Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Characterization of the human gelatinase : a collagen binding domain Steffensen, Bjorn 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1997-251667.pdf [ 15.55MB ]
Metadata
JSON: 831-1.0088121.json
JSON-LD: 831-1.0088121-ld.json
RDF/XML (Pretty): 831-1.0088121-rdf.xml
RDF/JSON: 831-1.0088121-rdf.json
Turtle: 831-1.0088121-turtle.txt
N-Triples: 831-1.0088121-rdf-ntriples.txt
Original Record: 831-1.0088121-source.json
Full Text
831-1.0088121-fulltext.txt
Citation
831-1.0088121.ris

Full Text

CHARACTERIZATION OF THE HUMAN GELATINASE A COLLAGEN BINDING DOMAIN By BJORN STEFFENSEN D.D.S., The Royal Dental College of Copenhagen, 1978 M.S., The University of Michigan, 1986 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Oral Biology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA June 1997 © Bjorn Steffensen, 1997 In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representative. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Oral Biology The University of British Columbia Vancouver, Canada 11 ABSTRACT Matrix raetalloproteinases (MMPs) can collectively degrade most extracellular matrix components during normal tissue remodeling and have also been implicated in pathological inflammatory diseases and in tumor cell invasive growth. Exploring the MMP-ligand interactions is important in understanding the function of these enzymes. Gelatinase A, like the other MMPs, is composed of distinct functional domains. One domain unique to the gelatinases consists of three fibronectin type II-like modules. The type II modules in fibronectin provide this molecule with gelatin binding properties. Therefore, the hypothesis was formulated that the fibronectin type II-like modules also provide gelatinase A with ligand binding. The tri-modular recombinant collagen binding domain of human gelatinase A (rCBD123) expressed in E. coli bound native type I collagen as well as denatured types I, IV and V collagen, elastin, and heparin. All of the gelatinase A type I collagen binding properties were found to reside in the CBD. However, rCBD123 did not bind several substrates including native type V collagen or fibronectin. Although gelatinase A can degrade basement membrane, rCBD123 did not bind laminin, fibronectin, SPARC, or matrigel. Binding site analysis further revealed that rCBD123 can bind at least two collagen molecules simultaneously. Whereas the major binding site in native type I collagen is in the telopeptide ends, collagen denaturation exposes multiple binding sites. Lysine residues were found to be important molecular determinants for ligand interactions of the CBD. Acetylation of rCBD123 lysine residues abolished heparin binding and reduced binding to collagen. Site-specific substitution of rCBD123 lysines with alanines demonstrated that K357 in the third module is required for heparin binding. Unaltered binding to other ligands by K357A and no change from wildtype protein in secondary structural components, as assessed by circular dichroism spectral analysis, confirmed that the loss of heparin binding was not a result of structural perturbation. These results together with structure modeling of the gelatinase A CBD indicated that two or more modules are required for heparin binding. Another mutant, K263A, demonstrated reduced saturation level binding to collagen but with an unchanged Kd for the interaction pointing to the presence of more I l l than one collagen binding sites in the domain. Studies of mechanisms for gelatinase A cell surface localization showed specific cell binding to coated rCBD123 which was inhibited by preincubation of cells with soluble rCBD123. The cellular proteins binding rCBD123 were characteristic of collagen by electrophoretic behavior and resistance to digestion by pepsin but not bacterial collagenase. In addition, cell binding to rCBD123 was abolished by treatment of the cells with collagenase, and was reduced on a collagen binding deficient mutant of rCBD123. That rCBD123 could compete progelatinase A from cultured human gingival fibroblasts and that cell binding to rCBD123 was blocked by a Bl-integrin specific antibody point to the formation of a gelatinase A/ native type I collagen/ 61-integrin cell surface attachment complex. Thus, the CBD is an essential gelatinase A ligand binding domain. CBD lysine residues are important molecular determinants of substrate specificities and module cooperativity is likely required for the.ligand interactions. Gelatinase A can be positioned via the CBD to cell surfaces where it may be stored poised for activation and proteolysis. iv TABLE OF CONTENTS PAGE ABSTRACT ii TABLE OF CONTENTS iv LIST OF FIGURES vi LIST OF TABLES viii LIST OF ABBREVIATIONS ix PREFACE xi ACKNOWLEDGMENTS xiii DEDICATION xv CHAPTER 1 INTRODUCTION OVERVIEW 1 EXTRACELLULAR MATRIX COMPONENTS AND INTEGRLNS 2 Collagenous Components 2 General structure and biosynthesis 2 Type I collagen 5 Type V collagen 6 Type IV collagen 7 Non-collagenous Components 8 Heparin and heparan sulfate proteoglycans 8 Proteoglycans in general 8 Heparin and heparan sulfate 9 Extracellular heparin and heparan sulfate proteoglycan . . . . 10 Cell surface-associated heparan sulfate proteoglycans 11 Biological effects of heparin and heparan sulfate 12 Heparin and heparan sulfate protein binding 13 Fibronectin 15 Laminin 17 Elastin 18 Integrins 20 MATRIX METALLOPROTEINASES 25 Matrix Metalloproteinases in General 25 MMP Subgroups and Substrate Specificities 26 Collagenases 26 Gelatinases 29 Stromelysins, matrilysin, and matrix metalloelastase 32 Membrane type metalloproteinases 33 Other MMPs 34 V MMP Activation 35 Tissue Inhibitors of MMPs (TTMPs) 36 MMP Domains 37 MMP Genes 40 MMP Three-dimensional Structure 42 Collagen Binding Domain Three-dimensional Structure 43 Cell Membrane Localization and Activation of Progelatinase A 46 DOMAINS AND LIGAND INTERACTIONS 48 Protein Modules and Domains 48 Ligand Interactions by Fibronectin Type II-like Modules 49 Ligand Binding Domains and Functional Properties of MMPs 52 PROBLEM STATEMENT AND RATIONALE FOR THESIS RESEARCH 56 CHAPTER 2 Extracellular matrix binding properties of recombinant fibronectin type EE-like modules of human 72-kDa gelatinase/type IV collagenase Introduction 58 Experimental Procedures 60 Results 67 Discussion 76 Summary 82 CHAPTER 3 The Contribution of Human Gelatinase A Collagen Binding Domain Lysine Residues to Ligand Interactions Introduction 96 Experimental Procedures 99 Results - •• 104 Discussion 107 Summary 112 CHAPTER 4 Integrin Dependent Cell Binding of Human Gelatinase A by the Fibronectin Type II-like Modules Introduction 124 Experimental Procedures 126 Results 131 Discussion 135 Summary 140 CHAPTER 5 Concluding Discussion 153 BIBLIOGRAPHY 164 APPENDICES Appendix 1 196 Appendix 2 197 Appendix 3 198 Appendix 4 199 Appendix 5 200 Appendix 6 201 vi LIST OF FIGURES PAGE FIG. 1.1 Potential mechanisms of gelatinase A interaction with cell surface receptors and pericellular matrix molecules 24 FIG. 1.2 Domain structure of the major matrix metalloproteinase subgroups . . . . 38 FIG. 1.3 Stereo images of fibronectin type II-like modules 45 FIG. 1.4 Ligand binding properties of gelatinase A domains 53 FIG. 2.1. SDS-PAGE and Western blot analysis of non-reduced, reduced, and carboxymethylated rCBD123 85 FIG. 2.2. Denatured type I collagen affinity chromatography of rCBD123 86 FIG. 2.3. rCBD123 affinity for type I collagen in different conformations 87 FIG. 2.4. Interaction of rCBD123 with native and pepsin-treated type I collagen . 88 FIG. 2.5. Gelatinase A and rCBD123 show similar binding to native and denatured type I collagen 89 FIG. 2.6. Competitive inhibition of rCBD123 binding by native, pepsin-treated, and denatured type I collagen 90 FIG. 2.7. Simultaneous occupancy of rCBD123 binding sites by type I collagen molecules 92 FIG. 2.8. Interaction of rCBD123 with native and denatured types IV and V collagens 93 FIG. 2.9. rCBD123 binding to elastin 94 FIG. 2.10. Interaction of rCBD123 with basement membrane proteins 95 FIG. 3.1. SDS-PAGE and Western blot analysis of wildtype rCBD123 and mutant proteins 115 FIG. 3.2. Far UV circular dichroism spectra for wildtype and mutant rCBD 123 proteins 116 FIG. 3.3. Heparin-Sepharose affinity chromatography of rCBD123 and acetylated rCBD123 117 FIG. 3.4. Heparin affinity chromatography of rCBD123 lysine mutants 118 FIG. 3.5. Denatured type I collagen affinity chromatography of acetylated rCBD123 119 FIG. 3.6. Interactions of wildtype and mutant rCBD123 proteins with denatured type I collagen 120 V l l FIG. 3.7. Interactions of rCBD123 and mutant proteins with native and denatured type I collagen 121 FIG. 3.8. Binding of rCBD123 and mutant proteins to denatured types IV and V collagens 122 FIG. 3.9. Elastin binding properties of wildtype and mutants of rCBD123 . . . . 123 FIG. 4.1. Attachment of human fibroblasts to rCBD123 and fibronectin 142 FIG. 4.2. Competition of cell attachment to rCBD123-coated surfaces by soluble rCBD123 143 FIG. 4.3. Spreading of human fibroblasts on rCBD123 and fibronectin 144 FIG. 4.4. Morphological differences between cells cultured on rCBD123 and fibronectin 145 FIG. 4.5. Cell attachment to rCBD123 is inhibited by antibodies to the 61 integrin subunit 146 FIG. 4.6. Cell attachment to rCBD123-type I collagen complex 147 FIG. 4.7. Effect of bacterial collagenase treatment on cell attachment to rCBD123 148 FIG. 4.8. Ligand blotting of rCBD123 to detergent solubilized cellular components 149 FIG. 4.9. Cell attachment to rCBD123 mutants with reduced binding to type I collagen (K263A) or heparin (K357A) 150 FIG. 4.10. The role of rCBD123 heparin binding sites in cell attachment 151 FIG. 4.11. Competitive dissociation of gelatinase A from human fibroblast cell layers by rCBD123 152 APP. 1. Oligonucleotides used for incorporation of lysine to alanine codon changes in pGYMX123 196 APP. 2. Relative reactivity of wildtype and mutant proteins in micro well substrate binding assay 197 APP. 3. Cell number by crystal violet staining and Coulter counter 198 APP. 4. Absence of crystal violet non-specific staining of coated protein in cell attachment assays 199 APP. 5. Bacterial collagenase cleavage of native type I collagen 200 APP. 6. Bacterial collagenase does not cleave rCBD123 201 vm LIST OF TABLES PAGE Table 1.1 Substrate specificities of MMP family members 27 Table 2.1 Interaction of rCBD123 with extracellular matrix proteins 84 Table 3.1 Molecular mass determination by electrospray mass spectrometry . . . 114 IX L I S T O F A B B R E V I A T I O N S AP-1: Activation protein-1 BSA: Bovine serum albumin CBB: Coomassie Brilliant Blue CBD: Collagen binding domain CD: Circular dichroism C domain: Carboxyl-terminal domain CBD: Collagen binding domain CBX: Cyanogen bromide fragment X CNBr: Cyanogen bromide Con A: Concanavalin A DMSO: Dimethyl sulfoxide (DMSO) DSPG: Dermatan sulfate proteoglycan DTT: Dithiothreitol ECL: Enhanced chemiluminescence ELISA: Enzyme-linked immunosorbent assay FGF: Fibroblast growth factor G A G : Glycosaminoglycan Gelatin: Denatured collagen M E M : Minimal essential medium MMP: Matrix metalloproteinase MT-MMP: Membrane type metalloproteinase NMR: Nuclear magnetic resonance spectroscopy PA: Plasminogen activator PAGE: Polyacrylamide gel electrophoresis PBS: Phosphate-buffered saline r: Recombinant X rCBD123: Recombinant collagen-binding domain consisting of gelatinase A fibronectin type II-like modules 1, 2, 3 SDS: Sodium dodecyl sulfate SEM: Scanning electron microscopy SPARC: Secreted protein with is acidic and rich in cysteine TIMP: Tissue inhibitor of matrix metalloproteinase TPA: 12-O-tetradecanoylphorbol-13-acetate TRE: TPA response element xi PREFACE This thesis is organized according to the "Instructions for the Preparation of Graduate Theses" from the Faculty of Graduate Studies, University of British Columbia. Part of the results described in this thesis have been published during the course of the graduate studies and additional manuscripts have been submitted for publication as described below. This material has been adapted to the format of the rest of the thesis. Contributions made by collaborators and technical assistance provided are detailed in the Acknowledgement section. Publications from the thesis . Steffensen, B., Wallon, U.M., and Overall, CM.: Extracellular matrix binding properties of recombinant fibronectin type II-like modules of human 72-kDa gelatinase/type IV collagenase. High affinity to native type I collagen but not native type IV collagen. J. Biol. Chem. 1995; 270: 11555-11566. Steffensen, B. and Overall, CM.: Integrin Dependent Cell Binding of Human Gelatinase A by the Fibronectin Type II-like Modules. J. Biol. Chem. {submitted) Steffensen, B., Maurus, R., and Overall, CM.: The Contribution of Human Gelatinase A Collagen Binding Domain Lysine Residues to Ligand Interactions. Identification of K357 by Site-specific Mutagenesis as an Important Heparin Binding Residue. J. Biol. Chem. {submitted) Abstracts from the thesis Steffensen, B., Wallon, U.M., Overall, CM.: The extracellular matrix binding properties of recombinant fibronectin type II-like domain from 72-kDa gelatinase. Mol. Biol. Cell (Spec. Iss.) 5: Abstr. # 2517, 1994. Steffensen, B., Wallon, U.M., Overall, CM.: The extracellular matrix binding properties of recombinant fibronectin type II-like domain from 72-kDa gelatinase. Gordon Conference on Matrix Metalloproteinases, NH, 1995. Steffensen, B., Wallon, U.M., Overall, CM.: Recombinant 72-kDa gelatinase/type IV collagenase collagen-binding domain interactions with types I and V collagens and fibronectin. J. Dent. Res. (Spec. Iss.) 74: #1532, 1995. Steffensen, B., Overall, CM.: Analysis of extracellular matrix binding properties of the fibronectin-like domain of 72-kDa gelatinase by alanine scanning mutagenesis. American Association for Cancer Research, Special Conference on Proteases and Protease Inhibitors, Panama City, FL, 1996. Steffensen, B., Overall, CM.: Novel mode of 72-kDa gelatinase cell binding by the collagen binding domain. International Conference on Inhibitors of Metalloproteinases in Development and Disease, Banff, Alberta, Canada, 1996. Xlll ACKNOWLEDGEMENTS To persevere a Ph.D. program is not merely the accomplishment of the individual student, but also a reflection of the supportive environment that is required for a successful completion. I extend my sincere gratitude to the many people that have supported my efforts and whom I hope to maintain as friends and colleagues in the years to come. First, I would like to acknowledge my supervisor Chris Overall for accepting me into his laboratory. I have appreciated his guidance of my activities from my first entry into the graduate program at UBC up to this point of completion of my thesis. Over five exciting years, I have had the opportunity to develop scientifically under his watchful eye. Events have led us through periods of stress and hard work, but also plenty of joy. These experiences have created ties that I hope will remain strong as a prelude for future scientific interactions and a continued personal friendship for many years to come. My appreciation also goes to the members of my supervisory committee, Donald M. Brunette, Veli-Jukka Uitto, and Michel Roberge, and graduate student supervisor, Doug Waterfield, who all gently, but decisively, guided and supported my efforts to assure that I gain the most from the time and efforts invested. The main part of my time was spent in the laboratory. Fortunately, through challenging experiments and long days, I could always turn to an outstanding group of individuals with whom I shared problems, successes, and failures. As a result, I was not left behind in the rapidly developing fields of science, computers, and alternative music. My thanks go to Aldrich Ong, Angus MacQuibben, Angela King, Doug Sam, Edwin Rydberg, Eric Tarn, Gayle Pelman, Heather Bigg, John Edmeston, Ling Chen, Maggie Wallon, Ringo Leung, Robin Abbey, Tim Lau, and last, but not least, Yili Wong. Special thanks go to my fellow students in the Oral Biology and Dental School graduate programs with whom I shared the intricacies of maneuvering through a graduate program. They were there to listen, to support, and to share those accumulated experiences of the student body that make such a big difference for the individual: Alan Young, Amir Ashique, Anak Iamaroon, Ananda Nithyanand, Brent Hehn, Chris Peck, Dan Dhanawansa, xiv Ed Putnins, Elisabeth Matovinovic, Kirsi Haapasalmi, Mernaz Izadnegahdar, Monica Tonndorf, Sarah Hamadi, and senior "students" Laslo Ratkay and Carol Oakley. While awaiting construction of the laboratory in the Faculty of Dentistry, I had the opportunity to work in the laboratories of Michael Smith. I greatly appreciated this special experience which exposed me to research in the field of protein engineering and allowed me to spend a year with an exciting and helpful group with whom I share fond memories: Fred Kisil, Guy Guillemette, Heather Merrilee, Hung Lee, Jeannette Beatty, Lindsay Eltis, Louis Lefebvre, Marianne Huyer (in mem.), Steven Rafferty, Sakura Iwagami, and Yip Ho. Research is not an isolated event but requires multidisciplinary collaborations. I have benefited from interactions with specialists in multiple scientific disciplines, computers, equipment, photography, and administration at many levels. From a large group, I shall mention just a few: Alan Lowe, Andre Wong, Bruce MacCaughey, Christian Sperantia, Cindy Goundry, Ed Yen, Hannu Larjava, Ingrid Ellis, Jim Firth, Lari and Leeni Hakkinen, Lesley Ellies, Lesley Weston, Linda Skibo, Linda Gerow, Marc Sasso, and Robert Maurus. I enjoyed teaching in the Graduate Periodontics Program, where an exciting group of graduate students, staff, and colleagues kept me alert to the developments in the field. My time in this program has provided me with many fond memories. My appreciation goes to Andrea Lynch, Avi Schetritt, Cameron Jones, Colin Wiebe, David French, Donnel McDonnel, Edward Chesko, Hannu Larjava, Heidi Handley, Henry Louie, John Silver, Laura Noche, Lee Colfer, Nathalie Pauletto, Pretty Somaya, Shelly Putnins, and Sonia Leziy. My thesis research benefitted from direct collaborations. U. Margaretha Wallon contributed the results of ligand binding assays presented in figure 2.3, 2.9, and table 2.1. Robert Maurus performed circular dichroism spectral analyses presented in figure 4.2. Finally, I gratefully acknowledge financial support from the Medical Research Council of Canada (Studentship 1992-1995 and Dental Fellowship 1995-1997), the UBC Travel Fellowship, and the S. Wah Leung Scholarship from the UBC Faculty of Dentistry. This support has permitted me to undertake these studies and further cemented my belief in the great benefits of international exchange in education. X V DEDICATION I wish to dedicate this thesis to my wife Jane who encouraged me in my desire to return for these graduate studies. Jane, only with your endearing support and love has this undertaking been possible and so rewarding. My profound gratitude also goes to my parents, Ebba and Peter Steffensen, for their continued support and belief in my educational pursuits. Finally, I also thank our families on both sides of the Atlantic Ocean who have been of invaluable support whenever we needed them. CHAPTER 1 INTRODUCTION OVERVIEW Remodeling of the extracellular matrix is an integral part of maintaining the functional structure of healthy tissues as well as of development, repair, and wound healing. In pathological conditions, such as inflammatory diseases or tumor expansion and invasive growth, degradation of extracellular matrix components occurs in response to the disease process. The matrix metalloproteinases (MMPs) are important in these processes and their activity is controlled by specific tissue inhibitors of metalloproteinases (TIMPs). The MMPs, which are characterized by a shared primary and tertiary structure, collectively can degrade most extracellular matrix components, yet vary significantly in substrate recognition and catalytic properties. The substrate specificities of individual MMPs reside in the distinct modules or domains that bind the substrates. Therefore, investigating the molecular determinants for ligand interactions between MMPs and extracellular matrix components is a fundamental components of gaining an understanding of enzyme function. The ligand interactions of the collagen binding domain (CBD) of gelatinase A (MMP2; 72-kDa gelatinase/type IV collagenase) with known substrates and other extracellular matrix proteins have been the focus of this thesis. The results of the experiments are presented in three sections: Characterization of the gelatinase A CBD ligand interactions (Chapter 2); Mutational analysis of contributions of lysine residues to ligand binding interactions (Chapter 3); and Cell surface localization of gelatinase A mediated by the CBD (Chapter 4). Each of these sections contains an introduction and a discussion focused on the specific research questions addressed. In addition, a Concluding discussion is provided in Chapter 5. As a prelude to this work, a review of current knowledge related to aspects of MMP interactions with extracellular matrix components is provided (Chapter 1). This Introduction has three major sections which introduce first the extracellular matrix components studied, then the family of MMPs, and finally aspects of protein domains and ligand interactions 1 Introduction pertinent to the present thesis. The extracellular matrix components reviewed are those studied in the first major part of the research (Chapter 2). In addition, attention to the known aspects of heparin and heparan sulfate proteoglycan protein interactions has been given due to the importance of lysine residues in the gelatinase A CBD heparin binding (Chapter 3). The integrin cell surface receptors are reviewed to provide a context for the results presented in Chapter 4 that revealed an important contribution of this class of receptors in binding gelatinase A to cell surfaces. The second introductory section on MMPs provides a general review of the major subgroups of the MMPs, their structure, activation mechanisms, and TIMPs. Substrate specificities of the individual enzymes are presented along with a discussion of the degradation characteristics that focuses on the gelatinases and the extracellular matrix components investigated in this thesis. Since homologous modules and domains are found in different proteins where they may provide characteristic and similar ligand binding properties, section three of the Introduction contains an overview of general aspects of functions of modules and domains with a more detailed consideration of the fibronectin type II-like protein modules that form the collagen binding domain of gelatinase A. The MMPs are good examples of how the contribution of distinct ligand binding domains may contribute to the overall function of the enzymes. Therefore, this section ends by describing the current status of the knowledge regarding the functional contributions of MMP domains to the function of the enzymes. Based on the review of the literature outlined, the Introduction concludes by presenting the Problem Statement and Rationale for Thesis Research. EXTRACELLULAR MATRIX COMPONENTS AND INTEGRINS COLLAGENOUS COMPONENTS General structure and biosynthesis Collagenous proteins are major constituents of all extracellular matrices. The 2 Introduction collagens compose a heterogenous class of different molecules that have previously been considered primarily of structural significance. However, it has become evident that collagens may also be involved, directly or indirectly, in cell attachment and differentiation, as chemotactic agents, as antigens in immunopathological conditions, and as hereditarily defective components that cause certain pathological conditions (Olsen, 1995; Linsenmayer, 1991). As an example of the latter, four mutations in the triple helical domain of fibrillar collagens have been described which are associated with osteogenesis imperfecta, Ehlers Danlos syndrome, and chondroplasia (Olsen, 1995; Vuorio and De Crombrugghe, 1990). The number of known collagen types to date is twenty and 37 collagen a-chains have been identified (Pihlajaniemi, 1995). Recently, the human types XVII, XVIII, and XIX collagens were cloned (Gatalica et al, 1997; Inoguchi et al, 1995; Oh et al, 1994) and it was determined that one type of epidermolysis bullosa is associated with specific mutations in this type XVII collagen gene (Gatalica et al, 1997). Collagens are frequently classified as "fibrillar collagens" (types I, II, III, V, XI) that form banded fibrils in tissues and non-fibrillar collagens. The non-fibrillar collagens are diverse in structure and include fibril-associated collagens with interrupted triple helices (FACITS) (types DC, XII, XIV, XVI, XIX) and collagens that form sheets or membranes (types IV, VIII, X) (Pihlajaniemi, 1995; Prockop and Kivirikko, 1995; Linsenmayer, 1991; Olsen, 1991). Unique structural features distinguish all these collagens from other molecules. Three polypeptide chains, a-chains, are left-handed poly-L-proline helices that assemble into a tight triple helix characterized by a right-handed twist. The collagens may be homotrimers or heterotrimers made up of identical or different a-chains, respectively (Piez, 1976). Examples of such a-chain combinations are the type I collagen which is composed of two al(I) and one a2(I) chains (al(I)2a2(I)) and types II and III collagens which are homotrimers containing three al(II)-chains (al(II)3) and three al(ffl)-chains (al(III)3), respectively. The helices are stabilized by interchain hydrogen bonds. These bonds are disrupted during thermal or chemical denaturation, resulting in unfolding of the triple helix. However, the 3 Introduction triple helix conformation is also such that the peptide bonds linking the adjacent amino acids are buried in the interior of the molecule. Therefore, the triple helical regions of the collagens are highly resistant to cleavage by most proteases (Linsenmayer, 1991). Amino acid sequencing has shown that collagen oc-chains have a repeating Gly-X-Y amino acid sequence with the proportion of Gly being approximately one third of the total number of residues. Although the sequence and representation of different amino acids likely affect the requirements for triple helical conformation, fiber formation, and crosslinking, no reproducible repeating pattern is evident. The positions X and Y can be filled by any amino acid but X is most frequently a proline. In the Y-position, vertebrate collagens contain two unique residues, hydroxyproline and hydroxylysine (Piez, 1976). There are differences in the proportion of amino acids between the cc-chains. This is exemplified by the presence of Cys in the al(IEI) and al(IV) chains but not in the al(I) or a2(I) chains. Such variation in amino acid composition between different a-chains is one of the underlying determinants for differences between collagen types (Burgeson, 1982). Certain aspects of collagen biosynthesis are unique (Olsen, 1991). The mRNA transcripts are first translated to prepro-a-chains in the ribosomes of the rough endoplasmic reticulum. Signal peptides are cleaved from the prepro-a-chains as they are transported through the membrane. During translation, some, but not all, proline and lysine residues at the X and Y positions are oxidized to form hydroxyproline and hydroxylysine by hydroxylases (prolyl hydroxylase and lysyl hydroxylase; co-factor requirements include molecular 0 2 and ascorbic acid). Hydroxyproline is critical for helix stability. Under conditions where the conversion of proline to hydroxyproline is blocked, the melting point for the triple helical collagen is reduced from 39 °C to 25 °C rendering the molecule susceptible to denaturation and enzymatic degradation (Rosenbloom and Harsch, 1973). Hydroxylysines can, upon further modifications, form crosslinks between chains after fibril assembly (Kagan and Trackman, 1991; Siegel, 1979). After chain synthesis, an extensive cysteine-rich COOH-terminal globular domain, that 4 Introduction is an extension of the triple helix, folds by intrachain disulfide bonds. Subsequently, three a-chains associate by the COOH-terminal globular domains and the triple helix formation occurs toward the NH2-terminal end of the a-chains. These procollagen molecules are secreted in vesicles. During conversion from procollagen to fibrillar collagen in the extracellular environment, the NH2- and COOH-terminal globular propeptide segments are removed by procollagen peptidases leaving each collagen a-chain of the triple helical portion of the molecules with non-helical extensions, the telopeptide segments (Olsen, 1991). The length of these non-helical telopeptides varies between the fibrillar collagens but is generally very short consisting of 11 to 27 amino acid residues. The fully processed collagen molecules then aggregate to form fibrils that are stabilized by crosslinking by lysine and hydroxylysines residues (Kagan and Trackman, 1991; Narayan etal., 1974). While this series of events during biosynthesis is typical of fibrillar collagen types I, U, and in, the degree and mode of processing and crosslinking varies greatly between the other collagens. Type I collagen This fibrillar collagen is the most abundant type of collagen in many adult connective tissues such as skin, bone, dentin, root cementum and periodontal ligament, and cornea. Type I collagen molecules (al(I)2a2(I)) are about 3000 A long and have a triple helix diameter of about 15 A. Each a-chain is approximately 1000 amino acids long. Glycine accounts for one-third of the total number of amino acids, reflecting the repeating Gly-X-Y triplets. In addition, both chains have about half of their proline content as hydroxyproline and some of their lysine content as hydroxylysine (Linsenmayer, 1991). In the fiber arrangement, the staggered aggregation of molecules yields the characteristic banding pattern of the type I collagen fibers. In addition to important structural functions, it has been shown that cells can bind directly to type I collagen (Ruoslahti and Pierschbacher, 1987). This binding occurs primarily via the a ^ and a ^ integrin receptors (Elices and Hemler, 1989; Wayner and Carter, 1987) and a site in type I collagen which was first localized in the al(I)-cyanogen bromide peptide 5 Introduction 7 (CB7) (Kleinman et al., 1976). Since cell binding was lost after digestion with collagenase the binding site is close to the mammalian collagenase cleavage site (Kleinman et al., 1978). The region around the mammalian collagenase cleavage site, position Gly775-Ile776, is lacking imino acids (proline and hydroxyproline) that are known to stabilize the collagen triple helix (Ramachandran and Ramakrishan, 1976). The cell binding site on type I collagen has later been defined to the amino acid sequence Arg-Gly-Asp-Thr (Dedhar et al., 1987) in studies using osteosarcoma cells. This sequence is similar to the Arg-Gly-Asp-Ser cell binding region of fibronectin (Pierschbacher and Ruoslahti, 1984; Pierschbacher et al., 1983). In addition to cell adhesive properties, type I collagen has binding sites for fibronectin (Oldberg and Ruoslahti, 1982; Engvall and Ruoslahti, 1977). One fibronectin binding site is located near the mammalian collagenase cleavage site (in CB7), approximately two-thirds from the NH2-terminal end of the al(I)-chain (Kleinman et al., 1976), whereas a second fibronectin binding site has been localized to the a 1(1) CB12 which is near the NH2-terminal end of the molecule (Guidry et al., 1990). Type I collagen also binds small proteoglycans that have been associated with regulation of the migration of neural crest cells (Perris and Johansson, 1990), and the glycosaminoglycan heparin at a site near the NH2-terminus of collagen (San Antonio et al., 1994). The influence of small proteoglycans bound to type I collagen on neural crest cell migration is a potentially important biological feature in the context of development of many facial structures (Perris and Johansson, 1990). Type V collagen Type V collagen is also categorized as a fibrillar collagen. Molecules of this collagen type generally consist of two al(V) and one a2(V) chains, but also contain, in some cases, an a3(V) chain (Fessler and Fessler, 1987). The variation in chain composition suggests that this collagen type may have several members. Localization and structure studies suggest that in most tissues type V collagen occurs in heterotypic fibrils along with type I collagen. Based on results from in vitro fibrillogenesis studies it has been proposed that type V collagen may play a role in regulating the type I collagen fiber diameter (Birk et al, 1990; 6 Introduction Adachi and Hayashi, 1985). Both collagen types occur in the quarter-staggered array, but the precise structural relationship between the two is yet unknown. However, immunolocalization following digestion of tissue sections with interstitial collagenase, gelatinase, or both to preferentially degrade type I or type V collagen showed that the type V collagen fibrils were protected from cleavage unless the type I collagen fibrils were first degraded (Fitch et al., 1988). This indicates that type V collagen fibrils are partly buried within the type I collagen fibrils. Type IV collagen Type IV collagen is a non-fibrillar collagen which is a major component of basement membranes and the only collagen type demonstrated so far in this localization (Timpl et al, 1981). Epithelial and endothelial cells are the primary sources of this collagen (Peltonen et al, 1989; Martin et al, 1985). It is thought that type IV collagen may exist in both heteropolymeric and homopolymeric forms (al(TV)2oc2(IV), al(IV)3, and a2(rV)3). The genes that encode type IV are in a head-to-head arrangement separated by 130 bp of sequence that acts as a promoter for both collagen TV genes (Ramirez and di Liberto, 1990). Type IV collagen is secreted and assembled extracellularly as a procollagen-like molecule in which each chain has an apparent molecular mass of 160-180 kDa. The assembled molecule has three major domains: the NH2-terminal triple helical domain, the central primarily triple helical domain, and the COOH-terminal globular domain. Twenty-six non-helical regions have been identified that provide extensive structural flexibility (Linsenmayer, 1991). Type IV collagen extracted from tumors has the appearance of rods (length of -4000 A) with a knob-like COOH-terminal end. Tetramers form by crosslinkage between overlapping "7S" NH2-terminal triple helical regions of four type IV collagen molecules. In this overlap region, the molecules are highly disulfide crosslinked and also resistant to digestion with both bacterial collagenase and with pepsin. Each molecule may further associate with a second molecule by interaction between their knob-like "NCI" non-helical COOH-terminal domains (Timpl, 1989). This mode of aggregation allows formation 7 Introduction of large sheet-like structures (Linsenmayer, 1991; Yurchenco and Schittny, 1990). It has been proposed that additional lateral associations between type IV collagen molecules may occur in the basement membrane, allowing expansion into the third dimension (Yurchenco and Schittny, 1990). Type IV collagen supports adhesion by epithelial cells (Herbst et al, 1988; Terranova and Lyall, 1986) and is also critical to cell migration and differentiation during embryogenesis and angiogenesis (Herbst et al, 1988). In addition, type IV collagen has binding sites for the extracellular matrix molecules laminin, heparan sulfate proteoglycan, and fibronectin (Desjardins and Bendayan, 1989; Grant etal, 1989). Degradation of type IV collagen occurs during regular remodeling of basement membranes in health and development, and is a critical feature of basement membrane penetration during tumor cell invasive growth and metastasis. The matrix metalloproteinase gelatinase A (also termed 72-kDa gelatinase/type IV collagenase), which is synthesized by many cells types, has also been localized at high levels on tumor cells and can degrade type IV collagen (Emonard et al, 1992; Zucker et al, 1990; Collier et al, 1988; Liotta et al, 1980). However, although type IV collagen is a substrate of gelatinase A only denatured, and not native, type IV collagen was bound by the collagen binding domain of the enzyme (Chapter 2, Steffensen et al, 1995). Thus, type IV collagen is a molecule that has an important structural role and serves additional functions in interactions with extracellular matrix molecules and cells. NON-COLLAGENOUS COMPONENTS HEPARIN AND HEPARAN SULFATE PROTEOGLYCANS Proteoglycans: General Aspects Proteoglycans are important extracellular matrix and cell membrane associated components that contribute to the integrity and structure of the extracellular matrix and, in addition, are involved in the regulation of such cellular events as cell adhesion, migration, and matrix synthesis (Wight et al, 1992). Proteoglycans are comprised of a protein core to 8 Introduction which one or more glycosaminoglycan (GAG) side chains are covalently attached. GAGs are long repeating linear polysaccharides composed of variously combined or various combinations of specific disaccharides (usually one hexuronic acid together with N-acetylglucosamine or N-acetylgalactosamine which contain one or two sulfate residues each). Four different GAGs are found in connective tissues - 1) chondroitin sulfate, 2) dermatan sulfate, 3) keratan sulfate, 4) heparan sulfate, and heparin (Kjellen and Lindahl, 1991). Heparin and heparan sulfate are of special interest in the context of this thesis and will be reviewed in more detail in the following sections. Proteoglycans are synthesized in the Golgi apparatus with formation initiated at protein core serine residues yielding a trisaccharide (galactosyl-galactosyl-xylosyl) linkage that connects the core protein covalently with the polysaccharide. The polysaccharide chain is subsequently elongated and modified in a stepwise manner resulting in highly sulfated proteoglycans (Kjellen and Lindahl, 1991; Lindahl et al, 1986). The variation in sulfation, acetylation, and epimerization contributes greatly to the diversity between glycosaminoglycans. Further, one core protein may be linked to different types of GAGs and the GAGs may vary in different cell types. Heparin and heparan sulfate Heparin and heparan sulfate are two GAGs that are structurally very similar and interact with a variety of proteins and peptides (Jackson etal., 1991; Lindahl, 1989). Heparin and heparan sulfate are composed of different populations of the same monomeric building blocks (Lindahl, 1989); with one monosaccharide generally being an amino sugar of either D-glucosamine or galactosamine and the other generally uronic acid in either the D-glucuronic acid or iduronic acid form. Due to the presence of carboxyl and sulfate groups both are highly negative in charge. In fact, the main differences between heparin and heparan sulfate result from different levels of modification of the polysaccharide. Heparin is heavily O-sulfated compared to heparan sulfate (Gallagher and Waler, 1985). 9 Introduction Extracellular matrix heparin and heparan sulfate proteoglycan Heparin is produced by mast cells and basophil leukocytes. Commercially available heparin is a highly sulfated degradation product of the native proteoglycan. Heparin which is more sulfated than heparan sulfate contains a higher proportion of iduronic acid/glucuronic acid. In fact, heparin is the most acidic polysaccharide in nature and in the human body. As a result, heparin has the ability to bind many cationic proteins and, thus, affects a multitude of biological activities involving angiogenesis, complement function, and growth factor-induced responses (Putnins et al., 1995; Ruoslahti and Yamaguchi, 1991; Lindahl, 1989). Heparan sulfate is a prominent GAG of proteoglycans found in the extracellular matrix or as components of cell membranes and pericellular matrix. By its binding of hormones, growth factors, and other factors, this proteoglycan plays an active role in the regulation of cell functions (Kjellen and Lindahl, 1991). Among the heparan sulfate proteoglycans that are secreted, perlecan has a large core protein (-400 kDa) (Noonan et al., 1988) and is detected primarily in basement membranes, but it can also be found in the stroma of tumors (Iozzo and Murdoch, 1996; Iozzo et al, 1994; Haggerty et al., 1992). In connective tissues, perlecan is produced by and located in the immediate surroundings of fibroblasts (Heremans et al, 1989). Perlecan binds laminin as well as basic fibroblast growth factor (bFGF), both components of the basement membrane (Gonzales et al, 1990). It has been proposed that release of bFGF during wound healing or cancer invasion may contribute to local neovascularization (Folkman and Shing, 1992). In addition, the heparan sulfate side chains of perlecan have affinity for fibronectin, fibrillar collagens, and vitronectin (Gallagher, 1989) and may therefore serve in mediating binding between different extracellular matrix proteins. Among several heparan sulfate proteoglycans that have been described, but not yet named, is a large proteoglycan (core protein -400 kDa) which is secreted by fibroblasts into the extracellular matrix (Heremans et al, 1990). This large proteoglycan has a binding site for fibronectin that may anchor the proteoglycan in the extracellular matrix. Although only 10 Introduction a few heparan sulfate proteoglycans have been identified and characterized, it is likely that many are present in the extracellular matrix due to the large number of potential combinations of GAGs and core proteins (Larjava et al, 1992). Cell surface-associated heparan sulfate proteoglycans Cell surface-associated heparan sulfate proteoglycans associate with the cell membrane by modes of interaction that include direct intercalation of the core protein in the plasma membrane, anchorage to a phospholipid, as is the case for glypican, or association by heparan sulfate with other molecules on the cell membrane (Yanagishita and Hascall, 1992). Important integral cell surface proteoglycans that have heparan sulfate chains and also possess core protein transmembrane domains include CD44, betaglycan, and the family of syndecans that now encompasses at least four members (Elenius and Jalkanen, 1994). Syndecans have attracted special attention because of their ability to bind several extracellular matrix components including fibrillar collagens, the heparin binding domain of fibronectin, and tenascin (Uitto and Larjava, 1991; Elenius et al, 1990). Therefore, syndecans may serve as cell surface receptors for extracellular matrix proteins. During tooth development syndecan-1 may bind tenascin that is also highly expressed in the mesenchymal tissue components (Vainio et al, 1988). Since they localize in cell-cell contact areas, syndecans may have functions analogous to integrins in cell-cell and cell-matrix interactions (Uitto and Larjava, 1991). Like syndecans, CD44 undergoes extensive glycosylation during biosynthesis (Jalkanen et al, 1988). Both syndecans and CD44 have conserved cytoplasmic domains with potential phosphorylation sites (Jalkanen et al, 1991) and may therefore participate in transmembrane signal transduction. Widely distributed in multiple cell types, CD44 mediates cell adhesion by binding via its heparan sulfate side chains to fibronectin as well as to collagen, laminin, and hyaluronan (Jalkanen and Jalkanen, 1992; Carter and Wayner, 1988). Additional cell surface binding sites for exogenous heparin and heparan sulfate have been demonstrated. This ligand/receptor type interaction can be destroyed by trypsin suggesting that the binding occurs by means of non-collagenous proteins (Wilson etal, 1990; 11 Introduction Biswas, 1988). The affinity of heparin for these binding sites exceeded that of heparan sulfate even though the amount of bound material was equivalent in the absence of binding competition (Biswas, 1988). Although the GAGs are widespread in the tissues, increased levels of expression and altered distribution of CD44 and syndecan-1 have been observed during wound healing (Oksala et al, 1995; Elenius et al, 1991) where they, along with other membrane receptors, may exert a variety of cellular control effects such as adhesion, migration, and growth factor recognition. The difference in degree of sulfation may be important in controlling heparin and heparan sulfate binding to cells. Liver hepatocytes preferentially bind heparin-like polysaccharides with high sulfate content over heparan sulfate (Kjellen et al, 1980) and, in addition, removal of either the O- or N-sulfates significantly reduced the binding of heparin to mouse uterine epithelial cells (Wilson et al, 1990). That the similarity in structure may also be reflected in the functions of heparin and heparan sulfate is illustrated by the observation that heparan sulfate isolated from endothelial cell surfaces, similar to heparin, has anticoagulant activity likely mediated through "heparin-like" sequences (Kjellen and Lindahl, 1991). Therefore, these molecules are commonly studied in parallel when biological effects are assessed. Biological effects of heparin and heparan sulfate Additional biological effects have been associated with these proteoglycans. Heparin is widely used as an anticoagulant. It has the capacity to potentiate nearly 1,000-fold the inhibitory effect of antithrombin III, a member of the serine protease inhibitor (serpin) family which inhibits several blood coagulation proteinases (thrombin, factor LXa, Xa, XIa, and kallikrein) (Jackson et al, 1991). This effect has been attributed to a conformational change of antithrombin II that occurs during interaction with a pentasaccharide of heparin and produces a high affinity state of the inhibitor (van Boeckel et al, 1994). Heparin exerts various effects on cells including enhanced release of collagenase from organotypic cultures of murine bone explants (Sakamoto et al, 1973) and, more recently, heparin has been found 12 Introduction to enhance the autolytic activation of human progelatinase A by binding to the C O O H -terminal hemopexin-like domain (C domain) of the enzyme (Crabbe et al., 1993). Heparin may also induce production of extracellular matrix proteins although the response varies between cell lines as exemplified by an increased collagen synthesis observed in smooth muscle cell cultures, but a decreased synthesis in chondrocyte cultures (Brown and Balian, 1987; Majack and Bornstein, 1984). In addition to collagen, increased expression of fibronectin, thrombospondin, and two non-collagenous proteins of 37-39 kDa has been reported in response to treatment of smooth muscle cells with heparin or heparan sulfate (Lyons-Giordano et al, 1987; Cochran et al, 1985; Majack and Bornstein, 1984). All of these proteins have binding sites for heparin (San Antonio et al, 1994; San Antonio et al, 1993; Barkalow and Schwarzbauer, 1991; Jackson etal, 1991). Heparin has also been shown to have high affinity for, and to modify, the effects of fibroblast growth factor (FGF) (Putnins et al, 1995; Ruoslahti and Yamaguchi, 1991). This effect of heparin occurs in conjunction with the interaction between FGF and its cellular receptors. Thus, heparin binding of FGF serves as protection from proteolysis and is, in addition, required for FGF to induce signal transduction via the FGF receptor (Klagsbrun, 1990; Seno et al, 1990). Heparin and heparan sulfate protein binding With the broad range of biological effects observed for heparin and heparan sulfate there has been considerable interest and progress in understanding the mechanism by which they bind to proteins. Heparin binding regions have been determined in a number of proteins including the apolipoprotein B-100, apolipoprotein E, vitronectin, platelet factor 4, fibronectin, and others. By sequence alignment it was found that these regions contain clusters of positively charged residues separated by uncharged amino acids (Cardin and Weintraub, 1989; Cardin et al, 1986). Among the heparin binding regions, two amino acid consensus sequences have been identified which are characterized by [XBBXBX] and [XBBBXXBX] where "B" indicate basic residues and "X" hydropathic residues (Cardin and Weintraub, 1989). The predominant amino acids in the "B" positions are the positively charged Lys, 13 Introduction followed by Arg, and, less frequently, His. There is also a relative abundance of the uncharged Asn, Ser, Ala, He, and Leu (Cardin et al, 1991; Cardin and Weintraub, 1989). Subsequent studies showed that synthetic peptides of the apolipoprotein B-100 consensus sequence [XBBBXXBX] bound heparin at the same ionic strength as a CNBr heparin-binding fragment of apolipoprotein B-100 (Hirose et al, 1987). Successive deletions demonstrated that four residues [XBBB] corresponding to Lys-Arg-Lys-Arg were essential for the binding. An alternative approach to determining the amino acids required for heparin or heparan sulfate binding was provided by Caldwell et al. (1996) who characterized those 7-mer peptides from a random peptide library that bound these proteoglycans. The library was generated using the 20 common amino acids. Peptides that bound heparin-Sepharose were enriched in Arg, Lys, Gly, and Ser. In comparison, the peptides interacting with heparan sulfate-Sepharose generally bound at a lower ionic strength (0.15 M versus 0.30 M NaCl) and contained primarily Arg, Gly, Ser, and Pro. These results indicated a difference in the requirement for binding site residues between the two proteoglycans and also differed to a certain extent from the consensus sequences generated from sequence alignments of heparin binding sites. However, spacing requirements were not accounted for by this experimental approach (Caldwell et al, 1996). Helical wheel diagrams based on binding regions of several proteins suggested that the basic residues segregate primarily to one side of the helical face (Cardin and Weintraub, 1989). Additional three-dimensional modeling of the binding regions based on periodicity of the binding region residues suggested that a characteristic a-helix conformation provides the critical structure for interaction with heparin (Margalit et al, 1993; Cardin and Weintraub, 1989). According to this model, all basic residues face to one side of the a-helix and, in this conformation, accommodate binding of a glycosaminoglycan pentasaccharide. Verifying this structural binding motif, a synthetic 19-residue peptide, that folded to form an a-helix, displayed the predicted binding properties for heparin (Ferran et al, 1992). However, to account for heparin binding sites that do not form an a-helix or lack a consensus sequence 14 Introduction it has been proposed that two basic amino acid residues located at each end of the binding regions, and independent of the residues between them, are characterized by a reproducible secondary structural orientation and distance (-20 A) (Margalit et al., 1993). In contrast to apolipoprotein B-100, synthetic peptides from the sequence in the fibronectin type III13 module, that contains the heparin binding consensus sequence, lack binding to heparin (Ingham et al., 1993). However, three-dimensional modeling of the binding region based on the coordinates from a homologous region of tenascin revealed that two distant regions containing basic residues are brought into close proximity by the folding of the protein and are arranged in such a way that they together form a cationic heparin binding "cradle" (Busby et al, 1995). Single residue mutations of six basic amino acids (Arg, Lys) to uncharged residues (Ser) all reduced the heparin binding. Therefore, these results suggest that heparin binding to fibronectin, and possibly to other proteins, may require not only the consensus sequence, but also a properly folded structure. Alternatively, it is possible that even basic residues that do not conform to one of the consensus sequences can contribute to a heparin binding structural unit through protein folding. FIBRONECTIN Fibronectin is a high molecular weight homo-dimeric glycoprotein with the two subunits, each -220 kDa, connected by a disulfide bond. Soluble fibronectin is produced by hepatocytes and is found at a concentration of 0.3 mg/ml in plasma (Tamkun and Hynes, 1983) and as extracellular fibronectin synthesized by various cell types in situ to form insoluble pericellular and intercellular fibrillar networks. Fibronectin is also associated with basement membranes. Although fibronectin is encoded by a single gene (Prowse et al, 1986), alternative splicing of the primary transcript results in at least 20 human isoforms that vary in their biological properties (Schwarzbauer et al, 1989; Hynes, 1985). Fibronectin is characterized by a modular primary structure containing types I, II, and III homology units. In addition, the molecule contains three modules, EDA, EDB, and IIICS, that vary as a result of the alternative splicing. The expression of different isoforms has been 15 Introduction shown to be altered in response to development, aging, wound healing, and treatment with growth factors in cell culture (Steffensen et al, 1992; Magnuson et al, 1991; ffrench-Constant et al, 1989). The individual modules alone, or in combination, contain binding sites for a large number of extracellular matrix components including heparan sulfate proteoglycans (Bentley et al, 1985), fibrinogen (Stathakis and Mosesson, 1977), collagen (Klebe, 1974), and DNA (Siddiqa et al, 1989). In addition, a number of integrins bind to fibronectin via the RGD sequence in the type III-10 module, and the LDV or REDV in the IIICS module (Ruoslahti, 1996; Guan and Hynes, 1990; Mould et al, 1990). Fibronectin also has binding sites for gelatinase A. Whereas the binding was not located in the gelatinase A CBD {Chapter 2, Steffensen et al. 1995), two chymotryptic fragments of fibronectin both bound the rC domain of gelatinase A (Overall et al, 1997; Wallon and Overall, 1997). The possession of binding sites for extracellular and cell surface molecules is involved in the well-known ability of fibronectin to bind and organize extracellular matrix and to serve in the adhesive interactions of cells (Milam et al, 1991; Klebe, 1974). In cell cultures, fibronectin is a