Proc. Natl. Acad. Sci. USA Vol. 90, pp. 2569-2573, April 1993 Biochemistry Structure-function relationship of human neutrophil : Identification of regions responsible for substrate specificity and general proteinase activity (/stromelysin/extracellular macromolecule) TOMOHIKO HIROSE*, CHRISTY PATTERSON*, TAYEBEH POURMOTABBEDt, CARLO L. MAINARDIt, AND KAREN A. HASTY*t Departments of *Anatomy and Neurobiology, and of tMedicine, University of Tennessee, Memphis, TN 38163 Communicated by Jerome Gross, December 16, 1992 (received for review September 10, 1992)

ABSTRACT The family of matrix metalloproteinases is a identity and 70% chemical similarity with human neutrophil family ofclosely related that play an inportant role in collagenase (MMP-8; NC). Preliminary data from other lab- physiological and pathological processes of matrix degrada- oratories have pointed toward the COOH-terminal domain as tion. The most distinctive characteristic of interstitial collage- a major determinant in substrate specificity. Windsor et al. nases (fibroblast and neutrophil ) is their ability to (16) have demonstrated that the cysteine residue immediately cleave interstitial coilagens at a single peptide bond; however, adjacent to the COOH terminus was essential for colla- the precise region of the responsible for this substrate genolytic activity of fibroblast collagenase (FC). Recently, specificity remains to be defined. To address this question, we Murphy et al. (17) have suggested that both COOH- and generated truncated mutants of neutrophil collagenase with NH2-terminal domains ofthe active enzyme contribute to the various deletions in the COOH-terminal domain and chimeric substrate specificity of collagenase and proposed a possible molecules between neutrophil collagenase and stromelysin and role of the COOH-terminal domain in substrate binding. Our assayed the expressed enzymes against type I and the data presented herein support the concept that the COOH- general substrate, casein. Our data suggest that substrate terminal domain of NC plays a major role in substrate specificity for interstitial collagen is determined by a 16-aa specificity. In addition, we have identified two loci in the sequence in the COOH-terminal domain of neutrophil colla- COOH-terminal domain, one governing collagenolytic activ- genase and is influenced by the integrity of a disulfide-defined ity and the other essential for general proteinase activity. loop at the COOH terminus for maximal activity. It was found that a relatively large region of 62-aa residues influenced the relative efficiency of collagenolytic activity. In addition to the MATERIALS AND METHODS region that conferred this specificity, a site at the COOH side Plasmid Construction ofTruncated Mutants ofNC (TrNCs). ofthe presumptive zinc-binding locus was found to be necessary The NC 7.2 cDNA containing a full-length coding region for for general catalytic activity. Mutation of a critical aspartic NC was used to create TrNCs with various deletions in the residue at position 253 within this area resulted in complete loss COOH-terminal sequence (6). The size of the TrNC is of proteolytic activity, suggesting that Asp-253 might function identified by amino acid residue numbers starting from the as one ofthe ligands for divalent cations, which are essential for initiating Met (Fig. 1). A premature stop codon was intro- enzymatic activity. duced by PCR. The primer (5'-GCTCGAATTCGGGC- TCGCCAGGGAAGGGCCCTACCC-3') complementary to The family of matrix metalloproteinases (MMPs) is a family the 5' end of NC 7.2 incorporated a unique EcoRI restriction of closely related enzymes that play an important role in a site and was used for construction of all the mutants. Primers variety ofphysiological and pathological processes, including at the 3' end contained sequences for a stop codon at various embryonic development (1), tumor invasion (2), and arthritis intervals and a unique Not I restriction site. The isolated (3, 4). The human MMP gene family contains at least two fragments were digested with EcoRI and Not I and then distinct interstitial collagenases (5, 6), three types of strome- ligated into these sites in the expression vector pcDNA I lysins (7-9), putative