Biochem. J. (2009) 422, 257–264 (Printed in Great Britain)
257
doi:10.1042/BJ20090405
Differential contributions of Glu96 , Asp102 and Asp111 to coagulation Factor V/Va metal ion binding and subunit stability Jina SONG*†‡, Kimberley TALBOT*†‡, Jeffrey HEWITT†§, Ross T. A. MACGILLIVRAY†§ and Edward L. G. PRYZDIAL*†‡1 *Canadian Blood Services, Research and Development, 2350 Health Sciences Mall, University of British Colombia, Vancouver, British Columbia, Canada V6T 1Z3, †Centre for Blood Research, 2350 Health Sciences Mall, University of British Colombia, Vancouver, British Columbia, Canada V6T 1Z3, ‡Pathology and Laboratory Medicine, 2350 Health Sciences Mall, University of British Colombia, Vancouver, British Columbia, Canada V6T 1Z3, and §Biochemistry and Molecular Biology, 2350 Health Sciences Mall, University of British Colombia, Vancouver, British Columbia, Canada V6T 1Z3
Blood coagulation FV (Factor V) is activated by thrombinmediated excision of the B domain, resulting in a non-covalent heterodimer, FVa (activated FV). Previous studies implicated Glu96 , Asp102 and Asp111 in the essential Ca2+ -dependent FVa subunit interaction. In the present study, FV E96A, D102A and D111A were purified and evaluated for function, subunit dissociation and metal ion binding. Chromogenic and clotting assays in the presence of procoagulant vesicles showed that each variant was inhibited (∼ 20–40 %). D111A was further inhibited (> 90 %) after cleavage by thrombin. Comparable function was observed on activated platelets. D111A inhibition correlated to spontaneous subunit dissociation and severely impaired Ca2+ binding. The Cu2+ interaction was also inhibited, suggesting interdependent Ca2+ and Cu2+ binding to FV. The parental FV (FV-810; wild-type human FV missing residues 811–1491) used here is fully active without proteolysis because the B domain is
truncated. Therefore, a FVa-like functional configuration exists for intact D111A independent of normal metal ion interactions. Unlike D111A, the thrombin-mediated FVa derived from E96A and D102A had only moderately enhanced subunit dissociation upon chelation and had normal metal ion binding. For FV810-, E96A- and D102A-derived FVa, loss of function after chelation significantly preceded subunit dissociation. This study defines the highly conserved segment spanning Glu96 –Asp111 in FV as multifunctional. Of the three amino acids evaluated, Asp111 is essential and probably functions through direct and indirect effects on Ca2+ and Cu2+ interactions. Glu96 and Asp102 individually influence FV/FVa by more subtle effects, possibly at the metal ion-dependent subunit interface.
INTRODUCTION
homology with ceruloplasmin and (2) the crystal structures of an inactivated form of bovine FVa lacking the A2 domain, FVai [11], and the human C2 domain [12], several theoretical models for intact human FVa have been generated [13–16]. Although these models do not agree on the specific amino acids involved in metal ion binding, there is consensus that an acidic A1 domain loop is involved in Ca2+ -binding. The loop encompasses Glu96 , Asp102 , Glu108 , Asp111 and Asp112 , and is predicted to be located at the A1/A3 domain interface. Most of the structural modelling studies also suggest that the bound Ca2+ and Cu2+ are located close to each other. Dissection of human FVa regions by plasmin-mediated fragmentation [17] and studies on mutant FV secreted into culture media [18] from our laboratory provided functional evidence that the amino acids spanning Glu96 –Asp111 contribute to the chelatorsensitive heterodimeric subunit complex. In particular, mutation of Asp111 to Ala appeared to destabilize the association of subunits after single-chain FV was activated to two-chain FVa. Mutation of any combination of amino acids within this region, not just acidic residues, had an inhibitory effect on function [14]. A study by Sorensen et al. [19] later demonstrated that a double mutation in the hypothesized Ca2+ -binding loop of FV (D111N/D112N) resulted in a rapid loss of activity upon treatment with thrombin. However, the direct involvement of this region with FV-bound metals has not been reported. To understand the importance of the predicted Ca2+ -binding loop in FV, we constructed and purified recombinant wild-type
In the presence of Ca2+ and an aPL (anionic phospholipid)containing membrane, coagulation FVa [activated FV (Factor V)] accelerates the activation of prothrombin to thrombin by the serine protease FXa (activated Factor X); this represents the critical step of clot generation [1]. Formation of the FVa–FXa–aPL–Ca2+ complex, termed prothrombinase, is dependent on the activation of FV, which circulates in plasma as a 330 kDa single chain procofactor [2]. FV is structurally and functionally homologous to FVIII (Factor VIII), sharing the domain structure NH2 -A1-A2B-A3-C1-C2-COOH [1,3]. Activation of FV by thrombin or FXa involves removal of the B domain [4]. The resulting FVa is a non-covalent heterodimer composed of a 105 kDa heavy chain (FVaH; domains A1-A2), and a 74/71 kDa light chain (FVaL: domains A3-C1-C2) [5], held together in the presence of divalent cations [6,7]. Although the FVaH and FVaL interaction is required for prothrombinase function [6], little is known about how the divalent cations precisely play their biochemical role. Bovine FV and FVa have a single high affinity site for Ca2+ [8], which must be occupied for FVa subunit association and consequent activity [7]. Ca2+ does not interact with either FVaH or FVaL alone [9], suggesting that the Ca2+ -binding pocket is formed by subunit association. FV also binds with high affinity to one Cu2+ atom per mole of protein [10], although an effect of Cu2+ binding on function has not yet been reported. Based on (1)
Key words: B domain, calcium, coagulation, copper, Factor V, subunit.
