BACTERIA, MICROBIAL COMMUNITIES AND ENGINEERING: STUDIES ON THE
MICROBIAL ECOLOGY OF SELECTED ENGINEERED SYSTEMS
By
Fan Yang

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
Environmental Engineering – Doctor of Philosophy
2013

ABSTRACT
BACTERIA, MICROBIAL COMMUNITIES AND ENGINEERING: STUDIES ON THE
MICROBIAL ECOLOGY OF SELECTED ENGINEERED SYSTEMS
By
Fan Yang
Environmental contaminants, such as soluble metal ions and agricultural wastes pose great
risks for both human health and ecosystems. To reduce these risks, environmental engineers
have developed remediation approaches that take the advantage of microbial communities
and populations. Understanding these microbial resources is instrumental to manage and
apply them in various engineered systems. In this dissertation, I study microbial communities and populations from three diﬀerent approaches and demonstrate how basic microbial
information can assist us in optimizing engineered systems.
The ﬁrst part of my dissertation focuses on understanding the genomic advantages of
Ralstonia pickettii strains, which allows them to adapt to high copper environments. We
have previously shown that these two strains were able to sequester a large amount of copper.
Hence, these two bacterial strains have a great potential in for application to industrial
wastewater treatment. Understanding the genomic evolution and adaptation behind the
copper binding phenomenon could unveil the industrial potential of these bacterial strains.
The second part of this dissertation focuses on understanding the role of anaerobic bacterial populations and communities in uranium immobilization. A large amount of research
has been conducted on identifying the bacterial communities involved in in situ uranium
immobilization. However, the extant of soil microbial diversity made it diﬃcult to identify
the most important speciﬁc populations. We employed enrichment culture methods to in-

crease the abundance of potential important bacterial populations and to link the community
functions.
Finally, I present a study on microbial communities in methane producing agricultural
waste co-bioreactors. Methane production has been reported as a highly cooperative reaction
between bacteria and archaea. Linking bacterial populations to speciﬁc functions would help
optimize agricultural waste degradation as well as alternative energy production.
I chose these three topics to emphasize the importance of microbial populations in engineered systems. By understanding the roles of individual bacteria populations as well as their
interactions with each other in a community, I hope to manage and utilize these microbial
resources to improve our living environment.

Copyright by
FAN YANG
2013

This thesis is dedicated to my parents, Hanying Yang and Ning Shen.

v

ACKNOWLEDGEMENTS

This dissertation marks the end of my journey as a Ph.D. student. Throughout the past ﬁve
years, there was excitement, frustration, as well as a tremendous amount of support from
people around me. I would like to begin the acknowledgement by thanking Dr. Terence L.
Marsh, who graciously took me in as an undergraduate research assistant, from where I grew
into loving microbial ecology. From 2004 to 2012, I have learned the importance of bacteria
and their applicabilities in sustainably managing our environments. I thank Dr. Thomas
C. Voice, who took the responsibility of being my major advisor in the department of Civil
and Environmental Engineering. It was a drastic change switching from microbiology to
engineering. I am grateful for Dr. Voice’s guidance on how to think and solve problems
like an engineer, which sets me up for applied science. I would like to thank my research
committee members, Dr. Alison Cupples and Dr. David T. Long, who have always been
supportive and taught me many valuable lessons through the past ﬁve years.
My thank goes to countless supportive friends. I’m especially thankful for Dr. Adina
Howe and Mike Howe’s hospitality, great advice and the tremendous amount of support
for the past one year and a half, which was deﬁnitely special and eventful. I would like
to thank Dr. Juliana Bordowitz and Dr. Zarraz May-ping Lee for their great friendship,
without whom, the transition to the Ph.D. program would not be easy. My thank goes
to the entire Ribosomal Database Project members, who have been extremely helpful in
preparing me for the big bioinformatic world. I truly appreciate the great relationships I
have with the past and current Tiedje group members. My special thank goes to the entire

vi

ROME lab, especially, Chirstina Harzman, Natasha Issacs-Cosgrove, Joe Duris, and Lisa
Reynolds Fogarty, without whom, I would not think graduate school is fun. I also would like
to acknowledge all of the hard work Katherine Ivens, Rachael Rossi, and Brian Lovett have
done as undergraduate research assistants. My thank also goes to all of the secretaries in
both the Department of Microbiology and Molecular Genetics and the Department of Civil
and Environmental Engineering, especially Lori Larner, without whom, I would not be able
to make it through the ﬁrst year.
Finally, I would like to thank my parents, Hanying Yang and Ning Shen, for the wonderful
opportunity they gave me to study abroad at Michigan State University. It has been a long
journey and I am thankful that I was not doing this alone.

vii

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
CHAPTER 1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 Bioremediaton of Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.1 Biosequestration of Metals . . . . . . . . . . . . . . . . . . . . . . . .
1.1.1.1 Metal Contamination at Torch Lake . . . . . . . . . . . . .
1.1.1.2 Adaptation of Raltonisa pickettii in High Copper . . . . . .
1.1.2 Bioimmobilization of Metals . . . . . . . . . . . . . . . . . . . . . . .
1.1.2.1 Metal Contamination at Oak Ridge Reservation . . . . . . .
1.1.2.2 Bacterial Community Proﬁling . . . . . . . . . . . . . . . .
1.1.2.3 Reconstruct Uranium Immobilizing Bacterial Communities .
1.2 Bioremediation of Agriculture Wastes . . . . . . . . . . . . . . . . . . . . . .
1.2.1 The Role of Class Clostridia in Methane Producing Anaerobic Cobioreactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 Managing Complex Microbial Communities for Sustainable Environment . .

1
1
3
6
6
11
12
14
14
16
19
20

CHAPTER 2 CO-EVOLUTION OF RALSTONIA PICKETTII STRAIN
12D AND 12J: GENOME REARRANGEMENTS AND
THEIR INFLUENCES ON THE EVOLUTION OF METAL
RESISTANCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.1 Genome Retrieval and Gene Annotations . . . . . . . . . . . . . . . . 26
2.3.2 Genome Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.3 Identiﬁcation of the Metal Resistant Genes . . . . . . . . . . . . . . . 27
2.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4.1 Genome Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4.2 Co-evolution of R. pickettii 12D and 12J . . . . . . . . . . . . . . . . 39
2.4.2.1 16S rRNA Gene Sequence Variations . . . . . . . . . . . . . 39
2.4.2.2 dif Site Sequence Variations . . . . . . . . . . . . . . . . . . 40
2.4.2.3 Filamentous Phage Elements in R. pickettii 12D and 12J . . 42
2.4.3 Genome Syntenies of R. pickettii 12D and 12J . . . . . . . . . . . . . 43
2.4.4 The Acquisition and Duplication of Resistance Determinants in R.
pickettii 12D and 12J . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.4.4.1 Metal Resistant Genes in R. pickettii 12D and 12J . . . . . 49
2.4.4.2 The Phylogenetic Lineage of czcA Copies in R. pickettii
12D and 12J . . . . . . . . . . . . . . . . . . . . . . . . . . 53

viii

2.4.4.3

2.5
2.6

The Phylogenetic
12D and 12J . .
Discussion . . . . . . . . . . . . .
Conclusion . . . . . . . . . . . . .

Lineage
. . . . .
. . . . .
. . . . .

of copA
. . . . .
. . . . .
. . . . .

Copies
. . . .
. . . .
. . . .

in R. pickettii
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .

58
60
65

CHAPTER 3 RESOURCE-INDUCED BACTERIAL STRUCTURE SHIFTS
AND THEIR INFLUENCE ON COMMUNITY FUNCTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.3.1 Anaerobic Media and Stocks . . . . . . . . . . . . . . . . . . . . . . . 71
3.3.2 Bacterial Community Enrichments . . . . . . . . . . . . . . . . . . . 71
3.3.3 Quantitative Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.3.4 DNA Extraction and Community Proﬁling . . . . . . . . . . . . . . . 73
3.3.5 Community Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.4.1 Monitoring the Enrichment Cultures . . . . . . . . . . . . . . . . . . 75
3.4.2 Overview of the Community Dynamics . . . . . . . . . . . . . . . . . 76
3.4.3 Shifts in the Enrichment Bacterial Community Structures . . . . . . 77
3.4.4 Shared Genera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.4.5 Functional Gene Proﬁles . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.4.6 Non-metric multidimensional scaling (NMDS) analysis . . . . . . . . 86
3.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
CHAPTER 4 THE ROLE OF FIRMICUTES IN URANIUM IMMOBILIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.3.1 Uranium Immobilization by FW107 and FW102-2 Enrichment Cultures 98
4.3.2 Quantitative Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.3.3 DNA Extraction and Sequecing Preparation . . . . . . . . . . . . . . 99
4.3.4 Community Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.3.5 Isolation of Potential Uranium Immobilizing Bacteria . . . . . . . . . 101
4.3.6 Growth Curves of Isolates . . . . . . . . . . . . . . . . . . . . . . . . 101
4.3.7 Isolate Identiﬁcations . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.3.8 Cell Suspension Assay . . . . . . . . . . . . . . . . . . . . . . . . . . 103
4.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
4.4.1 Soluble U6+ Immobilization by FW107 and FW102-2 Enrichments . 103
4.4.2 Bacterial Community Compositions of FW107 and FW102-2 Enrichments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.4.3 Potential U6+ Immobilizers . . . . . . . . . . . . . . . . . . . . . . . 111
4.4.4 U6+ Immobilization By Isolates . . . . . . . . . . . . . . . . . . . . . 117

ix

4.5
4.6

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

CHAPTER 5 PHYLOGENETIC COMPOSITION OF BACTERIA IN
ANAEROBIC CO-DIGESTION OF DAIRY MANURE
AND CORN STOVER . . . . . . . . . . . . . . . . . . . . . .
5.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.1 Anaerobic Bioreactors . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.2 DNA Extraction and Targeted Gene Ampliﬁcations . . . . . . . . . .
5.3.3 Preparation for Pyrosequencing and Construction of Clone Libraries .
5.3.4 Sequence and Community Analyses . . . . . . . . . . . . . . . . . . .
5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.1 The Inﬂuence of HRT and Corn Stover on System Performance . . .
5.4.2 Phylogenetic Comparisons of Anaerobic Bioreactor Bacterial Communities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.3 Bacteroidetes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.4 Hydrogen Producing Population Proﬁles . . . . . . . . . . . . . . . .
5.4.5 Non-metric multidimensional scaling (NMDS) analysis . . . . . . . .
5.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

122
122
123
124
124
125
126
127
128
128
130
140
143
148
150
159

APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
APPENDIX A COMPARATIVE GENOMICS OF RALSTONIA PICKETTII 12D AND 12J . . . . . . . . . . . . . . . . . . . . .
A.1 Metabolic Pathways of R. pickettii 12D and 12J . . . . . . . . . . . . . . . .
A.2 Clustering of R. pickettii 12D and 12J and other 23 Proteobacteria . . . . .
A.3 Genetic Elements in R. pickettii 12D and 12J . . . . . . . . . . . . . . . . .
A.4 Variation of R. pickettii 12D and 12J dif Sites . . . . . . . . . . . . . . . . .
A.5 Genome Syntenies of R. pickettii 12D and 12J . . . . . . . . . . . . . . . . .

162
162
163
164
165
167

APPENDIX B ANAEROBIC ENRICHMENTS . . . . . . . . . . . . . . 169
B.1 Media Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
B.2 Functional Gene Proﬁles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
APPENDIX C URANIUM IMMOBILIZATION BY
C.1 Uranium-azide Assay . . . . . . . . . . . . . . . . . . .
C.2 5X Gitschier Buﬀer . . . . . . . . . . . . . . . . . . . .
C.3 Growth Curves of Isolates in Minimal Media . . . . . .

FIRMICUTES
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .

. .
. .
. .
. .

172
172
173
173

APPENDIX D BACTERIAL COMMUNITIES IN ANAEROBIC COBIOREACTORS . . . . . . . . . . . . . . . . . . . . . . . . 176
D.1 Deep Sequencing of Bacterial Communities in Anaerobic Co-bioreactors . . . 176
D.2 Anaerobic Co-bioreactor Bacterial Community Compositions at Phylum Level 178
x

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

xi

LIST OF TABLES

Table 2.1

Table 2.2

Table 3.1

Gene summaries of R. solanacearum GMI1000, C. metallidurans CH34,
R. pickettii 12D and 12J. . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

The abundance of Co-Zn-Cd, Cu, and Hg resistant genes in R. pickettii
12D and 12J, R. solanacearum GMI1000, and C. metallidurans CH34. . .

51

Ecological indices of iFRC enrichment bacterial communities. O represents original soil sediment samples. Fe, NO3, and U supplemented
enrichment cultures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

78

Table A.1 Side-by-side genome comparison of R. pickettii 12D and 12J genetic elements by genomic scaﬀolds. . . . . . . . . . . . . . . . . . . . . . . . . . 164
Table B.1 The composition of the modiﬁed FWM. . . . . . . . . . . . . . . . . . . . 169
Table B.2 The composition of the mineral mix (100X).

. . . . . . . . . . . . . . . . 169

Table B.3 The composition of the vitamin mix (1000X).

. . . . . . . . . . . . . . . 170

Table C.1 Ingredients for 200 ml of 5X Gitschier Buﬀer.

. . . . . . . . . . . . . . . 173

Table D.1 Measurements of sampling size and estimation of ecological indices of
bacterial community in six anaerobic bioreactors. Calculations were
based on the complete linkage clustering of chimera removed and trimmed
bacterial 16S rRNA gene sequence alignments at 97% similarity. N represents the total number of non-chimera sequences that passed quality
trimming. Shannon index (H’) and Simpson’s index (D) were expressed
in eH and 1-D, respectively to emphasize the diﬀerences. C represents
the sampling coverage via gene targeted pyrosequencing. . . . . . . . . . . 176

xii

LIST OF FIGURES

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

Figure 2.8

R. pickettii 12D genome scaﬀolds. The distribution of putative horizontally transferred genes (HGT genes) identiﬁed by JGI-IMG are marked
as black dashes in rings 4-10. The scaﬀold sizes are in units of Mb. For
interpretation of the references to color in this and all other ﬁgures, the
reader is referred to the electronic version of this dissertation. . . . . . .

31

R. pickettii 12J genome scaﬀolds. The distribution of putative horizontally transferred genes (HGT genes) identiﬁed by JGI-IMG are marked
as black dashes in rings 4-9. The scaﬀold sizes are in units of Mb. . . .

35

Minimal evolution phylogenetic tree of β-Proteobacteria dif sequences.
The numbers on branches are bootstrap values. The scale at the left
hand lower corner represents the distance. . . . . . . . . . . . . . . . . .

41

Filamentous phage p12J gene compositions and their homologs on R.
pickettii 12D and 12J chromosomes. Fragments of the same fragments
represent homologous genes. Genes colored yellow with a red border
represent genes that share no homology to p12J Orf7 but are orthologs
to each other. Grey colored fragments represent other phage related
genes (not p12J related). White colored blocks are other non-phage
related genes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

42

Genome syntenies between R. pickettii 12D and 12J scaﬀolds. Panel
A, 12D_C1, 12J_C1 and 12D_P2; Panel B, 12D_C2 and 12J_C2;
Panel C, 12D_P3 and 12J_P1. Arrows indicate location of ori C on
the chromosomes. Blocks with the same color represent the regions
that are similar (connected by lines). The amount of color indicates the
percentage of identity. White areas within the blocks indicate absence
of a homologous sequence. . . . . . . . . . . . . . . . . . . . . . . . . .

44

BLAST results mapped onto the 1.72 Mb-1.94 Mb region of R. pickettii
12J_C1. The top scale indicates the length of the query region. Similar
genes are shown underneath the query sequence. The short thin black
bars indicate that the next stretch of sequence is on the opposite strand
of DNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

47

Phylogenetic position of R. pickettii 12D and 12J czcA genes. The
numbers on the branches represent bootstrap values. . . . . . . . . . . .

54

Phylogenetic position of R. pickettii 12D and 12J copA genes. The
numbers on the branches represent bootstrap values. . . . . . . . . . . .

59

xiii

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

Figure 3.8

The concentration of Fe2+ (panels A and B), lactate and acetate (panels
C and D) in each ferric citrate enrichment cultures. The ﬁrst generation
enrichments are presented in panel A and C and the second generation
enrichments are presented in panel B and D. Error bars represent the
diﬀerences between two replicates. . . . . . . . . . . . . . . . . . . . . .

79

The top ten (summed across all samples) most abundant genera identiﬁed via RDP multiclassiﬁer. Sample replicates were labeled as "_1"
and "_2". Original soil sediment samples were denoted with "_O." . .

80

The shared genera among FW107 samples. The color of sample font is
the same as the cognate ellipse. . . . . . . . . . . . . . . . . . . . . . . .

81

The shared genera among FW102-2 samples. The color of sample font
is the same as the cognate ellipse. . . . . . . . . . . . . . . . . . . . . .

82

The shared genera among FW102-3 samples. The color of sample font
is the same as the cognate ellipse. . . . . . . . . . . . . . . . . . . . . .

83

The functional gene proﬁles of the original soil sediment and enrichment
samples. The functional gene signals of all samples (columns) were
standardized to the same amount. Each functional gene (row) was then
scaled among all 21 samples to show the diﬀerences in abundance. Green
indicates high abundance and red indicates low abundance. . . . . . . .

85

NMDS analysis of NO− enrichments (grey), Fe3+ enrichments (pink),
3
U6+ enrichments (yellow), and original samples (green). Each ellipse
was drawn around each group centroid based on a group distance standard error at 95% conﬁdence level. The changes in genera abundances
that drive comparative community diﬀerences are indicated with red
arrows. The arrows point to the areas of high abundance. The red
arrows indicate the abundance of genera that changed signiﬁcantly (P
≤ 0.001). The speciﬁc genera are shown in red according to the group
labeling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

88

NMDS analysis of NO− enrichments (grey), Fe3+ enrichments (pink),
3
U6+ enrichments (yellow), and original samples (green). Each ellipse
was drawn around each group centroid based on group distance standard
error at 95% conﬁdence level. The changes in functional gene (with the
speciﬁc microorganism from which the probes were designed) abundance
were correlated to the shifts observed in community structures. The
red arrows indicate the abundance of functional genes that changed
signiﬁcantly (P ≤ 0.001). The speciﬁc genera are shown in red according
to the group labeling. The arrows point to the areas of high abundance.

89

xiv

NMDS analysis of NO− enrichments (grey), Fe3+ enrichments (pink),
3
and U6+ enrichments (yellow). Each ellipse was drawn around each
group centroid based on group distance standard error at 95% conﬁdence
level. The changes in genera abundance were correlated to the shifts
observed in community structures. The colored arrows indicate the
abundance of genera that changed signiﬁcantly (red: P ≤ 0.001, blue:
P ≤ 0.05, and green: P ≤ 0.01). The genera names are shown in the
same color according to the group labeling. The arrows point to the
areas of high abundance. . . . . . . . . . . . . . . . . . . . . . . . . . .

90

Figure 3.10 NMDS analysis of NO− enrichments (grey), Fe3+ enrichments (pink),
3
and U6+ enrichments (yellow). Each ellipse was drawn around each
group centroid based on group distance standard error at 95% conﬁdence level. The changes in functional gene abundance were correlated
to the shifts observed in community structures. The red arrows indicate the abundance of functional genes that changed signiﬁcantly (P
≤ 0.001). The functional genes and the speciﬁc microorganisms from
which the probes were designed are shown in the red according to the
group labeling on the side. The arrows point to the areas of high abundance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

91

Figure 3.9

Figure 4.1

The removal of soluble U6+ in FW107 and FW102-2 enrichments. The
high soluble U(VI) concentration in the control culture (No Cells) was
due to the excess amount of uranyl acetate carried over from the syringe
needle during inoculation. . . . . . . . . . . . . . . . . . . . . . . . . . . 106

Figure 4.2

Bacterial community shifts with the increasing of incubation time in
FW107 U-PBAGW microcosm. . . . . . . . . . . . . . . . . . . . . . . . 107

Figure 4.3

Bacterial community shifts with the increase of incubation time in FW1022 U-PBAGW microcosm. . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Figure 4.4

Distribution of [FeFe]-hydrogenase containing bacteria in FW107 and
FW102-2 grown in 1/2X R2B media (U 0 days) and in uranyl acetate
enrichments (U 64 days). The bacterial populations represented by the
small fractions of each pie chart are listed on the right hand side with
color blocks corresponding to the colors found in each pie chart. The
distribution of genera Clostridium (cyan color) and Desulfotomaculum
(purple color) are shown as the outer ring of each pie chart. . . . . . . . 109

Figure 4.5

The phylogenetic distribution of the anaerobic isolates based on their
16S rRNA gene sequences. The numbers on the branches represent the
bootstrap values (1000 permutations). . . . . . . . . . . . . . . . . . . . 112

Figure 4.6

The growth curves of FRC isolates in 1/2X R2B under anaerobic condition (H2 :CO2 :N2 3:10:87). See next page for legends. . . . . . . . . . 114
xv

Figure 4.7

REP-BOX-(GTG)5 -PCR proﬁles of potential uranium immobilizing bacteria. Branches labeled Dhaf represent control samples. The rest represents enrichment isolates. . . . . . . . . . . . . . . . . . . . . . . . . . . 116

Figure 4.8

Soluble U6+ removal by Firmicutes isolates (from FW107) and FW107
uranium enriched community after 70 hours (T70) of incubation at room
temperature in an anaerobic chamber (H2 :CO2 :N2 3:10:87). Sample
"neg" represents a control culture without the addition of any bacterial
cells. The error bars represent measurements of two replicates. . . . . . 118

Figure 5.1

The total solids (TS) reduction eﬃciency (diagonal and grid lined columns)
and the production rate of biogas (grey and white columns) in the DM
and DM+CS bioreactors at four HRTs: a HRT of 30 days (HRT30d),
a HRT of 40 days (HRT40d) and a HRT of 50 days (HRT50d). Solid
diamonds and solid triangles represent the amount of methane produced
in the DM and DM+CS bioreactors, respectively. . . . . . . . . . . . . . 129

Figure 5.2

The abundance and diversity (OTU at 97% sequence similarity) of phyla
Bacteroidetes (panel A1), class Clostridia (panel B1), and Escherichia /
Shigella (panel C1) in each bioreactor bacterial community. The distribution of OTUs among samples are presented as Venn diagrams in panel
A2 (Bacteroidetes), B2 (Clostridia) and C2 (Escherichia/Shigella). The
colors of the bioreactor/HRT font corresponds to an area bounded by
the same color. Labels HRT30d, HRT40d and HRT50d represent bioreactors with a HRT of 30, 40, and 50 days, respectively. . . . . . . . . . . 131

Figure 5.3

The abundance of phyla Bacteroidetes (vertical stripes) in each bioreactor bacterial communities. The distribution of classiﬁed and unclassiﬁed Bacteroidetes within phylum Bacteroidetes were calculated and
displayed in left panel (white and diagonal striped bars) and right panel
(black and gridded bars) respectively. Labels HRT30d, HRT40d and
HRT50d represent bioreactors with a HRT of 30, 40, and 50 days, respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

Figure 5.4

The comparisons of Bacteroidetes clusters (97% sequence similarity)
between the DM+CS bioreactor with a HRT of 40 days and other
bioreactor samples. Labels HRT30d, HRT40d and HRT50d represent
bioreactors with a HRT of 30, 40, and 50 days, respectively. Symbol
∗ denote peaks that are clusters of unclassiﬁed Bacteroidetes that are
highly abundant in all DM+CS bioreators (#11 and #14), in DM+CS
bioreactors at long HRTs (#43), and in the DM+CS bioreactor with a
HRT of 40 days (#69). . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

xvi

Figure 5.5

The shifts of abundance of the three most abundant [FeFe]-hydrogenases
containing bacterial groups in anaerobic bioreactors fed with DM and
DM+CS. Labels HRT30d, HRT40d and HRT50d represent bioreactors
with a HRT of 30, 40, and 50 days, respectively. . . . . . . . . . . . . . . 146

Figure 5.6

The abundance of [FeFe]-hydrogenase containing bacteria in anaerobic
bioreactor samples compared to the abundance pattern of unclassiﬁed
Bacteroidetes (background) determined by 16S rRNA gene sequencing.
The distribution of clusters #11, #14, #43 and #69 were also plotted as
part of unclassiﬁed Bacteroidetes within each sample. Labels HRT30d,
HRT40d and HRT50d represent bioreactors with a HRT of 30, 40, and
50 days, respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Figure 5.7

NMDS analysis of the bacterial community in anaerobic bioreactors
(based on 16S rRNA gene sequences at 97% similarity cutoﬀ). The
green arrows indicate the increase of biogas production (Biogas), HRT
(HRT) and dairy manure degradation rate (Degradation). The red and
blue arrows indicate bacterial genera that shifted signiﬁcantly, where
P≤0.001 and P≤0.01, respectively. The black, yellow and purple arrows
represent the [FeFe]-hydrogenase phylogenetic aﬃliations that shifted
signiﬁcantly (P≤0.001, P≤0.05, and P≤0.01, respectively). All arrows
point towards the abundant range of the gradients. Labels HRT30d,
HRT40d and HRT50d represent bioreactors with a HRT of 30, 40, and
50 days, respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

Figure 5.8

Diagram (adapted from Schink et al. [140]) of microbial corporation
during methane generation in DM and DM+CS bioreactors operated at
HRT 40 days. The thickness of the arrow represent the relative abundance of microbial populations possibly associated with the methanogensis step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

Figure A.1 Distance tree of R. pickettii 12D and 12J in relationship to 23 other
Proteobacteria based on their 16S rRNA gene sequences. The bar in
the bottom left corner represents the distance scale. . . . . . . . . . . . . 163
Figure A.2 The sequence variations of dif sites of R. pickettii 12D and 12J. The
consensus reveals the conserved the regions. . . . . . . . . . . . . . . . . 165
Figure A.3 Genome syntenies of R. pickettii 12D and 12J revealed by a DNA sequence dotplot. Blue dots represent matching genes on the same direction strands. Red dots represent matching genes on the opposite
direction strands. The scaﬀold nucleotide coordinates are ranked from
left to right on the x-axis and bottom to top on the y-axis. . . . . . . . 168

xvii

Figure B.1 Functional gene proﬁles of original soil communities and the enriched
communities (replicate 1). Symbol "O" indicates original communities
and "Fe", "NO3" and "U" indicate supplemented electron acceptors in
enrichments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Figure C.1 Standard curve of the uranium-azide assay. The error bars represent
data from two replicates. . . . . . . . . . . . . . . . . . . . . . . . . . . 172
Figure C.2 The growth curves of FRC isolates (ﬁrst replicate) in minimal media
supplemented with sodium pyruvate alone (blue) and minimal media
supplemented with sodium pyruvate and uranyl acetate (red) under
anaerobic condition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
Figure C.3 The growth curves of FRC isolates (second replicate) in minimal media
supplemented with sodium pyruvate alone (blue) and minimal media
supplemented with sodium pyruvate and uranyl acetate (red) under
anaerobic condition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Figure D.1 Estimation of community coverage by rarefaction analysis. Rarefaction curves of anaerobic bioreactor 16S rRNA gene sequences were constructed based on the estimation of OTUs and ChaoI indices at 97%
similarity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Figure D.2 The distribution of the top three most abundant phyla (wide columns)
in bioreactor samples. The narrow columns represent the abundance
of genera Escherichia/Shigella (checker box), group unclassiﬁed Bacteroidetes (diagonal stripes) and class Clostridia (blacks) in each community. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

xviii

CHAPTER 1
INTRODUCTION

Environmental engineering technologies frequently take advantage of microbial communities
and populations. My dissertation is focused on the investigations of microbial communities
involved with bioremediation of metal contaminated sites and the anaerobic digestion of farm
waste. In both cases, increasing our knowledge base regarding the microbial communities
that are instrumental in each of these activities is critical to eﬃcient remediation and longterm management. In this background section I will ﬁrst discuss bioremediation of metals
followed by anaerobic digestion of farm wastes.

1.1

Bioremediaton of Metals

Metals are naturally occurring elements that are important components of the earth’s crust
and have played signiﬁcant roles in the development of human civilizations. The discovery
and utilization of metals signify the development of diﬀerent culture eras associated with the
history of human civilizations, the Bronze Age, the Iron Age, the Steel Age, and ﬁnally the
Nuclear Age. Each era is represented by the mining and smelting of a particular metal ore,
e.g., Cu, Fe, iron alloys, and U. As the primary method to expose and collect metals in the
earth crust, mining became one of the most important industries in human history [55].
Although the development of mining was instrumental in the evolution of human civilizations, mining does massive damage to the environment. Commonly known issues such
as land erosion and formation of sinkholes may accompany mining processes. The toxic

1

chemicals (e.g., sulfuric acid, cyanide, alkyl xanthate salt) used during mining to enrich speciﬁc metals (e.g., Cu, Au) can contaminate surrounding soil, surface and subsurface water
bodies. The metal-containing tailing solutions along with naturally occurring oxidation on
exposed metals, can mobilize metals from the earth’s crust and cause further contamination of the surrounding environment via river and aquifers. Although the toxicity of metals
varies, ingestion of soluble metals, especially heavy metals, can be life threatening. Hence,
the immense impact of mining on biodiversity and ecosystems cannot be ignored.
The mobility and toxicity of metals in the environment are controlled by local acidbase and redox reactions. Shifts in pH can largely inﬂuence metal mobilities. Take the
most abundant transition metal, iron, as an example. Oxidized iron (Fe3+ ) precipitates as
Fe(OH)3 or other Fe3+ minerals at near-neutral pH but it becomes more soluble when pH
decreases. Therefore, Fe3+ is mobile in acidic environments. Environmental redox reactions
control the formation and dissolution of minerals. For example, iron can be oxidized to Fe3+
by exposure to oxygen. In the presence of sulﬁde or reduced humics, Fe3+ can be readily
reduced to Fe2+ . At neutral pH, Fe2+ is much more soluble than Fe3+ . Hence, in the form
of Fe2+ , iron can be readily absorbed by living organisms. For this reason, Fe2+ is more
toxic than Fe3+ at near-neutral pH [8].
Bacteria contribute substantially to metal mobility and toxicity. The growth of bacteria can change local pH. Through metabolic processes, such as denitriﬁcation, bacteria
can increase the pH of local environments. When this pH shift occurs in an acidic habitat,
the precipitation of metal ions would occur due to the increase of pH [187]. Also through
metabolic processes, such as oxidation of pyrite by Acidithiobacillus species, bacteria can decrease the pH of the local environment. This decline of pH may result in metal leaching from
2

rocks increasing the mobility and toxicity of heavy metals [54]. Bacteria can also physically
reduce the concentration of mobile metal ions via biosequestration. Due to their large surface to volume ratios and the composition of the cell envelopes, bacteria (alive or dead) can
naturally absorb mobile metal ions [178]. Studies have also shown that some bacteria can sequester a large amount of metals in the outer envelope of the cell and thus remove metal ions
from aqueous solution [90, 136, 191]. Finally, bacteria play an essential role in environmental redox processes. Some bacteria can derive energy by transferring electrons from organic
carbon or inorganic compounds to electron acceptors, including transition metals, metalloids and actinides. Studies have demonstrated that anaerobic bacteria, such as Geobacter
metallireducens, Shewanella oneidensis, and Desulfuromonas acetoxidans, can dissimilatorily
reduce metal ions (e.g., from Fe3+ to Fe2+ ) under anoxic conditions [65]. These indirect and
direct interactions between bacteria and metals have been observed across a wide spectrum
of heavy metals, including Mn, Cu, Cr, Co, Hg, and U. The unique attributes of bacterial
metal sequestration and immobilization have been cultivated by environmental engineers for
remediation of metal contamination.

1.1.1

Biosequestration of Metals

Biosequestration of metals herein refers to metal adsorption and accumulation by either live
or treated bacteria. Biosorption usually refers to the passive binding of metals to biological
materials. The sorptivity of metals by the cell envelope is independent of active metabolism
but directly related to sorbent surface area. The composition of the bacterial cell envelop
also contributes greatly to metal speciﬁcity and binding aﬃnity [139]. Due to their small
sizes, bacteria naturally have high surface to volume ratios. Escherichia coli cells achieve
3

a surface area of over 21 m2 /g dry cell weight [108, 158]. With appropriate treatments,
including physical crushing, treating and caustic treatment, bacterial biomass can be an
eﬃcient adsorbent [177]. Indeed, treated bacterial cells are preferred over live bacterial cells
because of convenience (e.g., little maintenance cost and easy to transport). The term bioaccumulation refers speciﬁcally to metabolism-dependent metal uptake. While some metals
at trace levels are required to maintain life (e.g., Se, Cu, Fe), toxic eﬀects can be observed
in bacteria when large amounts of metals are present. Research has shown that 500 µM of
Cu2+ (0.125 g/L) is toxic to E. coli. The ability to grow at elevated levels of copper are
largely due to resistance mechanisms including bioaccumulation, membrane bound e-ﬂux
pumps and metal transporters [147]. The active metal accumulation process is primarily
driven by eﬄux pumps that relocate metals from the interior to the exterior of cells. Active
bioaccumulation is likely to be slower than biosorption and subject to growth condition,
such as narrow ranges of pH, temperature and salt content. However, bioaccumulation has
a great potential in metal remediation because living metal sequestering bacteria have high
speciﬁcity in metal binding and can regenerate biomass via growth [95].
Biosequestration can be practically applied in metal removal/recovery from industrial
waste streams [177]. Wastewater produced from various industries, including mining, leather
working, energy production, and electroplating, contain diﬀerent heavy metals. Discharging
this waste into the environment poses serious risk (e.g., toxic metals like Hg and radioactive metals like U) as well as being a waste of resources (e.g., precious metals like Au).
Conventional methods like chemical precipitation and coagulation are able to reduce metal
concentrations in waste streams only down to the mg/L level. Ion-exchange techniques can
remove and recover metals from wastewater to µg/L levels, however, the cost of exchange
4

resins is high and the resins are prone to poisoning by organics and solids in the wastewater.
Membrane-based processes (e.g., reverse osmosis) are eﬀective at obtaining pure eﬄuents
but are extremely expensive for at large ﬂow rates and they are prone to failure as a results
of membrane fouling [81, 177]. Compared to conventional metal removal/recovery methods,
biosequestration oﬀers several potential advantages and may oﬀer the potential of a costeﬀective alternative method. Bacterial biomass can be relatively inexpensive and can serve
as a sink to all metals except the alkali cations (e.g., Na+ and K+ ). Biosorbents developed
from bacteria biomass have a broad range of metal binding aﬃnities that could be exploited
in selective metal removal/recovery. For example, Ogi et al. showed that Shewanella algae can be used in rare metal indium recovery from solutions at low concentrations. This
study revealed that at pH 3.8, S. algae cells were able to remove 100% of the soluble indium
in solution at concentrations up to 0.2 mol/m3 (229 mg/l). Indium concentrates could be
prepared from indium bound to cells by burning at approximately at 800◦ C [115]. Several
species of the genus Bacillus have shown to be eﬀective at removing Cu and Pb from aqueous solutions [122, 165, 133]. As reported by Tsuruta, Arthrobacter sp., Bacillus sp. and
Lactobacillus sp. isolated from uranium deposits removed 10.9-98.3% of U in solution. Some
of the Lactobacillus isolates were able to remove uranium from seawater as well. A selective
accumulation study conducted using one of the Arthrobacter isolates revealed that UO2+ is
2
preferentially accumulated over six other metal ions including Cu2+ , Mn2+ , Co2+ , Zn2+ ,
Cd2+ and Ni2+ [164]. Although biosequestration is still under development and has not
yet been applied to real wastewater treatment problems, these initial studies suggest that
biosequestration could be a viable process.

5

1.1.1.1

Metal Contamination at Torch Lake

Torch Lake, located in Houghton county, MI, has been on the National Priorities List since
1986. The area formerly known as the Copper Country, was well-known for its metallic
copper ore that was extensively mined from 1890’s to 1969. During the mining period, the
2,700 acre Torch Lake served as a waste disposal site for milling, smelting and leaching
chemicals and by-products. The U.S. Environmental Protection Agency estimated that
approximately 20% of the lake volume was made up of copper mill stamp sands (∼200
million tons). It was also estimated that the contaminated sediments are up to 70 ft deep
and contain up to 5,500 ppm of copper. The ecosystem around the lake was further damaged
by a cupric ammonium carbonate (a common herbicide) spillage (∼27,000 gallons) in 1972
and the disposal of industrial processing water, which contained copper and ammonium that
are 2,400 times and 100 times of the local sewage allowable limits, respectively, in the early
1980’s. With the discovery of extensive ecological degradation and three impairments of
beneﬁcial use (ﬁsh tumors, degraded benthos and ﬁsh consumption advisories), Torch Lake
was also enlisted as a Great Lakes Areas of Concern in 1987. The area has been under
remediation construction since 2003. While tumorous ﬁsh were no longer captured, the
recovery of the area is minimal [169].