very potent attachment factor which promotes attachment, spreading, and migration of many cell types (Huttenlocher et al, 1995; Klebe, 1974). Fibroblasts and epithelial cells can migrate across fibronectin-coated surfaces (Uitto et al., 1992; Singer, 1979). The mechanism of binding between the cells and the fibronectin substrate involves cellular integrin receptors. Interestingly, fibronectin was left behind by migrating cells (Uitto et al, 1992). Moreover, fibronectin is believed to mediate indirect cell-matrix interactions through binding to collagen and proteoglycans (Ruoslahti, 1988; Dufour et al, 1986; Yamada, 1983). Of special interest to the research of this thesis are the collagen binding properties of fibronectin that have been localized to a segment of fibronectin that contains the only two type II modules of the molecule (Skorstengaard et al, 1994; Litvinovich et al, 1991; Banyai et al, 1990; Owens and Baralle, 1986; Balian et al, 1979). Fibronectin type D-like modules have been identified in a number of other proteins including the bovine seminal fluid proteins PDC-109 (Esch et al, 1983) and BSP-A3 (Seidah et al, 1987), the coagulation factor XII 16 Introduction (MacMullen and Fujikawa, 1985), the insulin-like growth factor II receptor/mannose 6-phosphate receptor (Lobel et al., 1987; Morgan et al., 1987), and the gelatinases A and B (Wilhelm et al., 1989; Collier et al, 1988). By protein engineering studies it has been found that recombinant (r) type II-like modules provide PDC-109 and gelatinase A and B with collagen binding properties (Steffensen et al, 1995; Banyai et al, 1994; Collier et al, 1992; Banyai and Patthy, 1991; Banyai et al, 1990). The ligand binding by fibronectin type II-like modules is discussed in more detail later in this review. LAMININ Laminin is a glycosylated protein with a high molecular mass (-890 kDa). The best characterized laminin form, laminin-1, is abundant in all basement membranes (Martin and Timpl, 1987) where it forms a dense network with type IV collagen, nidogen, proteoglycans, and other molecules of this structure (Yurchenco and Schittny, 1990; Leblond and Inoue, 1989). Beside the role of a structural component of basement membranes, many other functions have been associated with laminin including cell attachment for epithelial and endothelial cells (Kuhl et al, 1986), wound healing (Campbell and Terranova, 1988), bacterial adherence (Switalski et al, 1987), stimulation of neurite outgrowth, and signal transduction in mammary epithelial cells (Streuli et al, 1995; Edgar et al, 1988; Edgar et al, 1984). Laminin is composed of three subunits: an a (A) chain (-440 kDa), a 6 (BI) chain (-220 kDa), and a y (B2) chain (-230 kDa) (Burgeson et al, 1994). So far, eight different laminin chains have been identified which generate seven different heterotrimeric forms (laminins 1-7) (Burgeson et al, 1994; Timpl and Brown, 1994). These units are organized in a cruciform-like structure. Like fibronectin, laminin consists of functional domains that have affinity for cells and extracellular matrix molecules (Buck and Horwitz, 1987; Paulsson, 1987; Hakomori et al, 1984; Yamada, 1983). Studies with proteolytically generated fragments, synthetic peptides, and antibodies have aided in identification of the functional domains. Binding sites have been identified for type IV collagen (Charonis et al, 1985), 17 Introduction heparan sulfate proteoglycans (Laurie et al, 1986), heparin (Sakashita et al., 1980), nidogen (Paulsson et al., 1987), and integrins (Gehlsen et al, 1988; Sonnenberg et al, 1988). Even though there is an RGD sequence, it is YIGSR in the Bl chain and IKVAV in the A chain that promote cell adhesion (Tashiro et al, 1989; Graf et al, 1987). Laminin also exists in isoforms whose expression differs among species, tissues, and developmental stages of organism (Sanes et al, 1990). Laminin contains several epidermal growth factor-like sequences that may be released with proteolytic fragments from the matrix with the potential of exerting growth factor-like effects on surrounding cells (Panayotou et al, 1989). Yet, the regulatory effects of laminin on the behavior of cells are not fully understood. Laminin may affect epithelial cells with different origins differently. Thus, spreading and migration of epithelial cells from skin and from periodontal ligament (epithelial cell rests of Malassez) is inhibited by laminin, whereas laminin serves as a chemoattractant for gingival epithelial cells (Uitto et al, 1992; Woodley et al, 1988). In addition, the cell regulatory properties vary, i.e. laminin-5 is important for epithelial cell migration, whereas laminin-1 inhibits migration. ELASTIN Elastin is an extracellular matrix protein which is the predominant component of mature elastic fibers. In addition, the fibers contain other microfibrillar proteins, lysyl oxidase, and proteoglycans. Elastin provides the fibers with elasticity. Elastin's hydrophobicity and extensive crosslinking are the reasons why it is highly resistant to denaturation and proteolysis - to such an extent that successful purification from other extracellular matrix components relies on hot alkali and autoclave treatment to eliminate all other contaminating components. Tropoelastin is synthesized in soluble form, post-translationally modified, and transported to the cell surface where cross-linking and fiber assembly occur. A 67-kDa elastin-binding protein has been identified which co-localizes with intracellular tropoelastin and likely serves a protective role during secretion and extracellular assembly (Hinek and 18 Introduction Rabinovitch, 1994). Assembly of elastin fibrils occurs in infoldings of the cell surface (Serafini-Fracassini, 1984). Then fibrils organize to generate elastin fibers. Structurally, elastin fibers appear as densely packed, randomly arranged networks of fine filaments (Mecham and Heuser, 1990). The apparently random crosslinked network of elastin chains found in a recoiled condition provides the high degree of elasticity observed. Elastin has a highly characteristic amino acid composition; the content of acidic and basic amino residues is low whereas the content of hydrophobic amino acids, particularly valine (-15%), is high. One third of the residues are glycine and one ninth proline. Elastin contains no hydroxylysine, but small amounts of hydroxyproline are present (Mecham and Heuser, 1991). Initial studies of tryptic peptides predicted that elastin is made up of alternating hydrophobic and crosslinking domains (Gray et al, 1973). This has been confirmed by determining the amino acid sequence (primary structure) of elastin from several mammalian species, including human, that share extensive homology (Indik et al, 1987). In the crosslinking regions, lysine residues occur in pairs within clusters of alanines (Indik et al., 1990). Two lysine residues are always separated by two or three alanine residues. The resulting KAAK or KAAAK motifs are consistent with an a-helix conformation (3.6 residues per turn) where the polypeptide chain forms the inner part of the helix and lysine side chains extend outward in a helical array where they spatially are in close proximity to one another (Mecham and Heuser, 1991). In a multi-step process lysine side chains are modified to generate intra- and inter-monomer crosslinks called desmosine and isodesmosine that are unique to elastin (Gray, 1977). Thus, tropoelastin molecules are crosslinked in three dimensions. In contrast to collagen, post-translational hydroxylation of proline residues may not be important (Kao et al, 1982; Rosenbloom and Cywinsky, 1976; Uitto et al, 1976). Mammalian monocytes and fibroblasts were first shown to have elastin receptors (Senior et al, 1980). Subsequently, three cell surface elastin binding proteins have been described: a protein of 120 kDa (elastonectin) expressed in smooth muscle cells (Hornebeck et al, 1986), a 59 kDa cell membrane protein on tumor cells (Blood et al, 1988), and a 67-19 Introduction kDa protein found on many cell types and which may be involved with transport and elastin fiber assembly (Hinek and Rabinovitch, 1994; Hinek et al, 1988). Elastin contains no RGD sequence and there is no evidence of interaction with integrins. However, both the 59-kDa and 67-kDa elastin binding proteins bind a hexapeptide stretch, VGVARG, found repeated several times in elastin (Senior et al, 1984). Interestingly, both laminin and type IV collagen compete for binding to the 67-kDa receptor (Mecham et al, 1989; Senior et al, 1989). The turnover rate of elastin is slow with very little remodeling occurring in adult life. However, in pathological conditions elastases (serine proteases) released from inflammatory cells or bacteria may degrade elastin. Moreover, human neutrophils release elastinolytic enzymes including human leukocyte elastase (HLE), cathepsin G, proteinase 3, and macrophage metalloelastase (Shapiro et al, 1993; Senior et al, 1991). In addition, the gelatinases A and B have been shown to degrade elastin fibers (Senior et al, 1991). It has recently been demonstrated that elastin binds to the collagen binding domain of these two metalloproteinases {Chapter 2, Steffensen et al, 1995) and that CBD domain deletion mutants of the enzymes do not have elastolytic activity (Shipley et al, 1996). INTEGRINS Interactions of cells with extracellular matrix components and other cells are mediated by a class of integral cell membrane receptor molecules, the integrins (Hynes, 1992; Ruoslahti, 1991; Albelda and Buck, 1990). The interactions between integrins and ECM components are important for cell adhesion and migration (Huttenlocher et al, 1995), for determining cell behavior in tumor cell invasion of tissues (Dedhar, 1990; Ruoslahti and Giancotti, 1989), and for trophoblast differentiation during embryo implantation and placenta development (Damsky et al, 1994). Structurally, integrins are heterodimeric molecules composed of non-covalently associated a- and 6-subunits that have molecular weights of 140-190 kDa and 105-130 kDa, respectively, (Ruoslahti, 1991; Hemler, 1990). Both a and 8 subunits contain a long N-20 Introduction terminal extracellular domain, a transmembrane domain, and a shorter COOH-terminal cytoplasmic domain. The extracellular domain of the a-subunits has several divalent cation binding sites which promote the association of a and 6 subunits required to form ligand binding sites (D'Souza et al, 1990). Several a-subunits are composed of two proteolytically processed disulfide-linked chains, and an additional approximately 180 amino acid insert ('!"-domain which is potentially collagen binding) is observed in the a 1 and a2 subunits (Ignatius et al, 1990; Ruoslahti and Giancotti, 1989). The 6 subunit cytoplasmic domain contains potential tyrosine phosphorylation sites which may be involved in integrin-mediated transmembrane signaling events (Yamada and Miyamoto, 1995; Tamkun et al, 1986). Some integrins are found on many cell types while others are restricted to specific cell types. Generally, the BI-containing integrins are expressed on several cell types and bind extracellular matrix proteins (Hemler, 1990; Wayner et al, 1988). In contrast, the 62-integrins are found only on leukocytes and mediate cell-cell interactions that are important in the immune system (Springer, 1990). The combinations of the subunits affect both function and expression levels in different cells. When 63 is combined with the allb subunit to form allb63, it is expressed primarily on platelets. In contrast, the 63-integrin av63 is found on a broad range of cell types (Cheresh et al, 1989; Phillips et al, 1988). Thus, the flexibility of integrin subunit combinations and the potential for alternative splicing of the subunits adds an additional level of complexity to integrin biology (Hynes, 1992). At this time, eight different 6-subunits and sixteen a-subunits are known, that combine to form at least twenty-two a6-heterodimers. Among the integrins in fibroblasts, the 61 and 63 sub-families are prominent in mediating interactions with extracellular matrix molecules. The al61 and a261 integrins bind laminin and collagen. In addition to binding these two proteins, the a3Bl can bind fibronectin. Other integrins that interact with fibronectin include with a4Bl, avBl, avB3, and a561, the classical fibronectin receptor. The av63 is further active in interactions with fibrinogen, thrombospondin, vitronectin, von Willebrandt factor, and gelatinase A, and av65 21 Introduction is a receptor for vitronectin (Brooks et al, 1996; Hynes, 1992; Ruoslahti, 1991; Albelda and Buck, 1990). An important and the most prominent recognition motif in proteins that interact with integrins is the RGD sequence which is characterized by residues with positive-neutral-negative charges. At least eight and possibly twelve different integrins bind the RGD motif, which serves as an important cell binding site for several of the major extracellular matrix proteins including fibronectin, vitronectin, fibrinogen, von Willebrand's factor, thrombospondin, laminin, entactin, tenascin, osteopontin, and bone sialoprotein (Ruoslahti, 1996). However, only a minority of the large number of proteins that contain the RGD motif actually mediate cell attachment and interaction with integrins. This is likely due to lack of access to the RGD that may not always be surface exposed or may be inaccessible due to conformational restraints. Other peptides with affinity for integrins generally display specificity for individual receptors. For example, a461 binds the alternatively spliced CS-1 site of fibronectin at LVD (Guan and Hynes, 1990; Mould et al, 1990). Arg and Asp are considered important residues in recognition sequences as they are part of oc2M integrin binding sites both on laminin (YGYYGDALR) and on type IV collagen (FYFDLR) (Underwood et al, 1995), and possibly also the type I collagen triple helix (Eble et al, 1993). Integrins mediate cell adhesion which is a requirement for function and differentiation of most cells. The initial step in cell adhesion is recognition of the substrate by specific integrins (Hynes, 1987). This event is then followed by cell spreading during which the cytoskeletal actin filaments and ligand bound integrins are reorganized along with generation of focal contact points (Albelda and Buck, 1990). This reorganization and presence of distinct contacts between the intra- and extra-cellular environment set the stage for transmembrane signaling and controls gene expression (Clark and Brugge, 1995; Yamada and Miyamoto, 1995; Chen et al, 1992) leading to defined cell functions in such processes as growth, differentiation, and remodeling (Roskelley et al, 1995; Hynes, 1992). In fact, a 22 Introduction considerable number of potential integrin-mediated pathways associated with cytoskeletal and cell adhesion changes have been described (Hannigan et al, 1996; Clark and Brugge, 1995; Roskelley et al., 1995; Yamada and Miyamoto, 1995). Two aspects of integrin ligand binding properties are currently of particular interest to understanding MMP function. First, it was recently demonstrated that gelatinase A may be localized directly to cell surfaces by a direct interaction between the COOH-terminal domain of the enzyme and the integrin receptor ccv63 (Brooks et al., 1996). This represents a new mode of enzyme localization that complements the known binding of progelatinase A to membrane type metalloproteinases (MT-MMPs), likely in a complex with TTMP-2 bound to the COOH-terminal domain of gelatinase A (Strongin et al, 1995), and binding of the progelatinase A/TPMP-2 complex to a proposed specific TTMP-2 cell surface receptor (Emmert-Buck et al, 1995) (Fig. 1.1). The direct interaction of the gelatinase A COOH-terminal domain with the integrin receptor ocv63 (Brooks et al, 1996) may have important implications in cell surface proteolytic events. For example, the av63 integrin is also a major receptor for the extracellular matrix protein vitronectin (Suzuki et al, 1986; Pytela et al, 1985). Occupation of this cell surface receptor by vitronectin or by blocking antibodies enhances expression of gelatinase A and cellular penetration of basement matrices in vitro (Seftor et al, 1993; Seftor et al, 1992). Secondly, the regulation of several MMPs is intricately tied to integrin-ligand interactions. Werb et al. (1989) demonstrated elevated collagenase and stromelysin gene expression in synovial fibroblasts plated on monoclonal antibodies to the fibronectin receptor (a5fll) (Werb et al, 1989). A similar response was observed in synovial fibroblasts seeded on a 120-kDa RGD-containing cell binding peptide from fibronectin. In contrast, seeding the cells on a different fibronectin peptide containing the CS-1 portion of the IIICS domain, which is located outside the 120-kDa fibronectin segment, suppressed the MMP expression (Huhtala et al, 1995). Antibodies to the integrin a481, which recognizes the CS-1 peptide, blocked the suppression by CS-1 of the MMP expression induced by the 120 kDa peptide. Thus, cooperative signaling via a581 and oc4Bl 23 FIG. 1.1. Potential mechanisms of gelatinase A interaction with cell surface receptors and pericellular matrix molecules. Cell surface localization of gelatinase A has been proposed to occur by interactions of the gelatinase A C domain or the CBD with (A) extracellular matrix components (glycosaminoglycans, fibronectin, collagen, and elastin), (B) cell bound collagen and fibronectin, or (C) cell membrane receptors (MT-MMP, TIMP-2 receptor, avB3 integrin). 24 Introduction is important in regulating MMP expression in synovial fibroblasts. In keratinocyte cultures, treatment with specific antibodies to the major fibronectin and collagen receptors including the 151 and a3 integrin subunits, but not the a261 integrin, resulted in a several-fold stimulation of the expression of gelatinase B independent of which substrate cells were plated on (Larjava et al., 1993). The regulation of fibroblast MMP expression appears to be slightly different for cells grown in collagen lattices with alBl providing down-regulatory signals and a2Gl mediating induction of interstitial collagenase expression (Langholz et al, 1995; Riikonen et al, 1995). Moreover, in the absence of the a261 integrin the cells lacked both of these functions (Riikonen et al, 1995). Together these studies demonstrate that cellular expression of several MMPs is closely associated with integrin-ligand interactions. MATRIX METALLOPROTEINASES Matrix Metalloproteinases: General Aspects The time span from the initial observations of collagenolytic activity in tadpoles to the present day rapidly expanding family of matrix metalloproteinase observed in multiple species ranges over only approximately 35 years. In 1962, Gross and Lapierre observed that the actively remodeling tadpole tails contained collagenolytic activity at neutral pH (Gross and Lapierre, 1962). Subsequent research revealed that the enzyme cleaved collagen molecules at a single site resulting in the characteristic 3/4 and 1/4 fragments (Gross and Nagai, 1965). The first report on collagenase in man originated from studies of human gingiva and bone (Fulmer and Gibson, 1966). Today, the MMPs have been firmly established as a subfamily that is evolutionarily a part of the metalloendopeptidase family (Woessner, 1994). The major characteristics of the MMPs are that; 1) they are secreted or released in latent forms as zymogens and can be activated by proteinases or by organomercurials with a reduction in molecular mass (10-12 kDa) due to loss of the prodomain; 2) their catalytic mechanism depends on zinc in the active site; 3) their activity is inhibited by tissue inhibitors 25 Introduction of metalloproteinases; 4) the MMPs cleave one or more components of the extracellular matrix; and 5) they have structural and sequence similarities that include the prodomain cysteine switch sequence PRCGVPD and the zinc binding HEXGHXXGXXHS/T sequence (Woessner, 1994; Woessner, 1991). MMP Subgroups and Substrate Specificities Collectively, the MMPs have specificity and catalytic activity against most major extracellular matrix components. Therefore, the MMPs are integral parts of the tissue remodeling processes that underlie extracellular matrix changes during development, tissue remodeling, and pathological conditions such as inflammatory disease and invasive growth into tissue by tumor cells (Birkedal-Hansen et al., 1993; Matrisian, 1992). As illustrated in Table 1.1, the MMPs can collectively degrade most major extracellular matrix components. Based on sequence homologies and domain structure the MMPs have been categorized into the major subgroups collagenases: gelatinases, stromelysins, membrane-type metalloproteinases, and others. The major domains found in the MMPs include the prodomain, the catalytic domain, the collagen binding domain having a fibronectin type II-like triple repeat (CBD), and the COOH-terminal domain (Fig. 1.2). The CBD is only found in gelatinases A and B whereas a transmembrane domain is characteristic of the MT-MMPs. The domain structure of the MMPs and the contributions of separate domains to enzyme function are reviewed in more detail in the sections "MMP Domains" and "Domains and Ligand Interactions." Collagenases Interstitial collagenase (MMP1), the first MMP to be cloned (Goldberg et al., 1986), and neutrophil collagenase (MMP8) belong to the collagenase subgroup. Amino acid sequence comparison shows 57% identity between these two collagenases with highly conserved regions including the zinc-binding active site (Hasty et al., 1990). However, neutrophil collagenase is more highly glycosylated and has a larger mass (75 kDa) than the glycosylated form of interstitial collagenase (57 kDa). Interstitial collagenase, like many 26 Introduction Table 1.11 Substrate specificities MMP family members Enzyme MMP# Substrates Collagenases Interstitial Neutrophil Collagenase 3 Collagenase 48 Gelatinases Gelatinase A Gelatinase B MMP1 MMP8 MMP 13 MMP 18 MMP2 MMP9 Col I2, II, III, VII, X , entactin, TN, aggrecan, proGel A, proGel B Col I, n , i n , aggrecan Col I; H, ffl3; IV, IX, X, XIV, TN, FN15; gelatin3 Col I, IV, V, VH, X, gelatins, FN, LM, aggrecan, elastin, proGel B Col I9, i n 9 , IV, V, gelatins, elastin, entactin, aggrecan Stromelysins, matrilysin, metalloelastase Stromelysin 1 Stromelysin 2 Stromelysin 3 Matrilysin Macrophage metalloelastase MMP 12 Membrane-type metalloproteinases MMP3 Col n, IV, IX, X, XI, gelatin, LM, FN, elastin, TN, aggrecan, proCol, proGel B, neutrophil proCol MMP10 Col IV, LM, FN, elastin, aggrecan MMP11 Col IV, FN, LM, aggrecan MMP7 Col IV, gelatin, LM, FN, entactin, elastin, aggrecan, TN, proGel A, proGel B, proCol Elastin MT1-MMP MT2-MMP7 MT3-MMP5 MT4-MMP6 Others Envelysin Enamelysin1011 MMP 14 MMP 15 MMP 16 MMP 17 Progel A; gelatin, casein, elastin, FN, LM, VN, DSPG4 Col i , n , m 1 7 ProGel A 1 6 ProGel A Gelatin MMP activity now identified as existing MMPs Telopeptidase12 MMP4 Col I, FN, (likely MMP13) 3/4-collagen endopeptidase13 MMP5 Col I, gelatin; 3/4-fragments col I, H, m , (rat MMP2) Acid MMP14 MMP6 Cartilage proteoglycan, insulin B chain (MMP3) Table information based on reviews by Sang and Douglas, 1996, Birkedal-Hansen et al., 1993, and Matrisian, 1992. Updated sources are referenced as footnotes. 2Abbreviations: Col, collagen; FN, fibronectin; LM, laminin, proCol, procollagenase; proGel, progelatinase; TN, tenascin; VN, vitronectin; DSPG, dermatan sulfate proteoglycan. 5Knauper et al, 1996,4Pei and Weiss, 1996,5Takino et al, 1995,6Puente et al, 1996, 7Will and Hinzmann, 1995,8Cossins et al, 1996,9Okada et al, 1992, 10Bartlett et al, 1996, "Overall and Limeback, 1988, 12Nakano and Scott, 1987, 13Overall and Sodek, 1987, 14Azzo and Woessner, 1986, 15Knauper et al, 1997, 16Butler et al, 1997, 17Ohuchi et al, 1997. 27 Introduction other MMPs, also occurs as a non-glycosylated form having a mass of 52 kDa. Interstitial collagenase is mainly produced by fibroblasts and macrophages whereas neutrophil collagenase is primarily synthesized by cells of the neutrophil lineage. However, recent work indicates that neutrophil collagenase mRNA is found in other cell types including articular chondrocytes (Cole et al, 1996). Interstitial collagenase is synthesized on demand in response to cytokines and growth factors causing a lag time before the enzyme is available (Birkedal-Hansen et al., 1993). In contrast, neutrophil collagenase is synthesized only during neutrophil maturation in the bone marrow and stored in granula until immediate release on demand. These two collagenases cleave all three a-chains of native type I, II, and in collagen triple helices at single sites (Gly775-Ile776 of a 1(1) and Gly775-Leu776 of a2(I)) yielding the characteristic 3/4 and 1/4 collagen cleavage fragments (Hasty et al, 1987; Welgus et al, 1981). Although the interstitial and neutrophil collagenases share substrate specificity (Table 1.1), comparisons using monomeric collagens in solution revealed that neutrophil collagenase cleaves type I collagen more efficiently than types II and III, whereas interstitial collagenase cleaves type III collagen at a higher catalytic rate than type I and n collagens (Hasty et al, 1987; Welgus et al, 1981). It should be noted that these differences are less distinct in studies of fibrillar collagen substrates (Hasty et al, 1987) which may be a result of intermolecular crosslinks leading to reduced access to cleavage sites by the enzymes (Birkedal-Hansen et al, 1993). In addition, interstitial collagenase has also been reported to cleave type VII and X collagens (Table 1.1). The specificity of the cleavage site is surprising because 27 other sites throughout the triple helix have an identical sequence (Gly-[Ile or Leu]-[Ala or Leu]) to that of the cleavage site. This has lead to the suggestion that other determinants such as structure and glycosylation in the region of the cleavage site may be important for the cleavage specificity (Fields et al, 1987). In a recent review, Fields and Van Wart (1992) proposed a model for the collagenase cleavage site with the following characteristics: 1) The four triplet region 28 Introduction preceding the cleaved bond is rich in imino acids (proline, hydroxyproline) yielding a tight, thermally stable triple helix, 2) The four triplets that follow the cleavage site are deficient in imino acids, less thermally stable, and form a looser triple helix, and 3) There is a high degree of hydrophobicity in the cleavage region. The localized, looser structure of the helix may allow the enzyme to access the Gly-X scissile bond. The other non-cleaved sites differ in one respect or other from the proposed model for the collagenase cleavage site (Fields and Van Wart, 1992). Two new collagenases have been cloned recently. Human collagenase-3 (MMP13) was cloned from a breast tumor cDNA library (Freije et al, 1994) and shares a more than 50% sequence identity with the interstitial and neutrophil collagenases. Biochemical characterization has shown that collagenase-3 preferentially hydrolyzes type II collagen and is 5-6 times less efficient in cleaving type I and III collagens compared to the interstitial and neutrophil collagenases (Knauper et al, 1996). Unlike the other collagenases, collagenase-3 efficiently degrades gelatin. In addition collagenase-3 has activity against tenascin, fibronectin, and type IV, IX, X, and XIV collagens, and a COOH-terminal domain deletion mutant of collagenase-3 retained telopeptidase activity but failed to cleave type I and II triple helical collagens (Knauper et al, 1997). Finally, a partial human cDNA sequence is now available for a fourth putative member of the collagenase subgroup (MMP 18). It has been designated collagenase-4 and is expressed in a variety of normal human tissues (Cossins et al, 1996), but the substrate specificity is not known at this time. Gelatinases The second major subgroup of the MMPs, the gelatinases, has two members, gelatinase A (72-kDa gelatinase, MMP2) (Huhtala et al, 1990; Collier et al, 1988; Salo et al, 1983) and gelatinase B (92-kDa gelatinase, MMP9) (Wilhelm et al, 1989). These two MMPs are structurally distinct from the other MMPs by containing three fibronectin type II-like modules inserted into the catalytic domains (Wilhelm et al, 1989; Collier et al, 1988). The domains comprising these three modules provide the enzymes with important ligand 29 Introduction binding properties (Chapter 2, Steffensen et al., 1995). (See also subsequent sections on structure and ligand interactions of the fibronectin type II-like modules and MMP domains). The two gelatinases show a 49% amino acid identity (Wilhelm et al., 1989) with the main difference between them being a 54 amino acid hinge region that is found only in gelatinase B. This region is located between the catalytic domain and the COOH-terminal domain and has sequence similarity with the a2-chain of type V collagen. Unlike gelatinase A, gelatinase B is glycosylated (Wilhelm et al., 1989). Gelatinase A is widely distributed. Although initially characterized in tumor cells, it has been reported to be constitutively expressed in many tissues by fibroblasts, chondrocytes, keratinocytes, endothelial cells, and monocytes (Salo et al., 1991; Murphy et al., 1989; Garbisa et al., 1986; Salo et al, 1983; Liotta et al., 1979). In comparison, gelatinase B is mainly expressed in polymorphonuclear leukocytes, monocytes, and alveolar macrophages (Hibbs et al., 1985; Mainardi et al, 1984). However, gelatinase B is also produced by keratinocytes (Salo et al., 1991; Wilhelm et al., 1989), malignant or transformed cell lines (Moll et al, 1990; Wilhelm et al, 1989), and human cytotrophoblasts (Fisher et al., 1989; Librach et al., 1991). The substrate specificity of the gelatinases is quite similar (Table 1.1). The two gelatinases efficiently degrade all types of denatured collagen oc-chains into small fragments by cleaving not only at the Gly-Leu and Gly-Ile bonds, but also at Gly-Glu, Gly-Asn, and Gly-Ser (Seltzer et al, 1990; Murphy et al, 1989; Nakano and Scott, 1987; Murphy et al, 1985). The position of hydroxyprolines may also contribute to defining the gelatinase cleavage sites since this residue is commonly found in a position five residues from the cleavage sites in cleavage fragments generated by proteolysis of cyanogen bromide peptides of type I collagen (Seltzer et al, 1990). Both gelatinases cleave type TV collagen into 140 and 125 kDa fragments with the scissile bonds located in the triple helical region (Gly^g-Ile^ in al(IV) and Gly464-Leu465 in a2(IV)) (Collier et al, 1988; Hostikka and Tryggvason, 1988; Fessler et al, 1984). However, questions have been raised regarding the efficiency of the 30 Introduction two gelatinases against type IV collagen in vivo because efficient in vitro degradation required conditions that favor denaturation or pepsin solubilization of the collagen (Mackay et al., 1990). In addition to denatured type I collagen, purified human and rat gelatinase A cleave full length, as well as 3/4 and 1/4 fragments, of native type I collagen (Okada et al., 1992; Sodek and Overall, 1992). The cleavage site is at the same Gly-Ile/Leu bonds in the oc-chains and and the proteinase has the same kinetic characteristics as those determined for cleavage by interstitial collagenase (Aimes and Quigley, 1995). Also, purified gelatinase B has been shown to cleave both type I and in collagen under conditions that do not denature the collagens (Okada et al, 1992). Gelatinase A also cleaves type VII and X collagens (Welgus et al, 1990; Seltzer et al, 1989) and both gelatinases are active against type V collagen (Niyibizi et al, 1994; Okada et al, 1990; Wilhelm et al, 1989; Collier et al, 1988; Murphy et al, 1985) and elastin (Senior et al, 1991; Okada et al, 1990; Murphy et al, 1985). Degradation by gelatinases of the triple helix of type V collagen yields fragments of 10 kDa or less (Niyibizi et al, 1994; Okada et al, 1990). In comparison, gelatinase A cleaves type X collagen, like interstitial collagenase, at two loci in the native type X triple helix (Gly92-Leu93 and Gly420-Ile421) yielding three major proteolytic fragments (Welgus et al, 1990a). Elastin, which is generally highly resistant to proteolytic degradation is sensitive to degradation by gelatinases A and B and, in addition, to neutrophil leukocyte-derived elastase, cathepsin G, and proteinase 3 (Senior et al, 1991). Activated gelatinases have, on a molar basis, approximately 30% activity in degrading elastin relative to the leukocyte elastase. Analyses of the degradation products have shown that the large, insoluble elastin aggregates are extensively degraded to fragments not detectable by SDS-PAGE (Senior et al, 1991). Whereas gelatinase A has the capacity to degrade fibronectin and laminin, gelatinase B does not possess activity against these proteins (Okada et al, 1990). The initial fibronectin degradation product has an apparent molecular weight of about 200 kDa, but extended incubation with synovial fibroblast gelatinase A produces six fragments with molecular 31 Introduction weights in the range of 65 to 180 kDa pointing to multiple internal cleavage sites. In comparison, increased electrophoretic mobility of both the A and B chains of laminin following digestion with gelatinase A suggests cleavage sites near the NH2- and/or the COOH-terminal ends in both of these chains (Okada et al, 1990). Degradation of articular proteoglycan in osteoarthritis and during aging is associated with an increase in MMP activity that is mainly ascribed to stromelysin-1. However, it has been shown that additional MMPs, including matrilysin, interstitial collagenase, as well as gelatinases A and B can degrade aggrecan core protein reducing the molecular mass to that of the chondroitin sulfate side chain (Nguyen et al, 1993). Gelatinase B demonstrated less activity against aggrecan core protein than the other MMPs and did not cleave link protein which, in comparison, was readily cleaved by gelatinase A (Nguyen et al, 1993). Together, these results show that the gelatinases have catalytic activities against a number of extracellular matrix components in addition to denatured collagens. Stromelysins, matrilysin, and matrix metalloelastase The stromelysin MMP subgroup has at least three members including stromelysins-1, -2, and -3 (MMP3, MMP10, MMP11). In addition, based on sequence similarities, matrilysin (PUMP-1, MMP7) (Quantin etal, 1989) and macrophage metalloelastase (MMP 12) (Shapiro et al, 1993; Shapiro et al, 1992) may be included in this subgroup. Thus, by amino acid sequence comparisons stromelysin-1 and -2 show 79% and 49% sequence identity with matrilysin, respectively (Henriet et al, 1992; Muller et al, 1988). However, a major difference is the absence of the COOH-terminal domain in matrilysin. The homologous rat counterpart of the human stromelysins was originally termed transin (Matrisian et al, 1986; Matrisian et al, 1985). Stromelysin-1 is secreted in both an un-glycosylated (57 kDa) and a glycosylated (60 kDa) form (Wilhelm et al, 1987). The stromelysin-3 gene has been cloned from a breast carcinoma cDNA library and is expressed in invasive breast cancers (Basset et al, 1990). Contrary to stromelysins-1 and -2 as well as most other MMPs, except the MT-MMPs, stromelysin-3 undergoes intracellular 32 Introduction activation. This mode of enzyme activation has been shown to be mediated by furin, a serine proteinase of the subtilisin group which cleaves a range of precursor proteins that contain the consensus sequence RXK/RR (Santavicca et al, 1996; Pei and Weiss, 1995). Stromelysin-1 cleaves a wide range of extracellular matrix components from both interstitial connective tissues and basement membranes (Table 1.1). These include proteoglycan core protein, gelatin, the non-helical and pepsin-sensitive globular domain of type IV collagen, types LX, X, and XI collagens, the non-helical NH2- and COOH-terminal peptides of type II collagen, the glycoprotein fibronectin yielding two 170 and 145 kDa fragments, and both chains of laminin (Sang and Douglas, 1996; Birkedal-Hansen et al, 1993; Murphy et al, 1991; Wilhelm et al, 1987; Chin et al, 1985). Elastin is also degraded by the stromelysin MMP subgroup members, and particularly efficiently by matrilysin and the macrophage metalloelastase (Shapiro et al, 1993; Shapiro et al, 1992; Murphy et al, 1991; Quantin et al, 1989). Stromelysin-2 and matrilysin have activity against some or all of the substrates of stromelysin-1 (Table 1.1). Although stromelysin-3 cleaves several extracellular matrix components, the activity is low (Welgus et al, 1990b). Importantly, stromelysin-1 and matrilysin can activate procollagenase, as well as progelatinases A and B (Shapiro et al, 1995; Murphy et al, 1987). Membrane type metalloproteinases The membrane type matrix metalloproteinases (MT-MMPs) constitute a separate subgroup of the MMPs by virtue of the fact that they are integral cell membrane proteins. The discovery of MT1-MMP (Sato et al, 1994) provided an longexplanation to the phenomenon of activation of progelatinase A which occurred in the presence of Con A (Overall and Sodek, 1990) or TPA-stimulated fibroblasts (Brown et al, 1990), or isolated membranes from these cells (Ward et al, 1991). (Discussed further in the section "Cell Membrane Localization and Activation of Gelatinase A"). Subsequently, three additional MT-MMPs have been identified and cloned (Puente et al, 1996; Takino et al, 1995; Will and Hinzmann, 1995). 33 Introduction The MT-MMPs are structurally characterized by a hydrophobic domain that is located as an extension to the COOH-terminal domain and serves to anchor the enzyme in the cell membrane. Thus deletion of this domain abolishes cell membrane positioning (Pei and Weiss, 1996; Cao et al, 1995). An additional unique -10 amino acid long motif is characteristic of all four MT-MMPs in addition to stromelysin-3 (Sang and Douglas, 1996). This sequence contains the RXK/RR consensus sequence which been shown to serve as an activation site in stromelysin-3 for furin (Pei and Weiss, 1995) and likely has a similar role in intracellular activation of MT-MMPs (Basbaum and Werb, 1996). Recently, however, an alternative extracellular activation mechanism was presented (Okumura et al., 1997) by which proMTl-MMP is transported to the cell membrane and, there, activated by extracellular plasmin. These results suggest that there may be both an intracellular and an extracellular activation pathway for MT1-MMP. Initially, the only function ascribed to MT1-MMP was that of proteolytic activation of progelatinase A (Sato et al., 1994). However, both native MT1-MMP and a transmembrane domain-deletion mutant of MT1-MMP not only mediated activation of progelatinase A indistinguishable from wildtype MT1-MMP, but also showed proteolytic activity against a range of extracellular matrix molecules including type I, II, and III collagens, gelatin, fibronectin, vitronectin, and laminin-1 (Table 1.1) (Ohuchi etal., 1996; Pei and Weiss, 1996). The cleavage of native type I collagen is at the mammalian cleavage site but MT1-MMP is 5 - 7 fold less efficient than interstitial collagenase in cleaving this substrate (Ohuchi et al, 1996). This shows that MT1-MMP, and potentially the other MT-MMPs, are not only active in activation of progelatinase A, but also may be directly involved with degradation of extracellular matrix components. Other MMPs MMP4 is also known as telopeptidase. It was partially purified from human gingival fibroblast cultures and shown to have activity against type I collagen telopeptides as well as fibronectin (Nakano and Scott, 1987; Scott et al., 1983). This gene encoding this enzyme has 34 Introduction yet to be cloned, but may represent MMP 13. A different endopeptidase (MMP5) was identified in human gingival crevicular fluid and rat osteosarcoma cultures (Overall and Sodek, 1987). This MMP showed activity against native 3/4 collagen fragments resulting from cleavage of type I, II, and HI collagens, native type I collagen, and gelatin. MMP5 was subsequently identified as rat gelatinase A (Sodek and Overall, 1992). Other studies of human articular cartilage extracts revealed an acid metalloproteinase (MMP6) that had the capacity to digest cartilage proteoglycans and also the B chain of insulin (Azzo and Woessner, 1986) and likely represented stromelysin-1. Finally, a recent report describes the cloning of a porcine tooth enamel-associated metalloproteinase (Bartlett et al, 1996) which was originally observed in small amounts at all stages of developing enamel matrix and characterized by gelatinolytic activity, but did not degrade amelogenins (Overall and Limeback, 1988). MMP Activation The MMPs are secreted in latent forms and require activation for catalytic activity. Activation involves disruption of the bond between the active site Zh^ and the Cys residue in the conserved prodomain sequence PRCGVPD and introduction of a molecule of water as the fourth ligand (Van Wart and Birkedal-Hansen, 1990). The mechanism has been referred to as the "cysteine switch" (Windsor et al, 1991; Springman et al, 1990). The importance of this coordinating bond is illustrated by spontaneous activation of the enzyme following site-specific mutations to introduce changes in the conserved propeptide sequence (Sanchez-Lopez et al, 1988). Disruption of the bond may be achieved chemically by various reagents including organomercurials, metal ions, thiol reagents, and oxidants to cause an opening of the cysteine switch (Birkedal-Hansen et al, 1993). A similar effect is observed with detergents such as SDS (Birkedal-Hansen and Taylor, 1982) and has been utilized in zymography analysis of MMPs. Common for these modes of activation is the initial absence of a change in mass although autocatalytic processing may occur subsequently (Grant et al, 1987). 35 Introduction Proteases (trypsin, chymotrypsin, plasmin, plasminogen activators (u-PA and t-PA), neutrophil elastase, and plasma kallikrein) may activate MMPs. Such proteolytic processing results in the removal of a segment of the prodomain NH2-terminal to the final cleavage site which results from subsequent autocatalytic cleavage to generate the final active form of the enzyme (Birkedal-Hansen et al., 1993; Nagase et al, 1990). The position of the final cleavage site varies between MMPs but is generally at Tyr or Phe residues approximately eight residues downstream of the conserved Cys residue (Birkedal-Hansen et al, 1993). MMPs may also proteolytically activate other MMPs. Examples are activation of procollagenase by stromelysins (Nicholson et al., 1989; Quantin et al, 1989; Murphy et al, 1987) and activation of progelatinase B by stromelysin-1 (Goldberg et al, 1992; Ogata et al, 1992) and gelatinase A (Friedman et al, 1995). In addition to the cell surface activation by the MT-MMPs (Butler et al, 1997; Strongin et al, 1995; Sato et. al, 1994), gelatinase A may autoactivate (Bergmann et al, 1995). Tissue Inhibitors of Matrix Metalloproteinases (TIMPs) Tissue inhibitors of matrix metalloproteinases are secreted proteins that are widely distributed in tissues and fluids and serve as specific inhibitors of MMPs. A specific inhibitor of collagenase and other MMPs was first cloned from murine fibroblasts (Edwards et al, 1986). Subsequently, TIMPs from several species have been identified, characterized, and the genes cloned, including those encoding four TIMPs from human (Greene et al, 1996; Kishnani et al, 1995; Leco et al, 1994; Wick et al, 1994; Pavloff et al, 1992; Stetler-Stevenson et al, 1990; Carmichael et al, 1986). TIMP-1 (30 kDa) (Stricklin and Welgus, 1983; Welgus et al, 1979) is glycosylated whereas TIMP-2 (21 kDa) (Goldberg et al, 1989) is not. The two proteins have approximately 40% amino acid sequence identity. Twelve cysteine residues in conserved positions are all involved in disulfide bond formation thereby dividing the TIMPs into two distinct domains, each with three internal disulfide-bonded loops (Williamson et al, 1990). The MMPs are inhibited by both TFMP-1 and -2 with a stoichiometry of 1:1. TIMPs 36 Introduction are tight-binding inhibitors (Kd of 10"9 to 10"10 M) of the active forms of all the MMPs, and TIMP-1 and TIMP-2 also form tight complexes with the latent forms of gelatinase B and A, respectively (Goldberg et al, 1989; Stetler-Stevenson et al, 1989). The occupation of the latent enzymes may serve to delay conversion to an active form (Howard et al, 1991) but recently, for gelatinase A, has also been shown to serve in a MT-MMP/TIMP-2/progelatinase A activation complex at cell surfaces (Strongin et al, 1995). The mechanism of binding is not yet well understood. However, it is known that gelatinase A has TIMP-2 binding sites on both the catalytic and COOH-terminal domains (Howard and Banda, 1991). In addition, the rC domain of gelatinase A has been found to contain two separate binding sites for the NH2- and COOH-terminal domains of TIMP-2 (Overall et al, 1997). TIMPs provide an important control mechanism for MMP activity. A disruption of the balance, as has been introduced into experimental in vitro and in vivo systems, may cause tissue damage and an expanded rate of tumor invasive growth (Edwards et al, 1996; Liotta and Stetler-Stevenson, 1991). Thus transfection of normal non-malignant cells with plasmids leading to production of antisense TIMP mRNA, and thereby reducing TIMP expression, conferred an oncogenic phenotype on the cells (Khokha et al, 1989). Also, TIMP knock-out transgenic mice show an accelerated growth and dispersion of tumors (Khokha et al, 1995). In addition to inhibiting MMPs, the TIMPs have been associated with cell growth stimulatory effects for non-erythroid cell types (Edwards et al, 1996; Birkedal-Hansen et al, 1993). This might be of significance in controlling vascular proliferation during tumor expansion. MMP Domains The MMPs are organized in distinct structural domains that are homologous between the MMPs, although there are differences in the number of domains present in each enzyme (Fig. 1.2) (Birkedal-Hansen et al, 1993). Following the signal peptide, that is removed in the endoplasmic reticulum, all MMPs have a prodomain (77 - 87 residues), which is the NH2-terminal domain of the secreted latent enzyme, and a catalytic domain. A COOH-terminal 37 Col lagenases IJiH Pro ' Catalytic: •( Hemopexin ) Gelat inase A d — • • ! Pro Gelatinase B BHI-l Pro IE5 ( Hemopexin ) Ca taMc : tPC Hemopexin ) Stromelysins [Pm\-\ Pro Catalvtta : : : : l ( Hemopexin ) Membrane-type metalloproteinases CEB-1 Pro I d L ^ Catalytic C H p m n p P Y i n 1 Others Pro- Catalytic: FIG. 1.2. Domain structure of the major matrix metalloproteinases subgroups. Abbreviations are: Pre, Signal peptide; Pro, prodomain; Fu, putative furin cleavage site; CBD, collagen binding domain; Zn, Zn++-binding active site; TM, transmem-brane domain; V, type V collagen-like region. 38 Introduction (-200 residues) is a constant finding in most MMPs and is connected with a flexible, variable length (5 - 50 residues) proline-rich hinge region to the catalytic domain. The COOH-terminal domain consists of four repeats with homology to hemopexin, a heme binding protein, and vitronectin (Jenne and Stanley, 1987). Matrilysin is the only MMP that does not possess the COOH-terminal domain and, therefore, has a considerably smaller molecular mass than the other MMPs. The catalytic domain contains a Zn^ binding site with three histidine residues in the highly conserved active site sequence HEXGHXXGXXH (Woessner, 1994). These three histidines bind a Zn""" ion. In the latent form of the enzyme, a fourth ligand site is occupied by a prodomain cysteine contained in the conserved PRCGXPD motif - disruption of the coordinating bond to this cysteine (cysteine switch) and introduction of a water molecule to occupy the fourth binding site are elements of enzyme activation (Springman et al, 1990; Van Wart and Birkedal-Hansen, 1990). Unlike other members of the matrix metalloproteinase family, the gelatinases A and B contain three 58-amino acid residue long modules that are highly homologous to the fibronectin type II-like modules and inserted into the catalytic domain (Wilhelm et al., 1989; Collier et al, 1988). In addition to fibronectin and gelatinases of the MMP family, fibronectin type II-like modules have been identified in the blood coagulation factor XII (MacMullen and Fujikawa, 1985), the mannose receptor (Taylor et al, 1990), the mannose-6-phosphate receptor (Morgan et al, 1987), and the bovine seminal fluid proteins PDC-109 and BSP-A3 (Seidah et al, 1987; Esch et al, 1983). Gelatinase B contains an additional 54 amino acid residue proline-rich sequence located adjacent to the zinc-binding domain on the carboxyl-terminal side (Huhtala et al, 1991; Wilhelm et al, 1989). The length of this sequence and 30-55% identity with a portion of the helical region of the collagen a2(V) chain suggest that the appearance of this domain may result from a recombinatorial event between the enzyme precursor gene and a collagen gene (exon length in collagen genes is commonly 54 bps) (Wilhelm et al, 1989). 