Type X Collagen Contains Two Cleavage Sites for a Vertebrate ...

Download PDF
2 downloads 0 Views 4MB Size Report
Comment
Oct 22, 1985 - curic acetate (39), rather than proteolytic activation. Fourth, other collagenase preparations, including one from rabbit corneal cells and another ...
THE

JOURNAL

OF

BIOLOGICAL CHEMISTRY

Vol. 261, No. Issue 9, of March 25, pp. 4184-4189,1986 Printed in U.S.A.

0 1986 by The American Society of Biological Chemists, Inc

Type X Collagen Contains Two CleavageSites for a Vertebrate Collagenase* (Received for publication, October 22, 1985)

Thomas M. Schmidl, Richard Mayneg, JohnJ. Jeffreyff, and Thomas F. Linsenmayer 11 From the $Departmentof Biochemistry, Rush-Presbyterian-St. Luke’sMedical Center, Chicago, Illinois 60612, the §Department 35294, the IIDiuision of of Cell Biology and Anatomy, Universityof Alabama at Birmingham, Birmingham, Alabama Dermatology, Departments of Medicine and Biological Chemistry, Washington University Schoolof Medicine, St. Louis, Missouri63110, andthe IlDepartment of Anatomy and Cellular Biology, Tufts University Medical Schools, Boston, Massachusetts 0211 1, and the Developmental Biology Laboratory, Massachusetts General Hospital, Boston, Massachusetts 02114

Type X collagen was cleaved at twosites by a puri- (3-5). In the next proximal region (zone 3), the chondrocytes fied human skin collagenase. Two experimental aphypertrophy and are eventually removed as the advancing proaches were used to identify the locationcleavof themarrow cavity is enlarged. age sites. First, native type X collagen was digested From the hypertrophic chondrocytes of zone 3 (6, 7) we with the enzyme, and the rotary-shadowed products have isolated an unusual short collagen molecule, now desigwere visualized in the electron microscope. The major nated as type X collagen. The molecule is composed of three collagenase fragment of typeX contained the epitope identical chains with M, = 59,000. As visualized in theelectron recognizedbyamonoclonalantibody(X-AC9).The microscope after rotary shadowing, the molecule has a length antibody was usedas a point ofreference to locate the of approximately 138 nm with a globular domain a t one end position of the cleavage fragment within the native X of the triple helix (8). Type X collagen has a relatively high molecule. Second, the digestion of radiolabeled type collagen substrateswas analyzed by gel electrophore- thermal stability for a collagen molecule, with a midpoint sis. The complete cleavage of type X generated three denaturation temperature (T,’) of 47 “C. Strong noncovalent products with 32-, 18-,and 9-kDa chains. The 32-kDa interactions hold the chains of the globular domain together peptides were present in a triple-helical conformation(9). Type X wasshown to be a distinct collagen type as demonstrated by its CNBr peptides and amino acid compoand demonstrated a midpoint denaturation temperature of 43 O C in CD experiments. The18-kDa peptide sition (6, 10). Other investigators have isolated similar if not contained the tyrosine-rich globular domain of theidentical mol- molecules including a molecule, originally termed G ecule. The9-kDa peptide was derived from the triple-collagen (11, 12), produced by chick sternal chondrocytes helical end of the native molecule. TypeX collagen was cultured in collagen gels, and a molecule isolated from rabbit cleavedmorerapidlybythevertebratecollagenase (13,14) and bovine (15,16) growth plates. Within the 19-day than was typeI1 collagen inin vitro solution studies. chick sternum (17-19) the molecule was present in the cephalic half, which will ossify,but was absent from the caudal half, whichwill remain cartilaginous. Capasso et al. (20) observed that type X was synthesized by chick chondrocytes Long bones are the major supportive elements of the ap- from the tibiotarsus andconfirmed our previous observations pendicular skeleton. During embryonic development, these that the molecule is synthesized in specific regions of the initially form as cartilaginous anlagen upon which, and even- tissue (21). They also reported that under their cell culture tually within which, bone is deposited. The growth of long conditions, type X collagen was deposited in the cell layer bones is dependent on theformative processes of cell division (22) and rapidly converted to a fragment with chains of Mr = and macromolecular synthesis, balanced with a concomitant 30,000. The capacity to synthesize type X was lost when the remodeling of existing elements. This enables the organ to chondrocyte phenotype was modulated by virus transformaincrease in size while maintaining its overall form and struc- tion (23), bromodeoxyuridine, or clonal growth to senescence tural strength. The longitudinal growth of long bones predominantly oc- (24). We have recently produced a monoclonal antibody which curs at the epiphyseal growth plates (1).The chondrocytes recognizes native type X collagen. This reagent has allowed within these regions mature inan organized linear continuum us to screen many tissues and determine that type X is present (2). In the most distal region (zone l), the small round only in the skeletal system (25). The exact location of the chondrocytes proliferate. Proceeding proximally, the extent molecule has been determined in the developing tibiotarsus of cell division decreases and the cells acquire a flattened (26) andthe sternum (18), as well asthe vertebrae and morphology with their long axes perpendicular to thatof the notochord (27). In all these tissues the type X collagen was diaphysis. The flattened cells (zone 2) have the highest rates associated with the older hypertrophic chondrocytes which of collagenand chondroitin sulfateproteoglycan biosynthesis soon wouldberemoved. Thus, type X may facilitate the degradation of the extracellular matrix in which it is depos* This work was supported by the National Arthritis Foundation, the Illinois Chapter of the Arthritis Foundation, a Biomedical Re- ited. To pursue this possibility, we have examined a possible search support grant, and Grant A“3564 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The abbreviations used are: T,, midpoint denaturation temperature; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

