University of Southern California Dissertations and Theses

Survival and evolution of Shewanella oneidensis MR-1: applications for microbial fuel cells

(USC Thesis Other)

Survival and evolution of Shewanella oneidensis MR-1: applications for microbial fuel cells

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2012 Meghann Adrienne Ribbens

SURVIVAL AND EVOLUTION OF SHEWANELLA ONEIDENSIS MR-1:
APPLICATIONS FOR MICROBIAL FUEL CELLS

by
Meghann Adrienne Ribbens

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MOLECULAR BIOLOGY)

August 2012

ii

Dedication

I dedicate this dissertation to my friends and family. Ann and AJ are without a doubt the
best parents ever. They’ve always been so supportive, especially of my life-long interest
in science. I have always been completely confident that, no matter what, they are there
for me, and I love them so much. Noah and Cathy are my two best friends from forever,
and they have been so generous with their love and support, even from afar. Thank you
all so much.

iii

Acknowledgements

I would like to thank my advisor, Prof. Steve Finkel. For a better advisor I could not have
asked. He knew when to guide me and when to let me take the bit between my teeth and
run with an idea. He was always available to answer my questions and give me new ideas
when I was stuck, but also trusted me when I came to him with bizarre results (which was
most of time).

Additionally, I would like to thank the members of my committee, past and present, for
helping me to become a better scientist. Prof. Ken Nealson was always enthusiastic and
offered insights I doubt I ever would have stumbled upon on my own; Prof. Steve
Goodman changed the way I thought about approaching a problem; Prof. Will Berelson,
for providing his unique insight; Prof. Katrina Edwards, for asking my favorite question
during my qualifying exam; and Bill Costerton, who first got me excited about biofilms.

My heartfelt thanks goes to our collaborators at the Naval Research Laboratory,
especially Dr. Brad Ringeisen, Dr. Lisa Fitzgerald and Dr. Justin Biffinger, who were not
only kind enough to include me in their work, but who are just great people.

Past and present members of the Finkel lab deserve a great deal of gratitude as well. I am
one of the fortunate few who works with her friends every day. I want to thank Dr. Slav
Palshevskiy for being so patient with me and showing me the ropes around lab. A huge
iv

“thanks” goes to Juliana Lima-Fricks, Kavita Chavan, Dr. Karin Kram and Lacey
Westphal for being so helpful and so much fun, and an extra-special thanks goes to Chris
Corzett, who’s been right there with me the entire way.

Finally, I wish to acknowledge the funding that made this work possible. This work was
supported in part by a MURI grant from the U.S. Air Force Office of Scientific Research.

vii

List of Tables

Table 4.1: Summary of notable amino acid pairs and their
effects on the four biofilms

95
Table 4.2: Summary of amino acid consumption, utility as
Stickland donors and acceptors, and predicted
effects.

100
Table A.1: Parameters for distinguishing between
laboratory assays and HTS assays
135

v

Table of Contents

Dedication

ii
Acknowledgements

iii
List of Tables vii

List of Figures

viii
Abstract

xi
Chapter 1: Introduction

1
Chapter 2: Long-term survival and evolution of planktonic
Shewanella oneidensis MR-1
11
2.1 Overview 11
2.2 Introduction 11
2.3 Materials and Methods 13
2.4 Results 16
2.5 Discussion 41
2.6 Conclusion

45
Chapter 3: Adaptive evolution of Shewanella oneidensis
MR-1 biofilms
47
3.1 Overview 47
3.2 Introduction 47
3.3 Materials and Methods 50
3.4 Results 56
3.5 Discussion 76
3.6 Conclusion

82
Chapter 4:Amino acid supplementation and biofilm
formation of Shewanella oneidensis MR-1
84
4.1 Overview 84
4.2 Introduction 84
4.3 Materials and Methods 86
4.4 Results 87
4.5 Discussion 95
4.6 Conclusion

102
Chapter 5: Conclusions and future directions

103
References 115
vi

Appendix A: Characterization of the electrochemically active
bacteria utilizing a high-throughput voltage-based
screening assay
132
A.1 Overview 132
A.2 Introduction 133
A.3 Materials and Methods 137
A.4 Results 141
A.5 Discussion 154

Appendix B: Simultaneous analysis of physiological and
electrical output changes in an operating microbial
fuel cell with Shewanella oneidensis
155
B.1 Overview 155
B.2 Introduction 156
B.3 Materials and Methods 159
B.4 Results 161
B.5 Discussion

172
Appendix C: The utility of Shewanella japonica for
microbial fuel cells
174
C.1 Overview 174
C.2 Introduction 174
C.3 Materials and Methods 178
C.4 Results 183
C.5 Discussion

196
Appendix D: Recipes and Strain List 197

vii

List of Tables

Table 4.1: Summary of notable amino acid pairs and their
effects on the four biofilms

95
Table 4.2: Summary of amino acid consumption, utility as
Stickland donors and acceptors, and predicted
effects.

100
Table A.1: Parameters for distinguishing between
laboratory assays and HTS assays
135

viii

List of Figures

Figure 1.1: Model microbial fuel cell.

2
Figure 1.2: Five phases of growth in planktonic E. coli
culture.

6
Figure 1.3: Stages of biofilm development.

7
Figure 1.4: Schematic of the Stickland reaction.

9
Figure 2.1: Long-term survival of S. oneidensis MR-1.

17
Figure 2.2: Expression of the GASP phenotype.

21
Figure 2.3: GASP competitions of older versus younger aged
populations.

25
Figure 2.4: Effect of spiking with lactate and fumarate.

27
Figure 2.5: GASP competitions of anaerobic Minimal
Medium spiked with fumarate.

29
Figure 2.6: Long-term survival and anaerobic Minimal
Medium cultures with doubled initial concentration
of fumarate.

31
Figure 2.7: Increasing concentrations of initial fumarate in
anaerobic Minimal Medium cultures.

33
Figure 2.8: Recovery of anaerobic Minimal Medium cultures
without the addition of electron acceptor.

34
Figure 2.9: “GASP swap.”

36
Figure 2.10: GASP competitions 30-day-old evolved
population and eight clonal subpopulations.

38
Figure 2.11: “Clone of a clone” GASP competition.

40
Figure 3.1: Schematic of biofilm aging.

52
Figure 3.2: Schematic of biofilm competition assays. 54
ix

Figure 3.3: Establishment of naïve and aged biofilm
populations on a virgin surface.

57
Figure 3.4: Competition of naïve versus aged biofilm
populations on a virgin surface.

59
Figure 3.5: Competition of naïve versus aged biofilm
populations for establishment within a 24-hour
biofilm.

62
Figure 3.6: Schematic of serially aged biofilms.

63
Figure 3.7: Establishment of unaged and serially aged
biofilm populations on a virgin surface.

64
Figure 3.8: Competition of naïve versus serially aged 3x10-
day-old biofilm populations on a fresh surface.

65
Figure 3.9: Competition of naïve versus serially aged 3x10-
day-old biofilm populations for establishment within
a 24-hour biofilm.

66
Figure 3.10: Competition of naïve versus serially aged
3x10+20-day-old biofilm populations for
establishment within a 24-hour biofilm.

67
Figure 3.11: Crystal violet staining of 10- or 20-day-old
biofilms with 10- or 20-day-old biofilm conditioned
media.

69
Figure 3.12: Heat maps of biofilm competitions in aged
biofilms.

71
Figure 3.13: Heat map of biofilm subclonal population
competitions.

74
Figure 4.1: Crystal violet staining of biofilms over a period
of hours and days.

89
Figure 4.2: Heat map of all replicates of all supplementary
amino acid pair combinations.

92
x

Figure A.1: Images of the pipet MFC.

142
Figure A.2: Performance of pipet MFCs.

143
Figure A.3: Images of the nine-well VBSA.

144
Figure A.4: Parellel current-interrupt data a 4-well VBSA
containing S. oneidensis MR-1.

146
Figure A.5: Current generated from the VBSA.

148
Figure A.6: Current generated from air exposed anodes
containing S. oneidensis MR-1 with selected electron
donors in a nine-well VBSA.
152

Figure B.1: Diagram of the nine-well VBSA.

132
Figure B.2: Average current output from S. oneidensis MR-
1-containing VSBA.

163
Figure B.3: ESEM images of the chemically fixed carbon
anode surfaces from VBSAs with lactate.

165
Figure B.4: Average current output from S. oneidensis MR-
1-containing VBSA.

169
Figure B.5: ESEM images of the chemically fixed carbon
anode surfaces from VBSAs with glucose.

171
Figure C.1: Images of S. japonica.

184
Figure C.2: Voltage-based screening assay (VBSA) current
output data.

186
Figure C.3: Current versus time chart for S. japonica in the
mini-MFC.

188
Figure C.4: HPLC data for the consumption of sucrose of S.
japonica in the mini-MFC.

190
Figure C.5: Cyclic voltammograms of uninoculated and
filtered S. japonica growth medium.

194

xi

Abstract
Microbial fuel cells are batteries in which microorganisms catalyze the conversion of
organic fuel (such as lactate) into protons and electrons that power a resistor (e. g., a light
bulb) before reducing the terminal electron acceptor (e. g., oxygen is reduced to water).
Great improvements in power production and efficiency have been made by engineering
inorganic components, such as the electrodes themselves, to be more efficiently utilized
by fuel cell-inhabiting organisms. However, other avenues for improvement may exist,
that is, engineering the fuel cell-inhabiting organisms themselves. We hypothesized that
Shewanella oneidensis MR-1, a model organism used for studying microbial fuel cells,
could be shown to evolve under physiological conditions which mimic those found in
microbial fuel cells. These physiological conditions include the planktonic lifestyle, the
biofilm lifestyle, and transient association between the two – that is, those cells that
rapidly detach from and reattach to the biofilm.

Here we show the Growth Advantage in Stationary Phase (GASP) phenotype conferred
by aging cells planktonically in conditions of abundant electron donor and acceptor, as
well as conditions of either electron donor or acceptor limitation. In general, the longer
cells are aged planktonically, the greater their advantage when competing in a similar
environment. A GASP-like phenotype is also conferred by aging cells in a biofilm for 10
days, though aging cells continuously within a biofilm for 20 days resulted in a
competitive disadvantage. To better understand cells that are transiently associated with
both lifestyles, we observed the rapid formation of, detachment from and reattachment to
xii

biofilms. Biofilm spontaneously form both where oxygen is plentiful and where it is
scarce. Oxygen-replete biofilms and oxygen-poor biofilms respond to different
supplementary amino acids. Response to amino acid supplementation also varies
according to the developmental stage of these biofilms. These data may offer insight into
the biology of microbial fuel cells, as well as guidance for physiological treatments and
methods of directed evolution that will improve microbial fuel cell performance.
1

Chapter 1: Introduction

The central question this work addresses is: can we engineer a living battery through
adaptive evolution? Microbial fuel cells capitalize on the ability of microorganisms to
convert organic material into electrons and protons. Like any other fuel cell, this
technology acts as a battery; its advantage is that the catalyst (the microbial community)
is self-renewing, so long as organic matter is provided. To improve the efficiency and
overall power output, we need to understand the biology that makes power production
possible. The adaptive nature of Shewanella oneidensis MR-1 is characterized separately
in a number of the different physiological environments that coexist within the microbial
fuel cell.

1.1 Microbial Fuel Cells
Microbial fuel cells (MFCs) use microorganisms to generate electrical current (Rabaey
and Rozendal, 2010). In a typical MFC, electrogenic bacteria populate the anode
compartment of the fuel cell, into which the fuel is injected. Fuels used in MFCs
typically fall within a range of organic material, from lactic acid to wastewater
(Watanabe, 2008; Biffinger, et al., 2009; Pant, et al., 2010). The nutrient fuel is then
catabolized, and the resulting electrons can be fed from the microorganism’s electron
transport chain to the anode. After passing through a resistor, electrons arrive at the
cathode where they reduce the terminal electron acceptor (frequently oxygen). MFCs can
be utilized for several applications, including biosensors (Mitcheson, 2010; Gong, et al.,
2

2011), biohydrogen production (Wang, et al., 2011) and the conversion of wastewater
into clean water and power (Lefebvre, et al., 2010; Fangzhou, et al., 2011). Several
physiological niches exist within the anode compartment of the MFC, as illustrated in
Figure 1.1. Specifically, microorganisms within the fuel cell may be (1) free-swimming
individuals (planktonic), (2) electrode-attached, sessile communities (biofilms), or
transitively associated with both.

Figure 1.1: Model microbial fuel cell. In the anode compartment, microorganisms
convert organic fuel into protons and electrons. Electrons are used to reduce the
electrode. These microorganisms can exist in several physiological states, including
the planktonic (1) and biofilm (2) lifestyles.

3

To understand how the biology of the MFC works as a whole, we sought to better
understand the different physiological states separately. Chapter 2 explores the long-term
survival and evolution of a particular MFC model organism in the planktonic lifestyle.
Chapter 3 investigates the nature of long-term survival and evolution of this organism in
the biofilm lifestyle. Chapter 4 hones in on shorter time scales, focusing on cells that
rapidly attach to and detach from the biofilm, and may therefore be considered
transitively associated with both the planktonic and biofilm lifestyles.

1.2 Shewanella oneidensis MR-1 as a model system
Shewanellae are Gram-negative rod-shaped bacteria that enjoy a ubiquitous distribution
throughout sedimentary, fresh water and marine environments (Fredrickson, et al., 2008).
In 1988, Shewanella oneidensis MR-1 (then named Alteromonas putrefaciens MR-1, and
later Shewanella putrefaciens MR-1) was the first organism shown to reduce manganese
oxide, making it a model organism for bacterial metal reduction (Myers and Nealson,
1988). Species of the Shewanella genus have since been shown to reduce a wide variety
of organic and inorganic, soluble and insoluble electron acceptors (Bencharit and Ward,
2005). Since its discovery, S. oneidensis MR-1 has been studied for its role in
environmental mineral cycling (Hau and Gralnick, 2007).

In addition to its role in biogeochemistry, S. onedensis serves as a model organism for a
variety of biotechnological applications. For example, it has been shown that one can
capitalize upon the exceptional ability of Shewanella species to reduce solid substrates. S.
4

oneidensis MR-1 was introduced into the anode chamber of a fuel cell to catalyze the
transfer of electrons from the fuel to the electrode (Kim, et al., 1999). Other organisms
have since been shown to reduce solid electron acceptors (Fredrickson and Zachara,
2008), but S. oneidensis MR-1 remains one of the best-characterized organisms for
studying this phenomenon (Logan, 2009). Due to its ability to respire solid substrates,
such as graphite electrodes, Shewanella oneidensis MR-1 is used as a model organism for
studying MFCs (Kim, et al., 2002; Bretschger, et al., 2007; El-Naggar, et al., 2008).

1.3 Evolution in MFCs
Two reasons it is important to understand the evolution of organisms relevant to the MFC
are: (1) to enhance power production through the application of specific selective
pressures and (2) to prevent adaptations that are deleterious to power production in
deployed MFCs. To date, enhanced power production via artificial selection has been
demonstrated using Escherichia coli (Zhang, et al., 2006), Geobacter sulferreducens (Yi,
et al., 2009) and mixed cultures (Ki, et al., 2008). Because of its importance as a model
organism in the MFC, we sought to better understand the evolution of S. oneidensis
MR-1.

We wanted to investigate evolution in a manner that best reflects the evolution occurring
within the MFC – that is, evolution as it relates to power production, Coulombic
efficiency (amount of organic material converted into electricity) and waste removal
(Borole, et al., 2011; Erable, et al., 2011). While evolution in other organisms has been
5

demonstrated via several approaches, batch culture conditions select for organisms that
can survive for days to years without the addition of nutrients (Dressaire, et al., 2011;
Marzan and Shimizu, 2011). This creates an environment that favors efficient utilization
of resources (including the detritus of expired organisms). Indeed, it has been shown that
E. coli mutants with enhanced ability to catabolize amino acids (compounds prevalent in
many types of wastewater MFCs are being designed to treat) are highly selective in batch
culture (Zinser and Kolter, 1999; Zinser and Kolter, 2000). Batch culture conditions
select for traits desirable in the MFC; therefore, we investigated the evolution of S.
oneidensis MR-1 in long-term batch culture.

In the laboratory setting, evolution of E. coli both in planktonic culture (Elena and
Lenski, 2003; Finkel and Kolter, 1999; Maharjan, et al., 2006) and in biofilms (Ponciano,
et al., 2009; Kraigsley and Finkel, 2009) has been demonstrated in several ways,
including via the expression of the GASP (Growth Advantage in Stationary Phase)
phenotype.

1.4 Growth Advantage in Stationary Phase
When aerated LB medium is inoculated with a small amount of planktonic E. coli, the
cells within that culture experience a five phase life cycle, as shown in Figure 1.2 (Finkel,
2006). In the first phase (lag), inoculated cells maintain a relatively low density.
Following lag phase, cell density increases exponentially; this phase is referred to as log
phase. The cells reach a maximum density and enter stationary phase, where cell
6

densities remain at roughly maximum for two to three days. After stationary phase, cells
enter death phase, during which time cell density drops roughly two to three orders of
magnitude. Remaining cells enter the long-term stationary phase. Without any addition of
nutrients, cells in long-term stationary phase maintain densities above the limit of
detection for years (Zambrano and Kolter, 1996).

Figure 1.2: Five phases of growth in planktonic E. coli culture. LB cultures
inoculated at low cell density experience lag (1), log (2) and stationary (3) phase.
Without the addition of nutrients, stationary phase is succeeded by death (4) and
long-term stationary (5) phase. Modified from Finkel, 2006.

It is during long-term stationary phase that the expression of the Growth Advantage in
Stationary Phase (GASP) phenotype is observed. The GASP phenotype is displayed
when aged cells out-compete unaged cells of an initially isogenic population during long-
term stationary phase (Zambrano, et al., 1993; Ramirez-Santos, et al., 2005; Pradhan, et
al., 2010; Sewell, et al., 2011; Helmus, et al., 2012). It is known that in E. coli,
spontaneous mutants are generated during outgrowth (Chavan and Finkel, in submission).
The selective pressures imposed in long-term stationary phase favor mutants in the
7

population that are better able to utilize the nutrients in spent media. Over time cells from
the aged population will increase in frequency and can drive the younger population to
extinction (Finkel, 2006).

1.5 Biofilm Lifestyle
Biofilms are surface-attached communities of microorganisms consisting of cells
surrounded by an exopolymeric substance (EPS) matrix (O’Toole, et al., 2000; Stoodley,
et al, 2002; Monroe, 2007; Kolter, 2010). These surface-attached communities are found
ubiquitously in nature, as well as synthetic surfaces such as indwelling medical devices
(Chen and Wen, 2011), ship hulls (Schultz, et al, 2011), and MFCs (Kim, et al., 2007).
As shown in Figure 1.3, the formation of the biofilm involves distinct stages of
attachment, proliferation, maturation and dispersal (Hall-Stoodley, et al., 2004).

Figure 1.3: Stages of biofilm development. To form a biofilm, a planktonic cell must
sense and attach to a surface (1), form an irreversible attachment (2), then
proliferate and recruit planktonic cells (3). It may then form an extracellular matrix
(4), and some cells may even detach and become planktonic, or reattach to an
existing biofilm (5). Modified from Monroe, 2007.

8

In order to maintain their intracellular electrochemical gradients, cells within these
communities must respond to a wide variety of stimuli, including nutrient and waste
concentrations, intercellular signaling molecules and terminal electron acceptor
availability (Karunakaran, et al., 2011). Due to their ability to respire a diverse host of
soluble and insoluble electron acceptors, Biofilms that include species such as those from
the Shewanella genus have an added level of complexity. Shewanella also have the
ability to donate electrons to remote electron acceptors via direct (e. g., nanowires) and
indirect (e. g., shuttles) means (Torres, et al., 2009). This has important ramifications for
electrode-associated biofilms, where the surface to which the community is attached
serves as the electron acceptor. With the electron donors (typically organic material in the
media) on one side of the biofilm and the electron acceptor on the other end, many cells
within the biofilm are physically distant from what they eat, or what they breathe, or both
(Rabaey, et al., 2007; Nealson and Finkel, 2011).

1.6 The Stickland Reaction
Long-term survival and the GASP phenotype have been shown to depend on the
utilization of detritus from cells that have expired during death phase (Finkel, 2006).
Indeed, the competitive advantage of certain GASP mutants has been attributed to
enhanced catabolism of amino acids (Zinser and Kolter, 1999; Zinser and Kolter, 2000).
Furthermore, amino acids are commonly found in wastewater used as fuel for MFCs
(Kahn and Wayman, 1964; Clauwaert, et al., 2008; Bhatnagar, et al., 2011). Therefore,
amino acids are of particular interest for the long-term survival and evolution of S.
9

oneidensis MR-1. In addition to their role as a carbon and energy source, amino acids
may play a role in respiration. In other organisms, it has been shown that some amino
acids can donate hydrogen, while others can accept hydrogen. Together, certain amino
acids pairs can generate ATP in a coupled reaction called the Stickland reaction, as
illustrated in Figure 1.4. Detrital amino acids may not only provide a carbon source for
long-term survival, but an alternative way to maintain intracellular electrochemical
gradients in S. oneidensis MR-1.

Figure 1.4: Schematic of the Stickland reaction. (http://en.wikipedia.org/wiki/
Stickland_fermentation, 2012)

1.7 Overview of Chapters
Chapter 2 investigates the long-term survival and evolution of S. oneidensis MR-1
planktonic cells. Specifically, the survivability and adaptability of cells grown in all four
combinations of aerobic and anaerobic, rich and Minimal Media was tested.
10

In Chapter 3, the competitive fitness of cells aged within a biofilm is evaluated. For
certain biofilm aging regimens, a series of assays were performed that demonstrate a
GASP-like phenotype of S. oneidensis MR-1 biofilms.

The mutability of biofilms from zero to 27 hours, and the effect of coupled amino acid
supplementation on these biofilms, is explored in Chapter 4. Biofilm density was
quantified by crystal violet staining.

A summary of the results is presented in Chapter 5, as well as direction for further
investigation. The genetic basis for the GASP phenotype in planktonic- and biofilm-aged
cells, the reason and genetic basis for a competitive disadvantage in certain biofilm-aged
strains, and the means by which certain amino acid pairs enhance biofilm staining, are
considered.
11

Chapter 2: Long-term survival and evolution of planktonic Shewanella
oneidensis MR-1

2.1 Overview
Shewanella oneidensis MR-1 is a model organism used to understand microbial fuel cell
(MFC) biology. Long-term survival in all four combinations of anaerobic and aerobic,
rich media and Minimal Media batch culture was characterized. The evolution of S.
oneidensis MR-1 was demonstrated under various laboratory conditions using the Growth
Advantage in Stationary Phase, or GASP, phenotype as a metric. Under conditions where
GASP was not demonstrated, electron donor and acceptor limitations were explored. This
approach can be used to adapt MR-1 to selective pressures relevant to MFCs.

2.2 Introduction
Microbial fuel cells are typically populated by both planktonic and biofilm-associated
microorganisms (Lanthier, et al., 2008). In either physiological state, microorganisms
must convert organic material into electrons and protons, and deliver the electrons to the
electrode. S. oneidensis MR-1 possesses a variety of strategies for reducing solid electron
acceptors (like the electrode), depending on its physiological state. When planktonic,
these bacteria can deliver the electrons to the electrode via direct contact (Harris, et al.,
2010), electron shuttles (Park and Zeikus, 2000; Rabaey, et al., 2005; Tang, et al., 2010;
Velasquez-Orta, et al., 2010), and possibly contact with the electrode-associated biofilm,
which can then deliver the electrons via its own means, including conductive nanowires
12

(Gorby, et al.; 2006; Reguera, et al., 2006). It is also possible that planktonic, MFC-
associated S. oneidensis MR-1 respires electron acceptors other than the electrode. For
example, planktonic cells may be scavenging oxygen that leaks into certain MFCs,
respiring detritus from dead cells (such as fumarate that was previously utilized in the
Krebs cycle), and perhaps donating protons and electrons to amino acids from dead cells
and the wastewater itself.

One of the attractive features of the microbial fuel cell is its potentially unlimited battery
life, as its catalyst (the microbial community) is self-reproducing. However, this feature
depends upon the ability of MFC microorganisms to survive long-term. It is known that
in aerobic, well-mixed E. coli batch culture, cells in long-term stationary phase cultures
can survive at detectable cell densities for years without the addition of nutrients
(Zambrano and Kolter, 1996; Kim, et al., 2001). For most microorganisms deployed in
MFCs, however, long-term survival in batch culture remains to be characterized. It is also
in long-term stationary phase that real-time evolution has been demonstrated in E. coli, e.
g., via expression of the growth advantage in stationary phase (GASP) phenotype (Farrell
and Finkel, 2003; Madan, et al., 2005). That is, when cells isolated from long-term
stationary phase E. coli batch culture are introduced as a minority into an overnight
culture of cells naïve to long-term stationary phase, aged cells take over the population
and drive the naïve cells to extinction (Zambrano, et al., 1993; Finkel and Kolter, 1999;
Finkel, 2006). If MFC organisms such as S. oneidensis MR-1 can survive long-term, their
13

ability to evolve in real-time (i. e., their ability to express the GASP phenotype) can also
be assessed.

In an effort to understand the planktonic lifestyle of S. oneidensis MR-1, I first
investigated long-term survival under several batch culture conditions. These conditions
were chosen to elucidate the role of electron donor and electron acceptor availability in
this organism’s survival. We observed distinct patterns of survival in each condition we
tested. Where survival was stymied, we determined whether electron donors and/or
electron acceptors were limiting. After establishing how population density changed over
time in these different environments, we probed changes in the population on the genetic
level by assaying for the GASP phenotype. Not only did we see that these populations
undergo genetic change over time; evolved populations display genetic and phenotypic
diversity.

2.3 Materials and Methods
2.3.1 Strains. All strains were derived from S. oneidensis MR-1 (provided by Ken
Nealson), or MR-1 marked by either a spontaneous rifampicin-resistant mutation that I
isolated or a kanamycin cassette knocked into the pilD gene (Saville, et al., 2010). A
spontaneous rifampicin-resistant mutant was isolated by taking 100µL of an overnight
culture of S. oneidensis MR-1 and spreading on LB agar plates (Difco) with 100µg/mL
rifampicin. The following day, a colony was picked and grown overnight in both 5mL LB
and 5mL LB with 100µL/mL rifampicin with aeration in a New Brunswick TC-7 roller
14

drum at 30
o
C. 100µL of the dense overnight cultures were frozen in 20% glycerol stocks.
The stability of rifampicin resistance in long-term batch culture was determined by
comparing cell counts on LB plates and LB plates with rifampicin over a period of
weeks. Drug resistant strains were chosen for their neutral fitness relative to each other
and to wild type in the conditions tested. That is, antibiotic-resistant strains were cultured
long term, and cell densities were comparable to those of wild-type. Furthermore, co-
cultures were initiated with equal numbers of kanamycin- and rifampicin-resistant cells,
and both strains maintained equal cell densities over the course of 30 days.

2.3.2 Culture media and chemicals. All cultures were inoculated from 20% glycerol
frozen stocks and grown overnight in 5mL Luria-Bertani (LB) broth with aeration. 5µL
of the appropriate overnight culture were used to initiate all long-term survival and
competition cultures. Long-term survival assays were performed in five different
conditions: aerobic LB, aerobic LB buffered with 100mM HEPES, anaerobic buffered
LB, aerobic Minimal Medium and anaerobic Minimal Medium. GASP competitions were
performed in the latter four conditions. Minimal Medium was prepared as described
(Bretschger, 2007). Except where indicated, anaerobic Minimal Media cultures were
supplemented with 30mM fumarate. Anaerobic LB and Minimal Medium cultures were
prepared by boiling media and sparging stoppered tubes with nitrogen before being
autoclaved.

15

2.3.4 Population density determination. To determine relative population densities of
antibiotic-resistant strains in competition, cultures were serially diluted and plated on
50µL/mL kanamycin or 100µL/mL rifampicin agar plates. Densities are expressed as
colony-forming units per milliliter (cfu/mL), and the limit of detection is 100 cfu/mL.

2.3.5 Long-term culture conditions and isolation of “evolved” populations.
Kanamycin- and rifampicin-resistant strains were inoculated in 5mL of the appropriate
medium and cultured at 30
o
C on a New Brunswick TC-7 roller drum to provide mixing in
the case of anaerobic cultures, and aeration and mixing in the case of aerobic cultures.
Aerobic cultures were incubated in 18x150mm test tubes. Anaerobic cultures were
incubated in 18x150mm Bellco test tubes with crimped rubber septa, allow for periodic
sampling via syringe. Except as noted, after inoculation of the outgrowth and competition
experiments, no additional nutrients were introduced. Each strain was inoculated in
nine replicates for each condition. The first set of triplicates was sampled and aliquots
frozen after ten days, the next set of triplicates was sampled and aliquots frozen after
twenty days, and the third set was sampled and aliquots frozen after thirty days. 100µL
aliquots were frozen in LB glycerol for storage at -80°C.

2.3.6 Clonal subpopulations. Clones of aged LB aerobic populations were isolated by
transferring a small amount of the frozen glycerol stock into 50uL LB. The medium was
spread onto an LB agar plate, incubated overnight at 30°C, and eight colonies were
picked. These colonies were grown overnight in LB and 100uL were frozen in 20%
16

glycerol stocks. Overnights were made from these stocks as described above, and these
overnights were used to initiate GASP competitions as described below.

2.3.7 GASP competitions. Competitions were conducted as previously described
(Finkel and Kolter, 1999). Briefly, 5µL of an LB overnight of aged strains, carrying
either a kanR or rifR marker, were introduced as a minority into a 5mL (1:1000 vol:vol
dilution) culture of unaged MR-1, carrying the alternate marker. Both antibiotic
resistance alleles are neutral in the absence of drug selection as described above.

2.4 Results
2.4.1 Long-term survival of MR-1 under aerobic and anaerobic conditions. We
sought to characterize the long-term survival of S. oneidensis MR-1. To do so, the
population densities were followed for several weeks under different media and
incubation conditions that varied the availability of electron donor and electron acceptor.
Figure 2.1 shows representative data for the long-term survival of S. oneidensis MR-1 in
unbuffered aerobic LB, buffered aerobic LB, anaerobic LB, aerobic Minimal Medium
and anaerobic Minimal Medium. Following inoculation at low density, all populations
show similar rates of logarithmic growth. In unbuffered aerobic LB, the day after
maximum cell density is reached, cultures enter death phase with cell densities falling
below the level of detection by day 2 and do not recover. In buffered aerobic LB,
anaerobic LB, aerobic Minimal Medium and anaerobic Minimal Medium, low density
17

inoculation is followed by logarithmic growth, stationary phase, death and long-term
stationary phase similar to E. coli (Gay, 1935).

Figure 2.1: Long-term survival of S. oneidensis MR-1. Cell densities in LB
(diamonds) and Minimal Medium (circles), aerobic (filled symbols), and anaerobic
(open symbols). ( ) aerobic unbuffered LB, ( ) aerobic buffered LB, ( ) anaerobic
buffered LB, ( ) aerobic Minimal Medium, and ( ) anaerobic Minimal Medium
batch culture. Asterisks (*) indicate the limit of detection. Data are representative.

While there was no addition of nutrients following inoculation, all buffered populations
persist for several weeks. Conditions of aerobic LB allow for the highest maximum
density; at roughly 5x10
9
cfu/mL, the overnight densities in aerobic LB are at least an
order of magnitude higher than that of any other condition. If the LB is buffered,
stationary phase lasts for an additional day before entrance into death phase, where the
population suffer a two to three order-of-magnitude decline. Cell densities then stabilize
at around 10
7
-10
8
cfu/mL in long-term stationary phase.
18

Of the conditions tested, anaerobic LB reaches the next highest maximum density,
roughly an order of magnitude lower the aerobic buffered LB. The timing of death phase
is similar to aerobic buffered LB, but the drop in cell density is less than two orders of
magnitude, and long-term stationary phase densities track aerobic buffered LB at an order
of magnitude lower to roughly equal numbers.

Maximum cell densities in aerobic Minimal Medium are only slightly lower than those of
anaerobic LB. Following stationary phase, these cultures show a more lingering decline
than LB cultures. However, aerobic Minimal Medium cell densities eventually recover,
and by the end of the long-term survival experiments the populations reach higher cell
densities than their LB counterparts.

Anaerobic Minimal Medium cultures reach a maximum density only slightly lower the
aerobic Minimal Medium, and the timing of stationary, death and long-term stationary
phases is similar to the other conditions. However, the decline in cell density during death
phase is at least four orders of magnitude – the most precipitous of all the conditions,
excepting only unbuffered aerobic LB. The majority of anaerobic Minimal Medium
cultures remain above the limits of detection, but persist at significantly lower cell
densities than any other culture condition.

19

2.4.2 Expression of the GASP phenotype. In Figure 2.2, populations derived from
cultures aged for 1, 10, 20 or 30 days are introduced as a minority, at a ratio of 1:1000
(vol:vol), into 1-day-old wild-type cultures expressing the reciprocal marker. While
these strains are perhaps not ideal for generating power, they provide the means for
assaying real-time evolution, a phenomenon relevant to the MFC. The first column
(Figure 2.2A, E, I and M) shows negative controls, in which 1-day-old populations were
added as a minority to the 1-day-old majority culture. After one day of incubation,
minority populations have not acquired sufficient advantage to compete in a 1-day
majority. The majority is therefore roughly one thousand times more likely to acquire the
beneficial mutation; this mutant then increases in frequency and drives less fit cells –
including the 1-day-old minority – to extinction. In aerobic LB, anaerobic LB and aerobic
Minimal Medium, 1-day-old minorities are unable to increase in density relative to the
majority (Figure 2.2A, E and I), and in some cases are driven below the limit of detection
(100cfu/mL). However, the anaerobic Minimal Medium negative control (Figure 2.2M)
does not yield results that would lend confidence to any GASP phenotypes we might see
in this condition; while both populations fluctuate, the majority population is unable to
gain a clear advantage over the minority.

20

Figure 2.2: Expression of the GASP phenotype. In A-D cells were aged and
competed in aerobic buffered LB; in E-H, cells were aged and competed in
anaerobic buffered LB; in I- L, cells were aged and competed in aerobic Minimal
Medium; and in M-P, cells were aged and competed in anaerobic Minimal Medium.
A, E, I and M are controls in which the 1-day old majority was competed against a
1-day old minority. In B, F and J, minority populations were aged for 10 days
before competing with the 1-day old majority. In C, G, K and N, minority
populations were aged for 20 days, and in D, J, L and O, minority populations were
aged for 30 days. Minority populations are represented by dashed lines and open
symbols; majority populations are represented by solid lines and filled symbols.
Corresponding symbol shapes indicate pairwise competitions. Data are
representative.
21

22

For each age and each marker, three strains were isolated and each was competed in
triplicate (n = 18 for each age and condition). While populations aged for one day are
unable to GASP, cells incubated for 10-days in aerobic LB, anaerobic LB and aerobic
Minimal Medium strains show a significant competitive advantage over their 1-day-old,
initially isogenic siblings (Fig. 2B, F and J). The longer the minority populations are aged
before competition, the more pronounced the advantage tends to be, as discussed below.

Aerobic LB minorities aged for 20 days overtake the unaged minority around the same
time as 10-day-olds (day 5) or even a day later. By the end of the experiment, however,
aerobic LB cultures consist of roughly equal numbers of unaged and 10-day-old cells
while population densities of 20-day-old cells can be approximately five orders of
magnitude higher than the unaged population. Cells aged for 30 days in aerobic LB
overtake the majority population a little earlier (around day 4) and by the end of the
experiment are approximately four orders of magnitude higher relative to the cell
densities of the unaged population.

