corno, gianluca, et al. bacterial diversity and morphology...

Bacterial diversity and morphology in deep ultraoligotrophic Andean lakes: Role of

UVR on vertical distribution

Gianluca Corno,a,* Beatriz E. Modenutti,b Cristiana Callieri,a Esteban G. Balseiro,b Roberto Bertoni,a

and Emanuele Caravatia

a CNR–Institute of Ecosystem Study, Verbania Pallanza, ItalybLaboratorio de Limnologıa, INIBIOMA, CONICET–UNComahue, Bariloche, Argentina

Abstract

We investigated the community composition and the relative size–shape distribution of bacterial assemblagesin deep transparent lakes, where bacteria may be affected by harmful ultraviolet radiation (UVR). Summersamples of the euphotic zones of nine ultraoligotrophic lakes located in different drainage basins between 40u279and 42u499S in the North Andean Patagonia were analyzed for relative diversity by denaturating gradient gelelectrophoresis and the bacterial morphology (free-living cocci and rods vs. filamentous forms) in relation toenvironmental factors (UVR, photosynthetically active radiation, temperature, nutrients, nanoflagellates, andciliate abundances). The overall bacterial community composition was similar in all lakes and over depth in eachlake. In contrast, the relative proportion of filaments to total bacterial biovolume was higher in the upper layers,which have higher UVR intensities (305–340 nm). Lakes with lower Kd305 (diffuse extinction coefficient, 305 nm)showed a greater proportion of filaments. Filament mean length in the upper layers was also significantly greaterthan at deeper levels. There was a significant correlation between UVR and filament proportion, whereas neithernanopredator abundances (nanoflagellates and ciliates), nor predation pressure, nor resources (dissolvednutrients) could explain the high proportion of filamentation in the near-surface layers. We propose that UVRplays a decisive role stimulating bacterial filamentation, particularly in lakes with high UVR penetration.

Bacteria are key organisms in biogeochemical andmetabolic processes, playing an important role in thebiodiversity of lakes (Logue and Lindstrom 2008). Thedegree of isolation and thereby the rate of exchange of cellsand genes between communities should have a greatimportance in shaping local microbial communities (Curtisand Sloan 2004). On the basis of the importance ofallochthonous carbon for lake ecosystem function (DelGiorgio and Peters 1994; Pace et al. 2004), lakes should beconsidered as part of a larger unit, i.e., the drainage basin(Soranno et al. 1999). In that sense, the distribution ofbacterioplankton taxa among different freshwater habitatscan be associated with catchment characteristics andhydrology (Lindstrom and Bergstrom 2004). The datacurrently available suggest that the bacterial communitycomposition in freshwaters is controlled by dispersal(biogeography) or environment (habitat quality), or both,and that these controls operate across a spatial range fromlocal to regional (Crump et al. 2007; Langenheder andRagnarsson 2007).

Among environmental factors, the solar radiation andparticularly ultraviolet radiation (UVR) have strong effectson the production, activity, and abundance of bacterio-plankton (Helbling et al. 1995; Sommaruga et al. 1997;Tranvik and Bertilsson 2001). Up to now few studies havebeen published that show evidence of the effects of UVR onbacterial community composition and morphologicaldistribution. In marine environments it was observed thatnot all bacteria assemblages were equally susceptible toUVR (Joux et al. 1999; Arrieta et al. 2000). In humic lakes,a slight change in taxonomic composition of bacterial

communities was demonstrated, in response to exposure tonatural solar radiation along a depth gradient (Langenhe-der et al. 2006).

In general, bacterial community composition andbacterial biomass are controlled by the interplay betweenseveral factors, such as organic and inorganic nutrients,predation by nanoflagellates (NF), and viral lysis (Thing-stadt 2000; Weinbauer et al. 2007; Corno and Jurgens2008). The dominant role of size-selective bacterivory byNF in shaping the size distribution and the composition ofbacterial assemblages has been observed in many studiesboth in marine and freshwaters (see Jurgens and Matz 2002for a review). Under high grazing pressure by NF, bacterialcell size is a highly adaptive trait allowing reducedmortality by the development of grazing-resistant mor-phologies such as filaments and aggregates as well as verysmall free-living cells, commonly defined as ultramicrobac-teria (a complete review on the field is given by Pernthaler2005). Drastic shifts in bacterial community compositionand morphological distribution have been observed inaquatic bacterial communities when exposed to elevatedgrazing pressure by NF, in relation to the productivity ofthe systems, and thus, to the amount of available nutrients:with high carbon limitation, filamentation is reduced,whereas the development of ultramicrobacteria is favored(Matz and Jurgens 2003; Boenigk et al. 2004). Raising theamount of organic carbon available (or shifting the C : Pratio), a small number of extremely plastic bacterial strainsdevelops in a large proportion of large morphotypes (e.g.,filaments, long chains of large rods), mostly inedible forNF, provoking a reduction in the overall diversity of thebacterial community (Corno and Jurgens 2008). Althoughfilamentation is a successful resistance strategy against NF* Corresponding author: [email protected]

Limnol. Oceanogr., 54(4), 2009, 1098–1112

E 2009, by the American Society of Limnology and Oceanography, Inc.

1098

grazing, it may be a response to other environmentalfactors. Strong variation in nutrient availability stimulatedfilament formation in strains belonging to the Cytophaga–Flavobacterium–Bacteroides (CFB) group (Hahn and Hofle2001). In contrast, a positive indirect effect of predatoryactivity by NF in extremely oligotrophic systems resultsfrom the release of a large amount of soluble exudatesduring grazing, support for bacterial production, and thusfor promoting the maintenance of higher bacterial diversityin systems highly limited by nutrients (Corno et al. 2008;Corno and Jurgens 2008). In natural environments thepresence of larger predators (large ciliates, zooplankton)may reduce (both by direct predation on larger bacteriaand by predation on NF) the effectiveness of filamentationas antipredation strategy in systems at high trophic stateswhere these organisms are present in large numbers(Pernthaler 2005). Moreover, early studies have observedthat exposure to UVR provokes a failure of cell division insensitive strains of Escherichia coli, resulting in theformation of greatly elongated filamentous cells (Kalleand Sivasubramanian 1967; Witkin 1967). More recently,morphological changes of the cyanobacterium Arthrospira(Spirulina) platensis exposed to solar UVR have also beenrecorded (Wu et al. 2005).

We studied nine Andean lakes characterized by hightransparency and high UVR penetration (Morris et al.1995; Balseiro et al. 2007) where planktonic organisms,living in the epilimnetic levels of the euphotic zone, areexposed to damaging ultraviolet wavelengths (UVA andUVB) (Alonso et al. 2004; Modenutti et al. 2004, 2005). In

these lakes, mixotrophic flagellates and ciliates dominatethe bacterivorous fraction though in low abundance(Modenutti and Balseiro 2002; Modenutti et al. 2005,2008). Furthermore, all these lakes exhibit high C : P ratios(Balseiro et al. 2008) in accordance with the light–nutrienthypothesis (Sterner et al. 1997). Under these conditions,bacterioplankton are exposed both to broad levels of UVRand to poor nutrient availability due to unbalancedelemental ratio. Thus, our hypothesis was that the differentlayers within the euphotic zone of the Andean lakes wouldhave a distinctive bacterial community composition (diver-sity and morphology) because of its differential UVRtransparency and penetration. The main points addressedin this paper are: (1) to describe the bacterioplanktoncommunity composition in three Andean basins (twoPacific and one Atlantic watershed), (2) to determinewhether environmental factors predict community compo-sition, and (3) to investigate the potential effect of UVR onbacterial diversity and morphological distribution.

Methods

Study area—The studied lakes are located at 40u279 and42u499S in the North Andean Patagonian region (Argen-tina) (Fig. 1, Table 1) corresponding to the glacial lakesdistrict of the southern Andes (Iriondo 1989). The climateis cool (mean annual temperature 8.7uC) with a predom-inance of strong westerly winds, and annual precipitationof 1500 mm (Paruelo et al. 1998). The vegetation corre-sponds to the Andean–Patagonic temperate forest repre-

Fig. 1. Map with the location (A) of the Andean studied lakes, and (B) the diagram of thebasins and watersheds.

