.1... n. ;
..“...._.9

.1, .

.32;
~ . . . 3:...

:u......._ ‘.....:
.311... .2 ..

 

 

 

.1; 5.1.3., .. .13.: 2322.:
L1 SIP...

 

3.1:.
.2...

 

u" "-1 rI-MO- amt

.J l ,
P’“ “fig.

:3 .13; \

Nfl'hfll u: r

'1

 

£509"

—————“

THEStS

3

200!

 

FLIBRARY &
tafiichlgan Sm:
Unwersstv

 

This is to certify that the
dissertation entitled

Microbial communities in pristine and

tetrachloroethylene-contaminated aquifer
sediment.

presented by

Merry Shannon Riley-Buckley

has been accepted towards fulfillment
of the requirements for

PhD degreein Crop and Soil Science

4M4 i/f

la‘en C? L MAID

Major professor

g //Z/(9{

MS U is an Affirmative Action/Equal Opportunity Institution

0-12771

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

 

 

DAgADUE DATE DUE DATE DUE
' 20 4
SEP Si?» . [grip 20 2004
0‘.

V 00} 00 95084“

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateOue.p65-p.15

 

 

 

 

 

 

MICROBIAL COMMUNITIES IN PRISTINE AND
TETRACHLOROETHYLENE-CONTAMINATED AQUIFER SEDIMENT

By

Merry Shannon Riley-Buckley

AN ABSTRACT OF A DISSERTATION

Submitted to
Michigan State University
In partial fiilfillment of the requirements
For the degree of
DOCTOR OF PHILOSOPHY
Department of Crop and Soil Sciences

2001

Dr. Terence L. Marsh and Dr. James M. Tiedje

ABSTRACT

MICROBIAL COMMUNITIES IN PRISTINE AND
TETRACHLOROETHYLENE-CONTAMINATED AQUIFER SEDIMEN T

By
Merry Shannon Riley-Buckley

The microbial communities of the saturated subsurface play an important role in
the quality of groundwater, a resource of incalculable value to human and ecological
concerns. Further, although the magnitude of the effect that subsurface microbes have on
global biogeochemical cycling is not known, estimates suggest that because of the sheer
numbers of subsurface bacteria, they likely play a significant role in the cycling of
bioactive elements. Despite these facts, subsurface microbial communities have been
little studied until recently. In an effort to extend current knowledge in this area, a study
of the microbial communities in clean and chemically-contaminated aquifer sediment was
undertaken. The diversity and composition of the microbial communities at two sites was
assessed using molecular techniques. At a pristine site in Oyster, Virginia, the members
of microbial communities in oligotrophic sediment were characterized by phylogenetic
methods. The bacterial community profile from oligotrophic sediment was compared to
that of the communities in a contrasting geochemical environment in the same formation.
Similarities were found to exist between the two communities. Bacterial community
diversity in aquifer sediment was markedly lower than that observed in surface soils. A
novel group of phylogenetically deeply-branching bacteria was detected at the site and
found to exist in eight out of ten sediments tested. The microbial communities at a
tetrachloroethylene-contarninated site in Oscoda, Michigan were also studied. Members

of communities in clean and contaminated sediment were described phylogenetically, and

found to be closely related to the B, y, and 5 Proteobacteria, the Acidobacteria,
Leptospirillum / Nitrospira, and Green nonsulfur bacteria. Within the plume, 16S
sequences closely related to the species Syntrophus were identified, suggesting that
interspecies hydrogen or acetate transfer played a role in carbon cycling at the site.
Bacterial diversity was found to be significantly higher in contaminated sediment than in
sediment from upstream of the plume. Finally, the distribution and abundance of a
sulfate-reducing bacterium which was detected within the plume was studied using the
emerging method of real-time PCR. The 168 sequence of this organism was found to
constitute only a small proportion of total bacterial l6S sequences at the site, and was

highly over-represented in the 168 clone library.

AKNOWLEDGEMENTS

Much of my success in graduate school I owe to my advisor Terry Marsh, who
presented a fantastic example of what a scientist could be by conducting exceptional
research while maintaining sanity and perspective. Thanks for your guidance. I will
remember: chop wood, carry water.

Thanks also go to the other faculty members who have helped me in my work,
including my committee members Dr. Jim Tiedje, Dr. Del Mokma, and Dr. Mike Klug.
Thanks also go to Dr. John Breznak for his help in designing the cultivation experiments.

Thanks to my parents Bert and Pauline, who, although they really wanted me to
be a brain surgeon, told me I should choose whatever path leads to my happiness and
fulfillment. How can I thank you enough for your unending encouragment and support?

And a final word of thanks goes to Dan who held me up in my periodic bouts with
frustration and self-doubt during these past five years. You are my rock. Thank you for

choosing to spend your life with me.

iv

TABLE OF CONTENTS

 

 

PAGE
LIST OF TABLES viii
LIST OF FIGURES )1
LIST OF ABBREVIATIONS xii
CHAPTER 1: MICROBIAL COMMUNITIES IN THE SATURATED
SUBSURFACE 1
1. Introduction 1
2. The importance of groundwater l
2.1. Human use of groundwater 1
2.2. Ecological significance of groundwater 2
2.3. Global nutrient cycles and the subsurface 2
3. Biogeochemistgr of the saturated subsurface 3
3.1. Aerobic reduction-oxidation processes 3
3.2. Anaerobic reduction-oxidation processes 7
3.3. Other processes not linked to energy conservation ll
4. Previously identified aguifer microbes 12
4.1. Eukarya 12
4.2. Archaea 14
4.3. Bacteria 16
5. Methods for analysis of microbial communities 22
5.1. Cultivation and characterization of isolates 22
5.2. Analysis of functional genes 22
5.3. Analysis of the metabolism of radio-labeled
substrates or electron acceptors 23
5.4. Analysis of membrane components - PLF A and FAME 23
5.5. Analysis of small subunit rRNA or small subunit rDNA 24
6. Specific aims of this work. 29
7. References 30
CHAPTER 2: A CU LTIVATION-INDEPENDENT ANALYSIS OF BACTERIAL
COMMUNITIES IN PRISTINE AQUIFER SEDIMENT 35
Introduction 35
Materials and methods 36
Sample collection 36
DNA extraction 38
16S rDNA T-RFLP 39
T-RFLP analyses 40
Cloning of 16S rRNA genes 41
Screening of rDNA clones by ARDRA 41

Sequencing of rDNA clones 42

Phylogenetic analyses 42
Results 43
Description of the study site 43
16S rDNA T-RFLP 43
Phylogenetic analysis 45
Discussion 46
References 66

CHAPTER 3: DETECTION OF A NOVEL, DEEPLY-BRAN CHIN G GROUP OF

BACTERIA IN PRISTINE AQUIFER SEDIMENT 70
Introduction 70
Materials and methods 72

Sample collection 72
DNA extraction 72
Primer design 72
Cloning of 16S rRNA genes 72
Screening of 16S clones by ARDRA 74
Sequencing of rDNA clones 75
Phylogenetic analyses 75
Results 76
Description of the study site 76
Phylogenetic analyses 76
Discussion 77
References 82

CHAPTER 4: A MOLECULAR COMPARISON OF MICROBIAL
COMIVIUNITIES IN PRISTINE AND TETRACHLOROETHYLENE—

CONTAMINATED AQUIFER SEDIMENT 83
Introduction 83
Materials and methods 85

Sample collection 85
DNA extraction 85
16S rDNA T-RFLP 86
T-RFLP analyses 87
Cloning of bacterial 16S rRNA genes 88
Screening of clones by ARDRA 88
Sequencing of rDNA clones 89
Phylogenetic analyses 89
Results 90
Description of the study site 90
Profiles of microbial communities in clean
and contaminated sediment 91
Bacterial diversity in clean and contaminated aquifer sediment 93

Discussion 101
References l 10

CHAPTER 5: QUANTIFICATION OF BACTERIAL, ARCHAEAL, AND
SPECIES-SPECIFIC 16$ GENES IN TETRACHLOROETHYLENE—
CONTAMINATED AQUIF ER SEDIIVIEN T USING REAL-TME PCR 114

Introduction 1 14

Materials and methods 116

Sample collection 116

DNA extraction 117

Primer design 118

Real-time PCR 1 l9

Controls 120

Results 121

Description of the study site 121

Primer specificity 122

Detection limits 124

Quantification limits 124
Quantification of bacterial, archaeal, eukaryal,

and lel SSU genes 125

Discussion 129

References 140

CHAPTER 6: CONCLUSIONS 144

References 149

APPENDIX A: CULTIVATION TECHNIQUES USED TO ISOLATE N CFA
GROUP I BACTERIA 150

APPENDIX B: T-RFLP DATA FROM CHAPTERS 2 AND 4 154
APPENDIX C: RAPID PHENOTYPIC CHANGE AND DIVERSIFICATION OF A

SOIL BACTERIUM DURING 1000 GENERATIONS OF EXPERINIENTAL
EVOLUTION 184

vii

Chi;

.L_

new
1.
Chi);
. I. _:
”Mu-5
ix,“ -
“-5 i.
:Wx' I
haul ‘

 

 

"1m ~- .
““Q‘u
l

l“

LIST OF TABLES

Chapter 1

Table 1 Bacterial isolates and 16S clones previously derived from the subsurface. The
division (or putative division) and genera are listed if determined for a given isolate or
sequence. All isolates and clones were phylogenetically identified by sequencing of the
16S gene, except isolates developed by Fries et al (24), which were identified through

fatty acid methyl ester (FAME) analysis. - Pages 19-21
Chapter 2

Table 1 Total number of terminal fragments detected in Hha I bacterial T-RFLP profiles
from Oyster, Virginia Page 44

Table 2 Phylogenetic placement of bacterial 16S sequences from the narrow channel
focus area in Oyster, Virginia Pages 47-49
Table 3 Presence/absence of deeply branching clone terminal restriction fragments in
TRFLP profiles from the narrow channel study site Page 60
Chapter 4

Table 1 Phylogenetic affiliations of 16S rDNA clones from the Bachman Road site and
BLAST search results Pages 99-100
Chapter 5

Table 1 Primer pairs for real-time PCR Page 119

Table 2 The numbers of bacterial, archaeal, eukaryal and clone le1 SSU targets per
gram of aquifer sediment (Wurtsmith, Bachman, and Oyster samples) or per ul of directly
extracted DNA (bioreactor and activated sludge samples), and the standard error of each
measurement. "% of total" is the percent of SSU genes that each kingdom contributed to
the total SSU genes detected for a given sample. The percent contribution to the total
number of bacterial 168 genes detected is reported for lel 168 sequences. The "+"
symbol indicates that le1 sequences were detected in a sample but not quantitated and

a "-" indicates that the group was not detected. Pages 127-128
Appendix A

Table 1 Media used for cultivation of NCFA group I. (Quantities are in units of g per L).
Store metals solution at 4 degrees C. Page 151

Table 2 Trace metals solution for growth media. (Quantities are in units of g per L)
Page 151

viii

All.

la‘:

 

 

Appendix B
Table 1 Chapter 3 data. Pages 154-169

Table 2 Chapter 4 bacterial T-RFLP data Pages 170-181

Table 3 Chapter 4 Archaeal T-RFLP data Pages 182-183

ix

LIST OF FIGURES

Chapter 1

Figure 1 Radial phylogram of selected bacterial divisions. Shaded wedges indicate
divisions for which representative isolates or 168 sequences have been recovered from
the subsurface. Adapted from Hugenholtz et al. (3 2). Page 18

Chapter 2

Figure 1 Map of Oyster, Virginia, indicating the locations of the south Oyster focus area
(SOFA) and the narrow channel focus area (NCFA). Below, the areas of the SOFA and
the NCFA have been expanded to indicate the positions of sampling points.

Page 37

Figure 2 a-e Maximum likelihood dendrograms of bacterial 16S sequences obtained from
the narrow channel focus area in Oyster, Virginia. Divisions are listed out side the
brackets for panels b through d and subdivisions are listed in panel a (Proteobacteria).
Optimality criteria used in bootstrap analysis of the sequences were: maximum
likelihood, maximum parsimony, and neighbor joining. Bifiircations supported
(bootstrap values >75%) by one optimality criteria but are only marginally supported (50-
75%) or not supported (<50%) by the other criteria are indicated with open circles.
Bifirrcations supported by two or three of the criteria are indicated with closed circles.
The number of characters in each analysis and the range of the mask (E. coli numbering)
were as follows: a) 330 characters, bases 216-484, b) 443 characters. bases 134-602, c)
332 characters, bases 116-527, d) 432 characters, bases 134-537, and e) 1165 characters,
bases 92-1390. Pages 50-54

Figure 3 Difi‘erences between T-RFLP profiles generated using two different reverse
primers. In comparing the profiles of template NC B2 (6 m) created with primer set (a)
27F/1392R and (b) 27F/1525R, we see the appearance and dissappearance of three
Significant peaks. The profiles of NC M3 (6 m) created using the same two primer sets
(c) 27F/1392R and (d) 27F/1525R are more comparable and all significant peaks are

found inboth profiles. Page 58
Chapter 3

FlSure 1 Map of Oyster, Virginia, indicating the location of the narrow channel focus
area and the relative positions of sampling points within the site. Page 73

Figure 2 Maximum likelihood dendrogram of NCFA group I 168 sequences obtained
(from the narrow channel focus area in Oyster, Virginia. Divisions are listed outside the
brackets. Optimality criteria used in bootstrap analysis of the sequences were: maximum
likelihood, maximum parsimony, and neighbor joining. Bifiircations supported
(bootstrap values >75%) by one optimality criteria but are only marginally supported (50-
75%) or not supported (<50%) by the other criteria are indicated with open circles.

Bifirrcations supported by two or three of the criteria are indicated with closed circles.
There~ were 1165 characters included in the analysis and the range of the mask (E. coli
numbering) covered bases 92-1390. Page 78

Chapter 4

Figure 1 A comparison of bacterial community T-RFLP profiles from contaminated (top -
IN) and pristine (bottom - 4A) aquifer sediment. In this comparison, each of the five
terminal fragments detected in the pristine sediment is also detected among the fi'agments
in the contaminated sediment profile. Page 92

Figure 2 a-d Maximum likelihood dendrograms of bacterial 16S sequences obtained from
the Bachman Road site in Oscoda, Michigan. Divisions are listed out side the brackets in
panels d and c (clone group OPB80, and Green non-sulfirr) and subdivisions are listed in
panels a (Proteobacteria), b (Proteobacteria), and c (clone group T78). Optimality criteria
used in bootstrap analysis of the sequences were: maximum likelihood, maximum
parsimony, and neighbor joining. Bifiircations supported (bootstrap values >75%) by one
optimality criteria but are only marginally supported (50-75%) or not supported (<50%)
by the other criteria are indicated with open circles. Bifurcations supported by two or
three of the criteria are indicated with closed circles. The number of characters in each
analysis and the range of the mask (E. coli numbering) were as follows: a) 436
characters, bases 113-600, b) 285 characters, bases 111-422, c) 414 characters, bases 98-
537, and d) 350 characters, bases 134—537. Pages 95-98

Chapter 5

Figure 1 a-d Standard curves used to interpret real-time PCR data. Panel a depicts
standard reactions with bacterial-specific primers, panel b depicts the archaeal standard
reactions, panel c depicts the eukaryal reactions, and panel (1 depicts the standard
reactions with clone lel-specific primers. Error bars indicate standard deviation among
triplicate standards. Page 126

Figure 2 The relative contributions of bacterial, archaeal, and eukaryal SSU rRNA genes
to total numbers of rRNA genes detected in DNA from activated sludge, bioreactor fluid,
PCB-contaminated aquifer sediment from the Bachman site, pristine aquifer sediment
from the Oyster site, and a contaminated aquifer at the former Wurtsmith Air Force Base.
Archaea] 168 was not detected in either of the activated sludge replicates or in sample
ODU4 6.3-6.4 m. Eukaryal 188 was detected only in Wurtsmith sediment ML3 (23.5 —
27.5 it) and in the activated sludge samples. Page 132

LIST OF ABBREVIATIONS

U = units of activity

bp = nucleotide base pairs
it = feet

DNA = deoxynucleic acid
RNA = ribonucleic acid

dNTP = deoxynucleotide triphosphate

xii

CHAPTER 1

MICROBIAL COMMUNITIES IN THE SATURATED SUBSURFACE

1. INTRODUCTION

The microbial communities of the saturated subsurface are complex, diverse, and, in
general, are not well characterized. The human need for quality groundwater resources is
expected to increase in the next century (50) even while more and more aquifers are
being contaminated by anthropogenic chemicals and human waste. Groundwater is also
a key component of aquatic ecosystems like lakes, rivers, and coastal areas, and the
chemical properties- of this water have a profound effect on ecosystem function.
Furthermore, the subsurface is a component of the global cycles of such bioactive
elements as carbon, nitrogen, sulfiir, and phosphorous, to name a few. The chemical
reactions that determine water quality and nutrient cycling in the subsurface are almost
exclusively mediated by microbes. In light of this fact, the importance of bridging the
gap in our understanding is apparent.

This chapter will review the importance of groundwater to human and ecological
concerns, some of the biogeochemical transformations that subsurface microbes carry
out, current knowledge about the distribution of different phylogenetic groups found in
the subsurface, and methods for analyzing subsurface microbial communities. Finally,
the goals and methods of my research into subsurface microbial communities are
described.

2. THE IMPORTANCE OF GROUNDWATER

2.1. Human use of groundwater

 

3.4..

Aw).

 

4"“.

~..
‘4

'7‘ ‘h;
5" l1.‘

It is difficult to overestimate the importance of groundwater in the United States. It is
estimated that 50% of all Americans and 98% of the population living in rural areas rely
on groundwater as their primary source of water (50). The most recent data available
indicate that over 77 billion gallons of groundwater is pumped per day in the United
States (4). Statistics from the US. Geological Survey indicate 19% of groundwater is
drawn for public drinking water and 63% is used for irrigation (4).
2.2. Ecological significance of groundwater

Approximately 99.7% of the world’s freshwater (not in the form of ice and snow) is held
in the subsurface. This groundwater serves as an important component of the hydrologic
cycle, particularly in regard to the cycling of freshwater. Precipitation on land eventually
becomes surface runoff which, in most geologic environments, infiltrates to the
subsurface. Subsurface water has a relatively high residence time, between two weeks
and 10,000 years, which may be longer than the residence times calculated for the ocean
(about 4000 years). Groundwater which is not held in the primary lithosphere (the rock
that forms the core of the earth) eventually exits the subsurface and discharges into lakes,
rivers, or wetlands, or it empties directly into the ocean. Each of these environments are
complex ecosystems, and groundwater flow rate and water quality have an impact on the
function of these ecosystems. The flow of subsurface water into surface ecosystems
influences water availability and habitat suitability, among other things, and the quality of
subsurface water in these systems often selects for the species that live there. So, the
processes that take place in subsurface environments are felt in other environments by
virtue of the role of groundwater in the hydrologic cycle.

2.3. Global nutrient cycles and the subsurface

The subsurface serves as a compartment of the global cycles of every known element of
biological importance, including carbon, iron, sulfirr, and oxygen. Since primary
production is thought to be insignificant in the subsurface (27), the dominant processes
are likely to be degradative (14, 38). The extent to which subsurface environments effect
global nutrient cycles is unknown.
3. BIOGEOCHEMISTRY OF THE SATURATED SUBSURFACE
Madsen and Ghiorse (3 8) have estimated that the total microbial biomass in subsurface
environments worldwide may be 40 times greater than in the top 1 m of soil. Moreover,
the subsurface microbiota catalyzes a vast array of chemical transformations that affect
global cycles of every bioactive element (27). Here, these processes are classified as
aerobic reduction-oxidation processes, anaerobic reduction-oxidation processes, or
processes not linked to energy conservation.

3.1. Aerobic reduction-oxidation processes

3.1.1. Degradation of organic compounds

Degradation of organic molecules in groundwater is crucial to efficient global cycling of
carbon. Carbon is fixed by green plants and other autotrophic organisms on the surface
and incorporated into plant tissues. Upon decomposition of these tissues by surface
microbes, soluble organic compounds are carried to the subsurface by infiltrating water.
In shallow or oxygenated systems, degradation of this dissolved carbon is largely
mediated by aerobic, heterotrophic bacteria. The major carbon end product of aerobic
heterotrophic metabolism is C02, a pivotal component of the carbon cycle. In this form,
carbon may either re-enter the atmosphere or it may be taken up by other microbial

processes, including methanogenesis and carbon fixation. Organic carbon concentrations

Act
t1.

 

are.”

Tie;

. .__
flyORO

than ...

0.
Fa: y
1.5."...

Pbflarj
ens.
b

Inn-H

au..\'.

 

T3: 3»

cA‘ “

“4'

“5 J58
l‘ .

CC." ‘_‘

t‘h“‘v'

.- f"?
\‘NLL
I

I
14..

can vary widely across different formations, but it is thought that most pristine aquifers
are carbon-limited (27, 38).

3.1.2. Nitrification
The energy conserving process by which nitrate is created from ammonia is referred to as
nitrification. Ammonia is generated by the anaerobic decomposition of plant and animal
matter. Generation of nitrate from ammonia is a two step process undertaken by different
genera: usually Nitrosomonas and Nitrobacter. Nitrosomas (or Nitrocystis) uses
ammonia as an electron donor in the overall energy-yielding processes:
NH; + V202 -) NH20H (rxn #1)
NH20H + 02 9 HNO2 + H20 (rxn #2)
Nitrite (N02') is used as an electron donor by Nitrobacter (or Nitrococcus)
species in the following reaction:
N02' + V202 9 N03' (rxn #3)
The final product, nitrate, can be assimilated by a variety of microorganisms, reduced,
and used as NH2 groups in amino acids and proteins. Alternatively, under anaerobic
conditions nitrate can be used as an electron acceptor by denitrifiers. Under the aerobic,
oligotrophic conditions observed in shallow local flow systems however, when nitrate is
introduced by agriculture or other practices it tends to accumulate (14).

3.1.3. Sulfide or elemental sulfur oxidation
Chemolithotrophic sulfide oxdizing bacteria have been shown to occur in a number of
subsurface systems (22). Coal spoil piles, in particular, encourage the growth of these
organisms, as they leach elemental sulfiir in an aerobic environment, where subsequent

bacterial activity creates acid mine drainage. In most systems, however, sulfur is found

in trace quantities and originates from minerals or from the decomposing organic matter
in the soil layer. The most widely studied sulfur oxidizer, Thiobacillus, uses sulfide or
elemental sulfur as an electron donor and molecular oxygen as an electron acceptor for

energy conservation:

2st + 02 -) 2H2O + 25° (rxn #4)
or
23° + 02 + 2H2O -> 2H2SO4 (rxn #5)

Reduced sulfiir species are typically available in anoxic areas where microbial activity
has reduced oxidized sulfur species. By definition, however, oxygen is limiting in these
zones, and for sulfur oxidation to proceed, these reduced species need to be transported to
an oxic environment. Alternatively, sulfide that is associated with the mineral substratum
can be utilized for energy-yielding reactions. For example, Chapelle (15) observed pyrite
oxidation in an aerobic subsurface system. The overall reaction is:

FeS2 + 15/4 02 + 7/2 H20 -) Fe(0H)3 + 4H* + 230.” (rxn #6)

In aerobic systems with a rich source of sulfide or sulfirr, sulfate will tend to accumulate.
This is seldom the case, however, as extensive sulfur oxidation will quickly deplete the
environment of oxygen and the buildup of sulfuric acid acts as a negative feedback on
sulfur oxidation. The end product of sulfide oxidation, sulfate, may accumulate in the
subsurface, it may be assimilated by other microorganisms for use in protein production,
or it may be used as a terminal electron acceptor for sulfate reducing bacteria in
anaerobic zones of the aquifer.

3.1.4. FeH oxidation

 

I
H)“.

val-L.»

61552?

 

'T. 1 a
HAIL) ::
2L .,

P. V
’Io
l'g‘filffl
C .
.’.
was J.-

Iron oxidizing bacteria, or “iron bacteria” are so common in groundwater that they are
considered a microbiological pest by the well water industry. In the saturated subsurface,
iron can be derived directly from mineral sources or from the breakdown of organic
matter. Ferrous iron is generally found in environments where anaerobic respiration has
taken insoluble ferric iron and converted it to soluble ferrous iron. Iron oxidizing
bacteria like Gallionella use ferrous iron as an electron donor and molecular oxygen as an
electron acceptor in the following overall reaction: ,
Fe2+ + 4H+ + 02 -> Fe3+ + 2H2O (rxn #7)
This Chemolithotrophic activity is observed most often in sediments that are anaerobic
but are exposed to the atmosphere either by the intrusion of a well or by natural discharge
into an aerobic environment like a river. Microbial iron oxidation is responsible for the
thick, slimy coatings of ferric iron observed in some wells or at bottom of a stream or
river where an anaerobic aquifer discharges.
3.1.5. C02 fixation
Chemolithotrophs, like those that oxidize SUlfiJT, iron, manganese, and hydrogen, can
utilize organic compounds as a carbon source in a manner much like chemoheterotrophs
do. Utilization of inorganic electron donors and acceptors and an organic carbon source
is called mixotrophy. Alternatively, many chemolithotrophs can autotrophically fix C02
through use of the Calvin cycle or the TCA cycle; this is known as chemolithoautotrophy.
3.1.6. H2 oxidation
Hydrogen oxidizing bacteria have been found in a deep subsurface sediment by
Fredrickson et al. (22). Hydrogen oxidation is coupled to aerobic respiration in the

following overall reaction:

 

 

m— ~

IAA.““

2H2 + 02 -) 2H20 (rxn #8)
In surface aquatic environments, where they have been most extensively studied,
hydrogen oxidizers are generally found at the interfaces of oxic and anoxic waters. In
these areas, H2 from anaerobic fermentation processes mixes with low concentrations of
oxygen, providing the substrates needed for the above energy-conserving reaction.

3.1.7. Oxidation of other reduced species — Mn, P, As, Sb, Mo, U, Se
Other reduced compounds may be used in energy conserving reactions, including
manganese, reduced phosphorous, arsenite, antimony, molybdenite, uranium, and
selenium. It is not known how important these processes are in the saturated subsurface,
but they may be most relevant in systems where high concentrations of these elements are
found.

~3.2. Anaerobic reduction-oxidation processes

3.2.1. Organics degradation
Organic compounds in anoxic environments may be degraded by fennenters or
anerobically respiring microbes. Many fermentation processes degrade high molecular
weight organic compounds to organic acids and hydrogen. These fermentation products
may be taken up as substrates by anaerobic respirers or methanogens that convert them to
carbon dioxide and methane.

3.2.2. Fermentative production of formate, acetate, lactate
Under anaerobic conditions, ferrnentative processes may play a large role in organics
cycling. Fermenters in the subsurface and other environments conserve energy by using

an organic compound as an electron donor and either organic compounds or protons as

 

Py‘o
l
-..

F;;'
4.».

323.67.

”Mr
15: ”,5.

‘

I‘_

 

 

 

the electron acceptor. Examples of ferrnentative processes follow. Here, lactate and
ethanol are produced from glucose:
C5H1206 9 CH3CHOHC00H + CH3CH20H + C02 (rxn #9)
In an unrelated reaction, lactate may be transformed to pr0pionate and acetate:
3CH3CH0HC00H -) 2 CH3CH2COOH + CH3C00H + C02 (rxn #10)
Protons serve as electron acceptor in the following fermentation of ethanol to acetate:
CH3CH20H + 2H20 -> CH3C00H + H“ + 4H2 (rxn #11)
Small organic molecules like acetate, propionate, and forrnate are suitable substrates for
anaerobically reSpiring bacteria. Hydrogen, too, is an extremely important substrate for
anaerobic respirers. The passage of these substrates on to anaerobic respirers allows
fermenters and respirers to co-exists under oligotrophic conditions.

3.2.3. Iron reduction
Iron may serve as a terminal electron acceptor for bacterial respiration in the absence of
oxygen. This process is thought to be relatively important in the saturated subsurface,
where mineral iron is often found in high concentrations and oxygen may be depleted.
Both autotrophic and heterotrophic iron reducers have been isolated, and potential
electron donors include volatile fatty acids, aromatics, and SO. Ferric iron is reduced in
the following overall reaction:
2Fe3+ + 2e -) 2Fe2+ (rxn #12)
Iron reduction is responsible for the solubilization of ferric iron in many aquifers, often to
such an extent that water treatment facilities must be created specifically for its removal.

3.2.4. Dissimilatory nitrate reduction / denitrification

 

“I"?
has...

E”
iii?» .7

533.5

 

 

Due to the paucity of nitrate in pristine groundwater, nitrate reduction is usually
considered an important process only in contaminated environments. Contamination may
originate from farm runoff or septic leachate or any of a number of different sources. In
the absence of the more energetically favorable electron acceptors oxygen and Fe (II),
nitrate can be reduced to dinitrogen by a single organism or stepwise by more than one
species as some species lack either nitrate reductase or nitrite reductase. Alternatively,
nitrate may be reduced to nitrite and then to ammonia in a process referred to as nitrite
ammonification. Dissimilatory nitrate reducers use nitrate as an electron acceptor and
they may use any of a number of organic metabolites or reduced sulfur such as S0 or H2S

as an electron donor. The steps of nitrate reduction are as follows.

N03' + 2H+ + 2c -> N02'+ H2O (rxn #13)
N02'+2H++2e-)N0+H20 (rxn #14)
2NO+2H++2e -)N20+H20 (rxn #15)
N20+2H*+2e-)N2+H20 (rxn #16)

The overall process of nitrite ammonification is:
NO2' + 7H++ 6e 9 NH3 + 2H2O (rxn #17)
It has been thought that denitrification occurs only under anoxic conditions, but
Robertson and Kuenen (43) have shown nitrate reduction in Thiosphaera pantotropha at
atmospheric oxygen concentrations.

3.2.5. Sulfate reduction
Sulfate reduction has been shown to be of great importance to the geochemistry of
certain saturated subsurface environments. Sulfate and other oxidized sulfiir species are

derived from degrading organic matter at the surface or from the microbial process of

g?

I." .;p

L~il¥\.

ta
~-.'$-~ .
. .

AAh'c\

I
Q0 ‘

I .
v 9
<I":fi‘. .-
vbuuu;
5.1-;
Hem“,
Mae...
ri~k‘v O
M

‘ cut .

FIOCESE

 

 

ll" em

1
A“

sulfur oxidation described previously. Saturated subsurface environments may have high
concentrations of sulfate where sulfirr minerals like pyrite are present, but most ofien
sulfate is present in small concentrations. When the more energetically favorable
electron acceptors oxygen, Fe (III), and nitrate have been depleted from an environment,
available sulfate may be used for anaerobic respiration. The overall reaction follows.
50." + 5H+ + 4e' -> HS’ + 4H2O (rxn #18)
Sulfate reduction results in the production of sulfide, which is toxic to most organisms,
including sulfate reducing bacteria. Dissolved metals like Fe (II), however, will tend to
precipitate sulfide, lowering the toxicity.

3.2.6. Methanogenesis
In the absence of any other terminal electron acceptor, C02 or acetate may be used for the
process of methanogenesis. This can be a significant process in many subsurface systems
where electron acceptors have been depleted by high activity of respirative organisms.
Stevens and McKinley (54) described the first subsurface ecosystem found in which
A autotrophic methanogenesis was the dominant metabolism. In methanogenesis, C02
produced by respirative organisms is used as a terminal electron acceptor and any of a
number of simple organic compounds or H2 (autotrophic methanogenesis) may be used as
en electron donor.
C02 + H2 9 CH4 + 2H20 (rxn #19)
Alternatively, acetate produced by respirers may be cleaved by methanogens in

aceticlastic methanogenesis:

CH3COOH 9 CH4 + C02 (rxn #20)

10

B";
k...

 

 

 

 

 

In subsurface systems where methane cannot cycle to an aerobic zone (where it could be
aerobically converted to C02) the carbon cycle is truncated and natural gas accumulates.
3.3. Other processes not linked to energy conservation

3.3.1. Production of organic and inorganic acids that encourage dissolution
Both heterotrophic and Chemolithotrophic microbes can contribute to mineral dissolution
by generation of acidic end-products. Heterotrophic metabolism in the subsurface
produces organic acids and C02 can hydrolyze to form carbonic acid in solution. Some
chemolithotrophs form large quantities of sulfuric acid or nitric acid. These acids can
react with minerals in the saturated subsurface, bringing about dissolution or alteration of
sediments.

3.3.2. Production of siderophores that encourage dissolution
Siderophores, particularly iron-specific siderophores, can cause dissolution of iron
minerals. Iron reducers like Geobacter metallireducens produce these chelators in order
to scavenge insoluble Fe (HI) from mineral surfaces. Chelated Fe (III) can be readily
utilized for metabolic processes.

3.3.3. Agents of concentration
Microbes may serve as agents of concentration in the subsurface. Certain species
actively concentrate inorganic matter intracellularly, as with intracellular sulfur
depOsition by sulfur bacteria. Passive accumulation of metals may occur on the sheath of
bacteria like Leptothrix and Sphaerotilus and in biofilms. Other species have been found
to encourage CaC03 precipitation outside of the cell (45, 60).

3.3.4. Agents of fractionation

11

Microbes may act as agents of fractionation by selective uptake of inorganic species.
Preferential action may take the form of differential use of metals as terminal electron
acceptors, as in the reduction of Mn (IV) before Fe (III). Other activities create stable
isotope fractionations. Microbes preferentially utilize light stable isotopes as substrates
and electron acceptors. The isotopes which have been shown to fractionate include: 328
and 34s, 12c: and 13c, 16c and 18C, 1‘N and 15N, 16o and 18o.
4. Previously identified aquifer microbes

4.1. Eukarya

4.1.1. Patterns of abundance

Eukaryotes are thought to be present in low numbers in the saturated subsurface. In
investigations in which they were observed, eukaryotic populations have been found to
be significantly smaller than those of bacteria or archaea (5, 9, 39, 59).
It has been asserted that populations of protozoa rely on the availability and numbers of
their bacterial prey (20) and from this fact it has been inferred that the low counts of
bacteria in the subsurface are the reason for the low observed numbers of protozoa (39).
Sinclair et al. (48) detected protozoa in sediment cores only in samples with bacterial
counts greater than 104 CFU/gdw. Madsen et al. (3 9) cultivated relatively high numbers
of protozoa in aquifer sediments contaminated with polyaromatic hydrocarbons (PAH’s)
compared to the counts observed in uncontaminated sediments from the same site (400
protozoa per gram versus <50 per gram). The authors assert that higher bacterial growth
rates observed in the contaminated sediments are the cause of the high protozoa counts.
Sinclair et al. (47) detected high numbers of protozoa in the unsaturated zone of a fuel-

contaminated site. Numbers were also high in the saturated zone where contaminated

12

 

TESL

prc.

l c

H,- .-
. (‘1‘.3‘

lH~I an
-»0..__

 

 

sediment met oxygenated groundwater. The authors argue that these observations are the
result of high oxygen sensitivity in protozoa. Using MPN techniques to examine
protozoa in sediments taken from within and outside of a plume of monoaromatic
hydrocarbons, Zarda et al. (61) noted high counts in surface soil, a decrease in numbers in
the vadose zone, and an increase in the saturated sediments. Strauss and Dodds (55)
observed a steep decrease in protozoa numbers with depth in a profile of a pristine site;
counts went from 105 protozoa/g dw in surface soils to 392 protozoa/g dw in saturated
sediments at a depth of 10.3 meters. Sinclair and Ghiorse (48) observed two patterns in a
pristine aquifer in Oklahoma. They noted both a decrease in protozoa counts with depth
and that deep layers with high permeability present an exception to this rule in that higher
protozoa counts were observed in coarser sediments. Using these observations, the
authors speculate that protozoa abundance is affected by pore neck size and forces that
carry these organisms fi'om the more heavily populated surface soils.

Only a few investigations have sought to quantify filngal populations in the subsurface.
Madsen et al. (39) observed on the order of 10 firngal CFU per gram in contaminated and
uncontaminated sediment from up to 10 m below the soil surface. Bone and Balkwill
(10) failed to observe yeast and other “eukaryotic forms” in sediment samples from
below 0.2 m deep. White et al. (59), failed to detect long-chain polyenoic fatty acids,
indicators of eukaryotes, in samples of aquifer clay from 400 meters below the surface.
Ludvigsen et al. (36) detected fatty acid esters which they attribute to microan or algae
in extracts of landfill-leachate contaminated sediment samples taken from 12 m below the
surface.

4.1.2. Identity of eukaryotic organisms

13

I”
"5

101g

9::-
{’1

W838
t ‘4'
Dim).

049.3:

on. a, -
may. ‘

“(5.. .‘

I‘LV‘L‘

 

The specific identity of subsurface eukaryotes has seldom been investigated beyond
rough estimates of identity by morphological examination. Sinclair and Ghiorse (48)
noted, using most-probable-number estimates, that ciliates decreased to extinction within
0.5 m of the surface in an uncontaminated site in Oklahoma. Flagellates and amoeba
were detected throughout the profile. In examining the culturable protozoa from a
hydrocarbon-contaminated plume near Hunxe, Germany, Zarda et al. (61) observed
ciliates and naked amoebae, but the dominant protozoa were flagellates. Studies
employing molecular techniques to phylogenetically identify subsurface
microeukaryoates have yet to be reported in the literature.
4. 2. Archaea
4.2.1. Patterns of abundance

While archaea have been isolated from enrichments of subsurface inoculum, few studies
have undertaken a description of the in situ abundance and activities of archaeal
populations in the subsurface. Bekins et al. (8) studied the distribution of several
microbial groups, including methanogens, in a contaminated aquifer in Bemidji,
Minnesota with most probable number (MPN) techniques. They found that methanogen
populations, like the other groups, had a greater affinity for sediment surfaces (85%) than
they did for the interstitial space. Furthermore, they detected 10-100 times more
methanogens per gram of sediment in the unsaturated zone than in the saturated zone,
possibly due to the presence of higher concentrations of nutrients in shallow sediments.
It is expected that, in most cases, the forces that determine the distribution of bacteria in
aquifers also determine the distribution of archaeal species.

4.2.2. Identity of subsurface archaea

14

spec.

 

u “‘1‘

.1.“

In their study of a contaminated aquifer at the former Wurtsmith Air Force Base near
Oscoda, Michigan, Dojka et al. (18) cloned archaeal 16S rRNA sequences from
anaerobic, jet-fire] contaminated sediments using both archaeal and universal primer sets.
They found archaeal sequences to be related to known methanogens and to an unknown
Crenarchaeote. They propose that aceticlastic methanogens detected at the site and a
known H2-producer, Syntrophus, also detected, may be involved in a syntrophic
association in which H2 is passed from Syntrophus to the methanogens. Chandler et a1.
(11) derived an archaeal 16S rRNA clone library from a deep subsurface sediment
extracted from beneath the Hanford Site in Washington state. The sediment had a
relatively high Eh (300 mV) and a high concentration of organic carbon (0.7% wt:wt).
The sequences were all most closely related to thermophilic Crenarchaea. The authors
speculate that these archaea may oxidize sulfur as their primary mode of energy-
conservation. Using taxon-specific oligonuceotide probes, Fry et al. (26) determined that
archaeal rRNA comprised 1.8% and 2.5% of the total rRNA in two deep, anaerobic,
alkaline aquifer sediments. They assert that these populations are likely to be
methanogenic since previous enrichments at the site uncovered methanogenic archaea.
Finally, Stevens and McKinley (54) have uncovered a subsurface environment that may
be dominated by lithoautotrophic methanogenesis. Using stable carbon isotope ratio
measurements, they determined that hydrogen of lithological origins was being
incorporated into methane, most likely through the metabolism of archaeal methanogens.
It is worth noting that two studies exist (40, 46) in which researchers have used specific
tools to detect Archaea in the subsurface (cultivation and rRNA analysis, respectively)

but failed to detect the presence of this group.

15

Di:

 

 

4. 3. Bacteria
4.3.1. Patterns of abundance

Data on the distribution of bacteria in the saturated subsurface is somewhat more
extensive than data on Eukarya or Archaea, but the picture is still incomplete. In studies
that measured direct microscopic counts and plate counts in pristine subsurface sediment
(9, 10, 40, 49, 59), bacterial numbers have been found to be lower than those observed in
the overlying soil. It is largely agreed that the relatively lower concentrations of organic
matter and nutrients in sediments below the soil horizons is the largest determining factor
for this observation (27). In sediment from a buried coal tar site, Madsen et al. (39)
found that chemical contamination resulted in increased numbers of viable bacteria, a
phenomenon observed in microbial communities from other environments as well (16,
19). There is conflicting data on the relative numbers of bacteria in unsaturated and
saturated subsurface sediments and the results are apparently not correlated to any
measured parameter. For example, Beloin et al. (9) found higher levels of microbial
biomass indicators in saturated sediment than in the upper zones but Kieft et al. (34)
found greater relative numbers of the same or similar indicators in unsaturated sediment
at their study site in Washington state. Other studies have failed to uncover a relationship
between saturation and bacterial counts (5, 10).

Sediment texture (and hence, permeability) may be a more important influence on
I'elative numbers of bacteria than degree of saturation. Albrechtsen (2) investigated the
l'€=lative numbers of viable bacteria in the various size fractions of a single aquifer
Sediment. It was found that although viable numbers are inversely proportional to grain

Size. a size cut-off exists below which numbers of viable bacteria drop off: the size

16

fraction 0.2-1.2 urn held only 0.01-0.04% of the total numbers of bacteria found in the
whole sediment. The highest numbers of viable bacteria per gram of sediment were
found in the 1.2-100 um fraction. Fredrickson et al. (23) uncovered the same
phenomenon in interleaved layers of sandstone and shale, but expressed their findings in
terms of the pore-throat size of the sediments examined. They observed a total
elimination of metabolic activity (as measured by 14C-labeled substrate uptake) in layers
with pore throat sizes below 0.2 um. They suggest that while inactive cells may be
maintained within sediment with low pore sizes, pore throats above 0.2 um are required
for sufficient nutrient flux to maintain active metabolism.

It is generally accepted in the field that the vast majority of bacteria in the saturated
subsurface are attached to the substratum, rather than suspended in the interstitial space
(3, 8), and further investigation has shown that sediment communities are distinct from
those of the surrounding groundwater (3, 35). These data suggests that groundwater
sampling does not serve as an appropriate surrogate for the more expensive and labor-
intensive sampling of subsurface sediment.

4.3.2. Identity of bacteria

Bacteria and 168 sequences previously isolated from the subsurface fall into these
phylogenetic divisions: Proteobacteria, High G+C Gram positive bacteria, Low G+C
Gram positive bacteria, candidate divisions OPS, 0P8, 0P10, and DP] 1,
Cytophaga/Flexibacter/Bacterioides, Acidobacteria, Spirochetes, Termite group I,
Verrucomicrobium, and Green nonsulfur. Figure 1 represents these groups in the context
of the currently accepted bacterial divisions and proposed divisions (32) and Table l is a

lists of these organisms and the most closely related genera, where determined.

17

    

§
'3 o g
e
2, a: as 3
o O 5 ' e e
0 r6 ~4 3 g 8‘ 4%)
~15 63 ga- 'o. e 9
” o~
“is 5" fire .29
. 0‘ ‘°
9. \ sad“ e‘"
I!» "a m ' 43°65? ‘9"? W
i, if. Cfi Cgfo’ Q
2 gm
// «0x
," ' w x/ ’ w FibIObac‘ct
.afii’ .

Green SUIfur
. . ...-. ~WRE55$%§W5534>3M

. “--..\ Cytophagales
4:; MIR-‘22:... hemNISID '
1E;- \ CUS

‘ flir-
, ._ 005919.?

Archaea 0'10

Figure 1 Radial phylogram of selected bacterial divisions. Shaded wedges
indicate divisions for which representative isolates or 168 sequences have
been recovered from the subsurface. Adapted from Hugenholtz et al. (32).

18

a.
:.

l.

Q Q Q ‘3 Q Q ‘3
‘rl ’yV ‘-_y ‘—< ’v-, ‘~. ’-

Nil.