metalloproteinase 1 (10), and two ge- (Invitrogen, San Diego). The plasmid DNA was purified from latinases, 72-kDa type IV collagenase (11, 12) and 92-kDa transformed MC1061.P3 by CsCl gradient ultracentrifuga- type V collagenase (13, 14). When the primary structures of tion, and each PCR fragment was sequenced by the dideox- MMPs are compared, it is apparent that they are structurally ynucleotide chain termination method (Sequenase kit; homologous molecules consisting of defined functional do- United States Biochemical) to confirm that there was no error mains (13, 15). All of these enzymes contain an essential introduced by PCR. catalytic zinc-binding domain and an NH2-terminal domain, Plasmid Construction of Chimeric Enzymes. Junctional which functions to preserve the latent state of the enzyme. A sites between collagenase and Strom in the chimeras are variable COOH-terminal domain is present in all but putative identified by numbers referring to the last amino acid residue metalloproteinase 1. of NH2-terminal collagenase. A plasmid containing an insert Despite the high degree ofsimilarity among these enzymes, of cDNA for Strom (the kind gift of Constance E. Brincker- the MMPs studied to date are distinct with respect to their hoff, Dartmouth Medical School, Hanover, NH) was used to substrate specificity. Interstitial collagenases (MMP-1 and construct chimeric mutants between NC and Strom. A re- MMP-8) cleave native interstitial . By comparison, striction enzyme site shared by the two enzymes at the same human stromelysin 1 (MMP-3; abbreviated as Strom) does locus was used for splicing their cDNA fragments. For not degrade interstitial collagens although it shares 52% Abbreviations: MMP, matrix metalloproteinase; NC, human neu- The publication costs of this article were defrayed in part by page charge trophil collagenase; Strom, human stromelysin 1; TrNC, truncated payment. This article must therefore be hereby marked "advertisement" mutant of NC; FC, fibroblast collagenase. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed.

2569 Downloaded by guest on September 26, 2021 2570 Biochemistry: Hirose et al. Proc. Natl. Acad. Sci. USA 90 (1993)

Cys 91 Cys 279 Cys 464 demonstrated by casein zymograms as described, with the NC i (467 aa) -.-m-M c exception that casein (2 mg/ml) was incorporated into the polyacrylamide gel (4).

TrNC 464 i., RESULTS I Substrate Specificity of TrNCs. In our initial experiments, TrNC 463 111101111 we constructed TrNCs with various deletions in the COOH terminus and transfected these mutants into COS-7 cells. The molecular mass of wild-type NC and the mutants TrNC 464 TrNC 262 _ and TrNC 463, lacking three and four terminal amino acid residues, respectively, was 66 kDa under reducing condi- FIG. 1. Schematic representation of the truncated mutants of tions. In contrast, TrNC 262, which lacks the COOH- NC. The size ofthe TrNC is identified by amino acid residue numbers starting from the initiating Met. The black boxes represent the terminal half of the molecule, migrated at 51 kDa. All the presumptive zinc-binding region, and the Cys residues (Cys-91, mutants showed catalytic activity against the general sub- Cys-279, and Cys-464) are indicated by vertical bars. strate casein; however, significant reduction in collagenolytic activity was seen with progressive deletion of the COOH- chimera NC 273-Strom, a 3' Aha II restriction site in NC was terminal sequence. TrNC 262 lost all collagenase activity generated by PCR. although it retained proteinase activity against casein (data Mutant NC-Strom-NC, in which the amino acid sequence not shown). TrNC 464 and TrNC 463 preserved collagenase of NC at positions 259-274 was replaced with the corre- activity as well as caseinase activity. Control cells trans- sponding polypeptide of Strom, was generated as follows. fected with the pcDNA I vector alone did not show any The primer 5'-ATGCCTGACGCCCAAACCCTGTGAC- collagenase or caseinase activities and were negative by CCCAGTTTGACA-3' was designed to create an Aha II site, Western blot analysis. To compare the specific activity, and the reverse primer was complementary to the 3'-flanking mutants and wild-type collagenases were quantitated by region