Abbreviations used: aPL, anionic phospholipid; FV, Factor V; FVa, activated FV; FVaH, FVa heavy chain; FVaL, FVa light chain; FVai , inactivated A2domainless bovine FVa; FV-810, wild-type human FV missing residues 811–1491; FVIII, Factor VIII; FXa, activated Factor X; PEG, poly(ethylene glycol); SUV, small unilamellar vesicle. 1 To whom correspondence should be addressed (email
[email protected]). c The Authors Journal compilation c 2009 Biochemical Society
258
J. Song and others
FV and mutants at Glu96 , Asp102 and Asp111 , resulting in the E96A, D102A and D111A variants. We report here for the first time that a single amino acid substitution, D111A, affects not only Ca2+ but also Cu2+ binding, consequently weakening inter-subunit stability and cofactor function. Compared with wild-type FV, E96A and D102A also have lower FVa activity. However, in contrast with D111A, the molecular basis of these reduced activities does not involve obvious changes to metal ion binding.
EXPERIMENTAL FV-810 and mutant expression and purification
FV-810 (wild-type human FV containing part of the B-domain amino acids 710–810, but missing amino acids 811–1491), was cloned into the pMT2 expression vector (pMT2-rFV) [4] and mutants were generated as described in [18]. For each mutant, two complementary oligonucleotides with the desired mutation were amplified using the following primers: 5 -GGTACAGTAAATTATCAGCCGGTGCTTCTTACCTTGACCAC-3 (E96A), 5 -AGGAGCTTCTTACCTTGCCCACACATTCCCTGCG-3 (D102A) and 5 -TCCCTGCGGAGAAGATGGCCGACGCTGTGGCTCCAG-3 (D111A). Transfection and conditioned media collection were performed as described [4]. Cell cultures were mycoplasmafree. All subsequent steps were carried out at 4 ◦C. The culture medium was concentrated and loaded onto a SP Sepharose Fast Flow column (Amersham) in 20 mM Mes buffer containing 0.1 M NaCl and 5 mM CaCl2 , pH 6.0. The column was eluted with a linear NaCl gradient from 0.1 to 0.6 M. Fractions containing FV activity were pooled, centrifuged, loaded onto Q Sepharose Fast Flow (Amersham) in HBS/Ca2+ (20 mM Hepes buffer containing 0.15 M NaCl and 2 mM CaCl2 , pH 7.4) and eluted with a linear NaCl gradient from 0.15 to 0.6 M. Fractions containing FV activity were pooled, centrifuged, concentrated and applied to a Superose 6 gel filtration column (Amersham) in HBS/Ca2+ . Fractions containing FV activity were concentrated and stored at − 80 ◦C.
Thrombin cleavage of FV-810 and mutants
FV-810 and mutants (200 nM) were incubated with thrombin (0.2 nM) in HBS/Ca2+ (5 mM) at 22 ◦C. The digests were stopped at various times by heating at 95 ◦C for 5 min in denaturing Laemmli sample buffer. Samples were subjected to SDS/PAGE and visualized by Coomassie Blue staining.
Binding of FV-810 and mutants to aPL
Binding curves using SUVs (small unilamellar vesicles) consisting of 25 % phosphatidylserine and 75 % phosphatidylcholine (20 μM) [20] incubated with various concentrations of FV-810 or mutants were obtained by light scattering analysis using a Varian Eclipse fluorescence spectrophotometer [21,22]. Scattering intensities were measured at 320 nm (excitation and emission) with a band pass of 5 nm at 22 ◦C. After the manual addition of FV-810 or each mutant, light scattering was followed for 5 min to ensure equilibration. The data were fitted to an equation assuming a single class of sites to derive apparent dissociation constants [K d (app)]. No detectable change in light scattering intensity was observed by SUV in the absence of FV, or by adding FV-810 or mutants without SUV. The changes in light scattering intensity due to dilution by continuous ligand titration were corrected. c The Authors Journal compilation c 2009 Biochemical Society
FV-810 and mutant prothrombinase activity
To evaluate prothrombinase function, saturating FXa (0.5 nM), SUV (50 μM), prothrombin (1.4 μM) and various concentrations of either FV-810 or mutants in HBS/Ca2+ with 0.1 % PEG [poly(ethylene glycol)] 8000 were incubated at 22 ◦C for 5 min. The reaction was terminated with EDTA (15 mM) and thrombin generation was quantified by S2238 (0.2 mM) cleavage using a Spectramax kinetic microplate reader (Molecular Devices) (1step assay). FV-810 or mutants (100 nM) were also evaluated after pretreatment with thrombin (5 nM) for 5 min, which was sufficient to cleave all of the FV to FVa (2-step assay). Hirudin (0.7 unit/ml) was added at an amount sufficient to inhibit only the thrombin in the pretreatment step. To evaluate prothrombinase function on activated platelets, various concentrations of intact or thrombinpretreated FV-810 or mutants were incubated with excess FXa (0.5 nM), activated platelets (5.4 × 104 /30 μl reaction volume) and prothrombin (1.4 μM) and thrombin generation was followed as above. The activated platelets were prepared by incubating fresh platelets (6 × 109 /ml) with A23187 (1 μM) for 10 min with gentle agitation. At the concentration of stimulated platelets used, the activity of FV released was minimal and was subtracted from all points. FV-810 and mutant clotting activity