4184

,

Collagenase Cleavage of Type X Collagen

mechanism for the degradation of the molecule. Since collagenase is the major enzyme responsible for the degradation of many types of collagen (28). we have examined the susceptibility of type X collagen to a vertebrate collagenase which was purified from human skin fibroblast cultures. We report here that, unlike other collagen molecules which are cleaved by the enzymeat a single locus, type X contains two cleavage sites for the enzyme. One cleavage site is very close to the epitope recognized by the X-AC9 antibody. In addition, type X appears to be degraded more rapidly than typeI1 collagen in the in vitro assays. EXPERIMENTALPROCEDURES

Collogen Isolation-"Ilpe X collagen was purified from mass chondrocyte cultures (9, 10) initiated from zones 2 and 3 of the chick embryonic tibiotarsus (5,29) aspreviously described. Two days after the initial plating, approximately 50% of the cells remained attached to the tissue culture dishes. Unlabeled type X collagen was purified from the harvested culture medium of the attached chondrocytes as previously described (10). Radiolabeled typesX and I1 collagens were isolated from the cultures of floating chondrocytes. The floating chondrocytes from the primary cultures were transferred to Petridishes and thesuspension cultures fed at weekly intervals. After 4 weeks of culture, the floating chondrocytes were pooled and split into two culture dishes. One dish, containing 2X lO'cells, was labeled with 0.5 mCi of ~-[2,3-'H]proline (23 Ci/mmol, Amersham Corp.) in 10 mlof medium. The medium was Dulbecco'smodified Eagle's containing 10% fetal calf serum (defined, Hyclone Laboratories), 50 pg/ml gentamycin, 50 pg/ml ascorbic acid, and 100 pg/ml 8-aminopropionitrilefumarate. The (43.3 Ci/ second dish was labeled with 0.5 mCi of ~-[3,5-~H]tyrosine mmol, Amersham Corp.) in 10 ml of medium as described above, but lacking tyrosine. The cells were labeled for 48 h and the culture medium collected. The collagen in the culture medium was precipitated twice with ammonium sulfate a t 30% of saturation. The precipitates were recovered by centrifugation for 30 min a t 10,OOO rpm and solubilized in 50 mM Tris, 0.4 M NaCI, 5 mM EDTA, 1 mM phenylmethanesulfonyl fluoride, and 1 mM N-ethylmaleimide at pH 7.5. The separation of type X from type I1 collagen was achieved by dialysis against increasing concentrations of sodium chloride in 0.5 M acetic acid. The type I1 collagen was recovered after precipitation against 0.7 M NaCI, while the type X was isolated when the NaCl concentration was raised from 1.2 to 2.0 M. Antibody Purification-The X-AC9 antibody is an IgCl molecule and was purified from spent hybridoma culture medium by chromatography on Protein A Sepharose (Bethesda Research Laboratories). A column (1 X 8 cm) containing 5 mlof Protein A Sepharose was equilibrated with 0.1 M sodium phosphate buffer, pH 8.2. Spent hybridoma medium was dialyzed against the same buffer and pumped over the column a t a rate of 10 ml/h. After the desired volume of medium was loaded, the column was washed with the pH 8.2 phosphate buffer until no absorbance a t 280 nm was detectable. The antibody was eluted in a single peak using 50 mM citric acid, 50 mM sodium phosphate buffer, pH 5.0. The tubes containing protein were pooled and dialyzed against 50 mM Tris. 0.2 M NaCl a t pH 7.5 and stored a t 4 "C. Collogenase Digestwns-Human skin collagenase was purified from human skin fibroblast cultures and purified according to the methods of Stricklin et al. (30). One hundred pl of enzyme, a t a concentration of 115 pg/ml, was activated by proteolytic treatment with 10 pg of trypsin (Sigma, diphenylcarbamyl chloride treated) for 10 min at 25 "C. The trypsin was neutralized with 40 pgof soybean trypsin inhibitor(Sigma) according to Stricklin et al. (30). The enzyme reactions received either 0, 0.5, 1, 2, or 4 pl of activated enzyme preparation. Ten thousand cpm of tritiated collagen substrate were added to each tube to give a total of 25 pl containing 10 mM CaCI2, 25 mM Tris, 0.2 M NaCI, pH 7.5. The digestions were performed a t 25 "Cfor 22 h and thereactions terminated by the addition of EDTA to 15 mM. Gel Electrophoresis-The collagenase digestion producta of the collagen preparations were separated by SDS-PAGE on a 12% polyacrylamide gel under reducing conditions (31). The gel was processed for fluorography (32) and exposed to a preflashed (33) x-ray film a t -70 'C. Rotary Shadowing-The products of a complete human skin col-

