Field-scale biofiltration: Performanceevaluation and microbial analysis

Item Type text; Dissertation-Reproduction (electronic)

Authors Jutras, Eileen Maura 1958

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this materialis made possible by the University Libraries, University of Arizona.Further transmission, reproduction or presentation (such aspublic display or performance) of protected items is prohibitedexcept with permission of the author.

Download date 15/08/2021 02:47:42

Link to Item http://hdl.handle.net/10150/282533

INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI

films the text directly fi'om the original or copy submitted. Thus, some

thesis and dissertation copies are in typewriter &ce, while others may be

fi-om any type of computer printer.

The quality of this reproduction is dependent upon the quality of the

copy submitted. Broken or indistinct print, colored or poor quality

illustrations and photographs, print bleedthrough, substandard margins,

and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete

manuscript and there are missing pages, these will be noted. Also, if

unauthorized copyright material had to be removed, a note will indicate

the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by

sectioning the original, beginning at the upper left-hand comer and

continuing fi'om left to right in equal sections with small overlaps. Each

original is also photographed in one exposure and is included in reduced

form at the back of the book.

Photographs included in the original manuscript have been reproduced

xerographically in this copy. Kgher quality 6" x 9" black and white

photographic prints are available for any photographs or illustrations

appearing in this copy for an additional charge. Contact UMI directly to

order

UMI A BeU & Howell Information Company

300 North Zed) Road, Ann Arbor MI 48I06-I346 USA 313/761-4700 800/521-0600

FIELD-SCALE BlOFILTRATION: PERFORMANCE EVALUATION AND

MICROBIAL ANALYSIS

by

Eileen Maura Jutras

A Dissertation Submitted to the Faculty of the

DEPARTMENT OF SOIL, WATER AND ENVIRONMENTAL SCIENCE

In Partial Fulfillment of the Requirements For the Degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

1 9 9 7

UMI Nxuober: 9814426

UMI Microform 9814426 Copyright 1998, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, Ml 48103

2

THE UNIVERSITY OF ARIZONA ® GRADUATE COLLEGE

As members of the Final Examination Committee, we certify that we have

read the dissertation prepared by Eileen Maura Jutras

entitled Field-Scale Biofiltration: Performance Evaluation and

Microbial Analysis

and recommend that it be accepted as fulfilling the dissertation

requirement for the Degree of Doctor of Philosophy

' I I ^^^

Ian L. Peppei^ J / - / Date /

^ \ ^ 2 b j ' i " )

Raina M. MaTller Date

Norva\ A Sinclair i Date

c- Christina K. Kennedy Date

Date

Final approval and acceptance of this dissertation is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

/ . . / ^ Dissertation Director Ian L. Pepper Date

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when, in his or her judgment, the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

4

ACKNOWLEDGMENTS

In the words of my esteemed colleague Mark Burr, "It will take you longer than you will ever know." Nothing about my Ph.D. has happened on my time scale, it has all been in the time scale of the big picture and HP. I am very grateful that through it all I've kept my sobriety my #1 priority even though the work demands seemed as though they were most important at times. I especially want to thank Beth for never doubting that I am a good scientist.

This work would not have been as much fun without the Pepper People to back me up. Mark for showing me how to let things roll off my back and giving me an appreciation for the Beatles. Marlowe for always loving her lesbian and reminding me that a blow pop a day keeps the blues away. T-Ro for reminding me that science really is fun. Karen, for without her there would be no Pepper lab. Debbie for putting up with my constant teasing. Kelly who has been an inspiration and showed me that there reaUy is life after a Ph.D. Tori for pulling me through. Scooter Van Dam I Love U Man. Carl and Felisa for organizing those Friday lunches. Thank you to the secretaries, Jan, Elenor, Judy, and our accovmtant Barb who have put up with my "platinum moments". A special thanks to Myma, Andy, Rita, and Sunhee in Christina Kennedy's lab who helped me through the last leg of the research.

A big thanks goes to Drs. Ian Pepper and Raina Miller who kept me going even in my darkest moments of doubt. Your constant challenges have paid off, even if I didn't like them at the time. Drs. Sinclair and Kennedy always had an open door for me to come and discuss ideas or problems.

TABLE OF CONTENTS

LIST OF ILLUSTRATIONS 6

LIST OF TABLES 7

ABSTRACT 8

INTRODUCTION 10

Statement Of The Problem 10 Literature Review 13

Traditional Techniques For Entimeration Of Environmental Bacteria 13 Cultural Methods 15 Microscopic Direct Counts 18 Recent Techniques For Enumeration 19 Identification Of Enviroiunental Bacteria 21 Possible Environmental Constraints To Biodegradation Of Petroleum Compoimds In Soil 25

Dissertation Format 29

PRESENT STUDY 31

APPENDIX A: FIELD-SCALE BIOFILTRATION OF GASOLINE VAPORS EXTRACTED FROM BENEATH A LEAKING UNDERGROUND STORAGE TANK 33

APPENDIX B: COMBINED CULTURAL AND NUCLEIC ACID BASED METHODS TO FOLLOW COMMUNITY DEVELOPMENT IN A BIOFILTER 69

APPENDIX C: A COMPARISON OF ARBITRARILY PRIMED PGR AND 16S RFLP TO DIFFERENTL\TE ENVIRONMENTAL BACTERIAL ISOLATES 101

REFERENCES 118

6

LIST OF ILLUSTRATIONS

Figure Page

1 Flow chart of the experimental design for evaluating the biofilter

performance and microbial analysis 12

7

UST OF TABLES

Table Page

1 Spectnmi of bacterial activity in environmental samples 13

?

I

8

ABSTRACT

Biofiltration has been shown to be an effective method for the remediation

of volatile organic compounds (VOC's), particularly petroleum vapors extracted

from the vadose zone. Many bacteria have the enzymatic pathways necessary

for aerobic mineralization of VCXI's to form ceU biomass, carbon dioxide and

water. Molecular methods such as nucleic acid hybridizations and the

polymerase chain reaction (PCR), are methods that can be applied to

environmental samples to characterize bacterial community structure and

function. The research presented here reports the use of a field-scale biofilter for

the remediation of unleaded gasoline vapors extracted from the vadose zone.

An evaluation of contaminant removal efficiency over a five month period

showed that the biofilter removed 90% of total petroleum hydrocarbons and

greater than 90% of the EPA priority pollutants benzene, toluene, ethylbenzene,

and xylene. The bacterial consortium in the biofilter medium readily adapted to

increased loading rates, and variations in temperature and moisture. A

combination of conventional cultural and molecular methods was used to track

the bacterial populations over the course of the experiment. Polymerase chain

reaction amplification of the small ribosomal subunit DNA sequence was used

for identification of bacterial isolates and the design of DNA hybridization

probes. Hybridization of these probes to community DNA samples taken from

9

the biofilter over time revealed changes in specific bacterial populations as

bioremediation occurred. Specifically, bacteria that could use gasoline, toluene,

ethylbenzene or xylene were prevalent throughout the biofilter. Bacterial

populations that could degrade xylene gradually increased over time, while the

overall total population size was the similar for the background samples as well

as samples taken at the end of the study.

1 0

CHAPTER 1

INTRODUCTION

Statement Of The Problem

As there is no one detection method that wiU ensure fail-safe tracking of

bacteria, the research presented in the appendices illustrate a combination of

conventional and molecxilar biological techniques to assess the bioremediation of

unleaded gasoline at a field site in Phoenix, AZ. A field scale biofiltration system

using a combination of diatomaceous earth and composted horse manure was

used to remediate unleaded gasoUne vapors extracted from the site. Since the

contaminant removal capacity of the filter was dependent on the bacterial

population, an evaluation of the effects of biodegradation on the bacterial

community in the compost was evaluated. The removal of total petroleum

hydrocarbons (TPH) and selected constituents, benzene, toluene, ethylbenzene,

and xylenes (BTEX) was monitored concurrently with microbial analysis.

The overall goal of this research was to identify and evaluate the potential

of bacteria for gasoline degradation. Molecular methods were used to assess the

genetic potential and the effect of bioremediation on the bacterial community

diversity. Specifically, the bacterial population of the compost-based biofiltration

system was a diverse consortium that was characterized by amplifying extracted

DNA with arbitrarily primed polymerase chain reaction (AP-PCR).

Amplification of 16S rDNA was used for identification and design of probes to

1 1

track bacterial succession as bioremediation progressed. The specific objectives

of this proposal included:

1. Evaluation of the removal of TPH and BTEX as a residt of biofiltration using

quantitative gas chromatography.

2. Enumeration and isolation of viable bacteria using a dilute nutrient medium,

R2A, for total heterotrophic plate counts and a mineral salts medium amended

with hexadecane, MSM C16, for estimating TPH degraders.

3. Identification of viable bacteria using Biolog''"^* Gram negative and positive

plates.

4. Fingerprinting isolates using AP-PCR and 16S PCR-RFLP.

5. 16S rDNA sequence information for the identification of bacteria and the

design of hybridization probes.

6. Extraction of compost bacterial community DNA using a direct lysis protocol.

7. The monitoring bacterial succession of the biofilter compost over time using

hybridization.

1 2

Bioremediation

Performance

Influent vapor sample

Effluent vapor sample

9 internal vapor samples

Gas chromatography

TPH

Benzene

Ethylbenzene

Toluene

Xylene

Biolog

BIOFILTER

12 sample dates Microbiological

Studies

9 compost samples

Serial dilution and plating

Selection of isolates Growth on sole carbon source

Unleaded gasoline

Benzene

Toluene

Ethylbenzene

Xylene

Direct lysis

16S rDNA amplification

Community DNA Dot blots

AP-PCR

Fingerprints

16S rDNA amplification

RFLP fingerprints

Sequence

Probe design

Probe community 16S

Figure 1: Flow chart of the experimental design for evaluating the biofilter

performance and microbial analysis.

1 3

Literature Review

Traditional Techniques For Enumeration Of Environmental Bacteria

Bacteria are ubiquitous in the environment. In soil samples, total

culturable bacterial can be as high as 10^° colony forming units per gram of dry

weight (cfu g-^) with aerobic bacteria comprising 50 - 70 percent of the total

(Turco and Sadowski, 1995). The diversity of bacteria within a gram of soU is

estimated to be as high as ten thousand species (Turco and Sadowski, 1995).

Bacteria in the environment have a spectnmi of activity:

Table 1: Spectrum of bacterial activity in environmental samples.

LIVE DEAD High Autochthonous Unbalanced Biotic True Zero

activity and growth: and dormancy: activity

Autochthonous bacteria are slow growing indigenous populations that degrade

organic matter. The zymogenous bacteria are fast growing indigenous

populations that quickly increase in numbers when nutrients become available.

AUochthonous bacteria are non-indigenous or introduced organisms present as

the result of the addition of sludge, animal waste, or genetically engineered

microbes (GEM's). These organisms are generally transient in the soil

zymogenous

bacteria

Moribund abiotic Sporulation

stress:

Rounding

cells

1 4

environment with relatively low siuvival over long periods due to competition

and antagonism with indigenous populations.

Survival for all bacteria in the environment is contingent on the biotic and

abiotic environmental conditions. In general, extremes of either category inhibit

overall bacterial populations though many bacterial species have adapted to

unique environments such as the Yellowstone hotsprings (Bams et al., 1989).

One of the biological factors that influence whether bacterial populations

flourish is competition for available carbon sources for metabolism. Bacteria as a

whole are able to metabolize a wide array of compounds in the environment.

Those bacteria that have enzymatic pathways for several compounds wiU out

compete those that have limited, specialized requirements. In a typical soil

profile the available carbon decreases with depth and so do the numbers of

bacteria (Boivin-Jahns et al., 1996; Fredrickson et al., 1995). The rhizosphere is an

area around plant roots where plant root exudates are available. This area

typically has increased numbers of bacteria that take advantage of the available

carbon source from plant exudates. Other bacteria-bacteria interactions such as

antagonism, commensalism, symbiosis, and mutualism aU affect bacterial

survival in the environment. Interactions of bacteria with plants, and bacteria

with other microorganisms especially predation of bacteria by protozoa and

microarthropods limit population growth.

1 5

Abiotic factors are chemical and physical properties of the soil that inhibit

bacterial growth. The redox potential of the soil influences the electron acceptors

that are available for respiration. Soils are variable and have micro sites with

limited oxygen thus Fe^^, NO3, SO4, and CO2 may be used as alternative terminal

electron acceptors for facultative and strict anaerobes. As well as available

carbon sources, other nutrients such as nitrogen and phosphorus need to be

present (Pena-Cabriales and Alexander, 1983). Soil pH outside the optimal range

of pH 6-8 becomes too acidic or too basic for many bacteria. Soil texture can

influence water activity and transport of bacteria. For example, desiccation

studies have shown that fine structured clays have more water holding capacity

than soils with high sand content which can increase the survival of bacteria

(Osa-Afiana and Alexander, 1982). Soils with pore sizes less than 2 nm restrict

bacterial transport.

Cultural Methods

Traditional methods of enumerating soil bacteria, serial dilution and

culturing on selective and differential media, are limited in that in most cases less

than 1 % of soil bacteria are culturable (Amann et al., 1995; Roszak and Colwell,

1987; and Ward et al., 1990). The Handbook of Microbiological Media (CRC) lists

several types of media for culturing bacteria (Atlas, 1995). Just the number of

1 6

different types of media needed to culture these organisms is an indication that

there is no one media that can be used to detect all viable bacteria in soil. A

second concern for any enimieration scheme is the sample size. What constitutes

a representative sample? How should it be divided to account for spatial

variability of an environmental sample? A third concern is that dilution of the

sample for plating results in dominant populations being over-represented while

obscure populations may be lost. In addition, the method of incubation can

adversely affect coimts. For example, within the soil there exist

microenvironments that are anaerobic. Some anaerobes can tolerate oxygen

while strict anaerobes must be cidtured in the total absence of oxygen.

Several assumptions are made for the dilution and plating technique:

1. One colony arises from one bacteriimi.

2. AH bacteria are brought into suspension.

3. All bacteria in the suspension grow.

The first assumption is discredited by the fact that bacteria are known to exist as

biofilms and are not free living organisms (Costerton et al., 1994). Although

surfactants and mechanical techniques are used to help disperse bacteria, a single

colony on an agar plate is usually the result of a clump of cells (Page et al., 1982).

Actinomycetes which have hyphae and are spore formers can generate several

1 7

colonies on an agar plate because dispersion techniques result in shearing of

h)^hae and transmittal of spores (VanDijk, 1979).

The second assumption is made to accoimt for the fact that bacteria in

particular colonize surfaces. They are attached to particle surfaces and it is

difficult to remove intact cells that are culturable (Ozawa and Yamaguchi, 1986).

Many bacteria are in close associations with plants and the rhizosphere effect has

shown that microbial counts can differ several orders of magnitude in proximity

to plant roots. Tj^ically, soil is sieved to remove plant debris and therefore the

bacteria associated with this debris are also removed.

The third assumption that all microbes will grow has been invalidated

with the concept of viable but nonculturable cells (Rosak and ColweU, 1987).

Viable but nonculturable cells are those cells that are injured or truly

nonculturable in the laboratory setting. The culturing techruque relies on the

ability of microbes to increase their numbers through cell division. As shown in

the spectrvmi above, cells in a state anything less than actively dividing may not

be culturable. This may be do to several factors including availability of proper

energy sources, injury to the cell, inadequate moisture, pH, and temperature. It

has been shown that plate counts of bacteria from soil increase when dilute

concentrations versus concentrated energy sources are used (Brendecke et al.,

1993). There can also be competition on the agar plate as there is in the

environment. Antagonism, spreading colonies, and metabolism that alters the

I S

local pH of the media all have the effect of limiting growth of bacteria. Serial

dilution and plating has persisted in part because the method is inexpensive,

easy to perform, and when used consistently, comparisons can be made between

samples.

Microscopic Direct Coimts

Microscopic direct counts of bacteria use fluorochrome staining such as,

3,6-bis[dimethylamino]acridiniimi chloride (acridine orange, AO) and 4',6-

diamidino-2 phenyHndole (DAPI) which bind to the DNA and/or RNA within a

cell. AO binds to DNA and RNA within a cell and fluoresces differentially based

on whether it is bound to double or single stranded nucleic acids. Double

stranded binding results in emission of green fluorescence and single stranded

binding or binding to particulates results in dye-dye interactions causing orange

fluorescence (Kepner and Pratt, 1994).

DAPI binds to A-T rich regions of DNA within a cell. When bound to

these regions it emits a blue or bluish-white fluorescence. When DAPI is

unbound or binds to material other than DNA it fluoresces yellow. The

differential emission helps to eliminate incorrectly coimting miss identified

particulates.

1 9

Epifluourescent microscopic direct counts of bacteria using fluorochrome

staining such as AO and DAPI, have resulted in estimations of 1-2 orders of

greater magnitude than spread plate counts (Kepner and Pratt, 1994). Direct

microscopic coxmts also give morphologic information about the bacteria and

their size. A limitation of AO and DAPI is that they both will bind to fine

particulate matter which can obstruct counts or they can bind to intact dead cells

resulting in over estimations. To determine if cells are viable, the sample can be

incubated with nalidixic acid which inhibits DNA synthesis while allowing the

cell to metabolize resulting in cells that are moribund and easily coimted in the

microscopic field (Kogure et al., 1978).

