The high-resolution crystal structure of porcine pepsinogen.

The structure of porcine pepsinogen at pH 6.1 has been refined to an R-factor of 0.173 for data extending to 1.65 A. The final model contains 180 solvent molecules and lacks density for residues 157-161. The structure of this aspartic proteinase zymogen possesses many of the characteristics of pepsin, the mature enzyme. The secondary structure of the zymogen consists predominantly of beta-sheet, with an approximate 2-fold axis of symmetry. The activation peptide packs into the active site cleft, and the N-terminus (1P-9P) occupies the position of the mature N-terminus (1-9). Thus changes upon activation include excision of the activation peptide and proper relocation of the mature N-terminus. The activation peptide or residues of the displaced mature N-terminus make specific interactions with the substrate binding subsites. The active site of pepsinogen is intact; thus the lack of activity of pepsinogen is not due to a deformation of the active site. Nine ion pairs in pepsinogen may be important in the advent of activation and involve the activation peptide or regions of the mature N-terminus which are relocated in the mature enzyme. The activation peptide-pepsin junction, 44P-1, is characterized by high thermal parameters and weak density, indicating a flexible structure which would be accessible to cleavage. Pepsinogen is an appropriate model for the structures of other zymogens in the aspartic proteinase family.     

Analysis of the interactions between streptokinase domains and human plasminogen.

The contrasting roles of streptokinase (SK) domains in binding human Glu1-plasminogen (Plg) have been studied using a set of proteolytic fragments, each of which encompasses one or more of SK's three structural domains (A, B, C). Direct binding experiments have been performed using gel filtration chromatography and surface plasmon resonance. The latter technique has allowed estimation of association and dissociation rate constants for interactions between Plg and intact SK or SK fragments. Each of the SK fragments that contains domain B (fragments A2-B-C, A2-B, B-C, and B) binds Plg with similar affinity, at a level approximately 100- to 1,000-fold lower than intact SK. Experiments using 10 mM 6-aminohexanoic acid or 50 mM benzamidine demonstrate that either of these two lysine analogues abolishes interaction of domain B with Plg. Isolated domain C does not show detectable binding to Plg. Moreover, the additional presence of domain C within other SK fragments (B-C and A2-B-C) does not alter significantly their affinities for Plg. In addition, Plg-binding by a noncovalent complex of two SK fragments that contains domains A and B is similar to that of domain B. By contrast, species containing domain B and both domains A and C (intact SK and the two-chain complex A1 x A2-B-C) show a significantly higher affinity for Plg, which could not be completely inhibited by saturating amounts of 6-AHA. These results show that SK domain B interacts with Plg in a lysine-dependent manner and that although domains A and C do not appear independently to possess affinity for Plg, they function cooperatively to establish the additional interactions with Plg to form an efficient native-like Plg activator complex.     

Evolution in the structure and function of aspartic proteases

Aspartic proteases (EC3.4.23) are a group of proteolytic enzymes of the pepsin family that share the same catalytic apparatus and usually function in acid solutions. This latter aspect limits the function of aspartic proteases to some specific locations in different organisms; thus the occurrence of aspartic proteases is less abundant than other groups of proteases, such as serine proteases. The best known sources of aspartic proteases are stomach (for pepsin, gastricsin, and chymosin), lysosomes (for cathepsins D and E), kidney (for renin), yeast granules, and fungi (for secreted proteases such as rhizopuspepsin, penicillopepsin, and endothiapepsin). These aspartic proteases have been extensively studied for their structure and function relationships and have been the topics of several reviews or monographs (Tang: Acid Proteases, Structure, Function and Biology. New York: Plenum Press, 1977; Tang: J Mol Cell Biochem 26:93-109, 1979; Kostka: Aspartic Proteinases and Their Inhibitors. Berlin: Walter de Gruyter, 1985). All mammalian aspartic proteases are synthesized as zymogens and are subsequently activated to active proteases. Although a zymogen for a fungal aspartic protease has not been found, the cDNA structure of rhizopuspepsin suggests the presence of a "pro" enzyme (Wong et al: Fed Proc 44:2725, 1985). It is probable that other fungal aspartic proteases are also synthesized as zymogens. It is the aim of this article to summarize the major models of structure-function relationships of aspartic proteases and their zymogens with emphasis on more recent findings. Attempts will also be made to relate these models to other aspartic proteases.     

On the activation of the canine pepsinogens.