4185

lagenase digestion of type X collagen or preparations of unlabeled type X collagen weremixed with purified X-AC9 antibody and dialyzed against a solution of 0.2 M ammonium bicarbonate. The samples were sprayed onto freshly cleaved mica and rotary shadowed with platinum and carbon (8). Electron micrographs were taken on a JEOL-100CX electron microscope. The micrographs were measured with a Science Accessories Graf Pen 3 sonic digitizer interfaced with aHewlett-Packard 9825A programmable calculator as previously described. CD Spectropolarimetry-An unlabeled preparation of type X collagen was digested with the vertebrate collagenase at 37'C in the same digestion buffer described above. The extent of the reaction was followed by SDS-PAGE and the reaction terminated when the 32kDa peptide was the largest detectable product. This major digestion product was recoveredfrom the reaction mixture by precipitation with ammonium sulfate to 30% of saturation. The precipitate was solubilized in 50 mM Tris, 0.4 M NaCl and dialyzed against 5 mM sodium phosphate, 0.2 M NaF, pH 7.5. The thermal stability of the molecule's triple helix was analyzed in a Cary 60 spectropolarimeter. The temperature of the protein solution (100 pglml) was raised in a water-jacketed cuvette at a rate of 30 *C/h and the ellipticity monitored a t 222 nm as previously described (9). RESULTS

Rotary Shadowing of Tvpe X CoUagen-AntMy ComplexThe type X collagen used in these studies was isolated from the medium of chondrocyte cultures, thus circumventing the need for proteolytic extraction. As isolated from this source, the molecule is composed of three identical chainswith M,= 59,000 and contains two distinct domains. The collagen triple helix accounts for three-fourths of the molecular mass while the additional one-fourth of the molecule is inaglobular configuration. This structure is easily recognized in rotaryshadowed preparationsas shown in Fig. 1. We have previously described some of the characteristics of a monoclonal antibody (X-AC9) which binds to the triplehelical domain of type X collagen (25). To determine the location of the epitope to which X-AC9 binds, we have performed rotary-shadowing analyses of the antibody-antigen complex. Preparations of the antibody, purified on Protein A Sepharose, were incubated with typeX collagen. The mixture was rotary shadowed and visualized in the electron microscope. Electron micrographs of representative molecules are shown in Fig. 1. For comparison, we have included a native molecule without a bound antibody in theupper left corner of Fig. 1. As we have reported previously (8),the molecule consists of a threadlike helical domain with a dense knoblike

FIG. 1. Electron micrograph of rotary-shadowed preparations of native typeX collagen anda monoclonal antibody, XAC9. A native type X molecule is shown in the upper left corner of the electron micrograph. The other eight type X molecules have the X-AC9 antibody bound to a single site approximately 19 nm from the globular end of the molecule. Bar, 100 nm; magnification, X 125,000.

