Proteolytic and non-proteolytic activation of human neutrophil progelatinase B.

The activation of human neutrophil progelatinase B (pro-HNG) by a variety of proteolytic and non-proteolytic activators has been investigated. A quantitative comparison of the activation efficiencies of treatments previously reported to activate pro-HNG or the related gelatinase B species produced by other cells demonstrates that stromelysin and trypsin are good activators. HgCl2 is a moderately effective activator, while p-chloromercuribenzoate and NaOCl are poor activators. It is also shown that human matrilysin and human fibroblast-type collagenase can activate pro-HNG by a mechanism that is very similar to that of stromelysin. Initially, these proteinases hydrolyze the Glu40-Met41 bond in the propeptide domain to generate an 88 kDa inactive HNG species. Collagenase also generates a 68 kDa HNG species through hydrolysis of the Ala74-Met75 bond. Ultimately, treatment with either matrilysin, collagenase or trypsin results in the production of a 65 kDa active form of HNG that arises from hydrolysis of the Arg87-Phe88 bond. This is the same active species produced on activation by stromelysin. This cleavage site is downstream of the 'cysteine-switch' residue located at position 80 and releases it, accounting for the permanent activation of the enzyme. These results suggest that matrilysin and collagenase may be physiologically relevant activators of pro-HNG and/or other progelatinase B species. Activation by HgCl2 produces an active 68 kDa enzyme due to autolytic hydrolysis of the Ala74-Met75 bond. This species retains the cysteine switch residue; however, it is shown that it is only active in the continued presence of HgCl2. Removal of the HgCl2 restores latency, indicating that this species is reversibly activated by HgCl2, which functions by complexing the sulfhydryl group of the cysteine switch residue and keeping it dissociated from the active site zinc atom. Thus, in spite of reports to the contrary, the cysteine switch mechanism can account for the latency and activation of pro-HNG.

Effect of metal ions on the fluorescence of leucine aminopeptidase and its dansyl-peptide substrates.

The effect of Cu(II), Ni(II), Zn(II), Mg(II), and Mn(II) on the fluorescence of porcine kidney cytosol leucine aminopeptidase and three of its dansyl(Dns) peptide substrates, Leu-Gly-NHNH-Dns, Leu-Gly-NH(CH2)2NH-Dns, and Leu-Gly-NH(CH2)6NH-Dns, has been investigated. These five metal ions were chosen for study because each binds to the regulatory metal binding site of leucine aminopeptidase. Since the binding is relatively weak, kinetic studies of the different metalloderivatives of the enzyme are normally carried out in the presence of large molar excesses of these metal ions that can potentially affect both the enzyme and substrate. The fluorescence of all of the dansyl-peptides, as well as several other dansyl species, is quenched by Ni(II) and Cu(II), but not by Mg(II), Mn(II), or Zn(II). The absorption spectra of these dansyl substrates are also perturbed by Ni(II) and Cu(II). The rate at which maximal quenching for some dansyl species is attained after mixing with Ni(II) and Cu(II) is slow and the quenching is reversed on addition of EDTA. These results indicate that the quenching is the result of complex formation between the fluorophores and these metal ions. The association constants for the metal complexes have been determined from Stern-Volmer plots. In addition to complex formation, Ni(II) and Cu(II) cause the degradation of Leu-Gly-NHNH-Dns through a two step mechanism involving loss of dansic acid. Ni(II) and Cu(II) also partially quench the fluorescence of leucine aminopeptidase through contact with its surface accessible Trp residues. These observations indicate that care must be taken in stopped flow fluorescence studies of reactions between this enzyme and its dansyl substrates to avoid adverse effects brought about by Ni(II) and Cu(II).

Resonance Raman spectra of bovine liver catalase: enhancement of proximal tyrosinate vibrations.

