Content uploaded by Rees Kassen
Author content
All content in this area was uploaded by Rees Kassen
Content may be subject to copyright.
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
10.1146/annurev.micro.58.030603.123654
Annu. Rev. Microbiol. 2004. 58:207–31
doi: 10.1146/annurev.micro.58.030603.123654
Copyright
c
2004 by Annual Reviews. All rights reserved
First published online as a Review in Advance on May 14, 2004
THE ECOLOGY AND
GENETICS OF MICROBIAL
DIVERSITY
Rees Kassen
1
and Paul B. Rainey
2,3
1
Department of Biology and Center for Advanced Research in Environmental Genomics,
University of Ottawa, Ottawa, ON K1N 6N5, Canada; email: rkassen@uottawa.ca
2
School of Biological Sciences, University of Auckland, Private Bag 92019,
Auckland, New Zealand; email: p.rainey@auckland.ac.nz
3
Department of Plant Sciences, University of Oxford, South Parks Road, Oxford,
OX1 3RB, United Kingdom
KeyWords fitness, genetic variation, niche, natural selection, microbial ecology
■ Abstract Natural communities of microbes are often diverse, a fact that is difficult
to reconcile with the action of natural selection in simple, uniform environments. We
suggest that this apparent paradox may be resolved by considering the origin and
fate of diversity in an explicitly ecological context. Here, we review insights into the
ecological and genetic causes of diversity that stem from experiments with microbial
populations evolving in the defined conditions of the laboratory environment. These
studies highlight the importance of environmental structure in governing the fate of
diversity and shed light on the genetic mechanisms generating diversity. We conclude
by emphasizing the importance of placing detailed molecular-level studies within the
context of a sound ecological and evolutionary framework.
CONTENTS
THE PROBLEM OF DIVERSITY AND THE NICHE EXCLUSION
PRINCIPLE .........................................................208
THE REVEREND DALLINGER AND MECHANISM IN
MICROBIOLOGY ...................................................209
TECHNIQUES AND METHODS ........................................210
THE STRUCTURE OF THE ENVIRONMENT AND THE
MAINTENANCE OF DIVERSITY ......................................211
Simple Environments Composed of a Single Niche .........................211
Complex Environments Composed of Many Niches ........................214
Complex Environments Created by the Growth of Competitors ...............217
Complex Environments Created Through the Growth of Predators .............218
Patterns of Diversity in a Collection of Communities ........................219
THE GENETIC CAUSES OF DIVERSITY IN AN ECOLOGICAL
CONTEXT ..........................................................219
0066-4227/04/1013-0207$14.00
207
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
208 KASSEN
RAINEY
Genetic Mechanisms Underlying Diversification Through Adaptive
Radiation .........................................................220
Fitness Effects of Mutations ...........................................221
CONCLUDING REMARKS: LIFE IN THE TANGLED BANK ................223
THE PROBLEM OF DIVERSITY AND THE NICHE
EXCLUSION PRINCIPLE
Most natural communities are highly diverse, and microbial communities appear
to be no exception (28, 38, 70, 100, 107, 126, 134). Application of molecular
techniques to the study of microbial diversity (98, 101) has revealed the existence
of an incredible variety of genotypes and species in all known habitats—many of
these types appear to be ecologically equivalent in terms of resource requirements
(107, 112). For example, recent theoretical (28) and empirical (125) analyses of
the microbial diversity of soils indicate that soils harbor in the order of 7000
different taxa at an abundance of approximately 10
9
cells per cubic centimeter.
The existence of such diversity demands explanation because it is at odds with the
theoretical expectation that natural selection eliminates all but the best-adapted
type under any given set of environmental conditions. How, then, can such high
levels of diversity be maintained in natural microbial communities?
Conventional approaches to microbiology have not supplied a satisfactory an-
swer. Indeed, microbial ecologists have tended to sidestep the problem, focusing
attention almost exclusively on the documentation of diversity. Ecologists and pop-
ulation geneticists, on the other hand, have long grappled with issues relating to the
maintenance of diversity: Much of the motivation for research in ecology, popula-
tion genetics, and evolution has been driven by the need to explain the paradox of
diversity (61, 86, 123). Despite different approaches (ecologists being concerned
with the diversity of species in communities and population geneticists with the
variety of genotypes within populations), both ecologists and population geneti-
cists have identified similar principles governing the fate of diversity in natural
environments.
Paramount among these is the niche exclusion principle (47, 54), which states
that a single niche can support no more than one type, whether it be a genotype
or a species. In this review, a niche is the variety of factors in an environment
that limit the growth of one type relative to others. A limiting factor may be an
essential resource or nutrient; a set of physical conditions such as temperature,
pH, or salinity; or the existence of refuges from predation. According to this view,
the environment can be interpreted as the number of niches available to a lineage
or lineages, and the number of types supported in an environment determined by
the number of niches (85, 122). As will become apparent, this statement requires
further qualification but nevertheless forms a useful starting point. Note also that
the concept of niche has a long and tortuous history, and we refer readers to books
by Elton and Whittaker, and Chase & Leibold (42, 132, 135) for a full account.
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 209
In this article we highlight studies that have investigated the emergence or fate
(or both) of microbial diversity. We draw primarily upon investigations that use
simple experimental populations of microbes propagated in highly controlled en-
vironments where the structure of the environment—i.e., the number of niches
and the manner in which they vary in space and time—is determined by the ex-
perimenter. We focus on these studies because, by virtue of their simplicity, they
provide insight into the mechanistic causes of diversity and its maintenance. Such
insights are difficult (if not impossible) to obtain from the analysis of complex
natural communities. Our concern here is with the origin and fate of genetic vari-
ance in fitness (see below). We do not consider the functional significance of this
diversity, which, although important, is beyond the scope of this work. Interested
readers may consult Loreau et al. (88) for entrance into this literature.
THE REVEREND DALLINGER AND MECHANISM
IN MICROBIOLOGY
For most microbiologists, “mechanism” refers to the biochemical and physiologi-
cal interactions occurring among genes and their products in a cell. To understand
diversity in populations and communities, it is necessary to broaden this perspec-
tive to incorporate fitness. Fitness, a measure of the ability of one genotype to leave
offspring relative to other types, is the ultimate arbiter of ecological success. The
mechanisms underlying changes in fitness are the interactions among genotypes
and between genotypes and their environment.
To illustrate this, consider an experiment reported by the Reverend William
Henry Dallinger, a Wesleyan clergyman and president of the Royal Microscopical
Society, over a century ago. In his president’s address to the Society for 1887,
Dallinger asked “whether it was possible by change of environment, in minute life-
forms, whose life-cycle was relatively soon completed, to superinduce changes of
an adaptive character, if the observations extended over a sufficiently long period”
(30, p. 191). He started with a collection of microorganisms (what he termed
monads) and grew them in a purpose-built incubator, gradually increasing the
temperature over seven years until the apparatus was accidentally destroyed. Just
before the accident Dallinger returned his experimental populations, which by now
were growing at temperatures well beyond their normal thermal tolerance limits,
to the temperature at which he had started the experiment. None grew. He reasoned
that the lines had adapted to the increasing temperatures by natural selection and
in so doing had lost the ability to grow at their normal temperature.
Note that mechanisms invoked to explain the results of this experiment are fairly
simple: Natural selection caused by increasing temperatures leads to adaptation,
and this comes at the cost of not being able to survive at the ancestral temperature.
No knowledge of the molecular genetics, biochemistry, or physiological causes
of adaptation is required to interpret these results, although such knowledge is
highly desirable. The fact that the derived populations could no longer grow in
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
210 KASSEN
RAINEY
the ancestral environment demonstrates that the changes that occurred during the
course of the selection experiment were genetically determined. Had Dallinger’s
experiment not been destroyed, he might well have proceeded to determine the
magnitude of improvement at elevated temperatures relative to the ancestral type,
tested for and directly measured tradeoffs in ecological performance at different
temperatures, and even documented diversity within evolving populations (its sta-
bility and changes in patterns of abundance through the course of the experiment).
The remarkable fact is that with relatively little additional development, the ap-
proach Dallinger took more than a century ago remains a valid and powerful means
of studying some of the most fundamental problems in biology. Moreover, such
analyses can provide insight into the mechanism of ecological and evolutionary
change—with the same kind of power and precision that a molecular microbiolo-
gist expects when tackling, for example, a problem in gene regulation.
