The refined crystal structure of the complex between human pepsin and a synthetic phosphonate inhibitor, Iva-Val-Val-Leu(P)-(O)Phe-Ala-Ala-OMe [Iva = isovaleryl, Leu(P) = the phosphinic acid analog of L-leucine, (O)Phe = L-3-phenyllactic acid, OMe = methyl ester], is presented. The structure was refined using diffraction data between 30 and 1.96 A resolution to a final R factor ( summation operator| |F(o)| - |F(c)| | / summation operator|F(o)|, where |F(o)| and |F(c)| are the observed and calculated structure-factor amplitudes, respectively) of 20.0%. The interactions of the inhibitor with the enzyme show the locations of the binding sites on the enzyme from S4 to S3'. Modeling of the inhibitor binding to porcine pepsin shows very similar binding sites, except at S4. Comparison of the complex structure with the structures of related inhibitors bound to penicillopepsin helps to rationalize the observed differences in the binding constants. The convergence of reaction mechanisms and geometries in different families of proteinases is also discussed.

Binding constants for complexes of variants of the ovomucoid inhibitor domain 3 from turkey (OMTKY3) and Streptomyces griseus protease B (SGPB) have been computed. On the basis of the crystallographically determined structures of the complexes, continuum electrostatic calculations have been carried out to evaluate the electrostatic contribution to the binding energy. The hydrophobic component was computed based on the change in the solvent accessible surface area on complex formation. These two terms were combined linearly and the parameters for the protein dielectric, atomic solvation parameter and a constant term were derived using a multivariate fit to the observed binding energies. The resulting fit shows a high correlation with a multiple coefficient of determination of 0.79. This indicates that 79% of the variation in the observed binding energies is explained by the electrostatic and hydrophobic terms. The analysis results in a protein dielectric of 8.2 and an atomic solvation parameter of 30 cal/mol Å2. As a test, these parameters were used to calculate the binding energies of complexes of chymotrypsin and of leukocyte elastase OMTKY3, as well as three other variants of OMTKY3 bound to SGPB. As these structures were not used for the multivariate fit, they serve as an independent check on the derived parameters. The calculated energies for the three new variants of OMTKY3 are in good agreement with the observed values. However, the binding energies of the other complexes are poorly predicted. This implies that the parameters that were obtained are not transferable. The possible causes for this lack of transferability are discussed.

The crystal structure of PR3, a serine proteinase from the azurophilic granules of human polymorphonuclear neutrophils, has been solved by molecular replacement using the human leukocyte elastase structure. The PR3 structure has been refined to anR-factor (=Σ∥Fo| − |Fc∥/ Σ |Fo|) of 0.201 for all data in the range of 10.0 to 2.2 Å resolution. The enzyme was crystallized in space groupP21with four molecules in the asymmetric unit (Vm≃ 2.6 Å/Da). The overall fold consists of two domains of β-barrel structures typical of the chymotrypsin family of serine proteinases. In general, the substrate binding sites, S4 to S3′, are more polar than comparable sites in the related proteinase, human leukocyte elastase. The experimentally observed preference of PR3 for small aliphatic residues at the P1 position of a substrate is explained by the Val to Ile substitution at position 190 when compared to the elastase structure. The substitution of Ala by Asp at position 213 at the back of S1 should not affect its specificity greatly, as the Asp side-chain points back into the interior of the protein. The PR3 structure includes a disaccharide unit (N-linked 2-acetamido-2-deoxy-β-D-glucopyranose and 1,6-linked α-L-fucopyranose) covalently attached to Asn159. The linear antigenic sites of PR3 reported to react with Wegener's granulomatosis autoantibodies occur in regions of the three-dimensional structure that may implicate the inactive pro-form of the enzyme in the pathogenesis of the disease.

The three‐dimensional crystal structure of human pepsin and that of its complex with pepstatin have been solved by X‐ray crystallographic methods. The native pepsin structure has been refined with data collected to 2.2 Å resolution to an R‐factor of 19.7%. The pepsin:pepstatin structure has been refined with data to 2.0 Å resolution to an R‐factor of 18.5%. The hydrogen bonding interactions and the conformation adopted by pepstatin are very similar to those found in complexes of pepstatin with other aspartic proteinases. The enzyme undergoes a conformational change upon inhibitor binding to enclose the inhibitor more tightly. The analysis of the binding sites indicates that they form an extended tube without distinct binding pockets. By comparing the residues on the binding surface with those of the other human aspartic proteinases, it has been possible to rationalize some of the experimental data concerning the different specificities. At the S1 site, valine at position 120 in renin instead of isoleucine, as in the other enzymes, allows for binding of larger hydrophobic residues. The possibility of multiple conformations for the P2 residue makes the analysis of the S2 site difficult. However, it is possible to see that the specific interactions that renin makes with histidine at P2 would not be possible in the case of the other enzymes. At the S3 site, the smaller volume that is accessible in pepsin compared to the other enzymes is consistent with its preference for smaller residues at the P3 position.

