UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Characterisation of the collagen binding domain of gelatinase A : involvement of specific residues in… Moore, Todd Robert 2001

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

Item Metadata

Download

Media
831-ubc_2001-0254.pdf [ 4.39MB ]
Metadata
JSON: 831-1.0090113.json
JSON-LD: 831-1.0090113-ld.json
RDF/XML (Pretty): 831-1.0090113-rdf.xml
RDF/JSON: 831-1.0090113-rdf.json
Turtle: 831-1.0090113-turtle.txt
N-Triples: 831-1.0090113-rdf-ntriples.txt
Original Record: 831-1.0090113-source.json
Full Text
831-1.0090113-fulltext.txt
Citation
831-1.0090113.ris

Full Text

C H A R A C T E R I S A T I O N O F T H E C O L L A G E N BINDING DOMAIN O F G E L A T I N A S E A: I N V O L V E M E N T O F SPECIFIC R E S I D U E S IN T H E FIBRONECTIN T Y P E II M O D U L E S IN S U B S T R A T E R E C O G N I T I O N by T O D D R O B E R T M O O R E B . S c , The University of British Columbia, 1999 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE D E G R E E OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Oral Biological and Medical Sciences) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 2001 © Todd Robert Moore, 2001 In presenting this thesis in partial fulfilment 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 representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Oral Biological and Medical Sciences The University of British Columbia Vancouver, Canada Date . . . A f x i M , . 2 M Abstract The matrix metalloproteinase (MMP) family of endopeptidases can collectively degrade many components of the extracellular matrix. Their proteolytic activities have been implicated in normal processes such as extracellular matrix turnover and certain pathological conditions such as periodontitis, arthritis, and tumour metastasis. Based upon domain composition gelatinase A is separated from the other MMPs by the insertion of three contiguous fibronectin type II modules into its catalytic domain. Since it was determined that a recombinant domain consisting of the three fibronectin type II modules bound to native type I collagen this domain was termed the collagen binding domain (CBD). The function of the C B D is to donate substrate binding exosites to allow for broader enzyme specificity. Considering that the C B D contains exosites for binding native/denatured collagen, this domain was subjected to site-directed mutagenic studies to elucidate essential residues involved in collagen binding. The binding properties of the collagen binding domain to denatured type I collagen (gelatin) was investigated using a recombinant protein constructed of the second and third fibronectin type II module (rCBD23). Ten mutations were performed within the rCBD23 protein. Since the binding of substrate to the CBD is via a hydrophobic pocket, the mutations F264A, F264Y, F266A, F266Y, F322A, F324A, F322Y, F324Y, F264A/F322A, and F266A/F324A were introduced into rCBD23 in order to determine the effect of removing hydrophobic character. i i It was found that there was a decrease in the binding of the mutant proteins that was proportional to the distance between the mutated residue side chain and a strictly conserved tryptophan at the base of the hydrophobic pocket. Complete abrogation of gelatin binding was observed in the double mutant F266A/F324A. The observed effect was proposed to be the result of destabilisation of the strictly conserved tryptophan that forms the base of the hydrophobic pocket. These studies have furthered our knowledge and understanding of the interactions of C B D with substrate. The present study, combined with results from other studies, could be used to synthesize compounds that are specific to the C B D and are able to irreversibly inhibit the binding of C B D to substrate. A drug that would preferentially block binding of the C B D to collagen is proposed to reduced the metastasis of tumour cells by abolishment of type IV collagen binding. The overall effect is decreased tumour migration from the primary site of development and therefore substantially decreasing tumour progression. iii TABLE OF CONTENTS ABSTRACT . . . ii TABLE OF CONTENTS iv LIST OF TABLES vi LIST OF FIGURES vii LIST OF ABBREVIATIONS viii ACKNOWLEDGEMENTS x CHAPTER 1 INTRODUCTION E X T R A C E L L U L A R MATR IX C O M P O N E N T S 1 Co l lagenous Components . 2 Genera l Structure and Biosynthesis 2 Col lagen Famil ies 3 Fibri l-Forming Co l lagens 6 Network-Forming Co l lagens 6 Non-col lagenous Components . 7 G lycosaminog lycans 7 Proteoglycans 8 Glycoproteins 8 E X T R A C E L L U L A R MATR IX R E M O D E L L I N G 13 MATR IX M E T A L L O P R O T E I N A S E S 14 Physio logy and Pathology 14 M M P Fami l ies and Domain Structure 14 Gelat inase A 17 Regulat ion 18 Substrate Specif icit ies 20 Metastas is 22 Gelat inase A-mediated E C M Remodel l ing 23 iv E X O S I T E S 24 F I B R O N E C T I N T Y P E II M O D U L E S 25 Deletion Mutant Studies 26 Character isat ion 28 Structural Determination 29 T H E S I S A I M S 39 Identification of Speci f ic Res idues Within the S e c o n d and Third Fibronectin Type l l-Like Modu les CHAPTER 2 Materials and Methods 41 CHAPTER 3 Resul ts 46 CHAPTER 4 Discuss ion 53 CHAPTER 5 Conc lus ion 63 References 6 5 List of Tables Table 1 Collagens 4 Table 2 Proteoglycans 9 Table 3 Fibronectin Type ll-Like Module Alignment 30 Table 4 Mass Spectrometry 48 Table 5 ELISA and DMSO Summary 61 vi List of Figures Figure 1 Collagen Structures 5 Figure 2 Fibronectin 11 Figure 3 Laminin 12 Figure 4 MMP Structures 15 Figure 5 Gelatinase A acitvation 21 Figure 6 S D S - P A G E analysis of samples 47 Figure 7 ELISAs 49 Figure 8 Mini-columns 51 Figure 9 VAST alignment 56 Figure 10 F322 and F324 Orientation 58 vii List of A b b r e v i a t i o n s BSA : Bovine serum albumin C domain: Carboxyl-terminal domain C B D : col lagen binding domain CD : Circular dichroism cDNA: Complementary DNA COL : col lagenous domains C O S Y : chemica l shift correlated spectroscopy D M S O : dimethyl sulfoxide DTT: Dithiothreitol ECL : Enhanced chemi luminescence E C M : extracellular matrix E G F : Epidermal growth factor FACITs: fibri l-associated col lagens with interrupted triple helicies F P L C : Fast protein liquid chromatography G A G s : g lycosaminoglycans Gelat in: Denatured col lagen FN : fibronectin K D a kilodaltons MCP -3 : monocyte chemoattractant protein-3 M M P s : matrix metal loproteinase M T - M M P Membrane-type M M P NC: noncol lagenous domains Nm Nanomenter NMR: Nuclear magnetic resonance r: Recombinant PAGE: Polyacrylamide gel electrophoresis PBS: phosphate-buffered saline PMSF: Phenylmethylsulfonyl fluoride TGF-(3 Transforming growth factor-(3 TIMP: tissue inhibitor of metalloproteinase TPA: 12-0-tetradecanoylphorbol-13-acetate SDS: Sodium dodecyl sulphate ix A c k n o w l e d g e m e n t s The road to this junction in my life has been long and very chal lenging. A s long ago as e lementary schoo l I remember the be low ave rage g rades and the wonderment of my teachers when they saw that I wa s underach iev ing at everything I tried to do scholast ical ly. The one thing that remained is they all be l ieved. They be l ieved that I could do this and they were right. F rom Mr. Brendzy in grade 5, to Mr. Rootman in grade 6, and to Mr. Baptie in grade 11 and 12, they all saw a spark inside that at some point in time would ignite. My path into biochemistry was paved during my studies at Langara Communi ty Col lege when my instructor Mike Holmwood took an afternoon to enlighten me as to what courses I needed to take in order to learn about the complexit ies of life at the molecular level. My curiosity for sc ience and biochemistry b lossomed when I became a member of the Overal l laboratory. Within the fertile conf ines of the Overal l lab, Dr. Chr is Overal l took me under his wing, kindled my inquisit iveness, and showed me that there is a plethora of knowledge to be extracted from our world and I thank him. My many thanks are extended to all of the members of the Overal l laboratory for their guidance during my time within the Overal l laboratory. To Gay le for helping to superv ise my 449 project, to Angus for keeping the lab energ ized, to Andrea for being my " C B D buddy", to Dr. B for keeping me inline, to Eric for "keeping it real", to Dennis for helping me with my project and being my buddy, to Cl ive for his tonnes of knowledge, and to Heidi (Boo) for our invigorated d iscuss ions and for being there whenever I needed her. T h e most important thanks go out to my parents and s i s ter for their understanding during my growing years. They stoically endured all of the above t imes in my life and are therefore to be congratulated for their ever endur ing presence during my years of study and also must too celebrate in the c losure of this chapter of the book that is my life, vene vidi vici. xi Introduction Extracellular Matrix Components It has now been determined that the extracellular matrix ( E C M ) constitutes many d iverse comp lex st ructures that prov ide a mechan i ca l scaf fo ld for ce l lu lar adhes ion and migration. This mechanica l scaffold determines the histochemical subst ruc tures spec i f i c to every organ thus provid ing ce l l s with the correct b io logical information for growth and differentiation (Aumai l ley and G a y r a u d , 1999). Intense recent work dedicated to the elucidat ion of important structures and funct ions of the E C M has now further estab l ished that the E C M profoundly inf luences the major cel lular p rocesses of growth, differentiation, migration, and apoptos is (Boudreau and J o n e s , 1999). The E C M is c o m p o s e d of a variety of versat i le proteins, g lycoproteins, and po lysacchar ides that are secre ted locally and assemb led into an organized meshwork in c lose assoc iat ion with the surface of the cell that produced them. In connect ive t issues, the E C M is frequently more abundant than the cel ls that it surrounds and thus is a major determinant of the phys i ca l p roper t ies of the t i ssue . C o l l a g e n f ibres form the archi tectura l f ramework of the vertebrate body with varying amounts being found in different organs and structures such as skin and bone, where they are found to be the major component . In contrast, in other t issues such as the brain and spinal cord, connec t i ve t issue is only a minor componen t of the organ (Aumai l ley and Gay raud , 1999). 1 In most connective tissues, the matrix molecules are secreted predominantly by fibroblasts (Streuli, 1999). There are two main classes of molecules in the E C M : (1) the fibrous proteins that perform either a structural function (collagen and elastin) or an adhesive role (fibronectin and laminin) and (2) the ubiquitously distributed glycosaminoglycans (GAGs) and proteoglycans that perform many roles within the extracellular space (Hay, 1991). However there are approximately 140 molecules that comprise the entire E C M , plus many that also have variants based on alternative splicing. C o l l a g e n o u s c o m p o n e n t s G e n e r a l S t ruc tu re a n d B i o s y n t h e s i s The most abundant structural component of the E C M is collagen. Collagen is secreted by connective tissue cells as well as a variety of other cell types (Alberts et al., 1994). The characteristic feature of collagen is its long, stiff, triple helical structure in which three collagen polypeptide a chains are wound around each other into a coiled-coil conformation: Collagen has an unusual amino acid composition in which glycine, proline and hydroxyproline are the dominant amino acids. Further characterization of the a chains shows that these amino acids are arranged into a repetitious tripeptide sequence G ly -X-Y , that forms the collagenous (COL) domains and noncollagenous portions (NC domains) of variable length and location. Within the repetitious motif, X is frequently Pro and Y is frequently an hydroxyproline. The placement of a glycine at every third position allows each individual a chain to pack very tightly, therefore creating a formidable challenge to proteolytic attack. The repeating Pro residues exclude 2 the possibility that the polypeptide chains in collagen could adopt an a-helical or (3-sheet conformation. Instead, individual collagen polypeptide chains assume a poly-L-proline-ll helical secondary structure that aggregates into three-stranded cables with a right-handed twist (Zubay et al., 1995). Based on differing primary sequence, 34 different a chains have been cloned and sequenced, several of them being differentially spliced, to yield even more diversity within the collagen superfamily. Up to now 19 different combinations of a chains (collagen types I-XIX) have been either identified or predicted to exist within the superfamily of vertebrate collagens (Aumailley and Gayraud, 1999) (Table 1). Collagen Families Collagens have the ability to form highly organized polymers. The different collagen types can be grouped into several c lasses: (1) the fibril-forming collagens (types I, II, III, V, and XI), (2) the network forming collagens (basement membrane type IV and types VIII and X), (3) the collagens that associate with the surface of fibrillar collagen and thus are referred to as the fibril-associated collagens with interrupted triple helicies (FACITs; that include types IX, XII, XIV, XVI and XIX), (4) the collagen that forms beaded filaments (type VI), (5) the collagen that forms anchoring fibrils for basement membranes (VII), (6) the collagens with transmembrane domains (XIII and XVII), and (7) the recently named multiplexins (multiple triple-helix domain and interruptions, XV and XVIII) (Prockop and Kivirikko, 1995) (Figure 1). For the focus of this thesis the two classes of collagens will be reviewed are: 1) the fibrillar collagens (with special 3 Table J: The Collagen Family and Subfamilies Molecular Composition Tissue Distribution Fibrillar collagen II faOOla III [0,(111)13 V [a 1 (V)] 2 a 2 (V)[a 1 (V)a 2 (V)a 3 (V)] XI [a 1(XI)a 2(XI)a 3(XI)] and chimeras between oc(V) and a(XI) Network-forming Collagen IV [ ^ ( I V J W I V ) VIII ? X Most connective tissue Cartilage, Vitreous humor Extensible connective tissues Tissues containing type I Tissues containing type II Basement Membranes Many tissues, especially endothelial Hypertrophic cartilage Fibrillar-associated collagen with interrupted triple helicies IX XII XIV XVI XIX [a 1(IX)a 2(IX)a 3(IX)] [a 1(XII)] 3 [a,(XIV)] 3 [a,(XVI)] 3 ? Beaded filament VI [^(VOa^VOc^VI)] Anchoring Fibrils VII fafVIIMa Transmembrane-containing domains XIII ? XVII [cc^XVIOk Multiplexins X V ? XVIII ? Tissues containing type II Tissues containing type I Tissues containing type I Many tissues Rhabdomyosarcoma cells Most connective tissues Anchoring fibrils Many tissues Skin hemidesosomes Many tissues Many tissues, especially liver and kidney 4 Figure 1 Schematic representing the different structure of collagen molecules. (1) Fibrillar collagens - Collagen fibrils align themselves in a staggered fashion along the axis of collagen fibril, (2) Network-forming collagens - The collagen