UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Platelet factor XIII contributes to the hemostatic process : factor XIII interactions at the platelet… Serrano, Katherine 2000

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

Item Metadata

Download

Media
831-ubc_2000-566196.pdf [ 18.69MB ]
Metadata
JSON: 831-1.0089876.json
JSON-LD: 831-1.0089876-ld.json
RDF/XML (Pretty): 831-1.0089876-rdf.xml
RDF/JSON: 831-1.0089876-rdf.json
Turtle: 831-1.0089876-turtle.txt
N-Triples: 831-1.0089876-rdf-ntriples.txt
Original Record: 831-1.0089876-source.json
Full Text
831-1.0089876-fulltext.txt
Citation
831-1.0089876.ris

Full Text

PLATELET FACTOR XIII CONTRIBUTES TO THE HEMOSTATIC PROCESS: FACTOR XIII INTERACTIONS AT THE PLATELET SURFACE A N D WITH THE INTERNAL PLATELET SKELETON by KATHERINE SERRANO B.Sc, The University of British Columbia, 1990 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Biochemistry and Molecular Biology) We accept this thesis as conforming to the required standard THE U N I V E R ^ T T Y ^ F T ^ T l l M t Q O L U M B I A July 2000 © Katherine Serrano, 2000 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of / 6 T R V / The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT During blood clot formation, the transglutaminase, FXIII , functions to chemically crosslink fibrin and a,2-antiplasmin to fibrin polymers. This strengthens and stabilizes the clot. Activated platelets accelerate these crosslinking reactions via the presentation of surface-bound active FXIII . The platelet cytoplasm contains approximately half of the FXIII content circulating in blood, yet the function of cytoplasmic FXIII is poorly understood at present. This thesis investigated the function of platelet FXIII in processes both outside and inside the platelet by studying the interactions of FXIII with fibrin(ogen) and with platelet cytoskeletal proteins. Surface associations determined using specific antibodies and flow cytometry indicate that FXIIIa association with platelets, occurred at early stages of activation, was partially reversible, correlated significantly with fibrinogen expression, and was enhanced after alkylation of the active-site cysteine. Immunoblotting experiments indicate that surface-associated fibrin(ogen) was crosslinked but not directly to the platelet surface. In aggregometry experiments, FXIIIa slightly enhanced the rate although not the final amount of aggregation. FXIII was present in the cytoskeletal fractions of the platelet lysate separated with centrifugation. When these fractions were immobilized on nitrocellulose membranes, activated but not resting FXIII bound to cytoskeletal proteins. With fluorescence microscopy, an actin polymerization-dependent, transient translocation of FXIII from a diffuse homogeneous distribution throughout the cytoplasm to the platelet periphery was observed upon platelet activation suggesting an association with cytoskeletal proteins. There i i was also a measurable increase in the transglutaminase activity of the cytoskeletal fractions of activated platelets. Immunoblotting analysis of such cytoskeletal fractions identified that the focal adhesion proteins, filamin and vinculin were crosslinked upon platelet activation. In summary, this study has demonstrated several new factors that affect the interactions of FXIII on the platelet surface, which could affect the promotion of fibrin polymer crosslinking by platelets. In addition to its role on the outside of platelets in stabilizing the fibrin clot, a role for FXIII on the inside of platelets in stabilizing focal adhesions can now be proposed. It appears that FXIII may enable clot retraction by strengthening structural elements required for such a process on both sides of the platelet plasma membrane. i i i T A B L E O F C O N T E N T S A B S T R A C T i i LIST OF T A B L E S v i i i LIST OF F I G U R E S ix LIST OF A B B R E V I A T I O N S x i i G L O S S A R Y xiv A C K N O W L E D G M E N T S xv Chapter 1 I N T R O D U C T I O N 1 1.1 Factor XIII 1 1.1.1 Factor XIII - The Early Days 1 1.1.2 Factor XIII in Coagulation and Hemostasis 2 1.1.3 Transglutaminase Activity 4 1.1.4 Factor X H I a Substrates 5 1.1.5 FXIII Structure 9 1.1.6 FXIII Activation 11 1.2 Platelets 15 1.2.1 Platelet Function 15 1.2.2 Platelet Structure 16 1.2.3 Platelet Activation 19 1.2.4 Platelet Adhesive Proteins and Their Receptors 20 1.2.5 Platelets and FXIII 22 1.2.6 Clot Retraction 24 iv 1.2.7 Platelet Cytoskeleton 25 1.3 Rationale and Objectives (. 29 Chapter 2 M A T E R I A L S A N D M E T H O D S 32 2.1 Materials 32 2.1.1 Proteins, peptides, and other chemicals 32 2.1.2 Antibodies 32 2.2 Platelet Isolation 33 2.3 Platelet Incubations for Flow Cytometric Binding Determinations 35 2.4 Pre-activation of rFXIII with Thrombin 36 2.5 FXIII Pre-incubation with Iodoacetamide 36 2.6 Platelet Flow Cytometry 37 2.7 Platelet Incubations for S D S - P A G E Analysis 38 2.8 Platelet Lysate Preparation 38 2.9 Platelet Cytoskeletal Fractionation 39 2.10 Separation of Platelet Proteins using S D S - P A G E 39 2.11 Gel Silver Staining 40 2.12 Western Blot Analysis of Platelet Proteins 41 2.13 Platelet Aggregometry 42 2.14 FXIII Overlay Experiments 43 2.15 Fluorescence Microscopy 44 2.16 FXIII Activity Assay on Platelet Lysate and Cytoskeletal Fractions 45 2.17 Lactate Dehydrogenase (LDH) Assay 46 2.18 Statistical Analysis 47 v Chapter 3 FXIII O N T H E O U T S I D E O F H U M A N P L A T E L E T S 49 3.1 Exogenous FXIIIa Binding to Isolated, Thrombin-Activated Platelets 49 3.2 Effect of Iodoacetamide on rFXIIIa Binding to Activated Platelets 54 3.3 Effect of rFXIIIa on Fibrin(ogen) Binding to Platelets 62 3.4 Effect of rFXIIIa on Aggregation 66 3.5 Endogenous FXIII Expression on Thrombin-Activated Platelets 73 3.6 Discussion 76 Chapter 4 FXIII A N D T H E P L A T E L E T C Y T O S K E L E T O N 86 4.1 FXIII in the Cytoskeletal Fractions of Platelets 86 4.2 rFXIII , rFXIIIa and r F X I I I ' C a 2 + Overlay Binding to Cytoskeletal Platelet Fractions 92 4.3 Intracellular Location of FXIII in Activated Platelets 97 4.4 Effect of Cytochalasin D on FXIII Translocation of Thrombin-Activated Platelets 104 4.5 FXIII Activity in Platelet Lysate and Cytoskeletal Fractions 107 4.6 Discussion 110 Chapter 5 P R O T E I N S C R O S S L I N K E D U P O N P L A T E L E T A C T I V A T I O N 119 5.1 The Crosslinking of Platelet Fibrin(ogen) 119 5.2 Cytoskeletal Proteins Crosslinked in Thrombin-Activated Platelets 125 5.3 Intracellular Location of Vincul in in Thrombin-Activated Platelets 131 5.4 Effect of Aggregation on Vincul in Association with the Platelet Cytoskeleton 133 5.5 Discussion 139 Chapter 6 S U M M A R Y 149 6.1 Summarizing Discussion 149 6.2 Future Directions 155 v i R E F E R E N C E L I S T LIST O F T A B L E S Table 1. FXIIIa substrates 6 Table 2. Adhesive protein receptors on platelets : 22 Table 3. Correlation between the number of small platelet aggregates and FXIII detection on single platelets 68 Table 4. The effect of rFXIIIa addition on platelet aggregation 72 Table 5. FXIII activity in resting and thrombin-activated platelet cytoskeleton 109 v i i i LIST O F FIGURES Figure 1. Coagulation cascade 3 Figure 2. Modified double displacement reaction catalyzed by FXIIIa 5 Figure 3. Graphical representation of the FXIII A-chain dimer structure 12 Figure 4. Mechanism of activation of Factor XIII 14 Figure 5. Platelet morphology 17 Figure 6. Model of the resting platelet cytoskeleton 26 Figure 7. Model of a Focal Adhesion Complex 29 Figure 8. Effect of G P R P on platelet aggregate formation and fibrinogen binding 50 Figure 9. rFXIIIa binding to thrombin-activated and non-activated platelets 51 Figure 10. The effect of E D T A on the binding of activated FXIII to activated platelets 53 Figure 11. rFXIII conversion in the platelet incubation mixtures 55 Figure 12. Effect of iodoacetamide on binding of rFXIIIa to thrombin-activated platelets. ..56 Figure 13. Iodoacetamide dose-response on rFXIIIa binding to platelets 58 Figure 14. Effect of iodoacetamide pre-incubation with FXIII on platelet FXIII binding 59 Figure 15. Effect of iodoacetamide on platelet activation as measured by the expression of platelet activation markers, CD62P and fibrinogen (Fbgn) 61 Figure 16. The effect of FXIIIa on the ability of E D T A to dissociate fibrin(ogen) from the platelet surface 63 Figure 17. The effect of rFXIIIa on the form of platelet-bound fibrin(ogen) 65 ix Figure 18. Effect of rFXIIIa on the number of platelet aggregates 67 Figure 19. Correlation between the expression of fibrinogen and FXIII on platelets 69 Figure 20. Fibrinogen and FXIII in the low speed pellet of thrombin-stimulated platelets... 70 Figure 21. Expression of endogenous FXIII on thrombin-activated platelets 74 Figure 22. FXIII in the low speed pellet and high speed pellet of platelets at 10 and 30 minute incubations 87 Figure 23. FXIII in the high speed pellet and low speed pellet of thrombin-activated platelets at 2 3/ 4 hours 89 Figure 24. Protein staining pattern in the high speed pellet of differentially incubated platelets 91 Figure 25. rFXIIIa overlay binding to platelet cytoskeletal proteins 93 Figure 26. FXI I I*Ca 2 + overlay binding to platelet membrane skeletal proteins 96 Figure 27. FXIII localization in thrombin-activated platelets at different activation time-points 98 Figure 28. FXIII localization in calcium ionophore-activated platelets at different activation time-points 101 Figure 29. FXIII localization in collagen-activated platelets at different activation time-points 103 Figure 30. Effect of cytochalasin D on FXIII localization in thrombin-activated platelets. 105 Figure 31. CD62 expression of platelets treated with cytochalasin D 106 Figure 32. FXIII activity in resting and thrombin-activated platelet lysate 108 Figure 33. The effect of thrombin-activation on the form of platelet-associated fibrin(ogen). 120 Figure 34. Crosslinking time-course for fibrin(ogen) on thrombin-activated platelets both in the presence and the absence of exogenous FXIII 123 Figure 35. Fibrinogen crosslinking in calcium ionophore activated platelets 124 Figure 36. CD62P expression of iodoacetamide-treated thrombin-activated platelets 127 Figure 37. Western blot analysis of filamin content in the cytoskeleton of platelets activated with thrombin 128 Figure 38. Western blot analysis of vinculin content in the cytoskeleton of platelets activated with thrombin 130 Figure 39. Vincul in localization in thrombin-activated platelets at different activation time-points 132 Figure 40. Association of a vinculin cleavage fragment with the cytoskeleton upon platelet aggregation 134 Figure 41. Association of vinculin with the platelet low speed pellet 137 Figure 42. Vincul in cleavage in whole platelet lysate 138 xi LIST O F A B B R E V I A T I O N S A C D A c i d citrate dextrose A D P Adenosine diphosphate B C A Bicinchoninic acid B S A Bovine serum albumin D M S O Dimethyl sulfoxide E D T A Ethylenediamine tetraacetic A c i d E G T A Ethylene glycol - bis (B-aminoethyl ether) - N , N , N ' , N ' - tetraacetic A c i d Fbgn Fibrinogen FITC Fluorescein isothiocyanate FXIII Factor XIII FXIIIa Activated FXIII FXIIIa* FXIII activated through thrombin cleavage FXIIIa 0 FXIII activated non-proteolytically GPIIb/IIIa Glycoprotein Ilb/IIIa G P R P Glycyl-prolyl-arginyl-proline H R P Horse radish peroxidase HSP High speed pellet I A A Iodoacetamide L D H Lactate dehydrogenase L S P L o w speed pellet M r Relative mobility M H T B Modified Hepes Tyrode's buffer M T B S T M i l k Tris-buffered saline containing Tween 20 N A D + Nicotinamide adenine dinucleotide (oxidized form) N A D H Nicotinamide adenine dinucleotide (reduced form) P B S Phosphate buffered saline P E Phycoerythrin . • P F A Paraformaldehyde P M S F Phenylmethylsulfonyl fluoride P P A C K Phenyl-prolyl-arginyl chloromethylketone PRP Platelet-rich plasma rFXIII Recombinant factor XIII rFXIIIa Thrombin-cleaved recombinant factor XIII rFXIII«Ca 2 + Recombinant factor XIII pre-incubated with calcium ions S D S - P A G E Sodium dodecyl sulfate-polyacrylamide gel electrophoresis T C A Trichloroacetic acid Xll l G L O S S A R Y Apyrase - A T P / A D P scavenger which assists in preventing platelets from becoming refractory to activation by A D P Benzamidine - Inhibitor of trypsin and trypsin-like enzymes Cytochalasin D - Inhibitor of actin polymerization that does not inhibit glucose transport D T T - Reducing agent used to ensure a free active-site sulfhydryl group for FXIII activity E D T A - Strong calcium ion chelator E G T A - M i l d calcium ion chelator G P R P - Fibrin polymerization inhibitor Iodoacetamide - Non-specific FXIII inhibitor which functions by alkylating the active-site cysteine Leupeptin - Protease inhibitor P M S F - Serine protease inhibitor P P A C K - Specific active-site thrombin inhibitor xiv A C K N O W L E D G M E N T S This project was greatly impacted upon by the kind support of many people. First and foremost I must thank my supervisor Dana for her generosity in all things, for the constant direction which she provided, and for her faith in me. M y respect for her as both a scientist and a person is great and I have learned much in terms of science and life both while under her supervision. I am also indebted to my thesis committee members Dr. Cedric Carter, Dr. Ross MacGil l ivray, and Dr. Rob McMaster for their helpful suggestions during my period at the bench as well as on the writing of the manuscript. Dr. Carter also helped to educate me on one of the more specialized uses for yeast. The recombinant FXIII and FXIII antibodies used throughout the project were kindly provided by Dr. Paul Bishop at Zymogenetics in Seattle and Dr. Rob West was helpful with information on FXIII inhibitors. The groundwork for this project was laid by Dr. Al l ison Cox, a former post-doctoral fellow of the lab. The friendship of the people in the lab both past and present w i l l always be very dear to me and I 'd like to thank them all for their contributions to the project as well as for their zest for life. In particular I 'd like to thank Elena, the other half of the FXIII team, for her help especially with flow cytometry, Maria for the helpful discussions we shared, Elisabeth for teaching me much about microscopy, Kat i for being a star at troubleshooting blotting experiments, V i c k y for helping me brush up on my Spanish and for her unwavering support, Nadine for the GPIb antibodies, N ikky for help with aggregometry experiments, and very specially Derek, my partner in the MonsterPrep, who tread a similar path to mine during his doctoral studies and who was always there for me. xv To my parents and to Dana xvi C H A P T E R 1 I N T R O D U C T I O N LI Factor XIII 1.1.1 Factor X I I I - The Ea r ly Days In the first half of this century it was observed that fibrin clots that were formed in the presence of calcium ions were insoluble in weak bases (Barkan and Gaspar 1923). It was also discovered that fibrin prepared from plasma was more resistant to fibrinolysis than fibrin prepared from the purified factors fibrinogen and thrombin (Tillett and Garner 1933). These findings would relate to the eventual discovery of a non-dialyzable, thermolabile serum factor that was necessary for the calcium-dependent conversion of fibrin to a form that was insoluble either in mild acid solution or urea (Robbins 1944; Lak i and Lorand 1948; Lorand 1948; Lorand 1950). It was purified a number of years later and many of its properties were examined (Loewy et al. 1957; Loewy et al. 1961a; Loewy et al. 1961b; Loewy a/. 1961c; Loewy et al. 196 Id). The first indication of how important this serum factor was in maintaining normal hemostasis came from the report of a young boy, deficient in this factor, who suffered both from umbilical bleeding after birth and from episodes of recurrent bleeding after minor traumas (Duckert et al. 1960). Three years later this factor was added to the list of clotting factors and so it was, that from being called serum factor, Laki-Lorand factor, L - L factor, fibrin-stabilizing factor (FSF), fibrinase, and fibrinoligase it would become known as factor XIII (FXIII), the terminal acting enzyme in the coagulation cascade. 1 1.1.2 Factor X I I I in Coagulat ion and Hemostasis When a blood vessel is damaged through injury, a rapid biological response is initiated to restore the vessel's integrity so as to limit the loss of blood. Small anuclear cells in the blood, 2 - 4 pm in diameter, called platelets adhere to the site of injury and coagulation is initiated. The coagulation cascade is a protein cascade whose function is the creation of a fibrin clot (Figure 1). Together with platelets, the fibrin clot acts as a physical barrier to seal shut openings in the damaged blood vessel. The central enzyme of the coagulation cascade, thrombin, cleaves small peptides (fibrinopeptides A and B) off the N-termini of the a- and p1-chains of fibrinogen monomers turning these monomers into fibrin (Bailey et al. 1951). The fibrin monomers are capable of self-polymerization, which they do in a half-staggered manner creating fibrin strands. Thrombin also cleaves FXIII at Arg37 - G l y 38, creating the activation peptide, and converting this enzyme into its active form, FXIIIa. In the final steps of the cascade, FXIIIa enzymatically crosslinks the fibrin strands together. The crosslinking of fibrin by FXIIIa gives increased strength to the fibrin clot and also greater resistance from fibrinolysis (Mockros et al. 1974; Lorand and Jacobsen 1962; McDonagh et al. 1971). Another important mechanism through which FXIIIa protects the fibrin clot from fibrinolysis is through the crosslinking of the most potent inhibitor of plasmin, a2-antiplasmin, directly to the fibrin polymers (Sakata and A o k i 1980). Due to the role of FXIIIa in the formation of these stabilizing interactions in the fibrin clot, a deficiency in FXIII can lead to bleeding problems and patients in this position can benefit from FXIII replacement therapy (Ottaviani and Mandelli 1966; Miloszewski and Losowsky 1975; Ikkala 1972). 2 Intrinsic Pathway Extrinsic Pathway "Tissue damage " Crosslinked Fibrin Figure 1. Coagulation cascade. In a wounded blood vessel, coagulation is initiated by the expression of tissue factor and thereafter continues via the intrinsic pathway. The generation of thrombin is key for the cleavage of fibrinogen to form fibrin and also for the activation of factor XIII . Fibrin monomers self-associate into polymers and factor XHIa crosslinks the polymers in a calcium-dependent process. Figure adapted from Hoffman et al. 2000, 1785. 3 1.1.3 Trans glutaminase Activity Factor XIII is a member of the enzyme family of transglutaminases (EC2.3.2.13) which when activated, catalyse the formation of e-(y-glutamyl)lysine crosslinks (Pisano et al. 1968) in a calcium-dependent manner (Lorand and Konishi 1964). Transglutaminase linkages have been proposed to occur universally. Besides being found in human fibrin clots, transglutaminase activity has been demonstrated in rhesus monkeys (Pisano et al. 1972), merino sheep wool (Asquith et al. 1970), guinea pig hair, porcupine quills (Harding and Rogers 1971), the vaginal plug created by postejaculatory clotting of rodent seminal plasma (Williams-Ashman et al. 1972), spiny lobsters (Fuller and Doolittle 1971), salmon (Murtaugh et al. 1973), birds (Cierniewski et al. 1975), Escherichia coli, the slime mold, Physarum polycephalum and the ciliate Paramecium aurelia (Loewy and Matacic 1979). Transglutaminases create this covalent crosslink through a modified double displacement reaction centered around a cysteine residue (Figure 2). The crosslink is formed between the side chains of a glutamine residue of one protein molecule and the lysine residue of another protein molecule. However, the amino functional group is not required to be attached to a peptide, as is the glutamine residue; any nucleophilic primary amine is reactive with the acyl enzyme intermediate (Folk and Finlayson 1977). Therefore, the reactive glutamine residue-containing protein dictates the specificity of this class of enzymes more than the lysine residue-containing protein (Greenberg et al. 1991). If there are no primary amines available, the acyl enzyme intermediate is hydrolyzed by a water molecule and the glutamyl group is released as a glutamate residue. Reactions that block the active-site of the transglutaminase cysteine residue, such as acetylation by iodoacetamide (Folk and Cole 1966; Holbrook et al. 1973), effectively inhibit enzyme activity. 4 FXIIIa peptide! S H Cys O G i n peptide. + NH, FXIIIa acyl enzyme intermediate peptide, peptide2 = \ FXIIIa + z \ peptide2 peptide, FXIIIa + O " S NH; S H Lys O H e (v-glutamyl)lysyl bond Figure 2. Modified double displacement reaction catalyzed by FXIIIa. Nucleophilic attack of the FXIII active site cysteine thiol group on the y-carboxamide group of a peptide-bound glutamine residue results in the formation of an enzyme acyl intermediate and the release of ammonia. In the second displacement the acyl enzyme intermediate reacts with the e-amino group of a peptide-bound lysine residue leading to the formation of an e(y-glutamyl)lysyl bond and the regeneration of the FXIII enzyme. 1.1.4 Factor X H I a Substrates FXIIIa has been documented to crosslink various proteins in purified systems but there have only been a few proteins which show evidence of having been crosslinked by FXIIIa at more physiological conditions, such as in clotting plasma (Table 1). Model synthetic peptides have been used to determine factors that dictate substrate specificity. In this manner, it has been shown that FXIIIa has precise requirements for certain amino acid side-chains on either 5 Table 1. FXIIIa substrates Proteins identified as substrates for factor XHIa and their crosslinking partners. Substrate Partner(s) System used References Fibrin Fibrin Others (see below) plasma Lorand et al. 1962 Fibrinogen Fibrinogen and Fibronectin purified proteins Procyk and Blomback 1988 Glucagon Primary amines purified system Loewy et al. 1966 Casein Primary amines purified system Dvilansky et al. 1970 Fibronectin Fibronectin Fibrin Staphylococcus aureus Collagen purified proteins plasma purified system purified proteins Mosher 1975 Mosher 1976 Mosher and Proctor 1980 Mosher 1984 a2-macroglobulin Primary amines plasma Mosher 1976 oi2-antiplasmin Fibrin a-chains plasma Sakataand Aoki 1980 Myosin Myosin purified proteins Cohen etal. 1979 Actin Actin purified proteins Cohen et al. 1980 Thrombospondin Fibrin and Thrombospondin plasma purified proteins Bale etal. 1985 von Willebrand factor Fibrin cc-chain Collagen plasma purified proteins Hada etal. 1986 Bockenstedt et al. 1986 Factor V Factor V Actin purified proteins purified proteins Francis et al. 1986 Wang et al. 1990 Vitronectin Vitronectin purified proteins Sane et al. 1988 Vinculin Fibrinogen purified proteins Asijeee/a/. 1988 Elastin Fibrin purified proteins Martine/ al. 1988 Osteopontin Primary amines Unidentified proteins purified proteins Prince et al. 1991 Type-2 plasminogen activator inhibitor Primary amines purified proteins Jensen et al. 1993 Plasminogen Plasminogen Fibronectin purified proteins Bendixen et al. 1993 Lipoprotein(a) Fibrinogen purified proteins Romanic etal. 1998 6 side of the substrate glutamine residue for efficient catalysis (Gorman and Folk 1984; Gorman and Folk 1981; Gorman and Folk 1980). The amine substrates have not shown similar specificity for extended interactions around the substrate lysine residue (Folk 1983). Fibrin has been the best-studied physiological FXIIIa substrate. Fibrin polymerization brings specific regions of the fibrin molecules in close proximity to each other. Glutamine residue 398 from one fibrin y-chain is crosslinked by FXIIIa to y-chain lysine residue 407 of a neighboring fibrin molecule and the reciprocal bond is created as well , forming a y-y dimer (Chen and Doolittle 1969; Doolittle et al. 1971). These y-y dimers appear early during the fibrin clotting reaction and they can be readily identified using electrophoretic analysis (McKee et al. 1970). In a slower reaction, the a-chains are crosslinked into a-polymers (McKee et al. 1970; Folk and Finlayson 1977; Lorand and Chenoweth 1969; Schwartz et al. 1973). There is also a small amount of a-y crosslinking and during prolonged crosslinking reactions y-chain multimers w i l l also eventually form (Mosesson et al. 1989; Shainoff et al. 1991; Siebenlist and Mosesson 1992). FXIIIa prefers fibrin to fibrinogen as a substrate (Lorand and Ong 1966) and there is no evidence that fibrinogen crosslinking is physiologically relevant. A s previously mentioned, the other hematologically important substrate for FXIIIa is the plasma glycoprotein a2-antiplasmin, a potent inhibitor of fibrinolysis which effects this result through direct inhibition of plasmin. Glutamine residue 2 of (X2-antiplasmin is crosslinked by FXIIIa to lysine residue 303 from the a-chain of fibrin at a frequency of about one a2-antiplasmin molecule per every 25 fibrin molecules (Tamaki and A o k i 1982) in a reaction 7 favored over the crosslinking of fibrin a-chains (Tamaki and A o k i 1981). (X2-antiplasmin crosslinking into the fibrin clot is very important for the construction of strong fibrin clots resistant to fibrinolysis (Sakata and A o k i 1982). Besides these hematological factors, FXIIIa has also been shown to crosslink the adhesive proteins, fibronectin, von Willebrand factor and thrombospondin in plasma systems. Fibronectin is a glycoprotein present in plasma, in the extracellular matrix and on cell surfaces, and it appears to be an important factor in tissue remodeling during wound healing (Mosher 1984). Fibronectin is crosslinked to fibrin a-chains through glutamine residue 3 and possibly through another glutamine residue as well (McDonagh et al. 1981; Fesiis et al. 1986). The crosslinking of fibronectin to fibrin has been shown to be important for maximal cell adhesion to a fibronectin-fibrin clot (Corbett et al. 1997). On fibroblast cell layers fibronectin is crosslinked to itself (Barry and Mosher 1989) and there is some evidence that it may also be crosslinked to collagen (Mosher 1984). The binding of fibronectin to fibrin does not appear to compete with that of ci2-antiplasmin (Tamaki and A o k i 1981). The adhesive protein thrombospondin is also crosslinked by FXIIIa in a plasma system. Although the incorporation of thrombospondin into a fibrin clot is not dependent on crosslinking, thrombospondin is crosslinked to fibrin and to itself during clot formation in plasma and has been suggested to contribute to the structure of the clot (Bale et al. 1985). Another adhesive protein crosslinked by FXIIIa in plasma is von Willebrand factor, a large multimeric glycoprotein that is involved in the primary adhesion of platelets to damaged blood vessel walls, von Willebrand factor is crosslinked to the fibrin a-chain during fibrin clots that form slowly, in both its monomeric and its disulfide-bonded polymeric forms (Hada et al. 1986). 8 Similar to results obtained with 012-antiplasmin crosslinking to fibrin, the crosslinking of von Willebrand factor to fibrin does not compete with that of fibronectin. There is some evidence that von Willebrand factor may also be crosslinked to collagen since this bond formation has been shown in a purified system (Bockenstedt et al. 1986). These interactions would be useful in stabilizing fibrin clot association with collagen fibrils in the subendothelium (Hada etal. 1986). Cytoskeletal proteins make up another class of possible substrate proteins for FXIIIa, although evidence of crosslinking at more physiological conditions remains to be shown. Some of these proteins include myosin (Cohen et al. 1979), actin (Cohen et al. 1980), and vinculin (Asijee et al. 1988). In instances where active FXIII may come in contact with these and other cytoskeletal proteins, such as in cells wherein the concentration of FXIII in the cytoplasm is significant, these interactions could be very important for the stability of cytoskeletal structures and contractile ability of the cell. 1.1.5 F X I I I Structure FXIII as originally identified in human plasma is a tetramer composed of two A-chains and two B-chains (Schwartz et al. 1973). The primary structure of both chains has been determined by c D N A cloning and amino acid sequence analysis (Ichinose et al. 1986a; Ichinose et al. 1986b; Grundmann et al. 1986; Takahashi et al. 1986). The A-chain m R N A is 3.9 kb which in its open reading frame codes for 731 amino acids. There is no identified hydrophobic leader sequence in the primary structure of the A-chains (Ichinose and Davie 1988) and it is not clear how FXIII is released into the plasma. It has been suggested that 9 FXIII may be released upon cell destruction (Kaetsu et al. 1996). The A-chains have a molecular mass of approximately 83 kDa each and are predominantly produced by cells of the bone marrow (Wolpl et al. 1987; Poon et al. 1989). The B-chain m R N A is 2.2 kb and codes for a protein of 641 amino acids composed predominantly of 10 "sushi domain" repeats. The B-chains have a molecular weight of approximately 80 kDa each and are produced in the hepatocytes of the liver (Wolpl et al. 1987). The concentration of A-chains in plasma is 0.13-0.16 u M whereas the concentration of B-chains in plasma is roughly double that of the A-chains at 0.26-0.28 u M (Yorifuji et al. 1988). In plasma, two A-chains and two B-chains associate together via non-covalent associations forming the approximately 340 kDa tetramer that circulates as a zymogen. A l l plasma A-chains are complexed in these tetramers while some of the B-chains are found to circulate freely (Yorifuji et al. 1988; K r o l l 1989). The enzymatic activity is found entirely in the A-chains, whereas the B-chains purportedly act as carriers for the A-chain dimers lengthening the half-life of A-chain circulation in plasma (Saito et al. 1990). Besides being found in plasma, FXIII is also normally present in the cytoplasm of cells such as platelets (Buluk 1955; Liischer 1957; Nachman and Marcus 1968), macrophages (Adany etal. 1985; Henriksson et al. 1985), monocytes (Muszbek et al. 1985; Henriksson et al. 1985), and fibroblasts (Fear et al. 1984; Nemeth and Penneys 1989). The FXIII molecules found inside of cells are composed only of the A-chains and exist as dimers. These cellular FXIII A-chains appear to have the identical primary sturcture as the plasma A-chains. Recombinant FXIII A-chains have been expressed in Escherichia coli (Board et al. 1990), Saccharomyces cerevisiae (Bishop et al. 1990; Jagadeeswaran and Haas 1990), and 10 Schizosaccharomyces pombe (Broker and Bauml 1989). These recombinant proteins behave similarly to the native purified protein. The crystal-structure of rFXIII A-chain dimers has been solved from crystals prepared from the zymogen state (Yee et al. 1994; Weiss et al. 1998), from the thrombin-cleaved state (Yee et al. 1995), and from a cation-bound state (Yee et al. 1996). In all cases, the resulting structure of the FXIII A-chain dimer resembles a puckered hexagon with an axis of symmetry formed at the interface where the two A -subunits meet (Figure 3). Each subunit is made up five distinct parts: the N-terminal activation peptide, the B-sandwich, the catalytic core, B-barrel 1, and B-barrel 2. The active-site is normally inaccessible to solvent, therefore, a conformational change is necessary for enzyme activation so that substrate may access the active-site (Yee et al. 1994). It was hypothesized that since the activation peptide of one FXIII A-chain subunit is positioned in front of the active-site of its dimer partner, a thrombin-cleavage would result in removal of the activation peptide allowing the access to the active-site. Unexpectedly, the activation peptide of thrombin-cleaved FXIII was found to remain in association with the FXIII dimer (Yee etal. 1995). 1.1.6 FXIII Activation FXIII is activated in a step-wise process involving cleavage by thrombin although it can also be activated in the absence of thrombin-cleavage (Lorand 1986). For the plasma FXIII tetramer the mechanism of activation begins with thrombin cleavage of the Arg37- Gly38 peptide bond (Takagi and Doolittle 1974). The B-subunits then dissociate in a calcium-dependent manner (Lorand and Konishi 1964) and a calcium-dependent conformational 11 Figure 3. Graphical representation of the FXIII A-chain dimer structure. Alpha helices and beta strands are drawn as cylinders and arrows, respectively, and the active site is marked with a gray sphere. This figure is reproduced from Muszbek et al. (1999). 