Recent Techniques For Enumeration

DNA based enumeration of environmental bacteria. Direct extraction of

DNA from soil bacteria has received increasing interest as a technique to detect

bacterial populations in environmental samples (More et al., 1994; Ogram et al.,

1987; Picard et al., 1992). The problems include meaningful analysis of the

extracted DNA, distinguishing prokaryotic from eukaryotic DNA, and also, free

from cellular DNA, the former being protected when sorbed to soil particles

(Lorentz and Wackemager, 1987). DNA-DNA hybridization studies suggest that

over 10,000 different species can exist in 1 gram of soil (Torsvik, 1978; Torsvik et

2 0

al., 1990). DNA extractions have been used for comparing the diversity of one

bacterial commiuiity to another rather than enumeration of bacteria present in a

sample (Ritz and Griffiths, 1994). To enumerate bacterial populations based on

the amount of DNA extracted, conversion factors for the amount of DNA per cell

must be used. A tj^ical E. coli chromosome is 4.7X10^ bases in length equaling 9

fg of DNA (Ingraham et al., 1983). The amount of DNA per cell varies for each

species and estimates for environmental bacteria are 2 - 4 fg of DNA per cell

(Krawiec and Riley, 1990). Also, the amount of DNA present in a cell can vary

because the bacterial chromosome can be dividing more than once resulting in

two to three times more DNA in the cell (Krawiec and Riley, 1990).

Ribosomal based enumeration of environmental bacteria. 16S ribosomal

RNA studies have burgeoned in the last decade. Fluorescently labeled RNA

probes have been designed to hybridize directly to the highly conserved regions

of the small ribosomal subunit for the universal detection Eubacteria and

Archaebacteria (Giovannoni, et al., 1988; Amann, et al., 1995). This method is

similar to AO or DAPI staining except that nucleic acid sequence information is

used to target specific groups of bacteria. Unlike other enumeration methods,

16S ribosomal RNA probing can also establish phylogenetic relationships

between bacterial using the sequence information. Since the probes are designed

to target the small subunit of the ribosome, a correlation can be made between

2 1

the amount of probe bound and the activity state of cell (DeLong, et al., 1989).

Actively metabolizing ceUs will have more ribosomes than non-metabolizing

ceUs.

Identification Of Environmental Bacteria

Conventional methods. Isolation of bacteria from the envirormient using

conventional methods of serial dilution and plating does not always result in a

positive identification of the organism. Bergy's manual provides a scheme for

identification based on carbon source utilization, and various enzymatic assays.

The recently developed Biolog'^" (Hayward, CA, USA) and APF'^ (Analytab

Products Inc., Plainview, NY, USA) systems are based on carbon source

utilization that results in a pattern for identification. Both of these methods were

intended for identifying clinical isolates not enviroiunental isolates and rely on a

data base which contains previously identified bacteria. Thus identification of

environmental isolates is often not possible.

Analysis of cell wall constituents has been used to identify bacteria.

Fouier transform infrared spectroscopy (h i IK) gives a spectral fingerprint based

on the cellular composition of the microorganism. Molecular bonds have specific

vibrational and rotational movements of their atoms which are detectable with

h i IK. The molecular composition of a bacterium wiU generate a specific spectral

pattern which can be compared to other known spectral patterns for

identification. The advantages to this approach are its ease of use and the

reproducibility of the spectra. The significant disadvantage is the need for the

development of a database of known organisms and their spectral patterns for

identification of unknown samples.

Flow cytometry and fatty acid analysis work in much the same way as

FTIR. A drawback of flow cytometry is the requirement of labeling the cell with

a fluorochrome by antibody-antigen interaction. This is similar to probing in that

a substance must be bound to the cell for identification. The labeled sample is

passed through a spectrum of light that will detect the fluorochrome. Fatty acid

methyl ester (FAME) analysis by gas chromatography (GC) does not require

labeling of the cell of interest (Guckert et al., 1991). Rather, isolates are grown on

a specific medium under controlled incubation time and temperature. The cells

are harvested and fatty acid extracts are prepared by saponification (Hemming,

1996; Mayberry and Lane, 1993). The resulting extract is passed through a gas

chromatograph and the chromatograph is compared to known isolates for

identification. Although the results are specific, proper sample preparation and

a database of known samples is required.

Molecular methods With the introduction of poljonerase chain reaction

(PGR) technology there has been the development of methods that take

advantage of the uniqueness of DNA sequences in various organisms for their

detection and identification in environmental samples (Pillai et al., 1991; Tsai and

Olsen, 1992; Steffan and Atlas, 1991). PGR is an enzymatic reaction amplifying

DNA with a pol)anerase enzyme derived from the Pol I enzyme of Thermus

aquatiais (Saiki et al., 1988). The standard PGR reaction uses two oligonucleotide

primers that are designed to target a specific region of interest. The PGR reaction

is cycled through temperature changes for denaturation, annealing and

extension.

Universal DNA primers have been designed to amplify the small suburut

of the ribosome (16S) for Eubacteria (Weisbvu"g et al., 1991). This methods differs

from the hybridization method noted above in that the DNA gene sequence is

the target compared to the small ribosomal subunit itself. An amplified 16S

rDNA PGR product is then sequenced and submitted to a data base for

identification. Ribosomal sequence work has the largest library of sequence data

for identifying organisms, but many sequences isolated from the environment

are not present (Larson et al., 1993). Many of the rDNA sequences in the data

base are from organisms that have been previously identified. Phylogenetic

relationships for uiudentified species are made by comparing sequence

information and assigning a bacterial species to a division or subdivision based

on phylogeny described by Woese and coworkers (Neefe et al., 1993). These

phylogenetic relationships can be used to interpret the diversity of extracted

DNA.

24

Repetitive extragenic palindromic (REP) elements and enteric repetitive

intergenic consensns (ERIC) sequences are repeat sequences interspersed

throughout the genomic DNA that were first found in the enteric bacteria E. coli

and S. typhimurium (Versalovic et al., 1991). REP sequences are thought to play a

role in the folding of the chromosome. Oligonucleotide primers for PCR have

been designed for the amplification of fingerprints based on these repetitive

sequences (Versalovic et al., 1991). The advantages of REP and ERIC PCR are the

generation of specific unique banding patterns because the location of these

sequences are different ^Nathin each genome of a particular species. Other

studies have shown that consensus sequences exist in other bacteria as well (de

Bruijn, 1992).

Unlike REP and ERIC, random amplified polymorphic DNA (RAPD) and

arbitrarily primed PCR (AP-PCR) do not require knowledge of specific sequence

information for designing primers (Williams et al., 1990; Welsh and McCleUand,

1990). Ten base oligonucleotides of arbitrary sequence are used with a low

annealing temperature to amplify fingerprints. It appears that plasmid DNA

does not affect banding patterns. Studies of Bacillus and Yersinia have shown

amplification of heat lysed cells before and after curing plasmids resulted in the

same banding pattern (Brousseau et al., 1993; Makino et al., 1994). This

technique may be useful in differentiating organisms that previously were

distinguished based on a physiologic trait due to the plasmid.

The primary advantage of AP-PCR is that there is no need for sequence

information when choosing primers to use. There are several problems that need

to be addressed when working vvith arbitrary primers. The single primer may

not generate a banding pattern that distinguishes two species, thus separate

amplifications with two or primers more may be required to identify organisms.

Also, the amount of DNA in the reaction can vary the banding pattern generated

and needs to be kept constant. Comparison between labs has shown that

banding patterns are not totally reproducible even when DNA, primers, and

optimized conditions were used (Penner et al., 1993) Due to the use of the short

primers, the reaction conditions must be optimized and maintained to prevent

artifactual variation (Jutras et al., 1995).

Many of the PCR based fingerprinting techniques, AP, REP, ERIC, etc.,

rely on the creation of a "library" of known organisms for comparison. Data

bases or libraries of known fingerprints need to be established for each type of

PCR used. AU the methods of fingerprinting would benefit from better

resolution and sensitivity which can be obtained with HPLC or capillary

electrophoresis (Marlowe et al., 1997).

Possible Environmental Constraints To Biodegradation Of Petroleum Compounds In Soil

26

When an underground storage tank leaks petroleum contaminants into

the environment there is the potential for microorganisms to mineralize the

contaminants to carbon dioxide and water. There are many biological, chemical,

and physical factors which will determine if bioremediation will occvir. If the

leak occurs below the surface, typically, these areas are paved virtually sealing

off the plume from oxygen. Anaerobic biodegradation of petroleum products

does occur, but it is a much slower process than aerobic biodegradation (Vogel

and Grabic-Galic, 1986). This section addresses the microbiologic aspects that

may hinder biodegradation processes in the soil environment.

Biotic and abiotic factors can positively or negatively influence the

biodegradation capacity of bacterial populations in the environment. Biotic

factors include microbial interactions of competition for the contaminant, and

enzyme specificity for the biodegradation. The bacterial populations that have

the metabolic capability to utilize the petroleum compound for energy wiU

increase while other bacterial populations will diminish due to toxicity or remain

at stationary levels (Atlas et al., 1991; Foght et al., 1987).

Bacterial degraders of low to moderate molecular weight aliphatic

compounds (Cio to C24) and single ring aromatics are easily isolated from soils

because these organisms have mono and dioxygenase enzymes for catabolic

degradation (Atlas, 1988). Monooxygenase is prevalent in several bacteria so

acclimation to and mineralization of these compounds is relatively fast. The lack

27

of previous exposure to more complex compounds will slow the rate of

biodegradation until organisms are acclimated or adapted to the compounds

(Cemiglia and Heitkamp, 1989; Leahy and Colwell, 1990; Zilli et al., 1993).

Recalcitrant byproducts that form as the result of partial breakdown of

contaminants may not be degraded by the microbial populations present.

Limited numbers of microorganisms in the enviroriment wiU restrict

biodegradation. Soil is spatially variable with micro-environments that are

vastly different within a few millimeters (Brock, 1971). Microbial counts

decrease with depth into the vadose zone where available carbon and nutrients

are limiting. Areas in the soil with low bacterial populations and oligotrophic

conditions may require a longer period of acclimation before biodegradation

proceeds. The presence of a compound may not be enough to induce enzyme

expression. Along with a carbon source, metabolizing microorganisms need

nutrients for incorporation of the metabolites into biomass. Soils, particularly in

the vadose zone, can be low in essential nutrients and may require the addition

of nutrients to enhance microbial biodegradation. The spatial variability of soil

makes it difficult to assess nutrient availability.

Abiotic factors such as soil pH outside of the range of 7.5 - 7.8 (Dibble and

Bartha, 1979), solubility of the contaminant, and hydrology of the area of

contamination are important factors that influence the biodegradability of

petroleum products. The main limiting factor for biodegradation of petroleum

28

compounds in soil is the availability of oxygen (Atlas, 1984; Borden et al., 1995;

Pepper et al., 1996; Zhou and Crawford, 1995). Aerobic microorganisms that can

easily degrade these compoimds decrease the available oxygen rapidly. The

resulting anoxic conditions may leave a large proportion of the original

contaminant to be degraded anaerobically. Diffusion processes are too slow to

replace the needed oxygen even a few centimeters below the siurface and results

in a persistence of the compounds (Gardiner et al., 1979, Pepper et al., 1996).

Although biodegradation is slow anaerobically for petroleum compounds, it is

possible particularly for the BTEX compounds (Vogel and Grabic-Galic, 1986).

The soil can act to protect the biodegradation of petroleimi compounds.

The viscosity of the petroleum and the porosity of the soil will determine the

depth of infiltration. Porosity will also limit microbial contact if the compound

can move into micro pores, less than 2 nm, thus away from microbes that are

typically larger than 2 |xm. Biodegradation occurs at interfaces; soil/compound,

water/compound, or air/compound. Bacteria are attached to soil particles at the

soil water interface. The compound may be attached to soil particles or move

with the soil water. If the compound is moving with the soil water, it can be

removed from the site of biodegradation. Sorption may either stimulate or

inhibit in situ biodegradation in soils. It is thought that sorption removes the

compound from the soil water in effect decreasing the concentration.

Microorganisms degrade compounds in solution rather than solids. The use of

29

surfactants can help dissolve and desorb the compounds making them readily

available to microorganisms (Zhang and Miller, 1992).

Toxicity can limit growth of microbes and therefore slow biodegradation

rates. Aliphatic compounds that are transformed into aliphatic acids (the acid

being more soluble than the hydrocarbon) and alcohols can interfere with

membrane integrity (Atlas, 1991). Short chain alkanes, less than CIO, can

intercalate the membrane making it more fluid, thus leaky. Aromatics can

inhibit growth rates and substrate consumption due to membrane interactions.

The solubility of the compound is proportional to the level of toxicity. The

buildup of toxic intermediates may occur and limit overall biodegradation.

Dissertation Format

This dissertation contains two manuscripts as appendices. Appendix 1 is

a report on the performance of a field-scale biofilter for the removal of gasoline

contaminants from a vapor stream. Appendix 2 is a report on the comparison of

conventional cultural and molecular methods used to analyze the bacterial

populations in the biofilter. Differences in formatting of the individual

manuscripts reflects the requirements of the different journals to which each

manuscript was submitted.

Each experiment has been primarily designed, implemented, and

interpreted by the candidate. Drs. I. L. Pepper and R. M. Miller are co-authors of

30

the papers and served to advise, but not design the candidates research. Cecil B.

Smart and Richard Rupert additional co-authors on the manuscript in Appendix

1, modified and adapted methods for gas vapor sampling and gas

chromatography analysis, but did not contribute to the development or

implementation of the microbial methods or analysis used in this research.

Each candidate for the advanced degree of Doctor of Philosophy in the

Department of Soil, Water and Environmental Science is expected to submit her

or his original research to peer-reviewed scientific journals for publication. By

using this alternative format, the manuscript in Appendix A has been published

in the journal Biodegradation, the manuscript in Appendix B is ready for

publication in the Journal of Microbial Ecology, and the manuscript in Appendix

C is ready for publication in the Journal of Microbiological Methods. Preparation

and submission of the manuscripts was undertaken by the candidate, Drs. I. L.

Pepper and R. M. Miller edited the manuscripts.

CHAPTER 2

PRESENT STUDY

The methods, results, and conclusions of this study are presented in the

papers appended to this dissertation. The following is a summary of the most

important findings in these papers.

Approximately 15000 L of unleaded gasoline were released into the

surrounding vadose zone from a leaking undergroimd storage tank. Initial

remediation was by soil vapor extraction and combustion which soon became

cost prohibitive, as added propane was required to reach the combustion limit of

the extracted vapors. As a cost effective alternative, a field-scale compost based

biofilter was used in conjunction with soil vapor extraction to remediate the

vadose zone. Results of a five month study showed that the biofilter removed

90% of total petroleum hydrocarbons (TPH) and >90% of the BTEX compounds

(benzene, toluene, ethylbenzene, xylene), achieving the stringent permit

requirements set at either 90% TPH reduction or less than 1.36 kg per day of

volatile organic compoimds (VOC's) released to the atmosphere. The bacterial

population in the biofilter was uniformly diverse throughout the biofilter,

suggesting that a consortium of bacteria was needed for efficient biodegradation.

The cost of biofilter set up and operation saved 90% in the first year alone of the

operating expenses incurred by soil vapor extraction and combustion.

Several bacteria isolated from the last sample period were further tested

in the laboratory for their ability to grow on unleaded gasoline, benzene, toluene,

ethylbenzene, and xylene (BTEX) as sole carbon sources. AP-PCR and 16S RFLP

fingerprinting were both used to distinguish isolates. AP-PCR produced unique

fingerprints from isolate DNA, but 16S RFLP fingerprints were vmable to

distinguish isolates at the species level. The AP-PCR and 16S RFLP information

was not able to distinguish bacteria in the community DNA. As a result, 16S

ribosomal sequence information was used to design DNA hybridization probes

of the isolates. The isolates chosen for probe design used either gasoline,

ethylbenzene, toluene, or xylene as sole carbon source. Dot blot hybridization

assays of community 16S rDNA that had been extracted over time from the

biofilter identified subtle shifts in the degradative bacterial consortium. By using

a combination of conventional cultural and molecular methods we were able to

track the bacterial succession of Arthrohacter spp., R. meliloti, and S. paudmobilis

that were degrading the contaminant vapor in tlie biofilter.

APPENDIX A

FIELD-SCALE BIOFILTRATION OF GASOLINE VAPORS EXTRACTED

FROM BENEATH A LEAKING UNDERGROUND STORAGE TANK

Eileen Maura Jutras^ Cecil M. Smart-, Richard Rupert\ Ian L. Pepper^ & Raina M. Miller^*

^Department of Soil, Water and Environmental Science University of Arizona

Tucson, AZ 85721.

^Greeley and Hanson 115 Broadway, Suite 13B

New York, NY 10006.

^1615 E. Monte Cristo Ave. Phoenix, AZ 85022.

'Corresponding author

Manuscript Published in Biodegradation 8:31-42

34

ABSTRACT

Approximately 15000 L of unleaded gasoline were released into the

surrounding vadose zone from a leaking underground storage tank. Initial

remediation was by soil vapor extraction and combustion which soon became

cost prohibitive, as added propane was required to reach the combustion limit of

the extracted vapors. As a cost effective alternative, a field-scale compost based

biofilter was used in conjunction with soil vapor extraction to remediate the

vadose zone. The biofilter was constructed on site using 4:1 diatomaceous

earthxomposted horse manure. Results of a five month study showed that the

biofilter removed 90% of total petroleum hydrocarbons (TPH) and >90% of the

BTEX compounds (benzene, toluene, ethylbenzene, xylene), achieving the

stringent permit requirements set at either 90% TPH reduction or less than 1.36

kg per day of volatile organic compounds (VOC's) released to the atmosphere.

