THE EVOLUTION OF A KEY INNOVATION IN AN EXPERIMENTAL POPULATION OF
ESCHERICHIA COLI: A TALE OF OPPORTUNITY, CONTINGENCY, AND CO-OPTION
By
Zachary David Blount

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Microbiology and Molecular Genetics
2011

ABSTRACT
THE EVOLUTION OF A KEY INNOVATION IN AN EXPERIMENTAL POPULATION OF
ESCHERICHIA COLI: A TALE OF OPPORTUNITY, CONTINGENCY, AND COOPTION
By
Zachary David Blount
The importance of historical contingency in evolution has been extensively debated over
the last few decades, but direct empirical tests have been rare.

Twelve initially identical

populations of E. coli were founded in 1988 to investigate this issue. They have since evolved
for more than 50,000 generations in a glucose-limited medium that also contains a citrate.
However, the inability to use citrate as a carbon source under oxic conditions is a species+

defining trait of E. coli. A weakly Cit variant capable of aerobic citrate utilization finally
evolved in one population just prior to 31,500 generations. Shortly after 33,000 generations, the
+

population experienced a several-fold expansion as strongly Cit variants rose to numerical
+

dominance (but not fixation). The Cit trait was therefore a key innovation that increased both
population size and diversity by opening a previously unexploited ecological opportunity.
+

The long-delayed and unique evolution of the Cit innovation might be explained by two
+

possible hypotheses. First, evolution of the Cit function may have required an extremely rare
+

mutation. Alternately, the evolution of Cit may have been contingent upon one or more earlier
mutations that had accrued over the population’s history. I tested these hypotheses in a series of
experiments in which I “replayed” evolution from different points in the population’s history. I
+

12

observed no Cit mutants among 8.4 x 10

ancestral cells, nor among 9 x 10

12

cells from 60

clones sampled in the first 15,000 generations. However, I observed a significantly greater

+

tendency to evolve Cit among later clones. These results indicate that one or more earlier
+

+

mutations potentiated the evolution of Cit by increasing the rate of mutation to Cit to an
accessible, though still very low, level.

+

The evolution of the Cit

function was therefore

contingent on the particular history of the population in which it occurred.
+

I investigated the Cit innovation’s history and genetic basis by sequencing the genomes
of 29 clones isolated from the population at various time points. Analysis of these genomes
revealed that at least 3 distinct clades coexisted for more than 10,000 generations prior to the
+

innovation’s evolution. The Cit trait originated in one clade by a tandem duplication that
produced a new regulatory module in which a silent citrate transporter was placed under the
control of an aerobically-expressed promoter. Subsequent increases in the copy number of this
+

new regulatory module refined the initially weak Cit phenotype, leading to the population
+

expansion. The 3 clades varied in their propensity to evolve the novel Cit function, though
genotypes able to do so existed in all 3, implying that potentiation involved multiple mutations.
My findings demonstrate that historical contingency can significantly impact evolution,
even under the strictest of conditions.

Moreover, they suggest that contingency plays an
+

especially important role in the evolution of novel innovations that, like Cit , require prior
construction of a potentiating genetic background, and are thus not easily evolved by gradual,
cumulative selection. Contingency may therefore have profoundly shaped life’s evolution given
the importance of evolutionary novelties in the history of life. Finally, the genetic basis of the
Cit+ function illustrates the importance of promoter capture and altered gene regulation in
mediation

the

exaptation

events

that

often

underlie

evolutionary

innovations.

Copyright by
ZACHARY DAVID BLOUNT
2011

This dissertation is dedicated to the memory of my great-grandmother, Geneva Mae Sikes
(February 5, 1905 – January 27, 1997), my brother, Nicholas Andrew Blount (April 2, 1982 –
April 17, 2008), and my aunt, Carmen Brightwell Poland (October 31, 1938 – March 18, 2011).

v

ACKNOWLEDGEMENTS

I think I first got the notion that I wanted to get a PhD when I was maybe six or seven. I
don’t quite recall at this point. I do remember it was during a time when I had gotten interested
in chemistry, and decided I wanted to be a chemist when I grew up (and an astronaut). My
mother told me around then that it took a lot of education, in fact a PhD to work in an area of
science. So I decided I was going to get one, a choice that was reinforced by the discovery of the
reruns of “Doctor Who” that GPTV showed on Saturday nights. I kept up with that decision
over the years, though what I would become a doctor of changed. I lost interest in chemistry and
got interested in physics, and maintained that interest until college, when I discovered that
physics required an enormous amount of math that I found utterly baffling. I had developed a
th

major interest in biology in 7 grade, when I first discovered evolution, and I had done very well
in the subject all through school. When I took AP biology in high school, I can recall Ms. Pugh
suggesting that I consider a career in biology, which I disregarded because I wanted to do
physics. So, upon learning that physics wasn’t going to work out, into biology I went. Here I
am, almost thirty years after first deciding to get a PhD, finally getting one. Wow.
This dissertation marks two distinct milestones. First, it provides an interim (because it
really isn’t done) capstone to a research project that began when I first rotated through the Lenski
lab back in 2004. Second, and in a larger sense, it marks the achievement of that childhood
dream, and the end of a long formal education that began in 1982.

I have a lot of people and

institutions who have helped me along the way to this achievement, and to whom I owe a debt of
gratitude that I would like to recognize. I will recognize those specific to the first milestone first,

vi

and then move on to the longer list of those who have contributed to the whole long journey. I
don’t think there is a length restriction on this section, so I will ignore the music if it comes up.
I have great fortune to have had Dr. Richard Lenski as my mentor. He is not only an
astonishingly brilliant scientist whose contributions will echo in evolutionary biology for
generations to come, but also one of the kindest and most fundamentally decent human beings I
have ever known. I want thank him for taking a chance on me, for his guidance, encouragement,
and support. Moreover, as an advisor, he has that special talent of being able to grant enormous
freedom to explore and strike out, and then knowing just when to provide just the right nudge to
ensure one goes in the right direction. It is difficult to describe, but he has helped me to grow
and develop confidence as a researcher and scientist in ways I hadn’t imagined possible when I
first entered the lab. I also want to thank him for being willing to accept, if not encourage little
eccentric projects like draping the lab with plants, putting up signs in the windows, not to
mention stockpiling used Petri plates for building tottering towers of infectious waste.
I have also had the benefit of a wonderful research advisory committee, which really
helped me to clarify my ideas over the years. I would like to thank the members of that
committee, Dr. Terrence Marsh, Dr. Robert Pennock, Dr. Charles Ofria, and Dr. Thomas
Schmidt, for their wisdom and good counsel.
I had a wonderful group of collaborators without whom the research detailed in this
dissertation could not have been done, and I thank them for all of their help and work. Dr. Chris
Borland was overseeing the citrate project when I came into the lab as a rotation student. Her
initial work on Cit+ laid a great foundation for what I ended up doing. She also trained me in the
ways of the Lenski lab, got me going in my work, and taught me skills I still use. I wish I had
gotten to work with her longer. Dr. Sean Sleight helped me greatly in the initial genetic analysis

vii

+

of the Cit trait. Dr. Carla Davidson put an enormous amount of time and effort into some
critical expression experiments that were important to telling the story found in Chapter 3. She
also became a very dear friend helped me in many ways – I don’t know how I would have gotten
through the last few years without her willingness to listen and reassure me. As far as the
research goes, I owe the biggest thank you to Dr. Jeff Barrick. Jeff is really one of the most
amazing brilliant people I have ever met, and his humility and self-effacement only barely makes
tolerable the blindingly Apollonian incandescent shine of his intellect and talent. Jeff is the
bioinformatics genius responsible for turning the jumbo-enormous amounts of genomics data
that provides the core of Chapter 3 into something comprehensible by way of his customdesigned data mining Breseq platform (which I will forever simply refer to as simply “The
Gnomes”. Were it not for him, I would still be trying to make sense out of all those letters. His
work ethic and dedication to the pursuit of so many questions have also provided me with much
inspiration. As an addendum to this section, I also want to thank Caroline Turner, who is delving
into the ecological aspects of the citrate story, for future collaboration. I look forward to
working with her on the next phase of work, and I love that her research means I get answers to
big questions without my having to do the work! Finally, I want to thank Mark Kauth, an
undergraduate who came into the lab initially as one of Jeff’s helpers, but who will be staying on
as a graduate student. He has provided critical last-minute help on a number of aspects of the
work covered in Chapter 3, and will likely be doing work on some aspect of the citrate story
later. Thank you for your help so far; I look forward to working with you in the future.
In addition to research collaborators, my work would not have been possible without the
efforts of the Lenski lab support staff. They are the ones who made sure that I had all the
supplies that I needed, who cleaned up after me, and who were willing to pour incredible

viii

numbers of plates, sometimes with little notice. They are the backbone of the lab, and without
them, everything would screech to a halt. The core of this staff is Neerja Hajela, the lab’s
manager and resident goddess.

Neerja oversees everything with calm, patience, care, and

attention. My work would have been absolutely impossible had it not been for her. Moreover,
during my time in the lab, she has become one of the best friends I have ever had, and really my
second mother. She has always been there to encourage, to listen, to comfort, to counsel, and to
just talk and joke around with. I love you, Neerja, and I am so grateful to have had your help,
guidance, and friendship. Assisting Neerja in keeping the lab going have been a number truly
wonderful undergraduate helpers, including Brian Chernoff, Florence Emanajo, Michelle Mize,
and Camorrie Bradley. They are the ones who washed the dishes, ran errands, prepped most of
the media, and generally kept things going. Y’all are awesome. I couldn’t have done it without
you. Finally, I want to single out one of those undergraduate assistances, Marwa Adawe, for
special thanks. Marwa was helper during the largest and most frenetic of my experiments,
including the big plating experiments discussed in Chapter 2. She did incredible amounts of
work without complaint or delay. Amidst all that work, she showed a wonderful curiosity about
everything in the lab, and was wonderfully patient for some pretty long-winded explanations.
Over the years, she has become a dear and close friend who is very much the sister I never had.
Marwa, you are an incredible person with a great mind and greater heart, and I know you will go
far. I love you, and I am so grateful to have had your help, and to have even simply known you.
I want to thank our computer tech Brian Baer for keeping the computers running,
answering stupid questions, and being patient with a functional Luddite, not to mention being
willing to loan various books and videos for years on end. Brian has also been master of
everything electronic, taken pictures, set up time-lapse videos, helped in setting up presentations,

ix

and been an all around jack-of-all trades. He is also one of the kindest, most helpful people I
have every known, in addition to being a good friend, and a great person to talk to about books,
movie, and random nerdy topics. He’s definitely made the lab a more fun place to work.
My work has been made possible also by the generous funding of a number of agencies
and organizations. Through grants to Dr. Lenski, my work has been primarily funded by the
National Science Foundation (NSF), including through the BEACON Center, and the Defense
Advanced Research Projects Agency (DARPA). In my own name, I have been supported by an
MMG travel grant (2005), two summer fellowships funded by the Ecology, Evolutionary
Biology, and Behavior (EEBB) program (2006 and 2007), a Barnett Rosenberg Fellowship (2008
– 2009), an American Society for Microbiology Student Travel Award (2009), a Rudolf Hugh
Fellowship (2008), a Duvall Family Award (2009), and a Ronald M. and Sharon Rogowski
Fellowship (2011).
Beyond its boss and support staff, the Lenski lab has been a wonderful place to work.
This is a place where a student can do really amazing science, and really feel like one is a part of
something big. Not only is this home to the LTEE, a truly epic experiment, but it has been the
nurturing ground of so many luminaries of experimental evolution. And it still nurtures a diverse
array of future luminaries. Therefore, one of the great things about the lab is the opportunity to
work with brilliant people who are examining fascinating questions. These undergraduate,
graduate, and postdoctoral colleagues have provided wonderful resources of intellect and
support, and who are always there to provide a different perspective, to read over or discuss a
paper, or to answer a question or lend a hand. I have already mentioned a number of them who
ended up playing a major role in my work, and I would like to additionally recognize all those
others with whom I have had the honor of working: Dr. Kristina Hillesland, Dr. Christopher

x

Marx, Dr. Dusan Misevic, Dr. Elizabeth Ostrowski, Dr. Chris Strelioff, Dr. Robert Woods, Dr.
Gabe Yedid, James Dittmar, Devin Dobias, Nathan Johns, Rohan Maddamsetti, Daniel Mitchell,
Dr. Jeff Morris, Christian Orlic, Mike Wiser, and Luis Zaman. I want to particularly thank Brian
Wade and Justin Meyer for a good bit of discussion and friendship over the years.
I gave a poster at the 2005 ASM general conference in Atlanta. During my session, a
man came up, looked over the poster, and listened to me give my spiel. When I was done, and
had answered his questions, he asked me if I knew how lucky I was to be working in the
Department of Microbiology and Molecular Genetics at MSU. He named a number of the
prominent researchers in the department, and declared it the “Athens of Microbiology”. I have
always been struck by that interaction, and, indeed, I have had amazing fortune as a doctoral
student to have studied in MMG, where I have gotten a first rate graduate education and have
been able to work with incredible people while doing cutting-edge science in a vibrant
environment. Moreover, needed resources are always at hand. This is a remarkable place, and if
it is the Athens of microbiology, it is Athens with the resources of Rome. I want to thank Dr.
Walter Esselman, the chair of the department, and the entire department faculty for making this
such a great place to learn and work.
I also doubt a debt of gratitude to a number of people who have generally helped or
simply brightened my day during my time at MSU, including Stephanie Eichorst, Uri Levine,
Fan Yang, Rhia Leveque, Jenifer Mayrberger, and Subrena Jones.
Finally, I want to thank some people I have never met, but whose work has greatly
influenced me and my work. First, Aristotle invented biology and provides a model for clear
thinking. Had it not been for him, I would likely have been a historian. Charles Darwin has
provided a source of great intellectual delight, as well as being a model of not only how to be a

xi

good and creative scientist, but also of how to be a good human being. It is remarkable just how
wonderful and helpful his writings are a century and half on. And then there is Stephen Jay
Gould. His writings and ideas have provided the impetus and intellectual foundation of much of
my work, and I think it is correct to say that my research has been contingent upon his earlier
efforts. Moreover, his extensive body of delightful, if often frustrating, writings have given me a
great number of ideas, moments of puzzled silence, and hours of entertainment. I regret that I
will never get to meet him and talk about contingency; he left the world far too soon.
On the broader journey that began long before I came to MSU, I have another myriad to
thank. First and foremost, I must thank my family for supporting me in so many ways during my
life. I am so grateful for their help, and I know how fortunate I have been to have had such a
supportive family.

My parents, Donna and David Blount, taught me to be resilient and

independent, and always encouraged me to follow my dreams. Mom taught me to love learning
and reading, and was a great emotional support through the years. Dad taught me a solid work
ethic, the value of hard work, and the lesson that any job worth doing is worth doing well – no
matter who might know. Working for Dad was not always pleasant or easy, but it gave me the
foundation that has helped me succeed.

My brothers, Jonathan and Nicholas, were good

companions growing up (usually), and they gave me many hours of fun, adventure, and
frustration. Nick passed away too soon, but he was a kind and gentle soul. Jonathan has been a
model of dedication and discipline, and I am grateful to him for the advice and help he has given
me over the years. My great-grandmother, Geneva Sikes, who I knew only as “Greatmommie”,
was the light of my childhood – a kindly presence who was always there to encourage, to
comfort, to listen, and to tell me stories of her life. I have missed her greatly since she left this
world. I live every day hoping that I make her proud. I want to thank my maternal grandparents,

xii

Frankie and John Davis, for their care, love, and support over the years. They have always been
there for me, and I couldn’t have asked for a better pair of grandparents. I thank my Aunt
Carmen and Uncle Tom for being such an important and nurturing part of my childhood. As the
dedication indicates, Aunt Carmen died recently, and I will miss her. She was one of the most
magical people I have ever had the pleasure of knowing. And, come to think of it, she also
introduced me to Doctor Who. And I also thank my nieces, Jordan and Taylor, and my nephews,
Bailey, Christian, and Xavier for their entertainment value. I love you all.
I have had a small, but close circle of dear friends – Tim Clemons, Lee Goodson, Laurie
Schwartz, Crystal Wenzler, Yasemin Tulu, and Matt and Karen Williams – who have been
important parts of my life for years. Thanks for the encouragement, willingness to listen, and the
years of friendship and fun. I hope I can one day repay you for it all.
I owe a deep debt of gratitude to my long-time significant other, Dr. Rae Braudaway, for
her help, love, and support over the years. She has been integral to getting me through a difficult
decade. I love you, Raeny.
I want to thank my cats, Aggie and Phrossy, for their companionship and what I choose
to interpret as affection.
I have had many, many teachers over the years who helped to get me here. Among that
great number, there is a select group of teachers who played a special role in my education and
life, to whom I owe a special debt. Paula Gleeson, my kindergarten teacher, gave me a good
start. Mary Ann Watson was my second grade teacher, and she did a lot to correct damage
inflicted by a truly ghastly first grade experience.

She gave me love a love of learning,

introduced me to Mozart and Bach, and put me on a solid academic path. One day I will find
you and let you know just how much good you did for that scared little boy. Lynn Pugh was my

xiii

high school chemistry and AP biology teacher. She was kind enough to take a particular interest
in me, and, as I mentioned above, was the first to recommend that I consider becoming a
biologist. She seems to have been correct in that recommendation. She was also the first teacher
to formally introduce me to many of the fundamental concepts of evolution, and to devote
significant class time to the subject. Dr. Thomas Tornabene was my introductory microbiology
professor at Georgia Tech, and he introduced me to the wonders of the microbial world. He had
an enthusiasm for the subject that was infectious, and that feeling really took me. And all these
years later, I still believe in Santa Claustridium. Dr. Dennis Grogan was my advisor for my
master’s work at the University of Cincinnati, and he really showed me the ropes of formal
scientific research. I was rather grateful for his patience. Finally, I want to thank Dr. James
Miller for his Nature and Practice of Science course here at MSU. That class was hands down
the most valuable I have ever taken during my graduate career. I do not exaggerate when I say it
taught me how to think like a scientist. My work has been much better-structured and thought
out than it would have been without Dr. Miller’s help. Indeed, I chose to work with Chris
Borland on the citrate project because the prospectus she put together of a possible project
conformed to Platt’s strong inference ideas, which he taught me to so respect. I honestly don’t
understand why that class, or at least something like it, isn’t required of every graduate science
program.
Finally, to She of the clear gray eyes, I give thanks for inspiration. May You forever
guide my head and my heart.

xiv

In all things of Nature, there is something of the marvelous.
– Aristotle, The Parts of Animals (350 BCE?)

There is grandeur in this view of life, with its several powers, having been originally breathed
into a few forms or into one; and that, whilst this planet has gone cycling on according to the
fixed law of gravity, from so simple a beginning endless forms most beautiful and most
wonderful have been, and are being, evolved.
– Charles Darwin, On the Origin of Species (1859)

It is hard to realize that the living world as we know it is just one among many possibilities; that
its actual structure results from the history of the earth. Yet living organisms are historical
structures: literally creations of history. They represent, not a perfect product of engineering, but
a patchwork of odd sets pieced together when and where opportunities arose. For the
opportunism of natural selection is not simply a matter of indifference to the structure and
operation of its products. It reflects the very nature of a historical process full of contingency.
– François Jacob, “Evolution and Tinkering” (1977)

xv

TABLE OF CONTENTS

LIST OF TABLES……………………………………………………………………….....…...xix
LIST OF FIGURES………………………………………………………………………….…..xx
KEY TO SYMBOLS AND ABBREVIATIONS…………………………………………....…xxii
CHAPTER 1
HISTORICAL CONTINGENCY AND EVOLUTION………………………………………......1
A Brief History of the Neglect of History in Evolutionary Biology………………………1
Stephen Jay Gould and “Replaying the Tape of Life”………..…………………...4
Difficulties in Gould’s Notions of Contingency…………………………………8
Reconsideration of Historical Contingency……………..……………………………….12
Contingency and Human History……………………………………………...…12
A historical vignette…..………………………………………………….12
Counterfactual analysis of history……………………………………….13
Properties of a Historically Contingent System……………………………….…16
Evolution and Contingency……………………………………………………...……….22
Empirical Studies of Historical Contingency in Evolution…….……………...…………26
Macroevolutionary Studies...…………………………………………………….28
Evolution Experiments……………………………………………...……………35
The E. coli Long-term Evolution Experiment……………………….………….37
An Open Ecological Opportunity..……………………………..………..43
History of the Inclusion of Citrate in DM Medium…..………………….44
E. coli has the Potential to Evolve a Means of Exploitation the Citrate
Opportunity...………………………………………………...…………..49
+
Cit Evolves in the Ara–3 Population...………………………………… 50
Research Discussed in this Dissertation……..………………….………. 53
Epilogue…………………………………………………………………….……………55
Increased Concern for Historical Contingency Has Developed in Parallel Across Multiple
Disciplines, and in the American Culture at Large………………………..……………..55
References………………………………………………………………………………..62
CHAPTER 2
HISTORICAL CONTINGENCY AND THE EVOLUTION OF A KEY INNOVATION IN AN
EXPERIMENTAL POPULATION OF ESCHERICHIA COLI…………………………………79
Abstract…………………………………………………………………………………..80
Background and Introduction……………………………………………………..……..82
Results……………………………………………………………………………87
+
Evolution of Cit Function in Population Ara–3…………………….......…….87

xvi

+

Historical Contingency in the Evolution of Cit …..………………………...…..92
Statistical Analysis of the Replay Experiments….…………………………...….97
+
Rates of Mutation to Cit and Quantification of the Potentiation Effect….….…99
Frequency-dependent Selection Maintains Ecological Diversity…..…………..103
Discussion and Future Directions………………………………………………………105
Methods and Materials………………………………………………………….………111
The Long-term Evolution Experiment..………………………….……………..111
Media and Culture Conditions………..……….…………………….………….111
Analysis of the Population Expansion……...….……………………….………112
+
Search from Cit Variants...……………………………………………………112
Growth Curves...………………………………………………………………..113
Replay Experiments…………….……………………………………………..113
+
Confirmation of Cit Variants….……………………………..………………113
Test for Frequency-dependent Interaction.……………………………………114
Acknowledgements…………………………………………………………..…………114
Supporting Information…………………………………………………………...…….114
SI Materials and Methods……………….…………………………………….114
Media and Culture Conditions…..……………………………………...114
Analysis of the Population Expansion…...……………………………..115
First Replay Experiment…..………………………………………..…..115
Second and Third Replay Experiments.…………………………...……116
+
Confirmation of Cit Variants...…………………………………..……118
Test for Frequency-dependent Interaction...………………………..…..118
References……………………………………………………………………...……….120
CHAPTER 3
GENOMIC ANALYSIS OF A NOVEL INNOVATION IN AN EXPERIMENTAL
POPULATION OF ESCHERICHIA COLI……………………………………………………..126
Abstract……………………………………………………………………………..…..127
Introduction and Background...………………………………………………...………128
Genome Sequencing and Phylogenetic History...………………………………………130
+
Evolution of the Cit Function…………………………………………………………134
+

Actualization of the Cit Function…………………………………….………..135
+

Refinement of the Cit Function……………………………………………..…143
+

Potentiation of Cit Evolution...………………………………………………..158
Perspective……………………………………………………………………...………172
Methods……………………………………………………………………….………..175
Evolution Experiment...…………………………………………………...……175
Genomic DNA Isolation………………………………………………………..176
Whole-genome Shotgun Re-sequencing and Mutation Detection.….………….176
Phylogenetic Reconstruction..…………………………………...……………..177
PCR Screens of cit Amplifications………..……………………...…………….178
Expression Experiments……………………………………..………………….178

xvii

Isogenic Strain Construction...………………………………...………………..181
Growth Trajectories..……………………………………………….…………..182
Fitness Assays…...……………………………………………..……………….182
Plasmid Construction…...………………………………..……………………..182
+
Identification of Cit -conferring Mutations in Replays...………...…………….183
Acknowledgements………………………………………………………..……………186
Appendix………………………………………………………………………………..188
References……………………………………………………………………..………..199

xviii

LIST OF TABLES

Table 1.1

Development of DM medium……………………………………………46

Table 2.1

Summary of replay experiments…………………………………………94

Table 2.2

Statistical analyses of three replay experiments………………………....98

Table 3.1

Historical Ara-3 clones subjected to whole genome sequencing……….131

Table 3.2

Results of PCR screens on whole-population samples for cit
amplification……………………………………………………………137

Table 3.3

Detection of cit amplification in Ara–3 clones………………………....138

Table 3.4

Point mutations in early Cit genomes…………………………………147

Table 3.5

IS insertions in early Cit genomes…………………………………….148

Table 3.6

Deletions in early Cit genomes……………………………………..…149

Table 3.7

Amplifications in early Cit genomes…………………………….……150

Table 3.8

Estimated citT module copy number…………………………………...157

Table 3.9

Phylogenetically informative mutations………………………………..159

Table 3.10

Mutations affecting cit region in Cit replay mutants………………….166

Table 3.11

Annotated differences between the genomes of Cit mutants and their Cit
parent clones……………………………………………………………167

Table 3.12

Primer pairs used in this study………………………………………….180

Table 3.13

Clones used in growth – trajectory experiments…………………..……185

+

+

+

+

+

+

–

–

Appendix Table A.1 Phylogenetic placement of Cit replay clones………………….189

xix

LIST OF FIGURES
+

Figure 2.1

Population expansion during evolution of the Cit phenotype…………..89

Figure 2.2

Growth of Cit and Cit cells in DM25 medium…………………….…..90

Figure 2.3

Alternative hypotheses for the origin of the Cit function……………....93

Figure 2.4

Mutation rates from Ara to Ara and Cit to Cit of the ancestor and a set
of potentiated clones…………………………………………………....101

Figure 2.5

Frequency-dependent selection allows stable co-existence of Cit and Cit
clones…………………………………………………………………...104

Figure 3.1

Phylogeny of Ara–3 population………………………………………...132

Figure 3.2

Total number of mutations relative to the ancestor…………………….133

Figure 3.3

Tandem amplification found in Cit genomes………………………….136

Figure 3.4

Annotation of sequence adjacent to the boundary of the cit amplification
+
found in Cit genomes………………………………………………….140

Figure 3.5

Expression levels from native citT, native rnk, and evolved rnk-citT
regulatory regions during aerobic metabolism………………………....142

Figure 3.6

phenotype in potentiated
New rnk-citT module confers Cit
background…………………………………………….…………..……144

Figure 3.7

Growth of early Cit clones in DM25 medium……………………...…145

Figure 3.8

Refinement of Cit phenotype by increased copy number of rnk-citT
module…………………………………………………………………..156

Figure 3.9

Key for placement of clones used in replay experiments that were sampled

–

+

+

–

+

–

+

–

+

+

+

+

+

before generation 20,000………………………………..……………...161
Figure 3.10

Key for placement of clones used in replay experiments that were sampled
from generations 20,000 or later………………………………………162

xx

+

Figure 3.11

Evidence of epistatic interactions in potentiation of Cit phenotype…..164

Figure 3.12

Mutations that produced Cit phenotype in 14 replay experiments…....173

Figure 3.13

Growth of Cit mutants derived in replay experiments………………...174

+

+

xxi

KEY TO SYMBOLS AND ABBREVIATIONS

–

Ara ………………………………………..………………….......Unable to use arabinose
+

Ara ………………………………………..…………………..Capable of using arabinose
–

–

+

+

Ara ##.....................Designation for one of six LTEE lines founded with an Ara ancestor
Ara ##....................Designation for one of six LTEE lines founded with an Ara ancestor
C1……………………………………………………………………....….Ara–3 clade one
C2………………………………………………………………...……….Ara–3 clade two
C3………………………………………………………………………..Ara–3 clade three
CI………………………………………………………………………Confidence interval
cit amplification………………………...…Gene amplification overlapping the cit operon
+

Cit ...................Aerobic citrate-utilizing phenotype or clone or group with that phenotype
–

Cit ……..…………………………………………..…….Unable to use citrate aerobically
CRP………………………………………………………...….Catabolite repressor protein
DM………………………………………….……Davis and Mingolini minimal salts base
DM25…………………………….……...DM-Base supplemented with 25 µg/mL glucose
DM0……………………………..…...………....………….DM-Base with no added sugar
DM500………………………………….…………...…DM-base with 500 µg/mL glucose
DM1000………………………………………………….DM-base with 1 mg/mL glucose
DNA……...…………………………………………………………Deoxyribonucleic acid
FS…………………………………………………………………………..……Frameshift

xxii

IS……………………………….………………………..…….Insertion sequence element
L……………………………………………………………………………….………Liter
LB……………………………………………………………………...Luria-Bertani broth
LTEE……………………………...……………………...Long-term evolution experiment
MC………………………………….……………………….……Minimal citrate medium
mL……………………………………………………………………………….....Mililiter
NC………………………………………………………………………………Noncoding
NS………………………………………………………………………...Nonsynonymous
OD420……………………………..….…………….Optical density at wavelength 420 nm
PCR……………………………………………………………..Polymerase chain reaction
P0………………………..Fraction of replicates producing no mutants in a fluctuation test
SNP……………………………………………………....Single nucleotide polymorphism
S……………………………………………………………………………....Synonymous
TA…………………………………………….…Tetrazolium arabinose indicator medium
TCA…………………………..…………Of or pertaining to the TriCarboxylic Acid cycle
UC……………………………………………………....……Unsuccessful clade of Ara–3
µ…………………………………………………………….………………...Mutation rate
µg……………………………………………………………….………………Microgram

xxiii

`

CHAPTER 1

HISTORICAL CONTINGENCY AND EVOLUTION

Evolution involves the interplay of chance and deterministic components in a sequential
process of descent with modification that occurs over time in organismal lineages that stretch
back unbroken over billions of years. Evolution is therefore an intrinsically historical process
(Darwin 1859, Mayr 1980, 1988, Bowler 2003, Gould 2002). Over the last few decades, the
importance of the historical nature of evolution, and particularly its impact on the repeatability of
evolutionary outcomes, has been among the most vibrant and far-reaching of debates within
evolutionary biology.

A BRIEF HISTORY OF THE NEGLECT OF HISTORY IN EVOLUTIONARY
BIOLOGY
The historical nature of evolution has been a core tenet of biology since Darwin.
However, though Darwin certainly recognized its potential importance (Darwin 1866, Beatty

1

2006a, 2008), the role of history in evolution was long given little emphasis or consideration.
This neglect is largely a reflection of how evolutionary biology developed after the publication
of On the Origin of Species. The failure of Darwin to gain wide acceptance of natural selection
led to the “Eclipse of Darwinism”. During this period much of evolutionary biology was
concerned with struggles to develop satisfactory theories of the core mechanisms evolution, with
neo-Lamarkism, orthogenesis, saltationism, and theistic evolution the dominant proposed
alternatives to natural selection (Huxley 1942, Provine, 1971, Bowler 2003, Larson 2004). The
conflict was exacerbated by early findings in genetics by Morgan, Pearson, de Vries, and
Bateson in the wake of the re-discovery of Mendel’s work, which added the mutation theory of
evolution to the other competing ideas. In the absence of a consensus on the fundamental
mechanism of evolution, consideration of contributing factors such as history and chance was
difficult during this period.

Moreover, one of the motivating concerns for the search for

alternatives to natural selection was the ongoing dominance of progressive, if not outright
teleological, notions of evolution under which chance and history were a priori debarred from
significant influence (Bowler 2003).
The conflict over the evolutionary mechanism was finally resolved during first half of the
20th century.

Between 1918 and the mid-1930’s Fischer, Haldane, and Wright developed

population genetics, which effectively integrated Mendelian genetics and natural selection
(Provine 1971, Mayr and Provine 1980, Bowler 2003).

Over the next decade, Mayr,

Dobzhansky, Simpson, Huxley, and a number of other luminaries integrated population genetics
with field biology and paleontology, resulting in the modern synthetic theory, which unified
biology via the acceptance that, “… gradual evolution can be explained in terms of small genetic
changes (“mutations”) and recombination, and the ordering of this genetic variation by natural

2

selection; and the observed evolutionary phenomena, particularly macroevolutionary processes
and speciation, can be explained in a manner that is consistent with the known genetic
mechanisms” (Mayr 1980).
Consideration of history’s role in evolution was not advanced by the synthesis. Due to its
basis in mechanistic models of population genetics, the issue of history in evolution had been
largely ignored [It must be acknowledged, though, that Wright’s shifting balance theory,
formulated during this period, does involve a historical component due to its emphasis on the
starting point of evolving populations, the chance role of genetic drift, and the effect of mass
selection during environmental change on genetic composition (Wright 1932, 1977, Pigliucci
2009)]. Moreover, during its consolidation in the 1950’s and early 1960’s, the synthesis, or
rather its interpretation, “hardened” around natural selection to the effective exclusion of other
aspects of evolution. Drift, for instance, was noticeably downplayed (Gould 2002, Bowler
2003). One effect of this was an overemphasis on the capacity of natural selection to mold most
any trait in most any way, which would mean that natural selection could effectively overwrite
prior history, rendering it irrelevant (Gould and Lewontin 1979, Griffiths 1996, Gould 2002,
Bowler 2003).
Beginning in the late 1960’s inklings of a renewed interest in the role of history and chance
in evolution began to develop. Richard Lewontin, for instance, published on the idea of
“historicity” in evolution, showing that the historical sequence of environmental fluctuations
can have profound effects on allele frequencies, even under conditions in which selection is
the only evolutionary force (1966).

Soon afterward, Kimura introduced neutral theory that

placed substantial emphasis on drift (1968). During this same period, a few paleontologists,
most notably Norman Newell, began to reexamine the issue of mass extinctions and their

3

potential association with catastrophic events, which suggested a possible role for singular
historical events in the evolution of life (Gould 2002, Bowler 2003, Allmon 2009). By the
mid-1970s, a number of biologists began to call for re-evaluation of evolution to better
explicitly consider the role history and chance (Beatty 2006).
Stephen Jay Gould and “Replaying the Tape of Life”
Among the biologists who began to this process of re-evaluation in the 1970’s was one of
Newell’s former graduate students, Stephen Jay Gould. By the early 1980’s, Gould had emerged
as the foremost proponent of the view that history often swamps the deterministic power of
natural selection to drive evolution to predictable ends, thereby rendering evolutionary outcomes
fundamentally quirky and unpredictable. Gould best elaborated this view in his 1989 book,
Wonderful Life, a prize-winning account of the Burgess Shale fossils written for the general
public. The principle narrative focus of the book was the re-evaluation of the Burgess Shale
fossils by Whittington, Briggs, and Conway Morris, which had shown that, contrary to the results
of earlier, less intensive studies, the ~500 million year old fossil organisms display a far greater
range of body plans than is currently seen across all extant phyla. Gould notes that each of the
extinct body plans represented an alternate solution to basic biological challenges, and that
diversification from each would have yielded descendents radically different from modern
organisms that evolved with the body plans that did survive. The modern living world is
therefore just one of many outcomes that could have resulted from a past that included the full
myriad of Burgess shale organisms. Gould thought this realization suggested a general feature of
evolution as a whole that he illustrated in a thought experiment that he called “replaying life’s
tape”:

4

You press the rewind button and, making sure you thoroughly erase everything
that actually happened, go back to the any time and place in the past – say, to the
seas of the Burgess Shale. Then let the tape run again and see if the repetition
looks at all like the original. If each replay strongly resembles life’s actual
pathway, then we must conclude that what really happened pretty much had to
occur.

But suppose that the experimental versions all yield sensible results

strikingly different from the actual history of life? (Gould 1989)

In Gould’s view, the second possibility would prove correct, and “any replay of the tape
would lead evolution down a pathway radically different from the road actually taken” (Gould
1989). This assertion was grounded in Gould’s view that the deterministic power of natural
selection only constrains evolution within broad channels of possible outcomes. Within these
broad channels are innumerable opportunities for possible evolutionary trajectories to outcomes
that vary greatly in their details (Gould 1989, 2002, Beatty 1995). In any replay, these nonnecessary details would be determined by chance historical events that would propagate down
the ensuing causal chain of evolutionary history, potentially leading to vastly different outcomes.
Therefore, because evolution involves an interaction of chance and necessity playing out over a
continuous history, evolutionary outcomes are in some sense crucially dependent upon the
precise, often chancy details of long, complex causal chains of antecedent states, making them
the unpredictable and quirky products of historical contingency (Gould 1979, 1989, 1993, 2002,
Allmon 2009). Gould thought that this role for historical contingency was the key feature of
evolution, and that evolution could not be understood except in light of it (Gould 1989, 1993,
2002, Allmon 2009, Bambach 2009). Consequently, evolution is not like the simpler natural

5

phenomena studied by chemistry and physics that are historically insensitive, and can thus be
explained by robust process explanations (Sterelny and Griffiths 1999). Instead, evolutionary
outcomes can only be understood as historical events subject to actual sequence explanations:
Historical explanations take the form of narrative: E, the phenomenon to be
explained, arose because D came before, preceded by C, B, and A. If any of these
earlier stages had not occurred, or had transpired in a different way, then E would
not exist (or would be present in a substantially altered form, E, requiring a
different explanation). Thus, E makes sense and can be explained rigorously as
the outcome of A through D … A historical explanation does not rest on direct
deduction from laws of nature, but on an unpredictable sequence of antecedent
states, where any major change in any step of the sequence would have altered the
final result (Gould 1989).
Needless to say, if correct, this perspective on evolutionary outcomes would have profound
ramifications for how both evolution and the history of life may be understood (Gould 1985,
Blaser 1999, Sterelny and Griffiths 1999, Beatty 2006, Gould 2002).
As Gould himself noted (1989), no evolutionary biologist dismisses chance history as
irrelevant to evolution, but few are willing to grant it as much sway as he does. His views have
thus engendered controversy and strong reactions. Richard Dawkins, coming from a position
more tightly focused on adaptation, has argued that the deterministic and directional nature of
natural selection is sufficient to drive evolutionary outcomes to similar highly adapted states
(Dawkins 1996). Simon Conway Morris (1998, 2003, 2010) makes similar arguments, but
emphasizes that the widespread occurrence of evolutionary convergence among lineages with
different evolutionary histories suggests that the range of viable evolutionary endpoints are

6

limited. Because “the evolutionary routes are many, but the destinations are limited”, evolution
is repeatable regardless of how contingent the routes taken may be on chance and history
(Conway Morris 2003). With evolution so constrained, were one to replay the tape of life
multiple times, the resulting outcomes would be markedly similar in their essential features, with
contingency only resulting in relatively trivial differences (Van Valen 1991, Maynard Smith
1992, Dennett 1995, Dawkins 1996, Conway Morris 2003, 2010). Such repeatability would
make evolution predictable and at least potentially subject to robust process explanations
(Sterelny and Griffiths 1999).
There has been one history of life on Earth, and replaying the tape of that history on the
planetary scale Gould envisioned is clearly not possible. This limitation has made resolution of
the contingency debate difficult, and explains at least part of its contentiousness.

Moreover,

given the involvement of chance at multiple points in evolutionary histories, contingency no
doubt plays at least some role in evolution. Indeed, the two hardened positions regarding
contingency in evolution obscure that the real issue is more a matter of the relative importance of
contingency in evolution, its effects, and the levels on which it works (Foote 1998). These are,
of course, questions that can only be answered by empirical studies, and the two decades since
the publication of Wonderful Life have seen numerous attempts to test Gould’s notions of
contingency (Gould and Woodruff 1990, Travisano et al 1995, Losos et al 1998, Emerson 2001,
de Queiroz and Rodriguez-Robles 2006, Vermeij 2006, Fukami et al 2007, Dick et al 2009).
These studies have illuminated a number of aspects of contingency in evolution, as I will discuss
later. Before doing so, however, there are some important conceptual issues that must be
addressed.

