which exceeded those numbers by an order of magnitude.

In oligotrophic systems, as well as in deep ocean waters, lysogeny should predominate over lytic infections as a survival strategy of phages at low host densities (Stewart and Levin 1984). This was confirmed by some studies, as higher percentage of lysogens was found in offshore compared to coastal environments (Weinbauer and Suttle 1999), and in deep compared to surface marine waters (Weinbauer et al. 2003). Our results show lower percentage of lysogeny in samples from deeper waters than from the upper ocean layers, with the highest values detected in DFM at stations influenced by eddies, although no statistically significant differences between depths were detected. Moreover, contrary to the results reported in Weinbauer et al. (2003), the percentage of lysogeny increased with increasing bacterial production, and no relationship with bacterial abundance was observed, suggesting that factors other than host density could also affect the prevalence of lysogeny.

Bacterial mortality - Losses of bacterial

production caused by viral infection prevailed over losses due to protists in 92% of the examined cases, with only two communities where mortality caused by protists exceeded that caused by viruses. In 25% of cases, phages caused more than 80% of losses of bacterial production, mostly in meso- and bathypelagic layers, but also in the DFM at stations influenced by anticyclonic eddies. These values are among the highest virus-induced bacterial mortality rates reported for oligotrophic waters (Wommack and Colwell 2000). In eastern tropical Atlantic Ocean the frequency of infected bacterial cells (which correspond to VMMBSS in this study), found by Winter et al. (2008), ranged between 2.4% in deeper waters (75 – 1000 m) and 8.3% in the mixed surface

(15 – 25 m) and DFM (50 m) layers. Virus-induced mortality of bacterioplankton measured by Weinbauer et al. (2003) in Mediterranean waters varied between 2.6% in deep waters (800 m to 2000 m) and 14.8% in surface waters (down to 100 m). On the other hand, Brum (2005) found that an important part of BSS was lysed due to viral infections at the oligotrophic station ALOHA, North Pacific Subtropical Gyre, with a loss of 3.2% of BSS h-1 (which corresponds to 63% d-1) at 5 m depth, and about 5 times more at 75 m. These examples show a great variability in rates of bacterial mortality due to viruses, which may be partly an effect of the variability of the methods used.

Protist-mediated mortality rates of bacteria obtained here were comparable to results from mesopelagic waters (down to 500 m) of the East Sea (Cho et al. 2000), ranging from 0.1 - 1.1 × 103 cells mL-1 h-1, Cho et al. (2000) concluded, however, that HNF were responsible for losses of most (up to 100%) of bacterial production in this region, but they did not evaluate experimentally mortality due to viruses. Indeed, a later study of Hwang and Cho (2002) reported that phages and protists had a comparable effect on bacterioplankton in that system. We observed a high grazing activity in transition, meso- and bathypelagic layers at stations under the influence of anticyclonic eddies, possibly related to the higher abundance of HNF and bacterial production therein.

Finally, we estimated the importance of protistan grazing vs. viral lysis of bacteria for the nutrient fluxes (C, P, N) within the microbial web, making a rough estimation of the POM shunted from bacterial cells to higher trophic levels by HNF, and to DOM by phages. Values obtained for the far-field stations were comparable to the range of fluxes of N and C shunted by viruses reported for the oligotrophic offshore waters of the Gulf of Mexico, with relatively steady-state conditions (Wilhelm et al. 1998). In that system viruses released 0.12 – 0.55 μg C L-1 d-1, compared to 0.65 ± 0.27 μg C L-1 d-1 estimated in our system (average for the whole water column, n = 8, Table 3). In

Boras et al.

18

contrast, fluxes of DOM assessed for the anticyclonic and cyclonic stations were similar or lower to those reported for the Strait of Georgia, British Columbia (Wilhelm and Suttle 2000), characterized by turbulently mixed waters. In those conditions the range of the C flux was 1.03

– 1.51 μg C L-1 d-1 in stratified waters, and 2.00

– 8.25 μg C L-1 d-1 in tidally mixed waters, while our estimations resulted in 0.75 ± 0.17 μg C L-1 d-1 at CE stations, and 1.38 ± 0.49 μg C L-1 d-1 at AE stations (averages from the epi- and transition layers, n = 4, Table 3). The release of DOM from bacterioplankton by cell lysis was slightly higher than the POM ingestion by HNF at far-field and anticyclonic stations in epipelagic and transition layers, although this trend was not statistically significant (Table 3). However, in meso- and bathypelagic layers of far-field stations, the nutrient fluxes shunted by viruses were significantly higher than shunted by protists (F1,6 = 15.5, p < 0.01; Table 3). This relation was not found for other stations. Finally, POM and DOM fluxes showed slightly higher values (although no significant) in eddy than at far-field stations, for all layers, indicating that these biological processes could be influenced by mesoscale features.

In summary, as expected in anticyclonic eddies HNF supported the highest pressure on bacteria, and in cyclonic eddies the lowest one. Furthermore, in anticyclonic eddies lysogeny was detected at all depths, however it was not always detected in deep waters at the other type of stations. Contrary to our expectations, loss of bacterial production due to viruses did not differ significantly among depths, except for the stations situated in anticyclonic eddies, where it was significantly higher in the epipelagic layer. In addition, low values of bacterial losses due to protists were detected at stations outside the influence of eddies. Our results indicate that viruses are responsible for losses of an important fraction of bacterial production, as well as for the release of significant amounts of C and nutrients to the DOM pool in the studied system. Furthermore, the importance of viruses in the bacterial C cycling was higher in eddies,

both cyclonic and anticyclonic, and in the deeper layers of mesopelagic zone. Regardless of the low bacterial abundances in deeper waters (1000 m), at these depths high viral production and higher percentage of bacterial losses due to viruses than due to grazers were found. Our findings indicate that viral lysis could be a dominant pathway for the flow of bacterial C in the NE Subtropical Atlantic, particularly significant in areas dominated by eddy activity.

Acknowledgments

We thank E. Martinez, E. Cabanillas, and A. Gil for helping in nanoflagellates’ and

ciliates’ counts, and R. Martínez and J. C.

Alonso for their help with nutrient analyses. We are also very grateful to the staff of Unidad de Tecnología Marina (Consejo Superior de Investigaciones Científicas) and the crew of the Buque de Investigación Oceanográfica - Hespérides for their help during the cruise. This study was supported by the following projects: ‘Remolinos Oceánicos y Deposición Atmosférica en la Corriente de Canarias (RODA)’, ‘Protozoos y Virus: Control de la

Biomasa y la Diversidad de Procariotas u su Repercusión en los Ciclos Biogeoquímicos en una Zona Costera del Mediterraneo Nor-Occidental (PROCAVIR)’, and ‘Aislamiento,

Identificación y Especificidad de Virus que Infectan a Microorganismos Marinos (MICROVIS)’ (CTM2004-06842 /MAR; CTM2004-04404-CO2-00/MAR; CTM2007-62140), funded by the Spanish Ministerio de Ciencia e Innovación. J. A. B.’s work was

supported by a Ph.D. fellowship from the Spanish Ministerio de Ciencia e Innovación (Formación de Profesorado Universitario grant), and M. M. S. by I3P-CSIC postdoctoral contract funded by the Fondo Social Europeo. This is a contribution to the European Network of Excellence EurOceans.


19

References

Arístegui, J., P. Tett, A. Hernández-Guerra, G. Basterretxea, M. F. Montero, K. Wild, P. Sangrá, S. Hernández-León, M. Cantón, J. A. García-Braun, M. Pacheco, and E. D. Barton. 1997. The influence of island-generated eddies on chlorophyll distribution: a study of mesoscale variation around Gran Canaria. Deep-Sea Res. I 44: 71-96.

Boras, J. A., M. M. Sala, E. Vázquez-Domínguez, M. G. Weinbauer, and D. Vaqué. 2009. Annual changes of bacterial mortality due to viruses and protists in an oligotrophic coasta environment (NW Mediterranean). Environ. Microbiol. 11: 1181-1193.

Brum, J. R. 2005. Concentration, production and turnover of viruses and dissolved DNA pools at Stn ALOHA, North Pacific Subtropical Gyre. Aquat. Microb. Ecol. 41: 103-113.

Brussaard, C. P. D. 2004. Optimisation of procedures for counting viruses by flow cytometry. Appl. Environ. Microbiol. 70:

1506-1513. Cho, B. C., S. C. Na, and D. H. Choi. 2000.

Active ingestion of fluorescently labeled bacteria by mesopelagic heterotrophic nanoflagellates in the East Sea, Korea. Mar. Ecol. Prog. Ser. 206: 23-32.

Christoffersen, K., O. Nybroe, K. Jürgens, and M. Hansen. 1997. Measurement of bacterivory by heterotrophic nanoflagellates using inmunofluorescence labeling of ingested cells. Aquat. Microb. Ecol. 13: 127-134.

del Giorgio, P. A., D. F. Bird, Y. T. Prairie, and D. Planas. 1996. Flow cytometric determination of bacterial abundance in lake plankton with the green nucleid acid stain SYTO 13. Limnol. Oceanogr. 41: 783-789.

del Giorgio, P. A., J. J. Cole, and A. Cimbleris. 1997. Respiration rates in bacteria exceed phytoplankton production in unproductive aquatic systems. Nature 385: 148-151.

Fagerbakke, K. M., M. Heldal, and S. Norland. 1996. Content of carbon, nitrogen, oxygen, sulphur and phosphorus in native aquatic and cultured bacteria. Aquat. Microb. Ecol. 10: 15-27.

Falkowski, P. G., D. Ziemann, Z. Kolber, and P. K. Bienfang. 1991. Role of eddy pumping in enhancing primary production in the ocean. Nature 352: 55-58.

Fuhrman, J. A. 1999. Marine viruses and their biogeochemical and ecological effects. Nature 399: 541-548.

Gasol, J. M., L. Alonso-Sáez, D. Vaqué, F. Baltar, M. L. Calleja, C. M. Duarte, and J. Arístegui. 2009. Mesopelagic prokaryotic bulk and single-cell heterotrophic activity and community composition in the NW Africa-Canary Islands coastal-transition zone. Prog. Oceanogr. doi:10.1016/j.pocean.2009.07.014.

Guixa-Boixereu, N. 1997. Abundance and dynamic of viruses in planktonic ecosystems. Ph.D. thesis. Universitat de Barcelona.

Guixa-Boixereu, N., J. I. Calderón-Paz, M. Heldal, G. Bratbak, and C. Pedrós-Alió. 1996. Viral lysis and bacterivory as prokaryotic loss factors along a salinity gradient. Aquat. Microb. Ecol. 11: 215-227.

Hansen, K., and F. Koroleff. 1999. Detemination of nutrients, p. 159-228. In K. Grasshoff, K. Kremling, and M. Ehrhardt [eds.], Methods of seawater analysis. Wiley-VCH, Weiheim.

Hara, S., I. Koike, K. Terauchi, H. Kamiya, and E. Tanoue. 1996. Abundance of viruses in deep oceanic waters. Mar. Ecol. Prog. Ser. 145: 269-277.

Hwang, C. Y., and B. C. Cho. 2002. Virus-infected bacteria in oligotrophic open waters of the East Sea, Korea. Aquat. Microb. Ecol. 30: 1-9.

Lochte, K., and O. Pfannkuche. 1987. Cyclonic cold-core eddy in the eastern North Atlantic. II. Nutrients, phytoplankton and bacterioplankton. Mar. Ecol. Prog. Ser. 39: 153-164.

Boras et al.

20

Magagnini, M., C. Corinaldesi, L. S. Monticelli, E. D. Domenico, and R. Danovaro. 2007. Viral abundance and distribution in mesopelagic and bathypelagic waters of the Mediterranean Sea. Deep-Sea Res. I 54: 1209-1220.

Middelboe, M., and P. G. Lyck. 2002. Regeneration of dissolved organic matter by viral lysis in marine microbial communities. Aquat. Microb. Ecol. 27: 187-194.

Murray, A. G., and G. A. Jackson. 1992. Viral dynamics: a model of the effects of size, shape, motion and abundance of single-celled planktonic organisms and other particles. Mar. Ecol. Prog. Ser. 89: 103-116.

Nelson, D. M., H. W. Ducklow, G. L. Hitchcock, M. A. Brzezinski, T. J. Cowles, C. Garside, R. W. Gould Jr., T. M. Joyce, C. Langdon, J. J. McCarthy, and C. S. Yentsch. 1995. Distribution and composition of biogenic particulate matter in a Gulf Stream warm-core ring. Deep-Sea Res. 32: 1347-1369.

Olaizola, M., D. A. Ziemann, P. K. Bienfang, W. A. Walsh, and L. D. Conquest. 1993. Eddy-induced oscillations of the pycnocline affect the floristic composition and depth distribution of phytoplankton in the subtropical Pacific. Mar. Biol. 116: 533-542.

Parada, V., G. J. Herndl, and M. G. Weinbauer. 2006. Viral burst size of heterotrophic prokaryotes in aquatic systems. J. Mar. Biol. Ass. U.K. 86: 613-621.

Parada, V., E. Sintes, H. M. van Aken, M. G. Weinbauer, and G. J. Herndl. 2007. Viral abundance, decay, and diversity in the meso- and bathypelagic waters of the North Atlantic. Appl. Environ. Microbiol. 73: 4429-4438.

Reinthaler, T., H. van Aken, C. Veth, J. Arístegui, C. Robinson, P. J. le B. Williams, P. Lebaron, and G. J. Herndl. 2006. Prokaryotic respiration and production in the meso- and bathypelagic realm of the eastern and western North Atlantic basin. Limnol. Oceanogr. 51: 1262-1273.

Rowe, J. M., M. A. Saxton, M. T. Cittrell, J. M. DeBruyn, G. M. Berg, D. L. Kirchman, D. A. Hutchins, and S. W. Wilhelm. 2008.

Constraints on viral production in the Sargasso Sea and North Atlantic. Aquat. Microb. Ecol. 52: 233-244.

Salat, J., and C. Marrasé. 1994. Exponential and linear estimations of grazing on bacteria: effects of changes in the proportion of marked cells. Mar. Ecol. Prog. Ser. 104:

205-209. Sherr, B. F., E. B. Sherr, and R. D. Fallon.

1987. Use of monodispersed, fluorescently labeled bacteria to estimate in situ protozoan bacterivory. Appl. Environ. Microbiol. 53:

958-965. Sherr, E. B., and B. F. Sherr. 2002.

Significance of predation by protists in aquatic microbial food webs. Ant. Van Leewen. 81: 293-308.

Simon, M., and F. Azam. 1989. Protein content and protein synthesis rates of planktonic marine bacteria. Mar. Ecol. Prog. Ser. 51: 201-213.

Stewart, F. M., and B. R. Levin. 1984. The population biology of bacterial viruses: Why be temperate? Theor. Popul. Biol. 26: 93-117.

Vaqué, D., J. M. Gasol, and C. Marrasé. 1994. Grazing rates on bacteria: the significance of methodology and ecological factors. Mar. Ecol. Prog. Ser. 109: 263-274.

Weinbauer, M. G., and M. G. Höfle. 1998. Size-specific mortality of lake bacterioplankton by natural virus communities. Aquat. Microb. Ecol. 15: 103-113.

Weinbauer, M. G., and C. A. Suttle. 1999. Lysogeny and prophage induction in coastal and offshore bacterial communities. Aquat. Microb. Ecol. 18: 217-225.

Weinbauer, M. G., C. Winter, and M. G. Höfle. 2002. Reconsidering transmission electron microscopy based estimates of viral infection of bacterioplankton using conversion factors derived from natural communities. Aquat. Microb. Ecol. 27: 103-110.

Weinbauer, M. G., I. Brettar, and M. G. Höfle. 2003. Lysogeny and virus-induced mortality of bacterioplankton in surface, deep, and anoxic marine waters. Limnol. Oceanogr. 48: 1457-1465.


21

Wells, L. E., and J. W. Deming. 2006. Significance of bacterivory and viral lysis in bottom waters of Franklin Bay, Canadian Arctic, during winter. Aquat. Microb. Ecol. 43: 209-221.

Wilhelm, S. W., M. G. Weinbauer, C. A. Suttle, and W. H. Jeffrey. 1998. The role of sunlight in the removal and repair of viruses in the sea. Limnol. Oceanogr. 43: 586-592.

Wilhelm, S. W., and C. A. Suttle. 2000. Viruses as regulators of nutrient cycles in aquatic environments, p. 551-556. In C. R. Bell, M. Brylinsky and P. Johnson-Green [eds.], Microbial biosystems: New Frontiers. Atlantic Canada Society of Microbial Ecology.

Wilhelm, S. W., S. M. Brigden, and C. A. Suttle. 2002. A dilution technique for the direct measurement of viral production: a comparison in stratified and tidally mixed coastal waters. Microb. Ecol. 43: 168-173.

Winget, D. M., K. E. Williamson, R. R. Helton, and K. E. Wommack. 2005. Tangential flow diafiltration: an improved technique for estimation of virioplankton production. Aquat. Microb. Ecol. 41: 221-232.

Winter, C., G. J. Herndl, and M. G. Weinbauer. 2004. Diel cycles in viral infection of bacterioplankton in the North Sea. Aquat. Microb. Ecol. 35: 207-216.

Winter, C., M. M. Moeseneder, G. J. Herndl, and M. G. Weinbauer. 2008. Relationship of geographic distance, depth, temperature, and viruses with prokaryotic communities in the eastern tropical Atlantic Ocean. Microb. Ecol. 56: 383-389.

Wommack, K. E., and R. R. Colwell. 2000. Virioplankton: viruses in aquatic ecosystems. Microbiol. Mol. Biol. Rev. 64: 69-114.

Paper V

Effect of ice melting on bacterial

carbon fluxes channelled by

viruses and protists in the Arctic

Ocean

Summary – Paper V

P a g e | 139

Resumen

Efecto del deshielo sobre los flujos de carbono bacteriano

canalizado por virus y protistas en el Océano Ártico

Artículo enviado para su publicación a Polar Biology (Effect of ice melting on bacterial carbon

fluxes channeled by viruses and protists in the Arctic Ocean)

El efecto del deshielo sobre los flujos de carbono bacteriano canalizado por virus y

protistas se evaluó en aguas del Mar de Groenlandia y del Océano glacial Ártico durante julio de

2007. El estudio se llevó a cabo principalmente en dos masas de agua: unas más calientes y más

salinas pertenecientes a Aguas Atlánticas (AW), y otras más frías y menos salinas, correspondientes

a Aguas Polares Superficiales mezcladas con AW (PSWm). En las AW se detectaron valores de

concentraciones de nutrientes (fosfato y nitrato+nitrito), y abundancias bacterianas

significativamente mas elevados que en las PSWm. En cambio en las PSWm se detectaron

mayores abundancias de virus. Se observó infección lisogenica en 7 casos de un total de 21

muestras, y constituía un porcentaje variable de la producción vírica total, desde un 33.9 % hasta

un 100 %. En las estaciones visitadas la mortalidad bacteriana debida a protistas (PMM) era más

alta que la mortalidad causada por los virus (VMM). Los protistas eliminaban un 45.4 ± 6.3% de la

producción bacteriana (BP) d-1, mientras que los virus fueron responsables de eliminar el 8.5 ± 6.9%

de BP d-1 a la profundidad de 0.1 m. En cambio, en el máximo profundo de fluorescencia,

correspondiente al máximo profundo de clorofila, los protistas consumían el 35.0 ± 15.7% de BP d-1

vs. el 17.8 ± 29.3% de BP d-1 de pérdidas debidas a la infección vírica. Ambos tipos de mortalidad

bacteriana (PMM y VMM) estaban negativamente correlacionados. En las aguas AW se

detectaron mayores VMM y menores PMM en comparación con aguas de mezcla de deshielo

(PSWm). Como consecuencia, más carbono bacteriano fue canalizado a niveles tróficos

superiores en PSWm que en AW. Así los virus canalizaron 2.63 ± 2.45 μgC l-1 d-1 en PSWm, y 4.27 ±

5.54 μgC l-1 d-1 en AW, mientras que los protistas ingirieron 13.05 ± 5.98 μgC l-1 d-1 en PSWm, y 8.91 ±


140 | P a g e

8.33 μgC l-1 d-1 en AW. Los resultados obtenidos sugieren que el proceso de deshielo debido al

cambio climático puede contribuir a un mayor aprovechamiento del carbono bacteriano hacía

niveles superiores vía protistas, con respecto a la materia orgánica disuelta y particulada

proveniente de la lisis celular aportada a la columna de agua del Océano Ártico.

Polar Biology, submitted

Effect of ice melting on bacterial carbon fluxes channeled by viruses

and protists in the Arctic Ocean

Julia A. Boras*1, M. Montserrat Sala1, Jesus M. Arrieta2, Elisabet L. Sà1, Jorge Felipe1, Carlos M. Duarte2, and Dolors Vaqué1 1 Institut de Ciències del Mar (CSIC), Passeig Marítim de la Barceloneta 37-49, 08003

Barcelona, Spain 2 Institut Mediterrani d'Estudis Avançats (CSIC – IUB), Miquel Marqués 21, 07190 Esporles

(Mallorca), Spain

Effect of sea ice melting on bacterial carbon fluxes channelled by phages and protists was evaluated in northern Greenland Sea and in Arctic Ocean waters, during July 2007. Two main water masses were studied: warmer and with higher salinity Atlantic Waters (AW), and colder and with lower salinity Polar Surface Waters from melted ice mixed with AW (PSWm). In AW waters significantly higher phosphate and nitrite+nitrate concentrations were detected, as well as higher bacterial abundances. In contrast, in PSWm waters, low nutrient concentrations, low bacterial abundance, but high viral abundances were observed. Losses of bacterial production (BP) caused by protists (PMMBP) were higher than those caused by phages (VMMBP) along the whole study, with average PMMBP of 45.4 ± 6.3% d-1 vs. VMMBP = 8.5 ± 6.9% d-1at the depth of 0.1 m, and 35.0 ± 15.7% of d-1 vs. 17.8 ± 29.3% of d-1 of PMMBP and VMMBP, respectively, in deep fluorescence maximum. PMMBP and VMMBP were significantly negatively correlated among them. In PSWm waters, significantly higher PMMBP and lower VMMBP compared with AW waters were detected. Consequently, significantly more bacterial carbon was channelled to the higher trophic levels in PSWm than in AW. Viruses channelled 2.63 ± 2.45 μgC l

-1 d-1 in PSWm and 4.27 ± 5.54 μgC l-1 d-1 in AW waters, and

protists ingested 13.05 ± 5.98 μgC l-1 d-1 in PSWm and 8.91 ± 8.33 μgC l

-1 d-1 in AW. We conclude that observed processes could moderate the carbon flow through microbial food webs in the scenario of the decrease of sea ice cover due to climate change in the Arctic region.

Keywords Viruses • Bacteria • Grazing • Viral infection • Arctic *Corresponding author: [email protected]

Boras et al.

2

Introduction

The Arctic Ocean ecosystem is thought to be particularly sensitive to global climate change (Cuffey et al. 1995; Fyfe et al. 1999). In fact, the consequences of changes in the Arctic climate are already clearly visible. Loss of ice cover occurs in parallel with warming of Arctic Ocean waters, and both processes have been especially pronounced since 2000 (Nghiem et al. 2007; Steele et al. 2008). The ice volume in the Arctic in winter experienced a net decrease of more than 42% since 2005 (Kwok et al. 2009). The large sea ice melting will inevitably lead to large changes in the biogeochemistry of Arctic waters (Chen et al. 2003). Large melt-water supply to the ocean reduces the salinity and temperature of the surface water (Heinrich 1988), and can be a source of additional pool of microorganisms (Thomas and Dieckmann 2002; Maranger et al. 1994), contaminants (Pfirman et al. 1995), and organic carbon (Krembs et al. 2002). Consequently, the cycling of organic matter in the Arctic region can be potentially affected by climate changes. However, no studies evaluating the impact of melting waters on microbial processes are available.

