- Email: [email protected]
[9] Characterization of dysfunctional factor VIII molecules
[9]
FACTORVIII CHARACTERIZATION
169
increases the sensitivity of this method. 6s This approach has distinguished a polymorphism in the Asian population in the first intron at base 192. 69 Acknowledgments Supported in part by the National Institutes of Health (HL-31193) and the March of Dimes Birth DefectsFoundation (6-463). ~sL. S. Lerman and K. Silverstein,this series, Vol. 155, p. 482. 69M. Tanimoto, M. Hamaguchi, T. Matsushita, M. Hamaguchi, T. Matsushita, K. Yamamoto, I. Sugiura,J. Takamatsu, and H. Saito, Blood74, 251a (1989).
[9] C h a r a c t e r i z a t i o n
of Dysfunctional Molecules
Factor
VIII
By LEoNW. HOYER Introduction Factor VIII (antihemophilic factor) participates in blood coagulation as an essential cofactor in the activation of factor X by factor IXa in the presence of phospholipid and calcium. It circulates in plasma as a large glycoprotein that is associated with von WiUebrand factor in a noncovalent complex. The mature single-chain protein of 2332 amino acids is proteolytically processed before entering the circulation as a heterodimer consisting of amino-terminal heavy chain polypeptides of 92 to 200 kDa and an 80-kDa carboxyl-terminal light chain. Factor VIII procoagulant activity is generated by thrombin cleavage of the heavy chain at Arg-372 and Arg-740 to yield 54- and 44-kDa fragments, and thrombin cleavage of the 80-kDa chain at Arg-1689 to yield a 72-kDa fragment (Fig. 1). 1 Hemophilia A (classic hemophilia) is an X chromosome-linked disorder of blood coagulation caused by deficient factor VIII activity. Although factor VIII procoagulant activity is consistently reduced or absent, hemophilia A is heterogeneous when characterized dinicaUy or immunologically,2 as well as when the specific molecular defects are identified) Dysfunctional, immunoreactive factor VIII-like protein can be detected in 1W. H. Kane and E. W. Davie, Blood71, 539 (1988). 2 L. W. Hoyer and R. T. Breckenridge,Blood32, 962 (1968). 3 E. G. D. Tuddenham, D. N. Cooper, J. Gitschier,M. I-Iiguchi,L. W. Hoyer, A. Yoshioka, I. R. Peake, R. Schwaab, K. Olek, H. H. Kazazian,Jr., J.-M. Lavergne,F. GianneUi,and S. E. Antonarakis,NucleicAcids Res. 19, 4821 (1991). METHODS IN ENZYMOLOGY, VOL. 222
Copyright © 1993 by Academic Press, Inc. All fights of reproduction in any form reserved.
170
MAMMALIAN COAGULATION FACTORS AND INHIBITORS 372
740
[9]
1669
8O
/
• C8
MAB038
! J 16D-9
FIG. 1. Schematic representation of human factor VIII showing its domain structure (A 1- A3, B, C 1, and C2) and three thrombin cleavagesites (arrows). Below this arc shown the thrombin cleavage fragments. The fragment mass values in kilodaltons correspond to those published by C. A. Fulchcr, J. R. Roberts, and T. S. Zimmcrrnan [Blood61, 807 (1983)]. The solid boxes represent the regions containing epitopes for the monoclonal antibodies used to identify factor VIII fragments. From Arai et al. ~ with permission.
10% of plasmas from patients with mild or moderate hemophilia A.4 These plasmas are termed cross-reacting material (CRM) positive. The remaining hemophilic plasmas are either CRM reduced, in which case the reduction in immunologicaUy detectable protein is comparable to reduction in functional activity, or CRM negative, i.e., there is no detectable factor VIII by sensitive immunoassays. Although the factor VIII gene is large (186 kb), and unrelated individuals have different mutations, there has been considerable progress in identifying the causative point mutations or deletions) Until recently, the specific point mutation or deletion could only be identified in half of the patients with severe hemophilia A using PCR amplification of genomic DNA followed by denaturing gradient gel electrophoresis.5 Rapid screening of all essential regions of the factor VIII gene using an mRNA-based method has been more successful, however, as it also identified the presence of frequent intron 22 mutations that cause defective joining of exons 22 and 23 in the mRNA. 6 Because these studies do not necessarily establish the basis for reduced procoagulant activity, we have developed a simple and sensitive method for the immunoisolation of factor VIII from plasma. ~ This method has been used to identify the functional defect in the factor VIii-like protein recovered from plasmas of patients with CRM-positive hemophilia A. 4 j. Lazarchick and L. W. Hoyer, J. Clin. Invest. 62, 1048 (1978). M. Higuchi, H. H. I(aTa7ian, Jr., L. Kasch, T. C. Warren, M. J. McGinniss, J. A. Phillips, III, C. K. Kasper, R. Janco, and S. E. Antonarakis, Proc. Natl..4cad. Sci. U.S.A. 88, 7405 (1991). 6 j. A. Naylor, P. M. Green, C. R. 1~i77a~ and F. Giannelli, Lancet 340, 1066 (1992). 7 M. Arai, H. Inaba, M. Higuchi, S. E. Antonarakis, H. H. I~aTaTian, Jr., M. Fujimaki, and L. W. Hoyer, Proc. Natl. Acad. Sci. U.S.A. 86, 4277 (1989).