07%; _
6a,!
5.1),, ,'
ND)“
,7 Play,
71%;;
7pm; I
715.7);

‘15 ,
.1”),

 

 

Table 1 Bacterial isolates and 16S clones previously derived from the subsurface. The
division (or putative division) and genera are listed if determined for a given isolate or
sequence. All isolates and clones were phylogenetically identified by sequencing of the
16S gene, except isolates developed by Fries et a1 (24), which were identified through
fatty acid methyl ester (FAME) analysis.

 

Division or putative Genera (if isolate or Method of
division determined) clone analysis reference
a Proteobacteria Erythromicrobium 16S clone sequencing (13)
a Proteobacteria Methylobacterium isolate FAME (24)
aProteobacteria Sphingomonas isolate sequencing (6)
a Proteobacteria Sphingomonas isolate sequencing (21)
a Proteobacteria Sphingomonas l6S clone sequencing (53)
a Proteobacteria 16S clone sequencing (18)
aProteobacteria 16S clone sequencing (18)
flProteobacteria Azoarcus 16S clone sequencing (18)
,6 Proteobacteria Burkholderia 16S clone sequencing (13)
fl Proteobacteria Burkholderia 16S clone sequencing (53)
fl Proteobacteria Burkholderia isolate sequencing (62)
,8 Proteobacteria Comamonas isolate sequencing (6)
,B Proteobacteria Comamonas l6S clone sequencing (13)
,6 Proteobacteria Comamonas isolate FAME (24)
,6 Proteobacteria Duganella l6S clone sequencing (l8)
,6 Proteobacteria H ydrogenophaga isolate FAME (24)
,6 Proteobacteria Janthinobacterium isolate FAME (24)
fl Proteobacteria Ralstonia l6S clone sequencing (53)
fl Proteobacteria Variovorax isolate sequencing (6)
fl Proteobacteria Variovorax isolate FAME (24)
[3 Proteobacteria 16S clone sequencing (18)
6 Proteobacteria Desulfovibrio 16S clone sequencing (41)
5 Proteobacteria Syntrophus 16S clone sequencing (18)
6 Proteobacteria 16S clone sequencing (18)
7 Proteobacteria Acinetobacter isolate sequencing (6)
7 Proteobacteria Acinetobacter isolate FAME (24)
Y Proteobacteria Acinetobacter l6S clone sequencing (41)
1’ Proteobacteria Actinobacillus isolate FAME (24)
7 Proteobacteria Legionella 16S clone sequencing (18)
YProteobacteria Methylomonas isolate sequencing (33)

19

Table 1 Continued

 

Division or putative Genera (if isolate or Method of
division determined) ‘ clone analysis reference
7Proteobacteria Shewanella 16S clone sequencing (41)
7Proteobacteria Thiamicrospira 16S clone sequencing (41)
Proteobacteria Pseudomonas isolate sequencing (6)
Proteobacteria Pseudomonas 16S clone sequencing (1 3)
Proteobacteria Pseudomonas isolate FAME (24)
Proteobacteria Pseudomonas l6S clone sequencing (53)
Hi G+C gram positive Arthrobacter isolate sequencing (6)
Hi G+C gram positive Clavibacter 16S clone sequencing (l 1)
Hi G+C gram positive Corynebacterium isolate FAME (24)
Hi G+C gram positive Micrococcus 16S clone sequencing (13)
Hi G+C gram positive Micrococcus isolate FAME (24)
Hi G+C gram positive Nocardia 16S clone sequencing (13)
Hi G+C gram positive Nocardia isolate FAME (24)
Low G+C gram positive Bacillus isolate sequencing (6)
Low G+C gram positive Bacillus 16S clone sequencing (13)
Low G+C gram positive Bacillus isolate FAME (24)
Low G+C gram positive Bacillus 16S clone sequencing (41)
Low G+C gram positive Bacillus 16S clone sequencing (53)
Low G+C gram positive Desulfosporosinus isolate sequencing (44)
Low G+C gram positive Desulfotomaculum 16S clone sequencing (18)
Low G+C gram positive Eubacterium l6S clone sequencing (41)
Low G+C gram positive Exiguobacterium l6S clone sequencing (18)
Low G+C gram positive Staphylococcus isolate FAME (24)
Low G+C gram positive Streptococcus isolate sequencing (6)
Low G+C gram positive 165 clone sequencing (l8)
OPIO l6S clone sequencing (18)
CPU 168 clone sequencing (18)
OPS 16S clone sequencing (18)
0P8 l6S clone sequencing (18)
Cytophaga/Flexibacter/
Bacterioides F lavobacter l6S clone sequencing (53)
Cytophaga/Flexibacter/
Bacterioides 16S clone sequencing (18)
CyIOphaga/Flexibacter/
Bacterioides Chryseobacterium 16S clone sequencing (18)

20

Table 1 Continued

 

Division or putative Genera (if isolate or Method of

division determined) clone analysis reference
Acidobacteria 16S clone sequencing (7)
Acidobacteria 16S clone sequencing (18)
Spirochetes 16S clone sequencing (18)
Termite group I 168 clone sequencing (18)
Verrucomicrobium 16S clone sequencing (18)
Green nonsulfizr l6S clone sequencing (18)

21

 

 

 

“1*—

H”...

Nina.

iiLg~
y

3‘3»,
“\14‘“'

3:09,)“,
A n

5. Methods for analysis of microbial communities
There are a number of methods available to study the microbial communities of the
saturated subsurface. Methods of community analysis are generally divided between
those which require cultivation and those which do not. Each method has its strong
points and its limits. In deciding which method to use, a researcher has to balance these
pros and cons and arrive at a choice that targets his or her objective the best. Some of the
more commonly used approaches are listed here.

5.1. Cultivation and characterization of isolates
The cultivation of subsurface microbes offers one route to describing the diversity of the
community. Isolates can be metabolically characterized to determine what
biogeochemical processes may be important in the ecosystem. Alternatively, the
substrate range of isolates may be determined to elucidate possible pathways present in
the system. Drawbacks include the possibility that isolates are poorly representative of
the metabolically active members of the community, since it has been reported that only
1 -IO% of environmental strains can be cultivated (52) and the fraction of culturable
bacteria in aquifer sediment can vary from sample to sample within the same formation
(5). Hence, isolates derived from the subsurface may, in fact, be representative of the
dominant metabolic groups present, but this has not been tested to date.

5.2. Analysis of functional genes
Testing for the presence of certain functional genes in the subsurface, whether they are
eXtracted in community DNA or recovered fiom the genomes of isolated organisms, is

another approach to understanding the metabolic features of the subsurface microbial

22

The
mitt)
field

it I

A I: ’
a,“
{.,.
4s“;
H ’t-\

flute

57.. -.
54¢

vu‘.

 

 

community. Functional genes can be used as indicators of certain processes, like
denitrification or toluene degradation for example, and the presence of these genes can be
tested in samples taken directly from the environment or from enrichment cultures (30).
One drawback to using this method is that in order to detect a given firnctional gene it
must have been previously identified and characterized, which can be laborious. Another
point of concern lies in the fact that detection of a given gene does not necessarily mean
that it is expressed under the current conditions. Furthermore, metabolic processes that
are not specifically targeted will not be detected, so pathways of importance may be
missed in this type of analysis.

5.3. Analysis of the metabolism of radio-labelled substrates or electron acceptors
The use of radioactively-labelled substrates or electron acceptors can be used to
investigate the metabolism of subsurface microbial communities on a bench scale or a
field scale. For example, Fredrickson et al. (23) used [14C] acetate, [14C] glucose, and
358042' in sediment microcosm studies to determine the viability and the sulfate reduction
capacity of the subsurface community. One drawback to using labeled substrates in
microcosm experiments is that one can only confirm the potential for a given metabolic
process, since in situ rates of a given reaction, given ambient concentrations of the
substance of interest, are not measured.

5.4. Analysis of membrane components — PLFA and FAME
Fatty acids and phospholipids found in the membranes of all microbes can be employed
in analyzing the composition of the subsurface community. In phospholipid fatty acid
(PLF A) analysis and fatty acid methyl ester (FAME) analysis, the lipids in a given

sediment are extracted and the polar lipids are separated from the mixture and analyzed

23

361‘;

'II

in
My

|
ya...

I“;

one:

Fun}...

..
‘ u‘¥ :

ii: I

misfit

 

- A
1’.

YnL;_"
*‘LLE s

IGOR:

 

(PLF A) or the lipids are esterified and analyzed (FAME). Signature molecules have been
identified that can indicate the presence of certain microbial groups (25). However, this
method provides only a low degree of resolution and allows the researcher to determine
the relative abundances of only those groups that have been cultivated, characterized, and
have distinct membrane lipids. Also, sample size considerations may be prohibiting, as
cell counts in subsurface sediment are often so low that large quantities of sediment are
needed for the extraction (>75 g) which may be expensive to acquire.
5.5. Analysis of small subunit rRNA or small subunit rDNA

The small subunit (SSU) rRNA gene and its transcripts, l6S rRNA in bacteria and
archaea and 18S rRN A in eukaryotes, are employed in many different methods of
community analysis. The SSU rRNA molecule serves in the ribosome as part of the
machinery of translation and its function has been found to be highly conserved among
all living organisms. There are a number of advantages to using 168 and 188 over other
molecular markers. Firstly, the SSU gene is present in all organisms. This means that
having the 168 sequence of a given organism allows researchers to place that organism
within the context of the universal tree of life. Also, some parts of the molecule are
highly conserved, so large evolutionary distances between kingdoms can be resolved, and
other parts are less conserved, allowing for finer resolution of genera and species.
Furthermore, the 168 molecule holds a great deal of phylogenetic information: each of
the 1500+ nucleotide bases has four possible identities (A, G, C, U). The 168 gene is not
translated into protein, so its sequence is not subject to the restrictions and complications
which translation into amino acids imposes. Finally, there are extensive databases of

known 16S sequences available (including the Ribosomal Database Project

24

 

 

i43nc

ins...

$1)ij

l' 5,“.
O4- I‘k‘a.

(II

E art:

IRS

l r .
u gE

fey-4;,
“will

1
eiffffnn.

the {536

http://www.cme.msu.edu/RDP/html/. and GenBank, available through the National
Center for Biotechnology Information http://www.ncbi.nlm.nih.gov/) which can be used
for comparing or identifying newly discovered 16S sequences.

There are drawbacks to the use of the 16S gene in community analysis. Biases may be
introduced by the polymerase chain reaction (PCR), which is often used to amplify the
community 168 sequences. Artifacts of the reaction may include preferential
amplification of minority sequences, fidelity problems in polymerization, chimeric
sequences, and founder effects (12, 51, 56, 57). Furthermore, 168 results may be difficult
to extrapolate to the environment because of lack of agreement between phylogenetic
identity and the metabolism of the organism (1). This is certainly a consideration when
employing 168 methods, but many phylogenetic groups (species, genera, and divisions)
have been shown to have consistent metabolic traits, allowing conjecture about the
processes that closely related organisms undertake (18). Also, 16S rDNA extracted from
environmental samples (soil, sediment, or water, for example) may have been derived
from dead or inactive cells, as it is unclear exactly how long DNA remains stable in the
environment. Each of these limitations must be acknowledged when using the 16S gene
in community analysis.

5.5.1. Amplified ribosomal DNA restriction analysis

In amplified ribosomal DNA restriction analysis (ARDRA) of microbial communities,
168 genes in community DNA are amplified through the PCR and digested with a
restriction enzyme. The resulting mixture of long and short fragments are separated by
electrophoresis to produce a profile of the 168 genes from the sample. ARDRA allows

the researcher to compare microbial communities from sample to sample, possibly to

25

Ar..-

kt.

C37

{HE'S
less

06(-

 

 

 

monitor the presence or absence of certain ribotypes or to observe whether the microbial
community at large is affected by a particular event (25). One drawback of this method
is the difficulty in identifying individual members of the community in a given
community profile. A single 168 type creates an unknown number of bands, so a band
cannot be correlated, with certainty, with a given ribotype.
5.5.2. Denaturing gradient gel elctrophoresis analysis
In denaturing gradient gel electrophoresis, or DGGE, 168 genes fiom the community are
amplified through PCR and separated on an acrylamide gel. A chemical denaturant in the
gel causes the 168 genes to be separated according to their base pair sequence. This
method allows many of the same analyses afforded by ARDRA analysis and provides a
less complicated community profile with a one-band-to-one—ribotype ratio. However,
DGGE is subject to poor reproducibility due to the vagaries of assebling the gradient gel.
5.5.3. Fluorescent in situ hybridization analysis

Fluorescent in situ hybridization microscopy, or FISH, is a method of analysis in which
cells are washed from the sample (sediment, in this case) and are stained with
fluorescently-tagged rRNA probes. The stained cells are viewed with an epifluorescence
Inicroscope which allows the user to apply varying wavelengths of incident light to cause
fluorescence of the chemical tags attached to the rRNA probes. The result is a series of
images in which cells that are tagged with each probe used are highlighted. The
Specificity of the probes is the key: they can be designed to target groups from the
Kingdom level to the strain level. Using this method, one can determine the numbers and
relative proportions of different phylogenetic groups within a sample. One drawback of

FISH is that the only groups that are identified and enumerated in a given sample are

26

Urn:_
“k.

6825,"

.-.-~-

 

those which the researcher has specifically targeted. For example, a particular division
which comprises a large part of the community will not be enumerated if the researcher
doesn’t know in advance to target it and design a specific probe for it. Furthermore,
some groups lack a consistent signature sequence in the 168 gene, preventing the design
of a probe to specifically target them.
5.5.4. rRNA hybridization analysis
The hybridization techniques in FISH were derived from earlier work with rRNA
hybridization in which whole rRNA fiom the microbial community is extracted and
attached to a membrane. The community rRNA is exposed to radioactively labeled
phylogenetic rRNA probes, and the amount of hybridization of a probe to the community
RNA is quantified. The concentration of these probes that attach to a given community
sample is proportional to the number of target sequences in the sample. Using rRNA
hybridization, the relative proportions of different phylogenetic groups in the community
can be established. This method has the same drawbacks as FISH, namely, that
untargeted phylogenetic groups and groups for which a probe cannot be designed are not
detected.
5.5.5. Cloning the 168 gene

The 168 DNA from a sediment sample can be amplified (with the PCR), cloned into
Escherichia coli, and analyzed to determine the identities of individual community
members. This method has proven to be a powerfiil tool in the discovery of previously
unknown bacteria and archaea and in assessing the distribution of known species. The

drawbacks to using 168 in community analysis (as described above) clearly apply to the

27

 

 

p——«¢
.

rt-
. .

A f.n
Lt.

V:
(I;
. (' i

C)
”o,

1
’H ‘0
‘ A
n...‘

10 i:

 

 

 

RFU

h‘ -- .
fit-v1“;

I... ‘
.,
“15C".

t¢.\

p-
'31 :p‘
C“
In," .
. ‘ I
II‘I‘L:
t,
"h
u .. :-
“A“

use of cloning, but it has allowed many researchers to explore the variety of community
members in the subsurface.
5.5.6. Terminal restriction fragment length polymorphism analysis
In terminal restriction fragment length polymorphism (T-RFLP) analysis, the 168 genes
from a microbial community are amplified using a primer that is fluorescently labeled at
the 5’ end. The amplified genes are then digested with a restriction enzyme and size-
separated by acrylamide gel electrophoresis. The terminal 168 fragments are
fluorescently tagged and can be detected and compiled to create a profile of the microbial
community. The length of any particular terminal fragment is dependent upon the
sequence of the original 168 gene, and the 168 sequence is dependent upon the identity
of the organism from which it was derived. So, in addition to the comparative analysis
that ARDRA allows, T-RFLP profiles can be analyzed using databases of 16S sequences.
The 168 sequences in these databases can be digested with restriction enzymes in silico
to determine the terminal 168 restriction fragment lengths of known species. Peaks in T-
RFLP profiles can be correlated to the restriction fragment lengths of known species or
phylogenetic groups. T-RFLP analysis is subject to the limitations inherent to all 168
methods, but it can be a usefirl tool for comparative community analysis.
5.5.7. Real-time PCR of SSU genes or functional genes

Real-time PCR is a molecular method in which the numbers of a targeted DNA sequence
are quantified by comparing the progress of a reaction containing an unknown number of
target sequences with the progress of standard reactions. The method has been used to
quantify a range of DNA sequences, including nitrite reductase (28), glycoprotein D of

human immunodeficiency virus (31), meningococcal-specific genes IS1106, ctrA, and

28

siaD (29), and large (37) and small subunit rRNA (17, 37, 42, 58). The method has great
potential for investigating microbial community structure through quantifying SSU genes
and fiinctional genes in environmental samples, but it is not yet widely used for this
purpose. The drawbacks of analyzing the distribution of functional genes and SSU genes
apply to real-time PCR.

6. Specific aims of this work

To achieve a better understanding of subsurface microbial communities, we have
undertaken a comparative analysis of the communities in pristine and chemically-
contarninated aquifer sediment by dissecting the communities of two different sites using
molecular, cultivation-independent approaches.

The specific aims of these projects were as follows:

1. Determine the community structure in several samples of pristine aquifer
sediment using T-RFLP of PCR amplified l6S rRNA genes.

2. Describe members of the bacterial community in a pristine subsurface sediment
by cloning and sequencing selected 168 sequences.

3. To use T-RFLP to assess the community-level differences between contaminated
and uncontaminated sediment sediments derived from the Bachman aquifer.
Intrinsic bioremediation of tetrachloroethylene (TCE) has been shown to occur
within the plume at this site.

4. To describe members of bacterial communities in these clean and contaminated
sediments by cloning 168 genes fi'om the site.

5. To elucidate the relative abundance of a 16S clone which has been recovered

from contaminated sediment at the Bachman site using real-time PCR.

29

10.

ll.

12.

13.

14.

REFERENCES

Achenbach, L. A., and J. D. Coates. 2000. Disparity between bacterial
phylogeny and physiology. ASM News. 66(December):p. 714.

Albrechtsen, H.-J. 1994. Distribution of bacteria, estimated by a viable count
method, and heterotrophic activity in different size fractions of aquifer sediment.
Geomicrobiology Journal. 12(4):253-264.

Alfreider, A., M. Krossbacher, and R. Psenner. 1997. Groundwater samples do
not reflect bacterial densities and activity in subsurface systems. Water Research.
31(4):832-840.

Alley, W. A., Thomas E. Reilly, and O. L. Franke. 1999. Sustainability of
ground-water resources, US. Geological Survey Circular 1186, US. Geological
Survey.

Balkwill, D. L., and W. C. Ghiorse. 1985. Characterization of subsurface
bacteria associated with 2 shallow aquifers in Oklahoma (USA). Applied and
Environmental Microbiology. 50(3):580-588.

Balkwill, D. L., R. H. Reeves, G. R. Drake, J. Y. Reeves, F. H. Crocker, M. B.
King, and D. R. Boone. 1997. Phylogenetic characterization of bacteria in the
subsurface microbial culture collection. FEMS Microbiol Rev. 20(3 -4):201-l6.
Barns, S. M., S. L. Takala, and C. R. Kuske. 1999. Wide distribution and
diversity of members of the bacterial kingdom Acidobacterium in the
environment. Applied and Environmental Microbiology. April, 1999;.
65(4):]731-1737.

Bekins, B. A., E. M. Godsy, and E. Warren. 1999. Distribution of microbial
physiologic types in an aquifer contaminated by crude oil. Microbial-Ecology.
May, 1999;. 37(4):263-275.

Beloin, R. M., J. L. Sinclair, and W. C. Ghiorse. 1988. Distribution and activity
of microorganisms in subsurface sediments of a pristine study site in Oklahoma
(USA). Microbial Ecology. 16(1):85-98.

Bone, T. L., and D. L. Balkwill. 1988. Morphological and cultural comparison of
microorganisms in surface soil and subsurface sediments at a pristine study site in
Oklahoma (U SA). Microbial Ecology. l6(1):49-64.'

Chandler, D. P., F. J. Brockman, T. J. Bailey, and J. K. Fredrickson. 1998.
Phylogenetic diversity of archaea and bacteria in a deep subsurface paleosol.
Microbial Ecology. July Aug. 36(1):37-50.

Chandler, D. P., J. K. Fredrickson, and F. J. Brockman. 1997. Effect of PCR
template concentration on the composition and distribution of total community
168 rDNA clone libraries. Molecular Ecology. 6(5):475-482.

Chandler, D. P., S. M. Li, C. M. Spadoni, G. R. Drake, D. L. Balkwill, J. K.
Fredrickson, and F. J. Brockman. 1997. A molecular comparison of culturable
aerobic heterotrophic bacteria and 168 rDNA clones derived fiom a deep
subsurface sediment. FEMS Microbiology Ecology. 23(2): 131-144.

Chapelle, F. H. 1993. Ground-Water Microbiology and Geochemistry. John
Wiley & Sons, Inc., New York.

30

 

‘7“
file
an
‘5‘!
:3
‘5

.14

 

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

Chapelle, F. H. 1985. Hydrogeology, digital solute-transport simulation, and
geochemistry of the Lower Cretaceous aquifer system near Baltimore, Maryland,
Maryland Geological Survey.

Cho, J.-C., and S.-J. Kim. 2000. Increase in bacterial community diversity in
subsurface aquifers receiving livestock wastewater input. Applied-and-
Environmental-Microbiology. March, 2000;. 66(3):956-965.

Corless, C. E., M. Guiver, R. Borrow, V. Edwards Jones, E. B. Kaczmarski,
and A. J. Fox. 2000. Contamination and sensitivity issues with a real-time
universal 16S rRNA PCR. Joumal-of-Clinical-Microbiology. [print] May, 2000;.
38(5):1747-1752.

Dojka, M., A., P. Hugenholtz, S. Haack, K., and N. Pace, R. 1998. Microbial
diversity in a hydrocarbon- and chlorinated-solvent-contaminated aquifer
undergoing intrinsic bioremediation. Applied and Environmental Microbiology.
Oct. 64(10):3869-3877.

Eloranta, P., and Kettunen, R. 1979. Phytoplankton [algae] in a watercourse
polluted by a sulphite cellulose factory. Ann Bot Fenn. Helsinki, Finnish
Biological Society Vanamo:338-350.

Fenchel, T. 1987. Ecology of protozoa: the biology of flee-living phagotrophic
protists. Science Tech Publishers, Madison, Wisconsin.

Fredrickson, J. K., D. L. Balkwill, M. F. Romine, and T. Shi. 1999. Ecology,
physiology, and phylogeny of deep subsurface Sphingomonas sp. Joumal-of-
Industrial-Microbiology-and-Biotechnology. Oct.-Nov., 1999;. 23(4-5):273-283.
Fredrickson, J. K., T. R. Garland, R. J. Hicks, J. M. Thomas, S. W. Li, and
K. M. McFadden. 1989. Lithotrophic and heterotrophic bacteria in deep
subsurface sediments and their relations to sediment properties.
GEOMICROBIOLOGY JOURNAL. 7(1-2):53-66.

Fredrickson, J. K., J. P. McKinley, B. N. Bjornstad, P. E. Long, D. B.
Ringelberg, D. C. White, L. R. Krumholz, J. M. Sufiita, F. S. Colwell, R. M.
Lehman, T. J. Phelps, and T. C. Onstott. 1997. Pore-size constraints on the
activity and survival of subsurface bacteria in a late Cretaceous shale-sandstone
sequence, northwestern New Mexico. Geomicrobiology Journal. 14(3): 183-202.
Fries, M., R., L. Forney, J., and J. Tiedje, M. 1997. Phenol- and toluene-
degrading microbial populations from an aquifer in which successfirl
trichloroethene cometabolism occurred. Applied and Environmental
Microbiology. 63(4): 1523-1530.

Fries, M., R., G. Hopkins, D., P. McCarty, L., L. Forney, J., and J. Tiedje, M.
1997. Microbial succession during a field evaluation of phenol and toluene as the
primary substrates for trichloroethene cometabolism. Applied and Environmental
Microbiology. 63(4): 15 15-1522.

Fry, N. K., J. K. Fredrickson, S. Fishbain, M. Wagner, and D. A. Stahl. 1997.
Population structure . of microbial communities associated with two deep,
anaerobic, alkaline aquifers. Applied and Environmental Microbiology.
63(4): 1498-1504.

Ghiorse, W. C., and Wilson, J.T. 1988. Microbial Ecology of the Terrestrial
Subsurface. Advances in Applied Microbiology. 33: 107-172.

31

 

. .x
i)

n a
.13

«.3

1.
1‘4

.‘v
‘3

 

~/
SJ

C1,

‘3

40

28.

29.

30.
31.
32.

33.

34.

35.
36.
37.

38.

39.

40.

Gruntzig, V., S. C. Nold, J. Zhou, and J. M. Tiedje. 2001. Pseudomonas
stutzeri Nitrite Reductase Gene Abundance in Environmental Samples Measured
by Real-Time PCR. Appl. Environ. Microbiol. 67(2):760-768.

Guiver, M., R. Borrow, J. Marsh, S. J. Gray, E. B. Kaczmarski, D. Howells,
P. Boseley, and A. J. Fox. 2000. Evaluation of the Applied Biosystems
automated Taqman polymerase chain reaction system for the detection of
meningococcal DNA. FEMS-Immunology—and-Medical-Microbiology. [print]
June, 2000;. 28(2): 173-179.

Guo, c., Sun, W., Harsh, J.B., Ogram, A. 1997. Hybridization analysis of
microbial DNA fi'om fiiel oil-contaminated and noncontaminated soil. Microbial
Ecology(34): 178-187.

Higuchi, R., C. Fockler, G. Dollinger, and R. Watson. 1993. Kinetic PCR
Analysis: real-time monitoring of DNA amplification reactions. Bio/Technology.
l 1(September): 1026- 1030.

Hugenholtz, P., M. Goebel Brett, and R. Pace Norman. 1998. Impact of
culture-independent studies on the emerging phylogenetic view of bacterial
diversity. Journal of Bacteriology. Sept. 180(18):4765-4774.

Kalyuzhnaya, M. G., V. N. Khmelenina, S. Kotelnikova, L. Holmquist, K.
Pedersen, and Y. A. Trotsenko. 1999. Methylomonas scandinavica sp. nov., a
new methanotrophic psychrotrophic bacterium isolated from deep igneous rock
ground water of Sweden. Systematic-and-Applied-Microbiology. Dec, 1999;.
22(4):565-572.

Kieft, T., L., J. Fredrickson, K., J. McKinley, P., B. Bj ornstad, N., S. Rawson,
A., T. Phelps, J., F. Brockman, J., and S. Pfiffner, M. 1995. Microbiological
Comparisons within and across Contiguous Lacustrine, Paleosol, and Fluvial
Subsurface Sediments. Applied and Environmental Microbiology. 61(2):749-757.
Koelbel Boelke, J., B. T ienken, and A. Nehrkorn. 1988. Microbial communities
in the saturated groundwater environment: 1. Methods of isolation and
characterization of heterotrophic bacteria. Microbial Ecology. 16(1): 17-30.
Ludvigsen, L., H. J. Albrechtsen, D. B. Ringelberg, F. Ekelund, and T. H.
Christensen. 1999. Distribution and composition of microbial populations in a
landfill leachate contaminated aquifer (Grindsted, Denmark). Microbial-Ecology.
April, 1999;. 37(3):197-207.

Lyons, S. R., A. L. Griffen, and E. J. Leys. 2000. Quantitative real-time PCR
for Porphyromonas gingivalis and total bacteria. Journal-of-Clinical-
Microbiology. [print] June, 2000;. 38(6):2362-2365.

Madsen, E. L., and W. C. Ghiorse. 1993. Groundwater microbiology:
subsurface ecosystem processes, p. 167-213. In T. E. Ford (ed.), Aquatic
Microbiology - An ecological approach. Blackwell Scientific Publications,
Boston.

Madsen, E. L., J. L. Sinclair, and W. C. Ghiorse. 1991. In situ biodegradation:
Microbiological patterns in a contaminated aquifer. Science. 252(5007):830-833.
Martino, D. P., E. L. Grossman, G. A. Ulrich, K. C. Burger, J. L.
Schlichenmeyer, J. M. Sufiita, and J. W. Ammerman. 1998. Microbial
abundance and activity in a low-conductivity aquifer system in east-central Texas.
Microbial Ecology. May June. 35(3):224-234.

32

 

4.

 

. t
. I

.4.
m

5].

{‘1

44'.

 

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

S3.

54.

Pedersen, K., J. Arlinger, S. Ekendahl, and L. Hallbeck. 1996. 168 rRNA
gene diversity of attached and unattached bacteria in boreholes along the access
tunnel to the Aspo hard rock laboratory, Sweden. FEMS Microbiology Ecology.
l9(4):249-262.

Pusterla, N., J. B. Huder, C. M. Leutenegger, U. Braun, J. E. Madigan, and
H. Lutz. 1999. Quantitative real-time PCR for detection of members of the
Ehrlichia phagocytophila genogroup in host animals and Ixodes ricinus ticks.
Joumal-of-Clinical-Microbiology. May, 1999;. 37(5): 1329-1331.

Robertson, L. A., and J. G. Kuenen. 1983. T hiosphaera pantotropha, new genus
new species a facultatively anaerobic, facultatively autotrophic sulfirr bacterium.
Journal of General Microbiology. 129(9):2847-2856.

Robertson, W. J., P. D. Franzmann, and B. J. Mee. 2000. Spore-forming,
Desulfosporosinus-like sulphate-reducing bacteria from a shallow aquifer
contaminated with gasolene. Joumal-of-Applied-Microbiology. Feb., 2000;.
88(2):248-259.

Schultze-Lam, S., and T. Beveridge. 1994. Nucleation of Celestite and
Strontianite on a Cyanobacterial S-Layer. Appl. Environ. Microbiol. 60(2):447-
453.

Shi, Y., M. D. Zwolinski, M. E. Schreiber, J. M. Bahr, G. W. Sewell, and W.
J. Hickey. 1999. Molecular analysis of microbial community structures in pristine
and contaminated aquifers: Field and laboratory microcosm experiments.
Applied-and-Environmental-Microbiology. May, 1999;. 65(5):2143-2150.
Sinclair, J., L., D. Kampbell, H., M. Cook, L., and J. Wilson, T. 1993.
Protozoa in subsurface sediments from sites contaminated with aviation gasoline
or jet fuel. Applied and Environmental Microbiology. 59(2):467-472.

Sinclair, J. L., and W. C. Ghiorse. 1987. Distribution of protozoa in subsurface
sediments of a pristine groundwater study site in Oklahoma (USA). Applied and
Environmental Microbiology. 53(5): 1 157-1 163.

Sinclair, J. L., S. J. Randtke, J. E. Denne, L. R. Hathaway, and W. C.
Ghiorse. 1990. Survey of microbial populations in buried-valley aquifer
sediments from northeastern Kansas (USA). Ground Water. 28(3):369-3 77.
Solley, W. B. R. R. P. H. A. P. 1998. Estimated use of water in the United States
in 1995, US. Geological Survey Circular 1200, US. Government Printing Office.
Speksnijder, A. G., G. A. Kowalchuk, S. De Jong, E. Kline, J. R. Stephen,
and H. J. Laanbroek. 2001. Microvariation artifacts introduced by PCR and
cloning of closely related l6S rRNA gene sequences. Appl Environ
Microbiol(67):469-472.

Staley, J. T., and A. Konopka. 1985. Measurement of in situ activities of
nonphotosynthetic microorganisms in aquatic and terrestrial habitats. Annu Rev
Microbiol. 39:321-46.

Stapleton, R. D., N. G. Bright, and G. S. Sayler. 2000. Catabolic and genetic
diversity of degradative bacteria from firel-hydrocarbon contaminated aquifers.
Microbial-Ecology. [print] April, 2000;. 39(3):21 1-221.

Stevens, T. 0., and J. P. McKinley. 1995. Lithoautotrophic microbial
ecosystems in deep Basalt aquifers. Science. 270(5235):450-454.

33

 

 

(A

 

55.

56.

57.

58.

59.

60.

61.

62.

Strauss, E. A., and W. K. Dodds. 1997. Influence of protozoa and nutrient
availability on nitrification rates in subsurface sediments. Microbial Ecology.
34(2):]55-165.

Suzuki, M., T., and S. Giovannoni, J. 1996. Bias caused by template annealing
in the amplification of mixtures of 168 rRNA genes by PCR. Applied and
Environmental Microbiology. 62(2):625-630.

von Wintzingerode, F., U. B. Goebel, and E. Stackebrandt. 1997.
Determination of microbial diversity in environmental samples: pitfalls of PCR-
based rRNA analysis. FEMS Microbiology Reviews. 21:213-229.

Weller, S. A., J. G. Elphinstone, N. C. Smith, N. Boonham, and D. E. Stead.
2000. Detection of Ralstonia solanacearum strains with a quantitative, multiplex,
real-time, fluorogenic PCR (TaqMan) assay. Applied-and-Environmental-
Microbiology. [print] July, 2000;. 66(7):2853-2858.

White, D. C., Smith, G.A., Gehron, M.J., Parker, J.H., Findlay, R.H., Martz,
R.F., and Fredrickson, EL. 1983. The groundwater aquifer microbiota:
Biomass, community structure, and nutritional status. Developments in Industrial
Microbiology. 24:201-21 1 .

Yates, K. K., and L. L. Robbins. 1995. Experimental evidence for a CaCO3
precipitation mechanism for marine Synechocystis. Bulletin de l'Institut
Oceanographique Monaco(Spec. Issue 14 Part 2):51-59.

Zarda, B., G. Mattison, A. Hess, D. Hahn, P. Hohener, and J. Zeyer. 1998.
Analysis of bacterial and protozoan communities in an aquifer contaminated with
monoaromatic hydrocarbons. FEMS Microbiology Ecology. Oct. 27(2): 141-152.
Zhang, H., S. Hanada, T. Shigematsu, K. Shibuya, Y. Kamagata, T.
Kanagawa, and R. Kurane. 2000. Burkholderia kururiensis sp. nov., a
trichloroethylene (TCE)-degrading bacterium isolated from an aquifer polluted
with TCE. International-Journal-of-Systematic-and-Evolutionary-Microbiology.
[print] March, 2000;. 50(6 part 2):743-749.

34

CHAPTER 2

A CULTIVATION-INDEPEN DENT ANALYSIS OF BACTERIAL
COMMUNITIES IN PRISTINE AQUIFER SEDIIVIEN T

INTRODUCTION

Groundwater is an important resource both economically and ecologically, and
the microbial communities that reside in the saturated subsurface play an integral role in
the quality of this resource. As groundwater filters through subsurface sediment,
microbes metabolize soluble organic and inorganic compounds, immobilizing or
mineralizing them and improving water purity (36). Subsurface microbes are also
responsible for the degradation of anthropogenic contaminants in groundwater (3, 31,
41).

To date, numerous studies of subsurface microbial communities have been carried
out, but by no means do we have a complete description of life in the saturated
subsurface. Early work has shown that life in saturated sediment is strictly microbial,
prokaryotic, and adapted to oligotrophic conditions. Most studies have examined the
cultivable fraction of the microbial community (6-8, 17, 19, 20, 23, 26, 27, 30, 38).
While these studies have provided a degree of insight into the physiology and numbers of
certain subsurface organisms, they are severely limited in the wider significance of their
findings because the culturable members do not necessarily reflect the entire community
or even a substantial fraction of the community (11, 16, 34, 39). These studies have
shown, however, that subsurface microbes represent a high degree of physiological and

morphological diversity (6, 7, 17, 23, 30, 38). Other investigations have contributed

35

information about abundance patterns in the subsurface, and they have begun to correlate
aquifer geochemistry and population abundance (6, 7, 17, 19, 20, 23, 30, 38). Specific
information on the phylogenetic groups and metabolic types present in uncontaminated
subsurface sediments is severley limited, however.

We have undertaken an analysis of the microbial communities in a pristine aquifer
on the eastern shore of Virginia, where sediments range fi'om aerobic and oligotrophic to
microoxic and relatively organic-rich. We have employed two nucleic acid-based
techniques, 16S rRNA T-RFLP and cloning, in an effort to compare and describe these
communities. While we observe similarities between the bacterial communities from the
two sites, there are also strong differences. Clones from the aerobic zone are
phylogenetically diverse and most show little homology with known species. We have
also detected two rDNA clones that show little relatedness to known phylogenetic
divisions and are deeply rooted in the bacterial phylogenetic tree.

MATERIALS AND METHODS
Sample collection The study site is located in Oyster, VA. Two areas of the aquifer
were sampled: the narrow channel focus area (N CFA) and the south Oyster focus area
(SOFA) (Figure l). A single intact core was aseptically collected from each sampling
point indicated in the diagram from the NCFA in October 1998 and from the SOFA in
April 1999. Cores were placed in plastic liners, divided, placed in gamma-irradiated
polypropylene tubes, and held at 4° C for seven days, after which they were stored at -
80°C. The sediment samples from the NCFA and SOFA were collected from 6.5 - 8.0

and 4.9 — 7.0 meters below the ground surface, respectively.

36

 

 

 

   
  

      
     
 
  
   

    

i
A

   
  
   

' ,. 1‘ 13:51,.”

Focus

   

 

 

 

 

 

South Oyster Focus Area (SOFA) o 2 4
Scale in meters
SO T2
. so S 14 .
SO 32 . Hydrological flow gradient NC M3
SO S 10 ——>
Narrow Channel Focus Area (NCFA) 0 2 4
Scale in meters
0 O 0
NC B2 NC 818 NC M3
Hydrological flow gradient
————>

 

 

Figure 1 Map of Oyster, Virginia, indicating the locations of the south Oyster
focus area (SOFA) and the narrow channel focus area (NCFA). Below, the areas
of the SOFA and the NCFA have been expanded to indicate the positions of
sampling points.

37

DNA extraction Microbial community DNA was extracted from NCFA sediment
samples by a protocol described by Zhou et al. (44) with modifications. Briefly, 15 ml
extraction buffer and proteinase K (final concentration 0.04 mg/ml) was added to 15 g of
sediment, and the mixture was incubated at 37°C for 30 minutes. Sodium dodecyl sulfate
(SDS) solution was added for a final concentration of 18 mg/ml, and the mixture was
incubated at 65°C for 2 hours, followed by three cycles of freezing in a dry ice and
ethanol bath and thawing in a 65°C water bath. The supernatant was collected and the
pellet was washed twice with extraction buffer and SDS and the washes and supernatant
were combined. This liquid was extracted twice with chloroform and isoamyl alcohol
(24:1 vol:vol) and the aqueous phases were combined. Nucleic acid was precipitated by
the addition of 0.6 volume of isopropanol, then resuspended in water and precipitated by
the addition of 2 volumes of ethanol. The pellet was resuspended in modified TE buffer
(10 mM Tris-HCL, 0.1 mM EDTA, pH 8) and used in polymerase chain reactions.

DNA was extracted from SOFA sediment samples with a Soil DNA Kit Mega
Prep (MoBio) according to the manufacturer’s instructions. We used 10 g of sediment
for each extraction. Isopropanol (0.7 volume) was added to the final DNA elution, and
the solution was incubated at 4°C overnight. The solution was centrifiiged at 12,000xg
for 30 min, the supernatant was decanted, and the pellet was resuspended in 200 ul of
water. The DNA was reprecipitated by addition of 0.15 volume of 3 M sodium acetate
and 2 volumes of ethanol. The solution was stored at -20°C overnight. Community
DNA was pelleted by centrifugation at 12,000xg for 30 min. The pellet was dried and

resuspended in 50 ul of water, and stored at -20°C until needed.

38

1';

C035

 

 

16S rDNA T-RFLP As sediment samples from the two sites were extracted by different
methods, the amount and purity of the genomic DNA in the final extraction volume
difl‘ered as well. Accordingly, the T-RFLP amplification mixture for each sample set was
adjusted independently.

NCFA sediment samples NC B2 (8 m), NC 518 (6 m), NC 818 (8 m), and NC M3
(6 m) were used for T-RFLP analyses. Bacterial 16S ribosomal genes were amplified
from bulk DNA in reaction mixtures that contained 1X PCR buffer (Perkin Elmer), 2 mM
MgC12, 0.2 mM of each dNTP, 0.2 mM of the reverse primer, 0.5 mM of the forward
primer, 8 ng bovine serum albumin per ul, approximately 0.08 ng/ul of template DNA,
and 0.02 U of AmpliTaq (Perkin Elmer) per ul.

SOFA sediment samples S0 B2 (5 m), SO BZ (6 m), SO T2 (5 m), SO T2 (6 m),
S0 814 (5 m), SO $14 (6 m), SO M3 (6 m), and SO 810 (4 m). For T-RFLP of the
SOFA DNA, bacterial 16S genes were amplified in the same manner as described for the
NCFA samples, with the following exceptions: the forward and reverse primer
concentrations were both 0.25 uM, the template concentration was 0.8 ng/ul and the
AmpliTaq concentration was 0.05 U/ul.

The forward primer for all reactions, 27F, which is specific for bacteria (5’-AGA
GTT TGA TCC TGG CTC AG-3’) (24), was labeled at the 5’ end with the
phosphoramidite dye 5-hexachlorofluorescein (“hex-labeled”). The reverse primer used
in SOFA amplifications was the universal primer 1392R (5’-ACG GGC GGT GTG TRC-
3’)(32). Amplifications of NCFA DNA contained either 1392R or the universal primer

152511 (sumo GAG GTG ATC CAG cc-3’) (5).

39

Reaction mixtures were incubated in a GeneAmp 2400 PCR System thermal
cycler (Perkin Elmer) at 94°C for 3 minutes, followed by 35 cycles at 94°C for 45
seconds, 60°C for 30 seconds, and 72°C for 90 seconds and a final extension step of 72°C
for 10 minutes.

We prepared three replicate amplifications of each NCFA DNA sample and two
replicates of each SOFA DNA sample. Replicate reaction mixtures were combined and
purified using WizardQ PCR purification columns (Promega) and eluted with a final
volume of 50 ul of modified TE buffer (0.1 mM EDTA, lOmM Tris, pH 8.0). For
restriction digests, 200 ng of purified PCR product were digested with 15 U of HhaI,
MspI, or RsaI (BMB) at 37°C for 3.5 hours. For electrophoresis, 1 ul of this mixture was
loaded on an acrylamide slab gel. The lengths of the terminal restriction fragments from
the amplified rDNA products were determined using an ABI Prism 377 DNA sequencer
and ABI software (PE Applied Biosystems) as described by Liu et al. (25). In order to
eliminate primer fragments from detection and to avoid inaccurate sizing of long
fragments, we excluded all fragments smaller than 30 bp and larger than 600 bp from the
analysis.

TRFLP analyses T-RFLP profiles were analyzed using GeneScan 3.1 software (PE
Applied Biosystems). For enumerating the restriction fragments in each profile, a
fluorescence intensity threshhold was set at 50 so that only fragments greater than this
intensity were included in further analysis. Comparisons between T-RFLP profiles were
conducted using the Genotyper software package (version 2.5, Perkin Elmer). Fragments
shorter than 30 base pairs (bp) or larger than 600 bp were excluded from analysis in order

to eliminate primer artifacts and to avoid the problems associated with identifying the

40

length of large fragments. The Hha I profiles were compared using the T-RFLP Profile
Analysis tool through the Ribosomal Database website

(http://www.cme.msu.edu/RDP/cgis/trflp.cgi?su=SSU), which determines the percent

 

similarity between two profiles by dividing the the number of terminal fragments present
in both profiles by the number of fragments in the profile with fewer fragments.

Cloning of 16S rRNA genes Sediment from the narrow channel site was selected for
16S rDNA cloning as this site represented a typical pristine, oligotrophic, aquifer. Three
sediments were used for cloning experiments: 818 (6 m), S18 (8 m), and B2 (8 m).
Bacterial 16S ribosomal genes were amplified from bulk DNA in reactions that contained
1X PCR buffer (Perkin Elmer), 2.5 mM MgC12, 200 M dNTPs, 0.2 mM of each primer,
8 ng bovine serum albumin per ul, and 0.02 U of AmpliTaq (Perkin Elmer) per ul. The
forward primer was 27F and the reverse primer was 1392R. The amount of template in
each amplification and the cycling conditions were the same as those used for T-RFLP
reactions. The PCR products were cloned using a TOPO TA Cloning Kit (Invitrogen
Corp.) in accordance with the manufacturers instructions. Plasmid DNA was extracted
and purified with a Qiagen Mini Plasmid-prep kit (Qiagen).