4186

Collagenase Cleavage of Type X Collagen

globular domain at one end of the molecule. Each of the remaining molecules shown in Fig. 1 has a single attached antibody which binds to an epitope located approximately 19 nmfrom the globular domain of the molecule. In rotaryshadowed preparations, the antibody appears to be larger and less dense than the terminal globular domain of the type X molecule itself. Of the 114 molecules examined, 89% of the type X molecules had an antibody associated with an equivalentsite nehr the globular domain. The location of the epitope as visualized in the rotary-shadowed complexesis consistent with our previous competition enzyme-linked immunosorbent assay experiments, which indicated that the epitope was conformation dependent and located within th. triple-helical domain of the molecule (25). Rotary Shodowing of Vertebrate Collagenase DigestedType X Collagen-Antibody Complexes-Previous data had shown that type X collagen is susceptible to vertebrate collagenase (6, 12). We have examined rotary-shadowed preparations to precisely identify the locations of the enzymatic cleavage site(s). The major collagenase cleavage product of type X collagen was a triple-helical fragment with chains of M , = 32,000 (6, 12). Preliminary data from enzyme-linked immunosorbent assay studies suggested that the 32-kDa fragment still contained the epitope recognized bythe antibody X-AC9. Therefore, we reasoned that rotary-shadowed preparations of the antibody bound to the products of collagenase digestion might allow direct visual assessment of the cleavage sit&). The results of such an experiment are shown in Fig. 2. The major collagenase cleavage product from a complete digestion of type X collagen has a length of 92 k 5 nm, a size whichfits well with that predicted for a helical fragment with 32-kDa chains. As can be seen in this figure, each of these fragments has a single antibody molecule bound close to one end. One hundred five molecules of similar length were examined and 102 had an antibody molecule bound to one end. Another common observation in such preparations is the cross-linking

FIG. 2. Electron micrograph of rotary-shadowed preparations of a vertebrate collagenase digestionof t y p e X collagen a n d the X-AC9 monoclonal antibody. After the digestion of type X collagen with human skin collagenase, the X-AC9 antibody was added and the mixture was rotary shadowed. The long arrows point tothe molecules that represent the major vertebrate collagenase cleavage products of type X. These molecules retain the epitope recognized by the X-AC9 antibody a t one end of the molecule. For comparison, native type X molecules with a bound antibody have been superimposed on the electron micrograph (outlined in the black rectangle) and aligned with respect to the bound antibody. The arrowhaads mark two X-AC9 molecules, each of which has crosslinked two collagenase cleavage products of type X collagen. Magnification, X 94,000.

-

1

2

3 4 6 8 7 8

"

8101112131416

-8 K

nc.3. SDS-PAGE of the humanskin collagenase digestion of type X a n d type I1 collagens. Each lane contained 1O.OOO cpm of collagen substrate. Lunes 1-5 contained [3H]proline-labeled type I1 collagen substrate; lanes 6-10 contained [3Hjproline-labeled type X collagen substrate; lanes 11-15 contained [3Hjtyrosine-labeled type X collagen substrate. The human skincollagenase was activated with trypsin, and all the assays were performed at 25 "C for 22 h. Lanes I, 6, and I 1 contained no enzyme; lanes 2, 7, and 12 contained 0.5 pl of enzyme; lanes 3,8, and 13 contained 1 p l of enzyme; lanes 4.9. and I4 contained 2 pl of enzyme; lanes 5, IO. and 15 contained 4 p l of enzyme. The vertebrate collagenase producta of type I1 collagen are marked on the left side of the gel, and the relative molecular masa of the collagenase prcducta of type X are shown on the r&t side of the gel.

of two of the 92-nm digestionproducta by one antibody molecule (demarcated by the arrowheads in Fig. 2): For purpose of comparison, two native molecules with a bound antibody (outlined by rectangles) have been superimposed in the electron micrograph shown inFig. 2. In each, the bound antibody was aligned withthe position of an antibody molecule on one of the cleavage products of type X. From this, it can be readily observedthat themajor cleavage product was excised fromthe centralregion of the type X triple helix. For this to occur requires at least two cleavages within the triple helix. In this set of experiments the overall length of the native molecule was 138f 5 nm. Therefore, one cleavage must occur a little less than 19 nm from the globular end of the molecule, whilethe second must take place approximately 27 nm from the triple-helical end of the molecule. SDS-PAGE Analyses of the Cleavage Process-The vertebrate collagenase of the type X collagen wasalso followed by SDS-PAGE and compared to the digestion of type I1 collagen. In order to follow the products of the enzymatic digestion, two different radiolabeled substrates were used. Chondrwere labeledwith [3H]prolinewhich would beenriched in the triple-helical domain or with ['Hltyrosine which is located primarily (90%) in the globular domain (10). A [3H]prolinelabeled preparation of type I1 collagen also was digested and its products, of known molecular weight,were used as standards. The three collagen substrates were digested under similar conditions using various concentrations of enzyme for22 h at 25 "C.The products of the digestions were separated by SDSPAGE as depicted in Fig. 3. The known molecular weightsof type I1 collagen and its collagenase producta (shown in lanes 1-5) wereused to calculate the sizes of the degradation Such cross-linking might be expected with the divalent potential of an antibody molecule; however, it is rarely seen in preparations of the antibody bound to the native type X molecule. In the native molecule, the additional 19 nm of helical structure and the attached globular domain may impartsteric hindrance which preventa the binding of two type X molecules to one X-AC9 antibody molecule.