7

Difficulties in Gould’s Notions of Contingency
As empirical studies of evolutionary contingency have accumulated, there have been
questions as to whether or not they actually engage with Gould’s ideas (McRae, 1993, Beatty
2006b, Turner 2011). Such questions are likely typical of attempts to examine any difficult and
controversial issues, but they seem to be exacerbated by the difficult in determining precisely
what Gould meant by contingency. Contingency was one of Gould’s favorite themes, and one of
the central foci of his later thinking on evolutionary theory. However, he never produced a
formal treatment in which he explicitly developed and clarified his ideas on contingency and its
sources (Gould 2002, Bambach 2009). Even in Gould and Woodruff (1990), his one technical,
empirical examination of historical contingency in production of an evolutionary outcome, he
does not formally define his concept. Moreover, most of Gould’s considerations of contingency
are presented informally and sometimes tangentially in non-technical writings published over
many years. While these writings, particularly his essays in Natural History, are delightful and
interesting, the descriptions of his concept of contingency they provide are imprecise, and lack
the sharply drawn explication that would have made them more amenable to empirical tests.
Adding to this difficulty, Gould’s description of contingency varies across these informal
treatments in potentially significant ways.
Inconsistency in Gould’s description of his notions of contingency may even be found in
Wonderful Life, which contains his most detailed consideration of the subject. Indeed, Beatty
(2006b) points out that Gould actually gives two different versions of contingency in Wonderful
Life. Gould presents the first when he asserts that replaying the “tape of life” would yield a
different result each time.

Beatty calls this version “contingency as unpredictability”, and

8

identifies it with Gould’s position of denying “that evolution by natural selection is sufficient to
guarantee the same evolutionary outcome, even given initially indistinguishable ancestral
lineages and indistinguishable environments…” This version shifted over time to suggest that
the unpredictability it deals with exists on a continuum (Beatty 2006). In the initial version of
the replay experiment in Wonderful Life, Gould suggests that a replay from any past time point
would yield a different outcome. He later amended this to specify that the replay would need to
be from a distant past time point (Gould 1990), a requirement that he later used to dismiss the
macroevolutionary repeatability among Anolis lizards identified by Losos et al (1998) as having
any real evidentiary relevance to contingency’s role in evolution (Beatty 2006b).

In the second,

labeled “contingency as causal dependence”, Gould emphasizes the dependence of outcomes on
the precise chain of underlying antecedent events.

In describing this version, Gould links

different outcomes to differences in the antecedent states. However, as McRae (1993) has
pointed out, there are inconsistencies in this version as well. Gould suggests at one point in
Wonderful Life that any change in an antecedent event will have significant consequences to the
evolutionary outcome, but later states that only a major change would do so. The two versions
also intersect at times, with Gould suggesting at one point that replaying from identical starting
points would yield different outcomes, but later specifying that that the starting point would have
to be changed for a different outcome to occur.
Turner (2011) discerns a third possible version of contingency in Gould’s concern for
chance in macroevolution and certain inconsistencies in his consideration of stochastic forces.
Gould had been part of a study group at the Marine Biological Lab at Woods Hole,
Massachusetts in the 1970’s that had also included David Raup, Thomas Schopf, Daniel
Simberloff, and Jack Sepkoski, which produced a series of papers that demonstrated how

9

stochastic models could replicate observed phyletic patterns, and suggesting a role for stochastic
processes in macroevolution (Raup et al 1973, Gould et al 1977). Gould often credited his
experiences with this group as having a seminal effect on his subsequent ideas about the
intrusion of random historical events into evolution (Gould 2002, Allmon 2009, Bambach 2009).
Despite the group’s emphasis on stochasticity, Gould later differentiated historical contingency
from stochastic processes, among which he specifically cited drift 1 (Gould and Woodruff 1990,
Beatty 2006b). However, in his final description of contingency, Gould specifically mentions
stochasticity as a contributing factor:
…the tendency of complex systems with substantial stochastic components, and
the intricate nonlinear interactions among components, to be unpredictable in
principle from full knowledge of antecedent conditions, but fully explainable after
time’s actual unfoldings (Gould 2002).
Turner thinks that these inconsistencies suggest that Gould’s actual view was that contingency
was synonymous with macroevolutionary stochasticity introduced by random species sorting and
the effects of mass extinction. Gould’s dismissal of microevolutionary stochastic processes such
as drift may then be related to a subsidiary notion of contingency at the microevolutionary level
that Gould thinks may be interesting, but not important over the whole of evolutionary history.
This comports well with Gould’s view that macroevolutionary forces swamp patterns introduced
by microevolution (Gould 1985, Allmon 2009).
These inconsistencies and general lack of clarity in Gould’s conception of contingency
present numerous difficulties.
1

First and foremost, different studies intended to study

Oddly, Gould explicitly describes bottleneck effects, mutations, and mutation order as

contributing to contingency despite their stochasticity (Gould and Woodruff 1990).

10

contingency can engage with different versions of Gould’s conception of contingency
(Beatty 2006b). This ambiguity can complicate the evaluation of individual studies, as well
as the comparison of multiple studies, leading researchers to talk past each other. Moreover,
if Gould did actually conceive of contingency as impacting evolution principally at the
macroevolutionary level, then his ideas should not actually be looked to for technical
guidance or evaluation of microevolutionary studies. Therefore, while Gould deserves credit
for bringing notions of historicity and contingency to the forefront of evolutionary thought
and to the attention of empirical researchers, it would perhaps be best to re-think the notion
of contingency in evolution independent of his speculations upon it. This divorce of the
concept from Gould’s writings would carry two benefits. First, it provide an opportunity to
develop a clearer, sharper concept of contingency that could then be used to guide research
and interpret empirical findings.

Second, it would explicitly divide consideration of

contingency’s role in evolution from Gould’s rather extreme advocated position that, like
Conway Morris’ position, ultimately comes down to a philosophical position (Dennet 1995,
Sterelny and Griffin 1999). In the following section, I make a brief first attempt at this
separation. I will begin by examining the general concept of historical contingency in human
history. From there, I will attempt to first define the characteristics necessary for a system to
be contingent, and then to derive a general definition of contingency from those features and
considerations of philosophical treatments of the subject. I will thereafter examine what
features of evolution render it subject to contingency based upon this definition, and then
review a number of empirical treatments of contingency in evolution. This approach will
allow me to set my empirical work with the evolution of aerobic citrate-using E. coli into

11

context prior to giving an extended background to lead into the empirical work presented in
chapters 2 and 3.

RECONSIDERATION OF HISTORICAL CONTINGENCY
Contingency and Human History

For want of a nail the shoe was lost.
For want of a shoe the horse was lost.
For want of a horse the rider was lost.
For want of a rider the battle was lost.
For want of a battle the kingdom was lost.
And all for the want of a horseshoe nail.
- Traditional Proverb

Cleopatra’s nose, had it been shorter, the whole face of the world would have been different.
-

Blaise Pascal, Pensées

But fortune, which has great power in all matters … causes great shifts in human affairs with just
a little disturbance.
- Gaius Julius Caesar, Commentaries on the Civil War
A Historical Vignette
During the fall of 1861, the US Army began operations to secure the Mississippi River. Early in
November, 5,000 Confederate soldiers crossed the river into Union-held Missouri, where they

12

made camp near a town named Belmont. To counter this incursion, Brigadier General Ulysses S.
Grant sailed from Cairo, Illinois with 3,100 men on November 6. On the following day, he
landed his force north of the city, and led them in an assault on the Confederate beachhead in
what would be his first combat leadership role of the war. After initially taking the Confederate
camp, Grant’s force came under heavy attack from Confederate reinforcements and an artillery
concentration on the opposite side of the river. Despite withering fire, Grant was able to guide
his men back to their transport ships for a retreat to the safety of Paducah, Kentucky. Shortly
after boarding one of the riverboats, Grant entered the captain’s quarters, and lay down on a sofa
to rest. He rose only a few minutes later to return to the deck. As he stepped away from the
sofa, a Confederate musket ball fired from shore pierced the wall of the cabin, and struck the
sofa precisely where his head had been moments before (Grant 1885). Grant later went on to
take Vicksburg by siege, opening the full length of the Mississippi to Union control, and to win
the battle of Chattanooga. Based on these successes, Lincoln made Grant Commanding General
of the United States Army. From that position, Grant awarded command of the Army of the
Tennessee to William Tecumseh Sherman, and developed a unified strategy of coordinated
attacks on Southern forces across multiple fronts. Sherman later took Atlanta, which contributed
to Lincoln’s reelection in 1864. Grant’s strategy ended up winning the war in 1865, and he went
on to be elected president in 1868 (McPherson 1988).
Counterfactual Analysis of History
One aspect of the role of contingency in history can be illustrated by use of
2

counterfactuals. A counterfactual is a type of historiographic thought-experiment that is used to
2

The first known use of a counterfactual is in Thucydides’ history of the Peloponnesian War

(1972), but Thucydides did not originate the concept. Psychological research over the last thirty

13

analyze causation, and involves the development of a “what if” scenario in an event or historical
circumstance is changed, and the consequences of that change on later events are examined
(Lewis 2001, Ferguson 2000, Day 2008). In this case, we will pose a counterfactual of the above
history in which Grant did not get up from that sofa after the battle of Belmont when he did, and
look at the putative consequences of this slight change in a minor event on the later event of his
election to the presidency. Assuming that he did not get up when he did, Grant would have
remained on the couch until the bullet entered the cabin, been struck in the head by the bullet. It
is impossible to assess the energy possessed by the bullet when it struck, but we will assume that
the resulting wound would have been fatal. Note that this event belongs to a particular class of
events with irreversible immediate consequences. Subsequent history would have had to have
unfolded without Ulysses S. Grant. Obviously, Grant dying on November 7, 1861 would have
precluded his being elected president in 1868. This is to say that his election is not resilient to
the counterfactual, and it is thus possible to say that Grant’s election to the presidency, as well as
any event in which he took part after November 7, 1861, was causally dependent up him having
gotten up from that sofa before the bullet entered the cabin.

Now consider a second

counterfactual, this being that Grant was killed at Shiloh, as he almost was in the actual history
(Flood 2006). Again, Grant’s election in 1868 would be precluded under the counterfactual
situation. Indeed, innumerable events, both large and minute, antecedent to Grant’s election can

years has shown that humans are inherently prone to counterfactual thinking (Kahneman and
Tversky 1982).

Indeed, as decisions of future action often involve deciding between the

potential consequences of multiple course of action, counterfactual thinking appears to play an
important role in learning (Kahneman & Miller 1986, Coricelli and Rustichini 2010).

14

be identified for which counterfactuals may be constructed to which the election would not be
resilient.
Now consider the effect of the first counterfactual on another event, this time Lincoln’s
reelection in 1864. In actual history, Grant played an important part in this event. Historians
generally credit Sherman’s capture of Atlanta with ensuring a change in public mood that
contributed greatly to the reelection. Sherman had his position at the head of the Army of the
Tennessee because Grant, who had been named Commanding General of the United States Army
by Lincoln, put him there. However, unlike the case of Grant’s presidential election, it is
difficult to say what effect the counterfactual would have on the fact of Lincoln’s election.
There are many plausible alternate scenarios in which Lincoln was reelected without Grant’s
involvement, suggesting some degree of counterfactual resilience, so we cannot logically argue
that Lincoln’s reelection was causally dependent upon Grant getting up from the couch when he
did. However, it is clear that the particular event of Lincoln’s reelection as it occurred in history
was causally dependent upon Grant getting up from the couch. In this case, we can say that there
is necessary causal sensitivity in the how of the later event to the earlier event, but not in the fact
of the later event to the earlier event. This situation could be called sensitive counterfactual
resilience indicative of causal relationship, but not necessary causal dependence. Innumerable
antecedent events, many of them of them tiny and of little or no obvious significance at the time,
can be identified that would display this same relationship to Lincoln’s reelection, just as there
are also innumerable antecedent events that can be identified to which it would display an
absolute lack of counterfactual resilience.
For virtually any other historical event, there is a similar range of antecedent events for
which counterfactuals may be posited to which the later event would be either not resilient or

15

sensitively resilient. Now consider that each of those events occurs in the context of many other
coincident events for which this is also true. Moreover, this historical context has emerged from
earlier historical contexts with the same sort of complex causal structure.

Consequently,

historical events are crucially dependent upon the sequential, and coincident occurrence of many
interacting and intersecting antecedent events to which they are causally connected to various
degrees (Bloch 1953, Ferguson 2000, Lewis 2001). Historical causation is thus bewilderingly
complex, with small, seemingly trivial, odd happenstance events often playing critical roles in
the chain of events that leads to an outcome.
The complexity of historical causation makes prediction of historical outcomes
notoriously difficult, as all the relevant causal factors and events cannot be accounted for from a
given starting point. Moreover, the involvement of small, otherwise trivial events lends a sense
that history could have been different had those tiny events occurred differently. There are two
notions of contingency that come out of these points. First, there is the idea of contingency as an
inability to understand historical outcomes from an initial historical point, provided sufficient
temporal distance due to the inability to consider all of the causal factors linking the initial point
to the outcome. The second, which I maintain is the relevant notion, is that there is true
indeterminacy from a given historical starting point, with the consequence that history matters in
determining what the outcome of that starting point is. With the possibility of different histories,
historical outcomes are only explicable in terms of the actual sequence of events that that lead to
them, and the causal relationship between those events (Ben-Menahem 1997, Ferguson 2000). I
maintain that this notion of historically sensitive outcomes is the core of historical contingency.

16

Properties of a Historical Contingent System
The above discussion indicates that, in order for a system to be subject to historical
contingency, it must be a system in which history is possible, and in which that history plays a
determinative role in outcome instantiation.

These requirements suggest that, at heart, a

historically contingent system must be one in which there are causal relationships between
temporally ordered sequences of events, as well as to the possibility of alternate outcomes
(Lamprecht 1971, McPherson 1988, Ferguson 2000). There therefore would seem to be at least
five properties necessary for contingency to be a feature of any temporally ordered system,
process, or phenomenon.

First, there must be a necessary causal relationship between an

outcome and at least some of its sequence of antecedent states. If an outcome is not necessarily
causally dependent upon any antecedent states, then it cannot be said to be contingent upon
them, and should be historically insensitive. Second, multiple outcomes must be possible from a
given initial state. Third, there must be at least as many possible causal chains as there are
possible outcomes such that each outcome may be reached by one or more chains, but no one
particular chain can result in multiple outcomes. The second and third requirements flow from
the first. A single possible outcome would require that any and all possible sequences of
3

antecedent states result in that outcome . In this case, the outcome would inevitably flow from
the initial state no matter the intermediate states, and would only have a causal relationship with
the initial state. (This situation would suggest a system that would display transient contingency
and contingent intermediate outcomes despite an inevitable final outcome, as might well be the
3

To avoid future confusion, when I speak of deterministic processes, this is what I have in mind.

If an outcome is guaranteed by a given initial state, then the process is deterministic in my
parlance. Determinisim, causality, and contingency are different, though related, things.

17

case with the universe as a whole.)

Alternately, were there a single possible sequence of

antecedent states, but multiple possible outcomes, the outcomes would be essentially random
with respect to, and therefore not causally related to their antecedents. Fourth, the multiple
causal chains cannot be fully reticulated, but must be at least partially independent. If they were
fully reticulated, then all outcomes would be reachable via all causal chains. This scenario
would effectively reduce down again to the outcome being random with respect to the underlying
causal chain.
While a contingent system must be causal in the way outlined by the first four criteria, as
a fifth criterion, it must also have a source of indeterminacy in order for alternate causal chains to
be initiated and alternate end states achieved.

Highly complex deterministic systems, for

instance, are subject to chaotic effects such that even small chance variations in their initial states
can result in wildly divergent and unpredictable outcomes (Ruelle 1991).

In the case of

historically contingent systems, however, there must be added inputs of indeterminacy beyond
that of the initial state in order for history to matter to the outcome. If this were not so, then
4

contingency would essentially be synonymous with highly complex deterministic causation . In
4

There is a distinction here that I like to think of as the difference between Fate and Fortuna. In

mythology, the Fates have a plan for all of history and everyone and everything within it. The
Fates are implacable, and their plan will come to be. Once they decide upon it, it will occur.
Also in mythology, there is Fortuna, the goddess of luck. She is not implacable, but fickle,
swayable, and capricious. Fortuna can change the course of events at any time she pleases. It is
possible that history may better be explained by the metaphor of Fate, where the outcomes of
history, strange and improbable though they might seem, are and have been inevitable. I do not
argue that this is not the case so. However, if this is so, then I don’t think that contingency can

18

considering this issue, Lamprecht (1971) traced contingency to the fact that outcomes result from
the non-necessary interaction of multiple independent causal chains that were causally
unconnected prior to their convergence. He illustrated this point with a particularly vivid and
informative example:
A former neighbor of mine went out into the woods on his farm the first morning
of deer season a few years ago. He is an excellent shot with a rifle, as I had had
occasion to observe frequently. He sighted a buck just after sunrise, aimed and
shot his rifle with confidence, and then saw the buck gallop off quite unharmed.
Puzzled by his failure to hit his quarry, he advanced slowly and carefully along
the line of his shot. And he found a small twig of a shrub which had been freshly
broken so that the upper part of the twig dangled loosely down. He concluded
that his bullet had broken the twig as the twig swayed in the breeze, and that the
bullet had been deflected by collision with the twig from the course it would have
taken otherwise…An unexpected factor had intruded into the course of a carefully
planned event. This factor, though having important causal consequences in the
issues of the event, was not in any way causally related to other factors in the
situation until it intruded with sudden efficacy. There was a causally continuous
sequence in the hunter’s intent, his sighting of the buck, his aiming and firing his
rifle, and the swift flight of the bullet out of the muzzle of the rifle. There was
also another causally continuous sequence in the movements of the buck through

be anything other than an illusory concept. Therefore, my discussion assumes that the Fortuna
metaphor is more appropriate, and that there really is indeterminacy in history that permits for
actual contingency. I may very well be quite wrong.

19

the woods. Those two sequences were brought together by the intent of the
hunter to shoot the buck. But there was a third causally continuous sequence in
the atmospheric conditions, the blowing of the breeze, and the swaying of the
twig in the breeze. And this third sequence was quite independent of the other
two sequences.
Note the similarity in the structure, though not the end result, of this story and the example from
the Civil War that I discussed earlier. In history and any other contingent system, not all of the
causal links leading to an outcome are necessarily related to or entailed by all the others. Only
the outcome must be necessarily related to all the antecedent events in its underlying causal
chain. This nonrelationship between causal links can result in the non-deterministic causal
chains that underlie outcomes seen in history (Lamprecht 1971). This, of course assumes that
the consequent interaction of the initially unrelated chains is not necessitated by the initial state.
Alternatively, indeterminism may also arise from one or more causal links being either affected
by or the product of chance or stochasticity (Day 2008). By this I mean that there must be some
events in the causal chains such that either their occurrence or instantiated state is probabilistic
rather than necessary given the initial state of the system.
These five criteria clearly suggest that the contingency of a system varies, and exists on a
continuum between high contingency on one end, and complete determinism on the other
(Desjardins 2011).

Contingency should decline as the number of possible, or effectively

reachable, outcomes declines, and increase as the number of possible outcomes increases.
Contingency will also decline as the proportion of possible causal chains leading to a single
outcome increases because the favored outcome will become increasingly counterfactually
resilient (Ben-Menahem 1997). After all, a system in which there are twenty possible end states

20

and 500 possible causal chains will display less contingency if 481 of those causal chains lead to
only one of the possible outcomes than if the number of causal chains were more evenly
distributed. Of course, a corollary of this principle will be that, as the number of causal chains
leading to a given outcome declines, that outcome becomes less counterfactually resilient, and
thereby more contingent. Another complication is that not all causal chains will be equally
likely, so that, even a large number of causal chains will not necessarily entail a highly
contingent system if all but one of those chains are highly unlikely. Similarly, the contingency
of a system should vary with the sources of indeterminacy, and the number of constituent causal
links subject to them. The actual contingency of a system will therefore be determined by the
number of possible end states, number of possible causal chains, the distribution of those chains
among the outcomes, the distribution of probabilities among the different chains, the reticulation
between the chains, and inputs of indeterminacy. Moreover, it is likely that contingency varies
depending on the length of the causal chain, and the temporal proximity between the starting
point and the end point. A final consideration is that the relevant contingency of a system will
also vary depending on how significantly the alternate outcomes vary between each other.
Given the above described criteria, it is possible to give a general description of a
historically contingent system.

Historical contingency should arise from any system of

temporally ordered events for which multiple outcomes are possible from a given initial state,
with the occurrence of a particular outcome causally dependent upon the occurrence of one or
more, but not all, series of events possible given the initial state, not all of which are necessarily
related to each other and/or in which chance plays a part, but that do form a discrete, at least
partially independent, causal chain necessarily related to the outcome. A historically contingent
outcome is one that requires a particular causal chain or set of causal chains possible from an

21

initial state, but not necessarily entailed by that initial state. As there are multiple possible
outcomes that are causally, but non-deterministically linked to the initial state, any prediction
regarding outcomes within a contingent system is necessarily uncertain, though potentially
probabilistic, depending on foreknowledge of the number of possible outcomes. Nonetheless,
whichever outcome results, it must be in principle fully explicable in terms of the events in its
causal chain, which is to say that precise recurrence of those events will result in the same
outcome (Lamprect 1971, Tucker 1999). It is important to note that this prior unpredictability
and posterior explicability are consequences of contingency, and not contingency itself.

EVOLUTION AND CONTINGENCY
Evolution is a historical process of descent along continuous organismal lineages with
5

hereditary modification over time. Perhaps unlike history , evolution is subject to a strongly
deterministic process.

Natural selection works systematically to adapt populations to their

prevailing environments, driving them toward local adaptive peaks (Wright 1932, 1988, Dawkins
1996b). However, chance inputs abound in evolution, and all hold the potential to contribute to
evolutionary contingency. Natural selection acts on heritable variation generated by random
mutation, the random reshuffling of recombination, or introduced by gene flow. Due to genetic
drift, even beneficial mutations may be lost, while detrimental mutations may rise to high
frequency. Multiple qualitatively similar beneficial mutations may arise, and the fixation of one
as opposed to another may hinge on slight fitness differences and chance events including
subsequent mutations. These alternative mutations may differ significantly in their pleiotropic
effects and epistatic interactions, with significant consequences for possible future evolutionary
5

Hegel, Marx, and other proponents of strongly deterministic theories of history would, of
course, disagree.
22

paths, including their reversibility and level of reticulation (Dollo 1893, Wright 1932, 1988,
Jacob 1977, Cooper and Lenski 2000, Ostrowski et al 2005, 2008, Weinreich et al 2005, 2006,
Cooper et al 2008, Bridgham et al 2009, Salverda et al 2011). Beneficial mutations will arise in
some random order, with potentially significant consequences for the fitness and range of
allowed future mutations (Mani and Clarke 1990, Lenski et al 1991, Weinreich et al 2005, Nosil
and Flaxman 2011).

The accumulation of neutral or even weakly detrimental mutations can

likewise impact future evolutionary possibilities (Huynen et al 1996, Wilke 2001, Lenski et al
2006).
Any adaptation is also local, and potentially rapid and capricious biotic or abiotic
environmental change can radically change fitness (Lewontin 1966, Gould 2002). The chance
factors discussed above can drive populations to distinct adaptive peaks that, while similar in
fitness in one environment, may have considerably different correlated fitnesses in other
environments (Travisano et al 1995, Cooper and Lenski 2000). These differences can have
potentially significant consequences for survival or extinction during environmental changes and
the “different rules” that govern mass extinction events (Lewontin 1966, Gould 1985, Jablonski
1986, Raup 1991).

Following mass extinctions, future evolutionary outcomes will be

constrained to those reachable from the surviving lineages, which may be a subset of prior extant
or reachable outcomes (Gould 2002, Sole et al 2002).
Is evolution a contingent process? Evolution is a temporally ordered process with causal
relationships between outcomes and initial states. As discussed above, though causal chains in
evolution are subject to the highly deterministic process of natural selection, chance is
nonetheless an integral and inescapable part of them. Moreover, the large number of chance
inputs necessarily entails multiple possible causal chains in evolution, though natural selection

23

and the de-reticulating effects epistasis clearly impact the likelihood or even possibility of the
assembly of those chains.
contingent process.

Evolution therefore fulfills four of the criteria necessary for a

Whether or not evolution is contingent in principle therefore rests on

whether or not there are multiple possible evolutionary outcomes.
Whether or not there are multiple possible outcomes depends upon how evolutionary
outcome is defined. The living world contains millions of diverse species and subspecies that
occupy distinct niches, and are characterized by a diverse array of traits (Wilson 1993). If
individual species may be considered to be different evolutionary outcomes, then the living
world is testament to the existence of alternate evolutionary outcomes reachable from the same
initial state that is the last universal common ancestor (Woese 1998). Evolutionary outcome may
also refer specifically to adaptive traits and/or suites of traits evolved by particular lineages,
and/or ecological interactions and ecological assemblages involving or composed of multiple
lineages, in which case there are also clearly alternate outcomes in general (Foote 1998). On a
smaller scale, using this definition, whether or not there are multiple possible evolutionary
outcomes will vary depending on the environment. In general, for any environment, there should
exist some set of possible allowed evolutionary outcomes that likely corresponds to the range of
possible ecological opportunities that exist in that environment [Note that on a small scale what
is mean by environment can get complicated because both biotic and abiotic aspects have to be
taken into account, as both will impact the range of possible ecological opportunities actually
available (Fukami et al 20007, Meyer and Kassen 2007)]. This set, which could be considered to
be the evolutionary potential of the environment, will define the total number of possible
evolutionary outcomes under those conditions. Given the aforementioned extreme diversity of
life on Earth, this set is likely extremely large for most habitable environments (Wilson 1993).

24

On this definition, and on most scales, there will exist many possible evolutionary outcomes,
making evolution at least potentially contingent.
However, whether or not there are multiple possible outcomes so defined will also
depend on the lineages involved and their prior evolutionary histories. Future evolution must be
based on the range of mutations possible and allowable from that ancestral state (Jacob 1977,
Burch and Chao 1999, Weinreich et al 2006, Salverda et al 2011). Prior evolution may preclude
some evolutionary outcomes by preventing access to their underlying causal chains due to
limited reticulation between chains, and the irreversibility of much of evolutionary change
(Dollo 1893, Muller 1939, Dawkins 1996a, Gould 2002, Yedid and Bell 2002, Weinreich 2006,
Bridgham et al 2009). At the same time, prior evolution may also result in easier access to some
outcomes than others. Moreover, new adaptations must be built on earlier structures, and prior
adaptations may be incidentally co-opted for other functions, with qualitatively similar
adaptations varying in their potential for cooption (Jacob 1977, 1982, Gould 1983, 1993, 2002,
Lenski et al 2003a). Prior evolutionary history will therefore potentiate some evolutionary
outcomes while constraining others (Jacob 1977, Gould 2002, Weinreich et al 2006, Salverda et
al 2011). For this reason, the range of evolutionary outcomes permitted by the environment will
almost certainly always be far larger than the number of outcomes possible from a given
evolutionary starting point.

This potential effect of past evolution has the interesting

ramification that future evolution may be rendered highly repeatable due to prior evolution
restricting outcomes, or else highly potentiating one particular outcome or set of outcomes
(Yedid and Bell 2002). Paradoxically, this means that deterministically repeatable outcomes on
this scale may be historically contingent, which can complicate interpretation.

25

An alternate view is of the highest scale, at which evolutionary outcome may be
construed to refer to the biosphere as a whole, with its full complement of filled niches and
adaptations. In this case, the relevant question of contingency is that of whether or not there
existed alternatives to the whole assemblage of the living world. There are two problems with
approaching the contingency of evolution from this perspective. First, given the possibility of
contingent intermediate states in a system with a single final possible outcome, the question is
likely unanswerable in principle while evolution is still going on. Second, there does not appear
to be any way of assessing whether or not there are possible alternate outcomes at this level
given the n = 1 problem. However, as a completely microbial biosphere does appear to be a
viable possibility, there may be at least some possible scope for contingency even at this scale
(Maynard Smith and Szathmary 1995, Gould 1996).

EMPIRICAL STUDIES OF HISTORICAL CONTINGENCY IN EVOLUTION

In my opinion, evaluating how contingent, complex, and chaotic history is can only
be done empirically … There is no philosophical superhighway to bypass the
careful historiographical study of concrete historical processes and events.
–

Aviezer Tucker (1999)

When one attempts to determine for a given trait whether it is the result of natural
selection or of chance … one is faced by an epistemological dilemma. Almost any
change in the course of evolution might have resulted by chance. Can one ever
prove this? Probably never.
- Ernst Mayr (1988)

26

The foregoing considerations strongly suggest that contingency is an intrinsic property of
evolution. However, they do not address the issue of the actual impact of contingency on the
evolutionary process, and the questions that pertain to it.

What scope, for instance, does

contingency have to impact or determine evolutionary outcomes?

To what degree does

cumulative selection along favorable paths canalize evolution toward particular outcomes? To
what extent is the evolution of particular traits, combinations of traits, ecological assemblages,
and ecological interactions contingent upon particular evolutionary or environmental histories?
Are some outcomes more contingent than others? If so, are there common features among them?
How much can the prior history of a lineage impact future evolution by restricting or potentiating
access to various evolutionary paths (causal chains) to given outcomes? Is natural selection a
robust enough search algorithm that it can always find the evolutionary path to a given outcome
no matter prior history? Even if a particular evolutionary outcome is effectively unreachable by
a particular lineage due to prior evolution history, does the extremely high number of lineages in
the living world with distinct evolutionary histories guarantee that certain outcomes will always
be found?
Addressing these fundamentally empirical questions requires investigation of three major
impacts that contingency has been predicted to have on evolution. First, as Gould suggested,
contingency should reduce evolutionary repeatability (Gould 1989, 2002). Second, evolutionary
outcomes should be history-sensitive, meaning they should be impacted by differences in
evolutionary and environmental history. Notably, variation in repeatability may arise from and
be explicable in terms of the sensitivity to these differences. Of course, different outcomes may
also arise from contingency inherent to microevolutionary processes that may leave no obvious

27

causal point of departure in the macroevolutionary record. Third, because contingency, while
causal, is non-deterministic, it cannot accelerate the occurrence of particular evolutionary
outcomes in the way that cumulative natural selection can (1996a,b). Consequently, the more
contingent an adaptation is upon a particular history of multiple mutations from a given starting
point, the longer the evolution of that adaptation may be expected to take (Foote 1998). While
related to the issue of repeatability, this impact deals more properly with evolutionary timing.
Evaluation of these predictions is essential to understanding the actual importance of
contingency in the evolutionary process. Over the past few decades, numerous empirical studies
have begun to examine these predictions at both the macroevolutionary and microevolutionary
levels, and using a variety of systems, including adaptive radiations on islands, and evolution
experiments with bacteria and digital organisms. The findings as a whole suggest that the
importance of contingency in evolution is not a simple question, likely reflective of the variation
of contingency depending on the exact properties of the system under study. In the following, I
review a number of these studies, focusing mainly on macroevolutionary studies. I will then
focus on the use of evolution experiments with microorganisms in research into contingency,
which will provide a basis for introducing the system and phenomenon addressed by the
experimental work that is presented in chapters 2 and 3.

Macroevolutionary Studies
Macroevolutionary studies constitute the broadest array of work yet done on contingency
in evolution, and have included investigation of all three predictions from contingency. Likely
reflective of Gould’s influence on the study of contingency, most macroevolutionary studies
have involved examination of repeatability on either a grand or a small scale. Grand-scale

28

studies have primarily focused on examining to what degree adaptations, traits, and other
biological properties have arisen independently over the course of evolution. Conway Morris
(2003, 2010) has compiled a large number of instances of striking evolutionary convergences
among a diverse array of lineages. Based on these findings, he has concluded that convergence
on a limited number of viable outcomes is the principle feature of evolution, and that
contingency therefore has little or no impact on the evolutionary process. However, Conway
Morris’ studies have been informal, have lacked a systematic and unbiased approach to
examining the actual relative prevalence of convergence in evolution, and have not rigorously
differentiated between convergent and parallel evolution.

Consequently, while his findings

certainly suggest that numerous biological properties can and do repeatedly evolve despite
significant differences in evolutionary history, his broader conclusions are in doubt.
Vermeij (2006) conducted a more systematic grand-scale study of major innovations,
which he defined as qualitatively new traits that either confer a fundamentally new function or
improve a preexisting function.

He examined 78 such innovations, their phylogenetic

distributions, and their times or origin. Among these innovations, Vermeij identifies 55 that
have demonstrably evolved multiple times, though he does so under loose definitions of
phylogenetic independence that dismisses the possible effects of deep homology on repetitive
evolution (Shubin et al 1997, 2009). The remaining 23 appear to have singular origins. Notably,
the singular innovations appear to have originated at far earlier time points, on average, than the
repeated innovations. Given that a significant proportion of the individual instantiations of
repeated innovations occur in small clades for which information would rapidly decay over time
due to extinction and the imperfection of the fossil record, Vermeij notes that ancient innovations
may be more likely to appear singular due to long-branch attraction and information loss

29

obscuring multiple origins. From this consideration, he suggests that current data are insufficient
to conclude that there are any truly singular innovations, though their existence cannot be
excluded on the basis of the existing data. However, given that most innovations originate in the
integration of pre-existing, repeatedly evolved components, their eventual production is likely
over long time scales. Moreover, some innovations require longer to evolve simply because they
are built on prior innovations.

Vermeij therefore concludes that significant ecological,

functional, and directional features of evolution are repeatable due to the iterative nature of
evolution and the long periods of time involved, while contingency has only a subsidiary role
that determines various details.
Dick and colleagues (2009) reached a similar conclusion in a study that addressed the
prediction that adaptations contingent upon particular evolutionary histories will take longer to
evolve. They examined the origin of two innovations in the cheilostome order of colonial
bryozoans. Cheilostomata arose more than 150 million years ago during the late Jurrasic. The
two innovations, a calcified costal shield protecting the frontal membrane, and an underlying,
water-filled ascus that corrects restrictions the shield places on function of the retractable feeding
organ, or lophophore, required prior evolution and subsequent modification of articulated
periopesial spines.

After the evolution of these spines, the emergence of the two innovations

showed long delays of 22 million years, and 33 million years, respectively. Nonetheless, the
innovations evolved multiple times independently. The authors suggested that such traits that
require particular histories or sets of histories to be actualized are ultimately very likely due to
the long periods of time involved in evolution. They are thus highly repeatable and predictable
features, with only the timing subject to historical contingency.

30

The most common approach to studying contingency’s effect on evolutionary
repeatability has been to examine small-scale, natural instantiations of Gould’s “replaying the
tape of life” thought experiment. These studies exploit situations in which evolution has played
out independently under similar conditions, from similar starting points, or both. These studies
have a number of benefits. First, they effectively permit tests of the repeatability prediction
under natural conditions.

Second, when differences between “replays” are identified, they

provide instances of contingent outcomes that can be studied for better insight into the conditions
under which contingency can drive divergence. Finally, they provide the prospect of at least
potentially identifying possible historical or environmental differences to which the different
outcomes may be causally attributed. The risk with such studies, of course, is that different
outcomes may in fact result from different selection regimes imposed by subtle environmental
differences between the “replicates”.
In one such study, Knapp et al. (2006) studied mesic grassland ecosystems that have
evolved independently in North America and South Africa.

Both systems evolved under

similar conditions subject to fire, grazing, and climatic variability, and had been previously
shown to have significant structural similarities. Knapp and colleagues found that this structural
convergence had been concurrent with convergence in the control of aboveground net primary
productivity (ANPP) by between-year variation in precipitation. However, despite this overall
functional convergence, the South African ecosystem showed greater ANPP sensitivity to
variation in precipitation during the early part of the year that could have major consequences if
climate change increases variability during these periods. The authors attribute this divergence
to intra-annual differences in soil moisture availability between the two locations. This study
shows that small differences in environment can lead to the evolution of slight divergences in

31

ecosystem function that can have potentially important long-term implications for a particular
ecosystem’s long-term stability and thereby not only its survival, but also that of its constituent
organisms.
Despite the value of studies like that of Knapp et al., rare instances of replicated adaptive
radiations on islands and in lakes are perhaps more useful and powerful study systems. First,
they often permit examination of larger numbers of “replicates”. Second, the radiations often
begin from very closely related ancestors, making it easier to examine the effect of contingency
on divergence from similar starting points. Among island systems, the Anolis lizards radiation
on Carribean islands has been best studied in the context of contingency. Assemblages of
multiple lizard species are found on the four islands of Cuba, Hispaniola, Jamaica, and Puerto
Rico. The species in these assemblages cluster into seven distinct sets of highly morphologically
similar ecomorphs that occupy similar ecological positions across the islands, with all but three
ecomorphs found on all islands. Losos et al. (1998) demonstrated that each ecomorph evolved
independently on each island, suggesting that natural selection and similar ecological
opportunities were sufficient to drive repeated evolutionary outcomes among the replicates. At
the same time, their analysis showed that ecomorphs rarely evolved multiple times on each
island, suggesting that ecomorph evolution was contingent upon whether or not the ecological
opportunity it suited had been taken. A similar pattern has been observed in the evolution of
bacterial ecotypes, suggesting a general phenomenon (Fukami et al 2007). These findings
suggested a constrained, incidental role for contingency in the radiations, with deterministic
factors dominant. Interestingly, a second genus of lizards, Phelsuma, or day geckos, that has
also experienced replicated adaptive radiations on islands in the Indian Ocean with highly similar
environmental conditions and presumably similar ecological opportunities. However, the gecko

32

radiations show significant differences in the ecomorphs evolved, the range of morphospace they
occupy, and how they partition resources relative Anolis (Harmon et al 2007, Losos 2010).
These differences suggest that significant features of the two radiations have been contingent
upon either subtle differences in the respective island environments, the prior evolutionary
histories of the genera, or an interaction of both factors (Losos 2010).
Striking radiations of holoarctic fish have also been identified in hundreds of postglacial
Canadian lakes colonized by marine ancestors within the last 12,000 years during periods of
marine inundation, providing opportunities to investigate a range of important evolutionary
issues (Schluter 1996). These radiations have involved a number of different species, but the
most intensively studied have been those involving the three-spine stickleback, Gasterosteus
aculeatus. The freshwater sticklebacks have shown substantial parallel evolution, including the
accumulation of numerous similar characteristics associated with reduced predation pressure and
exposure to new ecological opportunities (Schluter et al 2010). One commonly evolved trait is
reduced armor plating, which has been found to be due to parallel fixation of an ancestral allele
of eda, the gene for ectodysplasin (Colosimo et al 2005, Schluter et al 2010). These freshwater
populations also often display two sympatric, reproductively isolated, ecomorphs, which occupy
distinct benthic and limnetic niches (McPhail 1994, Schluter 1996). This degree of parallelism
and repeated instances of evolution of the ecomorphs strongly suggests deterministic underlying
factors, with ecological speciation driving the divergences.