fibres organise themselves according to their NC domains to form a mat-like structure, (3) FACITs - These collagens are found on the surface of the complexed fibrillar collagens, (4) Beaded filament collagens, (5) Anchoring fibrils - anchors the basement membrane to the underliyng type IV collagen and laminin of the ECM, (6) Collagens with transmembrane domains, and (7) The multiplexins. Figure adapted from Prockop and Kivirikko (1 995) 5 interest dedicated to type I col lagen) and 2) network forming co l lagens (primarily basement membrane type IV col lagen). Fibr i l -Forming C o l l a g e n s The fibril-forming col lagens are similar in s ize and structure. T h e s e co l lagens are syn thes ized and secreted as larger precursor molecu les , the procol lagens, and are subsequen t l y N- and C- termina l ly p r o c e s s e d by spec i f i c ex t race l lu lar proteinases to co l lagens. This process ing results in formation of mature col lagen mo lecu les that contain large (300 nm) tr iple-hel ical doma ins of approximately 1000 am ino ac i ds that are terminated by very short N C s e q u e n c e s , the te lopept ides . In a p rocess te rmed f ibr i l logenes is , the co l l agen mo lecu les assemb le into cross-str iated fibrils where each molecule is d isp laced about one-quarter of its length relative to its nearest neighbor (Figure 1). Th is results in the formation of two major t i ssue-spec i f i c fibril lar po lymers : the type II co l lagen conta in ing fibri ls typical of car t i lage and the co l l agen I conta in ing fibri ls of interstitial connect ive t issues. The other types of fibril forming co l lagens have a more varied t issue distribution (Prockop and Kivirikko, 1995) (Table 1). Network-Forming C o l l a g e n s T h e network- forming co l l agens inc lude the fami ly of type IV co l l agens that constitute basement membranes, and the respect ive type VIII and X co l lagens of endothel ial and hypertrophic cart i lage. The C O L domain of a type IV co l lagen molecule is longer than that found in the fibril-forming co l lagens and cons is ts of - 1 4 0 0 amino ac ids of the - G l y - X - Y - motif that are frequently interrupted by short 6 N C s e q u e n c e s . T h e N-terminus of a co l lagen molecu le conta ins a smal l N C domain and the C-terminus contains a larger more signif icant N C domain of - 2 3 0 amino ac ids . Type IV col lagen molecules are capab le of self assemb ly into net-like structures whereby the co l lagen monomers d imer ize at the C-terminal and tetramerize at the N-terminus. In addit ion to these end-to-end interactions, the tr ip le-hel ical doma ins intertwine to form superco i led structures (Prockop and Kivirikko, 1995) (Figure 1). N o n c o l l a g e n o u s c o m p o n e n t s -G l y c o s a m i n o g l y c a n s T h e g l y c o s a m i n o g l y c a n s ( G A G s ) a re unb ranched p o l y s a c c h a r i d e cha ins compr ised of repeating sugar -aminosugar d isacchar ide subuni ts that are highly negatively charged . The G A G s can be broken into four main groups based on their residue l inkage, residue number and the location of the sulphate groups: (1) hya luronan, (2) chondroit in and dermatan su lphate, (3) heparan sulphate and hepar in , and (4) keratan su lphate . Due to the large number of negat ively charged sulphate groups, posit ively charged cat ions (Na + ) are attracted to the G A G s and in turn set up an osmot ic gradient that c a u s e s water to diffuse into the extracel lular s p a c e thus creat ing hydrostatic turgor pressure . Therefore, within the E C M the G A G s adopt a highly extended conformation that occup ies a very large volume relative to their mass . The G A G chains fill most of the extracellular space . Except for hyaluronan, all of the G A G s are found covalent ly at tached to a core protein in the form of a proteoglycan (Alberts et al., 1994). 7 P r o t e o g l y c a n s Proteoglycans can be considered to constitute a distinct subset of noncollagenous glycoproteins containing G A G side chains that range in weight from 10 to 600 kDa (Table 2). The associated G A G side chains of proteoglycans perform a major transient role in chemical signaling between cells by constituting extracellular gels of varying pore size and charge density that serve as a selective sieves which regulate the trafficking of macromolecules and cells according to their size and/or charge (Aumailley and Gayraud, 1999). It has been shown that several growth factors can be bound by either the G A G side chains or the core protein of proteoglycans as illustrated respectively by fibroblast growth factor binding to heparan sulphate (Moscatelli, 1987) and TGF-(3 binding to the core protein of decorin (Fukushima, et al., 1993). Proteoglycans have also been shown to bind and regulate the activities of proteases and protease inhibitors via different mechanisms including: sequestration, steric hindrance, pooling, protection from proteolytic degradation, and alteration of local concentration for more effective presentation to cell surface receptors (Streuli, 1999). G l y c o p r o t e i n s In addition to collagens and proteoglycans, many glycoproteins are building blocks of the E C M (Table 2). Elastin and fibrillins are ubiquitous proteins of connective t issues that form the so-called elastic fibres, in which elastin corresponds to the amorphous material and fibrillins, together with other small proteins, constitute the associated 10-12 nm microfibrils (Aumailley and Gayraud, 8 Table 2: Structural glycoproteins and proteoglycans of the E C M Glycoproteins Interstitial connective tissue Fibronectins Tenascins Fibrillins Elastin Microfibril-associated Matrilins Thrombospondins Basement Membrane Laminins Nidogen/entactin Fibulin Proteoglycans Small leucine-rich proteoglycans Decorin Biglycan Fibromodulin Lumincin Epiphycan Modular proteoglycans (hyaluronan- and lectin-binding) Aggreacan Versican Neurocan Brevican 9 1999). Fibri l l ins are a family of proteins consist ing almost entirely of epidermal growth factor motifs. The best s tudied E C M glycoprotein is f ibronectin (FN) . The molecu le has two subunits (250-280 kDa) disulphide bonded to each other to form a dimer with two 50 nm a rms (Figure 2). The amino terminus of F N conta ins three types of repetit ive modu les : 12 modu les of type I homo logy , 2 modu les of type II homology, and 15-17 modules of type III homology. The latter contr ibutes up to half of the mo lecu le and is present in var iab le numbers due to the comp lex alternative spl ic ing of a single F N gene thus giving rise to at least 20 variants in humans (Schwarzbauer and Sech ler , 1999). A s a dimeric l igand, fibronectin can induce receptor clustering by binding two or more integrins. T h e s e integrin-FN c lusters create a relatively high local concentrat ion of f ibronect in at the cel l su r face thus he lp ing to promote F N self assoc ia t i on and FN-cy toske le ta l interactions. Col laborat ions between the f ibronectin matrix and growth factors can regulate cell differentiation and migration (Schwarzbauer and Sech ler , 1999). T h e most abundant g lycoprote ins of basemen t m e m b r a n e s are the laminins (Figure 3). Lamin ins have been impl icated in many b io log ica l p r o c e s s e s , including cel l adhes ion , migration, and differentiation (Hay, 1991). Lamin ins are constructed from three genetical ly distinct polypept ides, the a , (3, and , y cha ins (Aumail ley and Gay raud , 1999). After assembly of the three distinct polypeptides the final mass of the laminin molecule can reach one million Daltons (Hay, 1991). Laminin has a number of binding sites including, one for type IV co l lagen, one for 10 Figure 2 I T a) CD « i B - C D CO CO C O T D " D O c o O S o> " D ^ C C ~ «) c TO * i CD • C C D ) c - o o g 5 ^ ^ . 2 > to £ a <D ~ • - > ^ CO a>^  §-E c ° 3 o •— CO CD "D • -^- C CO c — E CD CO CO CO 5 ' c 0 D ) — i C J CD CD in Q.co CD CD . E E -CD CO CD c CO ^ CD W o £ o5 E t ; ^ CO o'c«Oco .— 3 CD £ -Q Q. . 2 3 D>cD +S -C O C 0 p CD Q. LU C L ^ — CO o ^ | E E c o o 11 Figure 3 Heparin Binding a-dystroglycan Binding T h e structure lamin in of is deta i led above . In laminin 5 the g l o b u l a r d o m a i n at the C - t e r m i n u s o f the m o l e c u l e b i n d s to sur face-assoc ia ted integrins to help facil itate h e m i d e s m o s o m e assoc ia t ion . 12 heparan sulphate, one for entactin, and two or more to laminin binding proteins on the cell surface (Alberts et al., 1994). E C M R e m o d e l l i n g P e r h a p s most intriguing are the dynamics involved in E C M remodel l ing. A somewhat static three d imensional descr ipt ion of E C M structure only begins to descr ibe the levels of complexi ty involved. Caut ion must be obse rved when thinking about the E C M b e c a u s e most often we are only looking at a smal l vo lume within the extracel lular space . The E C M componen ts act in concert to regulate many p rocesses such as cellular differentiation, morphogenes is , motility, and wound heal ing. The E C M can affect cell behaviour in two main ways . One mechan ism is to directly regulate cell behaviour through c e l l - E C M interactions v ia ei ther growth factor modulat ion of cel lu lar r e s p o n s e s or receptor med ia ted s igna l ing. The s e c o n d m e c h a n i s m is through the harbour ing of factors that regulate growth, surv ival , and differentiation (Streuli , 1999). Th is mechan i sm exploi ts the use of remodel ing e n z y m e s to control the re lease of matr ix-bound growth factors that subsequent ly control cel lular differentiation. Indeed, a large number of enzymes are involved in E C M remodel ing, including the t issue ser ine p ro teases and the large family of matrix meta l lopro te inases ( M M P s ) . T h e s e e n z y m e s act as broad spect rum pro teases for major E C M degradat ion events that occur during t issue remodel ing (Streuli, 1999). 13 Matrix Meta l loprote inases P h y s i o l o g y a n d Patho logy The matrix meta l loprote inases ( M M P s ) are a d iverse group of z inc contain ing e n d o p e p t i d a s e s that are c a p a b l e of deg rad ing many c o m p o n e n t s of the extracellular matrix. M M P s have been implicated in a wide range of physiological p rocesses such as : normal wound heal ing, bone remodel ing, uterine resorption, t rophoblast implantat ion and a n g i o g e n e s i s ( N a g a s e and W o e s s n e r , 1999). Alternatively, M M P s have a lso received a t remendous amount of work devoted towards elucidat ing their involvement in many pathological p r o c e s s e s such as chronic inf lammatory and degenerat ive d i s e a s e s a s well a s tumor metastas is (Pa rks and M e c h a m , 1998). Per turbat ions in co l l agen turnover and E C M degradat ion are signature features of periodontit is, arthritis, renal d i seases , and the metastatic spread of cancer cel ls (Overall and Sodek , 1987, Bresn ihan et al., 1999, Fon tana and De lmas , 2000, and Const igan et al., 1995). The proteolytic activit ies of M M P s are precisely control led during activation and are inhibited by e n d o g e n o u s inhibitors such a s the a-macrog lobu l ins and t issue inhibitors of metal loproteinases (TIMPs) (Nagase and Woessner , 1999). M M P Fami l ies a n d D o m a i n Structure To date 23 different human M M P s have been identified and or c loned. The M M P fami ly s h a r e s s igni f icant s e q u e n c e homology and a c o m m o n mul t i -domain organizat ion (Figure 4). All of the M M P s are synthes ized as prepropept ides with an approximately 20-amino ac id s ignal peptide to direct intracellular trafficking. 14 Figure 4 Matrilysin 1 and 2 (MMP-7 and -26) Zn 2 + Collagenases (MMP-1, -8, -13, and -18), Stromelysins (MMP-3 and 10), Metalloelastase (MMP-12), Enamelysin (MMP-20), and MMP-19 Zn2+ M Stromelysin 3 (MMP-11) M | Z n 2 + M Membrane-Type MMPs (MMP-14, -15, -16, -17, -24 and -25) • I Zn2+ M Gelatinases (MIVP-2 and -9) Signal Peptide o Fibronectin Type II-like Module Pro-Domain O-linked Oligosaccharaide Domain Catalytic Domain • Furin Recognition Linker Region Hemopexin-like Domain Transmembrane Domain Cytoplasmic Tail S c h e m a t i c representa t ion of the M M P d o m a i n s t ructure. I l lustrates the i nc reas ing complex i t y of the M M P fami ly through gene t i c recomb ina t ion of ind iv idual d o m a i n s . 1 5 Excep t for the fur in-act ivated membrane- type M M P s (MT1-6 - M M P s ) and s t romelys in-3, M M P s are secre ted as z y m o g e n s being mainta ined in a latent state by an ~ 80 res idue N-terminal p ro-domain (Bode et al., 1999). The propept ide domain conta ins the uniquely conse rved s e q u e n c e P R C G ( V / N ) P D whe reby the abso lu te l y c o n s e r v e d C y s c h e l a t e s the cata ly t ic z i nc thus mainta in ing e n z y m e latency (this C y s has been referred to a s the "cysteine switch") (Nagase and Woessne r , 1999). Immediately fol lowing the propeptide is an approximately 170 amino acid catalytic domain , which conta ins the catalytic z inc, a structural z inc and two to three structural ca lc ium ions (Bode era/., 1999). T h e act ive si te, l ike all prote inases, performs the two-fold funct ion of binding substrate and catalyzing peptide bond hydrolysis (Overal l , 2000). The eff iciency of these act ions def ines the substrate specificity of the proteinase by determining enzymic activity towards particular substrates. The catalytic domain contains the highly c o n s e r v e d z inc-b ind ing s e q u e n c e H E X X H X X G X X H and a c o n s e r v e d meth ion ine wh ich forms a un ique 'Met- turn ' structure charac ter is t i c of the metzincin superfami ly (Bode era/. , 1999). The catalytic activity of the M M P s is contr ibuted by a water molecule that is polar ised by a conse rved glutamic ac id res idue posi t ioned at the bottom of the act ive site cleft (Crabbe et al., 1994). Except for the matri lysin's ( M M P - 7 and M M P - 2 6 ) all of the M M P s are secre ted with a C- terminal hemopex in doma in . T h e C-terminal hemopex in domain is approx imate ly 210 amino ac ids long and has an el l ipsoidal d isk shape . The hemopex in domain has an overal l four b laded (3-propeller structure of pseudo-fourfold structure, with each blade consist ing of four anti-paral lel (3-strands and an p-helix (Bode etal., 1999). Unequivocal ev idence implicating the C-domain in 16 substrate recognit ion and subsequent c leavage has been demonstrated through the C-domain interactions of the co l lagenases (MMP-1 , -8 , and -13), where the C -doma in has been shown to be required for c l eavage of nat ive triple hel ical co l lagen (Murphy and Knauper , 1997) and in ge lat inase A for the c leavage of chemok ines (McQuibban et al., 2000). Ge la t inase A and B ( M M P - 2 and M M P - 9 , respect ively) are unique in that they have three cont iguous f ibronectin type II repeats inserted into their catalytic domains . Th is domain exhibited high affinity binding to gelatin and was initially termed the gelatin binding domain (Banyai and Patthy, 