Collagenase Cleavage of Type X Collagen products of type X (shown in lanes 6-15). Lanes 6-10 contain digestions of [3H]proline-labeledtype X collagen; lanes 11-15 show digestions of the [3H]tyrosine-labeledmaterial. The reaction products were identified on the basis of their enzyme susceptibility, relative content of proline and tyrosine, and their molecular weights. At the highest enzyme concentration, shown in lane 10,the digestion of proline-labeled type X was nearly complete. The major reaction products, which increased with increasing enzyme concentrations, had chains of32 kDa, 18 kDa, andthree closelyspaced peptides of approximately 9 kDa. The sum of molecular weights of the digestion products is very close to the size of the substrate chains (MI = 59,000). The 32-kDa fragment was the major product and also the most obvious product in the rotaryshadowed preparations. The 9-kDa peptides (lane 10) must be cleavedfrom the triple-helical end of the molecule, because it was labeled more strongly with [3H]proline (10) than was the 18-kDa peptide. Conversely, the 18-kDa chains (lane 15) probably contain the entire globular end of the molecule, since these products were more enriched in [3H]tyrosine than the other peptides of 9 kDa. With the use of lower enzyme concentrations, the original type X chains, digestion intermediates, and thefinal products of the vertebrate collagenase digestion were present (shown in lane 7). The largest intermediates in theenzyme digestion mixture had chains of 50 kDa. Other intermediates included a group of three peptides of approximately 41 kDa. The intermediates may be generated when the type X molecule was cleavedat only one of the two cleavagesites. The 50-kDa intermediate resulted from the removal of the 9-kDa fragment from the chains of the native molecule, while the 41-kDa peptides were generated by the loss of the 18-kDa fragment from the globular end of the molecule. The absence of the 41and 32-kDa products in theenzyme reaction containing [3H] tyrosine-labeled substrate (lanes 11-15) is consistent with these peptide assignments. The two products would have lost their tyrosine-rich globular region and would not be readily detectable. Several of the collagenase cleavage products appeared as doublets and triplets. We have previously observed similar multiple-band patterns in the native and pepsin-resistant forms of type X (6). This heterogeneity may be generated by small proteolytic cleavages of the two ends of the molecule during its purification. This interpretation is reinforced by the observation that different preparations of type X contained different amounts of such multiple bands. In Fig. 3, for example, proteolytic cleavage differences in the two type X substrate preparationswere evident in theregion of the 18kDa product. The [3H]proline-labeled preparation showed only one band in lanes 7-10, while the [3H]tyrosine-labeled preparation showed a doublet in the 18-kDa region of lanes

resulting collagenase products have a reduced thermal stability (28). To test if this is also true for type X collagen, we examined the thermal stability of the 32-kDa major collagenase product of type X collagen. A type X collagen preparation was completely digestedwith vertebrate collagenase, and the 32-kDa fragment was recovered by precipitation with 30% ammonium sulfate. The thermal stability of the fragment was analyzed by raising the sample to progressively higher temperatures, while monitoring the positive ellipticity of the sample at 222 nm in a spectropolarimeter. As shown in Fig. 4, the midpoint of the triple helix denaturation occurred at 43 "C.This T,,, is slightly less than thatof the native molecule, whose T, is 47 "C (9). DISCUSSION

