Factor VIII, a cofactor of the intrinsic clotting path- way, is proteolytically inactivated by the vitamin K- dependent serine protease, activated protein. C in a reaction ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY (c)
VOl. 266, No. 4, Issue of February 5, pp. 2172-2177, 1991 Printed in U,S.A.
1991 by The American Society for Biochemistry and Molecular Biology, Inc
von Willebrand Factor MediatesProtection of FactorVI11 from Activated Protein C-catalyzed Inactivation* (Received for publication, September 20, 1990)
Philip J. Fay$B, Jean-Valery Coumanss, and FrederickJ. Walker 11 From the §Hematology Unit, Departmentof Medicine and Department of Biochemistry, University of Rochester School of Medicine and Dentistry, Rochester, New York 14642 and IlAmerican Red Cross Blood Services, Connecticut Region, and the Departments of Medicine and Laboratory Medicine, University of Connecticut, Farmington, Connecticut 06032
Factor VIII, a cofactor of the intrinsic clotting path- vWfl (4,5 ) . Activation of factor VI11 by thrombin dissociates way, is proteolytically inactivated by the vitamin K- factor VI11 fromvWf, thus allowing its participation as a dependent serine protease, activated protein C in a cofactor in the membrane-dependent activation of factor X reaction requiring Ca2+ and a phospholipid surface. by the vitamin K-dependent serine protease factor IXa (see Factor VI11 was inactivated15 times faster than factorRef. 6 for review). Protein C, alsoa vitamin K-dependent VI11 in complex with either von Willebrandfactor plasma protein (7), when activated to a serine protease (acti(vWf) or the largehomodimeric fragment, SPIII (vWf vated protein C), becomes a potent inhibitor of blood coaguresidues 1-1365). Free factor VI11 or factor VI11 in lation (8, 9). The basis of this anticoagulant activity is the complex with a smaller fragment, SPIII-TI(vWf res- phospholipid-dependent proteolytic inactivation of factors Va idues 1-272),were inactivated at the same rate, suggesting that this effect was dependent upon the sizeof (10-12) and VIII(a) (12-15). Factor VI11 is synthesized asa single-chain precursor repfactor VIII-vWf complex rather thanchanges in factor resented by the domain structure Al-AZ-B-A3-Cl-C2 with VI11 brought about by occupancy of the vWf-binding site. Thrombin cleavage of the factor VI11 light chain heterodimers formed as a result of proteolysis at the B-A3 to remove the vWf-binding site eliminated the protec-junction plus additionalcleavages within the B domain (16). tive effectsof vWf.In theabsence of phospholipid, high The factor VI11 heavy chain is minimally represented by the size heterogeneity levels of the protease inactivated bothfree and vWf- Al-AZ domainsbutexhibitssignificant bound factor VI11 at equivalent rates. Using the same resulting from the presence of some or all of the contiguous B domain, whereas the light chain corresponds to the A3-Clconditions, isolated heavy chains and the heavy chains of the of factor VI11 were proteolyzed at similar rates. Taken C2 domains derivedfrom theCOOH-terminalend precursor. The intact heterodimeric structure is essentialfor together, these results suggested that, in the absence of phospholipid, inactivation of factor VI11 is inde- cofactor functioninthatthesubunits of factor VI11 are pendent of factor VI11 light chain and furthersuggest dissociated by chelating reagents resulting inloss of clotting that vWf did not mask susceptible cleavage sites in the activity (1,17). While little is known of the role of the heavy cofactor. Solution studies employing fluorescence en- chain in factor VI11 function, the light chain has been obergy transfer using coumarin-labeled factor VI11 (flu- served tocontainsites for binding of vWf ( l ) , activated orescence donor) and synthetic phospholipid vesicles protein C (18), and phospholipid (19). labeled with octadecyl rhodamine (fluorescence accep- Recent studies have indicated that association the of factor tor) indicated saturable binding and equivalent extents VI11 with vWf protects the cofactor from activated protein C of donor fluorescence quenching for factor VI11 alone (20, 21). Inthisreport we have examinedthe effects of or when complexed with SPIII-T4. However, complex- multimeric vWf and vWf fragments containing the factor ing of factor VI11 with eithervWf or SPIII eliminated VI11 bindingsiteontheinteractions of factor VI11 with its binding to the phospholipid. Since a phospholipid activated protein C and phospholipid vesicles. Our results are surface is required for efficient catalysis by the proconsistent with a mechanism where protection of the cofactor tease, these results suggest that vWf protects factor when bound to vWf results from a reduced affinity of the VI11 by inhibiting cofactor-phospholipid interactions. complex for the phospholipid surface. MATERIALS AND METHODS