1991). It was later found by Stef fensen et al. that the f ibronectin type II modu les of gelat inase A a lso bound avidly to type I co l lagen and thus lead the authors to term this domain as the col lagen binding domain (CBD) (Steffensen et al. 1995). Adopt ing the terminology of Stef fensen, in this thesis the recombinant form of the C B D will be termed r C B D 1 2 3 , with the r des ignat ion referring to recombinant protein and 123 referring to a construct contain ing modu les 1, 2, and 3. In the gelat inase A molecule, the domain is s imply referred to a s the C B D . A s will be descr ibed in greater detail later, the function of the C B D is to donate substrate binding exos i tes to help confer subst ra te speci f ic i ty to the gelat inases (Bode et al., 1999) and to possib ly function in triple he l icase activity (Overal l , 2000). Gelat inase A Gela t i nase A (also known as M M P - 2 , 7 2 - k D a ge la t inase and 7 2 - k D a type IV co l l agenase ; E C 3.4.24.24) w a s first descr ibed by Liotta et al. in 1979 as an enzyme secre ted by a metastat ic murine tumor that w a s capab le of degrading 17 soluble type IV co l lagen into 1/4 N-terminal and 3/4 C-terminal f ragments (Liotta, et al., 1979). Approximately ten years later a c D N A exp ressed c lone of the type IV co l lagenase was constructed and the expressed product was found to have a molecular weight of 72 k D a (Coll ier et al., 1988). Further character isat ion of the exp ressed protein lead researchers to c lassi fy the 7 2 - k D a co l l agenase into the matrixin family of prote inases and subcategor ised it a s the s e c o n d M M P to be d iscovered. Ge la t inase A is unique in that it has a ubiquitous t issue distribution, is uncharacter ist ical ly act ivated at the cel l sur face (Overal l and Sodek , 1990), and it is constitutively expressed by many cell types, including gingival f ibroblasts (Parks and M e c h a m , 1998). Regulation R e g u l a t i o n of g e l a t i n a s e A o c c u r s both t r ansc r i p t i ona l l y a n d post -transcriptionally. Unl ike other M M P s , transcription of ge lat inase A is not readily induced by agents such as 12-0- tet radecanoylphorbol-13-acetate (TPA) (Overall e r a / . , 1991) or interleukin 1 a (IL-1 a), both of which have been shown to induce transcription at gene and promoter e lements. The promoter regions of gelat inase A show marked dif ferences from the other M M P s in that this region lacks the well known transcription activator sequences for activator protein-1 and polyomavirus enhancer A-b inding protein 3. Th is region a lso contains a unique noncanonica l T A T A box, as well as a putative enhancer element located -223 to -422 relative to the translational start site (Parks and M e c h a m , 1998). Unl ike other M M P s , the gelat inase A promoter a lso lacks upstream TGF-(3 inhibitory e lements. Overal l et al. (1991) found that gelat inase A transcription was not s u p p r e s s e d but slightly 18 upregu la ted by TGF- (3 . T h e p r e s e n c e of an act ivator prote in-2 e lement contributes to cell-type specif ic express ion of both gelat inases. Cel lu lar regulation of progelat inase A activation is very important in the initiation of E C M turnover. Post- t ranscr ipt ional ly , ge la t inase A regulat ion occu rs v ia increased m R N A stability. Ev idence for this was found by Overa l l et al. (1991) where it w a s d i s c o v e r e d that addi t ion of T G F - P , wh ich normal ly reduces proteolytic activity through reduced prote inase synthes is and inc reased T I M P express ion , increased gelat inase A secret ion by human f ibroblasts and rat bone cel ls. After addit ion of T G F - p it was found that total cel lular gelat inase A m R N A was found to increase approximately 1.5- to 2.2-fold. Th is increase in gelat inase A m R N A was attributed to a concomitant increase in ge la t inase A stability (t 1 / 2 i nc rease from 46 to 150 h), hence providing ev idence for post- transcr ipt ional regulation of gelat inase A express ion. Wh i le most act ive M M P s are inhibited by T I M P binding to their act ive s i tes, extracel lu lar p roge la t inase A is often found spec i f ica l ly c o m p l e x e d with the endogenous inhibitor T I M P - 2 . Si te-di rected and delet ion mutant s tudies have loca l ised the T I M P - 2 binding site on M M P - 2 to charged res idues within the C -terminal domain (Overal l et al., 1999 a,b, Butler era/., 1999). It has been found that negative charges on the C-terminal tail of T IMP-2 (and possibly T IMP-4) help facil i tate T I M P binding to the C -doma in of proge la t inase A , thus conferr ing binding specificity to progelat inase A (Bigg era/., 1997, 2001). The progelat inase A / T I M P - 2 assoc ia t ion regulates cel l -sur face activation of progelat inase A v ia a membrane -assoc ia ted act ivator membrane- type 1 - M M P ( M M P - 1 4 ) , that is C -19 terminally bound to the cell membrane (Sato et al., 1996). Th is work s temmed from the observat ion that the lectin concanava l in A induced the endogenous activation of gelat inase A (Overal l and Sodek , 1990). Through interactions with the N- te rm inus of T I M P - 2 , m e m b r a n e - b o u n d M T 1 - M M P l o c a l i s e s the progelat inase A / T I M P - 2 complex at the cell sur face. The tr imolecular complex of proge la t inase A / T I M P - 2 / M T 1 - M M P a l lows for a s e c o n d act ive M T 1 - M M P to proteolytically c leave the N-terminal propeptide of progelat inase A at A s n 37-Leu 38 to give rise to a 64 k D a intermediate that is autoproteolytically c leaved to fully act ive gelat inase A (Woessner and Nagase , 2000) (Figure 5). Upon addit ion of r C B D to concanaval in- t reated f ibroblasts there is an increase in the amount of ge la t inase A act ivat ion that w a s attr ibuted to i nc reases in p roge la t inase A shuffl ing into the cellular activation pathway v ia d isp lacement of progelat inase A b o u n d to the ce l l s u r f a c e th rough py in teg r i n / co l l agen /CBD interact ions (Steffensen et al., 1998) (Figure 5). Substrate Specificities Gela t inase A is capab le of c leaving a number of peptide bonds such as G ly -Va l , G l y - L e u , G ly - l l e , G l y - G l u , G l y - A s n , and G l y - S e r in gelat in to p roduce smal l pept ides. Bes ides gelatin and soluble types I, IV and V co l lagen, gelat inase A can a lso c leave type VII and X co l lagen, and elast in. Interesting to note is that both gelat inase A and B can only c leave pepsin-solubi l ised type IV co l lagen, both enzymes being unable to c leave native type IV co l lagen. T h e s e findings s e e m to raise s o m e doubt about whether the gelat inases are true "type-IV co l lagenases" in vivo. Indeed, whi le the first c l eavage site of so lub le type IV co l lagen by 20 Figure 5 Schemat ic representation of gelt inase A activation. First, an MT1 - M M P binds to the N terminus of T IMP-2 (1). The free -C O O H terminus of the T I M P is avai lable to interact with the C-domain of progelat inase A (2). These interactions help to position progelat in-ase A at the cell surface for activation by a second active M T 1 - M M P (3). Upon M T 1 - M M P dependent gelat inase A c leavage, active gelat inase A is re leased from the cell surface (4). Upon addition of r C B D 1 2 3 , pericellularly-bound gelat inase A is re leased from type I co l lagen Stef fensen etal., (1998). 21 gelat inase A occurs 1/4 of the d is tance from the amino terminus, the c leavage site within full-length native type IV co l lagen remains to be determined (Parks and M e c h a m , 1998). In addit ion to the ability of gelat inase A to degrade native E C M structures such as f ibronect in and vi t ronect in, ge la t inase A has been s h o w n to p o s s e s s |3-sec re tase activity, be ing ab le to c l eave the p -sec re tase site within amylo id protein precursor 695. Gela t inase A is a lso capable of c leaving laminin-5, which, through interactions with integrins, is essent ia l for adhes ion of epithelial cel ls to the basement membrane. C l e a v a g e of laminin-5 l iberates a pro-motility cryptic site that induces ce l ls to migrate (Parks and M e c h a m , 1998). Ge la t i nase A funct ions most of the t ime to modulate the structure of the E C M therefore regulating p rocesses such as cell migration, development, and morphogenes is . Metastas is Three essent ia l events have been def ined in tumour migrat ion: (1) tumour cell adhes ion to E C M componen ts , (2) E C M degradat ion by proteolys is , and (3) tumour cel l migrat ion into the degraded a rea (Kleiner and Ste t le r -S tevenson, 1999) . Ben ign epi thel ia l tumours are a lways cha rac te r i sed by an intact basemen t m e m b r a n e that encapsu la tes the neop las t ic ep i the l ium from the connect ive t issue stroma, whereas malignant epithelial tumors have an ill-defined basemen t m e m b r a n e that forms an incomplete barr ier wh ich faci l i tates the sp read neoplast ic ce l ls (Parks and M e c h a m , 1998). Th is ability of metastatic tumour cel ls to initially t raverse the encapsulat ing basement membrane and to subsequen t l y c r o s s the subepi the l ia l basemen t m e m b r a n e of the vascu la r 22 sys tem al lows the tumour cel ls to indiscriminately circulate throughout the body. Indeed, the ability of tumour ce l ls to degrade the type IV co l lagen conta ined within the basement membrane is paramount in the determinat ion of metastatic potential. Ev idence for the enha need express ion of ge la t inase A in human tumours is derived from experimental studies correlating gelat inase A express ion to tumour grade (Dav ies et al., 1993) . Immunocytochemica l and in situ hybridizat ion s tudies have shown that there is a local inc rease in ge la t inase A express ion surrounding many human tumours (Rooprai et al., 1998). Accord ing ly , benign prol i ferat ive d i so rde rs of t hese t i s s u e s usua l l y s h o w a low or negat ive immunoreactivi ty to ant ibodies directed against ge lat inase A (Still et al., 2000). Upon tumour cel l migration, gelat inase A has been shown to play an important role in ang iogenes is (Kleiner and Stet ler -Stevenson, 1999). T h e s e f indings have been substant iated by the use of a gelat inase A knockout mouse where it was found that there w a s both a dec rease in tumour metastas is and ang iogenes is , thus illustrating the in vivo potential of gelat inase A (Itoh e ra / . , 1998). However , there were very few phenotypic dif ferences in the knockout mouse , which is likely a reflection of the omnivorous potential of the M M P s . Gelat inase A-Mediated E C M R e m o d e l l i n g Regulat ion of extracellular matrix degradat ion by gelat inase A is governed by the net ba lance of progelat inase A secret ion, express ion , act ivat ion, and local T I M P concentrat ion. The fact that increased gelat inase A secret ion around tumours precedes the acquisit ion of an invasive phenotype suggests that the activation of 23 gelat inase A makes a signif icant contribution to an invas ive phenotype (Parks and M e c h a m , 1998). To support this idea, N o m u r a et al. (1996) used gelatin z y m o g r a p h y to s h o w that act ive forms of ge la t inase A and B were more frequently detected in human gastric carc inoma t issue, and that the activation of the zymogen form of gelat inase A , but not that of ge lat inase B, correlated well with the degree of local invas ion and lymphat ic permeat ion . To further this notion, regulation of gelat inase A activity by T I M P - 2 can serve as a means to regulate the extracel lular activity of gelat inase A . A s descr ibed before, T I M P - 2 select ively b inds to progelat inase A and inhibits ge la t inase A . Consequent l y , T I M P - 2 can supp ress invas ion, metastas is , neovascu lar isa t ion , and growth of some rodent and human tumours (Kleiner and Stet ler -Stevenson, 1999). Exosites O n e of the keys for pro teases to exhibit a high speci f ic activity ( k c a / K M ) is the ability to bind efficiently to substrate (small K M ) . In order to inc rease binding avidity, enzymes contain discrete binding subsi tes, termed "exosi tes", that are far removed from the catalyt ic site (Overa l l , 2000) . Exos i t es are found within modules or domains of unrelated proteins that have, through the p rocess of gene t ransposi t ion, inserted themse lves into distinct po lypept ides, thus conferr ing a protein greater evolut ionary potent ial by funct ioning to i nc rease subst ra te recognit ion and therefore the spectrum of the prote inase. Indeed, gelat inase A has evo lved to contain severa l exos i tes that function in substrate recognit ion. Recent ly it was d iscovered through the use, of a yeast-2-hybrid sc reen that the C -domain of gelat inase A interacts with the chemok ine monocyte chemoattractant 24 protein-3 ( M C P - 3 ) subsequent ly facilitating gelat inase A-dependent c leavage of M C P - 3 to M C P - 3 (5-76) that antagonizes chemotax is (McQu ibban et al., 2000). Ge la t inase A and B are a lso structurally separated from the M M P family by the substrate binding exos i tes that are located on the three cont iguous f ibronectin type II modules that are inserted into their catalytic domains (Banyai and Patthy, 1991) Fibronectin Type II Modules Ev idence for the involvement of the fibronectin type II modules of gelat inase A in the binding of gelatin was first descr ibed by Banya i and Patthy ( F E B S lett. 1991). Th is paper demonstrated for the first time that a recombinant protein consist ing of only the three fibronectin type II repeats from gelat inase A could bind gelatin. Similar ly, it was shown by Col l ier et al. that the gelatin binding site of gelat inase B res ides within its f ibronectin type II repeats (Coll ier et al., 1992). Banya i et al. subsequent ly purif ied wi ld-type f ibronectin type II modu les , delet ion mutants lacking either the first, second , or third module (DELpgalco l l 23 , DELpga lco l l 13, and DEL(3galco l l12, respect ive ly) , and the s ing le modu les (DELpga l co l l 1, DELpga l co l l 2, and DELpga lco l l 3). F rom these studies, they found that each f ibronect in type II modu le w a s c a p a b l e of sa turab le b ind ing to a gelat in-S e p h a r o s e 4 B co lumn . Sca t cha rd ana lys i s of b inding a s s a y s s h o w e d that ge la t in-Sepharose p o s s e s s e d a range of binding si tes with different affinities for the recombinant prote ins. C o m p a r i s o n of assoc ia t i on cons tan ts i l lustrated marked d i f ferences in recombinant protein affinity for gelat in . Recomb inan t proteins containing two of the fibronectin type II modules bound more avidly than 25 any of the individual modules. Corresponding ly , wild-type domain competed off both of the single deletion mutants, thus displaying the highest affinity for gelatin. Th i s s u g g e s t e d that the three f ibronect in type II modu les of ge la t inase A cooperate in gelatin binding. The important quest ion raised was whether the fibronectin type II modules acted