Anaerobic LB cultures show a cross-over pattern (that is, the time at which the minority
population overtakes the majority) similar to aerobic LB (day 5 for 10-day-old minorities,
day 6 for 20-day-olds and day 4-5 for 30-day-olds). By the end of the experiment, cells
aged for ten days in anaerobic LB are at two to four orders of magnitude higher densities
relative to the unaged population, while 20-day-olds are consistently four orders-of-
23

magnitude higher, and 30-day-olds are three to four orders of magnitude higher densities,
but more quickly reach these higher relative densities.

Cells aged ten days in aerobic Minimal Medium do not overtake the unaged minority
until day 10 to 15, and by the end of the experiment, the relative population densities of
unaged and 10-day-old cells are frequently within an order of magnitude. 20-day-old
cells more consistently overtake the unaged majority by day 11, and by the end of the
experiment cell densities are a least ten-fold higher than the unaged population. When
aged for thirty days in aerobic Minimal Medium, cells overtake the majority even earlier
(around day 9) and by the end of the experiment are four to five orders of magnitude
higher than the unaged population.

Consistent with negative control experiments, 20- and 30-day-old anaerobic Minimal
Medium strains fluctuate with the 1-day-old majority population, but population densities
remained comparable. Overnights inoculated from frozen stocks of anaerobic Minimal
Medium cultures aged 10 days do not grow, thus 10-day-old minority populations are
inoculated and remain below the limit of detection in competition with the 1-day-old
majority. Data are representative, and consistent when markers are swapped (data not
shown).

Figure 2.2 shows clear differences in competitive phenotypes between cells aged for 1,
10, 20 or 30 days in a given medium (except for anaerobic Minimal Medium). In E. coli,
24

it is known that GASP mutants sweep through the population by day 10, but these
mutants accrue additional advantageous mutations when left to incubate in long-term
stationary for more extended periods. As a result, cells isolated from 20-day-old cultures
outcompete cells from 10-day-old cultures, and 30-day-old cells outcompete 20-day-old
cells. In general, the longer S. oneidensis MR-1 cells are aged in buffered aerobic LB,
anaerobic LB and aerobic Minimal Medium, the sooner these aged cells overtake naïve
populations, outnumbering the naïve cells by the end of the experiment. We therefore
hypothesized that older populations have accrued more advantageous mutations than
younger populations, and that older populations will display the GASP phenotype when
introduced as minorities in competitions where younger aged populations are the
majority. To test this, 20-day-old aerobic LB cells were added as a minority into
overnight cultures of 10-day-old aerobic LB populations, and 30-day-old aerobic LB cells
were added as a minority into overnight cultures of 20-day-old aerobic LB populations.
As shown in Figure 2.3, 20-day-old populations outcompete 10-day-old populations, and
30-day-old populations outcompete 20-day-old populations. These data support the
hypothesis that the longer S. oneidensis MR-1 cells are aged in long-term stationary
phase, the more advantageous mutations they accrue for GASP.
25

Figure 2.3: GASP competitions of older versus younger aged populations. Aerobic LB competitions of (A) 1-day-old
(solid black lines, filled symbols) versus 10-day-old (dotted lines, open symbols), (B) 10-day-old (dotted lines) versus
20-day-old (dashed lines, open symbols), or (C) 20-day-old (dashed lines, filled symbols) versus 30-day-old (solid
gray lines, open symbols). Open symbols denote minority populations, while filled symbols lines denote majority
populations. Data are representative.
26

2.4.3 Addition of electron donor, electron acceptor, or both to anaerobic Minimal
Medium. As shown in Figure 2.1, anaerobic Minimal Medium populations decline
several orders of magnitude more than anaerobic LB populations during death phase. We
hypothesized that either electron donor or electron acceptor concentration was limiting in
Minimal Medium. To test this, cultures at day 2 or day 3 (preceding death phase) were
spiked with lactate (an electron donor), fumarate (an electron acceptor), or both.

Figure 2.4A and B show that cultures treated with lactate on day 2 or 3, do not differ
significantly from the controls in either the timing of death phase, nor does additional
lactate ameliorate the 100,000-fold reduction in cell density. This result indicates that
cells do not die because they lack electron donor. When fumarate is added on day 2
(Figure 2.4C), stationary phase is prolonged for two days. When fumarate is added on
day 3 (Figure 2.4D), culture densities decrease, but recover in the middle of death phase
to stationary phase densities, and remain at those densities for an additional five days.
Ultimately, final population densities fall to levels comparable to the control. When both
lactate and fumarate are added on day 2 (Figure 2.4E), the benefits of increasing electron
acceptor concentration are abolished; electron acceptor is again limiting. The addition of
both lactate and fumarate on day 3 (Figure 2.4F) slightly prolongs stationary phase, with
an effect intermediate between the control and the addition of fumarate alone.
27

Figure 2.4: Effect of spiking anaerobic Minimal Medium cultures with lactate and
fumarate. Lactate was added to a final concentration of 15mM on day 2 (A) and day
3 (B). Fumarate was added to a final concentration of 30mM on day 2 (C) and day 3
(D). Lactate and fumarate were added to final concentrations of 15mM and 30mM,
respectively, on day 2 (E) and day 3 (F). Dotted lines and open symbols indicate
supplemented cultures, while solid lines and filled symbols are the unsupplemented
control cultures.

2.4.4 Anaerobic Minimal Medium GASP with fumarate spiking. It is possible GASP
mutants are present in the population at the time of inoculation, but are lost in a kind of
population bottleneck that occurs during the particularly severe death phase that occurs in
28

these cultures, after which only one in 100,000 cells may survive. As shown in Figure
2.4, adding fumarate to anaerobic Minimal Medium prolongs stationary phase, although
cell densities in long-term stationary phase are roughly equal to those of the control.
Prolonging stationary phase may ameliorate the population bottleneck of death phase, in
which case, GASP mutants could be isolated from anaerobic Minimal Medium cultures.
To test this, cells were isolated from 10-, 20- and 30-day-old Minimal Medium cultures
that were spiked with fumarate on day 2. These isolates were frozen and regrown
overnight in LB before being introduced as minorities into 1-day-old anaerobic Minimal
Medium cultures. However, in the subsequent GASP competition experiments, cultures
were not spiked with additional fumarate. These competitions are shown in Figure 2.5.

29

Figure 2.5: GASP competitions of anaerobic Minimal Medium spiked with
fumarate. Populations were aged in anaerobic Minimal Medium spiked with
fumarate on day 2, isolated from a 30-day-old cultures (open symbols, dashed lines),
and added as minorities into 1-day-old (filled symbols, solid lines) anaerobic
Minimal Medium cultures. Asterisks (*) indicate the limit of detection. Markers
were swapped, and data are representative.

Cells aged in anaerobic Minimal Medium supplemented with fumarate on day 2 and aged
for 30 days do not behave significantly differently from cells aged in anaerobic Minimal
Medium cultures that were not spiked with fumarate. Furthermore, unlike cells aged and
competed in aerobic LB, anaerobic LB and aerobic Minimal Medium, cells aged in
spiked anaerobic Minimal Medium never drive down the cell densities of the unaged
strain.

30

2.4.5 Long-term survival of anaerobic Minimal Medium initiated with increased
initial fumarate concentrations. As shown in Figure 2.4, anaerobic Minimal Medium
cultures are electron acceptor limited. While the addition of fumarate before death phase
prolongs stationary phase, it does not alter the number of cells lost during death phase, or
affect long-term stationary phase cell densities. We hypothesized that if initial
concentrations of electron acceptor (fumarate) were increased, S. oneidensis MR-1 would
be able to accumulate more biomass (greater maximum density), allowing for more
nutrients to be cannibalized after death (increased cell density during stationary phase).
To test this hypothesis, the concentration of fumarate was doubled (from 30mM to
60mM) at inoculation, and cell densities were monitored (Figure 2.6).

31

Figure 2.6: Long-term survival and anaerobic Minimal Medium cultures with twice
the initial concentration of fumarate. S. oneidensis MR-1 kanR (solid lines) and rifR
(dotted lines) were inoculated into nine and eight, respectively, 5mL cultures of
anaerobic Minimal Medium cultures with twice the normal (30mM) amount of
fumarate, for an initial concentration of 60mM fumarate. Asterisks (*) indicate the
limit of detection.

Increasing the initial concentration of fumarate prolonged stationary phase; indeed,
stationary phase lasted longer when an extra dose of fumarate was added at the beginning
of the experiment than when it was added on day two or three. However, under
conditions of increases fumarate, death phase still resulted in losses of cell density of
roughly four orders of magnitude, and long-term stationary phase cell densities were
similar to all other anaerobic Minimal Medium cultures. Because neither death phase nor
long-term stationary phase were significantly affected, adding more fumarate at the
32

beginning of the experiment would likely not address the question of whether GASP
mutants arise in anaerobic Minimal Medium cultures before or after death phase.
Anaerobic Minimal Medium cultures with increased initial concentrations of fumarate
behave similarly to anaerobic Minimal Medium cultures supplemented with fumarate;
thus, like supplemented cultures, we do not expect increased initial fumarate cultures to
yield GASP populations.

Further increases in initial fumarate concentrations proved to be poisoning, as shown in
Figure 2.7. Thus, increasing initial concentrations of fumarate is not an effective
approach for reducing the severity of death phase in anaerobic Minimal Medium cultures,
or increasing cell densities during long-term stationary phase.

33

Figure 2.7: Increasing concentrations of initial fumarate in anaerobic Minimal
Medium cultures. Cultures were initiated with 0mM (blue), 50mM (cyan), 100mM
(green), 150mM (yellow), 200mM (orange), 250mM (red) or 300mM (purple)
fumarate, all in triplicate. Asterisks (*) indicate the limit of detection.

2.4.6 Recovery of cells cultured in anaerobic Minimal Medium from death phase
without the addition of electron acceptor. As shown in Figures 2.2M-P, 2.5 and 2.6,
death phase in anaerobic Minimal Medium cultures is often so precipitous that cell
densities quickly fall below the limit of detection of 100 cfu/mL. However, after 7 days
of continuous incubation without the addition of nutrients, these cultures recover at least
an order of magnitude, and frequently much more. Examples of this phenomenon are
highlighted in Figure 2.8.
34

Figure 2.8: Recovery of anaerobic Minimal Medium cultures without the addition of
electron acceptor. Without the addition of nutrients (including electron acceptor)
anaerobic Minimal Medium cultures cell density recovers from death phase, even if
cell density falls below the limit of detection. Asterisks (*) indicate the limit of
detection. Symbols are meant to distinguish replicates. Data are representative.

2.4.6 Competitive fitness of aged cells competing in other environments. It is possible
that the competitive advantage displayed by cells recovered from aged cultures is the
result of mutations within a single locus, e. g., mutations analogous to those that attenuate
the expression of rpoS in E. coli. If this is the case, cells aged in aerobic Minimal
Medium cultures, for example, should display the GASP phenotype when introduced as a
minority into a 1-day-old aerobic LB culture, and vice versa. To test this, different 30-
day-old aerobic Minimal Medium-aged cells were added as minorities into 1-day-old
aerobic LB cultures (Figure 2.9A and B), and different 30-day-old aerobic LB-aged cells
were added as minorities into 1-day-old aerobic Minimal Medium cultures (Figure 2.9B
35

and C). In both cases, some strains display the GASP phenotype in conditions dissimilar
from the conditions in which they were aged. Still other strains were maladapted to
conditions other than the conditions in which they were aged. These results indicate that
different mutations are likely responsible for the GASP phenotypes displayed by each of
the strains aged for 30 days within a given condition, and different mutations are likely
present in strains aged in different conditions.

36

Figure 2.9: “GASP swap.” Two different 30-day-old aerobic Minimal Medium
strains (A and B, solid lines and symbols) are introduced as minorities into three
separate 1-day-old aerobic LB cultures (dashed lines and open symbols) each. Two
different 30-day-old aerobic LB strains (C and D, dashed lines and open symbols)
are introduced as minorities into three separate 1-day-old aerobic Minimal Medium
cultures (solid lines and symbols) each. Markers were swapped and data are
representative.

2.4.7 Phenotypic diversity of clones isolated from a GASP population. We wanted to
determine whether a GASP population was clonal, in which case we would predict that
clones isolated from the population would display a GASP phenotype similar to the
original population. To test this, eight clones from a GASP population were isolated.
Each of the eight clones was introduced as the minority in a GASP competition (all eight
clones in triplicate). The GASP phenotypes of the clones were compared to the
phenotype of the population from which they were derived. Figure 2.10A shows the
GASP competitions of the original population (kanamycin-resistant, aged 30 days in
37

aerobic LB), and Figure 2.10B-I show GASP competitions of the eight clonal
subpopulations. While each clone competition is consistent between replicates, the
variation between clones indicates that the original GASP population is diverse. For
example, this original 30-day-old population overtakes the majority, and both the aged
and unaged populations persist at comparable densities. Four of the clones behave
similarly to the original population (Figure 2.10B, D, E, and I), but two clones drove the
unaged population below detectable levels in one replicate (Figure 2.10C and I). Three of
the clones show a stronger GASP phenotype than the original population, increasing in
frequency up to a million-fold relative to the unaged population (Figure 2.10F, G and H).

Figure 2.10: GASP competitions 30-day-old evolved population and eight clonal
subpopulations. GASP competition of the original population aged 30 days in
aerobic LB (A) and eight populations cloned from the original (B-I). Aged strains
are introduced as minorities (open symbols, dashed lines), into 1-day-old majority
cultures (filled symbols, solid lines).
38

39

Clones from 10- and 20-day-old aerobic LB populations also show phenotypic diversity
(data not shown). However, none of the clones show the strong GASP phenotype seen in
k30a clone #7 (Figure 2.10H), for example. If such strong GASP mutants are present in
10-day-old populations, they are likely present at relatively low frequencies. Because
clones were chosen at random, they are believed to be representative of the total
population; i. e., if three of the eight clones display a strong GASP phenotype (as in
Figure 2.10F, G and H), it is likely that a strong GASP mutant is present at a frequency of
37.5% of the total population. The fact that none of the eight 10-day-old clones show a
strong GASP phenotype indicates that these mutants, if present at all, likely compose less
than 12.5% of the total population at 10 days.

To further examine these subclonal populations, clones were isolated from competitions
of a 30-day-old subclonal population. During the competition, once the subclonal
minority overtook the 1-day-old majority and was present at greater cell densities for two
time points in a row, colonies from the titer plates were picked, grown overnight and
frozen in 20% glycerol stocks. These frozen stocks were then used to inoculate
overnights, of which 5µL aliquots were introduced into 1-day-old cultures.

40

Figure 2.11: “Clone of a clone” GASP competition. Strain isolated from a
competition of a 30-day-old aerobic LB clone (open symbols, dashed lines) is
competed against a 1-day-old majority (filled symbols, solid lines) in triplicate.
Asterisks (*) indicate the limit of detection.

As shown in Figure 2.11, this “clone of a clone” overtakes the 1-day-old population soon
after day 3 (the earliest crossover of any aged population), and drives the majority
population to extinction. Of all the aged strains isolated the “clone of a clone” displays
the strongest GASP phenotype. These results support the notion that planktonic GASP
populations accrue more advantageous mutations the longer they are aged. Additionally,
these data indicate that under the conditions tested, repeated selections allow for the
isolation of increased fitness.

41

2.5 Discussion
In unbuffered LB, death phase is so precipitous that the cell densities fall below the level
of detection and do not recover. This is likely due to the fact that carbohydrate in LB is
quickly exhausted, whereas amino acids are in excess. Once the carbohydrate is
consumed, much of the amino acid is utilized as a carbon and energy source, producing
ammonium that basifies the medium (Sezonov, et al., 2007). This increase in pH is likely
fatal to S. oneidensis MR-1. However, under conditions where cell densities remain
relatively high following death phase – both with and without oxygen – we observe the
evolution of S. oneidensis MR-1 in real time, as reflected by the appearance of mutant
cells expressing the GASP phenotype. Buffered aerobic LB-, anaerobic LB- and aerobic
Minimal Medium-evolved strains displayed the GASP phenotype.

The longer S. oneidensis MR-1 is aged in a given planktonic environment, the greater its
competitive advantage, indicating that additional beneficial mutations are accrued after
the first GASP mutant increases in frequency within the population. This hypothesis is
supported by the fact that 20-day-old cells outcompete 10-day-old cells, and 30-day-old
cells outcompete 20-day-old cells in aerobic LB, though these results have yet to be
confirmed in anaerobic LB and aerobic Minimal Medium. Some GASP populations have
a competitive advantage in environments other than the environment in which they
evolved, while other, similar strains do not have a competitive advantage in other
environments. These results have yet to be confirmed with anaerobically aged strains and
42

environments. However, the data collected indicate that, like E. coli (Farrell, Fant,
Jessner and Finkel, unpublished data), more than one mutation in S. oneidensis MR-1 can
give rise to the GASP phenotype.

Within each GASP population, clonal subpopulations display diverse GASP phenotypes.
Furthermore, strains displaying stronger GASP phenotypes are isolated from repeated
selection. Taken together, these data indicate that GASP populations are genotypically
diverse, and that S. oneidensis MR-1 is amenable to directed evolution.

Cells incubated in anaerobic Minimal Medium do not demonstrate GASP. It is known
that spontaneous mutations arise differently when bacteria are grown in different media
(Koskiniemi, et al., 2010); therefore, it is formally possible that GASP phenotype-
conferring mutations are not being generated in anaerobic Minimal Medium. However,
the fact that such mutations arise in anaerobic LB and aerobic Minimal Medium renders
this possibility unlikely. At this point, it is unknown whether advantageous mutants are
present in overnight populations of S. oneidensis MR-1, or if they arise during long-term
stationary phase. It is possible that advantageous mutations are generated during
outgrowth, but because anaerobic Minimal Medium death phase culls the population by
roughly five orders of magnitude, aging populations experience a bottleneck in which
GASP mutants are lost before they have a chance to increase in frequency within the
population. A second possibility is that competitively advantaged mutants do not appear
in the population until long-term stationary phase, but the low population density and the
43

frequency at which advantageous mutants arise makes the appearance of a GASP mutant
within thirty days unlikely.

The precipitous death phase in anaerobic Minimal Medium was delayed by the addition
of electron acceptor (fumarate), but not the addition of electron donor (lactate) or the
addition of both the electron donor and acceptor together. Future experiments are needed
to determine if additional fumarate prolongs stationary phase in anaerobic LB. Although
additional electron acceptor prolongs stationary phase in these cultures, the final cell
densities were the same as the control. Thus, cells aged in anaerobic Minimal Medium
spiked with fumarate eventually experienced the same population bottleneck as cells aged
in anaerobic Minimal Medium cultures that do not receive additional fumarate. If GASP
mutants are present in the population before death phase, the prolonged stationary phase
of fumarate-spiked does not allow for them to increase in frequency before death phase
decimates the population. If GASP mutants are generated after death phase, the addition
of fumarate does not increase long-term stationary phase cell densities, so the probability
of an advantageous mutant arising spontaneously within 30 days is still low. Some strains
aged in aerobic LB or aerobic Minimal Medium display GASP in conditions in which
they were aged, but do not display GASP in other conditions; it is therefore formally
possible that cells isolated from fumarate-spiked anaerobic Minimal Medium do not
display GASP in anaerobic Minimal Medium, but do display GASP in fumarate-spiked
anaerobic Minimal Medium. However, it is known that in E. coli, no cell division (or cell
death) occurs during stationary phase (Roostalu, et al., 2008); thus, it is unlikely that
44

prolonging stationary phase in anaerobic Minimal Medium cultures would have increased
the frequency of GASP mutants in the population, were they introduced with the 30-day-
old minority population. Further investigation is needed to determine whether GASP
mutants are present in anaerobic Minimal Medium cultures before death phase, or if they
are generated after death phase. If GASP mutants are lost as a result a population
bottleneck, increasing the overall population density (by increasing the volume of the
cultures) should allow for enough of those mutants to survive 99.99% losses. For
example, if GASP mutants are present at a frequency of 100 cells/mL (1 in every million
cells) than the culture volume would need to be at least 100 mL to expect one mutant to
survive death phase. Increasing the population density after death phase (possibly by
frequent spiking with fumarate post-death phase), should allow for a greater chance of
generating GASP mutants arising the population if indeed such mutants are not generated
until after cells enter long-term stationary phase.

Frequently death phase in anaerobic Minimal Medium results in cell densities falling
below the limit of detection. Although no nutrients (including electron acceptor) are
added, the cultures recover and reach at densities well above the limit of detection.
Given that the above data shows electron acceptor is limiting in these cultures, it is likely
that the death of the majority of cells in these cultures releases electron acceptor which
the surviving cells respire. Future investigation is needed to determine if surviving cells
can utilize dead cell detritus (such as the fumarate found in the citric acid cycle, or
45

possibly amino acids used in the Stickland reaction (Stickland, 1934) as an electron
acceptor, and if this is indeed the cause of recovery in cell density.

While the mutations that confer the GASP phenotype in S. oneidensis MR-1 have not
been mapped, it is clear that evolved populations are genotypically and phenotypically
diverse. From the GASP competitions of the clones of the evolved population in Figure
2.10 and 2.11, it is clear that at least three genotypes exist within the evolved population,
and probably more. Further investigation, possibly via whole genome resequencing, is
required to elucidate the specific mutations responsible for the competitive advantage of
these clones under the conditions in which they are competed. To our knowledge, this is
the first evidence of real-time, heritable adaptation in S. oneidensis MR-1. Now that
evolution has been demonstrated, it is possible to capitalize on this ability and direct the
evolution of S. oneidensis MR-1 in a way that will enhance the performance of the MFCs
this organism inhabits.

2.5 Conclusion
After reaching maximum cell densities around 10
9
(aerobic LB) or 10
8
cfu/mL (anaerobic
LB, aerobic Minimal Medium and anaerobic Minimal Medium), planktonic batch
cultures of S. oneidensis MR-1 enter death phase. In aerobic LB, anaerobic LB and
aerobic Minimal Medium, death phase results in 1-2 orders-of-magnitude reductions in
cell density, and surviving cells maintained cell densities around 10
6
-10
7
cfu/mL
throughout long-term stationary phase. In anaerobic Minimal Medium, death phase
46

results in at least 4 orders-of-magnitude reduction in cell density, and surviving cells
maintain densities around 10
4
cfu/mL throughout long-term stationary phase (Figure 2.1).

S. oneidensis MR-1 displays the GASP phenotype in all of the above conditions, except
anaerobic Minimal Medium (Figure 2.2). Stationary phase in anaerobic Minimal Medium
is prolonged by the addition of electron acceptor (Figure 2.4 and 2.6), but prolonging
stationary phase did not allow for the isolation of GASP mutants (Figure 2.5).

In general, the longer a population is aged planktonically, the more competitive it is in
long-term stationary phase (Figure 2.2). Long-term stationary phase populations appear
to be continuously evolving, as evidenced by the fact that older-aged populations out-
compete younger-aged populations (Figure 2.3). The GASP phenotype is conferred by
different types of mutations, some of which confer a competitive advantage in
environments other than the one in which cells were aged, others of which do confer a
competitive advantage in different environments (Figure 2.9). GASP populations are
diverse (Figure 2.10) and tractable to directed evolution via serial aging (Figure 2.11).

47

Chapter 3: Adaptive evolution of Shewanella oneidensis MR-1 biofilms

3.1 Overview
Shewanella oneidensis MR-1 is a model organism used for understanding electrically
conductive biofilms, including electrode-associated biofilms found in microbial fuel cells
(MFCs). Here we observe changes in competitive fitness in S. oneidensis MR-1 biofilms
in the laboratory setting within a few days. Cells harvested from 10-day-old biofilms
express a competitive advantage over cells naïve to the biofilm, whereas cells harvested
from 20-day-old biofilms are competitively disadvantaged relative to their naïve
counterparts, and 30-day-old biofilms are roughly equal to naïve cells. This pattern is
consistent in all strain backgrounds tested. The advantages and disadvantages of these
aging regimens were only apparent in competition, similar to the growth advantage in
stationary phase (GASP) .phenotype exhibited by planktonically grown cells. The ability
of S. oneidensis MR-1 biofilms to evolve has broad implications for natural and synthetic
environments, including environments found within applied technologies such as
microbial fuel cells.

3.2 Introduction
Biofilms are surface-attached communities of microorganisms consisting of cells
surrounded by an extracellular matrix (Monroe, 2007; Bazaka, et al., 2011; Mosier and
Cady, 2011; Trevors, 2011). In order to maintain their intracellular electrochemical
gradients, cells within these communities must respond to a wide variety of stimuli,
48

including nutrient and waste concentrations, intercellular signaling molecules and
terminal electron acceptor availability (Karunakaran, et al., 2011). Biofilms that include
species from the Shewanella genus have an added level of complexity due to their ability
to respire a diverse assortment of soluble and insoluble electron acceptors. Shewanella
species also possess the ability to donate electrons to remote electron acceptors via direct
(e. g. nanowires) (El-Naggar, et al., 2008; Torres, et al., 2009; Yi, et al., 2009) and
indirect (e. g. shuttles) means (Torres, et al., 2010; Yong, et al., 2011). This has important
ramifications for electrode-associated biofilms, where the surface to which the
community is attached serves as the electron acceptor. With the electron donors (typically
organic material in the media) on one side of the biofilm and the electron acceptor on the
other end, many cells within the biofilm are physically distant from what they eat, or
what they breathe, or both (Rabaey, et al., 2007; Nealson and Finkel, 2011).

Electrode-associated biofilms are often found in microbial fuel cells. Microbial fuel cells
(MFCs) utilize microorganisms (often including electrogenic bacteria such as S.
oneidensis MR-1) to generate electrical current (Erable, et al., 2010). These
microorganisms populate the anode compartment of the fuel cell, into which fuel
(typically organic material, such as lactic acid or wastewater) is injected. The nutrient
fuel is catabolized, and the resulting electrons can be transferred from the microbes’
electron transport chain to the anode. Electrons flow from the anode, through a resistor,
to a cathode where they reduce the terminal electron acceptor. S. oneidensis MR-1 is
often used as a model organism for studying MFCs because of its ability to respire solid
49

substrates, including graphite electrodes frequently utilized in MFCs (Myers and
Nealson, 1990; Turick, et al., 2009; Osman, et al., 2010; Salas, et al., 2010; Plymale, et
al., 2012).

Evolution in biofilms has been demonstrated in a number of organisms, including E. coli
(Kraigsley and Finkel, 2009, Ponciano, et al., 2009, Bernstein, et al., 2011), P.
aeruginosa (Lujan, et al., 2011), Bulkholderia cenocepacia (Poltak and Cooper, 2011)
and multispecies biofilms (Hansen, 2007). Evolution has also been shown in MFC
electrode-associate biofilms (Zhang, et al., 2006; Ki, et al., 2008; Yi, et al., 2009; Hou, et
al., 2011). Electricity production in MFCs is intimately, if complexly, related to
electrode-associated biofilms (Biffinger, et al., 2009; McLean, et al., 2010; Rosenbaum,
et al.; 2010). To better understand evolution of electrically conductive biofilms on a
conductive surface, we first sought to elucidate how S. oneidensis MR-1 biofilms change
over time on an electrically inert surface (glass). In other words, to understand biofilms in
which the sources of electron donor (within the medium) and acceptor (within the
surface) are physically distant from each other, we sought to understand biofilms in
which the sources of electron donor and acceptor are both within the medium.

In the laboratory setting, the accumulation of heritable adaptations in both planktonic
culture and in biofilms has been demonstrated through several methods, including via the
expression of the GASP (growth advantage in stationary phase) phenotype. The GASP
phenotype is displayed when aged cells outcompete unaged cells of an initially isogenic
50

population during long-term stationary phase. Long-term stationary phase imposes
selective pressures that favor mutants within the population that are better able to utilize
the nutrients in spent media. Over time these mutants from aged populations increase in
frequency and outcompete younger populations (Zambrano, et al., 1993; Finkel, 2006). In
Chapter 2, GASP mutants were isolated from and competed in well-mixed planktonic
cultures. Here the evolution of S. oneidensis MR-1 is examined by ascertaining whether
mutants isolated from stationary phase biofilms display a GASP-like phenotype when
competing in static culture.

3.3 Materials and Methods
3.3.1 Strains. All strains were derived from S. oneidensis MR-1 (provided by Ken
Nealson), or MR-1 marked by either a spontaneous rifampicin-resistant mutation or a
kanamycin cassette knocked into the pilD gene (Saville, et al., 2010). Drug resistant
strains were chosen for their neutral fitness relative to each other and to wildtype in the
conditions tested. A spontaneous rifampicin-resistant mutant was isolated by taking
100µL of an overnight culture of S. oneidensis MR-1 and spreading on LB agar plates
(Difco) with 100µg/mL rifampicin. The following day, a colony was picked and grown
overnight in both 5mL LB and 5mL LB with 100µL/mL rifampicin with aeration in a
New Brunswick TC-7 roller drum. 100µL of the dense overnight cultures were frozen in
20% glycerol stocks. The stability of rifampicin resistance was determined by comparing
cell counts on LB plates and LB plates with rifampicin over a period of several weeks.

51

3.3.2 Culture media and chemicals. All cultures were inoculated from 20% glycerol
frozen stocks and grown at 30
o
C overnight in 5mL Luria-Bertani (LB) broth with
aeration in a New Brunswick TC-7 roller drum. To inoculate the biofilms, the appropriate
mixture of overnights was diluted (1:100 v:v) of Minimal Media (Bretschger, 2007).

3.3.3 Biofilm incubation and harvesting. Acid-washed borosilicate glass test tubes were
inoculated with 1mL of the diluted cell mixture, and incubated statically at 30
o
C without
the addition of nutrients. After the appropriate amount of time (10, 20, or 30 days), the
supernatant was removed, and biofilms were washed three times with Minimal Medium
osmolytes. Cells were then removed from the biofilm via mild sonication, and 100µL
were frozen in 20% glycerol stocks. Freezing and several generations of planktonic
growth “resets” the physiology of biofilm-harvested cells; any adaptations to the biofilm
made via the accumulation of certain proteins, for example, are likely abolished. Thus,
any differences in competitive fitness are due to heritable changes. This procedure is
illustrated schematically in Figure 3.1 Up to three populations for each age and each of
the two markers were isolated and assayed as described below.
52

Figure 3.1: Schematic of biofilm aging.

3.3.4 Biofilm assays. Single-population biofilm formation was assayed as described
above, i. e., 10µL of an LB overnight inoculated from the appropriate frozen glycerol
stock was diluted in 1mL of Minimal Medium and incubated statically at 30
o
C in glass
test tubes for 48 hours. The supernatant was removed, biofilms washed, and cells were
released from the biofilm via sonication. Cell counts were determined by serial dilution
on LB plates. Instead of inoculating with 10µL of a single population, competitions were
inoculated using 5µL of each population (“naïve” and “aged”), and cell counts were
determined by serial dilution on LB plates with the appropriate antibiotic (50µL/mL
kanamycin or 100µL/mL rifampicin agar plates) (Kraigsley and Finkel, 2009). In
competitions for establishment on virgin surfaces, this mixture was added to 1mL of fresh
53

media, and incubated statically for 48 hours before harvesting. In competitions conducted
in the presence of existing biofilms, test tubes were inoculated with 10µL of MR-1
without an antibiotic resistance marker diluted in 1mL of Minimal Medium. After 24
hours or static incubation at 30
o
C, during which MR-1 was allowed to form a biofilm,
5µL of each of the two competing populations were inoculated into the media, and
biofilm cells were harvested after an additional 48 hours of incubation. Figure 3.2 is a
schematic illustrating procedures from a competition for a virgin surface (Figure 3.2A)
and competition for a biofilm surface (Figure 3.2B).

Figure 3.2: Schematic of biofilm competition assays. Procedures for competitions
conducted in the presence of a virgin surface (A) and an existing biofilm (B) are
illustrated above.

54

55

3.3.5 Biofilm Index. As presented in Figures 3.4-5 and 3.8-10, the Biofilm Index corrects
for inequities in the number of planktonic cells present at inoculation, and is calculated as
follows: Biofilm Index=(B
aged
/B
naïve
)/(P
aged
/P
naïve
), where B
aged
and B
naïve
are cell yields
derived from cells aged in or naïve to a biofilm, respectively, within the experimental
biofilm at the time of harvesting, and P
aged
and P
naïve
are the densities of planktonic cells
at inoculation (Kraigsley and Finkel, 2009). A value of >1 indicates that a greater number
of “aged” biofilm-harvested cells are present in the nascent biofilm, relative to their
inoculation ratio, whereas a value of <1 indicates that a greater number of cells “naïve” to
the biofilm are present. All assays were performed with markers swapped, each in
triplicate, and all experiments were repeated at least once.

3.3.6 Crystal violet staining. Biofilms were stained with 0.1% crystal violet (CV) for 30
minutes, then rinsed with diH
2
O to remove excess stain (O’Toole and Kolter, 1998a).
After stained biofilms were photographed, 0.5mL of 95% ethanol was added for 15
minutes to solubilize the CV staining the lower the biofilm band in each tube. After the
solubilized stain was removed and quantified, an additional 1.25mL of 95% ethanol was
added to solubilize the CV staining the upper biofilm band. Absorbance was measured at
590nm.

3.3.7 Clonal subpopulations. Clones of 20-day-old biofilm populations were isolated
by transferring a small amount of the frozen glycerol stock into 50uL LB. The medium
was spread onto an LB agar plate, incubated overnight at 30°C, and eight colonies were
56

picked. These colonies were grown overnight in LB and 100uL were frozen in glycerol
stocks. Overnights were made from these stocks as described in section 3.3.2, and these
overnights were used to initiate biofilm competitions as described in section 3.3.3 and
3.3.4.

3.4 Results
We hypothesized that static incubation of S. oneidensis MR-1within a biofilm selects for
mutants via adaptive evolution, as evidenced by the expression of a phenotype akin to the
GASP phenotype. To test this hypothesis, a series of experiments were performed to
investigate the expression of three potential phenotypes, none of which are mutually
exclusive: (1) biofilm-aged cells form biofilms with higher cell densities when introduced
to a virgin surface, as compared to biofilms formed by biofilm-naïve cells; (2) when
competing to establish a biofilm on a virgin surface, biofilm-aged cells are present at
higher cell densities within the nascent biofilm than their biofilm-naïve competitors; and
(3) biofilm-aged cells are present at higher cell densities than their biofilm-naïve
competitors when naïve and aged populations compete for space within an existing
biofilm.

3.4.1 Single-population biofilms on a virgin surface. One possible phenotype is that
cells aged in a biofilm form more prolific biofilms (that is, biofilms with greater numbers
of cells) when introduced to a virgin surface. As shown in Figure 3.3, cells aged in a
biofilm do not form biofilms with significantly different cell counts than naïve cells.
57

Thus, aging cells in a biofilm does not confer an advantage towards forming biofilms on
a virgin surface.

Figure 3.3: Establishment of naïve and aged biofilm populations on a virgin surface.
Average counts of kanamycin- and rifampicin-resistant naïve strains (gray), all 10-
day-old, all 20-day-old, and all 30-day-old strains harvested from 48-hour biofilms
established on a virgin surface.