Screening of rDNA clones by ARDRA The plasmid inserts of approximately 40 clones
from each library were amplified using the PCR conditions described for cloning, with
roughly 30 ng of purified plasmid DNA template per 25 ul reaction. Five ul of rDNA
PCR products were digested with 10 U of the 4-base specific restriction enzyme Cfo I in
1X NEB bufl‘er (New England Biolabs) in a final volume of 15 ul for 3 hrs and 30 min at
37°C. Digested DNA fragments were separated by acrylamide gel (10% acrylamide)

electrophoresis and visualized by silver staining. Restriction fragment lengths were

41

 

 

 

estimated by comparison with molecular weight standards which ranged from eight to
587 bp (DNA molecular weight marker V2, Roche Molecular Biochemicals), and

plasmids with unique patterns were selected for partial sequencing of the rDNA insert.

Sequencing of rDNA clones Plasmid inserts from selected rDNA clones were amplified
using the same conditions described above for cloning. Amplified rDNA inserts were
purified by using Ultrafree MC Millipore 30,000 NMWL filter units (Millipore)
according to the manufacturer’s instructions. Sequencing was performed using the ABI
Prism BigDye Terminator Cycle Sequencing Reaction Kit and an ABI Prism 377 DNA
sequencer (PE Applied Biosystems) according to the manufacturer’s directions. Primers
for sequencing included 27F, 355F (5’-ACT CCT ACG GGR SGC AGC-3’) (4), 536R
(5’- GWA TTA CCG CGG CKG CT -3’) (33), 1100R (5’-AGG GTT GCG GTG GTT G-
3’) (ref. TLM?), and 1392R. Fifty-two (52) clones were partially sequenced using 27F
and five clones that were of particular phylogenetic interest were fiilly sequenced.

Phylogenetic analyses All sequenced clones were checked for chimeric sequences using
the Ribosomal Database Project II CHIMERA_CHECK version 2.7
(http://www.cme.msu.edu/RDP/html/index.html). Seven chimeras were detected among
the sequences and were eliminated from further analysis. Sequences were aligned against
close relatives in the Ribosomal Database Project 11 release 8.0 (28) using Arb software
(40). Alignments were refined by visual inspection and percent similarity to known
isolates and previously cloned sequences was determined. For analysis, a mask was
generated to exclude highly variable regions where the alignment was uncertain or where

one or more of the sequences had an alignment gap. All presented dendrograms were

42

"

 

E

 

constructed using the fastDNAml (maximum likelihood) function in the Arb software
package (40). The robustness of tree topologies was tested through loo-replicate
bootstrap resampling of sequences using evolutionary distance and the optimality criteria
of maximum likelihood and maximum parsimony (PAUP* version 4.0b8, written by
David L. Swofl‘ord). Evolutionary distance variables and corrections were selected

through the use of Modeltest (3 5).

RESULTS

Description of the Study Site The site is located in the village of Oyster, Virginia, on
the Delmarva Peninsula. The aquifer underlying Oyster is a shallow, unconfined
formation and is currently the focus of a study on bacterial transport by the US.
Department of Energy’s Natural and Accelerated Bioremediation Research Program.
Two areas were sampled for this study: the narrow channel focus area (NCFA) and the
south Oyster focus area (SOFA) (Figure 1). The NCFA lies to the south-southwest of the
village of Oyster. The water in this part of the formation was oxic (5.5 mg 02/1) and low
in organic carbon (1000 ppb). The SOFA lies directly to the south of Oyster. Here, the
groundwater was hypoxic (0.52 mg 02/1) due to upwelling suboxic groundwater and was
comparatively high in organic carbon (36,000 ppb). The water table at both sites was
roughly 1.5 to 2.0 meters below the ground surface.

168 rDNA T-RFLP We were able to identify up to 39 terminal fragments per T-RFLP
profile of Oyster sediment bacterial communities. On average, Hha I and Msp I
digestions produced more bands per profile (20 and 21, respectively), than did Rsa I (16).

See Table l for a list of the number of terminal fragments in each Hha I T-RFLP profile.

43

Table 1 Total number of terminal fragments detected in Hha I
bacterial T-RFLP profiles from Oyster, Virginia

 

 

number of

South Oyster sediment terminal fragments
S0 B2 (5 m) 21

SO T2 (4 m) 18

SO T2 (6 m) 15

SO S14 (5 m) 16

SO S14 (6 m) 23

SO S10 (4 m) 16

SO M3 (6 m) ' 8
Narrow channel sediment, primer set

NC B2 (6 m), 8F/1392R 29

NC 518 (6 m), 8F/1392R 18
NC S18 (8 m), 8F/1392R 25

NC M3 (6 m), 8F/1392R 20
NC B2 (6 m), 8F/1525R 22
NC 818 (6 m), 8F/1525R 39
NC S18 (8 m), 8F/1525R 13

NC M3 (6 m), 8F/1525R 14

44

Within the SOFA samples we can identify five profiles which are highly similar.
Each of the eight terminal fragments present in the SO M3 (6 m) profile are present
among the fragments in four other profiles: SO B2 (5 m), S0 814 (5 m), SO T2 (4 m) and
SO T2 (6 m). The Msp I and Rsa l digestions of these samples also display a high
degree of similarity (data not shown). Profiles SO T2 (6 m) and S0 814 (5 m) are also
highly analogous (93% identical fragments). Overall, each of the south Oyster profiles
showed no less than 40% identity with any of the other profiles.

Within the narrow channel samples, we note that the NC M3 (6 m) profiles are
markedly divergent from the other profiles. The 27F/1525R NC M3 (6 m) profile, in
particular, has a maximum of 28% similarity with other narrow channel profiles.

There was a degree of similarity between the SOFA and NCFA community
profiles. Terminal restriction fragments common to both sites included: Hha I 206 bp,
369 bp, 567 bp, and 573 bp, Msp I 494 bp, Rsa I 119 bp and 475 bp. Profiles from the
two sites shared as many as 50% of the same fragments.

Phylogenetic analysis The phylogenetic diversity of the bacterial community in the
narrow channel study site was determined through the analysis of three clone libraries of
16S ribosomal RNA genes. Figure 1 is a schematic diagram of the focus area and the
points from which sediment was extracted for use in the clone study. Bulk DNA was
extracted from aseptically-collected sediment samples and primers specific for bacterial
l6S sequences were employed to amplify the sequences of interest. In order to select
unique 16S clones for sequencing, 130 clones were screened using ARDRA (10). The
number of unique ARDRA patterns from each library was quite high: 26 out of 40

screened in the NC 818 (6 m) library (65%), 39 / 45 in the NC 818 (8 m) library (87%),

45

and 31 / 35 in the NC B2 (8 m) library (89%). There was a degree of overlap between
libraries, and we were able to discern 89 different ARDRA patterns among the 120
clones screened. We sequenced 15 clones from NC 818 (6 m), 4 clones from NC S18 (8
m), and 12 clones from NC BZ (8 m).

Table 2 lists the phylogenetic affiliations of the narrow channel focus area clones
as determined using the Arb software package with the RDP II 8.0 database and lists the
closest 168 sequence match in the GenBank database as determined by a BLAST search.
Figure 2 (a-e) is a series of maximum likelihood dendrograms depicting the phylogenetic
placement of 18 of these SSU rDNA clones. .

There are a number of clones that may be representatives of novel phylogenetic
divisions. Two clones, in particular, cluster together but fail to show a distinct
association with any known phylogenetic group (Figure 2e). These clones are apparently
deeply rooted in the bacterial tree and in Figure 2e they are shown in the context of other
deeply rooted genera and several later-diverging species. We will refer to these clones as
“NCFA group I”. None of the cloned sequences in this study are identical to known

species or previously discovered sequences.

DISCUSSION
Many studies of bacterial communities in uncontaminated subsurface
environments have failed to evaluate the phylogenetic diversity of these communities as
they have been limited by the bottleneck of cultivation, which prevents detection of
species that resist isolation. This can be a significant limitation when one considers that

90-99% of environmental species may be uncultivable (3 9) and that this number may be

46

. .nl 1 l\ Ila illit-
ll||||

at: guild! uticlt:lfirv ghv ~.h-—h

Nun‘s p...v.._ Kara-79:32.7. .fQ»— -~.-.h

U sup—.5 tyne uCQnCUU-ll 9.. 7v:

:7... v~>€A~

N av~r~ Enh.

 

«we 82 288 88888 55888: 2688 5% 6+0 E m S <52
m8 , w . 80.2 a 82:82 258 :88 96 E m : <82
:95
Rm 38.288822 58888: 358 :88 0+0 E m 2 <52
. V , q . V . ,_ 9.9m E one? 2380
E: 888.8 888085 82 .8588... 8588823 m m <82
. , . w, x 658883. \ 888082   g
Rm 8.88 88880 5588823 \ 888882 m w <82
_ , , 530883588, . . ; . . . , w . . .
., mwm 33 898006»: 19523:: , «EEBOEHE \ 8838.92 m A. «EUZ
mmm N218; 82888 52888: a58888.5 m m N 85.62
RN 8-2.8. 825888 8888: 588888 8 m 2.62
. , w . w . . on w . w , . .
cg -2835 8:588 5.2383 6_.88.8885 8 m v <52
.88 . . . 2.85 :58 . 88m at. 286 h a . .m <82
28 o mange «:88 882. a
Now 8:23 _ 8880882 .88 . 8 w 3:62
q .2825. . , ., .   j _
, 8. 8882...... s... 8838 _ as a as, I ,e . 882.22: . ; a ._ . .82
6.86m 033.8 £2550 «memo—o 838% w 283233. 48835 69 oc/no 00/80 0:20
o a S 25w 8 880 Pd J J
.52 8 . . _ 6 Q (600 @o.

33.5 $55: 820 E Eaten
42% .«o 3:08:80
EEwHS .8820 E was £50m 3520 38.8: 05 88m $0538 m2 32803 no 258083 owoaowoanm N 035.

47

 

A‘vru::..~:»v. V» N wv~a-~.~.

 

5N 03208303048. M0:0..030:3< 00300:: m mm <m02
umm 035830332 Ht0830£8< 00300:: S a w a 4202
an 032083832 m0.2—0:030:03». 00300:: N on <m02
80::8 33080 :03 3:080 0 0N0 . o . 30238 0:0
N03 .. a 0 m 0 H A \..N ma 8:083:03 paw—8030882 w m 3 «$202
8 205:3 0:833 205:3
mm: .555 80me .3 8:083:32 .2 $00 .33: .2 $3 3308000032 02083088: > N E 4202
8». 5 020 888083 83088: , 88882 0 : <32
3:050:38 0:38 _ 3530308 0:8w 3330308
Nam . . . “.2 0:30.82 .xcmm 833832 833052 m 02 <202
N000 ,
8.: 288 8588 :8 588.85 8888.2 \ 8.338883 0 m 2 <32
. 0:833 33.: 3020:0302
R: 3 8888882 .820 888582: .0 +0 E m E <32
, , , . 0:833
mm: 3:308 0830.35 3:308 08385.2. $00 mgomfiofi 0830.593 3 2 «£02
. . . . . 62392 880 0+0 5 . . .. . . , 1
88m 028.8 0.5300 380.0 3383 w 33333 .3330 60v /..o 0:20
and 05 S 0338 3000—0 0U 0U 9Joo

8:

988800 N 2.8.:

48

 

 

«~0Lv3Cm0C.v. v» .H .lififi

 

 

 

2 0:08 <202

: w 34502

 

q 8 2 :88 <32 : Om <32
:0 m M3080m Hmuomvm .
Sm 0:08 0% 0008880800882 00:30:: m mm <202
omm EC 33020 082088 00880:: m mm (£02
3 583: 9 5280808: 8808: m R <32
mum cowlflm 0:08 =3 083380 00880:: m cm «£02
30 Emmi. 0:08 008.803 m mm «£02
30883

000 8808:8080 ”000082008020 00880:: m 0m «£02
08083 08380: :005
00883

N00 8808:8080 00000308020 00880:: m mm <m02
M08083 853-3: :080

800m 0388: 083800 3088 033800 w 333033 .3335 J60 00/00 0%. 0:20
ADM 05 8 0288: 30020 @oo (Q0 @00

688800 N 2%:

49

Pseudomonas stutzeri
NCF A clone 18 'Y

NCFA clone 19
I, seudomonas azotoformans
NCF A clone 2 —

Figure 2 a {Pseudomonas putida

 

 

 

 

 

 

 

l'\

 

. Leptothrix discophora
r—C—Aquaspirillum gracile
_—____ NCFA clone 1 B

Nitrosomonas marina
Azoarcus denitrificians =
NCFA clone 5
I é NCFA clone 4

> Desulfomonile tiedjei
N CF A clone 6 5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

—— Geobacter metallireducens
Desulfovibrio africanus

Desulforhabdus amnigenus

 

 

 

Azorhizobium caulinodans

 

0.10

 

Figure 2 a-e Maximum likelihood dendrograms of bacterial 16S sequences
obtained from the narrow channel focus area in Oyster, Virginia. Divisions are
listed out side the brackets for panels b through d and subdivisions are listed in
panel a (Proteobacteria). Optimality criteria used in bootstrap analysis of the
sequences were: maximum likelihood, maximum parsimony, and neighbor
joining. Bifurcations supported (bootstrap values >7 5%) by one optimality
criteria but are only marginally supported (SO-75%) or not supported (<50%) by
the other criteria are indicated with open circles. Bifurcations supported by two or
three of the criteria are indicated with closed circles. The number of characters in
each analysis and the range of the mask (E. coli numbering) were as follows: a)
330 characters, bases 216-484, b) 443 characters. bases 134-602, c) 332
characters, bases 116-527, (1) 432 characters, bases 134-537, and e) 1165
characters, bases 92-1390.

50

Figure 2 b

N CFA clone 16
r—E‘ Nitrospira mascoviensis
Nitrospira marina

 

 

J

 

—<>

 

é“ Thermodesulfovibrio TGL-LSI

NCFA clone 15 Leptospirillum/
Leptospirillum ferrooxidans Nirrospira

Thermodesulfovibrio islandicus
Magnetobacterium bavaricum

 

 

Sulfalobus solfataricus

 

0.10

51

 

Figure 2 c Rhodococcus rhodochrous '

£orynebacterium jeikeium
Nocardia brasiliensi

Mycobacterium mageritense

Nocardioides simplex

Streptomyces coelicolor
r‘ NCFA clone 13

frirthrobacter psychrolactophilus
Arthrobacter oxydans

Propionibacterium granulosum Gram Positive
Geodermatophilus obscurus bacteria
Frankia (strain Dryas)
cidothermus cellulolyticus

NCFA clone 12

NCFA clone 11
NCFA clone 10

Rubrobacter radiotolerans

ubrobacter xylanophilus

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Escherichia coli

0.10

52

' re 2 d _.
Flgu NCFA clone 9

clone C002, AF013515 (soil)
I clone MC27, X68466 (soil)
Acidobacterium capsulatum

—l_c10ne MC22, X68463 (soil)
NCFA clone 8 F ibrobacter/

3
ECF A clone 7 Acidobacteria

clone OCS146, AF001653 (marine)
l £clone SAR406, U34043' (marine)

F ibrobacter succinogenes

 

 

 

 

 

 

 

 

 

 

WCHBl-Ol, AF050597 (aquifer)
Escherichia coli “'—

 

 

 

0.10

 

53

Figure 2e

 

 

EHydrogenobacter thermophilus
Aquifex pyrophilus

Q Deinococcus radiophilus
Thermus thermophilus

 

 

NCFA clone 30

l—
ENCFA clone 31

 

 

 

V— Verrucomicrobium spinosum
[—_C Prosthecobacter vanneervenii
' Planctomyces limnophilus
E Chlorobium limicola
Chlorobium vibrioforme
4 Spirochaeta smaragdmae

Nocardioides plantarum
Propionibacterium acnes

Streptomyces armeniacus
Lactobacillus reuteri

Staphylococcus saccharolyticus
F usobacterium nucleatum

Syntrophobacter pfennigii
Geobacter metallireducens

H elicobacter pylori
Bradyrhizobium japonicum

FBurkholderia glathei
Escherichia coli

0.10

 

 

 

 

 

  
  
  
  
   

 

 

 

54

F ervidobacterium gondwanense
J I Geotoga subterranea
Thermotoga maritima

Thermotogales

 

_‘—— Thermomicrobium roseum
Chloroflexus aurantiacus

_J
Planctomyces group

Aquifex

Green nonsulfur
bacteria

NCFA group I

Coprothermobacter proteolyticus C. proteolyticus group

Prosthecobacter

 

Green sulfur

 

Spirochetes

bacteria

High GC
Gram positive

 

Low GC
‘ Gram positive

 

F usobacteria

Proteobacteria

 

Piper
ol‘gotrg
Cfl‘vliCT;
metabc

indepe:

 

newt.

 

91 a1
biO‘an

fikfi~c

lheSe
berm

higher in low-carbon environments because of the unsuitability of traditional media for
oligotrophic species. Furthermore, while certain isolates can be derived from a given
environmental sample, they are not necessarily representative of the phylogenetic,
metabolic, or physiological types that dominate that environment. Using cultivation-
independent techniques avoids this ‘bottleneck’ in detecting the dominant members of a
microbial community.

In this study, we examined 12 sediments from a pristine aquifer using the culture-
independent methods T-RFLP and 16S rDNA cloning and sequencing. One strength of
T—RFLP analysis lies in its use in quantifying the microbial diversity in a given sample.
Although it is impossible to extrapolate an exact measure of community diversity from T-
RFLP profiles (42), it is possible to use T-RFLP to compare diversity between two or
more samples. T-RFLP employing the bacterial-specific primer 27F and the restriction
enzyme Hha I has been performed on samples from a number of environments. Soil
typically presents 60-80 unique phylotypes (29). Clement et al (13) found between 30
and 39 fragments in clean and petroleum-contaminated sand from a coastal oil field. Liu
et al. (25) detected 21 fragments in activated sludge, 33 fragments in a glucose-fed
bioreactor, 36 fragments in the termite gut, and 33 fragments in trichloroethylene and jet-
fiiel-contaminated aquifer sediment taken from the unsaturated capillary fringe. T-RFLP
profiles of the pristine aquifer sand from Oyster reveal fewer phylotypes than most of
these previously studied environments. In the narrow channel focus area we detected
between 13 and 39 phylotypes in the Hha I digestion, and profiles fi'om the south Oyster

focus area reveal between 8 and 23 phylotypes.

55

In comparison to soil (29), contaminated aquifer sediment (25), and the
contaminated sand tested by Clement et al. (13), the microbial communities from the
pristine aquifer at the Oyster study sites are relatively low in diversity. Lower diversity is
consistent with the relatively lower nutrient and organic carbon concentrations in Oyster
aquifer sediment. The greater the range of potential carbon sources, the greater the
number of specialist populations aimed at exploiting them (12). Water-saturated aquifer
sediment is also a less complex microbial environment than soil, sand, or unsaturated
sediment. Unsaturated particles bear innumerable different microniches where moisture,
nutrients, and dissolved gas concentrations can vary widely across small distances.
Water saturation creates a bridge between sediment particles, diminishing gradients of
nutrients and gases and allowing microbes to move more freely. The result is a more
consistent microbial environment in the interstitial space, fewer microniches for
specialized organisms, and possibly a lower microbial diversity. It can be argued that the
water-saturated environment in Oyster aquifer sediment is less complex than the
environments tested previously with T-RFLP, and that this is one reason for the lower
diversity we detected there.

It is apparent from an inspection of the data that there was a degree of similarity
between narrow channel and south Oyster T-RFLP profiles, indicating that although the
geochemistry was difl’erent at the two sites, the microbial communities may have had
some populations in common. T-RFLP profiles from the SOFA and the NCFA were
generated using the same amount of DNA and had comparable amounts of total
fluorescence per profile, allowing comparisons to be drawn between the two data sets.

The two sites shared as many as 50% of the same fragments in the Hha I digestion and

S6

had a comparable number of fragments in each profile, suggesting that many of the same
populations may be found at both sites. This was surprising in light of the fact that the
two sites had very different concentrations of oxygen and organic carbon. The two sites
are within the Wachapreague formation, however, and share the same overlying soil type
(14). It is reasonable to expect that sediment origin and soil nutrients have a determining
effect on aquifer microbial community structure, and we believe that the sinrilarity we
observe in the profiles may be due to one or both of these factors.

It appears that the primer set used to amplify the 168 genes for T-RFLP may be
an important determinant of the resulting community profile. Two different reverse
primers, 1392R and 1525R, were used on identical DNA templates from the NCFA in
order to study the effect of the reverse primer on T-RFLP results. For comparing profiles
generated by the two primer sets we considered only those fragments that exceeded 150
fluorescence units in height as peaks below this threshold may not be consistently
detectable in replicate profiles. Figure 3 is a comparison of T-RFLP profiles from NC
B2 (6 m) and NC M3 (6 m) sediment generated using these primers and the restriction
enzyme Hha I. We observe a significant effect in the NC B2 profiles: both the presence
and intensity of certain terminal fragments are affected. The NC B2 (6 m) profile
generated using 1392R is missing two fragments that are present in the 1525R profile: 95
and 384 bp and the 1525 profile is apparently missing fi'agment 477 bp. The NC M3 (6
m) profiles, however, are more comparable and we detected all fiagments of over 150
fluorescence units in both profiles. It is difficult to assess the relative specificity of the
two primers since most sequences in the RPDH database lack sequence information

beyond base 1400, but it was noted that of the 947 matches to primer 1525R, 910 of these

57

110 220 330 440 I 550

400 s
200 a
O

400
200 i

400C 2

200

0 MW
400 d j

200

o A L A L l A I __ /\. th

Figure 3 Differences between T-RFLP profiles generated
using two different reverse primers. In comparing the
profiles of template NC BZ (6 m) created with primer set (a)
27F/1392R and (b) 27F/1525R, we see the appearance and
dissappearance of three significant peaks. The profiles of NC
M3 (6 m) created using the same two primer sets (0)
27F/1392R and (d) 27F/1525R are more comparable and all
significant peaks are found inboth profiles.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

58

also matched the primer 1392R, suggesting that the two primers have similar
specificities. In light of the differences between 1392R and 1525R profiles, it is clear
that carefirl consideration should be given to the selection of a reverse primer for T-RFLP
analysis. In the interest of making meaningfirl comparisons between profiles, the same
reverse primer should be used in all reactions.

Cloned 16S sequences from the Oyster narrow channel site represent six major
bacterial divisions (Figure 2a-e), and 28 out of 31 sequence types are sufficiently
divergent (<97% similar to known species) to constitute new taxa. A number of
sequences fall within the B, y, and 6 Proteobacteria (Figure 2a). Other clones were
affiliated with the division Leptospirillum / Nitrospira (Figure 2b), the Gram positive
bacteria (Figure 2c), and the division Fibrobacter / Acidobacteria (Figure 2d). In
addition, we have detected ten sequence types (Table 2) which exhibited relatedness
values of less than 0.573 when compared to sequences in the RDP 11 version 8.0 database
and relatedness scores less than 557 when compared to sequences in GenBank using the
BLAST tool at the NCBI website (1). These sequences may represent novel divisions as
they fail to show a distinct relationship to any of the known divisions.

One group of unidentified clones, NCFA group I, was of particular interest as
analysis of the 16S sequence places them among the groups previously identified as
deeply rooted in the bacterial phylogenetic tree. The ARDRA pattern displayed by these
clones was found in all libraries (Table 3). Among the clones analyzed using ARDRA,
we uncovered four with this fragment pattern in the B2 (8 meters) library, three in the NC

818 (6 meters) library, and five in the NC S18 (8 meters library). We have explored the

59

Table 3. Presence/absence of deeply branching clone terminal restriction
fragments in TRFLP profiles from the narrow channel study site

 

Restriction enzyme Frangent (bp) B2 (8 m) S18 (6 m) S18 (8 m) M3 (6 m)
Hha I 61 (Type II) + + + -
Msp I 515 (Type I) + + + -
Msp I 150 (Type II) + + + -
Rsa I 498 (Type I, II) + + + -

 

6O

diversity of this group in a firrther investigation using a 16S primer set to clone NCFA
group I sequences from narrow channel sediment DNA (see chapter 3).

There was a degree of redundancy in the narrow channel clone libraries. Among
the 120 clones examined with ARDRA, we observed 31 restriction fragment patterns that
appear in the libraries more than once. For example, the restriction pattern for the
undefined clone NCFA 21 was found in all three libraries, the pattern belonging to the
“T78” clone type was found in both the NC 818 (6 meters) and the NC S18 (8 meters)
libraries, and the pattern belonging to NCF A group I was found in all three libraries.

Redundancy in the clone libraries was not unexpected. The geochemistry of the
narrow channel study site was more or less consistent across all sampled boreholes (21).
Furthermore, similarity analysis of T-RFLP profiles from the site indicates that the
sampled microbial communities are markedly similar. There are also several ARDRA
patterns and clones that appear to be limited to a single clone library. It is possible that
these clones are, in fact, present in each of the other libraries, but were not among the 120
clones screened or the 52 clones sequenced.

Phylogenetic data is not always a reliable indicator of the metabolism or
physiology of a given organism (2), but some bacterial divisions and genera possess a
consistent metabolic profile that can be extrapolated to clones found in a given
environment. In the narrow channel clone libraries, three out of the 31 clone types have
relationships with divisions and genera of this kind, and therefore we may speculate
about their role in the environment. Clone 13 is closely related to the Hi G+C gram
positive genera Arthrobacter (Figure 2c) and is 98% similar to Arthrobacter oxidans.

Arthrobacter is a common soil bacterium which has proven to be resistant to desiccation

61

and starvation. Many Arthrobacter species have been found previous studies of the
saturated subsurface (8, 15) and it is thought to be an important group in this environment
because of the broad substrate range observed in many Arthrobacter isolates and the high
numbers of these bacteria observed in the studies conducted by Balkwill et al. (8), and
Crocker et al. (15). Clones 18 and 19 from the S18 (8 meters) library are
phylogenetically placed within the Pseudomonas group of the y Proteobacteria, and are
most closely related to P. stutzeri and P. azotoformans. Pseudomonas strains have been
isolated from subsurface environments previously (8) and are best known for their strictly
respiratory metabolism and their broad range of carbon substrates.

Clones 7, 8, and 9 from the B2 (8 m) library group phylogenetically with the
Acidobacteria of the Fibrobacter / Acidobacter division. To date, there exists only a
single cultivated representative of the Acidobacteria, but the phylogenetic diversity of
this group is on a par with that of other bacterial divisions (22). Acidobacterial sequences
have previously been found in environments ranging from soil to hot spring mats to lake
water (9, 43) and so little is currently known about this group that speculation about their

' role in the aquifer community is impossible.

NCFA clones 15 and 16 fall within the division Leptospirillum / Nitrospira which
has a few cultivated species but mostly consists of rRNA sequences derived from a range
of different environments. Known members of this group include Leptospirillum
ferrooxidans, a lithotrophic, acidophilic species implicated in the evolution of acid mine
drainage, and Nitrospira moscoviensis and Nitrospira marina, both nitrite-oxidizers. The
relatively high nitrate concentration in this formation (3.8 — 7.1 mg N/l) (21) may be

linked to the presence of these Nitrospira-like bacteria.

62

Previous cultivation-independent studies of microbial communities from
uncontaminated subsurface formations provide a counterpoint to this work. Chandler et
al. (1997, 1998) developed clone libraries from sediment extracted from a deep formation
at the Hanford site in Washington state. As in the current study, they found clone types
related to Pseudomonas and Arthrobacter but they also uncovered sequences related to
Bacillus, Micrococcus, Nocardioides, Clavibacter, Comamonas, Erythromicrobium, and
Burkholderia. In another investigation of a deep formation, Fry et al. (18) used an rRNA
hybridization approach to characterize the bacterial community. They found the
sediment communities to be largely bacterial (64.4-92%) with a portion of that group
comprised of gram positive species (7.6 — 11.7%). Studies of shallow subsurface
formations are perhaps more relevant to the current work. Shi et al. (37) used rRNA
hybridization techniques to study the community from a shallow, sandy aquifer in
Wisconsin. Like Fry et al. (18) they found the community from pristine sediment to be
largely bacterial (Archaeal and Eukaryal signals were not detectable). As in our study,
Shi et al. found the B and y Proteobacteria to form a substantial part of the community
(43%), that sulfate-reducing genera comprise 15-18%, and that high G+C gram positive
bacteria comprise 5-10% of the community. Unlike our study, in which we failed to
uncover or Proteobacteria] clones, Shi et al. found the or Proteobacteria to comprise 35%
of the community. It is unknown why our clone library (and those of Chandler et al. (10,
11)) failed to contain a single or proteobacterial sequence, while it seemed to be such a
significant group in another shallow, sandy, aerobic aquifer. One explanation may lie in

a key difference between the two environments: organic matter concentrations. The

63

organic carbon concentration at the narrow channel site was 1 ppm while organic matter
concentrations in the Wisconsin sediment ranged from 2,000 to 32,000 ppm.
The same forward primer, 27F, was used to amplify bacterial 16S genes for T-
RFLP and for cloning. Hence, one should be able to predict the terminal restriction
fragment length of a given clone from the sequence of the cloned 16S genes. The
terminal restriction fragments for NCF A group I was determined in silica and we were
able to discern the terminal fragment that correlated with this group in each of the
profiles from the narrow channel study site except NC M3 (Table 3). A clone library was
generated from NC M3 DNA, and 40 clones were screened by ARDRA in the same
manner described for the other sediments. None of the NC M3 clones possessed the
NCFA group I signature ARDRA profiles (data not shown). However, we were unable to
detect the terminal fragments of other NCFA clones in the T-RFLP profiles. This may be
due to a number of factors. Firstly, many of the cloned sequences lacked 'a restriction site
in the 400 bases of sequence available for analysis, so we were unable to match these
with terminal fragments in the T-RFLP profiles. If the original, pre-digestion sequence
held a restriction site between bases 400 and 600 it would be detected with T-RFLP.
Secondly, it is possible that the labeled forward primer, 27F*, has an effect on the PCR,
causing a biased reaction that, while providing a reproducible profile of the community,
does not provide the same product that amplification with the unlabeled primer does.
The microbial communities of the saturated subsurface play an integral role in the

firnctioning of aquatic ecosystems, but to date the composition of these communities has
been studied in only a limited number of investigations. We have found that the diversity

if the microbial communities in a shallow aquifer was significantly lower than that

64

observed in soil, a fact that may be attributable to the comparative dearth of organic
carbon in aquifer sediment or to a lower habitat complexity. Furthermore, differences in
the geochemistry between our two sampling sites do not entirely eliminate the
community-level similarities that may have arisen from the common origin of the

sediments. So, while flow in shallow aquifers is affected by precipitation events, the

microbial communities may remain stable. The presence of Arthrobacter and‘

Pseudomonas strains in the NCFA sediments may be a determinant of this stability: both
groups are known for their broad substrate ranges, enabling them to endure the wide
nutrient fluctuations that are likely in such a shallow formation. Further work in other
pristine formations should be conducted in order to determine whether these groups are

ubiquitous and whether they remain numerically stable over the course of short-term

moisture and nutrient fluctuations.

65

 

 

 

 

10.

11.

12.

13.

14.

REFERENCES
National Center for Biotechnology Information (NCBI) Basic Local Alignment
Tool [Online] Available http://www.ncbi.nlm.nih.gov/BLAST/.
Achenbach, L. A., and J. D. Coates. 2000. Disparity between bacterial
phylogeny and physiology. ASM News. 66(December):p. 714.
Aelion, C. M., and P. M. Bradley. 1991. Aerobic biodegradation potential of
subsurface microorganisms from a jet fuel-contaminated aquifer. Applied and
Environmental Microbiology. 57(1):57-63.
Amann, R. L, L. Krumholz, and D. A. Stahl. 1990. Fluorescent-oligonucleotide
probing of whole cells for determinative, phylogenetic, and environmental studies
in nricrobiology. Journal of Bacteriology. l72(2):762-770.
Amann, R. I., W. Ludwig, and K. H. Schleifer. 1995. Phylogenetic
identification and in situ detection of individual microbial cells without
cultivation. Microbiol-Rev. 59(1):143-69.
Balkwill, D. L., J. K. Fredrickson, and J. M. Thomas. 1989. Vertical and
horizontal variations in the physiological diversity of the aerobic
chemoheterotrophic bacterial microflora in deep southeast coastal plain
subsurface sediments. Applied and Environmental Microbiology. 55(5):1058-
1065.
Balkwill, D. L., and W. C. Ghiorse. 1985. Characterization of subsurface
bacteria associated with 2 shallow aquifers in Oklahoma (USA). Applied and
Environmental Microbiology. 50(3):580-588.
Balkwill, D. L., E. M. Murphy, D. M. Fair, D. B. Ringelberg, and D. C.
White. 1998. Microbial communities in high and low recharge environments:
Implications for microbial transport in the vadose zone. Microbial Ecology.
March April. 35(2): 156-171.
Barns, S. M., S. L. Takala, and C. R. Kuske. 1999. Wide distribution and
diversity of members of the bacterial kingdom Acidobacterium in the
environment. Applied and Environmental Microbiology. April, 1999;.
65(4):]731-1737.
Chandler, D. P., F. J. Brockman, T. J. Bailey, and J. K. Fredrickson. 1998.
Phylogenetic diversity of archaea and bacteria in a deep subsurface paleosol.
Microbial Ecology. July Aug. 36(1):37-50.
Chandler, D. P., S. M. Li, C. M. Spadoni, G. R. Drake, D. L. Balkwill, J. K.
Fredrickson, and F. J. Brockman. 1997. A molecular comparison of culturable
aerobic heterotrophic bacteria and 16S rDNA clones derived from a deep
subsurface sediment. FEMS Microbiology Ecology. 23(2): 13 1-144.
Cho, J.-C., and S.-J. Kim. 2000. Increase in bacterial community diversity in
subsurface aquifers receiving livestock wastewater input. Applied-and-
Environmental-Microbiology. March, 2000;. 66(3):9S6-965.
Clement, B., G., L. Kehl, E., K. Debord, L., and C. Kitts, L. 1998. Terminal
restriction fragment patterns (TRFPs), a rapid, PCR-based method for the
comparison of complex bacterial communities. Journal of Microbiological
Methods. Jan. 31(3):]35-142.
Cobb, P. R., and d. W. Smith. 1984. Soil Survey of Northampton Connty,
Virginia. USDA-SCS. US. Government Printing Office, Washington, DC.,

66

 

 

 

 

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Cracker, F. H., J. K. Fredrickson, D. C. White, D. B. Ringelberg, and D. L.
Balkwill. 2000. Phylogenetic and physiological diversity of Arthrobacter strains
isolated from unconsolidated subsurface sediments. Microbiology-Reading.
[print] June, 2000;. 146(6): 1295-1310.

Dunbar, J., S. Takala, M. Barns Susan, A. Davis Jody, and R. Kuske Cheryl.
1999. Levels of bacterial community diversity in four arid soils compared by
cultivation and 16S rRNA gene cloning. Applied-and-Environmental-
Microbiology. April, 1999;. 65(4):]662-1669.

Frederickson, J. K., D. L. Balkwill, J. M. Zachara, S. M. Li, F. J. Brockman,
and M. A. Simmons. 1991. Physiological diversity and distributions of
heterotrophic bacteria in deep cretaceous sediments of the Atlantic coastal plain.
Applied and Environmental Microbiology. 57(2):402-411.

Fry, N. K., J. K. Fredrickson, S. Fishbain, M. Wagner, and D. A. Stahl. 1997.
Population structure of microbial communities associated with two deep,
anaerobic, alkaline aquifers. Applied and Environmental Microbiology.
63(4): 1498-1504.

Gehlen, M., H. J. Trampisch, and W. Dott. 1985. Physiological
characterization of heterotrophic bacterial communities from selected aquatic
environments. Microbial Ecology. ll(3):205-220.

Ghiorse, W. C., and Balkwill, D.L. 1983. Enumeration and morphological
characterization of bacteria indigenous to subsurface environments.
Developments in Industrial Microbiology. 24:213-224.

Golder-Associates. 1998. Field Summary Report, groundwater and sediment

characterization, South Oyster Site, Oyster, Virginia,
Hugenholtz, P., M. Goebel Brett, and R. Pace Norman. 1998. Impact of

culture-independent studies on the emerging phylogenetic view of bacterial

diversity. Journal of Bacteriology. Sept. 180(18):4765-4774.
Koelbel Boelke, J., B. Tienken, and A. Nehrkorn. 1988. Microbial communities

in the saturated groundwater environment: 1. Methods of isolation and
characterization of heterotrophic bacteria. Microbial Ecology. l6(1):17-30.

Lane, D. J. 1991. 16S/23S rRNA sequencing, p. 115-175. In E. S. a. M.
Goodfellow (ed), Nucleic acid techniques in bacterial systematics. John Wiley
and Sons, Inc., New York, NY.

Liu, W. T., T. L. Marsh, H. Cheng, and L. J. Forney. 1997. Characterization of
microbial diversity by determining terminal restriction fi'agment length
polymorphisms of genes encoding 16S rRNA. Applied and Environmental
Microbiology. Nov. 63(1 1):4516-4522.

Ludvigsen, L., H. J. Albrechtsen, D. B. Ringelberg, F. Ekelund, and T. H.
Christensen. 1999. Distribution and composition of microbial populations in a
landfill leachate contaminated aquifer (Grindsted, Denmark). Microbial-Ecology.
April, 1999;. 37(3): 197-207.

Madsen, E. L., J. L. Sinclair, and W. C. Ghiorse. 1991. In situ biodegradation:
Nficrobiological patterns in a contaminated aquifer. Science. 252(5007):830-833.
Maidak, B. L., J. R. Cole, T. G. Lilburn, C. T. J. Parker , P. R. Saxman, J. M.
Stredwick, G. M. Garrity, B. Li, G. J. Olsen, S. Pramanik, T. M. Schmidt,

67

 

 

 

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

and J. M. Tiedje. 2000. The RDP (Ribosomal Database Project) continues.
Nucleic Acids Res. 28: 173-174.

Marsh, T. L. 1999. Terminal restriction fragment length polymorphism (T-
RFLP): an emerging method for characterizing diversity among homologous
populations of amplification products. Curr Opin Microbiol. 2(3):323-7.

McCarthy, C. M., and L. Murray. 1996. Viability and metabolic features of
bacteria indigenous to a contaminated deep aquifer. Microbial Ecology.
32(3):305-321.

Nishino, S. F., J. C. Spain, L. A. Belcher, and C. D. Litchfield. 1992.
Chlorobenzene degradation by bacteria isolated from contaminated groundwater.
Applied and Environmental Microbiology. 58(5): 1719-1726.

Olsen, G. J., D. J. Lane, S. J. Giovannoni, N. R. Pace, and D. A. Stahl. 1986.
Microbial ecology and evolution: a ribosomal RNA approach. Annu Rev
Microbiol. 40:337-65.

Pace, N. R. S., D.A.; Lane, D.J.; Olsen, G.J. 1986. The analysis of natural
microbial populations by ribosomal RNA sequences. Adv. Microb. Ecol. 9:Jan-
55.

Pedersen, K., J. Arlinger, S. Ekendahl, and L. Hallbeck. 1996. 16S rRNA
gene diversity of attached and unattached bacteria in boreholes along the access
tunnel to the Aspo hard rock laboratory, Sweden. FEMS Microbiology Ecology.

l9(4):249-262.
Posada, D., and K. A. Crandall. 1998. MODELTEST: testing the model of

DNA substitution. Bioinformatics. 14((9)):817-818.

Pusch, M., D..Fiebig, I. Brettar, H. Eisenmann, B. K. Ellis, L. A. Kaplan, M.
A. Lock, M. W. Naegeli, and W. Traunspurger. 1998. The role of micro-
organisms in the ecological connectivity of running waters. Freshwater Biology.

Nov. 40(3):453-495.
Shi, Y., M. D. Zwolinski, M. E. Schreiber, J. M. Bahr, G. W. Sewell, and W.

J. Hickey. 1999. Molecular analysis of microbial community structures in pristine
and contaminated aquifers: Field and laboratory microcosm experiments.

Applied-and-Environmental-Microbiology. May, 1999;. 65(5):2143-2150.
Sinclair, J. L., S. J. Randtke, J. E. Denne, L. R. Hathaway, and W. C.

Ghiorse. 1990. Survey of microbial populations in buried-valley aquifer

sediments from northeastern Kansas (USA). Ground Water. 28(3):369-377.
Staley, J. T., and A. Konopka. 1985. Measurement of in situ activities of
nonphotosynthetic microorganisms in aquatic and terrestrial habitats. Annu Rev

Microbiol. 39:321-46.
Strunk, 0., and W. Ludwig. 1997. ARB: Software for phylogenetic analysis.

Technical University of Munich, Munich, Germany.
Suflita, G. D. M. a. J. M. 1985. Microbial metabolism of chlorophenolic

compounds in ground water aquifers. Environmental Toxicology and Chemistry.

4:751-758.
Tiedje, J. M., S. Asuming—Brempong, K. Nusslein, T. L. Marsh, and S. J.

Flynn. 1999. Opening the black box of soil microbial diversity. Applied Soil
Ecology. 13:109-122.

68

 

 

 

43.

44.

Wise, M., G., J. V. McArthur, and J. Shimkets Lawrence. 1997. Bacterial
diversity of a Carolina bay as determined by 16S rRNA gene analysis:
Confirmation of novel taxa. Applied and Environmental Microbiology.
63(4):]505-1514.

Zhou, J., M. A. Bruns, and J. M. Tiedje. 1996. DNA recovery from soils of
diverse composition. Applied and Environmental Microbiology. 62(2):316-322.

69

 

 

 

 

 

 

CHAPTER 3
DETECTION OF A NOVEL, DEEPLY-BRANCHIN G GROUP OF BACTERIA IN
PRISTINE AQUIFER SEDIMEN T
INTRODUCTION

Efforts to reveal microbial diversity have been greatly aided by the advent of

cultivation-independent methods of community analysis. By circumventing the

cultivation step, which can prevent the detection of >99% of microbes (1), molecular
methods have afforded insight into the distribution of the recognized species of bacteria
and archaea and have revealed previously undetected microbial diversity in multiple
environments (5). The 16S gene, in particular, has been a useful tool in identifying the
distribution of organisms which resist cultivation.

There are currently 36 accepted bacterial phylogenetic divisions, wherein a
division is defined as a related group of 168 sequences which show no specific relation to
any other branch in the bacterial phylogenetic tree (5). Of these, 13 are represented
solely by 16S rRNA sequences and lack cultivated representatives (5). Most divisions,
including the Acidobacteria, Nitrospira, Verrucomicrobia, and the Planctomycetes,
contain one or more cultivated strains but are largely composed of cloned sequences from
environmental DNA. The 16S sequence diversity within these divisions is comparable to
that observed in the more established divisions in which many strains have been
cultivated and characterized (5). The discovery of new sequences and new divisions by
cultivation-independent methods continues to illustrate that the breadth of the bacterial

tree is not limited to the cultivated species on hand, and that much of the phylogenetic

7O

 

 

 

(and therefore metabolic and physiological) diversity of the microbial world has yet to be

illuminated.

We have identified a novel group of 16S sequences in aquifer sediment fiom a
study site in Oyster, Virginia. Using bacterial-specific primers, we amplified bacterial
l6S sequences fi’om whole community DNA and cloned the resulting sequences (see
Chapter 2). This group is composed of two distinct types of clones which together
comprise roughly 13% of all clones in three bacterial clone libraries. These sequences,
referred to as NCFA group I, fail to show a specific relationship to any of the established

bacterial divisions. Furthermore, they appear to be deeply-rooted in the bacterial

phylogenetic tree, suggesting that NCF A group I may have diverged from rest of the

divisions at an early point in the evolution of the kingdom. The divisions at the root of
the bacterial tree, including Aquifex, 771ermotogales, and F ervidobacterium, consist of
thermophilic species (3). It has been found that bacteria closely related to thermophiles
may abound in mesophilic environments (10), but mesophilic relations to thermophiles
are not well represented in the literature. The detection of deeply-branching phylogenetic

diversity in a mesophilic environment such as the Oyster aquifer can serve to expand our

understanding of the root of the bacterial tree.