Various concentrations of FV-810 or mutants (50 μl) in HBS/Ca2+ were pre-incubated with FV-deficient plasma (50 μl, Biopool) at 37 ◦C for 1 min in a 1-stage assay. Coagulation was initiated by the addition of Innovin (100 μl, Dade Behring) and the time to clot formation was monitored using an ST4 analyser (Diagnostica Stago). Conversion of FV-810 and mutants to FVa by thrombin was conducted as above in 2-stage assays and evaluated for clotting activity. The clotting times observed were converted into activity units using a human reference plasma standard curve assayed under identical conditions [23]. FV-810 and mutant subunit dissociation after thrombin cleavage
FVaH dissociation from FVaL bound to SUV was monitored by following a decrease in light scattering as previously reported [24] using a Varian Eclipse fluorimeter. To rapidly convert all of the SUV (20 μM)-bound FV-810 or mutants into FVa, a higher concentration of thrombin was used (100 nM). Thrombin-cleaved FV-810 and mutant cofactor activity upon chelation
FV-810 or mutants (100 nM) were treated with thrombin (100 nM). Within 1 min, EDTA (5 mM) was added and assayed for clotting activity at various times as above. Ca2+ and Cu2+ bound to FV-810 and mutants
Excess divalent metal ions were removed by extensive dialysis of FV-810 and mutants against HBS-Me (divalent cation-free HBS). HBS-Me was produced by treatment with Chelex-100 resin (Bio-Rad). Samples of FV were then dialysed against HBS-Me with added Ca2+ (100 nM) and Cu2+ (100 fM). Prior to analysis, samples were quantified by optical density and diluted 10-fold in 0.1 % metal-free HNO3. FV-810 and mutants were also pretreated with thrombin in HBS/Ca2+ before dialysis as above. All samples assayed for Ca2+ and Cu2+ were within the linear range of detection by a Varian SpectraAA 300 Zeeman graphite furnace atomic absorption spectrometer. Metal ion standard curves were
Metal ions in FVa subunit stability
Figure 1
259
Cleavage of FV-810 and mutants by thrombin
FV-810 (A), E96A (B), D102A (C) and D111A (D) (200 nM) were treated with thrombin (0.2 nM) in HBS/Ca2+ (5 mM) at 22 ◦C. At the indicated time (min), aliquots were taken from the mixture and the reaction was stopped with Laemmli sample buffer. Samples were separated by SDS/8 % PAGE and stained with Coomassie Blue.
Figure 3 Figure 2
Binding of FV-810 and mutants to aPL
FV-810 (), E96A (䊉), D102A (䉫) or D111A (䊏) was added to SUV (20 μM) and binding was monitored by light scattering analysis at 22 ◦C. Light scattering was measured approximately 10 s after each addition and binding kinetics were measured for 5 min. The intensities after binding equilibrated were plotted against the concentrations of FV-810 or mutants.
obtained with 0 μM, 0.0625 μM, 0.125 μM and 0.25 μM of Ca2+ or Cu2+ .
Activity of FV-810 and mutants
(A and B) The plasma clotting activity was measured of FV-810 or mutants by pre-incubating with FV-deficient plasma (50 μl) for 1 min in 37 ◦C. Innovin (100 μl) was then added and time to clot formation was monitored. (C and D) FV-dependent prothrombinase activity was measured in a reaction mixture containing FXa (0.5 nM), SUV (50 μM), prothrombin (1.4 μM) and various concentrations of FV-810 or mutants in HBS/Ca2+ (5 mM)/0.1 % PEG at 22 ◦C. The reaction was stopped after 5 min with EDTA (15 mM) and thrombin generation was quantified using S2238 (0.2 mM) and compared with a purified thrombin standard curve. (E and F) As in (C) and (D) respectively, except A23187-activated platelets (5.4 × 104 /30 μl of reaction volume) were used instead of SUV. (A, C and D) 1-Stage assays with no thrombin addition. (B, D and F) In 2-stage assays, FV-810 and mutants (100 nM) were pre-cleaved with thrombin (5 nM) for 5 min and stopped by adding hirudin (0.7 unit/ml). The activated samples were then diluted to the indicated concentrations. Symbols: FV-810 (), E96A (䊉), D102A (䉫) and D111A (䊏).
RESULTS Conversion of FV mutants into FVa by thrombin is normal
SDS/PAGE (Figure 1) showed that FV-810 and mutants were comparably converted within 15 min into FVaH (∼ 105 kDa) and FVaL (∼ 74 kDa) through the anticipated activation intermediate (150 kDa). These data demonstrated that the targeted substitutions did not significantly affect proteolysis of FV-810 by thrombin and established the experimental conditions required for complete proteolysis to FVa. Binding of FV mutants to aPL is normal
Figure 2 shows that the binding isotherms of the mutants were indistinguishable from FV-810 with derived K d (app) values of 1.4–1.6 × 10−8 M. These data demonstrated that mutation did not alter the capacity of the FV mutants to bind aPL.