3.4.2 Competition for a virgin surface. A second possible phenotype is that cells aged
in a biofilm are better at establishing biofilms on a virgin surface when competing against
naïve cells. Figure 3.4 shows a ratio of aged to naïve cells harvested from a biofilm
established 48 hours after an equal number of both populations were introduced to a
virgin surface. In general, cells aged a biofilm for 10 days show a slight advantage over
their naïve counterparts (median value log
10
(cfu/mL) = 0.3, or two-fold higher than
58

naïve) and cells aged 30 days in a biofilm perform roughly equal to naïve cells (median
value log
10
(cfu/mL) = -0.2, or within two-fold of naïve). Interestingly, cells aged in a
biofilm for 20 days are significantly less represented in the mixed biofilm population than
their naïve counterparts (median value log
10
(cfu/mL) = -1.2, or 15-fold lower than naïve).
Therefore, aging 10 days in biofilm confers a slight advantage when competing for
establishment of a biofilm surface, but aging 20 days in a biofilm confers a competitive
disadvantage, and cells harvested from 30-day-old biofilms are similarly competitive
compared to cells naïve to the biofilm lifestyle.

59

Figure 3.4: Competition of naïve versus aged biofilm populations on a virgin
surface. Biofilm indices of all rifampicin- (white columns) and kanamycin-resistant
(gray columns) 10-day-old (A), 20-day-old (B) and 30-day-old (C) strains against
naïve cells of the opposite marker. Each column represents a single competition.
60

3.4.3 Competition for a biofilm surface. Our third hypothesis is that cells aged in a
biofilm are better at competing for establishment in an existing biofilm, as is the case for
E. coli cells aged in biofilms (Kraigsley and Finkel, 2009). Figure 3.5 shows the ratio of
aged biofilm cells to naïve cells competing for establishment in a 24-hour biofilm.
Similar to competition for a virgin surface, aging cells for 10 days in a biofilm confers a
competitive advantage (median value log
10
(cfu/mL) = 0.9, or eight-fold higher than
naïve), and aging for 30 days is roughly equal to naïve (median value log
10
(cfu/mL) =
-0.2, or within two-fold of naïve). However, aging cells in a biofilm for 20 days confers a
significant disadvantage (median value log
10
(cfu/mL) = -0.5, or three-fold lower than
naïve).

In general, cells aged in a biofilm performed better in the presence of a 1-day-old biofilm
than in the presence of a virgin surface. In the presence of a virgin surface, for example,
10-day-old biofilm-harvested cells are at best ten times more abundant than their naïve
counterparts. In the presence of a biofilm, on the other hand, 10-day-old biofilm-
harvested cells are often at least ten times more abundant than naïve cells, and are present
in excess as much as 1,000 times. Similarly, 20-day-old biofilm-harvested cells are less
competitive in when competing in the presence of a virgin surface than in the presence of
a biofilm. That is, 20-day-old biofilm-harvested cells are often 100 times less present
than naïve cells when forming a biofilm on a virgin surface, and occasionally 10,000
times less present. In the presence of an existing biofilm, 20-day-old biofilm-harvested
61

cells almost always account for at least 1% of the attaching population; naïve cells never
account for more than 99% of the attaching population. Furthermore, 20-day-old biofilm-
harvested cells never account for less than 0.01% (naïve cells never account for more
than 99.99%) of the attaching population.

62

Figure 3.5: Competition of naïve versus aged biofilm populations for establishment
within a 24-hour biofilm. Biofilm indices of all rifampicin- and kanamycin-resistant
10-day-old (A), 20-day-old (B) and 30-day-old (C) strains against naïve cells of the
opposite marker. Each column represents a single competition.
63

In an attempt to determine if the difference in fitness is due to the total amount of time of
incubation, cells were aged serially. That is, cells were aged in a biofilm for 10 days,
harvested, frozen, grown overnight and aged in a new biofilm for 10 days. This was done
for three iterations, for a total of 30 days of aging in a biofilm. In addition, descendants of
the 3x10-day-old serially aged strains were aged for an additional 20 days. Figure 3.6
shows the lineages of aged biofilm populations.

Figure 3.6: Schematic of serially aged biofilms. Strain names refer to their antibiotic
resistance marker (“k” for kanamycin, “r” for rifampicin), the number of days of
continuous incubation (10, 20 or 30), and the independent replicates for each aging
regimen (a-c).

3.4.4 Serially aged biofilm cells. Serially aged 3x10-day-old biofilm populations were
introduced to a virgin surface, similar to the experiments that generated Figure 3.3, and
the results are shown in Figure 3.7. As with populations aged continuously in a biofilm,
64

serially aged biofilm populations show no significant difference in their ability to form
biofilms on a virgin surface.

Figure 3.7: Establishment of unaged and serially aged biofilm populations on a
virgin surface. Cell counts of cells harvested from biofilms established from both
(rifampicin- and kanamycin-resistant) naïve strains, all rifampicin-resistant 3x10
strains and all kanamycin-resistant strains were averaged.

In experiments similar to those presented in Figure 3.4, cell densities of serially-aged
biofilm cells were competed against biofilm-naïve cells for establishment of a biofilm on
a virgin surface. The results of these experiments are shown in Figure 3.8. Unlike
continuously-aged biofilm cells, serially-aged biofilm cells demonstrate a competitive
disadvantage when competing for a virgin surface.
65

Figure 3.8: Competition of naïve versus serially aged 3x10-day-old biofilm
populations on a virgin surface. Biofilm indices of each of the six 3x10 strains in
competition with naïve cells were averaged.

Similar to experiments in Figure 3.5, serially aged 3x10-day-old biofilm cells were
competed against naïve cells within a 1-day-old biofilm. As shown in Figure 3.9, serially
aged biofilm cells show a marked advantage when competing for establishment within an
existing biofilm. Thus, while serially aged cells are less competitive than continuously
aged or naïve cells in the presence of a virgin surface, they are more competitive than
both in the presence of a biofilm. This advantage is far more pronounced than any of the
populations aged continuously within a biofilm.
66

Figure 3.9: Competition of naïve versus serially aged 3x10-day-old biofilm
populations for establishment within a 24-hour biofilm. Biofilm indices of each of
the six 3x10 strains in competition with naïve cells were averaged.

While continuously aging cells in a biofilm for 20 days confers a competitive
disadvantage, aging cells in 10-day increments ameliorates the detrimental effect of aging
continuously for 20 days on competitiveness. To determine whether continuous aging for
20 days adversely affects competitiveness of serially aged biofilm populations, 3x10-day-
old cells were aged for an additional 20 days continuously. As shone in Figure 3.10, an
additional 20 days of aging abrogates the advantage gained by serially aging. Indeed, in
five out of the six serially aged strains, 20 days of continuous aging turned highly
competitive strain backgrounds into competitively disadvantaged cells.
67

Figure 3.10: Competition of naïve versus serially aged 3x10+20-day-old biofilm
populations for establishment within a 24-hour biofilm. Biofilm indices of each of
the six 3x10+20 strains in competition with naïve cells were averaged.

The data from Figure 3.4, 3.5 and 3.10 indicate that something particular about 20 days
of continuous aging within a biofilm confers a heritable disadvantage in competition. We
hypothesized that between 10 and 20 days of continuous biofilms, cells excrete waste or
secrete a signal that encourages cells to leave the biofilm. To test this, several biofilms
were inoculated, and then several more were inoculated 10 days later. After the first set
of biofilms were 20 days old and the second set were 10 days old, the supernatant from
each set was removed and filter sterilized. Half of the 10-day-old biofilm media was then
added three 10-day-old biofilms, while the other half was added to three 20-day-old
68

biofilms. At the same time, half of the 20-day-old media was returned to three 20-day-old
biofilms, while the other half was added to three 10-day-old biofilms. After an additional
two days of incubation, all twelve biofilms were stained with crystal violet. As shown in
Figure 3.11, there were no significant differences between the six 10-day-old biofilms
and between the six 20-day-old biofilms. These results indicate that there is nothing in
the biofilm-conditioned media that alters biofilm staining. Because crystal violet staining
is often used as a proxy for the number of cells within a biofilm, these results are
interpreted to mean that biofilm-conditioned media does not alter biofilm cell density.

Figure 3.11: Crystal violet staining of 10- or 20-day-old biofilms with 10- or 20-day-
old biofilm conditioned media.

69

70

Cells aged continuously for 20 days were are less competitive than naïve cells in the
presence of a virgin surface or a one-day-old biofilm. However, neither of these are the
environments in which 20-day-old biofilm-harvested cells spent the majority of their time
evolving. We hypothesized that 20-day-old biofilm cells have an advantage when
competing for space within a biofilm that more closely resembles the one in which they
evolved, that is, when competing for space within a 20-day-old biofilm. To test this,
biofilms were inoculated 30, 20 and 10 days before competitions were initiated. Naïve
cells were mixed with cells harvested from either a 10-, 20- or 30-day-old biofilm. Each
of these cell mixtures were then aliquoted into each of the 10-, 20- and 30-day-old
biofilms, and competitions were harvested two days later. A heat map of the results of
these competitions is presented in Figure 3.12.

71

Figure 3.12: Heat maps of biofilm competitions in aged biofilms. The order of
magnitude difference of each competition is indicated by color; the more the aged
strain is present, the warmer the color (see key). Three separate strains (a, b and c)
for each drug resistance marker (kanR and rifR) and each age (10, 20 and 30 days)
were competed in triplicate, as shown.

In general, cells aged in a biofilm do not demonstrate an advantage over naïve cells when
competing for space with an aged biofilm, and 20-day-old cells are the most
disadvantaged. The notable exception is in one of the six 20-day-old biofilm-harvested
strains, k20a, which loses against naïve cells when competing from space on a virgin
competition
in 10-day-old
biofilm
#1
competition
in 20-day-old
biofilm
#1
competition
in 30-day-old
biofilm
#1
competition
in 10-day-old
biofilm
#2
competition
in 20-day-old
biofilm
#2
competition
in 30-day-old
biofilm
#2
competition
in 10-day-old
biofilm
#3
competition
in 20-day-old
biofilm
#3
competition
in 30-day-old
biofilm
#3
72

surface or a 1-day-old biofilm. However, in the presence of a 10-, 20- or 30-day-old
biofilm, k20a is at least 100-fold more present than its naïve adversary. These results
indicate that at least one of the six 20-day-old biofilm-harvested strains has likely
generated a mutation distinct from the other 20-day-old strains. Nevertheless, the fact that
five out of the six 20-day-old strains were disadvantaged when competing within a 20-
day-old biofilm, and the sixth strain had a competitive advantage when competing within
any aged (10-, 20- or 30-day-old) biofilm, controverts the hypothesis that cells harvested
from a 20-day-old biofilm are specifically adapted to 20-day-old biofilms.

3.4.5 Subclones of 20-day-old biofilm-harvested populations. As shown in Chapter 2,
subclonal populations of 30-day-old planktonic GASP populations behave significantly
differently from each other when each is introduced into a GASP competition. Given the
difference in competitive fitness between k20a in the presence of aged biofilms, and the
populations harvested from the other five 20-day-old biofilms, we hypothesized that, like
clonal subpopulations from planktonic GASP populations, clonal subpopulations from
20-day-old bioflm-harvested populations would display diverse competition phenotypes.
Eight clones were isolated from each of the six 20-day-old biofilm-harvested populations,
and each clone was competed in triplicate against naïve cells in the presence of a virgin
surface (similar to experiments used to generate Figures 3.4 and 3.8) or within an existing
biofilm (similar to experiments used to generate Figures 3.5 and 3.9). After 48 hours of
competition, biofilms were harvested and cfu/mL from differential antibiotic plating were
73

used to calculate the biofilm indices. The results of these experiments are presented in a
heat map (Figure 3.13).

74

Figure 3.13: Heat map of biofilm clonal subpopulation competitions. Eight clones
isolated from k20a (A), k20b (C), k20c (E), r20a (B), r20b (D) and r20c (F) biofilm
populations. Biofilm indices for all replicates of all experiments were calculated. If
biofilm-harvested cells were present at 10 times lower concentrations than naïve
cells (relative to their initial inoculation), the competition was marked green; the
worse biofilm-harvested cells “lost” in competition, the colder the color (see key).
75

With the exception of a few clones (for example, k20a clone #3), most of the eight clones
for each of the six populations harvested from 20-day-old biofilms behave similarly. As
seen in Figure 3.4 and 3.5, there is a great deal of variation in Biofilm Indices between
replicates of each aged biofilm strain. By comparison, clones of these populations show a
great deal of consistency between replicates. Naïve cells represent 99.99% of biofilms
formed in the presence of a virgin surface, while biofilm clones represent only 0.01% of
these nascent biofilms. Similarly, biofilm clones comprise only 1% of nascent biofilms in
1-day-old biofilm surface competitions, with naïve cells outnumbering clones 100 to one.
In contrast, 20-day-old populations as a whole tend to represent roughly 1% of nascent
biofilms in virgin surface competitions, and roughly 10% of nascent biofilms in 1-day-
old biofilm surface competitions. That is, 20-day-old populations mostly lose to naïve
cells in these competitions, but they form a larger minority of the biofilm than their
clones; most individual clones are much worse losers than the population as a whole.
Furthermore, 20-day-old biofilm-harvested populations occasionally outcompete naïve
cells for space on a virgin surface or within an existing biofilm; their clonal
subpopulations, on the other hand, never win. Thus, 20-day-old populations are more
competitive than most of the clones isolated; therefore, the mutants responsible for the
20-day-old populations’ competitive phenotype are likely present in these populations at
very low frequencies.

76

Taken together, these data indicate that, with few exceptions, there is little phenotypic
variation between clonal subpopulations derived from 20-day-old biofilm-harvested cells.
This differs from clonal subpopulations isolated from planktonic GASP populations,
which vary greatly in competitive phenotype.

3.5 Discussion
Aging S. oneidensis MR-1 in a biofilm does not make these cells better at establishing a
biofilm on a virgin surface. These results are consistent with the fact that aged cells spend
most of the 10, 20 or 30 days evolving in the presence of a biofilm, and relatively little
time in the presence of a virgin surface.

While biofilm biomass is not significantly different, aging cells within a biofilm causes
changes in fitness when these cells compete for a virgin surface. That is, although the
total cell count of the nascent biofilm does not differ, the proportion of aged cells and the
proportion of naïve cells are not equal. Cells aged for 10 days in a biofilm show a slight
advantage when competing for a virgin surface, but 20-day-old biofilm populations are
significantly less fit when competing against naïve cells for establishment on a virgin
surface. If the cells are aged for 30 days, their fitness is on par with that of naïve cells;
that is, the additional 10 days of continuous aging ameliorates the fitness deficit conferred
by 20 days of aging, but does not fully recover the increased fitness phenotype of the 10-
day-old cells.

77

A similar pattern is seen when cells aged in a biofilm competed against naïve cells for
space within an existing biofilm. Continuous aging for 10 days conferred a competitive
advantage, but by 20 days this advantage is abolished – 20-day-old biofilm cells are
competitively disadvantaged, and after 30 days cells have only just recovered from this
disadvantage to be competitively comparable to naïve cells. Although the trend of
competitive fitness and aging regimen is similar whether the competitions are conducted
in the presence of a virgin surface or a 1-day-old biofilm, in general, biofilm-aged cells
performed better in the presence of a biofilm than in its absence. Again, this is consistent
with the fact that biofilm-aged cells spent most of their time evolving in the presence of a
biofilm, and little time evolving in the presence of a virgin surface.

Although 20 days of continuous aging makes biofilm populations less competitive, and
an additional 10 days only recovers enough competitiveness to make cells as competitive
as naïve cells, serially aging biofilms for 10 days three times averts this loss of fitness.
While serially aged populations are worse at competing for a virgin surface, they are
better at competing within the environment in which evolved: the biofilm. Because 3x10
strains are better competitors than their 10-day-old progenitors, it appears there is at least
a partially additive fitness benefit from the additional serial 10-day incubations. This is
consistent with the results of Chapter 2 (Figures 2.2 and 2.3), which indicate that cells
recovered from 30-day-old planktonic cultures have accrued more advantageous
mutations than 10- and 20-day-old planktonic cells. Further investigation is needed to
determine the genetic basis for the competitive advantage of the 10-day-old and 3x10-
78

day-old populations. However, unlike planktonic cells, attempts at directing the evolution
of MFC-inhabiting biofilms may prove more fruitful if the cells in question are
occasionally removed from the MFC biofilm and reinoculated into a new MFC, rather
than aged continuously aged for longer periods of time.

The increased fitness conferred by three serial 10-day incubations is abrogated by an
additional 20 days of continuous aging. Together, these data suggest that 20 days of
continuous aging confers a competitive disadvantage, regardless of strain background.
While it is believed that between 10 and 20 days of static incubation, some cells leave the
biofilm, it is unclear why they do so. It may be that between 10 and 20 days, biofilm cells
have become starved (either for electron donor, electron acceptor, or both), and leave the
biofilm in search of nutrients. Alternatively, new detrital nutrients may become available
between 10 and 20 days, or cells may have developed better ways of utilizing nutrients of
expired cells (via increased expression enzymes involved in amino acid catabolism, for
example). However, switching media conditioned from 10- and 20-day-old biofilms did
not significantly alter biofilm staining; it is unlikely that anything in the medium –
nutrients or signals – is encouraging cells to leave the biofilm, or select for mutations that
disadvantage these cells in competition. Nevertheless, the results of crystal violet staining
should be confirmed with planktonic and biofilm cell counts. It should also be noted that,
although these populations change over time, both biofilm and planktonic populations are
detected. While moving from the biofilm to a planktonic lifestyle may become desirable,
79

these two populations may have a carrying capacity that prevents all cells from becoming
planktonic.

Still other possibilities exist that might cause the competitive disadvantage of 20- and
3x10+20-day-old biofilm cells. Further investigation is needed to determine the genetic
basis for the competitive disadvantage conferred by 20 days of continuous aging within a
biofilm. Based on the data presented in Figure 3.13, 20-day-old biofilm-harvested
populations are not likely to be genotypically diverse, as subclonal populations did not
show diverse competitive phenotypes (save for a few exceptions). It is possible that when
a mutation conferring a competitive advantage to cells aged in a biofilm for 10 days
appears, it sweeps through the population, driving down the population’s genetic
diversity. Clearly, 20 days of continuous aging within a biofilm applies a strong selective
pressure that favors cells that are competitively disadvantaged in the presence of a new
biofilm. This selective pressure may cull much of the diversity from the population.
Competing clonal subpopulations isolated from 10-day-old biofilm-harvested strains will
determine if phenotypic diversity is present at 10 days and lost thereafter, or if it is not
present in cells harvested from 10-day-old cells.

One clone isolated from one of the 20-day-old biofilm-harvested populations is of
particular interest. While all of these populations are competitively disadvantaged in the
presence of virgin surface or a 1-day-old biofilm, one such population (k20a) has a
competitive advantage in the presence of 10-, 20- and 30-day-old biofilms. Furthermore,
80

one of the eight subclonal populations isolated from k20a (#3) is significantly better in
competition than the other seven clones. This clone may be responsible for the
competitive advantage of k20a in the presence of aged biofilms, an advantage not seen in
the other five similarly aged populations. Whole genome resequencing may assist in
determining the genetic difference between k20a clone #3, the other clones of k20a and
the other populations harvested from 20-day-old biofilms. Furthermore, this method may
also help elucidate the genetic basis of the virgin- and new biofilm-surface competitive
disadvantage conferred by 20 days of continuous aging within a biofilm. It may also help
determine the genetic basis of the competitive disadvantage in the presence of aged (10-,
20- and 30-day-old) biofilms.

At this time, it is unknown by what mechanism cells aged in a biofilm outcompete their
naïve counterparts for establishment on a virgin surface or within an existing biofilm. In
the case of the virgin surface competitions, it may be that aged cells are better at initial
attachment, but are otherwise prevented from accruing more biofilm biomass (possibly
due to nutrient limitations in the media). Another possibility is that aged populations
excrete a signal that discourages unaged cells from forming or proliferating within a
biofilm. These possibilities are not mutually exclusive, and other mechanisms may exist.
Microscopy experiments will need to be performed to determine how aged cells are able
to outcompete unaged cells when colonizing a virgin surface. Comparing the genotypes
of 10-day-old and 3x10-day-old biofilms may be fruitful in determining the genetic basis
of increased competitiveness in the presence of a virgin surface, as the former is
81

competitively advantaged in the presence of both virgin and 1-day-old biofilm surfaces,
while the latter is only advantaged in the presence of a biofilm surface.

In the case of the existing biofilm, it could be that evolved cells are better able to occupy
interstitial spaces within an established biofilm (that is, they are better able to “colonize
the valleys”) (Lima, Acevedo and Finkel, unpublished). Another possibility is that
evolved cells aren’t necessarily better at attaching to existing biofilms, but once they do,
they are better at proliferating within the biofilm. A third, non-mutually exclusive
possibility is that evolved cells are better at insinuating themselves within heavily
populated areas of the biofilm and perhaps even displacing cells already occupying these
spaces. Future experiments using fluorescently labeled aged and naïve populations will
determine the mechanism by which aged cells are able to outcompete unaged cells.

To the best of our knowledge, this is the first demonstration of real-time evolution in S.
oneidensis MR-1 biofilms. The heritable phenotype only manifests an advantage in
competition, similar to a GASP-like phenotype. The phenotypes we see in biofilms are
distinct from those of mutants isolated from any of our planktonic cultures, but may offer
useful guidance for directed evolution in microbial fuel cells. These biofilms were
incubated in the presence of an electrically inert surface; future work is needed to
determine whether a similar pattern of competitive fitness is conferred by aging cells in
the presence of an electrically conductive surface. Furthermore, future investigation is
82

needed to determine whether the use of surfaces poised at different electrical potentials
results in biofilm phenotypes similar to the phenotypes presented here.

3.6 Conclusion
Aging cells in a biofilm for 10 days confers a competitive advantage, aging for 20 days
confers a competitive disadvantage, and aging for 30 days renders S. oneidensis MR-1
competitively neutral (Figure 3.4 and 3.5). While this general trend holds whether
biofilm-aged cells compete in the presence of a virgin surface or in the presence of a 1-
day-old biofilm surface, biofilm-aged cells tend to be more competitive in the presence of
a 1-day-old biofilm. According to the median values, 10-day-old biofilm-aged cells
outnumber naïve cells 2-to-1 and 8-to-1 when competing for establishment on a virgin
and biofilm surface, respectively. Cells aged in a biofilm for 20 days are outnumbered by
naïve cells 15-to-1 and 3-to-one (medians) in the presence of a virgin and biofilm surface,
respectively, and 30-day-old biofilm-aged cells are within a 2-fold range (that is, within
experimental error) of the naïve cells in both conditions.

Aging 20 days within a biofilm confers a competitive disadvantage regardless of strain
background, but this can be avoided by serial aging. That is, cells aged serially three
times for ten days were competitively disadvantaged (Figure 3.8) in the presence of a
virgin surface (unlike cells aged 10- or 30-days continuously), but in the presence of a 1-
day-old biofilm surface, these serially-aged strains are the most competitively advantaged
of any we’ve isolated to date (Figure 3.9). But whether naïve or competitively-
83

advantaged serially-aged strains are aged continuously for an additional 20 days, 20 days
of continuous aging within a biofilm confers a competitive disadvantage (Figure 3.10).

Eight clones picked at random from a planktonic GASP populations yielded at least four
distinct phenotypes (2.10). Indeed, some clones have a greater competitive fitness than
the parent population. By contrast, the eight clones within each 20-day-old biofilm-aged
population have – with few exceptions – similar phenotypes (3.13). All 20-day-old
biofilm clones are competitively disadvantaged, and most are poorer competitors than the
parent population. Aged biofilm populations appear to be less phenotypically diverse than
aged planktonic populations.

84

Chapter 4: Addition of amino acids and biofilm formation of
Shewanella oneidensis MR-1

4.1 Overview
Due to their ability to respire distant and diverse electron acceptors, Shewanella species
play an important role within biofilm communities, whether these biofilms are attached to
environmental or technological surfaces, such as the electrodes of microbial fuel cells.
However, little is known about how Shewanella biofilms form and disperse, even over
relatively short timescales. Here we show the rapid formation, detachment and
reattachment of Shewanella oneidensis MR-1 biofilms, and their stratification over an
oxygen gradient. Because of their myriad potential roles (e. g., in catabolism, anabolism,
respiration via the Stickland reaction, signaling), amino acids were added singly and in
pairs to test their effect on developing and established biofilms. The addition of certain
pairs of amino acids affect biofilm development, while other pairs affect biofilms only
when added after biofilms are established. Furthermore, some amino acid pairs only
affect oxygen-replete biofilms, while others only affect oxygen-deprived biofilms. These
data indicate a role for amino acids in the dynamic and physiologically distinct nature of
S. oneidensis MR-1 biofilms.

4.2 Introduction
Biofilms are surface-attached communities of microorganisms consisting of cells
surrounding by an extracellular matrix (Monroe, 2007; Bazaka, et al., 2011; Mosier and
85

Cady, 2011; Trevors, 2011). The formation of the biofilm involves distinct stages of
attachment, proliferation, maturation and dispersal (Hall-Stoodley, et al., 2004). Each of
these stages is the result of environmental cues and differential gene expression working
in concert (Karatan and Watnick, 2009).

Much remains to be elucidated regarding the stages of biofilm formation and dispersal of
the Shewanella genus. The genus is not only ubiquitous in natural environments (lakes,
sediments and oceans) (Fredrickson, et al., 2008), but is also found human-created
surfaces (Waters, et al., 2009; Biffinger, et al., 2011), and is even shown to be involved in
pathogenesis (Yim, et al., 2010). Not only is their distribution pervasive, but Shewanella
species play a special role in the environments they occupy. Their ability to conduct
electrons to distant (Torres, et al., 2009) and varied (Deutschbauer, et al., 2011) acceptors
make Shewanella physiology of special interest from both an environmental and a
technological perspective (Nealson and Saffarini, 1994). Here we examine the
attachment, proliferation, detachment and reattachment of S. oneidensis MR-1 biofilms as
a function of time and nutrient availability.

Organisms within a biofilm face different physiological challenges than their planktonic
counterparts, and biofilms deploy unique strategies for coping with these challenges. For
example, it has been shown in single species (Rani, et al., 2007) and multispecies
biofilms (Yu, et al., 2004; de la Rosa and Yu, 2005; Ahimou, et al., 2007) that dissolved
oxygen decreases with increasing biofilm depth. It is likely that at least some cells within
86

our biofilms experience microaerobic or even anaerobic conditions. These biofilms must
employ strategies other than the reduction of oxygen to maintain intracellular
electrochemical gradients.

In addition to their ability to utilize a variety of soluble and insoluble electron acceptors,
species within the Shewanella genus may have other ways of generating ATP. In the
genus Clostridium, certain pairs of amino acids can be used in what is known as the
Stickland reaction, where the oxidation of one amino acid is coupled with the reduction
of another amino acid. In the process, ADP is converted to ATP by the addition of
inorganic phosphate (Stickland, 1934).

To test the hypothesis that biofilm-associated cells utilize amino acids for respiration,
both S. oneidensis MR-1 biofilms that had access to abundant oxygen, as well as biofilms
with reduced oxygen abundance, were supplemented with all 20 amino acids singly and
in pairs.

4.3 Materials and Methods
4.3.1 Strains, culture media and chemicals. S. oneidensis MR-1 was provided by Ken
Nealson. All cultures were inoculated from 20% glycerol frozen stocks and grown at
30
o
C for 18 hours in 5mL Luria-Bertani (LB) broth with aeration in a New Brunswick
TC-7 roller drum. Of these 18-hour cultures, 5µL were used to inoculate 5mL of fresh LB
87

cultures, which were incubated on the roller drums for an additional 18 hours. At the end
of the incubation, the cultures were diluted 1:100 (v:v) of Minimal Medium (Bretschger,
2007), and this mixture was used to inoculate all biofilms.

4.3.2 Biofilm incubation and harvesting. Acid-washed borosilicate glass test tubes were
inoculated with 1mL of the diluted cell mixture, and incubated statically at 30
o
C. In
biofilms that were supplemented with amino acids, amino acids were added to the 1mL
cultures to a final concentration of 10mM. Amino acids were either added at inoculation,
and biofilms harvested after 10 hours of incubation, or added after 10 hours of
incubation, and harvested following an additional 17 hours of incubation (27 hours total).
At the end of the experiment, biofilms were stained with 0.1% crystal violet for 30
minutes, then rinsed with diH
2
O to remove excess stain. After stained biofilms were
photographed, 0.5mL of 95% ethanol was added for 15 minutes to solubilize the CV
staining the lower the biofilm band. After the solubilized stain was removed and
quantified, an additional 1.25mL of 95% ethanol was added to solubilize the CV staining
the upper biofilm band. Absorbance was measured at 590nm.

4.4 Results
To characterize the nature of S. oneidensis MR-1 biofilms, spontaneous biofilm
formation was assessed. These biofilms were then observed following addition of amino
acids, either when the biofilms were first initiated, or after the biofilms were well-
established.
88

4.4.1 Biofilm formation. Figure 4.1 shows biofilms sacrificed after 4, 6, 8, 10, 12, 14,
16, 18, 20, 22, or 24 hours of incubation (Figure 4.1A) and stained with crystal violet.
Additionally, 27 biofilms were initiated at the same time, and three were sacrificed and
stained after 1, 2, 3, 5, 7 10, 14, 20, or 33 days. One of the most striking features of
biofilms grown in these conditions is the spontaneous formation of two rings. The top
ring is located at the air-medium interface, where oxygen and nutrients are both relatively
plentiful. The bottom ring consistently forms roughly 1cm below the air-medium
interface.

Figure 4.1: Crystal violet staining of biofilms over a period of hours and days.
Biofilms were inoculated simultaneously, and three were sacrificed every two hours
during the 24-hour timecourse (A) or every few days during the 33-day timecourse
(B).
89

90

The staining of the top and bottom ring can change significantly between two-hour
intervals, and relative to each other. While the precise timing of these fluctuations varies
from experiment to experiment, the rapid and reversing shifts in staining of both rings is
reproducible. Based on these data, we believe that a large portion of the biofilm detach,
only to reattach to the biofilm within a very short time span. These shifts also occur over
timescales relevant to biofilm evolution, as shown in Chapter 3.

We hypothesized that the rapid detachment and reattachment of biofilm cells was in
response to changes in the availability of certain nutrients within the media. Because of
the myriad of roles amino acids play in growth, metabolism, and even respiration, media
were supplemented with 10mM of each of the twenty amino acids singly and pair-wise.
To assess the effect of amino acid supplementation on initial biofilm formation, the
media was supplemented at inoculation, and biofilms were harvested at a time when
biofilms were found to be consistently dense (10 hours). To investigate the effect of
amino acid supplementation on established biofilms, amino acids were added 10 hours
after inoculation and biofilms were stained after an additional 17 hours of incubation
(when the age of the biofilms was 27 hours total).

A heat map of the all the amino acid combinations are presented in Figure 4.2. Because
of the variations in staining between experiments, all data were normalized to the
unsupplemented control for each experiment. For each experiment, if the normalized data
fell within the range of the triplicate control, the replicate was considered neutral
91

(yellow). If the replicate fell below the lowest value of the controls, the biofilm was
declared to be thinner (red), and if it fell above the range of the controls, the biofilm was
declared to be thicker (blue). Both the top and bottom rings were considered separately
for all experiments. Thus, for each amino acid pair, four conditions were examined: (1)
the bottom biofilm ring when amino acids were added at inoculation and biofilms were
harvested at 10 hours, (2) the top biofilm ring when amino acids were added at
inoculation and biofilms were harvested at 10 hours (from the same tube as 1), (3) the
bottom biofilm ring when amino acids were added 10 hours after inoculation and
biofilms were harvested after a total of 27 hours, and (4) the top biofilm ring when amino
acids were added 10 hours after inoculation and biofilms were harvested after a total of
27 hours (from the same tube as 3). For each amino acid pair in Figure 4.2, the bottom
left quadrant contains up to six replicates for the first condition (bottom 10 hours), the top
left quadrant is for the second condition (top 10 hours), the bottom right quadrant is for
the third condition (bottom 27 hours), and the top right quadrant is for the fourth
condition (top 27 hours).

Figure 4.2: Heat map of all replicates of all supplementary amino acid pair
combinations. Each biofilm was either declared below the range of the controls
(red), within the range of the controls (yellow), or above the range of the controls
(blue). Each experiment was done in triplicate, and most of the experiments were
repeated. In the key, “amino acid 1” refers to the amino acid by vertical column,
and “amino acid 2” refers to the amino acid by horizontal row. For each amino acid
combination, the upper left six replicates correspond to the three top biofilm rings
(a-c) in each experiment (1 and 3) when amino acids are added at inoculation and
harvested after 10 hours of incubation. The lower left six replicates are the
corresponding bottom biofilm rings (1a-c and 3a-c). The upper and lower right six
replicates correspond to the top and bottom biofilm rings, respectively, when amino
acids were added after biofilms were established for 10 hours and harvested after 27
hours incubation total (2a-c and 4a-c).
92

93

Some combinations, such as asparagine plus isoleucine, had a generally negative effect
on the biofilms (as illustrated in the right half of the first line of Table 4.1), while some
combinations, such as histidine plus tyrosine or isoleucine plus lysine, had a generally
positive effect (as illustrated in the left half of the first line of Table 4.1).

Other combinations only have an effect on biofilms that were already established (27
hour), such as cysteine plus methionine or glutamine plus glycine, which have a positive
effect, and alanine plus methionine or asparagine plus glutamate, which have a negative
effect. Still other combinations only affect the biofilms if added at inoculation (10 hours),
such as arginine plus methionine (positive), or cysteine alone (negative).

Regardless of when the amino acids were added, certain combinations affect only the top
ring, such as alanine plus valine (positive) or asparagine plus histidine (negative). Yet
other combinations affect the bottom ring at both times, such as threonine plus valine
(positive) and alanine plus glycine (negative).

Still other combinations only had a positive effect on one of the four conditions, such as
asparagine plus serine on the bottom 10-hour biofilm, asparagine plus tyrosine on the top
10-hour biofilm, alanine plus proline on the top 27-hour biofilm, and aspartate plus
glutamine on the bottom 27-hour biofilm.

94

Certain combinations only had a negative effect on one of the four biofilms, such as
asparagine plus glycine on the bottom 10-hour biofilm, leucine alone on the top 10-hour
biofilm, proline plus tyrosine on the top 27-hour biofilm, and alanine plus leucine on the
bottom 27-hour biofilm.

From these data, it is clear that no amino acid has a general effect on any of the biofilms
tested. For example, while alanine has a positive effect on established top biofilms in the
presence of valine or proline, it does not have this effect when it is added alone, or in the
presence of the other amino acids. Given that unestablished, established, top and bottom
rings were each affected separately by certain amino acid combinations, it appears that
these four biofilms are physiologically distinct. The notable amino acid pairs and their
effects are summarized in Table 4.1.
95

Table 4.1: Summary of notable amino acid pairs and their effects on the four
biofilms. Icons at the top right of each entry indicate which biofilms are affected
(bottom right quadrant - bottom 10 hour biofilm; top right quadrant – top 10 hour
biofilm; bottom left quadrant – bottom 27 hour biofilm; top left quadrant – top 27
hour biofilm), and whether the effect is positive (blue), neutral (yellow), or negative
(red).

4.5 Discussion
Given the high relative abundance of electron acceptor (oxygen) and nutrients at the air-
medium interface, the presence of the top ring is unsurprising. The consistent presence of
the bottom ring, however, is unexpected. In the absence of the upper biofilm and
planktonic cells scavenging oxygen, the lower ring still experiences reduced oxygen
bottom threonine + valine alanine + glycine
positive negative
developing arginine + methionine cysteine
established
cysteine + methionine
glutamine + glycine
alanine + methionine
asparagine + glutamate
all
histidine + tyrosine
isoleucine + lysine
asparagine + isoleucine
developing bottom asparagine + serine asparagine + glycine
established bottom aspartate + glutamine alanine + leucine
developing top asparagine + tyrosine leucine
established top alanine + proline proline + tyrosine
top alanine + valine asparagine + histidine
96

availability. At this temperature and salinity, the concentration of dissolved oxygen in
uninoculated media is roughly half of that of air (Geng and Duan, 2010). It is therefore
likely that the differences in oxygen availability makes the top and bottom biofilms
physiologically distinct. This is supported by the unique effect certain amino acid
combinations have on the two biofilms, both when they are first forming, and after they
are established.