39 Introduction The four membrane-type metalloproteinases (MT-MMPs) contain a characteristic hydrophobic region that plays a critical role in positioning the MT-MMPs in the cell membrane (Puente et al, 1996; Takino et al., 1995; Will and Hinzmann, 1995; Sato et al., 1994), and an insertion of nine residues between the propeptide and the catalytic domain ending with the RXR/KR consensus sequence that in stromelysin-3 is essential for activation by furin (Pei and Weiss, 1995). The MMPs require Ca** ions for activity and putative Ca"1-1' binding sites have been proposed in Glu-Asp rich regions of the catalytic domain (Lepage and Gache, 1990). There is also recent evidence that Ca*4' bound centrally in the COOH-terminal domain of gelatinase A (Libson et al., 1995) is required for interaction with ECM proteins (Wallon and Overall, 1997). In addition, the COOH-terminal domain may also bind a Zn++. MMP Genes As with the protein structure, the gene structure is also conserved among the MMPs. The exons and introns align well between the MMPs corresponding to the modules that they encode, with the main difference in size reflected by varying lengths of the introns (Collier et al, 1991; Huhtala et al., 1991; Huhtala et al., 1990). This lends additional support to the proposed development from a common ancestral gene. The genes for rabbit and human interstitial collagenase (Collier et al., 1988; Fini et al, 1987), rat transin 1 and 2 [highly homologous with the human stromelysin 1 and 2] (Breathnach et al, 1987; Matrisian et al., 1986), and human macrophage metalloelastase (Belaaouaj et al., 1995) contain ten exons and nine introns and have a size of 8 - 12 kb. The NH2-terminal part of the MMPs, including the prodomain and the catalytic domain, is encoded by the first five exons. Exon 5 encodes the Zn -^binding active site and exons 6-10 give rise to the hemopexin-like COOH-terminal domain of the MMPs. The gelatinase A and B genes contain three additional exons inserted after exon 4, each of which encodes one of the three fibronectin type II-like modules (Huhtala et al, 1991; Huhtala et al., 1990; Wilhelm et al, 1989; Collier et al, 1988). Although the size of the gelatinase B gene, in spite of the 40 Introduction additional exons, corresponds to the other MMP genes (7.7 kb), the gelatinase A gene is characterized by larger introns and is approximately 27 kb in size. In addition, alignment of the mouse and human gelatinase B showed a conserved exon/intron structure with some differences in the length of the introns (Masure et al, 1993; Huhtala et al, 1991). The gene structures for the MT-MMPs are not yet available, but predictions from available DNA sequences suggest a modular structure as in the other MMPs. The genes for human interstitial collagenase, stromelysin-1, and macrophage metalloelastase have been assigned to chromosome 11 (Belaaouaj et al, 1995; Spurr et al, 1988), while the genes for the human gelatinase A and B reside on chromosome 16 (Collier et al, 1991; Huhtala et al, 1990). It has been demonstrated that the promoters of the human and rabbit interstitial collagenases, the rat stromelysin-1 and human stromelysin-1 and -3, macrophage metalloelastase, and gelatinase B have TATA motifs (Belaaouaj et al, 1995; Huhtala et al, 1991; Quinones et al, 1989; Angel et al, 1987; Fini et al, 1987; Matrisian et al, 1986). Different from these genes, gelatinase A does not have a TATA motif (Huhtala et al, 1990) or a TPA response elements (TRE) where AP-1 proteins, that increase significantly in response to TPA, can bind (Huhtala et al, 1991; Huhtala et al, 1990; Angel et al, 1987; Fini et al, 1987; Matrisian et al, 1986). The lack of a TRE in gelatinase A concurs with the observation that human collagenase, stromelysin-1, and gelatinase B are induced by TPA but that this compound does not have an effect on gelatinase A expression (Sato and Seiki, 1993; Mackay et al, 1992; Huhtala et al, 1990; Overall and Sodek, 1990; Angel et al, 1987). The AP-1 and PEA3 motifs are found in most MMPs and mediate responsiveness to growth factors, oncogenes, and phorbol esters as observed in human macrophage metalloelastase, interstitial collagenase, stromelysin-1, and matrilysin (Wilson and Matrisian, 1996; Belaaouaj et al, 1995; Schorpp et al, 1995). In addition to AP-1, the c-Ets factors have been reported to activate interstitial collagenase and stromelysin-1 genes through specific promoter elements (Wasylyk et al, 1991; Gutman and Wasylyk, 1990). However, c-Ets 41 Introduction factors do not appear to be important in regulation of gelatinase B (Sato and Seiki, 1993). In addition to AP-1 and PEA3 sites (Gum et al, 1996), gelatinase B contains a TRE as well as Sp-1 and NF-kappa B transcription factor binding elements that are required in this enzyme for c-Jun/c-Fos, TPA, and TNFoc induced promoter activity (Sato and Seiki, 1993; Frisch and Morisaki, 1990). The promoter of gelatinase A which is unique among the MMPs by not having AP-1 or PEA3 sites appears to utilize different regulatory factors. Like gelatinase B, gelatinase A has Sp-1 binding elements (Frisch and Morisaki, 1990). In addition, an enhancer element, r2, binds a protein with binding specificity very similar to that of the transcription factor AP-2 (Frisch et al, 1990). Human stromelysin-3 also does not have an AP-1 binding site but evidence for activation by retinoic acid through a retinoic acid response element points to regulation of this enzyme by specific factors associated with tissue remodeling (Anglard et al, 1995). MMP Three-Dimensional Structure Structures have been determined for several zinc-containing proteinases including thermolysin (Matthews et al, 1972; Matthews et al, 1972), a serralysin from Pseudomonas aeruginosa (Thayer et al, 1991), the digestive enzyme astacin from crayfish (Gomis-Ruth et al, 1993; Bode et al, 1992), the snake venom metalloproteinase adamalysin II (Gomis-Ruth et al, 1993) and, among the human matrix metalloproteinases, the recombinant catalytic domains from fibroblast collagenase (Barkakoti et al, 1994), neutrophil collagenase (Stams et al, 1994), and stromelysin-1 (Gooley et al, 1994). These structures are reflected in the following summary. In the "metzincin" superfamily, members of the serralysin, astacin, matrix metalloproteinase, and adamalysin subfamilies have been found to have a similar topology of the catalytic domains (reviewed by Stocker et al, 1995). The global fold is formed by a five-stranded 6-sheet structure with three a-helices arranged in sequential order. The consensus motif HEXXHXXGXXH is characteristic of the four proteins forming an active 42 Introduction site Zn"1"1" binding motif. The HEXXH residues forms a helix exposing the two His which bind the catalytic Zn"1-". A descending strand gains flexibility from a Gly and contains the third active site His. This strand continues into the characteristic "Met-turn" motif which contains a Met that, located beneath the active site Zn++, appears to be essential for the structural integrity of the active site. The active site Zn** is the only metal ion bound to astacin. However, in adamalysin the catalytic domain has an additional Ca""" ion, and the neutrophil and fibroblast collagenases have two Zn"1-" and two Ca** ions that are considered to be of structural importance (Stoecker et al., 1995). There is developing evidence that single amino acid changes may significantly affect the active site pockets of otherwise similar structures. Thus, an Arg to Leu change between human fibroblast and neutrophil collagenase yielded an enlarged SI' pocket (Stams et al., 1994). Thermolysin has some structural similarity to the metzincins but deviates structurally by having only HEXXH in the Zn*4" binding site and a Glu in place of His as the third Zn"1-" binding ligand separated by 20 amino acids from the second His. The crystal structure of the COOH-terminal domain of gelatinase A was solved recently (Libson et al, 1995). It shows a four-bladed 6-propeller protein with the four blades arranged around a channel that contains a bound structural Ca** and a Na+Cl" ion pair. Each blade of the propeller is a 6-sheet comprised of four anti-parallel 6-strands, with the innermost 6-strands being parallel and forming the channel. The COOH-terminal domain is bound to the catalytic domain by a flexible linker. Based on the close homology of gelatinase A with other metalloproteinases, it is reasonable to assume that the three-dimensional structure corresponds to those solved for other MMPs. However, compared to collagenase, the gelatinases contain the collagen binding domain, formed by the three fibronectin type II-like modules, inserted seven residues upstream of the HEXXHXXGXXH consensus motif. Collagen Binding Domain Three-Dimensional Structure The structure for the fibronectin type II-like modules of gelatinase A has not been 43 Introduction solved yet. However, the NMR structure is known for the fibronectin type II-like module (PDC-109/b) found in the bovine seminal fluid protein PDC-109 (Constantine et al., 1992; Constantine et al., 1991). The primary structure of PDC-109/b displays a high level of amino acid homology with the three type II-like modules of gelatinase A - 40%, 34%, and 48%, respectively. In addition, from CD spectral analysis, the estimated secondary structures of the type II modules of gelatinase A have been found to be in agreement with predictions based on the homology with the PDC-109 modules (Banyai et al, 1996). Therefore, predictions of the gelatinase A CBD structure made in this thesis have been based on the available PDC-109/b coordinates (Brookhaven Protein Data Bank). NMR structural analysis (Constantine et al, 1992; Constantine et al, 1991) showed that PDC-109/b contains a central core defined by two antiparallel 6-sheets and two irregular loop regions (Fig. 1.3). Starting from the N-terminus, an extended strand connects to a pair of antiparallel 6-strands that form one 6-sheet. The two strands are connected by a short turn involving two Gly. 6-strand 2 connects by an extended strand and an irregular loop structure to the third 6-strand. A large irregular loop structure connects the third and fourth 6-strands that pair to form a second antiparallel 6-sheet. Two disulfide bridges lie approximately perpendicular to each other connecting four cysteines that are highly conserved in the fibronectin type II-like modules of gelatinase A and B, fibronectin, and PDC-109. In addition, PDC-109/b forms a cavity largely characterized by a cluster of aromatic residues which are solvent-accessible and define a hydrophobic surface. It was proposed from this molecular modeling that Tyr7, Trp26, Tyr33, Asp34, and Trp39 were important in ligand binding (Constantine et al, 1992; Constantine et al, 1991). Analyses of the collagen binding domain position in context of the known structure of catalytic domains of MMPs (Banyai et al, 1996) have led to a proposed model according to which the collagen binding domain is inserted at the end of the active site cleft, close to the S'j pocket of the enzyme. The model suggests that the three fibronectin type II-like modules interact tightly forming a groove that is lined up with the catalytic domain active site 44 Gelatinase A module 2 Gelatinase A module 3 FIG 1.3. Stereo images of fibronectin type II-like modules. Three-dimensional structure of the fibronectin type II-like modules from bovine seminal fluid protein PDC109b. Coordinates from Brookhaven Protein Data Bank, accession number 1PDC (Constantine et al., 1991, 1992). Positions of lysine residues substituted with alanine residues in gelatinase A CBD are indicated {Chapter 3) View the stereo images using a stereo viewer. Alternatively, position your left and right eyes over the left and right images, respectively. Then, starting very close to the paper, move slowly back until the two superimposed images are in focus and appear as a three-dimensional image. 45 Introduction groove and may serve to orient and bind substrate molecules to present the scissile bond to the active site of the enzyme. Cell Membrane Localization and Activation of Progelatinase A Progelatinase A is unique among the MMPs in that it can be activated by a mechanism that involves stimulated cell membranes. Initially it was observed that progelatinase A was processed to 62, 59, and 43 kDa active forms following stimulation of human fibroblasts with Con A (Brown et al., 1993; Overall and Sodek, 1992; Overall and Sodek, 1990) or tumor cells with TPA (Azzam and Thompson, 1992; Emonard et al., 1992; Brown et al, 1990). To localize the site of activation, it was subsequently found that whole cell lysates and plasma membranes of Con A stimulated fibroblasts, but not the conditioned medium from the same cells, had the capacity to process progelatinase A (Ward et al., 1991). This suggested that a component of the cell membrane was involved in the endogenous activation of progelatinase A. This component has later been identified as MT1-MMP (Sato et al, 1994) and an additional three MT-MMPs have been cloned (Puente et al., 1996; Takino et al., 1995; Will and Hinzmann, 1995). Recently, it has also been demonstrated that recombinant catalytic domain of MT2-MMP can initiate activation of progelatinase A (Butler et al., 1997). The MT-MMPs possess as an extension to their COOH-terminal domain a hydrophobic domain that anchors the enzymes in the cell membrane and deletion of this domain abolishes the positioning in the cell membrane (Pei and Weiss, 1996; Cao et al., 1995). One model for cell surface positioning of progelatinase A during activation entails TIMP-2 bridging the COOH-terminal domain of progelatinase A and MT1-MMP (Strongin et al., 1995). This could explain the requirement for the progelatinase A COOH-terminal domain for membrane activation (Ward et al., 1994; Murphy et al, 1992). This model also explains why the presence of excess TIMP-2, but not TIMP-1, or isolated rC domain of gelatinase A reduces the cell membrane binding as well as activation of progelatinase A (Overall et al., 1997; Ward et al, 1994; Ward et al., 1991). Binding studies have recently demonstrated that gelatinase A rC domain binds the NH2- as well as the COOH-terminal 46 Introduction domains of TIMP-2 (Overall et al, 1997) opening new options for complex formation during activation. In addition, Chapter 4 of this thesis describes results of investigations of a new mechanism for cell surface localization of gelatinase A/progelatinase A that is mediated by the CBD rather than the COOH-terminal domain. The relative importance of the CBD and the COOH-terminal domain for progelatinase A complex formation with MT-MMPs and activation has been demonstrated by the unaltered activation in gelatinase A CBD deletion mutant enzyme but no activation in a COOH-terminal domain deletion mutant of the enzyme (Ward et al., 1994; Murphy et al., 1992). It has been reported that TIMP-2 as well as a complex of progelatinase A/TIMP-2 can bind directly to cell surfaces with a Kd of 2.5 nM and 30,000 sites/cell (Emmert-Buck et al, 1995). TIMP-2 binding to cells can further elicit signal transduction events as illustrated by the rTIMP-2 induced production of cAMP and cell proliferation that can be inhibited by an adenylate cyclase inhibitor (Corcoran and Stetler-Stevenson, 1995). Based on these results it has been proposed that a specific TIMP-2 receptor can mediate enzyme binding (Emmert-Buck et al, 1995). Direct binding of gelatinase A via the COOH-terminal domain to the av63 integrin receptor was demonstrated recently (Brooks et al, 1996). The COOH-terminal domain of gelatinase A is homologous to vitronectin which is a major extracellular matrix ligand for the avB3 integrin receptor (Suzuki et al, 1986; Pytela et al, 1985). Interestingly, occupation in vitro of this cell surface receptor by blocking antibodies or by vitronectin enhances the cellular penetration of basement membrane matrices and the expression of gelatinase A (Seftor et al, 1993; Seftor et al, 1992). It is not yet known, whether occupation of the ocv63 receptor is associated with altered gelatinase A expression. However, incubation of cells with other integrin ligands such as the fibronectin CS1 and heparin binding peptides or with blocking antibodies to the 61, a3, a461, and oc561 integrin profoundly affects cellular expression of several MMPs (Huhtala et al, 1995; Larjava et al, 1993; Werb et al, 1989). The understanding of cell membrane localization of MMPs is developing and it is 47 Introduction becoming evident that several direct and indirect modes of interaction between the enzyme and the cell membrane components are being utilized for enzyme binding. DOMAINS AND LIGAND INTERACTIONS Protein Modules and Domains Most of the extracellular matrix proteins are characterized by a multidomain structure. Among the prominent multidomain extracellular matrix proteins are such molecules as fibronectin, tenascin, and laminin [See Bork et al. (1996) for a detailed review of structure and distribution of modules in extracellular proteins]. Individual domains have been associated with specific functions including receptor recognition, metal ion binding, oligosaccharide binding, and interaction with other ECM proteins (Engel, 1996). Fibronectin has a modular structure composed of type I, II, and III modules, and the alternative spliced EDA, EDB, and IIICS regions (Hynes, 1990). Although certain functional properties like cell binding may reside in short stretches of amino acid motifs within single modules, such as the RGD integrin binding site, collagen binding requires the involvement of both type I and type II modules (Skorstengaard et al, 1994). Thus, several similar or different modules may together form functional domains. Identical or highly homologous motifs have been identified in a variety of different proteins - modules with homology to the fibronectin type I modules are present in tissue-type plasminogen activator (t-PA), the blood coagulation factor XII, and hepatocyte growth factor activator (Bork et al, 1996). Fibronectin type II-like modules are present in the gelatinases A and B (Wilhelm et al, 1989; Collier et al, 1988), the bovine seminal fluid proteins PDC-109 and BSP-A3 (Seidah et al, 1987; Esch et al, 1983), and several other proteins. From an evolutionary standpoint, there are strong indications that duplication, deletion, or exchange of modules within or between proteins occurred as a result of exon-shuffling and that this development is primarily a feature of evolutionarily younger proteins of eukaryotes (Patthy, 1996). Exon shuffling has likely played a major role in the development of proteases 48 Introduction of the fibrinolytic, blood coagulation, complement cascades (Patthy, 1993). The zinc metalloprotease matrixins also have a conserved modular structure (Fig. 1.2) that has lent support to the conclusion that exon shuffling was important in the evolution of this family of proteinases from a common ancestral protease (Patthy, 1994; Patthy, 1991). Ligand Binding by Fibronectin Type II-like Modules. The observation was made that a 42 kDa proteolytic fragment of fibronectin that contains four type I modules and the two only type II modules in the molecule ( I g - I I j - I L T y - I g -I9) binds denatured type I collagen (Litvinovich et al., 1991; Engvall et al., 1981; Balian et al., 1979). This property has been valuable in subsequent studies aimed at elucidating structural elements important to this function. It remains controversial whether type II modules alone can bind gelatin. Owens and Baralle (1986) found among recombinant fibronectin collagen binding domains of varying length that a 14 amino acid flanking sequence from the adjacent type I module (I7) was required for binding of the type II modules to gelatin. In agreement with this result, other investigators observed that a minimum fragment of Ig - I^-I^-Iy was needed to achieve binding to gelatin Sepharose by affinity chromatography (Skorstengaard et al., 1994). In contrast, E. coli 6-galactosidase fusion proteins containing only type II modules from fibronectin or PDC-109 bound immobilized gelatin (Banyai et al., 1990). These studies point to a need for domain-domain or module-module interactions between type I and type II modules in fibronectin for binding to gelatin. In contrast to fibronectin, isolated gelatinase A and B fibronectin type II-like modules alone can bind native and denatured type I collagen. Recombinant proteins that consisted of triple- (Banyai et al., 1994; Banyai and Patthy, 1991), double-, or single-modular (Banyai et al., 1994) constructs of gelatinase A fibronectin type II-like modules all bound to gelatin, although with different affinity. Strongest gelatin affinity was observed for the tri-modular constructs, followed by the combinations of modules 2+3, 1+3, 1+2, and with very low binding of individual modules. Especially module 1 of gelatinase A was found to bind weakly. There is also variation in affinity for gelatin among the gelatinase B fibronectin type 49 Introduction II-like modules with strong binding observed for module 2, but a very low contribution of module 1 to the binding (Collier et al, 1992). Recent work has also shown that the tri-modular collagen binding domain from gelatinase A possesses not only binding properties for gelatin but also for native and denatured types I, V, and X collagens, denatured type II and IV collagens, elastin, and heparin (Abbey et al, 1997; Overall et al, 1997'; Chapter 2, Steffensen et. al, 1995). The results also support the concept of module cooperation in interaction with enzyme substrates and other extracellular matrix proteins by the observation that truncated collagen binding domains consisting of modules 1+2 or 2+3 have reduced collagen binding properties and do not bind elastin (Abbey et al, 1997; Overall et al, 1997). There is some uncertainty regarding the need for intact disulfide bonds between the four conserved Cys residues that are present in each fibronectin type II module and have the potential to form two bonds per module. Balian et al (1979) found that a 42 kDa proteolytic collagen binding fragment of fibronectin that had been reduced only or reduced and carboxymethylated did not bind to a gelatin-Sepharose column. However, when the untreated fragment was bound on the gelatin-Sepharose column it could not be eluted with dithioerythreitol (DTT) as a reducing agent. This suggests that the disulfide bonds in fibronectin are required for ligand binding but once bound, the protein-protein binding stabilizes the conformation. Reduction and alkylation, and reduction alone, also abolished gelatin binding for recombinant collagen binding peptide containing the two fibronectin type II modules (Owens and Baralle, 1986) and the proteolytic ally generated 42-kDa fibronectin fragment (Isaacs et al, 1989), respectively. In our studies of the gelatinase A rCBD, we found that reduction and carboxymethylation, but not reduction alone, abolished the binding to type I collagen {Chapter 2, Steffensen et al, 1995). Later, it was shown that CD spectra of gelatinase A CBD after reductive cleavage and alkylation of the disulfide bonds in guanidinium are characteristic of an unordered, denatured form of the protein (Banyai et al, 1996). Together, it is likely that disulfide bonds have a role in providing conformational integrity but that the type U modules are structurally robust, in particular when associated 50 Introduction with a ligand. To determine residues in fibronectin which are important to gelatin binding, chemical treatment to block the potential contribution of specific residues to the binding interaction showed that Lys as well Glu and/or Asp residues profoundly affect the binding to gelatin (Isaacs et al., 1989). However, treatment of Tyr residues had no such effects in these studies. Collier et al. (1992) found that module 2 of gelatinase B CBD bound strongly to gelatin compared to the other two modules and, therefore, analyzed the contribution by individual amino acid residues of module 2 to binding by alanine scanning mutagenesis. Substitution of a Cys residue in the third 6-strand abolished the binding, pointing to contribution of disulfide bonds for maintaining proper functional folding of the protein. Next, mutations were performed to alter residues located in the two predicted loop structures that connect B-strands 2 and 3, and 3 and 4, respectively. According to Constantine et al. (1992), these loops enclose in PDC109/b a shallow depression with a hydrophobic surface that potentially may serve as the gelatin binding site. Single amino acid substitution with Ala in place of selected, conserved Arg, Asp, and Tip residues in the loop connecting B-strands 2 and 3, followed by Asn and Tyr residues (not conserved) in the loop connecting B-strands 3 and 4 resulted in a decrease or complete loss of binding to gelatin (Collier et al., 1992). The changes in gelatin binding properties were primarily attributed to alterations of the binding site resulting from the substitutions. Interestingly, while the wildtype gelatinase B CBD module 1 did not bind gelatin, substitution of two pairs of amino acid residues in module 1 to those present in the high-affinity module 2 recovered gelatin binding. One pair of residues (Leu, Pro) was in the loop connecting B-strands 2 and 3 of module 1 whereas the other pair (Ala, Ala) was located NH2-terminal to 8-strand 1 in close proximity to the predicted position of Lys263 of gelatinase A module 2. (The effects of substitution of Lys263 in gelatinase A CBD with alanine are described in Chapter 3). Thus, single amino acid residues contribute to the overall gelatin binding properties of the individual modules. In addition, the recently proposed structure for the CBD of 51 Introduction gelatinase A (Banyai et al., 1996) emphasized the importance of understanding the functional properties of the complete tri-modular CBD. Thus, in monitoring changes by circular dichroism spectral analyses during guanidinium hydrochloride-, urea-, or thermally-induced unfolding, the tri-modular CBD was found to be significantly more stable than individual or bi-modular proteins. Therefore, individual amino acid substitutions might affect the trimodular CBD differently than was observed for single modules. Ligand Binding Domains and Functional Properties of MMPs Expression of recombinant MMPs with domain deletions, as well as expression of isolated domains or modules, has made it possible to assign functional contributions to distinct domains or modules (Fig. 1.4). The COOH-terminal domains of several MMPs have been shown to provide binding to type I collagen. A purified isolated autocatalytic carboxyterminal fragment of collagenase has been shown to bind native type I collagen (Bigg et al, 1994). This interaction concurs with the observations that catalytic domains alone from human fibroblast and neutrophil collagenases do not cleave native type I collagen (Schnierer et al., 1993; Murphy et al., 1992; Windsor et al, 1991; Clark and Cawston, 1989) and points to an important function for the COOH-terminal domains of these enzymes in cleavage of collagen in the triple helical conformation. In comparison, recombinant truncated human collagenase and stromelysin-1 with deletion of the COOH-terminal domains maintained their ability to degrade casein, gelatin, and a peptide substrate (Murphy et al, 1992). Because stromelysin also can bind to collagen by its COOH-terminal domain, a hybrid enzyme was engineered which consisted of the catalytic domain of collagenase combined with the COOH-terminal domain of stromelysin-1. The catalytic properties of this hybrid collagenase corresponded to those of the collagenase catalytic domain alone but, in spite of binding to native type I collagen, the hybrid did not have the capability to cleave this substrate. This result confirmed the importance of the collagenase COOH-terminal domain in positioning the cleavage site of the triple helical type I collagen molecule with a precise orientation relative to the active 52 Fibronectin Denatured collagens GAGs Native type I, V, X TIMP-2 collagens a v & 3 integrin Procollagens Heparin Type IV ^ Heparan-sulfate collagen? /^ ffipV Elastin \ ^ I (V) 0 1 V 3 il S^^ "':r ^ C-domain I M ^ H Linker The Collagen Binding Domain ^^^^ w (rCBD123) ^ Catalytic domain FIG 1.4. Ligand binding properties of gelatinase A domains. This figure summarizes the localization of confirmed and potential sites of ligand interaction for gelatinase A (see also text). Chapter 2 describes the results of investigations of ligand binding properties for the collagen binding domain of gelatinase A. 53 Introduction catalytic site of collagenase (Murphy et al., 1992). Various deletions in the COOH-terminal domain of neutrophil collagenase have pointed to a 16 amino acid sequence as the critical region for collagen binding, while a larger 62 amino acid region affected the efficiency of the collagenolytic activity (Hirose et al., 1993). Contrary to that observed for the collagenases, the COOH-terminal domain of gelatinase A has little or no contribution to the collagen binding properties of the enzyme as determined in binding assays using isolated rC domain (Overall et al, 1997). This explains why the gelatinase A COOH-terminal domain has negligible significance for catalysis of collagen as gelatinase A with COOH-terminal domain deletions retains full enzymatic activity for this substrate (Fridman et al, 1992). The isolated rC domain of gelatinase A also binds plasma fibronectin, which is a substrate of the enzyme, and heparin (Overall et al, 1997; Wallon and Overall, 1997). Tissue inhibitors of metalloproteinases form complexes with activated MMPs by interactions with both the catalytic domain and the COOH-terminal domain (Bigg et al, 1994; Birkedal-Hansen et al, 1993). In addition, the gelatinases in latent forms can also bind TIMPs as exemplified by progelatinase A that complexes with TPMP-2 (Willenbrock et al, 1993; Fridman etal, 1992; Ward etal, 1991; Goldberg etal, 1989) and progelatinase B that can complex with TTMP-1 (Wilhelm et al, 1989). The importance of the MMP COOH-terminal domains in binding of TIMPs is illustrated by a reduced sensitivity of gelatinase A to inhibition by TTMP-2, and lower affinity of stromelysin-1 for TTMP-1, for COOH-terminal domain deletion mutants of these enzymes (Baragi et al, 1994; Fridman et al, 1992). This is supported by the finding that isolated gelatinase A rC domain binds TTMP-2 (363; Wallon and Overall, 1997; Fridman et al, 1992). The role of the COOH-terminal domain in binding TIMP-2 is further important to cell membrane activation of gelatinase A (Cao et al, 1995), an event which does not occur in COOH-terminal domain deletion mutants (Ward et al, 1994; Murphy et al, 1992). Little is known about the function of a 54-amino acid proline-rich domain that has 54 Introduction some homology to the collagen chain oc2(V) and is only found in gelatinase B (Wilhelm et al., 1989). However, gelatinase B with deletion of the type V collagen-like domain cleaved denatured type I collagen, as well as native types V and XI collagen (Pourmotabbed, 1994). There is substantial evidence that in gelatinase A and B the predominant collagen binding properties reside in the CBD (Chapter 2, Steffensen et ah, 1995, Murphy et al., 1994, Banyai et al, 1994, Collier et al, 1992, Banyai and Patthy, 1991, Wilhelm et al, 1989, Collier et al, 1988). This domain is composed of three fibronectin type II-like modules which have maintained the ability to bind denatured type I collagen that is observed in fibronectin (Banyai et al, 1990). Characterization of isolated rCBD from human gelatinase A has demonstrated that this domain accounts for most, if not all, of the native and denatured type I collagen binding properties of human gelatinase A and also provides the enzyme with binding to denatured types TV; V, and X collagens, native types V and X collagens, elastin, and heparin (Abbey et al, 1997; Overall et al, 1997; Chapter 2, Steffensen et al, 1995). The importance of the CBD to substrate binding has been illustrated with CBD deletion mutants of gelatinase A. The deletion of CBD led to loss of binding to type I collagen but not to TTMP-1. Although the catalytic activity in this mutant enzyme corresponded to that of the wildtype gelatinase A for a peptide substrate, the activity against casein was reduced by 50% and against type I collagen by 90% (Murphy et al, 1994). A CBD deletion mutant of gelatinase B also abolished the type I gelatinolytic activity (Pourmotabbed, 1994). These results stress the importance of CBD for substrate positioning and successful cleavage. Interestingly, both the wildtype and the truncated gelatinase A were activated by Con A treated fibroblasts suggesting that the CBD is not directly required for membrane-mediated activation (Murphy et al, 1994). An analogy to the loss of ability to cleave collagen in the absence of the CBD is found with respect to elastin cleavage. The CBD provides gelatinase A with elastin binding properties (Chapter 2, Steffensen et al, 1995) but, after deletion of the CBD, gelatinases A or B neither bound gelatin nor retained elastinolytic activity (Shipley etal, 1996). 55 Introduction PROBLEM STATEMENT AND RATIONALE FOR THESIS RESEARCH Considerable evidence has linked MMP expression by tumor cells with the processes of tumor expansion and invasive growth leading to metastasis. In addition, MMPs are important in processes of tissue remodeling and development. Due to its widespread expression in human tissues and its capacity to degrade several extracellular matrix components, gelatinase A is considered an important component in these processes. However, little is known about the interactions between gelatinase A and known substrates of the enzyme. Therefore, discerning the molecular determinants of substrate specificity and ligand interactions is important for understanding the role of this enzyme in both physiological and pathological processes. Type II modules in fibronectin contribute to its gelatin binding properties. Gelatinase A contains a domain composed of three fibronectin type II-like modules. Therefore, the hypothesis was that this domain in gelatinase A also binds gelatin. To precisely identify whether this specific domain was involved in gelatin binding, the domain was expressed as a recombinant domain in E. coli. Since gelatinase A degrades several additional extracellular matrix components, the characterization of ligand binding interactions of the domain was extended to also study interactions with other extracellular matrix components. The experiments established that several collagen types, elastin, and heparin bound specifically to the recombinant domain of gelatinase A, then designated the collagen binding domain. The structural importance of collagens and elastin in the extracellular matrix is well established and heparin has important biological functions. Moreover, heparin is structurally similar to heparan sulfate which is a biologically important side chain of several proteoglycans including perlecan found in basement membranes. By its ability to degrade other basement membrane components, including type IV collagen, the CBD interactions with heparin might reflect interactions with heparan sulfate during basement tumorigenic membrane degradation by gelatinase A. However, the molecular basis for gelatinase A binding of collagen, elastin, and heparin is not known. Since lysine residues have been 56 Introduction identified in heparin binding sites of several proteins and since the lysine residues in fibronectin are crucial to ligand binding by type II modules, chemical treatment of the CBD as well as site-specific mutagenesis were applied to define the potential contributions of individual lysine residues to gelatinase A CBD ligand binding properties. It has become clear that events in the vicinity of the cell surface are important to proteolytic events involving gelatinase A. Although gelatinase A cell surface binding occurs via both membrane type metalloproteinases and an integrin receptor, the mechanisms of binding are not fully understood. Because native type I collagen and heparin bound specifically to the CBD, the possibility was explored that extracellular matrix components could mediate cell surface localization of gelatinase A via interactions with the CBD. A cell attachment assay was applied to define the interactions of the cell surface components with the CBD because the assay permitted defined modification of conditions. This permitted analysis of individual pericellular and extracellular matrix components using competitive attachment assays, protein elimination by enzymic digestions, and receptor blocking antibodies. In addition, ligand blotting using solubilized cell extracts was used to define more precisely the behavior of interacting proteins. Thus, this series of experiments was designed to characterize the gelatinase A ligand interactions mediated by the CBD, to identify molecular determinants for the ligand binding, and to ascertain the potential role of the CBD in cell surface binding of gelatinase A. 57 CHAPTER 2 Extracellular Matrix Binding Properties of Recombinant Fibronectin Type II-like Modules of Gelatinase A. INTRODUCTION A central characteristic of metastatic tumor cells is their ability to degrade and penetrate basement membranes. Considerable evidence has linked elevated matrix metalloproteinase expression by many tumor cells with these processes (Liotta et al, 1980). Type IV collagen is the major structural component of basement membranes (Crouch et al, 1980; Timpl et al, 1994). Therefore, the type IV collagenolytic activity of either of two MMPs, the gelatinases A and B (MMP 2/72-kDa gelatinase and MMP 9/92-kDa gelatinase, respectively), is viewed as a critical component of the metastatic process. The gelatinases A and B are also important in other processes such as embryogenesis and tissue remodelling (reviewed (Overall, 1991; Matrisian, 1992; Sodek and Overall, 1992)), osteoclastic activity (Reponen et al, 1994), enamel formation (Overall and Limeback, 1988), cytotrophoblast invasion (Fisher et al, 1989), and lymphocyte cell migration (Weeks et al, 1993). These enzymes also assist in completing the collagenolytic cascade by degrading the three denatured a-chains of cleaved collagen. Accordingly, discerning the molecular determinants of substrate specificity of gelatinase A is important in understanding the role of this enzyme in physiological and pathological processes. MMPs share a basic primary and tertiary structure (Birkedal-Hansen et al, 1993; Murphy and Docherty, 1992; Matrisian, 1992; Blundell, 1994). MMPs have a highly conserved Zn -^binding histidine triad and a Ca -^binding motif in the catalytic domain. A free cysteine in a highly conserved sequence in the prodomain coordinates with the catalytic zinc (II) ion and is responsible for enzyme latency (Van Wart and Birkedal-Hansen, 1990; Springman et al, 1990). A hemopexin/vitronectin-like carboxyl domain binds the specific tissue inhibitors of MMPs (TIMPs) (Fridman et al, 1992; Howard and Banda, 1991; 58 Human gelatinase A collagen binding domain Goldberg et al, 1989; Murphy et al, 1992) and, in gelatinase A, binds cell membranes (Murphy et al, 1992) on Con A-activated fibroblasts (Murphy et al, 1992; Overall and Sodek, 1990; Overall and Sodek, 1992). The hemopexin/vitronectin-like carboxyl domains of collagenase and stromelysin also bind native type I collagen (Murphy et al, 1992; Windsor et al, 1991; Clark and Cawston, 1989). Removal of this domain from collagenase ablates collagenolysis but not catalytic competence — that is, the truncated collagenase lacking the carboxyl domain still degrades synthetic peptide substrates and casein, but not native type I collagen (Windsor et al, 1991; Clark and Cawston, 1989). In contrast, the hemopexin-like domain of gelatinase A does not bind collagen (Murphy et al, 1994). In addition to the main structural elements characteristic of the MMPs, both gelatinase A and gelatinase B contain 3 tandem copies of a 58-amino acid residue fibronectin type II-like module positioned immediately N-terminal to the zinc binding site (Wilhelm et al, 1989; Collier et al, 1988). Fibronectin, a modular extracellular matrix glycoprotein, is composed of repeating homologous domains (Type I, U, and III) that bind a number of extracellular matrix proteins, fibrin, and cells (Ruoslahti, 1988). Although it is controversial whether fibronectin type II modules alone bind denatured type I collagen (Skorstengaard et al, 1994; Ingham etal, 1989; Owens and Baralle, 1986; Ruoslahti, 1988), Banyai etal (1990) reported that a recombinant type II module from fibronectin and a type II-like module in bovine seminal fluid protein PDC-109 binds gelatin. Subsequent work demonstrated that the type II-like modules in both gelatinase A (Banyai et al, 1994; Banyai and Patthy, 1991) and gelatinase B (Collier et al, 1992) also bind denatured type I collagen. The individual fibronectin type II-like modules in gelatinase B (Collier et al, 1992) and gelatinase A (Banyai et al, 1990) show differential binding specificities for denatured type I collagen. Therefore, binding specialization of the different fibronectin type II-like modules in gelatinase A may also have occurred to generate exosites specific for the other collagens and extracellular matrix molecules degraded by the enzyme including native types IV, V, VII, and X collagens, elastin, and fibronectin (Collier et al, 1988; Seltzer et al, 1990). 59 Human gelatinase A collagen binding domain To further understand the function of the structural elements of gelatinase A, we have characterized the binding properties of the fibronectin-like domain of human gelatinase A to a number of the enzyme's substrates, reconstituted basement membrane, TIMP-1, and other extracellular matrix proteins. Reported here are experiments which establish that, in addition to binding denatured type I collagen, a recombinant fibronectin-like domain from human gelatinase A, encompassing all three type II-like modules, binds with high affinity to denatured types IV and V collagens and elastin. Although human gelatinase A cleaves native type IV collagen but not native type I collagen, surprisingly, the fibronectin-like domain avidly binds native type I collagen but not native type IV collagen or other basement membrane components. Thus, in addition to fulfilling the criteria as an exosite for a number of substrates, the fibronectin-like domain of gelatinase A may have an ancillary role as an extracellular matrix localization domain by virtue, in particular, of its native type I collagen binding properties. EXPERIMENTAL PROCEDURES Extracellular Matrix Proteins, Antibodies, and Chromatography Media—Acid soluble native type I collagen was prepared from rat tail tendons as described by Piez (Piez, 1967) by extraction with 0.5 M acetic acid and differential precipitation with 1.7 M NaCl. Pepsin-treated type I collagen was prepared by digestion of the acid-soluble type I collagen with pepsin (Sigma) at pH 2.0, 4 °C for 20 h, then precipitated with 1.7 M NaCl, redissolved in 0.15 M acetic acid, and lyophilized. Gelatin was prepared from acid-soluble type I collagen (non-pepsin treated) by heat denaturation at 56 °C for 30 min. [14C]Glycine-labeled type I collagen, with a specific activity of 3.5 x 108 dpm/mg, was prepared by metabolic labelling and purified from conditioned cell medium by pepsin-digestion and NaCl precipitation as described previously (Overall and Sodek, 1987). To confirm the native collagen content of the metabolically-labeled preparation, the labeled type I collagen was incubated with 0.1 or 0.01 ug/ml trypsin (type XII bovine pancreas, Sigma) (enzyme to substrate ratio -1:2 and 60 Human gelatinase A collagen binding domain 1:20) for 19 h at 20 °C. Intact protein was then precipitated in 10% (w/v) trichloroacetic acid/1% (w/v) tannic acid for 2 h at 0 °C and the pellets collected by centrifugation at 10,000 x g for 20 min at 0 °C (Overall and Sodek, 1987). The trypsin-digested denatured type I collagen content was determined by scintillation counting of the trichloroacetic acid/tannic acid-soluble protein fraction and calculated to constitute <9% of the radiolabeled type I collagen preparation. Fibronectin was purified from human serum by affinity chromatography over gelatin-Sepharose (Dedhar et al., 1993). After extensive washes with loading buffer (50 raM Tris, 0.2 M NaCl, pH 7.4), non-specifically bound proteins were eluted with 1 M NaCl. Bound fibronectin was eluted with 10% (v/v) dimethyl sulfoxide (DMSO) in loading buffer, analyzed for purity by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), and quantitated by spectroscopy at 280 nm. Cyanogen bromide al(I) collagen fragments 2, 7, and 8 and SPARC (secreted protein which is acidic and rich in cysteine) were gifts from Dr. J. Sodek (University of Toronto, Toronto, Ontario, Canada). Elastin and horse heart myoglobin were from Sigma; laminin and murine type IV collagen were obtained from Gibco BRL; human type V collagen and Matrigel® were from Becton Dickinson. TLMP-1 was from by Dr. T. Cawston (Aldenbrooks Hospital, Cambridge, UK) and human gelatinase A cDNA (hl91) (Huhtala et al., 1990) was provided by Dr. K. Tryggvason (University of Oulu, Oulu, Finland). Gelatin-Sepharose 4B, heparin-Sepharose CL-6B, CM-Sepharose Fast Flow, and chelating Sepharose 6B were from Pharmacia LKB Biotechnology. Affi-Gel 10 and 15 and goat anti-rabbit polyclonal IgG (H+L) antibody conjugated with alkaline phosphatase were from Bio-Rad. Polyclonal antibody to gelatinase A was generously provided by Dr. W. Stettler-Stevenson (The National Cancer Institute, NIH, Bethesda, MD). All other reagents were commercially available analytical grade reagents. DNA Amplification by the Polymerase Chain Reaction—DNA corresponding to exons 5, 6, and 7 of human gelatinase A, which code for the three fibronectin type II-like modules, was 61 Human gelatinase A collagen binding domain amplified as one product from cDNA h 191 by the polymerase chain reaction. The amplified DNA was defined by the 5' exon/intron border of exon 5 and the 3' border of exon 7. The 5' primer included a Nhel site for directed cloning (5'ATCGCTAGCGTGGTCCGTGTGAAG TATGG3'). Two stop codons and a PstI restriction site were included in the 3' primer (5'AGGTACCTGCAGTCATTATTGGTCAGGGCAGAAGCCCCACTT3'). Expression Vector Construct—The pGYMX vector, a modified form of pGYM (Guillemette et al, 1991) derived from pEMBL18, contains part of the ribosomal binding site of the phage T7 gene 10 leader sequence (Olins et al, 1988). The expressed recombinant protein is a fusion product that includes an N-terminal 16-amino acid flanking sequence (MATS(H)6IEGRAS) comprised of an N-terminal initiating methionine, a metal ion binding polyhistidine tract, a Factor Xa specific cleavage site (IEGR), and the residues alanine and serine coded for by the Nhel site (Eltis et al, 1994). The 549 bp PCR product was purified from the PCR reaction mix by Magic PCR Preps (Promega) and the insert and vector were digested with PstI and Nhel. Gel-purified vector and PCR DNA were then ligated overnight at 15 °C using T4 DNA ligase and 10 mM rATP (Gibco BRL). The fidelity of the PCR and ligation reactions used to produce the construct pGYMX123 were verified by double-stranded chain-termination DNA sequencing (Sanger et al, 1978). Bacterial Expression and Purification of rCBD123—Escherichia coli, transformed with pGYMX123, were grown in SB (1% (w/v) trypticase peptone (Becton Dickinson), 0.8% (w/v) yeast extract (Becton Dickinson), 0.5% (w/v) NaCl, pH 7.4) with ampicillin (100 ug/ml), harvested by centrifugation, and rinsed with 20 mM NaP;, 0.5 M NaCl, 1 mM EDTA, pH 8.0. The rinsed bacterial pellet was resuspended in 10 volumes of rinse buffer containing 5 ug/ml DNase (Boehringer-Mannheim) and 1.0 mM phenylmethylsulfonyl fluoride (Sigma) and lyzed with 1 mg/ml lysozyme (Sigma) for 2 h at 37 °C. Inclusion bodies were pelleted by centrifugation at 10,000 x g, 4 °C for 15 min and, after washing, dissolved in guanidine-hydrochloride buffer (6 M GuHCl, 0.1 M NaPi5 10 mM Tris, pH 8.0) under non-reducing conditions to favor dissolution of only non-intermolecularly crosslinked recombinant protein. 62 Human gelatinase A collagen binding domain SDS-PAGE confirmed that the residual pellet predominantly contained the intermolecularly crosslinked recombinant protein fraction. To facilitate disulphide bond formation required for correct protein folding, the extracted denatured protein was dialyzed against aerated 0.1 M sodium borate buffer, pH 10.0 for 2 h at 20 °C (Luck et al, 1992). The guanidinium was then removed and the pH lowered by dialysis against chromatography buffer (100 mM NaPj, 0.5 M NaCl, pH 8.0) at 4 °C. The refolded soluble protein was separated from denatured aggregates by centrifugation at 10,000 x g for 15 min at 4 °C. rCBD123 was purified by chromatography over a Zn^ -chelate column (Vt 2 ml) equilibrated in chromatography buffer at 4 °C. After sample loading and extensive washes in chromatography buffer, non-specifically bound bacterial protein was removed with 1.0 M NaCl and the bound protein eluted with 250 mM imidazole in chromatography buffer. Fractions containing recombinant protein were pooled and applied to a gelatin-Sepharose 4B column (Vt 10 ml). Non-specifically bound material was eluted with 1.0 M NaCl and the gelatin-binding recombinant protein eluted with 7% (v/v) DMSO in 50 mM Tris, pH 7.4. The eluted protein fractions were equilibrated against 50 mM Tris, pH 7.4 by dialysis to remove the DMSO, quantitated spectrophotometrically and by the Bradford protein assay (Bio-Rad), and then stored frozen at -70 °C until use. Only that fraction of the rCBD123 preparation which bound gelatin-Sepharose was used in subsequent characterization. Affinity Chromatography—To evaluate potential interactions of rCBD123 with extracellular matrix proteins, affinity columns of native type I collagen, pepsin-treated native type I collagen, native type IV collagen, laminin, and myoglobin as a control were prepared with Affi-Gel 10 and 15 according to the manufacturers instructions and, typically, at two pH conditions. Types I and IV collagens and pepsin-treated type I collagen coupled best to Affi-Gel 10 in 0.1 M acetic acid with a 3 h coupling reaction at 4 °C. Non-specific binding sites were blocked with 1.0 M ethanolamine, pH 8.0 at 4 °C for 2 h. Gelatin-Sepharose 4B and heparin-Sepharose CL-6B were also used for the assays. Affinity chromatography was performed using mini-columns (Vt 25 ul) in 50 mM Tris 63 Human gelatinase A collagen binding domain buffer, pH 7.4 as described previously (Overall et al., 1989). A standardized amount of rCBD123 (40 ug, 1.9 nmol) in 100 ul was loaded on each column and experiments under each condition were typically performed in triplicate at least twice. Elution strategies included step-gradients of NaCl from 0.1 to 1.0 M and DMSO from 1 to 10% (v/v) in 50 mM Tris buffer, pH 7.4. Washes (20 - 40 x V,) with the appropriate chromatography buffer were performed after sample loading and between different eluants. Fractions were collected and analyzed by SDS-PAGE at a constant dilution relative to the loaded sample volume to facilitate assessment of binding as described in detail below. Columns eluted with DMSO were not reused. The specificity of rCBD123 to collagen interaction was also assessed by competition assays. Typically, rCBD123 was incubated for 90 min at 20 °C with the competing ligand at various mole ratios between 1:0 to 1:9 before loading the reaction mixture onto either gelatin-Sepharose or native type I collagen affinity columns and eluting as described. The denatured and native type I collagen binding properties of gelatinase A were assessed by affinity chromatography of the enzyme in 1.5 ml aliquots of conditioned culture medium from rat osteoblastic cells (Overall, 1995). Chromatography and elution conditions were as described above for rCBD123. To ascribe the type I collagen binding properties of the parental gelatinase A to a specific domain, the enzyme was bound to native type I collagen columns and the binding then competed with rCBD123 (40 ug; 2 nmol) in chromatography buffer. Elutes were assayed for gelatinase A by enzymography as described below. Reduction and Carboxymethylation ofrCBD123—To reduce intramolecular disulphide bonds in rCBD123, DTT was added to rCBD123 to a final concentration of 65 mM (>100-fold moles of molecules excess per disulphide bond) for 30 min at 20 °C. This concentration of DTT was maintained in all buffers in column assays. rCBD123 was also reduced and carboxymethylated according to Hollecker (1990) (Creighton, 1990) and Creighton (1990) (Hollecker, 1990), modified as follows; rCBD123 was first equilibrated in denaturation buffer (8.0 M urea, 0.5 M Tris-HCl, 2 mM EDTA, pH 8.1) by gel filtration on a 10 DG column (Vt 64 Human gelatinase A collagen binding domain 10 ml) (Bio-Rad) and then reduced by addition of 100-fold molar excess of DTT (65 mM) over the estimated disulfide bond content and incubated at 50 °C for 1 h. After cooling to 20 °C, the alkylating agent, iodoacetic acid, was added to a 2-fold molar excess over DTT (130 mM) and reacted at 20 °C for 30 min. The reduced and carboxymethylated protein was then equilibrated in 50 mM Tris, pH 7.4 by chromatography over a 10 DG column and assessed by SDS-PAGE. Column assays with reduced and carboxymethylated rCBD123 were performed in the absence of reductant under buffer conditions identical to those described for non-reduced rCBD123. SDS Polyacrylamide Gel Electrophoresis and Enzymography—Proteins were separated by SDS polyacrylamide gel electrophoresis according to Laemmli (1970). Protein samples were analyzed without reduction or with the addition of 65 mM DTT and heating at 95 °C for 5 min. Gels were stained with Coomassie brilliant blue R250 at 42 °C and protein bands were quantitated by laser densitometry at 633 nm (LKB Ultrascan XL). For enzymography, non-reduced protein samples were electrophoresed in 10% (w/v) polyacrylamide gels containing 100 ug/ml heat-denatured acid-soluble type I collagen. Gels were processed as previously described (Overall and Limeback, 1988). Briefly, after electrophoresis gels were equilibrated in 5% (v/v) Triton X-100, incubated in assay buffer (50 mM Tris, 200 mM NaCl, 5 mM CaCL.) for 2-4 h at 37 °C, and the cleared bands, identifying the position of gelatinase A, revealed by counterstaining of the gelatin in the gels by Coomassie brilliant blue R250. Reduced molecular weight markers used were rabbit muscle phosphorylase B (97 kDa), BSA (67 kDa), chicken egg ovalbumin (43 kDa), bovine carbonic anhydrase (29 kDa), horse heart myoglobin (18.8 kDa), chicken egg-white lysozyme (14.4 kDa), and bovine insulin (6.2 kDa) (Sigma). Western Blot Analysis—Proteins were transferred to Immobilon-P PVDF membranes (Millipore) after separation by SDS-PAGE. Transferred rCBD123 was then reacted with a polyclonal anti-gelatinase A antibody diluted 1/1,000 in TBS/Tween with 1% (w/v) BSA for 65 Human gelatinase A collagen binding domain 1 h, washed, and conjugates detected using enhanced chemiluminescence (ECL) reagents and Hyperfilm (Amersham). Microwell Substrate Binding Assay—To screen for potential interaction with rCBD123, a number of known substrates of gelatinase A as well as other extracellular matrix proteins were coated as films in 96-microwell plates. Proteins included native type I collagen, pepsin-treated native type I collagen (telopeptide-free), and heat-denatured type I collagen (gelatin); collagen a 1(1) cyanogen bromide fragments 2, 7 and 8; native and heat-denatured types IV and V collagens; elastin, Matrigel®, laminin, fibronectin, SPARC, and TTMP-1. Myoglobin and BSA served as negative control proteins for the assays. Microliter plates were coated overnight at 4 °C with 10 pmol protein/well (typically 1-5 ug) in coating buffer (15 mM NajCCv 35 mM NaHC03, 0.02% (w/v) NaN3, pH 9.6). Consistent and equal binding of protein coated to micro wells was confirmed by incorporation of [14C]glycine-labeled native and denatured type I collagens with the coating solutions as appropriate and scintillation counting of the unbound supernatants. Matrigel® and native type I collagen were also prepared as three-dimensional gels as follows. Gels of reconstituted basement membrane were produced with Matrigel® diluted 1:3 in phosphate buffered saline (100 ul/well) at 4 °C and incubated for 30 min at 37 °C to assure solidification. Type I collagen fibrillar gels were prepared by dissolving 10 pmol acid-soluble native collagen in 100 ul 50 mM Tris, 200 mM NaCl, pH 7.0 followed by incubation for 30 min at 37 °C. Plates coated with protein films or gels were rinsed and then blocked with 2.5% (w/v) BSA in phosphate buffered saline for 1 h at 20 °C. After further extensive rinses with phosphate buffered saline, serially diluted rCBD123 in 50 mM Tris, pH 7.4 (1 nmol to 0.125 pmoywell = 10 uM - 1.25 nM) was added for 1 h at 20 °C. The plates were rinsed thoroughly and bound rCBD123 detected with a 1/1,000 dilution of polyclonal antibody raised against human gelatinase A followed by reaction with a 1/5,000 dilution of alkaline phosphatase conjugated goat anti-rabbit antibody (BioRad). All antibody reactions and washes were performed in buffers containing 0.05% (v/v) Tween 20. For quantitation, p-nitrophenyl phosphate di-sodium (Sigma) was added as 66 Human gelatinase A collagen binding domain substrate and, to ensure linearity of the assay, the color intensity was determined at various times by measuring the absorbance at 405 nm in an automated ELISA plate reader (BioRad). Negative controls consisted of reaction mixture minus rCBD123, primary antibody, or secondary antibody. [14C]Glycine-Labelled Collagen Binding Assay—To determine the potential for simultaneous binding of two or more molecules of type I collagen by rCBD123, a sandwich type assay was employed. rCBD123:[14C]glycine-labeled type I collagen was quantitated in those rCBD123 complexes that could also bind unlabelled collagen films on microwell plates. Specifically, rCBD123 was serially diluted in 50 mM Tris, pH 7.4 from 10,000 to 0.6 nM and incubated with 4000 dpm native or heat-denatured [14C]glycine-labeled type I collagen (0.03 and 0.1 pmol, respectively) for 1 h at 20 °C. The reaction products were then transferred to microwell plates coated with 10 pmolAvell of unlabelled native or denatured type I collagen as appropriate. That is, to avoid ligand displacement by competing substrates rCBD123 that had been incubated with [14C]glycine-native type I collagen was transferred to plates coated with native type I collagen, and rCBD123 that had been incubated with [14C]glycine-labelled denatured type I collagen was transferred to denatured type I collagen-coated plates. After 1 h incubation at 20 °C, unbound material was removed by three rinses with 50 mM Tris, pH 7.4. Bound rCBD123:[14C]glycine-labeled type I collagen complexes were dissociated by addition of 10% (v/v) Me2SO in 50 mM Tris, pH 7.4 for 30 min and then transferred to Scintillant and quantitated by scintillation counting. Binding reactions were performed in duplicate for eight serial dilutions with appropriate controls for both native and denatured type I collagens. RESULTS Characterization of Recombinant Gelatinase A CBD123 Protein—To determine the contribution of the fibronectin-like domain in gelatinase A to the substrate binding and functional properties of the enzyme, the three fibronectin type II-like modules of human 67 Human gelatinase A collagen binding domain gelatinase A were expressed in E. coli as a single recombinant fusion protein, denoted as rCBD123. Approximately 18% of total E. coli protein was rCBD123, localized predominantly to inclusion bodies after 18 h of culture. Typically, the yield of purified gelatin binding rCBD123 from 1,000 ml E. coli culture was approximately 15 mg. Non-reduced rCBD123 electrophoresed with an apparent Mr of 21.1 kDa whereas the reduced protein migrated with an apparent Mt of 22.1 kDa (Fig. 2.1 A), identical to that predicted from the sequence. Reduced and carboxymethylated rCBD123 electrophoresed with an apparent Mrof 24.1 kDa. The 1.0 kDa decrease of apparent Mr in non-reduced compared with reduced samples indicated that rCBD123 was folded and contained intact disulfide bonds (Kaderbhai and Austen, 1985). To further substantiate that the purified protein was indeed the recombinant product, an antibody to human gelatinase A was found to react with the recombinant protein (Fig. 2. IB). Moreover, antibodies raised against the fibronectin type II-like module, to a synthetic peptide corresponding to the poly-His tract of the fusion protein, and to a peptide in the third fibronectin type II-like module, all reacted strongly with the purified protein (not shown). Recombinant protein in the gelatin-Sepharose DMSO elute was predominantly monomeric with low amounts of dimers (<4%) present as assessed by SDS-PAGE of non-reduced protein samples (Fig. 2.1 A). However, ECL-Western blots of the purified rCBD123, exposed in the non-linear range, also revealed the presence of relatively small quantities of multimers (Fig. 2.1B). The high percentage of monomeric rCBD123 attests to the efficiency of the purification and refolding strategy that sacrificed final yield for the selection of protein that had never entered the intermolecularly crosslinked pool. Multimeric species remained largely undissolved by the guanidinium or were separated later by precipitation during protein refolding steps. Recombinant CBD 123 had gelatin-binding properties that were evident from the specific binding of rCBD123 to gelatin-Sepharose, but not Sepharose G10. Unrelated recombinant proteins containing the same NH2-terminal fusion peptide expressed from the same vector did not bind to gelatin, excluding the possibility that binding was by the fusion 68 Human gelatinase A collagen binding domain peptide. Lastly, a ID 'H-NMR spectrum of rCBD123 revealed H N and H a chemical shift dispersion indicative of fi-sheet secondary structure, and several upfield shifted methyl resonances indicative of a folded structure (not shown). This indicated that the rCBD123 preparation was homogenous and had folded to form an organized compact domain. As for any recombinant protein, until the three-dimensional structure of the rCBD123 and the fibronectin-like domain in the context of the full length enzyme is known, it cannot be stated that the folding of the recombinant protein is identical to the natural protein. Therefore, to confirm the validity of our approach we further characterized the binding of the rCBD123 to various proteins for comparison with the properties of the full length parental enzyme. Characterization of rCBD 123 Binding to Denatured Type I Collagen—To investigate the nature of the binding of the rCBD123 to denatured type I collagen, the gelatin-Sepharose purified recombinant protein was applied to minicolumns (in triplicate) of gelatin-Sepharose and the binding characterized by differential elution (Fig. 2.2A). No recombinant protein was detected in the unbound or wash fractions from the columns indicating that all the rCBD123 was bound to the gelatin-Sepharose. Step elution with NaCl using 0.1 M increments from 0 to 1.0 M was ineffective in eluting the protein showing that electrostatic interactions were not solely responsible for the binding affinity. rCBD123 was eluted from gelatin columns by DMSO with a characteristic peak elution at -3% (v/v) DMSO indicating that hydrophobic interactions were important in the binding interaction or that disruption of the three-dimensional structure of the rCBD123 by DMSO caused the protein to dissociate. 