Evolution in Microbesacademic.uprm.edu/~lrios/4368/Evol_in_microbes2013.pdfof evolution (37). Their...

BB42CH21-Kussell ARI 3 April 2013 17:42

Evolution in MicrobesEdo KussellCenter for Genomics and Systems Biology, Department of Biology, Department of Physics,New York University, New York, New York 10003; email: [email protected]

Annu. Rev. Biophys. 2013. 42:493–514

The Annual Review of Biophysics is online atbiophys.annualreviews.org

This article’s doi:10.1146/annurev-biophys-083012-130320

Copyright c© 2013 by Annual Reviews.All rights reserved

Keywords

bacteria, yeast, fitness, cooperation, physiology, ecology, evolutionarydynamics, experimental evolution

Abstract

This review presents a broad survey of experimental microbial evolution,covering diverse topics including trade-offs, epistasis, fluctuating conditions,spatial dynamics, cooperation, aging, and stochastic switching. Emphasis isplaced on examples that highlight key conceptual points or address the-oretical predictions. Experimental evolution is discussed from two pointsof view. First, population trajectories are described as adaptive walks on afitness landscape, whose genetic structure can be probed by experiments.Second, populations are viewed from a physiological perspective, and theirnongenetic heterogeneity is examined. Bringing together these two view-points remains a major challenge for the future.

493

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


Contents

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494EXPERIMENTAL EVOLUTION: EXAMPLES AND COUNTEREXAMPLES . . 495

Surprises in Long-Term Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495Evolutionary Trajectories of Individual Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496Fitness Landscapes and the Epistatic Genome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497Evolutionary Trade-Offs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497Adaptation to Fluctuating Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498Spatial Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500

INTERLUDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501QUANTITATIVE PHYSIOLOGY OF MICROBES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502

Exponential Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502Physiological Heterogeneity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503Stochastic Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504Aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505Collective Physiology: Quorums, Swarms, and Biofilms. . . . . . . . . . . . . . . . . . . . . . . . . . . 506

A FINAL EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509

INTRODUCTION

Microbes present our best chances to study the general principles of evolution. Short generationtimes, small genome sizes, and powerful genetic tools make for ideal organisms in laboratoryevolution experiments. Microbes can serve as model systems to study most of the major aspectsof evolution (37). Their vast species richness, physiological and behavioral sophistication, andintricate ecologies encompass a wide diversity of evolutionary questions.

Despite these advantages, the interpretation of microbial evolutionary experiments can oftenbe difficult. Numerous studies have shown that, when given sufficient rounds of evolution, mi-crobial populations eventually adapt to surmount diverse challenges. Beyond this nearly universalobservation, quantitative prediction of most aspects of the adaptation process remains beyond ourreach. For example, we cannot presently predict how long adaptation will take or in what formit will appear, how many mutations will occur, or how adapted the organisms will become. Asin many complex systems, one does not expect to predict trajectories precisely, because of theirintrinsically stochastic dynamics. Yet, with a theory in hand, one would like to predict correctlythe distribution of outcomes. Theory should reveal the structure of dynamical laws that govern theadaptation process while enabling the construction of accurate, predictive models of real systems.

Do dynamical laws even exist in complex evolutionary systems? In recent experiments on closedmicrobial ecosystems, species’ densities were continuously observed over the timescale of months.Data collected on replicate populations demonstrated that their independent trajectories comprisea meaningful ensemble whose dynamical laws could be extracted by analyzing fluctuations (36).

Understanding how evolutionary dynamics arises from organismal physiologies and interac-tions poses a challenging open question. In this review, we introduce the major components ofthis puzzle and survey quantitative studies that could provide a basis for its solution. This reviewis meant to inspire many new investigations—both theoretical and experimental—by presentinga wide survey of topics.

494 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


Fitness: a measure ofan individual’sexpected reproductiveoutput and survival

LTEE: long-termevolution experimentconducted on 12populations ofEscherichia coli

Stationary phase:a physiological statethat cells enter whennutrients have beenexhausted

EXPERIMENTAL EVOLUTION: EXAMPLES AND COUNTEREXAMPLES

Evolutionary experiments involve propagating populations over a large number of generationsunder controlled conditions (see Supplemental Figure 1; follow the Supplemental Materiallink from the Annual Reviews home page at http://www.annualreviews.org). Populations areassayed in time to determine changes in their composition. Comparisons with founder strains aremade to quantify the overall degree of adaptation. DNA sequencing of specific genes or entiregenomes is used to determine the mutational differences among evolved strains.

Although specific experiments differ in many respects, all seek to explain observed changesby developing appropriate models of organismal fitness. The concept of fitness is an abstraction,similar to the concept of energy in physics. As such, it comes in different flavors and is measuredin different ways, summarized in Supplemental Figure 1. Fitness is a property of individualorganisms; however, its measurement often involves large numbers of cells. For this reason, in anyevolution experiment, a model of population dynamics must be used to infer fitness values frommeasurements. Fitness measurements are therefore limited by both experimental noise and modelinaccuracy. In the following sections we discuss key evolution experiments, focusing on examplesin which theoretical predictions were verified experimentally, and on counterexamples in whichtheory failed to explain or predict observations.

Surprises in Long-Term Evolution

In a famous and still-running experiment begun in 1988, 12 populations of Escherichia coli havebeen propagated by daily dilution into fresh media (50). In this long-term evolution experiment(LTEE), growth conditions consist of a single utilizable carbon source, glucose, which limits themaximal daily population size (Figure 1a). A single day consists of six or seven generations (i.e.,cell doublings), followed by cells entering stationary phase until the next dilution. By now, thisexperiment has progressed for over 50,000 generations.

Whole genome sequences at six time points from the first 40,000 generations of a singlepopulation yielded some surprises (6). In the first 2,000 generations, this population’s average

10,000 20,000 30,000 40,0000

Mutators

Ara-2 Ara+3 Ara-4 Ara-1Ara+6

Citrateutilization

Ara-3

Cannibalism

Ara-2

Generations

λ phageresistance

Ara-5Ara+2Ara+1Ara-3

Mut

atio

ns

Relative fitness

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

0

10

20

30

40

50

10,0000 15,0005,000 20,000

Generations

a Major events in the long-term evolution experiment (LTEE) b Fitness and mutations in population Ara-1

00 20,000 40,000

200400600800

Figure 1Major events and dynamics in the long-term evolution experiment (LTEE). (a) The LTEE on 12 populations of Escherichia coli, labeledAra− (1–6) and Ara+ (1–6), has been running for over 50,000 generations (50). The time and population in which different adaptationsoccurred are indicated. (b) Dynamics of fitness ( green squares) and mutations (blue circles) in a single population. The inset showsmutations on a different scale, as this population acquired a mutator phenotype around 26,500 generations. Panel b is adapted withpermission from Reference 6.

www.annualreviews.org • Evolution in Microbes 495

Supplemental Material

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.

http://www.annualreviews.org

http://www.annualreviews.org/doi/suppl/10.1146/annurev-biophys-083012-130320


Mutator: a bacterialstrain with elevatedmutation rate

Phenotype:the physiological ormorphological state ofa cell

fitness had rapidly improved by about 25%. Subsequently, the adaptation rate slowed down, andby 20,000 generations the population had only reached a 45% improvement (see Figure 1b).Genome sequences, however, revealed that the rate at which mutations became established wasrelatively constant, in seeming contradiction with the declining rate of adaptation. Neutral theorypredicts a constant rate of establishment of new mutations, when these have negligible effects onfitness, yet several independent lines of evidence firmly rule out neutrality in these experiments.The observed behavior could be rationalized by assuming that the availability and fitness benefitof mutations are different in the early versus later parts of the experiment (see Fitness Landscapesand the Epistatic Genome, below).

Another surprise was the appearance of mutator phenotypes in five populations (51, 91). Earlierexperiments had shown that mutators can outcompete nonmutator strains (15, 16). Nevertheless,spontaneous emergence of mutators in long-term evolution was largely unexpected, as mutatorswere thought to suffer a high deleterious mutational load. Theory suggests that mutators may bemaintained at low levels within all populations and occasionally sweep to fixation with a beneficialmutation. Likewise, antimutators can subsequently outcompete mutators, leading to dramaticvariations in mutation rate during the evolution of a single population (22). Such dynamics havenot yet been observed, although the next 50,000 generations may hold new surprises.

While the 12 populations have been growing on glucose, the growth medium contained anothercarbon source, citrate, which these strains could not transport into the cell. Unexpectedly, onepopulation evolved the ability to grow on citrate (denoted by Cit+) around 33,000 generations(10). This caused a dramatic increase in population size, because cells could continue to groweven after glucose had been depleted. Most interestingly, coexistence of Cit+ and Cit− cells hasthereafter been maintained in this population. Upon dilution into fresh medium, Cit− cells initiallygrow faster on glucose but plateau at lower density than Cit+ cells. The constancy of daily freshmedium enables this long-term coexistence.

These experiments illustrate a breadth of evolutionary phenomena that can occur in even thesimplest environments. To understand their causes, further studies have sought to characterize indetail evolutionary trajectories and fitness effects, which we now describe.