12 change results in exposure of the FXIII active-site (Figure 4 A ) . The nomenclature for FXIII activated through thrombin cleavage is FXIIIa* or A2* . Although perhaps not particularly relevant physiologically, at calcium ion concentrations greater than 50 m M , plasma FXIII can also attain enzymatic activity in the absence of thrombin cleavage (Credo et al. 1978). FXIIIa 0 or A 2 0 is the nomenclature for FXIII activated in this manner. Fibrin plays a regulatory role in the activation of plasma FXIII , with polymerized fibrin increasing the efficiency of proteolysis by thrombin (Naski et al. 1991; Janus et al. 1983). In a feedback mechanism, once greater than 40% of the fibrin y-chains have been crosslinked by FXIIIa, fibrin no longer acts as a co-factor for activity (Lewis et al. 1985). This safeguards against overproduction of FXIIIa. Since cellular FXIII is found only as an A-chain dimer with the B-chain not being present, the method of activation of cellular FXIII obviously does not involve a calcium-dependent B -chain dissociation step. Cellular FXIII can be activated in a manner similar to that of plasma FXIII activation, involving thrombin cleavage and a subsequent calcium-dependent conformational change to expose the active-site giving rise to FXIIIa* or A 2 * (Figure 4 B) . The physiologic relevance of this is not known but it may be meaningful in the event of release of cellular FXIII to extracellular milieu containing thrombin. For FXIII to become active inside cells, activation would most probably proceed in a thrombin-cleavage independent manner. This makes the discovery of activation of cellular FXIII with high Ca (above 50 m M ) (Credo et al. 1978) or low C a 2 + (2 m M ) and high salt (1 M ) (Polgar et al. 1990) in the absence of thrombin giving rise to FXII Ia 0 or A 2 ° an important observation. Indeed, the demonstrated 13 ( f e V A \ Thrombin y/ ^ ^ 7 B / " A \ Thrombin ^) high [Ca2*] l o w /A*? [Ca2*] high [salt] low [Ca 2 + ] Figure 4. Mechanism of activation of Factor X I I I . Plasma factor XIII , assembled from two A-subunits and two B-subunits, is converted to the active form A2*, through thrombin cleavage in the presence of low calcium ion concentrations, or to the active form A 2 ° , with high calcium ion concentrations (Panel A ) . Cellular factor XIII , made up of two A-subunits only, can be activated with thrombin in the presence of low calcium ion concentrations leading to the formation of A2* or with high calcium ion concentrations in the presence o f low salt concentrations or with high salt concentrations in the presence of low calcium ion concentrations to A2 0 (Panel B) . 14 activation of cellular FXIII in thrombin-activated platelets occurs in the absence of proteolysis (Muszbek et al. 1993; Muszbek et al. 1995). It is not known what calcium-dependent conformational change takes place to expose the active-site cysteine. In the crystal structure of a calcium-bound form of FXII I solved by Yee et al. (1996) there was no major conformational change seen, indicating that calcium ion binding is not sufficient to generate the change in conformation required to expose the active site. The binding of substrate may be involved in conformational change as suggested by the activation enhancement seen in the presence of fibrin. A n alternate mechanism has been proposed whereby two rare non-proline cis peptide bonds, one of which is located close to the active site, may act as a molecular switch between two conformations of the enzyme (Weiss etal. 1998). 1.2 Platelets 1.2.1 Platelet Function Platelets, small anucleate cells that circulate through the body in the blood, were first recognized to play a part in thrombosis over 100 years ago (Osier 1874). They are of primary importance in the body's ability to maintain vascular integrity during wound healing and repair (hemostasis) but they may also be involved in unwanted clot formation i f the hemostatic balance is shifted (thrombosis). When blood vessel endothelium is damaged, elements present at the site of subendothelial exposure such as von Willebrand factor and collagen provide sites for adhesion of platelets. This reversible adhesion initiates the formation of the primary hemostatic plug. A signal is transmitted in the adhered platelets, 15 which results in irreversible adhesion, platelet spreading, platelet aggregation, and release of biologically active molecules from platelet stores. Some of these released biomolecules act to recruit more platelets to the site of injury and rapidly a barrier is formed which prevents loss of blood from the vessel wound. During platelet activation, the platelet surface becomes procoagulant, and initiation of the coagulation cascade on the platelet surface generates thrombin. Thrombin generation leads to the formation of fibrin polymers associated with the platelet plug, as well as resulting in further activation of platelets. Thrombin also activates FXIII which crosslinks and stabilizes the fibrin clot (section 1.1.2). This final phase of consolidation renders the hemostatic plug impermeable to leakage of plasma from the vessel. 1.2.2 Platelet Structure Originally considered by some as cellular debris, platelets are the smallest cells in the blood circulation, measuring between 2 and 4 um. Their concentration ranges from 150 x 10 9 to 450 x 10 9 platelets per liter of blood with approximately 30% of platelets pooled in the spleen at any one time (Aster 1966). Platelets live for about nine days before being removed from the circulation. Regulation of platelet levels is based on total platelet mass (de Sauvage et al. 1996). Platelets originate from megakaryocytes in the bone marrow and to a lesser degree, also in the lung and spleen. Megakaryocyte proplatelet projections are released into the marrow sinus and they then fragment intravascularly creating on the order of 1000 platelets each (Becker and De Bruyn 1976; Zucker-Franklin and Petursson 1984). Because of this unique method o f formation, platelets do not have nuclei o f their own and they have limited ability for synthesizing protein. Platelets however, do contain many specialized organelles, formed in the megakaryocyte precursor cell, unique to their function (Figure 5). 16 Dense tubular system Surface-connected canalicular system Mitochondria Membrane Lysosome g-granules *Fibrinogen *Fibronectin *Factor V *Vitronectin *Thrombospondin *von Willebrand factor *a2-antiplasmin Platelet factor 4 P-thromboglobulin Platelet-derived Growth factor Tumor Growth factor (3 Plasminogen Activator Inhibitor-1 Glycoprotein Ilb/IIIa High Molecular Weight Kininogen CI-inhibitor P-selectin Factor XI Protein S ADP ATP Serotonin Microtubules Cytoplasm *Cytoskeletal proteins Lactate Dehydrogenase Factor XIII Glvcocalyx Glycoprotein Ilb/IIIa Glycoprotein Ib-IX-V Glycoprotein IV Glycoprotein Ia/IIa Glycoprotein Ic/IIa Vitronectin receptor Thrombin receptor ADP receptor Fc receptor Figure 5. Platelet morphology. Some of the organelles of a resting platelet and their contents. The * indicates platelet proteins that can be substrates for FXIIIa. 17 The membrane of the resting platelet is asymmetrically maintained with phosphatidylserine sequestered in the inner leaflet of the membrane. This maintains the platelet surface as non-procoagulant until activation when the expression of phosphatidylserine on the platelet surface produces a procoagulant surface on which coagulation proceeds. Platelets also have two discrete membrane systems not found in other blood cells: the surface-connected canalicular system and the dense tubular system. The surface-connected canalicular system, as well as supplying a source of membrane, provides a ready connection for transfer of molecules from deep inside the platelet to outside and vice versa (White 1974). The dense tubular system, which originates from the rough endoplasmic reticulum of the megakaryocyte precursor (White 1972), sequesters the second messenger, C a 2 + , (Cutler et al. 1978) until rapid calcium ion mobilization is required for activation. The dense tubular system also contains enzymes necessary for prostaglandin synthesis (Gerrard et al. 1976). The microtubule ring, largely composed of the protein tubulin, is thought to maintain the discoid shape of resting platelets (White 1968). Microfilaments, also structure-determining elements of the platelet, found as small irregular networks in resting platelets are rapidly reorganized for cytoskeletal rearrangements upon platelet activation. A s in other cells, mitochondria function to supply platelets with the energy required for their various functions. Dense bodies, so-called because of their electron-dense appearance when viewed with electron microscopy, contain nucleotides, calcium ions, and serotonin, which are released upon platelet activation through fusion of these bodies with the surface-connected canalicular system. One of these released molecules, A D P is an important platelet agonist important for aggregation. The most abundant platelet organelle is the granule. There are two types of platelet granules: lysosomes containing acid hydrolases and a-granules. Alpha-granules, 18 like dense bodies, fuse with the platelet surface-connected canalicular system upon platelet activation, releasing their contents. In these granules, platelets carry many of the proteins and molecules required for their hemostatic function. Many of the proteins released, such as fibrinogen, fibronectin, vitronectin, thrombospondin, and von Willebrand factor (Figure 5) have corresponding receptors on the platelet surface to which they can subsequently bind, mediating adhesion or aggregation responses of the platelet. When released, the local concentration of some of these molecules increases quite significantly. Interestingly, the adhesive proteins released can all behave as substrates for FXIII (Table 1). 1.2.3 Platelet Activation In their resting state, platelets exist in a smooth discoid form and when activated they round up, extend filopodia, and release many bioactive molecules. On a suitable surface, platelets wi l l spread, and when they come in contact with each other, they aggregate together. These events are independent and to a certain extent involve separate activation pathways. The extent of platelet activation depends on the agonist used, with thrombin being the most potent activator of platelets (Davey and Luscher 1967). Thrombin activates platelets through a unique and interesting mechanism involving the proteolytic cleavage of a G protein-coupled membrane receptor. Thrombin cleaves an N -terminal peptide from the receptor unmasking a new N-terminus (Vu et al. 1991). This new N-terminal sequence behaving as a tethered ligand, binds back to the receptor (Chen et al. 1994) and in so-doing sends a signal of activation through the coupled G protein to the platelet. There are two known protease-activated receptors on human platelets, PAR-1 and 19 P A R - 4 (Kahn et al. 1998; Kahn et al. 1999). P A R - 1 is the major mediator for activation of platelets with low concentrations of thrombin (1 nM) and both P A R - 1 and P A R - 4 play important roles for rapid platelet activation at high concentrations of thrombin (30 nM) (Kahn et al. 1999). Platelets activated by thrombin show an unevenly distributed rise in internal calcium ion concentration localized to the pseudopods, periphery and core. This correlates with the redistribution of filamentous actin and glycoprotein Ilb/IIIa. This rise in calcium ion concentration in platelets upon activation is due both to influx of external calcium ions and release of calcium ions into the cytoplasm from internal stores in the dense tubular system. It has been suggested that both the redistribution of glycoprotein Ilb/IIIa and platelet shape change are regulated by local changes in calcium ion concentration which control cytoskeletal reorganization (Ariyoshi and Salzman 1996). 1.2.4 Platelet Adhesive Proteins and Their Receptors The formation of the primary hemostatic plug begins with recognition of various binding sites in the subendothelium by specific receptors on the platelet surface. Many of the platelet receptors for adhesive proteins are integrins. Integrins all have one a-subunit and one P-subunit forming a non-covalent heterodimer complex. The platelet membrane glycoproteins responsible for initial platelet adhesion are the Pi-integrins: glycoprotein Ia/IIa (012P1), glycoprotein Ic/IIa ((X5P1), and V L A - 6 (o^Pi). These receptors bind collagen, fibronectin, and laminin respectively (Staatz et al. 1989; Saelman et al. 1994; Piotrowicz et al. 1988; Sonnenberg et al. 1991). Another platelet glycoprotein involved in the early 20 adhesion response is glycoprotein Ib- IX-V. This receptor, not an integrin, binds to von Willebrand factor immobilized in the subcellular matrix (Table 2). This association is particularly important in areas of disturbed flow where a conformational change in von Willebrand factor is involved in the binding of platelets. The importance of this receptor in adhesion is demonstrated by a severe reduction in adhesion ability under shear conditions with blood from patients who suffer from a lack of glycoprotein Ib in Bernard-Soulier syndrome (Weiss et al. 1978). At a copy number of about 50,000 per platelet, glycoprotein Ilb/IIIa (dubPa) is the most abundant integrin on the platelet surface. When platelets are activated, the glycoprotein Ilb/IIIa receptor complex undergoes a conformational change that exposes a binding site for fibrinogen (Shattil et al. 1985b). This transfer of information from intracellular reactions to the extracellular components of the glycoprotein Ilb/IIIa receptor is known as 'inside-out' signaling. Due to its symmetry, fibrinogen can bind to glycoprotein Ilb/IIIa receptors on separate platelets, bridging these platelets and thereby mediating aggregation. Fibrinogen has at least three different binding sites for glycoprotein Ilb/IIIa: a six amino acid sequence at the carboxy-terminal end of the y-chain of fibrinogen vital for aggregation and two three amino acid sequences (RGD) located in the Aa-chains. Other proteins containing the R G D motif are also able to bind to glycoprotein Ilb/IIIa. Some other ligands for glycoprotein Ilb/IIIa include fibronectin, vitronectin, and von Willebrand factor (Table 2). A related receptor, the vitronectin receptor ( 0 ^ 3 ) also binds proteins via an R G D motif and therefore besides binding vitronectin, this receptor also binds fibronectin, von Willebrand factor and fibrinogen although with different affinities than the glycoprotein Ilb/IIIa receptor. 21 Table 2. Adhesive protein receptors on platelets. Receptor Ligand(s) Alternate Designations Function GPIa/IIa Collagen a 2 pi,VLA-2 Adhesion GPIc/IIa Fibronectin a 5pi, VLA-5 Adhesion VLA-6 Laminin a 6 Pi Adhesion GPIb-IX-V von Willebrand factor Adhesion (shear) GPIV Collagen Thrombospondin GPIIIb Adhesion Vitronectin receptor Vitronectin (RGD) Fibronectin von Willebrand factor Fibrinogen Thrombospondin avp3 Adhesion GPIIb/IIIa Fibrinogen Fibronectin (RGD) Vitronectin von Willebrand factor Collagen aiibP3 Adhesion Aggregation Spreading and Clot retraction 1.2.5 Platelets and FXIII 1.2.5.1 FXIII on the Platelet Surface Thrombin-cleaved FXIII binds to the surface of activated but not resting platelets (Greenberg and Shuman 1984; Kreager et al. 1988). Consequently, surface-bound FXIII can be utilized as a marker for the detection of platelet activation (Devine 1990) and is found on 22 the surface of circulating platelets in patients with peripheral vascular disease (Devine et al. 1993). This binding is mediated through activation of the platelet membrane protein, glycoprotein Ilb-IIIa (Cox and Devine 1994). The binding of active FXII I to active platelets may function to localize and enrich the amount of FXIIIa crosslinking activity to sites of active clot formation. In support of this hypothesis, platelet-associated FXIIIa facilitates fibrin chain crosslinking (Joist and Niewiarowski 1973; McDonagh and McDonagh 1972; Francis and Marder 1987) as well as incorporation of a.2-antiplasmin into growing fibrin clots (Devine and Bishop 1996; Reed et al. 1992; Sakata and A o k i 1980). The crosslinking of a.2-antiplasmin into fibrin clots makes the clot more resistant to proteolysis by plasmin and thereby helps to stabilize the clot (Sakata and A o k i 1982; Reed et al. 1992). In the enhancement of both of these crosslinking reactions, it appears that activated plasma FXIII bound to the surface of platelets may be physiologically more relevant than platelet-derived FXIII itself (Hevessy et al. 1996). 1.2.5.2 FXIII Inside Platelets Approximately half of the body's circulating content of FXIII is found in platelets (Buluk 1955; Kiesselbach and Wagner 1966; Nachman and Marcus 1968; McDonagh et al. 1969; Dvilansky et al. 1970; Day and Solum 1973; Broekman et al. 1975; Lopaciuk et al. 1976). Using an immunoelectronmicroscopic technique, it was seen that the cytoplasmic FXIII was distributed diffusely throughout the whole of the cytosol (Sixma et al. 1984). There has also been one report of the presence of FXIII in platelet a-granules (Marx et al. 1993). Unlike plasma FXIII , which is known to play an important role in clot stabilization, the very significant amount of FXIII found inside the platelet does not yet have a defined role in 23 platelet function. Different lines of evidence suggest that intracellular FXIII may be playing a role in the stabilization of platelets as they form the hemostatic plug. In the early to mid-1980's Cohen et al. documented the crosslinking of platelet proteins by a Ca 2 +-dependent transglutaminase when platelets were stimulated with calcium ionophore (Cohen et al. 1981), with thrombin, or through extensive storage (Cohen et al. 1985). Some of the proteins being crosslinked appear to belong to the group of platelet surface membrane proteins (Cohen et al. 1985). In addition, a crosslinked matrix generally composed of cytoskeletal proteins was identified in platelets and the amount of s-(y-glutamyl)lysine bonds was seen to increase with thrombin stimulation (Harsfalvi et al. 1991). These high molecular weight protein complexes were not formed in either the absence of calcium ion or of thrombin. A D P and epinephrine did not induce crosslinking whereas collagen did so only faintly after a long incubation. The transglutaminase activity in platelets is attributed to cellular FXIII since there is an absence of these highly crosslinked protein polymers in thrombin-stimulated FXIII-deficient platelets (Muszbek et al. 1993). 1.2.6 Clot Retraction After clot formation and stabilization, the clot retracts, squeezing out extra fluid and resulting in marked reduction of clot volume. This process is believed to be a primary step in the clearance of a thrombus, which initiates wound healing. Clot retraction is mediated through interactions between fibrinogen and platelets and can be modeled in vitro with fibrinogen, platelets, and thrombin. Ultrastructure determinations indicate that fibrin strands and platelets orient themselves in the direction of force generation. This suggests that tension develops as platelets simultaneously attach to and spread along fibrin strands, and contract 24 (Cohen et al. 1982). The glycoprotein Ilb/IIIa receptor is required for platelet-mediated clot retraction as determined by inhibition of clot retraction with glycoprotein Ilb/IIIa antagonists and by the fact that Glanzmann's thrombasthenia platelets do not support clot retraction (Carr et al. 1995; Cohen et al. 1982). Analysis of sites on fibrinogen which are important for platelet-mediated clot retraction have revealed that neither of the R G D sequences on the A a -chains nor the binding region on the carboxy-terminal portion of the y-chain are required. These results are consistent with the hypothesis that there exists a novel platelet-binding site on fibrin responsible for clot retraction (Rooney et al. 1998; Cohen et al. 1989). FXIII -mediated crosslinking of fibrin strands is also necessary for the contraction process (Cohen et al. 1982). For the platelet contractile forces to be translated to the fibrin strands through the glycoprotein Ilb/IIIa receptor, there must be a substantial intracellular connection between glycoprotein Ilb/IIIa and the cytoskeleton. This is thought to be mediated by focal adhesion complexes. The importance of these focal adhesion complexes in clot retraction is demonstrated by a corresponding relaxation of clot retraction with the cleavage of focal adhesion molecules by calpain (Schoenwaelder et al. 1997). 1.2.7 Platelet Cytoskeleton Platelets have a cytoskeleton that gives them their form, provides localization points for intracellular elements and allows communication between different parts of the cell. The cytoskeleton is composed of at least three distinct organizations: a central scaffolding based on actin fibrils, a membrane skeleton lying just beneath the cell membrane that provides attachment points for transmembrane receptors, (Fox 1985b) and a tubular coil describing the circumference of the platelet. (White 1971; White et al. 1984) (Figure 6). The different 25 GPIb-IX-V Cortical actin filaments Figure 6. Mode l of the resting platelet cytoskeleton. Some of the cytoskeletal elements of the resting platelet. In the resting platelet most of the actin is not filamentous. F-actin can be found near the membrane as short actin-filaments containing spectrin, composing the membrane skeleton. The GPIb-IX-V receptor is connected to the cortical actin filaments through filamin (actin-binding protein). The GPIIb/IIIa receptor is not associated with the resting platelet cytoskeleton. elements of the cytoskeleton are in constant communication in order to maintain the discoid form of resting platelets circulating through the blood. These dynamic elements of the 26 cytoskeleton are poised for a platelet activation signal, which wi l l initiate major cytoskeletal rearrangements. It is not surprising when one considers the role of the platelet that almost 50% of all platelet proteins are contractile proteins. Act in itself comprises 20-30% of the total platelet protein (Harris and Weeds 1978) of which approximately 40% is found in its filamentous form in resting platelets, with the rest as globular monomers (Fox et al. 1984). Short actin filaments are located in the region of submembraneous microfilaments and longer actin filaments radiate out from the center of the platelet. These longer cortical actin filaments are involved in the attachment of GPIb to the resting platelet cytoskeleton (Hartwig and DeSisto 1991) (Figure 6). The platelet cytoskeleton is responsible for the generation of forces required for platelet shape change, granule release, aggregation, and clot retraction (Harris 1981). The generation of tension, the extension of the cell membrane, and movement of structures within the platelet are accomplished through generation of mechanical energy by contractile proteins such as filamentous actin, myosin, a-actinin, tropomoysin, and vinculin. These proteins all harness the energy released from A T P hydrolysis. Assembly of actin filaments is associated with the earliest physical responses of surface-activated platelets and the filamentous actin content of activated platelets increases to between 60%> and 80% of the total actin (Jennings etal. 1981; Carlsson et al. 1979). The membrane skeleton increases its associations with the cortical actin filaments and proteins associated with the membrane skeleton are increasingly detectable in the cytoskeletal fraction. These increased associations between the two 27 separate actin skeletal structures may in part be due to the formation of focal adhesion complexes. Focal adhesions involve the clustering of ligand-bound adhesion receptors with the assembly of cytoskeletal proteins on the cytoplasmic side of the integrins connecting the extracellular matrix with the intracellular actin network (Figure 7). Some of the cytoskeletal proteins involved in forming the intracellular attachment site between the integrin cytoplasmic tail and actin filaments are vinculin, talin, paxillin, a-actinin, and tensin. The signalling molecules p p l 2 5 F A K and pp60 c" s r c are also found in association with focal adhesions. Vincul in is a membrane skeletal protein found in focal adhesions closely associated with another membrane protein, talin. It has a globular head region and a tail-like end. Vincul in is known to translocate to the cytoskeletal fraction upon platelet aggregation. Talin, spectrin, glycoprotein Ilb/IIIa and pp60c-src show a similar distribution indicating a possible association with the membrane skeleton (Fox et al. 1993). Talin is localized throughout the cytoplasm of resting platelets, but after platelet activation talin is found at the cytoplasmic side of the plasma membrane (Beckerle et al. 1989). 28 Adhesive Protein Integrin Membrane -Capping protein a-actinin Talin Paxillin Vinculin Actin filaments Figure 7 . Model of a Focal Adhesion Complex. A model for the focal adhesion complex of a generic adhesive protein integrin, such as glycoprotein Ilb/IIIa. 1.3 Rationale and Objectives The general objective of this thesis was to gain a greater understanding of the function of platelet FXIII both on the outside and the inside of the platelet. The demonstrated binding of activated FXIII to the platelet surface (Greenberg and Shuman 1984) contributes to the strength and stability of platelet-fibrin clots (Hevessy et al. 1996) and may be relevant in the stabilization of adhesive protein interactions with both platelets and the subendothelial matrix. Although it is not known exactly which specific interactions are responsible for 29 binding of FXIIIa to the platelet surface, it is known that both activated glycoprotein Ilb/IIIa and fibrinogen are important mediators of the process (Cox and Devine 1994). Expression of endogenous platelet FXIII on the platelet surface (Kreager et al. 1988) also contributes to fibrin and a2-antiplasmin crosslinking (Hevessy et al. 1996) although the mechanism for release and expression of this endogenous FXIII pool is not known. Considering the importance of FXIII presented by the platelets in normal hemostasis, it was important to learn more about the interactions between FXIIIa and the platelet surface. I hypothesized that FXIII binds to platelets through fibrinogen bound to glycoprotein Ilb/IIIa and can crosslink fibrinogen directly to the platelet surface or to itself on the surface of platelets. The following specific aims were designed to test this hypothesis: 1. To determine factors important for the binding of FXIII to platelets. 2. To determine the nature of the interactions between FXIIIa and fibrinogen on the platelet surface. 3. To determine whether endogenous platelet FXIII could contribute to surface fibrinogen crosslinking. Platelets contain approximately half of the circulating amount of FXIII (McDonagh et al. 1969, Dvilansky et al. 1970) yet there is very little known about what role this cytoplasmic pool of FXIII may be playing inside the cell. It is activated in a slow, non-proteolytic manner in thrombin-stimulated platelets (Muszbek et al. 1995) and is responsible for the calcium ion-dependent crosslinking of platelet proteins in calcium ionophore-stimulated, thrombin-stimulated, and energy depleted platelets (Cohen et al. 1981; Cohen et al. 1985). The presence of cytoskeletal proteins, tubulin and filamin in the crosslinked matrix of 30 thrombin-stimulated platelets has been noted (Harsfalvi et al. 1991) and FXIIIa can catalyze the crosslinking of cytoskeletal proteins, myosin, actin, and vinculin in vitro (Cohen et al. 1979; Cohen et al. 1980; Kahn and Cohen 1981; Asijee et al. 1988). These discoveries have led researchers to hypothesize that intracellular FXIII is involved in the stabilization of cytoskeletal structures after platelet activation. The following specific aims tested this hypothesis: 1. To determine whether FXIII could be interacting with cytoskeletal proteins upon platelet activation. 2. To determine whether FXIII was localized to areas of cytoskeletal rearrangement. 3. To ascertain whether enzymatically active FXIII was associated with the cytoskeleton. 4. To identify which cytoskeletal proteins were being crosslinked upon platelet activation. 31 C H A P T E R 2 M A T E R I A L S A N D M E T H O D S 2.1 Materials 2.1.1 Proteins, peptides, and other chemicals Recombinant human placental FXIII produced in a Saccharomyces cerevisiae yeast expression system (Bishop et al. 1990) was the gift of Zymogenetics (Seattle, W A , U S A ) . In both functional and immunological tests rFXIII produced in Saccharomyces cerevisiae has been shown to behave the same as human FXIII A-subunits (Bishop et al. 1990; Rinas et al. 1990). N,N-dimethyl casein (bovine), G P R P , and P P A C K were obtained from the Calbiochem-Novachem Corporation (La Jolla, C A , U S A ) . Apyrase (grade VII), iodoacetamide, leupeptin, and other chemicals were obtained from the Sigma-Aldrich Corporation (St. Louis, M O , U S A ) . a-Thrombin was the gift of Dr. John W . Fenton II (New York State Department of Health, Albany, N Y , U S A ) . 2.1.2 Antibodies Polyclonal rabbit anti-FXIII antibodies, which were subsequently conjugated with fluorescein isothiocyanate (FITC), were the kind gift of Zymogenetics. Polyclonal rabbit anti-human FXIII A-chain antibodies were obtained from the Calbiochem-Novachem Corporation (La Jolla, C A , U S A ) and polyclonal goat anti-human FXIII antibodies were obtained from American Diagnostica (Greenwich, C T , U S A ) . Monoclonal anti-actin (C4) 32 antibodies were from Chemicon International Inc. (Temecula, C A , U S A ) . Polyclonal goat anti-filamin antibodies and monoclonal anti-vinculin (VIN-11-5) antibodies were obtained from the Sigma-Aldrich Corporation. HRP-conjugated donkey anti-rabbit, antibodies, H R P -conjugated goat anti-mouse antibodies, and FITC-conjugated rabbit IgG whole molecule control antibodies were obtained from Jackson ImmunoResearch Laboratories Inc. (West Grove, P A , U S A ) . FITC-conjugated polyclonal rabbit anti-fibrinogen antibodies were obtained from the Dako Corporation (Carpenteria, C A , U S A ) . FITC-conjugated monoclonal anti-CD62P (CLB-Thromb/6), PE-conjugated monoclonal anti-CD42b (SZ2), F ITC-conjugated mouse I g G l and PE-conjugated mouse I g G l isotype control antibodies were purchased from Immunotech (Marseille, France). 2,2 Platelet Isolation To accurately assess the effects and interactions of FXIII with platelets, it was necessary to remove platelets from plasma in order to separate them from the high levels of FXIII found in plasma. The method chosen for platelet isolation, modified from a procedure by Mustard et al. (1989), consisted of a series of centrifugation steps optimized to rapidly obtain a good yield of platelets that had undergone a minimal amount of activation. Washed platelets were prepared fresh on the day of experimentation from volunteer donors who had not ingested aspirin in the week previous to donation. To minimize handling-induced activation of the platelets, all buffers in contact with the platelets were warmed to room temperature before use and all washing steps proceeded in polypropylene tubes. Whole blood was collected via a lOcc syringe into acid citrate dextrose ( A C D ) : 11.5 m M citric acid monohydrate, 88.5 m M trisodium citrate dihydrate, 111 m M dextrose, p H 6.0, anticoagulant at a ratio of 1:9 33 ACD:b lood or in a sodium citrate Vacutainer tube (Becton Dickinson, Franklin Lakes, N J , U S A ) . The first few milliliters of blood drawn were discarded to avoid contamination with the tissue factor contained in the skin plug. The blood was subjected to centrifugation at 140 x g for 15 minutes to collect the red blood cells, and the resulting platelet-rich plasma (PRP) was removed and diluted 1:1 with A C D . The platelets were sedimented at 500 x g for 15 minutes and washed twice in modified Hepes Tyrode's buffer ( M H T B ) : 10 m M Hepes, 137 m M N a C l , 2.7 m M KC1, 0.4 m M N a H 2 P O 4 . H 2 0 , 1.2 m M N a H C 0 3 , 0.1% dextrose, 0.2% bovine serum albumin (BSA) , p H 6.5 (500 x g, 10 minutes). The M H T B was supplemented with 1 U / m l apyrase immediately before use in the washing steps. The platelets were resuspended in M H T B , the platelet count was obtained on a Coulter S T K R blood cell counter (Miami, F L , U S A ) , and the count was adjusted to obtain an incubation concentration of 2 x 10 8 platelets/ml unless otherwise indicated. The platelets were then allowed to 'rest' for 30 to 60 minutes at 37°C to recover from handling. These platelets were minimally activated as determined by the expression of platelet granule-release marker, CD62P, measured using flow cytometry, contained less than 5% contamination by red blood cells, and less than 0.1 % contamination by white blood cells as measured on the cell counter. This procedure routinely produced around 4 ml of washed platelets from a starting volume of 10 ml of anticoagulated blood. For the preparation of gel-filtered platelets, collection proceeded as described above in the washing procedure. Instead of being processed with repeated centrifugation steps however, they were passed over a 19 x 1 cm Sepharose C L 2 B (Pharmacia Biotech, Uppsala, Sweden) 34 column equilibrated with M H T B to separate them from plasma. The turbid fractions were collected, adjusted to the desired concentration and kept at 37°C for use. 2.3 Platelet Incubations for Flow Cytometric Binding Determinations One hundred to five hundred ul suspensions of freshly isolated platelets resuspended to a concentration of 2 x 10 8 platelets/ml in M H T B were recalcified with 2 . 5 - 1 0 m M C a C l 2 and 25 u M G P R P was added so they would not aggregate when activated. The optimal concentration of G P R P to use with washed platelets was determined with titration to be 25 -35 u M . In this range, the least amount of aggregation occurred and inhibition of fibrinogen binding was only ~ 9% (Figure 8). Platelets were then treated with or without 1 - 1 0 U/ml thrombin and an equal amount of buffer was added to the non-treated samples. After the incubation, 1-10 u M P P A C K was added to all samples to stop any thrombin activity. For experiments in which the reversibility of binding was studied, the samples were then incubated in the presence or absence of 15 m M ethylenediamine tetraacetic acid ( E D T A ) for 30 minutes at 37°C. The samples were prepared for flow cytometry and run within 4 hours of platelet collection. For the flow cytometry of whole blood, 5 ul of whole blood collected in sodium citrate containing Vacutainer tubes was supplemented with 1.25 m M G P R P . The samples to be activated were incubated with 35 U/ml thrombin for 10 minutes and the reaction was stopped with P B S : 8% N a C l , 0.2% KC1, 0.2% K H 2 P 0 4 , and 0.53% N a 2 H P 0 4 (5 m M phosphate, 140 m M NaCl) containing 1% P F A . 