Articles in PresS. Am J Physiol Regul Integr Comp Physiol (April 29, 2009). doi:10.1152/ajpregu.90562.2008
Microbial Experimental Evolution
Albert F. Bennett§ and Bradley S. Hughes
Department of Ecology and Evolutionary Biology
University of California, Irvine, California 92697-2525
Running head: Microbial Experimental Evolution
§ To whom all correspondence and queries should be addressed: [email protected];
PHONE: 949-824-5315; FAX: 949-824-3035
Copyright © 2009 by the American Physiological Society.
Microbial Experimental Evolution
Abstract:
Microbes have been widely used in experimental evolutionary studies because they
possess a variety of valuable traits that facilitate large-scale experimentation. Many
replicated populations can be cultured in the laboratory simultaneously along with
appropriate controls. Short generation times and large population sizes make microbes
ideal experimental subjects, ensuring that many spontaneous mutations occur every
generation and that adaptive variants can spread rapidly through a population. Another
highly useful experimental feature is the ability to preserve and store ancestral and
evolutionarily derived clones. These can be revived in parallel to allow the direct
measurement of the competitive fitness of a descendant in comparison to its ancestor.
The extent of adaptation can thereby be measured quantitatively and compared
statistically by direct competition among derived groups and with the ancestor. Thus,
fitness and adaptation need not be matters of qualitative speculation, but are
quantitatively measurable variables in these systems. Replication allows the
quantification of heterogeneity in responses to imposed selection and thereby statistical
distinction between changes that are systematic responses to the selective regime and
those that are specific to individual populations.
Keywords: acidity, adaptation, bacteria, experimental evolution, fitness, laboratory
natural selection, microbe, pH
2
Microbial Experimental Evolution
I. Introduction: The Krogh Principle and Experimental Evolution
In pursuing an experimental study, a primary and crucial decision that a
physiologist must make is the choice of the subject organism. Nearly a century ago,
August Krogh discussed the importance of this decision and articulated what
subsequently has become known as the Krogh Principle. This states that the choice of the
subject organism should be dictated by the experimental convenience that it brings to the
proposed study (34). Convenience here is not meant to simply imply how easily the
subject can be obtained or high frequency of use in other studies. Rather this refers to
investigator consideration of a subject's unique biological properties that can best foster a
direct and clear experimental design and an unambiguous result and interpretation (8).
The Krogh Principle has been particularly important in guiding successful research in
neurophysiology and comparative physiology, but should not be limited to those fields or
to physiological studies alone.
Historically, evolutionary biology has been dominated by descriptive, theoretical,
and comparative studies (21). While these have their individual strengths and insights,
they lack the rigor and strong inference that the experimental method can bring to
scientific investigations (20, 47). The emergence of experimental evolution has
permitted the examination of fundamental hypotheses and assumptions about how
evolution works (23). In addition, experimental evolutionary studies produce new
organisms with unique combinations of properties that can become the subjects of new
investigations into mechanism and function (8). A broad taxonomic array of organisms
has been used in these studies, from vertebrates to viruses. Each of these groups has
experimental advantages and limitations. Here we argue, however, that the Krogh
3
Microbial Experimental Evolution
Principle suggests the unique utility of microbes for understanding general issues and
patterns in evolution, for testing evolutionary hypotheses, and for investigating the
breadth of adaptive solutions to environmental change (8).
This article discusses the unique experimental advantages of microbes for
experimental evolutionary studies and presents an application of this method for the study
of evolutionary responses to environments of different acidities.
II. The Microbial Advantage
The hallmarks of successful evolutionary experimentation are replication and
control. Replication effectively allows the experimenter to copy the same experiment
over and over independently, so that pattern and significance of results can be analyzed.
Control is important in two senses: for regulation of all aspects of the newly imposed
selective environment and for continued maintenance of parallel lineages in the ancestral
environment. Therefore the ability to maintain numerous independent experimental
populations under tightly regulated conditions is inherently advantageous. Similarly
advantageous are the large population sizes and rapid reproduction. Microbes (and
viruses) excel in all of these features and more.
1. Replication of lineages.
Many types of microbes have been successfully domesticated for easy liquid
culture in laboratory culture-ware or chemostats. The numbers of replicate lines that can
4
Microbial Experimental Evolution
be cultured are essentially limited only by the desire, patience, and resources of the
experimenter. Dozens or even hundreds of populations can be propagated
simultaneously. Evolution in different environments can be structured and analyzed
simultaneously. For instance, it is feasible to examine adaptation to a range of different
nutrients and different concentrations of the same nutrient at the same time. Once
established and maintained without cross-contamination, each replicate is evolutionarily
independent.
There are two distinct analytical advantages that are permitted by the ability to
replicate lineages in this manner. First, experimental replicates cultured under identical
conditions can be analyzed statistically to document significant evolutionary change and
to place confidence limits on that change (35). Such an analysis can permit the
quantitative documentation of adaptation to a new environment by demonstrating a
statistically significant improvement in reproductive output within the group of lines
evolving in that environment. For instance, a group of replicated lineages of bacteria
maintained at 42°C for 2000 generations evolved significantly greater fitness in that
temperature that did groups evolving at the ancestral temperature of 37°C or a novel but
colder temperature of 32°C (5,6). The magnitude of this improvement was demonstrated
to be greater in this stressful environment than that attained by either of the other groups
in their selective environments. Such analyses and conclusions regarding group
differences are possible only because of replication of experimental lineages within each
of these environments.
Replication of lineages, secondly, permits an analysis of the diversity of potential
evolutionary solutions and outcomes, given identical starting conditions. It gives us the
5
Microbial Experimental Evolution
ability to investigate the range of potential outcomes and adaptive solutions that might
occur, rather than being confined to the analysis of a single outcome ex post facto, as
characterizes most evolutionary analyses of the natural world (24). From such replicated
experimental evolutionary studies, it appears that there is no single adaptive pathway, no
single predetermined evolutionary outcome, even for identical clonal populations
adapting to very simple environments (e.g., 6, 9, 46, 58). Such long term evolutionary
studies have demonstrated that a variety of phenotypes with very different underlying
genetic mechanisms appear among the replicated lineages.