of NC 7.2 cDNA. The amplified cDNA fragments image analysis ofWestern blots (Fig. 2). The activity ofTrNC coding for the COOH-terminal collagenase were digested 464 was essentially identical to wild-type NC; in contrast, with Aha II and Sph I and ligated with the 83-bp Bsm I-Aha removal of the COOH-terminal cysteine (TrNC 463) resulted II Strom fragment and with the 838-bp EcoRI-Bsm I se- in a loss of nearly 62% of collagenase activity. quence in pUC19 coding for the NH2-terminal half of NC. Substrate Specificity of the Recombinant Chimeric En- Plasmid Construction for Site-Directed Mutants of NC. zymes. The diminished collagenolytic activity of TrNC 262 Site-directed mutagenesis of Asp-252 and Asp-253 to Gly or and TrNC 463 confirmed the importance of the COOH- Glu was performed using PCR with mutated oligonucleotide terminal region. To preserve the tertiary configuration of the primers and cut with EcoRI and Bsm I to give an 838-bp COOH terminus, we constructed chimeric enzymes by fusing fragment. The plasmid pUC19 containing NC 7.2 cDNA was the cDNA encoding the COOH-terminal region of Strom (which also contains a disulfide-defined COOH-terminal digested with EcoRI and Bsm I, and the excised 838-bp loop) to the NC 7.2 cDNA the fragment was replaced with the fragment encoding NH2- PCR-amplified fragments. terminal sequence. Equivalent amounts of the different chi- Transfection of the Mutant cDNAs into COS-7 Cells. The meras exhibited similar catalytic activity against casein with cDNAs subcloned in the expression vector pcDNA I were the exception of NC 258-Strom, which showed enhanced transiently transfected into COS-7 cells by the DEAE- caseinase activity as compared to wild-type collagenase. dextran/chloroquine-mediated procedure (18). In some ex- Furthermore, both exhibited the property of autocatalysis periments, to normalize the amount of synthesized enzymes, upon activation, which is characteristic of metalloprotein- the transfected cells were labeled with 500 ,uCi (1 Ci = 37 ases. The most striking difference between these chimeras GBq) of L-[35S]methionine (DuPont), and the intensity of the was in their ability to degrade native collagen; NC 326-Strom band corresponding to each enzyme was measured on auto- radiographs of SDS/polyacrylamide gels by using a comput- B er-based image analysis (Desktop SPARC, Sun Microsys- A tems, Milpitas, CA; Quantity One, Huntington Station, NY). Western Blot was as kDa Analysis. Immunoblotting performed 97 described (4) with specific antibodies [C44, raised against 66-- synthetic peptide MNFISLFWPSLPTGIQAAYEDFD within the COOH-terminal half of NC; anti-NC antibody, 45- generated against purified collagenase from human neutro- phil; and anti-Strom antibody (provided by C. E. Brincker- 1 2 3 45 ...... _ hoff, Dartmouth Medical School, Hanover, NH), raised against recombinant human Strom]. In some experiments, Western blots were scanned and analyzed with a computer- FIG. 2. Western blot analysis (A) and collagenase assay (B) of based image analysis to normalize the amount of expressed culture supernatants from transfected cells with truncated mutants of enzyme in further assays. NC. Lane 1, pcDNA I vector without an insert; lane 2, wild-type NC; Assays for Enzymatic Activity. Collagenase activity of lane 3, TrNC 464; lane 4, TrNC 463; lane 5, TrNC 262. (A) Western culture supernatants from transfected cells was analyzed by blot analysis probed by C44 antibodies under nonreducing condi- an assay with a soluble collagen substrate as described (19) tions. (B) Collagenase assay using soluble type I collagen as sub- and electrophoresed on 8% polyacrylamide gels to identify strate. The arrowheads indicate three-quarter-length degradation degradation products. Positive images of the gel on a fine products of al and a2 chins. Significant reduction in collagenase film were with a activity was seen with progressive deletion of the COOH terminus. grain positive (Kodak) analyzed computer- Note that Western blots probed by C44 antibodies were analyzed by based image analysis, and degradation was calculated by using a computer-based image analysis to normalize the amount of measuring the disappearance of the al chain, using various enzyme among wild-type NC, TrNC 464, and TrNC 463. TrNC 262 dilutions ofenzymes to ensure that substrate degradation was did not show any enzymatic activity against type I collagen, even linear. Casein-degrading activity in culture supernants was when increased amounts of the enzyme were assayed. Downloaded by guest on September 26, 2021 Biochemistry: Hirose et al. Proc. Natl. Acad. Sci. USA 90 (1993) 2571