35 2.4 Pre-activation ofrFXIII with Thrombin For some experiments preactivated rFXIIIa was added to the platelet incubations and was prepared as follows: rFXIII was incubated with 10 U / m l thrombin and 10 m M C a C b for 10 minutes at room temperature after which the thrombin was inactivated by the addition of 10 u M P P A C K . The preactivated rFXIIIa was prepared shortly before use and kept on ice until that time. For some experiments a thrombin*PPACK control was also prepared in the identical manner, substituting M H T B for rFXIII. 2.5 FXIIIPre-incubation with Iodoacetamide FXIII was pre-incubated with iodoacetamide for some of the flow cytometry binding studies of FXIII and platelets. A 1 ml solution in Hepes buffer containing 150 n M rFXIII, 10 m M CaCb , and 1.35 m M iodoacetamide was incubated for five minutes at room temperature. A handling control, which contained all the ingredients except for the iodoacetamide, was processed identically. The solutions were filtered in Centricon-30® concentrators (Amicon) at 5000 x g for 15 minutes after which 1 ml M H T B was added to each sample. The filtration procedure was repeated 5 times for a final dilution of the starting material of approximately 25000X to remove any excess iodoacetamide. After the final centrifugation, the samples were removed, a small volume rinse of the Centricon-30® concentrators was added to the samples and the protein concentrations were determined using a Bio-Rad Protein Assay (Bio-Rad Laboratories, Hercules, C A , U S A ) based on the Bradford dye-binding procedure (Bradford 1976). 36 2.6 Platelet Flow Cytometry Five ul of platelets at 2 x 10 8 platelets/ml or 10 ul of fixed whole blood and 5 ul fluorescein isothiocyanate (FITC)-conjugated antibodies and/or phycoerythrin (PE)-conjugated antibodies were incubated in a final volume of 50 ul with M H T B . Saturating concentrations of each antibody were used. F ITC rabbit anti-FXIII was used at a final concentration of 2.5 ug/ml, F ITC rabbit anti-fibrinogen was used at a final concentration of 15.7 ug/ml, FITC monoclonal anti-CD62P was used at a final concentration of 2.5 ug/ml, and P E monoclonal anti-CD42b was used at the recommended package insert dilution of 20 u l / 5x l0 5 cells. The control antibodies were used at the same final concentrations as those of the corresponding test antibodies. After a 30 minute incubation at room temperature shielded from ambient light, the platelets were diluted with 1 ml of 0.2% formaldehyde in saline or for the fixed whole blood samples, with 250 ul buffer. These samples were passed through a Coulter EPICS® X L - M C L model flow cytometer (Miami, F L , U S A ) for analysis. The platelets were confirmed as a single population and a bitmap was drawn around them based on their forward scatter and side scatter characteristics. A second bitmap corresponding to platelet aggregates was also included for some experiments. For a few experiments, only events positive for the pan-platelet marker, CD42b (GPIba), were analyzed to assure that only platelets were assessed. This procedure was discontinued, as the population of washed platelets was homogeneous and deemed not to require further identification. The fluorescence measurement gates were set to include approximately the upper 5% of signals obtained with F ITC- and/or PE-conjugated control antibodies and above. The signals collected within the gates for the test samples were counted as positive events for the 37 antibody marker. The percentage of positive events as well as the mean fluorescence of the positive population were recorded. 2.7 Platelet Incubations for SDS-PA GE Analysis Five hundred microliters to 1 ml of washed platelets at a concentration of 2 x 10 8 platelets/ml were treated with or without 10 U/ml thrombin for various different times in the presence of 10 m M C a C U or 14 m M E D T A . In some experiments, the platelets were pre-incubated with 270 u M iodoacetamide for 10 minutes at 37°C to inhibit possible crosslinking activity. A s iodoacetamide is capable of crossing the cell membrane (McAbee and Grinnell 1982) both internal and external inhibition of FXIIIa would be achieved. For activated yet non-aggregated platelets, 25 u M G P R P was added to the platelets before activation and they were incubated in the absence of stirring. For aggregated samples, G P R P was omitted from the incubation mixture and the platelets were incubated on a rotator. For those samples in which rFXIII was added to the incubation mixture it was ascertained that under the conditions used, most of the rFXIII was cleaved by thrombin to its activated state (rFXIIIa) by 5 minutes with over half having already been converted by 2 minutes (Figure 11). The thrombin was inactivated with 10 u M P P A C K . 2.8 Platelet Lysate Preparation Platelets were lysed at 0°C with ice-cold buffer containing 1% Triton X-100 which disrupts the plasma membrane releasing the contents of the cytoplasm. For the purpose of isolating the different platelet skeletal fractions for S D S - P A G E , the platelet lysis buffer contained enzyme inhibitors: 10 m M Tris, 150 m M N a C l , 1% Triton X-100, 2 m M ethylene glycol - bis 38 (B-aminoethyl ether) - N , N , N ' , N ' - tetraacetic acid ( E G T A ) , 200 ug/ml leupeptin, 10 m M benzamidine, 200 u M P M S F , 250 ug/ml iodoacetamide, p H 7.4. For the purpose of measuring cytoplasmic FXIII activity, the platelet lysis buffer contained an added reducing agent but did not contain iodoacetamide: 10 m M Tris, 150 m M N a C l , 1% Triton X-100, 200 ug/ml leupeptin, 10 m M benzamidine, 200 u M P M S F , 2 m M dithiothreitol, pH 7.4. 2.9 Platelet Cytoskeletal Fractionation Studies have shown that it is possible to separate the cytoskeletal and membrane skeletal fractions of the platelet through a differential centrifugation method (Fox et al. 1988). Platelet lysate was subjected to centrifugation in a Beckman Microcentrifuge (Beckman Coulter, Fullerton, C A , U S A ) at 12,400 x g for 7 minutes at 4°C. This resulted in a pellet referred to as the low speed pellet (LSP) that contains platelet cytoskeletal proteins. The supernatant was subjected to centrifugation in a Beckman Tabletop Ultracentrifuge at 100000 x g for 3 hours at 4°C. This resulted in a pellet referred to as the high-speed pellet (HSP) that contains platelet membrane skeletal proteins. The L S P was washed once with buffer containing 1% Triton X-100 before continuing the analysis. 2.10 Separation of Platelet Proteins using SDS-PAGE For gel electrophoretic studies the platelet cytoskeletal fractions were solubilized in l /20 t h their original volume with reducing sample buffer: 62.5 m M Tris (pH 6.8), 2% SDS, 5% B-mercaptoethanol, 10% glycerol and 0.004% bromophenol blue or a solubilization buffer: 62.5 m M Tris (pH 6.8), 6 M urea, and 2% SDS. The samples were then heated in 39 boiling water for 5 minutes. If the urea solubilization buffer was used, it was possible to determine the protein concentration of the samples using the Bicinchoninic acid ( B C A ) protein assay from Pierce (Rockford, IL, U S A ) . Afterwards, B-mercaptoethanol at a final concentration of 5% and bromophenol blue at a final concentration of 0.05% were added to these samples and they were subjected to boiling again for 2 minutes. Discontinuous SDS-polyacrylamide gel electrophoresis was performed according to Laemmli (1970) on the Mini-gel Protean II apparatus from Bio-Rad (Hercules, C A , U S A ) . Gels with 15 wells were poured containing a 6 or 7.5% polyacrylamide running gel depending on the molecular weight fractionation range desired. The samples were loaded onto the gel and a current was applied until the dye front approached the bottom of the gel. The gel proteins were either stained directly with silver or transferred to nitrocellulose for further evaluation with Western blot technique. The apparent molecular weight of the protein bands could be determined by comparison with the relative mobilities of known protein standards run on the same gel. 2.11 Gel Silver Staining The enhanced silver-staining protocol of Heukeshoven and Dernick (1988) was used. Briefly, after electrophoresis, the polyacrylamide gels to be stained for protein content were soaked anywhere between 2 hours and overnight in a solution composed of 10% acetic acid and 30%) ethanol. Three 20 minute rinses with 30%> ethanol were followed by 2 quick rinses with water and a 1 minute incubation in 250 ug/ml sodium dithionite. After three more washes with water, the gel was incubated for half an hour in a 0.2%o silver nitrate solution containing 0.0028%) formaldehyde and then rinsed twice with water. The developing solution: 40 mg/ml potassium carbonate, 0.0168%) formaldehyde, and 3.17 ug/ml sodium 40 thiosulfate, was added to the gel with gentle agitation until the development had proceeded long enough to see the protein bands (approximately 2 x 5 minutes). The development was stopped with the addition of 5% acetic acid and the gel was stored in distilled water until dried onto filter paper. 2.12 Western Blot Analysis of Platelet Proteins The identity and different forms of platelet proteins were determined with an immunodetection method. The proteins separated on S D S - P A G E gels were transferred to nitrocellulose membranes for a total of 100 Volt hours using a Bio-Rad Transblot apparatus (Bio-Rad Laboratories, Hercules, C A , U S A ) . The membranes were processed as described in the protocol booklet for the Enhanced Chemiluminescence ( E C L ) kit (Amersham Pharmacia Biotech, Uppsala, Sweden). Briefly, the membranes were blocked with Tris-buffered saline (TBS): 10 m M Tris, 150 m M N a C l (pH 7.4) containing 0.1% Tween 20 supplemented with 5% Carnation skim milk powder ( M T B S T ) for a minimum of 2 hours. The blocked membranes were then incubated for 1 hour with an M T B S T solution containing antibodies specific for the protein in question at the appropriate dilution determined previously by titration. This was followed by 3 washes in T B S 0.1%> Tween 20 and a 1 hour incubation with an M T B S T solution containing horseradish peroxidase (HRP)-conjugated antibodies directed against the primary antibodies. After 5 washes the membranes were taken to a dark room and the images visualized with the E C L kit on Hyperfilm (Amersham Pharmacia). The apparent molecular weight of the identified proteins could be determined by comparison with the relative mobilities of known rainbow protein standards (Amersham Pharmacia) run on the same gel. To re-probe a membrane with an antibody of different 41 specificity, the membrane was stripped of proteins. This proceeded as described in the protocol booklet for the E C L kit. Briefly, the membrane was incubated in a shaking water bath at 50°C for 30 minutes with stripping buffer: 62.5 m M Tris (pH 6.7), 100 m M B-mercaptoethanol, and 2% SDS. It was then washed 2 X 1 0 minutes in large volumes of T B S 0.1% Tween 20 and finally blocked for 1 hour with T B S containing skim milk powder and Tween 20. Antibody incubations proceeded as described above. 2.13 Platelet Aggregometry Platelet aggregometry was conducted on a Chrono-Log Corporation whole blood lumi-aggregometer (Havertown, P A , U S A ) . Aggregation was measured via optical transmittance. Two hundred and fifty microliters of washed platelets resuspended in M H T B were supplemented with 1 mg/ml fibrinogen and equilibrated with stirring on the aggregometer to obtain a base line reading. Pre-activated rFXIIIa or the rFXIIIa activation mix lacking rFXIII ( thrombin-PPACK) was added to the platelets. After 1 minute 7 u M calcium ionophore (A23187) was added to activate the platelets and the aggregation was followed for exactly 6 minutes. The aggregometer software, Aggro/Link version 4.75 calculates both the percent amplitude measurements and the slope of aggregation. The percent amplitude corresponds to the amount of aggregation in percentage where the starting platelet suspension is set as 0%> and the buffer is set at 100%. In this experiment, the aggregation reactions had reached a plateau by 6 minutes and the percent amplitude was measured at this time. The slope of aggregation is measured at the steepest point of the aggregometry curve and indicates the fastest rate at which aggregation proceeds. 42 Aggregometry for the vinculin cleavage experiment was conducted in larger aggregometry cuvettes to allow for sub-sampling during the course of aggregation without interference of the aggregometry measurements. Three milliliters of washed platelets in M H T B were placed in cuvettes and zeroed against M H T B in the aggregometer. The aggregations were initiated by the addition of 10 U / m l thrombin containing 10 m M C a C b in the presence or absence of stirring. A t various time-points, 500 pi aliquots of the platelet incubation were removed from the ongoing aggregation and placed directly into 500 pi ice-cold lysis buffer: 2% Triton X-100, 10 m M Tris, 150 m M N a C l , 4 m M E G T A , 400 ug/ml leupeptin, 400 p M P M S F , 20 m M benzamidine, 2.7 m M iodoacetamide, and 20 u M P P A C K . The platelet aggregation response was checked before and after the experiment to be sure that the platelets had not lost any activity over the time of the experiment. The L S P cytoskeletal fractions were prepared from the platelet lysate samples for Western blot analysis with monoclonal anti-vinculin antibodies. 2.14 FXIII Overlay Experiments Platelet cytoskeletal fractions were separated on 7.5% polyacrylamide S D S - P A G E gels and then transferred to nitrocellulose membranes as described above. The membrane was blocked overnight in T B S containing skim milk powder and Tween 20 and for the calcium ion addition experiment cut into lane-width strips so that each strip could be incubated with a different solution. The membrane/membrane strips were then incubated in T B S containing skim milk powder and Tween 20 containing rFXIII, rFXIIIa or rFXIII pre-incubated with calcium ions (rFXIII»Ca 2 + ) for 2 hours rocking gently at room temperature. rFXIIIa was prepared by incubating 6.1 u M rFXIII in T B S with 10 U / m l thrombin in the presence of 43 10 m M Ca for 10 minutes after which the thrombin was inactivated with the addition of 10 u M P P A C K . rFXIII»Ca 2 + was prepared by incubating 61 n M rFXIII in T B S containing the desired amounts of CaCh and N a C l for 1 hour at 37°C. The membrane/membrane strips were then processed as described in the method insert for the Amersham E C L kit. Briefly, the membrane/membrane strips were washed with T B S 0.1% Tween 20, incubated with a primary rabbit antibody against FXIII (1:2000) for 1 hour, washed again, incubated with an HRP-conjugated secondary donkey anti-rabbit antibody (1:10000) for 1 hour, washed again, and visualized with the Amersham E C L kit. 2.15 Fluorescence Microscopy Platelets were collected, treated and activated as indicated in the figure legends of each experiment. For all experiments, at the end of the specified incubation times an equal volume of 4% paraformaldehyde in P B S was added to the incubation mixtures and they were left to fix for 30-60 minutes at 37°C. The fixed platelets were permeabilized with 0.2% Triton X-100 in P B S for 5-10 minutes at 37°C and washed twice in P B S by centrifugation at 750 x g for 5 minutes. For all the antibody incubations, the platelets were resuspended with P B S supplemented with 0.1% B S A . To label the cellular FXIII , the permeabilized washed platelets were incubated with FITC-conjugated rabbit anti-FXIII at an antibody concentration of 25 ug/ml for 30 minutes. To label the cellular vinculin, the permeabilized platelets were incubated with monoclonal anti-vinculin at a concentration of 18 ug/ml for 1 hour followed by two washes in P B S and a secondary antibody incubation (FITC-conjugated goat anti-mouse) for 30 minutes. FITC-conjugated rabbit IgG (anti-FXIII control) and mouse IgG followed by FITC-conjugated goat anti-mouse (anti-vinculin control) 44 used at the same concentrations as the specific test antibodies were negative for staining. The platelets were viewed with an OPTIPHOT-2 Nikon microscope (Tokyo, Japan) with episcopic-fluorescence (mercury lamp), EFD-3 and phase contrast attachments. A 100X oil immersion objective with a numerical aperture of 1.25 was used. A n attached Pixera Professional digital camera (Pixera Corporation, Los Gatos, C A , U S A ) captured both phase contrast and fluorescence images. 2.16 FXIII Activity Assay on Platelet Lysate and Cytoskeletal Fractions Washed platelets resuspended to 2 - 4 x 108 platelets/ml in T B S , p H 7.4 containing 35 u M G P R P and 2.5 m M C a 2 + were treated with or without 1 U / m l thrombin for 30 minutes at 37°C. P P A C K was then added to the thrombin-activated samples at a final concentration of 10 u M and incubated for a further 10 minutes at room temperature. Thrombin that had been similarly pre-incubated with P P A C K was added to the 'resting' samples to control for the presence of thrombin in the platelet lysate. The platelets were lysed on ice with an equal volume of ice-cold activity assay 2 X lysis buffer: 10 m M Tris, 150 m M N a C l , 2% Triton X-100, 20 ug/ml leupeptin, 400 u M P M S F , 20 ug/ml benzamidine, and 4 m M dithiothreitol. Platelet lysis using sonication instead of lysis buffer was utilized for one of the experiments. In this case, a Braun-Sonic 1510 sonicator (Braun Instruments, South San Francisco, C A , U S A ) with a needle probe was used ( 3 x 5 second pulses with 5 second intervals) to lyse platelet samples kept at 0°C over ice. The cytoskeletal fraction was obtained with a 10 minute, 12,400 x g centrifugation at 4°C and washed once with T B S . To generate the positive control, 10 U / m l thrombin was added to a lysate sample prepared in the 45 absence of enzyme inhibitors and incubated for 30 minutes at 37°C. This was done to activate all possible FXIII present in the lysate. The FXIIIa activity was measured with a putrescine incorporation filter-paper assay modified from Lorand et al. (1972). FXIIIa wi l l crosslink the standard small molecule substrate, 1 4C-putrescine, to casein. The amount of crosslinking can then be quantified after trapping and washing the casein on filters by subsequently counting the radioactivity associated with each filter. Twenty microliters each of the positive control lysate, the test lysate, the entire cytoskeletal pellet in T B S , or T B S alone (background control) were mixed with 30 pi of the substrate mixture at 30 second intervals and incubated for 30 minutes at 37°C. The final reaction volume contained 5 uCi/ml 1 4C-putrescine, 0.88% N,N-dimethylcasein and 2.5 m M CaCl2. The reaction was stopped and the dimethylcasein was precipitated with 4 ml ice-cold 10% trichloroacetic acid (TCA) . The samples were poured over G F / C glass fiber filters (Whatman, A n n Arbor, M I , U S A ) on a sampling manifold (Millipore, Bedford, M A , U S A ) which had been pre-blocked with 2% B S A in T B S . The filters were rinsed 3 times with ice-cold 5% T C A , once with ethanol:acetone (1:1), and finally with acetone. Once dry, the filters were counted in 5 ml Cytoscint™ES* ( ICN Pharmaceuticals Inc., Costa Mesa, C A , U S A ) for 10 minutes each on a Philips P W 4700 liquid scintillation counter (Fisher Scientific, Atlanta, G A , U S A ) 1 4 C program. 2.17 Lactate Dehydrogenase (LDH) Assay The amount of platelet lysis occurring during the platelet incubations for fibrin(ogen) crosslinking experiments was determined by measuring the release of the cytoplasmic marker 46 lactate dehydrogenase (LDH) . Three hundred microliters of washed platelets in M H T B were activated with 10 U / m l thrombin in the presence of 10 m M C a C b for 10 minutes and then sedimented at 500xg for 10 minutes. A 100% lysis control was generated by resuspending the same number of platelets in 0.3% Triton X-100 T B S . A n L D H assay kit from S I G M A Diagnostics (St. Louis, M O , U S A ) was used to measure the L D H release in the platelet incubation supernatants according to the reverse reaction: pyruvate + NADH—LDH ^lactate + NAD+. In a fume hood, 1 ml of the pyruvate substrate was added to a vial containing 1 mg N A D H . This mixture was incubated for 3 minutes at 37°C. One hundred microliter aliquots of this mixture were transferred to nine 12x75 mm test tubes, capped, and further incubated for another 3 minutes at 37°C. One hundred microliters of the incubation supernatant or of the control samples were added and incubated for 30 minutes at 37°C. The color reagent was then mixed in and incubated at room temperature for an additional 20 minutes. Finally, 1 ml of 0.4 N N a O H was mixed in. After 5 minutes, 200 pi from each tube were transferred into a microtiter plate and the plate was read at a wavelength of 450 nm. The test sample absorbance values were transformed to percent L D H activity by comparison with the buffer control (0%> lysis) and the 100% lysis control absorbances. 2.18 Statistical Analysis A l l error bars depicted on graphs represent the standard deviation of at least three independent measurements. The paired t-test was used to analyze differences between paired samples of one factor and differences between groups with more than one variable were analyzed using A N O V A , two-factor with replication software in the Microsoft Excel Data 47 Analysis ToolPak. The a posteriori test used to identify where the differences lay i f any, was Newman-Keuls' multiple-range test. Significance levels for the correlation coefficients were obtained from the published statistical table (Rohlf and Sokal 1981, 168). For all statistical analyses, differences between groups were considered significant at less than a 5% chance of rejecting the null hypothesis (p < 0.05). 48 C H A P T E R 3 FXIII O N T H E OUTSIDE O F H U M A N P L A T E L E T S 3.1 Exogenous FXIIIa Binding to Isolated, Thrombin-Activated Platelets The binding of activated FXIII to the platelet surface can be detected using flow cytometry (Kreager et al. 1988). The fibrin polymerization inhibitor, G P R P is used to keep platelets from aggregating and in so doing allows analysis of individual platelets on a flow cytometer. The optimal concentration of G P R P for use in this washed platelet system was determined to be between 25 and 35 u M as that was the concentration that least affected platelet fibrin(ogen) binding while still inhibiting platelet aggregation (Figure 8). The F ITC-conjugated polyclonal anti-FXIII antibody used to detect this association does not distinguish between the zymogen form (FXIII) and the activated form (FXIIIa) of factor XIII; however, it is known that activated FXIII binds specifically to platelets while the zymogen FXIII does not (Greenberg and Shuman 1984). Figure 9, panel A shows that FXIIIa binds to a similar extent to both thrombin-activated platelets and to washed platelets not treated by thrombin. A n average of 67% of platelets activated by thrombin had FXIIIa detectable on their surface and 57%o of platelets not activated by thrombin also had FXIIIa on their surface; this difference was not statistically different. However, taken together, both of these percentages were significantly higher than the approximate 10%> of platelets with detectable FXIIIa on their surface, treated with or without thrombin in the absence of rFXIIIa. The average mean fluorescence of anti-FXIII antibody binding to platelets in the presence of rFXIIIa and activated by thrombin was 9.6 mean fluorescence intensity units whereas for those not 49 10000 [GPRP] uM Figure 8. Effect of GPRP on platelet aggregate formation and fibrinogen binding. Washed platelets in M H T B were incubated at room temperature with lOU/ml thrombin in the presence of l O m M C a C ^ and varying concentrations of G P R P . A t 10 minutes, the incubation was stopped with 1%PFA. The fixed platelets were analyzed on a flow cytometer for the ratio of platelet singlets to aggregates (A) and for the amount of fibrinogen expression (B) using F ITC anti-fibrinogen antibodies. 50 a 100 x u. si .9 t « Vj O) 1-Q. X 1/1 — "a. 80 60 40 20 0 9.6il.9 12.347.5 2.9±1.7 2.9±2.1 2.5 thrombin + rFXIIIa + resting + ft ioo 0 V s-a. x at — ct "a 80 60 40 5 20 0 thrombin + rFXIIIa + resting + Figure 9. rFXIIIa binding to thrombin-activated and non-activated platelets. Platelets in M H T B were incubated with or without 1 U / m l thrombin in the presence of 2.5 m M C a C l 2 and 25 u M G P R P for 5 minutes after which 305 n M rFXIIIa or a thrombin»PPACK control was added for a further 10 minute incubation before washing twice. The resting samples were those platelets that had not been handled after being isolated. Platelets were double-labeled with P E anti-CD42b and F ITC anti-FXIII (Panel A ) or P E anti-CD42b and FITC anti-CD62P (Panel B) antibodies for flow cytometric evaluation. Mean fluorescence intensity values for the positive populations are shown above their respective bars. The difference between samples incubated with rFXIIIa and those incubated without rFXIIIa was significant in panel A (p < 0.01, n = 3, A N O V A ) . The difference between samples treated with thrombin and those not treated with thrombin was significant in panel B (p < 0.01, n = 3, A N O V A ) . 51 activated by thrombin it was 12.3. The mean fluorescence readings were also significantly higher in both samples containing rFXIIIa than in samples not containing rFXIIIa (2.9) although not statistically different from each other. Figure 9, panel B , which illustrates CD62P expression, confirms that the platelets treated with thrombin had been activated as an average of greater than 80% of them did release their alpha-granules, whereas platelets that did not receive treatment with thrombin had not released their alpha-granules to the same extent (25 - 40%>). Approximately 25% of the washed platelets left to rest expressed the CD62P marker. Looking at the individual experiments for platelets not treated with thrombin and containing rFXIIIa, a subset of between 15%> and 55% of platelets were able to bind FXIIIa in spite of not being activated to the point of alpha-granule release. The reversibility of FXIIIa binding to platelets was also examined. Since the binding of FXIIIa to activated platelets is known to be mediated by glycoprotein Ilb/IIIa (Cox and Devine 1994), washed platelets activated with thrombin were incubated with or without pre-activated rFXIIIa and the platelets were then incubated with E D T A in conditions known to dissociate the glycoprotein Ilb/IIIa receptor (Shattil et al. 1985a). The binding of rFXIIIa to the platelets was seen to be partially reversible as indicated by the ability of E D T A to allow some but not all of the rFXIIIa to dissociate from the platelet surface (Figure 10A). Approximately 20% less platelets expressed FXIIIa after the E D T A incubation and the signal mean intensity of the remaining positive cells decreased from 17.1 to 3.74. Overall there was about a 5-fold loss in fluorescence signal associated with FXIIIa binding on the platelet surface with the E D T A incubation. A decrease in surface-bound FXIIIa with E D T A was also seen with the activated platelets incubated in the absence of rFXIIIa although to a much 52 > 8.= 3 Q — 100 80 60 40 20 0 thrombin rFXIIIa E D T A + + * 1 7 3.7 + + + + + + 100 > — ^ | U a c: 80 60 40 20 0 thrombin rFXIIIa E D T A + + + + + + + + Figure 10. The effect of E D T A on the binding of activated FXIII to activated platelets. Washed platelets in MHTB, pH 7.4 were activated with 1 U/ml thrombin in the presence of 2.5 mM CaCl2. After a one minute incubation, the thrombin was inactivated with PPACK. The samples were then incubated for 10 minutes at 37°C in the presence or absence of 61 nM rFXIIIa. rFXIIIa was not added to the non-thrombin treated samples. Finally, the samples were treated with or without 15 mM EDTA for 30 minutes at 37°C. The percentage of platelets expressing FXIII (A) and CD62P (B) was determined using flow cytometry. The mean intensity of the signal is shown inside the corresponding bars (* p < 0.01, A N O V A , n = 3). 53 smaller scale. A s well , in similar results to those seen in, Figure 9, Figure 10A shows that greater than 90% of the platelets activated by thrombin in the presence of exogenously added rFXIIIa had FXIII on their surface, whereas in the absence of exogenous rFXIIIa, ~ 4% of platelets express any FXIII . This is slightly more than in the case of platelets not activated by thrombin, upon which the gates for positive events were set at just under 2% of the platelets. The activation of the platelets was confirmed by measuring the expression of the degranulation marker, CD62P. 100% of the platelets activated with thrombin were positive for CD62P whereas only 26% of the platelets not activated by thrombin were positive for CD62P (Figure 10B). E D T A treatment resulted in an unexplained increase in granule-release under the conditions used, as 65% of the platelets not stimulated by thrombin but with E D T A treatment were positive for the CD62P marker. This EDTA-dependent increase in granule-release was observed on at least six separate occasions. 3.2 Effect of Iodoacetamide on rFXIIIa Binding to Activated Platelets Since iodoacetamide would be used in subsequent experiments to inhibit FXIII activity it was important to determine the effect of iodoacetamide on the platelet FXIII interaction. Washed platelets in an incubation mixture containing 305 n M rFXIII, 10 m M C a 2 + , and 25 u M G P R P were incubated with 10 U/ml thrombin in the presence or absence of iodoacetamide. The cleavage of rFXIII to its active form, rFXIIIa, was confirmed under these conditions to be mostly complete by five minutes of incubation (Figure 11). Surprisingly, iodoacetamide was seen to enhance the amount of FXIII detectable on the platelet surface (Figure 12). In the absence of exogenously added rFXIII , iodoacetamide did not affect the amount of FXIII detectable on the platelet surface. In three separate but similar experiments, inclusion of 54 FXIII FXIIIa 0 2' 5' 10' 30' 2h 0 2' 5' 10' 30' 2h Figure 11. rFXIII conversion in the platelet incubation mixtures. Washed platelets were incubated with 10 mM CaCh, 50 ug/ml rFXIII, and 10 U/ml thrombin. At various time-points (0, 2 minutes, 5minutes, lOminutes, 30minutes, and 2hours) the thrombin was inactivated with 10 p M PPACK. The platelets were pelleted at 10000 x g for 10 minutes and the incubation supernatant was prepared for SDS-PAGE and subsequent analysis with silver stain (A) and with a Western blot for rabbit anti-human FXIII A-chain (B). Molecular weight markers (kDa) are shown at the left. 55 — 12 -, l 10 B « 8 E 6 B 4 o fi 2 -S 0 rFXIII -IAA + + + Figure 12. Effect of iodoacetamide on binding of rFXIIIa to thrombin-activated platelets. Washed platelets in M H T B were incubated at 37°C with or without 305 n M rFXIII and with or without 1.35 m M iodoacetamide ( IAA) in the presence of 10 m M C a C h , 25 p M G P R P , and 10 U / m l thrombin in M H T B buffer. The thrombin was inactivated after 30 minutes with 10 u M P P A C K and the platelets were incubated with F ITC anti-FXIII antibodies for flow cytometric evaluation. These results show the results of one experiment and are representative of three similar experiments. 56 iodoacetamide led to an increase in surface-bound FXIII between 1.5 to 4 times as much as that in the absence of iodoacetamide. The varying conditions were: 1.35 m M or 11 m M iodoacetamide, 305 or 610 n M rFXIII, T B S or M H T B buffer, 5 or 10 m M C a C l 2 , o 1.5 - 2 x 10 platelets/ml, and the presence or absence of G P R P . This consistent and unexpected result was further investigated with an experiment in which the concentration of iodoacetamide was varied. A dose-response curve was obtained (Figure 13) wherein the signal mean intensity of rFXIIIa binding to thrombin-activated platelets in the absence of any iodoacetamide was measured at ~ 7 and increased with increasing iodoacetamide until a plateau between 10 and 11 was reached. This plateau was achieved at an iodoacetamide concentration of ~ 1.35 m M after which no further increase was obtained. Since iodoacetamide was added to the incubation mixture that contained both rFXIII and platelets the next question asked was whether the iodoacetamide enhancement of FXIIIa binding to the platelet was due to the interaction of iodoacetamide with FXIII or directly with the platelet. Iodoacetamide was incubated with rFXIII and then washed away before incubation of the rFXIII with the platelets. There appeared to be no difference in either the percentage of platelets expressing FXIII or in the mean fluorescence intensity signal between platelets incubated with rFXIII pre-incubated with iodoacetamide or platelets incubated with rFXIII that had not been treated with iodoacetamide (Figure 14). The ability of the rFXIIIa that had been pre-incubated with iodoacetamide to crosslink fibrin was the same as that of rFXIIIa which had not been pre-incubated with iodoacetamide indicating that the active-site cysteine had not been adequately alkylated during the iodoacetamide pre-incubation (data not shown). 57 ••a e « H *s ^ fl b-cu fl fl es cu 12 10 8 6 4 2 0 -•— activated! • resting 5000 10000 Iodoacetamide (uM) Figure 13. Iodoacetamide dose-response on rFXIIIa binding to platelets. Washed platelets in M H T B were incubated at 37°C with 305 n M rFXIII and 10 U / m l thrombin in the presence of 10 m M C a C ^ , 25 u M G P R P and varying concentrations of iodoacetamide (0, 1.35 u M , 13.5 u M , 135 u M , 1.35 m M , and 11 mM) . The thrombin was inactivated after 30 minutes with 10 u M P P A C K and the platelets were incubated with FITC anti-FXIII antibodies for flow cytometric evaluation. The resting control platelets did not have any additives but were diluted with an equivalent amount of M H T B . 