UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Coagulation factor V : pathology and biochemistry Song, Jina 2008

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

Item Metadata

Download

Media
24-ubc_2008_fall_song_jina.pdf [ 2.51MB ]
Metadata
JSON: 24-1.0067024.json
JSON-LD: 24-1.0067024-ld.json
RDF/XML (Pretty): 24-1.0067024-rdf.xml
RDF/JSON: 24-1.0067024-rdf.json
Turtle: 24-1.0067024-turtle.txt
N-Triples: 24-1.0067024-rdf-ntriples.txt
Original Record: 24-1.0067024-source.json
Full Text
24-1.0067024-fulltext.txt
Citation
24-1.0067024.ris

Full Text

COAGULATION FACTOR V: PATHOLOGY AND BIOCHEMISTRY by Jina Song B.Sc., Simon Fraser University, 2002  A THESIS SUBMITTED N PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES  (Pathology)  THE UNWERSITY OF BRITISH COLUMBIA (Vancouver) August 2008  ©Jina Song, 2008  ABSTRACT ROLE OF GLU96, ASP1O2 AND ASP111 IN FACTOR V  Coagulation factor V activity is unmasked by thrombin-mediated excision of the central B domain resulting in a noncovalent heterodimer, factor Va. To understand the role of individual amino acids in maintaining the Ca -dependent 2 subunit interaction, G1u96 (E96A), AsplO2 (D1O2A) or Asplil (D111A) were mutated because of known effects on chelator sensitivity. The primary clotting activity of each mutant was reduced by “40%. Demonstrating at least two distinct inhibition mechanisms, only D111A was further inhibited by thrombin pre-treatment consistent with spontaneous subunit dissociation and severely inhibited Ca 2 binding. The parental factor V construct used here has a truncated B domain that does not require excision for activity. Therefore inhibition of D111A by thrombin-cleavage reveals a new B domain function that maintains factor V in a factor Va-like configuration independent of Ca 2 binding. In addition to Ca , factor V binds Cu 2 , but with unknown function. Unexpectedly, D111A 2 also lost detectable Cu t Finding that a single amino acid substitution 2 simultaneously alters Ca 2 and Cu 2 suggests an interdependent metal ion binding site. Unlike D111A, the thrombin-mediated factor Va derived from E96A and D1O2A was stable, had only moderately faster subunit dissociation upon chelation and had normal metal ion binding. Thus, the current study defines the highly conserved acidic segment spanning G1u96-Aspll2 in factor V as multifunctional. Of the three amino acids I evaluated, Asplil is essential and  11  likely functions through direct and indirect metal ion interactions. G1u96 and AsplO2 individually influence factor V/Va function by more subtle effects at the metal ion-dependent subunit interface. FACTOR V-DEFICIENT PATIENT  A factor V-deficient patient due to Y1702C mutation has been studied. The patient however did not suffer from severe bleeding despite of undetectable levels of plasma and platelet factor V. A close inspection of the patient’s blood coagulation cascade showed that the lack of available factor V was compensated by other factors that influence the intrinsic pathway. This finding suggests that the commonly observed phenotypic differences shown among factor V-deficient patients with the same genotypes may be due to existing hypercoagulant factors that influence the outcome of the disease.  111  TABLE OF CONTENTS Astr  ii iv vii  ......................................................  1abIe of Contents ....................... . ............. . List of Tables  ...........  List of Figures Abbreviations . .. .. ........ AcknoA,ledgernents  p......  ....•..................  p.  .........u  Viii X  ...........xiv  1.1 Hemostasis 1.2 Blood coagulation 1.2.1 Activation and propagation 1.2.2 Prothrombinase complex 1.2.3 Factor X/Xa 1.2.4 Prothrombin/thrombin 1.2.5 Anionic phospholipids 1.3 Factor V/Va 1.3.1 Primary structure 1.3.2 Activation 1.3.3 Interaction with anionic phospholipids 1.3.4 Platelet factor V/Va 1.3.5 Cofactor function in prothrombinase 1.3.6 Mechanism of inhibition 1.3.7 Tertiary structure 1.4 Role of metal ions in factor V/Va 1.4.1 Essential metal ions 1.4.2 Calcium 1.4.3 Copper 1.4.4 Calcium and copper in factor V/Va 1.4.5 Metal-dependent subunit association 1.5 Factor V pathology 1.5.1 Factor V deficiency 1.5.2 Regulatory defects 1.6 Scope of this thesis  1 1 2 2 5 6 8 10 11 11 14 15 16 17 19 20 23 23 23 24 25 26 33 33 34 36  1ateriaIs and riiethods 2.1 Materials 2.2 Proteins 2.3 Molecular biology of factor V 2.3.1 Protein expression 2.3.2 Site-directed mutagenesis 2.3.3 Stable expression of FV-810 and mutants 2.3.4 Purification of FV-810 and mutants  38 38 39 39 39 42 44 45  1. Introduction  2.  ............  iv  2.4 Thrombin-mediated cleavage 2.5 Binding to anionic phospholipid vesicles 2.6 Activity assay 2.6.1 Prothrombinase assay 2.6.2 Clotting Assay 2.7 Effect of chelation 2.7.1 Chelator-induced subunit dissociation 2.7.2 Chelator-induced cofactor activity loss 2.8 Metal analysis 2.8.1 Inductively coupled plasma-mass spectrometry 2.8.2 Graphite furnace atomic absorption spectrometry 2.9 Factor V-deficient patient study 2.9.1 Blood sample preparation 2.9.2 Isolation of DNA from leukocytes 2.9.3 DNA sequencing 2.9.4 Functional factor V level in plasma and platelet 2.9.5 Immunological factor V level in plasma and platelets 2.9.6 Investigation of blood coagulation pathway 2.9.7 Investigation of intrinsic pathway  .46 46 48 48 48 49 50 50 50 50 51 52 52 54 54 54 55 56 58  3. Role of G1u96, AsplO2 and Aspill in Factor V 3.1 Hypothesis and specific goals 3.2 Results 3.2.1 Mutagenesis and production of factor V 3.2.2 Thrombin-mediated cleavage 3.2.3 Binding to anionic phospholipid vesicle 3.2.4 Cofactor function 3.2.5 Subunit dissociation 3.2.6 Chelator-induced cofactor activity loss 3.2.7 Metal ion measurement 3.3 Discussion  60 60 60 60 62 64 67 67 72 73 77  .......................  4. Factor V—deficient patient ......... .. .. .. ..... ..... ............. 91. 4.1 Hypothesis and specific goals 91 4.2 Results 93 4.2.1 Plasma and platelet preparation 93 4.2.2 Identification of mutation 93 4.2.3 Factor V activity in plasma and platelets 93 4.2.4 Factor V antigens in plasma and platelets 95 4.2.5 Compensation by blood coagulation pathway 95 4.2.6 Blood proteins in intrinsic pathway 100 4.3 Discussion 100 5. Surrlrilary  ....  ...............  107  V  5.1 Role of G1u96, AsplO2 and Asplil in factor V 5.2 Factor V-deficient patient References  .  ........  110 111 113  vi  LIST OF TABLES  Table 1.1 Predicted Ca 2 and Cu 2 coordinating amino acid residues  27  Table 2.1 Conditions used for PCR for mutagenesis  43  Table 4.1 Intrinsic/contact phase pathway blood protein activity  101.  vii  LIST OF FIGURES Figure 1.1 Blood coagulation  3  Figure 1.2 Human FX/FXa  7  Figure 1.3 Human prothrombin/thrombin  9  Figure 1.4 Human FV/FVa  12  Figure 1.5 FV A domain homology model  28  Figure 1.6 FVa homology model  30  Figure 1.7 Bovine FVa structure  31  Figure 2.1 Functional map of plasmid pMT2-rFVdt  41  Figure 3.1 Homologous region spanning 96-112 in the Al domain in FV, FVIII and ceruloplasmin from various species  61  Figure 3.2 DNA sequencing of pMT2 containing FV-810 and its variants  63  Figure 3.3 Thrombin-mediated conversion of FV-810 and mutants  65  Figure 3.4 Binding of FV-810 and mutants to anionic phospholipid vesicles.. 66 Figure 3.5 Prothrombinase activity of FV-810 and mutants  68  Figure 3.6 Clotting activity of FV-810 and mutants  69  Figure 3.7 Subunit interaction of FV-810 and mutants  70  Figure 3.8 Western blot analysis of FV-810 subunit dissociation  72  Figure 3.9 Loss of cofactor activity of thrombin-cleaved FV-810 and mutants upon chelation  74  Figure 3.10 ICP-MS standard curves  76  Figure 3.11 GFAAS standard curves  78  vi”  Figure 3.12 Bound Ca 2 and Cu 2 in FV-810 and mutants  .79  Figure 3.13 Effects of RI G1u96, AsplO2 or Asplil mutation on metal binding  and subunit association  81  Figure 4.1 Pedigree of plasma FVa Activity  92  Figure 4.2 DNA sequencing  94  Figure 4.3 FV clotting activity in plasma and platelets  96  Figure 4.4 FV/FVa antigens in plasma or platelets  97  Figure 4.5 Compensation by blood coagulation pathway  99  Figure 4.6 Conserved Tyr1702 and Pro1618 in the A domains  103  Figure 4.7 Conserved Tyr1702 in the A domains  104  Figure 4.8 Role of Tyr1702  105  Figure 4.9 Tyr1702 and nearby cysteine residues  107  Figure 5.1 Model: Interdependent metal ion binding facilitates subunit  interaction  112  ix  ABBREVIATIONS  APC: activated protein C aPL: anionic phospholipids APMSF: 4-amidinophenylmethane-sulfonyl fluoride hydrochloride app Kd: apparent dissociation constant APTT: activated partial thromboplastin time BHK: baby hamster kidney Da: Dalton which is a small unit of mass used to express atomic and molecular masses. It is defined to be one twelfth of the mass of an unbound atom of 12 C at rest and in its ground state D-MEM/F12: Dulbecco’s modified Eagle’s medium/F- 12 DII: dithiothieitol ECL: enhanced chemilumi nescence reagent EDTA: ethylenediamine tetraacetic acid CI: phenylalanyl-L-prolyl-L-arginine chloromethyl ketone 2 FPR-CH Fil: prothrombin Flia: thrombin  FV: factor V FVa: activated factor V FVaH: factor Va heavy chain FVaL: factor Va light chain FVa: inactivated factor Va  x  Fy11: factor VII FVIIa: activated factor VII FVIII: factor VIII FVIIIa: activated factor VIII FIX: factor IX FIXa: activated factor IX FX: factor X FXa: activated factor X FXI: factor XI  FXIa: activated factor XI FXII: factor XII FXIIa: activated factor XIIa FXIII: factor XIII FXIIIa: activated factor XIII GFAAS: graphite furnace atomic absorption spectrometry Gla: gamma-carboxyglutamic acid HBS: hepes buffered saline Hepes: 4-(2-hydroxyethyl)- 1-pi perazi neethanesulfonic acid HK: high-molecular weight kininogen ICP-MS: inductively coupled plasma-mass spectrometry ITS: insulin-transferrin-sodium selenite supplement Kb: kilobase  xi  Kcat: first-order rate constant, also called the turnover number  Kd: dissociation constant kDa: kilodalton 1Km: Michaelis constant LB: Lauria-Bertani LMV: large multilamellar vesicle mAb: monoclonal antibody M: Molarity or molar concentration which denotes the number of moles of a given substance per litre of solution Mr: molecular mass which is the mass of one molecule of that substance relative  to the unified atomic mass unit u (equal to 1/12 the mass of one atom of 12 C) PC: phosphatidyicholine PC: protein C PCR: polymerase chain reaction PE: phosphatidylethanolamine PEG 8000: polyethylene glycol with Mr 8000 PKK: prekallikrein PPP: platelet-poor platelet PRP: platelet-rich platelet PS: phosphatidylserine PT: prothrombin time PVDF: polyvinylidine difluoride  xii  SDS-PAGE: sodium dodecyl sulphate polyacrylamide gel electrophoresis SUV: small unilamellar vesicle TBST: Tris buffered saline plus 0.5% Tween-20 TF: tissue factor  xiii  ACKNOWLEDGEMNETS  I would like to first thank my supervisor Dr. Ed Pryzdial for his support and patience. In addition I would like to thank the following people for their contributions to this thesis: Fellow laboratory members, Kimberley Talbot, Scott Meixner, Dr. Michael Sutherland, Amanda Vanden Hoek, Edwin Gershom and Ji Ae Jun for assistance and helpful discussions. My supervisory committee, Dr. Ross MacGillivray, Dr. Cedric Carter and Dr. Cohn Fyfe for guidance and useful suggestions. Lastly, my family and friends for being there for me at times when I needed support.  xiv  1. INTRODUCTION 1.1 HEMOSTASIS  Hemostasis is the body’s normal physiological control of bleeding by blocking any vascular opening to ensure blood fluidity and blood vessel integrity [1]. Intact blood vessels normally constitute a non-leaking system achieved by structural integrity of the vessel wall with endothelial cells playing the major role. The endothelial cells are also important in preventing inappropriate adherence and activation of platelets and the coagulation cascade [2]. Upon vascular damage, hemostasis is responsible for changing blood from a fluid to solid state to arrest bleeding from damaged blood vessels rapidly in a localized fashion. The overall events include complex interactions between blood vessels, platelets and coagulation factors that result in the production of a clot at the site of injury and a subsequent clot removal by fibrinolysis [3]. The balance between the formation and the removal of a clot is essential to maintain proper blood flow. Any upset in this balance can result in either bleeding (hemorrhage) or unnecessary blood clot formation (thrombosis) [2]. Hemostasis consists of three distinct phases (primary hemostasis, blood coagulation and fibrinolysis) that are interrelated and strictly regulated in order to successfully close vessel wounds and promote vascular healing [2]. Primary hemostasis involves vessel constriction caused by temporary local contraction of vascular smooth muscle. This slows blood flow to the injury site allowing the formation of a platelet plug to initially block the vessel opening. This short-lived  and unstable plug is solidified through a series of reactions collectively referred to as blood coagulation, transforming soluble fibrinogen into insoluble fibrin. The third phase of hemostasis involves fibrinolysis to finally remove fibrin deposits after damaged vessel walls have been restored [2,4].  1.2 BLOOD COAGULATION 1.2.1 Activation and propagation  Blood coagulation is the result of a cascade of reactions involving clotting factors as inactive precursors referred to as zymogens that are activated by limited proteolysis by the functional factor immediately preceding them (Figure 1.1). Many of the active forms are serine proteases that act on a further inactive precursor. The ultimate goal of blood coagulation is the formation of insoluble fibrin fibers by thrombin [5]. All components necessary for the initiation of coagulation except for tissue factor (TF) and anionic phospholipid (aPL) are found in plasma [2]. Injury triggers the initiation of the coagulation cascade by exposing TF to blood allowing the subsequent binding of activated factor VII (FVIIa). Although factor VII (Fy11) is activated by its product factor Xa (FXa), a trace amount of FVIIa appears to be available in plasma at all times to interact with TF [6]. The FVIIa TF complex activates the zymogens, factor IX (FIX) and factor X (FX) to the protease, activated factor IX (FIXa) and FXa. The limited amount of FXa produced  generates thrombin  that activates platelets and  cleaves the  2  • • • • • .•  . . .  I’  H  Ii  H IUTTI’1 II  INTRINSIC PATHWAY  EXTRINSIC PATHWAY  1ii  ri  ii  Il  PROTHROM BINASE  4, Fibrin Clot Figure 1.1 Blood coagulation. This diagram summarizes the blood coagulation pathway. Blood proteins are shown in circles. Blood coagulation involves two distinct upstream pathways, extrinsic and intrinsic. After initiation, they converge to form the prothrombinase complex. Since thrombin generation is essential for stable hemostasis, the importance of all individual components that makes up the prothrombinase complex including FV can be stressed.  3  procofactors factor V (R/) and factor VIII (FVIII) [7,8]. FIXa and FXa, with their respective cofactors activated factors VIII (FVIIIa) and V (FVa), form the tenase and prothrombinase complexes that activate FX and prothrombin respectively on the surface of activated platelets [9]. This leads to the local generation of a significant amount of FXa and thrombin. The resulting thrombin further amplifies  its own formation by activation of platelets, RI, FVIII and factor XI (FXI) [2]. The activated factor XI (FXIa) enhances FIXa generation, which activates FX, therefore enforcing thrombin generation. Thrombin cleaves fibrinogen [10] and factor XIII (FXIII) to form a cross-linked fibrin clot [11]. Alternatively to the TF-dependent activation of coagulation traditionally referred to as the extrinsic pathway, coagulation can also be initiated via the contact phase of activation leading to the intrinsic pathway when factor XII (FXII), prekallikrein (PKK) and high-molecular weight kininogen (HK) bind to kaolin, glass or another artificial surface. When bound, activation of FXII and prekallikrein occurs yielding activated factor FXII (FXIIa) and kalllikrein (KK), respectively. FXIIa triggers coagulation via the sequential activation of FXI, FIX, FX and prothrombin [2]. Activation of FXII is not required for hemostasis, since patients with deficiency of FXII, prekallikrein or HMWK do not bleed. However, patients with FXI deficiency tend to have a mild bleeding disorder implying that FXI is involved in hemostasis [12]. Historically, two pathways have been considered to be important in the conversion of fibrinogen to fibrin, the intrinsic and extrinsic pathways converging  4  to activate the common pathway [13,14]. The intrinsic pathway was considered for many years to be the primary initiator of the coagulation cascade. However, it is now known that the major pathway involved in coagulation initiation is the extrinsic pathway, which is rapidly activated following trauma. The intrinsic pathway is thought to maintain the coagulation process while tissue regeneration occurs. Although the concept of the intrinsic and extrinsic pathways was used for many years to separately model blood coagulation, more recent evidence has shown that the pathways are highly interconnected. As discussed, the TF-FVIIa complex activates both FX and FIX. Also, patients with severe Fy11 deficiency bleed even though the intrinsic pathway is intact. Likewise, the severe bleeding associated with deficiencies of FVIII or FIX would not be expected if the extrinsic pathway alone was sufficient for hemostasis.  1.2.2 Prothrombinase complex FXa and its cofactor FVa assemble on cell membranes containing aPL in  the presence of Ca 2 to form a prothrombinase complex that is responsible for converting prothrombin to thrombin [15]. As shown in Figure 1.1, the prothrombinase complex is located at the convergence of the intrinsic and the extrinsic pathways. Activated platelets and endothelial cells expose aPL on their membrane surface providing a site of assembly for the prothrombinase complex [16-18]. FXa and prothrombin interact with the membrane through their amino terminal domains that contain y-carboxyglutamic acid (Gla) residues. These Gla  5  residues require Ca 2 to bind to aPL-containing membranes [19]. Unlike FXa and prothrombin, FVa binds to membranes independent of the presence of Ca 2 [20,21]. It has been shown that the initial formation of separate FXa:aPL and FVa:aPL is kinetically favorable for complex formation. The membrane-bound FXa and FVa then rapidly rearrange on the membrane surface to form an enzyme:cofactor complex [22,23] with FVa increasing the affinity for FXa:aPL interaction by 100-fold [15,24]. Although FXa can activate prothrombin directly in the absence of FVa, aPL and Ca , the rate of activation is very slow and does 2 not achieve the immediate physiologic response required to arrest blood loss. The rate of thrombin activation by the prothrombinase is about 300000 times higher than by FXa alone [25,26], which is essential for hemostasis.  1.2.3 Factor X/Xa  The blood coagulation serine protease, FXa, circulates as a vitamin K-dependent zymogen, FX, at a concentration in human plasma of 170 nM with an Mr of 58000 with a heavy (Mr 42000) and a light (Mr 16000) chain covalently associated through a disulfide bond [27,28] (Figure 1.2). Vitamin K is necessary for proper post-translational carboxylation of FX to produce Gla residues that allow  binding  of Ca 2  to  adhere  to  cell  membranes  with  exposed  phosphatidylserine (PS) [29,30]. Ca 2 binding to the Gla domain induces a large structural change from an unfolded and non-functional domain to the functional one that is tightly folded and able to bind to cell membranes [31]. The active  6  FX  IctJ  rptease’  Tenase  FXa  I  Figure 12 Human FX/FXa.  The activation of human FX to FXa involves the excision of an activation peptide (Act) from the NH -terminal fragment of the heavy chain between Arg5l and 2 11e52. The light chain (red) contains Gla domain residues and the heavy chain (green) contains the serine protease domain.  7  enzyme FXa (Mr 48000) is produced by the excision of a carbohydrate-rich activation peptide (Mr 12000) from the NH -terminal fragment of the heavy chain 2 between Arg5l and 11e52 [32]. The light chains of FX and FXa contain the Gla domain and the heavy chains of FX and FXa contain the serine protease domain [33]. FX participates in the common phase blood coagulation cascade at the juncture of the intrinsic and extrinsic pathways [2]. In the extrinsic pathway, it is cleaved by FVIIa in the presence of TF [34], which triggers coagulation after injury. In the intrinsic pathway, FX is cleaved to FXa by FIXa in the presence of FVIIIa, Ca 2 and aPL [35,36], which amplifies thrombin production after initiation  of coagulation. Although FXa is the only endogenous prothrombin activator, physiologically  substantial  thrombin  production  occurs  only  within  the  prothrombinase complex. There are two predominant inhibitors that control FXa in blood, the tissue factor pathway inhibitor and antithrombin [35,37-39].  1.24 Prothrombin/thrombin The final key serine protease in the blood coagulation cascade, thrombin, circulates in blood as a vitamin K-dependent zymogen, prothrombin, at a concentration of 1.4 pM with an Mr of 72000 [40]. Activation of prothrombin to thrombin results from proteolytic cleavages at Arg271/Thr272 and Arg320/11e321 by the prothrombinase complex [41] (Figure 1.3). Depending on the order of peptide bond cleavage, two intermediate products, meizothrombin  8  Prothrombin  s—s Arg271 /Thr27_’  Fragment 1.2  Prethrombin-2  Arg32O/lle27’>’\  \.<r32Oil 1e272  Meizothrombin  /<rg271/Thr272  Figure 1.3 Human prothrombin/thrombin.  The activation of human prothrombin to cc-thrombin involves the cleavage of two peptide bonds. The cleavage at Arg271/Thr272 followed by the cleavage at Arg320/11e321 yields prethrombin-2 and fragment 1.2 as intermediates. The cleavage at Arg320/1le321 followed by the cleavage at Arg271/Thr272 yields meizothrombin as an intermediate.  