FV mutant cofactor activity is inhibited
Both prothrombinase chromogenic and clotting assays were performed to evaluate the function of FV-810 and mutants (Figure 3) without (1-stage assay) or with (2-stage assay) thrombin pretreatment. The chromogenic experiments were conducted using either synthetic SUVs (Figures 3C and 3D) or activated platelets (Figures 3E and 3F) as a source of aPL for assembly of prothrombinase. The cofactor activity of each mutant was similarly inhibited by 20-40 % in 1-stage assays. Only D111A exhibited different activity when 2-stage prothrombinase and clotting assays were performed, which resulted in ∼ 90 % inhibition. Comparable results were obtained regardless of whether aPL or activated platelets were used in chromogenic assays, verifying contributions of Glu96 , Asp102 and Asp111 to FVa function when prothrombinase is assembled on a physiological membrane. c The Authors Journal compilation c 2009 Biochemical Society
260
J. Song and others
Figure 5
Chelator-induced inactivation of FV-810 and mutants
FV-810 (), E96A (䊉), D102A (䉫) or D111A (䊏) (100 nM) was bound to SUV (20 μM) in HBS/Ca2+ (5 mM) and cleaved by thrombin (100 nM) for 5 min. EDTA (5 mM) was then added at various time points and aliquots were taken and assayed for clotting activity. The rate of activity loss was fitted to a single phase exponential decay function (solid line).
of metal ions (Figure 4G). EDTA chelation did not result in further light scattering changes (Figure 4H), indicating that FVa subunits were already completely dissociated. These results clearly demonstrated that the reduction in cofactor activity of D111A is due to spontaneous subunit dissociation.
Chelator-mediated cofactor inhibition precedes subunit dissociation
Figure 4
Subunit stability of FV-810 and mutants after thrombin treatment
Subunit interaction of FV-810 or mutants (100 nM) bound to SUV (20 μM) in HBS/Ca2+ (5 mM) was monitored by light scattering analysis at 22 ◦C upon addition of thrombin (100 nM) to FV-810 (A), E96A (C), D102A (E) or D111A (G) and upon subsequent addition of EDTA (5 mM) to thrombin-cleaved FV-810 (B), E96A (D), D102A (F) or D111A (H). The light scattering profile was fitted to a single phase exponential decay function (solid line).
Having observed that the chelator-mediated FVa subunit dissociation of E96A, D102A and D111A is accelerated, we investigated the rate of cofactor activity loss under identical conditions in parallel. Clotting assays (Figure 5) revealed that FV-810, E96A, D102A and D111A had functional decay −1 −1 rates of 0.9 + 0.08 min−1 , 1.1 + − 0.07 min , 1.1 + − 0.09 min and −−1 respectively. As expected, the relative cofactor 3.1 + 0.1 min − activity of mutants at 0 min compared with FV-810 was inhibited (Figures 3 and 4). Interestingly, the loss of function for FV-810, E96A and D102A was markedly faster than their corresponding subunit dissociation rates (Figure 4), suggesting that metal ions may independently affect FVa function and subunit association. D111A clotting activity was negligible within 1 min of EDTA treatment. This observation was anticipated based on the spontaneous subunit dissociation observed by light scattering (Figure 4).
Subunit stability of FV mutants is differentially altered
Ca2+ and Cu2+ binding to only D111A is inhibited
When FV-810 and its mutants E96A and D102A were cleaved with thrombin in the presence of metal ions (Figures 4A, 4C and 4E), no subunit dissociation was detected, which is consistent with the similarity of 1-stage and 2-stage functional assays (Figure 3). Upon EDTA chelation, FV-810, E96A and D102A subunits dissociated with first-order rates of 0.026 + 0.0002 min−1 , − −1 −1 + and 0.041 0.0003 min respectively 0.051 + 0.0007 min − − (Figures 4B, 4D and 4F). Thus, upon chelation, E96A and D102A had moderately faster FVaH/FVaL dissociation compared with FVa-810. These observations from repeated experiments suggested that Glu96 and Asp102 contribute to the metal-dependent subunit interaction of FVa. In contrast to FV-810, E96A and D102A, the rapid conversion of FV D111A into FVa D111A by thrombin resulted in rapid −1 subunit dissociation (2.1 + − 0.05 min ) even in the presence
Graphite furnace atomic absorption spectrometry was used to compare the amount of Ca2+ (Figures 6A and 6B) and Cu2+ (Figures 6C and 6D) bound to FV-810 or mutants before (Figures 6A and 6C) and after (Figures 6B and 6D) cleavage by thrombin. Both non-cleaved and pre-cleaved FV-810 were shown to bind to Ca2+ and Cu2+ in a ∼ 1:1 molar ratio, which is consistent with the observations made by others [8,10]. E96A and D102A were identical to FV-810. A notable difference was observed for D111A, which had a significantly reduced number of moles of Ca2+ bound per mole of protein (no thrombin, 0.17:1; plus thrombin, 0.14:1). These observations quantitatively confirm the direct involvement of FV/FVa Asp111 in Ca2+ binding. The simultaneous measurement of Cu2+ revealed that mutation of Asp111 also reproducibly decreased its stoichiometry (no thrombin, 0.54:1; plus thrombin, 0.04:1). These data indicate a
c The Authors Journal compilation c 2009 Biochemical Society
Metal ions in FVa subunit stability Table 1
261
Predicted amino acids involved in human FVa metal ion binding
Reference
Model basis
Ca2+ -binding site(s)
Cu2+ -binding site
Villoutreix et al. [13] Pellequer et al. [14]
Human ceruloplasmin A domain homology [38] Human ceruloplasmin A domain homology [38] and human C2 domain crystal [12] Bovine FVai crystal Humanized bovine FVai A domains [11] and human C2 domain crystal [12] Humanized bovine FVai [11] and ceruloplasmin A domain homology [38]
(1) Glu96 , Glu108 , Asp102 , Asp111 , Asp112 (1) Glu148 , Asn149 , Asp1577 (2) Glu1572 , Asp1576 , Asp1583 (1) Lys93 , Glu108 , Asp111 , Asp112 (1) Lys93 , Glu108 , Asp111 , Asp112 (2) Asp1579 , Glu1572 , Asp1576 , Asp1583 (1) Lys93 , Glu108 , Asp111 , Asp112
(1) His85 , His1815
Adams et al. [11] Orban et al. [16] Lee et al. [15]
(1) His85 , His1815 , His1817 (1) His1815 , His1817 , Asn1857 * (1) His1815 , His1817 , Asn1857 , Glu1859 (1) His147 , Asn1857 , Thr1858 , Glu1859 , Gln1864
* Aspartate for homologous bovine residue.
