CHIMERIC MOUSE HUMAN ANTIBODIES WITH SPECIFICITY TO HIV ANTIGENS

BACKGROUND OF THE INVENTION

FIELD OF THE INVENTION

This invention relates to genetically-engineered antibodies with specificity for the Human Immunodeficiency Virus (HIV) and their derivatives, nucleotide and protein sequences coding therefor, as well as methods of obtaining and manipulating such sequences.

BACKGROUND ART

Monoclonal antibody (mAb) technology has greatly impacted current thinking about human disease management. The elegant application of cell to cell fusion for the production of mAbs by Kohler and Milstein (Nature (London) 256:495. 1975) spawned a revolution in biology equal in impact to that of recombinant DNA cloning. MAbs produced from hybrido as are already widely used in diagnostic and basic scientific studies. Their efficacy in the treatment of human diseases, including viral and microbial infections, to a large extent remains to be demonstrated.

Although they display exquisite specificity and can influence the progression of human disease, mouse mAbs, by their very nature, have limitations in their applicability to human medicine. Most obviously, since they are derived from mouse cells, they are recognized as foreign protein when introduced into humans and elicit immune responses. Similarly, since they are distinguished from human proteins, they are cleared rapidly from circulation. Finally, mouse antibodies may not be recognized as effectively as human antibodies by human effector cells or molecules.

Technology to develop human mAbs that could circumvent these particular problems has met a number of obstacles. In

many cases, human mAb-producing cell lines are obtained from Epstein Barr Virus (EBV) immortalized cells derived from human blood and may therefore not be useful for scale-up and production of human pharmaceuticals. Immortalized human antibody-producing cell lines may also carry human viral sequences including the Human Immunodeficiency Virus.

In addition, since human viruses are thought to have evolved in the face of the human immune response, key antigens may not be recognized by the human immune system. Such antigens would not be expected to elicit useful immune responses in man. In contrast, those viral antigens that are immunogenic in mice can be used for the production of mouse mAbs which may have therapeutic utility in humans. The novel, chimeric antibodies of this invention, have been created using both the hybridoma and genetic engineering technologies to provide reagents. The chimeric antibodies and products derived therefrom should have significant utility for the treatment and diagnosis of human disease.

SUMMARY OF THE INVENTION

The invention provides engineered chimeric mouse-human antibodies of desired variable (V) region specificity able to recognize certain antigens encoded by HIV and selected human constant (C) region properties, produced after gene cloning and expression of light (L) and heavy (H) chains. The chimeric antibodies and their derivatives may have applicability in the treatment and diagnosis of individuals infected with the Human Immunodeficiency Virus (HIV). The cloned immunoglobulin gene products and their derivatives can be produced in mammalian or microbial cells.

The invention provides cDNA sequences coding for immunoglobulin chains comprising a human C region and a non- human V region. The immunoglobulin chains are both heavy and light.

EI

The invention provides sequences as above, present in recombinant DNA molecules, in vehicles such as plasmid vectors, capable of expression in desired prokaryotic or eukaryotic hosts.

The invention provides hosts capable of producing, by culture, the chimeric antibodies and methods of using these hosts.

The invention also provides individual chimeric immunoglobulin chains, as well as complete assembled molecules having human C regions and mouse V regions with specificity for HIV antigens, wherein both V regions have the same binding specificity.

Among other immunoglobulin chains and/or molecules, provided by the invention are:

1. An antibody with monovalent specificity for an HIV antigen, i.e., a complete, functional immunoglobulin molecule comprising:

(a) Two different chimeric H chains, one of which comprises a V region with anti-viral specificity, and

(b) Two different L chains, with the corresponding specificities as the V regions of the H chains. The resulting heterobifunctional antibody would exhibit monovalent binding specificity toward a viral antigen.

2. Antibody fragments such as Fab, Fab', and F(ab')2- Genetic sequences, especially cDNA sequences, coding for the aforementioned combinations of chimeric chains are also provided herein.

