Evolutionary Dynamics

Environmental fluctuations and the design of proteins: a hypothesis.  In addition to folding and carrying out biochemical activities, natural proteins must also be evolvable; that is, they must have the capacity to generate heritable variation as conditions of selection fluctuate in the environment. In principle, this property of evolvability places unique and non-trivial selective pressures on genotypes, such that only a subset of sequences that can fold and function can also maintain the capacity for adaptation on the functional range and time-scales at which the environment changes. This line of thinking immediately raises the interesting proposition that the design of evolved proteins must not be fixed, but instead must fundamentally depend on the statistical history of fluctuations in conditions of selection. Indeed, we propose that the physical design of proteins embeds the statistics of environments.

Support for this hypothesis comes from several lines of work. First, we showed through deep mutational scanning in the TEM-1 β-lactamase that the extent and pattern of mutational sensitivity of proteins (or, robustness) and the capacity to adapt to new functions depends on the strength of selective pressure applied.  The higher the selective pressure, the less robust and evolvable is TEM-1. Furthermore, by examining the spatial pattern of adaptive mutations in the PDZ domain, we have argued that fitness under fluctuating conditions of selection depends on the existence of allosteric mutations that are neutral for the current environment but gain-of-function for alternative environments. Such “conditionally neutral” mutations can pre-exist as standing genetic variation and enable adaptation as selection conditions vary randomly.  This work leads to an interesting idea:  if allostery generates conditional neutrality and conditional neutrality enriches fitness in fluctuating environments, then it follows that the origins of allostery might lie not in its traditional role in functional regulation, but simply in the capacity of proteins to evolve. If so, proteins would come “pre-wired” for allosteric control as a consequence of the statistics of their evolutionary history, with functional regulation essentially riding on this preëxisting architecture. The concept that the origins of allostery lie in evolvability in fluctuating conditions of selection is a clear exposition of the hypothesis that environments can control the design of proteins.  Finally, theory and simulation in simple models for evolving networks in the laboratory of Olivier Rivoire show clearly that the topology of constraints between sequence positions depends fundamentally on the rate at which the condition of selection fluctuates.

A series of papers that make this argument: (1) (6) Kussell E et al. Physical review letters. 2006 97:068101, (2) Lee J, et al. Science. 2008 322:438-42 (3) Halabi N, et al. Cell. 2009 138:774-86, (4) Reynolds KA, et al. Cell. 2011 147:1564-75, (5) McLaughlin Jr RN, et al. Nature. 2012 491:138, (6) Stiffler MA, et al. Cell. 2015 160:882-92, (7) Hemery M, Rivoire O. Physical Review E. 2015 91:042704,  (8) Raman AS, et al. Cell. 2016 166:468-80.

The future: How can we test the basic hypothesis that the evolutionary design of proteins is a function of the statistical history of environments?  Our sense is that this can only be done through forward laboratory evolution, where we “watch” the evolution of a protein of interest under conditions in which we control all the relevant parameters – population size, mutation rate, and the statistics of selection. In addition it is important to be able to do this experiment with many replicates so that we can observe what is reproducible and what is idiosyncratic to particular trajectories. In current work, we are developing and extending phage-assisted continuous evolution (PACE), originally developed in the laboratory of David Liu (Harvard University) for this purpose.  With this technology, we should be able to quantitatively examine the evolution of protein function in constant or fluctuating environments, a platform for understanding if and how the dynamics of selection is embedded in the physical design of proteins. Stay tuned…