9  (Arg320/11e321 cleavage first) and prethrombin-2 (Arg271/Thr272 cleavage first) can exist [42,43]. Meizothrombin is the main intermediate If activated by the prothrombinase complex whereas prethrombin-2 is the major intermediate observed without the cofactor, FVa. The same final products, fragment 1.2 and thrombin are produced from both intermediates [44]. Alpha-thrombin, the principal product from the activation of prothrombin following injury, is a serine protease heterodimer with an Mr of 37000, composed of an A chain (Mr 6000) and a B chain (Mr 31000) [41]. The B chain contains the serine protease domain. Thrombin is multifunctional in our body capable of playing roles as both a procoagulant and an anticoagulant [39,45]. When sufficient amount of thrombin has been generated, thrombin can bind to thrombomodulin and act as an anticoagulant through the activation of protein C (PC) to shut off blood coagulation [46].  1.2.5 Anionic phospholipid  The exposure of aPL is restricted to sites of vascular damage, where cells are triggered to flip aPL, mainly phosphatidylserine (PS) from the inner to the outer leaflet of the plasma membrane [2], strictly localizing the assembly of prothrombinase complex. Activation of platelets, monocytes and endothelial cells are considered to be the most important physiological sources of aPL, of which platelets constitute the major source that is trapped directly within a fibrin clot [47]. In the case of the prothrombinase, aPL accelerates thrombin generation by  10  providing a site of association for FXa, FVa and prothrombin in the presence of 2 [18], thereby increasing the reactant concentrations [48]. Binding to aPL Ca also enhances catalytic efficiency [15]. In addition to a net negative charge, the polar head group composition of the phospholipid membrane has been shown to affect procoagulant activity. The membranes of coagulation-facilitating cells are composed mainly of PS, phosphatidylcholine (PC) and phosphatidylethanolamine (PE) [49]. The content of PS has been shown to directly affect the procoagulant quality of the membrane [50]; therefore, the majority of studies that require the assembly of coagulation proteins have used membranes composed of 75% PC and 25% PS [21]. In addition to PS, FVa inactivation by activated protein C (APC) is shown to also require PE [51-53]. Thus, subtle phospholipid composition changes of membrane  may  send  differential  signals  to  trigger  procoagulant and  anticoagulant reactions.  1.3 FACTOR V/VA  1.3.1 Primary structure The human FV gene is 80 kb in length located on chromosome 1 at q2125 and contains 24 introns giving rise to a 6800 base mRNA upon transcription [54-56]. FVa circulates in blood as a single chain procofactor with a concentration of 20 nM and an Mr of 330000 [23,57] (Figure 1.4). The amino acid sequence consists of a 28 amino acid leader peptide and a mature protein of  11  ha 709  A. FV ..  ..  . ...  .  V  j  Ila 1545 V  Ha 1018 V • •.  SH  ..  HSI..’ 1 S  900 PSDLLLLKQSNSSKILVGRWHLASEKGSYEIIQD TDEDTAVNNWUSPQNASRAWGESTPLANKPG KQSGHPKFPRVRHKSLQVRDGGKSRLKKSQFL IKTRKKKKEKHTHHAPLSPRTFHPLRSEAYNT 1030  B. Thrombin-mediated activation  2ifl C. FVa  IVIe  Figure 1.4 Human FV/FVa. (A) Diagram of the organization of the human FV molecule is shown. The arrows on top represent the activation cleavage sites by thrombin (ha). The positions of posttranslational modifications and free cysteines are shown; =N-linked glycosylation site,. =sulfation site, ¶ =phosphorylation site, SH =free cysteines. Amino acid sequence from 900 to 1030 of the B-domain is shown in a gray box with basic residues highlighted in red. (B) Schematic diagram of R/ cleavages mediated by thrombin is shown. The arrows indicate sequential cleavages. (C) Diagram of human FVa molecule is shown. FVa is a divalent cation (Me)-hinked heterodimer composed of FVaH and FVaL (FVaH, green; FVaL, red). The B domain is represented as a solid black line.  12  2196 amino acids [55]. The domain structure of RI is organized with triplicated A domains, a B domain and duplicated C domains H -A1-A2-B-A3-C1-C2-CO 2 (NH ) [55]. The A and C domains of R/ share homology with the A and C domains of human FVIII. The A domains of FV and FVIII are homologous to a copperbinding protein, ceruloplasmin and the C domains of RI and FVIII are homologous to a slime mold protein, discoidin. The B domain, however, is not well conserved and is released at sites of injury primarily by thrombin. The B domain is large (984 amino acids) and contains 2 tandem repeats of 17 amino acids and 31 tandem repeats of 9 amino acids [58]. It is heavily glycosylated and has no apparent involvement in the cofactor activity of RI. However, it has been shown to serve as a substrate for the transglutaminase activity of FXIIIa [59] and implicated in anticoagulation [60] and thrombin recognition [61]. RI undergoes multiple post-translational modifications including sulfation, phosphorylation and glycosylation (Figure 1.4). Sulfation is shown to be important for thrombin cleavage-mediated activation and function [62,63]. Phosphorylation is demonstrated to be crucial for inactivation by activated protein C (APC) [64]. R/ contains 37 N-linked glycosylation sites of which 25 are located in the B domain [65]. It also contains 19 Cys residues; 14 of 19 Cys are involved in disulfide bridges forming several loops and the remaining five are present as free Cys (two in the heavy chain, one in the B domain and two in the light chain) [66,67].  13  1.3.2 Activation  The procofactor,  RI  circulates in blood with little or no activity. At sites of  injury, FV is activated primarily by thrombin through a series of cleavages at Arg709, ArglOl8 and Arg1545 which results in the excision of a B domain as two fragments, 71 kDa (residues 710-1018) and 150 kDa (residues 1019-1545) in size [68] (Figure 1.4). The cleaved product is a heterodimer comprised of a heavy chain (FVaH:Mr 104000) and a light chain (FVaL:Mr 74000/71000) with noncovalent association requiring divalent cations. The heavy chain is composed of the Al domain (residues 1-317) and the A2 domain (residues 318-663) and the light chain is composed of the A3 domain (residues 1546-1883), the Cl domain (residues 1884-2036) and the C2 domain (residues 2037-2196) [55,69]. The proteolytic sites in human RI and bovine RI are the same although the order of cleavage seems to be different [7,57,70]. The RI activator from Russell’s viper venom cleaves only at two sites, ArglOl8 and Arg1545, yielding an active molecule with two components, 150 kDa and 74/71 kDa in size. The increase in cofactor activity in RI treated with thrombin or an activator for RI from Russell’s viper venom seems to correlate with the generation of a 74/71 kDa component. Therefore the cleavage at Arg1545, which liberates FVaL is thought to be crucial for FVa function [71]. Although the cleavage at Arg 1545 gives the full cofactor activity, how proteolysis directly confers FVa cofactor activity is not yet clear. A FV variant that contains a partially deleted B domain (811-1491) was shown to have full  14  procoagulant activity without requiring proteolysis [72]. This suggests that the B domain may inhibit procoagulant activity by simply masking binding sites on FV for FXa, prothrombin or both. The region within the B domain that may be involved in stabilizing the FV procofactor function has been found to be located within residues 902 to 1033 (Figure 1.4). The region from residues 963 to 1008 contains a cluster of basic amino acids and is well conserved among mammals [73] suggesting its importance in function.  1.3.3 Interaction with anionic phospholipid To localize clot formation to the site of injury, prothrombinase components including FV/FVa bind to cells that expose aPL on their surface associated with damaged vascular tissue. The interaction of FV/FVa with aPL is mediated through FVaL and is Ca 2 independent [50] unlike the Gla-containing proteins. Studies using fragmented bovine FVa have demonstrated that the A3 [74] and the A3/C2 [75] domains interact with aPL while experiments with human FVa using a series of recombinant deletion [76,77], antibodies [78,79] and site-directed mutagenesis [79-81] have implicated the C2 domain in aPL binding. The binding of the protein and the membrane is demonstrated to involve both ionic and hydrophobic interactions, the A3 domain is shown to interact by a hydrophobic mechanism and the C2 domain by an ionic mechanism [75]. Overall, it seems likely that binding of FVa to membrane occurs in two stages; initially there is a diffusion rate-limiting FVa:membrane interaction  15  mediated through the C2 domain followed by a hydrophobic contribution through the A3 domain to insure a stable FVa:aPL interaction [67]. Estimates of the dissociation constant for the interaction of FVa and aPL-containing vesicles vary depending on the methodologies used: 10-11 M by kinetic studies [82], i0 M by equilibrium binding measurements [21], 10-8 M by nonequilibrium sedimentation [83] and i0 M by light scattering [20]. The dissociation constant for FVa to platelets is shown to be 10b0 M [84].  1.3.4 Platelet factor V/Va  Approximately 20 % of FV contained in whole blood is found in the agranules of platelets [85] where it is complexed to multimerin 1, and secreted upon platelet activation bound to microparticles [86]. Unlike plasma FV which is synthesized in liver [87-90], the origin of platelet FV is unclear; however, it is believed to be from both plasma uptake and megakaryocytic synthesis [91]. Platelet  RI  is stored and released as a partially proteolyzed molecule with  significant cofactor activity prior to activation by FXa or thrombin. While human plasma RI undergoes a 25-to 30-fold increase in cofactor activity following full activation by thrombin [7], the cofactor activity of human platelet FV is increased only 1- to 10-fold, depending upon the method of platelet RI release [92-96]. Thrombin activation of platelet RI yields a cleaved product similar to thrombin cleaved plasma RI [97]. RI from platelets is activated by FXa approximately 50100 times more effectively than thrombin [98]. This suggests that in the early  16  stages of blood coagulation, platelet FV may be more important in the prothrombinase complex assembly since it is abundant on the activated platelet surface, partially active and can be activated by preformed FXa [99]. Platelet FV is different from plasma FV in that the cleavage of B domain to release a light chain is at Tyr1543 instead of Arg1545, it is O-glycosylated at Thr402, it is not phosphorylated at Ser692 and it is 2-3-fold more resistant to the inactivation by APC [100]. The critical importance of platelet RI in hemostasis can be emphasized by some clinical evidence. Two patients with acquired inhibitors of FV: one with inhibitors against both plasma and platelet FV [101] and the other with inhibitors only against plasma FV [102]. Only the patient with the inhibitors for both RI suffered bleeding. The patient with the inhibitors only against plasma RI was shown to have completely neutralized plasma RI but normal platelet RI. This patient did not show any bleeding complications, which suggests that platelet RI is sufficient to maintain adequate hemostasis.  1.3.5 Cofactor function in prothrombinase Formation of the RIa:FXa:aPL:Ca 2 prothrombinase complex is essential in hemostasis. A fundamental contribution of FVa to prothrombinase function is its ability to retain FXa on the surface of a membrane [22]. Without FVa, the duration of retention of FXa on a membrane is short-lived with an affinity (Kd) for aPL vesicles of approximately 10.6107 M [24,103-105]. The Kd’s for the  17  binding of FVa to aPL average about i0 M [15], considerably lower than that of FXa. In the absence of phospholipid, the FVa and FXa interaction is also relatively weak (Kd, 0.8 !iM) [24,103,106]. When both FVa and FXa interact on the phospholipid membrane, the Kd for the binding of FVa and FXa is 10b0 M with a 1:1 stoichiometry [21]. FVa, therefore, increases the affinity for FXa:aPL interaction by about 100-fold; however, FXa has no effect on the FVa and aPL interaction  [24].  The  FXa:FVa:aPL  complex  cleaves  membrane-bound  prothrombin to produce thrombin localized principally on the activated platelet membrane surface. As a consequence, the Km for prothrombin activation is decreased by 100-fold. FVa also increases the enzymatic rate of the conversion of prothrombin to thrombin by 1000-fold [15]. A decrease in the Km for interaction with aPL and an increase in the Kcat for the enzymatic reaction result in the overall enhancement of the rate of FXa-catalyzed thrombin generation by 300000-fold [261. It is not clear whether the enhancement in Kcat is a result of an alteration of FXa active site, prothrombin as a substrate or alteration in both FXa and prothrombin when bound to FVa. Prothrombin is converted to thrombin by two proteolytic cleavages. Prothrombinase cleaves prothrombin yielding meizothrombin as an intermediate, with subsequent cleavage resulting in thrombin and fragment 1.2 [107]. Without FVa, prothrombin is cleaved to prethrombin-2 and fragment 1.2 before it is cleaved to form thrombin [44,108,109]. This suggests that FVa not only enhances the rate of prothrombin activation but also influences the pathway of  18  cleavage, directing it through the meizothrombin pathway. It has also been proposed that prothrombin might be converted directly to thrombin without proceeding through the free intermediates meizothrombin and prethrombin-2 [1101.  1.3.6 Mechanism of inhibition Blood coagulation proceeds through a series of reactions to generate thrombin, which is responsible for generating a fibrin clot. Thus, thrombin generation must be controlled to avoid unnecessary formation of thrombin to prevent thrombosis. The protein C anticoagulant pathway is a well-studied regulatory mechanism in controlling the production of thrombin. When enough thrombin has been generated, it binds to thrombomodulin on the surface of the vascular endothelium, together activating the vitamin K-dependent zymogen, PC, to activated protein C (APC) [111]. APC is a serine protease that catalyses the proteolytic inactivation of FVa and FVIIIa to limit further thrombin generation. APC inactivates membrane-bound human FVa heavy chain at Arg506, Arg306 and Arg679 producing an inactive FVa species [1121. Initial cleavage at Arg506 reduces FVa activity by decreasing the affinity for FXa and reducing  Kcat;  however, does not result in complete inactivation. The subsequent aPL dependent cleavage of Arg306, however, completely inactivates FVa resulting in the dissociation of the A2 domain [112-114]. Human FV is also inactivated by APC in the presence of aPL-containing membrane. The sequential cleavage sites  19  of FV are Arg306, Arg506, Arg679 and Lys994, different from FVa. The initial cleavage site at Arg306 is shown to inactivate FV [112]. Platelet FV is 2-3-fold more resistant to inactivation by APC compared to plasma FV [100]. Phosphorylation at Ser692 increases the rate of FV inactivation [64], and since platelet FV is resistant to phosphorylation at this site, it cannot be cleaved as efficiently by APC.  1.3.7 Tertiary structure  Structural studies of intact FV/FVa by X-ray crystallography have not been successful. Early gel filtration and sedimentation equilibrium studies suggest that FV is not globular but rather a rod-like molecule with an axial ratio of 25:1. The subunits of thrombin-activated FV are more globular, each subunit with an axial ratio of 5:1 [115-117]. The studies based on scanning electron microscopy proposed that FV has a multidomain structure with three globular domains linked via thin spacers to a somewhat larger central domain [118-120]. Three-dimensional structural data available for FVa can be summarized as follows: a theoretical model of the A domains based on the X-ray crystal structure of ceruloplasmin [121], a theoretical model of FVa based on the crystal structures of ceruloplasmin and the C2 domain of FV [122], a homology model of the C domains of FVa based on the galactose oxidase binding domain [123], a crystal structure of the C2 membrane binding domain [124], a crystal structure of activated protein C-inactivated bovine FVa (FVa) lacking the A2 domain [125]  20  and a recently proposed human FVa model derived from the bovine FVa 1 structure and the human ceruloplasmin structure [1261. The structure of human ceruloplasmin, a plasma copper-binding protein, has been solved [127]. The A-domains of FV and FVIII share important sequence identity with ceruloplasmin allowing a theoretical model (994 residues) for the A domains of FV/FVa (residues 1-656 and 1546-1883) to be built [121]. This model indicated that three A domains were arranged in a triangular fashion with each domain made up of two cupredoxin-type folds and the organization of these domains should remain the same before and after activation. In this model, a type II Cu 2 was hypothesized to bind at the Al and A3 interface and Ca 2 in the Al domain near the Al and A3 interface. The homology model of the A domains based on the crystal structure of ceruloplasmin and the C domains based on the crystal structure of the C2 domain of FV [122], was proposed that lacks 46 amino acid residues from the carboxyl terminus of the heavy chain (residues 664-709). This model proposed a complete structure of FVa bound to a putative phospholipid membrane. According to the model, FVa had a windmill shape with the six-bladed wheel formed by the A domain trimer and the cylindrical mill formed by the Cl domain stacked above the C2 domain. Each A domain is composed of two cupredoxin-like 13-barrels. According to this model the binding to phospholipid was predicted to be facilitated by the C2 domain. A type II Cu 2 was predicted to bind at the Al and A3 interface. Two putative binding locations  21  for Ca 2 were predicted, one at the Al and A3 interface and the other in the A3 domain. Although the X-ray crystal structure of intact FV/FVa is not available, the crystal structures of bovine FVa 1 [124] and the C2 domain of human RI are known [125]. The C2 crystal structures show that the C2 domain exhibits a 13barrel motif of eight antiparallel strands arranged in two f3-sheets of five and three strands packed together. This creates a scaffold of three protruding loops. The upper part of the barrel structure contains several salt bridges and the lower part contains basic residues. The bovine FVa crystal structure, which lacks the A2 domain, demonstrated that each A domain is comprised of two linked cupredoxin-like p-barrels and C domains have a distorted jelly-roll p-barrel with a great similarity between the Cl and C2. The structure of the C domains from the bovine FVa structure is similar to the crystal structure of recombinant human RI C2 domain [124,125]. As seen in the homology model [122], the Al and A3 domains from the bovine FVa 1 structure are arranged around a pseudo-threefold axis like windmill blades on the C domains. The bovine FVa 1 structure is different from the homology model however in that the Cl is not stacked above the C2 domain but rather the C domains are side by side, suggesting that both domains maybe involved in membrane binding.  22  L4 ROLE OF METAL IONS IN FACTOR V/VA 1.4.1 Essential metal ions  The distribution of salts in the blood and other body fluids reflects the oceanic nature as life on earth first evolved in the sea. The ionic environment both inside and outside of a cell is mainly controlled by Na, K, Ca, Mg, and Cl, which were abundant in the primeval seas. The trace elements required by all organisms, Fe, Cu, Zn, Mn, Mo and Co appear to have been limited by their availability in the early oceans, in concentrations 1O51O6 times less than the above abundant ions in sea water [1281. These trace metal ions have evolved to be essential components of a wide variety of metalloproteins with roles including enzymatic catalysis, electron transfer reactions, genetic and metabolic regulation, structural stabilization and transport process [129]. Since the beginning of life, metal ions have played important roles, therefore, it is not unexpected that many proteins have evolved to require metal ions for their structure and function.  1.4.2 Calcium  The concentration of blood calcium is about 2.3 mM. Of this, approximately 1.2 mM is free ionized Ca , about 0.8 mM is bound to albumin 2 and about 0.3 mM is complexed by low molecular weight compounds [130]. Although free ionized Ca -binding 2 2 is approximately 1.2 mM, extracellular Ca proteins have dissociation constants for Ca 2 2 ranging from i0 to i0 M. Ca  23  plays structural roles in many proteins and also is required for such purposes as enzyme activation or membrane binding of proteins. 2 favors oxygen ligands that are provided by carbonyl groups of the Ca peptide bond or by side chains in proteins. Ca 2 is often shown to be coordinated by 7 or 8 ligands in crystals with extremely irregular bond lengths and angles. The irregularity of the coordination of Ca 2 exists between proteins. Some proteins contain evolutionarily conserved Ca -binding motifs such as an EF 2 hand. The EF-hand conformation consists of two turns of x-helix, a twelveresidue loop containing six Ca -coordinating ligands and two more turns of x 2 helix. Thirty-two distinct subfamilies of proteins contain EF-hands and about twothirds of EF-hand domains bind Ca . Many other proteins that bind to Ca 2 2 are neither homologous to one another nor to the EF-hand protein Ca -binding 2 patterns but bind to Ca 2 with high affinity and selectivity [131]. According to known structural data, the Ca 2 binding of FV does not involve a common EF hand motif.  1.4.3 Copper  The concentration of copper in plasma is about 20 pM [1321. This concentration does not represent the physiologically available copper pool because about 10 pM copper is incorporated into ceruloplasmin in a nonexchangeable way and 2.8 pM is bound to albumin, 1.9 pM to transcuprein, and 3.6 pM is bound to low molecular weight components like amino acids in an  24  exchangeable way. Copper may be in the Cu , Cu 1 2 or Cu 3 state. The proteins that contain Cu 2 can be in three distinguishable states; type I resulting in a blue protein, type II having no color but is detectable by electron paramagnetic resonance and type III resulting in a non-blue and non-electron paramagnetic resonance-detectable protein [129]. The type I Cu 2 proteins have a fairly flat tetrahedral Cu 2 centre, with nitrogen and suihydryl ligands. Type II and III Cu 2 coordinate with oxygen or nitrogen ligands contained within a protein rather than sulhydryl ligands [133].  