The experiment also highlights a second issue of importance for microbiologists
interested in understanding the causes of adaptation and diversity, namely that
evolution can be studied on timescales that are amenable to experimentation and
analysis. This should come as no surprise, as the rapid generation times and large
population sizes of microbial populations are ideal conditions for natural selection.
Interestingly, Darwin himself missed the opportunity to test experimentally his
theory of adaptation by natural selection: In a letter to Dallinger dated July 2, 1878,
Charles Darwin wrote: “I did not know that you were attending to the mutation
of lower organisms under changed conditions of life; and your results, I have no
doubt, will be extremely curious and valuable. The fact which you mention about
their being adapted to certain temperatures, but becoming gradually accustomed
to much higher ones, is very remarkable. It explains the existence of algae in hot
springs. How extremely interesting an examination under high powers on the spot,
of the mud of such springs would be” (30, pp. 191–192).
TECHNIQUES AND METHODS
The techniques for studying adaptation and diversity in the laboratory have not
changed much from Dallinger’s time. They are conceptually straightforward and
borrowed directly from quantitative genetics, agronomy, and animal breeding (4,
37, 41, 43, 64, 79, 104).
Briefly, a population is allowed to grow and reproduce in an environment that
is defined by the experimenter. Most commonly, these environments are batch
cultures of simple media (82, 90, 95), although continuous cultures (33, 97, 110)
and plate cultures (65, 71, 76) are sometimes used. The duration of the experiment
depends on the theory being tested. For ecological experiments, it is common to
start with a highly diverse population (or collection of species) and follow the fate
of diversity over the course of a few tens of generations (94, 116). Evolutionary
experiments typically begin with a single strain and last for hundreds or thousands
of generations, new variation arising during the course of the experiment through
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 211
de novo mutations. Variation that arises through mutation or pre-exists in the
population generates competition among genotypes as the population grows and
uses up the available resources. The type that is fittest, in the sense of being the
best competitor under the prevailing conditions of growth, comes to dominate
the population through natural selection. This leads to a change in the genetic
makeup of the population and the steady replacement of one genotype by another.
The evolved populations therefore out-compete the ancestor in the environment in
which they have been selected.
Fitness itself can be measured in either of two ways. The first estimates growth
parameters such as r, the intrinsic rate of growth, and K, the carrying capacity,
in pure culture. The second involves competing each genotype against a common
ancestor and measuring the rate of competitive exclusion (37, 82). The latter is
preferable because it is an integrated measure that takes into account adaptation to
the abiotic and biotic conditions of growth. To estimate the quantity of diversity
in a population, it is necessary to isolate individual genotypes from the population
and estimate the variance in fitness among them.
THE STRUCTURE OF THE ENVIRONMENT
AND THE MAINTENANCE OF DIVERSITY
Ecologists and population geneticists have long-suspected that the structure of the
environment is connected to the maintenance of diversity (64). It is easy to see why
this might be so. Natural selection should eliminate all but the fittest type under
any given set of growth conditions, leading to a loss of diversity. Reasonably, then,
if the environment were not uniform but instead a series of distinct niches (or
patches) that differed in their conditions of growth, then different types may be
favored in each niche, and so diversity maintained according to the niche exclusion
principle.
The variety of available niches, and therefore the variety of types supported,
may be determined by the physical structure of the environment, for example, by
the patchy distribution of essential resources. The variation in fitness of genotypes
in response to the abiotic environment is known in population genetics as genotype-
by-environment interaction, or GxE (66, 129). Alternatively, new niches may arise
through the actions of organisms themselves. In such instances diversity is main-
tained through genotype-by-genotype interactions, or GxG (4). As we shall see,
experimental studies with microbial populations in the laboratory provide support
for both kinds of interactions.
Simple Environments Composed of a Single Niche
Imagine a single genotype of an asexual microorganism inoculated into a simple
environment composed of a single niche to which it is initially not well adapted. As
the population grows, mutations arise randomly with respect to fitness and lead to
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
212 KASSEN
RAINEY
genetic diversification. While the vast majority of new mutations are deleterious,
a small fraction will be beneficial, in the sense that they confer a higher rate
of replication. This growth rate advantage may manifest in different ways. For
example, some variants may have a shorter lag time, higher intrinsic growth rates
owing to more efficient metabolism, or higher yields for a given unit of resource.
Regardless, some of these more fit genotypes will come to dominate the population
(for a discussion of constraints on the fixation of beneficial mutations in asexual
populations, see References 32 and 48). This is the process of periodic selection,
first documented more than half a century ago (2, 3, 97), whereby beneficial
mutations are substituted through the population, supplying a transient source of
genetic variation that is eventually exhausted as the population nears evolutionary
equilibrium.
It is no trivial matter to study periodic selection in microbial populations, as
it requires identifying beneficial mutants when they first arise at low frequency
in a population and following their fate through time. Atwood et al. (2) achieved
this by monitoring fluctuations in the frequencies of his+ prototrophs relative to
numerically dominant his- auxotrophs in evolving Escherichia coli populations.
The authors observed a saw-tooth trajectory of increases and sudden decreases
in the frequency of the minor his+ type, and reasoned that the increases were
due to recurrent mutation and the decreases were caused by the purging effect
associated with the “sweeping” of successive beneficial mutations that arose in the
numerically dominant his- population.
Direct support for the operation of periodic selection and concomitant diversity-
purging effects come from two recent molecular-level analyses. The first, by
Notley-McRobb & Ferenci (96), repeats and extends the classic work of Atwood
et al. by tracking the frequency of T5-resistant cells (the equivalent of the his+
marked population used by Atwood et al.) during adaptation to a glucose-limited
chemostat. They observed the expected saw-tooth pattern and also showed that
decreases in the number of T5-resistant cells corresponded exactly to the sweep
of beneficial mutants (Figure 1). Interestingly, they also noted the maintenance of
diverse alleles at the selected locus following a sweep, suggesting the existence of
different genetic routes conferring similar increases in fitness in these populations.
Going a step further, Wichman et al. (133) sequenced multiple genomes of
the DNA bacteriophage φX174 after adaptation to a novel host at high tempera-
tures. This allowed the identification of genetic changes and provided a means of
tracing the frequency of each change during the course of the experiment. Wich-
man et al. detected more than a dozen nucleotide substitutions and found strong
evidence, on the basis of parallel evolution of the same nucleotide changes and
the rapid substitution of unique changes, that most of these substitutions were
adaptive. The pattern of substitution indicated a series of rapid sweeps occurring
one after the other in a manner consistent with the model of periodic selection
described above (Figure 2). Of particular note is that in each population at the end
of the experiment, just a single genotype was present. Taken together, these exper-
iments are entirely consistent with the niche exclusion principle and suggest that,
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 213
Figure 1 Periodic selection in populations of E. coli evolving in a simple, uniform
environment. Note the saw-tooth pattern (solid line)inthe frequency of the rare T5-
resistant mutant. Decreases in T5 resistance are correlative with the selective sweep
of the beneficial mgl-con mutant (dashed line). Reproduced with permission from
Notley-McRobb and Ferenci (96).
as a general rule, simple, uniform environments cannot maintain genetic variation
in fitness in the long term (104).
In apparent contradiction to the niche exclusion principle, however, a number
of studies have detected substantial phenotypic and molecular variation in popu-
lations evolving in what should be single-niche environments. For example, DNA
fingerprinting of insertion sequence (IS) elements in two populations of E. coli
propagated for over 10,000 generations in glucose-limited minimal medium (79)
revealed that 11 of 11 clones in one population and 10 of 13 clones in another had
unique IS fingerprints (102). Clearly, diversity existed in these populations, but
whether this molecular variation reflected ecologically relevant genetic variation
in fitness needed to be determined.
To address this issue, Elena & Lenski (40) estimated the genetic variation in
fitness for six replicate E. coli populations at generation 10,000 by competing 25
independently isolated clones from each population against their common ancestor.
Statistically significant genetic variation in fitness was detected, with two random
clones from the same population differing on average by ∼4%. Although this
difference was relatively minor in comparison to the roughly 50% increase in
fitness relative to the ancestor since the start of the selection experiment (it also
varied among replicate populations), it was higher than would be expected through
the operation of periodic selection alone in a single-niche environment.