Reagents-Human factor VI11 concentrate (KoateTM)was a genFactor VIII, a plasma protein absent or defective in indiA, circulates as a series of Me2+- erous gift from the Cutter Division of Miles Laboratories. ~ - 1 - T o vidualswithhemophilia sylamido-2-phenylethyl chloromethyl ketone-treated trypsin(bovine linkedheterodimers (1-3) innoncovalentassociationwith pancreas) was purchased from Sigma and further purified by reverse phase high performance liquid chromatography using a Vydac CIS * This work was supported in partby National Institutesof Health column (5 pm, 0.45 X 25 cm) developed with a linear gradient from Grant HL-38199 (to P. J. F.) and HL-40328 (to F. J. W.) and by the 20-50% acetonitrile in 0.1% trifluoroacetic acid. Staphylococcus auAmerican Heart Association Grant 90-0643 (to P. J. F.). The costs reus V8 protease was purchased from ICN ImmunoBiologicals. Huof publication of this article were defrayed in part by the paymentof man 01 thrombin was obtained from Enzyme Research Laboratories. OR were purchased from Molecular page charges. This article must therefore be hereby marked “adver- The fluorescent probes CPM and tisement’’ in accordance with 18U.S.C. Section 1734 solely to indicate The abbreviations used are: vWf, von Willebrand factor; CPM, this fact. 7-dimethylamino-3-[4’-maleimidophenyl]-4-methylcoumarin; OR, j EstablishedInvestigator of the American Heart Association. Address correspondence to: Hematology Unit, P. 0. Box 610, Univer- octadecyl rhodamine B; Hepes, 4-(2-hydroxyethyl)-l-piperazineethsity of Rochester Medical Center, 601 Elmwood Ave., Rochester, NY anesulfonic acid PC,L“-phosphatidylcholine; PS, L“-phosphatidyl-Lserine; SDS, sodium dodecyl sulfate. 14642. Tel.: 716-275-7987. Fax: 716-473-4314.
2172
v Wf Protection of Factor VIII from Activated Protein C Probes. Le-Phosphatidylcholine (bovine brain) and La-phosphatidylL-serine (bovine brain) were purchased from Sigma. Proteins-Bovine activated protein C was prepared as described previously (22). Factor VIII, factor VI11 subunits, and vWfwere prepared from human factor VI11 concentrates as described previously (23, 24). Thrombin-cleaved factor VI11 light chain was prepared from intact factor VI11 heterodimers as follows. Factor VI11 (390 pg) in 5 ml of 0.01 M histidine (pH 6.0), 0.15 M NaCl, and 0.003 M CaClz was treated with human a-thrombin (26 pg, 73 units) for 10 min at 22" C. Clotting assays indicated that factor VI11 had been maximally activated. To the thrombin-cleaved factor VI11 was added 0.5 mM diisopropyl fluorophosphate and 0.03 M EDTA and the reaction mixture incubated overnight. The thrombin-cleaved light chain was isolated following chromatography on a Mono S column (HR5/5) using conditions identical to those described for native factor VI11 light chain (23). SDS-polyacrylamide gel electrophoresis (silver-stained) showed no uncleaved light chain present in the modified light chain preparation. Modified light chain (10.4 pg) was reconstituted with native heavy chains (38 pg; approximately 2-fold molar excess) in 0.44 ml of 0.01 M Hepes (pH 7.2), 0.4 M NaCl, and 0.03 M MnClz for 2 h at 22" C. The activity of the reconstituted factor VI11 prepared from the thrombin-cleaved light chain plus native heavy chains was approximately 140 units/ml. Assuming an upper limit of 66 pg/ml for the concentration of reassociated factor VIII, this level of activity indicated a minimum specific activity of approximately 2 units/pg. Fragment SPIII was prepared from the S. aureus V8 protease digest of vWf and isolated by Mono Q (HR5/5) column chromatography as previously described (24). Fragment SPIII-T4 was prepared from a tryptic digest of SPIII (24). Preparations of SPIII-TI used in the present study effectively blocked fluorescence energy transfer between N-pyrene maleimide-labeled factor VI11 and CPM-SPIII (24), indicating a direct interaction between this fragment and factor VIII. Protein concentrations were determined by the method of Bradford (25). SDS-polyacrylamide gel electrophoresis was performed as previously described (24). Assays-Factor VI11 activity was measured using a one-stage clotting assay. Inactivation of factor VI11 by activated protein C was carried out as described previously (15). Samples were removed from reaction mixtures at the indicated times and assayed for residual factor VI11 activity. For experiments involving the effects ofvWf (derived fragments) on factor VI11 inactivation, factor VI11 was preincubated with a 10-fold molar excess of vWf (fragment) for 1 h a t 22 "C prior to reaction with activated protein C. Fluorescent Labeling of Factor VU-Factor VI11 (90-240 pg/ml) in 0.01 M histidine (pH 6.0), 0.05 M NaC1,0.005 M CaClZ,0.01% Tween-20, and 50% (v/v) ethyleneglycol wasreacted with the sulfhydryl-specific fluorophore CPM. Reactions containeda 20-50-fold molar excess of CPM, assuming a mean molecular mass of 220 kDa for factor VIII heterodimers (17). Prior to reaction, CPM was dissolved in a