Legumain forms from plants and animals differ in their specificity.

We purified forms of legumain from a plant source (seeds of kidney bean, Phaseolus vulgaris) and a mammal (kidney of pig, Sus scropha) for comparison of their properties. Both forms were found to be stable only under moderately acidic pH conditions, and were maximally active at about pH 6; the plant enzyme was somewhat less stable and had a slightly higher pH optimum. With benzyloxycarbonyl-Xaa-Ala-Asn-aminomethylcoumarylamide substrates, the two forms of legumain showed distinctly different specificities for the P3 residue, the plant legumain preferring amino acids with bulky hydrophobic side chains because of lower Km values. Both forms of legumain were highly specific for hydrolysis of asparaginyl bonds in the arylamide substrates and in neurotensin. Aspartyl bonds were hydrolysed about 100-fold more slowly with lower pH optima. Potential substrates containing other amino acids structurally similar to asparagine were not hydrolysed. There were clear differences in specificity of hydrolysis of protein substrates. The plant legumain differed from pig legumain in its action on tetanus toxoid C-fragment, cleaving at Asn97 but not at Asn337, and produced more extensive digestion of phaseolin. The plant form of legumain was much more weakly inhibited by egg-white cystatin than was the mammalian form.

Inhibition of mammalian legumain by some cystatins is due to a novel second reactive site.

We have investigated the inhibition of the recently identified family C13 cysteine peptidase, pig legumain, by human cystatin C. The cystatin was seen to inhibit enzyme activity by stoichiometric 1:1 binding in competition with substrate. The Ki value for the interaction was 0.20 nM, i.e. cystatin C had an affinity for legumain similar to that for the papain-like family C1 cysteine peptidase, cathepsin B. However, cystatin C variants with alterations in the N-terminal region and the "second hairpin loop" that rendered the cystatin inactive against cathepsin B, still inhibited legumain with Ki values 0.2-0.3 nM. Complexes between cystatin C and papain inhibited legumain activity against benzoyl-Asn-NHPhNO2 as efficiently as did cystatin C alone. Conversely, cystatin C inhibited papain activity against benzoyl-Arg-NHPhNO2 whether or not the cystatin had been incubated with legumain, strongly indicating that the cystatin inhibited the two enzymes with non-overlapping sites. A ternary complex between legumain, cystatin C, and papain was demonstrated by gel filtration supported by immunoblotting. Screening of a panel of cystatin superfamily members showed that type 1 inhibitors (cystatins A and B) and low Mr kininogen (type 3) did not inhibit pig legumain. Of human type 2 cystatins, cystatin D was non-inhibitory, whereas cystatin E/M and cystatin F displayed strong (Ki 0.0016 nM) and relatively weak (Ki 10 nM) affinity for legumain, respectively. Sequence alignments and molecular modeling led to the suggestion that a loop located on the opposite side to the papain-binding surface, between the alpha-helix and the first strand of the main beta-pleated sheet of the cystatin structure, could be involved in legumain binding. This was corroborated by analysis of a cystatin C variant with substitution of the Asn39 residue in this loop (N39K-cystatin C); this variant showed a slight reduction in affinity for cathepsin B (Ki 1.5 nM) but >>5,000-fold lower affinity for legumain (Ki >>1,000 nM) than wild-type cystatin C.

Pig kidney legumain: an asparaginyl endopeptidase with restricted specificity.

Legumain was recently discovered as a lysosomal endopeptidase in mammals [Chen, Dando, Rawlings, Brown, Young, Stevens, Hewitt, Watts and Barrett (1997) J. Biol. Chem. 272, 8090-8098], having been known previously only from plants and invertebrates. It has been shown to play a key role in processing of the C fragment of tetanus toxin for presentation by the MHC class-II system [Manoury, Hewitt, Morrice, Dando, Barrett and Watts (1998) Nature (London) 396, 695-699]. We examine here the specificity of the enzyme from pig kidney by use of protein, oligopeptide and synthetic arylamide substrates, all determinations being made at pH 5.8. In proteins, only about one in ten of the asparaginyl bonds were hydrolysed, and these were mostly predicted to be located at turns on the protein surface. Bonds that were not cleaved in tetanus toxin were cleaved when presented in oligopeptides, sometimes faster than an equivalent oligopeptide based on a bond that was cleaved in the protein. Legumain cleaved the bait region of rat alpha1-macroglobulin and was 'trapped' by the macroglobulin, as most other endopeptidases are, but did not interact with human alpha2-macroglobulin, which contains no asparagine residue in its bait region. Glycosylation of asparagine totally prevented hydrolysis by legumain. Specificity for arylamide substrates was evaluated with reference to benzyloxycarbonyl-Ala-Ala-Asn-aminomethylcoumarin, and the preference for the P3-position amino acid was Ala>Tyr(tertiary butyl)>Val>Pro>Phe=Tyr>Leu=Gly. There was no hydrolysis of substrate analogues containing mono- or di-N-methylasparagines, l-2-amino-3-ureidopropionic acid or citrulline in the P1 position. We conclude that mammalian legumain appears to be totally restricted to the hydrolysis of asparaginyl bonds in substrates of all kinds. There seem to be no strong preferences for particular amino acids in other subsites, and yet there are still unidentified factors that prevent hydrolysis of many asparaginyl bonds in proteins.