2. Ancestral controls.
It is essential to determine whether observed changes in the evolving populations
are actually responses to the new selective environment or if those changes would have
occurred anyway in the ancestral environment. That is, is there continuing adaptation to
ancestral culture conditions that could be misconstrued as adaptation to the new selective
environment? With microbes there are several ways to minimize this concern. First, the
ancestral population can be drawn from lineages already maintained in and well adapted
to the laboratory culture conditions, thereby minimizing further evolutionary
improvement. Second, replicated populations can be cultured under ancestral conditions
in parallel with those changed to novel environments. The extent of further adaptation to
ancestral culture conditions can therefore be empirically analyzed and compared to that in
novel selective environments. In our temperature adaptation experiments, a line of
bacteria was first maintained at a culture temperature of 37°C for 2000 generations before
being used as the ancestor for other experimental groups, including a continuing
6
Microbial Experimental Evolution
experimental control group at 37°C. During the preconditioning 2000 generations, its
relative fitness improved about 33% (35); during the subsequent 2000 generations, fitness
increased only 1/10th as much (5). Extensive adaptation of the ancestral lineage to culture
conditions thus made it much easier to interpret fitness improvement in the other groups
as being temperature and not culture specific.
3. Environmental controls.
Because microbes can be maintained and cultured in supplemented liquid media
using standard laboratory culture-ware, it is possible to create highly replicable
environments within and among laboratories. Spatial heterogeneity can be eliminated by
agitation. Any environmental factor or pair of factors can be experimentally altered from
the ancestral condition, while otherwise maintaining every other feature of the ancestral
environment. The novel experimental environment can be constant or variable, selecting
respectively for specialists or generalists (e.g., 5). The experimental and ancestral
environments thus can be specified, controlled, and replicated to an extraordinary degree.
4. Enormous population size.
In a few milliliters of liquid culture, microbial populations numbering hundreds of
millions or even billions of individuals can be grown and maintained. For evolutionary
studies, there is always a problem if population size is too low: genetic drift, the fixation
of traits at random, becomes a major concern when too few individuals compete and
reproduce. Most experimental investigators prefer populations in the hundreds, or even a
thousand, as a minimum to avoid this effect. Populations of those magnitudes can be
7
Microbial Experimental Evolution
difficult to maintain and manage for some kinds of organisms, especially vertebrates, but
are easily obtained using microbes.
There is also a secondary and perhaps less obvious advantage in having truly huge
population sizes: the probability of the emergence of mutational novelty. During DNA
replication, there is a low but finite probability of an error in replication. The probability
of the occurrence of such a mutation is the product of population size and the mutation
rate, and thus larger populations facilitate the greater chances of obtaining a beneficial
mutation on which selection can operate. In long term evolutionary studies on microbial
populations of large size, it is probable that every single base pair within the genome has
experienced a mutation at least once and has been subjected to consequent selection (38.).
5. Rapid reproduction.
Selection operates on individuals, favoring or disfavoring their reproduction. The
more rapid reproduction is, the more quickly favored genotypes spread through the
population. For multicellular organisms, generation times may minimally be days or
weeks; for microbes, minutes or hours. Beneficial mutations may therefore spread
rapidly in microbial populations: mutational genetic adaptation was observed to occur
and increase mean population fitness within only 100 to 200 generations, 15 to 30 days,
of high temperature selection (4). Conversely, harmful mutations are also more rapidly
eliminated. Because of their very short generation times, evolutionary experiments on
microbes may feasibly span thousands or tens of thousands of generations (e.g., 38,46).
Reproduction is clonal, so an intact genome can be passed directly to the descendants
without being deconstructed by genetic recombination.
8
Microbial Experimental Evolution
6. Ancestral preservation.
Because cultures of microbes survive being frozen and thawed again, they can be
preserved more or less indefinitely in a non-evolving frozen state and then revivified.
The ability to maintain viable frozen cultures means that an entire evolutionary sequence
can be preserved, reconstructed, and analyzed. It also means that the ancestor and its
distant descendant can be compared contemporaneously at any later date (35). With
microbes, it is not necessary to make every conceivable measurement on an ephemeral
ancestral state at the beginning of an experiment. It is also possible to make
measurements on the ancestral state with technologies that were not developed until after
the experiment was initiated and to continue to apply new technologies to the
experimental series indefinitely into the future.
7. Fitness measurement.
Because the ancestor and its descendant can be brought together in the same place
at the same time, they can be directly compared. Of special interest, their relative
reproductive potential can be assayed in direct competition. Placed together in the same
environment and given access to resources, the number of offspring produced by the
ancestor and the descendant can be differentiated and a quantitative measurement of their
relative fitness in that environment can be obtained (35). In most experimental systems,
evolutionary biologists have to be content with measuring and comparing presumptive
fitness components, that is, phenotypic traits assumed a priori to be of importance within
the novel selective environment e.g., renal clearance capacity or nerve conduction rate.
9
Microbial Experimental Evolution
Such trait-by-trait measurements may give some insight into some mechanisms
underlying evolutionary change. However, their impact on adaptive evolutionary change
cannot be quantified and the contribution of unsuspected or unmeasured traits may be
missed entirely. In determining evolutionary success, organisms should not be
fragmented and studied as a series of isolated traits; it is only within the context of their
integrated functioning in the intact organism that their importance should be assayed (3).
The best measure of such integrated function within an evolutionary context is their
summed effect on reproductive potential. Relative fitness is the measure of the evolved
line's population growth compared to that of the ancestral line, when the two lines are
mixed together and directly compete in the same flask for shared nutrients. Assayed in
the novel selective environment, an improvement in fitness of the descendant relative to
the ancestor provides an unambiguous demonstration of evolutionary adaptation and an
integrated quantitative measurement of the degree of improvement resulting from all
accumulated genetic changes. Once adaptive fitness has been determined, its
components and genetic basis can be studied. It is also possible to assay fitness in other
environments, including the ancestral one, and determine if evolutionary adaptation has
been accompanied by a loss of reproductive potential in non-selective environments.
Such correlated decrements in fitness are termed trade-offs and are widely anticipated in
evolutionary theory; they can be directly determined in microbial evolutionary
experiments (e.g., (9)).
8. Genetic basis of adaptation.
10
Microbial Experimental Evolution
The genomes of dozens of species of microbes have been completely sequenced
and many more are anticipated. It is thus feasible to analyze the exact genetic changes
associated with adaptation in a microbial evolutionary experiment (e.g., 46,57).
Microbial genetic structures are relatively easy to manipulate, so the presumptive
phenotypic effects of any novel mutation or rearrangement can be independently verified
experimentally. Microbial genomes are relatively small, and the functions of many genes
are already well understood, facilitating the interpretation and understanding of the
functional basis of adaptation. After completing the evolutionary selection, along with
measurement of fitness and of physiological performance, the investigations of the
evolutionary physiologists can be extended through characterization of the genetic
architecture of the adaptations. For example, a study of high temperature adaptation
employed microarray technologies, both to measure evolved changes in gene expression
and for detection of large insertion and deletion events, to reveal candidate genes
associated with the higher expressions that occur during high temperature tolerance (49).