small volume of dry dimethylformamide and reactions were run overnight at 4 "C in the dark. The unbound fluorophore was separated from the modified protein by gel filtration using a PD-10 column (Pharmacia LKB Biotechnology Inc.) equilibrated in 0.02 M Hepes (pH 7.2), 0.15 M NaCl, and 0.003 M CaCl2. The molar ratio of probe bound to protein ranged from approximately 1.5 to 2.8 for the CPM-factor VIII, which possessed similar clotting activity when compared with the unmodified protein. Preparation of Phospholipid Vesicles-Vesicles (41 PCPs) were prepared according to the method of Husten et al. (26). OR-PCPs also were prepared as above using approximately 20 pg of OR per 2.5 mg (total) phospholipid. The final concentration of phospholipid was determined by an elemental phosphorous assay (27). The concentration of OR incorporated into P C P s vesicles wasdetermined using an extinction coefficient of 95,400 M" cm" at 564 nm. Fluorescent Measurements-Fluorescent measurements were made using a SPEX Fluorolog 212 spectrofluorometer. Samples were excited at 387 nm and emission spectra taken using 15-nm slit widths for both excitation and emission monochrometers. Data were collected over the appropriate wavelength ranges at 1-nm increments and a 1-s integration time. Energy transfer was measured for the CPM-factor VI11 (fluorescence donor) and OR-PCPs (fluorescence acceptor) pairing. Determinations for each concentration of acceptor involved spectral analysis for three samples: (i) CPM-factor VI11 plus unlabeled PCPs, (ii) unlabeled factor VI11 plus OR-PCPs, and (iii) CPM-factor VI11 plus OR-PCPs. Thecorrected emission spectra for the above samples were integrated from 470 to 560 nm. Donor quenching was calculated from the area ratio of sample iii to samples i + ii. All reactions (0.3 ml) were carried out at room temperature in
2173
buffer containing 0.02 M Hepes (pH 7.2), 0.15 M NaC1,0.003 M CaCl2, and 0.01 M lysine HCI. For reactions containing vWf or vWf-derived fragments, (CPM-) factor VI11 was first preincubated with the vWf for 1 h at 22" C in the reaction buffer prior to the addition of (OR-) P C P s vesicles. RESULTS
Effects of vWf on Factor VIII Inactivation-Catalytic amounts of activated protein C inactivated factor VI11 in the presence of Ca2+ anda phospholipid surface. However, prior incubation of factor VI11 with vWf reduced the rate of inactivation (Fig. 1). For this analysis, factor VI11 (mean molecular mass= 220 kDa) was reacted witha 10-fold molar excess of vWf for 1h prior to reaction with activated protein C. This ratio of vWfifactor VI11 and incubation time were chosen to ensurecompletebinding of factor VI11 (24). Underthese reaction conditions, the time required for enzymatic-catalyzed inactivation of 50% the factor VI11 was increased about 15fold (from -2 min to -30 min). To determine if protection resulted from either occupancy of the vWf-binding site on factorVI11 or depended upon the mass of the bound substrate, similar analyses were performed using two fragments of vWf, the homodimeric fragment SPIII (vWf residues 1-1365; subunit molecular mass -170 kDa) and the monomeric fragment SPIII-T4 (vWf residues 1-272; -34 kDa) (28). These fragments, which retain the factor VIIIbinding site (29), form stoichiometric complexes with factor VI11 (24). A similar protectiveeffect from activated protein C action to the factor VIII-vWf complex was observed forfactor VI11 bound to the SPIII homodimer. Here again, the rate of inactivation of the factor VIII-SPIII complex was approximately 15-fold slower compared with that observed for free factor VIJI. However, complexing of the smaller SPIII-T4 with factor VI11 offered no protection from activated protein C in that essentially similar reaction rateswere observed for free factor VI11 and the factor VIII-SPIII-T4 complex. These results suggested that the protection from activated protein C observed for factor VI11 complexed with either intact vWf or SPIII did not result from either occupancy of the vWfbinding site in factor VI11 or potential conformationalchanges induced in factorVI11 following binding, but insteadsuggested a mechanismdependent upon the size of the ligand that occupies the vWf-binding site. Results shown in Fig. 2 indicatedthat when the vWf-
0 0
10
20
30
TIME (Minutes)
FIG. 1. Effect of vWf (-derived fragments) on activated protein C-catalyzed inactivation of factor VIII. Factor VI11 (12 pg/ ml; approximately 60 units/ml) was reacted with either buffer alone (0), vWf (147 pg/ml; B),SPIII (88 pg/ml; A),or SPIII-T4 (21 pglml; 0 )for 1 h at room temperature. Concentrations of the vWf substrate represented a IO-fold molar excess relative to factor VIII. These samples were then treated with activated protein C (1.2 pg/ml), and aliquots were removed at theindicated times and assayed for residual factor VI11 activity. Data points represent the mean of at least two separate experiments.
