Solution structure of the catalytic domain of human collagenase-3 (MMP-13) complexed to a potent non-peptidic sulfonamide inhibitor: binding comparison with stromelysin-1 and collagenase-1

The full three-dimensional structure of the catalytic domain of human collagenase-3 (MMP-13) complexed to a potent, sulfonamide hydroxamic acid inhibitor (CGS 27023) has been determined by NMR spectroscopy. The results reveal a core domain for the protein consisting of three alpha-helices and five beta-sheet strands with an overall tertiary fold similar to the catalytic domains of other matrix metalloproteinase family members. The S1' pocket, which is the major site of hydrophobic binding interaction, was found to be a wide cleft spanning the length of the protein and presenting facile opportunity for inhibitor extension deep into the pocket. Comparison with the reported X-ray structure of collagenase-3 showed evidence of flexibility for the loop region flanking the S1' pocket in both NMR and X-ray data. This flexibility was corroborated by NMR dynamics studies. Inhibitor binding placed the methoxy phenyl ring in the S1' pocket with the remainder of the molecule primarily solvent-exposed. The binding mode for this inhibitor was found to be similar with respect to stromelysin-1 and collagenase-1; however, subtle comparative differences in the interactions between inhibitor and enzyme were observed for the three MMPs that were consistent with their respective binding potencies.

Identification of a splice variant of neutrophil collagenase (MMP-8)

We have identified a splice variant of human neutrophil collagenase (MMP-8) transcript (MMP-8alt) that has a 91 bp insertion between codons for amino acid residues 34 and 35 of MMP-8 cDNA. This splice variant encodes an open reading frame for a 444 residue protein, lacking a secretory signal sequence. Our data suggested that, as opposed to the original MMP-8, the translation product of MMP-8alt is not a secreted protein; nevertheless, it is enzymatically active. Further studies aimed at identifying the physiological substrates of MMP-8alt protein may lead to uncover novel roles it plays in cellular physiology.

Human HtrA, an evolutionarily conserved serine protease identified as a differentially expressed gene product in osteoarthritic cartilage

The human homologue of the Escherichia coli htrA gene product was identified by the differential display analysis of transcripts expressed in osteoarthritic cartilage. This transcript was identified previously as being repressed in SV40-transformed fibroblasts (Zumbrunn, J., and Trueb, B. (1996) FEBS Lett. 398, 187-192). Levels of HtrA mRNA were elevated approximately 7-fold in cartilage from individuals with osteoarthritis compared with nonarthritic controls. Differential expression of human HtrA protein was confirmed by an immunoblot analysis of cartilage extracts. Human HtrA protein expressed in heterologous systems was secreted and exhibited endoproteolytic activity, including autocatalytic cleavage. Conversion by mutagenesis of the putative active site serine 328 to alanine eliminated the enzymatic activity. Serine 328 was also found to be required for the formation of a stable complex with alpha1-antitrypsin. We have determined that the HtrA gene is highly conserved among mammalian species: the amino acid sequences encoded by HtrA cDNA clones from cow, rabbit, and guinea pig are 98% identical to human. In E. coli, a functional htrA gene product is required for cell survival after heat shock or oxidative stress; its role appears to be the degradation of denatured proteins. We propose that mammalian HtrA, with the addition of a new functionality during evolution, i.e. a mac25 homology domain, plays an important role in cell growth regulation.

Inhibition of interleukin-1alpha-induced cartilage oligomeric matrix protein degradation in bovine articular cartilage by matrix metalloproteinase inhibitors: potential role for matrix metalloproteinases in the generation of cartilage oligomeric matrix protein fragments in arthritic synovial fluid

OBJECTIVE

To determine whether matrix metalloproteinases (MMPs) degrade cartilage oligomeric matrix protein (COMP) to produce fragments similar to those found in synovial fluid (SF) from patients with arthritis.

METHODS

COMP fragments were generated in vitro by treating (a) bovine articular cartilage with interleukin-1alpha (IL-1alpha), (b) purified bovine COMP with MMPs, and (c) articular cartilage with MMPs. The fragments generated in each case were analyzed by Western blot, using an antibody to the C-terminal heptadecapeptide of COMP.