However, the lakes in which

repeated stickleback divergences occurred are restricted to a small portion of the freshwater
stickleback range, outside of which only single ecomorphic populations are seen. As these lakes
are not environmentally different from those containing single ecomorphs, McPhail (1993)
hypothesized that a historical event might have been responsible for the difference. He later

33

identified the lakes with sympatric pairs as those subjected to a second marine inundation ~15002000 years after the initial submergence. Later genetic analysis of these populations by Taylor
and McPhail (2000) were consistent with a hypothesis that the sympatric pairs were deterministic
outcomes contingent upon the environmental history of repeated marine submergence providing
multiple waves of marine stickleback colonization. Indeed, the results of this study, those of the
two lizard radiations, and the observation that adaptive radiation does not always occur under
conditions in which it might be expected suggest that interplay between historical contingencies
and deterministic factors determines important aspects of adaptive radiations, including their
very occurrence (Taylor and McPhail 2000, Seehausen 2007, Losos 2010).
This interplay is also demonstrated by the findings of Emerson (2000), who investigated
the sensitivity of later evolutionary outcomes on prior evolution in Southeast Asian fanged frogs.
Her analysis of fifty monophyletic species strongly suggests that the evolution of small fang size
later drove repeated evolution of small body size, terrestrial reproduction, and female-biased size
sexual dimorphism. This finding supports the view that prior evolution can potentiate certain
outcomes, either by constraining future evolution under certain conditions to a limited range of
general outcomes, or by altering the morphological “cross-section” exposed to deterministic
environmental and ecological forces.
Emerson’s findings, as well as those from the adaptive radiations support the conclusions
that that environmental and evolutionary historical contingencies can actually promote repeated
evolutionary outcomes in particular instances by providing a context that is effectively a part of
the complement of deterministic forces in evolution. Supportive of this, Yedid and Bell (2002),
demonstrated that digital organisms lacking a prior history were able to evolve to a wide variety
of unrepeatable, well-adapted end states even in an extremely simple environment, but became

34

more constrained to particular end states as they accumulated history. Ultimately, these findings
imply that contingency determines what is repeatedly evolved, which poses a problem for
interpreting repetition as evidence of evolution as deterministic. Indeed, it would seem that this
interpretation would be valid on a large scale only if one could demonstrate that the repetition
was of all outcomes possible for the environment. If the repetition were of a constrained set of
the full possible range, then it would actually support an interpretation of the importance of
contingency, as the evolutionary constraint that underlay the repetition would have been due to
prior history having reduced the range of probable outcomes. On a small scale, repetition would
be supportive of the deterministic primacy of natural selection only if the repeated outcome were
to correspond to the single most fit end state available in the environment. What these studies
generally suggest, therefore, is that, beyond certain almost inevitable general features of living
things, contingency and determinism have been tightly intertwined in evolution. History, by
eliminating evolutionary totipotence, can produce starting points from which the range of
outcomes is sufficiently constrained that certain outcomes are almost certain to result from them.
It appears that robust process explanations can be used to predict and explain evolutionary
outcomes given starting conditions, but those conditions are contingent, limiting predictability of
evolution in the case of most particular instances.

Evolution Experiments
The rise in interest in the role of contingency in evolution fortuitously coincided with the
development of evolution experiments with fast-growing microorganisms, including bacteria,
algae, viruses, and unicellular fungi as an accepted means with which to address fundamental
evolutionary questions (Elena and Lenski 2003).

35

Indeed, these experimental systems have

proven to be among the best and most powerful systems with which to examine the issue of
contingency in evolution.

While lacking the complexity and temporal scope of

macroevolutionary subjects, evolution experiments with microorganisms have numerous
advantages that compensate for these deficits. Microorganisms are easy to handle, grow, and
maintain in large populations under laboratory conditions, eliminating many of the logistical
hurdles associated with field-based macroevolutionary work. This logistical simplicity also
allows populations to be propagated over long periods of time. The value of this capacity is
amplified by the rapid generation times of the organisms, which permits many generations of
evolution over short time scales. A typical experiment can therefore be used to study hundreds
to thousands of generations of evolution.
Most of the model microorganisms used are clonal, meaning that multiple populations
can be founded from defined and isogenic points. Combined with the small space needed to
maintain large microbial populations, this means that high levels of true evolutionary replication
may be achieved. With microorganisms, significant stretches of the same tape of life really can
be replayed multiple times, and done so simultaneously. Due to the defined starting point,
selection in these replicates will act upon de novo variation introduced by mutation, rather than
upon standing variation. Samples of evolving populations can be frozen to provide a complete
fossil record of evolution. These frozen samples remain viable effectively indefinitely, allowing
evolutionary changes to be identified and examined by direct comparison of ancestral,
intermediate, and evolved organisms. Clonality also means that stable genetic markers may be
used to differentiate organisms from different time points in an evolutionary history, which in
turn permits fitness changes to be tracked by use of direct competition assays (Lenski et al 1991).
Moreover, the wealth of genetic tools, including advanced sequencing technologies, permit ready

36

identification and manipulation of genetic changes, which may be directly linked to phenotypic
changes (Herring et al 2003, Bentley 2006, Hegreness and Kishony 2007, Barrick et al 2009,
Barrick and Lenski 2009).

This capacity affords a far great precision in examining and

comparing evolutionary trajectories and outcomes than is possible the natural world.
A final major advantage to these experiments is the high level of control that is possible.
As already mentioned, unlike in nature, the genetic composition of founding populations can be
controlled to eliminate confounding initial variation. Conversely, initial genetic states and prior
evolutionary history can be manipulated to examine their effect on subsequent evolution
(Travisano et al 1995). Similarly, important population genetic factors such as mutation supply,
population size, and bottleneck stringency may be manipulated (Chao and Cox 1983, Burch and
Chao 2000a, b, De Visser et al 1999, Elena et al 2001, Perfeito et al 2007). Moreover, both
biotic and abiotic aspects of the environment of the experiment can be manipulated to examine
their contingent effects on evolution (Levin and Lenski 1985, Bennett and Lenski 1993,
Bohannan and Lenski 2000, Sleight et al 2006, Fukami et al 2007, Meyer and Kassen 2007).
Consequently, extrinsic features of the environment may be held highly stable and constant,
eliminating the effects of environmental variability that can confound the examination of other
aspects of evolution.
The E. coli Long-Term Evolution Experiment
Among the microbial evolution experiments that have been used to examine contingency
in evolution, the E. coli Long Term Evolution Experiment (LTEE) is the longest running and
most intensively yet undertaken. Begun in 1988, the experiment was explicitly conceived and
designed to examine not only evolutionary repeatability, but also the dynamics of evolutionary
change and the relationship between phenotypic and genomic change (Lenski et al 1991, Lenski

37

2004).

Twelve replicate populations were founded from the same E. coli B clone, REL606

(Lenski et al 1991). Save for a neutral single nucleotide polymorphism in half of the populations
+

that confers the capacity to metabolize the sugar arabinose (Ara ), all twelve populations were
initially genetically identical. Every day, 1% of each population is diluted into 9.9 mL of DM25,
a minimal salts medium amended with 139 mM glucose as a carbon and energy source (Davis
and Mingioli 1950). Under this these conditions, the populations each undergo approximately
6.64 doublings during a 24-hour period, resulting in a 100-fold expansion to a daily maximum
population size that was initially ~5 x 108 (Lenski et al 1991, Lenski 2004).

Each population

has experienced more than 50,000 generations of evolution over the 23 years of the experiment
so far. Samples of each population have been frozen every 500 generations throughout the
experiment.
The LTEE examines evolution under highly simplified conditions (Lenski et al 1991,
Lenski and Travisano 1994, Lenski 2004). The founding clone is strictly asexual and lacks
natural competence. Moreover, its genome lacks any plasmids or active prophages (Jeong et al
2009, Studier et al 2009). The populations are each maintained in separate culture vessels under
strict isolation from each other and outside organisms, eliminating the possibility for gene flow.
Evolution in the LTEE is therefore purely by the operation of genetic drift and natural selection
upon variation generated de novo by mutation (Lenski et al 1991, Lenski 2004). Moreover, the
populations are evolved under stable environmental conditions of 37°C with aeration. The only
perturbations the populations experience are the highly regular “seasonality” caused by glucose
exhaustion and bacterial alteration of the medium, and the brief period they spend at room
temperature without aeration which undergoing the daily transfers.

38

Consequently, the LTEE

permits the examination of the isolated effects of contingency inherent to the core evolutionary
processes of mutation, drift, and natural selection.
Indicative of the power of natural selection to drive repeated evolutionary changes, the
populations have shown significant and pervasive parallel evolution over the course of the
experiment. All twelve populations have evolved significantly higher fitness relative to the
ancestor under the conditions of the experiment. All showed a pattern of rapid fitness gain
during the early generations that then decelerated markedly (Lenski and Travisano 1994, Lenski
2004). Other parallels have included reduced lag times upon transfer, increased maximal growth
rates on glucose, increased average cell size, and decreased peak population density relative to
the ancestor (Lenski and Travisano 1994, Vasi et al 1994, Lenski 2004, Philippe et al 2009). All
populations have experienced metabolic specialization, as indicated by the loss of some or all of
their capacity to grow on a number of substrates (Cooper and Lenski 2000). In the most striking
case, all populations lost the capacity to metabolize ribose due to similar IS150-mediated
deletion events (Cooper and Lenski 2000, Cooper et al 2001). The populations also have also
experienced parallel changes in gene expression (Cooper et al 2003), global protein profiles
(Pelosi et al 2006), global regulatory networks (Philippe et al 2007, Crozat et al 2011), epistatic
interactions that affect the CRP regulon (Cooper et al 2008), and resistance to phages T6* and λ
(Meyer et al 2010). Moreover, ten of the populations have evolved increased DNA supercoiling
(Crozat et al 2005, 2010). At least three genes have experienced substitutions in all twelve
populations (Woods et al 2006, Cooper et al 2001), and a number of others have experienced
substitutions in multiple populations (Crozat et al 2005, Cooper et al 2003, Pelosi et al 2006,
Woods et al 2006, Philippe et al 2009, Barrick et al 2009, Crozat et al 2010). Given that most
genes across the populations have not accumulated any mutations (Lenski et al 2003), these

39

genetic parallelisms likely reflect the action of natural selection. Indeed, a number of these
parallel mutations have been verified to confer increased fitness under the experimental
conditions (Cooper et al 2001, Barrick et al 2009, Philippe et al 2009).
Amidst this remarkable parallelism, the LTEE populations have also diverged
evolutionarily. While there is parallelism in a number of the adaptive genetic changes identified
across the populations, there are still variations in the full complement of adaptive changes in
each population (Woods et al 2006, Stanek 2009). Moreover, even when the same genes are
mutated in multiple populations, the details and locations of the changes usually differ between
the populations (Crozat et al 2005, Cooper et al 2003, Pelosi et al 2006, Woods et al 2006,
Cooper et al 2001). Each population accumulated substantial and divergent gross genomic
changes, including IS insertions, inversions, and deletions (Papadopoulos et al 1999). A number
of the populations have evolved mutator phenotypes due to mutations in methyl-directed
mismatch repair pathway and other genes, resulting in accelerated genomic divergence
(Sniegowski et al 1997, Barrick and Lenski 2009, Barrick et al 2009). Significant betweenpopulation variation in relative fitness was detected after 2000 generations (Lenski et al 1991),
and persisted even after 8,000 more generations of evolution (Lenski and Travisano 1994). A
hundred-fold higher between-population variation in correlated fitness change on maltose and
lactose than on glucose was also detected after 2000 generations (Travisano et al 1995b). After
20,000 generations, decay in unused catabolic functions also varied greatly between populations
(Cooper and Lenski 2000).
Most of the divergence observed between the populations suggests that contingency has
played a relatively constrained role in LTEE population’s evolution. Indeed, natural selection
appears to have repeatedly driven the populations to ascend adaptive peaks that are highly

40

similar in the stable, simple LTEE environment, if not in their correlated properties in other
environments (Travisano 1995b, Cooper 2000). Of course, deterministic, repeated evolution is
to be expected if either the environment affords few ecological opportunities, or if the prior
evolutionary history of the starting population either restricts or precludes discovery of most of
the available opportunities or highly predisposes evolution to exploit one opportunity.

By

design, the glucose resource environment of DM25 affords one major ecological opportunity.
The large amount of parallelism with mostly subtle divergence in the LTEE suggests that the
ancestor was predisposed to adapt to the exploitation of this opportunity, and to do so via a
number of similar evolutionary paths. However, there are at least two populations that have
shown much more significant divergence. Both have been based on finding and exploiting
ecological opportunities in DM25 that have not been readily accessible via either the majority of
available evolutionary paths, or by those most likely to be followed in the LTEE, suggesting that
these niche discoveries have been historically contingent outcomes.
One population evolved to exploit an unforeseen ecological opportunity in the LTEE
produced in part by an interaction of a sort that the design of the LTEE was intended to
eliminate. The amount of glucose present in DM25 is much lower than is typically added to
defined bacteriological media (Atlas 2004). The use of such a low amount is a byproduct of the
experimental goal of examining evolutionary dynamics and the relationship between phenotypic
and genotypic changes without unnecessary ecological complexities (Lenski et al 1991, Lenski
2004, Lenski personal comments). At the time the experiment was begun, and to some extent
still today, the easiest way to assess these dynamics involved fitness assays in which evolved
populations are competed against their ancestor (Lenski et al 1991). To permit detection of
meaningful patterns of fitness change that could be related back to the sweep of individual

41

beneficial mutations through the populations using this method, it was necessary to simplify the
evolutionary dynamics. Reducing the medium’s glucose amendment was expected to achieve
this simplification in two ways. First, by reducing the maximum size of the population, it was
expected to reduce the possibility of clonal interference, a phenomenon in which multiple
mutations arise simultaneously in different, competing lineages within the population (Gerrish
and Lenski 1998, de Visser and Rozen 2006). Second, it was hoped that reduced glucose
availability would restrict the opportunity for evolution of persistent ecological complexity based
on cross-feeding or other bacterial niche construction interactions (Helling et al 1987, Tuner et al
1996, Pfeiffer and Bonhoeffer 2004).

While low glucose amendment did reduce these

complicating factors, neither was fully eliminated (Barrick and Lenski 2009).
The remaining potential for crossfeeding interactions, combined with the seasonality of
the LTEE environment, afforded an open ecological opportunity that one population, Ara-2, has
evolved to exploit. Ara-2 contains two monophyletic ecotypes that are referred as “S” and “L”,
based on their consistent production of relatively small and large colonies, respectively (Rozen
and Lenski 2000, Rozen et al 2005).

The two ecotypes originated prior to generation 6,000,

and have stably coexisted since then due to negative frequency-dependent selection based on
ecological differentiation between the two. In the mixed population, L ecotype cells display
higher fitness during growth on glucose, but suffer higher mortality during the seasonal
starvation phase that occurs following glucose exhaustion. S ecotype cells, by contrast, have
lower fitness on glucose, but display increased fitness and decreased mortality during starvation
in co-culture with L. This difference reflects S cells having improved capacity to scavenge
resources from L cells that lyse during starvation (Rozen et al 2009). The scavenging capacity of
S appears to have arisen in part, due to a loss of activity in RpoS, the stationary phase sigma

42

factor (Rozen et al 2009). Defects in RpoS have previously been observed in scavenging
phenotypes that evolve as part of the GASP response, and are typically observed to confer
improved substrate range, increased competitiveness at low substrate concentration, but reduced
competitiveness at high concentration (Finkel 2006). The genetic basis of this activity change
has yet to be determined, but it evolved very early in the S ecotype’s history. Indeed, the
existence of the Ara-2 balanced polymorphism and exploitation of the starvation season
opportunity appears to have been contingent upon this loss of RpoS activity. Similar loss of
RpoS activity has not been observed in any other LTEE populations. Its selective benefit, and
thereby the S ecotype, therefore appear to be likewise contingent upon the L ecotype having
improved growth on glucose via a mechanism that led to increased mortality during starvation
(Rozen 2009). The evolutionary divergence of Ara–2 from the other populations therefore
appears to have been an historically contingent product of multiple coincident events that opened
an ecological opportunity based on niche construction and environmental seasonality, and then
permitted exploitation of that opportunity.
This dissertation is concerned with the second population known to have evolved in a
manner highly divergent from the pattern of the other LTEE populations. This population is
known as Ara–3. Like Ara–2 it evolved to take advantage of an ecological opportunity not
easily accessed. Unlike Ara–2, however, Ara–3’s divergence was based on discovery of a preexisting niche.
An Open Ecological Opportunity
From the beginning of the experiment, the LTEE has contained an open ecological
opportunity. In addition to 139 µM glucose, the DM25 medium in which the populations are
propagated also contains 1700 µM citrate. However, this abundant potential second carbon

43

source is inaccessible to the bacteria. Though E. coli has a complete TCA cycle that permits
internal metabolism of citrate during aerobic growth on other substrates (Lara and Stokes 1952),
it is incapable of transporting and using citrate as a carbon and energy source under oxic
conditions. Moreover, this characteristic is highly stable. Spontaneous mutants of E. coli
capable of aerobic growth on citrate (Cit+) are extraordinarily rare, with only one having ever
–

been reported in the literature (Hall 1982). Consequently, the Cit phenotype has long been
regarded as one of the key defining characteristics of E. coli as a species (Koser 1923, 1924,
Scheutz and Strockbine 2005).
History of the Inclusion of Citrate in DM Medium
While citrate constitutes an open ecological opportunity in the LTEE, it was not included
in DM25 for that purpose. Citrate has been a standard component of the DM basal salts solution
since the medium was first developed and described (Davis and Mingioli 1950, Carlton and
Brown 1981) (Table 1.1). DM developed out of a medium formulated by Bernard Davis for use
in his experiments with the penicillin method of enriching for auxotrophic mutants of E. coli W
(Davis 1949). This method exploits the fact that penicillin kills only actively dividing cells,
permitting its addition to a culture growing in minimal medium to select for non-dividing
auxotrophs. Davis sought a defined medium in which auxotrophs could not grow, cultures could
4

be started from small inocula (10 to 10 cells), and proliferating cells would be efficiently killed
by penicillin. Davis notes that satisfaction of the second and third criteria required increasing the
concentrations of potassium and magnesium above those found in previously established E. coli
media such as F Medium (Cohen and Anderson 1946) and Medium 12 (Sahyun et al 1936).
During this development, Davis also experimented with citrate amendment, and noted that, “A
small amount of sodium citrate (0.01%) was also found to substitute for the high concentrations

44

+

++

of K and Mg , and was sometimes added to the minimal medium since it augmented the
bactericidal rate of penicillin.” The initial recipe was later refined, and a reduced concentration
of citrate was reported as a standard ingredient in the DM recipe reported by Davis and Mingioli
(1950). The penicillin method articulated by Davis (and Joshua Ledeberg, who developed his
own version independently), soon came to enjoy wide use. As a consequence, DM came to be an
accepted standard minimal medium amenable for use in a variety of experiments with a number
of different organisms. This wide usage entailed occasional alterations (Vogel 1955). Indeed,
the canonical version of DM currently used most widely is a variant developed for use with
Bacillus subtilis that contains 33% less KH2PO4 (Carlton and Brown 1981).
The biological role of citrate in DM medium is related to iron availability. Under
oxic conditions at neutral pH, iron predominantly occurs as an insoluble and biologically
inaccessible ferric hydroxide polymer. Under conditions typical of bacteriological media, the
equilibrium concentration of the soluble free ferric ion is ~10

-12

µM (Hartmann and Braun

1981). Citrate binds free ferric ions to form stable, soluble ferric dicitrate chelates. This
chelation activity increases the amount of iron in solution and renders it accessible to E. coli,
which has a dedicated, high affinity ferric dicitrate transport system (Cox et al 1970, Frost and

45

Component

K2SO4
MgSO4•7H2O
(NH4)SO4
NaHPO4•7H2O
KCl
K2HPO4
NaNH4HPO4•4H2O

Table 1.1 │ Development of DM Medium
DM for B.
ProtoDM
E Medium
subtilis
DM
Davis
Davis and Vogel 1955
Carlton
1949
Mingioli
(E
and
1950
Brown
medium*)
1981
1g
–
–
–

DM25
Lenski et
al 1991

–

0.025 g

0.1 g

0.2 g

0.1 g

0.1 g

–

1g

–

1g

1g

6g

–

–

–

–

4g
–

–
7g

–
–

–
7g

–
7g

–

–

3.5 g

–

–

–
3g
10 g
2g
KH2PO4
Thiamine
–
–
–
–
Sodium citrate
–
0.5 g
2g
0.5 g
Adjusted pH
7.6
7.0
–
–
Ammonium lactate
2g
–
–
–
Glucose
–
5.4 g
5g
5g
All quantities given are for 1 L of medium
*Values given for 1x dilution. Standard recipe is for 50x stock solution.

46

2g
0.002 g
0.5 g
–
–
0.025 g

Rosenberg 1973, Hussein et al 1981, Härle et al 1995). E. coli can also acquire iron using an
alternate ferric enterochelin uptake system (Cox et al 1970, Langman et al 1972), and is capable
of growing quite well in citrate-free media such as M9. However, iron solubility is affected by
+

+

+

+

Na and K concentrations, and a Na to K ratio of 3:1 or higher has been reported to be
necessary to maintain the solubility of iron at a level that permits efficient iron acquisition by E.
+

+

coli in the absence of citrate (Langman et al 1972). While M9’s Na to K ratio of 3.2:1 fits
this criterion, DM’s is 1:18.6, suggesting that iron availability is likely low in the absence of
citrate. E. coli can sometimes be cultured in DM medium without citrate, but anecdotal reports
suggest that these cultures are prone to periodic instability and failure to propagate during serial
transfer (Lenski personal communication, Blount personal observation). This instability could
be explained by the cultures experiencing greater sensitivity to normal fluctuations in iron
6

concentration in DM without citrate .
6

It is interesting that iron levels in the water used by Davis for medium preparation were

probably higher than they are in more modern laboratories. Davis developed the penicillin
method during his time as the head of the US Public Health Service Tuberculosis Research
Laboratory (Lochhead and Menninger 1949, Maas 1999). Davis’ lab was located in the Kipps
Bay Yorkville Health Center building on the campus of the Cornell University Medical Center.
This building was constructed in 1937 with iron water pipes (Lewis Wetstein, former head of
Cornell Medical Center Facilities Operations Department via personal communication with
Elizabeth M. Shepard, Acting Head of Medical Center Archives of New York-Presbyterian
Hospital/Weill Cornell). Copper pipes did not come into common use until the late 1940s, with
modern plastic plumbing much later.

It is unclear how much this technological shift has

47

While the need to ensure iron bioavailability explains the presence of citrate in DM25,
there remains the question of its high concentration in the medium. This is a particularly
important question, as it addresses the size of the ecological opportunity presented by the citrate
in the medium used in the LTEE. Frost and Rosenberg (1973) determined that the concentration
of dissolved ferric dicitrate rapidly declines as the ratio of citrate to total iron falls below 20:1,
and that the E. coli ferric dicitrate transport system ceases to function at citrate concentrations
below 10 µM. Moreover, they reported maximal growth at a citrate concentration of 1000 µM in
a medium supplemented with 30 mM glucose. The level of citrate found in DM25 is therefore
significantly higher than is necessary to prevent iron limitation even in a medium with far more
glucose. This great excess of citrate in DM25 may be explained by three basic points. First, the
formulation of DM was designed specifically for use in the penicillin method of obtaining
auxotrophic mutants, and Davis makes clear that the addition of citrate was made purely to
improve the killing efficiency of penicillin (1949). Given that E. coli was at the time thought to
be metabolically inert toward exogenous citrate, it is unlikely that Davis was concerned with
reducing citrate to the lowest level necessary to have the desired effect.

Second, the DM

minimal salt recipe provided by Davis and Mingioli (1950) was developed for supplementation
with 11.1 mM glucose. At this level of glucose, the citrate concentration is slightly less than
three times higher than strictly necessary. Third, the ferric dicitrate iron acquisition system was
not identified until 1973 (Frost and Rosenberg), while Davis and Mingioli wrote in 1950. DM
was developed for a particular experiment under particular conditions, and prior to the discovery
of the biological role of citrate in the medium or of the concentrations strictly necessary for it to
mattered, but it does not appear to be often taken into account during preparation of growth
media from recipes developed at much earlier points in time.

48

fulfill that role, and its formulation reflects this. Unless required for a particular experiment or
other pragmatic reason, the recipes of growth media are rarely revised, and this is particularly
true of minimal salts solutions. It is therefore unsurprising that the level of citrate in DM was
neither changed after the medium came into common use in experiments beyond that for which it
was originally intended, nor adjusted following the identification of the ferric dicitrate iron
acquisition system.

The high concentration of citrate in DM25, and the large ecological

opportunity it has presented in the LTEE, is therefore a matter of historical accident and
7

conservative lab practice .
E. coli has the Potential to Evolve a Means of Exploiting the Citrate Opportunity
The only known barrier to aerobic citrate utilization by E. coli is the inability to transport
citrate into the cell under oxic conditions (Hall 1982, Reynolds and Silver 1983, Pos et al 1998).
+

Indeed, numerous atypical Cit E. coli have been isolated from agricultural and clinical samples,
all of which have been found to harbor plasmids that encode foreign citrate transporters (Ishiguro
et al 1978, Ishiguro et al 1979). Hall (1982) reported the only known case of a spontaneous Cit
7

+

Peter Reeves of the University of Sydney and John Roth of the University of California at

Davis both offered an alternate historical origin to that described (personal communications).
According to this explanation, DM was developed from a Salmonella medium that was originally
kept in 50x stocks. At such a concentration, it was found that the amount of citrate had to be
increased to keep MgSO4 in solution. This explanation is likely incorrect, and appears to be
based upon confusion of DM, which is specifically formulated for a 1x concentration, with the
similarly constituted E medium (Table 1.1), which Vogel (1955) developed from DM following
his use of it during a collaboration with Davis (1952).

49

mutant of E. coli. His findings suggested that either a complex mutation, or multiple mutations,
activated one or more cryptic genes that expressed a citrate transporter, though he was unable to
identify the genes involved. E. coli is also known to encode a dedicated citrate transporter, CitT,
which permits fermentative anaerobic growth on citrate if a cosubstrate is available to provide
reducing power (Lutgens and Gottschalk 1980). Pos et al. (1998) demonstrated that high-level
constitutive expression of citT from a multi-copy plasmid was sufficient to confer a Cit

+

phenotype in the genetic background of the K-12 strain their studied. However, in its native
genomic context, citT is present in a single copy, and does not appear to be expressed under oxic
conditions (Pos et al 1998).
These findings suggest that E. coli has the potential to evolve a Cit+ phenotype.
Presumably, REL606, the ancestor of the LTEE populations shared this potential. However, the
+

extreme rarity of Cit mutants of E. coli makes clear that this evolutionary potential is not easily
+

actualized. Moreover, Hall’s (1982) findings indicate that the evolution of Cit is complex, and
requires multiple steps.
+

Cit Evolves in the Ara–3 Population
+

While the evolution of a Cit variant in one or more of the LTEE populations was
considered to be a possibility during the experiment, no additional effort was made to test for
that potential beyond seeing what happened in the LTEE itself. However, screens for Cit

+

variants occur regularly as incidental byproducts of experimental procedures. For example, prior
to the freezing of population samples every 500 generations, dilutions of each population are
spread on minimal glucose (MG) and minimal arabinose (MA) agar plates. These media are

50

both derived from DM, and therefore contain substantial amounts of citrate. However, these
platings are not especially effective as a means of screening for rare Cit
populations should show grown on MG, and half on MA.
+

+

mutants, as all

However, they do present an
–

incidental means of screening for Cit variants in the six Ara populations that should not
normally produce colonies on MA. The best screen has been the use of DM25 itself. Due to the
+

high concentration of citrate, the evolution of a Cit variant that could effectively exploit the
citrate opportunity would be signaled by a visually obvious increase in turbidity. Of course, as
part of precautions against contamination, turbidity checks are made of each population at the
time of transfer. If a population is found to display elevated turbidity, a sample is plated on TA,
MA, and MG medium, and the resulting colonies are scrutinized for evidence of outside
contamination. In those cases in which increased turbidity occurs and no evidence of outside
+

contamination is detected, tests may then be done for Cit variants.
On Saturday January 25, 2003, Tim Cooper performed LTEE transfer 4971. During the
turbidity check, he found that population Ara–3 was markedly more turbid than the other
populations.

In accordance with protocol, he spread samples of Ara–3 and Ara–4 (for

comparison) on TA plates to check for contaminants and to verify the elevated turbidity. Though
significantly more colonies grew from the Ara–3 sample than from that taken from Ara–4, all
displayed morphology, color, and odor consistent with E. coli. Given that all prior instances of
elevated turbidity had been due to contamination, however, the population was terminated
despite the lack of definitive evidence of contamination.
The Ara–3 population was restarted on March 21, 2003 from the sample frozen at
generation 33,000. On April 10, 2003, Neerja Hajela, while performing LTEE transfer 5044,

51

reported that Ara–3 again showed increased turbidity. Refrigerated samples of the population
from the April 8 and 9 transfers were plated on TA. Again, no obvious contaminant colonies
were observed, though samples from the April 9 transfer showed a ~3.5-fold increase in colony
number. Ten randomly chosen colonies were tested for resistance to phages T5 and T6, and all
r

s

showed the T5 , T6 phenotype characteristic of REL606-derived clones. Similarly, Gram stains
showed only Gram-negative coccal-rods consistent with E. coli. The population’s turbidity
continued to rise over subsequent transfers, peaking at approximately 10-fold higher than the
+

other populations on April 13. At this point, the evolution of a Cit variant was considered, and
samples from the April 11 transfers were plated on minimal citrate (MC) plates, and inoculated
into DM0 (DM25 lacking glucose). The DM0 cultures showed turbidity consistent with growth
after 48 hours of incubation, and well-formed colonies were observed on the MC plates at 72
hours, demonstrating that there were members of the population capable of using the citrate in
the medium. To provide a further check, population samples from April 8 through 14 were
plated on TA, and 20 to 400 typical colonies for each day were tested on TA, MC, and Simmon’s
citrate agar. While none of the clones checked for April 8 showed evidence of growth on citrate,
70% to 97% of clones tested from subsequent sample dates showed both growth on MC and a
+

positive reaction on Simmon’s citrate agar. DNA sequencing later showed that Cit clones
isolated from the population have the same mutations in the pykF and nadR genes possessed by
clones from earlier generations of the Ara–3 population, and which distinguish the population
+

from the other LTEE populations (Woods et al 2006). These results established that a Cit
variant of E. coli had indeed evolved in the Ara–3 population.

52

Research Discussed in this Disseration
The following chapters of this dissertation detail the first stages of research into the origin
and evolution of this novel metabolic innovation, which has been the most dramatic change
+

observed over the course of the LTEE so far. In Chapter 2, the Cit phenotype, the essentials of
its invasion and rise to high frequency, and the initial impact of its numerical dominance on the
+

population’s structure and ecology are described in more detail. Notably, the rise of Cit variants
to high frequency did not result in the extinction of clones displaying the ancestral Cit

–

–

phenotype. Instead, Cit variants persisted in the population until at least generation 40,000,
+

indicating that the evolution of Cit was also a cladogenic event.
The larger thrust of Chapter 2 deals with the question of the role of historical contingency
+

+

in the evolution of the Cit phenotype. Certainly, Cit evolution fits two expected features of a
historically contingent adaptation (Foote 1998). First, it evolved only once among the twelve
populations. Second, its evolution was long-delayed, occurring 16 years into the experiment,
and after more than 30,000 generations of evolution. In the core of Chapter 2, I describe a series
of experiments in which I replayed the tape of Ara–3’s evolution from various time points in its
+

history. These replay experiments establish that the Cit trait is indeed a historically contingent
innovation that was causally dependent upon the prior occurrence of one or more mutations. The
+

prior mutation(s) potentiated the actualization of the Cit trait by increasing the rate of mutation
+

to Cit from the immeasurably low ancestral rate to a rate that, while still low, rendered the
function mutationally accessible.

53

+

Cit therefore offers a remarkable opportunity to study the evolution of a demonstrably
historically contingent innovation in a context in which the details of the underlying history may
be precisely determined. Evolution in the LTEE is restricted to the operation of only the core
evolutionary processes of mutation, drift, and natural selection. In this context, evolutionary
history is effectively the accumulation of mutations.

Due to the advent of efficient, cost-

effective whole-genome sequencing technologies, all mutations that have occurred along a line
of descent may be identified (Bentley 2006, Hegreness and Kishony 2007, Barrick et al 2009,
Barrick and Lenski 2009). Combined with the capacity to move mutations into relevant genetic
backgrounds (Herring 2003), this technology holds the prospect of not only discovering the
putative historical sequence of events underlying Cit

+

evolution, but also experimentally

reconstructing that sequence to empirically identify the causal contributions of each event to the
innovation’s origin. While this approach has been done previously at the level of protein
evolution (Weinreich et al 2006, Ortlund et al 2007, Bridgham et al 2009), it has not been done
at the whole-genome level. Such experiments will not only help to better understand historical
contingency, but also the origin of novel innovations from the modification of pre-existing
structures, traits, and genetic information (Mayr 1960, Jacob 1977, 1982, Pigliucci 2008).
Chapter 3 describes the steps taken to realize this opportunity. The focus of the chapter
are the findings of an extensive program of whole-genome sequencing of dozens of clones,
including both Cit
population.

+

–

and Cit , isolated from numerous points in the history of the Ara–3

The resulting data allowed reconstruction of the evolutionary history of this

population. The phylogeny shows that the population was heterogeneous for most of its history.
–

Three Cit clades had emerged by 20,000 generations that co-existed until after the evolution of

54

+

+

the Cit innovation. The Cit clones form a coherent, monophyletic clade within one of these
–

+

three Cit clades. The data also led to the discovery that the Cit function’s immediate cause
was a tandem gene duplication event that produced a new regulatory module in which the citrate
transporter gene, citT, is placed under the control of a promoter that directs expression during
aerobic growth. Increased copy number of this regulatory module subsequently refined the
+

+

initially weak Cit phenotype, leading to the rise of Cit variants to numerical dominance in the
population. Further analysis of this mutation and the distribution of potentiated clones within the
population phylogeny suggest that potentiation involved at least two mutations, one of which
+

appears to have improved expression and thereby potentiated the evolution of Cit through
functional epistatic interactions.

EPILOGUE

Increased Concern for Historical Contingency Has Developed in Parallel Across Multiple
Disciplines, and in the American Culture at Large

The increased interest in historical contingency that arose in the later 20th century was not
limited to biology, but instead appears to have been a general phenomenon across many fields of
science, the humanities, and popular culture. Here I will give a quick overview of this more
general phenomenon, which would seem to be ripe for further study by those who might be so
inclined. I make no claim to understand the origins of this increased interest, though I suspect
the rise of science fiction and the increased historical self-awareness of western culture in the

55

post-World War I era has something to do with it.

Nonetheless, I find the phenomenon

remarkable, and feel it proper to take note of it here. If nothing else, the pervasiveness of interest
in contingency in academia and the broader culture will no doubt serve to spur further
widespread work on the subject in a variety of areas.
During the period during which Gould was advocating for a better appreciation of
contingency in evolution, similar interest in the concept arose in psychology. As previously
suggested, because contingency underlie the occurrence of a single outcome from among
multiple possible outcomes, it is related to counterfactual historical thinking, or the consideration
of alternate historical realities that is the source of the power of what historian and writer Harry
Turtledove (2001a) has called “… those two mournful little words … What if …” (Lewis 1986).
In 1982, the cognitive psychologist and Nobel prize-winning economist Daniel Kahneman and
Amos Tversky published a paper that suggested that counterfactual thinking is a core heuristic
feature of human thought. Counterfactual thinking has since become the subject of considerable
research (Mandel et al 2005). Counterfactual thinking appears to be particularly important in
consideration of personal history, where it often carries significant emotional content due to
linkage to regret and guilt (Ferguson 2000, Coricelli and Rustichini 2010).

Moreover, as

decisions regarding future actions often involve deciding between the potential consequences of
multiple courses of action, counterfactual thinking appears to play an important role in learning
(Kahneman and Miller 1986, and Morris and Moore 2000, Coricelli and Rustichini 2010).
As Beatty has noted (2006b), during the same period, sociology saw increased interest in
history and historical approaches to their area of study. This manifested in part in attempts to
discern boundaries between sociology and history due to the field’s interest in the sociological
dimension of history (Abrams 1983, Goldthorpe 1991, Bryant 1994). However, it has also

56

perhaps more importantly led to growth in the use of counterfactual approaches to causality in
sociological phenomena (Sobel 1995, Harding 2003, Morgan 2007).

Similar growth in

counterfactual analysis has also occurred in economics over the same period (Cowan and Foray
2002, Cartwright 2007). Interestingly, the use of counterfactuals in economics goes back to the
1950’s, during which interest in the role of history in biology first began to stir (Cartwright
2007).
Increased use of counterfactuals to trace causation, together with elevated appreciation
for contingency has also been seen in professional historiography. This movement has been
particularly noteworthy because, though Thucydides and Livy were the first known users of
counterfactuals (Thucydides 1972, Livius 1982), historians have traditionally viewed
counterfactuals with skepticism, if not disdain and outright hostility, and continue to do so
(Thompson 1978, Bulhof 1999, Ferguson 2000). Ferguson (2000) traces this hostility to four
factors. First, most historians have considered their jobs to chronicle and explicate what did
occur, and not what might have occurred. Second, strongly deterministic, even teleological
approaches to history have a long history within the field, as typified by the Hegelian and
Marxist traditions born in the 19th century. Third, even among historians who do not ascribe to
deterministic approaches, there is significant concern for the principle of causality in history that
many see as threatened by consideration of contingency, which has often been conflated with a
lack of causality (Ben-Menahem 1997). Finally, there is concern that counterfactuals encourage
reductionistic historical explanations in which contributing causal factors in historical outcomes
are mistaken for singular causes (see also Bulhof 1999, Fischer 2002).
Despite this traditional hostility, a few instances of the use of counterfactuals by
academic historians may be found during the 20th century. The earliest notable example is the

57

collection of counterfactual historical essays, including one by Winston Churchill, If It Had
Happened Otherwise (1931). Another, more serious effort is economic historian Robert Fogel’s
1964 book, Railroads and American Economic Growth, in which the impact of railroads on
America’s economic expansion was assessed by positing a counterfactual in which canal
development had continued in their stead. However, these early instances were sporadic at best,
and inspired few followers. Then, in the 1980’s and 90’s, roughly coincident with Gould’s push
for reconsideration of contingency in evolution, professional historians began to show an
increased willingness to use counterfactuals (Rosenfeld 1999, Bulhoff 1999, Fergusson 2000,
Fischer 2002). This has been true of both works designed for the general public (Crowley and
Ambrose 2000), as well as more academic works, in which counterfactuals have been
increasingly considered for use as legitimate historiographic tools that help to provide better
appreciation of the role of contingency in history (Reisch 1991, Ferguson 2000, 2007). It is
unclear if this growth was in any way tied to Gould’s work, though it is interesting that he
specifically mentioned McPherson’s explicit consideration of contingency’s role in the Civil War
in his magisterial The Battle Cry of Freedom (1988). Reciprocally, Fergusson has an extended
discussion of Gould’s ideas in the introduction of Virtual History (2000). While the interest in
contingency in history and biology may have initially developed independently, they are clearly
8

now intertwined to some degree .
Perhaps the most obvious sign of the increased interest in the contingent nature of history
during the late 20th century, however, is to be found in the enormous growth of “alternate
history”. Alternate history is a sub-genre of science fiction that originated in the 1930’s and
8

I can assert this with certainty, at least on one side of the ledger, as I have routinely consulted

counterfactual historical works during the course of the research I report in this dissertation.