Acidification induces a conversion of canine pepsinogens by a sequential mechanism to the active pepsins. Activation in the presence of pepstatin, which strongly inhibits the pepsins but does not prevent the first step of activation, allows the isolation of the peptide released in this first step. This peptide inhibits the milk clotting activity of canine and also porcine pepsin. Canine pepsins obtained in the absence of pepstatin were characterized by amino acid composition, molecular weight, and activity against hemoglobin and milk and compared with those of other mammalian pepsins.     

Structure and function of microplasminogen: I. Methionine shuffling, chemical proteolysis, and proenzyme activation.

We have cloned and expressed microplasminogen (mPlg), consisting of the N-terminal undecapeptide of human glu-Plg spliced to its proenzyme domain. This truncated (approximately 28 kDa) proenzyme retained the distinctive catalytic activities of the larger parent. Replacement of M residues followed by M shuffling permitted subsequent scission by site-directed chemical proteolysis (in CNBr/formic acid) without impairing any of the protein's characteristic properties. Activation of chymotrypsinogen-related zymogens occurs by limited proteolysis; the newly liberated, highly conserved N-terminus (VVGG) forms a salt bridge with an aspartyl residue immediately upstream of the active site serine. The role of both of these elements in mPlg activation was probed using protein engineering and site-directed proteolysis to alter the length and amino acid composition of the N-terminus, and to replace the aspartate. All modifications affected both Km and Kcat. The results identify some structural parameters of the N-terminus required for proenzyme activation.     

Fungal yapsins and cell wall: a unique family of aspartic peptidases for a distinctive cellular function

A novel class of aspartic peptidases known as fungal yapsins, whose first member ScYps1p was identified more than a decade ago in Saccharomyces cerevisiae, is characteristically modified by the addition of a glycophosphatidylinositol moiety and has a preference for cleaving substrates C-terminally to mono- and paired-basic residues. Over the years, several other members, first in S. cerevisiae and then in other fungi, have been identified. The implication of fungal yapsins in cell-wall assembly and/or remodelling had been suspected for many years. However, it is only very recently that studies performed on S. cerevisae and Candida albicans have confirmed their importance for cell-wall integrity. Here, we review 16 years of research, covering all fundamental aspects of these unique enzymes, in an effort to track their functional significance. We also propose a nomenclature for fungal yapsins based on their sequence identity with the founding members of this family, the S. cerevisiae yapsins.     

Crystal structure of an activation intermediate of cathepsin E

Cathepsin E is an intracellular, non-lysosomal aspartic protease expressed in a variety of cells and tissues. The protease has proposed physiological roles in antigen presentation by the MHC class II system, in the biogenesis of the vasoconstrictor peptide endothelin, and in neurodegeneration associated with brain ischemia and aging. Cathepsin E is the only A1 aspartic protease that exists as a homodimer with a disulfide bridge linking the two monomers. Like many other aspartic proteases, it is synthesized as a zymogen which is catalytically inactive towards its natural substrates at neutral pH and which auto-activates in an acidic environment. Here we report the crystal structure of an activation intermediate of human cathepsin E at 2.35A resolution. The overall structure follows the general fold of aspartic proteases of the A1 family, and the intermediate shares many features with the intermediate 2 on the proposed activation pathway of aspartic proteases like pepsin C and cathepsin D. The pro-sequence is cleaved from the protease and remains stably associated with the mature enzyme by forming the outermost sixth strand of the interdomain beta-sheet. However, different from these other aspartic proteases the pro-sequence of cathepsin E remains intact after cleavage from the mature enzyme. In addition, the active site of cathepsin E in the crystal is occupied by N-terminal amino acid residues of the mature protease in the non-primed binding site and by an artificial N-terminal extension of the pro-sequence from a neighboring molecule in the primed site. The crystal structure of the cathepsin E/pro-sequence complex, therefore, provides further insight into the activation mechanism of aspartic proteases.     

Unusual zymogen-processing properties of a mutated form of prochymosin

Site-specific mutagenesis of the gene encoding bovine prochymosin was used to produce a mutated zymogen in which seven contiguous amino acids of the N-terminal propeptide had been deleted and an eighth residue had been substituted. This altered region spans the normal site of autocatalytic proteolysis that occurs at the same time as (enzymatic) activation of prochymosin at acidic pH. Activation of the mutated zymogen at pH 4.5 was extremely slow, and cleavage occurred at an unusual Ser-Lys bond in the propeptide of the zymogen. The mutated prochymosin incubated at pH 2 generated the usual pseudochymosin by cleavage of the normal Phe-Leu bond, but at a rate severalfold slower than the authentic zymogen. These results indicate that even after deletion of seven of 42 amino acids of the propeptide the mutant protein could still assume a prochymosin (zymogen) structure, although these changes did result in striking differences in acid-catalyzed activation and processing reactions at one but not the other of the two processing sites of prochymosin.     