singly to bind gelatin, i.e. single binding sites on each of the modu les , or, do the modu les spatial ly orient themse lves so as to form a cooperat ive binding cleft. Banya i et al. add ressed this quest ion through competi t ion exper iments where it w a s found that pa i rs of non-over lapp ing f ragments , s u c h a s DEL(3galcoll 13/DEL(3galcoll 2 or DEL(3galcoll 23 /DELpga lco l l 1, were able to compete each other off gelat in. Th is sugges ted to the authors that different type II units may bind to the s a m e sites on gelatin and that different f ibronectin type II modules do not have unique binding si tes and that there w a s s o m e promiscui ty in their binding to gelatin (Banyai e ra / . , 1994). Indeed, Stef fensen et al. (1995) showed through the use of C N B r f ragments of gelatin that the C B D bound to multiple sites on the cc-chains. F ibronect in T y p e II M o d u l e Deletion Mutant S tud ies To further a s s e s s the role of the fibronectin type II domain Murphy et al. (1994) looked at this from the perspect ive of full-length gelat inase A . The propert ies of the ge la t inase A delet ion mutant A V 1 9 1 . Q 3 6 4 ( represent ing remova l of all the fibronectin type II repeats) were compared with those of full-length gelat inase A . Firstly, activation of full-length gelat inase A and the deletion mutant A V 1 9 1 . Q 3 6 4 both occurred through the single canon ica l intermediate s tage. T h e activity of both 26 proteins towards the synthet ic substrate M c a P L A N v a D p a A R w a s found to be very simi lar. Th is i l lustrates that comple te removal of the catalyt ic doma in -inserted f ibronectin type II repeats does not reduce catalyt ic activity towards an active si te-binding substrate. The deletion mutant A V 1 9 1 . Q 3 6 4 w a s shown to have only 5 0 % of the activity of full-length gelat inase A towards (3-casein, only 1 0 % of the activity of full-length gelat inase A towards gelatin (Murphy e ra / . , 1994), and a total l oss of e las t ino ly t ic act ivi ty (Sh ip ley et al., 1996) . E n z y m e - l i n k e d immunosorbent a s s a y s decis ively proved that the binding of gelat inase A to type I co l lagen could be local ised to the fibronectin type II repeats s ince the deletion mutant was unable to bind to native type I co l lagen coated wel ls whereas full-length gelat inase A avidly bound type I co l lagen (Murphy e ra / . , 1994). Through delet ion mutant and fibronectin competi t ion a s s a y s Al lan et al. (1994) provided further ev idence implicating the role of the fibronectin type II domains in binding type I and IV co l lagen. It was shown that co l lagenase 1 and stromelysin 1 could be s imul taneously bound to gelat inase A-bound type I co l lagen. Therefore the authors conc luded that ge la t inase A has different b inding speci f ic i t ies than stromelysin 1 and co l lagenase 1, both of which bind co l lagen fibrils through their C - te rm ina l h e m o p e x i n d o m a i n s (A l lan et al. 1994) . T h e s e da ta ful ly substant iated the role of the fibronectin type II repeats in affording gelat inase A the ability to bind gelatin. 27 F ibronect in T y p e II M o d u l e Character isat ion Character isat ion of r C B D 1 2 3 substrate binding showed that the fibronectin type II domain w a s capab le of binding denatured type I co l lagen in the micromolar range (Stef fensen e r a / . 1995). It was a lso shown that r C B D 1 2 3 avidly bound elast in, hepar in, native type I, V , and X co l lagens denatured type II, IV, V , and X co l lagens, but neither native type IV and V co l lagen, f ibronect in, reconsti tuted basemen t m e m b r a n e , laminin nor S P A R C , all of wh ich are subs t ra tes of ge lat inase A (Stef fensen et al. 1995. , Overa l l et al. 1997. , and A b b e y et al., 2001). Native type I co l lagen, which is not efficiently degraded by gelat inase A , compe ted with gelatin for a shared binding site on r C B D 1 2 3 . r C B D 1 2 3 a lso d isp laced full-length gelat inase A bound to native type I co l lagen (Steffensen et al., 1995, 1998). If e a c h of the three modu les con ta ined a s ing le co l lagen binding site, it was plausible to suggest that r C B D 1 2 3 could bind more than one molecu le of co l lagen . To test this, Ste f fensen et al. incubated [14C]-glycine-label led co l lagen in a molar e x c e s s of r C B D 1 2 3 . C o m p l e x e s of r C B D 1 2 3 bound to the rad io labe led l igand were then found to bind unlabel led native or denatured type I co l lagen coated on microwell plates. Th is demonstrated for the first time that r C B D 1 2 3 p o s s e s s e d at least two co l lagen-b ind ing s i tes that cou ld be s imul taneously occup ied by two col lagen molecules. The binding of r C B D 1 2 3 to a1(l) co l lagen cyanogen bromide fragments shows that there are multiple binding s i tes con ta ined within the a1 (I) co l lagen f ragment (Stef fensen et al., 1998). T h e s e f indings indicated that there w a s a potential for the f ibronectin type II 28 modu les to funct ion in faci l i tat ing e n z y m e loca l isat ion in connec t i ve t issue matricies. Fibronect in T y p e II M o d u l e Structural Determinat ion Recent work on the f ibronectin type II. repeats of ge la t inase A has focused on determining: 1) the overal l structure of the three modu les and 2) the res idues involved in the binding cleft and/or site(s). Th is body of work w a s b a s e d on previous studies that were used to model f ibronectin type II modu les to proteins that share high primary sequence homology. Fibronect in type II modu les are a c o m m o n structural motif containing two disulphide br idges l inking C y s residue pairs of the first and third C y s residues and the second and fourth C y s residues. A distant homology has been d iscovered between fibronectin type II modules and the three-disulphide kringle units found in thrombolytic and fibrinolytic enzymes . Th i s s u g g e s t s that these two types of modu les evo l ved f rom a c o m m o n ances te ra l protein. Interestingly, with the except ion of the ge la t inases , the f ibronectin type II modu les appear only as s ingle modu les or in pairs. T w o fibronectin type II modules are found in fibronectin (Owens et al., 1986), bovine seminal fluid protein-109 (Esch e r a / . , 1983) and bovine semina l p lasma protein B S P - 3 0 (Calvet te , J . J . , 1996) - both co l l agen b ind ing prote ins. A s ing le f ibronect in type II modu le is found in the m a n n o s e m a c r o p h a g e receptor precursor (Taylor e ra / . , 1990), phosphol ipase A 2 receptor (Ishizaki etal., 1994), m a n n o s e 6 -phospha te receptor (Morgan etal., 1987) , hepa tocy te growth activator (M iyazawa et al. 1993), and blood coagulat ion factor XII (Hageman factor)(McMullen and Fuj ikawa, 1985) (Table 3). 29 Table 3: Primary Sequence Alignment of Selected Fibronectin Type ll-Like Repeats - o <o i i i i i i i i - o LO I I I I I • I I I - o I I I I I I I I - o cn I i I i i i i i - o CM I I I I I I I I - o fa X P-i CM O U U fa fa fa W O O >< fa >* W W W •a cn o Q Q Q a < « H Q " " EH • Pi w a w • Q W P-I P-I P-I u u o fa fa o o o >< & fa W W PH w Pi p Q Q Q Pi p p p p Pi o • a En P CM CM u u fa fa o o fa EE p w p p Pi CO p p u u & & Pi p o o p p CO O Pi Pi o o P w EH CO EH EH o u <! EH CO CO >H ^ CO W a w o o a p fa fa H EH fa fa CM CM fa fa CX > U U P-i P W W o o p < i< CO O O P O Pi ' ' CO CM p fa fa fa p p EH CO CO CO CO O o o O X H H > p EH EH EH H U U O u P EH P EH fa fa fa P K >H >n CO Pi O O S O O fa >* 1_| {H 1 1 1 1 I-* fa fa 1 1 fa 1 1 fa CM CM CM P-I fa fa fa fa > > > > O o u u w fa < << u o PH W a Pi P-I • o X Pi co o o Pi w ffi co H EH o u w < ffi p P s C Pi Pi o X o o Pi fa fa CM CM fa fa x Pi o o CM CM w Pi o o p w < p p 5 p p P fa Pi O CO W fa Pi > EH > u u fa fa ffi fa X >H 5 CO a w a o s s £ fa a > fa fa CM CM fa fa s > o u CM CM EH W O O X P < H H H Q) CD ss EH EH 55 £ fa fa H CN m H OJ ro p P PQ P P P p q p q p ra m c q O O u u U o J J J 1 1 1 < C < < m m p q rH rH rH TH I H I H a ) CD a ) a ) OJ CD o o o o o o A & si si si P P p ^ , > u 1 1 H CD CM rH EH Pi U < cn CO rH u CD < < CN Pi CM CM CM m <; u fa fa fa CM CM p CO CO fa o p CM m ra X CM S CO • " D o co co — <D CD —> _ — E co __: CD •« -CD * " >> = lo =r o o >^ .y -c: .E o o i l l _ T to T 3 CD CD CD co co CD to c o _ o °i5 c-c ^ i— — CD M - C0 C O H -c c ° C D O CD O c o £ CL co CD X o a) to ^ c o o CC • CD O CD CO : TD CD CD CO c o to c o -CD 6 S c c 1 CD CD "5 3 0 Cons tan t ine et al. (1992) first determined the solut ion structure and l igand binding propert ies of P D C - 1 0 9 domain b by N M R spec t roscopy . Th is paper loose ly out l ined the core structure of P D C - 1 0 9 b a s be ing fo rmed of two antiparallel (3 sheets and two irregular loop regions at the top and bottom of the molecu le . T h e cavity formed by the backbone is largely fi l led by a c luster of aromat ic res idues, including Tyr 11, Tyr 30, Tyr 37, Trp 42 , and Tyr 45 , all of which are so lvent -access ib le and define an exposed hydrophobic sur face (note: for cons is tency I will refer to the residue number a s being counted from the beginning of each module as illustrated in Tab le 3). Th is was the first descript ion of a putative hydrophobic binding sur face. The res idues P h e 9, Thr 20, and lie 22 form the foundat ion for the aromat ic c luster. C o n c l u s i o n s regarding the b ind ing si te st ructure were der ived f rom l i gand- induced r e s o n a n c e shifts observed v ia N M R . The Trp 30 H6 and H7 resonances were most sensi t ive to 3 ,3-d imethy lbuty lamine (an N-terminal leuc ine mimic) . T h i s res idue w a s therefore hypothes ized to be either in direct contact with the l igand or indirectly affected by a change in orientation of the Tyr 11 and Tyr 37 rings, both being hypothes ized to be likely disturbed by direct l igand contact. The s ide cha ins of Thr 20, lie 22 , and A l a 40 were hypothes ized to surround the putative binding sur face and therefore any disturbances exper ienced by these res idues are likely to be secondary effects. Banya i et al. (1996) were the first group to model the structure and doma in -domain interactions of the fibronectin type II modules from gelat inase A to bovine semina l fluid protein P D C - 1 0 9 . F rom circular d ichro ism (CD) spec t ra it was determined that the f ibronect in type II doma in of ge la t inase A cons is ted of 31 approximately 3 2 % (3-sheet and 1 9 % (3-turn with no detectable a helix. (3galcoll 1, pgalcoll 2, and pgalcoll 3 were est imated to contain 3 0 % , 3 0 % , and 3 1 % (3-sheet, and , 2 8 % , 2 7 % , and 2 3 % P-turn s t ruc tures, respec t ive ly . T h e es t imated secondary structures of the fibronectin type II repeats were, as expec ted , very similar to that of the second type II domain of bovine semina l fluid protein P D C -109 doma in b. C D spec t roscopy , differential scann ing microcalor imetry, and urea- induced denaturat ion of recombinant Pgalcoll 23 and Pgalcoll 123 were used to s h o w that the thermal stabi l i ty of the s e c o n d type II modu le is significantly increased by its interactions with the first and third modules. It was postulated from these results, that the tandem modules permit little flexibility in their relative orientat ion at phys io log ica l temperatures. B a n y a i et al. further p roposed that all three f ibronectin type II modu les contr ibute to binding of a single substrate molecule. It must be noted that this hypothesis differed from the observat ion of Stef fensen et al. (1995), where it was shown that r C B D 1 2 3 was capab le of binding to more than one molecule of co l lagen. Banya i et al. further p roposed a mode l whereby the three f ibronect in type II m o d u l e s form an extension of the substrate binding cleft of gelat inase A (Banyai etal., 1996). Until 1997, the solution structure of the first human fibronectin type II module had not yet been determined. Pickford et al. determined the structure of the first type II module from human fibronectin by 2 D N M R spec t roscopy and found that the structure was very similar to that of the shorter f ibronectin type II module from P D C - 1 0 9 domain b. In both the first f ibronectin type II module and P D C - 1 0 9 b , the two double-s t randed p sheets are oriented approximately perpendicular to 32 each other, with the cleft between them being occup ied by the s ide cha ins of the invariant and highly conserved aromatic residues. The disulphide bonds were in c lose proximity to one another, and were located on the opposi te face of the second aromatic cluster |3 sheet. B a s e d upon primary s e q u e n c e homology, the invariant P h e 18 and Trp 39 (corresponding to P h e 9 and Trp 30 of P D C - 1 0 9 b ) were complete ly buried in the f ibronectin type II structure and const i tuted the core of the molecule. The highly conserved residues Leu 13, Tyr 20, P h e 25 , Tyr 46, Tyr 52, and Phe 54 were packed around this core, with the s idecha ins of the four res idues Tyr 20, Tyr 46, Tyr 52 , and P h e 54 (corresponding to Tyr 11, Tyr 35 , Trp 4 3 , a n d Tyr 4 5 in P D C - 1 0 9 b ) c lus tered on one face of the modu le forming a so l ven t -exposed hydrophobic sur face. W h e n the authors c lose ly examined the N M R - d e r i v e d model the p resence of a poss ib le l igand-binding pocket in the hydrophobic surface was noted with the Trp 39 indole ring forming the pocket f loor (Pickford et al., 1997). Th is was the first descr ip t ion of the hydrophobic sur face as being a hydrophobic pocket and has lead to further work a imed at e luc idat ing the contact res idues . However , not just hydrophob ic interactions are involved in binding to gelatin. In the absence of chaotropes, both full-length and a 42 k D a fibronectin fragment that contains the f ibronectin type II repeats bound gelat in and cou ld be eluted from immob i l i sed gelat in by a reduction in p H . Th is sugges ted direct involvement of cha rged aspartate and glutamate residues in binding. Elution by the mild denaturant dimethylformamide at concentrat ions lower than that required for elution by guanid ine, suggested an important role for hydrogen bonding in substrate binding (Pickford e ra / . , 1997). 