Penicillopepsin, the aspartyl protease from the mould Penicillium janthinellum, has had its molecular structure refined by a restrained-parameter least-squares procedure at 1.8 Å resolution to a conventional R-factor of 0.136. The estimated co-ordinate accuracy for the majority of the 2363 atoms of the enzyme is better than 0.12 Å. The average atomic thermal vibration parameter, B, for the atoms of the enzyme is 14.5 Å2. One determining factor of this low average B value is the large central hydrophobic core, in which there are two prominent clusters of aromatic residues, one of nine, the other of seven residues. The N and C-terminal domains of penicillopepsin display an approximate 2-fold symmetry: 70 residue pairs are topologically equivalent, related by a rotation of 177 ° and a translation of 1.2 Å. The analysis of the secondary structural features of the molecule reveals non-linear hydrogen bonding. In penicillopepsin, there is no difference in the mean hydrogen-bond parameters for the elements of α-helix, parallel or antiparallel β-pleated sheet. The mean values for these structural elements are: NO, 2.90 Å; NHO, 1.95 Å; NĤO, 160 °. The average hydrogen-bond parameters of the reverse β-turns and the 310 helices are distinctly different from the above values. The analysis of sidechain conformational angles χ1 and χ2 penicillopepsin and other enzyme structures refined in this laboratory shows much narrower distributions as compared with those compiled from unrefined protein structures. The close proximity of the carboxyl groups of Asp33 and Asp213 suggests that they share a proton in a tight hydrogen-bonded environment (Asp33OD2 to Asp213OD1 is 2.87 Å). There are several solvent molecules in the active site region and, in particular, O39 forms hydrogen-bonded interactions with both aspartate residues. The disposition of the two carboxyl groups suggests that neither is likely to be involved in a direct nucleophilic attack on the scissile bond of a substrate. The average atomic B-factors of the residues in this region of the molecule are between 5 and 8 Å2, confirming the proposal that conformational mobility of the active site residues has no role in the enzymatic mechanism. However, conformational mobility of neighbouring regions of the molecule e.g. the "flap" containing Tyr75, is verified by the high B-factors for those residues. The positions of 319 solvent sites per asymmetric unit have been selected from difference electron density maps and refined. Thirteen have been classified as internal, and several of these may have key roles during catalysis. The positively charged Nζ atom of Lys304 forms hydrogen bonds to the carboxylate of Asp14 (internal ion pair) and to two internal water molecules O5 and O25. The protonated side-chain of Asp300 forms a hydrogen bond to Thr214O, 2.78 Å, and is the recipient of a hydrogen bond from a surface pocket water molecule O46. There is no possibility for direct interaction between Asp300 and Lys304 without large conformational changes of their environment. The intermolecular packing involves many protein-protein contacts (66 residues) with a large number of solvent molecules involved in bridging between polar residues at the contact surface. The penicillopepsin molecules resemble an approximate hexagonal close-packing of spheres with each molecule having 12 "nearest" neighbours.

The x-ray crystal structure of recombinant human renin has been determined. Molecular dynamics techniques that included crystallographic data as a restraint were used to improve an initial model based on porcine pepsinogen. The present agreement factor for data from 8.0 to 2.5 angstroms (Å) is 0.236. Some of the surface loops are poorly determined, and these disordered regions border a 30 Å wide solvent channel. Comparison of renin with other aspartyl proteinases shows that, although the structural cores and active sites are highly conserved, surface residues, some of which are critical for specificity, vary greatly (up to 10 Å). Knowledge of the actual structure, as opposed to the use of models based on related enzymes, should facilitate the design of renin inhibitors.