Resonance Raman spectra of native bovine liver ferri-catalase have been obtained in the 200-1800 cm-1 region. Excitation at a series of wavelengths ranging from 406.7 to 514.5 nm has been used and gives rise to distinct sets of resonance Raman bands. Excitation within the Soret and Q-bands of the heme group produces the expected set of polarized and nonpolarized porphyrin modes, respectively. The frequencies of the porphyrin skeletal stretching bands in the 1450-1700 cm-1 region indicate that catalase contains only five-coordinate, high-spin heme groups. In addition to the porphyrin modes, bovine liver catalase exhibits bands near 1612 and 1520 cm-1 that are attributable to ring vibrations of the proximal tyrosinate that are enhanced via resonance with a proximal tyrosinate----Fe(III) change transfer transition centered near 490 nm. Similar bands have been observed in mutant hemoglobins that have tyrosinate axial ligands and in other Fe(III)-tyrosinate proteins. No resonance Raman bands have been observed that can be attributed to degraded hemes. The spectra are relatively insensitive to pH over the range of 5-10, and the same spectra are observed for catalase samples that do and do not contain tightly bound NADPH. Resonance Raman spectra of the fluoride complex exhibit porphyrin skeletal stretching modes that show it to be six coordinate, high spin, while the cyanide complex is six coordinate, low spin. Both the azide and thiocyanate complexes, however, are spin-state mixtures with the high-spin form predominant.

Human neutrophil collagenase.

Human neutrophil collagenase (HNC) has been purified from extracts of fresh and outdated buffy coats and from the exudates of phorbol myristate acetate-stimulated neutrophils. The HNC present in the starting material from such preparations can be either latent or active, or have an approximate molecular weight of 75 or 58 kDa, depending upon whether the extraction buffer contains protease inhibitors and/or antioxidants. The purification of these different forms of HNC is described and is made possible by taking appropriate precautions to stabilize the HNC. For example, a purification protocol is described that allows the purification to homogeneity of the active and PCMB-active latent 58 kDa forms of HNC in high yield with specific collagenase activities that greatly exceed that of trypsin-activated human fibroblast collagenase (HFC). The pattern of activation of the latent 58 and 75 kDa species by trypsin, organomercurials and oxidants has been investigated. HNC is shown to preferentially hydrolyze type I over types II and III collagens in solution. The specificity of HNC toward the hydrolysis of 60 octapeptides has been examined and compared with HFC. HNC is shown to be a glycoprotein that contains complex N-linked oligosaccharides.

Resonance Raman evidence for the activation of dioxygen in horseradish oxyperoxidase.

Resonance Raman spectroscopy has been employed to investigate the molecular bases for the markedly different properties of horseradish oxyperoxidase and oxymyoglobin. The porphyrin core of oxyperoxidase is slightly more expanded with the iron atom closer to the porphyrin plane, and there is greater iron d pi-to-oxygen pi backbonding compared to oxymyoglobin. The iron-oxygen (stretching or bending) bands are observed at 570 and 562 cm-1, respectively, for oxymyoglobin and oxyperoxidase, and the iron-His stretching bands have been tentatively identified at 276 and 289 cm-1, respectively. It is suggested that the stronger iron-His bond in oxyperoxidase facilitates greater iron d pi-to-oxygen pi backdonation by raising the energy of the iron d pi orbitals closer to the energy of the oxygen pi orbitals. This weakens the O-O bond and activates dioxygen for use as an electron acceptor in the peroxidase-oxidase reaction.

Purification to homogeneity of latent and active 58-kilodalton forms of human neutrophil collagenase.