Figure 6
Ca2+ and Cu2+ binding to FV-810 and mutants
Various concentrations of FV-810 (), E96A (䊉), D102A (䉫) or D111A (䊏) in HBS/Ca2+ (100 nM)/Cu2+ (100 fM) were diluted in 0.1 % HNO3 . The amount of Ca2+ (A) and Cu2+ (C) bound to FV-810 or mutants was measured by graphite furnace atomic absorption spectrometry. In order to assess metal ions bound to FVa, FV-810 or mutants (10 μM) was cleaved with thrombin (500 nM) for 5 min at 22 ◦C and diluted to the concentrations indicated. They were then dialysed against HBS/Ca2+ (100 nM)/Cu2+ (100 fM) and diluted in 0.1 % HNO3 . The amount of Ca2+ (B) and Cu2+ (D) bound to cleaved FV-810 or mutants was measured by graphite furnace atomic absorption spectrometry.
complicated FVa subunit interaction mechanism by suggesting that correct co-ordination of one metal ion may facilitate that of the other. DISCUSSION
Ca2+ -dependent interactions between the subunits of FVa are essential for stability and function [25], yet little is known about the facilitating amino acids. Five FVa models have been published based on: (i) the crystal structure of ceruloplasmin which shares homology with FVa A-type domains; (ii) the crystal structure of bovine FVai , which lacks the A2 domain; and (iii) the human recombinant C2 domain crystal structure. Shown in Table 1 are the amino acids predicted to co-ordinate with Ca2+ and Cu2+ for each model. None of these models completely agree on the amino acids directly involved in metal binding, adding to the uncertainty
of how metal ions facilitate FVa subunit interactions. Although Ca2+ or Cu2+ binding to amino acids in only the A1 domain or A3 domain or both have all been proposed, most models suggest that the highly conserved acidic residues spanning Glu96 to Asp112 in the A1 domain play a role in the Ca2+ interaction and that His1815 and His1817 in the A3 domain are involved in Cu2+ binding. None predict that residues between Glu96 and Asp112 are involved directly in the Cu2+ interaction. The models are supported in part by functional experiments using fragmented [17] and mutated [18,19] FVa, which demonstrated a role for the Glu96 –Asp112 segment in the chelatorsensitive subunit association. This region is represented in most of the models by a Ca2+ -binding loop and is depicted in Figure 7 (blue peptide backbone) according to the most recent model [15], which is based on both the humanized bovine FVai crystal and ceruloplasmin homology. Consistent with a functional role in mediating the FVaH–FVaL interaction, the loop is shown at the A1-A3 interface. Since Ca2+ and Cu2+ are proximal to each other in this model, it is plausible that both could affect subunit interactions by influencing Glu96 –Asp112 configuration relative to the A3 domain. Compared to other amino acids in the Ca2+ -binding loop, our preliminary work using recombinant FV secreted into tissue culture medium suggested that single point mutations of Glu96 , Asp102 and Asp111 (Figure 7; wireframe) to alanine had the greatest effect on FVa function [18]. To understand further how Ca2+ and Cu2+ facilitate FVa activity, we correlated metal ion binding to purified E96A, D102A and D111A with function in the present study. The parental FV-810 produces wild-type FV but has a truncated B domain that does not require proteolytic activation for cofactor function [4]. This unique reagent enabled us to investigate the Ca2+ dependence of the intra-molecular FVaH–FVaL domain interaction, which is thought to be fully formed prior to excision of the B domain [4]. Figure 8(A) summarizes our findings. Consistent with Figure 7, the proposed Ca2+ -binding loop [11] is shown projecting from the FVaH domain within FV-810, and association with the FVaL domain and Cu2+ is shown co-ordinating with both the FVaH and FVaL domains. Graphite furnace atomic absorption spectroscopy revealed a stoichiometry of 1:1 FV-810:Ca2+ , confirming previous literature for the high-affinity binding site [8]. The same molar ratio was derived for Cu2+ , which was also reported previously [10]. Of the mutations we studied, altering Asp111 had the most profound effect on FV/Va function. As predicted by the structural model (Figure 7), the interaction of D111A with Ca2+ was severely inhibited, suggesting that Asp111 provides a pivotal contact with Ca2+ . This conclusion is supported by most of the FVa models (Table 1) and is represented in Figure 8(B) as FV with no Ca2+ -binding loop. Interestingly, D111A also had impaired Cu2+ c The Authors Journal compilation c 2009 Biochemical Society
262
J. Song and others
Figure 7
Structural model of FVa metal binding
2+
The Ca - and Cu2+ -binding sites of human FVa are shown based on the model proposed by Lee et al. [15]. Glu96 , Asp102 , Asp111 and His147 in standard CPK (Corey–Pauling–Koltun) colour are depicted with side chains. Only the peptide backbone structure is shown for the rest of the molecule. The domains are coloured uniquely and the acidic loop spanning Glu96 to Asp112 is coloured blue. The inset shows the entire FVa structure, identifying the enlarged region (boxed). The Figure was produced using Rasmol2 software.
