WO1990005744A1 - Hdl-binding proteins - Google Patents

Abstract

Substantially pure mammalian proteins having the following characteristics are disclosed: (a) molecular weight of approximately 110,000 Daltons; (b) bind HDL3; (c) bind apoA-I proteoliposomes and apoA-II proteoliposomes; and (d) contain approximately 10 % carbohydrate by weight. The present invention also discloses a substantially pure protein having the following characteristics: (a) a molecular weight of approximately 38,000 Daltons; (b) binds HDL3; and (c) contains approximately 331 amino acid residues. DNA sequences encoding these proteins, as well as cells transfected or transformed with an expression vector containing such a DNA sequence are also disclosed. Antibodies that bind to these proteins, methods for isolating an HDL3-binding protein and methods for detecting a genetic defect in an HDL-binding protein in a patient are also disclosed.

Description

Description

HDL-BINDING PROTEINS

Statement of Government Interest

This invention was developed in part with support from National Institutes of Health Grant 5 R01 HL 31194-05. The Government has certain rights in this invention.

Cross-Referenσe to Related Application

This application is a continuation-in-part of U.S. Serial No. 07/347,855, filed May, 5, 1989, which is a continuation-in-part of U.S. Serial No. 07/273,388, filed November 18, 1988, which application is pending.

Technical Field

The present invention relates to cellular binding proteins, DNA sequences encoding them, methods for their production, an . assay systems employing them. More specifically, it relates to binding proteins for high-density lipoprotein, to assay systems useful for identifying high-density lipoprotein analogs, and to the analogs themselves.

Background of the Invention

Accumulation of cholesterol in the vascular endothelium is believed to be a key event in the development of atherosclerosis (for review, see Ross, New Enσ. J. Med. 314:488-97. 1986). Early in the development of atherosclerotic lesions, the artery walls are penetrated by cholesterol-containing atherogenic particles. These particles are believed to be derived from low-density lipoproteins (LDL) or to be minor subpopulations of sterol-rich lipoproteins. They are recognized as foreign by macrophages, which pick up the particles, integrate the cholesterol, and thereby become foam cells. Accumulation of foam cells is the first stage of fatty streak formation. These fatty streaks then enlarge through the gradual accumulation of lipid- containing acrophages and smooth muscle cells, eventu¬ ally developing into fibrous plaques rich in smooth muscle cells.

The structure and activity of plasma lipo¬ proteins have been reviewed by Gotto et al. fMeth. in Enzvmoloα 1£_3:3=41, 1986). High-density lipoproteins (HDL) are a class of plasma lipoproteins which consist of cholesterol, cholesteryl esters, phospholipids, triglycerides, and apolipoproteins (principally apoA-I and apoA-II) . Two main subclasses of HDL (HDL₂ and HDL₃) may be separated on the basis of density. A third subclass, designated HDL^ or HDL_C, contains apolipopro- tein E, but not apoA. HDL have been implicated in the transport of cholesterol (principally in the form of cholesteryl esters) from peripheral .tissues to the liver, where the cholesterol is catabolized or excreted (Glomset, Lipid Res. 9.:155-67, 1968). Other studies (Oram et al., J. Clin. Invest. 72.:1611-21, 1983) have shown that HDL can promote the efflux of cholesterol from cultured cells, presumably by binding to those cells and, through a second messenger response, stimu¬ lating the transfer of intracellular cholesterol across the cell membrane. The cholesterol is then absorbed by HDL or other plasma protein(s) and transported to the liver. In addition, HDL have been shown to bind to steroidogenic tissues in rats and are believed to participate in the delivery of cholesterol to these tissues. Thus, their biological role is complex and appears to involve a variety of tissue-specific effects. HDL carry about one-third of the total serum cholesterol in humans.

HDL have been shown to bind to extrahepatic nonsteroidogenic cells in culture with high affinity in a saturable manner. This binding is up-regulated by loading cells with unesterified cholesterol. Although this suggests that the HDL-induced efflux of cholesterol from cultured cells is mediated through binding of HDL to a cell surface receptor, the existence of such a receptor is disputed in the literature and conflicting data have been reported. Chen et al. fJ. Biol. Chem. 255:9162-67, 1980) demonstrated that rat testes have a specific high-affinity binding site for HDL which is not sensitive to proteolytic enzymes, suggesting that the testes and other steroidogenic tissues of rats obtain cholesterol from HDL by a cellular process not involving receptor proteins. Gwynne and Hess (J. Biol. Chem. 255:10875-83, 1980) found that rat adrenal cells possess a specific saturable uptake mechanism for HDL cholesterol, but the cell binding mechanism of HDL was not characterized. Fidge et al. (Biochem. Biophvs. Res. Co m. 129:759-65. 1985) and Fidge fFEBS Lett. 199:265- 68, 1986) isolated 78,000 Dalton HDL-binding proteins from membranes of several cell types. Tabas and Tall (J . Biol. Chem. 25.9:13897-905, 1984) studied the asso¬ ciation of HDL₃ with various cell types and concluded that this association is not mediated by specific ligand and receptor proteins, but involves lipid interactions. Mendel et al. ⁽J . Biol. Chem. .263:1314-19, 1988) studied the effects of radiation inactivation on HDL-binding sites on human fibroblasts and concluded that they were composed of peptides too small to be considered biologic receptors. Given the role of HDL in the removal of cholesterol from peripheral cells, HDL or HDL agonists have a potential application in the treatment of atherosclerosis. Indeed, there is a negative correla¬ tion between HDL levels and the incidence of coronary artery disease. HDL agonists could be used to stimulate the translocation of cholesterol from intracellular membranes to the cell surface, where it would be picked up by HDL or other transport protein(s) . Through the use of an assay system employing an HDL-binding protein, these agonists can be identified. Suitable agonists include antibodies to the HDL-binding protein and other protein and nonprotein molecules.

In addition, a cloned DNA fragment encoding an HDL-binding protein could be used to identify individu¬ als at risk for developing atherosclerosis. Through analysis of restriction fragment length polymorphisms (RFLPs) or DNA sequences, deficiencies in HDL-binding proteins could be identified. Furthermore, monoclonal antibodies could be used to screen blood cells for the absence of functional HDL-binding protein(s) .

Disclosure of the Invention

Briefly stated, the present invention discloses substantially pure mammalian proteins having the following characteristics: (a) molecular weight of approximately 110,000 Daltons; (b) bind HDL₃; (c) bind apoA-I proteoliposomes and apoA-II proteoliposomes; and (d) contain approximately 10% carbohydrate by weight. The proteins in certain embodiments may also be charac¬ terized as having an isoelectric point between about 5.8 and 6.0, and may be derived from bovine or human cells. Within a related aspect, the present invention discloses an isolated DNA molecule encoding a protein having the following characteristics: (a) a molecular weight of approximately 100,000 Daltons exclusive of carbohydrate; (b) binds HDL₃; and (c) binds ApoA-I and ApoA-II proteoliposomes. Within one embodiment, the DNA sequence encodes the amino acid sequence shown in Figure 3 from glycine, amino acid number ^'2, to glutamic acid, amino acid number 160.

Within another embodiment, the DNA sequence encodes the amino acid sequence shown in Figure 4 from methionine, amino acid number 1, to glutamic acid, amino acid number 1195. Within yet another embodiment, the DNA sequence encodes the amino acid sequence shown in Figure 5 from methionine, amino acid number 1, to isoleucine, amino acid number 1292.

The present invention also discloses methods for isolating an HDL₃-binding protein, comprising: (a) solubilizing mammalian cell membrane proteins to produce an enriched fraction; (b) exposing the enriched fraction to an antibody that binds to a protein that binds HDL₃, to produce a bound fraction and an unbound fraction; (c) recovering the bound fraction; and (d) isolating the HDL₃ binding protein from the bound fraction. The method may also include the step of load¬ ing the mammalian cells with cholesterol prior to the step of solubilizing. In another aspect of the present invention, an isolated DNA molecule encoding a protein having the following characteristics is disclosed: (a) a molecular weight of approximately 38,000 Daltons; (b) binds HDL3; and (c) contains approximately 331 amino acid residues. Within one embodiment, the protein comprises the amino acid sequence shown in Figure 1 from glycine, amino acid 2, to glutamic acid, amino acid 331.

Within a related aspect of the present invention, a substantially pure protein is disclosed having the following characteristics: (a) a molecular weight of approximately 38,000 Daltons; (b) binds HDL3; and (c) contains approximately 331 amino acid residues. Within certain embodiments, the protein reacts with polyclonal antisera to a 110,000-Dalton HDL3-binding protein from bovine aortic endothelial cells.

Cells transfected or transformed with an expression vector comprising a promoter operably linked to a DNA sequence encoding a protein as described above are also disclosed. In addition, antibodies that bind to the proteins described above are also disclosed. In yet another aspect of the present invention, methods for detecting a genetic defect in an HDL-binding protein in a patient are disclosed. The methods generally comprise: (a) isolating DNA from a 5 patient; (b) cleaving the DNA with a restriction endo- nuclease to produce DNA fragments; (c) hybridizing the DNA fragments to a DNA molecule comprising a portion of the nucleotide sequence of Figure l, the nucleotide sequence of Figure 3, the nucleotide sequence of Figure

10 4, the nucleotide sequence of Figure 5, or complementary

. to a portion of one of these sequences, wherein the DNA molecule is at least about 18 nucleotides in length; and

(d) detecting a pattern of hybridization of the DNA fragments and the DNA molecule, wherein the pattern is

15 diagnostic for the presence or absence of a genetic defect in an HDL-binding protein.

These and other aspects of the present invention will become evident upon reference to the following detailed description and attached drawings. •20

Brief Description of the Drawings

Figure 1 illustrates the nucleotide sequences of the coding strands of two cDNA clones encoding HDL-binding protein, together with the inferred amino

25 acid sequence of one of the clcfnes. Numbers above the lines refer to nucleotide positions. Numbers below the lines refer to amino acid positions.

Figure 2 illustrates the construction of the plasmid Zem217.

30 Figure 3 illustrates the nucleotide sequence of a cloned cDNA encoding a portion of a human HDL- binding protein, together with the deduced amino acid sequence. Certain restriction enzyme sites are shown. The terminal Eco RI sites are the result of the cloning

35 procedure. Numbers at the ends of lines refer to nucleotide position. Numbers below the lines refer to amino acid position. The arrow indicates the point of divergence in the sequences of clones 12B-1 and HEL 13.

Figure 4 illustrates the nucleotide sequence of a cloned DNA, designated HEL .13, encoding a human HDL-binding protein together with the deduced amino acid sequence. Certain restriction sites are shown. Numbers at the ends of the lines refer to nucleotide position. Numbers below the lines refer to amino acid position.

