Protein oligomerization rescues bacterial fitness

Most random mutations are destabilizing, and they affect fitness by compromising structural integrity of proteins that results in decreased activity and toxicity due to aggregation. Surprisingly, destabilizing mutations can also induce oligomerization, thus preserving proteins' solubility and function.

HFSP Long-Term Fellow Shimon Bershtein and colleagues
authored on Thu, 12 April 2012

One of the fundamental problems of modern biology is that there is no integrated model that relates the microscopic properties of proteins to the macroscopic level of evolutionary biology. Natural selection acts on organisms and populations (phenotype, fitness), and the laws of physics and chemistry act on proteins that are often far removed from the level of phenotype. What is lacking is an understanding of the rules that map interactions at one level onto interactions at a higher level, a genotype-phenotype map. Or, more specifically, how changes in molecular properties of proteins (stability being most important) induced by mutations are translated into biological fitness.

Earlier attempts to determine the genotype-phenotype relationship (GPR) were either within the ‘’phenotype first’’ paradigm, whereby phenotypic traits were first selected for, and the accompanying genomic changes later evaluated, or “mutations first” experiments, in which new phenotypes are obtained by random mutagenesis. Both approaches tried to jump the canyon separating sequences and phenotype in ‘’one shot’’. Despite substantial efforts, however, these past approaches provided only limited understanding of genotype-phenotype relationship, mainly because they did not allow to relate a single substitution event to a particular phenotypic trait. Here we undertook an entirely novel ‘’bottom up’’ approach. We explored the fitness landscape one amino acid change at a time with full control of what mutations do to the molecular properties of a mutated protein – allowing, therefore, to pinpoint fitness effects to specific changes in protein stability, activity or intracellular abundance (Fig.1 in the paper).

Specifically, distribution of fitness effects was measured upon controlled site–directed chromosomal mutagenesis of an E. coli’s gene folA encoding dihydrofolate reductase (DHFR). DHFR catalyzes an electron transfer reaction from NADPH to 7,8-dihydrofolate to form 5,6,7,8-tetrahydrofolate that serves as a cofactor for a number of one-carbon-transfer reactions, and is essential for the synthesis of purines, thymidylate, and several amino acids. Being an essential protein, perturbations in the microscopic properties of DHFR induced by mutations are directly reflected in the fitness of E. coli’s strains carrying these mutations. In total, 27 E.coli strains carrying mostly destabilizing mutations in folA gene were created. We found no significant correlation between protein stability and its catalytic activity, and between catalytic activity and fitness in a limited range of variation of catalytic activity observed in mutants. Strong correlation found between stability of mutants and their intracellular abundance is suggestive of an active role of protein homeostatic machinery in maintaining intracellular concentration of proteins. Significant correlation between intracellular abundance of soluble DHFR and fitness was observed when cells grew at 30C. The picture was quite different at 42C: mutant DHFR proteins in a few strains aggregated, rendering them nonviable but the majority exhibited fitness higher than wild type.We found that mutational destabilization of DHFR proteins in E. coli was counterbalanced at 42C by soluble oligomerization that restores their structural stability and protects from aggregation.This finding adds a completely underappreciated new dimension to our understanding of the genotype-phenotype relationship: protein stability mediates fitness effects through perturbations in solubility.

Our study shows that fitness effects of mutations, while hardly rationalizable at the level of sequence variation,  can be ‘’projected’’ on a small number of ‘’axes’’ reflecting coarse-grained biophysical properties of proteins, such as stability or intracellular abundance, and the fitness landscape in the space of such coarse-grained properties appears ‘’smooth’’. Significant correlation exists between fitness and several coarse-grained quantities. This is good news for biophysics-based multi-scale modeling of evolution. The bad news is that it is still challenging to postulate such correlations a priori – first, simple equilibrium statistical mechanics considerations may not be applicable here because cellular environment represents an active medium where energy consuming machinery acts on proteins. Second, unexpected equilibrium mechanisms, such as soluble oligomerization may intervene.


Soluble oligomerization provides a beneficial fitness effect on destabilizing mutations. Bershtein, S., Wu, W., Shakhnovich, E.I. Proc Natl Acad Sci U S A. 2012 Mar 27;109(13):4857-62. Epub 2012 Mar 12.

Pubmed link