The biofilter showed the capacity to readily adapt to changing environmental

conditions such as increased contaminant loading, and variations in temperature

and moisture. The bacterial population in the biofilter was uniformly diverse

throughout the biofilter, suggesting that a consortium of bacteria was needed for

efficient biodegradation. The cost of biofilter set up and operation saved 90% in

the first year alone of the operating expenses incurred by soil vapor extraction

and combustion.

35

INTRODUCTION

A recent study estimates that 300,000 soil and groundwater contamination

sites across the United States exist as the result of leaking fuel storage tanks

(Dowd, 1994). Such spills have resulted in contamination of the vadose zone,

with the potential for leaching into the saturated zone and spreading further

with groundwater flow. Among the contaminants present in significant

quantities in these sites are benzene, toluene, ethyl benzene, and xylene (BTEX);

aU U.S. EPA priority pollutants.

The site dociunented in this research is typical of many found in the

United States. A connecting pipe to an undergroxmd gasoline storage tank

ruptiired, and over a two year period released approximately 15000 L of

unleaded fuel into the surroimding vadose zone. Soil vapor extraction (SVE),

known to be an effective method for removing gasoUne from unsaturated soils,

was implemented at the site (Miller, 1992). Initially, the extracted vapors were

treated by incineration using an internal combustion engine. During this time

levels of extracted vapors were high (approximately 110,000 jag L"^), but after

approximately 3 months of treatment, the concentrations of extracted TPH were

reduced to approximately 9800 fig L"^. As treatment progressed it became

apparent that the treatment cost was becoming prohibitive. This was a result of

36

two factors. First, as the TPH concentration decreased, it became necessary to

add significant quantities of propane to achieve the vapor combustion limit.

Second, it became apparent that low levels of vapors would be extracted far

beyond the initial estimate of a two-year treatment for the site.

An alternative to incineration is biofiltration. Biofiltration uses a solid

medium to support microbial populations: Populations can, in turn, degrade the

contaminants of a waste gas stream passed through the mediiun. The use of

biofiltration for odor control and removal of volatile organics is well

documented in Europe (Leson & Winer, 1991; Atlas, 1995) and to a lesser extent

in the United States (Bohn, 1992). In this study, biofiltration was chosen as an

alternative, low cost method for treating the extracted vapors. Maricopa County,

AZ issued a permit of operation for the biofilter which required that the

following minimum standards be met; 90% reduction of contaminants or less

than 1.36 kg per day of volatile organic compounds (VCXI's) released to the

atmosphere.

Successful bioremediation of petroleum hydrocarbons is known to

depend on biotic as well as abiotic factors including temperature, moisture

content, oxygen, and concentration of contaminant. The objectives of this

research were to 1) to measure the effects of the operating parameters including

flow rate, contaminant concentration, temperature and moisture on TPH and

BTEX removal in a field scale biofilter and, 2) to enumerate the bacterial

37

population in the biofilter medium and select bacterial isolates that degrade TPH

and BTEX. Biofilter performance was evaluated over a five-month period, during

which time three influent contaminant concentrations were sequentially stepped

through four flow rates.

MATERIALS AND METHODS

Biofilter Design The biofilter was a 9.75 m (32 ft) half-cylinder (diameter, 2.44

m (8 ft)) constructed from a used underground steel storage tank. Polyvinyl

chloride (PVC) piping (Payless Hardware, Phoenix, AZ) perforated at 46 cm

intervals was placed along the length of the biofilter bottom. A 30 cm layer of

washed pea gravel (Pioneer Landscape, Gilbert, AZ) was placed over the PVC

piping. A 91 cm layer of biofilter medium [a 4:1 mixture of diatomaceous earth

celite No. 379 (Synergistic Performance Corp., Oakland, CA) and composted

horse manure with mulch (GRO-GREN Supply, Phoenix, AZ)] was placed over

the pea gravel. Finally, a 15 cm layer of a 4:1 mixture of coconut based 4 x 10

mesh activated carbon and composted horse manure was placed on top (Figure

1). The biofilter material had a pH of 7.3 (determined by the saturated paste

method. Page et al., 1982), a nitrogen content of 0.35%, total carbon content of

4.0% and organic carbon content of 3.2% (determined by high temperature

38

combustion using a Carlo Era 1500 Nitrogen, Carbon, Sulfur Analyzer, Artiola,

1990).

Water was delivered to the biofilter material using soaker hoses that were

placed at a 20 cm depth in the biofilter and the imit was watered when

necessary. Three sets of vapor collection ports (1.3 cm PVC capped at top and

bottom with perforations at either 15, 46, or 76 cm), tensiometers (Soilmoisture

Equipment Corporation, models 2725ARL06, 2725ARL18, and 2725ARL36,

Goleta, CA), and thermocouples (Thermx Southwest, S-Class 18U, San Diego,

CA) were emplaced across the length (north, center, south) of the cell at 15, 46,

and 76 cm depths (Figure 1). In order to maintain a mass balance within the

biofilter, the unit was kept as airtight as possible. This was done by covering the

biofilter unit with a heavy duty tarpaulin cover that was securely taped to the

unit sides. Three 61 cm x 61 cm pl5rwood doors were built into the biofilter cover

to provide access for sampling.

Contaminated vapors were piped from the vapor extraction unit into the

bottom of the biofilter using 5 cm PVC piping for upflow operation. A soil vapor

extraction exhauster (Invincible Air Flow Systems, Invincible model 300,

Scottsdale, AZ) was used to extract the gasoline vapors from the vadose zone.

Airflow from each extraction well and ambient air flow was controlled by an

inline manifold which was used to adjust the concentration of influent

contaminant.

Biofilter Operation The vadose zone consisted of fine- to medium-grained sand,

and the water table was approximately 65 m below the surface. Exploratory

wells had been drilled to depths of 26-28 m to define the extent of the plume, and

these were used as vapor extraction weUs. Biofilter operation was evaluated at

three different influent concentrations (500, 1200, and 2700 |ag L-^) sequentially

stepped up through four different flow rates (25, 55,75, and 95 m-^h*^). Although

the experiment was designed for twelve stepped loading rates, only nine were

actually tested, as shown in Table 1, due to breaches of the permit requirements.

Each contaminant concentration cycle was started at the lowest flow rate and

allowed to equilibrate for 7 days. Subsequent flow rates were equilibrated for 5

days. When the cycle was complete for one concentration, the influent

concentration was increased and the cycle was started again at the lowest flow

rate. For each sampling event, influent and effluent vapor samples were

analyzed as well as vapor samples from the 15,46, and 76 cm depths at the north,

center, and south sites of the biofilter (Figure 1). In addition, temperature and

moisture tension data were recorded, and biofilter medium cores were collected

from each sampling site for microbiological and gravimetric moisture content

analysis.

Due to heavy rains in the Phoenix area, the biofilter was shut down from

day 94 to day 111 because the above ground junction point of the extraction

wells became flooded with water. During this time a minimal flow of ambient

40

air, approximately 3.5-5 m^h-^, was maintained. The experiment resumed on day

111 when the flow was set to 27.3 m^h"^ and the influent vapor concentration was

adjusted to 2700 (ig L ^ The system was allowed to equilibrate for 7 days at this

setting.

Gasoline Vapor Analysis Biofilter vapor samples were collected periodically

from each sampling port by inserting a 0.31 cm diameter stainless steel rod into

one of the collection ports. The tube was cormected via tubing to a vacuum

pump. The port was flushed for approximately 20 seconds and then a Tedlar^'^'

bag (Analytical Technologies, Inc., Phoenix, AZ) was filled. Tedlar bags were

stored in a dark plastic garbage bag to inhibit photodegradation of gasoUne

components.

Vapor samples were analyzed by gas chromatography (GC) using EPA

method 8020/8015 modified volatile aromatics, including gasoline. A Hewlett

Packard Co. 5890C gas chromatograph equipped with an ultra alloy-5 (5%

phenylmethylsilicone) capillary column 30 m in length, inner diameter of 0.5

mm, and film thickness of 1.5 |am (Quadrex Corp., UA5-30V-1.5F, New Haven,

CT) was used to analyze gas samples. The carrier gas was ultra high purity

helium set at a flow rate of 30 ml min-^. Hydrocarbons were detected using a

flame ionization detector (FID) with ultra high purity hydrogen and breathing

41

quality air set at flow rates of 30 ml min-^ and 300 ml min-^ respectively. Briefly,

operating conditions were; initial temperature (T), 50° C; initial hold time, 8 min;

ramp, 12° C min*^; final T, 220° C; final hold time, 2 min; FID T, 250° C. The GC

was calibrated using benzene, toluene, ethylbenzene, xylene, and gasoline

standards (Alltech Assoc., Inc. Deerfield, IL). A 4-point calibration curve was

analyzed for BTEX and gasoline. BTEX mass standards analyzed were 25, 50,

100, and 300 ng (EPA method 8020/8015). Gasoline mass standards analyzed

were 500, 2500, 5000, and 10,000 ng (EPA method 8020/8015). An internal

standard, a,a,a - trifluorotoluene, was injected at a mass of 0.25 ^.g, and a

surrogate standard, 4,4-bromofluorobenzene was injected at a mass of 1.25 |ag.

Elimination Rate Rates of contaminant removal were determined from

influent and effluent air phase concentrations, airflow rate, and the volume of

the biofilter material using the following equation (Hodge & Devinny, 1994):

ER = (Cinn - Cefn) X Q Volume

where; ER = elimination rate, g m*^ h"^; Cinn = influent concentration g m-^; Ccff =

effluent air phase contaminant concentration, g m-^; Q = flow rate m^ h-^; Volume

= volume of biofilter packing material, m^.

Microbial Analysis Biofilter medium cores were taken using a 2.54 cm

Oakfield® probe (Oakfield, Inc.) that was sterilized with ethanol between sample

collections. Core samples were placed into Whirlpak® (Nasco) bags and stored

on ice for transport to the laboratory. Serial dilutions in 0.1% peptone were

made starting with 10 g samples of biofilter medium and used to determine

viable plate coxmts. Viable counts of biofilter microorganisms were performed

using two media, R2A® (Difco Laboratories, Detroit, MI) for total heterotrophic

counts, and mineral salts medium (MSM) with hexadecane (MSM-Cw) as sole

carbon and energy source for aliphatic hydrocarbon degraders. MSM contained

the foUowing ( L-^): Na2HP04 (4.0g), KH2PO4 (l.Og), NHiCl (l.Og), MgS04 (0.2g),

yeast extract (O.OOSg), ammonium iron(in) citrate (O.OOSg), and CaCb (0.01 g).

Hexadecane was added at a concentration of 0.1% (w^). Plates were incubated

at 25° C for 5 days and entunerated. Colonies of interest from the spread plates

of each type of media were re-streaked for isolation. Acridine orange direct

counts were made to determine total bacterial numbers (Brendecke et al., 1993).

To evaluate the ability of isolated bacteria to degrade unleaded gasoline, a

total of 27 and 28 isolates from the first and last sample day respectively were

used to inoculate 10 ml of MSM broth amended with either 0.1% or 0.01% (w^)

of unleaded gasoline. These were incubated at 25° C with shaking and growth

was indicated by turbidity. To determine degradation of BTEX, isolates were

incubated at 25° C on MSM plates in an individual atmosphere of each of the

43

BTEX compounds. Biolog® was used to identify these isolates as well as other

bacteria isolated during the experiment.

RESULTS

Gasoline Vapor Analysis Table 1 shows the removal of TPH during the course

of this study. Removal at the 15 cm depth (just below the activated carbon

layer), which represents biological removal, ranged from 74.2 to 89.4%.

However, the actual effluent which was further treated by the top 15 cm layer of

activated carbonxomposted horse manure mixture reached 88.6 to 91.1% TPH

removal, allowing permit requirements to be met. A removal of greater than

88% was achieved from day 55 (the start of effluent sampUng) through day 87.

At this point the contaminant concentration was 1440 ^g L"^ and the flow rate

was increased to 75.9 m^h"'. At this loading rate of 5.0 X10^ g m-^ h*^, only 78.5%

TPH total effluent was removed. This was considered breakthrough and the

final and highest flow rate for this concentration was not tested. Similarly, for

the highest contaminant concentration tested (2755-2785 |ig L-^) breakthrough

occurred immediately for the third flow rate that was to be tested (74.8 m^h-')

44

representing a loading rate of 7.0 XICP g h ^ As a result the experiment the

discontinued without further sampling.

The fate of TPH and the four individual BTEX components moving

through the biofilter was similar (as shown in Figures 2 and 3 which show the

removal of TPH and toluene). In general, the contaminant removal increased as

the contaminant traveled through the biofilter. However, the first four sampling

periods for the both total TPH and individual total BTEX compounds showed

>89% removal irrespective of depth. The removal of benzene, ethylbenzene, and

xylene was very similar to toluene with total removal >90% (data not shown).

Elimination rate Elimination rates were calculated to analyze the efficiency of

the biofilter in removal of TPH and BTEX over the course of this study. As

shown in Figure 4, elimination rates increased over the course of the study for all

components measured. The relationship between loading rate and elimination

rate is shown in Figxure 5. The elimination rate consistently increased as the

loading rate increased, indicating a correlation (Figxu^e 5).

Temperature and Moisture Analysis Temperature measurement showed that

although the temperature was uniform throughout the biofilter, the overall

temperature increased steadily over the course of the 5 month experiment

ranging from a low at day 55 of 11.0 ± 0.3° C to 29.1 ± 2.2° C on the final sampling

45

day. This was not surprising since the experiment ran from November until

April. Figure 6 shows a plot of the elimination rate (from the influent to 15 cm)

for TPH and toluene as a function of temperature at the 15 cm depth. In general,

elimination rate increased with temperature and regression of these data gave an

r- = 0.7. Benzene, ethylbenzene, and xylene had even less correlation yielding r^

values <0.7 (data not shown).

Soil moisture varied within the biofilter as determined by either

tensiometer readings or gravimetric analysis. The tensiometer readings were

calibrated using water retention curves determined from a sample of the

compost (Danielson and Sutherland, 1995). Gravimetric moisture

determinations ranged between 50 - 90% for all samples and fluctuated by as

much as 30% at some sample points over the course of the experiment (data not

shown).

Microbial Analysis Following construction, the biofilter was flushed with

ambient air for 35 days at 55 m^h-^ to establish and equilibrate baseline microbial

populations. Viable counts performed on Days 0, 5, 10, 16, 28, and 35 were

similar and averaged 1.3 X10^ ± 2.3 X 10^ CFU per gram of dry compost on R2A

medium. Viable counts on MSM-Cie mediimi were approximately 1/2 log lower

averaging 5.7 X 10^ ± 8.3 X 10^ CFU per gram of dry compost. Total direct

microscopic counts were approximately 2 logs higher than viable counts.

46

averaging 9.7 X 10^ ± 1.8 X 10® cells per gram of dry compost. On day 35, the

flow rate was lowered to 26.3 m%-^ and contaminant vapors were introduced at

the lowest concentration. Resxilts of periodic microbial plate coimts are shown in

Figure 7 for MSM-Cie- R2A counts showed a similar pattern but were 1/2 log

higher (data not shown). No significant difference in either viable plate counts

or in direct counts was observed between the beginning and end of this

experiment with direct counts averaging 3.3 X 10^ ± 9.6 X 10^ for the last day.

However, there were significant fluctuations in viable plate counts during the

experiment, for example MSM-C16 counts dropped by 1 log from day 28 to day

42 and then steadily increased again until Day 70, followed by a second decrease.

These fluctuations appear to be related to changes in flow rate, and therefore

loading rate, which showed similar patterns over the time course of the

experiment (Figure 6).

There were approximately 13 colony types with distinct differences in

color and morphology, isolated on R2A media during each sampling.

Differences in morphology and coloration were more limited on the MSM-Cie

media; residting in 9 - 10 distinct colony types each sampling period. Re-

streaking isolates from MSM-C16 onto R2A resulted in some colonies expressing

coloration after transfer. Gram stains were made from all isolated colonies and

showed that Gram negative and positive organisms were present in equal

numbers. The Gram stain information was used to screen the isolates further for

47

Biolog identification. Successful Biolog identification was limited. Only six

positive identifications were made from the eighteen isolates from the last

sample day; those isolates that were able to use the unleaded gasoline as the sole

carbon source. Horse manure inhibits growth of fungi, so no fungal inhibitors

were added to either media. There were some fungal colonies on the plates, but

these were not included in the viable coimts. Actinomycetes were included in

the viable coimts, and were more prevalent on the MSM-C16 medium, although

they were not the dominant population present.

Sole carbon source utilization studies of the bacterial isolates resulted in

three groupings: 1) Isolates that could use unleaded gasoline, as well as each

BTEX compound, as sole carbon source (see the top third of Table 2); 2) Isolates

that could use one or more of the compoimds (see the middle of Table 2); and 3)

Isolates that were unable to use any of the compounds as sole carbon source for

growth (see the bottom third of Table 2). In the first group, there was a total of

eleven isolates from day 0 and six from day 123. Of the eleven isolates from day

0 there were four pairs that exhibited the same morphology. One isolate had the

same morphology, on the first sample day as on the last sample day. In the

second grouping, there was a total of 26 isolates from day 0 and day 123. There

were five pairs of isolates that had similar colony morphology from the first

sample day and the last sample day. There was one pair from day 0 and a set of

five from day 123 that appeared to be the same. The last group had eight isolates

48

from day 0 and eight isolates from day 123, with two pairs that were similar

between samplings.

DISCUSSION

Effect of Influent Concentration A total of nine loading rates were tested

starting with the lowest concentration and flow rate. At the lowest

concentration, 440-500 fig L ^ 88% TPH removal occurred immediately, at the 76

cm depth (see days 55 - 70, Table 1 and Figure 2). This was likely due to high

initial rates of sorption that occurred until sorption sites were saturated. At the

two successive influent concentrations tested, 1040-1440 and 2755-2785 |ag L ',

removal of contaminant occurred at all depths. Removal ranged between 74.2

and 89.4% at the 15 cm depth, and was attributed to microbial activity.