Figure 8 Effects of FV Glu96 , Asp102 and Asp111 mutation on metal binding, subunit association and function (A) Before and after cleavage by thrombin, FV-810, E96A and D102A are shown to bind one mole of Ca2+ and Cu2+ as previously reported [8,10]. Upon chelation, thrombin-cleaved FV-810, E96A and D102A rapidly result in an inhibited heterodimeric species, which slowly dissociates. (B) In contrast, D111A has no Ca2+ and partially inhibited Cu2+ binding, leading to spontaneous subunit dissociation and complete loss of Cu2+ binding after cleavage by thrombin.
binding, and yet the site was predicted to be unrelated to Asp111 according to the bovine FVai crystal structure and the earlier ceruloplasmin homology model [11,13]. Although we cannot exclude the possibility of direct contact between Asp111 and Cu2+ at the present time, the further refined model that combines the crystal and homology structures (Figure 7) suggests that Ca2+ co-ordination with Asp111 is linked indirectly to Cu2+ binding through His147 (wireframe). By using our methodology, the Cu2+ c The Authors Journal compilation c 2009 Biochemical Society
stoichiometry is approx. 50 % and is consequently represented in Figure 8(B) as having a partial binding site. Thus, altered Ca2+ and Cu2+ binding to intact D111A provides an explanation for the inhibited activity in 1-stage prothrombinase and clotting assays (∼ 40 % of FV-810). After excision of the partial B domain by thrombin, light scattering experiments demonstrated that, unlike the FVa derived from FV-810 (FVa-810), which is stable in the presence of metal ions, the FVaH/FVaL subunits derived from D111A undergo spontaneous dissociation. Although a single Ca2+ and Cu2+ bound to FVa-810, thrombin cleavage of D111A resulted in the complete loss of bound Cu2+ in addition to a negligible Ca2+ stoichiometry that was also observed for uncleaved D111A. The finding that a single amino acid substitution affects both Ca2+ and Cu2+ binding may imply that interdependent metal ion-co-ordinating sites contribute to stabilize the non-covalent association of subunits. Since spontaneous subunit dissociation occurs after thrombin cleavage of D111A, the concomitant reduction observed for Cu2+ binding also suggests that FVaH/FVaL interactions may be required to form the Cu2+ -co-ordinating site. Combined with a similar conclusion made for Ca2+ binding [9,26], FVa subunit association may involve an intricate mechanism in which Ca2+ , Cu2+ , FVaH and FVaL interactions are all interdependent. Given the paradigm that Ca2+ is essential for FVa activity [27] and that the D111A variant of FV-810 has negligible Ca2+ -binding, the finding that this 60 %-active mutant retained significant cofactor function was unexpected. These results suggest that in addition to masking FV cofactor activity [4], the B domain facilitates a functional configuration between the FVaH and FVaL domains even in the absence of proper metal ion co-ordination (Figure 8B). After this stabilizing effect of the B domain is excised from D111A, the importance of metals is highlighted by complete inhibition and rapid dissociation of the non-covalent subunits. To evaluate the multi-functional role of the Ca2+ -binding loop implied by our earlier preliminary studies [18], we investigated the effect of mutating Glu96 and Asp102 to alanine on FV/FVa function. Unlike D111A, the stoichiometry for Ca2+ and Cu2+ bound to E96A or D102A was shown here to be identical to FV-810. E96A and D102A nevertheless had reduced cofactor activity. The structural model (Figure 7) indicates that both or either could contribute directly to inter-subunit contact. Although the interaction of FVaH and FVaL derived from thrombin-cleaved E96A and D102A was stable in the presence of metal ions, mutation moderately enhanced chelator-mediated dissociation. The close proximity of Glu96 and Asp102 to the probable Ca2+ binding site may thus confer a subtle effect on metal ion coordination that enhances the susceptibility to chelation and subunit dissociation due to EDTA. For human FVIII, Cu2+ enhances the affinity between subunits [28]. Ca2+ binds to a region in FVIII nearly identical to the Ca2+ -binding loop of FV [29,30] and results in conformational changes that probably participate in subunit association and in the formation of active cofactor. Roles for Cu2+ have not been reported for human FVa subunit association and function. For bovine FVa, Ca2+ alone appears to facilitate reassociation of isolated subunits, giving rise to full function [7]. In our studies, it was not possible to completely remove Cu2+ bound to FV-810 by extensive dialysis against a chelating resin (results not shown). Therefore, the involvement of Cu2+ in FVaH–FVaL association and function of human FVa cannot be excluded at this time and may be especially difficult to assess if indeed there is metal ion interdependent binding in FVa as hypothesized here. The chelator-induced subunit dissociation rate measured quantitatively for FVa-810 in the current work (t 12 ≈ 35 min) was
Metal ions in FVa subunit stability
comparable with that previously reported for human FVa [18]. Interestingly, this was approx. 10-fold slower than the rate of activity loss we measured, which is similar to several previously published observations using analogous methods in solution [6,8,31]. Although these reports differ from results acquired using a unique solid-phase activity assay [17], the preponderance of evidence in the literature, combined with the results presented in this study, suggest that FVa cofactor function is lost due to chelation prior to subunit dissociation (Figure 8A). A similar mechanism was proposed for FVIII by Fay and co-workers, who observed formation of inactive heterodimer in the absence of metal ions [28]. Only in the presence of both Ca2+ and Cu2+ did these authors observe a high FVIII-specific activity. Further experimentation is required to confirm that metal ion binding to FVa may not only contribute to subunit association but also to distinctly induce a functional conformation. Although extremely rare, X-linked FVIII deficiency is approximately two orders of magnitude more prevalent than the homozygous deficiency of FV. A relatively large number of mutations associated with FVIII deficiency have consequently been identified and have enabled correlations between individual amino acids and function [17]. Patients carrying substitutions in the probable Ca2+ -binding loop in FV have not been reported. However, homologous mutations have been identified in FVIII, which result in haemophilia A [32–35]. In the present study, coagulation assays revealed nearly 100 % inhibition after cleavage of D111A by thrombin, in contrast with only a partial loss of activity without thrombin pretreatment. To our knowledge, this 1-stage/2-stage assay discrepancy has not been reported clinically for FV but has been shown in naturally occurring haemophilia A mutations [36,37]. This raises the possibility that FV mutation resulting in impaired metal binding or subunit interface contact may be a previously overlooked basis of pathology.
