How synonymous mutations alter enzyme structure and function over long time scales

Abstract

The specific activity of enzymes can be altered over long time scales in cells by synonymous mutations, which change an mRNA molecule's sequence but not the encoded protein's primary structure. How this happens at the molecular level is unknown. Here, we resolve this issue by applying multiscale modeling to three E. coli enzymes - type III chloramphenicol acetyltransferase, D-alanine-D-alanine ligase B, and dihydrofolate reductase. This modeling involves coarse-grained simulations of protein synthesis and post-translational behavior, all-atom simulations as a test of robustness, and QM/MM calculations to characterize function. We first demonstrate that our model accurately predicts experimentally measured changes in specific activity due to synonymous mutations. Then, we show that changes in codon translation rates induced by synonymous mutations cause shifts in co-translational and post-translational folding pathways that kinetically partition molecules into subpopulations that very slowly interconvert to the native, functional state. These long-lived states exhibit reduced catalytic activity, as demonstrated by their increased activation energies for the reactions they carry out. Structurally, these states resemble the native state, with localized misfolding near the active sites of the enzymes. The localized misfolding involves noncovalent lasso entanglements - a topology in which the protein backbone forms a loop closed by noncovalent native contacts which is then threaded by another portion of the protein. Such entanglements are often kinetic traps, as they can require a large proportion of the protein to unfold, which is energetically unfavorable, before they disentangle and attain the native state. The near-native-like structures of these misfolded states allow them to bypass the proteostasis machinery and remain soluble, as they exhibit similar hydrophobic surface areas as the native state. These entangled structures persist in all-atom simulations as well, indicating that these conclusions are independent of model resolution. Thus, synonymous mutations cause shifts in the co- and post-translational structural ensemble of proteins, whose altered subpopulations lead to long-term changes in the specific activities of some enzymes. The formation of entangled subpopulations is therefore a mechanism through which changes in translation elongation rate alter ensemble-averaged specific activities, which can ultimately affect the efficiency of biochemical pathways and phenotypic traits.

A Newly Identified Class of Protein Misfolding in All-atom Folding Simulations Consistent with Limited Proteolysis Mass Spectrometry

“…Relative to their time-scale of protein folding, these entangled states are not long-lived kinetic traps even at 310 K. Based on the previous coarse-grained simulation results, this is to be expected 1 , as ubiquitin and NTD are atypically small (the median protein length in yeast is 379 residues 15 ). In such very small proteins, entanglements compose a large proportion of the total protein structure present – and hence, it is easier to disentangle as most of the protein structure is misfolded and less stable.…”

“…The misfolded states observed in the aforementioned simulations involved either the gain of a non-native entanglement (i.e., the formation of an entanglement that is not present in the native ensemble, Table S1) or the loss of a native entanglement (i.e., an entanglement present in the native state fails to form, Table S1) [1][2][3] . This newly predicted class of misfolding offers an explanation for the decades old observations that non-functional protein molecules can persist for long-time scales in the presence of chaperones and not rapidly aggregate nor be degraded [7][8][9] .…”

“…Only a small number of mechanisms intrinsic to a protein's primary structure are known to cause monomeric protein misfolding (Table 1). Recently, high-throughput, coarse-grained simulations of protein synthesis and folding of the E. coli proteome have suggested there exists an additional widespread mechanism of misfolding [1][2][3] . Proteins can populate off-pathway misfolded states that involve a change in entanglement 1 .…”

“…The radius of gyration of these states are comparable to the native state (within 5.5%, 5.2%, and 2.4%, on average, respectively, of the native state Ubiquitin, λ-repressor, and 4-Diphosphocytidyl-2-C-Methyl-D-Erythritol Kinase values), indicating the entangled and native states are of similar sizes. These entangled states are estimated to have a solubility similar to the native state (Table S2) according to a model that accounts for surface area chemical properties 1 . This indicates these misfolded states are likely to remain as soluble in solution as the native fold.…”

“…Several mechanisms intrinsic to a protein's primary structure are known to cause monomeric protein misfolding (Table 1). Recently, high-throughput, coarse-grained simulations of protein synthesis and folding of the E. coli proteome have suggested there exists an additional widespread mechanism of misfolding [1][2][3] : proteins can populate off-pathway misfolded states that involve a change in non-covalent lasso entanglements 1,2 . This type of entanglement is defined by the presence of two structural components: a loop formed by a protein backbone segment closed by a non-covalent native contact, and another protein segment that is threaded through and around this loop [4][5][6] in some cases multiple times (Figs.…”