58

focuses on exploration of various “what if” scenarios of worlds in which history played out
differently than in reality (Rosenfeld 2002, see www.uchronia.net for an exhaustive listing).
Examples include the persistence of the Roman empire (Silverberg 2003), complete destruction
of the European population by the Black Death (Robinson 2002), avoidance of the American
Revolution (Sobel 1973, Dreyfuss and Turtledove 1997), a Confederate victory in the American
Civil War (Moore 1953, Turtledove 1997), and an Axis victory in World War II (Dick 1962,
Harris 1992, Turtledove 2003a), with these last two being perhaps the most popular (Fergusson
2000, Rosenfeld 2002). Alternate histories commonly feature an explicit “point of divergence”
from which the altered timeline under exploration diverged.

In many cases, the point of

divergence traces to a change in seemingly inconsequential event, in which case it is sometimes
9

10

called a “Jonbar Hinge” , or a horseshoe nail . For instance, Harry Turtledove’s “Southern
Victory” series (Turtledove 1997, 1998, 1999, 2000, 2001b, 2002, 2003b, 2004, 2005, 2006,
2007) describes an alternate world in which the Confederacy won the Civil War, and traces
world history through to a Union victory in an alternate World War II. The Jonbar Hinge of this
universe takes the form of a Southern soldier picking up a dropped copy of Lee’s Special Order
191 to the Army of Northern Virginia, when in reality this copy was found by Union soldiers,
leading to the critical battle of Antietam (McPherson 2002). By way of another example, in Ray
Bradbury’s highly influential short story, “A Sound of Thunder” (1952), the death of a single

9

The term Jonbar Hinge comes from Jack Williamson’s 1938 novella, “The Legion of Time”, in

which history leads to either a utopian or dystopian future depending entirely upon whether a 12
year old boy named John Barr happens to pick up a magnet or a pebble while playing one day.
10

After the nursery rhyme quoted earlier in this chapter.

59

butterfly in the Jurassic has subtle, but significant effects on human history. While alternate
history stories are principally intended for purposes of entertainment, more literary entries are
concerned with exploration of various ideas, with contingency itself and the consequent fragility
of history being common (Rosenfeld 2002). Though long a significant part of science fiction,
alternate history experienced a flowering beginning in the late 1980’s that accelerated in the
1990’s, and continues in the early part of the 21st century. This growth is most evident from the
increasing pace of publication of alternate history works that is catalogued at the Uchronia
database (www.uchronia.net).
Increased interest in contingency has also manifested in other cultural media.

On

television, alteration of history has been a common plot point in numerous episodes of the
various incarnations of Star Trek

11

(1966 – 1969, 1973 – 1974, 1987 – 1994, 1993 – 1999, 1995

– 2001, 2001 – 2005) and The Twilight Zone (1959 – 1964, 1985 – 1989, 2002 – 2003). Other
shows had the alteration of history via contingency as part of the central premise of the story.
Quantum Leap (1989 – 1993), for instance, involved the adventures of a physicist, Sam Beckett,
who each week “leapt” into the body of a person from a different time period, with the charge to
“put right what once went wrong”. The suggestion was generally that the small changes Sam
made in each person’s life was somehow having a larger, though hidden, impact on history in
11

The most notable example is what is perhaps the most highly regarded episode of the original

series, City on the Edge of Forever (1967), in which Dr. McCoy accidentally goes back in time
to 1930’s America.

During his time there, he saves the life of Edith Keeler, with the

consequence that World War II ends with an Axis victory. The remainder of the episodes
revolves around the moral dilemma of having to rectify this alteration and set right the time line,
which requires Keeler’s death.

60

general.

The later Journeyman (2007), featured a similar premise.

Similarly, historical

contingency and couterfactuals have been key concepts in a variety of movies, including, of
course, It’s a Wonderful Life (1946). As with alternate history titles, these films have been more
frequent since the 1980’s, and have included Back to the Future (1985) and its sequels (1989,
1990), Groundhog Day (1993), Frequency (2000), The Butterfly Effect (2004), CSA:

The

Confederate States of America (2004), Inglourious Basterds (2009), and Never Let Me Go
(2010). On a smaller level of the contingencies that affect individual lives, Sliding Doors (1998),
examines the consequences of a character catching or missing a subway train, while The Family
Man (2000) traces the effects consequent to the alternative of a single decision in one man’s life.

61

REFERENCES

62

REFERENCES
1. Abrams, P. (Ed) Historical Sociology (Cornell University Press, Ithaca, NY, 1983).
2. Allmon, W.D. The structure of Gould, In Allmon, W.D., Kelley, P.H. & Ross, R.M.
(Eds.) Stephen Jay Gould: Reflections on His View of Life (Oxford University Press,
New York, NY, 2009).
3. Atlas, R.M. Handbook of Microbiological Media (CRC Press, Boca Raton, FL, 2004).
4. Barrick, J.E. & Lenski, R.E. Genome-wide mutational diversity in an evolving
population of Escherichia coli. Cold Spring Harbor Symposia on Quantitative Biology
74, 119-129 (2009).
5. Barrick, J.E., Yu, D.S., Yoon, S.H., Jeong, H., Oh, T.K, Schneider, D, Lenski, R.E. &
Kim, J.F. Genome evolution and adaptation in a long-term experiment with Escherichia
coli. Nature 461, 1243–1247 (2009).
6. Barrick, J.E., Kauth, M.R., Strelioff, C.C. & Lenski, R.E. Escherichia coli rpoB mutants
have increased evolvability in proportion to their fitness defects. Molecular and
Biological Evolution 27, 1338–1347 (2010).
7. Beatty, J. The evolutionary contingency thesis, In Wolters, G. Lennox, J.G. &
McLaughlin, P. Concepts, Theories, and Rationality in the Biological Sciences
(University of Pittsburgh Press, Pittsburgh, PA, 1995).
8. Beatty, J. Chance variation: Darwin on orchids. Philosophy of Science, 73, 629–641
(2006a).
9. Beatty, J. Replaying life’s tape. The Journal of Philosophy 103, 336–362 (2006b).
10. Beatty, J. Chance variation and evolutionary contingency: Darwin, Simpson, The
Simpsons, and Gould. In Ruse, M. & Werkmeister L.T. (Eds). The Oxford Handbook of
Philosophy of Biology (Oxford University Press, New York, NY, 2008).
11. Beatty, J. & Desjardins, E.C. Natural selection and history. Biology and Philosophy 24,
231–246 (2009).
12. Ben-Menahem, Y. Historical contingency. Ratio 10, 99–107 (1997).
13. Bennett, A.F. & Lenski, R.E. Evolutionary adaptation to temperature. II. Thermal niches
of experimental lines of Escherichia coli. Evolution 47, 1–12 (1993).
14. Bentley, D.R. Whole-genome resequencing.
Development 16, 545–552 (2006).

63

Current Opinion in Genetics and

15. Blaser, K. The history of nature and the nature of history: Stephen Jay gould on science,
philosophy, and history. The History Teacher, 32, 411–430 (1999).
16. Bloch, M. The Historian’s Craft. (Vintage Books, New York, NY, 1953).
17. Bohannan, B.J.M. & Lenski, R.E. The relative importance of competition and predation
varies with productivity in a model community. American Naturalist 156, 329-340
(2000).
18. Bowler, P.J. Evolution: The History of an Idea. (University of California Press, Berkley
and Los Angeles, CA, 2003).
19. Bradbury, R. A sound of thunder. (1952). Reprinted in: R is for Rocket (Doubleday and
Company, New York, NY, 1962).
20. Bridgham, J.T., Ortlund, E.A. & Thornton, J.W. An epistatic ratchet constrains the
direction of glucocorticoid receptor evolution. Nature 461, 515–519 (2009).
21. Bryant, J.M. Evidence and explanation in history and sociology: critical reflections on
Goldthorpe’s critique of historical sociology. British Journal of Sociology 65, 3-19
(1994).
22. Bulhof, J. What if? Modality and history. History and Theory 38, 145–168 (1999).
23. Burch, C. L. & Chao, L. Evolution by small steps and rugged landscapes in the RNA
virus φ-6. Genetics 151, 921–927 (2000a).
24. Burch, C.L. & Chao, L. Evolvability of an RNA virus is determined by its mutational
neighborhood. Nature 406, 625–628 (2000).
25. Carlton, B.C., and Brown, B.J. Gene mutation. In (Ed.) P. Gerhardt, Manual of Methods
for General Bacteriology. (American Society for Microbiology, Washington, DC, 1981).
26. Cartwright, N.D. Counterfactuals in economics: a commentary. In Campbell, J.K,
O’Rourke, Silverstein, H. (Eds). Causation and Explanation (MIT Press, Cambridge,
MA, 2007).
27. Chao, L. & Cox, E.C. Competition between high and low mutation strains of
Escherichia coli. Evolution 37, 125–134 (1983).
28. Cohen, S.S. and Anderson, T.F. Chemical studies on host-virus interactions I: The effect
of bacteriophage adsorption on the multiplication of its host, Escherichia coli B. Journal
of Experimental Medicine 84, 511–523 (1946).
29. Colosimo, P.F., Hosemann, K.E., Balabhadra, S., Villarrea, G. Jr., Dickson, M.,
Grimwood, J., Schmutz, J., Myers, R.M., Schluter, D. & Kinglsey, D.M. Widespread

64

parallel evolution in sticklebacks by repeated fixation of ectodysplasin alleles. Science
5717, 1928–1933 (2005).
30. Conway Morris, S. The Crucible of Creation: The Burgess Shale and the Rise of
Animals (Oxford University Press, Oxford, UK 1998).
31. Conway Morris, S. Life’s Solution: Inevitable Humans in a Lonely Universe
(Cambridge University Press, Cambridge, UK, 2003).
32. Conway Morris, S. Evolution: like any other science it is predictable. Philosophical
Transactions of the Royal Society of London B: Biological Sciences 365, 133–145
(2010).
33. Cooper, V.S. & Lenski, R.E. The population genetics of ecological specialization in
evolving E. coli populations. Nature 407, 736–739 (2000).
34. Cooper, V., Schneider, S., Blot, M. & Lenski, R. Mechanisms causing rapid and parallel
losses of ribose catabolism in evolving populations of E. coli B. Journal of Bacteriology
183, 2834–2841 (2001).
35. Cooper, T.F., Rozen, D.E. & Lenski, R.E. Parallel changes in gene expression after
20,000 generations of evolution in E. coli. Proceedings of the National Academy of
Sciences (USA) 100, 1072–1077 (2003).
36. Cooper, V.S., Remold, S.K., Lenski, R.E., & Schneider, D. Expression profiles reveal
parallel evolution of epistatic interactions involving the CRP regulon in Escherichia coli.
PLoS Genetics 4, e35 (2008).
37. Coricelli, G. & Rustichini, A. Counterfactual thinking and emotions: regret and envy
learning. Philosophical Transactions of the Royal Society of London B: Biological
Sciences 365, 241–247 (2010).
38. Cowan, R. & Foray, D. Evolutionary economics and the counterfactual threat: on the
nature and role of counterfactual history as an empirical tool in economics. Journal of
Evolutionary Economics 12, 539–562 (2002).
39. Cox, G., Gibson, F., Luke, R., Newton, N., O’Brien, I., and Rosenberg, H. Mutations
affecting iron transport in Escherichia coli. Journal of Bacteriology 104, 219–226
(1970).
40. Crowley, R. & Ambrose, S. (Eds). What If? The world’s foremost military historians
imagine what might have been (Berkeley Publishing Group, New York, NY 2000).
41. Crozat, E, Philippe, N., Lenski, R.E., Geiselmann, J. & Schneider, D. Long-term
experimental evolution in Escherichia coli. XII. DNA topology as a key target of
selection. Genetics 169, 523–532 (2005).

65

42. Crozat, E., Winkworth, C., Gaffe, J, Hallin, P.F. Riley, M.A, Lenski, R.E. & Schneider,
D. Parallel genetic and phenotypic evolution of DNA superhelicity in experimental
populations of Escherichia coli. Molecular Biology and Evolution 27, 2113–2128
(2010).
43. Crozat, E., Hindre, T., Kuhn, L, Garin, J., Lenski, R.E. & Schneider, D. Altered
regulation of the OmpF porin by Fis in Escherichia coli during an evolution experiment
between B and K12 strains. Journal of Bacteriology 193, 429–440 (2011).
44. Darwin, C.R. On the Origin of Species (John Murray, London, UK, 1859).
45. Davis, B.D. The isolation of biochemically deficient mutants of bacteria by means of
penicillin. Proceedings of the National Academy of Sciences (USA) 35, 1–10 (1949).
46. Davis, B.D. & Mingioli, E.S. Mutants of Escherichia coli requiring methionine or
vitamin B12. Journal of Bacteriology 60, 17–28 (1950).
47. Dawkins, R. The Blind Watchmaker, Second Edition (Norton, New York, NY, 1996a).
48. Dawkins, R. Climbing Mount Improbable (Norton, New York, NY, 1996b).
49. Day, M. Philosophy of History: An Introduction (Continuum, London, UK, 2008).
50. de Queiroz, A. & Rodriguez-Robles, J.A. Historical contingency and animal diets: The
origins of egg eating in snakes. The American Naturalist 167, 684–694 (2006).
51. De Visser, J.A.G.M., Zeyl, C.W., Gerrish, P.J., Blanchard, J.L, & Lenski, R.E.
Diminishing returns from mutation supply rate in asexual populations. Science 283, 404406 (1999).
52. De Visser, J.A.G.M. & Rozen, D.E. Clonal interference and the periodic selection of
new beneficial mutations in Escherichia coli. Genetics 172, 2093–2100 (2006).
53. Dennett, D. Darwin’s Dangerous Idea (Simon and Schuster, New York, NY, 1995).
54. Desjardins, E. Historicity and experimental evolution. Biology and Philosophy, In press.
55. Dick, M.H., Lidgard, S, Gordon, D.P. & Mawatari S.F. The origin of ascophoran
bryozoans was historical contingent but likely. Proceedings of the Royal Society B:
Biological Sciences 276, 3141–3148 (2009).
56. Dick, P.K. The Man in the High Castle (Putnam, New York, NY, 1962).
57. Dreyfuss, R. & Turtledove, H.N. The Two Georges (Tom Doherty Associates, New
York, NY, 1997).

66

58. Elena, S.F., Sanjuan, R., Borderia, A.V. & Turner, P.E. Transmission bottlenecks and the
evolution of fitness in rapidly evolving RNA viruses. Genetics and Evolution 1, 41-48
(2001).
59. Elena, S.F. & Lenski, R.E. Evolution experiments with microorganisms: the dynamics
and genetic bases of adaptation. Nature Reviews Genetics 4, 457-469 (2003).
60. Emerson, S.B. A macroevolutionary study of historical contingency in the fanged frogs
of Southeast Asia. Biological Journal of the Linnean Society 73, 139–151 (2001).
61. Enslin, W. The emperor and the imperial administration. In Baynes, N.H. and Moss,
H.St.L.B. Byzantium: An introduction to East Roman civilization. (Clarendon Press,
Oxford, UK, 1961).
62. Ferguson, N. Introduction. In Ferguson, N. (ed). Virtual History: Alternatives and
Counterfactuals (Basic Books, New York, NY, 2000).
63. Ferguson, N. The War of the World (Penguin, New York, NY, 2007).
64. Finkel, S.E. Long-term survival during stationary phase: evolution and the GASP
phenotype. Nature Reviews Microbiology 4, 113–120 (2006).
65. Fischer, D.H. Editor’s note In McPherson, J.M. Crossroads of Freedom: Antietam
(Oxford University Press, New York, NY, 2002).
66. Flood, C.B. Grant and Sherman (Harper Perennial, New York, NY, 2006).
67. Fogel, R.W. Railroads and American Economic Growth: Essays in Econometric History
(The Johns Hopkins University Press, Baltimore, MD, 1964).
68. Foote, M. Contingency and convergence. Science 280, 2068–2069 (1998).
69. Frost, G., and Rosenberg, H. The inducible citrate-dependent iron transport system in
Escherichia coli K-12. Biochimical et Biophysica Acta 330, 90–101 (1973).
70. Fukami, T., Beaumont, H.J.E., Zhang, X-. X. & Rainey, P.B. Immigration history
controls diversification in experimental adaptive radiation. Nature 446, 436–439 (2007).
71. Gerrish, P.J. & Lenski, R.E. The fate of competing beneficial mutations in an asexual
population. Genetica 102/103, 127-144 (1998).
72. Goldthorpe, J.H. The uses of history in sociology: reflections on some recent tendencies.
British Journal of Sociology 42, 211 – 30 (1991).
73. Gould, S.J. Hen’s Teeth and Horse’s Toes. (Norton, New York, NY 1983).

67

74. Gould, S.J. The paradox of the first tier: an agenda for paleobiology. Paleobiology 11,
2–12 (1985).
75. Gould, S.J. Wonderful Life (Norton, New York, NY, 1989).
76. Gould, S.J. The Individual in Darwin’s World: The Second Edinburgh Medal Address
(Edinburgh University Press, Edinburgh, UK, 1990).
77. Gould, S.J. Eight Little Piggies (W.W. Norton & Company, Inc., New York, NY, 1993).
78. Gould, S.J. Full House (Random House, New York, NY, 1996).
79. Gould, S.J. The Structure of Evolutionary Theory (The Belknap Press of Harvard
University Press, Cambridge, MA, 2002).
80. Gould, S.J., Raup, D.M., Sepkoski, J.J. Jr., Schopf, T.J.M. & Simberloff, D.S. The shape
of evolution: A comparison of real and random clades. Paleobiology 3, 23–40 (1977).
81. Gould, S.J. & Lewontin, R.C. The spandresl of San Marco and the Panglossian
paradigm: a critique of the adaptationist programme. Proceedings of the Royal Society
of London B: Biological Sciences 205, 581–598 (1979).
82. Gould, S.J. & Woodruff, D.S. History as a cause and area effects: An illustration from
Cerion on Great Inagua, Bahamas. British Journal of the Linnean Society, 40, 67–98
(1990).
83. Griffiths, P.E. The historical turn in the study of adaptation. British Journal of the
Philosophy of Science 47, 511–532 (1996).
84. Hall, B. 1982. Chromosomal mutation for citrate utilization by Escherichia coli K-12.
Journal of Bacteriology 151, 269–273 (1982).
85. Harding, D.J.
Counterfactual models of neighborhood effects:
the effect of
neighborhood poverty on dropping out and teenage pregnancy. American Journal of
Sociology 109, 676 – 719 (2003).
86. Härle, c., Kim, I., Angerer, A., and Braun, V. Signal transfer through three
compartments: transcription initiation of the Escherichia coli ferric citrate transport
system from the cell surface. The EMBO Journal 14, 1430–1438 (1995).
87. Harmon, L.J., Harmon, L.L. & Jones, C.G. Competition and community structure in
diurnal arboreal geckos (genus Phelsuma). Systematic Biology 57, 562–573 (2007).
88. Harris, R. Fatherland (Random House, New York, NY, 1992).

68

89. Hartmann, A. and Braun, V. Iron uptake and iron limited growth of Escherichia coli K12. Archives of Microbiology 130, 353–356 (1981).
90. Hegreness, M. & Kishony, R. Analysis of genetic systems using experimental evolution
and whole-genome sequencing. Genome Biology 8, 201 (2007).
91. Helling, R.B., Vargas, C.N. & Adams, J. Evolution of Escherichia coli during growth in
a constant environment. Genetics 116, 349–358 (1987).
92. Herring, C., Glasner, J., and Blatner, F. Gene replacement without selection: regulated
suppression of amber mutations in Escherichia coli. Gene 331, 151–163 (2003).
93. Hussein, S., Hantke, K., and Braun, V. Citrate-dependent iron transport system in
Escherichia coli K-12. European Journal of Biochemistry 117, 431–437 (1981).
94. Huxley, J.S. Evolution: The Modern Synthesis (Allen and Unwin, London, UK, 1942).
95. Huynen, M.A., Stadler, P.F. & Fontana, W.B. Smoothness within ruggedness: The role
of neutrality in adaptation. Proceedings of the National Academy of Sciences (USA) 93,
397–401 (1996).
96. Ishiguro, N., Oka, C. & Sato, G. Isolation of citrate positive variants of Escherichia coli
from domestic pigeons, pigs, cattle, and horses.
Applied and Environmental
Microbiology 36, 217–222 (1978).
97. Ishiguro, N., Oka, C., Hanazawa, Y. & Sato, G. Plasmids in Escherichia coli controlling
citrate-utilizing ability. Applied and Environmental Microbiology 28, 956–964 (1979).
98. Jacob, F. Evolution and tinkering. Science 196, 1161–1166 (1977).
99. Jacob, F. The Possible and the Actual (University of Washington Press, Seattle, WA,
1982).
100. Jeong, H., Barbe, V., Lee, C.H., Vallenet, D., Yu, D.S., Choi, S.H., Couloux, A., Lee,
S.W., Yoon, S.H., Cattolico, L., Hur, C.G, Park, H.S, Ségurens, B., Kim, S.C., Oh, T.K.,
Lenski, R.E, Studier, F.W., Daegelen, P. & Kim, J.F. Genome sequences of Escherichia
coli B strains REL606 and BL21(DE3). Journal of Molecular Biology 394, 644–653
(2009).
101. Kahneman, D. & Miller, D. Norm theory:
Psychological Review 93, 136–153 (1986).

Comparing reality to its alternatives.

102. Kahneman, D. & Tversky, A. The simulation heuristic. In Kahneman, D., Slovic, P, &
Tversky, A. (Eds). Judgement under uncertainty: heuristics and biases. (Cambridge
University Press, New York, NY, 1982).

69

103. Keegan, J. The First World War. (Vintage, New York, NY, 2000).
104. Kimura, M. Evolutionary rate at the molecular level. Nature 217, 624–626 (1968).
105. Knapp, A.K, Burns, C.E., Fynn, R.W.S., Kirkman, K.P., Morris, C.D. & Smith, M.D.
Convergence and contingency in production-precipitation relationships in North
American and South African C4 grasslands. Oecologia 149, 456–464 (2006).
106. Koser, S.A. Utilization of the salts of organic acids by colon-aerogenes group. Journal
of Bacteriology 8, 493–520 (1923).
107. Koser, S.A. Correlation of citrate-utilization by members of the colon-aerogenes group
with other differential characteristic and with habitat. Journal of Bacteriology 9, 59–77
(1924).
108. Lamprect, S. Contingency in nature. Philosophy and Phenomenological Research 32, 1–
14 (1971).
109. Langman, L., Young, I.G., Frost, G.E., Rosenberg, H, and Gibson, F. Enterochelin
system of iron transport in Escherichia coli: Mutations affecting ferric-enterochelin
esterase. Journal of Bacteriology 112, 1142–1149 (1972).
110. Lara, F. Stokes, J. Oxidation of citrate by Escherichia coli. Journal of Bacteriology 9,
59–77 (1952).
111. Larson, E.J. Evolution (Modern Library, New York, NY, 2004).
112. Lenski, R. Phenotypic and genomic evolution during a 20,000-generation experiment
with the bacterium Escherichia coli. Plant Breeding Reviews 24, 225–265 (2004).
113. Lenski, R.E., Rose, M.R., Simpson, S.C. & Tadler, S.C. Long-term experimental
evolution in Escherichia coli. I. Adaptation and divergence during 2,000 generations.
The American Naturalist 138, 1315–1341 (1991).
114. Lenski, R. & Travisano, M. Dynamics of adaptation and diversification: 10,000generation experiment with bacterial populations. Proceedings of the National Academy
of Sciences (USA) 91, 6808–6814 (1994).
115. Lenski, R.E., Ofria, C., Pennock, R.T. & Adami, C. The evolutionary origin of complex
features. Nature 423, 139–144 (2003a).
116. Lenski, R.E., Winkworth, C.L. & Riley, M.A. Rates of DNA sequence evolution in
experimental populations of Escherichia coli during 20,000 generations. Journal of
Molecular Evolution 56, 498–508 (2003b).

70

117. Lenski, R.E., Barrick, J.E. & Ofria, C. Balancing robustness and evolvability. PLoS
Biology 4, e428 (2006).
118. Levin, B.R. & Lenski, R.E. Bacteria and phage: a model system for the study of the
ecology and co-evolution of hosts and parasites. In Rollinson, D. & Anderson, R.M.
(Eds). Ecology and the Genetics of Host-Parasite Interactions. (Academic Press,
London, UK, 1985).
119. Lewis, D.K. Counterfactuals. (Blackwell, Oxford, UK, 2001).
120. Lewontin, R.C. Is nature probably or capricious? BioScience 16, 25–27 (1966).
121. Livius, T., Radice, B (Translator). Rome and Italy: Books Vi-X of The History of Rome
from its Foundation (Penguin Putnam Inc, New York, NY, 1982).
122. Lochhead, R. & Menninger, R. About the Department of Public Health and Preventative
Medicine. Cornell Medical Journal 4, 25–27 (1949).
123. Losos, J.B., Jackman, T.R., Larson, A., de Queiroz, K. & Rodriguez-Schettino, L.
Contingency and determinism in replicated adaptive radiations of island lizards. Science
279, 2115–2118 (1998).
124. Losos, J.B. Adaptive radiation, ecological opportunity, and evolutionary determinism.
The American Naturalist 175, 623–639 (2010).
125. Lutgens, M., and Gottschalk, G. Why a co-substrate is required for anaerobic growth of
Escherichia coli on citrate. Journal of General Microbiology 199, 63–70 (1980).
126. Maas, W.K. Bernard David Davis: 1916-1994. In Biographical Memoirs, Volume 77
(National Academy Press, Washington, D.C., 1999).
127. Mandel, D.R., Hilton, D.J. & Catellani, P. (Eds). The Psychology of Counterfactual
Thinking. (Routledge, New York, NY, 2005).
128. Mani, G.S, & Clarke, B.C. Mutational order: A major stochastic process in evolution.
Proceedings of the Royal Society of London B: Biological Sciences 240, 29–37 (1990).
129. Maynard Smith, J. Taking a chance on evolution. New York Review of Books May 14,
234–236 (1992).
130. Maynard Smith, J. & Szathmary, E.
University Press, Oxford, UK, 1995).

The Major Transitions in Evolution (Oxford

131. Mayr, E. The emergence of evolutionary novelties. In (Ed) Tax, S. Evolution after
Darwin (University of Chicago Press, Chicago, 1960).

71

132. Mayr, E. Some thoughts on the history of the evolutionary synthesis. In (Ed.) Mayr, E. &
Provine, W.B. The Evolutionary Synthesis (Harvard University Press, Cambridge, MA,
1980).
133. Mayr, E. Toward a New Philosophy of Biology (Harvard University Press, Cambridge,
MA, 1988).
134. McPhail, J.D. Ecology and evolution of sympatric sticklebacks (Gasterosteus): origins
of the species pairs. Canadian Journal of Zoology 71, 515–523 (1993).
135. McPhail, J.D. Speciation and the evolution of reproductive isolation in the sticklebacks
(Gasterosteus) of south-western British Columbia. In (Eds.) Bell, M.A. & foster, S.A.,
The Evolutionary Biology of the Threespine Stickleback (Oxford Science Publications,
Oxford, UK, 1994).
136. McPherson, J.M. Battle Cry of Freedom: The Civil War Era (Ballantine, New York,
NY, 1988).
137. McPherson, J.M. Crossroads of Freedom: Antietam (Oxford University Press, New
York, NY, 2002).
138. McRae, M.W. Stephen Jay Gould and the contingent nature of history. Clio 22, 239–250
(1993).
139. Meyer, J.R. & Kassen, R. The effects of competition and predation on diversification in
a model adaptive radiation. Nature 446, 432–435 (2007).
140. Meyer, J.R., Agrawal, A.A., Quick, R.T., Dobias, D.T., Schneider, D. & Lenski, R.E.
Parallel changes in host resistance to viral infection during 45,000 generations of relaxed
selection. Evolution 64, 3024–3034 (2010).
141. Moore, W. Bring the Jubilee (Random House, New York, NY, 1953).
142. Morgan, S.L. Counterfactuals and Causal Inference: Methods and Principles for Social
Research (Cambridge University Press, New York, NY, 2007).
143. Morris, M.W. & Moore, P.C. The lessons we (don’t) learn: counterfactual thinking and
organizational accountability after a close call. Administrative Science Quarterly 45, 737
– 765 (2000).
144. Muller, H.J. Reversibility in evolution considered from the standpoint of genetics.
Biological Reviews of the Cambridge Philosophical Society 14, 261–280 (1939).
145. Newell, N.D. Revolutions in the history of life. Geological Society of America Special
Paper 89, 62–91 (1967).

72

146. Nosil, P. & Flaxman, S.M. Conditions for mutation-order speciation. Proceedings of the
Royal Society of London B: Biological Sciences 278, 399–407 (2011).
147. Ortlund, E.A., Bridgham, J.T, Redinbo, MR. & thrnton, J.W. Crystal structure of an
ancient protein: evolution by conformational epistasis. Science 317, 1544–1548 (2007).
148. Ostrowski, E.A., Rozen, D.E. & Lenski, R.E. Pleiotropic effects of beneficial mutations
in Escherichia coli. Evolution 59, 2343–2352 (2005).
149. Ostrowski, E.A., Woods, R.J. & Lenski, R.E. The genetic basis of parallel and divergent
phenotypic responses in evolving populations of Escherichia coli. Proceedings of the
Royal Society of London B: Biological Sciences 275, 277–284 (2008).
150. Papadopoulos, D., Schneider, D., Meier-Eiss, J., Arber, W., Lenski, R.E. & Blot, M.
Genomic evolution during a 10,000-generation experiment with bacteria. Proceedings of
the National Academy of Sciences (USA) 96, 3807–3812 (1999).
151. Pelosi, L., Kuhn, L, Guetta, D., Garin, J., Geiselmann, J., Lenski, R.E. & Schneider, D.
(2006). Parallel changes in global protein profiles during long-term experimental
evolution in Escherichia coli. Genetics 173, 1851–1869 (2006).
152. Perfeito, L, Fernandes, L, Mota, C, & Gordo, I. Adaptive mutations in bacteria: high
rate and small effects. Science 317, 813–815 (2007).
153. Pfeiffer, T. & Bonhoeffer, S. Evolution of cross-feeding in microbial populations. The
American Naturalist 163, E126–135 (2004).
154. Philippe, N., Crozat, E., Lenski, R.E. & Schneider, D. (2007). Evolution of global
regulatory networks during a long-term experiment with Escherichia coli. BioEssays 29,
846–860 (2007).
155. Philippe, N., Pelosi, L., Lenski, R.E. & Schneider, D. Evolution of penicillin-binding
protein 2 concentration and cell shape during a long-term experiment with Escherichia
coli. Journal of Bacteriology 191, 909–921 (2009).
156. Pigliucci, M. What, if anything, is an evolutionary novelty? Philosophy of Science 75,
887–898 (2008).
157. Pigliucci, M. An extended synthesis for evolutionary biology. Annals of the New York
Academy of Sciences 1168, 218–228 (2009).
158. Pos, K., Dimroth, P., and Bott, M. The Escherichia coli citrate carrier CitT: a member of
a novel eubacterial transporter family related to the 2-oxoglutarate/malate translocator
from spinach chloroplasts. Journal of Bacteriology 180, 4160–4165 (1998).

73

159. Provine, W.B. The Origins of Theoretical Population Genetics (University of Chicago
Press, Chicago, IL, 1971).
160. Raup, D.M, Gould, S.J., Schopf, T.J.M, & Simberloff, D.S. Stochastic models of
phylogeny and the evolution of diversity. Journal of Geology 81, 525–542 (1973).
161. Raup, D.M. Extinction: Bad Genes or Bad Luck (Norton, New York, NY, 1991).
162. Reisch, G. Chaos, history, and narrative. History and Theory 30, 1–20 (1991).
163. Reynolds, C., and Silver, S. Citrate utilization by Escherichia coli: plasmid-and
chromosome-encoded systems. Journal of Bacteriology 156, 1019–1024 (1983).
164. Robinson, K.S. The Years of Rice and Salt (Bantam, New York, NY, 2002).
165. Rosenfeld, B. Why do we ask “what if?”: Reflections on the function of alternate
history. History and Theory 41, 90–103 (2002).
166. Rozen, D.E. & Lenski, R.E. Long-term experimental evolution in Escherichia coli. VIII.
Dynamics of a balanced polymorphism. The American Naturalist 155, 24–35 (2000).
167. Rozen, D.E., Schneider, D. & Lenski, R.E. Long-term experimental evolution in
Escherichia coli. XIII. Phylogenetic history of a balanced polymorphism. Journal of
Molecular Evolution 61, 171–180 (2005).
168. Rozen, D.E., Philippe, N., de Visser, J.A. & Schneider, D. Death and cannibalism in a
seasonal environment facilitate bacterial coexistence. Ecology Letters 12, 34–44 (2009).
169. Ruelle, D. Chance and Chaos. (Princeston Science Library, Princeton, NJ, 1991).
170. Sahyun, M., Beard, P, Schultz, E.W., Snow, J., and Cross, E. Growth stimulating factors
for microorganisms. The Journal of Infectious Diseases 58, 28–44.
171. Salverda, M.L.M., Dellus, E., Gorter, F.A., Debets, A.J.M., van der Oost, J., Hoekstra,
R.F., Tawfik, D.S. & de Visser, J.A.G.M. Initial mutations direct alternative pathways of
protein evolution. PloS Genetics 7, e1001321 (2011).
172. Scheutz, F., and Strockbine, N.A. Genus I. Escherichia, Castellani and Chalmers 1919.
In (Ed.) Garrity, G.M., Brenner, D., Kreig, N.R., Staley, J.R. Bergey’s Manual of
Systemaic Bacteriology, Volume Two: The Proteobacteria (Springer, New York, NY,
2005).
173. Schluter, D. Ecological speciation in postglacial fishes. Philosophical Transactions of
the Royal Society of London B: Biological Sciences 351, 807–814 (1996).

74

174. Schluter, D., Marchinko, K.B., Barrett, R.D.H., & Rogers, S.M. Natural selection and the
genetics of adaptation in threespine stickleback. Philosophical Transactions of the Royal
Society of London B: Biological Sciences 365, 2479–2486 (2010).
175. Seehausen, O. Chance, historical contingency, and ecological determinism jointly
determin the rate of adaptive radiation. Heredity 99, 361–363 (2007).
176. Shubin, N., Tabin, C. & Carroll, S. Fossils, genes, and the evolution of animal limbs.
Nature 388, 639–648 (1997).
177. Shubin, N., Tabin, C. & Carroll, S. Deep homology and the origins of evolutionary
novelty. Nature 457, 818–823 (2009).
178. Silverberg, R. Roma Eterna (Roc, New York, NY, 2003).
179. Sleight, S.C., Wigginton, N.S. & Lenski, R.E. Increased susceptibility to repeated freezethaw cycles in Escherichia coli following long-term evolution in a benign environment.
BMC Evolutionary Biology 6, 104–108 (2006).
180. Sniegowski, P., Gerrish, P. & Lenski, R. Evolution of high mutation rates: separating
causes from consequences. BioEssays 22, 1057–1066 (1997).
181. Sobel, M.E. Causal inference in the social and behavioral sciences. In Arminger, G.,
Clogg, C.C. & Sobel, M.E. (Eds). Handbook of Statistical Modeling for the Social and
Behavioral Sciences (Plenum Press, New York, 1995).
182. Sobel, R. For Want of a Nail (MacMillan, New York, NY, 1973).
183. Sole, R.V., Montoya, J.M. & Erwin, D.H. Recovery after mass extinction: evolutionary
assembly in large-scale biosphere dynamics. Philosophical Transactions of the Royal
Society of London B: Biological Sciences 357, 697–707 (2002).
184. Squire, J.C. (Ed) If It Had Happened Otherwise: Lapses into Imaginary History
(Longmans, Green, London, UK, 1931).
185. Stanek, M.T., Cooper, T.F. & Lenski, R.E. Identification and dynamics of a beneficial
mutation in a long-term evolution experiment with Escherichia coli. BMC Evolutionary
Biology 9, 302–315 (2009).
186. Sterelny, K., & Griffiths, P. Sex and Death: An Introduction to the Philosophy of
Biology (University of Chicago Press, Chicago, IL, 1999).
187. Studier, F.W., Daegelen, P., Lenski, R.E., Maslov, S. & Kim, J.F. Understanding the
differences between genome sequences of Escherichia coli B strains REL606 and
BL21(DE3) and comparison of the E. coli B and K-12 genomes. Journal of Molecular
Biology 394, 644–652 (2009).

75

188. Taylor, E.B. & McPhail, J.D. Historical contingency and ecological determinism interact
to prime speciation in sticklebacks, Gasterosteus. Proceedings of the Royal Society of
London, Series B: Biological Sciences 267, 2375–2384 (2000)
189. Tetlock, P.E. & Belkin, Aaron (Eds.) Counterfactual Thought Experiments in World
Politics:
Logical, Methodological, and Psychological Perspectives. (Princeton
University Press, Princeton, NJ, 1996).
190. Thompson, E.P. The Poverty of Theory and Other Essays (Merlin, London, UK, 1978).
191. Thucydides, Warner, R. (Translator), Finley, M.I. (Ed).
The History of the
Peloponnesian War: Revised Edition (Penguin Putnam Inc, New York, NY, 1972).
192. Travisano, M., Mongold, J.A, Bennett, A.F. & Lenski R.E. Experimental tests of the
roles of adaptation, chance, and history in evolution. Science 267, 87–90 (1995a).
193. Travisano, M., Vasi, R. & Lenski, R. Long-term experimental evolution in Escherichia
coli. III. Variation among replicate populations in correlated responses to novel
environments. Evolution 49, 189–200 (1995b).
194. Tucker, A. Historiographical counterfactuals and historical contingency. History and
Theory 38, 264–278 (1999).
195. Turner, D.D. Gould’s replay revisited. Biology and Philosophy 26, 65–79 (2011).
196. Turner, P.E., Souza, V. & Lenski, R.E. Tests of ecological mechanism promoting the
stable coexistence of two bacterial genotypes. Ecology 77, 2119–2129 (1996.
197. Turtledove, H.N. How Few Remain (Ballantine, New York, NY, 1997).
198. Turtledove, H.N. The Great War: American Front (Ballantine, New York, NY, 1998).
199. Turtledove, H.N. The Great War: Walk in Hell (Ballantine, New York, NY, 1999).
200. Turtledove, H.N. The Great War: Breakthroughs (Ballantine, New York, NY, 2000).
201. Turtledove, H.N. Introduction. In (Ed.) Turtledove, H.N & Greenberg, M.H. The Best
th
Alternative History Stories of the 20 Century (Del Rey, New York, NY, 2001a).
202. Turtledove, H.N.
2001b).