1.4.4 Calcium and copper in factor V/Va  It has been established that both bovine FV and FVa have a single high affinity site for Ca 2 [134]. Equilibrium binding studies have shown that bovine FV has a single extremely high affinity (Kd  <  10 nM) and two lower affinity (Kd  60 jiM) Ca -binding sites. In contrast, bovine FVa contains a single Ca 2 -binding 2 site (Kd  =  24 pM) and several lower affinity sites [135]. Atomic absorption and  emission spectroscopy studies have demonstrated that bovine and human FV/FVa also bind to one mole per mole of type II Cu 2 [136]. The Ca 2 may  stabilize the interaction between FVaH and FVaL [68,135]. When bovine FVa was treated with a chelating agent, EDTA, subunit dissociation was observed with the loss of cofactor activity. Upon adding divalent metal ions such as Ca , 2 , Mn 2 2 but not Mg 2 and Ba , subunits reassociated and the 2 , Co 2 Cd 2 and Sr  25  activity was restored [137]. Subunit reassociation has not been reported for human FVa, and may be preclusive.  1.4.5 Metal-dependent subunit association Although the metal-dependent FVaH and FVaL interaction is required for prothrombinase function [137] and thus critical for effective hemostasis, little is known about the precise biochemical role played by divalent cations. FVaH and FVaL are noncovalently associated in the presence of divalent cations with a -binding 2 dissociation constant of 6 nM [68]. This finding suggests that the Ca pocket is formed by the FVaH and FVaL association and Ca 2 does not interact with either subunit alone [135]. Occupancy of the Ca -binding site has been 2 correlated to FVa subunit association and consequent activity [68]. Terbium )-binding studies have shown that there is a single functionally important 3 (Tb -binding site on FVa located primarily on FVaH [138]. Tb 2 Ca 3 is widely used to probe Ca 2 binding because of its similar ionic radius and coordination properties 2 on to those of calcium and it is intrinsically fluorescent. The effect of type II Cu function has not yet been reported for FV. The amino acid residues involved in metal coordination according to various structural models are summarized in Table 1.1. A theoretical model for the human FV A-domains based on ceruloplasmin homology (Figure 1.5) indicated that Ca 2 and Cu 2 may be close to each other. The predicted Ca 2 binding site was surrounded by the acidic Al domain residues, G1u96, AsplO2,  26  Authors Villoutreix et al. (1998)  Pellequeret al. (2000)  Adams et al. (2004)  Orban et al. (2005)  Leeetal. (2007)  Model description  Ca 2 binding residues  Cu 2 binding residues  A homology model ofA domains based on the crystal structure of ce ru loplasm in  i. Glu96, G1u108, Aspi 02, Asplil, Aspll2 (Al)  1. His85 (Al), His1815 (A3)  I. Glu148, Asnl49(Al), Aspi 577 (A3)  1. His85 (Al), Hisl8l5, Hisl 817 (A3)  A theoretical model with A domains based or ceruloplasmin structure and C domains based or the crystal structure of human FV C2 domain A crystal structure of APC-cleaved bovine EVa (EVa)  2. Glul 572, GIul 576, Glul 583 (A3) I. Lys93, Glul 08, Aspi 11, Aspll2_(Al)  A model based on I. Lys93, the model by GlulO8, Pellequeretal. Asplli, with C domains AspI 12 (Al) remodeled by 2. AspI 579, replacing them Glu1572, with the C Glul 576, domains found in Glul 583 the bovine EVa (A3) structure  1. Hisl8l5, Hisi 817, Asn1857 (A3) I. HisI 815, Hisl8l7, Asnl857, Glul 859 (A3)  1. His147(Al), 1. Lys93, A theoretical AsnI 857, model based on G1u96, G1u108, Asplil, ThrI 858, the bovine EVa Aspli2 (Al) Glu1859, structure with the G1nl864 (A3) A2 domain modeled in based on ceruloplasmin  Table 1.1 Predicted Ca 2 and Cu 2 coordinating amino acid residues.  27  Figure 1.5 FV A domain homology model.  The Ca - and Cu 2 -binding sites are shown in the homology model of FV A 2 domains by Villoutreix et al. (FVaH, green; FVaL, red). The metal-coordinating amino acid residues are shown in white.  28  GlulO8, Asplll and Aspll2, which are near the interface with the A3 domain. On the other hand, modeling predicted that Cu 2 traverses the A1-A3domain interface by coordinating with His85, His1815 and Hisl8l7 [1211. The homology model by Pellequer et al. (Figure 1.6) predicted FVa binding to two 2 and one Cu Ca , with the binding sites of one Ca 2 2 and Cu 2 linking the Al and A3 domains [122]. The crystal structure of bovine FVa 1 (Figure 1.7) later confirmed the validity of the theoretical human models, although the amino acids implied to be in direct contact with metal ions were somewhat different; Ca 2 was coordinated by the Al domain residues, Lys93, GlulO8, Aspill and Aspll2, and Cu 2 only by the A3 domain residues His1802, His1804 (His1815, His1817 in human) and Asp1844 (Asn1857 in human). Both metal ions were located near the A1-A3 interface in the crystal [125]. Dissection of human FVa regions by plasmin-mediated fragmentation [139] and preliminary mutational [140] studies done in our laboratory provided functional evidence that highly conserved amino acids spanning 96-111 in the Al domain contribute to the chelator-sensitive heterodimeric subunit complex. In particular, mutation of Asplll to Ala appeared to destabilize the association of subunits after single chain FV was activated to two chain FVa when evaluated antigenically in culture media. A study by Sørensen et al. later demonstrated that double mutation in the hypothesized Ca 2 binding region of  RI,  AsplllAsn/Aspll2Asn, resulted in a  rapid activity loss upon treatment with thrombin due to subunit dissociation [141]. The direct involvement of this region on FV-bound metals has not been  29  Figure 1.6 FVa homology model. The Ca - and Cu 2 -binding sites are shown in the FVa homology model by 2 Pellequer et al. (FVaH, green; FVaL, red). The metal-coordinating amino acid residues are shown in white.  30  Figure 1.7 Bovine FVa structure.  The 2 Ca and Cu -binding sites are shown in the crystal structure of bovine 2 1 structure by Adams et al. (FVaH, green; FVaL, red). The metal-coordinating FVa amino acid residues shown in white are human numbers.  31  reported. Although FV and FVIII have the same homologous domain organization, the latter procofactor circulates in plasma as a non-covalent heterodimer [56]. Studies by Fay and colleagues demonstrated that unlike FVa, Ca 2 had only minor effects on the FVIH intersubunit association, but greatly increased the specific cofactor activity. In these studies Cu 2 was observed to predominantly facilitate the subunit-subunit affinity [142]. These investigators also showed that mutation at several acidic amino acid residues within 110-126, homologous to residues 94-110 of the Al domain of FV, cause a significant reduction (or complete loss) of Ca 2 binding, suggesting an involvement in Ca 2 coordination [143]. Two crystal structures of FVIII have been recently published [144,145]. Two bound Cu 2 were observed in both structures located near the Al and A3 interface. The Ca -binding pocket that is homologous to the Ca 2 -binding loop 2 in FV were identified in both structures. The coordinating amino acid residues are GlullO (homologous FV amino acid residues are shown in parentheses; Glu96), Aspll6 (AsplO2), Asp125 (Aspill) and Asp126 (Aspll2). The positions of these  metal ions support a vital role for maintaining the structural integrity of the interdomain interface.  32  1.5 FACTOR V PATHOLOGY 1.5.1 Factor V deficiency  The importance of FVa is highlighted by deficiencies in the precursor, FV, resulting in a bleeding disorder referred to as parahemophilia [146,147]. Although rare, X-linked FVIII deficiency, hemophilia A, is at least an order of magnitude more prevalent than the homozygous FV deficiency, which has an estimated frequency of less than one in one million [148]. Due to its low frequency in the population and the large size of the FV gene, little is known about the molecular basis underlying  RI  deficiency.  RI deficiency, first discovered by Owren in 1947, is an autosomal recessive bleeding disorder and is found in several ethnic backgrounds [149]. The phenotypic consequence of human parahemophilia varies. Heterozygotes are often asymptomatic while homozygotes show mild, moderate or severe bleeding symptoms [150]. Unlike humans, homozygous mice deficient in RI die before or soon after birth [151]. FV is essential for generating physiological levels of thrombin for survival [22]; therefore, complete R/ deficiency in humans is expected to be an early embryonic-lethal condition, as in mice. Conversely, individuals with no FV synthesis have either manageable or no pathology associated with bleeding. There seems to be poor correlation between RI levels and clinical phenotype [152-157]. These observations imply that there may be compensatory mechanisms in humans to mask the severity of the complete lack of RI.  33  Although described for the first time in 1947, the first molecular basis of a mutation causing the disease was only elucidated in 1998 [149]. Since then, more deficiency-causing mutations throughout the entire coding region of FV have been described. Currently, 65 mutations including point mutations, truncating (nonsense and frameshift) mutations and mutations of the splice sites for the introns and over 700 polymorphisms are documented in the database (http://www.lumc.nh/4010/research/factor V gene.html)  [158].  When  the  distribution of these mutations is investigated, nonsense and frameshift mutations are almost equally distributed throughout the protein. However, missense mutations are found throughout the protein except in the B domain suggesting the lower sequence constraints of the B domain for function. The majority of these mutations have been found in unique families. The mutation that substitutes Tyr1702 to Cys leading to Tyrl7O2Cys (Y1703C) is the most reported mutation occurring in diverse populations representing the only recurrent FV mutation at least in the Italian population [159].  1.5.2 Regulatory defects APC deficiency. Heterozygous protein C deficiency is relatively common  occurring in about 0.2 % of population [160]. Clinical studies demonstrate that individuals with heterozygous deficiency have increased risk of venous thrombosis [161], presumably due to the inability to proteolytically inactivate F\//Va. Homozygous deficiency, which results in no circulating protein C is  34  associated with life-threatening thrombotic complications that can be only treated by protein C supplementation [162-1641. APC resistance. APC resistance is characterized by a poor anticoagulant  response to APC in plasma associated with the increased risk for thrombosis. It was first reported in 1993 by Dahlbäch et al. to cause a reduced anticoagulant response to the addition of APC to the plasma of patients with a family history of venous thrombosis [165]. A year later, a single point mutation in the FV gene (G1691A), which results in the substitution of Arg506 by GIn was found to be responsible [166,167]. Since the discovery,  (Leiden  mutation)  p,Leiden  has been  shown to be the most frequent cause of APC resistance and occur at least ten times more common than other genetic defects that cause thrombosis [168]. Since FVa lacks the initial APC-cleavage site at Arg506, it is inactivated by APC at about one tenth the rate of normal FVa. However, FVaL is still inactivated through Arg306 cleavage [111]. The most common thrombotic manifestations of  p,jLeiden  are venous thrombosis and pulmonary embolism [169],  with a 5-fold increased risk of a thrombotic episode in heterozygous individuals and a 70-fold increased risk in homozygous individuals. The link between arterial thrombosis and the existence of studies have identified  pjLeiden  p,jLeiden  is yet unclear, however, some recent  as a risk factor for arterial thrombosis [170,171]  and atherosclerosis [172]. Since the discovery of have been identified.  LeIden, 1 p  pjCambndge  and  other genetic risk factors for thrombosis  pjHong Kong  also contain mutations that affect  35  APC cleavage, Arg306 to Thr in [174]. Both  jCambridge  and  cambridge  zjNong Kong 1  [173] and Arg306 to Gly in  are not as resistant to APC as  t’° Kong pjLeiden  although the cleavage at Arg306 is crucial in the FV inactivation by APC. Later it was shown that in the absence of the Arg306 cleavage site, APC still inactivated FV by cutting at an alternative site close to Arg306 [175] explaining the unexpectedly mild APC resistance associated with mutation at this site. Other factors apart from genetic defects that impair the response to APC have been reported. FV has five unpaired Cys and when the level of homocysteine  is elevated, these Cys can  incorporate homocysteine at  physiologically relevant concentrations associated with hyperhomocysteinemia, consequently resulting in a form of acquired APC resistance [176,177].  1.6 SCOPE OF THIS STUDY  The general scope of this study is to understand both the biochemistry and pathology regarding human FV/FVa. The main part of the study focuses on more detailed biochemistry on the role of metal ions in FV/FVa. The poorly understood, yet essential metal ion-dependent intersubunit interaction was studied by using a site-directed mutagenic approach. The effects of mutating several predicted Ca -binding residues of FV on subunit interaction, function 2 and metal binding were investigated. The other part of the project focuses on the clinical observation made in a patient who only suffers from a minor bleeding tendency, but has an insignificant amount of plasma FV. This study was  36  conducted cooperatively with other members at the UBC Center for Blood Research. The names of contributors will be mentioned in the Materials and Methods section.  37  2. MATERIALS AND METHODS  21. MATERIALS 4-(2-hydroxyethyl)- 1-piperazineethanesulfonic ethylenediami netetraacetic  acid  (EDTA),  acid  (Hepes),  phosphatidylserine  (PS),  phosphatidylcholine (PC), polyethylene glycol 8000 (PEG 8000), benzamidine and 4-amidinophenyl methane-sulfonyl fluoride hydrochloride (APMSF) and calcium ionophore A23187 were obtained from Sigma-Aldrich. Hirudin, aprotinin, and D phenylalanyl-L-prolyl-L-arginine  chloromethyl  ketone  CI) 2 (FPR-CH  were  purchased from Calbiochem. Optima grade nitric acid (F-1N0 ) was purchased 3 from Fisher Scientific. Opti-MEM, Dulbecco’s modified Eagle’s medium/F-12 (D MEM/F12), geneticin, trypsin, L-glutamine, ampicillin, Albumax I and penicillin streptomycin were purchased from Gibco and insulin-transferrin-sodium selenite supplement (ITS) was from Roche. QuikChange XL site-directed mutagenesis kit was from Strategene and Dh5ci-competent cells and LipofectAMlNE 2000 were from Invitrogen. Miniprep and synthesized oligonucleotides were from Qiagen. Factor V deficient plasma and hemostasis reference plasma were purchased from Biopool and Innovin was from Dade Behring. Automated APTT was from General diagnostics. SP Sepharose Fast Flow,  Q Sepharose Fast Flow and Superose 6  beads for purification were purchased from GE healthcare. Wild type human RI minus the residues 811-1491 was cloned into the pMT2 expression vector (pMT2-rFV) as described [72], and baby hamster kidney (BHK) cells were from Dr. MacGillivray (Centre for Blood Research, University of British Columbia,  38  Vancouver, BC, Canada). Small unilamellar vesicles (SUV) and large multilamellar vesicles (LMV) consisting of 75:25 PC:PS were prepared and quantified as described [1401.  2.2 PROTEINS  Human plasma-derived FVa, FXa, prothrombin and thrombin, and monoclonal antibody (mAb) specific for human FVaH (AHV-5146) were purchased from Haematologic Technologies. FVa A3 domain-specific mAb (14H12) was generated by Dr. R. Lemieux (Héma-Québec, Québec City, Québec, Canada), for which in-house recombinant FVaL produced using a baculovirus expression system and  purified by electroelution was the immunogen.  Peroxidase-conjugated goat anti-mouse IgG (Jackson ImmunoResearch) was used as a secondary detection antibody combined with the chemiluminescent ECL-Plus detection system (GE Healthcare) as a substrate.  2.3 MOLECULAR BIOLOGY OF FACTOR V 2.3.1 Protein expression  The human FV was expressed in BHK cells with the plasmid pMT2. The recombinant wild type RI (FV-810) expressed in the study had the partially truncated B domain (811-1491 deleted). This construct is fully cleavable by thrombin, although it is not a true procofactor in that it exhibits functional activity comparable to natural FVa [72]. It has been shown by several groups  39  that recombinant FV engineered to be deficient in the entire or partial B domain retains the cofactor activity [8,73,178,1791. Since the B domain is a large polypeptide (120 kDa) with many complicating post-translational modifications, many groups have eliminated or truncated the B domain for improved expression. The coding sequence of FV-810 was amplified from the plasmid pMT2-rFV-810 generously provided by Dr. Rodney M. Camire, University of Pennsylvania (Figure 2.1). The construct was transformed into DH5cL-competent cells and plated on Lauria-Bertani (LB) plates containing ampicillin (50 mg/L). A single colony was selected based on the resistance to ampicillin and grown in LB broth (5 ml) supplemented with ampicillin (50 mg/L) while shaking at 250 rpm at 37 °C overnight. Overnight-grown colony (200 j.il) was added to LB broth (100 ml) supplemented with ampicillin (50 mg/mI) and grown while shaking at 250 rpm at 37 °C overnight. DNA was then extracted using a maxiprep kit. The concentration of isolated DNA was determined by absorbance at 260 nm and the purity was checked by running on 1 % agarose gel. It was then fully sequenced to confirm the correctness of the entire FV gene in the pMT2 expression plasmid (Nucleic Acid Protein Service Unit, University of British Columbia, BC, Canada). The following primers, purchased from Qiagen, that span the entire RI gene were used for sequencing: 5’-GGTGGCCGCGTCCATCTGGT-3’ (FVSEQ1 F) 5’-CCCACCCTCAGTTAGGGTCC-3’ (FVSEQ2R) 5’-AGGACAGTGGACCCACCCAT-3’ (FVSEQ3F)  40  ‘‘fI1 (72) SV4O ori/enh Ampr  AdMLP TPL ‘Vs  pMT2-rFVdt (B-less)  Bsu361 (2537)  9935 bp  VAI  SV4O Poly A hFactor V DHFR  SnaBi (4226)  Figure 2.1 Functional map of plasmid pMT2-rFVdt. Expression construct (pMT2-rFVdt), which contains human FV minus the residues 811-1491 cloned into pMT2 vector, was used to create wild type FV, FV-810. This construct was used to create E96A, D1O2A and D111A using site-directed m utagenesis.  41  5’-TGAACACGA I i i I GAGTGTG-3’ (FVSEQ4R) 5’-ACATTATAAGAAAGTTATGT-3’ (FVSEQ5 F) 5’-CTAGMTTCATGGAAGTTAA-3’ (FVSEQ6 R) 5’-ACCATCCACTTCACTGGGCA-3’ (FVSEQ7F) 5’-GTAATAATTTCTTCTGTTTC-3’ (FVSEQ8R) 5’-ACTATGCTGAAATTGATTAT-3’ (FVSEQ9 F) 5’-CAGGCAAGCTGTAGATCATC-3’ (FVSEQ 1 OR) 5’-CTTACTAlTrATGACCI I IG-3’(FvSEQllF)  5’-TTATAGGCTCGAGTTGGAGA-3’ (FVSEQ1 2R) 5’-TGGCAGATCTrCAAAGGGAA-3’ (FVSEQ13F) 5’-ACGCTAGGATTGCCGTCAAG-3’ (FVSEQ14R)  2.3.2 Site-directed mutagenesis The single point mutants were generated using the QuikChange II XL sitedirected mutagenesis kit. The pMT2-rFV-810 expression construct was used as a template in the mutagenesis reaction. The polymerase chain reaction (PCR) was performed with the following program: 95°C/i mm; 18 cycles at 95°C/50 sec, 60°C/50 sec, 68°C/lO mm; 68°C/7 mm. The PCR conditions for a standard reaction are reported in Table 2.1. For each mutant, two complementary oligonucleotides with the desired mutation were amplified using the following primers purchased from Qiagen: 5’-GGTACAGTAAATTATCAGCCGGTGCTTCTTACCTTGACCAC-3’ (E96A),  42  Final Concentration  Reagent  Concentration of stock  Buffer  lOx  5  lx  Template  10 ng/pI  1  0.2 ng/pl  Forward primer  100 ng/pI  1.25  2.5 ng/pl  Reverse primer  100 ng/pl  1.25  2.5 nglpl  Volume  dNTPs  1  QuikSolution  3  ddH2O DNA polymerase  -  2.5 U/pl  -  36.5 1  0.05 U/pI  Table 2.1. Conditions used for PCR for mutagenesis.  43  5’-AGGAGCTTCTTACCTTGCCCACACATTCCCTGCG-3’ (Dl 02A), 5’-TCCCTGCGGAGAAGATGGCCGACGCTGTGGCTCCAG-3’ (Dli 1A). Each PCR product was digested Dpn I restriction enzyme (10 U/pl, 1 p1), transformed into DH5ct-competent cells and plated on LB plates containing ampicillin (50 mg/L). A single colony was selected from each mutant and grown in LB broth with ampicillin (50 mg/L). DNA was extracted using a maxiprep kit, quantified and fully sequenced to confirm the presence of the mutation and the fidelity of the rest of the clone by DNA sequencing (Nucleic Acid Protein Service Unit, University of British Columbia, BC, Canada).  2.3.3 Stable expression of FV-810 and mutants Transfection  of  FV  expression  plasmids  was  performed  using  LipofectAMiNE 2000 reagent according to the manufacturer’s instructions with pcDNA as the selectable marker plasmid. DNA (14 pg) and pcDNA (lpg) were suspended in 500 p1 of Opti-MEM. LipofectAMlNE 2000 (30 WI) was suspended in the same volume of Opti-MEM. DNA was mixed with LipofectAMiNE 2000 for 20 minutes at 22 °C and the mixture was used to transfect 80 % confluent BI-IK cells. After 5-6 hours, the media was changed to D-MEM/F12 supplemented with 5% fetal bovine serum and 1% L-glutamine. After 24 hours, cells were removed with trypsin/EDTA (0.25% trypsin and 1mM EDTA) and cultured in the above media with 1 mg/mI Geneticin. 10-14 days later, 24 Geneticin resistant colonies from each construct were selected for expansion in 24-well plates. Individual  44  wells were then tested for protein production by western blot analysis and clotting activity. Selected clones were expanded into triple flasks and cultured in D-MEM/F12 supplemented with ITS, 5 mM CaCI , and 1.0 mg/mI Albumax I. 2 Conditioned media was collected for 14-21 days, centrifuged, and stored at -80 °C in the presence of protease inhibitors (10 mM benzamidine, 10 pM APMSF).  