A Collagenase polypeptide of Strom (boxed in Fig. 4). This replacement Activity resulted in loss of collagenase activity although substantial caseinase activity was preserved (data not shown). NC N_~ o-c 100 % Proteinase Activity of the Site-Directed Mutants of NC. Comparing the structure of the family of metailoproteinases, Strom I Z0 this region is flanked on the COOH side by the cysteine- defined loop and on the NH2 side by the highly conserved Bsm sequence DDXXGIXXXYG (positions 252-262). Our hy- NC 258- was that the residues in this Strom Z 0 pothesis aspartic region might Ball function by preserving a critical tertiary structure or by participating in divalent cation binding (calcium or zinc) NC 326- 7 16.1 (20-22). Site-directed mutagenesis of the cDNA clone of NC Strom 7.2 replacing Asp-253 with the nonpolar amino acid residue B Gly resulted in a loss of enzymatic activity against native collagen as well as casein (Table 1), although the protein was NC 273-Strom 0.23 present in the supernatants as shown by Western blot anal- Dra ysis. The conservative substitution of Glu for Asp at residue 253 (NC D253E) preserved 35.9o of coilagenase activity as NC 301-Strom Ml I 0.94 well as caseinase activity, suggesting that the COOH side chain of Asp-253 is critical to coordinate with metal ions such FIG. 3. Mosaic structure of the chimeric enzymes between NC as zinc and calcium. In contrast, site-directed mutagenesis and Strom. The shaded and open boxes, respectively, represent changing Asp-252 to Gly or Glu preserved enzymatic activity polypeptide regions of NC and Strom, and the black boxes indicate the presumptive zinc-binding region. Junctional sites between NC against collagen and casein. Interestingly no significant dif- and Strom in the chimeras are identified by numbers that refer to the ference was seen in reduction of caseinase activity between last amino acid residue of NH2-terminal collagenase, and the restric- NC D252E and NC D252G, although these mutations reduced tion enzyme sites are indicated. Collagenase activity was determined collagenase activity to 55.6% and 57.8%, respectively, com- relative to that of recombinant wild-type NC, and these values are pared with the wild-type NC. Thus, Asp-253 is critical in the given at the right. (A) First series of the experiments about chimeric expression of general proteinase activity, whereas the adja- enzymes. (B) Two additional chimeras constructed after we focused cent Asp-252 is not essential. our attention on the 62-aa sequence (at positions 259-320) critical for substrate specificity. DISCUSSION cleaved soluble collagen, whereas NC 258-Strom had no We have investigated substrate specificity using TrNCs and collagenase activity. The observation that NC 326-Strom chimeric enzymes between NC and Strom. The remarkable retained coilagenase activity, whereas NC 258-Strom did not sequence homology between NC and Strom allowed the suggested that a 68-aa sequence at positions 259-326 was generation of chimeras that maintained general conforma- critical for substrate specificity. tional integrity. Our data indicate that substrate specificity Since the COOH-terminal six residues of this intervening for interstitial collagen is defined within a 62-aa sequence in sequence are identical between NC and Strom, we thus the COOH-terminal domain of NC and is influenced by the focused our attention on the remaining 62 aa. Two additional integrity ofa disulfide-defined loop at the COOH terminus for chimeras whose junctional sites were located within the maximal activity. Interestingly, a highly conserved region region were constructed (Fig. 3) and exhibited similar cata- preceding the 62-residue sequence was found to be important lytic activity against casein (data not shown); however, for general catalytic activity. Mutation of a