Evolutionary Trajectories of Individual Mutations

The dynamics of individual mutations are difficult to measure in general, because they initially arisespontaneously in single cells within huge populations. Once mutants reach appreciable frequency,dynamics can be tracked in various ways. For example, by mixing equal proportions of bacterialcells labeled with two different tags, beneficial mutation trajectories were observed in the finalstages of their dynamics (35).

In yeast populations, clever experimental design allows earlier stages to be measured, startingfrom frequencies of 0.1% (47). This was achieved by constructing fluorescent markers for a generalphenotype, sterility, which occurs frequently by mutation and confers a significant fitness advan-tage when cultures are propagated asexually. Automated liquid handling and daily measurementsof marker frequencies allowed hundreds of populations to be propagated for 1,000 generations.From these trajectories, the initial fitness advantage of sterile mutants was measured to be 2% (onaverage). On their own, however, sterile mutations confer an average advantage of 0.6% or 1.5%,in two different genetic backgrounds that were used. This finding implies that within evolvingpopulations, sterile mutants reach appreciable frequencies when they appear in already fitter back-grounds, i.e., in combination with other mutations. These experiments support recent theoreticalanalyses of adaptation in the presence of multiple mutations (21, 28, 31, 85).

496 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


Epistasis: anyinteraction betweenmutations that affectsfitness

Fitness landscape:a function on thehigh-dimensionalmetric space of allgenomes that assigns afitness to each genome

Clone: a geneticallyhomogeneouspopulation derivedfrom a single cell

Fitness Landscapes and the Epistatic Genome

Given that combinations of mutations, rather than mutations in isolation, are a major determinantof evolutionary dynamics, studies have probed the nature of interactions between mutations,also known as epistasis. These interactions are the basis for the structure of a fitness landscape,an abstract high-dimensional space in which genomes are viewed as nodes connected by edgescorresponding to mutations. As populations evolve, they move across the landscape influencedby its fitness peaks and valleys. Direct analogies exist between populations dynamics on fitnesslandscapes and thermodynamics of physical ensembles on energy landscapes (20, 68).

To gauge the structure of fitness landscapes, strains carrying all possible combinations of severalmutations have been genetically constructed, and their fitness was measured (17, 38, 83, 103). Inone study, a single gene responsible for β-lactam antibiotic resistance in E. coli was mutated atup to five locations known to enhance resistance to the antibiotic cefotaxime, yielding 25 = 32different strains (103). The combination of all five mutations yielded a 10,000-fold increase inresistance. Among all possible mutational paths to attain these five mutations, it was found that85% of paths were likely to be extremely rare as they included steps that decreased fitness or didnot improve it (77). Different results were obtained in a second study, which investigated fourmutations that arose in separate genes when Methylobacterium extorquens was evolved to employ aforeign pathway for growth on methanol (17). In this case, all mutational pathways were strictlyuphill in fitness (i.e., the fitness landscape was smooth).

In these two studies, each mutation can exist in multiple backgrounds whose fitness had beenmeasured (17). Hence it was possible to assess whether the fitness of the genetic background is atall predictive of the change in fitness due to each mutation. The methanol experiments revealeddiminishing returns epistasis: It was found that the higher the background fitness, the lower theexpected gain of each mutation. In contrast, for β-lactam adaptation, the effect of a mutation wasnot predictable on the basis of the background fitness. A third study examined in a similar way thefirst five mutations to fix in the LTEE (38) and found that four mutations exhibited diminishingreturns while a fifth mutation exhibited synergy (i.e., the higher the background fitness, the largerthe mutation’s benefit).

Dramatic consequences of epistasis were seen in experiments initiated from two different typesof clones, EW (eventual winner) and EL (eventual loser), isolated from early stages of the LTEE(105). The EW clones had mutations that later became fixed in the population; however, earlyon these clones were less fit than EL clones by approximately 6%. Replicate evolution experi-ments started from each clone showed that, despite their initial fitness deficit, EW clones adaptedmuch faster than EL clones, generating detectible beneficial mutants significantly earlier in mostreplicates. This behavior was traced to a difference between EW and EL clones in the DNA topoi-somerase I gene (topA), which modulates DNA supercoiling with epistatic effects on many genes.

Theoretical analysis predicts how adaptation speed on fitness landscapes depends on the formof epistasis (45). When fitness landscapes exhibit antagonistic epistatic interactions, theory showsthat although adaptation slows down, mutations still accrue at a nearly constant rate. On this basis,the data from the LTEE (see Figure 1b) were used to infer the structure of the epistatic land-scape, which was consistent with strong antagonistic epistasis (45). Additional results on epistaticinteractions are discussed in Reference 71.

Evolutionary Trade-Offs

Because microbes must survive under a large range of environments, natural selection may befaced with trade-offs: Improved performance in one set of conditions could adversely affect


Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


(Bacterio)phage:a virus that infectsbacteria

performance in another. Diverse evolutionary adaptations have been explained on the basis oftrade-offs. Experiments in microbes allow several of these explanations to be tested.

The effect of temperature on growth rate has been observed in bacteria isolated from dif-ferent natural environments, with different species displaying characteristic temperature-growthprofiles (11). To investigate trade-offs, replicate populations of E. coli were grown at differenttemperatures (32◦C, 37◦C, or 42◦C) for 2,000 generations and then transferred to growth at 20◦Cfor an additional 2,000 generations (8). Fitness measurements showed that nearly all populationsimproved significantly in growth rate at 20◦C, but the degree of improvement had no correlationwith their ancestral growth temperature. The evolved lines were then assayed for growth at 40◦C,with surprising results. Whereas a majority of populations exhibited a trade-off, one-third did not,and one population exhibited a statistically significant anti-trade-off, showing fitness gains at bothtemperatures. Moreover, the magnitude of fitness gains at 20◦C exhibited no correlation with themagnitude of fitness losses at 40◦C.

Experiments that probed trade-offs for growth on two different carbon sources revealed similarresults (48): A trade-off was found in a majority of populations, yet one-fourth of populationsexhibited an anti-trade-off, improving in growth on both media. Similar results were found fortrade-offs related to aging in bacteria (3) (see Aging, below). From these experiments, it appears thatthe trade-off concept might be applicable only in some statistical sense—in any given experiment,the trade-off may or may not be apparent. While molecular constraints may impose an overallstatistical trade-off on the adaptive landscape, a significant number of evolutionary paths mightavoid trade-offs altogether.

The LTEE presents a beautiful illustration of the often counterintuitive nature of trade-offs.E. coli populations had evolved for 45,000 generations, adapting to growth on glucose in theabsence of bacteriophages. Whereas the ancestral strain was susceptible to infection by λ phage,the evolved strains had developed complete resistance to it despite never encountering λ phagein their evolution (63). The reason for this unexpected result was illuminating. λ Phage infectsE. coli through LamB, an outer membrane receptor that is used to import maltose into the cell.Adaptation to growth on glucose involved mutations that turned off the maltose pathway andits receptor, which subsequently led to λ phage resistance. Many additional examples regardingtrade-offs and optimality are seen in evolutionary studies of bacteriophage T7 (12).

Adaptation to Fluctuating Environments

Evolutionary trade-offs are most relevant when adaptation occurs under fluctuating environments.Nutrient fluctuations, for example, are one large class of environmental changes to which bacteriahave adapted. Expression and regulation of sugar operons, such as the lac operon, allow bacteria togrow on carbon sources other than glucose. When conditions change, for example when glucose isdepleted and lactose becomes available, cells express the lac operon, producing the permease LacY,which imports lactose into the cell, and the β-galactosidase LacZ, which hydrolyzes the lactosemolecule. In the absence of lactose, the operon is repressed by the repressor LacI to alleviate thecost of expression (23).

Experiments have begun to probe how regulation evolves under fluctuations. A recent studyfollowed 2,000 generations of E. coli populations growing under four different regimes, bothconstant and fluctuating: (a) constant glucose, (b) constant lactose, (c) constant glucose and lactose,and (d ) daily alternation of glucose and lactose (79). The evolved strains exhibited a variety ofchanges in the lac operon, from lower induction thresholds, to loss of bimodality, to complete lossof regulation. Different mutations in lacI were observed, as well as mutations in genes with no

498 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


0

1

2

–0.5

0.0

0.5

1.0

1.5

2.0

0.0

0.5

1.0

1.5

2.0

106 107

–1 0 1 2

108

106 107 108

Expression level (a.u.)

Trade-off curves

Growth rates in two environments

25 mg ml–1

40 mg ml–1

80 mg ml–1

0.15%

0.25%

0.40%

Antibioticconcentration

Sucroseconcentration

F cm

(h–1

)

F cm

(h–1

)

F sucrose

(h–1

)

Fsucrose (h–1)

Figure 2Trade-offs due to varying gene expression in two different environments (76). (Left) Growth rates ofEscherichia coli were measured at different levels of induction of the engineered operon, which confers bothresistance to chloramphenicol (cm) (top) and toxicity in the presence of sucrose (bottom). (Right) Themeasurements, replotted on the two growth-rate axes shown, yielded curves of different concavities. Graylines correspond to isoclines of the average growth rate when environments alternate for equal periods oftime. Arrows indicate the position of the optimal expression level, which is seen to be intermediate for thered curve and extreme for the blue curve. Adapted with permission from Reference 76.

previously known role in lactose metabolism. Interestingly, populations that grew in the alternatingenvironment typically evolved to constitutively express the operon, losing regulation.

