Advertisement

Rapid Determination of Substrate Specificity of Clostridium histolyticum β-Collagenase Using an Immobilized Peptide Library*

Open AccessPublished:November 27, 2001DOI:https://doi.org/10.1074/jbc.M111042200
      The molecular basis of the substrate specificity of Clostridium histolyticum β-collagenase was investigated using a combinatorial method. An immobilized positional peptide library, which contains 24,000 sequences, was constructed with a 7-hydroxycoumarin-4-propanoyl (Cop) fluorescent group attached at the N terminus of each sequence. This immobilized peptide library was incubated with C. histolyticum β-collagenase, releasing fluorogenic fragments in the solution phase. The relative substrate specificity (k cat/K m) for each member of the library was determined by measuring fluorescence intensity in the solution phase. Edman sequencing was used to assign structure to subsites of active substrate mixtures. Collectively, the substrate preference for subsites (P3–P4′) of C. histolyticumβ-collagenase was determined. The last position on the C-terminal side in which the identity of the amino acids affects the activity of the enzyme is P4′, and an aromatic side chain is preferred in this position. The optimal P1′–P3′ extended substrate sequence is P1′-Gly/Ala, P2′-Pro/Xaa, and P3′-Lys/Arg/Pro/Thr/Ser. The Cop group in either the P2 or P3 position is required for a high substrate activity with C. histolyticum β-collagenase. S2 and S3 sites of the protease play a dominant role in fixing the substrate specificity. The immobilized peptide library proved to be a powerful approach for assessing the substrate specificity of C. histolyticum β-collagenase, so it may be applied to the study of other proteases of interest.
      MMP
      matrix metalloprotease
      CPG
      controlled pore glass
      Cop-OH
      (7-hydroxylcoumarin-4-yl)propanoic acid
      AMP
      3-aminopropyl
      ACP
      6-aminocaproic
      Fmoc
      (9-fluorenylmethoxy)carbonyl
      Collagenases, a small group of the matrix metalloprotease (MMP)1 family, are highly specific proteases capable of causing hydrolytic cleavage in the triple-helical region of collagen molecules (
      • Mookhtiar K.A.
      • Van Wart H.E.
      ,
      • Mallya S.K.
      • Mookhtiar K.A.
      • Van Wart H.E.
      ). In contrast to mammalian collagenases, which cleave the collagen helix at a single site, bacterial collagenases attack multiple sites along the helix. The bacterium Clostridium histolyticum produces several collagenases, and seven types have been identified to date. These enzymes are among the most efficient known for collagen cleavage. Because of their unique activities,C. histolyticum collagenases have found broad application in the isolation of specific cell types from attendant connective tissue (
      • Wolters G.H.
      • Vos-cheperkeuter G.H.
      • Lin H.C.
      • van Schilfgaarde R.
      ,
      • Klock G.
      • Kowalski M.B.
      • Hering B.J.
      • Eiden M.E.
      • Weidemann A.
      • Langer S.
      • Zimmermann U.
      • Federlin K.
      • Bretzel R.G.
      ). Recently, class I C. histolyticumcollagenases were used to anchor signaling molecules to collagen-containing tissues, presenting a great potential for targeted drug delivery of anti-arthritis and anti-cancer reagents (
      • Nishi N.
      • Matsushita O.
      • Yuube K.
      • Miyanaka H.
      • Okabe A.
      • Wada F.
      ,
      • Matsushita O.
      • Koide T.
      • Kobayashi R.
      • Nagata K.
      • Okabe A.
      ).
      Despite the wide use of these unique proteases, the substrate specificities that are central to their proteolytic function are not well understood. In previous studies (
      • Steinbrink D.R.
      • Bond M.D.
      • Van Wart H.E.
      ,
      • Yiotakis A.
      • Hatgiyannacou A.
      • Dive V.
      • Toma F.
      ), synthetic peptide substrates derived from the native sequence of collagen were used to probe the substrate specificities of each subsite around the scissile bond of these proteases. However, the synthesis and assay of single substrates are so tedious that these studies provided only limited substrate specificity profiles for C. histolyticumcollagenases.
      Combinatorial approaches have recently been used to address the identification of substrate recognition sites in biologically important molecules (
      • Smith M.M.
      • Shi L.
      • Navre M.
      ,
      • Harris J.L.
      • Backes B.J.
      • Leonetti F.
      • Mahrus S.
      • Ellman J.A.
      • Craik C.S.
      ,
      • Isalan M.
      • Patel S.D.
      • Balasubramanian S.
      • Choo Y.
      ,
      • Tolstrup A.B.
      • Duch M.
      • Dalum I.
      • Pedersen F.S.
      • Mouritsen S.
      ,
      • Dekker N.
      • Cox R.C.
      • Kramer R.A.
      • Egmond M.R.
      ). A support-bound peptide library is one of the major types of substrate libraries and has proved useful for rapid determination of substrate specificity for several proteases such as cysteine proteases and napsin A (
      • St Hilaire P.M.
      • Willert M.
      • Juliano M.A.
      • Juliano L.
      • Meldal M.
      ,
      • Rosse G.
      • Kueng E.
      • Page M.G.
      • Schauer-Vukasinovic V.
      • Giller T.
      • Lahm H.W.
      • Hunziker P.
      • Schlatter D.
      ). A support-bound peptide library has been used successfully to examine the specificity of human fibroblast collagenase (
      • Singh J.
      • Allen M.P.
      • Ator M.A.
      • Gainor J.A.
      • Whipple D.A.
      • Solowiej J.E.
      • Treasurywala A.M.
      • Morgan B.A.
      • Gordon T.D.
      • Upson D.A.
      ).
      In the present study, we have screened the relative substrate activities (k cat/K m) of an immobilized heptapeptide library with C. histolyticumβ-collagenase (EC 3.4.24.3), a class I C. histolyticumcollagenase, which is present in most commercial preparations. The relative k cat/K m values and sequences obtained for the active substrates provide a substrate specificity profile of this enzyme from the P3 to P4′ positions. Our results provide the most complete substrate specificity profile available for C. histolyticumβ-collagenase.