Cloning and expression of mouse legumain, a lysosomal endopeptidase.

Legumain, a recently discovered mammalian cysteine endopeptidase, was found in all mouse tissues examined, but was particularly abundant in kidney and placenta. The distribution in subcellular fractions of mouse and rat kidney showed a lysosomal localization, and activity was detectable only after the organelles were disrupted. Nevertheless, ratios of legumain activity to that of cathepsin B differed considerably between mouse tissues. cDNA encoding mouse legumain was cloned and sequenced, the deduced amino acid sequence proving to be 83% identical to that of the human protein [Chen, Dando, Rawlings, Brown, Young, Stevens, Hewitt, Watts and Barrett (1997) J. Biol. Chem. 272, 8090-8098]. Recombinant mouse legumain was expressed in human embryonic kidney 293 cells by use of a vector containing a cytomegalovirus promoter. The recombinant enzyme was partially purified and found to be an asparagine-specific endopeptidase closely similar to naturally occurring pig kidney legumain.

An asparaginyl endopeptidase processes a microbial antigen for class II MHC presentation.

Foreign protein antigens must be broken down within endosomes or lysosomes to generate suitable peptides that will form complexes with class II major histocompatibility complex molecules for presentation to T cells. However, it is not known which proteases are required for antigen processing. To investigate this, we exposed a domain of the microbial tetanus toxin antigen (TTCF) to disrupted lysosomes that had been purified from a human B-cell line. Here we show that the dominant processing activity is not one of the known lysosomal cathepsins, which are generally believed to be the principal enzymes involved in antigen processing, but is instead an asparagine-specific cysteine endopeptidase. This enzyme seems similar or identical to a mammalian homologue of the legumain/haemoglobinase asparaginyl endopeptidases found originally in plants and parasites. We designed competitive peptide inhibitors of B-cell asparaginyl endopeptidase (AEP) that specifically block its proteolytic activity and inhibit processing of TTCF in vitro. In vivo, these inhibitors slow TTCF presentation to T cells, whereas preprocessing of TTCF with AEP accelerates its presentation, indicating that this enzyme performs a key step in TTCF processing. We also show that N-glycosylation of asparagine residues blocks AEP action in vitro. This indicates that N-glycosylation could eliminate sites of processing by AEP in mammalian proteins, allowing preferential processing of microbial antigens.

Cloning, isolation, and characterization of mammalian legumain, an asparaginyl endopeptidase.

Legumain is a cysteine endopeptidase that shows strict specificity for hydrolysis of asparaginyl bonds. The enzyme belongs to peptidase family C13, and is thus unrelated to the better known cysteine peptidases of the papain family, C1 (Rawlings, N. D., and Barrett, A. J. (1994) Methods Enzymol. 244, 461-486). To date, legumain has been described only from plants and a blood fluke, Schistosoma mansoni. We now show that legumain is present in mammals. We have cloned and sequenced human legumain and part of pig legumain. We have also purified legumain to homogeneity (2200-fold, 8% yield) from pig kidney. The mammalian sequences are clearly homologous with legumains from non-mammalian species. Pig legumain is a glycoprotein of about 34 kDa, decreasing to 31 kDa on deglycosylation. It is an asparaginyl endopeptidase, hydrolyzing Z-Ala-Ala-Asn-7-(4-methyl)coumarylamide and benzoyl-Asn-p-nitroanilide. Maximal activity is seen at pH 5.8 under normal assay conditions, and the enzyme is irreversibly denatured at pH 7 and above. Mammalian legumain is a cysteine endopeptidase, inhibited by iodoacetamide and maleimides, but unaffected by compound E64 (trans-epoxysuccinyl-L-leucylamido-(4-guanidino)butane). It is inhibited by ovocystatin (cystatin from chicken egg white) and human cystatin C with Ki values < 5 nM. We discuss the significance of the discovery of a cysteine endopeptidase of a new family and distinctive specificity in man and other mammals.