Primary production in the Arctic Ocean in spring and summer plays an important role in the global carbon cycling (Stein and Macdonald 2004). Several studies showed that an active microbial food web in this region influences the transfer of carbon from dissolved pool to metazoan food webs and the deep sea (Wheeler et al. 1996). Within the microbial community, bacteria constitute a dynamic and important element of the planktonic food webs in Arctic waters (Middelboe et al. 2002), and bacterial production can correspond to up to 35 - 150% of primary production (Wheeler et al. 1996). Some studies suggest that bacterial production in summer reflects a response of the bacterial community to the release of organic carbon (Yager et al. 2001; Middelboe et al. 2002),

and particulate organic matter (POM; Hodges et al 2005) during the phytoplankton blooms.

Bacterial carbon can enter into the trophic web transferred by nanoprotists, or be released to the water column due to viral infections and consequent cell lysis. Thus, bacterial predators’ pressure could shape the carbon fluxes through the system, and control the carbon load to the deep ocean waters (Fuhrman 1999). Despite the potential importance of these processes in Arctic Ocean, to our best knowledge very few studies have evaluated protistan and viral impact on bacterial assemblages in pelagic communities in that region. In available reports, protists were shown to be responsible for removing a variable percentage of bacterial production, from not detected in pre- and post-bloom situations (Anderson and Rivkin 2001; Vaqué et al. 2008), to >200% of bacterial production during the early summer bloom (Maranger et al. 1994; Laurion et al. 1995). Shunt of bacterial carbon and nutrients to the dissolved and particulate phases through viral infections and cell lysis found in Arctic waters was equal (Steward et al. 1996), or much higher (Wells and Deming 2006) than protistan predation, and showed a high variation as well. Phages were overall responsible for removing from <1% (Steward et al. 2007), to 100% of bacterial production (Wells and Deming 2006). Furthermore, no reports on the special type of viral infection, a lysogeny, are available for Arctic waters. Lysogeny would be an important factor in bacterial carbon fluxes, and could be especially frequent in Arctic Ocean due to low host abundances (Jiang and Paul 1998; Weinbauer et al. 2003). The main goal of this study was to evaluate the effect of the Arctic sea ice melting on bacterial carbon fluxes shunted by protists and phages to respectively, higher trophic levels and pools of particulate and dissolved organic matter (DOM) in the water column. For this purpose, we measured bacterial mortality due to phages and protists in 12 stations in the northern Greenland Sea

Effect of ice melting on bacterial mortality

3

Fig. 1 Map of the area studied during the ATOS-I cruise. Colour symbols indicate type of waters detected in the deep fluorescence maximum (DFM) layer: black symbols - Atlantic Water (AW), grey symbols - Polar Surface Water (PSW) from melting mixed with AW (PSWm), open symbols: PSW from the recent melting. At all stations, the PSWm was detected at the surface (0.1 m). and the Arctic Ocean during the ATOS-I cruise in summer 2007. In that year the lowest ice cover for the last 40 years, and an unusually warm ocean surface were recorded (Steele et al. 2008; Zhang et al. 2008), providing a high input of melting waters. Owing to the dilution of the seawater by melted ice, we expected to find lower nutrient concentrations and lower bacterial production in areas under the influence of melting waters. Consequently, lower bacterial losses due to

phages and protists would be found due to lower host abundance and fitness. For the same reason, we inferred that protists would be overall more important than phages in removing bacterial production, and lysogeny should account for an important percentage of viral infections. Materials and methods

Study area and sampling

The study was carried out at stations located in the northern Greenland Sea and in the Arctic Ocean (78 – 80.5ºN, 2.3ºW – 16.5ºE, Fig. 1), on board of the R/V BIO-Hespérides during the cruise ATOS-I, from 27 June to 28 July 2007. Twelve stations were sampled to perform bacterial mortality measurements (Fig. 1). At each station, temperature, salinity and fluorescence were recorded down to 100 m depth using a CTD, mounted on a General Oceanics rosette sampler, equipped with 12 litter Niskin bottles. The water masses were classified into relatively warm and salty (>34) Atlantic Water (AW), Polar Surface Water from the recent melting (PSW) with temperature <2 ºC and low salinity (<34), and PSW mixed with AW (PSWm) with temperature >2 ºC and low salinity (<34). Samples for microbial abundances and processes were taken at 0.1 m (from a rubber boat), 1 m (with manually managed Niskin bottles), 5 m, and down at intervals of 15 – 20 m, including the deep fluorescence maximum (DFM) layer. Samples for the determination of the dissolved inorganic phosphate and nitrate + nitrite concentrations were kept frozen until analyzed in a Bran+Luebbe AA3 autoanalyser, following standard spectrophotometric methods (Hansen and Koroleff 1999).

Boras et al.

4

Table 1. Abundances of microorganisms along the depth profile at stations studied during the ATOS-I cruise. Values presented for three type of waters: AW – Atlantic Waters, PSW – Polar Surface Waters from the recent melting, PSWm – PSW mixed with PSWm. nd - not determined, VA – viral abundance (106 viruses ml-1), BA – bacterial abundance (106 cells ml-1), VBR – virus-bacterium ratio, HNF - heterotrophic nanoflagellates (103 cells ml-1), PNF – phototrophic nanoflagellates (103 cells ml-1). * indicates the deep fluorescence maximum (DFM) layer.

Station Depth Water type VA BA VBR PNF HNF

St.6

0.1 1 5 21 37* 61 81

PSWm PSWm PSWm AW AW AW AW

2.10 1.53 1.70 2.06 1.83 0.76 0.49

1.18 1.13 1.13 1.07 1.44 0.53 0.39

1.8 1.3 1.5 1.9 1.3 1.4 1.3

10.70 7.28

13.00 22.40 16.80

2.66 0.45

2.11 4.32 3.56 7.54 2.03 0.94 0.20

St.9

0.1 1 5 16* 30 51 72


1.87 2.99 2.55 2.56 2.33 1.59 1.51

1.66 1.81 2.39 2.67 2.31 1.77 1.11

1.1 1.6 1.1 1.0 1.0 0.9 1.4

14.60 17.10 10.50

8.32 19.00

7.82 3.85

1.94 4.96

14.00 2.74 3.06 3.54 1.70

St.12

0.1 1 5 10* 31 50 70

PSWm PSWm PSWm PSWm PSWm PSWm AW

3.70 2.33 2.10 1.84 1.62 1.11 1.03

0.95 1.65 1.91 2.10 1.93 1.72 1.19

3.9 1.4 1.1 0.9 0.8 0.6 0.9

5.66 6.33 0.12 4.99

11.70 2.37 2.22

6.09 4.14 1.95 1.80 1.95 0.57 1.02

St.15

0.1 1 5 20* 40 71 101

PSWm PSWm PSWm PSWm PSWm PSWm AW

1.74 2.38 1.52 1.55 1.44 1.20 1.34

nd 1.80 1.70 1.99 1.65 1.11 0.86

nd 1.3 0.9 0.8 0.9 1.1 1.5

16.70 22.00 27.40 29.40 20.00 11.00 11.80

4.76 1.19 0.82 1.08 0.40 0.53 0.80

St.18

1 6 27* 41 60 84

PSWm PSWm PSWm PSWm AW AW

1.26 1.34 1.13 1.06 0.64 0.50

0.87 1.06 1.23 1.06 0.92 0.78

1.4 1.3 0.9 1.0 0.7 0.6

9.41 13.50 11.60

7.15 4.65 3.94

0.32 0.23 0.43 0.51 0.36 0.11

St.19

3 12* 26

PSWm PSWm PSWm

3.34 3.17 3.24

0.84 0.81 0.81

4.0 3.9 4.0

9.00 14.10 11.00

1.00 0.95 0.24

Continues on the next page.


5

Table 1. Continuation.

Station Depth Water type VA BA VBR PNF HNF

St.20

0.1 1 5 10 26* 51 80

PSWm

PSWm PSWm PSWm PSWm AW AW

4.15 3.98 4.15 4.42 3.75 1.80 1.70

0.91 0.90 1.22 1.27 2.27 1.14 0.82

4.6 4.4 3.4 3.5 1.6 1.6 2.1

12.40 12.30

nd nd

15.10 7.22 5.33

0.60 0.51

nd nd

0.65 0.45 0.20

St.23

0.1 1 5 10 24* 37 51

PSWm PSWm PSWm PSWm AW AW AW

6.79 5.48 7.00 7.28 6.88 6.81 1.61

1.01 1.08 1.30 1.67 2.02 2.47 1.32

6.7 5.1 5.4 4.4 3.4 2.8 1.2

15.90 21.50 16.50 26.70 47.80 25.40

7.78

0.87 0.64 0.31 0.83 1.32 0.89 0.57

St.26

0.1 1 5 16 27* 41 62


4.93 3.46 3.98 5.82 6.87 2.09 1.40

0.77 0.55 1.18 1.93 2.65 1.94 0.99

6.4 6.3 3.4 3.0 2.6 1.1 1.4

19.20 29.70 21.40 40.40 45.00

8.13 11.60

0.29 0.86 0.10 0.76 0.54 0.68 0.79

St.33

0.1 1 5 10 23* 50 81


2.69 1.24 1.95 2.22 1.90 0.96 0.32

0.42 0.37 0.59 1.24 1.71 0.89 0.32

6.5 3.3 3.3 1.8 1.1 1.1 1.0

10.60 9.24 6.68

11.60 12.10

6.46 2.91

0.17 0.58 0.54 1.74 1.20 0.61 0.61

St.39

0.1 1 6 33 39* 51 71


2.24 1.43 5.29 0.84 1.13 0.81 0.49

0.82 0.46 1.16 2.87 2.13 1.59 1.16

2.7 3.1 4.5 0.3 0.5 0.5 0.4

7.14 4.81 9.69

12.60 7.59 4.45 4.44

0.19 0.12 0.38 0.59 0.30 0.21 0.13

St.43

0.1 1 5 10 20* 41 61

PSWm PSWm PSWm PSWm PSW PSW AW

2.49 1.66 1.57 1.44 0.76 0.37 0.53

0.38 0.53 0.51 0.50 0.53 0.22 0.41

6.6 3.1 3.1 2.9 1.4 1.7 1.3

9.45 12.00 12.20 13.40 15.50

4.36 1.83

0.29 1.25 0.55 1.18 1.61 0.46 0.10

Boras et al.

6

Fig. 2 Temperature (Temp.) and salinity at the surface (0.1 m) and in the deep fluorescence maximum (DFM) layer detected in Greenland Sea and Arctic Ocean during the ATOS-I cruise. Colours indicate the type of waters: AW – Atlantic Water, PSWm - Polar Surface Water from melting mixed with AW, PSW – Polar Surface Water from the recent melting. No samples for the surface waters for St.18 and St.19 were analyzed

Microorganisms’ abundances

In situ viral and bacterial abundances were determined by flow cytometry. Subsamples (2 ml), taken at 7 selected depths, were fixed with glutaraldehyde for viruses (0.5% final concentration), or paraformaldehyde for bacteria (1% final concentration). Samples of viruses were fixed at 4ºC during 15 - 30 min, then quick frozen in liquid nitrogen and stored at -80ºC as described in Brussaard (2004). Samples of bacteria were analyzed on board immediately after fixation. Counts were made on a FACSCalibur (Becton & Dickinson) flow cytometer. Samples of viruses were stained with SYBR Green I, and run at a medium flow speed (Brussaard 2004), and samples of bacteria were stained with a DMS (dimethyl sulfate)-diluted SYTO13, and run at a low speed using 50 l of 0.92 m yellow-green latex beads as an internal standard (del Giorgio et al. 1996).

In situ nanoflagellate abundance was determined by epifluorescence microscopy (Olympus BX40-102/E at 1000 ×). Subsamples (100 ml) were taken at 7 selected depths, fixed with glutaraldehyde (1% final concentration), filtered through 0.6 m black polycarbonate filters, and stained with DAPI (4,6-diamidino-2-phenylindole) at a final concentration of 5 μg ml

-1(Sieracki et al. 1985) . Heterotrophic and phototrophic nanoflagellates (HNF and PNF) were distinguished under ultraviolet radiation (UV) and blue light (B2 filter). At least 20 - 100 of HNF and PNF were counted per sample.

Bacterial and viral production, and bacterial

mortality

Bacterial (BP) and viral (VP)

production, and bacterial losses due to protists (PMM) and viruses (VMM) were obtained using water samples from 2 selected depths: 0.1 m, and DFM. Due to the weather


7

Fig. 3 Viral, bacterial and flagellates’ abundances at the surface (0.1 m) and in deep fluorescence

maximum (DFM) detected during the ATOS-I cruise. Colours indicate the type of waters: AW – Atlantic Water, PSWm - Polar Surface Water from melting mixed with AW, PSW – Polar Surface Water from the recent melting. HNF – heterotrophic nanoflagellates, PNF – phototrophic nanoflagellates. No samples for the surface waters for St.18 and St.19 were analyzed.

Boras et al.

8

Table 2. Significant correlation coefficients among variables. Temp.: temperature, VA: viral abundance, BA: bacterial abundance, VBR: virus-bacterium ratio, HNF: heterotrophic nanoflagellate abundance, PNF: phototrophic nanoflagellate abundance, VPL: lytic viral production, %VPLG: % of lysogenic VP in total VP, BP: bacterial production, VMM: virus-mediated mortality of bacteria, PMM: protist-mediated mortality of bacteria, BSS – bacterial standing stock. conditions, no measurements were carried out for 0.1 m at stations 18, 19, and 26.

Bacterial mortality due to protists was evaluated following the fluorescent labeled bacteria (FLB) disappearance method (Sherr et al. 1987). The FLB were prepared with a culture of Brevundimonas diminuta (Spanish Type Culture Collection, http://www.cect.org/index2.html). For each grazing measurement, duplicates (1 litter of seawater each) and one control (1 litter of virus-free water) were prepared in 2-litter polycarbonate bottles. All bottles (control and duplicates) were inoculated with FLB at 20% of in situ bacterial concentration. Bottles were incubated in a thermostatic chamber, simulating in situ temperature and in the dark during 48 h. Samples for evaluation of bacterial and FLB's abundances were taken at the beginning and at the end of the incubations. Abundances of bacteria and FLB were assessed by epifluorescence microscopy (Olympus BX40-102/E; a 1000 × magnification). To this end, aliquots of 20 ml sample were filtered through 0.2 μm black polycarbonate filters, and stained with DAPI at a final concentration of 5 μg ml-1 (Sieracki et al. 1985). Natural bacteria were identified by their blue fluorescence when excited with UV radiation, while FLB were identified by their yellow-green fluorescence when excited with blue light. Control bottles showed no decrease of FLB during the experiments.

Grazing rates of bacteria were calculated following the equations of Salat and Marrasé (1994), based on the specific grazing rate (g) and the specific net growth rate (a).

Variables N R P

Depth - Temp. 77 0.440 <0.01 Depth - Salinity 79 0.742 <0.01 PO4

-3 - Depth 59 0.693 <0.01 PO4

-3 - Temp. 57 0.593 <0.01 PO4

-3 - Salinity 59 0.739 <0.01 NO2

-+NO3- - Depth 61 0.659 <0.01

NO2-+NO3

- - Temp. 59 0.571 <0.01 NO2

-+NO3- - Salinity 61 0.727 <0.01

Fluorescence - Depth 56 0.308 <0.05 Fluorescence - Temp. 54 0.400 <0.01 Fluorescence - Salinity 56 0.368 <0.01 VA - Depth 79 -0.442 <0.01 VA - Salinity 79 -0.514 <0.01 VA - PO4

-3 59 -0.441 <0.01 VA - NO2

-+NO3- 61 -0.332 <0.01

VA - BA 78 0.394 <0.01 BA - Temp. 76 0.485 <0.01 BA - Salinity 78 0.378 <0.01 BA - NO2

-+NO3- 61 0.659 <0.01

BA - Fluorescence 56 0.525 <0.01 VBR - Depth 78 -0.625 <0.01 VBR - Salinity 78 -0.635 <0.01 VBR - PO4

-3 59 -0.624 <0.01 VBR - NO2

-+NO3- 61 -0.614 <0.01

VBR - Fluorescence 56 -0.320 <0.02 VBR - VA 78 0.663 <0.01 HNF - Temp. 77 0.276 <0.02 HNF - BA 78 0.249 <0.05 PNF - Temp. 77 0.267 <0.02 PNF - Fluorescence 56 0.330 <0.02 HNF - PNF 79 0.712 <0.01 VPL - PNF 18 0.513 <0.05 VPL - %VPLG 18 -0.484 <0.05 VMMBSS - VPL 42 0.475 <0.01 VMMBP - VPL 42 0.321 <0.05 PMMBSS - VBR 20 0.489 <0.05 PMMBSS - BP 21 0.634 <0.01 PMMBSS - %VPLG 21 0.569 <0.01 PMMBP - Depth 21 -0.439 <0.05 PMMBP - VBR 20 0.472 <0.05 PMMBP - BP 21 0.485 <0.05 PMMBP - VMMBP 21 -0.817 <0.01


9

Net bacterial production (BPN) in the incubation bottles was obtained:

BPN = BA0 × (eat – 1) [cells ml-1 d-1],

and grazing rate (G) was calculated:

G = (g / a) × BPN [cells ml-1 d-1].

Protist-mediated mortality of bacteria (PMM) was calculated as the percentage of bacterial standing stock (BSS) and BP:

PMMBSS = (G × 100) / BA0 [% d-1],

PMMBP = (G × 100) / BP [% d-1].

To determine viral production and bacterial losses due to phages, we followed the virus reduction approach (Wilhelm et al. 2002). Briefly, 1 litter of seawater was prefiltered through 0.8 m pore size cellulose filter (Whatman), and then concentrated by a spiral-wound cartridge (0.22 m-pore-size, VIVAFlow 200), obtaining 40 ml of bacterial concentrate. Virus-free water was collected filtering one litter of seawater using a cartridge of 100 kDa molecular mass cutoff (VIVAFlow 200). A mixture of virus-free water (160 ml) and bacterial concentrate (40 ml) was prepared and distributed into 4 sterile falcon plastic tubes. Two of the tubes were maintained without any manipulations as controls, while in other two, mitomycin C (Sigma) was added (1 µg ml-1 final concentration) as inducing agent of the lytic cycle in prophages. All falcon tubes were incubated in a thermostatic chamber simulating in situ temperature and in the dark, during 12 h. Samples for viral and bacterial abundances were collected at time zero and every 4 h of incubation, fixed with glutaraldehyde and stored as described before for viruses. Viruses and bacteria from viral production incubations were counted by flow cytometry. The number of viruses released by bacterial cell (burst size, BS) was estimated from viral production incubations, as in Middelboe and Lyck (2002), Wells and Deming (2006), and Boras et al. (2009). Increase of viral abundance during short time

intervals (4 h) was divided by a decrease of bacterial abundance over the same period of time. We assumed that the bacterial production and viral decay in this time interval were negligible. The estimated BS ranged from 1 to 59 viruses per cell.

Virus-mediated mortality of bacteria (VMM) was calculated as previously described in Weinbauer et al. (2002) and Winter et al. (2004). Briefly, viral increase in the control tubes represents lytic viral production (VPL), and the difference between viral increase in the mitomycin C treatments and VPL gives the lysogenic production (VPLG). To compare the values of VPL and VPLG from different incubations, we multiplied them by the bacterial correction factor calculated from the in situ and T0 bacterial abundances (0.2 – 0.9; Winget et al. 2005). Then, we calculated the rate of lysed cells (RLC) dividing VPL by burst size (BS, Guixa-Boixereu 1997). RLC was used to calculate VMM as a percentage of bacterial standing stock (VMMBSS):

VMMBSS = (RLC × 100) / BA0 [% d-1],

where BA0 is the initial bacterial abundance in the viral production determination. Assuming that percentage of losses of BSS due to viruses is the same in falcon tubes and in the grazing bottles, we used VMMBSS to calculate the rate of lysed bacteria in the grazing measurements (RLCG, cells ml-1 d-1):

RLCG = (VMMBSS × BAG) / 100,

where BAG is bacterial abundance in the grazing bottles at time 0. Using RLCG, VMM as a percentage of bacterial production (VMMBP) could be calculated:

VMMBP = (RLCG × 100) / BP [% d-1],

where BP is bacterial production. Finally, bacterial production was

calculated as the sum of net bacterial production (BPN), grazing rate (G) and rate of lysed cells (RLCG):

Boras et al.

10

BP = BPN + G + RLCG [cells ml-1 d-1].

Statistical analyses

The Shapiro-Wilk W-test was used to check normal distribution of data, and data were logarithmically transformed prior to analyses if necessary. One-way ANOVA for normal distributions, Wilcoxon test for non-normal distributions, and Tukey-Kramer HSD test

were used to evaluate the differences between the three types of water masses, or between depths. Pearson correlation analysis was performed to determine the relationships between the various properties examined. These statistical analyses were performed using the JMP program.

Fig. 4 Bacterial production, and lytic and lysogenic viral production, detected at the surface (0.1 m) and in deep fluorescence maximum (DFM) in Greenland Sea and Arctic Ocean. Each value corresponds to experimental duplicates, bars indicate the maximum and minimum values of duplicates. The percentages of lysogeny in total viral production are shown. Colours indicate the type of waters: AW – Atlantic Water, PSWm - Polar Surface Water from melting mixed with AW, PSW – Polar Surface Water from the recent melting. Due to the weather conditions no experiments were performed for the surface water at St.19 and St.26 to determine PMMBP, and at St.18, St.19, and St.26 to determine VMMBP


11

Results

Oceanographic conditions

Polar Surface Water mixed with

Atlantic Water (PSWm) was found at the surface (0.1 – 5 m) of all stations, and Atlantic Water (AW) was detected in deeper waters (Table 1). Polar Surface Water (PSW) was detected once, at St.43, at the depth of 20 - 40 m (Fig. 1). Water temperature at the surface was 2.6 ± 1.7ºC, and overall increased with depth (Table 2). At the DFM layer (from 20 to 40 m), water temperature was 5.3 ± 0.7ºC in AW, 4.2 ± 1.5ºC in PSWm, and -1.4ºC in PSW (Fig. 2). Water salinity at the surface was 33.4 ± 0.9, and increased with depth (Table 2). At DFM, salinity was 35.0 ± 0.1 in AW, 34.6 ± 0.3 in PSWm, and 33.9 in PSW (Fig. 2). At stations where PSWm was detected from the surface to DFM (PSWm stations), salinity was significantly lower at 0.1 m than in DFM (F1, 13 = 9.86, P < 0.01). Both parameters were significantly higher in AW than in PSWm (F1, 85 = 14.08, P < 0.01 for temperature, F1, 85 = 14.08, P < 0.01 for salinity). The average phosphate and nitrites+nitrates concentrations in the upper 50 m layer were 0.22 ± 0.14 μM, and 2.27 ± 2.61

μM respectively. Both PO4-3 and NO2

-+NO3-

were significantly higher in AW than in PSWm water masses (Tukey-Kramer, R = 0.24, P = 0.05, and R = 0.51, P = 0.05, respectively), but no differences among AW and PSW, and PSW and PSWm waters were detected. Concentrations of both nutrients increased with depth, temperature and salinity (Table 2).