The molecular structure of the complex between bovine pancreatic α-chymotrypsin (EC 3.4.4.5) and the third domain of the Kazal-type ovomucoid from Turkey (OMTKY3) has been determined crystallographically by the molecular replacement method. Restrained-parameter least-squares refinement of the molecular model of the complex has led to a conventional agreement factor R of 0.168 for the 19,466 reflections in the 1.8 Å (1 Å = 0.1 nm) resolution shell [I ≥ σ(I)]. The reactive site loop of OMTKY3, from Lys131 to Arg211 (I indicates inhibitor), is highly complementary to the surface of α-chymotrypsin in the complex. A total of 13 residues on the inhibitor make 113 contacts of less than 4.0 Å with 21 residues of the enzyme. A short contact (2.95 Å) from Oγ of Ser195 to the carbonylcarbon atom of the scissile bond between Leu181 and Glu191 is present; in spite of it, this peptide remains planar and undistorted. Analysis of the interactions of the inhibitor with chymotrypsin explains the enhanced specificity that chymotrypsin has for P′3 arginine residues. There is a water-mediated ion pair between the guanidinium group on this residue and the carboxylate of Asp64. Comparison of the structure of the α-chymotrypsin portion of this complex with the several structures of α and γ-chymotrypsin in the uncomplexed form shows a high degree of structural equivalence (root-mean-square deviation of the 234 common α-carbon atoms averages 0.38 Å). Significant differences occur mainly in two regions Lys36 to Phe39 and Ser75 to Lys79. Among the 21 residues that are in contact with the ovomucoid domain, only Phe39 and Tyr146 change their conformations significantly as a result of forming the complex. Comparison of the structure of the OMTKY3 domain in this complex to that of the same inhibitor bound to a serine proteinase from Streptomyces griseus (SGPB) shows a central core of 44 amino acids (the central α-helix and flanking small 3-stranded β-sheet) that have α-carbon atoms fitting to within 1.0 Å (root-mean-square deviation of 0.45 Å) whereas the residues of the reactive-site loop differ in position by up to 1.9 Å (Cα of Leu181). The ovomucoid domain has a built-in conformational flexibility that allows it to adapt to the active sites of different enzymes. A comparison of the SGPB and α-chymotrypsin molecules is made and the water molecules bound at the inhibitor-enzyme interface in both complexes are analysed for similarities and differences.

The structure of α-lytic protease, a serine protease produced by the bacterium Lysobacter enzymogenes, has been refined at 1.7 Å resolution. The conventional R-factor is 0.131 for the 14,996 reflections between 8 and 1.7 Å resolution with I ≥ 2σ(I). The model consists of 1391 protein atoms, two sulfate ions and 156 water molecules. The overall root-meansquare error is estimated to be about 0.14 Å. The refined structure was compared with homologous enzymes α-chymotrypsin and Streptomyces griseus protease A and B. A new sequence numbering was derived based on the alignment of these structures. The comparison showed that the greatest structural homology is around the active site residues Asp102, His57 and Ser195, and that basic folding pathways are maintained despite chemical changes in the hydrophobic cores. The hydrogen bonds in the structure were tabulated and the distances and angles of interaction are similar to those found in small molecules. The analysis also revealed the presence of close intraresidue interactions. There are only a few direct intermolecular hydrogen bonds. Most intermolecular interactions involve bridging solvent molecules. The structural importance of hydrogen bonds involving the side-chain of Asx residues is discussed. All the negatively charged groups have a counterion nearby, while the excess positively charged groups are exposed to the solvent. One of the sulfate ions is located near the active site, whereas the other is close to the N terminus. Of the 156 water molecules, only seven are not involved in a hydrogen bond. Six of these have polar groups nearby, while the remaining one is in very weak density. There are nine internal water molecules, consisting of two monomers, two dimers and one trimer. No significant second shell of solvent is observed.

We have determined the crystal structure of the molecular complex between Streptomyces griseus protease B (SGPB), a bacterial serine protease, and the third domain of the ovomucoid inhibitor from turkey. Restrained-parameter least-squares refinement of the structure with the 1.8-A intensity data set has resulted in an R factor of 0.125. The carbonyl carbon atom of the reactive bond between Leu-18 and Glu-19 in the inhibitor lies at a distance of 2.71 A from the O gamma atom of the nucleophilic Ser-195 in SGPB; this distance is 0.5 A shorter than a normal van der Waals contact. Unlike the reactive bond in the pancreatic trypsin inhibitor complexed with bovine trypsin, the Leu--Glu bond of the ovomucoid inhibitor is not distorted from planarity towards a pyramidal configuration.