Mutational and structural analysis of L-N-carbamoylase reveals new insights into a peptidase M20/M25/M40 family member

Abstract

N-Carbamoyl-L-amino acid amidohydrolases (L-carbamoylases) are important industrial enzymes used in kinetic resolution of racemic mixtures of N-carbamoyl-amino acids due to their strict enantiospecificity. In this work, we report the first L-carbamoylase structure belonging to Geobacillus stearothermophilus CECT43 (BsLcar), at a resolution of 2.7 Å. Structural analysis of BsLcar and several members of the peptidase M20/M25/M40 family confirmed the expected conserved residues at the active site in this family, and site-directed mutagenesis revealed their relevance to substrate binding. We also found an unexpectedly conserved arginine residue (Arg(234) in BsLcar), proven to be critical for dimerization of the enzyme. The mutation of this sole residue resulted in a total loss of activity and prevented the formation of the dimer in BsLcar. Comparative studies revealed that the dimerization domain of the peptidase M20/M25/M40 family is a "small-molecule binding domain," allowing further evolutionary considerations for this enzyme family.

Fig 1

(A) Dimeric arrangement of BsLcar showing the positions of the cobalt cations (magenta) and the isopropanol and cacodylate molecules (red and white). (B) Domain architecture of monomeric BsLcar, showing the α/β/α sandwich catalytic domain (green and orange) connected through the hinge region to the βαββαβ dimerization domain (red and purple). Secondary structure elements not shown by Pymol (β8 and β10) but calculated by PROMOTIF (30) through the PDBsum server (34) were inserted manually. (C) Superposition of BsLcar chains A and B, showing the differences produced by domain motion.

Fig 2

(A) Extended eight-stranded antiparallel β-sheet formed by dimerization of BsLcar chains A (gold) and B (purple). The relative positions of the binding residues Arg²⁸⁶, His²²⁵, and Asn²⁷³ and the newly identified important residues for l-carbamoylases (Arg²³⁴, Thr²⁶¹, and Asp²⁸⁴) are shown. (B) Close-up view of the same residues showing environmental distances (colors are the same as in panel A).

Fig 3

(A) SEC-HPLC analysis of the purified wild-type and mutated BsLcar species and of a gel filtration standard (Bio-Rad) (dotted line). The peaks corresponding to the gel filtration standard are as follows: for thyroglobulin, 670 kDa; for bovine gamma-globulin, 158 kDa; for chicken ovalbumin, 44 kDa; and for equine myoglobin, 17 kDa. Data for wild-type BsLcar (line 1) and the R234A mutant (line 7) appear as solid lines, whereas data for the rest of the mutants (lines 2 to 6) appear as dashed lines. Numeration is as follows: 1, wild-type BsLcar; 2, H225A mutant; 3, N273A mutant; 4, R286A mutant; 5, D284E mutant; 6, T261A mutant; and 7, R234A mutant. The inset corresponds to native PAGE analysis of mutated species, following the same numeration as in the SEC-HPLC analysis. (B) Interactions found in the Arg²³⁴-Thr²⁶¹ environment. The hydrogen bond between both residues appears in blue, whereas other inter- and intrachain interactions are represented in red and black, respectively.

Fig 4

(A) Comparison of the SMBD structures of dimeric BsLcar, monomeric PepD (PDB ID 3MRU), and monomeric PepV (PDB ID 1LFW). Colors represent a spectrum (blue to red) of the order of the secondary elements. The Arg-His-Asn triad is shown as sticks and surrounded by dashed lines. (B) Dimeric arrangement of the ACT domains from d-3-phosphoglycerate dehydrogenase (PDB ID 1YGY) showing similarity to the arrangement of dimeric members of the peptidase M20/M25/M40 family. (C) Stand-alone SMBD belonging to the carboxysome shell protein (PDB ID 2A18), with an Arg residue (sticks) in the same relative position as that for the peptidase M20/M25/M40 family. (D) Topology diagram of the ACT-like domain showing its arrangement and the putative insertion zone where domain fusion would take place.

Fig 5

Schematic representation of the putative evolution of the peptidase M20/M25/M40 family through domain insertion of ACT-like domains into an ancient cc-peptidase. Divergent evolution of the mono-SMBD peptidase and double-SMBD PepD-type ancestor and convergent evolution of the PepD-type ancestor are suggested from the structural findings of this work.