Figure 5 illustrates the nucleotide sequence of a cloned DNA, designated HEL 13/8 splice, encoding a human HDL-binding protein, together with the deduced amino acid sequence. Certain restriction sites are shown. Numbers at the ends of the lines refer to nucleotide position. Numbers below the lines refer to amino acid position.

Figure 6 shows partial restriction maps of human HDL-binding protein cDNA clones. Restriction sites are designated as: RI, Eco RI; P, Pst I; B, Bam HI; G, Bgl II; N, Nar I; A, Apa I; H, Hind III; V, Pvu II; S, Sst I; X, Xma I. The terminal Eco RI sites are the result of the cloning procedure.

Figure 7 illustrates the construction of the vectors Zem229R and ZMB4. Symbols used are: DHFR, mouse dihydrofolate reductase gene; SV40p, SV40 promoter; SV40t, SV40 terminator; MT-1, mouse metallothionein-1 promoter; MLP, adenovirus 2 major late promoter; and SS, splicing signals.

Best Mode for Carrying Out the Invention Prior to setting forth the invention, it may be helpful to an understanding thereof to set forth definitions of certain terms to be used herein.

Complementary DNA or cDNA: A DNA molecule or sequence which has been enzymatically synthesized from the sequences present in an mRNA template, or a clone of such a molecule. Expression Vector: A DNA molecule or a clone of such a molecule which has been modified through human intervention to contain segments of DNA which are combined and juxtaposed in a manner which would not otherwise exist in nature and which contains at least one DNA sequence to be expressed in a host cell and sequences which facilitate such expression, including a transcription promoter and terminator. Replication of an expression vector in a host cell commonly occurs autonomously, due to the presence of a replication origin on the vector, or may occur by integration into the host chromosome.

The present invention provides HDL-binding proteins and methods of making these proteins, as well as useful polypeptide fragments of these proteins. The- HDL-binding proteins and polypeptides are isolated from cells which naturally produce them, or, preferably, are made by expressing cloned DNA sequences in recombinant cells. The invention provides these proteins in substantially pure form, that is, at least 90% pure, and in one embodiment, free of other proteins of mammalian origin. The HDL-binding proteins and polypeptides thus produced are useful in test systems for screening poten¬ tial HDL analogs, including agonists and antagonists of HDL. In another embodiment of the present invention, the binding proteins and polypeptide fragments thereof are used as antigens to generate antibodies which may be used as HDL analogs. HDL analogs, including antibodies, other proteins, peptides and nonprotein compounds, may be used therapeutically in the treatment of hypercholes- terolemia and atherosclerosis. HDL analogs may also be used to screen patients for defects or deficiencies in HDL-binding proteins and the HDL receptor pathway.

The present invention also provides DNA sequences encoding HDL-binding proteins. These DNA sequences are useful in producing the HDL-binding proteins as noted above, and are also useful in the diagnosis of genetic defects in HDL-binding protein production. Such defects, which may be associated with a high risk for the development of atherosclerosis, may be detected by using the DNA sequences as probes for the presence of restriction fragment length polymorphisms (RFLPs) . These polymorphisms are the result of changes in DNA sequences and can be detected as differences in the sizes of fragments generated after digestion of DNA with restriction endonucleases. Thus, a cDNA encoding an HDL-binding protein or a fragment of such a cDNA can be used to probe for RFLPs in digests of genomic DNA isolated from, for example, blood leukocytes. Gener¬ ally, probes will be at least about 14, preferably at least 18, nucleotides in length. Human subjects with various forms of hyper- and hypo-lipidemia are screened to identify particular RFLPs that are diagnostic for disorders in lipoprotein metabolism and predisposition to atherosclerosis. RFLPs show Mendelian inheritance patterns and can be used to study the segregation of genes in family studies. Methods for carrying out RFLP analyses are generally known in the art and are described, for example, by Rees et al. (J . Clin. Invest. 2_.5:1090-95, 1985). Alternatively, the DNA sequences provided by the present invention can be used as specific probes or primers in other conventional screening systems.

Briefly, HDL-binding proteins are isolated from mammalian cells by solubilization of membranes followed by a combination of conventional column and slab gel chromatography techniques, including affinity chromatography using HDL apoproteins covalently bound to a suitable insoluble matrix (e.g., Sepharose, Pharmacia, Piscataway, N.J.). The HDL-binding proteins may also be purified by immunoaffinity chromatography using anti- bodies or antisera prepared as described hereinafter. Monoclonal antibodies are particularly preferred. Preferred sources of HDL-binding proteins include aortic endothelial cells and fibroblasts. Cells are preferably loaded with cholesterol by incubation in cholesterol- or LDL-containing medium prior to protein isolation.

Alternatively, the HDL-binding proteins may be produced by expressing cloned DNA sequences in recombi- nant cells. It is preferred to use cloned cDNA, because the use of intron-containing genomic DNA can result in aberrant expression, particularly in microorganism host cells. Suitable DNA sequences may also be synthesized according to conventional methods (see, for example, Caruthers et al., U.S. Patent No. 4,458,066; Itakura, in Trends in Biochemical Science. Elsevier Biochemical Press, 1982) based on the amino acid sequence of an isolated HDL-binding protein. Two such amino acid sequences are disclosed herein.

Immunogenic polypeptide fragments may be prepared from HDL-binding proteins by digestion with proteolytic enzymes or other agents, (e.g., cyanogen bromide) . Alternatively, peptides may be produced, by expression of cloned DNA sequences or by standard chemical synthesis techniques.

For expression in recombinant cells, a DNA sequence encoding an HDL-binding protein is inserted into a suitable expression vector, which in turn is inserted into appropriate host cells. The method of insertion will depend upon the particular host cell chosen. Methods for transfecting mammalian cells and transforming bacteria and fungi with cloned DNA are well known in the art. Suitable expression vectors will comprise a promoter which is capable of directing the transcription of a cloned DNA sequence in a host cell and a functional transcription termination site. In some instances, it is preferred that expression vectors further comprise an origin of replication, sequences which^" regulate and/or enhance expression levels, and one or more selectable markers, depending on the host cell selected. Suitable expression vectors may be derived from plasmids, RNA and DNA viruses, or cellular DNA sequences, or may contain elements of each.

Preferred prokaryotic hosts for use in carrying out the present invention are strains of the bacteria Escherichia coli. although Bacillus and other genera are also useful. Techniques for transforming these hosts and for expressing foreign DNA sequences cloned in them are well known in the art (see, for example, Maniatis et al. (eds.). Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1982). Vectors used for expressing foreign DNA in bacterial hosts will generally contain a selectable marker, such as a gene for antibiotic resis¬ tance, and a promoter which functions in the host cells. Appropriate- promoters include the trp (Nichols and Yanofsky, Meth. in Enzvmology 101:155- 1983), lac (Casadaban et al., J. Bact. 143:971-80- 1980), TAC (Russell et al.. Gene _2j0:231-43,1982) , and phage λ promoters. Plasmids useful for transforming bacteria include pBR322 (Bolivar et al., Gene .2:95-113, 1977), the pUC plasmids (Messing, Meth. in Enzvmology 101:20- 77, 1983; and Vieira and Messing, Gene _19_:259-68, 1982), pCQV2 (Queen, J. Mol. Ap l. Genet. 1-10, 1983), and derivatives thereof. Eukaryotic microorganisms, such as the yeast

Saccharo yces cerevisiae or filamentous fungi, including Aspergillus, may also be used as host cells. Particu¬ larly preferred species of Aspergillus include A. nidulans. A. niger, A_. orvzae. and A___ terreus. Techniques for transforming yeast are described by Beggs

(Nature 275:104-08. 1978) and Hinnen et al. fProc. Natl.

Acad. Sci. USA .31:1740-47, 1984). Expression vectors for use in yeast include YRp7 (Struhl et al., Proc.

Natl. Acad. Sci. USA 26:1035-39, 1979), YEpl3 (Broach et al.. Gene 8.:121-33, 1979), pJDB248 and pJDB219 (Beggs, ibid.), and derivatives thereof. Such vectors will generally comprise a selectable marker, such as the defective selectable marker POT1 (Kawasaki and Bell, EP 171,142; Murray et al., U.S. Patent No.4,766,073) , or a nutritional marker, such as the LEU2 gene. Preferred promoters for use in yeast expression vectors include promoters from yeast glycolytic genes (Hitzeman et al., J. Biol. Chem. .255:12073-80, 1980; Alber and Kawasaki, J. Mol. Appl. Genet. 1 :419-34, 1982; Kawasaki, U.S. Patent No. 4,599,311) or alcohol dehydrogenase genes (Young et al., in Hollaender et al. (eds.). Engineering of Microorganisms for Chemicals. New York:Plenum, 1982, p. 335; and Ammerer, Meth. in Enzvmology 101:192-201. 1983) . To facilitate purification of a foreign protein produced in a yeast transformant and to obtain, when necessary, proper disulfide bond formation, a signal sequence from a yeast gene encoding a secreted protein is joined in the proper reading frame, to the sequence encoding the foreign protein. Particularly preferred signal sequences include those encoding the pre-pro region of the MFαl gene (Kurjan and Herskowitz, Cell 10:933-43, 1982; Singh, EP 123,544; and Kurjan et al., U.S. Patent No. 4,546,082) and the secretory peptide portion of the BAR1 gene (MacKay et al., U.S. Patent No. 4,613,572; MacKay, EP 220,689; MacKay et al., U.S. Patent Application 07/270,933). Transformed cells are cultured in appropriate media containing carbon and nitrogen sources, as well as other nutrients or selective agents which may be required by the particular strain. Yeast host cells transformed with plasmids containing the PO 1 selectable marker may be cultured in complex media containing glucose as a carbon source.

Higher eukaryotic cells may also serve as host cells in carrying out the present invention. Cultured mammalian cells, such as the BHK, CHO, NS-1, SP2/0, HL60 and J558L cell lines, are preferred. These and other suitable cell lines are available from a variety of sources, such as the American Type Culture Collection, Rockville, Md. The tk" BHK cell line tk~tsl3 ( aechter and Baserga, Proc. Natl. Acad. Sci. USA 79:1106-10- 1982) is a particularly preferred adherent cell line. Expression vectors for use in mammalian cells will comprise a promoter capable of directing the transcrip¬ tion of a cloned gene or cDNA. Particularly preferred promoters include the SV40 promoter (Subra ani et al., Mol. Cell Biol. 1:854-64, 1981), the metallothionein gene (MT-1) promoter (Palmiter et al.. Science 222:809- 14, 1983; Palmiter et al., U.S. Patent No. 4,579,821), the mouse kappa gene promoter (Bergman et al., Proc. Natl. Acad. Sci. USA _8L:7041-45, 1984), and the adenovirus major late promoter (Berkner and Sharp, Nuc. Acids Res. 11:841-57, 1985) . Also contained in the expression vectors is a transcription terminator, located downstream of the insertion site for the DNA sequence to be expressed. A preferred terminator is the human growth hormone (hGH) gene terminator (DeNoto et al., Nuc. Acids Res. .9:3719-30, 1981). In addition, vectors will preferably contain enhancer sequences appropriate to the particular host cell line.