u Wf Protection of Factor VIII from Activated Protein
2174 - 100
*
C
1
Ph 5 so
0
10 20 TIME (Minutes)
FIG.2. Effect of vWf on inactivation of a modified factor VI11 lacking the vWf binding domain. Thrombin-cleaved factor VI11 light chain was reconstituted witha 2-fold molar excess of native heavy chainsas described under“Materialsand Methods.” This modified factor VI11 (20 pg/ml in total factor VI11 protein; approximately 25 units/ml) was reacted with either buffer alone (0)or vWf (146 pg/ml; B)for 1 h a t room temperature followed by reaction with activated protein C (1.2 pglml). Aliquots were removed at the indicated times following protease addition and assayed for factor VI11 activity. Data points represent the meanof two experiments.
binding site was removed from factor VIII, the protective effect of vWf toward activated protein C-catalyzed inactivation was abolished. For this analysis, factor VI11 was treated with thrombin to effect removal of the NH2-terminal peptide (residues 1649-1689) (30) containing the vWf-binding site from the factor VI11 light chain. Intact factor VI11 rather than factor VI11 light chain was used to prepare the thrombincleaved light chain because proteolysis of the isolated subunit was slow and failed to yield a fully cleaved product. Subsequent reconstitution of native factor VI11 heavy chains with the thrombin-cleaved light chain resulted in theformation of a proteolytically modified factor VI11 with similar (stable) clotting activity (specific activity -2 units/pg) when compared with native factor VI11 but altered only in the light chain so that it lacked the ability to bind vWf. Equivalent rates of activated protein C-catalyzed inactivation of this modified factor VI11 were observed in assaysperformed in the absence of vWf or following a preincubation with a 10-fold excess of vWf. Thus protection is achieved only when factor VI11 complexes with vWf. Multiple binding domains have been localized within the light chain of factor VI11 and include regions that bind vWf (residues 1670-1684) (31), phospholipid (residues 2303-2332) (32), and activated protein C (residues 2009-2018) (33). The results presented above are suggestive of two possible mechanisms by which the protective effect of vWf is dependent on the size of the binding ligand. The firstpossibility is that vWf can sterically block access to theactivated protein Cbinding site or mask the cleavage sites located on the heavy chain. The second possibility is that vWf can block the interaction of factor VI11 with membranes and preventformation of catalytic complex which would include protease, Ca2+,phospholipid, and factor VIII. The requirement for phospholipid in the formation of the catalytic complex can be overcome by high concentrations of activated protein C (Fig. 3). In the absence of phospholipid, an approximate 15-fold molar excess of enzyme relative to factor VI11 resulted in a rate of inactivation similar to that observed using catalytic levels of enzyme in the presence of phospholipid (see Fig. 1). Further, under these conditions of a high ratio of enzyme to substrate andno phospholipid, the rate of factor VI11 inactivation either in the presence or absence of vWf were equivalent. This experiment suggested that vWf did not mask the cleavage sites in the heavy chain from activated protein C.
20
10
0
30
30
TIME (Minutes)
FIG.3. Effect ofvWf on factor VI11 inactivation in the absence of phospholipid. Factor VI11 (13.6 pg/ml; approximately 70 units/ml) was reacted for 1 h a t room temperature with either buffer (0)or vWf (149 pg/ml; M). Samples were then treated with buffer but lacking activated proteinC (46 pg/ml) in the standard assay phospholipid.Aliquots were assayed at the indicatedtimes. Data points represent the meanof three separate experiments.