8 M urea was also effective in eluting rCBD123 from gelatin-Sepharose (not shown). The specificity of the interaction between rCBD123 and gelatin was further verified by competition studies. When rCBD123 was pre-incubated with gelatin, subsequent binding to gelatin-Sepharose was inhibited in a concentration-dependent manner (Fig. 2.2B). At a rCBD123:gelatin mole ratio of 1:9, there was quantitative recovery of rCBD123 in the unbound and wash fractions. At lower mole ratios, progressively increasing amounts of rCBD123 were bound to the gelatin-Sepharose and could be eluted with DMSO. 69 Human gelatinase A collagen binding domain Disulphide Crosslinks in rCBD123 are Not Required for Gelatin Binding—The fibronectin type II-like modules of gelatinase A each contain four cysteines. Thus, the fibronectin-like domain has the potential to form six disulfide bonds. As indicated by the shift in MT on reduction (Fig. 2.1), rCBD123 was judged to contain intact disulphide bonds. To determine the structural requirement of intact disulfide bonds for gelatin binding, we examined the interaction of rCBD123 to denatured type I collagen after reduction with DTT. To prevent reformation of the disulphide bonds after reduction, a > 100-fold mole excess of DTT was maintained in all chromatography buffers. This treatment had no effect on the binding of rCBD123 to denatured type I collagen. Further, the reduced rCBD123 showed an elution profile from gelatin-Sepharose identical to that of non-reduced rCBD123 (not shown, n=3). Therefore, it appears that rCBD123 can retain a biologically functional folded structure at 22 °C in the absence of disulphide crosslinks. However, when rCBD123 was reduced and carboxymethylated, gelatin binding was lost (Fig. 2.1) indicating that either structural perturbation or alterations in charge introduced by the carboxymethylation of cysteines were responsible for the loss of gelatin binding properties. rCBD123 Binds Native Type I Collagen—Human gelatinase A does not cleave nor has it been reported to bind native type I collagen. However, since fibronectin binds native type I collagen near the tissue collagenase cleavage site (Kleinman et al, 1981; Kleinman et al, 1976; Guidry et al, 1990), we assessed rCBD123 for binding to native type I collagen. Using microwell substrate binding assays with native type I collagen coated as a two-dimensional film, rCBD123 was found to bind to native type I collagen in a saturable manner (Fig. 2.3). In comparison, binding was somewhat stronger to denatured type I collagen with an apparent Kd in the uM range. In these experiments, an equal number of moles of molecules of denatured collagen a-chains and native type I collagen were coated per well. Since there are three a-chains per triple helical molecule of native type I collagen, each well of native type I collagen contained three times the number of moles of a-chains as the denatured collagen coated wells. Thus, together with the slightly lower binding of rCBD123 70 Human gelatinase A collagen binding domain to native type I collagen, this showed that there were at least three times fewer binding sites per mole of a-chains when in the native triple helical conformation than in the denatured conformation. This indicates that the triple helical structure masks binding sites on the constituent a-chains. Even though identical amounts of native type I collagen were used to form fibrils (10 pmol), a substantially higher amount of rCBD123 bound to the fibrillar gel form than to either native or denatured type I collagen coated as films (Fig. 2.3). These results may reflect improved access to binding sites on all faces of the type I collagen which are otherwise occluded when the collagen is adhered to plastic surfaces. • . To confirm that rCBD123 bound native type I collagen in solution, affinity chromatography on mini-columns coupled with native type I collagen was performed (Fig. 2.4A). No rCBD123 was detected in the wash fractions, only a trace was found in the 1.0 M NaCl elute, and essentially all the rCBD123 was recovered using DMSO in the chromatography buffer with a peak elution at -2% (v/v) DMSO. Notably, this concentration of DMSO was reproducibly lower than that required to elute rCBD123 from denatured type I collagen (-3%). In contrast to the studies shown in Fig. 2.3 where the native type I collagen contained intact telopeptides, rCBD123 did not bind pepsin-cleaved native type I collagen coupled to Affi-Gel as evidenced by the quantitative recovery of rCBD123 in the unbound and wash fractions after chromatography (Fig. 2.4B). Since pepsin digestion removes the collagen amino- and carboxyl-terminal telopeptides, this shows that the telopeptide regions contribute significantly to the rCBD123 binding of native type I collagen. Repeated column assays consistently failed to detect rCBD123 binding to pepsin-treated native type I collagen. However, the highly sensitive microwell substrate binding assay showed some, although weaker, binding of rCBD123 to pepsin-treated type I collagen (not shown; see also Fig. 2.7). rCBD123 Binding to Native and Denatured Type I Collagen Mimics the Binding of Full-Length Gelatinase A—To verify that the binding interaction observed between the rCBD123 71 Human gelatinase A collagen binding domain domain and type I collagen corresponds to the properties of the full-length enzyme, we compared the binding of gelatinase A to denatured and native type I collagen. Gelatinase A in conditioned cell culture medium was chromatographed over gelatin-Sepharose. Enzyme in the column fractions was revealed by enzymography. Specific binding of gelatinase A to gelatin was found with a peak elution from the column at -2% (v/v) DMSO (Fig. 2.5A). This concentration of DMSO was slightly lower than for the elution of rCBD123 from gelatin (-3% (v/v)) (Fig. 2.2A) and may reflect slight modifying effects of domaimdomain interactions on substrate binding when the CBD is in the context of the whole protein. Gelatinase A also bound specifically to native type I collagen affinity columns with a peak elution at 2% (v/v) DMSO (Fig. 2.5B) which corresponded closely to that found for rCBD123 binding to native type I collagen (Fig. 2.4). In competition studies, rCBD123 was found to quantitatively displace gelatinase A from native type I collagen (Fig. 2.5C). Thus, together with the similarities in the binding and elution profiles of rCBD123 and the full length parental enzyme to denatured and native type I collagen this indicated that essentially all of the binding properties of gelatinase A to these proteins could be ascribed to the CBD domain. Native and Denatured Type I Collagen Compete for the Same Binding Sites on rCBD123—To determine whether native and denatured type I collagen bind at the same site on rCBD123, competition studies were performed. Pre-incubation of rCBD123 with native and denatured type I collagen reciprocally blocked binding of rCBD123 to denatured and native type I collagen columns, respectively, in a concentration-dependent manner (Fig. 2.6). At a mole ratio of rCBD123:denatured type I collagen of 1:9, all the rCBD123 applied to a native type I collagen column remained in the unbound or wash fractions (Fig. 2.6A) and no material was recovered in the DMSO elutes. With decreasing mole amounts of denatured type I collagen, progressively more rCBD123 bound the native type I collagen column. Native type I collagen similarly inhibited the binding of rCBD123 to gelatin-Sepharose in a concentration-dependent manner, but the inhibitory effect was markedly less consistent 72 Human gelatinase A collagen binding domain } { with a lower apparent Kd of; interaction (Fig. 2.6B). At a rCBD123:native type I collagen mole ratio of 1:9, -40% of the loaded rCBD123 was recovered in the unbound fraction or in the washes with the native type I collagen. The remainder of the rCBD123 was recovered from the gelatin-Sepharose by elution with DMSO. Pepsin-treated native type I collagen did not block rCBD123 binding to gelatin-Sepharose as evidenced by the absence of rCBD123 in both the unbound and the wash fractions (Fig. 2.6C) and the quantitative recovery of rCBD123 from the gelatin-Sepharose by elution with DMSO at the expected concentration (not shown). This is consistent with the lower affinity interaction between rCBD123 and pepsin-treated collagen relative to denatured type I collagen. Thus, these results showed that native and denatured type I collagen competed for binding sites on rCBD123 that were either the same or closely positioned so that binding of one molecule sterically blocked interaction with a second. To exclude the latter possibility, rCBD123 was bound to a native type I collagen affinity column and then competed with denatured type I collagen (Fig. 2.6D). Under these conditions, rCBD123 was displaced from the native type I collagen and recovered in the column elute with the denatured type I collagen. These results point to the likelihood that identical binding sites on rCBD123 bind both denatured and native type I collagen. However, because there are a number of rCBD123 binding sites on collagen with potentially different affinity (see later), these experiments do not conclusively rule out the possibility that separate, but overlapping, binding sites for denatured and native type I collagen occur on rCBD123. rCBD123 Binds More Than One Collagen Molecule Simultaneously—If each of the three individual modules of rCBD123 contain one collagen binding site, then the triple repeat of the fibronectin type II-like modules in rCBD123 has the potential to contain three collagen binding sites. Thus, it is possible that one or more molecules of collagen may be bound at a time to the rCBD or that one molecule of rCBD123 may bind two or three binding sites on collagen simultaneously. To test this, [14C]glycine-labeled type I collagen was incubated with molar excess of rCBD123 (from 1:1 to 1:26,000). These ratios were chosen to favor 73 Human gelatinase A collagen binding domain formation of rCBD123:radiolabeled collagen complexes at a single rCBD123 binding site. Complexes of rCBD123 bound to the radiolabeled ligand were then found to bind unlabelled native or denatured type I collagen coated on microwell plates (Fig. 2.7). Control samples of radiolabeled collagen alone did not bind to the collagen films. Thus, this demonstrated that rCBD123 possesses at least two collagen binding sites that can be simultaneously occupied by two collagen molecules. The rCBD123 binding to pepsin-treated [14C]glycine-labeled native type I collagen observed here can not be fully accounted for by the -9% denatured collagen content in the preparation. This confirms that although a major binding site on native type I collagen resides in the telopeptide region, additional weaker binding sites are located in the triple helical region of the collagen molecule. al(I) Collagen Chains Contain Multiple Binding Sites for rCBD123—The reduced binding of rCBD123 to native type I collagen after pepsin-cleavage indicated that an important site of interaction was present in the non-helical telopeptide end segments (Fig. 2.4, 6). This was confirmed by the binding of rCBD123 to the a 1(1) cyanogen bromide fragment 2, which is derived from the N-terminal telopeptide (Table I). Since rCBD123 also bound denatured type I collagen that had been pepsin-treated (Fig. 2.7) this indicated that additional higher affinity binding sites for rCBD123 may be present in the denatured a-chains that are masked by the triple helical structure of native type I collagen. Therefore, we screened other non-overlapping al(I) cyanogen bromide fragments for their interaction with rCBD123. al(I)-CB7, which encompasses the tissue collagenase cleavage site and which also contains the a 1(1) binding site for fibronectin (Guidry et al, 1990; Kleinman et al, 1981; Kleinman et al, 1976), and al(I)-CB8, a large fragment derived from the NH2-terminal portion of the triple helical region, both bound rCBD123 with an avidity similar to that for denatured type I collagen (Table I). This confirmed our previous analyses which indicated that multiple binding sites occur along the a 1(1) chains. Thus, rCBD123 binds to native type I collagen at sites in the telopeptide ends, at least one of which is on al(I)-CB2. In addition, rCBD123 74 Human gelatinase A collagen binding domain binds to lower affinity sites along the triple helical region of the molecule. Upon unfolding of the triple helix, higher affinity binding sites for rCBD123 are exposed, some of which reside on the a 1(1) cyanogen bromide fragments 7 and 8. rCBD123 Binds Denatured Type TV and V Collagens, and Elastin—To determine the contribution of the fibronectin-like domain of gelatinase A to the binding of other substrates of the enzyme, types IV and V collagens and elastin were assessed for rCBD123 binding. The micro well substrate binding assay revealed binding of rCBD123 to both denatured types IV and V collagens (Fig. 2.8) that was approximately equivalent to that seen to denatured type I collagen. These results showed that the type II-like modules have specificity for several collagen types in their denatured form. However, in contrast to native type I collagen, rCBD123 showed weaker binding to both native types IV and V collagens (Fig. 2.8) indicating that rCBD123 does not bind non-specifically to all collagens. Affinity chromatography of rCBD123 on native type IV collagen-affi-Gel affinity columns (n=6) confirmed the lack of binding to native type IV collagen (Fig. 2.10). Elastin, another substrate of gelatinase A, was also found to be bound by rCBD123 (Fig. 2.9). rCBD123 Interaction with Other Basement Membrane Components and TIMP-1—Since gelatinase A degrades type IV collagen in the basement membrane and is important in basement membrane penetration by metastatic cells, we investigated the binding properties of rCBD123 to reconstituted basement membrane and its components. rCBD123 did not bind reconstituted basement membrane (Matrigel®) coated as a film in microwell substrate binding assays (Table I). In light of the differences observed between rCBD123 binding to type I collagen films and to fibrillar gels (Fig. 2.3), we also analyzed binding to gels of Matrigel®. However, rCBD123 did not bind to the gel preparations of Matrigel® either (Table I). Since it was possible that binding sites for rCBD123 in reconstituted basement membrane were masked by interaction between the components of Matrigel®, binding to individual basement membrane components was also examined. Laminin was studied both by column affinity assays and as films in microwell substrate binding assays. Although no 75 Human gelatinase A collagen binding domain rCBD123 binding to laminin was detected on substrate columns (Fig. 2.10), a weak reaction was detected in the microwell substrate binding assay at high rCBD123 concentrations (Table I). Whereas rCBD123 eluted from control columns in the unbound or wash fractions, rCBD123 was consistently retarded during chromatography over heparin-Sepharose and was gradually released during extended washes and multiple elutions with 1.0 M NaCl (Fig. 2.10). DMSO did not elute any further rCBD123. Since rCBD123 did not bind to the negatively charged CM-Sepharose (Fig. 2.10) this indicated that the weak interaction with heparin was specific. SPARC (osteonectin) and fibronectin, other components of basement membrane, were also analyzed by microwell substrate binding assays, but rCBD123 did not bind either of these components (Table I). Since TIMP-1 interacts with active MMPs to inhibit enzymic activity and the fibronectin-like domain in gelatinase A is located immediately N-terminal of the zinc binding a-helix in the active site, the binding of rCBD123 to TIMP-1 was assayed. However, no binding was observed (Table I). DISCUSSION An intriguing feature of the gelatinases A and B is the presence of three fibronectin type II-like modules within the catalytic domain N-terminal to the Zn(II) binding a-helix. To determine if this fibronectin-like domain has binding specificity for extracellular matrix molecules that are degraded by gelatinase A, we have studied the isolated triple repeat expressed as a recombinant protein. E. coli was a suitable host for protein expression since this domain is not naturally glycosylated. The recombinant domain, rCBD123, exhibited high binding to type I collagen in both its native and denatured forms and to denatured types IV and V collagens. Although rCBD123 could bind multiple sites along the denatured a(I) chains and, with lesser affinity, to the native triple helical segment of type I collagen, rCBD123 did not exhibit general nonspecific collagen binding properties. For instance, type IV and V collagens were only 76 Human gelatinase A collagen binding domain weak ligands in their native forms. In addition to being a potent gelatinase, gelatinase A is a type IV and type V collagenase (Collier et al., 1988), but does not cleave native type I collagen. Therefore, it was surprising that rCBD123 bound native type I collagen with an apparent Kd in the micromolar range. To confirm that this was also a property of the parental enzyme, we showed that gelatinase A bound native type I collagen. Moreover, its elution profile was similar to that of the rCBD123. Similarity in binding and elution profiles were also shown with denatured type I collagen. Importantly, rCBD123 was found to compete for gelatinase A bound to native type I collagen confirming that the recombinant domain had folded in a biologically equivalent manner to the natural domain. Thus, our studies indicate that most, if not all, of the native and denatured type I collagen binding properties of gelatinase A reside in the fibronectin-like domain of the enzyme that we accordingly designated the collagen binding domain. Domaimdomain interactions with flanking or distant sites in the primary sequence of the parent molecule may influence the functional properties of the domain or its constituent modules. Intermodular interactions occur between the four type I and two type II modules within the 42-kDa gelatin-binding fragment of fibronectin (Litvinovich et al, 1991). Indeed, it remains controversial whether or not individual type II modules in fibronectin alone can bind gelatin or whether interactions with flanking sequences or intact adjacent type I modules are required for gelatin binding. Owens and Baralle (Owens and Baralle, 1986) found that the type II2 module required a 14-amino acid residue flanking sequence from the adjacent type I module (I7) for gelatin interaction. In addition, proteolytically-generated fragments of the 42-kDa fibronectin gelatin-binding fragment that retained gelatin binding properties contained both type I and type II modules (Litvinovich et al, 1991). Although recombinant fibronectin type II domains bind gelatin (Banyai et al, 1990), Skorstengaard et al. (1994) were unable to confirm the gelatin binding properties of individual type II modules. Their results demonstrated that LTIjILJ^  was the smallest recombinant segment of fibronectin that 77 Human gelatinase A collagen binding domain retained gelatin-binding property. Thus, these studies indicate that in fibronectin gelatin binding requires cooperative interactions between the type I and type II modules. However, in gelatinase A, it is apparent that type II-like modules alone can bind type I collagen. Banyai and Pathy (1991) and Banyai et al. (1994) have previously used naturally occurring restriction endonuclease sites in the gelatinase A cDNA as delineators for fusion protein constructs containing fibronectin type II-like modules. However, since these sites do not correspond exactly to the exon borders their constructs contained codons encoding 13 amino acid residues from the catalytic domain of gelatinase A. Because the active site residues may have contributed to the gelatin binding properties of these recombinant fusion proteins, we avoided such an overlap by first defining precisely the borders of the fibronectin-like domain from analysis of the genomic DNA. We then produced recombinant cDNA that encoded a protein encompassing only the three fibronectin type II-like modules of gelatinase A. Further, to minimize the potential effects of large fusion components such as 8-galactosidase and to facilitate protein purification, rCBD123 was engineered to contain only ,a short amino-terminal fusion peptide that included six histidines. Poly-histidine containing fusion proteins have been used extensively to facilitate recombinant protein purification and there is considerable evidence that this tag neither affects structure nor function of the recombinant protein (Janknecht et al., 1991; Takacs and Girard, 1991; Dobeli et al, 1990). With the same expression vector used for rCBD123, we too have found that other recombinant proteins containing the identical (His)6-tag do not bind type I collagen. Therefore, the ligand binding properties observed for rCBD123 can be attributed solely to the fibronectin type II-like modules of rCBD123 and not to the (His)6-containing fusion peptide. The lower binding of rCBD123 with native type I collagen compared to denatured type I collagen indicates that rCBD123 either recognizes different sites on the native and denatured forms of type I collagen or that the rCBD123 interaction with collagen is conformation dependent. That is, rCBD123 may strongly bind identical sites on the unfolded a-chains but display a weaker interaction with the same sites if partially masked by the triple 78 Human gelatinase A collagen binding domain helix. To map the rCBD123 binding site on native type I collagen, the telopeptide end groups were removed by pepsin digestion. Pepsin treatment greatly diminished the binding of collagen to rCBD123 and pepsin-treated native type I collagen was also ineffective in competing for rCBD123 binding with native type I collagen. Therefore, this indicates that an important binding site is located in the telopeptide sequences. Confirming this was the observation that the al(I)-CB2 fragment from the NH2-terminal telopeptide bound rCBD123. Since the binding affinity for al(I)-CB2 was similar to that for denatured type I collagen, but greater than for native type I collagen, this also indicated a conformation-dependent interaction of the telopeptide region with the rCBD123. Additional studies showed weak binding of rCBD123 to pepsin-treated native type I collagen films and to [14C]glycine-labeled pepsin-treated native type I collagen in solution. Therefore, in addition to a strong binding site for rCBD123 in the telopeptide region, additional weaker binding sites are present in the triple helical portion of the native type I collagen molecule. It is unknown if these sites are identical to those of denatured type I collagen. If identical, this suggests that either the chain flexibility of denatured collagen potentiates binding or that key recognition residues are partially masked by the triple helix. Indeed, the binding of fibronectin to native type I collagen on CB7 near the collagenase cleavage site is analogous (Jilek and Hoermann, 1979; Engvall et al., 1981). Here, the low imino acid content of the collagenase cleavage site (Highberger et al, 1979) produces a localized, slight relaxation of the triple helix in this region due to reduced interchain hydrogen bonding. This region relaxes further with increasing temperature which facilitates increased fibronectin access to the binding site. The binding of rCBD123 to cd(I) CB2, 7, and 8 demonstrated the presence of at least three rCBD123 binding sites on the od(I) collagen chain. In view of this, binding kinetics and dissociation constants are difficult to interpret. In comparison, fibronectin binds only to individual a-chains in the vicinity of the tissue collagenase cleavage site on CB7 (Kleinman et al, 1981; Kleinman et al, 1976) and to a second lower affinity site on the a2(I) chain within the triple helix near the NH2-terminus on a2(I) CB4 (Guidry et al, 1990). The 79 Human gelatinase A collagen binding domain apparent Kdfor the rCBD 123:gelatin interaction is similar to that for the fibronectin 42-kDa gelatin-binding fragment, 3 x 10"6 M (Ingham et al., 1988). Therefore, the differences in binding site location and number between the archetypical fibronectin type II module and the type II-like modules in gelatinase A suggest that sequence differences have resulted in collagen binding modules with altered specificities. This commonly occurs with other protein modules which have been used many times in different proteins to perform similar or, through evolutionary divergence, different tasks (Reid and Day, 1989; Campbell et al., 1990; Williams and Barclay, 1988). Moreover, Collier et al. (1992) found that the individual type II-like modules in the 92-kDa gelatinase had different gelatin affinities and our own preliminary experiments with individual CBD modules from gelatinase A show that rCBD module 1 has essentially no gelatin-binding properties (Overall et al. unpublished). It remains to be determined whether or not differential gelatin-binding of the type II modules is an indication of module specialization that has generated new binding sites for the different extracellular matrix molecules that rCBD123 binds. The demonstration that two molecules of either denatured or native type I collagen can bind simultaneously to rCBD123 showed that there was more than one collagen binding site per rCBD123 and that steric clashes by either the substrate or the individual type II-like modules do not prevent additional binding site occupancy in the domain. Therefore, the CBD in gelatinase A has the potential to form complexes with connective tissue components. Considering the near universal constitutive synthesis and distribution in tissues of the gelatinase A, the enzyme might even fulfill a minor structural role in connective tissue networks. More likely though, gelatinase:gelatin complexes may serve to increase gelatinolysis by localizing substrate prior to cleavage. Indeed, a catalytically competent mutant of gelatinase A with deletion of the fibronectin-like domain had a 90% reduced rate of gelatinolysis compared with the wild type enzyme (Murphy etal, 1994). Nonetheless, the biological role of this domain in relation to the gelatinolytic properties of these two MMPs is uncertain since Collier et al. (1992) reported that DMSO had no effect on gelatinolysis at 80 Human gelatinase A collagen binding domain a concentration that disrupted the 92-kDa gelatinase type EE-like module binding to gelatin. In lieu of the high affinity binding of rCBD123 to denatured types I, IV and V collagens and native type I collagen, we hypothesized that the rCBD123 would also bind native types IV and V collagens which are substrates of gelatinase A. However, only weak binding to the native forms of these collagens was detected by the microwell substrate binding assay and no binding of rCBD123 was detected on native type IV collagen columns. The high sensitivity of the microwell plate assay may have detected a small amount of denatured collagen present in the preparation or which formed during adherence to the plastic. Thus, based on these experiments it is likely that the native triple helical structure of types IV and V collagens masks rCBD123 binding sites on the a-chains thereby preventing binding to the native forms of these collagens. Since the type IV collagen was not prepared by pepsin treatment, rCBD123 also does not significantly bind the large non-helical domains of type IV collagen that are removed by pepsin digestion. Therefore, we infer that exosites for native type IV collagen, if present, reside elsewhere in the gelatinase A molecule. Other potential ligands for rCBD123 binding were screened among the known substrates of gelatinase A. rCBD123 binding to reconstituted basement membrane was low. To ensure that masking of binding sites on the basement membrane proteins was not occurring in the gel, several constituent components were individually assessed for binding by rCBD123. Consistent with the lack of basement membrane binding, rCBD123 did not bind native type IV collagen, laminin, or fibronectin. SPARC, another protein found in basement membranes (Mann et al, 1987) and a variety of connective tissues (Domenicucci et al, 1988; Mason et al, 1986; Wasi et al, 1984), particularly those undergoing rapid remodelling and containing unfolded collagen (Salonen et al, 1990), also did not interact with rCBD123. Therefore, substrate binding by the fibronectin-like domain is not necessary for proteolytic cleavage of native types IV or V collagens or fibronectin by gelatinase A. In contrast, rCBD123 binds insoluble elastin, a substrate of the enzyme. Since the fibronectin-type II like modules are located immediately to the N-terminal 81 Human gelatinase A collagen binding domain side of the Zn(II)-binding a-helix in the catalytic domain of gelatinase A, it was possible that these modules play a role in the interaction of the enzyme with TIMP-1. However, we did not find any evidence for such an interaction of the rCBD123 with TIMP-1 which is consistent with the lack of binding to TIMP-1 by a fibronectin-like domain deletion mutant of gelatinase A (Murphy et al, 1994). The extracellular matrix binding properties of the CBD in gelatinase A has important implications for the function and clearance of the enzyme under physiological and pathological conditions. We have previously proposed that, as a consequence of the opposite regulation of gelatinase A and collagenase by TGFB-1, gelatinase A has a formative role in connective tissue deposition in addition to its role in connective tissue degradation (Overall et al, 1989). The range of extracellular matrix proteins bound by rCBD123 shown here suggests the possibility of an additional role for this domain in gelatinase A — that is one of extracellular matrix localization. The interaction with heparin also indicates that gelatinase A may bind heparin-sulphate proteoglycans on the cell surface and to perlecan in basement membranes (Noonan etal, 1991; Kallunki and Tryggvason, 1992). Indeed, gelatinase A can be extracted from developing tooth enamel (Overall and Limeback, 1988) and insoluble bone (Overall, Domenicucci, and Sodek, unpublished observations). A consequence of connective tissue localization may be that cell migration through basement membranes and other connective tissues would be facilitated for those cells expressing the cell membrane associated MT-MMP (Sato et al, 1994), a newly cloned gelatinase A activator. We suggest that gelatinase A may be localized in connective tissue matrices by binding a number of matrix proteins through the CBD, and remain poised for activation during tissue remodelling or in pathology. SUMMARY Gelatinase A is an important matrix metalloproteinase in the degradation of basement membranes and denatured collagens. These proteolytic processes are required for pathologic 82 Human gelatinase A collagen binding domain tissue destruction and physiologic tissue remodelling. To investigate the molecular determinants of substrate specificity of this enzyme, a 21 kDa domain of gelatinase A, consisting of three tandem fibronectin type II-like modules, was expressed in E. coli. Similar to the full length gelatinase A and the type II modules in fibronectin, the recombinant fibronectin type II-like domain of this proteinase bound denatured type I collagen with an apparent Kd in the uM range. This domain, designated the collagen-binding domain, possesses at least two collagen binding sites that can each be simultaneously occupied. rCBD123 also avidly bound elastin and denatured types IV and V collagens, but neither native types IV and V collagens nor fibronectin, all of which are substrates of the enzyme. Although gelatinase A is involved in basement membrane degradation, rCBD123 also did not bind reconstituted basement membrane, laminin, or SPARC. Native type I collagen, which is not degraded by gelatinase A, competed with gelatin for a shared binding site on rCBD123. rCBD123 also displaced full-length gelatinase A bound to native type I collagen, further demonstrating that the collagen binding properties of the recombinant domain closely mimicked those of the full length enzyme. Since rCBD123 showed reduced binding to pepsin-cleaved type I collagen, either or both of the collagen telopeptide ends contain recognition sites for the gelatinase A fibronectin-like domain. This was confirmed by the avid binding of rCBD123 to the al(I) collagen cyanogen bromide fragment CB2 from the NH2-terminal telopeptide. rCBD123 also bound al(I)-CB7, which encompasses the fibronectin binding site, and to cd(I)-CB8, a fragment not bound by fibronectin. Thus, type I collagen contains multiple binding sites for rCBD123 which are partially masked by the triple helical conformation of native collagen and fully exposed upon unfolding of the triple helix. The potential of the fibronectin-like collagen binding domain of gelatinase A to bind extracellular matrix proteins may facilitate enzyme localization in connective tissue matrices. 83 Human gelatinase A collagen binding domain TABLE 2.1 Interaction of rCBD123 with extracellular matrix proteins Binding of rCBD123 (10 pmol (0.1 uM) and 512 pmol (5.1 uM) in 100 ul of 50 mM Tris-HC1, pH 7.4) to proteins coated on 96-well microliter plates was assayed in the linear range of the microwell substrate binding assay as described under "Experimental Procedures" {Chapter 2). Relative binding' t 0.1 uM 5.1 uM Denatured type I collagen 1.0 1.0 al(I)-CB2 0.93 1.0 al(I)-CB2 1.17 1.11 al(I)-CB2 1.07 0.99 Matrigel® gel" 0.04 0.02 Matrigel® film 0 0 Laminin 0.13 0.19 SPARC 0.04 0.19 Fibronectin 0 0 TIMP-1 0 0.01 Myoglobin 0.08 0.03 aBinding is expressed relative to the binding of rCBD123 to denatured type I collagen, defined as 1.0, and background corrected. bMatrigel® was coated as a film or as a three-dimensional gel in the base of the microwell plates as described under "Experimental Procedures." 84 A i I I I i I - - - - + + + + + + CM H - - + + + + + + + + DTT 97. 67 43-29 18.4-> — rCBD123 fr B 97_, p 67 -> 43 -> 29-> 18.4-+ 14.4-* rCBD123 fr FIG. 2.1. SDS-PAGE and Western blot analysis of non-reduced, reduced, and carboxymethylated rCBD123. Samples of rCBD123 (1 ug/lane) were analyzed by SDS-PAGE on 15% (w/v) cross-linked polyacrylamide minislab gels under non-reducing {-DTT) or reducing {+DTT) conditions and stained with Coomassie brilliant blue R250 {Panel A) or transferred by Western blotting to PVDF membranes and reacted with an anti-gelatinase A polyclonal antibody {Panel B) as described under "Experimental Procedures" {Chapter 2). The antibody reaction to rCBD123 was consistently greater for non-reduced samples compared to reduced or reduced and carboxymethylated (CM+) samples. rCBD123 migrated with a relative molecular mass of 21.1 kDa non-reduced, 22.1 kDa reduced, and 24.1 kDa reduced and carboxymethylated. Panel A, reduced and carboxymethylated rCBD123 was analyzed for gelatin binding by affinity chromatography over mini-columns of gelatin-Sepharose as described under "Experimental Procedures" {Chapter 2). Reduced and carboxymethylated rCBD123 before chromatography {B) was fully recovered in the unbound {U) and wash 1 (W7) and wash 2 {W2) fractions and did not bind to gelatin as shown by the absence of protein in the 1.0 M NaCl (N) and 10% (v/v) DMSO (£>) elutes. M r, molecular weight markers x 10"3 as indicated; fr, dye front. 85 A %DMSO MrB U N 0 1 2 3 4 5 6 7 8 97 ~ : ' 67 -» —' 4 3 - * m ' 29-»w • • \ ^ — ! <-rCBD123 18.4—> «§ 14.4 6.2-» B 1 :0.3 1: 1 W -i r 1:3 1 :9 W W W Mr B U 1 2 B U .1 2 B U 1 2 B U 1 2 97_>:_ 67-*;-43-» <> 29 a-chains -rCBD123 18.4-»'«-14.4 6.2^ FIG. 2.2. Denatured type I collagen affinity chromatography of rCBD123. Recombinant CBD123 (40 ug, 1.9 nmol in 100 ul) was loaded (B) onto mini-columns of gelatin-Sepharose and chromatographed as described under Experimental Procedures. After extensive washes with chromatography buffer, 1.0 M NaCl was applied to the column in chromatography buffer followed by a step gradient of DMSO from 1 to 8% (v/v) in chromato-graphy buffer. Elutes were analyzed by 15% SDS-PAGE gels under reducing conditions (Panel A). rCBD123 bound avidly to gelatin-Sepharose as shown by the absence of rCBD123 in unbound fractions (U), washes (not shown), or the 1.0 M NaCl (AO elute. Peak elution of rCBD123 was at 3% (v/v) DMSO. The specificity of the interaction between rCBD123 and gelatin was confirmed by a competition assay in which rCBD123 was incubated with increasing mole-ratios of gelatin (1:0.3 to 1:9) for 90 min at 20 °C prior to loading onto gelatin-Sepharose columns. The elutes were analyzed under reducing conditions on 15% SDS-PAGE gels. Panel B shows that competition with increasing mole amounts of gelatin caused both a graduated increase in the amount of rCBD123 in the unbound (U) and the two wash fractions (W; 1, 2) and a corresponding decrease in the column-bound rCBD123 (not shown). The positions of collagen a-chains in the unbound and wash fractions from the competition reaction, rCBD123, and the kDa of marker protein standards (Mr) are indicated. 86 1.5-E c LO o CD O c CO _Q o CO _Q < O Type I fibrils • Type I denatured • Type I native A Myoglobin Concentration rCBD123 (JLIM) FIG. 2.3. rCBD123 affinity for type I collagen in different conformations. Type I collagen as a three-dimensional gel (Type I fibrils), an absorbed film (Type I native), and as heat denatured oc(I)-chains (Type I denatured) was coated on the bases of wells in 96-microwell plates as described under "Experimental Procedures" (Chapter 2). None of these collagens had been pepsin-treated. Myoglobin was coated as a control protein for non-specific staining. Serially diluted rCBD123 was added to the wells in a volume of 100 Lil and the amounts of bound rCBD123 quantitated after reaction with anti-gelatinase A and alkaline phosphatase conjugated secondary antibody by reading the absorbance of the reaction mixture at 405 nm. Data are plotted as the mean values of two experiments. 87 A W NaCl %DMSO B U 1 2 3 1 2 0 1 2 3 4 5 6 43-> 29-> ^ <_rCBD123 18.4 14.4-> ; • ' 6 - 2 ~ > ^ 43-^ 29 —> i .. — <-rCBD123 18.4-> ; . .14.4-» j 6.2-4 U FIG. 2.4. Interaction of rCBD123 with native and pepsin-treated type I collagen. Recombinant CBD123 (40 ug, 1.9 nmol in 100 ul) was chromatographed over Affi-Gel 10 columns coupled with native type I collagen (Panel A) or pepsin-treated native type I collagen (Panel B) as described under "Experimental Procedures" (Chapter 2). Columns were washed thoroughly with chromatography buffer before elution with buffer containing 1.0 M NaCl. A step gradient from 0 to 10% (v/v) DMSO was then applied in chromatography buffer. Whereas rCBD123 bound to native type I collagen and required 1 to 2% (v/v) DMSO for elution, rCBD123 binding to pepsin-treated type I collagen could not be detected. Sample before chromatography (B), unbound (U), washes 1, 2, and 3 (W), 1.0 M NaCl elutes 1 and 2 (NaCl), and DMSO (%DMSO) fractions (0 to 6% shown) were analyzed under reducing conditions on 15% SDS-PAGE gels. The positions of rCBD123 and the kDa of marker protein standards are indicated. 88 B %DMSO N 0 1 2 3 4 5 10 FIG. 2.5. Gelatinase A and rCBD123 show similar binding to native and denatured type I collagen. 1.5 ml of conditioned cell culture medium containing gelatinase A was chromatographed over minicolumns of Affi-Gel 10 coupled to denatured (Panel A) or native (Panel B) type I collagen. After sequential washes with chromatography buffer and with 1 M NaCl (AO in chromatography buffer, bound enzyme was eluted with a step gradient of DMSO (%DMSO) from 1 to 10% (v/v) in chromatography buffer. Column fractions (non-reduced) were analyzed on 10% SDS-PAGE gels containing 100 ug/ml heat-denatured type I collagen (gelatin). After enzymography (see "Experimental Procedures" (Chapter 2)), gelatinase was detected by counterstaining the gelatin in the substrate-impregnated gels with Coomassie brilliant blue R-250. Gelatinase A bound to both denatured (Panel A) and native (Panel B) type I collagen columns and was released from both columns with DMSO in chromatography buffer. The Mrs of the non-reduced latent (66 kDa) and activated (66, 59 kDa) forms of gelatinase A (Overall and Sodek, 1990) are indicated. Panel C, gelatinase A (B, before chromatography) was bound to native type I collagen columns and excess unbound enzyme removed by sequential washes in chromatography buffer (M0 and 1 M NaCl in chromatography buffer (N). Bound gelatinase A was then competed off the native type I collagen with 40 jig rCBD123 in chromatography buffer (rCBD123) and residual bound enzyme eluted with 10% DMSO. The rCBD123 used for competition was recovered with the enzyme-containing fractions as indicated. 89 Human gelatinase A collagen binding domain See figure 2.6 on the following page. FIG. 2.6. Competitive inhibition of rCBD123 binding by native, pepsin-treated, and denatured type I collagen. rCBD123 was incubated before chromatography with increasing mole ratios of denatured type I collagen (Panel A), native type I collagen (Panel B), or pepsin-treated native type I collagen (Panel C) prior to loading onto native type I collagen (Panel A) or gelatin-Sepharose columns (Panels B and Q. Mole ratios are as indicated (1:0.3 to 1:9) (1:0 controls not shown, see Fig. 2A and 4A for examples). The unbound material (LO and chromatography buffer washes (W) were analyzed under reducing conditions by SDS-PAGE using 15% gels. The elutes from the 1.0 M NaCl and DMSO step gradient elution are not shown. Native and denatured type I collagen reciprocally inhibited rCBD123 binding to denatured (Panel B) or native (Panel A) type I collagen, respectively, in a concentration-dependent manner. In contrast, pepsin-treated type I collagen did not inhibit the binding of rCBD123 to gelatin-Sepharose. Panel D, rCBD123 (B) was applied to a native type I collagen column and the column was washed with chromatography buffer (W) and 1 M NaCl (N). rCBD123 was found to be eluted after competition with denatured type I collagen in chromatography buffer (El). B, rCBD123 before chromatography; E2, wash; E3, 10% DMSO in chromatography buffer elute. The positions of rCBD123, a (I) collagen-chains (a-chains), and the kDa of marker protein standards (Mr) are indicated. 90 1:0.3 1:1 1:3 1:9 W W W W B U 1 2 B U 1 2 B U 1 2 B U 1 2 97. 67-43-29-18.4-14.4-• a-chains -rCBD123 6.2-» B 97. 67=? 43-» 29 -* 18.4-14.4-6.2-* 1:0.3 1:1 1:3 w' w w , w B U 1 2 B U 1 2 B U 1 2 B U 1 2 . a-chains -rCBD123 1:0.3 1:1 1 :3 1 :9 W W W W B U 1 2 B U 1 2 B U 1 2 B U 1 2 97 67-» 43-* 29-* 18.4-* 14.4-* 6.2-* . a-chains -rCBD123 D 97-67-43-29-M r B U W N W 1 2 3 ;1 ^ WT-18.4-* 14.4-* 6.2-* • a-chains -rCBD123 FIG. 2.6. Competitive inhibition of rCBD123 binding by native, pepsin-treated, and denatured type I collagen. (For full legend see previous page) 91 E Q. "D (D C o O 2000 1500-g 1 0 0 0 -500-• [1 4C] Type I denatured O [ 1 4C] Type I native 2.5 5.0 7.5 10.0 Concentration rCBD123 (uJvl) FIG. 2.7. Simultaneous occupancy of rCBD123 binding sites by type I collagen molecules. [14C]Glycine-labeled native and denatured pepsin-treated native type I collagen were incubated with molar excess of serially diluted rCBD123 for 1 h. The reaction mixtures containing radiolabeled and collagen:rCBD123 complexes were then transferred to microwell plates coated with unlabelled native or denatured type I collagen, respectively and incubated for 1 h. After removal of unbound material by rinses with phosphate buffered saline containing 0.5% (v/v) Tween 20, pH 7.4, the bound material was eluted with 10% (v/v) DMSO in 50 mM Tris, pH 7.4 for 30 min and the radioactivity measured by scintillation counting. Duplicate data from one experiment is shown. rCBD123 simultaneously bound radiolabeled and unlabelled denatured type I collagen and, to a lesser extent, native type I collagen. This indicates that the triple repeat of fibronectin type II-like modules in rCBD123 has the potential to bind at least two molecules of collagen concurrently. 92 Concentration rCBD123 ( J I M ) FIG. 2.8. Interaction of rCBD123 with native and denatured types IV and V collagens. Equal amounts of native and denatured type IV collagen, native and denatured type V collagen, and denatured type I collagen were coated as films on 96 well microtiter plates as described under "Experimental Procedures" (Chapter 2). Serially diluted rCBD123 was added for 2 h. After rinses, bound rCBD123 was detected using a polyclonal anti-gelatinase A antibody and quantitated at 405 nm after reaction of alkaline phosphatase-conjugated second antibody with substrate. Data points represent means of two experiments. 93 FIG. 2.9. rCBD123 binding to elastin. rCBD123 binding to films of elastin and denatured type I collagen coated on microtiter plates was determined by the microwell substrate binding assay as described under "Experimental Procedures" (Chapter 2). The rCBD123 binding to elastin and denatured type I collagen was comparable over a range of rCBD123 concentrations (5.1 LiM to 6.25 nM). Data points are averages of two experiments. 94 Laminin r w Type IV collagen Myoglobin i , . w II W 97->: 67 —> 43-» « 29 _» — 18.4-> — 14.4-» — H B U 1 2 N D B U 1 2 N D B U 1 2 N D • rCBD123 6.2 -» Heparin sulfate W N -\ r CM Sepharose Control W W B U 1 2 3 4 1 2 3 D B U 1 2 N D B U 1 2 N D 97-67-43-29. 18.4-> 14.4-> 6.2 -» • rCBD123 FIG. 2.10. Interaction of rCBD123 with basement membrane proteins. The binding of rCBD123 to basement membrane components and controls was determined using mini-affinity columns of Affi-Gel coupled with laminin and native type IV collagen. Heparin-Sepharose and CM-Sepharose were also used as described under "Experimental Procedures" (Chapter 2). The negative controls consisted of Affi-Gel 10 reacted with myoglobin or buffer alone (no protein), and blocked as for the other coupling reactions. The negatively charged CM-Sepharose served as a control for specific interaction of rCBD123 with heparin-Sepharose. Equal amounts of rCBD123 were loaded (B) on each column and unbound protein fractions (U), chromatography buffer washes (W, 1-4), and 1.0 M NaCl elutes (iV, 7-3) were analyzed by SDS-PAGE on 15% (w/v) cross-linked polyacrylamide minislab gels under reducing conditions. The positions of rCBD123 and the kDa of marker protein standards are indicated. Other than for the heparin-Sepharose column, quantitative recovery of rCBD123 in the unbound and wash fractions occurred in all affinity columns. This was verified by the absence of any bound protein in 10% (v/v) DMSO (D). In contrast, rCBD123 was retarded by the heparin-Sepharose with protein consistently observed in unbound, wash, and 1.0 M NaCl elutes. 