RESULTS

IL-1alpha stimulation of cartilage resulted in a fragmentation of COMP, which was inhibited by MMP inhibitors CGS 27023A and BB-94. Isolated, recombinant MMPs rapidly degraded purified COMP, as well as COMP residing in cartilage. Several COMP fragments produced in vitro had similar electrophoretic mobility to those in SF of patients with arthritis.

CONCLUSION

MMPs may contribute to the COMP fragments found in vivo. Quantitation of MMP-specific fragments may be useful in the evaluation of MMP inhibitors in patients with arthritis.

The expression, refolding, and purification of the catalytic domain of human collagenase-3 (MMP-13)

We have cloned, expressed, and purified a recombinant C-terminal truncated form (residues Glu103-Asn274) of human collagenase-3 (MMP-13) in Escherichia coli. The molecule contains the catalytic domain of the enzyme and is expressed almost exclusively as inclusion bodies. Using a combination of rapid dilution and diafiltration, the enzyme has been successfully refolded from these inclusion bodies. The protein was purified to homogeneity using cation-exchange and size-exclusion chromatography. The purified enzyme is a monomer with a Mr of approximately 19,600 and was characterized using a variety of techniques including, SDS-PAGE, RP-HPLC, LC-MS, amino acid analysis, and dynamic light scattering. Microheterogeneity at the NH2-terminus of the refolded, purified protein disappeared after incubating for 30-60 min at 37 degreesC. The enzyme was highly active using a fluorescent peptide substrate and was found to release S-GAG from bovine nasal cartilage chips.

Bioactive conformation of stromelysin inhibitors determined by transferred nuclear Overhauser effects

The transferred nuclear Overhauser effect has been used to determine the biologically active conformations of two stromelysin inhibitors. Both inhibitors used in this study were hydroxamic acids generated via chemical synthesis. These structures, representing the conformation of each inhibitor bound to stromelysin, superimposed with excellent agreement. The study also provided information on the shape and orientation of the S2' and S1' pockets of the enzyme relative to thermolysin. Comparisons were made between stromelysin and thermolysin inhibitors to critically examine thermolysin as a template for stromelysin-inhibitor design. The enzyme-bound conformations of these stromelysin inhibitors were determined for use as a template in conformationally restricted drug design.

Biochemical and molecular characterization of stromelysin synthesized by human osteoarthritic chondrocytes stimulated with recombinant human interleukin-1

OBJECTIVE

To study the biochemical and molecular characterization of stromelysin synthesized by human chondrocytes derived from osteofemoral heads.

METHODS

First passage human chondrocyte cultures were incubated with recombinant human interleukin-1 alpha or recombinant human interleukin-1 beta (10-1000 pg ml-1) for either 24 or 48 hrs. The medium compartment of these cultures was assayed for stromelysin activity. Total cellular RNA was used to determine: (i) the molecular structure of the stromelysin synthesized by these cells; and (ii) whether or not these chondrocytes expressed the Type II procollagen gene (COL2A1).

RESULTS

Human osteoarthritic chondrocytes released into the medium on enzyme requiring tryspin activation that possessed Substance P (SP) cleaving activity. SP cleaving activity was completely inhibited by EDTA. Casein zymography showed lysis zones produced by trypsin-activated chondrocyte culture medium that co-migrated with casein lysis zones produced by recombinant human prostromelysin. The majority of SP cleaving activity was eluted from a Zn-Sepharose column with 0.25 M glycine. Enzyme activity eluted from Zn-Sepharose produced casein lysis zones which co-migrated with lysis zones produced by recombinant human prostromelysin. Immunoblotting revealed the presence of prostromelysin (M(r), 55-57 kDa) in the pooled chondrocyte culture media applied to Zn-Sepharose and in the 0.25 M glycine eluate. Trypsin-activation converted prostromelysin to a mature stromelysin form (M(r), 45-47 kDa). Polymerase chain reaction (PCR) amplification of human chondrocyte cDNA demonstrated COL2A1 transcripts. A PCR product of expected size (680 bp) was produced by amplification of chondrocyte cDNA using stromelysin-1 oligonucleotide primers. The cloned and sequenced PCR product showed 100% homology between the chondrocyte stromelysin-1 mRNA-derived cDNA and the stromelysin-1 mRNA-derived cDNA of cultured human synovial, gingival and skin fibroblasts.