Distribution and abundance of

microorganisms

Average fluorescence in the upper 50 m at twelve stations was 9.98 ± 11.93 (relative units), and was significantly higher in DFM than at 0.1 m (F1, 13 = 50.65, P < 0.01), reaching there values from 2.44 at St.33 to 41.22 at St. 43 (data not shown). The DFM

changed its depth from 9.5 m at St.12, to 39.7 m at St.39. No significant differences in fluorescence between the three types of waters were detected, and mean values for AW and PSWm were very similar. Fluorescence increased with an increase of depth, temperature and salinity. Viral abundances (VA) ranged between 0.32 × 106 and 7.28 × 106 viruses ml-1 (Table 1), and bacterial abundances (BA) varied from 0.22 × 106 to 2.87 × 106 cells ml-1 (Table 1). Viral abundance showed significant negative correlation with depth, salinity, PO4

-3, and NO2

-+NO3-, while BA was positively

correlated with temperature, salinity, NO2-

+NO3-, and fluorescence (Table 2). VA and

BA showed positive correlation with each other (Table 2). Both parameters were significantly different among the three types of waters (Table 3), with the highest values of VA in the PSWm, and the highest BA in AW. Analysis of differences among pairs of water masses showed, that VA was significantly higher in PSWm than AW, and in PSWm than in PSW waters (Table 3), and BA was significantly higher in AW than in PSW, and in PSWm than in PSW water masses (Table 3). At PSWm stations BA was significantly higher in DFM than at 0.1 m (F1, 12 = 6.35, P < 0.05, Fig. 3). These differences were not detected for VA. Virus to bacterium ratio (VBR) varied between 0.6 and 6.7 (Table 1), decreasing significantly with depth, salinity, PO4

-3, and NO2-+NO3

- and fluorescence, and increasing with VA (Table 2). The lowest VBR was overall found between 40 and 100 m. The VBR was significantly different among water types (F2, 85 = 6.53, P < 0.01), with the highest values in the PSWm, and the lowest in the AW. At PSWm stations, significantly higher VBR was found at 0.1 m than in DFM (F1, 12 = 6.81, P < 0.05). The HNF reached the lowest abundances of 94 cells ml-1 at 60 m, and the highest of 1.4 × 104 cells ml-1 at 5 m (Table 1). The PNF abundances varied from 4.48 × 102 cells ml-1 at 80 m, to 4.78 × 104 cells ml-1 in DFM (Table 1, Fig. 3). The HNF abundance was

Boras et al.

12

Fig. 5 Protist- and virus-mediated mortality of bacteria, as a percentage of bacterial production (PMMBP and VMMBP) detected at the surface (0.1 m) and in deep fluorescence maximum (DFM) during the ATOS-I cruise. Bars indicate the maximum and minimum values of duplicates. Colours indicate the type of waters: AW – Atlantic Water, PSWm - Polar Surface Water from melting mixed with AW, PSW – Polar Surface Water from the recent melting. Due to the weather conditions no experiments were performed for the surface water at St.19 and St.26 to determine PMMBP, and at St.18, St.19, and St.26 to determine VMMBP.

positively correlated with BA and PNF abundance, and both HNF and PNF abundances decreased with temperature (Table 2). No significant differences in PNF and HNF abundance among the three types of waters were found. Bacterial and viral production

Bacterial production (BP) ranged

between 0.22 × 106 and 1.34 × 106 cells ml-1 d-

1 at 0.1 m, and between 0.15 × 106 and 2.08 × 106 cells ml-1 d-1 in DFM (Fig. 4). In DFM, BP was significantly different among the three types of waters (Table 3), with the highest values in the PSWm, and the lowest in the PSW. Analysis of differences among pairs of water masses showed, that BP was significantly higher in PSWm than PSW, and in AW than in PSW waters (Table 3). At PSWm stations, BP was significantly higher


13

in DFM than at the surface (F1, 12 = 4.86, P < 0.05; Fig. 4).

Lytic viral production (VPL) varied between 0.08 × 106 and 1.29 × 106 viruses ml-

1 d-1 at 0.1 m, and between 0.09 × 106 and 4.15 × 106 viruses ml-1 d-1 in DFM (Fig. 4). The VPL was not detected once at 0.1 m, and twice in the DFM. In 7 out of 21 viral production measurements, prophage induction was detected, and the percentage of lysogenic viral production (%VPLG) with respect to the total VP varied from 33.9% in DFM to 100.0% at 0.1 m (Fig. 4). The VPL was positively correlated with PNF abundance, and negatively with %VPLG (Table 2). Bacterial mortality

Protist-mediated mortality of bacteria (PMM) was not detectable once, at St.39, in DFM. Losses of BSS (PMMBSS) were from 9.08 to 62.76% d-1 at 0.1 m, and from 2.92 to 65.91% d-1 in DFM (data not shown). The PMMBSS was positively correlated with BP, VBR, and %VPLG (Table 2). Losses of BP due to protists (PMMBP) ranged from 37.13 to 55.08% d-1 at 0.1 m, and from 14.42 to 52.09% d-1 in DFM (Fig. 5). The PMMBP was negatively correlated with depth, and positively with VBR and BP (Table 2). Both PMMBSS and PMMBP were significantly different between the three types of stations (F2, 41 = 4.61, P < 0.02, and F2, 41 = 4.22, P < 0.05, respectively), with the highest values in PSWm, and the lowest in PSW. Analysis of differences among pairs of water masses showed, that PMMBP was significantly higher in AW than PSWm waters (Table 3).

Virus-mediated mortality of bacteria (VMM) was not detected once at 0.1 m (St.33), and twice in DFM (St.26 and 43). Losses of bacterial standing stock (VMMBSS) varied from 1.70 to 29.57% d-1 at 0.1 m, and from 1.07 to 25.94% d-1 in DFM (data not shown). Losses of bacterial production due to viruses (VMMBP) varied between 2.19 and 23.56% d-1 at 0.1 m, and between 2.12 and 50.61% d-1 in DFM (Fig. 5). At St.39, VMMBP

reached 100%, and grazing was not detectable. The VMMBSS and VMMBP were positively correlated with VPL (Table 2). The VMMBP was significantly different among the three water types (Table 3), with the highest values in the AW, and the lowest in the PSW. Analysis of differences among pairs of water masses showed, that VMMBP was significantly higher in AW than PSWm waters (Table 3). VMMBP and PMMBP were negatively correlated among them (Table 2). Discussion

Methodological considerations

Methods used in this study to evaluate bacterial mortality are widely recognized as one of the best suited methods for this purpose, yet they have their suite of assumptions and uncertainties. In most cases they gave high reproducibility of results among duplicates, however, in some experiments they resulted in relatively high range of values, as observed e.g., for PMMBP (St.43, DFM), or VPL (St.6, DFM).

The FLB disappearance technique (Sherr et al. 1987) was used to measure the grazing rates on bacteria by the whole community of protists. The bulk disappearance of FLB as tracers of natural bacteria used in the present study is, in our view, the best option in studies where the dynamics of the whole bacterial assemblage is targeted. In addition, it introduces the lowest possible water manipulation, which could be a potential source of artifacts. This technique has been widely discussed in Vaqué et al. (1994).

Bacterial production was estimated as the sum of the net increase of bacterial abundance, and losses of bacteria due to protists and viruses. This method was employed in previous studies, and results of BP were comparable, although slightly higher, with those obtained with in situ measurements of BP by 3H-leucine incorporation method (Boras et al. 2009).

Boras et al.

14

Table 3. Results of ANOVA test showing significant differences for principal biological variables among the three types of waters studied. Differences between paired types of waters were analysed by comparison of means using the Tukey-Kramer (T.-K.) HSD test. ns – not significant (p > 0.05), AW – Atlantic Waters, PSW – Polar Surface Waters from recent melting, PSWm – PSW mixed with AW, VA and BA – viral and bacterial abundances, BP – bacterial production, VMMBP – virus-mediated mortality of bacteria, PMMBP – protist-mediated mortality of bacteria.

Water type Test

Variable

VA BA BP VMMBP PMMBP

N F, R P N F, R P N F, R P N F, R P N F, R P

All

AW - PSW

AW - PSWm

PSW - PSWm

ANOVA

T.-K.

T.-K.

T.-K.

78

30

76

51

11.54

ns

0.10

0.18

<0.01

ns

0.05

0.05

77

30

76

51

5.05

0.13

ns

0.09

<0.01

0.05

ns

0.05

43

14

42

32

13.44

0.35

ns

0.52

<0.01

0.05

ns

0.05

41

14

40

30

4.22

ns

0.09

ns

<0.05

ns

0.05

ns

43

14

42

32

4.21

ns

2.17

ns

<0.05

ns

0.05

ns


15

Estimations of bacterial mortality due to viruses are subjected to the calculation of the burst size. We estimated the BS taking into account the increase of viral abundance and the decrease of bacterial abundance in the same period of time (4 h) in viral production measurements. We assumed that the only cause of bacterial abundance decrease over short time periods was cell lysis, and we did not take into account the viral decay and bacterial production during this time interval. Burst size values reported in this study, between 1 and 59, are within the range obtained for this system (Steward et al. 2007). Also, bacterial losses caused by lysis measured in our study are the potential losses, as we did not consider grazing on infected cells by protists in our calculations (González and Suttle 1993).

The virus reduction approach used to assess VP and VMM is based on the assumption that all viral production observed during the incubation time is a result of infections prior to incubation, no new infections occur, and that both filtration and incubation, do not induce lysogenic bacteria. This method allows direct observation of changes in viral abundance over time, and the distinction between production of virulent and temperate phages (Winget et al. 2005). Detection of lysogeny is based on lysis induction by mitomycin C. Although this agent is not sufficient to induce the lytic cycle in all prophages in some conditions (Weinbauer and Suttle 1999), this method is widely used and mitomycin C is better suited than other inducing agents, hence the obtained results are comparable to other studies. Considering that we could perform only one determination of viral production and grazing in PSW waters, in the further discussion we will rather focus on differences among AW and PSWm water masses. Sea ice conditions in summer 2007

The Arctic sea ice reduction has been recognized as an indicator of the Arctic

climate change (Arctic Climate Impact Assessment 2004). In the last decades a gradual reduction of extension and thickness of sea ice in Arctic were recorded (Nghiem et al. 2007), and in the last 4 years, the total ice volume in the winter has suffered a net loss of 6300 km3, which is more than 42% of the total Arctic ice coverage (Kwok et al. 2009). In 2007, the year in which this study was run, unprecedented low ice coverage and high water temperatures were recorded in the Arctic. Monthly ice extent for September 2007 was 23% smaller than in September 2005 (Stroeve et al. 2008), and was the lowest since 1970 (Zhang et al. 2008). Between 2007 and 2008, the average winter sea ice volume experienced losses of 3000 km3 (Kwok et al. 2009). Reduction of the sea ice extension was parallel to the warming of Arctic Ocean since 1965, particularly marked since 2000 (Steele et al. 2008). In summer 2007 sea surface temperature anomalies in this region reached 5ºC. Thus, our study was run in the scenario of extreme ice melting in the Arctic Ocean, which brought us an opportunity of study of microbial processes in conditions which will likely repeat in the future. Microorganisms’ abundances

Viral abundances reported in this

study are within the range found by other authors in Arctic Ocean (Middelboe et al. 2002), although higher values were also reported. Payet and Suttle (2008) found abundances reaching 2.0 × 107 viruses ml-1 in summertime in southern Beaufort Sea and Amundsen Gulf, which are one order of magnitude higher than detected in this study. Overall, our results agree with suggestions that viral concentrations at high latitudes are about 10 times lower than in temperate surface waters (Steward et al. 2007). Bacterial abundances in our study were similar or slightly higher than values found by other authors (Middelboe et al. 2002; Vaqué et al. 2008), and were overall of the same order of magnitude than viral abundances. This led to

Boras et al.

16

the low virus-bacterium ratio (from 0.4 to 6.7), which was among the lowest reported for the Arctic Ocean. Yager et al. (2001) found the VBR from 1 to 20, yet overall VBR from 10 to 70 were reported for Arctic surface waters in summer (Steward et al. 1996; Steward et al. 2007). The HNF and PNF abundances were among the values reported previously for Arctic waters (Sherr et al. 1997), reaching in occasions high concentrations of 104 cells ml-1 in the surface, which were one order of magnitude higher than abundances found by Sherr et al. (1997, 2003) and Vaqué et al. (2008). Bacterial and viral activities

Bacterial production detected at the

visited stations was in the range of values found in Arctic Ocean by other authors (e.g., Müller-Niklas and Herndl 1996; Middelboe et al. 2002) during summer. Some previous works report also much lower BP, of the order of magnitude of 103 cells ml-1 d-1 (Sherr and Sherr 2003), or much higher, of the order of magnitude of 107 cells ml-1 d-1 (Garneau et al. 2008). These differences among studies confirm that significant inter-annual variability of microbial processes exist in the Arctic Ocean, as was previously suggested by Sherr and Sherr (2003). In our study, bacterial abundance and BP were higher in AW than in PSWm waters, although the difference was not significant for BP. This would suggest that sea ice melting in different years could be one of factors influencing this inter-annual variability observed in different studies.

Lytic viral production was within the range of VPL found earlier in the Arctic Ocean in summer. Middelboe et al. (2002) found VP from 0.11 × 106 to 13.2 × 106 viruses ml-1 d-1 in the Buffin Bay, and Steward et al. (1996) reported VP from 0.39 × 106 to 14 × 106 viruses ml-1 d-1 for Bering and Chukchi Seas, which is the highest value recorded in the Arctic. However, it has to be mentioned that in our study, the VPL values of 106 viruses ml-

1 d-1were detected only on three occasions,

mainly in AW, and values of 104 to 105 viruses ml-1 d-1 were detected at the remaining stations (Fig. 4). These values are much lower than those observed in temperate waters (e.g., Rowe et al. 2008; Boras et al. in press).

The percentage of lysogenic production in total VP ranged from 34 to 72%, and was similar to those detected in subtropical Northeast Atlantic (Boras et al. in press) and in coastal Mediterranean (Boras et al. 2009). In one case, %VPLG accounted for 100% of VP, and at this station no mortality due to phages was detected. Negative correlation between VPL and %VPLG could suggest that the infection strategy decision is driven by an external factor, like low host abundances (Jiang and Paul 1998; Weinbauer et al. 2003). Bacterial mortality Bacterial mortality caused by viral infections in summer was found to be overall low in surface waters of Arctic Ocean. In previous studies phages accounted for losses of 1 to 11% of BP in central Arctic Ocean (Steward et al. 2007), 2 - 10% of BP in low productive Chukchi Sea (Steward et al. 1996), 10 – 23% of BP in productive waters of Bering Sea (Steward et al. 1996), and 6 – 28% of BP in Buffin Bay (Middelboe et al. 2002). Data of bacterial mortality caused by protistan grazing presents high variability, depending on the phytoplankton bloom occurrence (Anderson and Rivkin 2001; Vaqué et al. 2008). Values from not detected in the pre-bloom (Anderson and Rivkin 2001), through 1 - 30% of BP losses in the summer (Steward et al. 1996), to 187% of BP in winter non-bloom (Anderson and Rivkin 2001) were observed. In the only study where viral and protistan predation were evaluated simultaneously in surface waters, the impact of those two predators was roughly equal, and they both accounted for ca. 25% of the total prokaryotic mortality in the Chukchi Sea (Steward et al. 1996). In our study, bacterial mortality caused by protists was overall higher than mortality caused by


17

phages, at both the 0.1 m layer and the DFM. In surface waters grazers removed 45.4 ± 6.3% of BP d-1, and viruses 7.6 ± 7.1% of BP d-1. In the DFM layer this proportion was less pronounced, with 35.0 ± 15.7% of BP d-1 and 17.8 ± 29.3% of BP d-1 removed by protists and viruses, respectively. Together, viruses and protists were responsible for the loss of 53.0 ± 5.8% of BP d-1 at 0.1 m, and 48.5 ± 11.6% of BP d-1 in the DFM. These values are higher than those reported by Steward et al. (1996), yet still do explain only half of the potential bacterial mortality in Arctic waters. This suggests that the obtained mortality values are underestimated, or some important sources of bacterial losses, like autolysis or predation by other organisms, are missed (Steward et al. 2007). In one case in our study, viruses were responsible for removing of 100% of BP d-1, and simultaneously assessed protistan grazing was not detected. Similar results were reported by Wells and Deming (2006) in bottom Arctic waters in winter. This suggests an antagonistic relation among those two predators in natural conditions (Manage et al. 2002; Maranger et al. 2002). Comparison of VMMBP and PMMBP in AW and PSWm showed that the inflow of melting water had an effect on the fluxes of bacterial carbon. Significantly more grazing was detected in PSWm than in AW. Consequently, significantly more carbon was transferred up into the higher trophic levels, to micro- and mesoplankton in PSWm than AW waters (Tukey-Kramer, R = 0.06, P = 0.05). Assuming the bacterial cell carbon content of 29 fgC cell-1 (Middelboe and Lundsgaard 2003) we calculated that protists ingested 13.05 ± 5.98 μgC l

-1 d-1 and 8.91 ± 8.33 μgC l-

1 d-1 in PSWm and AW, respectively. In contrast, viral lysis was higher in AW than in PSWm, and thus in AW waters more, although not significantly, bacterial carbon was short-cut to the DOM and POM pools, favouring the losses of carbon by bacterial respiration. We calculated that viruses channelled 2.63 ± 2.45 μgC l

-1 d-1 in PSWm and 4.27 ± 5.54 μgC l

-1 d-1 in AW waters.

This situation could be a result of higher nutrient concentrations in AW than in PSWm, which would enhance primary production in AW. The fact, that with significantly higher nutrients concentrations we have found equal mean fluorescence in those two types of water could suggest that we have run our experiments in the post-bloom situation in AW waters. We can assume that in this scenario higher concentration of DOM and POM were present in AW than in PSWm waters, which could lead to higher, although not significantly, bacterial production in AW than PSWm. This hypothesis is confirmed by significantly higher bacterial abundances in AW than PSWm. Higher bacterial host abundance favoured viral infections (Murray and Jackson 1992), which was reflected by higher VMM in AW than in PSWm. In contrast, in PSWm waters, where nutrients were diluted by melting water, less DOM and POM from the phytoplankton would reach the water. Consequently, lower bacterial production and significantly lower bacterial abundances were detected there. This could be a cause of lower viral lysis in PSWm, as it is known that phages infect preferentially active hosts (Steward et al. 1996; Weinbauer et al. 2003). Lower VMM would favour protistan grazing. Furthermore, a positive correlation among percentage of lysogeny and PMM would confirm less favourable trophic conditions in PSWm, as it is known that lysogenic rather than lytic type of infection occurs in lower nutrient concentrations (Stewart and Levin 1984). Finally, closer analysis of bacteria-virus relation showed, that in PSWm no significant correlation among those two parameters exist, suggesting that relatively high viral community abundance in PSWm was due to viruses other than bacteriophages. In contrast, in AW a significant positive correlation among viruses and bacteria was detected (N = 28, R = 0.708, P < 0.01), thus the majority of viruses there were bacteriophages, which would provide and additional explanation of higher VMM in AW waters. The observed processes could

Boras et al.

18

moderate the process of carbon flow through microbial food webs in the scenario of the decrease of sea ice cover due to climate change in the Arctic region (Kirchman et al. 2009). Thus, sea ice melting, changing the physicochemical conditions of surrounding waters, can finally modify the carbon fluxes in the Arctic ecosystem. Conclusions Lytic viral production found in our study was lower than values found in temperate waters. In contrast, the levels of lysogeny were comparable with levels found elsewhere, accounting for 100% of viral production in one occasion. Our results showed that the sea ice melting process decreased nutrient concentrations in the surface Arctic waters, which in consequence altered microbial activity. Higher viral abundances and protist-mediated mortality rates, as well as lower bacterial abundances were detected in melting than in unaffected waters. This implicates a higher transport of organic carbon and nutrients along the trophic web via protists, and lower shunt of bacterial carbon to the DOM and POM pools in regions under the influence of melted ice. We conclude that the dominance of protistan grazing in melting waters could moderate the negative effect of global climate changes on the flow of carbon to higher levels of the trophic webs in the surface Arctic Ocean. Acknowledgments We are very grateful to Marta Álvarez for very helpul classification of water masses, to Andreu Saura for his help with the samples analyasis, and to the staff of UTM (CSIC) and the crew of the R/V - Hespérides for their help during the cruise. This study was supported by the following projects: ATOS, PROCAVIR and MICROVIS (POL2006-00550/CTM, CTM2004-04404-CO2-00/MAR; CTM2007-62140), funded by the Spanish Ministerio de Ciencia e Innovación. J. A. B.’s work was

supported by a Ph.D. fellowship from the Spanish Ministerio de Ciencia e Innovación (FPU grant), and M. M. S. by I3P-CSIC postdoctoral contract funded by the Fondo Social Europeo. This is a contribution to the European Network of Excellence EurOceans.

References

Anderson MR, Rivkin RB (2001) Seasonal patterns in grazing mortality of bacterioplankton in polar oceans: a bipolar comparison. Aquat Microb Ecol 25: 195-206

Arctic Climate Impact Assessment (2004) Impacts of a warming Arctic. Cambridge University Press, pp 140

Boras JA, Sala MM, Baltar F, Arístegui J, Duarte CM, Vaqué D (in press) Effect of viruses and protists on bacteria in eddies of the Canary Current region (subtropical Northeast Atlantic). Limnol Oceanogr

Boras JA, Sala MM, Vázquez-Domínguez E, Weinbauer MG, Vaqué D (2009) Annual changes of bacterial mortality due to viruses and protists in an oligotrophic coastal environment (NW Mediterranean). Environ Microbiol 11: 1181-1193

Brussaard CPD (2004) Optimisation of procedures for counting viruses by flow cytometry. Appl Environ Microbiol 70: 1506-1513

Chen C-TA, Liu K-K, Macdonald R (2003) Continental Margin Exchanges. In: Fasham MJR (ed) Ocean Biogeochemistry – The role of the ocean carbon cycle in global change. Global Change – The IGBP Series, Springer, Germany, pp 53-98

Cuffey KM, Clow GD, Alley RB, Stuiver M, Waddington ED, Saltus EW (1995) Large Arctic temperature change at the Wisconsin-Holocene Glacial transition. Science 270: 455-458


19

del Giorgio PA, Bird DF, Prairie YT, Planas D (1996) Flow cytometric determination of bacterial abundance in lake plankton with the green nucleid acid stain SYTO 13. Limnol Oceanogr 41: 783-789

Fuhrman JA (1999) Marine viruses and their biogeochemical and ecological effects. Nature 399: 541-548

Fyfe JC, Boer GJ, Flato GM (1999) The Arctic and Antarctic oscillations and their projected changes under global warming. Geophys Res Lett 26: 1601-1604

Garneau M-E, Roy S, Lovejoy C, Gratton Y, Vincent WF (2008) Seasonal dynamics of bacterial biomass and production in a coastal arctic ecosystem: Franklin Bay, western Canadian Arctic. J Geophys Res, 113, C07S91, doi: 10.1029/2007JC004281

González JM, Suttle CA (1993) Grazing by marine nanoflagellates on viruses and virus-sized particles: ingestion and digestion. Mar Ecol Prog Ser 94: 1-10

Guixa-Boixereu N (1997) Abundance and dynamic of viruses in planktonic ecosystems. Ph.D. thesis. Barcelona University.

Heinrich H (1988) Origin and consequences of cycling ice rafting in the northeast Atlantic Ocean during the past 130,000 years. Quaternary Res 29: 142-152

Hansen K, Koroleff F (1999) Detemination of nutrients. In: Grasshoff K, Kremling K, Ehrhardt M (eds) Methods of seawater analysis. Wiley-VCH, Weiheim, pp 159-228

Hodges LR, Nasreen B, Hollibaugh JT, Yager PL (2005) Illustrating the importance of particulate organic matter to pelagic microbial abundance and community structure – an Arctic case study. Aquat Microb Ecol 40: 217-227

Hope HG (1983) Significance of exoenzymatic activities in the ecology of brackish water: Measurements by means od methylumbelliferyl-

substrates. Mar Ecol Prog Ser 11: 299-308

Jiang SC, Paul JH (1998) Significance of lysogeny in the marine environment: studies with isolates and a model of lysogenic phage production. Microb Ecol 35: 235-243

Kirchman DL, Morán XAG, Ducklow H (2009) Microbial groeth in the polar oceans – role of temperature and potential impact of climate change. Nat Rev Microbiol 7: 451-459

Krembs C, Eicken H, Junge K, Deming JW (2002) High concentrations of exopolymeric substances in Arctic winter sea ice: implications for the polar ocean carbon cycle and cryoprotection of diatoms. Deep-Sea Res I 49: 2163-2181

Kwok R, Cunningham GF, Wensnahan M, Rigor I, Zwally HJ, Yi D (2009) Thinning and volumen loss of the Arctic Ocean sea ice cover: 2003 – 2008. J Geophys Res 114: doi: 10.1029/2009JC005312

Laurion I, Demers S, Vezina AE (1995) The microbial food web associated with the ice algal assemblage: biomass and bacterivory of nanoflagellate protozoans in Resolute Passage (High Canadian Arctic). Mar Ecol Prog Ser 120: 77-87

Manage PM, Kawabata Z, Nakano S-I, Nishibe Y (2002) Effect of heterotrophic nanoflagellates on the loss of virus-like particles in pond water. Ecol Res 17: 473-479

Maranger R, Bird DF, Juniper SK (1994) Viral and bacterial dynamics in Arctic sea ice during the spring algal bloom near Resolute, N.W.T., Canada. Mar Ecol Prog Ser 111: 121-127

Maranger R, del Giorgio PA, Bird DF (2002) Accumulation of damaged bacteria and viruses in lake water exposed to solar radiation. Aquat Microb Ecol 28: 213-227

Boras et al.

