.1

"A?

40-6 a”

-52.

w

31- "‘22
. 19:724.: ,,

£3“ . «
“*4“:

9...“: .3.-
-.. .w.

1'
”A...

‘1‘: - :
m tram
J.."':L'.'Fk 'G-
V‘." ’:

I 3
. , ‘4
“33?:‘55’

v?

“.43 L:
~

‘ '-u u
L .
1‘. Q?‘ huh“

. 'QVR

 

.2...’

MIC CHI IGAN STATE

IIIII III II IIIIIIIIIIIIIIIIIIIIIIIIIIII

23 010219024

II

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

MICROBIAL BIOFILM COMMUNITIES IN GRANULAR
ACTIVATED CARBON FLUIDIZED BED REACTORS FOR
TREATMENT OF GROUNDWATER CONTAMINATED WITH
AROMATIC HYDROCAEEEQNEd by

ARTURO ANDRES MASSOL—DEYA

has been accepted towards fulfillment
Of the requirements for

 

 

 

Ph.D. degreein MICROBIOLOGY
I
(I w \ I
Major p‘ofessor d

Cake/9:71

MSU i: an Affirmative Action/Equal Opportunity Instiluu'on 0-12771

——————_.

 

LIBRARY
Mlchlgan State
Unlverslty

 

 

 

PLACE II RETURN Boxwromavombchockoum ywncord.
TO AVOID FINES Mum on or More data duo. '

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU tummmwomlm
Wm:

, ..

MICROBIAL BIOFILM COMMUNITIES IN GRANULAR ACITVATED
CARBON FLUIDIZED BED REACTORS FOR TREATMENT OF GROUNDWATER
CONTAMINATED WITH AROMATIC HYDROCARBONS

By

Arturo Andres Massol—Deya

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHIIDSOPHY

Department of Microbiology

1994

 

 

IIII .IIlllII'. .IIIII'

ABSTRACT

MICROBIAL BIOFILM COMMUNITIES IN GRANULAR ACTIVATED
CARBON FLUIDIZED BED REACTORS FOR TREATMENT OF GROUNDWATER
CONTAMINATED WITH AROMATIC HYDROCARBONS

By

Arturo Andrés Massol-Deya

The role and stability of non-indigenous populations on the structure and
composition of biofilms in aerobic granular activated carbon fluidized bed reactors (GAC-
FBR) treating groundwater contaminated with monoaromatic hydrocarbons were
examined. The biofilm was composed of approximately 109 to 1010 total cells per gram of
GAC. Comparison of community fingerprints of amplified 16S rDNA obtained by
restriction endonuclease analysis and fatty acid methyl ester profiles (FAME) of seeded
community with GAC colonized by the indigenous aquifer community suggested there was
little influence of the non-indigenous populations on community structure. The biofilm
community was composed of at least 12 different coexisting bacterial populations.
Estimates of their maximum specific growth rate (um) half-saturation constant (K3) and
maximum substrate utilization rates (It) obtained by fitting substrate depletion cm'ves to the
integrated Moncd equation revealed a high diversity of toluene growth kinetics by these
microbial populations. Partial 168 ribosomal RNA sequences were determined for each of
the three co—dominant toluene degrading strains found among all laboratory operated FBRs
studied with the aim of establishing phylogenetic affinities of the biofilm community.
Similarity and evolutionary distance matrix analysis suggest closed affiliation of these
strains to three subclasses of the Proteobacteria, the alpha, beta and gamma subdivision
respectively. In addition, scanning electron microscopy (SEM) and confocal scanning laser

microscopy (CSLM) were used to study the development, organization, and structure of

the biofilm community. During the early stages of colonization, microcolonies were
primarily observed in crevices and other regions sheltered from hydraulic shear forces.
Eventually, these microcolonies grew over the entire surface of the GAC. This lead to the
development of discrete discontinuous multilayer biofilm structures. Cell free channel-like
snucmmsofvafiabksizeswaeobservedmmmcomectmesurfacefilmwimdtedeep
innerlayer's. Thisappearedtoincreasethebiological sm‘faceareaperunitvolumeratio,
which may facilitate transport of substrates in and wastes products out of deep regions of
the biofilm at rates greater than possible due to diffusion alone. These architectm'al features
were also observed in biofilms from four field scale GAC reactors that were in commercial
operation treating petroleum contaminated groundwaters. This suggests that formation of
cell-free channel structures and their maintenance may be a general microbial strategy to
deal with the problem of limiting diffusive transport in thick biofilms typical of fluidized
bed reactors. Understanding of the forces and principles which regulate biofilm
development will help in efforts to model, control, and exploit biofilm related processes of
industrial, medical, and environmental importance.

Covyfight by
ARTURO ANDRES MASSOL-DEYA
1994

This dissertation is dedicated to
my parents Alexis and Tinti,
my wife Vicky,
and my daughter Corali

 

 

 

 

 

 

 

ACKNOWLEDGMENTS

I am grateful to all the people who has profoundly affected my scientific
development, in particular for the encomagemeut, support and insights of Francisco
Fuentes, my undergraduate advisor, and James M. Tiedje and Robert F. Hickey, my major
advisors. I am also grateful to David Odelson and Larry Forney for their special help
throughout my graduate studies. I thank the members of my advisory committee: Loren
Snyder, Suzanne Thiem, and John Breznak for providing wealth of useful suggestions.
Special thanks is given to Joanne Whallon for help with confocal microscopy and comments
on manuscripts. I would also like to thank Rolland Weller who's inspiration has always
been present. I thank Zhou Jizhong for his help with the 168 rRNA sequencing and
phylogenetic analysis, Helen Garchow for performing all the FAME analysis, Luis Rios-
Hernandez for his help on the isolation and characterization of biofilm bacterial populations,
and the Center for Electron Optics for their help with SEM. I also thank my lab colleagues
James Champine, James Cole, Robert Sanford, Joanne Ghee-Sanford, and Marcos Fries for
teaching me science and English, and for their ever good will and help. The staff and
members of the Center for Microbial Ecology have always been very helpful. I am also
thankful for financial assistance from the National Institutes of Health, Fellowship No.
GM14047-04, and the National Science Foundation, Grant No. BIR-912m06 to the Center
for Microbial Ecology.

I am specially thankful to my family for support and motivation and Vicky Ortiz, my
wife, who always had encouraging and understanding words and who's struggles and

battlesarethesameasmine.

TABLE OF (DNTENTS

LIST OF TABLES
LIST OF FIGURES
Chapter 1. AN INTRODUCTION

Chapter II. SUCCESSION AND CONVERGENCE OF
BIOFILM COMMUNITIES IN FIXED-FILM REACTORS
TREATING BENZENE, TOLUENE, AND p—XYLENE
CONTAMINATED GROUNDWATER

Chapter III. ISOLATION AND CHARACTERIZATION OF
BENZENE, TOLUENE, AND P-XYLENE DEGRADERS FROM
AEROBIC BIOFILMS IN GRANULAR ACTIVATED CARBON
FLUIDIZED BED REACTORS

Chapter IV. CHANNEL STRUCTURES IN AEROBIC
BIOFILMS FROM FIXED-FILM REACTORS TREATING
CONTAMINATED GROUNDWATER

Appendix A. BACTERIAL COMMUNITY FINGERPRINTTNG
OF AMPLIFIED 168 AND 16-238 RIBOSOMAL DNA GENE
SEQUENCESAND RESTRICTION ENDONUCLEASE
ANALYSIS (ARDRA)

Appendix B. DIFFERENTIAL AMPLIFICATION OF rRNA
GENES FOR THE DETECTION AND IDENTIFICATION OF
PSEUDOMONAS CEPACIA (G4) AND P. PICKETTII (PKOl)

LIST OF REFERENCES

19

58

81

LIST OF TABLES

CHAPTERII

Table l. GAC-FER startup phase and performance at steady-
state conditions.

Table 2. Relative abundance (96 of total) of fatty acids of the
GAC-FER communities.

Table 3. Relative abundance (96 of total) of fatty acids from
samplescollectedatthebottomandtopofthecolumnbedheight
level of TOLS reactor at day 101.

CHAPTERIII

Table 1. Start-up conditions of fluidized bed reactors used in
this study.

Table 2. Description and comparison of toluene growth kinetic
coefficients of inoculum strains and dominant GAC-FER isolates.

CHAPTER IV
Table 1. Description of GAC-FBRS used in this study.

Table 2. Relative abundance (% of total) of fatty acids in the
microbial communities of five different GAC-FER systems.

APPENDIX B

Table l. Bacterialstrainsstudiedandreactionofspecie-
specific PCR amplification.

29

36

37

45

52

76

92

LIST OF FIGURES

CI-IAPTERI

Figure l. Processesinbicfilmaccumulation:cellsuanspcrtto
the substratum, reversible cell adsorption in equilibrium with
desorption, irreversible cell attachment by secretion of biological
active adhesives or by special cell structures, growth and biofilm
accumulation, sloughing of fragments of the film, and recolonization
of open surfaces.

Figure 2. Classical biofilm model illusu'ating substrates flux
from the liquid phase through a mass transfer layer with width L and
intothebiofilmmatrixwithwidthuanddensityXf (afterCriddleet
al., 1991).

CI-IAPTERII

Figure 1. Schematic diagram showing setup conditions of
FBRs used for the experiments. Filter system was absent in control
FBR (TOLNS and BTX.NS) as well as in the inoculated FBR
('I‘OLS) after day 53. Illustration not drawn to scale.

Figure 2. Typical profile of dissolved oxygen and toluene
concentration with depth in the bioreactor column.

Figure 3. Observed dissolved oxygen consumption (mg/L),
toluene effluent concentration (mg/L). and cell density (cells/g of
GAC) in the TOL.S reactor inoculated with three different bacterial
strains. The bioreactor was fed amended fiter-sterilized groundwater
for 53 days to allow the establishment and stabilization of the
inoculum members.

Figure 4. Densities of inoculated Strains as determined by
selective plate counts over time in the TOLS reactor. CFU, colony
forming units.

22

28

31

Figure 5. Community fingerprint of bioreactor seeded with
bacterial strains G4, PKOl, and Pan. The filtration system was
removed on Day 53. Shown is a HpaII digestion of the amplified
168 rDNA from reference cultures (Lanes G4, PKOl, and Pan),
and GAC community (Lanes Day 0, 22, 38, 53, 90, and 112). The
sizes ofmarker DNA fragments (lanes SIZE) (in base pairs) are
indicated on the left.

Rm 6. Restriction fragment length polymorphisms of
amplified 16S rDNA from S.BTX reactor at day 159. Samples
correspondto the 20, 60, and90% reactorcolumn bed heightflanes
B, M, and T, respectively) digested with HaeIII, Hinfl, and Msel.
ThesizesofmarkerDNAfragmentsaanesS) (inbasepairs)are
indicated

on the left.

Figure 7. Restriction fragment length polymorphisms of
amplified 168 rDNA from mature GAC communities digested with
HaeIII, Him, and MseI. The lane designations correspond to the
bioreactor treatments (see Materials and Methods). The sizes of
marteéDNA fragments (lanes SIZE) (in base pairs) are indicated on
the

Figure 8. Genomic fingerprints of bacterial isolated from GAC
communities. Shown is a Mad digestion of the amplified 16S
rDNA. Lanes SIZE show the DNA size markers, indicated (in base
pairs) on the left.

Figure 9. Principal components analysis showing temporal
variations of FAME profiles for GAC communities. Reactors:
toluene or benzene, toluene, and p—xylene fed systems of seeded
(TOLS and BTX.S) or aquifer colonized (TOL.NS and BTX.NS)
biofilm communities. Parameters are listed in Table 2.

CHAPTERIII

Figure 1. REPfingerprintpatternsofbacterialisolatesfrom
GAC-biofilm communities. DNA molecular weight standards (lanes
MW) are WindIII, pUCl9/Taq-I-pUC19/Sau3AI (Stratagene, La
Jolla, CA). The pattern of PCR products generated of chromosomal
DNA from inoculum strains PKOl and Pan as well as a positive
control (P. aeruginosa) are included.

Figure 2. Phyl enetic distance tree produced from partial 16S
rRNA sequences (81 1090 positions). Representative members of
the various subgroups of the Proteobacten'a and the dominant toluene

X

32

33

34

35

38

51

degradingbacterialpopulationsfromGACbiofilmcommunitiesare
shown. 168 rRNA sequenceofBacillus subtilr's wasusedasan
outm-

CHAPTERIV

Figure l. Biofilm development in the toluene-fed reactor. The
bed height reflects overall biofilm thickness (Open circles); dissolved
oxygen uptake is a measurement of biological activity (black circles).
Gray arrows are the times at which the microscopic image shown
weretaken. Blackarrow showsthefirsttimeatwhichbiomasswas
controlled by application of shear forces.

Sign 2. SEM micrograph of sterile GAC particle showing
' t surface zones: S, smooth; R, rough; C, cavity. Bar, 5 tun.

Figure 3. SEM micrograph of a 3-day-old microcolony mainly
growing over rough areas and within cavities of the GAC. Bar, 5 pm.

Figure 4. SEM micrograph of a 12-day-old colonized granule
from the toluene fed reactor. Most ofthe surface has been colonized
with actively dividing rod cells. However, there is smooth (arrow)
zone yet to be colonized. Bar, 5 um.

Figure 5. Series of confocal micrographs showing biofilm
growth on GAC particles in a toluene fed reactor. Each extended
focus view was constructed by computer overlay of the optical
sections in a 2 series. (A) 16—day-old; (B) 63.day-old; and (C) 77-
day-old biofilm structure; bar, 0.5 mm. The biofilm thickness
increased from 15 to 300 um between 16 and 77 days as determined
by dark field microscopy.

Figure 6. Confocal micrographs of a 77—day-old biofilm
showing the nature of the channel structures in the toluene fed
reactor. (A) horizontal (x, y) optical thin section, 82.6 um from the
top of the biofilm (the horizontal line indicates the optical cutting
position for the sagittal section); and (B) sagittal (x, 2) view showing
the vertical cell disuibution and the cell-free channel reaching up to
the surface of the film.

Figure 7. Confocal micrograph of the void spaces created by
protozoa (arrow) grazing on bacteria in a 153-day-old biofilm
community. At this point, the bioreactor had gone through tluee
shearing treatments for biomass removal and the biofilm structure

53

65

67

70

displaysnoneofthedistinctiveverticalcellarrangementpattern
observedintheearly stage. Bar,5 um. 71

Figure 8. Horizontal (x, y) optical sections of biofilm

communities from different field-scale FBRs. Note the spacing

between cell aggregates and cell arrangements. (A) TC reactor

system with channel-like structure partially colonized by cocci-

shaped cells. (B) BTEX fed system showing a 'coral-reef'

appearance of bacterial growth. (C) Biofilm structure of JEN system

showing bacterial aggregates separated by well-defined void space

boundaries (arrow). (D) Micrograph from the PW system showing

bacterial distribution and partitioning by cell-free spaces. 72

Figure 9. Horizontal (x, y) optical section (lower half of image)

and sagittal (x, 2) optical section (upper half of image) of a 30-day-

old field reactor site (TC) showing a fully hydrated biofilm with cell-

free spaces and a discontinuous base film (arrow). 73

Figure 10. Laser scanning darkfield photomicrograph ofa77-
day-old bacterial GAC coated particle showing variauon of the
biofilmthicknessonthesamecolonizedsurface.Bar,15ttm. 74

Figure 11. Principal components analysis showing variation in
FAME patterns in five different GAC-FBR systems: TOL, BTEX,
TC, PW, and JEN. 78

APPENDIX A

Figure 1. Schematic representation of the ribosomal RNA

operon showing approximate localization of priming binding sites for

PCR amplification. The spacer region varies in length (e. g., number

of tRNA's), and sequences among the different operons withina

single organism (Brosius et al., 1981). 84

Figure 2. Products of PCR amplification from microbial

communities samples. (A) The 16S rDNA amplification yielded a

1,500 bp fragment from axenic culture (lane B) as well as bioreactor

samples (lanes C to F). (B) Results of 16—238 rDNA intergenic

amplification of the same samples yielded various fragments of

different sizes. Phage lambda (A) DNA digested with PstI was used

asmarkerflanes A), indicated (inbasepairs) ontheleft. 86

Figure 3. Photograph of ethidium bromide stained agarose gel.
Lane MW: molecular weight DNA markers, indicated (in base pairs)
on the left. Lane ¢ DNA: PCR amplification of water. All other

lanes are HpaII restriction endonuclease digestion of PCR amplified

xii

16S rDNA from: G4, Pseudomonas cepacia; PKOl, P. pikettii; and
Pan, P. purida; Day 0, 13, 29, and 65 are bioreactor samples
inoculated with the same strains of bacteria at Day 0 and challenged
by aquifer populations after day 19. One Interpretation of the results
is that Pan becomes the dominant population after Day 29.
However, this erroneous interpretation is the result of multiple
water populations that happen to have the same restriction

gments of Pan when HpaII is used. Differences will be more
readily detectable when a second or third restriction endonuclease is
put to use. 89

APPENDIX B

Figure 1.Ribosomal RNA sequence alignments (corresponding

to positions in the Escherichia coli 16S rRNA, 998 to 1017) showing

target primer binding sites for strains G4 and PKOI. Sequences are

either results from Jizhong et al., unpublished or taken from the

ribosomal data base. 91

Figure 2. Difl'erential amplification of bacterial 16S rRNA
genes. Lanes A and S, 3. Fr: I standards (in base pairs); lanes B and
LPCRwithouttemplate; lanesCtoI-I, anthoOarePCR
amplification of Escherichia coli, strain G4, strain PKOl , strain
Pan , P. fluorescens, and P. aeruginosa using universal primers pA
and pH, and using the universal primer pA combined to the specie-
specific oligonucleotide primer pPKV3 respectively; and lanes P to R
are PCR amplification using pA and pPKV3 primers of 1:1, 1:10,
3:1 1:1)00 PKOl template DNA in the presence of non-specific DNA
W1 . 93

Figure 3. Differential PCR amplification of P. cepacia (G4)

from a toluene treating bioreactor community inoculated with the

same strain of bacteria using primers pA and pG4V3. Lanes A and

F, 1. PSI I standards (in base pairs); lane B, PCR without template;

lanes C, strain G4; lane D, bioreactor community, day 63; and lane

E, 1:1 ofthesamecommunity DNA de4. 94

Chapter I
AN INTRODUCTION

OVERVIEW

Bacteria in nature have a marked tendency to colonized solid-liquid interfaces
(Fletcher, 1979; Costerton et al., 1987; Kolenbrander et al., 1990). The immobilized cells
grow, divide, and produce extracellular polymers which frequently form a framework of
cdlsembeddcdmanmganicmauixofmicrobialuiginaappin-Scoudeostertm.
1989). Biofilms form on virtually all surfaces immersed in natural aqueous systems
including biological surfaces (e.g., aquatic plants) or abiological (e.g., metals). In natural
environments, the surface accumulation is not necessarily uniform in time or space. For
example, in the case of algal mats found in hot springs or modern lagoonal environments, a
thickbiofilm (approximately3lX)-400mm)cancontainbothaerobicandanaerobiczones
due to oxygen diffusion limitations (W and et al., 1989). Furthermore, diurnal changes in
the different compartments are affected by the production of oxygen in photosynthetic
processes during daylight hours.

The formation of biofilm communities on sm'faces has some very important
economic and ecological implications. Biofilm processes are studied because of the
scientific and technological applications of these systems. These systems are stable
biological units that can transform a remarkably wide range of chemical substrates (Holm et
al., 1992; Wolfaardt et al., 1994). These characteristics have been exploited in sewage and
wastewater management systems (e.g., trickling filters) and in the development of new
bioremediation technologies for detoxification of contaminated sites (e.g., biological
activated carbon fluidized bed reactors). However, little information is available on the
possible roles of microbial diversity, population interactions, and abiotic factors on the
maintenanceofstr'ucttueandfunction withinthesecommunities. Duetothestrong

interrelation between the microbial processes, substrate removal, community structure and

1

2
composition, a practice which neglects these facts is deemed to cause erroneous design and

improper Operation of biological waste water treatment facilities and/or bioreclamation of

contaminated sites. Theunderstandingoftheseinteractiveprocesseswillallowusto

namthediffemncesbaweenhbaataymodelsandactualprocessesinnatum. Without

a rigorous and systematic understanding of these ecological relationships then accurate

mathematical models can not be formulated. These models will be extremely valuable for

engineering biofilm processes. Our knowledge of the forces and principles involved in
regulation of biofilm development will help in efforts to manipulate and exploit these
processes for technological and ecological advantages.

Biofilms are Significant in bioengineering and fundamental microbiology. Thus,
conceptual and methodological approaches are required which will aid to better describe
and control biofilm processes. The effects of many biological and abiotic factors on
structuringofbiofilmsarestilltobedetermined. Studiesareneededontheimpactof
environmental conditions on microbial community structure. At the same time, we need to
evaluate the impact of population interactions (e. g., extinction vs. coexistence;
allochthonous populations vs. indigenous populations), and how such interactions dictate
community structure. These principles which regulate biofilm development should help in
efl'orts to model, control, and exploit biofilm related processes of industrial, medical, and
environmental importance.

This study was organized around the following objectives:

i. to examine the fate of allochthonous populations in preemptive colonization of a porous
media and after being challenged by indigenous groundwater bacteria in an aerobic
GAC-FBR

ii. to examine the role and stability of seeded pioneer populations

iiitoexaminetheroleandstabilityofsecondarycolonizers

iv. to evaluate the effects of substrate composition changes on community structure

v. to isolate and characterize BTX degrading groundwater populations

BIOFILMS
WWW" Biofilmdevelopsasamfltofscvflal
concurrent processes: (i) transport of cells to the substratum, (ii) adsorption of cells, (iii)
growth and metabolic processes at the colonized surface, (iv) detachment ofportions ofthe
film, and (v) recolonization (Figure 1).

 

13:5” ©\/”
lfilglg

Clean 7 Reversible
Substratum Tramport Adsorpt'mn Desorption

 

 

 

 

 

 

Figure 1. Processes In biofilm accumulation: cells transport to the substratum,

reversible cell adsorption In equilibrium with desorption, irreversible cell attachment by
secretion of biological active adhesives or by special cell structures, growth and biofilm

accumulation, sloughing of fragments of the film, and recolonization of open surfaces.

In general, microbial colonization begins as planktonic cells approache a
substratum. These processes may occm‘ by either passive transport (diffusion or advective
transport, Brownian motion), and/or active biologically controlled transport (flagella
mediated motility, and/or swarming motility) (Griddle et al., 1991). Active transport of
motile cells can also be affected by positive or negative chemotactic responses to gradients

4
of some biologically relevant compounds. As the bacterium meets the surface, reversible

adhesion of cells to the substratum initially occurs via physicochemical processes. Surface
properties, liquid phase and environmental conditions such as temperature or pH influence
the degree of reversible sorption (Fletcher, 1979; Mills and Maubery, 1981; Fletcher and
McEldowney, 1983; McEldowney and Fletcher, 1987; Van Loosdrecht et al., 1990). In
addition, biological factors such as species composition and interspecific interactions, and
their cell surface properties have also been shown to affect levels of adhesion (Warren et
al., 1992). For example, the ecological relevance of biological interactions among
genetically distinct cell types during colonization and coaggregation of oral bacteria have
been reviewed (Kolendrander et al., 1990).

During the reversible adhesion phase, organisms might become detached by
moderate shear forces and still capable of exhibiting Brownian motion and/or flagellar
motility (I-Iarkes et al., 1992). Meanwhile, irreversible adhesion is characterized by active
microbial processes in which production of extracellular polymers, fibrils, and adhesives
mediate cell-substratum interactions (Beachey, 1988; Vandevivere and Kirchman, 1993).
Thesehrcludespecialceflularsu'ucmressuchaspihfoundonmembersofdre
Enterobacten‘aceae and Pseudomonadaceae and many others, and holdfasts which occur on
some prosthecate bacteria. Subsequent cell binary fission leads to the formation of
microcolonies and eventual biofilm formation. Such biofilms consist of bacterial cells
embedded in a matrix of their own exopolysaccharide glycocalyces which helps to nmintain
the film structme.

A biofilm consists of microorganisms immobilized at a surface and an overlaying
liquid or gas layer. Sometimes the substratum can also serves as nutrient rate-limiting step
for microbial growth as illustrated in wood decomposition. On the other hand, a
continuous liquid phase which may fill and interconnect fraction of the biofilm volume with
the bulk liquid solution, contain different dissolved and suspended particulate materials and

clearly influence diffusion processes of substrates, nutrients, waste products, and gases

5
into the film matrix. Thus, the liquid compartment can directly affect biofilm behavior

(e.g., microbial activity) primarily as a result of the mixing patterns (fluid dynamics), and
system geometry. The gas phase may be absent in some biofilm systems. Oxygen can in
these systemsbeprovideviaaeration. Mirdngpattansmayalsofacilitateremovalof
gaseous microbial products such as C02 and CH4 from inside the biofilm compartment of
anaerobic biofilms.

 

MASS
BULK LIQUID TRANSFER
LAYER
substrate
flux ——)
X
T {>2
L
water
flow

 

 

 

 

Figure 2. Classical biofilm model illustrating substrates flux from the liquid phase
through a mass transfer layer with width L and into the biofilm matrix with width Lf and
density Xf (after Criddle et al., 1991).

Descriptions of substrate utilization processes in biofihns have been based on
idealized biofilm models consisting of an homogenous matrix of bacteria and their
polymeric materials with uniform density, Xf (mg/cm3), uniform thickness, Lf (cm), and a
laminar transfer layer, L (cm) through which substrate is transported from the btrlk liquid
phase (Figure 2). Within the biofilm, substrate moves by molecular diffusion and is
consumed locally. This model has been used for derivation of differential equations to
describe substrate removal kinetics in biofilms (Williamson and McCarty, 1976; Rittrnann
and McCarty, 1980; Criddle et al., 1991). This provides the methodological tools for
determining rate limiting processes and prediction power.