Latent and active 58-kDa forms of human neutrophil collagenase (HNC) have been purified to homogeneity. Buffy coats were extracted in the presence and absence of phenylmethanesulfonyl fluoride to generate crude starting preparations that contained latent and active HNC, respectively. The buffers used in preparing these extracts and for all subsequent chromatographic steps contained NaCl at a concentration of 0.5 M or greater, 0.05% Brij-35, concentrations of CaCl2 of 5 mM or greater, and (when feasible) 50 microM ZnSO4 to stabilize the HNC. The collagenase activity in the buffy coat extracts was adsorbed to a Reactive Red 120-agarose column at pH 7.5 in 0.5 M NaCl and was eluted when the NaCl concentration was increased to 1 M. The active and p-(chloromercuri)benzoate-activated latent enzymes were next adsorbed to a Sepharose-CH-Pro-Leu-Gly-NHOH affinity resin in 1 M NaCl at pH 7.5 and desorbed at pH 9 to give a fraction containing only HNC and a small amount of neutrophil gelatinase. The latter enzyme was removed by passage over a gelatin-Sepharose column in 1 M NaCl at pH 7.5. The purified samples of active and latent HNC were obtained with typical cumulative yields of 32 and 82% and specific activities toward soluble rat type I collagen at 30 degrees C of 7200 and 12,000 micrograms min-1 mg-1, respectively. These specific activities are markedly higher than previously reported for HNC. Both active and latent HNC exhibit a single band on sodium dodecyl sulfate-polyacrylamide gel electrophoresis both in the presence and in the absence of 2-mercaptoethanol. The mobility of latent HNC is consistent with a molecular weight of approximately 58K, with the active form exhibiting a slightly lower (less than 1-2K) molecular weight.

The cysteine switch: a principle of regulation of metalloproteinase activity with potential applicability to the entire matrix metalloproteinase gene family.

The general applicability of the "cysteine-switch" activation mechanism to the members of the matrix metalloproteinase (MMP) gene family is examined here. All currently known members of the MMP gene family share the characteristic that they are synthesized in a latent, inactive, form. Recent evidence suggests that this latency in human fibroblast collagenase (HFC) is the result of formation of an intramolecular complex between the single cysteine residue in its propeptide domain and the essential zinc atom in the catalytic domain, a complex that blocks the active site. Latent HFC can be activated by multiple means, all of which effect the dissociation of the cysteine residue from the complex. This is referred to as the "cysteine-switch" mechanism of activation. The propeptide domain that contains the critical cysteine residue and the catalytic domain that contains the zinc-binding site are the only two domains common to all of the MMPs. The amino acid sequences surrounding both the critical cysteine residue and a region of the protein chains containing two of the putative histidine zinc-binding ligands are highly conserved in all of the MMPs. A survey of the literature shows that many of the individual MMPs can be activated by the multiple means observed for latent HFC. These observations support the view that the cysteine-switch mechanism is applicable to all members of this gene family. This mechanism is unprecedented in enzymology as far as we know and offers the opportunity for multiple modes of physiological activation of these important enzymes. Since conditions in different cells and tissues may match those necessary to effect one of these activation modes for a given MMP, this may offer metabolic flexibility in the control of MMP activation.

Hydrolysis of dansyl-peptide substrates by leucine aminopeptidase: origin of dansyl fluorescence changes during hydrolysis.

The origin of the fluorescence changes observed in stopped-flow experiments of the hydrolysis of three 5-(dimethylamino)naphthalene-1-sulfonyl-(dansyl) peptide substrates by porcine kidney cytosol leucine aminopeptidase has been investigated. The substrates used all have the potential to accept energy from aromatic residues of the enzyme via resonance energy transfer when they are bound as enzyme-substrate complexes, indicating that fluorescence changes due to the buildup and decay of such intermediates are possible. However, the fluorescence of these substrates differs from that of the products, and direct excitation of their dansyl groups during hydrolysis can also be responsible for the observed fluorescence changes due to changes in the concentrations of free substrate and product. The dansyl fluorescence changes observed with excitation wavelengths near 280 nm are not accompanied by quenching of the enzyme fluorescence, as would be expected if there were enzyme-to-substrate energy transfer. The magnitude of the maximal fluorescence change at a fixed concentration of substrate is also independent of the enzyme concentration. Furthermore, the excitation profile for the fluorescence changes shows that they arise from direct excitation of the dansyl group. Thus, there is no energy transfer in these reactions, and the fluorescence changes observed arise from direct excitation of the dansyl group and reflect the instantaneous concentration of substrate. This behavior contrasts sharply with that for the reaction of carboxypeptidase A with dansyl-Gly-Tyr, which has been studied as a positive control for an energy-transfer system.(ABSTRACT TRUNCATED AT 250 WORDS)

