Activity of matrix metalloproteinase-9 against native collagen types I and III Heather F. Bigg 1 , Andrew D. Rowan 1 , Michael D. Barker 2 and Tim E. Cawston 1 1 Musculoskeletal Research Group, Institute of Cellular Medicine, The Medical School, Newcastle University, UK 2 Division of Genomic Medicine, Academic Unit of Pathology, University of Sheffield, Medical School, UK Collagens are the major structural proteins of connect- ive tissues such as skin, bone, cartilage and tendon. Interstitial collagen types I, II and III are the most abundant, and the native triple helical structure of these molecules makes them highly resistant to proteo- lysis. However, collagenases of the matrix metallopro- teinase (MMP) family [1] cleave native collagen types I, II and III at a specific site in all three chains of the triple helix, approximately three-quarters of the length from the N-terminus. The action of these col- lagenase enzymes is therefore critical for the initiation of collagenolysis. Once initiated, the cleaved helix unwinds at physiological temperatures and becomes susceptible to degradation by other, less-specific pro- teinases. MMP collagenases are active at neutral pH and play a highly important role in collagen degrada- tion in vivo. The mammalian MMP collagenases cur- rently include the ‘classical’ collagenases, MMP-1, MMP-8 and MMP-13 [2–4] and also the gelatinolytic enzyme, MMP-2 [5–7], and MMP-14 (MT1-MMP) [8], a member of the membrane-type subclass of MMPs. MMP-9 (also known as gelatinase B, 92 kDa gela- tinase or 92 kDa type IV collagenase, EC 3.4.24.35) shares a close structural similarity with MMP-2 [9,10]. It was originally identified as a gelatinolytic enzyme produced by polymorphonuclear leukocytes [11] and subsequent studies have demonstrated secretion in the latent form (proMMP-9) by a variety of cell types. It has also been implicated in the pathogenesis of several human diseases, including arthritis [12–15]. Unlike other MMPs, MMP-9 and MMP-2 both contain three fibronectin type II repeats inserted into the catalytic Keywords arthritis; collagen I; collagen III; collagenase; matrix metalloproteinase-9 Correspondence T. E. Cawston, Musculoskeletal Research Group, 4th Floor, Catherine Cookson Building, The Medical School, Framlington Place, Newcastle University, Newcastle- upon-Tyne, NE2 4HH, UK Fax: +44 191 2225455 Tel: +44 191 2225397 E-mail: t.e.cawston@ncl.ac.uk Website: http://www.ncl.ac.uk/medi/ research/rheumatology/ (Received 3 November 2006, revised 20 December 2006, accepted 22 December 2006) doi:10.1111/j.1742-4658.2007.05669.x Interstitial collagen types I, II and III are highly resistant to proteolytic attack, due to their triple helical structure, but can be cleaved by matrix metalloproteinase (MMP) collagenases at a specific site, approximately three-quarters of the length from the N-terminus of each chain. MMP-2 and -9 are closely related at the structural level, but MMP-2, and not MMP-9, has been previously described as a collagenase. This report inves- tigates the ability of purified recombinant human MMP-9 produced in insect cells to degrade native collagen types I and III. Purified MMP-9 was able to cleave the soluble, monomeric forms of native collagen types I and III at 37 °C and 25 °C, respectively. Activity against collagens I and III was abolished by metalloproteinase inhibitors and was not present in the concentrated crude medium of mock-transfected cells, demonstrating that it was MMP-9-derived. Mutated, collagenase-resistant type I collagen was not digested by MMP-9, indicating that the three-quarters ⁄ one-quarter locus was the site of initial attack. Digestion of type III collagen generated a three-quarter fragment, as shown by comparison with MMP-1-mediated cleavage. These data demonstrate that MMP-9, like MMP-2, is able to cleave collagens I and III in their native form and in a manner that is char- acteristic of the unique collagenolytic activity of MMP collagenases. Abbreviations APMA, p-aminophenylmercuric acetate; MMP, matrix metalloproteinase; TIMP, tissue inhibitor of metalloproteinases. 1246 FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS domain which are thought to mediate the ability to bind collagen [16,17]. However, this domain does not appear to be essential for the collagenolytic activity of MMP-2 [7]. MMP-9, unlike MMP-2, also contains an additional 54-amino acid proline-rich insertion, homol- ogous to the a 2 chain of type V collagen [10]. To date, MMP-9 has not been described as a collagenase. Sev- eral previous studies have investigated its ability to digest native collagen types I, II and III using enzyme from a variety of sources, both natural and recombin- ant [5,6,18–23]. Three of these demonstrated an inabil- ity to degrade soluble native collagen I at 22 or 25 °C [5,6,18]; a lack of digestion at 37 °C was additionally reported by Murphy et al. [18]. However, another study [20] has shown digestion of soluble native colla- gen I at 30 and 37 °C. Four previous studies have examined the digestion of collagen II and all report no degradation by MMP-9 [6,18,21,22]. However, investi- gation into the ability to digest native collagen III has produced disparate results. Three reports [6,21,22] des- cribe degradation at 22, 25 or 27 °C, whilst two others [18,20] report digestion at 37 but not 30 °C [20] or no digestion at 25 or 37 °C [18]. Furthermore, none of these previous studies has investigated the initial MMP-9-mediated cleavage site of collagens I and III. The possibility of contamination with another MMP collagenase is very difficult to exclude when working with a natural source. However production of recom- binant protein can also present problems with respect to correct folding of the enzyme, particularly when prokaryotic cells are used. Both of these issues can be avoided by