We have identified a new protein fold--the alpha/beta hydrolase fold--that is common to several hydrolytic enzymes of widely differing phylogenetic origin and catalytic function. The core of each enzyme is similar: an alpha/beta sheet, not barrel, of eight beta-sheets connected by alpha-helices. These enzymes have diverged from a common ancestor so as to preserve the arrangement of the catalytic residues, not the binding site. They all have a catalytic triad, the elements of which are borne on loops which are the best-conserved structural features in the fold. Only the histidine in the nucleophile-histidine-acid catalytic triad is completely conserved, with the nucleophile and acid loops accommodating more than one type of amino acid. The unique topological and sequence arrangement of the triad residues produces a catalytic triad which is, in a sense, a mirror-image of the serine protease catalytic triad. There are now four groups of enzymes which contain catalytic triads and which are related by convergent evolution towards a stable, useful active site: the eukaryotic serine proteases, the cysteine proteases, subtilisins and the alpha/beta hydrolase fold enzymes.

Based on the recently determined X-ray structures of Torpedo californica acetylcholinesterase and Geotrichum candidum lipase and on their three-dimensional superposition, an improved alignment of a collection of 32 related amino acid sequences of other esterases, lipases, and related proteins was obtained. On the basis of this alignment, 24 residues are found to be invariant in 29 sequences of hydrolytic enzymes, and an additional 49 are well conserved. The conservation in the three remaining sequences is somewhat lower. The conserved residues include the active site, disulfide bridges, salt bridges, and residues in the core of the proteins. Most invariant residues are located at the edges of secondary structural elements. A clear structural basis for the preservation of many of these residues can be determined from comparison of the two X-ray structures.

The three-dimensional structure of the lumenal domain of the lectin-like chaperone calnexin determined to 2.9 A resolution reveals an extended 140 A arm inserted into a beta sandwich structure characteristic of legume lectins. The arm is composed of tandem repeats of two proline-rich sequence motifs which interact with one another in a head-to-tail fashion. Identification of the ligand binding site establishes calnexin as a monovalent lectin, providing insight into the mechanism by which the calnexin family of chaperones interacts with monoglucosylated glycoproteins.

The structure of Candida rugosa lipase in a new crystal form has been determined and refined at 2.1 A resolution. The lipase molecule was found in an inactive conformation, with the active site shielded from the solvent by a part of the polypeptide chain-the flap. Comparison of this structure with the previously determined "open" form of this lipase, in which the active site is accessible to the solvent and presumably the substrate, shows that the transition between these 2 states requires only movement of the flap. The backbone NH groups forming the putative oxyanion hole d o not change position during this rearrangement, indicating that this feature is preformed in the inactive state. The 2 lipase conformations probably correspond to states at opposite ends of the pathway of interfacial activation. Quantitative analysis indicates a large increase of the hydrophobic surface in the vicinity of the active site. The flap undergoes a flexible rearrangement during which some of its secondary structure refolds. The interactions of the flap with the rest of the protein change from mostly hydrophobic in the inactive form to largely hydrophilic in the "open" conformation. Although the flap movement cannot be described as a rigid body motion, it has very definite hinge points at Glu 66 and at Pro 92. The rearrangement is accompanied by a cis-trans isomerization of this proline, which likely increases the energy required for the transition between the 2 states, and may play a role in the stabilization of the active conformation at the water/lipid interface. Carbohydrate attached at Asn 351 also provides stabilization for the open conformation of the flap. Keywords: crystallography; interfacial activation; lipasesLipases of known 3-dimensional structure show significant similarities in their topologies and conform in full or in part to the a/fl hydrolase fold ( Ollis et al., 1992). The catalytic machinery of lipases consists of a serine protease-like triad, Ser-His-Asp/Glu, and their hydrolysis of ester bonds of triacylglycerols is thought to involve an enzymatic mechanism similar to that of the serine proteases (Chapus et al., 1988). The activity of lipases is dramatically enhanced by the presence of a lipid/water interface (Desnuelle, 1972), and it is now well accepted that the phenomenon of interfacial activation involves a conformational change in the enzyme. This has been documented for 2 lipases, one from Rhizomucor miehei (Brzozowski et ai., 199i) and another from human pancreas (van Tilbeurgh et al., 1993), in which binding of an inhibitor or substrate analog causes a conformational rearrangement of 1 or more loops near the active site, exposing the serine nucleophile to the solvent and creating also an oxyanion hole in the process.A similar rearrangement was also proposed for larger lipases from Geotrichum candidurn (GCL;Schrag & Cygler, 1993) Candida rugosa (CRL; Grochubki et ai., 1993). These 2 enzymes show -40% amino acid sequence identity, and their overall 3D structures are very similar (Grochulski ...

. The interfacial activation of Pseudomonas lipases involves conformational rearrangements of surface loops and appears to conform to models of activation deduced from the structures of fungal and mammalian lipases. Factors controlling the conformational rearrangement are not understood, but a comparison of crystallization conditions and observed conformation suggests that the conformation of the protein is determined by the solution conditions, perhaps by the dielectric constant.

The Ser-His-Asp triad is a well known structural feature of the serine proteases. It has also been directly observed in the catalytic sites of two lipases, whose high-resolution three-dimensional structures have been determined 1,2. Lipases show a wide variety of sizes, substrate and positional specificities, and catalytic rates 3. They achieve maximal catalytic rates at oil-water interfaces. The fungus Geotrichum candidum produces several different forms of lipases, two of which have been purified to homogeneity 4,5. Two lipase genes have been identified, cloned and sequenced 6,7. Both code for proteins of 544 amino acids with a total relative molecular mass of about 60,000 (Mr 60K). The two forms are 86% identical. Their isoelectric points differ slightly, being between 4.3 and 4.6. About 7% of the total Mr is carbohydrate. Until now, only a low resolution structure of GCL has been reported 8, but no high resolution structure has followed. We now report the three-dimensional structure of a lipase from G. candidum (GCL) at 2.2 A resolution. Unlike the other lipases and serine proteases, the catalytic triad of GCL is Ser-His-Glu, with glutamic acid replacing the usual aspartate. Although the sequence similarity with the other two lipases is limited to the region near the active-site serine, there is some similarity in their three-dimensional structures. The GCL is also an alpha/beta protein with a central mixed beta sheet whose topology is similar to that of the N-terminal domain of human pancreatic lipase. As in the other lipases 1,2, the catalytic site is buried under surface loops. Sequence comparisons with proteins from the cholinesterase family suggest that they also contain the Ser-His-Glu triad.

No abstract

Serine/threonine protein kinases of the Ste20/PAK family have been implicated in the signalling from heterotrimeric G proteins to mitogen-activated protein (MAP) kinase cascades. In the yeast Saccharomyces cerevisiae, Ste20 is involved in transmitting the mating-pheromone signal from the betagamma-subunits (encoded by the STE4 and STE18 genes, respectively) of a heterotrimeric G protein to a downstream MAP kinase cascade. We have identified a binding site for the G-protein beta-subunit (Gbeta) in the non-catalytic carboxy-terminal regions of Ste20 and its mammalian homologues, the p21-activated protein kinases (PAKs). Association of Gbeta with this site in Ste20 was regulated by binding of pheromone to the receptor. Mutations in Gbeta and Ste20 that prevented this association blocked activation of the MAP kinase cascade. Considering the high degree of structural and functional conservation of Ste20/PAK family members and G-protein subunits, our results provide a possible model for a role of these kinases in Gbetagamma-mediated signal transduction in organisms ranging from yeast to mammals.