Role of UVR on bacterial communities 1099

sented in the lakes’ shores mainly by Nothofagus dombeyi(Mirb.) Blume and Austrocedrus chilensis (D.Don) Florin etBoutleje. The area is within the boundaries of two nationalparks: Nahuel Huapi and Los Alerces. The area ischaracterized by a profuse hydrographic system includinglarge and deep lakes (Zmax $ 150 m). The nine studiedlakes are Lakes Espejo, Correntoso, Nahuel Huapi, andGutierrez (Nahuel Huapi Basin) of the Atlantic watershed,Lakes Mascardi Catedral and Mascardi Tronador (MansoBasin) and Lakes Rivadavia, Menendez, and Futalaufquen(Futaleufu Basin), of the Pacific watershed (Fig. 1). Thetwo arms of Lake Mascardi (Catedral and Tronador) wereconsidered two lakes since the two arms have a differenthydrology due to the three headwater glaciers (Manso,Overo, and Alerce glaciers) that drain only into MascardiTronador (Modenutti et al. 2000).

The nine lakes exhibit a warm monomictic thermalregime, with thermal stratification during late spring andsummer. Penetration of photosynthetically active radiation(PAR) and UVR is high and dissolved organic carbonconcentration is low (Morris et al. 1995). Their trophicstatus was described as ultraoligotrophic (Modenutti et al.1998; Perez et al. 2007).

Sampling—Sampling was carried out during summer(15–22 January 2008) in a central sampling point located atthe deepest part of each lake. Vertical profiles (0 to 70 m)of UVR bands (305, 320, 340, and 380 nm), PAR (400–700 nm), and temperature were measured with a PUV 500Bsubmersible radiometer (Biospherical Instruments). Oxy-gen concentration was measured with a YSI (model 85)oxygen meter in discrete samples each 5-m depth. Watersamples of 2 liters were collected with a Ruttner bottlewithin the euphotic zone (50%, 10%, and 1% of surfacePAR irradiance); in Lake Nahuel Huapi a sample at 0.1%of surface PAR was also taken and in Lake Menendez weonly sampled at 0.5 m. In the field, a volume of 300 mLwas filtered through GF/F filter for chlorophyll a (Chl a)determination; filters were stored in the dark at 4uC untilreaching the laboratory. A volume of 100 mL was sampledfor NF enumeration and 60 mL for total bacteriaquantification; these samples were preserved by addingfiltered (0.2 mm) formaldehyde solution (2% v/v). A volumeof 150 mL was preserved with acid Lugol solution for

ciliate enumeration. The remaining sampled volume(,1.4 liters) was transferred to acid-washed polypropylenecontainers. Containers were kept in darkness and thermallyinsulated and were carried immediately to the laboratory todetermine nutrient concentrations. All samplings werecarried out in triplicate, at midday, 1 h before astronomicnoon.

Laboratory determinations—Total phosphorus (TP) andtotal nitrogen (TN) were determined directly on unfilteredlake water; total dissolved phosphorus (TDP) was deter-mined on GF/F filtered lake water. For TP and TDPdeterminations the samples were digested with potassiumpersulfate at 125uC at 1500 hPa for 1 h and the concen-trations were analyzed through the ascorbate-reducedmolybdenum method. TN was determined followingValderrama (1981). Dissolved organic carbon (DOC)concentration was measured on 50 mL of filtered lakewater (GF/F precombusted filters) with a Shimadzuanalyzer (TOC 5000A). The Chl a was extracted in hotethanol following Nusch (1980) and measured with a 10-AU fluorometer (Turner Design).

Total bacteria and NF enumerations were performed bystaining with fluorochrome 49,6-diamidino-2-phenylindole(final concentration 2% v/v, Porter and Feig 1980).Bacteria counting was performed on polycarbonate blackmembrane filters (0.2-mm pore size) at 12503 magnificationin an Olympus BX50 epifluorescence microscope using UVlight. Different bacteria morphologies were discriminated(free-living cocci and rods vs. filamentous forms). Whenbacteria developed filamentous forms .7 mm, they wererecorded as filaments that could be considered resistant topredation by NF (Corno 2006). A minimum of 500 totalbacteria per sample was counted and processed with animage analysis system (Image ProPlus; Media Cybernetics).

NF enumeration was done on 1-mm polycarbonate blackmembrane filters. Cells were counted by epifluorescencemicroscopy at 12503 magnification, using both UV andblue light, to distinguish heterotrophic nanoflagellates(HNF) from mixotrophic nanoflagellates (MxFl). Enumer-ation of the ciliate Ophrydium naumanni, a well-knownpredator of bacteria in these lakes (Modenutti and Balseiro2002; Modenutti et al. 2008), was performed following theUtermohl technique with an inverted microscope (Olympus

Table 1. Location and morphometric characteristics of the nine ultraoligotrophic Andean lakes (Argentina) studied. Horizontal barsseparate the different basins.

Lakes LocationAltitude (m above

sea level) Area (km2) Zmax (m) Watershed

Espejo 40u409000S 71u429000W 800 30.00 245 AtlanticCorrentoso 40u269420S 71u329070W 764 19.50 .100Nahuel Huapi 40u269540S 71u339320W 764 557.00 464Gutierrez 41u079570S 71u189310W 785 16.40 111

Mascardi Catedral 41u079310S 71u179240W 750 39.20 118 PacificMascardi Tronador 41u079050S 71u209340W 750 39.20 218

Menendez 42u409000S 71u509000W 600 55.7 287 PacificRivadavia 42u369390S 71u389580W 560 21.70 147Futalaufquen 42u499080S 71u429440W 520 44.60 168

1100 Corno et al.

IX70) using 50-mL Utermohl chambers and was carriedout by scanning the entire surface of the chamber at 2003magnification.

DNA extraction and amplification—Water samples(20 mL) from each lake were prefiltered onto 20-mm meshnet and then on NucleoporeH filters, 0.2-mm pore size,25 mm diameter. Nucleopore filters were then collected andstored at 220uC in phosphate-buffered saline, and trans-ported to the Consiglio Nazionale delle Ricerche–Instituteof Ecosystem Study laboratories. Deoxyribonucleic acid(DNA) was then extracted from the filters using Ultra-cleanH Soil Isolation kits (MoBio) following the producerprotocol for high yields and stored in 30 mL of Tris–ethylenediamine tetra-acetic acid (EDTA) buffer at 220uCuntil use.

Aliquots (1–2 mL) of extracted DNA were applied to thefollowing mix to give a 50-mL polymerase chain reaction(PCR) mixture: 25 mL of GoTaqH Mastermix (Promega),0.5 mL of bovine serum albumin, 2 mL of the forwardprimer 357F-GC (a 40-base pair [bp] GC clamp was addedin position 59), and 2 mL of an equal mixture of the reverseprimers 907R and 907Rm, all 10 mmol L21 concentrated.These primers were selected according to Sanchez et al.(2007) for being the most sensitive and less subjected toerrors in estimating water bacteria diversity. PCR details:initial denaturation was at 96uC for 5 min, amplificationwas carried out using 10 touch-down cycles includingdenaturation at 94uC for 45 s, annealing at 65uC for 45 sdecreased by 1uC for each cycle, and extension at 72uC for2 min. This was followed by additional 20 cycles ofdenaturation at 94uC for 45 s, annealing at 55uC for 45 s,extension at 72uC for 2 min, and a final elongation step at72uC for 5 min (Muyzer et al. 1998). PCR products fromreplicates were pooled and then purified with QIAquick PCRpurification kits (Qiagen). Finally, they were stored at 220uCuntil denaturing gradient gel electrophoresis (DGGE) runs.

Analysis by DGGE—DGGE was carried out as using aDcodeTM instrument and gradient former model 475according to the manufacturer’s instructions (Bio-Rad).The denaturing gradient was formed with two 9%acrylamide (acrylamide-bis 37.5 : 1) stock solutions (Pro-mega) in 13 Tris acetate–EDTA (TAE: 20 mmol L21 Tris,10 mmol L21 acetate, 0.5 mol L21 EDTA, pH 7.4). The100% denaturant solution contained 40% formamide and7 mol L21 urea. Perpendicular DGGE analyses werecarried out to select appropriate electrophoresis conditions:the gels were made then with denaturing gradients rangingfrom 35% to 60%. A volume of 13 mL of purified PCRproducts was mixed with 3 mL of loading dye beforeloading. Gels were run in 13 TAE at 60uC for 16 h at100 V and stained with SYBR Green I nucleic acid gel stain(Cambrex) for 30 min. Digitized DGGE images from thehigh-resolution UV illuminator Gel DOC XR system (Bio-Rad) were analyzed with Quantity One software (Bio-Rad).Bands occupying the same position in the different lanes ofthe gels were identified.