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
214 KASSEN
RAINEY
At least three mechanisms could potentially support such low but significant
levels of genetic variation in fitness in these populations. First, the populations
may have still been adapting to the conditions of growth, albeit at a reduced
rate relative to the start of the experiment. Any genetic variation present in the
population may, then, reflect the ongoing substitution of beneficial mutations, as in
the φX174 experiment above. None of the populations, however, showed evidence
for increases in mean fitness between generation 10,000 and generation 10,500,
as would be expected if beneficial mutations were being substituted. Second, the
populations may have reached an evolutionary equilibrium with the majority of new
mutations being deleterious and kept at low frequency by natural selection (52). The
fortuitous evolution of two “mutator” populations with high mutation rates caused
by defective methyl-directed mismatch repair (115) permitted this idea to be tested
directly. The prediction that the mutator populations would retain higher levels of
genetic variation in fitness at mutation-selection balance was not supported. Finally,
diversity may have been maintained by negative-frequency-dependent selection,
the fitness of a genotype being highest when rare (because resources are most
abundant) but not when common (resources are rare and competition intense)
(92). Invasion experiments of each clone against the population from which it
came revealed a small but statistically significant fitness advantage when rare,
consistent with the operation of negative-frequency-dependent selection through
cross-feeding interactions (111). These results suggest that, given sufficient time,
selection may lead to the evolution of new kinds of interactions among genotypes
even in the simplest of environments, a phenomenon discussed in more detail
below.
Complex Environments Composed of Many Niches
A heterogeneous environment is one composed of many niches that may vary in
either space or time, and at different scales relative to the generation time of the
organism concerned (64). The connection between environmental heterogeneity
and diversity has a long history in ecology and population genetics (85, 86, 109),
although direct tests of theory have been comparatively recent (64).
Evolutionary theory predicts that in a spatially heterogeneous environment, se-
lection favors the emergence of ecological specialists, different types being adapted
to different niches. Ecological specialists by definition trade off enhanced com-
petitive advantage in one niche against reduced competitive ability in another. The
existence of trade-offs prevents competitive exclusion by one type and renders
coexistence of organisms possible (66, 104). The evolution of ecological special-
ization can be readily achieved in the laboratory by allowing a single founding
population to grow and reproduce for many generations in contrasting environ-
ments. Bell and Reboud (6, 108), for example, founded two populations from
a single genotype of the unicellular chlorophyte Chlamydomonas reinhardtii that
were serially propagated for approximately 1000 generations either as phototrophs
in the light or as heterotrophs in the dark in an environment supplemented with
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 215
Figure 3 The evolution of ecological specialists through selection in different en-
vironments. Each pair of points connected by a line represents the fitness of a single
genotype of Chlamydomonas in the light (x-axis) and in the dark (y-axis) that has been
selected in the light (open circles) and in the dark (filled circles) for approximately
1000 generations. Note that specialization to one environment is accompanied by a
fitness trade-off in the other environment. Reproduced with permission from Bell &
Reboud (6).
organic carbon. The original genotype grew well in the light and relatively poorly
in the dark. By the end of the experiment, the light lines showed little improvement
in the light and lost the ability to grow in the dark, while the dark lines grew much
better than the ancestor in the dark and relatively poorly in the light compared
with the light lines (Figure 3). Trade-offs of this sort are commonly observed in
laboratory experiments (39, 64). Indeed these sorts of fitness trade-offs across en-
vironments form the basis for the widespread use of vaccines, which have been
adapted to laboratory conditions and therefore have lost the ability to cause disease
in humans (39).
The evolution of ecological specialization and fitness trade-offs across environ-
ments is a necessary but not sufficient condition for the maintenance of diversity
in a heterogeneous environment. This point is made by a second experiment by
Reboud & Bell (108) in which the light and dark specialists were mixed together
and then propagated for a further 200 generations in either spatially or tempo-
rally varying environments. Spatial variation was simulated by growing the mixed
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
216 KASSEN
RAINEY
population in the light and in the dark simultaneously, collecting samples from
each, combining them together, and then redistributing them to light and dark
flasks again. Temporal variation involved growing the original mixture first in one
environment and then transferring a sample to the other environment for the next
growth cycle. Theory predicts that spatial variation would be much more effec-
tive than temporal variation at maintaining diversity because it offers refuges for a
rare type well-adapted to one of the environmental patches. In contrast, temporally
varying environments offer no such refuge—the type favored being one that grows
best across all conditions. The results bore out this prediction: Populations exposed
to the spatially varying regime retained the mixture of light and dark specialists
with which they were founded, whereas generalist types, capable of growing well
under both light and dark conditions, evolved when the environment varied tempo-
rally. Thus the pattern of environmental variation—in time or in space—is central
to the maintenance of diversity.
The combination of ecological specialization and spatial heterogeneity main-
tains diversity when otherwise it would be lost. Diversity in Reboud and Bell’s
experiment, for example, was rapidly lost when the mixed population was prop-
agated solely in either the light or the dark, a result that has been observed many
times previously (64). It is not clear from their experiment whether diversity would
be ultimately lost in the spatially heterogeneous environment as well. In theory this
need not happen. Models of selection in heterogeneous environments suggest that,
under the right conditions, diversity can be permanently and stably maintained
through negative-frequency-dependent selection (83, 93), the fitness of a genotype
being higher when rare than when common. For this to happen, fitness trade-offs
must exist, as before, and the population size must be regulated at the level of the
local patch.
Evidence of trade-offs and the concomitant operation of negative-frequency-
dependent selection has been detected and measured in laboratory populations of
various microbes (8, 64, 80, 104). Rainey & Travisano (106) documented the emer-
gence of diversity in spatially heterogeneous environments. Genetically identical
founding populations of Pseudomonas fluorescens were propagated in heteroge-
neous (spatially structured) environments, afforded by incubating broth-containing
microcosms without shaking, and homogeneous (spatially unstructured) environ-
ments, created by incubating identical microcosms under a continuously shaken
regime. Diversity, in the form of niche-specialist genotypes, emerged rapidly in
spatially heterogeneous microcosms, but not in spatially homogeneous micro-
cosms. Taking ecologically diverse populations from spatially structured micro-
cosms and propagating them under the spatially heterogeneous regime or the spa-
tially homogeneous regime provided a test of the role of spatial heterogeneity in
supporting diversity. Diversity was rapidly lost once heterogeneity was reduced,
supporting the idea that heterogeneity itself is a primary cause of diversity and
necessary for the continued maintenance of diversity.
In the P. fluorescens experiments, niche specialization and trade-offs are read-
ily observable by eye (106); however, additional experiments were performed to
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 217
test the operation of negative-frequency-dependent selection—the mechanism ex-
pected to maintain diversity given the existence of obvious trade-offs. A series of
pair-wise “invasion from rare” experiments showed that in most instances niche-
specialist genotypes could increase in frequency relative to a common type, even
though initially present at just 1/100 the density of the common type. The diversity
is therefore protected and coexistence of genotypes is assured.
It is difficult to say with certainty whether conditions for the maintenance of
diversity in laboratory populations also operate in the field, especially for microbial
populations of which little is known about the dispersal rate and the degree of local
adaptation. Clearly, most natural environments are often heterogeneous, although
whether this heterogeneity occurs on the appropriate spatial scale for selection
to maintain diversity remains to be seen. Some indication of the scale at which
selection acts has been obtained by Belotte et al. (7), who isolated single colonies
from soil samples collected from a 10 × 10 m plot in an old-growth forest in
southern Quebec, Canada. The fitness of isolates was typically greatest when grown
on soil-water extracts made from the same soil from which they had been isolated,
and fitness was lowest when grown on soil extracts made from other samples,
suggesting that the scale of local adaptation in these bacterial communities is on
the order of 1 to 10 m. This and other recent studies suggest that spatial structure
of the form that could stably maintain diversity is a common feature of microbial
communities (23, 55, 118, 130, 131).
Complex Environments Created by the Growth
of Competitors
It is reasonably easy to envisage GxE interaction for fitness arising through se-
lection in response to different chemical resources or spatially distinct niches (84,
121). However, it is also possible for new sorts of interactions to arise as the re-
sult of the growth of genotypes themselves, even in the absence of physically or
chemically distinct niches. The growth of one genotype creates a new niche for
another genotype, and so diversity can be understood in much the same way as
before, in which the number of niches determines the number of types supported
in the community.
An example of such GxG interactions comes from the work of Adams and
colleagues (56), who performed long-term selection experiments with chemostat-
propagated populations of E. coli grown on a single limiting carbon source. Reg-
ular sampling from the chemostat populations revealed phenotypically distinct
colony types on agar plates, indicating that the once genetically uniform pop-
ulations had become polymorphic. Analysis of competitive interactions among
genotypes showed that the polymorphism was stable and maintained by density-
dependent processes. This surprising finding appeared at first to contradict the
niche exclusion principle and prompted further physiological and genetic studies.