6
In order to address biodegradation issues in fixed-film systems, betta' knowledge

of biological reaction kinetics and improved structural models are required. Cluistensen
and co-workers (1989) have shown that the structure of a biological film used for
wastewatertreatmentwill malargeextentbedeterrninedbythetransportationofsubstrate
andreactionproductsinsidethefilm. Adescriptionofthesystembasedonsubstrate
remalkineticsalonewillfailsincethereactionkineticscausesinevitablechangesinthe
biofilm structure. Design rules for substrate removal in biological films therefore must
include carelations between substrate removal kinetics and the structure and development
of the film.

\/Rele¥mgf_bigfi1ms - The famation of biofilm communities on Stufaces has
sane important economic, technical and ecological implications. The importance of
attached microbial community processes are manifest in many forms, from beneficial
processesr’n Wenvironmenttoundesirable roles asdisease agents (Characklis and
Wilderer, 1989; Characklis and Marshall, 1989; Costerton et al., 1987). The fouling of
heat exchanger tubes, cooling towers, and wastewater pipes by biofilms increases heat
transfer resistance, energy losses, fluid frictional resistance and will reduce the overall
perfa'mance eventually rendering these systems almost inoperative . The resultant energy
loss and reduced performance, along with the need for regular cleaning, cost hundreds of
millions of dollars per year on a worldwide scale (Marshall, 1992). In addition, microbial
processes can accelerate deterioration of metallic substrata wherein sulfate-reducing and
acidogmicbaaefiammemaaobiczawmmebiofilmmemlmwrfacecaninidamcamsim
by causing cathodic depolarization at the metal surfaceWy reducing the equipment
lifetime. Health risks associated with biofilm accumulation on teeth and gums, urinary
tract, intestines, and after insertion of prosthetic devices into the human body have been
shown (Costerton et al., 1987). J

Many transformation processes in nature are accomplished most efficiently by
mixtures of populations. Biofilms in natural aquatic ecosystems frequently influence water

7
quality by affecting dissolved oxygen content, nutrient fluxes, and serving as sink for

many toxic and hazardous materials. For years, engineers have exploited these
characteristics in fixed-film systems for extraction and biological oxidatiar of organic and
inorganic contaminants found in danestic, agricultural and industrial wastewaters (e.g.,
fluidized bed reactors). These systems are based on particles that provide a large surface
arufabiofilmawumtflafimmdmeabflityofflnseesmblishedpopulafionsmmsfama
wide range of chemicals. Biotechnological applications of biofilm include the extraction
of minerals (biohydrometallurgy), while in other industrial uses the productivity (yield) and
stability of the biotransfa'mation processes are improved.

W - Attached microbial cells can play important and not well
understood roles in the maintenance of ecosystem stability. For example, biofilm
comrmrnities can act as microbial reservoirs, aiding in the repopulation of surrounding
waters when favorable conditions for growth are present (Fletcher, 1979). The
predominance of attached growth within a broad range of environmental conditions
suggests that this mode of growth must provide an important selective advantage. Possible
advantages associated with attachment and biofilm fa'mation have been reviewed (Breznak
et al., 1984). First, fixation and preservation of position in preferential zones while
preventing extinction by eluding washout near food sources. Accumulation of substrates at
the solid surface and their subsequent availability for microbial uptake may also be an
influencing factor for cells attaching to surfaces. Another possible advantage can be the
optimization of substrate and waste transpat from the attached cells by either diffusive
and/0r advective transport. Planktonic cells will tend to be transpated within their locus of
Sturounding liquid. Therefae suspended cells will benefit mainly fiom diffusion alone.
Protection against predators has been observed in biofilm systems. This is probably the
result of sequestering of prey and diminished of area exposure to attack. Finally, attached
communities can create and maintain microenvironmental conditions which may be

advantageous for growth or survival but which are not achievable under suspended

8
planktaric states (for example, maintenance of anaerobic zones, optimal pH, as well as

avoidance of toxic or temperature shocks, and desiccation).

Biofilm formation is limited mainly by the availability and transpat of nutrients
from the surrounding aqueous phase (Marshall, 1992). Druing early stages of biofilm
formation however, oxygen and nutrient diffusive processes are normally not limiting
factors. Once multilayers ofbacterial cells become sequestered in their own polymeric
matrix,diffusiar beconres amajorfacta'indetermining tlremicrobial structrueatthe
colonized surface. Microsensor technology applied to relatively thick aerobic biofilms has
shown the existence of steep oxygen and nitrate gradients demonstrating the importance of
diffusive transport processes (Dalsgaard and Revsbach, 1992). These gradients develop as
a result of rapid oxygen, nitrate, and nutrient consumption by aerobic bacteria, at a' near
the water-biofilm interface, and the comparatively low rate of diffusion of oxygen and
nitrate through the biofilm matrix.

The chemical and physical activities of these microbial communities produce a
homogeneous system at the colonized surface (Lappin-Scott and Costerton, 1989).
Organismslivingwithinthebiofilmhaveatendencytobemaestableonatemporaland
spatial basis than planktonic populations. A general observation has been the resilience of
such communities against perturbations or potentially lethal factors such as biocides in the
nearby liquid phase. T‘IuWafiyimportant in the biomfical treatrrrentfilofilm
W (Blenkinsopp et al., 1992; Brown and Gauthier, 1993).

Microaganisms are distinguished from multicellular a'ganisms by their ability to
existasisolatedurricellularformsatleastinpartoftlreirlifecycle. The studyofrrricrobes

 

 

can thus contribute to our understanding of intrinsic cell properties (e.g., control of gene
expression). The influence of interspecific interactions on corrrrnunity structuring is a
cartinuing problem in microbial ecology. Current ecological theaies are based heavily on
the concept of competitive exclusion and the assumption that communities exist at

canpetitive equilibrium. Therefae, success or survival is based mainly on the

9
physiological fitness of each competitor in a stable uniform environment. Since

competitive equilibriumrequires thattheratesofchangeofallcompetitorsbe zero, and
since the physical environment, predation and other factas are changing constantly, it
seems likely that true equilibrium rarely occurs in nature (MacArthur and Levins, 1967;
Garden et al., 1969; Lewin). There is increasing evidence that competitive exclusion
occurs rarely in nattual systems (McNaughton, 1978; Dykhuizen et al., 1983; Wiens,
1983; McCormick et al., 1991). Even in kinetically distinct bi-species competition
experiments using steady-state chemostats, the exclusiar of one species is rare, perhaps
due to either resource or space partitioning (Levin, 1972; Veldkamp et al., 1983).

Dworkin (1991), has suggested that horizontal interactions among microbial
populations are an integral feature of the multicellular nature of biofilnrs. These interactions
take place when: (i) a multicellular more has to be established, (ii) a multicellular activity
has to be accomplished, or (iii) information has to be distributed among the populations.
While a rich body of mathematical equilibrium tlreaies has been developed, they are not
necessarily valid for solutions to natural communities (Fredrickson and Stephanopoulos,
1981; Veldkamp et al., 1983). The effects of many abiotic and biotic factors on community
suucuuesfiHmedsmbedemrminedandmaemfamadonisneededonhaimnmlconuol

levels on community Structming.

BIOFILMS IN POROUS MEDIA

 

Groundwater is one of the principal water sources for domestic, industrial, and irrigation
uses. Its cartamination by aromatic and alkyl substituted aromatic hydrocarbons such as
benzene, toluene and xylenes (BTX) has become a serious threat to human health. Some
pathologies have been attributed to BTX contaminated groundwater including benzene-
irrduced carcinogenesis (e.g., leukaemogenesis) and non-specific effects on the central
nervous system of vertebrates (Dean, 1978, 1985). Leakage from underground gasoline

10
and petroleum storage tanks, and distributions systems are the major causes of this

pollution. Release of such compounds to the environment, either naturally or accidentally,
can severely deteriorate water quality and have adverse effects on biological systems.
These effects (Moae et al., 1974) include lethal toxicity, which is ascribed to impairnrent
of essential cellular and subcellular processes (e.g., modified membrane integrity and
transpat systems) as well as alterations of physiological or behavioral activities (e.g.,
abnamal responses to physical and chemical stimuli).

The fate ofthe‘trace levels ofthese organics in aquifer systems is governed by
physical (solubility, adsorption, hydrogeology, volatilizatiar), chemical and biological
factors. On a long term basis, biodegradation in aquatic environments appears to be the
maja' attenuation mechanism (Atlas, 1981; Davis and Carpenter, 1990; Leahy and Colwell,
1990; Kcnopka and Turco, 1991). A diverse group of heterotrophic bacteria are known to
posses the metabolic ability to utilize BTX compounds as their sources ofcarbon and
energy (Olsen et al., 1989; Ridgway et al., 1990; Young, 1990; Fredrickson etal., 1991).
Several metabolic pathways, either aerobic or anaerobic, have been described (Evans et al.,
1991; Hutchins, 1991; Edwards et al., 1992). For toluene alone, five different aerobic
pathways are currently known (Zylstra and Gibson, 1991). In general, the different
pathways for aromatic oxidation have initial conversion steps that are carried out by
different enzymes. Such enzymes activate the ring structure by the introduction of oxygen
atornstoform alimited numberofcamnoninterrnediates; protocatechuate,cathecolsand
substituted cathecols. These are the substrates for ring cleavage and further metabolism to
TCA cycle intermediates (Van der Meer, et al., 1992).

The rates of microbial transformation are generally governed by physical processes
(mass transfer and diffusion) as well as the biological process itself (substrate transpat and
enzymatic machinery). Mass transfer limitations a' bioavailability of substrates (both
electron W and electron accepta) or nutrients however, may limited removal of low

concentrations of BTX by suspended bacteria in aquifers. At substrate concentrations at

11
whichgrowthanddecayratesareequaLnetgrowthwillnotoccur. Intlreabsenceofan

alternative energy source, establishment of deer populations will not take place
(Rittmann and McCarty, 1980; Criddle et al., 1991). This threshold substrate
carcentration mybeinexcessoftherecommendedstandardforpotable water. Atthe
monenawelackcompleteurxlersmndingofdrefacmthatpromaemcalduamea
biodegradability in nature. In order to gain that knowledge it is necessary to evaluate closer
the interaction of abiotic factors with biological processes.

Thecostsmexecuteammedialpmgrammgroundwaterconmminawdsitescanbe
staggering and funds available to clean up the number of seriously contarrrinated areas are
insufficient. The costs of physical and chemical remedial programs for the average site can
range from several thousand to millions of dollars per site (Hickey et al., 1991; Evans,
1994). There are only two courses of action: clean up fewer sites, or develop more
econorrrical technologies that significantly reduce the cost of reclamation. One attractive
alternative is the use of biological treatment, which is generally mac econanical and may
significantly reduce the cost of clean up.

The success or failure of biorernediation technology is determined by a variety of
environmental factas including temperature, pH, nutrients, trace elements, electron
acceptors, and substrate composition and bioavailability. The interactions of these factas
arebetterunderstoodandcanbemanipulatedinordertooptimize microbial growth,
survival and performance. Aerobic granular activated carbon fluidized bed reactors (GAC-
FBR) have been used for the reclamation of BTX contaminated groundwater in laboratory
and field scale systems (Ridgway et al., 1990; Hickey et al., 1991; Jing et al., 1992). Low
effluent concentrations of these compounds are achieved in cost effective pump and treat a'
enhanced in situ systems. The GAC-FBR process is a fixed-film biological treatment
which destroys the contaminant, coupled with the adsorption properties of the activated
carbon, fa' ensuring robust high capacity treatment. High surface area per unit volume
provides for bacterial colonization to ensure a high-rate biological treatment. Other

12
advantages include, the ability to achieve low-ppb BTX effluents in shat detention times

(4-10 minutes); no off-gas production; ability to treat low-concentration waste waters; and
biological reduction and adsorption in a Single reactor. The biological factors involved
however, are poorly understood. An improved understanding oftlrese factors will help to
define microbial behavior for a wider range of conditions.

 

populations with the appropriate catabolic pathways are present, environmental restoration
may consist of only optimizing growth conditions for the selected group of microbes.
However, reliance on these populations may be inadequate, especially in sites where
environmental conditions may not be suitable for growth or survival of those
microorganisms. Degradation of specific compounds can be accomplished or enhanced by
introducing organisms with unique physiologic traits at contaminated sites or in
bioreactas. Moreover, the current understanding of the mechanisms involved in
xenobiotic degradation at the enzyme and genetic level (Kukor and Olsen, 1990; Zylstra
and Gibson, 1991) have led research on the construction of recombinant organisms with
novel routes of degradatiar with the use of cloned genes. Many microbial strains have
been genetically manipulated by incorporation of extra catabolic genes or by controlling
their regulatory mechanism (Lehrbach and Trmmis, 1983). In situ experiments have
shown that such microbes can be successfully used for site remediation under selective
pressure (Chakrabarty, 1986). Nonetheless, biosafety concerns persist with respect to the
deliberate release of genetically manipulated populations to the environment and this
remains an area of debate (Tiedje et al., 1989; Dunster, 1992).

Regardless of the possible shat-term beneficial effects of using transgenic
organism to bioremediate contaminated sites, important ecological issues must first be
considered. These include persistence and reproduction of the introduced strain,
interactions with other organisms and with its environment, and possible effects of the
introduced population upon ecosystem function and stability. Our current insufficient

13
understanding of survival strategies ofmicroorganisms in natural systems (Roberson and

Firestone, 1992; Pipke et al., 1992; Siegele and Kolter, 1992; Smets et al., 1993),
increasesrisksassociatedwithreleaseoftransgenic microbes. Although manyoftlrese
mcombinantmicmagamgnswihprobablybelessfitmanindigemuspopulafions,
important exceptions may arise and many generations may pass befae the organisms
disappear completely. Ftuthermore, the observed ubiquity of biofilms, resilience to
paturbations,high cell densityandcloseproximitytooneanother,arrdthecreation and
maintenance of unique microenvironmental conditions give rise to irnpatant ecological
considerations since these communities may serve as potential refuges for non-native
organisms (W agner-Diibler et al., 1992; Warren et al., 1992).

Although natural selection will tend to increase fitness of the released transgenic
microbial population, genetic exchange can certainly be an alternative pathway for the
fixation of a novel trait in a community (Orvos et al., 1990; Stewart and Sinigalliano, 1990;
Ramos-Gonzalez et al., 1991; Romanowski et al., 1991, 1992; Barkay et al., 1993).
Exchange of genetic material among microbial populations has been known to occur for
mac than 50 years (Avery et al., 1944). Bacterial genomic plasticity and gene transfer
mechanisms have beendescribed: (i) transferbydirectcelltocellcontactbetween adoncr
and a recipient cell (conjugation) (Freter, 1983; lppen—Ihler, 1989), (ii) uptake of naked
DNA (transformation) (Stewart, 1989), and (iii) phage-mediated transfer (transduction)
(Kokjohn, 1989). The role of gene transfer in the ecology of microbial communities and
promoting physiological diversity are not known. Carjugal gene transfer in a wide range
of natural environmental carditions, and subsequent generation of new novel phenotypes
has been noticed in aquatic environments (Paul et al., 1991; Barkay et al., 1993). While
the occurrence of plasmid transfer between marine bacteria in aqueous environments has
been observed, there is evidence showing significantly higher transfer frequencies among
cells attached to bead surfaces in biofilm communities (Angles et al., 1993). Also, when
naked nucleic acids are adsorbed onto surfaces, resistance against endonuclease

14
degradation and subsequent bacterial transformation has been repated (Romanowski et al.,

1991). Thus, the influence of non-indigenous populations and their genes on function and
snucnneofsmfacecolmizedccnmunidesisanimpamntaspectofecobgicalfisk
assessment of deliberate release of transgenic organisms to the environment.

" -- Methodologies must be
preceded by a framewak of critical scientific questions for which answers are required in
ordertogain understandingofrelevantecologicalandpractical issues. The selectionof
analytical methods will also depend upon a desired level or scale of resolution (molecular-,

 

cellular-, population-, community-, or system-level). Furthermore, the level of
observations can also be subdivided into properties, processes, and biotic-abiotic
components.

A multiplicity of traditional microbiological culture-dependent techniques,
molecular-based methods, and microscopical procedures are currently available for
ecological studies of biofilm systems (Costerton, 1983; Marshall, 1986; Characklis and
Wilderer, 1989; lad and Costerton, 1990; Ney et al., 1990; Caldwell et al., 1992). A
qualitative and quantitative description of the microbial canposition can partially be
assessed by using traditional enrichment-isolation procedures, plate counts, or liquid
culture techniques. A small fraction of the total diversity however, seems to be represented
on those bacterial inventaies (Wayne et al., 1987; Torsvik et al., 1990; Ward et al., 1992).
Thus, success may be limited by the failure to perceive the niches of many microorganisms
or to reproduce these niches in artificial culture environments. In addition, culturing
bacteria from biofilms requires dissociation of the cells by potentially harmful procedures.
The application of gradient chambers for cell isolation and characterimtion represents an
important improvement for culttning previously described unculturable organisms
(Wolfaardt et al., 1993; Emerson et al., 1994). Multidimensional gradients can be build to
cover a broad range of physic and chemical parameters creating heterogeneous

envirarments for better simulation of natural conditions.

15
Molecularmethodologiesbasedintheuseofbiochemicalmarkers allow the study

of natural communities free of the bias inherent in cultured dependent methods. Molecular
probes targeted to specific biomarkers (DNA, ribosomal RNA, proteins, lipids) are useful
tools for the characterization of phylogenetic, enzymatic, and cell surface composition of
microbial populations (Adachi et al., 1993; McSweeney etal., 1993). In particular, the
extensive phylogenetic work on ribosomal RNA championed by Woese (1987, 1993,
1994) has spanned new frontiers for microbial community analysis. Ribosomal RNA
based methods fa' analysis of community structtue and composition have been
successfully applied to the study of natural systems without the need of culturing the
members of the community (Weller et al., 1991; Spring et al., 1992; Amosti and Repeta,
1994). They have also led to the development of nucleic acid hybridization probes suitable
for spatial analysis of specific populations in name (Anrman et al., 1990; Wagner et al.,
1994). In addition to direct comparison of the nucleic acid sequences (Weller and Ward,
1989; Britschgi and Giovannoni, 1991), numerous groups have used the rapid method of
PCR amplification of this gene as well as the complete rRNA locus as a simple method for
identification of bacterial species and genera (BOttger, 1989; Reysenbach et al., 1992;
Muyzer et al., 1993) . This technique allows a rapid synthesis of desired sequences from
genomic DNA by DNA chain extension from two opposing primers catalyzed by a DNA
polymerase (Mullis and Falcona, 1987; Keohavong and Thilly, 1989; Steffan and Atlas,
1991). In this latter procedure, the amplified DNA (rDNA) is also subjected to restriction
analysis (Moody and Tyler, 1990; Hook et al., 1991; Vaneechoutte et al., 1992; Massol-
Deya et al., in press; see Appendix A). The resulting restriction fiagrnent pattern is then
used as a fingerprint for the identification of bacterial genomes or as quick assessment of
genotypic changes of community samples, over time or between different locations. The
procedures have the potential to differentiate between closely related organisms.
Nonetheless, difi‘erential rRNA recoveries from microbial populations of a natural systems
can introduce bias to the community analysis. The potential bias introduced by sampling

16
problems, cell lysis, or nucleic acids recovery efficiency need to be examined on an

Fatty acids derived from phospholipid components of the cellular membranes of
microorganisms have been used to estimate microbial biomass and provide insight into the
taxonomic and functional diversity ofnatural microbial communities (Vestal and White,
1989; Rajendran et al., 1992; Frostegard et al., 1993). Community-level fatty acid analysis
basically entails the extraction of fatty acids from a sample with organic solvents followed
byanalysisofcertainfractionsoftheextractedmaterial. Thisapproachcangerreratesimple
and reproducible irrdicatas of the relative similarities between microbial communities.
Interpretations of community-level fatty acid profiles have been based on two general
assumptions: (i) that phospholipids comprise a relatively constant proportion of cell
biomass, and (ii) fatty acid composition between taxonomic groups results in "signatures"
which can be detected and interpreted (Vestal and White, 1989). Phospholipid contents
however, varies across taxa (Haack et al., submitted), and lipid and fatty acid profiles are
known to vary quantitatively with changes in the growth environment (Kamimura et al.,
1993; NordstrOm and Laakso, 1992; Nordstrom, 1993). Based on these factors,
interpretation of community fatty acid profiles in terms of biomass or taxonomic
composition, must be viewed with caution until knowledge of the quantitative and
qualitative distribution of fatty acids over a wide variety of taxa, and effects of growth
conditions on fatty acid profiles, is more extensive.

Microbial activity measmements can also be defined at different levels of resolution
dependent upon the questions posed. Ftu'thermore, these measurements can be categaized
as either in situ activity and potential activity. Direct in situ measurements of microbial
activity offers the advantage of obtaining dynamic informatiar of the system. Some of the
available techniques include radiolabeling/autoradiography, radia'espirometry, use of dyes
or inhibitors, and chemical analysis of transfa'mation products. These direct activity

measurements nonetheless have their own problems including penetration of substrates,

17
dyes and reagents through the film matrix. This is particularly true when molecular probes

are use for in situ measurements. Microsensor technology has been one of the most
powerful analytical techniques for characterizing chemical and metabolic activity gradients
in microbial films (Binnerup et al., 1992; Dalsgaard and Revsbach, 1992; Ktihl and
Jorgensen, 1992; De Beer et al., 1993). It can be used to measure high resolution depth
profiles of chemical species of biological relevance such as 02 (oxygenic respiration, or
gross photosynthesis), N20 (denitrification), H+ (proton production/consumption), NO3'
(net nitrate production/consumption), H28 (sulfidogenesis, or anoxygenic photosynthesis),
NIL:+ (ammonia production/consumption), and H2 (hydrogen production/consumption).
Complications however occur due to the diffusive nature of surface films, leading to
fluctuations in water flows, and three dimensional related problems. The microbial
potential activity in biofilm systems can be measrncd by the use of confocal scanning laser
microscopy coupled to fluaescently labeled probes, chemical or electrode analysis of
transformation products by microsectioning the biofilm, and immunological and nucleic
acid probes (Amman et al., 1990; Wagner et al., 1994). Nevertheless, these approaches
present a rather static picture of the metabolic processes at a single moment in time.

Finally, microscopy is one of the best ways to understand the spatial cell
organization in a biofilm. Microscopic techniques are relatively easy to understand and
highly infamative. Observations can answer questions about morphology, cell
arrangements, surface and attachment structures, extracellular materials, and when
canbined with specific probes or dyes, identity and activity of individual cells. Advance
methods currently used for direct visualization of biofilm structures includes X-ray
nricroanalysis (e.g., location of elements in the biofilm), and scanning electron microscopy
(e. g., distribution and morphology of cells at surfaces). Although electron microscopy
offers superb resolution of ultrastructure details (Marcelino and Bensosn, 1992), it is
damaging to specimens and suffers from fixation and sectioning artifacts.

1 8
Nondestructive experimental techniques must be used to examine biological

samples in their natural state. Confocal scanning laser microscopy (CSLM) has recently
emerged as the solution to this requirement. CSLM is a rapidly advancing technology
which offers detailed visualization of thick biological samples in which out-of-focus blur is
essentially absent (Wilson, 1989; Wright et al., 1993). First, the shallow depth of the field
(0.5-1.5 pm) of confocal microscopes allows information to be collected from a well
definedopticalsectionratherthanfiommostofthespecimenasintheconventionalhght
microscope. Second, the CSLM optically sections specimens so that physical sectioning
artifacts observed with light and electron microscopy are elinrinated. Because optical
sectiaring is noninvasive, living as well as fixed cells can be observed with greater clarity.
Examination of specimens can be accanplished while maintaining the integrity of the three-
dimcnsional matrix and with minimum sample processing. Therefae, it is no longer
necessary to disrupt the community superstructure when studying the molecular or
belravioralaspectsoftheirecology. AnodreradvantageofCSLMisthatspecimenscanbe
optically sectioned not only perpendicular to the optical axis of the microscope (x, y), but
also vertically (parallel to the optical axis of the microscope) (Carlsson and Liljeborg, 1989;
Shotton, 1989). Stacks of optical sections taken from successive focal planes can be
reconstructed to produce a three-dimensional view of the sample. Thus, CSLM not only
bridges the gap between light and electron microscopy, but provides a means to observe

structural components and ionic fluctuations of living cells and tissues in three dimensions.

Chapter II
SUCCESSION AND CONVERGENCE OF BIOFILM COMMUNITIES IN FIXED-FILM

REACTORS TREATING BENZENE, TOLUENE, AND p—XYLENE
CONTAMINATED GROUNDWATER

BACKGROUND

Reclamation of contaminated subsurface resources have focused on the use of
microorganisms. The catabolic potential of multiple bacterial populations to utilize aromatic
and alkyl substituted aromatic hydrocarbons either aerobically or anaerobically as carbon
and energy source has been described (Leahy and Colwell, 1990; Ridgway et al., 1990;
Aelion and Bradley, 1991, Hutchins, 1991). When indigenous populations with the
appropriate catabolic pathways are present, environmental restoration may consist of only
optimizing growth conditions for this selected group of microbes. However, reliance on
these populations may be inadequate, especially in sites where environmental conditions
may not be suitable for growth or survival of these microorganisms. Degradation of
contaminating compounds can also be accomplished by introducing and selecting for
organisms with unique physiologic traits at contaminated sites a' in bia'eactors where
groundwater extraction and treatment is required to prevent migration of the contaminant
plume.

Assessment of the fate and influence of introduced populations into heterogeneous
rrricrobial ecosystems has ecological and practical significance. Although the processes of
species succession has been studied intensively in ecology (Gorden et al., 1969; Sirnberloff
and Wilson, 1969; Andrews et al., 1987; Kinkel et al., 1989; McCormick et al., 1991),
little is known about contributing processes on pathways of community development
within microbial populations in natural habitats. There is direct experimental evidence
suggesting that interspecific canpetition within a given trophic level may be an important

19

20
determinant in structuring microbial communities (McArthur and Levins, 1967; McCormick
et al., 1991).