A density profile through each lane was established, thebands detected (regarded as operational taxonomic units

[OTUs]), and the relative contribution of each band to thetotal band signal was calculated in the lane after applying arolling disk as background subtraction. Bands with arelative intensity of ,0.5% of the total intensity of the lanewere disregarded. Remaining bands, considered as validbands, were used to assess the relative gross diversity ofeach sample. A distance matrix of the lanes was calculatedusing the Bray–Curtis coefficient and a cluster (unweightedpair-group mean average [UPGMA]) was inferred using thesoftware MEGA4 (Molecular Evolutionary Genetic Anal-ysis).

Sequencing of selected DGGE bands—From the DGGEgels, dominant bands were excised with a sterile scalpel,placed in 50 mL of sterile water, and incubated at 4uCovernight for diffusion of DNA into the water. Onemicroliter of each solution containing excised bands wasreamplified using the same primers and the same PCRconditions as described above except for using 357F primerwithout GC-clamp. PCR products were purified usingQIAquick PCR purification kit (Qiagen). Part (3 mL) of thePCR product was run on a 0.8% high-quality agarose gel tocheck the purity and confirm the melting behavior of theexcised band. The inserts (550 bp) recognized as pure weresequenced by Macrogen (Korea) using primer 357F.

Data analysis—Nutrient concentrations, N : P and C : Pratios, and bacteria abundances and biovolume wereanalyzed through two-way analysis of variance (ANOVA)(lakes and depths) and, when significant, an overallpairwise multiple-comparison Tukey test was applied.Nucleotide sequences were compared with sequences inthe GenBank database (Benson et al. 1999) using theBLAST program. Sequences from GenBank with highsimilarity to the excised DGGE band sequences were usedas reference sequences for the design of the phylogenetictree. The partial 16S rDNA sequences from this study aswell as reference sequences were optimally aligned (Fengand Doolittle 1987) before tree construction. A phyloge-netic tree was constructed by using the UPGMA methodwith BioEdit Sequence Alignment Editor 7.0 software.Nucleotide sequences obtained in this study have beendeposited in the GenBank database under accessionnos. FJ238449 to FJ238462.

Bacterial community composition (DGGE bands) be-longing to different lakes and depths were analyzedthrough principal component analysis (PCA). Then,bacterial communities (lakes and depths) were analyzedthrough a multivariate redundancy analysis (RDA) usingan environmental matrix composed of ecological factors:diffuse attenuation coefficients of PAR (400–700 nm) andUVR (305, 320, 340, and 380 nm); phosphorus, nitrogen,DOC, and Chl a concentrations; total prokaryote, NF, andO. naumanni abundances. These analyses were carried outwith CANOCO 4.5.

The relative grazing pressure by NF on bacterialassemblages was estimated through the mean prey–predator ratio (Fenchel 1986) as the ratio between thenumber of edible bacterial cells ,7 mm and NF cells.Bacteria morphology (percentage of filaments) as depen-


dent variable and UV irradiance at each sampled depth,NF number, dissolved C : P atomic ratios, and predationpressure by NF as independent variables were analyzedthrough forward stepwise regression with SigmaStat 3.5software. The surface sample of Lake Menendez was onlyconsidered for the DGGE analysis.

Results

Lake features—All the lakes were found to be thermallystratified, with a thermocline varying between 32 m (LakeNahuel Huapi) and 6 m (Lake Espejo) (Table 2). Dissolvedoxygen was always at saturation (100%) levels. In all thelakes, the 1% of surface PAR (Z1% PAR) exceeded thethermocline depth (Ztherm) (Table 2); thus the epilimnion,the metalimnion, and the upper part of the hypolimnionwere included in the euphotic zone. In particular, lakesbelonging to the Nahuel Huapi Basin of the Atlanticwatershed exhibited a higher UVR transparency (see Kd

values in Table 2), whereas lakes belonging to the Pacificwatersheds were comparatively less transparent (Table 2)(t-test t 5 24.172, p 5 0.0058 for the 305-nm band; t 524.446, p 5 0.0043 for the 340-nm band). Lakes Espejo,Correntoso, and Gutierrez from the Nahuel Huapi Basinpresented an epilimnion entirely exposed to the band of305 nm (UVB), as this band reached the thermocline. LakeMascardi Catedral was the most transparent among thelakes of the Pacific watershed, and presented an epilimnionalmost entirely exposed to the 340-nm band (UVA)(Table 2, Z1% 340). In the studied area the summer is adry and sunny season. According to data collected by theGUV 510 (Biospherical Instruments) installed at theUniversity of Comahue, the mean daily integrated irradi-ance for each wavelength (305, 320, 340, and 380 nm andPAR) for the whole month (January 2008) resulted in 85%to 90% of the maximum value for the daily integratedirradiance of each band registered for the whole month.Although light data used in all the analyses were obtainedat the moment of sampling, we considered, on the basis ofthe GUV data mentioned above, that they represent meanlight conditions during the sampling interval in January2008.

The lakes showed very similar nutrient features inde-pendent of their lake basin. TP concentrations were verylow in all lakes, ranging from 2.6 to 6.1 mg L21, whereasTN showed very little variation (55.0 mg L21 in LakeCorrentoso to 107.3 mg L21 in Lake Gutierrez). TN andTP concentrations did not show significant differencesbetween lakes and in the water column (two-way ANOVA,p . 0.05). The N : P ratios (total atomic) were always above25, reaching values higher than 70. Significant differencesbetween lakes were found (two-way ANOVA F7,14 5 9.14,p 5 0.0003), and Lake Espejo showed the highest value (aposteriori Tukey test p , 0.03 for Lake Espejo with otherlakes) (Table 2). DOC presented also minor variationsamong lakes and values were always below 0.6 mg L21;nevertheless these variations seemed to be enough toaccount for the differences in the two studied UVRabsorbance bands. The C : P ratio (dissolved) varied from991 in Lake Gutierrez to 409 in Lake Mascardi Tronador.

Table2.

Th

erm

ocl

ine

dep

th(Z

therm

)a

nd

dep

tho

f1%

surf

ace

irra

dia

nce

s(Z

1%

)o

fU

VR

(30

5a

nd

34

0n

m)

an

dP

AR

,li

gh

ta

tten

ua

tio

nco

effi

cien

ts(K

d)

of

UV

R(3

05

an

d3

40

nm

),a

nd

PA

R,

tota

lp

ho

sph

oru

s(T

P),

an

dto

tal

nit

rog

en(T

N)

con

cen

trati

on

s,a

nd

C:P

ato

mic

rati

oo

fth

ed

isso

lved

fra

ctio

n(D

OC

an

dT

DP

)o

fth

en

ine

lak

es.

TP

an

dT

Nco

nce

ntr

ati

on

sa

nd

C:P

an

dN

:Pra

tio

sa

reg

iven

as

mea

n6

SE

of

the

thre

est

ud

ied

dep

ths.

Ho

rizo

nta

lb

ars

sep

ara

teth

ed

iffe

ren

tb

asi

ns.

n.a

.,d

ata

no

ta

va

ila

ble

.

La

kes

Zth

erm

(m)

Z1%

30

5

(m)

Z1%

34

0

(m)

Z1%

PA

R

(m)

Kd

30

5

(m2

1)

Kd

34

0

(m2

1)

Kd

PA

R

(m2

1)

TP

(mg

L2

1)

TN

(mg

L2

1)

C:P

dis

solv

eda

tom

icra

tio

N:P

tota

la

tom

icra

tio

Esp

ejo

61

11

34

40

.42

0.3

50

.10

2.6

60

.19

80

.46

3.2

45

286

54

706

6.9

Co

rren

toso

13

10

15

46

0.4

60

.31

0.1

03

.66

0.3

45

5.4

61

.70

57

56

31

346

2.8

Na

hu

elH

ua

pi

32

11

17

45

0.4

20

.27

0.1

04

.96

0.2

46

7.3

61

1.5

39

046

18

13

26

5.1

Gu

tier

rez

11

10

14

33

0.4

60

.33

0.1

44

.26

0.2

11

07

.36

10

.16

99

16

10

15

76

7.0

Ma

sca

rdi

Ca

ted

ral

12

71

13

10

.66

0.4

20

.15

3.4

60

.46

70

.46

7.3

86

106

66

496

10

.7M

asc

ard

iT

ron

ad

or

85

71

30

.92

0.6

60

.35

6.1

62

.08

98

.26

20

.22

40

96

53

96

4.4

Men

end

ezn

.a.

n.a

.n

.a.

n.a

.n

.a.

n.a

.n

.a.