Another study used microarrays and PCR to identify the genetic duplications and
deletions involved in the specialized adaptations to two different sugars and to a mixture
of the sugars, revealing genetic mechanisms of antagonistic pleiotropic trade-offs of
specialization that likely inhibited evolution of generalists (59).
No experimental organism is ideal for all types of studies, physiological or
evolutionary. For instance, for some questions concerning evolutionary changes in
intercellular, organ or systems function and integration, microbes have nothing to offer,
and other experimental organisms (e.g., nematodes, (53)) may be more useful. However,
11
Microbial Experimental Evolution
for understanding the diversity and mechanisms of evolutionary adaptation and for
testing evolutionary theory, they offer the very significant experimental advantages
enumerated above. Microbes have been used in hundreds of experimental evolutionary
studies (see reviews by (8, 13, 17, 36, 37, 43)). Here we exemplify this type of research
with some of our recent work on adaptation of a bacterium to environmental pH. We
first however review the very first microbial evolutionary study.
III. The First Experimental Study of Microbial Evolution
The first microbial evolution experiment was conducted more than a century ago
by an ordained Methodist minister, the Reverend Dr. William Dallinger (15, 26). In
addition to his religious duties, Dallinger was an avid microscopist and served as
President of the Royal Microscopical Society. An evolutionist, he set out to investigate
whether it was possible to induce thermal adaptive change in an organism with a short
life cycle. Dallinger cultured rapidly propagating protozoans for seven years while
gradually increasing temperature. Starting at 16°C, he slowly raised the temperature to
23°C, at which point many of the population died, so temperature was held constant for
three months until regeneration rates recovered and thermal increments were slowly
continued. After seven years the protozoan cultures were able to tolerate temperatures of
70°C. The experiment was ended due to equipment failure. Interestingly it was noted
that the higher temperature adapted organisms died when placed back at the ancestral
temperature of 16°C, suggesting that a trade-off of low temperature survival accompanied
adaptation to high temperatures. While this experiment lacked many of the elements
12
Microbial Experimental Evolution
today deemed necessary for experimental evolution studies (e.g., replication,
contamination controls), it was widely regarded as an important demonstration of
evolution at the time, including reports in the popular press (14) and encouragement from
Charles Darwin.
IV. Experimental Evolutionary Adaptation to Environmental pH
Adaptation to physical environmental factors has been a major theme in microbial
experimental evolutionary studies. Beginning with Dallinger, most of this work has
concentrated on temperature as the selective environmental variable (e.g., (5, 11)) and
this work has been reviewed elsewhere (8). Here we discuss some of our recent
experimental work on adaptation to the environmental factor of pH. Environmental
acidity is both a great physiological challenge and evolutionary stress to many enteric
organisms because of their requisite passage through the extreme acidity of the stomach.
Our experimental organism, the bacterium Escherichia coli, is able to tolerate brief
exposure to a pH as low as 2.0. Because of this extreme acid tolerance, as few as 10 E.
coli cells are able to colonize a host, whereas Salmonella, with an acid tolerance of about
pH 3.0, requires over 10,000 cells to produce an infection (2).
1. Evolution in constant environments
In our experiments (29-31), a bacterial clone with an historical evolutionary
exposure to pH 7.0 was used to found groups of lineages in a variety of environments of
constant (pH 5.3, 6.3, 7.0, and 7.8) or variable (Cycled and Random, see below) pH.
13
Microbial Experimental Evolution
These pH environments encompassed more than a 300-fold range in hydronium ion
concentration (15-5000 nM). Derived lineages were propagated in six-fold replication
for 300 days to produce 2,000 generations in serial dilution culture within these new
selective environments. All other aspects of historical culture conditions were maintained
constant, and rigorous controls to detect external or cross-contamination were employed.
At the end of the experiment, evolutionary change in relative fitness of experimentally
evolved lineages was calculated as the ratio of the logarithm of population growth
doublings achieved by the experimental competitor compared to that of the common
ancestor (35). A relative fitness value (W) significantly greater than 1 signifies improved
evolutionary fitness, W=1 means that fitness did not change, and W <1 indicates a loss of
fitness. The populations propagated at pH 7.0 did not demonstrate a further significant
increment in relative fitness, indicating that the ancestral clone was already well adapted
to that pH under those culture conditions.
Of the lines that evolved in constant pH environments (see Table 1), the pH 5.3
group had the largest average increment in fitness (W=1.20) in its selective environment,
followed by the pH 7.8 group (W=1.08). All six lines of the former group significantly
increased fitness and only 4 of the latter group did. These two environments are stressful
to the ancestral bacteria, and adaptation proceeded more rapidly and extensively at these
pHs that it did at less stressful ones.
The mere occurrence of evolutionary adaptation, which is a fundamental
expectation of evolutionary biology and commonplace in microbial evolution
experiments, is not nearly as interesting as the more complex patterns of correlation
revealed when fitness is assayed in other, non-selective pH environments. Expectations
14
Microbial Experimental Evolution
of trade-offs, traditionally considered a necessary cost accompanying fitness gain in one
environment as a fitness loss in other environments, are generally anticipated by
evolutionary theory. The patterns of trade-offs we observed experimentally were actually
far less predictable and more complex than expected. For instance, adaptation to the
extreme acid and alkaline environments yielded highly divergent trade-off patterns
between the two groups (see Table 2). The pH 7.8 group had only 2 of its 6 lines with a
trade-off at pH 5.3. In contrast, all 6 lines of the pH 5.3 group showed significant tradeoffs at pH 7.8 (mean W=0.77). Evidently, further adaptive gains in the acidic
environment seem to proceed by mechanisms that negatively impact the physiological
ability to handle more alkaline conditions. Evolution in the alkaline environment
however does not seem to have a similar effect; indeed one of the alkaline lines
significantly improved its performance in the acid environment. Patterns and
predictability of trade-offs are proving to be more complex than evolution theory
predicted, and individual exceptions can be found even when the general and statistical
trend suggests their presence (9). The diversity of trade-off responses probably reflects a
diversity of underlying adaptive mechanisms in each independent lineage. Loss of fitness
in alkaline environments by all of our acid adapted lines may be an interesting study
system to determine if there is a common mechanism or series of mechanisms underlying
acid adaptation.