2.3.4 Purification of FV-810 and mutants  Three litres of conditioned media was thawed at 37 °C followed by the addition of protease inhibitors (10 pM APMSF, 1 pM FPR-CH CL, 0.5 pg/mI 2 aprotinin). All subsequent steps were carried out at 4 °C. The medium was centrifuged at 15000 g for 30 mm  and concentrated 5-fold on a 100 kDa  molecular mass cut-off concentrator. The concentrate was then centrifuged at 15000 g for 30 mm  and loaded onto a SP Sepharose Fast Flow column  equilibrated with 20 mM MES, 0.1 M NaCI, and 5 mM CaCI2, pH 6.0, at a flow rate of 1 mI/mm. The column was then washed with 50 ml of equilibration buffer and eluted with a 30 ml linear NaCI gradient (from 0.1 to 0.6 M), collecting 1 ml fractions. Fractions containing FV activity were pooled, dialyzed against 20 mM Hepes, 0.15 M NaCI, and 5 mM CaCl2, pH 7.4 (HBS-i-Ca ), and centrifuged at 2 15000 g for 30 mm. The dialysate was then loaded onto a  Q Sepharose Fast Flow  column equilibrated with HBS-i-Ca , at a flow rate of 1 mI/mm. The column was 2 then washed with 50 ml of equilibration buffer and eluted with a 30 ml linear NaCI gradient (from 0.15-0.6 M), collecting 1 ml fractions. Fractions containing  45  FV activity were pooled, centrifuged at 15000 g for 30 mm, concentrated to 1 ml . 2 and loaded onto a Superose 6 gel filtration column equilibrated in HBS+Ca The column was eluted at 0.5 mI/mm  collecting 0.5 ml fractions. Fractions  containing FV activity were pooled, centrifuged at 15000 g for 30 mm  and the  supernatant concentrated to 0.5 ml. The concentration of purified protein was cm at 280 nm [72]. The M estimated using the extinction coefficient of 1.54 1 protein purity was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, 8% acrylamide) followed by staining with Coomassie Brilliant Blue R-250. The purified protein was then stored in small fractions at -80 °C and thawed once.  2.4 THROMBIN-MEDIATED CLEAVAGE FV-810 and mutants (200 nM) were incubated with thrombin (0.2 nM) in  2 at 22 °C. The digests were stopped at different times by heating at 95 HBS+Ca °C for 5 minutes in denaturing Laemmli sample buffer containing dithiothieitol (DTT, 12.5 mg/mI). The digests were then subjected to SDS-PAGE (8% acrylamide) and stained with Coomassie Brilliant Blue R-250.  2.5 BINDING TO ANIONIC PHOSPHOLIPID VESICLES  Binding curves using fixed concentrations of SUV (20 pM) and varying concentrations of FV-810 or mutants were obtained by light scattering analysis using a Varian Eclipse fluorescence spectrophotometer, similar to that described  46  previously [20,50]. Scattering intensities were measured at 320 nm (excitation and emission) with a band pass of 5 nm. These readings were performed using a 0.5 X 0.5 cm 2 quartz cuvette with the sample chamber maintained at 22 °C. Buffers were degassed and filtered using 0.22 pm filters to reduce solution scatter. Protein was centrifuged in an Eppendorf microfuge for 5 mm at 4 °C to remove possible scattering material. Light scattering intensities of 20 pM SUV (140 p1) alone or with additions of various concentrations of FV-810 or mutants were measured. The increase in scattering intensity due to addition of FV-810 or each mutant was followed for 5 mm as quickly as possible after manual addition (“10 sec). The intensities after binding equilibrated (less than 5 mm) were then graphed against the concentration of FV-810 or mutants. The data were fitted by a single binding site function to derive the apparent dissociation constant (app Kd). No detectable changes in light scattering intensity were observed when the addition of SUV in the absence of FV or the addition of FV-810 or mutants in the absence of SUV was made. The changes in light scattering intensity due to dilution by continuous ligand titration were corrected. The correlation of light scattering to FV:aPL binding for FV-810 and D111A was confirmed initially by detecting FV-810 or mutants remaining bound to LMV after centrifugation using Western blot analysis as described [140].  47  2.6 ACTIVITY ASSAY 2.6.1 Prothrombinase assay FV-810 or mutants (pre-cleaved by thrombin or non-cleaved) were tested  for their ability to enhance thrombin generation of FXa by prothrominase. Saturating FXa (0.5 nM), SUV (50 pM), prothrombin (1.4 pM) and various concentrations of either FV-810 or mutants in HBS+Ca 2 with 0.1% PEG 8000 were mixed at 22 °C. Following a 5 mm  incubation, the reaction was stopped  with EDTA (15 mM) and thrombin generation was followed by S2238 chromogenic substrate (0.2 mM) cleavage using a Spectramax kinetic microplate reader (Molecular Devices). For some experiments, FV-810 or mutants (100 nM) were pre-treated with thrombin (5 nM) for 5 mm, which was sufficient to cleave all of the FV to FVa. Hirudin (0.7 U/mI) was added at an amount to stop only the thrombin included in the pre-treatment step and dilutions were made appropriately to be assayed for prothrombinase activity. All points were done in duplicate and the entire experiment repeated at least once for reproducibility.  2.6.2 Clotting assay  FV-810 or mutants (pre-cleaved by thrombin or non-cleaved) were tested for their ability to generate a fibrin clot. Various concentrations of FV-810 or mutants in HBS+Ca 2 were pre-incubated with 5Oul of FV-deficient plasma in 37 °C for 1 mm. At 1 mm, Innovin (100 p1) was added to initiate coagulation and time to clot formation was monitored using an ST4 analyzer (Diagnostica Stago).  48  For experiments with pre-cleavage, FV-810 or mutants (100 nM) were cleaved with thrombin (5 nM) for 5 mm. Proteolysis was then stopped with hirudin (0.7 U/mi) and appropriate dilutions were made to be assayed for clotting activity. The clotting times observed were converted to equivalent FV clotting activities assuming that undiluted plasma F\/, 20 nM [37], has a clotting activity of 1 U/mI. All assays were conducted in duplicate and the entire experiment was repeated for reproducibility.  2.7 EFFECT OF CHELATION 2.7.1 Chelator-induced subunit dissociation FVaH dissociation from FVaL bound to SUV was monitored by following a  decrease in light scattering as previously reported [114] using a Varian Eclipse fluorometer. To instantly convert all of the FV-810 or mutants bound to SUV (20 pM) to FVa, a high concentration of thrombin was used (100 nM). This avoided complicating the observed FVa subunit dissociation by thrombin cleavage kinetics. The changes in light scattering intensities upon addition of thrombin (100 nM) to FV-810 or mutants (100 nM) incubated with fresh SUV (20 pM) were continuously monitored until complete, then the changes upon addition of EDTA (5 mM) were similarly monitored. The decrease in light scattering due to FVaH dissociation from the FVaL-SUV complex was fitted by a single-phase exponential decay function. In control experiments no detectable changes in light scattering intensity were seen upon the addition of 1) thrombin and EDTA in the absence of  49  FV-810 or mutants, 2) buffer instead of thrombin to FV-810 or mutants, or 3) EDTA to non-cleaved FV-810 or mutants bound to SUV. The correlation of light scattering to FVa subunit dissociation for FV-810 was confirmed initially by detecting FVaH remaining bound to LMV after centrifugation using western blot analysis as described [140].  2.7.2 Chelator-induced cofactor activity loss  To correlate subunit dissociation with loss of activity, FV-810 or mutants (100 nM) were treated as above with thrombin (100 nM). After the addition of EDTA (5 mM), samples were diluted to 0.5 nM and assayed for clotting activity at various times. The clotting times observed were converted to equivalent FV clotting activities assuming that undiluted plasma FV, 20 nM [37], has a clotting activity of 1 U/mI. All points were done in duplicate and the entire experiment was repeated for reproducibility. The activity reduction observed was fitted by a single-phase exponential decay function.  2.8 METAL ANALYSIS 2.8.1. Inductively coupled plasma-mass spectrometry  In an attempt to optimize the conditions for measuring Ca 2 and Cu 2 bound to FV-810 and mutants, inductively coupled-plasma mass spectrometry (ICP-MS) was first utilized to measure the amount of divalent metal ions bound to commercially available natural human FVa (Haematologic Technologies).  50  Experiments were conducted with a Perkin Elmer-SCIEX ELAN inductively coupled plasma-mass spectrometer. All labware was treated with 3% HCI/3% 3 and I-INC  sample manipulation was conducted in a metal-free plastic  containment hood. Two different buffers were prepared for the experiments; 20 mM Hepes and 150 mM NaCI, pH 7.4 (buffer 1), 20 mM Hepes, 150 mM NaCI, 2 mM CaCI 2 and 20 pM CuCl , pH 7.4 (buffer 2). Commercially available natural 2 human FVa was diluted to 100 pg/mI in buffer 1 or buffer 2 and dialyzed overnight at 4 °C with at least two buffer changes. The concentration of the dialyzed natural FVa against each buffer was checked by absorbance at 280 nm. All samples were prepared by further diluting natural FVa to 1 ng/ml, 10 ng/ml and 100 ng/ml in 1 % HN0 . Both buffers were also similarly diluted to be used 3 as negative controls. The calibration curves were each prepared for 44 2 and Ca 2 with concentrations, 0 ppb, 1 ppb, 10 ppb and 50 ppb for 44 Cu 2 and 0 ppb, Ca 1 ppb and 10 ppb for Cu . The amount of 44 2 Ca, Mg, Al, Ti, Cr, Mn, Co, Cu, Zn, Se, P, S and Ni bound to natural FVa was simultaneously determined. Both standards and samples were measured in quintuplicate.  2.8.2 Graphite furnace atomic absorption spectrometry  Graphite furnace atomic absorption spectrometry (GFAAS) was next employed to measure Ca 3 at 2 and Cu . All labware was treated with 1 °k HNO 21 50 °C for 5 hours and sample manipulation was conducted in a metal-free plastic containment hood. Buffer background was sufficiently reduced of divalent cations  51  for differential measurement, by extensive dialysis of FV-810 and mutants against divalent cation-free FIBS (HBS-Me) to remove all excess divalent metal ions at 4 °C. The divalent cation-free buffer (500 ml) was prepared by dialyzing HBS (no added Ca j against Chelex 100 (20 g). Samples of FV were then 2 further dialyzed against HBS-Me made with 100 nM Ca 2 and 100 fM Cu . Prior 2 to analysis, samples were further diluted 10-fold in 0.1% metal-free HNO . FV 3 810 and mutants were compared to the respective FVa by pre-treatment with 2 before dialysis as above. All samples assayed for Ca thrombin in HBS+Ca 2 and 2 were within the linear range of detection by a Varian SpectraAA 300 Cu Zeeman graphite furnace atomic absorption spectrometer. Metal ion standard curves were obtained with 0 ppb (0 pM), 2.5 ppb (0.0625 pM), 5 ppb (0.125 pM) and 10 ppb (0.25 pM) of Ca 2 and Cu t Both Ca 2 2 and Cu 2 were measured with a slit width of 0.5 nm and a lamp current of 4 mA with sampler automixing and background correction on. The Ca 2 measurement was carried out with a wavelength of 422.7 nm and Cu 2 with 327.4 nm. All readings were measured in duplicate and the concentrations of purified proteins were measured by absorbance at 280 nm to be exact.  2.9 FACTOR V-DEFICIENT PATIENT STUDY 2.9.1 Blood sample preparation  Before the collection of blood, a donor consent document was signed by all participants. Platelets were isolated via a washing method [1801. To minimize  52  handling—induced activation of the platelets, all buffers to contact the platelets were warmed to 22 °C before use and all washing steps proceeded in polypropylene tubes. Whole blood was collected from the patient, his wife and son in 2.7 ml sodium citrate tubes. For the washing procedure, the blood was subjected to centrifugation at 140 g for 15 mm  at 22 °C to sediment the red  blood cells and the resulting platelet-rich plasma (PRP) and platelet-poor plasma (PPP) were removed. The PPP was centrifuged again to further isolate plasma. Once plasma was separated from other blood components, it was kept at 4°C for immediate use or frozen at —80°C for storage to avoid protein degradation. The PRP was centrifuged to sediment platelets at 500 g for 15 mm 22 °C and washed twice in 10 mM trisodium citrate, 30 mM dextrose and 120 mM NaCI, pH 6.5 (CGS) with the addition of 11.5 mM citrate acid monohydrate, 88.5 mM trisodium citrate dihydrate and 111 mM dextrose, pH 6.0 (ACD) before centrifuging to ensure the recovery of platelets from the centrifugation. The CGS was supplemented with 1 U/mI apyrase immediately before use in the washing steps. The platelets were finally resuspended in HBS+Ca . The platelets were then 2 allowed to rest for 30 mm  at 37 °C to recover from handling. Freshly isolated  platelets were kept at 22 °C and used within eight hours of isolation. (By Dr. Katherine Serrano, Centre for Blood Research, UBC)  53  2.9.2 Isolation of DNA from leukocytes The FV gene is located on chromosome lq, spanning about 80 kb of DNA and containing 25 exons and 24 introns. DNA was isolated from peripheral blood leucocytes using the protocol from Charge Switch® gDNA imI blood kit (Invitrogen) according to the manufacturer’s instructions. (By Michael Ho and Jeffery Hewitt, Centre for Blood Research, UBC)  2.9.3 DNA sequence Analysis Primers were designed for the 25 FV exons (exons 1-12 exon 13  -  B-domain and exons 14-25  -  -  FVaH, A1-A2;  EVaL, A3-C1-C2). The FV exons and  flanking intron sequences were amplified by PCR. Purified PCR product was subjected  to  automatic  DNA  sequence  analysis.  Blast  searches  (http://www.ncbi.nlm.nih.gov/BLAST/) were performed to identify the mutations in the FV gene. (By Michael Ho and Jeffery Hewitt, Centre for Blood Research, U BC)  2.9.4 Functional factor V level in plasma and platelets To confirm the activity level of plasma FV in the patient, his son and his wife, their plasma was tested for its ability to generate a fibrin clot. To check the FV activity level in platelets, platelets were treated with calcium ionophore A23187 (1 lJM) at 22 °C for 10 mm  while gently shaking to release platelet FV.  Both plasma- and platelet-derived FV were diluted in HBS+Ca 2 and pre  54  incubated with FV-deficient plasma (50 WI) in 37 °C for 1 mm. At 1 mm, Innovin (100 WI) was added to initiate coagulation and time to clot formation was monitored using an ST4 analyzer. The observed clotting times were converted to clotting activity using a standard curve constructed with assayed reference human plasma under identical conditions as the samples and converted to approximate FV concentration assuming the normal plasma concentration of 20 nM [37]. All assays were completed in duplicate and reproduced on a separate day.  2.9.5 Immunological factor V level in plasma and platelets To check the antigenic levels of FV, plasma and platelets from the patient, his son and wife were subjected to Western blot analysis. Prior to activation, platelets were treated with ionophore A23187 (1 pM) for 10 mm at 22 °C while gently shaking. Both plasma and platelet RI were then activated with thrombin (10 nM) in the presence of Ca 2 (5 mM) for 5 mm  at 37 °C. Both resting and  activated plasma and platelet RI were reduced by the addition of denaturing Laemmli sample buffer containing dithiothreitol (DTT, 12.5 mg/mI) and heated at 95 °C for 5 mm. They were then subjected to SDS-PAGE (4-15% acylamide) followed by transfer to polyvinylidine difluoride (PVDF) membrane. PVDF membranes were blocked with 5% skim milk in 0.05 M Tris-HCI, 150 mM NaCI and 0.05% v/v Tween-20, pH 8.0 (TBST) for 60 mm, then incubated with either 2 pg/mI anti-RI light chain (cL-FVaL) monoclonal antibody or 0.5 pg/mI anti-RI  55  heavy chain (cL-FVaH) monoclonal antibody for 90 mm  in TBST/0.1% skim milk  at 22 °C. Membranes were washed in TBST, incubated with horseradish peroxidase-conjugated goat anti-mouse antibody for 60 mm, washed again in TBST  with  subsequent  antigen  FV/FVa  detection  using  ECL-Plus  chemiluminescent substrate and documentation using a ChemiGenius instrument (PerkinElmer). The positions of FV and FVa were determined by comparison to commercially available purified FV/FVa (not shown). (By Kimberley Talbot, Centre for Blood Research, UBC)  2.9.6 Investigation of blood coagulation pathway To investigate if the patient’s lack of FV is compensated by other factors, the intrinsic and extrinsic pathways of the clotting cascade of the patient were separately  examined  using  TF-  or  contact  phase-dependent  initiation,  respectively. Effects on the intrinsic pathway were evaluated using an activated partial thromboplastin time (APTT)-clotting assay. Commercially available APTT reagent contains rabbit brain phospholipids and micronized silica to activate blood coagulation. A prothrombin time (PT)-clotting assay that uses recombinant TF reconstituted into SUV (Innovin) to initiate clotting was used. Each pathway was investigated by three separate experiments to see if any components in the patient plasma accelerated clot formation as compared to the commercially available RI-depleted plasma. These experiments were designed to identify the  56  possible compensating components in the patient plasma that may alleviate the patient’s bleeding tendency. Titration of purified FV against FV-depleted plasma. To investigate the  extrinsic pathway, purified FV-810 (50 WI), diluted to various concentrations (040 nM), was mixed with either patient or normal plasma immunodepleted of FV (50 p1) for 1 mm  at 37 °C. At 1 mm, Innovin (100 WI) was added to initiate  coagulation and time to clot formation was monitored using an ST4 analyzer. For the intrinsic pathway, patient or normal plasma immunodepleted of  RI  (50 WI)  was mixed with commercial APTT reagent (50 WI) at 37 °C for 5 mm (300 Sec). At 270 sec, 10 p1 of various concentrations of purified recombinant RI (0  -  40  nM) was added. At 300 sec, CaCI 2 (50 pl, 25 mM) was added to initiate coagulation and time to clot formation was monitored using an ST4 analyzer. All assays were completed in duplicate. Titration of normal plasma against FV-depleted plasma. For the extrinsic  pathway, various amount of either patient or RI-depleted plasma (0 was incubated with normal plasma (0  -  100 pL) for 1 mm  -  100 pL)  at 37 °C. At 1 mm,  Innovin (100 WI) was added to initiate coagulation and time to clot formation was monitored using an ST4 analyzer. For the intrinsic pathway, patient or RI depleted plasma (0  -  50 pL) was pre-mixed with normal plasma (0  -  50 pL) and  50 pL of the mixture was incubated with APTT reagent (50 p1) for 5 mm at 37 °C. At 5 mm, CaCI 2 (50 p1, 25 mM) was added to initiate coagulation and time to  57  clot formation was monitored using an ST4 analyzer. All assays were completed in duplicate. Titration of FL/-depleted plasma against 10 % normal plasma. In this experiment, different concentrations of patient plasma or RI-depleted plasma were added to normal plasma (10 %) to observe if the addition of more concentrated patient plasma would accelerate clotting. Normal plasma at 10 °k was chosen because of the previous results that showed the maximum difference in clotting times between the patient and RI-depleted plasma. To investigate the extrinsic pathway, 90 pL of variously diluted patient plasma or RI-depleted plasma (0  -  100 %) was incubated with 10 pL of 10 % normal plasma for 1 mm  at 37 °C. At 1 mi  Innovin (100 p1) was added to initiate coagulation and time  to clot formation was monitored using an ST4 analyzer. For the intrinsic pathway, 45 pL of patient or RI-depleted plasma (0  -  100 %) was pre-mixed  with 10 °h normal plasma (5 pL) and 50 pL of the mixture was incubated with APTT reagent (50 WI) for 5 mm at 37 °C. At 5 mi  2 (50 WI, 25 mM) was CaCI  added to initiate coagulation and time to clot formation was monitored using an ST4 analyzer. All assays were completed in duplicate.  2.9.7 Investigation of intrinsic pathway  The results from the previous experiment led me to further investigate the patient’s clotting factors focusing mainly on the intrinsic pathway in a clinical  58  laboratory. The examined proteins were FXII, Fy11, FVIII, FIX, FX, FXI, HK, PKK and antithrombin. (By Dr. Cedric Carter)  59  3. ROLE OF GLU96, ASP1O2 AND ASP111 IN FACTOR V 3.1 HYPOTHESIS AND SPECIFIC GOALS  In order to understand the importance of the predicted Ca -binding 2 region in RI function, subunit interaction and metal binding, the roles of three amino acids, G1u96, AsplO2 and Asplil, that are predicted to be involved in direct Ca 2 binding according to the homology model [1211 were investigated. The hypothesis was that the three amino acid residues are involved in divalent cation-dependent function. To address this hypothesis, the specific goals were: 1. Construct and purify recombinant RI mutated at G1u96, AsplO2, and Aspill to Ala giving Glu96Ala (E96A), AsplO2Ala (D1O2A) and AspillAla (D111A).  2. Determine the effect of mutation on thrombin-mediated proteolysis. 3. Determine the effect of mutation on binding to anionic phospholipid. 4. Determine the effect of mutation on cofactor activily. 5. Determine the effect of mutation on subunit interaction. 6. Determine the effect of mutation on metal coordination  3.2 RESULTS 3.2.1 Mutagenesis and production of FV  According to ceruloplasmin homology modeling, G1u96, AsplO2 and Asplil potentially interact with Ca 2 along with a few more nearby acidic amino residues [121] (Figure 3.1). A previous study from the Pryzdial laboratory had  60  96  102  4.  +  112 108 111  +  +  Human FV Mouse FV Bovine FV Pig FV Human FVII Mouse FVI II Dog FVIII LKNMASPVGVS KSQTYV Pig FVI II RPYTNSHGTYYKEF3AIYTDFQRPKVTYMLLATE Human Ceruloplasmin LNLASArIGVTYT KE’xAVrI\TTTDFQRKVLPGQQYVYV Mouse Ceruloplasmin Rat Ceruloplasmin FKNKADKPLSIHPKL1ASYIfl-ITFPAEKI\flDAVAPGREYTYE FRNKADKPLS IHPQFEPASY11Dê-ITFPAERE4DIDAVAPGEEYTYE FKNKAHKPLSIHAKF1EPASYDI-ITLPMEKIYIDIDAVAPGQEYTYE FRNKADKPLSIHPKF1EASYDTFLVEKLvIDIDAVAPGQEYTYE LKNMASHPGVSYWKAAEYE1DfTSQREKE4DIDKVF?GGSHTYV LKNMASHVGVSYWKAPDEYEE1TSQMEKEIDIDKVFPGESHTYV  Figure 3.1 Homologous region spanning 96-112 in the Al domain in P1111 and ceruloplasmin from various species.  RI,  The amino acid sequence of human FV implicated in the Ca -dependent FVaH 2 FVaL association by previous studies from our laboratory and homology modeling are shown along with the sequences of the homologous regions from other species and homologous proteins. The five acidic residues implicated by homology modeling to constitute a potential Ca -binding site are indicated by 2 the arrows. Conserved sequences are indicated in red. The amino acid residues that were mutated in the study are indicated in blue boxes.  