33 Sticht et al. (1998) cont inued work ing with f ibronect in type II doma ins and subsequent ly determined the solution structure of the g lycosy lated second type two module of f ibronectin. The averaged structure of the second fibronectin type II module was found to contain a short a helix and three antiparallel p sheets , all of which partly enc lose a c luster of aromat ic res idues . T h e topology of the s e c o n d f ibronect in type II module was found to be s imi lar to that of the first f ibronectin type II module . In both f ibronectin type II modu les , the d is tance between the N and C termini is short, but relative to the first f ibronectin type II modu le , the N terminus of the s e c o n d f ibronect in type II modu le is shif ted towards residues Gly 11 and C y s 14, thus forming a double-st randed antiparallel P-sheet (Sticht et al., 1998). A s is now becoming a general trend in the overal l structure of the f ibronect in type II and type II modu les , many of the highly conse rved aromat ic res idues in the f ibronectin type II modu le form a solvent-exposed hydrophobic surface. A s with the first f ibronectin type II module, it was found that there was a depress ion within the hydrophobic sur face of the second type II module, at the bottom of which is the invariant t ryptophan residue (Trp 39). St icht et al. cont inued to model the first and s e c o n d f ibronect in type II modu les together. After sc reen ing 300 ,000 r a n d o m i z e d modu le pai rs for favourable non-covalent interactions there w a s no s ingle lowest energy model obta ined. A l though it w a s not poss ib le to unambiguous ly ascer ta in a s ingle model led fibronectin type II module pair, one consistent feature regarding gelatin binding a rose during model l ing procedures. First, in all ana l yzed structures a minimum d is tance of 14.2 A is observed between the core Trp res idues of the two f ibronectin type II modules, which, accord ing to the authors, sugges ts that 34 l igand-binding hydrophobic pockets form discrete binding si tes. Th is was in stark contrast to Banya i et al. who had very shortly before p roposed that the three fibronectin type II modules of gelat inase A acted together in forming an extended binding pocket (Banya i et al., 1996) . T h e notion that there are d iscre te hydrophobic binding si tes conta ined within f ibronectin type II modu les helps to support the f indings by Stef fensen etal., where it was shown that r C B D 1 2 3 was c a p a b l e of b ind ing two sepa ra te m o l e c u l e s of c o l l a g e n s imu l t aneous l y (Steffensen etal., 1998). Even with the flourish of activity dedicated to determining the solution structure of the first and second fibronectin type II modules, the determination of the solution structure of one or all of the f ibronect in type II modu les from ge la t inase A remained to be comple ted. In 1999, Br iknarova et al. determined, for the first t ime, the solut ion structure of the s e c o n d f ibronect in type II modu le from ge la t inase A . It w a s found that the module cons is ted of two short doub le-s t randed ant iparal lel (3 shee ts ar ranged approx imate ly perpend icu lar to each other and three large irregular loops. The first (3 sheet and the loops were a r ranged a round the s e c o n d (3 sheet , forming a large cavi ty f i l led with the aromatic s idecha ins of Phe 16, Phe 18, Phe 20, Tyr 25 , Tyr 37, Trp 39, Tyr 46, Tyr 52, and P h e 54. Phe 18 was totally buried within the molecule and Phe 20, Tyr 37, Trp 39, Tyr 46, Tyr 52 , and P h e 54 formed an ex tended hydrophobic sur face, whereas P h e 16 faced the other s ide of the molecu le . The disulphide bridge and both the N and C termini were located at the back of the (3 sheet, on the other s ide of the molecule (Briknarova etal., 1999). 35 T o a s s a y for c h a n g e s in spect ra l charac te r upon subst ra te comp lexa t i on , Br iknarova et al. used a synthetic peptide with the c o n s e n s u s sequence ( P P G ) 6 and moni tored the chemical-shi f t c h a n g e s induced by ( P P G ) 6 binding v ia 2D chemica l shift corre lated spec t roscopy ( C O S Y ) and 1 H - 1 5 N heteronuclear shift cor re la ted s p e c t r o s c o p y exper imen ts . R e s i d u e s with the most per turbed backbone amide resonances were limited to a band that ran d iagonal ly ac ross the front s ide of the s e c o n d f ibronect in type II modu le and compr i sed the aromat ic c luster and surrounding a reas . B a c k b o n e amide r e s o n a n c e s from aromat ic res idues from the front s ide of the molecu le were found to be more sens i t ive to pept ide binding than r e s o n a n c e s of aromat ic res idues that are internal or face the back. It w a s conc luded that the bound pept ide ex tended along the aromatic cluster, interacting with P h e 20, Tyr 37, Trp 39, Tyr 46 , Tyr 52 , and P h e 54 on the front face of the modu le . C o m p a r i n g l igand binding res idues of the s e c o n d fibronectin type II module of ge lat inase A to that of the ear l ier reported interact ing res idues in P D C - 1 0 9 b and the first and s e c o n d fibronectin type II modules, cont inues to highlight the involvement of a spatial ly conse rved hydrophobic pocket with the c o n s e n s u s s e q u e n c e - ( F / Y ) 2 0 - W 3 9 - Y 4 6 -( Y / F ) 5 2 - F 5 4 ) . Until now, much of the work with the f ibronect in type II modu les has been focused on determining the overal l structure of the molecule. Indeed, structural s tudies have lead researchers to infer speci f ic res idue involvement in l igand binding, but little work has been afforded to elucidating the direct involvement of l igand- in teract ing res i dues . To rda i a n d Pat thy a d d r e s s e d this i s sue by performing site-directed mutagenesis of the second fibronectin type II module of 36 gela t inase A . T h e mutat ions were mode led from the p roposed hydrophobic binding sur face of P D C - 1 0 9 domain b and included res idues: R 1 9 L , R 1 9 L / R 3 8 L , Y 2 5 A , Y 3 7 A , R 3 8 L , Y 4 6 A , D 4 9 A , K 5 0 G , K 5 0 R , and Y 5 2 A of the s e c o n d fibronectin type II module of gelat inase A . During chracterisat ion of the mutants, it was found that the mutants could be divided into three categor ies based upon their structural and functional integrity. The first category, which conta ined the correctly refolded mutants R19L , R 3 8 L , K 5 0 G , K 5 0 R , and R 1 9 L / R 3 8 L , showed no impai red affinity for gelat in or al terat ion of their T m a n d denaturation midpoints. T h e s e data lead the authors to conc lude that these res idues are neither involved in substrate recognit ion nor structural integrity. The second group of mutants, Y 2 5 A , Y 4 6 A , D49A, and Y 5 2 A , were structurally compromised as determined by C D spec t roscopy and therefore were devo id of any gelatin binding (Tordai and Patthy, 1999). Indeed, the authors were unable to make any conc lus ions about the interact ions of these res idues with gelat in subst rate, however, the incorrect folding of these mutants is likely attributable to a loss of hydrophobic character and therefore suppor ted the impor tance of a structural hydrophobic core. The third category of mutants, compr is ing only Y 3 7 A , was correctly folded as determined by C D spec t roscopy and thermal and chemica l denaturat ion and s h o w e d severe l y dep le ted gelat in affinity. T h e authors conc luded that the only plausible reason for the reduction in gelatin affinity was that this residue is directly involved in gelatin recognit ion. Th is was the first direct ev idence implicating a contact site in the fibronectin type II module of gelat inase A . From the model led structure, this residue was p laced on the right hand rim of the hydrophobic pocket and has, for the first t ime, provided direct exper imental 37 ev idence for the involvement of the hydrophobic pocket in l igand binding. S ince the authors were only using a model led structure, quest ions a rose a s to where exactly in the fibronectin type II module did this residue lie. In June of 1999, the X-ray crystal structure of gelat inase A was resolved to 2.8 A. Th is a l lowed researchers to directly examine the gelat inase A molecule, hence al lowing for a determinat ion of the spat ial orientation of the severa l doma ins found within ge la t inase A . T h e structure of ge la t inase A conf i rmed that the fibronectin type II modules were compr ised of a pair of (3 sheets , each composed of two antiparallel strands, connected through a short a helix. It was a lso shown that the two (3 shee ts form a hydrophobic pocket that is a c c e s s i b l e from the outs ide of the molecu le . A cis-prol ine fol lowing the first (3 sheet w a s part of a hairp in turn that or iented the sur round ing a romat i c s i de c h a i n s into the hydrophobic pocket. Interestingly, it was found that in full-length gelat inase A the propeptide residue P h e 37 was inserted into the hydrophobic pocket of the third fibronectin domain . The propeptide F37 was a lso found to interact with the third fibronectin type II module v ia a hydrogen bond and a salt bridge. It was surmised by the authors that this interaction mimicked the interaction between the third f ibronectin type II module and gelatin that b a s e d on b iochemica l ev idence is predicted to have all three kinds of interactions (Pickford et al., 1997). The final p iece of ev idence uncovered the relative spatial orientation of the hydrophobic binding si tes located in each of the three f ibronectin type II modu les . It w a s found that the hydrophobic binding sites did not orient themse lves towards each other to form an extended binding cleft, but on the contrary, they turn outwards 38 as in what the author termed " a three-pronged f ishhook" (Morgunova et al., 1999). Aim of Present Work The speci f ic a im of this thesis is to identify new amino ac id res idues in the C B D that are involved in co l lagen binding in order to der ive the co l lagen binding c o n s e n s u s site. N e w si tes for mutagenes is have been se lec ted from 1) the results of our first round of mutagenes is (Steffensen et al., 2001) and 2) a 3D mode l of the C B D made us ing a "protein thread ing" a lgor i thm and energy minimizat ion (Stef fensen etal., 2001) . The mutation K 2 6 3 A s h o w e d reduced co l lagen binding. Th is identif ied a loop on the C B D between two conse rved cys te ines that is important for co l lagen interact ion. Of note, in this loop C x F P F x F x G x x Y x S ( / S x ) C is the longest conserved intra-cysteine sequence in all three C B D modu les of ge la t inase A and B; except that in ge la t inase B the p lacement of the S e r res idues is Y S x C . Lys263 is the nonconserved x in C B D 2 after the first C y s . Th is highl ights this region a s an important part of the mo lecu le . T h e requi rement for D M S O to elute bound C B D from co l lagen indicates that hydrophobic interactions are involved in binding. The. only other absolute ly conse rved aromat ic res idues in the C B D are a Trp adjacent to the third C y s and a Phe in the G C F P sequence at the end of this domain . However, their p lacement ad jacent to the c o n s e r v e d C y s res idues wou ld indicate a structural role for these residues. In order to a s s e s s the role of c o n s e r v e d pheny la lan ine res idues in gelat in binding, we used a C B D 2 3 recombinant protein s ince the individual modu les 39 proved to be difficult to express and purify. W e mutated the pheny la lan ines at posit ions 264 and 266 in module 2 and the corresponding res idues in module 3, 322 and 324, to a lan ines (numbered in relation to the full- length enzyme and both in posit ions 16 and 18 in Tab le 3). Th is a l lowed for a determination of the effect of reducing the aromat ic character of this site by exchang ing the large hydrophob ic s ide cha ins of pheny la lan ine for the sma l le r methyl g roups of a lan ine . S e c o n d l y , we mutated the pheny la lan ines to ty ros ines in order to conserve the aromat ic character and to determine the effect of introducing a phenol ic hydroxyl group that has the potential to hydrogen bond . Final ly, we introduced cor respond ing double mutat ions, F 2 6 4 A / F 3 2 2 A and F 2 6 6 A / F 3 2 4 A , into both of the conserved C B D modules. The double mutations were predicted to remove residual binding contributed by the unmutated wild type module in the paired constructs and unequivocal ly allow for the assessmen t of the role of these residues in gelatin binding. 40 Chapter 2 Materials a n d M e t h o d s Site-Directed Mutagenesis - The Escherichia coli strain D H 5 a w a s used as a host strain for p G Y M X - C B D 2 3 p lasmid isolat ion and p ropaga t ion . Us ing p G Y M X - C B D 1 2 3 a s a template, p G Y M X - C B D 2 3 w a s cons t ruc ted by P C R ampl i f icat ion of the respect ive region as desc r ibed prev ious ly (Abbey et al., 2001). The fol lowing mutations were introduced into the third co l lagen binding module of gelat inase A : F322A, F 3 2 2 Y , F 3 2 4 A and F 3 2 4 Y using the mutagenic primers: F 3 2 2 A 5 ' - G C C C A G G A A A G T G A A G G G G G C G A C A C A G G G G G C - 3 ' and 5 ' - G C C C C C T G T G T C G C C C C C T T C A C T T T C C T G G G C - 3 ' ; F 3 2 2 Y 5 ' - G C C C A G G A A A G T G A A G G G G T A G A C A C A G G G G G C - 3 ' and 5 ' - G C C C C C T G T G T C TAC C C C T T C A C T T T C C T G G G C - 3 ' ; F 3 2 4 A 5' -G C C C A G G A A A G T G G C G G G G A A G A C A C A G G G G G C - 3 ' and 5 ' - G C C C C C T G T G T C T T C C C C G C C A C T T T C C T G G G C - 3 ' ; and F 3 2 4 Y 5 ' - G C C C A G G A A A G T G T A G G G G A A G A C A C A G G G G G C - 3 ' and 5 ' - G C C C C C T G T G T C T T C C C C TAC A C T T T C C T G G G C - 3 ' . D H 5 a w a s the cel l line transformed with all mutagenesis products for the purposes of preparing plasmid stocks for sequenc ing and further experiments. The primers used for sequenc ing inc luded H I E G R , M 1 3 R e v e r s e and M 1 3 - 1 7 Forward . Cor rec t ly exchanged nucleot ides were determined by d ideoxysequenc ing . Simi lar p rocedures were fol lowed by And rea Connor for the preparation of corresponding mutations within the second col lagen binding domain of gelat inase A (F264A, F 2 6 4 Y , F266A , and 41 F266Y) . Us ing the mutated F 2 6 4 A and F 2 6 6 A p G Y M X - C B D 2 3 constructs and the above F 3 2 2 A and F 3 2 4 A pr imers, the double mutants F 2 6 4 A / F 3 2 2 A and F 2 6 6 A / F 3 2 4 A were p roduced by Denn is L e e . T h e mutants F 3 2 2 A , F 3 2 4 A , F 3 2 2 Y , F 3 2 4 Y , F 2 6 4 A / F 3 2 2 A , a n d F 2 6 6 A / F 3 2 4 A w e r e pur i f ied a n d character ized by myself as descr ibed below. Purification of Recombinant Gelatinase A Domains and Antibodies - r C B D 2 3 ( A l a 2 4 9 - G l n 3 6 4 ) of human gelat inase A was exp ressed in Escherichia coli BL21 gold and purified by Ni 2 + -che la te and gelatin Sepha rose chromatography. A 5 ml seed culture of p G V M X - C B D 2 3 - t r a n s f o m e d BL21 Go ld cel ls was grown overnight at 37 °C in L B media (1% (w/v) Bactotryptone, 0 .5% (w/v) Yeas t extract, 1% (w/v) NaCI , pH 7.5) with 40 u.g/ml ampici l l in. 7 x 600 ml superbroth cultures (1% (w/v) Bactotryptone, 0 .8% (w/v) Yeas t extract, 1% (w/v) NaCI , 0 . 