The invention also provides for a genetic sequence, especially a cDNA sequence, coding for the V region of desired specificity of an antibody molecule H and/or L chain with or without human C regions, linked to a sequence coding for a polypeptide different than an immunoglobulin chain (e.g., an enzyme). These sequences can be assembled by the methods of

the invention, and expressed to yield mixed-function molecules.

The use of cDNA sequences is particularly advantageous over genomic sequences (which contain introns), in that cDNA sequences can be expressed in bacteria or other hosts which lack appropriate RNA splicing systems.

BRIEF DESCRIPTION OF THE FIGURES

FIGURE 1. Nucleotide sequence of the coding strand for the 2E12 mouse L chain V region (V|_). Shown is the nucleotide sequence from the end of the oligo-dC tail to the J _κ 2 - C _κ junction. Also shown is the amino acid sequence deduced from the nucleotide sequence. Shown in bold are the oligonucleotides used for site directed utagenesis and the sites at which restriction site modifications were made.

FIGURE 2. Nucleotide sequence of the coding strand for the 2E12 mouse H chain V region (Vμ) . Shown is the nucleotide sequence from the end of the oligo-dC tail to the JJ_J2 - C^l junction. Also shown is the amino acid sequence deduced from the nucleotide sequence. Shown in bold are the oligonucleotides used for site directed mutagenesis and the sites at which restriction site modifications were made.

FIGURE 3. Construction scheme for the chimeric mouse-human Chimerl H chain mammalian expression plasmid, pING3112. The variable region from the cDNA clone pClG-12 was engineered to be compatible with the eucaryotic expression plasmid pING2227. Plasmid pING2227 contains the following gene reguTatory elements useful for expression in mammalian cells: 1) the IgG H chain enhancer element, 2) an Abelson LTR promoter, 3) the SV40 16S splice site, and 4) the IgG H chain polyadenylation signal sequence. It also contains the entire human IgGl C region from pGMH-6 (Liu, A. Y., et al . , Proc. Natl. Acad. Sci., USA 84:3439-3443, 1987). pING3112 contains the neomycin phosphotransferase gene which allows for G418 selection in transfected cells.

FIGURE 4. Construction scheme for the chimeric mouse-human Chimerl L chain mammalian expression plasmid PING3110. The V region from the cDNA clone pClK-7 was engineered to be compatible with the eucaryotic expression plasmid pING1712. Plasmid pING1712 contains the following gene regulatory elements useful for expression in mammalian

«_ U. BST.TUT

cells: 1) the IgH enhancer element, 2) the Abelson LTR promoter, 3) the SV40 16S splice site, and 4) a human kappa (K) polyadenylation signal sequence. It also contains the entire human C _κ region, Liu A.Y., et al . , supra, and the GPT gene which allows for mycophenolic acid resistance in transfected cells.

FIGURE 5. Construction scheme for the bacterial Chimerl Fab expression plasmid pING3127. Plasmid pING3127 contains the following elements useful for expression in E. coli: 1) the araC gene, 2) the inducible araB promoter, 3) the dicistronic Fd and K Chimerl genes fused to the pelB leader sequence, 4) the trpA transcription termination sequence, and 5) the tetR gene, useful for selection in E. coli. The. specific cloning steps are described in detail in Example 3.

FIGURE 6. Yeast expression plasmids for Fab expression. Shown are the yeast expression plasmid pING3114 containing the Chimerl L chain gene fused to the yeast PGK promoter, invertase signal sequence and PGK polyadenylation signal (a); the similar yeast plasmids pING3117 and pING3137 containing the Fd gene (b); the yeast expression plasmid pING3118 containing the L chain promoter/leader fusion with PGK transcription termination signal (c); similar yeast plasmid pING3138 containing the Fd gene (d); and the final 2 gene yeast expression plasmid pING3142 (e) .

FIGURE 7. Nucleotide sequence of the coding strand for the 2G12 mouse Vμ region. Shown is the nucleotide sequence from the end of the oligo-dC tail to the J^3 - Cμl junction. Also shown is the amino acid sequence deduced from the nucleotide sequence. Shown in bold are the oligonucleotides used for site directed mutagenesis and the sites at which restriction site modifications were made.