61  screened all the possible conserved acidic residues within the divalent cationsensitive region for their ability to coordinate the FVaH-FVaL interaction in FVa (143). According to this study E96A, D1O2A and D1I1A mutants had the greatest effect on FV function [140]. To study the roles of these three amino acid residues in FV, I generated recombinant wild type (FV-810) and the E96A, D1O2A and D111A mutants by site-directed mutagenesis. The correctness of the mutations was confirmed by DNA sequence analysis (Figure 3.2). The vectors were stably expressed in BHK cells, and the recombinant FV mutants were secreted into serum-free culture medium. Typical expression levels were ri5OO pg/I indicating that the mutations had no significant effect on expression levels. The recombinant FV molecules were subsequently purified. SDS-PAGE and Coomassie Blue staining showed that FV-810 and mutants had an estimated molecular mass of r’200 kDa. The antigenicity and clotting functions were confirmed by Western blots and clotting assays, respectively. The Western blots confirmed that the Mr was “.‘200 kDa using either commercial anti-FV antibody (AHV-5146) or a monoclonal antibody produced by our lab against baculovirus expressed human FVaL (141-112). The concentration of purified protein was measured by absorbance at 280 nm and semi-quantitative Western blots.  3.2.2 Thrombin-mediated cleavage To ascertain that mutation of FV-810 did not affect the thrombin-mediated cleavage pattern and to establish the conditions needed for complete proteolysis  62  Asplil  AsplO2  G1u96  CAG  CAG  AAG  —  5’  A  —  160  170  180  —  200  190  210  ‘ J  3’  A1a96  3’ —  ACAGCGrCorCcA ?C T?C?CCGCGGGAATGTGTGGTcAGoTAAoAAQcAccoQcTo 160 170 180 190 200 210  B  3’  Alal 02  5’  3’ 5’ CACAGCGTc0TCC?CTTC ?CCGCOGGAA70?0TGGGCAGGThGAp.GcAcc?TCTo  C 160  170  180  190  200  210  3’  Alal 11 5’ 3’  3’ 1?C tCC0CkGQ9AA?Q?O7QQTcAAGG?AkQaAGcAcCc.oJ  5’  Figure 3.2 DNA sequencing of pMT2 containing FV-810 and its variants.  DNA sequence chromatograms from amino acid residue 95 to 114 are shown. (A) FV-810 sequence is shown with the 3-nucleotide codons for G1u96, AsplO2 and Aspill underlined (B) GM was replaced by GCC at amino acid residue 96 creating GIu96Ala (E96A). (C) GAC was replaced by GCC at amino acid residue 102 creating AsplO2Ala (D1O2A). (C) GAC was replaced by GCC at amino acid residue 111 creating AspillAla (D111A).  63  to FVa, a thrombin-treatment time course was performed in the presence of t SDS-PAGE (Figure 3.3) showed that FV-8l0 and mutants were comparably 2 Ca converted within 15 mm into FVaH (“105 kDa) and FVaL Q’74 kDa) through the anticipated activation intermediate (150 kDa), demonstrating that the targeted substitutions do not significantly affect proteolysis by thrombin.  3.2.3 Binding to anionic phospholipid vesicle Binding of FV and FVa to an anionic phospholipid membrane (aPL) has been shown in the past to involve the C2 domain of FVaL [80,81]. Therefore, mutation in the Al domain would not be anticipated to affect the interaction with aPL. This was confirmed, experimentally by measuring the binding of FV-810 and mutants to aPL-containing SUV by light scattering. This technique allowed to obtain dissociation constants by equilibrium binding measurements without having to label FV with probes that maybe alter its lipid-binding properties. Figure 3.4 A shows the binding isotherms of the mutants were indistinguishable from FV-810, with derived apparent dissociation constants of 1.64 E96A, 1.58  x  x  108 M;  108 M; D1O2A, 1.39 x 10-8 M; D111A and 1.37 x 10-8 M; FV-810.  These data confirm that there was no evident alteration in the capacity to bind aPL due to mutation. Western blot analysis confirmed that the observed light scattering correlated with the binding of FV-810 and D111A to aPL (Figure 3.4 B).  64  FV 150 kDa FVaH FVaL FV  A  B  C  D  ..-.  —  —  150 kDa FVaH FVaL 0  2  4 15 40  I  0  2  4  1540(min)  Figure 3.3 Thrombin-mediated conversion of FV-810 and mutants.  FV-810 (A), E96A (B), D1O2A (C) and D111A (D) (200nM) were treated with (5 mM) at 22 °C. At the indicated time (mm), 2 thrombin (0.2 nM) in HBS+Ca aliquots were taken from the mixture and the reaction was stopped by adding SDS-containing Laemmli sample buffer. Samples were separated by SDS-PAGE (8°h) and stained with Coomasie blue.  65  F  25  -J  A  01’ 0  I  25 [FV] (nM) 1  2  3  50  4  FV 150 kDa  120 kDa B  Figure 3.4 Binding of FV-810 and mutants to anionic phospholipid vesicles. (A) Various concentrations of FV-810 (Is), E96A C.), D1O2A () or D111A (.) were added to PCPS (20 jiM) and binding was monitored by light scattering analysis at 22 °C. Light scattering was measured approximately 10 sec after each addition and binding kinetics was measured for 5 mm. The intensities after binding equilibrated were graphed against the concentrations of FV-810 or mutants. (B) The correlation of light scattering to FV-810:aPL and D111A:aPL binding was confirmed initially by detecting FV-810 or D111A remaining bound to LMV after centrifugation using Western blot analysis probed with x-FVaL mAb (14H12). Lanes 1 and 3 are LMV only. Lanes 2 (FV-810) and 4 (D111A) show pelleted proteins after incubation with LMV for 10 mm.  66  3.2.4 Cofactor function The cofactor function of FV-810 and mutants was investigated by prothrombinase (Figure 3.5) and clotting assays (Figure 3.6). Both assays were performed with non-cleaved (1-stage assay) or thrombin pre-cleaved (2-stage assay) FV-810 and mutants. The residual thrombin was inactivated with the minimum amount of hirudin, which had no effect as determined by combining thrombin and hirudin prior to addition in control experiments (not shown). Both prothrombinase and clotting activity of the mutants were reduced in both 1-stage and 2-stage assays. The magnitude of reduced cofactor function of E96A and D1O2A is shown to be similar in 1-stage and 2-stage assays. The reduction in cofactor activity of D111A, however, was much greater in 2-stage assays according to both prothrombinase and clotting assays.  3.2.5 Subunit dissociation To determine the effect of mutation on FVaH-FVaL association, FV-810 or mutants were equilibrated with aPL-containing SUV and subunit interaction was monitored by light scattering (Figure 3.7). When FV-810, E96A and D1O2A were cleaved with thrombin in the presence of metal ions (Figure 3.7 A,C,E), no subunit dissociation was detected as expected from the lack of a 1-stage:2-stage assay discrepancy. Upon EDTA chelation, subunits of FV-810, E96A and D1O2A dissociated with dissociation rates of 0.026 ± 0.0002 min , 0.051 ± 0.0007 min 1 1 and 0.041 ± 0.0003 min , respectively (Figure 3.7 B,D,F). Thus, upon chelation, 1  67  30  15  0  0’ L..  a,  a, 0  0  I  2  0  1  2  30  C .0  E 0  I  I-  15  [FV](nM) Figure 3.5 Prothrombinase activity of FV-810 and mutants.  (A) Thrombin generation catalyzed by prothrombinase was measured in a reaction mixture containing FXa (0.5 nM), PCPS (50 pM) prothrombin (1.4 pM) and various concentrations of FV-810 (tX), E96A (.), D1O2A (0) or D111A (.) in (5 mM)+PEG 8000 at 22 °C. Following a 5 mm incubation, the reaction 2 HBS+Ca was stopped with EDTA (15 mM) and thrombin generation was measured chromogenically using S2238 (0.2 mM) as a substrate. All assays were done in duplicate. (B) FV-810 (Lx), E96A (.), D1O2A (0) or D111A (.) (100 nM) was pre cleaved with 5 nM thrombin for 5 mm. Activation was stopped after 5 mm by adding hirudin (0.7 U/mI). The activated samples were then diluted to various concentrations and assayed as in (A).  68  0.3  0.2  0.1 D > LI  0 ‘I  0  2  4  0  2  4  I [FV] (nM)  Figure 3.6 Clotting activity of FV-81O and mutants. (A) Various concentrations of FV-810 (Li), E96A (.), D1O2A () or D111A (.) were pre-incubated with RI-deficient plasma (50 pi) for 1 mm in 37 °C. At 1 mm, Innovin (100 pi) was added to initiate coagulation and time to clot formation was monitored. All assays were done in duplicate. (B) FV-810 (1k), E96A (.), D1O2A () or D111A (.) (100 nM) was pre-cleaved with thrombin (5 nM) for 5 mm. Activation was stopped after 5 mm by adding hirudin (0.7 U/mI). The activated samples were then diluted to various concentrations and assayed as in (A).  69  5  A  B  4 3 2  1 0  0  20 40 60 800 C  50 100 150 2( D  20 40  50 100 150 2( F  ...  U, 4  ,  z I  .0 I  00  60  .  a)  800 E  •‘“,4  Cu C.) Cl) 0) -J  n  •*•.  20 40  5  60 8 00 G  50 1001502C H  ..z u0  I  2  3 40 20 40 60 80 Time (mm)  Figure 3.7 Subunit interaction of FV-810 and mutants.  Subunit interaction of FV-810 or mutants (100 nM) bound to PCPS (20 pM) was monitored by light scattering analysis at 22 °C. Light scattering was measured upon addition of thrombin (100 nM) to FV-810 (A), E96A (C), D1O2A (E) or D111A (G). After measurement, EDTA (5 mM) was added to cleaved FV-810 (B), E96A (D), D1O2A (F) or D111A (H) and light scattering was measured upon addition. The scattering decay observed was represented by solid lines fitted to a single-phase exponential decay function to the data where appropriate.  70  E96A and D1O2A had significantly faster FVaH-FVaL dissociation rates compared to activated FV-810 (p  <  0.0001; three experiments), suggesting that G1u96 and  AsplO2 are involved in the metal-dependent subunit interaction of FVa. This faster FVaH-FVaL dissociation observed for E96A and D1O2A compared to FVa 810 indicated that altered subunit interactions contributed to the reduced cofactor activity. We cannot, however, exclude the possibility that other FXa cofactor properties of FV are affected. Contrary to FV-810, E96A and D1O2A, the rapid conversion of D111A to FVa by thrombin resulted in subunit dissociation with a rate of dissociation, 2.14 ±  0.049 min 1 even in the presence of metal ions (Figure 3.7 G). EDTA chelation  did not result in further changes (Figure 3.7 H). These data clearly demonstrate that the reduction of cofactor activity of D111A is due to spontaneous subunit dissociation and explains our observation that D111A activity appears far more inhibited 2-stage than 1-stage assays. Western blot analysis (Figure 3.8) confirmed the correlation of light scattering to FVa-810 subunit dissociation by detecting the aPL-bound FV-810 both by cc-FVaH (lane 1) and ct-FVaL (lane 3) antibodies. After thrombin mediated cleavage followed by 5 hours of EDTA treatment, FVaH was not detected (lane 2) while FVaL was (lane 4) confirming that FVaH and FVaL were dissociated.  71  1  2  3  4  FV 150 kDa 120 kDa  FVaL  Figure 3.8 Western blot analysis of FV-81.O subunit dissociation.  The correlation of light scattering to FVa-810 subunit dissociation was confirmed initially by detecting FVaH remaining bound to LMV after centrifugation using western blot analysis. FV-810 was bound to LMV initially (lanes 1,3). After thrombin-mediated activation followed by 5 hours of EDTA incubation, FVaH was not detected (lane 2) while FVaL was detected (lane 4). Lanes 1 and 2 were probed with x-FVaH mAb (AHV-1546). Lanes 3 and 4 were probed with c-FVaL mAb (14H12).  72  3.2.6 Chelator-induced cofactor activity loss Having observed that the chelation-mediated FVa subunit dissociation of E96A, D1O2A and D111A is accelerated, we next investigated the rate of chelation-mediated cofactor activity loss conducted under identical conditions (Figure 3.9). Clotting assays revealed that FV-810, E96A, D1O2A and D111A had functional decay rates of 0.9 ± 0.08 min , 1.1 ± 0.07 min 1 , 1.1 ± 0.09 min 1 1 and 3.1 ± 0.1 min , respectively (Figure 3.9). As expected, the relative cofactor 1 activity of mutants at 0 mm compared to FV-810 was inhibited as shown (Figure 3.5 and 3.6). Interestingly, the loss of function for FV-810, E96A and D1O2A was markedly faster than their corresponding subunit dissociation rates (Figure 3.7) suggesting that metal ions may independently affect FVa function and subunit association. D111A clotting activity was negligible within 1 mm  of EDTA  treatment. This observation was anticipated based on the spontaneous subunit dissociation observed by the light scattering (Figure 3.7 G).  3.2.7 Metal ion measurement To directly investigate the effect of mutation on metal ions bound to FV/FVa, trace element techniques (ICP-MS and GFAAS) were utilized to compare on a molar basis the amount of Ca 2 and Cu 2 bound to FV-810 or mutants. ICP-MS is a widely used trace element technique since it can carry out rapid multi-element determinations with high sensitivity [181]. Unfortunately, this technique has a limitation in detecting several elements that overlap a major  73  D > U >1 ‘I  ‘I-I  C.)  C 0  C-) 0  2.5  5.0  Time (mm)  Figure 3.9 Loss of cofactor activity of thrombin-cleaved FV-810 and mutants upon chelation. D1O2A () or D111A (.) (100 nM) was bound to PCPS (20 tM) in HBS+Ca (5 mM) and cleaved by thrombin (100 nM) for 5 mm. EDTA (5 2 mM) was then added at various time points, aliquots were taken and diluted appropriately to be assayed for clotting activity. The rate of activity loss observed was represented by solid lines fitted to a single-phase exponential decay function to the data. FV-810 (Li), E96A  (.),  74  spectrum produced by ions generated from argon gas used in the protocol. Of Ca (97 % abundant) suffers an these elements, the most abundant Ca isotope 40 Ca isotope (2 % interference from 40 Ar. In order to avoid this problem, the 44 abundant) was measured instead. The calibration curves were constructed for Ca (Figure 3.10 A) and Cu (Figure 3.10 B). They both showed linear relations 44 within  the  concentrations  used.  Various  metal  ions  were  measured  simultaneously but only Cu 2 was detected to bind to natural FVa by this method. The blank intensily was too large to effectively measure the amount of 44 Ca bound due to its low abundance by this method. In order to measure this successfully, I would need about bOX more concentrated protein. Also the samples in buffer 2 showed very large background noise so could not be used for the study. ICP-MS showed that FVa, at a concentration of 10 ng/ml, bound to 2 in an approximately 1:1 molar ratio, which has been observed previously Cu [1361. However, at other concentrations, the background noise was too large for 2 and accurate measurement. Since I could not use this method to measure Ca the method yielded very low signal to noise ratios, I decided to use GFAAS instead. GFAAS is a highly sensitive spectroscopic technique that provides excellent detection limits for measuring concentrations of metals in aqueous and solid samples. This technique has a smaller working range (1-100 pg/L) compared to ICP-MS (0.01-1000 pg/L), takes longer analysis time than ICP-MS and has high matrix interference, but has no spectral interferences and is more cost-effective  75  C.) 0) Cl) 3  0  C.)  10  0  0)  E  20  40  30  50  B  U)  200000 (I)  150000  100000 50000 0 0  2.5  5  7.5  10  [Me] (ppb)  Figure 3.10 ICP-MS standard curves. The calibration curves were each prepared for (A) 44 Ca and (B) Cu with concentrations, 0 ppb, 1 ppb, 10 ppb and 50 ppb for 44 Ca and 0 ppb, 1 ppb and 10 ppb for Cu. (A) The 44 Ca standard curve showed a linear relationship within the concentrations used, y=1492.5x ± 142.1. (B) The Cu standard curve showed a linear relationship within the concentrations used, y=23729x ± 788.5.  76  compared to ICP-MS. Since GFAAS can analyze both divalent metal ions of interest, it was used to compare the moles of Ca 2 (Figure 3.12 A,B) and Cu 2 (Figure 3.12 C,D) bound to FV-810 or mutants before (Figure 3.12 A,C) and after (Figure 3.12 B,D) conversion to FVa by thrombin. All sample measurements were within the linear regions of the calibration curves (Ca ; Figure 3.11 A, Cu 2 ; 2 Figure 3.11 B). Both non-cleaved and pre-cleaved FV-810 were shown to bind 2 and Cu Ca 2 in an approximately 1:1 molar ratio, which is consistent with the observations made by others [134,136]. E96A and D1O2A were identical to FV 810. A sharp difference was observed for D111A, which was found to have a significantly reduced number of moles of Ca 2 bound per mole of protein (FV; 0.17:1, FVa; 0.14:1). These observations quantitatively confirm the direct involvement of FV/FVa Aspill in Ca 2 binding. The measurement of Cu 2 unexpectedly revealed that mutation of Asplil also reproducibly decreased its stoichiometry  (RI;  0.54:1, FVa; 0.04:1). These data indicate a complicated FVa.  subunit interaction mechanism by suggesting that correct coordination of one metal ion may facilitate that of the other.  3.3 DISCUSSION  -dependent interactions between the non-covalent subunits of FVa 2 Ca are essential for effective coagulation [33], yet little is known about the facilitating amino acids. In an attempt to quantitatively demonstrate the effects of mutating potential Ca -binding residues, G1u96, AsplO2 and Aspill, I 2  77  1.4  1.2  0.8 C.)  a)  U)  0.6  C.)  0.4  I..  0.2  0  U)  -o  0  2.5  5  7.5  10  0  2.5  5  7.5  10  a)  .0 0 U) .0  ‘I.’ .  0) -J  [Me] (ppb)  Figure 3.11 GFAAS standard curves.  The calibration curves were each prepared for (A) Ca 2 and (B) Cu 2 with concentrations, 0 ppb (0 pM), 2.5 ppb (0.0625 pM), 5 ppb (0.125 pM) and 10 ppb (0.25 pM). (A) The Ca 2 standard curve showed a linear relationship within the concentrations used, y=0.1254x ± 0.0308. (B) The Cu 2 standard curve showed a linear relationship within the concentrations used, y=0.0051x ± 0.0004.  78  12AB.  0 0  57510 FV(nmoles)  Figure 3.1.2 Bound Ca ’ and Cu 2 21 in FV-810 and mutants.  Various concentrations of FV-810 (1k), E96A (.), D1O2A () or D111A (.) were dialyzed extensively against HBS-Me then dialyzed again against HBS+Ca (10O 2 (1OO fM). They were then further diluted 10-fold in 0.1% HNO 2 nM)+Cu . The 3 amounts of Ca (A) and 2 2 (C) bound to FV-810 or mutants was measured by Cu graphite furnace atomic absorption spectrometry. In order to assess metal ions bound to pre-cleaved FV-810 and mutants, FV-810 (Ii), E96A (.), D1O2A () or D111A (.) (10 1 iM) was cleaved with thrombin (500 nM) for 5 mm at 22 °C and diluted to various concentrations. They were then dialyzed against (100 nM)+Cu 2 HBS+Ca (100 fM) and further diluted 10-fold in 0.1% HNO 2 . The 3 amounts of Ca 2 (B) and Cu 2 (D) bound to cleaved FV-810 or mutants was again measured by graphite furnace atomic absorption spectrometry. All readings were done in duplicate.  79  measured metal ion binding after mutation to Ala and correlated this to function. I measured the amount of Ca 2 and Cu 2 bound to E96A, D1O2A and D111A expecting to see the difference in stoichiometry at least for Ca 2 in all three mutants. However, this was only observed with D111A. Consequently a 1stage/2-stage discrepancy and spontaneous subunit dissociation were only seen for D111A. For these reasons, D111A and, E96A and D1O2A will be discussed separately. The parental FV-810 I used in this study produces wild type FVa after thrombin cleavage, but has a truncated B domain that does not require proteolytic activation for cofactor function [72,182]. This unique reagent enabled us to investigate the Ca -dependence of the intramolecular FVaH-FVaL domain 2 interaction, which is thought to be fully formed prior to excision of the B domain [72,1821. Figure 3.13 A depicts the proposed Ca -binding loop (Lys93-Aspll2) 2 [125] as projecting from the FVaH domain within FV-810 and associated with the FVaL domain. Structural models based on the homology with a known crystal structure of ceruloplasmin [121] and the crystal structure of bovine FVa [125] suggested amino acids, some overlapping, within Lys93-Aspll2 to interact with 2 forming a loop that might be necessary for the A1-A3 interaction. Inferred Ca from the crystal structure of bovine FVa 1 [125], the Cu -binding site is 2 represented as coordinating to only the FVaL domain, although discrepancies exist to homology models [121]. GFAAS revealed a stoichiometry of 1:1 RI , confirming previous measurements for the high affinity binding site 2 81O:Ca  80  A: FV-810, E96A, DIO2A  slow  inhibited  B: DIIIA  ha fast  0©  inhibited  Figure 3.1.3 Effects of FV G1u96, Aspl.02 or Asplil mutation on metal binding and subunit association. (A) Before and after cleavage by thrombin, FV-810, E96A and D1O2A are shown to bind one mole of Ca 2 and Cu 2 as previously reported [134,1361. Upon chelation, thrombin-cleaved FV-810, E96A and D1O2A rapidly result in an inhibited heterodimeric species, which slowly dissociates. (B) In contrast, D111A has no Ca 2 and partially inhibited Cu 2 binding, leading to spontaneous subunit dissociation and complete loss of Cu 2 binding after cleavage by thrombin.  81  [134]. The same molar ratio was derived for Cu , which was also reported 2 previously [136]. As predicted by the X-ray structure of bovine FVa 1 [125] and homology models based on ceruloplasmin [121], mutating Aspill to Ala was found to have lost the interaction with Ca , which is represented in Figure 3.13 2 B as having no Ca -binding loop. Surprisingly, D111A also had impaired Cu 2 2 binding although the site was predicted to be unrelated to Aspill [121,125]. By our methodology, the Cu 2 stoichiometry was approximately 50% and is consequently represented in Figure 3.13 B as having a partial binding site. The “50 °h Cu 2 binding in FV D111A was observed under different Cu 2 concentrations excluding the possibility that the experimental Cu 2 concentration used coincided with a dissociation constant of Cu 2 and FV. In these experiments, it should be noted that the extensive dialysis of FV against Cu 2 free buffer still allowed Cu 2 to remain bound. Therefore, it is possible that the truncated B-domain of FV D111A held FVaH and FVaL in an orientation that -binding site with lower affinity. An alternative explanation 2 created a partial Cu -binding site when the 2 is that only half of FV D111A was able to form a Cu partial B-domain was intact. My experiments could not distinguish between these possibilities, however, clearly showing that the Cu 2 binding of D111A was altered by mutation at the same time as Ca . 