Unraveling the causes of adaptive benefits of synonymous mutations in TEM-1 β-lactamase

Abstract

While synonymous mutations were long thought to be without phenotypic consequences, there is growing evidence they can affect gene expression, protein folding, and ultimately the fitness of an organism. In only a few cases have the mechanisms by which synonymous mutations affect the phenotype been elucidated. We previously identified 48 mutations in TEM-1 β-lactamase that increased resistance of Escherichia coli to cefotaxime, 10 of which were synonymous. To better understand the molecular mechanisms underlying the beneficial effect of these synonymous mutations, we made a series of measurements for a panel containing the 10 synonymous together with 10 non-synonymous mutations as a reference. Whereas messenger levels were unaffected, we found that total and functional TEM protein levels were higher for 5 out of 10 synonymous mutations. These observations suggest that some of these mutations act on translation or a downstream process. Similar effects were observed for some small-benefit non-synonymous mutations, suggesting a similar causal mechanism. For the synonymous mutations, we found that the cost of resistance scales with TEM protein levels. A resistance landscape for four synonymous mutations revealed strong epistasis: none of the combinations of mutations exceeded the resistance of the largest-effect mutation and there were synthetically neutral combinations. By considering combined effects of these mutations, we could infer that functional TEM protein level is a multi-dimensional phenotype. These results suggest that synonymous mutations may have beneficial effects by increasing the expression of an enzyme with low substrate activity, which may be realized via multiple, yet unknown, post-transcriptional mechanisms.

ΦX174 Attenuation by Whole Genome Codon Deoptimization

“…For example, of 48 cefotaxime resistance mutations in TEM-1 β -lactamase that were identified from a randomly mutagenized pool of variants, 10 were silent (98) . One mutation increased nitrocefin hydrolysis by 7.5-fold, probably by increasing the amount of functional TEM-1 (99 A good predictor of translational throughput is mRNA secondary structure (21) . Tightly folded mRNAs are traversed by the ribosome more slowly than unfolded mRNAs, although faster translation is not necessarily beneficial.…”

The fitness landscapes of translation

“…4). The combination of strong sign-epistatic interactions with extended plateaux of approximately constant resistance levels is indeed a visually striking feature of the experimental data set that motivated this work [7]. A quantitative comparison between the model and the data is however beyond the scope of this article and will be presented elsewhere.…”

“…Here we explore possible mechanisms that could explain sign epistatic interactions between synonymous mutations. In line with the experimental observations [7], we assume that the effect of the synonymous mutations on organismal fitness is mediated by the efficiency of protein translation. The process of translation can be described by stochastic particle models of exclusion type, which have been used in the field for more than 50 years [10,11,12,13,14,15].…”

“…A panel of 10 synonymous point mutations was identified which significantly increase the resistance against cefotaxime, as quantified by the drug concentration at which a fraction of 10 −4 of bacterial colonies survives [6]. Moreover, several of these mutations display strong nonlinear interactions in their effect on resistance [7]. Of particular interest are interactions referred to as sign-epistatic [8].…”

“…If efficiency is identified with the stationary ribosome current, we prove rigorously that sign epistasis is not possible. In contrast, if efficiency is taken to be the average speed of a ribosome (equivalently, the inverse of the ribosome travel time), then plausible scenarios giving rise to sign-epistatic interactions are readily found, and fitness landscapes that are qualitatively similar to those observed experimentally in [7] can be constructed [17]. We conclude, therefore, that there are (at least) two different fitness landscapes of translation, one of which is always simple, whereas the other displays a complex structure of hierarchically organized neutral networks.…”