In order to elucidate the diversity of NCFA group I, we designed a primer set that
specifically targets these 16S sequences and used it to construct clone libraries from
several sediment samples taken from the Oyster site. The clone libraries revealed a tight

cluster of deeply-branching phylotypes from points broadly dispersed across the study

site.

71

 

 

 

MATERIALS AND METHODS

Sample collection Samples were aseptically collected from the narrow channel focus
area (NCFA) in Oyster, Virginia in October 1998 and August 2000. Cores were placed
in plastic liners, divided, placed in gamma-irradiated polypropylene tubes, and held at 4°
C for seven days, after which they were stored at -80°C. Sediment samples were
collected from 6.5 — 8.0 meters below the ground surface. Figure l is a diagram of the
site and sampling points.

DNA extraction DNA was extracted from sediment samples with a Soil DNA Kit Mega
Prep (MoBio) according to the manufacturer’s instructions, using 10 g of sediment for
each extraction. To concentrate the final solution, 0.04 volumes of 5 M NaCl and 2
volumes of ethanol was added and the mixture was centrifuged at 9,000 x g for 30
minutes. The pellet was dried and resuspended in 50 ul of water and stored at -20°C until
needed.

Primer design A 16S rRNA primer was designed to specifically target the group of
interest using the probe design firnction in the Arb genetic data environment software

(12). The primer was compared against the sequences in the RDP database version 8

using the probe match firnction through the RDP website

 

(http://www.cme.msu.edu/RDP/cgis/probe match.cgj?su=SSU) to ensure specificity.
The target-specific primer was designated 984R (5’ - ATC CAG CAT GTC AAA CCC
TG-33.

Cloning of 168 rRNA genes Ten sediments were used for cloning experiments: S18 (6
m below the surface), S9 (6 m), M3 (6 m), and B2 (8 m) sediments were collected in

October 1998. oouz (6 m), ODU2 (7 m), ODU3 (5 m), ODU3 (6 m), ODU4 (6 m), and

72

 

      

 

   
  

   

      

.' , " .
g . NA", "6“. ' a .
. «a. ' .. .
. x: " {‘3’ M
7.33"} “A5109; :22") "’f
I“ ’ ; . '3? .

  
 
 
  

-’- ~ .r m” 5.7.:1'5‘ r.
._ i 40'. . .. ”3.: » _-._'.,' --T.'~:...a'...'.-".~ '97". . .47. _ 3.. ‘w'*‘
-: - -..: ”"1’41‘?-<~‘~4§’-‘5 5} .. --. .'” "

. . .‘ .. “'0'. jg...‘ .‘.-...- 7 '
:'I" ‘ ‘ .3“ g " fl /‘ {-0.}. .'. ‘ '. '-ac . ‘ ' -
fl: . "’ ' ' 29"," ~22. ' - w“. . ‘ .'
' ..'.. ‘I\‘.“
. - >0 .'.‘..'.. A, an
....:.-..;...-

_ . . . .
. ' N
{55¢ "t A». '..-' .
‘ ...I ._ . . _. : . I...“ _. . .v _ .
’ .'.'. La ' .' ' . .'. . "
'. x . _ . . ': {' ._
. _ _ .. . . .
I ‘ u u .‘ . 'I . .
.. , ,
. . )

  

      

   

  

 

Narrow Channel Focus Area (N CFA)

. O 2 4
——-——
ODU4 Scale in meters

. Q . .NCM3

NC B2 NC S18 ODU3

. Hydrological flow gradient
oouz >

 

 

 

 

Figure 1 Map of Oyster, Virginia, indicating the location of the narrow channel focus
area and the relative positions of sampling points within the site.

73

ODU4 (7 m) were collected in August 2000. Bacterial 16$ ribosomal genes were
amplified from bulk DNA in triplicate 100 ul reactions that contained 1X PCR buffer
(Perkin Elmer), 5 mM MgC12, 0.2 mM of each dNTP, 0.2 mM of each primer, 8 ng
bovine serum albumin per 111, 30 ng template DNA, and 0.02 U of AmpliTaq (Perkin
Elmer) per ul. To target bacterial 168, the forward primer 27F, which is specific for
bacteria (5’-AGA GTT TGA TCC TGG CTC AG-3’) (6) and the universal primer 1392R
(5’-ACG GGC GGT GTG TRC-3’) (8) were used. The triplicate reactions were
combined and purified using a Wizard® PCR Prep (Promega) according to the
manufacturer’s instructions and DNA was eluted in a final volume of 50 ul. The purified
PCR products were used as the template for nested PCR reactions using the NCFA group
I specific primer set.

NCFA group I sequences in the bacterial 16S rDNA were amplified under the
same conditions as those described for amplifying the bacterial l6S rDNA except that 50

ng template DNA was used per reaction. Triplicate reactions were combined for cloning

into the vector.

The PCR products were cloned using a TOPO TA Cloning Kit (Invitrogen Corp.)
in accordance with the manufacturers instructions. Plasmid DNA was extracted and
purified with a Qiagen Mini Plasmid-prep kit (Qiagen).

Screening of 168 clones by ARDRA The plasmid inserts of all clones were amplified
using the PCR conditions described for cloning, with 30 ng of purified plasmid DNA
template per 25 ul reaction. Five ul of rDNA PCR products were digested with 10 U of
the 4-base specific restriction enzyme Cfo I in 1X NEB buffer (New England Biolabs) in

a final volume of 15 111 for 3 hrs and 30 min at 37°C. Digested DNA fragments were

74

separated by agarose (3.5% agarose) electrophoresis and visualized by staining with

ethidium bromide and illumination on a UV transilluminator. Restriction fragment

lengths were estimated by comparison with molecular weight standards which ranged
from eight to 587 bp (DNA molecular weight marker V2, Roche Molecular
Biochemicals), and clones with unique patterns were selected for partial sequencing.
Sequencing of rDNA clones Plasmid inserts from selected rDNA clones were amplified
by the PCR using the same conditions described above for cloning. Amplified rDNA
inserts were purified by using Wizard® PCR purification columns (Promega) according to
the manufacturer’s instructions. Sequencing was performed using the ABI Prism BigDye

Terminator Cycle Sequencing Reaction Kit and an ABI Prism 377 DNA sequencer (PE

Applied Biosystems) according to the manufacturer’s directions. The primer 27F was

used for sequencing six clones and four of these were selected for sequencing with the

primer 984R.
Phylogenetic analyses All sequenced clones were checked for chimeric sequences using

the Ribosomal Database Project II CHECK_CHIMERA version 2.7 (7). Two chimeras

were detected among the clones and were eliminated from analysis. Sequences were

aligned against close relatives in the Ribosomal Database Project release 8.0, and percent
similarity to previously cloned NCFA I sequences was determined. For analysis of
sequenced clones, a mask was generated which excluded all ambiguous positions.

Dendrograms were generated using the maximum likelihood algorithm (F astDNAML) in

the Arb genetic data environment package (12). The robustness of dendrogram

topologies was tested by bootstrap resampling of trees using evolutionary distance (beta

75

version 4.0b6 of PAUP*, written by David Swofl‘ord; distance algorithm) with settings

selected by execution of the dataset with ModelTest (9).

RESULTS
Description of the Study Site The site is located in the village of Oyster, Virginia, on

the Delmarva Peninsula. The aquifer is currently the focus of a study on bacterial

transport by the US. Department of Energy’s Natural and Accelerated Bioremediation
Research Program. The aquifer is a shallow, unconfined, sandy formation (4). The
narrow channel focus area (N CFA) lies to the south-southwest of the village of Oyster
(Fig. 1). The water in this part of the formation is oxic (5.5 mg 02/1) and low in organic
carbon (1000 ppb) (4). The water table lies 1.5 to 2.0 meters below the ground surface
(4)-
Phylogenetic analyses The NCFA group I-specific primer 984R did not match any other
bacterial or archaeal 16S sequences in the RDP 11 database (version 8) (7). Using the
27F/984R primer pair we were able to amplify NCFA group I 168 sequences from eight
of the 10 168 DNA samples tested. A separate clone library was developed from each of
these PCR products. We isolated 18 clones from each library and analyzed each of these
clones with ARDRA, wherein we detected six different ARDRA patterns. A
representative of each type was sequenced with a single pass using the primer 27F. Two
of the ARDRA types (for which we had only one representative) proved to be chimeric
and were eliminated from further analysis. One representative of each of the other four
ARDRA types was sequenced using both the 27F primer and 984R, providing over 970

bases of sequence for each of the clones. From prior work with a bacterial primer set (see

76

Chapter 2), we have 400-500 bases of nucleotide sequence of eight NCFA group I clones

and 1400 bases of sequence from five NCF A group I clones.
In the former study (see Chapter 2), we detected two distinct types of NCFA

group I sequences using the bacterial primer set. Types 1 and 2 are reproducibly

monophyletic and are 91% identical in 16S sequence. In the bacterial clone libraries,
type 1 sequences were found only in the B2 (6 m) library and type 2 sequences were
found in both the S18 (6 m) library and in the S18 (8 m) library. In the present study, we
were able to recover type 1 sequences from all of our clone libraries that were created
using the NCFA group I-specific primer set (27F/984R). We did not detect type 2
sequences in these clone libraries. 1

NCF A group I sequences group deeply within the bacterial phylogenetic tree,

showing a degree of relatedness to Coprothermobacter. Figure 2 is a maximum

likelihood dendrogram of type 1 and type 2 sequences in the context of selected bacterial

l6S sequences. Bootstrap analysis with the methods of neighbor joining (100

replicates), maximum parsimony (100 replicates), and maximum likelihood (25
replicates) was used to determine the reproducibility of the branchings. Bifurcations that
were supported by two or three of these methods (>75% of bootstrap resampling) are

indicated with a closed circle. Open circles indicate branchings which were supported by

one method.

DISCUSSION
Diversity of a deeply-branching group of bacteria We have identified two closely

related 16$ rDNA sequences in aquifer sediment which branch deeply within the

77

 

0.1

L—
E Escherichia coli

 

 

F ervidobacterium gondwanense
‘ I Geotoga subterranea
Thermotoga maritima —
. EHydrogenobacter thermophilus
Aquifex pyrophilus

—L—Thermus thermophilus

 

 

Thermotogales

Aquifex

 

Deinococcus radiophilus

 

Thermomicrobium roseum

 

 

 

ENCFA clone 31

‘ Chloroflexus aurantiacus

Green nonsulfur
bacteria

 

 

N CFA clone 30

N CFA group I

 

 

 

1—

 

 

 

Coprothermobacter proteolyticus C. proteolyticus group

 

 

 

 

 

 

 

 

 

 

  
  
 

 

 

 

 
  
 

 

 

 

Verrucomicrobium spinosum Prosthecobacter
€— Prosthecobacter vanneervenii
Planctomyces limnophilus Planctomyces group
E Chlorobium limicola Green sulfur
Chlorobium vibrioforme bacteria
Spirochaeta smaragdinae Spirochetes
Nocardioides plantarum —‘ High GC
Propionibacterium acnes Gram positive
Streptomyces armeniacus _
{Lactobacillus reuteri LOW GC
Staphylococcus saccharolyticus _ Gram P035171“?
F usobacterium nucleatum Fusobacteria
S yntrophobacter pfennigii —‘
Geobacter metallireducens
Helicobacter pylori Proteo bacteria
Bradyrhizobium japonicum
Burkholderia glathei

O

 

' Figure 2 Maximum likelihood dendrogram of NCFA group I 16S sequences
obtained from the narrow channel focus area in Oyster, Virginia. Divisions are
listed outside the brackets. Optimality criteria used in bootstrap analysis of the
sequences were: maximum likelihood, maximum parsimony, and neighbor
joining. Bifurcations supported (bootstrap values >75%) by one optimality
criteria but are only marginally supported (SO-75%) or not supported (<50%) by
the other criteria are indicated with open circles. Bifurcations supported by two or
three of the criteria are indicated with closed circles. There were 1165 characters
included in the analysis and the range of the mask (E. coli numbering) covered

bases 92- l 390.

78

bacterial phylogenetic tree. They are closely related (91% identical) and reproducibly

monophyletic. We first discovered this group among the sequences in bacterial clone
libraries (see Chapter 2), where we found type 1 clones in the B2 (6 m) library and type 2
sequences in the S18 (6 m) and the S18 (8 m) libraries. We were unable to recover either
type of clone from a library that we created using DNA from sample borehole M3 (6 m).
Here, we report the development and application of a 16S rDNA primer set which was
specifically designed to ampify these sequences in DNA extracted from sediment at the
site. We were unable to amplify these sequences in bulk DNA directly extracted from
Oyster sediment, but through a nested PCR approach we were able to amplify the target
sequences in 16S rDNA from eight of 10 sediments taken from the site. These samples
included 818 (6 m), S9 (6 m), M3 (6 m), B2 (8 m), ODU2 (6 m), ODU2 (7 m), ODU3 (5
m), and ODU3 (6 m). We were unable to amplify DNA from the samples taken from
borehole ODU4.

According to our phylogenetic analysis, NCF A group I sequences diverged from
the bacterial tree early in the history of the kingdom. The species at the base of the
bacterial phylogenetic tree are of interest as their characteristics may point toward the
nature of the ancestor of all bacteria and toward the nature of the universal ancestor.
Unfortunately, initial attempts to acquire this organism in culture have been unsuccessful
(data not shown), so the metabolism of NCFA group I bacteria remains unknown. We do

have an understanding, however, of the environment where these sequences are found,
which can point toward certain characteristics that the group are likely to possess.
Groundwater in the aquifer at the narrow channel site is aerobic, so dominant microbial

groups at the site are likely to respire aerobically or ferrnentatively. Furthermore, the

79

average temperature at the site is 17°C, so NCFA group I is most certainly a mesophile.
It has been argued that because the currently known species at the root of the bacterial
tree are largely thermophilic, the bacterial ancestor was likely a therrnophile (2). NCFA
group I may constitute the earliest mesophiles in the bacterial kingdom, as there are
currently no mesophilic isolates in the divisions at the base of the phylogenetic tree:
Aquifex, Thermodesulfobacteria, C oprothermobacter, and Yhermotogales. If other, non-
thermophilic organisms are found to root deeply in the bacterial tree, lines of reasoning
on the nature of the bacterial ancestor may have to be modified.

It is interesting to note that while NCFA group I sequences were not among the
clones in the bacterial clone library for sediment from borehole M3 (6 m), we were able
to amplify these sequences from M3 (6 m) DNA using the specific primer set. It is
possible that this group was present in low numbers in the M3 (6 m) sediment and that its
168 sequence was not amplified and / or cloned in sufficient numbers because the
community DNA was dominated by other species. This scenario is plausible, as the
terminal 168 restriction fi'agment that corresponds to this group was not detected in M3
(6 m) bacterial T-RFLP profiles while it was detected in both B2 (6 m) and S18 (6 m) and
$18 (8 m) profiles (see Chapter 2).

We were unable to amplify NCFA group I sequences in DNA extracted from
sediments ODU4 (6 m) and ODU4 (7 m), suggesting that this group is not present at this
sampling location or that it may have been present in numbers too low to detect.

It was expected that amplification of bacterial 16S DNA with the NCFA group I
specific primer set would allow us to uncover further diversity within this group. This

was not the case; we were only able to find type 1 sequences among the 144 clones

80

analyzed. Even in sediment samples in which we previously identified only type 2
sequences ($18 (6 m) and S18 (8 m)), we detected only type 1 clones. The clone-specific
primer 984R is an exact match for both type 1 and type 2 sequences and should have
amplified type 2 sequences present in the reactions. In an effort to avoid the bias caused
by chance amplification artifacts, we carried out three separate amplifications of each
template and then combined them before use in the cloning steps, but despite this step,
we only recovered type 1 sequences fi'om each sediment.

Concerns have been raised over the validity of unique sequences detected in the
cloning of environmental DNA. Speksnijder et al. (1 l) have shown that clusters of
unique, closely related rRNA sequences may be introduced in a PCR by errors cause by
the polymerase. Errors include microvariations fiom the original sequence and chimeric
assemblies fi'om two or more sequences. We find it unlikely that NCF A group 1
sequences are artifacts of the amplification process. We were able to extract NCFA
group I sequences in 11 separate clone libraries (three bacterial libraries and eight NCFA-
specific libraries) which were developed using 11 different polymerase chain reactions.
Rather, we propose that NCFA group I sequences represent a widespread population at
the narrow channel focus area and, as such, they likely play a role in biogeochemical
cycling at the site. Future work to isolate these sequences from other subsurface sites in
Oyster or from other environmental media may prove usefiil in establishing the

distribution and importance of this group.

81

.)

L’J

 

10.

ll.

12.

REFERENCES
Amann, R. L, W. Ludwig, and K. H. Schleifer. 1995. Phylogenetic
identification and in situ detection of individual microbial cells without
cultivation. Microbiol-Rev. 59(1): 143-69.
Burggraf, S., G. J. Olsen, K. O. Stetter, and R. Woese Carl. 1992. A
phylogenetic analysis of Aquifex pyrophilus. Systematic and Applied
Microbiology. 15(3):352-356.
Giovannoni, S. J., M. S. Rappé, D. Gordon, E. Urbach, and K. G. Field. 1996.
Ribosomal RNA and the evolution of bacterial diversity, p. pp63-85. In P. S.
D.McL. Roberts, G. Alderson, and M. Collins, Eds (ed), Evolution of Microbial
Life. Cambridge University Press.
Golder-Associates. 1998. Field Sampling Plan, Aerobic flow-cell installation,
narrow channel focus area, South Oyster Site, Oyster, Virginia,
Hugenholtz, P., M. Goebel Brett, and R. Pace Norman. 1998. Impact of
culture-independent studies on the emerging phylogenetic view of bacterial
diversity. Journal of Bacteriology. Sept. 180(18):4765-4774.
Lane, D. J. 1991. l6S/23S rRNA sequencing, p. 115-175. In E. S. a. M.
Goodfellow (ed.), Nucleic acid techniques in bacterial systematics. John Wiley
and Sons, Inc., New York, NY.
Maidak, B. L., J. R. Cole, T. G. Lilburn, C. T. J. Parker , P. R. Saxman, J. M.
Stredwick, G. M. Garrity, B. Li, G. J. Olsen, S. Pramanik, T. M. Schmidt,
and J. M. Tiedje. 2000. The RDP (Ribosomal Database Project) continues.
Nucleic Acids Res. 28: 173-174.
Olsen, G. J., D. J. Lane, S. J. Giovannoni, N. R. Pace, and D. A. Stahl. 1986.
Microbial ecology and evolution: a ribosomal RNA approach. Annu Rev
Microbiol. 40:337-65.
Posada, D., and K. A. Crandall. 1998. MODELTEST: testing the model of
DNA substitution. Bioinformatics. l4((9)):817-81 8.
Schouten, S., E. C. Hopmans, R. D. Pancost, J. Damste, and S. Sinninghe.
2000. Widespread occurrence of structurally diverse tetraether membrane lipids:
Evidence for the ubiquitous presence of low-temperature relatives of
hypertherrnophiles. Proceedings-of-the-National-Academy-of-Sciences-of-the-
United-States-of-America. [print] December 19, 2000;. 97(26):l4421-14426.
Speksnijder, A. G., G. A. Kowalchuk, S. De Jong, E. Kline, J. R. Stephen,
and H. J. Laanbroek. 2001. Microvariation artifacts introduced by PCR and
cloning of closely related 16S rRNA gene sequences. Appl Environ
Microbiol(67):469-472.
Strunk, O., and W. Ludwig. 1997. ARB: Software for phylogenetic analysis.
Technical University of Munich, Munich, Germany.

82

CHAPTER 4
A MOLECULAR COMPARISON OF MICROBIAL COMMUNITIES IN
PRISTINE AND TETRACHLOROETHYLENE-CONTAMINATED
AQUIFER SEDIMEN T
INTRODUCTION

Chemical contamination of groundwater is a serious, ongoing problem in the
United States, where it is estimated that 40% of the population relies on municipally
supplied groundwater and 40 million people draw their household water from private
wells (3). In the last 20 years, it has been observed that the microbial communities within
these formations are capable of transforming many of these substances through intrinsic
bioremediation (3 9), that is, without the addition of exogenous nutrients or organisms.
Substances including crude oil (8), polyaromatic hydrocarbons (31), landfill leachate
(30), trichloroethylene (TCE) (20), dichloroethylene (DCE) (21), and vinyl chloride (VC)
(9) have been shown to be degraded by aquifer microbial communities either in situ or in
microcosms. Intrinsic bioremediation of these and other compounds is a feasible option
for the treatment of many sites, and its success is well documented (10, 14, 15, 45).

Contamination by tetrachloroethylene, or PCE, because of its relatively high
density and recalcitrance in aerobic surface sediments, is a particular problem in
groundwater ecosystems. Sources from Air Force bases to chemical companies to dry
cleaning establishments have been implicated in PCE release and, once released, surface
contamination moves rapidly to the underlying groundwater. The US. Environmental
Protection Agency has determined that of the 992 sites currently listed on the National

Priority List, 344 of them have been found to have contaminated the underlying

83

groundwater with PCE (1). It has been shown that PCE is dechlorinated either partially
or completely by isolates and consortia derived from environmental samples, and
intrinsic bioremediation is a promising mode of remediation for contamination at a
number of sites (10, 15).

The aquifer at the Bachman Road site, located in Oscoda, Michigan, is
contaminated with PCE released from a dry cleaning establishment. Studies indicate that
the indigenous microbial community at the site is responsible for a measured decrease in
PCE in the groundwater observed over a 14-year period (2). Bacteria capable of coupling
the dehalogenation of chlorinated hydrocarbons to respiration, or reductive
dehalogenators, have been isolated from sediments at the site and are thought to carry out
the intrinsic remediation observed there.

Due to the demonstrated potential of intrinsic bioremediation in treating
groundwater at contaminated sites, an increasing amount of work is being done to better
understand the microbial communities responsible for these chemical transformations.
However, because of the difficulty in aseptic collection of subsurface sediment, the
microbial communities in aquifers have been little studied until recently (12).
Furthermore, the communities responsible for intrinsic bioremediation in the
contaminated subsurface are better understood for their functional attributes (compounds
degraded, rate of transformation, limiting nutrient, etc.) than for their species-level
composition. In an effort to further our knowledge of both subsurface communities and
of the communities involved in intrinsic bioremediation, we have undertaken a molecular
comparison of the microbial communities in clean and contaminated aquifer sediment

from the Bachman Road site. We have compared the bacterial communities in the two

84

types of sediments using terminal restriction fragment length polymorphism (T-RFLP) to
uncover broad differences in community diversity. We also developed and compared
bacterial 16S rDNA clone libraries from both pristine and contaminated sediment, and
phylogenetically identified 35 of these clones. Both of these analyses have been directed
toward gaining a better understanding of the key differences between communities
exposed to contamination and those removed from the source and between a community
engaged in active reductive dechlorination of PCB and a community unexposed to the

contaminant.

MATERIALS AND METHODS
Sample collection The study site is located in Oscoda, Michigan, on the shore of Lake
Huron. A single intact core was aseptically collected from each sampling point in May,
1998. Sample site 4 was located in pristine sediment upstream of the plume and sample
sites 1, 2, and 3 were located within the center of the plume. Cores were extracted from
11 to 19 feet below the surface, divided into two-foot segments and homogenized. The
core segments used in these experiments were: 11 - 13 ft (4A, lAt, 2At, and 3At) and 17
- 19 ft (4D, le, 2Bb, and 3Bb). Samples were then divided into acid-washed glass jars .

and held at 4° C for one week, then transferred to gamma-irradiated polypropylene tubes

and held at -80°C until use.

DNA extraction Microbial community DNA for use in T-RFLP and cloning
experiments was extracted from sediment samples by a protocol described by Zhou et al.
(47) with modifications. Briefly, 15 ml extraction buffer and proteinase K (final

concentration 0.04 mg/ml) was added to 15 g of sediment, and the mixture was incubated

85

at 37°C for 30 minutes. Sodium dodecyl sulfate (SDS) solution was added for a final
concentration of 18 mg/ml, and the mixture was incubated at 65°C for 2 hours, followed
by three cycles of freezing in a dry ice and ethanol bath and thawing in a 65°C water
bath. The supernatant was collected and the pellet was washed twice with extraction
buffer and SDS and the washes and supernatant were combined. This liquid was
extracted twice with chloroform and isoamyl alcohol (24:1 vol/vol) and the aqueous
phases were combined. Nucleic acid was precipitated by the addition of 0.6 volume of
isopropanol, then resuspended in water and precipitated by the addition of 2 volumes of
ethanol. The pellet was resuspended in modified TE buffer (10 mM Tris-HCL, 0.1 mM
EDTA, pH 8) and used in polymerase chain reactions.

DNA for real-time PCR reactions was extracted from sediment samples with a
Soil DNA Kit Mega Prep (MoBio) according to the manufacturer’s instructions, with 10
g of sediment in each extraction. DNA in the final elution volume was precipitated by
the addition of 0.04 volumes 5 M NaCl and two volumes of ethanol. The solution was
stored at -20°C overnight. Community DNA was pelleted by centrifugation at 10,000xg
for 30 min. The pellet was dried, resuspended in 50 ul of water, and stored at -20°C until
needed.
16S rDNA T-RFLP For T-RFLP experiments, DNA was extracted from duplicate
samples of sediment 1At, 2At, 2Bb, and 4A. Bacterial 16S ribosomal genes were
amplified from bulk DNA in reaction mixtures that contained 1X PCR buffer (Perkin
Elmer), 2 mM MgC12, 0.2 mM of each dNTP, 0.2 mM of the reverse primer, 0.5 mM of
the forward primer, 8 ng bovine serum albumin per ul, approximately 0.08 ng/ul of

template DNA, and 0.02 U of AmpliTaq (Perkin Elmer) per ul.

86

The bacterial-specific forward primer 27F (5’-AGA GTT TGA TCC TGG CTC
AG—3’) (27) was used for all amplifications of bacterial 16S genes and was labeled at the
5’ end with the phosphoramidite dye S-hexachlorofluorescein (“hex-labeled”) *(Operon
Technologies, Inc.) The reverse primer used in bacterial amplifications was the
universal primer 1392R (5’-ACG GGC GGT GTG TRC-3’) (35). The primers used to
amplify archaeal 16S genes in DNA samples were 21F (5’-TTC CGG TTG ATC CYG
CCG GA-3’) (6) and 915R (5’-GTG CTC CCC CGC CAA TTC CT-3’) (5), both of
which are specific for archaea.

Reaction mixtures were incubated in a GeneAmp 2400 PCR System thermal
cycler (Perkin Elmer) at 94°C for 3 minutes, followed by 35 cycles at 94°C for 45
seconds, 60°C for 30 seconds, and 72°C for 90 seconds and a final extension step of 72°C
for 10 minutes.

Triplicate reaction mixtures for each duplicate extraction were combined and
purified using Wizard“) PCR purification columns (Promega) and eluted with a final
volume of 50 ul of modified TE buffer (0.1 mM EDTA, IOmM Tris, pH 8.0). For
restriction digests, 200 ng of purified bacterial PCR product were digested with 15 U of
HhaI, MspI, or RsaI (BMB) and 200 ng of archaeal products were digested with 15 U of
Hae HI, Hha I, or Rsa I (BMB) at 37°C for 3.5 hours. The lengths of the terminal
restriction fragments from the amplified rDNA products were determined using an ABI
Prism 377 DNA sequencer and ABI software (PE Applied Biosystems) as described by
Liu et al. (28).

TRFLP analyses T-RFLP profiles were analyzed using GeneScan 3.1 software (PE

Applied Biosystems). For enumerating the restriction fragments in each profile, a

87

 

 

 

Ill;

lllCi'

m:
and
re»;
COT:
do:
1113.".
Plas

Sm

111‘. 1':
l

 

fluorescence intensity threshold was set at 50 so that only peaks above this intensity were
included in further analysis. Fragments shorter than 30 base pairs (bp) or larger than 600
bp were excluded from analysis in order to eliminate primer artifacts and to avoid the
problems associated with identifying the length of large fragments. Only peaks present in
both replicate profiles were used for further analysis, and identical peaks were taken to be
those within 0.5 bp size of one another. The Hha 1 profiles were compared using the T-
RFLP Profile Analysis tool available through the Ribosomal Database website
(http://wwwcmemsu.edulRDP/ggis/trflp.cgflsu=SSU), which determines the percent
similarity between two or more profiles. The number of peaks present in both profiles
was divided by the number of peaks in the profile with fewer peaks.
Cloning of bacterial 16S rRNA genes Two sediments from the Bachman road site were
used for cloning experiments: 4D, which was taken up gradient of the plume, and 1Bb,
which came from within the contaminant plume. Bacterial 16S ribosomal genes were
amplified from bulk DNA in reactions that contained 1X PCR buffer (Perkin Elmer), 5
mM MgC12, 1 mM dNTPs, 0.2 mM of each primer, 8 ng bovine serum albumin per ul,
and 0.02 U of AmpliTaq (Perkin Elmer) per ul. The forward primer was 27F and the
reverse primer was 1392R. The amount of template in each amplification and the cycling
conditions were the same as those used for T-RFLP reactions. The PCR products were
cloned using a TOPO TA Cloning Kit (Invitrogen Corp.) in accordance with the
manufacturers instructions. Plasmid DNA was extracted and purified with a Qiagen Mini
Plasmid-prep kit (Qiagen).
Screening of rDNA clones by ARDRA The plasmid inserts of 40 clones from each

library were amplified using the PCR conditions described for cloning, with roughly 30

88

111‘

usi

pm

as

ng of purified plasmid DNA template per 25 ul reaction. Five ul of rDNA PCR products
were digested with 10 U of the 4-base specific restriction enzyme Cfo I in 1X NEB buffer
(New England Biolabs) in a final volume of 15 ul for 3 hrs and 30 min at 37°C. Digested
DNA fragments were separated by acrylamide gel ( 10% acrylamide) electrophoresis and
visualized by silver staining. Restriction fragment lengths were estimated by comparison
with molecular weight standards which ranged from eight to 587 bp (DNA molecular
weight marker V2, Roche Molecular Biochemicals), and plasmids with unique patterns
were selected for partial sequencing of the rDNA insert.
Sequencing of rDNA clones Plasmid inserts from selected rDNA clones were amplified
using the same conditions described above for cloning. Amplified rDNA inserts were
purified by using Ultrafree MC Millipore 30,000 NMWL filter units (Millipore)
according to the manufacturer’s instructions. Sequencing was performed using the ABI
Prism BigDye Terminator Cycle Sequencing Reaction Kit and an ABI Prism 377 DNA
sequencer (PE Applied Biosystems) according to the manufacturer’s directions. Primers
for sequencing included 27F, 355E (5’-ACT CCT ACG GGR SGC AGC-3’) (5) and
536R (36). Forty-two clones were partially sequenced using 27F and five clones that
were of particular phylogenetic interest were fully sequenced.
Phylogenetic analyses All sequenced clones were checked for chimeric sequences using
the Ribosomal Database Project II CHECK_CHIMERA version 2.7 (32). One chimera
. was detected among the sequences and eliminated from fiirther analysis. Sequences were
aligned against close relatives in the Ribosomal Database Project release 8.0 using the
Arb sofiware package (41), and percent similarity to known isolates and previously

cloned sequences was determined. For phylogenetic analysis of sequenced clones, a

89

mask was generated which excluded all positions with alignment gaps or character

uncertainty.

RESULTS

Description of the study site The study site is located in the town of Oscoda, Michigan
and the aquifer at the site is a sandy, unconfined formation which has been contaminated
with tetrachloroethylene (PCE) released from a dry cleaning establishment periodically
over a number of years. Flow within the aquifer is directed eastward toward Lake Huron,
and the hydraulic conductivity of the sediments ranges fi'om 10 to 50 ft/day (2). It has
been determined in previous studies that microorganisms are responsible for a sustained
degradation of PCE observed at the site (2). In the center of the plume, hydrogen and
redox data indicate that a gradient of conditions exists which favors halorespiring
organisms at shallow and intermediate depths (nine to 16 feet below the surface) and
sulfate reducers and methanogens in deeper regions (16 to 20 feet below the surface) (2).
Up gradient of the plume the aquifer is aerobic and oligotrophic.

Profiles of microbial communities in clean and contaminated sediment Profiles of
the bacterial communities in clean and contaminated aquifer sediment revealed a marked
difference in the number of detectable phylotypes in each location. For example, in
digestions with the enzyme Hha I, uncontaminated sediment yielded 5 reproducible
terminal fragments and contaminated sediment 2Bb yielded 8 fragments, 2At yielded 12
fragments, and lAt yielded 26 fragments. The number of ribotype bands in the
uncontaminated profiles was compared to the numbers observed in the contaminated

profiles using a Student’s unpaired t-test. The results indicate that the two data sets are

90

significantly different, that is, there is a 0.026 probability that the two data sets were
generated by chance from the same original data set.

The bacterial profiles were compared on the basis of common peaks using the
online analysis tool provided by the Ribosomal Database Project II (RDP H) available at
http://wwwcmemsu.edu/RDP/gis/trflp.cgi?su=SSU. In comparing the Hha 1 profiles
fi'om clean and contaminated sediment with the RPD II T-RFLP analysis tool, the number
of terminal fragments present in both profiles was divided by the number of fragments in
the profile with the fewest bands. This analysis revealed that all or most of the peaks
present in the uncontaminated profile are present among the fragments in the
contaminated profiles. For example, each of the five terminal fragments observed in the
Hha I uncontaminated community profile (37, 206, 370, 567, and 569 bp) are present in
the 1At and 2At profiles (Figure l), and four of the five were observed in the 2Bb profile.
Certain of these terminal fi'agments that were common to both pristine and contaminated
sediment were detected in all of the community profiles: the 495 bp fragment in the Msp I
profiles and the 476 bp fragment in the Km I profiles. By the RDP similarity measure,
the community of sediment sample 1At was 83% similar to 2At and 75% similar to 2Bb
and the profile of 2At was 75% similar to ZBb. Comparable similarities were observed in
the Msp I and Rsa I digestions (data not shown). F ragrnents present in each of the three
contaminated profiles but absent from the uncontaminated profile were: Hha I fragments

95,381, and 548 bp, Msp I fragments 149, 528, and 530 bp, and Rsa I fragments 57, 459,
and 496 bp.
Using the archaeal-specific primer set, we amplified archaeal 16S genes in DNA

extracted fiom contaminated sediment, but we were unable to amplify archaeal sequences

91

400

200

400

200

110 220 330 440 5?0

 

 

 

 

 

 

 

 

 

 

Figure l A comparison of bacterial community T-RFLP profiles
from contaminated (top - 1At) and pristine (bottom - 4A) aquifer
sediment. In this comparison, each of the five terminal
fragments detected in the pristine sediment is also detected
among the fragments in the contaminated sediment profile.

92

 

 

 

in DNA from uncontaminated sediment. We were able to detect between 10 and 11
terminal restriction fragments in each of the Hae III T-RFLP profiles, between eight and
12 in the Hha profiles, and four to seven in the Rsa I profiles. The Hoe 111 archaeal
profiles of sediments 1At, 2At, and 3At were 91% identical in each pair-wise comparison
(which were carried out as described for the bacterial profiles). These profiles were
markedly different from the 2Bb profile, which had 40%, 50%, and 50% identical
fragments, respectively. The TAP-T-RFLP utility available at the RDP 11 website
http://www.cme.msu.edu/RDP/html/analyses.html was used to identify the archaeal
terminal restriction. fragments in the Bachman profiles which correspond to known
organisms. The profiles of 1At, 2At, and 3At displayed terminal fragments (Hha I 69 and
71, Msp I 196, Rsa I 608 bp) that match those of Methanococcus vanellii (Hha I 69, Msp
I 194, Rsa I 606 bp) and M. voltae (Hha I 71, Msp I 196, Rsa I 608 bp). The profile of
sediment 2Bb presented fragments (Hha I 242, Msp I 327, Rsa I 262 bp) that match those
predicted for M. themolithotrophicus and M. aeolicus (Hha I 242, Msp I 327, Rsa I 263
bp). However, the other terminal fragments detected in the archaeal profiles did not
match the sequences in the RDP II database.
Bacterial diversity in clean and contaminated aquifer sediment The bacterial
diversity in both contaminated and uncontaminated sediment at the Bachman Road site
was evaluated by analyzing 16S rDNA clone libraries from sediments 1Bb and 4D,
respectively. Bulk DNA was extracted from aseptically-collected sediment samples and
the 16S genes therein were amplified using bacterial specific primers. These genes were
then cloned and 38 clones from each library were analyzed by ARDRA. There was no

overlap between the two libraries as measured by ARDRA: none of the ARDRA patterns

93

were detected in both libraries. Among the 38 clones from uncontaminated sediment we
detected 24 distinct ARDRA patterns (63% uniqueness) and in the clone library from
contaminated sediment we detected 25 distinct patterns (66% uniqueness). Fifteen clones
from pristine sediment and 20 clones from the contaminated sediment were selected for
single-pass sequencing. The sequences were aligned to the 16S sequences in the RDP 11
Version 8.0 with the Arb auto align function. Alignments were refined by visual

inspection.

Figure 2 ad depicts five maximum likelihood dendrograms in which clones from
clean and PCE-contaminated sediment are depicted. Figure 2 a depicts the B and y
Proteobacteria clones from the site, Figure 2 b depicts the 6 Proteobacteria clones, and
Figure 2 c depicts the Green nonsulfixr bacterial clones and the clone related to the
putative division OPBBO. Figure 2 (1 depicts the clones that are most closely related to
the A cidobacteria and the division Leptospirillum/Nitrospira. Table 1 lists the Bachman
site 168 clones, organized phylogenetically, their phylogenetic afiiliations in the RDP H
database and their nearest neighbor in the GenBank database as determined by a BLAST
search. Only one clone from the Bachman site showed a high degree of similarity
(>97%) to a previously discovered 16S sequence in the RDP H database: clone 1Bbl was

97% similar to the clone WCHBl-67 (GenBank accession #: AF050536) which was first

discovered at the Wurtsmith Air Force base by Dojka et al. (18).

Clone sequences were digested in silico to determine whether the restriction
fiagments of these clones could be identified among the bands in the corresponding T-
RFLP profiles. Temrinal restriction fragments consistent with those estimated for 5

Proteobacteria] clones 1Bb 2, 1Bb 4, 1Bb 11, and 1Bb 35 (HhaI 94 i2, Msp I 69 :2,

94

1213' 2 a Bachman clone 4D 18—-
{13th clone 4D 11
Bachman clone 4D 2
Hydrogenophaga flava
Rhodoferaxfermentans B
Janthinobacterium lividum

Nitrosomonas marina
Burkholderia brasilensis

Ralstonia eutropha
Thiocapsa roseopersicina

a:
fihromatium okenii 'y

J 1_Achromatium oxaliferum
achman clone 4D 3

 

   
  
  
 
   

 

 

 

 

 

 

 

 

 

 

Campylobacter hominis

0.10

Figure 2 a-d Maximum likelihood dendrograms of bacterial 16S sequences obtained
from the Bachman Road site in Oscoda, Michigan. Divisions are listed out side the
brackets in panels d and c (clone group OPB80, and Green non-sulfur) and
subdivisions are listed in panels a (Proteobacteria), b (Proteobacteria), and c (clone
group T78). Optimality criteria used in bootstrap analysis of the sequences were:
maximum likelihood, maximum parsimony, and neighbor joining. Bifurcations
supported (bootstrap values >75 %) by one optimality criteria but are only marginally
supported (SO-75%) or not supported (<50%) by the other criteria are indicated with
open circles. Bifurcations supported by two or three of the criteria are indicated with
closed circles. The number of characters in each analysis and the range of the mask (E.
coli numbering) were as follows: a) 436 characters, bases 113-600, b) 285 characters,
bases 111-422, c) 414 characters, bases 98-537, and d) 350 characters, bases 134-537.

95

Fig. 2 b

 

 

 

0.10

 

 

 

 

Bachman clone 4D 27

7 Bachman clone 1Bb 4—_
Bachman clone 18b 11

Bachman clone 1Bb 35
yntrophus gentianae
Syntrophus buswellii _,

Bachman clone 1Bb 2
Desulfobacca acetoxidans

 

 

 

Bachman clone 1Bb 1

 

— Desulfacinum infernum
Bachman clone 4D 20

___Bachman clone 1Bb 14
-— Bacillus subtilis

Bachman clone 1Bb 26
achman clone 1Bb 10 Bdellovibrio
Bachman clone 1Bb 24 stolpii

clone BAO7, UEU81641 (anaero. digestor)
Bdellovibrio stolpii
~———Bachman clone 1Bb 18
Desulfuromonas acetoxidans

F Desulfobulbus rhabdoformis
AE'Desulfobulbus propiom'cus Desulfobulbus

clone WCHBl-67, AF050536 (W urtsmith aquifer) group

 

group

Syntrophus group

 

propionicus

 

 

 

96

Fig. 2 C achman clone 1Bb 38 —‘

achman clone 1Bb 31
clone OPBSO, AF027092 (hot spring) Clone group
clone SJA-l7 l, M00950] (bioreactor) OPB8O

achman clone 1Bb 19
clone SJ A-176, AJOO9504 (bioreactor)

Bachman clone 1Bb 27 _

 
 
 
  
  

 

 

— i

 

[— Bachman clone 4D 24
Bachman clone 1Bb 25
clone GCA112, AF154100 Clone group Green non-sulfur
Bachman clone 1Bb 37 T78 bacteria
clone SJA-15, AJOO9453 (bior :actor)

 

 

 

 

) Thermus thermophilus
"C Chloroflexus aggregans

Azospirillum brasilense

0.10

97

Fig. 2 d Acrdobacterzum capsulatum

clone MC22 X68463 (soil)
clone TRB82, AF047646 (pyrite)
Bachman clone 1Bb 30

clone SJA-149, AJ009495 (bioreactor)
clone RB 30, Z95720 (soil)
clone GFPl, AF 130858 (hot spring)
Bachman clone 4D 16
Bachman clone 4D 9
Bachman clone 4D 32
Bachman clone 4D 1
Bachman clone 1Bb 5
Thermodesulfovibrio islandicus
leptospirillumferrooxidans
Nitrospira moscoviensis

 

 

Azorhizobium caulinodans

0.10

98

Acidobacteria

Leptospirillum /
Nitrospira

Table l Phylogenetic affiliations of 16S rDNA clones from the Bachman Road site

and BLAST search results

Clone Division / subdivision
4D 18 BProteobacteria

4D 11 BProteobacteria

4D 2 BProteobacteria
4D 3 y Proteobacteria

le 26 6 Proteobacteria , Bdellovibrio

stolpii group

le 10 6 Proteobacteria , Bdellovibrio

stolpii group

le 24 6 Proteobacteria , Bdellovibrio

. stolpii group
1Bb 18 6 Proteobacteria

1Bb 4 6 Proteobacteria, S yntrophus

group . . ’
1Bb 11 6 Proteobacteria, Syntrophus

. group * ,
1Bb 3 5 6 Proteobacteria, S yntrophus
group ,
1Bb 2 6 Proteobacteria

1Bb ’1 6 Proteobacteria ,
Desulfobulbus propionicus

group
4D 27 6 Proteobacteria
4D 20 6 Proteobacteria
18b 14 , 6 Proteobacteria
le 3’8 ’ Clone group OPB80
1Bb 31 Clone group OPB80

1Bb l9 Clone group OPBSO ’

Closest GenBank relative
Rhizosphere soil bacteria M252688

Rhizosphere soil bacteria AJ252688

Lake sediment clone AF320923
Rhizosphere soil bacteria A123 2812

Marine sediment clone AF3 54153
Marine sediment clone AF 3 541 53

Trichlorobenzene consortium
M 009448
Methanogenic consortium
AF254402
Coal tar waste groundwater
AF 3 5 1220
’ Marine sediment clone AF 35 1238

Coal tar waste groundwater
AF 35 1238

Desulfobacca acetoxidans

AF 00267 1 p .,
WCHBl-67 Wurtsmith aquifer
clone AF050536

Agricultural soil bacterium

A80252620

F erromanganous micronodule clone
, AF 293008 ’

Coal tar waste groundwater

AF351231 . j

Trichlorobenzene consortium

AJOO9504

Trichlorobenzene consortium

AJOO9504 . ,

Trichlorobenzene

consortiumAJOO9501

99

Score

643

814

954
512

683

701

344

388

1271

738

2313

1456

1863

772

486

738

655

662

377

Table 2 (Continued)

 

 

Clone Division / subdivision Closest GenBank relative Score
1Bb 27 Clone group OPBSO Trichlorobenzene consortium 500
. AJOO949O
4D 24 Green non-sulfur , Clone group River sediment clone AF 141416 592
T78
1Bb 25 Green non-sulfilr , Clone group Benzene consortium clone 648
T78 AF 029149.]
le 37 Green non-sulfur , Clone group Trichlorobenzene consortium 671
T78 AJOO9453
13b 30 Fibrobacter / Acidobacteria, Soil clone AF 010095 370
Acidobacteria group
4D 16 Fibrobacter /Acidobacteria, Soil clone Z95 732 759
Acidobacteria, Clone group
RB4O
4D 9 Fibrobacter / Acidobacter, Agricultural soil bacterium 978
Acidobacteria, Env. Clone AJ252654
MB1228 group
4D 1 Fibrobacter /Acidobacteria, PCB-contaminated soil clone 569
Acidobacteria group A] 292585
4D 32 Fibrobacter /Acidobacteria, PCB-contaminated soil clone 648
Acidobacteria group AJ 292585
18b 5 Leptospirillum / Nitrospira Lake sediment clone AF 320959S1 785
4D 38 Undefined Deep subsurface clone AF 005747 442
1Bb 3 Undefined Desulfovibrio sp. "Bendigo A" 327
AF131324
1Bb 28 Undefined Grassland soil AF0783 79 266
le 6 Undefined Aerotherrnobacter marianas 240
AB011495
4D 30 ’ Undefined Aquatic clone AF 3 17743 901
4D 12 Undefined Uranium waste clone M296569 422
13b 33 Undefined Unidentified therrnophile AJ131537 336
100

164 :2, 511 i2, and Rsa I 57 i2 bp) were detected among the bands in profiles of
communities from contaminated sediment (Hha I 95 bp, Msp 67, 164, 512, and Rsa I 57).
These clones were limited to the library created from contaminated sediment DNA and
their signature restriction fragment sizes were not detected in the T-RFLP profile from
pristine sediment. We were unable to consistently detect (i.e. in all digestions) the

restriction fragments of clones from pristine sediment in the T-RF LP profiles.