1.1.1.2

Adaptation of Raltonisa pickettii in High Copper

Eleven Ralstonia pickettii strains were previously isolated from the highly contaminated
Torch Lake sediments, where copper concentration ranged from 1,000 to 2,500 ppm. Previous studies revealed that all eleven R. pickettii isolates were Cu resistant, while seven of

6

them were also resistant to Zn and Cd and the rest were resistant to Ni and to low levels
of Cd. Konstantinidis et al. [80] observed that the colonies of these R. pickettii isolates
obtained a greenish blue color when growing on copper supplemented agar plates. It was
hypothesized that these isolates could uptake copper from the medium and deposit it in
the outer cell envelope [80]. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) supported this hypothesis and revealed drastic changes to the cells
when the isolates were grown in the presence of copper, including a substantial increase
of the amount of extracellular material and the deposition of high-density materials in the
outer envelope. Electron diﬀraction scanning of the copper-grown cells revealed the presence
of a large amount of copper, which was not observed in the cells grown in the absence of
copper. X-ray absorption near edge structure (XANES) and Extended X-ray absorption
ﬁne edge structure (EXAFS) analyses suggest that the cell associated copper is likely to
be an organic bound tetrachedral cupric complex and the copper sequestration process is
not via (hydr)oxide mineral precipitation [191]. A systematic copper sequestration study
revealed that the isolate representatives R. pickettii strain 12D (Cu-Ni-Cd resistant) and
strain 12J (Cu-Zn-Cd resistant) are well adapted to the high copper environment and the
binding mechanisms were diﬀerent between the viable and heat-killed cells. The live cells
can actively sequester copper up to 27.44 (12D) and 38.19 (12J) mg Cu/g dry weight of
cells, which is 1.5 to over 2 times more than what heat-killed 12D and 12J cells adsorb. The
study also suggests that distinct binding/transporting mechanisms are involved in copper
sequestration by live R. pickettii 12D and 12J cells and that live 12J cells bind signiﬁcantly
more copper than live 12D cells [191]. These attributes suggest that R. pickettii strain 12D
and 12J may be exploitable in the remediation of copper.
7

From the copper sequestering studies performed on R. pickettii strain 12D and 12J, we
concluded that the active uptake of copper by living R. pickettii cells is likely a display of
a detoxiﬁcation process associated with metal resistance systems. In a 2005 review, Silver
et al. noted that bacteria developed metal resistance systems to overcome the toxicity of
various toxic metal ions, including Zn2+ , Ni2+ , Cu2+ , Cd2+ , Co2+ , and Hg2+ . The majority
of these systems function as energy-dependent (e.g., ATPase and chemiosmotic ion/proton
exchange) eﬄux pumps and some also involve metal-binding proteins. For example, a classic
Cd2+ , Zn2+ , and Co2+ resistance system (Czc) is a polypeptide chemiosmotic system. It is
comprised of membrane polypeptide chemiosmotic eﬄux pumps, inner and outer membrane
proteins, and periplasmic coupling proteins to form channels between the cytoplasm and
outside of cells. To reduce metal toxicity, this system utilizes a proton gradient to transfer
excess amounts of Cd2+ , Zn2+ , and Co2+ from inside to outside of the cell [147]. Exploiting
the mechanisms of metal resistance will increase our understanding of how bacteria adapt
and thrive in high metal environments. This information is also important in evaluating
the potential and optimizing the ability of live bacterial cells in removing/recovering metals
from industrial wastewater.
One can envision a number of mechanisms for the acquisition of metal resistance, including the gradual selection of point mutations to develop a new functionality, the duplication/ampliﬁcation of existing metal resistance genes to increase resistance level, and the
acquisition of additional resistance determinants via horizontal gene transfer. Point mutations occur randomly and take a long period of time to evolve a new function from an old
one. Starting from small genetic modiﬁcations, point mutations alter a single base at a time.
Many point mutations are synonymous and therefore will not change the amino acids, which
8

determine genetic functions. Slowly, point mutations may accumulate and eventually change
the amino acids and modify the old function. This process may accelerate under selective
pressure in that suitable mutants adapt to the new environmental factors better [92]. In
conclusion, point mutation is a prolonged process for evolving a completely new function.
Unlike point mutations, gene duplication and ampliﬁcation (GDA) allows bacteria to gain
resistance rapidly [135]. GDA, described by its name, allows a bacterium to evolve by duplicating its own existing resistance determinants. This mechanism confers increased levels
of resistance by increasing the number of metal resistance gene copies, thereby increasing
the levels of metal resistance gene expression [135]. Bacteria can also obtain new functions
from other bacteria (sometimes across species, genus, even domain boundaries), referred
to as horizontal gene transfer (HGT), through transformation (foreign gene direct uptake),
transduction (foreign gene insertion by bacterial phages), and conjugation (plasmid transfer
via cell-cell contact). Both GDA and HGT provide a road to rapid acquisition of resistant
genes. Because they both depend on existing resistance determinants, GDA and HGT are
distinguishable on a sequenced genome. The presence of plasmids or phage indicates that
HGT may have occurred via conjugation or transduction. A stretch of genes on a bacterial
genome with diﬀerent guanosine/cytosine (GC) content compared to the rest of the genome
also can be used to detect putative foreign genes inserted. Similarly, the possible occurrence
of GDA could be identiﬁed by ﬁnding multiple copies of the same gene, frequently repeated
in tandem, in a genome. It is also possible that HGT of the same gene occurred several
times, which would result in multiple copies of the same gene on chromosome. To conﬁrm
GDA and diﬀerentiate from HGT, one has to rely on phylogenetic analyses of gene copies to
reveal the evolutionary lineages of each copy [92, 114, 135].
9

We have demonstrated that R. pickettii strain 12D and 12J have a great potential in
copper removal/recovery from wastewater streams. It has also been shown that these two
strains shared similar attributes (e.g., same species, ability to thrive in high copper environments, and active sequestration of copper) but have important diﬀerences (e.g., diﬀerent
genome sizes, 12J is resistant to Zn but not Ni, 12D is resistant to Ni but not Zn, 12J
grows better in the presence of copper and sequesters more copper than 12D). Hence, these
two strains are great model organisms for studying how genetic element variations inﬂuence
the manifestation of biosequestration by living bacteria. Konstantinidis et al. previously
observed that R. pickettii 12D and 12J are resistant to multiple metals [80]. Therefore, it
is likely that these two bacterial strains obtained multiple metal resistance systems. From
both metal resistance and copper sequestration studies, we observed that R. pickettii 12D
and 12J are diﬀerent both genetically and physiologically. It was observed that these two
strains belonged to diﬀerent genomovar groups and strain 12J appeared to adapt to high
copper environment better than strain 12D [80, 191]. Hence, we suspect that R. pickettii
12D and 12J either acquired diﬀerent metal resistant determinants or developed diﬀerent
regulation mechanisms during evolution. In their 2009 review, Sandegren et al. suggested
that GDA is a common adaptive mechanism of bacteria to various antibiotics [135]. Chan
et al. observed that GDA and HGT contributed greatly to the development of virulence,
antibiotic and metal resistance among Staphylococcus isolates [21]. Thus, as an adaptation
mechanism, we expect that GDA and HGT played a substantial role in R. picket 12D and
12J.

10

1.1.2

Bioimmobilization of Metals

Besides biosequestration, metals also can be removed from aqueous solution via bioimmobilization. Bioimmobilization of metals herein refers to 1) co-precipitation of metals
and metabolic by-products and 2) precipitation of metals via dissimilatory reduction. Coprecipitation of metals and bacterial metabolic by-products describes a chemical process
where soluble metal cations combine with sulﬁdes [82, 183] or phosphate compounds to form
insoluble metal complexes [130]. Large amounts of sulﬁde in the subsurface result from the
reduction of anthropogenically deposited sulfate (e.g., mining). Anaerobic sulfate-reducingbacteria (SRB) obtain metabolic energy (δG = -190 kcal/mol) by reducing sulfate to sulﬁde.
Sulﬁde is subsequently complexed with soluble heavy metal ions to form precipitates [121].
Besides sulﬁde, biogenic phosphate has been observed to be eﬀective in bacterial cell associated metal precipitations. Phosphate is liberated from bacterial storage polyphosphate
(e.g., Acinetobacter spp., Citrobacter sp.) via phosphatase activities when the growth condition changes, such as changing from aerobic to anaerobic metabolism [9, 91, 97]. Research
has revealed that with the presence of soluble heavy metals, the secreted phosphate forms
cell-bound metal-phosphate precipitates [9, 91, 97]. Co-precipitation of heavy metals occurs
in accord with the chemistry solubility rules and it does not change the oxidation state of
metal ions. Compared to co-precipitation of metals and metabolic by-products, metal precipitation via dissimilatory reduction changes the oxidation state of metal ions directly via
redox reactions. Similar to the conversion of sulfate to sulﬁde, anaerobic bacteria conserve
energy by enzymatically transferring electrons to metals ions. While the reduced form of iron
(Fe2+ ) and manganese (Mn2+ ) are more bioavailable than their oxidized forms (i.e., Fe3+

11

and Mn4+ ), other heavy metals are more insoluble in the reduced state (e.g., Se6+ VS. Se0 ,
Cr6+ VS. Cr3+ , U6+ VS. U4+ ). The dissimilatory reduction and precipitation of heavy metals immobilizes metals in situ without additional growth requirements, such as the presence
of sulfate and the need of changing growth conditions. However, the disadvantage of this
dissimilatory metal reduction is that the reduced metals can readily be reoxidized to their
oxidized state by the exposure to oxygen. In contrast, the co-precipitates of heavy metal and
sulﬁde or phosphate are relatively stable in the presence of oxygen. These metabolism dependent characteristics of bioimmobilization are particularly attractive to in situ subsurface
metal remediations for it is sustainable and relatively inexpensive for a large area.

1.1.2.1

Metal Contamination at Oak Ridge Reservation

Historical weapons development and improper disposal of inorganic, organic and radioactive
wastes have caused extensive areas of subsurface contamination at the Oak Ridge Reservation site (Oak Ridge National Laboratory, TN). For 31 years, the unlined S-3 pond was used
to store these toxic wastes. It was later discovered that much of the original contaminants
are now associated with soil and subsurface sediments and became secondary contaminant
sources encompassing an expanding area. The challenge of remediation at Oak Ridge Integrated Field Research Challenge (iFRC) site is the combination of mixed contaminants,
large subsurface area and low pH environment. The wide-spread contamination of both
soil and groundwater makes traditional remediation methods, including in situ containment
and solidiﬁcation, mechanical separation, and pyrometallurgical processes, unfeasible [106].
While the low pH environment is not favorable for either chemical remediation or bioimmobilization, increasing pH in situ would cause the precipitation of aluminum and calcium
12

and result in clogging of subsurface pores. In addition, denitriﬁcation is likely to occur if
growth conditions are optimized because of the high concentration of nitrate (133 mM) in
the groundwater. in situ denitriﬁcation should be avoided because bacteria will preferably
use nitrate rather than U6+ and hence inhibit uranium reduction [187].
Wu et al. [187] explored the approach of combining ex situ denitriﬁcation and in situ
bioimmobilization by installing a nested recirculation system at the Oak Ridge iFRC S-3 site.
The system creates a pH neutral zone by pump and treat, where groundwater is extracted
for above-ground pH adjustment, removal of clogging agents and bulk denitriﬁcation. To
maintain required ﬂow rates, treated water was mixed with tap water prior to reinjection.
The biostimulation zone is located in the center of the pH neutral zone, where ethanol was
injected as a carbon source and electron donor for bioimmobilization of uranium [187]. The
system was able to maintain a stable near-neutral pH for over 500 days. In addition to
removing bulk Al3+ , Ca2+ , and nitrate, ﬂushing also reduced soluble uranium from over
100 µM to ∼2-7 µM in 69 days. Soluble uranium concentration in the biostimulation zone
persisted at around 5 µM until day 177 due to the presence of residue nitrate but decreased
to less than 1 µM by day 268 and to levels below EPA drinking water standard (0.126 µM)
after 2 year of stimulation. Meanwhile, the uranium content in the biostimulation zone soil
increased accordingly and mixed U6+ and U4+ were detected via x-ray absorption near-edge
structure spectroscopy (XANES) [188]. Bacterial community studies of the biostimulated soil
revealed iron reducing bacteria (FeRB), sulfate reducing bacteria (SRB) and denitriﬁers were
grown in the treatment area. Bacteria belonging to groups that were previously identiﬁed
as uranium reducers (e.g., Geobacter, Clostridium) were abundant in the stimulation zone.
Cardenas et al. also observed that the abundance of the SRB group (e.g., Desulfovibiro,
13

Desulfosporosinus) followed the U6+ reduction pattern and suggested that SRB played a
prominent role in uranium immobilization at Oak Ridge iFRC site [16, 17].

1.1.2.2

Bacterial Community Proﬁling

Field studies conducted at Oak Ridge iFRC site proved that in situ bioimmobilization of
uranium is a suitable strategy. However, complex ﬁeld conditions complicate the identiﬁcation of uranium immobilizing bacteria due to the presence of other metals and minerals.
It was also extremely diﬃcult to study the interactions between diﬀerent bacterial strains
in response to the presence of uranium. Hence, we took the approach of enriching uranium
immobilizing bacterial consortia from the iFRC and anaerobicly isolating potential uranium
immobilizers in the laboratory.

1.1.2.3

Reconstruct Uranium Immobilizing Bacterial Communities

In general, microbes exist as communities in the environment, where the diversity of species
reﬂects the diversity of nutrients and the abiotic conditions. In such a diverse community,
interactions between species is complex and includes mutualism, commensalism, competition
and parasitism. One of the most studied relationships is the syntrophism between bacteria
and archaea in methanogenesis. The Methanobacillus omelianskii culture Barker et al.[140]
isolated was identiﬁed later as a co-culture of S strain and M.o.H strain. The two strains
could not be separated because each strain depended on the other under the isolation conditions. Strain S ferments ethanol and strain M.o.H scavenges H2 produced by strain S. The
collaborative process generates methane as well as 112 kJ of energy per mole of methane
produced [140]. Besides methanogenesis, Kato et al. [74] reported a mixed-culture composed

14

of cellulolytic Clostridium and non-cellulolytic bacteria that can degrade lignocellulose more
eﬀectively than isolates. The study showed that it is critical to have non-cellulolytic bacteria scavenge cellulose derivatives and metabolites, maintain the anaerobic condition, and
neutralize the pH of the system. It was also revealed that not all non-cellulolytic isolates
contributed to cellulose degradation [74]. A further study by the same group showed that
the four strains in the assembled cellulolytic community formed a complex network [75].
Besides the collaborative eﬀort on consuming metabolites, a recent study demonstrated that
electrons can also be transferred between two diﬀerent bacterial species. Kato et al. cocultured Geobacter sulfurreducens and Thiobacillus denitriﬁcans. The study showed that G.
sulfurreducens transfered electrons from acetate to T. denitriﬁcans, where the electron was
further passed onto nitrate to produce ammonia [76]. These studies suggest that microbial
interactions and cooperations are important and cannot be ignored in engineered microbial
systems.
Uranium immobilizing bacterial communities have been studied using two approaches:
bottom-up and top-down. A bottom-up approach refers to the strategy of studying the system from individual components, where individual bacterial species are isolated and studied
separately. Under strictly controlled conditions, one can determine the uranium immobilization ability of individual bacterial isolates. This approach is eﬀective in understanding cell
physiologies of each microorganism. However, the bottom-up approach suﬀers from a focus
on strains that are easily cultivated while neglecting the often more numerous and diﬃcult
to cultivate strains. A top-down approach is also used in studying uranium immobilizing
communities in order to correlate the total diversity of the community using cultivationindependent techniques with the abiotic conditions (e.g., pH, uranium concentration, etc.).
15

This community information is used to tentatively identify the important bacterial species
responsible for uranium immobilization. This approach is extremely powerful because the
cultivation-independent approach reveals nearly all populations. However, top-down approaches cannot be used to precisely evaluate population functions due to the large number
of variables and in addition, physiology does not always follow phylogeny. Zengler et al.
made the point that both top-down and bottom-up approaches should not be treated separately but, rather, should be integrated [193]. It was suggested that bacterial community
analyses should be used as a guide for cultivation and in-depth understanding of individual
populations. In return, the comprehensive understanding at the cellular level would assist
in prediction of complex biological systems.
Integrating the two approaches, reconstruction of uranium immobilizing bacterial communities would give insights on population interactions in uranium bioremediation systems.
As part of my dissertation, I investigated the uranium immobilizing bacterial communities
with a top-down approach. I also cultivated individual bacterial isolates and studied their
abilities in uranium immobilization.

1.2

Bioremediation of Agriculture Wastes

Agriculture is a signiﬁcant part of the US economy. The Food and Agriculture Organization
of the United Nations (FAO) estimates that approximately $0.2 trillion worth of production
(0.67 billion tons) was contributed by agriculture in the US alone in 2010. The top three
produced commodities based on value were indigenous cattle meat, dairy products and maize.
Maize was also the most abundantly produced based on quantity (at least three times more

16

than other commodities) in the US in 2010. The large level of agricultural production
resulted a contaminant amount of waste production. Approximated 6.7 million tons of
agriculture wastes were produced in 2010 in the US (estimated based on 1993 facts) [44].
The consequence of improper treatment and disposal of agricultural wastes can be serious.
Manures produced by diary and swine farms often contain human pathogens and can be life
threatening if the wastes come into contact with water sources. The direct disposal of wastes
that are rich in nutrients into the environment can alter the local ecosystems, causing for
example, algal blooms. Crop wastes, such as corn stalks are much harder to degrade due
to their lignin and cellulose content. Under strict regulations, wastes may be recycled by
applying them to ﬁelds as fertilizer or they can be used to generate biofuels (e.g., bioethanol).
Currently, the most common treatments for agricultural waste are composting and lagoons
[166]. Although the typical farm waste management strategy includes waste storage and
decomposing, it does lead to concerns regarding air and water quality and greenhouse gas
emissions [46].
Methane is a potent greenhouse gas and has a unique role in global climate change and
carbon cycling. Petit et al. [125] revealed the high correlation between temperature variation
and atmospheric green house gas concentration from the Vostock ice-core record. The past
420 kyr of record also suggested that the present day atmospheric CH4 concentrations are
unprecedentedly elevated [125]. As the main component of natural gas, methane is naturally
produced by syntrophic microorganisms (i.e., hydrogen producing bacteria and methane
generating archaea) during degradation of organic matter via anaerobic methanogenesis [6,
140]. Methanogenesis generates the least amount of energy (∼15% of aerobic degradation
of the same organic molecule) and occurs as the last step in microbial anaerobic respiration.
17

Methane production is common in organic-rich sediments [140]. Bastviken et al. [6] estimated
that the global freshwater methane emission rate corresponds to at least 0.65 Pg of C (CO2
equivalent, assuming 1 kg of CH4 equals to 25 kg of CO2 over a 100 year period). Besides the
naturally emitted methane, anthropogenic methane emission rose substantially along with
the development of industry. Stern and Kaufmann [153] estimated that the total amount
of global methane emitted from anthropogenic activities increased from 79.3 Tg in 1860 to
371 Tg in 1994 (0.5 Pg and 2.3 Pg CO2 equivalent C, respectively). Over 50% of this was
emitted from agricultural activities including livestock production and rice farming [153].
While the terrestrial land surface has been considered an important carbon sink (2.6±1.7 Pg
of C), the amount of natural and anthropogenic methane emissions oﬀsets the continental
green house gas balance. Thus the impact of global methane emission rate on global climate
change should not be overlooked [6].
Besides being a major contributor to global climate change, methane is also an alternative energy source for heat and electricity [46, 78]. With decreasing of fossil fuel discovery
rates, the need for renewable energy source increases [15, 66]. Anaerobic digestion of farm
wastes not only eliminates some of the environmental concerns of traditional farm waste
management methods, such as oder and water contamination, it also can be used to generate and capture methane rich biogas. However, system optimization is critical for anaerobic
digestion. Operational parameters, including temperature, HRT, and the carbon to nitrogen
ratio of the substrate have been studied extensively [42, 46, 110, 120]. While these studies
provide insights on maintaining an eﬃcient anaerobic bioreactor, the microbial communities
and species interactions responsible for methane production are not well understood. In this
section, I will focus on studying the microbial community aspects of methane producing
18

anaerobic co-bioreactors.

1.2.1

The Role of Class Clostridia in Methane Producing Anaerobic Co-bioreactors

Anaerobic methanogenesis is a group eﬀort between bacteria and archaea. Methanogens,
a group of archaea, are primary responsible for methane production. However, the process
cannot be carried out without the help of bacteria [6, 140]. Schink et al. [140] noted in a
recent review that methanogenic degradation of complex organic matter involves a mutual
dependency of ﬁve groups of microorganisms, primary fermentative bacteria, secondary fermentative bacteria, homoacetogenic bacteria, hydrogen oxidizing methanogens, and acetatecleaving methanogens. The concentration of hydrogen in the system plays a critical role
in determining how methane is produced. When the hydrogen partial pressure is less than
10−4 bar, majority of the methane is generated exclusively via the cooporation of primary
fermenters and either H2 oxidizing methanogens or acetate cleaving methanogens. Under
conditions where hydrogen accumulates in the system (e.g., excess substrate), the secondary
fermenters and homo-acetogenic bacteria contribute to methane production by converting
hydrogen to acetate. This alternative process shifts the methane producers from hydrogen
oxidizing methanogens to acetate-cleaving methanogens [140].
Bacteria class Clostridia categorizes a group of anaerobic bacteria with versatile metabolic
pathways. Species belonging to this class are noted for their ability to ferment various substrates, including saccharides, amino acids, and cellulose. Cellulose degradation is particularly relevant as it is a major component of plants.

Thus, it is widely abundant,

rich in carbon and yet diﬃcult to decompose. Studies have shown that several Clostridium species can decompose and partially hydrolyzed cellulose [36, 74]. Interestingly, many
19

members of Clostridia are also homo-acetogens, such as C. methoxybenzovorans, C. thermoautotrophicum and C. thermoaceticum. These attributes of class Clostridia suggest that
this group of bacteria may play a substantial role in anaerobic degradation of agricultural
wastes and methane generation. In fact, research studies have repeatedly reported that
Clostridia populations were abundant in hydrogen and methane producing anaerobic bioreactors [23, 116, 119, 126, 179]. Bacterial hydrogen production from protons and electrons is
catalyzed by hydrogenases. There are two major groups of hydrogenase, [FeFe]-hydrogenase
and [NiFe]-hydrogenase, which is composed of two iron atoms and one iron and one nickel
atom, respectively. The two groups of hydrogenases are phylogenetically diﬀerent. While
[NiFe]-hydrogenases widely occur in bacteria and archaea, [FeFe]-hydrogenases are found
only in anaerobic bacteria and eukaryotes [11, 103]. It is known that [NiFe]-hydrogenases
are more abundant in bacteria than [FeFe]-hydrogenase. However, it was suggested that
most of [NiFe]-hydrogenases are involved in hydrogen uptake. This is consistent with the observation that [NiFe]-hydrogenases are often found in hydrogen consuming bacteria [10, 176].
In contrast, [FeFe]-hydrogenase is 10-100 times more active than [NiFe]-hydrogenase and energetically prefers to catalyze the reduction of protons to produce hydrogen [47, 189]. Therefore, I hypothesize that class Clostridia is a substantial player in anaerobic co-bioreactors
and is the dominant bacterial group in methane producing anaerobic co-bioreactors.

1.3

Managing Complex Microbial Communities for Sustainable Environment

Microorganisms reside in various environments, such as soil, water, air, and the intestines
of living organisms. Among all microorganisms, bacteria and archaea are highly abundant

20

on earth. Whitman et al. estimated that there are 4-6×1030 bacteria and archaea cells in
terrestrial and marine environment [184]. It was estimated that there are at least 4.5×106
diﬀerent bacterial species globally and 2×104 platonic marine archaea taxa [31]. These
bacteria and archaea are the important components of primary global productivity (e.g.
photosynthesis) and geochemical cycling. To date, approximately 2.9×105 bacterial and
archaeal strains have been isolated and approximately 1.3×104 bacterial and archaeal taxa
have been identiﬁed [40]. The bacterial and archaeal diversity we have discovered is less
than 0.3% of the total diversity based on conservative estimates. Although our knowledge
is still limited, studies have shown that microbial communities are valuable resources and
eﬀectively managing these resources are the essential to a sustainable environment [175].
Human beings has taken advantage of microbial systems for millennia, dating back to
early wine and cheese making. More recently we have used bacterial communities and populations to remediate organic wastes, produce commodity chemicals and contain subsurface
contamination. Understanding the interactions among microbial populations is a vital step
in managing microbial resources. Microbial populations react to system changes and undesirable shifts in a microbial community can result in system failures. As an example,
sludge bulking in wastewater treatment plants is caused by the excessive growth of ﬁlamentous microorganisms in activated sludge. While the presence of ﬁlamentous microorganisms
helps in the degradation of organic matter, when abundant, ﬁlamentous microorganisms
prevent activated sludge from settling, resulting in failure of the wastewater treatment system [61, 131]. Recent advances in our understanding of the human microbiome also suggest
that the changes in composition of bacterial communities, which reside in diﬀerent parts of
human body, can alter human health [4, 127, 150, 173]. Microbial ecology studies the micro21

bial population interactions, community structures and interactions with the environment
including changes due to external factors. Knowing more about these interactions will help
us to better utilize microbial resources and control engineered systems.
Each topic of my dissertation involves a diﬀerent microbial community with a unique
energetic strategy. The lake sediment community of Torch Lake has been under extreme
pressure from unnaturally high contamination of metals and the R. pickettii strains isolated
from these sediments are uniquely adapted to high copper concentrations. Thus, these strains
can provide insight into how microbial populations adapt to high metal concentrations. The
sediment communities of Oak Ridge National Laboratory iFRC sites have responded to
metal, acid and hydrocarbon contamination and many strains isolated from these communities can immobilize uranium. We have extended the community analyses of these sediments,
as well as identiﬁed new strains capable of uranium immobilization. This work contributes
to our understanding of the interactions between uranium and microbial populations and
provide insights to in situ bioremediation of uranium. Finally, The microbial community of
an anaerobic bioreactor managing farm waste has a complex community of bacteria and archaea that removes organic matter and generates methane. We have identiﬁed a community
structure that correlates with production of a large amount of methane. In each instance,
knowing the conditions under which the community operates optimally is advantageous to
the engineer. By "optimally" we refer to an outcome that is successful for long-term bioremediation and sustainability. In most cases knowing ﬁrst, who is there (i.e. what species
are present and their growth preferences) and second, what is the metabolic strategy of
the community, can assist engineers in designing and controlling an environment that imposes suﬃcient constrains on the community such that it becomes predictable. Successful
22

management requires predictable outcomes.

23

CHAPTER 2
CO-EVOLUTION OF RALSTONIA PICKETTII STRAIN 12D AND 12J:
GENOME REARRANGEMENTS AND THEIR INFLUENCES ON THE
EVOLUTION OF METAL RESISTANCE

2.1

Abstract

The analysis of natural divergence of phylogenetically closely related bacteria residing in
the same habitat can reveal ﬁne-scale evolution events. We have previously reported on
Ralstonia pickettii strains (12J and 12D) isolated from the same cubic centimeter of coppercontaminated lake sediment that bioaccumulate copper. While these strains have nearly
identical 16S rRNA gene sequences (99% similar), they have distinct genomic structures on
which we report here. Strain 12J has two chromosomes, a single plasmid, and a ﬁlamentous phage, while strain 12D has two chromosomes, three plasmids, and a larger genome by
359,629 bases. Both strains have adapted to the high concentrations of metals using gene duplication and horizontal transfer of metal resistance genes. Plasmids and phage have played
a role in the genetic divergence of the strains. Plasmid 2 of strain 12D contains 31 copper and cobalt-zinc-cadmium resistance genes. This plasmid was inserted into the primary
chromosome and initiated, we postulate, a cascade of genetic events that have continued the
divergence of the two strains. Subsequent events included diﬀerences in horizontally transferred genes, acquisition of a ﬁlamentous phage, and diﬀerences of transposase carriage, with
ISXo4 and IS4 unique to strain 12J and the mutator type of transposase unique to strain
12D. Five transposase types are found in both strains. The genomic evolution of strains
12D and 12J provide a view into a collection of dynamic events shaping and reshaping the
24

Ralstonia genome.

2.2

Introduction

Microdiversity describes a group of bacterial populations found in the same environmental
niche that are phylogenetically closely related (ie. same species) but physiologically distinct
[1, 124, 141]. Studying genomic variations within a bacterial population can give us insights
into: 1) the impacts of the environmental factors on the bacteria populations; 2) the incidences of horizontal gene transfer (HGT) and gene duplication and ampliﬁcation (GDA),
that led to the divergence; 3) the extent and location of plasticity within the genome; 4) and
the functional potential of the microdiversity [21, 98, 124, 135, 146].
We have a collection of bacteria strains from a Ralstonia pickettii microdiversity. Two
representative strains, R. pickettii 12D and 12J were isolated from within the same 1 cm3
of heavy metal (especially copper) contaminated lake sediment (16 cm below the sediment
surface at a lake depth of 37 m from Torch Lake, MI). Konstantinidis et al. [80] showed
that these two strains are 99% identical according to their 16S rRNA gene sequences. Previous research showed that both strains thrived in copper supplemented media with minor
diﬀerences in growth rates and copper sequestration abilities [191]. Metal resistance tests
revealed that R. pickettii 12D is Cu-Ni-Cd resistant and 12J is Cu-Zn-Cd resistant [80]. We
also found that R. pickettii 12J contains ﬁlamentous phage (p12J) particles and strain 12D
is p12J resistant [191]. These physiological diﬀerences indicate that R. pickettii 12D and
12J have distinct genome structures.
By studying the genomic variations of these two R. pickettii strains we will better un-

25

derstand the evolutionary consequences of a high copper environment, produce a detailed
description of the genomic changes that have occurred between strains and improve our
understanding of strain divergence and speciation. Herein, we report our ﬁndings on the
genomic diﬀerences, evidence of HGT and GDA, identiﬁcation of metal resistance determinants and the evolutionary events that led to the diﬀerentiation of R. pickettii 12D and
12J.

2.3
2.3.1

Materials and Methods
Genome Retrieval and Gene Annotations

The genomic DNA of R. pickettii 12D and 12J were isolated in our laboratory and sequenced
by Joint Genome Institute (JGI) as reported previously [191]. The genome sequences of R.
pickettii 12D and 12J were retrieved from GeneBank database by using the following IDs,
NC_012849, NC_012851, NC_012855−57, NC_010678, NC_010682-83, and NC_005131.
The basic gene annotations were obtained from Joint Genome Institute Integrated Microbial
Genomes (JGI-IMG) database. Basic Local Alignment Search Tool (BLAST) and JGI-IMG
genome BLAST was used to identify and conﬁrm gene annotations (E-value ≤ 10−2 , protein
identity ≥ 30%, unless speciﬁed).

2.3.2

Genome Comparisons

Functional groups identiﬁed as clusters of orthologous groups (COG) were used to ﬁnd
shared and unique genes present in R. pickettii 12D and 12J. Metabolic pathway information was retrieved from MetaCyc [19] and Kyoto Encyclopedia of Genes and Genomes

26

(KEGG) database [73]. Genome replication origins (ori C) of R. pickettii 12D and 12J were
retrieved from the DoriC database [48]. The replication resolution sites (dif ) of 12D and
12J were located by searching for sequences comparable to the R. solanacearum GMI1000
dif sequence by BLAST (BLASTN). The phylogenetic positions of strain 12D and 12J were
analyzed, along with a collection of β-Proteobacteria dif sequences [18], by using Molecular
Evolutionary Genetics Analysis (MEGA) [161]. Brieﬂy, the dif sequences were aligned by
ClustalW and a minimal evolution (ME) phylogenetic tree was constructed with the following parameters: the Jukes-Cantor substitution model assuming uniform rate among sites,
gaps/missing data treated as pairwise deletions, and ﬁnally clustering with the ME heuristic
method by allowing close-neighbor-interchange. Bootstrapping (1000 replications) was used
to test the conﬁdence level of branching.
Genome statistics (e.g., copies of rRNA coding genes, protein coding genes, horizontally
transferred genes, etc.) of R. pickettii 12D and 12J and phylogenetically related bacteria (R.
solanacearum GM1000 and Cupriavidus metallidurans CH34) were retrieved from JGI-IMG
and compared based on their annotation. Transposases encoding sequences were identiﬁed
by JGI-IMG. The genomic syntenies of R. pickettii 12D and 12J were assessed by using
Mauve [33], a multiple genome alignment tool.

2.3.3

Identiﬁcation of the Metal Resistant Genes

Metal (Co-Zn-Cd, Cu, and Hg) resistant genes in R. pickettii 12D and 12J were extrapolated
from JGI-IMG based on the metal resistant genes identiﬁed in C. metallidurans CH34.
The abundances of these gene copies in R. pickettii strains were compared to those found
in C. metallidurans CH34 and R. solanacearum GMI1000. The evolution lineage of the
27

main subunit of these resistance determinant coding genes in R. pickettii 12D and 12J
was determined by comparing to similar coding sequences found in 23 other Proteobacteria
species retrieved from JGI-IMG (Figure A.1). The phylogenetic analysis of these genes was
conducted by using MEGA. Brieﬂy, the DNA sequences were aligned via ClustalW and the
bootstrapped (1000 replications) ME trees were constructed with the following parameters:
the Jukes-Cantor substitution model assuming uniform rate among sites, gaps/missing data
treated as pairwise deletion, and close-neighbor-interchange for ME heuristic method.

2.4
2.4.1

Results
Genome Overview

JGI-IMG annotations identiﬁed 1936 and 1945 COG groups in 12D and 12J, respectively.
Between the two strains, 1876 COG groups are shared. KEGG data revealed that the
metabolic pathway maps of both strains are highly similar with minor diﬀerences (Appendix
A). MetaCyc revealed that both strain 12D and 12J contain several detoxiﬁcation pathways involved in arsenate reduction, cytoplasm Hg2+ reduction, methylglyoxal detoxiﬁcation, cyanate degradation and superoxide radical degradation (Appendix A). R. pickettii
12D and 12J clustered closely to R. solanacearum GMI1000 and Cupriavidus metallidurans
CH34 based on their 16S rRNA gene sequences (Figure A.1). Basic comparisons among
these four strains showed that R. pickettii 12D and 12J are smaller in genome size compared
to R. solanacearum GMI1000 and C. metallidurans CH34. R. pickettii 12J has the least
number of bases and protein coding genes (Table 2.1). Multiple copies of rRNA genes are
found in all four bacterial strains. Both R. pickettii strains contain three copies of the 16S

28

rRNA genes. IMG annotations indicate that all four strains contain comparable numbers of
orthologs and paralogs and R. pickettii 12D and 12J contain the least number of horizontally
transferred genes. All four strains contain many signal peptide and transmembrane protein
encoding genes (Table 2.1).
The complete genome sequences revealed that the R. pickettii 12D genome is composed
of a primary chromosome (12D_C1) of 3.6 Mb, a secondary chromosome (12D_C2) of
1.3 Mb, and three plasmids of 0.39 Mb (12D_P1), 0.27 Mb (12D_P2), and 0.051 Mb
(12D_P3) (Figure 2.1). R. pickettii 12J contains a primary chromosome (12J_C1) of 3.9
Mb, a secondary chromosome (12J_C2) of 1.3 Mb, and a plasmid (12J_P1) of 0.08 Mb
(Figure 2.2). R. pickettii 12J_C1 is approximately 0.3 Mb larger than 12D_C1. The
secondary chromosomes of both strains are approximately the same size and substantially
smaller than the primary chromosomes. R. pickettii 12D_P3 and 12J_P1 are comparable
in size. Two large plasmids (P1 and P2) in 12D are absent in 12J. We also found that strain
12J contains ﬁlamentous phage particles (p12J), which is absent in strain 12D.
Horizontally transferred foreign genes (JGI-IMG prediction) are scattered across all
genome scaﬀolds and are twice as abundant in 12D than 12J (Table A.1). A total of 123 genes
were predicted as being horizontally transferred genes in R. pickettii 12J and the majority
of them are in 12J_C1 (Table A.1). Out of the 287 horizontally transferred genes in strain
12D, 115 are in 12D_C1 and 157 are in 12D_P1 (Table A.1). The JGI-IMG gene proﬁler
revealed that strain 12D and 12J share only one gene among all 410 predicted horizontally
transferred genes (JGI-IMG gene proﬁler with E-value ≤ 1E-5 and protein identity ≥ 80%).
This gene is annotated as an IS30 family integrase catalytic region (locus Rpic12D_1028
and Rpic_2522).
29

Table 2.1: Gene summaries of R. solanacearum GMI1000, C. metallidurans CH34, R. pickettii 12D and 12J.
Genome Features

R. solanacearum
GMI1000

C. metallidurans
CH34

R.pickettii
12D

R. pickettii
12J

Bases
GC (%)
Total Number of Genes
Protein Coding Genes
rRNA (5S, 16S, 23S)
tRNA

5810922
66.98
5204
5120
4,4,4
57

6913352
63.53
6430
6340
2,4,4
62

5685358
63.3
5518
5452
3,3,3
54

5325729
63.65
5092
5027
3,3,3
55

Orthologs
Paralogs
Signal Peptides
Transmembrane Proteins
Horizontally transferred

4774
1106
1232
1188
271

(91.74%)
(21.25%)
(23.67%)
(22.83%)
(5.21%)

5951
1753
1568
1514
510

30

(92.55%)
(27.26%)
(24.39%)
(23.55%)
(7.93%)

4951
1089
1349
1339
287

(89.72%)
(19.74%)
(24.45%)
(24.27%)
(5.20%)

4774
1106
1286
1291
123

(93.75%)
(21.72%)
(25.26%)
(25.35%)
(2.42%)

0
3.6
0.3

3.3

0.6

3.0

0.9

C1

2.7

1.2

2.4
2.1

1.8

1.5

Figure 2.1: R. pickettii 12D genome scaﬀolds. The distribution of putative horizontally transferred genes (HGT genes) identiﬁed by JGI-IMG are marked as black
dashes in rings 4-10. The scaﬀold sizes are in units of Mb. For interpretation of the
references to color in this and all other ﬁgures, the reader is referred to the electronic
version of this dissertation.

31

Figure 2.1 (cont’d)

1.30
1.2

0.1
0.2

1.1

0.3

1.0

C2
0.4

0.9
0.5

0.8
0.6

0.7

32

Figure 2.1 (cont’d)

0

0.3

0.25

P1

0

0.1

0.05

0.2

0

0 5
.0

0.04

P2

0
.2

0.01

0
.1

0.03

0 5
.1
33

P3

0.02

Figure 2.1 (cont’d)

1
2
3
4
5
6

oriC
xe
rC
xe
rD
d
if
d
naA
re A
p

7
8

10
1
1
1
2

RingLe e
g nd

9

1. Ge sonforwards
ne
trand
2. Ge sonre rs trand
ne
ve es
3. RNA g ne
e s(tRNAsg e rRNAsre othe black)
re n,
d,
r
4. HGT g ne
e sfromArchae
a
5. HGT g ne
e sfromEukaryota
6. HGT g ne
e sfromVirus s
e
7. HGT g ne
e sfromphylumothe thanProte
r
obacte
ria
8. HGT g ne
e sfromclas
sothe thanBe
r
ta-prote
obacte
ria
9. HGT g ne
e sfromorde othe thanBurkhode
r
r
riale
s
10. HGT g ne
e swithbe t hitstog ne
s
e sfromGam a-prote
m
obacte
ria
andnohitstog ne
e sfromBe
ta-prote
obacte
ria
1 GC conte
1.
nt
12. GC s w
ke

34

0
0.3

3.6

0.6

3.3

3.0

0.9

C1

1.2

2.7

2.4

1.5
1.8

2.1

Figure 2.2: R. pickettii 12J genome scaﬀolds. The distribution of putative horizontally transferred genes (HGT genes) identiﬁed by JGI-IMG are marked as black
dashes in rings 4-9. The scaﬀold sizes are in units of Mb.