expression in insect cells, since these do not appear to produce collagenolytic metalloproteinases. This report therefore investigates the ability of purified recombinant human MMP-9 produced in insect cells to cleave native collagen types I and III; in addition, the initial MMP-9-mediated cleavage site of these sub- strates is investigated for the first time. Results Characterization of purified, recombinant, human proMMP-9 produced in insect cells Purified, recombinant, human MMP-9 expressed in insect cells was used to examine the ability of this enzyme to cleave native collagens I and III. Recombin- ant human proMMP-9 was purified from the condi- tioned medium of pIB-proMMP-9-transfected insect cells and characterized by silver staining, western blot- ting and zymography (Fig. 1). Silver staining under reducing conditions (+ bme) revealed two bands with apparent molecular masses of 85 and 61 kDa, which were identified as MMP-9 by western blotting with two anti-MMP-9 sera (Fig. 1A). Human proM- MP-9 from natural sources has a M r of 92 kDa [10,21,24] and contains both N- and O-linked carbohy- drate [10]. The molecular mass of the unglycosylated protein is 81 kDa [10]. The 85-kDa species may therefore be a differentially glycosylated full-length form of proMMP-9, but it is also possible that trunca- tion of the polypeptide occurs during expression and 97.4 66.2 45 MAB911 A560/8 M r M r (kDa) (kDa)(kDa) SILVER STAIN W. BLOT + M r 97.4 66.2 45 βme SILVER STAIN 98 64 50 A560/8 (kDa) M r βme W. BLOT 148 MAB911 - ZYMOGRAM APMA - + ++++ 31 148 98 64 50 36 31 36 M r M r βme AB C Fig. 1. Characterization of recombinant human proMMP-9 by silver staining, western blotting and zymography. (A,B) Purified recombinant human proMMP-9 (0.5 lg) was run reduced (A, + bme) and nonreduced (B, – bme) on 10% SDS ⁄ PAGE gels followed by silver staining (SILVER STAIN) or transfer to nitrocellulose (W. BLOT) as described in Experimental procedures. Western blots were probed with a polyclonal sheep anti-(porcine MMP-9 serum) (A560 ⁄ 8, 2 lgÆmL )1 ) and a monoclonal mouse anti-(human MMP-9) serum (MAB911, 2 lgÆmL )1 ). M r , the positions of molecular mass markers (kDa) are shown. The individual bands present in nonreduced lanes (B) are marked by arrows. (C) Recombinant human proMMP-9 (1.6 ng) was run nonreduced (– bme) on a 10% gelatine zymogram (ZYMOGRAM) either with (+) or without (–) prior activation by 0.67 m M APMA (APMA) for 90 min at 37 °C. H. F. Bigg et al. MMP-9 activity against collagens I and III FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS 1247 purification. The 61-kDa species is clearly a truncated MMP-9 fragment, on account of its lower molecular mass and immunoreactivity to two anti-MMP-9 sera (Fig. 1A). Silver staining and western blotting under nonreduc- ing conditions (– bme) (Fig. 1B) revealed additional higher molecular mass bands, which therefore appear to be disulphide-bonded MMP-9 complexes. In addi- tion, the 85-kDa species displayed heterogeneity when run nonreduced, since it migrated as three separate bands; this may result from differential disulphide bond formation. Zymography (nonreducing condi- tions, – bme) (Fig. 1C) revealed a similar pattern of gelatinolytic bands, except that a single band only was seen for the 85-kDa species. This indicates that the additional nonreduced forms of this species lack gela- tinolytic activity, which may be due to incorrect disul- phide bond formation. The 61-kDa form appears to be an active site-containing fragment, as it has gelatinoly- tic activity (Fig. 1C). Activation of the proMMP-9 with p-aminophenylmercuric acetate (APMA) prior to zymography increased the migration of all species, including the 61-kDa fragment, therefore demonstra- ting that all species were proenzyme forms (Fig. 1C). Purified, recombinant, human MMP-9 cleaves soluble, native type I collagen The ability of purified recombinant human MMP-9 produced in insect cells to cleave type I collagen was tested using soluble substrate at 37 °C (Fig. 2). The collagen retained its native, triple helical structure under these assay conditions, as it remained resistant to trypsin (Figs 2A.T). The activity of the trypsin was confirmed by total lysis of denatured substrate (Fig. 2A,T,denat). Importantly, preparation of the type I collagen did not include pepsin digestion, as this may result in increased susceptibility to gelatinolysis at 37 °C. Furthermore, trypsin sensitivity is a reliable indicator of whether the collagen is susceptible to a gelatinolytic attack, as progressive heat denaturation of the collagen at increasing temperatures shows that resistance to trypsin is lost under the same conditions as resistance to gelatinolysis (data not shown). Extensive digestion of the type I collagen b and a chains was seen in the presence of either MMP-1 or proMMP-9 when combined with APMA. MMP-1- mediated cleavage did not generate the characteristic three-quarter length fragments seen at lower incubation temperatures, because at 37 °C these cleavage products spontaneously denature and are susceptible to further MMP-1-mediated gelatinolytic degradation. For the same reason, no fragments at all were detected in the presence of MMP-9, as this enzyme is a potent gelatinase. No collagen digestion was seen without APMA activation of the proMMP-9 (Fig. 2A, – APMA, + proMMP-9). However, some conversion of b 12 dimers to a 1 and a 2 monomers is apparent in this lane, to give an increased level of both the a chains and a slightly increased mobility of the a 2 chain. This indi- cates the presence of a 2 chain telopeptidase activity, resulting from a low level of spontaneous proMMP-9 activation during the assay. Digestion of the a 2 N-ter- minal telopeptides of native type I collagen by MMP-9 has been reported previously [21]. MMP-9-mediated digestion (both