Understanding trade-offs is key to explaining how a lack of regulation can be selectively ad-vantageous in fluctuating conditions. To this end, another recent study employed an engineeredoperon under the control of the lac promoter, which confers a cost or a benefit under two con-ditions (76). Expression of the operon provides resistance to the antibiotic chloramphenicol (cm)but leads to a toxic effect in the presence of sucrose. By artificially inducing the operon to differentexpression levels, it was possible to sensitively quantify growth rates Fsucrose and Fcm under the twoconditions (see Figure 2). As expression levels are changed, the growth rates trace out a curve inthe Fsucrose−Fcm plane. The shape of these curves, their concavity or convexity, depends sensitivelyon the concentrations of sucrose and chloramphenicol. And, importantly, the shape determineswhether the optimal expression level in fluctuating conditions will occur at an intermediate levelor at one of the extremes, which correspond to loss of regulation. These measurements formed


Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


a basis for highly successful selection experiments, in which an anti-Lac repressor (with invertedregulatory behavior) was evolved in three rounds of mutagenesis and selection (76).

Spatial Dynamics

Evolutionary dynamics can differ substantially between well-mixed populations and spatially dis-tributed or otherwise subdivided populations. In well-mixed populations, resources are global andthe rise of a beneficial mutation in a subpopulation can rapidly be felt by the rest of the population.In subdivided populations, resources are local and individuals are in direct competition only withtheir immediate neighbors. If a beneficial mutation arises in one part of the population, the restof the population may not feel any effect for a long time, because local resources remain unaf-fected. This increases the time to fixation for beneficial mutations while allowing the populationto maintain more genetic diversity.

To test whether the greater diversity afforded by subdivision could be beneficial, separatepopulations of yeast cells were propagated and periodically mixed (44). After 550 generationsof evolution at different migration rates between populations, fitness measurements showed thatwell-mixed populations were consistently fitter than subdivided populations. It was concludedthat the fitness landscape in these experiments must be relatively smooth, favoring fast fixation ofbeneficial mutations instead of more thorough exploration of the landscape.

Spatial trajectories of microbial evolution have been intensively studied over the past fiveyears, both theoretically and experimentally (32, 33, 41–43, 70). The experimental model consistsof microbial colonies of either bacteria or yeast cells inoculated on nutrient-rich plates and grownover several days. The initial inoculation contains a 50:50 mixture of two strains differing only ina neutral fluorescent marker. After growth, the plates are imaged and the spatial pattern of themarkers is analyzed.

Visually arresting patterns of spatially segregated domains of each marker appear, after aninitial growth period, and continue to propagate in time (see Figure 3a). This behavior, known

1 mma b

Figure 3Dynamics of spatial growth. (a) Two identical strains of Escherichia coli, fluorescently labeled green or red,were inoculated at the center. Their growth leads eventually to segregated domains at the colony edge, atwhich only a single strain (color) is present. (b) Beneficial mutations are seen via emergence of sectors.(Top) A green sector emerged spontaneously in a growing yeast colony consisting of a 5:1 mixture ofred:green cells. (Bottom) Sectors are seen when a strain with a known beneficial mutation (red ) was mixedwith wild-type cells ( green) in a 1:40 mutant–to–wild type ratio. Adapted with permission from Reference 33.

500 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


as genetic demixing, occurs because of random motion and reproduction at the edge of a growingcolony (32, 70). Physical models reveal that although demixing occurs at the colony boundary,the expansion of the boundary allows diversity to be maintained indefinitely (33, 41). The finitenumber of domains maintained can be computed exactly, and its predicted dependence on the sizeof the initial inoculation has been verified by experiments using two different bacterial species (43).Growth patterns of E. coli on plates have a chiral component, in which the entire population twistscounterclockwise when viewed from below the plate. The effect of this chirality was subtractedoff and the remaining motion of boundaries was seen to be diffusive (43).

In the presence of selective differences between strains, spatial growth patterns exhibit charac-teristic shapes, such as bulges, at the colony edge (Figure 3b). The boundary of these bulges takesthe form of logarithmic spirals, a theoretical prediction that was verified in experiments usingyeast cells (42). A labeled sterile yeast mutant was mixed at a low starting frequency with wild-typecells. When sterile sectors appeared, their opening angles and shapes allowed estimates of theirfitness advantage to be made.

These experiments reveal at a new level of detail how selection and drift are influenced byspatial structure, as well as the essential components of their theoretical description. At the sametime, they provide new tools for measuring selective differences between competing lineages.

INTERLUDE

By this point, thanks to the author’s biased selection of examples and counterexamples, the readermay be forming some opinions. On the one hand, examples of predictions verified by experimenthave been presented. On the other hand, counterexamples in which key experiments negate theo-retical expectations, or yield wholly unexpected results, have been discussed. Individual opinionson this state of affairs will vary and may include the following:

1. There is no problem with theory—evolutionary experiments are simply not sufficiently wellcontrolled.

2. There is no problem with experiments—theory is not (yet) sufficiently complex to accountfor all the phenomena.

3. There is no problem at all—for every observed evolutionary phenomenon, there alreadyexists a theory to explain its behavior.

4. There is a big problem—we have no real theory, and therefore we do not know whichexperiments to do.

One feature of purely mathematical models of evolutionary dynamics is that they abstract awayphysiology, replacing it with a collection of selective coefficients and a fitness landscape. Suchabstraction would appear to be a good thing: One would like to describe universal aspects ofevolution and ignore the messy details of biology. Physical theory is particularly good at suchabstraction. The equations of electromagnetism, for example, make no mention of details such aswires, light bulbs, or iPads. Yet the striking advantage of physical theory is that one knows howto put the details back in. Indeed, Maxwell’s equations provide the ultimate guide for derivingexplicit, predictive models of real systems. In mathematical evolutionary theory, abstraction isoften a one-way street: We do not know how to put the details back in.

Are there more physical routes to abstraction in evolutionary theory? In the second half of thisreview, we consider physiological aspects of microbial evolution as a starting point to inspire newtheoretical directions.


Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


Chemostat: a devicethat maintains abacterial population ina proliferative state at afixed growth rate

Turbidostat: a devicethat maintains abacterial population ina proliferative state at afixed density

QUANTITATIVE PHYSIOLOGY OF MICROBES

The quantitative basis of experimental evolution lies in the complex dependence of growth rateson physiological states and external conditions. We survey major facets of microbial physiologythat are critical for evolutionary experiments, focusing on recent quantitative studies with relevanttheoretical aspects.

Exponential Growth

Exponential growth is achieved by bacterial populations when nutrients are at sufficiently highconcentrations while cell densities remain sufficiently low. Such growth can be maintained inchemostats, turbidostats, and microfluidics devices. Monod’s experiments yielded a simple rela-tionship between steady-state growth rate, λ, and the concentration of a limiting nutrient N:

λ = λmax NN + K

, 1.

where λmax and K are parameters that depend on the specific nutrient as well as the growth medium(65). One rationalizes Monod’s law by imagining bacterial cells as nutrient absorbers that undergosimple molecular binding at specific receptors. Internal enzymes convert nutrients into cellularbiomass, which in simple cases may involve a single bottleneck reaction. The biomass flux is thendetermined by Michaelis-Menten kinetics, which leads to Monod’s law.

The simple form of this growth law hides much complexity. How do the parameters λmax andK depend on the internal state of the cell? What are the molecular determinants of the maximalgrowth rate, and how is it established? These questions were recently addressed on the basis oftwo general observations related to the (mass) ratio of cellular RNA to protein, r (see Figure 4):(a) The ratio r exhibits strong positive correlation with the growth rate λ when nutrient qualityis varied (87). (b) The ratio r exhibits strong negative correlation with λ when protein translationrates are varied, via antibiotics or genetic mutations (88).

ϕ Voltages

κ Conductances

λ Net current = the growth rate

b Increasing the metabolic conductance, κn

c Decreasing the ribosomal conductance, κt

a

Metabolicresistor

0.50.8

0.6

0.4

0.2

0

0.4

0.3

0.2

0.1

00 0.5 1 1.5 2

RNA

/pro

tein

rati

o, r

RNA

/pro

tein

rati

o, r

Growth rate, λ (h–1)

0 0.5 1 1.5 2

Growth rate, λ (h–1)

Ribosomalresistor Nutri

ent qualit

y

Chloramphenicollevel

κn

κt

ϕRmax

ϕP

λϕR

Figure 4Model of steady-state exponential cell growth and measurements under different nutrient and antibiotic conditions. (a) Ohm’s law canbe applied to describe exponentially growing cells (88). (b,c) Symbol shapes and colors correspond to a wide variety of nutrientconditions. In panel b, plot points correspond to different growth media; in panel c, plot points correspond to increasingchloramphenicol levels in different media (88). The (mass) ratio of RNA to protein, r, is directly proportional to φR , the fraction of theproteome dedicated to ribosomes. Because ribosomal RNA constitutes the dominant fraction of all bacterial RNA, stoichiometricrelations dictate that φR = ρ·r , where ρ is estimated to be ≈0.76. Adapted with permission from Reference 88.