[9]
FACTORVIII CHARACTERIZATION
171
Procedure
Preparation of lmmunoadsorbent Most studies have been carried out using IgG prepared by the caprylic acid methods from the plasma of a patient with a high-titer (3600 Bethesda units) factor VIII autoantibody. The IgG is coupled to cyanogen bromideactivated Sepharose CL-2B at a concentration of 4 mg/ml of settled gel volume.9 Three other high-titer (570-3000 Bethesda units) human antifactor VIII antibodies (as well as autoantibodies) and a murine monoclonal antifactor VIII heavy chain antibody have also been successfully used as the immunoadsorbent when coupled to agarose beads.
Immunopurification of Factor VIII After washing, the immunoadsorbent beads are suspended in Tris-buffered saline (TBS): 0.15 M NaC1, 4 m M CaCI2, 20 m M Tris-HC1, pH 7.4. Prior to characterization, normal or hemophilia A patient plasmas are brought to 0.8 M NaC1 and 0.5% Tween 80, and 0.6-2.5 ml is incubated with 20 #1 of immunobeads overnight at room temperature with gentle aggitation. The amount of factor VIII protein immunoabsorbed on the beads can be estimated by measuring the difference between the factor VIII antigen (VIII: Ag) in the plasma and in the Oostadsorption supematant fluid? The beads are then washed into a 0.8 X 4-era polypropylene column (Bio-Rad, Roekville Center, NY) using 20 ml of 50 m M imidazole, 40 m M CaC12, 5% (v/v) ethylene glycol, 0.5% (v/v) Tween 80, pH 6.4, and further washed with 5 ml TBS. The beads are then transferred into a 1.0-ml conical polystyrene tube, the wash buffer removed by centrifugation, and factor VIII eluted by adding 40/tl of 0.125 M Tris-HC1, pH 6.8, containing 20 mg/ml sodium dodecyl sulfate (SDS), 10% (v/v) glycerol, and 0.05 mg/ml bromphenol blue. After 1 hr incubation at 37 ° with gentle agitation, the tube is centrifuged and the eluted factor VIII is recovered in the supernatant.
Thrombin Treatment of Factor VIII Some aliquots of immunopurified factor VIII were treated with thrombin before the SDS elution step. Thrombin, diluted to 10 units/ml with TBS, is added to these aliquots of washed factor VIII immunoadsorbent beads in sufficient quantity to achieve the desired concentration (0.2-2 s M. Steinbuch and R. Audran, Arch. Biochem. Biophys. 134, 279 (1969). 9 R. Ax~n, J. Porfith, and S. Ernbick, Nature (London) 214, 1302 (1967).
172
MAMMALIAN COAGULATION FACTORS AND INHIBITORS
[9]
units of thrombin/unit VIII:Ag) and the mixture is incubated at room temperature for 0.5 to 60 min with gentle agitation before adding the SDS elution buffer.
Immunoblotting of Factor VIII Eluted factor VIII (40/A) is analyzed by discontinuous SDS-polyacrylamide gel electrophoresis (PAGE) using 1.5-ram-thick 5 - 12% (w/v) polyacrylamide slab gels7 and is transferred to nitrocellulose sheets (Schleicher & Schuell, Inc. Keene, NH) using a Bio-Rad Tram-Blot cell. After transfer, the nitrocellulose sheet is blocked by incubation at 37 ° for 1 hr with borate-buffered saline (BBS) pH 7.8,4 containing 30 mg/ml bovine serum albumin (BSA) and 0.05% (v/v) Tween 20. The sheet is then incubated with antifactor VIII monoclonal antibodies (1/zg IgG/ml) in BBS containing 10 mg/ml bovine serum albumin and 0.05% (v/v) Tween 20 (dilution buffer). After overnight incubation at room temperature, the sheet is washed three times with BBS containing 0.05% (v/v) Tween 20 (washing buffer) and then incubated with tz~I-labeled affinity-purified sheep antimouse immunoglobulin in dilution buffer (80,000 cpm/ml). After a 4-hr incubation at room temperature, the sheet is again washed three times with washing buffer, dried, and visualized by autoradiography at - 7 0 ° for 48 to 72 hr using Kodak XK-I film (Eastman Kodak Co., Rochester, NY). Comparable results have been obtained using chemiluminescence detection. ~° In this case, alkaline phosphatase-streptavidin and biotin-labeled goat antibody to mouse IgG are used to detect the monoclonal antibodies, and the enzyme complex is identified with Lumi-Phos 530 (Lumigen, Inc., Detroit, MI).