20

Middelboe M, Lyck PG (2002) Regeneration of dissolved organic matter by viral lysis in marine microbial communities. Aquat Microb Ecol 27: 187-194

Middelboe M, Nielsen TG, Bjørnsen PK (2002) Viral and bacterial production in the North Water: in situ measurements, batch-culture experiments and characterization and distribution of a virus-host system. Deep-Sea Res II 49: 5063-5079

Middelboe M, Lundsgaard C (2003) Microbial activity in the Greenland Sea: role of DOC lability, mineral nutrients and temperature. Aquat Microb Ecol 32: 151-163

Murray AG, Jackson GA (1992) Viral dynamics: a model of the effects of size, shape, motion and abundance of single-celled planktonic organisms and other particles. Mar Ecol Prog Ser 89: 103-116

Müller-Niklas G, Herndl GJ (1996) Dynamics of bacterioplankton during a phytoplankton bloom in the high Arctic waters of the Franz-Joseph Land archipelago. Aquat Microb Ecol 11: 111-118

Nghiem SV, Rigor IG, Perovich DK, Clemente-Colón P, Weatherly JW, Neumann G (2007) Rapid reduction of Arctic perennial sea ice. Geophys Res Lett 34, L19504, doi: 10.1029/2007GL031138

Payet JP, Suttle CA (2008) Physical and biological correlates of virus dynamics in the southern Beaufort Sea and Amundsen Gulf. J Mar Sys 74: 933-945

Pfirman SL, Eicken H, Bauch D, Weeks WF (1995) The potential transport of pollutants by Arctic sea ice. Sci Total Environ 159: 129-146

Rowe JM, Saxton MA, Cittrell MT, DeBruyn JM, Berg GM, Kirchman DL, Hutchins DA, Wilhelm SW (2008) Constraints on viral production in the Sargasso Sea and North Atlantic. Aquat Microb Ecol 52: 233-244

Salat J, Marrasé C (1994) Exponential and linear estimations of grazing on bacteria: effects of changes in the proportion of marked cells. Mar Ecol Prog Ser 104: 205-209

Sherr BF, Sherr EB, Fallon RD (1987) Use of monodispersed, fluorescently labeled bacteria to estimate in situ protozoan bacterivory. Appl Environ Microbiol 53: 958-965

Sherr EB, Sherr BF, Fessenden L (1997) Heterotrophic protists in the Central Arctic Ocean. Deep-Sea Res II 44: 1665-1682

Sherr BF, Sherr EB (2003) Community respiration/production and bacterial activity in the upper water column of the central Arctic Ocean. Deep-Sea Res I 50: 529-542

Sherr EB, Sherr BF, Wheeler PA, Thompson K (2003) Temporal and spatial variation in stocks of autotrophic and heterotrophic microbes in the upper water column of the central Arctic Ocean. Deep-Sea Res 50: 557-571

Sieracki ME, Johnson PB, Sieburth JMcN (1985) Detection, enumeration and sizing of planktonic bacteria by image analyzed epifluorescence microscopy. Appl Environ Microbiol 49: 799-810

Steele M, Ermold W, Zhang J (2008) Arctic Ocean surface warming trends over the past 100 years. Geophys Res Lett 35, L02614, doi: 10.1029/2007GL031651

Stein R, Macdonald RW (2004) Organic carbon budget: Arctic Ocean vs. Global Ocean. In: Stein R, Macdonald RW (eds) The organic carbon cycle in the Arctic Ocean. Springer, Germany, pp 315-322

Steward GF, Smith DC, Azam F (1996) Abundance and production of bacteria and viruses in the Bering and Chukchi Seas. Mar Ecol Prog Ser 131: 287-300

Steward GF, Fandino LB, Hollibaugh JT, Whitledge TE, Azam F (2007) Microbial biomass and viral infections of heterotrophic prokaryotes in the sub-


21

surface layer of the central Arctic Ocean. Seep-Sea Res I 54: 1744-1757

Stewart FM, Levin BR (1984) The population biology of bacterial viruses: Why be temperate? Theor Popul Biol 26: 93-117

Stroeve J, Serreze M, Drobot S, Gearheard S, Holland M, Maslanik J, Meier W, Scambos T (2008) Arctic sea ice extent plummets in 2007. EOS 89: 13-20

Thomas DN, Dieckmann GS (2002) Antarctic sea ice – a habitat for extremophiles. Science 295: 641-644

Vaqué D, Gasol JM, Marrasé C (1994) Grazing rates on bacteria: the significance of methodology and ecological factors. Mar Ecol Prog Ser 109: 263-274

Vaqué D, Guadayol Ò, Peters F, Felipe J, Angel-Ripoll L, Terrado R, Lovejoy C, Pedrós-Alió C (2008) Seasonal changes in planktonic bacterivory rates under the ice-covered coastal Arctic Ocean. Limnol Oceanogr 53: 2427-2438

Weinbauer MG, Suttle CA (1999) Lysogeny and prophage induction in coastal and offshore bacterial communities. Aquat Microb Ecol 18: 217-225

Weinbauer MG, Winter C, Höfle MG (2002) Reconsidering transmission electron microscopy based estimates of viral infection of bacterioplankton using conversion factors derived from natural communities. Aquat Microb Ecol 27: 103-110

Weinbauer MG, Brettar I, Höfle MG (2003) Lysogeny and virus-induced mortality of bacterioplankton in surface, deep,

and anoxic marine waters. Limnol Oceanogr 48: 1457-1465

Wells LE, Deming JW (2006) Significance of bacterivory and viral lysis in bottom waters of Franklin Bay, Canadian Arctic, during winter. Aquat Microb Ecol 43: 209-221

Wheeler PA, Gosselin M, Sherr E, Thibault D, Kirchman DL, Benner R, Whitledge TE (1996) Active cycling of organic carbon in the central Arctic Ocean. Nature 380: 697-699.

Wilhelm SW, Brigden SM, Suttle CA (2002) A dilution technique for the direct measurement of viral production: a comparison in stratified and tidally mixed coastal waters. Microb Ecol 43: 168-173

Winget DM, Williamson KE, Helton RR, Wommack KE (2005) Tangential flow diafiltration: an improved technique for estimation of virioplankton production. Aquat Microb Ecol 41: 221-232

Winter C, Herndl GJ, Weinbauer MG (2004) Diel cycles in viral infection of bacterioplankton in the North Sea. Aquat Microb Ecol 35: 207-216

Yager PL, Connelly TL, Mortazavi B, Wommack KE, Bano N, Bauer JE, Opsahl S, Hollibaugh JT (2001) Dynamic bacterial and viral response to an algal bloom at subzero temperatures. Limnol Oceanogr 46: 790-801

Zhang J, Steele M, Lindsay R, Schweiger A, Morison J (2008) Ensemble 1-Year predictions of Arctic sea ice for the spring and summer 2008. Geophys Res Lett 35, L08502, doi: 10.1029/2008GL033244

Discussion

Discussion

P a g e | 165

5. DISCUSSION

his thesis is based on the investigation performed in three different marine systems, i.e.

coastal zone of the NW Mediterranean Sea, the subtropical Atlantic Ocean, and the

Arctic Ocean. In all systems mortality of bacteria caused by viral infections and protistan grazers

was assessed, as well as a number of other parameters, like nutrient concentrations,

microorganisms’ abundances, viral and bacterial production, and bacterial diversity. Detailed

findings and their discussion for each system are presented in five scientific articles included in the

thesis. The present chapter includes a general discussion of the results obtained in each paper,

highlighting the main conclusions of the thesis.

5.1. VIRUS-MEDIATED MORTALITY OF BACTERIA IN THE DIFFERENT SYSTEMS STUDIED

Differences in the environmental parameters among the systems studied

The election of the Mediterranean Sea, and the Atlantic and the Arctic Oceans as systems

for these investigations, provided an opportunity to address questions on the impact of the system

trophy and productivity on viral activity. Nitrite and nitrate concentrations in surface waters (0 – 100

m) were significantly different among the three systems, as well as fluorescence (Table 5.1), yet not

significant differences were found for phosphate concentration. The highest NO2-+NO3-

concentrations were detected in the Arctic Ocean (Paper V), and were significantly higher than

values in the Atlantic Ocean (Paper IV), where the lowest concentrations were detected (Table

5.1). Also, higher fluorescence units, as indicator of phytoplankton biomass, were observed in

Arctic than in Atlantic waters, while Mediterranean waters presented overall low chlorophyll a

concentrations (Paper I).

T


166 | P a g e

Table 5.1. Results of ANOVA test showing significant differences for pooled chemical and biological variables

(Paper I, Paper IV, Paper V) among the surface waters (0 – 100 m) of the three systems studied. Differences

between paired systems were analysed by comparison of means using the Tukey-Kramer (T.-K.) HSD test. ns –

not significant (p > 0.05), nd – not determined, Ar. – Arctic Ocean, At. – Atlantic Ocean, Med. – Mediterranean

Sea, fluor. – fluorescence, VA – viral abundance, BA – bacterial abundance, VBR – virus-bacterium ratio, HNF –

heterotrophic nanoflagellates’ abundance, BP – bacterial production, BS – burst size, VPL – lytic viral

production, VMMBP and PMMBP – virus- and protist-mediated mortality of bacteria as a % of BP. No significant

differences were observed for protist-mediated mortality of bacteria (PMMBP).No data of fluorescence for the

Mediterranean Sea were available.

Variable

System

Test

n

F, r

p

NO2-+NO3- All

Ar. – At.

At. – Med.

Ar. – Med.

ANOVA

T.-K.

T.-K.

T.-K.

142

119

31

134

5.34

0.14

ns

ns

<0.01

0.05

ns

ns

fluor. Ar. – At. ANOVA 129 51.90 <0.0001

VA All

Ar. – At. At. – Med.

Ar. – Med.

ANOVA

T.-K. T.-K.

T.-K.

303

279 155

172

139.12

0.30 0.43

0.82

<0.0001

0.05 0.05

0.05

BA All

Ar. – At.

At. – Med.

Ar. – Med.

ANOVA

T.-K.

T.-K.

T.-K.

294

270

147

171

83.27

0.34

0.25

ns

<0.0001

0.05

0.05

ns

VBR All Ar. – At.

At. – Med.

Ar. – Med.

ANOVA T.-K.

T.-K.

T.-K.

294 270

147

171

197.24 11.12

5.76

18.83

<0.0001 0.05

0.05

0.05

BP All

Ar. – At.

At. – Med.

Ar. – Med.

ANOVA

T.-K.

T.-K.

T.-K.

51

27

30

45

6.63

0.15

0.10

ns

<0.005

0.05

0.05

ns

BS All Ar. – At.

At. – Med.

Ar. – Med.

ANOVA T.-K.

T.-K.

T.-K.

51 27

30

45

28.92 ns

0.05

0.77

<0.0001 ns

0.05

0.05

VPL All

Ar. – At.

At. – Med.

Ar. – Med.

ANOVA

T.-K.

T.-K.

T.-K.

48

24

30

42

26.32

ns

0.19

0.73

<0.0001

ns

0.05

0.05

VMMBP All Ar. – At.

At. – Med.

Ar. – Med.

ANOVA T.-K.

T.-K.

T.-K.

51 27

30

45

9.80 21.08

8.71

ns

<0.0005 0.05

0.05

ns

Discussion

P a g e | 167

Bacterial mortality caused by viruses

The highest mean virus-mediated mortality of bacteria (VMM) in surface waters (0 – 100 m)

was detected in the Atlantic Ocean, and the lowest in the Arctic Ocean (Table 5.1, Fig. 5.1). Losses

of bacterial production (VMMBP) in the studied systems reached respectively 60.5 ± 26.3 % d-1

(Paper IV), 26.6 ± 22.2 % d-1 (Paper I), and 13.2 ± 22.3 % d-1 (Paper V) in Atlantic, Mediterranean and

Arctic waters. These results are within the range of values obtained in other studies (Wommack &

Colwell 2000, and references therein). The great variability of VMM among different studies can be

observed even for the same study system (e.g., the eastern subtropical Atlantic, Rowe et al. 2008,

Winter et al. 2008, and Paper IV), which may be partly due to spatial (depths) or temporal

heterogeneity (seasonal) of the systems, added to a variety of the methods used. Nevertheless, it

can be clearly observed that viruses can cause important bacterial losses in natural pelagic

communities.

Fig. 5.1. Mean values of the principal biological

variables found in the Arctic Ocean (Ar.; Paper

V), Atlantic Ocean (At.; Paper IV), and

Mediterranean Sea (Med.; Paper I). VA – viral

abundance, BA – bacterial abundance, VBR –

virus-bacterium ratio, BS – burst size, BP –

bacterial production, VPL – lytic viral production,

VMMBP and PMMBP – virus- and protist-mediated

losses of BP.


168 | P a g e

Viral lysis vs. protistan grazing

Protist-mediated mortality (PMM) was less variable than VMM, and no significant

differences among systems were detected for the surface layer (0 – 100 m, Table 5.1). Losses of

bacterial production due to grazers (PMMBP) were the highest in the Arctic Ocean (39.7 ± 13.2 % d-

1; Paper V), followed by the Mediterranean Sea (33.2 ± 21.8 % d-1; Paper I), and the Atlantic Ocean

(22.2 ± 19.4 % d-1; Paper IV). Thus, it can be concluded that the dominance of viruses or protists as

a source of bacterial mortality depended on the type of studied system, with the clear dominance

of PMM in the productive Arctic Ocean, strong dominance of VMM in the oligotrophic Atlantic

Ocean, and roughly similar impact of viruses and protists in the coastal oligotrophic Mediterranean

Sea.

Fig. 5.2. Correlation (red equation) among the pooled virus-

(VMMBP) and protist-mediated mortalities of bacteria (PMMBP) in

three different systems. Correlations for each system (Paper I,

Paper IV, Paper V) are presented in different colours (blue for the

Arctic Ocean, black for the Atlantic Ocean, purple for the

Mediterranean Sea).

Pooling the data from the three systems together, a significant negative relationship

among VMM, and both PMM and abundance of heterotrophic nanoflagellates was observed

(Table 5.2). This result was obtained also analyzing each system separately, although for the

Mediterranean the relation was not significant (Fig. 5.2). The negative correlation among VMMBP,

and both PMMBP and HNF abundance seems to contradict results obtained in Paper III, which

suggest synergistic interactions among those two predators of bacteria. In Paper III, and a number

of other studies (Šimek et al. 2001, Weinbauer et al. 2003, Sime-Ngando & Pradeep Ram 2005,

Weinbauer et al. 2007), an enhancement of viral lysis was observed in the presence of protists. It

was suggested, that this is an effect of an increased bacterial fitness due to DOM input from the

grazers activity (Šimek et al. 2001). Yet, in Paper III was also observed an antagonistic interaction

among the two predators, since a negative correlation among viral and protistan abundances

was detected. This could suggest the competition for the same prey, as was previously suggested

Discussion

P a g e | 169

by some field studies (Maranger et al. 2002), and theoretical models (Miki & Yamamura 2005). Thus,

results obtained from the pooled data could suggest that indeed, in natural communities, the

synergistic interactions among phages and protists are blurred by the competition for prey, and by

other factors deciding on the dominance of one of those two sources of bacterial mortality.

Table 5.2. Significant correlation coefficients among variables from Paper I, Paper IV, and Paper V. VA – viral

abundance, temp. – temperature, fluor. – fluorescence, BA – bacterial abundance, VBR – virus-bacterium

ratio, HNF – heterotrophic nanoflagellates, BP – bacterial production, BS – burst size, VPL – lytic viral production,

%VPLG – percentage of lysogenic VP in the total VP, VMMBP – virus-mediated mortality of bacteria as % of BP,

PMMBP – protist-mediated mortality of bacteria.

Variables n r p Variables n r p

VA - depth 456 -0.612 <0.0001 BS - depth 45 0.612 <0.0001

VA - temp. 229 0.586 <0.0001 BS - temp. 68 0.393 <0.001

VA - salinity 221 -0.179 <0.01 BS - salinity 45 0.479 <0.001

VA - NO2-+NO3- 150 -0.328 <0.0001 BS - fluor. 39 -0.616 <0.0001

VA - PO4-3 140 -0.349 <0.0001 BS - BA 66 -0.326 <0.01

VA - fluor. 198 0.519 <0.0001 BS - VBR 66 0.427 <0.0005

BA - depth 442 -0.675 <0.0001 BS - HNF 69 -0.368 <0.002

BA - temp. 223 0.165 <0.02 BS - BP 69 -0.430 <0.0002

BA - salinity 215 -0.373 <0.0001 VPL - temp. 66 0.646 <0.0001

BA - fluor. 193 0.819 <0.0001 VPL - VA 66 0.566 <0.0001

VA - BA 465 0.636 <0.0001 VPL - BA 63 0.259 <0.05

VBR - depth 441 0.125 <0.01 VPL - VBR 63 0.489 <0.0001

VBR - temp. 223 0.667 <0.0001 VPL - BS 66 0.582 <0.0001

VBR - salinity 215 0.451 <0.0001 VPLG - BS 56 -0.354 <0.01

VBR - NO2-+NO3- 149 -0.198 <0.02 VMMBP - depth 45 0.671 <0.0001

VBR - PO4-3 139 -0.181 <0.05 VMMBP - salinity 45 0.548 <0.0001

VBR - fluor. 193 -0.528 <0.0001 VMMBP – fluor. 39 -0.588 <0.0001

VBR - VA 465 0.485 <0.0001 VMMBP - VA 69 -0.444 <0.0001

VBR - BA 465 -0.230 <0.0001 VMMBP - BA 66 -0.569 <0.0001

HNF - depth 116 -0.538 <0.0001 VMMBP - HNF 69 -0.594 <0.0001

HNF - salinity 116 -0.186 <0.05 VMMBP - BP 69 -0.565 <0.0001

HNF - fluor. 93 0.633 <0.0001 VMMBP - BS 69 0.291 <0.02

HNF - VA 140 0.450 <0.0001 PMMBP - depth 45 -0.504 <0.001

HNF - BA 137 0.693 <0.0001 PMMBP - salinity 45 -0.356 <0.02

BP - depth 45 -0.721 <0.0001 PMMBP - PO4-3 43 -0.302 <0.05

BP - temp. 68 0.364 <0.005 PMMBP - VA 69 0.249 <0.05

BP - salinity 45 -0.458 <0.002 PMMBP - BA 66 0.347 <0.005

BP - fluor. 39 0.808 <0.0001 PMMBP - HNF 69 0.385 <0.002

BP - VA 69 0.720 <0.0001 PMMBP - BP 69 0.326 <0.01

BP - BA 66 0.862 <0.0001 PMMBP - VMMBP 69 -0.508 <0.0001

BP - HNF 69 0.774 <0.0001


170 | P a g e

Along the studies presented in this thesis, viruses and protists were together responsible for

removing from 52.9 ± 13.9 % of BP d-1 in the Arctic (Paper V), through 59.7 ± 29.2 % of BP d-1 in the

Mediterranean (Paper I), to 83.6 ± 21.1 % of BP d-1 in the Atlantic waters (Paper IV; Table 5.3). The

losses of ≥100 % of bacterial production have been reported in some studies (e.g., Guixa-Boixereu

et al. 2002, Almeida et al. 2001, Paper I, Paper IV). However, more often a lack of balance

between bacterial production and mortality was found (e.g., Steward et al. 1996, Hwang & Cho

2002, Paper I, Paper V; Table 5.3), which could suggest that obtained results (1) were measured at

the moment when bacterial production was exceeding losses, (2) were underestimated due to

the used methodology, and/or (3) some additional sources of bacterial mortality, like predation by

larger predators or autolysis are also important factors of BP losses.

Organic carbon flux mediated by phages

Lysis of bacterial cells, consequence of viral infections, can be a source of DOM and POM

in the water column. This DOM can be utilized by other bacteria, closing in this way the viral short-

cut of organic carbon (OC) in the system (Bratbak et al. 1990). Results of the present study

revealed, that the mean input of OC to the surface waters (0 – 100 m) due to viral activity was

similar in the three types of systems, with values of 2.97 ± 3.63 μg C L-1 d-1 (from not detectable to

16.82 μg C L-1 d-1) in the Arctic Ocean (Paper V), 2.34 ± 2.07 μg C L-1 d-1 (from 0.06 to 6.40 μg C L-1 d-

1) in the Atlantic Ocean (Paper IV), and 2.51 ± 2.93 μg C L-1 d-1 (from 0.16 to 12.14 μg C L-1 d-1) in the

Mediterranean Sea (Paper I; assuming the cell-carbon factor of 12 fg C cell-1 for the Atlantic

Ocean and Mediterranean Sea, Simon & Azam 1989, and 29 fg C cell -1 for the Arctic Ocean,

Middelboe & Lundsgaard 2003). These mean values are much higher than the range of 0.12 – 0.55

μg C L-1 d-1 found to be released during the cell lysis in the oligotrophic offshore waters of the Gulf

of Mexico (Wilhelm et al. 1998), but are comparable or even lower than values detected in the

Strait of Georgia, British Columbia, characterized by turbulent mixed waters (range of 1.03 – 1.51 μg

C L-1 d-1 for stratified waters, and 2.00 – 8.25 μg C L-1 d-1 for tidally mixed waters, Wilhelm & Suttle

2000). The shunt of OC to the larger predators via protists was more variable among the three

systems. In the Arctic Ocean this C flux was almost four times higher than the flux via phages, with

mean value of 11.34 ± 7.16 μg C L-1 d-1 (from not detectable to 25.35 μg C L-1 d-1), favouring the

productivity of higher trophic levels in this region (Paper V). In the Mediterranean the flux of OC by

protists was similar to that of phages, accounting for 3.12 ± 3.35 μg C L-1 d-1 (from not detectable to

16.09 μg C L-1 d-1; Paper I), and in the Atlantic Ocean it was lower than the flux via viruses, with a

mean value of 0.88 ± 0.92 μg C L-1 d-1 (from not detectable to 2.41 μg C L-1 d-1), enhancing the

recycling of carbon within the microbial loop in the oligotrophic oceanic waters (Paper IV).

Discussion

P a g e | 171

Table 5.3. Bacterial production losses caused by protists and viruses detected simultaneously in different

systems and methods used to determinate prokaryotic mortality caused by viruses. BDR: bacterial decay rate;

VDR: viral decay rate; TEM: determination of frequency of visibly infected cells by transmission electron

microscopy; VDA: virus dilution approach; VRA: virus reduction approach; nd – no detected; *bacterial loss

rate in % h-1 (for details see reference); # bacterial mortality in h-1 or ## in d-1 (for details see reference).

