THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1985 by The American Society of Biological Chemists, Inc

Vol. 260, No. 19, Issue of September 5, pp. 10714-10719,1985 Printed in U.S.A.

Functional Characterization of Human Blood Coagulation Factor XIa Using Hybridoma Antibodies*

(Received for publication, December 3, 1984)

Dipali Sinha$, Alice Koshy, Frances S. Seaman, and Peter N. Walsh From the Thrombosis Research Center, Department of Medicine, Temple University School of Medicine, Philadelphia, Pennsylvania 19140

During the initiation of intrinsic coagulation factors XI and XIa interact intimately with several other co- agulation proteins (factor XIIa, high M, kininogen, and factor IX) as well as with the platelet surface. To help elucidate these complex intramolecular interactions, we have prepared a collection of monoclonal antibodies directed against various epitopes in factor XI. We have utilized these reagents to isolate factor XI and the light chain of factor XIa on affinity columns, and to probe structure-function relationships involved in the inter- actions of factor XIa with factor IX. The isolated light chain of factor XIa retained >90% of its amidolytic activity against the oligopeptide substrate pyro-Glu- Pro-Arg-pNA (S-2366), but only 3.8% of its clotting activity in a factor XIa assay and 1% of its factor IX activating activity in an activation peptide release as- say. This suggests that regions of the heavy chain are required for development of coagulant activity and specifically for the interaction of factor XIa with factor IX. To test this hypothesis, the effects of three of the monoclonal antibodies (5F4, 1F1, and 3C1) on the function of factor XIa were examined. The results show that in a clotting assay the light chain-specific antibody (5F4) inhibits 100% of the factor XIa activity, whereas of the heavy chain-specific antibodies, one (3C1) inhibits 75% and another (1F1) only 17%. Sim- ilarly in the factor IX activation peptide release assay, antibody 5F4 inhibits 100% of the factor XIa activity, whereas 3 C 1 inhibits 75% and 1F1 inhibits 33%. We conclude that regions located in the heavy chain, in addition to those in the light chain, are involved in the interaction of factor XIa with factor IX and in the expression of the coagulant activity of factor XI.

Factor XI is a coagulation protein present in plasma in a precursor form. It participates in the early or contact phase of blood coagulation (1). Since a hemorrhagic state can result from its deficiency, it undoubtedly has a key role in the regulation of blood coagulation (2). Our knowledge concerning the functions and molecular properties of factor XI expanded

*This research was supported by Health and Human Services Grants HL 14217 and HL 25661, Grant CTR 1389 from the Council for Tobacco Research Inc., a grant from the W. W. Smith Charitable Trust, a grant from the American Heart Association, Pennsylvania Affiliate, and by Biomedical Research Grant SO7 R05417 from the Division of Research Resources, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ T o whom reprint requests should be sent at: Thrombosis Re- search Center, Temple University Health Sciences Center, 3400 North Broad St., Philadelphia, PA 19140.

after its isolation from plasma to homogeneity (3-5). Factor XI, present in human plasma at a concentration of 4-6 pg/ ml, migrates with an apparent molecular weight of 160,000 on SDS’ gels and consists of two identical disulfide-linked poly- peptide chains. During the activation of factor XI by factor XIIa an internal peptide bond in each of the two chains is cleaved giving rise to a pair of disulfide-linked heavy and light chains with molecular weights of 50,000 and 30,000, respec- tively (4-6). Factor XI circulates in plasma in a noncovalent complex with high M , kininogen, which is required for the binding of factor XI to negatively charged surfaces and for its proteolytic activation to factor XIa (7, 8). Activated platelets can also promote the proteolytic activation of factor XI (9) and possess high-affinity, specific receptors for factors XI (10) and XIa (11). Factor XIa then remains surface-bound and recognizes factor IX as its normal substrate (12). In this calcium-dependent reaction, factor IX is cleaved at an argi- nine-alanine bond and at an arginine-valine bond to release an M , = 10,000 activation peptide (13, 14). Although it is known that the proteolytic activity of factor XIa is associated with its light chain while the heavy chain is involved in complex formation with high M , kininogen very little is known about the molecular mechanisms involved in the ac- tivation of factor XI by factor XIIa, in the binding of factor XI to high molecular weight kininogen and to activating surfaces, or in the activation of factor IX by factor XIa. To help elucidate these complex intramolecular interactions, we have prepared a collection of monoclonal antibodies directed against various epitopes in factor XI. We have utilized these reagents to isolate factor XI and the light chain of factor XIa, and to probe structure-function relationships involved in the interactions of factor XIa with factor IX.

MATERIALS AND METHODS’

RESULTS

Characterization of Monoclonal Antibodies and Isolation of the Functional Light Chain of Factor Xla-In order to isolate the fully functional light chain of factor XIa, the enzyme was reduced and alkylated under mild conditions. The cleavage products under such conditions contained some unreacted

~

The abbreviations used are: SDS, sodium dodecyl sulfate; PEG, polyethylene glycol.

Portions of this paper (including “Materials and Methods,” Table I, and Fig. 1) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Doc- ument No. 84M-3641, cite the authors, and include a check or money order for $4.40 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.