Preparation and reconstitution with divalent metal ions of class I and class II Clostridium histolyticum apocollagenases.

Both gamma- and zeta-collagenases from Clostridium histolyticum are fully and reversibly inhibited by 1,10-phenanthroline at pH 7.5 in the presence of 10 mM CaCl2 with KI values of 0.11 and 0.040 mM, respectively. The inhibition is caused by removal of the single, active-site Zn(II) present in each of these enzymes. The nonchelating analogue 1,5-phenanthroline has no effect on the activity of either enzyme. Dialysis of the enzymes in the presence of 1,10-phenanthroline, followed by back dialysis against buffer containing no chelating agent, gives the respective apocollagenases. Both apoenzymes can be instantaneously and fully reactivated by the addition of 1 equiv of Zn(II). Variable amounts of activity are restored to both apocollagenases by Co(II) and Ni(II) and to gamma-apocollagenase by Cu(II). The activity titration curve for gamma-apocollagenase with Co(II) and Scatchard plots for the reconstitution of gamma-apocollagenase with Cu(II) and Ni(II) and of zeta-apocollagenase with Ni(II) and Co(II) indicate that all activity changes are the result of binding of a single equivalent of these divalent metal ions at the active site of the collagenases. Cd(II) and Hg(II) do not restore measurable activity to either apoenzyme.

Preparation by direct metal exchange and kinetic study of active site metal substituted class I and class II Clostridium histolyticum collagenases.

Active site metal substitutions for both gamma- and zeta-collagenases from Clostridium histolyticum have been made by direct metal exchange. The incubation of Co(II), Cu(II), Ni(II), Cd(II), and Hg(II) with these native collagenases results in changes in activity that parallel those observed for the reconstitution of the respective apoenzymes with these metal ions. For both collagenases, the exchange reactions with Co(II) and Cu(II) are complete within 1 min. However, the changes in activity observed on addition of Ni(II), Cd(II), and Hg(II) to gamma-collagenase and Cd(II) and Hg(II) to zeta-collagenase are time dependent. The kinetic parameters Kcat and KM have been determined for each of the active metallospecies. The substitution of the active-site metal ion in gamma-collagenase results in changes in both kcat and KM, while the effect observed in zeta-collagenase is primarily on KM. This suggests that there are differences in the mechanisms of these two collagenases, at least with respect to the role of the zinc ion in catalysis.

Functional constituents of the active site of human neutrophil collagenase.

A series of chemical modification reactions has been carried out to identify functional constituents of the active site of human neutrophil collagenase. The enzyme is reversibly inhibited by the transition metal chelating agent 1,10-phenanthroline, and inhibition is fully reversed by zinc. Removal of weakly bound metal ions by gel filtration inactivates collagenase, and activity is fully restored on immediate readdition of calcium. The enzyme is unaffected by reagents that modify serine, cysteine, and arginine residues. However, reaction with the carboxyl reagents cyclohexylmorpholinocarbodiimide and Woodward's Reagent K lowers the activity of the enzyme substantially. Acetylimidazole inactivates the enzyme, but activity is completely restored on addition of hydroxylamine. The enzyme is also inactivated by tetranitromethane, indicating that it contains an essential tyrosine residue. Acylation of collagenase with diethyl pyrocarbonate, diketene, acetic anhydride, or trinitrobenzenesulfonate inactivates the enzyme, and activity is not restored on addition of hydroxylamine, indicating the presence of an essential lysine residue.

