Comparative analysis of mutational robustness of the intrinsically disordered viral protein VPg and of its interactor eIF4E

Abstract

Conformational intrinsic disorder is a feature present in many virus proteins. Intrinsically disordered regions (IDRs) have weaker structural requirement than ordered regions and mutations in IDRs could have a lower impact on the virus fitness. This could favor its exploration of adaptive solutions. The potyviral protein VPg contains IDRs with determinants for adaptation to its host plant. To experimentally assess whether IDRs are more resistant to mutations than ordered regions, the biologically relevant interaction between mutant libraries of both VPg and the eukaryotic translation initiation factor 4E (eIF4E) and their respective wild type partner was examined using yeast two hybrid assay. Our data shows that VPg is significantly more robust to mutations than eIF4E and as such belongs to a particular class of intrinsically disordered proteins. This result is discussed from the standpoint of IDRs involvement in the virus adaptive processes.

Fig 1. Distribution of the number of non-synonymous mutations in sequenced variants sampled from each VPg and eIF4E mutant library.

Black dots: number of NS mutation per kb retrieved from each of the 32 variants sequenced. a, b and c are 3 significantly different statistical groups (p-value cut-off at 0.01, Kruskal-Wallis test).

Fig 2. Distribution along eIF4E and VPg sequences of all mutations found in the 32 randomly chosen variants sequenced in each of the six libraries.

A) eIF4E; B) VPg. NS: non-synonymous; S: synonymous; INDEL: insertion/deletion; STOP: mutation involving the appearance of STOP-codons. For comparative purpose, number of mutations is displayed rather than percentages of mutated sites in the coding sequence, as the two proteins are of comparable length.

Fig 3

Comparative mutational robustness of VPg (dark grey) and eIF4E (light gray). Box plots display of the dispersion of remaining functional variants of VPg (dark grey) and eIF4 (light gray) within the low, med and high libraries. Yeasts transformed with “low”, “medium” and “high” libraries of variants were matted with yeasts transformed with the wild type partner encoding gene. The resulting diploid yeasts containing the genes of both protein partners were selected on a non-interaction selective medium (-LW). The corresponding number of colonies, τ, represented the whole population of variants. Yeast colonies able to grow on stringent interaction selecting medium (-LWHA) were counted on 3 to 5 independent plates (Three independent panning). The corresponding number, σ, was divided by τ. ρ = σ /τ refers to the rate of remaining functional variants within the mutants libraries relative to the starting whole population of variants. ρ is indicative of the mutational robustness of the protein considered. (See Material and Methods for more details). (t-test p-value <0.01). Inset: the rate of remaining functional variants versus mutation accumulation (black circles and gray circles, VPg and eIF4E library survival rates respectively). The mono exponential model of fitness decline as a function of mutation accumulation was fitted to the experimental data ρ = EXP(-k*n) with n, the averaged mutation number per protein. Best fit: k = 0.274±0.011 (VPg black line) and k = 0.412±0.016 (eIF4E gray line).