1 200
-
2
3
4
5
6
7
8
9 10111213 ._
’4J= 68-
4529-
FIG. 4. SDS-polyacrylamide gel electrophoresis of isolated factor VI11 heavy chainsandintactfactor VI11 following interaction with activated proteinC. Reaction mixtures (120 pl) contained either factor VI11 heavy chains (15 pg/ml), heavy chains (15 pglml) plus100 pg/ml phospholipid, or intact factor VI11 (20 pg/ ml). Activated protein C (50 pg/ml) was added to each reaction and incubated at 37 “C. Aliquots were removed at the indicated times, denatured, subjected to electrophoresis, and thegel stained with silver nitrate. Lanes 1, 5 , and 9 represent heavy chains, heavy chains plus phospholipid, and intact factorVIII, respectively, prior to addition of protease. Lanes 2-4, 6-8, and 10-12 represent 5-, 15-, and 60-min time points following protease addition to reactions shown in lanes 1 , 5 , and 9,respectively. Lane 13 shows activated protein C alone. M , is X io?
In the absence of phospholipid, we observed that cleavage of the heavy chain was also light chain-independent. High concentrations of activated protein C resulted in proteolysis of the isolated factor VI11 heavy chains (Fig. 4). The apparent rate of degradation, as judged by disappearance of heavy chains and appearance of terminal fragments of 48 and 23 kDa, was independent of the presence of phospholipid. Furthermore, for reactions run in the absence of phospholipid, the rate of cleavage of isolated heavy chains was similar to that observed for the heavy chains of intact factor VIII. This result suggested that inactivation of factor VI11 by activated protein C, in the absence of phospholipid, occurs by a mechanism independent of factor VI11 light chain. These experiments suggest that theeffect of vWf on activated protein Ccatalyzed inactivation of factor VI11 is to alter a parameter that is dependent upon the presence of the light chain. Effects of u Wf on Factor VZZI-Phospholipid ZnteractionsFluorescence energy transfer techniques were applied to determine if the results obtained from clotting assays, namely
v Wf Protection of Factor VIII from Activated Protein C
2175
protection of factor VI11 from activated protein C when bound to vWf and SPIII but not to SPIII-T4, were compatible with the effects of these substrates on the interaction between factor VI11 and phospholipid vesicles. For these experiments, factor VI11 was modified with the sulfhydryl-specific fluorophore, CPM (fluorescence donor). We have previously used this probe to modify factor VI11 subunits and have shown incorporation into residues in both the heavy chain (CYS~’~) (23). Intact factor VI11 incorpoand the light chain (CYS’~~’) rated on average about 2 mol of probe per molof protein, suggesting sites onbothsubunits were modified. Clotting 30 0 10 20 activity of the CPM-modifiedprotein was similar to thenative PCPs (uM) material (data not shown). Phospholipid vesicles (41PCPs) FIG. 6. Effect of vWf (-derived fragments) on the CPMwere modified with OR at a ratio of approximately 1 mol of probe per 300 mol of phospholipid monomer for use as the factor VIII/OR-PCPS pairing. (CPM-) factor VI11 (0.5 pg) was fluorescence acceptor. Substitution of OR-phospholipid for incubated with either buffer (O), 6.1 pg of vWf (a),3.9 pg of SPIII (A),or 0.8 pg of SPIII-T4 ( 0 )for 1 h a t room temperature. The unmodified phospholipid used in assays monitoring the acti- indicated amounts of (OR-) PCPswere then added and thereaction vated protein C-catalyzed inactivation of factor VI11 yielded mixtures (0.3 ml) were subjected to fluorescence analysis. Relative equivalent results (data not shown) indicating that presence fluorescence wasdetermined asdescribed under “Materials and Methof the probe did not alter the interactions of the lipid with ods.” Data points represent the meanof at least two determinations. factor VI11 or activated protein C. The OR-PCPs had an excitation spectrum that overlapped the emission spectrum subunit which is cleaved to several intermediate and terminal of CPM-factor VI11 (Fig. 5). Thus thefluorescence of CPM- digest products (15). However, our earlier observations (18) factor VI11 should be quenched upon binding to the labeled that (i) heavy chain alone was not a substrate for activated phospholipid vesicles. However, no donor fluorescence protein C, (ii) isolated light chain inhibited inactivation of quenching would be observed if prior complexing of CPM- factor VIII, and (iii) a (phospholipid-dependent) binding of factor VI11 with vWf (or a vWf-derived fragment) prevented light