2.9

60

.34

44

.06

0.2

0n

.a.

n.a

.R

iva

da

via

18

47

26

1.1

50

.66

0.1

85

.66

0.4

76

8.0

67

.20

60

06

11

286

4.7

Fu

tala

ufq

uen

25

68

33

0.7

70

.58

0.1

44

.96

0.4

29

2.6

61

4.5

75

946

38

436

9.7

1102 Corno et al.

We found significant differences in C : P (two-way ANOVAF7,14 5 7.46, p , 0.0008) between lakes: Lake Gutierrezand Nahuel Huapi exhibited higher values than the otherlakes (a posteriori Tukey test p , 0.05 for Gutierrez with allother lakes except Nahuel Huapi and for Nahuel Huapiwith Mascardi Tronador and Futalaufquen).

Bacterial community composition (DGGE bands)—Thegross bacterial diversity was determined by DGGE profilesfor different depths and for all examined lakes. Samplesbelonging to different depths within lakes showed similaroverall richness and species composition (two-way AN-OVA p . 0.05) (Fig. 2A,B). In two lakes (Rivadavia andFutalaufquen) bacterial richness was exactly the same atthe three depths, but also in all other lakes the changes werelimited to one or two different OTUs per depth. On averagethe measured bacterial richness reached 14 significantOTUs per sample. However, significant differences weredetected in richness between lakes (two-way ANOVA F7,14

5 53.76, p , 0.0001): the highest richness was measured inLake Mascardi Catedral, with about 20 OTUs per sample,whereas Lake Futalaufquen showed the lowest richness,with 11 OTUs (the unique surface sample of LakeMenendez with only 8 OTUs was not considered forstatistical analysis). Lake Rivadavia and Lake Espejoexhibited relatively rich diversity, with an average of 17and 15 OTUs, respectively.

From the sequencing of DGGE bands excised from gelsit was possible to determine between 40% and 80% of thetotal diversity estimated by the fingerprints (Fig. 3). Grosstaxonomic composition of bacterial community was relatedto the lake and independent from sampling depth. Allmajor bacterial groups except c-Proteobacteria weredetected by our analyses. Actinobacteria were present inall lakes and sampling depths, with percentages on totalband signal usually about 10% except in Lake MascardiTronador, where they reached 18%. Two main actinobac-terial OTUs were recognized (Fig. 3), from DGGE bands 1(common in all samples) and 3 (present only in lakesNahuel Huapi, Espejo, Mascardi Tronador, and Menen-dez). The limited size of the excised fragment (about550 bp) did not allow further assessment on speciescomposition and its relative importance within the studiedcommunities.

Only one b-Proteobacteria sequence was found (band 6),but it was common to all lakes, accounting always for morethan 10% (about 30% in lakes Futalaufquen and Menen-dez). Also a single a-Proteobacteria was determined (band9), present in all lakes but Espejo and Nahuel Huapi, withproportions always lower than 10%. The group CFB wasthe most diverse, represented by 4 OTUs (bands 2, 4, 5, and10). Bands 2, 4, and 5 were common and abundant in alllakes, whereas band 10 was characteristic of Lakes NahuelHuapi and Gutierrez. The overall proportion of CFBvaried from less than 20% in Lake Futalaufquen to morethan 50% in Lake Correntoso. Band 8 belonged toCyanobacteria and was common in Nahuel Huapi (.10%at all depths) and occurred in Lake Correntoso (about 3–8%) and Mascardi Catedral (4%). Finally, a Verrucomi-crobia strain (band 7) was determined (always around 10%)

in all lakes except Mascardi Tronador, Menendez, andCorrentoso (50% and 10% PAR).

Bacterial community composition (multivariate analysis)—The PCA analysis on OTUs showed a similar bacterialcomposition of the euphotic zone for most of the lakes(Fig. 4A). Total variance explained by the first three axes was62.1% (first 33%, second 15.6%, and third 13.2%). Lakeswere grouped without any pattern, indicating that neitherwatershed (Pacific or Atlantic) nor basin (Nahuel Huapi,Futaleufu, and Manso) is central in determining the grossbacterial community composition (Fig. 4A).

The similarity of the bacterial communities belonging todifferent lakes and depths was confirmed by applying aRDA multivariate analysis, and biological variables in theenvironmental matrix (Fig. 4B). The three axes explained54.8% of total variance among RDA species. The variablesof higher contribution in the first axis (27.9% of variance)were those related to light climate (UVR all wavelengths),the second axis (15.8%) was mainly explained by PAR,DOC, and Chl a, and the third axis (11.1%) by P.

Microbial abundances, biovolumes, and cell morphologydistribution—Bacterial abundances ranged between 0.42 60.02 3 106 cells mL21 in Lake Espejo at 50% PAR and 1.466 0.001 3 106 cells mL21 in Lake Nahuel Huapi at 10%PAR (Table 3). Statistical analysis of bacterial abundancebetween lakes or depth showed significant differences in thetwo factors (two-way ANOVA, lakes F7,14 5 7.41, p 50.0008 and depths F2,14 5 4.99, p 5 0.023). A significantreduction of bacterial abundance in the upper level (50%PAR) of water column was observed (a posteriori Tukeytest p 5 0.018). Total biovolume of bacterial communitiesranged between 0.74 6 0.06 3 105 mm3 mL21 in LakeGutierrez at 1% PAR and 2.34 6 0.03 3 105 mm3 mL21 inLake Gutierrez at 50% PAR (Table 3). Statistical analysisshowed that total biovolume did not vary significantlybetween lakes or depths (two-way ANOVA p . 0.05).

We investigated the distribution of cell morphologyalong the euphotic zone, recognizing small free-livingbacteria (cocci and rods shorter than 7 mm) and longerbacterial filaments (or chains of rods). The relativeproportion of filament biovolume to total bacterialbiovolume increased at 50% PAR depth (Fig. 5A) (two-way ANOVA, lakes F7,14 5 6.26, p 5 0.002; depth F2,14 519.91, p , 0.001 and a posteriori Tukey test, 50% PARhigher than other depth, p , 0.001). Overall the relativeproportion of filaments was 34.4% at 50% PAR and thenreduced to 17.1% at 10% PAR and to 9.7% at 1% PAR(Fig. 5B). The number of filaments also varied significantlybetween lakes and depths (two-way ANOVA, lakes F7,14 59.81, p , 0.001; depth F2,14 5 27.95, p , 0.001) (Fig. 5C).Lake Gutierrez exhibited the highest filament abundance (aposteriori Tukey test p , 0.0005 with all other lakes) andthe number of filaments was significantly higher at 50%PAR for all lakes (a posteriori Tukey test 50% vs. 10% orvs. 1% p 5 0.005). In addition, we found that filamentlength also increased toward the surface (Table 3, Fig. 5D),resulting in the lengths at 50% PAR depth significantlylonger than those present at deeper layers (one-way


Fig. 2. Richness and gross community compostion of the bacterial assemblages. (A) Richness is measured as number of significantOTUs per sample as obtained from DGGE profiles; different colors refer to different depths (percentage of surface PAR). (B) Grossrelative bacterial community composition is obtained by recognizing sequences of excised bands from DGGE gels: the importance of eachOTU (grouped when belonging to the same phylogenetic group) was obtained by its relative intensity in the DGGE gel.

1104 Corno et al.

Fig. 3. Phylogenetic tree showing the affiliation of the 16S rRNA encoding gene sequences retrieved from DGGE bands (about550 bp) from Andean Lakes (AL Band) with reference sequences obtained from GenBank (accession numbers are indicated). The scalerepresents the percentage of estimated sequence divergence.


