Experimental Evolution

Most experiments in microbial evolution are conceptually simple. Populations are established (often from single clones), then propagated in a controlled and reproducible environment for many generations. A sample of the ancestral population is stored indefinitely (for example, frozen at –80 °C), as are samples from various time points in the experiment. After a population has been propagated for some time, the ancestral and derived genotypes can be compared with respect to any genetic or phenotypic properties of interest, which provides information on the dynamics of the evolutionary process and the extent of evolutionary change. Importantly, adaptation can be quantified by measuring changes in FITNESS in the experimental environment, in which fitness reflects the propensity to leave descendants. — Santiago F. Elena & Richard E. Lenski (2003)

Microbial evolution experiments are important for a number of reasons. First, they are a good way to characterize organisms towards understanding the ecological function of their adaptations. Second, microbes are useful model organisms for ascertaining basic evolutionary principles. This occurs as a consequence of easily attained large population sizes, ease of handling (including archiving) along with relative ease of molecular characterization, small genomes that allow ready full-genome sequencing, asexuality, etc. (Elena and Lenski, 2003). Third, directed evolution is an important tool in industrial as well as medical microbiology, where specific properties in organisms may be enhanced or reduced such as toward vaccine development. In this section I briefly consider the second aspect, experimental microbial evolution done for the sake of better understanding evolution, including, perhaps especially, microbial evolution.

In experimental evolution studies, evolutionary change is most easily tracked in terms of new alleles becoming fixed, where the fixing of new alleles is an important measure of evolutionary change. This occurs first because fixation is not instantaneous but instead follows a delay. Indeed, these delays can be quite long between mutational generation of a new allele and its fixation, indeed if it is ever fixed, with establishment of a polymorphism coming in between.

The mechanisms that can contribute to such delay during experimental microbial evolution studies are discussed by Elena and Lenski (2003). The first consideration is that even alleles that display greater fitness than other alleles that are found within a population most likely will display long delays between generation and fixation, even when considering selection alone. These delays occur because typically the fitness differences between these greater-benefit alleles and their competitor alleles will be relatively small and therefore, per generation, changes in frequency of alleles also will be relatively small.

A second consideration is that drift can operate even in very large populations, especially on newly arising alleles. That is, despite the potentially large size of a population, the population size of a newly arising allele will be just one. Indeed, with population sampling as in the course of serial transfer experiments, the likelihood of an allele’s transfer will be a function of frequency of the allele within the population. For example, for an allele present in a single copy, if 10% of a population is transferred, then in nine out of ten transfers the allele would be lost (whereas in one out of ten transfers, on average, the allele would be increased in frequency tenfold). The result is that even given multiple possible alleles that could give rise to increases in a population’s average fitness, if a substantial number of those alleles are lost to genetic drift following their mutational generation, then the rate at which the population fixes such fitness enhancing alleles will be inherently slower.

A third consideration is the process of clonal interference, which is called clonal because it is something that operates with greater efficiency the lower the potential for sex to separate alleles from each other. Here fixation can be delayed, potentially indefinitely, because different alleles at different loci can all contribute to fitness, resulting in numerous genotypes with similar fitnesses and therefore a lower potential for any given genotype to rise in frequency to fixation. Indeed, via periodic selection and genetic hitchhiking, a given allele, even if highly beneficial individually, could be lost from a population because it happens to be found within an organism whose aggregate fitness, that stemming from all of its alleles, is relatively low.

These latter points can be restated in terms of adaptive landscapes. That is, different genotypes rising in frequency to fixation can take different paths up the same adaptive peak (i.e., convergence to a common, high benefit genotype may occur). Alternatively, different genotypes may take paths up different adaptive peaks (i.e., convergence to a single, high benefit genotype does not occur). If the different genotypes do not display differences in fitness during these climbs up either the same or different peaks, or instead if they reach to the top of different but nonetheless similarly tall peaks, then there will be little to distinguish the the genotypes selectively resulting in no one genotype outcompeting the others and thereby becoming fixed within the population.

Lastly, it is expected that there exist more mutations conferring small fitness gains than there will be mutations conferring large fitness gains. Since the latter are more likely to rise to fixation, the total number of fixation events, especially per locus, is expected to be relatively low in number. Thus, beneficial mutations likely occur quite often whereas fixation events, by contrast, are comparatively rare. In addition, fixation events can be a consequence not just of deterministic mechanisms but also of stochastic ones such as hitchhiking , the latter even in relatively large populations. Thus, even when alleles conferring substantial benefits become fixed within clonal populations, they may at the same time confer fixation on lower benefit or even detrimental alleles, thereby reducing the overall fitness benefit to the population of fixing the substantially beneficial allele.

Passage Through Time

Evolution inherently possesses a time variable and experimental microbial evolution is observed over the course of the propagating of microorganisms over time. Natural evolution, too, occurs in the course propagation of organisms over time. The means by which this propagation occurs results in a diversity of ecological as well as evolutionary circumstances that can result in a variety of evolutionary outcomes.

With microbial evolution as this occurs in the laboratory, the two basic approaches are batch versus continuous culture. In either case it is often true that organisms are being propagated within environments that differ from those in which the organisms either originally evolved or had been propagated previously within the laboratory. If allowed to, these populations will adapt to their new circumstances, and with clonal organisms the result will be numerous genetically diverging lineages all of which are in competition for prominence with the larger population. What determines the winners of these competitions can be summarized in terms particularly of the fitness associated with individual genotypes.

Various mechanisms can delay individual genotypes from coming to the fore within these populations, such as attaining fixation. Prior to fixation, populations can be viewed as continuously diverging genetically while also losing to extinction less fit as well as especially rare genotypes. This divergence halts, temporarily, when one genotype becomes fixed within the population since at that this point all genetic variation is either eliminated, other than the fixed genotype, or otherwise is reduced to very low levels. Also at this point, however, mutational divergence again takes place, starting predominantly with the fixed genotype, though to a lesser degree with rare, additional genotypes as well.

Again there will be jockeying among genotypes towards fixation, with mutational acquisition of beneficial alleles increasing the likelihood of that fixation occurring. Ultimately, as the population become increasingly well adapted to its environment, bigger-benefit mutations will become increasingly rare, as too can the rate of increase of population fitness. It is likely that these various mechanisms occur as well in nature, though within ecologically more complicated environments, and with more profound consequences than simply fixation of particularly fit genotypes (e.g., evolution of abilities to digest novel substrates such as pollutants, evolution of pathogen virulence, or indeed evolution of microorganisms into macroorganisms). In addition, these basic pathways towards adaptation that are based on steady accumulation of beneficial mutations can be substantially accelerated via the acquisition of beneficial alleles from other lineages via horizontal gene transfer.