Interactions of different inhibitors with active-site aspartyl residues of HIV-1 protease and possible relevance to pepsin

The importance of the active site region aspartyl residues 25 and 29 of the mature HIV-1 protease (PR) for the binding of five clinical and three experimental protease inhibitors [symmetric cyclic urea inhibitor DMP323, nonhydrolyzable substrate analog (RPB) and the generic aspartic protease inhibitor acetyl-pepstatin (Ac-PEP)] was assessed by differential scanning calorimetry. DeltaT(m) values, defined as the difference in T(m) for a given protein in the presence and absence of inhibitor, for PR with DRV, ATV, SQV, RTV, APV, DMP323, RPB, and Ac-PEP are 22.4, 20.8, 19.3, 15.6, 14.3, 14.7, 8.7, and 6.5 degrees C, respectively. Binding of APV and Ac-PEP is most sensitive to the D25N mutation, as shown by DeltaT(m) ratios [DeltaT(m)(PR)/DeltaT(m)(PR(D25N))] of 35.8 and 16.3, respectively, whereas binding of DMP323 and RPB (DeltaT(m) ratios of 1-2) is least affected. Binding of the substrate-like inhibitors RPB and Ac-PEP is nearly abolished (DeltaT(m)(PR)/DeltaT(m)(PR(D29N)) > or = 44) by the D29N mutation, whereas this mutation only moderately affects binding of the smaller inhibitors (DeltaT(m) ratios of 1.4-2.2). Of the nine FDA-approved clinical HIV-1 protease inhibitors screened, APV, RTV, and DRV competitively inhibit porcine pepsin with K(i) values of 0.3, 0.6, and 2.14 microM, respectively. DSC results were consistent with this relatively weak binding of APV (DeltaT(m) 2.7 degrees C) compared with the tight binding of Ac-PEP (DeltaT(m) > or = 17 degrees C). Comparison of superimposed structures of the PR/APV complex with those of PR/Ac-PEP and pepsin/pepstatin A complexes suggests a role for Asp215, Asp32, and Ser219 in pepsin, equivalent to Asp25, Asp25', and Asp29 in PR in the binding and stabilization of the pepsin/APV complex.     

Matriptase Expression and Zymogen Activation in Human Pilosebaceous Unit

Studies of human genetic disorders and mouse models reveal the important roles of matriptase in hair growth. Here, we investigate matriptase expression and zymogen activation in hair follicles. We show: 1) layer-dependent distribution patterns, with much higher matriptase expression in cells of the outer root sheath and matrix cells of the hair bulb than in cells of the inner root sheath; 2) cycle-dependent expression patterns, with matriptase expressed in the anagen and catagen phases of the hair lifecycle, but not in the telogen phase; 3) reduced expression of the matriptase inhibitor, HAI-1, in the catagen phase, suggesting increased proteolytic activity in this phase; and 4) definitive matriptase zymogen activation patterns, with the highest matriptase activation observed in matrix cells and outer root sheath cells in the isthmus/bulge region. In sebaceous glands, matriptase is highly expressed in basal and ductal cells, with much lower expression in the differentiated, lipid-filled cells of the interior. We also show that matriptase potently activates hepatocyte growth factor (HGF) in vitro, and that the HGF receptor, c-Met, is co-expressed in those cells that express activated matriptase. Our observations suggest that the matriptase-HGF-c-MET pathway has the potential to be engaged, primarily in proliferative cells rather than terminally differentiated epithelial cells of the human pilosebaceous unit.     