2. Evolution in variable environments
Microbial experimental evolution can also explore adaptation in varying
environments in order to investigate such physiological and evolutionary questions as
15
Microbial Experimental Evolution
whether variable environments select for increased phenotypic flexibility or if
acclimation is necessarily beneficial (30). It has been repeatedly argued that temporally
variable environments should select for increased phenotypic flexibility (19, 41, 42, 56).
This hypothesis that evolution in a temporally varying environment would increase
phenotypic flexibility was not supported when tested in variable thermal environments in
experimental lineages of E. coli (7,39,40). However, there was significant support for
this hypothesis when we tested our lineages that evolved in temporally variable
environmental pH. A Cycled group of 6 lines was propagated for 2,000 generations
under conditions that regularly alternated each day between pH 5.3 and pH 7.8. When
competed against the group that had evolved at constant 5.3 pH, the Cycled group had a
transitional advantage of 59% when tested in an environment that alternated between pH
5.3Æ7.8Æ5.3 over the course of three days (see Table 3). When the same competition
was carried out against the group that had evolved in constant pH 7.8, the transitional
advantage was still greater in the Cycled group, but by a more modest 15%. This
evidence from variable pH evolution consistently supports the prediction that phenotypic
flexibility is increased by temporally varying environments. It also suggests significant
differences in physiological evolution between thermal and pH tolerance and between
acid and base tolerance.
We also tested the hypothesis that plasticity is most favored when selective
habitats are regular and are spread equally and randomly across habitats (22, 52). We
created a Random group of six lines that were cultured in pH environments that
fluctuated stochastically among 5.3, 6.3, 7.0, and 7.8 pH conditions. Differences between
the regularly Cycled and Random groups (see Table 3) were not significant in the
16
Microbial Experimental Evolution
transitional plasticity achieved, thus the data was not directly supportive of the hypothesis
that random exposure enhances plasticity. However, the Random group's adaptation was
accompanied by a significantly lower fitness in constant pH regimes, offering some
limited support for this hypothesis that equal regularity and strength of selective habitats
favor plasticity. We also note that it remains an open question, whether long-term
laboratory selection, under the constant conditions that occur during adaptation to the lab,
has substantially modified acclimation ability in comparison to wild-type strains of
natural bacterial systems.
This experiment also tested the beneficial acclimation hypothesis (39, 40),
specifically that acclimation in an environment beneficially enhances performance in that
same environment (27, 48). Reliable testing of this hypothesis is often complicated in
multicellular organisms by behavioral effects (28) or variant life stages (55).
Acclimation effects may span three generations (57), and multigenerational exposure to
acclimation conditions is desirable (8), which was easily achieved by employing
microbial evolution for experimental design. Although physiologists have argued that
acclimation in an environment is expected to benefit performance in that same
environment (27, 48), our study found only mixed support for the beneficial acclimation
hypothesis with results having substantial heterogeneity of performance in response to
acclimation among the fluctuating and constant groups tested (See Table 4). While the
pH 5.3 group demonstrated beneficial acclimation, the pH 7.8 group had significant
decrements in performance after base acclimation. Interestingly, Table 4 also shows that
the Cycled and Random groups had very small or insignificant acclimation effects,
partially supporting a different hypothesis, that temporal variation will evolve reduced
17
Microbial Experimental Evolution
sensitivity to acclimation. The rapidly obtained and highly controlled experimental
results testing these four fundamental physiological evolution hypotheses further
illustrate the power and efficiency of microbial experimental evolution to test
evolutionary physiological hypotheses.
3. A Physiological Environmental Laboratory Analog
Microbial experimental evolution also has the potential to contribute to multiple
fields simultaneously through interdisciplinary approaches. An example is illustrated
here with an integrative study that touches on evolution, public health, and ecology, while
extending controlled laboratory methods to model natural physiological and
environmental processes (31). In this experiment, evolutionary patterns of bacterial
growth and survival fitness were measured in a complex multi-stress environmental
laboratory Analog. This study examined evolutionary trade-offs in a series of
environments and identified ecologically important abiotic factors in the coastal
ecosystem that pose a potential public health threat for transmission of pathogenic strains
of E. coli between the coastal environment and human host. Although serious health
concerns over E. coli have focused largely on its transmission through cattle (44),
pathogenic E. coli is now finding its way into recreational waters (18), with rapid urban
development in coastal regions increasing volumes of sewage discharge and urban runoff
to the coastal ocean (1, 45). E. coli has been found persisting in tropical freshwater
environments (33), suggesting that a natural isolate may have become genetically adapted
to that environment. Although E. coli has long been thought to die off rapidly in
seawater and to lack the ability to grow and proliferate in the ocean (51), the question of
18
Microbial Experimental Evolution
which environmental factors might eventually cause this organism to adapt to seawater
warrants consideration. Testing this question directly in the natural environment would
be confounded with variables of uncontrolled biological interactions, so this study (32)
sought to establish a controlled laboratory analog methodology to determine the roles that
various abiotic environmental factors may play in E. coli's potential to evolve increased
survival and growth in seawater and a human host (see Figure 1).
In seeking a more complex sequential combination of multi-stress environments
than had previously been used, we designed a "Host to Coast Environmental Analog",
which attempted to balance environmental realism with experimental tractability. The
Analog simulated abiotic environments in an 11-day sequence through the (Host) small
intestine and colon, transitioning through the (Coasttransition) sewer and into the (Coast)
seawater, reentering the (Host) human stomach and back into the small intestine (Table 5,
see (31,32) for full discussion). This laboratory-based system (31,32) approximates
parameters of pH, temperature, salinity, and nutrients at natural environmental levels,
within the constraints of utilizing the standard accepted media and methods employed for
microbial evolutionary experimentation. In regard to the stomach component, for
instance, it is possible to estimate pathogen survival and growth under approximate
gastric conditions (54). Although an earlier experimenter's attempt with a pH decreasing
immediately below 2.0 could not detect culturable cells (10), our experimental design
carefully controlled conditions to approximate actual physiological conditions more
closely and reflect bacterial growth challenges more accurately. First, a more reasonable
initial stomach pH of 6.8 was chosen for the Analog's stomach acid cycle (See Figure 2),
because that is typical immediately after meal intake rather than at fasting, when pH is
19
Microbial Experimental Evolution
below 2.0 (16, 50), and thereby avoids a total loss of culturable bacteria. Secondly, the
Analog models the acid cycle exposure time according to stomach acid concentrations of
young healthy adults, with average time of about 120 minutes for 50% of their stomach
contents to empty (12). Finally, only very gradual acidification with human physiological
pH concentration of cephalic phase parietal cell secretion (25) and alkalization with
shaking achieves similarity to human gastric conditions, and yields successful survival
and tractable measurements of fitness. Replicates of several different laboratory evolved
bacterial clones were independently run through the Analog [the Acid, Alkaline, and
Cycled lines discussed above and both 42ºC (5) and 14ºC (Bennett, unpublished data)
thermal specialists], as well as natural E. coli isolates sampled from an urban sewer and
coastal seawater.
