Search

X-ray studies of acid proteases indicate a bilobal structure with a well defined active site cleft. An intramolecular two-fold symmetry axis relates two topologically similar domains and the active site residues. A possible mechanism for evolution by gene duplication, divergence and gene fusion is presented.

The structure of alpha-lytic protease, a serine protease produced by the bacterium Lysobacter enzymogenes, has been refined at 1.7 A resolution. The conventional R-factor is 0.131 for the 14,996 reflections between 8 and 1.7 A resolution with I greater than or equal to 2 sigma (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 A. The refined structure was compared with homologous enzymes alpha-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.

Crystals of the serine protease inhibitor CI-2 from the high-lysine barley Hiproly, suitable for high-resolution X-ray crystallographic analysis have been grown from ammonium sulphate solution by vapour diffusion. The space group is P 622 with unit cell dimensions a = b = 69·1(1) A, c = 52·9(1) A and γ = 120°. There is one molecule of molecular weight 9100 per asymmetric unit.

Penicillopepsin is an acid protease produced by the mold Penicillium janthinellum at pH’s less than 4.1 (1). Enzyme production occurs after the mycelial growth has ceased and sporulation has begun (2). The specificity and catalytic mechanism of penicillo-pepsin are very similar to those of porcine pepsin (3). The two active site aspartic acid residues, Asp-32 and Asp-215, occur in peptide sequences of at least eight amino acid residues which are almost identical in penicillopepsin, pepsin and chymosin (1,4–10). In addition, there is an overall 32% identity of amino acid sequence between penicillopepsin and porcine pepsin; a tentative sequence alignment of these two enzymes is given in Table I using this sequence numbering of porcine pepsin (5). These facts indicate that the fungal enzyme penicillopepsin is an evolutionary homologue of the mammalian acid proteases.

Publisher Summary This chapter explores what is known about Troponin C (TnC) from crystallographic studies and reviews how this information can be exploited to understand the nature of the conformational transition that the molecule undergoes upon Ca 2+ binding. TnC belongs to a class of intracellular proteins that are functionally dependent upon Ca 2+ concentration. It plays a key role in the regulation of muscle contraction as part of the regulatory complex troponin. This complex is formed from three protein molecules: troponin T (TnT), which binds to the tropomyosin of the thin filament; troponin I (TnI), which is the subunit that inhibits the Mg 2+ -activated ATPase of actomyosin; and TnC, the subunit which can bind a total of four Ca 2+ ions, two with high affinity (Ka≈10 7 M -1 ) to the C-terminal domain of the molecule and the other two with low affinity (Ka≈10 5 M -1 ) to the N terminal domain. It is the binding of Ca 2+ to the low-affinity binding sites (Ca 2+ -specific sites) that triggers muscle contraction, whereas the high affinity sites (Ca 2+ /Mg 2+ sites) are always occupied by metal ions under physiological conditions. Therefore, the N-terminal domain is considered the regulatory domain of TnC. Ca 2+ binding to this domain is associated with conformational changes to TnC that are transferred to other components of the muscle system, leading to the activation of muscle.

Publisher Summary This chapter outlines the structural information obtained from crystallographic studies of Troponin C (TnC) from turkey skeletal muscle. TnC is one of three proteins that make up troponin, a complex that plays a regulatory role in muscle contraction. TnC is a dumb-belled shaped 75A long molecule, made up of two domains. Each domain is made of two Ca2+-binding motifs of helix-loop-helix and the two motifs in a pair are internally related by an approximate noncrystallographic 2-fold rotation. The inter domain helix is an unusual aspect of the TnC molecule, not found so far in any other protein except calmodulin. TnC crystal structure provides the only example of Ca2+ binding loops in both Ca2+-bound and Ca2+-free states. The inter-domain helix is a unique feature of TnC and calmodulin. The viability of the transition from the observed Ca2+ free state to the modeled Ca2+ bound state was confirmed by carrying out a guided energy minimization. There is at least one pathway through which the opening of the inter-helix angles and the rearrangement of the Ca2+ binding loops can be accomplished smoothly.