We studiedmicrobial successioninaerobic fluidizedbedreactas (FBR) using
granularactivatedcarbon (GAC)asthebiomasscarrier. Thesurface adhesion ofbacterial
populations with subsequent cell binary fission and exopolymer production, leads to the
fa'mation of biofilms (Costerton et al., 1987; Warren et al., 1992). FBRs are now in
connnercial operation for biaemediation of groundwater contaminated with benzene,
toluene, etlrylbenzene, xylenes (BTEX) and other substituted aromatic hydrocarbons
(Hickey et al., 1991). In turn, our ability to understand the function and behavia' of
mutual and introduced populations in these reactors is of particular relevance since they are
key environmental mediators of the transformations of these chemicals.

Microbial succession was examined using a reactor initially colonized with three
well-studied toluene degrading bacterial strains which were then challenged by errderrric
aquifer organisms residing in the groundwater. Mature biofihn communities from this
reactor were then compared to communities of reactors colonized directly by aquifer
organisms. In addition, the influence of changes in carbon sauce on community structure
was evaluated. Classical growth dependent microbiological methods in combination with
genanic community fingerprints and community-level fatty acid analysis were used to
describe the relative genetic and physiologic relatedness among these biofilm communities.
We found convergence of populations in the seeded and the natural aquifer colonized GAC
communities when toluene was used as sole carbon source. Although all inoculated strains
were outcompeted by the indigenous groundwater populations, survival of one of the
inoculated strains in a mature l 13—day-old biofilm community was observed. This
demonstrated the potential of biofilms as refuges for allochthonous populations. In
addition, an studies showed that the available carbon sources are important factors in

structuring of the community.

 

consisted of l-L sterile glass columns (2.54 cm diameter x 195 cm long) and were operated
as a one-pass upflow system without recycle at a flow rate of 0.2 l/min (Figure 1).
Groundwater from a deep aquifer underlying the MSU campus was amended with toluene
or a mixture ofberrzene, toluene and p-xylene (BTX 1:1:1) at a organic loading rate of 5.4
kg COD/m3-day. In addition, the influent was supplemented with pure oxygen (11 - 13
mg/l inlet concentration), and a nutrient solution (NH4C1 & KI-12P04) to achieve a ratio of
30:5:1 (carbon : nitrogen : phosphorous). Granular activated carbon (GAC; Calgon
Filtrasorb 400, Calgon Company, Pittsburgh, PA) was used as the carrier matrix.
Approximately 100 g dry weight of GAC per reactor was washed with deionized water to
remove carbon fines, and then sterilized by autoclaving (2 h at 121°C) before use. The
GAC had a geometric mean diameter of 0.7 mm and an overall density of 1.6. The
following FBR systems were operated in parallel for this study:

(i) Reactor colonized by indigenous aquifer microbial populations with toluene as
the sole carbon source (referred to as TOL.N S reactor [toluene.non-sterileD.

(ii) Reactor colonized by indigenous aquifer microbial populations with BTX added
as the sole carbon sources (referred to as BTX.NS reactor).

(iii) Reactor seeded with three toluene degrading species (initially referred to as
TOL.S). This bioreactor was operated in sequential stages: (a) day 0 to 53, filter-sterilized
groundwater with toluene as the sole carbon source, (b) day 53 to 124, the filter was
removed and groundwater with toluene continued to feed the reactor, and (c) day 124 to
175, no filtered groundwater was fed but with BTX rather than toluene as the sole carbon
sources. This latter configuration is referred to as the BTX.S reactor [8, seeded]. The
serial filter-sterilimtion system consisted of a 0.30 pm fiber filter, an activated carbon filter
and two 0.45 um membrane filters [Polycap 75 ASW/Bell].

22

 

 

 

 

 

 

 

 

  
    

Effluent
Sarnplin
Port
i R_e-cycI; _
Inoculum \
|
I
I
I Biofilm
— Sampling
I Ports
|
I
' J
I
I Influent Sampling Port
I Toluene WC]
| 3 mg/L KH2P04
I Mixer COD/N/P
100:5:1
| r-
I quflow
— Oxygen
Reservior Oxygen
Mixer
Carbon Filter
QPeristaltic Pump
0.30 um Filter

Aquifer

Schematic diagram showing setup conditions of FBRs used for the
experiments. Filter system was absent in control FBR (TOL.NS and BTX.NS) as well as
in the inoculated FBR (TOL.S) after day 53. Illustration not drawn to scale.

23

W’ Smn' and andim' uhritialcolonrzatron' ' ofthefilter-sterilized
FBRCI‘OLSreactor) usedthreewellcharacterizedtoluenedegrading strains:
Pseudomonas cepacia, strain G4 which carries the toluene ortho-hydroxylase pathway
(Folsom et al., 1990); P. pickettii, strain PKOI which carries the toluene meta-hydroxylase
pathway (Kaphammer et al., 1991); and P. putida, strain Pan which carries the toluene
methyl-hydroxylase pathway on plasmid pWWl (Franklin et al., 1981; Harayama et al.,
1989). These strains were grown to mid exponential phase in basal salt rrredium (BSM)
(Owens and Keddie, 1969) with toluene as the sole carbon source. Cultures were
enumerated by acridine orange direct counts, mixed in equal numbers (approximately 109
cells per culture), and used to inoculate the sterile GAC reactor. After recirculation of the
inoculum for 2 h, cell adsorption to the carbon was apparent since the liquid phase change
from turbid to clear. At this time continuous fading of toluene amended filter-sterilized
groundwater was started.

Selectivemcdiawereusedtodeterminetlrepopulation sizesofinoculantstrainsin
the filter-sterilized FBR. All media were made with distilled water and solidified with 2.2%
(w/v) Nobel agar. For strain G4, 1% (w/v) d-sorbitol was used as sole carbon source in
basal salt media (BSM) (Owens and Keddie, 1969; see Chapter III); for strain PKOI, 10
ug/ml of kanamycin and y-hydroxybutyric acid (1% w/v) was used in BSM; and for strain
Pan the medium was composed of (g/l): sucrose, 10; NaHCO:;, 1; MgSO4-7H20, 5;
K2HP04, 2.3; N-laurylsarcosine sodium salt (SLS), 1.2; 20 mg trimetlroprim (after
sterilization); 10 ml glycerol. Biofilm samples were dispersed, diluted and plated on these
media and incubated at 25°C.

Wm: - Physicochemical analyses of all FBR, including
measurements of toluene or BTX, oxygen consumption, temperature, and pH, were
generally conducted at 2 to 3 day intervals. For BTX determinatiar, 5 ml samples of
influent and effluent groundwater was collected in 10 ml vials, supplemented with
approximately 0.10 ml of 6 N HCl, sealed with Teflon-coawd butyl rubber septa and

24
aluminumcrimps,andstaedat4°Cuntil use. BTXcompoundswereanalyzedusingagas
chromatograph (Varian 3700) equipped with a flame ionization detecta' (GC/FID), a DB-
624 capillary column (1&W Scientific, Folson, CA) and a head space samjrler. The vials
waeequihaatedm40°C,dwcaumnm90°Canddreinjecmranddetecmrat2(XI°C.
Helium was the carrier gas. For dissolved oxygen (DO) and pH measurements, samples
collectedfromoverflowfunnel with builtinstirringbarwcre immediately analyzed with a
polarographic electrode coupled to an Orion digital pH/millivolt meter (Orion Model 611).
Dissolved oxygen consunrption was calculated as the difl‘erence between the influent and

effluent quantities.

 

' v.0 [k --Biofilmsampleswere
collected aseptically from column sampling ports every 2 to 7 days. Cells were removed
and homogenized from the activated carbon (approximately 0.2 g wet weight) by repetitive
va'texing (three times for 45 s) in 1 ml of cell extraction buffer (Warren et al., 1992).
Viable bacterial numbers from GAC samples were determined using medium R2A (Difco,
Detroit, MI). R2A plates were incubated at 25°C for 3 to 4 days and total counts and
relative densities of different colony types were recaded. For determining cells per GAC,
the extracted GAC sample was desiccated at 195°C for a week or until repetitive weight
readings were obtained in an analytical balance. Isolation of the major BTX degrading
populations from the GAC-FBRs was done as described elsewhere (see Chapter III). Total
bacterial cell counts were also determined by acridine orange (A0) epiflourescence
microscopy using the method of Zimmerman et al. (1978) as modified as follows. Biofilm
homogenates, were fixed in 0.1% (v/v) formaldehyde (final concentration) and then staed
at 4°C until processed. Dilutions were made and stained with 0.01% (w/v) A0 [final
concentration] for at least 10 min. Appropriate volumes were filtered through black
polycarbonate filters of 0.2 tun pore size and 47 mm in diameter (Nucleopae Corp.,
Pleasanton, CA). Filters were pre-wetted by passing approximately 1 ml of a 0.1% (v/v)

Tween 80 solution. Excess stain was removed by a last filtration of 3 ml of 25% (v/v)

25
isopropanol. Thedryfiltersweremountedarglassslides. Adropofimrnersionoilwas

placed on top of the filters, followed by examination under epifluaescent illumination. A
nrinirrrumof 10randomly selectedfields with 30to75cellseach werecountedarrdaveraged
for each subsample. A Leitz phase microscope, equipped with epifluorescent condensers
for reflected light excitation was used for the observation and enumeration of microbial
cells.

WW - Nucleic acid extraction (Maniatis et al., 1982)
waspa'famedfromcehpefletsofbactaialisolatesaGACcomunifiesflratwae
resuspended in 100 ul of 25% (w/v) sucrose-10 mM EDTA-50 mM Tris (pH 8.0). TB
buffer(40trl)containinglysozyme (5 mg/ml)wasaddedalonewith40trlof250mM
EDTA (pH 8.2). After incubation at 37°C for 10 min, the cells were Iysed with the addition
of250ttlof2% (w/V) sodiumdodecyl sulfate (SDS),andtreatedwith RNaseA [5|Llof2
mg/ml stock] and pronase E [10 ul of 10 mg/ml in distilled water] for 60 min at 37°C.
Samples were then amended with 120 pl NaCl (5 M), 85 It] TE (pH 8.0), and 80 ul of 10%
(w/v) CTAB solution (cetylttimethylammonium bromide) and incubated at 65°C for 20 nrin.
The cell lysate was then extracted once with an equal volume of phenol-chlaoform—isoamyl
alcohol (25:24: 1, v/v/v), twice with chloroform-isoamyl alcohol (24:1, v/v) followed by a
final extraction with chloroform. The nucleic acid was precipitated by the addition of 2.5
volumes of 97% (v/v) ethanol, washed with 70% (v/v) ethanol, dried under vacuum,
resuspended in TE buffer and quantified by measming the optical density at 260 nm
(Maniatis et al., 1982).

Extracted DNA was subjected to amplification of the ribosomal RNA genetic loci by
the polymerase chain reaction (PCR). Universal primers were used to amplify an
approximately 1,500 bp fragment using a forward primer which corresponds to nucleotide
positions 49 to 68 (5'-AGAGT'ITGATCCPGGCTCAG-3'; primer A) of Escherichia coli
16S rRNA and a reverse primer corresponding to the complement of positions 1541 to 1518
(5-AAGGAGGTGATCCAGCCGCA-3'; primer H) (Ulrike et al., 1989). All primers were

26
synthesizedwithanAppliedBiosystemDNA synthesizerattheMacromolecularStructme

and Sequencing facility at Michigan State University. In general, amplification was done in
100pltotalreactionvolumecontaining lmngoftotalcommunityextractedDNAor
bacterialisolatesastemplate, 1 pMofeachprimer,250pMofeachdeoxynucleoside
triphosphate, 10 pl of 10X Taq buffer, and 2.5 units of T 04 DNA polymerase (Perkin-
ElmerCetus, Norwalk,CI')wasused. PCRamplificaticnwasperformedinanautomated
thermal cycler (96(1) Perkin-Elmer Cetus, Nawalk, CT) with an initial denaturation [92°C,
130 s], followed by 30 cycles of denaturation [92°C, 70 s], annealing [48°C, 30 s], and
extension [72°C, 130 s], and a single final extension [72°C, 6 min]. An aliquot of 5 pl was
run in 0.7% agarose gel to evaluate the quality of the amplified fragment. In general,
amplification yielkd greater than 10 pg of PCR product.

Canmunity (8 pl, 5X concentrated by ethanol precipitation) or bacterial (13 pl)
amplifiedDNA wasdigestedat 37°Cforatleast3 bin 15 pl final volumewithtetrameric
site-specific restriction endonucleases (HaeIII, HpaII, Msel, or Hirtfl; Boelrringer
Mannheimr, Indianapolis, IN). The 168 rDNA restriction fragments were then separated by
electrophoresis on 4% (w/v) NuSieve 3:1 agarose gel (FMC, Indianapolis, IN) in 0.5x
TAE [Tris acetate-EDTA] at 7 volts per cm, and stained in 0.5 pg per ml ethidium bromride
solution. The resulting restriction fragment patterns were then used as a fingerprint to detect
similar communities.

Species-specific primers complementary to the V3 region of the 16S ribosomal RNA
gene of strains G4 and PKOl were constructed for a more sensitive and specific detection of
these strains by PCR amplification (see Appendix B). PCR was performed as described
above but at an annealing temperature of 55°C in 50 pl reaction volumes. After
amplification, 10 pl of the PCR reaction was separated by electrophoresis on 0.7% (w/V)
agarose gel in 1X TAE buffer.

WWW~Twaym°fufi®bifl
canmunities were analyzed using the Microbial ID, Inc. (MIDI; Newark, DE). Duplicate

27
samples of cells collected from fresh GAC samples (approximately 0.1 g wet weight) were

treated as follows: (i) whole community cell preparations were saponified at 1m°C with 1
m1 methanolic NaOH (15% (w/v) NaOH in 50% (v/v) methanol), (ii) esterification of the
fatty acids at 80°C with 2 ml 3.25 N HCl in 46% (v/v) methanol, (iii) extraction of the fatty
acid methyl esters (FAME) into 1.25 ml 1:1 (v/v) methyl-tert-butyl ether/hexane, (iv)
aqueous wash of the organic extract with 3 ml 1.2 % (w/v) NaOI-I, and (v) analysis of the
washed extract by gas chromatography equipped with a flame ionization detector.

Themeanconcenuafionarxlstmdmddeviafionsaswenaspfincipalcomponents
analysis for FAME profiles were determined by using SYSTA'I"'M version 5.1 (SYSTAT
Inc., Evanston, IL).

RESULTS

Wu The influentcellconcentrationofthe groundwaterin theopen
granular activated carbon-fluidized bed reactors (GAC-FBR) was approximately 107 total
cell counts per liter. These microbial populations were a natural and continuous inoculum
source for the colonization of the no seed GAC of the toluene and BTX fed reactors
(TOL.NS and BTX.NS FBRs, respectively), as well as a challenge to the seeded toluene
reactor (TOLS) after removal of the filter-sterilization system. The groundwater at the
bioreactor influent was 13°C (:1: l) and a pH of 7.1 (:t 0.3) throughout the course of the
experiments.

MW - Toluene (TOL.NS reactor) and BTX (BTX.NS reactor) fed
reactors were operated in parallel at the same organic loading rate. A description of the
startup conditions and overall performance is given in Table 1. After a 5 to 7 day lag
phase, dissolved oxygen (DO) consumption increased steadily as did biomass accumulation
on the GAC particles. Steady-state conditions were assumed to have been attained when
DO consumption was constant. This value was also commensurate with the oxygen
required for the complete biodegradation of the added toluene. Steady-state conditions were
generally achieved after 15 days, at this time substrate removal was 98 to 99% . A typical

28
profileofDOandtduenecmcmuafionvdmdepthmdnmacmrueaunentcdumnisshown

inFrgure 2. Ascanbeseen,morethan80%ofthetoluenewasconsumedinthefirst30%
of GAC bed height. Similar results were obtained with the column fed BTX as substrate

(data not shown).

Dissolved oxygen (mg/L)
Toluene (mg/L)

 

ommwmm
95 Bed Height

Figure 2. Typical profile of dissolved oxygen and toluene concentration with depth in
the bioreactor column.

W -— To determine the influences of non-indigenous populations on
community structure, equal numbers (2 x 107 cells/g of GAC) of Pseudomonas cepacr'a
(strain G4), P. pickettii (strain PKOI), and P. putida (strain Pan) preinduced for toluene
metabolism were initially inoculated into the sterilized FBR. The reactor was fed filter-
sterilized groundwater with toluene to allow the establishment ofthe seeded populations
(TOLS). A lag phase of 14 days was observed before significant oxygen consumption or
biomass accumulation occurred (Figure 3). A phase of transient toluene acclimation
occurred from day 14 to 18. This was followed by a 10 day growth phase.

29

aégsfithsgfiaeopsagaps

.83385 has 996 mm .8588 33 EB»? 8u§.§a 2:. n

.8on..— e5 .0529 6:85.. chum «
.8Qhu=8§A§§3§§§EV§Z~

 

 

 

 

Ma 88.9 «86 Ed 9:. KS n: - as tum§+

8 33.9 Sod 3.9 8.”. 883 as - mm mem=8+

ea 28.8 916 GE: and 3 - m— 833,—. mm - c :58 .235 .6 mean.

a 38.9 :36 find 86 e - n Xhm wa bung Khmdz

mm 23.8 08.9 Rudy m3. 5 - e 883,—. an .355. 49—62
32 zmm “a :82 xmm e :32 A53
23— SEE 3B8. Engage A933 8.58 cognac 08:8 sous—38n—

ESEom Ecsfim 8998 3203.5 and :85 we Ewen..— 85385 .898“
“Mm m .5 0:033.
.233 3.3.5308 3 08588.8.“ 23 cause 93.5» mmm¢<0 A 035.

30

 

 

   
      
  

 

    
     
 

 

rr
8 :T Trj I I frj I Y I I 1" rt 1 I I I 1 I 10
7 3 La [Growth] Steady-State} '5
6 L
E 0
$0 5 :' 109 §
: on
a 4 - 10' 9"
3 "r ' ----O-Tolueneefi1uent E3
2 :. J —°--DOCcnsumpticn 7 O
: —D—CeflDerBity 10
1 :-
0 h . . . .. .- . . 106
0 10 20 30 40 50

Days after inoculation

Figure 3. Observed dissolved oxygen consumption (mg/L), toluene effluent
concentration (mg/L), and cell density (cells/g of GAC) rn the TOL. S reactor inoculated
with three different bacterial strains. The bioreactor was fed amended fiter-sterilized
groundwater for 53 days to allow the establishment and stabilization of the inoculum
members.

Analysis of specific population types by selective plating dming the lag phase
(Figure 4), showed a decrease in the densities of strains G4 and PK01 but growth of strain
Pan . During the pseudo steady-state phase strain Pan was the sole recoverable culture.
The increase in cell density ofPan corresponded to an increase in D0 uptake, which
confirmed that toluene oxidation was occurring. At steady-state, the average DO consumed
was 5.70 mg/l and toluene effluent concentration was 0.140 mg]! for a 96% removal
efficiency (Table 1). After 35 days, however, new bacterial types appeared on R2A plates,
which suggested that groundwater strains may have breached the filter. On day 53, the
filtration system was removed from the TOL.S reactor to allow aquifer populations to fieely
enter the already colonized GAC. A lower steady-state effluent toluene concentration was
noted, declining from 0.140 to 0.063 mg]! (Table 1) after the filter was removed. The DO
consumption remained at the previous steady-state level. From these mattue biofilm

31
communities, 12 different toluene-degrading cultures were isolated. Inoculum strain Pan

was recovered from a 113-day-old GAC sample by eruichment for p-xylene degradation.

10"

 

T T U T I I I I I I V T U V I U I T T

   
 
      
   
   

10‘

+04
——u .. .PKOl
.u... ...POWI

 

CFU/g of GAC

lfiLLlllLl

0 5 10 15 20
Days after inoculation

Figure 4. Densities of inoculated strains as determined by selective plate counts over
time in the TOL.S reactor. CFU, colony forming units.

To evaluate the effects of carbon sources on community structure, the toluene
selected community (TOL.S reactor) was switched from groundwater amended with toluene
to groundwater amended with an equal mixture of BTX compounds (BTX.S reactor, Table
1) at the sameorganic loadingrate. NochangesinthecarbonremovalrateorDOuptake
were observed. Although similar bacterial types to those isolated from toluene fed reactors
were obtained, changes in isolation frequency of particular types were noted (see Chapter
III).

WWW - Characterization of bacterial isolates and
community composition was performed by restriction fi'agment length patterns of amplified
rRNA genes from bacterial or community DNA. Greater than 65% of the cells were lysed
as determined by direct fluorescent cell counts. This method yielded approximately 60 ug

32
DNA/g of GAC. The DNA was of high molecular weight (>23.1 kb) and could be digested

by restriction endonucleases. Each of the inoculated strains showed a distinct restriction
fragment pattern with the amplified rDNA when digested by HpaII (Figure 5, lanes G4,
PKOl, and Pan). These three patterns could also be discerned in mixed samples taken
after inoculation (Figure 5, Day 0). By day 22 however, DNA isolated from GAC biofilm
showed only a Pan-like pattern. This finding is consistent with the results obtained from

5”) 9° \\
‘9‘ o“ «2‘50 ”‘29 9'9 01d 9'” 0°“ 9°“ 919‘ sf“

350 -

222 -

126 -

 

 

Figure 5. Community fingerprint of bioreactor seeded with bacterial strains G4,
PKOl, and Pan. The filtration system was removed on Day 53. Shown' rs a HpaII
digestion of the amplified 16S rDNA from reference cultures (Lanes G4, PKOl, and
Pan), and GAC community (Lanes Day 0, 22, 38, 53, 90, and 112). The sizes of
marker DNA fragments (lanes SIZE) (in base pairs) are indicated on the left.

selective plating (see Chapter III). Further changes in band pattern profiles can be seen
over time, reflecting further changes in community composition. An intermediate band
pattern was observed at day 38 before the community profiles remained stable (Figure 5,
lanes Day 53, 90, and 112). Interestingly, band patterns similar to those generated from G4

33
and PKOl can be seen at these latertime points. Because ofthe pure culture results (Figure

4), we suspected that they were from different populations that shared restriction sites with
G4 and PKOl and successfully made it through the filter barrier. In order to examine
whether these strains were present, we used the species-specific oligonucleotide printers for
these strains for PCR amplification. Samples from time zero showed amplification products
for these species specific primers, but bands for G4 and PKOl were not detected at days
22, 38, 53, 90, or 112, confirming that the inoculated strains were not successful colonizers

at these stages.

HaeIII Hinfl M391
S B M T B M T B M T S

 

Figure 6. Restriction fragment length polymorphisms of amplified 16S rDNA from
S.BTX reactor at day 159. Samples correspond to the 20, 60, and 90% reactor column
bed height (lanes B, M, and T, respectively) digested with HaeIII, Him, and MseI. The
sizes of marker DNA fragments (lanes S) (in base pairs) are indicated on the left.

To determine whether position within the bioreactor treatment column lead to
differences in patterns of bacterial composition, samples collected from the bottom (B),
middle (M) and top (T) of the GAC bed height were subjected to RFLP analyses (Figure

34
6). Identical RFLP patterns were observed at all of these heights with all of the three

endonucleases tested.

HaeIII Hian MseI

as $95 .9 $5 $99 8 a“? an. 9
NA «0v 9* «0v 6* «0» 95* «OV'gs‘l‘gng‘f‘ gov‘qs‘f‘gai

      

Figure 7. Restriction fragment length polymorphisms of amplified l6S rDNA from
mature GAC communities digested with HaeIII, Hinfl, and MseI. The lane designations
correspond to the bioreactor treatments (see Materials and Methods). The sizes of marker
DNA fragments (lanes SIZE) (in base pairs) are indicated on the left

The composition of mature GAC communities in all FBRs was also compared by
RFLP analysis (Figure 7). Similar DNA patterns were observed between the seem
(T OL.S) and naturally colonized (T OL.NS) toluene fed reactors and between the seeded
(BTX.S) and naturally colonized (BTX.NS) BTX fed reactors with each of the three
endonucleases. The toluene and BTX fed communities seem to show some differences
from each other as particularly noted with the MseI restriction endonuclease digestion in
which two signature fragments of approximately 460 and 410 base pairs were more

apparent in the two BTX fed reactors. A thorough evaluation of faint bands as well as the

35
comparison by other enzymes, however, showed no band pattern differences between the

systems fed the two substrates. Identical RFLP fingerprints were also observed for strains
isolated from the two GAC communities (data not shown). Band patterns of these isolates

correspond with most of the bands observed in the community fingerprints (Figure 8).

$411;

«Effie 969‘“

b\ b\ 41’
$0 s) C3” (,0 Ssfiv‘i‘s‘gc

’0’
(LOQM

Qcfi C3\O

2,645 -
i

1.605 -
11,198 -

676-
517—

350 -
222 -

 

Figure 8. Genomic fingerprints of bacterial isolated from GAC communities. Shown
is a MseI digestion of the amplified 168 rDNA. Lanes SIZE show the DNA size markers,
indicated (in base pairs) on the left.

W315 -- Reactor community composition was also compared
by fatty acid methyl ester (FAME) profile analysis (Table 2). The coefficient of variation
for most fatty acids was generally below 10% for duplicate analyses. At least 22 fatty acids
were identified among the different bioreactor treatments including saturated, unsaturated,
and methyl and hydroxyl substituted fatty acids. Fatty acids 16:1 w7c and 18:1
w7c/m9t/(n12t have been reported to be characteristic of the genus Pseudomonar (Haack et
al., submitted).