Another striking feature of the data is the rapid dissociation and reattachment of cells to
and from the biofilm. It is possible that a subset of the population is transiently associated
with the biofilm. As discussed in Chapter 3, it is unknown at this time whether exodus
from the biofilm is a response to stressful conditions, or whether biofilm formation is a
strategy for coping with stress. However, the data presented above show that the behavior
of these cells in this system can be explained, at least in part, by the availability of certain
amino acids in the medium.

Supplementary amino acids could play a number of different roles in our system: in (1)
catabolism, as a carbon and energy source, in (2) anabolism, as for example the building
blocks for expressed proteins, as (3) proton and electron donors in the production of ATP,
and possibly other functions.

It is possible that in the top biofilms, where oxygen is plentiful but cells have been
consuming nutrients, certain amino acids may be preferentially scavenged as a carbon
97

source. As determined by mass spectrometry, in shaking minimal media cultures,
supplementary asparagine, cysteine, glutamate, glutamine, histidine, isoleucine, leucine,
serine and threonine were each consumed immediately along with lactate, and all but
asparagine, cysteine and histidine were completely exhausted by 20-35 hours (Hwang,
2002). Alanine, arginine, glycine and valine are each utilized after lactate is partially
consumed (around 10 hours), while phenylalanine and tyrosine were only used after
lactate was exhausted (starting at 20 and 15 hours, respectively), and lysine, methionine,
proline and tryptophan were not consumed at all (Hwang, 2002).

Asparagine plus histidine has a negative effect on the top ring (both at 10 and 27 hours),
and proline plus tyrosine has a negative effect on established top rings (top 27 hours).
These data support the notion that conditions of plentiful nutrients encourage the
planktonic lifestyle. However, phenylalanine by itself has a positive effect on established
top biofilms, supporting the notion that plentiful nutrients encourage biofilm association.

Surprisingly, glutamine plus glycine has a positive effect on established biofilms, while
glycine plus histidine and histidine plus tyrosine have a positive effect on established top
rings. The top ring should be consuming glutamine, histidine and threonine immediately;
however, biofilm staining may not immediately (or ever) reflect aerobic consumption of
these amino acids. Furthermore, cysteine plus methionine and isoleucine plus lysine had a
positive effect on established biofilms, even though cysteine and isoleucine should be
consumed immediately, while lysine and methionine should not be consumed at all. At
98

this point it is unclear if the top or established biofilms utilize any or all of these amino
acids as carbon sources, or if these effects are the result of some other response, such as
respiration.

In bottom biofilms, where there is less oxygen available, it is possible that certain amino
acid pairs are preferred for the Stickland reaction, in which electrons are transferred
between amino acid pairs within the cytoplasm. This may allow cells to generate ATP
and maintain their electrochemical gradients where electron acceptors may be scarce. In
the Clostridium genus, alanine, arginine, aspartate, cysteine, glutamate, isoleucine, lysine,
methionine, serine and valine act as electron donors, while glycine and proline act as
electron acceptors, and leucine, phenylalanine, threonine, tryptophan and tyrosine can act
as both donors and acceptors (Nisman, 1954; Ramsay, 1997). While none of these appear
to have universal effects, certain combinations do affect the bottom ring, such as
threonine plus valine (positive) and leucine plus proline (negative). At this point, it is
unclear if the bottom biofilms are using amino acids in redox reactions.

However, this does not exclude the possibility that bottom biofilms may also use
supplementary amino acids as a carbon and energy source. Because the production of
ATP per unit carbon source is less efficient when oxygen is not the terminal electron
acceptor, anaerobic cells require more carbon source to produce the same amount of ATP
as their aerobic counterparts. Furthermore, it is known that certain amino acids can be the
sole carbon source when electron acceptors other than oxygen are available. For example,
99

S. oneidensis MR-1 can use alanine, arginine, glycine, isoleucine, leucine, lysine,
methionine or valine as the sole carbon source when the only electron acceptor is TMAO
(Ringø, et al., 1984). However, of these amino acids, only valine has a uniquely bottom
biofilm effect, and only in the presence of threonine. Thus, it is unclear if bottom biofilms
use amino acids as a carbon source.

Amino acids shown to be consumed aerobically were predicted to positively affect the
top biofilm; if consumed immediately, they were predicted to affect top developing (top
10 hour) biofilms, while if they were consumed after 10 hours, they were predicted to
affect top established (top 27 hour) biofilms. Because reduced oxygen availability makes
bottom biofilm cells electron acceptor limited, amino acids known to serve as electron
donors in the Stickland reaction were predicted to have a negative effect on bottom
biofilms. Similarly, amino acids known to be electron acceptors in the Stickland reaction
were predicted to have a positive effect on electron acceptor-limited biofilms, i. e.,
bottom biofilms. The roles each amino acid plays in consumption and respiration, as well
as their the predicted function in our biofilms, is summarized in Table 4.2.

100

Amino acid
Aerobically
consumed
1

Anaerobically
consumed
2

Electron
donor
3,4

Electron
acceptor
3,4

Predicted
effect
Alanine after 10hrs √ √

Arginine after 10hrs √ √

Asparagine immediately

Aspartate √

Cysteine immediately √

Glutamate √

Glutamine immediately

Glycine after 10hrs √ √ √

Histidine immediately

Isoleucine immediately √ √

Leucine immediately √ √ √

Lysine never √

Methionine never √ √

Phenylalanine after 20hrs √ √

Proline never √

Serine immediately √

Threonine immediately √ √

Tryptophan never √ √

Tyrosine after 15hrs √ √

Valine after 10hrs √ √

Table 4.2: Summary of amino acid consumption, utility as Stickland donors and
acceptors, and predicted effects. Icons at the top right of each entry indicate which
biofilms are affected (bottom right quadrant - bottom 10 hour biofilm; top right
quadrant – top 10 hour biofilm; bottom left quadrant – bottom 27 hour biofilm; top
left quadrant – top 27 hour biofilm), and whether the effect is positive (blue),
neutral (yellow), or negative (red).Data collated from Hwang, 2002 (1), Ringø, et
al., 1984 (2), Nisman, 1954 (3), and Ramsay, 1997 (4).
101

Further investigation is needed to determine the role certain amino acid pairs play in
developing, established, oxygen-replete and oxygen-limited biofilms. One approach is to
apply mass spectrometry to these cultures (Li, et al., 2012). This method can assess the
concentration of amino acids in the media, as well as respiratory byproducts predicted to
arise from the Stickland reaction. Another approach is to monitor
13
C-labeled amino acids
in situ via NMR coupled with confocal microscopy to monitor both the spatial
distribution of the amino acids and the structure of the biofilm (Majors, et al., 2005;
McLean, et al., 2008). This will determine whether certain amino acids are being
consumed. It may also show whether cysteine plus methionine, for example, encourages
planktonic cells to join the biofilm, or encourages further proliferation of the biofilm. It
may determine whether alanine plus methionine discourages planktonic cells from
joining the extant biofilm, or discourages the proliferation of cells already associated with
the biofilm.

The above data clearly demonstrate the ability of S. oneidensis MR-1 to rapidly form,
detach from and reattach to biofilms. Under these conditions, biofilm rings spontaneously
form both where oxygen is plentiful, and where it is restricted. The timing of detachment
and reattachment differ between these two rings, and the two rings respond to different
supplementary amino acids. Taken together, these data suggest that the two rings are
physiologically distinct.

102

4.6 Conclusion
When incubated statically in the presence of an electrically inert surface (glass) S.
oneidensis MR-1 biofilms form spontaneously both where oxygen is plentiful and where
it is scarce. Both biofilm populations are highly dynamic, and a significant number of
cells detach from these biofilms within short (2 hour) intervals, only to reattach a few
hours later (Figure 4.1). Future experiments are needed to determine whether biofilms
attached to electrically active surfaces are similarly dynamic.

Oxygen-replete and oxygen-limited biofilms respond to different pairs of amino acids.
For example, asparagine plus histidine affects the biofilm ring at the air-medium
interface, both when biofilms are developing and after biofilms have established. By
contrast, threonine plus valine only affects biofilms that form approximately a centimeter
below the air-medium interface. That these biofilms respond to different amino acids
support that notion that oxygen-replete and oxygen-poor biofilms are physiologically
different.

Developing and established biofilms respond to different amino acid pairs. For example,
arginine plus methionine affects both oxygen-rich and oxygen-poor biofilms, but only if
added when biofilms are first forming. By contrast, asparagine plus glutamate affects
biofilms irrespective of oxygen availability, but only if added after the biofilms have
established. That biofilms respond to different amino acid supplementation at different
times indicate that different stages of biofilm development are physiologically distinct.
103

Chapter 5: Conclusions and Future Directions

5.1 Planktonic long-term survival and evolution
The long-term survival and evolution of planktonically-grown Shewanella oneidensis
MR-1 is demonstrated in a number of different conditions (Figure 2.1). Not only do cells
display the GASP phenotype when aged in cultures replete with oxygen and nutrients
(aerobic LB), but also where either electron donor (nutrients) or electron acceptor
(oxygen) are limited (aerobic Minimal Medium and anaerobic LB, respectively). For
each condition, 18 independently aged populations all indicate the same general trend:
the longer cells are aged in a given condition, the greater the advantage they enjoy when
competing in that condition (Figures 2.2 and 2.3). To our knowledge, this is the first
demonstration of real-time evolution of the Shewanella genus. Furthermore, GASP
populations were determined to be phenotypically and genotypically diverse (Figure
2.10).

Cells cultured in conditions of both electron donor and electron acceptor restriction
(anaerobic Minimal Medium) experience a precipitous reduction in viable cell counts
during death phase, maintain relatively low density during long-term stationary phase,
and did not display the GASP phenotype (Figure 2.1 and 2.2). Because the addition of
electron acceptor (fumarate), but not donor (lactate) or acceptor and donor together,
prolongs stationary phase, we conclude that the peculiar nature of death phase in these
conditions is due electron acceptor limitation (Figure 2.4). However, while increasing
104

initial concentrations of fumarate or adding additional fumarate during stationary phase
delays death phase, these treatments do not alter the proportion of cells lost when death
phase finally occurred, nor increase cell densities during long-term stationary phase.
GASP populations are not isolated from anaerobic Minimal Medium cultures in which
stationary phase was prolonged (Figure 2.5); however, it is unlikely that such mutants
arise during stationary phase.

5.1.1 Planktonic long-term survival and evolution future work: genetic basis
of GASP
Now that several evolved strains of S. oneidensis MR-1 have been isolated, the genetic
basis of their competitive advantage can be addressed. In E. coli, GASP can be conferred
by the attenuation of the stress-induced RNA polymerase sigma factor, RpoS; however,
RpoS attenuation is not required for GASP (Zinser and Kolter, 2000; Martínez-García, et
al., 2003). Here, several lines of evidence suggest that many different mutations confer
the GASP phenotype in S. oneidensis MR-1. Firstly, several of the eight clonal
subpopulations isolated from a GASP population display markedly different GASP
phenotypes (Figure 2.10). Secondly, if the mutation that conferred GASP was the same in
all populations, populations aged in aerobic Minimal Medium would be expected to
outcompete the 1-day-old majority in aerobic LB, and vice versa, but this is not always
the case (Figure 2.9). Thirdly, like E. coli (Finkel and Kolter, 1999; Zinser and Kolter,
1999; Zinser and Kolter, 2000; Zinser and Kolter, 2004; Finkel, 2006), older aged
populations of S. oneidensis MR-1 outcompete younger aged populations (Figure 2.3).
105

Given the increased fitness we see the longer S. oneidensis is aged, and the fact that it is
likely that these cells have also accrued multiple mutations responsible for their
competitive advantage. Genomic resequencing on the clonal subpopulations is underway
to determine the genetic basis of the GASP phenotype in planktonically aged S.
oneidensis MR-1. Identified candidates can then be cloned into unaged genetic
backgrounds to confirm their conferral of the GASP phenotype.

5.1.2 Planktonic long-term survival and evolution future work: amino acid
supplementation in anaerobic Minimal Medium
The addition of an electron acceptor (fumarate) prolongs stationary phase in anaerobic
Minimal Medium cultures (Figure 2.4 and 2.6). However, even in anaerobic Minimal
Medium cultures that are not spiked with fumarate, cell densities partially recover after
death phase, even after falling below the limit of detection (Figure 2.8). These data
suggest that previously inaccessible electron acceptor is made available after death phase.
Surviving cells may be using the detritus of expired cells as electron acceptor, such as
fumarate from the TCA cycle, or possibly amino acids from degraded proteins. Based on
the data presented in Chapter 4, S. oneidensis MR-1 is hypothesized to utilize certain
amino acids as electron acceptors. One way to test this is to spike anaerobic Minimal
Medium cultures with filter-sterilized post-death phase anaerobic Minimal Medium, as
well as individual amino acids, instead of fumarate. The results of these experiments will
determine whether post-death phase recovery in anaerobic Minimal Medium cultures is
106

the result of new electron acceptor availability, and if S. oneidensis MR-1 can use amino
acids as electron acceptors.

5.1.3 Planktonic long-term survival and evolution future work: GASP in
anaerobic Minimal Medium
Of all the conditions tested, the only one in which a GASP mutant was not isolated was
anaerobic Minimal Medium. Mutation frequencies are affected by culture conditions
(Koskiniemi, et al., 2010); thus, it is formally possible that GASP mutants are not
generated in anaerobic Minimal Medium. However, the fact that GASP mutants arise in
conditions similar to anaerobic Minimal Medium (i. e., anaerobic LB and aerobic
Minimal Medium) renders this possibility unlikely. Therefore, we hypothesize that GASP
mutants appear in the population either before or after death phase. If GASP mutants are
present in anaerobic Minimal Medium populations at densities of less than 10,000
cfu/mL, then they are likely to become extinct during death phase before the selective
pressures of long-term stationary phase allow these mutants to increase in frequency. If
GASP mutants are not generated until after death phase, the low long-term stationary
phase population density makes the spontaneous appearance of an advantageous mutant
unlikely to occur within the span of the 30-day experiment. Increasing the total size of the
population (e. g. from 5mL to 50mL) will test the former possibility; even if only one in
10,000 cell survive death phase and GASP mutants are present at a frequency of 1,000
cfu/mL, there should still be roughly five GASP mutants per 50mL culture that survive
107

the population bottleneck. Maintaining higher population densities (perhaps by periodic
spiking of fumarate) after death will test the latter possibility.

5.2 Biofilm evolution
In addition to planktonic evolution, the evolution of S. oneidensis MR-1 in the biofilm
lifestyle was demonstrated. Generally, the longer cells are aged planktonically, the more
competitive they were in that environment. This trend does not hold in biofilm-aged cells.
Cells aged in a biofilm for 10 days are more competitive, but 20-day-old biofilm-aged
cells are outcompeted by cells that never before saw a biofilm, and aging for 30 days only
partially restored the GASP-like phenotype conferred by 10 days of aging (Figure 3.4 and
3.5). Intriguingly, this trend does not just hold when biofilm-aged cells compete in an
environment similar to that in which they spent the majority of their time evolving (a 1-
day-old biofilm); the trend also holds when biofilm-aged cells compete in an environment
in which they spent relatively little time (a virgin surface). Serially-aged (3x10-days)
biofilm cells lose to naïve cells in the presence of a virgin surface (Figure 3.8), but in the
presence of an existing biofilm, they are the most competitive strains isolated (Figure
3.9). Like 10-day-old strains, the competitive advantage of 3x10-day-old cells is
abolished by an additional 20 days of continuous aging (Figure 3.10). The latter result
suggests that 20 days of continuous aging within a biofilm encourages a particular
adaptation that makes cells disadvantaged in our competitions.

108

5.2.1 Biofilm evolution future work: genetic basis of GASP
We isolated several strains that have an advantage when competing for space within a
biofilm in the presence of a virgin surface (Figure 3.4) or an existing biofilm (Figure 3.5).
The genetic basis for this advantage can now be addressed. Whole-genome resequencing
is one approach which can be used to discover mutations that may be responsible for the
competitive advantage seen in 10-day-old and 3x10-day-old biofilm-harvested cells. On a
virgin surface, 3x10-day-old cells are less competitive than naïve cells (Figure 3.8),
whereas their 10-day-old progenitors are more competitive; however, in the presence of
an existing biofilm, 3x10-day-olds are more competitive than their 10-day-old
progenitors (Figure 3.9). Taken together, these data suggest that multiple mutations are
responsible for the phenotypes we see in 3x10-day-old cells. Given the apparently low
phenotypic diversity of clones isolated from 20-day-old biofilm-harvested populations
Figure 3.13), continuously aged biofilm cells, whole genome resequencing may also be
fruitful in determining the genetic basis of competitive disadvantage in the presence of a
1-day-old biofilm. That is, this approach may elucidate the important genetic difference
between 20-day-old and 3x10+20-day cells (Figure 3.10), and biofilm-naïve cells.
Furthermore, it appears while most 20-day-old biofilm-harvested cells are disadvantaged
regardless of whether they are competing for space in a virgin surface, within a 1-day-old
biofilm or an aged biofilm, some 20-day-old biofilm-harvested cells have a competitive
disadvantage in the presence of a virgin surface or a 1-day-old biofilm, but a competitive
advantage in the presence of aged biofilms (Figure 3.12). If the genetic basis of the latter
109

could be determined, it could be cloned into a 3x10-day-old background; such genetically
engineered strains might have an advantage in all of our competition environments.

5.2.2 Biofilm evolution future work: fluorescent microscopy
While Chapter 3 shows that cells aged in a biofilm outcompete naïve cells, it is unknown
precisely how they do so. Cells aged in a biofilm may be better at attaching to virgin and
biofilm surfaces than their naïve counterparts. Alternatively, aged cells may attach in a
manner similar to naïve cells, but once attached are more prolific. These possibilities are
not mutually exclusive, and still other mechanisms may be responsible for the
competitive advantage seen particularly in 10-day-old and 3x10-day old biofilm cells. To
address this, fluorescently labeled strains would need to be aged in a biofilm, harvested
and competed. For example, 10-day-old cells labeled with cyan fluorescent protein (CFP)
could be competed against naïve cells labeled with yellow fluorescent protein (YFP), and
vice versa (Kraigsley and Finkel, 2009). The competitions could either be monitored in
situ via time-lapse movies, or examined via post-competition autopsy. Likewise, these
experiments can also be used to elucidate the nature of the competitive disadvantage
conferred by aging cells continuously in a biofilm for 20 days.

5.2.3 Biofilm evolution future work: directed evolution
Now that we have insight into the evolution of S. oneidensis MR-1 biofilms on an
electrically inert (glass) surface, future work is needed to understand how this organism
evolves on electrically conductive and poised surfaces. Microbes have been genetically
110

engineered to understand and improve their performance within microbial fuel cells.
Directed evolution offers the ability to “reverse engineer” microbes for improved MFC
performance. That is, spontaneous mutants that provide higher electrical current or
increased Coulombic efficiency can be screened for or selected. These improved
performers can then be probed (via whole genome resequencing, for example) to discover
what mutants confer these improvements. This approach allows for the possible
discovery of genes not previously considered for their involvement in power production.

In terms of power production, the best biofilm competitors (3x10-day-olds) did not
significantly differ from the wildtype when introduced into the MFC (data not shown). In
light of the fact that not all populations evolved in aerobic Minimal Medium enjoyed a
competitive advantage in aerobic LB (Figure 2.9), it is perhaps not surprising that strains
evolved in a glass test tube were not advantaged on the graphite electrode of the MFC.
However, the data presented here indicate that cells aged within the MFC will adapt to
the MFC environment. Furthermore, data presented in Chapter 2 indicate that continuous
aging within the MFC might be adaptive for planktonic cells. Data from Chapter 3
indicate that shorter aging periods, followed by harvesting and reinoculation, may prove a
more successful strategy for adapting electrode-associate biofilms.

5.3 Biofilm formation and amino acid supplementation
On acid-washed glass surface, S. oneidensis MR-1 rapidly forms two visible biofilms:
one at the air-medium interface, where oxygen is relatively abundant, and roughly 1cm
111

below the air-medium interface, where oxygen is relatively scarce (Figure 4.1). Within
relatively short periods of time (between 2 hour time points in our time course), a
significant portion of the cells associated with the biofilms detach, and later reattach.
These biofilms are physiologically different depending on their developmental stage, and
respond to different supplementary amino acids. While none of the 20 amino acids tested
had a universal effect, certain combinations not only affected developing and established
biofilms differently, but also had differential effects on the top and bottom rings (Figure
4.2). Further investigation is needed to determine what role (catabolic, anabolic,
respiratory, signaling, et cetera), these amino acids play. However, given the abundance
of amino acids in many types of wastewater, these effects should be attended to when
using MFCs as part of the wastewater treatment process.

5.3.1 Biofilm formation and amino acid supplementation future work: mass
spectrometry
While the medium is well defined at the time of inoculation, it is unknown how the
medium changes as the biofilm develops. Mass spectrometry allows for the examination
of the supernatant, and quantify the amount of lactate and supplementary amino acids
have been consumed, what products result from consumption or reduction (such as
acetate or pentanoic acid, respectively), what detrital nutrients are being made available
and possibly what signal molecules are being secreted. Furthermore, mass spectrometry
can be applied to the biofilms themselves (via a washing and sonication procedure similar
to that used in Chapter 3) to get a metabolomic and proteomic profile of the cells and the
112

extracellular matrix. For profiling amino acids, fatty acids and alcohols, either gas
chromatography-mass spectrometry (GC-MS) or liquid chromatography-mass
spectrometry (LC-MS) can be applied. However, GC-MS is preferred for analyzing less
polar compounds, such as lipids, while LC-MS is preferred for analyzing more polar
compounds such as nucleotides and organic acids (Agilent Technologies, Inc., 2007).
Techniques can be chosen to address specific compounds of interest, and may elucidate
why, for example, 20 days of continuous aging within a biofilm confers a competitive
disadvantage, regardless of strain background. It may also reveal whether certain amino
acids are hydrogen acceptors.

5.3.2 Biofilm formation and amino acid supplementation future work: radio-
labeled amino acids
It is clear that different amino acids and amino acid pairs have distinct effects on S.
oneidensis MR-1 biofilms depending on the developmental stage, as well as the
physiological conditions of the biofilm. These amino acids could be affecting biofilms
via a number of means, such as providing a carbon and energy source, producing ATP
via the Stickland reaction, integrating into anabolism, and still other possibilities. One
way to demonstrate an organism’s ability to use certain amino acids as hydrogen donors
is to observe the reduction of such chemicals as methylene blue; to demonstrate the use
of amino acids as hydrogen acceptors, the oxidation of reduced phenosafranin, for
example, may be observed (Nisman, 1954). However, this method cannot be directly
applied to our biofilms. Alternatively, potassium cyanide or sodium arsenite may be
113

added to inhibit amino acid dehydrogenases and amino acid reductases, respectively, but
these may have additional effects in our system. To determine if Stickland fermentation is
occurring in our biofilms, media can be supplemented with one of the amino acids
labeled with deuterium. At the end of 27 hours, nuclear magnetic resonance (NMR) can
be performed on cells and media to see if deuterium atoms have moved from the labeled
amino acid to the unlabeled amino acid. As an alternative, supplementary amino acids
can have
13
C-labeled carboxyl groups, to determine whether amino acids have been
converted into carboxylic acids, and thus indicate their role as hydrogen donors (Majors,
et al., 2005; McLean, et al., 2008). The latter approach can also determine whether
supplementary amino acids are incorporated into anabolic processes, and whether these
amino acids are involved catabolism.

5.3.3 Amino acids and biofilms future work: microbial fuel cell
supplementation
The addition of a few nutrients has a profound impact on our biofilms. What remains to
be seen is whether similar supplementation has an effect on current production in MFCs.
The occasional addition of certain amino acids, or even wastewater known to have high
concentrations of certain amino acids, may prove a cost-effective way to increase power.
Furthermore, measuring current against the addition of amino acid pairs may inform the
nature of MFC biofilms. For example, if the addition of alanine plus valine has a positive
effect on current production, and the addition of asparagine plus histidine has a negative
effect, it will indicate that electrode-associated biofilms behave similarly to the top ring
114

biofilms in our test tubes. That is, electrode-associated biofilms behave like biofilms
grown in conditions of relative electron donor abundance. Furthermore, it will indicate
that crystal violet staining in test tubes is a good proxy for current production in MFCs.

5.4 Conclusions
This work presents the first demonstration of real-time evolution of the environmentally
ubiquitous and technologically useful organism, Shewanella oneidensis MR-1.
Populations of free-swimming (planktonic) individuals, as well as populations of surface-
attached (biofilm) communities, display a GASP-like phenotype, though similar aging
regimens yield different phenotypes in the two lifestyles. Furthermore, a significant
portion of statically cultured cells were shown rapidly to shift from one lifestyle (e. g.
planktonic) to the other (e. g. biofilm). Evolution was demonstrated in nutrient-rich and
nutrient-limited conditions, as well as conditions oxygen abundance and oxygen paucity.
Cells that develop biofilms where oxygen is prevalent detach and reattach at different
times than cells that develop biofilms where oxygen is scarce. Like biofilms at different
stages of development, oxygen-unrestricted and oxygen-limited biofilm respond to
different combinations of supplementary amino acids. These results have important
implications for microbial fuel cell-inhabiting S. oneidensis MR-1, both in terms of
physiological adaption and directed evolution.

115

References

Abboud, R.; Popa, R.; Souza-Egipsy, V.; Giometti, C.; Tollaksen, S.; Mosher, J. J.;
Findlay, R. H.; Nealson, K. H. Low-temperature growth of Shewanella oneidensis MR-
1. 2005. Appl Environ Microbiol. 71:811-6

Aelterman, P.; Rabaey, K.; Clauwaert, P.; Verstraete, W. Microbial fuel cells for
wastewater treatment. 2006. Water Sci Techno. l 54 8, 5th World Water Congress:
Wastewater Treatment Processes. 2006: 9–15

Agilent Technologies, Inc. “Considerations for selecting GC/MS or LC/MS for
metabolomics.” 2007. http://www.chem.agilent.com/Library/selectionguide/Public/5989-
6328EN.pdf. Accessed Februrary 7, 2012.

Ahimou, F.; Semmens, M. J.; Haugstad, G.; Novak, P. J. Effect of protein,
polysaccharide, and oxygen concentration profiles on biofilm cohesiveness. 2007. Appl
Environ Microbiol. 73:2905-10

Andrieux, C. P.; Blocman, C.; Dumas-Bouchiat, J. M.; M’Halla, F.; Saveant, J. M.
Homogeneous redox catalysis of electrochemical reactions. Part V. Cyclic voltammetry.
1980. J Electroanal Chem Interfacial Electrochem. 113:19–40

Angenent, L. T.; Karim, K.; Al-Dahhan, M. H.; Wrenn, B. A.; Domiguez-Espinosa,
R. Production of bioenergy and biochemicals from industrial and agricultural wastewater.
2004. Trends Biotechnol. 22:477–85

Bazaka, K.; Crawford, R. J.; Nazarenko, E. L.; Ivanova, E. P. Bacterial extracellular
polysaccharides. 2011. Adv Exp Med Biol. 715:213-26

Beliaev, A. S.; Klingeman, D. M.; Klappenbach, J. A.;Wu,L.; Romine, M. F.; Tiedje,
J. M.; Nealson, K. H.; Fredrickson, J. K.; Zhou, J. Global transcriptome analysis of
Shewanella oneidensis MR-1 exposed to different terminal electron acceptors. 2005. J
Bacteriol. 187:7138–45

Bencharit, S.; Ward, M. J. Chemotactic responses to metals and anaerobic electron
acceptors in Shewanella oneidensis MR-1. 2005. J Bacteriol. 187:5049-53

Bernier, S. P.; Ha, D. G.; Khan, W.; Merritt, J. H.; O’Toole, G. A. Modulation of
Pseudomonas aeruginosa surface-associated group behaviors by individual amino acids
through c-di-GMP signaling. 2011. Res Microbiol. 162:680-8

116

Bhatnagar, D.; Xu, S.; Fischer, C.; Arechederra, R. L.; Minteer, S. D. Mitochondrial
biofuel cells: expanding fuel diversity to amino acids. 2011. Phys Chem Chem Phys.
13:86-92

Biffinger, J. C.; Byrd, J. N.; Dudley, B. L.; Ringeisen, B. R. Oxygen exposure
promotes fuel diversity for Shewanella oneidensis microbial fuel cells. 2008. Biosens
Bioelectron. 23:820-6

Biffinger, J. C.; Fitzgerald, L. A.; Ray, R.; Little, B. J.; Lizewski, S. E.; Peterson, E.
R.; Ringeisen, B. R.; Sanders, W. C.; Sheehan, P. E.; Pietron, J. J.; Baldwin, J. W.;
Nadeau, L. J.; Johnson, G. R.; Ribbens, M.; Finkel, S. E.; Nealson, K. H. The utility
of Shewanella japonica for microbial fuel cells. 2011. Bioresour Technol. 102:290-7

Biffinger, J. C.; Pietron, J.; Bretschger, O.; Nadeau, L. J.; Johnson, G. R.; Williams,
C. C.; Nealson, K. H.; Ringeisen, B. R. The influence of acidity on microbial fuel cells
containing Shewanella oneidensis. 2008. Biosens Bioelectron. 24:906-11

Biffinger, J. C.; Ray, R.; Little, B. J.; Fitzgerald, L. A.; Ribbens, M.; Finkel, S. E.;
Ringeisen, B. R. Simultaneous analysis of physiological and electrical output changes in
an operating microbial fuel cell with Shewanella oneisensis. 2009. Biotechnol. Bioeng.
103:524-31

Biffinger, J. C.; Ribbens, M.; Ringeisen, B. R.; Pietron, J.; Finkel, S.; Nealson, K. H.
Characterization of electrochemically active bacteria (EAB) utilizing a high-throughput
voltage-based screening assay. 2009. Biotechnol Bioeng. 102:436–44

Bond, D. R.; Holmes, D. E.; Tender, L. M.; Lovley, D. R. Electrode-reducing
microorganisms that harvest energy from marine sediments. 2002. Science. 295:483-5

Borole, A.; Hamilton, C.; Vishnivetskaya, T. Enhancement in current density and
energy conversion efficiency of 3-dimensional MFC anodes using pre-enriched
consortium and continuous supply of electron donors. 2011. Bioresour Technol.
102:5098-104

Bouhenni, R. A.; Vora, G. J.; Biffinger, J. C.; Shiodkar, S.; Brockman, K.; Ray, R.;
Wu, P. K.; Johnson, B. J.; Biddle, E. M.; Marshall, M. J.; Fitzgerald, L. A.; Little,
B. J.; Fredrickson, J. K.; Beliaev, A. S.; Ringeisen, B. R.; Saffarini, D. A. The role of
Shewanella oneidensis MR-1 outer membrane structures in extracellular electron transfer.
2010. Electroanalysis. 22:856–64

117

Bretschger, O., A. Obraztsova, C. A. Sturm, I. S. Chang, Y. A. Gorby, S. B. Reed, D.
E. Culley, C. L. Reardon, S. Barua, M. F. Romine, J. Zhou, A. S. Beliaev, R.
Bouhenni, D. Saffarini, F. Mansfeld, B. H. Kim, J. K. Fredrickson, and K. H.
Nealson. Current production and metal oxide reduction by Shewanella oneidensis MR-1
wild type and mutants. 2007. Appl Environ Microbiol. 73:7003-12

Burbaum, J. J.; Sigal, N. H. New technologies for high-throughput screening. 1997.
Curr Opin Chem Bio. 1:72-8

Burnes, B. S.; Mulberry, M. J.; Dichristina, T. J. Design and application of two rapid
screening techniques for isolation of Mn(IV) reduction-deficient mutants of Shewanella
putrefaciens. 1998. Appl Environ Microbiol. 64:2716-20

Call, D. F.; Wagner, R. C.; Logan, B. E. Hydrogen production by Geobacter species
and a mixed consortium in a microbial electrolysis cell. 2009. Appl Environ Microbiol.
75:7579–87

Cabral, M. P.; Soares, N. C.; Aranda, J.; Parreira, J. R.; Rumbo, C.; Poza, M.;
Valle, J.; Calamia, V.; Lasa, I.; Bou, G. Proteomic and functional analyses reveal a
unque lifestyle for Acinetobacter baumannii biofilms and a key role for histidine
metabolism. 2011. J Proteome Res. 10:3399-417

Carmona-Martinez, A. A.; Harnisch, F.; Fitzgerald, L. A.; Biffinger, J. C.;
Ringeisen, B. R.; Schröder, U. Cyclic voltammetric analysis of the electron transfer of
Shewanella oneidensis MR-1 and nanofilament and cytochrome knock-out mutants.
2011. Bioelectrochemistry. 81:74-80

Chang, I. S.; Moon, H.; Bretschger, O.; Jang, J. K.; Park, H. I.; Nealson, K. H.;
Kim, B. H. Electrochemically active bacteria (EAB) and mediator-less microbial fuel
cells. 2006. J Microbiol Biotechnol. 16:163-77

Chaudhuri, S. K.; Lovley, D. R. Electricity generation by direct oxidation of glucose in
mediatorless microbial fuel cells. 2003. Nat Biotechnol. 21:1229–32

Chavan, K.; Finkel, S. E. Effect of population density on genotypic diversity of
evolving E. coli cultures and expression of the GASP phenotype. 2012. In submission.