critical Asp at significant reduction in collagenase activity was seen with position 253 resulted in complete loss of proteolytic activity, progressive replacement ofthe region with the corresponding whereas the adjacent Asp-252 was found not to be essential. polypeptide of Strom. A comparative examination of this Our initial finding was that the deletion ofthe terminal Cys region revealed a striking difference between the colla- at position 464 substantially reduced collagenase activity, genolytic metalloproteinases (MMP-1 and MMP-8) and suggesting that a loop formed by the disulfide interaction of noncollagenolytic proteinase MMP-3 (Fig. 4). In Strom, this Cys-464 and Cys-279 influenced the degradation of native critical region contains a 9-aa insert. To test the hypothesis collagen. This is consistent with the previous observation by that the 9-aa insert and surrounding residues might affect Windsor et al. (16) that the site-directed mutagenesis of FC collagenase activity of the enzyme, a mutant NC-Strom-NC replacing Cys-259 or Cys-447 with Ser (their numbering starts was constructed by replacing the amino acid sequence of NC from the Phe after removal of the signal peptide) results in at positions 259-274 (16 residues) with the corresponding complete loss of substrate specificity against collagen, al- 252 I I 286 NC (MMP-8) DDIDGICAIYGLSSNPIO ...... PTGPSIPKPCDPSLTFD FC (MMP-1) DDIDGIQAIYGRSQNPVQ ...... PIGPQTPKACDSKLTFD Strom (MMP-3) DDINGIdsLyGPPPDSPETPLVPTEPVPPEPGrPANCDPALSFD 287 v 326T NC (MMP-8) AITTLRGEILFFKDRYFWRRHPQLQRVEMNFISLFWPSLP FC (MMP-1) AITTIRGEVMFFKDRFYMRTNPFYPEVELNFISVFWPQLP Strom (MMP-3) AVSTLRGEILIFKDRHFWRKSLRKLEPELHLISSFWPSLP FIG. 4. Comparison of amino acid sequences surrounding the region critical for substrate specificity against type I collagen. Numbers refer to the amino acid residues of NC starting from the initiating Met. Junctional sites between NC and Strom in the NC-Strom constructs (NC 258-Strom, NC 273-Strom, NC 301-Strom, and NC 326-Strom) are indicated by the arrowheads. The conservative Cys residues (at position 279 in NC) are in boldface type. In NC-Strom-NC chimera, 16 residues at positions 259-274 were replaced with the corresponding polypeptide of Strom (boxed). Downloaded by guest on September 26, 2021 2572 Biochemistry: Hirose et al. Proc. Natl. Acad. Sci. USA 90 (1993) Table 1. Enzymatic activity of the site-directed mutants of NC A B Collagenase Caseinase Enzyme Sequence activity, % activity Wild-type NC DDIDGIQ 100 + NC D252E EDIDGIQ 55.6 + NC D252G GDIDGIQ 57.8 + NC D253E DEIDGIQ 35.9 + NC D253G DGIDGIQ 0 - Substituted amino acid residues are indicated in boldface type.

though in our study the mutant TrNC 463 retained some collagenase activity. Further truncation of the molecule completely dissociated general catalytic activity from spec- ificity for type I collagen. TrNC 262 lost all collagenase FIG. 5. Models of NC for substrate recognition and for ligands to activity but retained caseinase activity, indicating that the divalent cations. The model represents activated enzyme after re- COOH-terminal domain, although essential for substrate moval of the propeptide domain, and the shaded bar indicates the 62-aa sequence, which is the critical region for substrate specificity. is not the specificity, essential for hydrolytic activity of The shaded circles represent triple-helical a chains of interstitial enzyme. collagen. It is possible that the region ofthe 62-aa sequence functions An alternative explanation for the above results is that the as a part of the collagen- with the COOH-terminal loop conformational integrity of the catalytic domain was altered stabilizing this conformation. There are two possibilities for the due to the absence of the cysteine loop. To eliminate this fourth ligand to the active-site zinc. (A) In this model His-217, possibility, we constructed chimeric enzymes by fusing His-221, and His-227 are coordinated to the zinc atom, and Asp-253 cDNA coding for the COOH-terminal half of Strom to the functions as a ligand to calcium, which is essential for enzymatic cDNA fragment of NC representing the NH2-terminal se- activity. (B) Asp-253 may function as the fourth ligand to the active-site zinc. quence. We