Amplification of Genomic DNA and Sequencing Sequence analysis of apparent thrombin cleavage site mutations is carried out with leukocyte DNA amplified using Taq DNA polymerase (Perkin-Elmer Cetus, Norwalk, CT). H The primers are chosen so that the specific factor VIII cleavage site region will be amplified.7,~2 Polymerase chain reaction (PCR) is performed in a 100-/zl volume containing 200-400 ng of genomic DNA, 400 nM of each PCR primer, 200 # M of each dNTP, and 2 units of Taq DNA polymerase in a buffer to A. P. Schaap, H. Akhavan, and L. J. Romano, Clin. Chem. (Winston-Salem, N.C.) 35, 1863 (1989). tt R. K. Saiki, D. H. G¢lfand, S. Stoffd, et al., Science 239, 487 (1988). t2 M. Arai, M. Higuchi, S. E. Antonarakis, et al., Blood75, 384 (1990).
[9]
FACTORVIII CHARACTERIZATION
173
composed of I0 m M Tris-HCl (pH 8.3), 50 mMKC1, 1.5 mMMgCI2, and 0.02% gelatin. Each of the 35 cycles consists of a 30-see denaturation at 94 o, thermal transition from 94 ° to the annealing temperature over 2 min, 45 sec of reannealing at 52-55 °, and a 90-see extension at 72 ° using a DNA Thermal Cycler (Perkin-Elmer Cetus, Norwalk, CT). Amplified DNA is desalted and excess dNTPs are removed by spin dialysis on a Centricon 30 (Amicon, Danvers, MA). The purified DNA is sequenced directly: 10 ng of SP7 sequencing primer (5'-GAT TTT GAC ATT TAT GAT-Y), end labeled with [?-32p]ATP using T4 polynucleotide kinase, is annealed with 80 ng of PCR product on ice after heat denaturation at 95 ° for 5 min. The reaction mixture is divided into four tubes containing 62/~M unlabeled dNTPs, 6.2/tM dideoxy (dd)NTPs, and 2 units of T7 DNA polymerase (Sequenase USB, Cleveland, OH) in the sequencing reaction buffer [25 m M Tris-HC1, pH 7.5, 10 m M MgCI2, 70 m M NaC1, 7 m M dithiothreitol (DTT)]. After incubation at 370C for 15 min, the reaction is stopped with 3/tl of a solution containing 95% formamide, 20 m M EDTA, 0.5 mg/ml bromphenol blue, 0.05% xylene cyanol FF. Samples are boiled for 3 min and electrophoresed in a 6% polyacrylamide/ 8 M urea gel at 58 W for 2 hr. Gels are then dried and exposed to Kodak XAR-5 film (Eastman Kodak Co., Rochester, NY) for 16 hr.
Denaturing Gradient Gel ElectrophoresisAnalysis of Amplified Genomie DNA For evaluation of factor VIII mutations that do not prevent thrombin cleavage, the initial evaluation is carried out using denaturing gradient gel electrophoresis (DGGE) of DNA fragments to which are attached a "GCclamp". 5 PCR products ( - 80 ng of each) from two patients are combined to form heteroduplexes. After heat denaturation at 95 ° for 5 rain, the DNA solution is slowly cooled to room temperature (> 30 rain) and subjected to DGGE under conditions determined empirically for each PCR product. 5 DNA is loaded onto a 6.5% polyacrylamide gel (14 × 19 cm, 0.75-ram thick) containing a linear gradient of denaturants and electrophoresed at 2 - 4 V/cm for 16-23 hr. The gradient difference in denaturants is 20% [100% denaturants = 7 M urea/40% (v/v) formamide]. Gels are then stained in ethidium bromide and photographed with a UV transilluminator. Purified GC-clamped PCR products that show abnormal migrating patterns on DGGE are then directly sequenced. 6,11,13
~3C. Wong, C. D. Dowling, R. K. Saiki, R. G. Higuchi, H. A. Erlich, and H. H. Kazazian, Jr., Nature (London) 330, 384 (1987).
174
MAMMALIAN COAGULATION FACTORS AND INHIBITORS
[9]
Results
Immuno&olation of Factor VIII Protein Factor VIII is immunoadsorbed from plasma using a human antibody to factor VIII that reacts with both heavy and light chain determinants.14 The adsorbent removes 65-95% of VIII:Ag in normal or CRM-positive plasmas. Because the interaction of human factor VIII with these antibodies is very avid, usual elution techniques such as low pH and chaotropic agents are rarely successful in separating the antifactor VIII from an immunoadsorbent. ~s For this reason, and because pH extremes destroy factor VIII immunogenicity, SDS is used to separate the bound factor VIII from the immunoadsorbent. Protein characterization is then accomplished by SDS-PAGE, followed by immunoblotting using monoelonal antifactor VIII. In general, human antifactor VIII antibodies are not satisfactory for this purpose because these plasmas have additional immunologic reactivities not related to factor VIII. Three monoclonal antibodies have been used in most studies: MAB038 (Chemicon, E1 Segundo, CA) binds to the 80kDa light chain fragment but does not react with the thrombin-cleaved 72-kDa light chain fragment; J16D-9 [an antibody prepared by Dr. C. Fulcher (Scripps Research Institute, La Jolla, CA) by immunization with a synthetic peptide containing factor VIII residues 2318-2332] detects both 80- and 72-kDa light chain fragments; and C8 (JR Scientific, Woodland, CA) detects factor VIII heavy chain fragments as large as 200 kDa and reacts strongly with the 92- and 44-kDa thrombin-cleaved heavy chain fragments (Fig. 1).7 The sensitivity of the method is such that both heavy and light chain factor VIII determinants can be detected when at least 0.5 units of VIII: Ag are present in a 20-ml plasma sample.