FIGURE 8. Nucleotide sequence of the coding strand for the 2G12 L chain mouse V region. Shown is the nucleotide sequence from the end of the oligo-dC tail to the J _κ l - C _κ junction. Also shown is the amino acid sequence deduced from

SUBSTITUTE SHEET

the nucleotide sequence. Shown in bold are the oligonucleotides used for site directed mutagenesis and the sites at which restriction site modifications were made.

FIGURE 9. Construction scheme for the chimeric mouse-human Chimer2 H chain mammalian expression plasmid, pING3004. The variable region from the cDNA clone pC2G-6 was engineered to be compatible with the eukaryotic expression plasmid pING2227. Plasmid pING2227 contains the gene regulatory elements described in the legend to Figure 3.

FIGURE 10. Construction scheme for the chimeric mouse-human Chimer2 L chain mammalian expression plasmid pING3005. The V region from the cDNA clone pC2K-14 was engineered to be compatible with the eucaryotic expression, plasmid pING1712. Plasmid pING1712 contains the gene regulatory elements described in the legend to Figure 4.

FIGURE 11. Construction scheme for the bacterial Chimer2 Fab expression plasmid pING3211. Plasmid pING3211 contains the elements useful for expression in E. coli described in the legend to Figure 5. The specific cloning steps are described in detail in Example 5.

FIGURE 12. Nucleotide sequence of the coding strand for the 1C11 mouse V|_| region. Shown is the nucleotide sequence from the end of the oligo-dC tail to the ^3 - Cμl junction. Also shown is the amino acid sequence deduced from the nucleotide sequence. Shown in bold is the oligonucleotide used for site directed mutagenesis and the sites at which restriction site modifications were made.

FIGURE 13. Nucleotide sequence of the coding strand for the 1C11 mouse V ₍ _ region. Shown is the nucleotide sequence from the end of the oligo-dC tail to the J _κ l - C _κ junction. Also shown is the amino acid sequence deduced from the nucleotide sequence. Shown in bold is the oligonucleotide

used for site directed mutagenesis and the sites at which restriction site modifications were made.

FIGURE 14. Construction scheme for the chimeric mouse-human Chimer4 H chain mammalian expression plasmid, pING2255. The variable region from the cDNA clone pC4M-8 was engineered to be compatible with the eucaryotic expression plasmid pING2227. Plasmid pING2227 contains the gene regulatory elements described in the legend to Figure 3.

FIGURE 15. Construction scheme for the chimeric mouse-human Chimer4 L chain mammalian expression plasmid pING2258. The V region from the cDNA clone pC4K-16 was engineered to be compatible with the eucaryotic expression plasmid pING1712. Plasmid pING1712 contains the gene, regulatory elements described in the legend to Figure 4.

FIGURE 16. Construction scheme for the bacterial Chimer4 Fab expression plasmid. This plasmid contains the elements useful for expression in E. coli described in the legend to Figure 5. The specific cloning steps are described in detail in Example 8.

FIGURE 17. Nucleotide sequence of the coding strand for the 4D12 mouse V^ region. Shown is the nucleotide sequence from the end of the oligo-dC tail to the J^4 - Cμl junction. Also shown is the amino acid sequence deduced from the nucleotide sequence. Shown in bold are the oligonucleotides used for site directed mutagenesis and the sites at which restriction site modifications were made.

FIGURE 18. Nucleotide sequence of the coding strand for the 4D12 mouse V[_ region. Shown is the nucleotide sequence from the end of the oligo-dC tail to the J _κ 5 - C _κ junction. Also shown is the amino acid sequence deduced from the nucleotide sequence. Shown in bold are the oligonucleotides used for site directed mutagenesis and the sites at which restriction site modifications were made.

FIGURE 19. Construction scheme for the chimeric mouse-human Chimer5 H chain mammalian expression plasmid,

SUBSTITUTE SHEEt

pING3126. The variable region from the cDNA clone pC5G-30 was engineered to be compatible with the eucaryotic expression plasmid pING2227. Plasmid pING2227 contains the gene regulatory elements described in the legend to Figure 3.