collagenolytic and telopeptidase) was -+ -+-+ MAB911 proMMP-9 + APMA MMP-1 MMP-13 β 11 β 12 α 1 α 2 b u f f e r M M P - 1 TT d e n a t -+ + + + +++ + proMMP-9 APMA E D T A s e r i n e / ++ + - + ++ + - T I M P - 2 E t O H 1 , 1 0 m o c k β 11 β 12 α 1 α 2 c y s t e i ne + - m o c k A B Fig. 2. Recombinant human MMP-9 cleaves soluble, triple helical type I collagen. (A) Soluble type I collagen from bovine skin (27 lgÆ lane )1 ) was digested for 72 h at 37 °C with buffer alone (buffer), 0.3 lg MMP-1 (MMP-1), 0.2 lg trypsin (T), 0.5 lg recombinant human proMMP-9 (proMMP-9) or 24 lg of protein from the concen- trated crude culture medium of mock-transfected insect cells (mock), in the absence or presence of 0.6 m M APMA, as indicated. Additional lanes also contained the following enzyme inhibitors: 6 m M EDTA (EDTA), 8 m M 1,10-phenanthroline (1,10), 2.3 lg TIMP-2 (TIMP-2), serine and cysteine protease inhibitors at the manufacturer’s recom- mended working strength (serine ⁄ cysteine) or the ethanol solvent used for 1,10-phenanthroline (EtOH). The efficacy of the trypsin was demonstrated by cleavage of denatured substrate (denat). The micr- ogram enzyme–substrate ratio of MMP-9–type I collagen is 1 : 54. The positions of the uncut collagen b and a chains (b 11 , b 12 , a 1 and a 2 ) are indicated. Cleavage of type I collagen by proMMP-9 combined with APMA was investigated in five separate experi- ments, with similar results each time. (B) The effect of 4 lg mono- clonal anti-(human MMP-9) serum (MAB911) on cleavage mediated by 0.2 lg MMP-1 (MMP-1), 0.2 lg MMP-13 (MMP-13) or 0.1 lg recombinant human proMMP-9 (proMMP-9) combined with APMA (+ APMA) is shown. The microgram enzyme–substrate ratio of MMP-9–type I collagen is 1 : 270. The positions of the uncut colla- gen b and a chains (b 11 , b 12 , a 1 and a 2 ) are indicated. MMP-9 activity against collagens I and III H. F. Bigg et al. 1248 FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS abolished by EDTA, 1,10-phenanthroline and tissue inhibitor of metalloproteinases (TIMP)-2, but not by serine and cysteine protease inhibitors, or the ethanol vehicle for the 1,10-phenanthroline (Fig. 2A), thereby demonstrating metalloproteinase-mediated activity. No cleavage was observed in the presence of crude insect cell culture medium (16.5-fold concentrate) conditioned by mock-transfected cells (chloramphenicol acetyl transferase vector) (Fig. 2A, mock), therefore excluding the possibility of a contaminating insect cell protease. Furthermore, collagenolytic digestion with a lower level of MMP-9 was blocked by a monoclonal anti-MMP-9 antibody, whereas cleavage mediated by MMP-1 or MMP-13 was unaffected or affected only slightly (Fig. 2B, – ⁄ +MAB911). Taken together, these data convincingly demonstrate that recombinant human MMP-9 is capable of cleaving native, trypsin-resistant, soluble type I collagen. The initial cleavage of type I collagen by MMP-9 is at the three-quarters ⁄ one-quarter locus A hallmark of MMP collagenolytic activity is the abil- ity to perform the initial cleavage of native substrate at the three-quarters ⁄ one-quarter site. To investigate the initial cleavage site of MMP-9-mediated type I col- lagen digestion, we examined its ability to digest mutated type I collagen which is completely resistant to collagenolytic cleavage, due to the mutation of Gln774 (P 2 ) and Ala777 (P¢ 2 ) of the a 1 (I) chain three- quarters ⁄ one-quarter site to proline. The wild-type a 2 (I) chain of each triple helix is also not cleaved, due to the presence of two mutated a 1 (I) chains [25]. Mutated type I collagen was not cleaved by MMP-1 or MMP-13, as expected (Fig. 3, mutated, MMP-1, MMP-13), although telopeptidase activity was evident in the presence of MMP-1. Under identical conditions, wild-type collagen I was digested by both these enzymes; the characteristic three-quarters fragments are not seen, because at 36 °C, these cleavage products spontaneously denature and are susceptible to further gelatinolytic degradation. Mutated type I collagen was also resistant to cleavage mediated by human recom- binant MMP-9 (Fig. 3, mutated, proMMP-9, APMA). Under identical conditions, wild-type collagen I was digested by MMP-9 (Fig. 3, wild-type, proMMP-9, APMA) to give extensive degradation without the appearance of partially digested fragments, as for Fig. 2. The resistance of the mutated type I collagen to MMP-9-mediated digestion demonstrates that this enzyme makes the initial cut at the three-quarters ⁄ one-quarter locus, which is a characteristic of MMP collagenolytic action. Importantly, these data also exclude gelatinolytic degradation of partially unfolded wild-type collagen I by MMP-9, as under the same conditions, this mechanism would also result in suscep- tibility of the mutated collagen. MMP-9 cleaves native, triple helical type III collagen to generate a 3/4 fragment The ability of recombinant human MMP-9 to cleave collagen type III was investigated in assays with sol- uble substrate and compared with the ability to cleave type I. Type III collagen was cleaved at 25 °Cby recombinant MMP-9, to produce a fragment with a similar mobility to that of the MMP-1-generated three-quarter piece (Fig. 4, type III, compare MMP-1 with + proMMP-9, + APMA) [26]. Digestion with proMMP-9 and APMA was abolished by EDTA, 1,10-phenanthroline and TIMP-2, but not by serine and cysteine protease inhibitors, or the ethanol vehicle for the 1, 10-phenanthroline, demonstrating metallo- proteinase-mediated cleavage. In addition, no digestion was observed with concentrated crude insect cell culture medium from mock-transfected cells (Fig. 4, type III, mock). A low level of cleavage was seen in the absence of APMA (Fig. 4, type III, – APMA, + proMMP-9), indicating some spontaneous activation of the proMMP-9 during the assay. Minor cleavage was