Agreement with the disulfide stretching frequency-conformation correlation of Sugeta, Go, and Miyazawa.

Two hypotheses have been advanced to explain the conformational dependence of the disulfide stretching frequency of the CCSSCC moiety in unstrained disulfides. Such compounds can adopt conformations that exhibit disulfide stretching bands near 510, 525, and 540 cm-1. Sugeta et al. [Sugeta, H., Go, A. & Miyazawa, T. (1973) Bull. Chem. Soc. Jpn. 46, 3407-3411] have attributed these three bands to conformations having none, one, and two trans conformations, respectively, about the C-S bonds. In apparent contradiction to this correlation, Van Wart and Scheraga [Van Wart, H. E. & Scheraga, H. A. (1976) J. Phys. Chem. 80, 1812-1823] found that dithioglycolic acid crystals, believed at the time to exist only in the trans conformation about both C-S bonds, exhibited the S-S stretching band near 510 cm-1. On the basis of this observation, it was suggested instead that the 510, 525, and 540 cm-1 bands arose from CCSSCC moieties having none, one, and two A conformations (i.e., those with CC-SS dihedral angles close to 30 degrees), respectively, about the C-S bonds. It has recently been shown by Nash et al. [Nash, C. P., Olmstead, M. M., Weiss-Lopez, B., Musker, W. K., Ramasubbu, M. & Parthasarathy, R. (1985) J. Am. Chem. Soc. 107, 7194-7195] that dithioglycolic acid can crystallize in two different forms, one with trans and the other with gauche conformations about both C-S bonds, and that these have S-S stretching bands at 536 and 510 cm-1, respectively. These results are confirmed here and it is shown that our earlier data were collected on the all-gauche (rather than the all-trans) form. Thus, the correlation proposed by Sugeta et al. is correct.

Inhibition of human neutrophil collagenase by gold(I) salts used in chrysotherapy.

Six gold(I) salts, some of which are used as drugs in chrysotherapy, are shown to be inhibitors of two forms of human neutrophil collagenase. The IC50 values vary over six orders of magnitude, the lowest being 3.5 nM for Myocrisin. Thus, inhibition is greatly affected by the identity of the ligands to the gold(I) atom. The inhibition of collagenase by these gold(I) salts may be a partial basis for their antiarthritic action.

Identification of Clostridium histolyticum collagenase hyperreactive sites in type I, II, and III collagens: lack of correlation with local triple helical stability.

The class I and II Clostridium histolyticum collagenases (CHC) have been used to identify hyperreactive sites in rat type I, bovine type II, and human type III collagens. The class I CHC attack both collagens at loci concentrated in the N-terminal half of these collagens starting with the site closest to the N-terminus. The class II CHC initiate collagenolysis by attacking both collagens in the interior to produce a mixture of C-terminal 62,000 and a N-terminal 36,000 fragments. Both fragments are next shortened by removal of a 3000 fragment. These results are very similar to those reported earlier for the hydrolysis of rat type I collagen by these CHC, indicating that the three collagens share many hyperreactive sites. Similar reactions carried out with the respective gelatins show that they are cleaved at many sites at approximately the same rate. Thus, the hyperreactivity of the sites identified must be attributed to their environment in the native collagens. N-terminal sequencing of the fragments produced in these reactions has allowed the identification of 16 cleavage sites in the alpha 1(I), alpha 2(I), alpha 1(II), and alpha 1(III) collagen chains. An analysis of the triple helical stabilities of these cleavage site regions as reflected by their imino acid contents fails to yield a correlation between reactivity and triple helical stability. The existence of these hyperreactive CHC cleavage sites suggests that type I, II, and III collagens contain regions that have specific nontriple helical conformations. The sequence of these sites presented here now makes it possible to investigate these conformations by computational and peptide mimetic techniques.