FIGURE 20. Construction scheme for the chimeric mouse-human Chimerδ L chain mammalian expression plasmid pING3132. The V region from the cDNA clone pC5K-4 was engineered to be compatible with the eucaryotic expression plasmid pING1712. Plasmid pING1712 contains the gene regulatory elements described in the legend to Figure 4.

FIGURE 21. Construction scheme for the bacterial Chimerδ Fab expression plasmid pING3139. Plasmid pING3139 contains the elements useful for expression in E. coli as described in the legend to Figure 5.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

GENETIC PROCESSES AND PRODUCTS

The invention provides chimeric antibodies that can be used for the treatment and diagnosis of individuals infected with HIV, either alone or in combination with other reagents. The chimeric antibodies contain mouse V regions which recognize certain antigens encoded by the HIV genome. In several of the . examples described herein, the chimeric antibodies recognize those antigens recognized by the mouse mAbs 2E12, 2G12, 1C11, and 4D12.

The method of production combines five elements:

1. Isolation of messenger RNA (mRNA) from the mouse B cell hybridoma line producing the monoclonal antibody, cloning and cDNA production therefrom;

2. Preparation of a full length cDNA library from purified RNA from which the appropriate V region gene segments of the L and H chain genes can be: (i) identified with appropriate probes, (ii) sequenced, and (iii) made compatible with a C gene segment.

3. Preparation of C region gene segment modules by cDNA preparation and cloning.

4. Construction of complete H or L chain coding sequences by linkage of the cloned specific immunoglobulin V region gene segments described in 2. above to cloned human C region gene segment modules described in 3.

5. Expression and production of chimeric L and H chains in selected hosts, including prokaryotic and eukaryotic cells.

One common feature of all immunoglobulin H and L chain genes and the encoded messenger RNAs is the so-called J region. H and L chain J regions have different sequences, but. a high degree of sequence homology exists (greater than 80%) among each group, especially near the C region. This homology is exploited in this invention and consensus sequences of H and L chain J regions were used to design oligonucleotides for use as primers for introducing useful restriction sites into the J region for subsequent linkage of V region segments to human C region segments.

C region cDNA module vectors prepared from human cells and modified by site-directed mutagenesis to place a restriction site at the analogous position in the human sequence were used. For example, one can clone the complete human C _κ region and the complete human Cγl region. An alternative method utilizing genomic C region clones as the source for C region module vectors would not allow these genes to be expressed in systems such as bacteria where enzymes needed to remove intervening sequences are absent. Cloned V region segments are excised and ligated to L or H chain C region module vectors. In addition, the human Cγl region can be modified by introducing a termination codon thereby generating a gene sequence which encodes the H chain portion of an Fab molecule.

STITUTE SHEET

The coding sequences with linked V and C regions are then transferred into appropriate expression vehicles for expression in appropriate hosts, prokaryotic or eukaryotic.

Two coding DNA sequences are said to be "operably linked" if the linkage results in a continuously translatable sequence without alteration or interruption of the triplet reading frame. A DNA coding sequence is operably linked to a gene expression element if the linkage results in the proper function of that gene expression element to result in expression of the coding sequence.

Expression vehicles include plasmids or other vectors. Preferred among these are vehicles carrying a functionally complete human Cμ or C|_ chain sequence having appropriate- restriction sites engineered so that any Vμ or V|_ chain sequence with appropriate cohesive ends can be easily inserted therein. Human Cμ or C _| _ chain sequence-containing vehicles are thus an important embodiment of the invention. These vehicles can be used as intermediates for the expression of any desired complete H or L chain in any appropriate host.

One preferred host is yeast. Yeast provides substantial advantages for the production of immunoglobulin H and L chains. Yeasts carry out post-trans!ational peptide modifications including glycosylation. A number of reco binant DNA strategies now exist which utilize strong promoter sequences and high copy number plasmids which can be used for production of the desired proteins in yeast. Yeast recognizes leader sequences of cloned mammalian gene products and secretes peptides bearing leader sequences (i.e., prepeptides) (Hitzman, et al . , 11th International Conference on Yeast, Genetics and Molecular Biology, Montpellier, France, September 13-17, 1982).