35

Figure 2.2 (cont’d)

0

1.2

0.1
0.2

1.1

1.0

0.3

C2

0.4

0.9

0.5

0.8
0.7

0.6

36

Figure 2.2 (cont’d)

0
0.08
0.01

0.07

P1

0.06

0.02

0.03

0.05
0.04

37

Figure 2.2 (cont’d)

1
2
3
4
5

oriC
xe
rC
xe
rD
d
if
d
naA
re A
p

6

10
1
1

7
8
9

RingLe e
g nd
1 Ge so
. ne nfo ards
rw
trand
2 Ge so
. ne nre rs trand
ve es
3 RNA g ne
.
e s(tRNAsg e rRNAsre o r RNAsblack)
re n,
d, the
4 HGT g ne
.
e sfro
mArchae
a
5 HGT g ne
.
e sfro
mEukaryo
ta
6 HGT g ne
.
e sfro
mVirus s
e
7 HGT g ne
.
e sfro
mphylumo r than Pro o
the
te bacte
ria
8 HGT g ne
.
e sfro
mclas the thanBe
so r
ta-pro o
te bacte
ria
9 HGT g ne
.
e sfro rde o r than Burkho riale
mo r the
lde
s
1 . GC co nt
0
nte
1 . GC s w
1
ke

38

Both R. pickettii strains are abundant in transposases (Table A.1). Strain 12J has 32
transposase genes which belong to seven transposase families and 12D has 33 transposase
genes that belong to eight transposase families. Five of these transpoase types are found in
both strains. Among the abundant transposases, type ISXo4 and IS4 are unique to strain
12J and the mutator type of transposase unique to strain 12D.
The genome replication initiation positions (ori C) are labeled as black arrows in Figure 2.1 and Figure 2.2. The estimated ori C positions are just upstream of dnaA (C1) or
repA (C2) genes. R. pickettii 12D_P1 also has a repA gene (Figure 2.1 P1 black hollow
arrow), which was not observed in the other plasmids of 12D and 12J. The identiﬁcation of
dif /xerCD system in both strains suggest that this is used to resolve dimers created during replication [56]. Each strain has a pair of xer CD located on the primary chromosomes
(Figure 2.1 and Figure 2.2 black and blue solid triangles). Each strain also contains two dif
sites, one is located on C1 and the other one is on C2 (Figure 2.1 and Figure 2.2 red solid
triangles). There are four copies of kfr A gene in each strain that presumably contribute to
plasmid stability. Two of the four kfr A copies are located on the secondary chromosomes,
and the other two were found on 12J_C1 and 12D_P1, respectively.

2.4.2
2.4.2.1

Co-evolution of R. pickettii 12D and 12J
16S rRNA Gene Sequence Variations

For the three copies of 16S rRNA genes in each R. pickettii strain, we observed that two
copies are located on the primary chromosome and the third copy is found on the secondary
chromosome. The three copies in strain 12J are all identical. The 16S rRNA gene copy on

39

the 12D secondary chromosome (12D_C2) is identical to the copies in 12J. The two copies
of 16S rRNA on the primary chromosome of 12D (12D_C1) are identical to each other but
diﬀerent to the copy on 12D_C2, with a distance of 0.001.

2.4.2.2

dif Site Sequence Variations

Many bacteria employ the dif /xer CD system to resolve dimers formed during chromosome
replication [18, 25, 34, 68, 109, 144]. This system is comprised of a set of recombinase encoding genes, xer C and xer D, and a 28 bp binding site (dif site) near the replication terminus
[18]. Genome sequences revealed that R. pickettii 12D and 12J also utilizes dif /xer CD system. We found each strain contains one set of xer CD genes and two dif sites (one on each
chromosome). The dif site sequences from the primary and secondary chromosomes of R.
pickettii vary substantially, although they presumably interact with the same sets of XerCD
protein. R. pickettii 12J_C1 dif appears to be the most diﬀerent. The dif sequences on the
secondary chromosomes are highly similar (Figure A.2). The ME phylogenetic tree revealed
that the R. pickettii dif sequences clustered together, away from other β-Proteobacteria dif
sequences (Figure 2.3). The dif sequence of 12J_C1 appears to be the most ancestral among
all four R. pickettii. We observed that 12D_C1 and both secondary chromosomes branched
out from 12J_C1 and the secondary chromosome dif sequences appeared to more distant to
12D_C1 than 12D_C1 (Figure 2.3).

40

Burkholderia mallei ATCC 23344 chr. 1
Burkholderia mallei ATCC 23344 chr. 2
46
Dechloromonas aromatica RCB
9
Ralstonia solanacearum GMI1000
7 Thiobacillus denitrificans
Bordetella pertussis Tohama I
35
Methylobacillus flagellatus KT
22
Rhodoferax ferrireducens DSM 15236
41
Nitrosomonas europaea ATCC 19718
Neisseria meningitidis MC58
Ralstonia pickettii 12J C1
94
Ralstonia pickettii 12D C1
Ralstonia pickettii 12D C2
94
0.05
Ralstonia pickettii 12J C2
98
71

70

Figure 2.3: Minimal evolution phylogenetic tree of β-Proteobacteria dif sequences. The numbers on branches
are bootstrap values. The scale at the left hand lower corner represents the distance.

41

2.4.2.3

Filamentous Phage Elements in R. pickettii 12D and 12J

Orf
1 2 34 5

p12J: p12Jp01-10

6 7 8 9 10

12J-C1_1: Rpic_1397-1405
12J-C1_2: Rpic_1391, 1393-95
Replication
initiation factor

12J-C1_3: Rpic_2540, 2542-4, 2546-9
Repressor

12J-C2: Rpic_4343-5, 4347, 4349-50

Resolvase
invertase

Repressor

12D-C1: Rpic12D_2133, 2142
Resolvase
invertase

12D-C2: Rpic12D_4471, 4475
Replication
initiation factor

phiLf Zonular
Protein occludens
toxin

Figure 2.4: Filamentous phage p12J gene compositions and their homologs on R.
pickettii 12D and 12J chromosomes. Fragments of the same fragments represent
homologous genes. Genes colored yellow with a red border represent genes that
share no homology to p12J Orf7 but are orthologs to each other. Grey colored
fragments represent other phage related genes (not p12J related). White colored
blocks are other non-phage related genes.

Filamentous phage p12J (NC_005131) is comprised of ten open reading frames (orf),
locus p12Jp01-10 (Figure 2.4 p12J). Three (orf8 - orf10) out ten of the p12J orf’s were
annotated as phage proteins but with little information. Orf5, orf7 and orf10 are the main
42

phage assembly proteins and orf7 was identiﬁed as a zonular occludens toxin coding gene.
We located four incidents of p12J insertion in R. pickettii 12J (Table A.1). It was observed
that the complete phage genome is located on 12J_C1, 464 bp upstream from the 12J_C1
dif site (Figure 2.4 12J-C1_1). Upstream adjacent to this complete phage genome is a
partial p12J copy with a non-p12J related phage replication initiation factor encoding gene
(Figure 2.4 12J-C1_2). The third insertion is also located on on 12J_C1 (Figure 2.4 12JC1_3). We found the fourth partial copy on 12J_C2 (Figure 2.4 12J-C2), which is located
931 bp downstream of 12J_C2 dif site. All three insertions in 12J contain at least four p12J
gene homologs, two out of which are primary phage components (orfs 5 & 7). In contrast,
only two p12J partial copies were found on 12D_C1 (Figure 2.4 12D-C1) and 12D_C2
(Figure 2.4 12D-C2). Both insertions are distal from R. pickettii 12D dif sites and appear
to be highly truncated by either non-phage related genes or other phage (non-p12J) related
genes. These 12D partial copies contain only two p12J homologs. Annotation showed that
one of the genes encodes zonular occludens toxin. However, it shares no synteny with p12J
orf7 (Figure 2.4 12D-C2).

2.4.3

Genome Syntenies of R. pickettii 12D and 12J

Genomic DNA sequence dotplot revealed that R. pickettii 12D and 12J share high synteny
(Figure A.3). Detailed genomic sequence alignments showed that R. pickettii 12J_C1 could
be viewed as homologs of R. pickettii 12D_C1 and 12D_P2. The 12D_P2 homolog was
located between 1.56 Mb and 1.95 Mb on 12J_C1, approximately 0.07 Mb downstream of
the dif site (Figure 2.5 A). Figure 2.5 panel B indicated where the genomes where rearranged. The replication initiation region of 12D_C2 appears to be drastically diﬀerent from
43

A. R. pickettii 12D C1

12D P2

R. p t ii 12J C1
icke t

B. R. pickettii 12D C2
R. p t ii 12J C2
icke t

C. R. pickettii 12D P3

R. p t ii 12J P1
icke t
Figure 2.5: Genome syntenies between R. pickettii 12D and 12J scaﬀolds. Panel A,
12D_C1, 12J_C1 and 12D_P2; Panel B, 12D_C2 and 12J_C2; Panel C, 12D_P3
and 12J_P1. Arrows indicate location of ori C on the chromosomes. Blocks with the
same color represent the regions that are similar (connected by lines). The amount
of color indicates the percentage of identity. White areas within the blocks indicate
absence of a homologous sequence.

44

annotated 0 position. If repositioned, both secondary chromosomes would start with repA
(12D_C2 is almost identical with 12J_C2, Figure refmauve B). The 12J_P1 and 12D_P3
alignment revealed a similar genome rearrangement pattern (Figure 2.5 C). We could not
locate the origins of replication in this study. The large plasmid of strain 12D (12D_P1)
showed little synteny with the other scaﬀolds and appears unique to itself.
The 0.39 Mb 12D_P2 homolog inserted into 12J_C1 is comprised of two distinct portions. The region from 1.56 Mb to 1.72 Mb appears to be abundant with low GC content
genes that encode secretion systems and cytoskeleton. Within this region, we observed two
copies of plasmid partition protein coding gene par B (locus Rpic_1509 and Rpic_1596)
[123]. The section from 1.72 to 1.95 Mb codes for multiple metal resistance genes [191]. This
region was extracted and annotated by hand using BLAST targeting the non-redundant protein database of GenBank. We obtained 408 hits with alignment scores over 200 with diverse
taxonomy indicating that the region is a phylogenetic mosaic of horizontally transferred genes
(Figure 2.6). The query sequence revealed itself as a "mosaic", where multiple genes from
diﬀerent bacteria were assembled to reconstruct the 0.23 Mb region of R. pickettii 12J_C1
(Figure 2.6). As we expected, 12D_P2 shares the most similarities. A fragment of 12J_C1
sequence (between 134000 and 150000) was highly similar (99% identity) to 25 genes on plasmid pMOL30 of C. metallidurans CH34 (0.71 Mb to 0.87 Mb). Finally, the KfrA protein
coding gene is also found in the query sequence, which is 100% identical to those kfr A genes
observed in R. pickettii 12J_C2, 12D_C2, and 12D_P1 (Figure 2.6). Within this region of
R. pickettii 12J_C1, we observed four copies of genes related to cobalt-zinc-cadmium resistance (czcA), one mercury resistance operon and three copper resistance operons. Besides
these speciﬁc metal resistances, this region also contains multiple copies of genes encoding
45

for heavy metal translocating P-type ATPase, ABC transporter, iron permease, cation diﬀusion transporter, and heavy metal transport/detoxiﬁcation protein. Scattered among these
resistance genes, were transposes, tyrosine-based site-speciﬁc recombinase, and hypothetical
protein encoding genes (Figure 2.6).

46

Figure 2.6: BLAST results mapped onto the 1.72 Mb-1.94 Mb region of R. pickettii 12J_C1. The top scale
indicates the length of the query region. Similar genes are shown underneath the query sequence. The short
thin black bars indicate that the next stretch of sequence is on the opposite strand of DNA.

47

Figure 2.6 (cont’d)

48

2.4.4

2.4.4.1

The Acquisition and Duplication of Resistance Determinants in R. pickettii 12D and 12J
Metal Resistant Genes in R. pickettii 12D and 12J

Annotations from the R. pickettii 12D_P2 and the inserted region of 12J_C1 revealed a
high abundance of ABC transporters, P-type ATPase, and metal (Cu, Co-Zn-Cd, and Hg)
resistance determinant encoding genes (Figure 2.6). To determine the number of metal resistant gene copies that reside in R. pickettii 12D and 12J, we employed IMG genome BLAST
to extrapolated possible metal resistant genes speciﬁcally related to Co-Zn-Cd, Cu and Hg
resistance. Co-Zn-Cd (czc operon), Cu (cop operon) and Hg (mer operons) resistant genes
found in C. metallidurans CH34 were used as reference genes. Two cop operon genes, copR
and copS, were not detected in C. metallidurans CH34. Thus, copRS found in Pseudomonas
aeruginosa LESB58 were used as references for identifying copR and copS. The abundance
and distribution of these genes are reported in Table 2.2. We observed multiple occurrence
of czc operon in all four bacterial strains. The major Co-Zn-Cd pump encoding gene, czcA
is the most abundant, which was found 13 times in CH34, 9 times in 12J, 8 times in 12D,
and 4 times in GMI1000. Two diﬀerent sets of czcBCD were located on all four strains. R.
pickettii 12J contains the most copies of czcD (13) and the least amount of czcC (2). R.
pickettii 12J contains the most amount of czcE, a copper binding protein coding gene [195],
and czcP, which is a transition metal eﬄux coding gene [138]. Of the four strains, only C.
metallidurans CH34 and R. pickettll 12J carry czcN while czcI was found only in CH34.
Copper resistant determinants were found in all four strains but were most abundant in R.
pickettii 12D and 12J. C. metallidurans contains two copies of each cop operon gene and

49

only one copy of each cop gene was found in R. solanacearum GMI1000. The major cop
operon component, copA, a Cu(I)-translocating P-type ATPase was observed 5 and 3 times
in 12J and 12D, respectively. The regulatory and signaling protein encoding genes of czc
and cop operon appear to be highly similar and were reported together. A mercury resistant
determinant is found in R. pickettii 12J and C. metallidurans CH34 only. Although the regulatory gene, mer R, was observed in both R. pickettii 12D and R. solanacearum GMI1000,
the gene encoding for mercury reductase (mer A) was absent from both. We identiﬁed two
copies of mer P, a mercuric transporter protein encoding gene, and one copy of mer A in
R. pickettii 12J. In contrast, three copies of mer A and four copies of mer P were found in
C. metallidurans CH34. The distribution of metal resistant genes on R. pickettii 12D and
12J genome revealed that majority of the metal resistant genes are located on 12D_P2 and
12J_C1. The secondary chromosomes of both strains carry the same number of the metal
resistant genes. No Co-Zn-Cd, Cu, or Hg resistant genes were found on R. pickettii 12D_P1
and 12D_P3. The total copy number of each metal resistant gene revealed that strain 12J
contains an equal or great number of copies than 12D, with the exception of czcC (Table 2.2).

50

Table 2.2: The abundance of Co-Zn-Cd, Cu, and Hg resistant genes in R. pickettii 12D and 12J, R. solanacearum GMI1000,
and C. metallidurans CH34.

czcA

P2
12D

C2
12D 12J

P3
12D

P1
12J

P1
12D

czcE
czcP
czcI
czcN

1
0
0
0
0
1
2
0
0
0
0

5
1
4
0
0
7
5a
4
3
0
1

4
3
2
0
1
1
0
1
2
0
0

3
0
1
1
1
0
0
0
0
0
0

3
0
1
1
1
0
0
0
0
0
0

0
0
0
0
0
0
0
0
0
0
0

1
0
0
0
0
0
1
0
0
0
0

0
0
0
0
0
0
0
0
0
0
0

cop Operon

C1
Metal Resistance Genes 12D 12J

copA
copB
copC
copD

0
0
0
0

5
4
4
3

3
4
4
2

0
0
1
1

0
0
1
1

0
0
0
0

0
0
0
0

0
0
0
0

czc/cop

czcR/copR
czcS/copS

0
1

5
6

2
2

0
1

0
1

0
0

0
0

0
0

mer Operon

czc Operon

czcB
czcC
czcD

mer A
0
1
0
0
0
0
0
mer E
0
0
0
0
0
0
0
mer P
0
2
0
0
0
0
0
mer R
3
5
0
2
2
0
0
mer T
0
1
0
0
0
0
0
a. Two out of ﬁve matches are shared by Rmet_5979 matches.
b. IMG genome BLAST with e-value of 1E-2 and at least 40% identity.

51

0
0
0
0
0

R. pickettii
12D

R. pickettii
12J

R. solanacearum
GMI1000

C. metallidurans
CH34

Reference
Locus

czcE
czcP
czcI
czcN

8
3
3
1
2
2
2
1
2
0
0

9
1
5
1
1
7
6
4
3
0
1

4
1
4
1
2
1
1
0
0
0
0

13
4
4
2
1
1
1
3
2
1
1

Rmet_5980
Rmet_5981
Rmet_4121
Rmet_5982
Rmet_4120
Rmet_5979
Rmet_3406
Rmet_5976
Rmet_5970b
Rmet_4595
Rmet_5984

cop Operon

Table 2.2 (cont’d)

copA
copB
copC
copD

3
4
5
3

5
4
5
4

1
1
1
1

2
2
2
2

Rmet_6112c
Rmet_6113
Rmet_6114
Rmet_6115

czc/cop

czcR/copR
czcS/copS

2
4

5
7

2
2

5
12

Rmet_5978/PLES_22631d
Rmet_5977/PLES_22621e

Metal Resistance Genes
czcA

mer Operon

czc Operon

czcB
czcC
czcD

mer A
0
1
mer E
0
0
mer P
0
2
mer R
5
7
mer T
0
1
c. IMG genome BLAST with e-value of 1E-5 and at
d. IMG genome BLAST with e-value of 1E-5 and at
e. IMG genome BLAST with e-value of 1E-2 and at

0
3
Rmet_6174
0
1
Rmet_6176
0
4
Rmet_6173b
5
12
Rmet_6171
0
3
Rmet_6172
least 50% identity. Matches contain all 3 Pfam domains.
least 60% identity. Both reference loci return the same results.
least 40% identity for PLES_22621.

52

2.4.4.2

The Phylogenetic Lineage of czcA Copies in R. pickettii 12D and 12J

Table 2.2 shows that copies of czcA are distributed on 12D_C1 (1), 12D_C2 (3), and
12D_P2 (4) in R. pickettii 12D and 12J_C1 (5), 12J_C2 (3), and 12J_P1 (1) in R. pickettii
12J. To determine their phylogenetic lineage, a collection of 85 czcA sequences from 25
Proteobacteria (Figure A.1) was used to construct a minimal evolution phylogenetic tree
(Figure 2.7). We observed that both R. pickettii 12D and 12J contain czcA genes that share
evolutionary lineages to copies found in C. metallidurans CH34, R. eutropha JMP134, R.
solanacearum Po82 and GMI1000. We observed that of eight copies of czcA genes in 12D, six
copies clustered with those from 12J (Figure 2.7 Group I-VI). The phylogenetic tree showed
that some copies of czcA from the same strain clustered together, such as 12D P2 copy 1
and copy2, and 12J C1 copy 4 and copy 5. One of the R. pickettii 12J czcA copy (12J C1
copy 1) appeared to be unique and did not cluster with the rest of the R. pickettii copies
(Figure 2.7).

53

Ralstonia eutropha H16 C2 (-)
Cupriavidus necator N-1 C2 (-)
82
Cupriavidus taiwanensis str. LMG19424 C2 (-)
Ralstonia solanacearum GMI1000 plasmid pGMI1000MP (+)
100
Ralstonia solanacearum Po82 megaplasmid (+)
41
Ralstonia eutropha JMP134 C2 (-)
100
Cupriavidus metallidurans CH34 C2 (-)
90
67
Cupriavidus metallidurans CH34 plasmid 1 (-)
100 Ralstonia pickettii 12J C1 (-) copy 1
79
Cupriavidus taiwanensis plasmid pRALTA (-)
Pseudomonas aeruginosa PA7 (+)
100
Pseudomonas fluorescens Pf-5 (-)
Pseudomonas syringae pv. tomato str. DC3000 (-)
99 100
Pseudomonas putida F1 (-)
93
Pseudomonas putida F1 (-)
38
Burkholderia xenovorans LB400 C2 (-)
Burkholderia sp. 383 C2 (-)
100
100
100
Burkholderia multivorans ATCC 17616 plasmid pTGL1 (-)
Burkholderia thailandensis E264 C2 (+)
99
Burkholderia mallei ATCC 23344 C2 (-)
100
100
100 Burkholderia pseudomallei 1710b C2 (-)
Cupriavidus necator N-1 plasmid BB1p (+)
Ralstonia pickettii 12D P2 (+) copy 1
Ralstonia pickettii 12D P2 (-) copy 2
100
Cupriavidus metallidurans CH34 plasmid 1 (-)
100
Cupriavidus metallidurans CH34 plasmid 2 (+)
100
100
100

Figure 2.7: Phylogenetic position of R. pickettii 12D and 12J czcA genes. The numbers on the branches
represent bootstrap values.

54

Figure 2.7 (cont’d)
100 Ralstonia pickettii 12J C2 (-) copy 2
Group I
Ralstonia pickettii 12D C2 (-) copy 3
Ralstonia eutropha JMP134 C2 (+)
Cupriavidus metallidurans CH34 C2 (-)
Curpiavidus metallidurans CH34 C1 (-)
Cupriavidus metallidurans CH34 C2 (-)
Burkholderia thailandensis E264 C2 (-)
100
Geobacter uraniumreducens Rf4 (+)
100 Ralstonia pickettii 12J C2 (-) copy 3
Group II
100
Ralstonia pickettii 12D C2 (-) coy 4
Ralstonia solanacearum Po82 megaplasmid (+)
Ralstonia solanacearum GMI1000 plasmid pGMI1000MP (-)
100
Cupriavidus metallidurans CH34 C2 (-)
Ralstonia eutropha H16 C2 (+)
100
100
Cupriavidus necator N-1 C2 (+)
100
Cupriavidus taiwanensis str. LMG19424 C2 (+)
Ralstonia eutropha JMP134 C2 (+)
56
Cupriavidus metallidurans CH34 C2 (+)
Cupriavidus metallidurans CH34 C2 (-)
Ralstonia solanacearum Po82 megaplasmid (-)
100
100 Ralstonia pickettii 12J C1 (+) copy 4
100
Ralstonia pickettii 12J C1 (-) copy 5
Ralstonia pickettii 12J P1 (-) copy 6
100
Group III
100 Ralstonia pickettii 12D C1 (-) copy 5
66
76

67
75

39
100

99

62

100

55

Figure 2.7 (cont’d)

Geobacter uraniumreducens Rf4 (-)
Geobacter lovleyi SZ (-)
Geobacter metallireducens GS-15 (+)
Ralstonia solanacearum GMI1000 plasmid pGMI1000MP (+)
100
Ralstonia solanacearum Po82 megaplasmid (+)
100
44
Ralstonia pickettii 12J C2 (-) copy 7
Group IV
100 Ralstonia pickettii 12D C2 (-) copy 6
100
Cupriavidus metallidurans CH34 C2 (+)
Ralstonia eutropha JMP134 C2 (+)
100
Cupriavidus taiwanensis str. LMG19424 C2 (-)
100
Ralstonia eutropha H16 C2 (-)
100
100 Cupriavidus necator N-1 C2 (-)
Geobacter lovleyi SZ (+)
Geobacter metallireducens GS-15 (-)
100
100
Geobacter uraniumreducens Rf4 (+)
100 Burkholderia mallei ATCC 23344 C1 (+)
Burkholderia pseudomallei 1710b CI (+)
Burkholderia multivorans ATCC 17616 C3 (+)
100
Burkholderia multivorans ATCC 17616 plasmid pTGL1 (+)
60
100

78

94
100

56

Figure 2.7 (cont’d)

Cupriavidus metallidurans CH34 plasmid 1 (+)
Cupriavidus metallidurans CH34 C2 (+)
100 100
Ralstonia eutropha H16 C1 (+)
Cupriavidus necator N-1 C1 (+)
Ralstonia solanacearum GMI1000 plasmid pGMI1000MP (+)
100
95
Ralstonia solanacearum Po82 megaplasmid (+)
100 Ralstonia pickettii 12J C1 (-) copy 8
Group V
98
Ralstonia pickettii 12D P2 (-) copy 7
Ralstonia pickettii 12J C1 (+) copy 9
82
Group VI
100
Ralstonia pickettii 12D P2 (-) copy 8
Pseudomonas putida F1 (+)
Pseudomonas aeruginosa PA7 (+)
100
Escherichia coli APEC O1:K1:H7 plasmid pAPEC-O1-R (+)
Escherichia coli APEC O1:K1:H7 (+)
100
Escherichia coli O157:H7 Sakai EHEC (+)
100
53 Escherichia coli O44:H18:042 EAEC (+)

0.05

57

2.4.4.3

The Phylogenetic Lineage of copA Copies in R. pickettii 12D and 12J

We obtained 24 copA sequences from 25 genomes reported in Figure A.1 and constructed a
minimal evolution tree (Figure 2.8). Fewer copA copies were found and they appeared to
be highly similar to those found in R. solanacearum Po82, C. metallidurans CH34, and γproteobacteria (Pseudomonas and Escherichia groups). We observed three incidences where
copA copies from 12J and 12D clustered together (Figure 2.8 Group I-III). R. pickettii 12J
C1 copy 2 clustered close to Group I and 12J C1 copy 5 appeared to be distant from all of the
other copies (Figure 2.8). R. pickettii 12D_P2 carries three copA copies which were clustered
into two distinct groups (Figure 2.8 12D P2 copy 1/2 and 12D P2 copy 3), suggesting the
occurence of multiple horizontal transfer events.

58

Ralstonia pickettii 12J C1 (+) copy 1
Group I
90
Ralstonia pickettii 12D P2 (+) copy 1
96
Ralstonia pickettii 12J C1 (-) copy 2
Ralstonia solanacearum Po82 megaplasmid (-)
76
Ralstonia solanacearum GMI1000 plasmid pGMI1000MP (+)
Ralstonia pickettii 12J C1 (-) copy 3
Group II
68
100
Ralstonia pickettii 12D P2 (+) copy 2
Cupriavidus metallidurans CH34 plasmid 1 (+)
Cupriavidus metallidurans CH34 C2 (-)
69
Ralstonia eutropha JMP134 C1 (-)
58
Ralstonia eutropha H16 C2 (-)
76
98
Cupriavidus taiwanensis str. LMG19424 C2 (-)
Burkholderia multivorans ATCC 17616 C1 (-)
Geobacter lovleyi SZ (+)
Ralstonia pickettii 12J C1 (+) copy 4
100
Group III
56
Ralstonia pickettii 12D P2 (-) copy 3
Pseudomonas fluorescens Pf-5 (-)
62
Pseudomonas aeruginosa PA7 (+)
99
Pseudomonas syringae pv. tomato str. DC3000 (+)
45
Escherichia coli APEC O1:K1:H7 plasmid pAPEC-O1-R (+)
97
Pseudomonas putida F1 (-)
69
100
Pseudomonas putida F1 (-)
Ralstonia pickettii 12J C1 (+) copy 5
100

0.05

Figure 2.8: Phylogenetic position of R. pickettii 12D and 12J copA genes. The
numbers on the branches represent bootstrap values.

59

2.5

Discussion

Studying the development of microdiversities gives insights on how environmental factors
inﬂuence the evolutionary events and how the speciﬁc adaptation strategies evolved. In this
study, we investigated the genomic diﬀerences of R. pickettii 12D and 12J and identiﬁed
possible evolutionary events that occurred as a result of the selective pressure of metal
contamination. Under the same amount of heavy metal stress, R. pickettii diverged and
evolved into two distinct genomovars. Herein, we report the genomic diﬀerences between the
two strains and the possible evolutionary events behind the divergence.
Interestingly, in what would appear to be a relatively constant environment, the genome
structures of R. pickettii 12D and 12J are drastically diﬀerent (5 scaﬀolds in 12D and 3
scaﬀolds in 12J). However, both strains share large amounts of genetic information (Figure 2.5 and Figure A.3). Figure 2.5 revealed that 12J_C1 is composed of 12D_C1 and
12D_P2, the secondary chromosomes of both strains are nearly identical, and 12J_P1 could
be viewed as 12D_P3 with the addition (or insertion) of 0.028 Mb of 12D_C1. The largest
plasmid 12D_P1 is unique to strain 12D. Thus R. pickettii 12D and 12J in fact share genetic
information across four genome scaﬀolds.
We observed substantial variations in the R. pickettii dif site sequences. Both R. pickettii
12D and 12J utilize the same set of xer CD genes, located on the primary chromosomes, to
resolve dimers created during replications on both the primary and secondary chromosomes.
Carnoy et al. [18] have reported that this is a common phenomenon observed in bacteria
with multiple chromosomes. It has also been shown that a single set of xer CD can bind to
distinct dif sequences [18, 170]. The ME tree of β-Proteobacteria dif sequences suggested

60

that the 12D_C1 dif copy is the closest to the ancestral R. pickettii copy and the other
three copies from 12D and 12J were derived from the dif sequence that is similar to 12J_C1
dif (Figure 2.3).
We have previously reported that live ﬁlamentous phage particles were isolated from
strain 12J and the intracellular replicative form of the virus [191]. This is consistent with
our genome analysis, where an intact copy of the ﬁlamentous phage genome was found on R.
pickettii 12J_C1. Besides the integrated p12J, there are three copies of p12J like sequences
in R. pickettii 12J (two on 12J_C1, and one on 12J_C2). These three copies contain at
least four genes that share syntenies with p12J genes (Figure 2.4). Two out of three vestigial
copies were integrated near 12J dif sites. Huber et al. [63] has shown that ﬁlamentous
phage can be integrated into bacterial chromosomes near dif sites only with the assistance
of dif /xer CD systems. This is consistent with our observation of the integrated ﬁlamentous
phage. Copy 12J-C1_3, which locates distant to the dif site, might be originally integrated
into ancestral R. pickettii and degraded through evolutionary events. Similarly, we located
two vestigial p12J copies in 12D_C1 and 12J_C2. These two copies resemble p12J in
structure but contain only two genes that share homology with p12J and their insertion
positions are distal to 12D dif sites. This observation suggests that copies 12D-C1 and 12DC2 were integrated into ancestral R. pickettii and later mutated (Figure 2.4). Unpublished
work in our lab demonstrated that R. pickettii 12D could not be infected by p12J. One could
speculate that the two vestigial p12J copies in 12D_C1 and 12J_C2 that resemble p12J in
structure might be contributing to the phage resistance.
The analyses of 16S rRNA gene sequences, dif sites variations, and p12J insertions suggest that the secondary chromosome was present in the common ancestor to 12J and 12D.
61

It has been frequently proposed that bacterial secondary chromosomes evolved from plasmids with insertions of house keeping genes [123, 149]. Supporting this hypothesis, we have
found that both secondary chromosomes of R. pickettii 12D and 12J contain the plasmid
stabilization protein encoding gene, kfr A. Cooper et al. [28] also demonstrated that secondary chromosomes evolve faster than primary chromosomes. Surprisingly, the secondary
chromosomes of R. pickettii 12D and 12J appeared to undergo many fewer changes than the
primary chromosomes (Figure 2.5). We do not know the reasons for this apparent discrepancy in Ralstonia.
One obviously major evolutionary event in R. pickettii 12D and 12J was the integration
of 12D_P2 into the ancestral R. pickettii primary chromosome. This inserted region is
0.07 Mb downstream of 12J_C1 dif site. The "mosaic" pattern of the inserted region in
12J_C1 indicates that the ancestral primary chromosome acquired more genes from other
bacteria after the insertion (Figure 2.6). With this in mind, it is logical to assume that
plasmid 12D_P2 transferred into the common ancestor and eventually became inserted into
chromosome 1 of what was to become 12J. This common ancestor also contained 12J_P1
(12D_P3) that has been preserved in the two lineages along with C1 and C2. A 0.03 Mb
portion of 12D_C1, which was missing in 12J_C1, was also found in 12J_P1 (Figure 2.5
C). This suggests that the insertion in 12J_P1 was due to the integration of 12D_P2 into
the ancestral primary chromosome. Hence, the ancestral R. pickettii likely contained two
chromosomes and two plasmids and lacked a virulent copy of the ﬁlamentous phage p12J.
Both R. pickettii 12D and 12J have multiple copies of Cu and Co-Zn-Cd resistant genes
(Table 2.2). The majority of these copies are located either on a plasmid (12D_P2) or the
insertion region that share high synteny to the plasmid (12J_C1). The phylogenetic trees of
62

czcA and copA genes showed that ancestral R. pickettii contained a large number of czcA
and copA genes. Both Figure 2.7 and Figure 2.8 revealed that several groups of czcA (Group
I, II, and VI) and copA (Group I-III) genes grouped together on the same branch of lineage.
These co-existing groups suggest that these gene copies existed in the ancestral R. pickettii
genome. Thus, the divergence of R. pickettii 12D and 12J appeared to be a recent event.
Horizontal gene transfer (HGT) and gene duplication/ampliﬁcation (GDA) played a substantial role in the R. pickettii 12D and 12J adaptation to the high copper environment. We
observed a large number of horizontally transferred foreign genes in both strains (Figure 2.1
and Figure 2.2). R. pickettii 12D P1 is especially abundant with foreign genes from Eukaryotes (Figure 2.1 and Table A.1). Previous studies suggested that HGT and GDA have been
frequently used by other bacterial strains (e.g. C. metallidurans and Staphylococcus sp.) to
acquire or amplify metal and antibiotic resistance genes [21, 98, 104, 135, 146]. Figure 2.5
showed that strain 12J obtained metal resistant genes, such as mer operon and czc operon,
from other β-Proteobacteria. The phylogenetic trees of czcA and copA genes revealed that
many copies of czcA and copA genes share lineages with R. metallidurans CH34, R. eutropha
JMP134, R. solanocerum Po82 and GMI1000 and γ-Proteobacteria (copA only). These close
lineages to other bacteria further conﬁrmed that many of these genes were obtained via HGT.
We also observed copies of czcA genes (12D P2 copy 2, 12J C1 copy 4 and 5, and Group V)
and copA genes (12J C1 copy 2) grouped together with other R. pickettii 12D and 12J czcA
and copA genes (Figure 2.7 and Figure 2.8). This suggests that these copies of genes were
duplicated to confer the metal stress.
We observed that R. pickettii 12D and 12J contain fewer copies of Co-Zn-Cd resistant
genes and more copies of Cu resistant genes compared to C. mtallidurans CH34, the model
63

organism to study the metal resistant determinants (Table 2.2) [67, 71, 105, 111, 160]. This
is interesting because C. metallidurans CH34 was isolated from a Zn rich environment and
R. pickettii 12D and 12J were isolated from a Cu rich environment [80, 171]. It is clearly
shown here that C. metallidurans contains more Co-Zn-Cd resistant genes because it was
isolated from a high Zn environment and R. pickettii 12D and 12J contains more Cu resistant
genes because it was isolated from a high copper environment. This is consistent with the
environmental conditions from which the strains were isolated and presumably reﬂecting the
selection pressures they were subjected to.
We observed that R. pickettii 12D and 12J contain a large number of cop gene copies
and they are more abundant in 12J than 12D (Table 2.2). We found four copies of czcE, a
copper binding protein encoding gene described by Zoropogui et al. in 12J and only three
in 12D [195]. Together, these high copy numbered copper resistant and binding genes could
elucidate the larger amount of copper sequestered by R. pickettii 12D and 12J [191]. The
larger copy numbers of czcE genes in 12J also correlates with the larger amount of copper
sequestered by 12J cells which was previously reported [191].
Passot et al. previously suggested that we are witnessing the emergence of the third
chromosome in 12J [123]. However, our results indicate that 12J actively integrates genes
rather than splitting chromosomes. Interestingly, we noticed that R. pickettii 12D_P1 is
large in size and contains a large amount of horizontally transferred genes and IS elements.
This plasmid is substantially diﬀerent from the other plasmids in 12D and 12J. It contains
kfr A and replication repA gene like the secondary chromosomes. Thus, it resembles a possible
early stage tertiary chromosome. All of this adaptation and evolution evidence suggests that
we are witnessing the co-evolution of two strains and the rising of diversity within the species.
64

The diﬀerent evolutionary strategies R. pickettii 12D and 12J took for adaptation in a heavy
metal environment might be the beginning of speciation.

2.6

Conclusion

In this study, we compare the genomic variations between two R. pickettii strains to reveal
the development of R. pickettii microdiversity in Torch Lake, Houghton, MI. Our results indicated that R. pickettii 12D and 12J acquired metal (eg. copper, cobalt-zinc-cadmium, mercury) resistant genes from other bacteria, such as R. solanocerum and C. metallodurans, via
horizontal gene transfer. Under the pressure of high copper in the environment, both strains
also increased the number of metal resistant gene copies via gene duplication/ampliﬁcation.
We found that a larger number of copper resistant gene copies (12J > 12D) corresponds
to a greater amount of actively sequestered copper (12J > 12D). The comparative genomic
analysis suggests that both strains co-evolved under the pressure of high copper. Gaining
the resistance to ﬁlamentous phage in R. pickettii 12D might be the beginning of the divergence. We predict that the ancestral R. pickettii had one primary chromosome (similar to
12J_C1 minus the insertion from 12D_P2), a secondary chromosome (similar to 12J_C2 or
12D_C2), two plasmids (similar to 12D_P2 and 12D_P3) and a ﬁlamentous phage particle.
After the divergence, strain 12J actively integrated mobile elements containing metal resistant genes and strain 12D maintained original plasmids and acquired another large plasmid
(12D_P1), which was largely composed of horizontally transferred foreign genes. We suspect
that 12D_P1 is undergoing the process of being transformed to a third chromosome. The
diﬀerences in metal-resistant gene copies and genome scaﬀolds suggested the development

65

of R. pickettii microdiversity might be the ﬁrst step in speciation. Our results also showed
that secondary chromosomes of R. pickettii 12D and 12J appeared to be much more stable
than the primary chromosomes during the evolution. This ﬁnding was to our surprise because it is the opposite of the observations of Cooper et al., which suggested that secondary
chromosomes mutate and evolve faster than primary chromosomes [28]. We suspect that
secondary chromosomes are preferably preserved in microdiversity but preferably modiﬁed
across species.

66

CHAPTER 3
RESOURCE-INDUCED BACTERIAL STRUCTURE SHIFTS AND THEIR
INFLUENCE ON COMMUNITY FUNCTIONS

3.1

Abstract

Subsurface uranium contamination poses serious health risks and is costly to remove by
traditional physical-chemical methods. A more cost-eﬀective method has proven to be in
situ immobilization of subsurface uranium via the microbially mediation. The full potential
of this approach would be facilitated by the identiﬁcation of all microbial species, including previously uncultured strains, that play a substantial role in uranium immobilization.
However, in situ studies are inﬂuenced by uncontrolled environmental factors and soil bacterial communities are complex. Thus, we used standard enrichment protocols on bacterial
communities from uranium contaminated sediments to select for strains capable of immobilizing uranium. Previous studies have indicated that many uranium reducing bacteria are
also able to reduce nitrate and iron. Therefore, minimal media supplemented with lactate
in combination with nitrate, ferric citrate, or uranyl acetate were used as the conditions
of enrichment. Culture independent bacterial community analyses showed that bacterial
communities shifted drastically upon enrichment and Firmicutes were observed in all enrichments. Genera Pyschrosinus and Clostridium were particularly abundant in uranyl acetate
enrichments and were commonly detected in both uranium and iron enriched communities.
Despite the drastic structural shifts, we observed a stable functional gene proﬁle in all enrichment communities. Our ﬁndings indicated that phylum Firmicutes might play a substantial

67

role in uranium immobilization. The stable functional gene proﬁles targeting over 44,000
genes of the enriched bacterial communities suggested that the net function of the bacterial
communities that resided in the uranium contaminated environments remained unchanged.