Chen, L.; Wen, Y. M. The role of bacterial biofilm in persistent infections and control
strategies. 2011. Int J Oral Sci. 3:66-73

Clauwaert, P.; van der Ha, D.; Verstraete, W. Energy recovery from energy rich
vegetable products with microbial fuel cells. 2008. Biotechnol Lett. 30:1947-51

118

Cooper, K. R.; Smith, M. Electrical test methods for on-line fuel cell ohmic resistance
measurements. 2006. J Power Sources. 160:1088-95

de la Rosa, C.; Yu, T. Three-dimensional mapping of oxygen distribution in wastewater
biofilms using an automation system and microelectrodes. 2005. Environ Sci Technol.
39:5196-202

Dressaire, C.; Redon, E.; Gitton, C.; Loubiere, P.; Monnet, V.; Cocaign-Bousquet,
M. Investigation of the adaptation of Lactococcus lactis to isoleucine starvation
integrating dynamic transcriptome and proteome information. 2011. Microb Cell Fact.
Suppl 1:S18

Deutscherbauer, A.; Price, M. N.; Wetmore, K. M.; Shao, W.; Baumohl, J. K.; Xu,
Z.; Nguyen, M.; Tamse, R. Davis, R. W.; Arkin, A. P. Evidence-based annotation of
gene function in Shewanella oneidensis MR-1 using genome-wide fitness profiling across
121 conditions. 2011. PLoS Genet. 7:e1002385

Diaz-Mochon, J. J.; Tourniaire, G.; Bradley, M. Microarray platforms for enzymatic
and cell-based assays. 2007. Chem Soc Rev. 36:449-57

DiChristina T. J.; Arnold R. G.; Lidstrom M. E.; Hoffmann M. R. Dissimilative iron
reduction by the marine eubacterium Alteromonas putrefaciens strain 200. 1988. Water
Sci Technol. 20:69–79

Driscoll, M. E.; Romine, M. F.; Juhn, F. S.; Serres, M. H.; McCue, L. A.; Beliaev, A.
S.; Fredrickson, J. K.; Gardner, T. S. Identification of diverse carbon utilization
pathways in Shewanella oneidensis MR-1 via expression profiling. 2007. Genome Inform
Ser. 18:287–98

El-Naggar, M.; Gorby, Y.; Xia, W.; Nealson, K. The molecular density of states in
bacterial nanowires. 2008. Biophys. J. 95:L10-2

Elena, S.; Lenksi, R. Evolution experiments with microorganisms: the dynamics and
genetic bases of adaptation. 2003. Nat Rev Genet. 4:457-69

Erable, B.; Etcheverry, L.; Bergel, A. From microbial fuel cell (MFC) to microbial
electrochemical snorkel (MES): maximizing chemical oxygen demand removal from
wastewater. 2011. Biofouling. 27:319-26

Fang, R.; Elias, D. A.; Monroe, M. E.; Shen, Y.; McIntosh, M.; Wang, P.; Goddard,
C. D.; Callister, S. J.; Moore, R. J.; Gorby, Y. A.; et al. Differential label-free
quantitative proteomic analysis of Shewanella oneidensis cultured under aerobic and
suboxic conditions by accurate mass and time tag approach. 2006. Mol Cell Proteomics.
5:714–725
119

Fangzhou, D.; Zhenglong, L.; Shaoqiang, Y.; Beizhen, X.; Hong, L. Electricity
generation directly using human feces wastewater for life support system. 2011. Acta
Astronautica. 68:1537-47

Farrell, M. J.; Finkel, S. E. The growth advantage in stationary-phase phenotype
conferred by rpoS mutations is dependent on the pH and nutrient environment. 2003. J
Bacteriol. 185:7044-52

Franks, A. E.; Nevin, K. P.; Jia, H.; Izallalen, M.; Woodard, T. L.; Lovley, D. R.
Novel strategy for three-dimensional real-time imaging of microbial fuel cell
communities: monitoring the inhibitory effects of proton accumulation within the anode
biofilm. 2009. Energy Environ Sci. 2:113–9

Fredrickson, J. K.; Zachara, J. M. Electron transfer at the microbe-mineral interface: a
grad challenge in biogeochemistry. Geobiology. 6:245-53

Fredrickson, J. K.; Romine, M. F.; Beliaev, A. S.; Auchtung, J. M.; Driscoll, M. E.;
Gardner, T. S.; Nealson, K. H.; Osterman, A. L.; Pinchuk, G.; Reed, J. L.;
Rodionov, D. A.; Rodrigues, J. L. M.; Saffarini, D. A.; Serres, M. H.; Spormann, A.
M.; Zhulin, I. B.; Tiedje, J. M. Towards environmental systems biology of Shewanella.
2008. Nat Rev Microbiol. 6:592-603

Finkel, S.; Kolter, R. Evolution of microbial diversity during prolonged starvation.
1999. Proc. Natl. Acad. Sci. U. S. A. 96:4023-7

Finkel, S. Long-term survival during stationary phase: evolution and the GASP
phenotype. 2006. Nat Rev Microbiol. 4:113-20

Ganesh, R.; Robinson, K. G.; Reed, G. D.; Sayler, G. S. Reduction of hexavalent
uranium from organic complexes by sulfate- and iron-reducing bacteria. 1997. Appl
Environ Microbiol. 63:4385–91

Gay, F. P. Bacteria growth and reproduction. Agents of Disease and Host Resistance,
Including Principles of Immunology, Bacteriology, Mycology, Protozoology,
Parasitology, and Virus Disease. 1935. Charles C. Thomas, Springfield. 1–38

Geng, M.; Duan, Zhenhao. Prediction of oxygen solubility in pure water and brines up
to high temperatures and pressures. 2010. Geochim Cosmochim Ac. 74:5631-40

Gong, Y.; Radachowsky, S. E.; Wolf, M.; Nielsen, M. E.; Girquis, P. R. Reimers, C.
E. Benthic microbial fuel cell as direct power source for an acoustic modem and seawater
oxygen/temperature sensor system. 2011. Environ Sci Technol. 45:5047-53
120

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.; Fredrickson, J. K.
Electrically conductive bacterial nanowires produced by Shewanella oneidensis strain
MR-1 and other microorganisms. 2006. Proc Natl Acad Sci USA. 103:11358-63

Hall-Stoodley, L.; Costerton, J. W.; Stoodley, P. Bacterial biofilms: from the natural
environment to infectious diseases. 2004. Nat Rev Microbiol. 2:95-108

Hamilton, S.; Bongaerts, R. J.; Mulholland, F.; Cochrane, B.; Porter, J.; Lucchini,
S.; Lappin-Scott, H. M.; Hinton, J. C. The transcriptional programme of Salmonella
enterica serovar Typhimurium reveals a key role for tryptophan metabolism in biofilms.
2009. BMC Genomics. 10:599

Harms, P.; Kostov, Y.; French, J. A.; Soliman, M.; Anjanappa, M.; Ram, A.; Rao,
G. Design and performance of a 24-station high-throughput microbioreactor. 2005.
Biotechnol Bioeng, 93:6–13

Harnisch, F.; Schroder, U. Selectivity versus mobility: separation of anode and cathode
in microbial bioelectrochemical systems. 2009. Chem Sus Chem. 2:921–6

Harnisch, F.; Schroöder, U.; Scholz, F. The suitability of monopolar and bipolar ion
exchange membranes as separators for biological fuel cells. 2008. Environ Sci Technol.
42:1740–6.

Harris, H. W.; El-Naggar, M. Y.; Bretschger, O.; Ward, M. J.; Romine, M. F.;
Obraztsova, A. Y.; Nealson, K. H. Electrokinesis is a microbial behavior that requires
extracellular electron transport. 2010. Proc Natl Acad Sci USA. 107:326-31

Hau, H. H.; Gralnick, J. A. Ecology and biotechnology of the genus Shewanella. 2007.
Annu Rev Microbiol. 61:237-58

Helmus, R. A.; Liermann, L. J.; Brantley, S. L.; Tien, M. Growth advantage in
stationary-phase (GASP) phenotype in long-term survival strains of Geobacter
sulfurreducens. 2012. FEMS Microbiol Ecol. 79:218-28

Hernandez, M. E.; Newman, D. K. Extracellular electron transfer. 2001. Cell Mol Life
Sci. 58:1562–71

Huang, J.; Sun, B.; Zhang, X. Electricity generation at high ionic strength in microbial
fuel cell by a newly isolated Shewanella marisflavi EP1. 2010. Appl Microbiol
Biotechnol. 85:1141–9
121

Hwang, J. S. Investigation of general metabolism in the metal reducer, Shewanella
oneidensis MR-1. 2002. Ph. D. Thesis, University of California, Berkeley. 62-74

Inglese, J.; Johnson, R. L.; Simeonov, A.; Xia, M.; Zheng, W.; Austin, C. P.; Auld,
D. S. High-throughput screening assays for the identification of chemical probes. 2007.
Nat Chem Biol. 3:466–79

Ishii, S. I.; Shimoyama, T.; Hotta, Y.; Watanabe, K. Characterization of a filamentous
biofilm community established in a cellulose-fed microbial fuel cell. 2008. BMC
Microbiol. 8:6–18

Ivanova, E. P.; Sawabe, T.; Gorshkova, N. M.; Svetashev, V. I.; Mikhailov, V. V.;
Nicolau, D. V.; Christen, R. Shewanella japonica sp. nov. 2001. Int J Syst Evol
Microbiol. 51:1027–33

Jelsbak, L.; Sogaard-Andersen, L. Cell behavior and cell–cell communication during
fruiting body morphogenesis in Myxococcus xanthus. 2003. J Microbiol Methods.
55:829–39

Kahn, L.; Wayman, C. Amino acids in raw sewage and sewage effluents. 1964. J Water
Pollut Con F. 36:1368-71

Karatan, E.; Watnick, P. Signals, regulatory networks, and materials that build and
break bacterial biofilms. 2009. Microbiol Molec Biol Rev. 73:310-47

Karunakaran, E.; Mukherjee, J.; Ramalingam, B.; Biggs, C. A. “Biofilmology”: a
multidisciplinary review of the study of microbial biofilms. 2011. Appl Microbiol
Biotechnol. 90:1869-81

Ki, D.; Park, J.; Lee, J.; Yoo, K. Microbial diversity and population dynamics of
activated sludge microbial communities participating in electricity generation in
microbial fuel cells. 2008. Water Sci Technol. 58:2195-201

Kim, B. H.; Kim, H. J.; Hyun, M. S.; Park, D. H. Direct electrode reaction of Fe(III)-
reducing bacterium, Shewanella putrefaciens. 1999. J Microbiol Biotechn. 9:127-31

Kim, B. H.; Chang, I. S.; Gadd, G. M. Challenges in microbial fuel cell development
and operation. 2007. Appl Microbiol Biot. 76:485-94

Kim, H. J.; Hyun, M. S.; Chang, I. S.; Kim, B. H. A microbial fuel cell type lactate
biosensor using a metal-reducing bacterium, Shewanella putrefaciens. 1999. J Microbiol
Biotechnol. 9:365–7

122

Kim, H. J.; Park, H. S.; Moon, S. H.; Chang, I. S.; Kim, M.; Kim, B. H. A mediator-
less microbial fuel cell using a metal reducing bacterium, Shewanella putrefaciens. 2002.
Enzyme Microb Tech. 30:145-52

Kim, W. S.; Park, J. H.; Ren, J.; Su, P.; Dunn, N. W. Survival response and
rearrangement of plasmid DNA of Lactococcus lactis during long-term starvation. 2001.
Appl Environ Microbiol. 67:4594-602

Kolter, R. Biofilms in lab and nature: a molecular geneticist’s voyage to microbial
ecology. 2010. Int Microbiol. 13:1-7

Koskiniemi, S.; Hughes, D.; Andersson, D. I. Effect of translesion DNA polymerases,
endonucleases and RpoS on mutation rates in Salmonella typhimurium. 2010. Genetics.
185:783-95

Kraigsley, A.; Finkel, S. Adaptive evolution in single species bacterial biofilms. 2009.
FEMS Microbiol Lett. 293:135-40

Kuznetsov, B. A.; Davydova, M. E.; Shleeva, M. O.; Shleev, S. V.; Kaprelyants, A.
S.; Yaropolov, A. I. Electrochemical investigation of the dynamics of Mycobacterium
smegmatis cells transformation to dormant, nonculturable form. 2004.
Bioelectrochemistry. 64:125–131

Kuznetsov, B. A.; Khlupova, M. T.; Shleev, S. V.; Kaprel’yants, AS, Yaropolov AI.
An electrochemical method for measuring metabolic activity and counting cells. 2006.
Appl Biochem Microbiol. 42:525–33

Lanthier, M.; Gregory, K. B.; Lovley, D. R. Growth with high planktonic biomass in
Shewanella oneidensis fuel cells. 2008. FEMS Microbiol Lett. 278:29-35

Larminie, J.; Dicks, A. Fuel Cell Systems Explained. 2003. John Wiley & Sons
Ltd.,West Sussex.

Leaphart, A. B.; Thompson, D. K.; Huang, K.; Alm, E. Wan, X. F.; Arkin, A.;
Brown, S. D.; Wu, L.; Yan, T.; Liu, X.; et al. Transcriptome profiling of Shewanella
oneidensis gene expression following exposure to acidic and alkaline pH. 2006. J
Bacteriol. 188:1633–1642

Lefebvre, O.; Uzabiaga, A.; Chang, I.; Kim, B.; Ng, H. Microbial fuel cells for energy
self-sufficient domestic wastewater treatment - review and discussion from energetic
consideration. 2010. Appl Microbiol Biotechnol. 89:259-70

123

Li, Y.; Champion, M. M.; Sun, L.; Champion, P. A.; Wojcik, R.; Dovichi, N. J.
Capillary zone electrophoresis-electrospray ionization-tandem mass spectrometry as an
alternative proteomics platform to ultraperformance liquid chromatography-electrospray
ionization-tandem mass spectrometry for samples of intermediate complexity. 2012. Anal
Chem. 84:1617-22

Lies, D. P.; Hernandez, M. E.; Kappler, A.; Mielke, R. E.; Gralnick, J. A.; Newman,
D. K. Shewanella oneidensis MR-1 uses overlapping pathways for iron reduction at a
distance and by direct contact under conditions relevant for biofilms. 2005. Appl Environ
Microbiol. 71:4414–26

Logan, B. E. Exoelectrogenic bacteria that power microbial fuel cells. 2009. Nat Rev
Microbiol. 7:375-81

Logan, B. E.; Hamelers, B.; Rozenda,l R.; Schröder, U.; Keller, J.; Freguia, S.;
Aelterman, P.; Verstraete, W.; Rabaey, K. Microbial fuel cells: Methodology and
technology. 2006. Environ Sci Technol. 40:5181–92

Logan, B. E.; Regan, J. M. Microbial fuel cells—challenges and applications. 2006.
Environ Sci Technol. 40:5172–80

Lovley, D.R. Bug juice: harvesting electricity with microorganisms. 2006. Nat Rev
Microbiol. 4:497–508

Lovley, D. R. Extracellular electron transfer: Wires, capacitors, iron lungs, and more.
2008. Geobiology. 6:225–31

Lovley. D. R.; Coates, J. D. Bioremediation of metal contamination. 1997. Curr Opin
Biotechnol. 8:285–9

Lovley, D. R.; Coates, J. D.; Saffarini, D. A.; Lonergan, D. J. “Dissimilatory iron
reduction.” In: Winklemann G, Carrano CJ, editors. 1997. Amsterdam, The Netherlands:
Harwood. Transition metals in microbial metabolism. 187–215

Lovley, D. R.; Phillips, E. J. P. Rapid assay for microbially reducible ferric iron in
aquatic sediments. 1987. Appl Environ Microbiol. 53:1536–40

Lovley, D. R.; Phillips, E. J. P. Novel mode of microbial energy metabolism: organic
carbon oxidation coupled to dissimilatory reduction of iron or manganese. 1988. Appl
Environ Microbiol. 54:1472–80

Madan, R.; Kolter, R.; Mahadevan, S. Mutations that activate the silent bgl operon of
Escherichia coli confer a growth advantage in stationary phase. 2005. J Bacteriol.
187:7912-7
124

Maharbiz, M. M.; Holtz, W. J.; Howe, R. T.; Keasling, J. D. Microbioreactor arrays
with parametric control for high-throughput experimentation. 2004. Biotechnol Bioeng.
85:376–81

Maharjan, R.; Seeto, S.; Notley-McRobb, L.; Ferenci, T. Clonal adaptive radiation in
a constant environment. 2006. Science. 313:514-7

Majors, P. D.; Mclean, J. S.; Pinchuk, G. E.; Fredrickson, J. K.; Gorby, Y. A.;
Minard, K. R.; Wind, R. A. NMR methods for in situ biofilm metabolism studies. 2005.
J Microbial Methods. 62:337-44

Manohar, A. K.; Bretschger, O.; Nealson, K. H.; Mansfeld, F. The use of
electrochemical impedance spectroscopy (EIS) in the evaluation of the electrochemical
properties of a microbial fuel cell. 2008. Bioelectrochemistry. 72:149–54

Marsili, E.; Baron, D. B.; Shikhare, I. D.; Coursolle, D.; Gralnick, J. A.; Bond, D. R.
Shewanella secretes flavins that mediate extracellular electron transfer. 2008. Proc Natl
Acad Sci USA. 105:3968–73

Martínez-García, E.; Tormo, A.; Navarro-Lloréns, J. M. GASP phenotype: presence
in enterobacteria and independence of sigmaS in its acquisition. 2003. FEMS Microbiol
Lett. 225:201-6

Marzan, L. W.; Shimizu, K. Metabolic regulation of Escherichia coli and its phoB and
phoR genes knockout mutants under phosphate and nitrogen limitations as well as at
acidic condition. 2011. Microb Cell Fact. 10:39

Mclean, J. S.; Majors, P. D.; Bilskis, C. L.; Reed, S. B.; Romine, M. F.; Fredrickson,
J. K. Investigations of the structure and metabolism within Shewanella oneidensis MR-1
biofilms. 2008. J Microbial Methods. 74:47-56

Mitcheson, P. D. Energy harvesting for human wearable and implantable bio-sensors.
2010. Conf Proc IEEE Eng Med Biol Soc. 2010:3432-6

Monroe, D. Looking for chinks in the armor of bacterial biofilms. 2007. PLoS Biol.
5:e307

Mosier, A. P.; Cady, N. C. Analysis of bacterial surface interactions using microfluidics
systems. 2011. Sci Prog. 94:431-5

Myers, C. R.; Nealson, K. H. Bacterial manganese reduction and growth with
manganese oxice as the sole electron acceptor. 1988. Science. 240:1319-21

125

Myers, C. R.; Nealson, K. H. Respiration-linked proton translocation coupled to
anaerobic reduction of manganese(IV) and iron(III) in Shewanella putrefaciens MR-1.
1990. J Bacteriol. 172:6232-8

Nealson, K. H.; Belz, A.; McKee, B. Breathing metals as a way of life: Geobiology in
action. 2002. Antonie van Leeuwenhoek. 81:215–22

Nealson, K. H.; Saffarini, D. Iron and manganese in anaerobic respiration:
environmental significance, physiology, and regulation. 1994. Annu Rev Microbiol.
48:311-43

Nealson, K. H.; Finkel, S. E. Electron flow and biofilms. 2011. MRS Bulletin. 36: 380-4.
Pant, D.; Van Bogaert, G.; Diels, L.; Vanbroekhoven, K. A review of the substrates
used in microbial fuel cells (MFCs) for sustainable energy production. 2010. Bioresour
Technol. 101:1533-43

Nisman, B. The Stickland reaction. 1954. Bacteriol Rev. 18:16-42

Okamoto, A.; Hashimoto, K.; Nakamura, R. Long-range electron conduction of
Shewanella biofilms mediated by outer membrane C-type cytochromes. 2011.
Bioelectrochemistry. 85:61-5

O’Toole, G.; Kaplan, H. B.; Kolter, R. Biofilm formation as microbial development.
2000. Annu Rev Microbiol. 54:49-79

O’Toole, G. A.; Kolter, R. Initiation of biofilm formation in Pseudomonas fluorescens
WCS365 proceeds via multiple, convergent signaling pathways: a genetic analysis. 1998.
Mol Microbiol. 28:449-61

O’Toole, G. A.; Kolter, R. Flagellar and twitching motility are necessary for
Pseudomonas aeruginosa biofilm development. 1998. Mol Microbiol. 30:295-304

Park, D. H.; Zeikus, J. G. Electricity generation in microbial fuel cells using neutral red
as an electronophore. 2000. Appl Environ Microbiol. 66:1292-7

Payne, A. N.; DiChristina, T. J. A rapid mutant screening technique for detection of
technetium [Tc(VII)] reduction-deficient mutants of Shewanella oneidensis MR-1. 2006.
FEMS Microbiol Lett. 259:282– 287

Plymale, A. E.; Bailey, V. L.; Fredrickson, J. K.; Heald, S. M.; Buck, E. C.; Shi, L.;
Wang, Z.; Resch, C. T.; Moore, D. A.; Bolton, H. Biotic and abiotic reduction and
solubilization of Pu(IV)O(2)∙xH(2)O((am)) as affected by anthraquinone-2,6-disulfonate
(AQDS) and ethylenediaminetetraacetate (EDTA). 2012. Environ Sci Technol. 46:2132-
40
126

Polcyn, D. S.; Shain, I. Theory of stationary electrode polarography for a multistep
charge transfer with catalytic (cyclic) regeneration of the reactant. 1966. Anal Chem.
38:376–382

Ponciano, J.; La, J.; Joyce, P.; Forney, L. Evolution of diversity in spatially structured
Escherichia coli populations. 2009. Appl Environ Microbiol. 75:6047-54

Pradhan, S.; Baidya, A. K.; Ghosh, A.; Paul, K.; Chowdhury, R. The El Tor biotype
of Vibrio cholerae exhibits a growth advantage in the stationary phase in mixed cultures
with the classical biotype. 2010. J Bacteriol. 192:955-63

Proft, T.; Baker, E. N. Pili in Gram-negative and Gram-positive bacteria – structure,
assembly and their role in disease. 2009. Life Sci Cell Mol. 66:613–635

Rabaey, K.; Boon, N.; Höfte, M.; Verstraete, W. Microbial phenazine production
enhances electron transfer in biofuel cells. 2005. Environ Sci Technol. 39:3401–8

Rabaey, K.; Rodriguez, J.; Blackall, L. L.; Keller, J.; Gross, P.; Batstone, D.;
Verstraete, W.; Nealson, K. H. Microbial ecology meets electrochemistry: electricity-
driven and driving communities. 2007. ISME J. 1:9-18

Rabaey, K.; Rozendal, R. Microbial electrosynthesis - revisiting the electrical route for
microbial production. 2010. Nat. Rev. Microbiol. 8:706-16

Ramirez-Santos, J.; Contreras-Ferrat, G.; Gomez-Eichelmann, M. Stationary phase
in Escherichia coli. 2005. Rev Latinoam Microbiol. 47:92-101

Ramasamy, R. P.; Gadhamshetty, V.; Nadeau, L. J.; Johnson, G. R. Impedance
spectroscopy as a tool for non-intrusive detection of extracellular mediators in microbial
fuel cells. 2009. Biotechnol Bioeng. 104:882–91

Ramasamy, R. P.; Ren, Z.; Mench, M. M.; Regan, J. M. Impact of initial biofilm
growth on the anode impedance of microbial fuel cells. 2008. Biotechnol Bioeng.
101:101–108.

Ramsay, I. R. Modelling and control of high-rate anaerobic wastewater treatment
systems. 1997. Ph. D. Thesis, Department of Chemical Engineering, the University of
Queensland, Brisbane, Australia.

Rani, S. A.; Pitts, B. ; Beyenal, H.; Veluchamy, R. A.; Lewandowski, Z.; Davidson,
W. M.; Buckingham-Meyer, K.; Stewart, P. S. Spatial patterns of DNA replication,
protein synthesis, and oxygen concentration within bacterial biofilms reveal diverse
physiological states. 2007. J Bacteriol. 189:4223-33
127

Rasmussen, K.; Lewandowski, Z. Microelectrode measurements of local mass transport
rates in heterogeneous biofilms. 1998. Biotechnol Bioeng. 59:302–9

Ray, R.; Little, B.; Wagner, P.; Hart, K. Environmental scanning electron microscopy
investigations of biodeterioration. 1997. Scanning. 19:98–103

Reguera, G.; Nevin, K. P.; Nicoll, J. S.; Covalla, S. F.; Woodard, T. L.; Lovley, D. R.
Biofilm and nanowire production leads to increased current in Geobacter sulferreducens
fuel cells. 2006. Appl Environ Microbiol. 72:7345-8

Ren, Z.; Steinberg, L. M.; Regan, J. M. Electricity production and microbial biofilm
characterization in cellulose-fed microbial fuel cells. 2008. Water Sci Technol. 58:617–22

Rezaei, F.; Xing, D.; Wagner, R.; Regan, J. M.; Richard, T. L.; Logan, B. E.
Simultaneous cellulose degradation and electricity production by Enterobacter cloacae in
a microbial fuel cell. 2009. Appl Environ Microbiol. 75:3673–8

Richter, H.; Lanthier, M.; Nevin, K. P.; Lovley, D. R. Lack of electricity production
by Pelobacter carbinolicus indicates that the capacity for Fe(III) oxide reduction does not
necessarily confer electron transfer ability to fuel cell anodes. 2007. Appl Environ
Microbiol. 73:5347–53

Ringeisen, B. R.; Henderson, E.; Wu, P. K.; Pietron, J.; Ray, R.; Little, B.;
Biffinger, J. C.; Jones-Meehan, J. M. High power density from a miniature microbial
fuel cell using Shewanella oneidensis DSP10. 2006. Environ Sci Technol. 40:2629–34

Ringeisen, B. R.; Ray, R.; Little, B. A miniature microbial fuel cell operating with an
aerobic anode chamber. 2007. J Power Sources. 165:591–7

Ringø, E.; Stenberg, E.; Strøm, A. R. Amino acid and lactate catabolism in
trimethylamino oxide respiration of Alteromonas putrefaciens NCMB 1735. 1984. Appl
Environ Microbiol. 47:1084-9

Roostalu, J.; Jõers, A.; Luidalepp, H.; Kaldalu, N.; Tension, T. Cell division in
Escherichia coli cultures monitored at single cell resolution. 2008. BMC Microbiol. 8:68

Rozendal, R. A.; Hamelers, H. V. M.; Buisman, C. J. N. Effects of membrane cation
transport on pH and microbial fuel cell performance. 2006. Environ Sci Technol.
40:5206–11

Salas, E. C.; Berelson, W. M.; Hammond, D. E.; Kampf, A. R.; Nealson, K. H. The
impact of bacterial strain on the products of dissimilatory iron reduction. 2010. Geochim
Cosmochim Acta. 74:574-83
128

Saville, R. M.; Dieckmann, N.; Spormann, A. M. Spatiotemporal activity of the mshA
gene system in Shewanella oneidensis MR-1 biofilms. 2010. FEMS Microbiol Lett.
308:76-83

Schröder, U. Anodic electron transfer mechanisms in microbial fuel cells and their
energy efficiency. 2007. Phys Chem Chem Phys. 9:2619–29

Scott, J. H.; Nealson, K. H. A biochemical study of the intermediary carbon metabolism
of Shewanella putrefaciens. 1994. J Bacteriol. 176:3408–11

Schultz, M. P.; Bendick, J. A.; Holm, E. R.; Hertel, W. M. Economic impact of
biofouling on a naval surface ship. 2011. Biofouling. 27:87-98

Serres, M. H.; Riley, M. Genomic analysis of carbon source metabolism of Shewanella
oneidensis MR-1: Predictions versus experiments. 2006. J Bacteriol. 188:4601–9

Sewell, D.; Laird, K.; Phillips, C. Growth advantage in stationary phase phenomenon in
Gram-positive bacteria. 2011. J Bacteriol. 193:1878-83

Sezonov, G.; Joseleau-Petit, D.; D’Ari, R. Escherichia coli physiology in Luria-Bertani
broth. J Bacteriol. 189:8746-9

Shantaram, A.; Beyenal, H.; Veluchamy, R. R. A.; Lewandowski, Z. Wireless sensors
powered by microbial fuel cells. 2005. Environ Sci Technol. 39:5037–42

Smotkin, E. S.; Jiang, J.; Nayar, A.; Liu, R. High-throughput screening of fuel cell
electrocatalysts. 2006. Appl Surf Sci. 252:2573–9

Stams, A. J. M.; de Bok, F. A. M.; Plugge, C. M.; van Eekert, M. H. A.; Dolfing, J.;
Schraa, G. Exocellular electron transfer in anaerobic microbial communities. 2006.
Environ Microbiol. 8:371–82

Stickland, L. H. Studies in the metabolism of the strict anaerobes (genus Clostridum):
the chemical reactions by which Cl. sporogenes obtains its energy. 1934. Biochem J.
28:1746-59

Stoodley, P.; Sauer, K.; Davies, D. G.; Costerton, J. W. Biofilms as complex
differentiated communities. 2002. Annu Rev Microbiol. 56:187-209

Storz, G.; Hengge-Aronis, R. Bacterial stress responses. 2000. Washington, DC: ASM
Press.

129

Sundberg, S. A. High-throughput and ultra-high-throughput screening: solution- and
cell-based approaches. 2000. Curr Opin Biotechnol. 11:47–53

Tang, X.; Du, Z.; Li, H. Anodic electron shuttle mechanism based on 1-hydroxy-4-
aminoanthraquinone in microbial fuel cells. 2010. Electrochem Comm. 12:1140-3

Tang, Y. J.; Laidlaw, D.; Gani, K.; Keasling, J. D. Evaluation of the effects of various
culture conditions on Cr(VI) reduction by Shewanella oneidensis MR-1 in a novel high-
throughput mini-bioreactor. 2006. Biotechnol Bioeng. 95:176–84

Taratus, E. M.; Eubanks, S. G.; DiChristina, T. J. Design and application of a rapid
screening technique for isolation of selenite reduction deficient mutants of Shewanella
putrefaciens. 2000. Microbiol Res. 155:79–85

Teal, T. K.; Lies, D. P.; Wold, B. J.; Newman, D. K. Spatiometabolic stratification of
Shewanella oneidensis biofilms. 2006. Appl Environ Microbiol. 72:7324–30

Tender, L. M.; Gray, S. A.; Groveman, E.; Lowy, D. A.; Kauffman, P.; Melhado, J.;
Tyce, R. C.; Flynn, D.; Petrecca, R.; Dobarro, J. The first demonstration of a
microbial fuel cell as a viable power supply: Powering a meteorological buoy. 2008. J
Power Sources. 179:571–5

Thormann, K. M.; Duttler, S.; Saville, R. M.; Hyodo, M.; Shukla, S.; Hayakawa, Y.;
Spormann, A. M. Control of formation and cellular detachment from Shewanella
oneidensis MR-1 biofilms by cyclic di-GMP. 2006. J Bacteriol. 188:2681-91

Thormann, K. M.; Saville, R. M.; Shukla, S.; Pelletier, D. A.; Spormann, A. M.
Initial phases of biofilm formation in Shewanella oneidensis MR-1. J. 2004. Bacteriol.
186:8096–104.

Trevors, J. T. Hypothesized origin of microbial life in a prebiotic gel and the transition
to a living biofilm and microbial mats. 2011. C R Biol. 334:269-72

Torres, C. I.; Krajmalnik-Brown, R.; Parameswaran, P.; Marcus, A. K.; Wanger,
G.; Gorby, Y. A.; Rittmann, B. E. Selecting anode-respiring bacteria based on anode
potentional: phylogenetic, electrochemical, and microscopic characterization. 2009.
Environ Sci Technol. 43:9519-24

Torres, C. I.; Marcus, A. K.; Lee, H.; Parameswaran, P.; Krajmalnik-Brown, R.;
Rittmann, B. E. A kinetic perspective on extracellular electron transfer by anode-
respiring bacteria. 2010. FEMS Microbiol Rev. 34:3-17

130

Turick, C. E.; Beliaev, A. S.; Zakrajsek, B. A.; Reardon, C. L.; Lowy, D. A.; Poppy,
T. E.; Maloney, A.; Ekechukwu, A. A. The role of 4-hydroxyphenylpyruvate
dioxygenase in enhancement of solid-phase electron transfer by Shewanella oneidensis
MR-1. 2009. FEMS Microbiol Ecol. 68:223

Velasquez-Orta, S. B.; Head, I. M.; Curtis, T. P. Scott, K.; Lloyd, J. R.; von
Canstein, H. The effect of flavin electron shuttles in microbial fuel cells current
production. 2010. Appl Microbiol Biotechnol. 85:1373-81

Venkata Mohan, S.; Veer Raghavulu, S.; Sarma, P. N. Influence of anodic biofilm
growth on bioelectricity production in single chambered mediatorless microbial fuel cell
using mixed anaerobic consortia. 2008. Biosens Bioelectron. 24:41–47

von Canstein, H.; Ogawa, J.; Shimizu, S.; Lloyd, J. R. Secretion of flavins by
Shewanella species and their role in extracellular electron transfer. 2008. Appl Environ
Microbiol. 74:615–23

Wan, X. F.; VerBerkmoes, N. C. McCue, L. A. Stanek, D. Connelly, H.; Hauser, L.
J. Wu, L.; Liu, X.; Yan, T.; Leaphart, A.; et al. Transcriptomic and proteomic
characterization of the Fur modulon in the metal-reducing bacterium Shewanella
oneidensis. 2004. J Bacteriol. 186:8385– 400

Wang, L.; Chen, Y.; Ye, Y.; Lu, B.; Zhu, S.; Shen, S. Evaluation of low-cost cathode
catalysts for high yield biohydrogen production in microbial electrolysis cell. 2011.
Water Sci Technol. 63:440-8.

Watanabe, K. Recent developments in microbial fuel cell technologies for sustainable
bioenergy. 2008. J Biosci Bioeng. 106:528-36

Waters, M. S.; Salas, E. C.; Goodman, S. D.; Udwadia, F. E.; Nealson, K. H. Early
detection of oxidized surfaces using Shewanella oneidensis MR-1 as a tool. 2009.
Biofouling. 25:163-72

Yi, H.; Nevin, K.; Kim, B.; Franks, A.; Klimes, A.; Tender, L.; Lovely, D. Selection
of a variant of Geobacter sulfurreducens with enhanced capacity for current production
in microbial fuel cells. 2009. Biosens Bioelectron. 12:3498-503

Yikrazuul. “Stickland fermentation.,” Wikipedia, The Free Encyclopedia. 2012.
http://en.wikipedia.org/wiki/Stickland_fermentation (accessed February 7, 2012)

Yim, S. Y.; Kang, Y. S.; Cha, D. R.; Park, D. W.; Youn, Y. K.; Jo, Y. M.; song, J. Y.
Sohn, J. W.; Cheong, H. J.; Kim, W. J.; Kim, M. J.; Choi, W. S. Fatal PD periotonitis,
necrotizing fasciitis and bacteremia due to Shewanella putrefacients. 2010. Perit Dial Int.
30:667-9
131

Yong, Y. C.; Yu, Y. Y.; Li, C. M.; Zhong, J. J.; Song, H. Bioelectricity enhancement
bia overexpression of quorum sensing system in Pseudomonas aeruginosa-inoculated
microbial fuel cells. 2011. Biosens Bioelectron. 30:87-92

Yu, T.; de la Rosa, C.; Lu, R. Microsensor measurement of oxygen concentration in
biofilms: from one dimension to three dimensions. 2004. Water Sci Technol. 49:353-8

Zambrano, M. M.; Siegele, D. A.; Almirón, M.; Tormo, A.; Kolter, R. Microbial
competition: Escherichia coli mutants that take over stationary phase cultures. 1993.
Science. 259:1757-60

Zambrano, M. M.; Kolter, R. GASPing for life in stationary phase. 1996. Cell. 86:181-
4

Zhao, F.; Harnisch, F.; Schröder, U.; Scholz, F.; Bogdanoff, P.; Herrmann, I.
Challenges and constraints of using oxygen cathodes in microbial fuel cells. 2006.
Environ Sci Technol. 40:5193–9

Zhao, F.; Slade, R. C. T.; Varcoe, J. R. Techniques for the study and development of
microbial fuel cells: an electrochemical perspective. 2009. Chem Soc Rev. 38:1926–39

Zhang, T.; Cui, C.; Chen, S.; Ai, X.; Yang, H.; Shen, Y.; Peng, Z. A novel
mediatorless microbial fuel cell based on direct biocatalysis of Escherichia coli. 2006.
Chem Commun. 21:2257-9

Zinser, E. R.; Kolter, R. Mutations enhancing amino acid catabolism confer a growth
advantage in stationary phase. 1999. J Bacteriol. 181:5800-7

Zinser, E. R.; Kolter, R. Prolonged stationary-phase incubation selects for lrp mutations
in Escherichia coli K-12. 2000. J Bacteriol. 182:4361-5

Zinser, E. R.; Kolter, R. Escherichia coli evolution during stationary phase. 2004. Res
Microbiol. 155:328-36

Zuo, Y.; Xing, D.; Regan, J. M.; Logan, B. E. Isolation of the exoelectrogenic
bacterium Ochrobactrum anthropi YZ-1 by using a U-tube microbial fuel cell. 2008.
Appl Environ Microbiol. 74:3130–7
132

Appendix A: Characterization of electrochemically active bacteria utilizing a
high-throughput voltage-based screening assay

This work appears essentially as published as
Biffinger, J. C.; Ribbens, M.; Ringeisen, B.; Pietron, J.; Finkel, S.; Nealson, K.
Characterization of electrochemically active bacteria utilizing a high-throughput voltage-
based screening assay. 2009. Biotechnology and Bioengineering. 102:436-44

A.1 Overview
Metal reduction assays are traditionally used to select and characterize electrochemically
active bacteria (EAB) for use in microbial fuel cells (MFCs). However, correlating the
ability of a microbe to generate current from an MFC to the reduction of metal oxides has
not been definitively established in the literature. As these metal reduction assays may
not be generally reliable, here we describe a four- to nine-well prototype high throughput
voltage-based screening assay (VBSA) designed using MFC engineering principles and a
universal cathode. Bacterial growth curves for Shewanella oneidensis strains DSP10 and
MR-1 were generated directly from changes in open circuit voltage and current with five
percent deviation calculated between each well. These growth curves exhibited a strong
correlation with literature doubling times for Shewanella indicating that the VBSA can be
used to monitor distinct fundamental properties of EAB life cycles. In addition, eight
different organic electron donors (acetate, lactate, citrate, fructose, glucose, sucrose,
soluble starch, and agar) were tested with S. oneidensis MR-1 in anode chambers exposed
to air. Under oxygen exposure, we found that current was generated in direct response to
additions of acetate, lactate, and glucose.