also seen with trypsin, in agreement with a previ- ous report demonstrating specific cleavage of native type III collagen with this enzyme (Fig. 4, type III, T) [27]. The more extensive digestion of denatured type III collagen by trypsin (Fig. 4, type III, T, denat) mutated wild-t y pe b u f f e r M M P - 1 TT d e n a t M M P - 1 3 p r o M M P - 9 bu f f e r M M P - 1 TT d e n a t M M P - 1 3 p r o M M P - 9 β 11 β 12 α 1 α 2 A P M A A P M A β 11 β 12 α 1 α 2 Fig. 3. Recombinant human MMP-9 cleaves native type I collagen at the three-quarters ⁄ one-quarter locus. Soluble type I collagen (27 lgÆlane )1 ) from bovine skin (wild-type) or mouse skin (mutated) was digested for 98 h at 36 °C with buffer alone, 0.6 lg MMP-1, 0.2 lg trypsin (T), 0.5 lg MMP-13 or 0.5 lg recombinant human proMMP-9 in the additional presence of 0.6 m M APMA. The effic- acy of the trypsin was demonstrated by cleavage of denatured sub- strate (denat). The positions of the uncut collagen b and a chains (b 11 , b 12 , a 1 and a 2 ) are shown. Cleavage of wild-type and mutated type I collagen by proMMP-9 combined with APMA was compared in two separate experiments, with similar results on each occasion. H. F. Bigg et al. MMP-9 activity against collagens I and III FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS 1249 confirms that all other conditions represent cleavage of native rather than denatured substrate. Taken together, these data clearly demonstrate the ability of recombinant human MMP-9 to cleave native type III collagen. Gel-scanning densitometry of the data in Fig. 4 indicates cleavage of 42% of the type III sub- strate by MMP-9. Under the same conditions as the type III assay, MMP-1 cleaved type I collagen to give characteristic three-quarter fragments, but no digestion was observed with proMMP-9 and APMA (Fig. 4, type I). This indi- cates that recombinant MMP-9 cleaves type III collagen more effectively than type I, as digestion of type III was seen at 25 °C whereas digestion of type I occurred only at the higher temperatures of 36 °Cor37°C (Figs 2 and 3). Recombinant MMP-9 was also able to digest colla- gen III at 35 and 36 °C, as well as at 25 °C; however, at 36 °C, extensive substrate digestion was also seen in the presence of trypsin, making it difficult to ascertain that the collagen retained its native conformation at this tem- perature (data not shown). Discussion A number of previous reports have investigated the ability of MMP-9 to degrade native collagen types I and III [5,6,18–23] with disparate results. In this study, recombinant human MMP-9 was expressed in insect cells and the ability of enzyme purified from this source to digest native collagens I and III was evaluated. Importantly, the possibility of contaminating, endog- enous collagenolytic activity was excluded, as shown by the lack of substrate cleavage seen with concentrated, crude insect cell culture medium from mock-transfected cells. The data in this report therefore conclusively demonstrate that MMP-9 is able to digest soluble, native collagen types I and III at 37 and 25 °C, respect- ively. Furthermore, the location of substrate cleavage sites was also investigated, demonstrating for the first time that MMP-9 attacks native collagens I and III initially at the three-quarters ⁄ one-quarter site. Several previous studies report that MMP-9 is unable to digest native collagen I [5,6,18,19,23]. In two of these [19,23], the precise assay conditions are not described and it is therefore difficult to compare these findings with the data reported here. Aimes and Quigley [5], Konttinen et al. [6] and Murphy et al. [18] performed assays at either 22 or 25 °C and reported no digestion at these temperatures, in agreement with the findings of this study. The latter study [18] also reported no degra- dation of native collagen I at 37 °C, but these data are described in the text only and therefore cannot readily be compared with the data reported here. In agreement with our study, a further report [20] describes digestion at both 30 and 37 °C, but in this case, the collagen I substrate was pepsin-treated and therefore possibly sus- ceptible to a gelatinolytic attack; furthermore, resist- ance to trypsin was not demonstrated. Three previous studies [6,21,22] have shown digestion of native collagen III by MMP-9 at 22, 25 and 27 °C, respectively, in agreement with the data reported here. However, another study [18] reported no degradation at either 25 or 37 °C. The discrepancy at 25 °C may be due to differences in the quantity of enzyme and assay period; although the amounts of substrate were similar, we used more enzyme (4.5·) in a longer assay (5·). A shorter assay time (5· less) with less enzyme (0.6·) may also explain the reported lack of digestion at 30 °C [20]; the same study indicated that collagen III is degraded under these conditions at 37 °C. Of the existing MMPs, MMP-9 is most closely rela- ted to MMP-2 at the structural level. The C-terminal, Type III β 11 α 1 b u f f e r M M P - 1 TT d e n a t APMA- ++++ + + E D T A E t O H 1 , 1 0 T I M P - 2 proMMP-9 m o c k + ++ ++ + + s e r i n e / c y s t e i n e m o ck - +- - p r o MM P - 9 T yp e I β 12 β 11 α 1 α 2 b u f f e r MM P - 1 TT d e n a t A P MA Fig. 4. Recombinant human MMP-9 cleaves soluble, triple helical type III collagen to generate a three-quarters ⁄ one-quarter fragment. Soluble type III or type I collagen (27 lgÆlane )1 ), as indicated, was digested for 98 h at 25 °C with buffer alone, 0.6 lg MMP-1, 0.2 lg trypsin (T), 0.5 lg recombinant human proMMP-9 or 24 lg of pro- tein from the concentrated crude culture medium of mock-trans- fected insect cells (mock) in the absence or presence of 0.6 m M APMA, as indicated. Additional lanes also contained the following enzyme inhibitors: 6 m M EDTA, 8 mM 1,10-phenanthroline (1,10), 2.3 lg TIMP-2, serine