0 0 1 % glycerol (v/v), pH 7.4) conta in ing ampici l l in (40 j ig/ml), were e a c h inoculated with a 1/1000 dilution of the s e e d culture. The cultures were incubated for 24 h at 37 °C. Ce l l s were harvested by centrifugation (7,000 x g for 10 minutes at 4 °C). The cel ls were s u s p e n d e d in a total of 100 ml of lysis buffer (20 m M N a 2 H P 0 4 , 50 m M NaCI , 2 m M MgCI, 5 j ig/ml D N a s e , 1 mg/ml lysozyme, 1 m M P M S F , pH 8.0) and incubated, with agitation, at 37 °C for 2 h. The samp le w a s then son icated, on i ce , for 3 x one minute in terva ls . Inc lus ion b o d i e s we re r e m o v e d by centrifugation (15,000 x g for 15 min at 4°C). Inclusion bodies were d isso lved in 200 - 300 ml of solubi l izat ion buffer (10 m M Tr is, 8 M Urea , pH 8.0) and stirred vigorously for 4 - 18 h. Upon pellet dissolut ion, 30 ml of N i 2 + - cha rged chelat ing Sepha rose was added to the solubil ization solut ion. The resin was batch- loaded 42 for 2 h to overn ight at 4 °C. T h e ba tch - loaded resin w a s recovered by centrifugation at 500 x g for 5 minutes at 4 °C. The resin was w a s h e d repeatedly with chromatography buffer (CB)(100 m M N a 2 H P 0 4 , 0.5 M NaCI , 8 M urea, pH 8.0) until the supernatant was clear. The resin was packed into a Kontes g lass co lumn (3 c m x 15 cm) and w a s h e d with 100 ml of C B . T h e co lumn was then washed with 250 ml_ of C B + 1 M NaCI and the 250 ml_ of C B + 1 M NaCI , pH 6.0. The co lumn w a s re-equi l ibrated with C B and subsequen t elut ions were performed by F P L C (B iochem-Pharma) . Non-speci f ical ly bound protein was pre-eluted in C B + 10 m M imidazole. Speci f ical ly bound r C B D mutant proteins were eluted with a 500 m M imidazole step gradient. P e a k fractions were pooled and diluted three-fold with C B and the sample was p laced in a dialysis bag ( 6 - 8 kDa cutoff). Protein was refolded against a 2 x vo lume of Vo l le rs buffer (18.2 M N a 2 C 0 3 , 24 m M N a H C 0 3 , 0.273 m M cyste ine/2.73 m M cyst ine, pH 10.0), with aerat ion, for 2 h at room temperature. After refolding the samp le w a s d ia lyzed against a 2 x vo lume of phosphate-buffered sal ine (PBS) (140 m M NaCI , 2.7 m M KCI, 4.3 m M N a 2 H P 0 4 - 7 H 2 0 , 1.5 m M K H 2 P 0 4 ) plus 0.273 m M cyste ine/2.73 m M cyst ine, pH 7.4, with aeration, for 2 h at room temperature. Refo lded protein was exhaust ive ly d ia lyzed against P B S to remove urea. Correct ly refolded protein w a s se lec ted for by us ing gelat in S e p h a r o s e affinity chromatography . The refo lded samp le w a s app l ied to a PBS-equ i l i b ra ted 7 ml gelat in Sepharose co lumn. Speci f ica l ly bound protein was eluted in a 20 ml 0 - 10 % dimethyl sulfoxide ( D M S O ) gradient. Fract ions containing r C B D 2 3 proteins were dia lyzed against P B S at 4 °C (2 x 2 h). Protein samp les were al iquoted, f rozen, and then stored at - 2 0 ° C . Po lyacry lamide gel e lectrophoresis ( P A G E ) in 1 5 % gels was 43 used to fol low the puri f icat ion of recombinant prote in. E lec t rospray m a s s s p e c t r o s c o p y w a s per fo rmed on a S C I E X AP I 300 (Perk in -E lmer ) m a s s spectrometer was used to a s s e s s fidelity of gene express ion . Rabbi t polyclonal ant ibody (a-H is 6 ) ra i sed aga ins t the N H 2 - t e r m i n a l H i s 6 fus ion tag on the recombinant proteins was affinity purified as before (Steffensen etal., 1998). Microwell Substrate Binding Assays for the Determination of Gelatin Affinity - 96 microwel l p lates were coa ted with 0.5 u.g/well of denatured type I co l lagen (gelatin). Unbound gelat in w a s removed by P B S + 0.05 % T w e e n 20 (v/v) washes and the plates were blocked with 2.5 % (w/v) B S A in P B S for 1 h at room temperature. After further washes with P B S / T w e e n , wild-type and mutant protein preparat ions were clarif ied by centirifugation at 10,000 x g for 5 min at 4 °C and quantitated at 280 nm immediately before use. S a m p l e s were serial ly diluted in 1 x P B S , from 5 \iM to 0.6 p M and al lowed to react for 1.5 h at room temperature. B a s e d upon the aromatic content of the protein, the protein concentrat ions were determined using extinction coeff icients (expressed as protein concentrat ion in mg/mL per unit of absorpt ion at 280 nm) calculated using the software program Protean (Lasergene) . After incubation, the plates were washed with P B S / T w e e n . Bound recombinant protein was detected with an ocHis 6 antibody dilution (1:500 in P B S + 0.01 % T w e e n 20 + 0.25 % B S A ) , fo l lowed by a react ion with alkal ine phosphate-conjugated goat anti-rabbit antibody (Bio-Rad) (1:1000 P B S dilution + 0.01 % T w e e n 20 + 0.25 % B S A ) . Both incubat ions were for 1 h at room temperature and plates were washed with P B S / T w e e n 20. For quantitation, /> nitrophenyl phosphate d isodium (Sigma) was added as a substrate. To ensure linearity of the a s s a y , the colour intensity w a s determined at var ious t imes by 44 measur ing the abso rbance (405 nm) in an automated plate reader. Negat ive contro ls for non-spec i f i c binding cons is ted of react ion mixtures subst i tut ing respect ive r C B D 2 3 proteins with recombinant horse heart myoglobin. The best curves were fitted to the data using S igmaPlo t 2000 ( S P S S ) . Absorpt ion va lues that we re d e c r e a s e d at h igh recomb inan t prote in c o n c e n t r a t i o n s we re disregarded when fitting the curve. FPLC Analysis to Assay for Gelatin Affinity - O n an F P L C mach ine a gelatin Sepha rose (1 ml_) (Pharmacia) co lumn was packed in a Kontes co lumn ( 1 x 1 0 cm) and equil ibrated in P B S . The absorpt ion basel ine was establ ished with P B S . Recombinant proteins were clarif ied by centrifugation at 10,000 x g for 5 min at 4 °C and quant i tated at 280 nm immediate ly before use . Individually, 500 u.1 samp les of proteins (50 u.g) were sequential ly loaded onto the co lumn, al lowed to transiently bind, and then eluted with D M S O . The co lumn w a s loaded, washed and eluted at a flow rate of 2.5 ml/min. After loading, the co lumn was washed with approximately 5 ml_ of P B S to reestabl ish a zero absorbance basel ine after the f low-through peak. Bound protein was eluted with a 20 ml 0 - 1 0 % D M S O gradient. Differential elution of bound protein was observed v ia superposi t ion of F P L C traces. 45 Chapter 3 RESULTS Recombinant Protein Expression - To invest igate the role of spec i f ic C B D res idues in subst rate recogni t ion, we used the p G Y M X - C B D 2 3 construct to express the delet ion mutant r C B D 2 3 (lacking C B D 1 ) (Abbey e r a / . , 2001) . A s il lustrated by S D S - P A G E analys is , wild-type and mutant r C B D 2 3 proteins were exp resse d in E. coli and purified to homogenei ty (Figure 6). Fideli ty of gene e x p r e s s i o n , conf i rmat ion of N-terminal meth ion ine p r o c e s s i n g , and samp le homogeneity was determined by mass spectrometry (Table 4). Microwell Substrate Binding Assays for the Determination of Gelatin Affinity - T h e relat ive affinity of wi ld-type and mutant r C B D 2 3 for gelat in w a s determined through the use of E L I S A - b a s e d a s s a y s . It w a s found that the apparent d issociat ion constant for wild-type r C B D 1 2 3 binding to immobi l ized gelatin was approx imate ly 1 x 10" 8 M , w h e r e a s the b ind ing of wi ld- type r C B D 2 3 w a s approximately 8 x 10 ' 8 M (Figure 7). The reduction in binding of approximately one order of magni tude is supported by the earl ier f indings of Stef fensen et al. (1998). Th is dec rease supports the hypothesis that cooperat iv i ty is involved in the binding of the C B D to substrate. The mutants F 2 6 4 A and F 3 2 2 A exhibited a 2.5 - 3 fold reduct ion in b inding, K d ~ 0.2 x 10 ' 6 and 0.24 x 10" 6 , respect ively. Binding of the cor responding Phe->Tyr mutants was simi lar to that of the wild-type r C B D 2 3 protein (Figure 7). T h e mutants F 2 6 6 A a n d F 3 2 4 A s h o w e d a greater reduction in gelatin binding as observed by a respect ive 6-fold and 13-46 Figure 6 14.4 kDa-S D S - P A G E analys is of purified protein samp les after gelatin Sepharose chromatography. The samples were electrophoresed under reducing condit ions on a 12.5 % Tris-Tricine gel . 47 Tab le 4: E lec t ros pray M a s s S p e c t r o m e t r y ol F R e c o m b i n a n t Prote ins Recombinant Protein Expected M r (Da) Found M r (Da) A Da + Met - Met rCBD12 Wt 15 094 14 963 14 961 2 rCBD23 Wt 14 680 14 549 14 547 2 F264A 14 604 14 473 14 477 -4 F264Y 14 696 14 565 14 564 1 F266A 14 604 14 473 14 477 -4 F266Y 14 696 14 565 14 566 -1 F322A 14 604 14 473 14 470 3 F322Y 14 696 14 565 14 567 -2 F324A 14 604 14 473 14 472 1 F324Y 14 696 14 565 14 566 -1 F264A/F322A 14 528 14 397 14 397 0 F266A/F324A 14 528 14 397 14 398 -1 E lec t rospray mas s spect rometry of recombinant prote ins. M a s s e s of the recombinant proteins were measured and compared to va lues ca lcu lated from the amino ac id composi t ion minus 4 Daltons for the loss of hydrogens during disulphide formation. r CBD12 and mutants F264A, F264Y, F266A, and F266Y were purified by A. Connor. The mutant F264A/F322A was purified by D. Lee. 48 F264A/F322A & F266A/F324A vs rCBD23 Wt 1.25 1.00 3 0.75 0.50 0.25 0.00 1e-11 1e-10 1e-9 1e-8 1e-7 1e-6 1e-5 Protein Concentration (M) Plots of the microwell substrate binding assays were used to determine the apparent dissociation constants. The figure shows that: (1) single alanine mutations had reduced affinity for denatured type I collagen, (2) tyrosine mutations maintained gelatin binding, (3) reduced gelatin binding was observed for F264A/F322A and (4) abrogated gelatin binding was observed for the F266A/F324A mutant. Representative horse heart myoglobin controls are shown in panels C and E. Plots A and B were made by A. Connor. 49 fold reduct ion in gelatin binding ve rsus wild-type r C B D 2 3 (Figure 7). Aga in , binding was essent ial ly unaltered in the corresponding Phe->Tyr mutants (Figure 7). In order to a s s e s s the contribution of the non-mutated domain (ie. C B D 3 in the F 2 6 4 A or F 2 6 6 A single mutants and C B D 2 in the F 3 2 2 A or F 3 2 4 A single mutant prote ins) , the co r respond ing doub le mutat ions F 2 6 4 A / F 3 2 2 A and F 2 6 4 A / F 3 2 4 A were made . The F 2 6 4 A / F 3 2 2 A mutant s h o w e d that b inding affinity to gelatin w a s reduced (K d ~ 0.46 x 10"6) relative to the single mutants F 2 6 4 A and F 3 2 2 A (K d ~ 0.2 x 10' 6 and 0.24 x 10~6), but not more than the single mutants F 2 6 6 A and F 3 2 4 A (K d ~ 0.5 x 10" 6 and 1.1 x 10" 6). The importance of posi t ions 266 and 324 in gelat in binding w a s shown by the doub le mutant F 2 6 6 A / F 3 2 4 A which showed no affinity for gelatin (Figure 7). Differential Elution of Wild-Type and Mutant rCBD23 from a Gelatin Sepharose Column - In order to further a s s e s s the ability of mutant doma ins to bind gelatin, we exploi ted the hydrophobic interactions during binding of f ibronectin type II modu les to gelat in . D M S O disrupt ion of these interact ions a l lows for the compar ison of elution profi les. Recombinant proteins were loaded onto a 1 ml gelatin Sepha rose column and were eluted with a 0 - 10 % D M S O gradient. The differential elution of mutant and wild-type r C B D 2 3 proteins w a s compared by superposi t ion of F P L C absorbance profiles (Figure 8). F 2 6 4 A and F 3 2 2 A eluted in 3.5 % and 3.0 % D M S O respect ively. The elution prof i les of the mutants F 2 6 4 Y and F 3 2 2 Y were very similar to that of wild-type r C B D 2 3 , reveal ing that the introduction of a hydroxyl group, which has the potential for hydrogen bond formation, did not alter aga in gelatin binding. The mutants F 2 6 6 A and F 3 2 4 A eluted at lower D M S O concentrat ions, 2.7 % and 3.0 % respect ively, than the 50 0 5 10 15 20 25 30 35 40 45 Fraction Number B 0.012 0.006 -0.006 •V. . i i 5 10 15 20 25 Fraction Number . 0.016 0.008 0.012 0.006 -0.006 0 5 10 15 20 25 30 35 4Q 45 Fraction Number 0 5 10 ; 15 . 20 25 Fraction Number : : • 0.006 rCBD23 Wt g F264A (A), F322A (B), F266A (C), F324A (D), F264A/F322A (E) 8 | F264Y (A), F322Y (B), F266Y (C), F324Y (D), F264Y/F322Y (E) DMSO Gradient Zero -0.006 DMSO mini-column elution profiles. 50 \SJQ of each protein was added to 1 ml_ gelatin Sepharose columns. The elution of the respective proteins is illustrated above. This figure illustrates that removal of hydrophobic character reduces the percentage of DMSO required to elute bound protein. Panel A shows the eluting concentrations of F264A and F264Y were found to be 3.3 and 4.3% respectively. The eluting concentration of wild-type rCBD23 was determined to be 4.2%. Panel B shows the eluting concentrations of F266A and F266Y were found to be 2.4 and 4.2% respectively. These data illustrate that removal of F266 hydrophobic character had a greater effect than removal of F264 hydrophobic character. Panel C shows the eluting concentrations of F322A and F322Y were found to be 2.8 and 4.7%, respectively. The eluting concentration of wild-type rCBD23 was determined to be 4.7%. Panel D shows the eluting concentrations of F324A and F324Y were found to be 3.0 and 4.6%, respectively. These data illustrate that removal of F322 hydrophobic character had a greater effect than removal of F324 hydrophobic character. Panel E shows that the mutant F264A/F322A avidly bound gelatin Sepharose, whereas the second double mutant F266A/F324A showed no affinity for gelatin Sepharose. Panels A and C were made by A. Connor. 51 F264A and F322A mutants. Aga in , introduction of hydroxy l g roups in the mutants F266Y and F324Y did not significantly alter the elution profile relative to that of the wild-type domain s ince the core hydrophobicity of the molecule was not perturbed. This result reiterates the previous observat ions for mutants F264 and F322 i l lustrating that the reduct ion of hydrophobic character dec reases gelatin affinity and that introduction of phenol ic hydroxyls does not alter binding. Involvement of the wild-type module in binding gelatin, that renders interpretation of the s ingle Phe mutants complex, was add res sed by apply ing the double mutants to a gelatin Sepharose column. The mutant F264A/F322A unexpectedly e luted at a D M S O concentrat ion greater than that of any of the s ingle A l a mutants ( D M S O % = 4.1). The second double mutant did not bind gelatin and was recovered only in the flow through (Figure 8). This is in agreement with our EL ISA results where F266A/F324A did not bind gelatin. 