In the absence of a crystal structure, the EFI Computation Core generated predictions by docking to homology models of an Agrobacterium radiobacter amidohydrolase superfamily member. The predicted substrates, a set of modified pterins, were then experimentally validated by the AH Bridging Project, providing a new annotation within the amidohydrolase superfamily. As for Hobbs, et al., this success provides additional support for the EFI’s large scale computation-based strategy for functional prediction.
Of the over 22 million protein sequences in the nonredundant TrEMBL database, fewer than 1% have experimentally confirmed functions. Structure-based methods have been used to predict enzyme activities from experimentally determined structures; however, for the vast majority of proteins, no such structures are available. Here, homology models of a functionally uncharacterized amidohydrolase from Agrobacterium radiobacter K84 (Arad3529) were computed on the basis of a remote template structure. The protein backbone of two loops near the active site was remodeled, resulting in four distinct active site conformations. Substrates of Arad3529 were predicted by docking of 57 672 high-energy intermediate (HEI) forms of 6440 metabolites against these four homology models. On the basis of docking ranks and geometries, a set of modified pterins were suggested as candidate substrates for Arad3529. The predictions were tested by enzymology experiments, and Arad3529 deaminated many pterin metabolites (substrate, k(cat)/K(m) [M(-1) s(-1)]): formylpterin, 5.2 × 10(6); pterin-6-carboxylate, 4.0 × 10(6); pterin-7-carboxylate, 3.7 × 10(6); pterin, 3.3 × 10(6); hydroxymethylpterin, 1.2 × 10(6); biopterin, 1.0 × 10(6); d-(+)-neopterin, 3.1 × 10(5); isoxanthopterin, 2.8 × 10(5); sepiapterin, 1.3 × 10(5); folate, 1.3 × 10(5), xanthopterin, 1.17 × 10(5); and 7,8-dihydrohydroxymethylpterin, 3.3 × 10(4). While pterin is a ubiquitous oxidative product of folate degradation, genomic analysis suggests that the first step of an undescribed pterin degradation pathway is catalyzed by Arad3529. Homology model-based virtual screening, especially with modeling of protein backbone flexibility, may be broadly useful for enzyme function annotation and discovering new pathways and drug targets.
Figure 1. Sequence similarity network for cog0402. Protein sequences for cog0402 were retrieved from NCBI, subjected to an All-by-All BLAST to determine overall sequence similarity to all of the other proteins within this network.(27) Each dot (node) represents an enzyme, and each connection between two nodes (an edge) represents those enzyme pairs that are more closely related than the arbitrary E-value cutoff (10–70). Groups are arbitrarily numbered; those groups with experimentally determined substrate profiles are as follows: (1) S-adenosylhomocysteine/5′-deoxy-5′-methylthioadenosine deaminase (red nodes); (2) guanine deaminase (orange nodes); (4) 8-oxoguanine/isoxanthopterin deaminase (green nodes); (6) cytosine deaminase (light blue nodes); and (8) N-formimino-l-glutamate deiminase (purple nodes).
Figure 2. The active site of E. coli CDA and residues predicted to be important for Arad3529. Residues drawn in white are conserved across all members of cog0402. Thr-66, Gln-156, and Asp-314 in CDA, displayed in green, differed significantly in the sequence of Arad3529 and may be important for substrate binding. Thr-66 corresponds to Lys-85, Gln-156 corresponds to Leu-177, and Asp-314 corresponds to Asn-332. In the X-ray structure of CDA, Gln-156 forms hydrogen bonds with the bound inhibitor; Asp-314 is found within 4 Å of the CDA inhibitor but only forms steric interactions with it; Thr-66 is more distant from the active site. The conversion of a large lysine in Arad3529 suggests that the Lys-85 may reach into the active site and play a role in substrate binding.
Figure 3. The active site conformations of four representative models of Arad3529 used in virtual screening. Model-1 is ranked the best by DOPE score among 500 homology models generated automatically by MODELLER. Model-2, Model-3, and Model-4 are different from Model-1 in Loop-1, Loop-2, and both Loop-1 and Loop-2, respectively.
Figure 4. (Top) The binding pose of pterin intermediate N3 (yellow stick) formed by the activated hydroxide attacking the C-2 atom on the si-face of the aromatic ring of the tautomer 2-aminopteridin-4(1H)-one, in the active site of Model-2 of Arad3529 (white stick). (Bottom) The binding pose of pterin intermediate N1 (yellow stick) formed by the activated hydroxide attacking the C-2 atom on the re-face of the aromatic ring of the tautomer 2-aminopteridin-4(3H)-one, in the active site of Model-3 of Arad3529 (white stick). Polar interactions between the docked pterin and the modeled active site are shown by red dashed lines.
Reprinted with permission from the Journal of the American Chemical Society.
© 2013 American Chemical Society.