Darwin errou feio fundamentalmente sobre a seleção natural: a evolução é um processo guiado e com previsão

terça-feira, maio 24, 2011

Evolutionary Genetics: Evolution with Foresight

Current Biology, Volume 21, Issue 10, R398-R400, 24 May 2011

Copyright © 2011 Elsevier Ltd All rights reserved.

10.1016/j.cub.2011.04.023

Authors

Merijn L.M. Salverda,J. Arjan G.M. de Visser

See Affiliations

Summary

Evolution has no foresight, but produces ad hoc solutions by tinkering with available variation. A new study demonstrates how evolution nevertheless prepares organisms for the future by increasing their evolvability.

Evolution produces ad hoc solutions for present problems rather than perfect designs for future needs. Evolution cannot follow a preconceived plan, because it lacks foresight. In the words of Francois Jacob: “evolution works like a tinkerer, not like an engineer [1]”. Examples like the vertebrate eye, where light passes through layers of nerves and blood vessels before reaching the retina, show the short-sightedness of evolution. A recent study [2] nevertheless demonstrates that evolution has ways to prepare organisms for the future.

How can this be? Evolution cannot predict future conditions, but it may produce organisms with an increased capacity to adapt to novel conditions, that is, with higher ‘evolvability’ [3]. Clearly, the intrinsic ability of organisms to evolve varies, both at short and longer time scales. For instance, microorganisms with increased mutation rate show faster short-term adaptation under some conditions [4]. 

Longer-term evolvability depends not only on the ability to produce variation, but also on how accumulated mutations interact to produce new adaptive phenotypes and functions. These interactions can be visualized by a fitness landscape, which plots fitness for all possible combinations of a set of mutations [5]. The genetic architecture of the organism determines whether its fitness landscape in a given environment is complex with multiple maxima or rather simple with few maxima and a smooth surface. Complex fitness landscapes impose stronger constraints on adaptation and decrease evolvability [3]. How much these constraints affect evolvability also depends on the efficiency of natural selection, thus on population dynamic parameters such as population size and mutation rate. Large populations more rapidly produce variants carrying multiple mutations that can evade constraints such as fitness valleys [6].

Given that there is genetic variation in evolvability, how can it evolve? This is not straightforward, as natural selection benefits organisms with high fitness and not those with increased evolutionary potential. In order to evolve by natural selection, variants with increased evolvability must be associated with direct or indirect fitness benefits. Direct positive effects on offspring fitness are unlikely, at least for short-term evolvability, because genotypes that produce relatively many beneficial mutations tend to be those with relatively low fitness [7]. Variants with increased evolvability thus rely on longer-term benefits arising from the association with rare beneficial mutations, which they produce at an increased rate. Such second-order selection due to hitchhiking with beneficial mutations (Figure 1) is also the mechanism by which mutators, i.e. mutants with an increased mutation rate, reach high frequency in microbial populations [8].

In the new study, Woods et al.[2] report a detailed demonstration of second-order selection of evolvability in a large population of the bacterium Escherichia coli. This population is part of an ongoing long-term evolution experiment with 12 populations in the laboratory of Richard Lenski that has been running for more than 20 years, equalling 50,000 generations [9]. Population samples from different time points were archived in the freezer, which allows them to go back and study particular instances of evolution with ‘replay’ experiments. This time, the authors wanted to understand the long-term fate of different clones, each carrying a different set of beneficial mutations, present at generation 500 in one of the 12 populations. After 1,500 generations, two of these mutations (in the genes topA and rbs) had become fixed, while others had perished. Woods and co-workers [2] thus divided the 500-generation clones into ‘eventual winners’ (EWs) and ‘eventual losers’ (ELs), based on whether or not they carried the topA and rbs mutations.