58 30 c OS U H E o i e 25 20 15 10 5 rFXIII pre-incubated with IAA +rFXIII -rFXIII resting Figure 14. Effect of iodoacetamide pre-incubation with FXIII on platelet FXIII binding. Washed platelets in M H T B were incubated with 10 U/ml thrombin in the presence of 61 n M rFXIII that had been pre-incubated with iodoacetamide ( IAA) , a sham-treated rFXIII control, or in the absence of rFXIII. The incubation mixtures also contained 10 m M CaCl2 and 25 u M G P R P except for the resting platelet samples, which had no additions. After a 30 minute incubation, 10 p M P P A C K was added to the thrombin-containing samples and the platelets were prepared for flow cytometry with F ITC anti-FXIII antibodies. Error bars indicate the standard deviation for incubations performed in triplicate. 59 A s iodoacetamide is known to inhibit many enzymes, the effect of iodoacetamide on platelet activation was also examined to find conditions where minimum effects on the platelet activation-state would be obtained. Increasing concentrations of iodoacetamide did inhibit platelet activation by thrombin as indicated by platelet surface expression of the activation markers, CD62P and fibrinogen (Figure 15). Iodoacetamide at a concentration of 270 u M , the concentration chosen for FXIIIa inhibition studies, showed minimal effect. A t this concentration of iodoacetamide the percentage of platelets expressing CD62P was not affected and the number of platelets expressing fibrinogen was inhibited by approximately one quarter. 60 Figure 15. Effect of iodoacetamide on platelet activation as measured by the expression of platelet activation markers, C D 6 2 P and fibrinogen (Fbgn). Whole blood was treated with 35 U / m l thrombin in the presence of 1.25 m M G P R P and increasing amounts of iodoacetamide (0, 270 u M , 5.4 m M , and 11 m M ) . After 10 minutes, the samples were fixed with 1% P F A and prepared for flow cytometric evaluation with F ITC anti-CD62P and FITC anti-fibrinogen antibodies. 61 3.3 Effect of rFXIIIa on Fibrin(ogen) Binding to Platelets At the same time as the FXIII binding on activated platelets incubated with rFXIIIa was assessed, the expression of fibrin(ogen) on the platelet surface was also investigated. It was thought, since both fibrin(ogen) and FXIIIa can bind to the platelet surface, that one of the mechanisms by which FXIIIa acts to strengthen and stabilize a growing clot may be in the crosslinking of fibrin(ogen) directly to the platelet surface. This was investigated in a system using washed platelets, with exogenously added rFXIIIa, where E D T A was used to dissociate fibrin(ogen) from the platelet surface. If the rFXIIIa did indeed crosslink fibrin(ogen) to the platelet surface then it was expected that less fibrin(ogen) would be removed by the action of E D T A from the platelet samples containing rFXIIIa. Figure 16 shows that this was not the case. E D T A treatment of thrombin-activated platelets incubated in the presence of rFXIIIa resulted in the loss of approximately 25% fibrin(ogen) from the platelet surface whereas those treated in the absence of rFXIIIa demonstrated a loss of around 15%> fibrin(ogen). If anything, the E D T A was able to remove even more of the fibrin(ogen) from those platelets incubated with rFXIIIa. Platelet activation was confirmed by measuring the percentage of platelets expressing the platelet granule-release marker, CD62P. Approximately 99%> of all the thrombin-activated platelets were positive for CD62P, ~ 25% of the washed platelets not stimulated by thrombin and not treated with E D T A were CD62P positive and ~ 65%> of the platelets not stimulated by thrombin but with E D T A treatment were CD62P positive (Figure 10B). The platelet-activating effect of E D T A treatment of non-thrombin activated platelets seen by an increase in CD62P expression was mirrored by an increase in fibrin(ogen) binding. The platelets not activated by thrombin show a relatively high amount of fibrin(ogen) binding due to the washing steps utilized to obtain the plasma-free platelets. 62 a CU > s ® OX O 8. cu "H. 100 80 60 40 20 -i-0 thrombin rFXIIIa E D T A + + + + + + + Figure 16. The effect of FXIIIa on the ability of E D T A to dissociate fibrin(ogen) from the platelet surface. Washed platelets resuspended in MHTB at a pH of 7.4 and supplemented with 2.5 mM CaCl2 were activated with 1 U/ml thrombin. After a one minute incubation, the thrombin was inactivated with 1 uM PPACK. The platelets were then incubated for 10 minutes at 37°C in the presence or absence of 61 nM rFXIIIa. Finally the samples were treated with or without 15 mM EDTA for 30 minutes at 37°C. The platelets were prepared for flow cytometry with FITC anti-fibrinogen antibodies. Error bars indicate the standard deviation of incubations performed in triplicate. Statistics included a two-factor A N O V A , n = 3 and Newman-Keuls' multivariable analysis. The first sample (no additives) was significantly lower (p < 0.05) than all other samples except for the 6 t h (+ thrombin, + rFXIIIa, + EDTA). The second sample (+ EDTA) was significantly different (p < 0.05) from all other samples except for the 4 t h (+thrombin, + EDTA). The third sample (+ thrombin) was significantly higher (p < 0.05) than all other samples except for the 5 t h (+ thrombin, + rFXIIIa). 6 3 Although it does not appear that rFXIIIa is acting to crosslink fibrin(ogen) directly to glycoprotein Ilb/IIIa on the platelet surface, it is clear from Western blot experiments, that the fibrin(ogen) bound to the platelet was crosslinked into higher molecular weight species in the presence of rFXIIIa (Figure 17). The fibrin(ogen) monomers Act, a and y present in the L S P of platelet samples incubated in the presence of iodoacetamide, a FXIIIa inhibitor, have disappeared in the samples incubated in the presence of rFXIII. In their stead can be seen a band migrating at approximately 93 kDa most probably representing the fibrin y-y dimer and high molecular weight complexes greater than 200 kDa at the top of the resolving gel and in the well of the stacking gel. 64 wells high <— molecular weight 200 — 97 — i— y-y dimer 30 — 1 2 3 Figure 17. The effect of rFXIIIa on the form of platelet-bound fibrin(ogen). Washed platelets in a suspension containing 25 u M G P R P and 10 m M CaCh were activated with 10 U/ml thrombin in the presence of 270 p M iodoacetamide (lane 2) or 305 n M rFXIII (lane 3). A non-activated control sample is shown in lane 1. After a 15 minute incubation at room temperature the thrombin was inactivated with 20 p M P P A C K and the cells were washed again. The platelets were lysed and the low speed pellet was resuspended in reducing sample buffer for electrophoresis on a 7.5% S D S - P A G E gel and Western blot analysis with rabbit anti-fibrinogen antibodies. Molecular weight marker positions (kDa) are indicated on the left. These results are representative of results obtained on nine separate occasions. 65 3.4 Effect of rFXIIIa on Aggregation During the course of performing flow cytometry experiments on washed platelets, it became apparent that there were more platelet aggregates detectable in samples containing exogenously added rFXIIIa than in those with no rFXIIIa present (Figure 18). The percentage of aggregates, on the order of 1%, was very low due to the presence of G P R P in the incubation mixture, which had been added to inhibit aggregation for flow cytometry. The percentage increase seen in the presence of pre-activated rFXIIIa was therefore also very low at approximately 0.5% and yet still statistically significant (p < 0.05). Furthermore, combining data obtained in three separate experiments from samples that had been variously treated with or without thrombin, with or without rFXIIIa addition, and with or without E D T A in the wash, the number of small platelet aggregates correlated positively with the amount of FXIII detected on single platelets (Table 3). This correlation (r = 0.39) was statistically significant (p < 0.05), and is most probably an under-estimation of the actual strength of the correlation as noting the r-values from the individual experiments, the data appear more strongly correlated. The discrepancy most probably relates to having drawn the platelet aggregate bitmap independently in each experiment so that the percentage of platelet aggregates can not be properly compared. There was also found to be a strong correlation (r = 0.89, p < 0.01) between the amount of FXIII detectable on the platelet surface and the amount of surface-bound fibrinogen (Figure 19). This correlation was mirrored in the amounts of FXII I and fibrin(ogen) associated with the low speed pellet of thrombin-stimulated platelets at increasing incubation time points (Figure 20). The increasing amount of fibrin(ogen) visible at the top of the gel in the loading wells (Figure 20, Panel A ) was matched by an increasing amount of FXIII in both the load wells and at approximately 66 2.5 5 2 es 0) •-J3 Q, 1.5 a> 1 0.5 0 • -rFXIIIa • +rFXIIIa thrombin treated non-thrombin treated Figure 18. Effect of rFXIIIa on the number of platelet aggregates. Washed platelets in MHTB supplemented with 2.5 mM CaCb. and 25 p M GPRP were treated with 1 U/ml thrombin or left untreated for 5 minutes. 305 nM rFXIIIa was added and after a further incubation of 10 minutes the thrombin was inactivated with 5 p M PPACK and the platelets were washed in MHTB. The percentage of platelet aggregates was determined using flow cytometry with a bitmap drawn at the area corresponding to platelet aggregates. The difference between samples containing or not containing rFXIIIa was significant (p < 0.05, two-factor A N O V A , n=3). 67 Table 3. Correlation between the number of small platelet aggregates and FXIII detection on single platelets. Washed platelets in MHTB, supplemented with 10 mM CaCb and 25 uM GPRP were incubated with or without 10 U/ml thrombin in the presence or absence of 305 nM rFXIIIa. After 15 minutes the thrombin was inactivated with 10 uM PPACK and the platelets were washed in the presence or absence of 15 mM EDTA. The platelets were fixed with 1% PFA and flow cytometric measurements were obtained. Correlation data r-value n Significance Experiment 1 60 C J c i- J ' 60 0 " ^ ( ) 0.5 1 Mnl FXIII r = 0.89 11 p<0.01 Experiment 2 % aggregates OOO 0 5 1 M n l FXIII 0 r = 0.87 12 p<0.01 Experiment 3 t3 6 60 A to 4 -60 T 60 I ' CO 5? 0 -0 5 1 Mnl FXIII 0 r = 0.45 11 n.s. Amalgamated data % aggregates OOO f—v-#*s#» ) 5 1 Mnl FXIII 0 r = 0.39 34 p < 0.05 68 Figure 19. Correlation between the expression of fibrinogen and FXIII on platelets. Washed platelets in M H T B , supplemented with 10 m M C a C l 2 , 25 p M G P R P and 305 n M rFXIIIa were incubated in the presence or absence of 10 U / m l thrombin. After 15 minutes the thrombin was inactivated with 10 p M P P A C K and the platelets were washed in the presence or absence of 15 m M E D T A . Flow cytometric measurements using F ITC anti-fibrinogen and F ITC anti-FXIII antibodies were obtained from three experiments conducted on separate days (r = 0.89, n = 34, p < 0.01). 69 A B 1 f l ft 2 0 0 2 0 0 -9 7 -6 9 -3 0 - ' 4 6 -3 0 ~ 0 2' 5' 10' 30 ' 0 2' 5* 10' 30* Figure 20. Fibrinogen and FXIII in the low speed pellet of thrombin-stimulated platelets. Washed platelets at 1 x 10 platelets/ml in Hepes buffered Tyrode's containing albumin ( H B T A ) were incubated with 6 U / m l thrombin, 10 m M C a C l 2 , and 305 n M rFXIII. A t various different time points (0, 2 minutes, 5 minutes, 10 minutes, and 30 minutes) the thrombin was inactivated with 10 p M P P A C K and the platelets were washed three times in H B T A . Platelet lysate was prepared and the low speed pellet was isolated, solubilized in reducing sample buffer and separated with S D S - P A G E . The samples were subjected to Western blot technique using rabbit anti-fibrinogen as the primary antibody (A) or rabbit anti-FXIII as the primary antibody (B) and peroxidase-conjugated anti-rabbit as the secondary antibody. Molecular weight markers (kDa) are indicated on the left of each blot. 70 85 kDa (Figure 20, Panel B) . The location of FXIII in the loading wells is not due to the chemical crosslinking of FXIII into high molecular weight polymers but rather to the incomplete solubilization of the cytoskeletal proteins in the absence of urea-containing solubilization buffer. Using a urea-containing buffer to solubilize the L S P fraction proteins, there is no FXIII detectable in the loading wells (data not shown). To examine the FXIII association with platelet aggregation further, aggregometry experiments were conducted both in the presence and absence of rFXIIIa. Samples containing exogenously added rFXIIIa tended to result in slightly higher aggregation levels than in those samples not containing rFXIIIa (Table 4). The increase in aggregation levels did not achieve statistical significance; yet, the rate of aggregation in samples containing rFXIIIa did increase significantly compared to those samples not containing rFXIIIa (p < 0.01). These aggregometry experiments of washed platelets indicated a moderate increase in rate for the progression of aggregation in the presence of 305 n M rFXIIIa, although the final amount of aggregation was not significantly altered. 71 Table 4. The effect of rFXIIIa addition on platelet aggregation. Washed platelets resuspended in M H T B were equilibrated in an aggregometer with rFXIIIa, or a thrombin-PPACK control. Aggregation was initiated with 7 u M calcium ionophore (A23187) and the aggregation response was followed for 6 minutes (n=3, two-factor A N O V A ) -rFXIIIa + rFXIIIa Significance Slope of aggregation 76.0 ± 4.6 92.5 ± 4 . 8 p<0.01 Percent amplitude 54.3 ± 6 . 8 62.0 ± 8 . 1 n.s. 72 3.5 Endogenous FXIII Expression on Thrombin-Activated Platelets Endogenous platelet FXIII originating not from plasma or an exogenous source but from the platelets themselves was also expressed on the surface of activated platelets. In experiments using washed platelets there was a significant increase in the measurable amount of FXIII on the platelet surface after thrombin-activation (Figure 21 A ) . Approximately 10% of activated platelets expressed measurable amounts of FXIII on their surface compared with approximately 1% of non-thrombin activated platelets. This increase was statistically significant and reproducible on 4 separate occasions. The surface expression of FXIII corresponded with the surface expression of the markers CD62P and fibrin(ogen) (Figure 21, Panels B & C). Approximately 15% of the washed non-thrombin treated platelets expressed CD62P, while after thrombin-activation greater than 99% of the platelets were positive for CD62P indicating that generally all the platelets had undergone an alpha-granule release response upon thrombin stimulation. The washed non-thrombin treated platelets similarly expressed approximately 15% fibrin(ogen), while after thrombin-activation approximately 45% of the platelets had fibrin(ogen) on their surface. This indicates that fibrinogen had been released from the platelet and that the platelet fibrinogen receptor, GPIIb/IIIa, had changed to its functionally active binding conformation. The washing procedure which tends to slightly activate a small proportion of the platelets before the thrombin addition and the fact that the only source of the fibrinogen is the platelets themselves most probably accounts for the fact that only half the platelets are positive for this marker. The FXIII did not appear to be originating from the cytoplasm of lysed platelets as there was no evidence of platelet lysis in similar conditions containing even higher concentrations of thrombin (see section 5.1). 73 Figure 21. Expression of endogenous FXIII on thrombin-activated platelets. Washed platelets were activated with 1 U/ml thrombin in the presence of 2.5 mM CaCb. After 10 minutes, the thrombin incubation was stopped with the addition of 10 uM PPACK. The platelets were washed and tagged with FITC anti-FXIII (Panel A), FITC anti-CD62P (Panel B), or FITC anti-fibrinogen (Panel C) for flow cytometric analysis. * p < 0.05, n = 3, paired t-test. 74 I 50 © ox c IS) a 40 1 30 20 2 io CU « a 0 no thrombin thrombin-activated isotype control 75 3.6 Discussion The discovery that FXIIIa binds specifically to activated platelets (Greenberg and Shuman 1984) has stimulated research into the contribution of platelet-associated FXIIIa to clot formation and stabilization. With this interaction, platelets effectively localize activated FXIII to the site of hemostatic plug formation, yet the nature of FXIIIa interaction with the platelet surface is not fully understood. In this study, the further characterization of FXIIIa interactions with the platelet surface and with fibrin(ogen) on the platelet surface was undertaken to gain a greater understanding of these processes that are important to hemostasis. When washed platelets were incubated in the presence of rFXIIIa, the rFXIIIa bound to washed platelets in relatively high amounts and equally well whether the platelets had been stimulated with thrombin or not. It is probable that in the absence of thrombin stimulation, the sheer stress of centrifugation and washing that the platelets were subjected to during the experiment induced FXIIIa binding sites on the platelets. Through the course of these experiments it was generally noted that platelets isolated from plasma through centrifugation expressed higher levels of the activation markers, CD62P and fibrinogen than platelets not subjected to centrifugation. Activation of platelet GPIIb/IIIa has been shown to occur under high shear stress using parallel-plate perfusion chambers (Holme et al. 1997). A s GPIIb/IIIa mediates FXIIIa binding to activated platelets (Cox and Devine 1994) it is possible that the washed platelets in this study bound FXIIIa in a mechanism mediated through the shear-induced activation of GPIIb/IIIa. Different agonists, used to activate platelets in vitro, differentially affect both the amount of FXIIIa that binds to platelets (Greenberg and Shuman 76 1984) and the percentage of platelets that bind FXIIIa (Cox and Devine 1994). Thrombin activation leads to approximately 100% of platelets binding FXIIIa in the greatest amounts, ionophore activation results in approximately 100% of platelets binding FXIIIa but at a much lesser amount, and adenosine diphosphate (ADP) activation results in almost no FXIIIa binding. The equivalent amount of FXIIIa binding seen on both non-thrombin-treated washed platelets and on thrombin-stimulated platelets (to up to 90%> of platelets) in this study indicates that sheer stress may be a potent inducer of platelet FXIIIa binding sites. Studies directly examining the binding of FXIIIa to platelets that have not been subjected to centrifugation and washing steps, under controlled shear conditions, would be required to elucidate conditions under which the apparent shear stress induction of FXIIIa binding to platelets could be occurring. The binding of rFXIIIa to platelets did not correlate with a-granule release; in separate experiments, 15%> to 55% of the non-thrombin-treated platelets that bound anti-FXIII antibodies had not released their a-granules as measured by CD62P expression. This indicates that platelets become competent to bind FXIIIa without requiring alpha-granule release. This has also been observed on platelets activated with either epinephrine (adrenaline) or collagen where 20%> - 25% of platelets bind FXIIIa although less than 5%> of them express the a-granule marker, CD62P (Cox and Devine 1994). The elevated levels of CD62P expression that were seen in washed non-thrombin-treated platelets (25%) to 40%) of platelets), although significantly lower than the levels seen in thrombin-stimulated platelets, are probably due to sheer-induced activation during the centrifugation and washing steps. There is evidence based on a reduced inhibition of fibrinogen binding with GPIIb/IIIa antagonists that a proportion of platelets prepared by a sedimentation and 77 washing technique undergo a-granule release during the washing procedure (Gralnick et al. 1991). As FXIIIa binding is mediated by the GPIIb/IIIa receptor (Cox and Devine 1994), further understanding of the nature of FXIIIa binding to platelets was gained by measuring the amount of surface-bound FXIIIa after uncoupling the GPIIb/IIIa receptor. The irreversible uncoupling of the GPIIb/IIIa heterodimer complex achieved by incubation of the platelets with E D T A for 30 minutes at 37°C (Shattil et al. 1985a) led to a five-fold decrease in the amount of FXIIIa binding to the platelet surface. However, the amount of FXIIIa binding did not return to baseline indicating that the binding was only partially reversible. The partial reversibility of FXIIIa binding to the platelet is also demonstrated in an experiment where addition of 100-fold molar excess unlabelled FXIIIa leads to only 10%-30% reversibility of FXIIIa binding (Greenberg and Shuman 1984). The partial reversibility of FXIIIa binding to activated platelets is reminiscent of the partial reversibility of binding seen with fibrinogen (Peerschke 1988), another protein whose binding to the platelet surface is mediated by GPIIb/IIIa. There was a strong correlation (r = 0.89) measured between the amount of FXIIIa binding and that of fibrin(ogen) binding to the surface of washed platelets which had been treated with or without thrombin and with or without E D T A . This suggests an association between these two proteins. In support of this hypothesis, GPIIb/IIIa antagonists that inhibit fibrinogen binding to platelets also inhibit FXIIIa binding and in a purified system, addition of fibrinogen to GPIIb/IIIa-coated wells enhances FXIIIa binding (Cox and Devine 1994). FXIIIa could be associating with the platelet in a manner very similar to that of fibrinogen or its association to the platelet may be mediated by fibrinogen. It has been reported that activated FXIII ( A 2 * chains) but not unactivated FXIII ( A 2 chains) bind to fibrin 78 (Procyk et al. 1993). If FXIIIa is binding to platelets through its interaction with fibrin(ogen), this is in agreement with the finding that FXIIIa but not FXIII binds to platelets (Greenberg and Shuman 1984). However, the model can not be as simple as a direct association between FXIIIa and fibrinogen on activated platelets because some agonists that induce fibrinogen binding such as A D P do not induce the binding of FXIIIa (Greenberg and Shuman 1984; Cox and Devine 1994). Perhaps a conformational change in platelet-bound fibrin(ogen), brought about by only certain agonists, or a change in the platelet surface itself near the fibrin(ogen) binding site is required to induce the FXIIIa binding sites. A possible candidate for a platelet surface molecule interacting with fibrin(ogen) to induce a FXIIIa binding site is GPIIb/IIIa. This member of the integrin family which is the fibrin(ogen) receptor is definitely in close proximity to fibrin(ogen), it undergoes a conformational change upon platelet activation, and FXIIIa can bind directly to immobilized GPIIb/IIIa in a purified system free of fibrin(ogen) (Cox and Devine 1994). In some experiments, iodoacetamide was included in the incubation mixture to inhibit FXIIIa activity. Surprisingly, the presence of iodoacetamide enhanced FXIIIa binding to the platelet. This suggested that a freely accessible active-site was not required for FXIIIa binding to activated platelets and that the platelet-binding site on FXIIIa may be separate from the active-site. This is a similar finding as for that of the interaction between FXIIIa and its fibrin substrate determined through studies with carbamylmethylated platelet FXIII (Hornyak and Shafer 1992). The requirement for activation of the FXIII molecule before binding to platelets, evidenced previously (Greenberg and Shuman 1984), suggests that a conformational rearrangement in the FXIII molecule upon activation confers to it the ability 79 to bind to platelets. The enhanced FXIIIa binding to a saturable level seen in the presence of iodoacetamide further suggested that occupation of the active site of FXIIIa rather than detrimentally affecting FXIIIa binding to the platelet, was inducing or stabilizing a conformation with increased ability to bind to platelets. Evidence from a FXIII antibody found in a patient with impaired fibrin crosslinking, indicates that a conformational change upon cleavage of the FXIIIa activation peptide is sufficient for exposure of a binding site for fibrinogen (Fukue et al. 1992). FXIII does undergo a conformational change upon activation, through cleavage of the activation peptide. A model for the conformational alteration, which would allow access to the active-site cysteine, was proposed based on the FXIII crystal structure (Yee et al. 1994). When exposed, the active-site cysteine can be alkylated with iodoacetamide; this has been used to titrate active-site cysteine exposure (Chung et al. 197'4; Curtis et al. 1973). A n alternate explanation for the enhanced binding of FXIIIa to thrombin-activated platelets observed in the presence of iodoacetamide could be that iodoacetamide was acting directly on the platelets to increase their ability to bind FXIIIa. A t the lowest iodoacetamide concentration to show enhanced rFXIIIa binding (1.35mM) there were some effects demonstrated in platelet activation marker expression: CD62P expression decreased a little (~ 4%) and fibrinogen binding decreased by almost 40%. Iodoacetamide showed the opposite effect on fibrin(ogen) binding to the platelet as that on activated FXIII . Since fibrin(ogen) appears to mediate the binding of FXIIIa this is opposite to what one would expect i f the iodoacetamide enhancement of FXIIIa binding is related to its action on the platelet FXIIIa binding site but it is still possible that iodoacetamide affects a site separate from that of fibrin(ogen) binding which may enhance FXIIIa binding. To distinguish whether iodoacetamide was acting on rFXIIIa or on the platelets, rFXIII was pre-80 incubated with iodoacetamide and any 'free' iodoacetamide was washed away before incubating activated rFXIIIa with thrombin-activated platelets. There was no difference in platelet binding of rFXIIIa pre-treated with iodoacetamide or without. This finding suggested that the iodoacetamide effect was due to its action on platelets. However, it was also possible that since the rFXIII had not been pre-activated with thrombin before treatment with iodoacetamide, the active-site had not been properly alkylated. Activation of FXIII allows exposure of the previously hidden active site cysteine (Curtis et al. 191 A; Hornyak et al. 1989). Also, the ability of rFXIIIa, which had been pre-incubated with iodoacetamide, to crosslink fibrin did not appear to be inhibited. This was not expected since iodoacetamide acts as a potent inhibitor of FXIIIa activity through alkylation of the active-site cysteine (Reinhardt 1981). A n experiment ensuring the alkylation of pre-activated rFXIIIa before incubation with platelets would be necessary to determine whether iodoacetamide enhances FXIIIa binding to platelets via occupation of the FXIIIa active-site or through some activity on the platelet itself. If occupation of the FXIIIa active-site does enhance the binding of FXIIIa to platelets, the availability of FXIIIa substrate may regulate the binding of FXIIIa to platelets. The co-localization of both FXIIIa and its physiologically most recognized substrate, fibrinogen, to the surface of platelets may be important for the formation and stabilization of platelet-rich thrombi. The ability of FXIIIa to crosslink fibrin to the platelet surface was examined as a contributing mechanism for clot stabilization. If FXIIIa were crosslinking fibrin to the platelet surface, one would predict that less fibrin would come off the platelet surface in the presence of E D T A when incubated with FXIIIa. However, there was no 81 increase in the amount of irreversible fibrin binding to the platelet detected in the presence of rFXIIIa as opposed to without rFXIIIa. Fibrin does not appear to be chemically crosslinked by rFXIIIa to the platelet surface. If not covalently attached, the irreversible association of fibrin(ogen) with platelets could be achieved through multiple non-covalent associations between fibrin(ogen) and the platelet surface after the initial reversible binding step (Marguerie et al. 1980). This irreversible association of fibrin(ogen) to the platelet surface may be important in the force generation required for clot retraction. Western blot analysis of the Triton X-100 insoluble fibrin(ogen) originating from the platelet and remaining associated with the platelet, however, does indicate crosslinking of fibrin(ogen) into high molecular weight complexes; y-y dimer formation and a-polymer formation (complexes greater than 200 kDa) are readily detectable. This indicates that although not crosslinked to the platelet, fibrin(ogen) is nonetheless crosslinked by FXIIIa into a matrix of protein. This together with the irreversible nature of fibrin(ogen) binding to the platelet results in a protein scaffolding resistant to fibrinolysis strongly attached to the platelet surface. This may be one of the ways FXIIIa contributes to clot retraction. Since platelet aggregation is dependent on the bridging of platelets by fibrinogen, the strong interrelation observed between fibrin(ogen) and FXIIIa on the surface of thrombin-activated platelets, led to investigation of the effect of FXIIIa on the rate and extent of platelet aggregation. The presence of rFXIIIa led to a significant increase in the number of platelet aggregates detected via flow cytometry, albeit at low levels due to the fibrin-polymerizing inhibitor, G P R P . This was true whether the platelets had been treated with thrombin or not. Also, the rate of aggregation when allowing platelets to aggregate fully, measured via platelet 82 aggregometry, was mildly yet significantly higher in samples containing rFXIIIa than in samples not containing rFXIIIa. However, a trend indicating a slight increase in the final amount of aggregation did not attain significance. These data suggest that FXIIIa had the capacity not necessarily to increase the extent of platelet aggregation but rather to accelerate the process of aggregation. Considering that FXIIIa crosslinks the fibrin polymers that bridge platelets in aggregate formation, this is not a surprising finding. Stabilization of any individual moment in the aggregation process should promote the process. The amount of FXIIIa bound to single platelets also correlated positively with the percentage of small platelet aggregates found in the same sample. Perhaps platelet aggregate formation is enhanced not only by FXIIIa but specifically by the FXIIIa bound to platelets. Future investigation of FXIIIa contribution to aggregation could include aggregometry experiments in platelet-rich plasma (PRP) using anti-FXIII antibodies to inhibit FXIIIa binding and activity. Using this method, platelets would not have to be subjected to the stresses of the wash procedure and more agonist responses could be evaluated. The origin of most of the surface-bound FXIIIa appears to be the aqueous media surrounding the platelets. In the above-mentioned experiments, rFXIIIa included in the incubation mixtures at concentrations similar to that of plasma FXIII (21.6 mg/L - Yorifuji et al. 1988) resulted in most activated platelets binding FXIIIa at relatively high amounts. The contribution of endogenous platelet FXIII to the measurable surface accessible pool of FXIII was examined in the absence of exogenously added rFXIIIa and it was found that there was an 8% increase in the number platelets with FXIIIa on the surface of thrombin-activated compared with non-thrombin-treated platelets. Endogenous platelet FXIII was somehow 83 released from the platelet to bind back to the surface or was possibly expressed on the membrane through another unknown mechanism. This possible membrane location of FXIII is supported by FXIII measurements in subcellular fractions of platelets wherein 4% of FXIII activity is recovered in the membrane fraction of platelets (Lopaciuk et al. 1976). The FXIII on the platelet surface did not appear to be originating from the cytoplasm of other inadvertently lysed platelets as L D H activity was not detectable in the supernatant of samples treated with even higher concentrations of thrombin. However, lysis of the platelets at levels below the sensitivity of the L D H assay can not be ruled out as a possible mechanism contributing to the available surface-associated FXIII . Other myeloid cells, including macrophages and monocytes, also release and express FXIII upon activation (Kradin et al. 1987; Kloczko et al. 1995) suggesting that the FXIII release observed in these studies is related to a biological process and not merely an artifact of the experimental set-up. The mechanism for release/expression of FXIII is unknown. There is one report of the presence of FXIII in platelet a-granules (Marx et al. 1993) which could possibly account for FXIII being released upon thrombin stimulation and subsequent a-granule release. Alternatively, there could also be some FXIII from the plasma in the O C S of platelets, which was not fully removed during the centrifugation and wash steps, available for activation by thrombin and subsequent binding to the platelet. There is an interesting report that the cytosolic platelet protein, matrix metalloproteinase-2 (MMP-2) is translocated upon collagen stimulation of platelets to the platelet plasma membrane and subsequently released depending on the level of platelet aggregation (Sawicki et al. 1998). This work demonstrates that there appears to be a mechanism in place for the translocation of cytosolic proteins to the surface of platelets. 