3.2

Introduction

Uranium is a naturally occurring radioactive metal that can be detected in rocks, soil and
streams. Due to weathering, oxidized uranium (U6+ ) is commonly found in the environment. Human beings are exposed to this background uranium at low levels via food and
water (∼0.7-1.1 µg/day) without potential health threats. However, human activities such
as mining and weapon development released elevated amounts of uranium in the environment. Among 1,517 former and current United States Environmental Protection Agency
(USEPA) National Priority List (NPL) sites, at least 54 sites were identiﬁed as uranium
contaminated [167, 168]. The uranium contaminants can permeate into ground water and
become a health threat to human beings. The Oak Ridge National Laboratory (ORNL)
integrated Field Research Center (iFRC) is an example of a site with extensive underground
uranium contamination. The uranium containing wastes penetrated an unlined waste pond
(S-3 area), contaminated the groundwater, and formed secondary contaminant sources with
the soil sediments in subsurface. This large area subsurface contamination site poses a great
challenge the development of cost-eﬀective remediation methods.
Lovley et al. [89, 88] discovered that some microorganisms were able to utilize oxidized
metals (e.g., Fe3+ , Mn4+ ) as electron acceptors to generate energy under anaerobic conditions. It was also shown and proven by Gorby et al. that U6+ could also be used as the

68

electron acceptor during bacterial anaerobic respiration [52]. The reduced uranium, U4+
is less stable (can be readily oxidized to U6+ ) but insoluble in water [14, 43]. Thus, by
decreasing the solubility of uranium (e.g., forming uraninite), uranium could be removed
from ground water. Several studies have also shown that iron and sulfate reducing bacteria have the potential to immobilize uranium via enzymatic activities [97, 102, 148, 186].
Encouraged by these laboratory results, researches on in situ bioremediation have been
performed at several Department of Energy (DOE) uranium contaminated research sites
[16, 17, 64, 77, 187, 188, 190].
Wu et al. conducted in situ uranium immobilization studies at Oak Ridge National
Laboratory (ORNL) integrated Field Research Challenge (iFRC) site S-3 area [187, 188].
Uranium reduction in this area is particularly challenging because of the inhibition caused
by pore water with low pH and high concentrations of nitrate and aluminum [187]. To
overcome the inhibition of uranium reduction, Wu et al. [187] designed a nested circulation
system, where a pH neutral zone (the outer loop) was created and a biostimulation zone (the
inner loop) was established within the pH neutral zone. Within the biostimulation zone, Wu
et al. successfully demonstrated that U6+ was immobilized in its reduced form (U4+ ) [188].
Cardenas et al. investigated the bacterial community shifts accompanied the reduction of
uranium in the biostimulation zone and suggested that sulfate reducing bacteria might play
a substantial role in uranium reduction [16, 17].
In situ bacterial community shifts reﬂecting uranium reduction are particularly informative in the identiﬁcation of novel uranium immobilizers. However, these populations are
usually diﬃcult to isolate and their activities cannot be directly conﬁrmed. In addition,
the in situ community dynamics are strongly inﬂuenced by the ﬁeld environmental factors
69

[35]. For these reasons, a controlled system that leads to cultivation of potential uranium
immobilizers is desirable.
Herein, we report our studies on identifying the potential uranium immobilizing populations by enriching uranium exposed soil sediments in the laboratory. Soil sediments were
collected from ORNL iFRC site, speciﬁcally within the biostimulation zone (FW102-2 and
FW102-3) and pH neutral zone (FW107) of the nested circulation system. Sample FW102-2
and FW102-3 were collected from the same monitoring well but at two depths (13.7 m bgs
and 12.2 bgs, respectively). Cardenas et al. [16] previously reported that soluble uranium
concentrations decreased in monitoring well FW102. No bacterial community studies have
been conducted in the pH neutral zone. All three samples contained a large amount of
uranium (700 ppm in FW107, 1600 ppm in FW102-2, 300 ppm in FW102-3). Thus, we hypothesized that all three samples contain strains capable of immobilizing uranium. Besides
uranium, the ORNL iFRC S-3 area where these samples were collected are also abundant in
nitrate and iron [187]. It is known thermodynamically that anaerobic energy yield ranks as
NO− >Fe3+ >U6+ . Research studies have also shown that uranium reducers are capable of
3
nitrate and iron reduction [2, 163]. Therefore, we enriched soil sediments in minimal media
supplemented with NO− , or Fe3+ , or U6+ . By doing this we expected to identify speciﬁc
3
bacterial community shifts associated with uranium enriching processes.

70

3.3
3.3.1

Materials and Methods
Anaerobic Media and Stocks

Minimal growth media (FWM and FWM-Fe) were modiﬁed based on previously reported
freshwater-medium and ferric-citrate media [87]. The detailed composition of FWM is listed
in Table B.1, Table B.2 and Table B.3. FWM-Fe medium is the same as FWM but supplemented with 56 mM of ferric citrate. Supplement stocks, sodium lactate (2 M) and sodium
nitrate (1 M) solutions were prepared by dissolving chemicals in deionized water to desired
concentrations. Stocks solutions were ﬁlter (0.22 µm) sterilized. Uranyl acetate stock (100
mM) was prepared by dissolving uranyl acetate powder in 30 mM pH7 sodium bicarbonate
solution. Filter sterilization was also applied. All media and stock solutions were bubbled
under N2 :CO2 (80:20) to eliminate oxygen.

3.3.2

Bacterial Community Enrichments

Three uranium contaminated soil samples (FW102-2, FW102-3, and FW107) were collected
anaerobically from the nested recirculation experimental site at Oak Ridge iFRC S-3 area
(denoted as "O" in sample labels). Bacterial consortia from these soil sediments (10%,
v/v) were enriched in FWM [86] supplemented with sodium lactate (20 mM) combined with
either sodium nitrate (10 mM, denoted as "NO3" in sample labels), or ferric citrate (56 mM,
denoted as "Fe" in sample labels) or uranyl acetate (0.25 mM, denoted as "U" in sample
labels) in duplicates at 15 ◦ C. One control culture without the addition of inocula was also
set up for each enrichment condition. The soil sediments were enriched for a total of 35 days
and another 20mM lactate was injected on the 17th day after sampling to provide additional

71

nutrients (after sampling). Two-ml samples were taken from ferric citrate enrichments and
passed through 0.22 µm HPLC certiﬁed syringe ﬁlters (Sun SRI Titan2 17mm Syringe Filters,
Thermo Fisher Scientiﬁc) on the 1st, 17th and 35th day of incubation. Samples were mixed
with HCl to obtain a ﬁnal concentration of 0.5 M HCl and stored at -20◦ C until being
processed. Half of the sample was used to monitor the depletion of lactate and half for
the reduction of Fe3+ . The enrichments were then used as inocula (1% v/v) for secondary
enrichments under the same conditions. Additional sodium lactate was added on the 13th
day. The ferric citrate enrichments were again monitored for 35 days. Bacterial consortia
samples were collected at the end of second enriching period (70th day) for 16S rRNA gene
and community functional gene proﬁling. From each enrichment, two 1 ml samples were
withdrawn and centrifuged at 10,000 RPM for 20 minutes in a Sorvall SS-34 rotor and the
pellet was stored at -20◦ C until extraction of community DNA. The remaining material was
processed using GeoChip [57].

3.3.3

Quantitative Analyses

The concentration of sodium lactate was determined using high-performance liquid chromatography (HPLC) with an Aminex HPX-87H column (polystyrene-divinylbenzene ion
exchange resins, Bio-Rad, Hercules, CA). Sodium lactate and acetate mixtures were used as
standards (0.04 µM - 0.4 µM) and freshly prepared in triplicate prior to sample analyses. A
mixture of acetylnitrile and sulfuric acid (5%) was used as the eluent.
Iron reduction was determined by the previously published ferrozine essay [154]. Brieﬂy,
0.1 ml of ﬁltered acidiﬁed ferric citrate sample was added to 0.9 ml 1 M HCl. This mixture
(0.1 ml) was then added to 1 ml ferrozine buﬀer (a mixture of 50 mM HEPES and 0.1%
72

ferrozine with a ﬁnal pH of 7) and incubated for 10 minutes at room temperature. After
incubation, 200 µl was added to a clear ﬂat-bottom Corning 96-well microtiter plate (Corning
Scientiﬁc, Tewksbury, MA) and absorbance was measured at 520 nm.

3.3.4

DNA Extraction and Community Proﬁling

Community DNA was extracted by using Powersoil DNA Extraction Kit (Mo Bio Laboratory, Carlsbad, CA). Two pyrosequencing chemistries (454 FLX and 454 Titanium) were
used to evaluate the bacterial 16S community structure. For samples in replicate 1, bacterial
16S rRNA gene V4-V5 regions (Forward 5’AYTGGGYDTAAAGNG3’ and Reverse 5’TACNVGGGTATCTAATCC3’) were ampliﬁed as previously described [157]. Samples from replicate 2 were ampliﬁed with primer sets developed by Human Microbiome Project (HMP)
targeting the V3-V5 region (357F 5’CCTACGGGAGGCAGCAG3’ and 926R: 5’CCGTCAATTCMTTTRAGT3’) of 16S rRNA genes. Both polymerase Chain Reactions (PCR)
were carried out by mixing 1X Taq reaction buﬀer, 2 mM MgSO4 , 0.1 mg/ml BSA, 0.2 mM
of each dNTP, 0.33 µM of each primer, 0.125 U/µl of Platinum High Fidelity Taq Polymerase
(Invitrogen, Carlsbad, CA), 10 ng DNA template and DNase and RNase free water. The
ampliﬁcation conditions include an initial denaturing step at 95◦ C for 5 minutes, followed
by 30 cycles of 3 temperature step consisted of denaturing at 95◦ C for 45 seconds, annealing
at 50◦ C for 45 seconds, and elongation at 72◦ C for 90 seconds, and a ﬁnal extension at
72◦ C for 5 minutes. All ampliﬁed 16S rRNA gene PCR products were visualized on 2%
Tris-Borate-EDTA (TBE) agarose gel. DNA fragments at ∼450 bp (V4-V5) and ∼569 bp
(V3-V5) were excised and extracted from the gels by using Qiagen Gel Extraction Kit and
further puriﬁed with QiaQuick PCR Product Puriﬁcation Kit (Qiagen, Valencia, CA). All
73

puriﬁed DNA products were quantitated with Quibit HS Double-stranded DNA Kit (Invitrogen, Carlsbad, CA) then composited at equal mass for a ﬁnal DNA concentration of 0.5
ng/µl. Pyrosequencing was performed at Michigan State University RTSF by using a Roche
454 GSFLX and 454 GSFLX Titanium Sequencers.
Functional gene proﬁling of the enriched consortia was performed at Oklahoma University.
Community DNA were extracted as previously described [172]. GeoChip 3.0 was used to
access the functional gene proﬁles of U6+ and NO− enriched samples in replicate 2. GeoChip
3
5.0 was used to access the functional gene proﬁles of the original soil sediment communities
and the Fe3+ , U6+ and NO− enriched communities in replicate 1 and Fe3+ and NO− enriched
3
3
communities in replicate 2. GeoChip 3.0 and GeoChip 5.0 data were combined and GeoChip
3.0 replicate 2 NO− enrichment data were omitted for analysis purpose.
3

3.3.5

Community Analysis

Bacterial 16S rRNA gene sequences obtained from pyrosequencing were processed and analyzed using the Ribosomal Database Project (RDP) Pyrosequencing pipeline [180]. Sequences that were shorter than 250 bp, containing any N’s in the sequencing region, or with
an average quality score less than 20 were discarded. Chimeras were removed from RDP processed sequences by using USEARCH (UCHIME reference mode with Silva gold alignment
database as reference). RDP complete linkage clustering was used to determine bacterial
operational taxonomic units (OTUs) based on 97% sequence similarity. To eliminate the
clustering artifacts, sequences were trimmed to the same length (144 bp model position)
based on RDP alignment proﬁles prior to clustering. RDP Multi-classiﬁer was used to identify each sequence to bacterial genus level (threshold 80). Any identiﬁed genus that has a
74

total count of 1 across all samples (21 samples) was omitted from analyses. RDP SeqMatch
was used for further identiﬁcation of bacterial clusters of interest.
OTU- and phylotype-based community analyses were performed by using statistical software R [128]. Ecological indices were estimated for each bioreactor sample based on clustered
sequences by using R package Vegan [117]. Speciﬁcally, community richness (ChaoI), diversity indices (Shannon H’) and community evenness (E) were calculated based on equations
previously described [94]. Diversity index Shannon H’ was expressed as magniﬁed H’ (eH )
to emphasize the diﬀerences among samples [94]. The sampling coverage (C) of each sample
was calculated based on Good’s method, C=1-n1 /N, where n1 represents number of OTUs
that was observed once and N is the total number of sequences in the sample [39]. R package
Vegan was also used to perform non-metric multidimensional scaling analysis (NMDS) to
correlate the dissimilarities among bioreactor samples and the shifts in phylotype abundance
[117]. To identify the shared genera among diﬀerent enrichments, genera found in both duplicated enrichments were extrapolated ﬁrst. Venn diagrams of these dereplicated genera
were then constructed by using the R package Vennerable [159].

3.4
3.4.1

Results
Monitoring the Enrichment Cultures

Three soil sediments were collected from locations containing 700 ppm (FW107), 1600 ppm
(FW102-2) and 300 ppm (FW102-3) uranium. Among three electron acceptor supplemented
enrichments, we observed growth in nitrate supplemented enrichments as early as the 5th
day. All enrichments grew equally well. Between the 15th and 30th day, we observed Fe3+

75

reduction in ferric citrate enrichments, where the culture turned dark brown. The change
of colors were ﬁrst observed in FW102-2 and FW102-3 enrichments. FW107 enrichments
changed color by day 30. A large amount of precipitates were formed upon the addition of
uranyl acetate. Hence, it was diﬃcult to diﬀerentiate the growth of the uranium enrichment
cultures from the turbidity due to precipitation. However, we did observe abundant live cells
under the microscope throughout the enrichment process.
The monitored ferric citrate enrichments revealed that iron was actively reduced to Fe2+
while sodium lactate was used as the electron donor and carbon source (Figure 3.1). No
growth was observed in the control microcosms and we did not detect Fe2+ (above background level) in these control cultures. Under the same enrichment condition, we observed
that the FW-107 culture reduced the least amount of Fe3+ (Figure 3.1 A and B) and accumulated the most amount of acetate (Figure 3.1 C and D). Some acetate accumulated in
FW102-2 and FW102-3 ferric citrate enrichments by the 13th day of incubation. However,
this acetate was consumed by the end of the incubation period (Figure 3.1 C and D). Second
generation enrichments were more eﬃcient at iron reduction and lactate utilization than the
ﬁrst generation.

3.4.2

Overview of the Community Dynamics

Bacterial communities of the original iFRC soil sediments and the enrichment cultures were
investigated based on bacterial 16S rRNA sequences. We obtained 703 to 8213 sequences
per sample. Although two diﬀerent sequencing chemistry were used, the ecological indices
indicated that the data obtain between two replicates are comparable (Table 3.1). Pyrosequencing was able to cover at least 97% of the sample community, except for FW1022_O,
76

the original soil sediment sample from well FW102-2, which also had the least number of
sequences (Table 3.1 column C). All of the original soil sediments appeared to be more even
and diverse than the enriched samples, as we expected. We also found that the original community in FW107 contained only 75 OTUs, considerably fewer than were found in FW102-2
and FW102-3. ChaoI, Magniﬁed Simpson (eH ) and evenness (E) indices also showed that
the original community of FW107 was much less diverse and more skewed than the other
two. Uranyl acetate and ferric citrate enrichments were the most skewed and least diverse
samples in both FW102 groups and FW107 group, respectively (Table 3.1).

3.4.3

Shifts in the Enrichment Bacterial Community Structures

A detailed comparison of the bacterial community structures showed substantial shifts among
enrichments (Figure 3.2). The ten most abundant genera summed from all 21 samples and
their distributions in each community were compared side-by-side. Consistent with the
ecological indices, we observed that community structures between replicates were similar,
however substantial diﬀerences were observed among enrichments. FW107 group appeared
to be diﬀerent from the others. FW107 original sample and enrichment samples were highly
dominated by one or two genera. The original soil sediment sample collected from FW107 was
abundant with Sulfuricurvum, which was not found in other samples. Over 80% of sequences
from the FW107 ferric citrate enrichments were unclassiﬁed Veillonellaceae. Psychrosinus
were found dominant in FW107 uranyl acetate enrichment cultures. FW102-2 and FW102-3
bacterial consortia enriched under the same condition appeared to be similar. While all
ten genera constituted less than 10% of the populations in FW102-2 and FW102-3 soil
sediment samples, Sulfuricurvum comprised of over 70% of the FW107 original community
77

Table 3.1: Ecological indices of iFRC enrichment bacterial communities. O represents original soil sediment samples. Fe, NO3, and U supplemented enrichment cultures.

Replicate

Total
Sequences

OTU

ChaoI

eH

E

C (%)

FW1022_O

0

703

146

286.05

53.33

0.80

88.76%

FW1022_Fe

1
2

6623
5891

57
72

91.20
107.00

13.48 0.64
14.86 0.63

FW1022_NO3

1
2

2293
8213

67
65

92.00
89.00

6.49
9.14

0.44
0.53

98.91
99.81

FW1022_U

1
2

1125
5152

16
44

19.75
49.14

1.95
3.81

0.24
0.35

99.47
99.83

FW1023_O

0

6078

203

309.24

26.87

0.62

98.70

FW1023_Fe

1
2

3700
4315

60
50

77.10
71.38

7.68
7.08

0.50
0.50

99.49
99.56

FW1023_NO3

1
2

878
5488

31
70

37.43
101.63

5.76 0.51
9.31 0.53

98.86
99.58

FW1023_U

1
2

6008
6055

53
49

78.67
62.00

4.09
5.38

0.35
0.43

99.63
99.79

FW107_O

0

1410

75

117.00

4.04

0.32

97.45

FW107_Fe

1
2

1664
2956

17
15

22.60
29.00

2.17
1.21

0.27
0.07

99.52
99.73

FW107_NO3

2
1

3596
4102

56
62

123.67
76.25

5.05 0.40
7.64 0.49

99.19
99.54

FW107_U

1
2

771
4679

32
46

36.67
50.20

3.87 0.39
3.40 0.32

98.96
99.85

99.71
99.64

sequences. A large amount of Geobacter were detected in the ferric citrate and uranyl acetate
enrichments. The proportion of these ten genera increased as the enrichment stress went from
nitrate to iron to uranium (Figure 3.2). We also observed that Psychrosinus was found in
all samples (except FW107 ferric citrate enrichments) and its abundance increased as the
enrichment stress increased (Figure 3.2).

78

A.
FW107
FW102-2
FW102-3

15
10
5
0

B.

Day 17
Lactate

Day 35

Acetate

15

FW107

10

FW102-2
FW102-3
Day 1

Day 1

25 D.
20

Lactate (mM)

Fe(II) (µM)

45
40
35
30
25
20
15
10
5
0

20

Day 35

1st Enrichment

C.

5
0

Day 17

Day 35

0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2

0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2

Acetate (mM)

Day 17

Day 1

Acetate

Acetate (mM)

25
Lactate (mM)

Fe(II) (µM)

45
40
35
30
25
20
15
10
5
0

Day 1

Day 17
Lactate

Day 35

FW107
FW102-2
FW102-3
FW107_lac_control FW102-3_lac_control
FW102-2_lac_control

2nd Enrichment

Figure 3.1: The concentration of Fe2+ (panels A and B), lactate and acetate (panels C and D) in each ferric
citrate enrichment cultures. The ﬁrst generation enrichments are presented in panel A and C and the second
generation enrichments are presented in panel B and D. Error bars represent the diﬀerences between two
replicates.
79

FW

FW

10 102
FW 22 2_
10 _NO O	  
2
3
FW 2_N _2	  
10 O3
FW 22 _1	  
10 _Fe
_
FW 22_ 2	  
10 Fe_
FW 22 1	  
10 _U_
22 2	  
FW _U
FW
_
10 102 1	  
FW 23 3_
10 _NO O	  
2
3
FW 3_N _2	  
10 O3
FW 23 _1	  
10 _Fe
_
FW 23_ 2	  
10 Fe_
FW 23 1	  
10 _U_
23 2	  
_U
FW FW _1
10 107 	  
FW 7_ _O
10 NO 	  
7_ 3_2
FW NO 	  
10 3_
FW 7_ 1	  
10 Fe_
2
FW 7_F 	  
10 e_1
FW 7_ 	  
10 U_2
7_ 	  
U_
1	  

Abundance	  (%)	  

100	  
90	  
80	  
70	  
60	  
50	  
40	  
30	  
20	  
10	  
0	  

Psychrosinus	  
Simplicispira	  
unclassiﬁed_Clostridiales	  

Geobacter	  
Pseudomonas	  
Clostridium	  XlVb	  

unclassiﬁed_Veillonellaceae	  
Sulfuricurvum	  
unclassiﬁed_Oxalobacteraceae	  

Figure 3.2: The top ten (summed across all samples) most abundant genera identiﬁed via RDP multiclassiﬁer.
Sample replicates were labeled as "_1" and "_2". Original soil sediment samples were denoted with "_O."

80

FW107_Fe

FW107_U

5

1

4
6

27

0

1
0

FW107_NO3

0

0

0

1
4

2

0
FW107_O

Figure 3.3: The shared genera among FW107 samples. The color of sample font is
the same as the cognate ellipse.

3.4.4

Shared Genera

We identify the bacterial genera that were shared among enrichments. Venn diagrams were
constructed on essential genera that were shared between the replicated cultures of FW107,
FW102-2 and FW102-3 samples. Among 51 genera identiﬁed in the original FW107 sediment
and enriched consortia, over 50% were unique to FW107 and only one (Sulfurospirillum) was
shared by all four samples (Figure 3.3). The unique genus in FW107 iron enrichment was
Trichococcus. FW107 uranium and nitrate enrichment cultures shared 5 genera, out of which,
one was also detected in iron enrichment consortia and the other four were Psychrosinus,
Pseudomonas and two groups of Clostridium. Five genera, unclassiﬁed Ruminococcaceae,
Oscillibacter, Clostridium sensu stricto, Clostridium IV, and Acidovorax were found only in
uranium enriched FW107 cultures. Unclassiﬁed Veillonellaceae was found in uranium, iron

81

FW1022_Fe

FW1022_U

0

10

0
1

47

3

4
3

FW1022_NO3

2

1

1
0

4
4

0
FW1022_O

Figure 3.4: The shared genera among FW102-2 samples. The color of sample font
is the same as the cognate ellipse.

and nitrate enriched consortia.
A large number of unique genera (47) were found in FW102-2 original soil sediment
sample (Figure 3.4). Uranyl acetate enrichment of FW102-2 appeared to be a subset of
FW102-2 ferric citrate enrichment. The four genera shared by ferric citrate, nitrate and
uranium enrichments were Sulfurospirillum, Sporomusa, Psychrosinus, and Geobacter.
FW102-3 samples contained a total number of 107 bacterial genera, 70 of which were
unique in the FW102-3 original soil sediment (Figure 3.5). Psychrosinus was found common
in iron, nitrate and uranium enrichment cultures. Geobacter and unclassiﬁed Veillonellaceae
were shared by all four samples (Figure 3.5). Unclassiﬁed Ruminococcaceae, unclassiﬁed Bacteroidetes, Sporomusa, Desulfosporomusa, Clostridium III were only found in uranyl acetate
enriched FW102-3 consortia. Among FW107, FW102-2 and FW102-3 uranium enrichments,

82

FW1023_Fe

FW1023_U

5

13

2
0

70

2

1
0

FW1023_NO3

5

2

2
0

0
5

0
FW1023_O

Figure 3.5: The shared genera among FW102-3 samples. The color of sample font
is the same as the cognate ellipse.

Sporomusa and Psychrosinus were the two numerically dominant taxa based on sequence
frequency.

3.4.5

Functional Gene Proﬁles

GeoChip proﬁles revealed that there were 1532 functional genes (between diﬀerent GeoChip
platforms) found in diﬀerent microorganisms. Four uranium enriched samples, FW107_U(2),
FW107_U, FW1022_U(2), and FW1023_U(2) appeared to be substantially diﬀerent from
the others (Figure 3.6). These four samples had many fewer abundant functional genes suggesting that uranium selects for a smaller collection of taxa. Three out of four samples were
analyzed by GeoChip3. Although both FW107 uranium enrichment duplicates clustered
close to each other, uranium enrichment duplicate 2 of FW102-2 and FW102-3 appeared

83

to be diﬀerent from duplicate 1 (Figure 3.6). Analysis on GeoChip5.0 enrichment culture
replicate 1 alone showed that the overall variation among all samples was small (Figure B.1).
Thus, we suspect the diﬀerent GeoChip platforms contributed substantially to the diﬀerences
among samples we observed.

84

Color Key

024

FW1022_NO3(2)

FW1023_NO3(2)

FW1023_NO3

FW1023_Fe(2)

FW1022_Fe(2)

FW107_NO3

FW1022_NO3

FW107_NO3(2)

FW107_O

FW1023_U

FW1022_U

FW1023_Fe

FW107_Fe

FW1023_O

FW1022_O

FW107_Fe(2)

FW1022_Fe

FW1023_U(2)

FW1022_U(2)

FW107_U

FW107_U(2)

Row
Z-Score

Figure 3.6: The functional gene proﬁles of the original soil sediment and enrichment samples. The functional
gene signals of all samples (columns) were standardized to the same amount. Each functional gene (row) was
then scaled among all 21 samples to show the diﬀerences in abundance. Green indicates high abundance and
red indicates low abundance.
85

3.4.6

Non-metric multidimensional scaling (NMDS) analysis

Non-metric multidimensional scaling (NMDS) analysis was performed to correlate the βdiversity of enrichment samples and bacterial community or functional gene shifts. The
original soil sediment sample community compositions appeared to be substantially diﬀerent
from the enriched communities (Figure 3.7). This diﬀerence also inﬂuenced the bacterial
genera and functional gene correlations substantially (Figure 3.7 and Figure 3.8). Both
Figure 3.7 and Figure 3.8 showed that the most signiﬁcant shifts were observed between
original soil sediment samples and the enrichments. Signiﬁcant decreases in bacterial genera,
including many unclassiﬁed bacteria were observed in the enriched cultures (Figure 3.7). The
functional gene proﬁle revealed that genes involved in the sulfur cycle (e.g., sox and dsr A)
and lincosamide resistance (linB) decreased signiﬁcantly in enriched cultures.
Within the enrichments we also noticed some signiﬁcant phylogenetic and functional
shifts among bacterial communities due to the enriching conditions (Figure 3.9 and Figure 3.10). In the nitrate enriched communities, unclassiﬁed bacteria (β-Proteobacteria,
Chitinophagaceae, and Comamonadaceae), Dechloromonas, and Simplicispira were found
signiﬁcantly more abundant than the other enrichments. We observed that unclassiﬁed Enterobacteriaceae, Clostridium XIVa, and unclassiﬁed Clostridiales were signiﬁcantly more
abundant in FW102-2 and FW102-3 Fe and U enrichment cultures. Unclassiﬁed Veillonellaceae, Sulfurospirillum and Trichococcus were signiﬁcantly more abundant in FW107 Fe
and U enrichments. The abundance of Geobacter shifted signiﬁcantly towards FW102-2
and FW102-3 U enrichments (Figure 3.9). The abundance of a large number of functional
genes shifted substantially among all enrichments (P ≤ 0.01, data not shown). We observed

86

the abundance of three groups changed most signiﬁcantly (P ≤ 0.001). Group III includes
seven diﬀerent bacterial functional genes, which appeared to be signiﬁcantly more abundant
in FW107 Fe enrichments. The abundance of chromate transporter encoding gene (chr A,
speciﬁcally found in Burkholderia xenovorans LB400) decreased signiﬁcantly from FW107
enrichments to FW102-2 and FW102-3 enrichments (Figure 3.10).

87

0.6

I

unclassified Deltaproteobacteria

0.4

FW107_NO3(2)

0.2

FW107_U

0.0
-0.2

FW1022_NO3

FW1023_O

FW107_NO3
FW1023_NO3
FW107_Fe

FW1023_Fe(2)

I

FW1023_U

II
FW1023_Fe

FW1023_U(2)

FW1022_Fe(2)
FW107_Fe(2)

FW1022_O

III

FW1022_Fe

FW1022_U

FW107_O

FW1022_U(2)

unclassified Cytophagaceae
Subdivision3 genera incertae sedis
unclassified Burkholderiaceae
unclassified Syntrophobacteraceae
unclassified Desulfuromonadales
unclassified Cystobacterineae
unclassified Cystobacteraceae
unclassified Opitutaceae
Gp6
Campylobacter
Arcobacter
Prevotella
Desulfobulbus
Roseomonas

III

-0.6

-0.4

II

FW1023_NO3(2)
FW107_U(2)

NMDS2

FW1022_NO3(2)

unclassified Desulfobulbaceae

-0.5

0.0

0.5
NMDS1

Figure 3.7: NMDS analysis of NO− enrichments (grey), Fe3+ enrichments (pink), U6+ enrichments (yellow),
3
and original samples (green). Each ellipse was drawn around each group centroid based on a group distance
standard error at 95% conﬁdence level. The changes in genera abundances that drive comparative community
diﬀerences are indicated with red arrows. The arrows point to the areas of high abundance. The red arrows
indicate the abundance of genera that changed signiﬁcantly (P ≤ 0.001). The speciﬁc genera are shown in red
according to the group labeling.
88

0.6
0.0
-0.2

II

2)
FW1023_Fe(

FW1023_U
FW1023_Fe

FW1023_U(2)
FW1022_U

II

III FW1023_O
FW1022_O

]

IV

FW1022_Fe

FW1022_Fe(2)
-0.4

sox Silicibacter pomeroyi DSS-3

_N
O3

FW107_U
FW107_NO3

FW1023_NO3
FW107_Fe

I

ohbAB Paracoccus denitrificans PD1222

III
adpB Kineococcus radiotolerans SRS30216

IV
gyrB Legionella pneumophila subsp. fraseri
linB Methylobacterium populi BJ001

FW107_O

FW107_Fe(2)

dsrA uncultured sulfate-reducing bacterium
nbaC Sclerotinia sclerotiorum 1980
linB Bradyrhizobium sp. BTAi1

FW1022_U(2)

-0.6

NMDS2

0.2

FW1023_NO3(2)
FW107_U(2)

I

FW
10
22

0.4

FW107_NO3(2)

FW1022_NO3(2)

-0.5

0.0

0.5

NMDS1
− enrichments
NO3

Figure 3.8: NMDS analysis of
(grey), Fe3+ enrichments (pink), U6+ enrichments (yellow),
and original samples (green). Each ellipse was drawn around each group centroid based on group distance
standard error at 95% conﬁdence level. The changes in functional gene (with the speciﬁc microorganism from
which the probes were designed) abundance were correlated to the shifts observed in community structures.
The red arrows indicate the abundance of functional genes that changed signiﬁcantly (P ≤ 0.001). The speciﬁc
genera are shown in red according to the group labeling. The arrows point to the areas of high abundance.

89

FW107_NO3(2)

IV Simplicispira

0.5

unclassified
Comamonadaceae

]

FW1023_NO3(2)

0.0

FW107_NO3

FW107_U
FW1
023_
Fe
FW
10
23 I
_F
e(
2)

V
VI

e
_F
22
10
FW
_U
23
10
FW
FW1022_U

-0.5

NMDS2

FW1023_NO3

FW107_Fe

FW1023_U(2)

FW107_Fe(2)

III

FW1022_U(2)

-1.0

-0.5

Geobacter

II

FW1022_Fe(2)

0.0

unclassified Enterobacteriaceae

II

Clostridium XlVa

FW1022_NO3

FW107_U(2)

VII

I

FW1022_NO3(2)

III

unclassified Clostridiales

IV

unclassified Betaproteobacteria
unclassified Chitinophagaceae
Dechloromonas

V

unclassified Veillonellaceae

VI

Sulfurospirillum

VII

Trichococcus

0.5

NMDS1
Figure 3.9: NMDS analysis of NO− enrichments (grey), Fe3+ enrichments (pink), and U6+ enrichments (yellow).
3
Each ellipse was drawn around each group centroid based on group distance standard error at 95% conﬁdence
level. The changes in genera abundance were correlated to the shifts observed in community structures. The
colored arrows indicate the abundance of genera that changed signiﬁcantly (red: P ≤ 0.001, blue: P ≤ 0.05,
and green: P ≤ 0.01). The genera names are shown in the same color according to the group labeling. The
arrows point to the areas of high abundance.
90

FW107_NO3(2)

FW1022_NO3(2)

I

FW1023_NO3(2)
0.5

II

I

phtA Comamonas testosteroni
FW1022_NO3

FW107_U(2)

FW107_NO3

FW107_U

0.0

FW107_Fe

III

FW1023_U

FW1023_Fe(2)
FW1023_Fe

FW1023_U(2)
FW107_Fe(2)
FW1022_U

-0.5

NMDS2

FW1023_NO3

FW1022_U(2)

-1.0

-0.5

II
ChrA Burkholderia xenovorans LB400

III

gyrB Legionella jamestowniensis
adpB Geobacillus thermodenitrificans NG80-2
tfdA Aspergillus fumigatus Af293
catB Planctomyces maris DSM 8797
GCoADH Sclerotinia sclerotiorum 1980
merP Idiomarina loihiensis L2TR
CadA Sinorhizobium meliloti 1021

FW1022_Fe

FW1022_Fe(2)

0.0

0.5

NMDS1
− enrichments
NO3

Figure 3.10: NMDS analysis of
(grey), Fe3+ enrichments (pink), and U6+ enrichments (yellow). Each ellipse was drawn around each group centroid based on group distance standard error at 95%
conﬁdence level. The changes in functional gene abundance were correlated to the shifts observed in community
structures. The red arrows indicate the abundance of functional genes that changed signiﬁcantly (P ≤ 0.001).
The functional genes and the speciﬁc microorganisms from which the probes were designed are shown in the
red according to the group labeling on the side. The arrows point to the areas of high abundance.

91

3.5

Discussion

The identiﬁcation of uncultured uranium immobilizers in sub-surface environments is instrumental in the optimization of in situ bioremediation. Our assessment of bacterial communities indicated, not surprisingly, that enrichment was an eﬀective method for identifying
bacterial populations that might play important roles in sub-surface uranium immobilization. We were able to reduce the bacterial diversity and identify common populations under
diﬀerent stressors (ie. iron, nitrate, and uranium) to evaluate their potentials.
From the ferric citrate enrichment cultures, we observed that FW107 enrichments were
not as eﬀective at iron reduction as the FW102-2 and FW102-3 enrichments. Bacterial community analysis revealed that FW107 iron enrichments were predominantly populated by
unclassiﬁed Veillonellaceae (Figure 3.2). Gihring et al. observed that although this group
was abundant in their enrichment cultures, Veillonellaceae did not contribute to uranium
reduction but to initial nutrient degradation in their study [50]. This might explain why
FW107 ferric citrate enrichments appeared to be less eﬀective at iron reduction and the
accumulation of acetate than FW102-2 and FW102-3 ferric citrate enrichments (Figure 3.1).
Thus we suspect that Veillonellaceae did not contribute to uranium immobilization directly
but was responsible for early nutrient consumption. The other two genera, Trichococcus and
Sulfurospirillum, observed in FW107 iron enrichments were reported to be associated with
sulfate and metal reducers [41, 152]. We suspect the iron reduction observed in FW017 ferric
citrate enrichments were mainly contributed by members of Trichococcus and Sulfurospirillum. In contrast to the FW107 iron enrichment culture, FW102-2 and FW102-3 ferric
citrate enrichments appeared to be much more diverse. Genus Geobacter, well-known for

92

its metal-reducing and acetate utilizing members [132], was found abundant in both sets of
consortia. The genus is likely a major contributor to iron reduction and acetate utilization
in FW102-2 and FW102-3 iron enrichments.
The shifts in the bacterial communities revealed that FW102-2 and FW102-3 enrichments
became abundant with Proteobacteriia while Firmacutes became numerically dominant in
FW107 enrichments (Figure 3.2). Venn diagrams revealed that uranium enrichments shared
genera with both iron and nitrate enrichments (Figure 3.3, 3.4, and 3.5). Among all of the
shared genera, Psychrosinus and Sporomusa, genera of phylum Firmicutes, were found in all
uranium enrichment cultures. A member of genus Psychrosinus was previously isolated from
near freezing temperature and identiﬁed as a lactate fermenting bacterium [137]. No evidence
has been shown that members of Psychrosinus are sulfate or metal reducers. However, the
Psychrosinus group was previously identiﬁed as Pelosinus according to RDP (spring 2012
update). We found the 16S rRNA sequences from our Psychrosinus genus were highly similar
to Pelosinus sp. UFO1, which was reported as a uranium reducer by Ray et al. [129]. Genus
Sporomusa has been reported to be found in various U6+ -reducing and sulfate-reducing
environments [7, 12, 113]. However, no speciﬁc members of this group has been identiﬁed as
a metal reducer.
The top ten most abundant genera in all samples revealed that the FW107 original soil
sediment sample was drastically diﬀerent from FW102-2 and FW102-3 (Figure 3.2). This was
expected as well FW107 was in the neutralized zone and was never stimulated [187, 188]. The
most signiﬁcant inﬂuence was the enrichment process. The NMDS analyses revealed that the
most signiﬁcant shifts were decreases in genera abundance (Figure 3.7 and functional gene
abundance (Figure 3.8) from original soil sediment samples to enrichment cultures. Analysis
93

of bacterial community shifts induced by the addition of NO− , Fe3+ and U6+ revealed that
3
bacteria belonging to phylum Firmicutes were signiﬁcantly more abundant in Fe3+ and U6+
consortia (Figure 3.9). The shifts in bacterial functional genes are less conclusive. Figure 3.10
suggests that the original soil sediment samples had a substantial impact on the functional
proﬁle variations. We suspect this result was also due to diﬀerent GeoChip platforms. When
eliminating GeoChip3.0 samples, we were surprised to observe that despite the drastic shifts
revealed by 16S rRNA proﬁles, little variation weas found in their functional gene proﬁle
(Figure B.1). This observation suggested that the underlying functional proﬁle of these
communities were resilient and highly similar. However, it is also possible that important
genes were not included on the GeoChip microarray. Thus, a diﬀerent and more eﬀective
method might be needed to investigate enrichment community function shifts.