{l

 

36

 

 

Table 2. Relative abundance (% of total) of fatty acids of the GAC-FBR communities.
Fatty Acidi . . . . .
Day 48 Day 98 Day 53 Day 129 S Day 151
Day 22
10:0 0 2.31 0 1.69 0 2.49
10:0 30H 3.28 5.94 2.84 1.64 6.37 7.75
12:0 3.91 21.62 1.59 21.11 4.35 25.07
12:0 20H 0.75 1.61 1.91 0.85 0 5.12
12:0 30H 0.91 1.00 1.42 0.65 1.99 3.99
14:0 2.20 8.24 1.02 9.43 0 7.68
i 15:0 0 1.39 1.24 0 2.55 0
15:0 0 0.53 0 0.69 0 0
14:0 20H 2.45 0.82 5.31 2.93 0 0
16:1 m7c 44.21 28.24 30.37 23.68 47.52 21.34
*i 15:0 20H, 16:1 m7t 0 0 0 0 14.04 8.76
16:0 15.60 11.74 7.12 9.64 17.43 9.85
17:1 «:60 0 0 2.64 0.50 0 0
17.0 CY CLO 0 1.51 0 0 0 2.37
i 16:0 30H 0 0 1.13 0.85 0 0
*18:2 m6,9c, a18:0 0 1.94 0 0.55 0 0
18:1 m9c 0 1.11 0 0 0 0
* 18:1 m7c,(o9t,(o12t 25.66 8.58 41.22 22.63 5.74 5.57
18:0 0 1.49 0 1.85 0 0
*20:4 0 1.24 0 0 0 0
m6,m9,to12,m15c
20:0 1.03 0 2.18 0.69 0 ‘ 0

 

1 Fatty acids are designated as the number of carbon atoms, number of double
bonds and position relative to the aliphatic (or) end of the molecule; prefixes i and
a, refer to iso and anteiso branching respectively; CY CLO and OH indicates
cyclopropane and hydroxyl substitutions respectively. Data are means of replicate
samples.

* a group that MIDI-FAME does not reliably separate.

37
SignificmtamountsofbothofflresefauyacidswereobsavedinauFBRsystems. Some

fatty acids were detectable in only aparticular treatment, for example, 14:0 20H in toluene
fed systems, and i15:0 20H/16:1 m7t in BTX amended reactors. There is also a trend ofa
higher percentage of fatty acid 12:0 in older samples regardless of carbon source. Although
thisfattyacidcanbefoundatlow levelsinsomePseudomonas species (Haacketal.,
submitted), it has been generally associated with microeukaryotic organisms (Shaw, 1966).
Yeasmwdprmomanswaeobsewedmmeolderandmmematmebiofilmflseeflrapter
IV).

Table 3. Relative abundance (% of total) of fatty acids from samples collected at the
bottom and top of the column bed height level of TOL.S reactor at day 101.

 

 

Fatty Acidr Bottom m
10:0 30H 2.10 0 2.20
12:0 3.95 2.97 0.48
14:0 2.14 1.98 0.01
14:0 20H 3.83 5.92 2.18
16:1 (070 41.20 35.04 18.97
16:0 11.60 8.36 5.25
17:1 (1)60 1.69 0 1.43
18:1 m7c,m9t,m12t 31.58 41.84 52.63
19:0 CYCLOm80 0 1.54 1.19
20:0 1.92 2.34 0.09

 

1 For fatty acid description, see Table 2.

MIDI-FAME profiles obtained from the bottom 20% and top 80% GAC bed height
of the TOL.S reactor system at day 101 (Table 3) shows higher variation (greater than
15%) relative to the mean for fatty acids 16:1 (070 and 18:1 m7c/to9t/012t than between
replicate subsarnples (less than 5%). Although the relative abundance of the FAMEs is

38
somewhat different between samples obtained from different positions within the GAC

column, the fatty acids present are virtually identical.

In addition to providing some information on taxonomic structure, community-level
FAMEprofileswereusedtodistinguishedtemporalandtreamrentvariation using
multivariate statistics. Principal component (PC) analysis was conducted using percentages
fm'eachfattyacidwhichappearedinanyofthecommunities. PCscoresobtainedbya
correlationmatrixanalysisarepresentedinfigure 9inthreedimensions,eachaxis
correspondingtothefirstthreePCs. ThefirstPCaccountedfor38% ofthevariancewith
high scores on fatty acids 10:0, 12:0, and 14:0. Separation into two classes of communities
can be observed, old GAC samples (above 0) and relatively young communities composed
mainly of bacteria (below 0).

TOL.NSO
Day 98

 

 

 

 

 

 

 

-2 _2

Figure 9. Principal components analysis showing temporal variations of FAME
profiles for GAC communities. Reactors: toluene or benzene, toluene, and p-xylene fed
systems of seeded (T OL.S and BTX.S) or aquifer colonized (TOL.NS and BTX.NS)
biofilm communities. Parameters are listed in Table 2.

39
ThesecondPCaccountedfor26%ofthevariancewithhighscoresonfattyacidsi15:0

20H/16:1 «fit and 16:0. Finally, the third PC accounted for 15% of the variation in the data
for a total cumulative of 79% with high scores on fatty acid 12:0 20H.

DISCUSSION

Indicators that reflect changes in phenotype and genotype are important for
describing succession, response to disturbances, and dynamics of natural communities.
Weexaminedthemicrobial structureandcompositionofGACcommunitiesbyrestriction
endonuclease analysis of amplified 16S ribosomal RNA gene fiagments and by
community-level FAME analysis. These fingerprinting analysis are simple, rapid and
efficient methods to quickly describe bacterial assemblages and interrelationships between
co-ocurring populations. While RFLP analyses were used as measurements of similarity
among communities and with isolates (Moody and Tyler, 1990; Hook et al., 1991; Adachi
et a1, 1993), fatty acid analyses provided us with insights into the taxonomic composition
(Federle et al., 1990; Frostegard et al., 1993a, 1993b).

The RFLP approach greatly simplifies the characterization ofcommunities over the
previous approach obtained analyzing several of individual samples of cloned l6S rRNA
libraries (Schmidt et al., 1991), because the information is obtained in one or a few
reactions. Disadvantages of our method, compared with analysis of individual cloned
samples, are potential bias during the amplification reaction, limitation of resolving power
on agarose gels for distribution of complex genotype mixtures in the community, and the
information is not quantitative. From a practical point of view, however, the complexity of
the community-level pattern can be controlled by the cutting frequency of the restriction
endonuclease used (Bickle and Kruger, 1993). This feature allowed us to quickly screen
and characterize cultures at different time points as well as to identify not yet isolated
cultures. The correlation between fingerprints determined fiom GAC samples and those
determined for isolated strains was acceptable. We do recognize , however, that

40
considerablephenotypic diversitycouldstill berepresentedbydifferentorganismsgiving
the same bands.

To determine the influence ofnon-indigenous populations on community structure,
threedifi‘erent toluenedegradingstrainswere seededtoasterilereactorfedtoluene
amended, filter-sterilized groundwater. After a prolonged lag phase successful colonization
was accomplished by only strain Pan. The growth phase of strain Pan in the filter-
sterilized reactor was immediately followed by colonimtion of some groundwater
populations which eventually made it through thefiltration system as noted by changes in
community fingerprints as well as by isolation of toluene degrading populations that were
not in the inoculum. To determine whether Pan was a dominant population after the
secondary colonization by native aquifer populations, Southern blots of total cormnunity
DNA from the TOL.S reactor samples were hybridizedto thexyIE gene probe which is
specific for the tol plasmid encoded pathway (Harayama et al., 1989). No cross
hybridization was observed at low stringency (Jizhong and Tiedje, personal
communication), suggesting that Pan was below the sensitivity for the hybridization
assay and therefore not a dominant member of the mature community structure. Thus,
Pan a non-native population that was present as single dominant member after the
controlled pioneer colonization phase of the GAC, was quickly outcompeted by the
endemic aquifer populations. The presence of strain Pan in a 113-day-old GAC sample
nonetheless, demonstrates that an allochthonous population can survive and avoid
exclusion from the biofilm community.

The DNA fingerprints of the stable biofilm community (TOL.S) and the non-seeded
cummunity (TOLNS) converged. In addition, identical bacterial populations were isolated
from the preinoculated GAC reactor after colonization by the aquifer populations (TOL.S)
and from the TOL.NS systems (see Chapter III) further confirming convergence of
populations in both reactor systems. These results suggest little influence of the pioneer

colonist as a determinant of community structure. Possible explanations for these

41
observations are the functional equivalency of strain Pan as a primary colonist. and/or

independenceovershortandlongtermcommunity structurefrompioneercolonizers.
Simepimeerpopubfiomamdkecflyintaacfingwimdnsubsuatummdsimemematme
biofilmcolonizablesurfaceiscells, superiorandbetteradaptedspeciescanccmpeteand
establish themselves even in GAC-surface limiting conditions. These results contrast with
dwseobsavedmmicrobialprimarysuccessionofplanhmicwmmunidesmaquafic
microcosmsinwhich therewasdirectional replacernentofspecies anddirecteffectsof
changesindwaderofcolmizemondwoumomeofcommunitysmmecCormicket
al., 1991).

Our results showed that community-level RFLP profiles reached band pattern
stability indicating that bacterial genetic stability was achieved and maintained on rnatme
colonized GAC communities. Furthermore, the bacterial composition of GAC particles
throughout the treatment column appears to be similar as noted by RFLP analysis.
SwitchingfromtclueneasthemaincarbonsoureetoBTXatthesameorganicloadingrate
however, appears to be an important driving factor in influencing community structtue.
Our RFLP analysis shows transition of the toluene selected community (TOL.S) to a BTX
structure (BTX.S) similar to that observed in the BTX.NS control reactor thus,
demonstrating the flexibility of the biofilm community to quickly adapt to new
environmental conditions.

The FAME community-level profiles from particles collected at different time points
and fiom different sections within the treatment column revealed differences in types and
abundance of fatty acids. Since the genomic fingerprints of the microbial community
suggested stable bacterial composition on temporal and spatial basis within the bioreactor,
changes in abundance of fatty acid appears to be influenced by differences in abundance of
the fatty acid-contributing microbial populations and/or physiological status rather than
changes in the basic bacterial composition. Generalizations about the distribution of certain
fatty acids across microbial taxa have been used for analysis of whole community fatty acid

42
profiles (Federle et al., 1990, Frostegard et al., 1993a, 1993b; Haack et al., submitted).

FanyacidsgenaanyassociawdwimpseudomonadsspedeswaefoundmaNndant
quantities in the GAC communities. These results are in agreement with preliminary
characterizationofsomeofthedominantGAC isolatedstrains (seeChapterIII). Therealso
appamdmbeaposidvecmdadmbetwemasignanuefanywidofmiaoeukaryomsand
ageofthebiofilm. Thisrelationshipcanbeatuibutedtotheobservedsecondary
cdauudonbyyeastandpromwaspecieswhichmquimpfimmymrfwemodificafimsand
longerreproductioncycles. Furthermore, theproportionsofthemonoenoic fatty acid 16:1
m7c, which is highly recognized as being primarily associated with prokaryotes (Vestal and
White, 1989), decreased in older GAC samples. The RFLP method used is sensitive to
only prohryows.

Crurent ecological theories are based heavily on the concept of competitive
exclusion and the assumption that communities exist at competitive equilibrium
(Fredrickson and Stephanopoulos, 1981). Therefore, success or survival is based mainly
on the physiological fitness of each competitor in a stable uniform environment. Since
competitive equilibrium requires that the rates of change of all competitors be zero, and
since the physical environment, predation and other factors are changing constantly, it
seems likely that equilibrium rarely occurs in nature (Levin, 1972; Wiens, 1983). In
FBRs, biofilm—GAC particles are being continuously exposed to different gradients. Since
theoveralldensityofthebiofilmmatrixislessthanthedensityoftheGAC,microbial
growth will tend to reduce the overall density of the biofilm-GAC particle. The particles,
accordingly slowly migrate up the profile of the FBR as the biofilm continues to grow.
Whenexcessbiomassisshearedfromthebiofilm,theparticlesbecomemoredenseand
migrate to the lower portion of the bed. The continuous process of growth and biomass
loss (e.g., shearing and sloughing of dead cells) produces a situation where the GAC
particles slowly migrate up and down the bioreactor treatment column. The net effect is a
gradual exposure, on a temporal basis, of each GAC-biofilm complex to changes in

43
substrate concentration. A change in location will represent a mild disturbance or

regeneration of a niche which may be advantageous for a given group. Therefore, GAC-
biofilmparficleswhichambcalimdmflwbonomofmebioreaauudnbewmpamyfiving
ataricherenergyzonecomparedmpardclesatflremiddleormpofdrereactor. The
constancy of fluctuation in the physical factors due to biological events will result in a state
of non-equilibrium, therefore, competitive equilibrium can never be reached, and
coexistence of trophic equivalent species is maintained.

In this study we combined community-level fingerprints to describe the relative
genetic and physiologic relatedness of biofilm community structure in GAC-FBRs. Using
restriction endonuclease analyses of community-level PCR amplified 16S rDNA and
analysis of isolates, we found convergence of the populations of the inoculated system and
the natural aquifer colonized GAC community. Although all the introduced strains were
outcompeted by the indigenous populations, survival of strain Pan at low density
demonstrates how biofilm systems can serve as refuges for allochthonous populations.
Useful indices can be obtained to distinguish communities and to characterize their
responses to environmental perturbations. It seems certain that by complementation with
other recently developed techniques such as in situ whole cell hybridization using
fluorescent targeted hybridization probes (DeLong et al., 1989; Amman et al., 1992) and
confocal microscopy (Caldwell et al., 1992) better estimates of physiological states, species
interactions, phylogenetic relationships and spatial cell distribution at various levels of
development will be possible. Understanding of microbial interactions in biofilm
communities are needed to improve biorernediation techniques and to better assess

influences of introduced non-native populations on community structure and performance.

Chapter III
ISOLATION AND CHARACTERIZATION OF BENZENE, TOLUENE, AND P-
XYLENE DEGRADERS FROM AEROBIC BIOFILMS IN GRANULAR ACTIVATED

CARBON FLUIDIZED BED REACTORS

BACKGROUND

Biological activated carbon fluidized bed reactors (FBR) have been used for
remediation treatmentofgroundwawrcontaminatedwith monoaromatic hydrocarbons
(Hickey et al., 1991; Voice et al., 1994). This cleanup approach relies primarily on
biologically mediated transformation of the contaminants. Bacterial populations will
colonize the surfaces of granular activated carbon (GAC) particles resulting in the formation
of a complex microbial layer or biofilm communities. The use of microorganisms for
bioreclamation hasrecentlybecornethefocusofattention becauseoftheirabilityfor
restoring such sites at considerable cost savings compared to technologies such as chemical
adsorption or incineration (Kobayashi and Rittmann, 1982). Although the indigenous
microbial p0pulations may be adapted to the existing site conditions, reliance on their
catabolic potential to transform a wide range of pollutants may be inadequate. For
example, the optimum strains capable of quick degradation at high or at low substrate
concentrations may be absent. Inoculation of contaminated sites or treatment systems with
pollutant-degrading strains may be an alternative in sites where environmental conditions
may not be suitable for growth, and groundwater extraction and treatment is required to
prevent migration of the contaminant plume. The objectives of this study were to isolate
and characterize BTX-degrading bacteria from mature biofilms to assess the inherent
substrate utilization kinetics diversity and the phylogenetic affinities among these naturally
occurring populations and introduced populations in GAC-FBRs operated under aerobic
conditions.

 

WK) - A description of the different GAC-FBRs used in this study is
presented in Table 1. In general, bioreactcrs consisted of l-L sterile glass columns (2.54
cm diameter x 195 cm long), operated as a one-pass upflow systems without recycle at a
flow rate of 0.2 Umin. Groundwater fiom a deep aquifer underlying the MSU campus
was amended with toluene or a mixture ofbenzene, toluene and p-xylene (BTX 1:1:1) at a
organic loading rate of 5.4 kg COD/m3day. The selection strategy employed for isolation
of BTX degrading populations from GAC particles included direct bacterial isolation of
homogenized GAC-biofilm samples by plating on non selective medium. The bacteria
were first dispersed and separated from the GAC particles (~0.2 g wet) by repetitive
vortexing (3 times for 45 seconds) in 1 ml of cell extraction buffer. This contained 0.001
M EGTA (ethylene glycol-tetraacetate [Sigma St. Louis, MO]), 0.0014 M Tween 20
(Sigma, St. Louis, MO), 0.01% (wt/vol) peptone and 0.007% (wt/vol) yeast extract in
phosphate buffer (Warren et al., 1992). Bacteria were isolated by diluting biofilm
homogenized samples on R2A media plates for oligotrophic bacteria (Difco, Detroit, MI).
Plates were prepared in duplicate, maintained at room temperature (about 24°C) in sealed
plastic bags, and examined at intervals for colony formation.

Table l. Start-up conditions of fluidized bed reactors used in this study.

 

 

Reactor Inoculum som'ccl Iengmmpmadm Carbon sources
desigpation (days)
TOL.S G4, PKOl, Pan and 0 - 124 toluene
aquifer populations
TOL.N S aquifer populations 0 - 98 toluene
BTX.NS aquifer populations 0 - 25 BTX2

 

1 G4, Pseudomonas cepacr'a (Folsom et al., 1990); PKOl, P. pickettii (Kaphammer et al.,
1991); Pan, P. putida (Harayama et al., 1989).

2 BTX; benzene, toluene, and p-xylene.

46
In addition, parallel liquid enrichments in Teflon coated serum bottles containing

amendedbenzene, toluene, orp-xylenein basal saltmedia (BSM) underaerobic conditions
were used, followed by a secondary transfer and plating onto R2A plates. The BSM
consisted of 15 mM NH4€1, 5.98 mM K2HPO4, 5.51 mM KH2PO4, 0.23 mM CaClz,
0.81 mM MgSO4-7H20, 1.71 mM NaCl with a mixture of trace metals (3.0 M)
composed of (g/L): EDTA, 5.0; Zn+, 0.5; Mn”, 0.14; Fe“, 0.1; CO”, 0.04; Cu“, 0.04;
Mo, 0.06; adjusted to pH 6.0 in destilled water (Owens and Keddie, 1969). The final pH
ofthemediawas7.1. Themedium(75ml)wasaddedtoserumbottles,sealedwith
Teflon-coated butyl rubber stoppers and aluminum crimps. Toluene was added with
microliter syringes to the sterilized medium (50 mg/L final), then aseptically inoculated with
two to three GAC-biofilm coated particles fiom each reactor system and incubated in an
orbital shaker at 25°C.

Bacterialcolmuesobminedfiomdirectorindirectisoladonpromcolswere screened
for differences in colony morphology, size, and pigmentation and were then purified by
repetitive plating (at least three times) and production of discrete, isolated colonies on solid
medium.

WW - We used repetitive exmscnic
palindromic (REP) sequences and the polymerase chain reaction (PCR) to fingerprint
bacterial genomes. REP oligonucleotides produced clearly reproducible and resolvable
DNA bands by agarose gel electrophoresis following PCR amplification (de Bruijn, 1992).
These DNA band patterns generate fingerprints for typing of bacterial cultures. REP-PCR
amplification was performed essentially as described previously by Versalovic et al. (1991).
The DNA sequence of the REP primers are: 5' - IIIICGICGICATCIGGC - 3', primer
REPlR-I, and 5' - ICGICI'I‘ATCIGGCCTAC - 3', primer REP2-I. Template DNA (50
ng) was used per reaction in a 2541] volume containing 50 pmol each of two opposing
primers, 1.25 mM deoxynucleosides triphosphate, and 2 U of AmpliTaq DNA polymerase
(Perkin-Elmer Cetus, Norwalk, CT). The amplifications were performed with an automated

47
dramflcycler(mcdel96mPakin-ElmaCemaNorwdkCDwidlanuudaldemnuafion

[95°C, 6 min], followed by 30 cycles of denaturation [94°C, 1 min], annealing [40°C, 1
min], and extension [65°C, 8 min], and a single final extension [65°C, 16 min]. After
amplification, IOuJoftheREPproductswereseparatedon 1% (w/V)agamscgelcontaining
1X TAE (40 mM Tris-HCl, pH 7.5, 5 mM sodium acetate, 1 mM EDTA [disodium
ethylenediaminetetraacetateD, stained in 0.5 [lg/ml ethidium bromide solution, and then
photographed in UV light with Polaroid Type 55 film.

«a -Axenicculturesweretestedfor
theirabilitytoutilizevariousmonoaromatichydrocarbonsasthe solecarbonandenergy
sourceinaerobic liquidmedia. Freshbacterialcolonies grownonRZAplateswere
resuspended and washed in BSM, and transfer to 10-ml sterile vials containing 5 ml of

 

BSM with 10 mg of toluene, benzene, or p-xylene per liter (final concentration) in triplicate
sets. The inoculated vials were sealed with Teflon-coated septa, and aluminum crimp
sealed, and incubated stationary for 2 weeks at 25°C. Substrate removal analysis was
performed using a gas chromatograph equipped with a flame ionization detector (GCIFID),
a DB-624 capillary column (J&W Scientific, Folson, CA) and a head space sampler. The
vialswereequilibratedat40°C, thecolumnat90°Candtheinjectoranddetectorat200°C
Helium was the carrier gas. Positive results were interpreted when more than 80% removal

relative to the control was observed.

 

biofilm populations were further characterized based on their ability to utilize toluene.
Duplicate sets of exponentially growing cultures in BSM containing 50 mg of toluene per
liter were used to determine maximum growth rates by optical density measurements
(ODsso). Further, the Monod growth kinetic parameters, half-velocity coefficient (K5) and
maximum specific rate of substrate utilization (k), were estimated by fitting progress curve
data to the integrated Monod equation, using a non-linear analysis. Both kinetic parameters
were calculated from a single substrate depletion curve as a function oftime (t). There is a

43
directrelationshipin theMonod equation (equation 1), betweenthenumberofcatalytic

units (biomass, X) and the rate of substrate utilization (dS/dt):

dS kXS

dt Ks+S

 

(equation 1)

If biomass is constant, the Monod equation can be simplified to the integrated form
for parameter estimation by non-linear regression:

K: S 1 .
k So k (S So) (eq on )

Ks SO 1
ln—+— -
X]: 5 x1550 S)

 

t:

Exponentially growing cells in BSM with toluene as sole carbon source (50 mg/L
final) were used for the kinetic assays. A cell suspension of axenic harvested cultures
(approximately 1010 cells/L) was transferred to fresh preoxygenated media with 3 mg/L
toluene (final concentration) and immediately drawn into a 50-ml gastight syringe
containing a small magnetic stir bar. The biomass concentration was determined
empirically for each culture (high enough not to increase significantly with the amount of
substrate added and low enough to allow collection of samples at reasonable time
intervals). All assays were performed at room temperature (24 to 25°C). Samples of 1 ml
were collected in 10-ml sterile vials every 5 to 15 min and fixed with approximately 0.05
ml of 7N HCl, sealed with Teflon-coated septa, aluminum crimp sealed and stored at 4°C
until head space analysis was performed (generally within a week). Cell numbers were
determined before and after the assay by direct plating on R2A medium. No significant
increased in the cell number was observed. Three of the cultures were independently
assayed twice; the obtained results for k and K5 fell within the 95% confidence interval.

 

ribosomal RNA gene was amplified from genomic DNA using the polymerase chain
reaction andclonedintoaplasmidvectorasdesaibedpreviously (Zhouet al.,
unpublished). The 168 rRNA gene was sequenced from both directions using an
automated fluorescent sequencer with the forward and reverse primers, which span
Escherichia coli 168 rRNA gene positions of 785 to 805 and 1115 to 1100, respectively.
The DNA sequences were analyzed using the programs in Genetics Computer Group
software package (Deveraux et al., 1984) and in PHYLIP phylogeny inference package
(Felsenstein, 1989).

RESULTS

MW - To isolate numerically dominant
bacteria from GAC-biofilm communities in fluidized bed reactor systems, the end point
dilution scheme was carried out using separated and dispersed biofilm bacteria from GAC
particles and plated on R2A solid media. Aerobic heterouophs were observed at a
concentration of 109-1010 CFU/g of GAC. Bacterial populations fi'om dispersed biofilms
were highly recoverable on R2A medium as noted by a high plating efficiency as compared
to total direct counts 00-45% plate counts/direct counts). Successful enrichments for
aerobic BTX degraders were routinely obtained within 24 h fiom GAC-bacterial coated
particlesasdeterminedbyincreasedcultme turbidity andsubstratedisappearance (higher
than 95% of total added substrate). Isolates were subcultured three additional times to
ensure culture purity. Several clones were retrieved that were able to grow aerobically on
at least one ofthe BTX compounds as sole source ofcarbon and energy. Isolates GF161,
GF261, GF262, and GF281 were numerically dominant among all the reactor treatments
constitutingmorethan90%ofthetotal biomassatfluctuatingratiosonatempcral and
spatial basis. Five of one hundred twenty recovered isolates from the different reactor

systems failed to transform any of the BTX substrates in batch culture conditions.