These studies found evidence of resource partitioning of the limiting glucose re-
source by the numerically dominant genotype (110). Partitioning of glucose into
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
218 KASSEN
RAINEY
acetate and glycerol, metabolic byproducts arising from metabolism of glucose,
provided ecological opportunity for the evolutionary emergence of mutants with
enhanced capacities to metabolize these byproducts. Indeed, the variant types that
arose (and arose repeatably; 128) had evolved, in one instance, an enhanced ca-
pacity to metabolize acetate and, in the other, an improved capacity to recover
glycerol.
Diversity can also be maintained through GxG interactions that occur over small
spatial scales and where strict competitive hierarchies are absent (29). Experimen-
tal studies of colicin-producing, colicin-resistant, and colicin-sensitive derivatives
of E. coli (71) show that in an environment allowing for localized interactions,
such as the surface of an agar plate, all three genotypes can be maintained. How-
ever, in an environment where opportunity for localized interactions is negligi-
ble, such as in a shaken (spatially homogeneous) broth culture, diversity is lost.
Similar dynamics are likely to be widespread in microbial communities, such
as biofilms (49, 50, 124), where opportunities exist for local interactions among
individuals.
One important distinction between GxG interactions and GxE interactions is
that the former can lead, at least in principle, to coevolution. Such interactions
have the potential to increase or decrease diversity (18, 20, 22), with the outcome
being dynamic and difficult to predict. The clearest examples of coevolution be-
tween microbial genotypes come from studies of predator-prey and host-parasite
interactions (13, 14, 34, 59). In principle, GxG interactions may lead to coevo-
lution between competitors as well, but, to our knowledge, this has not yet been
documented in experimental populations of microbes.
Complex Environments Created Through
the Growth of Predators
Our discussion has so far focused on explanations for diversity in simple commu-
nities composed of different genotypes or species competing for similar resources.
Natural communities may be much more complex, however, with many trophic
levels. These more complex communities can be understood as a special case
of GxG interactions, in which the ecological interactions between predator and
prey are severely asymmetric: positive for the predator and negative for the prey.
This represents a strong selective pressure on the prey population to evolve means
of escaping predation. Several experiments with bacterial prey and their phage
predators have documented the spontaneous evolution of phage-resistant bacteria
from a population of sensitive types (9, 21, 81). Interestingly, both sensitive and
resistant prey genotypes can coexist in the presence of phage if there is a trade-off
between growth rate and susceptibility to predation: Resistant genotypes avoid
predation but grow slower than sensitive genotypes in the absence of phage. The
magnitude of this trade-off can have profound impacts on the relative frequencies
of the different types in the community (9). Extrinsic features of the environment
can also determine the variety of trophic interactions in a community. These issues
are explored in more detail by Jessup et al. (63).
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 219
Patterns of Diversity in a Collection of Communities
Diversity in metazoan communities often follows regular patterns across environ-
mental gradients in nature (109), and there seems little reason to expect that the
same might not be true in microbial communities. In a recent review, Horner-
Devine et al. (58) provide evidence to support this contention. They show that
molecular variation in bacterial communities often changes in regular ways across
different kinds of habitats, within habitats that differ in their structural diversity,
and across gradients of primary productivity and disturbance. These results need
to be interpreted with caution because they are often based on molecular variation
that may not always accurately reflect the quantity of genetic variation in fitness.
Still, the fact that some of these patterns bear a striking resemblance to those
observed in metazoans suggests that similar processes may be at work.
Our own work has explored two of these patterns: those caused by alterations in
productivity (67) and disturbance (12). Along gradients of both factors, species di-
versity is typically humped-shaped, with a peak at intermediate rates of production
and disturbance. Traditionally, ecologists had thought that different mechanisms
were responsible for the two patterns. By examining diversity of P. fluorescens pop-
ulations cultured along gradients of both productivity and disturbance, it appears
that both patterns may be underlain by the same mechanism, namely competition
among niche specialists in a heterogeneous environment. The key features deter-
mining diversity in this system are diversifying selection—different niche special-
ists favored in different niches—and the relative fraction of individuals contributed
by each niche to the total population. Diversity can only be maintained if each niche
contributes similar numbers of individuals to the total population; if one niche pro-
duces many more individuals than the other, diversity cannot be maintained. We
have shown, furthermore, that the same mechanism may govern the level of di-
versity achieved during adaptive radiation, in which a single lineage diversifies to
form a range of niche specialists (68), suggesting that ecological factors can act as
major constraints on diversification over evolutionary timescales.
The Pseudomonas experiments lend support to the idea that environmental
structure is important in determining the level of diversity achieved not only within
a community, but also across communities in a landscape. These patterns may be
further modulated by a variety of factors such as parasites (10, 13), predators (9,
69), migration among populations in a landscape (27), and the contingent nature
of the interactions among types underlying the assembly of communities (46). The
extent to which these factors determine patterns of diversity in natural microbial
communities remains an empirical issue.
THE GENETIC CAUSES OF DIVERSITY
IN AN ECOLOGICAL CONTEXT
The ultimate source of all genetic variation is, of course, mutation. New genetic
variants may also be introduced into the community through migration or created
by recombination. These mechanisms serve to increase the amount of heritable
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
220 KASSEN
RAINEY
variation from which natural selection sorts under different ecological conditions
and so can be readily understood within the framework outlined above. This ap-
proach is not very useful, however, for understanding in detail the genetic mech-
anisms underlying the origin of diversity or the impact these mechanisms have
on the range of variation available to natural selection within a given ecological
context. We turn to these problems in this section.
Genetic Mechanisms Underlying Diversification
Through Adaptive Radiation
The vast majority of phenotypic and ecological diversity on the planet has arisen
during successive adaptive radiations, that is, periods in which a single lineage
rapidly diversifies to generate multiple niche-specialist types (35, 77, 113, 114).
Microbiologists tend not to think of microorganisms as undergoing adaptive ra-
diation, but there is no reason to exclude them from this general statement—in
fact the rapid generation times and large population sizes characteristic of many
microbial populations suggest that microorganisms may be particularly prone to
bouts of adaptive radiation (6, 90, 106, 127).
Diversification during adaptive radiation requires that a lineage first gain access
to a niche in which it has a selective advantage over the ancestral type. The pheno-
typic changes favored by selection during these early stages of adaptive evolution
typically involve increases in levels of expression of enzymes that are initially
limiting to growth in the novel environment. These higher levels of expression
allow the population size to increase, which then sets the stage for subsequent di-
vergence through modification and refinement of the existing genome. Uncovering
the genetic changes responsible for diversification thus represents a special case
of the more general problem of the genetics of adaptive evolution and phenotypic
innovation (19, 41, 78, 103, 120). Here we briefly mention two mechanisms re-
sponsible for the initial invasion of a novel niche, gene duplication and changes in
gene regulation.
Sonti & Roth (117) have documented the selection of stable duplications in
the araE locus of Salmonella typhimurium,agene coding for a permease used to
import arabinose, sorbitol, and malate. Under standard conditions, duplications in
araE occur at rates of about 10
−3
to 10
−4
,but they were elevated approximately
2000-fold when cultures were grown under arabinose limitation on chemostats
for roughly 200 generations. Similar results were obtained when cultures were
selected under sorbitol and malate limitation as well. These experiments show that
duplications, particularly in genes involved in nutrient acquisition, are favored
under selective growth (1, 57).
Once duplicate genes are fixed in a population, one copy is free to accumulate
mutations while the other retains its original function. This permits the lineage
possessing the duplication to take advantage of new and unexploited resources
while not forgoing the ability to grow on the original substrate. This may not
always be the case; one of the duplicated copies may be lost, or mutations may
occur in both thus destroying complementary functions by a process known as
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 221
subfunctionalization (53, 89). Nevertheless, gene duplication has been implicated
as a pervasive mechanism to explain the evolution of novel metabolic capabili-
ties (87), the redundancy of eukaryotic genomes (51), and the evolution of novel
morphological features in plants and animals (103).
The second mechanism facilitating access to novel environments has its roots in
developmental genetics, in which molecular analyses have provided compelling
evidence that phenotypic innovations often arise through changes to regulatory
rather than structural genes (120). This is because it is easier for mutations to
generate variation in the quantity of enzyme production, which is often controlled
by regulatory genes, rather than in the specificity of those enzymes themselves.