84 The translocation of platelet FXIII to the outside of platelets has many implications for the contribution of clot stabilization by platelets. 85 C H A P T E R 4 FXIII A N D T H E P L A T E L E T C Y T O S K E L E T O N 4.1 FXIII in the Cytoskeletal Fractions of Platelets During the cytoskeletal rearrangements that take place after activation of a platelet, certain soluble cytoplasmic proteins associate with the cytoskeleton. To determine whether FXIII associates with the platelet cytoskeleton during activation, different cytoskeletal fractions of platelets were examined for the presence of FXIII using Western blot technique. The low speed pellet (LSP) and high speed pellet (HSP) fractions corresponding roughly to the cytoskeletal fraction and the membrane skeletal fraction respectively, were obtained through differential centrifugation. FXIII was present in similar amounts in the L S P of platelets activated either with thrombin, calcium ionophore (A23187), or A D P after 30 minutes of activation (Figure 22A). FXIII was detected at equal levels in the L S P (panel B) and H S P (panel C) fractions of platelets kept untreated (Figure 22, lanes 4 & 5) or treated with thrombin (Figure 22, lanes 6 - 13). The amount of FXIII in these fractions did not vary whether platelets were incubated for 10 minutes (Figure 22B and C, lanes 6, 7, 10, and 11) or 30 minutes (Figure 22B and C, lanes 8, 9, 12, and 13). Neither did the amount of FXIII vary whether platelet aggregation was inhibited (Figure 22B and C, lanes 6-9) or allowed (Figure 22B and C, lanes 10-13). The presence of iodoacetamide also had no effect on the amount of FXIII detectable in the L S P and HSP of these platelets (Figure 22B and C, odd-numbered lanes versus even-numbered lanes). Apparently, the presence of FXIII in the L S P and HSP 86 A kDa 200-97 _ • » ( 9 v » 69-46-30-1 2 3 B C 4 5 6 7 8 9 10 11 12 13 4 5 6 7 8 9 10 11 12 13 Figure 22. FXIII in the low speed pellet and high speed pellet of platelets at 10 and 30 minute incubations. Panel A corresponds to the L S P of washed platelets incubated with one of three different agonists for 30 minutes: 10 U/ml thrombin (lane 1), 19.1 p M A23187 (lane 2), or 10 p M A D P (lane 3). The L S P was subjected to S D S - P A G E on 7.5% acrylamide resolving gels and Western blot analysis was performed using rabbit anti-FXIII and peroxidase-conjugated donkey anti-rabbit as the secondary antibody. In panels B and C platelets were pre-incubated with iodoacetamide (odd-numbered lanes) or not (even-numbered lanes) and were either left untreated by thrombin (lanes 4 and 5), thrombin-activated in the presence of G P R P for 10 minutes (lanes 6 and 7), thrombin-activated in the presence of G P R P for 30 minutes (lanes 8 and 9), thrombin-activated with stirring for 10 minutes (lanes 10 and 11), or thrombin-activated with stirring for 30 minutes (lanes 12 and 13). The L S P (B) and H S P (C) fractions from these platelets were subjected to S D S - P A G E on 7.5% acrylamide resolving gels and Western blot analysis was performed using goat anti-FXIII as the primary antibody and peroxidase-conjugated rabbit anti-goat as the secondary antibody. 8 7 fractions of platelets was not affected by varying activation of the platelets. FXIII was not detectable in the supernatant fractions of Triton X-100 lysates (data not shown). However, at 23/4 hours of activation FXIII was seen to be enriched in the H S P of activated compared to resting platelets in the presence of E D T A (Figure 23 A and B , lanes 1 and 2, left half). This increase of FXIII in the HSP also appeared in platelets incubated with thrombin in the presence of calcium ions with or without exogenous rFXIII (Figure 23A and B , lanes 3 and 4, left half). The amount of FXIII in the HSP correlated with the relative amount of actin in each sample (Figure 23C, left half) suggesting a possible association between cytoplasmic FXIII and the membrane skeleton of platelets. A t this relatively long incubation time endogenous platelet FXIII was not detectable in the L S P (Figure 23, right half) in contrast to experiments carried out with thrombin incubation times of 30 minutes or less (Figure 22 and Figure 33) suggesting that the association of FXIII with the cytoskeletal fraction is transitory. The FXIII band in lane 3 of the L S P fractions of Figure 23B is most probably rFXIIIa as it migrated at the position for the thrombin-cleaved form of FXIII (FXIIIa). The protein staining pattern of the H S P of platelets indicated that the amount of protein sedimenting in the H S P decreased slightly between 0 and 3 hours of incubation in 15 m M E D T A (Figure 24A). In the presence of 10 U / m l thrombin and 10 m M C a C b the amount of protein sedimenting in the HSP decreased substantially with time. In contrast, and similarly to what was seen in Figure 23, in the presence of both 10 U / m l thrombin and 15 m M E D T A , the amount of protein sedimenting in the HSP increased at 3 hours of incubation. 88 Figure 23. FXIII in the high speed pellet and low speed pellet of thrombin-activated platelets at 23A hours. Washed platelets were incubated with 14 m M E D T A (lane 1), 10 U / m l thrombin and 14 m M E D T A (lane 2), 10 U / m l thrombin, 10 m M C a C l 2 , and 305 n M rFXIII (lane 3), or 10 U/ml thrombin and 10 m M C a C l 2 (lane 4). After 2 % hours, the thrombin was inactivated with the addition of 10 p M P P A C K . The platelets were washed and lysed as in "Materials and Methods" excepting that the lysis buffer was buffered with 5 m M phosphate instead of Tris and did not contain protease inhibitors. The HSP (left half of gel) and the L S P (right half of gel) were prepared as described under "Materials and Methods". A : 6% polyacrylamide S D S - P A G E gel silver stained for total protein. B : Western blot probed with rabbit anti-FXIII as the primary antibody and peroxidase-conjugated donkey anti-rabbit as the secondary antibody. C: Western blot probed with monoclonal anti-actin as the primary antibody and peroxidase-conjugated anti-mouse antibody as the secondary antibody. Molecular weight markers (kDa) are shown on the left. 89 B kDa 2 0 0 - , 116-97— 4 fibrin y-y dimer 1 — FXIII FXIIIa 6 6 -4 5 " \ fibrin(ogen) J H L X X J P ) monomers 1 2 3 4 1 4 3 2 HSP LSP 2 0 0 -9 7 - • - •*•»%« m t = F x m 6 9 -1 2 3 4 1 2 3 4 HSP LSP 2 0 0 -9 7 -6 9 -4 6 -3 g - ~~mm»m — actin 1 2 3 4 1 2 3 4 HSP LSP 90 Figure 24. Protein staining pattern in the high speed pellet of differentially incubated platelets. Washed platelets were incubated for various different times (0, 10 minutes, lhour, and 3 hours) with 15 m M E D T A (A), 15 m M E D T A and 10 U / m l thrombin (B), or 10 m M C a C l 2 and 10 U / m l thrombin (C). The platelets were lysed and the H S P was prepared as described in "Materials and Methods" excepting that the lysis buffer was buffered with 5 m M phosphate instead of T B S and did not contain protease inhibitors. The H S P was subjected to S D S - P A G E on 7.5% P A G E gels. Note the decreasing intensity o f protein staining with increasing time at the position of the arrows in panels A and C and the increasing intensity of protein staining at the arrow in panel B . 91 4.2 rFXIII, rFXIIIa and rFXIII'Ca Overlay Binding to Cytoskeletal Platelet Fractions Cytoskeletal and membrane skeletal fractions that were resolved in 7.5% S D S - P A G E gels and transferred to nitrocellulose membranes were overlaid with either inactive rFXIII or thrombin-activated rFXIIIa. While thrombin-activated rFXIIIa was seen to bind to many proteins in both the cytoskeletal and membrane skeletal fractions (Figure 25A), inactive rFXIII did not appear to bind very much to any of the proteins (Figure 25B). The amount of rFXIIIa binding to proteins in the cytoskeletal fractions was seen to increase as the platelet activation state increased from resting platelet L S P (lane 1), to activated platelet L S P (lane 2), to aggregated platelet L S P (lanes 3 and 4). There was a concurrent decrease in rFXIIIa binding to proteins in the membrane skeletal fraction from resting platelet H S P (lane 1), to activated platelet H S P (lane 2) to aggregated platelet HSP (lanes 3 and 4). This agrees with experiments that show a shift of cytoskeletal proteins from the H S P to the L S P as platelets are activated and then aggregated (Fox et al. 1993). Most of the rFXIIIa seen binding to proteins in the L S P was determined by Western blot to be bound to the fibrin(ogen) Act, a , and B monomers and the y-y dimers: bands with a M r of 74, 71, 55, and 100 kDa respectively (Figure 25 C). Much of the rFXIIIa seen binding to proteins in the H S P was identified as binding to actin by Western blot analysis (Figure 25D). Two other proteins migrating at relative mobilities greater than 200 kDa, one at ~ 180 kDa and one at 55 kDa also served as binding partners in the H S P for the rFXIIIa. Activation of FXIII through cleavage by thrombin enhances its ability to bind to immobilized platelet fibrinogen and cytoskeletal proteins. 92 Figure 25. rFXIIIa overlay binding to platelet cytoskeletal proteins. The L S P and H S P of platelets which were kept resting (lane 1), thrombin-activated for 10 minutes (lane 2), or thrombin-aggregated for 10 or 30 minutes (lanes 3 and 4, respectively) were separated with S D S - P A G E and transferred to nitrocellulose membranes. The membranes were overlaid with thrombin-activated rFXIIIa (A) or non-activated rFXIII (B). The membranes were then probed with rabbit anti-FXIII as the primary antibody and peroxidase-conjugated donkey anti-rabbit as the secondary antibody. Western blots of the stripped membranes were performed with rabbit anti-fibrinogen as the primary antibody and peroxidase-conjugated donkey anti-rabbit as the secondary antibody (C) or with rabbit anti-actin as the primary antibody and peroxidase-conjugated donkey anti-rabbit as the secondary antibody (D). The dark band at ~ 90 kDa visualized in panels A and B corresponds to detection of endogenous FXIII in the platelet cytoskeletal fractions. Molecular weight markers are shown on the left. 93 A B C 46 — 3 0 - •*»• «•>* ** 1 2 3 4 H S P 200 — 92.5. 69 — 46 — 30 — mmm 1 2 3 H S P 200 — H M W fibrin(ogen) complexes y-y dimer A a a Bp/p/y actin 1 2 3 4 L S P 1 2 3 4 L S P H M W fibrin(ogen) complexes D 92.5-69 46-30-2 3 H S P y-y dimer mm m * w Wm W W 1 2 3 4 L S P A a a Bp/p/y actin 94 In a similarly constructed experiment, the resting platelet membrane skeletal fraction proteins were resolved with S D S - P A G E , transferred to nitrocellulose membranes and overlaid with rFXIII»Ca 2 + . A s the calcium ion concentration in the incubation mixture was increased from 0 to 100 m M C a 2 + , the amount of rFXIII»Ca 2 + binding to proteins in the membrane skeletal fraction also increased (Figure 26, lanes 1-5). A t 100 m M C a 2 + the amount of rFXIII«Ca 2 + binding approached that seen with a thrombin-activated rFXIIIa standard (Figure 26, lanes 5 and 7). A t 1M C a 2 + the rFXIII«Ca 2 + binding increased non-specifically as the darkened background and the decrease in some of the band densities suggest. Increasing the salt concentration with N a C l instead of C a C b gave very different results. The ability of FXIII to bind to proteins in the membrane skeletal pellet at 2 m M C a 2 + was only modestly enhanced with the addition of N a C l in increasing concentration from 0 to 0 .6M N a C l (Figure 26, lanes 8-11). Above this N a C l concentration, fuzzy bands of lower intensity indicate the binding was impaired (Figure 26, lanes 12-14). A t no N a C l concentration tested did the binding of FXIII approach the amount of binding seen with high calcium ion concentrations indicating that this is not a simple ionic strength effect. High concentrations of calcium ions appear necessary to induce FXIII to bind to immobilized platelet cytoskeletal proteins. The dark band running at ~ 90 kDa, consistently seen in all the lanes, is the endogenous FXIII from the platelet fractions visualized through the inherent properties of the method. 95 rFXIIIa* Figure 26. FXIII'Ca overlay binding to platelet membrane skeletal proteins. The H S P of washed platelets was separated with S D S - P A G E and transferred to nitrocellulose membrane. The membrane was cut into strips and the strips were overlaid with FXIII* C a 2 + at increasing calcium ion concentrations (lanes 1-6: 0, 10"5 M , 1 m M , l O m M , 0.1 M , 1 M Ca 2 + ) , or at a constant 2 m M C a 2 + with increasing N a C l concentrations (lanes 8-14: 0, 0.2 M , 0.4 M , 0.6 M , 0.8 M , 1 M , 2 M NaCl) . The membranes were then probed with rabbit anti-FXIII as the primary antibody and peroxidase-conjugated donkey anti-rabbit as the secondary antibody. Lane 7 overlaid with pre-activated rFXIIIa served as a positive control. The dark band at ~ 90 kDa corresponds to detection of endogenous FXIII in the platelet cytoskeletal fractions. Molecular weight markers are shown on the left. 96 4.3 Intracellular Location of FXIII in Activated Platelets Intracellular FXIII is distributed homogeneously throughout the platelet cytoplasm (Sixma et al. 1984). Under a fluorescence microscope, specific antibody binding to FXIII was seen as a diffuse signal throughout the entire platelet (Figure 27, top). However, once platelets were activated with thrombin, the FXIII signal translocated to the periphery of the cell where the phase contrast micrographs showed thin ruffle-like extensions and/or pseudopod formations. This was seen to occur by one minute of activation. The signal started gradually diffusing back into the central cytoplasmic compartment by 5 minutes of activation and proceeded with a mostly diffuse pattern regained by 30 minutes (Figure 27). A t the one hour time-point the staining pattern appeared the same as that of time 0. Results obtained with a control FITC-conjugated non-specific antibody were negative. In the phase contrast micrographs of platelets prepared in this way, thin ruffles were visible only in the samples activated for 1 minute with platelets activated at longer time-points achieving a more compact shape, although still indicating pseudopods. Small platelet aggregates were evident. A similar transient movement of the FXIII was seen when thrombin-activated, fixed platelets were immobilized on poly-L-lysine coated slides for the washing and permeabilization steps. These results show that thrombin activation of platelets appears to result in a transient FXIII translocation to the periphery where rapid cytoskeletal changes are known to be in progress. 97 Figure 27. FXIII localization in thrombin-activated platelets at different activation time-points. Washed platelets supplemented with 2.5 m M C a 2 + were washed and separate incubations were activated with 1 U / m l thrombin in the presence of 35 u M G P R P . A t the end of the specified incubation times these were fixed, permeabilized and stained with 25 ug/ml of rabbit anti-FXIII antibodies. Phase contrast micrographs (1440X magnification) are shown on the left and corresponding fluorescence micrographs are shown on the right. Bar in first micrograph corresponds to 5 pm. These results are representative of three separate experiments. 98 time 0 1 min. 5 min. 10 min. 30 min. 1 hour 99 To determine whether FXIII translocation was a purely thrombin-specific response, the action of alternate agonists was also examined. The FXIII in platelets activated with calcium ionophore was also translocated to the cell periphery by one minute of activation and again diffused slowly back into the central cytoplasmic compartment with a homogeneously diffuse pattern observed by 30 minutes of activation (Figure 28). These pattern changes can be seen even though the calcium ionophore results in a drastic reduction in the platelet number as seen in the phase contrast micrographs. In contrast, with the exception of one or two platelets, the platelets activated with collagen did not show rim-pattern staining for FXIII (Figure 29) indicating that the FXIII was not as readily mobilized with this agonist. In the phase contrast micrographs of these platelets, it was apparent that the platelets did not change shape in the same manner as those activated with thrombin or calcium ionophore; they did not contract or form aggregates, although they did send out pseudopods. The same results were also obtained when a higher concentration of collagen (5 pg/ml) was used. 100 Figure 28. FXIII localization in calcium ionophore-activated platelets at different activation time-points. FXIII localization in calcium ionophore-activated platelets at different activation time-points. PRP was diluted with platelet-poor plasma to a final platelet concentration of 2 x 10 8 platelets/ml and separate incubations were activated with 20 pg/ml calcium ionophore (A23187). A t the end of the specified incubation times these were fixed, permeabilized and stained with FITC-anti-FXIII antibodies. Phase contrast micrographs (1440X magnification) are shown on the left and corresponding fluorescence micrographs are shown on the right. Bar in first micrograph corresponds to 5 pm. These results are representative of two separate experiments. 101 1 min. 5 min. 10 min. 30 min. 1 hour 102 Figure 29. FXIII localization in collagen-activated platelets at different activation time-points. PRP was diluted with PBS to a final platelet concentration of 2 x 10 8 platelets/ml and separate incubations were activated with 2 pg/ml collagen. A t the specified incubation times, these were fixed, permeabilized and stained with FITC-anti-FXIII antibodies. Phase contrast micrographs (1440X magnification) are shown on the left and corresponding fluorescence micrographs are shown on the right. Bar in first micrograph corresponds to 5 pm. These results are representative of two separate experiments. 103 4.4 Effect of Cytochalasin D on FXIII Translocation of Thrombin-Activated Platelets Cytochalasins inhibit actin polymerization in intact cells (Fox and Phillips 1981) and so they have been used to help dissect the contribution of cytoskeletal rearrangements in relation to other cellular events. Cytochalasin D was used here to understand the contribution of polymerizing actin in the FXIII translocation previously observed in thrombin-activated platelets. The phase contrast micrographs in Figure 30 show that cytochalasin D somewhat inhibited the shape change which is characterized by ruffling and pseudopod formation. The fluorescence micrographs show that cytochalasin D completely abolished the translocation of FXIII to the platelet periphery seen at 1 minute of thrombin activation. A dimethyl sulfoxide ( D M S O ) control, run in parallel to ensure that the inhibition was not due to the cytochalasin D carrier, showed no inhibition of FXIII translocation. Aliquots of the fixed platelets (before permeabilization) were washed and stained with FITC-conjugated anti-CD62P antibodies after washing and examined by flow cytometry to determine whether the cytochalasin D was affecting the platelet activation pathway defined by platelet release. The percentage of cells expressing the CD62P marker was approximately the same for platelets activated either in the presence or absence of cytochalasin D (Figure 31). Cytochalasin D while not affecting the platelet granule release mechanism inhibited FXIII translocation to the cell periphery. Thus, FXIII translocation required active polymerization of actin. 104 thrombin thrombin + cytochalasin D Figure 30. Effect of cytochalasin D on FXIII localization in thrombin-activated platelets. Washed platelets were pre-incubated with or without 10 ug/ml cytochalasin D or the same volume of D M S O for 5 minutes at 37°C. They were also pre-incubated with 35 u M G P R P i f they were to be activated. The platelets were activated with 1 U / m l thrombin for 1 minute at 37°C, fixed, permeabilized, stained with FITC-anti-FXIII antibodies and viewed under the fluorescent microscope at 1440X magnification. Bar in first micrograph corresponds to 5 pm. Similar results were obtained with 100 ug/ml cytochalasin B . 105 Q U u ,© 50 40 CU :! 3 0 a 20 cu — thrombin cytochalasin D DMSO + + + + Figure 31. CD62 expression of platelets treated with cytochalasin D. Washed platelets were pre-incubated without cytochalasin D , with 10 ug/ml cytochalasin D in D M S O , or with the same volume of D M S O for 5 minutes at 37°C. They were also pre-incubated with 35 u M G P R P i f they were to be thrombin-treated. The platelets were stimulated with 1 U / m l thrombin for 1 minute at 37°C, fixed and prepared for flow cytometry with PE-conjugated anti-CD62 antibodies. Values represent the average of duplicate measurements. 106 4.5 FXIII Activity in Platelet Lysate and Cytoskeletal Fractions Evidence that FXIII was associating with the platelet cytoskeleton during thrombin activation led to the question of whether this cytoskeletally-associated FXIII was also in its active configuration. It has been shown previously that FXIII activity in the lysate of activated platelets is greater than that of resting platelet lysate (Muszbek et al. 1995). Confirmation of these results using thrombin-activated platelets is demonstrated in Figure 32. When the FXIII activity in the cytoskeletal fractions of platelets was measured there was a significant increase seen in the FXIII activity associated with the cytoskeleton of thrombin-stimulated versus non-thrombin-stimulated platelets (Table 5). This activity was relatively small when compared with the total potential FXIII activity in 1 x 10 8 platelets (positive control). FXIII activity in the cytoskeleton of non-thrombin-activated platelets was undetectable, as the results were no different from those of background. The thrombin used to stimulate platelets was inactivated with P P A C K before platelet lysis. Thrombin that had been pre-incubated with P P A C K was added to the non-thrombin-stimulated platelets before platelet lysis to control for inadvertent activation of platelet FXIII by any residual thrombin activity. Platelet activation by thrombin was confirmed by measuring CD62P expression on the flow cytometer. 107 250 200 o "a 3 150 ? 100 a 50 non-thrombin-stimulated thrombin-s timulate d Figure 32. FXIII activity in resting and thrombin-activated platelet lysate. Platelets were incubated in the absence or the presence of thrombin. After 30 minutes, P P A C K was added to the sample stimulated with thrombin and thrombin which had been pre-incubated with P P A C K was added to the 'resting' sample. The lysate from both samples was prepared and tested for FXIII activity using a 1 4C-putrescine incorporation assay. These data represent 7 independent incubations, carried out on three separate days, normalized for platelet count (p < 0.05, one-tailed, paired t-test). 108 Table 5. FXIII activity in resting and thrombin-activated platelet cytoskeleton cpm/10 platelets Background 117 ± 19 Non-activated 112 db 16 Thrombin-activated 186 ± 13 * Positive control 72065 ± 2 1 4 3 1 The cytoskeleton isolated from platelets treated with or without 1 U / m l thrombin was tested for FXIII activity using a 1 4C-putrescine incorporation filter-paper assay. The FXIII activity available when all FXIII in the lysate of the same number of platelets is activated by thrombin is shown as the positive control. The non-activated control contains P P A C K -inactivated thrombin to control for possible residual thrombin activity in the platelet lysate. These results are representative of 3 separate experiments. * p < 0.05, thrombin-activated versus non-activated, one-tailed paired t-test. 109 4.6 Discussion Approximately half of the FXIII circulating in humans is located in the platelet cytoplasm and a role for this relatively large pool of FXIII has yet to be ascribed. Platelet FXIII is separate from the plasma FXIII pool that is involved in strengthening and stabilization of forming fibrin clots. In this study, interactions between platelet FXIII and platelet cytoskeletal proteins were investigated to ascertain whether FXIII crosslinking activity could be involved in stabilizing cytoskeletal protein interactions. Taking advantage of the insolubility of actin polymers in Triton X-100, the cytoskeletal and membrane skeletal fractions of platelets can be fractionated through differential centrifugation (Jennings et al. 1981; Fox 1985b). Depending on the nature of their associations in a detergent system, some soluble proteins that associate with the cytoskeleton also become sedimentable (Fox and Phillips 1981; Fox et al. 1993). FXIII was detected in both the cytoskeletal pellet and the membrane skeletal pellet of washed platelets treated for 10 or 30 minutes in the presence or absence of thrombin. FXIII was also detected in the cytoskeletal fraction of platelets activated for 30 minutes with calcium ionophore or A D P . This indicates that FXIII may associate with both the cytoskeleton and the membrane skeleton of platelets. Alternatively, since FXIII was not detected in the supernatant fractions of Triton X-100 lysates, it is also possible that the presence of FXIII in these fractions is due to an inherent insolubility of FXIII in Triton X-100-containing buffer. However, after a 23/4 hour incubation with thrombin endogenous platelet FXIII was no longer identified in the cytoskeletal fractions, only in the membrane skeletal fractions. These results suggest that FXIII may have been transiently associated with the cytoskeleton at between 10 and 30 110 minutes of activation and that this association was no longer present at 2 3 /4 hours. A transient association between soluble myosin and the cytoskeleton of thrombin-activated platelets has previously been identified with similar fractionation techniques (Fox and Phillips 1982). In that case, the transient association of myosin with the platelet cytoskeleton appears to be mediated through the phosphorylation of myosin and occurs at a shorter time frame with maximal association at around 1 minute of activation and dissociation achieved by around 2 minutes of activation. Washed platelets incubated at 37°C for 1 - 3 hours after centrifugation regain the discoid shape characteristic of resting platelets. During this resting procedure the amount of actin that sediments at low g-forces decreases, presumably due to the dissociation of actin networks formed during centrifugation (Fox et al. 1984). This delayed platelet morphology recovery may account for why FXIII is detectable in the cytoskeletal fraction of non-agonist stimulated platelets incubated with E D T A for 30 minutes yet not in the cytoskeletal fraction of platelets incubated with E D T A for almost 3 hours. Perhaps in the absence of platelet agonist addition the washing procedure imparts enough stimuli for FXIII association with the cytoskeleton. A t the prolonged incubation time (IV* hours) FXIII was detectable in the membrane skeletal fractions of both thrombin-treated and non-thrombin-treated platelets. There was a visible increase in the amount of FXIII detected in the membrane-skeletal fraction of thrombin-stimulated washed platelets compared to non-thrombin-stimulated platelets, which corresponded with an increase in the amount of detectable actin. The incubations were conducted in the presence of E D T A , and therefore the aggregation-dependent decrease in amount of protein normally seen in the membrane skeleton of thrombin-activated platelets (Fox et al. 1993) was not evident. In fact, in the presence of E D T A , the amount of protein detected in the membrane skeleton of thrombin 111 activated platelets increased at the 2 3/ 4 hour time-point. The corresponding increase of both actin and FXIII content in the membrane skeletal fraction of thrombin-stimulated platelets suggests that there may be an association between FXIII and actin and that the presence of both of these proteins in the platelet membrane skeleton depends on cytoskeletal rearrangements initiated with thrombin activation. Further characterization of the interactions between FXIII and platelet cytoskeletal or membrane skeletal proteins was obtained by examining the binding of rFXIII and rFXIIIa to proteins from these two fractions immobilized on nitrocellulose membranes. Activation of rFXIII by thrombin cleavage greatly enhanced the ability of rFXIII to bind to immobilized proteins of platelet cytoskeletal and membrane skeletal fractions. The two proteins to which FXIIIa bound the most were identified as fibrin(ogen) (both monomers and crosslinked complexes) and actin. Although fibrinogen is not a cytoskeletal protein, it has been shown to associate with the cytoskeleton (Tuszynski et al. 1984) and it does sediment with the platelet cytoskeletal proteins of activated platelets. Therefore, the results regarding FXIIIa binding to fibrin(ogen) in the cytoskeletal fraction wi l l be discussed in this chapter. The reduction of fibrin(ogen) did not compromise the ability of FXIIIa to bind to it. The large amount of rFXIIIa seen binding to fibrin(ogen) is not a surprising finding considering that fibrin(ogen) is the primary substrate of plasma FXIIIa. The minimal binding of rFXIII indicates that rFXIIIa binding to immobilized fibrin(ogen) is specific. Compared to the large amount of rFXIIIa binding, the minimal amount of binding seen with non-cleaved rFXIII to fibrin(ogen) agrees with a previous study where measurements made in a solid phase assay 112 indicate that the affinity of FXIII binding to microtiter plate-coated fibrinogen is 25-fold less than that of FXIIIa to fibrinogen (Achyuthan et al. 1996). Also , evidence for the binding of FXIIIa but not of FXIII to fibrin clots is found in a study by Procyk et al. (1993). In contrast, there have been studies that have detected the association of non-activated zymogen FXIII A-chains with immobilized fibrinogen (Greenberg and Shuman 1982); a previous study also utilizing the overlay technique with purified FXIII A-chains and immobilized fibrin(ogen) demonstrates the specific binding of non-cleaved FXIII to fibrin(ogen) (Mary et al. 1987). In this study the binding of rFXIII to fibrin(ogen) even though much less than that of rFXIIIa is still detectable. It is likely that both FXIII and FXIIIa bind to immobilized fibrin(ogen) but that since the affinity of FXIIIa is so much higher than that of FXIII (Achyuthan et al. 1996), when studies have compared the two, the binding of FXIII has been deemed to be insignificant. There is no detectable binding of FXIII A-chains to fluid phase fibrinogen (Siebenlist et al. 1996). The ability for FXIII/FXIIIa to bind only to immobilized fibrin(ogen) implies that a conformational change or a density change in fibrin(ogen) creates a binding site on fibrin(ogen) for FXIII/FXIIIa. Evidence supporting a conformational change in fibrinogen through immobilization comes from a study demonstrating that a monoclonal antibody which binds only to the D-dimer fibrin degradation product in solution wi l l bind to both purified fragment D monomers and D-dimer when immobilized on nitrocellulose membranes (Devine and Greenberg 1988). The epitope hidden in the soluble form of fibrin is expressed in crosslinked fibrin cleaved by plasmin or when fibrin is immobilized on nitrocellulose. The platelet surface may be the location for just such a conformational change (Greenberg and Shuman 1984). A s well , the increase in binding of FXIIIa over FXIII suggests that cleavage of FXIII is involved in creating a conformation 113 with a much higher affinity for fibrin(ogen). These differences may serve as regulatory mechanisms to localize enzymatically active FXIIIa to areas of active deposition of fibrin(ogen). The ability of rFXIIIa to bind to immobilized actin further supports the evidence that activated FXIII associates with the platelet membrane skeleton. Furthermore, the minimal binding of rFXIII to actin suggests that rFXIIIa binding to actin is specific and that the activation of the FXIII molecule imparts a conformation with greater affinity for binding to actin similar to that seen for fibrin(ogen). Purified platelet actin serves as a substrate for thrombin-activated FXIII (Cohen et al. 1980; Kahn and Cohen 1981) and there is also evidence for the platelet FXIII crosslinking of 1 4C-histamine to actin in calcium ionophore stimulated platelets over long incubation periods (Cohen et al. 1981). It is therefore possible that when FXIII is activated in the platelet cytoplasm through the activation of platelets, it takes on a conformation that allows it to bind to actin and through this localization, effects its crosslinking activity on cell cytoskeletal and membrane skeletal components. It is always possible that when proteins are immobilized onto nitrocellulose membranes, the act of binding to nitrocellulose may cause them to unfold somewhat or not refold properly, and in the process they could expose amino acids that would normally not be exposed. Although the binding of rFXIIIa to immobilized fibrin(ogen) and actin could be an artifact, non-cleaved rFXIII did not bind very well to the same proteins suggesting that the interaction between rFXIIIa and immobilized fibrin(ogen) and actin was not due only to the aberrant folding of these proteins. 