DISCUSSION

T-RFLP was used to generate replicated profiles of the bacterial communities in
three contaminated sediments and one pristine sediment from the Bachman Road site.
The number of ribotypes detected was consistently higher in contaminated sediment than
in sediment taken upstream of the PCE plume. Ribotype diversity detected by T-RFLP
may be used to derive an estimate of absolute bacterial diversity in a sample, or to
determine the relative diversity among a group of samples (3 3). The detection of greater
numbers of terminal fragments in contaminated sediment profiles is an indication that a
relatively higher bacterial diversity exists in PCE-contaminated sediment at the Bachman
Road site.

The effect of chemical contamination on the diversity of bacterial communities
has yet to be clearly resolved (42). In most cases, the influence of contamination appears
to be due either to the toxicity of the material (11, 26, 37), to the utility of the
contaminant as a nutrient or carbon source (13), or to a combination of these effects (17,
25) but in any case, the detailed effects of either type of contaminant are not predictable.

There are indications that PCE and its degradation products may be toxic to bacterial

101

populations (4), but despite its potential toxicity, certain species are capable of linking
dechlorination of PCB and TCE to energy conservation (34). Active microbial
dechlorination of PCB has been detected at the Bachman site, and is the dominant
terminal electron-accepting process in sediment at shallow and intermediate depths in the
aquifer (as determined by measuring geochemical conditions) (2). Furthermore, PCE
dumped at the Bachman site likely contained the co-contaminants usually present in spent
dry cleaning PCE (2) including lipids, waxes, hydrocarbon solvents, detergents, and
starch which may serve as carbon sources for the microbial community. Hence, it is
possible that both toxicity and nutrient effects may influence diversity patterns at the
Bachman Road site. The increase in ribotype diversity could, then, be due in part to the
co-contaminant carbon sources present in the waste PCE, as they may serve to boost the
numbers of formerly small heterotrophic populations and have the overall effect of
increasing the observed genetic diversity. PCE, too, could have affected diversity by
serving as an electron acceptor for reductively dechlorinating populations which were
present in only small numbers in pristine sediment. The new advantage gained by these
populations would foresee ably increase their numbers and render them detectable with
T-RFLP. The compounds that both heterotrophs and dechlorinators release can, in turn,
serves as nutrients for other populations, creating ripples down the food chain.

Most of the ribotypes observed in the profile of the uncontaminated sediment
community at the Bachman site were detected among the ribotypes found in the
contaminated sediment, an observation which may be explained in two ways.
Populations which were dominant in the Bachman sediment prior to exposure to PCE

may have persisted in the altered environment where they either carried out their normal

102

metabolic fiinctions or adjusted their functions to suit the new nutrients and carbon
sources. While the number of bacteria in these populations remained steady, the numbers
of bacteria belonging to other groups that are capable of exploiting the new resources
may have expanded. This is consistent with the fact that the nutrients and carbon sources
present in the groundwater are not displaced by infiltrating PCB and co-contaminants but,
rather, are supplemented by these additional sources. Alternatively, some or all of the
dominant groups in pristine sediment may have been sensitive to the toxic effects of PCB
and either ceased metabolism or died while smaller populations and invading species
increased in numbers. Unfortunately, it is impossible to distinguish between these two
hypotheses, since one drawback of the direct extraction of DNA is that dead and inactive
populations may be detected (22), as their DNA is extracted and amplified along with
that of active, viable cells.

Whereas the T-RF LP profiles from clean and contaminated sediment convey a
degree of overlap between the two communities, the clone libraries are clearly very
distinct. Among the 49 ARDRA patterns identified, no patterns common to both libraries
were detected, suggesting a significant shift in the community structure. Considering the
differences in geochemical conditions between the sediments 4D and 1Bb, namely
difi’erences in oxygen saturation, available carbon, and PCB concentrations, this is not an
unexpected result. This contrast was reflected in the phylogenetic groups identified
among the clones from the site. For example, we were able to detect one 7 and three B
Proteobacteria] clones in the uncontaminated sediment, but these groups were not
represented among the clones from contaminated sediment. Furthermore, we detected a

number of sequences which are closely related to 6 Proteobacteria in both of the clone

103

 

libraries, but only sequences from the contaminated sediment library grouped with known
sulfate reducers and anaerobic syntrophs. Clone 1Bb 1 was identical to the 16S sequence
of a sulfate reducing isolate, STP23 (AJOO6620), discovered in the sediment of an
oligotrophic lake (3 8). The isolate is capable of using hydrogen, acetate, formate,
propionate, pyruvate, lactate, succinate, and ethanol as electron donors and sulfate,
sulfite, and thiosulfate as electron donors. This species may be of particular significance
in environments undergoing intrinsic bioremediation, as a nearly identical (97%) 16S
sequence was also derived from the methanogenic zone of a contaminated aquifer less
than five miles away at the Wurtsmith Air Force Base (18). (We have explored a method
to evaluate the numbers of this species in sediment from the site; see Chapter 5.)
Sequences related to the 6 Proteobacteria genus Syntrophus have also been identified in
the clone library fi'om contaminated sediment. Below, we will discuss the role that
syntrophs may play in the cycling of carbon at the Bachman site. While the differences
between the clone libraries from clean and contaminated aquifer sediment may reflect a
real divergence between the two sites, it is also possible that our sampling of these
diverse communities was too modest to detect the presence of a significant overlap. The
similarity of the T-RFLP profiles from the two sites is consistent with this possibility.

Five of the Bachman clones, one from contaminated sediment and four from
pristine sediment, were shown to have a reproducible phylogenetic affiliation with the
proposed division Acidobacteria and were 80-85% identical to Acidobacterium
capsulalum. Many representative sequences of this division have been identified in
various environments from marine sediment (46) to volcanic cinders and soil (7), and it is

thought to play a role in many communities. Unfortunately, there is only one cultivated

104

representative to represent this group so the unifying points of metabolism among these
organisms are unknown and we cannot extrapolate the characteristics of the
Acidobacteria-related populations at the Bachman site.

Clone sequences related to both Clone Group T78 (within the division Green-non
sulfur bacteria) and Clone Group OPB80 sequences were likewise found in both
libraries. Originally isolated from activated sludge (40), Clone Group T78 has
representatives from a hot spring (24), an anaerobic bioreactor (44), and lake ice (23).
Clone Group OPB80 sequences come from equally diverse environments (24, 44). As in
the case of Acidobacterium-related clones, no unifying metabolism has been identified
for this group, so it is impossible to make estimates of the processes that these species
carry out in the Bachman aquifer.

Using archaeal-specific primers, we were able to amplify archaeal 16S genes in
bulk DNA extracted from sediment within the plume, but we were unable to amplify
archaeal sequences in DNA from uncontaminated sediment. This suggests that archaea
were either present in very low abundance or absent altogether from the aerobic,
oligotrophic sediment upstream of the plume. Contaminated sediments, on the other
hand, are anoxic and have elevated organic carbon concentrations. Indeed, archaea are
expected to dominate the microbial community in deeper sediments, as geochemical data
indicates that methanogenesis and ’sulfate reduction are the dominant terminal electron
accepting processes in this zone of the aquifer (2). Analysis of archaeal T-RFLP profiles
suggests that Methanococcus species or a closely related but hitherto unidentified genera

may be an important group at the site. Methanococcus isolates identified to date have

105

A—n-l

been found to utilize hydrogen and formate as electron donors and to generate methane
strictly by autotrophic means, that is, through uptake of C02 and not acetate.

It is interesting to note the rift between archaeal T-RFLP profiles fi'om shallower
sediments 1At, 2At, and 3At (ll — 13 ft) and the profile from the deeper sediment 2Bb
(17 - 19 fl). Geochemical measurements at the site indicated that conditions favoring
reductive dehalogenators predominate in the zone where 1At, 2At, and 3At were taken
while sulfate-reducing to methanogenic conditions predominate where 2Bb was drawn
(2). We were able to detect archaea, presumably methanogens, in the shallow sediments
dominated by reductive dehalogenators, however, T-RFLP profiles of archaea show that
these archaeal populations are different from those seen in the deeper sediments where
methanogenesis has a stronger influence.

Knowing the phylogenetic affiliations of certain Bachman clones allows us to
speculate about the cycling of nutrients within the deeper regions of the plume. The
presence of sequences closely related to the genera Syntrophus is a particularly intriguing
point, indicating that interspecies hydrogen or organic acid transfer may be an important
process. Syntrophus-related species in the plume likely metabolize organic acids,
producing acetate and C02 as well as H2 which can be taken up and used as a valuable
electron donor by other species. Both geochemical indicators and the presence
of 6 Proteobacteria] clones which are closely related to known sulfate-reducers indicate
that sulfate reducers are present in this sediment as well. The sulfate-reducing isolates
most closely related to our clones, Desulfobacterium (13b 2) and STP23 (le 1), are
known to use hydrogen and simple acids as electron donors and they may be coupled via

transfer of these compounds to the Syntr0phus-related species at the site. Furthermore,

106

T-RFLP profiles of the archaeal community indicate that Methanococcus or a related
species may be present in the contaminated sediments. The geochemical indicators at the
site favor the presence of methanogens as well. Methanococcus and methanogenic mixed
cultures isolated from other contaminated aquifers (8) are known to carry out autotrophic
methanogenesis from C02 and H2 or other electron donors. Thus, if autotrophic
methanogenesis is the dominant type of methanogenesis at the site, then H2 and C02
produced by Syntrophus-related species can be taken up by these archaea. Alternatively,
if the methanogens at the site are acetoclastic they may use the acetate provided by the
Syntrophus—related species to drive methane production, leaving H2 to the sulfate-
reducers. Hence, depending on which type of methanogen is dominant at the site,
methanogens may have been competing with sulfate reducers for H2 (autotrophs) or
acetate (acetoclasts).

In their investigation of the bacterial and archaeal communities at the nearby
Wurtsmith AFB site, Dojka et a1. (18) likewise discovered Syntrophus-related clones and
uncovered archaeal clones closely related to known acetoclastic isolates. The authors
proposed that the two groups interacted via transfer of acetate and rejected the hypothesis
that autotrophic methanogens were involved in the process. The authors cite previous
work in mesophilic and psychrophilic environments indicating that acetoclastic
methanogenesis dominates at temperatures below 20°C and in these cases available
hydrogen is taken up by homoacetogens. We do find it unlikely, however, that CO2 and
H2 are funneled to homoacetogenic populations in the Bachman aquifer, as we were
unable to identify sequences related to known homoacetogens in our clone library.

Further, and it has been shown that homoacetogens are at a thermodynamic disadvantage

107

and are hence incapable of out competing methanogens and sulfate reducers for H2 except
in a limited number of cases (16).

Furthermore, it should be noted that whatever organisms were the recipients of
interspecies transfer of organic acids and hydrogen, they may have been involved in the
dechlorination of PCB to TCE, cDCE, VC, and ethene which has been detected in this
sediment. Reductive dehalogenators have been isolated from the Bachman Road site,
where they are thought to be the dominant organisms carrying out these transformations
(2). In microcosm studies using sediment from the site, it has been found that the activity
of these processes is limited by electron donors, namely lactate for the conversion of PCB
to cDCE and H2 for the conversion of cDCE to ethene (2). It is possible that Syntrophus-
related species provide the necessary hydrogen for reductive dehalogenators to carry out
the transformation of cDCE to ethene. While reductive dehalogenators are at a
thermodynamic advantage over sulfate reducers and methanogens in scavenging H2 and
should out compete these organisms, reductive dechlorination of chlorinated ethenes has
been observed in methanogenic mixed cultures (4, 19, 43) and this activity may not be
due to the activity of methanogenic populations (29). So, if reductive dehalogenators
were involved in syntrophy in the deeper sediments of the Bachman site, they likely
scavenged H2 from Syntrophus-related species while sulfate reducers and acetoclastic
methanogens competed for syntrophically-produced acetate.

In summary, T-RFLP analyses of the bacterial communities in sediment from the
Bachman Road site revealed a distinct increase in phylogenetic diversity with the onset of
contamination by PCE and co-contarninants, an effect which was likely brought about by

higher concentrations of organic carbon available in the contaminated sediments.

108

Furthermore, we detected all or most of the terminal restriction fragments from the
uncontaminated sediment profiles among the terminal fragments in the contaminated
profiles, indicating that either the dominant populations in clean sediment persisted with
the onset of contamination or that the limitations of the method permitted the extraction
and detection of dead and inactive populations. In contrast, the clone libraries did not
reveal an overlap between the two communities, but instead suggested that a significant
shift in the microbial community structure took place with the onset of contamination.
Overall, both results point to a dramatic change in the composition of the microbial
community and it is possible that any overlap that may exist between the two sites was
not detected in our relatively small clone libraries. Finally, the discovery of 16S rDNA
genes within the plume that are closely related to the genus Syntrophus indicated that
other species in the sediments may have been involved in the syntrophic exchange of
hydrogen, C02, and acetate. It is possible, in such a scenario, that reductive
dehalogenators were the beneficiaries of interspecies hydrogen transfer, allowing the
dechlorination of cDCE to ethene. If this is the case, then efforts to ameliorate
contamination at the Bachman site and others may be served by providing Syntrophus
species with small organic acids required for the production of hydrogen for reductive
dehalogenators. Further work to better define the relationship between these populations
should be carried out to clarify the best mode of action to treat contamination by

chlorinated hydrocarbons.

109

REFERENCES

1.

2.

10.

ll.

12.

13.

14.

US. Environmental Protection Agency Superfund Website -
http://www.epa.gov/superfund/sites/query/basic.htm.

Adriaens, P., L. Abriola, M. Barcelona, B. Fathepure, K. Hayes, J. Tiedje, P.
Kurt, F. Lofller, and E. Petrovskis. 1998. Application of surfactant-enhanced
aquifer remediation (SEAR) and enhanced halorespiration technologies at the
Bachman Road residential wells site, Oscoda, M., A proposal submitted to the
Michigan Department of Environmental Quality.

Alley, W. A., Thomas E. Reilly, and 0. L. Franke. 1999. Sustainability of
ground-water resources, US. Geological Survey Circular 1186, US. Geological
Survey.

Alvarez-Cohen, L., and P. L. McCarty. 1991. Effects of toxicity, aeration, and
reductant supply on trichloroethylene transformation by a mixed methanotrophic
culture. Applied and Environmental Microbiology. 57(1):228-235.

Amann, R. I., L. Krumholz, and D. A. Stahl. 1990. Fluorescent-oligonucleotide
probing of whole cells for determinative, phylogenetic, and environmental studies
in microbiology. Journal of Bacteriology. l72(2):762-770.

Amann, R. 1., W. Ludwig, and K. H. Schleifer. 1995. Phylogenetic
identification and in situ detection of individual microbial cells without
cultivation. Microbiol-Rev. 59(1): 143-69.

Barns, S. M., S. L. Takala, and C. R. Kuske. 1999. Wide distribution and
diversity of members of the bacterial kingdom Acidobacterium in the
environment. Applied and Environmental Microbiology. April, 1999;.
65(4):]731-1737.

Bekins, B. A., E. M. Godsy, and E. Warren. 1999. Distribution of microbial
physiologic types in an aquifer contaminated by crude oil. Microbial-Ecology.
May, 1999;. 37(4):263-275.

Bradley, P. M., F. H. Chapelle, and J. T. Wilson. 1998. Field and laboratory
evidence for intrinsic biodegradation of vinyl chloride contamination in a Fe(III)-
reducing aquifer. Journal of Contaminant Hydrology. May. 31(1-2):111-127.
Brigmon, R. L., N. C. Bell, D. L. Freedman, and C. J. Berry. 1998. Natural
attenuation of trichloroethylene in rhizosphere soils at the Savannah River Site.
Journal of Soil Contamination. July. 7(4):433-453.

Castro, 1. V., E. M. Ferreira, and S. P. McGrath. 1997. Effectiveness and
genetic diversity of Rhizobium leguminosarum bv. Trifolii isolates in Portuguese
soils polluted by industrial effluents. Soil Biology and Biochemistry. 29(8):]209-
1213.

Chapelle, F. H. 1993. Ground-Water Microbiology and Geochemistry. John
Wiley & Sons, Inc., New York.

Cho, J.-C., and S.-J. Kim. 2000. Increase in bacterial community diversity in
subsurface aquifers receiving livestock wastewater input. Applied-and-
Environmental-Microbiology. March, 2000;. 66(3):956-965.

Cho Jong, S., J. T. Wilson, D. C. Digiulio, J. A. Vardy, and W. Choi. 1997.
Implementation of natural attenuation at a JP-4 jet 1118] release after active
remediation. Biodegradation. 8(4):265-273.

llO

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

Clement, T. P., C. D. Johnson, Y. Sun, G. M. Klecka, and C. Bartlett. 2000.
Natural attenuation of chlorinated ethene compounds: Model development and
field-scale application at the Dover site. Journal-of-Contaminant-Hydrology.
March 31, 2000;. 42(2-4):113-140.

Cord-Ruwisch, R., H. J. Seitz, and R. Conrad. 1988. The capacity of
hydrogenotrophic anaerobic bacteria to compete for traces of hydrogen depends
on the redox potential of the terminal electron acceptor. Archives of
Microbiology. l49(4):350-357.

Derry, A. M., W. J. Staddon, and J. T. Trevors. 1998. Functional diversity and
community structure of microorganisms in uncontaminated and creosote-
contaminated soils as determined by sole-carbon-source-utilization. World
Journal of Microbiology and Biotechnology. July. l4(4):57 1-57 8.

Dojka, M., A., P. Hugenholtz, S. Haack, K., and N. Pace, R. 1998. Microbial
diversity in a hydrocarbon- and chlorinated-solvent-contaminated aquifer
undergoing intrinsic bioremediation. Applied and Environmental Microbiology.
Oct. 64(10):3869-3877.

Freedman, D. L., and J. M. Gossett. 1989. Biological reductive dechlorination
of tetrachloroethylene and trichloroethylene to ethylene under methanogenic
conditions. Applied and Environmental Microbiology. 55(9):2144-2151.

Fries, M., R., L. Forney, J., and J. Tiedje, M. 1997. Phenol- and toluene-
degrading microbial populations from an aquifer in which successfirl
trichloroethene cometabolism occurred. Applied and Environmental
Microbiology. 63(4): 1523-1530.

Gao, J., and R. S. Skeen. 1999. Glucose-induced biodegradation of cis-
dichloroethylene under aerobic conditions. Water-Research. Aug, 1999;.
33(12):2789-2796.

Gasol, J. M., U. L. Zweifel, F. Peters, J. A. Fuhrman, and A. Hagstrom. 1999.
Significance of size and nucleic acid content heterogeneity as measured by flow
cytometry in natural planktonic bacteria. Applied-and-Environmental-
Microbiology. Oct., 1999;. 65(10):4475-4483.

Gordon, D. A., J. Priscu, and S. Giovannoni. 2000. Origin and phylogeny of
microbes living in permanent Antarctic lake ice. Microbial-Ecology. April. '
39(3): 197-202.

Hugenholtz, P., C. Pitulle, K. L. Hershberger, and N. R. Pace. 1998. Novel
division level bacterial diversity in a Yellowstone hot spring. Journal of
Bacteriology. Jan. 180(2):366-376.

Juck, D., T. Charles, L. G. Whyte, and C. W. Greer. 2000. Polyphasic
microbial community analysis of petroleum hydrocarbon-contaminated soils from
two northern Canadian communities. FEMS-Microbiology—Ecology. [print]
September, 2000;. 33(3):241-249.

Kandeler, E., C. Kampichler, and 0. Horak. 1996. Influence of heavy metals
on the functional diversity of soil microbial communities. Biology and Fertility of
Soils. 23(3):299-306.

Lane, D. J. 1991. 16S/23S rRNA sequencing, p. 115-175. In E. S. a. M.
Goodfellow (ed.), Nucleic acid techniques in bacterial systematics. John Wiley
and Sons, Inc., New York, NY.

111

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

Liu, W. T., T. L. Marsh, H. Cheng, and L. J. Forney. 1997. Characterization of
microbial diversity by determining terminal restriction fragment length
polymorphisms of genes encoding 16S rRNA. Applied and Environmental
Microbiology. Nov. 63(11):4516-4522.

Loffler, F., K. Ritalahti, and J. Tiedje. 1997. Dechlorination of Chloroethenes
Is Inhibited by 2-Bromoethanesulfonate in the Absence of Methanogens. Appl.
Environ. Microbiol. 63(12):4982-4985.

Ludvigsen, L., H. J. Albrechtsen, D. B. Ringelberg, F. Ekelund, and T. H.
Christensen. 1999. Distribution and composition of microbial populations in a
landfill leachate contaminated aquifer (Grindsted, Denmark). Microbial-Ecology.
April, 1999;. 37(3): 197-207.

Madsen, E. L., J. L. Sinclair, and W. C. Ghiorse. 1991. In situ biodegradation:
Microbiological patterns in a contaminated aquifer. Science. 252(5007):830-833.
Maidak, B. L., J. R. Cole, T. G. Lilburn, C. T. J. Parker , P. R. Saxman, J. M.
Stredwick, G. M. Garrity, B. Li, G. J. Olsen, S. Pramanik, T. M. Schmidt,
and J. M. Tiedje. 2000. The RDP (Ribosomal Database Project) continues.
Nucleic Acids Res. 28:173-174.

Marsh, T. L. 1999. Terminal restriction fragment length polymorphism (T-
RFLP): an emerging method for characterizing diversity among homologous
populations of amplification products. Curr Opin Microbiol. 2(3):323-7.

McCarty, P. L. 1997. Breathing with chlorinated solvents. Science. Jun 6;.
276(5318):1521-2.

Olsen, G. J., D. J. Lane, S. J. Giovannoni, N. R. Pace, and D. A. Stahl. 1986.
Microbial ecology and evolution: a ribosomal RNA approach. Annu Rev
Microbiol. 40:337-65.

Pace, N. R. S., D.A.; Lane, D.J.; Olsen, G.J. 1986. The analysis of natural
microbial populations by ribosomal RNA sequences. Adv. Microb. Ecol. 9:Jan-
55.

Sandaa, R, Anne, V. Torsvik, 0. Enger, L. Daae Frida, T. Castberg, and D.
Hahn. 1999. Analysis of bacterial communities in heavy metal—contaminated
soils at different levels of resolution. FEMS-Microbiology-Ecology. Nov., 1999;.
30(3):237-251.

Sass, H., E. Wieringa, H. Cypionka, H.-D. Babenzien, and J. Overmann.
1998. High genetic and physiological diversity of sulfate-reducing bacteria
isolated fiom an oligotrophic lake sediment. Archives of Microbiology. Oct.
l70(4):243-251.

Sims, J. L., J. M. Suflita, and H. H. Russell. 1992. Ground Water Issue; In-situ
Bioremediation of Contaminated Ground Water, EPA/540/S-92/003, Superfund
Technology Support Center for Ground Water, US. Environmental Protection
Agency.

Snaidr, J., R. Amann, L Huber, W. Ludwig, and K. H. Schleifer. 1997.
Phylogenetic analysis and in situ identification of bacteria in activated sludge.
Applied and Environmental Microbiology. 63(7):2884-2896.

Strunk, 0., and W. Ludwig. 1997. ARB: Software for phylogenetic analysis.
Technical University of Munich, Munich, Germany.

112

42.

43.

44.

45.

46.

47.

Trevors, J. T. 1998. Bacteria biodiversity in soil with an emphasis on
chemically-contaminated soils. Water Air and Soil Pollution. 101 Jan.(1-4):45-
67.

Vogel, T. M., and P. L. McCarty. 1985. Biotransformation and
tetrachloroethylene to trichloroethylene, dichloroethylene, vinyl chloride and
carbon dioxide under methanogenic conditions. Applied and Environmental
Microbiology. 49(5): 1080-1083.

von Wintzingerode, F., B. Selent, W. Hegemann, and U. B. Goebel. 1999.
Phylogenetic analysis of an anaerobic, trichlorobenzene-transforrning microbial
consortium. Applied and Environmental Microbiology. Jan. 65(1):283-286.
Weiner, J. M., and D. R. Lovley. 1998. Rapid benzene degradation in
methanogenic sediments from a petroleum-contaminated aquifer. Applied and
Environmental Microbiology. May. 64(5): 1937-1939.

Wise, M., G., J. V. McArthur, and J. Shimkets Lawrence. 1997. Bacterial
diversity of a Carolina bay as determined by 16S rRNA gene analysis:
Confirmation of novel taxa. Applied and Environmental Microbiology.
63(4): 1505-1514.

Zhou, J., M. A. Bruns, and J. M. Tiedje. 1996. DNA recovery from soils of
diverse composition. Applied and Environmental Microbiology. 62(2):316—322.

113

CHAPTER 5
QUANTIFICATION 0F BACTERIAL, ARCHAEAL, AND SPECIES-SPECIFIC
16S GENES IN TETRACHLOROETHYLENE-CONTAMINATED AQUIFER
SEDIMEN T USING REAL-TIME PCR
INTRODUCTION

Detecting the presence and abundance of particular microbial populations in
environmental media can aid in evaluating the relative importance of that group in an
ecosystem. Where chemical contamination has affected the microbial community,
detecting the presence of certain groups can serve to achieve a better understanding of the
processes that the community performs (15) or to measure ecological risk (37). Intrinsic
bioremediation, the remediation of chemical contamination in the environment by the
indigenous microflora, is becoming a widely-used alternative for the treatment of such
sites (IO-12, 46, 48). Gaining a better understanding of the abundance and distribution of
microbial species that are active in these remediation processes affords researchers with
another tool to evaluate the potential effectiveness of the indigenous community.
Methods currently used for the purpose of detecting microbial populations of interest,
including rRNA hybridization analysis, microscopy, cultivation, and quantitative PCR,
each have their own individual set of advantages and limitations.

The emerging method of real-time PCR offers another approach to detecting and
quantifying microbial populations. First developed by Higuchi et a1 (25), the method
entails tracking the amplification of a target sequence in a polymerase chain reaction.
The increasing number of target sequences may be signaled through the release of an

intragenic fluorescent probe or through the use of a non-specific fluorescent nucleic acid

114

dye. The initial number of targets in the reaction can be extrapolated through comparing
the evolution of the fluorescent signal with known standards. Real-time PCR has been
used to quantify a range of DNA sequences, including nitrite reductase (22), glycoprotein
D of human immunodefficiency virus (25), meningococcal-specific genes IS1 106 , ctrA,
and siaD (23), and large (28) and small subunit rRNA (13, 28, 38, 43, 47).

A prior study of the microbial communities in sediment from the Bachman Road
site, a tetrachlorothylene (PCE) -contaminated aquifer in Oscoda, Michigan which is
undergoing intrinsic bioremediation, has uncovered the 16S sequence of a bacterium
closely related to sulfate-reducing species of the 6 Proteobacteria (see Chapter 4).
Indeed, an identical 165 sequence has since been found in a sulfate-reducing isolate from
lake sediment (41), confirming the presumed metabolism of this organism. In our
examination of the community in contaminated aquifer sediment, we found this sequence
to comprise 26% of the 16S sequences in our bacterial clone library. In a separate study,
a nearly identical 16S sequence was discovered in an aquifer at the former Wurtsmith Air
Force base, approximately five miles away from our study site (15). There is evidence
that intrinsic bioremediation is active in the Wurtsmith aquifer, which is contaminated
with hydrocarbons and chlorinated solvents, including PCE. The discovery of this clone
at two sites undergoing intrinsic bioremediation suggested that this organism might play a
role, either directly or indirectly, in the treatment of these sites.

As the process of cloning can introduce biases into the analysis of microbial
communities (27, 44), we were incapable of judging the true representation of the
organism 1Bb1 in sediment at the Bachman and Wurtsmith sites from the composition of

the clone libraries. In this study, we designed a species-specific 16S rDNA primer set to

115

detect and quantify lel in sediment from the Bachman and Wurtsmith sites through
real-time PCR. Furthermore, using kingdom-specific primers, we were able to measure
the numbers of bacterial, archaeal, and eukaryal SSU rRNA genes in these samples and in

sediment from a pristine aquifer, activated sludge, and bioreactor fluid.

MATERIALS AND METHODS

Sample collection The Bachman Road site is located in Oscoda, Michigan, on the shore
of Lake Huron. A single intact core was aseptically collected fiom each point where
injection and extraction wells were installed in August, 2000. Two injection wells (north
and south) and one extraction well were installed and the cores from these points were
named accordingly: IW-N, IW-S, and EW. Sampling points were arrange in a triangle:
IW-N and IW-S were six feet apart, arranged perpendicularly to the hydrologic flow.
Sample point EW was located 10 feet down gradient from the midpoint between the
injection wells. Cores were extracted from 11 to 19 fl below the surface, divided into
two-foot segments and homogenized. Samples were then divided into acid-washed glass
jars and held at 4° C for one week, then transferred to gamma-irradiated polypropylene
tubes and held at -80°C until use. The core segments used in these experiments were:
IW-N11-13 ft,EW11 - 13 fi,IW-N17-l9ft,IW-S 17—l9ft,andEW17-19fi.

Samples ML3 (15.5-17.5 ft) and ML3 (23.5-27.5 ft) from the Wurtsmith site were
aseptically collected in November, 1999 in the manner described by Dojka et al. (15)
from a point less than 10 fi downgradient of the extraction point used in that study.
Sediment from Oyster, Virginia, ODU4 8.8-8.9 m was extracted in August 2000 as

described in Chapter 2. Bioreactor sample LS1 was extracted from the bioreactor

116

described by Fernandez et al. (19) in September, 1999 and stored at -80° C until use. The
activated sludge sample was extracted from the benchtop reactor described by Marsh et
al. (32) and stored at -80° C until use.
DNA extraction DNA was extracted from sediment samples with a Soil DNA Kit Mega
Prep (MoBio) according to the manufacturer’s instructions. We used 10 g of sediment in
each extraction except in the case of the two Oyster sediment samples where we used 15
g. DNA in the final elution volume was precipitated by the addition of 0.04 volumes 5 M
NaCl and two volumes of ethanol. The solution was stored at -20°C overnight.
Community DNA was pelleted by centrifugation at 10,000xg for 30 min. The pellet was
dried, resuspended in 50 ul of water, and reprecipitated using 0.04 volumes of 5M NaCl
and two volumes of ethanol. Again, the pellet was dried, resuspended in 50 ul of water,
and stored at 4°C. In order to examine the reproducibility of the method two sediment
samples were selected for replication: two 10 g samples were extracted for sediments
IW-N (11 — 13 ft) and IW-S (l7 — 19 ft).

DNA was extracted from bioreactor and activated sludge samples with a Soil
DNA Isolation Kit (MoBio) according to the manufacturer’s instructions. Bioreactor and
activated sludge DNA was stored at 4° C until use. In order to examine reproducibility in
the non-sediment samples, a single aliquot of bioreactor material was divided into two
separate samples prior to extraction and analysis: 1.5 ml of bioreactor fluid was pelleted,
resuspended in 200 ul of water and divided. The same division was carried out with the
activated sludge sample, wherein a single pellet of 0.25 g was resuspended in 200 ul and

divided.

117

Primer design The primer pairs for real-time PCR are listed in Table 1. Primers were
selected to Optimize base pairing with target sequences, achieve a Tm close to 60° C, and
minimize amplicon length. The 1Bb1-specific primer 712R was designed using the
probe design function in the Arb software package (42).

Real-time PCR The amplification of target sequences during the PCR was detected
using the 7700 Sequence Detector (PE Applied Biosystems) to monitor the increase in
fluorescence caused by the binding of SYBR Green dye (PE Applied Biosystems) to
strands of double-stranded nucleic acid. The fluorescence of SYBR Green was
normalized by comparison with an internal fluorescent reference dye, 6-carboxy-X-
rhodamine. The threshold value, determined by the operator, is the fluorescent intensity
at which the increase in fluorescence in the reactions is approximately log-linear. The
CT, or threshold cycle, is the cycle at which the normalized fluorescence in the reaction
tube exceeds the threshold value. The CT was measured and compared to the CT values
in the standard curve to determine the number of targets in a given reaction.

All optimization steps were carried out in a final volume of 25 ul and all sample
and standard reactions were 50 ul. Each PCR reaction contained 1X SYBR Green PCR
Master Mix (SYBR Green I Dye, AmpliTaq Gold® DNA Polymerase, dNTPs with dUTP,
passive reference dye 6-carboxy-X-rhodamine, magnesium, and PCR buffer) (PE
Applied Biosystems). The concentration of each primer was Optimized separately and
the following concentrations were used: 1108F 300 nM; 1132Rb 300 nM; 967F 50 nM;
1132Ra 50 nM; 348F 300 nM; 552R 50 nM; 569F 300 nM; 694R SOnM. Standard DNA
and sample DNA was diluted in water containing 1 ng/ul lambda phage DNA (New

England BioLabs) as a carrier to prevent DNA loss to surfaces. MicroAmp optical tubes

118

 

been a: 23% 88885058520 $8

E . N2 839.. 58080650500 3% cesarzmz 8% 5:203

an 82% 8005080005230 man

as ”.8 we? 2 65085605003 .33 $35
$3... $5 w: 22 _ 6058588563 3am _

5 5 $33 053006000855. ”see 3352

as 2: 7m: 2 00585000580 3&2 a

E 3 mg 732 650505580502... ”so: Beam
855%» Gnu fiwfi: action 5th SEE mo €33 coconmm 0:8: c255

zoom—95

Mom 083-32 com «can harm ~ 033.

119

and caps were used (PE Applied Biosystems). The PCR cycling conditions were as
follows: 10 min at 95°C followed by 40 cycles of 155 at 95°C and l min at 60°C. A set
of standard reactions with known numbers of target SSU rRNA genes was run
simultaneously with every set of sample reactions to generate a standard curve.

Samples were diluted to a final concentration of 1:100, 1:200, 1:400, and 1:800 in

reactions with bacterial, eukaryal, and 1Bbl-specific primers and 1:200, 1:400, 1:800,
and 121600 in reactions with archaeal primers. The only exception was the activated
sludge DNA which was diluted to a final concentration of 111000, 122000, and 124000 for
reactions with eukaryal primers. Each template dilution was run in triplicate. By
comparison to the standard curve, the number of targets per reaction was determined and
linear regression analysis (Excel, Microsofl) was used to extrapolate these 12 values to
determine the target number in the undiluted sample.
Controls The DNA of the following species was used as positive control template and in
the generation of standard curves: Escherichia coli (bacterial primers), Methanococcus
jannaschii (archaeal primers), and Saccharomyces cerevisiae (eukaryal primers). For the
primer set specific for the cloned 168 sequence lel, the standard template was plasmid
pCRG2.1-TOPO (Invitrogen) with the cloned le1 insert.

To test for specificity, bacterial, archaeal, and clone lel-specific primer sets
were compared to the species available in the RDP 11 database (version 8) (30). Primers
used to amplify eukaryotic sequences were compared to the sequences in the Arb
database (release December 1998).

In order to test whether the presence of high concentrations of non-target template

affected specific amplification, a series of control reactions were carried out. For each

120

primer pair, a set of reactions containing a range of concentrations of standard template
was spiked with 0.004 ng/ul of DNA extracted from non-target species. The following
non-target templates were used: M. jannaschii (bacterial primers) and E. coli (archaeal,
eukaryal, and lel-specific primers). Standard templates are listed above, and the
concentrations of standard templates in these reactions were as follows: 2.3 X 10'5 to 3.0
X 10'3 ng-ul'1 (bacterial primers), 3.1 X 10.5 to 4.0 X 10’3 ng-ul'l (archae primers), 1.6
x10" to 2.1 x 1023 ng-ur‘ (eukaryal primers), and 1.6 x 10'9 to 2.7 x 1025 ng-ur‘
(1Bb1-specific primers). To further test the sensitivity of the primers to non-target
template, these templates were tested, in the absence of target template, in a series of
control reactions. The concentrations of non-target templates were as follows: 4.0 X 10'4
to 4.0 ng-ul‘l (bacterial primers) and 3.0 X 10" to 3.0 ng°ul'l (archae, eukaryal, and
1Bb1-specific primers).
RESULTS

Description of the study site The study site is located in the town of Oscoda, Michigan.
The aquifer at the site is a sandy, unconfined formation contaminated with
tetrachloroethylene (PCE) released from a dry cleaning establishment. Flow within the
aquifer is eastward toward Lake Huron. It has been determined in previous studies that
microorganisms were responsible for a sustained degradation of PCE observed at the site.
In the center of the plume, hydrogen and redox data indicate that a gradient of conditions
exists which favors halorespiring organisms at shallow and intermediate depths (eight to
12 feet below the surface) and sulfate reducers and methanogens in deeper regions (16 to

19 feet below the surface) (2).

121

Several samples were used as controls for this study, including sediment from the
Wurtsmith Air Force base, sediment from Oyster, Virginia, activated sludge, and
methanogenic bioreactor fluid. A bacterial 16S clone which was 97% identical to lel
was identified at the Wurtsmith Air Force base site in a prior study (15). Samples from
this site were selected in order to test for the relative abundance of clone 1Bbl in this
sediment. Sediment from Oyster, Virginia was selected as representative of a pristine
aquifer community. Activated sludge was selected to provide a sample known to contain
high numbers of microeukaryotes. Bioreactor fluid was selected to provide a well-
characterized microbial community of archaea and bacteria.

Primer specificity All primer sets were found to be highly specific for their target
groups. Out of 15208 bacterial sequences in the database, the bacterial primer 1108F
matched 10,658 and primer 1132R-b matched 9,731. Neither bacterial primer matched
any archaeal sequences in the database. Out of 1,173 archaeal sequences in the database,
the archaeal primer 967F matched 703 and primer 1132R-a matched 197. Neither
archaeal primer matched any of the bacterial sequences in the database. The clone
specific reverse primer matched 23 bacterial sequences in the database, namely Ihermus,
Desulfocapsa, Desulfobacterium, and Desulfofiistis species and several unidentified
clones. The forward primer used in conjunction with the clone-specific primer is
universal and matched 10,895 of the bacterial and archaeal sequences in the database. As
the RDP database did not contain eukaryal sequences, the eukaryal primers were
compared to sequences in the Arb database (release December, 1998). The forward
primer 348F matched 2,144 out of 32,270 eukaryal sequences in the database. The

universal reverse primer matched 13,905 sequences in the database, 2,527 of which were

122

eukaryal. PE Biosystems (35) recommends that for real-time PCR applications the last
five bases at the 3’ end of the primers should have no more than two G and/or C bases.
Further, the protocols recommend avoiding runs of identical nucleotides. In designing
primers for the SSU rRNA gene it was necessary to violate these recommendations.

In control reactions to test for the effect of non-target DNA template on the
amplification of target template, a high concentration of non-target DNA (0.004 ng/ul)
was added to reactions with variable concentrations of target template. The archaeal and
eukaryal-specific primer pairs were found to be insensitive to the presence of 0.004 ng/ul
of E. coli DNA in all reactions. The 1Bbl-specific primer set was sensitive to the
presence 0.004 ng/ul of E. coli DNA when the number of 1Bb1 16S genes in the reaction
was less than 60. This effect was deemed unimportant in the application at hand, as 60
targets per reaction was outside our reliable detection limits. The bacterial primer set was
sensitive to the presence of 0.004 ng/ul M. jannaschii DNA only when bacterial template
was less than 0.0004 ng/ul (30,000 bacterial 16S genes). Again, this effect was
considered to be insignificant in the current investigation: working backwards from the
numbers of targets detected in our sample reactions, most reactions held significantly
more than 0.0004 ng/ul bacterial DNA, and at lower bacterial DNA concentrations
archaeal DNA was equally dilute. Furthermore, for sample reactions in which the
bacterial template concentration was below this threshold value, reactions with 2 X and
4X as much sample template showed a linear relationship (R2 > 0.9) in regression
analysis to the dilute sample. Hence, as sample DNA was diluted, archaeal DNA did not

compete with bacterial DNA as the amplification template.

123

In testing our primer pairs against variable concentrations of non-target template
DNA in the absence of target template, archaeal, eukaryal, and 1Bb1-specific primer sets
were found to be insensitive to the presence of non-target template at concentrations at
or below 0.3 ng/ul. The bacterial primer set amplified non-target M jannaschii DNA at
all concentrations tested (4.0 X 10“‘ to 4.0 ng-ul”). However, in light of the fact that the
primer set was insensitive to non-target template when bacterial DNA was present in
concentrations greater than 0.0004 ng/ul, this effect was deemed irrelevant.
Detection Limits In the absence of template, reactions of each primer pair (except the
archaeal primers) were found to result in non-specific amplification, most likely a primer-
dimer phenomenon. The CT of these reactions was substantially greater than that of the
lowest set of standards. In order to score a given group as “detected” in a sample, the CT
of all of the 12 data points for that sample must have exceeded this background level. In
addition, the response in the calculated concentration of targets in these reactions must
have displayed a linear response to template dilution in order for the tested group to be
determined as “detected”. A linear response was defined as an R2 value greater than 0.9
in a linear regression plot of the calculated concentrations of each of the 12 data points.
Quantification Limits In order to accurately quantify the number of 16S targets in a
given sample, at least half of the data points must lie within the quantifiable range of the
standard curve. The quantifiable range was between 2,000 and 250,000 targets per
reaction for the bacterial primer set, between 2,000 and 250,000 for the archaeal primer
set, between 15,000 and 1,000,000 for the eukaryal primer set, and between 250 and
250,000 for the lel-specific primer set. The number of 18S genes in the genome of S.

cerevisiae, which was used as a template for the eukaryote standard curve, varies

124

between 100 to 200 c0pies per genome (1). Hence, by comparison to the standard curves
we were able to estimate the number of 188 targets in a given sample to within a factor of
two.

