Dissection of the role of a Src homology 3 domain in the evolution of binding preference of paralogous proteins

Fig. 1.. Characterization of the interactomes of Myo3 and Myo5. a) Schematic representation of the experimental design. The baits are shown as circles and preys are shown as rectangles. PPIs are represented with lines, and the color of the prey nodes shows the binding preference of each paralog. The left panel represents the WT PPI network of the paralogs, and the right panel shows the PPI network of the SH3-depleted variants. The comparison between experiments allows us to identify SH3-dependent PPIs. b) Myo3 and Myo5 PPI network. The bait proteins are displayed in the center of the network. The edges show the SH3 contribution to the PPI: Δ(PPIWT–PPIdepleted) median score. The vertical dashed line separates the SH3-dependent, for one or both paralog, and SH3-independent preys as determined by Benjamini–Hochberg-corrected Wilcoxon tests (P < 0.05). Paralog-specific partners are indicated by a colored dot in the nodes (Myo3: blue, Myo5: red). Node colors indicate functions (Cherry et al. 2012). White nodes are abundance controls. c) Scatter plots comparing the PPI scores of the optSH3 paralog variants with the SH3-depleted paralogs. Myo3 data are shown on the left, and Myo5 data are shown…

Fig. 2.. SH3 domains are not sufficient to explain the PPI preferences of the paralogs. a) Expression level of extant paralogs and SH3-depleted protein variants tagged with GFP. The median cell fluorescence signal (arbitrary units) was measured for 5,000 individual cells from 3 independent cultures (individual data point). The negative control is a strain not expressing GFP (strain BY4741). Student's t-test P-values are shown above each comparison. b) Comparison of the optSH3 variants interaction profiles of Myo3 and Myo5. The colors represent the SH3 dependency of the PPIs. Triangles are divergent PPIs between the paralogs. c) Schematic representation of the expected PPI network of the swapSH3 and AncA paralog variants compared with the extant PPI network under different evolutionary scenarios. The PPIs are represented by lines, and the colors of the nodes show the extant paralog partners. The swapSH3 determines if the contribution of the native SH3 is essential for the paralog-specific PPIs. The PPIs observed with AncA identify which paralog-specific PPI was gained or lost following the duplication event. d) and e) Comparison of the PPIs of Myo3 and Myo5 containing various SH3s. The…

Fig. 3.. Ancestral SH3 domains affect the PPIs of Myo3 and Myo5. a) Simplified phylogenetic relationship of MYO3/MYO5 sequences where nodes represent an extant SH3 (blue and red) or an ancestral SH3 that was reconstructed, with an estimation of the time at which these domains existed. b) Superimposed extantMyo3 SH3 structure (blue, PDB:1RUW) and the AlphaFold2-predicted structures of each reconstructed ancestral SH3 showing high structural conservation as estimated by RMSD. c) Heatmap showing the PPIs of SH3 variants inserted in Myo3 and Myo5 with each of their interacting partners (preys, y-axis). The preys are ordered by SH3-depleted Myo5 PCA signal and by interaction type with the paralogs (right annotation, cyan: SH3-dependent for one or both paralogs, gray: SH3-independent, black: abundance controls). The tile color corresponds to the median PPI score, and the red x marks correspond to a significant difference (P < 0.05, Wilcoxon test, Benjamini–Hochberg correction) in interaction strength compared with the control interaction (dashed red frame). Missing data are shown in gray.

Fig. 4.. Contribution of SH3 binding to their motifs estimated by molecular docking. a) Superimposed structures of the 10 best-scored docking structures of the Osh2 peptide binding to Myo3 SH3. The peptides show a variety of binding conformations, whereas the SH3 structures remain similar. b) Scatter plots showing the relationship between the computed median ΔGs of the docked peptide-SH3 complex and the corresponding PPI scores of the extant paralogs as measured by DHFR PCA. Yellow data points correspond to proteins for which the proline motifs were validated in vivo (Fig. 1e). Spearman correlation coefficients (r) and P-values appear at the top of each plot. c) Violin plot comparing the distribution of median ΔG values for each SH3s. Paired Wilcoxon tests P-values are labeled on top of the violin plot.

Fig. 5.. Characterization of free SH3 binding profiles. a) Schematic representation of the DHFR PCA experiment using free SH3s with an inducible promoter. The transcription factor GEM binds to β-estradiol and induces the expression of the SH3 coding sequence fused with the DHFR F[1,2] via the GAL1 promoter. The expression of the construct increases with the amount of β-estradiol, leading to an increase of the DHFR PCA signal. b) Scatter plot comparing the PPI profile of the free paralogous SH3s and that of the complete paralogs (entire proteins with the optSH3s). Spearman correlation coefficients (r) and P-values for each experiment are shown. c) Boxplot of the PPI scores for several preys (x-axis) with 3 SH3 variants. Kruskal–Wallis tests compare the scores among the 3 variants for each prey. P-value significance levels are labeled at the top of each group (ns: P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001).