114 The overlay experiments were repeated with rFXIII that had been pre-incubated with calcium ions at concentrations known to activate FXIII in vitro to determine if rFXII ICa 2 + could also bind to membrane skeletal proteins. Increasing concentrations of C a 2 + led to a substantial increase in the amount of FXIII binding to membrane skeletal proteins: in particular, two proteins at M r ~ 35 and ~ 46 kDa and a large protein of ~ 220 kDa. Increasing concentrations of NaCl however, did not result in very much of an increase in FXIII binding, 2+ indicating that the large increase in binding seen with increasing Ca concentrations was specific for calcium ions and was not just due to increasing salt concentration. Increasing NaCl concentration has also previously been shown to negatively affect the amount of FXIII binding to fibrinogen-latex beads (Greenberg and Shuman 1982). These overlay experiments indicate that in a manner similar to that of thrombin cleavage, 100 m M C a 2 + appears to induce a conformational change in the FXIII molecule that allows it to bind to immobilized proteins of the platelet membrane skeletal fraction. To a much lesser degree 0.6 M NaCl together with 2 mM C a 2 + allowed for some binding of FXIII A-chains. It is interesting that the incubation at 1M NaCl containing 2 mM Ca , conditions reported to activate FXIII A-chains (Polgar et al. 1990), resulted in much lower binding compared to incubation at 100 mM Ca 2 + , conditions which also activate FXIII A-chains. It appears that the proposed 9+ conformational change induced by Ca which allows for FXIII activity as well as FXIII binding is different from the proposed conformational change induced by NaCl which allows for FXIII crosslinking activity, but not for much binding. The binding seen at plasma concentrations of C a 2 + indicates that when released into the plasma from platelets, FXIII could gain the ability to bind to immobilized fibrinogen. There was no detectable FXIII 115 binding in the absence of calcium ions or at the intracellular calcium ion concentration of thrombin-activated platelets (10"5 M ) . For an association between cytoplasmic FXIII and membrane skeletal proteins to occur other cellular factors must be involved. Upon activation by thrombin or the calcium ionophore A23187, the platelet cytoplasmic distribution of FXIII rapidly changed from a diffuse distribution throughout the cell to one of FXIII enrichment at the periphery of the activated platelet. In effect, this change localized the cytoplasmic FXIII to one of the areas of the platelet undergoing dynamic cytoskeletal reorganization. The translocation of FXIII to the platelet periphery was a transient event, with most peripherally-associated FXIII seen at 1 minute of platelet stimulation and a gradual return of FXIII to the central area of the platelet seen after 5 minutes. The generally diffuse distribution of FXIII was regained after about 1 hour. The movement of FXIII to the platelet periphery was not as readily demonstrated in platelets activated with collagen. Morphologically these platelets extended pseudopodia but did not appear to contract indicating that the cytoskeleton had not undergone the same cytoskeletal changes as those platelets activated by thrombin or calcium ionophore. FXIII mobilization appears to require a specific stimulation signal; these differences may be important in regulating the desired platelet responses to different activation signals. Complete inhibition of FXIII translocation with the actin-polymerization inhibitor, cytochalasin D indicates that active actin polymerization was required for FXIII translocation to occur. In addition, since cytochalasin D treatment did not inhibit a-granule release, the translocation of FXIII appears not to be directly tied to a-granule release. A n interesting study has identified a specific interaction between heat shock protein 27 (HSP27) and FXIII (Zhu et al. 1994a). HSP27 is 116 phosphorylated upon platelet activation and rapidly associates with the cytoskeleton (Zhu et al. 1994b). They reported a diffuse distribution of FXIII throughout the cytoplasm changing so that FXIII is localized both to the platelet periphery and the very center of activated platelets corresponding with HSP27 location. The authors propose that HSP27 may be escorting FXIII to the platelet cytoskeleton. The translocation of FXIII to the platelet periphery during major cytoskeletal rearrangements suggests that i f activated, FXIII could be involved in stabilizing the cytoskeletal proteins and thereby stabilizing a forming clot from the inside of platelets. Perhaps the translocation of FXIII to the platelet periphery is also part of the process leading to expression of FXIII on the outside of thrombin-aggregated platelets. Even though a mechanism for allowing the protein to cross the platelet membrane has yet to be identified, a similar course of events has been identified in the translocation of platelet protein M M P - 2 from a random cytosolic distribution to the extracellular space (Sawicki et al. 1998). FXIII transglutaminase activity in both platelet lysate and cytoskeletal fractions increased significantly upon activation of platelets by thrombin. This confirms that platelet FXIII gains a measurable amount of transglutaminase activity in thrombin-activated platelets (Muszbek et al. 1995). Additionally, this activity associated with the platelet cytoskeleton, increasing the amount of evidence that indicates that activated FXIII associates with the cytoskeleton. The amount of measurable FXIIIa 0 activity associated with the cytoskeleton was on the order of 0.1% of the total potential FXIII activity in the same number of platelets. The amount of measurable FXII Ia 0 transglutaminase activity in platelet lysate samples was about 1 - 5%> of the total potential activity in general agreement with previously published results (Muszbek 117 et al. 1995). Therefore, about 10% of the FXIII transglutaminase activity generated through platelet activation was found in association with the platelet cytoskeleton. Although this appears to be quite a small percentage, the concentration of FXIII inside of platelets is high enough that this could be quite a significant amount of active transglutaminase activity, especially considering that it does not appear to be spread evenly throughout the platelet but rather localized to the area beneath the membrane (chapter 4.3). 118 CHAPTER 5 PROTEINS CROSSLINKED UPON PLATELET ACTIVATION 5.1 The Crosslinking of Platelet Fibrin(ogen) Upon stimulation of washed platelets with a supranormal concentration of thrombin (10 U/ml), in the absence of either plasma FXIII or exogenously added rFXIII , platelet-associated fibrin(ogen) was crosslinked in a manner very similar to that seen in the presence of exogenously added rFXIII and was inhibited by E D T A (Figure 33). In the presence of 305 n M rFXIII (lane 1) a ~ 100 kDa protein band immunoreactive with anti-fibrinogen and corresponding to the calculated relative mobility of the fibrin y-y dimer was present and the A a and a fibrin(ogen) monomers were not visible. The presence of high molecular weight protein bands (greater than 200 kDa) immunoreactive with anti-fibrinogen which was often visible in the blots but which did not show up very well in this particular blot suggests that these a-monomers have been crosslinked into higher molecular weight fibrin(ogen) species. In the presence of E D T A there was little crosslinking evident (lane 3) and the fibrin(ogen) A a , a , and B p subunits could be readily identified in both the silver stained gel and the fibrinogen Western blot. In the absence of both rFXIII and E D T A (lane 2) the presence of the y-y dimer band indicates that crosslinking similar to that seen in the presence of rFXIII occurred. However, the A a and a-chains were still visible, not having been crosslinked. 119 Figure 33. The effect of thrombin-activation on the form of platelet-associated fibrin(ogen). Washed platelets in a final volume of 1 ml were activated with 10 U / m l thrombin. The incubations contained 10 m M C a C b and 305 n M rFXIII (lane 1), 10 m M C a C l 2 (lane 2), or 14 m M E D T A (lane 3). After 10 minutes rotating at room temperature the thrombin was inactivated with 10 u M P P A C K and the platelets were washed three times and lysed. The L S P of each sample was prepared, washed, resuspended in reducing buffer and boiled for 15 minutes. The samples were resolved on 6% polyacrylamide S D S - P A G E gels and subjected to silver stain (A) or Western blot technique using rabbit anti-fibrinogen as the primary antibody (B) or rabbit anti-FXIII as the primary antibody (C) and peroxidase-conjugated anti-rabbit as the secondary antibody. These results are representative of 6 separate experiments. 120 A kDa 200 — 116. 66 45 — fibrin y-y dimer FXIII FXIIIa fibrinogen A a fibrin(ogen) BpVpVy B 200 — 97. fibrin y-y dimer 46 — 200 — 97. fibrinogen A a fibrin a fibrin(ogen) BpVpVy 1 2 3 < FXIII 4 — FXIIIa 46 — 1 2 3 121 To determine whether the fibrin(ogen) A a and a-chains were subject to the same crosslinking activity as that of the y-chains in the absence of exogenous FXIII , a fibrin(ogen) crosslinking time-course spanning 3 hours was performed (Figure 34). With increasing incubation time there was an increasing amount of high molecular weight complexes containing fibrin(ogen) that were so large they did not manage to enter the stacking gel (lanes 1-4). There was a corresponding decrease in the amount of fibrin(ogen) A a and a-chain monomers with the fibrinogen Aa-chain disappearing sooner than the fibrin a-chain. A t 3 hours in the absence of exogenous FXIII (lane 4) the crosslinked banding pattern began to strongly resemble the banding pattern of samples incubated for short times in the presence of rFXIII at plasma concentrations of FXIII (lane 6). This crosslinking activity was completely inhibited by iodoacetamide (lane 8). To determine whether the FXIII crosslinking activity was only present in platelets activated with thrombin, experiments using calcium ionophore and A D P as alternate agonists were performed. There were some faintly discernible high molecular weight polymers formed when platelets were activated with calcium ionophore that were not detectable when identically activated platelets were pre-incubated with the FXIII activity inhibitor, iodoacetamide (Figure 35, lanes 3 and 4). These high molecular weight fibrinogen crosslinked species were similar to those found in thrombin-activated platelets (lane 1). Using A D P as the platelet agonist, no high molecular weight fibrinogen polymers were detected (data not shown). 122 loading wells gel interface hL IL Nk } high molecular weight fibrin(ogen) containing complexes fibrin y-y dimer fibrinogen A a fibrin oc fibrin(ogen) Bp/p fibrin y 1 2 3 4 5 6 7 8 Figure 34 . Crosslinking time-course for fibrin(ogen) on thrombin-activated platelets both in the presence and the absence of exogenous FXIII. Washed platelets at a final volume of 1 ml were activated with 10 U/ml thrombin in the presence of 10 mM CaC^ and set to incubate on a rotator. After the incubation the platelets were lysed and the LSP of equal numbers of platelets was prepared. The LSP was then subjected to Western blot analysis using rabbit anti-fibrinogen as the primary antibody and peroxidase-conjugated goat anti-rabbit as the secondary antibody. Lanes 1 - 4 correspond to platelet incubations performed in the absence of 305 nM rFXIII and stopped at times 0, 10 minutes, 1 lA hours, and 3 hours, respectively. Lanes 5 - 7 correspond to platelet incubations performed in the presence of rFXIII and stopped at times 0, 10 minutes, and 3 hours, respectively. Lane 8 corresponds to platelets incubated in the presence of both rFXIII and 14 mM iodoacetamide and stopped at the 3 hour time point. 123 crosslinked fibrin polymers 1 2 3 4 Figure 3 5 . Fibrinogen crosslinking in calcium ionophore activated platelets. Washed platelets in a final volume of 1 ml were pre-incubated with (even-numbered lanes) or without (odd-numbered lanes) 270 p M iodoacetamide for 10 minutes at 37°C. They were then activated with 10 U/ml thrombin (lanes 1 and 2) or with 19 p M calcium ionophore (A23187) (lanes 3 and 4) for 30 minutes at room temperature with stirring. Platelets were washed once and the LSP was prepared. The samples were resolved on 7.5% polyacrylamide SDS-PAGE gels and subjected to Western blot technique using rabbit anti-fibrinogen as the primary antibody and peroxidase-conjugated anti-rabbit as the secondary antibody. Molecular weight markers (kDa) are depicted on the left. 124 To determine whether the FXIIIa-like crosslinking activity was originating from release and consequent activation of FXIII through platelet lysis, the question of platelet lysis was investigated. Lactate dehydrogenase (LDH) is an enzyme found in the cytoplasm of platelets and was used as a marker of platelet lysis. Uti l izing the same incubation conditions as those used in the above crosslinking experiments, platelet lysis was not detected, as there was no measurable L D H activity in the reaction supernatants. The L D H assay used was capable of measuring the lysis of as low as 2 x 105 platelets/100 pi , which corresponds to 1% of the total number of platelets in the incubation mixture. If there was platelet lysis during the experiment, it would have been on the order of less than 1%. Other proteins found on the outside of platelets were examined to see i f they were also being crosslinked upon platelet activation by thrombin in the presence of rFXIIIa. There was no evidence for crosslinking of a.2-antiplasmin, GPIIb/IIIa, GPIba or FXIII /FXIIIa itself detected in similar experiments (data not shown). 5.2 Cytoskeletal Proteins Crosslinked in Thrombin-Activated Platelets Cohen et al. (1985) noticed that many platelet proteins were crosslinked in platelets stored for long periods. Thrombin-activation of platelets also leads to the crosslinking of platelet proteins (Cohen et al. 1985; Harsfalvi et al. 1991). From my work with the platelet cytoskeletal fraction, it was suspected that some of these proteins might belong to the platelet cytoskeleton. To identify some of the proteins which are being crosslinked, platelets were treated with or without 10 U / m l thrombin in the presence or absence of the thiol ester inhibitor, iodoacetamide. Platelet activation by thrombin was confirmed by measuring 125 CD62P expression with flow cytometry. At the concentration used (1.35 mM) , iodoacetamide did not markedly affect the expression of this activation marker (Figure 36). These results for washed platelets were in agreement with the study looking at the effect of iodoacetamide on platelet CD62P expression in whole blood (Figure 15). The cytoskeleton of these platelets was analyzed via Western blot probed for various cytoskeletal proteins. Since the cytoskeletal fraction was solubilized with urea, immunoreactive bands migrating at a higher than expected apparent molecular weight are indicative that the protein in question has been chemically crosslinked. In Figure 37, utilizing anti-filamin antibodies, one can see that filamin is crosslinked to at least 4 high molecular weight species above the 200 kDa and 250 kDa filamin antibody reactive bands in all platelet samples which were activated with thrombin in the absence of iodoacetamide (lanes 3,5,7 and 9). Platelets pre-incubated with iodoacetamide (lanes 4, 6, 8 and 10) and non-activated platelets (lane 1) demonstrated a complete lack of filamin crosslinking. There was virtually no difference in crosslinking seen at the two different time points tested; the level of filamin crosslinking appeared the same after a 30 minute incubation as it did after only 10 minutes of incubation (lane 5 compared with lane 9 and lane 3 compared with lane 7). There did, however, seem to be an increase of filamin crosslinking in the samples that were allowed to aggregate (lanes 7 and 9). In these samples there was protein detectable in the loading wells, too big to even enter the stacking gel, which was not present in the non-aggregated samples. A s well , the amount of the filamin subunits was diminished in the aggregated samples compared with the non-aggregated samples, as presumably they are used up in becoming higher molecular weight species (lanes 7 and 9 compared with lanes 3 and 5). In summary, filamin was extensively crosslinked in platelets activated with thrombin and to a slightly greater degree i f those 126 cu V C — 0 a > "S3 o a. 100 80 60 40 v a "5. 20 0 iodoacetamide thrombin time + 30 minutes + + + 10 minutes + + + 30 minutes Figure 36. CD62P expression of iodoacetamide-treated thrombin-activated platelets. Washed platelets in MHTB were pre-treated with or without 1.35 mM iodoacetamide for 10 minutes at 37°C. The platelets to be activated were then incubated with 10 U/ml thrombin, in the presence of 10 mM CaCl 2 and 35 uM GPRP for 10 or 30 minutes. 10 uM PPACK was used to inactivate the thrombin and the samples were prepared for flow cytometry with FITC-conjugated anti-CD62P antibodies. 127 Figure 37. Western blot analysis of filamin content in the cytoskeleton of platelets activated with thrombin. Platelets pre-incubated with iodoacetamide (even-numbered lanes) or without iodoacetamide (odd-numbered lanes) were either left untreated by thrombin (lanes 1 and 2), thrombin-activated in the presence of G P R P for 10 minutes (lanes 3 and 4), thrombin-activated in the presence of G P R P for 30 minutes (lanes 5 and 6), thrombin-activated with stirring for 10 minutes (lanes 7 and 8), or thrombin-activated with stirring for 30 minutes (lanes 9 and 10). The L S P from these platelets was subjected to S D S - P A G E on 7.5% polyacrylamide resolving gels and Western blot analysis was performed using goat anti-filamin as the primary antibody and peroxidase-conjugated rabbit anti-goat as the secondary antibody. Molecular weight markers are indicated on the left and high molecular weight crosslinked complexes are indicated with arrows. 128 platelets were allowed to aggregate. This crosslinking was inhibited by the thiol ester inhibitor, iodoacetamide. In the same experiment, vinculin was also seen to be crosslinked in thrombin-activated platelets forming at least 4 high molecular weight complexes (155 kDa, 182 kDa, and two greater than 200 kDa) (Figure 38, lane 9). Again, iodoacetamide completely inhibited the formation of high molecular weight complexes involving vinculin (lanes 8 and 10). In the case of this protein, however, the detection of crosslinking was dependent on having allowed the platelets to aggregate, as the thrombin-activated platelets incubated in the presence of G P R P showed no signs of crosslinking (lanes 3 and 5). There was also more crosslinking detected in samples allowed to aggregate for 30 minutes as compared to those allowed to aggregate only for 10 minutes (lane 9 versus 7). In general, there was more vinculin detectable in the cytoskeleton of platelets activated in the presence of stirring than in the cytoskeleton of platelets activated in the presence of G P R P (lanes 7 through 10 versus lanes 3 through 6). This corresponds to a shift of vinculin from the membrane skeleton to the cytoskeleton as seen with other proteins whose associations with the cytoskeleton increase with increasing platelet activation and aggregation (Jennings et al. 1981). The very strong signal seen migrating at around 93 kDa of the aggregated platelet samples may indicate the association of a vinculin cleavage product with the cytoskeleton dependent on aggregation. Other cytoskeletal proteins examined for the possibility of crosslinking, included tubulin, actin and spectrin. N o high molecular weight crosslinked species for these proteins were identified as a result of platelet activation under this set of conditions (data not shown). 129 Figure 38. Western blot analysis of vinculin content in the cytoskeleton of platelets activated with thrombin. Platelets pre-incubated with iodoacetamide (even-numbered lanes) or without iodoacetamide (odd-numbered lanes) were either left untreated by thrombin (lanes 1 and 2), thrombin-activated in the presence of G P R P for 10 minutes (lanes 3 and 4), thrombin-activated in the presence of G P R P for 30 minutes (lanes 5 and 6), thrombin-activated with stirring for 10 minutes (lanes 7 and 8), or thrombin-activated with stirring for 30 minutes (lanes 9 and 10). The L S P from these platelets was subjected to S D S - P A G E on 7.5% polyacrylamide resolving gels and Western blot analysis was performed using monoclonal anti-vinculin as the primary antibody and peroxidase-conjugated goat anti-mouse as the secondary antibody. Molecular weight markers are indicated on the left and high molecular weight crosslinked complexes are indicated with arrows. These results are representative of four separate experiments. 130 5.3 Intracellular Location of Vinculin in Thrombin-Activated Platelets A s FXIII was associated with the cytoskeletal fraction, demonstrated an increased activity in the cytoskeleton, and was previously reported to crosslink a membrane skeletal protein, vinculin, in vitro, (Asijee et al. 1988) the next question posed was whether vinculin may be found in the same cellular areas as FXIII . In Figure 39 the distribution of vinculin in resting platelets is seen throughout the platelet with punctate staining indicating a clustering of a few molecules of vinculin. After 1 minute of activation by thrombin, the intensity of the signal increases and is generally translocated to the cellular periphery. This peripheral staining was found in discrete groupings and correlates well with the formation of focal adhesions. This pattern is still visible at the 10 minute activation time-point although the signal intensity decreases some. Although vinculin possesses a non-homogeneous distribution unlike that seen for FXIII in the previous chapter (section 4.3), they are both found throughout the cytoplasm of resting platelets and they are both localized to the platelet edges of thrombin-activated platelets at the same time points. This places FXIII in the same general area as vinculin both in resting and thrombin-activated platelets. 131 10 min. Figure 3 9 . Vinculin localization in thrombin-activated platelets at different activation time-points. Washed platelets supplemented with 2.5 m M C a 2 + were washed and separate incubations were activated with 1 U/ml thrombin in the presence of 35 uM GPRP. At the end of the specified incubation times these were fixed, permeabilized and stained for vinculin using monoclonal anti-vinculin as the primary antibody and FITC-conjugated goat anti-mouse as the secondary antibody. Phase contrast micrographs (1440X magnification) are shown on the left and corresponding fluorescence micrographs are shown on the right. Bar in first micrograph corresponds to 5 pm. Similar results were obtained upon calcium ionophore activation of the platelets. 132 5.4 Effect of Aggregation on Vinculin Association with the Platelet Cytoskeleton A n interesting and novel observation was made whilst examining the possibility of vinculin crosslinking upon thrombin activation of platelets. In stirred platelet samples which had been allowed to aggregate but not in those samples which were activated and not stirred, a vinculin cleavage product with a relative mobility of approximately 95 kDa was associating with the cytoskeletal fraction (Figure 38 lanes 7 and 9 versus lanes 3 and 5). To corroborate these findings, a time course study was employed whereby platelet aggregation and vinculin cleavage were simultaneously monitored (Figure 40). Association of the vinculin cleavage fragment with the platelet cytoskeletal fraction was seen by 5 minutes of aggregation and was strongest at 10 minutes of aggregation. This corresponded to 46% and 54% platelet aggregation respectively, with the maximum aggregation of this sample at 56%>. B y 20 minutes the association was no longer visible. The amount of the 130 kDa band corresponding to the full-length vinculin molecule also increased corresponding to the degree of platelet aggregation up to 10 minutes and then decreased in amount at 20 minutes. This increase was mirrored by the 150 kDa band immunoreactive with anti-vinculin antibodies, which most probably corresponds to meta-vinculin (Asijee et al. 1990; Turner and Burridge 1989; Bruin et al. 1991). The sample that was not stirred failed to aggregate to a normal level, achieving only 12%> platelet aggregation by 20 minutes. In this sample there was no visible vinculin cleavage product associating with the cytoskeletal fraction at any time point. Similar results were obtained for a third sample which contained G P R P in the incubation mixture (results not shown). Vincul in association with the cytoskeleton of washed platelets was dependent on the presence of fibrinogen. Much more vinculin associated with the cytoskeleton of washed platelets i f the thrombin-incubation contained exogenously added 133 Figure 40. Association of a vinculin cleavage fragment with the cytoskeleton upon platelet aggregation. Washed platelets were activated with 10 U/ml thrombin in the presence of 10 m M CaCb.. The aggregation trace for platelets activated with stirring is shown in Panel A and the aggregation trace for platelets activated in the absence of any stirring is shown in Panel B . Aliquots of the platelets were removed at 45 seconds, 2 minutes, 5 minutes, 10 minutes, and 20 minutes and placed directly in ice-cold lysis buffer as described in materials and methods (2.13). The low speed pellet fraction was isolated for Western blot analysis with monoclonal anti-vinculin as the primary antibody and peroxidase-conjugated goat anti-mouse as the secondary antibody (Panel C). Note vinculin cleavage product appearing after 5 minutes in stirred but not unstirred samples. The lysate lane corresponds to the platelet lysate of a non-activated platelet sample. 134 A B 135 fibrinogen at 0.2 mg/ml (Figure 41, lane 5). Vincul in was present in relatively low concentrations in a thrombin-activated stirred sample but was not visible in thrombin-treated samples that were not stirred or at one-hour incubation time-point. To determine whether the vinculin cleavage fragment was present in resting cells or whether platelet activation was required for its formation, experiments using whole platelet lysate were conducted (Figure 42). The PRP sample failed to indicate the presence of the vinculin cleavage fragment, while all samples treated with thrombin did indicate the presence of at least two and possibly three vinculin cleavage fragments. The washed platelet sample, which was not treated with thrombin, lacked two of the three vinculin cleavage fragments but the smaller of the three fragments could be faintly detected. 136 200 -46 -1 2 3 4 5 6 7 35 p M GPRP - + - - - - -0.2 mg/ml fibrinogen - - - + - -10 U/ml thrombin + + + + + + stirring - - + + + + incubation time 10' 10' 10' 10' 10' 10' 1 hr platelet lysate Figure 41. Association of vinculin with the platelet low speed pellet. Washed platelets at a concentration of 2 x 108 platelets/ml in MHTB were incubated in various conditions at room temperature. The platelets were lysed and fractionated and the LSP was dissolved and separated with SDS-PAGE. Western blot analysis using monoclonal anti-vinculin as the primary antibody and peroxidase-conjugated anti-mouse as the secondary antibody was performed. Lane 8 does not correspond to the LSP but rather is a control for total platelet lysate vinculin. Molecular weight markers (kDa) are shown on the left. 137 200-92.5-46-30-7 8 35 u M GPPvP - + - - - -40 ug/ml fibrinogen - - - - - + 10 U/ml thrombin - + + + + + + stirring - - + + + + incubation time 10' 10' 10' 5' 10' 20' 10 PRP Figure 42. Vinculin cleavage in whole platelet lysate. Washed platelets at a concentration of 2 x 108 platelets/ml in M H T B were incubated in various conditions at room temperature. The platelets were pelleted, resuspended in urea sample buffer and separated with SDS-PAGE. Western blot analysis using monoclonal anti-vinculin as the primary antibody and peroxidase-conjugated anti-mouse as the secondary antibody was performed. Lane 1 corresponds to platelets pelleted from PRP and solubilized immediately after collection. Molecular weight markers (kDa) are shown on the left. 138 5.5 Discussion The y-chains of platelet-associated fibrin on thrombin-activated platelets were crosslinked into the y-y dimers characteristic of FXIII crosslinking activity in the absence of exogenously added rFXIIIa (Figure 33 A and B, lane 2). These results confirm the findings of Hevessy et al. (1996) and McDonagh and McDonagh, Jr. (1972). This endogenous crosslinking activity appeared to act either more specifically on the fibrin y-chain or was slower than when rFXIIIa had been added as evidenced by the relatively large amount of non-crosslinked Aa/a-chains in this sample. Platelet fibrinogen and platelet FXIII are located in different compartments, with fibrinogen located in the platelet a-granules (Pham et al. 1983) and FXIII in the cytoplasm (Day and Solum 1973; Broekman et al. 1975; Lopaciuk et al. 1976; Sixma et al. 1984) although there may also be some FXIII in the a-granules (Marx et al. 1993). The crosslinking of fibrin y-chains was consistent with the hypothesis that some amount of platelet FXIII moves from the inside to the outside of platelets upon platelet activation (chapter 3.5) so that it can be in the same location as fibrin(ogen) after thrombin-induced a-granule release. This translocation of FXIII does not appear to result from the non-specific release of FXIII through platelet lysis, as there was an absence of detectable platelet L D H activity. When the form of FXIII in these samples was examined by Western blot, there was no detectable amount of thrombin-cleaved FXIIIa in the absence of exogenously added rFXIII. The absence of thrombin-cleavage of FXIII, suggested two separate possibilities: either the concentration of FXIII released into the supernatant was too low for detection with Western blotting technique or platelet-associated FXIII was acting on fibrin(ogen) in a manner not accessible to or protected from thrombin. The identification of a small amount of iodoacetamide-inhibitable fibrinogen crosslinking, associated with platelets 139 activated by calcium ionophore in the absence of thrombin, strengthens the argument that thrombin-activation of released platelet FXIII is not the only possibility for the FXIII crosslinking activity seen acting on platelet fibrin(ogen). That the crosslinking of fibrinogen on platelets activated with calcium ionophore or A D P was only very faintly detectable or not detectable at all could be related to either the inability of these agonists to fully activate and release platelet FXIII or to the decrease in platelet bound fibrinogen and the lack of conversion of fibrinogen to fibrin (the preferred FXIII substrate) in the absence of thrombin. The movement of FXIII to the outside of platelets is supported by findings of Cohen et al. (1985) who were able to measure the release of 1 - 3% of FXIII from thrombin-activated platelets into the supernatant with an activity assay. As well , another platelet protein found in association with the surface of platelets, (X2-antiplasmin is crosslinked by endogenous platelet FXIII (Francis and Marder 1987; Reed et al. 1992). The only other known platelet transglutaminase, a calmodulin-dependent transglutaminase is not a likely candidate for this activity for a couple of reasons; it acts preferentially on fibrin(ogen) a-chains instead of the y-chains preferred by active FXIII and it is not inhibited by iodoacetamide as is active FXIII (Puszkin and Raghuraman 1985). Additional evidence that this crosslinking originates from platelet FXIII comes from studies showing that anti-FXIII antibodies inhibit fibrin and a2-antiplasmin crosslinking (Reed et al. 1992) and that FXIII-deficient platelets do not induce the crosslinking of fibrin or a 2-antiplasmin (Hevessy et al. 1996) or indicate any transglutaminase activity (Boda et al. 1989). The report for the specific release of another cytoplasmic protein, M M P - 2 concurrent with granule release (Sawicki et al. 1998) raises the question of whether cytoplasmic FXIII could be released from the platelet in a similar manner. 