2 FV-810 and D111A function and metal binding were further compared after excision of the B domain by thrombin. Light scattering experiments demonstrated that unlike cleaved FV-810, which is stable in the presence of  82  metal ions, D111A undergoes spontaneous subunit dissociation once cleaved into FVaH/FVaL by thrombin. While metal binding to cleaved FV-810 was unchanged,  thrombin cleavage of D111A resulted in the complete loss of bound Cu , which 2 implies that FVaH/FVaL subunit interactions are needed to create the Cu 2 binding site. The finding that a single amino acid substitution affects both Ca 2 and Cu 2 binding, suggests that interdependent metal ion-coordinating sites may be necessary for subunit stability. An amino acid network existing between the two bound metal ions may explain the binding interdependency between Ca 2 and Cu . The bovine FVa 2 crystal structure suggests that Cu 2 binding is coordinated by l-1is1802, His1804 (His1815, His1817 in human) and Asp1844 (Asnl857 in human) which are all in the A3 domain. Conversely, the theoretical models suggest the putative Cu 2 binding site involves His85 in the Al domain, and His1815 and His1817 in the A3 domain, and therefore, potentially stabilizing the Al-A3 subunit interface directly. The latter homology models are more favorable of the amino acid network between Ca 2 proposed here. 2 and Cu  In this context, the Ca 2 binding site  could be linked through hydrogen bonding to His85, which in turn is predicted to bind Cu . According to this model, it is possible that the binding sites of Ca 2 2 and Cu 2 communicate through His85. It should be noted that this proposed network is not clearly present according to the bovine FVa 1 crystal structure which predicts that the bound Cu 2 is too far from any potential ligands in the Al domain. Instead, Cu 2 may be located adjacent to the Ca 2 binding site near the  83  A1-A3 interface. We cannot say at this time if either model is correct. To explain  inconsistencies between the bovine FVa crystal structure and our studies involving human recombinant R/, possible species discrepancies may exist, or APC-inactivation of FVa may cause altered conformation at the metal-binding interface. A recently published theoretical model based on the bovine FVa X-ray crystal structure with the A2 domain (ceruloplasmin-based) modeled into the human FVa structure to maximize subunit contacts, suggests that Cu 2 is coordinated by both the Al and the A3 domain [1261. Due to these obvious discrepancies between the existing models, it is impossible to ascertain the location of the metal ion binding sites and how they are interconnected. At 2 present, the data presented here suggest the possibility of interdependent Ca and Cu 2 binding, supported by the ceruloplasmin homology model. Given the paradigm that Ca 2 is essential for FVa activity [15] and D111A has insignificant Ca 2 binding, retention of considerable cofactor function (“60%) was unexpected. A functional configuration between the FVaH and FVaL domains in D111A is implied, even though the metal ions are not properly coordinated (Figure 3.13 B). This observation was enabled by use of the truncated B domain, which does not require excision for clotting activity. Interestingly, excision by thrombin caused inhibition of D111A due to subunit dissociation. Therefore, in addition to the known function of concealing RI cofactor activity [72,182], the results presented here indicate a new role for the B domain in stabilizing interactions between the FVaH and FVaL domains  84  independent of normal Ca 2 binding. Further experimentation is required to evaluate if B domain-mediated folding of RI contributes to metal ion binding site formation, which our results suggest are critical for stabilizing functional FVaH/FVaL interactions only after generation of FVa. In addition to D111A, I mutated Glu96 and AsplO2 which are also found in the Ca -binding loop. These amino acids are not predicted by the bovine FVa 2 crystal structure to be directly involved in Ca 2 binding [125] although homology modeling based on ceruloplasmin predicts that they are [121]. The bovine FVa 1 crystal structure furthermore predicts that G1u96 participates in the FVa subunit interaction by forming a direct hydrogen bond with His1804 in the A3 domain [125]. Unlike D111A, the stoichiometry for Ca 2 and Cu 2 bound to E96A or D1O2A was shown here to be identical to FV-810. E96A and D1O2A nevertheless had reduced cofactor activity and moderately enhanced chelator-mediated subunit dissociation rates, strongly supporting an indirect role in metal ion binding. Combined with the crystallography [125], these data can be interpreted as Glu96 and AsplO2 having an involvement in intersubunit contact, which is affected by mutation through a subtle effect on metal binding that increases the susceptibility to chelation. At this time, we cannot ascertain the roles of G1u96 and AsplO2 in the intersubunit stability of FVa. A few suggestions are made as follows to explain the faster chelator-mediated dissociation and reduced cofactor activity observed for E96A and/or D1O2A: 1) enhanced accessibility of the chelator; 2) direct but partial involvement in Ca 2 coordination; 3) indirect  85  involvement in Ca 2 coordination but a direct role in intersubunit contact. Despite the unaltered stoichiometry measured in the current report, weaker metal binding due to mutation is possible. Deriving equilibrium dissociation constants for metal binding would help to resolve these possibilities. However, dissociation constants could not be measured by the method used for the following reasons: 1. Cu 2 bound to FV/FVa could not be released even after extensive dialysis; 2. Heterodimeric FVa dissociates without the adequate metal ions present and for human FVa has not been shown to be reversible. Further studies to measure metal binding affinities are required to ascertain the roles of G1u96 and AsplO2. However, since the chelation-mediated subunit dissociation of thrombin-treated E96A and D1O2A was accelerated compared to FV-810, it is most probable that they are involved indirectly in Ca 2 coordination. To our knowledge, a role for Cu 2 in FV function has not been reported. Whether the bound Cu 2 is fundamentally involved with FVa cofactor activity, with the assembly of FVa subunit heterodimer or both is not known. However, in human FVIII, Cu 2 enhances the affinity between subunits [142]. Ca 2 binds to a -binding loop of FV [143,1441 and 2 region in FVIH nearly identical to the Ca results in conformational changes that likely participate in the intersubunit association and in the formation of active cofactor. Roles for these metal ions in human FVa subunit association and function have not been reported. For bovine FVa, Ca 2 alone appears to facilitate reassociation of isolated subunits giving rise  to full function [68]. In our studies, it was not possible to completely remove  86  2 bound to FV-810 by extensive dialysis against a chelating resin (not Cu shown). Therefore, the involvement of Cu 2 in intersubunit association and function of human FV/FVa cannot be excluded at this time. In the current report D111A, as a consequence of affected Ca 2 and Cu 2 binding, resulted in rapid dissociation of FVaH and FVaL even in the presence of metal ions resulting in cofactor function reduction. This implies the possible involvement of Ca , Cu 2 2 or both in FVaH-FVaL interaction with the possibility of the interdependent binding of Ca 2 and 2 Cu facilitating FV/FVa function and subunit stability. The loss of FVa cofactor activity is expected upon subunit dissociation induced by chelation. However, it is not clear whether the activity loss occurs as a consequence of subunit dissociation or the loss of metal ions. The chelator induced subunit dissociation rate measured quantitatively for cleaved FV-810 in the current work (t 112  “  35 mm) was comparable to that previously reported  qualitatively for human FVa [140]. Interestingly, this was approximately 10-fold slower than the rate of activity loss we measured, which is similar to several previously published  observations using analogous methods in  solution  [134,137,183]. Although these reports differ from results acquired using a unique solid-phase activity assay [139], the preponderance of evidence in the literature, combined with the data presented here suggest FVa cofactor function is lost due to chelation prior to subunit dissociation (Figure 3.13 A). The most probable explanation is that the loss of Ca 2 and Cu 2 upon chelation results in inactive heterodimeric FVa. This inactive species is short lived eventually losing  87  its intersubunit contact. A similar mechanism was supported by Wakabayashi et al. [142] who observed formation of an inactive FVIII heterodimer in the absence of metal ions. Only in the presence of both Ca 2 and Cu , did these 2 authors observe a high FVIII specific activity. It is speculated, therefore, as in FVIII, that FVa subunits may interact very weakly in the absence of metal ions  but require both Ca 2 and Cu 2 for subunit interaction and function. Unlike in FVa, the B domain in R/ is shown to link the FVaH and FVaL in a partially functional conformation independent of metal ions. These data suggest that the subunit interaction of R/ requires metal ions and/or the structural integrity of the B domain. Compared to homozygous RI deficiency, a relatively large number of mutations associated with X chromosome-linked FVIII deficiency has been identified and has enabled correlations between individual amino acids and function [184]. Patients carrying substitutions in the probable Ca -binding loop 2 in RI have not been reported. However homologous mutations have been identified in FVIII corresponding to GlullOVal (homologous RI amino acid residues are shown in parentheses; G1u96), Aspll6Tyr (AsplO2), Glul22Lys (G1u108), and Aspl26His (Aspll2), which result in hemophilia A [185-188]. Homologous RI mutation without compensating factors would be predicted to be lethal, and consequently the corresponding naturally occurring mutations are unknown. In the current study, both plasma clotting and purified prothrombinase  88  assays revealed nearly 100% inhibition after cleavage of RI D111A by thrombin, in contrast to only a partial loss of activity without thrombin pre-treatment. This type of 1-stage/2-stage assay discrepancy has not been reported clinically for FV to our knowledge but has been shown in hemophilia A [189,1901. Pipe and co workers have identified hemophilia A-causing naturally occurring mutations that result in accelerated subunit dissociation in FVIII. However these mutations have been localized to residues predicted from homology models to facilitate direct subunit contact and not metal-binding [189,190]. These mutations in the A1-A2, A2-A3, and A1-A3 interfaces were characterized to disrupt interdomain interactions resulting in the increased rate of inactivation of FVIIIa due to A2 subunit dissociation. Unlike FVa, the A2 subunit in FVIIIa is associated with the Al subunit through a weak ionic interaction and the dissociation of the A2 subunit from the A1/A3-C1-C2 heterodimer results in the loss of procoagulant activity [191]. Therefore, thrombin-mediated activation increases the A2 domaindissociation more pronouncedly in the mutants resulting in the activity loss. This phenotype is accurately identified by 1-stage clotting assays being far less inhibited than 2-stage assays in which pre-activation of the procofactor enables significant subunit dissociation prior to conducting the assay. The laboratory generated single mutant D111A behaves similarly to those FVIII mutants due to the loss of entire heavy chain, confirming the absolute requirement for intact intersubunit interaction for function. This raises the possibility that FV mutation resulting in impaired metal binding or subunit interface contact may be a  89  previously overlooked basis of pathology.  90  4. FACTOR V-DEFICIENT PATIENT 4.1 HYPOTHESIS AND SPECIFIC GOALS  A patient who is an 82-year-old male of Scottish-descent was identified in 1954 as having no detectible FV activity in plasma (<2 %). The pedigree of the patient is shown in Figure 4.1. The patient tends to bruise easily and after significant trauma, dental extraction or prostatectomy required special treatment to stop bleeding. However, he has never showed spontaneous bleeding. The initial hypothesis in order to rationalize his minimal bleeding tendency despite the apparent lack of an essential coagulation protein was that the patient’s deficiency of plasma FV is compensated by the platelet FV pool. To address this hypothesis, the specific goals were: 1. Determine the DNA sequence of the entire FV gene of the patient. 2. Investigate the patient R/ antigenic and activity levels in both his plasma and platelets. Failing the hypothesis that platelet FV alleviates bleeding, an alternate hypothesis is that the patient’s lack of RI is compensated by hypercoagulation in another aspect of the coagulation pathway. To address this amended hypothesis, the specific goals were: 1. Investigate the patient’s intrinsic pathway by activated partial thromboplastin time (APTT)-clotting assays. 2. Investigate the patient’s extrinsic pathway by prothrombin time (PT) clotting assays.  91  <2%  Figure 4.1 Pedigree of plasma FVa Activity.  Hospital laboratory analysis of FV activity identified the patient (red) as having less than 2% plasma factor V function, while the wife was normal. Their four heterozygous children had FVa function of ‘5O% and are clinically asymptomatic. (square, male; circle, female)  92  4.2 RESULTS 4.2.1. Plasma and platelet preparation  Platelets were minimally activated as determined by the expression of platelet granule-release marker, CD62P, measured using flow cytometry. The purified platelets contain less than 5 % contamination by red blood cells and less than 0.1 % contamination by white blood cells as measured on a cell counter. Platelets from all three donors were diluted to 2.1X10 /ml. 8  4.2.2 Identification of mutation  In order to identify R/ gene mutation(s) in the patient, nucleotide sequences of all 25 exons and exon-intron boundaries of the FV gene were analyzed. Results from sequencing of the FV gene revealed that the patient had a homozygous mis-sense mutation in exon 15 (A5279G) changing the codon for Tyr1702 to Cys (Y1702C), which has been previously reported (Figure 4.2). Thus, significant effects on hemostasis caused by only a single amino acid mutation in RI, predicts a significant functional role for Y1702.  4.2.3 Factor V activity in plasma and platelets  Conventional clotting assays were used to assess the plasma and platelet R/ function of the patient, his wife and his son. There was no detectible activity level of FV in both plasma and platelets of the patient. The heterozygous asymptomatic child of the patient had ‘‘50 % FV compared to normal in plasma  93  T  T  280 G  C  C  T  2  ‘I  C  T  A  C  290 T  C  Nucleotide substitution  fromAtoG  Figure 4.2 DNA sequencing. DNA sequence chromatogram shows the A5279G mutation in exon 15. The sense strand shows that the patient is homozygous for the A —* G point mutation. The nucleotide substitution is marked by the arrow.  94  and ‘30 % FV compared to normal in platelets (Figure 4.3). The assay had a detection limit down to 0.004 nM FV, which is 0.02 % of the physiological concentration. This method was 100-fold more sensitive than the initial clinical assays that concluded that the FV level was less than 2%. Therefore, we concluded that the patient has less than 0.02 % FV activity in both plasma and platelets.  4.2.4 Factor V antigens in plasma and platelets Western blot analysis was used to assess the plasma and platelet FV antigenic level of the patient, his wife and his son. The data demonstrate that the patient has no FV/FVa, the wife has a normal level, and the son has about 50 % in plasma (Figure 4.4 C,D). In platelets, similar results were shown (Figure 4.4 A,B) concluding that platelet FV was not protecting from bleeding. These results suggest that the patient has non-functional FV, either due to a gross conformational change or impaired secretion. The lower detection limit by this method is about 0.1 ng RI, 0.1 % of the physiological concentration. As a result, I conclude that the patient has less than 0.8 % FV antigens in both plasma and platelets.  4.2.5 Compensation by blood coagulation pathway To address the hypothesis that the lack of detectable FV function and antigen was being compensated by opposing hypercoagulation, the patient’s  95  25  20  A  I  B -E•  +  0.75  15 0.5 LL.  10 0.25  5 0  __  P  W  S  U  -  P  W  S  Figure 43 FV clotting activity in plasma and platelets.  Plasma (A) and platelet (B) FV function was assessed for wife (W), son (S) and patient (P) by conventional clotting assays using FV-depleted plasma initiated by a source of Ca , tissue factor and anionic phospholipid (Innovin). The observed 2 clotting times were converted to equivalent FV clotting activities assuming that 20 nM FV has a clotting activity of lU/mI.  96  wsP  wsP  wsP  wsP  FV FV220 FVaH FVaL B  FV FV220  FVaH FVaL D  C Resting  Activated  Resting  Activated  Figure 4.4 FV/FVa antigens in plasma or platelets. Western blot analysis showed reduced resting or activated platelet (A,B) or plasma (C,D) FV from wife (W), son (S) and patient (P) detected with either x FVaL (A,C) or cL-FVaH (B,D) mAb. The positions of FV and FVa were determined by comparison to purified controls (not shown).  97  coagulation pathway was divided into the intrinsic and extrinsic pathways and evaluated by clotting assays. Three variations of an experiment to test for hypercoagulation were conducted, including 1) titration of purified FV against either patient or RI-depleted plasma, 2) titration of normal plasma against either patient or RI-depleted plasma, and 3) titration of either patient or RI-depleted plasma against 10 % normal plasma. These assays were designed to compare the clotting ability of patient plasma and commercially available RI-depleted plasma. If there is a compensating hypercoagulant component in the patient plasma, differences will be detected by all three methods. There was no difference in the extrinsic pathway between the patient and RI-depleted plasma in all three experiments (Figure 4.5 B,D,F). Unlike the extrinsic pathway, hypercoagulation was obvious when evaluating patient plasma coagulation initiated by the contact phase of activation. When purified FV-810 was added, patient plasma was shown to significantly accelerate clotting compared to RI-depleted plasma (Figure 4.5 A). When patient or RI-depleted plasma and normal plasma were mixed in various ratios, patient plasma again sped up the clot formation compared to RI-depleted plasma (Figure 4.5 C). Lastly, when higher concentrations of patient plasma were added to 10 % normal plasma, the clotting time was shortened whereas addition of higher concentrations of RI-depleted plasma resulted in no change (Figure 4.5 E). These results strongly suggest that some component in the  98  0  C.,  20  40  0  20  40  50  75 100  a, Cl,  a,  E 0) 0  C-)  0  30  60  90 U  E 60  15  4o  10  9fl  0  25  50  75 100  5  F  0  25  [FV] (nM)  Figure 4.5 Compensation by blood coagulation pathway.  The intrinsic (A, C, E) and extrinsic (B, D, F) pathways of the patient (.) were compared to commercially available RI-depleted plasma (11). Purified FV was titrated against either patient or FV-depleted plasma (A, B). Normal RI was titrated against either patient or RI-depleted plasma (C, D). Finally either patient or RI-depleted plasma were titrated against 10 % normal plasma (E, F).  99  patient plasma that participates in the intrinsic part of the blood clotting cascade may be a hypercoagulant compensating for the lack of FV in the patient.  4.2.6 Blood proteins in the intrinsic pathway  The activity levels of the examined blood proteins, FXII, Fy11, FVIII, FIX, FX, FXI, HK, PKK and antithrombin were normal. The results are summarized in  Table 4.1.  4.3 DISCUSSION  Since homozygous FV deficiency is exceedingly rare, little is known about the genetic variables leading to this bleeding diathesis. In the current study, the patient who has been previously characterized in the clinical laboratory to have no detectible FV activity in plasma (< 2%) was studied further. Despite the lack of FV in plasma, the patient only suffers from a minor bleeding tendency strongly suggesting the existence of additional genetic components or environmental factors that may control the severity of his disease. DNA sequence analysis of his FV gene revealed that he carries a single amino acid substitution, Y1702C, in the A3 domain; he is homozygous for this mutation. To explain the patient’s unexpected minimal bleeding tendency, his  RI  level in plasma and platelets and  other blood proteins were tested. My hypothesis was that either the platelet FV or other  blood  proteins compensate for  his  lack of FV  in  plasma.  100  Blood protein  Activity (% normal)  Antithrombin  91  FXII  95  FX  90  FVlll  108  FVII  123  FIX  100  FXI  67  HK  81  PKK  109  Table 4.1 Intrinsic/contact phase pathway blood protein activity.  101  The FV Y1702C mutation has been reported previously and it is believed to be the most frequent cause of FV deficiency in the Italian population [159]. My observations suggest an additional ethnic distribution. The Y1702C FV was shown to possibly impair secretion when transiently transfected in COS-1 cells, resulting in 1.8 % of the wild type FV in clotting activity and antigenic level [192,193]. Castoldi et al. investigated the expression of Y1702C FV at the mRNA and protein levels in a patient. They showed that the  RI  Y1702C mRNA level was  normal but the R/ antigen level was undetectable suggesting that the Y1702C mutation impairs the posttranslational modifications and/or secretion and/or stability of secreted FV [194,195]. Tyr1702 in RI is conserved in all three A domains of human RI, FVIII and ceruloplasmin [195] (Figure 4.6). Tyr1702 and conserved Tyr residues in other A domains are shown in the homology model in Figure 4.7. When the role of Tyr1702 was investigated by inspection of the three-dimensional structure of ceruloplasmin, it was shown that the homologous residue, Tyr860, was buried in the domain core interacting with a Pro residue, Pro791, via two hydrogen bonds. Pro791 in ceruloplasmin is homologous to Pro1618 in RI and is also conserved in all A domains (Figure 4.8 A). Also Tyr1702 and Pro1618 are conserved in bovine  RI  forming hydrogen bonds  shown in the crystal structure of FVa 1 (Figure 4.8 B). Thus, the R/ Y1702C  mutation is understood to potentially cause protein instability by disrupting the A3 domain scaffold [195]. Also, introduction of a new Cys residue can interfere with the correct disulfide bridge formation between nearby Cys residues [195].  