3.6

Conclusion

We studied the shifts in bacterial communities, obtained from uranium contaminated soil
sediments (i.e., FW107, FW102-2, FW102-3) and bacterial community shifts induced by
supplemented resources (i.e., NO− , Fe3+ , and U6+ ). The enriched community structures
3
changed drastically according to bacterial 16S rRNA sequences. We observed that Geobacter
was abundant in the FW102-2 and FW102-3 U6+ enrichment cultures, while Psychrosinus
was abundant in the FW107 U6+ enrichment cultures. Psychrosinus was also found common
in all enrichment cultures at various levels of abundance. The functional gene proﬁles of the
enrichments and original soil samples appeared to be highly similar despite the large changes
we observed in 16S rRNA based community structure analysis. Our observations suggest that

94

the function potential of these bacterial community requires did not change as much as its
corresponding phylogeny structure. According to this study, we hypothesize that Firmicutes
play an important role in subsurface uranium immobilization.

95

CHAPTER 4
THE ROLE OF FIRMICUTES IN URANIUM IMMOBILIZATION

4.1

Abstract

The observation of uranium immobilization by anaerobic bacterial communities is a composite result of bacterial interactions. To understand the bacterial interactions and their impact
on uranium immobilization, it is not only important to study the bacterial community structural changes but it is also critical to study individual isolated members of the community. In
uranyl acetate supplemented enrichment cultures seeded with microbial communities from
uranium contaminated sediments, we observed that the concentration of soluble uranium
decreased and the abundance of Firmicutes bacteria increased as a function of time. We
hypothesized that members of Firmicutes play an important role in subsurface uranium
immobilization. We employed standard cultivation techniques under strictly anaerobic conditions and isolated 51 bacterial strains from the uranium enrichment cultures. Bacterial
16S rRNA gene sequences revealed that these isolates belong to four distinct phylogenetic
clades of Firmicutes. However, comparative analysis of growth rates and genome structure
using REP-BOX-GTG5 -PCR proﬁles indicated that these isolates are distinct strains and
might react diﬀerently in the presence of soluble uranium. We screened ten isolates (based
on their 16S rRNA identiﬁcations and REP-BOX-GTG5 -PCR proﬁles) for their abilities to
immobilize uranium and six isolates were more proﬁcient at removing soluble uranium than
the original enrichment community.

96

4.2

Introduction

Bacterial phylum Firmicutes are comprised of a group of ancient and metabolically versatile
anaerobic bacteria [38, 96, 99]. They were frequently found in various environments and
reported in bacterial enrichments. Many members of this group have also been identiﬁed as
sulfate and metal reducing bacteria using carbon sources as electron donors [3, 32, 45, 58,
60, 118]. Members of Firmicutes, such as Clostridia are also well-known for their ability to
produce hydrogen via [FeFe]-hydrogenase (HydA) [10, 103]. In resource scarce subsurface
environments, metabolic hydrogen could be readily used as an electron donor to support
bacterial growth and dissimulatory reduction (eg. Fe, Mn, and etc.) [79, 151, 194].
We have found large numbers of Firmicutes 16S rRNA gene sequences in iFRC (integrated Field Research Challenge site at Oak Ridge National Laboratory, TN) original soil
sediment samples as well as our sequentially enriched microcosms (see Chapter 3). In the
previous chapter, we were able to identify potential genera that might play important roles
in subsurface uranium immobilization. In this chapter we report on the isolation and characterization of strains capable of immobilizing U6+ . We monitored the decrease of soluble
U6+ , the corresponding changes in the U6+ enriched bacterial communities, and isolate potential U6+ immobilizers. We hypothesize that members of Firmicutes play a substantial
role in U6+ removal in iFRC.

97

4.3
4.3.1

Materials and Methods
Uranium Immobilization by FW107 and FW102-2 Enrichment Cultures

Enrichment cultures with uranyl acetate as the terminal electron acceptor (Chapter 3) were
used as a starting culture. To obtain an initial high cell density, they were ﬁrst enriched
anaerobically in 1/2X R2B medium at 15◦ C for 10 days. The 1/2X R2B enrichments were
then centrifuged at 5000 RPM to collect all of the cells. Cell pellets were washed twice in
30 mM anaerobic sodium bicarbonate buﬀer (pH 7. cells were collected via centrifuging in
between) inside of an aerobic chamber (2.5% H2 ). Final washed cell pellets were resuspended
in anaerobic PIPES buﬀered artiﬁcial ground water media (PBAGW) supplemented with 10
mM sodium pyruvate [69] to a ﬁnal OD600 of 0.01. Microcosms were set up in duplicate
and incubated at 15◦ C. Over a period of 64 days, the concentration of soluble U6+ was
determined by a spectrophotometric U6+ -azide assay [145]. All samples were taken from
microcosms anaerobically. Samples for HPLC and U6+ analysis were passed through 0.22
µm HPLC certiﬁed syringe ﬁlters (Sun Titan2, Fisher Scientiﬁc) under anaerobic condition.
Community analysis samples were also taken from one set of the microcosms (replicate 1)
on days 0 (1/2X R2B cultures), 10, 20 and 64 of incubation with uranyl acetate.

4.3.2

Quantitative Analyses

A uranyl-azide assay was modiﬁed from a previously described method [145]. Brieﬂy, HNO3
was added to ﬁltered samples to a ﬁnal concentration of 4.9%. Acidiﬁed samples (40 µl) were
then added to sodium azide solution (3M, 160 µl) in a clear bottom 96-well microtiter plate
(Corning Scientiﬁc, Tewksbury, MA). Absorbance measurements were performed at 340 nm.

98

To correlate the absorbance and U6+ concentration, a standard curve was constructed (in
duplicate) for each measurement (Figure C.1). This assay was found to be valid for soluble
U6+ ranging from 5 µM to 1000 µM.

4.3.3

DNA Extraction and Sequecing Preparation

Community DNA was extracted by using Powersoil DNA Extraction Kit (Mo Bio Laboratory, Carlsbad, CA). Pyrosequencing chemistries (454 FLX) were used to evaluate the
bacterial 16S community structures. Bacterial 16S rRNA gene V4-V5 regions (Forward
5’AYTGGGYDTAAAGNG3’ and Reverse 5’TACNVGGGTATCTAATCC3’) were ampliﬁed
as previously described with 454 FLX chemistry [157]. Polymerase Chain Reactions (PCR)
were carried out by mixing 1X Taq reaction buﬀer, 2 mM MgSO4 , 0.1 mg/ml BSA, 0.2 mM
of each dNTP, 0.33 µM of each primer, 0.125 U/µl of Platinum High Fidelity Taq Polymerase
(Invitrogen, Carlsbad, CA), 10 ng DNA template and DNase and RNase free water. The
ampliﬁcation conditions include an initial denaturing step at 95◦ C for 5 minutes, followed by
30 cycles of 3 temperature steps consisting of denaturing at 95◦ C for 45 seconds, annealing
at 50◦ C for 45 seconds, and an elongation at 72◦ C for 90 seconds, and a ﬁnal extension at
72◦ C for 5 minutes following the completion of 30 cycles. Ampliﬁed 16S rRNA gene PCR
products were visualized on 2% Tris-Borate-EDTA (TBE) agarose gel. DNA fragments at
∼450 bp were excised and extracted from the gels by using Qiagen Gel Extraction Kit. The
extracted products were further puriﬁed by using a QiaQuick PCR Product Puriﬁcation Kit
(Qiagen, Valencia, CA). All ﬁnal puriﬁed DNA products were quantitated by using Qubit
HS Double-stranded DNA Kit (Invitrogen, Carlsbad, CA) prior to combining at equal mass
for a ﬁnal DNA concentration of 0.5 ng/µl. Pyrosequencing was performed at Michigan
99

State University RTSF by using Roche 454 GSFLX Sequencer.
FeFe-hydrogenase (HydA) genes were ampliﬁed as described previously [11] and cloned
with Topo Cloning (Invitrogen, Carlsbad, CA) and sequenced at the MSU-RTSF. Brieﬂy, the
primers used were FeFe-272F (5’GCHGAYMTBACHATWATGGARGA3’) and FeFe-427R
(5’GCNGCYTCCATDACDCCDCCNGT3’) [11]. Each HydA PCR consists of a mixture of
1X Taq reaction buﬀer, DNase and RNase free water, 2 mM MgCl2 , 0.1 mg/ml BSA, 0.2
mM of each dNTP, 2 µM of each primer, 0.05 U/µl of Platinum Taq Polymerase (Invitrogen,
Carlsbad, CA), and 20 ng DNA template. The ampliﬁcation conditions include an initial
denaturing step at 94◦ C for 4 minutes, followed by 35 cycles of 3 temperature steps consisting
of denaturing at 94◦ C for 60 seconds, annealing at 56.5◦ C for 60 seconds, and an elongation
at 72◦ C for 90 seconds, and a ﬁnal extension at 72◦ C for 20 minutes following the completion
of 35 cycles. All PCR products were stored at -20◦ C until needed. HydA PCR products
(∼500 bp) were conﬁrmed via electrophoresis (1% Tris-Acetate-EDTA agarose gel). Clone
libraries of FeFe-hydrogenase were constructed by using Invitrogen Topo PCR2.1 Cloing
Kit (Invitrogen, Carlsbad, CA) following the vendor’s protocol. The cloned fragments were
puriﬁed and sequenced at RTSF via an ABI 3730 Genetic Analyzer.

4.3.4

Community Analysis

Bacterial 16S rRNA gene sequences obtained from pyrosequencing were processed and analyzed by using Ribosomal Database Project (RDP) Pyrosequencing pipeline [180]. Sequences
that were shorter than 250 bp, containing any N’s in sequencing region, or with an average
quality score less than 20 were discarded. Chimeras were removed from RDP processed sequences by using USEARCH (UCHIME reference mode with Silva gold alignment database
100

as reference). RDP Multi-classiﬁer was used to identify each sequence to bacterial genus
level (threshold 80). RDP SeqMatch was used for further identiﬁcation of bacterial clusters
of interest. To eliminate the clustering artifacts, sequences were also trimmed to the same
length (250 bp) based on alignment proﬁles prior to RDP complete linkage clustering.
[FeFe]-hydrogenase gene sequences were identiﬁed by BLAST against NCBI non-redundant
amino acid database. The BLAST E-value (expect value) was set at E < 10−5 . Top 50 best
matches were generated for each sequence. The annotated match with the best E-value was
then selected to be used to determine bacterial species information.

4.3.5

Isolation of Potential Uranium Immobilizing Bacteria

Potential uranium immobilizers were isolated from FW107 and FW102-3 uranyl acetate
enrichments on R2A agar plates in a Coy Lab anaerobic chamber ((H2 :CO2 :N2 = 3:10:87).
A volume of 100 µl enrichment culture from each microcosm was initially spread on R2A
plates anaerobically. Individual colonies were then picked and restreaked for isolation. Final
isolates were stored in 7.5% DMSO at -80◦ C.

4.3.6

Growth Curves of Isolates

The growth of each isolate under anaerobic conditions (H2 :CO2 :N2 = 3:10:87) was determined by measuring turbidities at 620 nm wavelength (OD620). Isolates were grown in 1/2X
R2B broth in 96 well microtiter plates for 63 hours at 25◦ C. To determine if any isolates can
utilize U6+ as an electron acceptor at 25◦ C, isolates were grown for 200 hours in minimal
media (30 mM NaHCO3 , pH 6.8) supplemented with 20 mM sodium pyruvate as a carbon
source and electron donor and 500 µM uranyl acetate. Isolates were also grown in minimal

101

media (30 mM NaHCO3 , pH 6.8) supplemented with 20 mM sodium pyruvate but without
addition of an electron acceptor to determine if these strains were capable of fermentation.
The turbidity measurements were performed under the same anaerobic conditions by using
a Bio-Rad PR 3100 TSC Microplate Reader (Bio-Rad, Hercules, CA)

4.3.7

Isolate Identiﬁcations

Final isolates were identiﬁed via 16S rRNA sequencing, which was carried out by the MSU
RTSF. Isolate cultures were supplied in 96 well microtiter plate and MSU RTSF carried
out 16S rRNA gene ampliﬁcation and sequencing (primer 27F, 5’AGAGTTTGATCCTGGCTCAG3’ was employed). REP-BOX-(GTG)5 -PCR was also used to diﬀerentiate isolates that are substantially diﬀerent at genomic level [49]. We used four primers, REP1RI (5’IIIICGICGICATCIGGC3’), REP2-I (5’IIICGNCGNCATCNGGC3’), GTG5 (5’GTGGTGGTGGTGGTG3’), and BOXA1R (5’CTACGGCAAGGCGACGCTGACG3’) for each
reaction and set up PCR reactions as previously described [174]. Brieﬂy, each 15 µl PCR
reaction consisted of 3 µl 5X Gitschier buﬀer (Table C.1), 1.5 µl 10% DMSO, 30 pmol of each
primers except BOXA1R (50 pmol), 0.75 µl of dNTP (25 mM of each), 0.85 mg/ml BSA,
0.24 µl of 5U/mul DNA Taq polymerase, 100 ng DNA template and water. To standardize
the REP-BOX-(GTG)5 -PCR proﬁle, Desulﬁtobacterium hafniense DCB-2 was also included
in each batch of reactions. Gel electrophoresis was conducted with 2% TAE (tris-acetateEDTA) agarose gels and a 72 Volt electric potential. Final REP-BOX-(GTG)5 -PCR ﬁles
were analyzed using Kodak Molecular Imaging Software.

102

4.3.8

Cell Suspension Assay

To evaluate the uranium immobilization abilities of the isolates, we selected nine strains
based on their REP-BOX-(GTG)5 -PCR proﬁles (presence and absence of DNA fragments).
All of these strains were isolated from FW107 uranium enriched culture. Cell suspension
assays [27, 88] were performed on these nine isolates along with the FW107 enrichment
culture and a false positive control (sterile media without the inoculation of cells). Brieﬂy,
individual isolates (-80◦ C freezer stocks) were inoculated into anaerobic 1/2X R2B broth
and grown to late exponential phase (OD600 ∼ 0.2). Cells were then washed with 30 mM
bicarbonate buﬀer twice (centrifugations at 13,000 g for 15 minutes were carried out at the
end of each wash). Final clean cell pellets were suspended in the reaction buﬀer (30 mM
sodium bicarbonate at pH 6.8, 20 mM sodium pyruvate, 1 mM uranyl acetate) to OD600
0.1. Suspended cells were incubated at 25◦ C for 70 hours in anaerobic chamber (H2 :CO2 :N2
= 3:10:87)

4.4
4.4.1

Results
Soluble U6+ Immobilization by FW107 and FW102-2 Enrichments

Over a period of 64 days, we observed that both FW107 and FW102-2 enrichments were able
to reduce the soluble U6+ concentrations from over 250 µM to under 150 µM (Figure 4.1).
The U6+ concentration variations at day 0 were likely due to the inoculation method. An
excess amount of uranyl acetate in the needle was injected, which was supposed to be retained. A slight increase in U6+ concentration was also observed during the early stage of
incubation (0-10 days) in both enrichment microcosms and the control culture. This could be

103

due to interactions between media and uranyl acetate. All four enrichment cultures yielded
a similar U6+ removal trend (Figure 4.1).

4.4.2

Bacterial Community Compositions of FW107 and FW102-2 Enrichments

Pyrosequencing results showed that genera Clostridium and Pelosinus were abundant in the
FW107 uranium enrichment culture. Together, they constituted over 80% of the community. As time increased, we observed an increasing amount of unclassiﬁed Clostridiales and
a decreasing amount of Clostridium (Figure 4.2). Unlike FW107, the FW102-2 uranium
enrichment culture was dominated by Tolumonas, which was initially found over 90% abundant. The relative abundance of this genus decreased as a function of time. Meanwhile, we
observed that Pelosinus (currently identiﬁed as Psychrosinus by RDP) became more abundant (from approximately 0 to almost 15%) as the incubation time increased (Figure 4.3).
The uranium enriched FW102-2 consortia initially (at time 0) contained few copies of
[FeFe]-hydrogenase genes initially. With the same amount of DNA template, the hyd A
amplicon fragment was barely visible (data not shown). Fewer hyd A clones were derived
from FW102-2 enrichments than from FW107 enrichments at time 0 (8 clones VS. 66 clones)
(Figure 4.4). The diﬃculty in [FeFe]-hydrogenase gene ampliﬁcation and cloning could be
due to either the lack of a hyd A containing population or experimental procedure bias. The
abundance of hyd A genes detected increased signiﬁcantly (to 50) after enriching in uranyl
acetate (Figure 4.4 FW102-2 U64days). While the changes in hydA clones obtained were
not quantitatively proven, it is likely due to the increase of [FeFe]-hydrogenase containing
bacteria after incubation.
A large number of hyd A sequences in the FW102-2 uranium consortia were found in
104

uncultured bacteria. In contrast, hyd A genes identiﬁed in both FW107 1/2X R2B enrichment
and the uranium enriched consortia belonged to two large genera groups, Clostridium and
Desulfotomaculum. We also observed that hyd A sequences from these two groups became
more abundant after the uranium enrichment process (Figure 4.4 FW107).

105

Figure 4.1: The removal of soluble U6+ in FW107 and FW102-2 enrichments. The high soluble U(VI) concentration in the control culture (No Cells) was due to the excess amount of uranyl acetate carried over from the
syringe needle during inoculation.

106

3.5
3
2.5
2
1.5
1
0.5

0

Pelosinus
Clostridium

Minor Groups

10

20

30

40

Time (days)

Sulfurospirillum
unclassiﬁed_
"Ruminococcaceae"
Trichococcus
Ralstonia
Desulﬁtobacterium

50

60

0

Minor Group Percentage

Pelonsinus/Clostridium Percentage

80
70
60
50
40
30
20
10
0

unclassiﬁed_Clostridiales
unclassiﬁed_Veillonellaceae
Herbaspirillum
Desulfosporosinus
Succinispira

Figure 4.2: Bacterial community shifts with the increasing of incubation time in FW107 U-PBAGW microcosm.

107

Pelosinus
Tolumonas
Minor Groups

16
14
12
10
8
6
4
2
0

0

10

20

30

40

Time (days)

Herbaspirillum
Clostridium
unclassiﬁed
Gammaproteobacteria
Sulfurospirillum

50

8
7
6
5
4
3
2
1
0
60

Minor	Group
	Percentage

Pelosinus Percentage

Tolumonas Percentage

100
90
80
70
60
50
40
30
20
10
0

unclassiﬁed_Clostridiales
Desulfosporosinus
unclassiﬁed_"Lachnospiraceae"
Microvirgula
unclassiﬁed_Enterobacteriaceae

Figure 4.3: Bacterial community shifts with the increase of incubation time in FW102-2 U-PBAGW microcosm.

108

Hyd Clone
OTU

8
6

50
11

66
8

66
12

FW102-2

Hyd Clone
OTU

FW107

U 0 days

U 64 days

Figure 4.4: Distribution of [FeFe]-hydrogenase containing bacteria in FW107 and
FW102-2 grown in 1/2X R2B media (U 0 days) and in uranyl acetate enrichments
(U 64 days). The bacterial populations represented by the small fractions of each
pie chart are listed on the right hand side with color blocks corresponding to the
colors found in each pie chart. The distribution of genera Clostridium (cyan color)
and Desulfotomaculum (purple color) are shown as the outer ring of each pie chart.

109

Figure 4.4 (cont’d)

110

4.4.3

Potential U6+ Immobilizers

We have isolated 51 anaerobic bacteria from FW107 and FW102-2 uranium enrichments. We
were able to identify 48 of them according to their 16S rRNA sequences. All of these isolates
belonged to phylum Firmicutes. RDP SeqMatch identiﬁed these isolates as Clostridium
XI, Clostridium XIVa, Clostridium sensu stricto, and Pelosinus. Phylogenetic analysis of
the isolates 16S rRNA sequences by Arb agreed with RDP SeqMatch and revealed that
29 isolates clustered closest to Clostridium sardiniense (Figure 4.5 Cluster I), 13 isolates
clustered closest to uncultured bacteria that are similar to Clostridium methylopentosum
(Figure 4.5 Cluster II), and 4 isolates clustered closest to Pelosinus fermentans (Figure 4.5
Cluster III). Two isolates, A04 and B03 did not belong to any of the clusters but appeared
more closely related to Cluster I than the other groups (Figure 4.5). Although the 16S rRNA
sequences indicated that majority of the isolates were highly similar, we observed substantial
diﬀerences in their growth rates at 25◦ C in anaerobic 1/2X R2B (Figure 4.6). Thus the 16S
rRNA sequence alone is not suﬃcient to diﬀerentiate these isolates. We employed REPBOX-(GTG)5 -PCR and the proﬁles revealed that isolates identiﬁed as the same bacterial
species according to their 16S rRNA sequences were substantially diﬀerent in their genome
organization based on REP-BOX-(GTG)5 proﬁles. These proﬁles were clustered and are
presented in Figure 4.7.

111

Figure 4.5: The phylogenetic distribution of the anaerobic isolates based on their 16S rRNA gene sequences.
The numbers on the branches represent the bootstrap values (1000 permutations).

112

Figure 4.5 (cont’d)

113

Figure 4.6: The growth curves of FRC isolates in 1/2X R2B under anaerobic condition (H2 :CO2 :N2 3:10:87).
See next page for legends.

114

Figure 4.6 (cont’d)

A1_big
E1
C12_small
B9_small
A12_big
A3_small
E3_big
D3_big
C10_small
B11

A5
E7
D7_big
C1_2
B3_(small)
A8
neg
D11
D5
C5

A11_small
A1_small
E2
C7_small
B9_big
A12_2_small
A3_big
E5
D11_2
C10_big

B1_2_small
A6
E9
C12_big
C2
B5_small
A9
A4_small
E6_small
D6

B8_small
A11_big
Blank
D8_small
C7_big
B9_2_small
A12_2_big
A10_small
neg2
D12

115

B12
B1_2_big
A2_small
E3_small
D2
C3
B5_big
B1_small
A4_big
E6_big

C6_small
B8_big
A7
E10_small
D8_big
C8
B9_2_big
B6
A10_big
nge3

C11
C1
A12_small
A2_big
E3_median
D3_small
C4_small
B10
B1_big

D7_small
C6_big
B2
A8_small
E10_big
D9
C9
C4_big
B7

D2
D7
A10
B12
E10
A4
A9
B11
B8
E6
D3
E7
A1
A3
B5_2
C4
C10
C11
Dhaf_32.50
Dhaf_1.8
Dhaf_17.32
Dhaf_51.68
Dhaf_1.16
Dhaf_redo
B9
E3
C12_A
D6
A2
B10
D8
B9_2
A12
C1
C2
E1
B6
E9
D12
A11
B3
A5
C8
D9
D11_2
B7
C5
D5
D11
C9
C3
C8_2
B2
B1
A8
C6
A7

4

3

2
Distance

1

0

Figure 4.7: REP-BOX-(GTG)5 -PCR proﬁles of potential uranium immobilizing bacteria. Branches labeled Dhaf represent control samples. The rest represents enrichment isolates.

116

4.4.4

U6+ Immobilization By Isolates

We observed that some isolates were able to grow anaerobically in minimal media supplemented with uranyl acetate and sodium pyruvate, while little or no growth was observed
in minimal media supplemented with sodium pyruvate alone (Figure C.2 B9 and D5, and
Figure C.3 A1, C9 and E2). We also observed isolates grew better in sodium pyruvate
media instead of uranyl acetate and sodium pyruvate media (Figure C.2 B10 and C1, and
Figure C.3 A10 and C1). However, due to the long incubation time in 96 well plates, we
could not obtain reliable results (Figure C.2 and Figure C.3). Therefore we employed REPBOX-(GTG)5 proﬁles and 16S rRNA identiﬁcations to dereplicate our isolate collection.
Within each bacteria group that was identiﬁed by 16S rRNA gene sequences, the isolates
were subgrouped based on their REP-BOX-(GTG)5 -PCR proﬁles. From FW107 consortia,
nine strains were selected based on their unique REP-BOX-(GTG)5 -PCR proﬁles and 16S
rRNA sequences combinations to examine their ability of anaerobic uranium immobilization. These nine isolates are shown in Figure 4.8 as A4 (Figure 4.5 A04), A5 (Figure 4.5
Cluster III), A8 (Figure 4.5 Cluster I), A9 (Figure 4.5 Cluster I), A11 (Figure 4.5 Cluster I),
A12 (Figure 4.5 Cluster I), B1 (Figure 4.5 Cluster I), B8 (Figure 4.5 Cluster II), and B10
(Figure 4.5 Cluster III). Sample FW107 represent the FW107 uranium enrichment culture.
The cell suspension assays were performed on these ten cultures to determine the amount
of soluble uranium removed. After 70 hours of incubation, we observed that isolates A11,
A12, B1 and B8 removed the largest amount of U6+ , some uranium removal was observed
in A8 and A9, A4 and A5 showed a comparable amount of removal to FW107, and little to
no uranium removal was observed in B10. (Figure 4.8).

117

1200

Soluble U(VI) (µM)

1000
800
T0_avg

600

T70_avg
400
200
0
neg

A4

A5

A8

A9

A11

A12

B1

B8

B10 FW107

Figure 4.8: Soluble U6+ removal by Firmicutes isolates (from FW107) and FW107 uranium enriched community after 70 hours (T70) of incubation at room temperature in an anaerobic chamber (H2 :CO2 :N2 3:10:87).
Sample "neg" represents a control culture without the addition of any bacterial cells. The error bars represent
measurements of two replicates.

118

4.5

Discussion

In this part of our study, we were able to monitor the reduction of soluble U6+ in FW107
and FW102-2 uranium enrichments. Consistent with our ﬁndings in studies performed in
Chapter 3, we observed that the abundance of Pelosinus (currently identiﬁed as Psychrosynus
by RDP) increased along with the enrichment time (Figure 4.2 and 4.3). In addition,
unclassiﬁed Clostridiales become more abundant in both enrichments during the same period,
while soluble U6+ decreased (Figure 4.1). Thus, it is likely that members of Pelosinus and
unclassiﬁed Clostridiales contributed to the U6+ removal processes.
[FeFe]-hydrogenase distributions in the enrichment cultures suggested that Clostridium
and Desulfotomaculum might be contributing to metabolic hydrogen production. However,
some members of these two groups are known uranium reducers [45, 72], their predominant
presence in the [FeFe]-hydrogenase gene communities might also be a simple statement of
their uranium reduction dependent growth from U6+ removal. The combined sequence data
from the 16S rRNA genes and [FeFe]-hydrogenases indicates that the Clostridia are numerically important members of the community with possible uranium reduction responsibilities,
as well as participants in the hydrogen economy of the bioreactors.
Although RDP SeqMatch and Arb phylogenetic analysis were consistent with each other
and showed that the isolates grouped into four major Firmicute categories, RDP SeqMatch
showed diﬀerent species identiﬁcations. Among 48 Firmicute bacterial isolates, RDP SeqMatch revealed that the majority of the sequences in Clostridium XIVa group were identical
to the 16S rRNA gene sequence of Desulfotomaculum guttoideum, which is a rarely studied
sulfate-reducing bacterium [51]. At 95% to 97% similarity, sequences in Clostridium sensu

119

stricto group were closely related to Clostridium acetobutylicum, Clostridium butyricum,
and Clostridium puniceum. While Clostridium acetobutylicum was recently reported as a
uranium reducing bacterium [32], the other two were mostly studied for their chemical producing ability, such as butanol and 1,3-propandiol [22, 62]. The Pelosinus isolates we have
identiﬁed were closely related to Pelosinus sp. strain UFO1, which was recently reported for
its role multi-mode uranium immobilization [129]. Out of 48 isolates, only one was identiﬁed
as Clostridium venationis in group Clostridium XI and no studies on anaerobic respiration
or metal immobilization have been reported on this species.
The above taxonomy information and literature reports suggest that studying Firmicutes
has a great potential in understanding bacterial community activities in U6+ removal. The
growth patterns and REP-BOX-(GTG)5 -PCR proﬁles suggested that these 48 isolates should
not be treated simply as four groups. Instead, it is important to test isolates based on their
REP-BOX-(GTG)5 -PCR proﬁles for their abilities to immobilize U6+ . Out of the ten isolates
we selected, four removed substantially more soluble uranium than the others, including the
FW107 uranium enriched community. It was to our surprise that the Pelosinus isolate A5
was not proﬁcient at uranium removal despite its high abundance in the enrichment cultures.
It was even more surprising to observe the drastic diﬀerences in uranium removal abilities
between strains (e.g., Cluster I A9 VS. Cluster I A11 in Figure 4.8). Considering the high
variabilities among isolates of the same species, it is very likely that we did not obtain a
uranium removal eﬃcient Pelosinus strain during the isolation processes.

120

4.6

Conclusion

We previously described our observations on the prominence of Firmicute bacteria in uranium supplemented enrichment cultures (Chapter 3). In this work, we studied the shifts
of bacterial communities during uranium immobilization. We have anaerobically isolated
and identiﬁed 48 bacterial strains, which belong to phylum Firmicutes. These isolates were
categorized into four diﬀerent species according to their 16S rRNA gene sequences. However,
their growth patterns and REP-BOX-(GTG)5 -PCR proﬁles indicated that they are physiologically and genomely distinct within each species. We evaluated the ability of uranium
immobilization of isolate representatives. Our results indicated that many isolates with close
phylogenetic aﬃliation to Clostridium sardiniense could immobilize uranium.

121

CHAPTER 5
PHYLOGENETIC COMPOSITION OF BACTERIA IN ANAEROBIC
CO-DIGESTION OF DAIRY MANURE AND CORN STOVER

5.1

Abstract

More than 160 million dry tons of animal waste are produced annually from animal farms
in the United States. These wastes pose a great environmental and health risk if not disposed of properly. Anaerobic digestion is an eﬀective method in breaking down animal
wastes to reduce the risks and it produces methane as an alternative energy source. Previous studies suggested that optimization of feed composition, HRT, and other operational
conditions can greatly improve total solids removal and increase methane productivity. Despite these improvements, the relationships between system performance and the underlying
bacterial community shifts are not well understood. Therefore, we constructed anaerobic
bioreactors fed with dairy manure (DS) and dairy manure supplemented with corn stover
(DM+CS) to compare the bacterial community structures and identify important bacterial populations correlating to total solids removal and methane production. Our results
suggested that unique uncultivated strains of Bacteroidetes played a substantial role in cellulose/hemicellulose degradation in the DM+CS bioreactors. Aminomonas paucivorans and
Clostria populations utilize secondary metabolites produced during cellulose/hemicellulose
degradation to generate hydrogen and acetate. Thus, promoting the growth of unclassiﬁed
Bacteroidetes, Clostridia, and A. paucivorans might optimize the eﬃciency of co-digestion
system and control the variability.

122

5.2

Introduction

Agriculture is a signiﬁcant part of the US economy. The Food and Agriculture Organization
of the United Nations (FAO) estimated that approximately $0.2 trillion of production was
contributed from agriculture in the US in 2010. The top three commodities based on production values were indigenous cattle meat, dairy products and maize. Maize was the most
produced commodity based on quantity (at least three times more than other commodities).
This large agricultural output resulted in considerable amounts of waste being produced.
Approximately 6.7 million tons of agriculture wastes were produced in 2010 in the US (estimated based on 1993 data) [44]. The wastes were mainly used as fertilizer for crop growth.
However, applying agricultural wastes, especially manures, to ﬁelds raises signiﬁcant environmental concerns of greenhouse gas emissions, spread of animal pathogens, and release of
excess nutrients to the environment, all of which may pose a great risk to human health and
the ecosystem [20, 24, 100, 112]. Hence, a controlled animal waste management strategy
that simultaneously treats the wastes and utilizes them for fuel/chemical production would
have both environmental and economic beneﬁts.
Anaerobic digestion can eﬀectively degrade farm wastes and concomitantly produce
methane, thereby reducing the health and environmental risks while producing energy [59].
Methane is produced via methanogenesis in anaerobic digestion. Methanogenesis is a syntrophic process that takes place under anaerobic conditions when organic matter is in high
concentration [140]. It is a highly cooperative process carried out by diﬀerent groups of
bacteria and archaea with distinct responsibilities [13]. Schink [140] schematically showed
that methanogenic degradation of organic matter was executed by three groups of microor-

123

ganisms. During anaerobic methanogenesis, organic substrates are degraded by fermentative
bacteria. Fermentation by-product hydrogen is used by hydrogen-oxidizing methanogens to
produce methane, while acetate-cleaving methanogens produce methane from acetate. Previous research has shown that the substrate C:N ratio inﬂuences the production of methane
in anaerobic digestion [46]. Several studies have demonstrated that a mixture of nutrient-rich
manures and carbohydrate-rich crop and food wastes are suitable substrates for anaerobic
digestion and support methanogenesis [30, 42, 46, 85, 110, 179]. While a large number of
studies were focused on optimizing system operational parameters, the microbial communities within the anaerobic digestion have not been not well explored. As pointed out by Schink
et al. [140], these microbial communities are the primary force behind methane production
and community shifts could eﬀect overall performance. Herein we report on a pilot study
describing microbial communities of anaerobic bioreactors fed with dairy manure (DM) and
corn stover (CS) using gene targeted metagenomics.

5.3
5.3.1

Materials and Methods
Anaerobic Bioreactors

Feedstock composition and the laboratory-scale anaerobic bioreactors (5 L working volume
anaerobic fermenters) were set up and described by Yue et al. [192]. Brieﬂy, six continuously
stirred tank reactors (CSTR) were operated at 35◦ C with hydraulic retention times (HRTs)
of 30 days, 40 days, and 50 days. The feed for diary manure bioreactors was 5% dry matter
and 95% water. The feed for manure and corn stover (DM+CS) bioreactors was a mixture
of dairy manure and corn stover in a ratio of 4:1 (wt:wt) and diluted to 5% DM using water.

124

Both feeds were homogenized using a blender prior to feeding. Analytical analyses of biogas,
total solids (TS), and ﬁber composition were reported by Yue et al. [192].

5.3.2

DNA Extraction and Targeted Gene Ampliﬁcations

Anaerobic bioreactors were fed with DM or DM+CS at three diﬀerent HRTs. Samples were
taken from each bioreactor once digestion reached the stabilized phase (after two HRTs). The
samples were centrifuged at 10,000 RPM for 20 minutes in a Sorvall SS-34 rotor and the pellet
was stored at -20◦ C until extraction of community DNA. DNA was extracted by using Powersoil DNA Extraction Kit (Mo Bio Laboratory, Carlsbad, CA). Polymerase Chain Reactions
(PCR) were carried out to amplify bacterial 16S rRNA gene (16S) and FeFe-hydrogenase
genes (HydA) with universal primer sets: 357F (5’CCTACGGGAGGCAGCAG3’) and 926R
(5’CCGTCAATTCMTTTRAGT3’) for 16S, and FeFe-272F (5’GCHGAYMTBACHATWAT
GGARGA3’) and FeFe-427R (5’GCNGCYTCCATDACDCCDCCNGT3’) for HydA [11].
Bacterial 16S rRNA genes were ampliﬁed with sequence-tagged primers developed by the
Human Microbiome Project (HMP) targeting the V3-V5 region of 16S rRNA genes. Brieﬂy,
1X Taq reaction buﬀer, 2 mM MgSO4 , 0.1 mg/ml BSA, 0.2 mM of each dNTP, 0.33 µM of
each primer, 0.125 U/µl of Platinum High Fidelity Taq Polymerase (Invitrogen, Carlsbad,
CA), and 10 ng DNA template were mixed with DNase and RNase free water for each
PCR reaction. The ampliﬁcation conditions include an initial denaturing step at 95◦ C for
5 minutes, followed by 30 cycles of 3 temperature steps consisting of denaturing at 95C for
45 seconds, annealing at 50◦ C for 45 seconds, and elongation at 72◦ C for 90 seconds, and a
ﬁnal extension at 72C for 5 minutes following the completion of 30 cycles. All PCR products
were stored at -20◦ C until sequencing at the MSU Research Technologies Support Facilities
125

(MSU-RTSF).
FeFe-hydrogenase genes were ampliﬁed as described previously [11] and cloned with Topo
Cloning (Invitrogen, Carlsbad, CA) and sequenced at the MSU-RTSF. Each HydA PCR consists of a mixture of 1X Taq reaction buﬀer, DNase and RNase free water, 2 mM MgCl, 0.1
mg/ml BSA, 0.2 mM of each dNTP, 2 µM of each primer, 0.05 U/µl of Platinum Taq Polymerase (Invitrogen, Carlsbad, CA), and 20 ng DNA template. The ampliﬁcation conditions
include an initial denaturing step at 94◦ C for 4 minutes, followed by 35 cycles of 3 temperature steps consisting of denaturing at 94◦ C for 60 seconds, annealing at 56.5◦ C for 60
seconds, and elongation at 72◦ C for 90 seconds, and a ﬁnal extension at 72◦ C for 20 minutes
following the completion of 35 cycles. All PCR products were stored at -20◦ C until needed.

5.3.3

Preparation for Pyrosequencing and Construction of Clone Libraries

The 16S rRNA gene PCR products were visualized on 2% Tris-Borate-EDTA (TBE) agarose
gel. DNA fragments at ∼569 bp were excised and extracted from the gels by using Qiagen Gel
Extraction Kit and further puriﬁed with QiaQuick PCR Product Puriﬁcation Kit (Qiagen,
Valencia, CA). All puriﬁed DNA products were quantitated with Quibit HS Double-stranded
DNA Kit (Invitrogen, Carlsbad, CA) then composited at equal mass for a ﬁnal DNA concentration of 0.5 ng/µl. Pyrosequencing was performed at Michigan State University RTSF
by using a Roche 454 GSFLX Titanium Sequencer.
HydA encoding gene sequence PCR products (∼500 bp) were conﬁrmed via electrophoresis (1% Tris-Acetate-EDTA agarose gel). Clone libraries of FeFe-hydrogenase were constructed by using Invitrogen Topo PCR2.1 Cloning Kit (Invitrogen, Carlsbad, CA) following
the vendor’s protocol. The cloned fragments were puriﬁed and sequenced at RTSF via an
126

ABI 3730 Genetic Analyzer.