50
W—Pmifiedbactaialcbneswerefmdrachmactefizedbydw

REP-Pal fingerprinting technique. The REP-PCR analysis revealed distinct DNA band
panermwhichdeadyaidedindeidendficafimmddifimdadonatahighlevelof
resolutionofdiva'sebacterialstrainsmgure 1). PCRwiththeREPprimersetand
chromosomal template DNA fiom the difi'erent cultures yielded multiple distinct DNA
productsofsizesrangingfromapproximately300to6,000basepairs. Thepatterns
generated were found to be very different for several strains examined from the same
system. However, similar amplified genomic band patterns were observed among
dominant isolates recovered fiom seeded (TOL.S) and non-seeded (TOL.NS and
BTX.NS) reactor treatments fed toluene or BTX compounds (e.g., Figure 1, lanes GF1T2
and GF2T3). This result is consistent with those results obtained by community-level
fingerprints in which convergence of seeded and non-seeded communities were observed.
Thus, 12 bacterial clones were identified that differed from one another in REP
fingerprints, and in other morphological and physiological properties. Only isolate
GFZXUhadabandpattemprofile similartoseededstrainPan. Thisclonewas
recovered from a secondary transfer of a p-xylene enrichment culture from a 113-day-old
GAC sample of the seeded toluene fed reactor system. Furthermore, clone GFZXU has the
same aromatic substrate range, similar distribution of contributing cellular fatty acids, and a
large plasmid of the same size of strain Pan (Massol-Deya et al., unpublished). From the
12 different GAC isolates, seven were selected for more intensive investigation in addition
to the three pseudomonads strains used to seed the sterile GAC-FBR.
We: -- The maximum substrate degradation rate (k) and half
velocity coefficient (Ks) for toluene utilization were determined by measuring toluene
disappearance rates of 5 mg of toluene per liter. The patterns of toluene utilization by the
different degrading populations were S~shaped and the data were analyzed by using
regression fits to the integrated form of the Monod equation (equation 2). Description and
comparison of the kinetic coefficients for toluene utilization by the inoculated strains and

51
aquifer indigenous GAC colonizers are presented in Table 2. The apparent values for k

ranged fiom 0.061 to 2.061 mg toluene/1010 cells-h for strains G4 and GF272
respectively. Observed K, values ranged from a low of 0.074 mg/L for strain GF161 to a
high observed value of 1.701 mg/L toluene for strain GF272. In general, all GAC isolated
strains examined showed a lower Kg and/or higher k values than the seeded populations.
Although strain GF272 has a relatively high K; constant, as compared to the rest of the
toluene degrading populations, it also has the highest substrate utilimtion rate constant.
Themajorityoftheisolatesshowedthecatabolic abilitytoutilizemorethanoneofthe
monoaromatic compounds tested. Estimates of pm as determined by turbidity
measurements on toluene ranged from 0.10 h‘1 (doubling time T4 = 6.9 h) to 0.52 h'1 (Ta =
0.93 h) (Table 2).

 

nucleotide bases, corresponding to E. coli 16S rRNA gene sequence positions of 810 to
1090, were determined for each of the three dominant GAC isolates. Similarity and
evolutionary distance matrix analysis suggest closed affiliation to three subclasses of the
Proteobacten'a, strain GF261 to the beta subgroup bacteria Comomonas testosterom' with
an actual similarity coefficient of 0.929, strain GF262 to the gamma subgroup bacteria
Pseudomonas putida with similarity of 0.989, and strain GF281 to the alpha subgroup
bacteria Zymomonas mobilr's with actual similarity of 0.897. A phylogenetic tree
constructed using neighbor-joining distance method shows this three-way relationship
(Figure 2). Very similar tree topologies were also obtained using maximum parsimony
and maximum likelihood methods (data not shown).

52

«C A? {23*

b\ \ o. \*l~\/\"r “3'0
Vi“;c3"\c3?“soci'tqslé’tfiflcft\cfi\Csd"C>"¢chs‘imc‘t ’0“ c8

      
  
   

   
   

‘1‘ .
OQ’LC)Q VC)‘. as (100981

   
   

  
 

  
 

<8 "3469;1’\ \ i
\ \Q‘Qol a

 

Figure 1. REP fingerprint patterns of bacterial isolates from GAC-biofilm
communities. DNA molecular weight standards (lanes MW) are UHindIII, pUCl9/I‘aq-I-
pUCl9/Sau3AI (Stratagene, La Jolla, CA). The pattern of PCR products generated of
chromosomal DNA from inoculum strains PKOl and Pan as well as a positive control
(P. aeruginosa) are included.

53

.etugaei 05 .«c mégfia dam—«5 «8539. <2”— ?88et m9 398 co 38m u
80:? E2— .x 8:029 .H 8588 .m _

 

 

 

- - - ~89 33 ._. .m an a=§§< E0
wnmd amNdv cad Anmedv awfifi 36.8 and X .H in... hangout gray
an." 38.3 ab." £653 SON 8:: and H. .m dam §§ouk Nan—U
anN Ammcdv wacd Aouodv 8N6 . X .H. .m Ngmmgfiam 6 59—0
25.2 38.8 awed Anwmdv man; 399 mud X .H .m «conceals» r NeNn—O
:9“. 83.8 wood 88.8 and 398 26 K .h. .m Neonatal—a a 380
awed €2.an #56 33.9 256 ASA: mmd X .H .m dun flight gammy
mace 33.8 wad awndv and 899 cvd X F GER—EN ~3~m
hde 82.8 had Cmcdv weed Amcdv 2.6 .H .m .8593 .K :5:
ozd 32.9 09.6 Smodv ~86 €de mud .H .m 3.88 .k #0

3.8 Be 5m: some UV 25: Se zetm.» meme—c
33- 3:8. 95 Ash—.38 22:8. 95 9...: ~88 1 883.6 Seam—809 859.—
uur H
3:8 «89.0 .5588 as mean... 8:385 no asoceooo one: are.» 83.9 co eaeaaoo as 88.53 .n 23.

“i, !' u; .' .

‘-
1 .)‘\'.
.-r.'
I
'1‘: s.

54

Alcaligenes entrophus
I I EPseudomonas cepacia
P. cepacia strain G4
I I GF261
Comomonas testosterom'

P. aeruginosa strain NIH 18
P. mendocina
P. putida strain Fl
GF262
P. putida strain 1’an
Erythrobacter longus strain OCh 101
l—Zymomonas mobilis subsp. mobilis
L—Z. mobilis subsp. pomaceae
GF281
Rhodospirillum rubmm strain ATH 1.1.1
P. diminuta
Bacillus subtilis
0.05

 

 

 

 

 

 

 

 

 

Figure 2. Phylogenetic distance tree produced from partial 16S rRNA sequences
(810-1090 positions). Representative members of the various subgroups of the
Proreobacten'a and the dominant toluene degrading bacterial populations from GAC biofilm
communities are shown. 168 rRNA sequence of Bacillus subtilis was used as an

outgroup.

55

DISCUSSION

DegradativebicfilmsinFBRsareccmplexccmmunitiescomposedcfanumberof
coexisting bacterial species. Twelve different toluene degrading populations were
recoveredfromthesereactorsystems. 'I‘hesamestrainswerealsoisolatedfromthereactor
seededwiththreePseudomonas spp. andfromtheBTX groundwaterfedreactor. These
results complement the previous observation of community convergence of the different
reactor treatments as determined by community-level fingerprints and fatty acid methyl ester
analysis (see Chapter II).

The successful use of microorganisms for biorernediation purposes requires the
selection of populations able to effectively degrade environmental pollutants over a wide
range of concentrations. The integrated Monod equation in conjunction with nonlinear
mgressionpmcedmeshasbeenpreviouslyusedmobtainesfimamsofkdesfmpme
cultrrres of organisms (Counotte and Prins, 1979; Greer et al., 1992; Robinson, 1985;
Robinson and Tiedje, 1983; Simkins and Alexander, 1984). This technique has several
advantages over estimation parameters from chemostat data, in that it is considerably faster
since estimates can be derived from a single depletion curve, and volatile and low water
soluble compounds can be feasible tested. Our data suggest that, in fluidized bed reactors,
a kinetic diversity for toluene utilization exists among naturally coexisting microbial
populations in GAC-biofilm communities. The lt/Ks ratio, which is the slope of the Monod
equation at lowest substrate concentration, offers a simple way of emphasizing both half-
saturation value and maximum rates (Healey, 1980). This ratio increases as It increases
and K. decreases, thus a higher ratio indicates a higher rate at lowest nutrient
concentrations, and therefore competitive advantage in the process and conditions being
considered as competition drives nutrient concentrations down. Despite the selective
opportunity of the introduced populations to colonized the GAC particles, they all were
outcompeted by the indigenous aquifer populations. The dominant aquifer populations

56
examinedtodateshowed betterratiosofk/thhantheintroduced strains. However, 1’an

whichwasflnmflyinuodwedpoptfladonmmssfunyrunainmdresystem(uaminm
membercfthematureccmmunity), hasarelatively highk/K.ratio. Thisresultsalso
mggestMendanicpcpuhdmsambeuaadapwdwthebcalenvirmmamlcondidms
and thus, competitively superior than the introduced populations. Maintenance of
funcdonaflyredundantspeciesindwbiofilmwmmmfityappemmbefavaedbyboth
biological and physical factors. The known development of substrate gradients through the
treatmentcolumnandbiofilmmatrixappearsto sustaintheestablishmentandcoexistenceof
kinetically diverse bacterial populations, possibly by niche partitioning.

Partial 16S rRNA sequences weredeterrnined foreachofthe threedorninant
toluene degrading strains with the aim of establishing phylogenetic affinities of the biofilm
community. The full sequence of this gene is not necessary for phylogenetic assignment
and it has been shown previously that phylogenetic trees constructed by using different and
limitedportionsofthe168rRNAsequencearethesameasthosegeneratedbyusingfull
sequences (Lane et al., 1985; Schmidt et al., 1991). We focused on a portion of the 168
rRNA gene that corresponds to nucleotides 810 to 1090 in the E. coli rRNA which contains
substantial sequence variability in conserved secondary structure for rapid sequence
comparisons. The rDNA sequences examined were affiliated with or closely related to
three groups of the Protcobacterr'a, the alpha, beta and gamma subdivisions. Members of
the beta and gamma subdivision such as many Pseudomonas spp. are common aerobic
toluene degraders (Zhou et al., unpublished). The alpha subdivision of Proteobacten'a
however, contains many genera of oligotrophic bacteria including Caulobactcr and
Hwhomicrobiwn but none of them have been previously described to grow aerobically on
aromatic hydrocarbons such as BTX. Thus, strain GF281 appears to be a phylogenetically
novel BTX degrading population.

Five pathways have been described for the metabolism of toluene by oxygen-
dependent reactions (Van der Meer et at., 1992; Zylstra and Gibson, 1991). Nucleic acid

57
probes for the first steps of all five aerobic degradation pathways were used in Dot blot

hybridization assaystodetermineifanygenehcrnologyexistbetweentheGACisolates
and the known pathways. None of the GAC cultures tested to date however, showed any
hybridization signal even though low stringency conditions were used (Marcos and Tiedje,
pascaal communication). These results suggests that more diversity exist at the gene
sequence of toluene oxidation pathways than what is currently known. In addition, GAC
isolate GF262aPscudomonaspua'da strain asdeterminedbyphysiological andpartial 16$
ribosomal RNA gene sequence analysis, appears to differs from those previously described
pathways in P. pua'da strain PpFl (meta ring cleavage) (Zylstra and Gibson, 1989) or
strain Pan (methyl group oxidation tol plasmid encoded pathway) (Harayama et al.,
1989).

In the present study we have shown the diversity of multiple toluene degrading
populations in aerobic biofilms of fluidized bed reactors. The observed coexistence of
these populations in such communities can be attributed to their growth kinetic diversity
and to the presence of substrate gradients through the treatment column and biofilm rmtrix.
It is certainly possible to complement the information obtained from the phylogenetic
analysis of axenic cultures with community-level analysis by selecting molecular based
probes for the different subclasses of the Proteobacten'a (Manz et al., 1992) with in situ
whole cell hybridization (Amann et al., 1990) and confocal microscopy (Caldwell et al.,
1992) to specifically follow, enumerate and study the distribution of these different
populations in the biofilm community. In order to address such important questions
however, non-destructive fixation and hybridization procedures must first be developed for
biofilm studies.

Chapter IV
CHANNEL STRUCTURES IN AEROBIC BIOFILMS FROM FIXED-FILM REACTORS
TREATING CONTAMINATED GROUNDWATER

BACKGROUND

The contamination of groundwater resources with aromatic and alkyl substituted
aromatic hydrocarbonshasbecomeaseriousthreattohumanhealthmr, 1985). Leakage
fiomundergroundgasolineandpeuoleumstcragetanksmnddisu'ibution systemsrepresent
a major source of this pollution. Pump and treat remediation techniques have often been
implemented to control contaminate plumes (Hickey et al., 1991; Voice et al., 1994).
Fixed-film biological systems developed on porous media have recently been employed for
thetreatmentphase. Eitherthrough attachmengcragglomeration, bacteriamayremainin
place within the porous medium forming biofilm communities. The granular activated
carbon-fluidized bed reactor (GAC-FBR) couples the absorptive and high surface area per
unit volume properties of activated carbon with biological treatment to obtain robust high-
capacity treatment (Ehrdardt and Rehm, 1989, Speitel et al., 1989; Hickey et al., 1991;
Voice et al., 1994). Under optimum conditions greater than 99% of the total applied organic
load can be removed using 5 to 10 min hydraulic retention times.

Past structural studies of aerobic biofilms by using traditional light microscopy,
transmission and scanning electron microscopy, show randomly distributed cells. This has
led to process models which assume generally homogenous biofilms (W illiarnson and
McCarty, 1976; Criddle et al., 1991). Bacterial cells living in a dense multilayer matrix
however, face the problem of limited transport of substrate and nutrients into the film inner
layers as well as export of waste products out of these deep regions. The use of
microsensor technology has shown the existence of steep oxygen gradients within aerobic
biofilms, thus demonstrating the importance of diffusion processes in biofilm systems
(Dalsgaard and Revsbech, 1992; Kiihl and Jargensen, 1992; Ramsing et al., 1993).

58

59
Temporal and spatial stability of biofilm communities against disturbances or lethal factors

have been demonstrated (Blenkinsopp et al., 1992; Brown and Gauthier, 1993). The
resilience of such systems to environmental changes has been largely attributed to the
interdependency of the community members for nutrient exchange, and to the creation and
maintenance of microenvironmental conditions which can be completely different from the
adjacent liquid phase (Brown et al., 1988; Criddle et al., 1991).

Withtheuseofconfocal scanninglasermicroscopy(CSLM),detailedexamination
ofmicrobial biofilms in nondestructive states has begun. Advantages ofCSLM to study
biofilm structures were described by Caldwell et al. (1992). Careful examination by CSLM
ofhbomwrybiofilmshasmvededmepresenceofspedficceudisuibufimpauansand
large inter-ml void spaces (Lawrence et al., 1991; Wolfaardt et al., 1994). In view of these
findings and the importance of efficient fixed-film processes, we investigated whether there
wa'e any structural traits in biofilms that might be important to biofilm functions. The
specific objectives were to describe structural patterns and properties of newly colonized
GAC in a laboratory reactor operated at an organic loading rate typical of full-scale reactors,
and to assess whether or not common spatial arrangement and cell distribution strategies

exist regardless of carbon sources or microbial composition in commercially operating field

 

one-liter working volume sterile glass column reactor (2.54 cm diameter x 195 cm long)
was used in this study. The bioreactor was operated as a one-pass up-flow system without
recycle at a flow rate of 0.2 liter per min. Groundwater from a deep aquifer underlying the
MSU campus was amended with 32 moles toluene per liter (3 ppm). In addition, the
influent was supplemented with pure oxygen (approximately 406 M final concentration),
and a nutrient solution (NH4C1 & KHgPO4) to achieve a ratio of 305:]

60
(carbonznitrogen:phosphcrous). About 100 g of sterile GAC (Calgon Filtrasorb 400,

Calgon Company, Pittsburgh, PA) was added to the reactor as adsorbent/biomass carrier.
The geometric mean diameter was approximately 0.7 mm with an average density of 1.6.
Thebedheight was controlledatthedesired level (maximumworking heightof 180cm) by
ashockhushhrgpmcedmetoparficlesneardretopofdrebedtommoveexcessbiomass
fromtheGAC. TheGAC,nowwithathinnerbiofilmreturnedbygravitytonearthe
bottomofthereactcrcolumn (becauseoftheincreaseinparticledensityduetoremovalof
most ofthe biomass) while the sheared biomass was carried away in the effluent stream.

In addition, we examined field-scale GAC-FBRs currently used for the treatment of
groundwater contaminated with complex mixtures of hydrocarbon compounds at different
sites in Michigan (Table 1). Samples were collected fiom mature biofilm communities and
stored at 4°C (generally less than 24 hours) until analysis was performed.

 

 

Table 1. Description of GAC-FBRs used in this study.
Length of
Bioreactor Location] Flow rm Carbon W W?" ”
(Umrn) sampling
(month)
TOL MSU, East Lansing, MI 0.2 Toluene 6
BTEX MBI, Lansing, MI 1.0 Benzene, toluene, 1.5
xylenes, & ethylbenzene
TC Traverse City, MI 60 - 80 Gasoline residues 1
PW MBI, Lansing, MI 1 - 2 Complex mixture of 12
aromatic hydrocarbons2
JEN Grand Rapids, MI 80 - 100 Gasoline residues 6

 

1 MSU, Michigan State University; MBI, Michigan Biotechnology Institute.
2 This waste stream contains 7% salt.

Analxticalmedm - Analyses of physicochemical parameters of the toluene fed
GAC-FBR were conducted at 2 to 3 day time intervals. Analyses included toluene

61
disappearanceoxygenconsumption,temperature, ande-I. FortoluenedeterminationJml
samplesofinfluentandeffluentgroundwaterwerecollectedin 10mlvials. Inordertostop
any further biodegradation or loss ofvolatiles, approximately 0.10 ml of6 N HCl was
added to each liquid sample. Tubes were sealed with Teflon-coated butyl rubber septa and
aluminum crimp seals, and stored at 4°C. Analysis was generally performed within a week.
Toluene was analyzed using a Varian 3700 gas chromatograph equipped with an automatic
headspace samplerandaflameionizationdetectorm). Samplequantification was
performed using external standards. A polarographic electrode (overflow funnel with built-
in stirring bar) coupled to an Orion digital pH/millivolt meter (Orion Model 611) was used to
measured dissolved oxygen (DO) concentration and pH. Dissolved oxygen consumption is
reported as the difference between the influent and effluent readings.

WW -- Microbial cells were removed from the activated carbon
(~02 g wet) and homogenized by repetitive vortexing (three times for 45 seconds) in 1 ml
of cell extraction buffer. The extraction buffer contained 0.001 M EGTA (ethylene glycol-
tetraacetate [Sigma]). 0.0004 M Tween 20 (Sigma), 0.01% (w/V) PCPtone and 0.007%
(w/V) yeaSt extract in phosphate buffer (Warren et al., 1992). Bacterial numbers from GAC
samples were determined by acridine orange (AO) epiflourescence microscopy using the
method of Zimmerman et al. (1978). Biofilm homogenates (1 ml), were fixed in 10 ml of
Milli Q filter-sterilized H20 with 0.1% (v/v) formaldehyde (final concentration) and then
stored at 4°C until processed. A Leitz microscope, equipped with an epifluorescent
condenser and a KP 490 blue excitation filter was used for the observation and enumeration
of microbial cells.

Microbial activityofbiofilmhomogenateswasdeterminedbymeasrmingratesof
degradation of toluene. Duplicate samples were incubated in basal salt medium (Owens and
Keddie, 1969) containing 109 pmol/L of toluene. The inoculated medium (5 ml) was added
to 10 ml vials, sealed with Teflon-coated butyl rubber stoppers and aluminum crimp seals.
Samples were incubated at 25°C, fixed at different time intervals with approximately 0.10

62
mlof6NHCLandtlrenheadspaceanalyzedfortoluenedisappearance;subsampleswere

amended with 0.1% (v/v) formaldehyde for A0 direct cell counts.

WuSamplesforscanningelecuonnnaoscopy (SEM)
were fixed with 0.1 M phosphate buffer (pH 7.0) containing 4% glutaraldehyde for 1 h,
washed with 0.1M phosphate buffer for 10 min, and then dehydrated through a graded
series of ethanol solutions (25, 50, 75, and 100% ethanol). The samples were then critical-
point dried and coated with gold. SEM micrographs were taken with a JSM 35C SEM
(JEOL, Ltd., Tokyo, Japan).

WWW - A Zeiss 210 laser scanning confocal
microscope (LSM) (Carl Zeiss Inc., Thornwood, NY) was used in this study. Dark field
images,forthepurposeofmeasuring biofilmthickness, wereacquiredinlaserscanning
transmitted mode using a Phase 3 condenser with a 10x (non-phase) objective. Tle
samples, which were stained with acridine orange (0.(X)5% final concentration) and
mounted in 0.1% agarose beds, were examined using the488 line ofadual line argon ion
laser with an LP520 barrier filter. The samples consisted of 10 to 15 biofilm—coated GAC
pardclesfieslflycouecedatdiffmentdmesfiomdebottommmdrebedheight Images
wererecordedon Kodak Tmax 100filmusingaMatrix Multicolorcomputerizedcamera unit
(Agfa Matrix, Orangeburg, NY) which received signals directly from the LSM.

Windeis - Total fatty acids of mimbial
communities were analyzed as previously described using the Microbial ID analytical system
(MIDI; Newark, DE) (Sasser, 1990; Haack et al., submitted). Duplicate samples of cells
colleced from fresh GAC samples (approximately 0.1 g wet weight) were treated as
follows: (i) whole cmnmunity cell preparations were saponified at 100°C with 1 ml
methanolic NaOH (15% (w/v) NaOH in 50% (v/v) methanol), (ii) esterification of the fatty
acids at 80°C with 2 ml 3.25 N HCl in 46% (v/v) methanol, (iii) extraction of the fatty acid
methyl esters (FAME) into 1.25 ml 1:1 (v/v) methyl-ten-butyl ether/hexane, (iv) aqueous

63
wash of the organic extract with 3 ml 1.2 % (w/v) NaOH, and (v) analysis of the washed

extract by gas chromatography.

Themeancoreenuadonsandstandarddeviafionsaswenasprimipdcomponents
analysis for FAME profiles were determined by using SYSTATTM version 5.1 (SYSTAT
Inc., Evanston, IL).

 

concentration ofbacterialcellsintheinfluentwaterfedtothetolueneamendedFBR system
(POL) was approximately 107 cells/L. These microbial populations were used as a natru'al
and continuous inoculum for the colonization of the virgin GAC. The inlet water
temperature was approximately 13°C (1 1) throughout the course of the experiment. The
pH of the bioreactor influent and effluent was neutral.

After an initial 5 to 7 day lag phase, the dissolved oxygen (DO) consumption
increased steadily as biomass accumulation began (Figure 1). Steady-state conditions were
assumed to have been attained when constant DO consumption occurred; this amount was
also commensurate with that required for complete oxidation of the added toluene. These
conditions were achieved after 15 days. The average DO consumed was of 129 trmol/L.
The effluent toluene concentration was 0.72 pmol/L. A 98% toluene removal efficiency
was observed.

Several observations suggest that toluene biodegradation commenced once
significant colonization of the GAC occurred. (i) Cell density as observed by bed height
increased (Figure 1) and direct cell counts (data not shown) increased coincident with
oxygen consumption, indicating that growth had occurred. (ii) Microbial populations
capable of utilizing toluene as sole carbon and energy source were readily isolated from the
colonized GAC. (iii) On average, the specific activity of toluene oxidation by these

communities was 9.9 :1: 0.4 pmol/109 cellsd, sufficient to account for total transformation

64
ofaddedsubstrate. TheinitiallagphasewaslikelyduetotolueneadsorptionbytheGAC

and the relatively cool temperature.

 

 

6 150
és 140
t; 130A
3.4 g
D 120..
as 35»
g 110%
32 3
> 100
'§
"31 90

o-‘r-‘rn so

0 10 20 30 40 50 60 70 80

Time (days)

Figure 1. Biofilm development in the toluene-fed reactor. The bed height reflects
overall biofilm thickness (open circles); dissolved oxygen uptake is a measurement of
biological activity (black circles). Gray arrows are the times at which the microscopic image
shown were taken. Black arrow shows the first time at which biomass was controlled by
application of shear forces.

Wum surfaceoftheGACparticles was
highly irregular (Figure 2). Close examination of the surface revealed the presence of
smooth open regions, large rough zones, and open cavities. The latter two provide larger
surface areas of sheltered regions and were the regions where microbial growth was initially
observed (Figure 3). After 12 days of continuous growth, the GAC surfaces were
completely covered with a layer of contiguous bacterial cells embedded within a polymeric
matrix (Figure 4). The bacteria present were mainly rod shaped (0.3 to 0.6 by 1 to 3 um)
and many ofthem were dividing. At this stage, the film depth reached 20 pm in some

65
areas. Thicker biofilms were particularly difficult to study by SEM because the extracellular

structures and bacterial cells collapsed during sample preparation.

 

Figure 2. SEM micrograph of sterile GAC particle showing different surface zones: S,
smooth; R, rough; C, cavity. Bar, 5 urn.

 

Figure 3. SEM micrograph of a 3—day—old microcolony mainly growing over rough
and within cavity areas of the GAC. Bar, 1 um.

 

Figure 4. SEM micrograph of a 12-day-old colonized granule from the toluene fed
reactor. Most of the surface has been colonized with actively dividing rod cells. However,
there is smooth zone (arrow) yet to be colonized. Bar, 5 run.

it t their rungs. --Forbiofilms

 

greater than 15 pm thick CSLM was preferred to study the development of the biofilm
superstructure on GAC particles. The following patterns are derived from observing 30 to
50 images from 10 to 15 granules examined at each of the sampled stages. Initially the
entire surface was colonized by a thin film (10 to 20 pm) which exhibited no structural
patterns (Figure 5a). However, after the biofilm had grown much thicker (63-days), a
mosaic pattern made up of separate lobes of cells was dominant (Figure 5b). At 77 days
this lobbed structure was not only maintained but even more pronounced showing that
further growth did not obliterate the channels separating the cell communities (Figure 5c).
These channels were typically 2 to 25 um wide. Close examination of some of these open

spaces revealed that they can extend from the top of the film to deep inner regions to form a

68
channel-like network (Figure 6). This overall biofilm structure was prominent after four to

six weeks ofoperation, when biofilm function had stabilized (Figure 1). Secondary
colonizers srehasyeastandthereaferpotomasweuasshearingofdeexcessbiomass
evennnflyappearsmdismrbdeinEgrityofdreagammmofilmsmenmedescfibed
above. TheWzoanfeedingonbactaialcdlsappearedtorecreatevoidspacesthat
resulted in conservation of channel-like structures (Figure 7).