Characterization of CRM-Positive Hemophilic Plasmas In most cases, the monoclonal antibodies are used together to characterize the immunoadsorbent eluates for CRM-positive plasmas. Figure 2 illustrates a typical study. The factor VIII fragments (before and after thrombin treatment) of plasmas ARC-2 and ARC-3 cannot be distinguished from those of normal plasma. In contrast, the ARC-1 pattern after thrombin treatment is distinctly different, even though the factor VIII protein directly immunoisolated from ARC-1 plasma has apparently normal fragments (lane 3, Fig. 2). The 92-kDa band is more prominent than that for normal factor VIII after incubation with thrombin, and the 44-kDa ~4C. A. Fuleher, S. D. G. Mahoney, J. R. Roberts, C. K. Kasper, and T. S. Zimmerman, Proc. Natl. Acad. Sci. U.S.A. 82, 7728 (1985). 25 j. p. Allain and D. Frommel, Blood42, 437 (1973).
[9]
FACTORVIII CHARACTERIZATION
NPP --
+
ARC2
ARC1 II
II --
+
ARC3 ii
-
+
175
I --
-IF
kDa --200
--92
--80 --72
!
1
/
2
3
4
5
~
6
7
-44
8
FIG. 2. Immunoadsorbed normal (NPP) and CRM-positive factor VIII fragments detected using the three monoclonal antibodies shown in Fig. 1. Thrombin-treated immunoisolated factor VIII is indicated as +; the -- lanes are factor VIII not incubated with thrombin. The patterns for ARC-2 and ARC-3 are the same as those for a normal plasma. The ARC-1 92-kDa factor VIII heavy chain fragment is not cleaved by thrombin at Arg-372 so that the 44-kDa fragment is not generated. From Arai et al. 7 w i t h permission.
fragment band is missing (lane 4, Fig. 2). ARC-1 light chain cleavage is normal, with loss of the 80-kDa band and the appearance of the 72-kDa fragment. This pattern identifies the failure of thrombin to cleave the ARC-1 factor VIII heavy chain, and it suggests that the absence of procoagulant activity is a result of this defect. PCR amplification of exon 8 of the patient's factor VIII gene identified a missense mutation that causes a substitution of histidine for Arg-372, abolishing thrombin cleavage at this site. ~ We have also identified two patients with a factor VIII light chain thrombin cleavage site mutation. In these instances, immunopurified fac-
176
MAMMALIAN COAOULATION FACTORS AND INHIBITORS
[9]
tor VIII has normal heavy chain cleavage, but no conversion of the 80-kDa light chain to its 72-kDa fragment. In both patients, PCR amplification of exon 14 of the factor VIII gene detected a missense mutation that causes a cysteine substitution for Arg-1689. This mutation prevents factor VIII light chain thrombin cleavage. 12 Mutations that change the factor VIII chain mass are also susceptible to characterization by this technique. To date, we have identified two nonfunctional factor VIII-like proteins with abnormal, slower moving heavy or light chains on SDS-PAGE. 16 Subsequently, the molecular defects have been identified by denaturing gradient gel electrophoresis screening of PCR-amplified factor VIII eDNA and sequencing the abnormal PCR products. In both cases, new N-glycosylation sites were identified. In one, substitution ofthreonine for Met-1772 in the factor VIII light chain creates a potential new N-glycosylation site at Asp-1770. A threonine substitution for Ile-566 generates a potential new N-glycosylation site in the second patient's factor VIII. In this case there is carbohydrate addition at Asp-564 in the A2 domain of the factor VIII heavy chain. The abnormal glycosylation was shown to be responsible for the reduced mobilities of the factor VIII chains on SDS-PAGE, as well as the loss of procoagulant activity. ~6 Summary Immunopurification and characterization of dysfunctional factor VIIIlike molecules in CRM-positive and CRM-reduced hemophilia A permit correlation of structural changes with molecular defects. The technique described here is sufficiently sensitive to characterize the molecular mass and enzymatic fragments of the factor VIII chains in patients with as little VIII: Ag as 0.05 units/ml. Specific abnormalities have been identified in 5 of the first 24 samples tested. In each case, the mutation responsible for factor VIII dysfunction has been determined by sequencing a part of the abnormal gene. Mutations have been identified that abolish critical thrombin cleavage sites or which generate new N-glycosylation sites. The technique provides a useful approach to the study of factor VIII structurefunction relationships, and it has the potential to clarify further the molecular basis of factor VIII procoagulant activity. Acknowledgments This work was supported in part by U.S. Public Health Service Grants HL36099 and HL 44336. 16A. M. Aly, M. I-liguehi, C. K. Kasper, H. H. l(aTZZi~an, Jr., S. E. Antonarakis, and L. W. Hoyer, Proc. Natl. Acad. Sci. U.S.A. 89, 4933 (1992).