140 Platelet-associated fibrin a-chains were also crosslinked in the absence of exogenously added rFXIIIa but a longer incubation time (IV2 hours) was required for the a-chain crosslinking to be realized. The more rapid disappearance of the fibrinogen Aa-chain compared with the fibrin a-chain is most likely due to thrombin cleavage of the Aa-chains to form a-chains and not because the platelet crosslinking activity preferred the Aa-chain to the a-chain as a substrate. Platelet activation experiments without exogenously added rFXIIIa achieved a similar degree of fibrin(ogen) crosslinking as those incubations that did contain rFXIIIa but were much slower in achieving this. In the presence of rFXIIIa, the platelet-associated fibrin(ogen) was rapidly crosslinked into very high molecular weight complexes with the almost complete crosslinking of fibrin a-chains by 10 minutes of activation. The degree of crosslinking in the absence of exogenous rFXIII at the longest incubation time point examined (3 hours) did not quite achieve the degree of crosslinking in plasma concentrations of FXIII at 10 minutes but did demonstrate the same crosslinking pattern. Therefore, it is likely that the activity of endogenous platelet FXIII is the same as that of plasma FXIII but is released from platelets at a much lower concentration than that of FXII I in plasma and would consequently require longer incubation times to attain similar crosslinking. In this study there was no evidence for the crosslinking of other proteins associated with the activated platelet surface: 012-antiplasmin, GPIIb/IIIa, GPIba or FXIII itself. Similarly, a report of gel-filtered platelets incubated with the supernatant of thrombin-activated platelets that contained approximately 1 - 3% active FXIII demonstrated a lack of polymer formation (Cohen et al. 1985). However, a study in which washed platelets were stimulated with 141 thrombin was able to demonstrate ci2-antiplasmin crosslinking into the platelet clot (Reed et al. 1992). It is likely that they were able to demonstrate a2-antiplasmin crosslinking due to the longer thrombin-incubation times of 1 hour, a higher platelet concentration in the incubation (4 x 10 8 platelets/ml) and a possibly more sensitive detection method using radioactive antibodies for the Western blot analysis. Since a2-antiplasmin is located in the platelet a-granules (Gogstad et al. 1983), it is reasonable to expect that as an active FXIII substrate it could be crosslinked along with fibrin to the fibrin polymers. When attention was turned to the crosslinking activity inside the platelet, two platelet cytoskeletal proteins were identified as being crosslinked upon thrombin-activation of platelets: filamin and vinculin. The crosslinking of filamin upon platelet activation was previously considered when it was found that filamin was detected in the urea insoluble matrices purified from thrombin-stimulated but not from 'resting' platelets (Harsfalvi et al. 1991). Furthermore, filamin is located in the same areas of both resting and activated platelets as is FXIII . FXIII was found homogeneously distributed throughout resting platelets and was translocated to the platelet periphery upon activation (chapter 4.3). Similarly filamin, which is homogeneously dispersed throughout resting platelets, is also found enriched at the platelet periphery of minimally activated platelets and both submembraneously and in the pseudopods of thrombin-activated platelets (Sixma et al. 1989). Filamin was crosslinked into at least 4 high molecular weight polymers of thrombin-stimulated platelets whether aggregation was allowed through stirring or inhibited by G P R P . When aggregation was allowed, there was a substantial increase in the amount of filamin crosslinking as seen by the development of polymers too large even to enter the stacking gel 142 and a greater decrease in the amount of filamin monomers in aggregated samples relative to non-aggregated samples. The increase in filamin crosslinking seen with platelet aggregation could be due to changes in the orientation or clustering of filamin upon aggregation which make filamin a better FXIII substrate or perhaps an aggregation signaling change is responsible for the difference in crosslinked filamin content. Evidence for the identity of the crosslinking activity originating from active FXIII came from the inhibition of this activity by iodoacetamide. This filamin crosslinking could be involved in strengthening the attachment of the platelet membrane skeleton to the plasma membrane as filamin binds directly to the cytoplasmic tail of platelet glycoprotein Iba serving as a link to actin filaments (Andrews and Fox 1992; Okita et al. 1985; Fox 1985a). A s well, filamin bridges the connection between Pi-integrins and actin (Loo et al. 1998; Pfaff et al. 1998) and although this interaction has yet to been demonstrated in platelets, these cells do have pVintegrins including the collagen receptor (GPIalla, (X2P1), the fibronectin receptor (GPIc-IIa, CI5P1), and the laminin receptor ( V L A - 6 , ( X 6 p i ) . Filamin is also involved in facilitating the high angle branching of cortical actin filaments through a leaf spring-like hinge (Gorlin et al. 1990; Hartwig et al. 1980) and it can be speculated that crosslinking by active FXIII at these branch points could be important in maintaining and/or strengthening the structure of aggregating platelets. The postulated role that FXIII plays in strengthening the connection between membrane glycoproteins and the platelet cytoskeleton could be very important for allowing the generation of forces required for clot retraction. The iodoacetamide-inhibitable crosslinking of the focal adhesion molecule, vinculin into at least four high molecular weight complexes implicates FXIII in the strengthening of platelet 143 focal adhesions upon thrombin-activation of platelets. Aggregation seemed to play a more important role in allowing vinculin to become crosslinked compared to filamin, as vinculin crosslinking was not detected at all when aggregation was inhibited. Also , vinculin crosslinking occurred more slowly than filamin crosslinking. The crosslinking of vinculin increased substantially between the 10 and 30 minutes time-points tested whereas filamin crosslinking was already quite pronounced at 10 minutes and did not increase much by 30 minutes. Perhaps, aggregation is required to bring enough vinculin into the cytoskeleton for crosslinking or to get vinculin into the correct orientation with relation to other cytoskeletal proteins for the ability of FXIII to crosslink this protein. Vincul in redistribution from the membrane skeleton to the cytoskeleton does depend on platelet aggregation (Fox et al. 1993; Asijee et al. 1990). Alternately, since a vinculin cleavage product of approximately 90-95 kDa became associated with the cytoskeleton upon aggregation perhaps it is this cleavage product which is preferentially crosslinked by active endogenous platelet FXIII . Purified platelet vinculin can serve as the acyl donor substrate of FXIIIa but likely does not have a FXIIIa substrate lysine side-chain (Asijee et al. 1988). Since there were four high molecular weight polymers containing vinculin in the cytoskeleton of thrombin-activated platelets it could be that one or more vinculin or vinculin cleavage molecules are being crosslinked to up to four other cytoskeletally-associated platelet proteins but that in all probability vinculin is not crosslinked to itself. Vincul in binds to actin (Ruhnau and Wegner 1988) and to talin (Burridge and Mangeat 1984), another focal adhesion protein, which in turn binds to the cytoplasmic side of glycoproteins on the platelet membrane. These interactions create a nidus of attachment for filamentous actin bundles. It is easy to imagine that these focus points for cytoskeletal proteins could provide solid points for the platelet to 144 attach either to other platelets during aggregation or to the exposed subendothelium of a damaged blood vessel. This connection is proposed to be required for the ability of platelets to exert tension on surfaces (Nachmias and Golla 1991). Specifically, vinculin is thought to promote platelet spreading through its postulated role in stabilizing focal adhesions and by transferring mechanical stresses that drive cytoskeletal remodeling (Ezzell et al. 1997). The chemical crosslinking of vinculin would further strengthen these interactions which could be very important for maintaining the integrity of the focal contact during the process of clot retraction. Co-localization studies looking at the simultaneous location of both FXIII and vinculin were not performed, as the platelets were too small for the available confocal microscopy setup to obtain sharp enough images. However, when the location of vinculin was determined with fluorescence microscopy, vinculin was seen to be located in the same general areas as FXIII . The distribution of vinculin in resting platelets was found to be generally homogeneous throughout platelets. L ike FXIII , vinculin was also found at the periphery of thrombin-stimulated platelets although unlike FXIII it stained in a characteristic punctate pattern. This punctate pattern of staining for vinculin has previously been identified in spread platelets (Takubo et al. 1998; Nachmias and Golla 1991; Rosenfeld et al. 1985) and corresponds to patches of vinculin located at the ends of filamentous actin bundles (Nachmias and Golla 1991). While the staining patterns of vinculin and FXIII indicate that they did not co-localize to the exact same places, it is clear that with the similar location of the two proteins, crosslinking of vinculin by FXIII was not excluded. 145 A vinculin cleavage fragment with a relative mobility of approximately 90-95 kDa was found in association with the cytoskeleton of platelets that had aggregated, but not with the cytoskeleton of platelets incubated in the absence of stirring whose aggregation was not induced. There was also a corresponding increase in the total amount of vinculin and meta-vinculin associated with the cytoskeleton upon platelet aggregation. In the presence of exogenously added fibrinogen, vinculin association with the cytoskeleton was much more pronounced than when the only source of fibrinogen was through release of internal stores of fibrinogen. The increase in association of vinculin to the platelet cytoskeleton in the presence of exogenously added fibrinogen was probably due to a stronger aggregation reaction induced by the extra fibrinogen. Vincul in was either not detected in the cytoskeleton of non-thrombin-stimulated platelets or in some experiments was detected but at lower amounts than in those samples that were stimulated with thrombin suggesting that the association of vinculin with the platelet cytoskeleton was mediated by an activation signal. The 90-95 kDa cleavage fragment of vinculin does not appear to be present in resting platelets at all as it was not detected in the whole platelet lysate of platelets isolated from PRP which had undergone minimal handling implying that cleavage of vinculin requires platelet activation. Similarly, it was reported that there is no evidence of vinculin proteolysis in fresh platelets whereas those stored for 48 hours at 4°C do indicate the accumulation of proteolytic vinculin fragments (Reid et al. 1993). The total platelet lysate of washed non-thrombin-treated platelets showed very little vinculin cleavage but there was a very faint band visible at approximately 80 kDa suggesting that the platelet wash procedure managed to activate the platelets enough to induce a very small amount of vinculin cleavage. A l l the platelet samples stimulated with thrombin contained proteolyzed vinculin fragments. The 146 stirred samples did not contain more of the cleavage fragments than the non-stirred samples suggesting that the increased amount of 90-95 kDa fragments associated with the cytoskeleton of aggregated platelets was due to preferred binding of this fragment to the cytoskeleton and not to an increase in proteolysis of vinculin. The vinculin cleavage detected in the experiments was not due to protease action on the vinculin molecule after platelet lysis as the lysis buffer was prepared fresh on the day of experiment and contained E G T A , leupeptin, P M S F , and benzamidine. A very interesting paper comparing the vinculin 90 kDa head fragment with the Listeria monocytogenes protein, A c t A required for the movement of the bacterium through infected cells identified a docking sequence in vinculin similar to the one found in A c t A , designated as an actin-based motility-1 ( A B M - 1 ) sequence (Laine et al. 1997). A c t A serves to recruit the host cell motor proteins, and actin-filaments form into rocket tails that propel the bacteria through the host cell. As a mechanism for the motility of Shigella flexneri, which does not have its own A B M - 1 sequence, the recruitment of vinculin and the subsequent unmasking of its A B M - 1 sequence through proteolysis has been proposed (Laine et al. 1997). In view of the findings in this study of proteolysis of vinculin upon platelet activation and a detectable association of the vinculin 90-95 kDa cleavage fragment with the cytoskeleton upon platelet aggregation, the proposed function of vinculin proteolysis in the actin-based motility of Shigella becomes very interesting. The proteolysis of vinculin upon platelet activation and the association of the vinculin 90-95 kDa cleavage fragment with the cytoskeleton may be involved in initiating an actin-based motility complex important for the cytoskeletal rearrangements involved in platelet aggregation similar to that initiated by the trailing end of Shigella. Taken together the results of this study suggest a possible model for the role of vinculin in promoting cell spreading, with activation of platelets leading to the 147 proteolysis of vinculin and association with the cytoskeleton. The proteolysis of vinculin would unmask the A B M - 1 sequence allowing the formation of the actin-based motility complex and cell spreading. FXIII recruited to the platelet periphery would then crosslink vinculin and filamin to other cytoskeletal proteins and strengthen the integrin-cytoskeleton connection to withstand the contractile forces generated during the ensuing clot retraction. This suggests a second mechanism for the regulation of vinculin in its associations with the cytoskeleton to that elucidated by Gilmore and Burridge (1996) whereby vinculin binding activity to cytoskeletal proteins is regulated by phosphatidyl-inositol-4-5-bisphosphate unfolding of the folded vinculin protein. 148 C H A P T E R 6 S U M M A R Y 6.1 Summarizing Discussion This thesis has investigated some of the interactions between FXIII and platelets with the goal of gaining a greater understanding of the function of FXIII both on the outer surface of the platelet and inside the platelet. Platelet-surface-associated FXIII plays an important role in fibrin clot stabilization through the significant acceleration of fibrin a-chain polymer formation and a2-antiplasmin crosslinking into the fibrin polymers. Although endogenous FXIII is capable of crosslinking fibrin polymers through some unknown mechanism of expression on the platelet surface, it appears that at physiologic conditions, the contribution of endogenous platelet FXIII in clot stabilization is minimal. A role for this very significant concentration of intracellular FXIII has yet to be identified. The binding of FXIII to the platelet surface was affected by low level platelet activation, glycoprotein Ilb/IIIa uncoupling, and by alkylation with iodoacetamide. In examining the surface-expression of rFXIII on the platelet surface, it was discovered that the wash procedure alone was enough to stimulate the platelets to express FXIII-binding sites. The washed platelets incubated with plasma concentrations of rFXIIIa, bound just as much rFXIIIa as maximally activated thrombin-stimulated platelets. Since the binding of FXIIIa to the platelet surface is mediated by activated glycoprotein Ilb/IIIa, and glycoprotein Ilb/IIIa is known to be activated through shear stress, the binding of rFXIIIa to washed platelets may be due in part to shear-induced glycoprotein Ilb/IIIa receptor activation. This method of receptor 149 activation may be physiologically relevant under conditions where high shear rates could be experienced by platelets, such as in partial occlusions at areas of vascular lesions. Perhaps this mechanism contributes to the elevated levels of FXIIIa found on patients with peripheral vascular disease (Devine et al. 1993). FXIII detection on the platelet surface did not correlate with granule release either in the studies described in this thesis or in published clinical studies (Devine et al. 1993; Levin et al. 2000). The irreversible uncoupling of the glycoprotein Ilb/IIIa receptor led to a five-fold decrease in rFXIIIa binding although it was not completely reversible. This behavior is consistent with the hypothesis that FXIIIa binds to fibrinogen bound to glycoprotein Ilb/IIIa i f one considers that fibrinogen also displays only a partial reversibility of binding to the platelet. As well , there was a significant correlation (r = 0.89) between the amount of fibrinogen binding and the amount of FXIII binding measured. However, since ADP-stimulation of platelets induces the binding of fibrinogen but not of FXIII , for the hypothesis that FXIII-binding is mediated by fibrinogen to be true, there would have to be another signal in addition to the binding of fibrinogen to platelets in order for FXIIIa to bind. This other signal could be a conformational change in the fibrinogen molecule or perhaps the conformational change of an associated protein such as glycoprotein Ilb/IIIa. Alternatively, it could be that the induction of fibrinogen binding sites is independent of the induction of binding sites for FXIIIa but that in thrombin-stimulated platelets or in the minimal activation of platelets through washing, these signals are activated in parallel and are therefore closely related. 150 Blockage of the active-site cysteine of FXIIIa through alkylation by iodoacetamide led to an unexpected increase in the binding of FXIII to platelets. This finding indicated that the binding site of FXIII for platelets is separate from the active-site. Considering that other research suggests platelet-bound FXIIIa to be active, it is reasonable to expect the active-site to be available while FXIIIa is platelet-bound. Furthermore, since the occupation of the active-site seemed to result in a conformation of FXIIIa which made binding to platelets energetically more favorable, it is conceivable that the binding of FXIIIa to platelets could by a similar mechanism result in easier access to the active-site and enhanced enzyme activity. Although FXIII binding is closely correlated with fibrinogen binding on the surface of thrombin-stimulated platelets and although fibrinogen is an excellent substrate for FXIIIa, there was no evidence that fibrin(ogen) was being crosslinked to the cell surface. This suggests a lack of suitable or accessible FXIIIa substrates on the platelet surface. The suitability of fibrin(ogen) bound to the platelet surface as a substrate for FXIIIa was, however, not in question, as platelet-surface-bound fibrin(ogen) was readily crosslinked into high molecular weight fibrin polymers. With fibrin strands playing such a fundamental role in clot retraction, the stabilization of these strands could be an important mechanism in which platelet-associated FXIIIa contributes to the process of clot retraction. Since aggregation is dependent on fibrinogen binding to the glycoprotein Ilb/IIIa receptor, and since the amount of FXIII binding correlated closely with that of fibrinogen binding, the effect of FXIIIa on aggregation was assessed. rFXIIIa was able to slightly increase the rate 151 of aggregation although the final amount of aggregation did not change compared with aggregation studies performed in the absence of rFXIIIa. Endogenous FXIII was readily detected on the surface of thrombin-stimulated platelets. Surface-expressed endogenous FXIII was enzymatically active as judged by the crosslinking of platelet-surface-associated fibrin molecules. The crosslinking activity was the same as that of thrombin-activated rFXIIIa, but the appearance of crosslinked fibrin species was slower in the case of endogenous FXIII activity. This thesis describes four independent lines of evidence suggesting that cellular platelet FXIII associates with the cytoskeleton upon thrombin stimulation of platelets: (1) Platelet FXIII is detected in the cytoskeletal fraction of thrombin-stimulated but not resting platelets at long incubation times. (2) rFXIIIa but not rFXIII binds to immobilized proteins of the cytoskeletal fraction. (3) Intracellular FXIII is translocated to the platelet periphery of thrombin-stimulated platelets but not of resting platelets. (4) Transglutaminase activity is increased in thrombin-stimulated but not in resting platelet cytoskeletal fractions. Investigation of the interactions between FXIII and the platelet cytoskeleton with fractionation of the cytoskeleton followed by S D S - P A G E analysis, indicated that FXIII was associated with the platelet cytoskeleton of washed platelets at time points less than 30 minutes but not greater than 2 2A hours. A t a very long incubation time point, the amount of FXIII detected in the membrane skeletal fraction of thrombin-stimulated platelets was much greater than that detected in the non-thrombin-stimulated platelet cytoskeleton. The amount 152 of FXIII detected in the cytoskeletal fraction also correlated with the amount of actin detected in the same samples. The association of activated rFXIII but not of inactive rFXIII with immobilized cytoskeletal proteins supported the hypothesis that activated FXIII associates with the platelet cytoskeleton. The identities of two of the immobilized proteins from the cytoskeletal fraction that bound rFXIIIa were assigned as fibrinogen and actin based on immunoreactivity and relative mobility. Similar to the binding of rFXIIIa on immobilized cytoskeletal proteins, non-thrombin-cleaved FXIII treated with calcium ions was also able to bind to cytoskeletal proteins. This is an important finding as FXIII is activated inside of cells in the absence of proteolysis and must therefore be able to bind to any potential substrate in the absence of proteolysis as well . Maximal binding was seen at a calcium ion concentration of 0.1 M and no binding was seen at the calcium ion concentration of activated platelets (10~5 M ) . This finding suggests that either the nitrocellulose membrane does not present the ideal conformation of these proteins, that calcium ions can achieve localized concentrations far greater than the average, or that another cellular factor besides C a 2 + is involved. These results are consistent with studies demonstrating that an increase in cytoplasmic concentrations of calcium ions leads to crosslinked polymer formation (Cohen et al. 1981). Another line of evidence supporting the association of intracellular FXIII with the cytoskeleton of thrombin-stimulated platelets came from microscopy studies depicting the transient translocation of intracellular FXIII from a diffuse homogeneous distribution throughout the resting platelet to a pattern of distribution localized to the platelet periphery of 153 activated platelets. In platelets the major cytoskeletal activity occurred close to the cytoplasmic side of the plasma membrane. FXIII translocation occurred by 1 minute of thrombin stimulation and was no longer evident by 1 hour of stimulation. This translocation was dependent on the active polymerization of actin. Evidence that FXIII may be escorted to the platelet cytoskeleton by heat shock protein 27 (HSP27) exists (Zhu et al. 1994a). Finally, the measurable transglutaminase activity associated with the cytoskeleton of thrombin-stimulated platelets was significantly greater than the activity associated with the cytoskeleton of non-thrombin-treated platelets and represents on the order of 0.1% of the total potential intracellular FXIII activity. Therefore in activated platelets, not only did FXIII appear to have translocated to areas of cytoskeletal rearrangements and formed an association with the cytoskeleton, but this cytoskeletally-associated FXIII also appeared to be enzymatically active. Identification of possible intracellularly activated FXIII (FXIIIa°) substrates came in the form of two actin-binding proteins: filamin and vinculin. These proteins were crosslinked into high molecular weight protein polymers in an iodoacetamide-inhibitable manner upon thrombin stimulation of washed platelets. The crosslinking of these proteins, which are involved in focal adhesion complex formation, could be important for the stabilization of the focal adhesion. In conclusion, the binding of FXIIIa to the platelet surface can occur at early stages of platelet activation, is partially reversible, is closely related to fibrinogen binding in thrombin-154 activated platelets, is involved in the crosslinking of fibrin polymers attached to the platelet surface, and can slightly enhance rates of aggregation. A role for the large amount of cytoplasmic platelet FXIII in the stabilization of focal adhesion complexes can be proposed involving the crosslinking of filamin and vinculin. Platelets have a physically demanding role in hemostasis and as such, they require a dynamic mode of action. Through the mechanisms of adhesion, aggregation, spreading, granule release and clot retraction, platelets play a vital role in protecting organisms from blood loss through sites of injury on blood vessels. FXIII crosslinking contributes to the final stages of clot formation and eventually to clot retraction through clot stabilization. Platelets localize FXIII to the site of would repair and through crosslinking reactions both on the outside of platelets with adhesive proteins and on the inside of platelets with cytoskeletal proteins, FXIII may be enabling clot retraction by imparting the structural strength required for such a process. 6.2 Future Directions Results from the work in this thesis open up exciting possibilities for future studies. The observations in this thesis hint at the possibility of shear force being involved in the expression of FXIIIa binding sites on platelets. Platelets could be subjected to controlled shear conditions mimicking those conditions found in the vasculature to ascertain whether shear is indeed a factor in the expression of platelet FXIIIa binding sites. Also, the identification of two intracellular cytoskeletal proteins as FXII Ia 0 substrates when platelets become activated suggest that there may be other cytoskeletal proteins that could be 155 crosslinked by intracellular FXIIIa 0 . Candidate proteins include myosin, which has been identified as a substrate for FXIIIa in vitro, and talin, another focal adhesion protein that associates with vinculin. In addition, FXIII-deficient platelets could be used to further dissect the contribution of intracellular crosslinking to the process of clot retraction with clot retraction measurements. 156 R E F E R E N C E LIST Achyuthan, K . E . , T. C. Rowland, P. J. Birckbichler, K . N . Lee, P. D . Bishop, and A . M . Achyuthan. 1996. Hierarchies in the binding of human factor XIII , factor XHIa , and endothelial cell transglutaminase to human plasma fibrinogen, fibrin, and fibronectin. Mol Cell Biochem. 162: 43-49. Adany, R., A . Belkin, T. Vasilevskaya, and L . Muszbek. 1985. Identification of blood coagulation factor XIII in human peritoneal macrophages. Eur J Cell Biol 38: 171-173. Andrews, R. K . , and J. E . B . Fox. 1992. Identification of a region in the cytoplasmic domain of the platelet membrane glycoprotein Ib-IX complex that binds to purified actin-binding protein. J Biol Chem 267:18605-18611. Ariyoshi , H . , and E . W. Salzman. 1996. Association of localized Ca2+ gradients with redistribution of glycoprotein Ilb-IIIa and F-actin in activated human blood platelets. Arterioscler Thromb Vase Biol 16: 230-235. Asijee, G . M . , A . Sturk, T. Bruin, J. M . Wilkinson, and J. W . ten Cate. 1990. Vincul in is a permanent component of the membrane skeleton and is incorporated into the (re)organizing cytoskeleton upon platelet activation. Eur J Biochem 189: 131-136. Asijee, G . M . , L . Muszbek, J. Kappelmayer, J. Polgar, A . Horvath, and A . Sturk. 1988. Platelet vinculin: a substrate of activated factor XIII . Biochim Biophys Acta 954: 303-308. Asquith, R. S., M . S. Otterbura, J. H . Buchanan, M Cole, J. C. Fletcher, and K . L . Gardner. 1970. The identification of epsilon-N-(gamma-L-glutamyl)-L-lysine cross-links in native wool keratins. Biochim Biophys Acta 221: 342-348. Aster, R. H . 1966. Pooling of platelets in the spleen: role in the pathogenesis of "hypersplenic" thrombocytopenia. J Clin Invest 45: 645-657. Bailey, K . , F. R. Bettelheim, L . Lorand, and W . R. Middlebrook. 1951. Action of thrombin in the clotting of fibrinogen. Nature 167: 233. Bale, M . D. , L . G . Westrick, and D. F. Mosher. 1985. Incorporation of thrombospondin into fibrin clots. J Biol Chem 260: 7502-7508. 157 Barkan, G. , and A . Gaspar. 1923. Zur frage der reversibilitat der fibringerinnung. II. Biochem Z 139: 291-301. Barry, E . L . R. and D . F. Mosher. 1989. Factor Xllla-mediated cross-linking of fibronectin in fibroblast cell layers: Cross-linking of cellular and plasma fibronectin and of amino-terminal fibronectin fragments. J Biol Chem 264: 4179-4185. Becker, R. P., and P. P. H . De Bruyn. 1976. The transmural passage of blood cells into myeloid sinusoids and the entry of platelets into the sinusoidal circulation: a scanning electron microscopic investigation. Am JAnat 145: 183-205. Beckerle, M . C. , D . E . Mil ler , M . E . Bertagnolli, and S. J. Locke. 1989. Activation-dependent redistribution of the adhesion plaque protein, talin, in intact human platelets. J Cell Biol 109: 3333-3346. Bendixen, E. , W. Borth, and P. C. Harpel. 1993. Transglutaminases catalyze cross-linking of plasminogen to fibronectin and human endothelial cells. J Biol Chem 268: 21962-21967. Bishop, P. D. , D . C. Teller, R. A . Smith, G . W. Lasser, T. Gilbert, and R. L . Seale. 1990. Expression, purification, and characterization of human factor XIII in Saccharomyces cerevisiae. Biochemistry 29: 1861-1869. Board, P. G . , K . Pierce, and M . Coggan. 1990. Expression of functional coagulation factor XIII in Escherichia coli. Thromb Haemost 63: 235-240. Bockenstedt, P., J. McDonagh, and R. I. Handin. 1986. Binding and covalent cross-linking of purified von Willebrand factor to native monomeric collagen. J Clin Invest 78: 551-556. Boda, Z . , G . Pfliegler, L . Muszbek, A . Toth, R. Adany, J. Harsfalvi, a. Papp, I. Tornai, and K . Rak. 1989. Congenital factor XIII deficiency with multiple benign breast tumours and successful pregnancy with substitutive therapy. Haemostasis 19: 348-352. Bradford, M . M . 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248-254. Broekman, M . J., R. I. Handin, and P. Cohen. 1975. Distribution of fibrinogen, and platelet factors 4 and XIII in subcellular fractions of human platelets. Br J Haematol 31: 51-55. Broker, M . , and Bauml, O. 1989. New expression vectors for the fission yeast Schizosaccharomyces pombe. FEBSLett 248: 105-110. 158 Bruin, T., G . M . Asijee, A . Prins, J. W. ten Cate, and A . Sturk. 1991. Subcellular distribution and phosphorylation of vinculin isoforms in human blood platelets. Thromb Haemost 65: 206-211. Buluk, K . 1955. A n unknown function of blood platelets. Pol TygLek 10: 191-198. Burridge, K . , and P. Mangeat. 1984. A n interaction between vinculin and talin. Nature 308: 744-746. Carlsson, L . , F. Markey, I. Blikstad, T. Persson, and U . Lindberg. 1979. Reorganization of actin in platelets stimulated by thrombin as measured by the DNase I inhibition assay. Proc Natl Acad Sci USA 76: 6376-6380. Carr, M . E . Jr., S. L . Carr, R. R. Hantgan, and J. Braaten. 1995. Glycoprotein Ilb/IIIa blockade inhibits platelet-mediated force development and reduces gel elastic modulus. Thromb Haemost 73: 499-505. Chen, J., M . Ishii, L . Wang, K . Ishii, and S. R. Coughlin. 1994. Thrombin receptor activation. Confirmation of the intramolecular tethered liganding hypothesis and discovery of an alternative intermolecular liganding mode. J Biol Chem 269: 16041-16045. Chen, R., and R. F. Doolittle. 1969. Identification of the polypeptide chains involved in the cross-linking of fibrin. Proc Natl Acad Sci USA 63: 420-427. Chung, S. I., M . S. Lewis, and J. E . Folk. 1974. Relationships of the catalytic properties of human plasma and platelet transglutaminases (activated blood coagulation factor XIII) to their subunit structures. J Biol Chem 249: 940-950. Cierniewski, C , T. Krajewski, and J. Klimczak. 1975. Comparative studies of fibrin polypeptide composition of two avian families: Sylvidae and Hirundinidae. Thromb Res 7: 297-303. Cohen, I., D . L . Burk, and J. G . White. 1989. The effect of peptides and monoclonal antibodies that bind to platelet glycoprotein Ilb-IIIa complex on the development of clot tension. Blood 73: 1880-1887. Cohen, I., C. T. L i m , D . R. Kahn, R. Glaser, J. M . Gerrard, and J. G . White. 1985. Disulfide-linked and transglutaminase-catalyzed protein assemblies in platelets. Blood 66: 143-151. Cohen, I., J. M . Gerrard, and J. G . White. 1982. Ultrastructure of clots during isometric contraction. J Cell Biol 93: 775-782. Cohen, I., T. Glaser, A . Veis, and J. Bruner-Lorand. 1981. Ca 2 +-dependent cross-linking processes in human platelets. Biochim Biophys Acta 676: 137-147. 159 Cohen, I., T. A . Blankenberg, D. Borden, D . R. Kahn, and A . Veis. 1980. Factor X l l l a -catalyzed cross-linking of platelet and muscle actin: regulation by nucleotides. Biochim Biophys Acta 628: 365-375. Cohen, I., L . Young-Bandala, T. A . Blankenberg, G . E . Siefring, Jr., and J Bruner-Lorand. 1979. Fibrinoligase-catalyzed cross-linking of myosin from platelet and skeletal muscle. Arch Biochem Biophys 192: 100-111. Corbett, S. A . , L . Lee, C. L . Wilson, and J. E . Schwarzbauer. 1997. Covalent cross-linking of fibronectin to fibrin is required for maximal cell adhesion to a fibronectin-fibrin matrix. J Biol Chem 272: 24999-25005. Cox, A . D. , and D . V . Devine. 1994. Factor X H I a binding to activated platelets is mediated through activation of glycoprotein Ilb-IIIa. Blood 83: 1006-1016. Credo, R. B . , C. G . Curtis, and L . Lorand. 1978. Ca 2 +-related regulatory function of fibrinogen. Proc Natl Acad Sci USA 75: 4234-4237. Curtis, C. G . , K . L . Brown, R. B . Credo, R. A . Domanik, A . Gray, P. Stenberg, and L . Lorand. 1974. Calcium-dependent unmasking of active center cysteine during activation of fibrin stabilizing factor. Biochemistry 13: 3774-3780. Curtis, C. G . , P. Stenberg, C. H . J. Chou, A . Gray, K . L . Brown, and L . Lorand. 1973. Titration and subunit localization of active center cysteine in fibrinoligase (thrombin-activated fibrin stabilizing factor). Biochem Biophys Res Commun 52: 51-56. Cutler, L . , G . Rodan, and M . B . Feinstein. 1978. Cytochemical localization of adenylate cyclase and of calcium ion, magnesium ion-activated ATP-ases in the dense tubular system of human blood platelets. Biochim Biophys Acta 542: 357-371. Davey, M . G. , and E . F. Luscher. 1967. Actions of thrombin and other coagulant and proteolytic enzymes on blood platelets. Nature 216: 857-858. Day, H . J., and N . O. Solum. 1973. Fibrinogen associated with subcellular platelet particles. Scand J Haematol 10: 136-143. de Sauvage, F. J., K . Carver-Moore, S. M . Luoh, A . Ryan, M . Dowd, D . L . Eaton, and M . W. Moore. 