American Empire:

Blood and Iron (Ballantine, New York, NY,

203. Turtledove, H.N. American Empire: The Center Cannot Hold (Ballantine, New York,
NY, 2002).

76

204. Turtledove, H.N. In the Presence of Mine Enemies (NAL, New York, NY, 2003a).
205. Turtledove, H.N. American Empire: The Victorious Opposition (Ballantine, New York,
NY, 2003b).
206. Turtledove, H.N. Settling Accounts: Return Engagement (Ballantine, New York, NY,
2004).
207. Turtledove, H.N. Settling Accounts: Drive to the East (Ballantine, New York, NY,
2005).
208. Turtledove, H.N. Settling Accounts: The Grapple (Ballantine, New York, NY, 2006).
209. Turtledove, H.N. Settling Accounts: In at the Death (Ballantine, New York, NY, 2007).
210. www.uchronia.net, accessed February 18, 2011.
211. Vasi, F., Travisano, M. & Lenski, R. Long-term experimental evolution in Escherichia
coli. II. Changes in life-history traits during adaptation to a seasonal environment. The
American Naturalist, 144: 432–456.
212. Vermeij, G. J. Historical contingency and the purported uniqueness of evolutionary
innovations. Proceedings of the National Academy of Sciences (USA) 103, 1804–1809
(2006).
213. Vogel, H.J., in McElroy, W.D., and Glass, B. In (Ed) McElroy, W.D. A Symposium on
Amino Acid Metabolism (The Johns Hopkins Press, Baltimore, MD, 1955).
214. Vogel, H.J. and Davis, B.D. Glutamic gamma-semialdehyde and delta-1-pyrroline-5carboxylic acid, intermediates in the biosynthesis of proline. Journal of the American
Chemical Society, 74, 109–112 (1952).
215. Weinreich, D.M, Watson, R.A. & Chao, L. Sign epistasis and genetic constrain on
evolutionary trajectories. Evolution 59, 1165–1174 (2005).
216. Weinreich, D.M., Delaney, N.F., DePristo, M.A. & Hartle, D.L. Darwinian evolution can
follow only very few mutational paths to fitter proteins. Science 312, 111–114 (2006).
217. Wilke, C.O. Adaptive evolution on neutral networks. Bulletin of Mathematical Biology
63, 715–730 (2001).
218. Williamson, J. The Legion of Time (Pyramid, New York, NY, 1967).
219. Wilson, E.O. The Diversity of Life (Norton, New York, NY, 1993).

77

220. Woese, C. The universal ancestor. Proceedings of the National Academy of Sciences
(USA) 95, 6854–6859 (1995).
221. Woods, R., Schneider, D., Winkworth, C.L., Riley, M.A. & Lenski, R.E. Tests of
parallel molecular evolution in a long-term experiment with Escherichia coli.
Proceedings of the National Academy of Sciences (USA) 103, 9107–9122 (2006).
222. Wright, S. The roles of mutation, inbreeding, crossbreeding, and selection in evolution,
in Jone, D.F. (Ed.) Proceedings of the VI International Congress of Genetics, (1932).
223. Wright, S. Evolution and the Genetics of Populations, Volume 3. Experimental Results
and Evolutionary Deductions. (University of Chicago Press, Chicago, IL, 1977)
224. Wright, S. Surfaces of selective value revisited. The American Naturalist 131, 115–123
(1988).
225. Yedid, G. & Bell, G. Macroevolution simulated with autonomously replicating computer
programs. Nature 420, 810–812 (2002).

78

CHAPTER 2

Blount, Z.D., Borland, C.Z., & Lenski, R.E. Historical contingency and the evolution of a key
innovation in an experimental population of Escherichia coli. Proceedings of the National
Academy of Sciences (USA) 105, 7899–7906 (2008).

79

HISTORICAL CONTINGENCY AND THE EVOLUTION OF A KEY INNOVATION IN
AN EXPERIMENTAL POPULATION OF ESCHERICHIA COLI

ABSTRACT
The role of historical contingency in evolution has been much debated, but rarely tested.
Twelve initially identical populations of E. coli were founded in 1988 to investigate this issue.
They have since evolved in a glucose-limited medium that also contains citrate, which E. coli
cannot use as a carbon source under oxic conditions. No population evolved the capacity to
exploit citrate for >30,000 generations, although each population tested billions of mutations. A
+

citrate-using (Cit ) variant finally evolved in one population by 31,500 generations, causing an
increase in population size and diversity. The long-delayed and unique evolution of this function
might indicate the involvement of some extremely rare mutation. Alternately, it may involve an
ordinary mutation, but one whose physical occurrence or phenotypic expression is contingent on
prior mutations in that population. We tested these hypotheses in experiments that “replayed”
+

evolution from different points in that population’s history. We observed no Cit mutants
12

among 8.4 x 10

12

ancestral cells, nor among 9 x 10

cells from 60 clones sampled in the first

15,000 generations. However, we observed a significantly greater tendency for later clones to
+

evolve Cit , indicating that some potentiating mutation arose by 20,000 generations. This
+

potentiating change increased the mutation rate to Cit , but did not cause generalized
hypermutability. Thus, the evolution of this phenotype was contingent on the particular history
of that population.

More generally, we suggest that historical contingency is especially

80

important when it allows key innovations that are not easily evolved by gradual, cumulative
selection.

81

BACKGROUND AND INTRODUCTION
At its core, evolution involves a profound tension between random and deterministic
processes.

Natural selection works systematically to adapt populations to their current

environments. However, selection requires heritable variation generated by random mutation,
and even beneficial mutations may be lost by random drift. Moreover, random and deterministic
processes become intertwined over time such that future alternatives may be contingent on the
prior history of an evolving population. For example, multiple beneficial mutations will arise in
some unpredictable order (Mani and Clarke 1990, Lenski et al 1991), and those that are
substituted first may differ from others in their pleiotropic effects and epistatic interactions
(Cooper and Lenski 2000, Cooper et al 2008), thus constraining some evolutionary paths while
potentiating other outcomes (Jacob 1977, Wright 1988, Yedid and Bell 2002, Weinreich et al
2005, Weinreich et al 2006). These accidents of history may even determine the survival or
extinction of entire lineages, given the capricious and sudden nature of some environmental
changes (Lewontin 1966, Gould 1985).
Stephen Jay Gould maintained that these historical contingencies make evolution largely
unpredictable. Although each change on an evolutionary path has some causal relation to the
circumstances in which it arose, outcomes must eventually depend on the details of long chains
of antecedent states, small changes in which may have enormous long-term repercussions (Gould
1989, 2002, Beatty 2006).

Thus, Gould argued that contingency renders evolution

fundamentally quirky and unpredictable, and he famously suggested that replaying the “tape of
life” from some point in the distant past would yield a living world far different from the one we
see today. Simon Conway Morris countered that natural selection constrains organisms to a

82

relatively few highly adaptive options, so that “the evolutionary routes are many, but the
destinations are limited” (Conway Morris 2003). He and others point to numerous examples of
convergent evolution as evidence that selection finds the same adaptations despite the vagaries of
history. Evolution may thus be broadly repeatable, and multiple replays would reveal striking
similarities in important features, with contingency mostly confined to minor details (Conway
Morris 2003, Van Valen 1991, Dawkins 1996, Bermeij 2006).
Of course, replaying life’s tape on the planetary scale is impossible, but careful
experiments can examine the role of contingency in evolution on a more modest scale (Beatty
2006, Travisano et al 1995).

To address the repeatability of evolutionary trajectories and

outcomes, the long-term evolution experiment (LTEE) with E. coli was started in 1988 with the
founding of 12 populations from the same clone (Lenski et al 1991). These populations were
initially identical except for a neutral marker that distinguished six lines from six others. They
have since been propagated by daily 1:100 serial transfer in DM25, a minimal medium
containing 25 mg/L glucose as the limiting resource (Lenski et al 1991, Lenski 2004).
Environmental conditions have been controlled, constant, and identical for all 12 lines. To date,
each population has evolved for over 44,000 generations, and samples have been frozen every
500 generations, providing a rich “fossil record” (Lenski and Travisano 1994). Moreover, these
samples remain viable, allowing us to perform simultaneous measurements and other
experiments with bacteria from different generations. The founding strain is strictly asexual, so
populations have evolved by natural selection and genetic drift acting on variation generated by
spontaneous mutations that have occurred during the experiment. Thus, the LTEE allows us to
examine the effects of contingency that are inherent to the core evolutionary processes of
mutation, selection, and drift.

83

Previous analyses of this experiment have shown numerous examples of parallel
phenotypic and genetic evolution. All twelve populations underwent rapid improvement in
fitness that decelerated over time (Lenski et al 1991, Lenski and Travisano 1994, Cooper and
Lenski 2000, Lenski 2004). All evolved higher maximum growth rates on glucose, shorter lag
phases upon transfer into fresh medium, reduced peak population densities, and larger average
cell sizes relative to their ancestor (Lenski 2004, Lenski and Travisano 1994, Vasi et al 1994,
Lenski and Mongold 2000, Novak et al 2006).

Ten populations evolved increased DNA

supercoiling (Crozat et al 2005), and those populations examined to date show parallel changes
in global gene-expression profiles (Cooper et al 2003, Cooper et al 2008, Pelosi et al 2006). At
least three genes have substitutions in all 12 populations (Woods et al 2006, Cooper et al 2001),
while several others have substitutions in many populations (Crozat et al 2005, Cooper et al
2003, Pelosi et al 2006, Woods et al 2006), even though most loci harbor no substitutions in any
of them (Lenski et al 2003). At the same time, there has also been some divergence between
populations. Four have evolved defects in DNA repair, causing mutator phenotypes (Cooper and
Lenski 2000, Sniegowski et al 1997).

There is subtle, but significant, between-population

variation in mean fitness in the glucose-limited medium in which they evolved (Lenski et al
1991, Lenski and Travisano 1994). In media containing other carbon sources, such as maltose or
lactose, the variation in performance is much greater (Travisano et al 1995). And while the same
genes often harbor substitutions, the precise location and details of the mutations almost always
differ between the populations (Crozat et al 2005, Cooper et al 2003, Pelosi et al 2006, Woods et
al 2006, Cooper et al 2001).
Throughout the duration of the LTEE, there has existed an ecological opportunity in the
form of an abundant, but unused, resource. DM25 medium contains not only glucose, but also

84

citrate at a high concentration. The inability to use citrate as an energy source under oxic
conditions has long been a defining characteristic of E. coli as a species (Koser 1924, Scheutz
and Strockbridge 2005). Nevertheless, E. coli is not wholly indifferent to citrate. It uses a ferric
dicitrate transport system for iron acquisition, though citrate does not enter the cell in this
process (Frost and Rosenberg 1973). It also has a complete tricarboxylic acid cycle, and can thus
metabolize citrate internally during aerobic growth on other substrates (Lara and Stokes 1952).
E. coli is able to ferment citrate under anoxic conditions if a co-substrate is available for
reducing power (Lutgens and Gottschalk 1980). The only known barrier to aerobic growth on
citrate is its inability to transport citrate under oxic conditions (Hall 1982, Reynolds and Silver
+

1983, Pos et al 1998). Indeed, atypical E. coli that grow aerobically on citrate (Cit ) have been
isolated from agricultural and clinical settings, and were found to harbor plasmids, presumably
acquired from other species, that encode citrate transporters (Ischiguro et al 1978, Ishiguro et al
1979).
+

Other findings suggest that E. coli has the potential to evolve a Cit phenotype. Hall
+

(Hall 1982) reported the only documented case of a spontaneous Cit mutant in E. coli. He
hypothesized that some complex mutation, or multiple mutations, activated cryptic genes that
jointly expressed a citrate transporter, although the genes were not identified. Pos et al. (Pos et al
1998) identified an operon in E. coli K-12 that apparently allows anaerobic citrate fermentation,
and which includes a gene, citT, encoding a citrate-succinate antiporter. High-level constitutive
expression of this gene on a multi-copy plasmid allows aerobic growth on citrate, but the native
operon has a single copy that is presumably induced only under anoxic conditions.

85

Despite this potential, none of the 12 LTEE populations evolved the capacity to use the
citrate that was present in their environment for over 30,000 generations. During that time, each
population experienced billions of mutations (Lenski 2004), far more than the number of
possible point mutations in the ≈4.6-million bp genome.

This ratio implies, to a first

approximation, that each population tried every typical one-step mutation many times. It must
+

be difficult, therefore, to evolve the Cit phenotype, despite the ecological opportunity. Here we
+

report that a Cit variant finally evolved in one population by 31,500 generations, and its
+

descendants later rose to numerical dominance. The new Cit function has been the most
profound adaptation observed during the LTEE, and has had major consequences. As we will
show, the population achieved a several-fold increase in size. Moreover, a stable polymorphism
emerged, with a Cit

–

minority coexisting with the new Cit

+

majority.

Interestingly, the

+

population that evolved the Cit function is not one that had previously become hypermutable. It
is also intriguing that this key innovation evolved so late in the experiment, given that the rate of
fitness improvement had declined substantially in all the populations (Cooper and Lenski 2000,
Lenski and Travisano 1994).
+

The long-delayed and unique evolution of the Cit phenotype might indicate that it
required some unusually rare mutation, such as a particular chromosomal inversion, that does not
scale with typical mutation rates. Alternately, the occurrence or phenotypic expression of the
+

mutation that generated the Cit function might depend on one or more earlier mutations, such
that its evolution was contingent upon the particular history of that population. Contingent
adaptations should tend to be complex and require multiple steps, some of which might not be

86

beneficial, at least not uniquely so given other advantageous paths. Otherwise, cumulative
selection would predictably favor the same steps, and the evolutionary path should be repeatable
(Dawkins 1996).

Contingent adaptations should thus display two characteristics.

First,

independent origins should be rare, because the same historical sequences would rarely recur
(Vermeij 2006).

Second, significant time-lags should occur between the presentation of

ecological opportunities or challenges and the evolution of those traits that confer adaptation to
those circumstances (Foote 1998).
Can the hypothesis of contingent adaptation be rigorously tested in the case of the
+

evolution of the Cit function? And can this hypothesis be formally distinguished from the
alternative explanation that this new function required an unusually rare mutation, but one that
was not historically contingent? The answer to both questions is yes, owing to certain features of
the LTEE in particular, and of bacteria in general, that allow us quite literally to replay the tape
of life from various points in a population’s history. We will first describe the emergence of this
new function, and then present our experiments to distinguish between the hypotheses of
mutational rarity and historical contingency in the origin of this key innovation.
RESULTS
+

Evolution of Cit Function in Population Ara–3.
The LTEE populations are transferred daily into fresh medium, and the turbidity is
checked visually for each one at that time. Owing to the low concentration of glucose in DM25
medium, the cultures are only slightly turbid when transferred. Occasional contaminants that
grow on citrate have been seen over the 20 years of this experiment, and the contaminated
cultures reach much higher turbidity owing to the high concentration of citrate that allows the

87

contaminating bacteria to grow to high density. (When contamination occurs, the affected
population is restarted from the latest frozen sample.)

After ~33,127 generations, one

population, designated Ara–3, had significantly elevated turbidity, which continued to rise for
+

several days (Figure 2.1). A number of Cit clones were isolated from the population and
checked for phenotypic markers characteristic of the ancestral E. coli strain used to start the
–

LTEE: all were Ara , T5-sensitive, and T6-resistant, as expected (Lenski et al 1991). DNA
+

sequencing also shows that Cit clones have the same mutations in the pykF and nadR genes as
do clones from earlier generations of the Ara–3 population, and each of these mutations
+

distinguishes this population from all the others (Woods et al 2006). Therefore, the Cit variant
arose within the LTEE and is not a contaminant.
The evolved Cit

+

variant grows to high density in DM0 (a citrate-only medium),

produces vigorous colonies on minimal citrate (MC) agar plates, and causes a positive color
change on Simmon’s citrate agar, all of which indicate that it can use citrate as a sole carbon
+

source. In DM25, Cit cells undergo a period of rapid growth on glucose that is followed by
slower growth on citrate (Figure 2.2). Also, growth on citrate is inhibited by the citrate analog 5+

fluorocitrate (data not shown), as was observed for the one previously reported Cit mutant of E.
coli (Reynolds and Silver 1983).
+

Cit clones could be readily isolated from the frozen sample of population Ara–3 taken at
+

generation 33,000. To estimate the time of origin of the Cit trait, we screened 1280 clones
randomly chosen from generations 30,000, 30,500, 31,000, 31,500, 32,000, 32,500 and 33,000

88

+

Figure 2.1. Population expansion during evolution of the Cit phenotype.
Samples frozen at various times in the history of population Ara–3 were revived, and three
DM25 cultures established for each generation. Optical density (OD) was measured for each
culture at 24 h. Error bars show the range of three values measured for each generation. For
interpretation of the references to color in this and all other figures, the reader is referred to the
electronic version of this dissertation.

89

Figure 2.2.

–

+

Growth of Cit (blue triangles) and Cit

(red diamonds) cells in DM25

medium.
Each trajectory shows the average OD for eight replicate mixtures of three clones, all from
generation 33,000 of population Ara–3.

90

for the capacity to produce a positive reaction on Christensen’s citrate agar, which provides a
+

sensitive means to detect even weakly citrate-using cells. No Cit cells were found in the
+

samples taken at 30,000, 30,500 or 31,000 generations. Cit cells constituted ~0.5% of the
population at generation 31,500, then 15% and 19% in the next two samples, but only ~1.1% at
+

generation 33,000. It appears that the first Cit variant emerged between 31,000 and 31,500
generations, although we cannot exclude an earlier origin.
frequency of Cit

+

The precipitous decline in the

cells just prior to the massive population expansion suggests clonal
–

interference (Gerrish and Lenski 1998), whereby the Cit subpopulation produced a beneficial
+

mutant that out-competed the emerging Cit subpopulation until the latter evolved some other
beneficial mutation that finally assured its persistence. The hypothesis of clonal interference
+

implies that the early Cit cells were very poor at using citrate, such that a mutation that
improved competition for glucose could have provided a greater advantage than did marginal
exploitation of the unused citrate.
+

Indeed, the Cit clones isolated from generations 32,500 and earlier grow much more
slowly on citrate than those from 33,000 generations and later. After depleting the glucose in
+

DM25, the earliest Cit clones grow almost imperceptibly, if at all, for many hours before they
+

begin efficiently using the citrate (data not shown), whereas later Cit clones switch to growth on
citrate almost immediately (Figure 2.2). Thus, the population expansion between generations
33,000 and 33,500 (Figure 2.1) was triggered by one or more mutations that improved citrate
+

utilization, rather than by the growth of the original Cit mutant. This finding also raises the

91

+

question of whether weak Cit mutants might exist in any other LTEE population. We therefore
screened the other 11 populations, in most cases using samples from generation 41,500. None of
220 cultures inoculated with heterogeneous population samples grew in glucose-free DM0, nor
did any of >3,500 clones show a positive reaction on Christensen’s citrate agar.

+

Historical Contingency in the Evolution of Cit .
+

We performed three experiments to test whether the evolution of the Cit

function

required an unusually rare mutation or, alternatively, was historically contingent and depended
on the prior evolution of a certain genetic background. All three experiments used clones
sampled from many generations of population Ara–3 to replay evolution starting from different
genetic backgrounds. The two hypotheses make different predictions about the temporal pattern
+

of Cit evolution during these replay experiments (Figure 2.3). According to the rare-mutation
+

hypothesis, Cit variants should evolve at the same low rate regardless of the time of origin of
the clone with which a replay started. By contrast, the historical-contingency hypothesis predicts
+

that the mutation rate to Cit should increase after some potentiating genetic background has
+

evolved. Thus, Cit variants should re-evolve more often in the replays using clones sampled
from later generations of the Ara–3 population.
In our first experiment, we performed the replays under the same conditions as the LTEE.
We isolated three random clones from each of twelve time points, from the ancestor to 32,500
+

generations (Table 2.1), and obtained neutral Ara mutants of each clone to embed as protection
against accidental cross-contamination during the experiment (Lenski et al 1991). In total, there

92

Mutation rate to Cit+

+

Figure 2.3. Alternative hypotheses for the origin of the Cit function.
–

+

According to the rare-mutation hypothesis, the probability of mutation from Cit to Cit was low
but constant over time. Under the historical-contingency hypothesis, the probability of this
transition increased when an earlier mutation arose that produced a genetic background with a
+

higher mutation rate to Cit .

93

Table 2.1. Summary of replay experiments
First experiment

Second experiment

Third experiment

Independent
Independent

Cit

+

Independent

+

Cit

+

Generation

Reps

Cit mutants

Reps

mutants

Reps

mutants

Ancestor

6

0

10

0

200

0

5,000

–

–

–

–

200

0

10,000

6

0

30

0

200

0

15,000

–

–

–

–

200

0

20,000

6

0

30

0

200

2

25,000

6

0

30

0

200

0

27,000

–

–

–

–

200

2

27,500

6

0

30

0

–

–

28,000

–

–

–

–

200

0

29,000

6

0

30

0

200

0

30,000

6

0

30

0

200

0

30,500

6

1

30

0

–

–

31,000

6

0

30

0

200

1

31,500

6

1

30

0

200

1

32,000

6

0

30

4

200

2

32,500

6

2

30

1

200

0

Totals

72

4

340

5

2,800

8

94

were 72 replay populations, six from each generation, each founded by a single clone. These
populations evolved for ~3,700 generations, and they were checked visually each day for the
elevated turbidity indicative of the Cit

+

phenotype.

We also tested samples on MC and

Christensen’s citrate agar plates every 250 generations, with incubation for up to a week. New
+

Cit variants evolved in four replay populations, all founded by clones from later generations of
the original population (Table 2.1).

These Cit

+

variants emerged between 750 and 3,700

generations of the replay experiment.
+

Our second experiment also looked for Cit mutants derived from clones sampled at
various times in the history of population Ara–3. This time, however, we incubated large
populations of cells on MC plates, enabling us to test more clones and more cells of each clone.
We also allowed a long incubation time to facilitate the growth and detection of very weak Cit

+

mutants, as well as mutations that might occur as cells sat starving on the plates. In this
8

experiment, ~3.9 x 10 cells of each of the same 68 clones used in the first replay experiment
were spread on each of five MC plates, and these 340 plates were then incubated for 59 days.
+

Five plates produced Cit mutants, and all used clones from generations 32,000 or 32,500 of the
original population (Table 2.1).

None of the particular clones that evolved Cit

+

in this

experiment did so in the first one, although there was overlap in the generations from which
those clones were sampled.
The third replay experiment was similar in design to the second, but on a larger scale.
We isolated 20 clones from each of 13 time points in the history of population Ara–3, again
through 32,500 generations. We generated and tested 10 replicate cultures of each evolved clone

95

10

and 200 replicates of the ancestor. Each culture grew to ~1-2 x 10

cells, which were pelleted
13

by centrifugation, spread on an MC plate, and incubated for 45 days. In total, ~4.0 x 10

cells

+

were tested for their ability to use citrate in this experiment. Eight additional Cit mutants
evolved from seven different clones (Table 2.1), with one clone yielding mutants in two replicate
+

cultures. Four clones that produced Cit mutants came from generations 31,000 and later, two
were from generation 27,000, and one (the one that produced two mutants) was from generation
+

20,000. We found no Cit mutants among any of the 200 ancestral cultures, nor among any of
the other 600 cultures that used clones isolated before generation 20,000.
+

Interestingly, 7 of the 8 plates that yielded a mutant Cit colony produced multiple
colonies, including one with 137 colonies.

This pattern illustrates the “jackpot” effect

discovered by Luria and Delbrück (1943), and it implies that mutations arose during the
+

population growth prior to plating on MC agar. On the other hand, the Cit colonies were not
observed until at least 8 days of incubation and, in one case, they were first seen after 28 days.
+

Such late appearances suggest that the mutations to Cit occurred after plating. One possible
explanation for this apparent discrepancy is that the mutants grow very slowly but, in fact, they
typically produce visible colonies in only two days when re-tested on MC plates. Another
–

potential explanation is that the high density of Cit cells on the plates interfered with the growth
+

–

and detection of emerging Cit colonies. To test this possibility, we seeded dense Cit cultures
+

+

with a few Cit cells prior to plating on MC agar, but the Cit colonies were seen after only two
+

+

to three days. The rapid growth of Cit colonies occurred even when the Cit cells had grown

96

+

on glucose, and not on citrate, prior to plating. These results imply that mutations to Cit

occurred after cultures were plated on MC agar. This conclusion, taken together with the jackpot
+

distribution of Cit mutants, indicates that the phenotypic change required two mutations, one of
which occurred during the culture growth prior to plating and the other after plating.

Statistical Analysis of the Replay Experiments.
+

All three experiments show the same tendency for Cit variants to evolve more often
from clones sampled in later than earlier generations of population Ara–3 (Table 2.1). To
calculate the significance of these data, we performed Monte Carlo re-sampling tests (shuffling
without

replacement)

using

the

Statistics101

Re-sampling

Simulator

version

1.0.6

(www.statistics101.net). For each experiment, we compared the observed mean generation of
+

those clones that yielded Cit variants to the mean expected under the null hypothesis that clones
from all generations have equal likelihood. The null thus corresponds to the rare-mutation
hypothesis laid out in the introduction. We ran one million re-sampling iterations for each
experiment.

The deviations from the null expectations range from marginally to highly

significant in the three experiments, and in all cases they support the historical-contingency
hypothesis, according to which clones from later generations have greater propensity to evolve
+

the Cit phenotype (Table 2.2). Although the third experiment was the largest, it was the least
+

significant owing primarily to the production of two Cit mutants by a 20,000-generation clone.

97

Table 2.2. Statistical analyses of three replay experiments
Mean generation
First

Second

Third

experiment

experiment

experiment

Expected

24,917

28,382

22,571

Observed

31,750

32,100

27,563

Monte Carlo pvalue

0.0085

0.0007

0.0823

of clones yielding
Cit

+

98

We also used the Z-transformation method (Whitlock 2005) to combine the probabilities from
our three experiments, and the result is extremely significant (p < 0.0001) whether or not the
+

experiments are weighted by the number of independent Cit mutants observed in each one.
Furthermore, the potentiation effect in later generations is underestimated by these tests, because
the number of cells trended lower in later-generation cultures owing to the evolution of larger
cells that reach lower population density (Vasi et al 1994, Lenski and Mongold 2000, Hall 1982,
Elena et al 1996).
These analyses compel us to reject the hypothesis that a rare mutation could have
+

produced a Cit variant with equal probability at any point in the LTEE. Some unusually rare
mutation might be involved, but its rarity does not provide a sufficient explanation for the unique
and exceptionally slow evolution of this new function during the LTEE. Our results instead
support the hypothesis of historical contingency, in which a genetic background arose that had an
+

increased potential to evolve the Cit phenotype.

+

Rates of Mutation to Cit and Quantification of the Potentiation Effect.
Given the fraction of cultures that produce no mutants (P0) and the number of cells per
culture (N), one can estimate the mutation rate as μ = –(ln P0) / N. The corresponding 95% CI
(confidence interval) can also be calculated from the Poisson distribution based on the
uncertainty in P0, which is much greater than the uncertainty in N. By comparing the mutation
+

rate to Cit between clones, one can in principle quantify the potentiation effect in the evolved
genetic background. However, we cannot unambiguously identify which evolved clones remain

99

+

non-potentiated owing to the very low mutation rate to Cit even in the potentiated clones. We
also cannot separate potentiated and non-potentiated clones by generation because some
generational samples may be polymorphic.
We therefore performed an independent series of fluctuation tests using seven clones that
+

yielded Cit mutants in at least one replay experiment. These additional tests permit an unbiased
+

estimate of the mutation rate to Cit in the potentiated background. We prepared 40 replicate
cultures for each potentiated clone, and 280 for the ancestor. Potentiated and ancestral cultures
10

had, on average, 1.1 and 1.5 x 10

cells, respectively, which were harvested and incubated on

MC agar plates for 45 days.
+

None of the ancestral cultures yielded any Cit mutants. We can nonetheless calculate an
upper limit to the mutation rate by noting that the Poisson distribution has a 5% probability of
yielding zero events when the expectation is three. With no more than 3 mutations among the
8.4 x 10

12

cells tested here and in the third replay experiment, the upper bound on the ancestral
+

mutation rate to Cit is 3.6 x 10

-13

per cell per generation (Figure 2.4). To the best of our

knowledge, this value is the lowest upper bound ever reported for a mutation rate that has been
experimentally measured. It is also probably too high because no mutations were actually
12

observed for the ancestor, nor were any found among another 9.0 x 10

cells of 60 clones

sampled through 15,000 generations; and because some cell turnover and other DNA activity
probably occurred during the many days that plates were incubated.

100

–

+

–

+

Figure 2.4. Mutation rates from Ara to Ara (blue diamonds) and Cit to Cit (red
squares) of the ancestor and a set of potentiated clones.
Error bars are 95% confidence intervals. See text for details.

101

+

Even among the potentiated clones, the rate of mutation to Cit is extremely low. Cit
-13

mutants arose in two of the 280 new cultures, giving an estimate of 6.6 x 10
-14

rate, with the 95% CI extending from 7.9 x 10

-12

to 2.4 x 10

+

for the mutation

(Figure 2.4). (Although the

upper bound of the CI for the ancestor overlaps the lower bound of the CI for the potentiated
clones, that upper bound does not overlap the point estimate for the potentiated clones, indicating
a significant difference and adding further support to the replay experiments.) The potentiated
+

genetic background thus increases the mutation rate to Cit at least two fold, and probably much
more. However, even this potentiated value represents an unusually low mutation rate. A
-10

typical mutation rate in E. coli is ~5 x 10

per bp per generation (Drake 1991). Such a low rate

+

suggests that the final mutation to Cit is not a point mutation but instead involves some rarer
class of mutation or perhaps multiple mutations.

The possibility of multiple mutations is

especially relevant given our evidence that the emergence of Cit

+

colonies on MC plates

involved events both during the growth of cultures prior to plating and during prolonged
incubation on the plates.
Another issue is whether the potentiating effect might indicate the evolution of a
generalized hypermutability in population Ara–3. Previous surveys of mutation rates in the 12
LTEE lines found that four had become mutators by generation 20,000, although population
Ara–3 retained the low ancestral mutability (Cooper and Lenski 2000, Sniegowski et al 1997).
To investigate this issue further, we performed another series of fluctuation tests using the same
–

+

seven potentiated clones to determine their rate of mutation from Ara to Ara , which serves as a
proxy for the background mutation rate (Sniegowski et al 1997).

102

This mutation reverts a

phenotype that was knocked out during the derivation of a predecessor to the ancestor of the
+

LTEE (Lenski et al 1991, Lederberg 1966). Sequence analysis indicates that most Ara mutants
have a mutation from GAC to GGC at codon 92 of the araA gene, restoring that codon to its
distant ancestral state. Forty cultures of each clone were grown in DM1000, and the cells were
spread on minimal-arabinose plates. At the same time, 280 cultures of the ancestor were tested
+

-10

in the same way. The ancestral mutation rate to Ara is 2.3 x 10
the P0 method (95% CI from 1.8 x 10

-10

-10

to 2.9 x 10
-10

rate for potentiated clones is also 2.3 x 10

per cell per generation using

), a typical rate for point mutations. The

(95% CI from 2.0 x 10

these estimates are within 1% of each other (Figure 2.4).

-10

to 2.7 x 10

-10

); in fact,

We conclude that general
+

hypermutability is not responsible for the elevated mutation rate to Cit in the potentiated clones.

Frequency-dependent Selection Maintains Ecological Diversity.
To this point, we have examined the evolutionary origin of the Cit

+

function in

+

population Ara–3. We now turn to its ecological consequence. The Cit phenotype did not
–

achieve fixation during the population expansion but, instead, Cit cells persisted as a minority.
–

+

When we mixed Cit and Cit clones at different initial frequencies, they stably coexisted over
–

many serial transfers (Figure 2.5). In these mixtures, the Cit cells gradually approached an
equilibrium frequency of roughly 1%, regardless of their initial frequency. We saw a transient
–

jump in the frequency of the Cit subpopulation on days 7 and 8, which was probably caused by
accidentally using a glucose-only medium on those days. After that perturbation, the populations

103

2

-

log10 Cit /Cit

+

3

1
0

-1
-2
-3
0

Figure 2.5.

1

2

3

4

5

6
7
Day

8

9

10 11 12 13

–

+

Frequency-dependent selection allows stable coexistence of Cit and Cit

clones.
Each trajectory shows the mean of five replicate cultures for seven different initial ratios of Cit
+

and Cit clones from generation 33,000. Error bars are 95% confidence intervals.

104

–

resumed their previous trajectories. Negative frequency-dependent selection thus maintains the
polymorphism.
–

+

This stable coexistence suggests that the Cit cells are superior to the Cit cells in
competition for glucose, allowing the former to persist as glucose specialists. Indeed, the Cit

–

+

cells have a shorter lag phase and higher growth rate on glucose than do the Cit cells (Figure
2.2).

These differences were also evident when we monitored the intra-day dynamics of
–

+

mixtures of Cit and Cit cells (data not shown).

DISCUSSION AND FUTURE DIRECTIONS
E. coli cells cannot grow on citrate under oxic conditions, and that inability has long been
viewed as a defining characteristic of this important, diverse, and widespread species. In a longterm experiment, we propagated 12 populations of E. coli, all founded from the same ancestral
strain, in a medium containing glucose, which is the limiting resource, and abundant citrate. For
more than 30,000 generations, none of them evolved the capacity to use the citrate, although
billions of mutations occurred in each population, such that any typical base-pair mutation would
have been tested many times in each one. It is clearly very difficult for E. coli to evolve this
–

+

function. In fact, the mutation rate of the ancestral strain from Cit to Cit is immeasurably low;
-13

even the upper bound is 3.6 x 10

per cell generation, which is three orders of magnitude

below the typical base-pair mutation rate. Nevertheless, one population eventually evolved the
+

–

Cit function, while all the others remain Cit after more than 40,000 generations.
We demonstrated that the evolution of this new function was contingent upon the history
of the population in which it arose. In particular, we showed that one or more earlier mutations

105

+

potentiated the evolution of this function by increasing the mutation rate to Cit , although even
the elevated rate is much lower than a typical mutation rate. The potentiated cells are not
generally hypermutable.

Rather, their potentiation is specific to the Cit

suggests two possible mechanisms.

+

function, which

One mechanism is epistasis, whereby the functional
+

expression of the mutation that finally yielded the Cit phenotype requires interaction with one
or more mutations that evolved earlier. A second possibility is that the physical production of
+

the mutation that produced the Cit phenotype requires some previous mutation that allows the
final sequence to be generated. For example, the insertion of a mobile genetic element creates
new sequences at its junctures, and one of these new sequences might then undergo a mutation
that generates a final sequence that could not have occurred without the insertion. The E. coli
genome has many insertion-sequence elements (Schneider et al 2002), some of which have been
active in the LTEE (Papadopoulos et al 1999, Schneider et al 2000, Schneider and Lenski 2004).
+

Whatever the mechanism, this potentiation made the Cit function mutationally accessible, and a
+

weak Cit variant emerged by 31,500 generations.
+

The origin of the Cit function also had profound consequences for the ecology and
subsequent evolution of that population. This new capacity was refined over the next 2,000
+

generations, leading to a massive population expansion as the Cit cells evolved to exploit more
+

efficiently the abundant citrate in their environment. Although the Cit cells continued to use
–

–

glucose, they did not drive the Cit subpopulation extinct because the Cit cells were superior
competitors for glucose. Thus, the overall diversity increased as one population gave rise

106

evolutionarily to an ecological community with two members, one a resource specialist and the
other a generalist.
+

The evolution of the new Cit function represents a key innovation that involves multiple
steps, and it provides an explicit demonstration of the importance of historical contingency in
evolution. It also transcends the phenotypic boundaries of a diverse and well-studied species,
and led to an ecological transition from a single population to a two-member community. Our
future research on this fascinating case of evolution in action will revolve around four themes:
genetics, physiology, ecology, and speciation.
What is the genetic basis of this evolutionary innovation? The emergence of the Cit

+

phenotype in population Ara–3 indicates at least two important genetic events: the origin of the
function in its weak form, and its subsequent refinement for efficient use of citrate. The replay
experiments indicate an even more complex picture that must involve, at a minimum, three
important genetic events. At least one mutation in the LTEE was necessary to produce a genetic
+

background with the potential to generate Cit variants, while the distribution and dynamics of
+

Cit mutants in fluctuation tests indicate at least two additional mutations are involved. To find
the relevant mutations, we will perform whole-genome re-sequencing, which has become a
powerful approach that is well suited to experimental evolution (Velicer et al 2006, Herring et al
2006, Hegreness and Kishony 2007). We expect to find dozens of mutations relative to the
ancestor (Lenski 2004), which will complicate identification of those changes that were
+

important specifically for the origin of the Cit function. However, some of the key changes
–

should become apparent if we also re-sequence a Cit clone from the same population around the

107

+

time that the Cit variants first emerged.

Once candidate genes and mutations have been

identified, we can examine the other 19 Cit

+

variants from our replay and mutation-rate

experiments for parallel changes.
We are especially eager to find the potentiating mutation or mutations. We want to know
whether the potentiating mutation interacts epistatically with a later mutation to allow expression
+

of the Cit function or, alternatively, whether it was physically required for the later mutation to
occur.

We also want to test whether the potentiating mutation was itself beneficial or,

alternatively, a neutral or deleterious change that fortuitously hitchhiked to high frequency. We
anticipate that identifying the potentiating mutation will be especially challenging, however,
because its only known phenotype is to increase the rate of production of certain mutants that are
themselves extremely rare.
Once we have identified all the relevant mutations, it might be interesting to model the
population dynamics that govern the emergence of this new function. Such a model would
require not only all the relevant mutation rates but also the ecological phenotypes of the mutants,
including their growth rates on glucose and citrate as well as their abilities to transition between
the two resources. A satisfactory model should also reflect the stochastic origin of mutations, the
role of random drift, and the possibility of alternative mutational paths to the phenotype of
interest (Weinreich et al 2006, Lenski et al 2003).
What physiological mechanism has evolved that allows aerobic growth on citrate? E.
coli should be able to use citrate as an energy source after it enters the cell, but it lacks a citrate
+

transporter that functions in an oxygen-rich environment. One possibility is that the Cit lineage
activated a “cryptic” transporter (Hall 1982), that is, some once-functional gene that has been

108

silenced by mutation accumulation. This explanation seems unlikely to us because the Cit

–

phenotype is characteristic of the entire species, one that is very diverse and therefore very old.
We would expect a cryptic gene to be degraded beyond recovery after millions of years of
disuse. A more likely possibility, in our view, is that an existing transporter has been coopted for
citrate transport under oxic conditions. This transporter may previously have transported citrate
under anoxic conditions (Pos et al 1998) or, alternatively, it may have transported another
substrate in the presence of oxygen. The evolved changes might involve gene regulation, protein
structure, or both (Mortlock 1984).
–

+

What will be the long-term fates of the coexisting Cit and Cit subpopulations? We
+

showed they stably coexist owing to the inferiority of the Cit cells in competition for glucose.
+

However, the Cit lineage might eventually acquire mutations that compensate for its inferior
performance on glucose, thus undermining the coexistence. Compensation for maladaptive sideeffects of adaptations, including resistance to phages and antibiotics, has been observed in many
other experiments with bacteria (Lenski 1988, Bouma and Lenski 1988, Schrag et al 1997,
+

Lenski 1997, Bjorkman et al 1998). Moreover, the Cit subpopulation is much larger than the
Cit

–

subpopulation, so it should experience more beneficial mutations even without

compensation. On the other hand, coexistence would be strengthened if selection in the Cit

+

subpopulation favors specialization on citrate. In the same way that we established multiple
populations for retrospective replays in this study, we can establish multiple communities to
examine their prospective evolution. We can also vary environmental factors, such as the

109

–

presence or absence of glucose, and the presence or absence of a Cit subpopulation, to
+

investigate how they influence the future evolution of the Cit lineage.
+

–

Will the Cit and Cit lineages eventually become distinct species? According to the
biological species concept widely used for animals and plants, species are recognized by
reproductive continuity within species and reproductive barriers leading to genetic isolation
between species (Mayr 1942). Although the bacteria in the LTEE are strictly asexual, we can
nonetheless imagine testing this criterion by producing recombinant genotypes. In particular, we
+

–

could move mutations that are substituted in the evolving Cit lineage into a Cit background to
test whether they reduce fitness in their ancestral context. One could also perform the reciprocal
+

experiment, though we anticipate more rapid evolution in the Cit

lineage because it has

acquired a key innovation that substantially changed its ecological niche. Such experiments
would require, of course, controls to examine the fitness effects of the same mutations in the
+

lineage where they arose. If the Cit lineage is indeed evolving into a new species, then we
expect, with time, that more and more of the beneficial mutations substituted in that lineage
–

would be detrimental in the ecological and genetic context of its Cit progenitor.
In any case, our study shows that historical contingency can have a profound and lasting
impact under the simplest, and thus most stringent, conditions in which initially identical
populations evolve in identical environments.