FACTORVIII CHARACTERIZATION
169
increases the sensitivity of this method. 6s This approach has distinguished a polymorphism in the Asian population in the first intron at base 192. 69 Acknowledgments Supported in part by the National Institutes of Health (HL-31193) and the March of Dimes Birth DefectsFoundation (6-463). ~sL. S. Lerman and K. Silverstein,this series, Vol. 155, p. 482. 69M. Tanimoto, M. Hamaguchi, T. Matsushita, M. Hamaguchi, T. Matsushita, K. Yamamoto, I. Sugiura,J. Takamatsu, and H. Saito, Blood74, 251a (1989).
[9] C h a r a c t e r i z a t i o n
of Dysfunctional Molecules
Factor
VIII
By LEoNW. HOYER Introduction Factor VIII (antihemophilic factor) participates in blood coagulation as an essential cofactor in the activation of factor X by factor IXa in the presence of phospholipid and calcium. It circulates in plasma as a large glycoprotein that is associated with von WiUebrand factor in a noncovalent complex. The mature single-chain protein of 2332 amino acids is proteolytically processed before entering the circulation as a heterodimer consisting of amino-terminal heavy chain polypeptides of 92 to 200 kDa and an 80-kDa carboxyl-terminal light chain. Factor VIII procoagulant activity is generated by thrombin cleavage of the heavy chain at Arg-372 and Arg-740 to yield 54- and 44-kDa fragments, and thrombin cleavage of the 80-kDa chain at Arg-1689 to yield a 72-kDa fragment (Fig. 1). 1 Hemophilia A (classic hemophilia) is an X chromosome-linked disorder of blood coagulation caused by deficient factor VIII activity. Although factor VIII procoagulant activity is consistently reduced or absent, hemophilia A is heterogeneous when characterized dinicaUy or immunologically,2 as well as when the specific molecular defects are identified) Dysfunctional, immunoreactive factor VIII-like protein can be detected in 1W. H. Kane and E. W. Davie, Blood71, 539 (1988). 2 L. W. Hoyer and R. T. Breckenridge,Blood32, 962 (1968). 3 E. G. D. Tuddenham, D. N. Cooper, J. Gitschier,M. I-Iiguchi,L. W. Hoyer, A. Yoshioka, I. R. Peake, R. Schwaab, K. Olek, H. H. Kazazian,Jr., J.-M. Lavergne,F. GianneUi,and S. E. Antonarakis,NucleicAcids Res. 19, 4821 (1991). METHODS IN ENZYMOLOGY, VOL. 222
Copyright © 1993 by Academic Press, Inc. All fights of reproduction in any form reserved.
170
MAMMALIAN COAGULATION FACTORS AND INHIBITORS 372
740
[9]
1669
8O
/
• C8
MAB038
! J 16D-9
FIG. 1. Schematic representation of human factor VIII showing its domain structure (A 1- A3, B, C 1, and C2) and three thrombin cleavagesites (arrows). Below this arc shown the thrombin cleavage fragments. The fragment mass values in kilodaltons correspond to those published by C. A. Fulchcr, J. R. Roberts, and T. S. Zimmcrrnan [Blood61, 807 (1983)]. The solid boxes represent the regions containing epitopes for the monoclonal antibodies used to identify factor VIII fragments. From Arai et al. ~ with permission.
10% of plasmas from patients with mild or moderate hemophilia A.4 These plasmas are termed cross-reacting material (CRM) positive. The remaining hemophilic plasmas are either CRM reduced, in which case the reduction in immunologicaUy detectable protein is comparable to reduction in functional activity, or CRM negative, i.e., there is no detectable factor VIII by sensitive immunoassays. Although the factor VIII gene is large (186 kb), and unrelated individuals have different mutations, there has been considerable progress in identifying the causative point mutations or deletions) Until recently, the specific point mutation or deletion could only be identified in half of the patients with severe hemophilia A using PCR amplification of genomic DNA followed by denaturing gradient gel electrophoresis.5 Rapid screening of all essential regions of the factor VIII gene using an mRNA-based method has been more successful, however, as it also identified the presence of frequent intron 22 mutations that cause defective joining of exons 22 and 23 in the mRNA. 6 Because these studies do not necessarily establish the basis for reduced procoagulant activity, we have developed a simple and sensitive method for the immunoisolation of factor VIII from plasma. ~ This method has been used to identify the functional defect in the factor VIii-like protein recovered from plasmas of patients with CRM-positive hemophilia A. 4 j. Lazarchick and L. W. Hoyer, J. Clin. Invest. 62, 1048 (1978). M. Higuchi, H. H. I(aTa7ian, Jr., L. Kasch, T. C. Warren, M. J. McGinniss, J. A. Phillips, III, C. K. Kasper, R. Janco, and S. E. Antonarakis, Proc. Natl..4cad. Sci. U.S.A. 88, 7405 (1991). 6 j. A. Naylor, P. M. Green, C. R. 1~i77a~ and F. Giannelli, Lancet 340, 1066 (1992). 7 M. Arai, H. Inaba, M. Higuchi, S. E. Antonarakis, H. H. I~aTaTian, Jr., M. Fujimaki, and L. W. Hoyer, Proc. Natl. Acad. Sci. U.S.A. 86, 4277 (1989).