Additional contaminant removal from 0.2 - 10.2% was achieved after flow

through the top activated carbon layer. Figure 5 shows that for the loading rates

used during this experiment the biofilter removed TPH and BTEX with equal

efficiency.

During the experiment, the two lowest measured total TPH removals

occurred at day 70 (88.6%) and day 87 (78.5%). These samples occurred at the

49

highest flow rates tested for each respective contaminant concentration, e.g., 96

m%-^ for the lowest contaminant concentration tested, and 76 m%-^ for the

intermediate contaminant concentration (Table 1). Breakthrough was also

observed at >55.4 m^h'^ for the highest concentration tested. Despite the fact that

the higher removal of TPH (Figure 5), the biofilter was not able to meet

performance requirements (90% TPH removal). These results indicate that

biofilter performance was not adequate at high flow rates.

Low flow rates also seemed to affect biofilter performance. For example,

days 55, 77 and 118 correspond to the lowest flow rate tested for each

contaminant concentration. On each of these days, biological removal (TPH

removal at 15 cm) was lower than at higher flow rates (Table 1). There are

several possible explanations for this behavior. The bacterial populations may

have required a longer acclimation period to accommodate the decreased flow

rate and simultaneous increase in contaminant concentration (Zilli et al., 1993).

In addition, Ottengraf (1986) has described the phenomena of reduced

contaminant removal, termed diffusion limitation, in his description of a

biophysical model of biofilter biofilms. At decreased contaminant levels, the

equilibrium concentration gradient formed around the medium used for

biofiltration is driven by the diffusion rate of the contaminant into the biolayer.

At low concentrations bioremediation is limited by this diffusion rate of the

contaminants into the biolayer (Ottengraf, 1986; Leson and Winer, 1991; Zilli et

50

al., 1996). Thus, at low contaminant concentration and high flow rate, residence

times are too short for diffusion. In summary, there is an intricate interplay

among the concentration, the flow rate and the bacteria that developed during

the five months of bioremediation and precluded our being able to fully evaluate

this complex field experiment.

Interestingly, the period of inoperation, from day 87 to day 111, did not

seem to diminish biofilter efficiency, which immediately retiimed to >90% when

contaminant vapors were reiatroduced. This finding concurs with other

published studies (Tang et al., 1995; Martin & Loehr 1996). Overall, flow rates

that were either too low (<50 m^h-^) or too high (<70 m^h-^) resulted in

suboptimal performance. Low flow rates resulted in decreased bacterial

populations except for the final phase of the study and high flow rates resulted

in residence times that were too short for the biodegradation of contaminant

vapors and thus, breakthrough occurred. Since the only oxygen available to the

biofilter for maintaining aerobic conditions was through the air manifold of the

influent line, a calculation was made to determine whether the amount of

available O2 in the vapor phase was sufficient for complete biodegradation. It is

known that TPH and BTEX are degraded by bacteria under aerobic conditions.

Assumptions for this calculation included: 1) 21% of vapor influent contained

oxygen for all loading rates; 2) 50% of the TPH carbon was mineralized to CO2;

51

and 3) octane (CsHis) was representative of TPH. Using these assumptions and

the relationship:

CgHis + O2 NH3 => C5H7NO2 + CO2 + H2O (TPH) (cell mass)

a mass balance was calculated for the amoimt of O2 needed for complete

oxidation of the compounds of interest. As an example, for each mole of TPH,

7.5 moles of O2 would be needed for oxidation. Based on this, there was an

excess of O2 available in the vapor phase for mineralization at all contaminant

concentrations and flow rates used. It should be noted that the diffusion of

oxygen into the liquid phase of the biofilm could also have been limiting

resulting in the development of anaerobic microsites within the compost

material.

Hodge and Devinny (1994) showed that biofiltration using activated

carbon as the filter medium had greater elimination rates than a compost that

was tested, but a lower degradation rate and buffering capacity which could

limit its long term effectiveness. Activated carbon that has been used for the

adsorption of contaminants also creates the need to dispose of contaminated

material. Column studies have shown that BTEX compounds were reduced by

at least 75% in mixed contaminant waste streams and that to improve removal a

series of biofilters may be required (Togna & Singh 1994). In this study, activated

carbon was placed over the biological component of the system to sorb any

undegraded contaminants. Composted horse manure was added to the

52

activated carbon with the intent to stimulate biodegradation of sorbed

compounds, but no evaluation of this was made during this study.

Effects of Temperature and Moisture on Bioremediation Zhou and Crawford

(1995) have shown that increasing the temperatiu^e increases the rate of

bioremediation due to the increased metabolism of microorganisms. In this

study, temperatures ranged from 11° C at the begirming to 29° C during the 85

day experiment. The increased temperature was correlated to increased

elimination rates for TPH and toluene (r^ = 0.7), but was not well correlated with

elimination rates for the other contaminants (data not shown). There are other

factors that could have influenced this result that were not considered. Changes

in temperature affects not only the microorganisms present, but also the

contaminants. As temperature increases, the amount of contaminant transferred

to the air phase also increases, but simultaneous increases in the flow rate can

decrease the transfer of contaminants resulting in diminished performance of the

biofilter (Zilli et al., 1993). While the biofilter adapted well to increasing

temperatures, preliminary studies in this system, showed that when the internal

temperature of the compost exceeded 45° C, the biofilter began to perform

poorly (data not shown). So the biofilter was shut down for July, August, and

September when daily temperatures ranged from 38 - 45° C. Such high

53

temperatures may cause increased toxicity of hydrocarbons to degrading

organisms (Atlas 1994).

Though adequate moisture is needed for microorganisms, saturated

conditions during gasoline degradation should be avoided, as degradation

proceeds more readily in aerobic conditions. The moisture content within the

biofilter fluctuated over a wide range. After an initial wetting with the soaker

hoses, the extracted soil vapor was sufficiently humid to maintain adequate

moisture in the system.

Bacterial Isolates Based on qualitative visual observation, the bacteria that

were isolated over the course of this experiment did not appear to vary

significantly. Some of these isolates were identified using Biolog, but one

problem with the use of the Biolog system is that it requires that organisms be

grown on specific t)^es of media prior to analysis. Many of the isolates did not

grow adequately on the recommended media to allow Biolog identification. For

example, of 19 isolates from sample day 123 that were able to grow with gasoline

as a sole carbon source, only 6 could be identified with Biolog. A second

example was an unusual sample dominated by yellow gram negative colonies

that was collected from the 46 cm depth on day 65. No conclusive identification

of this isolate was made.

54

Some isolates that were identified by Biolog included Rhizohium meliloti

which was isolated prior to hydrocarbon input, as weU as Rhizobium loti isolated

after contaminant was introduced. We found this interesting since composted

horse manure, which presumably contained digested alfalfa was used as the

inoculum. Rhizobia have been shown to degrade hydrocarbon compoimds

(Shimp et al., 1993). Other potential hydrocarbon degrading isolates include

Bacillus insolihis, B. licheniformis, and B. pasterii which were present at various

sample depths within the biofilter. Pseudomonas shitzerii was also prevalent

throughout the biofilter and persisted during the sampling period indicating that

these organisms were distributed within the biofilter.

The bacteria isolated and tested for their sole carbon source use showed

that there were a total of eight bacteria that appeared to be the same from the

background sample as the acclimated sample. Of those isolates that could

degrade all of the contaminants only one was the same from days 0 and 123, the

rest of the isolates from this group changed over the course of the experiment.

The numbers of bacteria isolated were similar from day 0 and 123. There were

several bacteria isolated on the same day on R2A and MSM that were the same.

Further experiments are being carried out using various molecular methods and

systematics to determine if these isolates are genetically the same.

55

Cost Comparison Table 3 shows a cost comparison of the biofiltration of

gasoline vapors compared to combustion of the said vapors at the Phoenix site

studied. As shown in this Table, even with all of the capital costs for the biofilter

imposed on the first year of operation, this technique saved 90% of the cost of

vapor combustion. Therefore expenses in subsequent years of operation should

provide even higher savings. Activated carbon was the most expensive material

cost, but was able to reduce emissions to within the permit requirements when

biological treatment alone was not sufficient.

CONCLUSIONS

Biofiltration of VOC's is a practical and cost effective method of

remediation. This field scale biofiltration system had an overall contaminant

removal of 90 - 98% for TPH and BTEX compoimds while operating over a wide

array of conditions: flow rate, contaminant concentration, temperature, and

moisture. Efficient operation of the biofilter required adequate flow rates, 50 to

70 m^h-^, allowing optimal residence times for the compounds. Contaminant

loading rate was also a factor in efficient operation of the biofilter system.

Temperatxire had a limited effect overall, though increased temperature did

correlate to an increase in elimination rate for TPH and toluene. Most

importantly, the bacterial population was uniform throughout the biofilter and

56

was active over a wide array of operating conditions. Few of the organisms

isolated were able to degrade all of the BTEX components suggesting that

optimal biofilter performance required a consortium of bacteria. The compost

based system with a final activated carbon layer successfully operated within

permit requirements with significant savings in operating costs.

ACKNOWLEDGMENTS

We thank Adria Bodour for her help in plating and identification of the biofilter

organisms.

57

REFERENCES

Artiola, JF (1990) Determination of carbon, nitrogen, and sulfur in soils, sediments, and wastes: A comparative study. Intern. J. Environ. Anal. Chem. 41:159-171.

Atlas, RM (1994) Microbial hydrocarbon degradation-bioremediation of oil spills. J Chem. Tech. Biotech. 52:149-156.

Atlas, RM (1995) Bioremediation. Chem. & Engin. News 73:32-42.

Bohn, H (1992) Consider biofiltration for decontaminating gases. Chem. Engin. Prog. 25:34-40.

Brendecke, JW, Axelson, RD, & Pepper, IL (1993) Soil microbial activity as an indicator of soil fertility: long-term effects of municipal sewage sludge on an arid soil. Soil Biol. Biochem. 25:751-758.

Danielson, RE & Sutherland, PL (1995) In: Klute, A (Ed.) Methods of Soil Analysis Part 1: Physical and Mineralogical. SSSA Inc., Madison, WI.

Dowd, RM (1994) Leaking underground storage tanks. Environ. Sci. Technol. 18:10.

Hodge, DS & Deviimy, JS (1994) Biofilter treatment of ethanol vapors. Environ Prog 13:167-173.

Leson, G & Winer, AM (1991) Biofiltration: An irmovative air pollution control technology for VOC emissions. Air & Waste Manage. Assc. 41:1045-1054.

Martin, FJ & Loehr, RC (1996) Effect of non-use on biofilter performance. J. Air & Waste Manage. Assc. 46:539-546.

Miller, ME, Pederson, TA, Kaslick, CA, & Hoag, G 1992 Soil vapor extraction column experiments on gasoline contaminated soil. USEPA/600/sr-92/170 Oct 92.

Ottengraf, SPP (1986) In: Rehm, HJ & Reed, G (Eds.) Biotechnology Vol. 8; VCH VerlagsgeseUschaft, Weinheim.

58

Page, AL, Miller, RH, & Keeny, DR. (1982) In: (Ed) Methods of soil analysis. Part 2, Chemical and microbiological, 2nd ed. Agonomy 9, ASA, SSSA Inc., Madison, WI.

Shimp, JF, Tracy, JC, Davis, LC, Lee, E, Huang, W, Erickson, LE, & Schnoor, JL. (1993) Beneficial effect of plants in the remediation of soil and groxmdwater contaminated with organic materials. Critical Rev. in Environ. Sci. and Tech. 23:41-77.

Tang, H-M, Hwang, S-J, & Hwang, S-C (1995) Dynamics of Toluene Degradation in Biofilters. Hazardous Waste & Hazardous Materials. 12:207-219.

Togna, PA & Singh, M (1994) Biological vapor-phase treatment using biofilter and biotrickling filter reactors: Practical operating regimes. Environ. Prog. 13:94-97.

Zhou, E & Crawford, RL (1995) Effects of oxygen, nitrogen, and temperature on gasoline biodegradation in soil. Biodegrad. 6:127-140.

ZiUi, M, Converti, A, Lodi, A, Del Borghi, M, & Ferraiolo, G (1993) Phenol removal from waste gases with a biological filter by Pseiidomonas putida. Biotechnol. Bioeng. 41:693-699.

Zilli, M, Fabiano, B, Ferraiolo, A. & Converti, A (1996) Macro-kinetic investigation on phenol uptake from air by biofiltration: Influence of superficial gas flow rate and inlet poUutant concentration. Biotechnol. Bioeng. 49:391-398.

59

Table 1: Total petroleum hydrocarbon (TPH) removal efficiency.

Contaminant TPff Removal TPH Removal Day Concentration Flow rate at 15 cm Total Effluent

(mg L-i) (m3h-i) (%) (%)

0 96.0 n.d.2 n.d. 28 0 96.0 n.d. n.d. 35 0 96.0 n.d. n.d. 55 470 26.3 79.6 89.4 60 500 56.1 88.4 90.0 65 480 70.2 89.4 89.6 70 440 96.0 88.6 88.6 77 1320 23.6 74.2 88.9 82 1040 52.0 87.7 89.9 873 1440 76.0 75.2 78.5 944 n.d. <5 n.d. n.d. 111 2755 27.4 n.d. n.d. 118 2755 27.4 79.8 90.0 1235 2785 55.4 86.5 91.1

^ The lower detection limit for TPH was 50 mg L-^. Each value at 15 cm is

the average of the concentration determined at the North, Center, and

South sampling sites.

- n.d. = not determined

^ Highest flow rate was not tested since permit conditions were not

achieved.

^ Due to heavy rains, the junction point of the extraction wells flooded

stopping vapor flow. The experiment was shut down for 17 days to allow

weUs to clear. At this time, the lowest flow rate with the highest influent

concentration was initiated and equilibrated for 7 days.

^ Higher flow rates were not tested.

60

Table 2: Carbon source utilization of bacterial isolates taken from Days 0 and

123.

Media and Isolate# Unleaded Benzene Toluene Ethyl- Xylene Dav 123 benzene

Xylene

MSM 4 & 6 + + + + +

MSM5&7&R2A12 + + + + +

MSM9&R2A13 + + + + +

MSM IQi MSM 5' + + + + +

R2A5&7 + + + +

R2A14 + + + + +

MSM 3 + + + + +

MSM 7 + + + + +

R2A5 + + + + +

R2A6 + + + +

R2A15 + + + +

MSMl R2A4 + - - +

MSM 3 R2A3 + - + +

R2A2 + - • . R2A4&6 MSM 1 + . - -

R2A10 + - + +

R2A9 MSM 11 + - . •

R2A17 R2A17 + - - -

MSM 2, 4, 6, R2A10 + - -

R2A11 + - + - +

R2A12 + - •

R2A16 + • •

MSM 2 & R2A 8 - - - . MSM 8 - . - -

R2A1 - . . R2A3 - • . R2An R2A9 . . R2A15 - . . •

R2A16 MSM 8 - . - . R2A1 - • • . R2A2 - . . R2A7 - . - . R2A8 - . - . MSM 10 & - •- - -

[solates listed in the same line had similar colony morphology and sole carbon source utilization patterns.

61

Table 3: Cost comparison of SVE with combustion and SVE with biofiltration^.

Combustion Biofiltration

Set up cost $1000 PVC piping $1500 diatomaceous earth $4000 activated carbon $500 composted horse

manure $250 tank^ and cutting

Monthly maintenance $200 monitoring^ $200 monitoring $17 electricity $17 electricity $5000 internal combustion

engine' $2000 propane

Year 1 of operation $86,604 $9854 Monthly maintenance $200 monitoring $200 monitoring

$17 electricity $17 electricity $5000 internal combustion $2000 propane

Subsequent years of $86,604 $2604 operation

^Does not include setup costs for extraction wells and assumes that these would be equal for each type of remediation.

2The tank for this experiment was previously used. ^Cost of monitoring 2 samples per month. ^Assumes that an engine would be rented. A VR Systems V-3 internal

combustion engine retails for approximately $75,000.

South Canlir North

9.75 m

Lateral View

End View

dmnnqt not to scait

Figure IrSchematic of the biofilter layout with lateral and end views.

15 cm 4:1 activated carbon:composted horse manure.

91 cm 4:1 diatomaceous earth:composted horse manure.

30 cm pea gravel.

63

100

130

Day

Figure 2: The percent removal of TPH from the 15, 46, and 76 cm depths as

averaged from three sample sites within the biofilter. The total

effluent removal, and error as indicated. The lower detection limit for

TPH was 50 |j,g L .

64

<u 60

130

Figure 3: The percent removal of toluene from the 15, 46, and 76 cm depths as

averaged from three sample sites within the biofilter. The total

effluent removal, and error as indicated. The lower detection limit for

toluene was 0.5 |ag L"^.

I

65

1e+1

1e+0 -

E o> 0) 2 ie-1 c o 13 c J LU

1e-2 -J

1e-3

^ ... v\\.

V

V

CT^

50 I

60 70 80

Day

I TPH •

^ Toluene A

c>^ Ethylbenzene v

-Cfyp^ Benzene • , Xi o Xylene O

<r"

V A-T-120 130

Figxire 4: Total elimination rate for all contaminants; TPH and BTEX.