ANOVA F2,20 5 22.29, p , 0.001, and a posteriori Tukeytest, 50% PAR longer than 10% or 1% PAR, p , 0.001).

Ecological factors and bacterial morphology—NF (MxFland HNF) and the peritrich ciliate O. naumanni wereanalyzed as bacterial predators. NF abundances atdifferent depths resulted in an increment toward 1% ofsurface PAR (Fig. 6). These predatory protists were mainlymixotrophs since they were dominated by Chrysochromu-lina parva and Rhodomonas lacustris. The proportion ofobserved MxFl was always higher than 90% of the totalflagellate number, whereas HNF never exceeded 10%. Themixotrophic ciliate O. naumanni also showed an increase inabundances at 1% of PAR (Fig. 6). As a proxy of relativegrazing pressure by HNF and MxFl on bacterial assem-blages the mean prey–predator ratio was calculated for alllakes and depths (Fig. 7). The calculated value for the

relative grazing effect approximately equaled or exceeded1000 bacteria NF21, with the exception of one sample eachfrom Espejo (50% PAR) and Gutierrez (1% PAR).

We analyzed the dependence of the proportion offilaments on different ecological factors: flagellates, C : Pratio (dissolved fraction), UVR (each band considered in adifferent analysis), and predation pressure (filaments : NFratio). The forward stepwise multiple regression analysisshowed that the only variable significant in the model wasUVR (for 305 nm R2 5 0.56, p , 0.0001; for 320 nm R2 50.52, p , 0.0001; for 340 nm R2 5 0.49, p 5 0.0001; for380 nm R2 5 0.46, p 5 0.0002). In all cases flagellateabundances, C : P ratio, and estimated predation pressurewere not statistically significant and therefore were notincluded in the model. Overall results showed a positiverelationship between the proportion of filaments and UVR(Fig. 8).

Fig. 4. (A) First three axes of the principal component analysis (PCA) on the bacterial community composition (DGGE bands)belonging to different lakes and depths. (B) First three axes of the multivariate redundancy analysis (RDA) with the vector resulted fromthe environmental variables. References: KdPAR: diffuse attenuation coefficients of PAR (400–700 nm); Kd305, Kd320, Kd340, and Kd380:diffuse attenuation coefficients UVR bands (305, 320, 340, and 380 nm); TP, total phosphorus; TDP, total dissolved phosphorus; TPP,total particulate phosphorus; TN, total nitrogen; DOC, dissolved organic carbon; Chl a, chlorophyll a concentration; Prok, totalprokaryote; NF, nanoflagellates; and Ophry: O. naumanni abundances.

Table 3. Bacterial abundances (106 cells mL21), biovolume (105 mm3 mL21), and filament length (mm). Values are expressed as mean6 SE.

Lakes

Bacterial number Bacterial biovolume Filament length

50% PAR 10% PAR 1% PAR 50% PAR 10% PAR 1% PAR 50% PAR 10% PAR 1% PAR

Espejo 0.4260.02 0.9660.02 0.9860.02 0.8860.06 1.2460.03 0.9960.01 18.4862.17 12.2660.84 12.1460.90Correntoso 0.8660.01 1.1060.01 1.4360.02 1.1960.04 1.1060.01 2.0060.03 19.0961.45 13.1260.82 14.4861.02Nahuel Huapi 1.4360.04 1.4660.00 1.3260.01 1.3360.03 1.2560.00 1.3260.01 19.0761.20 13.7660.85 12.6960.83Gutierrez 0.5360.00 0.6460.01 0.5560.02 2.3460.02 1.6360.02 0.7460.05 18.0663.12 13.5460.77 12.4860.70Mascardi Cat. 0.7160.00 0.9160.00 0.8960.08 0.9760.05 1.1860.01 1.7660.00 14.6060.76 12.6260.65 No filamentsMascardi Tr. 1.1360.05 1.1860.04 1.3260.01 1.2260.16 1.1360.08 0.9560.11 15.7260.89 12.6260.58 12.2060.82Rivadavia 0.9660.06 0.7860.14 1.4360.01 1.0560.13 0.7660.12 1.0660.00 16.0260.83 13.3561.11 13.5360.95Futalaufquen 1.0360.01 1.0760.02 1.3260.03 1.3260.02 0.9660.04 1.4860.03 14.8360.98 14.3060.76 14.7261.35

1106 Corno et al.

Discussion

In our survey of Andean lakes from Atlantic and Pacificwatersheds only slight changes in the gross bacterial

community composition were detected between lakes,allowing the conclusion that the gross composition washomogeneous throughout the geographic region investigat-ed. The group of CFB was the most diverse and important

Fig. 5. (A) Filament (.7 mm) relative importance and distribution at different depths. The percentage of filament biovolumes (inblack) on the total bacteria is given for each lake and depth, and (B) mean of all samples for each depth. (C) Absolute filamentabundances are reported as number of cells mL21. (D) Mean filaments length (mm); data are mean (6SD) of cells .7 mm among lakes.

Fig. 6. Overall predator abundances at three depths (percentage of surface PAR). Totalnanoflagellate (NF) values in the left graph are given by the sum of mixotrophic (MxFl) andheterotrophic NF (HNF) abundances. Right graph reports overall Ophrydium abundances. Givenvalues are mean (6SD) of all lakes.


in terms of relative abundance, together with Actinobac-teria, present in all samples and in all depths. In particular,in two of the most transparent lakes (Nahuel Huapi andEspejo Z1% 305 % 11 m and Z1% PAR % 44 m), threeDGGE bands of Actinobacteria were present. The impor-

tance of this class in Andean lakes can be related to its highUVR resistance, also observed by Warnecke et al. (2005) inmountain lakes. The multivariate analyses on the basis ofDGGE bands grouped the lakes without a clear pattern onthe basis of different basin or watershed, indicating that

Fig. 7. Predation pressure (vertical bars) by nanoflagellates (NF) in relation to NF (blackcircles) and Ophrydium (crossed triangles) abundances. Lake depths with a relation lower than1000 bacteria NF21 are considered NF controlled.

Fig. 8. Regression model between UVR intensity (mW cm22 nm21) at different wavelengths(305, 320, 340, 380 nm) and the relative proportion of filaments on total bacterial biovolume.

1108 Corno et al.

geographical distance at regional scale did not primarilydetermine the bacterial composition of these lakes.Additionally, the RDA analysis showed that bacterialcommunity composition appears to be regulated by localfactors operating with lake transparency (UVR and PAR),followed by P and DOC concentrations. In this sense,Logue and Lindstrom (2008) indicated that regionalprocesses act on a much larger spatial and temporal scale,and therefore are difficult to incorporate into ecologicalstudies. We cannot exclude that a larger data set withcomparisons across different environments and areas couldresult in stronger evidence of a direct correlation betweenbiogeography and the gross bacterial diversity of thetransparent Andean lakes.

Multivariate analyses (PCA and RDA) grouped thebacterial communities of the same lake regardless ofsampling depth in the euphotic zone, which includedepilimnion, metalimnion, and hypolimnion. Thus, ourhypothesis of different bacterial community compositionalong the euphotic zone was not supported by the analyzeddata. The absence of a vertical variability in bacterioplank-ton gross composition reflected the homogenous distribu-tion of chemical parameters such as nutrient and dissolvedoxygen. In particular, the lack of a vertical variability indissolved oxygen concentration, which has been indicatedas a potential source of niche selectivity in bacterioplank-ton (Urbach et al. 2001; Dominik and Hofle 2002; Zwisleret al. 2003), might contribute to this relatively uniformdistribution of the bacterial groups.