For expression in cultured mammalian cells, expression vectors containing cloned DNA sequences are introduced into the cells by appropriate transfection techniques, such as calcium phosphate-mediated transfec¬ tion (for example, Graham and Van der Eb, Virology 5_2:456-67, 1973, as modified by Wigler et al., Proc. Natl. Acad. Sci. USA 72:3567-70, 1980) or electropora- tion (Neumann et al. EMBO J. 1:841-45, 1982). In a typical transfection, a DNA-calcium phosphate precipi¬ tate is formed and then applied to the cells in the presence of medium containing chloroquine (100 βl ) . The cells are incubated for four hours with the precipitate, followed by a two-minute, 15% glycerol shock. A portion of the cells take up the DNA and maintain it inside the cell for several days. A small fraction of the cells integrate the DNA into the genome or maintain the DNA in nonchromosomal nuclear structures. These transfectants can be identified by cotransfection with a gene that confers a selectable phenotype (a selectable marker) . Preferred selectable markers include the DHFR gene, which imparts cellular resistance to methotrexate (MTX) , an inhibitor of nucleotide synthesis; or the bacterial neomycin resistance gene, which confers resistance to the drug G-418, an inhibitor of protein synthesis. After the host cells have taken up the DNA, drug selec- tion is applied to select for a population of cells that are expressing the selectable marker at levels high enough to confer resistance.

Transfected mammalian cells are cultured in serum-containing or serum-free media containing appro- priate nutrient and antibiotic supplements. Suitable media are available from commercial suppliers or may be prepared according to published recipes (see, e.g., catalogs of the American Type Culture Collection) .

Coamplification can be used to increase expression levels in cells transfected with the DHFR marker. High concentrations of MTX are^" added to the culture medium at the time of initial selection, or may be added by sequentially increasing the concentration of MTX in the medium, followed by repeated cloning by dilu- tion of the drug-resistant cell lines after each ampli¬ fication step. Cells which express the DHFR marker are then selected and screened for production of the HDL- binding protein.

Screening of transfected cells may be done by the ligand blotting technique (Graham and Oram, J. Biol. Chem. 262:7439-42. 1987). Briefly, cell fractions or culture media are separated by electrophoresis on polyacrylamide slab gels. The separated proteins are then transferred to nitrocellulose by electrophoresis. The nitrocellulose blots are blocked to prevent nonspecific binding, then incubated with the labeled ligand of interest (e.g., HDL₃) . Alternatively, cells may be screened by a functional assay that measures efflux of labeled cholesterol from cells in the presence of HDL. Briefly, transfected cells are incubated in sterol-rich medium to load cells with cholesterol, and intracellular pools of excess cholesterol are radio- labeled with a synthetic precursor (e.g., ³H~mevalono- lactone) . HDL-mediated translocation of intracellular ³H-cholesterol to the cell surface and into the medium is indicative of HDL receptor activity (Slotte et al., J. Biol. Chem. 262:12904-907. 1987).

As noted above, antibodies to HDL-binding proteins may be used as HDL agonists and antagonists and for the identification and purification of HDL-binding proteins. Monoclonal antibodies are generally preferred for these purposes. HDL-binding proteins or fragments thereof are used to make monoclonal antibodies according to methods generally described in the literature. Protein fragments useful in immunization will be of sufficient size to be immunogenic, preferably at least six amino acids in length, more preferably about fifteen amino acids in length. Preferred subject animals are mice or rats, with BALB/c mice being particularly preferred. The appropriate animals are immunized with a preparation of the HDL-binding protein, preferably a pure or partially pure preparation. Preferably, the animals are immunized with at least 100 ng each of the protein preparation, most preferably greater than 500 ng each. For immunization, the protein is preferably adsorbed to a solid-phase matrix, preferably to nitro- cellulose paper. The matrix is then introduced into the animal. Techniques for introduction of the adsorbed protein include implantation (U.S. Patent No. 4,689,220) or solubilization of the solid phase and injection of the solubilized material (Knudsen, Anal. ^• Biochem. 142:285-88, 1985). The solid-phase matrix may be solu¬ bilized in an appropriate organic solvent (e.g., DMSO) and either mixed with adjuvant or saline, or injected directly. Alternatively, the protein may be injected in the absence of a solid matrix and/or adjuvant. Injec¬ tion or implantation may be intraperitoneal, subcuta¬ neous, intramuscular or intravenous, preferably intraperitoneal. The animals may also be injected with adjuvant, such as Freund's adjuvant. Single or multiple booster immunizations are used. Between one and seven days prior to the fusion date, preferably on days one through four, intravenous injections of the HDL-binding protein may be given daily. When using small peptides as immunogens, it is preferred to conjugate the peptides to a hapten, such as keyhole limpet hemocyanin. Methods for conjugation are well known in the art.

Between one and seven days, preferably four days, after the administration of the final booster immunization, spleens or portions thereof are harvested from the immunized animals. At this time, the lymph nodes may also be harvested and included in the cell preparation. The harvested organs are minced using techniques which disrupt the structure of the organ but which are not detrimental to the lymphocytes. The organs are preferably minced with scissors, pressed through a mesh screen, and mixed with growth medium to enrich the preparation for lymphocytes. The minced and strained tissue is harvested by centrifugation, then mixed with growth medium to form a cell suspension. The red blood cells may be lysed by adding a hypotonic or hypertonic solution to the cell suspension. A preferred method for cell lysis is to add distilled water to the suspensions and quickly return the suspensions to an isotonic state with sodium chloride. Any remaining tissue may be removed by filtration through gauze.

The harvested cell suspension is then mixed with a myeloma cell line, preferably one which is syngeneic with the immunized animal. Myeloma cell lines from various species are widely available through, for example, American Type Culture Collection, Rockville, Md. Myeloma cell lines commonly used include PcS63Ag8 (ATCC TIB 9), SP2/0-Agl4 (ATCC CRL 1581), FO (ATCC CRL 1646), and 210-RCY-Agl (Galfre et al.. Nature 277:131. 1979) . A preferred cell line is P3/NSl/l-Ag4-l (ATCC TIB 18), hereinafter referred to as NS-1. The NS-1 cells are preferably tested to determine the cloning efficiency of the strain. This may be done by cloning out the NS-1 strain by limiting dilution and carrying out test fusions with the individual NS-1 clones to find candidates with the highest fusion efficiencies. The myeloma cells are cultured in an appropriate mammalian cell growth medium, a variety of which are generally known in the art and available from commercial sources. Mammalian cell lines are routinely grown between 36"C and 40^βC under conditions which maintain an optimal pH between 6.0 and 8.0, preferably about pH 7.2. The pH of the culture may be maintained through the use of a variety of buffer systems known in the art. A preferred buffer system involves growing the cells in a bicarbon- ate buffer in a humidified incubator containing CO,, preferably about 7% C0₂-

The fusion between the lymphocytes from the immunized animal and the myeloma cells may be carried out by a variety of methods described in the literature. These methods include the use of polyethylene glycol (PEG) (Brown et al., J. Biol. Chem. 255:4980-83 _r 1980) and electrofusion (Zimmerman and Vienken, J♦ Membrane Biol. .62:165-82, 1982; an electrofusion generator is commercially available from Biotechnologies and Experimental Research, Inc., San Diego, Calif.).

Following the fusion, the cells are plated into multi-well culture plates, preferably 96-well plates. A reagent which selectively allows for the growth of the fused myeloma cells over the unfused cells is added to the culture medium. A preferred selection technique uses HAT (hypoxanthine, aminopterin. thymidine) selection. Other selection techniques may also be used, depending on the myeloma cell line used.

Alternative methods of producing monoclonal antibodies utilize in vitro immunization techniques. Lymphocytes may be harvested from lymphoid organs, such as spleens or lymph nodes, or from whole blood as peripheral blood lymphocytes. The lymphocytes are put into culture in the presence of the HDL-binding protein. Often immunostimulatory polypeptides will be added to the culture medium concurrently. At various times following the culturing of the lymphocytes in vitro, the lymphocytes are harvested and fused with the myeloma cell line as described above.

Other techniques for producing and maintaining antibody-secreting lymphocyte cell lines in culture include viral transfection of the lymphocytes to produce transformed cell lines which will continue to grow in culture. Epstein-Barr virus (EBV) has been used for this technique. EBV-transformed cells do not require fusion with myeloma cells to allow continued growth in culture.

Thymocytes may be used as a feeder layer to condition the medium for the fused cells. Thymocytes may be prepared from juvenile mice less than eight weeks old. The thymus glands are harvested, and cell suspen¬ sions are prepared as described above. Alternatively, peritoneal macrophages or nonimmune spleen cells may be used as a feeder layer. Another alternative is to use conditioned medium from thymocytes or macrophages. At an appropriate time following the fusion day, the fused cells (hybridomas) are then analyzed for the production of antibody against the antigen of choice. This can be done by a wide variety of tech¬ niques, including Western blot, ELISA, immunoprecipita- tion, influence on biological activity assays, and immunocytochemical staining. These techniques and others are well described in the literature. (See, for example, J. G. R. Hurrell (ed.). Monoclonal Hybridoma Antibodies: Techniques and Applications. Boca Raton, Fla.:CRC Press Inc., 1982.) Hybridomas which secrete antibodies of interest are maintained in culture. The cells are expanded in culture and, at the same time, may be cloned in such a manner as to obtain colonies originating from single cells. This provides for the onoclonality of the antibodies obtained from the hybridomas. A wide variety of techniques exist for cloning cells, including limiting dilution, soft agar cloning, and fluorescence-activated cell sorting.

Once clones of cells are obtained, they are reassayed for the production of the antibody of interest. These cells are then expanded in culture to allow for the production of larger amounts of the antibody. Methods for expansion of the cells include maintaining the cells in culture, placement of the cells in a bioreactor or other type of large-scale cell culture environment,, or culturing the cells using various agar or gelatin carrier matrices. Antibodies are then isolated from the cell culture media.