Rat thimet oligopeptidase: large-scale expression in Escherichia coli and characterization of the recombinant enzyme.

The coding sequence for rat testis thimet oligopeptidase (TOP) (EC 3.4.24.15) was placed under the control of the T7 polymerase/promoter system. Cultures of Escherichia coli transfected with the resulting plasmid expressed the enzyme as a soluble cytoplasmic protein. Medium-scale cultures allowed isolation of the enzyme in quantities of tens of milligrams. The availability of the recombinant enzyme permitted the determination of such chemical properties as epsilon 280 (48,960), zinc content (2 atom/molecule) and available thiol content (8-10/molecule) for TOP. The recombinant enzyme showed the catalytic activities previously reported for the naturally occurring enzyme, so that we can now conclude with confidence that these are all due to TOP and there is no need to postulate the existence of separate 'Pz-peptidase' or 'endo-oligopeptidase A' enzymes.

Thimet oligopeptidase specificity: evidence of preferential cleavage near the C-terminus and product inhibition from kinetic analysis of peptide hydrolysis.

The substrate-size specificity of human thimet oligopeptidase (EC 3.4.24.15) was investigated with oligomers of glycyl-prolyl-leucine (GPL)n where n = 2, 3, 4 and 5. These peptides were cleaved only at Leu-Gly bonds to give GPL as the single final product. Hydrolysis was most rapid with (GPL)3 and slowest with (GPL)5. The more water-soluble oligomers of Gly-Hyp-Leu showed the same trend. (Gly-Hyp-Leu)6 was not hydrolysed, consistent with the previous finding that substrates larger than 17 amino acids are not cleaved by thimet oligopeptidase. The cleavage of (GPL)3 to GPL fitted a sequential first-order model. First-order kinetics were unexpected as the initial substrate concentration was greater than Km. The anomaly was also seen during the cleavage of bradykinin and neurotensin, and in these cases first-order behaviour was due to potent competitive inhibition by the C-terminal product. The sequential mechanism for (GPL)3 breakdown by thimet oligopeptidase does not discriminate between initial cleavages towards the N- or C-terminus. As isoleucine is an unfavourable residue in P1, substrates were made in which selected leucine residues were replaced by isoleucine. GPL--GPI--GPL (where--represents the bond between the tripeptide units) was resistant to hydrolysis and GPI--GPL--GPL was cleaved only at the -Leu-Gly- bond. Experiments with isoleucine-containing analogues of (Gly-Hyp-Leu)4 showed that thimet oligopeptidase preferred to cleave these peptides near the C-terminus.

Immunoglobulin E antibodies to papaya proteinases and their relevance to chemonucleolysis.

STUDY DESIGN: Levels of four papaya cysteine proteinases were determined in Chymodiactin, a pharmaceutical preparation of chymopapain (EC 3.4.22.6) used in chemonucleolysis for the treatment of sciatica. Twelve sera known to contain immunoglobulin E antibodies to Chymodiactin were assayed for immunoglobulin E antibodies to these enzymes. OBJECTIVES: The goal of the study was to determine what contribution each of the four proteinases makes to the allergic response that occasionally occurs during injection of a damaged intervertebral disc with chymopapain preparations. SUMMARY OF BACKGROUND DATA: The occurrence of an allergic reaction during chemonucleolysis implies prior sensitization to components of the injected enzyme solution. The latex of the unripe fruit of the papaya plant Carica papaya, from which chymopapain is purified, contains another three immunologically distinct cysteine proteinases: 1) caricain (EC 3.4.22.30), 2) glycyl endopeptidase (EC 3.4.22.25), and 3) papain (EC 3.4.22.2). METHODS: A dot-blot immunoassay was developed to quantify each enzyme in Chymodiactin. Total serum immunoglobulin E levels and specific immunoglobulin E antibody levels to each of the four papaya cysteine proteinases were assayed by an enzyme-linked immunoassay in 12 sera containing immunoglobulin E antibodies to Chymodiactin. RESULTS: Chymodiactin contained 70% chymopapain, 20% caricain, 4% glycyl endopeptidase, and 0.1% papain. Immunoglobulin E antibodies to all four proteinases were found in most of the 12 sera, but in varying proportions. Antibodies to glycyl endopeptidase were predominant in eight sera, and the mean amounts of immunoglobulin E directed against each protein were: glycyl endopeptidase, 4.21 IU/ml; caricain, 2.9 IU/ml; chymopapain, 1.97 IU/ml; and papain, 1.39 IU/ml. Total serum immunoglobulin E levels showed little correlation with immunoglobulin E responses to Chymodiactin. CONCLUSIONS: The results suggested that removal of glycyl endopeptidase and caricain from pharmaceutical preparations of chymopapain may help reduce the incidence of allergic reactions during chemonucleolysis.