CONCLUSIONS

By several criteria, human osteoarthritic chondrocytes synthesized stromelysin which was biochemically and antigenically identical, and molecularly homologous with human fibroblast stromelysin-1. These results suggest that a quantitative imbalance between stromelysin-1 and endogenous stromelysin-1 inhibitors rather than the transcription of a new stromelysin gene is the mechanism underlying the increased proteoglycan degradation seen in osteoarthritic cartilage.

Stabilization of C5a receptor--G-protein interactions through ligand binding

Binding of biotin-C5a to the C5a receptor in membrane fragments followed by detergent solubilization and purification with streptavidin-agarose affinity chromatography resulted in the isolation of a receptor complex with associated G-proteins. In contrast, when receptor was detergent-solubilized in the absence of C5a and purified by affinity chromatography with Affigel-C5a, G-proteins did not copurify. Since the results indicate that receptor ligation stabilized the receptor--G-protein interaction to allow purification of the complex, the findings emphasize the dynamic nature of the C5a receptor-effector interactions. When biotin-C5a-ligated receptor was purified from a mouse cell line overexpressing recombinant human receptor, both Gialpha2 and Gialpha3 subunits copurified, confirming that multiple transducing systems are linked to the C5a receptor. The method of stabilization of receptor-transducer complexes offers the opportunity to further elaborate the interactions of the C5a receptor with diverse transducing elements and second messenger systems.

A stromelysin assay for the assessment of metalloprotease inhibitors on human aggregated proteoglycan

Human proteoglycan was aggregated to an immobilized hyaluronan solid phase on a 96-well ELISA plate. This complex was then degraded by recombinant human stromelysin. The remaining proteoglycan fragments were detected using a monoclonal antibody probe directed against the chondroitin sulfate (CS) region of the core protein. Stromelysin degraded the aggregate in a time and dose dependent manner as reflected by the loss of the CS epitope. Assay sensitivity was 0.125 U/well with total loss of the CS epitope occurring at 4 U/well. o-phenanthroline (IC50 = 52 microM) and U24522 (IC50 = 9 microM) inhibited degradation, while phosphoramidon did not. Serine and cysteine protease inhibitors had no effect. A comparative analysis of this assay with a reference method, substance P assay, gave similar inhibitor profiles. The use of aggregated human proteoglycan (native conformation) as a substrate, may better reflect how stromelysin inhibitors behave in the presence of complex substrates such as cartilage matrix.

Induction of C-reactive protein by cytokines in human hepatoma cell lines is potentiated by caffeine

Induction of C-reactive protein (CRP) by conditioned medium from lipopolysaccharide-stimulated human monocytes in two human hepatoma-cell lines, Hep 3B and NPLC/PRF/5, was potentiated 3-6-fold by the methylxanthine caffeine. The induction observed in the presence of conditioned medium plus caffeine was as much as 180-fold, comparable with that seen after many stimuli in vivo. This potentiation was accompanied by an increase in the levels of CRP mRNA. By contrast, no potentiating effect on CRP induction by conditioned medium was found when we tested theophylline, forskolin, 8-bromo cyclic AMP or two Ca2+ ionophores, namely ionomycin and A23187. None of the above compounds, including caffeine, when tested alone, had any detectable effect on the synthesis and secretion of CRP. Our previous study [Ganapathi, May, Schultz, Brabenec, Weinstein, Sehgal & Kushner (1988) Biochem. Biophys. Res. Commun. 157, 271-277], employing defined cytokines, had shown that induction of CRP in Hep 3B cells requires IL(interleukin)-6 plus IL-1, whereas, in the NPLC/PRF/5 cell line, IL-6 alone is effective. Caffeine similarly potentiated induction of CRP by these defined cytokine signals in these two cell lines. Changes in synthesis of other acute-phase proteins, including serum amyloid A (SAA), alpha 1-proteinase inhibitor, alpha 1-antichymotrypsin and albumin, induced by conditioned medium or, in some cases, by IL-6 and/or IL-1 alpha, were only minimally affected by caffeine. Thus these results indicate that the mechanism by which caffeine potentiates CRP induction by cytokines appears to be independent of increases in intracellular concentrations of the two second messengers, cyclic AMP and Ca2+; the precise nature of this mechanism is unclear at the present time. Our results also indicate that the intracellular mechanisms by which cytokines regulate synthesis of CRP may differ from those regulating synthesis of some other acute-phase proteins. The differential response of CRP and SAA to caffeine is of particular interest, since induction of both of these two major acute-phase proteins can be accomplished by identical extracellular signals.