and cysteine protease inhibitors at the manu- facturer’s recommended working strength or the ethanol solvent used for 1,10-phenanthroline (EtOH). The efficacy of the trypsin was demonstrated by cleavage of denatured substrate (denat). The microgram enzyme–substrate ratio of MMP-9–collagen is 1 : 54. The positions of the uncut collagen b and a chains (b 11 , b 12 , a 1 and a 2 ) are indicated. Cleavage of type III collagen by proMMP-9 com- bined with APMA was investigated in four separate experiments, with similar results each time. MMP-9 activity against collagens I and III H. F. Bigg et al. 1250 FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS hemopexin-like domain of MMP-2 is essential for its collagenolytic activity [7], as is also the case for the ‘classical’ collagenases [28–33]. In contrast, the three fibronectin type II repeats, which are shared with MMP-9 but not other MMPs, are not absolutely required for collagenolysis, although their presence does enhance activity [7]. MMP-9, unlike MMP-2, contains a unique, 54 amino acid, proline-rich inser- tion, immediately before the hinge region, which is homologous to the a 2 chain of type V collagen [10]. This extra domain, the function of which is currently unknown, may play a role in the ability of MMP-9 to cleave native collagens. MMP-2 cleaves all three of the interstitial collagen types I, II and III [5–7] with a sub- strate preference of type III > type II > type I [6]. The data shown here indicate that MMP-9 also has a clear substrate preference for type III above type I; digestion of type II was not investigated, but in con- trast to MMP-2, all previous studies report a lack of cleavage by MMP-9 [6,18,21,22]. The data presented in this report demonstrate for the first time that MMP-9 performs the initial cleavage of native collagens I and III at the three-quarters ⁄ one- quarter site. Type I collagen, which is completely resistant to collagenolysis due to mutation of residues close to the three-quarters ⁄ one-quarter site in the a 1 (I) chain, was not digested by MMP-9. In addition, com- parison of the fragments generated by MMP-1 and MMP-9-mediated cleavage of type III collagen indica- ted digestion at the three-quarters ⁄ one-quarter locus by MMP-9. In this respect, MMP-9-mediated collagen cleavage resembles that of the MMP collagenases, since cleavage at this site is a hallmark of these enzymes. However, MMP-9 does not completely fit the description of a stereotypical MMP collagenase, as it does not appear to cleave type II collagen [6,18,21,22]; all the existing mammalian collagenases are able to cleave all three interstitial collagen types [2–4,6,8,34]. MMP-9 was also unable to cleave soluble type I colla- gen at 25 °C, but this finding does not preclude a phy- siological role in the degradation of native soluble substrate at 37 °C and the ability of MMP-9 to digest the latter in vitro is clearly demonstrated in this study. A clear association of MMP-9 with rheumatoid and inflammatory arthritis and a correlation between syn- ovial fluid levels of proMMP-9 and the collagen degra- dation product, hydroxyproline, in rheumatoid arthritis [35] indicate a role for MMP-9 in these dis- eases in the destruction of type I collagen in bone and type III collagen in synovium. MMP-9 levels in rheu- matoid and inflammatory synovial fluids are higher than those from noninflammatory or osteoarthritic patients [12–14]; furthermore, net gelatinolytic activity is not found in osteoarthritic synovial fluids but does occur in approximately one-quarter of rheumatoid samples [15]. Rheumatoid synovial tissue also contains significantly more MMP-9 than osteoarthritic syno- vium, although increased levels are observed in osteo- arthritic tissue with a heightened inflammatory response [12]. In addition, immunostaining of MMP-9 in experimental models of inflammatory arthritis shows a correlation with disease progression [36]; further- more, antibody-induced arthritis, which is one of the murine models of rheumatoid disease, is less severe in MMP-9 knockout mice [37]. Immunolocalization studies also show intense stain- ing for MMP-9 in the osteoclasts of both normal and rheumatoid knee joints [38], suggesting a role in both normal and pathological bone resorption. Osteoclasts also produce very low levels of TIMP-1 compared with MMP-9 [39]; the majority of MMP-9 produced should therefore be free from complex formation with this inhibitor. The gelatinolytic activity of MMP-9 is opti- mal at pH 7.5, but 50–80% of the full activity is retained at pH 5.5–6.0 [38]. MMP-9 is also activated by acidic conditions [38,40]. These observations, com- bined with its ability to solubilize collagen from demin- eralized bone particles [41] and to produce visible degradation of the collagen fibrils, as shown by elec- tron microscopy [38], strongly suggest a role in colla- gen degradation during bone resorption. Digestion of the a 2 N-terminal telopeptides of native type I collagen by MMP-9 [21] is also likely to be important in bone collagen resorption, as highly cross-linked type I colla- gen, such as that found in bone collagen fibrils, is reported to be resistant to enzymes such as MMP-1, without prior removal of the cross-link-containing ter- minal peptides by telopeptidases. Surprisingly, the bone histology of MMP-9 knockout mice does not reveal a lack of osteoclastic resorption of mineralized matrix [42], but other enzymes may be compensating for its absence. Indeed, another study [43], using a spe- cific neoepitope antibody, detected collagen fragments resulting from a three-quarters ⁄ one-quarter cut along the spicules of trabecular bone in the developing tibiae of wild-type, but not MMP-9 knockout mice, provi- ding in vivo evidence for the involvement of MMP-9 in bone collagen turnover. Experimental procedures Antibodies