AUTHOR CONTRIBUTION Jina Song designed and conducted experiments and wrote the first manuscript draft as part of a doctoral program. Kimberley Talbot conducted initial experiments and mutant preparation and assisted with manuscript revision. Jeffrey Hewitt was instrumental in mutant preparation. Ross MacGillivray provided project guidance, critical data analysis and key manuscript revisions. Edward Pryzdial directed the project and wrote the manuscript with Jina Song.
ACKNOWLEDGEMENTS We are grateful to Dr Rodney Camire (University of Pennsylvania) for critical review of the manuscript prior to submission and for providing pMT2-rFV. Dr. Lee Pedersen (University of North Carolina) is thanked for helpful discussion and providing the atomic co-ordinates for the FVa structural model. Dr Dominique Weis and Bert Mueller (Pacific Centre for Isotope and Geochemical Research, University of British Columbia) are acknowledged for use of the graphite furnace atomic absorption spectrometer and technical training. Dr Michael Murphy (University of British Columbia) and Dr William Sheffield (McMaster University) provided us with insightful comments.
FUNDING This work was supported by the Heart and Stroke Foundation of British Columbia and a Yukon Grant in Aid (to E. L. G. P.). We acknowledge the use of equipment funded by the Canadian Foundation for Innovation and administrative support provided by the Michael Smith Foundation awarded to the University of British Columbia Centre for Blood Research.
REFERENCES 1 Kalafatis, M., Egan, J. O., van’t Veer, C. and Mann, K. G. (1997) The regulation of clotting factors. Crit. Rev. Eukaryot. Gene Expr. 7, 241–180 2 Krishnaswamy, S., Nesheim, M. E., Pryzdial, E. L. and Mann, K. G. (1993) Assembly of prothrombinase complex. Methods Enzymol. 222, 260–280
263
3 Kane, W. H. and Davie, E. W. (1988) Blood coagulation factors V and VIII: structural and functional similarities and their relationship to hemorrhagic and thrombotic disorders. Blood 71, 539–555 4 Toso, R. and Camire R. M. (2004) Removal of B-domain sequences from factor V rather than specific proteolysis underlies the mechanism by which cofactor function is realized. J. Biol. Chem. 21643–21650 5 Suzuki, K., Dahlback, B. and Stenflo, J. (1982) Thrombin-catalyzed activation of human coagulation factor V. J. Biol. Chem. 257, 6556–6564 6 Esmon, C. T. (1979) The subunit structure of thrombin-activated factor V. Isolation of activated factor V, separation of subunits, and reconstitution of biological activity. J. Biol. Chem. 254, 964–973 7 Krishnaswamy, S., Russell, G. D. and Mann, K. G. (1989) The reassociation of factor Va from its isolated subunits. J. Biol. Chem. 264, 3160–3168 8 Hibbard, L. S. and Mann, K. G. (1980) The calcium-binding properties of bovine factor V. J. Biol. Chem. 255, 638–645 9 Guinto, E. R. and Esmon, C. T. (1982) Formation of a calcium-binding site on bovine activated factor V following recombination of the isolated subunits. J. Biol. Chem. 257, 10038–10043 10 Mann, K. G., Lawler, C. M., Vehar, G. A. and Church, W. R. (1984) Coagulation Factor V contains copper ion. J. Biol. Chem. 259, 12949–12951 11 Adams, T. E., Hockin, M. F., Mann, K. G. and Everse, S. J. (2004) The crystal structure of activated protein C-inactivated bovine factor Va: Implications for cofactor function. Proc. Nat. Acad. Sci. U.S.A. 101, 8918–8923 12 Macedo-Ribeiro, S., Bode, W., Huber, R., Quinn-Allen, M. A., Kim, S. W., Ortel, T. L., Bourenkov, G. P., Bartunik, H. D., Stubbs, M. T., Kane, W. H and Fuentes-Prior, P. (1999) Crystal structures of the membrane-binding C2 domain of human coagulation factor V. Nature 402, 434–439 13 Villoutreix, B. O. and Dahlback, B. (1998) Structural investigation of the A domains of human blood coagulation factor V by molecular modeling. Prot. Sci. 7, 1317–1325 14 Pellequer, J.-L., Gale, A. J., Getzoff, E. D. and Griffin, J. H. (2000) Three-dimensional model of coagulation factor Va bound to activated protein C. Thromb. Haemost. 84, 849–857 15 Lee, C. J., Lin, P., Chandrasekaran, V., Duke, R. E., Everse, S. J., Perera, L. and Pedersen, L. G. (2007) Proposed structure models of human factor Va and prothrombinase. J. Thromb. Haemost. 