5.3.4

Sequence and Community Analyses

Bacterial 16S rRNA gene sequences obtained from pyrosequencing were processed and analyzed by using the Ribosomal Database Project (RDP) Pyrosequencing pipeline [26]. Sequences that were shorter than 250 bp, containing any N’s in sequencing region, or with an
average quality score less than 20 were discarded. Chimeras were removed from RDP processed sequences by using USEARCH (UCHIME reference mode with Silva gold alignment
database as reference). RDP Multi-classiﬁer was used to identify each sequence to bacterial
genus level (threshold 80). RDP SeqMatch was used for further identiﬁcation of bacterial
clusters of interest. To eliminate the clustering artifacts, sequences were also trimmed to the
same length (250 bp) based on alignment proﬁles prior to complete linkage clustering at the
RDP.
[FeFe]-hydrogenase gene sequences were identiﬁed by BLAST against the NCBI nonredundant amino acid database. The top 50 best matches were generated for each sequence.
The annotated match with the best expected value (E < 10−5 ) was selected in determining
bacterial species information.
OTU- and phylotype-based community analyses were performed by using statistical software R [128]. Ecological indices were estimated for each bioreactor sample based on clustered
sequences by using the R package Vegan [117]. Speciﬁcally, community richness (chaoI), diversity indices (Shannon H’ and Simpson’s D) and community evenness (E) were calculated
based on equations previously described [94]. The sampling coverage (C) of each sample was
calculated based on Good’s method, C=1-n1 /N, where n1 represents the number of OTUs
127

that is a singleton and N is the total number of sequences in the sample [39]. The R package
Vegan was also used to perform non-metric multidimensional scaling analysis (NMDS) to
correlate the dissimilarities among bioreactor samples and the shifts in phylotype abundance
[117]. Shared OTUs among samples were calculated and plotted by using the R package
Vennerable [159].

5.4
5.4.1

Results
The Inﬂuence of HRT and Corn Stover on System Performance

For bioreactors fed with DM, a substantial increase on biogas productivity was observed
once the HRT increased from 30 to 40 days (Figure 5.1). The biogas productivity of the
DM bioreactors increased from 209 to 399 ml/g TS loading/day. A similar increase was
also observed in the DM+CS bioreactors (from 270 to 497 ml/g TS loading/day). Further
increases in HRT had no signiﬁcant (P > 0.05) inﬂuence on the biogas productivity for either
feeds. The TS reduction for both feeds also increased with the increases in HRT (Figure 5.1).
The HS reduction increased from 27% (HRT of 30 days) to 50% (HRT of 50 days) in the DM
bioreactors. A similar trend was observed in the DM+CS bioreactors (from 35% at a HRT
of 30 days to 53% at a HRT of 50 days). In addition, digestion performance demonstrated
that under the same HRT, both biogas productivity and TS reduction were signiﬁcantly (P
< 0.05) higher in the DM+CS bioreactors than in the DM bioreactors, presumably due to
the introduction of fast-hydrolyzing components from corn stover in DM+CS feed [192].

128

Figure 5.1: The total solids (TS) reduction eﬃciency (diagonal and grid lined columns) and the production
rate of biogas (grey and white columns) in the DM and DM+CS bioreactors at four HRTs: a HRT of 30 days
(HRT30d), a HRT of 40 days (HRT40d) and a HRT of 50 days (HRT50d). Solid diamonds and solid triangles
represent the amount of methane produced in the DM and DM+CS bioreactors, respectively.

129

The trend of biogas productivity and TS reduction indicated that bioreactor performance
underwent signiﬁcant (P < 0.05) changes from 30 days of HRT to 40 days or longer of HRT.
The biogas productivities for both DM and DM+CS bioreactors were maximized at 40 days
of HRT. Longer HRTs did not increase the biogas productivity but signiﬁcantly more TS
was removed (P < 0.05). In spite of diﬀerent biogas production rates and TS reduction
levels from diﬀerent bioreactors, the methane content of biogas from all six bioreactors fell
within the narrow range from 65.7% to 67.4%, and this was relatively stable throughout the
experiment. The pH measurements indicated that all six bioreactors were maintained at
near neutral condition (pH 6.58-6.94).

5.4.2

Phylogenetic Comparisons of Anaerobic Bioreactor Bacterial Communities

Bacterial 16S rRNA gene targeted sequencing revealed that we sampled over 98% of each
anaerobic bioreactor community (Table D.1 and Figure D.2). The evenness of the bacterial
communities were relatively stable (0.68 to 0.78) but bacterial diversity shifted drastically
from sample to sample. All diversity indices indicated that the DM+CS bioreactor with a
HRT of 40 days had the most diverse bacterial community (Table D.1).

130

40

50
45
40
35
30
25
20
15
10
5
0

A1.

35
30
25
20
15
10
5
0
HRT30d

HRT40d

Number of OTU's

Bacteroidetes Abundance (%)

45

HRT50d

Figure 5.2: The abundance and diversity (OTU at 97% sequence similarity) of phyla Bacteroidetes (panel A1), class Clostridia
(panel B1), and Escherichia / Shigella (panel C1) in each bioreactor bacterial community. The distribution of OTUs among
samples are presented as Venn diagrams in panel A2 (Bacteroidetes), B2 (Clostridia) and C2 (Escherichia/Shigella). The
colors of the bioreactor/HRT font corresponds to an area bounded by the same color. Labels HRT30d, HRT40d and HRT50d
represent bioreactors with a HRT of 30, 40, and 50 days, respectively.

131

Figure 5.2 (cont’d)

A2.
7
0

0

1
0

0
2

0

0

0

0

0

1 1

DM HRT40d
0

0

0

DM+CS HRT50d
0

DM+CS HRT30d
0

0
0
0

0

9

0
0

2

0 1

0

0

2

0

3
1

0 0
2

0

0

1

1

0

0
0

0

DM HRT30d 11

0
1

0

0

0

0
0
10

0 0
1

6

1 12
2
5

DM HRT50d

Bacteroidetes Venn Diagram

132

DM+CS
HRT40d

30
25

B1.

DM
DMCS

120
100

DM_OTU
20

DMCS_OTU 80

15

60

10

40

5

20

0

0

HRT30d

HRT40d

133

HRT50d

Number of OTU's

Clostridia Abundance (%)

Figure 5.2 (cont’d)

Figure 5.2 (cont’d)

B2.
15
0

0

3

1
1

1

0
2

2

2 9

0

1

DM+CS HRT30d

DM HRT40d
1
0

0

DM+CS HRT50d
0

0

4

2

1

1

0
1

3

3

1 0

1

2

0

1

1
3

2 1
6

5

0

2

1
0

1

0

4 4

1

DM HRT30d

7

1
4

0
1

0

0
11

1
18

3 34
5
8

DM HRT50d

46

Clostridia Venn Diagram

134

DM+CS
HRT40d

30
Abundance (%)

Escherichia/Shigella

35

9
8

C1.

7

25

6

20

5

15

4
3

10

2

5

1

0

0

HRT30d

HRT40d

135

HRT50d

Number of OTU's

Figure 5.2 (cont’d)

Figure 5.2 (cont’d)

C2.
1
0

0

0

0
0

0

0
0

0

1

0 0

DM HRT40d
1
1

0

DM+CS HRT50d

0

0

0

0

0

1

0

DM+CS HRT30d

0
1

2

0

0

0 0

0

0

0

0

0

0

0

0 0

0 0

0 0

0

DM HRT30d

0

0

0

0

0

0 0

0
0

0

2

0

0
0
1

0

DM HRT50d

Escherichia/Shigella
Venn Diagram

0
0

0

DM+CS
HRT40d

Bacterial taxa were determined based on 16S rRNA genes via RDP Multi-Classiﬁer with
a minimal bootstrap value of 80 (assigned as unclassiﬁed if score is smaller than 80). A total
of 16 bacterial phyla were assigned and approximately 5% of total sequence was categorized

136

as unclassiﬁed bacteria (3-7% unclassiﬁed bacteria in each sample). At the genus level, a
total of 198 bacterial groups (137 classiﬁed and 61 unclassiﬁed) were identiﬁed. Firmicutes,
Bacteroidetes, and Proteobacteria were the most abundant phyla in all six anaerobic bioreactor samples (Figure D.2 wide columns). Within these three phyla, Clostridia, unclassiﬁed
Bacteroidetes, and Escherichia/Shigella were found to be highly abundant (Fgiure D.2 thin
columns). We observed that Clostridia comprised 74-84% of the Firmicutes in the bioreactors. Phylum Bacteroidetes was largely composed of unclassiﬁed Bacteroidetes, except in the
DM a HRT of 30 days bioreactor that had abundant Alkaliﬂexus, Petrimonas, unclassiﬁed
Porphyromoadaceae and Flavobacterium. Although the abundance of phylum Bacteroidetes
was similar across the six bioreactors, the abundance of unclassiﬁed Bacteroidetes increased
as the HRT increased. In the DM+CS bioreactor with a HRT of 40 days, 35.45% of the
community populations were identiﬁed as unclassiﬁed Bacteroidetes, representing 87% of
the Bacteroidetes populations. Within phylum Proteobacteria, we observed that genera
Escherichia/Shigella were consistently abundant and its abundance variation reﬂected the
abundance of Proteobacteria except in the DM+CS bioreactor with a HRT of 40 days. Thus,
the population of Escherichia/Shigella increased as a function of HRT and on average constituted 17% of the bacterial community except in the DM+CS bioreactor with a HRT of 40
days, where it decreased to 1.53%. Over 50% of the phylum Proteobacteria in the DM+CS
bioreactors with a HRT of 30 days, the DM bioreactors with a HRT of 40 days and the
DM+CS bioreactors with a HRT of 50 days were identiﬁed as Escherichia/Shigella, while
37.71% and 21.21% of Proteobacteria were classiﬁed as Escherichia/Shigella in the DM bioreactor with a HRT of 30 days and DM+CS bioreactor with a HRT of 40 days, respectively
(Figure D.2). Besides the changes in abundance, we also observed the diversity shifts within
137

phylum Bacteroidetes, class Clostridia, and genera Escherichia/Shigella (Figure 5.2). When
operated at the HRTs of 30 days and 50 days, phylum Bacteroidetes was more abundant
in DM bioreactors than in DM+CS bioreactors. Phylum Bacteroidetes also appeared to be
less abundant in both DM and DM+CS bioreactors under the HRT of 50 days than when
they were operated with a HRT of 30 days. At a HRT of 40 days, the abundance of phylum
Bacteroidetes was substantially larger in DM+CS bioreactors than DM bioreactors (40.38%
in DM+CS a HRT of 40 days VS. 21.47% in DM a HRT of 40 days). The Bacteroidetes
diversity (OTU) pattern followed its abundance trend (Figure 5.2 A1). A larger numbers
of OTUs were observed in DM bioreactors than in DM+CS bioreactors at HRTs of 30 and
50 days. The number of OTUs decreased in DM and DM+CS bioreactors slightly as HRT
increased from 30 to 50 days. The lowest Bacteroidetes diversity was observed in the DM
bioreactor at a HRT of 40 days (25 OTUs). In contrast, the DM+CS bioreactor at a HRT
of 40 days revealed the largest Bacteroidetes diversity (43 OTUs). A total nine OTUs were
shared by all six samples. The proportion of unique OTUs in each sample ranged from 19%
in the DM+CS bioreactor at a HRT of 50 days bioreactor to 32% in DM a HRT of 30 days
and DM a HRT of 50 days bioreactors (Figure 5.2 A2). Similar to phylum Bacteroidetes,
the abundance of class Clostridia decreased when HRT increased from 30 days to 50 days
(Figure 5.2 B1). When operated at HRTs of 30 and 50 days, class Clostridia was more
abundant in DM bioreactors than DM+CS bioreactors. The abundances of class Clostridia
were similar in the DM (25.5%) and DM+CS (26.6%) bioreactors operated at a HRT of
40 days. The diversity within class Clostridia followed the same trend as observed in the
abundance proﬁles at HRTs of 30 and 50 days. At a HRT of 40 days, class Clostridia was
substantially more diverse in the DM+CS bioreactor than in the DM bioreactor. With an
138

abundance diﬀerence of 1.1% at a HRT of 40 days, the DM+CS bioreactor was observed
to have 105 OTUs of Clostridia, while only 75 Clostridia OTUs were detected in the DM
bioreactor (Figure 5.2 B1). There were four Clostrdia OTUs shared by all six samples. A
large portion of the OTUs were unique to each sample, ranging from 20% in DM a HRT of 50
days and DM+CS a HRT of 50 days bioreactors to 42% in DM a HRT of 30 days bioreactor
(Figure 5.2 B2). Unlike phylum Bacteroidetes and class Clostridia, the abundance of genera
Escherichia/Shigella increased in the bioreactors as HRT increased from 30 to 50 days (Figure 5.2 C1). It was also observed that the genera Escherichia/Shigella was more abundant
in the DM+CS bioreactors than in the DM bioreactors when operational HRTs were 30 and
50 days. Escherichia/Shigella were the least abundant (1.53%) in the DM+CS bioreactor
operated at HRT 40 days. In contrast, the DM bioreactor operated at a HRT of 40 days
was substantially more abundant with Escherichia/Shigella (19.55%). The diversity proﬁle
revealed that when the abundance of Escherichia/Shigella increased from 5.82% (the DM
bioreactor with a HRT of 30 days) to 19.55% (the DM bioreactor a HRT of 40 days), the
number of OTUs stayed the same (6). However, the Escherichia/Shigella OTUs decreased
from 8 to 3 in DM+CS bioreactors when the Escherichia/Shigella abundance decreased substantially at a HRT of 40 days. Finally, the number of Escherichia/Shigella OTUs remained
the same (from 3 to 4) when their abundance increased from 1.53% (the DM+CS bioreactor
with a HRT of 40 days) to 29.56% (the DM+CS bioreactor with a HRT of 50 days) (Figure 5.2 C1). Among these OTUs, two were found in all six samples. Two of six OTUs were
unique to the DM bioreactor with a HRT of 30 days, and one of eight was unique to the
DM+CS bioreactor with a HRT of 30 days and the DM bioreactor with a HRT of 50 days.
No unique Escherichia/Shigella OTUs were found in the DM bioreactor with a HRT of 40
139

days, the DM+CS bioreactor with a HRT of 40 days and the DM+CS bioreactor with a
HRT of 50 days (Figure 5.2 C2).

5.4.3

Bacteroidetes

Bacteroidetes was one of the most abundant phyla in the bioreactor samples described above.
The genera abundance revealed that unclassiﬁed Bacteroidetes was also predominant and
highly variable (Appendix C Figure C.2). Sequences identiﬁed as members of phylum Bacteroidetes by RDP Classiﬁer were extracted and separated into groups of unclassiﬁed Bacteroidetes and classiﬁed Bacteroidetes (all taxon groups that could be identiﬁed further than
phylum Bacteroidetes). The abundance of phylum Bacteroidetes in bacterial communities
varied from 21% to approximately 40% (Figure 5.3 vertical striped area). We observed that
the composition of phylum Bacteroidetes shifted from classiﬁed Bacteroidetes to unclassiﬁed
Bacteroidetes. Across six bioreactor samples, unclassiﬁed Bacteroidetes comprised as much
as 88% of the phylum Bacteroidetes (the DM+CS bioreactor a HRT of 40 days). The least
amount of unclassiﬁed Bacteroidetes was observed in the DM bioreactor with a HRT of 30
days, where only 15% of Bacteroidetes were unable to be identiﬁed further than the phylum level. The composition of phylum Bacteroidetes revealed that unclassiﬁed Bacteroidetes
were more abundant than classiﬁed Bacteroidetes in the DM+CS bioreactors regardless of
HRT (Figure 5.3). However, in the DM bioreactor samples, classiﬁed Bacteroidetes were
substantially more abundant than unclassiﬁed Bacteroidetes at HRT 30 days (the DM bioreactor at a HRT of 30 days) and became much less abundant in the DM bioreactor with a
HRT of 50 days. At the same retention time, unclassiﬁed Bacteroidetes were more abundant in the DM+CS bioreactors than in the DM bioreactors and vice versa for classiﬁed
140

Bacteroidetes. It was also observed that unclassiﬁed Bacteroidetes increased in abundance
as the HRT increased in both the DM and DM+CS bioreactors, while the abundance of
classiﬁed Bacteroidetes decreased (Figure 5.3).
At 97% similarity, sequences in phylum Bacteroidetes were clustered into diﬀerent groups.
It was observed that phylum Bacteroidetes was abundant (40%) and diverse (over 40 OTU)
in the DM+CS bioreactor operated at a HRT of 40 days (Figure 5.2A). In contrast, Bacteroidetes was the least abundant and diverse (approximately 25 OTU) in DM bioreactor
operated at a HRT of 40 days. The Bacteroidetes clusters (at 97% similiarity) in the DM+CS
bioreactor with a HRT of 40 days were compared to the Bacteroidetes clusters in other samples and four distinct clusters (#11, #14, #43, and #69) were identiﬁed (Figure 5.4). These
four clusters were abundant in the DM+CS bioreactor with a HRT of 40 days where together
they comprised of 62% of the Bacteroidetes populations in the DM+CS bioreactor with a
HRT of 40 days. These clusters were among the unclassiﬁed Bacteroidetes. Cluster #69
appeared unique to the DM+CS bioreactor with a HRT of 40 days as it was not detected in
other samples (the DM bioreactor with a HRT of 30 days, the DM bioreactor with a HRT of
40 days, the DM bioreactor with a HRT of 50 days and the DM+CS bioreactor with a HRT
of 50 days) or detected with a 0.3% abundance (the DM+CS bioreactor with a HRT of 30
days). Cluster #11 and #14 were observed in the DM+CS bioreactor samples (the DM+CS
bioreactor with a HRT of 30 days, the DM+CS bioreactor with a HRT of 40 days and the
DM+CS bioreactor with a HRT of 50 days) at high abundance (6-27% and 15-29% of the
phylum Bacteroidetes in each sample, respectively). These two clusters were also observed
in the DM bioreactor samples (the DM bioreactor with a HRT of 30 days, the DM bioreactor
with a HRT of 40 days and the DM bioreactor with a HRT of 50 days) but at a substantially
141

Abundance of Bacteroidetes (%)
Classified Bacteroidetes
100

80

60

40

20

Unclassified Bacteroidetes
0

20

40

60

80

100

DM

HRT
30d

DM+CS
DM

HRT
40d

DM+CS
DM

HRT
50d

DM+CS
0
20 40 60 80 100
Bacteroidetes Abundance
in Communities (%)

Total

DM

DM+CS

Figure 5.3: The abundance of phyla Bacteroidetes (vertical stripes) in each bioreactor bacterial communities.
The distribution of classiﬁed and unclassiﬁed Bacteroidetes within phylum Bacteroidetes were calculated and
displayed in left panel (white and diagonal striped bars) and right panel (black and gridded bars) respectively.
Labels HRT30d, HRT40d and HRT50d represent bioreactors with a HRT of 30, 40, and 50 days, respectively.

142

lower abundance (1-7% and 0.4-2% of the phylum Bacteroidetes in each sample, respectively)
but became more abundant as the HRT increased. Thus, the abundance of #11 and #14
could be ranked as the DM bioreactor with a HRT of 30 days<the Dm bioreactor with a
HRT of 40 days<the DM bioreactor with a HRT of 50 days and the DM+CS bioreactor with
a HRT of 30 days<the DM+CS bioreactor with a HRT of 40 days<the DM+CS bioreactor
with a HRT of 50 days. Over 11% of the Bacteroidetes abundance was contributed by cluster
#43. This cluster was present in the DM+CS bioreactors operated at HRT greater than
40d (the DM+CS bioreactors at HRT of 40 and 50 days). All three populations were highly
similar to uncultured Bacteroidetes and no isolates with similar 16S rRNA sequences could
be found (RDP Sequence Match).

5.4.4

Hydrogen Producing Population Proﬁles

We obtained 45 to 76 HydA gene clones (total 352 HydA gene clones). The coverage based
on Good’s method ranged from 72%-94% with a median of 87%. A total of 49 hydrogenase
OTUs were detected of which 28 were aﬃliated with class Clostridia. Three major taxonomic
aﬃliations (phylum Synergistetes, class Clostridia ad phylum Bacteriodetes) were identiﬁed
that accounted for 64-86% of the detected hydrogenases sequences.
The abundance of Synergistetes-like [FeFe]-hydrogenase genes increased from less than
15% to over 30% in the DM bioreactors and remained over 35% in the DM+CS bioreactors
as the HRT increased. The largest abundance of Synergistetes (42.86%) was observed in
the DM+CS bioreactor operated at a HRT of 40 days. The abundance of Clostridia-like
[FeFe]-hydrogenase genes decreased in the DM bioreactors (40.68% to 20%) and increased
in the DM+CS bioreactors (18.42% to 36.92%) as the operational HRT increased. Phylum
143

Figure 5.4: The comparisons of Bacteroidetes clusters (97% sequence similarity)
between the DM+CS bioreactor with a HRT of 40 days and other bioreactor samples.
Labels HRT30d, HRT40d and HRT50d represent bioreactors with a HRT of 30, 40,
and 50 days, respectively. Symbol ∗ denote peaks that are clusters of unclassiﬁed
Bacteroidetes that are highly abundant in all DM+CS bioreators (#11 and #14),
in DM+CS bioreactors at long HRTs (#43), and in the DM+CS bioreactor with a
HRT of 40 days (#69).

144

Bacteroidetes-like [FeFe]-hydrogenase genes appeared to be more abundant in the DM bioreactors than in the DM+CS bioreactors. The overall abundance of Bacteroidetes decreased
as the HRT increased (Figure 5.5).
Detailed abundance analysis at the species level revealed that Aminomonas paucivorans
(main contributors of Synergistetes-like [FeFe]-hydrogenase genes) was the most abundant
[FeFe]-hydrogenase containing population (Figure 5.6). Hence, A. paucivorans shared the
same pattern as phylum Synergistetes abundance trend. [FeFe]-hydrogenase genes similar
to those of Moorella thermoacetica, a member of class Clostridia, were also found abundant
in the DM+CS bioreactors. M. thermoacetica like [FeFe]-hydrogenase genes appeared to be
more abundant in the DM+CS bioreactors, except in the DM+CS bioreactor with a HRT
of 30 days. It was the most abundant hydrogenase phylotype in the DM+CS bioreactor
with a HRT of 40 days (18.37%). We also observed that the abundance patterns of A.
paucivorans- and M. thermoacetica-like hydrogenase genes were similar to the abundance
pattern of unclassiﬁed Bacteroidetes determined based on 16S rRNA gene sequences (Figure
5.6).

145

Abundance (%
)

45
40
35
30
25
20
15
10
5
0
45
40
35
30
25
20
15
10
5
0

DM
DM+CS

Synergistia

HRT30d

HRT40d
45
Clostridia
40
35
30
25
20
15
10
5
0

HRT30d HRT40d HRT50d

HRT30d

HRT50d

Bacteroidetes

HRT40d

HRT50d

Figure 5.5: The shifts of abundance of the three most abundant [FeFe]-hydrogenases
containing bacterial groups in anaerobic bioreactors fed with DM and DM+CS.
Labels HRT30d, HRT40d and HRT50d represent bioreactors with a HRT of 30, 40,
and 50 days, respectively.

146

Unclassified Bacteroidetes
#43
#69
#14
#11

% Abundance in Community
(16S rRNA Sequences)

40

Moorella thermoacetica ATCC 39073

35
30
25
20
15
10
5
0

HRT30d
HRT50d
HRT40d

HRT30d

HRT50d

50
45
40
35
30
25
20
15
10
5
0

HRT40d

DM+CS

DM

Figure 5.6: The abundance of [FeFe]-hydrogenase containing bacteria in anaerobic
bioreactor samples compared to the abundance pattern of unclassiﬁed Bacteroidetes
(background) determined by 16S rRNA gene sequencing. The distribution of clusters
#11, #14, #43 and #69 were also plotted as part of unclassiﬁed Bacteroidetes
within each sample. Labels HRT30d, HRT40d and HRT50d represent bioreactors
with a HRT of 30, 40, and 50 days, respectively.
147

% Abundance in Community
(HydA Sequences)

Aminomonas paucivorans DSM 12260

5.4.5

Non-metric multidimensional scaling (NMDS) analysis

Non-metric multidimensional scaling (NMDS) analysis was performed on the six anaerobic
bioreactor samples based on the complete linkage clustering of 16S rRNA gene sequences
(Figure 5.7). The DM bioreactor community shifts correlated strongly with changes in HRT
while the DM+CS reactors had additional factors that diminished any correlation with HRT.
The high rates of biogas production observed in the DM+CS bioreactor with a HRT of 40
days correlates with the unique community composition that we detected in this reactor
(Figure 5.7).
Fitting the detected bacterial genera to the bioreactor community distances revealed that
the abundance of Petrimonas decreased signiﬁcantly as the HRT increased and corn stover
was added. The abundance of several genera, including Caryophanon, Blastopirellula, Ketogulonicigenium, unclassiﬁed Syntrophaceae and unclassiﬁed Bacteroidetes were also found
signiﬁcantly associated with the DM+CS bioreactor with a HRT of 40 days (Figure 5.7
red and blue arrows). Fitting of the [FeFe]-hydrogenase gene sequence data showed that
the abundance of those most closely aﬃliated with Clostridium lentocellum and Candidatus Cloacamonas acidaminovoransa were found signiﬁcantly associated with the DM+CS
bioreactor with a HRT of 40 days. The abundance of several [FeFe]-hydrogenase sequences
closest to members of Clostridia, such as Clostridium therocellum, and Bacteroidetes, such
as Bacteroides sp. 20_3, shifted signiﬁcantly along the HRT gradient and upon the addition
of corn stover.

148

0.0

Candidatus Cloacamonas acidaminovoransa
P ≤ 0.001
DM+CS
Caryophanon
HRT40d
Blastopirellula
P ≤ 0.001
Ketogulonicigenium
Unclassified
Unclassified Syntrophaceae
Bacteroidetes
Biogas
P ≤ 0.01
Petrimonas
DM
P ≤ 0.001
DM
HRT40d

DM
HRT30d P ≤ 0.05

HRT50d

Bacteroides sp. 1_1_14
Coprococcus catus GD/7

−
0.2

NMDS2

0.2

0.4

Clos tridium le ntoce llum DS M 5427

DM+CS
HRT50d

DM+CS
HRT30d

P ≤ 0.01

HRT

Degradation

−
0.4

Clostridium
P ≤ 0.001
thermocellum
Bacteroides sp. 20_3
Clostridiales bacterium 1_7_47_FAA
DSM 2360

Odoribacter splanchnicus DSM 20712

−
0.6

−
0.4

−
0.2

0.0

0.2

0.4

NMDS1
Figure 5.7: NMDS analysis of the bacterial community in anaerobic bioreactors (based on 16S rRNA gene
sequences at 97% similarity cutoﬀ). The green arrows indicate the increase of biogas production (Biogas), HRT
(HRT) and dairy manure degradation rate (Degradation). The red and blue arrows indicate bacterial genera
that shifted signiﬁcantly, where P≤0.001 and P≤0.01, respectively. The black, yellow and purple arrows represent the [FeFe]-hydrogenase phylogenetic aﬃliations that shifted signiﬁcantly (P≤0.001, P≤0.05, and P≤0.01,
respectively). All arrows point towards the abundant range of the gradients. Labels HRT30d, HRT40d and
HRT50d represent bioreactors with a HRT of 30, 40, and 50 days, respectively.
149

5.5

Discussion

Anaerobic digestion is an eﬀective method of processing farm wastes. Methane produced
from digestion can be collected and used as alternative energy source to reduce the system’s
carbon footprint. Schink et al. [140] emphasized the importance of the microbial community
in anaerobic digestion and that undesirable structural shifts the community might result in
little or no methane production. Thus it is critical to understand the roles and distributions
of fermentative bacteria within anaerobic bioreactors so that both digestion and methane
production can be optimized. The inverted relationship observed between dairy manure
degradation rate and biogas/methane production rate could be explained by the abundance
of fermentative bacteria. Schink et al. [140] previously reported that a co-culture of Caloramator proteoclasticus (a glutamate fermenter, which produces acetate and hydrogen) and
Methanobacterium thermoautotrophium (a hydrogen oxidizing methanogen) would produce
methane only when the abundance of C. proteoclasticus was low. With a relatively high concentration of C. proteoclasticus, the co-culture was eﬀective at degradation but not biogas
production [140, 162]. Thus a community shift caused a reduction of methane production.
The addition of corn stover in anaerobic digestion appeared to improve overall system
performance (Figure 5.1) in that it maximized biogas production and enhanced total solids
reduction. Biogas methane content remained relatively stable across the six bioreactors
(66-67%) but overall biogas production increased signiﬁcantly in the DM+CS bioreactors.
The latter is consistent with other reported studies that the addition of crop remnants was
beneﬁcial to methane production [30, 42, 46, 85, 110, 179]. The increase in biogas production
rate could be related to cellulose and hemicellulose degradation as observed by Yue et al.

150

DM+CS
DM

Polymers
Unclassified
Bacteroidetes
Monomers
Fatty acids
Alcohols
Lactate
Succinate
Escherichia/Shigella
Aminomonas paucivorans
Clostridia

H2, CO2,
HCOO-,
CH2-R

Clostridia

Methanobacterium
CH4

Acetate

Methanosarcina
Methanosaeta

Figure 5.8: Diagram (adapted from Schink et al. [140]) of microbial corporation
during methane generation in DM and DM+CS bioreactors operated at HRT 40
days. The thickness of the arrow represent the relative abundance of microbial
populations possibly associated with the methanogensis step.

151

[192]. Thus it may be that cellulose and hemicellulose degrading bacteria and associates
contributed to the generation of hydrogen and acetate which in turn supported the growth
of methanogens.
Members of class Clostridia were frequently reported in other anaerobic digestion studies. Class Clostridia describes a diverse group of anaerobic bacteria with versatile metabolic
pathways, many members of which have been reported to have high cellulolytic activity
and are capable of both hydrogen consumption and acetate production through the reductive acetyl-CoA pathway [185]. For these reasons we expected to see a substantial amount
of class Clostridia in our anaerobic bioreactors. A recent metagenomic study on methane
producing community reported by Wirth et al. revealed that 36% of the community were
Clostridia and 23 out of 40 most frequently found bacterial species in the community were
members of class Clostridia [185]. In another similar study, Schl¨ter et al. showed that class
u
Clostridia was highly abundant (52%) in a production-scale biogas plant fermenter [142]. As
expected, we observed a large amount of Clostridia in both the DM and DM+CS bioreactors.
However, class Clostridia appeared to be more abundant in the DM bioreactors than in the
DM+CS bioreactors, except in the DM+CS bioreactor with a HRT of 40 days (Figure D.2
B). Clustering analysis revealed that the populations within this group varied substantially
from condition to condition and only a few populations were common in all six anaerobic
bioreactors (Figure 5.2 B2). Hence, populations within class Clostridia were dynamic and
diﬀerent members of the group assumed diﬀerent responsibilities (Figure 5.2 B2). Our observation of unpredictable shifts within Clostridia revealed by 16S rRNA sequences agreed with
the conclusion by Werner et al. that Clostridia populations were relatively unstable compared to other bacterial populations in their methanogenic bioreactor bacterial communities
152

[182]. Thus, the abundance of Clostridia as a group based on their 16S rRNA sequences
could not be used to correlate methane production in the anaerobic bioreactors. In contrast,
hydrogen producing populations identiﬁed by [FeFe]-hydrogenase gene clone libraries were
correlated with anaerobic digestion and methane production. For example, we observed that
[FeFe]-hydrogenase genes aﬃliated with M. thermoacetica (a member of Clostridia) were
more abundant in the reactors that were supplemented with corn stover (Figure 5.6). However, genus Moorella was not detected according to the 16S rRNA gene proﬁle. Moreover,
one could infer from the high abundance of Clostridia-like [FeFe]-hydrogenase genes observed
in the DM bioreactors at HRT of 30 and 40 days that Clostridia played a crucial role in hydrogen production in the DM bioreactors. Clostridia-like [FeFe]-hydrogenase genes became
more abundant in the DM+CS bioreactor as the HRT increased, perhaps indicating that an
increased amount of time was required to reduce TS. Class Clostridia could be producing
hydrogen from secondary fermentation of cellobiose or glucose generated by other bacteria
from primary cellulose/hemicellulose degradation (Figure 5.8).
Phylum Bacteroidetes constituted approximated 21-40% of each anaerobic bioreactor
bacterial community in this study. Like class Clostridia, Bacteroidetes populations were also
commonly reported in anaerobic digestion but are usually less abundant. Wirth et al. observed approximately 4% of Bacteroidetes in their metagenomic analysis [185]. Schlüter et
al. [142] reported that there were 16% of Bacteroidetes in the production-scale biogas plant
fermenter. Wang et al. [181] observed a slightly larger amount of phylum Bacteroidetes
(up to 23%). Members of Bacteroidetes including Cytophaga sp., Sporocytophaga myxococcoides, Flavobacterium johnsoniae have been reported as cellulose/hemicellulose degraders
[29, 83, 107]. A recent study conducted by Stursov˘ et al. [155] revealed that Bacteroidetes
a
153

populations played a substantial role in natural cellulose/hemicellulose degradation in forest
soil. We observed that phylum Bacteroidetes showed a similar abundance pattern to class
Clostridia, where they were generally more abundant in the DM+CS than in the DM bioreactors (Figure D.2 A). However, unlike Clostridia, a substantial amount of unclassiﬁed Bacteroidetes were observed in the DM+CS bioreactors. In the DM+CS bioreactor with a HRT
of 40 days, over 80% of the phylum Bacteroidetes could not be classiﬁed further than phylum
level (Figure 5.3). Clustering of all Bacteroidetes sequences at 97% cut oﬀ revealed that three
unique unclassiﬁed Bacteroidetes populations were associated with the DM+CS bioreactors.
These three populations contributed substantially to the shifts from classiﬁed to unclassiﬁed bacteria within phylum Bacteroidetes observed in this study. RDP SeqMatch results
revealed that these three groups of unclassiﬁed Bacteroidetes were highly similar to bacterial
16S rRNA gene clones found in methanogenic environments, speciﬁcally, domestic wastewater enriched microbial fuel cells (#43, SeqMatch score=0.998), biogas slurry (#11, SeqMatch
score=0.966), Kinneret lake sediments (#14, SeqMatch score=0.983), methanogenic zone of
a hydrocarbon- and chlorinated-solvent contaminated aquifer (#69, SeqMatch score=0.988)
and an anaerobic bioreactor fed with butyrate and sulphate enriched paper mill wasterwater
(#69, SeqMatch score=0.988, unpublished work) [37, 84, 134, 143]. Thus based on prior
observations as well as ours, these three Bacteroidetes clusters can play an important role
in some methanogenic environments. Moreover, their high abundance in corn stover supplemented bioreactors suggests that cellulose and hemicellulose may play a role in selecting
for their presence. We observed that the dairy manure degradation rate increased along
with the increase in HRT and groups #11, #14 and #43 became more abundant in the
DM+CS bioreactors operated at long HRTs. Cluster #69 was observed only in the DM+CS
154

bioreactor with a HRT of 40 days and at considerable abundance, which suggested that this
speciﬁc group of Bacteroidetes might contribute to the high biogas production rate in the
DM+CS bioreactor with a HRT of 40 days. Agreeing with our prediction, NMDS analysis
revealed that unclassiﬁed Bacteroidetes shifted signiﬁcantly (P ≤ 0.01) towards the DM+CS
bioreactor with a HRT of 40 days (Figure refnmds). Besides cellulolytic activities, members of phylum Bacteroidetes have also been identiﬁed as hydrogen producers. While the
16S rRNA gene based survey suggested that members of Bacteroidetes played an important
role in the anaerobic digestion, the [FeFe]-hydrogenase clone library proﬁle revealed that
Bacteroidetes-like [FeFe]-hydrogenase genes were not as abundant as would be expected if
they were linked to these important Bacteriodetes populations. It was observed that a larger
amount of Bacteroidetes-like [FeFe]-hydrogenase genes were detected in bioreactors despite
the diﬀerences in their operational HRT. In addition, the higher abundance of Bacteroideteslike hydrogenase genes were observed in the DM bioreactors than in the DM+CS bioreactors
under all three HRTs. These trends indicate that classiﬁed Bacteriodetes populations directly
contributing to the hydrogen economy through [FeFe]-hydrogenases were not responding to
the presence of corn stover, as were other unclassiﬁed populations of Bacteriodetes. Correlating 16S rRNA gene analysis to the [FeFe]-hydrogenase gene study, we speculate that
classiﬁed Bacteroidetes contributed to methane production in anaerobic bioreactors by generating hydrogen during dairy manure degradation and unclassiﬁed Bacteroidetes participated
solely in primary degradation of cellulose and hemicellulose. The degraded products, such
as cellobiose and glucose, could be then consumed by secondary fermenters (e.g., Clostridia
and Escherichia/Shigella for hydrogen generation (Figure 5.8).
While we expected to observe class Clostridia and phylum Bacteroidetes in abundance
155

in anaerobic bioreactors, it was surprising to observe genera Escherichia/Shigella as one
of the dominant groups in the DM+CS bioreactor. These genera include species that are
commonly found in animal gastrointestinal tracks, including pathogens. They are not well
known for cellulose and hemicellulose degradation but frequently excel at metabolism of
sugars. Therefore, Escherichia/Shigella are not likely contributing to cellulose and hemicellulose degradation despite their high abundance in the DM+CS bioreactors. It is known that
Escherichia/Shigella are found in feces, studies have shown that anaerobic fermentation is effective at reducing their abundance, especially pathogenic populations [70, 101]. We have observed an increase in both the abundance and the diversity of genera Escherichia/Shigella as
HRT increased from 30 to 50 days. More surprisingly, the abundance of Escherichia/Shigella
was generally higher in the DM+CS bioreactors than in the DM bioreactors, except the
DM+CS bioreactor with a HRT of 40 days (Figure D.2 C). Such a large abundance of Escherichia/Shigella group was not observed in other previously reported anaerobic bioreactor
studies. SeqMatch revealed that these Escherichia/Shigella sequences are similar to those
of Shigella and non-pathogenic E. coli. The increase in Escherichia/Shigella abundance
and diversity suggested that facultative fermenters may survive and even actively grow in
anaerobic bioreactors, and be further enriched by increasing C:N ratio (the more corn stover
in the feed the higher C:N ratios). The smallest populations of Escherichia/Shigella were
observed in the DM+CS bioreactor with a HRT of 40 days (1.53% of community). The
small abundance of Escherichia/Shigella was likely oﬀset by the large abundance of class
Clostridia and phylum Bacteriodetes.
The high abundance of Synergistia-like hydrogenase genes (at least 35%) observed in all
of the DM+CS bioreactors suggested that this group of bacteria were highly active in cel156

lulose/hemicellulose rich environments. The increase in Synergistia like [FeFe]-hydrogenase
genes in the DM bioreactors along with the increase in the HRT suggested that phylum
Synergistia grew on dairy manure degradation products. At species level, clone libraries
revealed that the abundance of Synergistia was solely contributed by Aminomonas paucivorans, an anaerobic bacterium that was previously identiﬁed as an amino-acid-degrader.
Baena et al. [5] described an A. paucivorans isolate that was able to grow on arginine,
histidine and glutamate when cultivated syntrophically with methanogen Methanobacterium
formicicum. The end products of the co-culture were propionate, acetate, CO2 and methane
[5]. There are no previous studies indicating that Aminomonas paucivorans is capable of
cellulose degradation. The available information on Aminomonas paucivorans indicate that
its high abundance in the DM+CS bioreactors was likely a result of growth due to secondary
fermentation of degraded dairy manure [5].
NMDS analysis revealed a complex network among bacterial communities in the methane
producing anaerobic bioreactors. Correlating 16S rRNA gene with [FeFe]-hydrogenase gene
diversities, we observed that the high abundance of unclassiﬁed Bacteroidetes was significantly associated with the DM+CS bioreactor with a HRT of 40 days that produced
the most methane. Besides unclassiﬁed Bacteroidetes, the abundance of Caryophanon,
Blastopirellula, Ketogulonicigenium and unclassiﬁed Syntrophaceae increased signiﬁcantly
in the DM+CS bioreactor with a HRT of 40 days. Little has been discovered regarding the
metabolic capabilities of Caryophanon and Blastopirellula. Members of Ketogulonicigenium
and Syntrophaceae were previously identiﬁed as fermenters [53, 93, 156]. Based on these
information, we speculate that unclassiﬁed Bacteroidetes played a substantial in primary
cellulose/hemicellulose degradation. It is likely that genera Caryophanon, Blastopirellula,
157

Ketogulonicigenium and unclassiﬁed Syntrophaceae participated in secondary fermentation.
NMDS analysis also showed that diﬀerent members of Clostridia and Bacteroidetes participated in diﬀerent activities among all six anaerobic bioreactors. For example, Bacteroides
sp. 1_1_14 appeared to be aﬃliated with hydrogen production during dairy manure digestion at a low HRT, while Bacteroides sp. 20_3 was found signiﬁcantly associated with
DM+CS digestion at an elevated HRT. Similarly, C. thermocellum like [FeFe]-hydrogenase
genes were found to be signiﬁcantly correlated with the low HRT bioreactors and Clostridiales bacterium and Odoribacter splanchnicus might play a signiﬁcant role during methane
production in DM+CS digestion at elevated HRT.
Although A. paucivorans- and M. thermoacetica- (a member of class Clostridia) like
[FeFe]-hydrogenase genes were highly abundant and varied in anaerobic bioreactors, NMDS
analysis indicated that the shifts of their abundance were not signiﬁcant across all bioreactors. However, we observed that the abundance of these two groups of [FeFe]-hydrogenase
genes followed the abundance pattern of unclassiﬁed Bacteroidetes determined based on 16S
rRNA gene sequences (Figure 5.6). We speculate that hydrogen producing population A.
paucivorans responds to the growth of unclassiﬁed Bacteroidetes. It agrees with our prediction that unclassiﬁed Bacteroidetes were responsible for cellulose/hemicellulose degradation
but not for hydrogen production. Subsequently, cellulose/hemicellulose degradation products, such as xylose, glucose and cellulobiose, were consumed by hydrogen producers (A.
paucivorans and M. thermoacetica).
Yue et al. [192] reported the archaeal community structures resided in these anaerobic
bioreactors, where they observed that the communities were dominated by Methanobacterium, Methanosarcina and Methanosaeta populations. Members of Methanobacterium
158

are known for producing methane via hydrogen oxidation. The presence of this group
of methanogens could explain the consumption of hydrogen produced by A. paucivorans,
Clostridia and Bacteroidetes. Accompanied to the production of hydrogen. The literature
indicates that many members of [FeFe]-hydrogenase containing bacteria generate acetate
as one of the ﬁnal products [11]. The presence of a large amount of Methanosarcina and
Methanosaeta suggests that acetate was a sole source for methane generation. Yue et al. [192]
also concluded that the community shifts from Methanosarcina dominant to Methanosaeta
dominant was due to the decrease in acetate concentration in the bioreactors as HRT increased. This also explains the decreasing trend in Methanobacterium abundance in both the
DM and DM+CS bioreactors as the HRT increased from 30 to 50 days. These observations
are consistent with the observation that the biogas/methane production rate approached
plateau when the operational HRT was more than 40 days.