      DISCUSSION

      A major advantage of the combinatorial peptide library approach used in the current study is that it permits rapid access to the relative k cat/K m values for the hydrolysis of a series of substrates. From the enzyme kinetics view, the relative k cat/K mvalues for the hydrolysis of a series of substrates are the correct kinetic quantities with which to evaluate the substrate specificity of the enzyme (
      • Nagase H.
      • Fields C.G.
      • Fields G.B.
      ). The relativek cat/K m values of the peptide samples in our peptide library with C. histolyticumβ-collagenase and the resulting structure assignments for active samples enable us to fix the substrate specificities for each subsite of this unique protease.
      The identity of amino acid at the P5′ position does not affect collagenase activity, but an aromatic amino acid residue is clearly preferred at the P4′ position. Thus, P4′ is the last position on the C-terminal side of the substrate that affects collagenase activity. Although an aromatic residue is preferred at the P4′ position, a strongly specific interaction between collagenase and the P4′ residue is not important for activity, because all other amino acids are also accommodated in this position.
      The subsites that play more critical roles in defining substrate specificity of collagenase are the P3–P3′ sites. The P3′ site has broad specificity, but there are some particular exceptions. For example, Asp, Glu, Asn, and Gln are never present at P3′ in active substrates. All other residues are tolerated in this position, although Lys, Arg, Ser, Thr, or Ala is preferred. At the P2′ position, the most strongly preferred amino acid is Pro. At P2′, His is inferior to Ser and Asn, whereas substitution with Lys or Arg greatly decreases the activity. It has been suggested that there is a hydrophobic S1′ pocket for MMPs (
      • Smith M.M.
      • Shi L.
      • Navre M.
      ). The substrate requirements at the P1′ position for many MMPS are modestly flexible if the amino acid residue in this position is hydrophobic. However, our results show that Gly is the only amino acid highly preferred at the P1′ position. Substitution of Gly by Ala decreases the hydrolytic activity about eight times, and no other amino acid residues were found at P1′ in the active substrates; this suggests a smaller hydrophobic space at the collagenase S1′ pocket than at that of other MMPs.
      On the subsites of the N-terminal side of the scissile peptide bond, all active substrates have the hydrophobic group Cop in either the P3 or P2 position, perhaps suggesting a critical role for P3 and P2 structures in positioning the substrate against the catalytic apparatus of collagenase. Cop is a bulky hydrophobic group. It has been reported that cinnamoyl and 4-nitrophenylalanyl are also preferred at the P3 and P2 positions. Like Cop, both cinnamoyl and 4-nitrophenylalanyl are hydrophobic groups. The preference for a bulky hydrophobic group in P3 and P2 indicates that collagenase has developed a large hydrophobic surface pocket at S3 and at S2. Tyr and Lys were also found at the P3 and P2 positions for sample Cop-AAAYKBA (Table IV). Apparently, Tyr and Lys are not the preferred residues because they occur in the minor cleavage.
      Collectively, the P3 through P4′ specificities are all important for hydrolysis of collagenase. Furthermore, the roles of these positions in aligning the substrate with the catalytic apparatus of the enzyme, which is the first step in proteolysis, can be evaluated in a comparative sense. Thus, the P3 and P2 positions play a significantly more critical role in defining substrate activity than do other positions. This result is illustrated by Cop-Ala-Ala-Ala-Gly-Pro-B-A (Table IV). As described above, Gly-Pro are the best residues for the P1′ and P2′ positions. If collagenase had developed S1′ and S2′ subsites as the determinative sites for alignment of the substrate with the catalytic apparatus, sample Cop-Ala-Ala-Ala-Gly-Pro-B-A would bind to the enzyme with Cop, Gly, and Pro in P4, P1′, and P2′, respectively. On the