Even from so simple a beginning, small

happenstances of history may lead populations along different evolutionary paths. A potentiated
cell took the one less traveled by, and that has made all the difference.

110

MATERIALS AND METHODS
The Long-term Evolution Experiment
The LTEE is described in detail elsewhere (Lenski et al 1991, Lenski 2004). Briefly, two
ancestral clones of E. coli B were each used to found six populations. The ancestors differ by a
+

–

+

single mutation that allows one of them to use arabinose (Ara ). Ara and Ara cells make red
and white colonies, respectively, on tetrazolium-arabinose (TA) plates, but the mutation is
neutral in the environment of the LTEE (Lenski et al 1991). The twelve populations have been
propagated for almost 20 years by daily serial dilution in DM25, a minimal salts medium that has
139 μM glucose and 1,700 µM citrate (Lenski et al 1991). Given 1:100 dilution and re-growth,
the populations achieve ~6.64 generations per day, and they have evolved for over 40,000
generations in this experiment to date. Every 500 generations, population samples are frozen at
–80ºC with glycerol added as a cryoprotectant. These samples contain all of the diversity present
in a population at that generation. Except as otherwise noted, all bacteria used in this study come
from samples of population Ara–3.

Media and Culture Conditions
DM0 is identical to DM25 except that it contains no glucose. MC plates have the same
formulation as liquid DM0, except that agar is added to solidify the medium and the
concentration of citrate is higher to allow large colonies. Prior to starting an experiment, bacteria
were preconditioned in the relevant experimental medium unless otherwise stated.
additional details, see Supporting Information (SI) Materials and Methods.

111

For

Analysis of the Population Expansion
When the increased turbidity was observed in population Ara–3, we froze samples of the
two previous days’ cultures, which had been temporarily stored at 4ºC. We also froze samples
after every transfer for several weeks. We later revived these and other samples taken between
generations 30,000 and 35,000 to quantify the population expansion.

For details, see SI

Materials and Methods.

+
Searches for Cit Variants
+

In order to determine when the earliest Cit variant arose in population Ara–3, we
randomly chose 1280 clones from samples frozen at generations 30,000, 30,500, 31,000, 31,500,
32,000, 32,500 and 33,000. We transferred colonies to Christensen’s citrate agar plates, which
were incubated for 10 days and examined daily for evidence of positive reactions. All putative
+

Cit clones were retested on Christensen’s agar for confirmation, and a subset was tested on MC
agar. All clones that had positive reactions on Christensen’s plates also grew on MC plates. In
+

order to find Cit cells that might be present at low frequency in the other 11 LTEE populations,
we tested their genetically heterogeneous full-population samples for growth in DM0. We
allowed the top of each sample to thaw, then inoculated 50 µl into LB broth. After overnight
growth, five 10-mL DM0 cultures were inoculated with 100 µL of 100-fold dilutions of each LB
culture. The DM0 cultures were incubated for 28 days with periodic checks for turbidity.

112

Growth Curves
–

+

Three representative Cit clones and three representative Cit clones from generation
–

33,000 were separately preconditioned in DM25. The Cit clones were then combined, as were
+

4

the Cit clones, and the mixtures were diluted in DM25 to ~3.4 x 10 cells per mL. Eight 200μL aliquots of each mixture were dispensed into randomly assigned wells in a 96-well plate, and
optical density (OD) at 420 nm wavelength was measured periodically using a VERSAmax
automated plate reader.

Replay Experiments
In the first replay experiment, populations evolved under the same conditions as the
LTEE. In the second and third replay experiments, cells were incubated on MC plates. We
rd

started the first replay experiment on the 3 anniversary of Stephen Jay Gould’s death; we ended
th

it on the 66 anniversary of his birth. For further details, see SI Materials and Methods.

+

Confirmation of Cit Variants
+

Each putative Cit variant was streaked on MC agar and Christensen’s citrate agar to
confirm its phenotype. Additional tests are described in the SI Materials and Methods.

113

Test for Frequency-dependent Interaction
–

We constructed mixtures of clones from generation 33,000 to test whether Cit and Cit

+

clones coexisted and, if so, whether they interacted in a frequency-dependent manner. For
details, see SI Materials and Methods.

ACKNOWLEDGEMENTS
We thank N. Hajela for years of outstanding technical assistance, and M. Adawe for help with
media preparation; together they poured some 20,000 agar plates used in our experiments. We
thank Francisco Ayala, Al Bennett, Rita Colwell, Bruce Levin, and Simon Levin for reviewing
our paper prior to publication. This research has been supported in part by grants from the
National Science Foundation (currently DEB-0515729) and the Defense Advanced Research
Projects Agency “Fun Bio” Program (HR0011-05-1-0057).

SUPPORTING INFORMATION

SI Materials and Methods
Media and Culture Conditions
The various DM media used to grow cells in different experiments are all based on the
same formulation as the DM25 used in the LTEE (Lenski et al 1991), with only glucose
concentration varied. DM25 contains glucose at 25 mg/L. DM500, DM1000, and DM2000 have
20, 40, and 80 times, respectively, more glucose than DM25, while DM0 contains no glucose.
MC plates have the same formulation as liquid DM0, except that agar is added at 16 g/L and

114

trisodium citrate dihydrate is increased to 4.5 g/L. Simmon’s and Christensen’s citrate agar
media are described elsewhere (Atlas 2006).
Prior to starting an experiment, unless stated otherwise, clones or population samples
were inoculated from frozen stocks into LB broth and incubated overnight. Next, the bacteria
were preconditioned by diluting them 10,000-fold into 10 mL of the relevant experimental
medium, and they were incubated for 24 h. Cultures were then diluted 100-fold into a second
volume of experimental medium, and again incubated for 24 h.
Analysis of the Population Expansion
When the increased turbidity was observed in population Ara–3, we froze samples of the
two previous days’ cultures, which had been temporarily stored at 4ºC. We also froze samples
after every transfer for several weeks. We later revived the samples from generations 30,000,
31,000, 32,000, 32,500, 33,000, 33,113, 33,120, 33,127, 33,133, 33,140, 33,180, 33,200, 33,293,
33,413, 33,500, 33,760, 34,000, 34,260, 34,500, 34,760, and 35,000 in LB, and preconditioned
them for 24 h in DM25. We then transferred each one into three replicate DM25 cultures. After
24 h, we moved 200-µL samples of each culture to a 96-well plate, and measured optical density
at 420 nm wavelength using a VERSAmax automated plate reader.
First Replay Experiment
Three clones were chosen at random from Ara–3 population samples from 10,000,
20,000, 25,000, 27,500, 29,000, 30,000, 30,500, 31,000, 31,500, 32,000, and 32,500 generations.
–

+

All these clones were Cit , including those from samples taken after the weak Cit variant
+

emerged. Spontaneous Ara mutants were derived from each clone by plating on minimalarabinose agar (Lensk et al 1991). Two populations were founded by each evolved clone, one
115

+

using the Ara– clone and one its Ara derivative. Six populations were founded by the ancestral
–

+

strain, including three each using the Ara and Ara

clones.

These 72 populations were

propagated by daily 1:100 dilutions in DM25, as in the LTEE except that these replays ran in
unshaken test-tubes. Each culture was inspected at every transfer for increased turbidity that
+

would indicate the re-evolution of the Cit phenotype. Undiluted samples from each culture
were spread on MC and Christensen’s citrate agar plates at ~250-generation intervals, at which
time population samples were frozen for long-term storage.

Excluding interruptions, this

experiment ran for 557 days, or ~3,700 generations with 6.64 generations per day.
Second and Third Replay Experiments
The second replay experiment used the same 68 clones as the first one. For each clone,
we inoculated five 10-mL cultures of DM500 with <60 cells. After 48 h, the cells in each culture
were pelleted by centrifugation (~5000 g for 10 min), and then re-suspended, with half of each
9

culture (~4 x 10 cells) spread on an MC plate and the other half on an unamended agar plate as
a control. The plates were incubated for 59 days with periodic checks for colonies.
For the third replay experiment, we isolated 20 Cit

–

clones each from the Ara–3

population samples from 5,000, 10,000, 15,000, 20,000, 25,000, 27,000, 28,000, 29,000, 30,000,
31,000, 31,500, 32,000 and 32,500 generations. Clones were chosen in order to maximize the
diversity of colony morphologies from each sample, in case only certain types could produce
+

Cit mutants. For those generations included in the first and second experiments, the three
clones used there were included among the 20 clones used here.

116

To facilitate handling and minimize possible confounding variables, we divided this third
experiment into 20 blocks of 14 clones each. All of the clones within a block came from
different generations, and the single ancestral clone was included in all 20 blocks. For every
clone in each block, we inoculated ten replicate 10-mL cultures with <60 cells. Seven blocks
8

9

used DM1000 medium, in which the bacteria achieved densities of ~8 x 10 to 1.5 x 10 cells
per mL. The other 13 blocks used DM2000, in which the cells reached densities about twice as
high. Cultures were incubated for 48 h with shaking for aeration. We then diluted a small
volume of every second culture on TA plates to estimate final cell numbers. Each culture was
pelleted, re-suspended, and spread in its entirety on an MC plate. These plates were incubated
for 45 days, with periodic checks for colonies.
+

The number of Cit mutants found in the second experiment led us to expect many more
mutants in our third replay experiment, but that expectation was not fulfilled. Subtle differences
in procedures or conditions evidently may affect the mutation rate or selective enrichment of
these mutants. We emphasize, however, that the same procedures and conditions were applied
simultaneously to clones sampled from all generations in each experiment. Also, in the second
and third replay experiments, some plates became contaminated during the repeated checks for
mutants, and they were discarded before the experiment ended.

Such contamination was

infrequent and haphazard, and so it did not affect our analysis.
With hindsight, we should probably have grown the cells to be spread on the MC plates
in the second and third experiments in medium without citrate, in order to avoid possible
+

selection for Cit mutants prior to plating. However, the presence of citrate in the growth
medium made no practical difference for several reasons: (i) the final mutations that gave the

117

Cit

+

phenotype arose on the plates; (ii) the same protocol was used for clones from all

generations; and (iii) the P0 method for estimating mutation rates is insensitive to the timing and
growth rate of mutants relative to non-mutants.
+

Confirmation of Cit Variants
+

Each putative Cit variant was streaked on MC agar and Christensen’s citrate agar to
confirm its phenotype. One colony was then selected from the MC plate and its Ara marker
status, sensitivity to phage T5, and resistance to phage T6 were checked to confirm that it was
derived from the ancestral E. coli B strain (Lenski et al 1991). We also sequenced the pykF and
+

–

nadR loci of Cit variants and their parental Cit clones to confirm single base-pair substitutions
that uniquely identify the Ara–3 population (Woods et al 2006).
Test for Frequency-dependent Interaction
–

+

We constructed mixtures of six clones from generation 33,000 to test whether Cit and Cit

clones coexisted and, if so, whether they interacted in a frequency-dependent manner. Three
+

–

+

clones were Cit and three were Cit , and the latter set carried the neutral Ara marker to help us
distinguish the two types using TA agar. The six clones were revived, preconditioned, and
6

combined in DM25, with ~3.4 x 10 cells in total used to inoculate each mixed culture. To test
for frequency-dependent effects, we made mixtures with seven different initial frequencies of
–

Cit cells: 0.1%, 1%, 10%, 50%, 90%, 99%, and 99.9%. Each initial frequency was replicated
five-fold, and all 35 mixed cultures were serially propagated with 100-fold dilution in DM25 for
13 days. As noted in the Results, we may have accidentally transferred these cultures into a

118

glucose-only medium for two days in the middle of this experiment. To estimate the abundances
–

+

of the Cit and Cit types, we plated aliquots of each mixed culture on TA agar at the start and at
every daily transfer.

119

REFERENCES

120

REFERENCES
1. Atlas, R.M. Handbook of Microbiological Media for the Examination of Food (CRC
Press, Boca Raton, FL, 2006).
2. Beatty, J. Replaying life’s tape. Journal of Philosophy 7, 336–362 (2006).
3. Bjorkman, J., Hughes, D., & Andersson, D.I. Virulence of antibiotic-resistant Salmonella
typhimurium. Proceedings of the National Academies of Science (USA) 95, 3949–3953
(1998).
4. Bouma, J.E. & Lenski, R.E. Evolution of a bacteria/plasmid association. Nature 335,
351-352 (1988).
5. Conway Morris, S. Life’s Solution (Cambridge Univ. Press, Cambridge, MA, 2003).
6. Cooper, V.S. & Lenski, R.E. The population genetics of ecological specialization in
evolving E. coli populations. Nature 407, 736–739 (2000).
7. Cooper, V.S., Schneider, S., Blot, M., & Lenski, R.E. Mechanisms causing rapid and
parallel losses of ribose catabolism in evolving populations of E. coli B. Journal of
Bacteriology 183, 2834–2841 (2001).
8. Cooper, T.F., Rozen, D.E., & Lenski, R.E. Parallel changes in gene expression after
20,000 generations of evolution in E. coli. Proceedings of the National Academy of
Sciences (USA) 100, 1072–1077 (2003).
9. Cooper, T.F., Remold, S.K., Lenski, R.E., & Schneider, D. Expression profiles reveal
parallel evolution of epistatic interactions involving the CRP regulon in Escherichia coli.
PLoS Genetics 4, e35 (2008).
10. Crozat, E., Philippe, N., Lenski, R,E., Geiselmann, J., & Schneider, D, Long-term
experimental evolution in Escherichia coli. XII. DNA topology as a key target of
selection. Genetics 169, 523–532 (2005).
11. Dawkins, R. The Blind Watchmaker (Norton, New York, NY, 1996).
12. Drake, J.W. A constant rate of spontaneous mutation in DNA-based microbes.
Proceedings of the National Academy of Sciences (USA) 88, 7160–7164 (1991).
13. Elena, S.F., Cooper, V.S., & Lenski, R.E. Punctuated evolution caused by selection of
rare beneficial mutations. Science 272, 1802–1804 (1996).
14. Foote, M.J. Contingency and convergence. Science 280, 2068–2069 (1998).
15. Frost, G.E. & Rosenberg, H. The inducible citrate-dependent iron transport system in
Escherichia coli K-12. Biochimica et Biophysica ACTA 330, 90–101 (1973).

121

16. Gerrish, P.J. & Lenski, R.E. The fate of competing beneficial mutations in an asexual
population. Genetica 102/103, 127–144 (1998).
17. Gould, S.J. The paradox of the first tier: an agenda for paleobiology. Paleobiology 11, 2–
12 (1985).
18. Gould, S.J. Wonderful Life (Norton, New York, NY, 1989).
19. Gould, S.J. The Structure of Evolutionary Theory (Belknap, Cambridge, 2002).
20. Hall, B.G. Chromosomal mutation for citrate utilization by Escherichia coli K-12.
Journal of Bacteriology 151, 269–273 (1982).
21. Hegreness, M. & Kishony, R. Analysis of genetic systems using experimental evolution
and whole-genome sequencing. Genome Biology 8, 201 (2007).
22. Herring, C.D. et al Comparative genome sequencing of Escherichia coli allows
observation of bacterial evolution on a laboratory timescale. Nature Genetics 38, 1406–
1412 (2006).
23. Hussein, S., Hantke, K., & Braun, V. Citrate-dependent iron transport system in
Escherichia coli K-12. European Journal of Biochemistry 117, 431–437 (1981).
24. Ishiguro, N., Oka, C., & Sato, G. Isolation of citrate positive variants of Escherichia coli
from domestic pigeons, pigs, cattle, and horse. Applied and Environmental Microbiology
36, 217–222 (1978).
25. Ishiguro, N., Oka, C., Hanazawa, Y., & Sato, G. Plasmids in Escherichia coli controlling
citrate-utilizing ability. Applied and Environmental Microbiology 38, 956–964 (1979).
26. Jablonski, D. Background and mass extinctions: the alternation of macroevolutionary
regimes. Science 231, 139-133 (1986).
27. Jacob F Evolution and tinkering. Science 196, 1161–1166 (1977).
28. Koser, S.A. Correlation of citrate-utilization by members of the colon-aerogenes group
with other differential characteristics and with habitat. Journal of Bacteriology 9, 59–77
(1924).
29. Lara, F.J.S. & Stokes, J.L. Oxidation of citrate by Escherichia coli. Journal of
Bacteriology 9, 59–77 (1952).
30. Lederberg, S. Genetics of host-controlled restriction and modification of
deoxyribonucleic acid in Escherichia coli. Journal of Bacteriology 91, 1029–1036
(1966).

122

31. Lenski, R.E. Experimental studies of pleiotropy and epistasis in Escherichia coli. II.
Compensation for maladaptive pleiotropic effects associated with resistance to virus T4.
Evolution 42, 433–440 (1988).
32. Lenski, R.E. The cost of antibiotic resistance – from the perspective of a bacterium. CIBA
Foundation Symposium 207, 131–140 (1997).
33. Lenski, R.E. Phenotypic and genomic evolution during a 20,000-generation experiment
with the bacterium Escherichia coli. Plant Breeding Reviews 24, 225–265 (2004).
34. Lenski, R.E., Rose, M.R., Simpson, S.C., & Tadler, S.C. Long-term experimental
evolution in Escherichia coli. I. Adaptation and divergence during 2,000 generations.
American Naturalist 138, 1315–1341 (1991).
35. Lenski, R.E. & Travisano, M. Dynamics of adaptation and diversification: a 10,000generation experiment with bacterial populations. Proceedings of the National Academy
of Sciences (USA) 91, 6808–6814 (1994).
36. Lenski, R.E. & Mongold, J.A. in Scaling in Biology, eds Brown J, West G (Oxford Univ.
Press, Oxford, UK), pp 221–235 (2000).
37. Lenski, R.E., Ofria, C., Pennock, R.T., & Adami, C. The evolutionary origin of complex
features. Nature 423, 139–144 (2003).
38. Lenski, R.E., Winkworth, C.L., & Riley, M.A. Rates of DNA sequence evolution in
experimental populations of Escherichia coli during 20,000 generations. Journal of
Molecular Evolution 56, 498–508 (2003).
39. Lewontin, R.C. Is nature probable or capricious? BioScience 16, 25–27 (1966).
40. Luria, S.E. & Delbrück, M. Mutations of bacteria from virus sensitivity to virus
resistance. Genetics 28, 491–511 (1943).
41. Lutgens, M. & Gottschalk, G. Why a co-substrate is required for anaerobic growth of
Escherichia coli on citrate. J Gen Microbiol 199, 63–70 (1980).
42. Mani, G.S. & Clarke, B.C. Mutational order: a major stochastic process in evolution.
Proceedings of the Royal Society of London Series B: Biological Sciences 240, 29–37
(1990).
43. Mayr, E. Systematics and the Origin of Species from the Viewpoint of a Zoologist
(Columbia Univ. Press, New York, 1942).
44. Mongold, J.A., Bennett, A.F., & Lenski, R.E. Evolutionary adaptation to temperature.
VII. Extension of the upper thermal limit of Escherichia coli. Evolution 53, 386–394
(1999).

123

45. Mortlock, R.P., ed. Microorganisms as Model Systems for Studying Evolution (Plenum,
New York, 1984).
46. Novak, M., Pfeiffer, T., Lenski, R.E., Sauer, U., & Bonhoeffer, S. Experimental tests for
an evolutionary trade-off between growth rate and yield in E. coli. American Naturalist
168, 242–251 (2006).
47. Pelosi, L. et al Parallel changes in global protein profiles during long-term experimental
evolution in Escherichia coli. Genetics 173, 1851–1869 (2006).
48. Papadopoulos, D. et al. Genomic evolution during a 10,000-generation experiment with
bacteria. Proceedings of the National Academy of Sciences (USA) 96, 3807–3812 (1999).
49. Pos, K.M., Dimroth, P., & Bott, M. The Escherichia coli citrate carrier CitT: a member of
a novel eubacterial transporter family related to the 2-oxoglutarate-malate translocator
from spinach chloroplasts. Journal of Bacteriology 180, 4160–4165 (1998).
50. Reynolds, C.H., & Silver, S. Citrate utilization by Escherichia coli: plasmid- and
chromosome-encoded systems. Journal of Bacteriology 156, 1019–1024 (1983).
51. Scheutz, F. & Strockbine, N.A. in Bergey’s Manual of Systematic Bacteriology, Volume
Two: The Proteobacteria, eds Garrity GM, Brenner DJ, Kreig NR, Staley JR (Springer,
New York, NY), pp 607–624 (2005).
52. Schneider, D. & Lenski, R.E. Dynamics of insertion sequence elements during
experimental evolution of bacteria. Research in Microbiology 155, 319–327 (2004).
53. Schneider, D., Duperchy, E., Coursange, E., Lenski, R.E., & Blot, M. Long-term
experimental evolution in Escherichia coli. IX. Characterization of IS-mediated
mutations and rearrangements. Genetics 156, 477–488 (2000).
54. Schneider, D. et al. Genomic comparisons among Escherichia coli strains B, K-12, and
O157:H7 using IS elements as molecular markers. BMC Microbiology 2, 18 (2002).
55. Schrag, S.J., Perrot, V., & Levin, B.R. Adaptation to the fitness costs of antibiotic
resistance in Escherichia coli. Proceedings of the Royal Society of London Series B:
Biological Sciences 264, 1287–1291 (1997).
56. Sniegowski, P.D., Gerrish, P.J., & Lenski, R.E. Evolution of high mutation rates:
separating causes from consequences. Nature 387, 703–705 (1997).
57. Travisano, M., Vasi, F., & Lenski, R.E. Long-term experimental evolution in Escherichia
coli. III. Variation among replicate populations in correlated responses to novel
environments. Evolution 49, 189–200 (1995).
58. Travisano, M., Mongold, J.A., Bennett, A.F., & Lenski, R.E. Experimental tests of the
roles of adaptation, chance, and history in evolution. Science 267, 87–90 (1995).

124

59. Van Valen, L.M. How far does contingency rule? Evolutionary Theory 10, 47–52 (1991).
60. Vasi, F., Travisano, M., & Lenski, R.E. Long-term experimental evolution in Escherichia
coli. II. Changes in life-history traits during adaptation to a seasonal environment.
American Naturalist 144, 432–456 (1994).
61. Velicer, G.J. et al. Comprehensive mutation identification in an evolved bacterial
cooperator and its cheating ancestor. Proceedings of the National Academy of Sciences
(USA) 103, 8107–8112 (2006).
62. Vermeij, G.J. Historical contingency and the purported uniqueness of evolutionary
innovations. Proceedings of the National Academy of Sciences (USA) 103, 1804–1809
(2006).
63. Weinreich, D.M., Watson, R.A., & Chao, L. Sign epistasis and genetic constraint on
evolutionary trajectories. Evolution 59, 1165–1174 (2005).
64. Weinreich, D.M., Delaney, N.F., DePristo, M.A., & Hartl, D.L. Darwinian evolution can
follow only very few mutational paths to fitter proteins. Science 312, 111–114 (2006).
65. Whitlock, M.C. Combining probability from independent tests: the weighted Z-method is
superior to Fisher's approach. Journal of Evolutionary Biology 18, 1368–1373 (2005).
66. Woods, R., Schneider, D., Winkworth, C.L., Riley, M.A., & Lenski, R.E. Tests of
parallel molecular evolution in a long-term experiment with Escherichia coli.
Proceedings of the National Academy of Sciences (USA) 103, 9107–9122 (2006).
67. Wright, S. Surfaces of selective value revisited. American Naturalist 131, 115–123
(1988).
68. Yedid, G. & Bell, G. Macroevolution simulated with autonomously replicating computer
programs. Nature 420, 810–812 (2002).

125

CHAPTER 3

Blount, Z.D., Barrick, J.E., Davidson, C.J., & Lenski, R.E. Genomic analysis of a key
innovation in an experimental E. coli population. Submitted for review.

126

GENOMIC ANALYSIS OF A KEY INNOVATION IN AN EXPERIMENTAL E. COLI
POPULATION

ABSTRACT
Evolutionary novelties have been important in the history of life, but their origins are
usually difficult to examine in detail. We previously described the evolution of a novel trait,
+

aerobic citrate utilization (Cit ), in an experimental population of Escherichia coli. Here we
analyze 29 genomes to investigate the history and genetic basis of this trait. At least three
+

distinct clades coexisted for more than 10,000 generations prior to its emergence. The Cit trait
originated in one clade by a tandem duplication that captured an aerobically-expressed promoter
for the expression of a previously silent citrate transporter. The clades varied in their propensity
to evolve this novel trait, although genotypes able to do so existed in all three clades, implying
that multiple potentiating mutations arose during the population’s history. Our findings illustrate
the importance of promoter capture and altered gene regulation in mediating the exaptation
events that often underlie evolutionary innovations.

127

BACKGROUND AND INTRODUCTION
Evolutionary novelties are qualitatively new traits that open up ecological opportunities
and thereby promote diversification (Mayr 1960, Piliucci 2008). These traits are thought to arise
largely by the exaptation of genes that previously encoded other functions (Pigliucci 2008, Jacob
1977, 1982, Gould and Vrba 1982) via such processes as domain shuffling (Patthy 1999), altered
regulation (True and Carroll 2002), and duplication followed by neo-functionalization (Zhang
2003, Bergthorsson et al 2007). Multiple mutations may be necessary to produce the new
function (Zhang 2003, Lenski et al 2003), and thus the potential to evolve that function may be
contingent on subtle genetic differences between species, populations, or even genotypes. A
complete understanding of the evolution of a novel trait requires explanation of its ecological
function, its mechanistic basis, the genetic changes that underlie its evolution, and the history of
the accumulated changes (Pigliucci 2008).
Evolution experiments with bacteria present unparalleled opportunities to assess these
issues (Elena and Lenski 2003, Lenski 2004).

Bacteria have rapid generations and large

populations, making it possible to observe the emergence of novel traits, while new technologies
allow discovery of mutations throughout their genomes (Bentley 2006, Hegreness and Kishony
2007, Barrick et al 2009, Barrick and Lenski 200). Samples of evolving populations can be
frozen and later revived, allowing phylogenies and mutational histories to be constructed from
ancestral, intermediate, and evolved genome sequences (Barrick et al 2009, Barrick and Lenski
2009).
Twelve populations of Escherichia coli have been propagated in the long-term evolution
experiment (LTEE) for over 40,000 generations in a glucose-limited minimal medium (Lenski
2004). The medium also contains abundant citrate, which is present as a chelating agent (Blount

128

et al 2008), but E. coli cannot exploit citrate as a carbon and energy source in the well-aerated
conditions of the experiment (Koser 1924, Lutgens and Gottschalk 1980). Indeed, the inability
to grow aerobically on citrate is a long-recognized trait that, in part, defines E. coli as a species21.
+

+

Spontaneous citrate-using (Cit ) mutants are extraordinarily rare (Hall 2982), but a Cit variant
evolved in one LTEE population (designated Ara–3) after 31,000 generations (Blount et al
+

–

2008). Cit cells became numerically dominant after 33,000 generations, although Cit cells
persisted. This shift was accompanied by a substantial increase in total population size owing to
the high concentration of citrate relative to glucose in the medium.
+

The eventual origin of the Cit trait was contingent upon one or more earlier mutations in
the population’s history. When evolution was replayed from clones isolated at various time+

points, clones from later generations had a significantly greater propensity to become Cit than
did the ancestor and other early clones (Blount et al 2008). This finding implied that a genetic
background had evolved that “potentiated” the evolution of this trait. In principle, this effect
could involve two distinct mechanisms. One possibility is that the rate of mutation, or certain
types of mutation, increased such that the required event was more likely to occur in later
generations. The other possibility is an epistatic interaction, such that the phenotypic expression
+

of a mutation that produced the Cit trait required one or more preceding mutations. Here we
report the results of extensive whole-genome re-sequencing that allowed us to reconstruct the
+

population’s history, identify and manipulate mutations underlying the Cit phenotype, and
elucidate the mechanistic basis for this novel trait.

129

GENOME SEQUENCING AND PHYLOGENETIC HISTORY
+

We sequenced 29 clones sampled at various generations from population Ara–3, including 9 Cit
–

and 3 Cit clones known to be potentiated (Blount et al 2008) (Table 3.1). Mutations including
single-nucleotide polymorphisms (SNPs) as well as deletions, insertions, and other
rearrangements were identified by comparing DNA sequence reads to the complete genome of
the ancestral strain, REL606 (Jeong et al 2009).
We constructed the population’s phylogenetic history from the presence-absence matrices
of all mutations in the sequenced genomes.

Figure 3.1 shows that the population was

polymorphic for most of its history. At least four major clades, each represented by multiple
clones, emerged before 20,000 generations. One of them, which we call UC (Unsuccessful
Clade), was not seen after 15,000 generations. The remaining clades C1, C2, and C3 coexisted
+

through the evolution of Cit and beyond. C1 includes four sequenced clones, the earliest from
generation 25,000, although a molecular clock (Figure 3.2) implies that C1 diverged from the
ancestor of C2 and C3 before 15,000 generations. C2 and C3 had diverged by generation
–

+

+

20,000. C3 includes both Cit clones and all the Cit clones, although the first Cit cells did not
arise until ~31,000 generations. Two mechanisms might explain the long-term coexistence of
–

the Cit lineages. First, they may all have acquired beneficial mutations without one gaining
enough advantage to fix (Fogle et al 2008). Alternatively, they might have filled subtly different
ecological niches, such as through the differential production of and growth on secreted
metabolites (Treves et al 1998, Rozen et al 2005).

130

Supplementary Table 3.1 │ Historical Ara–3 Clones Subjected to
Whole Genome Sequencing
Generation
0
2,000
5,000
10,000
15,000
20,000

25,000
30,000

31,500

32,000
32,500

33,000

34,000
36,000
38,000
40,000

Clones
REL606
REL1166A
ZDB409
ZDB429
ZDB446
ZDB458
ZDB464*

Clade
N/A
?
?
UC
UC
(C1,C2)
(C1,C2)

ZDB467
ZDB477
ZDB483
ZDB16
ZDB357
ZDB199*
ZDB200
ZDB564
ZDB30*
ZDB172
ZDB158
ZDB143
CZB199
CZB152
CZB154
ZDB83
ZDB87
ZDB96
ZDB99
ZDB107
ZDB111
REL10979
REL10988

(C1,C2)
C1
C3
C1
C2
C1
C2
Cit+
C3
Cit+
C2
Cit+
C1
Cit+
Cit+
Cit+
C2
Cit+
C1
Cit+
C2
Cit+
C2

131

Clade 1
Cit Replays

Clade 2

Clade 3

2/55

+

2/97

8/37
+

New Cit Clade
40,000

30,000

CZB152

ZDB30
ZDB564
ZDB172

35,000

ZDB200

ZDB199

Mutator Evolves

+

Cit
Evolves
25,000

20,000

UC
15,000

10,000

5,000

0

Figure 3.1 │ Phylogeny of Ara–3 population. Symbols at branch tips indicate placement of
29 sequenced clones; identifying labels are shown for clones mentioned in main text and figures.
Shaded areas and colored symbols identify major clades. Fractions above the tree show the
number of clones identified as belonging to the clade that yielded Cit+ mutants during replay
experiments (numerator) and the corresponding total number of clones used in those experiments
(denominator).

132

Figure 3.2│ Total number of mutations relative to the ancestor. Number of mutations
identified in each sequenced genome relative to the ancestor. Colors correspond to clade of
+

origin (green = Clade 1, blue = Clade 2, orange = Clade 3, red = Cit ). Circles are non-mutator
genomes, while squares correspond to genomes with mutator genotypes. Lines are maximum
likelihood estimates of Poisson rates of mutation accumulation in non-mutator (solid) and
mutator (dashed) genomes from a model used to estimate branching times.

133

+

The sequenced Cit clones from generation 36,000 and later have a SNP in the mutS gene
that produces a premature stop codon and truncates the MutS protein, thereby causing a defect in
methyl-directed mismatch DNA repair (Glickman and Radman 1980). We sequenced mutS from
+

three Cit clones from each of generations 33,000, 34,000, 35,000, 36,000, 37,000 and 38,000,
and found this SNP in 0, 0, 2, 3 and 3 clones, respectively, indicating that the mutation arose
+

–

after the origin of the Cit lineage. None of the 20 Cit genomes have this mutation. As a
+

consequence, Cit genomes from later generations accumulated SNPs much faster than the Cit

–

+

and early Cit genomes (Figure 3.2). Mutator phenotypes evolved in several other populations
in the LTEE (Sniegowski et al 1997, Barrick et al 2009), so this change was not unique to the
+

Cit lineage.

+

EVOLUTION OF THE CIT FUNCTION
The evolution of the Cit

+

trait involved three successive processes: potentiation,

actualization and refinement. Potentiation refers to the evolution of a genetic background in
+

+

which the Cit function became accessible by mutation; the ancestor’s rate of mutation to Cit
was inaccessibly low, on the order of 10
extremely weak Cit

+

-13

per cell-generation or lower (Blount et al 2008). An

variant had emerged by 31,500 generations, which represents the

actualization step. Finally, the new function was refined, allowing the efficient exploitation of
+

citrate, the rise of the Cit subpopulation to numerical dominance, and the substantial expansion

134

of the population size.

In the following sections, we identify and examine the effects of

mutations involved in these processes. We begin with the second and third processes, before
turning to potentiation.

+

Actualization of the Cit function
One reason that E. coli cannot grow aerobically on citrate is its inability to transport
+

citrate (Hall 1982, Lara and Stokes 1952, Pos et al 1998). The origin of the Cit phenotype
+

therefore required expression of a citrate transporter. All 9 sequenced Cit genomes have
multiple (2 to 9) tandem copies of a 2933-bp segment that includes part of the citrate
fermentation (cit) operon (Figure 3.3a). The amplified segment contains two genes: rna encodes
rnase I (Zhu and Deutsher 1992), and citT encodes a broad-spectrum C4-di- and tri-carboxylic
acid transporter that functions in fermentation as a citrate-succinate antiporter (Pos et al 1998).
The boundary upstream of citT is in the 3' end of the citG gene, which encodes
triphosphoribosyl-dephospho-CoA synthase, while the boundary downstream of rna is in the 5'
end of rnk, which encodes a regulator of nucleoside diphosphate kinase (Shankar et al 1995).
This tandem amplification is not present in the ancestor or any of the 20 sequenced Cit

–

+

genomes, and it is only found in population samples after the evolution of the Cit lineage (Table
3.2). PCR screens also failed to detect this segment in any of 27 Cit– clones from generations
+

33,000 through 40,000, whereas it was found in all 33 Cit clones from the same generations
(Table 3.3).

135

a

b

Figure 3.3 │ Tandem amplification in Cit+ genomes. a. Ancestral arrangement of citG, citT, rna, and
rnk genes. b. Altered spatial and regulatory relationships generated by the tandem amplification.

136

Supplementary Table 3.2 │ Results of PCR screens on
whole-population samples for cit amplification
Generation
25,000

30,000

31,500

32,000

32,500

33,000

33,500

34,000

36,000

cit Amplification
Detected?
–
–
–
–
–
–
–
–
–
–
–
+
–
–
–
+
+
+
+
+
+
+
+
+
+
+
+

Replicate
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C

137

Table 3.3 │ Detection of cit amplification in Ara–3 clones
cit amplification Detected?
Generation
Clone
Phenotype
31,500
ZDB25
No
ZDB26
No
ZDB27
No
Yes

ZDB566

+
+
+

ZDB28
ZDB29
ZDB30

-

No
No
No

ZDB172

Yes

ZDB179

+
+
+

ZDB31
ZDB32
ZDB33

-

No
No
No

ZDB143

Yes

ZDB145

+
+
+

CZB199
CZB204
CZB205

-

No
No
No

CZB151

Yes

CZB154

+
+
+

ZDB83
ZDB84
ZDB85

-

No
No
No

ZDB86

+
+
+
+
+
+

Yes

ZDB564
ZDB565
32,000

ZDB173
32,500

ZDB144
33,000

CZB152
34,000

ZDB87
ZDB88
35,000

ZDB89
ZDB90
ZDB91

138

Yes
Yes

Yes
Yes

Yes
Yes

Yes
Yes

Yes
Yes
Yes
Yes
Yes

Table 3.3 │ Detection of cit amplification in Ara–3 clones
cit amplification Detected?
Generation
Clone
Phenotype
ZDB92
No
Yes

ZDB98

+
+
+
+
+
+

ZDB99
ZDB100

-

No
No

ZDB101

Yes

ZDB102

+
+

ZDB103
ZDB104
ZDB105
ZDB106

-

No
No
No
No

ZDB107

Yes

ZDB109

+
+
+

ZDB110
ZDB111

-

No
No

ZDB112

Yes

REL10981

+
+
+
+

REL10988
REL10989
REL10990

-

No
No
No

ZDB93
ZDB94
36,000

ZDB95
ZDB96
ZDB97

37,000

38,000

ZDB108

40,000

REL10979
REL10980

139

Yes
Yes
Yes
Yes
Yes

Yes

Yes
Yes

Yes
Yes
Yes

Amplification mutations can alter the spatial relationship between structural genes and
regulatory elements, potentially causing altered regulation and novel traits (Usakin et al 2005,
Adam et al 1993, Bock and Timmis 2008, Whoriskey et al 1987). The structure of the cit
+

amplification suggests that the Cit trait arose from an amplification-mediated promoter capture
(Figure 3.3b, Figure 3.4).

The amplification joined upstream rnk and downstream citG

fragments, producing an rnk-citG hybrid gene expressed from the upstream rnk promoter.
Because the citT and citG genes are normally monocistronic, the downstream copy of citT should
therefore be co-transcribed with the hybrid gene. If the rnk promoter directs transcription under
oxic conditions, then the new rnk-citT regulatory module might allow CitT expression during
+

aerobic metabolism and thereby confer a Cit phenotype (Pos et al 1998).
To test this hypothesis, we first examined the capacity of the rnk-citT module to support
citT expression under oxic conditions. We engineered a low-copy (1-2 per cell), lux-based
reporter plasmid, Plux::rnk-citT, that has an rnk-citT module in which citT was replaced by the
luxCDABE reporter operon. We constructed two additional plasmids, Plux::rnk and Plux::citT, in
which the reporter was under the control of the native upstream regulatory regions of rnk and
citT, respectively. We transformed each plasmid into REL606, the ancestral strain; ZDB30, a
+

potentiated C3 clone from generation 32,000; and ZDB172, a weakly Cit clone from generation
32,000. We measured Lux expression during growth and stationary phase under oxic conditions
(Figure 3.5). The native citT regulatory region showed no Lux expression (above background) in
any strain, indicating that citT is normally silent under oxic conditions. The native rnk regulatory
region expressed Lux, with a peak around the transition into stationary phase. Lux expression

140

Figure 3.4 │ Annotation of sequence adjacent to the boundary of the cit amplification
+
found in Cit genomes.

141

ZDB30

REL606

ZDB172

Expression (cps)

1000

100

10
0

4

8 12 16 20 0

4

8 12 16 20 0

4

8 12 16 20

Figure 3.5 │ Expression levels from native citT, native rnk, and evolved rnk-citT regulatory
regions during aerobic metabolism. Average time course of luciferase expression for the
ancestral strain REL606 (left) and evolved clones ZDB30 (center) and ZDB172 (right), each
transformed with reporter plasmids Plux::citT (blue), Plux::rnk (black) and Plux::rnk-citT (red).
Each curve shows the average of four replicates.