95 CHAPTER 3 The Contribution of Human Gelatinase A Collagen Binding Domain Lysine Residues to Ligand Interactions. INTRODUCTION The family of matrix metalloproteinases collectively can degrade most extracellular matrix proteins and, therefore, in a regulatory balance with the specific tissue inhibitors of metalloproteinases, is important for tissue remodeling (Birkedal-Hansen et al, 1993; Woessner, 1991). In addition, MMPs are found at high levels in the inflammatory diseases of periodontitis and arthritis and contribute to tumor cell invasion (Liotta and Stetler-Stevenson, 1991; Murphy and Hembry, 1992; Birkedal-Hansen, 1993; Sodek and Overall, 1992). The MMPs are characterized by a high degree of homology between the constituent structural domains (Matrisian, 1992; Birkedal-Hansen et al, 1993). Outside the catalytic domain, exosites on the COOH terminal hemopexin-like domain serve in recognition and binding of substrates and are required for substrate cleavage (Murphy et al, 1992; Windsor et al, 1991; Bigg et al, 1994; Wallon and Overall, 1997; Overall et al, 1997). In human gelatinase A, exosites on the three fibronectin type II-like repeats that form the collagen binding domain account for most, if not all, of the native and denatured type I collagen binding properties of the enzyme (Chapter 2, Steffensen et al, 1995; Collier et al, 1992; Banyai et al, 1994; Banyai and Patthy, 1991), and for binding to native types V and X collagens, denatured types IV, V, and X collagens, elastin, and heparin (Chapter 2, Steffensen et al, 1995; Abbey et al, 1997). The importance of the CBD is illustrated by the complete loss of elastinolytic activity of gelatinase A (Shipley et al, 1996) and a 90% reduction in cleavage of denatured type I collagen (Murphy et al, 1994) following deletion of this domain. In previous studies, we demonstrated that heparin binds specifically to the gelatinase A CBD expressed as a recombinant protein consisting of the three fibronectin type II-like 96 Heparin binding by the gelatinase A CBD modules (rCBD123) (Chapter 2, Steffensen et al, 1995). The ionic interactions of rCBD123 bound to heparin-Sepharose were disrupted by 0.25 M NaCl. That this binding was specific and not merely the result of charge interactions was confirmed by the absence of rCBD123 binding to the negatively charged CM-Sepharose (Chapter 2, Steffensen et al, 1995). We have subsequently shown that a second higher affinity heparin binding site resides in the COOH terminal hemopexin-like domain of human gelatinase A (Wallon and Overall, 1997; Overall et al, 1997). Although gelatinase A has been demonstrated to degrade link protein and the core protein of aggrecan (Nguyen et al, 1993), it is not yet known whether this enzyme can cleave the core protein of heparan sulfate proteoglycans. Structurally, heparin closely resembles heparan sulfate with both consisting of repeating disaccharide units but differing, mainly, by their degree of sulfation (Gallagher and Waler, 1985). Although the two glycosaminoglycans are not identical, both possess anticoagulatory activity, compete for the same cell surface binding sites, and bind similar molecules (Busby et al, 1995; Kjellen et al, 1980). Therefore, the binding of gelatinase A and, in particular, the CBD to heparin may also reflect similar interactions with heparan sulfate chains of secreted proteoglycans such as perlecan found in the basement membrane (Noonan et al, 1988) and in intercellular spaces of tumor cells (Iozzo et al, 1994). This binding is of interest because gelatinase A has been localized at high levels on tumor cells during basement membrane penetration and invasive growth and has the potential to degrade important basement membrane components such as type IV collagen (Tryggvason et al, 1993; Liotta et al, 1980). In addition, perlecan binds basic fibroblast growth factor (Gonzales et al, 1990). Release of growth factors from perlecan during basement membrane degradation may increase bioavailability and stimulate cellular responses such as neovascularization during tumor expansion and invasion (Folkman and Shing, 1992). Exploring the heparin binding properties of gelatinase A is therefore important in gaining an improved understanding of the role of gelatinase A in physiological and pathological processes. The heparin binding sites on a number of different proteins contain the consensus 97 Heparin binding by the gelatinase A CBD sequences XBBXBX or XBBBXXBX where B positions generally are occupied by the basic amino acid residues lysine and arginine separated by hydrophobic residues (X) with high frequency of leucine, glycine, isoleucine, and alanine (Cardin et al., 1991; Cardin and Weintraub, 1989). There is also a relative abundance of aspartic acid and serine in the X positions (Cardin and Weintraub, 1989). While this binding site configuration holds true in studies based on sequence alignments, heparin and heparan sulfate binding peptides from a random peptide library were enriched not only in arginine and lysine, but also in glycine and serine (San Antonio et al., 1992). Although the gelatinase A CBD contains eleven lysines, the proposed consensus binding site motifs do not occur in this domain. However, like the fibronectin type III-13 module where protein folding positions non-contiguous charged residues to create a heparin binding "cationic cradle" (Busby et al., 1995), even distantly located lysines in CBD may contribute to form the heparin binding site. Little is also known about the molecular basis for the CBD module interactions with collagen. In addition to binding of gelatin (Collier et al., 1992; Steffensen et al, 1995; Banyai et al., 1994; Banyai and Patthy, 1991), our previous studies of the gelatinase A CBD also demonstrated binding to native types I and V and denatured types I, IV, and V collagens, as well as to elastin (Chapter 2, Steffensen et al., 1995). The molecular determinants of CBD interactions with gelatin have been addressed only once before in the CBD of human gelatinase B. Collier et al. (1992) demonstrated that substitution of individual charged residues in the two major loops connecting the predicted B-strands and hydrophobic residues within the B-strands of isolated CBD module 2 reduced gelatin binding. However, it is not clear to what extent the functional alterations of these mutations may have been caused by structural changes. Since very little is understood about the molecular determinants of the gelatinase A interaction with substrates and ligands and because lysine residues are important for heparin binding in many proteins, we investigated the effects of lysine modifications on the gelatinase A CBD binding to heparin and to other known ligands. Acetylation to block all surface lysines confirmed the contribution of these residues to CBD binding of heparin and type I 98 Heparin binding by the gelatinase A CBD collagen. Subsequent substitutions of individual lysines with alanines showed that specific changes of individual lysines were sufficient to abolish heparin binding completely or to reduce collagen binding. EXPERIMENTAL PROCEDURES Extracellular Matrix Proteins—Acid soluble native type I collagen was extracted from rat tail tendons with 0.5 M acetic acid and purified by differential precipitation with 1.7 M NaCl (Piez, 1967). Gelatin was prepared from type I collagen by heat denaturation at 56 °C for 30 min. Bovine nuchal ligament elastin and horse heart myoglobin were from Sigma; murine type IV collagen (non-pepsin treated) was obtained from GIBCO BRL and Becton Dickinson; and human type V collagen (non-pepsin treated) was from Becton Dickinson. rCBD123 Expression Vector Construct—The pGYMX123 expression vector for the three fibronectin type II-like modules (CBD 1,2,3) of human gelatinase A, previously prepared from h72-kDa gelatinase cDNA (Chapter 2, Steffensen et al., 1995), was used for the expression of rCBD123 and for site-specific mutagenesis. The rCBD123, corresponding to residues 191 to 364 of human gelatinase A, contains a short N-terminal 16-amino acid fusion sequence that includes an N-terminal initiating methionine and a metal ion binding (His)6 tract. Site-Specific Mutagenesis—The sequences of pGYMX123 encoding six lysines in CBD modules 2 and 3 were mutated to code for alanine by site-specific mutagenesis according to the method of Kunkel (Kunkel et al, 1993). Uracil-containing single-stranded template DNA was prepared from the E. coli strain RZ1032 (duf, ung~) using the M13-based helper phage R408. Phage were collected from the bacterial supernatant by centrifugation at 10,000 x g and precipitated with 3.5% (w/v) polyethylene glycol 8000, 0.4 M NaCl. Capsids were denatured by phenol/chloroform-isoamyl alcohol extraction and the phagemid DNA ethanol precipitated. In the mutagenesis reactions, template DNA was annealed with 5'-phosphorylated oligonucleotides (0.5 pmol) incorporating the lysine to alanine codon changes (Appendix 1) in annealing buffer (10 mM MgCl2, 1 mM DTT, 50 mM NaCl, 10 mM Tris-99 Heparin binding by the gelatinase A CBD HC1, pH 8.0) at 55 °C for 5 min followed by slow cooling to 22 °C. The reaction buffer was then adjusted to 7.5 mM MgCl2, 3 mM DTT, 38 mM Tris-HCl, pH 7.5 and incubated with Klenow fragment (1 U), T 4 DNA ligase (1 U), dNTPs (0.5 mM), and rATP (0.5 mM) for 2 h at 22 °C. E. coli strain DH5ocF' was transformed with the mutagenesis reaction DNA and colonies were screened for the mutant plasmids by DNA sequencing (Sanger et al, 1978). The fidelity of the mutation and the entire coding sequence were then verified for each mutant plasmid by double-stranded chain-termination DNA sequencing (Sanger et al, 1978). Expression and Purification of Recombinant Proteins—rCBD123 and mutants were expressed in E. coli and purified as described previously in detail (Chapter 2, Steffensen et al., 1995) by methods producing correctly folded recombinant protein (Luck et al, 1992). In brief, ten E. coli strains were initially screened for expression of the mutant CBDs from which Le392F' was selected and then grown in 3.6 1 cultures of 1% (w/v) trypticase peptone, 0.8% (w/v) yeast extract, 0.5% (w/v) NaCl, pH 7.4 containing 100 ug/ml ampicillin for 18 h at 37 °C. Inclusion bodies were pelleted by centrifugation at 10,000 x g after bacterial lysis with lysozyme and dissolved in 6 M guanidine-hydrochloride under non-reducing conditions to favor dissolution of only non-intermolecularly disulfide cross-linked recombinant protein. Disulfide bond exchange was performed by dialysis against aerated 0.1 M sodium borate buffer, pH 10 (Luck et al, 1992) and the proteins were then refolded by stepwise dialysis in chromatography buffer (100 mM NaPi5 0.5 M NaCl, pH 8.0). The refolded soluble protein was then separated from denatured or highly crosslinked aggregates by centrifugation. rCBD123 and mutants proteins were purified by chromatography over a Zn2+-chelate column (chelating Sepharose 6B was from Pharmacia) and elution with imidazole in chromatography buffer. The eluted protein was then applied to a gelatin-Sepharose 4B (Pharmacia) column and bound protein stepwise eluted with 7% (v/v) DMSO in 50 mM Tris, pH 7.4. To separate the correctly folded and crosslinked recombinant proteins from a small amount of remnant intermolecularly crosslinked forms, a second Zn2+-chelate column was used. Monomeric proteins eluted with 100 mM imidazole in chromatography buffer whereas 100 Heparin binding by the gelatinase A CBD the multimers were retained on the column at this imidazole concentration. Purity was monitored by SDS-PAGE (see below) and the purified protein was equilibrated against 50 mM Tris, pH 7.4 by dialysis. The protein concentration was determined spectrophotometrically at 280 nm using a molar extinction coefficient of 2.6 x 104 M"1 cm"1 on a spectrophotometer (Beckman DU 640), and also by the Bradford protein assay (Bio-Rad). Protein samples were snap-frozen in liquid nitrogen and stored at -70 °C until use. Electrospray Mass Spectrometry—To confirm the correct molecular mass of the wildtype rCBD123 and change in mass after the lysine to alanine substitutions in the mutant proteins, the wildtype and all mutants were analyzed by electrospray mass spectrometry. Recombinant proteins (10 pg) were desalted and concentrated on a PLRP-S column and then injected directly for analysis on a SCIEX API 300 (Perkin Elmer). Circular Dichroism Spectral Analysis—To ensure that the mutations did not introduce structural alterations relative to the wildtype protein, far UV circular dichroism spectral analyses were performed on the wildtype and mutant proteins. Measurements were made using a Jasco Model J-720 spectropolarimeter calibrated with ammonium-d-camphor-10-sulfonate. Protein samples (0.5 mg/ml) were equilibrated against 10 mM NaPj, pH 7.0, and placed in a water-jacketed, cylindrical quartz cuvette (pathlength 0.1 cm) that was temperature regulated with a Neslab Model RT-110 circulating water bath. All spectra (190-250 nm) were collected at 25 °C using a scan rate of 20 nm/min, a 16-s time constant, and a spectral slit of 0.5 nm. Measurements were accumulated over five scans and computer averaged. The percentage distributions of secondary structure elements were estimated with a 5% error level from the CD spectra using the K2D program (Andrade et al, 1993). This is a non-linear method that evaluates the secondary structure with a neural network approach and extracts secondary structure information from a data set of 200-240 nm CD spectra of 24 standard proteins. Chemical Modification of Lysines—Lysines in the wildtype rCBD123 were acetylated with acetic anhydride (Habeeb and Atassi, 1970; Isaacs et al, 1989) after equilibration in reaction 101 Heparin binding by the gelatinase A CBD buffer (0.15 M NaCl, 0.2 M NaHC03, pH 8.5) on size exclusion spun columns containing Sepharose G-10 (V tl ml). Acetylation was performed by incubation with 10- to 100-fold molar excess of acetic anhydride over the lysine content in rCBD123 for 6 h at 22 °C. The acetylated protein was separated from the reaction mixture by spun columns equilibrated in 50 mM Tris, pH 7.4 and then analyzed by SDS-PAGE to confirm the upshift in apparent molecular weight induced by lysine acetylation. Affinity Chromatography—To evaluate any differences in the interactions of the wildtype and mutant rCBD123s with extracellular matrix proteins, affinity columns were prepared. Heparin-Sepharose CL-6B and type I gelatin-Sepharose 4B were from Pharmacia. Elastin mini columns (Vt 100 ul) consisted of a 1:1 mixture of elastin and Sephadex G-10 (Sigma). A standardized amount of wildtype or mutant rCBD123 (25 ug, 1.2 nmol) in 100 ul was loaded on each mini column in 50 mM Tris buffer, pH 7.4 as described previously (Chapter 2, Steffensen et al., 1995; Overall et al., 1989). Experiments for each matrix were typically performed in duplicate in at least two separate experiments. Elution strategies included step-gradients of NaCl from 0.1 to 1.0 M and DMSO from 1 to 10% (v/v) in 50 mM Tris buffer, pH 7.4. Thorough washes (20 - 40 x Vt) with chromatography buffer were performed after sample loading and between different eluants. Fractions were collected and analyzed by SDS-PAGE at a constant dilution relative to the loaded sample volume to facilitate assessment of protein recovery. Columns eluted with DMSO were not reused. Electrophoresis and Western Blotting—Proteins were separated by SDS-PAGE according to Laemmli (Laemmli, 1970) using cross-linked polyacrylamide minislab gels. Protein samples were analyzed without reduction or with the addition of 65 mM DTT and heating at 95 °C for 5 min. Proteins were detected by staining with Coomassie Brilliant Blue R-250. For Western blot analyses recombinant proteins were transferred to Immobilon-P PVDF membranes (Millipore) after separation by SDS-PAGE and then detected with aCBD123 (see below) or with affinity purified oc(His)6 antibody raised in rabbits against the (His)6-containing NH2-terminal fusion peptide (Wallon and Overall, 1997). Bound secondary 102 Heparin binding by the gelatinase A CBD anti-rabbit antibodies conjugated with horseradish peroxidase were detected using enhanced chemiluminescence reagents and Hyperfilm (Amersham). Reduced high range molecular weight markers used were rabbit muscle phosphorylase B (97 kDa), bovine serum albumine (67 kDa), chicken egg ovalbumin (43 kDa), bovine carbonic anhydrase (29 kDa), horse heart myoglobin (18.4 kDa), chicken egg-white lysozyme (14.4 kDa), and bovine insulin (6.2 kDa) (Sigma). Prestained, high range molecular standards (GIBCO BRL) were used for Western blotting. Polyclonal Antibodies—Polyclonal antibody (aCBD123) was raised by injecting rabbits with bacterial inclusion bodies of rCBD123 together with purified rCBD123 adsorbed to aluminum hydroxide (Harlow and Lane, 1988; Glenny et al, 1926). Inclusion bodies were prepared from bacterial lysozymal lysates by extraction with 5% (w/v) sarcosyl, 0.5 M NaCl, 20 mM NaPj, pH 8.0 and separated by alternating centrifugation at 10,000 x g at 4 °C and resuspension until SDS-PAGE analysis showed the preparation to be free of residual bacterial host proteins. After developing a high titer immune response to rCBD123 (as determined by ELISA), whole sera from rabbits were collected and polyclonal antibody was purified by affinity chromatography over rCBD123 coupled to AffiGel 10. Affinity purified aCBD123 reacted specifically with rCBD123 but not with bacterial proteins in western blots of the bacterial lysate from which the rCBD123 was prepared. ccCBD123 also did not cross-react with BSA, myoglobin, or rat tail type I collagen in ELISA assays. Microwell Substrate Binding Assay—Quantitation of the wildtype or mutant rCBD123 binding to extracellular matrix proteins was by the microwell substrate binding assay (Chapter 2, Steffensen et al, 1995). In brief, proteins were coated in 96-microwell plates for 18 h at 4 °C with 10 pmol protein/well (typically 0.5 pg). Myoglobin and BSA served as negative control proteins. Washed plates were then blocked with 2.5% (w/v) BSA in PBS for 1 h at 22 °C. After thorough rinses with PBS, serially diluted wildtype and mutant rCBDs in 50 mM Tris, pH 7.4 (2 x 10"6 to 2 x 10"9 M) were added for 1 h at 22 °C. After further rinses, bound proteins were detected with a 1/250 dilution of the ct(His)6 polyclonal antibody 103 Heparin binding by the gelatinase A CBD and secondary alkaline phosphatase-conjugated goat anti-rabbit antibody (IgG H+L) (BioRad) (1/500) with p-nitrophenyl phosphate di-sodium (Sigma) added as substrate. The color intensity was determined at 405 nm in a microplate reader (Thermomax, Molecular Devices) and binding expressed relative to saturation level binding of wildtype rCBD123 (=1.0). Negative controls consisted of reaction mixture minus rCBD123, primary antibody, or secondary antibody. The reactivity of the antibodies to wildtype and mutant proteins was equivalent as assessed by determining the reaction intensity using a constant concentration of the primary antibody added to serially diluted rCBD123 and mutant proteins (Appendix 2). RESULTS Expression and Characterization of Recombinant rCBD123 Lysine Mutants—Wildtype, four single-lysine mutants (K263A, K330A, K343A, K357A), and one double lysine mutant (K298/299A) were expressed in E. coli using expression and purification protocols that previously have yielded functional rCBD123 protein (Chapter 2, Steffensen et al., 1995). The purified mutant proteins migrated with apparent molecular weights that were very close to that of wildtype rCBD123 (22.1 kDa reduced and 21.1 kDa non-reduced) (Fig. 3.1 A). All recombinant proteins were identified in Western blot analyses by the polyclonal antipeptide antibody oc(His)6 which reacts with the (His)6-containing NH2-terminal fusion tag (Fig. 3. IB), and also by the new affinity purified polyclonal antibody, aCBD123 (Fig. 3.1C). Interestingly, a(His)6 reacted stronger with the reduced form of the proteins whereas the aCBD123 preferentially recognized conformational dependent epitopes present on non-reduced protein. The electrophoretic behavior of the double-mutant K298/299A deviated from wildtype rCBD123 and electrophoresed with an apparent molecular weight of 23.5 kDa reduced and 20.3 kDa non-reduced. To ensure that this did not represent a misfolded form of the protein, K298/299A was expressed and purified twice from separate clones but in each case identical electrophoretic behavior was observed. That the slight shifts in migration of some mutants 104 Heparin binding by the gelatinase A CBD were apparently the result of the charge differences introduced by the mutations and not unexpected molecular weight changes was verified by electrospray mass spectrometry which showed well-defined peak distributions consistent with a single species of protein in all preparations. These analyses also confirmed the molecular mass predicted from the amino acid sequences and verified a change in mass of 57 Da for each lysine to alanine substitution (Table 3.1). The results indicated that the N-terminal methionine had been processed in the wildtype and mutant proteins. The CD spectra of rCBD123 and mutant proteins shown in Fig. 3.2 revealed the presence of 6-sheet components and an absence of a-helix content. Secondary structure elements were estimated to be 48% 6-component and 52% random coil. The spectra showed no differences between wildtype and mutant proteins except for K263A (Fig. 3.2). Although the scans pointed to a slight increase in the 6-sheet component for this mutant, the changes were not detectable by quantitation of secondary structure components. rCBD123 Lysines Contribute to Heparin Binding—rCBD123 binds heparin specifically by an ionic interaction requiring elution with 0.25 M NaCl to disrupt the interaction (Fig 3.3A). To determine the potential contribution of lysines in heparin binding, we first acetylated these basic residues in the rCBD123. Analysis by affinity chromatography demonstrated complete loss of binding to heparin-Sepharose by acetylated rCBD123 (Fig. 3.3B) compared to untreated wildtype rCBD123 (Fig. 3.3A). Acetylated rCBD123 was quantitatively recovered in the unbound and first wash fractions with no recovery of any protein following elution with up to 1 M NaCl or 10% DMSO in chromatography buffer. The elution fractions were analyzed both by Coomassie Brilliant Blue staining and, to avoid false negative results, by the more sensitive Western blot analysis (Fig. 3.3B). This experiment demonstrated that lysines residues are crucial to heparin binding by the collagen binding domain. To define the contributions of individual rCBD123 lysines to heparin binding, mutants of rCBD123 containing substitutions of the positively charged lysines to uncharged hydrophobic alanines were analyzed in affinity chromatography assays. Four of the five 105 Heparin binding by the gelatinase A CBD mutant proteins, K263A, K298/299A, K330A, and K343A, bound specifically to heparin-Sepharose affinity columns (Fig. 3.4A) and displayed elution profiles unaltered from that of wildtype rCBD123; overloaded protein was retarded by the heparin column and was released in wash fractions with bound protein being eluted by 0.25 M NaCl. In contrast, the mutant K357A protein did not bind heparin (Fig. 3.4B) with all loaded protein being recovered in the unbound and initial wash fractions (N=3, n=4) showing the pivotal contribution of this lysine to heparin binding. rCBD123 Lysine Contributions to Collagen and Elastin Binding-—We also investigated the binding of acetylated rCBD123 to gelatin-Sepharose. Although acetylated rCBD123 bound specifically to gelatin-Sepharose affinity columns, binding was reduced with the protein consistently eluting at a lower DMSO concentration (2 to 3% (v/v)) (Fig. 3.5A) than that required for elution of wildtype rCBD123 (3 to 4% (v/v)) (Fig. 3.5B). In order to determine if K357A, the heparin binding knockout mutant, also showed reduced gelatin binding, the gelatin-Sepharose elution profiles of this and the other mutants were determined. K357A (Fig. 3.6B), like K330A, K343A, and K298/299A, all showed unaltered gelatin binding properties. However, K263A, repeatedly showed a reduced affinity for gelatin with a peak elution at 3% (v/v) DMSO compared to 3 to 4% (v/v) for wildtype rCBD123 (Fig. 3.6A, C). To quantitate more precisely the interactions of the rCBD123 mutant proteins with gelatin, the more sensitive and quantitative microwell substrate binding assay (Chapter 2, Steffensen et al., 1995) was applied for studies of binding to both native and denatured type I collagen. The Kds of interaction between mutants and type I collagen (Fig. 3.7), as determined by four-parameter curve fitting procedures, showed very small deviations from the wildtype rCBD123, all being in the ranges of 2.0-4.7 x IO"7 M and 3.1-6.4 x 10"7 M for native and denatured forms of type I collagen, respectively. However, the level of saturation of K263A binding was consistently lower than the wildtype and the other mutants by approximately 25% to native (Fig. 3.7A) and 20% to denatured (Fig. 3.7B) type I collagen. Together with the affinity column elution profiles, this points to K263 being involved in the 106 Heparin binding by the gelatinase A CBD binding of rCBD123 to type I collagen. In addition, the binding interactions of the mutant proteins with denatured types IV and V collagen (Fig. 3.8) showed only small differences in the Kds that were in the ranges of 3.3-7.0 x 10"7 and 5.9-11.7 x 10"7 M, respectively. Since neither K263A nor the other mutants had altered levels of saturation binding to type IV (Fig. 3.8A) or type V collagen (Fig. 3.8B) this indicates that the effect of collagen binding by the lysine substitution K263A is selective to type I collagen. The binding of wildtype and mutant rCBD123 proteins to elastin was also compared. All proteins had the same elution profile from elastin affinity columns with a peak elution at 1% (v/v) DMSO in the chromatography buffer with little or no difference between the mutants (Fig. 3.9). DISCUSSION We have previously identified the heparin binding properties of the CBD and COOH terminal domain of human gelatinase A (Wallon and Overall, 1997; Overall et al., 1997; Chapter 2, Steffensen et al., 1995). Based on the requirement for basic residues in canonical heparin binding sites, this study focused on the surface exposed lysine residues of the gelatinase A CBD. Chemical treatment with acetic anhydride to preferentially acetylate lysines and to block their contribution to ligand binding fully abolished rCBD123 binding to heparin-Sepharose and reduced binding to denatured type I collagen. This pointed to a functional contribution of lysine residues in the heparin and type I collagen binding properties of the CBD. Nonetheless, the positions of the lysines and arginines in the CBD do not conform to any of the published heparin binding consensus sequences. However, recently Busby et al. (1995) observed that the known consensus sequence, XBBXBX, present in the heparin binding site of the fibronectin type III-13 module was not sufficient for heparin binding. Three-dimensional modeling of this module based on a homologous region of tenascin (Leahy et al., 1992) together with mutagenesis studies revealed that basic residues from distant sites in the primary structure of the fibronectin type III-13 module were brought 107 Heparin binding by the gelatinase A CBD into close proximity in the folded protein to form a heparin binding site (Busby et al., 1995). Because there is no contiguous heparin binding sequence in the CBD and no patch of positively charged residues can be predicted in the tertiary protein structure of single modules (Constantine et al, 1992) it is likely that basic residues from two or all three of the fibronectin type II-like modules of the CBD contribute to forming a heparin binding site. Therefore, in these studies we analyzed the tri-modular rCBD123 protein. In addition, we have recently observed that bi-modular gelatinase A CBD constructs consisting of modules 2 and 3 possess heparin binding properties unaltered from rCBD1232. Hence, we focused on substitution of all six lysines in these two CBD modules of the rCBD123 as a first step in determining more precisely the heparin binding site in the gelatinase A CBD. The K357A mutant was characterized by undetectable binding to heparin confirming the acetylation results regarding the importance of lysines for CBD binding of heparin. Interestingly, the affinity of this mutant for other ligands (types I, IV, and V collagens, and elastin) did not deviate from the wildtype protein indicating that the effects were selective and not global. That the absence of heparin binding in K357A was not merely due to a reduction in net positive charge of the protein was further shown by analysis of three additional single substitutions (K263A, K330A, K343A) and one double mutation (K298/299A) which showed unaltered heparin binding properties. This provides further support for the specificity of the interactions and the importance of K357 in heparin interactions. Although the mutant proteins K357A and K263A displayed reduced binding of only specific ligands (heparin and type I collagen, respectively) the introduction of alanines in place of lysine could have induced structural alterations despite the low probability of steric clashes from alanine substitutions (Cunningham and Wells, 1989). However, the far UV CD spectra for rCBD123 and all mutant proteins reported here were comparable to those recorded for the gelatinase A CBD by others (Banyai et al, 1996). In the present experiments, the secondary structural elements were estimated to be 48% 6-component and 52% random coil for the wildtype and all the mutant rCBD123s. Slight differences in CD spectra were detected only for K263A 108 Heparin binding by the gelatinase A CBD where a small increase at the 198 nm and a decrease at the 224 nm maxima indicated a potentially increased amount of 6-sheet component in this mutant protein. Helical fractions can generally be quantitated with good accuracy from CD spectra, but the estimates of 8-sheet and random coil components are inherently less precise (Greenfield, 1996; Woody, 1995). Therefore, although quantitation of the secondary structural component content was unable to distinguish changes, a minor structural perturbation cannot be excluded as a contributing factor for the altered type I collagen binding properties of K263A. Although the three-dimensional structure of the CBD has yet to be solved, the structural coordinates from the PDC-109 fibronectin type II-like module b (Constantine et al, 1992; Constantine et al, 1991), which has a high degree of amino acid identity with the three CBD modules of gelatinase A (40%, 34%, and 48%, respectively), was used to model the CBD modules. PDC-109 module b is characterized by four 6-strands that form two central antiparallel 8-sheets with no oc-helices. The 6-strands are connected by two irregular loops that define a large, partially exposed, hydrophobic surface which may be a potential ligand binding site. In addition, two disulfide bonds connect cysteine residue pairs 1-3 and 2-4 which are highly conserved among the fibronectin type II-like domains. This model predicts that in CBD module 2 K263 is in the second position NH2-terminal to 6-strand 1 and exposed on the opposite side of the molecule relative to K298 and K299 which are located in the loop connecting 6-strands 3 and 4. K298 and K299 are exposed on the same surface of the module as R281, which is positioned in the loop connecting 6-strands 2 and 3, but separated from this residue by the hydrophobic groove and with a distance of 10-14 A. Collier et al. (1992) previously have shown that substitutions of charged residues in these two loops of gelatinase B CBD module 2 were effective in reducing the gelatin affinity. However, these investigators modified only one lysine in gelatinase B, corresponding to K298 in the gelatinase A CBD, without a major effect on gelatin affinity. Because the double lysines in the positions 298 and 299 of the gelatinase A CBD potentially might constitute the double basic residue part of a heparin binding site (Cardin and Weintraub, 1989), a double alanine 109 Heparin binding by the gelatinase A CBD substitution was introduced here. However, the K298/299A did not have altered ligand binding properties for heparin, gelatin, or the additional ligands investigated in the present experiments and not considered in the previous mutagenesis study. Structure modeling of the CBD module 3 places K357, which was found to be crucial for heparin binding, adjacent to R356 in the loop connecting 6-strands 3 and 4. As in the second module of the CBD, the five basic residues in the third module are not predicted to form a positively charged cluster. In fact, K357 and R356 are predicted to be in close spatial proximity to three negatively charged aspartate residues in positions 353, 354, and 355. Two additional basic residues in module 3, K343, and R339, found in the loop between 6-strands 2 and 3 are separated by distances of approximately 20 A from K357 and are positioned on opposite aspects of the major hydrophobic surface. The third lysine in position 330 is found on the opposite side of the molecule relative to the other lysines and arginines. This distribution of the basic residues to distantly placed sites (10-20 A) of the CBD modules is unlike the positioning observed in the fibronectin type 111-13 module where the basic residues are folded together in close proximity to form a heparin binding site (Busby et al, 1995). Therefore, it is likely that a single module of the gelatinase A CBD is not by itself responsible for heparin binding and that several surface exposed basic residues from two or all three modules form the heparin binding site when packed together upon folding of the CBD. This is also supported by our preliminary observations that bi-modular CBD constructs have heparin binding properties (Overall et al, 1997; Abbey et al, 1997). Until the three dimensional structure of the CBD is determined, the structural nature of the heparin binding site in the CBD remains unknown. In the present ligand binding experiments, we determined that the saturation level binding between the K263A mutant protein and type I collagen was reduced. That lysines are important for collagen binding has also been shown in the type II-1 and II-2 modules of fibronectin by chemical modification (Isaacs et al, 1989). Moreover, the predicted position of K263 lies NH2-terminal to 6-strand 1 in module 2 placing this residue in proximity to the 110 Heparin binding by the gelatinase A CBD predicted positions of two alanines in the fibronectin type II-like module 1 of gelatinase B. This module in gelatinase B was found not to bind gelatin; however, substitution of these A,A in module 1 to K,R, as found in module 2 of gelatinase B, increased binding to gelatin (Collier et al, 1992). Thus, the presence of basic residues in this segment of the CBD modules appears to be important for collagen binding. The reduced binding to collagen for the K263A mutant protein occurred without a concurrent change in the Kd for the interaction relative to wildtype rCBD123. This may reflect the presence of more than one collagen binding site on the CBD which can compensate for a disrupted collagen binding site in module 2 involving K263. This notion is supported by our previous observation that rCBD123 can simultaneously bind at least two molecules of native or denatured type I collagen (Chapter 2, Steffensen et al. 1995). However, it is not yet fully understood how several modules contribute to each collagen binding site and whether the binding sites have different ligand binding specificities. From studies of gelatinase A single-, bi, and tri-modular CBD constructs (Overall et al, 1997; Abbey et al, 1997; Banyai et al, 1994), there is evidence for module cooperativity with strongest ligand affinity found in the intact tri-modular wildtype structure. In addition, we have found that rCBD12 and rCBD23 have reduced Kds for type I collagen binding by an order of magnitude compared to rCBD123, that CBD 12 does not bind type V collagen, and that elastin binding requires the presence of all three modules (Overall et al, 1997; Abbey et al, 1997). In comparison, the module interactions are different for gelatinase B CBD where module 2 alone can bind gelatin stronger than any combination of the modules including the wildtype tri-modular CBD (Collier et al, 1992). The principle of modular cooperativity for gelatinase A CBD has also been supported by Banyai et al. (1996) who have proposed a model in which the three fibronectin type II-like modules of gelatinase A together form a collagen binding hydrophobic groove lined by aromatic residues. Although this is an attractive model, it is associated with significant uncertainty to predict the tri-modular structure of the CBD and its relationship to the catalytic domain since the structural coordinates of the CBD have not yet been assigned. I l l Heparin binding by the gelatinase A CBD In addition, the model does not explain the potential for binding of two separate molecules of collagen simultaneously (Chapter 2, Steffensen et al., 1995). Together these mutational ligand binding experiments have demonstrated that lysine residues of the human gelatinase A CBD play important roles in the interactions with heparin and also for type I collagen. In particular, a single lysine substitution at position 357 was sufficient to abolish the binding of heparin. Overall, our results support the notion that heparin and collagen binding by the CBD occurs through contributions from the three modules of this domain. The interaction of gelatinase A via the CBD or the COOH domain (Wallon and Overall, 1997) with heparin, and most likely with heparan sulfate side chains of proteoglycans, has important biological implications. It provides a mechanism of localization for gelatinase A in the extracellular matrix or the pericellular matrix where the enzyme via the CBD can bind to collagen (Chapter 4) and possibly heparin and heparan sulfate. Indeed, heparin binding by gelatinase A has been shown to increase cell surface activation of the enzyme (Crabbe et al, 1993). Binding of gelatinase A to heparan sulfate side chains of proteoglycans might also facilitate basement membrane type IV collagen and proteoglycan core protein degradation to facilitate cellular invasion during tumorigenesis and angiogenesis. SUMMARY The matrix metalloproteinases are important for metastatic tissue penetration by tumor cells and also in inflammatory diseases and tissue remodeling. In these processes, binding of gelatinase A to extracellular matrix components is required for localization of the enzyme and degradation of substrates. To explore further the interactions of the gelatinase A fibronectin type II-like modules, that constitute the collagen binding domain, with substrates and ligands we investigated the role of lysine residues in CBD binding to heparin and extracellular matrix proteins. Lysine acetylation in recombinant CBD completely abolished binding to heparin and reduced the binding affinity for denatured type I collagen. Site-112 Heparin binding by the gelatinase A CBD specific lysine to alanine substitutions demonstrated that one lysine (K357) located in the third fibronectin type II-like module of the CBD is absolutely required for heparin binding. Circular dichroism spectral analysis confirmed that the mutant K357A and wildtype rCBD were characterized by comparable amounts of secondary structural elements. In addition, K357A displayed the same binding properties to type I, IV, and V collagens, and elastin as the wildtype protein. A second mutant protein, K263A was associated with a reduced level of binding at saturation to type I collagen but without a concurrent change in the K d which ranged from 2 to 6 x 10"7 M for native and denatured type I collagen. Neither of the mutants with altered heparin or type I collagen affinity nor three additional mutants (K330A, K343A, K298/299A) displayed changes in the binding properties to types denatured IV or V collagen, or to elastin. This indicates that the lysine mutations produced alterations in binding that were ligand specific. Interpretation of these results based on structural data for fibronectin type II-like modules points to cooperativity between the three modules of the gelatinase A CBD in the interactions with substrates of the enzyme and other ligands. 113 Heparin binding by the gelatinase A CBD TABLE 3.1 Molecular mass determination by electrospray mass spectrometry The molecular masses of the wild type recombinant collagen binding domain (rCBD123) and five lysine mutants determined by electrospray mass spectrometry. Proteins from wildtypeb Mass" Mass decrease compared to wildtype Predicted massc + N-terminal - N-terminal methionine methionine*1 rCBD123 21,218 - 21,361 21,212 K263A 21,160 58 21,304 21,155 K298/299A 21,105 113 21,247 21,098 K330A 21,161 57 21,304 21,155 K343A 21,159 59 21,304 21,155 K357A 21,162 56 21,304 21,155 a Mass (Da) measured by electrospray mass spectrometry on a PESCEEX API 300. b The decrease in mass by a lysine (K) to alanine (A) substitution is 57 Da. 0 Mass predicted from the amino acid sequence. d Predicted decrease in mass if N-terminal methionine is removed by intracellular processing. 114 ffff r , Mr+ + + + + + / / / / / / DTT B 97 — > , «M»' 67-? *** 43-» 29-4.**. rs;r::'S«iii(»'((ja 18.4 —> 14.4 Wjgm ^H9* g^pn T ^ ^ ^ ^ ^ P P r 97 _> 67 43-4 29-> 18.4-> 14.4-4 • «M» — mm *m mm 67 ^ 43-4 29 -> 18.4-> 14.4-4 FIG. 3.1. SDS-PAGE and Western blot analysis of wildtype rCBD123 and mutant proteins. Samples (1 (ig/lane) of wildtype (WT) and rCBD123 mutant proteins were analyzed by SDS-PAGE on 15% cross-linked polyacrylamide minislab gels under reducing C+DTT) or non-reducing (-DTT) conditions and stained with Coomassie Brilliant Blue R250 (Panel A). Proteins transferred to PVDF membranes for Western blot analysis were recognized by polyclonal anti-peptide antibody <x(His)6 raised against the (His)6-containing N-terminal fusion peptide present on both the rCBD123 and mutants (Panel B) and by aCBD123, a polyclonal antibody raised against rCBD123 (Panel Q as described under "Experimental Procedures" (Chapter 3). M„ molecular weight markers (kDa) are indicated. 115 FIG. 3.2. Far UV circular dichroism spectra for wildtype and mutant rCBD123 proteins. Spectra (190-250 nm) of protein samples (0.5 mg/ml) in 10 mM NaPi, pH 7.0 were collected at 25 oC with a scan rate of 20 nm/min, a 16 s time constant, and a spectral slit of 0.5 nm. The spectrum for each protein shows the unsmoothed average of five scans of each sample. 116 W NaCl D A M r B U 1 2 3 4 1 2 3 4 10 %7^~ ~ ~~~~ 4 3 - * 29-> 1 8 . 4 - » ^ > B 97-> 67 -> t o - * I 29-> 18.4-> 14.4-* rCBD123 Ac/rCBD123 FIG. 3.3. Heparin-Sepharose affinity chromatography of rCBD123 and acetyl-ated rCBD123. Protein samples (25 ng, 1.2 nmol in 100 (xl) were overloaded (B) onto the heparin-Sepharose columns (Vt 25 (il) which were then washed (W) extensively with chromatography buffer before elution with a step gradient of 0-1.0 M NaCl (NaCl, 1-4 corresponding to 0.25, 0.50, 0.75, and 1.0 M) and then 10% (v/v) DMSO (D) in the chroma-tography buffer. Samples were analyzed by 15% SDS-PAGE gels under reducing conditions. Untreated rCBD 123 (rCBD123) (Panel A) was retarded by the columns and was eluted by repeated washes (W; 1-4). However, complete elution of bound rCBD123 required 0.25 M NaCl (NaCl; 1). In contrast, all loaded (B) acetylated rCBD123 (Ac/rCBD123) (Panel B) was recovered in the unbound (U) or initial wash fractions (W; 1-2) with none recovered in 0.25 M NaCl (NaCl; 1). The elution profile for acetylated rCBD123 was analyzed by Western blotting (Panel B) to confirm results of Coomassie Brilliant Blue staining of 15% SDS-PAGE gels (not shown). The kDa of protein standards (Mr) are indicated. 117 §7=1 43-> 29-> 18.4-> 14.4 -> 6.2-> W NaCl DMSO B U 1 2 3 4 1 2 3 4 0 10 B •K330A 43-> 29-> 18.4-> 14.4-> 6.2-> <-K357A FIG. 3.4. Heparin affinity chromatography of rCBD123 lysine mutants. Equal amounts of wildtype rCBD123 and mutants (25 ug, 1.2 nmol in 100 ul) were loaded (fi) onto mini columns (Vt 50 ul) of heparin-Sepharose as described under "Experimental Procedures" (Chapter 3). Columns were washed thoroughly (W) and a step gradient of 0-1 M NaCl (NaCl, 1-4, corresponding to 0.25, 0.5, 0.75, and 1.0 M) was applied followed by 10% (v/v) DMSO (%DMSO) in chromatography buffer. Fractions were collected and analyzed by SDS-PAGE on 15% minislab gels under reducing conditions. Mutant K330A (Panel A) is representative of four mutants (K263A, K298/299A, K330A, K343A) which had elution profiles identical to that of wildtype rCBD123. In contrast, the mutant K357A (Panel fi) did not bind heparin, with all loaded protein being quantitatively recovered in the unbound (U) and initial wash fractions (W; 1). The kDa of molecular weight standards (Mr) are shown. 118 W %DMSO W t B U 1 2 N 0 1 2 3 4 5 6 10 97. 67. 43-29-18.4 14.4 B <-Ac/rCBD123 671; 43-> 29-> 18.4-» 14.4-» -rCBD123 FIG. 3.5. Denatured type I collagen affinity chromatography of acetylated rCBD123. rCBD123 was acetylated as described under "Experimental Procedures" (Chapter 3) to discern the contribution of lysines to gelatin binding. Equal amounts (25 ug; 12 nmol) of acetylated (Ac/rCBD123) (Panel A) and wildtype rCBD123 (Panel B) were loaded (B) onto gelatin-Sepharose mini columns (Vt 25 ul), followed by thorough rinses (W) with chroma-tography buffer, before elution with 1 M NaCl (N) and then by a step gradient of 0 to 10% (v/v) DMSO (%DMSO) in chromatography buffer. All fractions were analyzed by SDS-PAGE on 15% minislab gels under reducing conditions. Peak elution of acetylated rCBD123 was at 2-3% (v/v) DMSO which was significantly lower compared to the wildtype rCBD123 which required 3-4% (v/v) DMSO for elution. Positions of molecular weight markers (kDa) are indicated. 119 W %DMSO A M r B U 1 2 N 0 1 2 3 4 5 6 97-4 "=r 67 -> f f " " : ; :- I 43 —> *•»• 29 -> **, ^ _ <-rCBD123 18.4 _> | p 14.4-* S 6.2-). B 97 -4 — 67 -» —> 43-* m 2 9 ^ _ — _ 4-K357A 18.4 _> M 14.4-> g 6.2-> c 9 7 67=4 * 43 -4 H 18.4_> 14.4-> 6.2-* K263A FIG. 3.6. Interactions of wildtype and mutant rCBD123 proteins with denatured type I collagen. Wildtype and mutant rCBD123 (25 ug, 1.2 nmol in 100 ul) were chromatographed over gelatin-Sepharose columns (Vt 25 ul) as described under "Experimental Procedures" (Chapter 2). After loading (B) of the proteins, unbound material was collected (U) and the columns were then washed thoroughly with chromatography buffer (W) before elution with 1 M NaCl (AO and a step gradient from 0 to 10% (v/v) of DMSO (%DMSO) in chromatography buffer. Analysis of elution fractions by SDS-PAGE on 15% minislab gels under reducing conditions demonstrated that wildtype rCBD123 (Panel A) and four of five mutants (K298/299A, K330A, K343A, K357A) had identical elution profiles - the heparin binding-deficient mutant K357A is shown (Panel B). In comparison, the mutant K263A was eluted from gelatin-Sepharose at the lower concentration of 3% (v/v) DMSO (Panel Q. The positions of rCBD123, mutant proteins, and the marker protein standards (kDa) are indicated. 120 Native Type I Collagen Denatured Type I Collagen K„ (xlO-8 M) o Wildtype 4.1 " • K263A 6.0 • K298/299A 6.4 10"9 10"8 10' 7 10"6 10"5 Protein Concentration (M) FIG. 3.7. Interactions of rCBD123 and mutant proteins with native and denatured type I collagen. In a microwell substrate binding assay, equal amounts of native or denatured type I collagen were coated in 96-microwell plates (10 pmol/well) and assayed for interaction with wildtype or mutant rCBD123 proteins added at a concentration range of 2 x 10"6 to 2 x 10"9, and 0 mM as described under "Experimental Procedures" (Chapter 3). Data points are averages of two separate experiments from the same plate and representative of five separate experiments. 121 Denatured Type IV Collagen 1.6 1.4 E5 1-2 H .£ 1.0 H tn Wildtype K263A K298/299A K330A K343A K357A 4.9 6.1 7.0 6.7 3.8 3.3 IO"3 10" io-: Denatured Type V Collagen 10"! 10"' 10"' 10-Protein Concentration (M) 10": FIG. 3.8. Binding of rCBD123 and mutant proteins to denatured types IV and V collagens. Serially diluted rCBD123 and mutant proteins (2 x 10"6 to 4 x 10"9, and 0 mM) were added to denatured types IV (Panel A) or V collagen (Panel B) coated at equal amounts (0.5 mg/well) as films in 96 well microtiter plates as described under "Experimental Procedures" (Chapter 3). Data points are means of two experiments from the same plate and representative of four separate experiments. 122 B W %DMSO H B U 1 2 3 N 0 1 2 3 4 5 9 7 , „ ,. 67=? 4 3 - » * 29 - » 18.4 _» ^ 1 4 . 4 = ? ^ <-rCBD123 # # ^ ^ 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 %DMSO 29 18.4 _> 14.4-> FIG. 3.9. Elastin binding properties of wildtype and mutants of rCBD123. Equal amounts (25 p.g, 1.2 nmol in 100 ul) of wildtype rCBD123 (Panel A) and five mutant proteins (Panel B) were loaded (B) onto elastin mini columns (Vt 100 ul; ratio of elastin:Sephadex-G 10 = 1:1) as described under "Experimental Procedures" (Chapter 3). Unbound protein (U) was collected and columns were then washed thoroughly with chromatography buffer (W; 1-3) and eluted first with 1 M NaCl (AO and then with a step gradient of 1-5% (v/v) DMSO in chromatography buffer (%DMSO; 0-5). Column fractions were analyzed under reducing conditions by SDS-PAGE using 15% cross-linked minislab gels. All proteins were eluted with a peak at 1% (v/v) DMSO (%DMSO) in chromatography buffer with no difference detected between wildtype (Panel A) and any of the mutants (Panels B). (U, unbound) The positions of rCBD123 and mutant proteins and kDa of molecular mass standards are indicated. 123 CHAPTER 4 Integrin Dependent Cell Binding of Human Gelatinase A by the Fibronectin Type II-like Modules INTRODUCTION Gelatinases A and B bind TIMP with the C-domain (Overall et al, 1997; Wallon et al, 1997; Fridman et al, 1992; Murphy et al, 1992; Howard and Banda, 1991; Goldberg et al, 1989) and interstitial and neutrophil collagenases utilize the C domain for binding and cleavage of native type I collagen, but not synthetic peptide or casein substrates (Schnierer et al, 1993; Murphy et al, 1992; Windsor et al, 1991; Clark and Cawston, 1989; Hasty et al, 1987). A different collagen binding domain (CBD) is found exclusively in gelatinases A and B (Steffensen et al, 1995; Banyai et al, 1994; Collier et al, 1992; Banyai and Patthy, 1991) and consists of three fibronectin type II-like modules inserted into the catalytic domain immediately NH2-terminal to the active site Zn2+-binding a-helix active site (Collier et al, 1992; Wilhelm et al, 1989; Collier et al, 1988). In addition to type I gelatin binding properties (Steffensen et al, 1995; Banyai et al, 1994; Murphy et al, 1994; Collier et al, 1992; Banyai and Patthy, 1991), our characterization of isolated recombinant CBD (rCBD123) from human gelatinase A (Chapter 2, Steffensen et al, 1995) demonstrated that this domain accounts for most, if not all, of the native and denatured type I collagen binding properties of human gelatinase A and also provides the enzyme with binding specificity to denatured types IV, V, and X collagens, native types V and X collagens, elastin, and heparin (Abbey et al, 1997; Overall et al, 1997; Steffensen et al, 1995). The importance of the domain is also indicated by deletion of the CBD which reduces the catalytic activity of gelatinase A for denatured type I collagen by 90% (Murphy et al, 1994) and abolishes both binding to and cleavage of elastin in gelatinases A and B (Shipley et al, 1996). Localized control of extracellular matrix degradation in the vicinity of the cell surface is important in pathological processes and also in tissue remodeling. Plasma membranes of various human cancer cells contain high levels of collagenolytic and gelatinolytic activities 124 Gelatinase A cell binding via the CBD (Emonard et al, 1992; Whitelock et al., 1991; Zucker et al., 1990; Zucker et al, 1987). Indeed, the biological behavior of such cancer cell lines has revealed a positive correlation between the expression levels of gelatinase A and the invasive potential (Zucker et al, 1992). Moreover, certain tumor cell lines, which themselves do not express gelatinase A, bind the enzyme to their cell membranes by a specific membrane-associated receptor (Emonard et al, 1992). Studies demonstrating activation of progelatinase A by cell membranes of concanavalin A (Con A) (Overall and Sodek, 1992; Brown et al, 1990; Overall and Sodek, 1990) or 12-0-tetradecanoylphorbol-13-acetate (TPA) (Ward etal, 1991; Brown etal, 1990) stimulated normal cells pointed to a specific mode of enzyme-cell interaction that utilizes the C domain of gelatinase A (Ward et al, 1994; Murphy et al, 1992) and the TIMP-2 (Ward et al, 1991). Recently, four MT-MMPs have been isolated and cloned (Puente et, al, 1996; Takino et al, 1995; Will and Hinzmann, 1995; Sato et al, 1994) which possess a hydrophobic transmembrane domain. MT-MMPs can activate progelatinase A at the cell surface (Strongin et al, 1995; Ward et al, 1994) in a proposed activation complex of gelatinase A, TTMP-2, and MT-MMP (Strongin et al, 1995). However, questions remain regarding the precise binding mechanisms between these components and, indeed, alternative interactions of the progelatinase A C-domain with other cell surface components such as specific TIMP-2 receptors (Emmert-Buck et al, 1995) and 0^63 integrin receptors (Brooks et al, 1996) have been identified. Interestingly, the 0 ^ 3 integrin receptor can also increase gelatinase A expression when bound to vitronectin or blocking antibodies (Seftor et al, 1993; Seftor et al, 1992). Thus, discerning the mechanisms of gelatinase A - cell binding is important for understanding the role of gelatinase A in these processes. Since the gelatinase A CBD is important for the extracellular matrix protein binding properties of gelatinase A we have previously proposed that such interactions may serve to localize the enzyme in tissues (Overall et al, 1997; Chapter 2, Steffensen et al, 1995). Because binding of gelatinase A to pericellular matrix proteins may also provide a mode of cell surface enzyme localization, we have studied the interaction between the gelatinase A 125 Gelatinase A cell binding via the CBD CBD and human fibroblasts as well as components of the pericellular matrix and cell membrane. Here, we report experiments which establish for the first time that a domain other than the C-domain of gelatinase A binds the enzyme to cell surfaces. The CBD was found to localize gelatinase A to cell surfaces by the formation of a gelatinase A/type I collagen/lV integrin receptor complex. EXPERIMENTAL PROCEDURES Preparation of Recombinant Collagen Binding Domain and Extracellular Matrix Proteins— rCBD123, encoded for by exons 5, 6, and 7 of gelatinase A, was expressed in E. coli and purified as described previously (Chapter 2, Steffensen et al., 1995). Correct molecular mass and homogeneity of recombinant protein preparations were verified by electrospray mass spectrometry using samples of protein desalted and concentrated on a PLRP-S column and then injected directly for analysis on a SCIEX API 300 (Perkin Elmer) mass spectrometer. Acid-soluble native type I collagen was prepared from rat tail tendons as described by Piez (Piez, 1967) by extraction with 0.5 M acetic acid and differential precipitation with 1.7 M NaCl. Bovine plasma fibronectin was purchased from Sigma and Chemicon. Porcine intestinal heparin was from Sigma, and bovine serum albumin (BSA) (fraction V, heat shock) was from Boehringer-Mannheim. Antibodies—Polyclonal antibody (aCBD123) was raised in rabbits against rCBD123 purified as described previously (Chapter 3) and injected with sarcosyl-extracted rCBD123 inclusion bodies. The aCBD123 antibodies were affinity purified over columns of rCBD123 coupled to AffiGel 10 (BioRad) and sensitivity and specificity of the aCBD123 confirmed using western blots of bacterial lysates from which rCBD123 was purified (not shown). There was also no cross-reaction to BSA, myoglobin, and rat tail type I collagen in enzyme-linked immunoabsorbance assays. Affinity purified anti-peptide polyclonal antibody (a(His)6) to the (His)6 sequence was prepared as described (Wallon and Overall, 1997). Monoclonal blocking antibody to the byintegrin subunit (MAM3) was a kind gift from Dr. K. M. Yamada 126 Gelatinase A cell binding via the CBD (National Institute of Dental Research, NIH, Bethesda, MD) and Oj-integrin subunit blocking antibody (P1E6) was from Life Technologies. Cell Culture—Early passage human gingival fibroblasts, isolated as described previously (Brunette et al., 1976), were kindly provided by Drs. D.M. Brunette and H.S. Larjava (University of British Columbia, Vancouver, Canada). Cells of passage 4-10 were used in all experiments. Cells were maintained subconfluent in a-minimal essential medium (a-MEM) (Gibco) containing 10% (v/v) newborn calf serum (Gibco) and antibiotics (100 ug/ml penicillin-G, 50 ug/ml gentamycin-sulfate (Sigma)) at 37 °C in a humidified atmosphere of 5% C0 2 and 95% air. To minimize proteolysis of cell membrane proteins, cells were harvested for cell attachment assays by addition of 0.2 mM EDTA supplemented with a low concentration of trypsin (0.05% (w/v)) in PBS for 30 - 60 seconds. Cell Attachment Assay—Interactions between rCBD123 and cell membrane components were analyzed using cell attachment assays. Tissue culture surface treated 96-microwell polystyrene plates were coated with two-fold serially diluted rCBD123 (25 to 0.25, and 0 ug/ml) in 100 ul phosphate-buffered saline (PBS) per well for 18 h at 4 °C unless otherwise indicated. After blocking with 10 mg/ml BSA for 30 min, 4 x 104 human fibroblasts were added per well in serum-free a-MEM (to avoid contributions to cell attachment from serum proteins) and