502 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


These observations suggested that r, as a measure of the cell’s state, underlies fundamentalgrowth laws, which were derived in Reference 88 as discussed below. Indeed, r is directlyproportional to a key quantity, φR, the fraction of protein mass dedicated to ribosomes (seeFigure 4). Theoretical analysis was rooted in a single basic fact: At steady-state growth, allconcentrations must grow exponentially at the same rate λ. The rate of increase of protein mass,dM/dt, is proportional to the number of ribosomes, which itself is proportional to φR·M . Oneobtains exponential growth of M at the rate λ = κtφR, where κt is the translational capacity ofribosomes, a measure of the protein production rate.

Cell growth is not wholly dictated by the ribosomes, and one expects an upper bound on φR,which was measured by Scott et al. (88) to be φmax

R ≈ 0.55. Apparently, growth would entirelycease if more than 55% of cellular protein mass were devoted to ribosomes. On this basis, itwas postulated that the proteome is partitioned into (at least) three components: the ribosomalfraction φR, the flexible fraction φP , and the fixed fraction φQ = 1−φmax

R . The size of the flexiblefraction, where metabolic enzymes reside, would adjust to the nutritional demands of the growthmedia. Because growth on a limiting nutrient involves bottleneck enzymatic reactions, dM/dt isproportional to the number of enzymes, scaling as φP ·M . This yields the exponential growthrate λ = κnφP , where κn is the nutritional capacity, a parameter determined by enzymatic rates,nutrient concentrations, and medium composition.

The two different expressions obtained for λ are subject to the constraint φR + φP = φmaxR .

By analogy with Ohm’s law, then, φR and φP act as voltages across resistors wired in series, withconductance κt and κn, respectively; and λ corresponds to the net current through the circuit(see Figure 4). If we increase the nutritional capacity, we increase the net current, hence thevoltage φR across the ribosomal resistor increases. This reveals the positive correlation of φR

and λ when growth rates are manipulated by varying the medium composition. Alternatively, ifwe increase the translational capacity, the voltage across the ribosomal resistor drops, revealingthe negative correlation of λ and φR when translation rates are manipulated. Expressing the netcurrent explicitly yields

λ = φmaxR

κtκn

κt + κn, 2.

which is a reformulation of Monod’s law by quite different reasoning, and with the more generalquantity κn in place of nutrient concentration.

This elegant formulation could provide a conceptual and quantitative basis to describe evolu-tion during steady-state growth. Mutations may independently alter translational and nutritionalcapacities, κt and κn, as well the cell’s growth potential φmax

R ; or, they may simultaneously affectcombinations of these complex parameters. The basic ideas on the logic underlying regulation ofproteome partitions have been discussed (39) and could form a theoretical basis for prediction.Generalization for growth under fluctuating conditions, away from the steady state, is an openavenue for research.

Physiological Heterogeneity

Microbial populations are known to be heterogeneous. Cells within the population exhibit a rangeof internal states, or phenotypes, that influence their physiological behaviors. Phenotypic statesare determined by many factors, including expression levels and activities of genes, methylationstates of promoters, concentrations of nutrients, and mechanical properties of cells. Cells are ableto switch between phenotypic states, which can strongly influence their reproduction and survival.Under stressful conditions, for example, such as high temperatures or starvation, bacteria activatestress response pathways, which protect cells and increase survival rates.


Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


When cells divide, their phenotypic states can be passed on to their progeny. The stability ofstates—their heritability—can vary dramatically among phenotypes. Phenotypes that are codeddirectly by DNA, e.g., the allelic states of genes, are the most heritable because stability is limitedonly by the mutation rate (∼10−10 per base pair per division in E. coli ). However, certain DNAsequences (e.g., simple repeat sequences) are hypervariable. Such repeats can mutate at ratesspanning orders of magnitude, from 10−8 to 10−4 per cell division, and are observed to vary betweendifferent bacterial strains (57). Repeats in promoters and coding regions can lead to spontaneousphenotypic switching and maintenance of heritable heterogeneity (66, 98). Even nongeneticallydetermined phenotypes, such as the induction state of the lac operon (97), or methylation statesof promoters (56), can be stably inherited for hundreds of generations.

Timescales of heritable physiological variation within microbial populations span a wide con-tinuum (80). Phenotypic variation can develop quickly within populations, indeed, much fasterthan genetic variation, and can be beneficial. For example, slow growth appears to be a gen-erally protective state. Measurements in yeast cells have shown that genetically identical cellsexhibit a wide distribution of growth rates, and that slow-growing cells exhibit increased toleranceagainst stress such as heat killing (55). The existence of heritable physiological variation couldstrongly influence the outcome of evolution experiments (see A Final Example, below). Severalwell-characterized examples of physiological heterogeneity are discussed in the following section.

Stochastic Switches

Stochastic switches are molecular mechanisms that allow cells to spontaneously switch their phe-notypic states without the use of sensory responsive pathways. Such switches have been studiedwidely in the field of pathogenic bacteria, where they are often involved in evading host immuneresponses (66, 96). Quantitative analysis of stochastic switches has advanced rapidly in recent yearsthrough single-cell microscopy and microfluidics-based studies (59). Closely related theoreticalworks have analyzed the benefits of switches in fluctuating environments (46, 95, 104). In particu-lar, these studies predict that the long-term growth rate of stochastic switching bacteria will dependsensitively on their switching rates, hence the rates themselves could be evolutionarily tuned.

The persistence switch in E. coli is a stochastic switch that maintains a subpopulationof slow-growing persister cells at frequencies of 10−6 in wild-type strains (5). Its molecularmechanism was recently elucidated (84). Persister cells are highly tolerant of certain antibioticsand other stresses and enable populations to survive in conditions of fluctuating antibiotics (54)or bacteriophages (75). Another well-characterized stochastic switch of evolutionary significanceis the competence switch of Bacillus subtilis, which activates a phenotypic state that allows cellsto take up DNA from their environment (93). When starved, cells stochastically choose betweenactivating the competence state and turning on sporulation genes to begin a developmentalprogram whose completion results in a highly resistant spore.

In a few cases, competition experiments have been conducted using strains with differentswitching rates. In the case of persistence, two strains were used—wild type and the hipA7 high-persistence strain (5). A more complete set of competitions was performed using a syntheticstochastic switch constructed in yeast cells (1). In these experiments, performed in turbidostats,fast-switching strains outcompeted slow-switching strains when environmental fluctuations wererapid, whereas the reverse was seen when fluctuations were slow, consistent with theoreticalpredictions.

These studies suggested that it may be possible to evolve stochastic switches in the lab, byexposing populations to fluctuating selection. Evolution of a reversible stochastic switch wasachieved using Pseudomonas fluorescens, which exhibits different niche-specialized phenotypes with

504 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


altered colony morphologies. By directly selecting for novel colony morphology, picking individ-ual colonies, and propagating over multiple rounds, eventually a strain emerged that stochasticallyswitched between opaque and translucent colonies (7). The ability to switch was traced to a singlenucleotide change in the gene carB, which is involved in arginine and pyrimidine biosynthesis. Itsrole in switching has not yet been revealed.

Stochastic switching rates are typically much larger than mutation rates, yet much smaller thancell division rates. Hence, they yield heritable phenotypic variation within populations, whichselection can act upon, with potentially large effects on evolutionary outcomes. How can wedetect their presence within microbial populations? Recent theoretical analysis has shown that bymeasuring the variance of fitness between single-cell histories, one can detect stochastic switches.This analysis is based on a path-integral formulation of populations in which the history of eachcell (and its ancestors) constitutes a set of trajectories (49). Cell divisions observed along eachhistory can be used to infer the fitness of cells and assign a fitness to each history. Large varianceof this historical fitness across the population is a sensitive detector of stochastic switching events.The description of populations as thermodynamic ensembles of individual histories can also begeneralized to include the effects of aging and cell cycles (99), as discussed in the following section.

Aging

Aging, in the evolutionary context, is the reduction in average reproductive output as a functionof an individual’s age. In microbes, aging has been studied in both yeast and bacteria (see, e.g.,4, 90). Numerous molecular mechanisms underlie aging in microbes, including accumulation ofmisfolded aggregated proteins, toxic compounds, and other damage to cellular components thatis not efficiently repaired. Upon cell division, damage is partitioned between the two cells.

Damage repair and asymmetric damage partitioning are two strategies that cells use to coun-teract damage accumulation. Both require investment of energy. For repair processes, ATP-dependent chaperones help refold misfolded proteins, and degradation and new synthesis areused for damaged cellular components such as membranes. Asymmetric partitioning likewise re-quires ATP-dependent transport of aggregates, such that after cell division one cell will carry lessdamage; that is, it will be rejuvenated relative to the other cell (82).

How much energy should cells invest in repair versus asymmetric partitioning? Naively, thequestion appears to involve accounting for energy of transport versus repair, with the simpleoutcome that the cheaper process would be preferred. However, these two strategies reap theirbenefits in very different ways. Efficient repair maintains all cells at some mean damage level,reducing cell-to-cell fluctuations. Asymmetric damage partitioning increases cell-to-cell variabilityby rejuvenating only one of the two cells at each division. Within an actively dividing population,selection can act to dilute out the cells that carry more damage, as these cells divide slower thancells with less damage.