For each set of unknown reactions, a set of known standard reactions was
assembled and used to define a standard curve. Figure 1 illustrates four of these standard
curves.

The fluorescence thresholds used to analyze the data were set to 0.02 for the

amplifications with bacterial primers, 0.015 for the archaeal primer set, 0.01 for the
eukaryal primer set, and 0.011 for the 1Bb1-specific primer set.
Quantification of bacterial, archaeal, eukaryal, and lel SSU genes The numbers of
bacterial, archaeal, eukaryal, and lel SSU genes detected in each sample and the
standard errors of each calculation are listed in Table 2. The percent of the total
calculated number of targets of SSU genes from archaeal, bacterial, and eukaryal sources
is also reported. We were able to identify clone 1Bb1 in five out of seven sediment
samples from the Bachman site and in both of the Wurtsmith sediment samples.
Furthermore, bacterial 168 genes were detected in all samples, archaeal 16S was detected
in all sediment samples and in both replicates of bioreactor fluid, and eukaryal 18S genes
were detected in saturated Wurtsmith sediment and in the activated sludge samples.

In aquifer sediment, bacterial 16S genes ranged from 2.02 X 106 to 1.72 X 107
168 genes per gram of sediment, archaeal measurements ranged from 1.20 X 105 to 4.94
X 107 genes per gram of sediment, and we measured 1.51 X 108 eukaryal 18S genes per

gram of sediment in Wurtsmith sediment ML3 (23.5 — 27.5 it). In Bachman samples, the

125

8&8 88m 33.2% 3% .8 8:52

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8+wé 8+m€ 3+m... mo+m€ No+mé 5+mé 8+w.w
. . . . . 9
u/Im/ t
Bantam 9
N
H
/ 8
/W .8
.R
M mm
8&8 88» <5: :8 co 88:2
8+m... 82m; 3.5.? 8.7m;
p h ON
/ a
m/ N W
commonam / mm o
/ em m..
0
/ mm M.
4/ em up
.. R m
on
8

(13) 91°40 P10113911”.

.3828». 8823b macaw ceramic 33:83
33:2: 23 comm £235 058%; am: one—o 5; £85on pawn—am 05 $296 w ESE 98 .maougoc

RES—:0 ofi fiofiow 0 $53 .9858: 335% House“ on... $2ro a 3:3 abate ommooamétfioan
firs genome 33:86.. weave a 38m .83 MUm cumin»: “2835 8 com: 823 Havana we fl onE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8&8 88» <72. 3mm .6 8.82
8+moo. w 8+woo€ 3&8. F
. t mm
/ 8
85° u N7“ mm
/H 1. mm
0 - vm
mm
on
3&8 85m «.ch Dmm do 8952

8+w€ mo+mé ¢o+m€ 8+mé

. . t

I 9

/ mr

/ om

. ../li a

88 o i am / N

K m 1. mm

A. . en

mm

(to) arofio 91011591111.

(13) 91°50 proqwru

126

 

 

- - - $3 .2 x RN ‘ 02 x of a 8% m< 882m .3
- - - $3 ..2 x a: 82 x 8: 28m m< 882m .8<
$28 82, x a: m.2 x88 . . $02 “2 x 3.8 82 x 83. a 8323 88885
888.3 82 x 8.x he x E 0%.: LNS x SN 82 x 8N 23:3 88885
.58 Sta 8555 <75 .33 SEQ 8955 <75 «Es—am Emma 0.9—Em
.3 .x. FEE—Sm _= \ 8035. .3 ..\e FEE—Sm _= \ flow—ah.
:5 m: x 8.8 m2 x o: x :28 x2.“ x 82 x 58 As awéevbao 858
. $3 a: x Nam 82 x 38 :3 x :.~ 82 x mg a EN. 03 32 £882.25
$38 n2 x :8 \[2 x «3 xx? x m3 h2x m: E 02. on: 32 58553
088.9” 82 x 3.8. 82 x m2 0838 x E.“ 82 x 8N E 2 .53m 558m
«03¢ _ 82 x 3m 82 x 89m _ :88 x 3% 82 x o: E 2 - 5 3m 5883
x xx: ”2 x 9% n2 x 3.8 $38 x Ga 82 x 8a a com? 2 -532 58.88
. $8.2 82 x a: 82 x w: _ $.88 x 82 82 x Ea . 13:: E - 5 mi: 8.583
$68 .82 x mam 82 x w: gnaw x «3 82 x o: E 2 432.2: .8583
$8.? 82 XS: 82 x 03 , 38.8 x :5 .82 x 8a a 8&3 2 - :sta 5885
$8.8m ”2 x 3m 82 x mi 888.2 x 38 82 x 83 _ .8318 2 - 52-2: 5:288
.83 .Shm 2.2568 :38 SEE 2.2.5.8 0353 Ewto 0.956
.8 ex. 235% w \ 8.8» m3 .8 .\.. 2.65% w \ 8.8” me—
.3292 3.58.5
.8883 8: mu? .3on 2: :2:
8:865 ..-.. m 98 335525 8: :5 29:8 8 E 388% 803 80538 SE :2: 83385 398% ..+.. 23. 835368 we

35 8m Bron—2 & 88038 macaw mg 3.883 .8 598:: 33 05 8 couzntuaoo “5ch 2t. .2953 cozm a 8m 3883
macaw Dmm :33 05 8 3:55:00 803:3 some 35 macaw Dmm .«o 83th 2.: mm L88 mo 8\9. 3088388 some ,«o 8:0
€3.88 2: 98 A8383 owvém 8883qu 98 888828 <75 388on 288% mo 3 hog 8 Ame—9:3 “890 can amen—cum
5:888:26 52:68 83:3 mo 8an 8a flows“ Dmm Em: vac—o can fibuxso 4828a .3588 no 838:: 2:. N 038.

127

 

 

:58 “S x 03 A“A: x 84 a com m< omega .3
£68 n2 x m: we x w: 5an m< omuam .8<
- . - = dam EA 88885
- . - H and 5A 3838mm
:38 Sta 8955 <2: 9385 ENE. 0.9—Em
..o .x. FEE—Sm _= \ 8935.
- - - - . - As 3 - we :50 9:0
+ + + 9&3 “2 x mom ”2 x a; 8 3m. 28 32 £§§B
+ + + - - - 8 wt- 3: 32 $553
$3 "2 x m3 m2 x 3.9 - . - 8 2 -526 5853
$3 NS x O? ,2 x a: - - 8 2 - 53m 5853
- . - - - = gm 8 2 - 5 mi: $853
+ + + - . - H com 8 2 - 5 TE 5555
«one "2 x 8.0 ,2 x of - . - a 2 -5245 5:205
- - .. - - = and a 2 - :v 2.5 5858
+ + + - . - H 9.x G 2 - 5 2-2: .5853
mop-0W mw— kahuna afloEm—VOm =39“ nob-H 235:5» o-QEam £me o-Qaam
.83 ..c .x. Env.—Sm w \ 85» m3 .3 .x. Eat—Sum w \ 85» mm—
EE 25.9 REE;

302388 N 25

128

16S gene of our target organism, lel, ranged from 9.64 X 103 to 1.66 X 104 genes per
gram of sediment.

S. cerevisiae DNA was used for building the standard curve with the eukaryotic
primer set, and the number of 188 genes in S. cerevisiae is thought to vary between 200
and 400 copies per genome (l). The numbers of eukaryal 188 genes listed in Table 2 are
conservative estimates based on the assumption that S. cerevisiae carried 200 188 genes
per genome. Less conservative estimates may be calculated by increasing these values by
a factor of two.

DISCUSSION

Real-time PCR has been shown to be a sensitive, precise method for quantifying
target nucleotide sequences in both clinical and environmental samples (9, 22, 25, 26, 28,
38, 47). Through the use of four kingdom- and species-specific primer sets, we have
used real-time PCR to detect and quantify the 16S rRNA genes in DNA extracted from
environmental microbial communities. Each of the primer sets was found to be highly
specific for the target group and we were able to detect and quantify as few as 2,000
bacterial, 2,000 archaeal, 8,000 eukaryal, and 250 SSU rRNA genes from the clone 1Bb]
per reaction.

In order to employ real-time PCR accurately, we validated three aspects of the
method: specificity, sensitivity, and reproducibility. The primer pairs used in this study
were found to be highly specific for their target groups. Furthermore, real—time PCR
exhibited sensitivity to even very low concentrations of target sequences: reactions were
sensitive to the presence of as few as 250 rRNA genes per tube. Finally, we evaluated

the reproducibility of the method from one DNA extraction replicate to another. We

129

extracted" and analyzed replicate sediment samples lW-N (1 1-13 it) I and II and 1W-S
(17-l9 ft) I and II, replicate bioreactor samples I and II, and replicate activated sludge
samples I and H. In general, a more reproducible detection of absolute numbers of
bacteria and archaea was observed in replicate sediment samples (replicate samples were
within 4-46% different) than in replicate activated sludge and bioreactor samples
(replicate samples within 11 and 86% different) which were derived from a single cell
pellet. However, in terms of the representation of bacterial and archaeal 16S calculated
for a given sample, the bioreactor replicates were more consistent (1-8% difference) than
the replicate sediment samples (7-26% difference). Replicate sediment samples did not
agree in the detection of clone sequence lel, suggesting that sediment heterogeneity,
even within replicate samples from the same core, can affect the detection of small
populations.

The primary focus of this study was the detection and quantitation of our target of
interest, cloned l6S sequence 1Bbl, in sediment from the contaminated aquifer at the
Bachman site. We were able to detect this sequence in five of the seven Bachman
sediments tested and in both Wurtsmith sediments (Table 2). Clone lel was not
detected in any of the samples from the aquifer in Oyster, Virginia, the bioreactor, or
activated sludge. Although this clone comprised 26% of a bacterial clone library derived
from contaminated sediment at the Bachman site (see Chapter 4), we determined that it
only comprised up to 0.6% of the total bacterial 168 in Bachman sediment. This
discrepancy may have been caused by a bias in the PCR reaction used to amplify 16$ for

cloning (for a review of these considerations, see reference (44)) or to bias in the cloning

130

process, including toxicity of vector inserts to the transformed host, or the choice of
cloning kit (27).

The discovery that the organism of interest, lel, constitutes less than 1% of the
bacterial community at the Bachman site does not preclude its importance in the natural
attenuation detected at the site. It is possible that the relative contribution of lel to the
activity of the microbial community is disproportional to the number of 16S genes
detected in the sediments. In one scenario, a high number of dead or inactive cells that
contribute their 16S genes in the extraction of DNA from a sample could reduce the
comparative importance of this species. It is possible that lel is directly involved in
reductive dechlorination at the site, as sulfate-reducing isolates and mixed cultures have
been found to reductively dechlorinate chlorinated ethenes (5, 20, 34). Further work with
the isolate most closely related to this strain, STP23 (41), may help to define the
processes that it undertakes at the Bachman site.

Bacterial 168 genes in Bachman sediment ranged from 2.02 X 106 to 3.70 X 106
per gram of sediment, archaeal 168 genes ranged from 9.37 X 105 to 3.66 X 106 per
gram, and eukaryal 188 genes were not detected. Bacterial 16S genes outnumbered
archaeal genes in these samples: the relative proportion of bacterial 16S genes to the total
number of sequences detected ranged from 50.3% to 70.4% (Figure 2). In Wurtsmith
sediment samples, bacterial 16S genes ranged from 2.22 X 106 to 1.72 X 107 per gram of
sediment, archaeal 16S ranged from 5.38 X 106 to 4.94 X 107 genes per gram, and
eukaryal 18S was determined to be 1.14 X 108 per gram in the deeper, saturated
sediment. Eukaryal 18S genes dominated the numbers of rDNA detected in the saturated

Wurtsmith sediment: they comprised 93.7% of the total SSU genes detected (Figure 2).

131

.8383. owes. H0308a 2: a Ha.W a E I 0va SSH 2028...

stew-:23 5 2:8 080880 33 m2 RES—5m .E Had-Wm VDQO 0388 5 H0 mow-8:08 owes—m 0033800 8: .3 5:30 5 080800
80: 33 mm: H8202 .095 0802 5‘ £28223 Hugo-H 05 3 H8800 080588000 a 28 .86 e895 2: 80.2 88508 83:3
00:85 .83 HaEHHBNm 8: 80¢ Haofivom H0503 wouecgacoo-mom .20: 530805 .0352 080388 80¢ <75 E c8088.
80% 56% .Ho 8098:: :38 8 800m 54% Dmm R5030 05 4008.80 €583.80 80:35:00 03228 8i. N oHHHmE

 

_me 35%-"HE m2 H322... l 8: Haeeoamfi

 

Xbo— .Xéo

 

H 38:82 Hm-H
HH 88:32 Hm-H
H angina m<

 

 

 

HH 88:32 3.

 

 

 

He 0.» - ”.8 :50
E n:- 2: SE

 

3H 3.“ - was 32

 

\. - x... \
..Nxfifihmkfl

H 28:82 E 2 - H: 2- 3H
HH eagaom E 2 - H: 2- 3H
8 2 - H-c 2- 3H

H 28:82 8 2 - :H WBH
HH 28:82 8 S - E 35H
8 2 - H: 3m

 

 

J|___l L__1Ll

€2-th 3m 1

20808on

owes—m c80>to<

«COEmUOv. heum>c

HHHoEEom fiHEthHB

80868 sac-£00m

 

132

In pristine, aerobic aquifer sediment from Oyster, Virginia, the number of
bacterial 16S far outstripped that of archaeal 16S. We measured 3.81 X 106 bacterial
168 genes and 1.2 X 105 archaeal 16S genes per gram of sediment, so bacterial 168
genes represented 98.3% of all those detected (Figure 2). Direct cell counts at the Oyster
site have measured approximately 107 cells per gram of sediment (21), and the relative
numbers of bacterial and archaeal rRNA genes measure in this study suggests that the
vast majority of theses cells are bacterial.

DNA extracted from bioreactor fluid held between 1.20 X 106 and 2.06 X 106
bacterial 16S genes and between 6.56 X 106 and 1.22 X 107 archaeal 16S genes per ul.
Hence, archaeal 16S comprised roughly 85% of the total SSU genes detected in
bioreactor fluid. Direct cell counts indicate that the bioreactor holds roughly 1010
cells/ml (16). The large discrepancy between cell counts and rRNA genes indicates that
the DNA extraction of the bioreactor cell pellet may have been inefficient, resulting in
the loss of genomic DNA during the process. DNA extracted from activated sludge held
between 1.5 x 105 and 1.36 x 106 copies ofbacterial 16S and 1.21 x 109 to 3.37 x 108
eukaryal 18S genes per ul, so eukaryotes represented over 99% of the total SSU genes
detected (Figure 2).

In our measurements of SSU rRNA genes, replicate bioreactor and activated
sludge samples varied by as much as 86% and replicate sediment samples varied by as
much as 46%. In light of this, it is advisable when using real-time PCR to detect and
quantify kingdom-level groups or small populations to thoroughly homogenize

environmental samples and to use replicate samples for extraction and analysis.

133

The relative numbers of prokaryotic rRNA genes detected in DNA from aquifer
sediment samples are consistent with the relative textures of these samples. Each of the
sediment samples was examined and their relative textures were visually compared to the
others in terms of relative texture (data not shown). The finest, sandiest textured
sediments, ODU4 (8.8 — 8.9 m) and ML3 (15.5 - 17.5 ft), yielded the highest numbers of
bacterial 16S, and the sediment samples with the coarsest texture, the Bachman
sediments, yielded the lowest numbers. Although the number of bacterial 16S targets in
the Oyster sediment sample ODU4 (8.8 — 8.9 m) was high, the organic carbon
concentration in Oyster sediment, 1,000 ppb (21), was lower than that measured in
sediment from either Bachman (as high as 2,400 ppb TCE alone (2)) or Wurtsmith (as
high as 13,650 ppm (15)). These facts suggest that for our samples, the texture of a given
sediment was a greater determinant of bacterial density than organic carbon
concentration.

The ratios of archaeal 16S genes to bacterial 16S genes measured in sediment
from the Bachman site and the Wurtsmith site and in biomass fiom the methanogenic
bioreactor were higher than ratios of rRNA derived fiom hybridization data. At the
Wurtsmith site, rRNA hybridization analysis has shown that archaeal rRNA comprises
between 1.2 and 1.6% of the total rRNA in unsaturated sediment and between 0.9 and
9.5% of the total in saturated sediment (48). Hybridization analysis of the community in
the methanogenic bioreactor has revealed that archaeal rRN A comprises up to 25% of the
total (16), and work on other methanogenic bioreactors has revealed similar ratios of
bacterial to archaeal rRNA (39). Of the combined numbers of archaeal and bacterial

genes detected, we observed percentages of archaeal 16S DNA to be 70.8 to 74.2% in

134

Wurtsmith aquifer sediment and 84.5 to 85.6% in the bioreactor. We have confirmed
these data by re-examining the concentration of DNA in reactions used to generate the
standard curve, verifying that the standard curve was calibrated correctly (data not
shown). We also ruled out the possibility that bacterial DNA affected amplification with
archaeal primers by assembling artificial community DNA samples, in which E. coli and
M. jannaschii DNA were combined in three different ratios. We were able to measure
the relative contribution of bacterial and archaeal 16S targets to within 8% of the correct
value (data not shown). Taking this line of investigation a step further, we developed
control reactions with archaeal primers, varying concentrations of archaeal template, and
4.0 X 10'3 ng/ul bacterial DNA. Amplification of archaeal 16S was not affected by the
presence of bacterial template. Furthermore, we verified that, in the absence of archaeal
template, the presence of bacterial template in reactions with archaeal primers did not
result in a false-positive amplification: we detected amplification greater than that seen in
no-template controls only when the amount of bacterial DNA in the reaction exceeded .
0.3 ng/ul. We estimated, from the number of bacterial and archaeal targets measured in
our samples, that there was a maximum of 0.1 ng/ul of total template DNA in our
reactions, which is well below this limit. Having validated the method of quantifying
archaeal 16S using real-time PCR, we consider our data on archaeal 16S targets in
aquifer sediment and bioreactor DNA to be accurate.

There are a number of possible explanations for the high archaeal:bacterial rDNA
ratios in Wurtsmith sediment and bioreactor samples measured in this study. Firstly,
differential cell retention times for bacteria and methanogenic archaea in the LS1

bioreactor may have affected these analyses. In the bioreactor, F420 fluorescent cells,

135

presumably methanogens, frequently formed aggregates. The cell retention time of
inactive methanogens (with low rRN A content and ambient DNA content) in aggregates
may have been affected by the differences in settling properties, leading to
overrepresentation of archaeal cells in samples taken from the reactor. This would
explain why high amounts of bacterial rRNA, which indicates metabolic activity, were
detected in hybridization analysis, while the scales tipped in favor of archaea in the
analysis of rDNA. The most likely explanation for the disparity between rRNA content
and rDNA content, however, lies in the possibility that archaea maintain a higher
rDNAerN A ratio than bacteria. Studies of Sulfolobus acidocaldarius and M. jannaschii
have shown that, because of a characteristically eukaryal cell cycle, these species carry
between 2 and 5 genome equivalents during stationary phase (6, 8, 31, 36). E. coli
carries, on average, one genome equivalent during stationary phase (17). If relatively
small populations of archaea in the Wurtsmith sediments and in the methanogenic
bioreactor maintained multiple genome equivalents while bacterial cells maintained a
single equivalent, comparisons of rDNA from these environments would be skewed in
favor of the archaea. More study of the archaeal cell cycle may confirm that the
maintenance of multiple chromosomes during stationary phase is ubiquitous among the
archaea, as the machinery of cell division observed in S. acidocaldarius and M.
jannaschii are thought to be shared by all members of the kingdom Archaea (7). Ifthis is
the case, then researchers using DNA-based comparisons, like real-time PCR, of archaeal
and bacterial populations will have to address the relationship between activity and

genome equivalents in members of these two kingdoms.

136

The detection of eukaryal 16S sequences in activated sludge DNA was expected,
as municipal wastewater treatment sludge is known to harbor high numbers of diverse
microeukaryotes (32). Eukaryal 16S genes were also detected in saturated sediment from
the Wurtsmith site. Again, this is not an unexpected result, as previous work using rRNA
hybridization has shown that eukaryal sequences comprise from 1.1 to 4.4% of the total
detected SSU rRNA in sediment from 7.5 m below the surface (48). In the same study,
rRNA from shallower, unsaturated sediment, was found to be 5% eukaryal. However, we
did not detect eukaryal sequences in the unsaturated sediment from the Wurtsmith site,
possibly due to temporal heterogeneity at the site.

The detection of eukaryotes in saturated sediment and their absence in unsaturated
sediment is contrary to the trend noted by Madsen et al. (29), who measured higher
numbers of culturable protists in unsaturated sediment than in saturated sediment from
the same contaminated aquifer. They conclude from prior work (18) that the higher
protozoan numbers in the unsaturated sediment indicate faster bacterial grth rates. If
higher protozoan counts do signify faster bacterial growth, then growth rates may be
higher in the saturated sediment at Wurtsmith than in the unsaturated sediment.

The numbers of eukaryotic 188 genes listed in Table 2 are conservative estimates
based on the assumption that the DNA standard template, S. cerevisiae, yields 200 18S
genes per genome (15 Mb). However, the number of 18S genes in the yeast genome may
be as high as 400 (1). If this is the case, the number of 188 genes detected in our
samples may be twice that reported in Table 2. In future work with real-time PCR, the
eukaryal-specific primer set could be used to more precisely quantify these targets in a

given sample through the development of a standard template for which the number of

137

SSU targets per ng of DNA is known. Since the number of 188 cperons per genome can
vary in eukaryotes, eukaryal 188 target sequence may have to be inserted into a cloning
vector in order to develop a reliable standard template.

Where detected, the number of eukaryal 18S genes far exceeds that of bacterial
and archaeal 16S genes (Table 2). In the activated sludge samples, eukaryal 18S
comprises 99.7 to 99.9% of all SSU genes detected. In activated sludge, protozoa
number about 5,000 per ml or 5% of the total dry weight (40). Further, seventy percent
of protozoa in activated sludge are ciliates (14), a group which has been found to yield as
many as 9,000 SSU genes per cell (45). Bacteria have been found to number as high as
1010 per ml in activated sludge. Assuming that eukaryotes maintain 9,000 copies of the
rRNA operon and bacteria maintain approximately 5 copies, and using the observation.
that 99.8 of the detected SSU genes are eukaryal, these data translate into approximately
four bacteria for every eukaryote. This is a distinct underestimate of measured ratios of
bacterial to eukaryal organisms in activated sludge (1010:5000), a discrepancy which may
be due to errors in the aforementioned assumptions or difl‘erential DNA extraction
efficiency from bacterial and eukaryal cells. In any event, eukarya are important
members of the microbial community in activated sludge but the differences in rRNA
copy number between bacteria and eukarya make population comparisons problematic.

In Wurtsmith sample ML3 (23.5 - 27.5 ft) eukaryal 188 genes comprise 95.2% of
all SSU genes detected. In the aforementioned study of saturated Wurtsmith sediment
communities, eukaryal rRNA was determined to represent from 1.1 to 4.4% of the total

detected SSU rRNA (48). The differences in rDNA and rRNA representation in the

138

larger community suggests that the relative contributions of eukaryotic 188 rDNA and
rRNA to total community pools are significantly different.

The use of a species-specific primer set in real-time PCR appears to hold a great
deal of promise for the detection and quantitation of microbial species of interest. Even
in conjunction with a universally-specific SSU primer, using the 1Bb1 primer set we
were able to quantify as few as 250 16S targets per reaction. Future work with species
and group-specific primer sets in real-time PCR could avoid some of the limitations of
rRNA hybridization analysis while achieving quantitation of a species or genus of
interest. For example, real-time PCR does not require handling of RNA, which can be
unstable and is easily degraded, or radioactively-labeled probes, which can be of concern
to individual or environmental health. However the potential for a disparity between the
number rRNA genes per cell and the activity of that organism may complicate
conclusions about the contribution of a species to the functioning of the whole
community drawn from real-time PCR data. Further work in comparing real-time PCR
with SSU rRNA and real-time PCR with a single-copy gene may help to correct for some
of this potential for error. Another area for exploration lies in combining reverse-
transcription PCR and real-time PCR, wherein SSU rRNA would be extracted from an
environmental sample and amplified by reverse transcriptase in a real-time PCR reaction.
This approach would avoid potential problems associated with tying rRNA gene
abundance to activity, while sensitively detecting the presence and abundance of groups
of interest in environmental samples. In any case, real-time PCR promises to be a
valuable method in the detection and quantification of SSU rRNA from environmental

communities.

139

10.

ll.

12.

13.

14.

REFERENCES
Munich Information Center for Protein Sequences. Yeast - Ribosomal RNAs.
[Online] Available http://www.mips.biochem.mpg.de/proj/yeast/rna/nna.html
April, 2001.
Adriaens, P., L. Abriola, M. Barcelona, B. Fathepure, K. Hayes, J. Tiedje, P.
Kurt, F. Loffler, and E. Petrovskis. 1998. Application of surfactant-enhanced
aquifer remediation (SEAR) and enhanced halorespiration technologies at the
Bachman Road residential wells site, Oscoda, M., A proposal submitted to the
Michigan Department of Environmental Quality.
Amann, R. I., L. Krumholz, and D. A. Stahl. 1990. Fluorescent-oligonucleotide
probing of whole cells for determinative, phylogenetic, and environmental studies
in microbiology. Journal of Bacteriology. l72(2):762-770.
Amann, R. I., W. Ludwig, and K. H. Schleifer. 1995. Phylogenetic
identification and in situ detection of individual microbial cells without
cultivation. Microbiol-Rev. 59(1):143-69.
Bagley, D. M., and J. M. Gossett. 1990. Tetrachloroethene transformation to
trichloroethene and cis-1,2- dichloroethene by sulfate-reducing enrichment
cultures. Applied and Environmental Microbiology. 56(8):2511-2516.
Bernander, R. 1998. Archaea and the cell cycle. Molecular Microbiology. Aug.
29(4):955-961.
Bemander, R. 2000. Chromosome replication, nucleoid segregation and cell
division in Archaea. Trends-in-Microbiology. [print] June, 2000;. 8(6):278-283.
Bernander, R., and A. Poplawski. 1997. Cell cycle characteristics of
thermophilic archaea. Journal of Bacteriology. l79(16):4963-4969.
Bowers, H. A., T. Tengs, H. B. Glasgow, Jr., J. M. Burkholder, P. A. Rublee,
and D. W. Oldach. 2000. Development of Real-Time PCR Assays for Rapid
Detection of Pfiesteria piscicida and Related Dinoflagellates. Appl. Environ.
Microbiol. 66(11):4641-4648.
Brigmon, R. L., N. C. Bell, D. L. Freedman, and C. J. Berry. 1998. Natural
attenuation of trichloroethylene in rhizosphere soils at the Savannah River Site.
Journal of Soil Contamination. July. 7(4):433-453.
Cho Jong, S., J. T. Wilson, D. C. Digiulio, J. A. Vardy, and W. Choi. 1997.
Implementation of natural attenuation at a JP-4 jet fiiel release after active
remediation. Biodegradation. 8(4):265-273.
Clement, T. P., C. D. Johnson, Y. Sun, G. M. Klecka, and C. Bartlett. 2000.
Natural attenuation of chlorinated ethene compounds: Model development and
field-scale application at the Dover site. Journal-of-Contaminant-Hydrology.
March 31, 2000;. 42(2-4):113-140.
Corless, C. E., M. Guiver, R. Borrow, V. Edwards Jones, E. B. Kaczmarski,
and A. J. Fox. 2000. Contamination and sensitivity issues with a real-time
universal 16S rRNA PCR. Journal-of-Clinical-Microbiology. [print] May, 2000;.
38(5): 1747-1752.
Curds, C. R. 1975. Protozoa, p. 203-268. In C. R. Curds and H. A. Hawkes (ed.),
Ecological Aspects of Used-Water Treatment, vol. 1 The Organisms and their
Ecology. Academic Press, London.

140

15.

16.
17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Dojka, M., A., P. Hugenholtz, S. Haack, K., and N. Pace, R. 1998. Microbial
diversity in a hydrocarbon- and chlorinated-solvent-contaminated aquifer
undergoing intrinsic bioremediation. Applied and Environmental Microbiology.
Oct. 64(10):3869-3877.

Dollhopf, S. L. Personal communication.

Drlica, K. 1987. The Nucleoid. In F. C. Neidhardt (ed.), Escherichia coli and
Salmonella typhimurium. American Society for Microbiology, Washington, DC.
Fenchel, T. 1987. Ecology of protozoa: the biology of free-living phagotrophic
protists. Science Tech Publishers, Madison, Wisconsin.

Fernandez, A. S., S. A. Hashsham, S. L. Dollhopf, L. Raskin, O. Glagoleva, F.
B. Dazzo, R. F. Hickey, C. S. Criddle, and J. M. Tiedje. 2000. Flexible
Community Structure Correlates with Stable Community Function in
Methanogenic Bioreactor Communities Perturbed by Glucose. Appl. Environ.
Microbiol. 66(9):4058-4067.

Gerritse, J., V. Renard, M. Pedro Gomes Teresa, A. Lawson Paul, D. Collins
Matthew, and C. Gottschal Jan. 1996. Desulfitobacterium sp. strain PCE], an
anaerobic bacterium that can grow by reductive dechlorination of
tetrachloroethene or ortho-chlorinated phenols. Archives of Microbiology.
165(2):132-140.

Golder-Associates. 1998. Field Sampling Plan, Aerobic flow-cell installation,
narrow channel focus area, South Oyster Site, Oyster, Virginia,

Gruntzig, V., S. C. Nold, J. Zhou, and J. M. Tiedje. 2001. Pseudomonas
stutzeri Nitrite Reductase Gene Abundance in Environmental Samples Measured
by Real-Time PCR. Appl. Environ. Microbiol. 67(2):760-768.

Guiver, M., R. Borrow, J. Marsh, S. J. Gray, E. B. Kaczmarski, D. Howells,
P. Boseley, and A. J. Fox. 2000. Evaluation of the Applied Biosystems
automated Taqman polymerase chain reaction system for the detection of
meningococcal DNA. FEMS-Immunology-and-Medical-Microbiology. [print]
June, 2000;. 28(2):173-179.

Hendriks, L., A. Goris, J. Neefs, Y. Van de Peer, G. Hennebert, and R. De
Wachter. 1989. The Nucleotide Sequence of the Small Ribosomal Subunit RNA
of the Yeast Candida albicans and the Evolutionary Position of the Fungi Among
the Eukaryotes. System. Appl. Microbiol. 12:223-229.

Higuchi, R., C. Fockler, G. Dollinger, and R. Watson. 1993. Kinetic PCR
Analysis: real-time monitoring of DNA amplification reactions. Bio/Technology.
l 1(September): 1026-1030.

Kessler, H. H., G. Muhlbauer, B. Rinner, E. Stelzl, A. Berger, H. W. Dorr, B.
Santner, E. Marth, and H. Rabenau. 2000. Detection of herpes simplex virus
DNA by real-time PCR. Journal-of-Clinical-Microbiology. [print] July, 2000;.
38(7):2638-2642.

Lueders, T., and M. Friedrich. 2000. Archaea] population dynamics during
sequential reduction processes in rice field soil. Applied-and-Environmental-
Microbiology. [print] July, 2000;. 66(7):2732—2742.

Lyons, S. R., A. L. Griffen, and E. J. Leys. 2000. Quantitative real-time PCR
for Porphyromonas gingivalis and total bacteria. Journal-of-Clinical-
Microbiology. [print] June, 2000;. 38(6):2362-2365.

141

29.

30.

31'.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

Madsen, E. L., J. L. Sinclair, and W. C. Ghiorse. 1991. In situ biodegradation:
Microbiological patterns in a contaminated aquifer. Science. 252(5007):830-833.
Maidak, B. L., J. R. Cole, T. G. Lilburn, C. T. J. Parker , P. R. Saxman, J. M.
Stredwick, G. M. Garrity, B. Li, G. J. Olsen, S. Pramanik, T. M. Schmidt,
and J. M. Tiedje. 2000. The RDP (Ribosomal Database Project) continues.
Nucleic Acids Res. 28:173-174.

Malandrin, L. 1999. Nucleoid structure and partition in Methanococcus
jannaschii: an archaeon with multiple c0pies of the chromosome. Genetics.
152:1315-1323.

Marsh, T., L., W. T. Liu, L. Forney, J., and H. Cheng. 1998. Beginning a
molecular analysis of the eukaryal community in activated sludge. Water Science
and Technology. Feb. March. 37(4-5):455-460.

Pace, N. R. S., D.A.; Lane, D.J.; Olsen, G.J. 1986. The analysis of natural
microbial populations by ribosomal RNA sequences. Adv. Microb. Ecol. 9:Jan-
55.

Pavlostathis, S. G., and P. Zhuang. 1993. Reductive dechlorination of
chloroalkenes in microcosms developed with a field contaminated soil.
Chemosphere. 27(4):585-59S.

PEBiosystems. 1999. SYBR (registered trademark) Green PCR Master Mix and
RT-PCR Protocol.

Poplawskki, A., and R. Bemander. 1997. Nucleoid structure and distribution in
thermophilic archaea. Journal of Bacteriology. Dec. l79(24):7625-7630.

Power, M., R. Van Der Meer Jan, R. Tchelet, T. Egli, and R. Eggen. 1998.
Molecular-based methods can contribute to assessments of toxicological risks and
bioremediation strategies. Journal of Microbiological Methods. April. 32(2):107-
119.

Pusterla, N., J. B. Huder, C. M. Leutenegger, U. Braun, J. E. Madigan, and
H. Lutz. 1999. Quantitative real-time PCR for detection of members of the
Ehrlichia phagocytophila genogroup in host animals and Ixodes ricinus ticks.
Joumal-of-Clinical-Microbiology. May, 1999;. 37(5):]329-1331. 7

Raskin, L., L. K. Poulsen, D. R. Noguera, B. E. Rittman, and D. A. Stahl.
1994. Quantification of methanogenic groups in anaerobic biological reactors by
oligonucleotide probe hybridization. Applied and Environmental Microbiology.
April(1241-1248).

Ratsak, C. H., B. W. Kooi, and H. W. Van Verseveld. 1994. Biomass reduction
and mineralization increase due to the ciliate Tetrahymena pyriforrrris grazing on
the bacterium Pseudomonas fluorescens. Water Science and Technology.
29(7): 1 19-128. '

Sass, H., E. Wieringa, H. Cypionka, H.-D. Babenzien, and J. Overmann.
1998. High genetic and physiological diversity of sulfate-reducing bacteria
isolated from an oligotrophic lake sediment. Archives of Microbiology. Oct.
l70(4):243-251.

Strunk, O., and W. Ludwig. 1997. ARB: Soflware for phylogenetic analysis.
Technical University of Munich, Munich, Germany.

142

43.

44.

45.

46.

47.

48.

49.

Takai, K., and K. Horikoshi. 2000. Rapid Detection and Quantification of
Members of the Archaeal Community by Quantitative PCR Using Fluorogenic
Probes. Appl. Environ. Microbiol. 66(11):5066-5072.

von Wintzingerode, F., U. B. Goebel, and E. Stackebrandt. 1997.
Determination of microbial diversity in environmental samples: pitfalls of PCR-
based rRN A analysis. FEMS Microbiology Reviews. 21:213-229.

Ward, J. G., P. Blomberg, N. Hoffman, and C. Yao Meng. 1997. The
intranuclear organization of normal, hemizygous and excision-deficient rRNA
genes during developmental amplification in Tetrahymena thermophila.
Chromosoma Berlin. 106(4):233-242.

Weiner, J. M., and D. R. Lovley. 1998. Rapid benzene degradation in
methanogenic sediments from a petroleum-contaminated aquifer. Applied and
Environmental Microbiology. May. 64(5): 193 7-1939.

Weller, S. A., J. G. Elphinstone, N. C. Smith, N. Boonham, and D. E. Stead.
2000. Detection of Ralstonia solanacearum strains with a quantitative, multiplex,
real-time, fluorogenic PCR (TaqMan) assay. Applied-and-Environmental-
Microbiology. [print] July, 2000;. 66(7):2853-2858.

West, T. G. 1999. Master's Degree. Central Michigan University, Mount
Pleasant, Michigan.

Wilmotte, A., G. Van der Auwera, and R. De Wachter. 1993. Structure of the
16S ribosomal RNA of the thermophilic cynobacterium Chlorogloeopsis HTF
('Mastigocladus laminosus HTF') strain PCC7518, and phylogenetic analysis.
FEBS Letters. 317:96-100.

143

CHAPTER 6

CONCLUSIONS

The terrestrial subsurface is a relatively new frontier for microbial ecology. Prior
to the advent of aseptic sampling of the subsurface in the 1980s little work was, in fact,
possible. It is becoming clear, however, considering the sheer size of the terrestrial
subsurface, subsurface microbes likely play a substantial role in global cycling of
bioactive elements. Furthermore, the ecological connectivity of groundwater with
surface systems means that the subsurface is not an isolated area, but instead has
significant impacts on macro-ecological systems. And the human demands on
subsurface water are substantial: we pump 77 billion gallons of groundwater per day in
the US alone. So, the subsurface and groundwater in particular are key underpinnings of
global nutrient cycles, ecosystem health, and human water requirements, and microbes
are largely responsible for the chemical properties of these resources. Understanding the
communities and the roles of individual populations within them will help us to manage
groundwater resources and to better understand the role of the terrestrial subsurface in
global nutrient cycling.

In an effort to firrther our understanding of subsurface microbial communities, the
microbial diversity in aquifer sediment from two different sites was described using
molecular, 16S rRNA-based approaches. The first site was a pristine, shallow, aerobic
formation in Oyster, Virginia. The narrow channel focus area, or NCFA, is currently the
focus of a Department of Energy — Natural and Accelerated Bioremediation study in

bacterial transport, and has been extensively characterized for this effort. The aquifer is

144

an unconfined formation consisting of unconsolidated to weakly ceinented, medium- to
fine-grained sand and pebbly sand. These sediments are, in decreasing order of
abundance: quartz, feldspar, clay, and iron and aluminum hydroxides, with the smaller
sized quarts particles encrusted with clay and hydroxide coatings (3). Dissolved oxygen
was measured at 5.5 mg/L and determined to be consistent across the study site, organic
carbon was determined to be 1000 ppb, and the groundwater pH was 6.1(1).

The second study site was a tetrachloroethylene (PCE)-contaminated aquifer in
Oscoda, Michigan referred to as the Bachman Road site. The Bachman aquifer is a
similar formation to that in Oyster, Virginia: it is a shallow, sandy formation, composed
of unconsolidated, medium-grained glacial sand. Within the plume, dissolved oxygen
concentrations range from 1.8 mg/L 8 ft below ground level to 0.2 mg/L at 19 feet, and
redox potentials range from —151 to —247 respectively. The concentration of organic
carbon has not been measured in these sediments, but PCE alone has been found to reach
as high as 10,000 ppb and within our sampling area the concentration of cis-
dichloroethylene (cis-DCE) was determined to reach as high as 3,600 ppb (2).
Measurements of groundwater pH within the plume range between 5.8 to 7.6 (2).

, The bacterial communities at both sites were assessed using terminal restriction
fragment length polymorphism analysis (T-RFLP). While analysis did reveal diverse
communities at both sites, with Oyster communities presenting as many as 39 terminal
restriction fragments in the Hha I digestion and Bachman contaminated sediment
presenting as many as 26 and uncontaminated sediment presenting 5 terminal fragments,
these communities appear to be less diverse than soil bacterial communities. Soil

bacterial communities present between 60 and 80 terminal fragments when analyzed

145

using T-RFLP with the same primers and restriction enzyme. This suggests that the
properties that the Oyster and Bachman sediments share that surface soils do not, most
significantly water saturation and low concentrations of recalcitrant organic carbon, are
strong determinants of bacterial diversity. Further work to relate indices of bacterial
diversity to environmental heterogeneity and to concentration and complexity of organic
carbon would prove valuable in better understanding these observations.

Furthermore, the phylogenetic identities of members of the bacterial communities
at both sites were determined using 16S cloning and sequencing. Both sites exhibited
previously unknown bacterial diversity: ten clones from Oyster and seven clones from
Bachman failed to show a specific relationship to any of the known bacterial divisions.
The subsurface is a relatively new environment for study, so it is not surprising that
molecular microbiological techniques should reveal novel groups. While there was no
overlap between the 16S sequences detected at Bachman and Oyster, three divisions
were detected in clone libraries fi'om both sites: Acidobacteria,
Leptospirillum/Nitrospira, and the Proteobacteria. It could be that the characteristics the
two sites share, including relatively low organic carbon concentrations, absence of light
energy, low nutrient concentrations, water saturation, and low seasonal temperature
fluctuations are driving these similarities.

The only isolated representative of the division Acidobacteria, Acidobacterium
capsulatum, is an acidophilic chemoorganotroph. Acidobacteria-related sequences have
been identified in many environments, including soil, subsurface sediment, and marine
sediment. The diversity of these sequences suggests that the division Acidobacteria is

likely quite a metabolically diverse group, but the lack of isolates from more

146

representative environments is limiting. Representatives of the division
Leptospirillum/Nitrospira have not previously been identified in the subsurface. Isolates
from three genera compose this division: Leptospirillum, Nitrospira, and
Ihermodesulfovibrio. Leptospirillum isolates are Chemolithotrophic sulfirr or iron
oxidizers from acid mine drainage. Nitrospira are chemolithoautotrophic, oxidizing
nitrite for energy, and Ihermodesulfovibrio is a thermophilic sulfate-reducer. It is not
known what roles the members of the divisions Acidobacteria and
Leptospirillum/Nitrospira play in the subsurface, but fiirther work to isolate members of
these groups from subsurface sediment may help to better understand their functions.
Isolates and 16S sequences of the division Proteobacteria, on the other hand, have been
identified in most studies of subsurface bacterial communities and appear to be
ubiquitous in these environments. There are innumerable isolates of this metabolically
diverse division, but the sequences fiom Bachman and Oyster suggest that our
knowledge about the diversity of the Proteobacteria is not complete, as most of these
sequences failed to show relationships to any of the known genera. Again, more work to
isolate novel Proteobacteria from subsurface sediment could help in our understanding
of the activity of these species at Bachman and Oyster.

Sampling at the Oyster and Bachman sites may be seen as inadequate in light of
the potential for sediment variation across very small distances to affect microbial
abundance (4) and presumably community structure. Ideally, numerous cores from a
particular sampling point (in a meter square area, for example) would be gathered and
analyzed equally, but this is simply not feasible in the case of most subsurface

investigations. Regardless, I find it unlikely that selective sampling of any environment

147

as complex as an aquifer will reveal the entirety of microbial diversity of that site. With
the tools currently available to researchers, we have no choice but to extract and analyze
a limited number of samples from and extrapolate those results appropriately.

The discovery of previously unidentified 16S sequences and sequences belonging
to divisions about which we still know very little is an indication that a great deal of
research remains to be carried out to characterize the microbial communities of the
saturated subsurface. Further work should be focused on isolating and characterizing
these organisms and defining their distribution in other subsurface environments using

molecular techniques.

148

REFERENCES

U.S. Department of Energy. 1998. South Oyster Science Plan; Field
Experimentation in Bacterial Transport; NABIR Acceleration Program Element,
Office of Energy Research, Office of Biological and Environmental Research,
Environmental Sciences Division.