Vertebrate collagenases are theprincipal enzymes involved in the degradation of collagen (34). Their mode of action, as first described by Gross and Lapiere (35), is to produce a single cleavage within the collagen triple helix. This reduces the thermal stability of the resulting cleavage products below the physiologicaltemperature of the organism and thusallows the removal of the denatured cleavage products by general proteolytic activity or by further gelatinolytic activity of the collagenase itself(36,37). We report here that purified human skin collagenase cleaves type X collagen in at least two sites within its triple-helical domain. This conclusion is based on two experimental approaches. The first was to visualize the collagenase products of type X in the electron microscope and compare these to native molecules. This was performed on rotary-shadowed preparations using a bound monoclonal antibody as a frame of reference. The second method was to follow the enzyme digestion of radiolabeled type X preparations by SDS-PAGE. These experimental approaches show type X collagen is cleaved at two sites within its triple helix. Similar results were obtained when only the pepsin-resistant domain was used as a substrate (data not shown). The enzyme digestions were carried out at 25 "C, approximately 22 "C below the denaturation temperature of the native molecule, thus ensuring the

X 60 80

-I

w

I

ae

40-

11-15.

The [3H]proline-labeledtype X and type I1 collagens were prepared from the same chondrocyte cultures, and, therefore, the relative susceptibility of the twocollagen types to a vertebrate collagenase may be compared under identical enzyme digestion conditions. We presume that thetwo collagen types were derived from the same intracellular proline pool; therefore, their specific activities should be the same. Equivalent amounts of radioactivity were added to each enzyme reaction mixture. By comparing the equivalent digestions in the corresponding lanes for each enzyme concentration (i.e. lanes 4 and 9),it can be seen that type X collagen was more rapidly degraded than was type 11. Vertebrate collagenase facilitates the degradation of types I, 11, and I11 collagens by cleaving the triple helix, and the

4187

20

-

I

I

I

I

38

41

44

47

TEMPERATURE ('(2) FIG. 4. Thermal denaturation of the major vertebrate collagenase cleavageproduct of type X collagen. After a digestion of type X collagen with human skin collagenase, the major 32-kDa collagenase product was isolated from the reaction mixture by ammonium sulfate precipitation. The molecule showed a CD spectrum with the characteristics of the collagen triple helix. The temperature of the protein solution was raised a t a rateof 30 "C/h and thepositive ellipticity monitored at 222 nm in a Cary 60 spectropolarimeter. The midpoint (T,)of the thermal denaturation curve was 43 "C.

Collagenase Cleavage of Type X Collagen

4188

integrity of the triple helix. It is unlikely that either of the cleavages is due to noncollagenase proteolytic activity within the digestion mixtures. First, the substrates themselves did not contain sufficient activity to degrade the type X collagen substrate during equivalent incubations. Second, the enzyme was a highly purified preparation (30, 38) showing no proteolytic activity against a variety of noncollagenous substrates. Third, thesame digestion products were obtained if the latent collagenase preparation was activated by aminophenyl mercuric acetate (39), rather than proteolytic activation. Fourth, other collagenase preparations, including one from rabbit corneal cells and another from human neutrophils, generated the same major digestion products from a type X collagen substrate. Although the experiments reported here demonstrate that multiple vertebrate collagenase cleavage sites exist within the triple-helical domain of type X collagen, we have not determined if the cleavage occurs between -Gly-Leu- or -Gly-Ileresidues ashas been shown for other collagenmolecules. However, we can speculate about the location of the vertebrate collagenase cleavage sites if our data are compared to recent sequence data derived from the analysis of the type X collagen gene (40). Assuming that the cleavage of type X is similar to other collagen types, the positions of the -Gly-Leu- and -GlyIle- bonds can be examined. In theregion of the cleavage site, near the amino-terminal end of the molecule, there are 5 possible cleavage positions. At the other cleavage site, near the carboxyl-terminal end of the triple helix, there are 3 possible cleavage positions. The most likely cleavage sites may be at the -Gly-Leu- bond, 92 residues from the amino terminus of the triple helix, and the-Gly-Ile-bond, 40 residues from the carboxyl terminus of the triple helix. Each of the bonds is in a -G$-X-Y- triplet which is separated by two additional amino acids from the next -Gly-X-Y- triplet. Besides theinterruption of the triple helix, the regions are deficient in imino acid residues which might decrease the stability of the triple helix (41, 42). Seqnence studies of the enzyme digestion products willbe necessary to determine exactly which bonds are cleaved. From the two sources of data, we have constructed a tentative model for the cleavage of type X collagen. This is schematically diagrammed in Fig. 5 and explained in the legend. The evidence presented here shows type X is cleaved in a

.

c

Acknowledgments-We thank Sandra H. Silvers and Michael H. Irwin for their technical assistance with the rotary shadowing and photography, Dr. Barbara Johnson-Wint for the rabbit corneal collagenase, Drs. David Schwartz and CharalamposArsenisfor the human neutrophil collagenase, and Drs. Jerome Gross and Klaus Kuettner for their helpful discussions.