This study (32) found that E. coli growth during the Host Analog is extremely
challenging, with culture count measurements undetectable during the stomach phase,
requiring that fitness assays be performed later during the phase simulating passage
through the small intestine to the colon (See Table 6). The natural Host Isolate strain,
sampled from an urban sewer, had the highest performance of relative growth fitness in
the Host Analog, which was expected and may offer further validity to the calibration
accuracy of Analog as a simulation of the human host. Of the experimental lines, the
Cycled pH line demonstrated the highest fitness in the Host Analog, along with the 14°C
line that also experienced a significant fitness gain (Table 6). Growth in the Coast Analog
is so challenging for E. coli that the experimental system was designed to measure
survival alone (See Table 7). Remarkably, growth did occur in one line (See Figure 3).
The surprisingly high significant seawater growth rate demonstrated in the Alkaline line
20
Microbial Experimental Evolution
(selected at pH 7.8) suggests what might be an important environmental health discovery:
prolonged alkaline exposure can lead to the evolution of increased survival and fitness in
seawater by E. coli. This study also contributed substantially to our understanding of the
complexity of trade-offs. While the 42°C Line had previously been shown to be a
generalist, exhibiting no trade-offs in any previously tested thermal environment (e.g.,
42°C=REL2051 in (6)), here it was shown to have a significant trade-off in the novel
environment of the Host Analog. This Line did not exhibit a trade-off in the Coast
Analog, actually demonstrating a relative fitness gain in the novel seawater environment.
The 14°C and Alkaline Lines were shown to be Host to Coast generalists, since their
respective gains in the Host (14°C Line) and Coast (Alkaline Line) Analog were not
accompanied by trade-offs of fitness in the reciprocal portion of the Analog. While the
42°C Line was shown to be a specialist for thermal fitness, the Alkaline Line seems to be
a generalist, an evolutionary "Jack-of-all-trades".
Future studies of this type of interdisciplinary application of microbial
experimental evolution may help inform many fields, such as the public health regulation
of wastewater management to prevent selection for strains of pathogenic E. coli
possessing generalist physiologies that could facilitate increased growth in coastal
seawater and infectivity in human hosts. As this type of approach develops further, we
may eventually be able to predict the development of diseases through understanding the
evolution of their bacterial or viral source.
V. Perspectives and Significance
21
Microbial Experimental Evolution
Microbes have proven to be exceptionally useful and tractable systems for
experimental evolutionary studies and may facilitate progress in the field of evolutionary
physiology. They are easy to obtain, store, and culture, either as mixed populations or
clones. Any experiment can effectively have as many replicated populations and controls
as the experimenter wishes. Growth rates are high and generation times are short, so that
advantageous mutations or genetic rearrangements can be selected for expeditiously.
Preservation of viable ancestral forms permits quantification of adaptation and trade-off
though competition experiments and measurement of differential reproduction and
fitness. For some species, there is an exceptionally large amount of information about
their genetics, molecular biology and biochemistry, and physiology. They are also
relatively easy to manipulate genetically and thereby permit experimental verification of
putative adaptive genetic effects that correlate with measures of physiological
performance. Using microbial systems in properly designed experiments, it is possible to
test evolutionary theory, to examine the diversity of adaptive physiological mechanisms
and responses to particular environments, and to select for new experimental organisms
with evolving physiologies of superior performance and fitness. These approaches can
also be applied to solve biological challenges, such as predicting the physiological
evolution of disease causing microbes as they adapt to environmental changes such as
global warming.
Acknowledgments
Supported by NSF Grant IOS 0748903 to AFB.
22
Microbial Experimental Evolution
Figure Legends:
Figure 1.
The Host to Coast Laboratory Analog system is used to predict evolutionary
patterns of Escherichia coli growth and survival in transmission between human host and
seawater. Illustration Shane Hunter Hughes, (32) Courtesy of Aquatic Microbial
Ecology, Vol. 53: 243–255, 2008
Figure 2.
pH of media versus time (hours: minutes) elapsed during the acid cycle of
the Host Analog. Short discontinuous horizontal bars indicate rapidly shifted pH levels
accompanying 0.9mL additions of 0.15M HCl acid at times 0:15, 0:30, 0:45, 1:00, 1:15,
and 1:30 during acidification and 0.5mL additions of 7.5% Solution NaHCO3 base at
times 2:00, 2:15, and 2:30 during neutralization. Continuous lines indicate sustained pH
levels of flasks, beginning with a mixture of 9.6mL of LB media and 0.4mL seawater
containing the competing bacteria, at time = 0:00 pH = 6.7, time = 1:30 pH = 1.93, and
time = 2:30 pH = 6.3 gradually increasing to 7.5 at elapsed time = 20:00 hours. (32)
Courtesy of Aquatic Microbial Ecology, Vol. 53: 243–255, 2008
Figure 3. Absolute survivorship percentage after five-days of exposure to Host Analog
filtered seawater (characterization analysis in Table 3) is shown for each experimental
line (Acid, Alkaline, pH Cycled, 14ºC, 42ºC, Host Isolate, and Coast Isolate) paired
alongside the survivorship of the matched Ancestor/Control it competed with in the same
flask. Survivorship values significantly higher than 100% (dashed line) indicate
population growth (black bars), survivorship values not significantly different from 100%
indicate population persistence (gray bars), and survivorship values significantly less than
23
Microbial Experimental Evolution
100% indicate population decline (white bars). (32) Courtesy of Aquatic Microbial
Ecology, Vol. 53: 243–255, 2008
24
Microbial Experimental Evolution
Tables:
Table 1. Relative fitness for constant pH evolved
groups tested directly in the their selective environments
after 2,000 generations.
1
Selective
Mean fitness
p
Line
pH
(±95%CI)1
values2
responses3
5.3
6.3
7.0
7.8
1.200 (±0.062)
1.050 (±0.009)
1.037 (±0.074)
1.076 (±0.105)
0.0002
<0.0001
0.1284
0.0615
6+
4+
3+
4+
0≈
2≈
2≈
2≈
0010-
The mean relative fitness (and 95% confidence interval)
for each group is calculated from six line means, each of
which is based on six replicates.