Gene for porcine pregnancy-associated glycoprotein 2 (poPAG2): its structural organization and analysis of its promoter

The pregnancy-associated glycoproteins (PAG) are abundant secretory products of the placental trophectoderm of ungulate species. They are structurally related to pepsin, having the capability to bind peptides. However, many cannot function as enzymes due to amino acid substitutions in and around the catalytic site. Here, we demonstrate that pigs, like cattle and sheep, but unlike equids, have multiple PAG genes. One of the transcribed porcine PAG (poPAG) genes, the one for poPAG2, was cloned. It had a nine-exon organization similar to that of other mammalian aspartic proteinase genes with an atypical TATA sequence. A total of 1.2 kbp upstream from exon 1 was sequenced. This region shared identity (> 65%) with the promoter regions of the bovine (bo) PAG1, boPAG2 and equine (eq) PAG genes, but not with other aspartyl proteinase genes, including that of pepsinogen A. Nor were there clear similarities to the promoters of other genes with trophoblast-specific expression. Of the different poPAG2 promoter constructs tested in transfection experiments in two human (JAr and JEG3) and one rat (Rcho) choriocarcinoma cell lines, only the shortest (-149 bp) was required to provide full expression of a luciferase reporter. Although this short promoter was not active in Cos-1 and L-929 cells, it was active in CHO cells, a transformed non-trophoblast hamster ovarian cell line. Co-transfection of Ets2 elevated the activity of this short promoter approximately six-fold in JAr cells, but, disruption of the two putative Ets sites did not alter the ability of Ets2 to transactivate the promoter. In the non-trophoblast cell lines, Ets2 failed to elicit any response. Ets2 responsiveness may be a common feature of most or all trophoblast-expressed genes, although in the case of poPAG2, the effect may be indirect.     

Crystal structure of human BACE2 in complex with a hydroxyethylamine transition-state inhibitor

BACE2 is a membrane-bound aspartic protease of the A1 family with a high level of sequence homology to BACE1. While BACE1 is involved in the generation of amyloid plaques in Alzheimer's disease by cleaving Abeta-peptides from the amyloid precursor protein, the physiological function of BACE2 is not well understood. BACE2 appears to be associated with the early onset of dementia in patients with Down's syndrome, and it has been shown to be highly expressed in breast cancers. Further, it may participate in the function of normal and abnormal processes of human muscle biology. Similar to other aspartic proteases, BACE2 is expressed as an inactive zymogen requiring the cleavage of its pro-sequence during the maturation process. We have produced mature BACE2 by expression of pro-BACE2 in Escherichia coli as inclusion bodies, followed by refolding and autocatalytic activation at pH 3.4. Using a C and N-terminally truncated BACE2 variant, we were able to crystallize and determine the crystal structure of mature BACE2 in complex with a hydroxyethylamine transition-state mimetic inhibitor at 3.1 angstroms resolution. The structure of BACE2 follows the general fold of A1 aspartic proteases. However, similar to BACE1, its C-terminal domain is significantly larger than that of the other family members. Furthermore, the structure of BACE2 reveals differences in the S3, S2, S1' and S2' active site substrate pockets as compared to BACE1, and allows, therefore, for a deeper understanding of the structural features that may facilitate the design of selective BACE1 or BACE2 inhibitors.     

Stromelysin-1: three-dimensional structure of the inhibited catalytic domain and of the C-truncated proenzyme.

The proteolytic enzyme stromelysin-1 is a member of the family of matrix metalloproteinases and is believed to play a role in pathological conditions such as arthritis and tumor invasion. Stromelysin-1 is synthesized as a pro-enzyme that is activated by removal of an N-terminal prodomain. The active enzyme contains a catalytic domain and a C-terminal hemopexin domain believed to participate in macromolecular substrate recognition. We have determined the three-dimensional structures of both a C-truncated form of the proenzyme and an inhibited complex of the catalytic domain by X-ray diffraction analysis. The catalytic core is very similar in the two forms and is similar to the homologous domain in fibroblast and neutrophil collagenases, as well as to the stromelysin structure determined by NMR. The prodomain is a separate folding unit containing three alpha-helices and an extended peptide that lies in the active site of the enzyme. Surprisingly, the amino-to-carboxyl direction of this peptide chain is opposite to that adopted by the inhibitor and by previously reported inhibitors of collagenase. Comparison of the active site of stromelysin with that of thermolysin reveals that most of the residues proposed to play significant roles in the enzymatic mechanism of thermolysin have equivalents in stromelysin, but that three residues implicated in the catalytic mechanism of thermolysin are not represented in stromelysin.     