Simple models have been used to explore how the overall shape of the aging profiles andrepair parameters influence the degree to which asymmetry is evolutionarily favored (2, 24).Applying conclusions from models to experiments is a much more delicate problem. For example,evolutionary models of aging often assume the existence of trade-offs, such that faster reproductionrate at young age comes at the expense of slower growth at older ages. A laboratory evolutionexperiment in which strong selection was applied to favor mutations that increase growth ratein young cells, however, yielded mutants in which even very old cells grew significantly faster thanin the original strain (3).

Detailed measurements of bacterial aging have become possible by innovative applications ofmicrofluidics and microscopy. By continuously imaging cell divisions over multiday periods, it


Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


Biofilm: a dense,surface-bound,structured microbialcommunity

was found that E. coli cells have the ability to maintain uniform growth rates for at least 100 celldivisions (100). After approximately 30 divisions, however, cell mortality becomes a significantfactor that affects an increasing proportion of cells. Interestingly, in the first 8–10 cell divisions,a small reduction in growth rate was seen in different experiments (58, 92, 100), an observationinterpreted as evidence of cellular aging (92), physiological adaptation (100), or trade-offs betweendamage and segregation processes (14, 81).

The interplay between cellular energetics, physiological variability, and selection within pop-ulations lies at the heart of these experiments. To elucidate its underlying principles, a statisticalmechanical approach has been formulated (99): Single-cell histories make up a space of conforma-tions (or trajectories), selection acting on histories behaves analogously to energy, and the intrinsicvariability in the timing of cell divisions results in entropy. Optimal lineages, histories that optimizea trade-off between selection and entropy, determine the population’s growth rate, a free-energy-like quantity. The strength of selection acting at different ages can be directly determined from thedistribution of ages along optimal lineages. Understanding how cellular thermodynamics influ-ence, and are influenced by, the thermodynamics of lineages poses another challenging theoreticalquestion.

Collective Physiology: Quorums, Swarms, and Biofilms

The coexistence of different cellular phenotypes within populations provides opportunities foradditional physiological behaviors, mediated by interactions between cells. These behaviors canbe cooperative, if cells engage in collective activities, such as swarming, creating biofilms, orproducing light; or specialized, if cells are capable of better exploiting certain metabolites, such asthe citrate phenotypes discussed above; or antagonistic, for example, when cells produce toxins tokill competitors or when cells cannibalize their neighbors (86). Theoretical treatment of dynamicsof interacting cells has revealed a rich range of new behaviors, which can often be described ingame-theoretical terms (19, 26). In this section, we focus on collective behaviors, which are oftenregulated by or subject to the overall density of the population.

Cooperation and cheating. When cells appear to cooperate, the potential for cheating is alwayspresent. Cells that do not pay the physiological costs associated with a collective activity canoften profit from the investments of neighboring productive cells. For example, to metabolizesucrose, yeast cells produce the enzyme invertase, which hydrolyzes the disaccharide into glucoseand fructose. Because the reaction occurs outside the cytoplasm, the monosaccharides can diffuseaway, becoming a public good available to other cells. Experiments using cheater mutants that donot produce invertase show that nonproducers have a competitive advantage in the presence ofproducers (29). Conversely, when producers are a minority, they increase in numbers against thecheaters. A stable coexistence between the two strains is rapidly attained. In the presence of excessglucose, however, producers lose any advantage and can be driven to extinction.

How are cooperative interactions sustained in the presence of cheaters? If conditions are right,population subdivision can enhance cooperation. Although cheaters outcompete cooperatorswithin each subpopulation, the overall growth of different subpopulations is enhanced bytheir fraction of cooperators. In a regime of periodic mixing and subdivision, it is possiblefor cooperators to increase in the population at large while their fractions decrease withineach subpopulation. This scenario, known as Simpson’s paradox, was recently constructed in asynthetic microbial system (18), which used quorum-sensing autoinducer molecules as a publicgood.

506 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


BEYOND PHYSIOLOGY: ECOLOGIES, INSIDE AND OUT

Metagenomics analyses are dramatically revealing the ecological structures of bacterial communities (27). Theseanalyses have uncovered a prevalence of horizontal gene transfer in bacteria much higher than previously thought;hence purely asexual evolutionary dynamics are an inadequate representation of real bacterial evolution (25, 53, 89)and models incorporating recombination must be considered (72). Bioinformatic analyses suggest that horizontaltransfer can strongly influence the evolution of cooperation in bacteria (74). Studying ecological interactions inquantitative laboratory experiments is a major frontier for future research (36).

An entirely different ecology exists in microbes—an ecology of mobile genetic elements, such as transposons,phages, and plasmids, that can move around within genomes and between cells. Their influence on adaptation ishighly significant, as seen in studies of yeast adapting to nutrient-limited chemostats (30). The evolutionary pressuresthat mobile genetic elements experience and exert on their hosts are highly nontrivial (13). Several striking examplesinvolve toxin-antitoxin modules in bacteria (108) and restriction-modification systems (40, 78). The coevolution ofphage and bacteria presents an additional set of challenges regarding self-nonself recognition and immunity (67).Recently discovered CRISPR systems in bacteria (61), which provide immunity against phage as well as memoryof past infections, have inspired new theoretical works (52, 102). The world of bacterial mobile elements overflowswith fascinating evolutionary complexity.

Quorums. Population density is sensed by bacteria using genetic pathways known as quorum-sensing networks. These networks include genes for production, export, and sensing of small-molecule signals, known as autoinducers. Using specific receptors that bind autoinducers, quorum-sensing networks measure concentrations and respond at specific autoinducer levels to regulatedownstream genes. Quorum sensing was first discovered in connection with a specialized organof a Hawaiian squid that harbors a population of light-emitting Vibrio fischeri bacteria (73, 101).Because light production is beneficial to the squid host only when a large number of bacteria arepresent, the bacteria use quorum sensing to activate the behavior only at high density, saving thecost of light production when cells are at lower density.

Bacteria typically express more than one quorum-sensing system, each specialized for a givenautoinducer molecule. Information from multiple autoinducer signals is integrated by cells andused to make internal decisions regarding which sets of genes to activate or repress (60, 94). Forexample, Vibrio harveyi uses three different autoinducers: One is produced by a large number ofbacterial species, another may be specific only to Vibrio bacteria, and the third is specific to theV. harveyi species. Intriguingly, numerous examples of interspecies molecular warfare have beendescribed in which one species specifically destroys the quorum-signaling molecules of another(a process named quorum quenching).

What is the evolutionary logic underlying integration of quorum-sensing signals? What arethe meanings that evolution has devised for different signaling molecules in bacteria? Thesequestions remain under intensive investigation, using concepts from communication theory(9, 62). Beyond their evolutionary logic, the presence of quorum-sensing networks in bacteriahas immediate implications for experimental evolution. When cell densities remain constant (e.g.,in chemostats or turbidostats), there may be no pressure to maintain these systems. However, whendensities change dramatically over time (e.g., for experiments in batch culture or on plates), fitnesslandscapes may be modulated by density-dependent factors mediated through quorum-sensingsystems, which can affect hundreds of genes.


Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


rhlA–

(green)

Wild type(red)

1 cm

Figure 5Swarming of Pseudomonas aeruginosa on a plate. Wild-type cells are capable of rhamnolipid production andswarming, whereas rhlA− cells cannot produce rhamnolipids. Initial inoculations of each strain wereseparated by 2 cm. Once wild-type cells began swarming, rhlA− cells were able to use the availablerhamnolipids to swarm as well. Adapted with permission from Reference 107.

Swarms. In collective behaviors, the connection between individual physiological costs and fit-ness can be particularly subtle. Swarming motility in the bacterium Pseudomonas aeruginosa pro-vides a key example. In order to swarm, P. aeruginosa cells produce and secrete large amounts ofrhamnolipid biosurfactants, preparing a wet surface on which the collective motion can take place.Nonproducer cheater mutants, deficient in rhamnolipid production, are fully capable of swarmingon the secretions of wild-type cells (see Figure 5).

Because the physiological cost of rhamnolipid production is high, it was expected that cheaterswould have an advantage swarming when mixed with wild-type cells. However, after passagingmixtures of the two cell types for multiple days on swarming plates, no fitness cost was seen—thetwo populations were maintained in constant proportions (107). Much further analysis revealedthat the production rate of rhamnolipids, though regulated by quorum sensing, is ultimatelydetermined by growth rate rather than density. When essential nutrients for cell growth (e.g.,nitrogen) are depleted, cells turn on rhamnolipid production, using excess carbon to build thebiosurfactants.

The physiological cost of production therefore does not translate into a fitness cost, due to thismetabolically prudent behavior (107): Production is regulated to occur under conditions in whichcells are not actively growing. For this reason, although they can be maintained, cheater mutantscannot invade the wild-type population. Just how these cells manage to integrate quorum-sensingsignals with other signals to produce a precise measurement of their own growth rate remains atantalizing question.