142

from the evolved rnk-citT module was much weaker, but there were small spikes in expression in
both ZDB30 and ZDB172 coincident with peak expression from the native rnk regulatory region.
These results confirm that the rnk-citT module can support citT expression during aerobic
metabolism.
+

We also tested whether the rnk-citT module could confer a Cit phenotype by inserting a
single copy of an amplification fragment containing the rnk promoter immediately upstream of
the chromosomal copy of citT (Figure 3.6a) in potentiated clone ZDB30.

This construct,

+

+

ZDB595, has a Cit phenotype, but it is extremely weak, similar to the earliest evolved Cit
variants (Figure 3.6b).

In the same medium used in the evolution experiment, ZDB595

experienced a lag of >60 hours after glucose depletion, followed by a short period of abortive
+

and inconsistent growth on citrate (Figure 3.6b,c). These data imply that the initial Cit type was
too weak to permit exploitation of citrate under the daily-transfer regime of the long-term
experiment. Nonetheless, ZDB595 had a small (1.0%) but significant fitness advantage over
ZDB30 in competition assays performed in the same environment (n = 10, t = 3.09, two-tailed P
= 0.0128).

+

Refinement of the Cit function
+

+

Given the extremely weak initial Cit phenotype, the rise of Cit cells to high frequency required
further mutations to refine the new function. These mutations are evidenced by the substantial
improvement on citrate of variants between 31,500 and 33,000 generations (Figure 3.7), shortly
before the population expansion. To understand this refinement, we examined the genomes of

a

c

b
-1

-2

-3

ln OD420

ln OD420

-2
-4
-5

-3
-4

-6
-5

-7
0

24

48 72 96 120 144 168
Incubation (h)

0

24

48 72 96 120 144 168

Incubation (h)

Figure 3.6 │ New rnk-citT module confers Cit+ phenotype in potentiated background. a.
Engineered construct containing the cit amplification junction with the rnk promoter, and the
gene arrangement following insertion of the construct into the chromosome. b. Average growth
trajectories in DM25 of potentiated clone ZDB30 (red), its isogenic construct with chromosomal
integration of the rnk-cit module ZDB595 (blue), 31,500-generation Cit+ clone ZDB564 (purple),
and 32,000-generation Cit+ clone ZDB172 (green). c. Growth trajectories of ZDB595 compared
to its parent. Light blue trajectories show heterogeneity among replicate cultures of ZDB595;
dark blue and red are averages of trajectories for ZDB595 and its parent, ZDB30, as in panel b
(except on different scale).

144

0
-1

lnOD420

-2
-3
-4
CZB152

-6

CZB154

ZDB143

-5

ZDB172

ZDB564

-7
0

24

48

72

96

Incubation (h)

+

Figure 3.7 │ Growth of early Cit clones in DM25.
Early Cit+ clones from population Ara–3 improved greatly in their capacity to grow on the citrate
in DM25 over time.

Cit+ clones from generations 31,500 (ZDB564), 32,000 (ZDB172), and

32,500 (ZDB143) showed little growth on citrate even after 96 h, while two clones from
generation 33,000 (CZB152 and CZB154) achieved substantial growth on citrate within 24 h.
This improvement in citrate utilization allowed Cit+ clones to rise from a minority of the
population to numerical dominance.

145

+

five Cit clones from before the population expansion, including one each from generations
31,500, 32,000 and 32,500 and two from generation 33,000. We focused on those mutations on
+

the line of descent for the Cit subpopulation, where that line is defined by the presence of the
+

same mutations in Cit clones from generations 34,000, 36,000 and 38,000 (Tables 3.4 – 7).
Among those mutations, two SNPs and one IS-element insertion were present in only one of the
33,000-generation genomes. The IS insertion and one SNP appear unrelated to growth on
citrate. The remaining SNP is in the regulatory region of dctA, which encodes a transporter of
succinate and other C4-di-carboxylic acids (Janausch et al 2002). This mutation may improve
recovery of succinate exported in exchange for citrate. However, the other 33,000-generation
clone showed stronger growth on citrate (Figure 3.7). Thus, the dctA mutation does not appear
to be responsible for the population expansion.
+

These early Cit genomes also underwent a progressive increase in cit copy number. The
earliest one had a tandem duplication, whereas later genomes had a tandem duplication within a
larger tandem duplication, a four-copy tandem array, a three-copy tandem array within a larger
duplication, and a nine-copy tandem array, respectively. Changes in amplification copy number
readily occur by recombination, and have been implicated in the refinement of other weak
functions (Andersson et al 1998, Reams et al 2010). These early changes all increased the
number of rnk-citT modules (Figure 3.8a, Table 3.8) and should therefore increase expression of
CitT. To test whether this increase caused the population expansion, we cloned the rnk-citT
module (Figure 3.8b) into the high-copy plasmid pUC19 (Yanisch-Perron et al 1985) and moved
it into the 32,000-generation potentiated clone ZDB30. The resulting strain, ZDB612, is strongly
+

Cit , rapidly transitions from glucose to citrate, and grows similarly to the 33,000-generation

146

+

CZB154 (33,000)

CZB152 (33,000)

ZDB143 (32,500)

ZDB172 (32,000)

ZDB564 (31,500)

Clone
(generation)

Product

Gene

Amino-acid
Change

Type

Nucleotide
Change

Genome
Position

Table 3.4 │ Point mutations in early Cit genomes

Acyl-CoA
thioesterase II /
predicted outer
membrane
lipoprotein
Predicted
peptidase /
aminobenoylglutamate
utilization
protein

447290 T  C

NC

N/A

tesB/
ybaY

1399744 T  C

NS

I228V

abgB

2241625 C  A

NS

A259S

ccmH

Heme lyase
subunit

2443160 C  A

S

N/A

glk

Glucokinase

dctA/
yhjK

C4dicarboxylate
transporter /
Predicted
diguanylate
cyclase

3612959 C T

NC

N/A

Mutations shown are those that were not uniformly found in all five Cit+ clones under study. Red
fill indicates presence of mutation. Crosshatching further indicates mutation was also present in
Cit+ clones from generations 34,000, 36,000, and 38,000. Gray fill indicates absence of mutation.

147

+

CZB154 (33,000)

CZB152 (33,000)

ZDB143 (32,500)

ZDB172 (32,000)

ZDB564 (31,500)

Clone
(generation)

Product

Gene

Element

Genome
Position

Table 3.5 │ IS-element insertions in early Cit genomes

Outer membrane
transporter/Ferric
enterobactin esterase
Periplasmic disulfide
isomerase/Thioldisulphide oxidase
Fused predicted
transporter subunits of
ABC superfamily: ATPbinding components

595335

IS150

fepA/fes

620126

IS150

dsbG

1028311

IS150

uup

2322345

IS186

menC

O-succinylbenzoate

2877315

IS150

kduD

2-deoxy-D-fluconate/3dehydrogenase

4252526

IS150

uvrA

Excinuclease ABC
subunit A

+

Mutations shown are those that were not uniformly found in all five Cit clones under study. Red
fill indicates presence of mutation. Crosshatching further indicates mutation was also present in
Cit

+

clones from generations 34,000, 36,000, and 38,000. Gray fill indicates absence of

mutation.

148

CZB154 (33,000)

CZB152 (33,000)

ZDB143 (32,500)

ZDB172 (32,000)

ZDB564 (31,500)

Clone
(generation)

Genes Deleted

Size (bp)

Genome
End

Genome
Start

Table 3.6 │ Deletions in early Cit+ genomes

hokE, insL-3,
590472

599560

9088

entD, fepA, fes,
ybdZ

590472

595335

4863

hokE, insL-3,
entD, fepA

1345210 1345211

1

yciR/rnb

3786737 3786738

1

Noncoding

+

Mutations shown are those that were not uniformly found in all five Cit clones under study. Red
fill indicates presence of mutation. Crosshatching further indicates mutation was also present in
Cit

+

clones from generations 34,000, 36,000, and 38,000. Gray fill indicates absence of

mutation.

149

599561

619885

1729052

666130

634746

1995783

3

(x2)

dnaK (internal fragment)

66569
(x2)

[entF], fepE, fepC, fepG
, fepD, ybdA,
fepB, entC, entE, entB,
entA, ybdB, cstA,cstA, y
bdD, ybdH, ybdL, ybdM,
ybdN, insB-8, insA-8,
dsbG, ahpC, ahpF,
ybdQ,ybdR, rnk, rna, cit
T, citG, citX, citF, citE,ci
tD, citC, insB-9, insA-9,
citA, citB, dcuC, crcA,
cspE, ccrB, ybeM,
tatE,
lipA,ybeF, lipB, ybeD, d
acA, rlpA, mrdB,mrdA, y
beA, ybeB, phpB, nadD,
holA,
rlpB, leuS, ybeL, ybeQ,
ybeQ, ybeR,
ybeV, hscC, rihA, gltL, i
nsJ-2, insK-2, [gltL]

14861
(x3)

[dsbG], ahpC, ahpF, yb
dQ, ybdR, rnk,
rna, citT, citG, citX, citF,
citE, citD, citC, insB9, insA-9

266731
(x3)

[ydhV], ydhY, ydhZ, pyk
F, lpp, ynhG,
sufE, sufS, sufD, sufC,
sufB, sufA, ydiH,
ydiI, ydiJ, ydiK, ydiL, ydi

150

CZB154 (33,000)

CZB152 (33,000)

ZDB143 (32,500)

ZDB172 (32,000)

ZDB564 (31,500)

Clone
(Generation)

12455

Genes
Duplicated

Genome End

12452

Size (bp)

Genome Start

Table 3.7 │ Amplifications in early Cit+ genomes

M, ydiM, ydiN,
ydiB, aroD, ydiF, ydiO, y
diP, ydiQ, ydiR,
ydiS, ydiT, ydiD, ppsA,
ydiA, aroH, ydiE,
ydiU, ydiV, nlpC, btuD,
btuE, btuC, himA,
pheT, pheS, pheM, rplT,
rpmI, infC, thrS,arpB, a
rpB, ECB_01690, ydiY,
pfkB, ydiZ,yniA, yniB, yn
iC, ydjM, ydjN, ydjO, ce
dA,katE, ydjC, celF, cel
D, celC, celB, celA,
osmE, nadE, ydjQ, ydjR
, spy, astE, astB,astD, a
stA, astC, xthA, ydjX, yd
jY, ydjY,
ydjZ, ynjA, ynjB, ynjC, y
njD, ynjE, ynjF,
nudG, ynjH, gdhA, ynjI,
topB, selD, ydjA,sppA, a
nsA, pncA, ydjE, ydjF, y
djG, ydjH,ydjI, ydjJ, ydjK
, ydjL, yeaC, yeaA,
gapA, yeaD, yeaE,
mipA, yeaG, yeaH,
yeaI, yeaJ, yeaK,
ECB_01757, yeaL,
yeaM, yeaN, yeaO,
yoaF, yeaP, yeaQ, yoa
G, yeaR, insL-4,
yeaS, yeaT, yeaU,
yeaV, yeaW, yeaX,
rnd, fadD, yeaY,
yeaZ, yoaA, yoaB,
yoaC, yoaH, pabB,

151

CZB154 (33,000)

CZB152 (33,000)

ZDB143 (32,500)

ZDB172 (32,000)

ZDB564 (31,500)

Clone
(Generation)

Genes
Duplicated

Size (bp)

Genome End

Genome Start

Table 3.7 │ Amplifications in early Cit+ genomes

yeaB,sdaA, yoaD,
yoaE, manX, manY,
manZ,yobD, yebN,
rrmA, cspC, yobF,
yebO,yobG, ECB_0179
7, kdgR, yebQ,
htpX,prc, proQ, yebR, y
ebS, yebT, yebU,yebV,
yebW, pphA, yebY,
yebZ, yobA, holE,
yobB, exoX, ptrB, yebE,
yebF, yebG, purT,
eda, edd, zwf, yebK,
pykA, msbB, yebA,
znuA, znuC, znuB,
ruvB ,ruvA, yebB, ruvC,
yebC, ntpA, aspS,
yecD, yecE, yecN,
yecO, yecP, torZ,torY,
cutC, yecM, argS,
yecT, flhE, flhA,flhB,
cheZ ,cheY, cheB,
cheR, tap, tar, cheW,
cheA, motB, motA, flhC,
flhD,yecG, otsA, otsB,
araH, araG, araF, yecI,
yecJ, yecR, ftn, yecH,
tyrP, yecA, leuZ,
cysT, glyW, pgsA, uvrC,
yvrY, insA-13,
insB-13, yedU, yedV,
yedW, yedX, yedY,
yedZ, yodA, yodB,
serU, yeeI, asnT,
yeeJ, yeeL, yeeL,
shiA, amn, yeeN

152

CZB154 (33,000)

CZB152 (33,000)

ZDB143 (32,500)

ZDB172 (32,000)

ZDB564 (31,500)

Clone
(Generation)

Genes
Duplicated

Size (bp)

Genome End

Genome Start

Table 3.7 │ Amplifications in early Cit+ genomes

3268052

2086894

3268052
(x2)

rpmA, rplU, ispB, nlp,
murA, yrbA, yrbB,
yrbC, yrbD, yrbE, yrbF,
yrbG, yrbH, yrbI,yrbK, y
hbN, yhbG, rpoN, yhbH,
ptsN, yhbJ, ptsO, yrbL,
mtgA, yhbL, arcB,yhcC
gltB, gltD, yhcG,
ECB_03080,yhcH,
nanK, nanE, nanT,
nanA, nanR,dcuD,
sspB, sspA, rpsI, rplM,
yhcM,yhcB, degQ
degS, mdh, argR, yhcN
,yhcO, yhcP, yhcQ,
yhcR, yhcS, tldD,
yhdP, rng, maf,
mreD, mreC, mreB,
yhdA, yhdH, accB,
accC, yhdT, panF,
prmA, yhdG, fis, yhdJ,
yhdU, envR, acrE,acrF,
yhdV, yhdW, yhdX,
yhdY, yhdZ,rrfF, thrV,
rrfD, rrlD, alaU, ileU,
rrsD,yrdA, yrdB, aroE,
yrdC, yrdD, smg, smf,
def, fmt, rrmB, trkA,
mscL, yhdL, zntR,yhdN,
rplQ, rpoA, rpsD,
rpsK, rpsM,rpmJ, prlA,
rplO, rpmD, rpsE,
rplR, rplF,rpsH, rpsN, r
plE, rplX, rplN, rpsQ,
rpmC,rplP, rpsC, rplV,
rpsS, rplB, rplW, rplD,
rplC, rpsJ, pioO, gspA,
153

CZB154 (33,000)

CZB152 (33,000)

ZDB143 (32,500)

ZDB172 (32,000)

ZDB564 (31,500)

Clone
(Generation)

Genes
Duplicated

Size (bp)

Genome End

Genome Start

Table 3.7 │ Amplifications in early Cit+ genomes

gspC, gspD,gspE,
gspF, gspG, gspH, gspI,
gspJ,gspK, gspL,
gspM, gspO, bfr, bfd,
chiA, tuf, fusA, rpsG,
rpsL, yheL, yheM,
yheN,yheO, fkpA, slyX,
slyD, kefB, yheR,yheS,
yheT, yheU, prkB,
yhfA, crp, yhfK,argD,
pabA, fic, yhfG, ppiA,
yhfC, nirB,nirD, nirC,
cysG, yhfL, yhfM, yhfN,
frlC,yhfQ, yhfR, yhfS,
insB-23, insA-23, yhfS,
yhfT, yhfU, yhfV, yhfW,
yhfX, yhfY, yhfZ,trpS,
gph, rpe, dam, damX,
aroB, aroK,hofQ, yrfA,
yrfB, yrfC, yrfD, mrcA,
yrfE,yrfF, yrfG, hslR,
hslO, yhgE, pckA, envZ,
ompR, greB, yhgF,
feoA, feoB, yhgG,yhgA,
bioH, yhgH, yhgI, gntT,
malQ,malP, malT,
rtcA, rtcB, rtcR, glpR,
glpG,glpE, glpD, yzgL,
ECB_03279, yzgL,glgP,
glgA, glgC, glgX,glgB,
asd, yhgN, gntU, gntK,
gntR, yhhW, yhhX,
yhhY,yhhZ, yrhA, yrhB,
ggt, yhhA, ugpQ, ugpC,
ugpE, ugpA, ugp, livF, li
vG, livM, livH,livK,
yhhK, livJ, rpoH, ftsX,

154

CZB154 (33,000)

CZB152 (33,000)

ZDB143 (32,500)

ZDB172 (32,000)

ZDB564 (31,500)

Clone
(Generation)

Genes
Duplicated

Size (bp)

Genome End

Genome Start

Table 3.7 │ Amplifications in early Cit+ genomes

CZB154 (33,000)

CZB152 (33,000)

ZDB143 (32,500)

ZDB172 (32,000)

ZDB564 (31,500)

Clone
(Generation)

Genes
Duplicated

Size (bp)

Genome End

Genome Start

Table 3.7 │ Amplifications in early Cit+ genomes

ftsE, ftsY,yhhF,
yhhL, yhhM, yhhN,
zntA, sirA,yhhQ,
dcrB, yhhS, yhhT, acpT,
nikA,nikB, nikC, nikD,
nikE, nikR, rhsB, yhhH,
yrhC, yhhI, yhhJ, yhiH,
yhiI, yhiJ, yhiKL,yhiM,
yhiN, pitA, yhiO, uspA,
yhiP, yhiQ,prlC, yhiR,
gor, arsR, arsB, arsC,
yhiS,slp, yhiF, yhiD,
hdeB, hdeA, hdeD,
yhiE,yhiU, yhiV, yhiW,
gadX, gadA, yhjA, treF,
yhjB, yhjC, yhjD, yhjE,
yhjG, yhjH, kdgK,yhjJ,
dctA, yhjK, bcsC, bcsZ,
yhjN, yhjO,yhjQ, yhjR,
yhjS, yhjT, yhjU, ldrD,
ldrD,ldrD, yhjV, dppF,
dppD, dppC, dppB,
dppA, proK, yhjW,
yhjX, yhjY, tag, yiaC,
bisC, yiaD, tkrA, yiaF,
yiaG, cspA, hokA,
insJ-4, insK-4
Mutations shown are those that were not uniformly found in all five Cit+ clones under study. Red
fill indicates presence of mutation. Crosshatching further indicates mutation was also present in
Cit+ clones from generations 34,000, 36,000, and 38,000. Gray fill indicates absence of mutation.

155

rnk-citT modules

a

b

8

BamHI

6

SalI

4
citT

2
0

rnk promoter

31 32 33 34 35 36 37 38 39 40
Generation (thousands)

ln OD420

c

0
-1
-2
-3
-4
-5
-6
-7
0

24

48

72

96

Incubation (h)
(

120
)

144

168

Figure 3.8 │ Refinement of Cit+ phenotype by increased copy number of rnk-citT module.
a. Change in rnk-citT module copy number in sequenced Cit+ clones over time. b. Structure of
rnk-citT module cloned into the high-copy plasmid pUC19. c. Growth trajectories in DM25 of
potentiated Cit– clone ZDB30 (red), 31,500-generation Cit+ clone ZDB564 (purple), 33,000generation Cit+ clone CZB152 (black), and ZDB612, a pUC19::rnk-citT transformant of ZDB30
(green). Trajectories are averages of six replicates for each strain.

156

Table 3.8│ Estimated citT module copy number

Gen
(k)

Clone

Anc citT

rnk-citT

Relative
rnk-citT

citT
module

Reference
Regions

Relative
citT

Predicted
Configuration

Coverage

Anc rnk

Junction Number

31.5

ZDB564

52

49

42

0.83

120.8

64.9

1.86

2×

32

ZDB172

45

39

82

1.95

139.3

23.8

5.86

2× (3×)

32.5

ZDB143

116

76

301

3.14

457.2

112.8

4.05

4×

CZB152

22

23

192

8.53

267.4

32.7

8.18

9×

CZB154

398 252

345

1.06

797.6

130.0

6.14

3× (2×)

34

ZDB83

57

43

125

2.50

214.7

62.9

3.42

4×

36

ZDB96

26

22

87

3.63

97.4

28.8

3.39

4×

38

ZDB107

51

32

133

3.20

204.7

60.0

3.41

4×

40

REL10979

36

32

73

2.15

115.6

32.5

3.56

4×

33

+

For each Cit genome, the number of new junctions per genome was estimated from the relative
number of reads supporting the new rnk-citT junction produced by the amplification versus the
number of reads supporting the ancestral rnk and citT junctions. The total number of citT
modules per genome was estimated by comparing read-depth coverage of the amplified citT
region to coverage of regions that appear to be single copy in all genomes (comprising ~20 kb
total including the ara operon and tufB gene). Together these data can be used to predict the
likely configuration of citT amplification copies in each genome. Examination of read-depth
coverage over a larger area supports the observation that there are nested amplifications in
CZB154 and ZDB172. For example, the CZB154 genome contains three copies of a larger
region, and each copy of that region contains two tandem copies with the usual rnk-citT junction.

157

clone CZB152 (Figure 3.8c). The increased number of rnk-citT modules thus appears to explain
+

the initial refinement of the Cit phenotype that led to the population expansion. In contrast to
+

the early variation in cit amplification, later Cit genomes have four-copy tandem arrays (Figure
3.8). Amplification mutations tend to be unstable (Andersson et al 1998, Reams et al 2010), and
further refinement of the Cit

+

phenotype may have favored more stable mutations.

The

+

evolution of the mutator phenotype in the Cit lineage complicates efforts to identify these later
mutations, but interesting candidates include SNPs in citT itself; gltA, which encodes citrate
synthase; and aceA, which encodes isocitrate lyase.

+

Potentiation of Cit evolution
+

Before the Cit phenotype appeared, the Ara–3 population had to evolve a background in
which the requisite new function was accessible by mutation. The ancestor’s mutation rate to
Cit

+

-13

was immeasurably low, with an upper bound of 3.6 x 10

per cell-generation.

Potentiation was demonstrated by ‘replay’ experiments using 270 clones sampled throughout this
+

population’s history (Blount et al 2008). The replays produced 17 additional Cit mutants that
derived from 13 clones, all from generation 20,000 or later. Fluctuation tests confirmed that
+

potentiated clones had increased mutation rates to Cit , although such mutations were still
extremely rare (Blount et al 2008).
The potentiating mutations are not known to confer any other phenotype, hence there is
no simple way to distinguish between potentiated and non-potentiated clones. Therefore, we first

158

ybaL
nadR
hemE
cspC

yaaH

leuA
tolR

arcB

gltA

Predicted transporter
with NAD(P)-binding
Rossmann-fold domain
Nicotinamide-nucleotide
adenyltransferase
Uroporphyrinogen
decarboxylase
Stress protein, member
of the CspA-family
Conserved inner
membrane protein
associated with acetate
transport
2-isopropylmalate
synthase
Membrane spanning
protein in TolA-TolQTolR complex
Hybrid sensory histidine
kinase in twocomponent regulatory
system with ArcA.
Aerobic respiration
control sensor
Citrate synthase

Associated
Clades

Evolved
Nucleotide

Ancestral
Nucleotide

Gene Product

Gene
Position

Gene

Genome
Position

Table 3.9 │ Phylogenetically informative mutations

475173

133

G

C

UC

4616538

1010

A

C

C1,C2,
C3

4177963

636

T

G

C1

1886011

4

C

A

C1

9972

521

T

G

C2,C3

85556

778

C

T

C2

756799

69

C

T

C2

3288053

208

G

C

C2

3288026

236

T

A

C3

734488

772

G

A

C3

159

examined the distribution of the 13 potentiated clones identified by the replay experiments using
ten mutations (Tables 3.9, 10) that differentiated clades UC, C1, C2 and C3 (Figure 3.1). We
also determined the distribution of the other 256 evolved clones used in the replays to assess the
coverage of the clades in those experiments (Figure 3.9, 10). Overall, 205 clones were assigned
to clades, including 12 potentiated clones (Appendix Table A.1).

Sixteen clones from

generations 15,000 and earlier were in clade UC. The others came from generations 20,000 and
later including 55 in C1, 97 in C2, and 37 in C3. Potentiated clones occurred in all three clades
with 8 in C3 and 2 each in C1 and C2. Nonetheless, this distribution is highly non-random (twotailed Fisher’s Exact Test comparing C3 with C1 and C2 combined, P < 0.0003; that test finds no
+

difference between C1 and C2, P = 0.6206), with most Cit mutants from the C3 clade that
+

produced the original Cit mutant. These data suggest that potentiation involved at least two
mutations, with one before these clades diverged and another in C3 (Figure 3.1).
Two distinct mechanisms might explain the potentiation effect. One is epistasis, whereby
an interaction between the potentiating background and the actualizing mutation was necessary
+

to produce the Cit phenotype. The second is that the background physically promoted the final
mutation; for example, a later rearrangement might require some prior genomic rearrangement.
+

If expression of Cit required earlier mutations, then the rnk-citT module should confer a weaker
+

Cit phenotype when moved to a non-potentiated background than to a potentiated background.
Alternatively, if potentiation facilitated the amplification event itself, then that module should
+

produce an equally strong Cit phenotype in both potentiated and non-potentiated backgrounds.

160

Clone
N

ybaL
G133C

Unsuccessful
Clade

C1, C2, C3 or
Unknown
nadR
A1010C

N

Y

Y

C1, C2 or C3

Unknown
N

yaaH
T521G

(C1, C2 or C3)

Y
C2 or C3

Figure 3.9 │ Key for placement of clones used in replay experiments that were sampled
before generation 20,000.
The phylogenetic placement of each clone isolated from the population prior to generation
20,000 for use in the replay experiments was determined by scoring the presence or absence of
informative mutations according to the key above.

161

b.
Clone
N

Y

yaaH
T521G

C1 or Unknown
N

hemE
T636G

C2 or C3
N

Y

gltA
C772T

Y

C1

C2

C3

Check for
cspC to
confirm

Check for leuA
C778T,
tolR C69T or arcB
G208C to confirm

Check for arcB
G236A to
confirm

Unknown

Figure 3.10 │ Key for placement of clones used in replay experiments that were sampled
from generations 20,000 or later.
The phylogenetic placement of each clone isolated from the population at generations 20,000 or
higher for use in the replay experiments was determined by scoring the presence or absence of
informative mutations according to the key above. Note the increased resolution that results
from greater divergence.

162

To test these predictions, we moved pUC19::rnk-citT into the ancestor and clones from
clades C1, C2 and C3, and examined their growth trajectories (Figure 3.11).

All four

transformants grow on citrate after depleting the available glucose. However, the transformants
of the ancestor and clones from C1 and C2 grow poorly on citrate, even with this high-copynumber plasmid, as evidenced by long lags while transitioning from glucose to citrate, low
yields, and inconsistent trajectories across replicates (Figure 3.10a-c). By contrast, ZDB612, a
transformant of the potentiated C3 clone ZDB30, grows much faster, more extensively, and
consistently across replicates (Figure 3.10d). These differences indicate epistatic interactions
between the rnk-citT module and mutations that distinguish the backgrounds. These data also
support the phylogenetic association between clade C3 and the strength of potentiation observed
in the replay experiments. We examined C3 for candidate mutations that may contribute to its
potentiation. A mutation in arcB, which encodes a histidine kinase (Gunsalus and Park 1994), is
noteworthy because disabling that gene up-regulates the TCA cycle (Niam et al 2009). That
mutation might interact with the rnk-citT module by allowing efficient use of citrate that enters
the cell via the CitT transporter.
+

Given the extremely weak Cit phenotype when the rnk-citT module was moved onto the
chromosome of the same potentiated C3 clone (Figure 3.6), any mutation that produced a single
copy of that module in the ancestor would presumably give either no growth on citrate or such
weak growth that no mutants would have been detected in the replay experiments. Indeed, we
were unsuccessful in moving the rnk-citT module to the ancestral chromosome. One step of that
+

failed process was to screen potential recombinants for a Cit phenotype; that approach was
successful, albeit difficult, in the potentiated C3 clone. Thus, the relative strength of the Cit

163

+

b

0
-1
-2
-3
-4
-5
-6
-7

ln OD420

ln OD420

a

0

0

96

d

0
-1
-2
-3
-4
-5
-6
-7

ln OD420

ln OD420

c

24
48
72
Incubation (h)

0

24
48
72
Incubation (h)

0
-1
-2
-3
-4
-5
-6
-7

96

24
48
72
Incubation (h)

96

0

24
48
72
I Incubation (h) )
b i
(h

96

0
-1
-2
-3
-4
-5
-6
-7

Figure 3.11 │ Evidence of epistatic interactions in potentiation of Cit+ phenotype. Growth
trajectories in DM25 of phylogenetically diverse clones transformed with the pUC19::rnk-citT
plasmid. a. Ancestral strain REL606 and its transformant ZDB611. b. C1 clone ZDB199 and
its transformant ZDB614. c. C2 clone ZDB200 and its transformant ZDB615. d. C3 clone
ZDB30 and its transformant ZDB612. In each panel, red and dark blue trajectories are averages
for the parent clone and its Cit+ transformant; light blue trajectories show the replicates for the
transformant.

164

phenotype produced by the high-copy plasmid in the various backgrounds (Figure 3.11) and the
+

relative success in generating any Cit phenotype with a chromosomal copy of the rnk-citT
module both support the hypothesis that potentiation depends, at least partly, on epistatic
interactions between the genetic background and the actualizing mutation that first generated that
module.
+

Further evidence regarding potentiation comes from the Cit mutants observed during the
replay experiments (Blount et al 2008). The physical-promotion hypothesis predicts that these
mutants would all have cit amplifications similar or identical to the original one. If epistatic
potentiation enhanced citT expression only from the rnk promoter, then the prediction would be
the same. However, if epistatic potentiation operated at some broader physiological level, then
+

the Cit replay mutants should have a variety of mutations that share only the property that they
enable expression of the citrate transporter in the oxic environment of the long-term experiment.
We examined 19 re-evolved Cit

+

mutants to identify the relevant mutations.

The region

containing citT was examined in all of them, and the genomes of six were sequenced. All 19
have mutations affecting citT, and most clearly put that gene downstream of a new promoter
+

–

(Table 3.10). Four fully sequenced Cit mutants were derived from Cit clones that were also
sequenced. Besides citT-related mutations, these sequenced mutants had 1-3 other mutations; no
gene was mutated in multiple cases, and none appear related to citrate utilization (Table 3.11),
+

supporting the inference that the citT mutations were responsible for all re-evolved Cit
phenotypes.

165

+

Table 3.10 │ Mutations affecting cit region in Cit replay mutants
Cit–
Cit+
Gen
Clade
Mutation affecting cit Region
Parent
Mutant
2978-bp tandem cit duplication that
ZDB285*
creates rnk-citT regulatory module.
20,000 ZDB464*
C3
2656-bp tandem cit duplication that
ZDB286
creates rnk-citT regulatory module.
2070-bp tandem cit duplication that
ZDB309
C3
ZDB288
creates rna-citT regulatory module.
27,000
~5,000-bp tandem duplication.
ZDB310
C3
ZDB290*
Basis of citT activation unknown.
30,500
ZDB20
C3
ZDB547 IS3 insertion in citG.
31,000

2706-bp tandem cit duplication that
creates rnk-citT regulatory module.
IS3 insertion in citG

ZDB390

C1

ZDB292

ZDB25

C3

ZDBr218

ZDB199

C1

ZDB279

ZDB28

C3

ZDB161

2745-bp tandem cit duplication that
creates rnk-citT regulatory module.
IS3 insertion in citG.

ZDB29

C2

ZDB166

IS3 insertion in citG.

ZDB164

IS3 insertion in citG.

ZDB165*

IS3 insertion in citG.

31,500

32,000

ZDB30*

C3

ZDB283*
ZDB294*

ZDB183

C3

ZDB281*
ZDB163

ZDB31

C3

ZDB546

32,500
ZDB549
ZDB32

C2

ZDB548

*Genome has been sequenced.

166

~568-kbp inversion that places
much of the cit operon under
control of the fimB promoter.
2663-bp tandem cit duplication that
creates rnk-citT regulatory module.
~14.3-kbp duplication. Basis of
citT activation unresolved.
2990-bp tandem cit duplication that
creates rnk-citT regulatory module.
Unknown rearrangement or
duplication affecting citT.
422-bp deletion in citG. Basis of
citT activation unresolved.
3144-bp tandem cit duplication that
creates rnk-citT regulatory module.

Description

848202

IS150
Insertion

Tandem
duplication
with one
junction in citG
and the other
between rna
and rnk;
presumed Cit+
actualizing
mutation
Insertion in a
coding region

1360374

IS150
Insertion

Insertion in a
coding region

ycjC

IS150
Insertion

Insertion in a
non–coding,
intergenic
region;
promoters not
disrupted

Site
between
nlp/murA

626102

3270443

2794 bp
duplication

citG

citT
rna
ybiS

167

Product(s)
Triphosphoribosul–
dephospho–
CoA
transferase
Citrate
transporter
Ribonuclease
I
Hypothetical
protein
DNA–binding
transcriptional
repressor
DNA–binding
transcriptional
activator of
maltose
metabolism/
UDP–N–
acetylglucosa
mine 1–
carboxyvinyltransferase

ZDB285

Mutation

ZDB464

Position

Gene or
Genes
Involved

Supplementary Table 3.11a │Pair 1: ZDB464 (Generation
+
20,000) and Cit Mutant ZDB285

Cit–
Parent
+
Cit
Mutant

Table 3.11 │ Annotated differences between genomes of four Cit+ mutants and
their Cit– parent clones

628716

Description

IS3 Insertion

Insertion in a
coding region;
presumed Cit+
actualizing
mutation

citG

ydgG
1651966

∆795 bp

Deletion of
multiple
coding regions

168

pntB

Product(s)
Triphosphoribosul–
dephospho–
CoA
transferase
Transporter
of quorum
signal AI–2
Pyridine
nucleotide
transhydrogenase,
β subunit

–

ZDB165

Mutation

ZDB30

Gene or
Genes
Involved

Position

Cit
Parent
+
Cit
Mutant

+

Table 3.11b │Pair 2: ZDB30 (Generation 32,000) and Cit Mutant
ZDB165

632864

687047

Description

568476 bp
inversion

Most of cit
operon
structural
genes placed
downstream of
fimB promoter;
new junctions
in citC and
between yjhA
and fimB;
+
presumed Cit
actualizing
mutation

IS150
Insertion

Gene
Product(s)

citC

fimB

169

Regulator of
fimA pili
subunit

nagE

Insertion in a
coding region

Citrate lyase
synthetase

Fused N–
acetyl
glucosamine
specific PTS
enzyme: IIC,
IIB, and IIA
components

–

Cit
Parent
+
Cit
Mutant
ZDB293

Mutation

Gene or
Genes
Involved

Position

ZDB30

+

Table 3.11c │Pair 3: ZDB30 (Generation 32,000) and Cit
ZDB293

626107

2661 bp
duplication

1607917

∆1 bp

1729055

∆2462 bp

2157881

∆12781 bp

Description

Gene
Product(s)

Tandem
Triphosphoduplication with
ribosul–
citG
junctions in
dephospho–
citG and
CoA transferase
between rna
Citrate
citT
and rnk;
transporter
presumed Cit+
actualizing
rna
Ribonuclease I
mutation
Single
Putative tail
nucleotide
ECB_1510 component of
deletion in a
prophage
coding region
Predicted oxidoIS150–
ydhV
reductase
mediated
deletion of
Predicted 4Fe–
multiple coding
ydhY
4S ferredoxin–
regions
type protein
Conserved
yehM
protein
Conserved
yehP
protein
Possible
yehQ
pseudogene
IS150–
Conserved
mediated
yehR
protein
deletion of
Conserved
multiple coding
yehS
protein
regions.
Predicted
sensory kinase
in two–
yehT
component
system with
YehU

170

Cit–
Parent
Cit+
Mutant
ZDB294

Mutation

ZDB30

Position

Gene or
Genes
Involved

Table 3.11d │ Pair 4: ZDB30 (Generation 32,000) and Cit+ Mutant
ZDB294

Description

Gene
Product(s)

Cit–
Parent
Cit+
Mutant
ZDB294

Mutation

ZDB30

Position

Gene or
Genes
Involved

Table 3.11d │ Pair 4: ZDB30 (Generation 32,000) and Cit+ Mutant
ZDB294

Predicted
sensory kinase
in two–
yehU
component
system with
YehT
MerR–like
yehV
regulator
Unknown
ECB 02057
function
Membrane
component of
an ABC
yehW
transporter
involved in
osmoprotection
Membrane
component of
an ABC
yehX
transporter
involved in
osmoprotection
Membrane
component of
an ABC
yehY
transporter
involved in
osmoprotection
Periplasmic
component of
an ABC
yehZ
transporter
involved in
osmoprotection
Red fill indicates presence of mutation. Gray fill indicates absence of corresponding mutation.

171

+

The re-evolved Cit mutants arose by diverse mutational processes, though all affected citT
(Table 3.10). Eight have citT duplications similar to the original case, although no two share the
same boundaries (Figure 3.12). In seven of these, the duplication generated a version of the rnkcitT module; in the other, the second citT occurs downstream of the rna promoter. Six mutants
have an IS3 element inserted in the 3' end of citG (Figure 3.11). IS3 carries outward-directed
promoter elements that can activate adjacent genes (Treves et al 1998, Charlier et al 1982). Two
mutants have large duplications that encompass all or part of the cit operon. One mutant has a
large inversion that places most of that operon downstream of the promoter for the fimbrial
regulatory gene fimB, while another has a deletion in citG that presumably formed a new
+

promoter. Thus, the Cit phenotype can evolve in potentiated backgrounds by a variety of
mutational processes that recruit several different promoters that enable CitT expression during
aerobic metabolism. It is also interesting that most of these replay mutants exhibit stronger
+

phenotypes (Figure 3.13) than the earliest Cit clones in the main experiment (Figure 3.6 – 7).

PERSPECTIVE
The evolution of citrate-utilization in an experimental E. coli population provided an
unusual opportunity to examine in detail the multi-step origin of a key innovation.

From

comparative studies it has long been long recognized that gene duplications play an important
creative role in evolution, particularly by generating redundancies that allow neofunctionalization (Gould and Vrba 1982, Taylor and Raes 2004, True and Carroll 2002, Zhang
2003, Bergthorsson et al 2007). Our findings highlight the less-appreciated capacity of gene

172

IS3 Insertion Sites

Figure 3.13 │ Mutations that produced Cit+ phenotype in 14 replay experiments. The red

box shows the boundaries of the 2,933-bp amplified segment that actualized the Cit+ function in
the original long-term population. Blue boxes show citT-containing regions amplified in 8
replays that produced Cit+ mutants. Vertical black lines mark 5 locations in citG where IS3
insertions produced Cit+ mutants in 6 other replays.