102  FVAI FVIIIAI CPAI  54 I S G L L G P T L Y A E 68WMGLLGPTI QAE 7OWLGFLGPII KAE  138 P C L T H I Y Y S H E N 152LCLTYSYLSHVD 154NCVTRIYHSHID  FVA2388D-GILGPIIRAQ471QCLTRPYYSDVD FVIIIA2 446 S G L G P L L Y GE 527R C L T R Y Y S S F V N CPA2 426HLG1 LGPVIWAE 5I4VCLAKMYYSAVD -  FVA316I2HLGILGPIIRAE1696ACRAWAYYSAVN FVIIIA3I755HLGL LGPYI RAE 1831DCKAWAYF SDVD CPA3 785 HLGI LGPQLHAD 854AC1 PWAYYSTVD  Figure 4.6 Conserved Tyr1702 and Pro1618 in the A domains. Alignment of sequences near the conserved Tyr and Pro residues of FV, FVIII and ceruloplasmin (CP) is shown. The conserved Tyr residues are in blue and pro residues are in red.  103  Figure 4.7 Conserved Tyr1702 in the A domains. Tyr1702 and its conserved Tyr residues in the A domains are shown in the homology model (Al domain, yellow; A2 domain, green; A3 domain, blue). The Tyr residues are shown in orange.  104  Figure 4.8 Role of Tyr1702.  Two hydrogen bonds between Tyr1702 and Pro1618 are shown in both the homology model (A) and the FVa model (B). The Y1702C mutation would eliminate this interaction resulting in structural instability.  105  Figure 4.9 shows Tyr1702 in the homology model (A) and the FVa 1 model (B). In both structures, there are nearby disulfide-bonded Cys residues in the A3 domain. The introduction of a novel Cys can interfere with the formation of this normal disulfide bond. Also there are free Cys residues that can potentially interact with the newly introduced Cys residue, disrupting the proper folding of the molecule. Therefore, these models strongly suggest that the Y1702C mutation causes deficiency of FV by gross folding alterations that may prevent secretion from cells. The discrepancy in bleeding phenotype has been observed in the past between a homozygous patient for Y1702C mutation and her asymptomatic brother with the same FV genotype [159]. It has long been understood that FV plasma levels are a poor indicator for a bleeding phenotype. Such phenotypic discrepancies are unexplained, leading to our hypothesis that either higher level of patient platelet FV may make up for the low plasma level or the existence of additional genetic components may compensate for the lack of FV in the patient. FV circulates in blood, partitioned between plasma (‘.‘80 %) and platelet x-granules (r’.’20 %). Plasma RI circulates as an inactive procofactor whereas platelet FV is found to be partially activated [85]. Since it is partially active and is close to the procoagulant surface of platelets, platelet RI is believed to be important for hemostasis for the assembly of the prothrombinase complex at a site of vascular injury. My initial hypothesis was that the lack of plasma RI in the patient is compensated by his platelet RI. However, I have failed to detect any  106  Figure 4.9 Tyr1702 and nearby cysteine residues. 1 model (B) showing the location of Tyr1702 The homology model (A) and the FVa (yellow) that is mutated to Cys in the patient (FVaH, green; FVaL, red). Other Cys residues (disulfide-bonded, blue; free, orange) are located nearby suggesting incorrect disulfide formation may occur. The A3 domains are circled.  107  patient platelet FV both antigenically or functionally suggesting that the lack of plasma FV is not compensated by platelet R/ in this patient. To test the alternative hypothesis that an additional difference between the patient and normal plasma is compensating for the patient’s lack of R/, coagulation activily was investigated by using both APTT- and PT-clotting assays. Several experiments were designed to compare the patient plasma and commercially available plasma that is immuno-depleted of FV to determine if the patient plasma accelerates clot formation in the latter. It was shown that when initiated through the contact phase by APTT reagent, the patient plasma significantly accelerated clotting compared to adding the same proportion of normal plasma to the immunodepleted FV plasma. Unlike APTT-clotting assays, PT-clotting assays showed no differences between the patient and FV-depleted plasma. These observations suggested that hypercoagulable components in the patient’s intrinsic pathway or contact phase might be compensating for his lack of FV. The common pathway of the patient is implied to be normal since only the APTT-clotting assays showed differences. If the common pathway was compensating, both pathways would be accelerated. I further examined the functional levels of blood proteins that could affect the patient’s coagulation cascade. The proteins examined included Fy11, FVIII, FIX, FX, FXI, FXII, HK, PKK and antithrombin. The activity levels of all of these  were shown to be normal by conventional clotting assays using deficient plasmas. Therefore, I have eliminated the possibility of these proteins compensating for  108  the absence of FV in the patient. Other genetic or environmental factors that affect his intrinsic pathway other than those tested blood proteins are believed to be the basis of his mild rather than severe bleeding diathesis. This current study suggests that the reason for phenotypic differences observed among some FV-deficient patients despite same genotypes may be due to other existing genetic or environmental factors that influence the blood coagulation pathway. Our observations demonstrate that the patient’s lack of detectable RI in the plasma and platelets is compensated by components that act in the intrinsic part of the blood coagulation pathway. It is also possible that a trace amount of RI that is not detectable by our experiments could be easing the bleeding tendency of the patient. A partial rescue of the lethal  RI-/  phenotype by introducing low levels of FV has been shown in mice suggesting that trace expression of RI could improve hemostasis significantly [196]. As in mice, it is probable that complete R/ deficiency in humans is not compatible with life since the physiological concentration of thrombin required for hemostasis is believed to be unachievable without RI [151]. I suggest that homozygous FV deficient patients without a bleeding tendency either posses a trace amount of RI that cannot be detected by conventional test or co-inherit risk factors for thrombophilia, which can influence the overall coagulation cascade, therefore positively counteract the clinical outcome of RI deficiency.  109  5. SUMMARY 5.1 ROLE OF GLU96, ASP1O2 AND ASP111 IN FACTOR V  The binding sites for Ca 2 are adjacent to each other and are 2 and Cu located near the Al and A3 domain interface (Figure 5.1). The data presented here suggest that only Asplil directly interacts with the Ca -binding site, which 2 in turn is interconnected to the Cu -binding site. Since D111A, lacking two 2 essential metal ions, causes the rapid loss of function when cleaved by thrombin due to subunit dissociation, Ca , Cu 2 2 or both are implied to be essential for both subunit interaction and function. Unlike D111A, the thrombin-mediated FVa derived from E96A and D1O2A was stable, had only moderately faster subunit dissociation upon chelation and had normal Ca 2 and Cu 2 binding. This study therefore defines the highly conserved acidic segment spanning G1u96-Aspll2 in FV as multifunctional. Of the three amino acids I evaluated, Aspill is essential and likely functions through direct and indirect metal ion interactions. G1u96 and AsplO2 individually influence FV/FVa function by subtle effects at the metal iondependent subunit interface. The parental FV construct used here has a truncated B domain that does not require removal for activity, but produces wild type FVa. Each mutant was ‘.‘40% inhibited, demonstrating specific amino acid involvement in function. Pre treating E96A or D1O2A with thrombin had no effect on activity. In contrast, D111A was inhibited by >90% after cleavage by thrombin, which was explained 2 binding. Thus, by spontaneous subunit dissociation and severely inhibited Ca  110  our results suggest that the B domain maintains FV in a FVa-like functional configuration. Physiological Ca 2 interactions appear to become vital only after B domain excision. -binding residues 2 Further studies with mutations in other potential Ca and hypothesized Cu -binding residues will be useful to investigate the 2 interconnection between the two divalent cations and the role of each individual divalent cation in human FV. Also, similar studies using the full length FV and its variants will provide more insights into the newly suggested role of the B domain.  5.2 FACTOR V-DEFICIENT PATIENT  We have demonstrated a case of severe plasma FV deficiency due to a homozygous Y1702C mutation. However the patient did not show severe symptoms that are normally associated with the deficiency. He had undetectable FV in his platelets eliminating the possibility that his platelet RI was compensating for the lack of plasma FV. The lack of available FV in the patient was compensated by other factors that influence the intrinsic blood coagulation pathway according to the observations made in this study. This finding suggests that the commonly observed phenotypic differences shown among FV-deficient patients with same genotypes may be due to other factors that influence the outcome of the disease. Further studies that analyze the intrinsic and contact pathway of the patient’s heterozygous child may confirm the observations made in this study. Also, showing the similar mechanism of compensation in other FV  111  deficient patients would verify the results shown here.  Figure 5.1 Model: Interdependent metal ion binding facilitates subunit interaction. The crystal structure of bovine FVa is shown (FVaH, green; FVaL, red). Glu96, AsplO2 and Asplil are shown in white. Both Ca 2 and Cu 2 are found near the Al and A3 interface. The data presented in the study show that D111A mutation affects both Ca 2 and Cu 2 binding suggesting that the change in Ca 2 coordination affects the Cu It is shown 2 coordination. here that both Ca 2 and 2 might be required for FVaH-FVaL interaction. On the contrary, E96A and Cu D1O2A mutations do not affect either Ca 2 or Cu 2 binding but result in faster subunit dissociation suggesting that G1u96 and AsplO2 individually influence FV/FVa function by more subtle effects at the metal ion-dependent subunit interface.  112  Reference List 1. Mann K G. Thrombosis: theoretical considerations. Am.J.Clln.Nutr. 1997; 65(suppl): 1657S-1664S. 2. Colman R W, Marder V J, Salzman E W, Hirsh J. Hemostasis and thrombosis: basic princio/es and clinical practice. 2nd edition J. B. Li ppi ncott Co.: Philadelphia, PA, 1987. 3. Mann K G, Krishnaswamy S, Lawson J H. Surface-dependent hemostasis. Semin.Hematol. 1992; 29: 213-226. 4. Bouma B N, Meijers J C M. Fibrinolysis and the contact systerm: a role for factor XI in the down-regulation of fibrinolysis. Thrombosis and Haemostasis 1999; 82: 243-250. 5. Mann K G, Nesheim M E, Church W R, Haley P, Krishnaswamy S. Surfacedependent reactions of the vitamin K-dependent enzyme complexes. Blood 1990; 76:1-16. 6. Kalafatis M, Egan J 0, van’t Veer C, Mann K G. The regulation of clotting factors. Critical Reviews in Euka,yotic Gene Expression 1997; 7: 241-180. 7. Suzuki K, Dahlback B, Stenflo J. Thrombin-catalyzed activation of human coagulation factor V. J.Biol.Chem. 1982; 257: 6556-6564. 8. Mann K G, Jenny R J, Krishnaswamy S. Cofactor proteins in the assembly and expression of blood clotting enzyme complexes. Annu.Rev.Biochem 1988; 57: 915-956. 9. Foster W B, Nesheim M E, Mann K G. The factor Xa-catalyzed activation of factor V. JBio/Chem. 1983; 258:13970-13977. 10. Bachman F. Fibrinolysis. In: Thrombosis and Haemostasis (eds. Verstraete M., Vermylen J., Lijnen H. R., Arnout J.). Leuven Universily Press: Leuven, 1987; 227-265. 11. Hornyak T J, Shafer J A. Role of Calcium Ion in the Generation of Factor XIII Activity. Biochemistry 1991; 30: 6175-6182. 12. Berube C, Ofosu F A, Kelton J G, Blajchman M A. A novel congenital haemostatic defect: combined factor VII and factor XI deficiency [see comments]. Blood Coagul.Fibrinolysis 1992; 3: 357-360. 13. Davie E W, Ratnoff 0 D. Waterfall sequnece of intrinsic blood clotting. Science 1964; 145: 1310-1312.  113  14. MacFarlane R G. An enzyme cascade in the blood clotting mechanism and its function as a biochemical amplifier. Nature 1964; 202: 498-499. 15. Mann K G, Kalafatis M. Factor V: a combination of Dr Jekyll and Mr Hyde. Blood 2003; 101: 20-30. 16. Scandura J M, Ahmad S S, Walsh P N. A binding site expressed on the surface of activated human platelets is shared by factor X and prothrombin. Biochemistry 1996; 35: 8890-8902. 17. Walker R K, Krishnaswamy S. The activation of prothrombin by the prothrombinase complex: The contribution of the substrate-membrane interaction to catalysis. J Biol Chem. 1994; 269: 27441-27450. 18. Giesen P L, Willems G M, Hemker H C, Hermens W T. Membranemediated assembly of the prothrombinase complex. J.Biol.Chem. 1991; 266 (28): 18720-18725. 19. Schwalbe R A, Ryan J, Stern D M, Kisiel W, Dahlb:ack B, Nelsestuen G L. Protein structural requirements and properties of membrane binding by gamma-carboxyglutamic acid-containing plasma proteins and peptides. 3 Biol.Chem. 1989; 264: 20288-20296. 20. Bloom J W, Nesheim M E, Mann K G. Phospholipid-binding properties of bovine factor V and factor Va. Biochemistry 1979; 18: 4419-4425. 21. Krishnaswamy S, Mann K G. The binding of factor Va to phospholipid vesicles. 3 Biol Chem. 1988; 263: 57 14-5723.  22. Krishnaswamy S, Jones K C, Mann K G. Prothrombinase complex assembly. Kinetic mechanism of enzyme assembly on phospholipid vesicles. 3 Biol Chem. 1988; 263: 3823-3834. 23. Kane W H, Lindhout M J, Jackson C M, Majerus P W. Factor Va-dependent binding of factor Xa to Human platelets. J.Bio/.Chem. 1980; 255: 11701174. 24. Krishnaswamy S. Prothrombinase complex assembly. Contributions of protein-protein and protein-membrane interactions toward complex formation. J.Biol.Chem. 1990; 265 (7): 3708-3718. 25. Pryzdial E L, Mann K G. The association of coagulation factor Xa and factorVa. JBiol.Chem. 1991; 266: 8969-8977.  114  26. Nesheim M E, Taswell J B, Mann K G. The contribution of bovine Factor V and Factor Va to the activity of prothrombinase. J Biol Chem. 1979; 254: 10952-10962. 27. Leytus S P, Chung D W, Kisiel W, Kurachi K, Davie E W. Characterization of a cDNA coding for human factor X. Proc.Natl.AcadSci.USA 1981; 81 (12): 3699-3702. 28. Fair D S, Revak D J, Hubbard J G, Girolami A. Isolation and characterization of the factor X Friuli variant. Blood 1989; 73: 2108-2116. 29. Stenflo J. Contributions of Gla and EGF-like domains to the function of vitamin K-dependent coagulation factors. Crit Rev.Eukaryot Gene Expr. 1999; 9: 59-88. 30. Nelsestuen G L, Shah A M, Harvey S B. Vitamin K-dependent proteins. Vitam.Horm. 2000; 58: 355-389. 31. Soriano-Garcia M, Padmanabhan K, de Vos A M, Tulinsky A. The Ca2+ ion and membrane binding structure of the Gla domain of Ca-prthrombin fragment 1. Biochemistry 1992; 31: 2554-2566. 32. Ahmad S 5, Rawala-Sheikh R, Walsh P N. Components and assembly of the factor X activating complex. Seminars in Thrombosis and Hemostasis 1992; 18: 311-323. 33. Davie E W, Fujikawa K, Kisiel W. The Coagulation Cascade: Initiation, Maintenance, and Regulation. Biochemistry 1991; 30:1030-1037. 34. Bajaj S P, Rapaport S I, Brown S F. Isolation and characterization of human factor VII. Activation of factor VII by factor Xa. J.Biol.Chem. 1981; 256: 253-259. 35. Bennett B. Coagulation pathways: interrelationships mechanisms. SeminJ-Iematol. 1977; 14: 301-318.  and  control  36. Huang Z-F, Wun T-C, Broze G J. Kinetics of factor Xa inhibition by tissue factor pathway inhibitor. J.Biol.Chem. 1993; 268 (36): 26950-26955. 37. Hockin M F, Jones K C, Everse S J, Mann K G. A model for the stoichiometric regulation of blood coagulation. J.Biol.Chem. 2002; 277: 18322-18333. 38. Osterud B, Miller-Andersson M, Abildgaard U, Prydz H. The effect of antithrombin III on the activity of the coagulation factors VII, IX and X. Thromb.Haemost 1976; 35: 295-304.  115  39. Abildgaard U. Heparin cofactor and antithrombin. Thromb.Diath.Haemorrh. 1975; 33: 38-42. 40. Lundblad R L, Uhteg L C, Vogel C N, Kingdon H S, Mann K G. Preparation and partial characterization of two forms of bovine thrombin. Biochem Biophys Res Commun. 1975; 66: 482-489. 41. Côte H C F, Bajzar L, Stevens W K, Samis J A, Morser J, MacGillivray R T A, Nesheim M E. Functional characterization of recombinant human meizothrombin and meizothrombin(desFl). J.Biol.Chem. 1997; 272 (10): 6194-6200. 42. Doyle M F, Mann K G. Multiple active forms of thrombin. IV. Relative activities of meizothrombins. 3 Biol Chem. 1990; 265: 10693-10701. 43. Esmon C T, Owen W G, Jackson C M. The conversion of prothrombin to thrombin. V. The activation of prothrombin by factor Xa in the presence of phospholipid. J BioI.Chem. 1974; 249: 7798-7807. 44. Esmon C T, Owen W G, Jackson C M. A plausible mechanism for prothrombin activation by factor Xa, factor Va, phospholipid, and calcium ions. JB1oI.Chem. 1974; 249: 8045-8047. 45. Jackson C M. Mechanisms of prothrombin activation. In: Hemostasis and thrombosis: basic princioles and din/cal practice 1st edition (eds. Colman R. W., Hirsh J., Marder V. J., Salzman E. W.). J. B. Lippincott Company: Philadelphia, 1982; 100-111. 46. Fay P F, Smudzin T M, Walker F J. Activated protein C-catalyzed inactivation of human factor VIII and factor Villa. J.BioI.Chem. 1991; 266: 20139-20145. 47. Bevers E M, Comfurius P, Zwaal R F A. Changes in membrane phospholipid distribution during platelet activation. Biochem Biophys Acta 1983; 736: 57-66. 48. Krishnaswamy S, Nesheim M E, Pryzdial E L G, Mann K G. Assembly of the prothrombinase complex. Methods Enzymol. 1994; 222: 260-280. 49. Higgins D L, Callahan P J, Prendergast F G, Nesheim M E, Mann K G. Lipid mobility in the assembly and expression of the activity of the prothrombinase complex. 3 Biol Chem. 1985; 260: 3604-3612. 50. Higgins D L, Mann K G. The interaction of bovine factor V and factor V derived peptides with phospholipid vesicles. 3 Biol Chem. 1983; 258: 6503-6508.  116  51. Smirnov M D, Esmon C T. Phosphatidylethanolamine incorporation into vesicles selectively enhances factor Va inactivation by activated protein C. JBiolChem. 1994; 269: 816-819. 52. Bakker H M, Tans G, Janssen-Claessen T, Thomassen M C, Hemker H C, Griffin J H, Rosing J. The effect of phospholipids, calcium ions and protein S on rate constants of human factor Va inactivation by activated human protein C. EuriBiochem 1992; 208:171-178. 53. Smirnov M D, Ford D A, Esmon C T, Esmon N L. The effect of membrane composition on the hemostatic balance. Biochemistry 1999; 38: 35913598. 54. Cripe L D, Moore K D, Kane W H. Structure of the gene for human coagulation factor V. Biochemistry 1992; 31: 3777-3785. 55. Jenny R J, Pittman D D, Toole J J, Kriz R W, Aldape R A, Hewick R M, Kaufman R J, Mann K G. Complete cDNA and derived amino acid sequence of human factor V. Proc.NatI.AcadSci.U.S.A. 1987; 84: 48464850. 56. Kane W H, Davie E W. Blood coagulation factors V and VIII: structural and Functional similarities and their relationship to hemorrhagic and thrombotic disorders. Blood 1988; 71: 539-555. 57. Nesheim M E, Foster W B, Hewick R, Mann K G. Characterization of Factor V activation intermediates. 3 Biol Chem. 1984; 259: 3187-3196. 58. Jenny R J, Mann K G. Factor V: a prototype pro-cofactor for vitamin Kdependent enzyme complexes in blood clotting. Bailleres,Clln Haematol. 1989; 2: 919-944. 59. Francis R T, McDonagh J, Mann K G. Factor V is a substrate for the transamidase factor XIIIa. 3 Biol Chem. 1986; 261: 9787-9792. 60. thorelli E, Kaufman R J, Dahibach B. The C-terminal region of the factor V B-domain is crucial for the anticoagulant activity of factor V. 3 Biol Chem. 1998; 273: 16140-16145. 61. Marquette K A, Pittman D D, Kaufman R J. The factor V B-domain provides two functions to facilitate thrombin cleavage and release of the light chain. Blood 1995; 86: 3026-3034. 62. Pittman D D, Tomkinson K N, Michnick D, Selighsohn U, Kaufman R J. Posttranslational sulfation of factor V is required for efficient thrombin  117  cleavage and activation and for full procoagulant activity. Biochemistry 1994; 33: 6952-6959. 63. Pittman D D, Tomkinson K N, Kaufman R J. Post-translational requirements for functional factor V and factor VIII secretion in mammalian cells. JBio/.Chem. 1994; 269:17329-17337. 64. Kalafatis M. Identification and partial characterization of factor Va heavy chain-kinase from human platelets. 3 Biol Chem. 1998; 273: 8459-8466. 65. Kumar H P M, Besman M J, Lundblad R L, Jenny N S, Mann K G. Carbohydrate analysis of plasma factor V and factor VIII. Thromb Haemost 1999; Suppl 82:102. 66. Xue J, Kalafatis M, Mann K G. Determination of the disulfide bridges in factor Va light chain. Biochemistry 1993; 32: 5917-5923. 67. Kalafatis M, Xue J, Lawler C M, Mann K G. Contribution of the heavy and light chains of factor Va to the interaction with factor Xa. BiochemLctry 1994; 33: 6538-6545. 68. Krishnaswamy 5, Russell G D, Mann K G. The reassociation of factor Va from its isolated subunits. J.Bio/.Chem. 1989; 264: 3160-3168. 69. Kane W H, Davie E W. Cloning of a cDNA coding for human factor V 1 a blood coagulation factor homologous to factor VIII and ceruloplasmin. Proc.Nat/.Acad.Sci.USA 1986; 83: 6800-6804. 70. Monkovic D D, Tracy P B. Activation of human factor V by factor Xa and thrombin. Biochemistry 1990; 29: 1118-1128. 71. Kalafatis M, Beck D 0, Mann K G. Structural requirement for expression of factor Va activity. J Biol Chem. 2003; 278: 33550-3356 1. 72. Toso R, Camire R M. Removal of B-domain sequences from factor V rather than specific proteolysis underlies the mechanism by which cofactor function is realized. J.Bio/.Chem. 2004; 21643-21650. 73. Zhu H, Toso R, Camire R M. Inhibitory sequences within the B-domain stabilize circulating factor V in an inactive state. J B/of Chem. 2007; 282: 15033-15039. 74. Kalafatis M, Jenny R J, Mann K G. Identification and characterization of a phospholipid-binding site of bovine factor Va. J B/of Chem. 1990; 265: 21580-21589.  118  75. Kalafatis M, Rand M D, Mann K G. Factor Va-membrane interaction is mediated by two regions located on the light chain of the cofactor. Biochemistry 1994; 33: 486-493. 