2
Significance values are based on one-tailed t-distribution
with μnull = 1 and d.f. = 5.
3
Each number followed by a +, ≈, or - indicates the number
of lines in the group with significant gains, no significant
changes, or a significant losses, respectively, in mean direct
fitness.
25
Microbial Experimental Evolution
Table 2. Trade-off patterns for Acid pH 5.3 group compared to Alkaline pH 7.8 group,
with means ± 95% confidence limits of relative fitness based on six replicate assays (line
means) or the six line means (group means).
Evolutionary pH
5.3 Evolved Acid Group (±95%CI)
7.8 Evolved Alkaline Group (±95%CI)
Assay pH
Direct: 5.3
Correlated: 7.8
Pattern1
Direct: 7.8
Correlated: 5.3
Pattern1
1st Line
1.239 (±0.048)
0.718 (±0.143)
T
0.897 (±0.107)
0.469 (±0.033)
-
2nd Line
1.179 (±0.064)
0.863 (±0.079)
T
1.102 (±0.097)
0.236 (±0.071)
T
3rd Line
1.089 (±0.043)
0.844 (±0.062)
T
1.171 (±0.086)
1.035 (±0.100)
N
4th Line
1.242 (±0.025)
0.560 (±0.089)
T
1.029 (±0.078)
1.130 (±0.078)
E
5th Line
1.224 (±0.080)
0.832 (±0.110)
T
1.150 (±0.053)
0.520 (±0.089)
T
6th Line
1.228 (±0.060)
0.820 (±0.059)
T
1.105 (±0.045)
1.050 (±0.070)
N
Group Mean
1.200 (±0.062)
0.773 (±0.122)
1.076 (±0.105)
0.740 (±0.396)
1
Shown for each line is the pattern of evolutionary change across the two environments
E (Exaptation): correlated fitness gain
N (No Trade-off): direct fitness gain with no correlated fitness loss
T (Trade-off): direct fitness gain but with correlated fitness loss
- (Negative Evolution Pattern): loss of direct and correlated fitness, is not
qualified as a trade-off since the loss does not accompany any gain
26
Microbial Experimental Evolution
Table 3. Transitional advantage of variable pH evolution, comparing adaptation within
cycled pH test regime between the variable Cycled and Random groups versus the
constant pH 5.3 or 7.8 groups.
Groups
Means:
Means:
Transitional Advantage
Cycled=C
Variable Groups
Constant Groups
of Variable Evolution
Random=R
Tested in
Tested in
Percentage Change
P
Compared
pH 5.3→7.8→5.3
pH 5.3→7.8→5.3
(Difference ±SE)a
(2-tailed)b
C – 5.3
1.120
.704
+59 (.417 ± .125)
.020
C – 7.8
1.120
.972
+15 (.148 ± .041)
.013
R – 5.3
1.087
.704
+54 (.384 ± .027)
.027
R – 7.8
1.087
.972
+12 (.115 ± .046)
.036
a
Percentage difference calculated by dividing variable by constant group relative fitness
(and absolute difference with SE of the difference) tested in the same overall regime,
based on six replicate lines in each experimental group.
lines in each experimental group, compared between the same overall regime.
b
Two-tailed probabilities were calculated using a t-test, assuming unequal variance,
with the null hypothesis that the mean difference equals 0.
27
Microbial Experimental Evolution
Table 4. Benefit of pH Acclimation
CIb
Pc
Acclimated Nonacclimated Acclimation SEa
Group
Acclimation Group Mean
Group Mean
Benefit
(Basic)
7.8→7.8pH
5.3→7.8pH
(Difference)
.708
.098
.209 ±.538 .033
pH 5.3
.610
1.010
1.175
.043 ±.111 .012
pH 7.8
-.165
1.147
1.143
039 ±.099 .914
.004
Cycled
1.035
1.073
.028 ±.072 .243
-.037
Random
(Acidic)
5.3→5.3pH
7.8→5.3pH
(Difference)
1.276
.939
.113 ±.292 .031
pH 5.3
.337
.809
.824
.037 ±.094 .700
pH 7.8
-.015
1.188
1.108
.020 ±.050 .010
.080
Cycled
1.142
1.103
.020 ±.051 .107
Random
.039
Note. Values shown represent the group means based on six-fold replication.
a
SE of the mean difference.
b
Confidence interval (CI) for the group mean difference
c
Two-tailed probabilities based on t distributions, with n – 1 = 5 df and the null
hypothesis that the mean difference equals 0.
28
Microbial Experimental Evolution
Table 5. Experimental Conditions of Temperature,
pH, and Culture Medium, as Sequenced in the
Host to Coast Environmental Laboratory Analog
Day of
Environment
Temp
pH
Culture
Analog
Simulated
Medium
Day 1
Proximal Æ
37°C 6.8Æ7.7 LB broth
HOST
Distal Small
Intestine
Day 2
Right Colon
Davis
37°C 7.0
HOST
Left Colon
Minimal
Days 3, 4
Sewer
20°C 6.8Æ8.2 LB broth
COASTTransition System
Days 53-9
Sea
Sea
14°C 8.2
COAST
Water
Water1
Day 104
Stomach
LB broth
37°C Acid
2
HOST
+HCL
Acidification
Cycle
6.8Æ2.0 +NaHCO3
Æ Intestine
Neutralization
2.0Æ6.8
Day 115
Proximal Æ
37°C 6.8Æ7.7 LB broth
HOST
Distal Small
Intestine
1
Seawater was sampled from Balboa pier in March of
2006 and filtered
2
Acid Cycle involves gradual acidification in the stomach
and gradual neutralization near the pancreatic duct in the
duodenum
3
First relative fitness assay
4
Second relative fitness assay
5
Third relative fitness assay
(32) Courtesy of Aquatic Microbial Ecology,
Vol. 53: 243–255, 2008
29
Microbial Experimental Evolution
Table 6. Host Analog Relative Growth Fitness
of Experimental Lines Tested in Competition
with the Ancestral Control Bacterium in the
Laboratory Environment of the Host Analog
on Days 10 and 11
P2
Experimental
Mean (±SE)1
E. coli Line
Relative
(2-tailed)
Tested
Growth Fitness Probability
Acid
<.001
.382 (±0.076)
Alkaline
.920
.997 (±0.028)
Cycled pH
<.001
1.186 (±0.019)
14°C
.008
1.154 (±0.036)
42°C
<.001
.556 (±0.014)
Host Isolate
<.001
2.117 (±0.096)
Coast Isolate
<.001
1.712 (±0.051)
1
Mean (and standard error of the mean)
based on six replicate tests for each evolved
or natural isolate experimental line.