System and

Geographical location Method Losses by

protists

Losses by

viruses Reference

Polar systems

Arctic

Bering & Chukchi Seas

integrated depths: bottom

50 m

Franklin Bay, Canadian Arctic

Greenland Sea, Arctic Ocean

TEM

VDA

VRA

1 – 40

1 – 27

nd – -0.004#

39.7 (nd – 55.1)

2 – 37

3 – 20

-0.006 – -0.015#

13.2 (nd – 100.0)

Steward et al., 1996

Wells and Deming, 2006

Paper V

Antarctic

Bellingshausen Sea

Bransfield Strait

Gerlache Strait

VDR 35.3

37.2

0.9

400

180

44

Guixa-Boixereu et al.,

2002

Eutrophic

Santa Monica, California, USA

Ria de Aveiro, Portugal marine waters

brackish waters

Masan Bay, Korea

Mediterranean Sea, (Masnou harbour,Spain)

BDR

VDA

TEM

VDR

1.8 – 2.2*

2.8 – 3.4*

69

73

41

120

1.2 – 1.4*

2.7 – 2.8*

36

59

9.4

100

Fuhrman and Noble, 1995

Almeida et al., 2001

Choi et al., 2003

Guixa-Boixereu et al., 1999a

Oligotrophic

Mediterranean Sea, Spain

Maditerranean Sea, Spain

East Sea, Korea

Mediterranean Sea, France

Mediterranean Sea, Spain

NW Atlantic Ocean NW Pacific Ocean, Japan

TEM

TEM

TEM

TEM

VRA

VRA VDA

31.0

36.5

74.0

19

9.5 (1.6 – 32.9)

26 – 80

33.9 (nd – 77.6) 32.4 (nd – 77.0)

22.5 (nd – 79.4)

0.05 – 0.13##

nd

nd

nd

17

13.1 (6.9 – 26.9)

nd - 18

12.2 (3.1 – 47.7) 40.9 (9.6 – 84.1)

60.6 (20.6 – 92.0)

0.53 – 0.98##

Guixa-Boixereu et al.,

1999b

Vaqué, unpubl. data

Guixa-Boixereu et al., 1996

Hwang & Cho, 2002

Bettarel et al., 2002

Paper I

Paper IV

Taira et al. 2009


172 | P a g e

5.2. LYSOGENY

Highly variable percentage of lysogenic viral production (%VPLG) of total viral production

(VP) was found along this study, ranging from not detectable to 100 %. Lysogeny was detected in

slightly fewer samples in the more productive Arctic Ocean (31.8 %, Paper V) than in low

productive Atlantic Ocean (45.8 %, Paper IV) and Mediterranean Sea (45.8 %, Paper I). No

differences in the %VPLG among the three types of systems were observed. The only significant

relationship of the pooled %VPLG data from Paper I, Paper IV and Paper V and other parameters

was the negative correlation with burst size (Table 5.2). This could suggest that probably some

factor that influences negatively BS, influences positively lysogeny. However in each system some

trends were observed. For example, a positive correlation with BP in the Mediterranean Sea and

the Atlantic Ocean (Paper I, Paper III, Paper IV), with fluorescence and VA in Atlantic waters

(Paper IV), and with protist-mediated bacterial mortality and lytic VP (VPL) in Arctic waters (Paper

V). These results could confirm the previous findings that lysogens have often better fitness than

uninfected cells (Lin et al. 1977), and suggest that lysogeny could, in fact, increase bacterial

activity. On the other hand, the negative correlation with VPL is in agreement with results of other

studies suggesting that lysogenic bacteria also gain protection from superinfection by other

bacteriophages (Saye & Miller 1989, Campbell 2006). The findings from this study are also in

agreement with the general view that lysogeny should be more frequent in oligotrophic than

eutrophic systems as a survival strategy of phages at low nutrient concentrations (Jiang & Paul

1994, Williamson et al. 2002), but however, they contradict the theory that postulates that more

lysogeny should be detected with lower bacterial production (Weinbauer et al. 2003). Certainly,

more studies on the factors driving lysogeny occurrence and the effect of lysogeny on bacterial

production are needed.

5.3. FACTORS SHAPING VIRAL MORTALITY OF BACTERIA

Bacterial abundance, production, and phylogenetic composition

Metabolic status of the host is critical for viral infection and proliferation (e.g. Steward et al.

1996, Weinbauer et al. 2003). Thus, higher virus-mediated mortality is expected with higher

bacterial production and abundance (Lenski 1988, Murray & Jackson 1992). In this study an

opposite trend was observed (Fig. 5.2). In spite of an increase of VMMBP with BP in Atlantic waters

(Fig. 5.2; Paper IV), pooled VMMBP of the three systems was negatively correlated with BA and BP

(Table 5.2), and lytic viral production (VPL) showed no relationship with those two parameters. Lack

of relation among BP and VPL was observed also by other authors in the Baltic Sea and Masan Bay,

Discussion

P a g e | 173

Korea, which concluded that the virus productivity was not dependent on bacterial abundance

and growth in those systems (Choi et al. 2003, Holmfeldt 2009). This could suggest that the fitness of

bacterial host is not a determining factor for viral infection and proliferation. Some studies indicate,

that the composition of the bacterial communit.y could be more important for viral productivity,

and consequently, bacterial mortality than the metabolic status of the bacterial hosts. For

example, Bouvier & del Giorgio (2007) found that the viral control of bacteria is not proportional to

the host density in some cases, as they found the rare bacterial groups to be more affected by

viral infections than the abundant strains. In other study, Holmfeldt (2009) observed that virus

production rates in the Bothnian Bay (Baltic Sea) were double those found elsewhere in the system,

although bacterial abundance and production were the lowest there. Regarding to the unique

bacterial and viral community composition in that system they suggested that viral production was

influenced by bacterial community composition.

Fig. 5.3. Correlation (red equation) among the pooled virus-mediated mortality of bacteria (VMMBP) and

bacterial abundance (BA), and production (BP) detected along the studies in three different systems.

Correlations for each system (Paper I, Paper IV, Paper V) are presented in different colours (blue for the Arctic

Ocean, black for the Atlantic Ocean, purple for the Mediterranean Sea).

Virus-bacterium ratio

Virus-bacterium rate (VBR) indicates the numerical predominance of viruses over bacteria,

which in marine pelagic waters is typically 5 – 10 (Weinbauer 2004). Intuitively, with higher VBR

more encounter rates are enhanced between viruses and bacteria, and more infections should

be detected. This should be correct, considering that bacteriophages constitute the great part of


174 | P a g e

virioplankton, as can be deduced from the positive correlation among bacterial and viral

abundances found in this (Table 5.2), and other studies (e.g., Jiang & Paul 1994, Payet & Suttle

2008). In this study, however, no significant relationship was found among VMMBP and VBR, and

thus viral abundance and VBR would be not a good predictor of bacterial mortality due to

phages (Fig. 5.3). Nevertheless, a positive correlation among pooled VBR, and both VPL and BS

values detected in this study was observed (Table 5.2), and higher VBR and BS were observed with

lower fluorescence (Table 5.2). This would suggest that viral production could increase due to an

increase of burst size with changing environmental conditions, which do not need to lead to an

enhancement of bacterial losses, and that VBR could be an indicator of VP.

Fig. 5.4. Correlation (red equation) among the pooled virus-mediated mortality of bacteria (VMMBP) and

fluorescence, and viral abundance (VA) detected along the studies in three different systems. Correlations for

each system (Paper I, Paper IV, Paper V) are presented in different colours (blue for the Arctic Ocean, black for

the Atlantic Ocean, purple for the Mediterranean Sea).

Trophic status of the system

It is believed that in oligotrophic systems, low rates of viral infection and low bacterial

losses due to viruses will be found due to lower host abundance (Murray & Jackson 1992, Thingstad

& Lignell 1997). Yet in this study an opposite trend was found, and the highest mean rates of

bacterial mortality due to viruses were found in systems with the lowest nitrite+nitrate

concentrations. Also, during the two-year seasonal study in the Mediterranean Sea (Paper I) similar

relation was observed. This could suggest, that low nutrient concentrations could enhance viral

infections, e.g. increasing occurrence of some structure on the bacterial cell wall (porins) that are

also crucial for the virus attachment (Lenski 1988, Poole & Hancock 1986, Paper I). Furthermore,

Discussion

P a g e | 175

pooled VMMBP values found along this thesis were negatively correlated with fluorescence (Fig.

5.4). Fluorescence values are an indicator of the phototrophic organisms’ abundance, which are

potential source of DOM and POM for bacteria. Obtained results suggest that an increase of the

organic matter (OM) supply increased bacterial fitness (higher BP, Table 5.2), which could, in

consequence, decrease bacterial susceptibility to infection (Wommack & Colwell 2000), or be

related with an increase of the rate of lysogenic infections (Lin et al. 1977). Similar existence of

high viral lysis rates in NW Pacific Ocean (Japan) was reported recently by Taira and colleagues

(2009). Concluding, findings of this thesis seems to not confirm hypotheses suggesting, that viral

control might be stronger in eutrophic than in oligotrophic systems (e.g. Bohannan & Lenski 2000,

Murray & Jackson 1992), and contrast with some field findings, that have shown high bacterial

mortality (up to 100 %) in nutrient-rich environments (e.g., Guixa-Boixereu et al. 1999 a, Almeida et

al. 2001), and low bacterial losses due to viruses, often not detectable, in oligotrophic systems

(e.g., Guixa-Boixereu et al. 1999 b, Bettarel et al. 2002; Table 5.3).

5.4. VIRAL INFECTIONS IN DEEP WATERS

In the Atlantic Ocean, viral production and bacterial mortality due to phages could be

measured for deep waters, from 700 to 1000 m (Paper IV). The relatively high viral abundances and

VBR found there were similar to those found in the same system by other authors (Parada et al.

2007) at depths >1000 m. Also, detected lytic VP was comparable or enen an order of magnitude

higher than those reported by Parada and colleagues (2007), and was found to be sufficient to

maintain the viral standing stock in deep waters. Results of this study indicate that bacteriophages

are an abundant and active part of plankton in deep oceanic waters. Although it was previously

found that lysogeny was more frequent in deep than surface marine waters (Weinbauer et al.

2003), no significant differences in the frequency of detection of lysogeny among depths were

observed (Paper IV), in spite of the low bacterial abundances in deep ocean layers.

5.5. EFFECT OF VIRAL INFECTIONS ON BACTERIAL PHYLOGENETIC DIVERSITY

Bacteriophages are believed to maintain or increase bacterial diversity, and prevent the

dominance of a few bacterial taxa in the community (Thingstad & Lignell 1997, Thingstad 2000).

Although in some studies not clear effect of viruses on bacterial diversity was observed (Hewson &

Fuhrman 2006), other authors provided arguments supporting the “killing the winner” hypothesis

(e.g., Hennes et al. 1995, Auguet et al. 2009, Holmfeldt 2009). The results of the studies included in

this thesis provide further evidence for a positive effect of viruses on bacterial apparent richness.


176 | P a g e

More important role of viruses than protists in shaping bacterial phylogenetic diversity and richness

during the seasonal study in the Mediterranean Sea (Paper II) was supported by the results of

Paper III, where a clear positive impact of viruses, and a negative of protists on the number of

bacterial taxa was observed. Furthermore, it was found that in conditions of heavy viral lysis, viruses

can be the main factor shaping the phylogenetic composition of bacterial community (Paper II).

5.6. OPEN QUESTIONS

Over two decades of intense studies on the marine bacteriophages ecology has brought

to over 1000 scientific papers treating many different aspects of this topic. The fast evolving

techniques, mostly molecular ones, brought an opportunity of answering more detailed and

sophisticated questions. In this doctoral thesis some of important ecological processes have been

clarified, however, new questions have arisen as well:

Which is the cause of the lack of balance between the measured bacterial production rates and

the losses due to viruses and protists?

Could the nutrients availability and consequent porins occurrence on the bacterial cell wall be the

factor influencing viral infections in natural communities?

Which are the factors driving the lysogeny occurrence in marine waters?

Which is the relation of lysogeny and bacterial fitness in natural environments?

Could lysogeny have an effect on phylogenetic structure of bacterial community?

Which is the impact of bacteriphages on particular bacterial strains?

The future studies in the complex field of viral ecology will help to resolve those questions.

5.7. MAIN THESIS CONCLUSIONS

The results obtained in the studies included in this thesis allow to conclude that:

1. Bacteriophages were an important source of bacterial mortality in natural marine systems,

being in some systems the main factor of bacterial production losses.

2. Higher rates of virus-mediated mortality of bacteria were found in oligotrophic system

(Atlantic Ocean), and lower in more productive system (Arctic Ocean), which would

suggest that the trophic status of the ecosystem could determine the rate of viral

infections.

Discussion

P a g e | 177

3. Fitness of bacterial host was not the determining factor for viral infection and proliferation.

4. Bacteriophages and protists were competitors for prey in natural systems, and a negative

relation existed between rates of bacterial mortality caused by those two predators.

However, in experimental conditions a synergistic interaction between both predators

could be observed, favouring bacterial losses due to viruses.

5. Lysogeny was more frequent in oligotrophic than eutrophic systems, and prevailed in

bacterial communities with higher fitness.

6. Bacteriophages enhanced bacterial community richness and phylogenetic diversity,

being a more important factor than protists in shaping community composition.

Resumen de la tesis

Thesis summary – Spanish version

P a g e | 181

6. RESUMEN DE LA TESIS (Thesis summary – Spanish version)

os virus son los entes más numerosos y activos de los ecosistemas aquaticos (marinos y

de agua dulce), con abundancias de hasta 107 virus por mililitro de agua, o 109 virus

por gramo de sedimento (Wommack & Colwell 2000). Debido a su abundancia, después de las

bacterias, son las partículas responsables de una elevada biomasa en el océano (aprox. 200 Mt,

Hambly & Suttle 2005), y probablemente representan el reservorio de diversidad genética más

grande en el océano (Rohwer 2003).

Debido a que la abundancia de los virus normalmente refleja la abundancia de sus

hospedadores, el tipo de virus más abundante en el océano los virus de bacterias son los virus de

las bacterias, los bacteriófagos. Los bacteriófagos (o fagos) son uno de los factores

potencialmente importantes en los cambios de las comunidades bacterianas, tanto en la

abundancia, como en productividad y diversidad. Por tanto, los cambios causados por los fagos

pueden modificar el metabolismo del carbono y sus flujos en las redes tróficas.

A pesar de la importancia de los procesos biológicos y biogeoquímicos impulsados por los

virus para el funcionamiento de los ecosistemas marinos, todavía existen áreas de la ecología de

los virus marinos que necesitan ser investigadas y/o profundizadas. Este trabajo tiene como

objetivo principal contribuir al conocimiento del rol de los bacteriófagos marinos en las redes

tróficas microbianas en los sistemas marinos y su contribución en los flujos de carbono bacteriano

en la columna de agua después de la lisis celular.

L


182 | P a g e

6.1. INTRODUCCIÓN

Bacteriófagos

Los bacteriófagos son partículas de dimensiones entre 30 y 60 nm, aunque se conocen

fagos mayores y menores (Weinbauer 2004). Se componen de ácido nucleico,

desoxirribonucleico (ADN) o ribonucleico (ARN), encerrado en un cápside de proteína.

Los virus no tienen metabolismo propio y para replicarse tienen que entrar dentro del

hospedador y usar la maquinaria y la energía de la célula infectada. La infección vírica ocurre

según un esquema general: (1) anclaje del virus a la superficie de la bacteria, (2) penetración de

la membrana celular e inyección de su genoma en el citoplasma del hospedador, (3) replicación

del material genético del virus, (4) formación de las cápsides con el material genético, y

consiguiente formación de los virus dentro de la célula hospedadora, (5) la lisis de la célula

hospedadora y liberación de una nueva progenie vírica (Campbell 2001). Este proceso es general

para todos los virus, pero la infección puede ir según distintas pautas y puede ser dividido en:

infecciones productivas (líticas y crónicas), reductivas (pseudolisogénicas y lisogénicas), y

destructivas (restrictivas y abortivas; Abedon 2008).

De acuerdo con el tipo de infección, los bacteriófagos presentan tres estrategias distintas

de replicación: productiva (lítica o crónica), lisogénica y pseudolisogénica. La estrategia lítica

lleva a la muerte de la bacteria infectada poco después de la inyección del genoma del virus.

Después de la infección, el genoma del virus redirige el metabolismo del hospedador hacia la

producción de nuevos fagos. La liberación de los fagos maduros está ligada a la destrucción de

la célula bacteriana. Algunos estudios sugieren que la infección lítica puede ser predominante en

las comunidades bacterianas naturales (Wilcox & Fuhrman 1994, Moebus 1983). La estrategia

crónica permite la liberación de nuevos virus por extrusión, pero no significa la destrucción

inmediata de la bacteria (Russel & Model 2006). Hasta la fecha existe solo un estudio sobre la

posible infección crónica in situ (Hofer & Sommaruga 2001). La estrategia lisogénica está

caracterizada por la larga relación entre el fago y el hospedador. El genoma del fago (ahora

llamado profago) forma parte del genoma del hospedador, se replica junto con él, y se transmite

a las nuevas generaciones bacterianas. La presencia del profago no causa la muerte de la

bacteria ni la producción de los nuevos virus, y a menudo mejora el estado fisiológico de la

bacteria. El ciclo lítico en el profago puede ser inducido por numerosos factores (p.e. el pH, la

radiación UV, los agentes químicos), y causa la muerte del hospedador. En los sistemas marinos la

lisogénia debería ser la forma dominante de infección vírica, debido a las condiciones que la

favorecen (baja concentración de nutrientes, tasas bajas de crecimiento bacteriano y bajas

abundancias relativas de los hospedadores). Así, algunos autores opinan que la lisogénia debe


P a g e | 183

ser la estrategia de replicación predominante en el océano (Miller 2006). Ackerman y DuBow

(1987) estimaron que entre 21 % y 60 % de las bacterias en el los sistemas naturales son lisogénicas.

No obstante, los estudios in situ en el campo no dan una respuesta clara. Por un lado, se ha

observado un mayor porcentaje de lisógenos en mar abierto que en ecosistemas costeros (Jiang

& Paul 1994, Weinbauer & Suttle 1999), y también durante los periodos de baja concentración de

fosforo y nitrógeno en el agua (Williamson et al. 2002). Por otro lado, Moebus (1983) encontró que

solamente el 10 % de los 300 fagos marinos que examinó presentaban la estrategia lisogénica.

Además, algunos autores han detectado muy pocas inducciones líticas de las bacterias

lisogénicas procedentes de las aguas oligotróficas alejadas de la costa, comparado con los

sistemas costeros (Jiang & Paul 1996), lo que introduce interrogantes en la teoría de la ocurrencia

de la lisogénia. El cuarto tipo de estrategia de replicación de los virus es la pseudolisogénia. En

este caso el genoma del virus (aquí llamado preprofago) está mantenido en la célula

hospedadora durante largo periodo de tiempo, pero no está integrado en el genoma de la

bacteria, y no se replica con él. Se cree, que la baja concentración de nutrientes impulsa la

replicación pseudolisogénica (Ripp & Miller 1997). Es posible que la pseudolisogénia sea un

fenómeno común entre los bacteriófagos (Ackerman & DuBow 1987) y en los sistemas marinos

(Moebus 1996), aunque hasta la fecha han sido descritos pocos fagos pseudolisogénicos

acuáticos (p.e. Moebus 1997).

Uno de los objetivos de esta tesis ha sido la evaluación de la frecuencia

relativa de las infecciones líticas y lisogénicas en los sistemas con distinta

trofia, como las aguas costeras del Mar Mediterráneo, el Océano Atlántico

subtropical y el Océano Ártico (Artículos I, IV y V).

Bacteriófagos como modificadores de las comunidades bacterianas

Los bacteriófagos pueden modificar las comunidades bacterianas, alterando la

abundancia de las bacterias, su productividad y diversidad y, como consecuencia, afectar los

flujos de nutrientes en dicho ecosistema.

La lisis vírica, junto con la depredación por protistas, es una causa importante de

mortalidad de las comunidades bacterianas acuáticas. Se calcula que los fagos eliminan

diariamente entre 10 % y 40 % de las bacterias en los sistemas marinos (Thingstad et al. 2008). Se

ha observado una gran variación en las pérdidas ocasionadas en la producción bacteriana, que

pueden ir desde no detectadas (p.e. Guixa-Boixereu et al. 1996) hasta más del 100 % (p.e. Guixa-

Boixereu et al. 2002). Los bacteriófagos han sido en distintos sistemas el factor principal de


184 | P a g e

pérdidas de bacterias (Guixa-Boixereu et al. 1999a, Wells & Deming 2006), tienían un impacto

comparable a los protistas (Fuhrman & Noble 1995, Hwang & Cho 2002), o de menor importancia

(Guixa-Boixereu et al. 1996, Choi et al. 2003). No obstante, a pesar del elevado número de

estudios que evalúan la mortalidad debida a los virus, la comparación de los resultados obtenidos

por distintos autores es difícil, debido a la gran variabilidad espacial y temporal de los sistemas

investigados, y en parte a la variedad de métodos usados.

Según varios estudios, el impacto de los fagos sobre la abundancia bacteriana varía con

el estado trófico del sistema. Los virus serían la causa más importante de mortalidad bacteriana

en los sistemas eutróficos, y los protistas en los sistemas oligotróficos. Esto está basado en dos

argumentos. Primero, los modelos matemáticos sugieren que la probabilidad de contacto entre el

virus y la bacteria incrementa con el incremento de la abundancia del hospedador (Murray &

Jackson 1992), favoreciendo la tasa de infecciones víricas y la mortalidad de bacterias. Segundo,

los fagos infectan preferentemente los hospedadores más activos y productivos (Lenski 1988). De

acuerdo con esto, en muchos estudios se ha observado que los virus eran la causa más

importante de mortalidad bacteriana en sistemas eutróficos (Weinbauer & Suttle 1999, Noble &

Fuhrman 2000), y los protistas en aguas más oligotróficas (Guixa-Boixereu et al. 1996, Bettarel et al.

2002). No obstante, otros autores encontraron más lisis que depredación por protistas en los

sistemas con baja concentración de nutrientes (Wells & Deming 2006).

Uno de los efectos más importantes de la lisis vírica de las células bacterianas es el aporte

de carbono orgánico, nutrientes y elementos traza al agua. Se estima, que el carbono

removilizado por los virus puede aportar aproximadamente entre un 5 % y un 7 % de la demanda

bacteriana de carbono en las aguas oligotróficas en mar abierto, 30 % en las aguas mesotróficas

costeras en la Bahía de México, y aproximadamente 80 % - 95 % en el mesotrófico Estrecho de

Georgia (Wilhelm & Suttle 2000), y hasta un 180 % en las aguas eutróficas del Mar Báltico (Riemann

et al. 2009, Holmfeldt 2009). En las aguas mezcladas verticalmente estos valores llegaron incluso a

140 - >1000 % de la demanda bacteriana de carbono (Wilhelm & Suttle 2000). Esto indica que la

lisis vírica puede ser un mecanismo clave para aportar materia orgánica disuelta (dissolved

organic matter DOM) en algunos sistemas marinos.

Las pérdidas de bacterias han sido evaluadas en distintos sistemas marinos

y comparadas con las pérdidas causadas por los nanoprotistas. Los

resultados de estos estudios se muestran en los Artículos I, IV y V.


P a g e | 185

Diversidad bacteriana

Los bacteriófagos pueden modificar la diversidad bacteriana de cuatro maneras

diferentes: (1) eliminando las bacterias que ganan la competición por los recursos, la hipótesis

“killing the winner (matar al ganador)”; (2) liberando el contenido de las células bacterianas al

medio durante la lisis (Weinbauer & Rassoulzadegan 2004); (3) cambiando la estructura de la

comunidad gracias a las infecciones lisogénicas; y (4) posibilitando el intercambio genético entre

las especies y cepas de bacterias.