Biofilms obtained from field GAC-FBRs also shared common organizational
structures with the laboratory biofilms. Bacterial density ranged from 109 to 1011 cells/g of
GAC in all biofilm communities studied. Mature biofilms exhibited patterns in which
internal contiguous cell-free spaces were commonly found in all four reactors studied (Table
1; Figure 8). For example, the biofilm superstructure of the Traverse City (TC) field-scale
FBR system exhibited contiguous void spaces opportunistically colonized by cocci shaped
cells (Figure 8a). In contrast, the benzene, toluene, ethylbenzene, and xylenes (BTEX) fed
system showed a 'coral-reef' like structure, where each finger of the biofilm formed of
similarly shaped cells has a particularly well defined boundary and distribution (Figure 8b).
Thedimensionsofthesefingersvariesfrom35t02mttmlongby20to40umthick. They
all appear to develop from a contiguous base film which completely covers the GAC
surface. Biofilm samples examined from two other treatment sites (PW and JEN) also
displayed unique spatial cell arrangements and contiguous internal void spaces (Figures 8c-
(1).

All thick biofrlms appear to have a contiguous (e.g., TOL) or semicontiguous (e.g.,
TC) base layer, which is recognized by the absence of channels or other void space. In the
30-day-old biofilm sample from the TC site, a discontinuous bacterial base film of 10- to
25-|.rm-deep was observed to cover the substratum from which cell-free spaces and channels
appear to develop (Figure 9). Biofilms from the other reactor systems also showed a base
film which ranged from 10- to 60-rrm-thick.

 

Figure 5. Series of confocal micrographs showing biofilm formation on GAC particles
by indigenous aquifer populations growing on toluene as sole source of carbon. Each
extended focus view was constructed by computer overlay of the optical sections in a 2
series. (A) l6—day-old; (B) 63-day-old; and (C) 77-day-old biofilm structure. The biofilm
thickness increased from 15 to 300 run between 16 and 77 days as determined by dark field
microscopy.

70

 

02'82. 8775”

Figure 6. Confocal micrographs of a 77-day-old biofilm showing the natrrre of the
channel structrues in the toluene fed reactor. (A) horizontal (x, y) optical thin section, 82.6
umfromthetopofthebiofilm (the horizontallineindicatestheopticalcuttingposition for
the sagittal section); and (B) sagittal (x, 2) view showing the vertical cell distribution and
the cell-free channel reaching up to the surface of the film.

71

 

Figure 7. Confocal micrograph of the void spaces created by protozoa (arrow) grazing
on bacteria in a 153-day-old biofilm community. At this point, the bioreactor had gone
through three shearing treatments for biomass removal and the biofilm structure displays
none of the distinctive vertical cell arrangement pattern observed in the early stage. Bar, 5
um.

 

Figure 8. Horizontal (x, y) optical sections of biofilm communities fiom different
field-scale FBRs. Note the spacing between cell aggregates and cell arrangements. (A) TC
reactor system, note also differentially packed cell aggregates and channel-like structure
partially colonized by cocci-shaped cells. (B) BTEX fed system showing a 'coral-reef
appearance of bacterial growth. (C) Biofilm structure of JEN system showing bacterial
aggregates separated by well-defined void space boundaries (arrow). (D) Micrograph from
the PW system showing bacterial aggregates and partitioning by cell-free spaces.

73

 

ZrXY-l

Figure 9. Horizontal (x, y) optical section (lower half of image) and sagittal (x, 2)
optical section (upper half of image) of a 30-day-old field reactor site (TC) showing a fully
hydrated biofilm with cell-free scpaces and a discontinuous base film (arrow).

 

Figure 10. Laser scanning dark field photomicrograph of a 77-day-old bacterial GAC
coated particle showing variation of the biofilm thickness within the same colonized surface.
Bar, 30 um.

75
The vertical dimensions of the different biofilm structures were examined by using

laser scanning dark field microscopy. We typically saw a high variation in biofilm thickness
even within the same colonized granule (Figure 10), e.g. 30 to 200 um.

 

desebiofilmcomumfieshaddmnhrsmwmrdfeammawemughtmevduaehowsimihr
theymightbein microbialcompositionbyFAME analysis. Therelativeabundanceof
FAMEsfrom tlediffaentGAC-FBRispresentedinTable 2. A totalofatleast 47 fatty
acids, including saturated, unsaturated, and methyl and hydroxyl substituted fatty acids
were identified among the samples. The greatest number and diversity of fatty acids were
found forthe communities in theFBRtreating [reduction waterbrinesfromanoil
producing field heavily contaminated with aromatic hydrocarbons (PW), and from the
aquifer contaminated with gasoline residues (TC) at Traverse City. The most abundant fatty
acids were 12:0, 14:0, 16:0, and 16:1 aflc at these two sites. By contrast, the community
with the least fatty acid diversity was the laboratory system fed only toluene (TOL), with
16:1 aflc, 18:1 m7c/m9c/m12t, and 16:0 as dominant fatty acids.

Principal component (PC) analysis of FAME profiles showed that the five GAC
communities were different (Figure 11). The first PC accounted for 41%, the second PC
for 25%, and the third PC for 14% of the variation in the data, for a cumulative total of
80%. PC scores obtained after analysis of the correlation matrix showed high scoring
(higher than 0.90) on fatty acids 13:0, a15:l, 15:0, i16:1, i16:0, 16:1 ar9c, 17:1 08c, 17:0,
and 18:1 m9c for PC]; 14:0, i15:0, and 16:1 (05c for PC2; and 10:0, i14:0, and 16:0 for
PC3.

76

Table 2. Relativeabmrdance(%oftotal)offauyacidsindiemicrobialcanmunifiesof
five different GAC-FBR systems.

 

 

 

W

Fatty acidl TDL BTEX TC PW JEN
1&0 0 6.30 0.95 1.12 1.88
i11:0 0 0 0 0 1.77
10.0 30H 3.28 7.40 3.40 0.39 3.27
12:0 3.91 43.65 14.20 12.46 25.91
i1 1:0 30H 0 0 0 0 1.68
13:0 0 0 0 0.45 0
12:0 20H 0.75 0 0 0 0
12:0 30H 0.91 1.07 0.14 3.32 0
i14:0 0 2.83 0 0.60 0
14:0 2.20 8.78 6.88 5.96 14.68
i15:1 0 0 1.20 0.42 0
a15:l 0 0 0 0.31 0
i15:0 0 1.93 1.03 1.80 3.44
a15.0 0 1.40 0.16 0.35 0.73
15:1 «Sc 0 0 0 0.41 0
15:1 m6c 0 0 0.23 0.57 0
15:0 0 0 0.34 3.06 0
14:0 20H 2.45 0 0 0 0
i14:0 30H 0 0 0.20 0 0
16:1 aflc alcohol 0 0 0.25 0 0
i16:l 0 0 0 1.40 0
i161) 0 0 0 4.89 0
16:1 m9c 0 0 0 6.33 0
16:1 (07¢ 44.21 18.43 39.87 8.80 11.52
*i 15:0 20H, 16:1 m7t 0 0 4.03 1.49 2.76
16:1 m5c 0 1.65 1.00 0.15 3.41
16:0 15.60 4.61 12.46 9.49 11.70
16.0 10 methyl 0 o 0 4.74 o
i15:0 30H 0 0 0.19 0 0
a17:0 0 0 0 0.38 0
17:1 ar8c 0 0 0.37 4.58 0

77

17:1t06c 0 0 0
17.0 crow 0 0 0.14
170 0 0 0.13
117.0 0 0 0.50
17.0 10 methyl 0 0 0.40
17:1m7c 0 0 0
16:0 3011 0 0 0
*18:3 o6c/o9c/o12c 0 0 0.10
1s:1o9c 0 0 1.14
*13:1 m7dor91/0012t 25.66 1.96 7.89
130 0 0 1.15
11703011 0 0 0.22
19:0CYCL0108c 0 0 0
*20:4 o6c/o9c/o12c/n15c 0 o 0.16
200 1.03 0 0

1.46
0
1.14
0.50
3.03
0
0.28
0.47
6.01
1.40
0.82
0.09
0.66
0.61
0

s—a
CSOGC

0.98
1.33

2.66
4.93
4.09

0.95
0
0

 

I Fatty acids are designated as tle number of carbon atoms, number of double bonds and
position relative to the aliphatic ((0) end of the molecule. The prefixes, i and a, refer to iso
and anteiso branching respectively; CY CLO indicates cyclopropane substitutions; OH
indicates hydroxyl substitutions; and other methyl branched fatty acids are designated upon

the methyl group position relative to the carboxylic acid.

2 The coefficient of variation for most fatty acids was generally below 10% for replicate

analyses.
* A group that MIDI-FAME does not reliably separate.

 

 

 

 

Figme 11. Principal components analysis showing variation in FAME patterns in five
different GAC-FBR systems: TOL, BTEX, TC, PW, and JEN.

DISCUSSION

Image analysis is potentially a powerful tool to understand microbial communities at
an organizational level, and CSLM and other new methods are beginning to allow this
potential to be realized. By using CSLM in this study of petroleum degrading films we
found that the classical model of random growth patterns leading to a uniform biofilm
structure with active cells at the surface was not supported. In contrast, large interaggregate
channels that extend well into the biofilm appear to be maintained for long periods of time.
Ifthismodel generallyprovestobetrue, which seems tobethecasein atleastafewother
types of biofilms examined (Lawrence et al., 1991; Wolfaardt et al., 1994), then
assumptions about uniformity of diffusion within biofilms and where active cells reside
should be reevaluated.

The high resolution and magnification power of SEM allowed detailed studies of the
early events of preemptive surface colonization. However, sample preparation (fixation,
dehydration, and embedding) introduces morphological artifacts for multilayer biofilm

79

analysis imposing intrinsic limitations on the film thicknesses which can be precisely
studied. CSLM however, allows the study of relatively thick film communities because of
itsabilitytoobtainclearoptical sectionsfreeofdefocusinterferencefromunderlyingor
overlying materials (Carlsson and Liljeborg, 1989; Carlsson et al., 1989; Caldwell et al.,
1992). The spatial relationship of cell arrangements, extracellular polymers, and spaces are,
therefore, unaffected.

TheGACofthelaboratoryscaleFBRsystemfedgrormdwateramendedwitb
toluene was rapidly colonized by the indigenous aquifer populations. SEM analysis showed
that colonization initially occumed in sheltered regions, e.g. on rough surfaces and in
cavities of the virgin GAC. This pattern of colonization is probably the result of the
adsorption properties of the GAC, fluid dynamics, and system geometry. Geesey et al.
(1979) have shown differential bacterial colonization in the crevices of suspended detrital
material in fast-moving rivers. Planar surfaces of suspended particles in fluidized bed
reactorsareexposedtogreatersm'face abrasion generatedby shearforcesandparticle
collisions. Eventually, however, continuous microbial growth resulted in contiguous
coverage of the entire surface.

OreethepardckwasfunycolmizedanddebiofilmgrewfiomZOttmtomum
inthickness, wenotedthatalargenumberofverdcalmrdhofizontalchannelspermeateddre
film. These channels extended fiom the surfaces into at least one halfthe depth ofthe film.
Thesmuctured,moreporousfilmwasbuiltonabasedfilmof10-to75-um. Toseeifthis
architectrual feature was the general case, especially outside oflaboratory controlled
systems, we examined four field-scale fluidized bed reactors treating petroleum
contaminated groundwaters. While the types of structures seen were more varied, the
common feature was that channels were permeating the biofilms in all reactors.

All GAC-FBR biofilms exhibited distinctive horizontal and vertical arrangements
with the following features: (i) complete GAC surface colonization and base films, (ii)
irregular film edges, (iii) variable film thickness within the same colonized granule, and (iv)

80
contiguous cell-free spaces. Open channels of variable size connected the outside film

surface and bulk solution with the deeper inner layers. This biofilm structure, which
increases the biological surface area per unit volume, may facilitate transport of substrates,
nuuimteandgasesdeeperintodemanixthanispossibleduetodiffusion alone. Active
bactaidcensindeepinnerhyersmayhelpwkeepthebasdfilmfirmlyamachedmde
substratum and, therefore, to be more resistant against uncontrolled sloughing of the biofilm
causedby shearstressortherapidstreamflowandparticlecollisions. Although tlriswas
not as extensively studied, late secondary colonization by yeast, fungi, and protozoa, and
mechanical biomasscontrol (tokeepthebiofilmthicknessfiomexceedingacertaindepth)
clearly affectedtheintegrityofthesesuuctmesinthebacterialbiofilm Whilewedonot
knowtheimpactofthese structuresonbiofilmperformance,wedoknowthatthereactor
performance, as measured by oxygen consumption remained constant.

It might be argued that this more porous structure is simply a result of random
growth, but there are important features that suggest that this is not likely to be the entire
explanation. First, the channels are maintained for many weeks, well beyond the time when
confluent growth should have obliterated the channels. Secondly, there are a diversity of
channel types which appears to be influenced by growth conditions and microbial
composition. Finally, the hydraulic stress of these systems will tend to select structures in
which the base film can be more active to maintain the GAC-film interaction thus, avoiding
destruction of itself. This suggests that some cell—cell communication occurs in these mixed
communities to form and maintain this architecture.

Community structure would be expected to be influenced by community
composition. We used FAME analysis of the community to determine whether the
communities of the one laboratory and four field reactors had different composition and
wletler some were more closely related. This method has been successfully applied by
several groups (Federle et al., 1990; Peltonen et al., 1992; Rajendran et al., 1992;
Frostegard et al., 1993; Haack et al., submitted) to examine microbial community

81
composition free of the bias inherent in culture dependent procedures (Wayne et al., 1987).

Usingprincipalcomponent analysiswedetermined thateachcommunityhadadistinct
composition. TheTOLaaboratory)andTCreactorswerethemo&similar. Thismaybe
dueedefactmmdeTOLmaaahaddesimpkstsubsuaefeedand,whflemeTCmacpr
hadamorecomplexfeed,itwasinoperationonlyonemonthandmaynothavecompleted
itssuccessiontoamorecomplexccmmunity. ReactorPWwastheonly systemthatwas
distinguishedbyPCl; thisreactorwastleonlyonewithahigh saltwaste stream
(approximately 7% w/v). In spite of the difi'erences in microbial composition among the
reactors, the channeled architecture was conserved.

Studies on anaerobic fixed-bed reactor community previously revealed integrative
networks of channel-like structures throughout the microbial granules which has been
interpreted as playing a role in transport of cooperan've substrates, nutrients and gases
(Robinson et al., 1984; Macleod et al., 1990; Wu et al., 1991). In the anaerobic food web
it is more clear why interspecific interactions between complementary trophic groups may
result in a structure that facilitates molecule transport. It is interesting however, that even
though aerobic oxidation of hydrocarbons might require few if any interspecific interactions,
that microbe-microbe communication and growth coordination seems to play a key role
during progressive biofilm establishment. Our CSLM results suggest conservation of
common architectrual strategies that appear to be biologically controlled perhaps to deal with
diffusion limiting processes and shear stress. This architectural arrangement of a biofilm
may represent a more optimal arrangement for influx of substrate, nutrients, and waste
transfer, and if me would have important implications on traditional concepural biofilm
models in which microbial growth and horizontal and vertical distribution was assumed to
be randomly determined. Further testing of this model and understanding of the forces and
principles which regulate biofilm development will help in efforts to model, control, and
exploit biofilm related processes of industrial, medical, and environmental importance.

APPENDIX A

APPENDIX A

BACTERIAL COMMUNITY FINGERPRINTING OF AMPLIFIED 168 AND 16-238
RIBOSOMAL DNA GENE SEQUENCES AND RESTRICTION ENDONUCLEASE
ANALYSIS (ARDRA)

The 168 and 238 rRNA genes have been utilized for phylogenetic analysis of both
prokaryotes and eucaryotes organisms (Woese, 1987). In addition to direct comparison of
the nucleic acid sequences, numerous groups have used the rapid method of polymerase
chain reaction (PCR) amplification of this gene (Muyzer et al., 1993) as well as the complete
rRNA locus (Jayarao et al., 1991; Vaneechoutte et al., 1992; Jensen et al., 1993) for a
simplemethod foridentificationofbacterial generaand species. In these latterprocedures,
the amplified DNA (rDNA) is subjected to restriction endonuclease analysis; this has been
termed ARDRA (Amplified Ribosomal DNA Restriction Analysis) (V aneechoutte et al..
1992). The resulting restriction fragment pattern is then used as a fingerprint for the
identification of bacterial genomes. This method is used on the principle that the restriction
sites on the RNA operon are conserved according to phylogenetic pattems.

Although ARDRA has been used for the characterization of bacterial isolates, in theory
this method can also be used for analyzing mixed bacterial populations. If the rDNA
fingerprint for individual bacteria in a mixture, e.g. community, were sufficiently different,
thenonecouldexaminetheamplifiedproductforaseriesofdistinctpatternresultingfrom
the different populations tlmt make up the community. Ribosomal DNA fingerprinting of
these communities could be used for a quick assessment of genotypic changes, possibly
over time or between different locations reflecting different environmental conditions. This
method involves the use of a pair of universal priming sequences for the PCR amplification
of either the 16S rRNA genetic loci or the intergenic regions of the 16S and 23S rRNA

82

83
genes. The latter regions exhibit a large degree of sequence, length and frequency variation

becausethespaceregionisnotwellconserved. IngeneraLARDRAusingthe l6SrRNA
geneswfllmsfltmasimplerpauanG-Sbandspagenomewhen4basesiw-specific
restriction endonucleases are used) than ARDRA using the 16-238 region. In this case
diffaent but related organisms may share common restriction fragments. Some fragments
maybeuniqueforaspecifworganismandwouldbediagmsdcforagivengroupof
microbes. ARDRAusingtbe l6—23SrDNAregionwillprovideamcrediverse substratefor
restriction analysis resulting in morecomplexbandpatternsandpotentially higherresolution
ofcommunitymembers. 'I‘hisversionofARDRAmaybeusefulinmoresimple
communities which are composed of closely related populations. In highly diverse
microbial communities, both of these methods may either yield too complex a pattern or too
little resolution of distinct bands. If this is the case, an alternative may be to use group
specificprimersforagiven bacterial grouportocarryouttheDNA amplificationunder
discriminatory conditions (e.g., increasing stringency conditions during the annealing step
of PCR). Further, resolution could also be obtained by hybridization with DNA probes
specific for the target groups. Finally, one may be able to dilute DNA samples from
communities so that only the more dominant member's genomes are present for ARDRA.
The level of resolution (band pattern complexity) is ultimate controlled by the frequency of
restriction sites and restriction endonucleases used.

This PCR based technique will allows for assessment ofmicrobial structure and
composition at the microbial scale. As little as picograms quantities of DNA (1-10 cells) can
be used in this method. In turn, changes or responses of the microbial system can be
monitored at the genotypic level without drastically disttn'bing the system by sampling.
Additionally this method is bias fiee of culture dependent methods, although it can be
applicable for characterization of axenic cultures as well as to identify non yet isolated
members.

34
EXPERIMENTAL APPROACH

A pairofhighly conserved flanking sequences areused forprimerbinding sites to
amplify the 16S ribosomal genes from microbial systems (Figure 1). The intragenic PCR
product (~1500 bp) is then used as substrate for restriction endonuclease digestion follow
by gel electrophoresis. In order to facilitate analysis and the resolution (ability to distinguish
twodifi‘uentpopmadms)ofdreassayafiathePCRamphficadm,maedmnmesepamte
restrictionenzymedigestionshouldbeused. SincecommunityamplifieerNAisan
indirect indicator of the diversity of sequences, sensitivity is maximized (ability to avoid
false negative) by concentrating the amplified rDNA or digested rDNA product before
electrophoresis. For the l6-23S rDNA intergenic region, amplification is carried out by
using a pair of opposite highly conserved 16S and 23S ribosomal primers (Figure 1)
followed by restriction enzyme digestion and gel electrophoresis.

 

168 238 SS
pA pHr
i : p2 SR 1
pH . I
spacerregrons '7

Figure 1. General schematic representation of the ribosomal RNA operon showing
approximate localization of priming binding sites for PCR amplification. The spacer region
varies in length (e.g., number of tRNA's), and sequences among the different operons
within a single organism (Brosius et al., 1981).

 

Primer Position Sequence (5' - 3') (Ulrike et al., 1989)
Designation (E.coli 168 rRNA)
pA 49-68 AGA G'I'I‘ 'I‘GA TCC TGG CI‘C AG
pH 1541-1518 AAG GAG GTG ATC CAGCCGCA

 

Primer Position Sequence (5' — 3')

Designation (Ecoli 16-238 rRNA)
pHr (reverse) 1518-1541 TGC GGC TGG ATC ACC TCC 'I'I‘
p238R01 1069- 1052 GGC TGC TI‘C TAA GCC AAC
PROCEDURES

WW - For mixed microbial systems. it is important “3
maximize cell lysis before amplification since lysis efficiency is unequal for different
bacterial types and physiological stages. Direct lysis using repetitive (5X) freeze (dry ice-
ethanol bath) and thaw (80°C water bath) cycles of the bacterial cells suspension can be
used for PCR. As an alternative, the use of community DNA (Maniatis et al., 1982) as
template for the reaction is suggested. This will ensure a more representative, reproducible
and controllable method to successfully carry out the amplification step.

1. Prepare Master solution containing:

ddH20 86.5 in
10X PCR Buffer 10.0 [11
lOOX dNTP's 1.0 ul
lOOX primer x 1.0 ul
100x primer y 1.0 ul
Taq DNA polymerase (5 U/ttl) 0.5 ul

2. Add 1 pl (approximately 100 ng) of bacterial or community DNA, or cell suspension
(approximately 10 to 104 cells). Total reaction volume is 100 rd.
3. Amplification with the following temperature profile:
Temperature Time
Denature 92°C 2 min, 10 sec

Melt 92°C
Anneal 48°C
Extend 72°C
Final extend 72°C
Soak 4°C

86

l min, 10 sec
30 sec (30-35 cycles)

2min10sec

6min lOsec

hold

4. After amplification, 5 to 10 pl is electrophoresed on 0.7% agarose gel made in TAE

buffer to examine the amplified products (Figure 2).

A

1 ,500

B
ABCDEF ABCDEF

 

Figure 2. Products of PCR amplification from microbial communities samples. (A)
The 16S rDNA amplification yielded a 1,500 bp fragment from axenic culture (lane B) as
well as bioreactor samples (lanes C to F). (B) Results of 16-238 rDNA intergenic
amplification of the same samples yielded various fragments of different sizes. Phage
lambda (1) DNA digested with PstI was used as marker (lanes A), indicated (in base pairs)

on the left.

87
Notes;
In. Maintain all solution on ice.
lb. Reagents shouldbeaddedinthatorder.
1c. For n number ofreactions, each solution volume has to be multiplied by n, and aliquot
it in 1: PCR vials. Always include a non-DNA control (e.g., water).
1d. Primersx&yarepA&pror16S rDNA amplificationorpHr&p23SR01 for 16-
238 rDNA amplification.
1e. Taq DNA polymerase from different sources may behave differently because of
different formulations, assay conditions, or unit definitions.
2. When cells are used as templates, they can be pre-treat (e.g., freeze & thaw) directly in
the PCR vials containing 86.5 ul of ddeO with the PCR buffer. The master solution of
nucleotides, primers and polymerase are then added before amplification.
3a. Optirml cycle temperatures may vary within different thermal cycler models. Given set
ofwmpaannesmddmeswaedaarmnedempiricaflyfmdeakinElmadramalcycler
model 9600.
3b. When cells are used as templates, melting steps should be performed at 94°C for 3
minutes. In addition, annealing should be done at 52°C.
3c. When DNA is used as template, 48°C provides low stringent annealing conditions, but
discriminatory enough to avoid non specific amplification.
Solutions;

- 10X PCR buffer

- 500 mM KCl

- 100 mM Tris-HCl, pH 8.3

- 15 mM MgC12
- 100x dNTP's solution

- 25 mM of each nucleotide (ATP, 11?, CTP, and GTP)

88
-Primers

— 100 W stock solution in TE buffer (10 mM Tris, 0.1 mM EDTA, pH 8)
- AmpliTaq DNA polymerase

- 5 Units/tr! (Perkin-Elmer Cetus, Norwalk, CT)
- 1x TAB electrophoresis buffer (pH 7.5)

- 40 mM Tris-HCl

- 5 mM sodium acetate

- 1 mM EDTA (disodium ethylenediaminetetraacetate)

 

- ems-u.» r ouvm‘ -- Digestion ofthe amplified
rDNA product with tandem tetrameric site-specific restriction endonucleases (e. g., HaelII,
Msel, Him, Sau3A, I-IpaII; Boehringer Mannheim, Indianapolis, IN) should yield a
restriction enzyme digestion pattern which can be discerned by gel electrophoresis. An
aliquot of the PCR product can be directly used for each digestion. Digest in the appropriate
restriction buffer (generally provided by all commercial companies in 10X concentration
when enzymes are purchased) at optimal temperature for 2—3 hours. As noted above with
community level amplification it is useful to concentrate the amplified product. This can be
accomplished by ethanol precipitation (1/2 volume 3 M sodium acetate with 2 volumes of
cold 100% ethanol at 20°C for 30 min to overnight) (Maniatis et al., 1982) before or after
digestion. The amount of DNA to be loaded per well is dependent of the limiting bands
which need to be resolved. The least amount of DNA that can be consistenly detected with
ethidium bromide staining is about 10 ng. DNA amounts greater than 100 ng will not be
resolved as a sharp clean band. The rDNA restriction fragments are then electrophoresed on
4% (w/v) NuSieve agarose gels (FMC, Rockland, ME) at 7 volts/cm containing 0.5x TAE,
and stained in 0.5 rig/ml ethidium bromide solution for 20 min (Figure 3). Finally, DNA
banding patterns database and numerical analysis can be performed with analytical
hardware/software package such as AMBIS MicroPM-l-M based system (Microbiology
Pattern Matching) (Hook et al., 1991).