[9]
FACTORVIII CHARACTERIZATION
171
Procedure
Preparation of lmmunoadsorbent Most studies have been carried out using IgG prepared by the caprylic acid methods from the plasma of a patient with a high-titer (3600 Bethesda units) factor VIII autoantibody. The IgG is coupled to cyanogen bromideactivated Sepharose CL-2B at a concentration of 4 mg/ml of settled gel volume.9 Three other high-titer (570-3000 Bethesda units) human antifactor VIII antibodies (as well as autoantibodies) and a murine monoclonal antifactor VIII heavy chain antibody have also been successfully used as the immunoadsorbent when coupled to agarose beads.
Immunopurification of Factor VIII After washing, the immunoadsorbent beads are suspended in Tris-buffered saline (TBS): 0.15 M NaC1, 4 m M CaCI2, 20 m M Tris-HC1, pH 7.4. Prior to characterization, normal or hemophilia A patient plasmas are brought to 0.8 M NaC1 and 0.5% Tween 80, and 0.6-2.5 ml is incubated with 20 #1 of immunobeads overnight at room temperature with gentle aggitation. The amount of factor VIII protein immunoabsorbed on the beads can be estimated by measuring the difference between the factor VIII antigen (VIII: Ag) in the plasma and in the Oostadsorption supematant fluid? The beads are then washed into a 0.8 X 4-era polypropylene column (Bio-Rad, Roekville Center, NY) using 20 ml of 50 m M imidazole, 40 m M CaC12, 5% (v/v) ethylene glycol, 0.5% (v/v) Tween 80, pH 6.4, and further washed with 5 ml TBS. The beads are then transferred into a 1.0-ml conical polystyrene tube, the wash buffer removed by centrifugation, and factor VIII eluted by adding 40/tl of 0.125 M Tris-HC1, pH 6.8, containing 20 mg/ml sodium dodecyl sulfate (SDS), 10% (v/v) glycerol, and 0.05 mg/ml bromphenol blue. After 1 hr incubation at 37 ° with gentle agitation, the tube is centrifuged and the eluted factor VIII is recovered in the supernatant.
Thrombin Treatment of Factor VIII Some aliquots of immunopurified factor VIII were treated with thrombin before the SDS elution step. Thrombin, diluted to 10 units/ml with TBS, is added to these aliquots of washed factor VIII immunoadsorbent beads in sufficient quantity to achieve the desired concentration (0.2-2 s M. Steinbuch and R. Audran, Arch. Biochem. Biophys. 134, 279 (1969). 9 R. Ax~n, J. Porfith, and S. Ernbick, Nature (London) 214, 1302 (1967).
172
MAMMALIAN COAGULATION FACTORS AND INHIBITORS
[9]
units of thrombin/unit VIII:Ag) and the mixture is incubated at room temperature for 0.5 to 60 min with gentle agitation before adding the SDS elution buffer.
Immunoblotting of Factor VIII Eluted factor VIII (40/A) is analyzed by discontinuous SDS-polyacrylamide gel electrophoresis (PAGE) using 1.5-ram-thick 5 - 12% (w/v) polyacrylamide slab gels7 and is transferred to nitrocellulose sheets (Schleicher & Schuell, Inc. Keene, NH) using a Bio-Rad Tram-Blot cell. After transfer, the nitrocellulose sheet is blocked by incubation at 37 ° for 1 hr with borate-buffered saline (BBS) pH 7.8,4 containing 30 mg/ml bovine serum albumin (BSA) and 0.05% (v/v) Tween 20. The sheet is then incubated with antifactor VIII monoclonal antibodies (1/zg IgG/ml) in BBS containing 10 mg/ml bovine serum albumin and 0.05% (v/v) Tween 20 (dilution buffer). After overnight incubation at room temperature, the sheet is washed three times with BBS containing 0.05% (v/v) Tween 20 (washing buffer) and then incubated with tz~I-labeled affinity-purified sheep antimouse immunoglobulin in dilution buffer (80,000 cpm/ml). After a 4-hr incubation at room temperature, the sheet is again washed three times with washing buffer, dried, and visualized by autoradiography at - 7 0 ° for 48 to 72 hr using Kodak XK-I film (Eastman Kodak Co., Rochester, NY). Comparable results have been obtained using chemiluminescence detection. ~° In this case, alkaline phosphatase-streptavidin and biotin-labeled goat antibody to mouse IgG are used to detect the monoclonal antibodies, and the enzyme complex is identified with Lumi-Phos 530 (Lumigen, Inc., Detroit, MI).