Yeast gene expression systems can be routinely evaluated for the levels of production, secretion and the stability of chimeric H and L chain proteins and assembled chimeric antibodies. Any of a series of yeast gene expression systems

SUBSTITUTE SHEET

incorporating promoter and termination elements from the actively expressed genes coding for glycolytic enzymes produced in large quantities when yeasts are grown in media rich in glucose can be utilized. Known glycolytic genes can also provide ver efficient transcription control signals. For example, the promoter and terminator signals of the phosphoglycerate kinase (PGK) gene can be utilized. A number of approaches may be taken for evaluating optimal expression plasmids for the expression of cloned immunoglobulin cDNAs in yeast (see Glover, D.M., ed., DNA Cloning, Vol. II, pp45-66, IRL Press, 1985).

Bacterial strains may also be utilized as hosts for the production of antibody molecules or antibody fragments described by this invention, E. coli K12 strains such as L coli 3110 (ATCC 27325), and other enterobacteria such as Salmonella typhimurium or Serratia marcescens, and various Pseudo onas species may be used.

Plasmid vectors containing replicon and control sequences which are derived from species compatible with a host cell are used in connection with these bacterial hosts. The vector carries a replication site, as well as specific genes which are capable of providing phenotypic selection in transformed cells. A number of approaches may be taken for evaluating the expression plasmids for the production of chimeric antibodies or antibody chains encoded by the cloned immunoglobulin cDNAs in bacteria (see Glover, D.M., ed., DNA Cloning, Vol . I, IRL Press, 1985).

Other preferred hosts are mammalian cells, grown in vitro or in vivo. Mammalian cells provide post-translational modifications to immunoglobulin protein molecules including leader peptide removal, folding and assembly of H and L chains, glycosylation of the antibody molecules, and secretion of functional antibody protein.

Mammalian cells which may be useful as hosts for the production of antibody proteins include cells of lymphoid

TIT TE SHEET

origin, such as the hybridoma Sp2/0-Agl4 (ATCC CRL 1581) or the myeloma P3X63Ag8 (ATCC TIB 9), and its derivatives. Others include cells of fibroblast origin, such as Vero (ATCC CRL 81) or CH0-K1 (ATCC CRL 61).

Many vector systems are available for the expression of cloned H and L chain genes in mammalian cells (see Glover, D.M., ed., DNA Cloning, Vol. II, ppl43-238, IRL Press, 1985). Different approaches can be followed to obtain complete H2L2 antibodies. It is possible to co-express H and L chains in the same cells to achieve intracellular association and linkage of H and L chains into complete tetrameric H2L2 antibodies. The co-expression can occur by using either the same or different plasmids in the same host. Genes for both H and L chains can be placed into the same plasmid, which is then transfected into cells, thereby selecting directly for cells that express both chains. Alternatively, cells may be transfected first with a plasmid encoding one chain, for example the L chain, followed by transfection of the resulting cell line with an H chain plasmid containing a second selectable marker. Cell lines producing H2L2 molecules via either route could be transfected with plasmids encoding additional copies of H, L, or H plus L chains in conjunction with additional selectable markers to generate cell lines with enhanced properties, such as higher production of assembled H2L2 antibody molecules or enhanced stability of the transfected cell lines.

POLYPEPTIDE PRODUCTS

The invention provides chimeric immunoglobulin chains, either H or L, with specificity toward viral antigens of the human immunodeficiency virus, HIV. A chimeric chain contains a C region substantially similar to that present in a natural human immunoglobulin, and a non-human V region having the desired anti-viral specificity of the invention. The invention also provides immunoglobulin molecules having H and

L chains associated so that the overall molecule exhibits the desired binding and recognition properties.

Various types of immunoglobulin molecules are provided: monovalent, divalent, or molecules with the invention's V region binding domains attached to moieties carrying desired functions. This invention also provides for "fragments" of chimeric immunoglobulin molecules, which include Fab, F(ab'), and F(ab')2 molecules. The invention also provides for "derivatives" of the chimeric immunoglobulins, which term includes those proteins encoded by truncated or modifed genes to yield molecular species functionally resembling the immunoglobulin fragments. The modifications include, but are not limited to, addition of genetic sequences coding for cytotoxic proteins such as plant and bacterial toxins and tumor necrosis factor (TNF). The fragments and derivatives can be produced from any of the hosts of this invention.