I

66

•ID

cc. c O

1 e + 1

1 e + 0 ^

) e - 1 ^

1 e - 2 - s

1 e - 3

5 0 0 ^ g L

1 e + 0 1 e + 1 1 e + 2

L o a d i n g R a t e g m ' ^ h '

1 e + 3 1 e + 4

1 e + 1

1 e + 0 —

1 e - 1 ^

1 e - 2

1 2 0 0 u g L

1 e + 1 1 e + 2 1 e + 3 1 6 + 4

L o a d i n g R a t e g m " ' h

Figure 5: Elimination rate versus loading rate for the low, 500 |xg L ^ and medium, 1200 ixg L'^, influent contaminant concentrations. Experimental sample values are shown with symbols: • TPH, •

benzene, A toluene, V ethylbenzene, 0 m&p-xylene, and o-xylene; regression as the line, and r^ as indicated. The low r^ for benzene at the medium concentration, 1200 |ig L*^, resulted from the total effluent removal showing a zero (Fig.5). Regressions were not determined for the highest concentration because there were only two data points taken before breakthrough ocairred. The linear regression equations used for all contamiaants were:

y = (9.5e - 4 ± 4.9e - 5) x + (1.8e - 3 ± 4.3e - 3) at 500 ng L-' y = (8.3e - 4 ± 1.9e - 4) x + (6.5e - 2 ± 1.2e - 1) at 1200 ng L '

67

1e+8 60-00- -0-0-1

£

1 1e+7 -I CD O !o 'a>

Li-1e+6 -

••

6-CK>-'

o-

CH

1e+5 ' 1 ' ' ' ' I ' 20 40

I 60 80

O

Q

- 1 — I — I — I — 1 — 1 — — r

100

I—I 1 1

120 130

r 90

r 80

- 70 ^ E 0}

-60 "S Q: : ? r50 §•

7 40 O

r 30

- 20

10

Day

Figure 6: Viable plate counts of bacteria isolated on MSM-C16 medium from

samples collected from the north 46 cm depth compared to the flow

rate over the course of the experiment.

68

E

•o a> E

o o

"Q U. o a> O) CO W a> > <

1 e + 8

1 e + 7 9 •.

1 e + 6

1 e + 5

1 e + 4 O-l

1 e + 3

1 e + 2

1 e + 1 I " ' • I " ' ' I " ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I • ' ' ' I ' ' y 0 10 20 30 40 50 60 70 80

D a y

E

T5 0) E

o A O U. O o •> to W o > <

1 e+8

1 e + 6

1 e + 5

1 e + 4

l e + 2

1 e + 1

0 10 20 30 40 50 60 70 80

O—I

1 e + 2

o CO

K S o LL.

i ' ' '

1 2 0 1 3 0

1 e + 1

1 e + 4

- i e + 3

<70) 0) <Q Q: O) c <0 o

X a .

^<-7—-1 2 0 1 3 0

1 e + 2

D a y

Figure 7: Average viable plate counts of bacteria isolated on MSM-C16 medium

from samples collected from the north biofilter sampling site.

A: Compares the plate counts with flow rates used.

B: Compares plate counts with loading rates used.

Depths as indicated and error bars are shown.

69

APPENDIX B

COMBINED CULTURAL AND NUCLEIC ACID BASED METHODS TO

FOLLOW COMMUNITY DEVELOPMENT IN A BIOFILTER

Eileen Maura Jutrast, Raina M. MiUer, and Ian L. Pepper* University of Arizona, Department of SoU, Water and Environmental Science,

Tucson, AZ 85721

'Corresponding author: University of Arizona Department of Soil, Water and Environmental Science 429 Shantz Building, #38 Tucson, AZ 85721 ipepper@ag.arizona.edu

tPresent address: Metropolitan Water District of Southern CaUfomia 700 Moreno Avenue La Verne, CA 91750 ejutras@mwd.dst.ca.ua

Intended for publication in Applied and Environmental Microbiology

70

ABSTRACT

A soil biofilter was constructed to treat gasoline vapor collected from a

subsvu*face site contaminated with gasoline. Vapors were vented through the

biofilter for a period of 118 days, during which time bacterial populations

degraded BTEX components. At designated sample times, bacterial isolates

capable of degrading gasoline and/or BTEX components were culturally

isolated. Attempts were made to identify isolates using the Biolog system, but

this system, which is clinically based, only identified 6 out of 18 environmental

isolates. In contrast, 16S rDNA identification using a combination of PCR

amplification, cloning and sequence analysis, resulted in identification of 16 of

the 18 isolates. Dominant species included Arthrobacter, Rhodococais and

Flavobacterium. Several probes were designed from selected isolates to foUow

isolate population development during the start-up and 118-day operation of the

biofilter. To accomplish this, probes were designed from isolate 16S rDNA

sequences and used to analyse community DNA obtained over time from soil

samples collected within the biofilter. Visualization of isolate population

development within the biofilter was achieved with dot blot hybridization. This

study demonstrates that a combination of cultural and nucleic acid based

methodologies can be used to follow specific population development in a

biofilter.

71

DMTRODUCnON

Microbial ecologists are challenged to find methods that will enable them

to better understand the distribution and diversity of bacteria in the

environment. Two of the general approaches used to study bacteria include

cultural and moleciilar methods. However, it is documented that conventional

cxiltural methods often detect less than one percent of bacteria in soil (Amann et

al., 1995; Roszak and Colwell, 1987; Ward et al., 1990). In addition, when serial

dilution and plating are used to isolate bacteria from environmental samples

they do not elucidate community structure. Cultural methods used for

identification purposes may require a combination of assays including Gram

staining, enzyme reactivity, and carbon source utilization, but these phenotypic

based assays can still result in the misidentification of bacteria (Smith-Vaughan

et al., 1995).

In contrast, molecular methodologies based on nucleic acid sequences

give researchers the potential to study bacteria without culturing allowing a

broader section of communities to be studied. A representative bacterial

communit}'^ DNA sample can be obtained from environmental samples by direct

lysis of bacterial cells in si hi, followed by extraction and purification of the DNA

(More et al., 1994; Picard et al., 1992; Tsai and Olson, 1992; Xia et al., 1995).

Polymerase chain reaction (PGR) is a sensitive and specific assay which can

detect the presence of specific bacteria within a community sample by the use of

72

species specific primers (Knabel and Crawford, 1995; Picard et al., 1992; Pillai et

al., 1991; Tsai and Olson, 1992). Alternatively, the small ribosomal subunit

associated with bacteria can be PCR amplified. These 16S rDNA sequences are

found in all known bacteria and contain phylogenetic information about bacteria

(Amann et al., 1995; Bams et al., 1994; Weisburg et al., 1991). The non-conserved

sequence regions of the 16S rDNA make it possible to design species specific

DNA probes that can be hybridized to community DNA from environmental

samples via southern blots or dot blots (Boivin-Jahns et al., 1996; Giovarmoni et

al., 1988). The advantage of using dot blot hybridization is that it may be

possible to use the relative signal intensity to quantitate the abimdance of the

species of interest within a particular sample. Thus the prevalance of an

individual species within different environmental samples can be evaluated,

provided equal amoimts of community nucleic adds are fixed to the membrane

(Ausubel et al., 1996).

Previously we used a field-scale biofilter to evaluate the effect of flow rate

and contaminant concentration on biodegradation efficiency of unleaded

gasoline vapors removed from the subsurface by soil vapor extraction (Jutras et

al., 1997). In this prior study, cultural methods used to determine viable bacterial

counts over the course of the experiment showed only slight differences in total

numbers of cidturable bacteria at the various sample sites within the biofilter

Qutras et al., 1997). Thus, although we h3^othesized that specific populations of

73

bacteria within the community were changing in response to bioremediation of

the contaminant vapors, cultural methods alone were not able to detect these

changes either quantitatively or qualitatively.

The objective of this study was to develop a method based on a

combination of cvdtural and molecvdar techniques to foUow specific bacterial

populations within the biofilter bacterial community. Our approach was to first

culturally isolate bacteria capable of degrading gasoline or individual BTEX

components. Next we compared identification of the isolates by a culturable

method (Biolog^w) ^^jth 16S rDNA sequencing. Finally, following definitive

identification we designed DNA probes for selected isolates based on small

ribosomal subunit sequence information, and used these probes to estimate the

relative abundance of each isolate within communit}' DNA via the use of dot blot

hybridizations.

MATERIALS AND METHODS

Bacterial Isolates The biofilter used in this study is diagrammed in Figure 1.

The biofilter was composed of a mixture of 80% diatomaceous earth that was

used as the support medium, and 20% composted horse manure as a nutrient

source and the bacterial inocxilum (see Jutras et al., 1997 for details). Over the

course of the study the biofilter was operated at different influent concentrations

74

of gasoline vapors, either 500,1500, or 2500 |j.g L-1, in combination with four flow

rates 25, 50, 75, or 90 m^ h*^. As shown in Table 1, total petroleum hydrocarbon

(TPH) removal ranged from 78.5 to 91.5% over the course of the study.

During biofilter operation the dominant culturable populations were

isolated from the established community in the biofilter compost medium by

serial dilution and plating as described previously (Jutras et al., 1997). These

bacterial isolates were screened for their ability to grow on gasoline or individual

BTEX components (benzene, toluene, ethylbenzene, or xylene) as a sole carbon

soxirce as described previously (Jutras et al., 1997). Ultimately, eighteen isolates

were obtained that degraded one or more of these contaminants. These isolates

were subjected to cultural and nucleic acid based methodologies to determine

their identity and follow their prevalence during biofilter operation. In addition

to cultural analysis of biofilter samples, soil samples were subjected to

community DNA extraction. Community DNA samples were analyzed with

constructed species specific probes to follow population development with time

in the biofilter. The cultural and molecular methodologies used in this study are

detailed in the following sections.

Biolojg Identification The TPH-degrading isolates obtained from the

biofilter were designated 1-18 and were analyzed cidturally using the Biolog

(Hayward, CA) system. This system, which uses carbon substrate oxidation

75

patterns to determine bacterial identity, was used following the manufacturer's

recommended protocols.

Nucleic Acid Based Identification Isolate DNA Extraction Isolates 1-18

were grown in 5 ml of R2A broth (based on the composition of Difco R2A agar,

Difco, Detroit, MI) L-1: yeast extract 0.5g, proteose peptone #3 0.5g, casamino

acids 0.5g, dextrose 0.5g, soluble starch 0.5g, sodiiun pyruvate 0.3g, potassium

phosphate, dibasic 0.3g, and magnesixmi sulfate 0.05g. Chromosomal DNA was

extracted from turbid cultures with a proteinase K and phenol chloroform

isoamyl alcohol (25:24:1) extraction (Ausubel et al., 1996).

Isolate 165 rDNA PCR Pol5Tnerase chain reaction with the primers

forward 5' AGA GTT TGA TCC TGG CTC AG 3' and reverse 5' ACG GTT ACC

TTG TTA CGA CTT 3' (synthesized by IDT, Inc., Coralville, IW) were used to

amplify a 1.5 kb fragment of Eubacterial 16S rDNA (Weisburg et al., 1991). The

optimized reaction conditions were as foUows: A mixture of the forward and

reverse primers, 100 pM, 200 jxM dNTP's (Pharmacia Biotec, Uppasala, Sweden),

IX Pfu buffer (Stratagene Inc., LaJoUa, CA), 2.5 U of Pfu DNA polymerase

enzyme (Stratagene Inc., LaJoUa, CA), and 10-50 ng of template DNA. The

GenAmp'^'^ PCR System 9600 (Perkin Elmer Cetus Corp., Norwalk, CT) was

programmed for the following amplification conditions: One cycle at 95° C for 1

76

min denaturation, followed by 30 cycles of 94° C 15 sec denaturation, 55° C 15

sec annealing, 72° C 30 sec extension, and a final 72° C 2 min extension cycle.

Amplification of 16S rDNA of an isolate results in a bulk ribosomal sequence that

can be used for identification (Cilia et al., 1996; Laguerre et al., 1996).

Cloning and Sequencing of Bacterial Isolates The 16S rDNA PCR

product for isolates 1, 2, 3, 4, 6, 7, 8,10,11,14,15, and 18 was cloned directly into

Stratagene pCR-Scripf^'^ SK(+) plasmid (Stratagene, La Jolla, CA). The 16S

rDNA PCR product for isolates 5, 9, 12, 13, 16, 17 were cloned directly into

Invitrogen pCR 2.1^" plasmid (Invitrogen, San Diego, CA). Four other known

bacterial isolates were used as positive controls and were also cloned. Two of

these known isolates had previously been isolated from sugar cane roots and

identified by fatty acid methyl ester analysis (FAME) at a commercial laboratory.

These were cloned and sequenced from the pCR 2.1 plasmid (Myma Sevilla,

Department of Plant Pathology, University of Arizona, personal

communication). The other two positive controls were ATCC isolates E.coli

15224 and Micrococais liiteiis 381, which were cloned into the Invitrogen plasmid.

After transformation, white colonies that were selected from LB, IPTG, X-gal,

and ampicillan plates were used to inoculate 25ml of SOC broth. Plasmids were

extracted using the QiaFilter midiprep system (Qiagen Inc., Chatsworth, CA).

Sequencing was performed using the ABI377 automated sequencer by personnel

77

of the Laboratory of Molecular and Systematic Evolution (LMSE, Tucson, AZ)

using the M13 forward and reverse primers resulting in sequences of both

strands. Whole, approximately 1500 base ribosomal sequences were assembled

using the Genetics Computer Group (GCG) (Genetics Computer Group Inc.,

Madison, WI) Sequence editor and Bestfit programs. The program FastA was

used to search for homology to other sequences in the GenBank database, and

used to identify isolates.

78

DNA Probe Design The 16S sequence information of the dominant isolates that

degraded either gasoline or one or more of the BTEX compounds (Table 2) were

chosen for DNA probe design. The ARTH probe was based on the 16S rDNA

sequence of isolate 6, Arthrobacter polychromogenes, which used gasoline, or xylene

as a sole carbon soiu^ce. The FVBL probe was from isolate 9, Flavohacterium

halustinutn, which used gasoline, toluene or xylene as sole carbon sources. The

RSP probe was from isolate 3, a Rhodococais sp., which was able to use gasoline

or all of the BTEX compounds individually as a sole carbon source. The SPEP

probe was based on the 16S rDNA sequence from isolate 4, Sphingomonas

paucimohilis, which used gasoline or xylene as a sole carbon source.

For the design of probes, the GCG program PUeUp was used to align all of

the isolate 16S nucleic acid sequences, and areas of non-conserved sequences

were used for probe design. Putative probes were then submitted to GenBank

using FastA to search for homology within the database to ensure that the

chosen sequences would hybridize to intended target sequences. In addition, a

Euhacterial (EUB) probe sequence was obtained from the literature (Kampfer et

al., 1996). This sequence is theoretically conserved among all known bacteria. In

all cases, probes were then synthesized at the University of Arizona.

Isolate Dot Blot Hybridization In order to test the probes, DNA isolated from

a variety of bacteria was subjected to dot blot hybridizations. Oligonucleotide

79

probes s5mthesized by IDT (Coralville, IW) were end labeled with a^2p[ATP]

(Easytides, MEN, Dupont, Boston, MA). Sephadex™ G-50 (Pharmacia Biotech,

Uppasala, Sweden) spin columns were used to remove unincorporated label. A

96 well vacuvun manifold was used to blot 5 |j,l of the isolate 16S rDNA PCR

reaction onto Hybond'^^^N+ positively charged nylon membranes (Amersham

Inc., Arlington Heights, IL). Prewash and hybridizations were performed at 42°

C for 30 and 60 min respectively, in Rapid-hyb'^'^ (Amersham Inc., Arlington

Heights, IL) hybridization buffer. Wash conditions were 5X SSC, 0.1% SDS at

25^ C for 20 min followed by two washes with IX SSC, 0.1% SDS at 42*^" C for

15 min. The dot blots were autoradiographed at -80 C for 1-3 days using

BIOMAX'TM film (Eastman Kodak, Rochester, NY).

Community DNA Direct Lysis Extraction Compost filter medium samples

were collected from nine sites within the biofilter over a five month time period

with various gasoUne vapor loading rates (Jutras et al., 1997). Bacterial

commimity DNA was extracted from compost samples by directly lysing

bacterial cells in sihi using the proteinase K, CTAB method of Xia et al. (1995)

with modifications to the final purification step. Briefly, 4 g (wet weight) of

compost medium was lysed at 65° C in an 100 nxM EDTA (Sigma Chemical, St.

Louis, MO), 500 mM Tris (FisherBiotech, Fairlawn, NJ) buffered solution, pH 8.0

which contained 2.5% sodiiun dodecyl sulfate (SDS, FisherBiotech, Fairlawn, NJ)

80

and 0.2 mg nil-l proteinase K (Promega, Madison, WI). Cell debris and sediment

was cleared by centrifugation and the supernatant transferred to a fresh tube for

incubation with 1% hexadecyltrimethylammoniimi bromide (CTAB,

FisherBiotech, Fairlawn, NJ). An equal volume of chloroform isoamyl alcohol

(24:1) (FisherBiotech, Fairlawn, NJ) was added to precipitate the protein-CTAB

complex. Ammoniimi acetate (Sigma Chemical, St. Louis, MO) at a final

concentration of 1.5 - 2.5 M was used to precipitate humic acids that remained in

solution. After removing the humic acids, the DNA was ethanol precipitated

and a phenol chloroform isoamyl alcohol (25:24:1) extraction was used as the

final purification of the DNA (Ausubel et al., 1996).

Community 16S rPNA Amplification The commxmity DNA PCR reaction

conditions were the same as those used for the isolate amplification. In all cases

10 ng of community DNA was used for template in the PCR reaction.