A preferred method for producing large amounts of antibodies involves growing the hybridoma cells in ascites fluid. The hybridomas are preferably isolated from the culture media by centrifugation and washed with an iso-osmotic solution, preferably phosphate-buffered saline. The cells are then resuspended in an iso- osmotic solution and injected into the peritoneal cavity of an appropriate host animal, preferably a mouse, and allowed to grow within the host animal. The host animal may receive a preinjection of pristane (2,6,10,14- tetramethylpentadecane) prior to the injection of the hybridoma cells, preferably seven to thirty days prior to the introduction of the hybridomas. Following growth of the cells in the peritoneal cavity, ascites fluid, containing the antibody of interest, is collected. Antibodies may be purified from conditioned media or ascites fluid by a variety of methods known in the art. These methods include ammonium sulfate precip¬ itation, ion-exchange chromatography (see Hurrell, ibid.), and high-pressure liquid chromatography using a hydroxylapatite support (Stanker et al., J. Immunol. Methods 26 15 , 1985). A preferred method for purifying antibodies from conditioned media or ascites fluid utilizes a commercially available Protein A-Sepharose CL-4B column (Pharmacia, Piscataway, N.J^*. ; Sigma, St. Louis, Mo.). Antibodies may be purified with these columns using conditions suggested by the manufacturer. Typically, the conditioned medium or ascites fluid is mixed with an equal volume of TNEN (20 mM Tris-base, pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5% NP-40) and applied to the column. The antibodies are eluted using a pH gradient. Preferably, the elution buffer comprises 0.1 M sodium citrate.

Antibodies may also be produced by reσombinant methods. Antibody gene sequences of interest are isolated from antibody-producing cells and are cloned and expressed according to known procedures. Recombi- nant methods facilitate the production of human mono¬ clonal antibodies and chimeric antibodies containing human constant regions. Methods for producing recombi- nant antibodies are disclosed by, for example, Neuberger et al. (WO 86/01533) and Winter (published European patent application EP 239,400).

Antibodies and other potential HDL agonists are screened in a multistep assay employing the isolated binding protein and/or recombinant cells which express the protein. Potential agonists are first screened for their ability to bind to the HDL-binding protein by the ligand blotting technique. Those compounds which are found to bind are then screened for their ability to stimulate transport of cholesterol from cells essen¬ tially as described above. Briefly, cellular choles- terol pools are labeled and the cells are incubated in the presence of the compound to be tested. Total lipids are then extracted from the medium, separated by thin- layer chromatography, and assayed for the presence of labeled cholesterol.

Proteins, including antibodies, that bind to HDL-binding proteins are useful within diagnostic compo¬ sitions for identifying defects in the synthesis and processing of HDL-binding proteins in a patient and for detecting structural defects in HDL-binding proteins. Given the role that HDL is believed to play in choles¬ terol transport and the prevention of atherosclerosis, diagnosis of such defects may be used to identify individuals at risk for atherosclerosis or to diagnose the particular condition responsible for atherogenesis.

To screen for genetic defects in HDL-binding proteins, cells are obtained from a patient and analyzed for the presence and structure of the proteins. Preferred cells in this regard include fibroblasts and leukocytes. Fibroblasts may be obtained from skin samples and are readily cultured in vivo . Leukocytes may be obtained from blood samples by standard procedures. Analysis of cell membrane components by Western blotting (Towbin et al., Proc. Natl. Acad. Sci. USA 76:4350-54■ 1979) with antibodies to HDL-binding proteins is used to detect structural defects or deficiencies. The use of a panel of antibodies that recognize different epitopes on a protein is particularly advantageous for detecting structural defects. Genetic defects may also be detected by probing cellular DNA from a patient with cloned DNA sequences encoding HDL-binding proteins or fragments thereof and analyzing the hybridization pattern, which is indicative of the presence or absence of a mutation. HDL agonists may be used to stimulate clear¬ ance of cholesterol in patients with low levels of HDL or defects in apoprotein structure. Peptide agonists will generally be combined with a physiologically acceptable carrier or diluent and administered by injec¬ tion, although small peptides may also be administered by ingestion or inhalation using conventional formula- tions. Chemical agonists will preferably be adminis¬ tered by ingestion.

Experimental Example 1 - Identification and Characterization of a Representative HDL-Binding Protein

LDL and HDL₃ were isolated from the plasma of healthy fasting subjects by sequential centrifugation, as described by deLalla and Gofman (Methods Biochem. Anal. 1:459-78, 1954). The HDL₃ fraction was subjected to heparin-Sepharose affinity chromatography to remove apoE-containing particles (Biesbroeck et al. , J. Clin.

Invest. 21:525-39 1983). Lipoproteins were iodinated by a modification of the monochloride procedure (Bilhei er et al., Biochem. Biophys. Acta 260:212-21- 1972). Proteoliposomes were prepared using egg lecithin and purified apoprotein A-I, A-II, A-IV or E by the cholate dialysis method (Chen and Albers, J. Li id Res. 23:680- 91, 1982) at a molar ratio of lecithin to apoprotein of 300:1. Cultured human skin fibroblasts, bovine aortic endothelial cells, Swiss 3T3 fibroblasts, and transformed macrophage-like cell lines (e.g., mouse J774 macrophages and human HEL cells) were plated into 150 mm culture dishes in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and were grown to confluency. Cells were loaded with cholesterol by incubating them for 24 to 48 hours with serum-free medium containing 0.2% bovine serum albumin plus either nonlipoprotein cholesterol added in ethanol or acetylated LDL (Oram, Meth. in Enzvmology 129B:645-59_f 1986) . Cell membranes were prepared from cholesterol- treated cells essentially as described by Basu et al. (J. Biol. Chem. 253:3852-56, 1978). Cell monolayers were washed three times with PBS plus 0.2% BSA at room temperature, and then chilled to 4^*C and scraped with a nylon policeman into 10 mM Tris-HCl, pH 7.4, containing 150 mM NaCl, 1 mM benzamidine, 0.5 mM EDTA, and 1 mM phenyl ethyl-sulfonyl fluoride dissolved in dimethyl- sulfoxide (TBS) . Cells were centrifuged at 200 x g for 5 minutes at 4^*C, resuspended in TBS, pooled, homogenized with two 5-second pulses of a Polytron homogenizer (Tekmar Co., Cincinnati, Ohio), and centrifuged at 800 x g for 10 minutes at 4'C. The 800 x g supernatant was centrifuged at 100,000 x g for 60 minutes at 4^*C, and the membrane pellet was stored at -70^*C. In a typical preparation, the total membrane protein yield, as quantified by the method of Lowry et al. (J . Biol. Chem. 191:265-75, 1951), from 20 150-mm dishes was approximately 2 mg. Membrane pellets were solubilized by needle aspiration in electrophoresis sample buffer and boiled for 3 minutes in the presence of 10% (v/v) _/3-mercap- toethanol. Fractions containing equal amounts of protein (generally 500 μg) per lane were electrophoresed on 7% polyacrylamide slab gels according to Laemmli (Nature 227:680-85, 1970). Separated proteins were transferred onto nitrocellulose (Schleicher and Schuell, 0.45 /zm) by electrophoresis (Towbin et al., Proc. Natl. Acad. Sci. USA 26:4350-54, 1979). Gels were calibrated using prestained or iodinated molecular weight standards.

To assay protein binding activity, nitrocellu¬ lose strips were incubated for 2 hours with blocking buffer (10 mM Tris-HCl, pH 7.4, containing 150 mM NaCl, 1% w/v BSA, 1% w/v Carnation nonfat dried milk, 0.01% v/v antifoam A, 50 μg/ml LDL) , followed by a further 2- hour incubation in the same buffer containing 5 g/ml ¹²⁵ι-HDL₃. Strips were washed once rapidly and then five times for 15 minutes with washing buffer (blocking buffer without LDL) . All blotting incubations and washes were at room temperature. After drying, protein bands were visualized by autoradiography. Autoradio- graphs revealed that the ¹²⁵i-HDL₃ bound to a major membrane protein having an apparent molecular mass of approximately 110,000 Daltons in all the aforementioned cell types. The relative migration and binding activity of this protein were not altered by boiling of the solubilized membranes or by pretreat ent with 10% (v/v) ,5-mercaptoethanol. As compared to cholesterol-loaded cells, cells incubated with cholesterol-free medium had lower or undetectable amounts of the 110 kDa protein, suggesting that its binding activity was up-regulated by cholesterol loading of cells. Preincubation of human arterial smooth muscle cells with increasing concentra¬ tions of cholesterol produced a dose-dependent increase in the binding activity of the 110 kDa protein. Trypsin treatment of the endothelial cell monolayer abolished the activity of the HDL-binding protein, indicating that it is located in the plasma membrane and is therefore available at the cell surface for interaction with HDL. Under these conditions, the cell monolayer remained intact; and cell morphology, as assessed by light microscopy, was unaffected.

Treatment of the 110^' kDa HDL-binding protein with endoglycosidase^" H followed by gel electrophoresis of the treated protein indicated a carbohydrate content of approximately 10% by weight.

Example 2 - Purification of HDL-Binding Protein

The HDL-binding protein described in Example 1 is prepared from microsomal membranes prepared from cultured bovine endothelial cells by solubilization in

1% Triton X-100. Solubilized proteins are passed through a wheat germ agluttinin column. The adsorbed glycoprotein is eluted in the presence of excess N-acetylglucosamine, dialyzed, and subjected to isoelec- tric focusing electrophoresis in the presence of 4 M urea. Proteins having the isoelectric point of the HDL- binding protein (pH 5.8 to pH 6»0) are isolated and purified by SDS-polyacrylamide gel electrophoresis. The position of the HDL-binding protein on the gel is identified by silver staining and confirmed by the ligand blot assay. The HDL-binding protein is then electroeluted from the gel. Protein purified by this method is at least 90% pure, as determined by silver staining of polyacrylamide gels. The protein may be further purified by HPLC as necessary. Alternatively, the HDL-binding protein may be purified by immuno- affinity chromatography of solubilized membranes using monoclonal antibodies produced as described herein. The purified protein is used for antibody preparation and protein characterization.

Example 3 - Cloning of cDNA and Genomic Seguences

Encoding HDL-Binding Proteins A. Preparation of a Lambda gtll Library from Human Fibroblast Cell RNA Total RNA was prepared from human fibroblast

(SK-5) cells cultured in the presence of 10% serum by the guanidine isothiocyanate method (Chirgwin et al., Biochemistry 18:52-94, 1979). Poly (A)⁺ RNA was obtained by two rounds of affinity chromatography on oligo (dT) cellulose (Maniatis et al., ibid.). First and second strand DNA was synthesized by either oligo (dT) or random priming for first strand synthesis (Frischer et al., Cell 42:1017-28, 1986) from 5 ug of 2 X selected poly (A)⁺ RNA, as described by Gubler and Hoffman (Gene 25:263-69, 1983) and Hagen et al. (Proc. Natl. Acad. Sci USA .33:2412-16, 1986). The double- stranded cDNA was methylated with Eco RI methylase (Maniatis et al., ibid.), passed over a Sepharose CL-6B (Pharmacia, Piscataway, N.J.) column, blunted with T₄ DNA polymerase (Maniatis et al., ibid.), ligated to Eco^"RI linkers, digested with Eco RI, and passed over a Sepharose CL-2B (Pharmacia) column. The cDNA was ligated to an Eco Rl-digested, phosphatased lambda gtll vector (Young and Davis, Proc. Natl. Acad. Sci USA £0:1194-98, 1983) and packaged in lambda packaging extract (Stratagene, Inc., La Jolla, Calif.). Plate lysate libraries were prepared (Maniatis et al. , ibid.) using Ej_ coli Y1088 as the host strain. These plate lysate libraries were stored over chloroform at 4^*C until used for ligand screening. The random-primed cDNA library was composed of approximately 11 million isolates, and the oligo (dT)-primed library was composed of approximately 13 million isolates.