A rabbit polyclonal anti-MMP-9 serum was purchased from Sigma-Aldrich (Gillingham, UK), and a mouse mono- clonal antibody against human MMP-9 (MAB911) was H. F. Bigg et al. MMP-9 activity against collagens I and III FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS 1251 from R & D Systems Europe Ltd (Abingdon, UK). A sheep polyclonal antibody against porcine MMP-9 (A560 ⁄ 8) was a generous gift from R. Hembry (University of East Anglia, Norwich, UK). MMP and TIMP proteins Human proMMP-1 was expressed in Escherichia coli, refold- ed and purified as previously described [44,45] and quantitat- ed by enzyme-linked immunosorbent assay [46]. Refolding caused activation of proMMP-1 to the fully active form or conversion to an intermediate lacking the polyhistidine tag and the first four amino acids of the proenzyme. Activation with APMA resulted in conversion of the intermediate form to the fully active enzyme. Enzyme assays were performed without APMA activation and therefore measured only the fully active form present in the refolded sample. The cDNA for human MMP-13 was generously provided by V. Kna ¨ uper (University of York, York, UK). Oligonu- cleotides to amplify the mature MMP-13 sequence were designed, incorporating an initiating ‘ATG’ in the forward primer and a ‘TAG’ stop codon in the reverse primer. Polymerase chain reaction was performed with Pfu DNA polymerase (Promega, Southampton, UK) and the frag- ment was ligated into pRSETA (Invitrogen Ltd, Paisley, UK), followed by expression in E. coli cells, inclusion body extraction and refolding as for proMMP-1 [44,45]. Active MMP-13 was quantitated by BCA assay (Pierce Biotechno- logy, Inc., Rockford, IL, USA) and stored in the presence of 10 )5 m CI-C, to prevent autocatalytic fragmentation. CI-C, a broad-spectrum peptide hydroxamic acid MMP inhibitor [47] was donated by SmithKline Beecham Phar- maceuticals (Harlow, UK). Human proMMP-9 cDNA, including the stop codon, was subcloned from pCOC-MMP-9 (a kind gift from G. Murphy, University of Cambridge, Cambridge, UK) into the pIB ⁄ V5-His-TOPO expression vector (Invitrogen). Following sequence confirmation, the pIB-proMMP-9 con- struct was transfected into Hi-5 insect cells (Invitrogen) grown in serum-free HyClone HyQ SFX-Insect culture medium (Perbio Science UK Ltd, Cramlington, UK). Highly expressing clones were identified by gelatine zymo- graphy [48] and dot-blot, using the anti-MMP-9 rabbit polyclonal serum from Sigma-Aldrich. One litre of culture supernatant from these cells was adjusted to pH 7.6 with Tris ⁄ HCl. Brij-35 and NaN 3 were added to 0.025% and 0.02% (w ⁄ v), respectively, and proMMP-9 was then puri- fied by affinity chromatography on gelatine-agarose (Sig- ma-Aldrich) according to [16]. The purified enzyme was quantitated using the absorbance at 280 nm combined with the predicted molar extinction coefficient, which was deter- mined using lasergene, Protean software (DNAStar, Inc, Madison, WI, USA). Recombinant TIMP-2 was purified by Ultrogel AcA44 gel filtration and heparin-Sepharose (GE Healthcare UK Limited, Chalfont St Giles, UK) chromatography from COS cell conditioned medium (supplied by British Biotech, Oxford, UK). Electrophoretic techniques Reduced or nonreduced (± 0.17 m 2-mercaptoethanol) sam- ples were heat-denatured (105 °C, 5 min) and analysed by SDS ⁄ PAGE on 10% gels, which were silver-stained using a Plus One TM kit (GE Healthcare UK Limited). Samples analysed by zymography were applied nonreduced and without heat denaturation to 10% gels copolymerized with 1mgÆmL )1 gelatine. Gels were renatured by two 10-min washes in 5% (v ⁄ v) Triton X-100 and digested for 15 h at 37 °C, followed by staining with Coomassie brilliant blue G-250. For western blotting, heat-denatured samples were run-reduced or nonreduced on 10% gels followed by transfer to nitrocellulose membranes (PROTRANÒ; Whatman plc, Brentford, UK). Blots were probed with primary antibodies as appropriate and horseradish peroxidase-conjugated secon- dary antibodies (Dako UK Ltd, Ely, Cambridgeshire, UK), followed by detection of immunoreactive bands by enhanced chemiluminescence (GE Healthcare UK Limited). Soluble collagen assay for collagenase activity Bovine skin type I collagen was prepared as previously described [49,50]; bovine type III collagen was a generous gift from V. Duance (Cardiff University, Cardiff, UK). Mutated type I collagen from murine skin was generously provided by S. Krane (Harvard Medical School and Massa- chusetts General Hospital, Boston, MA, USA). This type I collagen is completely resistant to collagenolytic cleavage, due to the mutation of Gln774 (P 2 ) and Ala777 (P¢ 2 ) of the a 1 (I) chain to proline. The wild-type a 2 (I) chain of each tri- ple helix is also resistant to cleavage, due to the presence of two mutated a 1 (I) chains [25]. Activity was measured using 27 lgÆsample )1 of collagen in 50 mm Tris ⁄ HCl, pH 7.6, 1 m glucose, 200 mm NaCl, 1mm CaCl 2 , 0.02% (w ⁄ v) NaN 3 . In some reactions, the collagen was denatured by incubation at 56 °C for 30 min prior to the assay. Samples were digested for 72 or 98 h at 25, 36 or 37 °C, followed by electrophoresis on 6.5% SDS ⁄ PAGE gels. Reactions were visualized by staining with Coomassie brilliant blue G-250. Protease inhibitor cocktail tablets containing serine and cysteine protease inhibitors (Roche Products Ltd, Welwyn Garden City, UK) were used at the manufacturer’s recommended working strength. Acknowledgements We thank Dr N. McKie for critically reviewing the manuscript. We are also grateful to Professor S. Krane MMP-9 activity against collagens I and III H. F. Bigg et al. 1252 FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS for generously providing the mutated type I collagen. This work was funded by the Arthritis Research Cam- paign, UK; The Health Foundation, UK; and the Dunhill Medical Trust, UK. References 1 Woessner JF Jr & Nagase H (2000) Matrix Metallopro- teinases and Timps (Sheterline, P, eds), Oxford Univer- sity Press, Oxford. 