1996. Physiological regulation of early and late stages of megakaryocytopoiesis by thrombopoietin. J Exp Med 183: 651-656. Devine, D . V . 1990. Novel markers for the detection of platelet activation. Transfus Med Rev 4:115-120. Devine, D . V . , and P. D . Bishop. 1996. Platelet-associated factor XIII in platelet activation, adhesion, and clot stabilization. Sem Thromb Hemost 22: 409-413. 160 Devine, D . V . , G . Andestad, D . Nugent, and C. Carter. 1993. Platelet-associated factor XIII as a marker of platelet activation in patients with peripheral vascular disease. Arterioscler Thromb 13: 857-862. Devine, D . V . , and C. S. Greenberg. 1988. Monoclonal antibody to fibrin D-dimer (DD-3B6) recognizes an epitope on the gamma-chain of fragment D . Am J Clin Path 89: 663-666. Doolittle, R. F., R. Chen, and F. Lau. 1971. Hybrid fibrin: proof of the intermolecular nature of - crosslinking units. Biochem Biophys Res Commun AA: 94-100. Duckert, F. , E . Jung and D . H . Shmerling. 1960. A hitherto undescribed congenital haemorrhagic diathesis probably due to fibrin stabilizing factor deficiency. Thromb Diath Haemorrh 5: 179-186. Dvilansky, A . , A . F. H . Britten, and A . G . Loewy. 1970. Factor XIII assay by an isotope method 1. Factor XIII (transamidase) in plasma, serum, leucocytes, erythrocytes, and platelets and evaluation of screening tests of clot solubility. Br J Haematol 18: 399-410. Ezzel l , R. M . , W . H . Goldmann, N . Wang, N . Parasharama, and D . E . Ingber. 1997. Vincul in promotes cell spreading by mechanically coupling integrins to the cytoskeleton. Exp Cell Res 231: 14-26. Fear, J. D . , P. Jackson, C. Gray, K . J. Miloszewski, and M . S. Losowsky. 1984. Localisation of factor XIII in human tissues using an immunoperoxidase technique. J Clin Pathol 37: 560-563. Fesus, L , J. L . Mestis, L . Muszbek, and V . E . Koteliansky. 1986. Transglutaminase-sensitive glutamine residues of human plasma fibronectin revealed by studying its proteolytic fragments. Eur J Biochem 154: 371-374. Folk, J. E . 1983. Mechanism and basis for specificity of transglutaminase-catalyzed e-(y-glutamyl) lysine bond formation. Adv Enzymol 54: 1-56. Folk, J. E . , and J. S. Finlayson. 1977. The e-(y-glutamyl)lysine crosslink and the catalytic role of transglutaminases. Adv Protein Chem 31: 1-133. Folk, J. E. , and P. W . Cole. 1966. Mechanism of action of guinea pig liver transglutaminase. I. Purification and properties of the enzyme: identification of a functional cysteine essential for activity. J Biol Chem 241: 5518-5525. Fox, J. E . B . 1985a. Identification of actin-binding protein as the protein linking the membrane skeleton to glycoproteins on platelet plasma membranes. J Biol Chem 260: 11970-11977. 161 Fox, J. E . B . 1985b. Linkage of a membrane skeleton to integral membrane glycoproteins in human platelets. J Clin Invest 16: 1673-1683. Fox, J. E . B . , L . Lipfert, E . A . Clark, C. C. Reynolds, C. D . Austin, and J. S. Brugge. 1993. On the role of the platelet membrane skeleton in mediating signal transduction. J Biol Chem. 34: 25973-25984. Fox, J. E . B . , J. K . Boyles, M . C. Berndt, P. K . Steffen, and L . K . Anderson. 1988. Identification of a membrane skeleton in platelets. J Cell Biol 106: 1525-1538. Fox, J. E . B . , J. K . Boyles, C. C. Reynolds, and D . R. Phillips. 1984. Act in filament content and organization in unstimulated platelets. J Cell Biol 98: 1985-1991. Fox, J. E . B . , and D . R. Phillips. 1982. Role of phosphorylation in mediating the association of myosin with the cytoskeletal structures of human platelets. J Biol Chem 257: 4120-4126. Fox, J. E . B . , and D . R. Phillips. 1981. Inhibition of actin polymerization in blood platelets by cytochalasins. Nature 292: 650-652. Francis, C. W. , and V . J. Marder. 1987. Rapid formation of large molecular weight a-polymers in cross-linked fibrin induced by high factor XIII concentrations: Role of platelet factor XIII . J Clin Invest 80: 1459-1465. Francis, R. T., J. McDonagh, and K . G . Mann. 1986. Factor V is a substrate for the transamidase factor XHIa . J Biol Chem 261: 9787-9792. Fukue, H . , K . Anderson, P. McPhedran, L . Clyne, and J. McDonagh. 1992. A unique factor XIII inhibitor to a fibrin-binding site on factor XI I IA. Blood 19: 6365-74. Fuller, G . M . , and R. F. Doolittle. 1971. Studies of invertebrate fibrinogen. II. Transformation of lobster fibrinogen into fibrin. Biochemistry 10: 1311-1315. Gerrard, J. M . , J. G . White, G . H . Rao, and D. Townsend. 1976. Localization of platelet prostaglandin production in the platelet dense tubular system. Am J Path 83: 283-298. Gilmore, A . P. and K . Burridge. 1996. Regulation of vinculin binding to talin and actin by phosphatidylinositol-4-5-bisphosphate. Nature 381: 531-535. Gogstad, G . O., H . Stormorken, andN. O. Solum. 1983. Platelet alpha 2-antiplasmin is located in the platelet alpha-granules. Thromb Res 31: 387-390. Gorlin, J. B . , R. Yamin, S. Egan, M . Stewart, T. P. Stossel, D . J. Kwiatkowski, and J. H . Hartwig. 1990. Human endothelial actin-binding protein (ABP-280, nonmuscle filamin): a molecular leaf spring. J Cell Biol 111: 1089-1105. 162 Gorman, J. J., and J. E . Folk. 1984. Structural features of glutamine substrates for transglutaminases: Role of extended interactions in the specificity of human plasma factor X H I a and of the guinea pig liver enzyme. J Biol Chem 259: 9007-9010. Gorman, J. J., and J. E . Folk. 1981. Structural features of glutamine substrates for transglutaminases: Specificities of human plasma factor X H I a and the guinea pig liver enzyme toward synthetic peptides. J Biol Chem 256: 2712-2715. Gorman, J. J., and J. E . Folk. 1980. Structural features of glutamine substrates for human plasma factor X H I a (activated blood coagulation factor XIII). J Biol Chem 255: 419-427. Gralnick, H . R., S. Williams, L . McKeown, G . Connaghan, B . Shafer, K . Hansmann, M . V a i l , and J. Fenton. 1991. Endogenous platelet fibrinogen surface expression on activated platelets. J Lab Clin Med 118: 604-613. Greenberg, C. S., P. J. Birckbichler, and R. H . Rice. 1991. Transglutaminases: multifunctional cross-linking enzymes that stabilize tissues. FASEB J 5: 3071-3077. Greenberg, C. S., and M . A . Shuman. 1984. Specific binding of blood coagulation factor XI I I a to thrombin-stimulated platelets. J Biol Chem 259: 14721-14727. Greenberg, C. S., and M . A . Shuman. 1982. The zymogen forms of blood coagulation factor XIII bind specifically to fibrinogen. J Biol Chem 257: 6096-6101. Grundmann, U . , E . Amann, G . Zettlemeissl, and H . A . Kupper. 1986. Characterization of c D N A coding for human factor XIIIa. Proc Natl Acad Sci USA 83: 8024-8028. Hada, M . , M . Kaminski , P. Bockenstedt, and J. McDonagh. 1986. Covalent crosslinking of von Willebrand factor to fibrin. Blood 68: 95-101. Harding, H . W. , and G . E . Rogers. 1971. Epsilon-(gamma-glutamyl)lysine cross-linkage in citrulline-containing protein fractions from hair. Biochemistry 10: 624-630. Harris, H . 1981. Regulation of motile activity in platelets. In Platelets in biology and pathology, 2, ed. J. L . Gordon, 473-500. Amsterdam: Elsevier/North Holland Biomedical Press. Harris, H . E . , and A . G . Weeds. 1978. Platelet actin: sub-cellular distribution and association with profilin. FEBSLett 90: 84-88. Harsfalvi, J., L . Fesiis, E . Tarcsa, J. Laczko, and A . G . Loewy. 1991. The presence of a covalently cross-linked matrix in human platelets. Biochim Biophys Acta 1073: 268-274. 163 Hartwig, J. H . and M . DeSisto. 1991. The cytoskeleton of the resting human blood platelet: structure of the membrane skeleton and its attachment to actin filaments. J Cell Biol 112:407-425. . Hartwig, J. H . , J. Tyler, and T. P. Stossel. 1980. Actin-binding protein promotes the bipolar and perpendicular branching of actin filaments. J Cell Biol 87: 841-848. Henriksson, P., S. Becker, G . Lynch and J. McDonagh. 1985. Identification of intracellular factor XIII in human monocytes and macrophages. J Clin Invest 76: 528-534. Heukeshoven, J., and Dernick R. 1988. Improved silver staining procedure for fast staining in PhastSystem Development Unit. I. Staining of sodium dodecyl sulfate gels. Electrophoresis 9: 28-32. Hevessy, Z . , G . Haramura, Z . Boda, M . Udvardy, and L . Muszbek. 1996. Promotion of the crosslinking of fibrin and a,2-antiplasmin by platelets. Thromb Haemost 75: 161-167. Hoffman, R., E . J. Benz Jr., S. J. Shattil, B . Furie, H . J. Cohen, L . E . Silberstein, and P. McGlave, ed., 2000. Hematology: Basic principles and practice. 3 r d ed. Chapter 102, Molecular basis of blood coagulation, by Furie, B . and B . C. Furie. New York, Edinburgh, London, Phildelphia, and San Francisco: Churchill Livingstone. Holbrook, J. J., R. D. Cooke, and I. B . Kingston. 1973. The amino acid sequence around the reactive cysteine residue in human plasma factor XIII. Biochem J 135: 901-903. Holme, P. A . , U . Orvim, M . J. A . G . Hamers, N . O. Solum, F . R. Brosstad, R. M . Barstad, a n d K . S. Sakariassen. 1997. Shear-induced platelet activation and platelet microparticle formation at blood flow conditions as in arteries with a severe stenosis. Arterioscler Thromb Biol 17: 646-653. Hornyak, T. J., and J. A . Shafer. 1992. Interactions of factor XIII with fibrin as substrate and cofactor. Biochemistry 31 423-429. Hornyak, T. J., P. D . Bishop, and J. A . Shafer. 1989. a-Thrombin-catalyzed activation of human platelet factor XIII: relationship between proteolysis and factor X H I a activity. Biochemistry 28: 7326-7332. Ichinose, A . , and E . Davie. 1988. Characterization of the gene for the a subunit of human factor XIII (plasma transglutaminase), a blood coagulation factor. Proc Natl Acad Sci USA 85: 5829-5833. Ichinose, A . , L . E . Hendrickson, K . Fujikawa, and E . W. Davie. 1986a. Amino acid sequence of the a subunit of human factor XIII . Biochemistry 25: 6900-6906. 164 Ichinose, A . , B . A . McMul l en , K . Fujikawa, and E . W. Davie. 1986b. Amino acid sequence of the b subunit of human factor XIII, a protein composed often repetitive segments. Biochemistry 25: 4633-4638. Ikkala, E . 1972. Transfusion therapy in congenital deficiencies of plasma factor XIII. Ann NY Acad Sci 202: 200-203. Jagadeeswaran, P., and P. Haas. 1990. Synthesis of human coagulation factor XIII in yeast. Gene 86: 279-283. Janus, T. J., S. D . Lewis, L . Lorand, and J. A . Shafer. 1983. Promotion of thrombin-catalyzed activation of factor XIII by fibrinogen. Biochemistry 22: 6269-6272. Jennings, L . K . , J. E . B . Fox, H . H . Edwards, and D . R. Phillips. 1981. Changes in the cytoskeletal structure of human platelets following thrombin activation. J Biol Chem 256: 6927-6932. Jensen, P. H . , L . Lorand, P. Ebbesen, and J. Gliemann. 1993. Type-2 plasminogen-activator inhibitor is a substrate for trophoblast transglutaminase and factor XIIIa. Transglutaminase-catalyzed cross-linking to cellular and extracellular structures. Eur J Biochem 214: 141-146. Joist, J. H . , and S. Niewiarowski. 1973. Retention of platelet fibrin-stabilizing factor during the platelet release reaction and clot retraction. Thromb Diath Haemorrh 29: 679-684. Kaetsu, H . , T. Hashiguchi, D . Foster, and A . Ichinose. 1996. Expression and release of the a and b subunits for human coagulation factor XIII in baby hamster kidney ( B H K ) cells. J Biochem 119: 961-969. Kahn, D . R., and I. Cohen. 1981. Factor XHIa-catalyzed coupling of structural proteins. Biochim Biophys Acta 668: 490-494. Kahn, M . L . , M . Nakanishi-Matsui, M . J. Shapiro, H . Ishihara, and S. R. Coughlin. 1999. Protease-activated receptors 1 and 4 mediate activation of human platelets by thrombin. J Clin Invest 103: 879-887. Kahn, M . L . , Y . W. Zheng, W. Huang, V . Bigornia, D . Zeng, S. Moff, R. V . Farese Jr., C . Tarn, and S. R. Coughlin. 1998. A dual thrombin receptor system for platelet activation. Nature 394: 690-694. Kiesselbach, T. H . , and R. H . Wagner. 1966. Fibrin-stabilizing factor: a thrombin-labile platelet protein. Am J Physiol 211: 1472-1476. Kloczko, J., J. Giedrojc, P. Radziwon, M . Galar, and M . Bielawiec. 1995. Stimulated monocytes release factor XIII subunit A and fibronectin in vitro. Roczniki Akademii Medycznej w Bialymstoku 40: 408-413. 165 Kradin, R. L . , G . W. Lynch, J. T. Kurnick, M . Erikson, R. B . Colvin , and J. McDonagh. 1987. Factor XIII A is synthesized and expressed on the surface of U937 cells and alveolar macrophages. Blood 69: 778-785. Kreager, J. A . , D . V . Devine, and C. S. Greenberg. 1988. Cytofluorometric identification of plasmin-sensitive factor XHIa binding to platelets. Thromb Haemost 60: 88-93. K r o l l , J. 1989. The subunit composition of factor XIII proteins in normal and factor XIII deficient plasma and serum analysed by line immunoelectrophoresis. Clin Chim Acta 179: 279-284. Laemmli, U . K . 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685. Laine, R. O., W . Zeile, F. Kang, D . L . Purich, and F. S. Southwick. 1997. Vincul in proteolysis unmasks an A c t A homolog for actin-based Shigella motility. J Cell Biol 138: 1255-1264. Lak i , K . , and Lorand L . 1948. On the solubility of fibrin clots. Science 108: 280. Levin, E. , J. Wu, D. V . Devine, J. Alexander, C. Reichart, S. Sett, and M . Seear. 2000. Hemostatic parameters and platelet activation marker expression in cyanotic and acyanotic pediatric patients undergoing cardiac surgery in the presence of tranexamic acid. Thromb Haemost 83: 54-59. Lewis, S. D . , T. J. Janus, L . Lorand, and J. A . Shafer. 1985. Regulation of formation of factor XHIa by its fibrin substrates. Biochemistry 24: 6772-6777. Loewy, A . G. , and S. S. Matacic. 1979. Presence of the epsilon-(gamma-glutamic)lysine crosslink in cellular proteins. Biochim Biophys Acta 576: 263-268. Loewy, A . G . , S. Matacic, and J. H . Darnell. 1966. Transamidase activity o f the enzyme responsible for insoluble fibrin formation. Arch Biochem Biophys 113: 435-438. Loewy, A . G. , K . Dunathan, R. Kr ie l , and H . L . Wolfinger, Jr. 1961a. Fibrinase. I. Purification of substrate and enzyme. J Biol Chem 236: 2625-2633. Loewy, A . G . , A . Dahlberg, K . Dunathan, R. Kr i e l , and H . L . Wolfinger, Jr. 1961b. Fibrinase. II. Some physical properties. J Biol Chem 236: 2634-2643. Loewy, A . G . , K . Dunathan, J. A . Gallant, and B . Gardner. 1961c. Fibrinase. III. Some enzymatic properties. J Biol Chem 236: 2644-2647. Loewy, A . G . , J. A . Gallant, and K . Dunathan. 196Id. Fibrinase. I V . Effect on fibrin solubility. J Biol Chem 236: 2648-2655. 166 Loewy, A . G. , C . Veneziale, and M . Forman. 1957. Purification of the factor involved in the formation of urea-insoluble fibrin. Biochim Biophys Acta 26: 670-671. Loo, D. T., S. B . Kanner, and A . Aruffo. 1998. Filamin binds to the cytoplasmic domain of the Pi-integrin. J Biol Chem 273: 23304-23312. Lopaciuk, S., K . M . Lovette, J. McDonagh, H . Y . K . Chuang, and R. P. McDonagh. 1976. Subcellular distribution of fibrinogen and factor XIII in human blood platelets. Thromb Res 8: 453-465. Lorand, L . 1986. Activation of blood coagulation factor XIII. Ann NY Acad Sci 485: 144-158. Lorand, L . 1950. Fibrin clots. Nature 166: 694-695. Lorand, L . 1948. A study on the solubility of fibrin clots in urea. Acta Physiol Hung 1: 192-196. Lorand, L . , L . K . Campbell-Wilkes, and L . Cooperstein. 1972. A filter paper assay for transamidating enzymes using radioactive amine substrates. Anal Biochem 50: 623-631. Lorand, L . , and D . Chenoweth. 1969. Intramolecular localization of the acceptor cross-linking sites in fibrin. Proc Natl Acad Sci USA 63: 1247-1252. Lorand, L . , and H . H . Ong. 1966. Labeling of amine-acceptor cross-linking sites of fibrin by transpeptidation. Biochemistry 5: 1747-1753. Lorand, L . , and K . Konishi . 1964. Activation of the fibrin stabilizing factor of plasma by thrombin. Arch Biochem Biophys 105: 58-67. Lorand, L . , and A . Jacobsen. 1962. Accelerated lysis of blood clots. Nature 195: 911-912. Lorand, L . , K . Konishi , and A . Jacobsen. 1962. Transpeptidation mechanism in blood clotting. Nature 194: 1148-1149. Liischer, E . F. 1957. E i n fibrinstabilisierender factor aus thrombozyten. Schweiz Med WochenschrZl: 1220-1221. Marguerie, G . A . , T. S. Edgington, and E . F. Plow. 1980. Interaction of fibrinogen with its platelet receptor as part of a multistep reaction in ADP-induced platelet aggregation. J Biol Chem 255: 154-161. Martin, M . T., F. Lefebvre, M . Rabaud, and P. V . Graves. 1988. Biochemical study of adduct synthesis between fibrin monomers and elastin. Biomaterials 9: 519-524. 167 Marx, G . , G . Korner, X . M o u , and R. Gorodetsky. 1993. Packaging zinc, fibrinogen, and factor XIII in platelet a-granules. J Cell Physiol 156: 437-442. Mary, A . , K . E . Achyuthan, and C. S. Greenberg. 1987. Factor XIII binds to the Act- and Bp1-chains in the D-domain of fibrinogen: an immunoblotting study. Biochem Biophys Res Commun 147: 608-614. McAbee, D . D . , and F. Grinnell. 1982. Thiol-sensitive sites in cell adhesion. Biochem J208: 473-478. McDonagh, R. P., J. McDonagh, T. E . Petersen H . C. Thogersen, K . Skorstengaard, L . Sottrup-Jensen, S. Magnusson, A . Del l , and H. R. Morris. 1981. Amino acid sequence of the factor XHIa acceptor site in bovine plasma fibronectin. FEBS Lett 127: 174-178. McDonagh, J., and R. P. McDonagh, Jr. 1972. Factor XIII from human platelets: effect on fibrin crosslinking. Thromb Res 1: 147-160. McDonagh, R. P. Jr., J. McDonagh, and F. Duckert. 1971. The influence of fibrin crosslinking on the kinetics o f urokinase-induced clot lysis. Br J Haematol 21: 323-332. McDonagh, J., R. P. McDonagh Jr., J . - M . Delage, and R. H . Wagner. 1969. Factor XIII in human plasma and platelets. J Clin Invest 48: 940-945. McKee , P. A . , P. Mattock, and R. L . H i l l . 1970. Subunit structure of human fibrinogen, soluble fibrin, and cross-linked insoluble fibrin. Proc Natl Acad Sci USA 66: 738-744. Miloszewski, K . , and M . S. Losowsky. 1975. Factor XIII concentrate in the long term management of congenital factor XIII deficiency. Thromb Diath Haemorrh 34: 323-324. Mockros, L . F. , W . W. Roberts, and L . Lorand. 1974. Viscoelastic properties of ligation-inhibited fibrin clots. Biophys Chem 2: 164-169. Mosesson, M . W. , K . R. Siebenlist, D . L . Amrani, and J. P. DiOr io . 1989. Identification of covalently linked trimeric and tetrameric D domains in crosslinked fibrin. Proc Natl Acad Sci USA 86: 1113-1117. Mosher, D . F. 1984. Cross-linking of fibronectin to collagenous proteins. Mol Cell Biochem 58: 63-38. Mosher, D . F. 1976. Action of fibrin-stabilizing factor on cold-insoluble globulin and a2-macroglobulin in clotting plasma. J Biol Chem 251: 1639-1645. 168 Mosher, D . F. 1975. Cross-linking of cold-insoluble globulin by fibrin-stabilizing factor. J Biol Chem 250: 6614-6621. Mosher, D . F., and R. A . Proctor. 1980. Binding and factor XIII a-mediated cross-linking of a 27-dilodalton fragment of fibronectin to Staphylococcus aureus. Science 209: 927-929. Murtaugh, P. A . , J. E . Halver, and J. A . Gladner. 1973. Cross-linking of salmon fibrinogen and fibrin by factor XIII and transglutaminase. Biochem Biophys Res Commun 54: 849-855. Mustard, J. F., R. L . Kinlough-Rathbone, and M . A . Packham. 1989. Isolation of human platelets from plasma by centrifugation and washing. Methods Enzymol 169: 3-11. Muszbek, L . , V . C. Yee, and Z . Hevessy. 1999. Blood coagulation factor XIII : Structure and function. Thromb Res 94: 271-305. Muszbek, L . , G . Haramura, and J. Polgar. 1995. Transformation of cellular factor XIII into an active zymogen transglutaminase in thrombin-stimulated platelets. Thromb Haemost 73: 702-705. Muszbek, L . , J. Polgar, and Z . Boda. 1993. Platelet factor XIII becomes active without the release of activation peptide during platelet activation. Thromb Haemost 69: 282-285. Muszbek, L . , R. Adany, G . Szegedi, J. Polgar, and M . Kavai . 1985. Factor XIII of blood coagulation in human monocytes. Thromb Res. 37: 401-410. Nachman, R. L . , and A . J. Marcus. 1968. Immunological studies of proteins associated with the subcellular fractions o f thrombasthenic and afibrinogenaemic platelets. Br J Haematol 15: 181-189. Nachmias, V . T. and R. Golla. 1991. Vincul in in relation to stress fibers in spread platelets. CellMotil Cytoskeleton 20: 190-202. Naski, M . C , L . Lorand, and J. A . Shafer. 1991. Characterization o f the kinetic pathway for fibrin promotion of alpha-thrombin-catalyzed activation of plasma factor XIII. Biochemistry 30: 934-941. Nemeth, A . J., and N . S. Penneys. 1989. Factor XIIIa is expressed by fibroblasts in fibrovascular tumors. J Cutan Pathol 16: 266-271. Okita, J. R , D. Pidard, P. J. Newman, R. R. Montgomery, and T. J. Kunicki . 1985. On the association of glycoprotein Ib and actin-binding protein in human platelets. J Cell Biol 100:317'-321. 169 Osier, W. 1874. A n account of certain organisms occuring in the liquor sanguinis. Proc R Soc Lond5: 692-734. Ottaviani, P., and F. Mandelli . 1966. Two cases of haemorrhagic disease arising from fibrin stabilizing factor deficiency. Hemostase 6: 317-324. Peerschke, E . I. B . 1988. Irreversible platelet fibrinogen interactions occur independently of fibrinogen alpha chain degradation and are not mediated by intact platelet membrane glycoprotein Ilb-IIIa complexes. J Lab Clin Med 111: 84-92. Pfaff, M . , S. L i u , D . J. Erie, and M . H . Ginsberg. 1998. Integrin P cytoplasmic domains differentially bind to cytoskeletal proteins. J Biol Chem 273: 6104-6109. Pham, T. D . , K . L . Kaplan, and V . P. Butler, Jr. 1983. Immunoelectron microscopic localization of platelet factor 4 and fibrinogen in the granules of human platelets. J Histochem Cytochem 31: 905-910. Piotrowicz, R. S., R. P. Orchekowski, D . J. Nugent, K . Y . Yamada, and T. J. Kunicki . 1988. Glycoprotein Ic-IIa functions as an activation-independent fibronectin receptor on human platelets. J Cell Biol 106: 1359-1364. Pisano, J. J., T. J., Bronzert, M . P. Peyton, and J. S. Finlayson. 1972. Epsilon-(gamma-glutamyl) lysine cross-links: determination in fibrin from normal and factor XIII-deficient individuals. Ann NY Acad Sci 202: 98-113. Pisano, J. J., J. S. Finlayson, and M . P. Peyton. 1968. Cross-link in fibrin polymerized by factor XIII : s-(y-glutamyl)lysine. Science 160: 892-893. Polgar, J., V . Hidasi, and L . Muszbek. 1990. Non-proteolytic activation of cellular protransglutaminase (placenta macrophage Factor XIII). Biochem J 261: 557-560. Poon, M . C , J. A . Russell, S. Low, G . D . Sinclair, A . R. Jones, W. Blahey, B . A . Ruether, and D . I. Hoar. 1989. Hemopoietic origin of factor XIII A subunits in platelets, monocytes, and plasma. Evidence from bone marrow transplantation studies. J Clin Invest 84: 787-792. Prince, C. W. , D . Dickie, and C. L . Krumdieck. 1991. Osteopontin, a substrate for transglutaminase and factor XIII activity. Biochem Biophys Res Commun 111: 1205-1210. Procyk, R., P. D . Bishop, and B . Kudryk. 1993. Fibrin - recombinant human factor XIII A -subunit association. Thromb Res 71: 127-138. Procyk, R., and B . Blomback. 1988. Factor XHI-induced crosslinking in solutions of fibrinogen and fibronectin. Biochim Biophys Acta 967: 304-313. 170 Puszkin, E . G . and V . Raghuraman. 1985. Catalytic properties of a calmodulin-regulated transglutaminase from human platelet and chicken gizzard. J Biol Chem 260: 16012-16020. Reed, G . L . , G . R. Matsueda, and E . Haber. 1992. Platelet factor XIII increases the fibrinolytic resistance of platelet-rich clots by acceleration the crosslinking of a 2 -antiplasmin to fibrin. Thromb Haemost 68: 315-320. Reid, D . M . C. E . Jones, C. Y . Luo, andN. R. Shulman. 1993. Immunoglobulins from normal sera bind platelet vinculin and talin and their proteolytic fragments. Blood 81:745-751. Reinhardt, G . 1981. a-Halogenmethyl carbonyl compounds as very potent inhibitors of factor XIIIA in vitro. Ann NY Acad Sci 370: 836-842. Rinas, U . , B . Risse, R. Jaenicke, M . Broker, H . E . Karges, H . A . Kupper, and G . Zettlmeissl. 1990. Characterization of recombinant factor XIIIa produced in Saccharomyces cerevisiae. Biotechnology (NY) 8: 543-546. Robbins, K . C. 1944. A study on the conversion of fibrinogen to fibrin. Am J Physiol 142: 581-588. Rohlf, F. J., and R. R. Sokal. 1981. Critical values for correlation coefficients. In Statistical tables. 2d ed. San Francisco and Oxford: W. H . Freeman and Co. Romanic, A . M . A . J. Arleth, R. N . Willette, and E . H . Ohlstein. 1998. Factor XIIIa cross-links lipoprotein(a) with fibrinogen and is present in human atherosclerotic lesions. CircRes 83: 264-269. Rooney, M . M . , D . H . Farrell, B . M . van Hemel, P. G . de Groot, and S. T. Lord. 1998. The contribution of the three hypothesized integrin-binding sites in fibrinogen to platelet-mediated clot retraction. Blood 92: 2374-2381. Rosenfeld, G . C , D . C , Hou, J. Dingus, I. Meza, and J. Bryan. 1985. Isolation and partial characterization of human platelet vinculin. J Cell Biol 100: 669-676. Ruhnau, K . , and A . Wegner. 1988. Evidence for direct binding of vinculin to actin filaments. FEBSLett 228: 105-108. Saelman, E . U . M . , H . K . Nieuwenhuis, K . M . Hese, P. G . de Groot, H . F. G . Heijnen, E . H . Sage, S. Williams, L . McKeown, H . R. Gralnick, and J. J. Sixma. 1994. Platelet adhesion to collagen types I through VIII under conditions of stasis and flow is mediated by GPIa/IIa (a 2pi-integrin). Blood 83: 1244-1250. Saito, M . , H . Asakura, T. Yoshida, K . Ito, K . Okafuji, T. Yoshida, and T. Matsuda. 1990. A familial factor XIII subunit B deficiency. Br J Haematol 74: 290-294. 171 Sakata, Y . , and N . A o k i . 1982. Significance of cross-linking of (X2-plasmin inhibitor to fibrin in inhibition of fibrinolysis and in hemostasis. J Clin Invest 69: 536-542. Sakata, Y . , and N . A o k i . 1980. Cross-linking of a 2-plasmin inhibitor to fibrin by fibrin-stabilizing factor. J Clin Invest 65: 290-297. Sane, D . C. T. L . Moser, A . M . M . Pippen, C J. Parker, K . E . Achyuthan, and C. S. Greenberg. 1988. Vitronectin is a substrate for transglutaminases. Biochem Biophys Res Commun 157: 115-120. Sawicki, G . , E . J. Sanders, E . Salas, M Wozniak, J. Rodrigo, and M . W . Radomski. 1998. Localization and translocation of M M P - 2 during aggregation of human platelets. Thromb Haemost 80: 836-839. Schoenwaelder, S. M . , Y . Yuan, P. Cooray, H . H. Salem, and S. P. Jackson. 1997. Calpain cleavage of focal adhesion proteins regulates the cytoskeletal attachment of integrin anbP3 (platelet glycoprotein Ilb/IIIa) and the cellular retraction of fibrin clots. J Biol Chem 272: 1694-1702. Schwartz, M . L . , S. V . Pizzo, R. L . H i l l , and P. A . McKee . 1973. Human factor XIII from plasma and platelets. Molecular weights, subunit structures, proteolytic activation, and cross-linking of fibrinogen and fibrin. J Biol Chem 248: 1395-1407. Shainoff, J. R., D . A . Urbanic, and P. M . DiBel lo . 1991. Immunoelectrophoretic characterizations of the cross-linking of fibrinogen and fibrin by factor XHIa and tissue transglutaminase. Identification of a rapid mode of hybrid alpha-/gamma-chain cross-linking that is promoted by the gamma-chain cross-linking. J Biol Chem 266: 6429-6437. Shattil, S. J., L . F. Brass, J. S. Bennett, and P. Pandhi. 1985a. Biochemical and functional consequences of dissociation of the platelet membrane glycoprotein Ilb-IIIa complex. Blood 66: 92-98. Shattil, S. J., J. A . Hoxie, M . Cunningham, and L . F. Brass. 1985b. Changes in the platelet membrane glycoprotein Ilb.IIIa complex during platelet activation. J Biol Chem 260: 11107-11114. Siebenlist, K . R., D . A . Meh, and M . W. Mosesson. 1996. Plasma factor XIII binds specifically to fibrinogen molecules containing y' chains. Biochemistry 35: 10448-10453. Siebenlist, K . R., and M . W. Mosesson. 1992. Factors affecting gamma-chain multimer formation in cross-linked fibrin. Biochemistry 31: 936-941. 172 Sixma, J. J., A . van den Berg, B . M . Jockusch, and J. Hartwig. 1989. Immunoelectron microscopic localization of actin, alpha-actinin, actin-binding protein and myosin in resting and activated human blood platelets. Eur J Cell Biol 48: 271-281. Sixma, J. J., A . van den Berg, M . Schiphorst, H . J. Geuze, and J. McDonagh. 1984. Immunocytochemical localization of albumin and factor XIII in thin cryo sections of human blood platelets. Thromb Haemost 51: 388-391. Sonnenberg, A . , K . R. Gehlsen, M . Aumailley, and R. Timpl . 1991. Isolation of o^Pi integrins from platelets and adherent cells by affinity chromatography on mouse laminin fragment E8 and human laminin pepsin fragment. Exp Cell Res 197: 234-244. Staatz, W. D. , S. M . Rajpara, E . A . Wayner, W . G . Carter , and S. A . Santoro. 1989. The membrane glycoprotein Ia-IIa ( V L A - 2 ) complex mediates the Mg++-dependent adhesion of platelets to collagen. J Cell Biol 108: 1917-1924. Takagi, T., and R. F. Doolittle. 1974. Amino acid sequence studies on factor XIII and the peptide released during its activation by thrombin. Biochemistry 13: 750-756. Takahashi, N . , Y . Takahashi, and F. W. Putnam. 1986. Primary structure of blood coagulation factor XIIIa (fibrinoligase, transglutaminase) from human placenta. Proc Natl Acad Sci USA 83: 8019-8023. Takubo, T., M . Hino, K Suzuki, and N . Tatsumi. 1998. Localization of myosin, actin, a-actinin, tropomyosin and vinculin in surface-activate, spreading human platelets. Biotech Histochem 73: 310-315. Tamaki, T., and N . A o k i . 1982. Cross-linking of a2-plasmin inhibitor to fibrin catalyzed by activated fibrin-stabilizing factor. J Biol Chem 257: 14767-14772. Tamaki, T., and N . A o k i . 1981. Cross-linking of a2-plasmin inhibitor and fibronectin to fibrin by fibrin-stabilizing factor. Biochim Biophys Acta 661: 280-286. Tillett, W. S., and R. L . Garner. 1933. The fibrinolytic activity of hemolytic streptococci. J Exp Med 5S: 485-502. Turner, C. E . , and K . Burridge. 1989. Detection of metavinculin in human platelets using a modified talin overlay assay. Eur J Cell Biol 49: 202-206. Tuszynski, G . P., E . Kornecki, C. Cierniewski, L . C. Knight, A . Koshy, S. Srivastava, S. Niewiarowski, and P. N . Walsh. 1984. Association of fibrin with the platelet cytoskeleton. J Biol Chem 259: 5247-5254. V u , T - K . H . , D . T. Hung, V . I. Wheaton, and S. R. Coughlin. 1991. Molecular cloning of a functional thrombin receptor reveals a novel proteolytic mechanism of receptor activation. Cell 64: 1057-1068. 173 Wang, D . L . , A . E . Annamalai, S. Ghosh, A . M . Gewirtz, and R. W . Colman. 1990. Human platelet factor V is crosslinked to actin by FXIIIa during platelet activation by thrombin. Thromb Res 57: 39-57. Weiss, H . J., V . T. Turitto, and H . R. Baumgartner. 1978. Effect of shear rate on platelet interaction with subendothelium in citrated and native blood. I. Shear rate—dependent decrease of adhesion in von Willebrand's disease and the Bernard-Soulier syndrome. JLab Clin Med 92: 750-764. Weiss, M . S., H . J. Metzner, and R. Hilgenfeld. 1998. Two non-proline cis peptide bonds may be important for factor XIII function. FEBS Lett 423: 291-296. White, J. G . 1974. Electron microscopic studies of platelet secretion. [Review] [140 refs] Prog Hemost Thromb 2: 49-98. White, J. G . 1972. Interaction of membrane systems in blood platelets. Am J Path 66:295-312. White, J. G . 1971. In Platelet aggregation, ed. J. Caen, 15-52. Paris: Masson et Cie. White, J. G . 1968. Effects of colchicine and vinca alkaloids on human platelets. I. Influence on platelet microtubules and contractile function. Am J Path 53:281-291. White, J. G . , S. M . Burris, D . Tukey, C. Smith, and C. C. Clawson. 1984. Micropipette aspiration of human platelets: influence of microtubules and actin filaments on deformability. Blood 64: 210-214. Williams-Ashman, H . G . , A . C. Notides, S. S. Pabalan, and L . Lorand. 1972. Transamidase reactions involved in the enzymic coagulation of semen: isolation of -glutamyl- -lysine dipeptide from clotted secretion protein of guinea pig seminal vesicle. Proc Natl Acad Sci USA 69: 2322-2325. Wolpl , A . , H . Lattke, P. G . Board, R. Arnold, T. Schmeiser, B . Kubanek, M . Robin-Winn, R. Pichelmayr, and S. F. Goldmann. 1987. Coagulation factor XIII A and B subunits in bone marrow and liver transplantation. Transplantation 43: 151-153. Yee, V . C , I. Le Trong, P. D . Bishop, L . C. Pedersen, R. E . Stenkamp, and D . C. Teller. 1996. Structure and function studies of factor XIIIa by x-ray crystallography. Sem Thromb Hemost 22: 377-384. Yee, V . C , L . C. Pedersen, P. D . Bishop, R. E . Stenkamp, and D . C. Teller. 1995. Structural evidence that the activation peptide is not released upon thrombin cleavage of factor XIII . Thromb Res 78: 389-397. 174 Yee, V . C , L . C. Pedersen, I. Le Trong, P. D . Bishop, R. E . Stenkamp, and D . C. Teller. 1994. Three-dimensional structure of a transglutaminase: human blood coagulation factor XIII. Proc Natl Acad Sci USA 91: 7296-7300. Yorifuji, H . , K . Anderson, G . W. Lynch, L . VanDe Water, and J. McDonagh. 1988. B protein of factor XIII : differentiation between free B and complexed B . Blood 72: 1645-1650. Zhu, Y . , L . Tassi, W. Lane, and M . E . Mendelsohn. 1994a. Specific binding of the transglutaminase, platelet factor XIII, to HSP27. J Biol Chem 269: 22379-22384. Zhu, Y . , S. O ' N e i l l , J. Saklatvala, L . Tassi, and M . E . Mendelsohn. 1994b. Phosphorylated HSP27 associates with the activation-dependent cytoskeleton in human platelets. Blood 84: 3715-3723. Zucker-Franklin, D. , and S. R. Petursson. 1984. Thrombocytopoiesis: analysis by membrane tracer and freeze-fracture studies on fresh human and cultured mouse megakaryocytes. J Cell Biochem 99: 390-402. 175 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0089876/manifest

Comment

Related Items