Antibodies, fragments or derivatives having chimeric H chains and L chains of the same or different V region binding specificity, can be prepared by appropriate association of the individual polypeptide chains, as taught, for example by Sears et al. (Proc. Natl . Acad. Sci. USA 72:353-357 (1975)). With this approach, hosts expressing chimeric H chains (or their derivatives) are separately cultured from hosts expressing chimeric L chains (or their derivatives), and the immunoglobulin chains are separately recovered and then associated. Alternatively, the hosts can be co-cultured and the chains allowed to associate spontaneously in the culture medium, followed by recovery of the assembled immunoglobulin, fragment or derivative.

USES

The chimeric antibodies of this invention, of which Chimerl, Chimer2, Chimer4, and Chimerδ are examples, recognize epitopes of antigens encoded by the HIV genome, including core antigens, reverse transcriptase, and envelope glycoproteins.

SUBSTITUTE SHEET

- lδ -

These HIV antigens, some of which are expressed on the surface of infected cells, include the glycoproteins encoded by the viral eny gene (gpl60, gpl20, gp41), and proteins encoded by the viral ig gene (pδδ, p4δ, p39, p24, and pl8), and βol gene (p6δ, pδl) (Musing, M.A. et al . , Nature 313:450-458 (1985)). The Chimerl, Chimer2, Chimer4, and Chimerδ antibodies recognize various of these antigens (see Examples, below) and can be used alone or in combination with antibodies that recognize other viral components.

The chimeric antibodies react with viruses, subviral particles, viral proteins, viral peptides, and HIV-infected cells expressing viral antigens and should exhibit therapeutic activity in infected individuals via normal immune and host defense mechanisms. The antibodies of this invention should be useful for treatment of asymptomatic HIV-infected individuals to prevent progression of the infectious process, as well as individuals with clinical symptoms of HIV infection (Acquired Immune Deficiency Syndrome (AIDS) and AIDS Related Complex (ARC)), by reducing viral burden and alleviating symptoms related to circulating viral antigens or virus- infected cells. Circulating HIV proteins including gpl20 and p24 are thought to be associated with immunological dysfunction. Administration of Chimerl, Chimer2, Chimer4, and Chimerδ, either singly or in combination therapy, should benefit individuals with low titers of anti-HIV antibodies.

Treatment of an individual infected with HIV using the antibodies, fragments or derivatives of this invention comprises parenterally admininstering a single or multiple doses of the antibody, fragment or derivative. The effective dose is a function of the individual chimeric antibody, the presence and nature of a conjugated therapeutic agent (see below), the patient and his clinical status, and can vary from about 10 ng/kg body weight to about 100 mg/kg body weight. The route of administration may include intravenous,

subcutaneous, intramuscular, intrapulmonary, intraperitoneal , intranasal, intrathecal, transder al or other known routes.

The antibodies of this invention can be adapted for therapeutic efficacy by virtue of their ability to mediate antibody-dependent cellular cytotoxicity (ADCC) and/or complement-dependent cytotoxicity (CDC) against HIV-infected cells. For these activities, either an endogenous source or an exogenous source of effector cells (for ADCC) or complement components (for CDC) can be utilized. Antibodies that effectively mediate ADCC or otherwise mark infected cells for destruction by host effector cells should be especially useful after acute HIV infection when relatively few host cells carry the virus.

The chimeric antibodies of this invention, their fragments, and derivatives can be used therapeutically as immunoconjugates (see for review: Dillman, R.O., Ann. Int. Med. 111:592-603 (1989)). They can be coupled to cytotoxic proteins, including, but not limited to Ricin-A, Pseudomonas toxin, Diptheria toxin, and TNF. Toxins conjugated to antibodies or other ligands, are known in the art (see, for example, Olsnes, S. et al . , Immunol. Today 10:291-295 (1989)). Plant and bacterial toxins typically kill cells by disrupting the protein synthetic machinery.