133

A.2 Introduction
The discovery of microbes that couple metal reduction to produce energy for anaerobic
growth (DiChristina et al., 1988; Lovley and Phillips, 1988; Myers and Nealson, 1988)
has generated a significant amount of research on the identification and manipulation of
metal reducing microbes for applications ranging from bioremediation of heavy metals
(Lovley and Coates, 1997) to harvesting electricity from biomass (Logan and Regan,
2006). Microbes capable of metal reduction within environmental samples are commonly
isolated and identified based on their ability to reduce certain transition metal or actinide
electron acceptors (iron and manganese oxides, technetium, uranium) by either
colorimetric assays (Lovley and Phillips, 1987) or plating on agar indicator supports
(Ganesh et al., 1997; Payne and DiChristina, 2006; Taratus et al., 2000). For example,
metal reduction assays were used to identify electrochemically active bacteria (EAB)
obtained from the anode of a microbial fuel cell (MFC) submerged in sediment from
Boston Harbor, MA (Bond et al., 2002). Reduction of solid metal oxides by EAB is
generally considered to be concomitant with electricity production from MFCs. However,
the mechanisms of extracellular electron transfer from bacteria to carbon electrodes and
insoluble metal oxides are still ill-defined (Chang et al., 2006; Stams et al., 2006).
Specifically, Shewanella (one of the two families of bacteria frequently used in pure
culture MFCs) express multiple pathways for electron transfer to graphite electrodes and
manganese/iron oxides (Bretschger et al., 2007).

134

In contrast to past results, recent experiments suggest that the connection between current
output from MFCs and metal oxide reduction for unidentified EAB in the environment is
tenuous. The first significant disconnect between metal reduction and electricity output
from MFCs was recently observed with Pelobacter carbinolicus (Richter et al., 2007).
Pelobacter are of interest for MFCs because of their phylogenetic relationship to the
other major bacterial family used in MFC research, Geobacteracae (Lovley et al., 1997).
It was observed that Pelobacter was capable of reducing Fe(III) oxides but did not
generate current in a MFC. Thus, the expression of multiple pathways for electron
transfer to different electron donors and the lack of current generation from a MFC by P.
carbinolicus brings into question how many other species of EAB can reduce metal
oxides yet may not be able to deliver electrons to a carbon electrode or vice versa.

To date, little has been published on rapidly screening electrochemically active biological
species for use as energy harvesters or biosensors. Routinely, multiple MFC experiments
are performed in serial (slow due to time between experiments) or in parallel by running
multiple larger scale MFCs at the same time. Since running a single full scale MFC
requires significant space and materials, then a single device with multiple wells would
be more efficient. Rapid screening methods are desirable, as the best candidates for a
given MFC application need to be identified quickly and accurately. High-throughput
screening (HTS) continues to be primarily focused on the automation, detection, and
miniaturization of assay technology (Burbaum and Sigal, 1997; Sundberg, 2000).
Bacterial metabolism can be significantly influenced by environmental stressors (Storz
135

and Hengge-Aronis, 2000), and correlating all of the potential variables and mutations
with current output would require a standardization between research groups in the MFC
field for every bacterial sample with a defined power output.

Several strategies have been used for HTS of both whole cell or catalytic activity from
enzymes and function (Diaz-Mochon et al., 2007). After the introduction of the 96-well
microtiter plate and spectrophotometric plate readers, a clear distinction arose between
HTS and traditional laboratory assays. Some of these differences are provided in Table
A.1 and were described in detail within a review by Inglese et al (2007). A noteworthy
difference between laboratory assays and HTS are a significantly reduced sample size for
HTS and a simple protocol. There are no HTS assays in the literature that monitor
biological function as it relates directly to current generating ability or power output.

Parameter Laboratory HTS
Protocol Can be complexed with numerous
steps
Less than 10 steps, simple,
addition only
Assay volume 0.1-1mL <1-400 µL
Reagents Quantity often limited, different
batches
Single batch, stable over
long time
Variable Time, substrate, compound Compound (mg quantity),
compound concentration
Assay container Tube, slide, microtiter plate, Petri
dish, cuvette, animal
Microtiter plate
Time Milliseconds-months Minutes-hours
Output Plate reader, size separation,
radioactivity
Plate reader (fluorescence,
luminescence, absorbance)
Table A.1: Parameters for distinguishing between laboratory assays and HTS
assays.

There has been some interesting work on assay technology to identify and manipulate
bacteria for bioremediation research. DiChristina and coworkers have published a rapid
136

screening routine for identifying Mn(IV) (Burnes et al., 1998), Tc(VII) (Payne and
DiChristina, 2006), and Se(IV) (Taratus et al., 2000) reduction by Shewanella mutants
using indicator plate assays. However, changes in thermal growth conditions cannot be
varied across plate assays and only 10–12 colonies can be analyzed at one time by
digitally imaging each plate separately. Miniature biological reactors have been
developed as a way to increase the number of growth and metabolic variables within a
single device (Harms et al., 2005; Maharbiz et al., 2004). For Shewanella specifically, a
high-throughput mini-bioreactor was fabricated for the rapid screening of growth
conditions (Tang et al., 2006). Their multi-component mini-bioreactor (10 mL volume)
generated 24 different growth conditions for S. oneidensis by changing the pH, O
2
/CO
2

content, and temperature in each well but relied on external analysis of growth rates and
metal reduction.

Researchers studying hydrogen/oxygen fuel cell catalysts have developed HTS methods
for screening potential catalysts for increased activity (current output) and stability
(Smotkin et al., 2006). Unlike hydrogen/oxygen fuel cells, biological fuel cells do not
require stringent catalyst preparation; thus making it possible to create a biological
reactor with a common cathode and catholyte directly from a basic batch reactor design.
A HTS assay for EAB will need to account for the bacterial conditioning of the anode
surface and gradual bacterial biofilm formation (Kim et al., 1999, 2002); a process that
can take several weeks.

137

The device described in this work is an operational prototype of an HTS assay that uses
real time voltage detection instead of metal reduction as an indicator for potential
microbial power output from MFCs. The design is based on general MFC principles
using a ferricyanide catholyte for each assay. Shewanella oneidensis strains MR-1 or
DSP10 were used for bacterial growth studies as well as power output from various
carbon fuels. The data collected in this platform resulted in the efficient determination of
energy harvesting potential compared to using large scale individual MFCs and enabled
multiple nutrients to be screened simultaneously for current and power output from
Shewanella.

A.3 Materials and Methods
A.3.1 Culture media and chemicals. Stock solutions of D-glucose (1.0 M), D-fructose
(0.5 M) were filter sterilized (0.2 mm cellulose nitrate filter). Stock solutions of sodium
lactate (1.95 M adjusted to pH 7), sodium acetate (1.95 M), 1% agar, sucrose, 2% starch,
sodium citrate (0.5 M) were sterilized by autoclave (13 min, 1218C). All VBSA
experiments were performed a minimum of three times and conclusions were drawn from
similar trends in each experiment.

A.3.2. Strains. The DSP10 and MR-1 strains of S. oneidensis were obtained from the
Nealson lab strain collection. Both strains were inoculated from –80
o
C frozen stock
cultures, and grown in 50 mL of Luria-Bertani (LB) broth with gentle shaking (100 rpm)
at 25
o
C in 125 mL flasks aerobically.
138

A.3.3 Dimensions and fabrication of pipet microbial fuel cell. Graphite felt (GF,
Electrosynthesis Company, Lancaster, NY, 15 mg) woven with a titanium wire was
pressed into the bottom half of a separated 1 mL pipetter tip (Fisher Scientific, Pittsburgh,
PA). This chamber was then attached to a pre-treated Nafion-117 membrane with 5 min
epoxy (Devcon, Danvers, MA). The upper (anode) chamber was attached to the opposite
side of the membrane with 5 min epoxy and 25 mg of graphite felt (woven with a
titanium wire) was pressed inside. The two chambers were attached permanently with
marine epoxy (Loctite, Avon, OH) on the outside of the device (Fig. A.1). A stationary
phase culture (500 mL) of S. oneidensis DSP10 (1-10
8
CFU/mL) was added to the anode
chamber serving as the anolyte. Three pipet MFCs were placed in a stirred
50mMpotassium ferricyanide dissolved in 100 mM pH 7 sodium phosphate buffer as a
standard catholyte. The cathodes from each independent fuel cell were connected in
parallel with titanium wire.

A.3.4 Dimensions and fabrication of the four- to nine-well voltage-based screening
assays (VBSAs). VBSAs containing four to nine wells (Fig. A.3) were constructed for
these experiments. The upper 4.3 cm x 4.3 cm VBSA frame was formed from a 1.3 cm
thick polysulfone polymer sheet (Trident Engineering Plastics, Bristol, PA). The frame
was fabricated from polysulfone because of its resistance to typical sterilization
temperatures (121–125
o
C) and machinability. The diameter of each well was 0.8 cm. A
second 0.5 cm thick polysulfone polymer sheet was cut to the same size frame as above
139

with mirroring 0.5 cm diameter holes to aid in supporting the separator to the main array.
The separator between the polysulfone sheets was a pretreated Nafion1-117 membrane
(deionized (DI) water, 3% hydrogen peroxide solution, 1M sulfuric acid, and DI water at
70
o
C for 1 h each). The Nafion-117 membrane was then hot pressed (5 min, 100 psi,
150
o
C) with Toray Carbon Paper (E-TEK, TGPH-090) connected with a titanium wire.
The anodes were fabricated from a titanium metal sheet (active electrode area, 0.3 cm x
0.3 cm) coated with a conductive carbon ink. The carbon ink contained 30 mg carbon
black, 300 mL 2-propanol, 300 mL 5% Nafion Solution in water, and 2mL of de-ionized
water. The ink was sonicated for 30 min prior to application (drop-cast method) to the
bottom half of one side of etched (1 M HCl, 80
o
C, 5 min) titanium foil (Goodfellow
Cambridge Limited, Huntingdon, England) cut into an ‘‘L’’ shape. Each titanium anode
was placed in a chamber. The entire device was then assembled immediately with zinc
plated screws and sterilized in an autoclave at 121
o
C for 13 min. The fully assembled
VBSA was autoclaved as one piece including the titanium electrodes in each anode
chamber to limit bacterial contamination. A conditioning period (<24 h) was required for
the wetting of the membrane electrode assembly in the VBSA device because it was hot
pressed to carbon paper and autoclaved prior to use which resulted in the membrane
drying.

The catholyte for each experiment was a filter sterilized 50 mM potassium ferricyanide
solution in 100 mM phosphate buffer (pH 7.0). Due to the well-defined electrochemical
properties of potassium ferricyanide and its presence in excess concentration, changes in
140

overall cell voltage and current output will be dictated by each anode. There was no
detectible cross-over of the ferricyanide into the anode chamber throughout the duration
of the experiment. Marine epoxy (Loctite) was coated around the junction between the
polysulfone sheets and Nafion-117 to protect the chambers from ferricyanide seepage and
over the zinc plated screws exposed to the ferricyanide catholyte. The container for the
VBSA (Fig. A.3C) was sterilized with 10% bleach and UV irradiation in the biosafety
hood prior to use. The VBSA was placed into the catholyte solution to complete the
device. The potential of the cathode was monitored versus Ag/AgCl (Analytical Sensors,
Inc., Sugarland, TX) using an ORION 330 electrochemical apparatus (Thermo Electron
Corp., Waltham, MA). All sterile manipulations were performed in a biosafety flow
hood.

A.3.5 Data acquisition. The voltages at open circuit or across a 100 kV resistor (in a
custom nine-resistor bank made for simultaneous measurements) were recorded with a
personal data acquisition device (I/O tech, personal daq/54) every 2 min. Ohm’s law was
used to convert voltage to current and to generate polarization curves. The polarization
curves for each pipet fuel cell were recorded by changing the external resistance of each
fuel cell independently.

A.3.6 Monitoring the growth of S. oneidensis with voltage and current. In a five-well
VBSA, 300 mL of sterile LB was inoculated with 100 mL of a culture containing 1-10
8

CFU/mL S. oneidensis MR-1 or DSP10 in LB. The growth of the cells was monitored
141

with time using the voltage output from each well versus the ferricyanide cathode system.
A 100 kV resistor was used during experiments for the determination of current output
for either growth or carbon source utilization. Growth experiments following current
output contained the redox mediator 9,10-anthraquinone-2,6 disulfonic acid (AQDS,
5mM) to eliminate changes in current derived from differential mediator secretion and
biofilm formation.

A.4 Results
A.4.1 Pipet microbial fuel cells. To date, a multi-anode/common cathode MFC has not
been reported in the literature. The initial design concept for the VBSA was brought to
practice by simply modifying 1 mL pipetter tips (Fig. A.1). Three pipet MFCs were
placed in a standard potassium ferricyanide catholyte. The ferricyanide concentration was
in excess so that large changes in localized concentration would not affect the working
potential of the cathode. Sodium lactate was added periodically to the anode chambers
containing S. oneidensis DSP10 to concentrations of 20–30 mM over the period of 4–6
days during the experiment.

142

Figure A.1: Images of the pipet MFC. A schematic (A) and image (B) of pipet MFC
operating with titanium wires are depicted.

The voltages across an external resistor (8,600 or 4,700 V) or at open circuit (OCV) were
recorded with time from three pipet MFCs (Fig. A.2A). There was a negligible difference
in voltage output between MFCs 1 and 3 under identical conditions, while MFC 2
exhibited a slight variation from the other two. In general, the use of GF in these designs
is limited because the connection between the titanium wire and GF deteriorates upon
wetting. Therefore, titanium– titanium connections were used for the anodes in the
subsequent VBSA design, resulting in a substantial decrease in variability for the VBSA
(below). The difference between MFCs 1/3 and MFC 2 was also observed when
polarization curves were calculated from each pipet MFC (Fig. A.2B). The average open
circuit voltages and short circuit currents were 0.65-0.05 V and 0.021-0.003 mA (Fig.
A.2B inset) from each MFC, respectively. The average power density from S. oneidensis
DSP10 in all three MFCs was 4,400-500 W/m3 (per volume) and 0.20-0.02 mW/m
2
(per
cross-sectional electrode surface area).

143

Figure A.2: Performance of pipet MFCs. Time-voltage curves (A) of three pipet
MFCs containing S. oneidensis MR-1 with connected cathodes and power (B) and
voltage (inset) versus current for individual pipet MFCs.

144

The power density (per surface area) of the pipet MFCs are considerably less than
previously published MFCs. The best comparison of power density is with a miniature
MFC (mini-MFC) (Ringeisen et al. 2006) because it was designed to maximize the
surface area of the electrode to volume of the chamber similar to the pipet MFC. Due to
small anode volume (500 mL) and the minimization of void volume (GF filled fuel cell
chambers), these pipet MFCs were able to be operated with air exposed anodes. Power
from aerobic cultures was also observed using the mini-MFC (Ringeisen et al., 2007).
The pipet MFC (as well as any MFC utilizing ferricyanide catholytes) could also be
operated under anaerobic atmospheres because the redox couple for ferricyanide is not
oxygen dependent. The percent deviation of all three pipet MFCs was 8% from the mean
voltage recorded. This initial pipet design highlights the concept that power generation
can be monitored in multiple biological air exposed anodes using a single cathode with a
well-defined redox active catholyte.

Figure A.3: Images of the nine-well VBSA. Schematic (A), operating VBSA (B) and
device container with electrical connects (C) are depicted above.

A.4.2 General VBSA design and properties. The experimental design of pipet MFCs
was used to fabricate an improved single frame batch reactor with a common cathode
(Fig. A.3). The total Ohmic resistance of an MFC can be quite pronounced considering
145

the optimal operating conditions (pH 1, 60
o
C) for Nafion membranes (Harnisch et al.,
2008; Rozendal et al., 2006; Zhao et al., 2006) are not used in these systems. Voltage
data collected from a fourwell VBSA containing S. oneidensis DSP10 inoculated with
lactate was used to calculate the total Ohmic resistance (RV, Eq. 1) of the system via the
current-interrupt method (Cooper and Smith, 2006). The initial change in voltage (DE)
(0.0112, Fig. A.4 inset) after operating the VBSA at 4 mA (I) resulted in a RV of 2,800-
100 V. This high resistance was most likely due to the low power output and the high
resistive state of the
R
Ω
= ΔE/I
Nafion MEA initially. However, the total Ohmic resistance decreased considerably after
25 h of operation (1,100-50 V). This was a common property with this prototype and
resulted in a mandatory 10–25 h conditioning period for consistent results to be obtained
from the VBSA. Conditioning the VBSA was accomplished by soaking each anode
chamber with either sterile water or the medium of interest. Variation derived from the
VBSA itself are mitigated by recording data against a control experiment run in parallel
with the experiments of interest. The percent deviation between each of the wells (5%)
was determined from deviations in the mean voltage recorded throughout triplicate
current-interrupt data sets (Fig. A.4).

146

Figure A.4: Parellel current-interrupt data a 4-well VBSA containing S. oneidensis
MR-1. Linear fit of voltage versus time (inset) data for total Ohmic resistance
calculation.

A.4.3 Growth of S. oneidensis MR-1 and DSP10 using the VBSA. A variety of
analytical methods (bioluminescence, immunology, microscopy, flow cytometry,
electrochemical) are currently being employed to monitor the safety and quality of food
and agricultural products from bacterial contaminants. There are few rapid and sensitive
methods to monitor bacterial load and activity with most direct assays relying on oxygen
respiration activity which is often inaccurate for anaerobic bacteria (Kuznetsov et al.,
2004). A recently published assay for the rapid analysis of bacterial growth and metabolic
activity used external mediators with integrated electrochemical methods for
Mycobacterium smegmatis (Kuznetsov et al., 2006). Their system consisted of a standard
three-electrode cell with working, reference, and platinum counter electrodes. The VBSA
could also be used to monitor bacterial activity similar to the three electrode
147

electrochemical systems by utilizing voltage changes with or without external mediators
depending on the bacterial species. Standard three electrode systems will eventually be
susceptible to biofouling of the reference and counter electrode considering they are both
placed in the bacterial culture. The VBSA is designed with a universal ferricyanide
cathode system separated from the bacterial culture by a Nafion1-117 membrane making
any changes in the cathode potential capable of being monitored independently by an
additional reference electrode (example: Ag/AgCl), if necessary, and isolated from the
bacteria of interest.

The growth of S. oneidensis MR-1 and DSP10 was examined with the VBSA using both
open circuit voltage (OCV) (Fig. A.5A) and current (Fig. A.5B) at pH 7. A six-well
VBSA was assembled with 300 mL cell-free LB media in each anode chamber. After the
initial conditioning period (25 h), 100 mL inocula from 1-10
8
CFU/mL MR-1 and DSP10
cultures were added to each well. The voltage data collected from each chamber was
substracted from a control experiment containing acellular LB. This data correction was
necessary considering the presence of background voltage changes due to LB and carbon
electrodes.

The data obtained at OCV would be the most rapid way to analyze the growth of
Shewanella considering once the maximum potential difference is obtained, general
activity could be monitored by a change in the OCV. There were large differences
observed between the growth rates calculated using the VBSA with OCV and standard
growth experiments (e.g., serial dilution or optical density). Typical OCVs reported in the
148

literature for MFCs using a ferricyanide catholyte and Shewanella (500–800 mV) were
observed approximately 1 h after the MR-1 culture was added and (after a 1.2 h lag
period) 1 h after the addition of DSP10 (Fig. A.5A). Typical doubling times (calculated
during exponential growth) for S. oneidensis range from 35 to 45 min under comparable
conditions to those used for the VBSA (Abboud et al., 2005; Biffinger et al., 2008b).
Doubling times calculated from the OCV data are 12 min for DSP10 and 18 min for MR-
1. These data suggest OCV is a poor indicator for bacterial growth. The growth curves
generated using OCV are more likely showing the bacterial conditioning (biofilm
formation, biosynthesis of redox mediators) of the electrode considering OCV is a
measure of the maximum working potentials between the anode and cathode. Since there
are no literature precedents for the rate bacteria condition an electrode, this metric should
be useful in screening new electrode materials and coatings.

Figure A.5: Current generated from the VBSA. Above are shown time versus open
circuit voltage (A) or current curves (B) generated from the growth of S. oneidensis
DPS10 and MR-1 using the VBSA.

Large changes in OCV can be useful for qualitative measurements of the bacterial culture
properties and electrode interactions, but OCV does not correlate to actual bacterial
149

growth. If an external resistor is placed in series with each anode in the VBSA and
connected to the universal cathode then each well will behave as individual MFCs. A
second set of experiments were designed to relate current output directly with S.
oneidensis growth in the presence of AQDS. An external mediator was added to
eliminate current responses due to biofilm formation and bacterial conditioning of the
electrode surface. The doubling times calculated for DSP10 and MR-1 during the first 10
h of data in Figure A.5B are 34 and 42 min, respectively. These doubling times were
calculated using the initial slope from the current increase after the addition of bacteria to
each anode well and are consistent with literature values (above). The data collected after
the rise in current shows a gradual decrease in current to a steady state 10 h after the
initial addition. Since current is being collected from these fuel cells, the gradual decrease
in current is in response to a decrease in viable carbon sources for Shewanella. This
gradual drop in current after 10 h is nothing like the OCV data (Fig. A.5A) that showed
no decrease in voltage until 25 h after the addition of bacteria (data not shown). The
actual currents that were generated by MR-1 and DSP10 can be deceiving considering
that both strains utilize LB in dissimilar ways for current output. However, both strains
utilize lactate the same and this was reflected in the identical maximum current output
observed from DSP10 and MR-1 in Figure A.5B when lactate was used as the sole
carbon source.

Several properties of EAB can be observed by using the VBSA with OCV or current
detection. These data suggest that OCV is a better metric for bacterial conditioning of
150

electrodes and surfaces and not for bacterial growth. Conversely, current appears to be
the best detection metric for bacterial growth rates. The use of external mediators is
optional for screening EAB, but should be used when the effects of bacterial conditioning
and biofilm formation are to be negated.

A.4.4 Screening for carbon substrate utilization by air exposed S. oneidensis MR-1.
The VBSA will be useful for both applied and basic research applications involving
biological electrochemically active species. Our interest in MFCs and Shewanella led us
to screen a variety of potential electron donors for current output. Until recently, only a
small number of organic electron donors (lactate, formate, pyruvate, amino acids,
hydrogen) have supported metal reduction and power output from MFCs using S.
oneidensis MR-1 under anaerobic atmospheres (Nealson et al., 2002). However, electron
donors (such as glucose) that were thought to be unusable for power output by
Shewanella have generated substantial power under an air exposed atmosphere using a
miniature MFC (Biffinger et al., 2008a). Therefore, the effects of oxygen exposure on S.
oneidensis MR-1 were tested with the VBSA prototype with various electron donors of
interest (agar, sucrose, starch, glucose, fructose, lactate, acetate, citrate) using current
detection. A 100 kV resistor was put in series with each anode well and connected to the
universal cathode to generate a current. The use of an unusually high resistor will allow
for comparisons to be made between EAB that produce both low and high currents, but
will not optimize power output.

151

There were four distinct time periods defined during the operation of the VBSA (Fig.
A.6). Each period was defined by the addition of the carbon source to a concentration of
8 mM in each well. Each anode well contained sterile water plus the electron donor for
the first 10 h to condition the VBSA. At the beginning of time period A (Fig. A.6), 1-10
8

CFU/mL S. oneidensis MR-1 cultures in pH 7 LB replaced the sterile water with 1 mM
electron donor. The control experiment contained the original S. oneidensis MR-1 culture
in LB and these data were used to standardize results between chambers. The largest
current output was generated by lactate (45 mA) over the next 35 h during time period (C
and D). Lactate produced comparable current after all additions following time period B.
This suggests that bacterial conditioning of the electrode was completed after 100 h of
operation.

152

Figure A.6: Current generated from air exposed anodes containing S. oneidensis
MR-1 with selected electron donors in a nine-well VBSA. Sections A-D indicate
when electron donor was added.

153

Of the remaining electron donors, only glucose and acetate generated consistent current
responses from MR-1. These currents were significantly less than the 45 mA generated
by lactate (-10 mA) but definite positive current output was observed upon addition of
glucose and acetate. Starch, agar, citrate, sucrose, and fructose did not show any
significant response upon addition of each electron donor. The minor decreases in current
at the beginning of each time period are from exposure to oxygen at the electrode during
the mixing of the electron donor and the anolyte. During time period B, MR-1 exposed to
starch, agar, sucrose, fructose, and citrate show significant increases in current but no
comparable response from subsequent additions. These isolated responses suggest that all
of the usable electron donors were exhausted resulting in either increased biofilm
formation on the electrode (due to nutrient limited conditions) or a cell death response
with the residual release of mediators and potential electron donors from cellular debris.
Experiments are being performed currently to determine the factors that generate current
spikes under carbon source limited conditions.

The results from this electron donor VBSA experiment show the complexity of bacterial
metabolism and highlight the positive attributes of this prototype. Not only was a unique
starvation behavior observed with respect to current output from starch, agar, citrate,
sucrose, and fructose (response at the end of time period (B), but lactate and acetate were
confirmed as electron donors that could be utilized by MR-1 under oxygen exposure for
current output from a MFC. In addition, glucose was identified as a potential electron
donor, but only after acclimation of the air exposed anodes over several hours.
154

A.5 Discussion
The VBSA is the first high throughput prototype using voltage and current output to
monitor fundamental EAB properties as well as screen for potential electron donors in
parallel for MFCs applications. The straightforward design for the VBSA makes it
applicable for both microbial and enzymatic fuel cell engineering research. The use of a
universal cathode and a well-defined catholyte (potassium ferricyanide) allows for small
changes in voltage (and current) to be analyzed between each chamber with excellent
reproducibility.

155

Appendix B: Simultaneous analysis of physiological and electrical output
changes in an operating microbial fuel cell with Shewanella oneidensis.

This work appears essentially as published as
Biffinger, J. C.; Ray, R.; Little, B. J.; Fitzgerald, L. A.; Ribbens, M.; Finkel, S. E.;
Ringeisen, B. R. Simultaneous analysis of physiological and electrical output
changes in an operating microbial fuel cell with Shewanella oneidensis. 2009.
Biotechnology and Bioengineering, 103:524-31

B.1 Overview
Changes in metabolism and cellular physiology of facultative anaerobes during oxygen
exposure can be substantial, but little is known about how these changes connect with
electrical current output from an operating microbial fuel cell (MFC). A high-throughput
voltage based screening assay (VBSA) was used to correlate current output from a MFC
containing Shewanella oneidensis MR-1 to carbon source (glucose or lactate) utilization,
culture conditions, and biofilm coverage over 250 h. Lactate induced an immediate
current response from S. oneidensis MR-1, with both air-exposed and anaerobic anodes
throughout the duration of the experiments. Glucose was initially utilized for current
output by MR-1 when cultured and maintained in the presence of air. However, after
repeated additions of glucose, the current output from the MFC decreased substantially
while viable planktonic cell counts and biofilm coverage remained constant suggesting
that extracellular electron transfer pathways were being inhibited. Shewanella maintained
under an anaerobic atmosphere did not utilize glucose consistent with literature
precedents. Operation of the VBSA permitted data collection from nine simultaneous S.
oneidensis MR-1 MFC experiments in which each experiment was able to demonstrate
organic carbon source utilization and oxygen dependent biofilm formation on a carbon
156

electrode. These data provide the first direct evidence of complex cellular responses to
electron donor and oxygen tension by Shewanella in an operating MFC at select time
points.

B.2 Introduction
Microbial fuel cells (MFCs) are electrochemical devices capable of generating an
electrical current directly from the oxidation of carbon electron sources using bacterial
metabolic pathways. These devices are currently being developed for a variety of
applications ranging from the generation of electricity using wastewater (Aelterman et
al., 2006; Angenent et al., 2004) to autonomous power sources for sensors (Shantaram et
al., 2005) and beacons (Tender et al., 2008). Different bacterial strains have evolved a
variety of strategies for delivering electrons to solid electron accepting materials (metal
oxides, carbon electrodes). For example, direct contact with the electrode surface is
required for Geobacter sulfurreducens to generate current from a MFC (Reguera et al.,
2006) while Shewanella oneidensis MR-1 can deliver reducing equivalents to electron
accepting surfaces without direct contact (Lies et al., 2005) using redox mediators
(Marsili et al., 2008; von Canstein et al., 2008). Several reviews have been published
recently that summarize the significant progress in understanding electron transport
pathways within electrochemically active bacteria (EAB) (Chang et al., 2006;
Fredrickson et al., 2008; Hernandez and Newman, 2001; Lovley, 2008; Schro¨der, 2007).
However, few studies have been published that address the complex real-time cellular
157

physiological changes that determine how EAB interact with anodes during MFC
operation (Lanthier et al., 2008).

Both environmental bacterial consortia and single strain MFCs are reported in the
literature (Logan et al., 2006). Working with single strains allows mechanistic and
physiological details to be observed directly. Two bacterial families, Geobacteracea and
Shewanellacea, are commonly used in pure culture MFC research. S. oneidensis MR-1 is
a facultative, anaerobic g-proteobacterium capable of dissimilatory metal reduction
(Myers and Nealson, 1988) as well as generating current within MFCs (Bretschger et al.,
2008; Kim et al., 2002; Ringeisen et al., 2006). Shewanella was chosen for this work
because of its adaptability to aerobic and anaerobic environments.

Glycolytic aerobic metabolic pathways in Shewanella have been identified through
genomic and proteomic studies (Beliaev et al., 2005; Driscoll et al., 2007; Fang et al.,
2006; Leaphart et al., 2006; Serres and Riley, 2006; Wan et al., 2004) and experimental
evidence linking current output with glucose metabolism was recently reported using
aerobic cultures of S. oneidensis DSP10 in a miniature MFC (Biffinger et al., 2008). Prior
to the aforementioned work, S. oneidensis was considered limited in the range of organic
electron sources (e.g., formate, lactate, pyruvate, amino acids) that could be used for
anaerobic metal reduction or current output from MFCs (Nealson et al., 2002). Bacterial
metabolic pathways dictate how different types of organic electron sources
(carbohydrates, linear carboxylic acids, polysaccharides) are utilized for current output
158

from MFCs. Therefore, in situ monitoring of both cellular and culture environment
conditions is important for improving the long-term survivability of MFC devices.

Since removing electrode samples from a continuously operating MFC is not practical
and running multiple laboratory scale MFCs under identical conditions for sampling
electrodes is unfeasible, no research has been reported on the study of microbial cellular
changes within an operating MFC. Direct real time measurements of biofilm formation
and coverage have been analyzed by nuclear magnetic resonance (NMR) (McLean et al.,
2008) and confocal microscopy (Teal et al., 2006) on transparent supports. However
experiments performed on non-conductive surfaces may not be germane to the conditions
in an operating MFC (Lanthier et al., 2008). Indirect real time analysis of biofilm
formation by electrochemical impedance spectroscopy (EIS) (Manohar et al., 2008;
Ramasamy Ramaraja et al., 2008) or utilizing pre-formed biofilms on electrodes placed in
MFCs (Venkata Mohan et al., 2008) have also been used to monitor biofilm dynamics but
are difficult to relate to actual biofilm coverage in an operating MFC.

In this study, a voltage based screening assay (VBSA) was used to monitor voltage
output from EAB under both closed and open circuit conditions (Biffinger et al., 2009).
Additionally, the use of a high-throughput assay for monitoring current output from
bacteria provided a pathway to correlate electrical current output with cellular and
metabolic changes; factors that have not been studied within an operating MFC to date.
The VBSA was used to monitor real-time current output as it relates to anaerobic and air-
159

exposed cultures, planktonic cell concentration, and extent of biofilm formation on a
carbon electrode at defined time points during the experiment. The combination of these
data resulted in a physiological description of how Shewanella respond to glucose and
lactate in the presence of oxygen in an operating MFC.

B.3 Materials and Methods
B.3.1 Culture media and chemicals. A stock solution of 1.95 M sodium lactate was
adjusted to pH 7.0 and sterilized by autoclaving for 15 min at 1218C. A D-glucose (1 M)
stock solution was sterilized with a 0.2 mm cellulose nitrate filter. Luria-Bertani (LB)
Broth (Miller) and LB/agar (Difco LB-Agar, Miller) was used for liquid cultures and
plates, respectively (Fisher Scientific, Inc, Pittsburgh, PA). The solvent for each solution
was Millipore 18 MV water. Serial dilutions for viable planktonic cell concentration
measurements were performed in phosphate buffered saline (pH 7.0) with 0.03% Triton-
X100 (Sigma–Aldrich, Milwaukee, WI).

B.3.2 Strains. S. oneidensis MR-1 (obtained from Dr. Kenneth Nealson (University of
Southern California, Los Angeles, CA)) was grown from a single colony isolated from a
LB/agar plate inoculated from a -80
o
C glycerol stock culture. A single colony was
transferred to 50 mL of LB broth and incubated aerobically at 25
o
C with gentle shaking
(100 rpm). Experimental cultures were sub-cultured after 20 h of growth three times
before being used in VBSA experiments. Anaerobic S. oneidensis MR-1 cultures were
created from an MR-1 culture, which was incubated aerobically for 48 h. These aerobic
160

stationary phase MR-1 cultures were then degassed with a nitrogen purge and shaken
(100 rpm) in an anaerobic chamber for 24 h at 25
o
C prior to MFC experimentation.

B.3.3 Construction and data acquisition. Dimensions and fabrication of the VBSA
were published previously (Biffinger et al., 2009). The anodes were singlesided carbon-
coated titanium flags and the cathode system was graphite paper in a 50 mM potassium
ferricyanide (dissolved in 100 mM phosphate buffer at pH 7.0). Each experiment was
completed in a nine-well VBSA apparatus depicted in Figure B.1. Experiments with no
addition of glucose or lactate were labeled as blank. Planktonic cell concentrations of
each well were determined from serial dilution of aliquots in phosphate buffered saline
with 0.03% Trition-100 and plated onto LB/agar with average cell counts reported for
glucose, lactate, and blank during the experiment. Once the electrode was removed for
environmental scanning electron microscopy (ESEM) fixation, the well was no longer
used for planktonic cell concentration determination. The voltages across a 100 kV
resistor bank (in a custom nine-resistor bank made for simultaneous measurements) were
recorded with a personal data acquisition device (I/O tech, personal daq/54) every 4 min.
Ohm’s law was used to convert voltage to current. Anaerobic (performed in a Coy
instruments anaerobic chamber) and aerobic (or air-exposed) experiments were
performed at 23
o
C.