2
Two-tailed probabilities were calculated
using the t-distribution with n – 1 = 5 degrees
of freedom; the null hypothesis was that the
mean fitness equaled 1.
(32) Courtesy of Aquatic Microbial Ecology,
Vol. 53: 243–255, 2008
30
Microbial Experimental Evolution
Table 7. Coast Analog Relative Survival Fitness
of Experimental Lines Tested in Competition
with the Ancestral Control Bacterium in the
Laboratory Environment of the Coast Analog
P2
Experimental
Mean (±SE)1
E. coli Line
Relative
(2-tailed)
Tested
Survival Fitness Probability
Acid
.263
1.144 (±0.279)
Alkaline
.010
2.909 (±0.469)
Cycled pH
.866
.996 (±0.020)
14°C
.570
1.050 (±0.083)
42°C
.010
1.124 (±0.030)
Host Isolate
<.001
.213 (±0.013)
Coast Isolate
<.001
1.457 (±0.029)
1
Mean (and standard error of the mean)
based on six replicate tests for each evolved
or natural isolate experimental line.
2
Two-tailed probabilities were calculated
using the t-distribution with n – 1 = 5 degrees
of freedom; the null hypothesis was that the
mean fitness equaled 1.
(32) Courtesy of Aquatic Microbial Ecology,
Vol. 53: 243–255, 2008
31
Microbial Experimental Evolution
References
1.
Ahn JH, Grant SB, Surbeck CQ, Digiacomo PM, Nezlin NP, and Jiang SC.
Coastal water quality impact of stormwater runoff from an urban watershed in southern
California. Environmental Science & Technology 39: 5940-5953, 2005.
2.
Audia, JP, Webb CC, and Foster JW. Breaking through the acid barrier: an
orchestrated response to proton stress by enteric bacteria. International Journal of
Medical Microbiology 291: 97-106, 2001.
3.
Bartholomew GA. Interspecific comparison as a tool for ecological
physiologists., In: New Directions in Ecological Physiology, edited by Feder ME,
Bennett AF, Burggren WW, and Huey RB: Cambridge University Press, Cambridge,
U.K., 1987, p. 11-35.
4
Bennett AF, Dao KM, and Lenski RE. Rapid evolution in response to high
temperature selection. Nature 346: 79-81, 1990.
5.
Bennett AF, Lenski RE, and Mittler JE. Evolutionary adaptation to
temperature. I. Fitness responses of Escherichia coli to changes in its thermal
environment. Evolution 46: 16-30, 1992.
6.
Bennett AF and Lenski RE. Evolutionary adaptation to temperature. II. Thermal
niches of experimental lines of Escherichia coli. Evolution 47: 1-12, 1993.
7.
Bennett AF and Lenski RE. Evolutionary adaptation to temperature: VI.
Phenotypic acclimation and its evolution in Escherichia coli. Evolution 51: 36-44, 1997.
8.
Bennett AF. Experimental evolution and the Krogh Principle: Generating
biological novelty for functional and genetic analyses. Physiological and Biochemical
Zoology 76: 1-11, 2003.
9.
Bennett AF and Lenski RE. An experimental test of evolutionary tradeoffs
during temperature adaptation. Proceedings of the National Academy of Sciences 104:
8649-8654, 2007.
10.
Beumer RR, de Vries J, and Rombouts FM. Campylobacter jejuni nonculturable coccoid cells. International Journal of Food Microbiology 15: 153-163, 1992.
11.
Bull JJ, Badgett MR, Wichman HA, Huelsenbeck JP, Hillis DM, Gulati A,
Ho C, and Molineux IJ. Exceptional convergent evolution in a virus. Genetics 147:
1497-1507, 1997.
12.
Clarkston WK, Pantano MM, Morley JE, Horowitz M, Littlefield JM, and
Burton FR. Evidence for the anorexia of aging: gastrointestinal transit and hunger in
healthy elderly vs. young adults. American Journal of Physiology 272: 243–248, 1997.
13.
Colegrave N and Collins S. Experimental evolution: experimental evolution and
evolvability. Heredity 100: 464-470, 2008.
14.
Daily Telegraph. open citation, note 50, p. 5. 19 Feb, 1887.
15.
Dallinger WH. The president's address. Journal of the Royal Microscopical
Society 10: 185-199, 1887.
16.
Dressman JB, Berardi RR, Dermentzoglou LC, Russell TL, Schmaltz SP,
Barnett JL, and Jarvenpaa KM. Upper gastrointestinal (GI) pH in young, healthy men
and women. Pharmaceutical Research 7:756–761, 1990.
17.
Dykhuizen DE. Experimental studies of natural selection in bacteria. Annual
Review of Ecology and Systematics 21: 378-398, 1990.
18.
Dziuban EJ, Liang JL, Craun GF, Hill V, Yu PA, Painter P, Moore MR,
Calderon RL, Roy SL, and Beach MJ. Surveillance for waterborne disease and
32
Microbial Experimental Evolution
outbreaks associated with recreational water-United States, 2003--2004. MMWR
Surveillance Summaries 55:1-24-30, 2006.
19.
Feder ME. Environmental variability and thermal acclimation in neotropical and
temperate zone salamanders. Physiological Zoology 51: 7-16, 1978.
20.
Fisher RA. The Design of Experiments. Edinburgh: Oliver & Boyd, 1935.
21.
Futuyma DJ. Evolutionary Biology, 3rd ed. Sinnauer Associates, Sunderland,
Massachusetts. 1998.
22.
Garland T and Kelly SA. Phenotypic plasticity and experimental evolution. The
Journal of Experimental Biology 209: 2344-2361, 2006.
23.
Garland T and Rose MR. Experimental Evolution: Concepts, Methods, and
Applications. University of California Press. In press.
24.
Gould SJ. Wonderful Life: The Burgess Shale and the Nature of History. New
York: Norton. 1989.
25.
Guyton AC and Hall JE. Textbook of Medical Physiology, 10th ed. W.B.
Saunders, Philadelphia, USA, 2000.
26.
Haas JW. The Reverend Dr William Henry Dallinger, F.R.S. Notes and Records
of the Royal Society of London 54: 56-65, 2000.
27.
Hochachka PW and Somero GN. Biochemical Adaptation. Mechanism and
Process in Physiological Evolution. Oxford University Press, New York, 2002.
28.
Huey RB, Hertz PE, and Sivervo B. Behavioral drive versus behavioral inertia
in evolution: a null model approach. The American Naturalist 161: 357-366, 2003.