observed that one of these chimeras had colla- genase the activity, whereas other did not, although both although other residues may be involved including the acidic mutants more could degrade the general substrate casein and amino acids aspartic and glutamic acid. showed autocatalysis upon aminophenylmercuric acetate In addition to zinc, MMPs also require calcium for their full treatment. Recently, Murphy et al. (17) used the same approach and found that neither chimera FC-Strom (corre- activity (33). However, unlike the interaction of MMPs with zinc, the residues involved in the interaction with calcium are sponding to our NC 258-Strom) nor Strom-FC had colla- unidentified. Our data document the importance of Asp-253 genase activity. In defining which amino acid residues are for general catalytic activity, which may suggest a role for critical for substrate specificity, we found that a relatively large region of the 62 residues affected the efficiency of this residue in the coordination with divalent cations. Two collagenase. The fact that NC 326-Strom showed only 16.1% possibilities for Asp-253 to serve as a ligand to a divalent collagenase activity as compared to wild-type enzyme indi- cation are shown in Fig. 5. Asp-253 may function as a ligand cates that the substrate-recognition site probably extends to to calcium, which is essential for maximal enzymatic activity, more distal COOH-terminal sequences. These data do not whereas His-227, His-217, and His-221 may be coordinated to exclude the role ofthe NH2 terminus in substrate recognition. the active-site zinc. A second, more remote possibility is that Interestingly, the replacement of the amino acid sequence in Asp-253 functions as the fourth ligand to the active-site zinc NC at positions 259-274 with the corresponding polypeptide along with His-217 and His-221. of Strom resulted in complete interruption of collagenase The techniques of cDNA cloning have led to the model of activity. Our data support the results of DeSouza and Bren- common functional domains within the individual enzymes of tani (23), who constructed inhibitory peptides based on their the MMP family. The refinement of these techniques have complementary hydropathy to the collagen cleavage site. The permitted the engineering of highly specific mutants to ex- hydropathy plot of their peptides is almost identical to the plore the role of structural motifs in the control ofidentifiable SSNPIQP sequence in the area of collagenase that we con- enzyme functions. Our data support the concept that the sider to be critical. domain(s) governing substrate specificity are distinct from It is clear that the MMPs are zinc metalloenzymes, and it those governing general catalytic activity. Further studies are is likely that the zinc moiety is catalytic in nature (24, 25). All likely to dissect more completely the relationships between of the cloned MMPs have been found to contain the metal- enzyme activity and protein structure. binding motif HEXXH, which is also shared by certain bacterial metalloproteinases (26). The two histidine residues The authors thank Dr. Constance E. Brinckerhofffor providing the in the HEXXH motif (located at 217-221 in NC) correspond Strom cDNA clone and the anti-Strom Ab and Virginia Ilardi and Michael J. Beauchamp for technical assistance. This work was to those of the zinc-binding site of (27), whose supported by grants from the National Institutes of Health (AI22603 x-ray structure has been elucidated (28). The third ligand is and AR39166) and the Arthritis Foundation. felt to be the conserved cysteine in the propeptide domain consistent with the attractive "cysteine switch" model in 1. Librach, C., Werb, Z., Fitzgerald, M., Chiu, K., Corwin, N., which the propeptide blocks the (29). The identity Esteves, R., Grobelny, D., Galardy, R., Damsky, C. & Fisher, of the fourth ligand of the active-site zinc has been more S. (1991) J. Cell Biol. 113, 437-449. speculative; however, the remarkable homology in HEXX- 2. Matrisian, L. M., Bowden, G. T., Krieg, P., Furstenberger, HXXGXXH sequence (located at residues 217-227 in NC) G., Briand, J.