Community Dot Blot Hybridization Community 16S rDNA samples were

arranged on the membrane so that each sample site within the biofilter could be

evaluated over time. Day 0 was prior to the introduction of vapors to the

biofilter (a loading rate of 0 gm-^h-^ Table 1). Day 55 corresponds to the

introduction of vapors at the lowest flow rate and concentration, thus a loading

rate of 562 gm-^h V Community blots were made for the lowest and highest flow

81

rate tested at each of the three contaminant concentrations used. The dot blot

conditions were the same as those described above for the isolate DNA.

RESULTS AND DISCUSSION

Isolate Identification Only six isolates, 4, 5, 9, 12, 13, and 18 could be

identified using Biolog, and of these, only three were similarly identified using

16S sequence homology (Table 2). It should be noted that the Biolog data base is

based upon clinical isolates and contains very few environmental isolates, so

these results are not surprising. However, the Biolog system is easy to use and

provides information on the metabolic capabilities of a bacterial isolate. Of the

three isolates similarly identified by Biolog and 16S rDNA sequencing, isolate

number 13 was identified to the genus level as Pseudomonas sp. (Biolog) and was

identified to the species level as P. mendocina using molecxdar methods. Foght et

al. (1996) also reported P. mendocina to be a common petroleum degrader,

although initially P. shitzeri was predominant. However, as degradation

continued and succession took place, P. mendocina became the dominant

bacterium. This isolate was able to use unleaded gasoline, benzene, toluene,

ethylbenzene, or xylene as a sole carbon source. Isolate 4 was also identified

similarly by Biolog and 16S rDNA sequencing as Sphingomojias paucimobilis

although it was only able to utilize gasoline and xylene. Finally, isolate 18 which

82

utilized all the carbon sources tested was identified by both methods as

Variavorax paradoxus.

Of the remaining 15 isolates, all but two were able to be identified by 16S

rDNA sequence homology. Identification showed that Arthrobacter,

Flavobacteriiim and Rhodococais were the dominant genera. These identifications

were based on the highest percent similarity scores which were 0.90 or better to

be considered an accurate match. For isolates 2, 6, and 15 there was only a tenth

of a percent difference between the highest similarity scores for identifications as

either A. oxydans, A. polychromogenes, or A. ilicis. In taxonomic studies of these

organisms, A. oxydans and A. polychromogenes have almost identical 16S

sequences, the same cell wall types, menaquinone compositions, and DNA G+C

contents (Koch et al., 1995). A. ilicis has a slightly lower DNA G+C content as

well as its 16S sequence being more related to A. oxydans and polychromogenes

than A. globiformis (Koch et al., 1994; Koch et al., 1995). Isolates 5 and 17 could

not be sequenced successfully even though several attempts were made to

sequence the 16S amplicon from these isolates. One of these isolates, #5, was

identified as Bacillus insolitus by Biolog.

The results of the GenBank data base search for homology to the 16S

rDNA sequence of the positive controls used in this study are shown in Table 3.

The positive controls, (two ATCC cultures and two environmental samples

83

identified by FAME were positively identified by the 16S rDNA sequenciig

method used in this study.

Probe Construction and Isolate Dot Blot Hybridization The sequences of all

probes used in the study are shown in Table 4. As a control, the 16S rDNA fi-om

the eighteen bacterial isolates was hybridized with the individual DNA probes.

The data fi-om GenBank showed that all probe sequences were theoretically

specific to their intended target. However, although hybridization conditions

were optimized for each probe, two probes, Flcwobacteriiim (FVBL) and

Rhodococcus (RSP) showed cross reactivity with several non-target sequences

(data not shown). The Eubacterial (EUB) probe, as expected, hybridized to all the

isolates (data not shown). The Arthrobacter (ARTH) probe targeted nucleotide

(nt) sites 1008-1031 of the small subunit of the ribosome and this probe

hybridized only to the isolates that had been identified as being Arthrobacter,

namely, 2, 6, and 15 (Figure 2A). Isolate 8 which was also identified as A.

globiformis is more distantly related to these organisms and has a different

peptidoglycan composition and is the likely reason the ARTH probe did not bind

to this isolate (Koch et al., 1994; Koch et al., 1995; Stackebrandt et al., 1995). The

Spliingomonas (SPEP) probe targeted the nt region 640-666 of the small subiuiit.

SPEP bound to isolate number 4 which had been identified as a Sphingomonas,

but also to isolate 14 which had been identified as Rhizobium meliloti (Figure 2B

84

and Table 2). A review of the literature did not reveal a phylogenetic placement

of Rhizobium nieliloti in comparison to S. paucimobilis and it was beyond the scope

of this work to create such a placement. These two isolates both used xylene as a

sole carbon source and were the only two isolates in this study classified as

Eubacteria; Proteobacteria; alpha subdivision.

Community Dot Blot Hybridization The EUB, ARTH and SPEP probes were

used to analyze community DNA extracted from biofilter samples. Of the

samples extracted, there were seven community samples that did not result in

PCR product formation (noted in Figs. 3 - 5). The EUB probe of the commvmity

samples resulted in strong hybridization intensities for all sample sites (Figure 3).

These strong, similar intensities compared favorably with enumeration data that

showed total heterotrophic (R2A) bacterial counts were 10^-10^ CFU g •

throughout the biofilter operation Qutras et al., 1997). The EUB probe results

suggest that bacteria were evenly dispersed among the sample sites within the

biofilter for each sample period.

In contrast to the EUB probe, the other hybridization probes, which

targeted specific populations, resulted in differing hybridization signal

intensities (Figs. 4 and 5). The ARTH probe was specific to isolates identified as

either A. oxydans or A. polychromogenes, isolates 2, 6, and 15 (Fig. 4). The overall

intensity of the signal strength for the ARTH probe was less than that of the EUB

85

probe confinning that Arthrobacter spp. were only a subset of the bacteria present

within the biofilter (Fig. 4). The ARTH probe hybridized to all of the community

samples from Day 0 and showed that Arthrobacter spp. continued to be present

throughout the biofilter operation. The hybridization signal from the probe

targeted to Arthrobacter sp. showed that while it was not the most dominant

population (compare the hybridization signal intensity of Figs. 4 and 5), these

bacteria remained as a stable component of the community throughout

biofiltration and they were present at all depths. Samples from the south (S) and

north (N) sample sites at the 46 cm depth along with the center (C) 76 cm depth

showed greater intensity compared to the other sample depths for Day 0 (Fig. 4).

Samples from C76 increased in signal intensity until Day 118. All the sample

sites had a diminished signal intensity for the last sample Day 123 although

removal of TPH remained high (91.1%). Wantanabe and Hino (1996) showed

that during bacterial succession, adaptation of microbes to a substate(s) peaks

early and that populations begin to dedine as stable populations compete for

available carbon sources.

Results from commvmity hybridization with SPEP were quite different

than with ARTH. There was no hybridization signal on Day 0 (prior to

contaminant introduction to the biofilter) for any of the samples (Fig. 5) although

microbial counts showed 10^ - 10^ bacteria present. The signal intensity

increased for many of the samples over the time course of the experiment. For

86

example, the center sample at the 76 cm depth (C76) signal strength increased

starting with the introduction of contaminant vapors on Day 55. This

corresponded to an increase in bacterial numbers and high removal of TPH

(Jutras et al., 1997). Hybridization signals were the strongest for Days 87 and 118

at sample sites S15, S76, C46 and C76. On these days (87 and 118) high xylene

removal occurred with values of 95.6 and 79.2% respectively (Jutras et al., 1997).

For the last sample period. Day 123, the signal intensity was diminished again for

aU sample sites within the biofilter. The results of the SPEP probe derived from

Sphingomoncis paiidmobilis, which was able to degrade gasoline and the typically

more recalcitrant xylene, were notable. Sphingomoncis spp. have been shown to

degrade a wide variety of aromatic compounds particularly xylene derivatives

(Fredrickson et al., 1995).

Overall, dot blot hybridizations using the 16S rDNA probes resulted in

confirmation that specific species capable of BTEX degradation were present in

the bacterial community within the biofilter over time. In this experiment, the

hybridization signal intensity could not be used to quantitate the number of

bacteria present because there are multiple copies of the 16S gene in bacteria. For

example, there are three copies in Rhizohium spp., seven in Escherichia coli and up

to ten copies in Bacillus suhtilis (Huber and Selenka-PobeU, 1994; Hill and

Hamish, 1981; Loughney et al., 1982).

87

CONCLUSIONS

In this study conventional cultural methods were needed to identify the

metabolic capabilities (i.e., sole carbon utilization) of bacteria and obtain the

ribosomal sequence information used in the design of DNA probes. The Biolog

system was only modestly successful in identifying bacteria, whereas the 16S

rDNA nucleic acid sequence information did result in bacterial identification.

Using a combination of cultural and nucleic acid based methods we were able to

assess changes in specific bacterial populations as bioremediation took place.

There are several advantages to using hybridization of community samples to

16S rDNA probes for tracking bacterial populations. The bacterial community to

be probed does not need to be cultured eliminating the biases of that method.

Several samples can be screened at one time and as used in this study, samples

taken over a period of time can be evaluated for changes. Several probes can be

used to assess various bacterial populations that make up the community and

identify which organisms are present and their relative abimdance. The

hybridization signal intensity can be related to the removal of contaminant when

the metabolic capabilities of the bacteria are known. Further work in the area of

probe design will enhance the utility of the method for tracking populations of

environmental bacteria that are important for the biodegradation of petroleum

compounds.

88

ACKNOWLEDGMENTS

We thank Myma Sevilla for supplying the bacteria isolated from sugar

cane roots and Drs. Christina Kermedy and Rita Colnagi, and Andy Green who

provided valuable advice on the cloning and hybridization techniques.

89

REFERENCES

Amann, R.L, W. Ludwig, and K.-H. Schleifer. 1995. Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiological Reviews 59:143-169.

Ausubel, F. M., R. Brent, R. E. Kingdom, D. D. Moore, J. A. Smith, S. G. Sideman, and K. Strul (Eds.). 1996. Current protocols in molecular biology. John Wiley & Sons, Inc., New York.

Bams, S.M., R.E. Fundyga, M.W. Jefferies, and N.R. Pace. 1994. Remarkable archaeal diversity in a Yellowstone National Park hot spring environment. Proceedings of the National Academy of Science USA 91:1609-1613.

Boivin-Jahns, V., R. Ruimy, A. Bianchi, S. Daumas, and R. Christen. 1996. Bacterial diversity in a deep-subsurface clay environment. Applied and Envirorunental Microbiology 62:3405-3412.

Cilia, v., B. Lafay, and R. Christen. 1996. Sequence heterogeneities among 16S ribosomal RNA sequences, and their effect on phylogenetic analysis at the species level. Molecular Biology and Evolution 13:451-461.

Fredrickson, J.K., D.L. BalkwiU, G.R. Drake, M.F. Romine, D.B. Ringelberg, and D.C. White. 1995. Aromatic-degrading Sphingomonas isolates from the deep subsurface. Applied and Environmental Microbiology 61:1919-1922.

Foght, J.M., D.W.S. Westlake, W. Johnson, and H.F. Ridgeway. 1996. Environmental gasoline-utilizing isolates and clinical isolates of Pseudomonas aeriginosa are taxonomically indistinguishable by chemotaxinomic and molecular techniques. Microbiology 142:2333-2340.

Giovaimnoni, S.J., E.F. de Long, G.J. Olsen, and N.R. Pace. 1988. Phylogenetic group- specific oligodeoxynucleotide probes for identification of single microbial cells. Journal of Bacteriology 170:720-726.

Hill, C.W. and B.W. Hamish. 1981. Inversion between ribosomal RNA genes of Escherichia coli. Proceedings National Academy of Science 78:7069-7072.

Huber, L. and S. Selenka-Pobell. 1994 Pulsed-field electrophoresis-fingerprinting, genome size estimation and rm loci number of Rhizobium galegae. Journal of Applied Bacteriology 77:528-533.

Jutras, E. M., I. L. Pepper, and R. M. Miller. 1997. Field scale biofiltration of gasoline vapors extracted from beneath a leaking undergrovmd storage tank. Biodegradation 8:31-42.

Kampfer, P., R. Erhart, C. Beimfohr, J. B5hringer, M. Wagner, and R. Amaim. 1996. Chciracterization of bacterial commimities from activated sludge: Culture-dependent niunerical identification versus in situ identification using group- and genus-specific rRNA-target oligonucleotide probes. Microbial Ecology 32:101-121.

90

Knaebel, D.B. and R.L. Crawford. 1995. Extraction and purification of microbial DNA from petroleum-contaminated soils and detection of low numbers of toluene, octane and pesticide degraders by multiplex polymerase chain reaction and Southern analysis. Molecular Ecology 4:579-591.

Koch, C., F.A. Rainey, and E. Stackebrandt. 1994. 16S rDNA studies on members of Arthrobacter and Micrococcus'. An aid for their future taxonomic restructuring. FEMS Microbiology Letters 123:167-172.

Koch, C., P. Schumann, and E. Stackebrandt. 1995. Reclassification of Micrococcus agilis (Ali-Cohen 1889) to the genus Arthrobacter agilis comb. Nov. and emendation of the genus Arthrobacter. International Journal of Systematic Bacteriology 45:837-839.

Laguerre, G., P. Mavinguli, M.-R. AUard, M.-P. Chamay, P. Louvrier, S.-I. Mazurier, L. Rigottier-Gois, and N. Amarger. 1996. Typing of Rhizobia by PGR DNA fingerprinting and PCR-restriction fragment length poljmiorphism analysis of chromosomal and symbiotic gene regions: Application to Rhizobium leguminosarm and its different biovars.

Loughney K., E. Lund, and J.E. Dahlbiurg. 1982. TRNA genes are found between the 16S and 23S rRNA genes in Bacilhis subtilis. Nucleic Acids Research 10:1607-1625.

More M. I., Herrick J. B., Silva M. C., Ghiorse W. C., and Madsen E. L. 1994. Quantitative cell lysis of indigenous microorganisms and rapid extraction of microbial DNA from sediment. Applied and Envirorunental Microbiology 60:1572-1580.

Picard C., Poisonnet C., Paget E., Nesme X., and Simonet P. 1992. Detection and enumeration of bacteria in soil by direct DNA extraction and

polymerase chain reaction. Applied and Environmental Microbiology 58: 2717-2722.

Pillai, S.D., K.L. Josephson, R.L. BaQey, C.P. Gerba, and LL. Pepper. 1991. Rapid method for processing soil samples for polymerase chain reaction

amplification of specific gene sequences. Applied and Environmental Microbiology 57:2283-2286.

Roszak, D.B. and R.R. Colwell. 1987. Survival strategies of bacteria in the natural environment. Microbiological Reviews 51:365-379.

Smith-Vaughan, H.C., K.S. Sriprakash, J.D. Mathews, and D.J. Kemp. 1995. Long PGR- Ribotjqping of nontypeable Haemophilus influenzae. Journal of Clinical Microbiology 33:1192-1195.

Tsai Y.-L. and Olson B. H. 1992. Rapid method for direct extraction of DNA from soil and sediments. Applied and Environmental Microbiology 57:1070-1074.

91

Watanabe, K. and S. Hino. 1996. Identification of a functionally important population in phenol-digesting activated sludge with antisera raised against isolated bacterial strains. Applied and Environmental Microbiology 62:3901-3904.

Ward, D.M., R. Weller, and M.M. Bateson. 1990. 16S rRNA sequences reveal numerous xmciiltured microorganisms in a natural community. Nature (London) 345:63-65.

Weisburg, W. G., S. M. Barns, D. A. Pelletier, and D. J. Lane. 1991. 16S RibosomalDNA amplification for phylogenetic study. Journal of Bacteriology 173:697-703.

Xia, X., J. BoUinger, A. Ogram. 1995. Molecular genetic analysis of the response of three soU microbial communities to the application of 2,4-D. Molecular Ecology 4:17-28.TabIe 1: The seven sample periods, corresponding contaminant loading rates and total petroleum hydrocarbon (TPH) removal efficiency compared in the hybridization study.

92

Table 1: The seven sample periods, corresponding contaminant loading rates

and total petroleum hydrocarbon (TPH) removal efficiency compared

in the hybridization study.

Day TPH influent Loading rate (gm-%-^) TPH Removal (%)

concentration

°0 0°^gTi 0 n.d.i

~55 500 ng L-i 562 89.4

70 1919 88.6

~fj 1500 ig L-i 1417 88.9

87 4974 78.5

~VL8 2500 ng L-i 3425 90.0

123 7011 91.1

^ n.d. = not determined

Table 2: Identification of isolates with Biolog or 16S sequence and their sole

carbon source use.

Isolate Biolog Identification 16S Identification Carbon Source^

1 nd'' Flaoobacterium indolitheticum G

2 nd Arthrobacter oxydans G, T, E,

3 nd Rhodococais sp. 16S rDNA G, B, T, E, X

4 Sphingomonas Sphingomonas paticimobilis G, X paiicimobilis

5 Bacillus insolihis 26°C nd G, B, T, E, X

6 nd Arthrobacter polychromogenes G, X

7 nd Micrococais roseus G

8 nd Arthrobacter globiformis G, E, X

9 Pseudomonas stutzeri Flavobacterium balustinum G, T, X

10 nd Promicromonospora citrea G, B, T, E, X

11 nd Klebsiella sp. G, B, T, E, X

12 Rhizobiiim loti Cytophaga diffliiens G

13 Pseudomonas sp. Pseudomonas mendocina G, T, E, X

14 nd Rhizobium meliloti G, X

15 nd Arthrobacter oxydans G, E

16 nd Rhodococcus sp. 16s rDNA G, B, T, E, X

17 nd nd G, E, X

18 Variovorax paradoxus Variaoorax paradoxus G, B, T, E, X

G = gasoline, B = benzene, T = toluene, E = ethylbenzene, X = xylene

No identification made

94

Table 3: The results of 16S rDNA sequence analysis for positive controls and two

unknown samples that were also identified by fatty acid methyl ester

(FAME) analysis.