89

'19 b5
§§lag$ xo‘b OiwaQfi 9rd \Q‘s‘i 9% $411

2.645 —

1 1,198 -

676 -

517 -

350 -
222

126 -

 

Figure 3. Photograph of ethidium bromide stained agarose gel. Lane MW: molecular
weight DNA markers, indicated (in base pairs) on the left. Lane 4’ DNA: PCR amplification
of water. All other lanes are HpaII restriction endonuclease digestion of PCR amplified l6S
rDNA from: G4, Pseudomonas cepacr'a ; PKOl, P. pikettii ; and Pan, P. putida ; Day 0,
13, 29, and 65 are bioreactor samples inoculated with the same strains of bacteria at Day 0
and challenged by aquifer populations after day 19. One interpretation of the results is that
Pan becomes the dominant population after Day 29. However, this erroneous
interpretation is the result of multiple groundwater populations that happen to have the same
restriction fragments of Pan when HpaII rs used. Differences will be more readily
detectable when a second or third restriction endonuclease rs put to use.

APPENDIX B

APPENDIX B

DIFFERENTIAL AMPLIFICATION OF rRNA GENES FOR THE DETECTION AND
IDENTIFICATION OF PSEUDOMONAS CEPACIA (G4) AND P. PICK EIT II (PKOl)

Assessmentofthefateofintroducedpopulationson natru'alecosystems has
ecological andpractical significance. Forexample, in the areaofbioremediation
technology, seva'al studies have suggested that by introducing and selecting organisms
with unique physiologic traits at contaminated sites or in bioreactors may stimulate
biodegradation (Chakrabarty, 1986). Pseudomonas cepacia (G4), and P. pickettr'i (PKOl)
are two well characterized benzene, toluene, and xylene degrading populations which can
also cometabolizc other toxic recalcitrant compounds such as trichloroethylene (TCE) fiom
the environment (Shields et al., 1989; Kukor and Olsen, 1990). Recently, we have
examined the use ofthese microorganisms in treatment of toluene contaminated
groundwater on biological activated carbon fluidized bed reactors and their successes
against indigenous aquifer populations.

Molecular-based methods can be faster, more sensitive and accurate than classical
plating techniques (Johnson, 1989). For example, molecular phylogeny of comparative
analysis of 16S rRNA genes (Gray et al., 1984) has been used to determine the species
composition of microbial communities, and several approaches have been used to obtain
rRNA sequences from natural systems free of bias inherent in cultivation procedures (Lane
et al., 1985; Schmidt et al., 1991; Weller et al., 1992). In one of these methods, the
polymerase chain reaction (PCR) is used for the synthesis and recovery of rRNA genes
from natural systems (Reysenbach et al., 1992; Muyzer et al., 1993) or axenic cultures
(Jayarao et al., 1991; Ward et al., 1992). We developed a differential amplification
procedure for the detection and rapid identification of strains G4 and PKOl that had been

released to a natural environment. This amplification was based on alignment of the 16S
90

91
rRNAsequencewithinaspecificvariableregionEigme l). Resultspresentedhere
illusu'atethespecificityandsensidvityoftheseamplificationprobes.

P. cepacia (strain G4) GGU CGG AAU CCC GAA GAG AU
P. pickeru'i (strain PKOl) GCC ACU AAC GAA GCA GAG AU
E. coli CCA CGG AAG UUU UCA GAG AU
P. cepacia GGU CGG AAU CCU GCU GAG AG
P. picketn'i GCC ACU AAC GAA GCA GAG AU
P. aeruginosa GCU GAG AAC UUU CCA GAG AU
P. testosteroui GGC AGG AAC UUA CCA GAG AU
P. mendocina GCU GAG AAC UUU CCA GAG AU

Figure 1. Ribosomal RNA sequence alignments (corresponding to positions in the
Escherichia coli 168 rRNA, 998 to 1017) showing target primer binding sites for strains
G4 and PKOl. Sequences are either results from Jizhong et al., unpublished or taken fiom
the ribosomal data base.

Table 1 showsthelistofbacterialsu'ainsusedinthisstudy. Axeniccultureswere
grown overnight on Luria broth (Difco, Detroit, MI) and harvested by centrifugation.
Genomic DNA (RNA-free) was isolated and purified as previously described (Maniatis et
al., 1982). Intragenic regions of the small-subunit rRNA was selectively amplified by
PCR using a forward primer complementary to a universally conserved region which
correspond to nucleotide positions 49 to 68 (5' - AGAGTTI‘GA TCCTGGCI‘CAG - 3‘,
noted as pA) (Ulrike et al., 1989) of Escherichia coli l6S rRNA and reverse species-
specific primers corresponding to the complement of positions 1017 to 998 (5' - GGUC
GGAAUCCCGAA GAGAU - 3', noted as pG4V3; and 5' - GCCACUAACXEAAGCAGA
GAU - 3', noted as pPKV3 specific for strains G4 and PKOl, respectively) (Jizhong et
al., unpublished). As a positive control, a universal conserved reverse primer binding site
(5' - AAGGAGGTGATCCAGCCGCA - 3', noted as pH; positions 1541 to 1518) (Ulrike
et al., 1989) was also used. All primers were synthesized with an Applied Biosystem
DNA synthesizer at the Macromolecular and Structure facilities at Michigan State

92
University. Template DNA (50 ng) was used in a 50-ttl reaction volume containing: 5 pl

of 10X reaction buffer (500 mM KCl, 100 mM Tris-HCI, pH 8.3, 15 mM MgC12 [Perkin-
ElmerCetus,Norwalk, C11), 100uMeachoftwoopposingPfim250uMofeach
deoxynucleoside trisphosphates (Perkin-Elmer Cetus, Norwalk, CT), and 1.5 U of Taq
DNA polymerase (Perkin-Elmer Cetus, Norwalk, CT). The amplification cycles used were
asfollows:denaturationat94°Cfor1min30sec;followedby 30cyclesofdenaturationat
92°Cfor1min30sec,annealingat52°Cfor30sec,andextensionat72°Cfor1min30
sec; anextensionat72°Cfor6min;andterminationat4°C. Amplificationproductswere
resolved by agarose gel electrophoresis in a 0.8% (w/V) agarose gel containing 1X TAB
(40 mM Tris-HCl, pH 7.5, 5 mM sodium acetate, 1 mM EDTA [disodium ethylene-
diaminetetraacetateD, and stained in 0.5 rig/ml ethidium bromide solution, and were
photographed in UV light with Polaroid Type 55 film.

Table 1. Bacterial su'ains studied and reaction of specie-specific PCR amplification.

 

 

Reaction with
Bacterial strain or isolate Descriptionl Lab. source or ref. $42333 pPKV3
PGseudbmonas cepacr'a beta Shields et al., 1989 + -
4
P. picketrir' PKOl beta Kukor et al., 1990 - +
P. mendocina KR gamrrna M. DeFlaun; Yen et al., - -
1991
P. panda Pan gamma Harayama et al., 1989 - -
Azoarcus sp. Tol 4 gamrrna Fries et al., submitted - -
P. fluorescens gamrm Bill Holben, Mich. St. - -
Univ.
P. aeruginosa gamma Bill Holben - -
Escherichia coli W3110 gamma Bill Holben - -
Enrerobacter sp. gamma Bill Holben - -
Agrobacterium alpha Bill Holben - -
radiobacr‘er
Ps. sp. gf 261 beta Lab. collection, Chaper III - -
Strain gf 281 alpha Lab. collection, Chaper III - -

 

Subdivision of the Proteobacteria.

93
In standard reactions in which PCR was canied out with universal eubacterial

primers, an approximately 1,500-bp amplified fragment was obtained for all the tested
strains (Figure 2, lanes 3—8). However, differential amplification with the pPKV3 specific
primer only showed an amplified product of the predicted size (approximately 1,000-bp)
for strain PKOI, while non of the other species showed product (Table 1; Figure 2, lanes
10-15). The addition of non-specific DNA has no major effects on the reaction, even when
the target DNA was 1% of the total (Figure 2, lanes 16-18). Similar results were obtained
with the pG4V3 primer for differential amplification of strain G4 (Table 1).

ABCDEFGHIJKLMNOPQRS

1.500 9

 

Figure 2. Differential amplification of bacterial 16S rRNA genes. Lanes A and S, i.
PstI standards (in base pairs); lanes B and I, PCR without template; lanes C to H, and J to
O are PCR amplification of Escherichia coli, strain G4, strain PKOl, strain Pan, P.
fluorescens, and P. aeruginosa using universal primers pA and pH, and using the universal
primer pA combined to the specie-specific oligonucleotide primer pPKV3 respectively; and
lanes P to R are PCR amplification using pA and pPKV3 primers of 1:1, 1:10, and 1:100
PKOl template DNA in the presence of non-specific DNA (Pan).

We also used this procedure for examining a biofilm community pre-seeded with P.
cepacr'a (G4) (see Chapter II). Figure 3 shows differential amplification of the 16S rDNA

94
intragenic region of strain G4. According to the test, strain G4 or closely related

populations were below detection limits in the marine biofilm community suggesting that
G4 probably was out competed during earlier events of community establishment. Our
observations were further supported by traditional microbiological methods and DNA-DNA
hybridization techniques with a catabolic gene probe derived from strain G4 (data not

shown).

 

Figure 3. Differential PCR amplification of P. cepacia (G4) from a toluene treating
bioreactor community inoculated with the same strain of bacteria using primers pA and
pG4V3. Lanes A and F, A Pst I standards (in base pairs); lane B, PCR without template;
lanes C, strain G4; lane D, bioreactor community, day 63; and lane E, 1:1 of the same

community DNA and G4

In conclusion, we developed a rapid PCR-based method as a useful presumptive
test for the detection of toluene degrading strains G4 and PKOl. Lack of amplification
would then imply that the template concentration of the target strain or phylogenetically
closed relatives were below the limit of sensitivity or that the template was absent, as was

the case for our DNA-negative controls. If positive results are obtained by PCR, then other

95
viability testing using cell culture techniques or DNA-DNA hybridization procedures are

recommended.

LIST OF REFERENCES

LIST OF REFERENCES

Adachi, Y., H. Watanabe, K. Tanabe, N. Doke, S. Nishimura, and T. Tsuge.
1993. Nuclear ribosomal DNA as a [robe for genetic variability in the jagnese pear
pathotype of Alternaria alternata. Appl. Environ. Microbiol. 59:3197-3 5.

Aelion, C. M., and P. M. Bradley. 1991. Aerobic biodegradation potential of
subsurface microorganisms from a jet fuel-contaminated aquifer. Appl. Environ.
Microbiol. 57:57-63.

Alvarez, P. J. J., and T. M. Vogel. 1991. Substrate interactions of benzene,
toluene, and para-xylene during microbial degradation by pure cultures and mixed culture
aquifer slurries. Appl. Environ. Microbiol. 57:2981-2985.

Amman, R. 1., L. Krumholz, and D. A. Stahl. 1990. Flourescent-oligonucleotide
probing of whole cells for determinative, phylogenetic, and environmental studies in
microbiology. J. Bacteriol. 172:762-770.

Amman, R. 1., J. Stromley, R. Devereux, R. Key, and D. A. Stahl. 1992.
Molecular and microscopic identification of sulfate-reducing bacteria in multispecies
biofilms. Appl. Environ. Microbiol. 58:614-623.

Andrews, J. H., L. L. Kinkel, F. M. Berbee, and E. V. Nordheim. 1987. Fungi,
leaves, and the theory of island biogeography. Microb. Ecol. 14:277-290.

Angles, M. L., K. C. Marshall, and A. E. Goodman. 1993. Plasmid transfer
baweenmafinebactaiaintheaquwusphaseandbiofilmsmmactarmcmcosms.AppL
Environ. Microbiol. 59:843-850.

Atlas, R. 1981. Microbial degradation of petroleum hydrocarbons: an
environmental perspective. Microbiol. Rev. 45:180-209.

Avery, O. T., C. M. MacIeod, and M. MaCarty. 1944. Studies on the chemical
.1719ature3 of tlge substance inducing transformation of pneumococcal types. J. Exp. Med.
:1 7-15 .

97

Awong, J, G. Bitton, and G. R. Chaudhry.1990.Microcosms for assesing
survival of genetically engineered microorganisms in aquatic environments. Appl. Environ.
Microbiol. 56:977-983.

Barkay, T., C. Liebert, and M. Gillman. 1993. Conjugal gene transfer to aquatic
gaitaiadetectedbydregenerafionofanewphenotypeAppL Environ. Microbiol. 59:807-

Beachey, E. H. 1988. Bacterial adherence, p. 1-4. In Molecular Mechanisms of
Microbial Adhesion, Switalski Hook and Beachey, Editors. Springer-Veriag, New York.

Bickle, T. A., and D. H. Kriiger. 1993. Biology of DNA restriction. Microbiol.
Rev. 57:434-450.

Binnerup, S. J., K. Jensen, N. P. Revsbech, M. H. Jensen, and J. Sorensen.
1992. Denitrification, dissirnilatory reduction of nitrate to ammonium, and nitrification rn a

bioturbated estuarine sediment as measured with 15"N and microsensor techniques. Appl.
Environ. Microbiol. 58:303-313.

Blenkinsopp, S. A., A. E. Khoury, and J. W. Costerton. 1992. Electrical
enhancement of biocide efficacy against Pseudomonas aeruginosa biofilms. Appl. Environ.
Microbiol. 58:3770-3773.

BOttger, E. C. 1989. Rapid determination of bacterial ribosomal RNA sequence by
direct sequencing of enzymatically amplified DNA. FEMS Microbiol. Lett. 65: 171-176.

Breznak, J. A., K. E. Cooksey, F. E. W. Eckhardt, Z. Filip, M. Fletcher, R. J.
Gibbons, H. , W.A.H. Guide T. Hattori, H. G. Hoppe, A. G. Matthysse, D. C. Savage, ,
and a.M. Shilo. 1984. Activity on Surfaces, Group Report, p. 203-222. In Microbial
Adhesion and Aggregation. Springer-Veflag, New York.

Britschgi, T. B., and S. J. Giovannoni. 1991. Phylogenetic analysis of a natural
marine bacterioplankton population by rRNA gene cloning and sequencing. Appl. Environ.
Microbiol. 57:1707-1713.

Brosius, J., T. J. Dull, D. D. Sleeter, and H. F. Noller. 1981. Gene organization
at; prarInarZy structure of a ribosomal RNA operon from Escherichia coli. J. Mol. Biol.
:1 -1 7.

Brown, M. R. W., D. G. Allison, and P. Gilbert. 1988. Resistance of bacterial
313031311118 to antibiotics: a growth-rate related effect? J. Antimicrobiol. Chemother. 22:777-

98

Brown,M.L.,andJ.J.Gauthier.1993.Celldensityand wthphaseasfactors
intheresistanceofabiofilmofPseudanonasaenrginosa (ATCC 853)toiodine.Appl.
Environ. Microbiol. 59:2320-2322.

Caldwell, D. E., D. R. Korber, and J. R. Lawrence. 1992. Confocal laser
' y and digital image analysis in microbial ecology. Adv. Microbiol. Ecol. 12:1-
67.

Caldwell, D. E., D. R. Korber, and J. R. Lawrence. 1992. Confocal laser
microscopy and digital image analysis in microbial ecology. Adv. Microb. Ecol. 12:1-67.

Carlsson, K., P. Wallen, and L. Brodin. 1989. Three-dimensional imaging of
neurons by confocal fluorescence microscopy. J. Microscopy 155:15-26.

Carlsson, K., and A. Liljeborg. 1989. A confocal laser microscope scanner for
digital recording of optical serial sections. J. Microscopy 153:171-180.

Chakrabarty, A. M. 1986. Genetic engineering and problems of environmental
pollution. Biotechnology 8:515-530.

Characklis, W. G., and P. A. Wilderer (eds.). 1989. Structure and Function of
Biofilms. J. Wiley and Sons, New York, p. 371.

Characklis, W.G., and K. C. Marshall (eds.). 1989. Biofilms. John Wiley and
Sons, New York, p. 796.

Christensen, F. R., G. Holm Kristensen, and J. la Cour Jansen. 1989. Biofilm
2structures--anrmpor'tan' tandneglectedparameterinwastewateru'eaunentWatSci.Tech.
1:805- 14.

Costerton, J. W. 1983. Direct ultrastructure examination of adherent bacterial
populations in natural and pathogenic ecosystems, p. 115-123. In Current Perspective in
Microbial Ecology, Klug and Reddy, Editors. Am. Soc. Microbiol., Washington, D. C.

Costerton, J. W., K.-J. Cheng, G. G. Geesey, T. I. Ladd, J. C. Nickel, M.
Dasgupta, and T. J. Mamie. 1987. Bacterial biofilms in nature and disease. Ann. Rev.
Microbiol. 41:435-464.

Counotte, G. H. M., and R. A. Prins. 1979. Calculation of Km and Vmax from
substrate concentration versus time plot. Appl. Environ. Microbiol. 38:758-760.

99

Criddle, C. 8., L. A. Alvarez, and P. L. McCarty. 1991. Microbial processes in
porousmediap. 639-691.!n'1‘ransportProcessesinPorousMedia,Bearand
Corapcioglu, Editors. Kluwer Academic Publishers, The Netherlands.

Dalsgaard, T., and N. P. Revsbech. 1992. Regulating factors of denitrification in
trickling filter biofilms as measured with the oxygen/nitrous oxide microsensor. FEMS
Microbrol. Ecol. 101:151-164.

Davis, J. W., and C. L. Carpenter. 1990. Aerobic biodegradation of vinyl chloride
in groundwater samples. Appl. Environ. Microbiol. 56:3878-3880.

De Beer, D., J. C. van der Heuvel, and 8. P. P. Ottenfraf. 1993. Microelectrode
meastuements of the activity distribution in nitrifying ° aggregates. Appl. Environ.
Microbiol. 59:573-579.

De Bruijn, F. J. 1992. Use of repetitive (repetitive extragenic palindromic and
enterobacterial repetitive intrageneric consensus) sequences and the polymerase chain
reaction to fingerprint the genomes of Rhizobium meliloti isolates and other soil bacteria.
Appl. Environ. Microbiol. 58:2180-2187.

Dean, B. J. 1978. Genetic toxicology of benzene, toluene, xylenes and phenols.
Mutation Res. 147 :75-97 .

Dean, J. B. 1985. Recent findings on the genetic toxicology of benzene, toluene,
xylenes and phenols. Mutation Res. 154:153-181.

DeLong, E. F., G. 8. Wickham, and N. R. Pace. 1989. Ph logenetic stains:
ribosomal RNA-based probes for the identification of single cells. 'ence 243: 1360-
1363.

Devereaux, J., P. Haeberli, and O. Smithies. 1984. A comprehensive set of
sequence analysis programs for the VAX. Nucleic Acid Res. 12:387-395.

Dunster, H. J. 1992. Risk assessment, p. 93-98. In The Release of Genetically
Modified Microorganisms-Regem 2, Duncan, Stewart-Tull, and Sussman, Editors.
Plenum Press, New York.

Dworkin, M. 1991. Introduction, p. 1-6. In Microbial Cell-Cell Interactions,
Dworkin, Editor. Am. Soc. Microbiol., Washington, D. C.

Dykhuizen, D. E., and D. L. Hartl. 1983. Selection in Chemostats. Microbiol.
Rev. 47:150-168.

100

Edwards, E. A., L. E. Wills, M. Reinhard, and D. Grbic-Galic.1992. Anaerobic
degradation oftoluene and xylene by aquifer microorganisms under sulfate-reducing
conditions. Appl. Environ. Microbiol. 58:794-800.

Ehrhardt, H. M., and I-L-J. Rehm. 1989. Semicontinuous degradation of phenol
% I3’seudonro3 nas punda’ P8 adsorbed on activated carbon. Appl. Microbiol. Biotechnol.
: 12- 17.

Emerson, D., R. M. Worden, and J. A. Breznak. 1994. A diffusion gradient
chamber fa' studying microbial behavior and separating microorganism. Appl. Environ.
Microbiol. 60:1269-1278.

Evans, P. J., D. T. Mang, K. S. Kim, and L. Y. Young. 1991. Anaerobic
degradation of toluene by a denitrifying bacterium. Appl. Environ. Microbiol. 57:1139-
1 145.

Evans, G. M. 1994. Price information and the hazardous waste remediation
industry, p. 84—87. In Twentieth Annual RREL Research Symposium. Cincinnati, OH,

Federle, T. W., R. M. Ventullo, and D. C. White. 1990. Spatial distribution of
microbial biomass, activity, community structure, and the biodegradation of linear
alkylbenzene sulfonate (LAS) and linear alcohol ethoxylate (LAB) in the subsurface.
Microbiol. Ecol. 20:297-313.

Felsenstein, J. 1989. PHYLIP -- Phylogeny inference package (Version 3.2).
Cladistics 5:164-166.

Fletcher, M. 1979. The attachment of bacteria to surfaces in aquatic environments,
p. 87-108. In Adhesion of Microorganisms to Surfaces, Ellwood, Melling and Rutter,
Editors. Academic Press, London.

Fletcher, M., and 8. McEldowney. 1983. Microbial attachmrent to nonbiological
srmfaces, p. 124-129. In Current Perspective in Microbial Ecology, Klug and Reddy,
Editors. Am. Soc. Microbiol., Washington, D. C.

Folsom, B. R., P. J. Chapman, and P. H. Pritchard. 1990. Phenol and
trichloroethylene degradation by Pseudomonas cepacr'a G4: kinetics and interactions
between substrates. Appl. Environ. Microbiol. 56:1279-1285.

Franklin, F. C. H., M. Bagdasarian, M. M. Bagdasarian, and K. N. Trmmis.
1981. Molecular and function analysis of the TOL plasmid pWWO fiom Pseudonronas
putida and cloning of genes for the entire regulated aromatic ring meta cleavage pathway.
Proc. Natl. Acad. Sci., USA 78:7458-7462.

101

Fredrickson, A. G., and G. Stephanopoulos. 1981. Microbial competition. Science
213:972-979.

Fredrickson, J. K., F. J. Brockman, D. J. Workman, S. W. Li, and T. 0.
Stevens. 1991. Isolation and characterization of a subsurface bacterium capable of growth
on tolggesnga naphthalene, and other aromatic compounds. Appl. Environ. Microbiol.

57:7 .

Freter, R. 1983. Factors affecting conjugal plasmid transfer in natural bacterial
communities, p. 105-114. In Current Perspective rn Microbial Ecology, Klug and Reddy,
Editors. Am. Soc. Microbiol., Washington, D. C.

Fries, M., and J. M. Tiedje. Personal communication.

Fries, M. R., J. Zhou, J. Chee-Sanford, and J. M. Tiedje. Submitted. Isolation,
characterization and distribution of denitrifying toluene degraders from a variety of habitats.
Appl. Environ. Microbiol.

Frostegard, A., A. Tunlid, and E. Bath. 1993. Phospholipid fatty acid
composition, biomass, and activity of microbial communities from two soil types
experimentally exposed to difi‘erent heavy metals. Appl. Environ. Microbiol. 59:3605-
3617.

Frostegard, A., E. Baath, and A. Tunlid. 1993. Shifts in the structure of soil
microbial communities in limed forests as revealed by phospholipid fatty acid analysis. Soil
Biol. Biochem. 25:723-730.

Geesey, G. G., and J. W. Costerton. 1979. Microbiology of a northern river:
bacterial distribution and relationship to suspended sediment and organic carbon. Can. J.
Microbiol. 25: 1058-1062.

Gorden, R. W. R. J. Beyers, E. P. Odurn, and R. G. Eagon. 1969. Studies ofa
simple laboratory microecosystem: bacterial activities in a heterotrophic succession.
Ecology 50: 86-100.

Gray, M. W., D. Sakoff, and R. J. Cedergren. 1984. On the evolutionary descent
of organisms and organelles: a global phylogeny based on a highly conserved structural
core in small subunit ribosomal RNA. Nucleic Acid Res. 12:5837-5852.

Greer, L. E., J. A. Robinson, and D. R. Shelton. 1992. Kinetic comparison of
seven strains of 2,4-Dichlorophenoxyacetic acid-degrading bacteria. Appl. Environ.
Microbiol. 58:1027-1030.

102

Gutell, R. R., N. Larsen, and C. R. Woese. 1994. Lessons from an evolving
1518111321668 and 23S rRNA structures from a comparative perspective. Microbiol. Rev.

Haack, S. K., H. Garchow, D. A. Odelson, L. J. Forney, and M. J. Klug.
Submitted. Microbial community analysis: accuracy, reproducibility and interpretation of
fatty 80:1 methyl ester profiles from model bacterial communities. Appl. Environ.

Harayama,S., M.Rekik,M.Wubbolts,K.Rose,R.A. Leppik, andK.M.
Timmis.1989.Characterizationoffivegenesintheupper-pathwayoperonoftolplasmid
[lrWWOfiosrglgseudmwnasputidaandidendficafionofdregeneproducmL Bacteriol.

71 '5048- 5

Harkes, G., J. Dankert, and J. Feijen. 1992. Bacterial migration along solid
surfaces. Appl. Environ. Microbiol. 58:1500-1505.

I-Iealey, F. P. 1980. Slope of the Monod equation as an indicator of advantage in
nutrient competition. Microb. Ecol. 5:281-286.

Hickey, R. F., D. Wagner, and G. Mazewski. 1991. Treating contaminated
groundwater using a fluidized-bed reactor. Remediation/Autumn :447-459.

Holm, P. E., P. H. Nielsen, H.-J. Albrechtsen, and T. H. Christensen. 1992.
hnpmmnceofunatmchedbactaiaandbaaefiaanaclwdmsedimnntindewrmdning
potentials for degradation of xenobiotic organic contaminants in an aerobic aquifer. Appl.
Environ. Microbiol. 58:3020-3026.