10714

Hybridoma Antibodies in Characterization of Factor XIa 10715

intact factor XIa and also single-chain factor XIa (due to cleavage of interchain disulfide bond(s)) in addition to heavy and light chains of factor XIa. To isolate the light chain of factor XIa from this reaction mixture we used a heavy chain- specific monoclonal antibody affinity column. Initially there- fore we characterized the three monoclonal antibodies, 5F4, 3C1, and 1F1, with respect to whether they are directed against the light chain or the heavy chain of factor XI. Iz5I- factor XI was activated to '251-factor XIa using bovine factor XIIa as described under "Materials and Methods.'' The en- zyme was then reduced in the presence of 1 mM dithiothreitol and subsequently alkylated using 5 mM iodoacetamide. Under these conditions complete reduction and alkylation occurred producing heavy and light chains as observed on 10% SDS- polyacrylamide gel electrophoresis (Fig. 2, Lane 2). Portions of this reduced and alkylated material were passed through the three different monoclonal antibody affinity columns (approximately 0.5-ml bed volume) equilibrated with phos- phate-buffered saline containing 1 mg/ml bovine serum al- bumin. Materials that did not adhere to the columns as well as those retained by the columns (eluted with 4 M guanidine HC1, pH 4.0) were run on a 10% SDS-polyacrylamide gel (Fig. 2). Lanes 4 and 5 represent the material not retained by 1F1 and 3C1 columns, respectively. Lanes 3 and 6 represent the materials retained by affinity columns 5F4 and 3C1, respec- tively. The light chain preparation, when tested using sub- strate S-2366 (23), appeared to have lost 75% of its activity. However, this experiment enabled us to establish that anti- body 5F4 is directed against the light chain of factor XIa

whereas antibodies 1F1 and 3C1 are directed against the heavy chain.

In order to isolate the fully active light chain, the reduction and alkylation procedure was performed under the following conditions. To 300 pg of factor XIa in 0.5 ml of 40 mM Tris/ succinate buffer, pH 8.3, containing 50 pg/ml Polybrene, 1.0 mM benzamidine, and 1 mM EDTA, 50 pl of 1 mM dithio- threitol was added and the reduction was allowed to proceed for 1 h at room temperature under nitrogen in the dark. The reduced material was alkylated by addition of 50 pl of iodoa- cetamide (5 mM), and the reaction was allowed to proceed for 1 h under the same conditions as for reduction. After dialysis against phosphate-buffered saline, the reduced and alkylated material was run on a 10% SDS-polyacrylamide gel (Fig. 3, Lane 2). This preparation was then passed through the heavy chain-specific 3C1 affinity column. The flow-through fraction was analyzed by SDS-gel electrophoresis (Fig. 3, Lane 3). It is apparent that the reduced and alkylated material contained, in addition to heavy and light chains, single-chain (Mr = 80,000) factor XIa, which resulted from cleavage of only the interchain disulfide bond(s). Since antibody 3C1 recognizes the heavy chain, only the light chain passed through the 3C1 column as a pure component. The purified light chain was analyzed for protein, amidolytic activity, procoagulant activ- ity and for its capacity to activate factor IX in comparison with factor XIa (Table 11). It is evident that, compared with intact factor XIa, the light chain retained >90% of its enzy- matic activity but only 4% of its clotting activity and 1% of its capacity to activate factor IX.

Effects of Monoclonal Antibodies on the Procoagulant Actiu- ity of Factor XZa-Since the biological activity of the light chain of factor XIa is much less than the intact molecule, one

' --. I

1 2 3 4 5 6 FIG. 2. SDS-polyacrylamide gel electrophoresis of reduced

and alkylated '251-factor XIa before and after passage through monoclonal antibody columns. Lane I , factor XIa; Lane 2, factor Xla after reduction and alkylation; Lune 3, material that came through 3C1 affinity column without sticking; Lane 4, material unre- tained by 1F1 affinity column; Lane 5, material eluted from 5F4 affinity column; Lane 6, material eluted from 3C1 affinity column.

-

I 2 3 FIG. 3. SDS-polyacrylamide gel electrophoresis of isolated

light chain of factor XIa. Lane I, factor XIa; Lane 2, factor XIa after mild reduction and alkylation; Lane 3, affinity purified light chain of factor XIa.

10716 Hybridoma Antibodies in Characterization of F a c t o r XIa

TABLE I1 Functional comparison of factor XIa and its light chain

Details of each measurement are described under “Materials and Methods.” Amidolytic activities were measured using the oligopeptide substrate pyro-Glu-Pro-Arg-pNA (S-2366). Three different concen- trations of factor XIa (70 nM, 35 nM, 17.5 nM) and of the light chain (37 nM, 18.5 nM, 8.5 nM) were used to measure the initial rate of release of pNA and the data in column 1 represent the average of these three values. One clotting unit is defined as the factor XI clotting activity contained in 1 ml of normal pooled plasma. In the %factor IX activation peptide release assays, the concentration of 3H-factor IX in the assay mixture was 2 pg/ml or 35 nM (specific radioactivity 0.42 X lo6 cpm/pg), that of factor XIa was 0.04 pg/ml or 0.5 nM and that of the light chain 0.4 pg/ml or 13.3 nM. Since factor XIa contains two light chains and two heavy chains comprising a total molecular weight of 160,000, the molar concentration of factor XIa was calculated assuming a molecular weight of 80,000. Molar concentration of the light chain was calculated assuming a molecular weight of 30,000.