chain and the enzyme suggested an integral role for the its association with the phospholipid surface. light chain in this catalytic mechanism. This paper further CPM-factor VI11 was titrated with OR-PCPs (Fig. 6). The emphasizes the importance of the light chain in the catalytic acceptor quenched the fluorescence of CPM-factor VI11 ap- mechanism. In this paper we observe that association of the proximately 12% and thiseffect was saturable. Incubation of light chain with phospholipid membranes is an essential step CPM-factor VI11 with vWf or SPIIIblocked subsequent donor in activated protein C-catalyzed inactivation of factor VIII. quenching following addition of OR-PCPs, whereas prior In addition, we have observed that cleavage of the isolated complexing of CPM-factor VI11 with SPIII-T4 yielded a result heavy chain does not appear to be accelerated by these memequivalent to that observed for free factor VIII. These data branes, suggesting that an activated protein C-phospholipid suggested that prior complexing of factor VI11 with either interaction is not sufficient for the rapid cleavage of substrate. vWf or SPIII prevented factor VI11 from binding the phos- Recently we have localized an activated protein C-binding pholipid vesicles. The above results, taken together with the region to light chain residues 2009-2018 (33) located near the results indicating protectionof factor VI11 from the protease COOH-terminal end of the A3 domain. when the cofactor was complexed with vWf or SPIII but not Also contained within the factor VI11 light chain (A3-C1SPIII-T4, support a model where vWf-mediated protection C2 domainal structure; residues 1649-2332) (16) are sites for from activated protein C resultsfrom inhibition of the factor binding vWf and phospholipid. vWf binds very near the NH2 VIII-phospholipid interaction. terminus of this subunit in that thrombin cleavage (at residue (or light chain) from vWf (34, 1689) dissociates factor VI11 DISCUSSION 35). Results of Foster et al. (31) have further localized the Limited proteolysis of factor VI11 by activated protein C vWf-binding region to residues 1670-1684. Recently, these correlates with the observed inactivation of cofactor function investigators have suggested that residues 2303-2332 mediate (13-15, 30). This proteolysis is restricted to the heavy chain the binding of factor VI11 to phospholipid (32). It is not known how these regions that bind protease, vWf and phospholipid are spatially oriented inthe folded protein. As a result of the multiple macromolecular interactions attributed to the light chain, one can envision several alternative mechanisms for the vWf-dependent protection of factor VI11 from activated protein C-catalyzed inactivation. Protection of factor VI11 was observed when the cofactor was complexed with multimeric vWf or the homodimeric SPIII but not with the smaller monomer, SPIII-T4. This result indicated a size dependence of the binding substrate with respect to protection, thus excluding the possibilities that occupancy of the vWf-binding site would either induce a 400 450 500 550 600 conformational change or in itself preclude interaction of WAVELENGTH (nm) factor VI11 with protease. Instead, the above result was comFIG. 5. Spectral relationship between CPM-factor VI11 and patible with the larger vWf substrates either sterically blockOR-PCPS. Corrected emission spectrum for CPM-factor VI11 excited a t 387 nm (. . . . .) and excitation spectrum for OR-PCPs ing access to the activated protein C-binding sites and/or masking the cleavage sites present in the heavy chain. This measured a t 650 nm (-) were determined at 1-nm increments. Fluorescence intensity is in arbitrary units. latter alternative was suggested by our previous studies which
2176
u W fProtection of Factor VIII from Activated Protein
C
indicated activated protein C initially cleaves factor VI11 a t have observed partial protection of factor VI11 (1.2 nM) from site(s) within the A2 domain of the heavy ch@n (15) and a theprotease(4 nM) by vWf, whereasnoprotection was close spatial separation of approximately 30 A between the observed following thrombin activation of factorVIII, the NH2-terminal region of vWf and this domain in the reconsti-latter result consistent with protection requiring a physical tuted factor VIII-vWf complex (24). Our experiments do not association between factor VI11 and vWf. Similar to thrombin-activated factor VI11 (factor VIIIa), allow us to exclude the possibility that vWf sterically blocks the proteolytically modified