I

X-AC9

9K

32K

unique fashion, and three-quarters of its helical domain is removed en bloc. Our preliminary data suggest that preparations of type X containcross-links in pepsin-sensitive regions of the molecule. Therefore, two collagenase cleavage sites may be necessary to remove the thermally stable centralregion of the type X triple helix. The sizes of the collagenase products reported here are interesting with respect to two recent reports of processing and turnover of type X collagen. One study reported that in cell and organ culture experiments the 59-kDa chains of type X were processed to 49-kDa chains (43). Although this size is close to the50-kDa collagenase intermediate of type X which we have observed, the investigators showed that thecleavage, which generated the 49-kDa form of the molecule, occurred within the globular domain of the molecule. Chondrocyte monolayers have been reported to degrade type X collagen resulting inthe generation of a 30-kDa product (22). However, this study reported that the 30-kDa material ". . .was not digested by invertebrate collagenase and most probably is a noncollagenous domain of .. ." type X. Since the nonhelical domain of type X has15-20-kDa chains, it is difficult to understand how 30-kDa chains couldbe derived from the molecule and not contain a triple-helical region which would besusceptible to bacterial collagenase. A more plausible explanation for this observation might be that the 30-kDa material was generated by vertebrate collagenase activity in the cultures and represents the thermally stable 32-kDa vertebrate collagenase product of type X. If this were the case, then the 30-kDa material should be derived from only the helical domain, should be sensitive to bacterial collagenase, but resistant to further cleavage by a vertebrate collagenase. Obviously further experimentation is required to resolve these potential discrepancies. Between days 10 and 19 of chick tibiotarsus development, type X collagen is located near the advancing boundary of the newly formed niarrow cavity (25, 26). Within a short period of time, the area of cartilage containing the type X molecule will beeroded to expand the cavity. The most likely structural function of the molecule in this region would be to supply a pattern for the removal of this cartilage. The susceptibility of the molecule to a vertebrate collagenase as shown here may be an integral aspect of this temporary pattern, especially considering the result that type X appearsto be more rapidly degraded by the enzyme than does type 11. Further experiments willbe necessary to identify the enzymes and cells which orchestrate this skeletal remodeling in uiuo.

18K

w"

FIG. 5. Schematic diagram of vertebrate collagenase digestion of type X collagen. The native type X collagen is approximately 138 nm in length. The molecule has a globular domain at one end of the molecule, and a monoclonal antibody, X-AC9, binds to the collagen triple helix about 19 nm from the globular end. Vertebrate collagenase cleaves the native molecule at two sites within its helical domain. The largest fragment is derived fromthe center of the helical domain. Its 32-kDa chains are 92 nm in length and contain the XAC9 epitope at one end. A smaller helical fragment, with 9-kDa chains, is released from the triple-helical end of the type X molecule. Another fragment, with 18-kDa chains, is released from the other end of the molecule and contains the entire globular domain of the molecule.

REFERENCES 1. Brighton, C. T. (1978) Clin. Orthop. Relut. Res. 1 3 6 , 22-32 2. Fell, H. B. (1925) J. Morphol. Physiol. 40,417-459 3. Oohira, A., Kimata, K., Suzuki, S., Takata, K., Suzuki, I., and Hoshino, M. (1974) J. Biol. Chem. 249,1637-1645 4. Kimata, K., Okayama, M., Oohira, A., and Suzuki, S. (1974) J . Biol. Chem. 249,1646-1653 5. Stocum, D. L., Davis, R. M., Leger, M., and Conrad, H. E. (1979) J. Embryol. Exp. Morphol. 5 4 , 155-170 6. Schmid, T. M., and Conrad, H. E. (1982) J. Biol. Chem. 2 5 7 , 12444-12450 7. Schmid, T. M., and Conrad, H. E. (1982) J. Biol. Chem. 2 5 7 , 12451-12457 8. Schmid, T. M., Mayne, R., Bruns, R. R., and Linsenmayer, T. F. (1984) J. Ultrastruct. Res. 86,186-191