Biofilms. Bacteria growing on surfaces are capable of forming dense, structured communitiesknown as biofilms. To establish a biofilm, many species produce and secrete extracellular poly-meric substances (EPS), which form an extracellular matrix structure supporting the cells. As withrhamnolipid synthesis in the case of swarming, EPS production involves fitness costs that could besubject to cheating by nonproducers. Recent experiments using mixtures of producers and non-producers demonstrate fitness costs of EPS production in liquid culture—i.e., in the absence of

508 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


actual biofilm structure—while revealing that producers rapidly outcompete nonproducers whengrowing within biofilms (69).

Remarkably, this behavior was previously predicted by detailed simulations of biofilms (106).Owing to spatial effects, the EPS structures themselves promoted the growth of producer cellsthrough formation of tower-like structures, very similar to the actual structures observed with con-focal microscopy (69). This enables producers to reach higher nutrient concentrations while suffo-cating their nonproducing neighbors. Rather than favoring nonproducing cheaters, competitionswithin biofilms may select for vigorous, efficient EPS production. Interestingly, experiments usingmixed-species biofilms have found that certain species exhibit specialized phenotypes, adapted forbiofilm growth in the presence of competitors, that can increase in frequency by selection (34).

A FINAL EXAMPLE

This review began with a series of key examples from evolutionary experiments and then shiftedto quantitative microbial physiology, surveying cellular processes that determine growth rates andform the basis for evolutionary selection.

A final example illustrates how physiological effects conspire to dramatically affect evolution.Populations of E. coli were coevolved with λ phage in a glucose-limited environment for 28days (64). Very quickly, within 8 days, mutants that downregulated the LamB outer membranereceptor, which the phage uses to infect cells, became fixed in the population. Notwithstanding thiscalamity for the phage, a tiny population of phage was still maintained after 8 days. These phageswere growing on a small subpopulation of E. coli that stochastically expresses the LamB receptor.Remarkably, sometime between 8 and 17 days, the phage population evolved an innovation: thenew ability to infect through OmpF, a different outer membrane receptor. Thus, the propagationof phage on a tiny subpopulation of E. coli maintained by stochastic switching enabled a majorevolutionary innovation. And this was only half the story (64).

In summary, elucidating the relationships between evolutionary dynamics and microbial phys-iology presents a series of worthy challenges for many future studies. Undoubtedly, achieving thisgoal will require the critical input of many different approaches across diverse fields and disciplines.

SUMMARY POINTS

1. Evolutionary experiments in microbes are probing the structure of fitness landscapes,testing quantitative predictions of adaptation rates and spatial dynamics.

2. Microbial physiology provides the mechanistic basis for growth rate differences on whichevolutionary selection acts.

3. Understanding quantitatively how selection and physiology interact remains a majorchallenge for future studies.

DISCLOSURE STATEMENT

The author is not aware of any affiliations, memberships, funding, or financial holdings that mightbe perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS

This work was supported by NIH grant R01-GM-097356.


Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


LITERATURE CITED

1. Acar M, Mettetal JT, van Oudenaarden A. 2008. Stochastic switching as a survival strategy in fluctuatingenvironments. Nat. Genet. 40:471–75

2. Ackermann M, Chao L, Bergstrom CT, Doebeli M. 2007. On the evolutionary origin of aging. AgingCell 6:235–44

3. Ackermann M, Schauerte A, Stearns SC, Jenal U. 2007. Experimental evolution of aging in a bacterium.BMC Evol. Biol. 7:126

4. Ackermann M, Stearns SC, Jenal U. 2003. Senescence in a bacterium with asymmetric division. Science300:1920

5. Balaban NQ, Merrin J, Chait R, Kowalik L, Leibler S. 2004. Bacterial persistence as a phenotypic switch.Science 305:1622–25

6. Barrick JE, Yu DS, Yoon SH, Jeong H, Oh TK, et al. 2009. Genome evolution and adaptation in along-term experiment with Escherichia coli. Nature 461:1243–47

7. Beaumont HJE, Gallie J, Kost C, Ferguson GC, Rainey PB. 2009. Experimental evolution of bet hedging.Nature 462:90–93

8. Bennett AF, Lenski RE. 2007. An experimental test of evolutionary trade-offs during temperature adap-tation. Proc. Natl. Acad. Sci. USA 104(Suppl. 1):8649–54

9. Bischofs IB, Hug JA, Liu AW, Wolf DM, Arkin AP. 2009. Complexity in bacterial cell-cell communi-cation: quorum signal integration and subpopulation signaling in the Bacillus subtilis phosphorelay. Proc.Natl. Acad. Sci. USA 106:6459–64

10. Blount ZD, Borland CZ, Lenski RE. 2008. Historical contingency and the evolution of a key innovationin an experimental population of Escherichia coli. Proc. Natl. Acad. Sci. USA 105:7899–906

11. Bronikowski AM, Bennett AF, Lenski RE. 2001. Evolutionary adaptation to temperature. VII. Effectsof temperature on growth rate in natural isolates of Escherichia coli and Salmonella enterica from differentthermal environments. Evolution 55:33–40

12. Bull JJ, Molineux IJ. 2008. Predicting evolution from genomics: experimental evolution of the bacterio-phage T7. Heredity 100:453–63

13. Burt A, Trivers R. 2006. Genes in Conflict: The Biology of Selfish Genetic Elements. Cambridge, MA: HarvardUniv. Press

14. Chao L. 2010. A model for damage load and its implications for the evolution of bacterial aging. PLoSGenet. 6:e1001076

15. Chao L, Cox EC. 1983. Competition between high and low mutating strains of Escherichia coli. Evolution37:125–34

16. Chao L, Vargas C, Spear BB, Cox EC. 1983. Transposable elements as mutator genes in evolution.Nature 303:633–35

17. Chou HH, Chiu HC, Delaney NF, Segre D, Marx CJ. 2011. Diminishing returns epistasis amongbeneficial mutations decelerates adaptation. Science 332:1190–92

18. Chuang JS, Rivoire O, Leibler S. 2009. Simpson’s paradox in a synthetic microbial system. Science323:272–75

19. Damore JA, Gore J. 2012. Understanding microbial cooperation. J. Theor. Biol. 299:31–4120. de Vladar HP, Barton NH. 2011. The contribution of statistical physics to evolutionary biology. Trends

Ecol. Evol. 26:424–3221. Desai MM, Fisher DS. 2007. Beneficial mutation selection balance and the effect of linkage on positive

selection. Genetics 176:1759–9822. Desai MM, Fisher DS. 2011. The balance between mutators and nonmutators in asexual populations.

Genetics 188:997–101423. Eames M, Kortemme T. 2012. Cost-benefit tradeoffs in engineered lac operons. Science 336:911–1524. Erjavec N, Cvijovic M, Klipp E, Nystrom T. 2008. Selective benefits of damage partitioning in unicellular

systems and its effects on aging. Proc. Natl. Acad. Sci. USA 105:18764–6925. Fraser C, Alm EJ, Polz MF, Spratt BG, Hanage WP. 2009. The bacterial species challenge: making

sense of genetic and ecological diversity. Science 323:741–46

510 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


26. Frey E. 2010. Evolutionary game theory: theoretical concepts and applications to microbial communities.Physica A 389:4265–98

27. Gilbert JA, Dupont CL. 2011. Microbial metagenomics: beyond the genome. Annu. Rev. Mar. Sci.3:347–71

28. Good BG, Rouzine IM, Balick DJ, Hallatschek O, Desai MM. 2012. Distribution of fixed beneficialmutations and the rate of adaptation in asexual populations. Proc. Natl. Acad. Sci. USA 109:4950–55

29. Gore J, Youk H, van Oudenaarden A. 2009. Snowdrift game dynamics and facultative cheating in yeast.Nature 459:253–56

30. Gresham D, Desai MM, Tucker CM, Jenq HT, Pai DA, et al. 2008. The repertoire and dynam-ics of evolutionary adaptations to controlled nutrient-limited environments in yeast. PLoS Genet. 4:e1000303

31. Hallatschek H, Korolev KS. 2009. Fisher waves in the strong noise limit. Phys. Rev. Lett. 103:10810332. Hallatschek O, Nelson DR. 2008. Gene surfing in expanding populations. Theor. Popul. Biol. 73:

158–7033. Hallatschek O, Nelson DR. 2010. Life at the front of an expanding population. Evolution 64:193–20634. Hansen SK, Rainey PB, Haagensen JAJ, Molin S. 2007. Evolution of species interactions in a biofilm

community. Nature 445:533–3635. Hegreness M, Shoresh N, Hartl D, Kishony R. 2006. An equivalence principle for the incorporation of

favorable mutations in asexual populations. Science 311:1615–1736. Dynamics withinclosed ecologies ismeasured, and analysisof fluctuations revealsits underlying laws.

