Probing the dimeric structure of porcine aminoacylase 1 by mass spectrometric and modeling procedures

Aminoacylase I is a zinc-binding metalloprotease catalyzing the hydrolysis of N-alpha-acylated L-amino acids; it presents altered expression levels in different renal and small cell lung carcinomas. A description of its redox and oligomerization state was achieved by combined biochemical and mass spectrometric procedures. A topological analysis of the enzyme structural architecture was derived from limited proteolysis and selective chemical modification experiments, using a broad range of proteases and chemical reagents. The analysis of the reaction products by different mass spectrometric techniques identified 26 amino acids as being accessible on the molecular surface, defining polypeptide regions exposed in the structure of the dimeric protein. The nature of the intermolecular contact zone between monomers was investigated by cross-linking reaction and mass mapping experiments. The cross-linked dimer was isolated, and the intermolecular cross-linked peptides were characterized, thus demonstrating the spatial proximity of Lys220 and Lys231 at the dimerization interface. Standard modeling procedures based on automatic alignment on the structure of members of the M20 peptidase family failed to produce a dimeric model consistent with experimental data. Discrepancies were observed mainly at the dimer interface and at loop regions. Therefore, a refined model for this dimeric protease was calculated by selecting the one able to generate a structure fully compatible with experimental findings, among all possible suboptimal sequence alignments. According to this model, each aminoacylase monomer consists of two domains: a globular catalytic subunit (residues 1-188 and 311-399) consisting of a beta-sheet sandwiched between alpha-helices and a second beta-sheet located on the surface, and the dimerization domain (residues 189-310) folding into a beta-sheet flanked on one side by two alpha-helices. These results indicate that reliable approaches such as limited proteolysis, selective chemical modification, and cross-linking coupled to mass spectrometry can be used to test and optimize molecular models of multimeric proteins and highlight problems in automatic model building.