Figure 8

Analysis of positional differences in protein families can reveal potential sites of functional specificity. Comparison of structurally equivalent positions in the multiple alignments of sequence homologs for two distant relatives, 2'-O ribose methyltransferase from T. thermophilus (PDB ID 1ipaA) and hypothetical protein Ybea from E. coli (PDB ID 1ns5A). A. The ribbon diagrams of the structurally similar domain fragments (1ipaA, residues 116–263, and 1ns5A, residues 2–146) drawn by BobScript, a modification of the MolScript program [50]. Similar secondary structure elements are colored in blue (α-helices) and yellow (β-strands). The C-terminal helix D, which is involved in dimerization, is highlighted in red. Ball-and-stick models of sidechains are shown for the residues that (i) have C^αdistance less than 2A in the superimposed structures of the two domains, and (ii) correspond to significantly different positions in the multiple sequence alignments of the compared families (P < 0.01). In the ball-and-stick models, C, N, O, and S atoms are shown in gray, blue, red, and yellow, respectively. B. Structure-based sequence alignment of the two domains. The α-helices and β-strands are displayed as arrows and cylinders, respectively. Highlighted: the protein positions that have significantly different residue content in the two corresponding multiple alignments of sequence homologs detected by PSI-BLAST. Cyan, spatially close protein positions, with C^αdistance in the superimposed structures less than 2A; gray, position pairs with higher C^αdistances.