76. Ortel T L, Devore-Carter D, Quinn-Allen M, Kane W H. Deletion analysis of recombinant human factor V. Evidence for a phosphatidylserine binding site in the second C-type domain. JBiol.Chem. 1992; 267 (6): 4189-4198. 77. Ortel T L, Quinn-Allen M A, Keller F G, Peterson J A, Larocca D, Kane W H. Localization of functionally important epitopes within the second C-type domain of coagulation factor V using recombinant chimeras. J BioI.Chem. 1994; 269: 15898-15905. 78. Ortel T L, Quinn-Allen M A, Charles L A, Devore-Carter D, Kane W H. Characterization of an acquired inhibitor to coagulation factor V. Antibody binding to the second C-type domain of factor V inhibits the binding of factor V to phosphatidylserine and neutralizes procoagulant activity. J Clln Invest 1992; 90: 2340-2347. 79. Izumi T, Kim S W, Greist A, Macedo-Ribeiro 5, Fuentes-Prior P, Bode W, Kane W H, Ortel T L. Fine mapping of inhibitory anti-factor V antibodies using factor V C2 domain mutants. Identification of two antigenic epitopes involved in phospholipid binding. Thromb Haemost. 2001; 85:1048-1054. 80. Kim S W, Quinn-Allen M A, Camp J T, Macedo-Ribeiro S 1 Fuentes-Prior P, Bode W, Kane W H. Identification of functionally important amino acid residues within the C2-domain of human factor V using alanine-scanning mutagenesis. Biochemistry 2000; 39:1951-1958. 81. Srivastava A, Quinn-Allen M A, Kim S W, Kane W H, Lentz B R. Soluble phosphatidylserine binds to a single identified site in the C2 domain of human factor Va. Biochemistry 2001; 40: 8246-8255. 82. Pusey M L, Mayer L D, Wei G J, Bloomfield V A, Nelsestuen G L. Biochemistry 1982; 21.: 5262-5269. 83. van de Waart P, Bruls H, Hemer H C, Lindhout T. Interaction of bovine blood clotting factor Va and its subunits with phospholipid vesicles. Biochemistry 1983; 22 (10): 2427-2432. 84. Tracy P B, Peterson J M, Nesheim M E, McDuffie F C, Mann K G. Interaction of coagulation factor V and factor Va with platelets. J Biol Chem. 1979; 254: 10354-10361. 85. Tracy P B, Eide L L, Bowie E J, Mann K G. Radioimmunoassay of factor V in human plasma and platelets. Blood 1982; 60: 59-63. 119  86. Magdeleyns E J P, Keuren J F W, Curvers J. Factor Va, bound to microparticles released during platelet storage, is resistant to inactivation by acitvated protein C. Transfusion 2007; 47:1880-1888. 87. Wilson D S, Salem H H, Mruk J S, Maruyama I, Majerus P W. Biosynthesis of coagulatoin factor-V by a human hepatocellular carcinoma cell line. J Cl/n.m vest. 2008; 73: 654-658. 88. Gewirtz A M, Keefer K, Doshi A E, Annamalai H C, Chiu R W, Colman R W. Biology of human megakaryocyte factor V. Blood 1986; 67: 1639-1648. 89. Gewirtz A M, Shapiro C, Shen Y M, Boyd R, Colman R W. Cellular and molecular regulation of factor V expression in human megakaryocytes. J Cell Physiol. 1992; 153: 277-287. 90. Giampaolo A, Vulcano F, Macioce G, Mattia A, Milazzo B L, Ciccarelli M C, Hassan H J. Factor V expression in platelets from human megakaryocytic culture. BritJ.Haematol. 2004; 128: 108-111. 91. Camire R M, Pollak E S, Kaushansky K, Tracey P B. Secretable human platelet-derived factor-V originates from plasma pool. Blood 2008; 92: 3035-3041. 92. Vicic W J, Lages B, Weiss H J. Release of human platelet factor V activity is induced by both collagen and ADP and is inhibited by aspirin. Blood 1980; 56: 448-455. 93. Osterud B, Rapaport S I, Lavine K K. Factor V activity of platelets: evidence for an activated factor V molecule and for a platelet activator. Blood 1977; 49: 819-834. 94. Breederveld K, Giddings J C, ten Cate J W, Bloom A L. The localization of factor V within normal human platelets and the demonstration of a platelet-factor V antigen in congenital factor V deficiency. Br] Haematol. 1975; 29: 405-412. 95. Kane W H, Mruk J 5, Majerus P W. Activation of coagulation factor V by a platelet protease. J Clln.Invest 1982; 70: 1092-1100. 96. Gould W R, Silveira J R, Tracy P B. Unique in vivo modifications of coagulation factor V produce a physically and functionally distinct platelet derived cofactor: characterization of purified platelet-derived factor V/Va. JB/olChem. 2004; 279: 2383-2393.  120  97. Viskup R W, Tracy P B, Mann K G. The isolation of human platelet factor V [published erratum appears in Blood 1987 Jul;70(1):339]. Blood 1987; 69: 1188-1195. 98. Monkovic D D, Tracy P B. Functional characterization of human plateletreleased factor V and its activation by factor Xa and thrombin. J Biol Chem. 1990; 265:17132-17140. 99. Monkovic D D, Tracy P B. Functional Characterization of human plateletreleased factor V and its activation by factor Xa and thrombin. 3 Biol Chem. 1990; 265: 17132-17140. 100. gould W R, Silveira J R, Tracy P B. Unique in vivo modifications of coagulation factor V produce a physically and functionally distinct plateletderived cofactor. 3 Biol Chem. 2004; 279: 2383-2393. 101. Grigg A P, Dauer R, Thurlow P J. Bleeding due to an acquired inhibitor of platelet associated factor V. Aust.N.ZJ Med 1989; 19: 310. 102. Nesheim M E, Nichols W L, Cole T L, Houston J G, Schenk R B, Mann K G, Bowie E J. Isolation and study of an acquired inhibitor of human coagulation factor V. 3 Clln Invest 1986; 77: 405-415. 103. Krishnaswamy S, Mann K G, Nesheim M E. The prothrombinase-catalyzed activation of prothrombin proceeds through the intermediate meizothrombin in an ordered, sequential reaction. 3 Biol Chem. 1986; 261: 8977-8984. 104. Tucker M M, Nesheim M E, Mann K G. Differentiation of enzyme and substrate binding in the prothrombinase complex. Biochemistry 1983; 22: 4540-4546. 105. Kung C, Hayes E, Mann K G. A membrane-mediated catalytic events in prothrombin activation. 3 Biol Chem. 1994; 269: 25838-25848. 106. Krishnaswamy S, Walker R K. Contribution of the prothrombin fragment 2 domain to the function of factor Va in the prothrombinase complex. Biochemistryl997; 36(11): 3319-3330. 107. Boskovic D S, Giles A R, Nesheim M E. Studies of the role of factor Va in the factor Xa-catalyzed activation of prothrombin, fragment 1.2prethrombin-2, and dansyl- L-glutamyl-glycyl-L-arginine-meizothrombin in the absence of phospholipid. 3 Biol Chem. 1990; 265:10497-10505.  121  108. Heldebrant C M, Butkowski R 3, Bajaj S P, Mann K G. The activation of prothrombin. II. Partial reactions, physical and chemical characterization of the intermediates of activation. J Biol Chem. 1973; 248: 7149-7163. 109. Mann K G, Yip R, Heldebrant C M, Fass D N. Multiple active forms of thrombin. 3. Polypeptide chain location of active site serine and carbohydrate. JBiolChem. 1973; 248:1868-1875. 110. Boskovic D S, Bajzar L 5, Nesheim M E. Channeling during prothrombin activation. J Biol Chem. 2001; 276: 28686-28693. 111. Suzuki K, Stenflo J, Dahlback B, Teodorsson B. Inactivation of human coagulation factor V by activated protien C. J B/of Chem. 1983; 258: 1914-1920. 112. Kalafatis M, Rand M D, Mann K G. The mechanism of inactivation of human factor V and human factor Va by activated protein C. J Biol.Chem. 1994; 269: 31869-31880. 113. Hockin M F, Cawthern K M, Kalafatis M, Mann K G. A model describing the inactivation of factor Va by activated protein C(APC):bond cleavage, fragment dissociation and product inhibition. Biochemistry 1999; 38: 6918-6934. 114. Mann K G, Hockin M F, Begin K J, Kalafatis M. Activated protein C cleavage of factor Va leads to dissociation of the A2 domain. J.BioI.Chem. 1997; 272: 20678-20683. 115. Laue T M, Johnson A E, Esmon C T, Yphantis D A. 23:1339-1348.  Biochemistry 1984;  116. Nesheim M E, Myrmel K H, Hibbard L, Mann K G. Isolation and characterization of single chain bovine factor V. J B/of Chem. 1979; 254: 508-517. 117. Laue T M, Johnson A E, Esmon C T, Yphantis D A. Structure of bovine blood coagulation factor Va. Determination of the subunit associations, molecular weights, and asymmetries by analytical ultracentrifugation. Biochemistry 1984; 23: 1339-1348. 118. Mosesson M W, Nesheim M E, DiOrio 3, Hainfeld 3 F, Wall 3 S, Mann K G. Studies on the structure of bovine factor V by scanning transmission electron microscopy. Blood 1985; 65: 1158-1162. 119. Dahlback B. Ultrastructure of human coagulation factor V. J.BioI.Chem. 1985; 260: 1347-1349.  122  120. Dahlback B. Bovine coagulation factor V visualized with electron microscopy: ultrastructure of the isolated activated forms and of the activation fragments. J.Bio/.Chem. 1986; 261: 9495-9501. 121. Villoutreix B 0, Dahlback B. Structural investigation of the A domains of human blood coagulation factor V by molecular modeling. Protein science 1998; 7:1317-1325. 122. Pellequer J-L, Gale A J, Getzoff E D, Griffin J H. Three-dimensional model of coagulation factor Va bound to activated protein C. Thrombosis and Haemostasis 2000; 84: 849-857. 123. Villoutreix B 0, Bucker P, Hofmann K, Baumgartner S, Dahlbach B. Molecular models for the two discoidin domains of human blood coagulation factor V. J Mol.Model. 1998; 4: 268-275. 124. Macedo-Ribeiro S, Bode W, Huber R, Quinn-Allen M A, Kim S W, Ortel T L, Bourenkov G P, Bartunik H D, Stubbs M T, Kane W H, Fuentes-Prior P. Crystal structures of the membrane-binding C2 domain of human coagulation factor V. Nature 1999; 402: 434-439. 125. Adams T E, Hockin M F, Mann K G, Everse S J. The crystal structure of activated protein C-inactivated bovine factor Va: Implications for cofactor function. PNAS USA 2004; 101: 8918-8923. 126. Lee C J, Lin P, Chandrasekaran V, Duke R E, Everse S 3, Perera L, Pedersen L G. Proposed structure models of human factor Va and prothrombinase. 1 Thromb.Haemost. 2007; 6: 83-89. 127. Zaitseva I, Zaitsev V, Card G, Moshkov K, Bax B, Ralph A, Lindley P. The x-ray structure of human serum ceruloplasmin at 3.1 A: Nature of the copper centres. Journal of Biological Inorganic Chemistry 1996; 1:15-23. 128. Goldberg E D. The sea. J. Wiley: New York, 1963.  129. Friedman M. Advances in experimental medicine and biology. Plenum Press: New York, 1973. 130. Siegel F L, Carafoli E, Kretsinger R H, Maclennan D H, Wasserman R H. Calcium-binding proteins. Elsevier/north-Holland: New York, 1980.  131. Kawasaki H, Krestsinger R H. Calcium-binding proteins 1: EF-hands. Academic Press: Liverpool, 1994. 132. Owen C.A. Physiological aspects of copper. Noyes Publications: New Jersey, 1982.  123  133. Owen C A, Jr. Biochemical aspects of copper Noyes Publications: New Jersey, 1982. 134. Hibbard L S, Mann K G. The calcium-binding properties of bovine factor V. J.BiolChem. 1980; 255: 638-645. 135. Guinto E R, Esmon C T. Formation of a calcium-binding site on bovine activated factor V following recombination of the isolated subunits. J.BiolChem. 1982; 257: 10038-10043. 136. Mann K G, Lawler C M, Vehar G A, Church W R. Coagulation Factor V contains copper ion. iBiol.Chem. 1984; 259:12949-12951. 137. Esmon C T. The subunit structure of thrombin-activated factor V. Isolation of activated factor V, separation of subunits, and reconstitution of biological activity. J.Bio/.Chem. 1979; 254: 964-973. 138. Laue T M, Lu R, Krieg U C, Esmon C T, Johnson A E. Ca2+-dependent structural changes in bovine blood coagulation factor Va and its subunits. Biochemistry 1989; 28: 4762-4771. 139. Zeibdawi A R, Pryzdial E L. Mechanism of factor Va inactivation by plasmin. Loss of A2 and A3 domains from a Ca2+-dependent complex of fragments bound to phospholipid. JBioI Chem. 2001; 276: 19929-19936. 140. Zeibdawi A R, Grundy J E, Lasia B, Pryzdial E L G. Coagulation factor Va G1u96-Asplll:A chelator-sensitive site involved in function and subunit association. Biochemistry 2004; 377:141-148. 141. Sorensen K W, Nicolaes G A, Villoutreix B 0, Yamazaki T, Tans G, Rosing J, Dahlback B. Functional properties of recombinant factor V mutated in a potential calcium-binding site. Biochemistry 2004; 43: 5803-5810. 142. Wakabayashi H, Koszelak M E, Mastri M, Fay P J. Metal ion-independent association of factor VIII subunits and the roles of calcium and copper ions for cofactor activity and inter-subunit affinity. Biochemistry 2001; 40: 10293-10300. 143. Wakabayashi H, Freas J, Zhou Q, Fay P J. Residues 110-126 in the Al domain of factor VIII contain a Ca 2 binding site required for cofactor activity. J. Biol. Chem. 2004; 279:12677-12684. 144. Shen B W, Spiegel P C, Chang C H, Huh J W, Lee J 5, Kim J, Kim Y H, Stoddard B L. The tertiary structure and domain organization of coagulation factor VIII. Blood 2008; 11.1.: 1240-1247.  124  145. Ngo J C, Huang M, Roth D A, Furie B C, Furie B. Crystal structure of human factor VIII: implications for the formation of the factor IXa-factor Villa complex. Structure. 2008; 16: 597-606. 146. Lak M, Sharifian R, Peyvandi F, Mannucci P M. Symptoms of inherited factor V deficiency in 35 Iranian patients. Br.J Haematol. 1998; 103: 1067-1069. 147. Peyvandi F, Duga 5, Akhavan S, Mannucci P M. Rare coagulation deficiencies. Haemophilla 2002; 8: 308-321. 148. Guasch J F, Cannegieter S, Reitsma P H, Veer-Korthof E T, Bertina R M. Severe coagulation factor V deficiency caused by a 4 bp deletion in the factor V gene. Br.J Haematol. 1998; 101: 32-39. 149. Murray J M, Rand M D, Egan J 0, Murphy 5, Kim H C, Mann K G. Factor VNew Brunswick: A1a221-to-Val substitution results in reduced cofactor activity. Blood 1995; 86: 1820-1827. 150. Lunghi B, Castoldi E, Mingozzi F, Bernardi F, Castaman G. A novel factor V null mutation detected in a thrombophilic patient with pseudo homozygous APC resistance and in an asymptomatic unrelated subject. Blood 1998; 92: 1463-1464. 151. Cui J, O’Shea K S, Purkayastha A A, Saunders T.L., Ginsberg D. Fatal haemdrrhage and incomplete block to embryogenesis in mice lacking coagulation factor V. Nature 1996; 384: 66-68. 152. Miletich J P, Prescott S M, White R, Majerus P W, Bovill E G. Inherited predisposition to thrombosis. Cell 1993; 72: 477-480. 153. Ehrenforth 5, Klarmann D, Zabel B, Scharrer I, Kreuz W. Severe factor V deficiency presenting as subdural haematoma in the newborn. Eur.J Pediatr. 1998; 157: 1032. 154. Totan M, Albayrak D. Intracranial haemorrhage due to factor V deficiency. Acta Paediatr. 1999; 88: 342-343. 155. Salooja N, Martin P, Khair K, Liesner R, Hann I. Severe factor V deficiency and neonatal intracranial haemorrhage: a case report. Haemophilla 2000; 6: 44-46. 156. Lane D A, Mannucci P M, Bauer K A, Bertina R M, Bochkov N P, Boulyjenkov V 1 Chandy M, Dahlback B, Ginter E K, Miletich J P, Rosendaal F R, Seligsohn U. Inherited thrombophilia: Part 2. Thromb Haemost. 1996; 76: 824-834.  125  157. Zoller B, Garcia d F, Hillarp A, Dahlback B. Thrombophilia as a multigenic disease. Haematologica 1999; 84: 59-70. 158. Vos H L. An online database of mutations and polymorphisms in and around the coagulation factor V gene. J Thromb Haemost. 2007; 5:185188. 159. Castoldi E, Lunghi B, Mingozzi F, Muleo G, Redaelli R, Mariani G, Bernardi F. A missense mutation (Y1702C) in the coagulation factor V gene is a frequent cause of factor V deficiency in the Italian population. Haematologica 2001; 86: 629-633. 160. Simioni P, Scudeller A, Girolami A. Factor V Leiden and thrombophilia [letter; comment]. N.Eng/J Med 1995; 332: 1382. 161. Simioni P, Kalafatis M, Tormene D. Abnormal propeptide processing resultling in the presence of two abnormal species of protine C in plasma. Thromb./-/aemost 2001; 86: 1017-1022. 162. Branson H E, Kats J, Marble R, Griffin J H. Inherited protein C deficiency and coumarin-responsive chronic relapsing purpura fulminans in a newborn infant. Lancet 1983; 19:1165-1168. 163. Estelles A, Garcia Plaza I, Dasi A. Severe inherited “homozygous” protein C deficiency in a newborn infant. Thromb Haemost. 1984; 52: 5356. 164. Seligsohn U, Berger A, Abend M. Homozygous protein C deficiency manifested by massive venous thrombosis in the newborn. N.EngI.J Med 1984; 310: 559-562. 165. Dahlback B, Carisson M, Svensson P J. Proc.NatI.Acad.Sci.USA 1993; 90: 1004-1008. 166. Odegaard B, Mann K G. Proteolysis of factor Va by factor Xa and activated protein C. J.BioI.Chem. 1987; 262: 11233-11238. 167. Kalafatis M, Mann K G. Role of the membrane in the inactivation of factor Va by activated protein C. J Biol Chem. 1993; 268: 27246-27257. 168. Bertina R M, Koeleman B P, Koster T, Rosendaal F R, Dirven R J, de Ronde H, van der Velden P A, Reitsma P H. Mutation in blood coagulation factor V associated with resistance to activated protein C [see comments]. Nature 1994; 369: 64-67. 169. Svensson P J, Dahlback B. Resistance to activated protein C as a basis for venous thrombosis [see comments]. N.EngI.J Med 1994; 330: 5 17-522.  126  170. Page C, Rubin L E, Gusberg R J. Arterial thrombosis associated with heterozygous factro V Leiden disorder, hyperhomocysteinimia and peripheral arterial disease:importance of synergistic factors. J Vasc.Surg. 2005; 42: 1014-1018. 171. de Paula Sabino A, Rlbeiro D D, Carvalho M G. Factor V Leiden and increased risk for arterial thrombotic disease in youing Brazilian patients. B/ood2006; 17: 271-275. 172. Rosendaal F R, Siscovick D S, Schwartz S M. Factor V Leiden increases the risk of myocardical infarction in young women. Blood 1997; 89: 28172821. 173. Williamson D, Brown K, Luddington R. Ractor V Cambridge:a new mutation(Arg306->Thr) associated with resistance to activated protein C. B1ood1998; 91: 1140-1144. 174. Chan W P, Lee C K, Kwinbg Y L. A novel mutation of Arg306 of factor V gene in Hong Kong Chinese. Blood 1998; 91:1135-1139. 175. Van der Neut Kolfschoten M, Dirven R J, Vos H L, Tans G, Rosing J, Bertina R M. Factor Va is inactivated by activated protien C in the absence of cleavage sites at Arg306. Arg506 and Arg679. J Biol Chem. 2004; 279: 6567-6575. 176. Gemmati D, Serino M L, Moratelli S, Man R, Ballerini G, Scapoli G L. Coexistence of antithrombin deflciency,factor V Leiden and hyperhomocysti nem ia in a th rom botic family. Blood Coagulat. F/br/no!. 1998; 9:173-176. 177. Kluijtmans L A, Boers G H, Verbruggen B, Trijbels F J, Novakova I R, Blomm H J. Homozygous cystathionine beta-synthase deficiency, combined with factor V Leiden or thermolabile methylenetetrahydrofolate reductase in the risk of venous thrombosis. Blood 1998; 91: 2015-2018. 178. Toso R, Camire R M. Removal of B-domain sequences from factor V rather than specific proteolysis underlies the mechanism by which cofactor function is realized. JB1oI Chem. 2004; 279: 21643-21650. 179. Nelsestuen G L, Kisiel W, Di Scipicio R G. Interaction of vitamin K dependent proteins with membranes. Biochemistry 1978; 17: 2134-2138. 180. Longhurst C M, White M M, Wilkinson D A, Jennings L K. A CD9, alphallbbeta3, integrin-associated protein, and GPIb/V/IX complex on the surface of human platelets is influenced by alphallbbeta3 conformational states. Eur.JBiochem. 1999; 263:104-111.  127  181. Tanner S D, Baranov V I, Bandura D R. Reaction cells and collision cells for ICP-MS. Spectrochimica Acta Part B 2002; 57:1361-1452. 182. Zhu H, Toso R, Camire R M. Inhibitory sequences within the B-domai stabilize circulating factor V in an inactive state. J.BioI.Chem. 2007; 282: 15033-15039. 183. Walker F J. Characterization of the interaction between the heavy and light chains of bovine factor Va. J.Biol.Chem. 1992; 267: 19896-19900. 184. Klinge J, Ananyeva N M, Hauser C A, Saenko E L. Hemophilia A-from basic science to clinical practice. Semin. Thromb.Hemost 2002; 28: 309-322. 185. Maugard C, Tuffery 5, Aguilar-Martinez P, Schved J F, Gris J C, Demaille J, Claustres M. Protein truncation test: Detection of severe haemophilia A mutation and analysis of factor VIII transcripts. Hum.Mutat. 1998; 11.: 18-22. 186. Repesse Y, Slaoui M, Ferrandiz D, Gautier P, Costa C, Costa J M, Lavergne J M, Borel-Derlon A. Factor VIII (FVIII) gene mutations in 120 patients with hemophilia A: detection of 26 novel mutations and correlation with FVIII inhibitor development. J. Thromb.Haemost. 2007; 5:1476. 187. Theophilus B D M, Enayat M 5, Williams M D, HIll G H. Site and type of mutations in the factor VIII gene in patients and carriers of Hemolphilia A. HaemophIlia 2001; 7: 381-391. 188. Liu M L, Murphy M E P, Thompson A R. A domain mutations in 65 haemophilia A families and molecular modelling of dysfunctional factor VIII proteins. Brit.J.Haematol. 1998; 103:1051-1060. 189. Pipe S W, Saenko E L, Eickhorst A N, Kemball-Cook G, Kaufman R J. Hemophilia A mutations associated with 1-stage/2-stage activity discrepancy disrupt protein-protein interactions within the triplicated A domains of thrombin-activated factor Villa. Blood 2001; 97: 685-691. 190. Hakeos W H, Pipe S W, Saenko E L, Kaufman R J. Hemophilia A mutations within the factor VIII A2-A3 subunit interface destabilize Villa and cause one-stage/two-stage activity discrepancy. Blood 2000; 96: 2288. 191. Fay P J, Haidaris P J, Smudzin T M. Human Factor-Villa Subunit Structure Reconstitution of Factor Villa from the Isolated A1/A3-C1-C2 Dimer and A2 Subunit. JBiolChem. 1991; 266: 8957-8962. -  192. Yamakage N, Ikejiri M, Okumura K, Takagi A, Murate T, Matushita T, Naoe T, Yamamoto K, Takamatsu J, Yamazaki T, Hamaguchi M, Kojima T.  128  A case of coagulation factor V deficiency caused by compound heterozygous mutations in the factor V gene. Haemophiia 2006; 12: 172178. 193. van Wijk R, Nieuwenhuis K, van den B M, Huizinga E G, van der Meijden B B, Kraaijenhagen R J, van Solinge W W. Five novel mutations in the gene for human blood coagulation factor V associated with type I factor V deficiency. Blood 2001; 98: 358-367. 194. Montefusco M C, Duga 5, Asselta R, Malcovati M, Peyvandi F, Santagostino E, Mannucci P M, Tenchini M L. Clinical and molecular characterization of 6 patients affected by severe deficiency of coagulation factor V: Broadening of the mutational spectrum of factor V gene and in vitro analysis of the newly identified missense mutations. Blood 2003; 102: 3210-3216. 195. Castoldi E, Simioni P, Kalafatis M, Lunghi B, Tormene D, Girelli D, Girolami A, Bernardi F. Combinations of 4 mutations (FV R506Q, RI H1299R, FV Y1702C, PT 20210G/A) affecting the prothrombinase complex in a thrombophilic family. Blood2000; 96:1443-1448. 196. Yang T L, Cui J, Taylor J M, Yang A, Gruber S B, Ginsburg D. Rescue of fatal neonatal hemorrhage in factor V deficient mice by low level transgene expression. Thromb Haemost. 2000; 83: 70-77.  129  

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.24.1-0067024/manifest

Comment

Related Items