Principal Investigator David Ollis Project r11

Research School of Chemistry Machine VP

Co-Investigators Allan Beveridge

Research School of Chemistry

Dienelactone Hydrolase: Protein Engineering

Dienelactone hydrolase (DLH) is an enzyme used by bacteria and fungi in the degradation of aromatic compounds. It is a monomeric protein of 25,000 Daltons that catalyses the hydrolysis of a lactone (dienelactone). The structure of the protein was determined using x-ray crystallography (1,2) and was found to be an alpha/beta protein. The active site contains a catalytic triad of three residues that are linked by H-bonds in a manner similar to that first found in the serine protease chymotrypsin; an enzyme which has quite different overall structures to that of DLH. When its structure was first determined, the topology of DLH was unique, however, since that time a number of other proteins have been found with catalytic domains that have essentially the same topology as DLH. Like DLH these other enzymes also have a catalytic triad that was very similar to that of DLH. DLH is an ideal protein with which to study protein structure and function and it is also an ideal system with which to pursue protein engineering. It is the simplest member of the a/b hydrolase fold enzymes so that information obtained for DLH will contribute to the understanding of numerous other enzymes.

What are the basic questions being addressed?

The active site of DLH, like the serine proteinases, consists of a catalytic triad. In the case of the serine proteases this triad consists of an asparate, a histidine and a serine which serves as the nucleophile. The triad found in DLH differ from that of the serine proteases in that the nucleophile is a cysteine residue. The triad found in DLH is unusual for two reasons. First, the DLH triad can be produced in the serine proteases using site directed mutagenesis--the resulting enzyme has virtually no activity. Secondly, the nucleophile in DLH can adopt two conformations and what appears to be the more favoured conformation is pointing away from the active site. These two anomalies can be resolved if it is argued that the substrate induces a conformational and chemical change in the nucleophile. The physical evidence to support this theory is scant, but the calculations described below give clear support for the theory.

What are the results to date and the future of the work?

Ab initio SCF/MP2 calculations (using Gaussian92) were performed on a small portion of the active site to demonstrate that the substrate activates the thiol group of the nucleophile, C123 (Cysteine-123), when it binds to the active site by reversing the relative acidities of E36 (glutamic acid 36) and C123. In the native enzyme E36 is ionised and forms a hydrogen bond with the thiol group of C123. Substrate binding destabilises the carboxylate anion of E36 forcing it to abstract the thiol proton. The thiolate anion then undergoes a significant conformational rearrangement before attacking the scissile bond of the substrate. Comparative calculations on both the native enzyme and the inhibitor bound complex have been performed by using a set of small molecules (e.g. imidazole to represent the side chain of histidine) to model the active site residues E36,C123,H202 (histidine 202), D171 (asparatic acid 171), R206 (arginine 206), S208 (serine 208), R81 (arginine 81) and the bound inhibitor. A comparison between the crystal structures of native DLH, and the mutant C123S DLH (in this mutant residue 123 has been converted from a cysteine to a serine) complexed with dienelactam indicates that the inhibitor (or substrate) induces a small but significant shift in the position of E36. Our calculations have now demonstrated that this small shift in the position of E36 is enough to reverse the relative acidities of E36 and C123, resulting in the formation of a highly nucleophilic thiolate anion. These calculations have therefore provided supporting evidence for the initial stages of the catalytic mechanism outlined above.

What computational techniques are used and why is the supercomputer required?

Gaussian92 has been used for all calculations; these calculations could not have been completed using computers within the Research School of Chemistry.

Publications

The theoretical work verifying this novel method of substrate activation has been accepted for publication in Protein Engineering.