52 CHAPTER 4 DISCUSSION It has been shown that there is a distant homology between type II domains and the three-disulphide kringle units of thrombolytic and fibrinolytic e n z y m e s (Patthy et al. 1984). The non-proteolytic kringle domains are known to mediate protein-protein interactions (Trexler and Patthy, 1983). It was found that the fibronectin type II domain bound co l lagen and similarly bovine semina l fluid protein P D C -109b bound to gelatin. Through the use of N M R studies, the prec ise structure of P D C - 1 0 9 b w a s e luc idated which contr ibuted to a better unders tanding of the f ibronectin type II doma ins (Constant ine etal., 1992). It w a s found that the mo lecu le con ta ined a hydrophob ic sur face that w a s hypo thes i zed to be a putative binding sur face for gelatin (Constant ine et al., 1992). Further N M R studies of the first f ibronectin type II domain showed that there w a s indeed a hydrophobic cluster on one side of the molecule and that the cluster contained a putative hydrophobic binding pocket (Pickford et al., 1997). F ive surrounding hydrophobic res idues and one strictly conserved tryptophan at the bottom of the pocket forms the hydrophobic pocket on the f ibronectin type II doma in . The general principle that a conserved primary sequence infers information about the amount of c o n s e r v e d tertiary structure a l lowed for model l ing of the s e c o n d f ibronect in type II module of ge la t inase A to P D C - 1 0 9 b (Tordai and Patthy, 1999). Until now there has been a limited number of studies looking at the direct involvement of res idues contained within the hydrophobic pocket. Therefore, in 53 the present work we undertook the chal lenge to systemat ical ly illustrate the roles of key amino ac ids in gelatin binding. The Overal l laboratory has previously demonstrated that chemica l modification of the C B D shows that lysine residues are solvent exposed and that one or more of the res idues is involved in heparin and gelatin binding (Steffensen et al., 2001). S i n c e r C B D 1 2 s h o w e d weaker co l lagen binding, the lys ine res idues in C B D modules 2 and 3 were targeted for a lanine substitution to detect binding si tes. Total loss of heparin binding by the K 3 5 7 A mutation conf i rmed the importance of this site in heparin binding. A second mutant protein, K 2 6 3 A , showed reduced levels of binding at saturation to type I co l lagen without a concurrent change in the apparent d issociat ion constant. Neither of the mutants with al tered heparin or type I co l lagen affinity nor the K330 , K343A , or K298 /299A mutants d isp layed any c h a n g e s in binding to denatured types IV or V co l lagen , or e last in. Th is indicates that the interaction of K263 is l igand speci f ic (Stef fensen e r a / . , 2001). T h e K 2 6 3 A mutat ion identif ied a loop on the C B D be tween two c o n s e r v e d cys te i nes that is important for co l l agen in teract ion. Of note, this loop, C x F P F x F x G x x Y x S ( / S x ) C (Table 3) is the longest c o n s e r v e d int ra-cyste ine s e q u e n c e in all three C B D modu les of ge la t inase A and B; except that in ge la t inase B the p lacement of the S e r re idues is Y S x C . T h e requirement for D M S O to elute bound C B D from col lagen indicates that hydrophobic interactions are involved in binding. Th is highlights these conse rved phenyla lan ine sites (Table 3). 54 Using the V A S T search serv ice offered by the National Center for Biotechnology Information, we were ab le to super impose the three C B D doma ins over the fibronectin type II domain (Figure 9). It can be seen that there is a high degree of structural homology between the three C B D molecu les and F N . Morgunova et al. have presented crysta l lographic data that orients the hydrophobic pockets within each of the C B D s on the outside of what has been l ikened to a three-prong f ishhook (1999). Upon looking at the superimposit ion of these individual domains it can be seen that the hydrophobic pocket is strictly conse rved , with only slight variation in the backbone regions outside the hydrophobic pocket. In our present work we made the mutants F264A, F 2 6 4 Y , F266A , F 2 6 6 Y , F322A, F322Y , F324A , F324Y , F 2 6 4 A / F 3 2 2 A and F 2 6 6 A / F 3 2 4 A . T h e s e mutations were des igned to remove the large aromat ic s ide cha ins of the pheny la lan ines by exchang ing the benzy l s ide chain of phenyla lan ine to the smal le r methyl s ide chain of a lanine. A lan ine mutations are typically conservat ive and do not disrupt the structure. T h e s e mutat ions were targeted against the first two conserved Phe residues in the conserved sequence C x F P F x F x G x x Y x S ( / S x ) C that has been impl icated to be involved in gelatin binding (Stef fensen e r a / , . 2001) (Table 3). Our mutagenes is plan preceded the report of the structure of ge lat inase A and this has subsequen t l y enab led us to determine the exac t or ientat ion of the phenyla lanine res idues within the C B D . The F264 and F322 res idues appear to be more buried within the (3-sheet backbone of the C B D molecu le and as such would be predicted to have a lesser effect on binding and possib ly play more of a role in contr ibuting to the structural stability of the molecu le . Indeed, we still observed binding in these alanine mutants that was slightly reduced relative to 55 Figure 9 F322 FN Type I I vlvQT rggnsnGALC HFPFLynnhN YTDCTSEgRR DNMKWCGTTQ NYDADQKFGF Cpma hGelA_CBDl qvvRV kygnadGEYC KFPFLfngkE YNSCTDTgRS DGFLWCSTTY NFEKDGKYGF Cphe hGelA_CBD2 t a l F T mggnaeGQPC KFPFRfqgtS YDSCTTEgRT DGYRWCGTTE DYDRDKKYGF Cpet hGelA_CDB3 tamST vggnseGAPC VFPFTfIgnK YESCTSAgRS DGKMWCATTA NYDDDRKWGF CPdq Three d imensional structural al ignment of the first (green), s e c o n d (blue), and third (pink) f ibronectin type ll-l ike modules of gelat inase A onto the first fibronectin type II doma in (brown). Th i s f igure i l lustrates that there is a high deg ree of three d imens iona l conservat ion in the f ibronectin type l l- l ike modu les . The red regions equate to highly conserved spatial organizat ion. 56 that of the wi ld-type domain . Th is indicated that mutat ions, wh ich alter the hydrophob ic co re , did not induce a fatal d isrupt ion of molecular , stabi l i ty. Nonethe less we consider it possib le that the removal of aromatic character at this posi t ion by the P h e - > A l a mutat ions cou ld reduce hydrophobic i ty within the binding pocket by destabi l isat ion of the Trp located at the bottom of the binding pocket. Microwel l substrate binding a s s a y s demonst ra ted that the wi ld-type r C B D 1 2 3 bound gelatin more avidly than wild-type r C B D 2 3 (Figure 7). Abbey e r a / . (2001) found that there was a 10-fold dec rease in the binding of r C B D 1 2 to denatured type I co l lagen. Tordai and Patthy (1994) d iscovered that removal of the middle domain of C B D 1 2 3 ( C B D 1 3 ) c a u s e d there to be a d e c r e a s e in the K a . Th is il lustrates that all of the three C B D modules facilitate the binding of the C B D to substrate. Delet ion mutant studies us ing 'a sol id phase a s s a y have lead to the observa t ion that the loss of a C B D modu le a c c o u n t s for a reduct ion of approximately one order of magnitude. The mutants F 2 6 4 A and F 3 2 2 A showed that the mutated domain contr ibuted to s o m e of the binding as there was an approx imate 2.5 - 3 fold reduction in binding for each of the proteins. If there were no contribution by the mutated domain , ie. if it were improperly fo lded, the s ing le wi ld- type doma in would bind and therefore accoun t for an ~10-fold reduction in binding. The F 2 6 4 A and F 3 2 2 A mutants show reduced , but not abrogated, binding within the mutated domain . Th is result w a s attributed to the fact that these residues are buried well within the molecule and that they are far removed from the Trp at the bottom of the hydrophobic pocket (closest d istance between the Trp and the Phe rings being 9.2 and 8.7 A, respectively)(Figure 10). 57 Figure 10 Crystal structure of the third fibronectin type II module from gelat inase A. Yel low residues highlight the conserved hydrophobic pocket that interacts with F37 of the propeptide (orange). The magenta res idues are F322 and F324, with F322 being located further away from the hydrophobic pocket than F324. 58 Tyros ine mutants were used to determine the effects of a conservat ive mutation on the module structure. A binding curve similar to that of wild-type r C B D 2 3 was s e e n for the cor respond ing Phe—>Tyr mutants (Figure 7) indicat ing that the introduction of a buried phenol ic hydroxyl group did not perturb the hydrophobic pocket-gelatin interaction. Poss ib ly these mutants will bind other substrates with altered affinity, but this has yet to be tested. The mutants F 2 6 6 A and F 3 2 4 A showed a greater reduction in gelatin binding as observed by a respect ive 6-fold and 13-fold reduction in gelatin binding versus wild-type r C B D 2 3 (Figure 7). The F266A mutants apparent d issociat ion constant showed that there is some residual binding by the mutated domain , converse ly the F 3 2 4 A mutant had a greater than 10-fold reduction in binding. The greater than 10-fold reduction in binding by F 3 2 4 A could be attributed to a disruption of substrate contact by the wild-type domain . Pe rhaps the mutant domain indirectly h inders subst ra te interact ions with the unmutated modu le (F igure 7). The res idues F266 and F324 are in c lose proximity to the Trp at the bottom of the hydrophobic pocket compared with F264 and F322 and therefore we suggest have a more profound effect on gelatin binding (closest d istance between the Trp and the Phe rings for F266 is 4.4 A and F324 is 4.1 A)(Figure 10). T h e 5.4-fold reduct ion in b inding of the F 2 6 4 A / F 3 2 2 A mutant s h o w e d that binding was reduced relative to the single mutants F 2 6 4 A and F 3 2 2 A but not more than the single mutants F266A and F324A. Th is further il lustrated that the F 2 6 4 A and F 3 2 2 A mutants had reduced binding and that the apparent K d ' s for the single mutants represented the average of the individual apparent K d ' s in the 59 bimodular mutants of r C B D 2 3 . The reduction in binding for the single mutants F 2 6 6 A and F 3 2 4 A was very similar to the reduction in binding when compar ing r C B D 1 2 3 to ei ther r C B D 1 2 or r C B D 2 3 . Individual muta t ions essen t ia l l y el iminated binding within that module. Indeed, introduction of both mutations into the r C B D 2 3 protein totally el iminated gelatin affinity (Figure 7). Differential elut ion of wi ld-type and mutant r C B D 2 3 f rom gelat in Sepha rose explo i ted the property of D M S O in disrupt ing hydrophobic interact ions. The d isc repancy in binding of the F 3 2 2 A and F 3 2 4 A mutants in the E L I S A a s s a y s and the gelatin S e p h a r o s e min i -co lumns may be a result of the exposure of a critical epitope on gelatin that is masked when the gelatin is bound to a microwell titre plate. The mutants F 2 6 4 Y and F 3 2 4 Y had an elution profile simi lar to the wild-type domain , thus reiterating our observat ions that the reduction of aromatic character d e c r e a s e s gelatin affinity and that introduction of a phenol ic hydroxyl group does not alter gelatin affinity. Us ing gelat in S e p h a r o s e chromatography the invo lvement of the wi ld-type module in the s ingle P h e mutants was add ressed . The mutant F 2 6 4 A / F 3 2 2 A eluted at a D M S O concentrat ion greater than that of any s ing le A l a mutants ( D M S O %= 3.9). The seemingly high elution concentrat ion of the double mutant F 2 6 4 A / F 3 2 2 A , which retained 83 % of its affinity for gelatin (Table 5), could be attributed to binding of the second module to gelatin binds gelatin at a capacity of 7 9 % that of wild-type protein (Table 5). The second double mutant did not bind gelatin being solely in the flow through materials. Th is is in agreement with our E L I S A resu l t s w h e r e F 2 6 6 A / F 3 2 4 A d id not b ind ge la t i n . F r o m our 60 Table 5 EL ISA and Elution Profile Data for mutants Recombinant Protein K d x 10"6 (M) [n] (K d Mutant) / (K d Wt) Eluting D M S O Concentrat ion (%) [n=2] D M S O % Mutant / D M S O % Wt rCBD123 Wt 0.028 0.3 nd nd rCBD23 Wt 0.084 1 4.2* 1.00 F264A 0.20 2.3 3.3* 0.79 F264Y 0.11 1.3 4.3* 1.02 F266A 0.49 5.8 2.4* 0.57 F266Y 0.093 1.1 4.2* 1.00 r CBD23 Wt 0.084 1 4.7* 1.00 F322A 0.24 2.8 2.8* 0.60 F322Y 0.19 2.2 4.7* 1.00 F324A 1.1 13 3.0* 0.64 F324Y 0.16 1.9 4.6* 0.98 F264A/F322A 0.46 5.4 3.9* 0.83 F266A/F324A 0 0 0* 0 "Column runs were performed by A. Connor, "Column runs were performed by T. Moore 61 character isat ion of the hydrophobic interactions of the fibronectin type II modules of gelat inase A we have determined that the removal of aromatic character from this region c a u s e s a loss of gelatin affinity. W e propose this may be through destabi l isat ion of the Trp located at the bottom of the hydrophob ic pocket . Ev idence for Trp 40 interactions will be determined in the future by a Trp—>Ala mutation. The present character isat ion of interactions between the hydrophobic pocket and substrate has cont inued to extend our knowledge of ge lat inase A substrate interactions. Further character isat ion of the C B D is required in order to fully unders tand substrate interact ions involving the contr ibut ion of polar and hydrogen bonding res idues , binding modal i t ies that are a l so be l ieved to be present in fibronectin type II module/gelatin interactions. 62 Chapter 5 C o n c l u s i o n The aromatic s ide cha ins of res idues F266 and F324 extend behind the Trp that l ies at the bottom of the hydrophobic pocket (Figure 10). There fore , these res idues are in c lose enough proximity that removal of their aromat ic character cou ld disturb the Trp that is bel ieved to form the base of a substrate binding pocket that is predicted to interact with substrate. Th is was best illustrated in our E L I S A data, where F 2 6 6 A bound gelatin only 41 % as well as F 2 6 4 A and F324A bound gelatin only 22 % as well as F322A. The most profound illustration of the d i f ference in gelat in b inding be tween the two res idues w a s the d i f ference between F 2 6 4 A / F 3 2 2 A and F 2 6 6 A / F 3 2 4 A where gelat in binding w a s reduced and then completely abrogated. The direction for future exper iments should focus on directly implicating Trp 39 in gelat in b ind ing . Mutat ing this res idue to an a lan ine wou ld remove the hydrophobic character at the bottom of the pocket which we hypothes ise would el iminate gelat in b inding. Upon complete mapping the hydrophobic pocket a m o r e g l o b a l v i e w of p o s s i b l e C B D i n v o l v e m e n t in ac t i va t i on a n d col lagenolys is /gelat inolys is could be invest igated. Th is would be accompl i shed by introducing the above effective point mutat ions and/or the putative Trp 39 mutation into the full length enzyme. The first experiment would be to follow the process of activation s ince removal of the hydrophobic pocket with the third C B D will in all l ikel ihood lead to a superact ivatable gelat inase A . Upon determination 63 of the eff iciency of gelat inase A binding to immobi l ized co l lagen through the use of E L I S A and gelat in S e p h a r o s e b ind ing/compet i t ion a s s a y s , an effect ive ge la t in /co l lagen-minus binding mutant would al low for a determinat ion of the m o d e of per ice l lu lar ge la t i nase A b ind ing and seques t ra t i on . Ac t i va ted g e l a t i n / c o l l a g e n - m i n u s g e l a t i n a s e A w o u l d d e m o n s t r a t e C B D -dependent/ independent substrates of gelat inase A . Th is would either reveal new subst rates s ince ge la t inase A may be bound and quarant ined by co l lagen or substrates that bind sole ly to the C B D . Th is C B D "minus" mutant would offer insight into the role of the C B D and a perspect ive into the mechan ism of col lagen triple hel icase activity. F rom the X- ray crystal structure of ge la t inase A it w a s found that F37 of the propeptide bound back to the third C B D . Two exper iments cou ld be des igned based on this observat ion: 1) synthesis of a propeptide ana logue consist ing 9-10 amino ac ids and 2) mutat ion of F 3 7 to either an a lan ine or a t ryptophan. Syn thes i s of a 9 to 10 res idue pept ide would al low for exper iments into the activation of gelat inase A . Ge la t inase A bound to the cel l sur face v ia C B D 1 and C B D 2 is ma in ta ined latent by the propept ide. Add i t ion of the synthet ic propeptide may d is lodge the C B D 3 - b o u n d propeptide indicating the importance of this interaction. If so , we propose that if the free propeptide l iberated by M T 1 -M M P c leavage of gelat inase A may compete for binding of the propeptide on the C B D and thus may ampli fy the act ivat ion of ge la t inase A . T h e F37 to A 3 7 mutat ion wou ld a d d r e s s this i ssue from another v iewpoint . If there is no alteration in activation, we would conc lude that they F37 interaction with C B D 3 is an artifact of crystal l ization. 