Characterization of a mitochondrial metallopeptidase reveals neurolysin as a homologue of thimet oligopeptidase.

We have isolated a metallopeptidase from rat liver. The peptidase is primarily located in the mitochondrial intermembrane space, where it interacts non-covalently with the inner membrane. The enzyme hydrolyzes oligopeptides, the largest substrate molecule found being dynorphin A1-17; it has no action on proteins, and does not interact with alpha 2-macroglobulin, and can therefore be classified as an oligopeptidase. We term the enzyme oligopeptidase M. Oligopeptidase M acts similarly to thimet oligopeptidase (EC 3.4.24.15) on bradykinin and several other peptides, but hydrolyzes neurotensin exclusively at the -Pro+Tyr- bond (the symbol + is used to indicate a scissile peptide bond) rather than the -Arg+Arg- bond. The enzyme is inhibited by chelating agents and some thiol-blocking compounds, but differs from thimet oligopeptidase in not being activated by thiol compounds. The peptidase is inhibited by Pro-Ile, unlike thimet oligopeptidase, and the two enzymes are separable in chromatography on hydroxyapatite. The N-terminal amino acid sequence of rat mitochondrial oligopeptidase M contains 19 out of 20 residues identical with a segment of rabbit microsomal endopeptidase and 17 matching the corresponding segment of pig-soluble angiotensin II-binding protein. Moreover, the rat protein is recognized by a monoclonal antibody against rabbit soluble angiotensin II-binding protein, all of which is consistent with these proteins being species variants of a single protein that is a homologue of thimet oligopeptidase. The biochemical properties of the mitochondrial oligopeptidase leave us in no doubt that it is neurolysin (EC 3.4.24.16), for which no sequence has previously been reported, and which has not been thought to be mitochondrial.

Thimet oligopeptidase: similarity to 'soluble angiotensin II-binding protein' and some corrections to the published amino acid sequence of the rat testis enzyme.

The deduced amino acid sequence of pig liver soluble angiotensin II-binding protein [Sugiura, Hagiwara and Hirose (1992) J. Biol. Chem. 267, 18067-18072] is similar over most of its length to that reported for rat testis thimet oligopeptidase (EC 3.4.24.15) by Pierotti, Dong, Glucksman, Orlowski and Roberts [(1990) (Biochemistry 29, 10323-10329]. We have found that homogeneous rat testis thimet oligopeptidase binds angiotensin II with the same distinctive characteristics as the pig liver protein. Analysis of the nucleotide sequences reported for the two proteins pointed to the likelihood that sequencing errors had caused two segments of the amino acid sequence of the rat protein to be translated out of frame, and re-sequencing of selected parts of the clone (kindly provided by the previous authors) confirmed this. The revised deduced amino acid sequence of rat thimet oligopeptidase contains 687 residues, representing a protein of 78,308 Da, and is more closely related to those of the pig liver protein and other known homologues of thimet oligopeptidase than that described previously.

Human thimet oligopeptidase.

We have purified human thimet oligopeptidase to homogeneity from erythrocytes, and compared it with the enzyme from rat testis and chicken liver. An antiserum raised against rat thimet oligopeptidase also recognized the human and chicken enzymes, suggesting that the structure of the enzyme has been strongly conserved in evolution. Consistent with this, the properties of the human enzyme were very similar to those for the other species. Thus human thimet oligopeptidase also is a thiol-dependent metallo-oligopeptidase with M(r) about 75,000. Specificity for cleavage of a number of peptides was indistinguishable from that of the rat enzyme, but Ki values for the four potent reversible inhibitors tested were lower. In discussing the results, we consider the determinants of the complex substrate specificity of thimet oligopeptidase. We question whether substrates containing more than 17 amino acid residues are cleaved, as has been suggested. We also point out that the favourable location of a proline residue and a free C-terminus in the substrate may be as important as the hydrophobic residues in the P2, P1 and P3' positions that have been emphasized in the past.

Quantification of peptide aldehyde ligands immobilized for the affinity chromatography of endopeptidases.

A method is described for quantification of aldehyde and aldehyde semicarbazone groups attached to an insoluble matrix. Semicarbazones are converted to aldehydes, and the aldehydes are coupled with 4-phenylazoaniline. The excess reagent is washed off, and the remainder then displaced with salicylaldehyde. The quantity of the phenylazoaniline/salicylaldehyde adduct is determined spectrophotometrically, allowing the calculation of the amount of aldehyde or semicarbazone per unit volume of the matrix. Analyses by the new method show that four matrices offered commercially for this type of immobilization differ greatly in coupling capacity and stability of the conjugate under conditions of affinity chromatography.

