The Pith: In evolution if you want to win in the long run you don’t want all your eggs in one basket, even if that’s the choicest basket. Sh*t happens, and you better have some back up strategies.

Diversity is a major question in evolutionary biology. In particular, why is there so much diversity, so that the tree of life manifests a multitude of morphs? Might there not be some supreme replicator which emerges from the maelstrom to conquer all before it? This is actually the scenario which unfolds in much of science fiction, with monomorphic grey goo eating everything in its path (a more aesthetically differentiated variant of the super-species emerges in Brian W. Aldiss’  Helliconia Winter). As it is, life on earth does not seem to be converging upon an optimum phenotype for all individuals. In contrast, it seems to be going in the opposite direction broadly speaking (thinking on billion year scales), with the shift from the monotony of communal cyanobacteria to the riotous diversity of tropical forest biomes and coral reefs.

There are many ways you might be able to explain this diversity. Temporal and spatial heterogeneity produces perpetually shifting selection pressures, resulting in transient morphs one after the other. Negative frequency dependent selection, whereby the fitness of a phenotype runs up against its own success. This dynamic is one of the drivers of the Red Queen Hypothesis; the evolutionary arms race in some cases witnessing the resurrection of old techniques against which defenses are no longer recalled. Then there is the possibility that the lack of natural selection as an efficacious evolutionary force could allow for the diversification of phenotypes through  random drift. Finally, it may simply be that the gusher of mutation is powerful enough that novelty overwhelms selection and drift’s attempt to pare it back.

A new paper in Nature offers up another possibility. It does so by examining the fact that biological diversity remains operative even within a homogenized chemostat. A chemostat in this context refers to a controlled environment where inputs and outputs are balanced to maintain constant equilibrium conditions for a bacterialculture. Therefore, an unbeatable strategy should emerge in this medium perfectly tailored to the environmental constants, resulting in a homogeneous biota to match. Empirically this is not what occurs. So some explanation is warranted.

Metabolic trade-offs and the maintenance of the fittest and the flattest:

How is diversity maintained? Environmental heterogeneity is considered to be important1, yet diversity in seemingly homogeneous environments is nonetheless observed…This, it is assumed, must either be owing to weak selection, mutational input or a fitness advantage to genotypes when rare…Here we demonstrate the possibility of a new general mechanism of stable diversity maintenance, one that stems from metabolic and physiological trade-offs…The model requires that such trade-offs translate into a fitness landscape in which the most fit has unfit near-mutational neighbours, and a lower fitness peak also exists that is more mutationally robust. The ‘survival of the fittest’ applies at low mutation rates, giving way to ‘survival of the flattest’…at high mutation rates. However, as a consequence of quasispecies-level negative frequency-dependent selection and differences in mutational robustness we observe a transition zone in which both fittest and flattest coexist. Although diversity maintenance is possible for simple organisms in simple environments, the more trade-offs there are, the wider the maintenance zone becomes. The principle may be applied to lineages within a species or species within a community, potentially explaining why competitive exclusion need not be observed in homogeneous environments. This principle predicts the enigmatic richness of metabolic strategies in clonal bacteria…and questions the safety of lethal mutagenesis…as an antimicrobial treatment.

A ‘quasispecies’ model is usually characterized by a high mutational rate relative to what we’re usually used to thinking about (where fidelity rates are higher). You can think of a quasispecies as clusters of genotypes separated by various mutational ‘steps.’ The biological variation within this context then consists of a set of clusters reshaped and buffeted by mutational froth. One important point to mention, the supplements to this paper are huge. And in some ways more informative than the paper itself, which is a ‘letter’, and so a spare five pages. (They’re free too!)

What the authors found is that there are two strategies which ‘win’ in extreme mutational environments. In a low mutation environment the highest fitness phenotype spreads and dominates. Consider for example a bamboo planet where bamboo has been the the vegetation for billions of years. You’re talking panda heaven. But obviously perturbing the environment even marginally can cause problems for the perfectly adapted pandas.

In this case the level of focus consists of the genotypes. There may be perfectly fit genotypes which can be shifted off their adaptive peak by only a few mutations. But this is obviously not an issue in a low mutation environment. In a world of only steaks, having the best steak knife maximizes your ability to eat steak. The homogeneous medium in the chemostat is less important here than the homogeneity of genotype enabled by the low level of inputs of new variants.

But what if you’re thrown into the jungle? Then a swiss army knife may be better. The ‘flat’ strategy has a lower fitness peak, but its nearby mutational neighbors don’t drop nearly as fast in relative terms. This means it’s robust to mutational events which might perturb it from the optimum. Going back to an analogy, imagine you’re on a pedestal. You’d like to be on the highest pedestal. But what if there’s a trade off between height and width of the area upon which you can stand? If it is a calm day, you’d pick the narrow but high pedestal. On a windy day, you might think the better of it and go for the pedestal with the widest base for you to position yourself upon. Even if you’re not reaching for the heights, it might be better not to risk falling to your death.

Where the two strategies can coexist in mixture is in the broad zone between low and high mutational environments, in concert with negative frequency dependence. Remember, this is all operative in a homogeneous environment. But here the variation parameter is endogenous to the system, in the form of the mutations. This prevents a super-dominant fit strategy from sweep all before it.

An interesting property of this model is that the zone of coexistence in terms of mutation rates expands as you add into the model more traits which might exhibit fitness trade offs. In this bacterial model they focused on the trade offs between yield of energy and rate of energy production, as well as affinity and transport of a receptor to a substrate. These are trade offs which are biophysically constrained, suggesting that at this level the evolutionary adaptive space is scaffolded by deep physical properties of the universe. You can go further with this in terms of speculation. Obviously more complex multicellular organisms likely have many more trade offs baked into the cake structurally than unicellular microbes, so it may be that the possibilities for diversity of morphs across a wide mutational zone increase as one ascends the complexity scale.

Finally, the authors finish with a rather cautionary note:

…Importantly, the survival of the fittest and flattest questions the safety of this therapy. Whereas recent models26 suggest that pathogen densities should decrease linearly as mutation rates increase, our work suggests that increasing mutation rates can increase both density and diversity by pushing the population to a slightly lower but much flatter fitness peak….

Makes me think of the More mutations = greater fitness post.

In many ways this sort of model reminds me about the arguments for the evolution of sex. In the short term uniform perfectly adapted lineages are the winning strategy, but over the long term these lineages tend to go extinct as conditions change. What looks good in the short run may not be a good bet in the long run. In the longest run we’re dead, and the universe will wind down, but evolutionary biology often operates in that medium scale zone between the de facto infinities of physics and the comprehensible periods of history. Ten million years is not something we can grasp intuitively, but it is far different from one hundred billion years.

Citation: Beardmore RE, Gudelj I, Lipson DA, & Hurst LD (2011). Metabolic trade-offs and the maintenance of the fittest and the flattest. Nature PMID: 21441905

March 29th, 2011 Tags: Evolution, Evolutionary Genetics, Genetics
by Razib Khan in Evolution | 5 comments | RSS feed | Trackback >