incubated for 90 min at 37 °C. Cells were then thoroughly rinsed with PBS and fixed with 4% (v/v) formaldehyde in PBS. The number of attached cells was quantitated by staining with 0.1% (w/v) crystal violet in 200 mM boric acid, pH 6.0 (Keung et al., 1989; Gillies et al, 1986). After rinses to eliminate free stain, cellular stain was dissolved in 10% (v/v) acetic acid and the optical density was measured at 590 nm in a microplate reader (Thermomax, Molecular Devices). Positive control wells were coated with fibronectin or collagen, or were non-blocked wells. Cell attachment to BSA-blocked wells in the absence of experimental protein coating was used as a negative control and, for data analyses, served to adjust for any non-specific cell attachment. Experiments were always performed in duplicate or triplicate and repeated several times, but results were only compared for 127 Gelatinase A cell binding via the CBD experiments on the same plate. Cell Morphology and Spreading Characterization—To determine the spreading characteristics of fibroblasts, 5 x 103 cells were seeded per well in 96-microwell plates on surfaces coated with two-fold serially diluted (50 to 0.25, and 0 ug/ml) rCBD123 or fibronectin. After 30, 60, and 120 min of incubation at 37 °C, cells were fixed in a final concentration of 4% (v/v) formaldehyde for 30 min at 22 °C and spreading was quantitated by phase-contrast microscopy as the proportion of spreading cells to total cells seeded (Hakkinen et al, 1994; Hahn and Yamada, 1979). Cells surrounded by a ring of lamellar cytoplasm were considered to be spreading. The morphology of fibroblast attachment and spreading on rCBD123 and fibronectin was further analyzed by scanning electron microscopy (SEM). Glass cover slips (1 cm2) were coated with rCBD123 or fibronectin and blocked with BSA prior to seeding 5 x 103 cells per slide in serum free a-MEM. After 1 or 2 h of incubation at 37 °C, the cells were rinsed with serum-free a-MEM and fixed with 2.5% (v/v) glutaraldehyde in PBS. The slides were subsequently processed for SEM by staining with 1% (w/v) osmium in PBS and then 2% (w/v) tannic acid, dried by critical point drying, and sputter-coated with gold. Specimens were analyzed on a Stereoscan 260 (Cambridge Instruments). Mechanisms of Cell Attachment—To determine the specificity of cell binding to rCBD123, cellular binding sites for rCBD123 were blocked by incubation of fibroblasts with a two-fold serial dilution of rCBD123 (50 to 0.25, and 0 ug/ml) in serum-free a-MEM for 1 h at 37 °C prior to cell attachment assays in wells coated with rCBD123. To analyze whether pericellular collagen or cell surface heparan sulfate proteoglycans may be involved in mediating the interaction between cells and rCBD123, 7.5 to 0.075 U/100 ul highly purified bacterial collagenase (clostridio-peptidase A, Type in, fraction A (EC 3.4.24.3), Sigma) was used to digest pericellular collagen. Pericellular heparan sulfate was digested with 0.1 and 0.01 U/ml highly pure heparinase (Flavobacterium. heparinum heparinase; Seikagaku Corporation) in the presence of 10 mM Ca2+-acetate and 0.1% (w/v) 128 Gelatinase A cell binding via the CBD BSA. After digestion in a-MEM at 37 °C for 30 min, enzymes were removed from the cells by repeated cell sedimentation (120 x g, 5 min) and resuspension in serum-free a-MEM (approximately 100,000-fold dilution) prior to cell attachment assays in wells coated with rCBD123 (25 ug/ml). Cell viability after enzyme treatments was determined by cell attachment in control wells of tissue culture treated plastic alone or fibronectin coated wells. To further assess the potential attachment of cells via pericellular heparan sulfate, wells were coated with rCBD123 and blocked with BSA and then incubated with 1 or 10 ug of heparin in 100 ul PBS per well for 1 h at 22° C to occupy rCBD123 heparin binding sites prior to cell attachment assays. PBS without heparin was added to control wells. Inhibition of cell attachment by blocking antibodies to the 6 r and the Oj-integrin subunits was analyzed to determine the potential role of this family of receptors in the rCBD123 to cell interaction. Fibroblasts were seeded in micro wells coated with 25 ug/ml rCBD123 in the presence of two-fold serially diluted (20 to 0.6, and 0 ug/ml) monoclonal antibody to the Bl-subunit (MAbl3) or ascites fluid containing antibody to the a2-subunit (P1E6) diluted 1/10 to 1/100. Affinity purified aCBD123 and a(His)6 served as control antibodies. Cell attachment was determined after 90 min incubation. Competition experiments—The presence or absence of gelatinase A bound to cell layers via the CBD was detected by competition experiments. Human fibroblasts were seeded in 96 microwell plates on tissue culture treated plastic surfaces. Quiescent cell layers were rinsed with PBS to remove unbound secreted enzyme that had accumulated in the conditioned medium. Gelatinase A was then competed from cell surfaces by incubation of cell layers with 250 or 25 ug/ml rCBD123 in serum free a-MEM for 5 min at 22 °C. Incubation with a-MEM alone served as control. After collection of released gelatinase A, the remaining cell-associated gelatinase A was collected from the cell layer by lysis with SDS-PAGE buffer, and both fractions were analyzed by enzymography. In a second experiment, fibroblasts were seeded in serum free a-MEM in uncoated plastic wells or wells coated with rCBD123 (25 ug/ml). After 18 h cells were rinsed three times with a-MEM and then incubated with or 129 Gelatinase A cell binding via the CBD without Con A (20 ug/ml) and soluble rCBD123 (100, 10, and 0 ug/ml) for 18 h. Conditioned medium was collected and non-reduced samples (-DTT) analyzed by enzymography. SDS-Polyacrylamide Gel Electrophoresis and Enzymography—Proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE) (Laemmli, 1970) after heating at 65 °C for 10 min with or without reduction by 65 mM dithioerythreitol (DTT). After separation, gels were stained with Coomassie Brilliant Blue R-250. For enzymography, non-reduced proteins were separated on 10% (w/v) polyacrylamide gels that contained 40 ug/ml gelatin and processed as described previously (Overall and Limeback, 1988). Gelatinase A was identified after counterstaining undigested gelatin with Coomassie Brilliant Blue R-250. The regular and prestained reduced high range molecular weight standards were from Gibco: Myosin (H-chain) (200 kDa), rabbit muscle phosphorylase b (97 kDa), BSA (67 kDa), chicken egg ovalbumin (43 kDa), bovine carbonic anhydrase (29 kDa), B-lactoglobulin (18.4 kDa) or horse heart myoglobin (18.8), and chicken egg-white lysozyme (14.4 kDa). Cell Protein Extraction and Ligand Blot Analyses—Confluent cultures of human fibroblasts were rinsed thoroughly with PBS and then lyzed with 50 mM octyl-6-D-thioglucopyranoside (Sigma) in PBS for 30 min at 15 °C (Pytela et al, 1987; Pytela et al, 1985). Cellular debris was sedimented by centrifugation at 10,000 x g for 15 min at 22 °C and detergent-solubilized protein was precipitated at 0 °C, collected by centrifugation at 10,000 x g for 10 min at 0 °C, and then stored at -20 °C until analyzed. Only minor amounts of protein remained in solution after precipitation (not shown). The protein pellet was dissolved in PBS, separated under non-reducing or reducing conditions by SDS-PAGE using 7.5% (w/v) gels, and transferred to Immobilon-P polyvinylidene difluoride (PVDF) membranes (Millipore). The membranes were incubated with 20 ug/ml rCBD123 in 150 mM NaCl, 10 mM Tris, pH 7.2 containing 0.2% (w/v) BSA for 1 h at 22 °C. After extensive washes, rCBD123 bound to proteins on the blot was detected using the aCBD123 antibody and enhanced chemiluminescence reagents (Amersham); 130 Gelatinase A cell binding via the CBD The rCBD123-binding proteins were characterized by digestion either alone or sequentially with pepsin (porcine stomach mucosal pepsin (EC 3.4.23.1), Sigma) and highly purified bacterial collagenase. Bacterial collagenase (4 U/100 ul) digestions were for 18 h at 37 °C, pH 7.0. Pepsin (0.1 mg/ml) digestions were at for 3 h at 15 °C, pH 2.0. An aliquot of the pepsin-treated sample was then brought to pH 7.0 and incubated with collagenase for 18 h at 37 °C. Control protein samples without proteases were exposed to identical pH and temperature conditions, and the efficiency of the enzyme digestions was also verified using BSA and type I collagen as substrates in control incubations. After digestion, the samples were separated by SDS-PAGE and probed for interaction with rCBD123 by ligand blotting as described above for non-protease treated detergent solubilized cell proteins. RESULTS Recombinant Protein Expression—The fidelity of expression of rCBD123 was confirmed by the predicted electrophoretic migration (22.1 kDa reduced and 24.1 kDa non-reduced) and absence of dimers and multimers as assessed by SDS-PAGE and Western blot with or without reduction by DTT (data not shown). Electrospray mass spectrometry of rCBD123 measured the mass to be 21,218 Da confirming N-terminal methionine processing of the recombinant protein (predicted mass 21,212 Da) and homogeneity of the protein preparation. The Collagen Binding Domain of Gelatinase A Mediates Cell Attachment—Human fibroblasts attached to rCBD123-coated microwells in a concentration dependent manner with half-maximal attachment found at a coating concentration of 150 nM (3.2 ug/ml) (Fig. 4.1 A). In comparison, cell attachment to fibronectin, used as a standard, was more efficient with a half-maximal concentration of 12 nM (5.3 ug/ml); approximately 1.4-fold more cells attached at saturation (Fig. 4.1B). To confirm the specificity of the cell interaction with rCBD123, incubation of fibroblasts with soluble rCBD123 (50 - 0.25 ug/ml in serum free a - M E M for 30 min at 37 °C prior to seeding in rCBD123 coated wells was found to inhibit attachment in a concentration dependent manner (Fig. 4.2). No attachment was observed in control wells 131 Gelatinase A cell binding via the CBD coated with 10 mg/ml BSA whereas cell attachment to tissue culture treated plastic alone or to type I collagen coated wells was equivalent to that on fibronectin under saturating conditions. Cell Spreading on rCBD123 is Numerically and Morphologically Different from Spreading on Fibronectin—After 30 min incubation, significantly fewer fibroblasts (23%) displayed cytoplasmic spreading on rCBD123 compared to fibronectin (50%) at coating concentrations of 10 ug/ml (Fig. 4.3A). Using 25 ug/ml coated protein, greater differences were manifest between rCBD123 and fibronectin with 23% and 90% of the cells spreading, respectively (Fig. 4.3B). After incubation for 60 and 120 min, maximal spreading reached plateau levels of 80 - 90% of all cells seeded on both substrates at each coating concentration. In addition to numerical differences, morphological differences between fibroblasts seeded on rCBD123 and fibronectin were apparent by phase contrast microscopy (not shown). Therefore, to gain detailed images of the cellular morphology, fibroblasts were studied by SEM (Fig. 4.4). After 1 h incubation, cells on fibronectin demonstrated typical extensive cytoplasmic spreading (Fig. 4.4A, C). In contrast, fibroblasts seeded on rCBD123 for 1 h had a rounded morphology with limited spreading of the cytoplasm (Fig. 4.4B). An apparent tighter adaptation to the surface and extension of delicate filopodia was observed after 2 h of incubation on rCBD123 (Fig. 4.4D) with the cells having a diameter of -50 um. This was clearly less than the typical diameter of -100 um for cells attached to fibronectin and resulted in a smaller total surface area for cells seeded on rCBD123. These results indicate that different attachment and spreading mechanisms are used by cells on rCBD123 compared with fibronectin. Anti-fil Integrin Antibody Blocks Cell Attachment to rCBD123—Cellular attachment to the major cell adhesion proteins including fibronectin and type I collagen is mediated primarily by integrin receptors containing the iVintegrin subunit (Hynes, 1992; Albelda and Buck, 1990). Therefore, to assess the potential for involvement of Bj-integrins in cell interactions with the CBD, cells were seeded in the presence of integrin subunit blocking antibodies in 132 Gelatinase A cell binding via the CBD rCBD123 coated wells and the effects on cell attachment analyzed. When 2.5 ug/ml blocking monoclonal antibody (MAb 13) specific for the Bj-integrin subunit was incubated with cells seeded in rCBD123-coated wells, a greater than 50% inhibition of attachment of fibroblasts to rCBD123 was observed that increased to 90% inhibition at concentrations exceeding 5 ug/ml (Fig. 4.5). In comparison, the a2-integrin subunit blocking antibody (P1E6), and affinity purified ocCBD123 and a(His)6 control antibodies had no significant blocking effects in this concentration range. Pericellular Collagen Mediates Cell Interactions with rCBD123—Since gelatinase A CBD binding to native type I collagen (Chapter 2, Steffensen et al., 1995) may represent a mode for cell attachment to rCBD123, plates coated with rCBD123 (coating concentration 1.5 and 25 ug/ml) and blocked with BSA were then incubated with soluble native type I collagen (1, 3, and 10 ug/well). A concentration dependent increase in cell attachment was found in rCBD123-coated wells that had been incubated with increasing amounts of native type I collagen (Fig. 4.6). At 10 ug collagen coupled to rCBD123 cell attachment levels approached that for wells coated with collagen alone. This indicates that integrin-bound pericellular collagen was the rate limiting component at early stages of cell attachment to rCBD123. To confirm the potential for cell bound collagen to mediate cellular attachment to rCBD123 and to discriminate from direct binding of rCBD123 with Bj integrins, cells treated with highly purified bacterial collagenase (0.075, 0.75, or 7.5 U/100 ul) for 30 min prior to plating in wells coated with 25 ug/ml rCBD123 showed a concentration dependent decrease in cell attachment (Fig. 4.7). 0.75 TJ collagenase was found to completely digest 100 ug purified native type I collagen after 15 min incubation at 37 °C and did not cleave the rCBD123 as assessed by SDS-PAGE (not shown). To confirm that this effect was due to collagen removal and not to degradation of integrins, cell attachment was rescued in a concentration dependent manner by addition of soluble native type I collagen to the rCBD123 coated wells (Fig. 4.7). The rescue was not complete, likely due to prolonged release of collagenase endocytosed during the enzymic treatment (Sodek and Heersche, 1981). Moreover, it is possible that the 133 Gelatinase A cell binding via the CBD a d d i t i o n o f n a t i v e t y p e I c o l l a g e n d i d n o t r e p l a c e a l l d i g e s t e d c o l l a g e n t y p e s p o t e n t i a l l y i n v o l v e d i n the c e l l b i n d i n g to r C B D 1 2 3 (Chapter 2, S t e f f e n s e n et al. 1995) . T h a t the d i g e s t i o n s w e r e n o t c y t o t o x i c w a s n o t o n l y c o n f i r m e d b y r e s c u e e x p e r i m e n t s , b u t a l so b y u n a l t e r e d c e l l a t t a c h m e n t to t i s sue c u l t u r e t reated p l a s t i c a l o n e a n d f i b r o n e c t i n c o a t e d w e l l s (not s h o w n ) . T o i d e n t i f y p r o t e i n s f r o m m e m b r a n e extract s o f h u m a n f i b r o b l a s t s that m a y b e i n v o l v e d i n the i n t e r a c t i o n s w i t h r C B D 1 2 3 , l i g a n d b l o t t i n g w a s p e r f o r m e d . O n t rans fer m e m b r a n e s , r C B D 1 2 3 r e a c t e d w i t h t w o d i s t i n c t p r o t e i n b a n d s w h i c h h a d a p p a r e n t m a s s e s o f 1 4 0 a n d 160 k D a u n d e r r e d u c i n g c o n d i t i o n s ( F i g . 4 .8) i n the a p p r o x i m a t e p o s i t i o n s o f the p r o c o l l a g e n c h a i n s o r s e v e r a l a o r fi i n t e g r i n s u b u n i t s . E n z y m e d i g e s t i o n s w e r e p e r f o r m e d to c h a r a c t e r i z e these b i n d i n g p r o t e i n s . B o t h b a n d s w e r e d e g r a d e d b y b a c t e r i a l c o l l a g e n a s e . T h e 140- a n d 1 6 0 - k D a b a n d s w e r e a l so p a r t i a l l y p e p s i n s e n s i t i v e b e i n g d e g r a d e d to p e p s i n res is tant , b u t c o l l a g e n a s e s e n s i t i v e , 112- a n d 1 2 6 - k D a p r o t e i n s that c o - m i g r a t e d w i t h c o l l a g e n a 1(1) a n d a 2 ( I ) c h a i n s w h i c h a l s o w e r e b o u n d b y r C B D 1 2 3 . L a s t l y , a r C B D 1 2 3 m u t a n t ( K 2 6 3 A ) w h i c h w e h a v e c h a r a c t e r i z e d as h a v i n g a r e d u c e d t y p e I c o l l a g e n b i n d i n g a f f i n i t y (Chapter 3) w a s f o u n d to h a v e a h i g h e r h a l f m a x i m a l c e l l a t t a c h m e n t c o n c e n t r a t i o n (3.5 u g / m l ) c o m p a r e d to the w i l d t y p e r C B D 1 2 3 (1 .7 u g / m l ) ( F i g . 4 .9 ) . C o l l e c t i v e l y these resul t s s t r o n g l y i n d i c a t e d that r C B D 1 2 3 b o u n d p e r i c e l l u l a r n a t i v e t y p e I c o l l a g e n a n d p r o c o l l a g e n a t t a c h e d to 6 , i n t e g r i n s i n the c e l l m e m b r a n e s . Cell Attachment to rCBD123 is not Heparan-Sulphate Dependent—In o u r p r e v i o u s c h a r a c t e r i z a t i o n o f the r C B D 1 2 3 b i n d i n g p r o p e r t i e s to e x t r a c e l l u l a r m a t r i x p r o t e i n s a n d c o m p o n e n t s (Chapter 2, 3, S t e f f e n s e n et al., 1995) w e f o u n d that r C B D 1 2 3 c o n t a i n s a l o w a f f i n i t y h e p a r i n b i n d i n g site i n a d d i t i o n to the c o l l a g e n b i n d i n g sites . T h e r e f o r e , to d e t e r m i n e the p o t e n t i a l f o r r C B D 1 2 3 to in te rac t w i t h c e l l m e m b r a n e a s s o c i a t e d h e p a r a n - s u l p h a t e p r o t e o g l y c a n s , c e l l s w e r e t reated w i t h h e p a r i n a s e p r i o r to p l a t i n g . H o w e v e r , e v e n at c o n c e n t r a t i o n s u p to 0.1 U /ml , h e p a r i n a s e t r e a t m e n t d i d n o t a l ter the c e l l a t t a c h m e n t to r C B D 1 2 3 c o m p a r e d to u n t r e a t e d c o n t r o l c e l l s ( F i g . 4 . 1 0 A ) . M o r e o v e r , h e p a r i n a d d e d to 1 3 4 Gelatinase A cell binding via the CBD rCBD123 coated wells to block rCBD123 heparin binding sites prior to cell seeding also did not reduce cell attachment from control levels (Fig. 4.10B). Finally, K357A, another mutant of rCBD123 that we have also generated which shows complete loss of binding to heparin-Sepharose (Chapter 3) did not display any differences in cell attachment compared to wildtype rCBD123 (Fig. 4.9). Therefore, pericellular heparan sulfate proteoglycans are unlikely to contribute to the cell interaction with rCBD123. The Collagen Binding Domain Mediates Binding of Gelatinase A to Cells—To investigate whether the CBD of gelatinase A can bind the natural secreted enzyme to cell surfaces, a competition experiment was performed using human fibroblasts cultured on plastic. After extensive washes with PBS to remove secreted unbound gelatinase A in the medium and cell layer, the fibroblasts were treated with 250, 25, or 0 pg/ml rCBD123 for 5 min in serum-free culture medium. Zymogram analysis demonstrated that the PBS washes efficiently removed all unbound gelatinase and, on subsequent incubation with rCBD123, gelatinase A was competitively released from the confluent cell layer in a concentration dependent manner (Fig. 4.11 A). The released cell-bound gelatinase A migrated under non-reducing conditions with an apparent molecular mass of 66 kDa (-DTT) corresponding to the latent form of the enzyme. No enzyme was detected in the control samples (+CBD, 0) that were not treated with rCBD123. Subsequent extraction of the cell layer with 4 M urea and 1% (w/v) SDS revealed that a large proportion of gelatinase A remained in the cell layer, either cell bound or in the secretory pathway. The rCBD123, either added to the culture medium or coated on the plastic wells, did not alter the Con A induced activation of progelatinase A (Fig. 4.1 IB). Overall, these experiments directly demonstrated that the collagen binding domain provides a mode of binding gelatinase A to cell surfaces in addition to other mechanisms of interaction involving the C-domain of gelatinase A. DISCUSSION Identification of cell surface proteinases and their mechanisms of membrane 135 Gelatinase A cell binding via the CBD attachment is currently of great interest. The recent cloning of membrane anchored MT-MMPs (Puente etal., 1996; Takino et al, 1995; Will and Hinzmann, 1995; Sato et al, 1994), and other cell surface metalloproteinases such as meprins (Johnson and Hersh, 1992; Butler et al, 1987), and members of the ADAM (A Disintegrin And Metalloprotease) family (Wolfsberg et al, 1995; Wolfsberg et al, 1993) points to a pericellular environment rich in proteinases and with a significant proteolytic potential. In addition, soluble secreted MMPs also have the potential to accumulate on the cell surface in cis, on the same cell, or in trans, on bystander cells. Indeed, cell surface localization of gelatinase A, as observed on tumor cell islands (Pyke et al, 1993; Tryggvason et al, 1993; Poulsom et al, 1992), in combination with localized enzyme activation places this important proteolytic activity under precise cellular control. Therefore, identification of the enzymes and cell attachment mechanisms involved in these processes is of fundamental importance in understanding the roles played by MMPs in both physiological tissue processes and pathological conditions. By studying fibroblast attachment to the recombinant CBD of human gelatinase A we have developed a new approach to explore alternative cell binding mechanisms of gelatinase A. Our results show that the CBD mediates cell surface localization of gelatinase A in unstimulated cells by binding pericellular collagen that in turn is anchored to cell membrane 61 integrins. In support of this model, detergent solubilized fibroblast cell surface components that were bound by rCBD123 on ligand blots were fully degraded by highly purified bacterial collagenase. Further, these proteins were partially pepsin sensitive, pointing to native collagen or procollagen mediating the cell binding interaction to rCBD123. Treatment of fibroblasts with bacterial collagenase also abolished the cell attachment to rCBD123 which was rescued in a concentration dependent manner by addition of native type I collagen. In addition, adding type I collagen to coated rCBD123 preceding cell seeding also produced a dose dependent increase in cell attachment ultimately reaching that of cells attached to collagen alone coated on plastic. This likely reflects an increased availability of collagen for cell surface receptors. Moreover, fibroblasts showed less attachment to a 136 Gelatinase A cell binding via the CBD rCBD123 mutant (K263A) which has reduced type I collagen binding properties (Chapter 3). Importantiy, the biological potential for gelatinase A interaction with cells via the CBD was shown by the competitive release of progelatinase A from fibroblasts by rCBD123. Therefore, the interaction between rCBD123 and cell surface components observed in the cell attachment assays is representative of gelatinase A binding the fibroblast cell surfaces and, likely, represents an in vivo mode of enzyme localization. To our knowledge, this is the first demonstration that a domain other than the C domain of gelatinase A has been shown to be involved in cell surface binding. Recent progress in understanding the interaction between gelatinase A and cell membranes followed the initial in vitro observations of enhanced activation of gelatinase A in cell cultures treated with ConA (Overall and Sodek, 1992; Overall and Sodek, 1990) or TPA (Brown et al, 1990). The activator was subsequently localized to cell membrane preparations (Ward et al., 1991) suggesting that an integral component of the plasma membrane was involved in the endogenous activation of gelatinase A. Activation did not occur with C domain deletion mutants of gelatinase A (Ward et al., 1994; Murphy et al., 1992) or in the absence of the TIMP-2 (Strongin et al., 1995) indicating an important role for the C domain interaction with stimulated cells. This cell membrane component was later identified as MT1-MMP (Sato et al., 1994) and a total of four MT-MMPs (Puente et al., 1996; Takino et al., 1995; Will and Hinzmann, 1995) have now been cloned. One model for the gelatinase A cell surface localization and activation involves TIMP-2 bridging the C domain of progelatinase A and the cell membrane associated MT-MMP (Strongin et al, 1995) with possible activation occurring by a second MT-MMP (Overall et al., 1997). Alternatively, since TIMP-2 as well as the progelatinase A/TIMP-2 complex can bind directly to cell surfaces, it has been proposed that a specific TIMP-2 receptor can mediate enzyme binding (Emmert-Buck et al, 1995). Although a TIMP-2 receptor has not yet been cloned, signal transduction events can be elicited by TIMP-2 binding (Corcoran and Stetler-Stevenson, 1995). Direct binding of gelatinase A via the C domain to a different cell 137 Gelatinase A cell binding via the CBD receptor, the integrin ot^, was also demonstrated recently (Brooks et al., 1996). Interestingly, the 0^ 63 integrin is a major receptor for the extracellular matrix protein vitronectin (Suzuki et al., 1986; Pytela et al, 1985), to which the gelatinase A C domain is homologous. Occupation of this cell surface receptor by vitronectin or by blocking antibodies enhances gelatinase A expression and cellular penetration of basement matrices in vitro (Seftor et al, 1993; Seftor et al, 1992). Although cellular expression of several MMPs is modified by incubation of cells with integrin ligands such as the fibronectin CS 1 and heparin-binding peptides or with blocking antibodies reacting specifically with the Bl5 oc3, a4fi1, a5iS: integrins (Huhtala et al, 1995; Larjava et al, 1993; Werb et al, 1989), it is not known whether binding of the C domain to the 0^ 63 integrin modifies gelatinase A expression. Nonetheless, the direct interaction of gelatinase A with this receptor (Brooks et al, 1996) has emphasized the important potential role of integrins in cell surface proteolytic events. The gelatinase A bound via the CBD to fibroblast cell layers of unstimulated cells was primarily of the 66-kDa (-DTT) latent form. Thus, the CBD mediated enzyme binding may provide a means of maintaining a pool of progelatinase A in the vicinity of the cell membrane poised for subsequent C domain interaction on cell stimulation and activation by MT-MMP in the C domain - TIMP-2 pathway. Although the binding of gelatinase A to pericellular collagen at the cell surface may co-localize the enzyme in proximity to MT-MMPs this was not sufficient for activation since rCBD123 added to Con A activated cells did not block activation of progelatinase A. This concurs with the observation, that both wildtype and a CBD deletion mutant of gelatinase A were activated by Con A treated fibroblasts suggesting that the CBD is not directly required for membrane-mediated activation (Murphy et al, 1994). In comparison, addition of recombinant gelatinase A C domain blocked ConA induced activation of progelatinase A in a concentration dependent manner (Overall et al, 1997). Contact interactions between integrin receptors and extracellular matrix components are also important for defining cellular behavior. The initial step in cellular adhesion is 138 Gelatinase A cell binding via the CBD substrate recognition via specific integrin receptors (Hynes, 1987). This early event typically is followed by cell spreading during which the cytoskeletal actin filaments and ligand bound integrins are reorganized with generation of focal contact points (Albelda and Buck, 1990). The linking of integrin to the extracellular matrix during initial attachment and spreading sets the stage for transmembrane signaling and control of gene expression (Clark and Brugge, 1995; Yamada and Miyamoto, 1995) required to define cell function in such processes as growth, differentiation, and remodeling (Roskelley et al, 1995; Hynes, 1992). Thus, it was of interest that fibroblasts seeded on rCBD123 spread at a slower rate compared to cells on fibronectin, displayed less spreading, and showed a distinctly different morphological appearance. As has been observed for short synthetic peptides of cell attachment proteins (Massia and Hubbell, 1991), reduced or delayed spreading may reflect a limited accessibility of binding sites. Alternatively, cell attachment to rCBD123 mediated through pericellular collagen may explain the differences in cell attachment and spreading compared to cells binding directly to fibronectin primarily via the (X5R1 integrin (Hynes, 1992; Albelda and Buck, 1990). In addition, the property of rCBD123 to simultaneously bind two molecules type I collagen (Chapter 2, Steffensen et al., 1995) provides a potential for cells to form attachments between cell bound collagen and extracellular matrix collagen via the CBD of gelatinase A. The high efficiency inhibition of cell attachment to rCBD123 by blocking the B^ integrin receptors with specific antibody points to the formation of a cell membrane-associated complex consisting of gelatinase A bound by the CBD to f^  integrins, most likely via a collagen bridge. Although direct binding of rCBD123 to Bj integrins is a possibility not entirely ruled out by our studies no RGD sequence is present in the gelatinase A CBD. In contrast, the gelatinase B CBD contains a RGD sequence which aligns with a RSDG sequence at position 339-341 in gelatinase A (Wilhelm et al., 1989). Since our results also show a specific interaction of rCBD123 with type I collagen it is likely that these observations together reflect a model in which the gelatinase A CBD interacts with 139 Gelatinase A cell binding via the CBD pericellular collagen bound to 61-integrins on the cell surface among which oclBl, a2Bl, and a3J31 have been ascribed a major role in binding type I collagen (Hemler, 1988). The rCBD123 interaction site with native type I collagen does not sterically block integrin attachment to the collagen indicating that the integrin and rCBD123 have different collagen binding sites. Based on these results, we propose that the collagen binding domain of gelatinase A mediates positioning of gelatinase A to cell surfaces of unstimulated cells through formation of a gelatinase A/Type I collagen/integrin receptor complex. SUMMARY The mechanism of cell surface localization of gelatinase A, a process important in enzymic activation, was investigated. Here we demonstrate that human fibroblasts attach specifically to the CBD of human gelatinase A comprising the three fibronectin type II-like modules. The mode of attachment was distinct from that to fibronectin as evidenced by reduced cell spreading and by scanning electron microscopy which revealed distinct differences in cell morphology. Incubation of fibroblasts with blocking antibodies to the 6 r integrin subunit, but not the o^ -integrin subunit, completely inhibited cell attachment to rCBD123 indicating that integrins contributed to the binding interaction. Cell attachment to rCBD123 but not fibronectin was reduced in a concentration dependent manner by bacterial collagenase treatment, with the attachment being partially recovered by addition of soluble native type I collagen. In addition, rCBD123, used as a probe on ligand blots of octyl-B-D-thioglucopyranoside-solubilized cell extracts, bound distinct protein bands of Mr 140 and 160 kDa potentially corresponding to either integrin subunits or procollagen chains. These proteins were bacterial collagenase sensitive and pepsin resistant, being converted to 112- and 126-kDa proteins that co-migrated with collagen ccl(I) and a2(I) chains indicating their identity as procollagen chains. Moreover, a site specific mutant of rCBD123 (K263A) with reduced collagen affinity was characterized by reduced cell binding, whereas a heparin-binding knockout mutant of rCBD123 (K357A) or treatment of cells with heparinase did not 140 Gelatinase A cell binding via the CBD alter binding. Although rCBD123 competed progelatinase A from unstimulated fibroblast cell layers in culture and thereby showed the biological significance of the interaction, rCBD123 added to cells did not alter Con A induced cell membrane activation of progelatinase A. Thus, these results demonstrate for the first time a cell attachment mechanism for gelatinase A that does not involve the C domain. Gelatinase A can be positioned on cell surfaces by a mechanism which utilizes the CBD and involves the formation of a gelatinase A/native type I collagen/Brintegrin attachment complex. 141 0 600 120018002400 0 20 40 60 80 100120 Protein Concentration (nM) FIG. 4.1. Attachment of human fibroblasts to rCBD123 and fibronectin. rCBD123 and fibronectin (50 to 0.25, and 0 ug/ml) were coated in PBS to 96-microwell plates for 18 h at 4 °C. After blocking with 10 mg/ml BSA, 4 x 104 fibroblasts were seeded per well in serum-free a-MEM and incubated for 90 min at 37 °C. Wells were rinsed with PBS and attached fibroblasts were fixed with 4% (w/v) formaldehyde and stained with crystal violet. Stain from attached cells was dissolved with 10% (v/v) acetic acid and cell attachment quantitated by the absorbance at 590 nm. Fibroblasts attached to rCBD123 and fibronectin in a saturable and concentration dependent manner. Data points are means of duplicate measure-ments and representative of three separate experiments. 142 I 20-rr 0-1 , 1 , 1 1 0 10 20 30 40 50 60 Soluble rCBD123 (ug/ml) FIG. 4.2. Competition of cell attachment to rCBD123-coated surfaces by soluble rCBD123. Human fibroblasts were incubated with soluble rCBD123 (50 to 1.5 (J-g/ml) in a-MEM for 30 min at 37 °C. The treated fibroblasts were then transferred to 96-well plates coated with 3.0 (Xg/ml rCBD123 and blocked with 10 mg/ml BSA. After incubation for 90 min, cell attachment was quantitated by the absorbance of dissolved crystal violet stain from attached cells as described under "Experimental Procedures" (Chapter 4). Pre-incubation with rCBD123 produced a concentration-dependent inhibition of cell attachment demonstrating specific binding to rCBD123. Data points are representative of two separate experiments. 143 0 30 60 90 120 Time (min) FIG. 4.3. Spreading of human fibroblasts on rCBD123 and fibronectin. Fibroblasts (5 x 103 per well) were seeded in 96-well plates coated with rCBD123 or fibronectin (10 or 25 U.g/ml) and blocked with 10 mg/ml BSA. After 30, 60, or 120 min incubation at 37 °C, cells were fixed with 3% (w/v) formaldehyde and the number of spreading cells in proportion to the number of seeded cells was determined by phase contrast microscopy. Each data point represents the mean of three random fields in each of triplicate wells. Error bars = SD. 144 FIG. 4.4. Morphological differences between cells cultured on rCBD123 and 3 fibronectin. Human fibroblasts (1 x 10 ) were seeded on glass coverslips coated with rCBD123 or fibronectin (25 ug/ml), blocked with BSA, and incubated at 37 °C. At different time points, the fibroblasts were fixed with glutaraldehyde and processed for SEM as described under "Experimental Procedures" (Chapter 4). After 1 and 2 h, cells on fibronectin (A, C) demonstrated extensive cytoplasmic spreading (arrows) accompanied by significant increases in surface area. In contrast, cells on rCBD123 were smaller after 1 h (B) and still displayed only limited spreading with extension of delicate filopodia (arrow heads) after 2 h (D). Bars = 25 mm 145 c CD E .c o co *=: < O 5 10 15 20 Antibody Concentration (Lig/ml) FIG. 4.5. Cell attachment to rCBD123 is inhibited by antibodies to the BI integrin subunit. 96-microwell plates were coated with rCBD123 (25 Llg/ml) and human fibroblasts then seeded in the presence of monoclonal blocking antibodies (20 to 1.25, and 0 Llg/ml) to the 81 integrin subunit (MAbl3) (afil-Integrin), the <x2 integrin subunit (P1E6) (ao2-Integrin) and, as controls, affinity purified polyclonal antibodies to the rCBD123 (aCBD123) and the (His)6-containing peptide in the N-terminal fusion tag of rCBD123 (oi(His)6). Protein concentration of 1/10 - 1/200 dilution of P1E6 monoclonal antibody was estimated from ascites fluid protein concentration. Cell attachment was quantitated after 90 min incubation as described under "Experimental Procedures" (Chapter 4). Antibodies to the 81 integrin subunit strongly inhibited cell attachment to rCBD123 with 50% inhibition occurring at 2 Llg/ml. Data points are means of duplicate measurements and representative of three separate experiments. 146 o < c CD £ o cd CD O 0.8-Type I Collagen Added (per well): [PBj L J I 1-5 25 | Controls Collagen rCBD123 (ug/ml) FIG. 4.6. Cell attachment to rCBD123-type I collagen complex. 96-microwell plates were coated with either rCBD123 (1.5 and 25 Llg/ml) or native type I collagen (1, 3, or 10 Lig/well). After blocking with BSA, additional rCBD123-coated wells were then incubated with native type I collagen (1, 3, or 10 fig/well). Tissue culture treated plastic alone (P) or blocked with BSA (B) served as positive and negative controls, respectively. Human fibroblasts (4 x 104) were seeded and cell attachment analyzed after 90 min as described under "Experimental Procedures" (Chapter 4). Formation of an rCBD123-type I collagen complex enhanced the cell attachment over rCBD123 alone in a dose-dependent manner. Data points are means of duplicate wells from three experiments. 147 FIG. 4.7. Effect of bacterial collagenase treatment on cell attachment to rCBD123. Human fibroblasts were treated with highly purified bacterial collagenase (0.075, 0.75, or 7.5 U/100 in serum-free a-MEM for 15 min at 37 °C. Collagenase treated fibroblasts (4 x 104) were then seeded in wells coated either with rCBD123 (25 |ig/ml) alone or rCBD123 (25 |ig/ml) with bound native type I collagen (1, 3, or 10 u.g/well). Cell attachment was quantitated by crystal violet staining after 90 min incubation as described under "Experimental Procedures" (Chapter 4). The collagenase treatment produced a concentration-dependent inhibition of cell attachment to rCBD123 that was rescued by addition of native type I collagen to the rCBD123 coated wells. Data are means of triplicate wells. 148 200 97- * 67 —•> - Pepsin - Collagenase B11 m fc^p12 t- port * - 3(X2 * ^ o 2 43-> 2 9 - * CB Control St FIG. 4.8. Ligand blotting of rCBD123 to detergent solubilized cellular components. Confluent cultures of human fibroblasts were rinsed with PBS and lysed with 50 mM 6-D-thioglucopyranoside for 30 min at 15 °C. The detergent-solubilized proteins were digested with pepsin or bacterial collagenase as indicated before separation by SDS-PAGE using 7.5% (w/v) minislab gels. After transfer to Immobilon PVDF membranes the blots were probed with 20 ug/ml rCBD123 for 1 h at 22 °C. Control membranes were processed identically but not incubated with rCBD123 (Control). Protein bound rCBD123 was detected with affinity-purified polyclonal antibody to rCBD123 (aCBD123) and visualized by enhanced chemiluminescence reagents. rCBD123 bound to distinct protein bands that were partially resistant to pepsin digestion, but degraded by bacterial collagenase. After pepsin treatment, the binding proteins co-migrated with the type I collagen standard (St). The kDa of marker protein standards are indicated. Untreated detergent solubilized protein separated on 7.5% (w/v) SDS-PAGE gel and stained with Coomassie Brilliant Blue (CB) 149 Protein Concentration (|ig/ml) FIG. 4.9. Cell attachment to rCBD123 mutants with reduced binding to type I collagen (K263A) or heparin (K357A). Wildtype rCBD 123 and two mutant rCBD 123 proteins characterized by reduced binding to collagen, K263A, or no binding to heparin, K357A, were coated (25 - 0.6, and 0 iig/ml) in 96-microwell plates. Human fibroblasts (4 x 104) were plated and cell attachment was quantitated after 90 min as described under "Experimental Procedures" (Chapter 4). While no differences were observed at saturating conditions, the half-maximal cell attachment concentration for K263A was 3.5 ug/ml compared to 1.7 ug/ml for both wildtype rCBD123 and K357A. The data points are means of duplicate wells and representative of four experiments. Bars = range. 150 o LO < C CD E .c o CO < O • -s • Control • 0.01 U/ml Heparinase A 0.1 U/ml Heparinase T T o in < C CD E .c o CO < O • Control • 1 |ig/well Heparin A 10 [ig/we\\ Heparin i i i i i 0 5 10 15 20 25 30 rCBD123 Concentration (u.g/ml) FIG. 4.10. The role of rCBD123 heparin binding sites in cell attachment. To assess the role of pericellular heparan on cell attachment to rCBD123, human fibroblasts were treated with heparinase (0.1 or 0.01 U/ml) in serum free a-MEM containing 10 mM sodium acetate and 1 mg/ml BSA for 30 min at 37 °C. Control cells were incubated under identical conditions but without heparinase. Treated fibroblasts were then seeded in wells coated with rCBD123 (25 to 0.25, and 0 iig/ml) and cell attachment quantitated as described under "Experimental Procedures" (Chapter 4). No differences in cell attachment were observed between heparinase-treated cells and untreated control cells (Panel A). Then, to occupy heparin binding sites, rCBD123-coated wells were incubated with heparin (1 or 10 ug/well) for 1 h at 22 °C. No heparin was added to control wells. After thorough PBS rinses, human fibroblasts were plated and cell attachment quantitated. No inhibition of cell attachment resulted from occupying rCBD123 heparin binding sites with exogenous heparin (Panel B). Data points are means of triplicate wells and representative of two identical experiments. 151 FIG. 4.11. rCBD123 competitively dissociates progelatinase A from human fibroblast cells but does not alter Con A induced activation. Human fibroblast cell layers on plastic surfaces were rinsed thoroughly with PBS (W; 1, 2, 3) representing sequential rinses to remove all unbound secreted enzyme that had accumulated in the conditioned culture medium (M). Cells were then incubated with rCBD123 in PBS for 5 min at 22 °C (Panel A). Incubation with 250 or 25 ug/ml rCBD123 (+rCBD; 250, 25) competitively released gelatinase A which electrophoresed with an apparent molecular mass of 66 kDa (-DTT) corresponding to the latent form of the enzyme. No gelatinase A was released in the absence of rCBD123 (+rCBD; 0). A significant proportion of gelatinase A remained in the cell layer or in the intracellular secretory pathway and was extracted from the corresponding wells (C; 250, 25, 0) with 4 M urea and 1% (w/v) SDS. In a second experiment (Panel B), human fibroblasts were seeded in uncoated plastic wells or in wells coated with rCBD123 (25 mg/ml) (rCBD123 coated, +). After reaching confluence, cells were rinsed with PBS and then incubated with serum free a-MEM with or without Con A (20 ug/ml) (Con A; ±) and 100 or 10 ug/ml rCBD123 (rCBD123 soluble, +; 100, 10) for 18 h. Samples were analyzed by enzymography using 10% SDS-PAGE gels copolymerized with 40 ug/ml heat-denatured type I collagen as described under "Experimental Procedures." The positions of rCBD123 and the kDa of the latent and active forms of gelatinase A are indicated. 152 CHAPTER 5 CONCLUDING DISCUSSION The MMPs are important in the physiological processes of embryogenesis and tissue remodeling, in pathological conditions such as the inflammatory diseases arthritis and periodontal disease, and during invasive tissue penetration by tumor cells. Although the different MMPs show considerable overlap in substrate specificity, there are distinct differences between several members of the family. An important determinant in the substrate specificity is the recognition and binding of the MMPs via distinct domains to the extracellular matrix components. Among the members of the family of MMPs only the gelatinases contain a domain composed of three fibronectin type II-like modules that are inserted into the catalytic domain. The study of this domain has been the focus of this thesis. Fibronectin, a multidomain molecule, contains two type EI modules which are localized in a region of the protein that is associated with gelatin binding (Litvinovich et al., 1991; Owens and Baralle, 1986; Balian et al, 1979). Fibronectin type II-like modules have been found in several other proteins, and provide gelatin binding properties in PDC-109 (Banyai et al., 1990) and human gelatinase B (Collier et al, 1992). Based on these observations, the hypothesis was formulated that the fibronectin type II-like modules serve an analogous function in human gelatinase A which is a potent gelatinase. The interactions between enzymes and ligands may be studied by using isolated enzyme domains or domain deletion mutants of the enzyme. Although studies of isolated recombinant modules or domains do not directly permit conclusions to be drawn regarding the function of the full-length enzyme, such studies have the advantage that one can assign binding sites to specific modules or domains. This experimental approach was selected for these studies. Recombinant protein expressed in E. coli and containing all three fibronectin type II-like modules of human gelatinase (rCBD123) bound specifically to gelatin (Chapter 2, 153 Concluding discussion Steffensen et al., 1995). This finding is in agreement with results of research published during the course of this research by others (Banyai et al., 1994; Banyai and Patthy, 1991). In addition, rCBD123 bound native type I collagen which has subsequently been recognized as a substrate of the gelatinase A (Aimes and Quigley, 1995; Sodek and Overall, 1992). In fact, our extended characterization of the rCBD123 ligand binding properties demonstrated that this domain was responsible for most if not all of the type I collagen binding properties of gelatinase A, and also of denatured types IV and V collagens, elastin, and heparin, but not fibronectin, laminin, SPARC, or TIMP-1 (Chapter 2, Steffensen et al., 1995). The binding to the broad range of collagens, including several in their native form, prompted the term "collagen binding domain" (CBD) and subsequent studies including the tri-modular rCBD123 have demonstrated that the domain is also important for enzyme binding to the type X collagen (Abbey et al., 1997; Overall et al., 1997). However, rCBD123 did not bind native type IV collagen in spite of this protein being a substrate for gelatinase A. Since the rC domain of gelatinase A also does not bind native type IV collagen (Overall et al., 1997; Wallon and Overall, 1997), the enzyme binding of this protein likely resides in the catalytic domain of the enzyme. The importance of the CBD for function of the enzyme has been demonstrated clearly in studies of CBD-deletion mutants of gelatinase A that showed an over 90 per cent reduction in gelatin degradation (Murphy et al., 1994) and complete loss of elastinolytic activity (Shipley et al., 1996). Interestingly, the collagen binding properties which reside in the CBD of the gelatinases are mediated in interstitial collagenase by the COOH-terminal domain without which collagenase cannot cleave native type I collagen (Bigg et al., 1994; Schnierer et al., 1993; Windsor et al., 1991; Clark and Cawston, 1989; Hasty etal., 1987). Previously, it had been demonstrated that a major fibronectin binding site on collagen is located in the cyanogen bromide fragment 7 (CB7) of the al(I)-chain near the mammalian collagenase cleavage site (Kleinman et al., 1978), with a second site in the al(I)-chain CB12 near the NH2-terminal part of the molecule (Guidry et al., 1990). In comparison, we found 154 Concluding discussion binding of rCBD123 to three tested fragments, CB2, CB7 and CB8, of the a 1(1) chains of denatured type I collagen, and the binding site analyses indicated an even larger number of potential binding sites. Further, in native type I collagen the major binding site for rCBD123 was located in the non-helical telopeptide segments (Chapter 2, Steffensen et al., 1995). Although the type II modules in gelatinase A, like those in fibronectin, provided the enzyme with collagen binding properties, the differences in localization and number of collagen binding sites point to domain specialization during evolution of the type II modules in fibronectin and gelatinase A in spite of their likely common ancestral origin (Patthy, 1996; Patthy, 1991; Patthy et al., 1984). The observation that rCBD123 bound specifically to heparin was of particular interest because heparin is structurally very similar to heparan sulfate. Thus, the CBD interaction with heparin might reflect similar interactions with heparan sulfate side chains of secreted proteoglycans such as perlecan found in the basement membrane (Noonan et al., 1988) and in the stroma of tumors (Iozzo et al., 1994). This is important because gelatinase A has been localized at high levels on tumor cells during penetration of basement membrane and invasive growth (Tryggvason et al, 1993; Liotta et al, 1980). Also, perlecan binds basic fibroblast growth factor which stimulates angiogenic cell proliferation (Folkman and Shing, 1992). Perlecan has not yet been shown to be a substrate for gelatinase A although several other proteoglycans are known to be degraded by MMPs. However, the potential for release of basic fibroblast growth factor by proteolysis of perlecan by tumor cell-associated gelatinase A during basement membrane penetration might induce local neovascularization which is a requirement for tumor expansion and invasion (Folkman and Shing, 1992). Although heparin binds a larger number of proteins including the type III-13 module of fibronectin (Busby et al, 1995) it was not known at the outset of these studies that this glycosaminoglycan could bind fibronectin type II modules. The consensus binding sequences (XBBXBX and XBBBXXBX) for a number of heparin binding proteins is rich in basic residues (B) that are separated with a characteristic spacing pattern by amino acid residues 155 Concluding discussion (X) that primarily are hydrophobic in nature (Cardin and Weintraub, 1989; Cardin et al., 1986). Sequence analyses revealed that such consensus sequences did not occur in the gelatinase A CBD. However, it has also been shown that basic residues distantly located in the primary sequence of heparin binding proteins may form a positively charged heparin binding "cradle" through protein folding (Busby et al., 1995; Mann et al., 1994). Therefore, because little is known about the molecular determinants for the binding of gelatinase A to heparin, the studies presented in Chapter 3 focused on the contributions of rCBD123 lysines to heparin binding. Acetylation of lysines fully abolished the rCBD123 binding to heparin and reduced binding to gelatin pointing to the functional importance of these basic residues in CBD heparin and type I collagen binding. To determine the contributions of individual lysines in the ligand binding interactions, site-specific substitution to alanines demonstrated that one lysine residue in position 357 was crucial for the rCBD123 binding to heparin. The specificity of the altered function was verified by the other mutations studied [K263A, K330A, K343A, K298/299A] which had unaltered affinity to heparin and the absence of structural perturbations for K357A relative to wildtype rCBD123, as determined by circular dichroism spectral analyses. Structure modeling of CBD module 3 based on coordinates from the PDC109b type II-like modules (Constantine et al., 1992; Constantine et al., 1991) predicted that K357 is located next to R356 in the loop connecting B-strands 3 and 4. However, these residues are not predicted to be part of a positively charged patch and three additional basic residues on different aspects of the third module of CBD are separated from K357 by distances of approximately 20 A. Therefore, it is unlikely that one single module is responsible for heparin binding, whereas surface exposed basic residues from two or all three modules may form the heparin binding site in the tertiary structure of the CBD. The principle of modular cooperativity for gelatinase A CBD has also been supported by Banyai et al. (1996) who proposed a model in which the three type II-like modules of gelatinase A together form a 156 Concluding discussion collagen binding hydrophobic groove lined by aromatic residues. A different lysine to alanine substitution in position 263 of CBD module 2, K263A, reduced saturation level binding to type I collagen (Chapter 3). This result concurs with earlier experiments which showed that lysine residues of the type II modules of fibronectin are important for maintaining intact collagen binding (Isaacs et al., 1989). In addition, the structural prediction places K263 in proximity to the positions of two alanines found in the fibronectin type II-like model 1 of gelatinase B. Module 1 in gelatinase B does not bind gelatin; however, substitution of these AA residues in module 1 to KR residues, as are found in the module 2 of gelatinase B, recovered the binding to gelatin (Collier et al., 1992). Therefore, the presence of basic residues in the segment NH2-terminal to B-strand 1, in which K263 is also located, appears to be required for collagen binding. The reduced binding to collagen for K263A occurred without a concurrent change in the Kd for the interaction compared to wildtype rCBD123. This may reflect the presence of more than one collagen binding site on the CBD which may have compensated for disrupted binding in the binding site involving K263. This explanation is also supported by our observation that rCBD123 can simultaneously bind at least two molecules of native or denatured type I collagen (Chapter 2, Steffensen et al. 1995). Evidence for module cooperativity is supported by results from studies of single-, bi-, and tri-modular gelatinase A CBD constructs (Abbey et al., 1997; Overall et al., 1997; Banyai et al., 1994) which showed that rCBD123 has an affinity for type I collagen that is about one order of magnitude stronger than the bi-modular constructs rCBD12 and rCBD23. Of note, these module interactions are different for the gelatinase B CBD where module 2 alone can bind gelatin even stronger than any combination of the modules including the tri-modular CBD (Collier et al., 1992). Together, these ligand binding experiments