La hipótesis “matar al ganador” (Thingstad & Lignell 1997, Thingstad 2000) sugiere que el

tipo de bacterias más abundante en el ecosistema estará el más afectado por las infecciones

víricas y, por tanto, los tipos menos abundantes serán beneficiados de la eliminación de los

competidores, y el incremento de la disponibilidad de los recursos favorecerá su crecimiento. Esta

hipótesis está basada en la suposición que (a) el crecimiento de las bacterias está limitado por la

disponibilidad de nutrientes, y (b) la abundancia bacteriana está controlada por la depredación

por protistas. Además, la hipótesis está basada en el modelo matemático que prevé que las tasas

de contacto entre el virus y su hospedador aumenta con el incremento de la abundancia

bacteriana (Murray & Jackson 1992). La hipótesis “matar al ganador” sugiere, que ni la tasa de

crecimiento, ni la abundancia de la comunidad bacteriana están controlados por virus, pero si la

diversidad bacteriana. Los fagos mantendrían la riqueza de las especies, previniendo que

ninguna especie en particular domine la comunidad, y permitiendo la existencia y crecimiento de

las especies bacterianas menos competitivas.

La hipótesis “matar al ganador” ha sido respaldada por numerosos estudios in situ. Las

oscilaciones de las abundancias de distintos fagos y sus hospedadores muestran una típica pauta

depredador-presa, con las abundancias totales de los virus y bacterias relativamente estables en

la Bahía de Chesapeake (Wommack et al. 1999 a, b). En numerosos estudios, la presencia de los

virus cambió la composición de la comunidad bacteriana (Weinbauer & Höfle 1998, Auguet et al.

2009) y de fitoplancton (Suttle 1992). A pesar de estos datos que apoyan dicha hipótesis, algunos

experimentos con comunidades bacterianas presentaron las respuestas negativas a la presencia

y actividad de los virus (p.e. Hewson & Fuhrman 2006), lo que sugiere que el impacto de los fagos

sobre la diversidad bacteriana es un proceso complejo, que puede depender de una gran

variedad de factores (p.e. medioambientales, lisogénia).

El efecto de la actividad de los virus sobre la composición de la

comunidad bacteriana ha sido evaluado in situ y experimentalmente en el

Mar Mediterráneo (Artículos II y III).


186 | P a g e

Bacteriófagos en el bucle microbiano

El papel central de las bacterias y sus depredadores, los protistas, en los flujos de materia

orgánica en las redes tróficas de los ecosistemas acuáticos fue descrito por primera vez en 1974

(Pomeroy 1974). Más tarde, esta idea fue mejorada cuando se postuló la existencia del “bucle

microbiano” (Azam et al. 1983, Ducklow 1983). El bucle microbiano es una parte fundamental de

todas las redes tróficas acuáticas. Permite la regeneración de los macronutrientes, sobre todo el

nitrógeno y fósforo, lo que permite el crecimiento del fitoplancton y el flujo de materia orgánica

hacía niveles tróficos superiores. Brevemente, el bucle microbiano se inicia con el consumo

bacteriano de materia orgánica de origen fitoplanctónico, y el posterior consumo de las

bacterias por protistas y estos por el zooplancton, que finalmente es consumido por los peces. El

proceso devuelve a la cadena trófica principal la materia orgánica que sería perdida para los

depredadores más grandes (Azam et al. 1983). Los componentes esenciales del bucle microbiano

marino son las bacterias heterótrofas, las principales consumidoras de la DOM en el océano

(Azam 1998), y los protistas herbívoros, los principales consumidores de la producción primaria en

los sistemas marinos. Los protistas heterótrofos pequeños (sobre todo flagelados, <5 μm) movilizan

el carbono bacteriano y del picofitoplancton (<5 μm) hacía protistas de mayor tamaño (sobre

todo ciliados), y finalmente hacía el zooplancton. El bactrioplancton incorpora de un 10 % a un 50

% de la producción primaria (Azam et al. 1983). Por otro lado, el carbono y los nutrientes

contenidos en las células bacterianas y de fitoplancton pueden ser devueltas al conjunto de la

DOM por la acción de sus respectivos virus.

La lisis vírica de los organismos fotosintéticos, como algas y cianobacterias, transfiere

aproximadamente entre un 6 % hasta un 26 % del carbono fijado durante fotosíntesis hacía la

DOM (Wilhelm & Suttle 1999). Durante la lisis de las bacterias, los fagos convierten la materia

orgánica encerrada den las células bacterianas a forma disuelta, que puede ser accesible

nuevamente para otras bacterias heterótrofas. Este proceso, denominado “el bucle vírico”

(Bratbak et al. 1992), excluye los niveles tróficos superiores de la utilización de este carbono. La

conversión de la biomasa bacteriana en DOM es un proceso muy eficiente, y la lisis vírica

representa por tanto una fuente de substrato rico en nutrientes para la producción bacteriana

(Bratbak et al. 1990, Middelboe & Jørgensen 2006). Además, la liberación de la DOM mediada

por los fagos resulta en un incremento de las pérdidas por respiración, pero también en un

incremento sustantivo de la producción bacteriana (Fuhrman 1999).

El efecto directo de la actividad de los virus consiste en la eliminación del sistema de las

presas potenciales para protistas. Los modelos indican que la presencia de los fagos siempre

disminuye el transporte del carbono hacía los protistas (Miki et al. 2008), lo que puede tener un

efecto profundo sobre toda la red trófica. Las evaluaciones teóricas calculan que, asumiendo


P a g e | 187

que la lisis vírica causa un 50 % de la mortalidad bacteriana, la producción bacteriana debida a

la DOM liberada aumenta un 27 %, y la exportación del carbono bacteriano hacía los

depredadores nanoplanctónicos, disminuye un 37 %, comparado con las redes tróficas sin lisis

vírica. Esto resulta en la pérdida neta del 25 % de la producción de nanoprotistas (Fuhrman 1992).

Cálculos similares, que incluyen también la infección vírica del fitoplancton (10 % de la

mortalidad) y las pérdidas de virus debidas al consumo por nanoprotistas (13 % de la producción

vírica), resultaron en un 33 % de incremento de la producción y respiración bacteriana, y en un 20

% de reducción de la producción de nanoprotistas (Fuhrman & Suttle 1993). Se cree que las

infecciones por fagos tienen un mayor impacto sobre la producción del mesozooplancton en

sistemas oligotróficos, donde predomina el reciclaje de la materia orgánica (Murray & Eldridge

1994). En estos ecosistemas las bacterias forman un porcentaje elevado de la biomasa (Fuhrman

et al. 1989), y la producción bacteriana constituye entre un 15 % y un 25 % de la nutrición del

mesozooplancton (desde 20 hasta 200 μm, Murray & Eldridge 1994). En cambio, en los sistemas

mesotróficos los productores primarios son la principal fuente de alimentación del

mesozooplancton. En estos sistemas, las pérdidas de bacterias debidas a los fagos causarán un

reducción menor de la producción del mesozooplancton (1.2 % - 7.4 %, Murray & Eldridge 1994).

El proceso de la lisis vírica permite también la retención de nutrientes en la zona eufótica

de los sistemas acuáticos, lo que puede ser especialmente importante en sistemas oligotróficos,

donde podría contribuir a disminuir el descenso de nutrientes hacia aguas profundas (Bratbak et

al. 1990, Fuhrman 1999). Debido al impacto que pueden producir sobre los ecosistemas naturales,

los virus han sido incluidos como electo clave en el bucle microbiano (Fuhrman 1992, Wommack

& Colwell 2000).

El aporte de los nutrientes desde las células bacterianas al agua mediado

por la lisis vírica ha sido evaluado a lo largo de esta tesis, y comparado

con el flujo de carbono hacia la red trófica mediado por protistas (Artículo

I, IV y V).


188 | P a g e

6.2. HIPÓTESIS Y OBJETIVOS DE LA TESIS

Hipótesis

A partir del conocimiento existente que ha sido resumido anteriormente, en la presente tesis se

formulan las siguientes hipótesis:

1. En los sistemas oligotróficos, las pérdidas de bacterias debidas a protistas son más altas que

las debidas a virus. Contrariamente, en los sistemas eutróficos las pérdidas de bacterias

causadas por fagos son similares o más importantes que las causadas por protistas. Por tanto,

los fagos podrían modificar los flujos de carbono en los sistemas marinos.

2. La lisogénia es el tipo dominante de la estrategia de replicación en los sistemas oligotróficos, y

la infección lítica de las bacterias predomina en los sistemas eutróficos.

3. Los virus son un factor importante en la modificación de la composición de la comunidad

bacteriana, y los protistas, debido a las interacciones sinérgicas o antagónicas con los fagos,

pueden influir sobre la diversidad bacteriana.

Objetivos

El objetivo principal de este trabajo era determinar los efectos de la actividad de bacteriófagos

sobre el bacterioplancton (su abundancia, pérdidas y diversidad), comparados con la actividad

de los protistas. Para ello se evaluaron los factores que influyen el proceso de la infección vírica

(lítica y lisogénica) sobre las bacterias y su depredación por protistas. Además, se ha determinado

el tipo de interacciones que existían entre estos dos depredadores de bacterias, y el efecto de

estas interacciones sobre las comunidades bacterianas. Objetivos detallados de la tesis:

1. Estimas del rango de la mortalidad bacteriana causada por las infecciones víricas en

distintos sistemas marinos, y:

comparación con la mortalidad bacteriana causada por la depredación de

protistas (Articulo I, IV y V),

determinación del carácter de las interacciones entre los fagos y protistas, y el

efecto de estas interacciones en las pérdidas bacterianas debidas a los fagos

(Artículo III),

evaluación la variabilidad de los tipos de infección vírica lisogénica en distintos

sistemas (Artículo I, IV y V),

estima del flujo del carbono orgánico a la columna de agua aportado por la lisis

bacteriana después de la infección por virus (Artículo I, IV y V).


P a g e | 189

2. Investigación del efecto del estado trófico del sistema sobre la tasa de infecciones víricas

(Artículo IV).

3. Determinación de la actividad de los fagos y protistas como responsables de los cambios

en la estructura de la comunidad bacteriana (Artículo II), y:

evaluación de los efectos de las posibles interacciones entre fagos y protistas

sobre la diversidad bacteriana (Artículo III).

6.3. DISCUSIÓN

Esta tesis está basada en los estudios llevados a cabo en tres distintos sistemas marinos: la

zona costera del Mar Mediterráneo, el Océano Atlántico subtropical, y el Océano Ártico. En todos

los sistemas se ha evaluado la mortalidad de bacterias causada por los virus y protistas, además

de la concentración de nutrientes, abundancias de microorganismos, producción vírica y

bacteriana, y diversidad bacteriana. Con el fin de determinar la producción vírica y la mortalidad

de bacterias debida a virus se ha empleado el método de la reducción de virus (Wilhelm et al.

2002, Weinbauer et al. 2002), y para determinar la depredación por protistas se ha utilizado el

método de desaparición de bacterias marcadas fluorescentes (FLB, fluorescent labeled bacteria;

Sherr et al. 1987, Vàzquez-Domínguez et al. 1999). Los detalles de los métodos están incluidos en el

Anexo de esta tesis. Los resultados y su discusión para cada uno de los sistemas se encuentran en

cinco artículos científicos incluidos en esta memoria. El presente capítulo incluye una discusión de

las generalidades encontradas en los tres sistemas marinos.

6.3.1. MORTALIDAD DE BACTERIAS DEBIDA A VIRUS EN DISTINTOS SISTEMAS

La elección del Mar Mediterráneo y los Océanos Atlántico y Ártico para las

investigaciones de esta tesis ha permitido evaluar el efecto de la trofia del sistema y su

productividad sobre la actividad de los virus. Las concentraciones de los nitratos+nitritos en las

aguas superficiales (0 – 100 m) eran significativamente diferentes entre los tres sistemas, igual que

la fluorescencia (Tabla 5.1, Capítulo 5), pero no se encontraron diferencias significativas en la

concentración de fosfatos. La concentración más elevada de NO2-+NO3- fue detectada en el

Océano Ártico (Artículo V), que era significativamente más alta que en el Océano Atlántico

(Artículo IV), donde se observaron las concentraciones más bajas (Tabla 5.1, Capitulo 5). Además,

los indicadores de biomasa fitoplanctónica (fluorescencia y clorofila a) mostraron valores distintos

en los tres sistemas: en aguas árticas obtuvieron valores más altos de fluorescencia que en el


190 | P a g e

Atlántico, mientras que en el Mediterráneo se detectaron bajas concentraciones de la clorofila a

(Articulo I).

Variación de la mortalidad bacteriana causada por fagos

El promedio más alto de mortalidad bacteriana causada por virus (VMM) en aguas

superficiales (0 – 100 m) fue detectado en el océano Atlántico, y el más bajo en el Océano

Ártico. Las pérdidas de producción bacteriana (VMMBP) en los sistemas estudiados era de 60.5 ±

26.3 % d-1 en el Océano Atlántico (Artículo IV), 26.6 ± 22.2 % d-1 en el Mar Mediterráneo (Artículo I)

y 13.2 ± 22.3 % d-1 en el Océano Ártico (Artículo V). Estos resultados están dentro del rango de

valores obtenidos en otros estudios, tanto en sistemas eutróficos, como en los oligotróficos

(Wommack & Colwell 2000). Se observó una gran variabilidad de la VMM entre distintos estudios,

incluso para un mismo sistema (p.e. para el Atlántico subtropical este, Rowe et al. 2008, Winter et

al. 2008, y Artículo IV), lo que puede ser en parte debido a la variabilidad espacial y temporal del

sistema, así como a los distintos métodos utilizadas en distintos estudios. A pesar de esto se puede

observar que los virus pueden ser una causa importante de la mortalidad de bacterias en las

comunidades pelágicas.

Lisis vírica vs. depredación por protistas

La mortalidad bacteriana debida a protistas (PMM) era menos variable que la VMM, y no

se observaron diferencias significativas entre las aguas superficiales de los tres sistemas (0 – 100 m,

Tabla 5.1, Capitulo 5). Las pérdidas de producción bacteriana debida a protistas (PMMBP) fueron

más altas en el Océano Ártico (39.7 ± 13.2 % d-1; Artículo V), seguidas por el Mar Mediterráneo

(33.2 ± 21.8 % d-1; Artículo I), y el Océano Atlántico (22.2 ± 19.4 % d-1; Artículo IV). Se puede

concluir entonces, que la mayor contribución al porcentaje de mortalidad bacteriana por parte

de los virus o los protistas dependía del tipo de sistema, con un mayor porcentaje de la PMM en el

productivo Océano Ártico, elevado porcentaje de la VMM en el Océano Atlántico, y un impacto

similar de virus y protistas en la zona costera del oligotrófico Mar Mediterráneo (Observatori de

Microbiologia Marina de Blanes).

Sintetizando todos los datos obtenidos en los distintos sistemas, se ha observado una

correlación negativa significativa entre las VMMBP y PMMBP (Tabla 5.2, Capítulo 5). Este patrón se

obtuvo también analizando los datos de cada sistema por separado, aunque en el Mar

Mediterráneo la correlación no era significativa. La correlación negativa entre las VMMBP y tanto

con PMMBP como con abundancia de nanoflagelados heterotrópicos, parece ser contradictoria

a los resultados del Artículo III, que indicaban en la mayoría de experimentos una interacción

sinérgica entre estos dos depredadores de bacterias. En el Artículo III y en otros estudios (Šimek et


P a g e | 191

al. 2001, Weinbauer et al. 2003, Sime-Ngando & Pradeep Ram 2005, Weinbauer et al. 2007), el

aumento de la lisis vírica fue observado en presencia de protistas. Esto podía ser debido a que la

actividad depredadora de los protistas, ya sea sobre bacterias heterótrofas, u otras presas pico-

nanplanctónicas aportaba DOM a las bacterias presentes, mejorando su estado fisiológico (Šimek

et al. 2001). Sin embrago, en el Articulo III se podía observar también una interacción negativa

entre los dos depredadores, y la correlación negativa entre sus abundancias. Esto apunta a una

competición por la presa, como ha sido sugerido por algunos estudios de campo (Maranger et al.

2002) y modelos teóricos (Miki & Yamamura 2005). Los resultados obtenidos en esta tesis sugieren

pues, que en las comunidades naturales las interacciones sinérgicas entre fagos y protistas

pueden estar enmascaradas por la competición por la presa, u otros factores que deciden sobre

la dominancia de estas dos causas de la mortalidad bacteriana.

En los estudios presentados en esta memoria, se ha visto que los virus y los protistas son

responsables conjuntamente de la eliminación de la producción bacteriana (BP) del 52.9 ± 13.9 %

d-1 en las aguas árticas (Artículo V), 59.7 ± 29.2 % de BP d-1 en el Mediterráneo (Artículo I), hasta

83.6 ± 21.1 % de BP d-1 en las aguas atlánticas (Artículo IV; Tabla 5.3, Capitulo 5). Algunos estudios

muestran pérdidas de producción bacteriana de ≥100 % (p.e. Guixa-Boixereu et al. 2002, Almeida

et al. 2001; Tabla 3, Artícluo I, Artículo IV). Sin embargo, frecuentemente se encuentra un

desequilibrio entre la producción y la mortalidad bacteriana (p.e. Steward et al. 1996, Hwang &

Cho 2002, Artículo I, Artículo V), la que se puede deber al momento de muestreo, los métodos

utilizados, o a causas adicionales de la mortalidad bacteriana (depredación por los

depredadores mayores o la autólisis).

Flujos del carbono mediados por fagos

La lisis de las células bacterianas, consecuencia de las infecciones víricas, son una fuente

de la DOM y materia orgánica particulada (particulate organic matter POM) en el agua. Esta

DOM puede ser utilizada de nuevo por otras bacterias, cerrando de esta manera el ciclo vírico

del carbono orgánico en el sistema (Bratbak et al. 1990). Los resultados de esta tesis revelan que

el aporte promedio del carbono orgánico en las aguas superficiales (0 – 100 m) debido a la

actividad de los fagos era similar en los tres sistemas estudiados, con valores de 2.97 ± 3.63 μg C L-1

d-1 (no detectable - 16.82 μg C L-1 d-1) en el Océano Ártico (Artículo V), 2.34 ± 2.07 μg C L-1 d-1 (0.06

- 6.40 μg C L-1 d-1) en el Océano Atlántico (Artículo IV), y 2.50 ± 2.90 μg C L-1 d-1 (0.16 - 12.14 μg C L-1

d-1) en el Mar Mediterráneo (Artículo I; asumiendo un factor de contenido de carbono de 12 fg C

por célula bacteriana en el Océano Atlántico y Mar Mediterráneo, Simon & Azam 1989, y 29 fg C

por célula en el Océano Ártico, Middelboe & Lundsgaard 2003). Estos valores promedios son

mucho más altos que el rango del carbono liberado por la lisis vírica en aguas oligotróficas de


192 | P a g e

mar abierto del Golfo de México (0.12 – 0.55 μg C L-1 d-1; Wilhelm et al. 1998), pero son

comparables o incluso más bajos a los valores encontrados en el Estrecho de Georgia, Colombia

Británica, caracterizadas por la mezcla turbulenta de las aguas (1.03 – 1.51 μg C L-1 d-1 para aguas

estratificadas, y 2.00 – 8.25 μg C L-1 d-1 para aguas mezcladas por mareas; Wilhelm & Suttle 2000).

El flujo del OC hacia los niveles tróficos superiores vía protistas fue más variable entre los tres

sistemas. En el Océano Ártico este flujo de C era casi cuatro veces más alto que el flujo vía fagos,

con un valor promedio de 11.34 ± 7.16 μg C L-1 d-1 (no detectable - 25.35 μg C L-1 d-1),

favoreciendo la productividad de los niveles tróficos más altos en esta región (Artículo V). En el

Mediterráneo, el flujo de carbono orgánico por protistas era similar al de los fagos, con un valor

medio de 3.12 ± 3.35 μg C L-1 d-1 (no detectable - 16.09 μg C L-1 d-1; Artículo I), y en el Océano

Atlántico era más bajo que el flujo vía viruses, con un promedio de 0.88 ± 0.92 μg C L-1 d-1 (no

detectable - 2.41 μg C L-1 d-1), aumentando el reciclaje de carbono en el bucle microbiano en las

oligotróficas aguas oceánicas (Artículo IV).

6.3.2. LISOGÉNIA

A lo largo de este estudio se ha encontrado un porcentaje altamente variable de

producción vírica lisogénica (%VPLG) respecto a la producción vírica total (VP). El rango de valores

de la %VPLG oscilaba entre “no detectable” y 100 %. En el Océano Ártico, que era el sistema más

productivo de los estudiados, la lisogénia fue detectada en un menor número de casos (31.8 %,

Artículo V) que en sistemas menos productivos como el Océano Atlántico (45.8 %, Artículo IV) y el

Mar Mediterráneo (45.8 %, Artículo I). Sin embargo no se encontraron diferencias significativas en

el porcentage %VPLG entre los tres sistemas. La única relación significativa entre %VPLG,

sintetizando los datos de los Artículos I, IV y V, y otros parámetros ha sido la correlación negativa

entre la %VPLG y el tamaño de explosión (burst size, BS; Tabla 5.2, Capitulo 5). Esto sugiere que

algún factor que influye negativamente sobre el BS, influye positivamente sobre la lisogénia. Sin

embargo, en cada sistema estudiado se observaban correlaciones positivas entre %VPLG y (1) BP

en el Mar Mediterráneo y Océano Atlántico (Artículo I, Artículo III, Artículo IV), (2) valores de

fluorescencia y abundancia de virus en aguas atlánticas (Artículo IV), o (3) mortalidad bacteriana

debida a protistas y la producción vírica lítica (VPL) en as aguas árticas (Artículo V). Ello

confirmaría los resultados de los estudios anteriores, que indicaban que las bacterias lisogénicas a

menudo son las que tienen un mejor estado fisiológico que las células no infectadas (Lin et al.

1977), y sugieren que la lisogénia podría incluso aumentar la actividad bacteriana. Por otro lado,

la correlación negativa con VPL concuerda con los resultados de otros estudios, que indican que

las bacterias lisogénicas adquieren una protección ante la sobre-infección por otros


P a g e | 193

bacteriófagos (Saye & Miller 1989, Campbell 2006). Los resultados de los estudios incluidos en esta

tesis están también de acuerdo con el concepto general, que la lisogénia debería ser más

frecuente en sistemas oligotróficos que en eutróficos como una estrategia de supervivencia ante

las bajas concentraciones de nutrientes (Jiang & Paul 1994, Williamson et al. 2002), pero

contradice a la teoría de que dice que en las comunidades bacterianas menos productivas el

grado de incidencia de la lisogenia es mayor (Weinbauer et al. 2003). Indudablemente, la

complejidad de estas relaciones sugiere que se precisen más estudios futuros para elucidar con

mayor detalle cual es la relación entre la lisogenia y la producción bacteriana.

6.3.3. FACTORES DETERMINANTES DE LA MORTALIDAD BACTERIANA DEBIDA A LOS VIRUS

Abundancia y producción bacteriana

Numerosos estudios (p.e. Steward et al. 1996, Weinbauer et al. 2003) muestran que el

estado fisiológico del hospedador es critico para la infección y proliferación vírica. Como

consecuencia, con una mayor producción y abundancia bacteriana se espera una mortalidad

bacteriana debida a virus más elevada (Lenski 1988, Murray & Jackson 1992). En los trabajos

presentados en esta memoria, no siempre un incremento de la BP iba unido a un incremento de

la VMMBP. Así, este patrón se cumple para las aguas atlánticas (Fig. 5.1, Capitulo 5; Artículo IV), no

observándose correlación significativa entre estas dos variables en el Océano Ártico, ni tampoco

en el Mar Mediterráneo. Sin embargo, cuando se juntan todos los valores de VMMBP de todos los

sistemas analizados, se obtiene una correlación negativa con la abundancia y producción

bacteriana (Tabla 5.2, Capitulo 5), mientras que la producción vírica lítica (VPL) no presenta una

relación significativa con estos dos parámetros. Esta falta de correlación entre la BP y la

mortalidad debida a virus o producción vírica ha sido observada también en otros estudios (Choi

et al. 2003, Holmfeldt 2009). Esto sugiere que el estado fisiológico del hospedador no es el único

factor determinante para la infección y proliferación vírica. Algunos autores proponen que la

composición de la comunidad bacteriana puede ser un factor más determinante para la

producción vírica, y en consecuencia para la mortalidad bacteriana, que la producción

bacteriana de la comunidad (Bouvier & del Giorgio 2007, Holmfeldt 2009).