Hook, L. A., D. A. Odelson, A. H. Bogardt, B. B. Hemmingsen, D. P. Labeda,
and M. T. MacDonell. 1991. Numerical analysis of DNA restriction fragment length
polymorphisms and whole-cell protein banding patterns: a means of bacterial identification
at the species and subspecies level. Newsletter (U SFCC) 21:1-10.

Hutchins, S. R. 1991. Biodegradation of monoaromatic hydrocarbons by aquifer
microorganisms using oxygen, nitrate, or nitrous oxide as the terminal electron acceptor.
Appl. Environ. Microbiol. 57: 2403-2407.

Ippen-Ihler, K. 1989. Bacterial conjugation, p. 32-72. In Gene transfer in the
environment,, Levy and Miller, Editors. New York, McGraw-I-Iill Publishing.

Jayarao, B.M., J. J. E. Dore, G. A. Baumbach, K. R. Matthews, and S. P.
Oliver. 1991 . Differentiation of Streptococcus uberr's from Streptococcus pararrberr's by

103

polymerase chain reaction and restriction fragments length polymorphism analysis of 168
ribom DNA. J. Clinical Microbiol. 29:2774-2778.

Jensen, M. A., J. A. Webster, and N. Straus. 1993. Rapid identification of bacteria
onthebasisofpolymerasechainreaction-amplifiedribosomalDNA spacer
polymorphisms. Appl. Environ. Microbiol. 59:945-952.

Jing, 8., X. Zhao, T. C. Voice, and R. F. Hickey. 1992. The importance of
absorptionin biological activatedcarbonfluidizedbedreactorsforthetreatmentof
water contaminated with volatile aromatic hydrocarbons. Master Thesis, Michigan
tate University:

Jizhong, Z., R. Weller, and J. M. Tiedje. Unpublished.
Jizhong J .-K., and J. M. Tiedje. Personal communication.

Johnson, J. L. 1989. Nucleic acids in bacterial classification, p. 2306-2309. In
Bergy's manual of systematic bacteriology, 8. T. Williams, M. E. Sharpe, and J. G. Holt,
Editors. The Williams & Wilkins Co., Baltimore.

Kamimura, K., H. Fuse, 0. Takimura, and Y. Yamaoka. 1993. Effects of growth
pressure and temperature on fatty acid composition of a barotolerant deep-sea bacterium.
Appl. Environ. Microbiol. 59:924—926.

Kaphammer, B. J., J. J. Kukor, and R. H. Olsen. 1991. Characterization of a
novel toluene degradative pathway from Pseudomonas pickettii PKOl, p. 571-572. In
Biodeterioration and Biodegradation, H.W. Rossmore, Editor. Elsevier Applied Science,
London.

Keohavong, P., and W. G. Thilly. 1989. Fidelity of DNA polymerase in DNA
amplification. Proc. Natl. Acad. Sci. (USA) 86:9253-9257.

Kinkel, L. L., J. H. Andrews, and E. V. Nordheim. 1989. Fungal immigration
dynamics and community development on apple leaves. Microb. Ecol. 18:45-58.

Kobayashi, H., and B. E. Rittrnann. 1982. Microbial removal of hazardous organic
compounds. Environ. Sci. Technol. 16:170A-l83A.

Kokjohn, T. A. 1989. Transduction: mechanim and potential for gene transfer in
the environment, p. 73-97. In Gene transfer in the environment, Levy and Miller, Editors.
New York, McGraw-Hill Publishing.

104

Kolenbrander,P. E., R. N. Andersen,andL. V. H. Moore. 1990.1ntrageneric
cmggregadmamongsuaimofhumanaalbacmimpowndalmkmprimarycobniudon
of the tooth surface. Appl. Environ. Microbiol. 56: 3890-3894.

Konopka,A., andR. Turco. 1991. Biodegradationoforganrc conggrsrndsin
vadoaezoneandaquifer sediments. Appl. Environ. Microbiol. 57:2260-

Kukor, J. J., and R. Olsen. 1989. Diversity of toluene degration following long
term exposure to BTEX in situ., In Biotechnology Biodegradation, p. 405-421.

Kukor, J. J., and R. H. Olsen. 1990. Molecular cloning, characterization, and
regulation of a Pseudomonas pickettr’i PKOl gene encoding phenol hydroxylase and
expression of the gene in Pseudonronas aeruginosa PAOlc. J. Bacteriol. 172:4624—4630.

Ktihl, M., and B. B. Jorgensen. 1992. Microsensor measurements of sulfate
mducfionandsulfideoxidadmhwmpadmiaobialcmanfiesofaerobicbiofilms
Appl. Environ. Microbiol. 58:1164-1174.

Ladd, T. 1., and J. W. Costerton. 1990. Methods for studying biofilm bacteria.
Meth. Microbiol. 22:285-310.

Lane, D. J., B. Pace, G. J. Olsen, D. A. Stahl. M. L. Sogin, and N. R. Pace.
1985. Rapid determination of 16S ribosomal RNA sequences for phylogenetic analyses.
Proc. Natl. Acad. Sci. USA 82:6955-6959.

Lappin-Scott, H. M., and J. W. Costerton. 1989. Bacterial biofilms and surface
fouling. Biofouling 1:323-342.

Larsen, N., G. J. Olsen, B. L. Maidak, M. J. McCaughey,R. Overbeck,T. J.
Macke,T. L. Marsh, andC. R.Woese. 1993. The ribosomal database projects. Nucleic
Acid Res. 21 (Supplement): 3021- 3023.

Lawrence, J. R., D. R. Korber, B. D. Hoyle, J. W. Costerton, and D. E.
Caldwell. 1991. Optical sectioning of microbial biofilms. J. Bacteriol. 173:6558-6567.

Leahy, J. G., and R. R. Colwell. 1990. Microbial degradation of hydrocarbons in
the environment. Microbiol. Rev. 54:305-315.

Lehrbach, P. R., and K. N. Timmis. 1983. Genetic analysis and manipulation of
catabolic pathways in Pseudomonas. Biochem. Soc. Symp. 48:191-219.

105

Levin,B. R. 1972. Coexistanceoftwo asexualstrainsonasingleresource.
Science 175: 1272-1274.

Lewin, R. 1983. Santa Rosalia was a goat. Science 221:636-639.

MacArthur, R. H., and R. Levins. 1967. The limiting similarity, convergence, and
divergence of coexisting species. Am. Nat. 101:377-385.

MacLeod, F. A., S. R. Giot, and J. W. Costerton. 1990. Layered structure of
bacterial aggregates produced in an upflow anaerobic sludge bed and filter reactor. Appl.
Environ. Microbiol. 56:1598-1607.

Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a
laboratory manual. New York, Cold Spring Harbor Laboratory.

Manz, W., R. Amman, W. Ludwig, M. Wagner, and K.-H. Schleifer. 1992.
Phylogenetic oligodeoxynucleotide probes for the major subclasses of Proteobacteria:
problems and solutions. System. Appl. Microbiol. 15:593-600.

Marcelino, N., and D. R. Bensosn. 1992. Scanning electron and light microscopic
study of microbial succession on Bethlehem 8t. Nectaire cheese. Appl. Environ.
Microbiol. 58:3448-3454.

Marshall, K. C. 1986. Microscopic methods for the study of bacterial behavior at
inert surfaces. J. Microbiol. Meth. 4:217-227.

Marshall, K. C. 1992. Biofilms: an overview of bacterial adhesion, activity, and
control at surfaces. ASM News 58:202-207.

Massol-Deya, A., R. F. Hickey, and J. M. Tiedje. Unpublished.

Massol-Deya, A., D. A. Odelson, R. F. Hickey, and J. M. Tiedje. In press.
Bacterial community fingerprinting of amplified 168 and 16-238 ribosomal DNA gene
sequences and restriction endonuclease analysis (ARDRA). In Molecular Microbial
Ecology Manual. Akkermans, Elsas and de Bruijn, Editors. Kluwer Academic
Publishers, The Netherlands.

McClure, N. C., J. C. Fry, and A. J. Weightrnan. 1991. Survival and catabolic
activity of natural and genetically engineered bacteria in a laboratory-scale activated-sludge
unit. Appl. Environ. Microbiol. 57:366-373.

106

McCormick, P. V., E. P. Smith, and J. Cairns. 1991. The relative importance of
gaguslgtigg versus community processes in microbial pimary succession. Hydrobiologia

McEldowney, 8., and M. Fletcher. 1987. Adhesion of bacteria from mixed cell
suspension to soil surfaces. Arch. Microbiol. 148:57-62.

McNaughton, S. J. 1978. Stability and diversity of ecological communities. Nature
(London) 274:251-253.

McSweeney, C. 8., R. I. Mackei, A. A. Odenyo, and D. A. Stahl. 1993.
Developmentof an oligonucleotide probe targeting 168 rRNA and its application for
detection and quantitation of the rumial bacterium Synergistes jonesii in a mixed-population
chemostat. Appl. Environ. Microbiol. 59:1607-1612.

Mills, A. L., and R. Maubery.1981. Effect ofmineral composition on bacterial
attachment to submerged rock surfaces. Microb. Ecol. 7:315-322.

Moody, S. F., and B. M. Tyler. 1990. Use of nuclear DNA restrictitni fragment

length polymorphisms to analyze the diversity of the Aspergillruflavus group: A. flavus,
A. parasiticus, and A. nomius. Appl. Environ. Microbiol. 56:2453-2461.

Moore, 8. F., and R. I. Dwyer. 1974. Effects of oil on marine organisms: a critical
assessment of published data. Water Res. 8:819-828.

Moyer, C. L., F. C. Dobbs, and D. M. Karl. 1994. Estimation of diversity and
community structure tluough restriction fragment length polymorphism distribution
analysis of bacterial 16S rRNA genes from a microbial mat at an active, hydrothermal vent
system, Loihi Seamount, Hawaii. Appl. Environ. Microbiol. 60:871-879.

Mullis, K. B., and F. A. Faloona. 1987. Specific synthesis of DNA in vitro via a
polymerase-catalyzed chain reaction. Meth. Enzymol. 155:335-350.

Muyzer, G., E. C. de Waal, and A. G. Utterlinden. 1993. Profiling of complex
microbial populations by denaturing gradient gel electrophoresis analysis of polymerase
clatgn reaction-amplified genes coding for 16S rRNA. Appl. Environ. Microbiol. 59:695-
7 .

Ney, U., A. J. L. Macario, E. Conway de Macario, A. Aivasidis, 8. M.
Schobertlr, and H. Sahm. 1990. Quantitative microbiological analysis of bacterial
community shifts in a high-rate anaerobic bioreactor treating sulfite evaporator condensate.
Appl. Environ. Microbiol. 56:2389-2398.

107

Nordstr'om, K. M., and S. V. Laakso. 1992. Effects of growth temperature on
fatty acid composition of ten Thermrrs strains. Appl. Environ. Microbiol. 58:1656-1660.

Nordstrbm, K. M. 1993.Effectoftemperatureonfattyacidconmositionofawhite
Thennus strain. Appl. Environ. Microbiol. 59:1975- 1976.

Olsen, R. H., J. J. Kukor, and M. D. Mikesell. 1989. Metabolic diversity of
microbial BTEX degradation under aerobic or anoxic conditions, p. 11-14. In
Biotreatrnent: theuseofmicroorganismsinthetreatmentofhazardousmaterialsand
hazardous wastes. Washington, D. C.

Olsen, G. J., and C. R. Woese. 1993. Ribosomal RNA: a key to phylogeny.
FASEB 7.

Orvos,D. R., G. H.Lacy, andJ. Cairns. 1990. GeneticallyengineeredErwinia
carotovora: survival, intraspecific competition, and effects upon selected bacterial genera.
Appl. Environ. Microbiol. 56:1689-1694.

Owens, J. D., and R. M. Keddie. 1969. The nitrogen nutrition of soil and herbage
coryneform bacteria. J. Appl. Bacteriol. 32:338-347.

Paul, J. H., M. E. Frischer, and J. M. Thurmond. 1991. Gene transfer in marine
water column and sediment microcosms by natural plasmid transformation. Appl. Environ.
Microbiol. 57:1509-1515.

Peltonen, R., W.-H. Ling, O. Htinninen, and E. Eerola. 1992. An uncooked vegan
diet shifts the profile of human fecal microflora: computerized analysis of direct stool
sample gas-liquid chromatography profiles of bacterial cellular fatty acids. Appl. Environ.
Microbiol. 58:3660-3666.

Pipke, R., I. Wagner-Dobler, K. N. Timmis. and D. F. Dwyer. 1992. Survival
and function of a genetically engineered pseudomonad in aquatic sediment microcosms.
Appl. Environ. Microbiol. 58:125-1265.

Rajendran, N., O. Matsuda, N. Imarnura, and Y. Urushigawa. 1992. Variation in
microbial biomass and community structure in sediments of eutrophic bays as determined
by phospholipid ester- linked fatty acids. Appl. Environ. Microbiol. 58:562-571.

Ramos-Gonzalez, M., E. Duque and J. Ramos. 1991. Conjugation transfer of
recombinant DNA in cultures and in soils: host range of Pseudomonas putida TOL
plasmids. Appl. Environ. Microbiol. 57:3020-3027.

108

Ramsing, N. B., M. Kiihl, and B. B. Jorgensen. 1993. Distribution of sulfate-
reducing bacteria, Oz,andH28 inphotosynthetic biofrlmndeterminedbyoligonucleotide
probes and microelecuodes. Appl. Environ. Microbiol. 59:3840-3849.

Reysenbach, A.-L., L. J. Giver, G. 8. Wickham, and N. R. Pace. 1992.
Differential amplification of rRNA mes by polymerase chain reaction. Appl. Environ.
Microbiol. 58:3417-3418.

Ridgway, H. F., J. Safarik, D. Phipps, P. Carl, and D. Clark. 1990.
Identification and catabolic activity of well-derived gasoline-degrading bacteria from a
contaminated aquifer. Appl. Environ. Microbiol. 56:3565-3575.

Rittrnann, B. E., and P. L. McCarty. 1980. Model of steady-state-biofilm kinetics.
Biotechnol. Bioeng. 22:2343-2357.

Roberson, E. B., and MK. Firestone. 1992. Relationship between desiccation and
exopolysaccharide production ina soil Pseudomonas sp. Appl. Environ. Microbiol.
58: 1284-1291.

Robinson, J. A., and J. M. Tiedje. 1983. Nonlinear estimation of Monod growth
kinetic parameters from a single substrate depletion curve. Appl. Environ. Microbiol.
45:145 -1458.

Robinson, R. W., D. E. Akin, R. A. Nordstedt, M. V. Thomas, and H. C.
Aldrich. 1984. Light and electron microscopic examination of methane-producing biofilms
from anaerobic fixed-bed reactors. Appl. Environ. Microbiol. 48:127-136.

Robinson, J. A. 1985. Determining microbial kinetic parameters using nonlinear
regression: advantages and limitations in microbial ecology. Adv. Microb. Ecol. 8:61-114.

Romanowski, G., M. G. Lorenz, and W. Wackernagel. 1991. Adsorption of
plasmid DNA to mineral surfaces and protection against DNase 1. Appl. Environ.
Microbiol. 57:1057-1061.

Romonowski, G., M. G Lorenz, G. Sayler, and W. Wackernagel. 1992.
Persistence of free plasmid DNA in soil monitored by various methods, including a
transformation assay. Appl. Environ. Microbiol. 58:3012-3019.

Sasser, M. 1990. Technical Note#102: Tracking a strain using the Microbial
Identification System. MIDI Inc.

109

Schmidt, T. M., E. F. DeLong, and N. R. Pace. 1991. Analysis ofa marine
picoplgnkton community by 168 rRNA gene cloning and sequencing. J. Bacteriol.
173:4 71-4378.

Shaw, R. 1966. The polyunsaturated fatty acids of microorganisms, p. 107-173. In
Nderk. in Lipid Research, R. Paoletti and D. Kritchevsky, Editors. Academic Press,
ew

Shields, M., 8. Montgomery, P. Chapman, 8. CuskeyandP. H. Pritchard. 1989.
Novel pathway of toluene catabolism in the trichloroethylene-degrading bacterium G4.
Appl. Environ. Microbiol. 55: 1624-1629.

Shotton, D. M. 1989. Confocal scanning optical microscopy and its applications
for biological specimens. J. Cell Sci. 94:175-206.

Siegele, D. A., and R. Kolter. 1992. Life after log. J. Bacteriol. 174:345-348.

Simberloff, D. 8., and E. O. Wilson. 1969. Experimental zoogeography of island:
the colonization of empty islands. Ecology 50:278-296.

Simkins, 8., and M. Alexander. 1984. Models for mineralization kinetics with the
variables of substrate concentration and population density. Appl. Environ. Microbiol.
47: 1299- 1 306.

Smets, B. F., B. E. Rittrnann, and D. A. Stahl. 1993. The specific growth rate of
Pseudomonas putida PAWl influences the conjugal transfer rate of the TOL plasmid. Appl.
Environ. Microbiol. 59:3430-3437.

Speitel, G. E., C.-J. Lu, M. Turakhia, and X.-J. Zhu. 1989. Biodegradation of
trace concentrations of substituted phenols in granular activated carbon columns. Environ.
Sci. Technol. 23:68-74.

Spring, 8. R., R. Amann, W. Ludwig, K. H. Schleifer. and N. Petersen. 1992.
Phylogenetic diversity and identification of non-culturable magnetotactic bacteria. System.
Appl. Microbiol. 15:116-122.

Steffan, R. J ., and R. M. Atlas. 1991. Polymerase chain reaction: applications in
environmental microbiology. Ann. Rev. Microbiol. 45:137-161.

Stewart, G. J. 1989. The mechanism of natural transformation, p. 139-164. In
Gelae zanmfer in the environment, Levy and Miller, Editors. New York, McGraw-Hill
Pu lis ' g.

110

Stewart, G. J., and C. D. Sinigalliano. 1990. Dewction of horizontal gene transfer
bymumalmansformadoninnafiveandinmoducedspeciesofbactedainmafinemd
synthetic sediments. Appl. Environ. Microbiol. 56:18181-1824.

Tiedje ,.J M., R.C.Colwell,Y. L. Grossman,R.F. Hodson,R.E. .Lenski,R.
N. Mack,andP. J. Regal. 1989. Theplannedintroductionofgeneticallyen
organisms: ecologicalconsiderartionsandrecomendatrons. Ecology 70:298-315.

Torsvik, V., J. Gokscyr, and F. L. Daae. 1990. High diversity in DNA of soil
bacteria. Appl. Environ. Microbiol. 56:782-787.

Ulrike, E., T. Rogall, H. Blocker, M. Emde, and E. C. Bottger. 1989. Isolation
anddirectcompletenucleotidedeterminationofentire genes. Characterintionofagene
coding for 168 ribosomal RNA. Nucleic Acids Res. 17:7843-7853.

Van der Meer, J. R., W. M. de Vos, 8. Harayama, and A. J. B. Zehnder. 1992.
Molecular mechanisms of genetic adaptation to xenobiotic compounds. Microbiol. Rev.
56:677-694.

Van Loosdrecht, M. C. M., J. Lyklema, W. Norde, and A. J. B. Zehnder. 1990.
Influence of interfaces on microbial activity. Microbiol. Rev. 54:75-87.

Vandevivere, P., and D. L. Kirchman. 1993. Attachment stimulates
exopolysaccharide synthesis by a bacterium. Appl. Environ. Microbiol. 59:3280-3286.

Vaneechoutte, M., R. Rossau, P. De Vos, M. Gillis, D. Janssens, N. Paepe, A.
De Rouck, T. Piers, G. Claeys, and K. Kersters. 1992. Rapid identification of bacteria of
the Conramonadaceae with amplified ribosomal DNA-resuiction analysis (ARDRA).
FEMS Microbiol. Lett. 93:227-234.

Veldkamp, H., H. van Gemerden, W. Harder, and H. J. Laanbroek. 1983.
Competition among bacteria: an overview, p. 279-290. In Current Perspective in Microbial
Ecology, Klug and Reddy, Editors. Am. Soc. Microbiol., Washington, D. C.

Versalovic, J ., T. Koeuth, and J. R. Lupski. 1991. Disuibution of repetitive DNA
sequences in eubacteria and application to fingerprinting ofbacterial genomes. Nucleic
Acids Res. 19:6823-6831.

Vestal, J. R., and D. C. White. 1989. Lipid analysis in microbial ecology.
BioScience 39: 535-541.

111

Voice, T. C., X. Zhao, J. Shi, and R. F. Hickey. 1994. Remediation of
contaminants with biological activated carbon systems, p. 122-126. In Abstract
softhe20thAnnualRRELResearchSymposium. EnvironmentalProtection
Agency, proceedin(€irrcirrnrrti.OIrI.

Wagner, M., R. Erhart, W. Manz, R. Amman, H. Lemmer, D. Wedi, and K.-H.
Scheifer. 1994. Development of an rRNA-targeted oligonucleotide probe specific for the
genus Acinetobacter and its application for in sitrr mnnitoring in activated sludge. Appl.
Environ. Microbiol. 60:792-800.

Wagner- -D6bler, 1., R. Pipke, K. N. Timmis. and D. F. Dwyer. 1992. Evaluation
ofaquatic sedimentmicmcosmsandtheiruseinassesingpossibkeffectsofinuoduced
microorganisms on ecosystem parameters. Appl. Environ. Microbiol. 58: 1249-1258.

Ward, D. M., R. Weller, J. Shiea, R. W. Castenholz, and Y. Cohen. 1989. Hot
spring microbial mats: anoxygenic and oxygenic mats of possible evoluitonary
significance, p. 3-15. In Microbial Mats, Cohen and Rosenberg, Editors. Am. Soc.
Microbiol., Washington, D. C.

Ward, D. M., M. B. Bateson, R. Weller, and A. L. Ruff-Roberts. 1992.
Ribosomal RNA analysis of microorganisms as they occur in natrrre. Adv. Microb. Ecol.
12:219-286.

Warren, T. M., V. Williams, and M. Fletcher. 1992. Influence of solid surface,
adhesive ability, and inoculum size on bacterial colonization in microcosm studies. Appl.
Environ. Microbiol. 58:2954-2959.

Wayne, L. G., D. J. Brenner, R. R. Colwell, P. A. D. Grimont, O. Kandler, M. I.
Krichevsky, L. H. Moore, W. E. C. Moore, R. G. E. Murray, E. Stackebrandt, M. P.
Starr, and H. G. Truper. 1987. Report of the ad hoc committee on reconciliation of
approaches to bacterial systematics. Int. J. Syst. Bacteriol. 37:463-464.

Weller, R., and D. M. Ward. 1989. Selective recovery of 168 ribosomal RNA
sequences from natural microbial communities in the form of cDNA. Appl. Environ.
Microbiol. 55:1818-1822.

Weller, R., J. W. Weller, and D. M. Ward. 1991. 16S rRNA sequences of
uncultivated hot spring cyanobacterial mat inhabitants retrived as randomly primed cDNA.
Appl. Environ. Microbiol. 57:1146-1151.

Weller, R., M. M. Betson, B. K. Heimbuch, E. D. Kopczynski, and D. M. Ward.
1992. Uncultivated cyanobacteria, Chloroflexwr-like inhabitants of a hot spring microbial
mat. Appl. Environ. Microbiol. 58:3964-3969.

112
Wiens, J. A. 1983. Competition or peaceful coexistence? Nat. History 3:30-34.

Williamson, K., and P. L. McCarty. 1976. A model ofsubstrate utilization by
bacterial films. J. Cont. Poll. Fed. 48:9-24.

Wilson, T. 1989. Optical sectioning in confocal fluorescent microscopes. J.
Microsc. 154:143-156.

Woese, C. R. 1987. Bacterial evolution. Microbiol. Rev. 51:221-271.

Wolfaardt, G. M., J. R. Lawrence, M. J. Hendry, R. D. Roberts, and D. E.
Caldwell. 1993. Development of steady-state diffusion gradients for the cultivation of
deer microbial consortia. Appl. Environ. Microbiol. 59:2388-2396.

Wolfaardt, G. M., J. R. Lawrence, R. D. Robar'ts, S. J. Caldwell, and D. E.
Caldwell. 1994. Multicellular organization in a degradative biofilm community. Appl.
Environ. Microbiol. 60:434-446.

Wright, S. J., V. E. Centonze, S. A. Snicker, P. J. DeVaries, S. W. Paddock, and
G. Schatten. 1993. Introduction to confocal microscopy and three-dimensional
reconstruction, p. 1-45. In Cell Biological Applications of Confocal Microscopy,
Matsumoto, Editor. Academic Press, New York.

Wu, W. M., R. F. Hickey, and J. G. Zeikus. 1991. Characterization of metabolic
performance of methanogenic granules treating brewery wastewater: role of sulfate-
reducing bacteria. Appl. Environ. Microbiol. 57:3438-3449.

Yen, K.-M., M. R. Karl, L. M. Blatt, M. J. Simon, R. B. Winter, P. R. Fausset,
H. 8. Lu, A. A. Harcourt, and K. K. Chen. 1991. Cloning and characterization ofa
Pseudomonas niendocina KRl gene cluster encoding toluene-4-monooxygenase. J.
Bacteriol. 173:5315-5327.

Young, L. Y. 1990. Anaerobic degradation of aromatic compounds, p. 487-521. In
Soil Biochemistry, J.-M. B. and G. Stotzky, Editors. Marcel Dekker, Inc., New York.

Zimmerman, R., R. Ituriiaga, and J. Becker. 1978. Simultaneous determination of
the total number of aquatic bacteria and the number thereof involved in respiration. Appl.
Environ. Microbiol. 36:926-935.

Zylstra, G. J., and D. T. Gibson. 1989. Toluene degradation by Pseudornonas
putida F1: Nucleotide sequence of the tolCICZBADE genes and their expression in
Escherichia coli. J. Biol. Chem. 264: 14940-14946.

113

Zylstra, G. J., and D. T. Gibson. 1991. Aromatic hydrocarbon degradation: a
molecular approach. Gen. Engin. 13:183-203.

"lilillillllllllllr