The chimeric antibodies of this invention can be conjugated to additional types of therapeutic moieties including, but not limited to, radionuclides, cytotoxic agents and anti-viral drugs, to treat individuals infected with HIV. Examples of radionuclides which can be coupled to antibodies and delivered in vivo to sites of antigen include ²¹² Bi, ¹³¹ I, l^Re, and ⁹⁰ Y, which list is not intended to be exhaustive. The radionuclides exert their cytotoxic effect by locally irradiating the cells, leading to various intracellular lesions, as is known in the art of radiotherapy.

Cytotoxic drugs which can be conjugated to antibodies and subsequently used for in vivo therapy include, but are not

SUBSTITUTE SHEET

limited to, daunorubicin, doxorubicin, ethotrexate, and Mitomycin C. Cytotoxic drugs interfere with critical cellular processes including DNA, RNA, and protein synthesis. Antiviral drugs which may be conjugated to antibodies and subsequently used for in vivo therapy include, but are not limited to, acyclovir, azidothy idine, adenine arabinoside, dideoxyinosine, and protease inhibitors. Antiviral drugs act by interfering with virus-specific enzymes such as reverse transcriptase, with viral proteases, and with the metabolism and incorporation of nucleotides into nucleic acids. For a fuller exposition of these classes of drugs which are known in the art, and their mechanisms of action, see Goodman, A.G., et al., Goodman and Gilman's THE PHARMACOLOGICAL BASIS OF THERAPEUTICS. 7th Ed., Macmillan Publishing Co., 1985).

The chimeric antibodies, fragments or derivatives of this invention may be advantageously utilized in combination with other chimeric antibodies, or with lymphokines or hemopoietic growth factors, etc., which serve to increase the number or activity of effector cells which interact with the antibodies.

Antibodies specific for HIV-infected cells should play an active role in limiting HIV spread within an individual. Earlier intervention in the infectious process may delay or prevent the later emergence of more pathogenic HIV strains or variants within an infected individual. By controlling viral replication, the antibodies, fragments or derivatives of this invention may permit the establishment and maintenance of a longer asymptomatic course which would improve the life of infected individuals.

Interventive therapy with the chimeric antibodies of this invention may also be useful either alone or in conjunction with other**anti-viral therapies to prevent accidental viral exposure from developing into symptomatic disease. Prophylactic immunization with the antibodies of this invention may effectively reduce the risk of a successful

SUBSTITUTE SHEET

viral infection upon challenge, much as other antibody preparations are used in the management of hepatitis and cytomegalovirus infections.

The chimeric antibodies, fragments, or derivatives of this invention, attached to a solid support, can be used to remove the virus, viral antigens, or virus-infected cells from fluids or tissue or cell extracts. In a preferred embodiment, they are used to remove HIV from blood or blood plasma products. In another preferred embodiment, the chimeric antibodies are advantageously used in extracorporeal immunoadsorbent devices, which are known in the art (see, for example, Seminars in Hematology, Vol. 26 (2 Suppl . 1) (1989)). Patient blood or other body fluid is exposed to the attached antibody, resulting in partial or complete removal of circulating HIV virions, viral antigens (free or in immune complexes), or virus-infected cells, following which the fluid is returned to the body. This immunoadsorption can be implemented in a continuous flow arrangement, with or without interposing a cell centrifugation step. See, for example, Terman, D.S. et al., J. Immunol. 117:1971-1975 (1976).

Specifically, the chimeric antibodies of this invention can be used for any and all uses in which the murine mAbs can be used, with the clear advantage that mouse-human chimeric antibodies are more compatible with the human body.

Having now generally described the invention, the same will be further understood by reference to certain specific examples which are included herein for purposes of illustration only and are not intended to be limiting unless otherwise specified.