Such mutations are also less likely to have deleterious pleiotropic effects (120).
An example of this comes from studies of the evolution of novel substrate
use in bacteria. Lin et al. (87) review a series of experiments stemming from the
pioneering studies of Wu and Mortlock (cited in Reference 87) on experimental
evolution of xylitol metabolism by Klebsiella aerogenes; xylitol being a substrate
not normally found in nature and on which wild-type strains cannot grow. Selection
in liquid medium containing xylitol as a sole carbon and energy source generated
a series of mutants (X1–X3) with increasing growth rates. Biochemical analysis
of the first mutant to appear, X1, showed that it constitutively expressed ribose
dehydrogenase (RDH), which is normally an inducible enzyme that hydrolyzes
ribitol into ribulose. RDH from X2, a second mutant derived from X1, is not
only constitutively expressed on xylitol but shows structural alterations to the
enzyme that increase its affinity for xylitol. Thus the evolution of novelty in this
system involves the deregulation of a pre-existing enzyme and the subsequent
modification of its structural properties. Changes in levels of enzyme expression are
also responsible for the evolution of acetate cross-feeding mutants of E. coli (128)
and niche specialization in the model P. fluorescens adaptive radiation (105, 119).
Fitness Effects of Mutations
The process of diversification characteristic of adaptive radiation involves adap-
tation to novel conditions of growth, which itself depends on the supply of new,
beneficial mutations. Mutations arise constantly by a variety of mechanisms (92),
but most of these are deleterious and never achieve high frequency in a popula-
tion. It is the much smaller class of beneficial mutations that provides the fuel for
adaptive evolution.
One approach to studying beneficial mutations is to characterize them through
a combination of laboratory selection and genetic analysis (16, 17, 25, 41). Bull
et al. (16) identified the genetic changes in the bacteriophage φX174 that permit
growth at high temperature after a single round of selection. Two of the three
major classes of mutation occurred at different locations within the capsid protein,
and the third concerned a gene expressing a protein involved in the development
of the mature virion. The precise biochemical and physiological effects of these
mutations are not known, but they seem to be involved in maintaining the stability
of the immature and mature virus particle at high temperature (36).
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
222 KASSEN
RAINEY
That these mutations were the ones responsible for adaptation was confirmed by
estimating the fitness of the different genotypes relative to the ancestor from which
they were derived. All three mutants conferred large fitness advantages relative to
the ancestor at the high temperature. However, adaptation to this novel environment
came at a cost: The mutants had reduced fitness (relative to the ancestor) in the low
temperature ancestral environment. This cost (in which a mutation is beneficial in
one environment but deleterious in another) is known as antagonistic pleiotropy and
forms one of two potential costs of adaptation—the other arising from mutations
that are neutral in the environment of selection but deleterious elsewhere (26, 91).
Finally, the beneficial effect of the mutations depended on the genetic background
in which they occurred. The original mutations were recovered from strains that
were not well adapted to growth in E. coli. When the mutations were transferred
into a genetic background that was well adapted to E. coli, the gains in fitness
were substantially smaller. Thus the fitness of a given mutation is contingent on
the genetic background in which it occurs.
A second approach to studying beneficial mutations is to focus on the statistical
properties of the variation among the mutations without regard to their mechanistic
bases. Fisher’s geometric model of adaptive evolution predicts that the distribu-
tion of fitness effects among beneficial mutations, both new and those ultimately
fixed by selection, follows a negative exponential, with many mutations of small
effect and few of large effect (45, 99). Imagine a population located at some dis-
tance from a fitness optimum, represented in Figure 4 by the intersection of the
Figure 4 A schematic of Fisher’s geometric model of adap-
tation. See text for a description.
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 223
x- and y-axes. Mutations of variable effect on the phenotype arise at random and
are represented by vectors of any direction and magnitude in space that have
the location of the original population as their origin. Only those mutations that
bring the population closer to the fitness optimum are favored by natural selec-
tion. Large mutations stand a greater chance of taking the population farther away
from the fitness optimum, or even overshooting it, and therefore are unlikely to
be viable. Small mutations, however, have a higher probability of bringing the
population closer to the optimum and therefore are more likely to contribute to
adaptation.
To study experimentally the distribution of fitness effects among beneficial
mutations, Imhof & Schlotterer (62) followed the evolution of 10 replicate E. coli
populations for 1000 generations starting from a single ancestral strain containing
a highly mutable microsatellite locus from Arabidopsis thaliana contained on a
plasmid. Population expansion generated a variety of microsatellite alleles that
were neutral with respect to the fitness of E. coli cells. Selective sweeps caused
by beneficial mutations were detected by monitoring the frequency of specific
microsatellite alleles to which particular beneficial mutations were genetically
linked. Isolation of the strains bearing the putative beneficial mutations and analysis
of their fitness relative to the population from which they were isolated (as opposed
to the usual practice of competing them against the unevolved ancestor) provided
an estimate of the selective advantage of each newly risen mutant. Taken together,
the distribution of fitnesses among these beneficial mutations was not statistically
different from a negative exponential distribution, as expected from theory.
CONCLUDING REMARKS: LIFE IN THE TANGLED BANK
It is interesting to contemplate a tangled bank, clothed with many plants of many
kinds, with birds singing on the bushes, with various insects flitting about, and
with worms crawling through the damp earth, and to reflect that these elaborately
constructed forms, so different from each other, and dependent upon each other in
so complex a manner, have all been produced by laws acting around us.
Charles Darwin (31, p. 429)
Microbial communities are highly diverse, a fact that demands both description
and explanation. Much progress has been made over the past 15 to 20 years on
the former; the experiments reviewed here take us some way toward the latter.
From a purely ecological perspective, diversity is supported most readily in the
presence of divergent natural selection: Different types are favored in different
niches. The variety of niches available depends largely on the physical structure of
the environment. Complex, heterogeneous environments provide more niches—
and therefore are more likely to maintain diversity—than simpler environments
with fewer niches. But the abiotic environment is not the sole determinant of niche
space: The biotic environment is also relevant, with new ecological opportunities
being continually created through the growth and activities of organisms.
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
224 KASSEN
RAINEY
There does not seem to be any serious limit to the ability of mutation to gen-
erate new variation upon which selection can act. Diversification in the face of
ecological opportunity occurs through at least two mechanisms, gene duplication
and deregulation of gene expression. Subsequent adaptation to novel environments
can be interpreted both in mechanistic terms, as a consequence of the biochemistry
and physiology of gene action, and in purely a statistical manner, in terms of the
distribution of mutational effects.
That being said, some qualifying statements are necessary. Perhaps the most
important is that we have ignored (quite deliberately) an important alternative
explanation for diversity, namely that much of the variation observed in nature is
neutral with respect to fitness, its fate determined by stochastic processes sorting
among ecologically equivalent types. The neutral theory was originally developed
to explain the greater than expected molecular variation revealed by studies of
protein and DNA polymorphism (73), but it has recently been extended to cover
species diversity in ecological communities (5, 60). We suspect that the unified
neutral theory of biodiversity and biogeography (60) may have particular relevance
to diversity in microbial communities. Note, however, that it cannot by definition
explain the existence of genetic variation in fitness—the focus of this article. In
this regard we need better estimates of the quantity of genetic variation in fitness
that exists in natural communities (44).
A number of lines of evidence suggest that such variation in fitness exists in na-
ture: The mere fact that some microbial species can be cultured and others cannot
points to the existence of genetic variance in fitness for the conditions of culture.
Schmidt and colleagues (74) provided evidence that ribosomal copy number influ-
ences the structure of microbial populations in soil. In addition, the composition
of microbial communities typically changes along environmental gradients, such
as temperature and depth in hot springs (130), different soil and moisture types
(11, 112) and geographical locations (55, 131), length of the gastrointestinal tract
(15), and space and time within oral cavities (75), suggesting that selection in these
different environments is often strong enough to lead to ecological specialization.
Direct, experimental evidence demonstrating adaptation to these different envi-
ronments, and the scale at which this occurs, can be obtained through reciprocal
transplant or explant experiments (7). Evidence can also be obtained by careful
manipulation and monitoring of the response of natural microbial populations to
perturbations (74).