Cociente virus-bacteria

El cociente virus-bacteria (virus-bacterium ratio, VBR) indica la predominancia numérica

de virus sobre bacterias, y en los sistemas marinos dicho cociente oscila entre 5 y 10 (Weinbauer

2004). Así, a mayor VBR, mayor probabilidad de encuentro entre virus y bacterias y, por tanto, má

infecciones deberían ser detectadas. Esto debería ser correcto considerando que los


194 | P a g e

bacteriófagos constituyen la mayor parte del virioplancton, lo que puede ser deducido de la

correlación positiva entre las abundancias bacterianas y víricas encontradas en este (Tabla 5.2,

Capitulo 5) y otros estudios (p.e. Jiang & Paul 1994, Payet & Suttle 2008). En el presente estudio no

se obtuvo una relación significativa entre VMMBP y VBR, por lo cual la abundancia de virus y el

VBR no se podrían considerar para predecir la mortalidad bacteriana debida a fagos. Sin

embargo, juntando todos los datos, se detectó una correlación positiva y significativa entre el VBR

y valores de VPL y BS, a la vez que los valores más altos de VBR y BS eran observados cuando los

valores de unidades de fluorescencia decrecían (Tabla 5.2, Capitulo 5). Esto sugiere que la

producción de virus (VPL) podría incrementar debido al aumento del tamaño de explosión celular

(BS), por cambios en las condiciones ambientales, lo cual no incrementaría las perdidas de

bacterias, pero si la producción vírica. En nuestro caso parece ser que el VBR seria útil para

predecir la VPL.

Estatus trófico del sistema

Está comúnmente aceptado que en los sistemas con bajas concentraciones de

nutrientes, clorofila a y abundancias de bacterias, se caracterizan por tasas bajas de infección

vírica y bajas pérdidas bacterianas debidas a fagos (Murray & Jackson 1992, Thingstad & Lignell

1997). Sin embargo, en la presente memoria se ha observado valores medios de mortalidad

bacteriana por virus más altos en sistemas con concentraciones de nutrientes (nitritos+nitratos)

más bajos (Artículo I, Artículo IV). Considerando los estudios por separado, en el Mar

Mediterráneo (ciclo estacional 2005-2006) se observaron bajas tasas de mortalidad bacteriana

por virus en paralelo con bajas concentraciones de clorofila a y nutrientes (Articulo I). Por tanto, el

nivel trófico del sistema, no siempre es un buen indicador del comportamiento y las estrategias de

vida de los microorganismos. Puede suceder que concentraciones bajas de nutrientes puedan

favorecer las infecciones víricas, p.e. incrementando la cantidad de unas estructuras llamadas

porinas en la pared celular bacteriana, las cuales son cruciales para el acoplamiento de los virus

en la superficie de la bacteria (Lenski 1988, Poole & Hancock 1986, Artículo I). Al mismo tiempo, los

valores de VMMBP encontrados a lo largo de esta tesis estaban negativamente correlacionados

con los valores de fluorescencia (Fig. 5.4, Capitulo 5). Los organismos fototróficos son una fuente

potencial de DOM y POM para las bacterias, y los resultados obtenidos sugieren que un mayor

aporte de la materia orgánica, generalmente de origen fitoplanctónico, elevaba la producción

bacteriana (Tabla 5.2, Capitulo 5), lo que podía en consecuencia disminuir la susceptibilidad de

las bacterias a la infección (Wommack & Colwell 2000), o incrementar la tasa de infecciones

lisogénicas. Parecido “incumplimiento” entre el estatus oligotrófico del sistema y altas tasas de

mortalidad por virus en el Océano Pacífico NO (Japón) han sido reportado recientemente por


P a g e | 195

Taira y coautores (2009). En resumen, los resultados de esta tesis no parecen confirmar las hipótesis

que sugieren que el control de los virus sobre el bacterioplancton podría ser más importante en los

sistemas eutróficos que en los oligotróficos (Bohannan & Lenski 2000, Murray & Jackson 1992), y

contrastan con los resultados de algunos estudios en campo obtenidos previamente (p.e. Guixa-

Boixereu et al. 1999 a, Almeida et al. 2001).

6.3.4. INFECCIONES VÍRICAS EN AGUAS PROFUNDAS

En el estudio llevado a cabo en el Océano Atlántico, la producción vírica y la mortalidad

bacteriana debida a los virus han sido medidas en profundidades de zonas meso-bathypelágicas

(700 – 1000 m; Artículo IV). Elevadas abundancias de virus y tasas VBR fueron detectadas a estas

profundidades, valores similares a los encontrados por otros autores en el mismo sistema (Parada

et al. 2007) a profundidades >1000 m. También la producción vírica lítica era comparable o más

alta (un orden de magnitud) que los valores de Parada y coautores (2007). Los valores de VPL

obtenidos en las profundidades estudiadas (Artículo IV) resultaron ser suficientes para mantener

altas abundancias víricas en dichas aguas profundas. Los resultados de este estudio indican que

los bacteriófagos son un elemento abundante y activo del plancton en aguas oceánicas

profundas. Aunque previamente se ha observado una mayor frecuencia de la lisogénia en aguas

marinas profundas que en las de superficie (Weinbauer et al. 2003), los datos del presente estudio

no muestran una diferencia significativa en el número de casos de lisogénia detectados entre

profundidades (Artículo IV), a pesar de la más baja abundancia bacteriana en zonas más

profundas.

6.3.5. EFECTO DE LAS INFECCIONES VÍRICAS SOBRE LA DIVERSIDAD FILOGENÉTICA DE LAS

BACTERIAS

Según las hipótesis y teorías realizadas por distintos investigadores, los bacteriófagos

mantienen o incrementan la diversidad de bacterias, y evitan la dominancia de pocos taxones

bacterianos en la comunidad (Thingstad & Lignell 1997, Thingstad 2000). Aunque en algunos

estudios no se ha encontrado un efecto claro de los virus sobre la diversidad bacteriana (Hewson

& Fuhrman 2006), otros autores proporcionaron argumentos que respaldan la hipótesis de “matar

al ganador” (p.e. Hennes et al. 1995, Auguet et al. 2009, Holmfeldt 2009). Los resultados de los

estudios incluidos en esta tesis ofrecen algunas evidencias para respaldar el efecto positivo de los

fagos sobre la riqueza bacteriana. El hallazgo de que los virus juegan un papel más importante

que los protistas en la determinación de la riqueza y diversidad bacteriana durante el estudio


196 | P a g e

estacional en el Mar Mediterráneo (Artículo II), ha sido respaldado por los resultados del Artículo

III, que muestran un efecto positivo de los virus, y negativo de los protistas sobre el numero de los

taxones bacterianos. Además, se ha encontrado que en condiciones de intensa lisis vírica, los virus

han jugado un papel principal en la modificación de la composición filogenética de las

comunidades bacterianas (Artículo II).

6.3.5. PREGUNTAS ABIERTAS

Los estudios sobre la ecología de los bacteriófagos marinos se iniciaron hace algo más de dos

décadas, y han tenido como resultado la publicación de más de 1000 artículos científicos sobre

distintos aspectos de la ecología vírica. La rápida evolución de las técnicas, sobre todo

moleculares, proporcionó la oportunidad de responder a preguntas más detalladas y complejas.

Esta tesis doctoral ha contribuido al conocimiento de algunos de los procesos ecológicos más

importantes pero, al mismo tiempo, ha generado nuevas preguntas:

¿Cuál es la causa de la falta de balance entre la producción bacteriana medida y las evaluadas

pérdidas de la producción causadas por los virus y protistas?

¿En que grado la disponibilidad de nutrientes y la consecuente aparición de porinas en la pared

celular bacteriana podría ser un factor importante para que se produjeran infecciones

bacterianas en las comunidades naturales?

¿Cuáles son los factores que determinan la infección lisogénica en aguas marinas?

¿Cuál es la relación entre la lisogénia y el estado fisiológico de las bacterias en los ecosistemas

naturales?

¿Podría tener la lisogénia un efecto sobre la estructura filogenética de la comunidad bacteriana?

¿Cuál es el impacto de los bacteriófagos en distintas especies bacterianas?

Futuras investigaciones en el complejo campo de la ecología vírica contribuirán a la resolución

de estas cuestiones aún sin respuesta.


P a g e | 197

6.4. CONCLUSIONES DE LA TESIS

Los resultados obtenidos en los estudios incluidos en esta memoria de tesis permiten concluir:

1. Los bacteriófagos son una causa importante de la mortalidad bacteriana en los sistemas

naturales marinos, siendo en algunos sistemas una causa principal de pérdidas de la

producción bacteriana.

2. El estatus trófico del sistema determinaba la tasa de infecciones víricas, con tasas de

mortalidad bacteriana debida a virus mayores en el sistema más oligotrófico (Océano

Atlántico), y más bajas en un sistema más productivo (Océano Árctico).

3. El buen estado fisiológico de los hospedadores no es un factor determinante para la infección

y proliferación vírica.

4. Las correlaciones negativas entre las tasas de mortalidad debidas a bacteriófagos y protistas

en sistemas naturales sugieren que ambos depredadores compiten por la presa. Este hecho

no excluye interacciones sinérgicas entre ambos.

5. La lisogénia es más frecuente en sistemas oligotróficos que en eutróficos, y prevalece en las

comunidades bacterianas con producción más elevada.

6. La lísis vírica favorece más la riqueza y diversidad filogenética de las bacterias que la

depredación por protistas.

Annex - Methodology

Annex - Methodology

P a g e | 201

7. ANNEX. METHODOLOGY

he experimental methods used along this thesis include both standard techniques for the

evaluation of bacterial and protists’ parameters (abundances, production, and grazing),

and other more recently applied methods employed for the estimation of viral parameters.

The techniques used for the evaluation of grazing by protists, abundances of bacteria and

protists, and bacterial diversity are detailed in Papers I - V in this thesis (the “Methods”

paragraphs), as well as in a wide collection of literature cited therein.

The methods applied to estimate viral parameters, as well as the calculations of protist-

mediated mortality of bacteria are relatively new and used by rather narrow group of specialists,

and thus are presented in detail in this annex.

T


202 | P a g e

I. VIRUS ENUMERATION BY FLOW CYTOMETRY

Enumeration of viral particles using flow cytometry was based on the method description in

Marie and colleagues (1999) and Brussaard (2004).

1. SAMPLE FIXATION

Sample is fixed to 0.5 % final concentration with glutaraldehyde 25 % filtered throw 0.2 µm filter, left

for 15 - 30 minutes at 4 ºC, deepened in liquid N2 and stored at -80 ºC (Marie et al. 1999).

2. SOLUTIONS

SYBR Green I working stock (200 x diluted commercial stock solution).

The commercial stock is thawed in the dark at room temperature, and then diluted 200 times in

sterile eppendorf tubes (5 µl stock solution + 995 µl prefiltered sterile Milli-Q water, MQ). The solution

is split in aliquots and kept in –20 ºC. After thawing, the solution is centrifuged during 2 - 4 minutes.

TE buffer (10:1 mM Tris:EDTA).

(a) Preparation of 0.5 M EDTA solution (18.6 g EDTA + 100 ml MQ). While mixing with the

magnetic stirrer, pH = 8 is adjusted by slowly adding solid NaOH.

(b) Preparation of 10x TE Tris:EDTA (100 mM Tris; 10 mM EDTA), 100 ml final volume (1.21 g Tris +

2 ml 0,5 M EDTA). Add MQ up to 100 ml, and while mixing with the magnetic stirrer, pH = 8

is adjusted by slowly adding HCl (37 %).

(c) Preparation of 1x TE buffer Tris:EDTA 10:1 mM (10 ml of 10x TE and 90 ml of fresh MQ). While

mixing with the magnetic stirrer, check the pH and, if necessary, adjust pH = 8 by slowly

adding HCl (37 %) or NaOH.

(d) Autoclave the final solution. Take care to mark the level before autoclaving. Fill the missing

volume with MQ water.

(e) Store at 4 ºC. Filter the autoclaved TE by 0.2 µm each day before use.

3. CLEANING CONTROL

Prepare controls (blanks) to be subtracted from counts:

0.5 ml of prefiltered TE + 5 μl of SYBR Green I.

Annex - Methodology

P a g e | 203

Incubate 10 minutes in a 80 ºC water bath, and 5 minutes at room temperature at dark.

Run the controls at medium speed for 1 minute. Note the number of events counted in the

virus region on the cytogram, which should be below 45 events per second. The value of the

control will be subtracted from the counts.

4. SAMPLES ANALYSIS

Samples are thawed between the hands or in a water bath (~ 37 ºC), since the thawing

procedure must be fast, 1 - 2 minutes for 2 ml cryovials, and the samples should not get warm.

After thawing the samples, they should be analyzed as soon as possible (the concentration of

viruses decreases rapidly after thawing).

Samples are diluted in TE. Typically dilutions are 50 - 100x, but depending on the samples they

are often diluted 5 - 10x.

Dilutions of 50x and 100x are prepared in two steps:

Dilution I: 450 μl of TE + 50 μl of sample (10x dilution)

Dilution II: 400 μl of TE + 100 μl of “Dilution I” (50x dilution)

Dilution III: 450 μl of TE + 50 μl of “Dilution I” (100x dilution)

Add 5 μl of SYBR Green I to each tube. Stir.

Incubation of the tubes for 10 minutes in a 80 ºC water bath. After that, tubes are left for 5

minutes in the dark at room temperature to cool down.

Sample is stirred and placed in FC (avoid the drops from the injection system).

Wait until the voltage is stable (in Status window) more or less 10 seconds and only THEN press

AQUIRE. Sample will run for 60 seconds at medium speed. The mean events per second

should be around 300-600. It should never be <75 or >800. If you could not reach these

numbers, change the dilution of your sample.

Note in your printed form after each running: the speed used, the average events per second,

the total number of events and the settings used.

5. RESULTS ANALYSIS

All the data after acquisition by the Browser are stored as Cell Quest Files. On the cytograms,

adjust the regions around the populations to count. Different groups of viruses can be seen on the

cytogram (Fig. 7.1).


204 | P a g e

Fig. 7.1. An example of cytogram with marked different populations of viruses.

Annex - Methodology

P a g e | 205

II. VIRUS-MEDIATED MORTALITY OF BACTERIA

Virus reduction approach (Weinbauer et al. 2002, Wilhelm et al. 2002) was followed to

determine viral production and bacterial losses due to phages.

One liter of seawater was prefiltered by 0.8 m pore size cellulose filter (Whatman), and next

concentrated by a spiral-wound cartridge (0.22 m pore size, VIVAFlow 200), obtaining 60 ml of

bacterial concentrate. Virus free water was collected filtering one liter of seawater using a

cartridge of 100 kDa molecular mass cutoff (VIVAFlow 200). A mixture of virus free water (240 ml)

and bacterial concentrate (60 ml) was prepared and distributed into 6 sterile falcon plastic tubes

(50 ml in each one). Three of the tubes were maintained without any manipulations as controls,

while in the other three, mitomycin C (Sigma) was added (1 µg ml-1 final concentration) as

inducing agent of the lytic cycle in lysogenic bacteria. All falcon tubes were incubated in a

thermostatic chamber simulating in situ temperature and light conditions, during 12 to 24 hours.

Samples for bacterial and viral abundances were collected at time zero and each hour during the

first six hours of the experiment, and at the end of the experiment. The choice of the sampling

period was made based on previous experiments that showed that almost 100 % of the lysis occurs

during the first six hours. Samples for viral and bacterial counts were fixed with glutaraldehyde and

stored. Viral and bacterial abundances were evaluated using flow cytometry.

The number of viruses released by one bacterial cell (burst size, BS) was estimated from viral

production experiments, as in Middelboe and Lyck (2002) and Wells and Deming (2006). The

increase of viral abundance during one hour of incubation was divided by the decrease of

bacterial abundance in the same period of time.

Virus-mediated mortality of bacteria (VMM) was estimated following the model presented by

Weinbauer and colleagues (2002) and Winter and colleagues (2004). Briefly, virus increase in the

control tubes represents lytic viral production (VPL), and an increase in the mitomycin C treatments

represents total (VPT), i.e. lytic plus lysogenic, viral production (Fig. 7.2). A difference between VPT

and VPL represents lysogenic production (VPLG). As during tangential flow filtration the loss of part

of bacterial in situ standing stock occurs, VPL and VPLG were multiplied by the bacterial loss factor

(Winget et al. 2005) to compare the values between different incubations.


206 | P a g e

Fig. 7.2. An example of changes in viral abundance during the incubation of the viral

production experiment. Difference between the maximum abundance in control and in treatment

with mitomycin C (Mit. C treatm.) indicates lysogenic viral production; difference between the

minimum and maximum viral abundance in the control indicates lytic viral production.

The percentage of lysogeny in total viral production was calculated:

%VPLG = (VPLG × 100) / VPT [%].

Following the method used by Guixa-Boixereu (1997), the rate of lysed cells (RLC) was obtained

dividing VPL by burst size (BS):

RLC = VPL / BS [cells ml-1 d-1].

RLC was used to calculate VMM as a percentage of bacterial standing stock (VMMBSS):

VMMBSS = (RLC × 100) / BA0 [% d-1],

where BA0 is the initial bacterial abundance in the viral production experiment.

Annex - Methodology

P a g e | 207

Assuming that percentage of losses of bacterial standing stock due to viruses is the same in falcon

tubes and in the grazing mesocosms, VMMBSS could be used to calculate the rate of lysed bacteria

during the grazing experiment (RLCG):

RLCG = (VMMBSS × BAG) / 100 [cells ml-1 d-1],

where BAG is bacterial abundance in the grazing bottles at time zero.

Finally, using RLCG, VMM as a percentage of bacterial production (VMMBP) could be calculated:

VMMBP = (RLCG × 100) / BP [% d-1],

where BP is total bacterial production.


208 | P a g e

III. PROTIST-MEDIATED MORTALITY OF BACTERIA

Bacterial mortality due to protists was evaluated following the fluorescent labeled bacteria

(FLB) disappearance method (Sherr et al. 1987). For each grazing experiment, duplicates (1 litter of

seawater each) and one control (1 litter of virus-free water) were prepared in 2-litter

polycarbonate bottles. All bottles (control and duplicates) were inoculated with FLB at 20% of the

natural bacterial concentration. The FLB were prepared with a culture of Brevundimonas diminuta

(strain obtained from the Spanish Type Culture Collection http://www.cect.org/index2.html). B.

diminuta was heat-killed and stained with DTAF (5-([4,6 dichlorotriazin-2yl) amino]-fluorescein;

Vázquez-Domínguez et al. 1999). Bottles were incubated in a thermostatic chamber, simulating in

situ temperature and in the dark. Samples for evaluation of bacterial and FLB's abundances were

taken at the beginning and at the end of the experiments. Abundances of bacteria and FLB were

assessed by epifluorescence microscopy (Olympus BX40-102/E; a 1000X magnification). To this end,

aliquots of 20 ml sample were filtered through 0.2 μm black polycarbonate filters, and stained with

DAPI at a final concentration of 5 μg ml-1 (Sieracki et al. 1985). Natural bacteria were identified by

their blue fluorescence when excited with UV radiation, while FLB were identified by their yellow-

green fluorescence when excited with blue light.

Grazing rates of bacteria were obtained following the mathematical model #3 of Salat

and Marrasé (1994), based on the specific grazing rate (g) and specific net growth rate (a):

g = ln (BA24 × BA0-1) [d-1]

a = ln (FLB0 × FLB24-1) [d-1],

where BA0 and FLB0, and BA24 and FLB24 are bacterial and the FLB abundance at time 0 and 24h of

the grazing experiment respectively.

Net bacterial production (BPN) in the incubation bottles was obtained:

BPN = BA0 × (eat – 1) [cells ml-1 d-1],

where BA0 is bacterial abundance at the beginning of the experiment; t is time of experiment.

Annex - Methodology

P a g e | 209

Following, grazing rate (G) was calculated:

G = (g / a) × BPN [cells ml-1 d-1].

Finally, protists-mediated mortality of bacteria (PMM) was calculated as the percentage of

bacterial standing stock (BSS) and bacterial production (BP) losses:

PMMBSS = (G × 100) / BA0 [% d-1]

PMMBP = (G × 100) / BP [% d-1].

References

References

P a g e | 213

A

Abedon, S.T. 2008. Phages, ecology, evolution. In: Bacteriophage Ecology: Population Growth, Evolution, and Impact of BacterialViruses. Abedon, S.T. (ed.). New York: Cambridge University Press, pp. 1-27.

Ackermann, H.-W. 2001. Frequency of morphological phage descriptions in the year 2000. Archives of Virology 146: 843-857.

Ackermann, H.-W. 2006. Classification of bacteriophages. In: The Bacteriophages. Calendar, R. (ed.). New York: Oxford University Press, Inc., pp. 8-16.

Ackermann, H.-W., and M.S. DuBow. 1987. Viruses of prokaryotes: general properties of bacteriophages. CRC Press, Inc., Boca Raton, Florida, USA.

Almeida, M.A., M.A. Cunha, and F. Alcântara. 2001. Loss of estuarine bacteria by viral infection and predation in microcosm conditions. Microbial Ecology 42: 562-571.

Arrieta, J.M., and G.J. Herndl. 2002. Changes in bacterial b-glucosidase diversity during a coastal phytoplankton bloom. Limnology and Oceanography 47: 594-599.

Auguet, J.C., H. Montanié, H.J. Hartmann, P. Lebaron, E.O. Casamayor, P. Catala, and D. Delmas. 2009. Potential effect of freshwater virus on the structure and activity of bacterial communities in the Marennes-Oléron Bay (France). Microbial Ecology 57: 295-306.

Azam, F. 1998. Microbial control of oceanic carbon flux: the plot thickens. Science 280: 694-696.

Azam, F., T. Fenchel, J.G. Field, J.S. Gray, L.A. Meyer-Reil, and F. Thingstad. 1983. The ecological role of water-column microbes in the sea. Marine Ecology Progress Series 10: 257-263.

B

Barksdale, L., and S.B. Arden. 1974. Persisting bacteriophage infections, lysogeny, and phage conversions. Annual Review of Microbiology 28: 265-299.

Beijenrick, M.W. 1898. Concerning a contagium vivum fluidum as a cause of the spot-disease of tobacco leaves. Verh Akad Wetensch. Amsterdam, II 6: 3-21.

Bergh, Ø., K.Y. Børsheim, G. Bratbak, and M. Heldal. 1989. High abundance of viruses found in aquatic environments. Nature 340: 467-468.

Bettarel, Y., J.R. Dolan, K. Hornak, R. Lemée, M. Masin, M.-L. Pedrotti, E. Rochelle-Newall, K. Simek, and T. Sime-Ngando. 2002. Strong, weak, and missing links in a microbial community of the NW Mediterranean Sea. FEMS Microbiology Ecology 42: 451-462.

Bohannan, B.J.M., and R.E. Lenski. 1999. Effect of prey heterogeneity on the response of a model food chain to resource enrichment. The American Naturalist 153: 73-82.

Bohannan, Bohannan, B.J.M., and R.E. Lenski. 2000. The relative importance of competition and predation varies with productivity in a model community. American Naturalist 156: 329-340.


214 | P a g e

Bouvier, T., and P.A. del Giorgio. 2007. Key role of selective viral-induced mortality in determining marine bacterial community composition. Environmental Microbiology 9: 287-297.

Bratbak, G., M. Heldal, S. Norland, and T.F. Thingstad. 1990. Viruses as partners in spring bloom microbial trophodynamics. Applied Environmental Microbiology 56: 1400-1405.

Bratbak, G., M. Heldal, T.F. Thingstad, B. Riemann, and O.H. Haslund. 1992. Incorporation of viruses into the budget of microbial C-transfer: a first approach. Marine Ecology Progress Series 83: 273-280.