with site-specific introduction of alanines in place of lysines demonstrated that the gelatinase A CBD lysines contribute to CBD ligand binding of heparin and type I collagen. In addition, the results support the notion that heparin and collagen binding by the CBD occurs through cooperativity of two or more of the three 157 Concluding discussion modules of the domain. Considerable interest is currently devoted to studying the mechanisms of cell surface positioning of gelatinase A. Several lines of evidence point to very precise and localized proteolytic actions of gelatinase A in the proximity of cell surfaces. First, cell membranes of human cancer cells contain high levels of proteolytic activities (Emonard et al., 1992; Zucker et al., 1990). Second, progelatinase A can be activated on cell surfaces by an interaction with Con A stimulated membrane-type metalloproteinases (Strongin et al., 1995; Ward et al., 1994; Overall and Sodek, 1990) in a complex containing progelatinase A, TTMP-2, and MT1-MMP (Strongin et al., 1995). However, the recent finding that progelatinase A may also be positioned by direct binding to the integrin ctv63 (Brooks et al., 1996) has shown that alternative modes of cell surface localization may be important in addition to that mediated by the MT-MMPs. To investigate cell interaction with human gelatinase A, we applied a cell attachment assay that allowed modifications of the conditions to analyze factors involved in the binding between cells and the enzyme (Chapter 4). Coated rCBD123 supported attachment by human gingival fibroblasts, and the specificity of the interaction was verified by the ability of soluble rCBD123 to block the attachment when incubated with the cells prior to plating. Several experiments indicated that cell attachment to rCBD123 occurred via pericellular collagen. In ligand blotting experiments, rCBD123, used as a probe, was found to bind to specific protein bands in detergent solubilized cell components. These bands displayed an electrophoretic behavior similar to that of type I procollagen chains and were pepsin resistant but sensitive to digestion with bacterial collagenase. The loss of binding to rCBD123 following treatment of the cells with bacterial collagenase, and recovery of binding by addition of exogenous native type I collagen, confirmed the involvement of pericellular collagen in the interaction. Moreover, fibroblasts showed less attachment to a rCBD123 mutant, K263A, in which a single lysine to alanine mutation had reduced the collagen binding properties (Chapter 3). Together, these experiments demonstrated cell surface localization 158 Concluding discussion of gelatinase A mediated by the CBD binding to pericellular collagen. Importantly, the biological potential for gelatinase A interaction with cells via the CBD was shown by competitive release of progelatinase A from fibroblasts by rCBD123. Therefore, the interactions between rCBD123 and cell surface components observed in the cell attachment assays are representative of gelatinase A binding the fibroblast cell surfaces and, likely, represent an in vivo mode of enzyme localization. The gelatinase A bound via the CBD was primarily of the 66 kDa latent form. Since the CBD-mediated enzyme binding could provide a means of maintaining a pool of progelatinase A in the vicinity of the cell membrane poised for activation by the MT-MMPs, we investigated the potential for an alternative mechanism of progelatinase A activation by MT-MMPs. However, Con A stimulated fibroblasts seeded on rCBD123 coated surfaces or incubated with soluble rCBD123 did not display any change in the progelatinase A activation pattern (Chapter 4) suggesting that the CBD is not required for activation of progelatinase A by MT-MMPs. This concurs with the observation, that both wildtype and a CBD deletion mutant of gelatinase A were activated by Con A treated fibroblasts suggesting that the CBD is not directly required for membrane-mediated activation (Murphy et al, 1994). In comparison, incubation of cells with recombinant gelatinase A C domain blocked Con A induced activation of progelatinase A in a concentration dependent manner (Overall et al, 1997) most likely by competing binding of the latent enzyme to the TIMP-2 of the activation complex (progelatinase A/TIMP-2/MT-MMP) (Strongin et al, 1995; Ward et al, 1994). Although this showed that the CBD is not needed for enzyme activation, the gelatinase A binding to pericellular collagen is also interesting in the light of the finding that the CBD has the capacity to bind at least two molecules of type I collagen simultaneously (Chapter 2, Steffensen et al. 1995). Thus, gelatinase A at the cell surface could be involved in localized proteolysis of one molecule of collagen while bound via the CBD to a second collagen molecule. Type I collagen is anchored to cell membranes primarily via the 61-group of integrins among which oclGl, oc2Bl, and a3Gl have been ascribed a major role in binding type I 159 Concluding discussion collagen (Albelda and Buck, 1990). The high efficiency inhibition of cell attachment to rCBD123 by blocking the 61-integrin receptors with specific function-perturbing antibody points to the formation of a cell membrane-associated complex consisting of the gelatinase A CBD bound to 81 -integrins. Direct binding of gelatinase A via the C domain to a different integrin, the av83, was demonstrated recently (Brooks et al., 1996). This receptor is also a major receptor for the extracellular matrix protein vitronectin (Suzuki et al, 1986; Pytela et al, 1985) to which the gelatinase A C domain is homologous. Occupation of the avB3 by vitronectin or blocking antibodies enhances gelatinase A expression and cellular penetration of basement membranes in vitro (Seftor et al, 1993; Seftor et al, 1992). In addition, the expression of several MMPs is modified by incubation of cells with integrin ligands such as the fibronectin CS1 and heparin-binding peptides or with blocking antibodies that react with the 81, a3, a461, or a581 integrins (Huhtala et al, 1995; Larjava et al, 1993; Werb et al, 1989). This emphasizes the important role of integrins in MMP regulatory and surface proteolytic events. Since our results demonstrated a specific interaction of rCBD123 with type I collagen (Chapters 2 and 4), it is likely that these observations together reflect a model in which the gelatinase A CBD mediates positioning of gelatinase A to cell surfaces through formation of a gelatinase A/type I collagen/integrin receptor complex. The results of the research described in this thesis highlight several major aspects of current concepts of the roles of matrix metalloproteinases in the maintenance and remodeling of the extracellular matrix. First, it is evident that distinct domains of these MMPs are. important in mediating the binding between the enzymes and their substrates that is required for cleavage to occur. Here, we have provided evidence for a role of the gelatinase A CBD as the major binding domain not only for gelatin but also for several other extracellular matrix components. Indeed, based on our reported observation that the CBD binds elastin, Shipley et al. (1996) showed that deletion of the CBD eliminated elastin binding and degradation by gelatinase A. The ligand binding properties of the CBD of gelatinase B have not yet been characterized to the same spectrum of potential ligands. However, there are 160 Concluding discussion good reasons to presume that this domain in gelatinase B may also bind more than just denatured type I collagen as described so far. In fact, deletion mutants of gelatinase B have shown that the CBD is required for binding of elastin (Shipley et al., 1996), a protein that we determined to be a ligand of recombinant gelatinase A CBD. As reviewed in the Introduction, studies of the MMP C domains have primarily addressed the importance of this domain for collagen cleavage in collagenase, for collagenase binding by stromelysin, and for binding of TIMP in the gelatinases. However, recent characterization of gelatinase A C domain ligand binding properties show that this domain may supplement the CBD by binding fibronectin and also heparin (Overall et al., 1997; Wallon and Overall, 1997). Nonetheless, there is still a need for a more complete characterization of the ligand binding properties of modules and domains of members of the MMP family. The present thesis research has also demonstrated that the gelatinase A CBD as an isolated domain maintains the type I collagen binding properties of the domain when it is an integral part of the full length gelatinase A. This justifies a continuation of experiments based on introduction of specific amino acid mutations into the isolated CBD aiming to define more precisely the ligand binding sites in the CBD. The mutational analyses described here showed that lysine residues are a requirement for the CBD binding of heparin (Chapter 3). Since heparin binding by the gelatinase A CBD, like other proteins that do not contain the heparin binding consensus sequences (Cardin et al, 1986), likely results from cooperation of basic residues located in different modules, studies of these other basic residues may provide a promising avenue for future studies. In comparison, the outcomes of the mutational analyses of CBD binding to collagen are more difficult to interpret. Because individual modules or bi-modular constructs of the CBD of both gelatinase A and B have shown collagen binding properties (Overall et al, 1997; Collier et al, 1992; Banyai and Patthy, 1991), a better understanding of the collagen binding site(s) may be best achieved by studies of such modules or constructs rather than the complete tri-modular CBD. Alternatively, multiple mutations could be introduced to block binding in several modules simultaneously. 161 Concluding discussion A second emerging theme is the expanding evidence for proteolytic events in the vicinity of cell surfaces. The demonstrated positioning of progelatinase A via binding of the CBD to an integrin/collagen cell surface complex (Chapter 4) supplements the presently known mechanisms of gelatinase A C domain binding indirectly to MT-MMPs via TTMP-2 (Strongin et al, 1995) or directly to the integrin ocvB3 (Brooks et al, 1996). Understanding the mechanisms of cell surface localization of the MMPs will be of increasing biological importance. This is exemplified by results from studies of gelatinase A in tumors which showed that MMPs may be expressed by stromal cells, but exert their proteolytic function on the surfaces of epithelial tumor cells (Poulsom et al, 1992). The mechanisms of gelatinase A translocation from one cell type to another is not yet known, but may be crucial to understanding the tumorigenic process and to developing inhibitory strategies. Interestingly, results from studies of transmembrane zone deletion mutants of MT-MMPs indicate that the truncated MT1-MMP has independent catalytic properties towards a range of extracellular matrix proteins and, thus, proteolytic functions besides that of activating progelatinase A (Ohuchi et al, 1996; Pei and Weiss, 1996). That cell surface proteolytic activities are not limited to those of the MMPs is illustrated by the emergence of interest in other cell surface molecules with proteolytic properties. These include the newly described family of ADAMs (a disintegrin and metalloproteinase) that may be involved in tumor necrosis factor processing (Wolfsberg et al, 1995), and the uPA/uPA-receptor interaction which is important to fibrin degradation in wound healing (Romer et al, 1996). In light of these developments, cell surface associated proteolytic events is a promising field for future research. In the described thesis experiments, a cell attachment assay was applied for studies of the mechanisms of binding between the gelatinase A CBD and cell surface components. This approach has been valuable previously for studying cell membrane receptor interactions with adhesion proteins. Since this provides an assay system in which individual potential binding factors may be modified experimentally, it may be valuable in future studies amied 162 Concluding discussion at delineating further the interactions of cell surface components with the gelatinase A CBD or other isolated domains or modules. We have addressed interactions with pericellular type I collagen and heparin, but not other types of collagen or cell surface associcated proteoglycans. Because the integral membrane molecules with proteolytic properties appear to be important in processes of tissue remodeling both in health and wound healing, and in tumor growth, investigations into their expression, regulation, and activation have a high probability of providing results with broad implications to the understanding of localized cell surface associated proteolysis. 163 BIBLIOGRAPHY Abbey RS, Steffensen B, and Overall CM. (1997a). Module cooperation in gelatinase A collagen binding domain, in preparation Abbey RS, Steffensen B, and Overall CM. (1997b). Protein engineering of the 72-kDa gelatinase collagen binding domain. J. Dent. Res. 76, Abstr. #561 Adachi E, Hayashi T. (1985). In vitro formation of fine fibrils with a D-periodic banding pattern from type V collagen. Coll. Relat. Res. 5, 225-232. Aimes RT, Quigley JP. (1995). Matrix metalloproteinase-2 is an interstitial collagenase. Inhibitor-free enzyme catalyzes the cleavage of collagen fibrils and soluble native type I collagen generating the specific 3/4-and 1/4-length fragments. /. Biol. Chem. 270,5872-5876. Albelda SM, Buck CA. (1990). Integrins and other cell adhesion molecules. FASEB J. 4, 2868-2880. Andrade MA, Chacon P, Merelo JJ, and Moran F. (1993). Evaluation of secondary structure of proteins from UV circular dichroism spectra using an unsupervised learning neural network. Protein. Eng. 6, 383-390. Angel P, Imagawa M, Chiu R, Stein B, Imbra RJ, Ramsdorf HJ, Jonat C, Herrlich P, and Karin M. (1987). Phorbol ester-inducible genes contain a common cis element recognized by a TPA-modulated trans-acting factor. Cell 49, 729-739. Anglard P, Melot T, Guerin E, Thomas G, and Basset P. (1995). Structure and promoter characterization of the human stromelysin-3 gene. J. Biol. Chem. 270, 20337-20344. Azzam HS, Thompson EW. (1992). Collagen-induced activation of the M r 72,000 type IV collagenase in normal and malignant human fibroblastoid cells. Cancer Res. 52, 4540-4544. Azzo W, Woessner JF. (1986). Purification and characterization of an acid metalloproteinase from human articular cartilage. J. Biol. Chem. 261, 5434-5441. Balian G, Click EM, Crouch E, Davidson JM, and Bornstein P. (1979). Isolation of a collagen-binding fragment from fibronectin and cold-insoluble globulin. J. Biol. Chem. 254, 1429-1432. Banyai L, Patthy L. (1991). Evidence for the involvement of type II domains in collagen binding by 72 kDa type IV procollagenase. FEBS Lett. 282, 23-25. Banyai L, Tordai H, and Patthy L. (1994). The gelatin-binding site of human 72 kDa type IV collagenase (gelatinase A). Biochem. J. 298, 403-407. Banyai L, Tordai H, and Patthy L. (1996). Structure and domain-domain interactions of the gelatin-binding site of human 72-kilodalton type IV collagenase (gelatinase A, matrix metalloproteinase 2). J. Biol. Chem. 271, 12003-12008. Banyai L, Trexler M, Koncz S, Gyenes M, Sipos G, and Patthy L. (1990). The collagen-binding site of type-II units of bovine seminal fluid protein PDC-109 and fibronectin. Eur. J. Biochem. 193, 801-806. Baragi VM, Fliszar CJ, Conroy MC, Ye QZ, Shipley JM, and Welgus HG. (1994). 164 Bibliography Contribution of the C-terminal domain of metalloproteinases to bindng by tissue inhibitor of metalloproteinases. C-terminal truncated stromelysin and matrilysin exhibit equally compromised binding affinities as compared to full-length stromelysin. J. Biol. Chem. 269, 12692-12697. Barkakoti N, Winkler FK, Williams DH, D'Arcy A, Broadhurst MJ, Brown PA, Johnson WH, and Murray EJ. (1994). Structure of the catalytic domain of human fibroblast collagenase complexed with an inhibitor. Nature Struct. Biol. 1, 106-110. Barkalow FJB, Schwarzbauer JE. (1991). Localization of the major heparin-binding site in fibronectin. J. Biol. Chem. 266, 7812-7818. Bartlett JD, Simmer JP, Xue J, Margolis HC, and Moreno EC. (1996). Molecular cloning and mRNA tissue distribution of a novel matrix metalloproteinase isolated from porcine enamel organ. Gene 183, 123-128. Basbaum CB, Werb Z. (1996). Focalized proteolysis: Spatial and temporal regulation of extracellular matrix degradation at the cell surface. Curr. Opin. Cell Biol. 8, 731-738. Basset P, Bellocq JP, Wolf C, Stoll I, Hutin P, Limacher JM, Podhajcer OL, Chenard MP, Rio MC, and Chambon P. (1990). A novel metalloproteinase gene specifically expressed in stromal cells of breast carcinomas. Nature 348, 699-704. Belaaouaj A, Shipley JM, Kobayashi DK, Zimonjic DB, Popescu N, Silverman GA, and Shapiro SD. (1995). Human macrophage metalloelastase. Genomic organization, chromosomal location, gene linkage, and tissue-specific expression. J. Biol. Chem. 270, 14568-14575. Bentley KL, Klebe RJ, Hurst RE, and Horowitz PM. (1985). Heparin binding is necessary, but not sufficient, for fibronectin aggregation. J. Biol. Chem. 260, 7250-7256. Bergmann U, Tuuttila A, Stetler-Stevenson WG, and Tryggvason K. (1995). Autolytic activation of recombinant human 72 kilodalton type IV collagenase. Biochemistry 34, 2819-2825. Bigg HF, Clark IM, and Cawston TE. (1994). Fragments of human fibroblast collagenase: interaction with metalloproteinase inhibitors and substrates. Biochem. Biophys. Acta 1208, 157-165. Bigg HF, Shi YE, Liu YE, Steffensen B, Overall CM. Specific, high-affinity binding of TPMP-4 to the COOH terminal hemopexin-like domain of human gelatinase A. TIMP-4 binds progelatinase A and the COOH terminal domain in a similar manner to TTMP-2. [In Press] J Biol Chem 1997; Birk DE, Fitch JM, Babiarz JP, Doane KJ, and Linsenmayer TF. (1990). Collagen fibrillogenesis in vitro: Interaction of types I and V collagen regulates fibril diameter. J. Cell Sci. 95, 649-657. Birkedal-Hansen H. (1993). Role of matrix metalloproteinases in human periodontal diseases. J. Periodontol. 64, 474-484. Birkedal-Hansen H, Moore WGI, Bodden MK, Windsor LJ, Birkedal-Hansen B, DeCarlo A, and Engler JA. (1993). Matrix metalloproteinases: A review. Crit. Rev. Oral Biol. Med. 4, 197-250. 165 Bibliography Birkedal-Hansen H, Taylor RE. (1982). Detergent-activation of latent collagenase and resolution of its component molecules. Biochem. Biophys. Res. Commun. 107, 1173-1178. Biswas C. (1988). Heparin and heparan sulfate binding sites on B-16 melanoma cells. J. Cell. Physiol. 136, 147-153. Blood CH, Sasse J, Brodt P, and Zetter BR. (1988). Identification of a tymor cell receptor for VGVAPG, an elastin-derived chemotactic peptide. J. Cell Biol. 107, 1987-1993. Blundell TL. (1994). Metalloproteinase superfamilies and drug design. Nature Struct. Biol. 1, 73-75. Bode W, Gomis-Ruth F-X, Huber R, Zwilling R, and Stoecker W. (1992). Structure of astacin and implications for activation of astacins and zinc-ligation of collagenases. Nature 358, 164-167. Bork P, Downing K, Kieffer B, and Campbell ID. (1996). Structure and distribution of modules in extracellular proteins. Quart. Rev. Biophys. 29, 119-167. Breathnach R, Matrisian LM, Gesnel MC, Staub A, and Leroy P. (1987). Sequences coding for part of oncogene-induced transin are highly conserved in a related rat gene. Nucleic Acids Res. 15, 1139-1150. Brooks PC, Stromblad S, Sanders LC, von Schalscha TL, Aimes RT, Stetler-Stevenson WG, and Quigley JP. (1996). Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrin ocv|33. Cell 85, 683-693. Brown CC, Balian G. (1987). Effect of heparin on synthesis of short chain collagen by chondrocytes and smooth muscle cells. J. Cell Biol. 105, 1007-1012. Brown PD, Kleiner DE, Unsworth EJ, and Stetler-Stevenson WG. (1993). Cellular activation of the 72 kDa type IV procollagenase/TIMP-2 complex. Kidney Int. 43, 163-170. Brown PD, Levy AT, Margulies IMK, Liotta LA, and Stetler-Stevenson WG. (1990). Independent expression and cellular processing of MT 72,000 type IV collagenase and institial collagenase in human tumorigenic cell lines. Cancer Res. 50, 6184-6191. Brunette DM, Melcher AH, and Moe HK. (1976). Culture and origin of epithelium-like and fibroblast-like cells from porcine periodontal ligament explants and cell suspensions. Archs. Oral. Biol. 21, 393-400. Buck CA, Horwitz AF. (1987). Cell surface receptors for extracellular matrix molecules. Ann. Rev. Cell Biol. 3, 179-205. Burgeson RE. (1982). Genetic heterogeneity of collagens. J. Invest. Dermatol. 79, 25s-30s. Burgeson RE, Chiquet M, Deutzmann R, Ekblom P, Engel J, Kleinman H, Martin GR, Meneguzzi G, Paulsson M, and Sanes J. (1994). A new nomenclature for the laminins. Matrix Biology. 14, 209-211. Busby TF, Argraves WS, Brew SA, Pechik I, Gilliland GL, and Ingham KC. (1995). Heparin binding by fibronectin module III-13 involves six discontinuous basic residues brought together to form a cationic cradle. J. Biol. Chem. 270, 18558-18562. 166 Bibliography Butler GS, Will H, Atkinson SJ, and Murphy G. (1997). Membrane-type-2 matrix metalloproteinase can initiate the processing of progelatinase A and is regulated by the tissue inhibitors of metalloproteinases. Eur. J. Biochem. 244, 653-657. Butler PE, McKay MJ, and Bond JS. (1987). Characterization of meprin, a membrane-bound metalloendopeptidase from mouse kidney. Biochem. J. 241, 229-235. Caldwell EEO, Nadkarni VD, Fromm JR, Linhardt RJ, and Weiler JM. (1996). Importance of specific amino acids in protein binding sites for heparin and heparan sulfate. Int. J. Biochem. Cell Biol. Campbell ID, Baron M, Cooke RM, Dudgeon TJ, Fallon A, Harvey TS, and Tappin MJ. (1990). Structure-function relationship in epidermal growth factor (EGF) and transforming growth factor-alpha (TGF-alpha). Biochem. Pharmacol. 40, 35-40. Campbell JH, Terranova VP. (1988). Laminin: molecular organization and biological function. J.Oral Path. 17, 309-323. Cao J, Sato H, Takino T, and Seiki M. (1995). The C-terminal region of membrane type matrix metalloproteinase is a functional transmembrane domain required for pro-gelatinase A activation. J. Biol. Chem. 270, 801-805. Cardin AD, Demeter DA, Weintraub HJR, and Jackson RL. (1991). Molecular design and modeling of protein-heparin interactions. Methods Enzymol. 203, 556-583. Cardin AD, Hirose N, Blankenship DT, Jackson RL, Harmony JAK, Sparrow DA, and Sparrow JT. (1986). Binding of high reactive heparin to human apolipoprotein A: Identification of two heparin-binding domains. Biochem.. Biophys. Res. Commun. 134, 783-789. Cardin AD, Weintraub HJR. (1989). Molecular modeling of protein-glycosaminoglycan interactions. Arteriosclerosis 9, 21-32. Carmichael DF, Sommer A, Thompson RC, Anderson DC, Smith CG, Welgus HG, and Stricklin GP. (1986). Primary structure and cDNA cloning of human fibroblast collagenase inhibitor. Proc. Natl. Acad. Sci. USA 83, 2407-2411. Carter WG, Wayner E. (1988). Characterization of a collagen-binding, phosphorylated, transmembrane glycoprotein expressed in nucleated human cells. J. Biol. Chem. 263, 4193-4201. Charonis AS, Tsilibary EC, Yurchenco PD, Furthmayr H, and Coritz A. (1985). Binding of laminin to type IV collagen: a morphological study. J. Cell Biol. 100, 1848-1853. Chen D, Magnuson VL, Hill S, Arnaud C, Steffensen B, and Klebe RJ. (1992). Regulation of integrin gene expression by substrate adherence. 7. Biol. Chem. 267, 23502-23506. Cheresh DA, Smith JW, Cooper HM, and Quaranta V. (1989). A novel vitronectin receptor integrin (alpha v beta x) is responsible for distinct adhesive properties of carcinoma cells. Cell 57, 59-69. Chin JR, Murphy G, and Werb Z. (1985). Stromelysin, a connective tissue-degrading metalloendoprptidase secreted by stimulated rabbit synovial fibrobalsts in parallel with 167 Bibliography collagenase. /. Biol. Chem. 260, 12367-12376. Clark EA, Brugge JS. (1995). Integrins and signal transduction pathways: The road taken. Science 268, 233-239. Clark IM, Cawston TE. (1989). Fragments of human fibroblast collagenase. Purification and characterization. Biochem. J. 263, 201-206. Cochran DL, Castellot, Jr., and Karnovsky MJ. (1985). Effect of heparin on vascular smooth muscle cells. II. Specific protein synthesis. J. Cell. Physiol. 124, 29-36. Cole AA, Chubinskaya S, Schumacher B, Huch K, Cs-Szabo G, Yao J, Mikecz K, Hasty KA, and Kuettner KE. (1996). Chondrocyte matrix metalloproteinase-8. Human articular chondrocytes express neutrophil collagenase. J. Biol. Chem. 271, 11023-11026. Collier IE, Bruns GAP, Goldberg GI, and Gerhard DS. (1991). On the structure and chromosome location of the 72- and 92-kDa human type IV collagenase genes. Genomics 9, 429-434. Collier IE, Krasnov PA, Strongin AY, Birkedal-Hansen H, and Goldberg GI. (1992). Alanine scanning mutagenesis and functional analysis of the fibronectin-like collagen-binding domain from human 92-kDA type IV. J. Biol. Chem. 267, 6776-6781. Collier IE, Smith J, Kronberger A, Bauer EA, Wilhelm SM, Eisen AZ, and Goldberg GI. (1988). The structure of the human skin fibroblast collagenase gene. J. Biol. Chem. 263, 10711-10713. Collier IE, Wilhelm SM, Eisen AZ, Marmer BL, Grant GA, Seltzer JL, Kronberger A, He C, Bauer EA, and 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. Constantine KL, Madrid M, Banyai L, Trexler M, Patthy L, and Llinas M. (1992). Refined solution structure and ligand-binding properties of PDC-109 domain b. A collagen-binding type II domain. J. Mol. Biol. 223, 281-298. Constantine KL, Ramesh V, Banyai L, Trexler M, Patthy L, and Llinas M. (1991). Sequence-specific ! H NMR assignments and structural characterization of bovine seminar fluid protein PDC-109 domain b. Biochemistry 30, 1663-1672. Corcoran ML, Stetler-Stevenson WG. (1995). Tissue inhibitor of metalloproteinase-2 stimulates fibroblast proliferation via a cAMP-dependent mechanism. J. Biol. Chem. 270, 13453-13459. Cossins J, Dudgeon TJ, Catlin G, Gearing AJH, and Clements JM. (1996). Identification of MMP-18, a putative novel human matrix metalloproteinase. Biochem. Biophys. Res. Commun. 228, 494-498. Crabbe T, Joannou C, and Docherty AJP. (1993). Human progelatinase A can be activated by autolysis at a rate that is concentration-dependent and enhanced by heparin bound to the C-terminal domain. Eur. J. Biochem. 218, 431-438. Creighton TE. (1990). Disulphide bonds between cysteine residues. In: Creighton, T.E. (ed.). Protein structure - a practical approach. Oxford University Press, Oxford, pp. 155-167. 168 Bibliography Crouch E, Sage H, and Bornstein P. (1980). Structural basis for apparent heterogeneity of collagens in human basement membranes: Type IV collagen contains two distinct chains. Proc. Natl. Acad. Sci. USA 77, 745-749. Cunningham BA, Wells JA. (1989). High-resolution epitope mapping of hGH-receptor interactions by alanine-scanning mutagenesis. Science 244, 1081-1085. D'Souza SE, Ginsberg MH, Burke TA, and Plow EF. (1990). The ligand binding site of the platelet integrin receptor GPIIb-IIIa is proximal to the second calcium binding domain ot its a subunit. J. Biol. Chem. 265, 3440-3446. Damsky CH, Librach C, Lim KH, Fitzgerald ML, McMaster MT, Janatpour M, Zhou Y, Logan SK, and Fisher SJ. (1994). Integrin switching regulates normal trophoblast invasion. Development. 120, 3657-3666. Dedhar S. (1990). Integrins and tumor invasion. BioEssays 12, 583-590. Dedhar S, Ruoslahti E, and Pierschbacher MD. (1987). A cell surface receptor complex for collagen type I recognizes the Arg-Gly-Asp sequence. J. Cell Biol. 104, 585-593. Dedhar S, Saulnier R, Nagle R, and Overall CM. (1993). Specific alterations in the expression of a3(3l and 006(34 integrins in highly invasive and metastatic variants of human prostate carcinoma cells selected by in vitro invasion through reconstituted basement membrane. Clin. Exp. Metastasis 11, 391-400. Desjardins M, Bendayan M. (1989). Heterogenous distribution of type TV collagen, entactin, heparan sulfate proteoglycan, and laminin among renal basement membranes as demonstrated by quantitative immunocytochemistry. J. Histochem. Cytochem. 37, 885-897. Dobeli H, Trzeciak A, Gillesen D, Matile H, Srivastava IK, Perrin LH, Jakob PE, and Certa U. (1990). Expression, purification, biochemical characterization and inhibition of recombinant Plasmodium falciparum aldolase. Mol. Biochem. Parasitol. 41, 259-268. Domenicucci C, Goldberg HA, Hofmann T, Isenman D, Wasi S, and Sodek J. (1988). Characterization of porcine osteonectin extracted from foetal calvariae. Biochem. J. 253, 139-151. Dufour S, Duband JL, and Thiery JP. (1986). Role of a major cell-substratum adhesion system in cell behavior and morphogenesis. Biol. Cell 58, 1-14. Eble JA, Golbik R, Mann K, and Kahn K. (1993). The albl integrin recognition site of the basement membrance collagen molecule [al(IV)]2a2(IV)]. EMBO J. 12, 4795-4802. Edgar D, Timpl R, and Thoenen H. (1984). The heparin-binding site of laminin is responsible for its effects on neurite outgrowth and neuronal survival. EMBO J. 3, 1463-1468. Edgar D, Timpl R, and Thoenen H. (1988). Structural requirements for the stimulation of neurite outgrowth by two variants of laminin and their inhibition by antibodies. J. Cell Biol. 106, 1299-1306. Edwards DR, Beaudry PP, Laing TD, Kowal V, Leco KJ, Leco PA, and Lim MS. (1996). The roles of tissue inhibitors of metalloproteinases in tissue remodelling and cell growth. Int. J. Obes. 20 (Suppl. 3), S9-S15. 169 Bibliography Edwards DR, Waterhouse P, Holman ML, and Denhardt DT. (1986). A growth-responsive gene (16C8) in normal mouse fibroblasts homologous to a human collagenase inhibitor with erythroid-potentiating activity: evidence for inducible and constitutive transcripts. Nucleic Acids Res. 14, 8863-8878. Elenius K, Jalkanen M. (1994). Function of the syndecans - a family of cell surface proteoglycans. J. Cell Sci. 107, 2975-2982. Elenius K, Salmivirta M, Inki P, Mali M, and Jalkanen M. (1990). Binding of human syndecan to extracellular matrix proteins. J. Biol. Chem. 265, 17837-17843. Elenius K, Vainio S, Laato M, Salmivirta M, Thesleff I, and Jalkanen M. (1991). Induced expression of syndecan in healing wounds. J. Cell. Physiol. 114, 585-595. Elices MJ, Hemler ME. (1989). The human integrin VLA-2 is a collagen receptor on some cells and a collagen/laminin receptor on others. Proc. Nad. Acad. Sci. USA 86, 9906-9910. Eltis LD, Iwagami SG, and Smith M. (1994). Hyperexpression of a synthetic gene encoding a high potential iron sulfur protein. Protein. Eng. 7, 1145-1150. Emmert-Buck MR, Emonard HP, Corcoran ML, Krutzsch HC, Foidart J-M, and Stetler-Stevenson WG. (1995). Cell surface binding of TIMP-2 and pro-MMP-2/TIMP-2 complex. FEBS Lett. 364, 28-32. Emonard HP, Remacle AG, Noel AC, Grimaud J-A, Stetler-Stevenson WG, and Foidart J-M. (1992). Tumor cell surface-associated binding site for the M,. 72,000 type IV collagenase. Cancer Res. 52, 5845-5848. Engel J. (1996). Domain organizations of modular extracellular matrix proteins and their evolution. Matrix Biol. 15, 295-299. Engvall E, Bell ML, and Ruoslahti E. (1981). Affinity chromatography of collagen on collagen-binding fragments of fibronectin. Coll. Relat. Res. 1, 505-516. Engvall E, Ruoslahti E. (1977). Binding of soluble form of fibroblast surface protein, fibronectin, to collagen. Int.J.Cancer 20, 1-5. Esch FS, Ling NC, Bohlen P, Ying SY, and Guillemin R. (1983). Primary structure of PDC-109, a major protein constituent of bovine seminal plasma. Biochem. Biophys. Res. Commun. 113, 861-867. Ferran DS, Sobel M, and Harris RB. (1992). Design and synthesis of a helix heparin-binding peptide. Biochemistry 31, 5010-5016. Fessler JH and Fessler LI. (1987). Type V collagen. In: Mayne, R. and Burgeson, R.E. (eds.). Structure and function of collagen types. Academic Press, New York. pp. 81-97. Fessler LI, Duncan KG, Fessler JH, Salo T, and Tryggvason K. (1984). Characterization of the procollagen IV cleavage products produced by a specific tumor collagenase. J. Biol. Chem. 259, 9783-9789. ffrench-Constant C, Van De Water L, Dvorak HF, and Hynes RO. (1989). Reappearance of an embryonic pattern of fibronectin splicing during wound healing in the adult rat. J. Cell 170 Bibliography Biol. 109, 903-914. Fields GB, Van Wart HE. (1992). Unique features of the tissue collagenase cleavage site in interstitial collagens. Matrix Suppl. 1, 68-70. Fields GB, Van Wart HE, and Birkedal-Hansen H. (1987). Sequence specificity of human skin fibroblast collagenase. Evidence for the role of collagen structure in determining the collagenase cleavage site. J. Biol. Chem. 262, 6221-6226. Fini ME, Plucinska IM, Mayer AS, Gross RH, and Brinckerhoff CE. (1987). A gene for rabbit synovial cell collagenase: member of a family of metalloproteinases that degrade the connective tissue matrix. Biochemistry 26, 6156-6165. Fisher SJ, Cui TY, Zhang L, Hartman L, Grahl K, Zhang GY, Tarpey J, and Damsky CH. (1989). Adhesive and degradative properties of human placental cytotrophoblast cell lines in vitro. J. Cell Biol. 109, 891-902. Fitch JM, Birk DE, Mentzer A, Hasty KA, Mainardi C, and Linsenmayer TF. (1988). Corneal collagen fibrils: Dissection with specific collagenases and monoclonal antibodies. Invest. Ophthalmol. Vis. Sci. 29, 1125-1136. Folkman J, Shing Y. (1992). Angiogenesis (Review). J. Biol. Chem. 267, 10931-10934. Freije JM, Diez-Itza I, Balbin M, Sanchez LM, Blasco R, Tolivia J, and 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. Fridman R, Fuerst TR, Bird RE, Hoyhtya M, Oelkuct M, Kraus S, Komarek D, Liotta LA, Berman ML, and Stetler-Stevenson WG. (1992). Domain structure of human 72-kDa gelatinase type IV collagenase characterization of proteolytic activity and identification of the tissue inhibitor of metalloproteinase-2 (TIMP-2) binding regions. J. Biol. Chem. 267, 15398-15405. Fridman R, Fuerst TR, Hoyhtya M, Oelkuc M, Kraus S, Komarek D, Liotta LA, Berman ML, and Stetler-Stevenson WG. (1992). Domain structure of human 72-kDa gelatinase/type IV collagenase. Characterizaiton of proteolytic activity and identification of the tissue inhibitor of metalloproteinase-2 (TIMP-2) binding regions. J. Biol. Chem. 267, 15398-15405. Friedman R, Toth M, Pena D, and Mohashery S. (1995). Activation of progelatinase B (MMP-9) by gelatinase A (MMP-2). Cancer Res. 55, 2548-2555. Frisch SM, Morisaki JH. (1990). Positive and negative transcriptional elements of the human type IV collagenase gene. Mol. Cell. Biol. 10, 6524-6532. Frisch SM, Reich R, Collier IE, Genrich LT, Martin G, and Goldberg GI. (1990). Adenovirus El A represses protease gene expression and inhibits metastasis of human tumor cells. Oncogene 5, 75-83. Fulmer HM, Gibson W. (1966). Collagenolytic activity in gingivae of man. Nature 209, 728-729. Gallagher JT. (1989). The extended family of proteoglycans: social residents of the pericellular zone. Curr. Opin. Cell Biol. 1201-1218. 171 Bibliography Gallagher IT, Waler A. (1985). Molecular distinctions between heparan sulphate and heparin. Analysis of sulphation patterns indicates that heparan sulphate and heparin are separate families of N-sulphated polysaccharides. Biochem. J. 230, 665-674. Garbisa S, Balin M, Daga-Gordini D, Fastelli G, Naturale M, Negro A, Semenzato G, and Liotta LA. (1986). Transient expression of type IV collagenolytic metalloproteinase by human mononuclear phagocytes. J. Biol. Chem. 261, 2369-2375. Gatalica B, Pulkkinen L, Li K, Kuokkanen K, Ryynanen M, McGrath JA, and Uitto J. (1997). Cloning of the human type XVII collagen gene (COL17A1), and detection of novel mutations in generalized atrophic benign epidermolysis bullosa. Am. J. Human Genet. 60, 352-365. Gehlsen KR, Dillner L, Engvall E, and Ruoslahti E. (1988). The human laminin receptor is a member of the integrin family of cell adhesion receptors. Science 241, 1228-1229. Gillies RJ, Didier N, and Denton M. (1986). Determination of cell number in monolayer cultures. Anal. Biochem. 159, 109-113. Glenny AT, Pope CG, Waddington H, and Wallace U. (1926). Immunological notes. J. Pathol. 29, 31-40. Goldberg GI, Manner BL, Grant GA, Eisen AZ, Wilhelm S, and He C. (1989). Human 72-kilodalton type IV collagenase forms a complex with a tissue inhibitor of metalloproteases designated TIMP-2. Proc. Natl. Acad. Sci. USA 86, 8207-8211. Goldberg GI, Strongin AY, Collier IE, Genrich LT, and Marmer BL. (1992). Interaction of 92-kDa type IV collagenase with the tissue inhibitor of metalloproteinases prevents dimerization, complex formation with interstitial collagenase, and activation of the proenzyme with stromelysin. /. Biol. Chem. 267, 4583-4591. Goldberg GI, Wilhelm SM, Kronberger A, Bauer EA, Grant GA, and Eisen AZ. (1986). Human fibroblast collagenase. Complete primary structure and homology to an oncogene transformation-induced rat protein. J. Biol. Chem. 261, 6600-6605. Gomis-Ruth F-X, Kress LF, and Bode W. (1993). First structure of a snake venom metalloproteinase: A Prototype for matrix metalloproteinases/collagenases. EMBO J. 12, 4151-4157. Gomis-Ruth F-X, Stoecker W, Huber R, Zwilling R, and Bode W. (1993). Refined 1.8 A x-ray crystal structure of astacin, a zinc-endopeptidase from the crayfish Astacus astacus L. Structure determination, refinement, molecular structure and comparison with thermolysin. J. Mol. Biol. 229, 945-968. Gonzales A-M, Buscaglia M, Ong M, and Baird A. (1990). Distribution of basic fibroblast growth factor in the 18-day rat fetus: Localization in the basement membranes of diverse tissues. J. Cell Biol. 110, 753-765. Gooley PR, O'Connell JF, Marcy Al, Cuca GC, Salowe SP, Bush BL, Hermes JD, Esser GC, Hagmann WK, Springer JP, et al. (1994). The NMR structure of the inhibited catalytic domain of human stromelysin-1. Nature Struct. Biol. 1, 111-118. Graf J, Iwamoto Y, Sasaki M, Martin GR, Kleinman HK, Robey FA, and Yamada Y. (1987). Identification of an amino acid sequence in laminin mediating cell attachment, chemotaxis, and receptor binding. Cell 48, 989-996. 172 Bibliography Grant DS, Leblond CP, Kleinman HK, Inoue S, and Hassell JR. (1989). The incubation of laminin, collagen IV, and heparan sulfate proteoglycan at 35 degrees C yields basement membrane-like structures. J. Cell Biol. 108, 1567-1574. Grant GA, Eisen AZ, Marmer BL, Roswit WT, and Goldberg GI. (1987). The activation of human skin fibroblast collagenase. J. Biol. Chem. 262, 5886-5889. Gray WR. (1977). Some kinetic aspects of crosslink biosynthesis. Adv. Exp. Med. Biol. 246, 461-466. Gray WR, Sandberg LB, and Foster JA. (1973). Molecular model for elastin structure and function. Nature 246, 461-466. Greene J, Wang M, Liu YE, Raymond LA, Rosen C, and Shi YE. (1996). Molecular cloning and characterization of human tissue inhibitor of metalloproteinase 4. J. Biol. Chem. 271, 30375-30380. Greenfield NJ. (1996). Methods to estimate the conformation of proteins and polypeptides from circular dichroism data. Anal. Biochem. 235, 1-10. Gross J, Lapierre CM. (1962). Collagenolytic activity in amphibian tissues: A tissue culture assay. Proc. Natl. Acad. Sci. USA 48, 1014-1022. Gross J, Nagai Y. (1965). Specific degradation of the collagen molecule by tadpole collagenolytic enzyme. Proc. Natl. Acad. Sci. USA 54, 1197-1204. Guan JL, Hynes RO. (1990). Lymphoid cells recognize an alternatively spliced segment of fibronectin via the integrin receptor alpha 4 beta 1. Cell 60, 53-61. Guidry C, Miller EJ, and Hook M. (1990). A second fibronectin-binding region is present in collagen alpha chains. J. Biol. Chem. 265, 19230-19236. Guillemette JG, Matsushima-Hibiya Y, Atkinson T, and Smith M. (1991). Expression in E. coli of a synthetic gene coding for horse heart myoglobin. Protein. Engng. 4, 585-592. Gum R, Lengyel E, Juarez J, Chen JH, Sato H, Seiki M, and Boyd D. (1996). Stimulation of 92-kDa gelatinase B promoter activity by ras is mitogen-activated protein kinase 1-independent and requires multiple transcription factor binding sites including closely spaced PEA3/ets and AP-1 sequences. J. Biol. Chem. 271, 10672-10680. Gutman A, Wasylyk B. (1990). The collagenase gene promoter contains a TPA and oncogene-responsive unit encompassing the PEA3 and AP-1 binding sites. EMBO J. 9, 2241-2246. Habeeb AFSA, Atassi MZ. (1970). Enzymic and immunochemical properties of lysozyme. Evaluation of several amino group reversible blocking reagents. Biochemistry 9, 4939-4944. Haggerty JG, Bretton RH, and Milstone LM. (1992). Identification and characterization of a cell surface proteoglycan on keratinocytes. J. Invest. Dermatol. 99, 374-380. Hahn S-HE, Yamada KM. (1979). Isolation and biological characterization of active fragments of the adhesive glycoprotein fibronectin. Cell 18, 1043-1051. 173 Bibliography Hakkinen L, Heino J, Koivisto L, and Larjava H. (1994). Altered interaction of human granulation-tissue fibroblasts with fibronectin is regulated by 0:5(31 integrin. Biochem. Biophys. Acta 1224, 33-42. Hakomori S, Fukuda M, Sekiguchi K, Carter WG. (1984). Fibronectin, laminin, and other extracellular glycoproteins. In: Piez, K.L. and Reddi, A.H. (eds.). Extracellular matrix biochemistry. Elsevier, New York. pp. 229 Hannigan GE, Leung-Hagesteijn C, Fitz-Gibbon L, Coppolino MG, Radeva G, Flimus J, Bell JC, and Dedhar S. (1996). Regulation of cell adhesion and anchorage-dependent growth by a new printegrin-linked protein kinase. Nature 379, 91-96. Harlow E, Lane D. (1988). Immunizations. Cold Spring Harbor Laboratory; 99p. Hasty KA, Jeffrey JJ, Hibbs MS, and Welgus HG. (1987). The collagen substrate specificity of human neutrophil collagenase. J. Biol. Chem. 262, 10048-10052. Hasty KA, Pourmotabbed TF, Goldberg GI, Thompson JP, Spinella DG, Stevens RM, and Mainardi CL. (1990). Human neutrophil collagenase. A distinct gene product with homology to other matrix metalloproteinases. J. Biol. Chem. 265, 11421-11424. Hemler ME. (1988). Adhesive protein receptors on hematopoietic cells. [Review]. Immunol. Today 9, 109-113. Hemler ME. (1990). VLA proteins in the integrin family: structures, functions, and their role on leukocytes. Annu. Rev. Immunol. 8, 365-400. Henriet P, Rousseau GG, and Eeckhout Y. (1992). Cloning and sequencing of mouse collagenase cDNA. Divergence of mouse and rat collagenase from the other mammalian collagenases. FEBS Lett. 310, 175-178. Herbst TJ, McCarthy JB, Tsilibary EC, and Furcht LT. (1988). Differential effects of laminin, intact type IV collagen, and specific domains of type TV collagen on endothelial cell adhesion and migration. J. Cell Biol. 106, 1365-1373. Heremans A, De Cock B, Cassiman J-J, Van den Berghe H, and David G. (1990). The core protein of the matrix-associated heparan sulfate proteoglycan binds to fibronectin. J. Biol. Chem. 265, 8716-8724. Heremans A, Van der Schueren B, De Cock B, Paulsson M, Cassiman J-J, Van den Berghe H, and David G. (1989). Matrix-associated heparan sulfate proteoglycan: Core protein-specific monoclonal antibodies decorate the pericellular matrix of connective tissue cells and the stromal side of basement membranes. J. Cell Biol. 109, 3199-3211. Hibbs MS, Hasty KA, Seyer JM, Kang AH, and Mainardi CL. (1985). Biochemical and immunological characterization of the secreted forms of human neutrophil gelatinase. J. Biol. Chem. 260, 2493-2500. Highberger JH, Corbett C, and Gross J. (1979). Isolation and characterization of a prptide containing the site of cleavage of the chick skin collagen alpha(I) chain by animal collagenases. Biochem. Biophys. Res. Commun. 89, 202-208. Hinek A, Rabinovitch M. (1994). 67-kD elastin-binding protein is a protective "companion" of extracellular insoluble elastin and intracellular tropoelastin. J. Cell Biol. 126, 563-574. 174 Bibliography Hinek A, Wrenn DS, Mecham RP, and Barondes SH. (1988). The elastin receptor is a galactoside binding protein. Science 239, 1539-1541. Hirose N, Blankenship DT, Krivanek MA, Jackson RL, and Cardin AD. (1987). Isolation and characterization of four heparin-binding cyanogen bromide peptides of human plasma apolipoprotein B. Biochemistry 26, 5505-5512. Hirose T, Patterson C, Pourmotabbed TF, Mainardi CL, and Hasty KA. (1993). Structure-function relationship of human neutrophil collagenase: Identification of regions responsible for substrate specificity and general proteinase activity. Proc. Natl. Acad. Sci. USA 90, 2569-2573. Hollecker M. (1990). Counting integral numbers of residues by chemical modification. In: Creighton, T.E. (ed.). Protein structure - a practical approach. Oxford University Press, Oxford, pp. 145-153. Hornebeck W, Tixier JM, and Robert L. (1986). Inducible adhesion of mesenchymal cells to elastic fibers: Elastonectin. Proc. Natl. Acad. Sci. USA 83, 5517-5520. Hostikka SL, Tryggvason K. (1988). The complete primary structure of the oc2 chain of human type IV collagen and comparison with the al(IV) chain. J. Biol. Chem. 263, 19488-19493. Howard EW, Banda MJ. (1991). Binding of tissue inhibitor of metalloproteinases-2 to two distinct sites on human 72 kDa gelatinase. Identification of a stabilization site. J. Biol. Chem. 266, 17972-17977. Howard EW, Bullen EC, and Banda MJ. (1991). Regulation of the autoactivation of human 72-kDa progelatinase by tissue inhibitor of metalloproteinase-2. J. Biol. Chem. 266, 13064-13069. Huhtala P, Chow LT, and Tryggvason K. (1990). Structure of the human type IV collagenase gene. J. Biol. Chem. 265, 11077-11082. Huhtala P, Eddy RL, Fan YS, Byers MG, Shows TB, and Tryggvason K. (1990). Completion of the primary structure of the human type IV collagenase preproenzyme and assignment of the gene (CLGY) to the q21 region of chromosome 16. Genomics 6, 554-559. Huhtala P, Humphries MJ, McCarthy JB, Tremble PM, Werb Z, and Damsky CH. (1995). Cooperative signaling by alpha 5 beta 1 and alpha 4 beta 1 integrins regulates metalloproteinase gene expression in fibroblasts adhering to fibronectin. J. Cell Biol. 129, 867-879. Huhtala P, Tuuttila A, Chow LT, Lohi J, Keski-Oja J, and Tryggvason K. (1991). Complete structure of the human gene for 92-kDa type IV collagenase. Divergent regulation of expression for the 92- and 72-kilodalton enzyme genes in HT-1080 cells. J. Biol. Chem. 266, 16485-16490. Huttenlocher A, Sandborg RR, and Horwitz AF. (1995). Adhesion in cell migration. Curr. Opin. Cell Biol. 7, 697-706. Hynes R. (1985). Molecular biology of fibronectin. Ann. Rev. Cell Biol. 1, 67-90. 175 Bibliography Hynes RO. (1987). Integrins: A family of cell surface receptors. Cell 48, 549-554. Hynes RO. (1990). Fibronectins. Springer-Verlag; New York. Hynes RO. (1992). Integrins: Versatility, modulation, and signalling in cell adhesion. Cell 69, 11-25. Ignatius MJ, Large TH, Houde M, Tawil JW, Barton A, Esch F, Carbonetto S, and Reichardt LF. (1990). Molecular cloning of the rat integrin al-subunit: a receptor for laminin and collagen. J. Cell Biol. I l l , 709-720. Indik Z, Yeh H, Ornstein-Goldstein N, Rosenbloom J. (1990). Structure of the elastin gene and alternative splicing of elastin mRNA. In: Sandell, L. and Boyd, C. (eds.). Genes for extracellular matrix proteins. Academic Press, New York. pp. 221-250. Indik Z, Yoon K, Morrow SD, Cicila B, Rosenbloom JC, and Ornstein-Goldstein N. (1987). Structure of the 3' region of the human elastin gene: Great abundance of Alu repetitive sequences and few coding sequences. Connect. Tissue Res. 16, 197-211. Ingham KC, Brew SA, and Isaacs BS. (1988). Interaction of fibronectin and its gelatin-binding domains with fluorescent-labeled chains of type I collagen. J. Biol. Chem. 263, 4624-4628. Ingham KC, Brew SA, and Migliorini MM. (1989). Further localization of the gelatin-binding determinants within fibronectin. Active fragments devoid of type II homologous repeat modules. J. Biol. Chem. 264, 16977-16980. Ingham KC, Brew SA, Migliorini SA, and Busby TF. (1993). Biochemistry 32,12548-12553. Inoguchi K, Yoshioka H, Khaleduzzaman M, and Ninomiya Y. (1995). The mRNA for alpha 1 (XIX) collagen chain, a new member of FACITs, contains a long unusual 3' untranslated region and displays many unique splicing variants. J. Biol. Chem. 117, 137-146. Iozzo RV, Cohen IR, Grassel S, and Murdoch AD. (1994). The biology of perlecan: the multifaceted heparan sulfate proteoglycan of basement membranes and pericellular matrices. Biochem. J. 302, 625-639. Iozzo RV, Murdoch AD. (1996). Proteoglycans of the extracellular environment: Clues from the gene and protein side offer novel perspectives in molecular diversity and function. FASEB J. 10, 598-614. Isaacs BS, Brew SA, and Ingham KC. (1989). Reversible unfolding of the gelatin-binding domain of fibronectin: Structural stability in relation to function. Biochemistry 28, 842-850. Jackson RL, Busch SJ, and Cardin AD. (1991). Glycosaminoglycans: Molecular properties, protein interactions, and role in physiological processes. Physiol. Rev. 71, 481-539. Jalkanen M, Jalkanen S, Bernfield M. (1991). Binding of extracellular molecules by cell surface proteoglycans. In: McDonald, J.A. and Mecham, R.P. (eds.). Biology of extracellular matrix: receptors for extracellular matrix. Academic Press, Orlando, pp. 1-37. Jalkanen S, Jalkanen M. (1992). Lymphocyte CD44 binds the COOH-terminal heparin-binding domain of fibronectin. J. Cell. Biol. 116, 817-825. 176 Bibliography J a l k a n e n S, J a l k a n e n M , B a r g a t z e R , T a m m i M , a n d B u t c h e r E C . ( 1 9 8 8 ) . B i o c h e m i c a l p r o p e r t i e s o f g l y c o p r o t e i n s i n v o l v e d i n l y m p h o c y t e r e c o g n i t i o n o f h i g h e n d o t h e l i a l v e n u l e s i n m a n . J. Immunol. 1 6 1 5 - 1 6 2 3 . J a n k n e c h t R , D e M a r t y n o f f G D , L o u J, H i p s k i n d R A , N o r d h e i m A , a n d S t u n n e n b e r g H G . ( 1 9 9 1 ) . R a p i d a n d e f f i c i e n t p u r i f i c a t i o n o f n a t i v e h i s t i d i n e - t a g g e d p r o t e i n e x p r e s s e d b y r e c o m b i n a n t v a c c i n i a v i r u s . Proc. Natl. Acad. Sci. USA 88, 8 9 7 2 - 8 9 7 6 . J e n n e D , S t a n l e y K K . ( 1 9 8 7 ) . N u c l e o t i d e s e q u e n c e a n d o r g a n i z a t i o n o f the h u m a n S - p r o t e i n g e n e : R e p e a t i n g p e p t i d e m o t i f s i n the " p e x i n " f a m i l y a n d a m o d e l f o r t h e i r e v o l u t i o n . Biochemistry 26, 6 7 3 5 - 6 7 4 2 . J i l e k F, H o e r m a n n H . ( 1979 ) . F i b r o n e c t i n ( C o l d - I n s o l u b l e G l o b u l i n ) , V I . I n f l u e n c e o f h e p a r i n a n d h y a l u r o n i c a c i d o n the b i n d i n g o f n a t i v e c o l l a g e n . Hoppe Seylers Z. Physiol. Chem. 360, 5 9 7 - 6 0 3 . J o h n s o n G D , H e r s h L B . ( 1 9 9 2 ) . C l o n i n g a rat m e p r i n c D N A r e v e a l s the e n z y m e i s a h e t e r o d i m e r . J. Biol. Chem. 267, 1 3 5 0 5 - 1 3 5 1 2 . K a d e r b h a i M A , A u s t e n B M . ( 1 9 8 5 ) . S t u d i e s o n the f o r m a t i o n o f i n t r a c h a i n d i s u l p h i d e b o n d s i n n e w l y b i o s y n t h e s i z e d b o v i n e p r o l a c t i n . R o l e o f p r o t e i n - d i s u l p h i d e i s o m e r a s e . Eur. J. Biochem. 153, 1 6 7 - 1 7 8 . K a g a n H , T r a c k m a n P . ( 1991 ) . P r o p e r t i e s a n d f u n c t i o n o f l y s y l o x i d a s e . Amer. J. Resp. Cell Mol. Biol. 5, 2 0 6 K a l l u n k i P , T r y g g v a s o n K . ( 1992 ) . H u m a n b a s e m e n t m e m b r a n e h e p a r a n su l fa te p r o t e o g l y c a n c o r e p r o t e i n : A 4 6 7 - k D p r o t e i n c o n t a i n i n g m u l t i p l e d o m a i n s r e s e m b l i n g e l e m e n t s o f the l o w d e n s i t y l i p o p r o t e i n r e c e p t o r , l a m i n i n , n e u r a l c e l l a d h e s i o n m o l e c u l e s , a n d e p i d e r m a l g r o w t h fac tor . J. Cell Biol. 116, 5 5 9 - 5 7 1 . K a o W W Y , B r e s s a n G W , a n d P r o c k o p D J . ( 1 9 8 2 ) . K i n e t i c o f i n c o r p o r a t i o n o f t r o p o e l a s t i n i n t o e la s t i c f i b e r s i n e m b r y o n i c c h i c k aorta . Connect. Tissue Res. 10, 2 6 3 - 2 7 4 . K e u n g W , S i l b e r E , a n d E p p e n b e r g e r U . ( 1 9 8 9 ) . Q u a n t i f i c a t i o n o f c e l l s c u l t u r e d o n 9 6 - w e l l p la tes . Anal. Biochem. 182, 16-19 . K h o k h a R , M a r t i n D C , a n d F a t a J E . ( 1995 ) . U t i l i z a t i o n o f t r a n s g e n i c m i c e i n the s t u d y o f m a t r i x d e g r a d i n g p r o t e i n a s e s a n d t h e i r i n h i b i t o r s . Cancer Metastasis Rev. 14, 9 7 - 1 1 1 . K h o k h a R , W a t e r h o u s e P , Y a g e l S, P e e y u s h K L , O v e r a l l C M , N o r t o n G , a n d D e n h a r d t D T . ( 1 9 8 9 ) . A n t i s e n s e R N A - i n d u c e d r e d u c t i o n i n m u r i n e T I M P l e v e l s c o n f e r s o n c o g e n i c i t y o n S w i s s 3 T 3 c e l l s . Science 243, 9 4 7 - 9 5 0 . K i s h n a n i N S , S t a s k u s P W , Y a n g T - T , M a s i a r z F R , a n d H a w k e s S P . ( 1 9 9 5 ) . I d e n t i f i c a t i o n a n d c h a r a c t e r i z a t i o n o f h u m a n t i s sue i n h i b i t o r o f m e t a l l o p r o t e i n a s e - e a n d d e t e c t i o n o f three a d d i t i o n a l m e t a l l o p r o t e i n a s e i n h i b i t o r a c t i v i t i e s i n e x t r a c e l l u l a r m a t r i x . Matrix Biol. 14, 4 7 9 - 4 8 8 . K j e l l e n L , L i n d a h l U . ( 1991 ) . P r o t e o g l y c a n s : S t r u c t u r e s a n d i n t e r a c t i o n s . Ann. Rev. Biochem. 60, 4 4 3 - 4 7 5 . K j e l l e n L , O l d b e r g A , a n d H o o k M . ( 1 9 8 0 ) . C e l l s u r f a c e h e p a r a n su l fa te . M e c h a n i s m s o f p r o t e o g l y c a n - c e l l a s s o c i a t i o n . J. Biol. Chem. 255, 1 0 4 0 7 - 1 0 4 1 3 . 177 Bibliography Klagsbrun M. (1990). The affinity of fibroblast growth factors (FGFs) for heparin; FGF-heparan sulfate interactions in cells and extracellular matrix. Curr. Opin. Cell Biol. 2, 857-863. Klebe RJ. (1974). Isolation of a collagen-dependent cell attachment factor. Nature 250, 248-251. Kleinman HK, Klebe RJ, and Martin GR. (1981). Role of collagenous matrices in the adhesion and growth of cells. J. Cell Biol. 88, 473-485. Kleinman HK, McGoodwin EB, and Klebe RJ. (1976). Localization of the cell attachment region in types I and II collagens. Biochem. Biophys. Res. Commun. 