By way of further qualification, we recognize that we have not considered the
effects of dispersal and recombination (gene flow) on the maintenance of diversity
in microbial populations. These are of particular relevance, as both factors limit
the ability of selection to maintain diversity in heterogeneous environments and
both dispersal and recombination are undoubtedly significant factors in microbial
communities. The trouble is we have little idea over what scales these operate or the
full magnitude of their effects (24). Nevertheless, there is scope for incorporating
these factors into future experiments.
Despite these caveats there can be little doubt that microbes, on account of their
rapid generation times and large population sizes, have a remarkable capacity to
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 225
evolve and diversify through natural selection. This demands that we pay attention
to the evolutionary context of microbes and the populations and communities of
which they are part. Microbial communities are evolving entities whose compo-
nents are genotypes and species and whose driving mechanisms are the interactions
among them. These interactions may take on any form, from antagonistic to ben-
eficial, and can themselves evolve, sometimes fortuitously and unexpectedly. For
example, consider the relationship between bacterial pathogens and disease. Single
cells rarely ever cause disease—it is a property of populations of cells and more
usually communities. Disease symptoms typically manifest once pathogen popu-
lations reach certain thresholds, and the likelihood that these are reached depends
on the nature of the interactions (antagonistic or synergistic) among individuals
and populations that comprise the community. Moreover, the composition of the
community, and therefore the severity of disease, can be modified by ecological
factors, such as nutrient supply and biomass disturbance, that in turn are influenced
by prevailing therapeutic regimes. Of course the severity of disease is also influ-
enced by evolution within the pathogen population occurring within the ecological
context defined by the host itself.
That we become more conversant with the central concepts of ecology and
evolution and begin to place our detailed molecular understanding of the lives
of microbes within this context is crucial. The techniques required to do this are
essentially the same as those introduced by Dallinger well over a hundred years ago
and require only that we pay closer attention to the interactions among genotypes
and species and their consequences at the population and community levels. We
suspect that when more is known about the variety of these interactions and the
complexity of the habitats in which they occur in nature, the resulting picture of
life in microbial communities will bear a striking resemblance to the one Darwin
gave of his tangled bank.
ACKNOWLEDGMENTS
The authors thank past and present members of the laboratory of experimental
evolution for stimulating discussion. R.K. thanks St. Hugh’s College, Oxford and
NSERC (Canada). P.B.R. is grateful to the BBSRC (United Kingdom) and NERC
(United Kingdom) for support.
The Annual Review of Microbiology is online at http://micro.annualreviews.org
LITERATURE CITED
1. Andersson DI, Slechta ES, Roth JR. 1998.
Evidence that gene amplification under-
lies adaptive mutability of the bacterial lac
operon. Science 282:1133–35
2. Atwood KC, Schneider LK, Ryan FJ.
1951. Periodic selection in Escherichia
coli. Proc. Natl. Acad. Sci. USA 37:146–
55
3. Atwood KC, Schneider LK, Ryan FJ.
1951. Selective mechanisms in bacteria.
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
226 KASSEN
RAINEY
Cold Spring Harbor Symp. Quant. Biol.
16:345–54
4. Bell G. 1997. Selection: The Mechanism
of Evolution.New York: Chapman & Hall
5. Bell G. 2001. Neutral macroecology. Sci-
ence 293:2413–18
6. Bell G, Reboud X. 1997. Experimental
evolution in Chlamydomonas. II. Genetic
variation in strongly contrasted environ-
ments. Heredity 78:498–506
7. Belotte D, Curien JB, Maclean RC, Bell
G. 2003. An experimental test of lo-
cal adaptation in soil bacteria. Evolution
57:27–36
8. Bohannan BJM, Kerr B, Jessup CM,
Hughes JB, Sandvik G. 2002. Trade-offs
and coexistence in microbial microcosms.
Antonie Van Leeuwenhoek 81:107–15
9. Bohannan BM, Lenski RE. 2000. Link-
ing genetic change to community evolu-
tion: insights from studies of bacteria and
bacteriophage. Ecol. Lett. 3:362–77
10. Brockhurst MA, Rainey PB, Buckling A.
2004. The effect of spatial heterogeneity
and parasites on the evolution of host di-
versity. Proc. R. Soc. London Sci. Ser. B
271:107–11
11. Buckley DH, Schmidt TM. 2001. Envi-
ronmental factors influencing the distri-
bution of rRNA from Verrucomicrobia in
soil. FEMS Microbiol. Ecol. 35:105–12
12. Buckling A, Kassen R, Bell G, Rainey PB.
2000. Disturbance and diversity in exper-
imental microcosms. Nature 408:961–64
13. Buckling A, Rainey PB. 2002. Antagonis-
tic coevolution between a bacterium and a
bacteriophage. Proc. R. Soc. London Sci.
Ser. B 269:931–36
14. Buckling A, Rainey PB. 2002. The role of
parasites in sympatric and allopatric host
diversification. Nature 420:496–99
15. Buddington RK, Weiher E. 1999. The
application of ecological principles and
fermentable fibers to manage the gas-
trointestinal tract ecosystem. J. Nutr. 129:
1446–50
16. Bull JJ, Badgett MR, Wichman HA.
2000. Big-benefit mutations in a bacterio-
phage inhibited with heat. Mol. Biol. Evol.
17:942–50
17. Burch CL, Chao L. 1999. Evolution by
small steps and rugged landscapes in the
RNA virus (phi) 6. Genetics 151:921–27
18. Caldarelli G, Higgs PG, McKane AJ.
1998. Modelling coevolution in multi-
species communities. J. Theor. Biol. 193:
345–58
19. Carroll SB. 2000. Endless forms: the evo-
lution of gene regulation and morpholog-
ical diversity. Cell 101:577–80
20. Case TJ, Taper ML. 1994. Coevolution
among competitors. In Oxford Surveys in
Evolutionary Biology, ed. DJ Futuyma,
J Antonovics, pp. 63–109. Oxford: Ox-
ford Univ. Press
21. Chao L, Levin BR, Stewart FM. 1977. A
complex community in a simple habitat:
an experimental study with bacteria and
phage. Ecology 58:369–78
22. Chase JM, Abrams PA, Grover JP, Diehl
S, Chesson P, et al. 2002. The interac-
tion between predation and competition:
areview and synthesis. Ecol. Lett. 5:302–
15
23. Cho JC, Tiedje JM. 2000. Biogeography
and degree of endemicity of fluorescent
Pseudomonas strains in soil. Appl. Envi-
ron. Microbiol. 66:5448–56
24. Cohan FM. 2002. Sexual isolation and
speciation in bacteria. Genetica 116:359–
70
25. Cooper TF, Rozen DE, Lenski RE. 2003.
Parallel changes in gene expression af-
ter 20,000 generations of evolution in Es-
cherichia coli. Proc. Natl. Acad. Sci. USA
100:1072–77
26. Cooper VS, Lenski RE. 2000. The popula-
tion genetics of ecological specialization
in evolving Escherichia coli populations.
Nature 407:736–39
27. Cuevas JM, Moya A, Elena SF. 2003.
Evolution of RNA virus in spatially
structured heterogeneous environments.
J. Evol. Biol. 16:456–66
28. Curtis TP, Sloan WT, Scannell JW.
2002. Estimating prokaryotic diversity
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 227
and its limits. Proc. Natl. Acad. Sci. USA
99:10494–99
29. Czaran TL, Hoekstra RF, Pagie L. 2002.
Chemical warfare between microbes pro-
motes biodiversity. Proc. Natl. Acad. Sci.
USA 99:786–90
30. Dallinger WH. 1887. The President’s Ad-
dress. J. R. Microsc. Soc. 7:184–99
31. Darwin C. 1890. The Origin of Species.
London: John Murray
32. de Visser JAGM, Zeyl CW, Gerrish PJ,
Blanchard JL, Lenski RE. 1999. Dimin-
ishing returns from mutation supply rate
in asexual populations. Science 283:404–
6
33. Dean AM. 1989. Selection and neutral-
ity in lactose operons of Escherichia coli.
Genetics 123:441–54
34. Depolo NJ, Giachetti C, Holland JJ. 1987.
Continuing coevolution of virus and de-
fective interfering particles and of vi-
ral genome sequences during undiluted
passages: virus mutants exhibiting nearly
complete resistance to formerly dominant
defective interfering particles. J. Virol. 61:
454–64
35. Dobzhansky T. 1951. Genetics and the
Origin of Species.New York: Columbia
Univ. Press
36. Dowell CE. 1980. Growth of bacterio-
phage Phi X 174 at elevated temperatures.