64 R e f e r e n c e s Abbey, R.S., Stef fensen, B., and Overal l , C M . (2001). Modu le cooperat ion in gelat inase A col lagen binding domain. In Preparation Alberts, B., Bray, D., Lewis, J . , Raff, M., Roberts, Watson , J . (1994). Molecular Biology of the Cel l (3 r d ed.), pp 971 - 1000. New York: Gar land Pul ishing Inc. Al lan, J.A., Docherthy, A . J . , Barker, P.J., Husk i sson, N.S., Reyno lds , J . J . , and Murphy, G . (1994). Binding of Ge lat inases A and B to Type-I Co l lagen and Other Matrix Components . Biochem.J. 309, 299-306. Aumail ley, M. and Gayraud B., (1998). Structure and biological activity of the extracellular matrix. J Mol Med76 253-265 Banyai , L. and Patthy, L. (1991). Ev idence for the involvement of type II domains in col lagen binding by 72 kDatype IV procol lagenase. FEBS Lett 282 23-25 Banyai , L. and Patthy, L. (1994). The gelatin-binding site of human 72 kDa type IV co l lagenase (gelatinase A). Biochem. J. 298 403-407 Banya i , L., Torda i . H., and Patthty, L. (1996) Structure and domain-domain interactions of the gelatin binding site of human 72-kilodalton type IV co l lagenase (gelatinase A, matrix metal loproteinase 2). J. Biol. Chem. 271 12003-12008 Bigg, H. F., Sh i , Y . E., Liu, Y . E., Stef fensen, B., and Overa l l , C. M. (1997). Specif ic, high affinity binding of t issue inhibitor of metal loproteinases-4 (TIMP-4) to the COOH- te rmina l hemopexin domain of human gelat inase A. T IMP-4 binds progelat inase A and the COOH-termina l domain in a similar manner to TIMP-2. J Biol Chem 272 15496-15500 65 Bode W. , Fe rnandez -Ca ta l an , C , T s c h e s c h e , H., G r a m s , F., N a g a s e H., and M a s k o s K. (1999). Structural propert ies of matrix meta l loprote inases. Cell Mol Life Sci 55 639-652 B o u d r e a u , N. J . and J o n e s , P. L , (1999) Ext racel lu lar matrix and integrin signal l ing: the shape of things to come. Biochem. J. 339 481 -488 Bresn ihan , B. (1999). Pa thogenes i s of joint d a m a g e in rheumatoid arthritis. J Rheumatol 26 717-719 Br iknarova, K., G r i shaev , A . , Banya i , L , Torda i , H., Patthy, L , and L l inas, M. (1999) T h e s e c o n d type II module from human matrix meta l lopro te inase 2: structure, function and dynamics. Structure FoldDesl 1235-1245 Brown, P. D., Levy, A . T., Margul ies, I. M. , Liotta, L. A . , and Ste t le r -Stevenson, W. G . , (1990). Independent express ion and cellular process ing of Mr 72,000 type IV c o l l a g e n a s e and interstitial co l l agenase in human tumor igen ic cel l l ines. Cancer. Res. 50 6184-6191 But ler, G . S . , Hut ton, M. , Wa t tam, B. A . , W i l l i amson , R. A . , Knauper , V . , Wi l lenbrock, F., and Murphy, G . , (1999) T h e specif ic i ty of T I M P - 2 for matrix meta l lopro te inases can be modif ied by s ingle amino ac id mutat ions. J. Bio.l Chem. 274 20391-20396 Ca lve te , J . J . , M a n n , K., S a n z , L., R a i d a , M. , and Top fe r -Pe te rsen , E. (1996). The primary structure of B S P - 3 0 K , a major l ipid-, gelat in-, and heparin-binding glycoprotein of bovine seminal p lasma. FEBS Lett. 399 147-152 Col l ier, I. E., Wi lhe lm, S . M., E i sen , A . Z . , Marmer, B. L , Grant, G . A . , Sel tzer, J . L , Kronberger , A . , He , C . S . , Bauer , E. A . , and Go ldbe rg , G . I. (1998). H-ras oncogene-t ransformed human bronchial epithelial cel ls (TBE-1) secrete a 66 single metal loprotease capable of degrading basement membrane co l lagen. J. Biol. Chem. 263 6579-6587 Coll ier, I. E., Krasnov, P. A. , . Strongin, A . Y , B i rkeda l -Hansen, H., and Goldberg , G . I., (1992). A lanine scanning mutagenes is and functional analys is of the fibronectin col lagen-binding domain from human 92 -kDa type IV co l lagenase. J. Biol. Chem. 267 6776-6781 Constant ine, K. L , Madr id , M. , Banya i , L , Trexler, M. , Patthy, L. and Ll inas, M. (1992) Re f ined solut ion structure and l igand-binding propert ies of P D C - 1 0 9 domain b. A col lagen-binding type II domain. J. Mol. Biol. 223 281-298 C o s t i g a n , M. , C h a m b e r s , D. A . and Boot -Handfo rd , R. P. (1995). C o l l a g e n turnover in renal d i sease . Exp Nephrol 3 114-121 Dav ies B., W a x m a n J . , W a s a n H., Abe l P. , Wi l l iams G . , K r a u s z T., Nea l D., T h o m a s D., Hanby A . and Balkwil l F. (1993) Leve ls of matrix metal loproteases in bladder cancer correlate with tumor grade and invasion. Cancer Res. 53 5365-5369 E s c h , F. S . , L ing , N. C , Boh len , P. , Y i n g , S . Y . , and Gu i l l emin , R. (1983). Pr imary structure of P D C - 1 0 9 , a major protein const i tuent of bov ine semina l p lasma. Biochem. Biophys. Res. Commun. 113 861-867 F o n t a n a , A . and D e l m a s , P . D . (2000). Marke rs of bone turnover in bone metastases. Cancer 88 2952-2960 Fukush ima, D., Butzow, R., Hi ldebrand A. , and Ruoslaht i , E. (1993). Local izat ion of t ransforming growth factor beta binding site in betag lycan. Compar i son with smal l extracellular matrix proteoglycans. J. Biol. Chem. 268 22710-22715 67 Hay, E .D. (1991) Ce l l Biology of Extracel lular Matrix (2 n d ed.), pp 124-127. New York: P len ium Press . Ishizaki, J . , Hanasak i , K., H igashino, K., K ish ino, J . , K ikuchi , N., Oha ra , O. , and Ar i ta , H. (1994). Mo lecu la r c lon ing of pancreat ic group I phospho l ipase A 2 receptor. J. Biol. Chem. 269 5897-5904 Itoh, T., Tan ioka , M., Y o s h i d a , H., Y o s h i o k a , T., Nishimoto, H., and Itohara, S . (1998). R e d u c e d ang iogenes is and tumor progress ion in gelat inase A-def ic ient mice. Cancer Res. 58 1048-1051 Kleiner, D. E. and Ste t le r -Stevenson, W . G . , (1999). Matrix metal loprote inases and metastasis. Cancer Chemother Pharmacol 43 S42-S51 McMu l l en , B. A . and Fuj ikawa, K., (1985). Amino ac id s e q u e n c e of the heavy chain of human alpha-factor XI la (activated H a g e m a n factor). J. Biol. Chem. 260 5328-5341 M c Q u i b b a n , G . A . , G o n g , J . H., Tarn, E. M., M c C u l l o c h , C . A . , C la rk -Lew is , I., and Overa l l , C M . (2000). Inflammation dampened by gelat inase A c leavage of monocyte chemoattractant protein-3. Science 289 1202-1206 M i y a z a w a , K., S h i m o m u r a , T., K i tamura , A . , K o n d o , J . , Mor imoto , Y . , and Ki tamura, N. (1993). Molecular cloning and sequence ana lys is of the c D N A for a human ser ine pro tease reponsib le for act ivat ion of hepatocyte growth factor. Structural similarity of the protease precursor to blood coagulat ion factor XII. J. Biol. Chem. 268 10024-10028 Morgan, D. O. , E d m a n , J . O , Standr ing, D. N., Fr ied, V . A . , Smi th , M. C , Roth, R. A . , a n d Rut ter W . J . (1987) Insul in growth factor II receptor a s a multifunctional binding protein. Nature 329 301-307 6 8 Morgunova, E., Tuutti la, A . , Bergmann , U., Isupov, M., Lindqvist, Y . , Schne ider , G . , and Tryggvason K. (1999). Structure of human pro-matrix metal loproteinase-2: activation mechan ism revealed. Science 284 1667-1670 Moscatel l i , D (1987). High and low affinity binding sites for bas ic fibroblast growth factor on cultured cel ls: absence of a role for low affinity binding in the stimulation of p lasminogen activator production by bovine capil lary endothel ial cel ls . J. Cell. Physiol. 131 123-130 Murphy, G . , Nguyen, Q. , Cockett , M. I., Atk inson, S . J . , A l lan , J . A . , Knight, C . G . , Wi l lenbrock, F., and Docher ty , A . J . (1994). A s s e s s m e n t of the role of the f ibronectin domain of ge la t inase A by ana lys is of a delet ion mutant. J. Biol. Chem. 269 6632-6636 Murphy, G . and Knauper , V . (1997). Relat ing matrix metal loproteinase structure to function: why the "hemopexin" domain? Matrix Biol. 15 511-518 N a g a s e , H and W o e s s n e r J . F., (1999) Matrix Metal loprote inases and T I P M s , pp 100-101. New York: Oxford University P ress . Nomura , H., Fuj imoto, N., Se ik i , M. , Ma i , M. and O k a d a , Y . (1996). E n h a n c e d p r o d u c t i o n of ma t r i x m e t a l l o p r o t e i n a s e s a n d a c t i v a t i o n of mat r ix metal loproteinase 2 (gelat inase A) in human gastr ic ca rc inomas . Int. J. Cancer. 6: 9-16 Ove ra l l , C . M. and S o d e k , J . , (1987). Initial charac te r iza t ion of a neutral metal loproteinase, active on native 3/4-col lagen fragments, syn thes ized by R O S 17/2.8 os teob las t i c ce l l s , per iodonta l f ibroblasts , and ident i f ied in g ingival crevicular fluid. J. Dent. Res. 6: 1271 -1282 69 Overa l l , C . M. and S o d e k , J . , (1990). C o n c a n a v a l i n A p roduces a matrix-degrada t i ve pheno type in human f ib rob las ts . Induct ion a n d e n d o g e n o u s activation of co l lagenase, 72 -kDa gelat inase, and Pump-1 is accompan ied by the suppress ion of the t issue inhibitor of matrix metal loproteinases. J. Biol. Chem., 265 2 1 1 4 1 - 2 1 1 5 1 Overa l l , C . M. , W r a n a , J . L. and Sodek , J . , (1991). Induction of formative and resorptive cel lular phenotypes in human gingival f ibroblasts by T G F - b e t a 1 and concanaval in A : regulation of matrix metal loproteinases and T I M P . J. Periodontal Res. 26 279-282 Overa l l , C . M. , W r a n a , J . L. and S o d e k , J . (1991) Transcr ip t iona l and post-t ranscr ip t iona l regu la t ion of 7 2 - k D a ge la t i nase / t ype IV c o l l a g e n a s e by t ransforming growth factor -beta 1 in human f ibrob lasts . C o m p a r i s o n s with co l lagenase and t issue inhibitor of matrix metal loproteinase gene express ion . J. Biol. Chem. 266 14064-14071 Overa l l , C . M., K ing , A . E. , S a m , D. K., O n g , A . D., Lau , T. T., Wa l lon , U. M. , DeClerck , Y . A . , and Atherstone J . (1999) Identification of the t issue inhibitor of metal loproteinases-2 (TIMP-2) binding site on the hemopex in carboxyl domain of human ge la t inase A by s i te-d i rected mu tagenes is . T h e h ierarch ica l role in binding T I M P - 2 of the unique cationic clusters of hemopex in modu les III and IV. J. Biol. Chem. 274 4421-4429 Overal l , C . M. , K ing , A . E. , B igg , H. F., McQu ibban , G . A . , Atherstone, J . , S a m , D. K., O n g , A . D., Lau , T. T.,. Wal lon , U. M, DeClerck , Y . A . , and E. Tarn (1999) Identification of the T IMP-2 binding site on the gelat inase A hemopexin C-domain 70 by site-directed mutagenes is and the yeast two-hybrid sys tem. Ann. N. Y. Acad. Sci. 878 747-753 O w e n s , R. J . & Baral le, F. E. (1986) Mapping the col lagen-binding site of human fibronectin by express ion in Escher ich ia col i . E M B O J 5: 2825-2830 Parks , W . C . and M e c h a m , R. P (1998) Matrix Meta l loprote inases, pp 85-105. New York: Academ ic P ress . Pickford, A . R., Potts , J . R., Bright, J . R., P h a n , I., and C a m p b e l l , I. D. (1997). Solu t ion structure of a type 2 modu le from f ibronect in: impl icat ions for the structure and function of the gelatin-binding domain. Structure 5 359-370 Prockop, D. J . and Kivirikko, K. I. (1995). Co l lagens : molecular biology, d i seases , and potentials for therapy. Annu. Rev. Biochem. 64 403-434 Roopra i , H. K., V a n Meter, T., Ruck l idge, G . J . , Hudson , L , Evera l l , I. P. and Pi lk ington, G . J . (1998). Compara t i ve ana lys is of matrix meta l loprote inases by immunocytochemistry, immunohistochemistry and zymography in human primary brain tumours. Int. J. Oncol. 13 1153-1157 Sato, H., Tak ino, T., Kinoshi ta, T., Imai, K., O k a d a , Y . , Ste t ler -Stevenson, W. G . , and Seik i , M. (1996) Ce l l surface binding and activation of gelat inase A induced by express ion of membrane-type-1 -matrix metal loproteinase ( M T 1 - M M P ) . FEBS Lett. 385 238-240 S c h w a r z b a u e r , J . E. and Sech le r , J . L. (1999). F ibronect in f ibr i l logenesis: a paradigm for extracellular matrix assembly . Curr. Opin. Cell. Biol. 11: 622-627 Ship ley, J . M. , Doyle, G . A. , F l iszar, C . J . , Y e , Q . Z . , J o h n s o n , L. L , Shap i ro , S . D., We lgus , H. G . and Senior , R. M. (1996) The structural bas is for the elastolytic 71 activity of the 9 2 - k D a and 7 2 - k D a ge la t inases. Ro le of the f ibronectin type II repeats. J. Biol. Chem. 271 4335-4341 Ste f fensen, B., Wa l l on , U. M. , and Overa l l , C . M. (1995). Extracel lu lar matrix binding propert ies of recombinant f ibronectin type II modu les of human 72 -kDa gelat inase/type IV co l lagenase. High affinity binding to native type I co l lagen but not native type IV co l lagen. J. Biol. Chem. 270 11555-11566 Ste f fensen, B., B igg , H. F., and Overa l l , C . M. (1998). The involvement of the fibronectin type II modules of human gelat inase A in cell sur face local izat ion and activation. J. Biol. Chem. 273 20622-20628 Stef fensen, B., Maurus, R., Ryberg, E., and Overa l l , C . M. (2001), in preparation. Sticht, H., P ick ford , A . R., Potts, J . R., and C a m p b e l l , I. D. (1998). Solut ion structure of the g lycosylated second type 2 module of f ibronect in. J. Mol. Biol. 276 177-187 Stil l , K., R o b s o n , C . N., Au tzen , P. , Rob inson , M. C , and Hamdy, F. C . (2000). Local izat ion and quantif ication of m R N A for matrix metal loproteinase-2 (MMP-2) and t issue inhibitor of matrix metal loproteinase-2 (TIMP-2) in human benign and malignant prostatic t issue. Prostate 42 18-25 Streul i , C . (1999). Extracel lu lar matrix remodel l ing and cel lu lar differentiation Curr. Opin. Cell Biol. 11 634-640 Taylor , M . E. , Cona ry , J . T., Lennar tz , M. R., S tah l , P. D., and Dr ickamer, K. (1990) Pr imary structure of the m a n n o s e receptor con ta ins mult iple motifs resembl ing carbohydrate-recognit ion domains. J. Biol. Chem. 265 12156-12162 Torda i , H. and Patthy, L. (1999). The gelat in-binding site of the s e c o n d type-l l domain of gelat inase A / M M P - 2 . Eur. J. Biochem. 259 513-518 72 Zubay, G . , Pa rson , W. , and V a n c e , D. (1995) Pr inc ip les of Biochemistry, pp 79-81. Dubuque, IA: W m C . Brown Publ ishers. 73 

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.831.1-0090113/manifest

Comment

Related Items