Enzyme Amidolytic Clotting 3H-factor IX activity activity activation

pmol p N A releasedl mol radiopeptide/ pmol enzymels mol enzymelmin

Factor XIa 112 21.4 4.8 Light chain 101 0.81 0.047

1Fl

3C1

I 1 I n I 5F4 I

IO 20 30 40 50 60

CONCENTRATION OF ANTIBODY (POIml) FIG. 4. Neutralization of factor XIa in the clotting assay by

the three monoclonal antibodies, 5F4, 3C1, and 1F1. Factor XIa (0.5 pg/ml) was incubated with either buffer or different concen- trations (2-40 pg/ml) of each antibody for 20 min before measuring the clotting activity.

obvious question is whether the conformation around the active site of the light chain is vital for full expression of its biological activity and whether this conformation is lost dur- ing reduction and alkylation procedures. Alternatively, the results presented in Table I1 suggest the possibility that the heavy chain of factor XIa makes a major contribution to the enzymatic activity of factor XIa as a factor IX activator. A complementary approach to the examination of this question involves the use of heavy chain-specific monoclonal antibod- ies to block putative functional sites essential for the inter- action of factor XIa with its macromolecular substrate, factor IX. Therefore, the procoagulant activity of factor XIa, deter- mined both in a factor XIa clotting assay and by its capacity to activate factor IX, was examined in the presence of various concentrations of heavy chain-specific and light chain-specific antibodies. As shown in Fig. 4 light chain-specific 5F4 inhib- ited 100% of the factor XIa procoagulant activity in a coagu-

lation assay whereas heavy chain-specific 3C1 caused 75% inhibition and 1F1 only 17%. The effects of the antibodies on the rate of activation of 3H-factor IX by factor XIa are shown i n Fig. 5. When the per cent of radiopeptide release at 10 min in the absence and presence of different concentrations of the antibodies was examined (Fig. 6), the three antibodies showed similar inhibition patterns to those in the coagulation assay (Fig. 4). Normal mouse IgG did not inhibit factor XIa either in the clotting assay or in the 3H-factor IX activation peptide release assay. It should be emphasized that the three antibod- ies used in these experiments showed equivalent binding to factor XI when tested in the solid-phase radioimmunoassay.

DISCUSSION

Factor XI is involved in complex interactions with at least three coagulation proteins in carrying out its functions in

30

20

I O 0 w I

WJ

W 0-

”“1 /- ” I

40-

30-

20-

10-

0 IO 20 30 40 SO 60

INCUBATION TIME (min) FIG. 5. Effects on the rate of activation :f ’H-factor IX by

factor XIa in the presence of various concentrations of light chain-specific antibody 5F4 (Panel A ) , heavy chain-specific antibody 3C1 (Panel B) , and heavy chain-specific antibody 1F1 (Panel 0. Factor XIa (0.4 pg/ml) was incubated with either buffer or various concentrations of the antibody solution prior to use in the activation of 3H-factor IX. The concentrations of the antibody utilized were: none (0); 8 pg/ml(O); 20 pg/ml (A); 40 pg/ml (A).

Hybridoma Antibodies in Characterization of Factor XIa 10717

W n

0 n 5F4 1 1 I I

10 20 30 40

ANTIBODY CONCENTRATION (pg/ml)

FIG. 6. Comparison of the three monoclonal antibodies in their abilities to neutralize the enzymatic activity of factor XIa. Factor XIa (0.4 pg/ml) was incubated with buffer or various concentrations of the antibody solutions (4-40 pg/ml) for 20 min at room temperature prior to its use in the 3H-factor IX activation assay. The final concentrations of factor XIa and 3H-factor IX in the 3H peptide release assay were 0.04 pg/ml and 0.2 Fg/ml, respectively.

hemostasis. Factor XIIa is the enzyme that activates factor XI (4-6). It circulates in plasma as a noncovalent complex with high M , kininogen which is a nonenzymatic cofactor in the surface-dependent activation of factor XI by factor XIIa (7, 8). Finally, factor XIa recognizes factor IX as its normal substrate and cleaves it at two internal peptide bonds with release of an activation peptide (13,14). Few details are known regarding intramolecular recognition sites of factor XI for these proteins. In the present study, we have focused our attention on the factor XIa-catalyzed activation of factor IX. The application of hybridoma technology has allowed us to generate monoclonal antibodies to different structural do- mains in human factor XI. These antibodies have proven to be extremely useful not only for rapid purification of factor XI, a trace protein in plasma, and in isolation of the light chain of factor XI, but also for the elucidation of some structure-function relationships involving the activation of factor IX by factor XIa.