Adriaens, P., L. Abriola, M. Barcelona, B. Fathepure, K. Hayes, J. Tiedje, P.
Kurt, F. Loffler, and E. Petrovskis. 1998. Application of surfactant-enhanced
aquifer remediation (SEAR) and enhanced halorespiration technologies at the
Bachman Road residential wells site, Oscoda, MI., A proposal submitted to the
Michigan Department of Environmental Quality.

Dong, H., T. C. Onstott, M. F. DeFlaun, M. E. Fuller, S. H. Streger, R. K.
Rithmel, and B. J. Mailloux. Transport of adhesion deficient bacteria in
heterogeneous porous media: relative dominance of physical vs. chemical control
on bacterial transport. Environmental Science and Technology (submitted for
publication). .

Zhang, C., V. Palumbo Anthony, J. Phelps Tommy, J. Beauchamp John, J.
Brockman Fred, J. Murray Chris, S. Parsons Brian, and J. P. Swift Donald.
1998. Grain size and depth constraints on microbial variability in coastal plain
subsurface sediments. Geomicrobiology Journal. July Sept. 15(3): 171-185.

149

APPENDIX A
CULTIVATION TECHNIQUES USED TO ISOLATE
N CFA GROUP I BACTERIA
INTRODUCTION

A unique group of 16S sequences was identified at the Narrow Channel focus
area at a study site in Oyster, Virginia, and was designated NCF A group I (see Chapters 2
and 3). Cultivation experiments were designed and implemented in an attempt to obtain
these organisms in culture, but were unsuccessful. These experiments are described here.

Briefly, aliquots of sediment from two depths within the same borehole were
serially diluted in minimal media with added sulfate, thiosulfate, nitrate, peptone, and
yeast extract. Replicate dilutions were held at 10° C under three different gaseous
conditions: anaerobic, microaerophilic with hydrogen, and rrricroaerophilic without
hydrogen. Cultures were periodically checked for the development of NCFA group I by
whole cell PCR of an aliquot of culture using primers specific for the target organisms.

Media was created by adding ingredients to water in order listed in Table 1 then
making up to the final volume. Table 2 lists the ingredients of the trace metals solution
used in the media. The mixture was boiled on a hot plate and cooled in an ice tub under a
flow of nitrogen gas. Under nitrogen, 18 ml of media was transferred to individual
Balsch tubes and sealed with rubber stoppers and aluminum crimp tops. The tubes were
cooled and 90 ul of 10% dithiothreitol (DTT) solution was added to the tubes to be used
for anaerobic cultures and rrricroaerophilic cultures under hydrogen. The tubes were

autoclaved for 45 minutes.

150

Table 1 Media used for cultivation of NCF A group I. (Quantities are in units of

g per L). Store metals solution at 4 degrees C.

NaCl
MgC12-6H20
KH2PO4
K2HPO4
NH4Cl

KC]
CaC12-2HZO
Yeast extract
Peptone

Agar

Na2SO4

KNO3
Na2S203-5H20
Resazurine
Trace metals solution

0.05
0.5
0.2
0.25
0.3
0.3
0.015
0.1
0.1

0.142
0.05
0.25
0.001
1 ml

Table 2 Trace metals solution for grth media. (Quantities are in units of

g per L)

Nitrotriacetic acid

PH 6.5 then add in order, dissolving each one:

F eCl3-3H20
MnC12-4H20
CoC12-6H20
ZnCl2
CuC12-2H20
H3BO4

NaCl
Na2SeO3-5H20
NiC12-6HZO
Na2Wo4-2H20

12.8

1.35
0.1
0.024
0.1
0.025
0.01

1
0.026
0.12
0.033

Two sediment samples were used for inoculation, both of which were recovered
from the site in August, 2000: ODU2 7.7-7.8 m and ODU2 6-7 m. Media was inoculated
in the anaerobic hood (90% nitrogen and 10% hydrogen atmosphere), keeping the media
and sediment on ice as temperatures exceeded 37° C inside. Approximately 2 grams of
sediment was added to the first tube in the series (1:10 dilution wzv), and the tube was
capped and mixed by shaking. The mixture was allowed to settle for 10 seconds, then 2
ml was transferred to the next tube in the series. The dilution series was carried out eight
times for a total of nine tubes. Each sediment was used as inoculum for three replicate
dilution series. The replicate series destined for anaerobic conditions was recapped in the
hood and handled anaerobically for all subsequent sampling. The replicate series
destined for microaerophilic conditions with hydrogen were loosely capped with culture
tube caps (not rubber stoppers) and transferred to gas pack jars with Campy Pak Plus
nricroaerophilic envelopes (Becton Dickinson Microbiology Systems) which provided a
nricroaerophilic environment with hydrogen and carbon dioxide. The microaerophilic
replicate series was loosely capped with culture tube caps. All replicate series were
placed at 10° C in the dark.

Cultures were sampled in the anaerobic hood every other day as they grew turbid,
up to 18 days afier inoculation. Two depths of each culture were sampled (1 cm fi'om the
meniscus and 1 cm from the bottom) using sterile Pasteur pipettes and 1.25 ul was used
for each of two PCR screening reactions using (1) the NCFA group l-specific primer set

(see Chapter 3) and (2) bacterial primers 27F and 1392R. The amplification conditions,

152

including cycle parameters and reagent concentrations are identical to those outlined in

Chapter 2 for bacterial 16S cloning.

153

APPENDIX B

T-RFLP DATA FROM CHAPTERS 2 AND 4

Table 1 Chapter 3 data.

SOUTH OYSTER FOCUS AREA SEDIMENT

SEDIMENT B2 (5 m)
Hha I

Peak area
32.73 536
33.9 1260
35.64 361
38.99 1858
39.98 1595
41.57 3140
48.88 2037
60.43 876
66.07 7207
181.14 935
182.11 302
182.92 467
183.57 1297
187.37 580
187.97 690
189.34 822
206.2 31812
370.75 1797
567.63 3358
568.83 3923
574.52 1007
SEDIMENT M3 (6 m)
Hha I
Terminal
Fragment Peak area
33.92 2493

Msp I

Terminal
Fragment

32.75
33.91
38.98
39.96
41.54
48.81
125.9
141.98
143.8
172.97
177.18
197.19
234.71
431.87
433.76
440.56
477.37
492.09
494.02

Msp I

Terminal
Fragment

34.11

154

Peak area

333
951
1369
1496
2098
151
7732
1556
3163
3879
1478
346
636

10567
37121

1733
8082
3043

31662

Peak area

1881

Rsa I
Terminal Peak
Fragment area
32.97 377
34.12 945
39.15 1406
40.12 1367
41.68 1917
79.99 281
82.08 3849
109.63 404
110.71 399
116.81 5058
119.55 30299
475.34 38551
482.51 2094
489.17 492
492.06 1207
492.7 1146
Rsa I
Terminal Peak
Fragment area
33.92 1338

Table 1 (cont’d)

39.16 1741
39.95 1825
41.52 2978
48.94 324
65.61 618
206.38 7413
569.07 9922
SEDIMENT S10 (5 m)
Fragment Peak area
34.11 3780
35.65 787
37.98 1926
39.16 2454
39.95 2961
41.52 5007
48.94 591
50.7 1513
65.99 2064
67.72 1637
196.93 27974
225.15 3378
227.3 6353
234.61 9779
372.07 2000
569.67 1231

39.95
41.52
48.94
140.02
433.14
435.02
493.99
495.58

Msp I
Terminal
Fragment

32.97
34.12
37.2
39.15
40.12
41.49
48.88
125.92
133.43
152.36
154.35
174.43
186.3
293.87
492.56
495.75
500.38
517.27

155

2610
2050
215
375
600
4649
5126

14597

Peak area

443
1956
352
1505
1732
2522
187
158
1459
609

15683

1512

584
4411
2521
1991
4023
1147

38.97 1022
39.95 1290
41.52 1540
117.61 381
119.7 5274
430.7 1043
431.48 1479
472.02 523
476.4 2632
477.75 2219
Rsa I
Terminal Peak
Fragment area
32.97 343
33.93 1463
39.15 1621
39.93 1283
41.49 2087
117.78 439
464.27 4241
467.61 319
476.18 26912

Table 1 (cont’d)

SEDIMENT $14 (5 m)
Hha I

32.94
33.91
38.98
39.96
41.54
48.81
65.74
77.28
202.37
203.46
206.58
210.19
368.76
370.59
566.59
569.01

SEDIMENT $18 (6 m)

32.56
33.91

37
37.99
39.17
39.96
41.54
49.01
89.84

Terminal Fragment Peak area

393
1377
1667
1512
2494

984
3966

980

781

983

29426

678

834
2187
3991

14108

Terminal Fragment Peak area

507
751
384
541
1202
1251
3056
332
1753

Msp I

Terminal
Fragment

32.96
34.11
39.16
40.14
41.52
126.07
138.35
140.16
142.13
432.34
434.23
459.1
460.86
477.64
492.08
494
494.95
600.2

Msp 1

Terminal
Fragment

33.92
39.16
39.95
41.52
84.73
86.25
88.9
126
140.35

156

Peak area

422
1331
1870
1756
1992
3408
846
2200
531
2391
7721
758
2392
1006
19866
29789
36754

2536

Peak area

205
583
756
1 107
960
1968
1690
1616
678

RsaI

32.96
34.11
39.16
39.95
41.52
100.83
117.61
119.54
430.55
471.85

Terminal Peak
Fragment area

287
1236
1704
1493
1695

571
3153
9311
4964
4996

475.89 46424

RsaI

Terminal Peak
F ragrnent area

31.99
33.92
38.97
39.95
41.52
67.92
89.28
116.8

518
435
918
1086
1310
745
717
3815

119.5410041

 

Table 1 (cont’d)

157

121.06 446 142.16 780 126.81 900
122.54 721 156.39 327 127.79 538
123.52 1364 158.56 489 292.17 951
150.74 2227 443.03 350 307.31 2629
164.43 371 446.19 1202 476.63 6897
165.44 964 449.21 1855 482.22 3710
206.06 19491 450.64 1363
206.85 25477 451.59 778
370.85 1059 452.23 811
374.2 520 453.02 672
543.3 1112 454.45 1560
566.14 4448 455.73 1489
569.07 20229 456.85 889
574.43 3728 457.48 867
458.28 3179
484.38 1069
493.36 9910
494.47 16520
495.59 48757
515.74 625
517.31 1226
600.12 4484
SEDIMENTT2(4 m)
HhaI MspI RsaI
Terminal Terminal Peak
Terminal Fragment Peak area Fragment Peak area Fragment area
32.75 396 33.91 838 31.82 359
33.91 655 38.98 1235 32.78 372
37.2 350 39.96 1170 33.93 407
38.96 1559 41.54 1580 39.13 1322
39.94 1817 67.67 1784 39.9 1500
41.5 2969 88.7 435 41.45 2010
48.72 5370 98.15 1991 81.8 1943
65.6 "2831 100.07 1096 89.24 1211
120.26 3483 100.83 1090 101.04 1061
175.56 817 103.62 1340 109.72 2288

Table 1 (cont’d)

158

206.32 33716 104.72 699 110.71 2416
369.66 12317 105.26 786 111.68 2834
374.08 2944 106.7 646 11612216
375.77 878 107.77 1319 118.8232803
397 1108 110.87 1531 170.56 3452
567.43 4881 112.1 750 171.64 3155
569.15 18831 113.16 872 427.14 4511
577.49 522 114.21 1078 475.05 31411
120.7 566
124.92 212
126.06 6920
127.37 3867
172.81 1262
432.22 23210
433.79 28376
443.91 1464
477.37 5117
490.48 677
492.4 5008
493.53 5929
495.45 28178
600.25 1764
SEDIMENTT2(6 m)
I-IhaI Mspl RsaI
Terminal Terminal Peak
Terminal Fragment Peak area Fragment Peak area Fragment area
33.91 2763 33.92 2570 33.92 1779
39.17 3036 39.16 2051 38.97 1644
40.16 2360 39.95 2078 39.95 1671
41.54 5005 41.52 3165 41.32 2137
48.81 243 67.72 401 117.46 3176
65.74 4150 138.42 360 119.57 6737
67.86 388 140.23 1638 430.55 5499
77.28 849 142.37 663 472.02 1971
202.56 346 434.23 2246 473.37 2798
203.65 1558 459.1 544 476.57 4117

Table 1 (cont’d)

206.32 8304 460.86 1725
368.75 542 492.4 10229
369.82 984 493.36 9765
566.56 2603 495.59 23181
568.98 14926 600.36 606

NARROW CHANNEL FOCUS AREA SEDIMENT

159

SEDIMENT B2 (8 m)
Hha I
Primer set 8F-1392R Primer set 8F-1525R
Terminal
Terminal fragment Peak area fragment Peak area
39.27 1311 30.33 9983
40.22 305 40.22 475
41.98 1127 41.82 1417
64.16 928 64.16 1747
69.59 441 92.72 2233
87.14 355 95.21 1032
92.72 3510 102.94 484
94.51 482 105.72 5621
103.2 3225 205.39 6079
105.53 994 206.91 3347
107.36 596 210.1 848
108.7 773 211.91 687
120.38 961 219.18 2084
121.55 521 220.59 1301
128.56 358 313.51 1378
136.16 1261 365.77 7154
138.51 485 379.34 2673
141.14 1024 382.62 756
205.36 1292 384.76 2317
207.03 763 487.75 3966
219.18 750 573.72 3075
365.78 819 581.14 819
373.14 1162

Table 1 (cont’d)
375.27

378.83

474.5

476.76

487.57

574.17

Msp I

Primer set 8F-1392R

1088
8222
3220
5737
4538
1231

Terminal fragment Peak area

39.27

67.99

70.38

86.98
103.41
105.41
106.33
107.38
108.71
120.35
121.51
129.93
136.05
138.38
139.84
141.01
161.09
163.45
175.02
180.31
206.69
225.63
282.78
289.66

478.2
488.54

1191
1047
1606
563
2800
372
473
597
761
871
464
2178
1207
378
2425
1040
4021
2927
426
478
427
734
576
456
927
975

31.16

69.68

86.89
129.99

139.9

143.4
144.13
148.22
161.05
162.66
163.97
171.85
180.17
186.13
205.52
225.46
282.66
289.65

292.6
297.22
474.29
479.08
488.84
495.48
499.08
505.04

160

Primer set 8F-1525R

Terminal

fi'agment Peak area

11259

504
3625
1389
1792

596
1147
1522

699

679

486

475
1647
1023

480
5176
5351

956
2541
2941
1423

674

392
7409

754
1253

Table 1 (cont’d)

RsaI

Primer set 8F-1392R

495.48
513.82
519.56
574.16

2639
3872
4851
1785

Terminal fragment Peak area

34.16
39.27
40.7
56.66
63.2
68.79
69.43
86.02
103.41
105.41
106.33
107.38
109
120.19
136
141.09
243.28
455.59
458.08
460.1
478.74
484.29
488.32
489.5
497.08
502.9

2203
878
1047
429
501
633
655
458
3263
276
550
680
1414
731
1344
840
2877
1502
3961
8185
4668
1492
307
1407

13878

6893

515.11
521.42
534.09

40.29

85.94
243.38
418.06
450.82

455.2
468.23
477.85
484.11
485.67
489.65
490.47
498.18

161

2292
650
395

Primer set 8F—1525R

Terminal

fragment Peak area

1097
2887
1843

946
1714
9893
2060

10305

3621
2392
5685

16611

5531

Table l (cont’d)
SEDIMENT 818 (6 m)
Hha I

Primer set 8F-1392R

37.91
38.73
40.52
61.39
68.73
86.66
92.53
103.33
211.9
370.15
375.68
379.1
389.12
402.45
487.73
550.11
570.14
573.74

Terminal fragment Peak area

345
494
1257
450
411
387
494
1566
1975
691
739
2046
959
2709
2133
575
579
2183

68.47
69.27
96.04
107.36
114.02
124.6
129.99
132.18
137.13
139.76
142.38
143.99
145.59
148.22
149.67
151.42
162.66
163.68
177.74
186
282.63
285.71
293.46
294.72
300.2
305.25
469.29
471.13
471.89
472.39
473.02
477.62

162

Primer set 8F-1525R

Terminal

fragment Peak area

657
940
1204
830
606
661
359
483
493
1854
4767
782
3385
827
416
880
813
760
421
1619
3688
1120
1136
3804
690
4696
807
681
361
377
919
883

Table l (cont’d)

163

485.2 482
491.58 3673
495.22 4159
507.41 1448
509.11 386
511.7 902
527.56 1243
589.32 1258
Msp I
Primer set 8F-1392R Primer set 8F-1525R
Terminal

Terminal fragment Peak area fragment Peak area
39.27 350 40.22 773
69.43 392 41.82 1979
70.22 715 62.24 262
87.14 808 79 1191
88.89 431 79.96 458
98.58 356 92.72 730
103.41 1281 95.38 1182
139.7 1561 105.72 407
142.33 981 154.14 578
145.4 979 184.14 512
147.74 407 211.8 10271
149.35 1456 223.6 4543
163.74 681 227.46 1379
305.51 591 230.27 2562
428.21 804 302.95 829
489.03 797 311.26 538
495.92 1687 313.52 2523
499.51 1116 363.79 675
514.3 978 365.89 4330
520.65 1511 370.09 462
528.62 1107 371.64 589
376.01 1816
378.13 781
382.67 1439

Table l (cont’d)

384.95 597
402.19 11289
410.04 643
487.73 3401
534.55 1739
573.88 3937

164

Rsa I
Primer set 8F-1392R Primer set 8F-1525R
Terminal
Terminal fragment Peak area fragment Peak area
34.32 523 41.02 1230
39.43 233 84.59 1231
40.86 528 86.02 2660
69.43 863 180.74 2211
86.02 1102 297.8 634
103.41 1754 418.38 467
105.41 540 437.48 2407
437.13 408 446.52 1228
449.98 1093 457.39 11823
458.09 3006 461.62 7231
461.94 1266 466.67 524
469.69 889 470.68 3984
471.31 528 473.39 1164
479.65 1980 474.41 625
481.62 2676 475.77 682
485.56 1683 479.33 3074
491.89 763 482.89 2502
493.15 880 484.75 1352
497.26 4305 485.6 1140
502.81 5433 486.78 2397
488.48 934
491.73 5665
492.99 9888
498.18 1304
499.91 2670

Table l (cont’d)
502.74 609

517.18 661

 

SEDIMENT S18 (8 m)
Hha I
Primer set 8F-1392R Primer set 8F-1525R
Terminal Terminal
fragment Peak area fragment Peak area
39.11 486 69.74 406
40.06 439 96.24 326
41.82 1059 124.44 497
62.24 707 139.75 707
69.43 265 142.38 2026
86.98 920 145.6 1103
92.72 702 185.86 662
103.2 2242 282.52 537
114.87 478 294.71 822
117.9 405 305.38 816
123.18 400 491.28 8601
124.06 681 495.4 1740
137.39 648 527.51 622
151.9 433
169.36 885
206.86 556
212.01 1583
225.77 511
370.13 797
379.06 1522
380.77 476
389.21 847
402.52 2089
488.06 1854
573.96 2869

165

 

Table 1 (cont’d)

Primer set 8F-1525R

Terminal

fragment Peak area
40.22 459
41.82 1056

79 600

95.38 336
211.87 3515
216.08 562
221.43 444
223.44 1962
230.28 763
313.32 665
365.96 668
382.67 641
402.54 5061
410.02 583
487.12 1726
572.71 4911

Msp I
Primer set 8F-l392R
Terminal
fragment Peak area
39.27 392
64.64 305
69.43 239
70.38 553
87.14 786
88.89 397
103.41 1815
114.76 694
123.95 494
136.29 463
137.16 658
139.64 1357
142.26 677
145.32 897
149.41 1249
151.6 853
163.76 460
169.35 614
468.61 573
489.02 739
495.76 4636
499.66 913
514.73 722
520.73 995
529.14 707
Rsa I
Primer set 8F -1392R
Terminal
fragment Peak area
34. 16 1825
39.27 356
40.7 412

Primer set 8F-1525R

Terminal

fragment Peak area
40.45 753
84.48 445
85.94 855

166

Table l (cont’d)

56.82 228
63.2 860
64.16 281
67.03 350
69.43 504
85.86 696
103.36 1349
117.89 393
190.74 386
458.86 1726
462.87 577
479.56 531
480.37 826
482.16 1747
486.73 1298
497.96 2285
504.21 3302
SEDIMENT M3 (6 m)
Hha I
Primer set 8F-1392R
Terminal
fragment Peak area
39.27 337
69.59 452
86.98 324
103.36 1117
107.51 686
204.18 1887
206.53 16851
234.12 1155
369.72 10465
469.53 1286
470.48 537
471.12 668
471.92 561

180.88

433.2
437.76
456.37
458.09
462.46
466.53
470.66
478.91
483.21
484.37
487.18
488.34
493.31
498.05
499.94

31.16

204.2
206.55
234.24
362.53
369.93
470.48
471.12
472.08
473.53
475.45
476.58
515.49

167

1087
1456
818
508
6597
2016
424
1170
5683
1175
915
1196
1666
8095
856
763

Primer set 8F-1525R

Terminal

fragment Peak area

8097
2830
31473
1450
611
11955
1614
439
1011
1122
571
330
1004

Table 1 (cont’d)

472.89 893
474.01 463
475.62 1045
515.01 515
515.65 312
516.78 876
566.8 118058
Msp I
Primer set 8F-1392R
Terminal
fragment Peak area
31.16 11215
39.84 373
69.84 488
87.21 308
103.24 1166
107.36 773
206.42 1284
369.18 858
420.99 2389
421.89 694
423.41 2374
493.84 152150
Rsa I
Primer set 8F-1392R
Terminal
fragment Peak area
34.32 1630
38.63 200
39.43 276
43.26 530
63.36 394
69.59 425
103.36 765
107.36 754
119.79 2910

566.94

158.92
206.56
419.24
420.74
422.41
430.16
493.86

119.81
329.02
397.66
399.48
401
433.17
469.24
476.12
651.4

168

115605

Primer set 8F-1525R

Terminal

fragment Peak area

604
877
2198
1073
1453
630
167712

Primer set 8F-1525R

Terminal

fragment Peak area

4655
974
5742
1691
3466
47840
445 '
66974
6278

Table 1 (cont’d)

395.19
396.53
397.57
398.32
398.91
400.26
403.24
433.36
475.77

1978
1836
854
700
1361
2177
726
22555
63608

169

Table 2 Chapter 4 bacterial T-RFLP data

170

SEDIMENT 1At
Hha I After eliminating
Data from 4/10 Data from 4/27 inconsistent peaks
Terminal Peak Terminal Peak
fragment area fragment area Terminal fragment
36.81 764 37 1386 37
38.14 290 58.18 445 58.18
57.91 300 59.85 398 59.85
59.62 275 95.14 2971 95.14
95.03 2656 97.01 473 98.48
98.62 479 98.48 636 155.84
155 1294 111.28 417 197.9
197.49 825 155.84 1711 201.42
201.33 3975 197.9 751 204.53
204.27 750 199.07 562 206.46
205.82 1574 201.42 4789 215.45
215.27 578 204.53 1232 218.94
218.94 1631 206.46 1743 232.84
233 2509 208.39 689 369.78
370.1 1592 215.45 751 374.41
373.97 599 218.94 2379 375.56
376.14 506 232.84 4412 380.78
381.12 4935 364.02 605 539.79
539.61 710 368.2 537 540.75
540.99 1013 369.78 2619 542.65
542.71 830 374.41 1134 544.57
544.27 1264 375.56 1552 548.42
546.7 304 377.59 1004 565.19
548.44 1739 379.48 2985 569.17
565.22 841 380.78 6001 581.07
567.19 756 406.8 703 585.65
569.16 1938 415.2 481
571.5 1347 539.79 2472
580.94 1815 540.75 2670
584.98 793 542.65 997
585.9 1071 544.57 3302
548.42 4334

Table 2 (cont’d)

Data from 4/10

34.33
34.33
75.17
77.64
93.24
149.35
155.66
157.83
161.86
163.54
165.73
172.8
207.25
230.96
448.35
450.15
463.68
466.17

Peak
area

3813
4219
518
393
635
630
1044
1231
628
1880
689
654
1256
546
2232
835
1071
1172

554.07 824
556.02 2011
560.26 1546
561.9 705
565.19 2587
569. 17 7632
574 555
575.01 347
576.85 1346
581.07 4862
585.65 4047
590.78 1704
599.42 544
Data from 4/27
Terminal Peak
fragment area
75.84 496
93.1 518
134.16 472
149.32 471
156. 16 1105
158.23 1489
161.9 779
163.82 1855
165.9 880
170.55 388
172.93 649
206.22 520
207.56 904
229.18 410
285.53 410
286.46 470
288.43 926
376.58 429

171

After eliminating
inconsistent peaks

Terminal fragment

75.84
93.1
149.32
156.16
158.23
161.9
163.82
165.9
172.93
448.48
450.29
463.91
466.19
481.85
482.97
491.97
495.76
505.44

Table 2 (cont’d)

481.62
482.87
490.84
492.34
494.17
496.17
505.81
507.64
512.44
514.09
519.54
520.86
521.69
522.51

523.5
524.98
527.12
527.78
529.59
532.22
534.03
568.91

Data from 4/10

33.38
34.14
34.33
56.8
121.94
131.64

514
1033
655
1469
1579
3018
1909
2762
2228
3940
1362
1157
792
801
1193
1635
356
782
861
804
656
999

Peak
area

1045
2653
4219
440
526
560

437.82
448.48
450.29
463.91
466.19
481.85
482.97
490.15
491.97
493.64
495.76
505.44
507.26
512.25
513.76
520.87
522.54

523.6
527.98

529.8
532.22
534.03
569.24

570.9

56.89
122.1
130.43
131.69
298.9
421.23

577
3242

830
1361
1364

873
1094

936
1885
1794
4103
2589
3204
2979
4077
3131
2075
2780
1326
1358
1099

698
1802

858

Data from 4/27

Terminal Peak
fragment area

595
827
308
830
892
628

172

507.26
512.25
513.76
520.87
522.54

523.6
527.98

529.8
532.22
534.03
569.24

After eliminating
inconsistent peaks

Terminal fragment

56.89
122.1
131.69
446.11
451.71
453.07

Table 2 (cont’d)

445.83
451.71
453.01
454.33
459.58
460.74

464.7
473.37
476.91
478.51
489.12
496.32
502.47

SEDINIENT 2At

Data from 4/10

36.62
59.86
69.97
81.55
95
143.23

144.05,

205.58
209
211.19
211.98
287.63
370.09
379.7

504
1746
667
1837
1186
998
1049
2001
1569
5441
2377
4727
553

Peak
area

1403
1495
599
403
2019
372
472
2064
1127
452
598
2686
4679
4066

428.68
446.11
451.71
453.07
454.14
459.46
460.98
464.64
466.17
466.78
473.07
476.63
478.73
489.04
495.93

502.3

.36.81
56.61
57.73
59.96
62.19
72.72
82.42
94.96
97.21
98.51

101.05

176.18

204.08

206.01

665

941
2973
1250
2857
2112
1593
1531

350

765
3250
2875
8997
3468
7416
1357

Data from 4/27

Terminal Peak
fragment area

23 82
402
621

2806
763
345
923

4799
396
410
474
776
586

3 626

173

454.14
459.46
460.98
464.64
473.07
476.63
478.73
489.04
495.93

502.3

After eliminating
inconsistent peaks

Terminal fragment

36.81

59.96

94.96
204.08
206.01
209.29
211.69
369.79
381.23
548.71
568.95
585.43

Table 2 (cont’d)
381.72 3744
548.85 4139
567.32 5288
569.311368
582.8 756
585.2 1744

Msp I
Data from 4/10
Terminal Peak
fragment area
65.34 582
127.91 689
131.15 1402

209.29 958
211.69 1079
341.57 930
369.79 6924
375.57 1154
379.05 6091
381.23 4245
406.67 431
407.25 490
484.25 759
538.84 2273
541.69 1058
543.91 860
548.71 8494
552.9 958
555.66 648
556.31 401
557.28 385
558.59 833
560.55 872
565.48 5393
568.95 26683
576.97 479
578.48 715
579.83 571
585.43 5995
589.18 715
590.73 375
596.25 622
Data from 4/27
Terminal Peak
fragment area
33.11 610
42.88 313
65.16 620

174

After eliminating
inconsistent peaks

Terminal fragment
65.16
128.28
131.59

Table 2 (cont’d)

149.18
162.53
207.88
209.75
288.12

432.3
466.01
494.01

891
1088
938
803
858
788
1423
7080

496.01 10282

502.82

513.1
528.11
529.26
598.26
600.15

Data from 4/10

33.38
54.12
56.8
130.09
430.54
459.89
460.89
474.61

1430
5351
1469
1413

738
1583

Peak
area

544
260
270
426
1496
1806
1382
3550

476.91 13210

128.28
131.59
149.61
150.88
162.93
165.16
207.7
288.44
290.41
431.77
438.95
465.88
493.65
495.62
501.23
502.29
511.38
513.34
515.46
528
529.81
599.06
600.8

33.11

56.98
284.44
430.34
459.46
460.99
474.53
476.47
480.66

491
1275
506
397
1 151
541
323
1004
445
968
471
1003
9200

12666

637
2003
952
3459
1380
1922
1572
669
1683

Data from 4/27

Terminal Peak
fragment area

335
525
578
3098
2249
2410
5253

13815

1915

175

149.61
162.93
207.7
288.44
431.77
465.88
495.62
502.29
513.34
528
529.81
600.8

After eliminating
inconsistent peaks

Terminal fragment

33.11

56.98
430.34
459.46
460.99
474.53
476.47
484.21
496.38

Table 2 (cont’d)

After eliminating
inconsistent peaks

Terminal fragment

484.34 719 484.21 1112
496.66 4619 494.86 2902
498.15 2039 496.38 4502
501.24 678
SEDIMENTZBb
HhaI
Data from 4/10 Data from 4/27
Terminal Peak Terminal Peak
fragment area fragment area
36.04 1725 35.16 1152
60.24 625 37 3995
62.17 299 55.77 312
91.23 573 56.7 334
95.36 1391 57.81 444
205.74 1099 59.85 1840
207.44 416 61.89 1229
287.79 1710 82.43 934
370.08 1930 86.6 613
373.17 246 91.3 1663
373.79 584 94.95 4230
377.51 1375 101.02 461
380.46 2387 143.21 331
381.86 1690 175.37 629
548.85 1439 176.13 1177
550.43 752 182.67 940
567.15 1456 206.11 3145
569.31 6839 208.96 1179
357.44 1850
359.81 660
362.29 547
363.01 844
369.64 6851
375.14 1302
376.87 3732
379.48 3913

176

61 .89
91.3
94.95
206.11
369.64
380.5
548.71
569. 12

Table 2 (cont’d)

380.5 5738
403.43 654
415.22 741
540.58 3869
543.91 1013
548.71 7291
554.85 311
556.14 568
559.4 942
561.04 551
563.34 836
565.31 2106
569.12 16391
574.95 655
576.63 247
577.81 697
581.69 1543
584.23 2458
587.3 1530
589.35 392
590.21 517
591.59 256
592.62 432
594.17 515
MspI
Data from 4/10 Data from 4/27
Terminal Peak Terminal Peak
fragment area fragment area
34.14 5938 64.73 300
43.12 242 65.84 661
65.07 708 93.1 297
93.05 272 . 139.68 818
139.3 698 149.32 656
149.18 765 151.23 1510
150.84 1005 158.87 330
159.51 799 160.3 673

177

Consistent peaks

Terminal fragment

64.73
93.1
139.68
149.32
151.23
174.79
207.65
217.12

Table 2 (cont’d)

161.86
163.54
174.56
207.97
210.02
217.16
491.84

494

1193
697
450

1158
626
885

1021

5568

496.01 11580

502.33
503.5
509.98
512.14
514.13
516.29
519.44
521.76
523.25
525.07
528.06
529.88
598.58
600.66
31

Data from 4/10

54.12
56.8
126.25
166.04
430.45

519
442
681
2227
3252
3993
961
967
490
375
1090
1940
807
1390

Peak
area

375
806
447
967
2161

432.7 1003

445.5

2419

162.38

164.3
174.79
176.33
207.65
208.99
217.12
465.87
491.82

493.8
495.62
500.47

502.6
509.87
512.13
514.25
516.37
519.69
521.66
523.47
525.13
528.15
529.81
598.87
600.61

56.98
166.04
424.61
430.43
445.59

448.3
449.36

1447
742
585
507
600
919

1019
869

1289

7094

11803

1212

1418

1047

2873

4345

4283

1447

1234
670
616

1934

2377
881

2509

Data from 4/27

Terminal Peak
fragment area

651
850
572
2621
331 1
420
464

178

491.82

493.8
495.62

502.6
509.87
512. 13
514.25
516.37
519.69
521.66
523.47
525.13
528.15
529.81
598.87
600.61

Consistent peaks

Terrrrinal fragment

56.98
166.04
430.43
445.59
451.17
451.77
452.83

Table 2 (cont’d)

447.13
450.72
451.87
452.85
455.47
456.95
459.58
464.37
467.35
470.53
474.43

1938
782
748

1002

1431
502

4295
765

1083
793

2903

476.91 16521

493.98
496.14
501.44

22

1786
7830
2565

SEDIMEN T 4A
Hha I
Data from 4/10

Terminal Peak
fragment area

37
40.08
41.04
56.39
69.79

205.58
287.46
369.89
567.19

604
454
716
339
524
3964
2430
2742
4785

569.36 22244

451 . 17
451.77
452.83

455.4
456.77
459.65
463.75
464.82
467.25
471.28
474.65
476.57
483.77

493.8
495.77
499.42
501.09

37
205.98
369.77
538.98
549.15
567.18
568.99
579.32
584.72

588.8
589.49

923
613
1042
1193
498
3754
348
468
1010
419
3158
16152
800
2428
7000
1084
2358

Data from 4/27

Terminal Peak
fragment area

773
7695
6867

715

626
6140

36653

579

539

317

459

179

455.4
456.77
459.65
464.82
467.25
474.65
476.57

493.8
495.77
501.09

Consistent peaks

Terminal fragment

37
205.98
369.77
567.18
568.99

Table 2 (cont’d)

Msp I
Data from 4/10
Terminal
fragment
40.09
41.05
207.97
209.99
491.84
493.84
496.01
598.53
600.61
Rsa I
Data from 4/10
Terminal
fiagment
33.38
40.09
126.5
430.56
474.43

Peak
area

729
951
499
684
455
7042
21256
796
1291

Peak

area
1751
3 19
504
4792
2048

Data from 4/27
Terminal Peak
fragment area
32.92 705
207.61 771
209.56 856
461.9 312
491.82 983
493.5 19572
495.47 48311
541.04 334
542.15 668
545.33 879
547.09 484
549.97 519
551.26 493
552.7 520
576.84 512
584.93 686
589.69 417
591.39 445
593.45 569
594.31 239
595.51 562
598.62 2829
600.52 3639
Data from 4/27
Terminal Peak
fragment area
126.58 638
421.02 592
430.46 19894
474.53 5718
476.63 47428

180

Consistent peaks

Terminal fragment

207.61
209.56
491.82

493.5
495.47
598.62
600.52

Consistent peaks

Terminal fragment

126.58
430.46
474.53
476.63

476.91 15222

488.71
499.45
500.97
584.98
589.72

543
430
492
1036
434

181

Table 3 Chapter 4 Archaeal T-RFLP data

SEDIMENT 1At
Hae III Hha I
Terminal Terminal
fragment Peak area fragment
69.23 1937 60.15
70.7 702 62.19
175.56 675 90.03
207.8 2298 175.56
208.99 6359 195.23
214.4 19069 224.11
240.69 19001 327.83
246.13 1864 338.14
272.52 1958 357.9
298.9 699 359.52
317.46 2716
SEDIMENT 2At
Hae III Hha I
Terminal Terminal
fragment Peak area fi'agment
69.46 377 62.08
70.74 1093 88.77
175.56 830 89.86
208.99 1918 175.56
214.55 18797 196.14
240.27 30076 327.68
245.96 1923 338.74
272.36 2627 358.78
298.76 2265
311.15 558
317.17 1456

Peak area

608
7258
1985

459
5898

712

22176
7577
2502
6201

Peak area

15787
563
1224
1041
719
14024
9229
9970

182

RsaI

Terminal
fragment

40.78
78.43
80.25
241.71
262.15
294.96

RsaI

Terminal
fragment

40.59
242.55
262.35
263.67
295.09

Peak area

2230
2272

716
2598
7115
6954

Peak area

3276
7278
1545
1049
9036

Table 3 (cont’d)
SEDIMENT 2Bb

38.12
191.34
209.89

214.4

240
242.15
246.37

311.1

316.96

SEDIMENT 3At

38.14
69.41
70.7
175.25
208.99
214.55
240.41
245.97
272.49
298.87
317.71

Peak area

627
8522
1267

45307
6221
2873

670

827
1529

Peak area

537
1139
778
698
5855
10713
36801
1996
2211
1302
1087

Hha I

Terminal
fragment

38.13
52.04
60.23
62.26
76.05
196.29
327.31
333.13
338.48
358.64

Hha I
Terminal
fragment

38.13
62.08
88.77
89.86
175.56
195.95
224.29
327.69
333.35
339
359.09
362.72

Peak area

1444
399
557

4191
766
897

48106

9991

4972

2394

Peak area

574
15397
828
3920
721
471
2502
17300
588
1287
15655
665

183

Rsa I
Terminal
fragment

40.78
91.12
242.4
258.21
262.24
295.28
297.1

RsaI

Terminal
fragment

40.78
242.55
262.35
295.11

Peak area
17001

429
3482
8992
7100
1326
1 151

Peak area

3509
773
2295

15264

APPENDIX C

Merry S. Riley, Vaughn S. Cooper, Richard E. Lenski, Larry J. Forney, and Terence L.
Marsh Rapid phenotypic change and diversification of a soil bacterium during 1000
generations of experimental evolution. Microbiology 2001 147: 995-1006.

184

 

Microbiology (2001), 147, 995-1006

Printed in Great Britain

 

Rapid phenotypic change and diversification of
a soil bacterium during 1000 generations of
experimental evolution

Merry S. Riley,"2 Vaughn 5. Cooper,‘ Richard E. Lenski,‘ Larry J. Forney‘

and Terence L. Marsh‘-3

Author for correspondence: Terence L. Marsh. Tel: +1 517 432 1365. Fax: +1 517 432 3770.

e-mail: marsht@pilot.msu.edu

 

1.2-3 Center for Microbial
Ecology‘, Department
of Crop and Soil
Science2 and
Department of
Microbiology’, Michigan
State University, East
Lansing, MI 48824. USA

‘ Department of Biology,
University of ldaho.
Moscow, ID 83844, USA

Evolutionary pathways open to even relatively simple organisms. such as
bacteria. may lead to complex and unpredictable phenotypic changes. both
adaptive and non-adaptive. The evolutionary pathways taken by 18
populations of Ralstonia strain W041 while they evolved in defined .
environments for 1000 generations were examined. Twelve populations
evolved in liquid media. while six others evolved on agar surfaces. Phenotypic
analyses of these derived populations Identified some changes that were
consistent across all populations and others that differed among them. The
evolved populations all exhibited morphological changes in their cell

envelopes. including reductions of the capsule in each population and reduced
prostheca-like surface structures in most populations. Mean cell length
increased in most populations (in one case by more than fourfold), although a
few populations evolved shorter cells. Carbon utilization profiles were variable
among the evolved populations, but two distinct patterns were correlated
with genetic markers introduced at the outset of the experiment. Fatty acid
methyl ester composition was less variable across populations, but distinct
patterns were correlated with the two physical environments. All 18
populations evolved greatly increased sensitivity to bile salts, and all but one
had increased adhesion to sand; both patterns consistent with changes in the
outer envelope. This phenotypic diversity contrasts with the fairly uniform
increases in competitive fitness observed in all populations. This diversity may
represent a set of equally probable adaptive solutions to the selective
environment: it may also arise from the chance fixation of non-adaptive
mutations that hitchhiked with a more limited set of beneficial mutations.

 

 

Keywords: Ralstonia, phenotypic radiation, diversity, mutation

 

 

INTRODUCTION

Micro-organisms have proven useful for studying the
mechanisms and consequences of evolution. By virtue of
their small size, fast growth rates and comparatively
simple genetic systems, questions that could not be
addressed with macro-fauna can be answered in a robust
and statistical manner with micro-organisms (Helling et
al., 1987; Dykhuizen, 1990, 1993; Lenski 86 Travisano,
1994; Lenski et al., 1998; Rainey 8c Travisano, 1998).
Moreover, because of the resilience of some micro-

 

Abbrevlatlons: FAME, fatty acid methyl ester; SEM. scanning electron
microscopy; TEM, transmission electron microscopy.

organisms to cryogenic preservation, a detailed his-
torical record of an evolving population can be preserved
and used comparatively in the dissection of evolutionary
processes (Lenski 8c Travisano, 1994).

Our present work extends previous investigations of the
experimental evolution of a soil bacterium (Korona et
al., 1994; Korona, 1996; Nakatsu et al., 1998). The
original intent of this work was to extend investigations
performed on domesticated Escherichia coli to a
recently isolated undomesticated soil bacterium. In
addition, two different environmental conditions were
used as selective regimes during the propagation of lines
derived from this isolate for 1000 generations. Twelve
replicate populations were maintained in a liquid shake

 

0002-4335 0 2001 SGM

185

M. S. RILEY and OTHERS

 

ask and six populations were cultivated on agar. The

former provided a mass-action environment that is
highly homogeneous throughout, while the latter pro-
vided a structured environment with complex gradients
of nutrients as well as metabolites produced by the
bacteria themselves. The soil bacterium used as the
ancestral founder of all these populations is strain
TFD41, which was identi ed as a Ralstonia species
based on its 16S rRNA sequence (Nakatsu et al., 1998).
Previous work described a substantial increase in
competitive tness of all 18 evolved populations,
measured relative to the common ancestor in the
experimental environments (Korona et al., 1994).
Variations in colony morphology were noted (Korona
et al., 1994), as were certain genome changes, both
chromosomal and plasmid encoded (Nakatsu et al.,
1998»

In this paper, we extend these observations with a
systematic examination of cell morphology, substrate
utilization, bile salts sensitivity, adhesion properties and
fatty acid methyl ester (FAME) analysis of the evolved
papulations and their ancestor. Our objectives were to
identify common phenotypic motifs that may indicate
parallel genetic adaptations, as well as differences among
populations within and between the two selective
regimes. The notion of an adaptive landscape (Wright,
1932, 1988) suggests that a population of organisms may
have many potential evolutionary solutions to a selective
challenge. It seems reasonable, too, that the diversity of
potential solutions would be correlated with the com-
plexity of the selective environment. Consistent with
this postulate is the observation that variation in colony
morphology was greater amongst the populations
evolved on solid media than those evolved in liquid
(Korona et al., 1994; Korona, 1996).

METHODS

Strains and media. Strain TFD41 was originally isolated from
soil based on its ability to grow on 2,4-dichlorophenoxyacetate
(Tonso er al., 1995). The isolation of streptomycin and
nalidixic acid resistant mutants of TFD41 and the maintenance
of 18 replicate populations for 1000 generations have been
described previously (Korona er al., 1994). The ancestor,
antibiotic resistant mutants and isolates from generation 1000
of the evolved populations were maintained at —80 °C in
glycerol stocks. Cultures derived from frozen stocks were
maintained on plates for no longer than 2 weeks. Cells were
routinely grown on either nutrient agar or RZA agar at 30 °C
unless otherwise indicated.

Microscopy. Light microscopy was performed on a Zeiss
Axioskop microscope. Cells were taken from freshly streaked
nutrient agar plates that had been incubated at 30°C for
48 72 h. The presence of outer capsule material was detected
with an India ink stain (Doetsch, 1981) and representative
micrographs were made at 1000x magni cation with an oil
immersion lens (Zeiss). Cells were evaluated for the presence
of capsule as well as cell shape and mean length. Images were
captured Son Kodak TMAX Im.