In the studied lakes the thermocline lies within theeuphotic zone; thus the sampling depths included twotemporarily isolated layers, epilimnion and hypolimnion.Organisms inhabiting each of these strata were affected bydifferential UVR and PAR irradiances. For instance, lakesof the Nahuel Huapi drainage area presented an epilimnionentirely exposed to UVB (305 nm) and UVA (340 nm)radiation. During summer stratification, bacterial assem-blages living in these water layers were isolated because ofthe sharp density gradient that prevented water mixing. Theresidence time for bacterioplankton and phytoplankton inthe surface waters was considered important to estimate thepotentially hazardous effect of UVR and PAR (Scully et al.2000) causing direct damage at the membrane cellular levels(Maranger et al. 2002). Futhermore, it has been document-ed that under calm, clear conditions (e.g., calm wind,summer season) the damaging effect might be enhanced(Modenutti et al. 2008). Under this scenario we observed asubstantial change in the morphological distribution of thebacterial assemblages, the relative proportion of thefilaments on total biovolume being higher in the upperlevels (at 50% surface PAR). Moreover, lakes with higherUVR at 50% PAR showed not only larger proportions offilaments at this depth but also longer filaments. It shouldbe noted that in most studied lakes (all but Rivadavia andFutalafquen, Table 2), the depth of 10% PAR lies belowthe epilimnion, and so the 50% PAR sample wouldrepresent the mean condition of bacteria dragged alongthe epilimnion. Although we analyzed the filamentationagainst the actual irradiance at each depth, it could beassumed that the observed filamentation at 50% PAR

depth is a consequence of the mean epilimnetic UVirradiance during the last week.

Filamentation is one of the most studied phenotypicantipredation strategies in aquatic environments (Jurgensand Matz 2002; Corno and Jurgens 2006), and it wasobserved in nutrient-rich environments as well as underlimiting conditions or starvation (Jiang and Chai 1996). Inparticular, predation by NF is considered one of the majorforces shaping aquatic bacterial communities, controllingboth biomass (Thingstad and Lignell 1997) and relativediversity (Corno and Jurgens 2008). System productivityand availability of nutrients also play a fundamental role inthe development of resistant bacterial morphotypes (Hahnand Hofle 2001; Matz and Jurgens 2003; Simek et al. 2003).Filaments either avoid predation by exceeding the prey sizethat most bacterivorous NF can feed on, or are grazed atsignificantly lower rates than smaller bacteria (Hanh andHofle 1998; Wu et al. 2005). However, the rate of flagellategrazing on bacteria may also be reduced when theorganisms are exposed to UVR (Sommaruga et al. 1996),suggesting that filamentation due to predation, under thiscondition of high UVR, should not be enhanced.

According to the estimated predation pressure on thebasis of Fenchel (1986), in most of the studied lakes,predation by NF had a limited effect. The bacteria predatorcommunity was composed mainly of MxFl and of theciliate O. naumanni, both showing an increase in abun-dances at 1% of PAR. Nevertheless, predation pressure wasrelatively constant at different depths, allowing theassumption of an even-size class and bacterial morphologydistribution. In contrast, in all lakes the measured,morphological distribution of bacteria was quite differentalong the water column, with a larger proportion offilaments in surface waters and a monotonic decrease withdepth. In the Andean lakes the high proportion offilamentous forms found at 50% PAR cannot be attributedto high predation pressure by NF or to high productivity.All the lakes are in fact unproductive systems and did notpresent a correlation between the predation pressure by NFand the proportion of filaments, thus suggesting that theeffect of these factors was not the main force enhancingfilamentation in these lakes. Another possibility would bethat the ciliate O. naumanni, which increases its abundancetoward the 1% PAR depth, decreases filament length atdeeper layers of the euphotic zone by preying upon them.However, previous work showed that O. naumanni preys onparticles ,1 mm (Modenutti and Balseiro 2002) andthrough catalyzed reporter deposition fluorescence in situhybridization (CARD-FISH) dye analysis no filamentswere observed in food vacuoles of this ciliate species(Modenutti et al. 2008). Therefore, the inverse relationbetween filament proportion and O. naumanni abundancewas not the consequence of a top-down effect.

It has also been proposed that the C : P ratio caninfluence filament formation (Matz and Jurgens 2003):systems relatively poor in C are usually unfavorable forfilamentation. Lakes Nahuel Huapi and Gutierrez have thehighest dissolved C : P ratio, and therefore are expected tosustain the largest populations of filaments, but thisrelation is more difficult to find in the other lakes where


C is relatively lower. In general, with such limitingconditions, it is difficult to predict which nutrients canbecome limiting at a certain step. In any case, it is clear thatNF predation, as well as Ophrydium abundance, andnutrient availability, either alone or in combination, aresufficient to explain the larger proportion and length offilaments occurring at higher UVR and PAR, and theconstant decrease along the water column in all lakes whereNF and ciliate abundances increased. In chemostatscontaining natural assemblages of freshwater bacteria andHNF, Corno and Jurgens (2008) found that under highHNF predation pressure, inedible bacterial biovolume(filaments) reached 25–63% of total biovolume. However,under very different situations (no differential predationpressure or nanopredator abundances) at the upper layersof the euphotic zone we obtained comparable results in theproportion of filaments. This result suggests that UVR canbe as important as nanopredators inducing filamentation innatural communities.

The strong correlation between UVR and filamentproportion suggests a direct effect of UVR on the filamentformation. We are aware of only a few studies attempting todocument the relationship between filamentation and UVintensity. Strikingly, a large number of bacterial filamentshas been reported in ultraoligotrophic high-altitude Alpinelakes (Wille et al. 1999), with environmental conditions (e.g.,the high effect of UVR in surface layers) rather similar tothose of Andean lakes, but the possibility of UV as a factorinfluencing filamentation was not mentioned. We alsonoticed that most studies analyzing filamentation due topredation were laboratory experiments without UVRexposure (Matz and Jurgens 2003; Corno and Jurgens2006). However, early laboratory studies indicated that astrain of E. coli exposed to UVR produced elongatedfilaments as a consequence of a failure in cell division (Kalleand Sivasubramanian 1967; Witkin 1967). More recently,UVR has also been reported as the cause of morphologicalchanges in cyanobacteria (Wu et al. 2005).

Aquatic organisms have evolved the capacity to synthe-size or accumulate UV-absorbing mycosporine-like aminoacids (MAAs) for protection against environmental UVR;however, small cell size itself limits the efficiency of suchUV-absorbing pigments (Shick and Dunlap 2002). Al-though there is considerable information about thepresence of MAAs in cyanobacteria (Sommaruga andGarcıa Pichel 1999; Castenholz et al. 2000), little is knownabout these UV-absorbing compounds in heterotrophicbacteria. Only in one marine Actinobacteria of the genusMicrococcus was the presence of MAAs observed (Araiet al. 1992). On the other hand, there is no direct evidenceof the benefit of bacteria elongation as a response to UVR.Nevertheless, theory predicts that smaller cells are moresensitive to UVB-mediated photodamage than larger ones(Garcıa Pichel 1994). Empirical evidence that supports thisassumption comes from phytoplankton studies where itwas observed that smaller cells are more sensitive to DNAdamage than larger ones (Karentz et al. 1991; Helbling et al.2005).

The fact that larger cells are less sensitive to UVBdamage allows us to assume that UVR would act as a

factor promoting bacterial filamentation. Andean lakeshave extremely transparent waters to UVR, and this factorhas been mentioned as responsible for decreases inphotosynthetic efficiencies (Modenutti et al. 2004), strongphotoinhibition (Callieri et al. 2007), zooplankton surfaceavoidance (Alonso et al. 2004), and the presence anddominance of species heavily pigmented and UVR resistant(Modenutti et al. 2005, 2008). In these systems, bacteriawould also show consequences of these high irradiances.

In this study, from several in situ data it has beendemonstrated for the first time the direct implication ofUVR in favoring bacterial filamentation. Our data alsosupport the assumption that filamentation results fromphenotypic plasticity and not from the selection of aspecific bacterial composition. The appearance of bacterialstrains able to develop different morphotypes as a responseto specific environmental factors reduces the potentialrichness of the bacterial community, resulting in a selectiontoward the most plastic bacteria.

AcknowledgmentsWe are grateful to the reviewers for their suggestions and

comments. We thank the authorities of Argentina National Parksfor permission to carry out this sampling program.

This work was carried out in a framework of collaboration projectConsejo Nacional de Investigaciones Cientıficas Tecnicas (CON-ICET)–Consiglio Nazionale delle Ricerche (CNR) (Argentina–Italy)and was supported by Fondo para la Investigacion Cientıfica yTecnologica (FONCyT) PICT 2007-01256 and 2007-01258.

References

ALONSO, C., V. ROCCO, J. P. BARRIGA, M. A. BATTINI, AND H.ZAGARESE. 2004. Surface avoidance by freshwater zooplank-ton: Field evidence on the role of ultraviolet radiation.Limnol. Oceanogr. 49: 225–232.