Of the three antibodies described here one (5F4) is directed against the light chain, whereas the two others (3C1 and 1F1) are directed against the heavy chain. Using an antibody 5F4 affinity column >1,300-fold purification can be achieved in a single step. In a second step, the remaining contaminants are absorbed using a DEAE A-50 column connected in tandem with a protein A column. This rapid two-step procedure yields a homogeneous, fully functional protein with a recovery of 41%. Isolation of the light chain of factor XIa has been facilitated through use of a heavy chain-specific 3C1 affinity column. When reduced and alkylated factor XIa is passed over this column, both the heavy chain and any unreacted material or partially reduced material containing the heavy chain are retained, and as a result the light chain is eluted from the column as a pure component. Although we have not yet attempted to isolate the heavy chain, we predict this can be accomplished by a similar technique. When the reduced and alkylated material is passed over a light chain-specific antibody column, the light chain, zymogen factor XI, and any partially reduced material containing light chains should be retained, whereas the heavy chain should pass through the column as a pure component.

Our results demonstrate that factor XIa, when reduced and alkylated under mild conditions retains >90% of its enzymatic activity as measured in an amidolytic assay. However, when tested in a clotting assay and in the activation peptide release assay for activation of factor IX, the reduced and alkylated material is considerably less active than intact factor XIa. We therefore tested the activity of the isolated XIa light chain for its capacity to hydrolyze the synthetic substrate pyro-Glu- Pro-Arg-pNA, its procoagulant activity, and its ability to activate purified factor IX. As shown in Table I1 when com- pared with the parent factor XIa molecule, the light chain is far less active both in the clotting assay and in the activation peptide release assay, while showing almost full amidolytic activity. The procoagulant activity of the light chain is only 4% of the parent factor XIa molecule on a molar basis, and in the activation peptide release assay for activation of factor IX it is only 1% as active as native factor XIa. Qualitatively similar results have been reported by van der Graaf et al. (28); however, their light chain preparation had only 0.05% of the activity of native factor XIa in the activation peptide release assay and 1% of the clotting activity of native factor XIa. We do not know the reason for this discrepancy although one possibility is that variations in the extent of conformational alteration resulting from reduction and alkylation account for the differences. Nonetheless it is evident that light chain preparations which are fully reactive against pyro-Glu-Pro- Arg-p-nitroanilide are far less active against factor IX. This difference in the activity of factor XIa and its light chain against factor IX is explained in a t least three different ways. One possibility is that the full enzymatic activity of factor XIa against factor IX is expressed only when the light chain and the heavy chain are joined as in the native molecule, i.e. the heavy chain has an important role in the interaction of factor XIa and factor IX. A second possibility is that the native conformation of the light chain which may be vital for its interaction with factor IX is destroyed during reduction and alkylation. A third possibility is that the dimeric nature of factor XIa with its two active sites (4) is important for the efficient cleavage of two peptide bonds in factor IX.

In order to provide data that might support one of these three hypotheses, we examined the effects of our heavy chain- specific antibodies on the enzymatic activities of factor XIa. We reasoned that chemical modification of factor XIa might alter the native conformation of the light chain even though conditions of reduction and alkylation are mild, and that retention of the amidolytic activity of the light chain ( i e . maintenance of an intact active site) does not exclude the possibility that conformational alterations may impair the interaction of the light chain with factor IX. An alternative and complementary way to examine the role of the heavy chain of factor XIa in the activation of factor IX is to test the effects of monoclonal antibodies directed against various ep- itopes in the heavy and light chains on the enzymatic prop- erties of factor XIa. As expected, the light chain-specific antibody (5F4) abolished the catalytic activity of native factor XIa in the coagulation assay in plasma and in the activation peptide release assay using purified factor IX. This result is consistent with the view that the active site is contained in the light chain, as previously sugested by our results (above) and those reported by van der Graaf et al. (28), that the isolated light chain possesses catalytic activity. These obser- vations support the same conclusion drawn from the original demonstration by Kurachi and Davie (4) that both diisopro- pylphosphofluoridate and antithrombin I11 bind to the light chain of factor XIa and the demonstration by Koide et al. (29) that the light chain contains an amino acid sequence

10718 Hybridoma Antibodies in Characterization of Factor X I a

typical of a serine active center. Similar studies with our heavy chain-specific antibodies

(1F1 and 3C1) are most revealing as regards the role of the heavy chain in the enzymatic activity of factor XIa. When these two antibodies were tested for their effects on factor XIa in a clotting assay and in the activation peptide release assay, one of the antibodies (3C1) caused 75% inhibition of procoagulant activity and 75% inhibition of factor IX-cleaving activity, whereas the other (1F1) resulted in only 17% inhi- bition of procoagulant activity and 33% of factor IX-cleaving activity. These results are consistent with the interpretation that (a) specific domain(s) in the heavy chain of factor XIa contribute(s) to the full expression of enzymatic activity. At least one heavy chain domain, a t or near the epitope recog- nized by antibody 3C1, is postulated to be involved in the interaction of factor XIa with factor IX. The large inhibition of factor XIa procoagulant or enzymatic activity by 3C1 could not possibly be due solely to steric hindrance since Fab' fragment of the same antibody (data not shown in this paper) showed similar inhibition. Since the heavy chain of factor XI contains binding sites for high M , kininogen, which serves as a nonenzymatic cofactor in the surface-dependent activation of factor XI by factor XIIa (4, 7), we consider it important to determine whether this high M, kininogen binding site on the heavy chain of factor XI also contributes to the enzymatic activity of factor XIa.