Low-dose, low-volume chemonucleolysis. A biochemical study.

Thirty patients with persistent sciatica due to a single disc protrusion were injected with 0.5 ml (2.0 mg) or 2.0 ml (8.0 mg) of chymopapain. Serial 24-hour urine samples were collected 1 day preinjection and for 5 days postinjection for glycosaminoglycan analysis. There was a significant increase in urinary glycosaminoglycan in both groups following chemonucleolysis, but no significant difference was found between the two groups in the total excretion of glycosaminoglycan during the 5-day period. Low-dose, low-volume chemonucleolysis may be as effective as standard doses in releasing glycosaminoglycan from intervertebral discs.

Human sputum cathepsin B degrades proteoglycan, is inhibited by alpha 2-macroglobulin and is modulated by neutrophil elastase cleavage of cathepsin B precursor and cystatin C.

The high-Mr alkali-stable form of cathepsin B was purified from purulent human sputum. It was shown to solubilize proteoglycan monomer entrapped in polyacrylamide at a rate comparable with that of human lysosomal cathepsin B. Like the enzyme from lysosomes, sputum cathepsin B was bound by human alpha 2-macroglobulin, which inhibited its action on proteoglycan. Cystatin C in purulent sputum was shown to be the N-terminally truncated form generated by neutrophil elastase cleavage, and sputum cathepsin B was only weakly inhibited by recombinant cystatin C that had been cleaved by neutrophil elastase in vitro. Addition of neutrophil elastase to mucoid sputum led to a 5-fold increase in cathepsin B activity concomitant with a lowering in Mr of the cysteine proteinase from 40,000 to 37,000, i.e. the size of the active enzyme purified from purulent sputum. It is concluded that the high-Mr form of cathepsin B present in purulent sputum is a functional proteinase, unlike similar forms of the enzyme secreted by mammary gland in organ culture. The activity of cathepsin B in sputum is modulated by neutrophil elastase, by a combination of inhibitor inactivation and zymogen activation.

The preparation of fully active chymopapain free of contaminating proteinases.

Chymopapain (EC 3.4.22.6) was purified from commercially available dried latex of papaya (Carica papaya) by extraction at acidic pH, cation-exchange chromatography and active site-directed affinity chromatography on immobilized alanyl-phenyl-alaninaldehyde semicarbazone, with elution by mercuric chloride. The product was found by immunoassay to be essentially free of the other cysteine proteinases from papaya, including papaya proteinase IV, and was fully active. The rate of alkylation of the active site cysteine of chymopapain by iodoacetate was found to be sufficiently rapid and selective for this reagent to be used as an active-site titrant.

Interactions of papaya proteinase IV with inhibitors.

Papaya proteinase IV (PPIV) is not inhibited by chicken cystatin, or human cystatins A or C, unlike most other proteinases of the papain superfamily. The enzyme inactivates chicken cystatin and human cystatin C by limited proteolysis of the glycyl bond previously shown to be involved in the inhibitory inactivity of the cystatins, but has no action on cystatin A. Contamination of commercial crystalline papain with PPIV accounts for the limited proteolysis of cystatins by 'papain' reported previously. PPIV is slowly bound by human alpha 2-macroglobulin. The enzyme is irreversibly inactivated by E-64, and by peptidyl diazomethanes containing glycine in P1 and a hydrophobic side-chain in P2. The reaction of PPIV with iodoacetate is extremely slow. PPIV is inhibited by peptide aldehydes despite the presence of bulky sidechains in P1, suggesting that these reversible inhibitors do not bind as substrate analogues.

Quantitative assessment of human proteinases as agents for chemonucleolysis.

A rabbit model system is described. It allows accurate measurement of the dose-dependent loss of glycosaminoglycan from the nucleus pulposus of lumbar intervertebral discs after injection of a proteinase. At the dose equivalent to that of chymopapain used in human chemonucleolysis, two human serine proteinases, cathepsin G and chymotrypsin, were as effective as chymopapain in removing up to 80% of the glycosaminoglycan from the disc. A cysteine proteinase, cathepsin B released no more than 45% of glycosaminoglycan. X-ray films clearly showed narrowing of the disc space when 30-40% of glycosaminoglycan was removed. The degradation of the nucleus pulposus was seen histologically as loss of toluidine blue metachromasia.