Modular Organization of Residue-Level Contacts Shapes the Selection Pressure on Individual Amino Acid Sites of Ribosomal Proteins

Abstract

Understanding the molecular evolution of macromolecular complexes in the light of their structure, assembly, and stability is of central importance. Here, we address how the modular organization of native molecular contacts shapes the selection pressure on individual residue sites of ribosomal complexes. The bacterial ribosomal complex is represented as a residue contact network where nodes represent amino acid/nucleotide residues and edges represent their van der Waals interactions. We find statistically overrepresented native amino acid-nucleotide contacts (OaantC, one amino acid contacts one or multiple nucleotides, internucleotide contacts are disregarded). Contact number is defined as the number of nucleotides contacted. Involvement of individual amino acids in OaantCs with smaller contact numbers is more random, whereas only a few amino acids significantly contribute to OaantCs with higher contact numbers. An investigation of structure, stability, and assembly of bacterial ribosome depicts the involvement of these OaantCs in diverse biophysical interactions stabilizing the complex, including high-affinity protein-RNA contacts, interprotein cooperativity, intersubunit bridge, packing of multiple ribosomal RNA domains, etc. Amino acid-nucleotide constituents of OaantCs with higher contact numbers are generally associated with significantly slower substitution rates compared with that of OaantCs with smaller contact numbers. This evolutionary rate heterogeneity emerges from the strong purifying selection pressure that conserves the respective amino acid physicochemical properties relevant to the stabilizing interaction with OaantC nucleotides. An analysis of relative molecular orientations of OaantC residues and their interaction energetics provides the biophysical ground of purifying selection conserving OaantC amino acid physicochemical properties.

Keywords: contact network; network motif; protein–RNA interaction; purifying selection; ribosome; stability.

Fig. 1.—

(A) Different OaantCs identified in the 70S ribosomal particle. In the second column, we show the different OaantCs. Left half of the panel enlists the amino acid data: 1) Column 3 represents the number of occurrences of each OaantCs, the number of cases where statistically significant signature of the respective amino acid chemical property being conserved is found is mentioned in the parenthesis, 2) average d_N/d_S ratio obtained at each OaantCs class (column 4), and 3) randomized evolutionary rates (column 5) and their ratios (column 6). Right half of the panel represents nucleic acid data: 1) Average Shannon entropy of nucleotides associated with each OaantCs class (column 7), 2) the randomized entropy values (column 8), and 3) their ratios (column 9). (B) Scaled occurrence bias of different types of amino acids in different contact OaantCs. The bias estimates are scales to the range 0–1 and are plotted as a colored heatmap. (C) The expected number of nucleotide contacts for individual amino acids is shown as bar-plots. (D) The heatmap represents the statistical bias of individual amino acids associated with different OaantCs classes being subjected to selection pressure conserving their chemical properties.

Fig. 2.—

OaantCs associated with various functional aspects of the 70S particle. Complex interactions such as stabilizing multiple ribosomal RNA domain interfaces, cooperativity phenomenon, and intersubunit bridging OaantCs with higher contact number. One OaantC can be associated with multiple types of interactions. For example, a major fraction of the OaantCs maintaining interprotein cooperativity also maintain the molecular packing among multiple rRNA domains. The frequency of occurrences of different OaantCs in the respective functional classes is mentioned in the parenthesis.

Fig. 3.—

Box-plot distributions of (A) synonymous substitution rate (d_S) and (B) the nonsynonymous substitution rate (d_N) estimated at amino acid sites associated with different OaantC class. Permutation Mann–Whitney U tests under the null hypothesis of equal mean are performed for all possible OaantC pair to investigate whether the distributions significantly differ from one another. The respective P values of the U tests are shown in terms of a colormap matrix (significance of each color is mentioned in the figure). (C–F) The percent changes of two physical (flexibility and volume) and two chemical properties (hydropathy and polarity) upon individual point-mutations are plotted as box-plot distributions for each OaantC class. Distribution plots for the four properties are presented in panel (C), (D), (E), and (F), respectively. Permutation Mann–Whitney U tests under the null hypothesis of equal mean are performed to investigate whether the distributions associated with various OaantCs significantly differ from one another. P values are represented as colormap matrices and significance of each color is mentioned in the figure.

Fig. 4.—

(A) Superimposed OaantCs with k = 5 are shown here as surface representation for positively charged, hydrophobic and polar amino acids. Surfaces are colored according to atom partial charges. (B) Molecular diagrams (stick representation) of “selected” OaantCs are shown for the three chemical categories. Stick colors are according to the amino acid partial charge scale provided in panel (A). At the top, the Arg12 of uL20 protein represents a positively charged “selected” OaantC that contacts five nucleotides. In the middle, Ala162 of uL3 represents a hydrophobic “selected” OaantC contacting five nucleotides. At the bottom, Asn23 of bL17 represents a polar OaantC. (C) Comparison of computed amino acid–nucleotide interaction free energies for selected and nonselected OaantCs (top: positively charged, middle: hydrophobic, bottom: polar). SE, selected; NS, not selected.