The expected scenario was that EW clones were already more fit than the EL clones at generation 500, but competition experiments showed that actually the opposite was the case. How, then, could the descendants of the EWs eventually have won the battle? Was this pure luck, or did these genotypes somehow have a higher potential to evolve? Replaying the experiment many times starting with the 500-generation EWs and ELs showed that the EWs indeed beat the ELs most of the time. As the authors found no evidence of altered mutation rates in either of the two types, they suspected a difference in interactions between the genetic background of both clones and new beneficial mutations. Using whole-genome sequencing and mutant construction, Lenski's team succeeded in identifying the epistatic cause of the observed pattern. It turned out that the ELs, like the EWs, had a beneficial mutation in topA, but in an amino acid adjacent to the one altered in the EW topA mutation — a seemingly trivial molecular difference, but one with far-reaching consequences. Another beneficial mutation (in spoT) that appeared between generation 500 and 1,000 turned out to be beneficial in the background of the EW topA allele, but neutral in the background of the EL topA mutation. Differences in epistatic interactions thus caused a difference in the evolvability of EW and EL clones.

What is particularly beautiful about the study by Woods et al.[2] is that it illustrates in detail how the selection of clones with increased evolvability depends crucially on the interplay between fitness landscape and population dynamics. Given the complex fitness landscape observed, the size and mutation rate of the evolving population set the limits for second-order selection of clones with increased evolvability. Had the population been smaller, the spoT mutation that rescued the EW clone might not have occurred before the clone went extinct, turning the clone into an ‘eventual loser’ instead of the topic of a research project. As the authors mention, a minimal requirement for second-order selection of evolvability is the simultaneous presence of multiple contending clones carrying different beneficial mutations. Beyond that, population size and mutation rate determine how far evolution can look into the future, that is, how many new beneficial mutations are allowed to accumulate before the EW clone fixes. These results support previous claims that there is ample opportunity for higher-order selection of evolvability in microbial populations, since often multiple beneficial mutations accumulate before they fix [10].

Several other studies have found evidence for the complexity of real fitness landscapes [11,12,13,14], and for the importance of population dynamic parameters for adaptation on these landscapes. For instance, it was found that small bacterial populations sometimes reached higher fitness than populations 50-fold larger in size, despite their lower fitness early on [15]. These results were explained by assuming that large populations adapt by using bigger-effect mutations — those that survive the competition [16] — which would sometimes lead to local maxima. Small populations use different mutations each time, some of which would lead to higher fitness maxima, particularly when steep slopes lead to low peaks [17]. A recent study with the enzyme β-lactamase found that alternative initial mutations repeatedly directed adaptation onto different mutational pathways [18]. Here, drift — the chance occurrence of the first mutation — was again important for evolvability, but it was the mutation with greatest benefit that directed evolution to a higher peak.

The study by Woods et al.[2] is about selection for evolvability over relatively short time scales, allowing a single fixation event. However, selection for increased evolvability may also happen at longer time scales involving multiple selective sweeps, but then as a result from competition between rather than within populations. Future studies should address the factors determining long-term evolvability, for which Woods et al.[2] provide an important framework, conceptually as well as methodologically.


+++++


Professores, pesquisadores e alunos de universidades públicas e privadas (300) com acesso ao site CAPES/Periódicos podem ler gratuitamente este artigo da Current Biology and de mais 22.440 publicações científicas.


+++++


NOTA CAUSTICANTE DESTE BLOGGER:


Que Darwin já era epistemologicamente, disso eu já sabia desde 1998 após ler o livro A caixa preta de Darwin, de Michael Behe. O meu ceticismo localizado sobre a teoria geral da evolução diz justamente respeito à capacidade da seleção natural como mecanismo evolucionário criar a diversidade e complexidade biótica encontradas na natureza. Darwin disse que a seleção natural era cega e não guiada.

Coitado de Darwin, o homem que teve a maior ideia que toda a humanidade já teve está a cada dia que passa se mostrando falido nas suas especulações transformistas: as evidências apontam noutra direção -- teleologia!!! Argh, isso é como assassinar Darwin morto!!!