B. cDNA Cloning by Ligand Blotting

Thirty plates (150 mm) of NZY agar were each overlayed with 10 ml of NZY 0.7% agarose containing E. coli Y1O90 cells adsorbed with approximately 50,000 λgtll that contained oligo (dT)-primed and random-primed human fibroblast cDNA inserted at the Eco RI site. The plates were incubated at 37^*C for 3 hours, overlayed with 137-mm nitrocellulose filters (Schleicher and Schuell) previously soaked in 10 mM IPTG (isopropyl -β-D- thiogalactopyranoside, Sigma^' Chemical Co., St. Louis, Mo.), and incubated at 37^*C overnight. The filters were washed twice for 15 minutes each at room temperature in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl, and washed for 2 hours at room temperature in ligand-blotting blocking buffer. The filters were transferred to fresh blocking buffer containing 5 μg/ml ¹²⁵I-HDL₃ and incubated for 2 hours at room temperature. The filters were washed three times for 15 minutes each at 10^*C in washing buffer, two times for 15 minutes each at 10*C in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl, 0.2% BSA, and once for 5 minutes at 10^βC in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl. The filters were air dried and exposed to X-ray film at room temperature for 10 days. Positive regions on the plates were picked, titered, and replated to give 10,000 and 50,000 plaques each on IL. coli Y1090 cells according to standard procedures (Maniatis et al., ibid.). IPTG- soaked filters were overlayed and screened as described above. Positive plaques were picked and replated to give 500 plaques each on I__ coli Y1090 cells. IPTG- soaked filters were overlayed and screened as described above. Plate lysate stocks were prepared as described by Maniatis et al. (ibid.) . Phage DNA was prepared by incubating 5 OD₆₀₀ units of EL. coli Y1090 cells infected with 5 x IO⁶ phage in 500 ml of NZY broth at 37^*C for 5 hours. The phage were recovered by precipitation with 1 M NaCl and 10% (w/v) polyethylene glycol 8000 at 4°C and centrifugation at 5000 rpm in a Sorvall GS3 rotor for 15 minutes. The phage precipitates were each dissolved in 5 ml of =1.50 gm/ml CsCl in 10 mM Tris-HCl, pH 7.4, 100. mM NaCl, 10 mM MgS0₄, and centrifuged in a Sorvall SW55 rotor at 35,000 rpm overnight. The CsCl-banded phage were recovered by ethanol precipitation, and DNA was extracted by resuspending the λ pellet in 850 μl of 50 mM Tris-HCl, pH 8.0, adding 100 μl of 100 mM Tris-HCl, pH 8.0, 100 mM EDTA, 0.1% SDS and 50 μl of 1 mg/ml proteinase K followed by incubation for 1 hour at 37^*C. The DNA was extracted twice with phenol:CHC1₃:isoamyl alcohol (24:24:1) (PCI), once with CHCI₃:isoamyl alcohol (24:1) (CI) , and precipitated with ammonium acetate and ethanol. The purified cDNAs were resuspended in 75 μl of 10 mM Tris, pH 8.0, 1 mM EDTA, digested with Eco RI at 37*C, and the cDNA inserts were recovered by electrophoresis on 0.8% agarose gels. The cDNA fragment from a positive phage clone from the random-primed library (designated 17R1E3) was recovered on NA45 (Schleicher and Schuell, Inc., Keene, N.H.), eluted in 1.5 M NaCl, 10 mM Tris-HCl, pH 8.0, 1 mM EDTA at 65*C for 20 minutes, PCI and CI extracted and ethanol precip¬ itated. The cDNA fragment was inserted at the Eco RI sites of M13mpl8 and pUC19, and the DNA sequence was determined by the method of Sanger et al. (Proc. Natl. Acad. Sci. USA 74.:5463-67, 1977). Results indicated that this clone contained nucleotides 257-828 of the sequence shown in Figure 1. The pUC19 cDNA clone was designated pM230-l.

A 2 kb human fibroblast cDNA was isolated by screening the human fibroblast λgtll cDNA library with the partial cDNA from plasmid pM230-l. All bacterio- phage platings, plaque purifications, and DNA purifica¬ tions were performed as described above. The 2 kb cDNA was excised from λ clone 18B and inserted into the Eco RI site of pUC19 to construct plasmid pM253-3„ The cDNA was also inserted into M13mpl8 and M13mpl9, and the DNA sequence was determined by the method of Sanger et al. (ibid.) . The nucleotide sequence of the cloned 2 kb cDNA and the inferred amino acid sequence are shown in Figure 1. While the amino acid sequence in Figure 1 is shown with an amino-terminal methionine residue, cellular processing may result in a protein having an amino-terminal glycine.

C. Construction of a Human Cosmid Library

The vector Zem217 was prepared from pJB8 (Lau and Kan, Proc. Natl. Acad. Sci. USA .30:5225-29, 1983) and pSV2-neo (Southern and Berg, J. Mol. APPI. Genet. 1:327-41, 1982) by combining the 3.4 kb Hind III fragment of pJB8, which contains the SV40 promoter and COS region, with the 2.65 kb Hind III-Eco RI fragment of pSV2-neo, which contains the neomycin resistance gene and SV40 terminator, into the Hind III and Eco RI sites of pUCl3 (see Figure 2) . High molecular weight DNA was isolated from the human acute lymphoblastic leukemia cell line MOLT-4 (Minowada et al., J. Natl. Cancer Inst. _49.:891-95, 1972; ATCC CRL 1582) . The DNA was partially digested with Sau 3A and size-fractionated on an NaCl gradient as described by Dillela and Woo (Focus 2 2, 1-4, 1985). DNA^" fractions containing 30 to 50 kb fragments were ligated into the Bam HI site of Zem217, and the ligation reaction was in vitro packaged into phage particles (using a packaging kit obtained from Stratagene, Inc. , La Jolla, Calif.) which were used to transfect E__s_ coli DHL The procedure gave rise to approximately 2.5 x IO⁶ ampicillin-resistant colonies which were harvested from plates, mixed together, aliquoted, and frozen at -70^*C. Staphylococcal growth was prevented by the addi¬ tion of vancomycin-HCl (Lilly) at approximately 3 mg/liter to ampicillin-containing media. The human cosmid library was screened with

³²P-labeled pM253-3 2-kb cDNA. Twenty plates (150 mm) of LB media containing 50 mg per liter of neomycin were inoculated with approximately 50,000 colonies each of the human cosmid library and incubated overnight at 37^*C. Nitrocellulose filter lifts of the plates were prepared and hybridized with the pM253-3 cDNA, and posi¬ tive colonies were purified according to standard methods (Maniatis et al., ibid.). The three positive clones obtained were designated pM280-16, pM280-17 and pM280-20.

D. cDNA Cloning

Polyclonal antisera were pooled from two mice which had been immunized with a partially purified preparation of the bovine 110,000 Dalton HDL-binding protein. The pooled sera were depleted of anti-E. coli antibodies by chromatography on CNBr-activated Sepharose

(Pharmacia, Piscataway, N.J.) conjugated to an extract of _ _ coli Y1090. Twenty plates (150 mm) of NZY agar were each overlayed with 10 ml of NZY 0.7% agarose containing _ \___ coli Y1090 cells adsorbed with approxi-

^• mately 50,000 λgtll phage containing random-primed human fibroblast cDNA inserted at the Eco RI site. The plates were incubated at 37^*C for 3 hours, then over¬ layed with 137 mm nitrocellulose filters, which had been previously soaked in 10 mM IPTG, and incubated at 37*C overnight. The filters were washed at room temperature in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl for 15 minutes; in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl, 3% BSA for 1 hour; and overnight in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl, 3% BSA containing the pooled, depleted antisera diluted approximately 1:500. The filters were then washed at room temperature in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl for 10 minutes; in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl, 0.1% NP-40 for 10 minutes; twice in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl for 10 minutes each; and in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl, 3% BSA for 10 minutes. The filters were then incubated for 2.5 hours in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl, 3% BSA with approximately 1,000,000 cpm per filter of ¹²⁵ι-iabeled affinity-purified . rabbit anti-mouse IgG _t (Organon Teknika-Capell, Malvern, Pa.). The filters were then washed at room temperature in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl for 10 minutes; in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl, 0.1% NP-40 for 10 minutes; and twice in 50 mM Tris-HCl, pH 8.5, 150 mM NaCl for 10 minutes each. The filters were air dried and autoradiographed at -70*C with an intensifying screen. Positive regions on the plates were plaque-purified as described above. The one positive clone obtained was designated 12B-1. Phage DNA was purified from clone 12B-1 as described above. Plaque 12B-1 was used to lysogenize E^. coli

Y1089. The infected bacteria were plated on LB- ampicillin media at 30^βC, then replica plated onto LB- ampicillin and incubated at 30^βC and 42^βC. Colonies that failed to grow at 42"C were lysogenized by induction with IPTG, and the 3-galactosidase fusion protein was purified by immuno-affinity chromatography using ProtoSorb lacZ adsorbent (obtained from Promega Biotec, Madison, Wis., and used as specified by the manufacturer) .

The purified fusion protein was dialyzed against 50 mM ammonium bicarbonate, lyophilized, resuspended in PBS and used to prepare polyclonal antisera. The protein (1.0 μq/μl) was adsorbed to two nitrocellulose discs using 10 μl per disc. The discs were dried at 37"C for 3-4 hours, then each disc was dissolved in 100 μl DMSO. The resulting solution was combined with Freund's adjuvant and injected intraperi- toneally into male Balb/c mice (obtained from Simonsen Lab, Inc., Gilroy, Calif.). The mice were immunized with an initial injection of 2 μg of protein per mouse, followed by seven booster injections of 2 or 4 μg each at 2- to 3-week intervals. To obtain antisera, blood was collected from the mice by retro-orbital bleeds and allowed to clot, and the serum was recovered.