Sample ATCC Identification 16S rDNA sequence identification

Positive #1 Escherichia coli Escherichia coli

Positive #2 Micrococcus luteiis Micrococcus hiteiis

FAME Identification

8M Spfiingomonas paiidniobilis Sphingomoms pauciniobilis

PALS Acetobacter diazotrophiais Acetobacter diazotrophicus

Table 4: The sequence of probes used for hybridization study.

Isolate Probe Sequence 5' - 3' Target 5'-3'

RSP" ACT CAC AGC TCA ACT GTG AGC

CTG CAG GCG

615-645

4 SPEPi' err TGA GAG TGG ATT GCT TGA

ACATC

640-666

6 ARTH*^ GAA GCG GTA ATA GCT GGA GAG

AG

1008-1031

9 FVBL'' CTT AAA TGG GAA TTG AGA GGT

TT

1004-1027

AU EUB^ GCT GCG TGG GGT AGG AGT 1022-1040

•• Rhodococcus ''Sphingomonas'^ Arthrobacter Flavobacterium ® Eubacterial

96

South Center North

15 cm

46 cm

76 cm

Activated carbon Icomposted horse manure

Diatomaceous earth composted horse manure

Pea gravel

Vapor influent

Lateral View Drawing Not To Scale

Figure 1: Schematic of the biofilter showing the location of sample sites for

microbial analysis.

>57

A

B

F-igure 2: Dot blot hybridization of isolate 16S rDNA. The numbers indicate the

isolate.

A. The hybridization probe was based on the non-conserved 16S

ribosomal sequence of Arthrohacter, ARTTi. The DNA probe bound

only to isolates 2, 6, and 15 which were identified as being of the genus

Artlirohacter (see Table 3).

B. The hybridization probe was based on the non-conserved 16S

ribosomal sequence of Sphingcmonas, SPEP. The DNA probe bound

only to isolates 4 and 14. Both of these isolates used either gasoline or

xylene as a sole carbon source.

'>8

Sample sites within the biofilter

S15 S46 S76 C15 C46 C76 N15 N46 N76

Day 0

Day 55

Day 70

Day 77

Day 87

Day 118

Day 123

Figure 3: Dot blot hybridization of community 16S rDNA with Eubacterial

(EUB) DNA probe. The sample sites within the biofilter are as

follows: Sample sites, S = south; C = center; N = north; sample

depths, 15 = 15 cm; 46 = 46 cm; 76 = 76 cm. The sample day and

corresponding loading rate for gasoline vapor are noted. indicates

samples that did not produce an amplification product.

Day 0

Day 55

Day 70

Day 77

Day 87

Day 118

Day 123

Figure 4: Dot blot hybridLzation of community 16S rDNA with Arthrobacter

(ARTH) DNA probe. The sample sites within the biofilter are as

follows: Sample sites, S = south; C = center; N = north; sample

depths, 15 = 15 cm; 46 = 46 cm; 76 = 76 cm. The sample day and

corresponding loading rate for gasoline vapor are noted. indicates

samples that did not produce an amplification product.

Sample sites within the biofilter

SIS S46 S76 CIS C46 C76 N15 N46 N76

100

Sample sites within the biofilter

S15 S46 S76 C15 C46 C76 N15 N46 N76

Day 55

Day 70

Day 77

Day 87

Day 118

Day 123

Figure 5: Dot blot hybridization of community 16S rDNA with Sphingomonas

(SPEP) DNA probe. The sample sites within the biofilter are as

follows: Sample sites, S = south; C = center; N = north; sample

depths, 15 = 15 cm; 46 = 46 cm; 76 = 76 cm. The sample day ^nd

corresponding loading rate for gasoline vapor are noted. indicates

samples that did not produce an amplification product.

101

APPENDIX C

A COMPARISON OF ARBITRARILY PRIMED PGR AND 16S RFLP TO

DIFFERENTIATE ENVIRONMENTAL BACTERIAL ISOLATES

Eileen Matira Jutras*t, Raina M. MiUer, and Ian L. Pepper University of Arizona, Department of Soil, Water and Environmental Science,

Tucson, AZ 85721

* Corresponding author t Present address; Metropolitan Water District of Southern California

700 Moreno Ave. La Verne, CA 91750 ejutras@mwd.dst.ca.us

Intended for publication in the Toumal of Microbiological Methods

102

ABSTRACT

Arbitrarily primed PGR (AP-PCR) and 16S rDNA restriction fragment

length polymorphism (RFLP) analysis were used to generate polymorphic

differences among various bacterial isolates as a basis for differentiation. Results

showed that AP-PCR fingerprinting discriminated between isolates, but

provided no information as to isolate identity. In contrast, 16S rDNA RFLP

analysis clustered the isolates into genus-specific groups, and while 16S RFLP

did not discriminate between isolates, subsequent analysis of 16S rDNA nucleic

acid sequence information did result in identification to the species level.

103

iNTRODUCnON

Since the inception of arbitrarily primed polymerase chain reaction (AP-

PCR), one use of the method has been as a diagnostic tool for the differentiation

of bacterial strains (Welsh and McClelland, 1990; McClelland et al., 1995;

Williams et al., 1990). The major advantage of AP-PCR in comparison to other

PCR methods is that prior sequence information about the target template is not

needed for the amplification of a diagnostic banding pattern or fingerprint. The

amplified products range in size and number and differences in the banding

patterns occur between genomes due to the different distances between primer

binding sites.

Another method that has been used to generate a diagnostic fingerprint is

actually a combination of two molecular methods: PCR amplification of the

small ribosomal subunit (16S rDNA) and restriction fragment length

polymorphism (RFLP). This combination (16S RFLP) has been used to

differentiate bacterial isolates important to the degradation of xenobiotics in the

environment (Jayarao et al., 1991; Schmidt, 1994).

Both of these types of fingerprinting methods, AP-PCR and 16S RFLP, can

be used to determine if environmental bacteria isolated via conventional

methods are, in fact, the same or different organisms. Use of these methods

eliminates the need to rely on phenotypic characteristics for identification.

104

Although AP-PCR and 16S RFLP have been used to identify organisms of

interest, their use requires a known reference standard of previously identified

isolates to interpret the resulting fingerprints.

The objective of this study was to compare two molecular diagnostic

methods, AP-PCR and 16S RFLP, and determine the utility of each method to

distinguish between bacteria isolated from envirorunental samples. 16S rDNA

sequencing was also used for the identification of bacteria.

MATERIALS AND METHODS

Bacterial Isolates The 18 bacterial isolates used in this study were isolated

from a biofilter compost medium by serial dilution and plating as described

previously Qutras et al., 1997).

DNA Extraction Isolates 1-18 were grown in 5 ml of R2A broth (based on the

composition of Difco R2A agar, Difco, Detroit, MI) L"^: yeast extract 0.5g,

proteose peptone #3 0.5g, casamino acids 0.5g, dextrose 0.5g, soluble starch 0.5g,

sodium pyruvate 0.3g, potassium phosphate, dibasic O.Sg, and magnesium

sulfate 0.05g. Chromosomal DNA was extracted from turbid cultures with a

105

proteinase K and phenol chloroform isoamyl alcohol (25:24:1) extraction

(Ausubel et al., 1996).

AP-PCR A single primer was used in each AP-PCR reaction to amplify

fingerprints of the isolated DNA as optimized previously (Jutras et al., 1995).

The five single primers used for amplification were: 615 5' CGT CGA GCG G 3';

620 5' TTG CGC CCG G 3'; 623 5' TGC GGG ACT G 3'; 625 5' CCG CTG GAG C

3'; 682 5' CTG CGA CGG T 3'. The AP-PCR reaction reagents were used mth

the following concentrations; IX buffer (3.5 mM MgClj, 10 mM Tris-HCl, 50 mM

KCl and 0.1 mg ml' gelatin), 200 nM dNTP, 0.4 jiM primer, 2.5 U AmpliTaq®

(Perkin-Elmer Cetus, Norwalk, CT) polymerase enzyme, and 10 ng of template

DNA. The Perkin Elmer GeneAmp^^ 9600 PCR System was programmed with a

94° C 15 sec denaturation step, a 45° C 15 sec annealing step, and 72° C 30 sec

extension step for a total of 35 cycles. The amplified fragments were

electrophoresed in 3% Metaphor^" (FMC, Rockland, ME) agarose gels at 2.7 V

cm' in IX TBE buffer. For AP-PCR analysis, a negative control was included and

contained all the reaction reagents except for the template; sterile deionized

water was substituted for template.

16S rPNA PCR Polymerase chain reaction with the primers forward 5' AG A

GTT TGA TCC TGG CTC AG 3' and reverse 5' ACG GTT ACC TTG TTA CGA

106

cm 3' (s)mthesized by IDT, Inc., CoralviUe, IW) were used to amplify a 1.5k base

fragment of Eubacterial 16S rDNA (Weisburg et al., 1991). The optimized reaction

conditions were as follows: A mixture of the forward and reverse primers, 100

pM, 200 nM dNTP's (Pharmacia Biotec, Uppasala, Sweden), IX Pfu buffer

(Stratagene Inc., Lajolla, CA), 2.5 U of Pfu DNA pol5nnerase enzyme (Stratagene

Inc., Lajolla, CA), and 50 ng isolate template DNA. The GenAmp^** PCR System

9600 (Perkin Elmer Cetus Corp., Norwalk, CT) was programmed for the

following amplification conditions: One cycle at 95° C for 1 min denaturation,

followed by 30 cycles of 94® C 15 sec denaturation, 55° C 15 sec annealing, 72° C

30 sec extension, and a final 72° C 2 min extension cycle.

165 RFLP Following 16S rDNA PCR amplification, 5 (xl of the reaction was

electrophoresed in a 1% SeaKem^"^ (FMC, Rockland, ME) agarose gel for

verification of product formation. After reviewing the literature, four base

recognition restriction enzymes (endonucleases) were chosen (Akopyanz et al.,

1992; Jayarao et al., 1991; Schmidt, 1994) and 25 ^1 of the amplicon was used for

restriction digest. The enzymes that gave consistently repeatable RFLP's were

Aiul, Mspl, Rsal, Sau3A, and TaqI (Promega, Madison, WI). The digested 16S

rDNA fragment was electrophoresed in 3% Metaphor™ (FMC, Rockland, ME)

agarose gel at 2.7 V cm • in IX TBE buffer.

107

16S rPNA Sequencing: The 16S rDNA PCR product for isolates 1, 2, 3,4, 6, 7,

8,10, 11,14, 15, and 18 was cloned directly into Stratagene pCR-Script^"*^ SK(+)

plasmid (Stratagene, La JoUa, CA). The 16S rDNA PCR product for isolates 5, 9,

12, 13, 16, and 17 was cloned directly into Invitrogen pCR 2.1^'*^ plasmid

(Invitrogen, San Diego, CA). After transformation, white colonies were selected

from LB, IPTG, X-gal, and ampicillan plates and inoculated into 25 ml cultures in

SOC meditun. Plasmids were extracted using the QiaFilter midiprep system

(Qiagen Inc., Chats worth, CA). Sequencing was performed using the ABI 377

automated sequencer by personnel of the Laboratory of Molecular and

Systematic Evolution (LMSE, Tucson, AZ) using the M13 forward and reverse

primers resulting in sequences of both strands. Whole, approximately 1.5 kb

ribosomal sequences were assembled using the Genetics Computer Group

(GCG) (Genetics Computer Group Inc., Madison, WI) Sequence editor and

Bestfit programs. The program FastA was used to search for homology to other

sequences in the GenBank database.

RESLTLTS

AP-PCR The chromosomal DNA extracted from eighteen bacterial isolates

was amplified by AP-PCR. Five separate reactions, each using a different single

108

primer, were used to generate diagnostic fingerprints of each isolate DNA. The

results of AP-PCR reactions for all 18 isolates are shown in Figure 1. This figure

shows the results from primer 682 as representative DNA fingerprints. AP-PCR

resulted in unique fingerprints for each of the eighteen isolates and confirmed

that environmental bacteria of the same genus, but different species do in fact

produce unique fingerprints (compare Figure 1 with Table 1). For example,

isolates 3 and 16 which were identified ordy as Rhodococcus sp. based on 16S

sequence data (Table 1), had distinctly different AP-PCR fingerprints confirming

that AP-PCR can distinguish between environmental isolates on the species level

(Brousseau et al., 1993; Jutras et al., 1995; Makino et al., 1994).

16S RFLP The 16S RFLP patterns generated for the eighteen bacterial isolates

resulted in similar banding patterns for several isolates thus less differentiation

between individual isolates. Table 2 shows that image analysis of the patterns

generated by the five restriction enzymes used resulted in four groupings

clusters. The clusters were similar for 4 of the enzymes: Alul, Mspl, Rspl, and

Taql. In contrast, 5aii3A banding patterns were unique for 16 out of the 18

isolates tested. Although 4 clusters were observed, many of the isolates had

unique banding patterns. For example, the banding patterns for isolates 1 and 4

were consistendy unique when digested with the five restriction enzymes used

in this study. Of the four clusters identified, cluster one consistently contained

109

isolates 2, 6, 7, 8, and 15. These isolates were identified as the genus Arthrobacter

except for number 7 which was identified as Micrococcus roseus (Table 1). These

bacteria have very sixnilar sequence homology for the 16S sequence (Koch et al.,

1995). Isolates 3 and 16 also gave the same fingerprint consistently for the

restriction enzymes used in this study except SaiiSA. These two isolates were

both identified as Rhodococciis sp. (Table 1). Isolates 9 and 13 had the same

banding pattern for Alul, Mspl and Rspl enzyme digests, but unlike the

previously mentioned clusters there did not appear to be any significance to the

similarity between fingerprints as number 9 was identified as Flaoohacteriuni

halustinum and number 13 as Pseudomonas mendocina.

16S rPNA Sequence Identification Nucleic acid sequence homology

identification of isolates 1 - 18 is shown Table 1 with the accession numbers

obtained from GenBank. These identifications were based on the highest percent

similarity scores, which were 0.90 or better to be considered an accurate match.

DISCUSSION

AP-PCR and 16S RFLP are very rapid techniques for laboratory diagnosis

and differentiation of bacterial isolates. However, the type of information

no

obtained from each method is limited to a presence/absence scoring of banding

patterns (Burr and Pepper, 1997). For AP-PCR, the presence of a banding pattern

that is the same for two isolates suggests that these isolates are the same species.

In contrast, for 16S RFLP, if the fingerprint is the same for two organisms

then this does not necessarily indicate that the isolates are the same species. The

16S rDNA sequences have conserved restriction sites, resulting in banding

patterns that clustered closely related organisms (i.e. the same for organisms of

the same genus), even when the organisms were different species. Thus, there is

not enough variability in the sequence data to differentiate various strains within

a species (I^guerre et al., 1996). While 16S rDNA RFLP analysis did not generate

unique banding patterns, it may be possible to generate unique RFLP banding

patterns by PCR amplification of the entire ribosomal operon followed by

restriction digest (Ibrahim et al., 1996 and Smith-Vaughan et al., 1995). For

example, Acinetobacter and Francisella, have been successfully differentiated using

this approach.

To provide information concerning the identity of an isolate, the 16S

rDNA PCR product can be used to obtain the nucleic acid sequence information.

In this study we reconstructed the whole ribosomal sequence to search for

homologous sequences in the GenBank database which resulted in similarity

scores that were greater than 0.90 for identification. Although AP-PCR can be

used to discriminate between isolates, a known sample must be amplified with

I l l

the same reaction conditions to make an identification. The construction of a

database of fingerprint patterns does not seem possible though because banding

patterns are subject to artifacts when different thermal cyclers are used (Permer

et al., 1993; Schweder et al., 1995). 16S RFLP on the other hand should result in

the same patterns from the restriction digest of the small ribosomal subunit thus

allowing the creation of a database that could be used for identifications.

AP-PCR or 16S RFLP of a bacterial community DNA sample should result

in diagnostic fingerprint of a mixture of bacteria (Malik et al., 1994). Although it

would seem that using moIecxUar techniques that either do not rely on specific

sequence information, AP-PCR, or are targeted for a universal sequence, such as

16S RFLP, would be ideal for community DNA analysis, AP-PCR and 16S RFLP

provide only qualitative information that populations are changing (or remain

stable) when applied to community DNA (data not shown). A Limitation

imposed by both AP-PCR and 16S RFLP for community analysis was the lack of

quantitative interpretations of the commvmity fingerprints.

1 1 2

REFERENCES

Akopyanz, N., N.O. Bukanov, T.U. Westblom, and D.E. Berg. 1992. PCR based RFLP analysis of DNA sequence diversity in gastric pathogen Helicobacter pylori. Nucleic Adds Research 20:6221-6225.

Ausubel, F. M., R. Brent, R. E. Kingdom, D. D. Moore, J. A. Smith, S. G. Sideman, and K. Strul (Eds.). 1996. Current protocols in moleoilar biology. John Wiley & Sons, Inc., New York.

Brousseau, R., A. Saint-Onge, G. Prefontaine, L. Masson, and J. Cabana. 1993. Arbitrary primer polymerase chain reaction, a powerful method to identify Bacillus thuringiensis serovars and strains. Applied and Environmental Microbiology 59:114-119.