Amplification of Genomic DNA and Sequencing Sequence analysis of apparent thrombin cleavage site mutations is carried out with leukocyte DNA amplified using Taq DNA polymerase (Perkin-Elmer Cetus, Norwalk, CT). H The primers are chosen so that the specific factor VIII cleavage site region will be amplified.7,~2 Polymerase chain reaction (PCR) is performed in a 100-/zl volume containing 200-400 ng of genomic DNA, 400 nM of each PCR primer, 200 # M of each dNTP, and 2 units of Taq DNA polymerase in a buffer to A. P. Schaap, H. Akhavan, and L. J. Romano, Clin. Chem. (Winston-Salem, N.C.) 35, 1863 (1989). tt R. K. Saiki, D. H. G¢lfand, S. Stoffd, et al., Science 239, 487 (1988). t2 M. Arai, M. Higuchi, S. E. Antonarakis, et al., Blood75, 384 (1990).
[9]
FACTORVIII CHARACTERIZATION
173
composed of I0 m M Tris-HCl (pH 8.3), 50 mMKC1, 1.5 mMMgCI2, and 0.02% gelatin. Each of the 35 cycles consists of a 30-see denaturation at 94 o, thermal transition from 94 ° to the annealing temperature over 2 min, 45 sec of reannealing at 52-55 °, and a 90-see extension at 72 ° using a DNA Thermal Cycler (Perkin-Elmer Cetus, Norwalk, CT). Amplified DNA is desalted and excess dNTPs are removed by spin dialysis on a Centricon 30 (Amicon, Danvers, MA). The purified DNA is sequenced directly: 10 ng of SP7 sequencing primer (5'-GAT TTT GAC ATT TAT GAT-Y), end labeled with [?-32p]ATP using T4 polynucleotide kinase, is annealed with 80 ng of PCR product on ice after heat denaturation at 95 ° for 5 min. The reaction mixture is divided into four tubes containing 62/~M unlabeled dNTPs, 6.2/tM dideoxy (dd)NTPs, and 2 units of T7 DNA polymerase (Sequenase USB, Cleveland, OH) in the sequencing reaction buffer [25 m M Tris-HC1, pH 7.5, 10 m M MgCI2, 70 m M NaC1, 7 m M dithiothreitol (DTT)]. After incubation at 370C for 15 min, the reaction is stopped with 3/tl of a solution containing 95% formamide, 20 m M EDTA, 0.5 mg/ml bromphenol blue, 0.05% xylene cyanol FF. Samples are boiled for 3 min and electrophoresed in a 6% polyacrylamide/ 8 M urea gel at 58 W for 2 hr. Gels are then dried and exposed to Kodak XAR-5 film (Eastman Kodak Co., Rochester, NY) for 16 hr.
Denaturing Gradient Gel ElectrophoresisAnalysis of Amplified Genomie DNA For evaluation of factor VIII mutations that do not prevent thrombin cleavage, the initial evaluation is carried out using denaturing gradient gel electrophoresis (DGGE) of DNA fragments to which are attached a "GCclamp". 5 PCR products ( - 80 ng of each) from two patients are combined to form heteroduplexes. After heat denaturation at 95 ° for 5 rain, the DNA solution is slowly cooled to room temperature (> 30 rain) and subjected to DGGE under conditions determined empirically for each PCR product. 5 DNA is loaded onto a 6.5% polyacrylamide gel (14 × 19 cm, 0.75-ram thick) containing a linear gradient of denaturants and electrophoresed at 2 - 4 V/cm for 16-23 hr. The gradient difference in denaturants is 20% [100% denaturants = 7 M urea/40% (v/v) formamide]. Gels are then stained in ethidium bromide and photographed with a UV transilluminator. Purified GC-clamped PCR products that show abnormal migrating patterns on DGGE are then directly sequenced. 6,11,13
~3C. Wong, C. D. Dowling, R. K. Saiki, R. G. Higuchi, H. A. Erlich, and H. H. Kazazian, Jr., Nature (London) 330, 384 (1987).
174
MAMMALIAN COAGULATION FACTORS AND INHIBITORS
[9]
Results
Immuno&olation of Factor VIII Protein Factor VIII is immunoadsorbed from plasma using a human antibody to factor VIII that reacts with both heavy and light chain determinants.14 The adsorbent removes 65-95% of VIII:Ag in normal or CRM-positive plasmas. Because the interaction of human factor VIII with these antibodies is very avid, usual elution techniques such as low pH and chaotropic agents are rarely successful in separating the antifactor VIII from an immunoadsorbent. ~s For this reason, and because pH extremes destroy factor VIII immunogenicity, SDS is used to separate the bound factor VIII from the immunoadsorbent. Protein characterization is then accomplished by SDS-PAGE, followed by immunoblotting using monoelonal antifactor VIII. In general, human antifactor VIII antibodies are not satisfactory for this purpose because these plasmas have additional immunologic reactivities not related to factor VIII. Three monoclonal antibodies have been used in most studies: MAB038 (Chemicon, E1 Segundo, CA) binds to the 80kDa light chain fragment but does not react with the thrombin-cleaved 72-kDa light chain fragment; J16D-9 [an antibody prepared by Dr. C. Fulcher (Scripps Research Institute, La Jolla, CA) by immunization with a synthetic peptide containing factor VIII residues 2318-2332] detects both 80- and 72-kDa light chain fragments; and C8 (JR Scientific, Woodland, CA) detects factor VIII heavy chain fragments as large as 200 kDa and reacts strongly with the 92- and 44-kDa thrombin-cleaved heavy chain fragments (Fig. 1).7 The sensitivity of the method is such that both heavy and light chain factor VIII determinants can be detected when at least 0.5 units of VIII: Ag are present in a 20-ml plasma sample.