Acknowledgments-We are grateful to Dr. Lei Mei Kuo and Dr. Helen Davis (Department of Microbiology, University of Pennsylva- nia) for advice and fruitful discussions in the production of monoclo- nal antibodies, to Dr. Linda C. Knight and Monica Kollman for their assistance in radiolabeling proteins, to Dr. Edward P. Kirby for providing bovine factor XIIa, to Dr. Petter Friberger, AB KABI Peptide Research, Molndal, Sweden, for providing the chromogenic substrate S-2366, and to Patricia Pileggi for typing the manuscript.

REFERENCES

1. Davie, E. W., Fujikawa, K., Kurachi, K., and Kiesel, W. (1979)

2. Rosenthal, R. L., Dreskin, 0. H., and Rosenthal, N. (1953) Proc.

3. Koide, T., Kato, H., and Davie, E. W. (1977) Biochemistry 16,

Adu. Enzmol. Relat. Areas Mol. Biol. 48, 277-318

SOC. Exp. Biol. Med. 82, 171-174

2279-2286

4. Kurachi, K., and Davie, E. W. (1977) Biochemistry 16, 5831-

5. Bouma, B. N., and Griffin, J. H. (1977) J. Biol. Chem. 252,6432-

6. Mannhalter, C., Schiffman, S., and Jocobs, A. (1980) J. Biol.

7. Mandle, R. J., Colman, R. W., and Kaplan, A. P. (1976) Proc.

8. Griffin, J. H., and Cochrane, C. G. (1976) Proc. Natl. Acad. Sci.

9. Walsh, P. N., and Griffin, J. H. (1981) Blood 57, 106-118

5939

6437

Chem. 255,2667-2669

Natl. Acad. Sci. U. S. A. 73, 4179-4183

U. S. A. 73, 2554-2558

10. Greengard, J. S., Walsh, P. N., and Griffin, J . H. (1981) Blood

11. Sinha, D., Seaman, F. S., Koshy, A,, and Walsh, P. N. (1984) J.

12. Wiggins, R. C., Bouma, B. N., Cochrane, C. G., and Griffin, J. H.

13. Fujikawa, K., Legaz, M. E., Kato, H., and Davie, E. W. (1974)

14. Osterud, B., Bouma, B. N., and Griffin, J. H. (1978) J. Biol.

15. DiScipio, R. G., Hermodson, M. A., Yates, S. G., and Davie, E.

16. Miletich, J. P., Broze, G. J., Jr., and Majerus, P. W. (1980) Anal.

17. Walsh, P. N., Bradford, H., Sinha, D., Piperno, J . R., and Tusz-

18. Bolton, A. E., and Hunter, W. M. (1973) Biochem. J. 133, 529-

19. Tuszynski, G. P., Knight, L., Piperno, J . R., and Walsh, P. N.

20. Van Lenten, L., and Ashwell, G. (1971) J. Biol. Chem. 246,1889-

21. Silverberg, S. A., Nemerson, Y., and Zur, M. (1977) J. Biol. Chem.

22. Proctor, R. R., and Rapaport, S. J . (1961) Am. J . Clin. Pathol.

23. Scott, C. F., Sinha, D., Seaman, F. S., Walsh, P. N., and Colman,

24. Laemmli, U. K. (1970) Nature (Lond.) 227, 680-685 25. Kohler, G., and Milstein, C. (1975) Nature (Lond.) 256,495-497 26. Shulman, M., Wilde, C. D., and Kohler, G. (1978) Nature (Lond.)

27. Kennett, R. H., and McKearn, T. J . (1980) in Hybridomas: A New Dimension in Biological Analysis (Kennett, R. H., Mc- Kearn, T. J., and Bechtol, K. B., eds) pp. 365-375, Plenum Press, New York

28. van der Graaf, F., Greengard, J . S., Bouma, B. N., Kerbiriou, D. M., Griffin, J. H. (1983) J. Biol. Chem. 258, 9669-9675

29. Koide, T., Hermodson, M. A., and Davie, E. W. (1977) Nature

58, 688 (abstr.)

Clin. Inuest. 73, 1550-1556

(1977) Proc. Natl. Acad. Sci. U. S. A. 74, 4636-4640

Biochemistry 13,4508-4516

Chem. 253,5946-5951

W. (1977) Biochemistry 16, 698-706

Biochem. 105,304-310

ynski, G. P. (1984) J. Clin. Inuest. 73, 1392-1399

539

(1980) Anal. Biochem. 106, 118-122

1894

252,8481-8488

35, 212-219

R. W. (1984) Blood 63,42-50

276,269-270

(Lond.) 266, 729-730

Supplemenrary ?larerial t o : FUNCTIONAL CHARACTERIZATION OF H W SLOOD COAGULATION FACTOR XIa USING HYBRIDOMA ANTIBODIES by Dipali Sinha, Alice Koshy, Frances S. Seaman and Peter N. Waleh

MATERIALS AND METHODS

and w m a s e d from Sigma Chemical Co.. St. LOULS, MO; Fisher Chemical Materials. All chemicals were the best grade commercially available

Co., Fairlawn, NJ; or J.T. Baker Chemical CO., Philllpsburg, NJ. Plasmas

Overland Park, KA. Betaphase scintillation fluid was obtalned from West Chem deflclent in coagulation factors were purchased from George King Biochemical,

Products, Sa" Diego, CA. Tritiated sodium borohydride (75 Cl/mmoll was purchased ~n crystalline form in sealed ampoules from New England Nuclear ,

acid] was from Calbiochem-Behring Corp. Crystallized bovlne S e r u m albumln Boston, MA. Ulfrol Hepes [4-12-Hydroxyethyli-l-piperazineethanesulfO"~~

(BSAI, benzamidine hydrochloride, soybean trypsin Inhibitor, sodium heparin and cephalin lrabbir5brain extract1 were purchased from sigma Chemlcal Co. (St. Louis. Mol. I-labeled Bolton and Hunter reagent was obtalned from New England Nuclear (Boston, MA). Acrylamide, sodium dodecyl Sulfate (5051, N,N' -methy lene -b iS -acry lamide . N,N,N',N'-tetramethylethylenediamine and cyanogen bromide activated sepharose were purchased from B10-Rad Laboratories (Richmond, CAI. The chromogenic sobstrate pyro-Glu-Pro-A~g-paranitroanilide 2 HCL ( S - 2 3 6 6 1 was a gift from Kabi Peptide Research (Stockholm. Sweden).

Purlficatron of Proteins. Factor XI was purlfled from 2L of human plasma by a m o d i f i G t m of the method of Bouma and Grlffin (51. The purified protein appeared homogenous on polyacrylamide gel electrophoresis in the presence of SDS as prevmusly published (111. and had a specific activity of 270 U/mg protein. It was stored I" 0.2 M sodium acetate, 0.6 M NaC1, pH 5.3 at -lOac. Human factor Ix was purified to a specrflc actlvlty Of 225 U/mg by a modificatzon of methods described by DiSCipiO et a1 ( 1 5 1 and by MlletlCh et a1 1161 and was a single band on SDS gel electrophoresis as previously pubdished (171. It was stored in 0 . 0 1 2 M Tr15, 0.045 M glyclne, pH 8 . 3 at -185 C. Bovine two-chaln factor XlIa w a s kindly provided by D r . E . P. Klrby of the ThrOmbOS1S Research Center, Temple University School of Medlclge, Philadelph~a, PA as a homogenous protein which hydrolyzed

pH 8.0 and 31'C. 3.18 x IO mo of the chromogenic substrate S-2302/mn/pg protein at

was activated by incubation at 37-C w i z bovine factor XIIa as prevlously described (111. TO 100 p l (100 wgl of purified factor XI (rn 0 . 2 M acetate

and this was follgwed by lncnbatlon with 6 p 1 ( 3 p g l of two-chaln bovine buffer, pH 5.31 concentrated Tris base was added to ad3ust the pH to 7.5

tion under reducing condition showed two major bands of Mr 50,000 and factor XIIa at 37 C for 3-4 hr5. SDS-gel electrophoresis Of this prepara-

30,000 respectively.

removed from the labeled protein by passage over a 1 ml G - 2 5 column 1 1 9 1 and the p7ocedure d e s z r a e m t o n and Hunter (18). Most of the free I was

to remove resldual "1. Factor Ix was labeled wlth tritium by a p r e v l o u s l y the proteln was the? ialyred further in the presence of OvalDumln (1 mg/mll

descrrbed modification (171 of the method described by Van Lenten and Ashwell

proteln was flnally purrfied by alkaline ge l electrophoresis, appeared (201 and adapted for bovine factor x by Silverberg et a1 1211. The labeled

homogenous as Judged by SDS g e l electrophoresis and fluorography as pre- vlously reported 1171, and >98% of the radioactivity was precipitable in 5 % trichloroacetic acld. The tritiated factor IX had a Speciflc radioactivity Of 415 CPM/ng and retalned >90% of its Coaqulant activitv compared wlth unlabeled factor Ix.

Preparation and characterlzatzon Of Factor XIa. eurlfied factor XI

Radtolabelln of Proteins. Factor XI was labeled with lZ5I Utilii$ng

modifications of the kaolin-activated partla1 thromboplastin tlme (221 using appropriate congenitally deflclent Substrate plasmas, and reBUltS were quantitated on double logarithmic plots of clotting times versus concen- tratlm of pooled normal plasma. The procoagulant activltres of factor X I a and its llght chaln were determlned uslng factor XI-deficient plasma. One hundred m~crolrters of factor XI-deflclent plasma was incubated w t h 100 p l of kaolin I S mg/ml I" sallnel, 100 p l of 0.2% inosithln in 20 mn Tris/sallne. pH 1.4 and 10 w & of plasma or sample + 90 ill of the above mentioned buffer for S mln at 37 C. Then 100 p 1 of 30 mM CaCl was added to initlate clot formation. The observed clottlng tlme was cdverted to clotting units by comparlson to the clotting activlties of ser~al dllutions of a normal pooled plasma.

Coagulatlon w. Factors IX and XI were assayed utrllzlng minor

Hybridoma Antibodies in Characterization of Factor XIa 10719

assay of factor XIa using the oligopeptide substrate pyro-G1U-Pro-Arg-pNA Amidolytic Factor XIa and its Liqht Chain. The amidolytic Assays of

has been described previously 1 2 3 ) . Ten p 1 of different concentrations of factor XIa 10.6 - 6.0 pg/ml) or its light chain 10.2 - 2.0 pg/ml) were added to 250 p l of 0.1 M Na phosphate, pH 7.6 containing 0.15 M NaCl and 1 mM EDTA and 15 p l of the substrate S-2366 (11.5 mM) in a 1 cm cuvette. The rate of hydrolysis was measured using a Gilford System 2600 spectrophotometer (Gilford, Oberlin, OH).

of ['-X was follovedy measurementof the trichloroacetic acid soluble actlvatlon peptkde released during activation by factor XIa as described previously (17). For each time point the assay mixture consisted of 64 p l of TriS (50 mM1 NaCl (100 M I , pH 7 . 5 (TBS) Containing 1 m g / m l bovine 9 rum albumin, 8 ;1 of 'H-factor IX (20 vg/ml, specific radioactivity 0.4 X 10% cpm/pgl and 8 "1 of factor XIa (0.2 - 0.4 pg/ml) or llqht chain 12 - 4 pg/ml). The reaction was stopped by adding 290 ul of an ~ c e cold mixture containing one part TBS and two parts 50 mM EDTA. pH 7.5. To this mixture was added 160 "1 of tce cold 150 trichloroacetic acid. This was kept on ice and vortexed vigorously and repeated for 2 min and then centrifuged at 1 0 , O U O g for 3 min in a bench top Brimman Model 3200 microfuge (Brinkman Instruments, InC., Westbury, NYI. Thereafter, 100 p l aliquot3 of the gupernatants were removed into 10 ml of Scintillation fluid and counted for

solution was replaced by 8 p l TBS. In each experiment assay background was Fullerton, CAI. In order to determine the assay background, 8 p 1 of enzyme

less than 20 of the total number of counts. The initial rate of release of

of the activation peptide had been released. the activation peptide was determined under conditions where less than 200

Radiometric Assay of Factor x Activation. The rate of activation

H In a BeCKman LS8000 scintillation counter (Beckman Instruments, Inc.,

Bio-Radbblnding assay according to inJtrUCtionS provided by the manu- Protein Analyses. Protein concentrations were determined by the

quancltated using an extinction coefflcient 01 1 4 for a 10 solution at facturer (Bio-Rad, Rlchmond, CA). Purifled monoclonal antlbodies were

280 nm. Polyacrylamide slab gel electrophoresis in SDS was done by the procedure of Laemmlt (24).

(RIA) was used to detect antibody aga~nst factor XI. Ninety-six well round Radioimmunoassay of Antifactor XI. A solid phase radioimmunoassay

bortom, flexible, polyvinyl chlorlde microtiter plates (Dynatech, Alexandria, VA I were Washed with ethanol and drled. To each well was added 300-500 ng of factor XI ~n 100 p l of 0.01 M sodium phosphate, 0.15 M NaC1, pH 7.4 (pho hate-buffered saline, PBSI. The plates were incubated for 16-20 hrs at d8;. The wells were washed With 0 . 5 0 BSA in PBS to remove any ""adsorbed antigen. Addrtlonal sites on the plates were coated with BSA by incubating with 200 p l of 0.50 BSA in PBS for 2 hrs a t room temperature. One hundred

at room temperature. The plates were washed with 0.1u albumin in PBS. "1 of each test sample was then added to the wells and incubated for 2 hrs

The bound specific antlbody was detected by the addition of 0.1 m l of 1251- labeled, affigity pucifted goat antI-mouse IgG diluted in 0.10 albumin in PBS to 1 X 10 cpm/ml. Afcer incubation at 37OC for 2 hrs, the plates were washed five times with 0.10 BSA in PBS. The toDs of the wells were cut off and the individual wells were assayed in a serine counter (Nuclear Enterprises, NE 1600, Scotland, U.K.).

Production of Monoclonal Antibodies. Murine nonoclonal antibodies

cellular hybridizatlon (251. BALB/C mice (Skin and Cancer Laboratory, to purified factor XI were prepared by immunlration of mice and subsequent

Philadelphia, PA) were immunized by three intraperitoneal injections of 50-60 pg of purlfied factor XI in complete Freund's adjuvant given one week apart. Seven to ten days after the last immunization the mice were boosted wlth one intravenous injection of the same amount of Antigen in PBS. Three

were removed aseptically and single cell suspensions were prepared in mnimal to five days after the final injection, the mice were sacrificed, the spleens

essential medium (HEM),(GIBCO Grand Island New Yock) containing 200 fetal cajf Serum. Cells (aPpr0xima;ely 10') from'a single spleen were mixed with 10 SP2/O-Ag14 mouse myeloma cells, pelleted and washed in medium containing no fetal calf serum. SP2/O-Ag14 is a hybridoma variant that does not syn-

myeloma line 125) by Shulman et al. 126). Cells were fused by a modification thesize immunoglobulin components and which was derived from the P3-X63-Ag8

of the procedure described by Kennett et a1 (27). The medium was removed the Pellet was loosened gently and 1 m l of 350 (vol/vol) polyethylene qlyl-01 1500 in HEM was added, pelleted and then diluted With 5 m l of medium Without Serum followed by addition of another 5 m l of medium with 200 fetal calf Serum. The tlme from addition of PEG to the dilutlon of the cell pellet was carefully controlled between 7-9 minutes. The cells in diluted PEG were pelleted and resuspended in 30 m l of HY medium IDulbecco's HEM with high glucose ( 4 . 5 g/literI, 100 NCTC 109 (Microbiological ASSOC., Walkersville, MD), 200 fetal calf serum, 0.15 mg/ml oxaloacetate (Sigma), 0.05 mg/ml pyruvate (Sigma), 0.2 U/ml bovine insulin (Sigma)]. In addition to the purine and pyrmidine bases present in the NCTC 109 medium thymidine (16 pH1 and hypoxanthine 10.1 mMI were added to the medium in whici the fused cells were plated. The 30 m l of cells were evenly suspended and distributed into 6 mlcroplates. The next day each well was fed with an equal volume of the same medium with aminopterin ( 0 . 8 pH) to make hypoxanthlneaminopterinthymidine IHAT) selective medium. The cells were fed twice a week with medium wlthout aminopterin. Colony growth was apparent beginning at day 14.

~ ~~

Tissu- supernatants from 20 hybrids were tezed for any anti- Isolation of Hybridoma Clones with Antifactor XI Activity.

Supernatants from ten colonies were found to be strgngly positive. Cells factor XI activity using the solid-phase radioimmunoassay described above.

supernatants positive in the RIA were ala, tested for their capacity to from these wells were frozen in liquid N at 5 X 10 cells per ampule.

neutralize factor XI in the clotting assay. Three colonies were then chosen for further cloning. One of these three 15F4) neutralized factor XI in the clotting assay very effectively, another one OC1I neutralized it partially, Whereas the third one (1F1) was very weak in its capacity to neutralire

was done by a limiting dilution method (27). Cells were grown for freezing factor XI. All three antibodies were equally potent in the RIA. Cloning

and passage into mice as an asciti tunor Pristane primed BALB/c mice were injected intraperitoneally with 10' cloned hybridoma cells. Ten to twelve

days later ascitic fluids were collected. Ten to twelve m l of ascitic fluids were obtained from each mouse. Antibodies were isolated from the ascitic fluids with 4 0 8 ammonium sulfate DreciDitation followed bv ael filtration using BiGel A-1.5 m. In some caies the antibody was further purified

8s > 980 pure by SDS-polyacrylamide gel electrophoresis and quantitated using AffiGel Blue (BioRad). The purified antibodies were characterized

using an extinction coefficient of 14 for a 10 solution at 280 nm. To make antibody affinity columns, approximately 3-4 mg of the purifled antibodies were coupled to 1 g of cyanogen bromide activated Sephsrose (Sigma) according to the instructions provided by the manufacturer.

m. Initially we developed the following procedure for the small scale Isolation of Human Factor XI Usinq a Monoclonal Antibody Affinity

centrlfugation at 10,000 g and applied to an antibody (5F4) affinity column Isolation of human factor XI. Fresh frozen plasma (200 m l ) was clarified by

(3.5 mlI eauilibrated in PBS. NO factor XI clottina activitv was detected in the floi through fraction. The column was then washed with epproximately 10 vol of the Same buffer and then with 0.2 M NaHC03/1 M NaCl pH~8.0 until ~

the optical density at 280 nm of the wash was < 0.1. The column was further washed with 10 "01s of 0.1 M qlvcine/HC1/1 M NaC1. DH 3.0. Factor x1 was eluted from the column with 4H guanidine-HC1 pH 4;0't25 ml) and immediately dialyzed VS. 40 mM Tcis-succinate buffer pH 8.3 containing 50 pg/ml polybrene 1 0 mM benrsmidine 1 mM EDTA and 0 020 NaN . The dialyzed material kas'passed through a'DEAE A-50 column'(5 m l ) donnected in tandem with a protein A column 12 m l ) equilibrated with the same buffer. The column was washed with the same buffer (the volume of the wash being twice the volume of the sample size). The effluent together with the wash was pooled and concentrated. The purification data from a single preparation of factor XI are summarized ~n Table I. Factor XI was purified 15,000 fold with a final yield of 410 by this two step procedure. The entire purification procedure was carried Out in a total of three days.

HUNAN FACTOR XI PURIFICATION

Protein Activity Purification Purification Total Activity Activity

trig) (Units) ( U Yield) I Fold I

Pool from Affinity Column 6.0 136 68 1,586

Pool from DIAE-ASO/ Protein A Column 0.380 82 41 15.105

...................................

Assuming protein concentration in plasm is approximately 70 mg/ml .

The purified protein had a specific activity of 216 units/mg protein and migrated as a slngle band at M r = 160,000 under nonreducing conditions and at M r = 80.000 under reducing conditions on SDS-gel electrophoresis (Figure 1).

MW x IO-'

200 - 116- 96- 68-

4 5 -

31 - 2 0 -

1 2 3 Figure 1: SDS-polyacrylamide gel (108) electrophoresis of affinity purified

factor XI. ~ a n e 1, molecular weight standards; lane 2, unreduced factor x11 lane 3, reduced factor XI.