Heterogeneous nature of the acute phase response. Differential regulation of human serum amyloid A, C-reactive protein, and other acute phase proteins by cytokines in Hep 3B cells

Because a number of different cytokines have been reported to regulate the synthesis of human, murine, and rat acute phase proteins (APP), we studied the effect of cytokines on production of several major human APP in a single system, the human hepatoma cell line Hep 3B. Conditioned medium (CM) prepared from human blood monocytes activated with LPS in the presence of dexamethasone led to substantial induction of serum amyloid A (SAA) and C-reactive protein (CRP) synthesis whereas the defined cytokines IL-1 beta, TNF alpha, and medium from a human keratinocyte cell line (COLO-16), containing hepatocyte-stimulating factor activity, failed to induce these two major APP. Induction of SAA and CRP was accompanied by an increase in concentration of their specific mRNA. Size fractionation of CM from activated monocytes by fast protein liquid chromatography indicated that SAA- and CRP-inducing activity eluted as a single peak with a Mr of approximately 18 kDa. alpha 1-Antitrypsin, which also failed to respond to IL-1 beta or TNF alpha, was induced by both CM and medium from COLO-16 cells. The induction of AT by CM was accompanied by an increase in specific mRNA. Induction of ceruloplasmin and alpha 1-antichymotrypsin and decrease in the synthesis of albumin was achieved by both CM and IL-1 beta. Ceruloplasmin and albumin responded in a comparable fashion to both TNF alpha and medium from COLO-16 cells; the response of ACT to these cytokines was not evaluated. These results indicate that human SAA and CRP are induced in Hep 3B cells by products of activated monocytes but not by IL-1 beta, TNF-alpha, or some hepatocyte-stimulating factor preparations and that a group of heterogeneous mechanisms are involved in the induction of the various human APP.

Heterogeneous nature of the acute phase response. Differential regulation of human serum amyloid A, C-reactive protein, and other acute phase proteins by cytokines in Hep 3B cells

Because a number of different cytokines have been reported to regulate the synthesis of human, murine, and rat acute phase proteins (APP), we studied the effect of cytokines on production of several major human APP in a single system, the human hepatoma cell line Hep 3B. Conditioned medium (CM) prepared from human blood monocytes activated with LPS in the presence of dexamethasone led to substantial induction of serum amyloid A (SAA) and C-reactive protein (CRP) synthesis whereas the defined cytokines IL-1 beta, TNF alpha, and medium from a human keratinocyte cell line (COLO-16), containing hepatocyte-stimulating factor activity, failed to induce these two major APP. Induction of SAA and CRP was accompanied by an increase in concentration of their specific mRNA. Size fractionation of CM from activated monocytes by fast protein liquid chromatography indicated that SAA- and CRP-inducing activity eluted as a single peak with a Mr of approximately 18 kDa. alpha 1-Antitrypsin, which also failed to respond to IL-1 beta or TNF alpha, was induced by both CM and medium from COLO-16 cells. The induction of AT by CM was accompanied by an increase in specific mRNA. Induction of ceruloplasmin and alpha 1-antichymotrypsin and decrease in the synthesis of albumin was achieved by both CM and IL-1 beta. Ceruloplasmin and albumin responded in a comparable fashion to both TNF alpha and medium from COLO-16 cells; the response of ACT to these cytokines was not evaluated. These results indicate that human SAA and CRP are induced in Hep 3B cells by products of activated monocytes but not by IL-1 beta, TNF-alpha, or some hepatocyte-stimulating factor preparations and that a group of heterogeneous mechanisms are involved in the induction of the various human APP.