For sequence analysis, phage DNA was recovered fro the 12B-1 recombinant and the cDNA was excised by Eco RI digestion. Eco RI fragments of 261 and 339 bp were inserted into the Eco RI site of pUC19, and the resultant plasmids were designated pM283-ll and pM283-10, respectively. The fragments were also inserted into M13mpl8 and sequenced by the method of Sanger et al. (ibid). The DNA sequence of the 12B-1 cDNA is shown in Figure 3. The 600 bp cDNA has an open reading frame that extends for its entire length, suggesting that it encodes an internal fragment of the HDL-binding protein. No significant matches were found to the cDNA sequence or the corresponding six potential translation products using the FASTA and TFASTA (Pearson and Littman, Proc. Natl. Acad. Sci. USA 85:2442-48, 1988) computer searches of the GenBank, EMBL and PIR databases. The individual Eco RI fragments from the 12B-1 phage clone were ³ P-labeled using random primers (Amersham Corp., Arlington Heights, 111.) and used to probe Northern blots (Thomas, Proc. Natl. Acad. Sci. USA 22:5201, 1980) of poly (A)⁺ RNA (Chirgwin et al.. Biochemistry 18:5294, 1979; Aviv & Leder, Proc. Natl. Acad. Sci. USA .69.:1408, 1972) from the mouse macrophage cell line J774, which had been cultured in the absence and presence of exogenous acetylated LDL (Ora et al., J. Biol. Chem. 262:2405-10. 1987). Identical RNA blot patterns were obtained, indicating that the cDNA encom¬ passes a natural Eco RI site. Both RNA preparations contained hybridizing mRNAs of 4.4 and ~6 kb, but these RNAs were approximately fourfold more abundant in the LDL-grown cells. Equal amounts of RNA were shown to be present in both samples by stripping the probe from the blot and rehybridizing the blot with a rat glyceralde- hyde-3-phosphate dehydrogenase cDNA probe. The 12B-1 fibroblast cDNA probe yielded .identical RNA blot patterns in human HepG2 liver carcinoma, bladder carcinoma, rhabdomyosarcoma, and normal mononuclear cells, indicating that the HDL-binding protein gene is expressed in many cell types.

One million recombinant Lambda gtll plaques containing oligo (dT)-primed human erythroleukemia (HEL) cell cDNA (Lopez et al. , Proc. Natl. Acad. Sci. USA _.34:5615-19, 1987) were screened for hybridization to the 12B-1 fibroblast cDNA. Nitrocellulose filters, each containing approximately 50,000 plaques, were hybridized with the ^{' 32}P-labeled cDNA in 50% formamide, 5x Denhardt's (1 mg/ml Ficoll, 1 mg/ml polyvinylpryroli- done, 1 mg/ml BSA), 5x SSPE (2Ox SSPE = 3.6 M NaCl, 200 mM NaH₂P0₄, 20 mM EDTA, pH 7.4), 0.1% SDS and 100 μg/ml denatured salmon sperm DNA (Maniatis et al. (eds.). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. , 1982, pp. 326-28). The filters were washed twice in 2x SSC, 0.1% SDS at room temperature for 30 minutes per wash; twice in 0.2x SSC, 0.1% SDS at 65^βC for 30 minutes per wash; in O.lx SSC, 0.1% SDS at 65^βC for 90 minutes; in O.lx SSC, 0.1% SDS, 50% forma ide at 65^*C for 90 minutes; and in O.lx SSC, 0.1% SDS at 65^*C for 2 hours. Two positive plaques, designated HEL 13 and HEL 8, were purified, phage DNA was obtained, and the cDNAs were excised by Eco RI digestion. Clone HEL 13 contained three Eco RI fragments, designated 13A, 13B and 13C, whose approxi¬ mate sizes were determined as 0.6, 0.8 and 3.6 kb, respectively. Each of the HEL 13 Eco RI fragments was inserted into the Eco RI site of pUC19, yielding plasmids pM288 (13A) , pM289 (13B) and pM290 (13C) . Clone HEL 8 contained two Eco RI fragments, designated 8A and 8B, whose approximate sizes were determined as 0.8 and 1.4 kb, respectively. The HEL 8 Eco RI fragments were each inserted into the Eco RI site of pUC19, yielding plasmids pM291 (8A) and pM292 (8B) . The intact HEL ..13 cDNA and flanking Lambda gtll DNA was excised by digestion with Kpn I and Sst I, and was inserted into pUC19 to construct plasmid pM293. DNA hybridization analysis demonstrated that the pM283-10 cDNA hybridized to the 13A and 8A fragments, whereas the pM283-ll cDNA hybridized to the 13C and 8B fragments. Restriction endonuclease maps were prepared for the cDNA fragments in plasmids pM288 to pM293. The maps indi¬ cated that the HEL 13 cDNA has a much longer 5' end and .differs from the HEL 8 and 12B-1 cDNAs at the 3' end by an apparent alternative splicing event. The comparative restriction maps are shown in Figure 6. Analysis of the HEL 13 DNA sequence showed the presence of fourteen internal repeats. Each repeat was shown to contain an internal consensus sequence bounded on either side by an amphipathic helix. A comparison between the HEL 8 and 12B-1 DNA sequences suggested that 12B-1 represents an internal fragment of HEL 8. Sequence analysis of the HEL 13 cDNA (Figure 4) also indicated the presence of a stop codon after nucleotide 381 (Figure 3) .

A comparison of the HEL 13 and HEL 8 DNA ^' sequences suggested that HEL 13 may be a variant of HEL 8. A computer-generated DNA sequence showing the HEL 13 sequence joined to the C-terminal DNA sequence of HEL 8, designated HEL 13/8 splice, is shown in Figure 5. A DNA sequence encoding the HEL 13/8 splice is generated as follows. Plasmid pM290 is digested with Eco RI and Aat II to isolate a 3.55 kb fragment comprising the 5' coding sequence from HEL 13. Plasmid pM305, comprising the cDNA fragment of HEL 8 subcloned as a partial Eco RI fragment into pUC19, is digested with Aat II and Bel I to isolate the approximately 0.4 kb fragment. Plasmid pM291 is digested with Bel I and Eco RI to isolate the approximately 0.39 kb fragment encoding the 3' coding region of HEL 8. The Eco RI - Aat II pM290 fragment, the Aat II - Bel I pM305 fragment and the Bel I - Eco RI pM291 fragment are joined with Eco Rl-linearized ZMB4 (Example 4) to construct a mammalian expression vector. The inserts are characterized for orientation with, respect to the promoter. A plasmid containing the HEL 13/8 splice DNA sequence in the correct orientation in the vector ZMB4 is designated HEL 13/8 splice-ZMB4.

Example 4 - Transfection of Mammalian Cells and Expression of HDL-Binding Protein The vector Zem229R was constructed as shown in Figure 7 from Zem229. Plasmid Zem229 is a pUC18-based expression vector containing a unique Bam HI site for insertion of cloned DNA between the mouse metallo- thionein-1 promoter and SV40 transcription terminator and an expression unit containing the SV40 early promoter, mouse dihydrofolate reductase gene, and SV40 terminator. Zem229 was modified to delete the two Eco RI sites by partial digestion with Eco RI, blunting with DNA polymerase I (Klenow fragment) and dNTPs, and re- ligation. Digestion of the resulting plasmid with Bam HI followed by ligation of the linearized plasmid with Bam HI-Eco RI adapters resulted in a unique Eco RI cloning site. The resultant plasmid was designated Zem229R.

The expression vector ZMB4 was constructed from Zem229R and pDX (Hagen et al., U.S. Patent No. 4,784,950) as shown in Figure 7. Zem229R was digested with Hind III and Eco RI, and the 520 bp fragment containing the SV40 and MT-1 promoters was removed. The large fragment of Zem229 was then joined to the ~1100 bp Hind III-Eco RI fragment of pDX, which contains the SV40 promoter/enhancer, the adenovirus major late promoter, and a set of splicing signals.

The HEL 13 cDNA was inserted into the Eco RI sites of the mammalian cell expression vectors ZMB4 and Zem229R by ligation of the following fragments: Eco RI - Bgl II from pM290-ll, Bgl II - Xma I from pM293-4, X a I - Eco RI from pM289-5 and Eco Rl-digested vector. The inserts were characterized for orientation with respect to the promoter. A plasmid containing the cDNA- in the sense orientation in ZMB4 was designated pM296-10. (Plasmid pM296-10 has been deposited with American Type

Culture Collection, Rockville, Md. , as an E___ coli HB101 transformant.) A plasmid containing the cDNA in the antisense orientation in ZMB4 was designated pM296-7. A plasmid containing the cDNA in the sense orientation in Zem229R was designated pM297-4, and a plasmid containing the cDNA in the antisense orientation was designated PM297-2 (Table 1) .

Baby hamster kidney (BHK) cells (tk~tsl3) in growth medium (Dulbecco's minimum essential medium [DMEM] containing 10% fetal bovine serum, 2 mM gluta ine, 1 mM sodium pyruvate, 50 μg/ l penicillin, 50 μg/ml streptomycin, and 10 μg/ml neomycin) were seeded into 10 cm² tissue culture dishes and grown overnight at 37^βC. Cells were transfected with plasmid DNA using the calcium phosphate precipitation method (Graham and Van der Eb, ibid). Three hundred μl of 10 mM Tris-HCl, 1 mM EDTA, pH 8.0, containing 0.5 M CaCl₂, 10 μg plasmid DNA and 12.5 μg salmon sperm DNA was added dropwise, with vortexing, to 300 μl of 2x HBS (280 mM NaCl, 50 mM Hepes, 1.5 mM Na₂HP0₄, pH 7.15). The mixture was left at room temperature for 30 minutes to allow precipitate formation.. Six hundred μl of precipitate was added to each 10 cm² dish of BHK cells, and the dishes were incubated overnight at 37^*C. The culture medium and precipitate were aspirated from the dishes, and 5 ml of 15% glycerol in DMEM was added for 3 minutes. The dishes were washed once with DMEM, and 10 ml of growth medium was added. After 24 hours the culture medium was changed to selection medium (DMEM containing 10% fetal bovine serum and 250 nM methotrexate) . Subsequently, the selection medium was changed every 3-4 days. Colonies of transfected cells formed within 10-14 days after transfection. Transfectant colonies were trypsinized and replated in selection medium for continued passage of the transfectant pools. The pools were assigned the designations shown in Table 1.

Cell membranes from cultured transfectant pools 2-22A, 2-22B, and 2-22C were prepared for ligand blot analysis (Graham and Ora , J. Biol. Chem. 262:7439- 42, 1987) . Membrane proteins were resolved on SDS polyacrylamide gels and transblotted to nitrocellulose membranes. Incubation of nitrocellulose strips with ¹²⁵ι-HDL₃ revealed the presence of a 110 kDa protein in membranes from all three transfeetants. Transfectant 2-22B (sense) membranes contained three times more of this protein than either 2-22A (anti-sense) or 2-22C (control) membranes. Monolayers of transfectants 2-22A, 2-22B, and

2-22C were chilled to 0^*C and incubated with either ¹²⁵I-HDL₃ or ¹²⁵I-LDL to test for cell-surface binding of these lipoproteins. Results indicated that transfectant 2-22B (sense) bound two to three times more ¹²⁵I-HDL₃ than either transfectant 2-22A or 2-22C, whereas binding of ¹²⁵I-LDL was the same for all three transfectants.