factor VI11 employed inthis the interaction between activatedprotein C and the light chain. Experimentally, this would be difficult to prove since study, composed of a thrombin-cleaved light chain bound to binding of activated protein C to the light chain is lipid- a native heavy chain was inactivated by the protease ata rate dependent and vWf inhibits the light chain interaction with independent of the presence of vWf. Furthermore, in the absence of vWf, the rateof inactivation of this form of factor lipid, it is not possible to set up an assay in which a lipidbound light chain-vWf complex could be used as a ligand for VI11 was similar to that observed for native factor VI11 (see Figs. 1 and 2). Although factor VI11 is a good substrate for protein C binding. However, vWf does not appear to mask bonds in the factor VI11 heavy chain from cleavage by acti- activated proteinC (15), Marlaret al. (12) reported thatfactor vated protein C. In theabsence of a phospholipid surface and VIIIa was inactivated 30-fold faster by the protease. Since at high protease levels, we observed similar rates of inactiva- this value was determined from comparison with the inactition of factor VI11 and factor VI11 complexed with vWf (Fig. vation of factor VIII/vWf, the rate enhancementobserved for in part, from its 3 ) and similar rates of proteolysis of free heavy chain and the the activated cofactorprobablyresulted, dissociation from vWf. However, preliminary data from our heavy chain of intactfactor VI11 (Fig.4). Theseresults suggested a surface-independent mechanism for cofactor in- laboratory suggest a significant increase in the rate of factor activation that was also independent of an activated protein VIIIa inactivation by the protease. Since thrombin cleaves one or both C-light chain interaction. These results suggest that the so- bothfactor VI11 subunitsduringactivation, cleavage events must result in increasing the reactivity of the lutionphaseinactivation of factor VIII, which is slow, is activated cofactor for the enzyme. The above result suggests independent of light chain. cleavage of the heavy chain, not light Factor VI11 binds tightly and reversibly to phospholipid that it is thrombin vesicles (36) and platelets (37) and it is when factor VI11 is chain, that disposes factor VIIIa torapidinactivation by surface-bound that it is an optimal substrate for activated activated protein C. It is not known what function(s) are impairedin activated protein C. Earlier results showed that high concentrations of is phospholipid (above 250 pg/ml) dissociated factor VI11 from protein C-cleaved factor VIII. Since the light chain subunit between not covalently altered it is unlikely that inactivation results vWf (38), indicatinganantagonisticrelationship vWf and phospholipid for factor VI11 binding. Thus vWf, by from altered phospholipid binding. Activated protein C cleavinterferingwiththefactor VIII-phospholipid interaction, age of factor Va reduces the affinity of the cofactor for both (39). By analogy, inactivated could potentially reduce the rate of cofactor inactivation by prothrombinandfactorXa factor VI11 would show reduced affinity for factor X and activated protein C. macromolecular interactions among We have employed fluorescence energytransfer techniques factor IXa. Studies on the components of the factor Xase enzyme complex and alterat o assess phospholipid binding of free factor VI11 and factor VI11 complexed to vWf, SPIII or SPIII-T4. Fluorescence data tions produced by activated protein C are currently in proindicated saturable bindingof both free factor VI11 and factor gress. VI11 complexed with SPIII-T4 to the phospholipid surface. Acknowledgments-We thank the Cutter Division of Miles LaboLevels of fluorescence quenching a t saturating levels of phos- ratories for the therapeutic concentratesused to prepare factor VI11 pholipid, an indicator of the distance separating donor and and vWf, Therese Smudzin for excellent technical assistance, and acceptorfluorophores, were equivalent for factor VI11 and Carol Weed for help in the preparation of the manuscript. factor VIII/SPIII-T4. Thus occupancy of the vWf-binding REFERENCES site by the 34-kDa SPIII-T4 fragment did not perturb the 1. Fass, D. N., Knutson, G. J., and Katzmann, J. (1982) Blood 59, factor VIII-phospholipid interaction, suggesting thatthe 594-600 phospholipid and vWf binding sites within the light chain 2. Anderson, L. O., Forsman, N., Huang, K., Larsen, K., Lundin, were not juxtaposedso closely that theywere mutually excluA., Pavlu, B., Sandberg, H., Sewerin, K., and Smart, J. (1986) sive. However, little if any donor fluorescence quenching was Proc. Natl. Acad. Sci. U. S. A. 83,2979-2983 observed for factor VI11 in complex with vWf or SPIII, indi3. Fay, P. J., Anderson,M.T., Chavin, S. I., and Marder, V. J. (1986) Biochim. Biophys. Acta 871, 268-278 cating no interaction of these complexes with the phospho4. Weiss, H. J., Sussman, I. I., and Hoyer,L. W. (1977) J . Clin. lipid surface. Thus the effects of vWf, SPIII, and SPIII-T4 Inuest. 