173

Set 3

ln OD420

ln OD420

ln OD420

-2
-2
-3
-3
-4
-4
-5
-5
-6
-6
-7
-7

ln OD420

Set 2

ln OD420

Set 1

ln OD420

0
0
-1
-1

ZDB285
ZDB285
ZDB292
ZDB292
ZDB163
ZDB163

0
0

0
0
-1
-1
-2
-2
-3
-3
-4
-4
-5
-5
-6
-6
-7
-7

0
-1
-1
-2
-2
-3
-3
-4
-4
-5
-5
-6
-6
-7
-7

24
24

ZDB286
ZDB286
ZDB279
ZDB279
ZDB548
ZDB548

48
48

ZDB288
ZDB288
ZDB294
ZDB294

72
72

ZDBr218 ZDB547
ZDBr218
ZDB547

0
0

ZDB161
ZDB161

ZDB166
ZDB166

ZDB165
ZDB165

24
24

ZDB164
ZDB164
48
48

ZDB281
ZDB281

ZDB546
ZDB546

ZDB290
ZDB290

0
0

96
96

72
72

96
96

ZDB283
ZDB283

ZDB549
ZDB549

24
24

48
72
48
72
Incubation (h)
Incubation (h)

96
96

Figure 3.12 │ Growth of Cit+ mutants derived in replay experiments. The 19 spontaneous
Cit+ mutants isolated during the course of replay experiments vary in how quickly they transition
from the glucose to the citrate and how well they grow on the citrate in DM25. Here mutants are
grouped according to the mutations responsible for the Cit+ phenotype. The eight mutants in Set
1 have tandem citT duplications similar to that which evolved in the original population. The six
mutants in Set 2 mutants have IS3 insertions in citG. The five mutants in Set 3 have a variety of
mutations, as described in the main text and Supplementary Table 12.

174

duplications to mediate exaptation events by altering gene-regulatory networks, especially by
promoter capture (Whoriskey et al 1987). The evolution of citrate-utilization also highlights that
such actualizing mutations are only part of the process by which novelties arise. Before some
new function can arise, it may be essential for a lineage to evolve a potentiating genetic
background that allows the actualizing mutation to occur or the new function to be expressed.
Finally, novel functions often emerge in rudimentary forms that must be refined by further
mutations and selection to exploit fully the ecological opportunities. This three-step process – in
which potentiation makes a trait possible, actualization makes the trait manifest, and refinement
makes it effective – is likely typical of many new functions.

METHODS
Evolution Experiment
The long-term experiment has been described in detail elsewhere (Lenski 2004, Lenski et
al 1991). In brief, twelve populations of E. coli B were started in 1988, and they have evolved
for more than 20 years under conditions of daily 100-fold dilutions in a minimal medium, DM25,
containing 139 µM glucose and 1700 µM citrate (Lenski et al 1991).

The populations

experience ~6.64 generations per day, and they had evolved for over 40,000 generations when
this study began. Every 500 generations, viable samples of each population were frozen at –
80ºC with glycerol added as a cryoprotectant. The focus of this study was population Ara–3, in
which the ability to grow aerobically on citrate evolved after more than 30,000 generations
(Blount et al 2008).

175

Genomic DNA Isolation
Clones from the Ara–3 population were revived from frozen stocks by overnight growth
in LB medium at 37°C with aeration. DNA was extracted and purified from several mL of each
culture using the Qiagen Genomic-tip 100/G kit (Qiagen, Hilden, DE).

Whole-genome Shotgun Re-sequencing and Mutation Detection
Clones were sequenced on Illumina GA, GA II, and GA IIx instruments. The resulting read
data have been deposited in the NCBI SRA database (SRP004752). Most genomic datasets
consist of single-end reads only, but additional mate-paired libraries were sequenced for ZDB30
and ZDB172. Reads were mapped to the reference genome of the ancestral strain (REL606)
(Jeong et al 2009), and mutations were predicted using the breseq computational pipeline
(Barrick and Knoester 2010). This pipeline detects point mutations, deletions, and new sequence
junctions that may indicate IS-element insertions or other rearrangements, as described in its
online documentation.

Large duplications and amplifications were predicted manually by

examining the depth of read coverage across each genome.
Mutation lists were further refined by manually reconstructing the most plausible series
of mutational events generating the genomic differences observed in each clone. This procedure
involved: (1) splitting predicted changes into multiple mutations based on phylogenetic
relationships, (2) assigning mutations to genomes when a subsequent mutational event prevented
their detection (e.g., a SNP within a later deletion), and (3) correcting false-positives for a
handful of difficult-to-predict mutations by examining read alignments and coverage in all
clones. After these procedures, 17 homoplasies remained in the phylogenetic tree, most of which

176

appear to indicate mutational hot-spots: 9 are IS-element insertions at specific sites, 7 are
insertions or deletions at the boundaries of IS-elements, and 1 is a single-base substitution not
associated with IS-elements. Given these signatures, it is likely that most or all of these
mutations occurred independently in multiple lineages.
The copy number and configuration of the citT module were estimated from the number
of reads overlapping the new rnk-citT sequence junction relative to the two original flankingsequence junctions, and from the average read-depth in this region relative to a single-copy
region (Table 3.8).

Phylogenetic Reconstruction
An initial parsimony-based tree was calculated using presence-absence data for all
mutational events in each genome using the dnapars program from PHYLIP. Branch lengths
were recalculated for the tree using an irreversible Camin-Sokal model (Camin and Sokal 1965),
because the ancestral states are known and reversions are extremely unlikely given the genome
size and number of mutations observed.

A maximum-likelihood model was then used to

estimate all branching times and two mutation rates, one for non-mutator branches and one for
mutator branches. This model fixed the generation when each clone was sampled and assumed
that mutations accumulated on each branch according to a Poisson process.
Ten mutations in nine genes (ybaL, nadR, hemE, cspC, yaaH, leuA, tolR, arcB, and gltA)
were identified as phylogenetically informative based on their association with particular clades
in the Ara–3 population (Table 3.9). Sanger sequencing of PCR-amplified gene fragments was
used to detect the presence or absence of these mutations in replay clones, which were then
mapped onto the population phylogeny according to the keys shown in Figure 3.9. Owing to the

177

large number of clones and genes under consideration, not all genes were sequenced for all
clones. Appendix Table 1 shows sequencing results and phylogenetic assignments. Table 3.16
provides sequences of the primer pairs used to amplify each locus.

PCR Screens for cit Amplifications
The cit amplification was detected in both population samples and clones by PCR
amplification across the rnk-citG junction using outward-directed primers specific to citT (Table
3.12). When screening population samples, three separate PCR reactions were run for each time
point, and the template used was a 1:10 dilution of the population sample frozen for that
generation.

Expression Experiments
+

The regulation of expression of the Cit phenotype was measured using luciferase-based
reporter constructs. The complete upstream regions for the native citT and rnk genes were PCRamplified from the cognate reporters from the E. coli transcriptional library (Zaslaver et al 2006)
using the primers pZE05 and pZE07.

The intergenic region of the rnk-citT module was

+

amplified from Cit clone ZDB172 using primers nctForward and nctReverse (Table 3.12). The
PCR products were cloned into the low-copy (1-2) plasmid pCS26-pac, which contains a
kanamycin-resistance gene and the luciferase operon (luxCDABE) (Bjarnason et al 2003). Each
reporter plasmid was transformed into clones REL606, ZDB30 and ZDB172.
Prior to the expression assays, strains were grown in a 96-well plate (BD Biosciences,
Bedford, MA, USA) in 200 µL per well of DM25 supplemented with 50 µg/mL kanamycin, with

178

Table 3.12│ Primer pairs used in this study
Gene or
Region
Amplified

Primer
Name

ybaLmut F
ybaLmut R
nadRmut F
nadRmut R
hemEmut F
hemEmut R
cspCmut F
cspCmut R
yaaHmut F
yaaHmut R
leuAmut F
leuA mutation
leuAmut R
tolRmut F
tolR mutation
tolRmut R
arcBmut F
arcB
mutation
arcBmut R
gltAmut F
gltA mutation
gltAmut R
citTout F
cit
amplification citTout R
ybaL
mutation
nadR
mutation
hemE
mutation
cspC
mutation
yaaH
mutation

citAmpJ
fragment

citAmpJ F
citAmpJ R

rnk promoter
nctForward
region of rnkcitT module
for
nctReverse
expression
studies
citT–citG F
citT-citG
fragment for
gene-gorging citT–citG R
citGfrag
citGfrag F
fragment for
gene-gorging citGfrag R

Primer Sequence
5’ CATCGCCCTGTTCCATCATTCCT 3’
5’ ACCCCGCTTATCACCACCATTGTT 3’
5’ ATGGTCGCGATTATGTCTTTTCAC 3’
5’ CGTTTCATCGCGGTTATCTCTG 3’
5’ GTGCCGGACGCGATGGGGTTAG 3’
5’ CACTGTCCGCCGCCTTTGGTA 3’
5’ GGGCAAATATCCGAACG 3’
5’ AGCCTTATATTGGTGCCTCAT 3’
5’ CTTTCGCGTCAGGTTGGTGTG 3’
5’ CCTGCCTGCGCCGGATGGTTAG 3’
5’ GAATGCGCCGCTGCCAACA 3’
5’ GCCTCAACCAGCGCGTAAACAAA 3’
5’ GCCTCAACCAGCGCGTAAACAAA 3’
5’ ACTTCCGCCACCACCTGCTCTG 3’
5’ TGTCGCGACCAAAGCCCATCA 3’
5’ GCCCTCGTCGTTCTTGCCATTGT 3’
5’ TGTGTTTAACGGAGCTGATTTCTT 3’
5’ GCTGGCGACCGATTCTAACTACCT 3’
5’ GTCCTGGGGTGATTATTTACGGCT 3’
5’ CAATAACGCAAATAGTAACCGCAA 3’
5’ TTTTTTGGATCCGGTTCGAATGCCCCCTTTTT 3’
5’ TTTTTTGTCGACGGTAACCCTGCGTATTTGACTGAA 3’
5’ AAAAAAGGATCCGACACCCATCACCACCAGT 3’
5’ AAAAAACTCGAGACGCCATCAACGCTCCGCTTTCT 3’
5’ AACCAGCCAGGCCCCATTTCAGC 3’
5’ AAAAAAGGATCCCACGCCTTGCCGCATTACCTCACT 3’
5’ TTTTTTGGATCCGGGGGTTCGAATGCCCCCTTTTT 3’
5’ GCACAAAGATATGGCGCTGGAAGA 3’

179

PCR
Prodcuct
Size
503
bp
459
bp
524
bp
412
bp
1030
bp
497
bp
400
bp
709
bp
626
bp
1807
bp
529
bp

707
bp

648
bp
694
bp

Table 3.12│ Primer pairs used in this study
Gene or
Region
Amplified
rnk promoter
and cit
amplification
junction
construct for
gene-gorging
rnk-citT
module for
cloning into
pUC19
citT internal
fragment for
Southern
hybridizations
Amplification
of genomic
region
immediately
upstream of
citT

Primer
Name

Primer Sequence

citT–citG
Gorge F

5’ TAGGGATAACAGGGTAATAACCAGCCAGGCCCCATTTCAGC 3’

citGfrag R
citTAmpX F
citTAmpX R
citTprobe F

5’ GCACAAAGATATGGCGCTGGAAGA 3’
5’ AAAAAAGGATCCGGGCAGCAACCGATTTAGG 3’
5’ AAAAAAGTCGACAACGCTCCGCTTTCTGC 3’
5’ AGCCGTAAATAATCACCCCAGGAC 3’

citTprobe R

5’ TTGCGGTTACTATTTGCGTTATTG 3’

citTupstrm R

1889
bp*

2490
bp

1173
bp

5’ CTCTCCCGCCGCGACTATTCA 3’

citTupstrm F

5’ CAATAACGCAAATAGTAACCGCAA 3’

*Length for fully assembled construct.
†

PCR
Prodcuct
Size

Length without deletions or insertions.

180

1264
bp†

constant shaking at 37°C for two 24-h cycles to ensure proper preconditioning. Fresh overnight
cultures were then diluted 100-fold into a black, clear-bottomed 96-well plate (9520 Costar;
Corning, Lowell, MA, USA), with 150 µL per well of DM25, and covered with breathable
sealing membrane (Nunc, Rochester, NY, USA) to prevent evaporation. Luciferase activity was
2

read in a Wallac Victor plate reader (Perkin Elmer Life Sciences, Boston, MA, USA) every 20
minutes for 19 h with 90 s of 2-mm orbital shaking prior to each reading.

Assays were

performed in quadruplicate in the same plate.

Isogenic Strain Construction
–

A single-copy chromosomal rnk-citT module was made in a Cit background using the
“gene gorging” allele transfer technique (Herring et al 2003). Owing to difficulties inherent to
the manipulation of amplified genes, we did not attempt to move an entire citT amplification
segment. Instead, we engineered a construct of the cit amplification junction containing the rnk
promoter from three smaller fragments that were PCR-amplified from the 32,000-generation Cit

+

clone ZDB172. The first fragment contained the cit amplification junction (citAmpJ), including
the rnk promoter region, and was PCR-amplified using the primers citTAmpJ F and citTAmpJ R
(Table 3.12). The second fragment contained the citT-citG junction (citT-citG) and was PCRamplified using the primers citT-citG F and citT-citG R. The third fragment contained sequence
internal to the citG gene (citGfrag) and was PCR-amplified using the primers citGfrag F and
citGfrag R. Each primer pair was designed with restriction sites that allowed ligation of the
fragments to a hybrid construct in which the cit amplification junction, including the rnk
promoter, was embedded within >500 bp of citG flanking sequence. The assembled fragment

181

was then PCR-amplified using the primers citT-citG Gorge F and citGfrag R (Table 3.12). The
primer citT-citG Gorge F incorporated an I-Sce-I restriction site necessary for the gene-gorging
procedure, which was then performed as described elsewhere (Herringg et al 2003).

We

screened for putative transformants by performing PCR and by testing for expression of a Cit

+

phenotype as indicated by a positive reaction on Cristensen’s Citrate Agar (Christensen 1949).
Successful constructs were then confirmed by Sanger sequencing.

Growth Trajectories
Strains of interest (Table 3.13) were revived from frozen stocks by growing them in LB,
and they were then preconditioned by two 24-h culture cycles in DM25.

Following

preconditioning, 100 µL of each culture was diluted into 9.9 mL of DM25, and then dispensed as
200 µL aliquots into randomly assigned wells in a 96-well plate. For all pUC19::rnk-citT
transformants, the medium was supplemented with 100 µg/mL ampicillin to ensure plasmid
retention.

Growth trajectories were replicated 6- to 8-fold for every strain.

To reduce

evaporation, strains were grown in the innermost 60 wells, while the outermost 36 wells were
filled with 300 µL of a saline buffer. In those cases when the assays lasted longer than 96 hours,
the buffer was replenished after 96 hours. OD420 was measured every 10 minutes by a VersaMax
automated plate reader (Molecular Devices, Sunnyvale, CA, USA). Plates were shaken orbitally
for 5 seconds prior to each measurement, but were otherwise stationary.

Fitness Assays
Relative fitness was measured in competition experiments described in detail elsewhere
(Lenski et al 1991). In brief, we inoculated 0.05 mL each of preconditioned cultures of ZDB595

182

+

and ZDB63, an Ara revertant of ZDB30, into 9.9 mL of fresh DM25 with 10-fold replication.
Starting densities of each strain were determined by dilution plating on tetrazolium arabinose
+

(TA) plates, on which the Ara– strain ZDB595 and the Ara strain ZDB63 made red and white
colonies, respectively. The competition cultures were then propagated for three daily transfer
cycles, after which the final densities of each strain were again determined by dilution plating on
TA plates. Relative fitness was calculated as the ratio of the realized net growth rates of the two
strains over the course of the competition experiment (Lenski et al 1991).

Plasmid Construction
A fragment containing the complete rnk-citT module was PCR-amplified from the
+

genome of Cit clone ZDB172 using the primers citTAmpX F and citTAmpX R (Table 3.12).
The fragment was then inserted into the cloning site of pUC19 (NEB, Ipswich, MA, USA). The
corresponding region of the plasmid, pUC19::rnk-citT, was confirmed by Sanger sequencing.
The plasmid was transformed into strains REL606, ZDB30, ZDB199 and ZDB200.

+

Cit

transformants were identified by positive reactions on Cristensen’s Citrate Agar (Christensen
1949).

+

Identification of Cit -conferring Mutations in Replays
+

Nineteen Cit mutants isolated during replay and related experiments (Blount et al 2008)
were analyzed to identify the responsible events. These mutants were first checked for large
changes in the region containing citT by Southern hybridization with citT-specific probes.

183

Table 3.13 │ Clones used in growth-trajectory experiments
Clone

Gen

Description

REL606
CZB152

0
33,000

Cit clone from main population

CZB154

33,000

Cit clone from main population

Ancestor
+
+

Growth Curve
Locations
Figure 3.10
Figure 3.7, 8
Figure 3.7

–

ZDB30

32,000

ZDB143

32,500

ZDB161

Potentiated Cit clone from
Clade 3
+
Cit clone from main population
+

32,000

Figure 3.6, 7, 10
Figure 3.7

+

Figure 3.12
Figure 3.12

+

Figure 3.12

+

Figure 3.12

+

Cit mutant of ZDB28

ZDB163

32,500

Cit mutant of ZDB31

ZDB164

32,000

Cit mutant of ZDB30

ZDB165

32,000

Cit mutant of ZDB30

ZDB166

32,000

Cit mutant of ZDB29

Figure 3.12

ZDB172

32,000

Cit clone from main population

Figure 3.6, 7
Figure 3.10

+

–

ZDB199

31,500

Potentiated Cit clone from
Clade 1

ZDB200

31,500

Cit clone from Clade 2

–

Figure 3.10

+

Figure 3.12

+

Figure 3.12

ZDB279

31,500

Cit mutant of ZDB199

ZDB281

32,000

Cit mutant of ZDB183

ZDB283

32,000

Cit mutant of ZDB30

ZDB285

20,000

Cit mutant of ZDB464

ZDB286

20,000

Cit mutant of ZDB464

ZDB288

27,000

Cit mutant of ZDB309

ZDB290

27,000

Cit mutant of ZDB310

+

Figure 3.12

+

Figure 3.12

+

Figure 3.12

+

Figure 3.12

+

Figure 3.12

+

Figure 3.12

ZDB292

31,000

Cit mutant of ZDB390

ZDB294

32,000

Cit mutant of ZDB30

ZDB546

32,500

Cit mutant of ZDB31

ZDB547

30,500

Cit mutant of ZDB20

+

Figure 3.12

+

Figure 3.12

+

Figure 3.12

+

Figure 3.12

+

ZDB548

32,500

Cit mutant of ZDB32

ZDB549

32,500

Cit mutant of ZDB31

Figure 3.12

ZDB564

31,500

Cit clone from main population

Figure 3.6, 7, 8

+

+

ZDB595

32,000

Cit isogenic construct of
ZDB30 in which rnk-citT
module promoter was inserted
into the chromosome
184

Figure 3.6

Table 3.13 │ Clones used in growth-trajectory experiments
Clone

Gen

ZDB611

0

ZDB612

32,000

ZDB614

31,500

ZDB615

31,500

ZDBr218

31,500

Description
pUC19::rnk-citT transformant of
REL606
pUC19::rnk-citT transformant of
ZDB30
pUC19::rnk-citT transformant of
ZDB199
pUC19::rnk-citT transformant of
ZDB200
+
Cit mutant of ZDB25

185

Growth Curve
Locations
Figure 3.10
Figure 3.8, 10
Figure 3.10
Figure 3.10
Figure 3.12

Genomic DNA was digested with EcoRV (NEB, Ipswich, MA, USA), and fragments were
separated on 0.8% agarose gels with a 1-kb ladder (NEB, Ipswich, MA, USA), and then
transferred to nylon membranes. Hybridizations were performed at 68ºC. The citT-specific
probe was an internal fragment amplified by PCR using the primers described in Table 3.16,
purified using an Illustra GFX PCR DNA purification kit (GE Healthcare, Little Chalfont,
Buckinghamshire, UK), and labeled using the DIG DNA labeling and detection kit (Roche,
Basel, Switzerland). Most, but not all, mutants showed enlarged citT bands. We then tried to
PCR-amplify across possible amplification boundaries of each mutant with the same outwarddirected citT primers used to screen for the amplification in the original Ara–3 population (Table
3.12). PCR products were thus obtained for 8 of the 19 clones. Sanger sequencing of these
products permitted resolution of novel junctions that were consistent with amplifications similar,
but not identical, to that in the original population. For 7 of the remaining 11 clones, PCR
products of altered size were obtained when the region immediately upstream of citT was
amplified, and Sanger sequencing determined that each altered product was caused by either a
small deletion or an IS3 insertion in citG. To identify mutations at other loci, as well as
mutations affecting the cit region, we re-sequenced the genomes of six mutants, as described
above. These genomes included three of the four mutants for which the mutations conferring the
+

Cit phenotype were not resolved using the approaches described above.

ACKNOWLEDGEMENTS
We thank N. Hajela, M. Kauth and S. Sleight for laboratory assistance, J. Meyer for
discussion, and C. Turner for comments on the manuscript. We thank the MSU Research
Technology Support Facility for sequencing services.

186

We acknowledge support from the

DARPA “Fun Bio” Program (HR0011-09-1-0055 to R.E.L), the US National Institutes of Health
(K99GM087550 to J.E.B.); the US National Science Foundation (DEB-1019989 to R.E.L)
including the BEACON Center for the Study of Evolution in Action (DBI-0939454), a Rudolf
Hugh Fellowship (to Z.D.B.), a DuVall Family Award (to Z.D.B.), and a Barnett Rosenberg
Fellowship (to Z.D.B).

187

APPENDIX

188

APPENDIX: Supplemental Data Table for Chapter 3
–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones

0
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
10
10
10
10
10

hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

REL606*
ZDB400
ZDB401
ZDB402
ZDB403
ZDB404
ZDB405
ZDB406
ZDB407
ZDB408
ZDB409*
ZDB410
ZDB411
ZDB412
ZDB413
ZDB414
ZDB415
ZDB416
ZDB417
ZDB418
ZDB419

Clade

Ancestor
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
(C1,C2,
C3)
(C1,C2,
C3)
(C1,C2,
C3)
n.d.
UC

ZDB1
ZDB2
ZDB3
ZDB420
ZDB421

189

–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones

10
10
10
10

hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

ZDB422
ZDB423

UC
n.d.
(C1,C2,
C3)
(C1,C2,
C3)
UC
n.d.
UC
UC
UC
UC
n.d.
n.d.
UC
n.d.
UC
n.d.
(C1,C2,
C3)
n.d.
UC
UC
n.d.
UC
n.d.
(C2,C3)
UC
UC
(C2,C3)
n.d.

ZDB424
ZDB425

10
10
10
10
10
10
10
10
10
10
10
15
15

ZDB426
ZDB427
ZDB428
ZDB429*
ZDB430
ZDB431
ZDB432
ZDB433
ZDB434
ZDB435
ZDB436
ZDB437

15
15
15
15
15
15
15
15
15
15
15

Clade

ZDB439
ZDB440
ZDB441
ZDB442
ZDB443
ZDB444
ZDB445
ZDB446*
ZDB447
ZDB448
ZDB449

ZDB438

190

–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones

15
15
15
15
15
15
15
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
25
25
25

hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

ZDB450
ZDB451
ZDB452
ZDB453
ZDB454
ZDB455
ZDB456
ZDB4
ZDB5
ZDB6
ZDB457
ZDB458*
ZDB459
ZDB460
ZDB461
ZDB462
ZDB463
ZDB464*
ZDB465
ZDB466
ZDB467*
ZDB468
ZDB469
ZDB470
ZDB471
ZDB472
ZDB473
ZDB7
ZDB8
ZDB9

Clade

UC
n.d.
UC
n.d.
n.d.
(C2,C3)
n.d.
(C2,C3)
(C2,C3)
C1
(C2,C3)
(C2,C3)
C1
(C2,C3)
(C2,C3)
(C2,C3)
(C2,C3)
(C2,C3)
(C2,C3)
C1
(C2,C3)
C1
(C2,C3)
(C2,C3)
(C2,C3)
(C2,C3)
(C2,C3)
(C2,C3)
C2
C1

191

–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones

25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
27
27
27
27
27
27
27
27
27
27
27
27
27

hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

ZDB474
ZDB475
ZDB476
ZDB477
ZDB478
ZDB479
ZDB480
ZDB481
ZDB482
ZDB483
ZDB484
ZDB485
ZDB486
ZDB487
ZDB488
ZDB489
ZDB490
ZDB300
ZDB301
ZDB302
ZDB303
ZDB304
ZDB305
ZDB306
ZDB307
ZDB308
ZDB309
ZDB310
ZDB311
ZDB312

Clade

C2
C1
C3
C1
C3
C3
(C2,C3)
C3
(C2,C3)
C3
(C2,C3)
C3
C3
C3
C3
C3
C3
C3
C2
C1
C1
C2
C2
C2
C2
C2
C3
C3
C3
C3

192

–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones

27
27
27
27
27
27
27
27
27
27
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28

hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

ZDB313
ZDB314
ZDB315
ZDB316
ZDB317
ZDB318
ZDB319
ZDB10
ZDB11
ZDB12
ZDB320
ZDB321
ZDB322
ZDB323
ZDB324
ZDB325
ZDB326
ZDB327
ZDB328
ZDB329
ZDB330
ZDB331
ZDB332
ZDB333
ZDB334
ZDB335
ZDB336
ZDB337
ZDB338
ZDB339

Clade

C1
C3
C2
C2
C3
C3
C2
C2
C1
C2
C2
C2
C1
C2
C2
C2
C2
C2
C2
C1
C2
C2
C2
C2
C3
C2
C2
C2
C2
C3

193

–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones

29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
30
30
30
30
30
30
30
30
30
30

hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

ZDB13
ZDB14
ZDB15
ZDB340
ZDB341
ZDB342
ZDB343
ZDB344
ZDB345
ZDB346
ZDB347
ZDB348
ZDB349
ZDB350
ZDB351
ZDB352
ZDB353
ZDB354
ZDB355
ZDB356
ZDB16*
ZDB17
ZDB18
ZDB357*
ZDB358
ZDB359
ZDB360
ZDB361
ZDB362
ZDB363

Clade

C3
C3
C2
C2
C2
C1
C2
C2
C2
C1
C1
C2
C1
C1
C2
C1
C2
C2
C1
C2
C1
C3
C3
C2
C2
C2
C1
C2
C2
C2

194

–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones

30
30
30
30
30
30
30
30
30
30
30.5
30.5
30.5
31
31
31
31
31
31
31
31
31
31
31
31
31
31
31
31
31

hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

ZDB364
ZDB365
ZDB366
ZDB367
ZDB368
ZDB369
ZDB370
ZDB371
ZDB372
ZDB373
ZDB19
ZDB20
ZDB21
ZDB22
ZDB23
ZDB24
ZDB374
ZDB375
ZDB376
ZDB377
ZDB378
ZDB379
ZDB380
ZDB381
ZDB382
ZDB383
ZDB384
ZDB385
ZDB386
ZDB387

Clade

C2
C2
C2
C2
C2
C2
C1
C2
C2
C2
C3
C3
C2
C2
C3
C1
C2
C2
C2
C1
C2
C2
C2
C1
C2
C2
C2
C1
C2
C2

195

–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones

31
31
31
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
31.5
32
32
32
32

hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

ZDB388
ZDB389
ZDB390
ZDB25
ZDB26
ZDB27
ZDB197
ZDB198
ZDB199*
ZDB200*
ZDB201
ZDB202
ZDB203
ZDB204
ZDB205
ZDB206
ZDB207
ZDB208
ZDB209
ZDB210
ZDB211
ZDB212
ZDB213
ZDB214
ZDB215
ZDB216
ZDB28
ZDB29
ZDB30*
ZDB183

Clade

C2
C2
C1
C3
C1
C3
C2
C1
C1
C2
C1
C2
C2
C1
C1
C2
C2
C2
C2
C2
C2
C2
C2
C2
C1
C2
C3
C2
C3
C3

196

–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones

32
32
32
32
32
32
32
32
32
32
32
32
32
32
32
32
32.5
32.5
32.5
32.5
32.5
32.5
32.5
32.5
32.5
32.5
32.5
32.5
32.5
32.5

hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

ZDB184
ZDB185
ZDB186
ZDB187
ZDB188
ZDB189
ZDB190
ZDB191
ZDB192
ZDB391
ZDB392
ZDB393
ZDB394
ZDB395
ZDB396
ZDB397
ZDB31
ZDB32
ZDB33
ZDB146
ZDB147
ZDB148
ZDB149
ZDB150
ZDB151
ZDB152
ZDB153
ZDB154
ZDB155
ZDB156

Clade

C1
C3
C1
C1
C2
C1
C2
C3
C2
C2
C2
C2
C1
C2
C2
C1
C3
C2
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1

197

–

Appendix Table A.1 │ Phylogenetic placement of Cit replay clones
hemE

cspC

yaaH

leuA

tolR

arcB

arcB

gltA

4

521

778

69

208

236

772

Ancestral
G
Nucleotide
Evolved
C
Nucleotide

636

Gene
Position

nadR

Clone

ybaL

Gen
(k)

1010

Gene

133

Phylogenetically Informative Mutations

A

T

C

T

C

C

G

T

G

C

G

A

G

T

T

C

A

A

32.5
ZDB157
32.5
ZDB158*
32.5
ZDB159
32.5
ZDB160
32.5
ZDB398
32.5
ZDB399
*Genome has been sequenced.

Clade

C3
C2
C2
C1
C1
C2

Red fill indicates presence of mutation has been established by sequencing. Gray fill indicates
absence of mutation has been established by sequencing. No fill indicates that the presence or
absence of mutation was not examined.
Clade refers to UC, C1, C2, or C3 as shown in Fig. 1 of the main text. When two or more clades
are grouped by parentheses, either the clone belongs to the basal group or the clone’s placement
could not be resolved further based on the available data. n.d. indicates that the clone belongs to
some other early clade or its placement could not be resolved based on the available data.

198

REFERENCES

199

REFERENCES
1.

Adam, D., Dimitrijevic, N. & Schartl, M. Tumor suppression in Xiphophorus by an
accidentally acquired promoter. Science 259, 816–819 (1993).

2.

Andersson, D.I., Slechta, E.S. & Roth, J.R. Evidence that gene amplification underlies
adaptive mutability of the bacterial lac operon. Science 282, 1133–1135 (1998).

3.

Barrick, J.E. & Knoester, D.B. breseq. http://barricklab.org/breseq (2010).

4.

Barrick, J.E. & Lenski, R.E. Genome-wide mutational diversity in an evolving population of
Escherichia coli. Cold Spring Harbor Symposia on Quantitative Biology 74, 1–11 (2009).

5.

Barrick, J.E., Yu, D.S., Yoon, S.H., Jeong, H., Oh, T.K., Schneider, D., Lenski, R.E., &
Kim, J.F. Genome evolution and adaptation in a long-term experiment with E. coli. Nature
461, 1243–1247 (2009).

6.

Bentley, D.R. Whole-genome resequencing. Current Opinion in Genetics and Development
16, 545–552 (2006).

7.

Bergthorsson, U., Andersson, D.I. & Roth, J.R. Ohno's dilemma: evolution of new genes
under continuous selection. Proceedings of the National Academy of Sciences (USA) 104,
17004–17009 (2007).

8.

Bjarnason, J., Southward, C.M. & Surette, M.G. Genomic profiling of iron-responsive
genes in Salmonella enterica Serovar Typhimurium by high-throughput screening of a
random promoter library. Journal of Bacteriology, 185, 4973–4982 (2003).

9.

Blount, Z.D., Borland, C.Z. & Lenski, R.E. Historical contingency and the evolution of a
key innovation in an experimental population of Escherichia coli. Proceedings of the
National Academy of Sciences (USA) 105, 7899–7906 (2008).

10. Bock, R. & Timmis, J.N. Reconstructing evolution: gene transfer from plastics to the
nucleus. BioEssays 30, 556–566 (2008).
11. Camin, J.H. & Sokal, R.R. A method for deducing branching sequences in phylogeny.
Evolution 19, 311–326 (1965).
12. Charlier, D., Piette, J. & Glansdorff, N. IS3 can function as a mobile promoter. Nucleic
Acids Research 10, 5935–5948 (1982).
13. Christensen, W.B. Hydrogen sulfide production and citrate utilization in the differentiating
of enteric pathogens and coliform bacteria. Research Bulletin Weld County Health
Department of Greeley Colorado 1, 3–16 (1949).
14. Elena, S.F. & Lenski, R.E. Evolution experiments with microorganisms: the dynamics and
genetic bases of adaptation. Nature Reviews Genetics 4: 457–469 (2003).

200

15. Fogle, C.A., Nagle, J.L. & Desai, M.M. Clonal interference, multiple mutations and
adaptation in large asexual populations. Genetics 180, 2163–2173 (2008).
16. Glickman, B.W. & Radman, M. Escherichia coli mutator mutants deficient in methylinstructed DNA mismatch correction. Proceedings of the National Academy of Sciences
(USA) 77, 1063–7 (1980).
17. Gould, S.J. & Vrba, E.S. Exaptation – a missing term in the science of form. Paleobiology
8, 4–15 (1982).
18. Gunsalus, R.P. & Park, S.J. Aerobic-anaerobic gene regulation in Escherichia coli: control
by the ArcAB and Fnr regulons. Research in Microbiology. 145, 437–450 (1994).
19. Hall, B.G. Chromosomal mutation for citrate utilization by Escherichia coli K-12. Journal
of Bacteriology 151, 269–273 (1982).
20. Hegreness, M. & Kishony, R. Analysis of genetic systems using experimental evolution and
whole-genome sequencing. Genome Biology 8, 201 (2007).
21. Herring, C.D., Glasner, J.D. & Blattner, F.R. Gene replacement without selection:
regulated suppression of amber mutations in Escherichia coli. Gene 311, 153–63 (2003).
22. Jacob, F. Evolution and tinkering. Science 196, 1161–1166 (1977).
23. Jacob, F. The Possible and the Actual. (Univ. Washington Press, Seattle, 1982).
24. Janausch, I.G., Zientz, E., Tran, Q.H., Kroger, A. & Unden, G. C4-dicarboxylate carriers
and sensors in bacteria. Biochimica et Biophysica Acta 1553, 39–56.
25. Jeong, H., Barbe, V., Lee, C.H., Vallenet, D., Yu, D.S., Choi, S.H., Couloux, A., Lee, S.W.,
Yoon, S.H., Cattolico, L., Hur, C.G., Park, H.S., Sequrens, B., Kim, S.C, Oh, T.K., Lenski,
R.E., Studier, F.W., Daegelen, P., & Kim, J.F. 2009. Genome sequences of Escherichia
coli B strains REL606 and BL21(DE3). Journal of Molecular Biology 394, 644–652
(2009).
26. Koser, S.A. Correlation of citrate-utilization by members of the colon-aerogenes group with
other differential characteristics and with habitat. Journal of Bacteriology 9, 59–77 (1924).
27. Lara, F.J.S. & Stokes, J.L. Oxidation of citrate by Escherichia coli. Journal of Bacteriology
63, 415–420 (1952).
28. Lenski, R.E. Phenotypic and genomic evolution during a 20,000-generation experiment with
the bacterium Escherichia coli. Plant Breeding Reviews 24, 225–265 (2004).
29. Lenski, R.E., Ofria, C., Pennock, R.T. & Adami, C. The evolutionary origin of complex
features. Nature 423, 139–144 (2003).

201

30. Lenski, R.E., Rose, M.R., Simpson, S.C. & Tadler, S.C. Long-term experimental evolution
in Escherichia coli. I. Adaptation and divergence during 2,000 generations. American
Naturalist 138, 1315–1341 (1991).
31. Lutgens, M. & Gottschalk, G.. Why a co-substrate is required for anaerobic growth of
Escherichia coli on citrate. Journal of General Microbiology 199, 63–70 (1980).
32. Mayr, E. “The emergence of evolutionary novelties”. In Evolution after Darwin Tax, S.
(Ed.) (University of Chicago Press, Chicago, 1960).
33. Nizam, S.A., Zhu, J, Ho, P.Y. & Shimizu, K. Effects of arcA and arcB genes knockout on
the metabolism in Escherichia coli under aerobic condition. Biochemical Engineering
Journal 44, 240–250 (2009).
34. Patthy, L. Genome evolution and the evolution of exon-shuffling – a review. Gene 238,
103–114 (1999).
35. Pigliucci, M. What, if anything, is an evolutionary novelty? Philosophy of Science 75, 887–
898 (2008).
36. Pos, K., Dimroth, P. & Bott, M. The Escherichia coli citrate carrier CitT: a member of a
novel eubacterial transporter family related to the 2-oxoglutarate/malate translocator from
spinach chloroplasts. Journal of Bacteriology 180, 4160–4165 (1998).
37. Reams, D., Kofoid, E., Savageau, M. & Roth, J. R. Duplication frequency in a population of
Salmonella enterica rapidly approaches steady state with or without recombination.
Genetics 184, 1077–1094 (2010).
38. Rozen, D.E., Schneider, D. & Lenski, R.E. Long-term experimental evolution in
Escherichia coli. XIII. Phylogenetic history of a balanced polymorphism. Journal of
Molecular Biology 61, 171–180 (2005).
39. Scheutz, F. & Strockbine, N.A. Genus I. Escherichia, Castellani and Chalmers 1919,
Bergey’s Manual of Systematic Bacteriology, Volume 2: The Proteobacteria, Garrity, G.M.,
Brenner, D.J., Kreig, N.R. & Staley, J.R. (Eds.). (Springer, New York, NY, 2005).
40. Shankar, S., Schlictman, D., & Chakrabarty, A.M. Regulation of nucleoside diphosphate
kinase and an alternative kinase in Escherichia coli: role of the sspA and rnk genes in
nucleoside triphosphate formation. Molecular Microbiology 17, 935–943 (1995).
41. Sniegowski, P.D., Gerrish, P.J. & Lenski, R.E. Evolution of high mutation rates in
experimental populations of E. coli. Nature 387, 703–705 (1997).
42. Taylor, J.S. & Raes, J. Duplication and divergence: the evolution of new genes and old
ideas. Annual Review of Genetics 38, 615–643 (2004).

202

43. Treves, D.S., Manning, S. & Adams, J. Repeated evolution of an acetate-crossfeeding
polymorphism in long-term populations of Escherichia coli. Molecular Biology and
Evolution 15, 789–797 (1998).
44. True, J.R. & Carroll, S.B. Gene co-option in physiological and morphological evolution.
Annual Review of Cell and Developmental Biology 18, 53–80 (2002).
45. Usakin, L.A., Kogan, G.L., Kalmykova, A.I. & Gvozdev, V.A. An alien promoter capture as
a primary step of the evolution of testes-expressed repeats in the Drosophila melanogaster
genome. Molecular Biology and Evolution 22, 1555–1560 (2005).
46. Whoriskey, S.K., Nghiem, V., Leong, P., Masson, J. & Miller, J.H. Genetic rearrangements
and gene amplification in Escherichia coli: DNA sequences at the junctures of amplified
gene fusions. Genes Development 1, 227–237 (1987).
47. Yanisch-Perron, C., Vieira, J., & Messing, J. Improved M13 phage cloning vectors and host
strains: nucleotide sequences of the M13mpl8 and pUC19 vectors. Gene 33, 103–119
(1985).
48. Zaslaver, A., Bren, A., Ronen, M., Itzhovitz, S., Kikoin, I., Shavit, S., Liebermeister, W.,
Surette, M.G., & Alon, U. A comprehensive library of fluorescent transcriptional reporters
for Escherichia coli. Nature Methods 3, 623–628 (2006).
49. Zhang, J. Evolution by gene duplication: an update. Trends in Ecology and Evolution 18,
282–298 (2003).
50. Zhu, L. & Deutsher, M.P. The Escherichia coli rna gene encoding RNase I: sequence and
unusual promoter structure. Gene 119, 1–6 (1992).

203