Cultured monolayers of transfectants 2-22A, 2-22B, and 2-22C were sterol-loaded by preincubation with medium enriched with cholesterol, then pulsed for three hours with ³H-mevalonolactone to biosynthetically radiolabel intracellular pools of sterol. Cells were then chased for four hours with medium containing either native HDL₃ or HDL₃ previously treated with trypsin to digest apoproteins, and the appearance of radiolabeled sterol in the medium was measured. Trypsin-treated HDL₃ was used as a control to monitor receptor-independent sterol efflux, since these particles do not interact with cell receptors. Results indicated that native HDL₃ promoted greater ³H-sterol efflux from transfectant 2-22B than did trypsin-treated HDL₃, whereas both treated and untreated HDL₃ removed the same amount of ³H-sterol from transfectants 2-22A and 2-22C.

Cultured transfectants 2-22A, 2-22B, and 2-22C were incubated with LDL containing ³H-cholesteryl esters to radiolabel intracellular sterol pools via a receptor- mediated lysosomal pathway. Cells were then chased for four hours with medium containing either no lipoprotein or HDL₃, and the appearance of ³H-sterol in the medium was measured. In the presence of HDL₃, transfectant 2-22B secreted two- to threefold more ³H-sterol than either transfectant 2-22A or 2-22C. Cholesterol-loaded monolayers of transfectants

2-22A, 2-22B, and 2-22C were incubated with medium containing either no lipoprotein or HDL₃. After six hours, cells were pulsed one hour with ¹⁴C-oleic acid, and incorporation of radiolabel into cellular cholesteryl esters was determined. This assay measures the relative degree of depletion of intracellular pools of free cholesterol mass available as substrate for the

- microsomal esterification enzyme. HDL₃ caused a two- to threefold greater depletion of this substrate pool in transfectant 2-22B than in either transfectant 2-22A or

2-22C.

Example 5 - Production of Antibodies Against an HDL-Binding Protein

E. coli Y1089 cells were lysogenized with phage clone 17RI-E3 essentially as described above. The hybrid ,3-galactosidase-HDL-binding proteins were IPTG- induced and resolved by polyacrylamide gel electrophore- sis (Young and Davis, ibid. ; Sikela and Hahn, Proc. Natl. Acad. Sci. USA 84.:³038-42, 1987). The hybrid protein was partially purified using a ?-galactosidase affinity substrate column (Sigma Chemical Co., St. Louis, Mo.) or purified using an Ε . coli £-galactosidase antibody column (Promega Biotech, Madison, Wis.) as specified by the supplier. The partially purified protein was used as an antigen for antibody production.

Balb/c mice were immunized multiple times with the hybrid ,5-galactosidase/HDL-binding protein molecule. The purified soluble molecule was blotted onto nitro¬ cellulose (Schleicher & Schuell, 0.45 μm) . The blots were dried at 37^*C, and the nitrocellulose was dissolved with DMSO. The sample was then mixed with Freund's adjuvant and injected into the peritoneal cavities of the mice. To obtain polyclonal antisera, blood was removed from the mice by retro-orbital bleeds and allowed to clot, and the serum was recovered. A second group of mice was immunized with partially purified membranes of bovine endothelial cells. Cell membranes isolated from bovine endothelial cells were separated by SDS-PAGE, then electrophoreti- cally transferred to nitrocellulose. The nitrocellulose sheets were incubated with ¹²5ι-iabeled HDL. A region of the nitrocellulose sheet corresponding to the 100 kDa range of the SDS gel was found to bind the radiolabeled HDL. The region of the nitrocellulose sheet correspond- ing to this area was then cut out and used for immuniza¬ tion of the mice. The nitrocellulose sheet was solubi¬ lized in DMSO, then mixed with Freund's adjuvant. This mixture was injected into the peritoneal cavities of the mice. This procedure was repeated multiple times. Polyclonal antisera were obtained by bleeding the mice and extracting the serum using conventional methods.

Hybridoma cell lines were generated by fusing lymphocytes from a mouse immunized with the isolated membranes of bovine endothelial cells with the NS-1 mouse myeloma cell line (ATCC TIB 18) .

To prepare for the fusion between the immunized mouse cells and the mouse myeloma cell line, the spleens and lymph nodes of the immunized mice were removed and minced with scissors on top of fine-mesh stainless steel screens. The minced tissues were washed through the fine screens into petri dishes with 10 ml RMPI 1640 (RPMI, GIBCO, Lawrence, Mass.) The remainder of the minced tissues was pressed through the screens with a spatula, and the screens were washed with 5 ml RPMI. To remove any remaining cell material, the bottoms of the screens were scraped and the cell material was added to the petri dishes.

The strained tissues were transferred to 50 ml centrifuge tubes. The petri dishes were washed with 10 ml RPMI to remove any remaining material. The cell suspensions were centrifuged for 10 minutes at 200 x g.

The supernatants were discarded, and the pellets were resuspended in 4 ml RPMI. After resuspension, 1 ml fetal calf serum (BioCell, Carson, Calif.) was added to each tube.

The red blood cell contaminants were lysed by adding 18 ml sterile distilled water to the cell suspen¬ sions. The mixtures were swirled quickly, and 5 ml 4.25% NaCl was added to each tube. The cell suspensions were centrifuged for 10 minutes at 200 x g. The super- natants were discarded, and the cell pellets resuspended in 10 ml RPMI.

To remove any remaining tissue material, the suspensions were filtered through two layers of sterile gauze into 50 ml tubes. The centrifuge tubes and gauze were rinsed with an additional 10 ml RPMI. Dilutions of the resultant cell suspensions were counted with a hemacytometer to determine the yield of lymphocytes. The prepared cells were kept at room temperature for approximately one hour until ready for use.

The NS-1 mouse myeloma cell line was used for the fusion. To optimize the fusion procedure, the NS-1 line was cloned out to isolate a clone with a high fusion efficiency. The NS-1 cells were cloned out by limiting dilution into 96-well microtiter plates at an average of five and ten cells per well in NS-1 medium (Table 2) + 2.5 x IO⁶ thymocytes/ml (prepared as described below) . The plates were incubated at 37*C, 7% CO₂ for ten days.

TABLE 2 NS-1 Medium

For a 500 ml solution:

5 ml 10 mM nonessential amino acids

(GIBCO, Lawrence, Mass.) 5 ml 100 mM sodium pyruvate (Irvine, Santa Ana, Calif.)

5 ml 200 mM L-glutamine (GIBCO) 5 ml 10Ox Penicillin/Streptomycin/Neomycin (GIBCO) 75 ml inactivated fetal calf serum

(BioCell, Carson, Calif.) 1 gm NaHC0₃

Add RPMI 1640 (GIBCO, Lawrence, Mass.) to a total volume of 500 ml.

Sterilize by filtration through a 0.22 μm filter.

10Ox HT Stock 38.5 mg thymidine

136.10 mg hypoxanthine (Sigma, St. Louis, Mo.)

Dissolve the thymidine and hypoxantine in distilled H₂0 and bring volume up to 100 ml. Warm the solution to 60°-70^βC to dissolve the solids. After the solids have dissolved, readjust the volume to 100 ml. Sterilize by filtration through a 0.22 μm filter. Store frozen at -20^βC.

lOOOx A Stock

17.6 ng aminopterin

Add sterile distilled water to the aminopterin and bring the volume to 50 ml. Add 1 N NaOH dropwise until the aminopterin dissolves. Bring the final volume to

100 ml with distilled H₂0. Sterilize by filtration through a 0.22 μm filter. Store frozen at -20°C.

5Ox HAT 50 ml 100X HT

5 ml 100Ox A stock 45 ml distilled H₂0

Sterilize the solution by filtration through a 0.22 μm filter. Store frozen at -20^βC. Freezing Medium

7 ml NS-1 medium

2 ml fetal calf serum

1 ml DMSO

Mix the ingredients and make fresh for each freezing.

On day 10, the cells were examined microscopi- cally and screened for wells containing single colonies. On the same day, 100 μl of fresh NS-1 medium containing 2.5 x 10 - thymocytes/ml was added to each well. On the fourteenth day, eight- of the most vigorously growing single' colonies were chosen to expand for fusion. The eight candidate colonies were transferred to individual 24-well plates containing 1.5-2 ml NS-1 medium + 2.5 x IO⁶ thymocytes/well. These plates were maintained at 37^*C, 7% C0₂ and the cells split at appro¬ priate intervals by expanding■ the cells to fresh wells of the plate and adding fresh NS-1 medium. This proce¬ dure was repeated until there was a sufficient number of cells to expand into a 75 cm² tissue culture flask. At this point, 1 x IO⁷ cells from the growing cultures were used to inoculate a 75 cm² tissue culture flask contain- ing 50 ml NS-1 medium. The flasks were incubated at 37^*C, 7% C0₂, until the cells reached a density of at least 5 x 10^s cells/ml. The cells were then harvested by centrifugation and diluted to a concentration of approximately 5 x 10⁶ cells/ml with freezing medium (Table 2) . The cells were divided into 1 ml aliquots and frozen stepwise first at -80^βC and then at -130^βC.

To assay the clones of NS-1 cells for fusion efficiency, one vial for each clone was quickly thawed in water held at 37^βC. The cells were inoculated into flasks containing NS-1 medium to a concentration of 2 x 10^s cells/ml. The cells were grown at 37°C with 7% C0₂. The cells were cut back to 2 x 10^s cells/ml daily. One day before fusion, two 75 cm² flasks containing 50 ml of cell culture were set up. Each of the candidate NS-1 clones was mixed with immunized mouse spleen cells and fused as described below. The results of the fusions showed that one of the single-colony isolates showed an increased fusion efficiency. This clone was designated clone F and was subsequently used as the myeloma cell line in all further fusions. Thymus glands obtained from baby mice were the source of the thymocytes which were used as a feeder layer to condition the culture media for the cell fusions. Thymus glands were obtained from 3- to 4-week- old Balb/c mice. The thymus glands were rinsed with NS- 1 medium and minced on a fine-mesh, stainless steel screen with scissors. The minced tissues were rinsed through the screen with 10 ml NS-1 medium into a petri dish. The thymus tissue was then pressed through the screen with a spatula into the petri dish. The screen was then washed with 10 ml NS-1 medium. The bottom of the screen was scraped to remove any adhered tissue, and the tissue was pooled in the petri dish. The strained tissue was transferred to a 50 ml centrifuge tube through two layers of sterile gauze. The petri dish and gauze were rinsed with an additional 10 ml NS-1 medium. The cells were centrifuged for 10 minutes at 200 x g. The supernatants were discarded and the pellets resuspended in 10 ml NS-1 medium. Dilutions of the cell suspensions were counted using a hemacytometer. The yield from two thymus glands was routinely about 400 million cells. The cells were stored at room tempera¬ ture until ready for use.