Secretion of rabbit C-reactive protein by transfected human cell lines is more rapid than by cultured rabbit hepatocytes

C-reactive protein (CRP) is a major acute phase protein in humans and rabbits. Its synthesis by the liver varies over a 1000-fold range depending on the presence and severity of inflammatory stimuli. In previous studies of synthesis and secretion of rabbit CRP, we showed that secretion becomes more efficient over the course of the acute phase response as CRP synthesis rates increase (Macintyre, S.S., Kushner, I., and Samols, D. (1985) J. Biol. Chem. 260, 4169-4173). The current studies were undertaken to help distinguish between two alternative explanations for this finding: 1) that secretion efficiency may simply be a property of the rate of synthesis and intracellular concentration of CRP or 2) that secretion may be regulated by separate intracellular mechanisms. A fusion gene containing the mouse metallothionein I promoter linked to the protein coding region of the rabbit CRP gene was introduced into the human hepatoma cell line, NPLC, and the nonliver cell line, HeLa. In this system a graded response of the mouse metallothionein I promoter following exposure to increasing zinc concentrations results in increasing CRP synthesis. Unlike hepatocytes from rabbits undergoing the acute phase response, we found that rabbit CRP was secreted by these transfected cell lines with a very high degree of efficiency which was independent of the rate of CRP synthesis. This finding implies that normal rabbit hepatocytes retard the secretion of CRP and that this inhibition is diminished as the acute phase response progresses. It further indicates that the relationship between changes in synthetic rate and efficiency of secretion of rabbit CRP is not a causal one and that synthesis and secretion of CRP by rabbit hepatocytes are regulated by independent intracellular mechanisms during the acute phase response.

Expression of envelope glycoproteins of human immunodeficiency virus by an insect virus vector

The envelope gene of human immunodeficiency virus was inserted into the genome of an insect virus vector (Autographa californica nuclear polyhedrosis virus). Upon infection of tissue culture cells, this recombinant virus produced immunoreactive polypeptides related to the envelope glycoproteins of human immunodeficiency virus. Serological survey indicates such polypeptides would be of value as antigens in diagnostics for acquired immunodeficiency syndrome.

Cloning and characterization of the gene for rabbit C-reactive protein

C-reactive protein (CRP), an acute-phase plasma protein of hepatic origin in man and rabbit, is a cyclic pentamer composed of five identical nonglycosylated Mr 22 500 subunits. We have isolated both cDNA and genomic clones for rabbit CRP. These clones were used as probes to demonstrate that when CRP synthesis is increased following an acute inflammatory stimulus, there is a corresponding increase in the level of accumulated CRP mRNA. The rabbit CRP gene is 2.6 kilobases in length containing a single intron of 252 base pairs (bp) which interrupts the codon for amino acid 2 in the protein. The mRNA for CRP contains a 5'-nontranslated region of 113 bp and a 3'-nontranslated region of 1550 bp. Sequencing of the protein-coding region of the gene indicates that the primary translation product contains a 20 amino acid N-terminal signal peptide. The deduced amino acid sequence is in general agreement with the published sequence [Wang, C. M., Nguyen, N. Y., Yonaha, K., Robey, F., & Liu, T.-Y. (1982) J. Biol. Chem. 257, 13610-13615] except in the region between amino acids 63 and 73. In this region, the sequence of both cDNA and genomic clones indicates the presence of 28 amino acids not previously reported. This alteration may be the result of genetic heterogeneity or an error in the reported protein sequence.

Acetate synthesis from carbon monoxide by Clostridium thermoaceticum. Purification of the corrinoid protein

A corrinoid protein has been purified from Clostridium thermoaceticum which is required for the synthesis of acetyl-CoA from carbon monoxide and methyltetrahydrofolate. The purified protein is an alpha beta dimer with subunit molecular weights of 34,000 and 55,000, respectively, and contains 0.69 mol of corrinoid/mol of dimer. The corrinoid protein is methylated in the presence of methyltransferase and methyltetrahydrofolate; methylation is on the cobalt of the corrinoid moiety of the protein. When 14C-methylated protein is incubated with Fraction F3, ATP, CoASH, and CO, [14C]acetyl-CoA is formed. Methylation of cobalamin (B12) is catalyzed by the methyltransferase but methylcobalamin does not substitute for the methylated corrinoid protein as the source of methyl in the formation of acetyl-CoA.