60,390-404 on the factor VIII-phospholipid interaction paralleled their 5 . Owen, W. G., and Wagner, R. H. (1982) Thromb. Diath. Hueeffects in the factor VIII-activated proteinC system, suggestmorrh. 27, 502-515 ing that protectionfrom protease resulted from inhibitionof 6. Kane, W. H., and Davie, E. W. (1988) Blood 71,539-555 7. Stenflo, J. (1976) J . Biol. C k m . 251, 355-363 cofactor binding tophospholipid. 8. Seegers, W., McCoy, L. E., Groben, H. D., Sakuragawa, N., and The protection offered factor VI11 by vWf from activated Agrawal, B. B. L. (1972) Thromb. Res. 1, 443-456 protein C-catalyzed inactivation has been the subject of two 9. Kisiel, W., Ericsson, L. H., and Davie, E. W. (1976) Biochemistry recent reports. Rick et al. (21) showed that vWf modestly 15,4893-4900 decreased the level of factor VI11 inactivation by about 20- 10. Kisiel, W., Canfield, W. M., Ericsson, L. H., and Davie, E. W. 30%. These studies used factor VI11 levels of 0.6-0.8 unit/ml (1977) Biochemistry 16, 5824-5831 and protease concentrations of up to 2.3 pg/ml. Assuming a 11. Walker, F. J., Sexton, P. W. and Esmon, C. T. (1979) Biochim. Biophys. Acta 571,333-342 specific activity for factor VI11of 5 unitslpg,this would 12. Marlar, R. W., Kleiss, A. J., and Griffin, J. H. (1982) Blood 59, indicate a (weight) ratioof pr0tease:substrate = 16. Thus, the 1067-1072 observed lack of significant protection was likely attributed 13. Vehar, G. A,, and Davie, E. W. (1980) Biochemistry 19,401-410 to thehigh levels of enzyme usedand reflected aphospholipid- 14. Fulcher, C. A., Gardiner, J. E., Griffin, J. H. and Zimmerman, T. S . (1984) Blood 63,486-489 independent inactivation of the cofactor. Koedam et al. (20)
v Wf Protection of Factor VIII from Activated ProteinC 15. Walker, F. J., Chavin, S. I., and Fay, P. J. (1987) Arch. Biochern. Biophys. 252,322-328 16. Vehar, G. A,, Keyt, B., Eaton, D., Rodriquez, H., O’Brien, D. P., Rotblat, F., Oppermann, H., Keck, R., Wood, W. I., Harkins, R. N., Tuddenham, E. G. D., and Capon, D. J. (1984) Nature 312,337-342 17. Fay, P. J. (1988) Arch. Biochem. Biophys. 262, 525-531 18. Fay, P. J., and Walker, F. J . (1989) Biochirn. Biophys. Acta 994, 142-148 19. Bloom, J. W. (1987) Thrornb. Res. 48,439-448 20. Koedam, J. A., Meijers, J. C. M., Sixma, J. J., and Bouma, B. (1988) J . Clin. Inuest. 82, 1236-1243 21. Rick, M. E., Esmon, N. L., and Krizek, D. M. (1990) J. Lab. Clin. Med. 115,415-421 22. Walker, F. J. (1980) J. Biol. Chem. 255.5521-5524 23. Fay, P. J., and Smudzin, T. M. (1989) J. Biol. Chern. 264,1400514010 24. Fay, P. J., and Smudzin, T. M. (1990) J. Biol. Chern. 265,61976202 25. Bradford, M. M. (1976) Anal. Biochern. 72, 248-254 26. Husten, E. J., Esmon, C. T., and Johnson, A. E. (1987) J. Bid. Chern. 262, 12953-12962 27. Skipski, V., Peterson, R., and Barclay, M. (1964) Biochern. J . 90, 374-378
2177
28. Marti, T., Rosselet, S. J., Titani, K., and Walsh, K. A. (1987) Biochemistry 26,8099-8109 29. Bahou, W. F., Ginsburg, D., Sikkink, R., Litwiller, R., and Fass, D. N. (1989) J . Clin. Inuest. 84, 56-61 30. Eaton, D., Rodriguez, H., and Vehar, G. A. (1986) Biochemistry 25,505-512 31. Foster, P. A., Fulcher, C. A., Houghten, R. A., and Zimmerman, T. S. (1988) J . Biol. Chern. 263, 5230-5234 32. Foster, P. A., Fulcher, C. A., Houghten, R. A., and Zimmerman, T. S. (1990) Blood 75, 1999-2004 33. Walker, F. J., Scandella, D., and Fay, P. J. (1990) J. Biol. Chern. 265,1484-1489 34. Hamer, R. J., Koedam, J . A., Besser-Visser, N. H., and Sixma, J. J. (1987) Eur. J. Biochern. 167, 253-259 35. Lollar, P., Hill-Eubanks, D. C., and Parker, C. G. (1988) J . Biol. Chern. 263, 10451-10455 36. Gilbert, G. E., Furie, B. C. and Furie, B. (1990) J . Biol. Chem. 265, 815-822 37. Nesheim, M. E., Pittman, D. D., Wang, J. H., Slonosky, D., Giles, A. R., and Kaufman, R. J . (1988) J. Biol. Chern. 263, 1646716470 38. Anderson, L-O., and Brown, J. E. (1981) Biochem. J . 206, 161167 39. Guinto, E. R., and Esmon, C. T. (1984) J . Biol. Chern. 259, 13986-13992