-P., Leroy, P. & Breathnach, R. (1986) Proc. seen throughout the MMP family and also seen in the other Natl. Acad. Sci. USA 83, 9413-9417. families of supports the third 3. Hasty, K., Reife, R., Kang, A. & Stuart, J. (1990) Arthritis (30-32) Rheum. 33, 388-397. histidine in this consensus sequence. The observation by 4. Hirose, T., Reife, R., Smith, G., Stevens, R., Mainardi, C. & Sanchez-Lopez et al. (15) that the site-directed mutagenesis Hasty, K. (1992) J. Rheumatol. 19, 593-599. of the third histidine of this consensus in rat transin resulted 5. Goldberg, G. I., Wilhelm, S. M., Kronberger, A., Bauer, in loss of proteinase activity supports the concept that this E. A., Grant, G. A. & Eisen, A. Z. (1986) J. Biol. Chem. 261, histidine functions as the fourth ligand to the active-site zinc, 6600-6605. Downloaded by guest on September 26, 2021 Biochemistry: Hirose et al. Proc. Natl. Acad. Sci. USA 90 (1993) 2573 6. Hasty, K., Pourmotabbed, T., Goldberg, G., Thompson, J., D. D., Seidman, J. G., Smith, J. A. & Struhl, K. (Wiley-Liss, Spinella, D., Stevens, R. & Mainardi, C. (1990) J. Biol. Chem. New York), pp. 16.13.1.-16.13.7. 265, 11421-11424. 19. Hasty, K. A., Hibbs, M. S., Kang, A. H. & Mainardi, C. L. 7. Whitham, S. E., Murphy, G., Angel, P., Rahmsdorf, H.-J., (1986) J. Biol. Chem. 261, 5645-5650. Smith, B. J., Lyons, A., Harris, T. J. R., Reynolds, J. J., 20. Ohlin, A.-K., Landes, G., Bourdon, P., Oppenheimer, C., Herrlich, P. & Docherty, A. J. P. (1986) Biochem. J. 240, Wydro, R. & Stenflo, J. (1988) J. Biol. Chem. 263, 19240-19248. 913-916. 21. Handford, P. A., Mayhew, M., Baron, M., Winship, P. R., 8. Saus, J., Quinones, S., Otani, Y., Nagase, H., Harris, E. D., Campbell, I. D. & Brownlee, G. G. (1991) Nature (London) Jr., & Kurkinen, M. (1988) J. Biol. Chem. 263, 6742-6745. 351, 164-167. 9. Basset, P., Bellocq, J., Wolf, C., Stoll, I., Hutin, P., Limacher, 22. Maslen, C. L., Corson, G. M., Maddox, B. K., Glanville, J., Podhajcer, O., Chenard, M., Rio, M. & Chambon, P. (1990) R. W. & Sakai, L. Y. (1991) Nature (London) 352, 334-337. Nature (London) 348, 699-704. 23. De Souza, S. J. & Brentani, R. (1992) J. Biol. Chem. 267, 10. Muller, D., Quantin, B., Gesnel, M.-C., Millon-Collard, R., 13763-13767. Abecassis, J. & Breathnach, R. (1988) Biochem. J. 253, 187- 24. Stricklin, G. P., Jeffrey, J. J., Roswit, W. T. & Eisen, A. Z. 192. (1983) Biochemistry 22, 61-68. 11. Collier, I., Wilhelm, S., Eisen, A., Marmer, B., Grant, G., 25. Birkedal-Hansen, H. & Taylor, R. E. (1982) Biochem. Biophys. Seltzer, J., Kronberger, A., He, C., Bauer, E. & Goldberg, G. Res. Commun. 107, 1173-1178. (1988) J. Biol. Chem. 263, 6579-6587. 26. Vallee, B. L. & Auld, D. S. (1990) Biochemistry 29, 5647-5659. 12. Huhtala, P., Chow, L. & Tryggvason, K. (1990) J. Biol. Chem. 27. Matthews, B. W., Weaver, L. H. & Kester, W. R. (1974) J. 265, 11077-11082. Biol. Chem. 249, 8030-8044. 13. Wilhelm, S., Collier, I., Marmer, B., Eisen, A., Grant, G. & 28. Matthews, B. W., Jansonius, J. N., Colman, P. M., Schoen- Goldberg, G. (1989) J. Biol. Chem. 264, 17213-17221. born, B. P. & Dupourque, D. (1972) Nature (London) New 14. Huhtala, P., Tuuttila, A., Chow, L., Lohi, J., Keski-Oja, J. & Biol. 238, 37-41. Tryggvason, K. (1991) J. Biol. Chem. 266, 16485-16490. 29. Van Wart, H. E. & Birkedal-Hansen, H. (1990) Proc. Natl. 15. Sanchez-Lopez, R., Nicholson, R., Gesnel, M.-C., Matrisian, Acad. Sci. USA 87, 5578-5582. L. M. & Breathnach, R. (1988) J. Biol. Chem. 263, 11892- 30. Dumermuth, E., Sterchi, E. E., Jiang, W., Wolz, R. L., Bond, 11899. J. S., Flannery, A. V. & Beynon, R. J. (1991) J. Biol. Chem. 16. Windsor, L., Birkedal-Hansen, H., Birkedal-Hansen, B. & 266, 21381-21385. Engler, J. (1991) Biochemistry 30, 641-647. 31. Takeya, H., Onikura, A., Nikai, T., Sugihara, H. & Iwanaga, 17. Murphy, G., Allan, J. A., Willenbrock, F., Cockett, M. I., S. (1990) J. Biochem. (Tokyo) 108, 711-714. O'Connell, J. P. & Docherty, A. J. P. (1992) J. Biol. Chem. 32. Delepelaire, P. & Wandersman, C. (1989) J. Biol. Chem. 264, 267, 9612-9618. 9083-9089. 18. Aruffo, A. (1987) in Current Protocol in Molecular Biology, 33. Seltzer, J. L., Welgus, H. G., Jeffrey, J. J. & Eisen, A. Z. eds. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, (1976) Arch. Biochem. Biophys. 173, 355-361. Downloaded by guest on September 26, 2021