D-Aminoacylases and Dipeptidases within the Amidohydrolase Superfamily: Relationship Between Enzyme Structure and Substrate Specificity

Abstract

Approximately one third of the genes for the completely sequenced bacterial genomes have an unknown, uncertain, or incorrect functional annotation. Approximately 11,000 putative proteins identified from the fully-sequenced microbial genomes are members of the catalytically diverse Amidohydrolase Superfamily. Members of the Amidohydrolase Superfamily separate into 24 Clusters of Orthologous Groups (cogs). Cog3653 includes proteins annotated as N-acyl-D-amino acid deacetylases (DAAs), and proteins within cog2355 are homologues to the human renal dipeptidase. The substrate profiles of three DAAs (Bb3285, Gox1177 and Sco4986) and six microbial dipeptidase (Sco3058, Gox2272, Cc2746, LmoDP, Rsp0802 and Bh2271) were examined with N-acyl-L-, N-acyl-D-, L-Xaa-L-Xaa, L-Xaa-D-Xaa and D-Xaa-L-Xaa substrate libraries. The rates of hydrolysis of the library components were determined by separating the amino acids by HPLC and quantitating the products. Gox1177 and Sco4986 hydrolyzed several N-acyl-D-amino acids, especially those where the amino acid was a hydrophobic residue. Gox1177 hydrolyzed L-Xaa-D-Xaa and N-acetyl-D-amino acids with similar catalytic efficiencies (~10⁴ M⁻¹s⁻¹). The best substrates identified for Gox1177 and Sco4986 were N-acetyl-D-Trp and N-acetyl-D-Phe, respectively. Conversely, Bb3285 hydrolyzed N-acyl-D-Glu substrates (kcat/Km ⁹́⁸ 5 x 10⁶M⁻¹s⁻¹) and, to a lesser extent, L-Xaa-D-Glu dipeptides. The structure of a DAA from A. faecalis did not help explain the substrate specificity of Bb3285. N-methylphosphonate derivatives of D-amino acids were inhibitors of the DAAs examined. The structure of Bb3285 was solved in complex with the N-methylphosphonate derivative of D-Glu or acetate/formate. The specificity of Bb3285 was due to an arginine located on a loop which varied in conformation from the A. faecalis enzyme. In a similar manner, six microbial renal dipeptidase-like proteins were screened with 55 dipeptide libraries. These enzymes hydrolyzed many dipeptides but favored L-D dipeptides. Respectable substrates were identified for proteins Bh2271 (L-Leu-D-Ala, kcat/Km = 7.4 x 10⁴ M⁻¹s⁻¹), Sco3058 (L-Arg-D-Asp, kcat/Km = 7.6 x 10⁵ M⁻¹s⁻¹), Gox2272 (L-Asn-D-Glu, kcat/Km = 4.7 x 10⁵ M⁻¹s⁻¹), Cc2746 (L-Met-D-Leu, kcat/Km = 4.6 x 10⁵ M⁻¹s⁻¹), LmoDP (L-Leu-D-Ala, kcat/Km = 1.1 x 10⁵ M⁻¹s⁻¹), Rsp0802 (L-Met-D-Leu, kcat/Km = 1.1 x 10⁵ M⁻¹s⁻¹). Phosphinate mimics of dipeptides were inhibitors of the dipeptidases. The structures of Sco3058, LmoDP and Rsp0802 were solved in complex with the pseudodipeptide mimics of L-Ala-D-Asp, L-Leu-D-Ala and L-Ala-D-Ala, respectively. The structures were used to assist in the identification of the structural determinants of substrate specificity.