Brum, J.R. 2005. Concentration, production and turnover of viruses and dissolved DNA pools at Stn ALOHA, North Pacific Subtropical Gyre. Aquatic Microbial Ecology 41: 103-113.

Brussaard, C.P.D. 2004. Optimisation of procedures for counting viruses by flow cytometry. Applied Environmental Microbiology 70: 1506-1513.

C

Campbell, A.M. 2001. Bacteriphages. In: Virology. Fields, B.N. et al. (eds.) Lippincott Williams & Wilkins Publishers, pp. 522-539.

Campbell, A.M. 2006. General aspects of lysogeny. In: The Bacteriophages. Calendar, R., Abedon, S.T. (eds.). Oxford: Oxford University Press, pp. 66-73.

Choi, D.H., C.Y. Hwang, and B.C. Cho. 2003. Comparision of virus- and bacterivory-induced bacterial mortality in the eutrophic Masan Bay, Korea. Aquatic Microbial Ecology 30: 117-125.

D

d’Herelle, F.H. 1917. Sur un microbe invisible antagoniste des bacillus dysentériques. C R Hebs Seances Acad Sci Paris 165: 373-390.

d’Herelle, F. 1921. Le microbe bactériophage, agent d’immunité dans la peste et le barbone. C R Hebs Seances Acad Sci Paris 172: 99.

d’Herelle, F. 1926. The bacteriophage and its behavior. Baltimore: Williams & Wilkins.

Ducklow, H.W. 1983. Production and fate of bacteria in the oceans. BioScience 33: 494-501.

E

Edlin, G., L. Lin, and R. Kudrna. 1975. λ-lysogens of E. coli reproduce more rapidly than non-lysogens. Nature 255: 735–7.

F

Fields, B.N., Howley, P.M., Griffin, D.E., Lamb, R.A., Martin, M.A., Roizman, B., Straus, S.E., Knipe, D.M. 2001. Virology, Vol.1. Lippincott Williams & Wilkins Publishers.

References

P a g e | 215

Fraenkel-Conrat, H., and R.C. Williams. 1955. Reconstruction of active tobacco mosaic virus from its inactive protein and nucleic acid components. Proceedings of the National Academy of Sciences of the United States of America – Physical Sciences 41: 690-698.

Fraenkel-Conrat, H., B. Singer, and R.C. Williams. 1957. Infectivity of viral nucleic acid. Biochimica et Biophysica Acta 25: 87-96.

Frye, J.G., S. Porwollik, F. Blackmer, P. Cheng, and M. McClelland. 2005. Host gene expression changes and DNA amplification during temperate phage induction. Journal of Bacteriology 187: 1485-1492.

Fuhrman, J.A. 1992. Bacterioplankton roles in cycling of organic matter: the microbial food web. In: Environmental Science Research, Vol. 43: Primary Productivity and Biogeochemical Cycles in the Sea. Falkowski, P.G., and Woodhead, A.D. (eds.). New York: Plenum Press, pp. 361-384.

Fuhrman, J.A. 1999. Marine viruses and their biogeochemical and ecological effects. Nature 399: 541-548.

Fuhrman, J.A., and C.A. Suttle. 1993. Viruses in marine planktonic systems. Oceanography 6: 50-62.

Fuhrman, J.A., and R.T. Noble. 1995. Viruses and protists cause similar bacterial mortality in coastal seawater. Limnology and Oceanography 40: 1236-1242.

Fuhrman, J.A., T.D. Sleeter, C.A. Carlson, and L.M. Proctor. 1989. Dominance of bacterial biomass in the Sargasso Sea and its ecological implications. Marine Ecology Progress Series 57: 207-217.

G

Guixa-Boixereu, N. 1997. Abundància i dinàmica dels virus en ecosistemes planctònics (doctoral thesis). Barcelona: Universitat de Barcelona.

Guixa-Boixereu, N., J.I. Calderón-Paz, M. Heldal, G. Bratbak, and C. Pedrós-Alió. 1996. Viral lysis and bacterivory as prokaryotic loss factors along a salinity gradient. Aquatic Microbial Ecology 11: 215-227.

Guixa-Boixereu, N., K. Lysnes, and C. Pedrós-Alió. 1999a. Viral lysis and bacterivory during a phytoplankton bloom in a coastal water microcosm. Applied Environmental Microbiology 65: 1949-1958.

Guixa-Boixereu, N., D. Vaqué, J.M. Gasol, and C. Pedrós-Alió. 1999b. Distribution of viruses and their potential effect on bacterioplankton in an oligotrophic marine system. Aquatic Microbial Ecology 19: 205-213.

Guixa-Boixereu, N., D. Vaqué, J.M. Gasol, J. Sánchez-Cámara, and C. Pedrós-Alió. 2002. Viral distribution and activity in Antarctic waters. Deep Sea Research II 49: 827-845.

H

Hambly, E., and C. Suttle. 2005. The viriosphere, diversity, and genetic exchange within phage communities. Curentr Opinion in Microbiology 8: 444-450.

Hammond, P.M. 1995. Magnitude and distribution of biodiversity. In: Global Biodiversity Assessment. Heywood, V.H. (ed.). Cambridge, UK: Cambridge University Press, pp. 107-91.


216 | P a g e

Hayashi, T., K. Makino, M. Ohnishi, K. Kurokawa, K. Ishii, K. Yokoyama, C.-G. Han, E. Ohtsubo, K. Nakayama, T. Murata, M. Tanaka, T. Tobe, T. Iida, H. Takami, T. Honda, C. Sasakawa, N. Ogasawara, T. Yasunaga, S. Kuhara, T. Shiba, M. Hattori, and H. Shinagawa. 2001. Complete genome sequence of enterohemorrhagic Escherichia coli O157:H7 and genomic comparison with a laboratory strain K-12. DNA Research 8: 11-22.

Helton, R.R., M.T. Cottrell, D.L. Kirchman, and K.E. Wommack. 2005. Evaluation of incubation-based methods for estimating virioplankton production in estuaries. Aquatic Microbial Ecology 41: 209-219.

Hennes, K.P., C.A. Suttle, and A.M. Chan. 1995. Fluorescently labeled virus probes show that natural virus populations can control the structure of marine microbial communities. Applied and Environmental Microbiology 61: 3623-3627.

Hewson, I., and J.A. Fuhrman. 2006. Viral impacts upon marine bacterioplankton assemblage structure. Journal of the Marine Biological Association of the United Kingdom 86: 577-589.

Hofer, J.S., and R. Sommaruga. 2001. Seasonal dynamics of viruses in an alpine lake: importance of filamentous forms. Aquatic Microbial Ecology 26: 1-11.

Holmfeldt, C. 2009. Regulation of bacterioplankton community composition – emphasis on the role of viruses (doctoral thesis). Kalmar, Sweden: University of Kalmar.

Hwang, C.Y., and B.C. Cho. 2002. Virus-infected bacteria in oligotrophic open waters of the East Sea, Korea. Aquatic Microbial Ecology 30: 1-9.

I

Ivanofsky, D. 1892. Concerning the mosaic disease of the tobacco plant. St Petersburg Acad Imp Sci Bul 35: 67-70.

Ivanovics, G., V. Gaal, E. Nagy, B. Pragai, and M. Simon, Jr. 1976. Studies on negacinogeny in Bacillus cereus. II. Bacillus cereus isolates characterized by prophage-controlled production of megacinA (Phospholipase A). Acta Microbiologica Academiae Scientiarum Hungaricae 23: 283-291.

J

Jiang, S.C., and J.H. Paul. 1996. Occurrence of lysogenic bacteria in marine microbial communities as determined by prophage induction. Marine Ecology Progress Series 142: 27-38.

Jiang, S.C., and J.H. Paul. 1998. Gene transfer by transduction in the marine environment. Applied and Environmental Microbiology 64: 2780-2787.

Jiang, S.C., and J.H. Paul. 1994. Seasonal and diel abundance of viruses and occurrence of lysogeny/bacteriocinogeny in the marine environment. Marine Ecology Progress Series 104: 163-172.

Jones, S.E., R.J. Newton, and K. D. McMahon. 2009. Evidence for structuring of bacterial community composition by organic carbon source in temperate lakes. Environmental Microbiology doi: 10.1111/j.1462-2920.2009.01977.x.

References

P a g e | 217

K

Karaolis, D.K.R., S. Somara, D.R. Maneval, Jr, J.A. Johnson, and J.B. Kaper. 1999. A bacteriophage encoding a pathogenicity island, a type IV-pilus and a phage receptor in cholera bacteria. Nature 199: 375-379.

Kirchman, D.L. 2008. Introduction and overview. In: Microbial Ecology of the Oceans. Kirchman, D.L. (ed.). Hoboken, New Jersey: John Wiley & Sons, Inc. pp. 1-26.

L

Lenski, R.E. 1988. Dynamics if interactions between bacteria and virulent bacteriophage. Advances in Microbial Ecology 10: 1-44.

Lin, L., R. Bitner, and G. Edlin. 1977. Increased reproductive fitness of Escherichia coli lambda lysogens. Journal of Virology 21: 554-559.

M

Mann, N.H., A. Cook, A. Millard, S. Bailey, and M. Clokie. 2003. Bacterial photosynthesis genes in a virus. Nature 424: 741.

Maranger, R., P.A. del Giorgio, and D.F. Bird. 2002. Accumulation of damaged bacteria and viruses in lake water exposed to solar radiation. Aquatic Microbial Ecology 28: 213-227.

Marie, D., C.P.D. Brussaard, R. Thyrhaug, G. Bratbak, and D. Vaulot. 1999. Enumeration of marine viruses in culture and natural samples by flow cytometry. Applied and Environmental Microbiology 65: 45-52.

Martinez-Molina, E., and J. Olivares. 1979. Antibiotic production by Pseudomonas reptilivora as a phage conversion. Canadian Journal of Microbiology 25: 1108-1110.

Middelboe, M., and P.G. Lyck. 2002. Regeneration of dissolved organic matter by viral lysis in marine microbial communities. Aquatic Microbial Ecology 27: 187-194.

Middelboe, M., and C. Lundsgaard. 2003. Microbial activity in the Greenland Sea: role of DOC lability, mineral nutrients and temperature. Aquatic Microbial Ecology 32: 151-163.

Middelboe, M., and N.O.G. Jørgensen. 2006. Viral lysis of bacteria: an important source of dissolved amino aids and cell wall compounds. Journal of the Marine Biological Association of the United Kingdom 86: 605-612.

Middelboe, M., A. Hagström, N. Blackburn, B. Sinn, U. Fischer, N.H. Borch, J. Pinhassi, K. Simu, and M.G. Lorenz. 2001. Effects of bacteriophages on the population dynamics of four strains of pelagic marine bacteria. Microbial Ecology 42: 395-406.

Miki, T., and S. Jacquet. 2008. Complex interactions in the microbial world: unexplored key links between viruses, bacteria and protozoan grazers in aquatic environments. Aquatic Microbial Ecology 51: 195-208.

Miki, T., and N. Yamamura. 2005. Intraguild predation reduces bacterial species richness and lossens the viral loop in aqutic systems: “kill the killing the winner” hypothesis. Aquatic Microbial Ecology 40: 1-12.


218 | P a g e

Miller, R.V. 2006. Marine Phages. In: The Bacteriophages, 2nd edition. Calendar, R. (ed.). New York: Oxford University Press, Inc., pp. 534-544.

Miller, R.V., and M.J. Day. 2008. Contribution of lysogeny, pseudolysogeny, and starvation to phage ecology. In: Bacteriophage Ecology: Population Growth, Evolution, and Impact of Bacterial Viruses. Abedon, S.T. (ed.). New York: Cambridge University Press, pp. 114-143.

Moebus, K. 1983. Lytic and inhibition responses to bacteriophages among marine bacteria, with special reference to the origin of phage-host systems. Helgolander Meeresuntersuchungen36: 375-391.

Moebus, K. 1996. Marine bacteriophage reproduction under nutrient-limited growth of host bacteria. 2. Investigations with phage-host system [H3:H3/1]. Marine Ecology Progress Series 144: 13-22.

Moebus, K. 1997. Investigations of the marine lysogenic bacterium H24.2. Development of pseudolysogeny in nutrient rich broth. Marine Ecology Progress Series 148: 229-240.

Munn, C.B. 2006. Viruses as pathogens of marine organisms – from bacteria to whales. Journal of the Marine Biological Association of the United Kingdom 86: 453-467.

Murray, A.G., and G.A. Jackson. 1992. Viral dynamics: a model of the effects of size, shape, motion and abundance of single-celled planktonic organisms and other particles. Marine Ecology Progress Series 89: 103-116.

Murray, A.G., and P.M. Eldridge. 1994. Marine viral ecology: Incorporation of bacteriophage into the microbial planktonic food web paradigm. Journal of Plankton Research 16: 627-641.

N

Noble, R.T., and J.A. Fuhrman. 2000. Rapid virus production and removal as measured with fluorescently labeled viruses as tracers. Applied and Environmental Microbiology 66: 3790-3797.

O

Ohnishi, M., K. Kurokawa, and T. Hayashi. 2001. Diversification of Escherichia coli genomes: are bacteriphages the major contributors? Trends in Microbiology 9: 481-485.

P

Parada, V., E. Sintes, H.M. van Aken, M.G. Weinbauer, and G.J. Herndl. 2007. Viral abundance, decay, and diversity in the meso- and bathypelagic waters of North Atlantic. Applied and Environmental Microbiology 73: 4429-4438.

Payet, J.P., and C.A. Suttle. 2008. Physical and biological correlates of virus dynamics in the southern Beaufort Sea and Amundsen Gulf. Journal of Marine Systems 74: 933-945.

Pearson, H. 2008. “Viriophage” suggest viruses are alive. Nature 454: 677.

Pomeroy, L.R. 1974. The ocean’s food web, a changing paradigm. BioScience 24: 499-504.

Poole, K., and R.E. Hancock. 1986. Phosphate-starvation-induced outer membrane proteins of members of the families Enterobacteriaceae and Pseudomonodaceae: demonstration of

References

P a g e | 219

immunological cross-reactivity with an antiserum specific for porin protein P of Pseudomonas aeruginosa. Journal of Bacteriology 165: 987-993.

R

Riemann,L., K. Holmfeldt, and J. Titelman. 2009. Importance of viral lysis and dissolved DNA for bacterioplankton activity in a P-limited estuary, northern Baltic Sea. Microbial Ecology 57: 286-294.

Ripp, S., and R.V. Miller. 1997. The role of pseudolysogeny in bacteriophage-host interactions in a natural freshwater environment. Microbiology 143: 2065-2070.

Rohwer, F. 2003. Global phage diversity. Cell 113: 141.

Rowe, J.M., M.A. Saxton, M.T. Cittrell, J.M. DeBruyn, G.M. Berg, D.L. Kirchman, D.A. Hutchins, and S.W. Wilhelm. 2008. Constrains on viral production in the Sargasso Sea and North Atlantic. Aquatic Microbial Ecology 52: 233-244.

Russel, M., and P. Model. 2006. Filamentous phage. In: The Bacteriophages. Calendar, R.L., and S.T. Abedon (eds), Oxford: Oxford University Press, pp. 146-160.

S

Salat, J., and C. Marrasé. 1994. Exponential and linear estimations of grazing on bacteria: effects of changes in the proportion of marked cells. Marine Ecology Progress Series 104: 205-209.

Saye, D.J., and R.V. Miller. 1989. The aquatic environment: consideration of horizontal gene transmission ina diversified habitat. In: Gene transfer in the Environment. Levy, S.B., and R.V. Miller (eds.), New York: McGraw-Hill, pp. 223-259.

Saye, D.J., O. Ogunseitan, G.S. Sayler, and R.V. Miller. 1987. Potential for transduction of plasmids in a natural freshwater environment: effect of plasmid donor concentration and a natural microbial community on transduction in Pseudomonas aeruginosa. Applied and Environmental Microbiology 53: 987-995.

Schlesinger, M. 1934. Zur Frage der chemischen Zusammensetzung des Bakteriophagen. Biochem Z 273: 306-311.

Sherr E., and B. Sherr. 2008. Understanding roles of microbes in marine pelagic food webs: a brief history. In: Microbial Ecology of the Oceans, Kirchman, D.L. (ed.). Hoboken, New Jersey: John Wiley & Sons, Inc, pp: 27-44.

Sherr, B.F., E.B. Sherr, and R.D. Fallon. 1987. Use of monodispersed, fluorescently labeled bacteria to estimate in situ protozoan bacterivory. Applied and Environmental Microbiology 53: 958-965.

Sieracki, M.E., P.B. Johnson, and J.McN. Sieburth. 1985. Detection, enumeration and sizing of planctonic bacteria by image analyzed epifluorescence microscopy. Applied and Environmental Microbiology 49: 799-810.

Simon, M., and F. Azam. 1989. Protein content and protein synthesis rates of planktonic marine bacteria. Marine Ecology Progress Series 51: 201-213.

Sime-Ngando, T., and A.S. Pradeep Ram. 2005. Grazer effect on prokaryotes and viruses in a freshwater microcosm experiment. Aquatic Microbial Ecology 41: 115-124.


220 | P a g e

Spencer, R. 1955. Marine bacteriophage. Nature 175: 690-691.

Steward, G.F., D.C. Smith, and F. Azam. 1996. Abundance and production of bacteria and viruses in the Bering and Chukchi Seas. Marine Ecology Progress Series 131: 287-300.

Suttle, C.A. 1992. Inhibition of photosynthesis in phytoplankton by the submicron size fraction concentrated from seawater. Marine Ecology Progress Series 87: 105-112.

Suttle, C.A. 2005. Viruses in the sea. Nature 437: 356-361.

Š

Šimek, K., J. Pernthaler, M.G. Weinbauer, K. Hornák, J.R. Dolan, J. Nedoma, M. Mašin, and R. Amann. 2001. Changes in bacterial community composition and dynamics and viral mortality rates associated with enhanced flagellate grazing in a mesoeutrophic reservoir. Applied and Environmental Microbiology 67: 2723-2733.

T

Taira, Y., M. Uchimiya, and I. Kudo. 2009. Simultaneous estimations of viral lysis and protozoan grazing on bacterial mortality using a modified virus-dilution method. Marine Ecology Progress Series 379: 23-32.

Thingstad, T.F. 2000. Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnology and Oceanography 45: 1320-1328.

Thingstad, T.F., and R. Lignell. 1997. A theoretical approach to the question of how trophic interactions control carbon demand, growth rate, abundance and diversity. Aquatic Microbial Ecology 13: 19-27.

Thingstad, T.F., G. Bratbak, and M. Hekdal. 2008. Aquatic phage ecology. In: Bacteriophage Ecology: Population Growth, Evolution, and Impact of Bacterial Viruses. Abedon, S.T. (ed.). New York: Cambridge University Press, pp. 251-280.

Twort, F.W. 1915. An investigation on the nature of the ultramicroscopic viruses. Lancet 189: 1241-1243.

V

Vázquez-Domínguez, E., J.M. Gasol, F. Peters, and D. Vaqué. 1999. Measuring the grazing losses of picoplankton: methodological improvements to the use of fluorescently labeled tracers combined with flow cytometry. Aquatic Microbial Ecology 20: 110-128.

W

Waldor, M., and J. Mekalanos. 1996. Lysogenic conversion by a filamentous phage encoding cholera toxin. Science 272: 1910-1914.

Weinbauer, M.G. 2004. Ecology of prokaryotic viruses. FEMS Microbiology Reviews 28: 127-181.

References

P a g e | 221

Weinbauer, M.G., and M.G. Höfle. 1998. Size-specific mortality of lake bacterioplankton by natural virus communities. Aquatic Microbial Ecology 15: 103-113.

Weinbauer, M.G., and C.A. Suttle. 1999. Lysogeny and prophage induction in coastal and offshore bacterial communities. Aquatic Microbial Ecology 18: 217-225.

Weinbauer, M.G., and F. Rassoulzadegan. 2004. Are viruses driving microbial diversification and diversity? Environmental Microbiology 6: 1-11.

Weinbauer, M.G., C. Winter, and M.G. Höfle. 2002. Reconsidering transmission electron microscopy based estimates of viral infection of bacterioplankton using conversion factors derived from natural communities. Aquatic Microbial Ecology 27: 103-110.

Weinbauer, M.G., I. Brettar, and M.G. Höfle. 2003. Lysogeny and virus-induced mortality of bacterioplankton in surface, deep, and anoxic marine waters. Limnology and Oceanography 48: 1457-1465.

Weinbauer, M.G., K. Hornák, J. Jezbera, J. Nedoma, J.R. Dolan, and K. Šimek. 2007. Synergistic and antagonistic effects of viral lysis and protistan grazing on bacterial biomass, production and diversity. Environmental Microbiology 9: 777-788.

Wells, L.E., and J.W. Deming. 2006. Significance of bacterivory and viral lysis in bottom waters of Franklin Bay, Canadian Arctic, during winter. Aquatic Microbial Ecology 43: 209-221.

Wilcox, R.M., and J.A. Fuhrman. 1994. Bacterial-viruses in coastal seawater: lytic rather than lysogenic production. Marine Ecology Progress series 114: 35-45.

Wilhelm, S.W., and C.A. Suttle. 1999. Viruses and nutrient cycles in the sea. BioScience 49: 781-788.

Wilhelm, S.W., and C.A. Suttle. 2000. Viruses as regulators of nutrient cycles in aquatic environments. In: Viruses as Regulators of Microbial Systems, Microbial Biosystems: New Frontiers, Proceedings of the 8th International Symposium on Microbial Ecology. Bell, C.R., M. Brylinsky, and P. Johnson-Green (eds). Halifax, Canada: Atlantic Canada Society for Microbial Ecology. pp. 551-556.

Wilhelm, S. W., M. G. Weinbauer, C. A. Suttle, and W. H. Jeffrey. 1998. The role of sunlight in the removal and repair of viruses in the sea. Limnology and Oceanography 43: 586-592.

Wilhelm, S.W., S.M. Brigden, and C.A. Suttle. 2002. A dilution technique for the direct measurement of viral production: a comparison in stratified and tidally mixed coastal waters. Microbial Ecology 43: 168-173.

Williamson, S.J., L.A. Houchin, L. McDaniel, and J.H. Paul. 2002. Seasonal variation in lysogeny as depicted by prophage induction in Tampa Bay, Florida. Applied and Environmental Microbiology 68: 4307-4314.

Winget, D.M., K.E. Williamson, R.R. Helton, and K.E. Wommack. 2005. Tangential flow diafiltration: an improved technique for estimation of virioplankton production. Aquatic Microbial Ecology 41: 221-232.

Winter, C., G.J. Herndl, and M.G. Weinbauer. 2004. Diel cycles in viral infection of bacterioplankton in the North Sea. Aquatic Microbial Ecology 35: 207-216.

Winter, C., M. M. Moeseneder, G. J. Herndl, and M. G. Weinbauer. 2008. Relationship of geographic distance, depth, temperature, and viruses with prokaryotic communities in the eastern tropical Atlantic Ocean. Microbial Ecology 56: 383-389.

Wommack, K.E., and R.R. Colwell. 2000. Virioplankton: viruses in aquatic ecosystems. Microbiology and Molecular Biology Reviews 64: 69-114.


222 | P a g e

Wommack, K.E., J. Ravel, R.T. Hill, and R.R. Colwell. 1999 a. Hybridization analysis of Chesapeake Bay virioplankton. Applied and Environmental Microbiology 65: 241-250.

Wommack, K.E., J. Ravel, R.T. Hill, J. Chun, and R.R. Colwell. 1999 b. Population dynamics of Chesapeake Bay virioplankton: total-community analysis by pulse-field gel electrophoresis. Applied and Environmental Microbiology 65: 231-240.

This thesis has been funded by the Spanish Ministry of Science and Innovation

through a PhD studentship under the program “Formación de Profesorado

Universitario (FPU)”.

digital.csic.esdigital.csic.es/bitstream/10261/101768/4/tesis_boras.pdf · institut de ciències...

Documents