Scanning electron microscopy (SEM) was performed on cells
grown on nutrient agar plates as described for light mi-
croscopy. The cells were xed in 4 °/o glutaraldehyde, mounted

on polylysine-coated coverslips, dehydrated in an ethanol
series (2.5%, 50 %, 7S % and 95%), critical-point dried and
sputter coated with gold. At each step the cells were treated as
gently as possible so as to preserve the integrity of the outer
cellular architecture. The mounted cells were viewed on a
JEOL scanning electron microscope at several different
magni cations. Images were captured and stored elec-
tronically. At least ten elds, at approximately 2500x
magni cation, were viewed for each lineage to obtain a
representative sample of morphology. Cell length of well-
isolated cells was measured from end to end, excluding surface
appendages. The mean cell length was calculated from a
minimum of25 cells in a minimum of two elds. The mean cell
length of an evolved population was deemed to depart
signi cantly from that of the ancestor if there was no overlap
between the respective meansi twice the standard error.

Transmission electron microscopy was performed on the
ancestor and three evolved populations. The cells were grown
as described above for light microscopy, scraped from the agar
plate and pelleted by centrifugation. The pellets were
resuspended in a 4% agar solution and allowed to solidify.
The agar was diced into lmrna squares then xed in 2%
glutaraldehyde buffered with 100 mM sodium cacodylate
(pH 7'2) for 2 h. The xed cubes were washed three times in
100 mM sodium cacodylate buffer (pH 72), post- xed in 1%
osmium tetroxide for 1'5 h and then washed four times in
deionized water. The specimens were dehydrated in a graded
acetone series and in ltrated in a graded Quetol resin series
and embedded. Preparations were stained with uranyl
acetate/ lead citrate. All preparation steps were conducted at
room temperature. Scanning and transmission electron micro-
scopies were performed at the Michigan State University
Center for Electron Optics.

BIOLOG assays. Carbon utilization pro les of the ancestor and
evolved populations were determined using BIOLOG GN
plates and the protocol provided by the vendor (modi ed to
accommodate a Ralstonia sp.). Brie y, an aliquot of an
overnight culture grown on 1/ 3x Trypticase soy broth was
transferred to fresh media and cells were grown at 30 °C in a
reciprocating shaker to mid-exponential phase. Cells were
harvested by centrifugation and resuspended in 0145 M NaCl
to an of 0D,,o 1'0. One lineage, L16, had anomalously low cell
density at an OD,” of 10 and therefore was adjusted to OD”.
20. Cells were then starved for 15 min at room temperature
prior to the inoculation of the microtitre plates with 150 pl cell
suspension per well. Each strain was tested in duplicate. The
plates were incubated at 30 °C and the OD,” measured using
a Bio-Kinetics microtitre plate reader (model EL 312E, Bio-
Tek Instruments) every 2h for a total of 8 h. Plates were
vigorously shaken prior to each reading and the ODm was
corrected for the zero substrate control well. Some substrates
had a variable response where one of the replicates was
unambiguously positive (005,, > 0-1 after zero substrate
correction) and the other borderline. These substrates are
indicated (see Fig. 3).

Analysis of BlOLOG data. BIOLOG optical density (OD)
readings from the plate reader were imported into a database
and fed into a Microsoft Visual Basic-Excel program that
approximated the area under the growth curve for each of the
9S substrates for each genotype replicate. Following the
general method of Guckert et al. (1996), the program calculates
a trapezoidal area that approximates the result of tting and
integrating each individual growth curve. This curve area
approach collapses several values from a grOWth curve into a
single value and integrates several properties (e.g. duration of
lag, growrh rate, yield). Individual values for each substrate,

 

186

Rapid phenotypic evolution

 

 

 

--» Fig. 1. Light and scanning electron micro-
scopy of the ancestor, Ralstonia strain TFD41,
and several evolved lines. (a) Light micro-
graph of ancestor. (b) Light micrograph
of evolved L1. d) Scanning electron
micrographs of ancestor. (e-h) Scanning
electron micrographs of evolved lines L6
L12, L16 and L18. Bars,

 

area , were subjected to a
hierarchical cluster analysis (SYSTAT, v. 7.0, SPSS) to determine

which we refer to as catabolic

the relationships tween the ancestors and thee eov v
lineages based on their patterns of carbon source utilization.
The area data were expressed on a continuous numeric scale,
and therefore normalized Euclidean distances were calculated
for our clustering analysis. Ward 3 linkage meth oda(W rd,
1963) was applied to adjust for covariance and to focus on
mean values within clusters. Other linkage methods were also
tested and none gave substantially different cluster patterns
(data not shown). The source codes for the above procedures
are available by request from the author (V. S. C. at
cooperva@pilot.msu.edu).

Bile salts MIC. Overnight cultures grown on 1/3x TSB were
diluted 1:100 into fresh broth containing a range of bile salts
(Difco) concentrations (01 g g "). The cultures were
incubated at 30 °C and scored for growth at 24, 48 and 72 h.
Any increase In turbidity above the initial OD was scored as a
positive. All evolved populations were tested In duplicate, and
no differences were detected between these rep itca es.

Adhesion. The adhesion assay was based on the retention of
bacterial cells in a saturated sand matrix supported in a
column and has been described previously (DeFlaun er al.,

1990). Brie y, A+N minimal medium (Wyndham, 1986)
supplemented with 30 mM L-aspartic acid (AN.asp) as the
sole carbon source was inoculated with the ancestor or an
evolved population from a fresh nutrient agar plate and grown
overnight at 30 °C In a reciprocating shaken? Prior screening
indicated that the ancestor and all 18 evolved lineages could
utilize L—aspartic acid. The culture was diluted into es AN-
asp and grown to late-exponential phase. The cells were
harvested by centrifugation, resuspended In phosphate
buffered saline (PBS) and adjusted by dilution with PBS to an
OD". of 05 06 (1 cm light path).

Bio-Rad Econocolumns (5 x 1-5 cm inner diameter) were
packed with 11 g ~50 +70 mesh sand (Sigma S-9887) and
rinsed with approximately 15 ml PBS. The cell suspension
(3-5 ml) was pipetted gently onto the column, an entire pore
volume was allowed to pass into the matrix (3'5 ml void
volume) and then ow was stopped. The loaded column was
incubated undisturbed at room temperature for 1 h to allow
cells to bind to the matrix. The column was then eluted with
14 ml PBS, collected In four fractions. The fraction of cells
eluted was determined spect rophotometrically In a 1c mpat
length quartz cuvette (Hewlett Packard 8452A Diode Array
Spectrophotometer) and the fraction of cells bound to the
column was calculated as fraction bund = l—total OD

 

187

M. S. RILEY and OTHERS

 

Table 1. Morphological characteristics of the evolved populations

 

Minimal medium with 2,4.dichlorophenoxyacetate as the carbon source was used for selection (see Methods).

 

 

 

 

Strain' Selection Colony Light SEM
typct microscowt
Cell surface Mean cell length 2(SE) ll
(nuns

Ancestor Soil + C-i- P+ + + + Capsule 1'95 014
L1, Str Liquid - C- P- Smooth 1'8 0'12
L2, Nal Liquid - C- P- Smooth 2'95! 059
L3, Str Liquid j: C— P- Smooth 5'64! 0'92
L4, Nal Liquid + C- P+ j: Capsule 1'39! 009
L5, Sn Liquid + C— P+ Smooth 4-08! 068
L6, Nal Liquid - C- P— Smooth 1'46! 013
L7, Str Liquid — C- P- Smooth 3'44! 052
L8, Nal Liquid - C- P- iCapsuIe 3'21! 0'46
L9, Str Liquid - C— P— Smooth 5'19! 0'81
L10, Nal Liquid - C- P- j: Capsule 3'59! 075
L11, Str Liquid - C- P- Smooth 3'23! 047
L12, Nal Liquid - C- P— Smooth 1'88 0'16
L13, Str Agar + C— P+ j: Capsule 4'53! 064
L14, Nal Agar — C- P- j: Capsule, blebs 1'97 034
L15, Str Agar i C— P— :t Capsule 1'51! 0'08
L16, Nal Agar - C- P- Smooth 9'14! 273
L17, Str Agar - C- P- Smooth 4'08! 07
L18, Nal Agar j: C—- P+ Smooth, blebs 1'93 207

 

 

’Str, streptomycin resistant; Nal, nalidixic acid resistant.

1' + , Highly mucoid; :t , intermediate mucoid; -, non-mucoidal.

# C+ , detectable capsule with India ink; C- , capsule not detectable with India ink; P-i- + , prostheca-like appendages present in greater
than 75% of cells; P+, prostheca-Iike appendages present in less than 25% of cells; P , prostheca-like appendages absent.

SMean of at least 25 cells from a minimum of two SEM elds; a total of ten elds of each population were viewed.

ll Standard error (SE) -= standard deviation/(Vii).
! Statistically different from the ancestor.

--r

 

eluted/total OD loaded. All populations were tested in
triplicate.

FAME analysis. This was conducted on a subset of the derived
lineages (L1, L5, L6, L8, L9, L11, L12, L16, L17 and L18), the
original ancestor as well as the Str" and Nal "derivatives of the
ancestor. Strains were streaked from —80 °C stocks onto
l/3x Trypticase soy agar, incubated at 30 °C and harvested
after 2 d by scraping the plates with a sterile spatula directly
into dry-heat-sterilized glass culture tubes. The resulting cell
pellet was stored at —80 °C until assayed. Each sample was
extracted and run in triplicate as described previously by
Sasser (1997). FAME analysis was performed on a Hewlett
Packard 5890 series 11 gas chromatograph equipped with a FID
detector. A Hewlett Packard Ultra2 column (cross-linked 5 °/o
phenyl methyl silicone) of 25 cm x 022 mm with 033 um Im
thickness was used with ultra—high purity hydrogen as the
carrier gas at 50 ml min'1 using a 50:1 split ratio. The initial
temperature, 170 °C, was ramped to 270 °C at a rate of 5 °C
min". The column was baked for 2 min at 300 °C after each
run. The injector and detector temperatures were maintained
at 250 °C and 300 °C, respectively. FAME data were analysed
with the Sherlock MIDI system and the Dend rogram program
(MIDI). The former identi es the closest microbial relative
based on the fatty acid pro les, while the latter compares the

lineages based on a clustering analysis using unweighted pair-
group method with arithmetic averages (UPGMA).

RESULTS
Microscopy

One of the rst phenotypic changes noted in the evolved
populations was colony morphology (Korona et al.,
1994). Because these differences were suggestive of
changes in capsule production, the cell morphology of
the ancestor and evolved populations was systematically
assessed using both light and electron microscopy. Fig.
1 (a) shows the cell morphology of the ancestor visualized
with a capsule counter-stain (India ink). Cells of the
ancestral strains were heavily encapsulated, as evidenced
by the light-bright halo surrounding the cells. In
addition, very long prostheca-like appendages were
present on most cells. All of the evolved populations
showed a substantial reduction in encapsulation as
judged by light microscopy. Four lineages (L4, L5, L13
and L18) showed some indication of prostheca-like
appendages but the length of the appendages and the

 

188

Rapid phenotypic evolution

 

 

 

Fig. 2. Transmission electron mlcrographs of the ancestor and
two evolved lines, L16 and L18. (a, b) Ancestor. (c) Evolved line
L16. (d) Evolved line L18. Bars. 1 pm.

 

number of cells with appendages were considerably less
than for the ancestor. Several lineages had cells that
were consistently longer than the ancestor. Cells from
population L18 are shown in Fig. 1(b). L18 is an example
not only of the lack of encapsulation but also of the
occasional presence of vestigial prosthecae. Light-bright

blebs were occasionally seen associated with the cell
surface or with a prostheca-like appendage, as can be
seen in this photomicrograph. The compiled data are
presented in Table 1.

Scanning electron micrographs of the ancestor and four
evolved populations are also shown in Fig. 1. he
ancestor (Fig. 1c and d) presented a complex contoured
surface with clear indications of heavy encapsulation as
well as prostheca-like appendages. The appendages were
not as long as those observed using light microscopy,
perhaps due to the more rigorous preparative steps
required by SEM. SEM micrographs of evolved popu-
lations L6, L12, L16 and L18 are shown in Fig. 1(e h).
These lineages do not display evidence of encapsulation.
Cells from the evolved populations L12 and L18 do not
Vary signi cantly in length from the ancestor (Table 1).

L16 was remarkable in that its mean cell length was 4'7
times the mean length of the ancestor and occasional
cells of 20 25 um length were observed. Frequently, one
terminus of the rod-shaped cell had greater light and
electron opacity than the other. L18 showed evidence of
vestigial prostheca-like appendages, some evidence of
capsule material, and the unusual light-bright blebs
mentioned above. L6 is one of three lineages that was
signi cantly shorter than the ancestor. Eight of the 12
lineages evolved in liquid and three of six evolved on
agar had signi cant increases in mean cell length,
ranging from 1'5 to 4'7 times that of the ancestor. No
signi cant differences between the original ancestor and
the streptomycin or nalidixic acid resistant mutants
were detected with microscopy. Table 1 gives the mean
cell lengths for all of the lineages based on SEM.

Transmission electron microscopy (TEM) was per-
formed to con rrn the differences in morphology be-
tween the ancestor and two lineages, and these images
are shown in Fig. 2. As viewed with TEM, L16 and L18
(Fig. 2c and d, respectively) had internal cellular details
similar to those seen in the ancestor (Fig. 2a and b),
including structures consistent with poly-fiehydroxy-
butyrate storage granules. The TEM images also con-

rm the integrity of the cytoplasm throughout the length
of the unusually long cells of L16.

BIOLOG

To monitor changes in the catabolic breadth of the
replicate lines, BIOLOG GN plates testing the ability to
metabolize 95 different carbon sources were used. The
salient features of the BIOLOG data (Fig. 3) are as
follows. First, the ancestor was capable of utilizing
43 49 of the 95 substrates in the BIOLOG GN plates in
a pattern consistent with the type strain description of
Ralstonia eutropba (formerly Alcaligenes eutropbus;
Krieg 6c Holt, 1984). Second, no two derived genotypes
had the same pattern of carbon utilization. Third, the
greatest range of utilization patterns appeared amongst
the populations evolved on agar. Fourth, while there
was tremendous phenotypic variation among the
evolved populations, there was one discernible pattern
detected among the lineages evolved in liquid as
described below.

The results of hierarchical cluster analysis on the
BIOLOG data using all 18 lineages evolved either in
liquid or on solid media are presented in Fig. 4. Note
that the antibiotic resistance marker determines the
primary grouping for the lineages evolved on liquid.
Lineages founded with the streptomycin-resistant an-
cestor (L1, L3, L5, L7, L9 and L11) group with the
original ancestor and the two antibiotic resistant
variants whereas those founded with the nalidixic acid
resistant strain (L2, L4, L6, L8, L10 and L12) group
separately. No such ancestor-dependent clustering is
seen among the lineages evolved on agar. This marker
effect was examined statistically by two t-tests: one
which divided all evolved populations based on their
antibiotic marker state, and another which only divided

 

189

M. S. RILEY and OTHERS

 

-lactic acid

-phenylalanine

-pmline

 

 

Fly. 3. For legend see facing page.

 

190

Rapid phenotypic evolution

 

ANC-Str"
L13

ANC
L14

L17
ANC-NalR
7

 

 

 

 

i I i I I
0.0 ol1 0-2 0-3 0-4 05 0'6
Distance

 

Fig. 4. Clustering based on catabolic profiles of the 18 evolved
lines, their two proximate ancestors that differ in antibiotic
resistance markers (ANC- Str“ and ANC—Nal"), and the original
Ra [stun/a strain TFDM (ANC). The tree was constructed aby
hierarchical cluster analysis based on the meth odon
(1963), adjusted for covariance, using SYSTAT v. 7. 01. See Tablaer 1
for strain identification. ANC indicates the original Ralstonia
strain TFD41, while ANC- Str" and ANC- Nal" denote antibiotic
resistance marker variants thereo .

 

 

 

~18
.. 15
3::
£10
3 8
2 6
a 4
2
°gegzsnsassssezeezeeee
(z‘ol JJ-JIJJ—l—ul—lJ
33' Liquid ' SolId '
<<
Lineage

 

Fig. 5. MIC of bile salts. Each strain was tested' In duplicate. The
MIC was defined as the concentration at which no detectable
increase in turbidity was 0 served after 5d at 30 °C. ANC
indicates the original Ralstonia strain TFD41, while ANC-Str"
an nd ANC-N “denote antibiotic resistance marker variants
thereof

 

liquid-evolved populations. We found that Nal“ popu-
lations had diminished breadth in carbon source
utilization compared to the Str“ populations, but the

 

 

 

 

 

a 1-0

3

..30-8

0:

8'50'6

h-..

EEO-4

h o-zi i i
II.

M i
3%23333335339251‘33‘329
(zwl _l_l_ll_l_l_J—J_J_JJ

'Ui . . I .
“2’2 LIqmd Solid
<<

Lineage

 

Fly. 6. Adhesion of ancestor and derived genotypes to a sand
substra et. Ce lls grown on L-aspa rtic acid were incubated In the
sand matrix for 1 h. Following elution, the fraction of cells no
attached to the sand was determined spectrophotometrically
and the fraction retained was calculated by subtr traction. All
assays were rformed In triplicate. Erro or barsin dicate standard
deviation ANC indicates the original Ralstonia strain TFD41
while ANC- Str" and ANC-Naln denote antibiotic resistance marker
variants thereof.

 

difference was only clearly signi cant among the liquid-
evolved populations (all evolved lines, 1 =1-91, d.f. =
16, P =0'074; liquid evolved only, I =3'59, d.f. =

P =0005 ; both two-tailed tests).

MIC of bile salts

ln Gram-negative bacteria, changes to the outer en-
velope can alter sensitivity to antibiotics and detergents.
To determine if any of the evolved populations changed
in this respect, the MIC of bile salts was determined (Fig.
5). The ancestor and its antibiotic-resistant mutants
were able to grow in media with bile salts concentrations
above 16g 1". In contrast, all 18 evolved populations
were substantially more sensitive, with growth inhi-
bition observed at concentrations be ow 6 g l'I in all
cases.

Adhesion

The ancestors and evolved populations displayed very
different adhesion properties (Fig. 6). All of the evolved
strains, with one exception (L5), had high affinity for the
sand matrix, which retained more than 50 °/o of the cells
applied to the column. The ancestral strains, on the
other hand, did not adhere strongly to the sand, with
about 80% of the cells loaded onto the column passing
through the sand matrix. L5, the only exception among
the evolved populations, behaved like the ancestor in
that 80 °/o of cells applied to the column were recovered
in the effluent.

 

Fly. 3. Carbon utilization patterns of the ancestor, Str" and Nal" mutants thereof, and the 18 evolved lines after 1000
generations. Black cells indicate the carbon source was consistently metabolized; grey cells indicate variability inth

catabolism of the particular carbon source across replicate BIOLOG plates. Carbon sources were sorted first by overall
mean (mean performance of all lineages on a particular substrate), in descending order, and the n by coefficient of
variation (the standard deviation divided by the overall mean), in descending order. The total number of carbon sources

used by each strain is showno

bottom row, includin ng both consistently positive and variable carbon sources. The
rirers

neth
lines that evolved In liquid media are grouped by their ancestral antibiotic resistance ma

 

191

M. S. RILEY and OTHERS

 

ANC-Nal"-c
L16-a
ANC-Nal'I-b
L16-b
ANC—a
ANC-c
ANC-b
L17-a

L17-c

L17-b
L18-a
L18-c
L18-b
ANC-StrR-a
ANC-Str"-b
ANC—Str’Lc
ANC-NaIR-a
L16-c

L1 1-b

L1 1-a

L1 1 -c

L6-a

L6-b

LS-c

LB-b

LB-a

1.8-c

L12-b

L9-b

L1 Z-a

L12-c

L9-c

LS-b

Ll-b

Ll-a

L1-c

L9-a

L54

L5~c

 

Solid
media

 

 

 

 

 

 

Liquid

 

 

wiwww

2-15 6* 10-77 150.

v v v r vvvvvvvvvvvv

Euclldian Distance

 

Fly. 7. Cluster analysis of FAME data from the ancestral strains
and ten evolved lineages. L1. L5, L6, L8. L9, L11 and L12 evolved
in liquid medium, whereas L16, L17 and L18 evolved on agar.
Each strain was tested in triplicate (denoted a, b and c). ANC
indicates the original Ralstonia strain TFD41, while ANC-Str"
and ANC—Nal' denote antibiotic resistance marker variants
thereof.

 

FAME analysis

FAME analysis of the ancestors and a subset of ten
evolved lineages consistently revealed nine peaks cor-
responding to fatty acids 14:0, 14:0 3-OH/16:1 iso 1,
16:1 w7c/15 iso 2-OH, 16:0, 17:0 cyclo, 16:0 2-OI-l,
18:1 w9c/a212t/w7c, 18:0 and 18:1 Z-OH. Note that in
this nomenclature a slash indicates that two or more
fatty acids co-elute under these conditions and they
are indistinguishable from one another. MIDI analysis
indicated that all lineages resemble Pseudomonas

pickettii or Burkholderia cepacia most closely among
those organisms in the database. A cluster analysis
of these data placed the ancestral strains and the
lineages evolved on solid medium together in one group
that was distinct from the lineages evolved in liquid (Fig.
7). The reproducibility of this major division is quite
high, as indicated by the frequent clustering of the three
replicates for each population. Only the Nal" parental
type showed variability in this regard. The distin-
guishing features in the FAME pro les that account for
the clustering were the relative areas of three peaks that
correspond to between three and six fatty acids. The
18 :1m7c/m9t/m2t peak was dominant in the liquid-
medium derived lineages, whereas 16 :1co7c/15 iso 2-
0H and 16:0 were the dominant peaks in the ancestor
and in those lineages evolved on agar.

DISCUSSION

After 1000 generations of evolution of Ralstonia strain
TFD41 in two different selective environments, we
observed several different patterns of phenotypic evol-
ution, depending on the particular trait: (1) changes
common to most or all derived lines, indicating par-
allelism; (2) changes that were speci c to the particular
selective environment; (3) changes that depended upon
which particular antibiotic-resistant variant of the
ancestor was used as the founder, indicating historical
contingency; and (4) changes that appeared more
stochastic and led to phenotypic diversity among the
evolved populations.

Morphological changes in the outer cell envelope
occurred in all 18 evolved lineages, including reductions
in the capsule (Table 1), increased sensitivity to bile salts
(Fig. 5), and with one exception, increased cell adhesion
to a sand matrix (Fig. 6). A genomic change also
occurred in all of the evolved lineages, including a
deletion of a genomic region detected by the loss of a
REP-PCR fragment (Nakatsu er al., 1998); it will be
interesting to determine whether these parallel genetic
and phenotypic changes are related to one another.
Changes that were speci c to the particular selective
environment were evident in the FAME pro les (Fig. 7),
as well as the observation that competitive tness was
more variable among the agar-evolved lineages than
among the liquid-evolved populations (Korona at al.,
1994). Evidence for historical contingency was found in
the carbon utilization patterns, which grouped the
liquid-evolved populations as a function of the
antibiotic-resistance marker borne by their immediate
ancestors (Fig. 4). Finally, substantial phenotypic di-
versity of cell morphology and carbon utilization
patterns was observed across all of the evolved popu-
lations, irrespective of differences in their selective
environments or parental genotype. For example, L16
and L18 were both evolved on agar, and both were
founded by the NatlR parental strain. L18 remained
similar in mean cell length (Table 1) and carbon
utilization pro le (Fig. 3) to the ancestor, whereas L16
has increased almost ve-fold in mean length and has a
substantially narrowed catabolic pro le.

 

192

Rapid phenotypic evolution

 

In broad outline, but not in detail, this mixture of
parallel and divergent patterns is similar to that seen in
a long-term evolution experiment with 12 replicate
populations of E. coli, each propagated in a de ned
glucose medium for thousands of generations (Lenski,
1995; Lenski et al., 1998). In that study, all the lines
improved in tness to a similar degree when in com-
petition against their ancestor in the glucose medium
(Lenski 8c Travisano, 1994). However when the
competitiveness of derived lines was tested against the
ancestor on different substrates, their performance was
much more variable, which implies a diversity of
underlying physiological changes (Travisano 8c Lenski,
1996). All 12 lines also increased their mean cell volume,
but the degree to which volume increased and the
resulting cell shape were variable (Lenski 8C Mongold,
2000). Substantial differences among the lines in genomic
mutation rates also arose, as several lines evolved defects
in methyl-directed mismatch repair (Sniegowski et al.,
1997) while other lines exhibited bursts of activity of
certain IS elements (Papadopoulos et al., 1999). How-
ever, unlike the present study with Ralstonia, in which
resistance mutations used as genetic markers in the
ancestor in uenced the path of subsequent evolution,
none of the differences among E. coli lines was strongly
affected by the arabinose-utilization marker that was
used to discriminate between those lines. Also, the main
long-term E. coli experiment used only a single selective
environment; however, when new lines derived from
these original populations evolved in different thermal
regimes, there emerged systematic differences among
them (Lenski, 1995; Mongold et al., 1996).

In considering the many changes to Ralstonia that have
accrued over 1000 generations, it is worthwhile to
consider in some detail the observed alterations to the
outer envelope. The outer envelope is the rst part of the
cell to interact with the external environment, serving
both as a barrier to unwanted factors (viruses, anti-
biotics, etc.) and as a conduit for essential resources.
Components of the outer envelope may therefore be the
most sensitive and responsive cellular constituents to the
selective environment. As Parke et al. (2000) put it: The
cell surface is where the rubber hits the road in bacterial
evolution. The components of the outer envelope
in uence shape, chemical resistance, adhesion and
metabolic properties of cells, and indeed all these
properties were substantially altered during 1000
generations of evolution in Ralstonia strain TFD41.
However, the underlying genetic and physiological bases
of these changes are not yet sufficiently understood to
interpret unambiguously the selective advantages con-
ferred by these traits, either in the laboratory experiment
or in nature. In the following paragraphs, we discuss
some possible adaptive scenarios.

Morphological changes

Both light and electron microscopy clearly show that all
the evolved populations have changed from the ancestral
morphology in terms of the cell-surface architecture

and, in many cases, cell length (Table 1). Most popu-
lations have lost all prostheca-like appendages and all
experienced either partial or complete loss of their
capsule. Variation in capsule production in bacteria is
not uncommon (Roberts, 1996; see also Fletcher,
1996; Whit eld 8c Valvano, 1993) and has been cited as
an induced response that may confer an advantage in
appropriate environments. In E. coli, for example, the
expression of capsule material is increased when cells
are subjected to an arid environment, thereby providing
protection against desiccation (Ophir 8C Glutnick,
1994). The cell surface also determines adhesive proper-
ties of the cell. A capsule can promote or prevent
adhesion depending upon the relative chemistries of the
capsule and substratum. Various extracellular polymers,
including polysaccharides, that promote adhesion have
been identi ed (Roberts,1996). Onestudy identi ed two
types of extracellular polymers in the same organism:
one that promotes attachment and the other that
promotes detachment (Wrangstadh et al., 1990). Non-
adhesive variants of Pseudomonas uorescens that
possess a capsule have also been described (Williams 8C
Fletcher, 1996). Simoni et al. (1998) speculate that
heterogeneity within bacterial populations expressed at
the level of the outer envelope can in uence their
transport in groundwater aquifers. Among the evolved
populations of Ralstonia in this study all but one showed
a substantial increase in adhesion to a sand matrix in
comparison to the ancestor. Inasmuch as all of the
evolved populations show reduced encapsulation, the
capsule of the ancestral soil-dwelling Ralstonia strain
must have served some function other than providing a
matrix for adhesion to silicates. Another possible bene t
for capsule-bearing strains is resistance to anti-bacterial
compounds (Nikaido, 1996). Consistent with this possi-
bility is the observation that, in a selective environment
lacking such compounds, all 18 populations evolved
increased sensitivity to bile salts (Fig. 5). Hence, in
Ralstonia, the primary bene t of the capsule in the soil
environment may be protection against chemicals and
desiccation rather than adhesion to soil particles.

Perhaps the most striking phenotypic changes were in
the mean cell length: 14 of 18 evolved lines showed
signi cant changes in cell length. Of the 12 liquid-
derived populations, eight are signi cantly longer than
the ancestor and two are shorter. The six lines evolved
on agar also include both shorter (one of six) and longer
(three of six) morphologies, including lineage L16 that
increased its mean length almost vefold. The unusually
long cells of this population were not caused by
incomplete septation, as TEM clearly showed a normal
cytoplasm (Fig. 2). Changes in cell length have been
correlated with survival strategies including, for
example, the avoidance of predation in the case of long

laments (Hahn et al., 1999; jiirgens et al., 1999), but it
is unclear what factors promoted such conspicuous
changes in this study. The fact that some lineages
evolved greater cell length and others became smaller,
while all showed substantial improvements in com-
petitive tness (Korona et al., 1994), suggests that cell

 

193

M. S. RILEY and OTHERS

length was not itself a direct target of selection in these
experiments.

Biochemical changes

Mechanisms for recognition and transport of nutrients
also have essential components in the cell envel0pe.
Mutations that alter either speci c transport complexes
or their supporting structural matrix could change
carbon utilization patterns. For example, in E. coli at
least 90 genes are required for biosynthesis of the outer
membrane and capsule which in turn provides structural
scaffolding for the products of at least 37 genes encoding
integral outer-membrane proteins (see for example the
E. coli genome database at http ://mbgd.genome.ad.jp).
If the absence of capsule is a selected phenotype, as
suggested by the loss of capsule in all lineages, then any
mutation in the biosynthetic pathways of outer envelope
synthesis might contribute the loss of capsule and the
resulting increase in tness. Given the structural and
biochemical complexity of the outer envelope, a number
of possible genotypic pathways leading to loss of the
capsule exist. The complex and variable carbon
utilization pro les of the evolved populations are
consistent with this scenario. Several populations have
much narrower pro les than their ancestors, including
all the lines founded from the N a1“ ancestor and evolved
in liquid as well as two lines evolved on agar surfaces
(L15 and L16), whereas another surface-evolved line
(L14) has a much broader pro le than its ancestor

(Fig. 3).

The capsule has been viewed as a molecular sieve that
can in uence the accessibility of a substrate to the cell
(Nikaido, 1996). Given the diverse and complex changes
in morphology and surface architecture (Fig. 1 and
Table 1), it is easy to imagine that carbon utilization
patterns would be somehow affected, but difficult to
discern any simple association between these classes of
data. We did detect the acquisition of catabolic activity
(Fig. 3). In particular the ability to metabolize Tween 40
was detected in several lineages. It seems likely that the
loss of the copious capsule produced by the ancestor
permitted increased access of some carbon sources to
existing metabolic capabilities of the cell (Nikaido,
1996). Indeed, the loss of capsule may facilitate the
assimilation of the substrate 2,4-dichlorophenoxy-
acetate on which the replicate lines were evolved.

There is, however, one striking pattern in the carbon
utilization pro les. In particular, a signi cant clustering
was detected amongst the 12 liquid-evolved populations
(Fig. 4), in which the two dominant clusters derive from
the two different antibiotic-resistant ancestors that were
each used to found half of the evolving lines (Korona
et al., 1994). In particular, all of the liquid-evolved
populations that were founded with the Nal“ ancestor
had much narrower catabolic pro les than did those
founded by the Strn ancestor (Fig. 3). It is important to
emphasize that this narrower pro le was not directly
attributable to the mutation that conferred Nal“,

because the Nal" ancestor had a catabolic breadth
comparable to the common ancestor. Moreover, this
association with the founding ancestor was not seen
with the agar-evolved lines. Interestingly, all six of the
lines that were initially Nal“ and evolved in liquid
reverted to sensitivity to nalidixic acid during the
evolution experiment, whereas all three surface-evolved
lines retained the Nal“ phenotype (Korona et al., 1994).
These data collectively indicate that the initial resistance
marker in uenced subsequent evolution in the liquid
environment, although the reasons for this remain
speculative. In E. coli, I‘vlalfl phenotypes can be produced
by mutations in the gyrB locus that encodes DNA gyrase
(Yamagishi et al., 1986) as well as by mutations in other
loci that cause reduced permeability of the outer
membrane (Hrebenda et al., 1985). If the nalidixic acid
resistance of the ancestral strain was expressed at the
level of the cell envelope, then a subsequent evolutionary
change in this structure (of the sort observed in the
evolved lines) could negate the phenotypic expression of
the resistance mutation.

FAME analysis indicated that the liquid-evolved lineages
changed from the ancestral composition, whereas the
surface-evolved lines have not (Fig. 7). In particular, all
six of the liquid-evolved lines that were tested had
changes in the relative amounts of three elution peaks,
corresponding to three to six co-eluting fatty acids,
compared to the ancestors and the three tested lineages
that had evolved on solid media. While these patterns
are consistent,"it is not obvious how these alterations
relate functionally to the two different selective regimes,
except perhaps to note that the greater deviation of the
liquid-evolved populations from the ancestral state
suggests that this regime was more different from the
natural environment than was the surface regime. It is
also interesting to point out that a Euclidean distance of
10 in the FAME analysis has been proposed as an
indication of species-level differences (MIDI). By this
standard, the liquid-evolved lineages have evolved into a
new species in only 1000 generations. This interpretation
is unreasonable, and so one might be tempted to suggest
that the experimental populations became contaminated
by some other species, but this possibility is rejected by
molecular genetic analyses that show all the evolved
lines, from both liquid and surface regimes, to be
derived from ancestral TFD41 (Nakatsu et al., 1998). In
our view, the idea that such differences in FAME
composition re ect species-level differences is evidently
awed, at least as a general proposition.

Concluding remarks

Considering all of the phenotypic traits together, we nd
remarkable both the extent of change from the ancestral
condition and the diversity among the derived lines. It is
tempting to suggest that all this diversity represents
different adaptive solutions achieved by each line in
response to the selective environment. An alternative
explanation is that a substantial fraction of the pheno-
typic diversity was the result of non-adaptive genotypic

 

194

I?

Rapid phenotypic evolution

 

changes that spread by hitchhiking (linkage) with
bene cial mutations. To the extent that the selective
environment was stressful and thereby mutagenic to the
bacteria, this explanation becomes more likely, as the
opportunity for neutral and even deleterious mutations
to hitchhike becomes greater the higher the genomic
mutation rate. It has been shown that some Ralstonia
strains undergo substantial stress-induced changes in
their genomes, including deletions and rearrangements
(Taghavi et al., 1997), and similar kinds of events were
seen in this evolution experiment with strain TFD41
(N akatsu et al., 1998). Indeed, recent evidence indicates
that Ralstonia strain TFD41 undergoes deletions similar
to those found in the evolved lines when subjected to
certain stresses, and these deletions occur at high
frequency (T. L. Marsh, unpublished data), so that they
could spread by hitchhiking. We do not know whether
batch culture (including lag and stationary phases, as
well as growth) on 2,4-dichlorophenoxyacetate repre-
sented a comparable stress, but it is certainly possible
that stress-induced genotypic changes contributed to the
phenotypic diversity observed (Finkel 8c Kolter, 1999).
We are currently assessing the extent to which stress-
induced changes in the TFD41 genome may have caused
changes in cell morphology and other traits during the
evolution experiment as well as determining the mol-
ecular bases of the increase in relative tness exhibited
by the derived lineages at 1000 generations.

ACKNOWLEDGEMENTS

The authors thank the Center for Electron Optics at Michigan
State University for assistance with the electron microscopy.
Special thanks to Carol Flegler, Uwe Danielowitz and Sally
Burns. We are grateful to the Kellogg Biological Station for
assisting with the FAME analysis and for the excellent
technical assistance of Tracy Simpson and Elizabeth
Seymoure. This work was supported by the Center for
Microbial Ecology through National Science Foundation
grant DEB-9120006.

REFERENCES

Defiaun, M. P., Tanzer, A. S., McAteer, A. L, Marshall, B. a Levy,
S. B. (1990). Development of an adhesion assay and charac-
terization of an adhesion-dc cient mutant of Pseudomonas
uarescens. Appl Environ Microbial 56, 112 119.

Doetsch, R. N. (1981). Determinative methods of light microscopy.
In Manual of Methods for General Bacteriology, pp. 21 33.
Edited by P. Gerhardt. Washington, DC: American Society for
Microbiology.

Dykhulzen, D. E. (1990). Experimental studies of natural selection
in bacteria. Annu Rev Ecol Syst 21, 373 398.

Dykhulzen, D. E. (1993). Chemostats used for studying natural
selection and adaptive evolution. Methods Enzymal 224, 613 631.

Flnltel, S. E. I: Kolter, R. (1999). Evolution of microbial diversity
during prolonged starvation. Proc Natl Acad Sci USA 96,
4023 4027.

Fletcher, M. (edltor) (1996). Bacterial Adhesion: Molecular and
Ecological Diversity. New York: Wiley-Lisa. ,

Gucltert. J. B., Carr. 6. J., Johnson, T. D., Hamm. B. 6., Davidson,
0. H. 8: Kamagal. Y. (1996). Community analysis by Biolog: curve

integration for statistical analysis of activated sludge microbial
habitats. ] Microbial Methods 27, 183 197.

Hahn, M. W., Moore, E. R. B. 8: Hbfle, M. G. (1999). Bacterial

lament formation, a defense mechanism against agellate
grazing: is growrh rate controlled in bacteria of different phyla?
Appl Environ Microbial 65, 25 35.

Helllng, R. B., Vargas, C. N. 8: Adams, J. (1987). Evolution of
Escherichia coli during growth in a constant environment.
Genetics I16, 349 358.

Hrebenda. J., Heleszlto, N., Brzostelt. K. I: Blelecltl, J. (1985).
Mutation affecting resistance of Escherichia coli K12 to nalidixic
acid. ] Gen Microbiol 131, 2285 2292.

Jilrgens. K., Pernthaler. J., Schalla, S. a Amann, R. (1999).
Morphological and compositional changes in a planktonic
bacterial community in response to enhanced protozoan grazing.
Appl Environ Microbiol 65, 1241 1250.

Karena, R. (1996). Adaptation to structurally different environ-
ments. Prac Roy Soc Lond B 263, 1665 1669.

Karona. 11., Nakatsu, C. N., Forney. L. J. a. Lenski, R. E. (1994).
Evidence for multiple adaptive peaks from populations of bacteria
evolving in a structured habitat. Prac Natl Acad Sci USA 91,
9037 9041.

Krleg. N. R. I: Holt, 1. G. (editors) (1984). Bergey 5 Manual of
Systematic Bacteriology, vol. 1. Baltimore: Williams 8c Wilkins.

lenski, R. E. (1995). Evolution in experimental populations of
bacteria. In Population Genetics of Bacteria, pp. 193 215. Edited
by S. Baumberg and others. Cambridge: Cambridge University
Press.

Lenskl, R. E. I: Marigold, J. A. (2000). Cell size, shape and tness
in evolving populations of bacteria. In Scaling in Biology, pp.
221 235. Edited by J. H. Brown 8c G. B. West. Oxford: Oxford
University Press.

Lenski, R. E. In Travlsano, M. (1994). Dynamics of adaptation and
diversi cation: a 10,000-generation experiment with bacterial
populations. Prac Natl Acad Sci USA 91, 6808 6814.

Lenski, R. E., Mongold. J. A., Snlegowsltl, P. D., Travlsano, M.,
Vasl, F.. Gerrlsh. P. .l. & Schmidt. T. M. (1998). Evolution of
competitive tness in experimental populations of E. coli: what
makes one genotype a better competitor than another? Antonie
Leeuwenhoek 73, 35 47.

Marigold. J. A., Bennett. A. F. a Lenski, R. E. (1996). Experimental
investigations of evolutionary adaptation to temperature. In
Animals and Temperature: Phenotypic and Evolutionary Ad-
aptation, pp. 239 264. Edited by I. A. johnston 8c A. F. Bennett.
Cambridge: Cambridge University Press.

Naltatsu, C. H., Karona. R., Lenski, R. E., De Brulln, F. J., Marsh,
T. L 8: Forney, L. .l. (1998). Parallel and divergent genotypic eval-
ution in experimental populations of Ralstonia sp. ] Bacterial
180, 4325 4331.

lealda, H. (1996). Outer membrane. In Escherichia coli and
Salmonella, pp. 29 47. Edited by F. C. Neidhardt. Washington,
DC: American Society for Microbiology.

Ophlr, T. In Glutnlck, D. (1994). A role for exopolysaccharides in
the protection of microorganisms from desiccation. Appl Environ
Microbial 60, 740 745.

Papadopoulos, D., Schneider, 0., Meler-Elss, J., Arbor, W., Lenski,
R. E. 8: Blot, M. (1999). Genomic evolution during a 10,000-
generatian experiment with bacteria. Proc Natl Acad Sci US A 96,
3807 3812.

Parke, D., D'Argenlo. D. A. a Omston, L N. (2000). Bacteria are
not what they eat: that is why they are so diverse. ] Bacterial I82,
257 263.

 

 

195

M. S. RILEY and OTHERS

 

Ralney, P. 8. 8i Travlsano. M. (1998). Adaptive radiation in a
heterogeneous environment. Nature 394, 69 72.

Roberts. l. S. (1996). The biochemistry and genetics of capsular
polysaccharide production in bacteria. Annu Rev Microbial 50,
285 315.

Sasser, M. (1997). MIDI technical note 101 (revised 1997).
Newark, DE: MIDI.

Simoni, S. E., Harms, H., Bosma, T. M. P. a Zehnder, A. l. B.
(1998). Population heterogeneity affects transport of bacteria
through sand columns at low ow rates. Environ Sci Technol 32,
2100 2105.

Snlegowsltl, P. D., Gerrlsh, P. J. 8i Lenski, R. E. (1997). Evolution
of high mutation rates in experimental populations of Escherichia
coli. Nature 387, 703 705.

Taghavl, S., Mergeay, M. a. van der Leila, D. (1997). Genetic and
physical maps of the Alcaligenes eutrophus CH34 megaplasmid
pMOL28 and its derivative pMOLSO obtained after temperature-
induced mutagenesis and mortality. Plasmid 37, 22 34.

Tonso, N. L. Matheson, V. G. a: Holben, W. E. (1995). Polyphasic
characterization of a suite of bacterial isolates capable of
degrading 2,4-D. Microb Ecol 30, 3 24.

Travisana, M. 8: Lenski, R. E. (1996). Long-term experimental
evolution in Escherichia coli. IV. Targets of selection and the
speci city of adaptation. Genetics 143, 15 26.

Ward, J. H. (1963). Hierarchical grouping to optimize an objecrive
function. ] Am Stat Assoc 58, 236 244.

Whitfield, C. & Valvano, M. A. (1993). Biosynthesis and exPression
of cell-surface polysaccharides in Gram negative bacteria. Adv
Microb Physiol 35, 136 246.

Williams, V. 8: Hatchet. M. (1996). Pseudomonas uorescens
adhesion and transport through porous media are affected by
lipapolysaccharide composition. Appl Environ Microbial 62,
100 104.

Wrangstadh, M., Szewzylt. U., Ostllng. J. 8: Kjelleberg, S. (1990).
Starvation-speci c formation of a peripheral exopalysaccharide
by a marine Pseudomonas sp. strain 59. Appl Environ Microbiol
56, 2065 2072.

Wright. S. (1932). The roles of mutation, inbreeding, crass-
breeding and selection in evolution. In Proceedings of the Sixth
international Congress afGenetics, pp. 356 366. Edited by D. F.
lanes. Menasha, WI: Brooklyn Botanic Garden.

Wright, S. (1988). Surfaces of selective value revisited. Am Nat
131,115 123.

Wyndham, R. C. (1986). Evolved aniline catabolism in Acineto-
bacter calcoaceticus during continuous culture of river water.
Appl Environ Microbiol 51, 781 789.

Yamaglshl. J.-1., Yoshida. N., Yamayoshi, M. & Nakamura, S.
(1986). Nalidixic acid-resistant mutations of the gyrB gene of
Escherichia coli. Mol Gen Genet 204, 367 373.

 

Received 20 June 2000; revised 20 November 2000; accepted 1 December
2000.

 

196

iiijiiijljiiiijiiii‘ijigijii