2 Welgus HG, Jeffrey JJ & Eisen AZ (1981) The collagen substrate specificity of human skin fibroblast collage- nase. J Biol Chem 256, 9511–9515. 3 Hasty KA, Jeffrey JJ, Hibbs MS & Welgus HG (1987) The collagen substrate specificity of human neutrophil collagenase. J Biol Chem 262, 10048–10052. 4 Freije JMP, Diez-Itza I, Balbin M, Sanchez LM, Blasco R, Tolvia J & Lopez-Otin C (1994) Molecular cloning and expression of collagenase-3, a novel human matrix metalloproteinase produced by breast carcinomas. J Biol Chem 269, 16766–16773. 5 Aimes RT & Quigley JP (1995) Matrix metalloprotei- nase-2 is an interstitial collagenase. J Biol Chem 270, 5872–5876. 6 Konttinen YT, Ceponis A, Takagi M, Ainola M, Sorsa T, Sutinen ME, Salo T, Ma J, Santavirta S & Seiki M (1998) New collagenolytic enzymes ⁄ cascade identified at the pannus-hard tissue junction in rheumatoid arthritis: destruction from above. Matrix Biol 17, 585–601. 7 Patterson ML, Atkinson SJ, Kna ¨ uper V & Murphy G (2001) Specific collagenolysis by gelatinase A, MMP-2, is determined by the hemopexin domain and not the fibronectin-like domain. FEBS Lett 503, 158–162. 8 Ohuchi E, Imai K, Fujii Y, Sato H, Seiki M & Okada Y (1997) Membrane type 1 matrix metalloproteinase digests interstitial collagens and other extracellular matrix macromolecules. J Biol Chem 272, 2446–2451. 9 Collier IE, Wilhelm SM, Eisen AZ, Marmer BL, Grant GA, Seltzer JL, Kronberger A, He C, Bauer EA & Goldberg GI (1988) H-ras oncogene-transformed human bronchial epithelial cells (TBE-1) secrete a single metalloprotease capable of degrading basement membrane collagen. J Biol Chem 263, 6579–6587. 10 Wilhelm SM, Collier IE, Marmer BL, Eisen AZ, Grant GA & Goldberg GI (1989) SV40-transformed human lung fibroblasts secrete a 92-kDa type IV collagenase which is identical to that secreted by normal human macrophages. J Biol Chem 264, 17213–17221. 11 Sopata I & Dancewicz AM (1974) Presence of a gelatin- specific proteinase and its latent form in human leuco- cytes. Biochim Biophys Acta 370, 510–523. 12 Ahrens D, Koch AE, Pope RM, Stein-Picarella M & Niedbala MJ (1996) Expression of matrix metalloprotei- nase 9 (96-kd gelatinase B) in human rheumatoid arthri- tis. Arthritis Rheum 39, 1576–1587. 13 Gruber BL, Sorbi D, French DL, Marchese MJ, Nuovo GJ, Kew RR & Arbeit LA (1996) Markedly elevated serum MMP-9 (gelatinase B) levels in rheumatoid arthritis: a potentially useful laboratory marker. Clin Immunol Immunopathol 78, 161–171. 14 Yoshihara Y, Nakamura H, Obata K, Yamada H, Hay- akawa T, Fujikawa K & Okada Y (2000) Matrix metal- loproteinases and tissue inhibitors of metalloproteinases in synovial fluids from patients with rheumatoid arthri- tis or osteoarthritis. Ann Rheum Dis 59, 455–461. 15 van den Steen PE, Proost P, Grillet B, Brand DD, Kang AH, van Damme J & Opdenakker G (2002) Cleavage of denatured natural collagen type II by neutrophil gela- tinase B reveals enzyme specificity, post-translational modifications in the substrate, and the formation of remnant epitopes in rheumatoid arthritis. FASEB J 16, 379–389. 16 Murphy G & Crabbe T (1995) Gelatinases A and B. Methods Enzymol 248, 470–484. 17 Bode W, Fernandez-Catalan C, Tschesche H, Grams F, Nagase H & Maskos K (1999) Structural properties of matrix metalloproteinases. Cell Mol Life Sci 55, 639–652. 18 Murphy G, Reynolds JJ, Bretz U & Baggiolini M (1982) Partial purification of collagenase and gelatinase from human polymorphonuclear leucocytes. Analysis of their actions on soluble and insoluble collagens. Biochem J 203, 209–221. 19 Kolkenbrock H, Ali HM, Hecker-Kia A, Buchlow G, Sorensen H, Hauer RW & Ulbrich N (1991) Characteri- zation of a gelatinase from human rheumatoid synovial fluid cells. Eur J Clin Chem Clin Biochem 29, 499–505. 20 Yasumitsu H, Miyazaki K, Umenishi F, Koshikawa N & Umeda M (1992) Comparison of extracellular matrix- degrading activities between 64-kDa and 90-kDa gelati- nases purified in inhibitor-free forms from human schwannoma cells. J Biochem 111, 74–80. 21 Okada Y, Gonoji Y, Naka K, Tomita K, Nakanishi I, Iwata K, Yamashita K & Hayakawa T (1992) Matrix metalloproteinase 9 (92-kDa gelatinase ⁄ type IV collage- nase) from HT 1080 human fibrosarcoma cells. Purifica- tion and activation of the precursor and enzymic properties. J Biol Chem 267, 21712–21719. 22 Watanabe H, Nakanishi I, Yamashita K, Hayakawa T & Okada Y (1993) Matrix metalloproteinase-9 (92 kDa gelatinase ⁄ type IV collagenase) from U937 monoblas- toid cells: correlation with cellular invasion. J Cell Sci 104, 991–999. 23 Pourmotabbed T, Solomon TL, Hasty KA & Mainardi CL (1994) Characteristics of 92 kDa type IV collagenase ⁄ gelatinase produced by granulocytic leuke- mia cells: structure, expression of cDNA in E. coli and enzymic properties. Biochim Biophys Acta 1204, 97–107. H. F. Bigg et al. MMP-9 activity against collagens I and III FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS 1253 24 Morodomi T, Ogata Y, Sasaguri Y, Morimatsu M & Nagase H (1992) Purification and characterization of matrix metalloproteinase 9 from U937 monocytic leu- kaemia and HT1080 fibrosarcoma cells. Biochem J 285, 603–611. 25 Wu H, Byrne MH, Stacey A, Goldring MB, Birkhead JR, Jaenisch R & Krane SM (1990) Generation of col- lagenase-resistant collagen by site-directed mutagenesis of murine pro1(I) collagen gene. Proc Natl Acad Sci USA 87, 5888–5892. 26 Miller EJ, Harris ED Jr, Chung E, Finch JE Jr, McCroskery PA & Butler WT (1976) Cleavage of type II and III collagens with mammalian collagenase: site of cleavage and primary structure at the NH 2 -terminal portion of the smaller fragment released from both collagens. Biochemistry 15, 787–792. 