other hand, if the requirement of a bulky hydrophobic group in either the S3 or S2 site of the enzyme were to play the more important role in positioning the peptide, the Gly and Pro of sample Cop-AAAGPBA should not be found at the P1′ and P2′ positions. In fact, this sample is hydrolyzed by collagenase in a single site, with splitting of the peptide into two segments, Cop-Ala and Ala-Ala-Gly-Pro-B-Ala. The peptide is therefore positioned across the catalytic apparatus with Cop, Gly, and Pro at P2, P3′, and P4′, respectively. From this result, we conclude that interactions at P3 and P2 are the determinative factors for substrate positioning with C. histolyticum β-collagenase. It should be noted that the primary specificity of many MMPs is dominated by a hydrophobic S1′ pocket (
      • Smith M.M.
      • Shi L.
      • Navre M.
      ). Although the P1′ position also demonstrated a high preference for small residues of Gly or Ala, this position is certainly not the dominant site for fixing the substrate specificity of this enzyme. It is reported that serine proteases from both the subtilisin and chymotrypsin families have a dominant S1 pocket (
      • Perona J.J.
      • Craik C.S.
      ). Perhaps the closest similarity to the collagenase for substrate binding is found in papain families, which utilize S2 pocket to fix specificity (
      • Turk D.
      • Guncar G.
      • Podobnik M.
      • Turk B.
      ). However, the high preference for Gly at P1′, which is seen for collagenase, is not observed in papains. The primary specificity of collagenase is different from most proteases, including many other MMPs (
      • Smith M.M.
      • Shi L.
      • Navre M.
      ). This may be the major reason for the unique activity of this enzyme for hydrolyzing collagen molecules that are resistant to most proteases.
      The results from the library cleavage analysis show a unique specificity fingerprint of collagenase. Previously, the action of this enzyme was studied primarily through the use of synthetic short peptides, which were derived from the sequence around the cleavage site of collagenase in the native collagen molecule. Those studies produced some valuable information on the substrate specificity of collagenase. For example, Van Wart and co-workers (
      • Steinbrink D.R.
      • Bond M.D.
      • Van Wart H.E.
      ) discovered a preference for unbranched hydrophobic side chains with Tyr > Phe > Ala at the P1 position. This point was not addressed in our study because of the specificity dominance of the P2 and P3 positions. However, because of the tedium of the synthesis and assay of individual peptides, the previous studies did not explore complete diversity at each position. Additionally, the peptides used in previous studies were exclusively short, none of them extending beyond P3′ on the C-terminal side of the scissile bond. In contrast, the use of the combinatorial peptide library, which provides complete diversity of amino acids beyond the P1position, surmounts the limitations present in the previous studies. The substrate specificity profile gathered is in agreement with known specificities and provides a “fingerprint” for the protease. The ability of the library to screen all possible substrates yields not only sensitive substrates but also suboptimal substrates, which enables us to determine the roles of individual positions with high preference in a comparative sense.
      In conclusion, we have used a unique and efficient approach for mapping the substrate specificity of the collagenase. However, the method will not be applicable to all proteases. The requirement of some proteases, such as aspartyl proteases, for interactions C-terminal to the cleavage site may limit the usefulness of these substrates. However, we have evidence that some other MMPs have the ability to hydrolyze the immobilized substrates (data not shown). The combinatorial library should be applicable to those proteases for rapid access to detailed specificity information, which can be used as a starting point in the design and synthesis of potent and selective inhibitors.