Burr, M.D. and I.L. Pepper. 1997. Variability in presence-absence scoring of AP-PCR fingerprints affects computer matching of bacterial isolates. Journal of Microbiological Methods 29:63-68.

Ibrahim, A., P. Gemer-Smidt, and A. Sjostedt. 1996. Amplification and restriction endonuclease digestion of a large fragment of genes coding for rRNA and a rapid method for discrimination of closely related pathogenic bacteria. Journal of Clinical Microbiology 34:2894-2896.

Jayarao, B. M., J. J. E. Dore, Jr., G. A Baimibach, K. R. Matthews, and S. P. Oliver. 1991. Differentiation of Streptococcus uteris from Streptococais parauberis by polymerase chain reaction and restriction fragment length polymorphism analysis of 16S ribosomal DNA. Journal of Clinical Microbiology 29:2774-2778.

Jutras, E. M., R. M. Miller, and I. L. Pepper. 1995. Optimization of AP-PCR for the identification of bacterial isolates. Joiunal of Microbiological Methods 24:55-63.

Jutras, E. M., I. L. Pepper, and R. M. Miller. 1997. Field scale biofiltration of gasoline vapors extracted from beneath a leaking underground storage tank. Biodegradation 8:31-42.

1 1 3

Koch, C., P. Schumann, and E. Stackebrandt. 1995. Reclassification of Micrococcus agilis (Ali-Cohen 1889) to the genus Arthrobacter as Arthrobacter agilis comb. nov. and emendation of the genus Arthrobacter. International Journal of Systematic and Bacteriology 45:837-839.

Laguerre, G., P. Mavinguli, M.-R. Allard, M.-P. Chamay, P. Louvrier, S.-I. Mazurier, L. Rigottier-Gois, and N. Amarger. 1996. Typing of Rhizobia by PGR DNA fingerprinting and PCR-restriction fragment length pol5nnorphism analysis of chromosomal and symbiotic gene regions: Application to Rhizobium leguminosann and its different biovars.

Makino, S.-I., Y. Okada, T. Maruyama, S. Kaneko, and C. Sasakawa. PCR-based random amplified polymorphic DNA fingerprinting of Yersina pseiidotuberadosis and its practical applications. Journal of Clinical Microbiology 32:65-69.

Malik, M., J. Kain, G. Pettigrew, and A. Ogram. 1994. Purification and molecular analysis of microbial DNA from compost. Journal of Microbiological Methods 20:183-196.

McGlelland, M., F. Mathieu-Daude, and J. Welsh. 1995. RNA fingerprinting and differential display using arbitrarily primed PGR. Trends in Genetics 11:243-246.

Permer, G.A., A. Bush, R. Wise, W. Kim, L. Domier, K. Kasha, A. Laroche, G. Scoles, S.J. Molnar, and G. Fedak. 1993. Reproducibility of random amplified polymorphic DNA (RAPD) analysis among laboratories. PGR Methods and Applications 2:341-345.

Schmidt, T.M. 1994. Fingerprinting bacterial genomes using ribosomal RNA genes and operons. Methods in Molecvdar and Cellular Biology 5:3-12.

Schweder, M.E., R.G. Shatters S.H. West, and R.L. Smith. 1995. Effects of transition interval between melting and annealing temperatures on RAPD analysis. BioTechniques 19:38-42.

Smith-Vaughan, H.C., K.S. Sriprakash, J.D. Mathews, and D.J. Kemp. 1995. Long PGR-Ribotyping of nontypeable Haemophilus influenzae. Journal of Clinical Microbiology 33:1192-1195.

1 1 4

Weisburg, W. G., S. M. Bams, D. A. Pelletder, and D. J. Lane. 1991. 16S Ribosomal DNA amplification for phylogenetic study. Journal of Bacteriology 173:697-703.

Welsh, J. and M. McClelland. 1990. Fingerprinting genomes using PCR with arbitrary primers. Nucleic Adds Research 18:7213-7218.

Williams J.G.K., A.R. Kubelik, K.J. Livak, J.A. Rafalski, and S.V. Tingey. 1990. DNA polymorphisms amplified by arbitrary primers useful as genetic markers. Nucleic Acids Research 18; 6531-6535.

1 1 5

Table 1: Identification of isolates based on 16S sequence with GenBank accession

numbers.

Isolate 16S Identification Accession number

1 Flavobacterium indolitlieticum M58774

2 Artlirobacter oxydans X83408

3 Rliodococciis sp. 16S rDNA X80616

4 Sphingotuoms paiicimobilis X94100

5 nd-" nd

6 Artlirobacter polychrouiogenes X80741

7 Micrococcus roseus X87756

8 Artlirobacter globifoniiis X80736

9 Flcwobacterium balustimim X58771

10 Promicromouospora citrea X83808

11 Klebsiella sp. X80684

12 Cytopliaga diffluens M58765

13 Pseiidoiiionas metidodm D84016

14 Rliizobiiini iiieliloti X67222

15 Artlirobacter oxydans X83408

16 Rliodococciis sp. 16s rDNA X80616

17 nd nd

18 Variovorax paradoxus D30793

•' nd: not determined

Table 2: Groupings of 16S RFLP banding patterns for the isolates. The isolates in

each cluster generated the same banding pattern for the restriction

enzyme indicated. Isolates that generated a unique banding pattern

when digested with a restriction enzyme are listed in the last column.

Erizyme Cluster 1 Cluster 2 Clusters Cluster 4 Unique

Patterns

Ahd 2,6,7,8,15 ^16 9/L3 1, 4, 5, {10-12},

14,17,18

Mspl 2,6,7,8,15 3,5,11,16 9,13,18 1, 4, 10, 12, 14,

17

Rspl 2, 6, 7, 8,15 3,16 9,13 12,17 1, 3, 4, 5, 10, 14,

16

5au3A 2,15 1, {3-14}, {16-18}

Taql 2, 6, 7, 8,15 3,14,16 9,10 5,11 1, 4, 12, 13, 17,

18

117

Figure 1: The eighteen isolates used in this study amplified with a single

arbitrary primer, 682. Lane headings indicate the number of the

isolate and M designates the 123 DNA marker

118

REFERENCES

Amaim, R.I., W. Ludwig, and K.-H. Schleifer. 1995. Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiological Reviews 59:143-169.

Atlas, R. 1988. Microbiology: Fundamentals and Applications. 2nd ed. MacMillan Publishing, New York.

Atlas, R.M., A. Horowitz, M. Krichevsky, and A.K. Bej. 1991. Response of microbial populations to environmental disturbance. Microbial Ecology 22:249-256.

Atlas, R. 1995. Handbook of microbiological media. CRC Press, Boca Raton, FL.

Boivin-Jahns, V., R. Ruimy, A. Bianchi, S. Daumas, and R. Christen. 1996. Bacterial Diversity in a deep-subsurface clay environment. Applied and Environmental Microbiology 62:3405-3412.

Borden, R.C., C.A. Gomez, and M.T. Becker. 1995. Geochemical indicators of intrinsic bioremediation. Ground Water 33:180-189.

Brendecke, J.W., R.D. Axelson, and I.L. Pepper. 1993. Soil microbial activity as in indicator of soil fertility: Long-term effects of mimicipal sewage sludge on arid soil. Soil Biology and Biochemistry 25:751-758.

Brock, T.D. 1971. Microbial growth rates in nature. Bacteriological Reviews 35:39-58.

Brousseau, R., A. Saint-Onge, G. Prefontaine, L. Masson, and J. Cabana. 1993. Arbitrary primer poljmierase chain reaction, a powerful method to identify Bacillus thuringiensis serovars and strains. Applied and Environmental Microbiology 59:114-119.

CemigUa, C.E. and M.A. Heitkamp. 1989. Microbial degradation of polycyclic aromatic hydrocarbons (PAH) in the aquatic environment. In: Metabolism of Polycyclic Aromatic Hydrocarbons in the Aquatic Environment. Varanasi, U. ed. CRC Press, Boca Raton, FL.

119

Costerton, J.W., Z. Lewandowski, D. deBeer, D. Caldwell, D. Korber, and G. James. 1994. Biofilms, the customized microniche. Journal of Bacteriology 176:2137-2142.

de Bruijn, F.J. 1992. Use of repetitive (repetitive extragenic palindromic and enterobacterial repetitive intergeneric consensus) sequences and the polymerase chain reaction to fingerprint the genomes of Rhizohiiim meliloti isolates and other soil bacteria. AppUed and Enviroiunental Microbiology 58: 2180-2187.

DeLong, E.F., G.S. Wickham, and N.R. Pace. 1989. Phylogenetic strains: Ribosomal RNA-based probes for the identification of single cells. Science 243:1360-1363.

Dibble, J.T. and R. Bartha. 1979. The effect of environmental parameters on the biodegradation of oil sludge. Applied and Envirormiental Microbiology 37: 729-739.

Fredrickson, J.K., D.L. Balkwill, G.R. Drake, M.F. Romine, D.B. Ringelberg, and D.C. White. 1995. Aromatic-degrading Sphingomonas isolates from the deep subsurface. Applied and Enviroiunental Microbiology 61:1917-1922.

Foght, J.M. and D. Westlake. 1987. Biodegradation of hydrocarbons in freshwater. Oil in freshwater: Chemistry, biology, and countermeasure technology, pp.217-230 Pergamon Press, New York.

Gardner, W.D., R.F. Lee, K.R. Tenore, and L.W. Smith. 1979. Degradation of selected polycyclic aromatic hydrocarbons on coastal sediments: Importance of microbes polychaete worms. Water Air Soil Pollution 11:339.

Giovannoni, S.J., E.F. DeLong, G.J. Olsen, and N. Pace. 1988. Phylogenetic group-specific oUgodeoxynucleotide probes for identification of single microbial cells. Journal of Bacteriology 170:720-726.

Guckert, J.B., D.B. Ringelberg, D.C. White, R.S. Hanson, and B.J. Bratina. 1991. Membrane fatty acids as phenotypic markers in the polyphasic taxonomy of methylotrophs within the proteobacteria. Journal of General Microbiology 137:2631-2641.

120

Hemming, B.C. 1996. Monitoring and identification of bacteria from agricultural environments by gas chromatography fatty add methyl ester (FAME) analysis. Microbe Inotech publication #112, Microbe Inotech Laboratories Inc. St Lotiis, MO, USA.

Ingraham, J. L., O. Maaloe, and F.C. Neidhardt. 1983. Growth of the bacterial cell, p. 1-48. Sinauer Associates, Inc., Simderland, MA.

Kepner, R.L. and J.R. Pratt. 1994. Use of fluorochromes for direct enumeration of total bacteria in environmental samples: past and present. Microbiological Reviews 58:603-615.

Kogure, K., U. Shimidu, and N. Taga. 1978. A tentative direct microscopic method for coimting living marine bacteria. Canadian Jotimal of Microbiology 25:415-420.

Krawiec, S. and M. Riley. 1990. Organization of the bacterial chromosome. Microbiological Reviews 54:502-539.

Larsen, N., G.J. Olsen, B.L. Maidak, M.J. McCaughey, R. Overbeek, T.J. Macke, T.L. Marsh, and C.R. Woese. 1993. The ribosomal database project. Nucleic Acids Research 21:3021-3023.

Leahy, J.G. and R.R. Colwell. 1990. Microbial degradation of hydrocarbons in the environment. Microbiological Reviews 54:305-315.

Lorentz, M.G., and W. Wackemagel. 1987. Adsorption of DNA to sand and variable degradation rate of adsorbed DNA. Applied and Environmental Microbiology 53:2948-2952.

Makino, S.-L, Y. Okada, T. Maruyama, S. Kaneko, and C. Sasakawa. PCR-based random amplified poljonorphic DNA fingerprinting of Yersina pseudotuberadosis and its practical applications. Journal of Clinical Microbiology 32:65-69.

Mayberry, W.R. and J.R. Lane. 1993. Sequential saponification/ acid hydrolysis/esterfication in a one-tube method with enhanced recovery of both cyclopropane and hydroxylated fatty acids. Journal of Microbiological Methods 18:21-32.

121

More M.I., J.B. Herrick, M.C. Silva, W.C. Ghiorse, and E.L. Madsen. 1994. Quantitative cell lysis of indigenous microorganisins and rapid extraction of microbial DNA from sediment. Applied and Environmental Microbiology 60:1572-1580.

Neefs, J.-M., Y. Van de Peer, P. De Rijk, S. Chapelle, and R. De Wachter. 1993. Nucleic Adds Research 21:3025-3049.

Ogram, A., G.S. Sayler, and T. Barkay. 1987. The extraction and purification of microbial DNA from sediments. Journal of Microbiological Methods 7:57-66.

Osa-Afiana, L.O. and M. Alexander. 1982. Clays and the survival of Rhizobium during desiccation. Soil Science Society Journal 46:285-288.

Ozawa, T., and M. Yamaguchi. 1986 Fractionation and estimation of particle-attached and unattached Bradyrhizobiiim japoniaim strains in soil. Applied and Environmental Microbiology 52:911-914.

Page, AL, Miller, RH, & Keeny, DR. (1982) In: (Ed) Methods of soil analysis. Part 2, Chemical and microbiological, 2nd ed. Agonomy 9, ASA, SSSA Inc., Madison, WI.

Pena-Cabriales, J.J. and M. Alexander. 1983. Growth of Rhizobium in unamended soils. Soil Science Society America Journal 47:81-84.

Penner, G.A., A. Bush, R. Wise, W. Kim, L. Domier, K. Kasha, A. Laroche, G. Scoles, S.J. Molnar, and G. Fedak. 1993. Reproducibility of random amplified poljonorphic DNA (RAPD) analysis among laboratories. PCR Methods and Applications 2:341-345.

Pepper, I.L., C.P. Gerba, and M.L. Brusseau. 1996. Pollution Science pp.81. Academic Press, San E>iego, CA.

Perry, J.J. 1984. Microbial metabolism of cyclic alkanes. In: Petroleum Microbiology, Atlas, R.M. ed. Macmillan Publishing Co., New York.

Picard C., C. Poisormet, E. Paget, X. Nesme, and P. Simonet. 1992. Detection and enumeration of bacteria in soil by direct DNA extraction and polymerase chain reaction. Applied and Environmental Microbiology 58: 2717-2722.

Pillai, S.D., K.L. Josephson, R.L. Bailey, C.P. Gerba, and I.L. Pepper. 1991. Rapid method for processing soil samples for polymerase chain reaction amplification of specific gene sequences. Applied and Environmental Microbiology 57: 2283-2286.

Ritz, K. and B.S. Griffiths. 1994. Potential application of a community hybridization technique for assessing changed in the population structure of soil microbial communities. Soil Biology Biochemistry 26:963-971.

Rosak, D.B. and R.R. Colwell. 1987. Survival strategies of bacteria in the natural environment. Microbiological Reviews 51: 365-379.

Saiki, R.K., D.H. Gelfand, S. Stoffel, K. MuUis. 1988. Primer directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-494.

Singer, M.E. and W.R. Finnerty. 1984. Microbial metabolism of straight-chain and branched aUcanes. In: Petroleiun Microbiology. Atlas, R.M. ed. Macmillan Co., NY.

Steffan R. J. and Atlas R. M. 1991. Pol5mierase chain Reaction: Applications in environmental microbiology. Annual Review of Microbiology 45:137-161.

Torsvik, V.L. and, J. Goksoyr. 1978. Determination of bacterial DNA in soil. Soil Biology Biochemistry 10: 7-12.

Torsvik, V.L., J. Gokso)^', and F.L. Daae. 1990. High diversity in DNA of soil bacteria. Applied and Environmental Microbiology 56:782-787.

Turco, R.F. and Sadowski, M.J. 1995. The microflora of bioremediation. H.D. Skipper and R.F. Turco (Ed.), Bioremediation Science and Applications, pp. 87-103. Soil Science Society of America special publication 43.

Versalovic, J., T. Koeuth, and J.R. Lupski. 1991. Distribution of repetitive DNA sequences in eubacteria and application to fingerprinting of bacterial genomes. Nucleic Acids Research 19:6823-6831.

Vogel, T.M. and D. Grabic-GaHc. 1986. Incorporation of oxygen from water into toluene and benzene during anaerobic fermentative transformation. Applied and Environmental Microbiology 52:200-.

Ward, D.M., R. Weller, and M.M. Bateson. 1990. 16S rRNA sequences reveal numerous imcultured microorganisms in a natural community. Nature (London) 345:63-65.

Weisburg, W.G., S.M. Bams, D.A. Pelletier, and D.J. Lane. 1991. 16S ribosomal DNA amplification for phylogenetic study. Journal of Bacteriology 173:697-703.

Welsh, J. and M. McClelland. 1990. Fingerprinting genomes using PGR with arbitrary primers. Nucleic Adds Research 18: 7213-7218.

Williams J.G.K., A.R. Kubelik, K.J. Livak, J.A. Rafalski, and S.V. Tingey. 1990. DNA polymorphisms amplified by arbitrary primers useful as genetic markers. Nucleic Adds Research 18: 6531-6535.

Zhang, Y. and R.M. Miller. 1992. Enhanced octadecane dispersion and biodegradation by a Pseiidomonas rhamnoUpid surfactant (biosurfactant). AppUed and Environmental Microbiology 58:3276-3282.

Zilli, M., B. Fabiano, A. Ferraiolo, and A. Converti. 1996. Macro-kinetic investigation on phenol uptake from air by biofiltration: Influence of superfidal gas flow rate and inlet pollutant concentration. Biotechnology Bioengineering 49:391-398.

Zhou, E. and R.L. Grawford. 1995. Effects of oxygen, nitrogen, and temperature on gasoline biodegradation in soil. Biodegradation 6:127-140.