36. Hekstra DR, Leibler S. 2012. Contingency and statistical laws in replicate microbial closedecosystems. Cell 149:1164–73

37. Kawecki TJ, Lenski RE, Ebert D, Hollis B, Olivieri I, Whitlock MC. 2012. Experimental evolution.Trends Ecol. Evol. 27:547–60

38. Khan AI, Dinh DM, Schneider D, Lenski RE, Cooper TF. 2011. Negative epistasis between beneficialmutations in an evolving bacterial population. Science 332:1193–96

39. Klumpp S, Zhang Z, Hwa T. 2009. Growth rate-dependent global effects on gene expression in bacteria.Cell 139:1366–75

40. Kobayashi I. 2001. Behavior of restriction-modification systems as selfish mobile elements and theirimpact on genome evolution. Nucleic Acids Res. 29:3742–56

41. Korolev KS, Avlund M, Hallatschek O, Nelson DR. 2010. Genetic demixing and evolution in linearstepping stone models. Rev. Mod. Phys. 82:1691–718

42. Korolev KS, Mueller MJI, Karahan N, Murray AW, Hallatschek O, Nelson DR. 2012. Selective sweepsin growing microbial colonies. Phys. Biol. 9:026008

43. Korolev KS, Xavier JB, Nelson DR, Foster KR. 2011. A quantitative test of population genetics usingspatiogenetic patterns in bacterial colonies. Am. Nat. 178(4):538–52

44. Kryazhimskiy S, Rice DP, Desai MM. 2012. Population subdivision and adaptation in asexual populationsof Saccharomyces cerevisiae. Evolution 66:1931–41

45. Kryazhimskiy S, Tkacik G, Plotkin JB. 2009. The dynamics of adaptation on correlated fitness landscapes.Proc. Natl. Acad. Sci. USA 106:18638–43

46. Kussell E, Leibler S. 2005. Phenotypic diversity, population growth, and information in fluctuatingenvironments. Science 309:2075–78

47. Lang GI, Botstein D, Desai MM. 2011. Genetic variation and the fate of beneficial mutations in asexualpopulations. Genetics 188:647–61

48. Lee M-C, Chou H-H, Marx CJ. 2009. Asymmetric, bimodal trade-offs during adaptation of methylobac-terium to distinct growth substrates. Evolution 63:2816–30

49. Leibler S, Kussell E. 2010. Individual histories and selection in heterogeneous populations. Proc. Natl.Acad. Sci. USA 107:13183–88

50. Lenski RE, Rose MR, Simpson SC, Tadler SC. 1991. Long-term experimental evolution in Escherichiacoli. 1. Adaptation and divergence during 2,000 generations. Am. Nat. 138:1315–41

51. Lenski RE, Winkworth CL, Riley MA. 2003. Rates of DNA sequence evolution in experimental popu-lations of Escherichia coli during 20,000 generations. J. Mol. Evol. 56:498–508

52. Levin BR. 2010. Nasty viruses, costly plasmids, population dynamics, and the conditions for establishingand maintaining CRISPR-mediated adaptive immunity in bacteria. PLoS Genet. 6:e1001171


Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


53. Levin BR, Cornejo OE. 2009. The population and evolutionary dynamics of homologous gene recom-bination in bacteria. PLoS Genet. 5:e1000601

54. Levin BR, Rozen DE. 2006. Non-inherited antibiotic resistance. Nat. Rev. Microbiol. 4:556–6255. Levy SF, Ziv N, Siegal ML. 2012. Bet hedging in yeast by heterogeneous, age-correlated expression of

a stress protectant. PLoS Biol. 10:e100132556. Lim HN, van Oudenaarden A. 2007. A multistep epigenetic switch enables the stable inheritance of

DNA methylation states. Nat. Genet. 39:269–7557. Lin WH, Kussell E. 2012. Evolutionary pressures on simple sequence repeats in prokaryotic coding

regions. Nucleic Acids Res. 40:2399–41358. Lindner AB, Madden R, Dernarez A, Stewart EJ, Taddei F. 2008. Asymmetric segregation of pro-

tein aggregates is associated with cellular aging and rejuvenation. Proc. Natl. Acad. Sci. USA 105:3076–81

59. Locke JCW, Elowitz MB. 2009. Using movies to analyse gene circuit dynamics in single cells. Nat. Rev.Microbiol. 7:383–92

60. Long T, Tu KC, Wang Y, Mehta P, Ong NP, et al. 2009. Quantifying the integration of quorum-sensingsignals with single-cell resolution. PLoS Biol. 7:640–49

61. Marraffini LA, Sontheimer EF. 2010. CRISPR interference: RNA-directed adaptive immunity in bacteriaand archaea. Nat. Rev. Genet. 11:181–90

62. Mehta P, Goyal S, Long T, Bassler BL, Wingreen NS. 2009. Information processing and signal inte-gration in bacterial quorum sensing. Mol. Syst. Biol. 5:325

63. Meyer JR, Agrawal AA, Quick RT, Dobias DT, Schneider D, Lenski RE. 2010. Parallel changes in hostresistance to viral infection during 45,000 generations of relaxed selection. Evolution 64:3024–34

64. Demonstrates theinfluence ofphysiologicalheterogeneity onevolutionarytrajectories.

64. Meyer JR, Dobias DT, Weitz JS, Barrick JE, Quick RT, Lenski RE. 2012. Repeatability andcontingency in the evolution of a key innovation in phage lambda. Science 335:428–32

65. Monod J. 1942. Recherches sur la Croissance des Cultures Bactriennes. Paris: Hermann & Cie66. Moxon R, Bayliss C, Hood D. 2006. Bacterial contingency loci: the role of simple sequence DNA repeats

in bacterial adaptation. Annu. Rev. Genet. 40:307–3367. Murray NE. 2002. Immigration control of DNA in bacteria: self versus non-self. Microbiology 148:3–2068. Mustonen V, Lassig M. 2009. From fitness landscapes to seascapes: non-equilibrium dynamics of selec-

tion and adaptation. Trends Genet. 25:111–1969. Nadell CD, Bassler BL. 2011. A fitness trade-off between local competition and dispersal in Vibrio cholerae

biofilms. Proc. Natl. Acad. Sci. USA 108(34):14181–8570. Nadell CD, Foster KR, Xavier JB. 2010. Emergence of spatial structure in cell groups and the evolution

of cooperation. PLoS Comput. Biol. 6:E100071671. Nagel AC, Joyce P, Wichman HA, Miller CR. 2012. Stickbreaking: A novel fitness landscape model

that harbors epistasis and is consistent with commonly observed patterns of adaptive evolution. Genetics190:655–67

72. Neher RA, Shraiman BI, Fisher DS. 2010. Rate of adaptation in large sexual populations. Genetics184:467–81

73. Ng W-L, Bassler BL. 2009. Bacterial quorum-sensing network architectures. Annu. Rev. Genet. 43:197–222

74. Nogueira T, Rankin DJ, Touchon M, Taddei F, Brown SP, Rocha EPC. 2009. Horizontal gene transferof the secretome drives the evolution of bacterial cooperation and virulence. Curr. Biol. 19:1683–91

75. Pearl S, Gabay C, Kishony R, Oppenheim A, Balaban NQ. 2008. Nongenetic individuality in the host-phage interaction. PLoS Biol. 6:957–64

76. Using a syntheticconstruct, trade-offs aredirectly controlled andused to evolve ananti-Lac repressor.

76. Poelwijk FJ, de Vos MGJ, Tans SJ. 2011. Tradeoffs and optimality in the evolution of generegulation. Cell 146:462–70

77. Poelwijk FJ, Kiviet DJ, Weinreich DM, Tans SJ. 2007. Empirical fitness landscapes reveal accessibleevolutionary paths. Nature 445:383–86

78. Qian L, Kussell E. 2012. Evolutionary dynamics of restriction site avoidance. Phys. Rev. Lett. 108:15810579. Quan S, Ray JCJ, Kwota Z, Duong T, Balazsi G, et al. 2012. Adaptive evolution of the lactose utilization

network in experimentally evolved populations of Escherichia coli. PLoS Genet. 8:e1002444

512 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


80. Rando OJ, Verstrepen KJ. 2007. Timescales of genetic and epigenetic inheritance. Cell 128:655–6881. Rang CU, Peng AY, Chao L. 2011. Temporal dynamics of bacterial aging and rejuvenation. Curr. Biol.