      Acknowledgments

      We thank Dr. Richard L. Schowen for valuable comments.

      REFERENCES

        • Mookhtiar K.A.
        • Van Wart H.E.
        Matrix. 1992; 1: 116-126
        • Mallya S.K.
        • Mookhtiar K.A.
        • Van Wart H.E.
        J. Protein Chem. 1992; 11: 99-107
        • Wolters G.H.
        • Vos-cheperkeuter G.H.
        • Lin H.C.
        • van Schilfgaarde R.
        Diabetes. 1995; 44: 227-233
        • Klock G.
        • Kowalski M.B.
        • Hering B.J.
        • Eiden M.E.
        • Weidemann A.
        • Langer S.
        • Zimmermann U.
        • Federlin K.
        • Bretzel R.G.
        Cell Transplant. 1996; 5: 543-551
        • Nishi N.
        • Matsushita O.
        • Yuube K.
        • Miyanaka H.
        • Okabe A.
        • Wada F.
        Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 7018-7023
        • Matsushita O.
        • Koide T.
        • Kobayashi R.
        • Nagata K.
        • Okabe A.
        J. Biol. Chem. 2001; 276: 8761-8770
        • Steinbrink D.R.
        • Bond M.D.
        • Van Wart H.E.
        J. Biol. Chem. 1985; 260: 2771-2776
        • Yiotakis A.
        • Hatgiyannacou A.
        • Dive V.
        • Toma F.
        Eur. J. Biochem. 1988; 172: 761-766
        • Smith M.M.
        • Shi L.
        • Navre M.
        J. Biol. Chem. 1995; 270: 6440-6449
        • Harris J.L.
        • Backes B.J.
        • Leonetti F.
        • Mahrus S.
        • Ellman J.A.
        • Craik C.S.
        Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 7754-7759
        • Isalan M.
        • Patel S.D.
        • Balasubramanian S.
        • Choo Y.
        Biochemistry. 2001; 40: 830-836
        • Tolstrup A.B.
        • Duch M.
        • Dalum I.
        • Pedersen F.S.
        • Mouritsen S.
        Gene (Amst.). 2001; 263: 77-84
        • Dekker N.
        • Cox R.C.
        • Kramer R.A.
        • Egmond M.R.
        Biochemistry. 2001; 40: 1694-1701
        • St Hilaire P.M.
        • Willert M.
        • Juliano M.A.
        • Juliano L.
        • Meldal M.
        J. Comb. Chem. 1999; 1: 509-523
        • Rosse G.
        • Kueng E.
        • Page M.G.
        • Schauer-Vukasinovic V.
        • Giller T.
        • Lahm H.W.
        • Hunziker P.
        • Schlatter D.
        J. Comb. Chem. 2000; 2: 461-466
        • Singh J.
        • Allen M.P.
        • Ator M.A.
        • Gainor J.A.
        • Whipple D.A.
        • Solowiej J.E.
        • Treasurywala A.M.
        • Morgan B.A.
        • Gordon T.D.
        • Upson D.A.
        J. Med. Chem. 1995; 38: 217-219
        • Furka A.
        • Sebestyen F.
        • Asgedom M.
        • Dibo G.
        Int. J. Pept. Protein Res. 1991; 37: 487-493
        • Ator M.A.
        • Beigel S.
        • Dankanich T.C.
        • Echols M.
        • Gainor J.A.
        • Gilliam C.L.
        • Gordon T.D.
        • Koch D.
        • Koch J.F.
        • Kruse L.I.
        • Morgan B.A.
        • Krupinski-Olsen R.
        • Siahaan T.J.
        • Singh J.
        • Whipple D.A.
        Hodges R. Smith J. Proceeding of the 13th America Peptide Symposium, Edmonton, Canada, June 13–18, 1993. Escom Science Publishers B.V., Leiden, The Netherlands1994: 1012-1015
        • Nagase H.
        • Fields C.G.
        • Fields G.B.
        J. Biol. Chem. 1994; 269: 20952-20957
        • Perona J.J.
        • Craik C.S.
        Protein Sci. 1995; 4: 337-360
        • Turk D.
        • Guncar G.
        • Podobnik M.
        • Turk B.
        Biol. Chem. 1998; 379: 137-147