ARAI, T., M. NISHIJIMA, K. ADACHI, AND H. SANO. 1992. Isolationand structure of a UV absorbing substance from the marinebacterium Micrococcus sp. AK-334.MBI Report. Mar.Biotechnol. Inst. Tokyo 113: 88–94.

ARRIETA, J. M., M. G. WEINBAUER, AND G. J. HERNDL. 2000.Interspecific variability in sensitivity to UV radiation andsubsequent recovery in selected isolates of marine bacteria.Appl. Environ. Microbiol. 66: 1468–1473.

BALSEIRO, E., B. MODENUTTI, C. QUEIMALINOS, AND M. REISSIG. 2007.Daphnia distribution in Andean Patagonian lakes: Effect of lowfood quality and fish predation. Aquat. Ecol. 41: 599–609.

———, M. S. SOUZA, B. MODENUTTI, AND M. REISSIG. 2008.Living in transparent lakes: Low food P : C ratio decreasesantioxidant response to ultraviolet radiation in Daphnia.Limnol. Oceanogr. 53: 2383–2390.

BENSON, D. A., M. S. BOGUSKI, D. J. LIPMAN, J. OSTELL, B. F. F.OUELLETTE, B. A. RAPP, AND D. L. WHEELER. 1999. GenBank.Nucleic Acids Res. 27: 12–17.

BOENIGK, J., P. STADLER, A. WIEDLROITHER, AND M. W. HAHN. 2004.Strain-specific differences in the grazing sensitivities of closelyrelated ultramicrobacteria affiliated with the Polynucleobactercluster. Appl. Environ. Microbiol. 70: 5787–5793.

CALLIERI, C., B. MODENUTTI, C. QUEIMALINOS, R. BERTONI, AND E.BALSEIRO. 2007. Production and biomass of picophytoplank-ton and larger autotrophs in Andean ultraoligotrophic lakes:Differences in light harvesting efficiency in deep layers.Aquat. Ecol. 41: 511–523.

1110 Corno et al.

CASTENHOLZ, R. W., F. GARCIA-PICHEL, AND M. P. BA WHITTON.2000. Cyanobacterial responses to UV-radiation, p. 591–611.In B. A. Whitton and M. Potts [eds.], The ecology ofcyanobacteria: Their diversity in time and space. KluwerAcademic Publishers.

CORNO, G. 2006. Effects of nutrient availability and Ochromonassp. predation on size and composition of a simplified aquaticbacterial community. FEMS Microbiol. Ecol. 58: 354–363.

———, E. CARAVATI, C. CALLIERI, AND R. BERTONI. 2008. Effectsof predation pressure on bacterial abundance, diversity, andsize-structure distribution in an oligotrophic system. J.Limnol. 67: 107–119.

———, AND K. JURGENS. 2006. Direct and indirect effects ofprotist predation on population size structure of a bacterialstrain with high phenotypic plasticity. Appl. Environ.Microbiol. 72: 78–86.

———, AND ———. 2008. Structural and functional patterns ofbacterial communities in response to protist predation alongan experimental productivity gradient. Environ. Microbiol.10: 2857–2871.

CRUMP, B. C., H. E. ADAMS, J. HOBBIE, AND G. W. KLING. 2007.Biogeography of bacterioplankton in lakes and streams of anarctic tundra catchment. Ecology 88: 1365–1378.

CURTIS, T. P., AND W. T. SLOAN. 2004. Prokaryotic diversity andits limits: Microbial community structure in nature andimplications for microbial ecology. Curr. Opin. Microbiol. 7:221–226.

DEL GIORGIO, P. A., AND R. H. PETERS. 1994. Patterns inplanktonic P-R ratios in lakes—influence of lake trophy anddissolved organic carbon. Limnol. Oceanogr. 39: 772–787.

DOMINIK, K., AND M. G. HOFLE. 2002. Changes in bacterioplank-ton community structure and activity with depth in aeutrophic lake as revealed by 5S rRNA analysis. Appl.Environ. Microbiol. 68: 3606–3613.

FENCHEL, T. 1986. The ecology of heterotrophic microflagellates.Adv. Microb. Ecol. 9: 57–97.

FENG, D. F., AND R. F. DOOLITTLE. 1987. Progressive sequencealignment as a prerequisite to correct phylogenetic trees. J.Mol. Evol. 25: 351–360.

GARCIA-PICHEL, F. 1994. A model for internal self-shading inplanktonic organisms and its implication for the usefulness ofultraviolet sunscreens. Limnol. Oceanogr. 39: 1704–1717.

HAHN, M. W., AND M. G. HOFLE. 1998. Grazing pressure by abacterivorous flagellate reverses the relative abundance ofComamonas acidovorans PX54 and Vibrio strain CB5 inchemostat cocultures. Appl. Environ. Microbiol. 64: 1910–1918.

———, AND ———. 2001. Grazing of protozoa and its effect onpopulations of aquatic bacteria. FEMS Microbiol. Ecol. 35:113–121.

HELBLING, E. W., E. BARBIERI, M. A. MARCOVAL, R. J.GONCALVES, AND V. E. VILLAFANE. 2005. Impact of solarultraviolet radiation on marine phytoplankton of Patagonia,Argentina. Photochem. Photobiol. 81: 807–818.

———, V. E. VILLAFANE, AND O. HOLM-HANSEN. 1995. Variabilityof phytoplankton distribution and primary productionaround Elephant Island, Antarctica, during 1990–1993. PolarBiol. 15: 233–243.

IRIONDO, M. H. 1989. Quaternary lakes of Argentina. Paleogeogr.Paleoclimatol. Paleoecol. 70: 81–88.

JIANG, X. P., AND T. J. CHAI. 1996. Survival of Vibrioparahaemolyticus at low temperatures under starvationconditions and subsequent resuscitation of viable, noncultur-able cells. Appl. Environ. Microbiol. 62: 1300–1305.

JOUX, F., W. H. JEFFREY, P. LEBARON, AND D. L. MITCHELL. 1999.Marine bacterial isolates display diverse responses to UV-Bradiation. Appl. Environ. Microbiol. 65: 3820–3827.

JURGENS, K., AND C. MATZ. 2002. Predation as a shaping force forthe phenotypic and genotypic composition of planktonicbacteria. Antonie Van Leeuwenhoek 81: 413–434.

KALLE, G. P., AND P. P. SIVASUBRAMANIAN. 1967. Filamentformation in ultraviolet-resistant mutants of Escherichia colistrain B-3. Radiat. Res. 30: 71–80.

KARENTZ, D., F. S. MCEUEN, M. C. LAND, AND W. C. DUNLAP.1991. Survey of mycosporine-like amino acid compounds inAntarctic marine organisms: Potential protection fromultraviolet exposure. Mar. Biol. 108: 157–166.

LANGENHEDER, S., E. S. LINDSTROM, AND L. J. TRANVIK. 2006.Structure and function of bacterial communities emergingfrom different sources under identical conditions. Appl.Environ. Microbiol. 72: 212–220.

———, AND H. RAGNARSSON. 2007. The role of environmental andspatial factors for the composition of aquatic bacterialcommunities. Ecology 88: 2154–2161.

LINDSTROM, E. S., AND A. K. BERGSTROM. 2004. Influence of inletbacteria on bacterioplankton assemblage composition in lakesof different hydraulic retention time. Limnol. Oceanogr. 49:125–136.

LOGUE, J. B., AND E. S. LINDSTROM. 2008. Biogeography ofbacterioplankton in inland waters. Freshw. Rev. 1: 99–114.

MARANGER, R., P. A. DEL GIORGIO, AND D. F. BIRD. 2002.Accumulation of damaged bacteria and viruses in lake waterexposed to solar radiation. Aquat. Microb. Ecol. 28: 213–227.

MATZ, C., AND K. JURGENS. 2003. Interaction of nutrientlimitation and protozoan grazing determines the phenotypicstructure of a bacterial community. Microb. Ecol. 45:384–398.

MODENUTTI, B., E. BALSEIRO, C. CALLIERI, C. QUEIMALINOS, AND

R. BERTONI. 2004. Increase in photosynthetic efficiency as astrategy of planktonic organisms exploiting deep lake layers.Freshw. Biol. 49: 160–169.