Collagenase Cleavage of Type X Collagen T. M., and Linsenmayer, T. F. (1984) Biochemistry 23, 9. Schmid, 553-558 T. M., and Linsenmayer, T. F. (1983) J. Bill. Chem. 10. Schmid, 258,9504-9509 11. Gibson, G. J., Schor, S. L., and Grant, M. E. (1982) J. Cell Biol. 93,767-774 12. Gibson, G. J., Kielty, C. M., Garner, C., Schor, S. L., and Grant, M. E. (1983) Biochem. J. 211,417426 13. Remington, M. C., Bashey, R. I., Brighton, C. T., and Jimenez, S. A. (1983) Collagen Relat. Res. 3,271-278 14. Sussman, M. D., Ogle, R. C., and Bahan, G. (1984) J. Orthop. Res. 2,134-142 15. Grant, W. T., Sussman, M. D., and Balian, G. (1985) J. Biol. Chem. 260,3798-3803 16. Remington, M. C., Bashey, R. I., Brighton, C. T., and Jimenez, S. A. (1984) Biochem. J. 224,227-233 17. Gibson, G. J., Beaumont, B. W., and Flint, M. H. (1984) J. CeU Biol. 99,208-216 18. Schmid, T. M., Gibney, E., and Linsenmayer, T. F. (1986) Ann. N. Y. Acad. Sci., in press G. J., and Flint, M. H. (1985) J. Cell Biol. 101, 277-284 19. Gibson, 20. Capasso, O., Gionti, E., Pontarelli, G., Ambesi-Impiombato, F. S., Nitsch, L., Tajana, G., and Cancedda, R. (1982) Exp. Cell Res. 142,197-206 21. Capasso, O., Tajana, G., and Cancedda, R. (1984) Mol. Cell. Biol. 4,1163-1168 22. Capasso, O., Quarto, N., Descalzi-Cancedda, F., and Cancedda, R. (1984) EMBO J. 3, 823-827 23. Gionti, E., Capasso, O., and Csncedda, R. (1983) J. Biol. Chem. 258,7190-7194 24. Mayne, R., Elrod, B. W., Mayne, P. M., Sanderson, R. D., and Linsenmayer, T. F. (1984) Exp. Cell Res. 151,171-182 25. Schmid, T. M., and Linsenmayer, T. F. (1985) J. Cell Biol. 100,

598-605

4189

T. M., and Linsenmayer, T. F. (1985) Den Biol. 107, 26. Schmid, 373-381 27. Linsenmayer, T. F., Gibney, E., and Schmid, T. M. (1986) Deu. Biol., in press 28. Gross, J. (1981) in Cell Biology of the Extracellular Matrix (Hay, E., ed) pp. 217-258, Plenum Press, New York 29. Kim, J. J., and Conrad, H. E. (1977) J. Biol. Chem. 252,82928299 30. Stricklin, G. P., Bauer, E. A., Jeffrey, J. J., and Eisen, A. Z. (1977) Biochemistry 16,1607-1615 31. Laemmli, U. K. (1976) Nature 227,680-685 32. Bonner, W. M., and Laskey, R. A. (1974) Eur. J. Biochem. 46, 83-88 33. Laskey, R. A., and Mills, A. D. (1975) Eur. J. B&hem. 56,335341 34. Woolley, D. E., Taylor, D. J., and Yoffe, J. R. (1984) Life Chem. Rep. 2,181-235 35. Gross, J., and Lapiere, C. M. (1962) Proc. N&l. Acad. Sci. U. S. A. 48,1014-1022 D. E., Glanville, R. W., Roberts, D. R., and Evanson, J. 36. Woolley, M. (1978) B.&hem. J. 169.265-276 37. Web&, H: G., Jeffrey, J. J.,’ Stricklin, G. P., and Eisen, A. Z. (1982) J. Bill. Chem. 257,11534-11539 38. Stricklin, G. P., Eisen, A. Z., Bauer, E. A., and Jeffrey, J. J. (1978) Biochemistry 1'7,2331-2337 39. Stricklin, G, P., Jeffrey, J. J., Roswit, W. T., and Eisen, A. Z. (1983) Biochemistry 22,61-68 40. Ninomiya, Y., Olsen, B. R., van der Rest, M., Schmid, T. M., and Linsenmayer, T. F. (1985) J. Cell Biol. 101, 7a 41. Miller, E. J., Harris, E. D., Jr., Chung, E., Finch, J. E., Jr., McCroskery, P. A., and Butler, W. T. (1976) Biochemistry 15,

787-792

42. Highberger, J. Biophys. Res. 43. Kielty, C. M., Grant, M. E.

H., Corbett, C., and Gross, J. (1979) Commun. 89,202-208 Kwan, A. P. L., Holmes, D. F., Schor, (1985) B&hem. J. 27,545-554

Biochem. S. L., and