B.3.4 Imaging of S. oneidensis MR-1 biofilms. Environmental scanning electron
microscopy (ESEM) of carbon surfaces on the titanium anodes was performed at the
161

Naval Research Laboratory, Stennis Space Center, (NRLSSC). Unattached biomass was
removed by washing each anode with three separate 1 mL aliquots of phosphate buffered
saline solution at the Naval Research Laboratory, Washington, DC (NRLDC). Each
anode was placed in 2 mL of 4% cacodylate buffered glutaraldehyde fixative (Ray et al.,
1997) in water at NRLDC and fixed for at least 24 h at 48
o
C prior to shipment to
NRLSSC without further manipulations. Anaerobic samples were fixed in the anaerobic
hood using degassed 4% cacodylate buffered glutaraldehyde fixative. For collecting
ESEM images, each anode was removed from the fixative and washed with 50 mL of
distilled water. After 2 min of gentle rinsing, each anode was placed on a mounting stub
on the Peltier cooling device inside the ESEM chamber. The anodes were kept wet/moist
by using the Peltier cooling device maintained at 48
o
C and a chamber water vapor
pressure between 4.5 and 5.5 torr. Water vapor was allowed to condense on the cooled
anodes to keep it moist while performing ESEM imaging. Liquid water was removed
from the top layer, several microns thick, to view the biofilm on the carbon surface of
each titanium anode. A gaseous secondary electron detector (GSED) was used to collect
the ESEM images of the wet/moist sample surface.

B.4 Results
The combination of time course results from ESEM images of electrode surfaces, viable
planktonic cell densities, and electrical current output generates a broader understanding
of how S. oneidensis interact with carbon electrode surfaces in an operating batch MFC.
Correlating the three parameters mentioned previously was made possible by using a
162

small modular array of identical MFCs operated in parallel. These data demonstrate
distinct cellular differences with carbon source utilization and oxygen tension as well as
providing insight into the role cellular physiology plays on current output from an
operating S. oneidensis MFC.

Figure B.1: Diagram of the nine-well VBSA. Depicted above is the experimental
setup for both aerobic and anaerobic studies. Electrodes were removed and
chemically fixed at time 1(t
1
), time 2(t
2
) and time 3(t
3
) with the carbon electron
source indicated in each well.

B.4.1 Lactate metabolism by S. oneidensis. Current output (Fig. B.2) from S. oneidensis
MR-1 was correlated to both planktonic cell density (Fig. B.2) and biofilm formation
(Fig. B.3) with lactate as the sole electron source. Subsequent additions of lactate over
163

the first 170 h for air exposed anodes resulted in a four-fold current increase (Fig. B.2A).
The remaining 100 h of the experiment resulted in a doubling of the current output. In
general, the current output doubled from successive additions of lactate to air-exposed
MR-1. The maximum current generated by anaerobic MR-1 with lactate (Fig. B.2B) was
eight-fold less than air-exposed MR-1 (Fig. B.2A). However, there were rapid current
responses (<4 min) from lactate additions for anaerobic cells.

Figure B.2: Average current output from S. oneidensis MR-1-containing VSBA.
Aerobic (a) or anaerobic (b) atmospheres with 10mM lactate as the sole electron
carbon source with baseline correction. Secondary axis reports planktonic cell count
with time in colony forming units (CFU/mL). Solid vertical line indicate when
164

lactate was added and block arrows designate when anode was removed and
chemically fixed for ESEM at t
1
, t
2
and t
3
.

The planktonic cell density remained constant for lactate (~8x10
8
CFU/mL) in both the
presence and absence of air (Fig. B.2A and B, respectively). Planktonic cell density in the
blank anode decreased exponentially after 70 h, correlating with LB nutrient depletion.
Biofilm formation was weak for all blank electrodes under aerobic (Fig. B.3A–C) and
anaerobic (Fig. B.5A–C) atmospheres. Electrodes from air-exposed anode chambers (Fig.
B.3G–I) showed significant biofilm coverage with a complete lawn of MR-1 formed over
the entire anode surface after 220 h of operation (Fig. B.3H). Anaerobic MR-1 did not
form a substantial biofilm in the presence of lactate (Fig. B.3D–F).
165

Figure B.3: ESEM images of the chemically fixed carbon anode surfaces from
VBSAs with lactate. Anodes removed at times indicated by block arrows in Figure
B.2 from acellular (a-c) and S. oneidensis MR-1 anaerobic (d-f) or air-exposed (g-i)
anode chambers with lactate as the sole carbon electron source. Scale bar is 10 µm.

The gradual increase in current with time (Fig. B.2A) correlated with the formation of
biofilm for MR-1 (Fig. B.3G–I). This gradual current increase is typically described as a
conditioning period where the bacteria modify the electrode surface for either bacterial
attachment or mediator release. However, when using air-exposed anodes this gradual
166

increase in current should also be attributed to a decrease in oxygen concentration at the
electrode surface, which would eliminate the competitive oxygen reduction reaction and
increase the Coulombic efficiency of the MFC. The Coulombic efficiency doubled as
substantial biofilm was formed on the electrode surface (Fig. B.2a). This concept of
oxygen gradients in biofilms was first demonstrated using direct microelectrode
measurements showing a decrease in oxygen concentration with increasing biofilm
thickness (Rasmussen and Lewandowski, 1998) and is consistent with these results.

Since S. oneidensis does not need to be in contact with electrode surfaces to deliver
electrons at a distance, planktonic cell density would impact current output significantly
for a Shewanella containing MFC. The viable planktonic cell count remained essentially
constant throughout the air-exposed experiment, and significant current was generated
immediately, even with sparse biofilm formation over the first 100 h of operation. We
observed little change in planktonic cell concentration with time (Fig. B.2A), but found a
significant increase in biofilm coverage on the anode (Fig. B.3). This experiment
demonstrated that the decrease in oxygen concentration at the anode and increased
number of bacteria near the electrode surface is primarily responsible for the gradual
increase in current from lactate. This colonization of the electrode is certainly enhanced
in the presence of oxygen when comparing ESEM images from MR-1 exposed to air
(Fig. B.3G–I) and anaerobic experiments (Fig. B.3D–F).

167

There has been only one other study that has monitored Shewanella growth and biofilm
formation with an active MFC carbon electrode as the sole electron acceptor (Lanthier et
al., 2008). Our results are consistent with their observation that planktonic biomass is
primarily responsible for current output from anaerobic S. oneidensis containing MFCs
but is not consistent for air-exposed cultures. It is clear from our results that substantial
biofilms of S. oneidensis MR-1 are formed with air-exposed anodes and lactate (Fig.
B.3G–I) in an operating batch MFC, while a significant biofilm is not formed under
anaerobic conditions (Fig. B.3D–F). Therefore, a lack of biofilm formation yet a
sustained planktonic cell concentration over time indicates Shewanella utilizes lactate as
a food source and our observations under anaerobic conditions indicate that planktonic
cells rather than direct cell-anode contact are primarily responsible for current output.

B.4.2 Glucose metabolism by S. oneidensis. Until recently, only a limited range of
organic electron sources were known which S. oneidensis could use for anaerobic metal
reduction or current output in a MFC (Fredrickson et al., 2008). Lactate is one such
electron source that has been utilized for studies of current production from a Shewanella
MFC (Kim et al., 1999, 2002; Ringeisen et al., 2006) and also was shown in the previous
section. However, the natural abundance of lactate is limited. Therefore, in order to use a
Shewanella-containing MFC in a variety of applications, such as an autonomous power
source for sensors, we must understand the physiological role naturally occurring electron
sources might play on bacteria. The results presented here demonstrate a simultaneous
168

time course analysis of the physiological and electrical output changes S. oneidensis
undergoes in an operational MFC with glucose as the sole electron source.

S. oneidensis MR-1 cultured and exposed to air within a MFC (Fig. B.4A) can utilize
glucose as an electron source for current production. However, repeated additions of
glucose resulted in a gradual increase in current over the first 150 h with a subsequent
decrease in current after this time period (Fig. B.4A). The addition of glucose did result
in smaller current increases after 150 h for MR-1 but significantly less than the maximum
current of 17 mA recorded in the first 150 h. This result is consistent with similar
experiments using air-exposed S. oneidensis DSP10 cultures in a flowing miniature MFC
(Biffinger et al., 2008) with glucose as the sole electron source. Significantly more of the
electrode surface was covered by MR-1 using glucose with oxygen exposure (Fig. B.5G–
I) than without (Fig. B.5D–fF). Planktonic cell density remained high for air-exposed
cells indicating that glucose was being utilized by Shewanella during this experiment.
169

Figure B.4: Average current output from S. oneidensis MR-1-containing VBSA.
Aerobic (a) or anaerobic (b) conditions with 10mM clucose as the sole electron
carbon source with baseline correction. Secondary axis reports planktonic cell count
with time in colony forming units (CFU/mL). Solid vertical lines indicate when
glucose was added and block arrows designate when anode was removed and
chemically fixed for ESEM at t
1
, t
2
and t
3
.

Current output after addition of glucose for air-exposed MR-1 was initially weak (<5
mA) but over the next 150 h generated approximately 17 mA (Fig. B.4A). Since
improvements in current output correlate with biofilm formation with air exposed
cultures, then maximizing biofilm formation is a major factor in optimizing Shewanella
containing MFCs. In general, there were only sparse biofilms formed when glucose was
170

the sole electron source under all conditions. Air-exposed MR-1 current vs. time data
(Fig. B.4a) is consistent with Shewanella utilizing glucose upon continued exposure to
oxygen. The present experiments show distinct cellular physiological responses from
repeated exposure to glucose as well as a gradual decrease in current after 170 h
consistent with previous results (Biffinger et al., 2008). The viable planktonic cell
concentration and biofilm coverage remained constant throughout the air-exposure
experiments, suggesting that repeated additions of glucose eventually down-regulates
extracellular electron transport pathways in favor of sustaining growth. The conservation
of energy for growth rather than biofilm formation on an electrode by anaerobic S.
oneidensis was reported recently, although a complete picture of bacterial growth
changes on the electrode was not provided (Lanthier et al., 2008).
171

Figure B.5: ESEM images of the chemically fixed carbon anode surfaces from
VBSAs with glucose. Anodes removed at times indicated by block arrows in Figure
B.4 from acellular (a-c) and S. oneidensis MR-1 anaerobic (d-f) or air-exposed (g-i)
anode chambers with lactate as the sole carbon electron source. Scale bar is 10 µm.

S. oneidensis is capable of reducing a wide range of electron acceptors, but only a small
number of electron donors have been utilized effectively in anaerobic environments
(Fredrickson, et al., 2008). The initial current response of S. oneidensis under anaerobic
conditions to glucose is attributed to the transition of aerobically cultured cells to an
172

anaerobic MFC environment. Since air exposed stationary phase MR-1 cultures were
degassed for anaerobic experiments, the decrease in current with time after 170 h is
consistent with the decreased expression of proteins in glycolytic pathways under low
oxygen levels (Scott and Nealson, 1994). This conclusion is also supported by the
repeated positive current responses to additions of glucose in the presence of oxygen
(Fig. B.4A), while anaerobic MR-1 cells did not generate any current response with
repeated glucose additions (Fig. B.4B). Planktonic cell densities also decreased with time
under anaerobic conditions and glucose exposure, while air-exposed cells were able to
maintain their cell density (Fig. B.4). All of these data indicate that glucose was
metabolized when exposed to air and not utilized efficiently under anaerobic conditions
in a single experiment.

B.5 Discussion
The miniature modular design of the VBSA resulted in the first time-lapse analysis
correlating cellular physiological responses to current output from an operating MFC.
Large differences in current output and physiology were observed between MFCs
utilizing air-exposed and anaerobic MR-1 cultures exposed to glucose and lactate. The
reduced response in current generation from lactate-exposed anaerobic S. oneidensis MR-
1 was fivefold greater than the current response from glucose-exposed anaerobic MR-1.
However, the sustainability of aerobic Shewanella cultures in the presence of glucose, a
naturally occurring electron source, is a promising result for developing long-term
autonomous sensors. Nonetheless, the fact that sustained current production has not been
173

demonstrated when glucose is the sole electron donor means that consortia will still be
necessary to achieve efficient energy harvesting by MFCs. These results demonstrate, for
the first time, the ability to correlate current output in relation to carbon source
utilization, culture conditions, and biofilm coverage in an operational MFC.

174

Appendix C: The utility of Shewanella japonica for microbial fuel cells

This work appears essentially as published as
Biffinger, J. C.; Fitzgerald, L. A.; Ray, R.; Little, B. J.; Lizewski, S. E.; Petersen, E.
R.; Ringeisen, B. R.; Sanders, W. C.; Sheehan, P. E.; Pietron, J. J.; Baldwin, L. J.;
Johnson, G. R.; Ribbens, M.; Finkel, S. E.; Nealson, K. H. The utility of Shewanella
japonica for microbial fuel cells. 2011.Bioresource Technology. 102:290-7

C.1 Overview
Shewanella-containing microbial fuel cells (MFCs) typically use the fresh water wild-
type strain Shewanella oneidensis MR-1 due to its metabolic diversity and facultative
oxidant tolerance. However, S. oneidensis MR-1 is not capable of metabolizing
polysaccharides for extracellular electron transfer. The applicability of Shewanella
japonica (an agar-lytic Shewanella strain) for power applications was analyzed using a
diverse array of carbon sources for current generation from MFCs, cellular physiological
responses at an electrode surface, biofilm formation, and the presence of soluble
extracellular mediators for electron transfer to carbon electrodes. Critically, air-exposed
S. japonica utilizes biosynthesized extracellular mediators for electron transfer to carbon
electrodes with sucrose as the sole carbon source.

C.2 Introduction
Developing carbon-neutral renewable energy sources is an important research area for
alternative power systems. Electricity generated from fuel cells has found large scale
application in all aspects of transportation as well as stationary power supplies (Larminie
and Dicks, 2003). Biological fuel cells are an alternative technology to commercial H
2
/O
2

proton exchange membrane fuel cells in that they operate at ambient temperatures, in
175

aqueous environments, and with minimal energy input. Microbial fuel cell (MFC)
technology has rapidly advanced in the last five years with an outpouring of both new
devices and identification of microbial strains or consortia that convert a variety of
carbon sources directly into electricity (Watanabe, 2008; Harnisch and Schroder, 2009;
Zhao et al., 2009). A particularly attractive facet of MFCs is their use of many fuels, even
wastewater, to generate power. This is a welcome relief to the stringent fuel purity
requirements of most conventional proton exchange membrane (PEM) fuel cells.
However, MFCs typically suffer from low current densities (Logan and Regan, 2006). To
address this issue, new electrode materials and/or new microbes or consortia need to be
developed.

Bacteria of the groups Shewanellaceae and Geobacteraceae are classic models in MFC
research because of the breadth of knowledge about their metabolism and versatility
(Lovley, 2006; Fredrickson et al., 2008). The compatibility of fuel types in a given MFC
is defined by the metabolism of the microorganism coupled to extracellular electron
transfer to generate electrical output but is not necessarily linked to metal reduction
(Richter et al., 2007). Geobacter-containing MFCs generate high Coulombic efficiencies
(Call et al., 2009) but require punctilious anaerobic conditions, limiting their
applicability. Conversely, Shewanella-containing MFCs can be operated with air-exposed
cultures. Shewanella sp. respire a wide range of inorganic and organic compounds
through many mechanisms, including the use of mediators, to facilitate electron transfer
outside the cell membrane (Schröder, 2007). Therefore, Shewanella sp. can reduce solid
176

substrates through indirect mechanisms, unlike Geobacter sp. which require direct
contact to the electrode surface (Lovley, 2006). Shewanella sp. are an attractive
bacterium for MFCs because they can biosynthesize redox mediators and operate under
diverse environmental conditions (Fredrickson et al., 2008). Systems relying on artificial
mediators to decrease the overpotential for electron transfer are limited by irreversible
deactivation and cost due to the periodic addition of the mediators. However, bacterial
systems that generate their own mediators for extracellular electron transfer are feasible
for autonomous power sources.

Shewanella oneidensis MR-1 is the archetype for the genus and is frequently used as a
model in MFCs. S. oneidensis MR-1 is a fresh water microbe and was initially isolated
from Lake Oneida, New York in the early 1980s (Myers and Nealson, 1988). Research in
both fundamental electron transfer pathways and systems analysis has been explored but
unfortunately, the bacterium is unable to survive in high salinity environments
(Fredrickson et al., 2008). Interest in generating electricity from marine environments
necessitates the use of marine strains of Shewanella for electricity production within
MFCs. Recently, a study has shown that Shewanella marisflavi EP1 is capable of
generating power at a high ionic strength (up to 8% NaCl) but only lactate was used as a
carbon source (Huang et al., 2010). Shewanella japonica is a marine microbe isolated
from mussels in the Sea of Japan (Ivanova et al., 2001). In addition to metabolizing agar,
S. japonica utilizes D-galactose, D-fructose, glucose, and sucrose in growth experiments
177

suggesting that these carbon sources could be converted to electricity, provided these
metabolic pathways are coupled to extracellular electron transfer processes.

One important advance that could increase the competitiveness of MFCs in energy
production applications would be the direct conversion of polysaccharides into
electricity. Only 2-3 strains of bacteria (as well some spontaneous mutants of these
strains) have been described in the literature that convert glucose directly to power within
MFCs (Chaudhuri and Lovley, 2003; Biffinger et al., 2008a,b; Zuo et al., 2008).
Enterobacter cloacae was the first native strain of bacteria shown to generate power from
cellulose in a MFC (Rezaei et al., 2009). Presently, di- and polysaccharide utilization can
only be accomplished by enhanced bacterial consortia (Ishii et al., 2008; Ren et al.,
2008). The agar-lytic capacity of S. japonica makes it a promising candidate for
generating power from polysaccharides.

This manuscript determines for the first time that S. japonica can convert di- and
polysaccharides into electrical current. The mechanism for electron transfer is probed as
well as the response of S. japonica to a carbon electrode surface in an operating MFC. An
analysis of all these data suggests that S. japonica could be used in MFCs utilizing
polysaccharides, but the strain may be limited by the presence of electron transfer
mediators not initially present in cultures.

178

C.3 Materials and methods
C.3.1 Culture media and chemicals. Stock solutions of sodium lactate (1.95 M), sodium
acetate (1.95 M), 1% cellobiose, sucrose (1 M), 2% starch, and sodium citrate (0.5 M)
were sterilized by autoclaving for 13 min at 121
o
C and adjusted to pH 7. A D-glucose (1
M) stock solution was sterilized with a 0.2 lm cellulose nitrate filter. Marine broth (MB)
and marine agar (MA) were used for liquid cultures and plates, respectively (Difco). The
solvent for each solution was Millipore 18 MX water.

C.3.2 Strains. S. japonica (ATCC: BAA-316) was grown from a single colony isolated
from MA inoculated from a –80
o
C glycerol stock culture. Stabs of S. japonica in MA
were used to inoculate liquid culture media in MFCs. Stabs were only used over a period
of one month. S. japonica was transferred to 50 mL of MB in a 125 mL flask and
incubated exposed to air at 27
o
C at 100 rpm. Experimental cultures from stabs and frozen
stocks were acclimated to growth in media by subculturing once after 30 h before being
used in miniature MFC and voltage based screening assay (VBSA) experiments.

C.3.3 VBSA construction and data acquisition. Dimensions and fabrication of the
VBSA were published previously (Biffinger et al., 2009). The diameter of each well was
0.8 cm with a depth of 1.3 cm. Fully assembled, each well contains a maximum of 600
lL. The anodes were constructed from a titanium metal sheet (active electrode area, 0.3 x
0.3 cm) coated with a conductive carbon ink. The carbon ink contained 30 mg carbon
black, 300 lL 2-propanol, 300 lL 5% Nafion Solution in water, and 2 mL of de-ionized
179

water. The cathode system was graphite paper in 50 mM potassium ferricyanide
(dissolved in 100 mM phosphate buffer at pH 7.0). Each experiment was completed using
a nine-well VBSA apparatus. Experiments with no addition of carbon source were
designated as blanks and were control experiments. Electrodes used for environmental
scanning electron microscopy (ESEM) imaging were removed after 100 h of operation.
Voltages were measured across a 100 kX resistor (in a nine-resistor bank made for
simultaneous voltage measurements) and were recorded with a high-resolution data
acquisition module (I/O tech, personal daq/54) every two minutes. The measured voltage
was converted to current using Ohm’s law (Voltage = Current* Resistance). Each set of
VBSA biofilm experiments were performed twice with each VBSA setup containing two
experiments per each carbon source.

C.3.4 Biofilm growth and staining. Overnight cultures of S. japonica and S. oneidensis
MR-1 were established by inoculating 5 mL MB or LB, respectively, from a frozen stock
in test tubes and were agitated in a cell culture roller. These cultures were diluted 100-
fold in the appropriate growth medium, and 1 mL aliquots were removed for static
incubation. After three days of static incubation at room temperature (20
o
C ± 1
o
C) for S.
japonica or 30
o
C for S. oneidensis MR-1, supernatants were discarded, and each tube was
stained with 1.25 mL 0.1% crystal violet for 30 min. Biofilms in tubes were rinsed to
remove excess stain, dried and imaged.

180

C.3.5 Miniature MFC setup and data collection. The general dimensions and setup for
the mini-MFC apparatus were described previously (Ringeisen, et al., 2006). Two
identical miniature MFC systems were operated simultaneously for each experiment. The
electrodes within the fuel cell chambers were low-density graphite felt (0.13 g,
Electrosynthesis Company, Lancaster, NY; 0.47 m2/g) and were connected with titanium
wires to an external load. The anode and cathode chambers were separated by Nafion-117
(The Fuel Cell Store). Membranes were pre-treated for 1 h each in hot de-ionized (DI)
water, 3% hydrogen peroxide, 1 M sulfuric acid, and DI water again. The anolyte and
catholyte were passed through the chambers at a flow rate of 1–2 mL/min using a
peristaltic pump. The catholyte for each fuel cell was a 50 mM potassium ferricyanide
solution in 100 mM phosphate buffer (pH 7.2) using uncoated graphite felt (GF)
electrodes. All fuel cells were run at 25 ± 1
o
C. Fuel cells were operated simultaneously,
inoculated from identical 50 mL cultures of S. japonica either with or without the
addition of riboflavin (1 lM). Riboflavin- exposed cultures were wrapped in aluminum
foil to limit ambient light exposure. Sucrose was added to a concentration of 2 mM in
each anode culture flask at 48 and 124 h into MFC operation. Samples were removed
periodically through each experiment for analysis by HPLC (Varian, Inc.) with a
refractive index detector. The mobile phase was a 0.005 M sulfuric acid solution and the
column was a PL Hi-Plex H+ ion exchange column, at 65
o
C, with a flow rate of 0.6
mL/min. Peaks for sucrose and acetate were calibrated using known standards.

181

C.3.6 Imaging S. japonica biofilms from the VBSA. Environmental scanning electron
microscopy (ESEM) of carbon surfaces on the titanium anodes was performed at the
Naval Research Laboratory, Stennis Space Center, MS (NRLSSC). Unattached biomass
was removed by washing each anode with three separate 1 mL aliquots of phosphate
buffered saline solution at the Naval Research Laboratory, Washington, DC (NRLDC).
Each anode was placed in 2 mL of 4% cacodylate buffered glutaraldehyde fixative (Ray,
et al., 1997) at NRLDC and fixed for at least 24 h at 4
o
C prior to shipment to NRLSSC
without further manipulations. ESEM imaging procedures were performed as previously
described (Biffinger, et al., 2009).

C.3.7 SEM and AFM imaging. A 125 mL Erlenmeyer flask containing 50 mL marine
broth was used to initiate a S. japonica culture from an agar stab. The culture was
incubated for 48 h at 25
o
C after which the 1 mL of cells was harvested by centrifugation
at 5K rpm for 3 min. The cell pellet was washed 5 times with 1 mL of distilled water. The
cells were resuspended with 1 mL distilled water and 2 µL was deposited onto a silicon
oxide wafer for imaging. SEM micrographs were collected using a LEO Supra 55
microscope using the in-lens detector and with the primary beam voltage set to 10 kV.
The primary beam voltage was chosen to give maximum contrast and reduced charging
of the sample. The sample prepared for SEM imaging was dehydrated in acetone. Atomic
force microscopy (AFM) images were collected using an Autoprobe CP Research AFM
equipped with a microlever contact mode tip (force constant of 0.01 N m
-1
), both
182

manufactured by Thermomicroscopes. The AFM was operated in contact mode with a 3
nN set point, 0.4 Hz scan rate, and 512 x 512 pixel resolution.

C.3.8 Identification of riboflavion from mini-MFC experiments. The HPLC analysis
of culture supernatants was described previously (Biffinger et al., 2008a,b). The
instrumentation included Agilent (1100 series) chromatography components (Santa
Clara, CA), a quaternary pump for mobile phase delivery and diode array detector (DAD)
for monitoring elutent and collecting UV–Vis spectra. An Altima Phenyl column (250
mm x 4.6 mm, 5 µm support; Alltech Assoc., Deerfield Ill) was used. The mobile phase
was formic acid (0.1%):methanol (70:30, 1 mL min
-1
). The eluent was monitored at 210
nm and absorbance was recorded for chromatograms. For the sample preparation, the
culture supernatants were first clarified using centrifugation to remove bacteria and
insoluble material. Following centrifugation, samples were concentrated using a C18
solid phase extraction (SPE) cartridge (100 mg scale, Suplelco, Bellefonte, PA) that had
been preconditioned using manufacturer’s recommendations. For SPE, 2 mL of culture
supernatant was passed through the resin, polar compounds were washed from the matrix
with an equal volume of water and then the non-polar molecules eluted with 0.5 mL
methanol, and then analyzed using HPLC.

C.3.9 Electrochemistry of supernatant for mediator expression. A subculture from a
four-day-old culture of S. japonica was grown for two days. The subculture was
centrifuged for 10 min (5500 g, 15
o
C) and the supernatant filtered through a 0.2 µm
polytetrafluoroethylene polytetrafluoroethylene (PTFE) filter prior to electrochemical
183

analysis. Blank MB samples were obtained from the same growth medium used for the S.
japonica culture. In a three-electrode electrochemical cell, a polished 1.6-mm diameter
polycrystalline gold electrode was used as the working electrode, Pt gauze was used as
the counter electrode and a Ag/AgCl reference electrode (Bioanalytical Systems)
completed the electrochemical cell. Potentials were corrected and reported versus the
reversible hydrogen electrode (RHE). The electrolytes (media) were continuously
sparged with either Ar or O
2
to maintain anoxic or aerobic conditions during
electrochemical background and electrocatalytic oxygen reduction measurements,
respectively. The working electrode was polished before each measurement using an
aqueous slurry of 0.05 lm alumina powder (Buehler) on a polishing cloth and sonicated in
DI water to remove residual alumina from the electrode surface. Cyclic voltammetric
scan rates were 100 mV/s.

C.4 Results
C.4.1 Imaging of S. japonica. Bacterial nanofilamentous appendages have been studied
for their role in electron transfer to solid substrates (Gorby et al., 2006; Reguera et al.,
2006), biofilm formation (Proft and Baker, 2009), or cellular communication (Jelsbak and
Sogaard-Andersen, 2003). Images of air exposed S. japonica were collected by either
SEM (Fig. C.1A and B) or AFM (Fig. C.1C) in contact mode. S. japonica is rod-shaped
(~1 µm in length) with multiple flagella and pilitype structures expressed around the cell
body. Generally, images of S. japonica compared to S. oneidensis MR-1 (data not shown)
prepared using identical protocols indicate that there are more nanofilamentous structures
184

expressed by S. japonica than S. oneidensis MR-1. Since pili are essential for biofilm
formation, it may be expected that S. japonica will form thick biofilms (Thormann et al.,
2004). Maximum lengths for flagella were between 10 and 15 µm while pili lengths were
between 3 and 4 µm. In addition to the filamentous appendages, S. japonica produced a
large sheath-like material (30–40 µm x 1 µm). Except for one example of S. oneidensis
MR-1 grown at 3
o
C (Abboud et al., 2005), sheath formation is not common for the
Shewanellacae family and the exact role of the structure is currently unknown.

Figure C.1: Images of S. japonica. SEM (A and B) and AFM (C) are depicted above.

C.4.2 Screening of nutrients for current production by S. japonica. The ability of a
bacterium to reduce inorganic oxides does not necessarily translate into current producing
capabilities (Richter et al., 2007). Current output from S. japonica was assessed using
185

monosaccharides (glucose, fructose), disaccharides (sucrose, cellobiose), carboxylic acids
(acetate, lactate, citrate), and a polysaccharide (starch) using a VBSA. Within this single
miniature modular platform, nine simultaneous MFCs could be operated against a single
cathode/reference electrode (Biffinger et al., 2009). For the 9- well VBSA experiments,
the amount of biofilm formation was determined at the end of each screening experiment
as well as the current generated from air-exposed anodes with time (data not shown).

Anodes containing S. japonica in the absence of riboflavin and presence of 1 µM
riboflavin were imaged by ESEM (data not shown). Fig. C.2A and C.2B presents the
VBSA current versus time data for each carbon source. All experiments were shielded
from light because of the sensitivity of riboflavin to ambient light. Riboflavin was chosen
because of its potential role in mediating electron transfer in S. oneidensis (Biffinger et
al., 2008; Marsili et al., 2008; von Canstein et al., 2008; Ramasamy et al., 2009). The
maximum current generated from non-mediated S. japonica MFCs was less than 0.3 µA
throughout the duration of the experiment; while mediated electron transfer resulted in
defined current spikes after the addition of carbon sources with the largest current output
after the addition of sucrose (>2 µA) (Fig. C.2B). The results from the nonmediated
VBSA (Fig. C.2A) indicate that none of the carbon substrates were used efficiently by S.
japonica for current output after 100 h. On the other hand, in the presence of 1 µM
riboflavin S. japonica was capable of generating current immediately from sucrose,
glucose, cellobiose, and to a lesser extent, starch (Fig. C.2B).
186

Figure C.2: Voltage-based screening assay (VBSA) current output data. High-
throughput food screening experiments with no mediator additive (a) and with 1
µM riboflavin (b). Addition of 10mM carbon sources is indicated by vertical lines.

Even though Shewanella sp. can deliver electrons through direct contact as well as at a
distance through external mediators, biofilm formation will generate higher current
densities due to the proximity of the microbe to the electrode surface (Franks et al.,
2009). ESEM images of electrode surfaces from each of the active anode chambers of the
VBSA confirms that there was no significant biofilm formation in most of the
unmediated MFCs, while sparse biofilms were observed with glucose-exposed S.
japonica (data not shown). There was an overall increase in cellular attachment to the
electrode surfaces when S. japonica was exposed to riboflavin with the following carbon
sources: acetate, cellobiose, and fructose. However, the opposite was true for glucose. In
addition, cellobiose and fructose exposure resulted in the formation of a sheath type
structure. Biofilm formation did not result in power output in the case of acetate,
indicating that S. japonica cannot use acetate as a carbon substrate for current production.

187

The lack of biofilm formation on the electrode contradicted results observed in the
culture tube. S. japonica forms thicker biofilms in standard culture tubes than S.
oneidensis MR-1 (data not shown). During the VBSA experiments there was no
significant biofilm formation on the electrode surface suggesting that S. japonica is
presumably experiencing voltage based repulsion to the electrode surface in an
operational MFC. Previous work with S. oneidensis MR-1 confirms that thick biofilms
can be observed using our current electrode preparation protocol (Biffinger et al., 2009;
Bouhenni et al., 2010) thus S. japonica does not form significant biofilms on active
electrodes. Therefore, under these particular conditions, the increase in pili expression, as
seen from the micrographs, does not lead to increased biofilm formation in an operational
MFC and thus may not play a direct role in electron transfer.

C.4.3 Utilization of sucrose for electricity output in the mini-MFC. The necessity for
the addition of homogenous organic redox active mediators (i.e., riboflavin) to produce
current responses is both promising and concerning. S. japonica utilized simple
carbohydrates, and both di- and polysaccharides to generate current in a MFC in the
presence of riboflavin but no significant current was observed without riboflavin. To
study the maximum current generated from S. japonica exposed to sucrose with and
without the addition of riboflavin, two additional MFC experiments were performed in
identical miniature MFCs (mini-MFC). The mini-MFC is an active flow system that
generates high current densities (per volume) and shows good reproducibility from MFC
to MFC as well as negates the need for substantial biofilm formation to generate
188

maximum power density (Ringeisen et al., 2006; Biffinger et al., 2007; Biffinger et al.,
2008). Two mini-MFCs were operated simultaneously (with and without the addition of 1
lM riboflavin) using the same catholyte to eliminate variation derived from the cathodic
reaction between experiments. The current generated from each mini-MFC versus time is
shown in Fig. C.3 using an 820 X external resistor on each MFC. Additions of sucrose
were performed at the onset of the experiment, at 48 and 124 h.

Figure C.3: Current versus time chart for S. japonica in the mini-MFC. Current
produced without (light gray line) and with (dark gray line) the addition of 1 µM
riboflavin. Addition of 10mM sucrose is indicated by vertical lines.

189

The current from the mini-MFC with 1 µM riboflavin was consistently higher than the
MFC without riboflavin during the first 75 h of operation. After the second addition of
sucrose (48 h of operation), there was an immediate current response for the MFC with
riboflavin that is consistent with the results from the VBSA experiment (Fig. C.2B).
However, after 75 h a reproducible increase in current density (per volume) as high as
0.66 mA/cm
3
was observed from the mini-MFC without riboflavin. The third addition of
sucrose at 124 h confirmed that sucrose was being utilized by S. japonica for
extracellular electron transfer.

Two very interesting phenomena were observed from the mini- MFC experiments. The
first was that S. japonica could generate twice the current (maximum current: 0.08 mA
compared to 0.04 mA for riboflavin) from sucrose without addition of riboflavin after 75
h. The second observation was that riboflavin addition could inhibit current generation
even while mediating electron transfer between S. japonica and electrode. This result
suggests that riboflavin at 1 µM could inhibit microbial metabolism or could be a result
of the microbe metabolizing riboflavin as a carbon source instead of sucrose at this
concentration. However, the relationship between riboflavin concentration and bacterial
metabolism should be explored further as riboflavin is being used to study fundamental
mechanisms of extracellular electron transfer in cyclic voltammetric experiments.

To determine the source of the increase in current after 75 h in the unmediated MFC, 1
mL aliquots were removed from both MFCs throughout the experiment for analysis by
190

HPLC with a refractive index detector. Fig. C.4 shows data from these aliquots collected
during operation of the mini-MFC (shown in Fig. C.3) at 48, 51, and 110 h (prior to the
first sucrose addition, 3 and 62 h after sucrose addition, respectively). The inserts in Fig.
4 are the calculated concentration of sucrose and acetate from the same mini-MFC
experiment by HPLC. Substrate utilization and metabolite formation in MFC anolytes
differed in response to addition of riboflavin. The mini-MFC without riboflavin
consumed all sucrose by hour 110 while generating 0.25 mM of acetate as a by-product
(Fig. C.4A). No other by-products were observed. These results were significantly
different from those obtained from the mini-MFC containing riboflavin as acetate was not
produced. The rate of sucrose metabolism was eight times faster than the MFC containing
riboflavin (Fig. C.4B). In addition, the differential consumption of sucrose was observed
by the consistently lower current generated from the riboflavin containing mini-MFC.