Synthesis of acetyl coenzyme A from carbon monoxide, methyltetrahydrofolate, and coenzyme A by enzymes from Clostridium thermoaceticum

Two purified fractions from Clostridium thermoaceticum are shown to catalyze the following reaction: CO + CH3THF + CoA ATP leads to CH3COCoA + THF. The methyltetrahydrofolate (CH3THF) gives rise to the methyl group of the acetyl-coenzyme A (CoA) and the carbon monoxide (CO) and CoA to its carboxyl thio ester group. The role of ATP is unknown. One of the protein fractions (F2) is a methyltransferase, whereas the other fraction (F3) contains CO dehydrogenase and a methyl acceptor which is postulated to be a corrinoid enzyme. The methyltransferase catalyzes the transfer of the methyl group to the methyl acceptor, and the CO is converted to a formyl derivative by the CO dehydrogenase. By a mechanism that is as yet unknown, the formyl derivative in combination with CoA and the methyl of the methyl acceptor are converted to acetyl-CoA. It is also shown that fraction F3 catalyzes the reversible exchange of 14C from [1-14C]acetyl-CoA into 14CO and that ATP is required, but not the methyltransferase. It is proposed that these reactions are part of the mechanism which enables certain autotrophic bacteria to grow on CO. It is postulated that CH3THF is synthesized from CO and tetrahydrofolate which then, as described above, is converted to acetyl-CoA. The acetyl-CoA then serves as a precursor in other anabolic reactions. A similar autotropic pathway may occur in bacteria which grow on carbon dioxide and hydrogen.

Purification of five components from Clostridium thermoaceticum which catalyze synthesis of acetate from pyruvate and methyltetrahydrofolate. Properties of phosphotransacetylase

A five-component enzyme system which catalyzes synthesis of acetylphosphate from methyltetrahydrofolate (CH3THF) plus pyruvate has been purified from the homoacetate-fermenting bacterium, Clostridium thermoaceticum. One of the components was identified as the low potential electron carrier, ferredoxin, and the other 4 protein components have been designated F1, F2, F3, and F4. F1, F2, and F4 have been purified to homogeneity and, as estimated by gel filtration, have native molecular weights of 88,100, 58,900, and 255,000, respectively, while the subunit molecular weights obtained by sodium dodecyl sulfate-polyacrylamide gel electrophoresis are 20,000, 25,500, and 120,000, respectively. F3 contains 3 to 4 protein bands and has not been characterized with respect to molecular weights. Acetylphosphate synthesis by the purified system is optimal at pH 6.0 and 65 degrees C and requires ATP, CoA, and, to a lesser extent, thiamin pyrophosphate and Fe2+. S-Adenosylmethionine is not required. The F1 component has been identified as phosphotransacetylase and in its absence, the product is acetyl-CoA. Some properties of the phosphotransacetylase are presented. A scheme is given indicating present views of the functions of the individual components.

Purification of carbon monoxide dehydrogenase, a nickel enzyme from Clostridium thermocaceticum

Carbon monoxide dehydrogenase (CO dehydrogenase) has been purified from the homoacetate-fermenting bacterium, Clostridium thermoaceticum. By use of 63Ni, it has been determined that the dehydrogenase is a metallo nickel enzyme. Nickel was rapidly taken up by the organism and most of the ingested metal was found to be incorporated into CO dehydrogenase. As estimated by gel filtration, the native enzyme has a molecular weight of 410,000. Ferredoxin and a membrane-bound b-type cytochrome, both obtained from C. thermoaceticum, are rapidly reduced by the enzyme in the presence of carbon monoxide and both are considered to be native electron carriers. FMN and Desulfovibrio vulgaris cytochrome c3 were also reduced by the enzyme, while spinach ferredoxin, FAD, NAD, and NADP were not. CO dehydrogenase activity was not appreciably affected by propyl iodide, methyl iodide, carbon tetrachloride, or metal chelators, but was reversibly inhibited by KCN. A method for the in situ assay of CO dehydrogenase in polyacrylamide gels is presented.