The three-dimensional structure of Streptomyces griseus protease B, SGPB, a bacterial serine protease, has the following catalytically important features in common with the chymotrypsin family of enzymes: the histidine-57 loop (42–58), the aspartate-102 loop (86–108), the serine-195 loop (191–220) and an S 1 subsite depression on the molecular surface involving residues 192A-193 and 215–218. The N-terminal half of the molecule (~75 residues) adopts a similar tertiary structure to that of the pancreatic enzymes, whereas those residues in the region 123–190 have a completely different polypeptide chain conformation. The tertiary-structural similarities, coupled with the sequence homologies, albeit slight, indicate that the serine proteases are members of an ancient class of enzymes having a primordial ancestor common to the mammals and bacteria. The conformational differences provide the first evidence of how protein molecules evolve their molecular architecture to acquire additional or altered functions; in the case of the pancreatic enzymes this new function is the zymogen precursor of the active form of the enzyme. These evolutionary changes involve the addition or deletion of external loops while still preserving the catalytically active structural core of the molecule.

A comparison of the solution n.m.r. structures of barley serine protease inhibitor 2 (BSPI-2) with the X-ray structures of both subtilisin complexed and native BSPI-2 is presented. It is shown that the n.m.r. and X-ray structures are very similar in terms of overall shape, size, polypeptide fold and secondary structure. The average atomic rms difference between the 11 restrained dynamics structures on the one hand and the two X-ray structures on the other is 1.9 +/- 0.2 A for the backbone atoms and 3.0 +/- 0.3 A for all atoms. The corresponding values for the restrained energy minimized mean dynamics structure are 1.5 and 2.4 A, respectively.

Three crystalline modifications of the bacterial serine peptidase Streptomyces griseus Protease B have been grown. A 4.5 Å resolution electron density map of one form has been computed from the multiple isomorphous replacement (MIR) phases deduced from two heavy metal derivatives plus the anomalous dispersion effects of one of the derivatives. The crystalline modification used was grown from 0.7–1.0 M KH2PO4 at pH 4.2. These crystals have space group P21212 and unit cell dimensions of a = 44.15 (5) Å, b = 108.72 (10) Å, and c = 37.28 (5) Å. The crystal asymmetric unit contains a protein mass of approximately 19 000 daltons. The electron density map, mean figure of merit 0.80, clearly shows the molecular boundary; relatively long stretches of extended chain are discernable. The active site has been identified from a difference electron density map computed using the MIR protein phases and the amplitude differences between a crystal of the enzyme inhibited at the active serine in solution by p-iodobenzenesulfonyl fluoride (PIPSYL) and those from a crystal of the native enzyme. In addition to showing the site of the PIPSYL binding, there is an apparent conformational change in which the histidine side chain moves away from the serine residue by approximately 4.3 Å.

The X-ray crystal structures of native penicillopepsin and of its complex with a synthetic analogue of the inhibitor pepstatin have been refined recently at 1.8-A resolution. These highly refined structures permit a detailed examination of peptide hydrolysis in the aspartic proteinases. Complexes of penicillopepsin with substrate and catalytic intermediates were modeled, by using computer graphics, with minimal perturbation of the observed inhibitor complex. A thallium ion binding experiment shows that the position of solvent molecule O39, between Asp-33(32) and Asp-213(215) in the native structure, is favorable for cations, a fact that places constraints on possible mechanisms. A mechanism for hydrolysis is proposed in which Asp-213(215) acts as an electrophile by protonating the carbonyl oxygen of the substrate, thereby polarizing the carbon-oxygen bond, a water molecule bound to Asp-33(32) (O284 in the native structure) attacks the carbonyl carbon as the nucleophile in a general-base mechanism, the newly pyramidal peptide nitrogen is protonated, either from the solvent after nitrogen inversion or by an internal proton transfer via Asp-213(215) from a hydroxyl of the tetrahedral carbon, and the tetrahedral intermediate breaks down in a manner consistent with the stereoelectronic hypothesis. The models permit the rationalization of observed subsite preferences for substrates and may be useful in predicting subsite preferences of other aspartic proteinases.

Much of our present understanding of macromolecular architecture has come from the extrapolation to large molecules of rules derived from analysis of small molecules. This is exemplified by the work of Pauling et al.1 on polypeptide conformation, which depended on the insight afforded by the crystal structures of individual amino acids and some dipeptides. Now that accurate protein crystal structures are available, it appears that most of the hydrogen-bonding patterns that are observed have already been found in the structures of small molecules. Thus we can learn much about weak interactions in proteins by analysing small molecule crystal structures. Here we comment on the existence of hydrogen bonds between two carboxyl groups, forming what amounts to the acid salts of monobasic carboxylic acids. We believe these interactions are more important in stabilizing the structures of protein crystals and multi-subunit complexes at low pH than has been realized previously. Furthermore, the interaction contributes to hydrolytic activity and may be involved in inter-subunit recognition.