Figure C.4: HPLC data for the consumption of sucrose of S. japonica in the mini-
MFC. MFCs run without (a) and with (b) 1 µM riboflavin from 48h (before sucrose
addition), 51 h (3 h after sucrose addition), and 110 h (62 h after sucrose addition) of
operation of mini-MFCs from Figure C.33. Figure inserts are calculated
concentration of sucrose and acetate from the HPLC data. Aliquot removal time
from mini-MFC (Figure C.3): 48 h (green trace), 51 h (red trace), 110 h (blue trace).

191

Results from the qualitative analysis of the MFC anolyte suggested that, like S.
oneidensis, S. japonica synthesizes a soluble redox mediator that could serve to facilitate
extracellular electron transfer or secondary metabolite for respiration. Comparison of the
culture supernatants and the MB medium revealed that riboflavin was identified in only
post-culture supernatants. No evidence of a similar compound was found in extracts from
the MB stock medium (data not shown). Absorbance maxima were observed at 222, 268,
370 and 448 nm by UV–Vis spectroscopy. The compound identified in culture
supernatants co-elutes with a riboflavin standard and the UV–Vis spectrum of the product
and the riboflavin standard shared identical features (data not shown). However, the
abnormally high current from the unmediated MFC compared to the mediated MFC
suggests that riboflavin is not the only redox active molecule in the culture supernatant.
Since there were sparse biofilms formed using all carbon sources from the VBSA
experiment and yet large increases in current after 75 h in the mini- MFC experiment,
external mediators were the primary mechanism for electron transfer by S. japonica.

C.4.4 Use of electroanalytical methods to compare mediation of oxygen reduction by
S. japonica-exposed supernatants and by riboflavin. Electroanalytical methods enable
determination of thermodynamics and kinetics of electrochemical reactions mediated by
freely diffusing redox species (Polcyn and Shain, 1966; Andrieux et al., 1980). Precise
determination of rate constants for reaction between the mediator and substrate and
electrochemical reduction/ oxidation potentials of both the mediator species and the
substrate require either large data sets which are fit to analytical expressions, or
192

comparison of smaller data sets to digitally simulated data. Comparison of the
voltammetric data generated in separate experiments will result in a qualitative
comparison between two mediators which catalyze the electrochemical oxidation or
reduction of the same chemical species. Using this sort of comparison, one can identify
whether the two mediators are the same species or different, provided that their catalytic
rates for the reaction of interest or their standard reduction potentials are sufficiently
different from one another.

Cyclic voltammetry was performed separately on S. japonica culture supernatant, sterile
MB, and MB spiked with different concentrations of riboflavin. Electrocatalytic
reduction of oxygen was used as a metric for redox activity. The simplest case of
mediated electrocatalysis by a freely diffusing mediator is described generally by the
following three equations (Polcyn and Shain, 1966; Andrieux et al., 1980):

A + ne
–
= B (1)
P + ne
–
= Q (2)
Q + A = P + B (3)

Eq. (1) is the direct reduction of the substrate of interest at the electrode surface. In the
present experiment, A represents molecular oxygen, and B is the reduced form of oxygen,
which will be hydrogen peroxide if n, the number of electrons in the reaction, is equal to
two, and water if n = 4. In Eq. (2), P is the oxidized form of the electrochemical mediator
and Q is the reduced form of the same, and the equation represents the reduction of P to
Q at the electrode surface. In the present experiments, P represents mediators generated
by S. japonica or an added mediator, such as riboflavin. Eq. (3) represents the
193

electrocatalytic event, where the reduced form of the mediator reacts with the substrate to
generate the reduced substrate and regenerate the oxidized mediator, represented in the
present experiment by the electrocatalytic reduction of oxygen by reduced forms of S.
japonica-derived mediators or added mediators.

Voltammograms for both the argon-saturated MB (voltammogram 1), and the argon-
saturated S. japonica growth medium (voltammogram 2), featured very small cathodic
peaks of ~2 µA at ~ –0.47 V (Fig. C.5A). The small peaks probably represent the
reduction of trace oxygen not entirely removed from the solution. The onset of hydrogen
evolution was evident around –0.65 V. The concentration of electrochemical mediator
was below the detection limit in the CV experiment in the absence of oxygen, and the
voltammograms from marine broth and culture supernatant were indistinguishable.
194

Figure C.5: Cyclic voltammograms of uninoculated and filtered S. japonica growth
medium. Unmodified (a) uninoculated control solution (curves 1 and 3, solid lines)
and filtered S. japonica growth medium (curves 2 and 4, dashed lines) under both
argon- (curves 1 and 2) and oxygen-saturated (curves 3 and 4). Scan rate: 100 mV/s.
Solutions with the addition (b): oxygen-saturated marine broth control solution
(curve 1, solid line), filtered S. japonica growth medium (curve 4, solid line), marine
broth control solution with 1 µM riboflavin (curve 2, dashed line), marine broth
control colution with 1 µM riboflavin (curve 2, dashed line), and acellular marine
broth control solution with 5 µM riboflavin (curve 3, dotted line). Scan rate: 100
mV/s.
195

Under oxygen-saturated conditions, oxygen reduction was evident in both media. In MB,
the oxygen reduction wave featured a cathodic peak at ~ –0.59 V (~ –28 µA amplitude),
and represents direct reduction of oxygen at the gold electrode (voltammogram 3). In the
S. japonica growth medium, the oxygen reduction wave was shifted positively by ~ 0.13
V, as evidenced by the ~ –28 µA cathodic peak at ~ –0.46 V, indicating the presence of a
biogenerated redox mediator (voltammogram 4). Fig. C.5B compares the mediated
electrocatalysis of oxygen reduction in S. japonica growth medium and in MB controls
spiked with riboflavin to concentrations of 0–5 µM. Riboflavin supplements caused
diminution of the cathodic peak derived from direct oxygen reduction at the electrode
(voltammograms 2 and 3) compared to the unmediated case (voltammogram 1) and the
onset of oxygen reduction at more positive potentials. When high riboflavin
concentrations (5 µM) were used, the electrocatalytic wave started to take the form of a
cathodic peak around –0.2 V (voltammogram 3).

The onset potential and the general shape of the voltammograms for electrocatalysis of
oxygen reduction in the riboflavinspiked controls, when compared to the voltammograms
for electrocatalytic oxygen reduction in the S. japonica growth medium (voltammogram
4), are very different. While a detailed electroanalytical determination of the kinetic and
thermodynamic parameters for the mediator present in the S. japonica growth medium is
beyond the scope of the present study, the data imply that both the reduction potential for
the mediator and the rate constants for the electron transfer between the mediator and
oxygen are different from those for riboflavin (Andrieux et al., 1980). It is possible that
196

some variant of riboflavin or collection of other organic redox active compounds or
proteins was synthesized by S. japonica and served as a redox mediator in its metabolic
pathway. Alternately, a different mediator altogether may be responsible for the
electrocatalysis of oxygen reduction. It is clear some sort of mediator was generated by S.
japonica, enabling the reduction of oxygen in the metabolic pathway of S. japonica and
the function of S. japonica in the MFC.

C.5 Discussion
The ability of S. japonica to use a diverse range of carbon sources (from
monosaccharides to sucrose to agar) indicates that it may have great promise in MFC for
marine environments. The reported data conclude that sucrose can be utilized for power
production from S. japonica. Additionally, riboflavin was observed in culture
supernatants and corresponded with significant increases in the current generated from a
MFC but may not be the only mediator present for mediating electron transfer from S.
japonica to carbon electrodes. Moreover, there was not a clear connection between
biofilm formation and current output suggesting that electron transfer from the bacterium
to the electrode surface was solely through mediated mechanisms not direct electrode
contact.

197

Appendix D: Recipes and Strain List
D. 1 Minimal Medium
Chemical Description FW g/L Formula Conc.(mM)
PIPES buffer 302.4 15.1 C8H18N2O6S2 50
Sodium hydroxide 40 3 NaOH

Ammonium chloride 53.49 1.5 NH4Cl 28.04
Potassium chloride 74.55 0.1 KCl 1.34
Sodium phosphate monobasic 138 0.6 NaH2PO4 H2O 4.35
Sodium chloride 5.844 5.8 NaCl 100
Mineral solution, 100X stock

10mL

see below
Vitamins solution, 100X stock

10mL

see below
Amino acid solution, 100X stock

10mL

see below
Sodium lactate 60% (w/w) syrup
98% pure, d=1.3g/mL, = 7M 112.1 2.54mL C3H5O3Na 18

Vitamin solution

Conc. nM
biotin (d-biotin) 244.3 0.002 C10H16N2O3S 81.87
folic acid 441.1 0.002 C19H19N7O6 45.34
pyridoxine HCl 205.6 0.01 C8H12ClNO3 486.38
riboflavin 376.4 0.005 C17H20N4O6 132.84
thiamine HCl 1.0 H2O 355.3 0.005 C18H18Cl2NOS 140.73
nicotinic acid 123.1 0.005 C6H5NO2 406.17
d-panthothenic acid, hemicalcium salt 238.3 0.005 C9H16NO5 1/2Ca 209.82
B12 1355.4 0.001
C63H88CoN14O14
P 0.74
p-aminobenzoic acid 137.13 0.005 C7H7NO2 364.62
thiotic acid 206.3 0.005 C8H14O2S2 242.37

Mineral solution

Conc. uM
nitrilotriacetic acid
(dissolve with NaOH to pH 8) 191.1 1.5 C6H9NO3 78.49
magnesium sulfate heptahydrate 246.48 3 MgSO4 7H2O 121.71
manganese sulfate monohydrate 169.02 0.5 MnSO4 H2O 29.58
sodium chloride 58.44 1 NaCl 171.12
ferrous sulfate helptahydrate 277.91 0.1 FeSO4 7H2O 3.6
calcium chloride dihydrate 146.99 0.1 CaCl2 2H2O 6.8
cobalt chloride hexahydrate 237.93 0.1 CoCl2 6H2O 4.2
zinc chloride 136.28 0.13 ZnCl2 9.54
cupric sulfate pentahydrate 249.68 0.01 CuCO4 5H2O 0.4
aluminum potassium disulfate
dodecahydrate 474.38 0.01 AlK(SO4)2 12H2O 0.21
boric acid 61.83 0.01 H3BO3 1.62
sodium molybdate dihydrate 241.95 0.025 NaMoO4 2H2O 1.03
nickel chloride hexahydrate 237.6 0.024 NiCl2 6H2O 1.01
198

sodium tungstate 329.86 0.025 Na2WO4 2H2O 0.76

Amino acid solution

Conc. mg/L
L-glutamic acid 2

2
L-arginine 2

2
DL-serine 2

2

D.2 Strain List
Names Species Anti
bioti
c
Growth Conditions SFS #
MR1 S. oneidensis MR-1 0001
k1 S. oneidensis MR-1 kan 0002
r1 S. oneidensis MR-1 rif 0003
aeLB k10a S. oneidensis MR-1 kan aerobic LB 10 days 0004
aeLB k10b S. oneidensis MR-1 kan aerobic LB 10 days 0005
aeLB k10c S. oneidensis MR-1 kan aerobic LB 10 days 0006
aeLB k20a S. oneidensis MR-1 kan aerobic LB 20 days 0007
aeLB k20b S. oneidensis MR-1 kan aerobic LB 20 days 0008
aeLB k20c S. oneidensis MR-1 kan aerobic LB 20 days 0009
aeLB k30a S. oneidensis MR-1 kan aerobic LB 30 days 0010
aeLB k30b S. oneidensis MR-1 kan aerobic LB 30 days 0011
aeLB k30c S. oneidensis MR-1 kan aerobic LB 30 days 0012
aeLB r10a S. oneidensis MR-1 rif aerobic LB 10 days 0013
aeLB r10b S. oneidensis MR-1 rif aerobic LB 10 days 0014
aeLB r10c S. oneidensis MR-1 rif aerobic LB 10 days 0015
aeLB r20a S. oneidensis MR-1 rif aerobic LB 20 days 0016
aeLB r20b S. oneidensis MR-1 rif aerobic LB 20 days 0017
aeLB r20c S. oneidensis MR-1 rif aerobic LB 20 days 0018
aeLB r30a S. oneidensis MR-1 rif aerobic LB 30 days 0019
aeLB r30b S. oneidensis MR-1 rif aerobic LB 30 days 0020
aeLB r30c S. oneidensis MR-1 rif aerobic LB 30 days 0021
anLB k10a S. oneidensis MR-1 kan anaerobic LB 10 days 0022
anLB k10b S. oneidensis MR-1 kan anaerobic LB 10 days 0023
anLB k10c S. oneidensis MR-1 kan anaerobic LB 10 days 0024
anLB k20a S. oneidensis MR-1 kan anaerobic LB 10 days 0025
anLB k20b S. oneidensis MR-1 kan anaerobic LB 10 days 0026
anLB k20c S. oneidensis MR-1 kan anaerobic LB 10 days 0027
199

anLB k30a S. oneidensis MR-1 kan anaerobic LB 10 days 0028
anLB k30b S. oneidensis MR-1 kan anaerobic LB 10 days 0029
anLB k30c S. oneidensis MR-1 kan anaerobic LB 10 days 0030
anLB r10a S. oneidensis MR-1 rif anaerobic LB 10 days 0031
anLB r10b S. oneidensis MR-1 rif anaerobic LB 10 days 0032
anLB r10c S. oneidensis MR-1 rif anaerobic LB 10 days 0033
anLB r20a S. oneidensis MR-1 rif anaerobic LB 10 days 0034
anLB r20b S. oneidensis MR-1 rif anaerobic LB 10 days 0035
anLB r20c S. oneidensis MR-1 rif anaerobic LB 10 days 0036
anLB r30a S. oneidensis MR-1 rif anaerobic LB 10 days 0037
anLB r30b S. oneidensis MR-1 rif anaerobic LB 10 days 0038
anLB r30c S. oneidensis MR-1 rif anaerobic LB 10 days 0039
aeMM k10a S. oneidensis MR-1 kan aerobic MM 10 days 0040
aeMM k10b S. oneidensis MR-1 kan aerobic MM 10 days 0041
aeMM k10c S. oneidensis MR-1 kan aerobic MM 10 days 0042
aeMM k20a S. oneidensis MR-1 kan aerobic MM 20 days 0043
aeMM k20b S. oneidensis MR-1 kan aerobic MM 20 days 0044
aeMM k20c S. oneidensis MR-1 kan aerobic MM 20 days 0045
aeMM k30a S. oneidensis MR-1 kan aerobic MM 30 days 0046
aeMM k30b S. oneidensis MR-1 kan aerobic MM 30 days 0047
aeMM k30c S. oneidensis MR-1 kan aerobic MM 30 days 0048
aeMM r10a S. oneidensis MR-1 rif aerobic MM 10 days 0049
aeMM r10b S. oneidensis MR-1 rif aerobic MM 10 days 0050
aeMM r10c S. oneidensis MR-1 rif aerobic MM 10 days 0051
aeMM r20a S. oneidensis MR-1 rif aerobic MM 20 days 0052
aeMM r20b S. oneidensis MR-1 rif aerobic MM 20 days 0053
aeMM r20c S. oneidensis MR-1 rif aerobic MM 20 days 0054
aeMM r30a S. oneidensis MR-1 rif aerobic MM 30 days 0055
aeMMr30b S. oneidensis MR-1 rif aerobic MM 30 days 0056
aeMM r30c S. oneidensis MR-1 rif aerobic MM 30 days 0057
anMM k10a S. oneidensis MR-1 kan anaerobic MM 10 days 0058
anMM k10b S. oneidensis MR-1 kan anaerobic MM 10 days 0059
anMM k10c S. oneidensis MR-1 kan anaerobic MM 10 days 0060
anMM k20a S. oneidensis MR-1 kan anaerobic MM 20 days 0061
anMM k20b S. oneidensis MR-1 kan anaerobic MM 20 days 0062
anMM k20c S. oneidensis MR-1 kan anaerobic MM 20 days 0063
anMM k30a S. oneidensis MR-1 kan anaerobic MM 30 days 0064
anMM k30b S. oneidensis MR-1 kan anaerobic MM 30 days 0065
200

anMM k30c S. oneidensis MR-1 kan anaerobic MM 30 days 0066
anMM r10a S. oneidensis MR-1 kan anaerobic MM 10 days 0067
anMMr10b S. oneidensis MR-1 rif anaerobic MM 10 days 0068
anMM r10c S. oneidensis MR-1 rif anaerobic MM 10 days 0069
anMM r20a S. oneidensis MR-1 rif anaerobic MM 20 days 0070
anMM r20b S. oneidensis MR-1 rif anaerobic MM 20 days 0071
anMM r20c S. oneidensis MR-1 rif anaerobic MM 20 days 0072
anMM r30a S. oneidensis MR-1 rif anaerobic MM 30 days 0073
anMM r30b S. oneidensis MR-1 rif anaerobic MM 30 days 0074
anMM r30c S. oneidensis MR-1 rif anaerobic MM 30 days 0075
bio k10a S. oneidensis MR-1 kan biofilm 10 days 0076
bio k10b S. oneidensis MR-1 kan biofilm 10 days 0077
bio k10c S. oneidensis MR-1 kan biofilm B 10 days 0078
bio k20a S. oneidensis MR-1 kan biofilm B 20 days 0079
bio k20b S. oneidensis MR-1 kan biofilm B 20 days 0080
bio k20c S. oneidensis MR-1 kan biofilm B 20 days 0081
bio k30a S. oneidensis MR-1 kan biofilm 30 days 0082
bio k30b S. oneidensis MR-1 kan biofilm B 30 days 0083
bio k30c S. oneidensis MR-1 kan biofilm B 30 days 0084
bio r10a S. oneidensis MR-1 rif biofilm 10 days 0085
bio r10b S. oneidensis MR-1 rif biofilm 10 days 0086
bio r10c S. oneidensis MR-1 rif biofilm 10 days 0087
bio r20a S. oneidensis MR-1 rif biofilm 20 days 0088
bio r20b S. oneidensis MR-1 rif biofilm 20 days 0089
bio r20c S. oneidensis MR-1 rif biofilm 20 days 0090
bio r30a S. oneidensis MR-1 rif biofilm 30 days 0091
bio r30b S. oneidensis MR-1 rif biofilm 30 days 0092
bio r30c S. oneidensis MR-1 rif biofilm 30 days 0093
k3x10a S. oneidensis MR-1 kan biofilm 3x10 days 0094
k3x10b S. oneidensis MR-1 kan biofilm 3x10 days 0095
k3x10c S. oneidensis MR-1 kan biofilm 3x10 days 0096
r3x10a S. oneidensis MR-1 rif biofilm 3x10 days 0097
r3x10b S. oneidensis MR-1 rif biofilm 3x10 days 0098
r3x10c S. oneidensis MR-1 rif biofilm 3x10 days 0099
k3x10a+20 S. oneidensis MR-1 kan biofilm 3x10+20 days 0100
k3x10b+20 S. oneidensis MR-1 kan biofilm 3x10+20 days 0101
k3x10c+20 S. oneidensis MR-1 kan biofilm 3x10+20 days 0102
r3x10a+20 S. oneidensis MR-1 rif biofilm 3x10+20 days 0103
r3x10b+20 S. oneidensis MR-1 rif biofilm 3x10+20 days 0104
r3x10c+20 S. oneidensis MR-1 rif biofilm 3x10+20 days 0105
aeLB k10a-1 S. oneidensis MR-1 kan clone 1 from aeLB k10a 0106
201

aeLB k10a-2 S. oneidensis MR-1 kan clone 2 from aeLB k10a 0107
aeLB k10a-3 S. oneidensis MR-1 kan clone 3 from aeLB k10a 0108
aeLB k10a-4 S. oneidensis MR-1 kan clone 4 from aeLB k10a 0109
aeLB k10a-5 S. oneidensis MR-1 kan clone 5 from aeLB k10a 0110
aeLB k10a-6 S. oneidensis MR-1 kan clone 6 from aeLB k10a 0111
aeLB k10a-7 S. oneidensis MR-1 kan clone 7 from aeLB k10a 0112
aeLB k10a-8 S. oneidensis MR-1 kan clone 8 from aeLB k10a 0113
aeLB k10c-1 S. oneidensis MR-1 kan clone 1 from aeLB k10c 0114
aeLB k10c-2 S. oneidensis MR-1 kan clone 2 from aeLB k10c 0115
aeLB k10c-3 S. oneidensis MR-1 kan clone 3 from aeLB k10c 0116
aeLB k10c-4 S. oneidensis MR-1 kan clone 4 from aeLB k10c 0117
aeLB k10c-5 S. oneidensis MR-1 kan clone 5 from aeLB k10c 0118
aeLB k10c-6 S. oneidensis MR-1 kan clone 6 from aeLB k10c 0119
aeLB k10c-7 S. oneidensis MR-1 kan clone 7 from aeLB k10c 0120
aeLB k10c-8 S. oneidensis MR-1 kan clone 8 from aeLB k10c 0121
aeLB k20a-1 S. oneidensis MR-1 kan clone 1 from aeLB k20a 0122
aeLB k20a-2 S. oneidensis MR-1 kan clone 2 from aeLB k20a 0123
aeLB k20a-3 S. oneidensis MR-1 kan clone 3 from aeLB k20a 0124
aeLB k20a-4 S. oneidensis MR-1 kan clone 4 from aeLB k20a 0125
aeLB k20a-5 S. oneidensis MR-1 kan clone 5 from aeLB k20a 0126
aeLB k20a-6 S. oneidensis MR-1 kan clone 6 from aeLB k20a 0127
aeLB k20a-7 S. oneidensis MR-1 kan clone 7 from aeLB k20a 0128
aeLB k20a-8 S. oneidensis MR-1 kan clone 8 from aeLB k20a 0129
aeLB k20b-1 S. oneidensis MR-1 kan clone 1 from aeLB k20b 0130
aeLB k20b-2 S. oneidensis MR-1 kan clone 2 from aeLB k20b 0131
aeLB k20b-3 S. oneidensis MR-1 kan clone 3 from aeLB k20b 0132
aeLB k20b-4 S. oneidensis MR-1 kan clone 4 from aeLB k20b 0133
aeLB k20b-5 S. oneidensis MR-1 kan clone 5 from aeLB k20b 0134
aeLB k20b-6 S. oneidensis MR-1 kan clone 6 from aeLB k20b 0135
aeLB k20b-7 S. oneidensis MR-1 kan clone 7 from aeLB k20b 0136
aeLB k20b-8 S. oneidensis MR-1 kan clone 8 from aeLB k20b 0137
aeLB k20c-1 S. oneidensis MR-1 kan clone 1 from aeLB k20c 0138
aeLB k20c-2 S. oneidensis MR-1 kan clone 2 from aeLB k20c 0139
aeLB k20c-3 S. oneidensis MR-1 kan clone 3 from aeLB k20c 0140
aeLB k20c-4 S. oneidensis MR-1 kan clone 4 from aeLB k20c 0141
aeLB k30a-1 S. oneidensis MR-1 kan clone 1 from aeLB k30a 0142
aeLB k30a-2 S. oneidensis MR-1 kan clone 2 from aeLB k30a 0143
aeLB k30a-3 S. oneidensis MR-1 kan clone 3 from aeLB k30a 0144
aeLB k30a-4 S. oneidensis MR-1 kan clone 4 from aeLB k30a 0145
aeLB k30a-5 S. oneidensis MR-1 kan clone 5 from aeLB k30a 0146
aeLB k30a-6 S. oneidensis MR-1 kan clone 6 from aeLB k30a 0147
aeLB k30a-7 S. oneidensis MR-1 kan clone 7 from aeLB k30a 0148
aeLB k30a-8 S. oneidensis MR-1 kan clone 8 from aeLB k30a 0149
202

aeLB k30b-1 S. oneidensis MR-1 kan clone 1 from aeLB k30b 0150
aeLB k30b-2 S. oneidensis MR-1 kan clone 2 from aeLB k30b 0151
aeLB k30b-3 S. oneidensis MR-1 kan clone 3 from aeLB k30b 0152
aeLB k30b-4 S. oneidensis MR-1 kan clone 4 from aeLB k30b 0153
aeLB k30b-5 S. oneidensis MR-1 kan clone 5 from aeLB k30b 0154
aeLB k30b-6 S. oneidensis MR-1 kan clone 6 from aeLB k30b 0155
aeLB k30b-7 S. oneidensis MR-1 kan clone 7 from aeLB k30b 0156
aeLB k30b-8 S. oneidensis MR-1 kan clone 8 from aeLB k30b 0157
aeLB k30c-1 S. oneidensis MR-1 kan clone 1 from aeLB k30c 0158
aeLB k30c-2 S. oneidensis MR-1 kan clone 2 from aeLB k30c 0159
aeLB k30c-3 S. oneidensis MR-1 kan clone 3 from aeLB k30c 0160
aeLB k30c-4 S. oneidensis MR-1 kan clone 4 from aeLB k30c 0161
aeLB k30c-5 S. oneidensis MR-1 kan clone 5 from aeLB k30c 0162
aeLB k30c-6 S. oneidensis MR-1 kan clone 6 from aeLB k30c 0163
aeLB k30c-7 S. oneidensis MR-1 kan clone 7 from aeLB k30c 0164
aeLB k30c-8 S. oneidensis MR-1 kan clone 8 from aeLB k30c 0165
aeLB r10b-1 S. oneidensis MR-1 kan clone 1 from aeLB r10b 0166
aeLB r10b-2 S. oneidensis MR-1 kan clone 2 from aeLB r10b 0167
aeLB r10b-3 S. oneidensis MR-1 kan clone 3 from aeLB r10b 0168
aeLB r10b-4 S. oneidensis MR-1 kan clone 4 from aeLB r10b 0169
aeLB k30c-6-1 S. oneidensis MR-1 kan clone 6 from aeLB k30c-6 0170
aeLB k30c-6-2 S. oneidensis MR-1 kan clone 6 from aeLB k30c-6 0171
aeLB k30c-7-1 S. oneidensis MR-1 kan clone 1 from aeLB k30c-7 0172
aeLB k30c-7-2 S. oneidensis MR-1 kan clone 2 from aeLB k30c-7 0173
aeLB k30c-7-3 S. oneidensis MR-1 kan clone 3 from aeLB k30c-7 0174
aeLB r10b-2-1 S. oneidensis MR-1 kan clone 3 from aeLB r10b-2 0175
aeLB r10b-2-2 S. oneidensis MR-1 kan clone 2 from aeLB r10b-2 0176
bio k20a-1 S. oneidensis MR-1 kan clone 1 from bio k20a 0177
bio k20a-2 S. oneidensis MR-1 kan clone 2 from bio k20a 0178
bio k20a-3 S. oneidensis MR-1 kan clone 3 from bio k20a 0179
bio k20a-4 S. oneidensis MR-1 kan clone 4 from bio k20a 0180
bio k20a-5 S. oneidensis MR-1 kan clone 5 from bio k20a 0181
bio k20a-6 S. oneidensis MR-1 kan clone 6 from bio k20a 0182
bio k20a-7 S. oneidensis MR-1 kan clone 7 from bio k20a 0183
bio k20a-8 S. oneidensis MR-1 kan clone 8 from bio k20a 0184
bio k20b-1 S. oneidensis MR-1 kan clone 1 from bio k20b 0185
bio k20b-2 S. oneidensis MR-1 kan clone 2 from bio k20b 0186
bio k20b-3 S. oneidensis MR-1 kan clone 3 from bio k20b 0187
bio k20b-4 S. oneidensis MR-1 kan clone 4 from bio k20b 0188
bio k20b-5 S. oneidensis MR-1 kan clone 5 from bio k20b 0189
bio k20b-6 S. oneidensis MR-1 kan clone 6 from bio k20b 0190
bio k20b-7 S. oneidensis MR-1 kan clone 7 from bio k20b 0191
bio k20b-8 S. oneidensis MR-1 kan clone 8 from bio k20b 0192
203

bio k20c-1 S. oneidensis MR-1 kan clone 1 from bio k20c 0193
bio k20c-2 S. oneidensis MR-1 kan clone 2 from bio k20c 0194
bio k20c-3 S. oneidensis MR-1 kan clone 3 from bio k20c 0195
bio k20c-4 S. oneidensis MR-1 kan clone 4 from bio k20c 0196
bio k20c-5 S. oneidensis MR-1 kan clone 5 from bio k20c 0197
bio k20c-6 S. oneidensis MR-1 kan clone 6 from bio k20c 0198
bio k20c-7 S. oneidensis MR-1 kan clone 7 from bio k20c 0199
bio k20c-8 S. oneidensis MR-1 kan clone 8 from bio k20c 0200
bio r20a-1 S. oneidensis MR-1 rif clone 1 from bio r20a 0201
bio r20a-2 S. oneidensis MR-1 rif clone 2 from bio r20a 0202
bio r20a-3 S. oneidensis MR-1 rif clone 3 from bio r20a 0203
bio r20a-4 S. oneidensis MR-1 rif clone 4 from bio r20a 0204
bio r20a-5 S. oneidensis MR-1 rif clone 5 from bio r20a 0205
bio r20a-6 S. oneidensis MR-1 rif clone 6 from bio r20a 0206
bio r20a-7 S. oneidensis MR-1 rif clone 7 from bio r20a 0207
bio r20a-8 S. oneidensis MR-1 rif clone 8 from bio r20a 0208
bio r20b-1 S. oneidensis MR-1 rif clone 1 from bio r20b 0209
bio r20b-2 S. oneidensis MR-1 rif clone 2 from bio r20b 0210
bio r20b-3 S. oneidensis MR-1 rif clone 3 from bio r20b 0211
bio r20b-4 S. oneidensis MR-1 rif clone 4 from bio r20b 0212
bio r20b-5 S. oneidensis MR-1 rif clone 5 from bio r20b 0213
bio r20b-6 S. oneidensis MR-1 rif clone 6 from bio r20b 0214
bio r20b-7 S. oneidensis MR-1 rif clone 7 from bio r20b 0215
bio r20b-8 S. oneidensis MR-1 rif clone 8 from bio r20b 0216
bio r20c-1 S. oneidensis MR-1 rif clone 1 from bio r20c 0217
bio r20c-2 S. oneidensis MR-1 rif clone 2 from bio r20c 0218
bio r20c-3 S. oneidensis MR-1 rif clone 3 from bio r20c 0219
bio r20c-4 S. oneidensis MR-1 rif clone 4 from bio r20c 0220
bio r20c-5 S. oneidensis MR-1 rif clone 5 from bio r20c 0221
bio r20c-6 S. oneidensis MR-1 rif clone 6 from bio r20c 0222
bio r20c-7 S. oneidensis MR-1 rif clone 7 from bio r20c 0223
bio r20c-8 S. oneidensis MR-1 rif clone 8 from bio r20c 0224
CN32 S. putrefaciens CN32 0225
SP200 S. putrefaciens SP200 0226
OS217 S. denitrificans OS217 0227
PV4 Shewanella PV4 0228
MR4 S. oneidensis MR-4 0229
MR7 S. oneidensis MR-7 0230
W3-18-1 ShewanellaW3-18-1 0231
SB2B S. amazonensis SB2B 0232
ana S. ana 0233
Sjap S. japonica 0234

Abstract (if available)

Abstract Microbial fuel cells are batteries in which microorganisms catalyze the conversion of organic fuel (such as lactate) into protons and electrons that power a resistor (e. g., a light bulb) before reducing the terminal electron acceptor (e. g., oxygen is reduced to water). Great improvements in power production and efficiency have been made by engineering inorganic components, such as the electrodes themselves, to be more efficiently utilized by fuel cell-inhabiting organisms. However, other avenues for improvement may exist, that is, engineering the fuel cell-inhabiting organisms themselves. We hypothesized that Shewanella oneidensis MR-1, a model organism used for studying microbial fuel cells, could be shown to evolve under physiological conditions which mimic those found in microbial fuel cells. These physiological conditions include the planktonic lifestyle, the biofilm lifestyle, and transient association between the two – that is, those cells that rapidly detach from and reattach to the biofilm. ❧ Here we show the Growth Advantage in Stationary Phase (GASP) phenotype conferred by aging cells planktonically in conditions of abundant electron donor and acceptor, as well as conditions of either electron donor or acceptor limitation. In general, the longer cells are aged planktonically, the greater their advantage when competing in a similar environment. A GASP-like phenotype is also conferred by aging cells in a biofilm for 10 days, though aging cells continuously within a biofilm for 20 days resulted in a competitive disadvantage. To better understand cells that are transiently associated with both lifestyles, we observed the rapid formation of, detachment from and reattachment to biofilms. Biofilm spontaneously form both where oxygen is plentiful and where it is scarce. Oxygen-replete biofilms and oxygen-poor biofilms respond to different supplementary amino acids. Response to amino acid supplementation also varies according to the developmental stage of these biofilms. These data may offer insight into the biology of microbial fuel cells, as well as guidance for physiological treatments and methods of directed evolution that will improve microbial fuel cell performance.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Evolution of E. coli in unstructured and structured environments

PDF

From fuel cells to single cells: electrochemical measurements of direct electron transfer at microbial-electrode interfaces

PDF

Electron transfer capability and metabolic processes of the genus Shewanella with applications to the optimization of microbial fuel cells

PDF

Investigating microbial biofilm community mediated processes on surfaces: from single cell genomics to community meta-omics

PDF

Physiological roles and evolutionary implications of alternative DNA polymerases in Escherichia coli

PDF

Mechanisms of long-term survival of Vibrio harveyi and Escherichia coli: linking damage and senescence

PDF

Identification of a new bacterial sensing mechanism: characterization of bacterial insoluble electron acceptor sensing

PDF

From cables to biofilms: electronic and electrochemical characterization of electroactive microbial systems

PDF

From single molecules to bacterial nanowires: functional and dynamic imaging of the extracellular electron transfer network in Shewanella oneidensis MR-1

PDF

Bioremediation of selenite using biological and bioelectrical systems

PDF

Bacterial nanowires of Shewanella oneidensis MR-1: electron transport mechanism, composition, and role of multiheme cytochromes

PDF

Applications of advanced electrochemical techniques in the study of microbial fuel cells and corrosion protection by polymer coatings

PDF

Dynamics of Escherichia coli propagation, biofilm formation and evolution

PDF

Studies on direct oxidation formic acid fuel cells: advantages, limitations and potential

PDF

Genomic and physiological characterization of Escherichia coli evolving in long-term batch culture

PDF

The effect of microbial load and autophagy on drosophila immunity and life span

PDF

Electrocatalysts for direct liquid-feed fuel cells and advanced electrolytes for lithium-ion batteries

PDF

Design and modification of electrocatalysts for use in fuel cells and CO₂ reduction

PDF

Diversity and dynamics of giant kelp “seed-bank” microbiomes: Applications for the future of seaweed farming

PDF

Electrokinetic transport of Cr(VI) and integration with zero-valent iron nanoparticle and microbial fuel cell technologies for aquifer remediation

Asset Metadata

Creator Ribbens, Meghann Adrienne (author)

Core Title Survival and evolution of Shewanella oneidensis MR-1: applications for microbial fuel cells

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Molecular Biology

Publication Date 08/02/2012

Defense Date 08/01/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag amino acid,biofilm,Evolution,GASP,growth advantage in stationary phase,microbial fuel cells,OAI-PMH Harvest,S. oneidensis MR-1,Shewanella

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Finkel, Steven E. (committee chair), Berelson, William M. (committee member), Goodman, Steven D. (committee member), Nealson, Kenneth H. (committee member)

Creator Email maribbens@gmail.com,ribbens@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-81950

Unique identifier UC11289237

Identifier usctheses-c3-81950 (legacy record id)

Legacy Identifier etd-RibbensMeg-1107.pdf

Dmrecord 81950

Document Type Dissertation

Rights Ribbens, Meghann Adrienne

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA