FACTOR VIIIR (VON ILLEBRAND FACTOR)

This invention relates generally to the field of structural gene cloning and the use of such genes in the

5 recombinant DNA-directed synthesis of desired protein products. More particularly, it relates to blood clotting

Factor VIIIR, its recombinant DNA-directed synthesis, and its use in the treatment of coagulation disorders, such as von illebrands's disease.

10 In general, recombinant DNA techniques have now become well known. See: Methods In Enzymology, (Academic Press) Volumes 65 and 68 (1979); 100 and 101 (1983) and the references cited therein, all of which are incorporated herein by reference. An extensive technical discussion

-\ r- embodying most commonly used recombinant DNA methodologies can be found in Maniatis, et al. , Molecular Cloning, Cold Spring Harbor Laboratory (1982) . Genes coding for various polypeptides may be clone ^" d by incorporating a DNA fragment coding for the polypeptide in a recombinant DNA vehicle,

2o e.g., bacterial or viral vectors, and transforming a suitable host. This host is typically an-Escherichia coli (,E. coli) cell line; however, depending upon the desired product, eukaryotic hosts may be utilized. Clones incorporating the recombinant vectors are isolated and may be grown and used to

25 produce the desired polypeptide on a large scale.

Several groups of workers have isolated mixtures of mRNA from eukaryotic cells and employed a series of three enzymatic reactions to synthesize double-stranded DNA copies of entire genes which are complementary to this mRNA mixture. In the first reaction, mRNA is transcribed to form a

5

single-stranded complementary DNA (cDNA) by an RNA-directed DNA polymerase, also called reverse transcriptase. Reverse transcriptase synthesizes DNA in the 5'-3' direction, utilizes deoxyribonucleoside 5'-triphosphates as precursors, and requires both a template and a primer strand, the latter of which must have a free 3'-hydroxyl terminus. Reverse transcriptase products, whether partial or complete copies of the mRNA template, often possess short, partially double-stranded hairpins ("loops") at their 3' termini. In ^tne second reaction, these "hairpin loops" can be exploited as primers for DNA polymerase. Preformed DNA is required both as a template and as a primer in the action of DNA polymerase. The DNA polymerase requires the presence of a DNA strand having a free 3'-hydroxyl group, to which new nucleotides are added to extend the chain in the 5'-3' direction. The products of such sequential reverse transcriptase and DNA polymerase reactions still possess a loop at one end. The apex of the loop or "fold-point" of the double-stranded DNA, which has thus been created, is substantially a single-strand segment. In the third reaction, this single-strand segment is cleaved with the single-strand specific nuclease Sl to generate a "blunt end" duplex DNA segment. This general method is applicable to any mRNA mixture, and is described by Buell, et al. , J. Biol. Chem. , 253; 2483 (1978).

The resulting double-stranded cDNA mixture (ds-cDNA) is inserted into cloning vehicles by any one of many known techniques, depending at least in part on the particular vehicle being used. Various insertion methods are discussed in considerable detail in Methods In Enzymology, 68:16-18, and the references cited therein.

1 Once the DNA segments are inserted, the cloning vehicle is used to transform a suitable host. In a common form of this technique, these cloning vehicles can impart an antibiotic resistance trait to the host. Such hosts are generally prokaryotic or eukaryotic cells. Generally at this point, only a few of the transformed or transfected hosts will contain the desired cDNA, and the sum of all transformed or transfected hosts constitutes a gene "library". In the principle, the overall ds-cDNA library created by this method

10 provides a representative sample of the coding information present in the mRNA mixture used as the starting material.

If an appropriate oligonucleotide sequence is available, it can be used to identify clones of interest as follows. Individual transformed or transfected cells are iς grown as colonies on nitrocellulose filter paper. These colonies are lysed; the DNA released is rendered single strands and then strongly bound to the filter paper after heating. Such a sheet is then incubated ^' with a labeled oligonucleotide probe which is complementary to the

2o structural gene of interest. The probe hybridizes with the cDNA for which it is complementary, and hybrids are identified by autoradiography. The corresponding clones are characterized in order to identify one, or a combination of clones which contain all of the structural information for

25 the desired protein. The nucleic acid sequence coding for the protein of interest is isolated and reinserted into an expression vector. The expression vector brings the cloned ^• gene under the regulatory control of a specific prokaryotic or eukaryotic control element which allows the efficient

20 expression (transcription and translation) of the cloned ds-cDNA. Thus, this general technique is only applicable to those proteins for which at least a portion of their amino acid or DNA sequence is known and for which an oligonucleotide probe is available. See, generally, Maniatis, et al., supra.

More recently, methods have been developed to identify specific clones by probing bacterial colonies with antibodies specific for the encoded protein or interest. This method can only be used with "expression vector" cloning vehicles since elaboration of the product protein is required. The structural gene is inserted into the vector adjacent to regulatory gene sequences that control expression of the protein. The cells are lysed, either by the vector or by chemical methods, and the protein detected by the specific

10 antibody and a labeling system such as enzyme immunoassay. An example of this is the lambda gt-, system described by Young and Davis, Proc. Nat'l. Acad. Sci. USA, 80: 1194-1198 (1983) and Young and Davis, Science, 22: 778 (1983).

The von illebrand Factor (vWF or Factor VIIIR) is

■ _j c a large, adhesive plasma glycoprotein which is instrumental in mediating the attachment of platelets to damaged areas of • the circulatory system. In the absence of normal v f function, there _. is a defect in formation of the "platelet plug" of primary hemostasis, and a bleeding disorder, von

2 _Q Willebrand's disease, results. Existing evidence strongly suggests that vWf serves as a carrier for Factor VIIIC, and in that capacity, prolongs its (Factor VIIIC's) circulatory half-time. As a consequence, in patients with very low levels of vWf, there is also a decrease in the level of circulating coagulation Factor VIIIC (anti-hemophilic factor) , even though they have- the genetic capability to produce normal amounts of the latter protein. Therefore, patients with severe von Willebrand's disease also experience difficulty in developing adequate fibrin deposition in a 0 manner similar to individuals with classical hemophilia A. Thus, vWf is involved in critical roles in both the earliest and later steps in normal hemostasis [see Hoyer, Blood, 58: 1-13 (1981) , and Zimmerman, et al. , in Progress in

5

1 He atology, E.B. Brown, ed. , Vol. XIII, New York: Grune and Stratton, pp. 279-309].

In plasma, vWf circulates as a series of self-aggregated structures ranging from dimers of a 225 kd 5 subunit to polymers containing more than 50 such subunits [Legaz, et a _, J. Biol. Che . , 248: 3946-3955 (1983); Hoyer and Shainoff, Blood, 55: 1056-1059 (1980); Ruggeri and Zimmerman, J. Clin. Invest., 65: 1318-1325 (1980)]. The results of various studies have established that the largest 0 structures are most important in the intrinsic hemostatic role of vWf (Zimmerman, e_t al. , supra) , while oligomers of all sizes can function equally as carriers for Factor VIIIC [Davies, et al. , Thromb. Res., 22: 87-93, (1983)]. In view of the complex structure of vWf and its dual hemostatic role, it is not surprising that several type of von Willebrand's disease have been described and that the inheritance patterns are often complex (Zimmerman, _et al. , supra) .

Factor VIIIR biosynthesis in endothelial cells is the major source of plasma vWf. The 225 kd vWf subunit found _Q in plasma and in endothelial cell culture medium is first synthesized in endothelial cell cultures as a 240 kd intracellular proprotein precursor (provWf) [Wagner and Marder, J. Biol, Chem., 258:.2065-2067 (1983); Lynch, et al. , Proc. Natl. Acad. Sci., 80: 2738-2742, (1983)]. While provWF 5 may be in-itially detected as a monomer, it rapidly dimerizes, apparently by disulfide bond formation. ProvWF dimers are the predominant intracellular form of the vWf gene product in endothelial cells. They do not self-associate further and are not secreted [Lynch, t _^ al. , J. Biol. Chem. 258: _Q 12757-12760 (1983)]. However, provWF dimers do undergo a series of post-translational modifications involving addition of a sulfated moiety and cleavage of sequence(s) specific to

5

1 the precursor. These events result in the formation of 225 kd and 260 kd subunits. Just prior to secretion from the cell, mixed dimers of these species associate to form the series of vWf multimers typically found in plasma. Brief Description of the Invention

The invention provides replicable expression vectors incorporating a DNA sequence encoding human Factor VIIIR or fragment thereof and a self-replicating host cell system transformed or transfected thereby. The host system 0 is usually of prokaryotic, e.g., E^ coli, B. subtilis, or eukaryotic cells.

The human Factor VIIIR or fragment thereof is produced by a process which comprises (a) preparing a replicable expression vector capable of expressing the DNA 5 sequence encoding human Factor VIIIR in a suitable host cell system; (b) transforming said host system to obtain a recombinant host system; (c) maintaining said recombinant host system.under conditions permitting expression of said Factor .VIIIR-encoding DNA sequence to produce human Factor _Q VIIIR protein; (d) recovering said human Factor VIIIR protein or fragment thereof.

Preferably, the Factor VIIIR-encoding replicable expression vector is made by preparing a double-stranded complementary DNA (ds-cDNA) preparation representative of 5 Factor VIIIR mRNA and incorporating the ds-cDNA into replicable expression vectors. The preferred mode of recovering the human Factor VIIIR comprises reacting the proteins expressed by the recombinant host system with a reagent composition comprising at least one binding protein _Q specific for Factor VIIIR.

5

The invention provides a process for the cloning of human factor VIIIR in a self-replication recombinant host system which comprises: a) providing a heterogeneous mixture of mRNA enriched in mRNA encoding for human VIIIR; b) preparing ds-cDNA complementary to said enriched mixture; c) incorporating said ds-cDNA into a vector; and e) identifying those vectors in which said ds-cDNA has been incorporated.

In a further embodiment the invention provides a process wherein ds-cDNA encoding, factor VIIIR is put into a replicable expression vector capable of expressing the DNA sequence encoding human Factor VIII-R in a suitable self-replicating recombinant host system; b. transforming said host system to obtain a recombinant and host system; c. maintaining said recombinant host system under conditions permitting expression of said Factor VIII-R encoding DNA sequence to produce human

Factor VIII-R; and d. identifying said human Factor VIII-R.

In a further embodiment the invention provides a cDNA comprising a region whose polynucleotide is substantially CTG CAG TAT GTC AAG GTG GGA AGC TGT AAG TCT GAA GTA GAG GTG GAT ATC CAC TAC TGC CAG GGC AAA TGT GCC AGC AAA GCC ATG TAC TCC ATT GAC ATC AAC GAT GTG CAG GAC CAG TGC TCC TGC TGC TCT CCG ACA CGG ACG GAG CCC ATG CAG GTG GCC CTG CAC TGC ACC AAT. GGC TCT GTT GTG TAC CAT GAG GTT CTC AAT GCC ATG GAG TGC AAA TGC TCC CCC AGG AAG TGC AGC AAG TGA GGCTGCTGCAGCTGCATGGGTGCCTGCTGCTGCCTGCCTT

GGCCTGATGTGGCCAGAGTGCTGCCAGTCCTCTGCATGTTGTGCTCTTGTGCCCTT CTGAGCCCACAATAAAGGCTGAGCTCTTATCTGCAAAAAAAAAAAAAAAAAAC CCCCCCCCCCC.

1 In a further embodiment the invention provides vectors containing cDNAs encoding VIII-R and host cells transformed thereby.

In a final embodiment the invention provides probes 5 useful for identification and isolation of DNA fragments containing the encoding region of human Factor VIII-R comprising the vector of Claim 31. Such probes have utility as a means for recovering from a mixture of cDNAs by hybridization those cDNAs containing full length copies of 0 the VIIIR gene, which then may be cloned and expressed in a suitable vector system. Detail Description of the Drawings

Figure 1 provides a diagramatic illustration of a restriction map of pDL34. For simplicity, only relevant sites in the insert cDNA are shown. "Eco" and "Ori" refer to the direction of the unique EcoRI site and origin or replication, respectively . of the pBR322 vector.. The insert is flanked by PstI sites resultant from the GC tailing method of cloning. Of the o original nine colonies identified with the polysome-generated probe, all had sequences which hybridized to the 250 bp fragment flanked by internal PstI sites and to the 160 bp fragment to the right of the PstI sites. Eight of them also had sequences hybridizing to a probe made from the 2 kb 5 fragment flanked by internal PstI sites, but all of these inserts were considerably smaller than that of pDL34. Two other cDNA's which were analyzed contained sequences extending to the right of the internal PstI sites similar in size to that of pDL34. The oligo dT probe hybridized to the _Q shaded area to the right of the SacI site (see Figure 2) .

To prepare the fragment for subcloning into pSP64, pDL34 was digested to completion with SacI, and then partially digested with PstI. The 2.3 kb PstI-SacI fragment

5

was extracted from a low melting point agarose gel, and ligated to Pstl/SacI digested pSP64. The resultant plasmid is designated pSP64-34. Figure 2 illustrates the hybridization of 32P-oligo dT to PDL34

A southern blot was prepared from various restriction digests of pDL34 which had been electrophoresed through a 2% agarose gel. Lane 1, SphI (There is one recognition site in both the vector and the insert.); Lane 2, SacI (there is a recognition site in the insert only.), and EcoRI (there is a recognition site in vector and none in the insert) . Some anomalous cleavage was noted under the conditions of the digestion, leading to the appearance of two additional minor DNA species, one of which hybridized to the probe.); Lane 3, PstI (Recognition sites flank the insert, and there are two sites within the insert); Lane 4, pBR322 DNA digested with Avail and EcoRI (Used as molecular size- markers.) . The probe used here, was-prepared by labeling oligo (dT) , _g with gamma- 32P-ATP and T. polynucleotide kinase. Hybridization was at 33°C in 4XSSC and 0.1% SDS. The probe specifically hybridized to the portion of pDL34 to the right of the SacI site, as displayed in Figure 1.

Figure 3 illustrates the hybridization of a fragment of pDL34 to Endothelial Cell RNA 10 ug of endothelial cell (Lane 1 and 2) or SV80 cell (Lane 3) total RNA were electrophoresed through a 0.8% agarose gel and transferred to nitrocellulose. Hybridization was with the nick translated, gel purified approximately 800 bp PvuII fragment of pDL34 (see Figure 1) . The markers at the left denote the migration positions of the 28S and 18S rRNAs. The size of the band in the endothelial cell RNA lanes was estimated to be 9.5 kb, based on the migration of the rRNA specj.es.

Fiqure 4 illustrates the translation of synthetic mRNA derived from pSP64-34 in a rabbit reticulocyte lysate

A. Translation reactions (50 ul) were incubated for 60 minutes at 30°C and prepared for immunoprecipitation as described in the Example. A 2 ul aliquot was directly electrophoresed on a 12.5% polyacrylamide gel under reducing conditions. Lane 1, a reaction mixture with no mRNA added.

Lane 2, a reaction mixture containing approximately 50 ug/ml synthetic pSP64-34 RNA. The Figure depicts a fluorogram of

10 the stained and dried gel. The major band which migrates approximately 15 kd is obscured by a combination of the radioactive globin background and an artifact caused by the intense Coomassie blue stain of the endogenous globin on the fluorographic enhancer. The molecular size markers indicated

-ι r- at the left are from the bottom: alpha-lactalbu in, 14,200; soybean trypsin inhibitor, 20,100; trypsinogen, 24,000; carbonic anhydrase, 29,000; glyceraldehyde-3-phosphate dehydrogenase, 36,000; and egg albumin, 45,000.

B_. Immunoprecipitation of translation products

2o derived from pSP64-34 with rabbit anti-vWF and the effect of preincubation of the antibody with highly purified plasma vWf. A 300 ul translation mixture was incubated with 50 ug/ml of pSP64-34 RNA for 1 hour at 30°C and 60 ul aliquots prepared for immunoprecipitation as detailed in Experimental 5 Procedures. Electrophoresis was on a 12.5% SDS polyacrylamide gel under reducing conditions. Lane 1, rabbit antiserum to human fibronectin, 3 ul. Lanes 2-5, rabbit antiserum to vWf, 3 ul. Where indicated, highly purified plasma vWf was preincubated for two hours with the rabbit 0 anti-vWf (Lane 3, 650 ng; Lane 4, 130 ng; Lane 5, 26 ng) and the -resulting antibody-antigen mixture was used to immunoprecipitate the translation reaction aliquots. No polypeptides were immunoprecipitated by rabbit anti-vWf if

5

-1 I- the synthetic RNA was omitted from the translation reaction

(data not shown) . Molecular size markers were the same as shown in Figure 4A.

Figure 5 illustrates the monoclonal anti-vWf Immunoprecipitation of pSP64-34 derived translation products

Translation mixtures were incubated, prepared from immunoprecipitation and electrophoresed as noted in Figure

4B. Immunoprecipitations were performed with:

A. Lane 1, rabbit anti-vWf; Lanes 2-4, monoclonal anti-vWf 0 antibodies AVW-3, AVW-2, and AVW-1, respectively, AVW-1 reacts preferentially with the two major species recognized by the polyclonal anti-vWf. Molecular size markers indicated on the left are from the bottom: cytochrome C, 12,500; soybean trypsin inhibitor, 20,100; trypsinogen, 24,000; _t=. glyceraldehyde-3-phosphate dehydrogenase, 36,000.

B. Lane 1, rabbit anti-fibronectin; Lane 2, monoclonal anti-t yroglobulin; Lanes 3-5 monoclonal anti-vWf antibodies RG32, .2.2, and 5.5, respectively. Antibodies 2.2 and 5.5 react with the same synthetic peptides as AVW-1 in Figure 5A. o ^Tne molecular size markers are the same as in Figure 4A.

C. Lane 1, rabbit anti-vWf; Lane 2, rabbit anti-fibronectin; Lane 3, monoclonal anti-thyroglobulin; Lanes 4 _^ -7, monoclonal anti-vWf AVW-1, 2.2, 5.5, and TGI. Monoclonal TGI reacts preferentially with the major 30 kd species and a minor 5 approximately 17 kd species recognized by the polyclonal antiserum, but not with the major bands recognized by antibodies AVW-1, 2.2 and 5.5. Another monoclonal anti-vWf, TG2 reacted in a similar fashion to TGl. Molecular size markers are as in Figure 4A. 0

5

Figure 6 provides the sequence of the 3' end of pDL34

Sequencing of pDL34 cDNA subcloned into M13mpl8 was performed by the method of Sanger, et al. , [Proc. Natl. Acad.

Sci. , USA, 74: 5463-5467 (1977)] with Amersham dideoxy sequencing reagents and standard procedures suggested by the supplier. The typical polyadenylation signal (AATAAA) is underlined. Note that the oligo dT binding region which is

18 bp downstream from the AATAAA contains exactly 18 dA residues, the same length as the primer used to initiate reverse transcription. The predicted amino acid sequence in each reading frame is shown until the. first termination codon (-) is reached. The underlined sequence matches exactly that reported for the C-terminus of plasma vWf. Detailed Description of the Invention As used herein, "human Factor VIIIR" denotes human

Factor VIIIR (von Willebrand's Factor) produced by cell or cell-free culture systems, in bioactive forms having the capacity to initiate normal blood coagulation as does Factor VIIIR native to the human plasma. Different alleles of Factor VIIIR may exist in nature. These variations may be characterized by difference (s) in the nucleotide sequence of the structural gene coding for proteins of identical biological function. In addition, the location and degree of glycosylation as well as other post-translationa-1 modifications may vary and will depend to a degree upon the nature of the host and environment in which the protein is produced. It is possible to produce analogs having single or multiple amino acid substitutions, deletions, additions, or replacements. All such allelic variations, modifications, and analogs resulting in derivatives of human Factor VIIIR which retain the biologically active properties of native human Factor VIIIR are included within the scope of this invention.

Recombinant vectors and methodology disclosed herein are suitable for use in host cells covering a wide range of prokaryotic and eukaryotic organisms. In general, of course, prokaryotics are preferred for the cloning of DNA sequences and in the construction of vectors. For example,

E^ coli K12 strain HB101 (ATCC No. 33694) is particularly useful. Of course, other microbial strains may be used.

Vectors containing replication and control sequences which are derived from species compatible with the host cell or system are used in connection with these hosts. The vector ordinarily carries an origin of replication, as well as characteristics capable of providing phenotypic selection in transformed cells. For example, E^ coli can be transformed by the vector pBR322, which contains genes for ampicillin and tetracycline resistance [Bolivar, et al. , Gene, 2: 95 (1977)].

These antibiotic resistance genes provide a means of identifying transformed cells. The expression vector-may also contain control elements which can be used by the vector f° ^r expression of its own proteins. Common prokaryotic control elements used for expression of foreign DNA sequences in E_^ coli include the promoters and regulatory sequences derived from the B-galactosidase and tryptophan (trp) operons of E_^ coli, as well as the pR and pL promoters of bacteriophage lambda. Combinations of these elements have also have been used (e.g., TAC, which is a fusion of the trp promoter with the lactose operator) . Other promoters have also been discovered and utilized, and details concerning their nucleotide sequences have been published enabling the skilled worker to combine and exploit them functionally.

Host cells can prepare human Factor VIIIR proteins which can be of a variety of chemical compositions. The protein is produced having methionine as its first amino acid

(present by virtue of the ATG start signal codon naturally existing at the origin of the structural gene or inserted before a segment of the structural gene) . The protein may also be intra- or extracellularly cleaved, giving rise to the amino acid which is found naturally at the amino terminus of the protein. The protein may be produced together with either its signal polypeptide or a conjugated protein other than the conventional signal polypeptide, the signal polypeptide of the conjugate being specifically cleavable in an intra- or extracellular environment. Finally, Factor VIIIR may be produced by direct expression in mature form without the necessity of cleaving away any extraneous polypeptide. This.would include any post-translational modifications required for biologic activity.

Recombinant host cells refer to cells which have been transformed with vectors constructed using recombinant DNA techniques. As defined herein. Factor VIIIR is produced as a consequence of this transformation. Factor VIIΪR is produced by such cells are referred to as "recombinant Factor VIIIR". Recombinant and Screening Methodology

The procedures below are but some of a wide variety of well established procedures to produce specific reagents useful in the process of this invention. The general procedure for obtaining a messenger RNA (mRNA) mixture is to prepare an extract from a tissue sample or to culture cells producing the desired protein, and to extract from a tissue sample or to culture cells producing ^" the desired protein, and to extract the mRNA by a process such as that disclosed by Chirgwin, et al. , Biochemistry, 18: 5294 (1979). The mRNA is enriched for poly(A) mRNA-containing material by

chromatography on oligo (dT) cellulose or poly(U) Sepharose, followed by elution of the poly(A) containing mRNA-enriched fraction.

The above pol (A) containing mRNA-enriched fraction is used to synthesize a single-strand complementary cDNA (ss-cDNA) using reverse transcriptase. As a consequence of DNA synthesis, a hairpin loop is formed at the 3' end of the DNA which will initiate second strand DNA synthesis. Under appropriate conditions, this hairpin loop is used to effect synthesis of the second strand in the presence of DNA polymerase and nucleotide triphosphates.

The resultant double-strand cDNA (ds-cDNA) is inserted into the expression vector by any one of many known techniques. In general, methods, etc., can be found in Maniatis, supra, and Methods in Enzymology, Vol. 65 and 68 (1980); and Vol. 100 and 101 (1983). In general, the vector is linearized by at least one restriction endonuclease, which will produce at least two blunt or cohesive ends. The ds-cDNA is ligated with or joined to the vector insertion site.

If prokaryotic cells or other cells which contain substantial cell wall material are employed, the most common method of transformation with the expression vector is calcium chloride pretreatment as described by Cohen, R.N., et al. , Proc. Nat'l. Acad. Sci. USA, 69: 2110 (1972) . If cells without cell wall barriers are used as host cells, transfection is carried out by the calcium phosphate precipitation method described by Graham and Van der Eb, Virology, 62: 456 (1973). Other methods for introducing DNA into cells such as nuclear injection or protoplast fusion, have also been successfully used. The organisms are then cultured on selective media and proteins for which the expression vector encodes are produced.

1 Clones containing part or all of the gene for

Factor VIIIR are identified with specific radiolabeled oligonucleotides directed against part or all of the Factor

VIIIR genome. These oligonucleotide probes are prepared by oligo (dT) _lg primed reverse transcription of polysome-isolated poly A RNA in the presence of

32 P-deoxynucleotide triphosphates or by nick translation.

Such clones are detected using in situ colony hybridization. Northern blotting, and Southern blotting. These procedures 0 are described in Maniatis, supra.

Clones containing the entire sequence of Factor VIIIR may be identified using specific antibodies directed against part or all of the Factor VIIIR protein. This - requires that the ds-cDNA be inserted into a vector 5 containing appropriate regulatory nucleic acid sequences adjacent to the insertion site. These regulatory sequences initiate transcription and translation of those ds-cDNA molecules inserted in the vector. Those clones containing - Factor VIIIR cDNA sequences correctly positioned relative to o ^tne regulatory sequences synthesize part or all of the Factor VIIIR protein.

Alternatively, in non-expressiohal cloning systems, clones containing- the Factor VIIIR cDNA can be confirmed by in vitro transcription of the putative cDNAs. The resultant 5 mRNAs are translated _in vitro in rabbit reticulocyte, wheat germ, or Xenopus laevis oocyte translation systems. Resultant Factor VIIIR protein production is determined by immunoprecipitation of the translation mixture supernatants using polyclonal or monoclonal antibodies specific for Factor VIIIR.

Deposits of a strain useful in practicing the invention

A deposit of a biologically pure culture of the following strains was made with the American Type Culture

Collection, 12301 Parklawn Drive, Rockville, Maryland on April 25, 1985 the accession number indicated was assigned after successful viability testing, and the requisite fees were paid. Access to said culture will be available during pendency of the patent application to one determined by the

Commissioner to be entitled thereto under 37 C.F.R. §1.14 and 35 U.S.C. §122. All restriction on availability of said culture to the public will be irrevocably removed upon the granting of a patent based upon the application and said culture will remain permanently available for a term of at least five years after the most recent request for the furnishing of a sample and in any case for a period of at least 30 years after the date of the deposit. Should the culture become nonviable or be inadvertently destroyed, it will be ^' replaced with a .viable culture(s) of the same taxonomic description.

Strain/Plasmid ATCC No. pDL34 53109

! EXAMPLE

Cells and Cell Culture Conditions

Early passage human umbilical vein endothelial cells were obtained from the laboratory of Dr. Michael

5 Gimbrone (Brigham and Women's Hospital, Boston). The cells were grown as previously described (Lynch, et al. , 1983, supra) . Specifically, they were seeded on a matrix of human fibronectin in Medium 199 supplemented with 20% calf serum and 150 ug/ml of endothelial cell growth factor. In 10 addition, heparin sulfate to 100 ug/ml was added to the culture medium [Thornton, et al. , Science, 222: 623-625

(1983)]. Cells were utilized between passages 2-20.

35 Metabolic labeling with S-methionine was performed as previously noted (Lynch, et al. , 1983, supra) .

!_; Production of RNA and Polysomes

Endothelial cells were grown in roller bottles for experiments involving production of RNA or polysomes. Total cellula-r RNA was isolated by the guanidinium isothiocynate" method of Chirgwin, et al. , [Biochemistry, 18: 5294-5299

2p (1979)], and poly A-containing RNA was purified by chromatography on poly U-Sephadex. Total RNA was size fractionated by velocity sedimentation in sucrose gradients as described [Aloni and Attardi, J\_ Mol. Biol. , 55: 271-276 (1971) ] , and material sedimenting faster than the central _C part of the 28S rRNA peak was pooled. vWf-specific polysomes were immunoisolated from an endothelial cell lysate prepared with 1% Nonidet P-40 by the general method of Shapiro and Young [J\_ Biol. Chem. , 256: 1495-1946 (1981) ] . A significant change was that the polysome-antibody mixture was passed over

30 a column of glycine-inactivated CNBr-Sepharose immediately before passage over the protein A-Sepharose immunoisolation column. This step helps to remove material that otherwise

5

binds non-specifically to protein A-Sepharose. vWf-specific poly A-containing RNA was isolated from the eluate of the protein A-Sepharose column by oligl-dT cellulose chromatograph . 5 cDNA Construction and Probe Labeling

A cDNA library was prepared in pBR322 from poly A-containing RNA isolated from total RNA greater that 28S by the GC tailing method using the detailed procedures described by Maniatis, et_ al. , supra. First strand cDNA was 0 synthesized from the RNA which had been poly U-Sephadex selected, using an oligo (dT) _lg primer (New England Biolabs) and avian myeloblastosis virus reverse transcriptase (Life Sciences, St. Petersburg, Fla.). The RNA was then removed by alkaline hydrolysis, and second strand DNA was synthesized by _c; hairpin loop priming following addition of E. coli DNA polymerase I large fragment. An additional incubation with reverse transcriptase was employed to achieve greater second strand lengths. The hairpin loop was removed with Sl nuclease and the ends repaired with DNA polymerase I large _Q fragment. Oligo dC tails of 10-15 residues were added with terminal transferase, and the cDNAs annealed to oligo dG-tailed pBR322. A cDNA library of approximately 60,000 independent recombinant bacteria was then constructed.

To synthesize a probe with which to screen this library, a polysome fraction from endothelial cells was prepared and incubated with monospecific rabbit anti-vWf IgG. Antibody-bound polysomes were then purified by immunoaffinity chromotography on protein A-Sepharose. Poly A-containing RNA was isolated from the purified polysomes and reverse 0 transcribed in the presence of oligo dT primer and all four alpha- 32-P-deoxynucleotide triphosphates (40 Ci/mmole) at 50 uM. The average probe length was approximately 300 bp.

5

1 Other 32-labeled DNA probes were prepared by nick translation with DNA polymerase I or 5' end-labeling with polynucleotide kinase (Maniatis, et al. , supra) . From the immunoisolated

RNA of approximately 5 x 10 8 cells, 2 x 105 cpm of reverse transcript were synthesized.

Library Screening

In situ colony hybridization. Northern and Southern blotting were performed as described by Maniatis, et al. , supra. In the initial screen of 3,000 colonies with this

10 probe, nine were found to hybridize specifically. Further analysis of the cDNAs from the selected colonies revealed that they all cross hybridized and shared a common end as defined by restriction enzyme and molecular hybridization analyses (see Figure 1) . The strong homology among the

-, r- cDNAs suggested that the probe generated from the immunoisolated polysomes was largely, if not completely, homogeneous. A restriction map of the 2.4 kb insert of the largest of these initial isolates, pDL34, is shown in Figure 1. 0 The presence of a common end in all of these clones suggested that it might contain the initiation site 'for the reverse transcription, and possibly, the DNA complement of the poly A tail of the mRNA. To investigate this possiblity,

32

P-oligo (dT) _lg was hybridized to pDL34 immobilized on a 5 .Southern blot (Figure 2) . From these results, it may be seen that a unique hybridization site is detected in pDL34. A single 3.9 kb band was detected in an SphI digest (Lane 1). There is a single SphI recognition site within the pDL34 insert (as shown on the map in Figure 1) as well as one in 0 the pBR322 vector. The probe hybridized to the fragment to the right of the SphI site present in the insert. Two DNA fragments were generated by digestion of pDL34 with the enzymes SacI (which cuts in the insert) and EcoRI (which cuts

5

1 in the vector). As shown in Lane 2, the probe hybridized to the 3.7 kb fragment, i.e., to the sequence on the right side of the SacI site. Lane 3 contains a PstI digest of pDL34, and no hybridization of oligo dT was noted. PstI digestion 5 produced three DNA fragments from the insert, as well as an intact pBR322 vector backbone. In other experiments, the small ( 160 bp) restriction fragment generated by PstI at the right end of the cloned insert failed to bind to nitrocellulose under the conditions employed for Southern 10 transfer.

Therefore, the oligo dT binding region of pDL34 is located in the approximately 50 bp region between the SacI site and the end of the cloned insert (also see Figure 6) . This region was detected in all of the initially isolated !_ clones. This finding suggests that the direction of the reverse transcription is from the oligo dT-hybridizing fragment to the left as denoted in Figure 1. It also suggests the orientation of the reading frame within the cDNA and implies that pDL34 is likely to be a complement of the 3' 20 end of its template mRNA.

Identification of pDL34 as a vWf cDNA Clone

To determine if pDL34 was derived from a large, endothelial cell-associated mRNA, a probe was made by nick-translation of an 800 bp internal PvuII fragment of its 25 insert (Figure 1) . The 32P-labelled DNA was then hybridized to a Northern blot containing total RNA from endothelial cells and SV80 cells (as established human cell line which does not express vWf) . As shown.in Figure 3, a single band of approximately 9.5 kb was detected with endothelial cell

30 RNA, and no bands were observed with the SV80 RNA. Hence, pDL34 is homologous to a large, apparently endothelial cell-specific RNA species.

35

Although pDL34 had properties consistent with it being a partial vWF cDNA clone, direct verification of this possibility presented a problem. The complete amino acid sequence of vWf was not published, and since it seemed likely 5 that pDL34 represented the C-terminal end of the protein. One would not easily be able to determine its identity by comparing DNA and protein sequences. However, a large number of monoclonal antibodies of vWf were available [Fulcher and Zimmerman, Proc. Natl. Acad. Sci. USA, 79: 1648-1652 (1982);

10 and Schullek, et a _, J._ Clin. Invest. , 73: 421-428 (1984)]. Therefore, it seemed possible that iii vitro translation of endothelial cell mRNA purified by hybridization to pDL34 and subsequent monoclonal anti-vWf immunoisolation of translation products would be a reasonable demonstration of the identity iς of pDL34. Since pDL34 was originally identified with a probe whose synthesis required prior selection with polyclonal anti-vWf, verification of its identify would require independent immunoisolation with vWf monoclonal antibody to eliminate any possibility that pDL34 encoded an impurity

2o present in the vWf used to raise the polyclonal antiserum. Primarily because of limited vWf mRNA availability, identification of vWf translation products in pilot translation studies with endothelial cell A-containing RNA was not possible. In an effort to eliminate this supply 5 problem, SP6 RNA polymerase was employed to generate synthetic mRNA from a pDL34 template- inserted downstream of a clonal SP6 promoter sequence [Melton, et al. , Nucleic Acids Res. , 12: 7035-7056 (1984)]. Since the 2.4 kb pDL34 insert is smaller than the predicted size of full -length vWf cDNA, _Q SP6-generated synthetic mRNA would be truncated, presumably at its 5' end. It would be expected to lack any specific ribosome binding sequences as well as the natural vWf

5

1 initiation codon. Therefore, if the synthetic RNA served as a template for translation initiation of protein synthesis would need to occur at one or more fortuitous, downstream methionine codons. 5 To produce the desired plasmid, the bulk of the pDL34 cDNA insert, ^' in the form of a 2.3 kb Pstl-SacI fragment, was inserted into the polylinker of pSP64 using the orientation predicted for sense transcription from the oligo dT binding results (see Figures 1 and 2). 2.3 kb synthetic

10 mRNA was then generated in vitro with SP6 polymerase. This RNA was added to a rabbit reticulocyte lysate containing ³⁵ S-m-thionine.

In vitro translation mixtures were prepared for immunoisolation by 20-fold dilution into 20 mM Tris, pH 8, iς 0.15 M NaCl containing 1% Nonidet P-40, 0.1% sodium dodecyl sulfate (SDS) , and 200 u/ml aprotinin. Phenylmethylsulfonyl fluoride was then added to 50 ug/ml. Diluted reaction mixtures were preabsorbed for 2-4 hours at 4°C with normal rabbit -serum and protein A-Sepharose for 1 hour at 4°C. When

2o monoclonal antibodies were employed rabbit antiserum to mouse IgG (Cappel) was added four hours prior to addition of protein A-Sepharose. SDS gel electrophoresis of Ln vitro translation products was performed in 12.5% polyacrylamide running gels with 4% polyacrylamide stacking gels as

2 described by Laemmli [Nature, 227: 640-655 C1970) ] . As seen in Figure 4A, translation of several discrete polypeptides was directed by the truncated, synthetic mRNA, although synthesis was relatively inefficient considering the quantity and quality of the RNA available. When such translation

O _Q reaction mixtures were subjected to immunoprecipitation by commercially available monospecific rabbit anti-vWf antibody, with rabbit anti-fibronectin antibody as a control, a series

35

of approximately 15-30 kd polypeptides was specifically precipitated by the former. No bands were precipitated by the antifibronectin antibody. The reaction of these polypeptides with vWf antiserum could be readily inhibited by 5 preincubation of the antiserum with highly purified vWf

(Figure 4B) . The rate of electrophoretic migration of these polypeptides varied slightly in different translation reactions, perhaps as a function of the degree of post-translational modification. However, the relative 0 migration rates and intensities of the major discrete species was constant.

A series of 30 monoclonal antibodies to vWf was then screened for reactivity with these in vitro translation products. Strong, specific reactions were seen with four of

-,ς these- (Figure 5). Weak reactions were noted with three others (not shown) . It should be noted, that the strongly reactive monoclonal antibodies recognized native vWf (antibodies 2.2, 5.5, AVW-1 in Figures 5A and 5B) and trypsin-treated vWf (antibody TGI in Figure 5C) . By

20 contrast, 16 of the 30 monoclonals screened recognized reduced and denatured vWf, but none of these reacted strongly with the in vitro translation products under the conditions employed. This suggests that the epitope(s) involved are capable of assuming their native structure despite their 5 occurrence on protein fragments. Recognition of these in vitro-generated polypeptides by some of a heterogeneous collection of monoclonal anti-vWf antibodies establishes pDL34 as an authentic vWf cDNA clone and the 9.5 kb mRNA identified in Figure 3 as vWf mRNA. _Q ^{Tne DNA} sequence of the 3* region of pDL34 is shown in Figure 6. There is a stop codon in each of the three possible reading frames, indicating that the 3' end of this

5

1 clonal insert is an untranslated region. The predicted amino acid sequence is shown in each of the three reading frames up to the first stop codon encountered. The final ten amino acids of the longest open reading frame are underlined and 5 correspond exactly to the decapeptide recently described as comprising the C terminus of plasma vWf (Tatani, e_t al. , 1984, abstract. Circulation 70, 11-210). This finding further corroborates the identification of pDL34 as a Vwf cDNA clone and indicates that the precursor-specific amino 0 acid sequences in pro-vWf extend from the N terminus of the plasma protein.

The oligo(dT) binding site of pDL34 may be seen as a stretch of 18 dA residues, followed immediately by 12 dCs synthesized during library construction. The hexanucleotide 5 AATAAA is located 18 bases upstream of the run of As. This is the characteristic poly(A) ^' addition sequence (Proudfoot and Browniee, Nature 263: 211-214, 1976) at an appropriate distance (Fitzgerald and Shenk, Cell 24: 251-260, 1981) for the detected dA tract- to be a portion of the poly(A) tail of o vWf mRNA. Thus, it is possible that the (dA),„ represents a portion of the poly(A) tai.l of vWf mRNA; in this case, the 3' untranslated region would extend for less than 200 bp. However, since the original primer for the cDNA library was oligo (dT) _lg it is also possible that the (dA),„ represents an 5 internal complement of the primer located in the 3' untranslated region of the vWf mRNA.

In attempting to account for the 9.5 kb observed as the approximate length of vWf mRNA, one must consider the minimum coding unit for plasma vWf to be 5.4 kb. However, _Q glycoproteins can be difficult to size accurately by discontinuous SDS-polyacrylamide gel electrophoresis; there have been a number of estimates for the vWf subunit larger than the 225 kd we reported. In particular, Legaz, et al.

5

1 [J. Biol. Chem. 248: 3946-3955 (1973)] have suggested that the size of the plasma vWf subunit is 250 kd, and they estimated that the protein contains 2100 amino acids. Similarly, pro-vWf can be construed to be 50 kd larger than the secreted subunit (Wagner and Marder, 1983) . Therefore, as much as 7.6 kb of mRNA sequence could be required to encode pro-vWf. It the AATAAA sequence determined here is the true poly (A) addition signal, then 1.7 kb of mRNA is still unassigned. While large 3' 0 untranslated regions are common among eukaryotic mRNAs, we are not aware of any reported 5 ' untranslated regions of this magnitude. For example, the 3' and the 5' untranslated regions of human factor VIIIC mRNA contain 1.8 and 0.2 kb, respectively. Thus, the protein is synthesized as an 5 even larger, currently unrecognized precursor; or the mRNA migrates anomalously in formaldehyde agarose gels; and/or vWf mRNA may have an unusually extensive 5' untranslated region. The advantages of the SP6 system for generating - intact mRNA to verify the identity of full-length cDNA clones o have been demonstrated previously (Toole, e al. , Nature 312: 342-347, 1984) . As is disclosed herein, a partial cDNA clone can be similarly used. It might be expected that this approach would be most useful for the identification of cDNA clones encoding the N terminus of a given protein. Such 5 cDNAs might be expected to yield more readily translated mRNAs because of the presence of the natural initiation codon and, possibly, ribosome binding sites. However, pDL34 contains the C terminus of vWf, as indicated by DNA sequence information (Figure 5). Thus, the presence of the natural 5' _Q end of the mRNA was not a requirement for its _j ln vitro translation. Although inefficient, the extent of _in vitro translation of pDL34 was adequate for the present purpose. Similarly, from ther odynamic considerations of the folding

of newly synthesized proteins, it might be expected that translation products lacking the N-terminal portion of the intact protein may not assume a native conformation.

However, all synthetic mRNA-directed polypeptides that can be identified on gels of total translation products were efficiently immunoprecipitated by the polyclonal serum raised against the native protein. Also, the reacting monoclonal antibodies recognize epitopes present on the native protein.

It is possible that the presence of the C terminus may have aided antibody recognition, as protein termini may be generally more antigenic than internal sequences (Walter, et al. , Proc. Nat'l. Acad. Sci. USA, 77: 5197-5200, 1980; Lerner, et al. , Cell 23: 309-10, 1981). However, the anti-vWf monoclonals recognized more than one subset of these translation products, so in the aggregate, their activity appears to be against epitopes from more than one distinct region. Thus, it is possible that some of the polypeptides do not contain the authentic C terminus and were immunoprecipitated by antibodies to internal epitopes. Finally, this method of verification may prove to have general applicability in defining the identity of certain cDNA cloning products in situations where antibodies to a protein are available and protein sequence information is not. Such a situation is not unusual for many large proteins and is particularly common in the study of certain cell-surface antigens.

The observation that cloned fragments which are subsequently transcribed and translated i_n vitro retain the ability to react with monoclonal antibodies directed thereto, provides a means for developing a convenient library of DNA fragments encoding epitopes of monoclonal antibodies. Epitopes, as used herein being synonomous with antigenic

determinent sites, are those regions of an antigen which are specifically reactive with the antigen combining site of an antibody molecule.