To prepare cells for fusion, the NS-1 clone F cells, grown as described above, were transferred to 50 ml centrifuge tubes and centrifuged for 10 minutes at 200 x g. The supernatants were discarded, and the cell pellets were resuspended and combined in 10 ml RPMI. The cell suspension was counted with a hemacytometer to monitor the cell viability as determined by trypan blue exclusion. The cell viability was determined to be greater than 95%. 5 For fusion, 2.5 x IO⁷ NS-1 clone F cells were added to the prepared immunized mouse cells (prepared as described above) . The mixed cells were centrifuged for 10 minutes at 200 x g. The supernatant was removed by aspiration with a Pasteur pipet attached to a vacuum

10. line. The cell pellet was resuspended in 100 μl RPMI and warmed in a water bath at 37'C.

One ml 50% polyethylene glycol (PEG-1500) solution in RPMI, was pH adjusted within the range of pH 7.0 to pH 8.0 with 20 μl 1% sodium bicarbonate. The PEG

15 solution was added to the cell suspension over a period of one minute with gentle stirring. The suspension was stirred for an additional minute. One ml of NS-1 medium was added over a period of one minute with gentle stirring. Over an additional period of ,one minute, 1 ml

20 of NS-1 medium was added to the suspension with gentle stirring. Eight ml NS-1 medium was added over a period of two minutes. The suspension was centrifuged at 125 x g at room temperature for 10 minutes. The supernatant was discarded, and 25 ml NS-1 medium was added to gently

25- resuspend the pellet. The cells were harvested gently into a pipet and transferred to a 175 cm² flask.

Four hundred million thymocytes (prepared as above) were added to the flask. The volume was adjusted to 160 ml with NS-1 medium, and the mixture was

30 incubated at 37^*C, 7% C0₂ for two to four hours.

After incubation, 3.2 ml of 50x HAT (Table 2) was added. The cell suspension was transferred to eight 96-well culture plates at 200 μl per well, using pipet tips with the ends cut off to avoid shearing forces.

35 The plates were incubated at 37^βC in a humid incubator with 7% C0₂. The plates were examined microscopically after three days to determine fusion efficiency, with the expectation of approximately five hybridoma colonies per well. The cells were fed after seven days by replacing 100 μl of the medium with fresh NS-1 medium containing 1 x HAT and 2.5 x 106 thymocytes per ml. The hybridomas were tested between days 9 and 14 for the production of specific antibodies.

The hybridoma cells were screened by Western blot (Tσwbin et al., ibid.) for the production of antibodies to the bovine endothelial cell membranes. Bovine endothelial cell membranes were separated by SDS- PAGE, electrophoretically transferred to nitrocellulose, and the nitrocellulose was incubated in blocking buffer to prevent nonspecific binding by the antibodies. Sections of the nitrocellulose sheet were then incubated with conditioned media from the hybridoma cell lines, and the presence of specific antibody was determined using ABC-peroxidase (Vector Laboratories, Burlinga e, Calif.) as described by the manufacturer. Several hybridoma cells secreting antibody to endothelial cell membranes were detected.

The antibodies are characterized for detection of the 110 kDa HDL-binding protein and for the recogni¬ tion of the recombinant HDL-binding protein.

The hybridoma cell lines producing antibodies of interest are cloned by limiting dilution to obtain colonies of cells originating from single cells. These colonies are expanded in culture for generation of large amounts of antibody.

Example 6 - HDL Agonist Screening Assay

Anti-HDL antibodies are screened for possible antagonistic and agonistic properties. Cells expressing HDL receptor in a constitutive manner (e.g., bovine aortic endothelial cells) or cells transfected with an expression vector encoding an HDL-binding protein are plated on microtiter plates and exposed to the anti- ■ bodies. The cells are then incubated with ¹²⁵ι-HDL to identify antibodies that compete for HDL binding to intact cells.

Parallel cultures of cells are used for the preparation of membrane proteins for ligand blot immunoblot analysis. Membrane proteins are bound to nitrocellulose strips, and the strips are incubated with antibody, then exposed to ¹²⁵ι-HDL to determine which antibodies block HDL binding to the isolated proteins. Antibodies that bind to isolated HDL-binding proteins without blocking HDL binding are designated as non- blocking antibodies. Both blocking and non-blocking antibodies are used in functional assays to assess their abilities to act as HDL agonists or antagonists. For the functional assay, cultured cells (e.g., fibroblasts, endothelial cells, or macrophages) are treated with ³H- mevalonolactone to radiolabel intracellular pools of cholesterol. . The cells are then exposed to media containing antibodies plus either native HDL or other cholesterol acceptor particles. . Other acceptor particles include chemically treated HDL that no longer interact with HDL receptors but are able to act as cholesterol acceptors (Brinton et al. , J. Biol. Chem. £1:495-503, 1986). The effects of antibodies on the efflux of labeled sterol indicate agonistic and antagonistic properties. Those antibodies that block the efflux-stimulating effects of HDL are HDL antagonists; those that stimulate sterol efflux in the presence of a sterol acceptor that does not bind HDL receptors are HDL agonists. The effects of the antibodies on translocation of radiolabeled sterol from intracellular sites to the cell surface in the presence or absence of native HDL are measured by determining their abilities to stimulate the clearance of sterol from sterol-laden macrophages. Macrophages loaded with acetylated LDL-derived cholesterol are incubated in medium containing the test antibody plus either native or chemically modified HDL particles, and the rate of disappearance of excess intracellular sterol is deter¬ mined as a function of time. These studies provide an indication of the therapeutic potential of individual agonist antibodies in promoting the clearance of excess 5 sterol from foam cells of the artery wall.

From the foregoing, it will be appreciated that, although specific embodiments of the invention have been described herein for purposes of illustration, various modifications may be made without deviating from 10 the spirit and scope of the invention. Accordingly, the invention is not limited except as by the appended claims.

15

20

25

30

>

35

Claims

Claims

1. A substantially pure mammalian protein having the following characteristics: a) molecular weight of approximately 110,000 Daltons; b) binds HDL₃; c) binds apoA-I proteoliposomes arid apoA-II proteoliposomes; d) contains approximately 10% carbohydrate by weight; and e) an isoelectric point between about 5.8 and 6.0.

2. A protein according to claim 1 - wherein said protein comprises a sequence of amino acids selected from the group consisting of the amino acid sequences as shown in Figure 4, from methionine, amino acid number 1, to glutamic acid, amino acid number 1195, Figure 5, from methionine, amino acid number 1, to isoleuceine, amino acid number 1292, and Figure 3, from glycine, amino acid number 2, to glutamic acid, amino acid number 125.

3. An isolated DNA molecule encoding a protein having the following characteristics: a) molecular weight of approximately 38,000 Daltons; b) binds HDL3; and c) contains approximately 331 amino acid residues.

4. A DNA molecule according to claim 3 wherein said protein comprises the amino acid sequence of Figure 1, from glycine, amino acid number 2, to glutamic acid, amino acid number 331.

5. A DNA molecule according to claim 3 wherein said molecule comprises a sequence of nucleotides as shown in Figure 1, from nucleotide 378 to nucleotide 1370.

6. A DNA molecule comprising a sequence of at least about 45 nucleotides of the DNA sequence shown in Figure 1, said sequence encoding an immunogenic peptide.

7. A substantially pure protein having the following characteristics: a) molecular weight of approximately 38,000 Daltons; b) binds HDL₃; and c) comprises a sequence of amino acids as shown in Figure 1, from glycine, amino acid number 2, to glutamic acid, amino acid number 331.

8. An isolated DNA molecule encoding a protein having the following characteristics: a) molecular weight of approximately 100,000 Daltons exclusive of carbohydrate; b) binds HDL ; and c) binds apoA-I and apoA-II proteoliposomes.

9. A DNA molecule according to claim 8 wherein said protein has an isoelectric point between about 5.8 and 6.0.

10. A DNA molecule according to claim 8 wherein said protein comprises a sequence of amino acids selected from the group consisting of the amino acid sequences as shown in Figure 4, from methionine, amino acid number 1, to glutamic acid, amino acid number 1195, Figure 5, from methionine, amino acid number 1, to isoleuceine, amino acid number 1292, and Figure 3, from glycine, amino acid number 2, to glutamic acid, amino acid number 125.

11. A DNA molecule according to claim 8 comprising a sequence of nucleotides selected from the group consisting of the nucleotide sequences as shown in Figure 4, from nucleotide 155 to nucleotide 3739, Figure 5, from nucleotide 155 to nucleotide 4030, and Figure 3, from nucleotide 10 to nucleotide 381.

12. Cells transfected or transformed with an expression vector comprising a promoter operably linked to a DNA sequence, according to any one of claims 3-5 or 8-11.

13. The cells of claim 12 wherein said cells are yeast cells or cultured mammalian cells.

14. An antibody that binds to a protein accordin to any one of claims 1, 2 or 7.

15. A method for detecting a genetic defect in a HDL-binding protein in a patient, comprising: isolating DNA from a patient; cleaving said DNA with a restriction endonuclease t produce DNA fragments; hybridizing said DNA fragments to a DNA molecul comprising a nucleotide sequence selected from the grou consisting of: a) at least about 18 contiguous nucleotides of th sequence shown in Figure 1; b) at least about 18 contiguous nucleotides of th sequence shown in Figure 3; c) at least 18 contiguous nucleotides of th sequence shown in Figure 4; d) at least 18 contiguous nucleotides of th sequence shown in Figure 5; and e) a sequence of about 18 nucleotides complemen tary to sequence (a) , sequence (b) , sequence (c) or sequenc (d) ; and detecting a pattern of hybridization of said DNA fragments and said DNA molecule, wherein said pattern is diagnostic for the presence or absence of a genetic defect in HDL-binding protein.

16. An immunogenic polypeptide comprising a portion of an amino acid sequence selected from the group consisting of the amino acid sequence of Figure 1, the amino acid sequence of Figure 3, the amino acid sequence of Figure 4, and the amino acid sequence of Figure 5, wherein said polypeptide is at least about 6 amino acids in length.

17. An oligonucleotide which .hybridizes to a nucleotide sequence selected from the group consisting of the nucleotide sequence of Figure 1, the nucleotide sequence of Figure 3, the nucleotide sequence of Figure 4, and the nucleotide sequence of Figure 5, wherein said oligonucleotide is at least about 18 nucleotides in length.

18. An oligonucleotide comprising a portion of the nucleotide sequence of Figure 1, Figure 3, Figure 4 or Figure 5, wherein said oligonucleotide is at least about 18 nucleotides in length.