MODENUTTI, B. E., AND E. G. BALSEIRO. 2002. Mixotrophic ciliatesin an Andean lake: Dependence on light and prey of anOphrydium naumanni population. Freshw. Biol. 47: 121–128.

———, ———, C. CALLIERI, AND R. BERTONI. 2008. Light versusfood supply as factors modulating niche partitioning in twopelagic mixotrophic ciliates. Limnol. Oceanogr. 53: 446–455.

———, ——— , ———, ———, AND C. P. QUEIMALINOS. 2005.Effect of UV-B and different PAR intensities on the primaryproduction of the mixotrophic planktonic ciliate Stentoraraucanus. Limnol. Oceanogr. 50: 864–871.

———, ———, AND C. P. QUEIMALINOS. 2000. Ciliate communitystructure in two South Andean lakes: The effect of lake wateron Ophrydium naumanni distribution. Aquat. Microb. Ecol.21: 299–307.

———, ———, ———, D. A. ANON SUAREZ, M. C. DIEGUEZ, AND

R. J. ALBARINO. 1998. Structure and dynamics of food web inAndean lakes. Lakes Reserv. Res. Manage. 3: 179–186.

MORRIS, D. P., AND oTHERS. 1995. The attenuation of solar UVradiation in lakes and the role of dissolved organic carbon.Limnol. Oceanogr. 40: 1381–1391.

MUYZER, G., T. BRINKHOFF, U. NUEBEL, C. SANTEGOEDS, H.SCHAEFER, AND C. WAWER. 1998. Denaturing gradient gelelectrophoresis (DGGE) in microbial ecology, p. 1–27. In A.D. L. Akkermans, J. D. van Elsas and F. J. Bruijn [eds.],Molecular microbial ecology manual. Kluwer AcademicPublishers.

NUSCH, E. A. 1980. Comparison of different methods forchlorophyll and phaeopigment determination. Arch. Hydro-biol. Beih. Ergebn. Limnol. 14: 14–36.

PACE, M. L., AND oTHERS. 2004. Whole-lake carbon-13 additionsreveal terrestrial support of aquatic food webs. Nature 427:240–243.


PARUELO, J. M., A. BELTRAN, E. JOBBAGY, O. SALA, AND R.GOLLUSCIO. 1998. The climate of Patagonia: General patternsand controls on biotics processes. Ecol. Aust. 8: 85–101.

PEREZ, G., C. QUEIMALINOS, E. BALSEIRO, AND B. MODENUTTI.2007. Phytoplankton absorption spectra along the watercolumn in deep North Patagonian Andean lakes (Argentina).Limnologica 37: 3–16.

PERNTHALER, J. 2005. Predation on prokaryotes in the water columnand its ecological implications. Nat. Rev. Microbiol. 3: 537–546.

PORTER, K. G., AND Y. S. FEIG. 1980. The use of DAPI foridentifying and counting aquatic microflora. Limnol. Ocea-nogr. 25: 943–948.

SANCHEZ, O., J. M. GASOL, R. MASSANA, J. MAS, AND C. PEDROS-ALIO. 2007. Comparison of different denaturing gradient gelelectrophoresis primer sets for the study of marine bacter-ioplankton communities. Appl. Environ. Microbiol. 73:5962–5967.

SCULLY, N. M., P. R. LEAVITT, AND S. R. CARPENTER. 2000.Century-long effects of forest harvest on the physicalstructure and autotrophic community of a small temperatelake. Can. J. Fish. Aquat. Sci. 57: 50–59.

SHICK, J. M., AND W. C. DUNLAP. 2002. Mycosporine-like aminoacids and related gadusols: Biosynthesis, accumulation, andUV-protective functions in aquatic organisms. Ann. Rev.Physiol. 64: 223–262.

SIMEK, K., H. HORNAK, M. MASIN, U. CHRISTAKI, J. NEDOMA, M.G. WEINBAUER, AND J. R. DOLAN. 2003. Comparing the effectsof resource enrichment and grazing on a bacterioplanktoncommunity of a meso-eutrophic reservoir. Aquat. Microb.Ecol. 31: 123–135.

SOMMARUGA, R., AND F. GARCIA-PICHEL. 1999. UV-absorbingmycosporine-like compounds in planktonic and benthic organ-isms from a high-mountain lake. Arch. Hydrobiol. 144: 255–269.

———, A. OBERLEITER, AND R. PSENNER. 1996. Effect of UVradiation on the bacterivory of a heterotrophic nanoflagellate.Appl. Environ. Microbiol. 62: 4395–4400.

———, I. OBERNOSTERER, G. J. HERNDL, AND R. PSENNER. 1997.Inhibitory effect of solar radiation on thymidine and leucineincorporation by freshwater and marine bacterioplankton.Appl. Environ. Microbiol. 63: 4178–4184.

SORANNO, P. A., AND oTHERS. 1999. Spatial variation among lakeswithin landscapes: Ecological organization along lake chains.Ecosystems 2: 395–410.

STERNER, R. W., J. J. ELSER, E. J. FEE, S. J. GUILDFORD, AND T. H.CHRZANOWSKI. 1997. The light : nutrient ratio in lakes: Thebalance of energy and materials affects ecosystem structureand process. Am. Nat. 150: 663–684.

THINGSTAD, T. F. 2000. Elements of a theory for the mechanismscontrolling abundance, diversity, and biogeochemical role oflytic bacterial viruses in aquatic systems. Limnol. Oceanogr.45: 1320–1328.

———, AND R. LIGNELL. 1997. Theoretical models for the controlof bacterial growth rate, abundance, diversity and carbondemand. Aquat. Microb. Ecol. 13: 19–27.

TRANVIK, L. J., AND S. BERTILSSON. 2001. Contrasting effects ofsolar UV radiation on dissolved organic sources for bacterialgrowth. Ecol. Lett. 4: 458–463.

URBACH, E., K. L. VERGIN, L. YOUNG, A. MORSE, G. L. LARSON,AND S. J. GIOVANNONI. 2001. Unusual bacterioplanktoncommunity structure in ultra-oligotrophic Crater Lake.Limnol. Oceanogr. 46: 557–572.

VALDERRAMA, J. C. 1981. The simultaneous analysis of totalnitrogen and total phosphorus in natural waters. Mar. Chem.10: 109–122.

WARNECKE, F., R. SOMMARUGA, R. SEKAR, J. S. HOFER, AND J.PERNTHALER. 2005. Abundances, identity, and growth stateof actinobacteria in mountain lakes of different UVtransparency. Appl. Environ. Microbiol. 71: 5551–5559.

———, R. SOMMARUGA, ———, ———, AND ———. 2005.Abundances, identity, and growth state of actinobacteria inmountain lakes of different UV transparency. Appl. Environ.Microbiol. 71: 5551–5559.

WEINBAUER, M. G., K. HORNAK, J. JEZBERA, J. NEDOMA, J. R.DOLAN, AND K. SIMEK. 2007. Synergistic and antagonisticeffects of viral lysis and protistan grazing on bacterialbiomass, production and diversity. Environ. Microbiol. 9:777–788.

WILLE, A., B. SONNTAG, B. SATTLER, AND R. PSENNER. 1999.Abundance, biomass and size-structure of the micro-bial assemblage in the high mountain lake Gossenkollesee(Tyrol, Austria) during the ice-free period. J. Limnol. 58:117–126.

WITKIN, E. M. 1967. The radiation sensitivity of Escherichiacoli B: A hypothesis relating filament formation andprophage induction. Proc. Natl. Acad. Sci. USA 57:1275–1279.

WU, H., K. GAO, V. E. VILLAFANE, T. WATANABE, AND E. W.HELBLING. 2005. Effects of solar UV radiation on mor-phology and photosynthesis of filamentous cyanobacteriumArthrospira platensis. Appl. Environ. Microbiol. 71:5004–5013.

ZWISLER, W., N. SELJE, AND M. SIMON. 2003. Seasonal patterns ofthe bacterioplankton community composition in a largemesotrophic lake. Aquat. Microb. Ecol. 31: 211–225.

Associate editor: Wade H. Jeffrey

Received: 06 November 2008Accepted: 27 February 2009

Amended: 12 March 2009

1112 Corno et al.

corno, gianluca, et al. bacterial diversity and morphology...

Documents