29.
Hughes BS, Cullum AJ, and Bennett AF. Evolutionary adaptation to
environmental pH in experimental lineages of Escherichia coli. Evolution 61: 1725-1734,
2007.
30.
Hughes BS, Cullum AJ, and Bennett AF. An experimental evolutionary study
on adaptation to temporally fluctuating pH in Escherichia coli. Physiological and
Biochemical Zoology 80: 406-421, 2007.
31.
Hughes BS. Spiraling Evolutionary Patterns in Cycling Environments: The
Experimental Evolution of Physiologically Superior Ecological Generalists. Ph.D.
Dissertation 2007.
Hughes BS. Evolutionary patterns of Escherichia coli growth in seawater
32.
determined with a Host to Coast Environmental Laboratory Analog. Aquatic Microbial
Ecology 53: 243-255, 2008.
33.
Jimenez L, Muniz I, Toranzos GA, and Hazen TC. Survival and activity of
Salmonella typhimurium and Escherichia coli in tropical freshwater. Journal of Applied
Bacteriology 67: 61–69, 1989.
34.
Krogh A. Progress of physiology. American Journal of Physiology 90: 243-251,
1929.
35.
Lenski RE, Rose MR, Simpson SC, and Tadler SC. Long-term experimental
evolution in Escherichia coli. I. Adaptation and divergence during 2000 generations. The
American Naturalist 138: 1315-1341, 1991.
36.
Lenski RE. Experimental evolution. In: Encyclopedia of Microbiology, Vol. 2,
edited by Lederberg J: Academic Press, San Diego CA. 1992, p. 125-140.
37.
Lenski RE. Evolution in experimental populations of bacteria. In: Population
Genetics of Bacteria (Society for General Microbiology Symposium Series), edited by S.
33
Microbial Experimental Evolution
Baumberg, P. Young, E. Wellington, and J. Saunders: Cambridge Univ. Press. 1995, p.
193-215.
38.
Lenski RE, Winkworth CL, and Riley MA. Rates of DNA sequence evolution
in experimental populations of Escherichia coli during 20,000 generations. Journal of
Molecular Evolution 56: 498-508, 2003.
39.
Leroi AM, Bennett AF, and Lenski RE. Temperature acclimation and
competitive fitness: an experimental test of the beneficial acclimation assumption.
Proceedings of the National Academy of Sciences 91: 1917-21, 1994.
40.
Leroi AM, RE Lenski, and Bennett AF. Evolutionary adaptation to
temperature.III.Adaptation of Escherichia coli to a temporally varying environment.
Evolution 48: 1222-29, 1994.
41.
Levins R. Evolution in Changing Environments. Princeton University Press,
Princeton, N.J. 1968.
42.
Levins R. Thermal acclimation and heat resistance in Drosophila species. The
American Naturalist 103: 483-499, 1969.
43.
Levin BR and Lenski RE. Coevolution in bacteria and their viruses and
plasmids. In: Coevolution, edited by Futuyma DJ and Slatkin M: Sinauer Associates,
Sunderland MA., 1983, p. 99-127.
44.
Naugle AL, Holt KG, Levine P, and Eckel R. Sustained decrease in the rate of
Escherichia coli O157:H7-positive raw ground beef samples tested by the Food Safety
and Inspection Service. Journal of Food Protection 69: 480-1, 2006.
45.
Noble RT, Weisberg SB, Leecaster MK, McGee CD, Dorsey JH, Vainik P,
and Orozco-Borbon V. Storm effects on regional beach water quality along the southern
California shoreline. Journal of Water and Health 1: 23-31, 2003.
46.
Ostrowski EA, Woods RJ, and Lenski RE. The genetic basis of parallel and
divergent phenotypic responses in evolving populations of Escherichia coli. Proceedings
of the Royal Society, London B 275: 277-284, 2008.
47.
Platt JR. Strong inference. Science 146: 347-353, 1964.
48.
Prosser CL. Adaptational Biology: Molecules to Organisms. John Wiley & Sons,
New York. 766p 1986.
49.
Reihle MM, Bennett AF, and Long AD. Differential patterns of Gene
Expression and Gene Complement in Laboratory-Evolved Lines of E. coli. Integrative
Comparative Biology 45:532-538, 2005.
50.
Russell TL, Berardi RR, Barnett JL, Dermentzoglou LC, Jarvenpaa KM,
Schmaltz SP, and Dressman JB. Upper gastrointestinal pH in seventy-nine healthy,
elderly, North American men and women. Pharmaceutical Research 10: 187–196, 1993.
51.
Savageau MA. Escherichia coli habitats, cell types, and molecular mechanisms
of gene control. The American Naturalist 122: 732–744, 1983.
52.
Scheiner SM.. Genetics and evolution of phenotypic plasticity. Annual Review of
Ecology and Systematics 24: 35-68, 1993.
53.
Strange K. Revisiting the Krogh Principle in the post-genome era:
Caenorhabditis elegans as a model system for integrative physiology research. Journal
of Experimental Biology 210: 1622-1631, 2007.
54.
Tamplin ML. Inactivation of Escherichia coli O157:H7 in simulated human
gastric fluid. Applied and Environmental Microbiology 71: 320-325, 2005.
34
Microbial Experimental Evolution
55.
Terblanche JS and Chown SL. The relative contributions of developmental
plasticity and adult acclimation to physiological variation in the tsetse fly, Glossina
pallidipes (Diptera, Glossinidae). The Journal of Experimental Biology 209: 1064-1073,
2006.
56.
Tsuji JS. Thermal acclimation of metabolism in Sceloporus lizards from different
latitudes. Physiological Zoology 61: 241-253, 1988.
57.
Watson MJO and Hoffman AA. Cross-generation effects for cold resistance in
tropical populations of Drosophila melanogaster and Drosophila simulans. Australian
Journal of Zoology 43: 51-58, 1995.
58.
Woods R, Schneider D, and Winkworth CL, Riley MA, and Lenski RA. Tests
of parallel molecular evolution in a long-term experiment with Escherichia coli.
Proceedings of the National Academy of Sciences, USA 103: 9107-9112, 2006.
59.
Zhong S, Khodursky A, Dykhuizen DE, and Dean AM. Evolutionary genomics
of ecological specialization. Proceedings of the National Academy of Sciences, USA 101:
11719-11724, 2004.
35
Figures:
Figure 1.
Microbial Experimental Evolution
Figure 2.
2
Microbial Experimental Evolution
Figure 3.
3