21:1813–1682. Rokney A, Shagan M, Kessel M, Smith Y, Rosenshine I, Oppenheim AB. 2009. E. coli transports aggre-

gated proteins to the poles by a specific and energy-dependent process. J. Mol. Biol. 392:589–60183. Rokyta DR, Joyce P, Caudle SB, Miller C, Beisel CJ, Wichman HA. 2011. Epistasis between beneficial

mutations and the phenotype-to-fitness map for a ssDNA virus. PLoS Genet. 7:e100207584. Rotem E, Loinger A, Ronin I, Levin-Reisman I, Gabay C, et al. 2010. Regulation of phenotypic vari-

ability by a threshold-based mechanism underlies bacterial persistence. Proc. Natl. Acad. Sci. USA 107:12541–46

85. Rouzine IM, Wakeley J, Coffin JM. 2003. The solitary wave of asexual evolution. Proc. Natl. Acad. Sci.USA 100:587–92

86. Rozen DE, Philippe N, Arjan de Visser J, Lenski RE, Schneider D. 2009. Death and cannibalism in aseasonal environment facilitate bacterial coexistence. Ecol. Lett. 12:34–44

87. Schaechter M, Maaloe O, Kjeldgaard NO. 1958. Dependency of medium and temperature of cell sizeand chemical composition during balanced growth of Salmonella typhimurium. J. Gen. Microbiol. 19:592–606

88. Bacterialexponential growth isanalyzed by analogywith Ohm’s law.

88. Scott M, Gunderson CW, Mateescu EW, Zhang Z, Hwa T. 2010. Interdependence of cell growthand gene expression: origins and consequences. Science 330:1099–102

89. Shapiro BJ, Friedman J, Cordero OX, Preheim SP, Timberlake SC, et al. 2012. Population genomics ofearly events in the ecological differentiation of bacteria. Science 336:48–51

90. Shcheprova Z, Baldi S, Frei SB, Gonnet G, Barral Y. 2008. A mechanism for asymmetric segregation ofage during yeast budding. Nature 454:728–34

91. Sniegowski PD, Gerrish PJ, Lenski RE. 1997. Evolution of high mutation rates in experimental popu-lations of E. coli. Nature 387:703–5

92. Stewart EJ, Madden R, Paul G, Taddei F. 2005. Aging and death in an organism that reproduces bymorphologically symmetric division. PLoS Biol. 3:295–300

93. Suel GM, Garcia-Ojalvo J, Liberman LM, Elowitz MB. 2006. An excitable gene regulatory circuitinduces transient cellular differentiation. Nature 440:545–50

94. Teng S-W, Schaffer JN, Tu KC, Mehta P, Lu W, et al. 2011. Active regulation of receptor ratios controlsintegration of quorum-sensing signals in Vibrio harveyi. Mol. Syst. Biol. 7:491

95. Thattai M, van Oudenaarden A. 2004. Stochastic gene expression in fluctuating environments. Genetics167:523–30

96. van der Woude MW, Baumler AJ. 2004. Phase and antigenic variation in bacteria. Clin. Microbiol. Rev.17:581–611

97. Vilar JMG, Guet CC, Leibler S. 2003. Modeling network dynamics: the lac operon, a case study. J. CellBiol. 161:471–76

98. Vinces MD, Legendre M, Caldara M, Hagihara M, Verstrepen KJ. 2009. Unstable tandem repeats inpromoters confer transcriptional evolvability. Science 324:1213–16

99. Wakamoto Y, Grosberg AY, Kussell E. 2012. Optimal lineage principle for age-structured populations.Evolution 66:115–34

100. Measures growthand cell death in E. colicells with a novelmicrofluidics device.

100. Wang P, Robert L, Pelletier J, Dang WL, Taddei F, et al. 2010. Robust growth of Escherichiacoli. Curr. Biol. 20:1099–103

101. Waters CM, Bassler BL. 2005. Quorum sensing: cell-to-cell communication in bacteria. Annu. Rev. CellDev. Biol. 21:319–46

102. Weinberger AD, Sun CL, Plucinski MM, Denef VJ, Thomas BC, et al. 2012. Persisting viral sequencesshape microbial CRISPR-based immunity. PLoS Comput. Biol. 8:e1002475

103. Weinreich DM, Delaney NF, DePristo MA, Hartl DL. 2006. Darwinian evolution can follow only veryfew mutational paths to fitter proteins. Science 312:111–14

104. Wolf DM, Vazirani VV, Arkin AP. 2005. Diversity in times of adversity: probabilistic strategies inmicrobial survival games. J. Theor. Biol. 234:227–53

105. Woods RJ, Barrick JE, Cooper TF, Shrestha U, Kauth MR, Lenski RE. 2011. Second-order selectionfor evolvability in a large Escherichia coli population. Science 331:1433–36


Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


106. Xavier JB, Foster KR. 2007. Cooperation and conflict in microbial biofilms. Proc. Natl. Acad. Sci. USA104:876–81

107. Xavier JB, Kim W, Foster KR. 2011. A molecular mechanism that stabilizes cooperative secretions inPseudomonas aeruginosa. Mol. Microbiol. 79:166–79

108. Yamaguchi Y, Inouye M. 2011. Regulation of growth and death in Escherichia coli by toxin-antitoxinsystems. Nat. Rev. Microbiol. 9:779–90

514 Kussell

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.

BB42-Frontmatter ARI 13 April 2013 7:12

Annual Review ofBiophysics

Volume 42, 2013Contents

Doing Molecular Biophysics: Finding, Naming, and Picturing SignalWithin ComplexityJane S. Richardson and David C. Richardson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Structural Biology of the ProteasomeErik Kish-Trier and Christopher P. Hill � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �29

Common Folds and Transport Mechanisms of SecondaryActive TransportersYigong Shi � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �51

Coarse-Graining Methods for Computational BiologyMarissa G. Saunders and Gregory A. Voth � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �73

Electrophysiological Characterization of MembraneTransport ProteinsChristof Grewer, Armanda Gameiro, Thomas Mager, and Klaus Fendler � � � � � � � � � � � � � � �95

Entropy-Enthalpy Compensation: Role and Ramifications inBiomolecular Ligand Recognition and DesignJohn D. Chodera and David L. Mobley � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 121

Molecular Mechanisms of Drug Action: An Emerging ViewJames M. Sonner and Robert S. Cantor � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 143

The Underappreciated Role of Allostery in the Cellular NetworkRuth Nussinov, Chung-Jung Tsai, and Buyong Ma � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 169

Structural Insights into the Evolution of the Adaptive Immune SystemLu Deng, Ming Luo, Alejandro Velikovsky, and Roy A. Mariuzza � � � � � � � � � � � � � � � � � � � � � 191

Molecular Mechanisms of RNA InterferenceRoss C. Wilson and Jennifer A. Doudna � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 217

Molecular Traffic Jams on DNAIlya J. Finkelstein and Eric C. Greene � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 241

vii

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


Advances, Interactions, and Future Developments in the CNS, Phenix,and Rosetta Structural Biology Software SystemsPaul D. Adams, David Baker, Axel T. Brunger, Rhiju Das, Frank DiMaio,

Randy J. Read, David C. Richardson, Jane S. Richardson,and Thomas C. Terwilliger � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 265

Considering Protonation as a Posttranslational ModificationRegulating Protein Structure and FunctionAndre Schonichen, Bradley A. Webb, Matthew P. Jacobson, and Diane L. Barber � � � � � 289

Energy Functions in De Novo Protein Design: Current Challengesand Future ProspectsZhixiu Li, Yuedong Yang, Jian Zhan, Liang Dai, and Yaoqi Zhou � � � � � � � � � � � � � � � � � � � � � 315

Quantitative Modeling of Bacterial Chemotaxis: Signal Amplificationand Accurate AdaptationYuhai Tu � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 337

Influences of Membrane Mimetic Environments on MembraneProtein StructuresHuan-Xiang Zhou and Timothy A. Cross � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 361

High-Speed AFM and Applications to Biomolecular SystemsToshio Ando, Takayuki Uchihashi, and Noriyuki Kodera � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 393

Super-Resolution in Solution X-Ray Scattering and Its Applications toStructural Systems BiologyRobert P. Rambo and John A. Tainer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 415

Molecular Basis of NF-κB SignalingJohanna Napetschnig and Hao Wu � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 443

Regulation of Noise in Gene ExpressionAlvaro Sanchez, Sandeep Choubey, and Jane Kondev � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 469

Evolution in MicrobesEdo Kussell � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 493

Protein Structure Determination by Magic-Angle Spinning Solid-StateNMR, and Insights into the Formation, Structure, and Stability ofAmyloid FibrilsGemma Comellas and Chad M. Rienstra � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 515

Structural Studies of RNase PAlfonso Mondragon � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 537

On the Universe of Protein FoldsRachel Kolodny, Leonid Pereyaslavets, Abraham O. Samson, and Michael Levitt � � � � � � 559

viii Contents

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.


Torque Measurement at the Single-Molecule LevelScott Forth, Maxim Y. Sheinin, James Inman, and Michelle D. Wang � � � � � � � � � � � � � � � � 583

Modeling Gene Expression in Time and SpacePau Rue and Jordi Garcia-Ojalvo � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 605

Mechanics of Dynamin-Mediated Membrane FissionSandrine Morlot and Aurelien Roux � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 629

Nanoconfinement and the Strength of BiopolymersTristan Giesa and Markus J. Buehler � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 651

Solid-State NMR of Nanomachines Involved in PhotosyntheticEnergy ConversionA. Alia, Francesco Buda, Huub J.M. de Groot, and Jorg Matysik � � � � � � � � � � � � � � � � � � � � � � 675

Index

Cumulative Index of Contributing Authors, Volumes 38–42 � � � � � � � � � � � � � � � � � � � � � � � � � � � 701

Errata

An online log of corrections to Annual Review of Biophysics articles may be found athttp://biophys.annualreviews.org/errata.shtml

Contents ix

Ann

u. R

ev. B

ioph

ys. 2

013.

42:4

93-5

14. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsid

ad d

e Pu

erto

Ric

o -

Rec

into

May

ague

z C

ampu

s on

01/

23/1

4. F

or p

erso

nal u

se o

nly.

Evolution in Microbesacademic.uprm.edu/~lrios/4368/Evol_in_microbes2013.pdfof evolution (37). Their...

Documents

Transcript of Evolution in Microbesacademic.uprm.edu/~lrios/4368/Evol_in_microbes2013.pdfof evolution (37). Their...