J. Gen. Virol. 49:41–50
37. Dykhuizen DE. 1990. Experimental stud-
ies of natural selection in bacteria. Annu.
Rev. Ecol. Syst. 21:373–98
38. Dykhuizen DE. 1998. Santa Rosalia re-
visited: Why are there so many species of
bacteria? Antonie van Leeuwenhoek 73:
25–33
39. Ebert D. 1998. Experimental evolution of
parasites. Science 282:1432–35
40. Elena SF, Lenski RE. 1997. Long-term
experimental evolution in Escherichia
coli VII: mechanisms maintaining genetic
variability within populations. Evolution
51:1058–67
41. Elena SF, Lenski RE. 2003. Evolution ex-
periments with microorganisms: the dy-
namics and genetic bases of adaptation.
Nat. Rev. Genet. 4:457–69
42. Elton C. 1927. Animal Ecology. London:
Sidgwick & Jackson
43. Falconer DS. 1981. Introduction to Quan-
titative Genetics. London: Longmann.
2nd ed.
44. Feldgarden M, Byrd N, Cohan FM. 2003.
Gradual evolution in bacteria: evidence
from Bacillus systematics. Microbiology
149:3565–73
45. Fisher RA. 1958. The Genetical Theory
of Natural Selection.New York: Dover
Publ.
46. Fukami T, Morin PJ. 2003. Productivity-
biodiversity relationships depend on the
history of community assembly. Nature
424:423–26
47. Gause GF. 1934. The Struggle for Exis-
tence. Baltimore: Williams & Wilkins
48. Gerrish PJ, Lenski RE. 1998. The fate
of competing beneficial mutations in an
asexual population. Genetica 102–103:
127–44
49. Gieseke A, Bjerrum L, Wagner M, Amann
R. 2003. Structure and activity of mul-
tiple nitrifying bacterial populations co-
existing in a biofilm. Environ. Microbiol.
5:355–69
50. Gieseke A, Purkhold U, Wagner M,
Amann R, Schramm A. 2001. Com-
munity structure and activity dynamics
of nitrifying bacteria in a phosphate-
removing biofilm. Appl. Environ. Micro-
biol. 67:1351–62
51. Gu ZL, Steinmetz LM, Gu X, Scharfe C,
Davis RW, Li WH. 2003. Role of dupli-
cate genes in genetic robustness against
null mutations. Nature 421:63–66
52. Haldane JBS. 1932. The Causes of Evolu-
tion. London: Longmans & Green
53. Hall BG. 2003. The EBG system of E.
coli: origin and evolution of a novel beta-
galactosidase for the metabolism of lac-
tose. Genetica 118:143–56
54. Hardin G. 1960. The competitive exclu-
sion principle. Science 131:1292–97
55. Haubold B, Rainey PB. 1996. Genetic and
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
228 KASSEN
RAINEY
ecotypic structure of a fluorescent Pseu-
domonas population. Mol. Ecol. 5:747–61
56. Helling RB, Vargas CN, Adams J. 1987.
Evolution of Escherichia coli during
growth in a constant environment. Genet-
ics 116:349–58
57. Hendrickson H, Slechta ES, Bergthors-
son U, Andersson DI, Roth JR. 2002.
Amplification-mutagenesis: evidence that
“directed” adaptive mutation and general
hypermutability result from growth with
a selected gene amplification. Proc. Natl.
Acad. Sci. USA 99:2164–69
58. Horner-Devine MC, Carney KM, Bohan-
nan BJM. 2004. An ecological perspec-
tive on bacterial biodiversity. Proc. R. Soc.
London Sci. Ser. B 271:113–22
59. Horodyski FM, Nichol ST, Spindler KR,
Holland JJ. 1983. Properties of DI par-
ticle resistant mutants of vesicular stom-
atitis virus isolated from persistent infec-
tions and from undiluted passages. Cell
33:801–10
60. Hubbell SP. 2001. The Unified Neutral
Theory of Biodiversity and Biogeography.
Princeton, NJ: Princeton Univ. Press
61. Hutchinson GE. 1961. The paradox of the
plankton. Am. Nat. 95:137–45
62. Imhof M, Schlotter C. 2001. Fitness ef-
fects of advantageous mutations in evolv-
ing Escherichia coli populations. Proc.
Natl. Acad. Sci. USA 98:1113–17
63. Jessup CM, Kassen R, Forde SE, Kerr B,
Buckling A, et al. 2004. Big questions,
small worlds: microbial model systems in
ecology. Trends Ecol. Evol. 19:189–97
64. Kassen R. 2002. The experimental evo-
lution of specialists, generalists, and the
maintenance of diversity. J. Evol. Biol. 15:
173–90
65. Kassen R, Bell G. 1998. Experimental
evolution in Chlamydomonas.IV. Selec-
tion in environments that vary through
time at different scales. Heredity 80:732–
41
66. Kassen R, Bell G. 2000. The ecology and
genetics of fitness in Chlamydomonas. X.
The relationship between genetic correla-
tion and genetic distance. Evolution 54:
425–32
67. Kassen R, Buckling A, Bell G, Rainey PB.
2000. Diversity peaks at intermediate pro-
ductivity in a laboratory microcosm. Na-
ture 406:508–12
68. Deleted in proof
69. Kaunzinger CMK, Morin PJ. 1998. Pro-
ductivity controls food-chain properties in
microbial communities. Nature 395:495–
97
70. Kent AD, Triplett EW. 2002. Microbial
communities and their interactions in soil
and rhizosphere ecosystems. Annu. Rev.
Microbiol. 56:211–36
71. Kerr B, Riley MA, Feldman MW, Bohan-
nan BJM. 2002. Local dispersal promotes
biodiversity in a real-life game of rock-
paper-scissors. Nature 418:171–74
72. Deleted in proof
73. Kimura M. 1983. The Neutral Theory
of Molecular Evolution. Cambridge, UK:
Cambridge Univ. Press
74. Klappenbach JA, Dunbar JM, Schmidt
TM. 2000. rRNA operon copy number
reflects ecological strategies of bacteria.
Appl. Environ. Microbiol. 66:1328–33
75. Kolenbrander PE. 2000. Oral microbial
communities: biofilms, interactions, and
genetic systems. Annu. Rev. Microbiol.
54:413–37
76. Korona R, Nakatsu CH, Forney LJ, Lenski
RE. 1994. Evidence for multiple adaptive
peaks from populations of bacteria evolv-
ing in a structured habitat. Proc. Natl.
Acad. Sci. USA 91:9037–41
77. Lack D. 1947. Darwin’s Finches. Cam-
bridge, UK: Cambridge Univ. Press
78. Leigh EG. 1999. The modern synthesis,
Ronald Fisher and creationism. Trends
Ecol. Evol. 14:495–98
79. Lenski R, Travisano M. 1994. Dynam-
ics of adaptation and diversification: a
10,000-generation experiment with bac-
terial populations. Proc. Natl. Acad. Sci.
USA 91:6808–14
80. Lenski RE. 1988. Experimental studies
of pleiotropy and epistasis in Escherichia
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 229
coli. I. Variation in competitive fitness
amongst mutants resistant to virus T4.
Evolution 42:425–32
81. Lenski RE, Levin BR. 1985. Constraints
on the coevolution of bacteria and viru-
lent phage: a model, some experiments,
and predictions for natural communities.
Am. Nat. 125:585–602
82. Lenski RE, Rose MR, Simpson SC, Tadler
SC. 1991. Long-term experimental evolu-
tion in Escherichia coli.I.Adaptation and
divergence during 2,000 generations. Am.
Nat. 138:1315–41
83. Levene H. 1953. Genetic equilibrium
when more than one ecological niche is
available. Am. Nat. 87:331–33
84. Levin BR. 1972. Coexistence of two asex-
ual strains on a single resource. Science
175:1272–74
85. Levins R. 1968. Evolution in Changing
Environments. Princeton, NJ: Princeton
Univ. Press
86. Lewontin RC. 1974. The Genetic Ba-
sis of Evolutionary Change.New York:
Columbia Univ. Press
87. Lin ECC, Hacking AG, Aguilar J. 1976.
Experimental models of acquisitive evo-
lution. BioScience 26:548–55
88. Loreau M, Naeem S, Inchausti P, eds.
2002. Biodiversity and Ecosystem Func-
tioning. Oxford: Oxford Univ. Press
89. Lynch M, Force A. 2000. The probabil-
ity of duplicate gene preservation by sub-
functionalization. Genetics 154:459–73
90. MacLean RC, Bell G. 2002. Experimental
adaptive radiation in Pseudomonas. Am.