5.6

Conclusion

In conclusion, the DM+CS bioreactor with a HRT of 40 days was unique for its distinct
community structure and high methane production rate. We predict that unclassiﬁed Bacteoridetes populations might act as primary fermenters and play a signiﬁcant role in cellulose/hemicellulose degradation. A. paucivorans and members of Clostridia, such as M.
thermoacetica are likely participating substantially in hydrogen and acetate production (Figure 5.8). The observations made by Yue et al. [192] suggested that methane production from
H2 and CO2 decreased as HRT increased. However, we observed [FeFe]-hydrogenase gene
containing populations at high abundance in bioreactors operated at large HRTs . The high-

159

est rate of methane production was detected in the bioreactor fed with both dairy manure
and corn stover at a HRT of 40 days. This bioreactor had a unique community with an
unclassiﬁed population of Bacteriodetes not found in other bioreactors. The population size
of Escherichia/Shigella was suppressed in this bioreactor while the hydrogen economy was
likely driven by Clostridia and A. paucivorans, the former through the reductive acetyl-CoA
pathway. We speculate that the acetate concentration was somewhat low at this retention
time, thus selecting for Mathanosaeta rather than Methanosarcina.

160

APPENDICES

161

APPENDIX A
COMPARATIVE GENOMICS OF RALSTONIA PICKETTII 12D AND 12J

A.1

Metabolic Pathways of R. pickettii 12D and 12J

We observed minor diﬀerences in the R. pickettii 12D and 12J available KEGG pathway maps. Strain 12J has glucose-1-phosphate thymidylyltrasferase, mannose-1-phosphate
guanylyltransferase / mannose-6-phosphate isomerase, and GDP-mannose 4,6-dehydratase
genes, which are missing in 12D (conﬁrmed with JGI-IMG BLAST). On the other hand,
R. pickettii 12D has N-acylneuraminate-9-phosphate synthase and 4,5-dihydroxyphthalate
decarboxylase genes that could not be identiﬁed by JGI-IMG BLAST in strain 12J.
MetaCyc showed that R. pickettii 12D and 12J contain several detoxiﬁcation pathways, including arsenate reduction, cytoplasm Hg2+ reduction, methylglyoxal detoxiﬁcation,
cyanate degradation and superoxide radical degradation. Among these detoxiﬁcation pathways, arsenate detoxiﬁcation was only mentioned in R. pickettii 12D by MetaCyc. However,
arsenate reductase encoding genes were identiﬁed in R. pickettii 12J via JGI-IMG BLAST.
Besides the detoxiﬁcation pathways, MetaCyc also revealed that both of R. pickettii 12D and
12J are capable of organic acid, aromatic compound and chlorinated compound degradation.
They also contain genes that are essential for assimilating inorganic elements like P, N and
S from ammonium, nitrate, phosphate and sulfate. We have experimentally conﬁrmed that
both 12D and 12J are capable of denitriﬁcation under anaerobic minimal growth conditions
(unpublished data).

162

A.2

Clustering of R. pickettii 12D and 12J and other 23 Proteobacteria

0.01

Geobacter metallireducens GS-15
Geobacter lovleyi SZ
Geobacter uraniireducens Rf4
Escherichia coli O157:H7 Sakai EHEC
Escherichia coli O1:K1:H7 AP EC
Escherichia coli O44:H18 042 EAEC
P seudomonas fluorescens P f-5
P seudomonas syringae pv. tomato DC3000
P seudomonas putida F1
Pseudomonas aeruginosa PA7
Cupriavidus necator J MP 134
Ralstonia eutropha H16
Cupriavidus necator N-1
Cupriavidus taiwanensis LMG 19424
Cupriavidus metallidurans CH34
Ralstonia solanacearum GMI1000
Ralstonia pickettii 12J
Ralstonia pickettii 12D
Burkholderia xenovorans LB400
Burkholderia thailandensis E264
Burkholderia cepacia 383
Burkholderia pseudomallei 1710b
Burkholderia mallei ATCC 23344
Burkholderia multivorans ATCC 17616

Figure A.1: Distance tree of R. pickettii 12D and 12J in relationship to 23
other Proteobacteria based on their 16S rRNA gene sequences. The bar in the
bottom left corner represents the distance scale.

163

A.3

Genetic Elements in R. pickettii 12D and 12J
Table A.1: Side-by-side genome comparison of R. pickettii 12D and 12J genetic elements by genomic scaﬀolds.
C1
12D 12J

P2
12D

C2
12D 12J

P3
12D

HGTa
115
118
9
4
3
2
Transposase
22
28
6
3
3
0
Filamentous Phageb
1
3
0
1
1
0
c
Unique Genes
63
74
1
0
5
0
a Number of genes that are identiﬁed by IMG as horizontally transferred.
b Mostly incomplete copies except one in 12J C1.
c All genes that are absent in the other strain based on IMG annotations.

164

P1
12J

P1
12D

2
1
0
0

157
2
0
15

A.4

Variation of R. pickettii 12D and 12J dif Sites

R. solanacearum GMI1000
R. pickettii 12D C1
R. pickettii 12D C2
R. pickettii 12J C1
R. pickettii 12J C2

C
A
A
G
A

CA
TT
TT
GT
TT

TCGC
TAAC
TAAC
TTTC
TAAC

A
A
A
A
A

10
TAA
TAA
TAA
TAA
TAA

TTTA
TGCA
CACA
TGCA
CACA

TC
AA
AA
AA
AA

T
T
T
T
T

20
TA TG
TA TG
TC TG
TA TG
TA TC

TTAA
AGAA
CG TC
AGAA
CGAA

A
A
A
A
T

T
T
T
T
T

Co ns e ns us
A TTTAACA TAA T+CAAA TTA TG +GAAA T
Figure A.2: The sequence variations of dif sites of R. pickettii 12D and 12J. The consensus reveals the conserved
the regions.

165

Four 28 bp dif site sequences obtained from R. pickettii 12D and 12J chromosomes
were compared along with the one from their close relative R. solanacearum GMI1000 (Figure A.2). All ﬁve sequences are conserved at positions 5, 8-11, 15, 18-19, 21 and 28. R.
pickettii 12J_C1 dif appears to be the most diﬀerent and is complementary to GMI1000 dif
at position 1, 2, 16 and 23. R. pickettii 12D_C1 dif is complementary to 12J_C1 (position
5 and 6). R. pickettii 12D C2 and 12J C2 dif are similar to 12D_C1 and the sequences are
complementary at position 25 (12D_C2 to 12D_C1) and position 22 and 27 (12J_C2 to
12D_C1) (Figure A.2).

166

A.5

Genome Syntenies of R. pickettii 12D and 12J

Base-to-base comparisons of R. pickettii 12D and 12J scaﬀolds revealed that the primary
chromosomes are in good agreement. The discontinued and gapped line in 12D_C2-12J_C1
frame indicates a DNA fragment was inserted into 12J_C1 (Figure A.3). This inserted
portion is highly homologous to 12D_P2 (Figure A.3 12D_P2-12J_C1). The disjointed
lines in Figure A.3 12D_C2-12J_C2 frame is an evidence of genome rearrangement, where
the origin of replication (ori C) appears to be relocated in 12D_C2 in relation to 12J_C2.
Figure A.3 also showed that the 12J_P1 sequence largely agrees with 12D_P3 and an
approximately 0.03 Mb fragment of 12D_C1 (Figure A.3). R. pickettii 12D_P1 shares few
syntenies with other scaﬀolds and appears to be unique to itself (Figure A.3).

167

P1
Rals tonia picke ttii 12J

C2

C1

C2

C1

P2 P1

P3

R a ls to nia p ic ke ttii 1 2 D
Figure A.3: Genome syntenies of R. pickettii 12D and 12J revealed by a DNA sequence dotplot. Blue dots
represent matching genes on the same direction strands. Red dots represent matching genes on the opposite
direction strands. The scaﬀold nucleotide coordinates are ranked from left to right on the x-axis and bottom to
top on the y-axis.

168

APPENDIX B
ANAEROBIC ENRICHMENTS

B.1

Media Preparation
Table B.1: The composition of the modiﬁed FWM.
Ingredients

Concentration

Mineral Mix (100X)
NaHCO3
NH4 Cl
NaH2 PO4 •H2 O
KCl
Adjust pH to 6.8-7 under

10 ml/L
2.5 g/L
0.25 g/L
0.6 g/L
0.1 g/L
N2 :CO2 =80:20.

Table B.2: The composition of the mineral mix (100X).
Ingredients

g/L

Nitrilo triacetic acid (C6 H9 NO6 ) 1.5
MgSO4 •7H2 O
3
MnSO4 •2H2 O
0.5
NaCl
1
FeSO4 •7H2 O
0.1
CaCl2 •2H2 O
0.1
CoCl2 •7H2 O
0.1
ZnCl2
0.13
CuSO4 •5H2 O
0.01
AlK(SO4 )2 •12H2 O
0.01
H3 BO3
0.01
Na2 MoO4
0.025
NiCl2 •6H2 O
0.024
Na2 WO4 •2H2 O
0.025

169

Table B.3: The composition of the vitamin mix (1000X).
Ingedients

mg/L

Biotin
Folic acid
Pyridoxine-HCl
Thamine-HCl•2H2 O
Riboﬂavin
Nicothinic acid
D-Ca-pantothenate
Vitamin B12
p-Aminobenzoic acid
Lipoic acid

2
2
10
5
5
5
5
0.1
5
5

170

Figure B.1: Functional gene proﬁles of original soil communities and the enriched
communities (replicate 1). Symbol "O" indicates original communities and "Fe",
"NO3" and "U" indicate supplemented electron acceptors in enrichments.

B.2

Functional Gene Proﬁles

GeoChip analysis on the original soil sediment samples and the enrichment replicate 1 revealed a total of 475 diﬀerent genes. The functional gene proﬁle patterns are highly similar
(Figure B.1).

171

APPENDIX C
URANIUM IMMOBILIZATION BY FIRMICUTES

C.1

Uranium-azide Assay

3.5
3.0

A340

2.5
2.0
1.5

y = 0.0032x - 0.0065
R² = 0.99965

1.0
0.5
0.0
0

200

400

600

800

1000

1200

U(VI) (µM)
Figure C.1: Standard curve of the uranium-azide assay. The error bars represent
data from two replicates.

172

C.2

5X Gitschier Buﬀer
Table C.1: Ingredients for 200 ml of 5X Gitschier Buﬀer.

To prepare 200 ml of 5X Gitschier combine:

Final concentration

16.6 ml of a 1 M (NH4 )2 SO4
83 mM (NH4 )2 SO4
67 ml of a 1 M Tris-HCl pH 8.8
335 mM Tris-HCl pH 8.8
6.7 ml of a 1 M MgCl2
33.5 mM MgCl2
1.3 ml of 1:100 dilution of a 0.5 M EDTA
33.5 µM EDTA
2.08 ml of a 14.4 M commercial stock of β-mercapto-ethanol 150 mM β-mercapto-ethanol
Adjust ﬁnal volume to 200 ml and autoclave for sterilization.

C.3

Growth Curves of Isolates in Minimal Media

173

0.6
0.4
0.2
0

A1_U

A1

A10_U

A10

A1
A3_U

A3

A4_U

A4

A3

OD620

0.6
0.4
0.2
0

A8_U

A8

B5_U

B5

C10_U

C10

C4_U

E10_U

B1

B8

C8(2)_U

B10

B9

B9_U

C2_U

C5_U

C5

C9_U

C9

50

A6

A6_U

B11_U

B11

B9(2)_U

E3

A7_U

B11

B2_U

C6

D3_U

E7_U

150

C7

C1_U

C1

C3_U

C3

C7_dup_U

C7_dup

C7

D5_U

D5

Neg_U

D8_U

D8

D5
E7

B2

C3

C7
D3

A7

A7

C1
C2c5

C7_U

A2

A2

B2

C2c5

E7

100

A6

B9(2)

C2c5_U

D3

E3_U

A2_U

B9(2)

C6

E3
0

B9

C2

C6_U

C9
E2

A5_dup

B10_U

A12(2)

A12(2)

C2

C5
C8(2)

E2_U

C12

C12
C5c2

A12(2)_U

B10

C12_U

E2
150

A5_dup_U

A12

A5

B8_U

C8(2)
E10

100

C11

C5c2
C8

A12

B8

C5c2_U

C8

A5

B1_U

C11
C4

C8_U

50

A5_U

A12_U

B1
B6

C11_U

C10

E10
0

A9

B6_U

A11

A11
A5

B6

C4
0.6
0.4
0.2
0

A4

A11_U

A9

A8
B5

0.6
0.4
0.2
0

A9_U

A10

Neg

Neg
0

50

100

150

Time (hours)
Figure C.2: The growth curves of FRC isolates (ﬁrst replicate) in minimal media supplemented with sodium
pyruvate alone (blue) and minimal media supplemented with sodium pyruvate and uranyl acetate (red) under
anaerobic condition.

174

D8

0.6
0.4
0.2
0

A05_U

A05

A1_U

A1

A5
A2_U

A2

A3_U

A3

OD620

A8_U

A8

A9_U

B8_U

0.6
0.4
0.2
0

C11_U

C12_U

C11
C6_dup_U

C6

0.6
0.4
0.2
0

D3_U

D3

E7

E7
0

50

100

D5

Neg_U

B1

B10

B9(2)_U

A6_U

B11_U

C2

C4_U

C7_U

C7

C8_U

C8

C8

D8_U

D8

D8

E10_U

0

50

E10

100

C1

C10_U

C4

C8(2)_U

150

E2_U

C10

C5

C9_U

C9

C9
E2

E3_U

E3

E3
0

50

100

Neg
150

Time (hours)

Figure C.3: The growth curves of FRC isolates (second replicate) in minimal media supplemented with sodium
pyruvate alone (blue) and minimal media supplemented with sodium pyruvate and uranyl acetate (red) under
anaerobic condition.

175

C10

C5
C8(2)

E2

B2

C5_U

C8(2)

E10

Neg

B2_U

B2

C4

C3

A7

A7
B11

C1

C1_U

C3

C3_U

A7_U

B11
B9(2)

A12(2)

A12(2)
A6

B9(2)

C2_U

A12(2)_U

A6

B10
B9

A12

A12
A5

B10_U

C7

D5

D3
E7_U

D5_U

A12_U

A5

C2

C6

C6

A5_U

B9
C12

C6_U

A4

B9_U

C12
C6_dup

A11

A11

B1_U

B8
C11

A11_U

B1
B8

B5

A10

A4_U

A9
B5

A10

A4
A9

A8
B5_U

A10_U

A3

A2
0.6
0.4
0.2
0

A1

150

APPENDIX D
BACTERIAL COMMUNITIES IN ANAEROBIC CO-BIOREACTORS

D.1

Deep Sequencing of Bacterial Communities in Anaerobic Cobioreactors

Table D.1: Measurements of sampling size and estimation of ecological indices of
bacterial community in six anaerobic bioreactors. Calculations were based on the
complete linkage clustering of chimera removed and trimmed bacterial 16S rRNA
gene sequence alignments at 97% similarity. N represents the total number of nonchimera sequences that passed quality trimming. Shannon index (H’) and Simpson’s
index (D) were expressed in eH and 1-D, respectively to emphasize the diﬀerences.
C represents the sampling coverage via gene targeted pyrosequencing.
Inﬂuent
HRT (days)
N
OTU
ChaoI
eH
E
1-D
C

100% DM
30
40
50
6729
5003
5328
230
187
171
240.62 208.94 190.50
70.62 49.30 36.10
0.78
0.75
0.70
0.97
0.94
0.91
99.64 99.46 99.49

80% DM + 20% CS
30
40
50
4939
5684
5323
163
277
134
193.67 332.80 146.21
40.64 73.62 27.30
0.73
0.76
0.68
0.91
0.97
0.87
99.51 98.89 99.64

To determine the phylogenetic structure of the microbial communities in anaerobic cobioreactors, we collected samples from all six bioreactors. Each anaerobic bioreactor had
completed 2 HRT cycles prior to withdrawing the samples. We employed 454-pyrosequencing
technology to phylogenetically characterize bacteria in each bioreactor based on bacterial 16S
rRNA gene sequences. This level of sampling permitted the detection of bacterial populations
that were minorities [16]. At least 4900 sequences per sample met quality prerequisites.
Operational Taxonomy Units (OTU) were calculated based on complete linkage clustering

176

300
250
200
150
100
0

50

Number of OTU's

DM HRT30d
DM+CS HRT30d
DM HRT40d
DM+CS HRT40d
DM HRT50d
DM+CS HRT50d
0

1000 2000 3000 4000 5000 6000 7000
Number of Sequences

Figure D.1: Estimation of community coverage by rarefaction analysis. Rarefaction curves of anaerobic bioreactor 16S rRNA gene sequences were constructed based on the estimation of OTUs and ChaoI indices at 97% similarity.

of aligned sequences at 97% similarity (Table D.1) (all clustering analyses in this paper
refer to clustering done with these parameters). Rarefaction analysis was conducted on
each sample by dividing rareﬁed number of OTU with rareﬁed ChaoI estimator. At 97%
sequence similarity, the rarefaction curves approached plateau (Figure D.1). Good’s nonparametric coverage estimator was used to determine the community coverage (Table D.1)
[39]. The results indicated that at least 98% of each bacterial community could be described
by sequences obtained (Table D.1). Magniﬁed Shannon diversity index (eH ) indicated that
177

the DM bioreactor with a HRT of 30 days and the DM+CS bioreactor with a HRT of 40
days were more diverse than the others, consistent with their Simpson’s indices (1-D) and
ChaoI estimations. Shannon evenness (E) indices revealed that the communities in all six
bioreactors were relatively skewed.
Pyrosequencing appeared to be an eﬀective method in studying anaerobic bioreactor
bacterial communities in depth (Figure D.1). Analysis of bacterial 16S rRNA gene sequences
revealed that 99% of the populations in each community were described phylogenetically.
Interestingly, we observed that the bacterial communities fed with DM appeared to be more
diverse than DM+CS bioreactor communities, except the DM+CS bioreactor with a HRT
of 40 days, the bioreactor with the highest methane production rate (Table D.1). This
suggested that bacterial diversities do not correlate with biogas production and that subtle
environmental factors may inﬂuence community structure and resulting biogas production
levels.

D.2

Anaerobic Co-bioreactor Bacterial Community Compositions
at Phylum Level

Phylum Bacteroidetes, Firmicutes and Proteobacteria were highly abundant in all six bioreactor samples. The populations from these three phyla comprised 83% or greater of all six
bioreactor communities (Figure D.2). In the DM bioreactor samples, the abundance of Proteobacteria increased and the abundance of Firmicutes decreased as HRT increased. The
abundance of phylum Bacteroidetes ﬂuctuated slightly with no correlation to HRT, shifting
from 33.21% in the DM bioreactor with a HRT of 30 days to 21.47% in the DM bioreactor
with a HRT of 40 days, and 28.40% in the DM bioreactor with a HRT of 50 days. Similar ob-

178

servations were made in the DM+CS bioreactor samples. Proteobacteria was more abundant
in the DM+CS bioreactor with a HRT of 50 days than the DM+CS bioreactor with a HRT
of 30 days. The abundance of Firmicutes decreased in the DM+CS bioreactor with a HRT of
50 days when comparing to the DM+CS bioreactor with a HRT of 30 days. The abundance
of Bacteroidetes stayed approximately the same (23.20% in the DM+CS bioreactor with a
HRT of 30 days and 21.87% in the DM+CS bioreactor with a HRT of 50 days). Phylum
Proteobacteria was abundant in the DM+CS bioreactor with a HRT of 30 days (44.60%)
and the DM+CS bioreactor with a HRT of 50 days (55.55%). However, unlike the rest of the
samples, the DM+CS bioreactor at a HRT of 40 days had a distinctive pattern. Only 7.23%
of the populations were assigned to phylum Proteobacteria in the DM+CS bioreactor with
a HRT of 40 days and 35.40% and 40.38% of the populations were assigned to Firmicutes
and Bacteroidetes, respectively (Figure D.2).

179

Abundance (%)

100
90
80
70
60
50
40
30
20
10
0

HRT30d
HRT50d
HRT40d
DM

HRT50d
HRT30d
HRT40d
DM+CS

Proteobacteria

Bacteroidetes

Escherichia/Shigella

Firmicutes

unclassified Bacteroidetes

Clostridia

Figure D.2: The distribution of the top three most abundant phyla (wide columns)
in bioreactor samples. The narrow columns represent the abundance of genera Escherichia/Shigella (checker box), group unclassiﬁed Bacteroidetes (diagonal stripes) and
class Clostridia (blacks) in each community.

180

BIBLIOGRAPHY

181

BIBLIOGRAPHY

[1]

Acinas, S. G., Klepac-Ceraj, V., Hunt, D. E., Pharino, C., Ceraj, I., Distel, D. L., and
Polz, M. F. Fine-scale phylogenetic architecture of a complex bacterial community.
Nature 430, 6999 (July 2004), 551–4.

[2]

Akob, D. M., Mills, H. J., Gihring, T. M., Kerkhof, L., Stucki, J. W., Anastácio,
A. S., Chin, K.-J., Küsel, K., Palumbo, A. V., Watson, D. B., and Kostka, J. E.
Functional diversity and electron donor dependence of microbial populations capable
of U(VI) reduction in radionuclide-contaminated subsurface sediments. Applied and
Environmental Microbiology 74, 10 (May 2008), 3159–70.

[3]

Akob, D. M., Mills, H. J., and Kostka, J. E. Metabolically active microbial communities
in uranium-contaminated subsurface sediments. FEMS Microbiology Ecology 59, 1
(Jan. 2007), 95–107.

[4]

Andersson, A. F., Lindberg, M., Jakobsson, H., Bäckhed, F., Nyrén, P. l., and Engstrand, L. Comparative analysis of human gut microbiota by barcoded pyrosequencing. PLoS One 3, 7 (Jan. 2008), e2836.

[5]

Baena, S., Fardeau, M., Ollivier, B., Labat, M., T, P., and De, A. Aminomonas paucivorans gen. nov., sp. nov., a mesophilic, anaerobic, amino-acid-utiIizing bacterium.
International Journal of Systematic Bacteriology 49 (1993), 975–982.

[6]

Bastviken, D., Tranvik, L. J., Downing, J. A., Crill, P. M., and Enrich-Prast, A.
Freshwater methane emissions oﬀset the continental carbon sink. Science (New York,
N.Y.) 331, 6013 (Jan. 2011), 50.

[7]

Blöthe, M., Akob, D. M., Kostka, J. E., Göschel, K., Drake, H. L., and Küsel, K.
pH gradient-induced heterogeneity of Fe(III)-reducing microorganisms in coal miningassociated lake sediments. Applied and Environmental Microbiology 74, 4 (Feb. 2008),
1019–29.

[8]

Borch, T., Kretzschmar, R., Kappler, A., Cappellen, P. V., Ginder-Vogel, M., Voegelin,
A., and Campbell, K. Biogeochemical redox processes and their impact on contaminant
dynamics. Environmental Science and Technology 44, 1 (Jan. 2010), 15–23.

[9]

Boswell, C. D., Dick, R. E., and Macaskie, L. E. The eﬀect of heavy metals and other
environmental conditions on the anaerobic phosphate metabolism of Acinetobacter
johnsonii. Microbiology (Reading, England) 145 (July 1999), 1711–20.

[10] Boyd, E. S., Hamilton, T. L., Spear, J. R., Lavin, M., and Peters, J. W. [FeFe]hydrogenase in Yellowstone National Park: evidence for dispersal limitation and phylogenetic niche conservatism. The ISME Journal 4, 12 (Dec. 2010), 1485–95.

182

[11] Boyd, E. S., Spear, J. R., and Peters, J. W. [FeFe] hydrogenase genetic diversity provides insight into molecular adaptation in a saline microbial mat community. Applied
and Environmental Microbiology 75, 13 (July 2009), 4620–3.
[12] Brodie, E. L., Desantis, T. Z., Joyner, D. C., Baek, S. M., Larsen, J. T., Andersen,
G. L., Hazen, T. C., Richardson, P. M., Herman, D. J., Tokunaga, T. K., Wan, J. M.,
and Firestone, M. K. Application of a high-density oligonucleotide microarray approach
to study bacterial population dynamics during uranium reduction and reoxidation.
Applied and Environmental Microbiology 72, 9 (Sept. 2006), 6288–98.
[13] Bryant, M. P., Wolin, E. a., Wolin, M. J., and Wolfe, R. S. Methanobacillus omelianskii,
a symbiotic association of two species of bacteria. Archiv für Mikrobiologie 59, 1 (Jan.
1967), 20–31.
[14] Burgos, W., Mcdonough, J., Senko, J., Zhang, G., Dohnalkova, a., Kelly, S., Gorby, Y.,
and Kemner, K. Characterization of uraninite nanoparticles produced by Shewanella
oneidensis MR-1. Geochimica et Cosmochimica Acta 72, 20 (Oct. 2008), 4901–4915.
[15] Campbell, C. The end of cheap oil. Scientiﬁc American 278, March (1998).
[16] Cardenas, E., Wu, W.-M., Leigh, M. B., Carley, J., Carroll, S., Gentry, T., Luo,
J., Watson, D., Gu, B., Ginder-Vogel, M., Kitanidis, P. K., Jardine, P. M., Zhou, J.,
Criddle, C. S., Marsh, T. L., and Tiedje, J. M. Microbial communities in contaminated
sediments, associated with bioremediation of uranium to submicromolar levels. Applied
and Environmental Microbiology 74, 12 (June 2008), 3718–29.
[17] Cardenas, E., Wu, W.-M., Leigh, M. B., Carley, J., Carroll, S., Gentry, T., Luo,
J., Watson, D., Gu, B., Ginder-Vogel, M., Kitanidis, P. K., Jardine, P. M., Zhou,
J., Criddle, C. S., Marsh, T. L., and Tiedje, J. M. Signiﬁcant association between
sulfate-reducing bacteria and uranium-reducing microbial communities as revealed by
a combined massively parallel sequencing-indicator species approach. Applied and
Environmental Microbiology 76, 20 (Oct. 2010), 6778–86.
[18] Carnoy, C., and Roten, C.-A. The dif/Xer recombination systems in proteobacteria.
PLoS One 4, 9 (Jan. 2009), e6531.
[19] Caspi, R., Altman, T., Dreher, K., Fulcher, C. a., Subhraveti, P., Keseler, I. M.,
Kothari, A., Krummenacker, M., Latendresse, M., Mueller, L. a., Ong, Q., Paley, S.,
Pujar, A., Shearer, A. G., Travers, M., Weerasinghe, D., Zhang, P., and Karp, P. D.
The MetaCyc database of metabolic pathways and enzymes and the BioCyc collection
of pathway/genome databases. Nucleic Acids Research 40, Database issue (Jan. 2012),
D742–53.
[20] Ceylan, A., Ertem, M., Ilcin, E., and Ozekinci, T. A special risk group for hepatitis E
infection: Turkish agricultural workers who use untreated waste water for irrigation.
Epidemiology and Infection 131, 1 (Feb. 2003), 753–756.

183

[21] Chan, C. X., Beiko, R. G., and Ragan, M. a. Lateral transfer of genes and gene
fragments in Staphylococcus extends beyond mobile elements. Journal of Bacteriology
193, 15 (Aug. 2011), 3964–77.
[22] Chatzifragkou, A., Aggelis, G., Gardeli, C., Galiotou-Panayotou, M., Komaitis, M.,
and Papanikolaou, S. Adaptation dynamics of Clostridium butyricum in high 1,3propanediol content media. Applied Microbiology and Biotechnology 95, 6 (Sept. 2012),
1541–52.
[23] Cheng, X.-Y., Li, Q., and Liu, C.-Z. Coproduction of hydrogen and methane via
anaerobic fermentation of cornstalk waste in continuous stirred tank reactor integrated
with up-ﬂow anaerobic sludge bed. Bioresource Technology 114 (June 2012), 327–33.
[24] Chern, E. C., Tsai, Y.-L., and Olson, B. H. Occurrence of genes associated with
enterotoxigenic and enterohemorrhagic Escherichia coli in agricultural waste lagoons.
Applied and Environmental Microbiology 70, 1 (Jan. 2004), 356–362.
[25] Clerget, M. Site-speciﬁc recombination promoted by a short DNA segment of plasmidR1 and by a homologous segment in the terminus region of the Escherichia coli chromosome. New Biologist 3, 8 (Aug. 1991), 780–788.
[26] Cole, J. R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R. J., Kulam-SyedMohideen, A. S., McGarrell, D. M., Marsh, T., Garrity, G. M., and Tiedje, J. M. The
Ribosomal Database Project: improved alignments and new tools for rRNA analysis.
Nucleic Acids Research 37, Database issue (Jan. 2009), D141–5.
[27] Cologgi, D. L., Lampa-Pastirk, S., Speers, A. M., Kelly, S. D., and Reguera, G. Extracellular reduction of uranium via Geobacter conductive pili as a protective cellular
mechanism. Proceedings of the National Academy of Sciences of the United States of
America 108, 37 (Sept. 2011), 15248–52.
[28] Cooper, V. S., Vohr, S. H., Wrocklage, S. C., and Hatcher, P. J. Why genes evolve
faster on secondary chromosomes in bacteria. PLoS Computational Biology 6, 4 (Apr.
2010), e1000732.
[29] Coughlan, M., and F, M. The cellulose-decomposing bacteria and their enzyme systems. In Prokaryotes: a handbook on the biology of bacteria, A. Balows, H. Truper, M. Dworkin, W. Harder, and K. Schleifer, Eds., 2 ed. Springer New York, 1992,
pp. 460–516.
[30] Cuetos, M. J., Fernández, C., Gómez, X., and Morán, A. Anaerobic co-digestion of
swine manure with energy crop residues. Biotechnology and Bioprocess Engineering
16, 5 (Sept. 2011), 1044–1052.
[31] Curtis, T. P., Sloan, W. T., and Scannell, J. W. Estimating prokaryotic diversity and
its limits. Proceedings of the National Academy of Sciences of the United States of
America 99, 16 (Aug. 2002), 10494–9.

184

[32] Dalla, E., Veeramani, H., Suvorova, E. I., Wigginton, N. S., Bargar, J. R., and Bernierlatmani, R. U(VI) reduction by spores of Clostridium acetobutylicum. Research in
Microbiologoy 161, 9 (2010), 765–771.
[33] Darling, A. C. E., Mau, B., Blattner, F. R., and Perna, N. T. Mauve: multiple
alignment of conserved genomic sequence with rearrangements. Genome Research 14,
7 (July 2004), 1394–403.
[34] Darling, A. E., Miklós, I., and Ragan, M. A. Dynamics of genome rearrangement in
bacterial populations. PLoS Genetics 4, 7 (Jan. 2008), e1000128.
[35] Davis, K. E. R., Joseph, S. J., and Janssen, P. H. Eﬀects of growth medium, inoculum
size, and incubation time on culturability and isolation of soil bacteria. Applied and
Environmental Microbiology 71, 2 (Feb. 2005), 826–34.
[36] Desvaux, M. Unravelling carbon metabolism in anaerobic cellulolytic bacteria. Biotechnology Progress 22, 5 (2006), 1229–38.
[37] Dojka, M., and Hugenholtz, P. Microbial diversity in a hydrocarbon- and chlorinatedsolvent-contaminated aquifer undergoing intrinsic bioremediation. Applied and Environmental Microbiology 64, 10 (1998), 3869–3877.
[38] Drake, H. L., Gössner, A. S., and Daniel, S. L. Old acetogens, new light. Annals of
the New York Academy of Sciences 1125 (Mar. 2008), 100–28.
[39] Esty, W. The eﬃciency of Good ’ s nonparametric coverage estimator. The Annals of
Statistics 14, 3 (1986), 1257–1260.
[40] Euzéby, J. List of prokaryotic names with standing in nomenclature, 2012.
[41] Evans, E. a., Evans, K. M., Ulrich, A., Ellsworth, S., and Khan, W. Anaerobic Processes. Water Environment Research 82, 10 (Jan. 2010), 1235–1287.
[42] Fang, C., Boe, K., and Angelidaki, I. Anaerobic co-digestion of by-products from sugar
production with cow manure. Water Research 45, 11 (May 2011), 3473–80.
[43] Fletcher, K. E., Boyanov, M. I., Thomas, S. H., Wu, Q., Kemner, K. M., and Löﬄer,
F. E. U(VI) reduction to mononuclear U(IV) by Desulﬁtobacterium species. Environmental Science and Technology 44, 12 (June 2010), 4705–9.
[44] Food and Agriculture Organization of the United Nations. FAOSTAT, 2012.
[45] Francis, A. J., Dodge, C. J., Lu, F., Halada, G. P., and Clayton, C. R. XPS and
XANES studies of uranium reduction by Clostridium sp. Environmental Science and
Technology 28, 4 (Apr. 1994), 636–9.
[46] Frear, C., Liao, W., Ewing, T., and Chen, S. Evaluation of co-digestion at a commercial
dairy anaerobic digester. CLEAN - Soil, Air, Water 39, 7 (July 2011), 697–704.