The active site of alpha-lytic protease has been studied with a number of synthetic peptides and compared with similar data published for elastase. The kinetic data indicate that the active site of alpha-lytic protease extends over at least six subsites (S4--S2'). This extended active site has the effect of increasing kcat/Km by more than 10(6)-fold on going from an N-acetylated amino acid amide to a hexapeptide, due mainly to increases in kcat. There are major differences between alpha-lytic protease and elastase, both in terms of the kinetic parameters for a number of substrates and in terms of the tertiary structure of the active site. The ability of the S1 subsite to interact with various P1 amino acid side chains differs markedly between the two enzymes, and can be rationalized in terms of the tertiary structural differences. For alpha-lytic protease, enzyme-substrate interactions made in subsite S2 appears to be of primary importance, whereas subsite S4 is most important for elastase. The tertiary structural homology of alpha-lytic protease with another bacterial serine protease, Streptomyces griseus protease A, has allowed detailed model building of a tetrapeptide Ac-Pro-Ala-Pro-Ala-OH at the active site. In this way, the subsites S1--S4 have been examined for alpha-lytic protease and compared to other serine proteases.

Hen egg white lysozyme was the first enzyme whose structure was determined by X-ray crystallography. The proposed mechanism based on this structure involves the distortion of the saccharide residue (2-acetamido-2-deoxy-D-muramic acid, NAM) in the natural substrate (an alternating beta (1 leads to 4) linked oligomer of 2-acetamido-2-deoxy-D-glucose (NAG) and NAM residues) bound to site D in the binding cleft. The importance of substrate distortion has prompted numerous enzymatic, chemical, theoretical, and physical studies, but there is little direct crystallographic evidence on the conformation of a NAM residue bound at site D. We now present the X-ray structure of the non-hydrolysed trisaccharide NAM-NAG-NAM bound in subsites B, C, D. Our interpretation of the 2.5-A resolution difference map does not involve distortion of this residue in site D. Comparison with the structure of the delta-lactone derived from tetra N-acetylchitotetraose (NAG)3NAL) bound to lysozyme suggests we may be looking at a Michaelis complex.

Crystals of troponin C from turkey skeletal muscle suitable for high resolution X-ray studies have been grown from a Ca2+-containing solution. These crystals diffract to at least 2.5 A resolution, and have space group P3(1)21 or P3(2)21 and the following cell dimensions: a = b = 66.6 A, c = 61.0 A, alpha = beta = 90 degrees, gamma = 120 degrees.

The crystal structure of 2,4-diselenouracil has been determined to provide some quantitative molecular information about the nature of hydrogen bonding exhibited by this compound in solution. The crystal is monoclinic, with space group P 2 1 /c, with a = 4.41 A., b = 14.90 A., c = 10.69 A., and β = 108.4°. The observed density is 2.42 Gm. cm. −3 and that calculated for four molecules in the unit cell is 2.33 Gm. cm −3 . The structure was determined by the heavy atom technique and refined by block diagonal least squares to an R value of 0.136. The compound was found to exist primarily in a zwitterionic amidic configuration rather than the keto form. The seleniums formed relatively strong hydrogen bonds of the form N (H) … Se with distances of 3.47 and 3.75 A.

α-LYTIC protease is one of several proteolytic enzymes isolated from culture filtrates of a soil bacillus Sorangium sp. (Myxobacter 405)1. Chemical and kinetic studies of this enzyme have shown many remarkable homologies with the pancreatic serine proteases chymotrypsins A and B, trypsin and elastase. For example, α-lytic protease has exactly the same sequence (.Gly.Asp.Ser*.Gly.Gly.) around the active serine as have the pancreatic enzymes1,2. Moreover, the sequence around the single histidine residue in the enzyme (Val.Thr.Ala.Gly.His.Cys.Gly.)3 closely matches the sequence around His 57 of chymotrypsins A4,5 and B6, and the similar sequences around the corresponding histidine residues of trypsin7,8 and pancreatic elastase9.