6, 83–89 16 Orban, T., Kalafatis, M. and Gogonea, V. (2005) Completed three-dimensional model of human coagulation factor Va. Molecular dynamics simulations and structural analyses. Biochemistry 44, 13082–13090 17 Zeibdawi, A. R. and Pryzdial, E. L. G. (2001) Mechanism of FVa inactivation by plasmin: Loss of A2 and A3 domains from a Ca2+ -dependent complex of fragments bound to phospholipid. J. Biol. Chem. 276 (23), 19929–19936 18 Zeibdawi, A. R., Grundy, J. E., Lasia, B. and Pryzdial, E. L. G. (2004) Coagulation factor Va Glu96-Asp111:A chelator-sensitive site involved in function and subunit association. Biochem. J. 377, 141–148 19 Sorensen, K. W., Nicolaes, G. A., Villoutreix, B. O., Yamazaki, T., Tans, G., Rosing, J. and Dahlback, B. (2004) Functional properties of recombinant factor V mutated in a potential calcium-binding site. Biochemistry 43, 5803–5810 20 Grundy, J. E., Lavigne, N., Hirama, T., MacKenzie, C. R. and Pryzdial, E. L. G. (2001) Binding of plasminogen and tissue plasminogen activator to plasmin-modulated factor X and factor Xa. Biochemistry 40, 6293–6302 21 Bloom, J. W., Nesheim, M. E. and Mann, K. G. (1979) Phospholipid-binding properties of bovine factor V and factor Va. Biochemistry 18, 4419–4425 22 Higgins, D. L. and Mann, K. G. (1983) The interaction of bovine factor V and factor V-derived peptides with phospholipid vesicles. J. Biol. Chem. 258, 6503–6508 23 Hockin, M. F., Jones, K. C., Everse, S. J. and Mann, K. G. (2002) A model for the stoichiometric regulation of blood coagulation. J. Biol. Chem. 277, 18322–18333 24 Mann, K. G., Hockin, M. F., Begin, K. J. and Kalafatis, M. (1997) Activated protein C cleavage of factor Va leads to dissociation of the A2 domain. J. Biol. Chem. 272, 20678–20683 25 Davie, E. W., Fujikawa, K. and Kisiel, W. (1991) The coagulation cascade: initiation, maintenance, and regulation. Biochemistry 30, 1030–1037 26 Laue, T. M., Lu, R., Krieg, U. C., Esmon, C. T. and Johnson, A. E. (1989) Ca2+ -dependent structural changes in bovine blood coagulation factor Va and its subunits. Biochemistry 28, 4762–4771 27 Mann, K. G. and Kalafatis, M. (2003) Factor V: a combination of Dr Jekyll and Mr Hyde. Blood 101, 20–30 28 Wakabayashi, H., Koszelak, M. E., Mastri, M. and Fay, P. J. (2001) Metal ion-independent association of factor VIII subunits and the roles of calcium and copper ions for cofactor activity and inter-subunit affinity. Biochemistry 40, 10293–10300 29 Wakabayashi, H., Freas, J., Zhou, Q. and Fay, P. J. (2004) Residues 110–126 in the A1 domain of factor VIII contain a Ca2+ binding site required for cofactor activity. J. Biol. Chem. 279, 12677–12684 c The Authors Journal compilation c 2009 Biochemical Society
264
J. Song and others
30 Shen, B. W., Spiegel, P. C., Chang, C. H., Huh, J. W., Lee, J. S., Kim, J., Kim, Y. H. and Stoddard, B. L. (2008) The tertiary structure and domain organization of coagulation factor VIII. Blood 111, 1240–1247 31 Walker, F. J. (1992) Characterization of the interaction between the heavy and light chains of bovine factor Va. J. Biol. Chem. 267, 19896–19900 32 Maugard, C., Tuffery, S., Aguilar-Martinez, P., Schved, J. F., Gris, J. C., Demaille, J. and Claustres, M. (1998) Protein truncation test: detection of severe haemophilia A mutation and analysis of factor VIII transcripts. Hum. Mutat. 11, 18–22 33 Repesse, Y., Slaoui, M., Ferrandiz, D., Gautier, P., Costa, C., Costa, J. M., Lavergne, J. M. and Borel-Derlon, A. (2007) Factor VIII (FVIII) gene mutations in 120 patients with hemophilia A: detection of 26 novel mutations and correlation with FVIII inhibitor development. J. Thromb. Haemost. 5, 1476 34 Theophilus, B. D. M., Enayat, M. S., Williams, M. D. and Hill, G. H. (2001) Site and type of mutations in the factor VIII gene in patients and carriers of Hemolphilia A. Haemophilia 7, 381–391 Received 9 March 2009/6 May 2009; accepted 12 June 2009 Published as BJ Immediate Publication 12 June 2009, doi:10.1042/BJ20090405
c The Authors Journal compilation c 2009 Biochemical Society
35 Liu, M. L., Murphy, M. E. P. and Thompson, A. R. (1998) A domain mutations in 65 haemophilia A families and molecular modelling of dysfunctional factor VIII proteins. Brit. J. Haematol. 103, 1051–1060 36 Pipe, S. W., Saenko, E. L., Eickhorst, A. N., Kemball-Cook, G. and Kaufman, R. J. (2001) Hemophilia A mutations associated with 1-stage/2-stage activity discrepancy disrupt protein-protein interactions within the triplicated A domains of thrombin-activated factor VIIIa. Blood 97, 685–691 37 Hakeos, W. H., Miao, H., Sirachainan, N., Kemball-Cook, G., Saenko, E. L., Kaufman, R. J. and Pipe, S. W. (2002) Hemophilla A mutations within the factor VIII A2-A3 subunit interface destabilize factor VIIIa and cause one-stage/two-stage activity discrepancy. Thromb. Haemost. 88, 781–787 38 Zaitseva, I., Zaitsev, V., Card, G., Moshkov, K., Bax, B., Ralph, A. and Lindley, P. (1996) The X-ray structure of human serum ceruloplasmin at 3.1 A˚: nature of the copper centres. J. Biol. Inorg. Chem. 1, 15–23