27 Miller EJ, Finch JE Jr, Chung E & Butler WT (1976) Specific cleavage of the native type III collagen molecule with trypsin. Similarity of the cleavage pro- ducts to collagenase-produced fragments and primary structure at the cleavage site. Arch Biochem Biophys 173, 631–637. 28 Clark IM & Cawston TE (1989) Fragments of human fibroblast collagenase. Purification and characterization. Biochem J 263, 201–206. 29 Murphy G, Allan JA, Willenbrock F, Cockett MI, O’Connell JP & Docherty AJP (1992) The role of the C-terminal domain in collagenase and stromelysin speci- ficity. J Biol Chem 267, 9612–9618. 30 Sanchez-Lopez R, Alexander CM, Behrendtsen O, Breathnach R & Werb Z (1993) Role of zinc-binding- and hemopexin domain-encoded sequences in the sub- strate specificity of collagenase and stromelysin-2 as revealed by chimeric proteins. J Biol Chem 268, 7238–7247. 31 Kna ¨ uper V, Osthues A, DeClerck YA, Langley KE, Bla ¨ ser J & Tschesche H (1993) Fragmentation of human polymorphonuclear-leucocyte collagenase. Biochem J 291, 847–854. 32 Hirose T, Patterson C, Pourmotabbed T, Mainardi CL & Hasty KA (1993) Structure–function relationship of human neutrophil collagenase: identification of regions responsible for substrate specificity and general protein- ase activity. Proc Natl Acad Sci USA 90, 2569–2573. 33 Kna ¨ uper V, Cowell S, Smith B, Lo ´ pez-Otin C, O’Shea M, Morris H, Zardi L & Murphy G (1997) The role of the C-terminal domain of human collagenase-3 (MMP- 13) in the activation of procollagenase-3, substrate spe- cificity, and tissue inhibitor of metalloproteinase interac- tion. J Biol Chem 272, 7608–7616. 34 Kna ¨ uper V, Lopez-Otin C, Smith B, Knight G & Mur- phy G (1996) Biochemical characterization of human collagenase-3. J Biol Chem 271, 1544–1550. 35 Tchetverikov I, Ronday HK, van El B, Kiers GH, Verzijl N, TeKoppele JM, Huizinga TWJ, DeGroot J & Hanemaaijer R (2004) MMP profile in paired serum and synovial fluid samples of patients with rheumatoid arthritis. Ann Rheum Dis 63, 881–883. 36 Stein-Picarella M, Ahrens D, Mase C, Golden H & Niedbala MJ (1994) Localization and characterization of matrix metalloproteinase-9 in experimental rat adju- vant arthritis. Ann NY Acad Sci 732, 484–485. 37 Itoh T, Matsuda H, Tanioka M, Kuwabara K, Itohara S & Suzuki R (2002) The role of matrix metalloprotei- nase-2 and matrix metalloproteinase-9 in antibody- induced arthritis. J Immunol 169, 2643–2647. 38 Okada Y, Naka K, Kawamura K, Matsumoto T, Nak- anishi I, Fujimoto N, Sato H & Seiki M (1995) Locali- zation of matrix metalloproteinase 9 (92-kilodalton gelatinase ⁄ type IV collagenase ¼ gelatinase B) in osteo- clasts: implications for bone resorption. Laboratory Invest 72, 311–322. 39 Grassi F, Cristino S, Toneguzzi S, Piacentini A, Facchini A & Lisignoli G (2004) CXCL12 chemokine up-regulates bone resorption and MMP-9 release by human osteoclasts: CXCL12 levels are increased in synovial and bone tissue of rheumatoid arthritis patients. J Cell Physiol 199 , 244–251. 40 Davis GE & Martin BM (1990) A latent M r 94,000 gelatin-degrading metalloprotease induced during differ- entiation of HL-60 promyelocytic leukemia cells: a member of the collagenase family of enzymes. Cancer Res 50, 1113–1120. 41 Garnero P, Borel O, Byrjalsen I, Ferreras M, Drake FH, McQueney MS, Foged NT, Delmas PD & Delaisse ´ J-M (1998) The collagenolytic activity of cathepsin K is unique among mammalian proteinases. J Biol Chem 273, 32347–32352. 42 Vu TH, Shipley JM, Bergers G, Berger JE, Helms JA, Hanahan D, Shapiro SD, Senior RM & Werb Z (1998) MMP-9 ⁄ gelatinase B is a key regulator of growth plate angiogenesis and apoptosis of hypertrophic chondro- cytes. Cell 93, 411–422. 43 Stickens D, Behonick DJ, Ortega N, Heyer B, Harten- stein B, Yu Y, Fosang AJ, Schorpp-Kistner M, Angel P & Werb Z (2004) Altered endochondral bone develop- ment in matrix metalloproteinase 13-deficient mice. Development 131, 5883–5895. 44 Bigg HF, Wait R, Rowan AD & Cawston TE (2006) The mammalian chitinase-like lectin, YKL-40, binds specifically to type I collagen and modulates the rate of type I collagen fibril formation. J Biol Chem 281, 21082–21095. 45 Zhang Y & Gray RD (1996) Characterization of folded, intermediate, and unfolded states of recombinant human interstitial collagenase. J Biol Chem 271 , 8015–8021. 46 Clark IM, Powell LK, Wright JK, Cawston TE & Hazleman BL (1992) Monoclonal antibodies against human fibroblast collagenase and the design of an MMP-9 activity against collagens I and III H. F. Bigg et al. 1254 FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS enzyme-linked immunosorbent assay to measure total collagenase. Matrix 12, 475–480. 47 Li J, Brick P, O’Hare MC, Skarzynski T, Lloyd LF, Curry VA, Clark IM, Bigg HF, Hazleman BL, Cawston TE et al. (1995) Structure of full-length porcine synovial collagenase reveals a C-terminal domain containing a calcium-linked, four-bladed beta-propeller. Structure 3, 541–549. 48 Woessner JF Jr (1995) Quantification of matrix metallo- proteinases in tissue samples. Methods Enzymol 248, 510–528. 49 Koshy PJT, Rowan AD, Life PF & Cawston TE (1999) 96-Well plate assays for measuring collagenase activity using 3 H-acetylated collagen. Anal Biochem 275, 202–207. 50 Cawston TE & Barrett AJ (1979) A rapid and reprodu- cible assay for collagenase using [1- 14 C]acetylated colla- gen. Anal Biochem 99, 340–345. H. F. Bigg et al. MMP-9 activity against collagens I and III FEBS Journal 274 (2007) 1246–1255 ª 2007 The Authors Journal compilation ª 2007 FEBS 1255 . like MMP-2, is able to cleave collagens I and III in their native form and in a manner that is char- acteristic of the unique collagenolytic activity of. Activity of matrix metalloproteinase-9 against native collagen types I and III Heather F. Bigg 1 , Andrew D. Rowan 1 , Michael D. Barker 2 and Tim E.