185

[47] Frey, M. Hydrogenases: hydrogen-activating enzymes. Chembiochem: A European
Journal of Chemical Biology 3, 2-3 (Mar. 2002), 153–60.
[48] Gao, P., Shi, C., Tian, J., Shi, X., Yuan, K., Lu, X., and Xu, G. Investigation on
response of the metabolites in tricarboxylic acid cycle of Escherichi coli and Pseudomonas aeruginosa to antibiotic perturbation by capillary electrophoresis. Journal of
Pharmaceutical and Biomedical Analysis 44, 1 (May 2007), 180–7.
[49] Gevers, D., Huys, G., and Swings, J. Applicability of rep-PCR ﬁngerprinting for
identiﬁcation of Lactobacillus species. FEMS Microbiology Letters 205 (2001), 31–36.
[50] Gihring, T. M., Zhang, G., Brandt, C. C., Brooks, S. C., Campbell, J. H., Carroll,
S., Criddle, C. S., Green, S. J., Jardine, P., Kostka, J. E., Lowe, K., Mehlhorn, T. L.,
Overholt, W., Watson, D. B., Yang, Z., Wu, W.-M., and Schadt, C. W. A limited
microbial consortium is responsible for extended bioreduction of uranium in a contaminated aquifer. Applied and Environmental Microbiology 77, 17 (Sept. 2011), 5955–65.
[51] Gogotova, G. I., and Vainshtein, M. B. Spre-Forming, sulfate-reducing bacterium
Desulfotomaculum Guttoideum sp. nov. Microbiology 52, 5 (1983), 618–622.
[52] Gorby, Y. a., Yanina, S., McLean, J. S., Rosso, K. M., Moyles, D., Dohnalkova, A.,
Beveridge, T. J., Chang, I. S., Kim, B. H., Kim, K. S., Culley, D. E., Reed, S. B.,
Romine, M. F., Saﬀarini, D. a., Hill, E. a., Shi, L., Elias, D. a., Kennedy, D. W.,
Pinchuk, G., Watanabe, K., Ishii, S., Logan, B., Nealson, K. H., and Fredrickson, J. K.
Electrically conductive bacterial nanowires produced by Shewanella oneidensis strain
MR-1 and other microorganisms. Proceedings of the National Academy of Sciences of
the United States of America 103, 30 (July 2006), 11358–63.
[53] Graber, J. R., Leadbetter, J. R., and Breznak, J. A. Description of Treponema azotonutricium sp. nov. and Treponema primitia sp. nov., the ﬁrst Spirochetes isolated from
termite guts. Applied and Environmental Microbiology 70, 3 (Mar. 2004), 1315–1320.
[54] Haferburg, G., and Kothe, E. Microbes and metals: interactions in the environment.
Journal of basic microbiology 47, 6 (Dec. 2007), 453–67.
[55] Hartman, H. Introductory mining engineering, 2 ed. No. 1980. Wiley, 2002.
[56] Hayes, F., and Sherratt, D. J. Recombinase binding speciﬁcity at the chromosome
dimer resolution site dif of Escherichia coli. Journal of Molecular Biology 266, 3 (Mar.
1997), 525–37.
[57] He, Z., Gentry, T. J., Schadt, C. W., Wu, L., Liebich, J., Chong, S. C., Huang, Z.,
Wu, W., Gu, B., Jardine, P., Criddle, C., and Zhou, J. GeoChip: a comprehensive
microarray for investigating biogeochemical, ecological and environmental processes.
The ISME Journal 1, 1 (May 2007), 67–77.
[58] He, Z., Piceno, Y., Deng, Y., Xu, M., Lu, Z., Desantis, T., Andersen, G., Hobbie,
S. E., Reich, P. B., and Zhou, J. The phylogenetic composition and structure of soil

186

microbial communities shifts in response to elevated carbon dioxide. The ISME Journal
6, 2 (Feb. 2012), 259–72.
[59] Hedegaard, M., and Jaensch, V. Anaerobic co-digestion of urban and rural wastes.
Renewable Energy 16 (1999).
[60] Herrera, L. K., and Videla, H. a. Role of iron-reducing bacteria in corrosion and
protection of carbon steel. International Biodeterioration and Biodegradation 63, 7
(Oct. 2009), 891–895.
[61] Heukelekian, H. Activated sludge bulking. Sewage Works Journal 13, 1 (1941), 39–42.
[62] Holt, R. A., Cairns, A. J., and Morris, J. G. Production of butanol by Clostridium
puniceum in batch and continuous culture. Applied Microbiology and Biotechnology
27, 4 (Jan. 1988), 319–324.
[63] Huber, K. E., and Waldor, M. K. Filamentous phage integration requires the host
recombinases XerC and XerD. Nature 417, 6889 (June 2002), 656–9.
[64] Hwang, C., Wu, W., Gentry, T. J., Carley, J., Corbin, G. A., Carroll, S. L., Watson,
D. B., Jardine, P. M., Zhou, J., Criddle, C. S., and Fields, M. W. Bacterial community
succession during in situ uranium bioremediation: spatial similarities along controlled
ﬂow paths. The ISME Journal 3, 1 (Jan. 2009), 47–64.
[65] Jahn, M. K., Haderlein, S. B., and Meckenstock, R. U. Reduction of Prussian Blue
by the two iron-reducing microorganisms Geobacter metallireducens and Shewanella
alga. Environmental Microbiology 8, 2 (Feb. 2006), 362–7.
[66] Jakobsson, K., Bentley, R., Söderbergh, B., and Aleklett, K. The end of cheap oil:
Bottom-up economic and geologic modeling of aggregate oil production curves. Energy
Policy 41 (Feb. 2012), 860–870.
[67] Janssen, P. J., Van Houdt, R., Moors, H., Monsieurs, P., Morin, N., Michaux, A.,
Benotmane, M. a., Leys, N., Vallaeys, T., Lapidus, A., Monchy, S., Médigue, C.,
Taghavi, S., McCorkle, S., Dunn, J., van der Lelie, D., and Mergeay, M. The complete
genome sequence of Cupriavidus metallidurans strain CH34, a master survivalist in
harsh and anthropogenic environments. PLoS One 5, 5 (Jan. 2010), e10433.
[68] Jensen, R. B. Analysis of the terminus region of the Caulobacter crescentus chromosome and identiﬁcation of the dif site. Journal of Bacteriology 188, 16 (Aug. 2006),
6016–9.
[69] Jeon, O.-H., Kelly, S. D., Kemner, K. M., Barnett, M. O., Burgos, W. D., Dempsey,
B. a., and Roden, E. E. Microbial reduction of U(VI) at the solid-water interface.
Environmental Science and Technology 38, 21 (Nov. 2004), 5649–55.
[70] Joshi, S., Parhad, N., and Rao, N. Elimination of Salmonella in stabilization ponds.
Water Research 7 (1973), 1357–1365.

187

[71] Julian, D. J., Kershaw, C. J., Brown, N. L., and Hobman, J. L. Transcriptional
activation of MerR family promoters in Cupriavidus metallidurans CH34. Antonie van
Leeuwenhoek 96, 2 (Aug. 2009), 149–59.
[72] Junier, P., Frutschi, M., Wigginton, N. S., Schoﬁeld, E. J., Bargar, J. R., and BernierLatmani, R. Metal reduction by spores of Desulfotomaculum reducens. Environmental
Microbiology 11, 12 (Dec. 2009), 3007–17.
[73] Kanehisa, M., Goto, S., Sato, Y., Furumichi, M., and Tanabe, M. KEGG for integration and interpretation of large-scale molecular data sets. Nucleic Acids Research 40,
Database issue (Jan. 2012), D109–14.
[74] Kato, S., Haruta, S., Cui, Z. J., Ishii, M., and Igarashi, Y. Eﬀective cellulose degradation by a mixed-culture system composed of a cellulolytic Clostridium and aerobic
non-cellulolytic bacteria. FEMS Microbiology Ecology 51, 1 (Dec. 2004), 133–42.
[75] Kato, S., Haruta, S., Cui, Z. J., Ishii, M., and Igarashi, Y. Network relationships of
bacteria in a stable mixed culture. Microbial ecology 56, 3 (Oct. 2008), 403–11.
[76] Kato, S., Hashimoto, K., and Watanabe, K. Microbial interspecies electron transfer via
electric currents through conductive minerals. Proceedings of the National Academy of
Sciences of the United States of America (June 2012), 22–23.
[77] Kelly, S. D., Kemner, K. M., Carley, J., Criddle, C., Jardine, P. M., Marsh, T. L.,
Phillips, D., Watson, D., and Wu, W.-M. Speciation of uranium in sediments before
and after in situ biostimulation. Environmental Science and Technology 42, 5 (Mar.
2008), 1558–64.
[78] Kiely, G., Tayfur, G., and Dolan, C. Physical and mathematical modelling of anaerobic
digestion of organic wastes. Water Research 31, 3 (1997), 534–540.
[79] Komlos, J., and Jaﬀé, P. R. Eﬀect of iron bioavailability on dissolved hydrogen concentrations during microbial iron reduction. Biodegradation 15, 5 (Oct. 2004), 315–25.
[80] Konstantinidis, K. T., Isaacs, N., Fett, J., Simpson, S., Long, D. T., and Marsh,
T. L. Microbial diversity and resistance to copper in metal-contaminated lake sediment.
Microbial ecology 45, 2 (Feb. 2003), 191–202.
[81] Kotrba, P. Microbial Biosorption of Metals, 1 ed. Springer New York, 2011.
[82] Labrenz, M. Formation of Sphalerite (ZnS) Deposits in Natural Bioﬁlms of SulfateReducing Bacteria. Science 290, 5497 (Dec. 2000), 1744–1747.
[83] Lednicka, D., Mergaert, J., Cnockaert, M., and Swings, J. Isolation and identiﬁcation of
cellulolytic bacteria involved in the degradation of natural cellulosic ﬁbres. Systematic
and Applied Microbiology 23 (2000), 292–299.
[84] Lefebvre, O., Nguyen, T. T. H., Al-Mamun, a., Chang, I. S., and Ng, H. Y. T-RFLP
reveals high β-Proteobacteria diversity in microbial fuel cells enriched with domestic
wastewater. Journal of applied microbiology 109, 3 (Sept. 2010), 839–50.
188

[85] Lehtomäki, A., Huttunen, S., and Rintala, J. Laboratory investigations on co-digestion
of energy crops and crop residues with cow manure for methane production: Eﬀect of
crop to manure ratio. Resources, Conservation and Recycling 51, 3 (Sept. 2007), 591–
609.
[86] Lovley, D. R., and Phillips, E. J. Organic matter mineralization with reduction of
ferric iron in anaerobic sediments. Applied and Environmental Microbiology 51, 4
(Apr. 1986), 683–9.
[87] Lovley, D. R., and Phillips, E. J. Novel mode of microbial energy metabolism: organic
carbon oxidation coupled to dissimilatory reduction of iron or manganese. Applied and
Environmental Microbiology 54, 6 (June 1988), 1472–80.
[88] Lovley, D. R., and Phillips, E. J. Bioremediation of uranium contamination with
enzymatic uranium reduction. Environmental Science and Technology 26, 11 (1992),
2228–2234.
[89] Lovley, D. R., and Phillips, E. J. Reduction of uranium by Desulfovibrio desulfuricans.
Applied and Environmental Microbiology 58, 3 (Mar. 1992), 850–6.
[90] Lu, W.-B., Shi, J.-J., Wang, C.-H., and Chang, J.-S. Biosorption of lead, copper and
cadmium by an indigenous isolate Enterobacter sp. J1 possessing high heavy-metal
resistance. Journal of Hazardous Materials 134, 1-3 (June 2006), 80–6.
[91] Macaskie, L. E., Empson, R. M., Cheetham, a. K., Grey, C. P., and Skarnulis, a. J.
Uranium bioaccumulation by a Citrobacter sp. as a result of enzymically mediated
growth of polycrystalline HUO2PO4. Science (New York, N.Y.) 257, 5071 (Aug. 1992),
782–4.
[92] Madigan, M. T., Martinko, J. M., Stahl, D., and Clark, D. P. Brock Biology of Microorganisms, 13 ed. Benjamin Cummings, 2010.
[93] Magot, M., Fardeau, M.-L., Arnauld, O., Lanau, C., Ollivier, B., Thomas, P., and Patel, B. Spirochaeta smaragdinae sp. nov., a new mesophilic strictly anaerobic spirochete
from an oil ﬁeld. FEMS Microbiology Letters 155, 2 (Jan. 1997), 185–191.
[94] Magurran, A. E. Measuring Biological Diversity. Malden, Ma.: Blackwell Pub., 2004.
[95] Malik, A. Metal bioremediation through growing cells. Environment International 30,
2 (Apr. 2004), 261–78.
[96] Martin, W. F. Hydrogen, metals, bifurcating electrons, and proton gradients: the early
evolution of biological energy conservation. FEBS letters 586, 5 (Mar. 2012), 485–93.
[97] Martinez, R. J., Beazley, M. J., Taillefert, M., Arakaki, A. K., Skolnick, J., and
Sobecky, P. a. Aerobic uranium (VI) bioprecipitation by metal-resistant bacteria isolated from radionuclide- and metal-contaminated subsurface soils. Environmental Microbiology 9, 12 (Dec. 2007), 3122–33.

189

[98] Martinez, R. J., Wang, Y., Raimondo, M. A., Coombs, J. M., Barkay, T., and Sobecky,
P. A. Horizontal gene transfer of PIB-type ATPases among bacteria isolated from
radionuclide- and metal-contaminated subsurface soils. Applied and environmental
microbiology 72, 5 (May 2006), 3111–8.
[99] Matson, E. G., Gora, K. G., and Leadbetter, J. R. Anaerobic carbon monoxide dehydrogenase diversity in the homoacetogenic hindgut microbial communities of lower
termites and the wood roach. PLoS One 6, 4 (Jan. 2011), e19316.
[100] McDaniels, T. L., Axelrod, L. J., Cavanagh, N. S., and Slovic, P. Perception of
ecological risk to water environments. Risk Analysis : An Oﬃcial Publication of The
Society for Risk Analysis 17, 3 (June 1997), 341–52.
[101] Menz, G., Vriesekoop, F., Zarei, M., Zhu, B., and Aldred, P. The growth and survival
of food-borne pathogens in sweet and fermenting brewers’ wort. International Journal
of Food Microbiology 140, 1 (May 2010), 19–25.
[102] Merroun, M. L., Nedelkova, M., Ojeda, J. J., Reitz, T., Fernández, M. L., Arias, J. M.,
Romero-González, M., and Selenska-Pobell, S. Bio-precipitation of uranium by two
bacterial isolates recovered from extreme environments as estimated by potentiometric
titration, TEM and X-ray absorption spectroscopic analyses. Journal of Hazardous
Materials 197 (Dec. 2011), 1–10.
[103] Meyer, J. [FeFe] hydrogenases and their evolution: a genomic perspective. Cellular
and Molecular Life Sciences 64, 9 (May 2007), 1063–84.
[104] Mijnendonckx, K., Provoost, A., Monsieurs, P., Leys, N., Mergeay, M., Mahillon, J.,
and Van Houdt, R. Insertion sequence elements in Cupriavidus metallidurans CH34:
distribution and role in adaptation. Plasmid 65, 3 (May 2011), 193–203.
[105] Mikolay, A., and Nies, D. H. The ABC-transporter AtmA is involved in nickel and
cobalt resistance of Cupriavidus metallidurans strain CH34. Antonie van Leeuwenhoek
96, 2 (Aug. 2009), 183–91.
[106] Mulligan, C., Yong, R., and Gibbs, B. Remediation technologies for metalcontaminated soils and groundwater: an evaluation. Engineering Geology 60, 1-4 (June
2001), 193–207.
[107] Mullings, R., and Parish, J. Mesophilic aerobic Gram negative cellulose degrading
bacteria from aquatic habitats and soils. Journal of Applied Microbiology, 1981 (1984).
[108] Neidhardt, F. Escherichia coli and Salmonella: cellular and molecular biology, 2 ed.
ASM Press, Washington, DC„ 1996.
[109] Neilson, L., Blakely, G., and Sherratt, D. Site speciﬁc recombination at dif by
Haemophilus inﬂuenzae XerC. Molecular Microbiology 31, 3 (2002), 915–926.
[110] Nges, I. A., Escobar, F., Fu, X., and Björnsson, L. Beneﬁts of supplementing an
industrial waste anaerobic digester with energy crops for increased biogas production.
Waste Management (New York, N.Y.) 32, 1 (Jan. 2012), 53–9.
190

[111] Nies, D. H. Heavy metal-resistant bacteria as extremophiles: molecular physiology
and biotechnological use of Ralstonia sp. CH34. Extremophiles: Life Under Extreme
Conditions 4, 2 (Apr. 2000), 77–82.
[112] Nolan, B. T., Ruddy, B. C., Hitt, K. J., and Helsel, D. R. Risk of Nitrate in Groundwaters of the United States - A National Perspective. Environmental Science and
Technology 31, 8 (Aug. 1997), 2229–2236.
[113] Nyman, J. L., Wu, H.-I., Gentile, M. E., Kitanidis, P. K., and Criddle, C. S. Inhibition of a U(VI)- and sulfate-reducing consortia by U(VI). Environmental Science and
Technology 41, 18 (Sept. 2007), 6528–33.
[114] Ochman, H., Lawrence, J. G., and Groisman, E. a. Lateral gene transfer and the
nature of bacterial innovation. Nature 405, 6784 (May 2000), 299–304.
[115] Ogi, T., Tamaoki, K., Saitoh, N., Higashi, A., and Konishi, Y. Recovery of indium
from aqueous solutions by the Gram-negative bacterium Shewanella algae. Biochemical
Engineering Journal 63 (Apr. 2012), 129–133.
[116] Ohmiya, K. Anaerobic Bacterial Degradation for the Eﬀective Utilization of Biomass.
Biotechnology and Bioprocess . . . (2005).
[117] Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., O’Hara, R. B.,
Simpson, G. L., Solymos, P., Stevens, M. H. H., and Wagner, H. Vegan: Community
Ecology Package, 2012.
[118] Oregaard, G., and Sø rensen, S. r. J. High diversity of bacterial mercuric reductase
genes from surface and sub-surface ﬂoodplain soil (Oak Ridge, USA). The ISME
Journal 1, 5 (Sept. 2007), 453–67.
[119] O’Sullivan, C. a., Burrell, P. C., Clarke, W. P., and Blackall, L. L. Structure of a
cellulose degrading bacterial community during anaerobic digestion. Biotechnology
and Bioengineering 92, 7 (Dec. 2005), 871–8.
[120] Pagés Díaz, J., Pereda Reyes, I., Lundin, M., and Sárvári Horváth, I. Co-digestion of
diﬀerent waste mixtures from agro-industrial activities: Kinetic evaluation and synergetic eﬀects. Bioresource Technology 102, 23 (Dec. 2011), 10834–40.
[121] Pagnanelli, F., Cruz Viggi, C., and Toro, L. Isolation and quantiﬁcation of cadmium removal mechanisms in batch reactors inoculated by sulphate reducing bacteria:
biosorption versus bioprecipitation. Bioresource Technology 101, 9 (May 2010), 2981–7.
[122] Pan, J.-h., Liu, R.-x., and Tang, H.-x. Surface reaction of Bacillus cereus biomass and
its biosorption for lead and copper ions. Journal of Environmental Sciences (China)
19, 4 (Jan. 2007), 403–8.
[123] Passot, F. M., Calderon, V., Fichant, G., Lane, D., and Pasta, F. Centromere binding
and evolution of chromosomal partition systems in the Burkholderiales. Journal of
Bacteriology 194, 13 (July 2012), 3426–36.
191

[124] Peña, A., Teeling, H., Huerta-Cepas, J., Santos, F., Yarza, P., Brito-Echeverría, J.,
Lucio, M., Schmitt-Kopplin, P., Meseguer, I., Schenowitz, C., Dossat, C., Barbe, V.,
Dopazo, J., Rosselló-Mora, R., Schüler, M., Glöckner, F. O., Amann, R., Gabaldón, T.,
and Antón, J. Fine-scale evolution: genomic, phenotypic and ecological diﬀerentiation
in two coexisting Salinibacter ruber strains. The ISME Journal 4, 7 (July 2010),
882–95.
[125] Petit, J., Jouzel, J., Raynaud, D., and Barkov, N. Climate and atmospheric history of
the past 420,000 years from the Vostok ice core, Antarctica. Nature (1999).
[126] Pobeheim, H., Munk, B., Müller, H., Berg, G., and Guebitz, G. M. Characterization
of an anaerobic population digesting a model substrate for maize in the presence of
trace metals. Chemosphere 80, 8 (Aug. 2010), 829–36.
[127] Qin, J., Li, R., Raes, J., Arumugam, M., Burgdorf, K. S., Manichanh, C., Nielsen,
T., Pons, N., Levenez, F., Yamada, T., Mende, D. R., Li, J., Xu, J., Li, S., Li, D.,
Cao, J., Wang, B., Liang, H., Zheng, H., Xie, Y., Tap, J., Lepage, P., Bertalan, M.,
Batto, J.-M., Hansen, T., Le Paslier, D., Linneberg, A., Nielsen, H. B. r., Pelletier,
E., Renault, P., Sicheritz-Ponten, T., Turner, K., Zhu, H., Yu, C., Li, S., Jian, M.,
Zhou, Y., Li, Y., Zhang, X., Li, S., Qin, N., Yang, H., Wang, J., Brunak, S. r., Doré,
J., Guarner, F., Kristiansen, K., Pedersen, O., Parkhill, J., Weissenbach, J., Bork, P.,
Ehrlich, S. D., and Wang, J. A human gut microbial gene catalogue established by
metagenomic sequencing. Nature 464, 7285 (Mar. 2010), 59–65.
[128] R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, 2012.
[129] Ray, A. E., Bargar, J. R., Sivaswamy, V., Dohnalkova, A. C., Fujita, Y., Peyton,
B. M., and Magnuson, T. S. Evidence for multiple modes of uranium immobilization
by an anaerobic bacterium. Geochimica et Cosmochimica Acta 75, 10 (May 2011),
2684–2695.
[130] Renninger, N., McMahon, K. D., Knopp, R., Nitsche, H., Clark, D. S., and Keasling,
J. D. Uranyl precipitation by biomass from an enhanced biological phosphorus removal
reactor. Biodegradation 12, 6 (Jan. 2001), 401–10.
[131] Reynolds, T. D., and Richards, P. A. Unit Operations and Processes in Environmental
Engineering, 2nd ed. PWS Publishing Company, Boston, MA, 1996.
[132] Roden, E. E., and Lovley, D. R. Dissimilatory Fe(III) Reduction by the Marine Microorganism Desulfuromonas acetoxidans. Applied and Environmental Microbiology
59, 3 (Mar. 1993), 734–42.
[133] Rodríguez-Tirado, V., Green-Ruiz, C., and Gómez-Gil, B. Cu and Pb biosorption on
Bacillus thioparans strain U3 in aqueous solution: Kinetic and equilibrium studies.
Chemical Engineering Journal 181-182 (Feb. 2012), 352–359.

192

[134] Roest, K., Heilig, H. G., Smidt, H., de Vos, W. M., Stams, A. J., and Akkermans,
A. D. Community analysis of a full-scale anaerobic bioreactor treating paper mill
wastewater. Systematic and Applied Microbiology 28, 2 (Mar. 2005), 175–185.
[135] Sandegren, L., and Andersson, D. I. Bacterial gene ampliﬁcation: implications for
the evolution of antibiotic resistance. Nature Reviews. Microbiology 7, 8 (Aug. 2009),
578–88.
[136] Sar, P., Kazy, S. K., and Singh, S. P. Intracellular nickel accumulation by Pseudomonas
aeruginosa and its chemical nature. Letters in Applied Microbiology 32, 4 (Apr. 2001),
257–61.
[137] Sattley, W. M., Jung, D. O., and Madigan, M. T. Psychrosinus fermentans gen.
nov., sp. nov., a lactate-fermenting bacterium from near-freezing oxycline waters of a
meromictic Antarctic lake. FEMS Microbiology Letters 287, 1 (Oct. 2008), 121–7.
[138] Scherer, J., and Nies, D. H. CzcP is a novel eﬄux system contributing to transition
metal resistance in Cupriavidus metallidurans CH34. Molecular Microbiology 73, 4
(Aug. 2009), 601–21.
[139] Schiewer, S., and Volesky, B. Biosorption process for heavy metal removal. In Environmental Microbe-Metal Interactions, D. R. Lovley, Ed. ASM Press, Washington,
DC„ 14, 2000, p. 329.
[140] Schink, B., and Stams, A. J. M. Syntrophism among prokaryotes. In The Prokaryotes,
M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt, Eds.
Springer New York, 2006, pp. 309–335.
[141] Schloter, M., Lebuhn, M., Heulin, T., and Hartmann, a. Ecology and evolution of
bacterial microdiversity. FEMS Microbiology Reviews 24, 5 (Dec. 2000), 647–60.
[142] Schlüter, A., Bekel, T., Diaz, N. N., Dondrup, M., Eichenlaub, R., Gartemann, K.H., Krahn, I., Krause, L., Krömeke, H., Kruse, O., Mussgnug, J. H., Neuweger, H.,
Niehaus, K., Pühler, A., Runte, K. J., Szczepanowski, R., Tauch, A., Tilker, A.,
Viehöver, P., and Goesmann, A. The metagenome of a biogas-producing microbial community of a production-scale biogas plant fermenter analysed by the 454pyrosequencing technology. Journal of biotechnology 136, 1-2 (Aug. 2008), 77–90.
[143] Schwarz, J. I. K., Eckert, W., and Conrad, R. Community structure of Archaea and
Bacteria in a profundal lake sediment Lake Kinneret (Israel). Systematic and Applied
Microbiology 30, 3 (Apr. 2007), 239–54.
[144] Sciochetti, S. A., Piggot, P. J., and Blakely, G. W. Identiﬁcation and characterization
of the dif Site from Bacillus subtilis. Journal of Bacteriology 183, 3 (Feb. 2001), 1058–
68.
[145] Sherif, F., and Awad, A. Spectrophotometric determination of uranium (VI) by the
azide reaction. Analytica Chimica Acta 26 (1962), 235–241.

193

[146] Shoeb, E., Badar, U., Akhter, J., Shams, H., Sultana, M., and Ansari, M. a. Horizontal
gene transfer of stress resistance genes through plasmid transport. World Journal of
Microbiology and Biotechnology 28, 3 (Mar. 2012), 1021–5.
[147] Silver, S., and Phung, L. T. A bacterial view of the periodic table: genes and proteins
for toxic inorganic ions. Journal of Industrial Microbiology and Biotechnology 32, 11-12
(Dec. 2005), 587–605.
[148] Sitte, J., Akob, D. M., Kaufmann, C., Finster, K., Banerjee, D., Burkhardt, E.-M.,
Kostka, J. E., Scheinost, A. C., Büchel, G., and Küsel, K. Microbial links between
sulfate reduction and metal retention in uranium- and heavy metal-contaminated soil.
Applied and Environmental Microbiology 76, 10 (May 2010), 3143–52.
[149] Slater, S. C., Goldman, B. S., Goodner, B., Setubal, J. a. C., Farrand, S. K., Nester,
E. W., Burr, T. J., Banta, L., Dickerman, A. W., Paulsen, I., Otten, L., Suen, G.,
Welch, R., Almeida, N. F., Arnold, F., Burton, O. T., Du, Z., Ewing, A., Godsy,
E., Heisel, S., Houmiel, K. L., Jhaveri, J., Lu, J., Miller, N. M., Norton, S., Chen,
Q., Phoolcharoen, W., Ohlin, V., Ondrusek, D., Pride, N., Stricklin, S. L., Sun, J.,
Wheeler, C., Wilson, L., Zhu, H., and Wood, D. W. Genome sequences of three
agrobacterium biovars help elucidate the evolution of multichromosome genomes in
bacteria. Journal of Bacteriology 191, 8 (Apr. 2009), 2501–11.
[150] Sommer, M. O. a., Dantas, G., and Church, G. M. Functional characterization of the
antibiotic resistance reservoir in the human microﬂora. Science (New York, N.Y.) 325,
5944 (Aug. 2009), 1128–31.
[151] Stams, A. J. M., de Bok, F. a. M., Plugge, C. M., van Eekert, M. H. a., Dolﬁng,
J., and Schraa, G. Exocellular electron transfer in anaerobic microbial communities.
Environmental Microbiology 8, 3 (Mar. 2006), 371–82.
[152] Stams, A. J. M., de Bok, F. a. M., Plugge, C. M., van Eekert, M. H. a., Dolﬁng, J.,
Schraa, G., Bok, F. A. M. D., and Eekert, M. H. A. V. Exocellular electron transfer
in anaerobic microbial communities. Environmental Microbiology 8, 3 (Mar. 2006),
371–82.
[153] Stern, D., and Kaufmann, R. Annual estimates of global anthropogenic methane
emissions: 1860-1994. Trends Online: A Compendium of Data on Global Change. Tech.
rep., Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory,
U.S. Department of Energy, Oak Ridge, Tenn., U.S.A., 1998.
[154] Stookey, L. L. Ferrozine—a new spectrophotometric reagent for iron. Analytical Chemistry 42, 7 (June 1970), 779–781.
[155] Stursová, M., Zifčáková, L., Leigh, M. B., Burgess, R., and Baldrian, P. Cellulose
utilization in forest litter and soil: identiﬁcation of bacterial and fungal decomposers.
FEMS Microbiology Ecology 80, 3 (June 2012), 735–46.

194

[156] Sugisawa, T., Miyazaki, T., and Hoshino, T. Microbial Production of L-ascorbic acid
from D-sorbitol, L-sorbose, L-gulose, and L-sorbosone by Ketogulonicigenium vulgare
DSM 4025. Bioscience, Biotechnology, and Biochemistry 69, 3 (2005), 659–662.
[157] Sul, W. J., Cole, J. R., Jesus, E. D. C., Wang, Q., Farris, R. J., Fish, J. a., and Tiedje,
J. M. Bacterial community comparisons by taxonomy-supervised analysis independent
of sequence alignment and clustering. Proceedings of the National Academy of Sciences
of the United States of America 108, 35 (Aug. 2011), 14637–42.
[158] Sundararaj, S., Guo, A., Habibi-Nazhad, B., Rouani, M., Stothard, P., Ellison, M.,
and Wishart, D. S. The CyberCell Database (CCDB): a comprehensive, self-updating,
relational database to coordinate and facilitate in silico modeling of Escherichia coli.
Nucleic Acids Research 32, Database issue (Jan. 2004), D293–5.
[159] Swinton, J. Vennerable: Venn and Euler area-proportional diagrams.
[160] Taghavi, S., Lesaulnier, C., Monchy, S., Wattiez, R., Mergeay, M., and van der Lelie,
D. Lead(II) resistance in Cupriavidus metallidurans CH34: interplay between plasmid
and chromosomally-located functions. Antonie van Leeuwenhoek 96, 2 (Aug. 2009),
171–82.
[161] Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., and Kumar, S. MEGA5:
molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28, 10
(Oct. 2011), 2731–9.
[162] Tarlera, S., Muxí, L., Soubes, M., and Stams, a. J. Caloramator proteoclasticus sp.
nov., a new moderately thermophilic anaerobic proteolytic bacterium. International
Journal of Systematic Bacteriology 47, 3 (July 1997), 651–6.
[163] Truex, M. J., Peyton, B. M., Valentine, N. B., and Gorby, Y. a. Kinetics of U(VI)
reduction by a dissimilatory Fe(III)-reducing bacterium under non-growth conditions.
Biotechnology and Bioengineering 55, 3 (Aug. 1997), 490–6.
[164] Tsuruta, T. Microbes and microbial technology. In Microbes and Microbial Technology,
I. Ahmad, F. Ahmad, and J. Pichtel, Eds. Springer New York, New York, NY, 2011,
pp. 267–281.
[165] Tunali, S., Çabuk, A., and Akar, T. Removal of lead and copper ions from aqueous
solutions by bacterial strain isolated from soil. Chemical Engineering Journal 115, 3
(Jan. 2006), 203–211.
[166] U.S. Environmental Protection Agency. Agriculture, 2012.
[167] U.S. Environmental Protection Agency. NPL Mining Sites, 2012.
[168] U.S. Environmental Protection Agency. NPL Site Status Information, 2012.
[169] U.S. Environmental Protection Agency. Torch Lake, 2012.

195

[170] Val, M.-E., Kennedy, S. P., El Karoui, M., Bonné, L., Chevalier, F., and Barre, F.X. FtsK-dependent dimer resolution on multiple chromosomes in the pathogen Vibrio
cholerae. PLoS Genetics 4, 9 (Jan. 2008), e1000201.
[171] Van Houdt, R., Monsieurs, P., Mijnendonckx, K., Provoost, A., Janssen, A., Mergeay,
M., and Leys, N. Variation in genomic islands contribute to genome plasticity in
Cupriavidus metallidurans. BMC Genomics 13, 1 (Jan. 2012), 111.
[172] Van Nostrand, J. D., Wu, W.-M., Wu, L., Deng, Y., Carley, J., Carroll, S., He, Z.,
Gu, B., Luo, J., Criddle, C. S., Watson, D. B., Jardine, P. M., Marsh, T. L., Tiedje,
J. M., Hazen, T. C., and Zhou, J. GeoChip-based analysis of functional microbial
communities during the reoxidation of a bioreduced uranium-contaminated aquifer.
Environmental Microbiology 11, 10 (Oct. 2009), 2611–26.
[173] Verberkmoes, N. C., Russell, A. L., Shah, M., Godzik, A., Rosenquist, M., Halfvarson,
J., Lefsrud, M. G., Apajalahti, J., Tysk, C., Hettich, R. L., and Jansson, J. K. Shotgun
metaproteomics of the human distal gut microbiota. The ISME Journal 3, 2 (Feb.
2009), 179–89.
[174] Versalovic, J., Schneider, M., De Bruijn, F. J., and Lupski, J. R. Genomic ﬁngerprinting of bacteria using repetitive sequence-based polymerase chain reaction. Methods in
Molecular and Cellular Biology 5, 1 (1994), 25–40.
[175] Verstraete, W., Wittebolle, L., Heylen, K., Vanparys, B., de Vos, P., van de Wiele,
T., and Boon, N. Microbial resource management: the road to go for environmental
biotechnology. Engineering in Life Sciences 7, 2 (Apr. 2007), 117–126.
[176] Vignais, P., Billoud, B., and Meyer, J. Classiﬁcation and phylogeny of hydrogenases.
FEMS Microbiology Reviews 25 (2001).
[177] Volesky, B. Detoxiﬁcation of metal-bearing eﬄuents: biosorption for the next century.
Hydrometallurgy 59, 2-3 (Feb. 2001), 203–216.
[178] Volesky, B., and Holan, Z. R. Biosorption of heavy metals. Biotechnology Progress 11,
3 (1995), 235–50.
[179] Wang, H., Lehtomäki, A., Tolvanen, K., Puhakka, J., and Rintala, J. Impact of crop
species on bacterial community structure during anaerobic co-digestion of crops and
cow manure. Bioresource Technology 100, 7 (Apr. 2009), 2311–5.
[180] Wang, H., Tolvanen, K., Lehtomäki, A., Puhakka, J., and Rintala, J. Microbial community structure in anaerobic co-digestion of grass silage and cow manure in a laboratory
continuously stirred tank reactor. Biodegradation 21, 1 (Feb. 2010), 135–46.
[181] Wang, H., Vuorela, M., Keränen, A.-L., Lehtinen, T. M., Lensu, A., Lehtomäki, A.,
and Rintala, J. Development of microbial populations in the anaerobic hydrolysis of
grass silage for methane production. FEMS Microbiology Ecology 72, 3 (June 2010),
496–506.

196

[182] Werner, J. J., Knights, D., Garcia, M. L., Scalfone, N. B., Smith, S., Yarasheski, K.,
Cummings, T. a., Beers, A. R., Knight, R., and Angenent, L. T. Bacterial community
structures are unique and resilient in full-scale bioenergy systems. Proceedings of the
National Academy of Sciences of the United States of America 108, 10 (Mar. 2011),
4158–63.
[183] White, C., Sharman, A. K., and Gadd, G. M. An integrated microbial process for
the bioremediation of soil contaminated with toxic metals. Nature Biotechnology 16, 6
(June 1998), 572–5.
[184] Whitman, W. B., Coleman, D. C., and Wiebe, W. J. Perspective prokaryotes : the
unseen majority. Proceedings of the National Academy of Sciences 95, June (1998),
6578–6583.
[185] Wirth, R., Kovács, E., Maróti, G., Bagi, Z., Rákhely, G., and Kovács, K. L. Characterization of a biogas-producing microbial community by short-read next generation
DNA sequencing. Biotechnology for Biofuels 5, 1 (June 2012), 41.
[186] Wu, Q., Sanford, R., and Löﬄer, F. Uranium (VI) reduction by Anaeromyxobacter
dehalogenans strain 2CP-C. Applied and Environmental Microbiology 72, 5 (2006),
3608–3614.
[187] Wu, W.-M., Carley, J., Fienen, M., Mehlhorn, T., Lowe, K., Nyman, J., Luo, J.,
Gentile, M. E., Rajan, R., Wagner, D., Hickey, R. F., Gu, B., Watson, D., Cirpka, O. a.,
Kitanidis, P. K., Jardine, P. M., and Criddle, C. S. Pilot-scale in situ bioremediation
of uranium in a highly contaminated aquifer. 1. Conditioning of a treatment zone.
Environmental Science and Technology 40, 12 (June 2006), 3978–85.
[188] Wu, W.-M., Carley, J., Gentry, T., Ginder-Vogel, M. a., Fienen, M., Mehlhorn, T.,
Yan, H., Caroll, S., Pace, M. N., Nyman, J., Luo, J., Gentile, M. E., Fields, M. W.,
Hickey, R. F., Gu, B., Watson, D., Cirpka, O. a., Zhou, J., Fendorf, S., Kitanidis, P. K.,
Jardine, P. M., and Criddle, C. S. Pilot-scale in situ bioremedation of uranium in a
highly contaminated aquifer. 2. Reduction of u(VI) and geochemical control of u(VI)
bioavailability. Environmental Science and Technology 40, 12 (June 2006), 3986–95.
[189] Xing, D., Ren, N., and Rittmann, B. E. Genetic diversity of hydrogen-producing
bacteria in an acidophilic ethanol-H2-coproducing system, analyzed using the [Fe]hydrogenase gene. Applied and Environmental Microbiology 74, 4 (Feb. 2008), 1232–9.
[190] Xu, M., Wu, W.-M., Wu, L., He, Z., Van Nostrand, J. D., Deng, Y., Luo, J., Carley,
J., Ginder-Vogel, M., Gentry, T. J., Gu, B., Watson, D., Jardine, P. M., Marsh,
T. L., Tiedje, J. M., Hazen, T., Criddle, C. S., and Zhou, J. Responses of microbial
community functional structures to pilot-scale uranium in situ bioremediation. The
ISME Journal 4, 8 (Aug. 2010), 1060–70.
[191] Yang, F., Pecina, D. a., Kelly, S. D., Kim, S.-H., Kemner, K. M., Long, D. T., and
Marsh, T. L. Biosequestration via cooperative binding of copper by Ralstonia pickettii.
Environmental Technology 31, 8-9 (2010), 1045–60.
197

[192] Yue, Z., Chen, R., Yang, F., MacLellan, J., Marsh, T., Liu, Y., and Liao, W. Eﬀects
of dairy manure and corn stover co-digestion on anaerobic microbes and corresponding
digestion performance. Bioresource Technology 128 (Nov. 2012), 65–71.
[193] Zengler, K. Central role of the cell in microbial ecology. Microbiology and Molecular
Biology Reviews 73, 4 (Dec. 2009), 712–29.
[194] Zhang, W., Culley, D. E., Scholten, J. C. M., Hogan, M., Vitiritti, L., and Brockman,
F. J. Global transcriptomic analysis of Desulfovibrio vulgaris on diﬀerent electron
donors. Antonie van Leeuwenhoek 89, 2 (Feb. 2006), 221–37.
[195] Zoropogui, A., Gambarelli, S., and Covès, J. CzcE from Cupriavidus metallidurans
CH34 is a copper-binding protein. Biochemical and Biophysical Research Communications 365, 4 (Jan. 2008), 735–9.

198