Characterization of CRM-Positive Hemophilic Plasmas In most cases, the monoclonal antibodies are used together to characterize the immunoadsorbent eluates for CRM-positive plasmas. Figure 2 illustrates a typical study. The factor VIII fragments (before and after thrombin treatment) of plasmas ARC-2 and ARC-3 cannot be distinguished from those of normal plasma. In contrast, the ARC-1 pattern after thrombin treatment is distinctly different, even though the factor VIII protein directly immunoisolated from ARC-1 plasma has apparently normal fragments (lane 3, Fig. 2). The 92-kDa band is more prominent than that for normal factor VIII after incubation with thrombin, and the 44-kDa ~4C. A. Fuleher, S. D. G. Mahoney, J. R. Roberts, C. K. Kasper, and T. S. Zimmerman, Proc. Natl. Acad. Sci. U.S.A. 82, 7728 (1985). 25 j. p. Allain and D. Frommel, Blood42, 437 (1973).
[9]
FACTORVIII CHARACTERIZATION
NPP --
+
ARC2
ARC1 II
II --
+
ARC3 ii
-
+
175
I --
-IF
kDa --200
--92
--80 --72
!
1
/
2
3
4
5
~
6
7
-44
8
FIG. 2. Immunoadsorbed normal (NPP) and CRM-positive factor VIII fragments detected using the three monoclonal antibodies shown in Fig. 1. Thrombin-treated immunoisolated factor VIII is indicated as +; the -- lanes are factor VIII not incubated with thrombin. The patterns for ARC-2 and ARC-3 are the same as those for a normal plasma. The ARC-1 92-kDa factor VIII heavy chain fragment is not cleaved by thrombin at Arg-372 so that the 44-kDa fragment is not generated. From Arai et al. 7 w i t h permission.
fragment band is missing (lane 4, Fig. 2). ARC-1 light chain cleavage is normal, with loss of the 80-kDa band and the appearance of the 72-kDa fragment. This pattern identifies the failure of thrombin to cleave the ARC-1 factor VIII heavy chain, and it suggests that the absence of procoagulant activity is a result of this defect. PCR amplification of exon 8 of the patient's factor VIII gene identified a missense mutation that causes a substitution of histidine for Arg-372, abolishing thrombin cleavage at this site. ~ We have also identified two patients with a factor VIII light chain thrombin cleavage site mutation. In these instances, immunopurified fac-
176
MAMMALIAN COAOULATION FACTORS AND INHIBITORS
[9]
tor VIII has normal heavy chain cleavage, but no conversion of the 80-kDa light chain to its 72-kDa fragment. In both patients, PCR amplification of exon 14 of the factor VIII gene detected a missense mutation that causes a cysteine substitution for Arg-1689. This mutation prevents factor VIII light chain thrombin cleavage. 12 Mutations that change the factor VIII chain mass are also susceptible to characterization by this technique. To date, we have identified two nonfunctional factor VIII-like proteins with abnormal, slower moving heavy or light chains on SDS-PAGE. 16 Subsequently, the molecular defects have been identified by denaturing gradient gel electrophoresis screening of PCR-amplified factor VIII eDNA and sequencing the abnormal PCR products. In both cases, new N-glycosylation sites were identified. In one, substitution ofthreonine for Met-1772 in the factor VIII light chain creates a potential new N-glycosylation site at Asp-1770. A threonine substitution for Ile-566 generates a potential new N-glycosylation site in the second patient's factor VIII. In this case there is carbohydrate addition at Asp-564 in the A2 domain of the factor VIII heavy chain. The abnormal glycosylation was shown to be responsible for the reduced mobilities of the factor VIII chains on SDS-PAGE, as well as the loss of procoagulant activity. ~6 Summary Immunopurification and characterization of dysfunctional factor VIIIlike molecules in CRM-positive and CRM-reduced hemophilia A permit correlation of structural changes with molecular defects. The technique described here is sufficiently sensitive to characterize the molecular mass and enzymatic fragments of the factor VIII chains in patients with as little VIII: Ag as 0.05 units/ml. Specific abnormalities have been identified in 5 of the first 24 samples tested. In each case, the mutation responsible for factor VIII dysfunction has been determined by sequencing a part of the abnormal gene. Mutations have been identified that abolish critical thrombin cleavage sites or which generate new N-glycosylation sites. The technique provides a useful approach to the study of factor VIII structurefunction relationships, and it has the potential to clarify further the molecular basis of factor VIII procoagulant activity. Acknowledgments This work was supported in part by U.S. Public Health Service Grants HL36099 and HL 44336. 16A. M. Aly, M. I-liguehi, C. K. Kasper, H. H. l(aTZZi~an, Jr., S. E. Antonarakis, and L. W. Hoyer, Proc. Natl. Acad. Sci. U.S.A. 89, 4933 (1992).