Rational Design of R-Keto Triglyceride Analogues as Inhibitors for Staphylococcus hyicus Lipase †,‡ Jan-Willem F. A. Simons, Ruud C. Cox, Maarten R. Egmond,* and Hubertus M. Verheij Department of Enzymology and Protein Engineering, Centre for Biomembranes and Lipid Enzymology, Institute of Biomembranes, Utrecht UniVersity, The Netherlands ReceiVed October 20, 1998; ReVised Manuscript ReceiVed January 27, 1999 ABSTRACT: We have synthesized a series of R-keto triglyceride analogues as inhibitors for the lipase from Staphylococcus hyicus (SHL). Hydrolysis at positions 1 and 2 was prevented by replacement of the ester bonds by nonhydrolyzable ether, carbamoyl, or amide bonds, and an R-keto fatty acid was introduced at position 3. Such compounds act as competitive inhibitors of SHL. Inhibition must be caused by the presence of the R-keto functions, since the compounds containing an ester or a hydroxyl group in position 3 did not inhibit the enzyme. We propose that our inhibitors react with the active site Ser of the lipase, hereby mimicking the tetrahedral intermediate occurring in substrate hydrolysis. We conclude that the localization of the R-keto triglycerides is very important for inhibition because only those compounds which are insoluble in water are lipase inhibitors. In addition, other specific protein-inhibitor interactions, probably with the carbonyl oxygen at position 1 and/or 2, improve inhibitor binding. This makes the compounds with amide or carbamoyl groups at positions 1 and 2 better inhibitors than their ether counterparts. The inhibitory power could be improved further by replacing the oxygen at position 3 by an amido group. The resulting inhibitor 1,2-diethylcarbamoyl-3-amido-R-ketododecanoylglycerol has a K i * value of 0.008 mol %, indicating that it binds approximately 125-fold tighter than the substrate. These results illustrate that effective lipase inhibitors can be designed by combining an R-keto group with good micellar solubility and optimal protein-inhibitor interactions. Lipases (glycerol ester hydrolases, EC 3.1.1.3) are active at the lipid-water interface where they degrade water- insoluble triglycerides. Not only triglycerides but also numerous synthetic lipids and phospholipids are hydrolyzed by lipases, often with stereo, positional, and/or chain length selectivity (1, and references therein). Moreover, lipases are very stable in organic solvents, and under these conditions, they can catalyze (trans)esterification reactions (2, 3). These properties make lipases suitable biocatalysts for the applica- tion in organic synthesis and as detergent additives, and at present, these enzymes are produced on a bulk scale for industrial purposes. The challenge for industry is the optimization and the de novo design of lipase substrate specificities, for catalyzing specific reactions. To achieve this, one requires a better understanding of how these enzymes function and, moreover, how they interact with substrate molecules at the molecular level. The determination of the first X-ray structure of a lipase from human pancreas has substantially improved our un- derstanding of lipases (4). This structure confirmed the presence of a “classical” Ser-Asp-His catalytic triad, com- parable to the active site of proteases, but in the lipase, the active site was blocked by surface loops, thereby making it inaccessible. The subsequently published structure of this lipase bound to a substrate micelle revealed that this inter- action results in the structural rearrangement of these loops, now making the active site accessible to substrate molecules (5). Comparable conformational changes were also observed for many other lipases, either complexed with active site- directed inhibitors (6, 7), bound to micelles (8), or crystallized under specific conditions (9, 10). As a result, the molecular basis for the opening of the active site is thoroughly understood today, but it is the next step in the catalytic event that determines the selectivity of the enzyme. In this step, the substrate molecule specifically interacts with the protein, resulting in the optimal orientation for chemical hydrolysis by the catalytic residues. These highly specific enzyme- substrate interactions are only poorly understood because the current understanding is based on X-ray structures of lipases inhibited by small monoalkyl organophosphonates. These structures revealed the presence of a so-called oxyanion hole, where during catalysis the negatively charged transition state intermediate is stabilized by specific hydrogen bonds with the protein. The alkyl moiety of the inhibitor appeared to be bound in either a hydrophobic binding pocket inside the protein or a groove on the surface of the protein (11-13). These observations provided the basis for the first models of lipase-triglyceride interaction and stereoselectivity, but for a full understanding of all protein-substrate interactions, the inhibitor ideally should have a structure that is as close † This research was carried out with the financial support from the BRIDGE-T-lipase and Biotech-G-lipase programmes of the European Communities under Contract N-BIOT-0194 NL and BIO2 CT94-3013. ‡ Abstract published previously [(1997) FASEB J. 11 (9)]. * To whom correspondence should be addressed: Department of Enzymology and Protein Engineering, CBLE, Utrecht Univer- sity, Padualaan 8, P.O. Box 80054, NL-3508 TB Utrecht, The Neth- erlands. Phone: +31-30-2533526. Fax: +31-30-2522478. E-mail: m.r.egmond@chem.uu.nl. 6346 Biochemistry 1999, 38, 6346-6351 10.1021/bi982498b CCC: $18.00 © 1999 American Chemical Society Published on Web 04/16/1999