72, 426-431. Kleinman HK, McGoodwin EB, Martin GR, Klebe RJ, Fietzek PP, and Woolley DE. (1978). Localization of the binding site for cell attachment in the a 1(1) chain of collagen. J. Biol. Chem. 253, 5642-5646. Knauper V, Cowell S, Smith B, Lopez-Otin C, O'Shea M, Morris H, Zardi L, and Murphy G. (1997). The role of the C-terminal domain of human collagenase-3 (MMP-13) in the activation of procollagenase-3, substrate specificity, and tissue inhibitor of metalloproteinase interaction. J. Biol. Chem. 272, 7608-7616. Knauper V, Lopez-Otin C, Smith B, Knight G, and Murphy G. (1996). Biochemical characterization of human collagenase-3. J. Biol. Chem. 271, 1544-1550. Kuhl U, Ocalan M, Timpl R, and von der Mark K. (1986). Role of laminin and fibronectin in selecting myogenic versus fibrogenic cells from skeletal muscle cells in vitro. Dev. Biol. 117, 628-635. Kunkel TA, Roberts JD, and Zakour RA. (1993). Rapid and efficient site-specific mutagenesis without phenptypic selection. Methods Enzymol. 154, 367-382. Laemmli UK. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. Langholz O, Rockel D, Mauch C, Kozlowska E, Bank I, Krieg T, and Eckes B. (1995). Collagen and collagenase gene expression in three-dimensional collagen lattices are differentially regulated by alpl and a2(3l integrins. J. Cell. Biol. 131, 1903-1915. Larjava H, Hakkinen L, and Rahemtulla F. (1992). A biochemical analysis of human periodontal tissue proteoglycans. Biochem. J. 284, 267-274. Larjava H, Lyons JG, Salo T, Makela M, Koivisto L, Birkedal-Hansen H, Akiyama SK, Yamada KM, and Heino J. (1993). Anti-integrin antibodies induce type rv collagenase expression in keratinocytes. J. Cell. Physiol. 157, 190-200. Laurie GW, Bing JT, Kleinman HK, Hassell JR, Aumailley M, Martin GR, and Feldmann RJ. (1986). Localization of binding sites for laminin, heparan sulfate proteoglycan and fibronectin on basement membrane (Type rV) collagen. J. Mol. Biol. 189, 205-216. Leahy DJ, Hendrickson WA, Aukhil I, and Erickson HP. (1992). Structure of a fibronectin type III domain from tenascin phased by MAD analysis of the selenomethionyl protein. Science 258, 987-991. 178 Bibliography Leblond CP, Inoue S. (1989). Structure, composition, and assembly of basement membrane. Am. J. Anat. 185, 367-390. Leco KJ, Khokha R, Pavloff N, Hawkes SP, and Edwards DR. (1994). Tissue inhibitor of metalloproteinases-3 (TIMP-3) is an extracellular matrix-associated protein with a distinctive pattern of expression in mouse cells and tissues. J. Biol. Chem. 269, 9352-9360. Lepage T, Gache C. (1990). Early expression of a collagenase-like hatching enzyme gene in the sea urchin embryo. EMBO J. 9, 3993-3012. Librach CL, Werb Z, Fitzgerald M L , Chiu K, Corwin N M , Esteves RA, Grobelny D, Galardy R, Damsky C H , and Fisher SJ. (1991). 92-kD type IV collagenase mediates invasion of human cytotrophoblasts. J. Cell Biol. 113, 437-449. Libson A M , Gitis A G , Collier IE, Marmer BL, Goldberg GI, and Lattman EE. (1995). Crystal structure of the hemopexin-like C-terminal domain of gelatinase A. Nature Struct. Biol. 2, 938-942. Lindahl TJ. (1989). Biosynthesis of heparin and related polysaccharides. In: Lane, D.A. and Lindahl, TJ. (eds.). Heparin, chemical and biological properties, clinical applications. Edward Arnold, London, pp. 159-189. Lindahl TJ, Feingold DS, and Roden L. (1986). Biosynthesis of heparin. Trends Biochem.. Sci. 11, 221-225. Linsenmayer TF. (1991). Collagen. In: Hay, E.D. (ed.). Cell Biology of Extracellular Matrix. 2nd ed. Plenum Press, New York. pp. 7-44. Liotta L A , Abe S, Gehron RP, and Martin GR. (1979). Preferential digestion of basement membrane collagen by an enzyme derived from a metastatic murine tumor. Proc. Natl. Acad. Sci. USA 76, 2268-2272. Liotta L A , Stetler-Stevenson WG. (1991). Tumor invasion and metastasis: An imbalance of positive and negative regulation. Cancer Res. 51 (Suppl.), 5054S-5059S. Liotta L A , Tryggvason K, Garbisa S, Hart I, Folts C M , and Shafie S. (1980). Metastatic potential correlates with enzymatic degradation of basement membrane collagen. Nature 284, 67-68. Litvinovich SV, Strickland DK, Medved L V , and Ingham K C . (1991). Domain structure and interactions of the type I and type U modules in the gelatin-binding region of fibronectin. All six modules are independently folded. J. Mol. Biol. 217, 563-575. Lobel P, Dahms N M , Breitmeyer J, Chirgwin JM, and Kornfeld S. (1987). Cloning of the bovine 215-kDa cation-independent mannose 6-phosphate receptor. Proc. Natl. Acad. Sci. USA 84, 2233-2237. Luck DN, Gout PW, Sutherland ER, Fox K, Huyer M , and Smith M . (1992). Analysis of disulphide bridge function in recombinant bovine prolactin using site-specific mutagenesis and renaturation under mild alkaline conditions: a crucial role for the central disulphide bridge in the mitogenic activity of the hormone. Protein. Engng. 5, 559-567. Lyons-Giordano B, Conaway H, and Kefalides NA. (1987). The effect of heparin on 179 Bibliography fibronectin and thrombospondin synthesis by human smooth muscle cells. Biochem. Biophys. Res. Commun. 148, 1264-1269. Mackay AR, Ballin M, Pelina MD, Farina AR, Nason AM, Hartzler XL, and Thorgeirson UP. (1992). Effect of phorbol ester and cytokines on matrix metalloproteinase and tissue inhibitor of metalloproteinase expression in tumor and normal cell lines. Invasion Metastasis 12, 168-184. Mackay AR, Hartzler JL, Pelina MD, and Thorgeirsson UP. (1990). Studies on the ability of 65-kDa and 92-kDa tumor cell gelatinases to degrade type IV collagen. J. Biol. Chem. 265, 21929-21934. MacMullen BA, Fujikawa K. (1985). Amino acid sequence of the heavy chain of human alpha-factor Xlla (activated Hageman factor). J. Biol. Chem. 260, 5328-5341. Magrtuson VL, Young M, Schattenberg DG, Mancini MA, Chen D, Steffensen B, and Klebe RJ. (1991). The alternative splicing of fibronectin pre-mRNA is altered during aging and in response to growth factors. /. Biol. Chem. 266, 14654-14662. Mainardi CL, Hibbs MS, Hasty KA, and Seyer JM. (1984). Purification of a type V collagen degrading metalloproteinase from rabbit alveolar macrophages. Coll. Relat. Res. 4, 479-492. Majack RA, Bornstein P. (1984). Heparin and related glycosaminoglycans modulate the secretorty phenotype of vascular smooth muscle cells. J. Cell. Biol. Mann DM, Rpmm E, and Migliorini M. (1994). Delineation of the glycosaminoglycan-binding site in the human inflammatory response protein lactoferrin. J. Biol. Chem. 269, 23661-23667. Mann K, Deutzmann R, Paulsson M, and Timpl R. (1987). Solubilization of protein BM-40 from a basement membrane tumor with chelating agents and evidence for its identity with osteonectin and SPARC. FEBS Lett. 218, 167-172. Margalit H, Fischer N, and Ben-Sasson SA. (1993). Comparative analysis of structurally defined heparin binding sequences reveals a distinct spatial distribution of basic residues. J. Biol. Chem. 268, 19228-19231. Martin GR, Timpl R. (1987). Laminin and other basement membrane components. Ann. Rev. Cell Biol. 3, 57-85. Martin RG, Timpl R, Muller PK, and Kuhn K. (1985). The genetically distinct collagens. Trends Biochem. Sci. 10, 235 Mason IJ, Taylor A, Williams JG, Sage H, and Hogan BLM. (1986). Evidence from molecular cloning that SPARC, a major product of mouse embryo parietal endoderm, is related to an endothelial cell 'culture shock' glycoprotein of Mr 43 000. EMBO J. 5, 1465-1472. Massia SP, Hubbell JA. (1991). An RGD spacing of 440 nm is sufficient for Integrin Oypymediated fibroblast spreading and 140 nm for focal contact and stress fiber formation. J. Cell Biol. 114, 1089-1100. Masure S, Nys G, Fiten P, Van Damme J, and Opdenakker G. (1993). Mouse gelatinase B: cDNA cloning, regulation of expression and glycosylation in WEHI-3 macrophages and gene 180 Bibliography organization. Eur. J. Biochem. 218, 129-141. Matrisian L M . (1992). The matrix-degrading metalloproteinases. BioEssays 14, 455-463. Matrisian L M , Glaichenhaus N, Gesnel M C , and Breathnach R. (1985). Epidermal growth factor and oncogenes induce transcription of the same cellular mRNA in rat fibroblasts. EMBO J. 5, 1679-1686. Matrisian L M , Leroy P, Ruhlman C, Gesnel M C , and Breathnach R. (1986). Isolation of the oncogene and epidermal growth factor-induced transin gene: complex control in rat fibroblasts. Mol. Cell Biol. 6, 1679-1686. Matthews BW, Colman PM, Jansonius JN, Titani K, Walsh K A , and Neurath H. (1972). Structure of thermolysin. Nature New Biol. 238, 41-43. Matthews BW, Jansonius JN, Colman PM, Schoenborn BP, and Dupourque D. (1972). Three-dimensional structure of thermolysin. Nature New. Biol. 238, 37-41. Mecham RP, Heuser JE. (1990). Three-dimensional organization of extracellular matrix elastic cartilage as viewed by quick freeze, deep etch electron microscopy. Connect. Tissue Res. 24, 83-93. Mecham RP and Heuser JE. (1991). The elastic fiber. In: Hay, E.D. (ed.). Cell Biology of Extracellular Matrix. 2nd ed. Plenum Press, New York. pp. 79-109. Mecham RP, Hinek A, Griffin GL, Senior R M , and Liotta L. (1989). The elastin receptor shows structural and functional similarities to the 67-kDa tumor cell laminin receptor. J. Biol. Chem. 264, 16652-16657. Milam SB, Steffensen B, Haskin C, Zardeneta G, Chen D, Magnuson V L , and Klebe RJ. (1991). Cell adhesion proteins in oral biology. Crit. Rev. Oral Biol. Med. 2, 451-491. Moll U M , Youngleib G L , Rosinski KB, and Quigley JP. (1990). Tumor promoter-stimulated Mr 92,000 gelatinase secreted by normal and malignant human cells: Isolation and characterization of the enzyme from HT1080 tumor cells. Cancer Res. 50, 6162-6170. Morgan DO, Edman JC, Standring DN, Fried V A , Smith M C , Roth RA, and Rutter WJ. (1987). Insulin-like growth factor II receptor as a multifunctional binding protein. Nature 329, 301-307. Mould AP, Wheldon L A , Komoriya A, Wayner EA, Yamada K M , and Humphries MJ. (1990). Affinity chromatographic isolation of the melanoma adhesion receptor for the IIICS region of fibronectin and its identification as the integrin alpha4,betal. J. Biol. Chem. 265, 4020-4024. Muller D, Quantin B, Gesnel M C , Millon-Collard R, Abecasis J, and Breathnach R. (1988). The collagenase gene family in humans consists of at least four members. Biochem. J. 253, 187-192. Murphy G, Allan JA, Willenbrock F, Cockett MI, O'Commell JP, and Docherty AJP. (1992). The role of the C-terminal domain in collagenase and stromelysin specificity. J. Biol. Chem. 267, 9612-9618. Murphy G, Cockett MI, Stephens PE, Smith BJ, and Docherty AJP. (1987). Stromelysin is 181 Bibliography an activator of procollagenase. A study with natural and recombinant enzymes. Biochem. J. 248, 265-268. Murphy G, Cockett MI, Ward RV, and Docherty AJP. (1991). Matrix metalloproteinase degradation of elastin, type IV collagen and proteoglycan. Biochem. J. 277, 277-279. Murphy G, Docherty AJP. (1992). The matrix metalloproteinases and their inhibitors. Amer. J. Resp. Cell Mol. Biol. 7, 120-125. Murphy G, Hembry RM. (1992). Proteinases in rheumatoid arthritis. J. Rheumatol. 19, 61-64. Murphy G, Hembry RM, McGarrity AM, and Reynolds JJ. (1989). Gelatinase (type TV collagenase) immunolocalization in cells and tissues: Use of an antiserum to rabbit bone gelatinase that identifies high and low M r forms. J. Cell. Sci. 92, 487-495. Murphy G, McAlpine CG, Poll CT, and Reynolds JJ. (1985). Purification and characterization of a bone metalloproteinase that degrades gelatin and types IV and V collagen. Biochem. Biophys. Acta 831, 49-58. Murphy G, Nguyen Q, Cockett MI, Atkinson SJ, Allan JA, Knight CG, Willenbrock F, and Docherty AJP. (1994). Assessment of the role of the fibronectin-like domain of gelatinase A by analysis of a deletion mutant. J. Biol. Chem. 269, 6632-6636. Murphy G, Ward RV, Hembry RM, Reynolds JJ, Kuhn K, and Tryggvason K. (1989). Characterization of gelatinase from pig polymorphonuclear leukocytes. A metalloproteinase resembling tumour type IV collagenase. Biochem. J. 258, 463-472. Murphy G, Willenbrock F, Ward RV, Cockett MI, Eaton D, and Docherty AJP. (1992a). The C-terminal domain of 72 kDa gelatinase A is not required for catalysis, but is essential for membrane activation and modulates interactions with tissue inhibitors of metalloproteinases. J. Biol. Chem. 267, 637-641. Murphy G, Willenbrock F, Ward RV, Cockett MI, Eaton D, and Docherty AJP. (1992b). The C-terminal domain of 72 kDa gelatinase A is not required for catalysis, but is essential for membrane activation and modulates interactions with tissue inhibitors of metalloproteinases. Biochem. J. 283, 637-641. Nagase H, Enghild JJ, Suzuki K, and Salvesen G. (1990). Stepwise activation of the precursor of matrix metalloproteinase 3 (stromelysin) by proteinases and (4-aminophenyl) mercuric acetate. Biochemistry 29, 5783-5789. Nakano T, Scott PG. (1987). Partial purification and characterization of a neutral proteinase with collagen telopeptidase activity produced by human gingival fibroblasts. Biochem. Cell Biol. 65, 286-292. Narayan A, Siegel RC, and Martin G. (1974). Stability and purification of lysyl oxidase. Arch. Biochem. Biophys. 162, 231 Nguyen Q, Murphy G, Hughes CE, Mort JS, and Roughley PJ. (1993). Matrix metalloproteinases cleave at two distinct sites on human cartilage link protein. Biochem. J. 295, 595-598. Nicholson R, Murphy G, and Breathnach R. (1989). Human and rat malignant-tumor-associated mRNAs encode stromelysin-like metalloproteinases. Biochemistry 182 Bibliography 28,5195-5203. Niyibizi C, Chan R, Wu J-J, and Eyre DR. (1994). A 92 kDa gelatinase (MMP-9) cleavage site in native type V collagen. Biochem. Biophys. Res. Commun. 202, 328-333. Noonan D M , Fulle A, Valente P, Cai S, Horigan E, Sasaki M , Yamada Y, and Hassell JR. (1991). The complete sequence of perlecan, a basement membrane heparan sulfate proteoglycan, reveals extensive similarity with laminin A chain, low density lipoprotein-receptor, and the neural cell adhesion molecule. J. Biol. Chem. 266,22939-22947. Noonan D M , Horigan EA, Ledbetter SR, Vogeli G, Sasaki M , Yamada Y, and Hassell JR. (1988). Identification of cDNA clones encoding different domains of the basement membrane heparan sulfate proteoglycan. J. Biol. Chem. 263, 16379-16387. Ogata Y, Enghild JJ, and Nagase H. (1992). Matrix metalloproteinase 3 (stromelysin) activates the precursor for the human matrix metalloproteinase 9. J. Biol. Chem. 267, 3581-3584. Oh SP, Warman M L , Seldin MF, Cheng SD, Knoll JH, Timmons S, and Olsen BR. (1994). Clonong of cDNA and genomic DNA encoding human type XVIII collagen and localization of the alpha 1 (XVIII) collagen gene to mouse chromosome 10 and human chromosome 21. Genomics 19, 494-499. Ohuchi E, Imai K, Fujii Y, Sato H, Seiki M , and Okada Y. (1996). Membrane type 1 matrix metalloproteinase digests interstitial collagens and other extracellular matrix macromolecules. J. Biol. Chem. 272, 2446-2451. Okada Y, Gonoji Y, Naka K, Tomita K, Nakanishi I, Iwata K, Yamashita K, and Hayakawa T. (1992). Matrix metalloproteinase 9 (92-kDa gelatinase/type IV collagenase) from HT 1080 human fibrosarcoma cells. Purification and activation of the precursor and enzymic properties. J. Biol. Chem. 267, 217.12-21719. Okada Y, Morodomi T, Enghild JJ, Suzuki K, Yasui A, Nakanishi I, Salvesen G, and Nagase H. (1990). Matrix metalloproteinase 2 from human rheumatoid synovial fibroblasts. Purification and activation of the precursor and enzymic properties. Eur. J. Biochem. 194, 721-730. Oksala O, Salo T, Tammi R, Hakkinen L, Jalkanen M , Inki P, and Larjava H. (1995). Expression of proteoglycans and hyaluronan during wound healing. J. Histochem. Cytochem. 43, 125-135. Okumura Y, Sato H, Seiki M , and Kido H. (1997). Proteolytic activation of the precursor of membrane type 1 matrix metalloproteinase by human plasmin. FEBS Lett. 402, 181-184. Oldberg A, Ruoslahti E. (1982). Interactions between chondroitin sulfate proteoglycan, fibronectin, and collagen. J. Biol. Chem. 257, 4859-4863. Olins PO, Devine CS, Rangwala SH, and Kavka KS. (1988). The T7 phage gene 10 leader RNA, a ribosome-binding site that dramatically enhances the expression of foreign genes in Eschericia coli. Gene 73, 227-235. Olsen BR. (1991). Collagen biosynthesis. In: Hay, E.D. (ed.). Cell Biology of Extracellular Matrix. 2nd ed. Plenum Press, New York. pp. 177-220. 183 Bibliography Olsen BR. (1995). New insights into the function of collagens from genetic analysis. Curr. Opin. Cell Biol. 7, 720-727. Overall C M . (1991). Recent advances in matrix metalloproteinase research. Trends Glycosci. Glycotech. 3, 384-399. Overall C M . (1995). Repression of tissue inhibitor of matrix metalloproteinase (TIMP-1) expression by all-frans'-retinoic acid in rat bone cell populations: Comparison with transforming growth factor-P'l. J. Cell. Physiol. 164, 17-25. Overall C M , Limeback H. (1988). Identification and characterization of enamel proteinases isolated from developing enamel. Biochem. J. 256, 965-972. Overall C M , Sodek J. (1987). Initial characterization of a neutral metalloproteinase, active on native 3/4-collagen fragments, synthesized by ROS 17/2.8 osteoblastic cells, periodontal fibroblasts, and identified in gingival crevicular fluid. J: Dent. Res. 66, 1271-1282. Overall C M , Sodek J. (1990). Concanavalin-A produces a matrix-degradative phenotype in human fibroblasts. Induction/endogenous activation of collagenase, 72 kDa gelatinase, and PUMP-1 is accompanied by suppression of tissue inhibitor of matrix metalloproteinases. J. Biol. Chem. 265, 21141-21151. Overall C M and Sodek J. (1992). Reciprocal regulation of collagenase, 72 kDa gelatinase and TIMP gene expression and protein synthesis in human fibroblasts induced by concanavalin A. In: Birkedal-Hansen, H. (ed.). Matrix Metalloproteinases and Inhibitors. Gustav Fischer, Stuttgart, pp. 209-211. Overall C M , Wallon U M , Steffensen B, De Clerk Y, Tschesche H, Abbey RS. (1997). Substrate and TIMP interactions with human gelatinase A recombinant COOH-terminal hemopexin-like and fibronectin type II-like domains: Both the N- and C-dpmains of TIMP-2 bind the C-domain of gelatinase A. In: Edwards, D., Hawkes, S., Khokha, R. (eds.). Inhibitors of Metalloproteinases in Development and Disease. Gordon & Breach (in press). Overall C M , Wrana JL, and Sodek J. (1989). Independent regulation of collagenase, 72-kDa progelatinase, and metalloendoproteinase inhibitor expression in human fibroblasts by transforming growth factor-p. / . Biol. Chem. 264, 1860-1869. Owens RJ, Baralle FE. (1986). Mapping the collagen-binding site of human fibronectin by expression in Escherichia coli. EMBO J. 5, 2825-2830. Panayotou G, End P, Aumailley M , Timpl R, and Engel J. (1989). Domains of laminin with growth-factor activity. Cell 56, 93-101. Patthy L . (1991). Modular exchange principles in proteins. Curr. Opin. Struct. Biol. 1, 351-361. Patthy L. (1993). Modular design of proteases of coagulation, fibrinolysis, and complement activation: implications for protein engineering and structure-function studies. Methods Enzymol. 222, 10-21. Patthy L. (1994). Exons and introns. Curr. Opin. Struct. Biol. 4, 383-392. Patthy L. (1996). Exon shuffling and other ways of module exchange. Matrix Biol. 15, 301-310. 184 Bibliography Patthy L, Trexlef M , Vali Z, Banyai L , and Varadi A. (1984). Kringles: modules specialized for protein binding. Homology of the gelatin-binding region of fibronectin with the kringle structures of proteases. FEBS Lett. 171, 131-136. Paulsson M . (1987). Noncollagenous proteins of basement membranes. Coll. Relat. Res. 7, 443-461. Paulsson M , Aumailley M , Deutzmann R, Timpl R, Beck K, and Engel J. (1987). Laminin-nidogen complex. Eur. 7. Biochem. 166, 11-19. Pavloff N, Staskus PW, Kishnani NS, and Hawkes SP. (1992). A new inhibitor of metalloproteinases from chicken: ChIMP-3. A third member of the TTMP family. J. Biol. Chem. 267, 17321-17326. Pei D, Weiss SJ. (1995). Furin-dependent intracellular activation of the human stromelysin-3 zymogen. Nature 375, 244-247. Pei D, Weiss SJ. (1996). Transmembrane-deletion mutants of the membrane-type matrix metalloproteinase-1 process progelatinase A and express intrinsic matrix-degrading activity. J. Biol. Chem. 271, 9135-9140. Peltonen J, Jaakkola S, Gay S, Olsen DR, Chu M-L, and Uitto J. (1989). Expression of extracellular matrix genes by cultured human cells: localization mRNAs and antigenic epitopes. Anal. Biochem. 178, 184 Perris R, Johansson S. (1990). Inhibition of neural crest cell migration by aggregating chondroitin sulfate proteoglycans is mediated by their hyaluronan-binding region. Dev. Biol. 137, 1-12; Phillips DR, Charo IF, Parise L V , and Fitzgerald L A . (1988). The platelet membrane glycoprotein Ub-IIIa complex. Blood 71, 831-842. Pierschbacher M , Hayman EG, and Ruoslahti E. (1983). Synthetic peptide with cell attachment activity of fibronectin. Proc. Natl. Acad, Sci. ,USA 80, 1224-1227. Pierschbacher MD, Ruoslahti E. (1984). Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature 309, 30-33. Piez KA. (1967). Soluble collagen and the components resulting from its denaturation. In: Ramachandran, G.N. (ed.). Treatise on collagen. Academic Press, London, pp. 207-252. Piez KA. (1976). Primary structure. In: Ramachandran, G.N. and Reddi, A:H. (eds.). Biochemistry of Collagen. Plenum Press, New York. pp. 1-44. Pihlajaniemi T. (1995). Kollagenfamiljen: nya medlemmar och nya tankar om deras biosyntes. Nord. Med. 110, 96-99. Poulsom R, Pignatelli M , Stetler-Stevenson WG, Liotta L A , Wright PA, Jeffery RE, Longcroft JM, Rogers L, and Stamp GWH. (1992). Stromal expression of 72 kda type IV collagenase (MMP-2) and TIMP-2 mRNAs in colorectal neoplasia. Am. J. Pathol. 141, 389-396. Pourmotabbed TF. (1994). Relation between substrate specificity and domain structure of 185 Bibliography 92-kDa type IV collagenase. Ann. NY Acad. Sci. 732, 372-374. Prockop DJ, Kivirikko KI. (1995). Collagens: Molecular biology, diseases, and potentials for therapy. Annu. Rev. Biochem. 64, 403-434. Prowse KR, Tricoli JV, Klebe RJ, and Shows TB. (1986). Assignment of the human fibronectin structural gene to chromosome 2. Cytogenet. Cell Genet. 41, 42-46. Puente XS, Pendas A M , Llano E, Velasco G, and Lopez-Otin C. (1996). Molecular cloning of a novel membrane-type matrix metalloproteinase from a human breast carcinoma. Cancer Res. 56, 944-949. Putnins EE, Firth JD, and Uitto V-J. (1995). Keratinocyte growth factor stimulation of gelatinase (matrix metalloproteinase-9) and plasminogen activator in histiotypic epithelial cell culture. J.-Invest. Dermatol. 104, 989-994. Pyke C, Ralfkjaer E, Tryggvason K, and Danoe K. (1993). Messenger RNA for two type TV collagenases is located in stromal cells in human colon cancer. Am. J. Pathol. 142, 359-365. Pytela R, Pierschbacher MD, Argraves S, Suzuki S, and Ruoslahti E . (1987). Arginine-glycine-aspartic acid adhesion receptors. Methods Enzymol. 144, 475-489. Pytela R, Pierschbacher MD, and Ruoslahti E. (1985a). A 125/115-kDa cell surface receptor specific for vitronectin interacts with the arginine-glycine-aspartic acid adhesion sequence derived from fibronectin. Proc. Natl. Acad. Sci. ,USA 82, 5766-5770. Pytela R, Pierschbacher MD, and Ruoslahti E . (1985b). Identification and isolation of a 140 kd cell surface glycoprotein with properties expected of a fibronectin receptor. Cell 40, 191-198. Quantin B, Murphy G, and Breathnach R. (1989). Pump-1 cDNA codes for a protein with characteristics similar to those of classical collagenase family members. Biochemistry 28, 5327-5334. Quinones S, Saus J, Otani Y, Harris ED, and Kurkinen M . (1989). Transcriptional regulation of human stromelysin. J. Biol. Chem. 264, 8339-8344. Ramachandran G N and Ramakrishan C. (1976). In: Ramachandran, G.N. and Reddi, A .H. (eds.). Biochemistry of collagen. Plenum Press, New York. pp. 45-81. Ramirez F, di Liberto M . (1990). Complex and diversified regulatory programs control the expression of vertebrate collagen genes. FASEB. J. 4, 1616-1623. Reid K B M , Day AJ. (1989). Immunol. Today 10, 177-180. Reponen P, Sahlberg C, Munaut C, Thesleff I, and Tryggvason K. (1994). High expression of 92-kD type TV collagenase (gelatinase B) in the osteoclast lineage during mouse development. J. Cell Biol. 124, 1091-1102. Riikonen T, Westermarck J, Koivisto L, Broberg A, Kahari V M , and Heino J. (1995). Integrin alpha 2 beta 1 is a positive regulator of collagenase (MMP-1) and collagen alpha 1(1) gene expression. J. Biol. Chem. 270, 13548-13552. Romer J, Bugge T H , Pike C, Lund LR, Flick MJ, Degen JL, and Dano K. (1996). Impaired 186 Bibliography wound healing in mice with a disrupted plasminogen gene. Nat. Med. 2, 287-292. Rosenbloom J, Cywinsky A. (1976). Inhibition of proline hydroxylase does not inhibit secretion of tropoelastin by chick aorta cells. FEBS Lett. 65, 246-250. Rosenbloom J, Harsch M . (1973). Hydroxyproline content determines the denaturation temperature of chick tendon collagen. Arch. Biochem. Biophys. 158, 478 Roskelley CD, Srebow A, and Bissell MJ. (1995). A hierarchy of ECM-mediated signalling regulates tissue-specific gene expression. Curr. Opin. Cell Biol. 7, 736-747. Ruoslahti E. (1988). Fibronectin and its receptors. Ann. Rev. Biochem. 57, 375-413. Ruoslahti E. (1991). Integrins. J. Clin. Invest. 87, 1-5. Ruoslahti E. (1996). RGD and other recognition sequences for integrins. Annu. Rev. Cell Dev. Biol. 12, 697-715. Ruoslahti E, Giancotti FG. (1989). Integrins and tumor cell dissemination. Cancer Cells 1, 119-126. Ruoslahti E, Pierschbacher MD. (1987). New perspectives in cell adhesion: RGD and integrins. Science 238, 491-497. Ruoslahti E, Yamaguchi Y. (1991). Proteoglycans as modulators of growth factor activities. Cell 64, 867-869. Sakamoto S, Goldhaber P, and Glimcher MJ. (1973). Mouse bone collagenase. The effect of heparin on the amount of enzyme released in tissue culture and on the activity of the enzyme. Calc. Tiss. Res. 12, 247-258. Sakashita S, Engvall E, and Ruoslahti E. (1980). Basement membrane glycoprotein laminin binds to heparin. FEBS Lett. 116, 243-246. Salo T, Liotta L A , and Tryggvason K. (1983). Purification and characterization of a murine basement membrane collagen-degrading enzyme secreted by metastatic tumor cells. J. Biol. Chem. 258, 3058-3063. Salo T, Lyons JG, Rahemtulla F, Birkedal-Hansen H, and Larjava H. (1991). Transforming growth factor-pi up-regulates type rv collagenase expression in cultured human keratinocytes. J. Biol. Chem. 266, 11436-11441. Salonen J, Domenicucci C, Goldberg HA, and Sodek J. (1990). Immunohistochemical localization of SPARC (Osteonectin) and denatured collagen and their relationship to remodelling in rat dental tissues. Archs. Oral. Biol. 35, 337-346. San Antonio JD, Lander AD, Karnovsky MJ, and Slayter HS. (1994). Mapping the heparin-binding sites on type I collagen monomers and fibrils. J. Cell Biol. 125, 1179-1188. San Antonio JD, Lander AD, Wright TC, and Karnovsky MJ. (1992). Heparin inhibits the attachment and growth of balb/c-3T3 fibroblasts on collagen substrata. J. Cell. Physiol. 150, 8-16. San Antonio JD, Slover J, Lawler J, Karnovsky MJ, and Lander AD. (1993). Specificity in 187 Bibliography the interactions of extracellular matrix proteins with subpopulations of the glycosaminoglycan heparin. Biochemistry 32, 4746-4755. Sanchez-Lopez R, Nicholson R, Gesnel M C , Matrisian L M , and Breathnach R. (1988). Structure-function relationships in the collagenase family member transin. J. Biol. Chem. 263, 11892-11899. Sanes JR, Engvall E , Butkowski R, and Hunter DD. (1990). Molecular homogeneity of basal laminae: isoforms of laminin and collagen IV at the neuromuscular junction and elsewhere. J. Cell. Biol. I l l , 1685-1699. Sang QA, Douglas DA. (1996). Computational sequence analysis of matrix metalloproteinases. J. Prot. Chem.. 15, 137-160. Sanger F, Coulson AR, Friedman T, Air G M , Barrell BG, Brown NL, Fiddes JC, Hutchinson III CA, Slocombe PM, and Smith M . (1978). The nucleotide sequence of bacteriophage 0X174. J. Mol. Biol. 125, 225-246. Santavicca M , Noel A, Angliker H, Stoll.I, Segain JP, Anglard P, Chretien M , Seidah N, and Basset P. (1996). Characterization of structural determinants and molecular mechanisms involved in pro-stromelysin-3 activation, by 4-aminophenylmercuric acetate and furin-type convertases. Biochem. J. 315, 953-958. Sato H, Seiki M . (1993). Regulatory mechanism of 92 kDa type IV collagenase gene expression which is associated with invasiveness of tumor cells. Oncogene 8, 395-405. Sato H, Takino T, Okada Y, Cao J, Shinagawa A, Yamamoto E, and Seiki M . (1994). A matrix metalloproteinase expressed on the surface of invasive tumor cells. Nature 370, 61-65. Schnierer S, Kleine T, Gote T, Hillemann A, Knauper V, and Tschesche H. (1993). The recombinant catalytic domain of human neutrophil collagenase lacks type I collagen substrate specificity. Biochem. Biophys. Res. Commun. 191, 319-326. Schbrpp M , Mattei M G , Herr I, Gack S, Schaper J, and Angel P. (1995). Structural organization and chromosomal localization of the mouse collagenase type I gene. Biochem. J. 308, 211-217. Schwarzbauer JE, Spencer CS, and Wilson CL. (1989). Selective secretion of alternatively spliced fibronectin variants. J. Cell Biol. 109, 3445-3453. Scott PG, Goldberg HA, and Dodd C M . (1983). A neutral proteinase from human gingival fibroblasts active against the c-terminal cross-linking region of type I collagen. Biochem. Biophys. Acta 114, 1064-1070. Seftor REB, Seftor EA, Gehlsen KR, Stetler-Stevenson WG, Brown PD, Ruoslahti E, and Hendrix MJC. (1992). Role of the integrin in human melanoma cell invasion. Proc. Natl. Acad. Sci. USA 89, 1557-1561. Seftor REB, Seftor EA, Stetler-Stevenson WG, and Hendrix MJC. (1993). The 72 kDa type IV collagenase is modulated via differential expression of ovpV and a5^ integrins during human melanoma cell invasion. Cancer Res. 53, 3411-3415. Seidah NG, Manjunath P, Rochemont J, Sairam MR, and Chretien M . (1987). Complete amino acid sequence of BSP-A3 from bovine seminal plasma. Homology to PDC-109 and 188 Bibliography to the collagen-binding domain of fibronectin. Biochem. J. 243, 195-203. Seltzer JL, Akers KT, Weingarten H, Grant GA, McCourt DW, and Eisen AZ. (1990). Cleavage specificity of human skin type IV collagenase (gelatinase). Identification of cleavage sites in type I gelatin, with confirmation using synthetic peptides. J. Biol. Chem. 265, 20409-20413. Seltzer JL, Eisen AZ, Bauer EA, Morris NP, Glanville RW, and Burgeson RE. (1989). Cleavage of type VII collagen by interstitial collagenase and type IV collagenase (gelatinase) derived from human skin. J. Biol. Chem. 264, 3822-3826. Senior RM, Griffin G L , Fliszar CJ, Shapiro SD, Goldberg GI, and Welgus HG. (1991). Human 92- and 72 kilodalton type IV collagenases are elastases. J. Biol. Chem. 266, ,7870-7875. Senior RM, Griffin GL, and Mecham RP. (1980). Chemotactic activity of elastin-derived peptides. / . Clin. Invest. 66, 859-862. Senior RM, Griffin GL, Mecham RP, Wrenn DS, Prasad K U , and Urry DW. (1984). Val-Gly-Val-Ala-Pro-Gly, a repeating peptide in elastin, is chemotactic for fibroblasts. and. monocytes. J. Cell Biol. 99, 870-874. Senior RM, Hinek A, Griffin GL, Pipoly DJ, Crouch EC, and Mecham RP. (1989). Neutrophils show chemotaxis to type IV collagen and its 7S domain and contain a 67 kD type IV collagen-binding protein with lectin properties. Amer. J. Resp. Cell Mol. Biol. 1,479-487. Seno M , Sasada R, Kurokawa T, and Igarashi K. (1990). Cafboxyl-terminal structure of basic fibroblast growth factor significantly contributes to its affinity for heparin. Eur. J. Biochem. 188,239-245. Serafini-Fracassini A. (1984). Elastinogenesis in embryonic and post-natal development. In: Ruggeri, A. and Motta, P.M. (eds.). Ultrastructure of the Connective Tissue Matrix. Nijhoff, The Hague, pp. 140-150. Shapiro SD, Fliszar CJ, Broekelmann TJ, Mecham RP, Senior RM, and Welgus HG. (1995). Activation of the 92-kDa gelatinase by stromelysin and 4-aminophenylmercuric acetate. Differential processing and stabilization of the carboxyl-terminal domain by tissue inhibitor of metalloproteinases (TTMP). J. Biol. Chem. 270, 6351-6356. Shapiro SD, Griffin GL, Gilbert DJ, Jenkins NA, Copeland NG, and Welgus HG. (1992). Molecular cloning, chromosomal localization, and bacterial expression of a murine macrophage metalloelastase. J. Biol. Chem. 267, 4664-4671. Shapiro SD, Kobayashi DK, and Ley TJ. (1993). Cloning and characterization of a unique elastolytic metalloproteinase produced by human alveolar macrophages. J. Biol. Chem. 268, 23824-23929. Shipley JM, Doyle GAR, Fliszar CJ, Ye Q-Z, Johnson L L , Shapiro SD, Welgus HG, and Senior R M . (1996). The structural basis for the elastolytic activity of the 92-kDa and 72-kDa gelatinases. Role of the fibronectin type II-like repeats. J. Biol. Chem. 271, 4335-4341. Siddiqa A, Ayad SR, and Dickinson P. (1989). The interaction between fibronectin and DNA. Anticancer. Res. 9, 373-381. 189 Bibliography Siegel RC. (1979). Lysyl oxidase. Zn?. Rev. Connect. Tissue Res. 8, 73-118. Singer II. (1979). The fibronexus: a transmembrane association of fibronectin-coating fibers and bundles of 5 nm microfilaments in hamster and human fibroblasts. Cell 16, 675-685. Skorstengaard K, Holtet T L , Etzerodt M , and Thogersen HC. (1994). Collagen-binding recombinant fibronectin fragments containing type II domains. FEBS Lett. 343, 47-50. Sodek, J, Heersche JN. (1981). Uptake of collagenolytic enzymes by bone cells during isolation from embryonic rat calvaria. Calcif, Tissue Int. 33, 255-260. Sodek J and Overall C M . (1992). Matrix metalloproteinases in periodontal remodelling. In: Birkedal-Hansen, H. (ed.). Matrix Metalloproteinases and Inhibitors: Gustav Fischer, Stuttgart, pp. 352-362. Sonnenberg A, Modderman PW, and Hogervorst F. (1988). Laminin receptor on platelets is the integrin VLA-6. Nature 336, 487-489. Springer TA. (1990). Adhesion receptors of the immune system. Nature 346, 425-434. Springman.EB, Angleton EL, Birkedal-Hansen H, and Van Wart HE. (1990). Multiple modes of activation of latent human fibroblast collagenase: Evidence for the role of a Cys 7 3 active-site zinc complex in latency and a "cysteine switch" mechanism for activation. Proc. Natl. Acad. Sci. USA 87, 364-368. Spurr NK, Gough A C , Gosden J, Rout D, Porteous DJ, Van Heyningen V, and Docherty AJ. (1988). Restriction fragment length polymorphism analysis and assignment of the metalloproteinases stromelysin and collagenase to the long arm of chromosome 11. Genomics 2, 119-127. Stams T, Spurlino JC, Smith DL, Wahl RC, Ho TF, Qoronfleh MW, Banks T M , and Rubin B. (1994). Structure of human neutrophil collagenase reveals large SI' specificity pocket. Nature Struct. Biol. 1, 119-123. ' ' Stathakis NE, Mosesson MW. (1977). Interactions among heparin, cold-insoluble globulin, and fibrinogen in formation of the heparin-precipitable fraction of plasma. J. Clin. Invest. 60, 855-865. . Steffensen B, Magnuson V L , Potempa C, Chen D, and Klebe RJ. (1992). a5 Integrin subunit expression changes during myogenesis. Biochem. Biophys. Acta 1137, 95-100. Steffensen B, Wallon U M , and Overall C M . (1995). Extracellular matrix binding properties of recombinant fibronectin type II-like modules of human 72-kDa gelatinase/type IV collagenase. High affinity binding to native type I collagen but not native type IV collagen. J. Biol. Chem. 270, 11555-11566. Stetler-Stevenson WG, Brown PD, Onisto M , Levy AT, and Liotta L A . (1990). Tissue inhibitor of metalloproteinases-2 (TIMP-2) mRNA expression in tumor cell lines and human tumor tissues. J. Biol. Chem. 265, 13933-13938. Stetler-Stevenson WG, Krutzsch HC, and Liotta LA. (1989). Tissue inhibitor of metalloproteinase (TIMP-2). A new member of the metalloproteinase inhibitor family. J. Biol. Chem. 264, 17374-17378. 190 Bibliography Stoecker W, Grams F, Baumann U, Reinemer P, Gomis-Ruth F-X, McKay DB, and Bode W. (1995). The metzincins - topological and sequential relations between the astacins, adamalysins, serralysins, and matrixins (collagenases) define a superfamily of zinc-peptidases. Protein Science 4, 823-840. Streuli CH, Schmidhauser C, Bailey N, Yurchenco P,.Skubitz AP, Roskelley C, and Bissell MJ. (1995). Laminin mediates tissue-specific gene expression in mammary epithelia. J. Cell Biol. 129, 591-603. Stricklin GP, Welgus HG. (1983). Human skin fibroblast collagenase inhibitor. Purification arid biochemical characterization. J. Biol. Chem. 258, 12252-12258. Strongin A Y , Collier IE, Bannikov G, Marmer BL, Grant GA, and Goldberg GI. (1995). Mechanism of cell surface activation of 72-kDa type IV collagenase. Isolation of the activated form of the membrane metalloprotease. J. Biol. Chem. 270, 5331-5338. Suzuki S, Argraves WS, Pytela R, Arai H, Krusius T, Pierschbacher MD, and Ruoslahti E. (1986). cDNA and amino acid sequences of the cell adhesion protein receptor recognizing vitronectin reveal a transmembrane domain and homologies with other adhesion protein receptors. Proc. Natl. Acad. Sci. ,USA 83, 8614-8618. Switalski L M , Murchison H, Timpl R, Curtiss in R, and Hook M . (1987). Binding of laminin to oral and endocarditis strains of viridans streptococci. J. Bacteriol. 169, 1095-1101. Takacs BJ, Girard M-F. (1991). Preparation of clinical grade proteins produced by recombinant DNA technologies. J. Immunol. Methods. 143, 231-240. Takino T, Sato H, Shinagawa A, and Seiki M . (1995). Identification of the second membrane-type matrix metalloproteinase (MT-MMP-2) gene from a human placenta cDNA library. MT-MMPs form a unique membrane-type subclass in the MMP family. J. Biol. Chem. 270, 23013-23020. Tamkun JW, DeSimone DW, Fonda D, Patel RS, Buck C, Horwitz AF, and Hynes RO. (1986). Structure of integrin, a glycoprotein involved in the transmembrane linkage between fibronectin and actin. Cell 46, 271-282. Tamkun JW, Hynes RO. (1983). Plasma fibronectin is synthesized and secreted by hepatocytes. J. Biol. Chem. 258, 4641-4647. Tashiro K, Sephel GC, Weeks B, Sasaki M , Martin GR, Kleinman HK, and Yamada Y. (1989). A synthetic peptide containing the IKVAV sequence from the A chain of laminin mediated cell attachment, migration, and neurite outgrowth. J. Biol. Chem. 264,16174-16182. Taylor M E , Conary JT, Lennartz MT, Stahl PD, and Drickamer K. (1990). Primary structure of the mannose receptor. contains multiple motifs resembling carbohydrate-recognition domains. J. Biol. Chem. 265, 12156-12162. Terranova VP, Lyall RM. (1986). Chemotaxis of human gingival epithelial cells to laminin: a mechanism for epithelial apical migration. J. Periodontol. 57, 311-317. Thayer M M , Flaherty K M , and McKay DB. (1991). Three-dimensional structure of the elastase of Pseudomonas aeruginosa at 1.5-A resolution. J. Biol. Chem. 266, 2864-2871. 191 Bibliography Timpl R. (1989). Structure and biological activity of basement membrane proteins. Eur. J. Biochem. 180, 487 Timpl R, Brown JC. (1994). The laminins. [Review] [38 refs]. Matrix Biology. 14, 275-281. Timpl R, Martin GR, Bruckner P, Wick G, and Wiedemann H. (1994). Nature of the collagenous protein in a tumor basement membrane. Eur. J. Biochem.. 84, 43-52. Timpl R, Weidemann H, VanDelden V, Furthmayr H, and Kuhn K. (1981). A network model for the organization of type TV collagen molecules in basement membranes. Eur. J. Biochem. 120,203-211. Tryggvason K, Hoyhtya M , and Pyke C. (1993). Type IV collagenases in invasive tumors. Breast Cancer Res. Treat. 24, 209-218. Uitto J, Hoffman H-P, and Prockop DJ. (1976). Synthesis of elastin and procollagen by cells from embryonic aorta. Differences in the role of hydroxyproline and the effects of proline analogs on the secretion of the two proteins. Arch. Biochem. Biophys. 173, 187-200. Uitto V-J, Larjava H. (1991). Extracellular matrix molecules and their receptors: An overview with special emphasis on periodontal tissues. Crit. Rev. Oral Biol. Med. 2, 323-354. Uitto V-J, Larjava H, Peltonen J, and Brunette D M . (1992). Expression of fibronectin and integrins in cultured epithelial cells. J. Dent. Res. 71, 1203-1211. Underwood PA, Bennett FA, Kirkpatrick A, Bean PA, and Moss BA. (1995). Evidence for the location of a binding sequence for the a2pi integrin of endothelial cells, in the pi subunit of laminin. Biochem. J. 309, 765-771. Vainio S, Jalkanen M , and Thesleff I: (1988). Cell surface proteoglycan and tenascin expression is induced by epithelial-mesenchymal interactions in embryonic mesenchyme. J. Cell Biol. 108, 1945-1953. van Boeckel C A A , Grootenhuis PDJ, and Visser A. (1994). A mechanism for heparin-induced potentiation of antithrombin III. Nature Struct. Biol. 1, 423-425. Van Wart HE, Birkedal-Hansen H. (1990). The cysteine switch: A principle of regulation of metalloproteinase activity with potential applicability to the entire matrix metalloproteinasel gene family. Proc. Natl. Acad. Sci. USA 87, 5578-5582. Vuorio E, De Crombrugghe B. (1990). The family of collagen genes. Annu. Rev. Biochem. 59, 837-872. Wallon U M , Overall C M . (1997). The COOH-terminal hemopexin-like domain of human gelatinase A (MMP-2) requires Ca 2 + for fibronectin and heparin binding: Binding properties of recombinant gelatinase A C-domain to extracellular matrix and basement membrane components. J. Biol. Chem. 272, 7473-7481. Wallon U M , Tschesche H, De Clerk Y, and Overall C M . (1997). J. Biol. Chem. Ward RV, Atkinson SJ, Reynolds JJ, and Murphy G. (1994). Cell surface-mediated activation of progelatinase A: demonstration of the involvement of the C-terminal-domain of progelatinase A in cell surface binding and activation of progelatinase A by primary fibroblasts. Biochem. J. 304, 263-269. 192 Bibliography Ward RV, Atkinson SJ, Slocombe PM, Docherty AJP, Reynolds JJ, and Murphy G. (1991). Tissue inhibitor of metalloproteinases-2 inhibits the activation of 72 kDa progelatinase by fibroblast membranes. Biochem. Biophys. Acta 1079, 242-246. Ward RV, Hembry RM, Reynolds JJ, and Murphy G. (1991). The purification of tissue inhibitor of metalloproteinases-2 from its 72 kDa progelatinase complex. Demonstration of the biochemical similarities of tissue inhibitor of metalloproteinases-2 and tissue inhibitor of metalloproteinases-1. Biochem. J. 278, 179-187. Wasi S, Otsuka K, Yao K - L , Tung PS, Aubin JE, Sodek J, and Termine JD. (1984). An osteonectin-like protein in porcine periodontal ligament and its synthesis by periodontal ligament fibroblasts. Can. J. Biochem. Cell Biol. 62, 470-478. Wasylyk C, Gutman A, Nicholson R, and Wasylyk B. (1991). The c-Ets oncoprotein activates the stromelysin promoter through the same elements as several non-nuclear oncoproteins. EMBO J. 10, 1127-1134. Wayner EA, Carter WG. (1987). Identification of multiple cell adhesion receptors for collagen and fibronectin inhuman fibrosarcoma cells possessing unique alpha and common beta subunits. J. Cell Biol. 105, 1873-1884. Wayner EA, Carter WG, Piotrowicz RS, and Kunicki TJ. (1988). The function of multiple extracellular matrix receptors in mediating cell adhesion to extracellular matrix; Preparation of monoclonal antibodies to the fibronectin receptor that specifically inhibit cell adhesion to fibronectin and react with platelet glycoproteins Ic-IIa. J. Cell Biol. 107, 1881-1891. Weeks BS, Schnaper HW, Handy M , Holloway E, and Kleinman HK. (1993). Human T lymphocytes synthesize the 92 kDa type rv collagenase (gelatinase B). J. Cell. Physiol. 157, 644-649. Welgus HG, Fliszar CJ, Seltzer JL, Schmid T M , and Jeffrey JJ. (1990a). Differential susceptibility of type X collagen to cleavage by two mammalian interstitial collagenases and 72-kDa type IV collagenase. J. Biol. Chem. 265, 13521-13527. Welgus HG, Fliszar CJ, Seltzer JL, Schmid T M , and Jeffrey JJ. (1990b). Differential susceptibility of type X collagen to cleavage by two mammalian interstitial collagenases and 72-kDa type IV collagenase. J. Biol. Chem 265, 13521-13527. Welgus HG, Jeffrey JJ, and Eisen AZ. (1981). The collagen substrate specificity of human skin fibroblast collagenase. J. Biol. Chem. 256, 9511-9515. Welgus HG, Stricklin GP, Eisen AZ, Bauer EA, Cooney RV, and Jeffrey JJ. (1979). A specific inhibitor of vertebrate collagenase produced by human skin fibroblasts. J. Biol. Chem. 254, 1938-1943. Werb Z, Tremble PM, Behrendtsen O, Crowley E, and Damsky CH. (1989). Signal transduction through the fibronectin receptor induces collagenase and stromelysin gene expression. J. Cell Biol. 109, 877^889. Whitelock JM, O'Grady RL, and Gibbins JR. (1991). Interstitial collagenase (matrix metalloproteinase 1) associated with the plasma membrane of both neoplastic and nonneoplastic cells. Invasion Metastasis 11, 139-148. 193 Bibliography Wick M , Burger C, Brusselbach S, Lucibello FC, and Muller R. (1994). A novel member of human tissue inhibitor of metalloproteinases (TTMP) gene family is regulated during GI progression, mitogenic stimulation, differentiation, and senescence. J. Biol. Chem. 269, 18953-18960. Wight TN, Kinsella M G , and Qwarnstrom EE. (1992). The role of proteoglycans in cell adhesion, migration, and proliferation. Curr. Opin. Cell Biol. 4, 793-801. Wilhelm SM, Collier IE, Kronberger A, Eisen AZ, Manner BL, Grant GA, Bauer EA, and Goldberg GI. (1987). Human skin fibroblast stromelysin: Structure, glycosylation, substrate specificity, and differential expression in normal and tumorigenic cells. Proc. Natl. Acad. Sci. USA 84, 6725-6729. Wilhelm SM, Collier IE, Manner BL, Eisen A Z , Grant GA, and 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. Will H, Hinzmann B. (1995). cDNA sequence and mRNA tissue distribution of a novel human matrix metalloproteinase with a potential transmembrane segment. Eur. J. Biochem. 231, 602-608. Willenbrock F, Crabbe T, Slocombe PM, Sutton CW, Docherty AJP, Cockett MI, O'Shea M , Brocklehurst K, Philips IR, and Murphy G. (1993); The activity of the tissue inhibitors of metalloproteinases is regulated by C-terminal domain interactions: A kinetic analysis of the inhibition of gelatinase A. Biochemistry 32, 4330-4337. Williams AF, Barclay AN. (1988). The immunoglobulin superfamily-domains for cell surface recognition. Annu. Rev. Immunol. 6, 381-405. Williamson RA, Marston FAO, Angal S, Koklitis P, Panico M , Morris HR, Carne AF, Smith BJ, Harris TJR, and Freedman RB. (1990). Disulphide bond assignment in human tissue inhibitor of metalloproteinases (TTMP). Biochem. J. 268, 267-274. Wilson C L , Matrisian L M . (1996). Matrilysin: a epithelial matrix metalloproteinase with potentially novel functions. Int. J. Biochem. Cell Biol. 28, 123-126. Wilson O, Jacobs A L , Stewart S, and Carson DD. (1990). Expression of externally-disposed heparin/heparan sulfate binding sites by uterine epithelial cells. J. Cell. Physiol. 143, 60-67. Windsor LJ, Birkedal-Hansen H, Birkedal-Hansen B, and Engler JA. (1991). An internal cysteine plays a role in the maintenance of the latency of human fibroblast collagenase. Biochemistry 30, 641-647. Woessner JF. (1991). Matrix metalloproteinases and their inhibitors in connective tissue remodeling. FASEB J. 5, 2145-2154. Woessner JF. (1994). The family of matrix metalloproteinases. In: Greenwald, R.A. and Golub, L . M . (eds.). Inhibition of matrix metalloproteinases. Therapeutic potential. The New York Academy of Sciences, New York. pp. 11-21. Wolfsberg TG, Bazan JF, Blobel CP, Myles DG, Primakoff P, and White JM. (1993). The precursor region of a protein active in sperm-egg fusion contains a metalloproteinase and a disintegrin domain: Structural, functional, and evolutionary implications. Proc. Natl. Acad. Sci. USA 90, 10783-10787. 194 Bibliography Wolfsberg TG, Primakoff P, Myles DG, and White JM. (1995). A D A M , a novel family of membrane proteins containing a disintegrin and metalloprotease domain: Multipotential functions in cell-cell and cell-matrix interactions. J. Cell Biol. 131, 275-278. Woodley DT, Bachmann PM, and O'Keefe EJ. (1988). Laminin inhibits human keratinocyte migration. J. Cell. Physiol. 136, 140 Woody RW. (1995). Circular dichroism. Methods Enzymol. 246, 34-71. Yamada K M . (1983). Cell surface interactions with extracellular materials. Ann. Rev. Biochem. 52, 761-799. Yamada K M , Miyamoto S. (1995). Integrin transmembrane signaling and cytoskeletal control. Curr. Opin. Cell Biol. 7, 681-689. Yanagishita M , Hascall V C . (1992). Cell surface heparan sulfate proteoglycans. J. Biol. Chem. 267,9451-9554. Yurchenco PD, Schittny JC. (1990). Molecular architecture of basement membranes. FASEB. J. 4, 1577-1590. Zucker S, Lysik R M , Malik M , Bauer BA, Caamano J, and Klein:Szanto AJP. (1992). Secretion of gelatinases and tissue inhibitors of metalloproteinases by human lung cancer cell lines and revertant cell lines: Not an invariant correlation with metastasis. Int.J.Cancer 52, 366-371. ; Zucker S, Moll U M , Lysik RM, DiMassimo EI, Stetler-Stevenson WG, Liotta L A , and Schwedes JW. (1990). Extraction of type IV collagenase/gelatinase from plasma membranes of human cancer cells. Int.J.Cancer 45, 1137-1142. Zucker S, Wieman JM, Lysik RM, Wilkie DP, Ramamurthy N, and Lane B. (1987). Metastatic mouse melanoma cells release collagen-gelatin degrading metalloproteinases as components of shed membrane vesicles. Biochem. Biophys. Acta 924, 225-237. 195 K263A ( 5 ' G G A A T G G A A A C G C G C A G G G C 3') K298/299A (5 'AAGCCATACT/GCCT/GCGTCGCGGTC 3') K330A (5 'CTCTCATATT/GCGTTGCCCAG 3') K343A ( 5 ' C A C C A C A T C J / G C T C C G T C A C T 3') K357A (5 'AAGCCCCACT/GCGCGGTCGTC 3') APPENDIX 1. Oligonucleotides used for incorporation of lysine to alanine codon changes in pGYMX123. Five degenerate 5'-phosphorylated oligonucleotides were used to generate four single and one double mutations in pGYMX123 by site-specific mutations (Chapter 3). The oligonucleotides were designed to also code for glutamate in place of lysine to generate swap mutations. However, only those proteins were expressed and tested which contained the lysine to alanine substitutions. Altered nucleotides are underlined. 196 APPENDIX 2. Relative reactivity of wildtype and mutant proteins in microwell substrate binding assay. To ensure that the detecting antibody (aHis6) had equal reactivity for wildtype rCBD123 and the mutant proteins, all proteins were coated at a two-fold serial dilution (1.0 to 0.01 |ig/well) in 96 microwell plates and detected using a constant concentration (1/500) of aHis 6 as described under "Experimental Procedures" (Chapter 3). The antibody showed equivalent reactivity to all tested proteins. Data points are averages of triplicate measurements and bars show SD. 197 Cell number (x "IO"3) APPENDIX 3. Cell number by crystal violet staining and Coulter counter. To verify that quantitation of cell number by crystal violet staining correlated with the cell number by Coulter counter, cells were seeded in 96 microwell plates at two-fold serial dilution (50,000 to 100, and 0 cells/well) in a-MEM. Cells in parallel wells were then either quantitated by crystal violet staining as described under "Experimental Procedures"(Chapter 4) or removed from the plates by mild trypsinization (0.05% trypsin, 0.2 mM EDTA) and counted using a Coulter counter. The cell numbers assessed by the two methods were characterized by a linear relationship confirming that quantitation by crystal violet staining was valid for quantitating number of cells bound. Each measurement point represents the mean of triplicate measurements by each method. Bars show SD for Coulter counter (horizontal) and crystal violet (vertical, not visible). 198 APPENDIX 4. Absence of crystal violet non-specific staining of coated protein in cell attachment assays Methods Assay conditions in this experiment of non-specific staining of coated protein by crystal violet corresponded to those used in the cell attachment assays (Chapter 4). rCBD123 was coated at 25 and 2.5 ug/ml for 18 h at 4 °C and followed by blocking with 10 mg/ml bovine serum albumin (BSA) for 20 min at 22 °C. BSA, also used for blocking in cell attachment assays, was coated at 10 mg/ml for 30 min at 22 °C. A non-coated and non-blocked plastic surface served as control for non-specific binding of the stain to plastic. After coating and blocking, surfaces were washed thoroughly with PBS, treated with fixative (4% formaldehyde, 5% sucrose, PBS) for 30 min at 22 °C and washed with dH 2 0. After air-drying, the plates were stained with crystal violet and quantitated in a microplate reader at 590 nm as described under "Experimental Procedures" (Chapter 4). Results rCBD123 (ug/ml) BSA (mg/ml) Plastic 25 2.5 0 10 -Mean 0.037 0.035 0.036 0.042 0.051 SD 0.003 0.002 0.001 0.001 0.001 The results demonstrated that binding of crystal violet to coated rCBD123 and BSA was equivalent to binding to plastic alone. Thus, this would produce a minimal background staining of equal magnitude for all wells. Therefore, the non-specific binding to plastic and to coated proteins was not considered in cell attachment assays. 199 15 min 45 min Mr Cs 1 2 3 4 1 2 3 4 HP u : immi 200-> • 4-^h\l 29 -> 97 - > w _ Collagenase 67-> 43 APPENDIX 5. Bacterial collagenase cleavage of native type I collagen. To determine experimental conditions for collagen cleavage to be applied in cell culture experiments, 100 (ig aliquots of native type I collagen in serum-free a-MEM were incubated with 0.075 U (2), 0.75 U (3), 7.5 U (4), or no (/) highly purified bacterial collagenase (Clostridiopeptidase A, Type III, fraction A; EC 3.4.24.3) (Sigma) for 15 or 45 min at 37 °C. Reaction products analyzed on 7.5% SDS-PAGE gels under reducing conditions demonstrated that native type I collagen was efficiently cleaved by 0.75 U of collagenase after 15 min. Positions of collagen a and p chains, kDa of molecular weight marker(Mr), collagenase (Cs), and dye front (fr) are indicated. 200 15 min 90 min M r 1 2 3 4 5 1 2 3 4 5 97 - » — T = T ~ ™ " fc= Collagenase 67 * 43 -> * 29-> — 18.4—> M » 14.4-* S APPENDIX 6. Bacterial collagenase does not cleave rCBD123. To verify that the inhibitory effects of bacterial collagenase on human gingival fibroblast attachment to rCBD123 did not result from cleavage of this protein, rCBD123 was incubated with highly purified bacterial collagenase (Clostridiopeptidase A, Type III, fractions A; EC 3.4.24.3) (Sigma). 5 ug aliquots of rCBD123 in serum-free a-MEM were incubated with bacterial collagenase at concentrations of 7.5 U (7), 0.75 U (2), 0.075 U (3), 0.0075 U (4), or without collagenase (5) in 100 ul reactions for 15 or 90 min at 37 °C. Analysis of reaction products on 15% SDS-PAGE gels under reducing conditions showed no signs of degradation of rCBD123 even at the highest concentration of bacterial collagenase after 90 min of incubation. Positions of rCBD123 and collagenase and kDas of molecular weight marker (Mr),and dye front (fr) are indicated. 201 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088121/manifest

Comment

Related Items