Nat. 160:569–81
91. MacLean RC, Bell G, Rainey PB. 2004.
The evolution of a pleiotropic fitness
tradeoff in Pseudomonas fluorescens.
Proc. Natl. Acad. Sci. USA 101:8072–77
92. Maynard Smith J. 1998. Evolutionary Ge-
netics. Oxford: Oxford Univ. Press
93. Maynard Smith J, Hoekstra R. 1980. Poly-
morphism in a varied environment: How
robust are the models? Genet. Res. 35:45–
57
94. McGrady-Steed J, Morin PJ. 2000. Bio-
diversity, density compensation, and the
dynamics of populations and functional
groups. Ecology 81:361–73
95. Mortlock RP. 1982. Metabolic acquisi-
tions through laboratory selection. Annu.
Rev. Microbiol. 36:259–84
96. Notley-McRobb L, Ferenci T. 2000. Ex-
perimental analysis of molecular events
during mutational periodic selections in
bacterial evolution. Genetics 156:1493–
501
97. Novick A, Szilard L. 1950. Experiments
with the chemostat on spontaneous mu-
tations of bacteria. Proc. Natl. Acad. Sci.
USA 36:708–19
98. Olsen GJ, Lane DJ, Giovannoni SJ, Pace
NR, Stahl DA. 1986. Microbial ecol-
ogy and evolution: a ribosomal RNA ap-
proach. Annu. Rev. Microbiol. 40:337–65
99. Orr HA. 2003. The distribution of fitness
effects among beneficial mutations. Ge-
netics 163:1519–26
100. Pace NR. 1997. A molecular view of mi-
crobial diversity and the biosphere. Sci-
ence 276:734–40
101. Pace NR, Stahl DA, Lane DJ, Olsen
GJ. 1986. The analysis of natural micro-
bial populations by ribosomal RNA se-
quences. Adv. Microb. Ecol. 9:1–55
102. Papadopoulos D, Schneider D, MeierEiss
J, Arber W, Lenski RE, Blot M. 1999.
Genomic evolution during a 10,000-
generation experiment with bacteria.
Proc. Natl. Acad. Sci. USA 96:3807–12
103. Raff RA. 1996. The Shape of Life:
Genes, Development, and the Evolution
of Animal Form. Chicago/London: Univ.
Chicago Press. 520 pp.
104. Rainey PB, Buckling A, Kassen R, Trav-
isano M. 2000. The emergence and main-
tenance of diversity: insights from experi-
mental bacterial populations. Trends Ecol.
Evol. 15:243–47
105. Rainey PB, Rainey K. 2003. The evolution
of cooperation and conflict in experimen-
tal bacterial populations. Nature 425:72–
74
106. Rainey PB, Travisano M. 1998. Adaptive
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
230 KASSEN
RAINEY
radiation in a heterogeneous environment.
Nature 394:69–72
107. Rapp´e MS, Giovannoni SJ. 2003. The un-
cultured microbial majority. Annu. Rev.
Microbiol. 57:369–94
108. Reboud X, Bell G. 1997. Experimental
evolution in Chlamydomonas. III. Evolu-
tion of specialist and generalist types in
environments that vary in space and time.
Heredity 78:507–14
109. Rosenzweig ML. 1995. Species Diversity
in Space and Time. Cambridge, UK: Cam-
bridge Univ. Press
110. Rosenzweig RF, Sharp RR, Treves DS,
Adams J. 1994. Microbial evolution in a
simple unstructured environment: genetic
differentiation in Escherichia coli. Genet-
ics 137:903–17
111. Rozen DE, Lenski RE. 2000. Long-term
experimental evolution in Escherichia
coli. VIII. Dynamics of a balanced poly-
morphism. Am. Nat. 155:24–35
112. Schloter M, Lebuhn M, Heulin T, Hart-
mann A. 2000. Ecology and evolution
of bacterial microdiversity. FEMS Micro-
biol. Rev. 24:647–60
113. Schluter D. 2000. The Ecology of Adap-
tive Radiation. Oxford: Oxford Univ.
Press
114. Simpson GG. 1953. The Major Features
of Evolution.New York: Columbia Univ.
Press
115. Sniegowski PD, Gerrish PJ, Lenski RE.
1997. Evolution of high mutation rates in
experimental populations of Escherichia
coli. Nature 387:703–5
116. Sommer U. 1999. Competition and coex-
istence. Nature 402:36–37
117. Sonti VR, Roth JR. 1989. Role of gene du-
plications in the adaptation of Salmonella
typhimurium to growth on limiting carbon
sources. Genetics 123:19–28
118. Souza V, Nguyen TT, Hudson RR, Pinero
D, Lenski RE. 1992. Hierarchical analy-
sis of linkage disequilibrium in Rhizobium
populations: evidence for sex? Proc. Natl.
Acad. Sci. USA 89:8389–93
119. Spiers AJ, Kahn SG, Bohannon J, Trav-
isano M, Rainey PB. 2002. Adaptive di-
vergence in experimental populations of
Pseudomonas fluorescens.I.Genetic and
phenotypic bases of wrinkly spreader fit-
ness. Genetics 161:33–46
120. Stern D. 2000. Evolutionary developmen-
tal biology and the problem of variation.
Evolution 54:1079–91
121. Stewart FM, Levin BR. 1973. Partitioning
of resources and the outcome of interspe-
cific competition: a model and some gen-
eral considerations. Am. Nat. 107:171–98
122. Tilman D. 1982. Resource Competition
and Community Structure. Princeton, NJ:
Princeton Univ. Press
123. Tilman D. 2000. Causes, consequences
and ethics of biodiversity. Nature 405:
208–11
124. Tolker-Nielsen T, Molin S. 2000. Spatial
organization of microbial biofilm commu-
nities. Microb. Ecol. 40:75–84
125. Torsvik V, Daae FL, Sandaa RA, Ovreas
L. 1998. Novel techniques for analysing
microbial diversity in natural and per-
turbed environments. J. Biotechnol. 64:
53–62
126. Torsvik V, Ovreas L, Thingstad TF.
2002. Prokaryotic diversity—magnitude,
dynamics, and controlling factors. Sci-
ence 296:1064–66
127. Travisano M, Rainey PB. 2000. Studies of
adaptive radiation using model microbial
systems. Am. Nat. 156:S35–S44
128. Treves DS, Manning S, Adams J. 1998.
Repeated evolution of an acetate-
crossfeeding polymorphism in long-term
populations of Escherichia coli. Mol.
Biol. Evol. 15:789–97
129. Via S, Lande R. 1985. Genotype-
environment interaction and the evolu-
tion of phenotypic plasticity. Evolution
39:505–22
130. Ward DM, Ferris MJ, Nold SC, Bateson
MM. 1998. A natural view of microbial
biodiversity within hot spring cyanobac-
terial mat communities. Microbiol. Mol.
Biol. Rev. 62:1353–70
131. Whitaker RJ, Grogan DW, Taylor JW.
31 Aug 2004 11:33 AR AR224-MI58-09.tex AR224-MI58-09.sgm LaTeX2e(2002/01/18)
P1: IKH
MICROBIAL DIVERSITY 231
2003. Geographic barriers isolate en-
demic populations of hyperthermophilic
archaea. Science 301:976–78
132. Whittaker RH. 1975. Communities and
Ecosystems. London: Macmillan
133. Wichman HA, Badgett MR, Scott LA,
Boulianne CM, Bull JJ. 1999. Differ-
ent trajectories of parallel evolution dur-
ing viral adaptation. Science 285:422–
24
134. Woese CR. 1987. Bacterial evolution. Mi-
crobiol. Rev. 51:221–71
135. Chase JM, Leibold MA. 2003. Ecological
Niches: Linking Classical and Contempo-
rary Approaches. Chicago: Univ. Chicago
Press
MICROBIAL DIVERSITY C-1
Figure 2 Molecular evolution of beneficial phage mutants in two populations, TX
and ID. Lines and letters represent the frequency of novel mutants on a given day.
Numbers refer to the order in which each mutant appeared. Note that a single geno-
type tends to dominate the population at any given time. Reproduced with permis-
sion from Wichman et al. (133).
HI-RES-MI58-09-Kassen.qxd 8/31/04 12:22 PM Page 1