Microtube batch protein crystallization: Applications to human immunodeficiency virus type 2 (HIV-2) protease and human renin

For therapeutically relevant targets, the evaluation of enzymes in complex with their inhibitors by cocrystallization and high resolution structural analysis has become a vital component of structure-driven drug design and development. Two approaches, hanging drop vapor diffusion and a novel microtube batch method, were utilized in parallel to grow crystals of recombinant HIV -2 protease and recombinant human renin in complex with inhibitors. In the case of HIV -2 protease in complex with a reduced amide inhibitor, crystallization was achieved only by the microbatch method. In the case of human renin, the addition of precipitant was required for crystal growth. The microbatch method described here is a useful supplementary or alternative approach for screening parameters and generating crystals suitable for high resolution structural analysis. © 1994 Wiley-Liss, Inc.     

The pepsinogen releasing effect of Helicobacter pylori lipopolysaccharide

Background. Helicobacter pylori lipopolysaccharide (LPS) affects pepsinogen release by a nontoxic mechanism. We hypothesized that this effect was characteristic of the organism and related to the clinical status of the strain. Materials and methods. LPS was isolated from 11 H. pylori strains whose pathogenic profile was known and four other nongastric bacteria. The effects of luminal LPS on guinea pig gastric mucosal pepsinogen release was evaluated using the Ussing chamber technique. CCK‐8 (10^?9M) was used as a positive control. Results. H. pylori LPS dose‐dependently stimulated pepsinogen release with a maximal stimulation at 250 µg/ml (~4500%; p < .001 vs. control). LPS from other Helicobacter or Campylobacter species had no effect on pepsinogen release. ANOVA demonstrated significant differences in the efficacies of pepsinogen release between the 11 clinical H. pylori strains (p < .0001) despite the fact that they were all cagA⁺ and 90% had the cytotoxic vacA subtype s1. Physical and chemical disruption of the LPS suggested that both the structure and the carbohydrate composition of this molecule may play a critical role in pepsinogen release. Polymyxin B partly (p < .03) inhibited and dephosphorylation completely inhibited (p = .0002) LPS‐stimulated pepsinogen release. Conclusion. Pepsinogen release is an innate property of all cagA⁺H. pylori LPS. The structure of the molecule and composition of side‐chains are important in this response which appears to be partially lipid A driven.     

Effect of omeprazole on secretion,synthesis and the gene expression of pepsinogen in the guinea pig stomach mucosa

The effects of omeprazole, a proton pump inhibitor, on gene expression, protein synthesis, intracellular storage and secretion of pepsinogen in guinea pig stomach were investigated. After treatment with omeprazole for five days, acid and pepsinogen secretion into the gastric lumen was significantly reduced. Concomitant with this, there was an increase in intracellular pepsinogen as demonstrated by increased pepsin activity in the gastric mucosa, more intense immunohistochemical staining by antibodies specific of pepsinogen and accumulation of secretory granules in the cells producing pepsinogen. In these cells, the amount for pepsinogen mRNA was reduced as revealed by Northern blotting and in situ hybridization. Ultrastructurally the endoplasmic reticulum of these cells was poorly developed, the findings being consistent with a reduction in protein synthesis. It appears that omeprazole inhibits the secretion of pepsinogen, increasing the intracellular store and leading to the reduction in gene expression probably by a feedback mechanism and consequent reduction in pepsinogen synthesis. Since these changes were most evident in the acid-secreting fundic gland mucosa, as compared with other mucosae secreting only pepsinogen, namely pyloric and duodenal mucosa, it appears probable that these changes are linked with omeprazole-induced reduction in the acid secretion.     

Circular zymogens of human ribonuclease 1

The endogenous production of enzymes as zymogens provides a means to control catalytic activities. Here, we describe the heterologous production of ribonuclease 1 (RNase 1), which is the most prevalent secretory ribonuclease in humans, as a zymogen. In folded RNase 1, the N and C termini flank the enzymic active site. By using intein‐mediated cis‐splicing, we created circular proteins in which access to the active site of RNase 1 is obstructed by an amino‐acid sequence that is recognized by the HIV‐1 protease. Installing a sequence that does not perturb the RNase 1 fold led to only modest inactivation. In contrast, the ancillary truncation of residues from each terminus led to a substantial decrease in the catalytic activity of the zymogen with the maintenance of thermostability. For optimized zymogens, activation by HIV‐1 protease led to a > 10⁴‐fold increase in ribonucleolytic activity at a rate comparable to that for the cleavage of endogenous viral substrates. Molecular modeling indicated that these zymogens are inactivated by conformational distortion in addition to substrate occlusion. Because protease levels are elevated in many disease states and ribonucleolytic activity can be cytotoxic, RNase 1 zymogens have potential as generalizable prodrugs.