SUBSTITUTE SHEET

EXAMPLE 1

A Chimeric Mouse-Human Immunoglobulin (Chimerl)

Produced by Mammalian Cells and Specific for an HIV Envelope Protein

The mouse mAb 2E12 (described as anti-gpl20,160 mAb by Yoshihara, P. et al., Proc. 4th Int'l. Conf. on AIDS. June, 1988, p237) was derived from a hybridoma fusion utilizing spleen cells from a BALB/c mouse immunized with an HIV viral lysate, and was then cloned in the presence of purified eny gene product. The Sp2/0 myeloma line was used as the fusion partner. Clone 2E12 produces immunoglobulin of the IgGl subclass. MAb 2E12 was reactive against cloned eny gene product by ELISA analysis. In addition, 2E12 was capable of immunofluorescent staining of HIV-infected cells. Western blot analysis of the 2E12 mAb against viral lysates demonstrated a predominant 120 kD band as well as a weaker 160 kD band, depending upon the lysate used. 1. Recombinant Plasmid and Bacteriophage DNAs

Oligo-dG tailed pBR322, pUC18, pUC19, M13mpl8, and M13mpl9 were purchased from BRL (Gaithersburg, MD) . DNA manipulations involving purification of plasmid DNA by buoyant density centrifugation, restriction endonuclease digestion, purification of DNA fragments by agarose gel electrophoresis, ligation and transformation of E. coli were as described by Maniatis, T., et al . , Molecular Cloning: A Laboratory Manual, (1982), or other standard procedures. Restriction endonucleases and other DNA/RNA modifying enzymes were purchased from Boehringer-Mannheim (Indianapolis, IN), BRL, and New England Biolabs (Beverly, MA).

TITUTE SHEET

2. RNA Purification and cDNA Library Construction One liter of 2E12 hybridoma cells at approximately 1 x 10 ⁶ cells/ml were collected by centrifugation and washed in 100 ml of PBS (8g NaCl, 0.2g KH ₂ P0 ₄ , l.lδg Na ₂ HP0 ₄ ,and 0.2g KCl per liter). The cells were centrifuged again, the cell pellet suspended in a solution of guanidine thiocyanate, and total cellular RNA was prepared from tissue culture cells by the method described in Maniatis, T., et al . , supra. Preparation of poly(A)+ RNA was as described by Maniatis, T., et al . , supra. Oligo-dT primed cDNA libraries were prepared from poly(A)+ RNA by the methods of Gubler, V., et al . (Gene 25:263 (1983)). The cDNA was dC-tailed with terminal deoxynucleotide transferase and annealed to dG-tailed pBR322. cDNA libraries were screened by hybridization (Maniatis, T., et al . , supra) with 2p_ι _a beled, nick-translated DNA fragments, i.e., for K clones with a mouse C _κ region probe and for H chain clones with a mouse IgGl C region probe.

The V|_ and Vμ region fragments from the full length cDNA clones, pClK-7 and pClG-12 respectively, were inserted into M13 bacteriophage vectors for nucleotide sequence analysis. The complete nucleotide sequences of the V region of these clones were determined (Figures 1 and 2) by the dideoxy chain termination method. These sequences predict V region amino acid compositions that agree well with the observed compositions, and predict peptide sequences which have been verified by direct amino acid sequencing of portions of the V regions.

The nucleotide sequences of the cDNA clones show that they are immunoglobulin V region clones as they contain amino acid residues diagnostic of V domains (Kabat, et aj_. ,

Sequences of Proteins of Immunological Interest, U.S. Dept. of HHS, 1983).

SUBSTITUTE SHEET

3. Construction of Chimeric Expression Plasmids

Expression vectors appropriate for the insertion of V|_| and V|_ gene modules to obtain expression of Chimerl were constructed. The L chain vector pING1712 was made by first making a plasmid DNA containing a test chimeric L chain gene (pING2122) and adding a mouse Abelson LTR promoter, a splice region, and a mouse genomic kappa region 3' to the polyadenylation signal. The H chain mouse enhancer 0.7 kb Xbal to EcoRI fragment from M13 M8alphaRX12 (Robinson, R.R., et al . , supra) was inserted into Xbal plus EcoRI cut M13mpl9. The enhancer-containing HindlH to Bglll fragment was inserted into the BgUI to HindlH region of pSH6, an E. coli recombinant plasmid DNA that contains unique Xhol, BgUI, and HindlH sites, in that order. The enhancer-containing Xbal to Xhol fragment was then inserted into the enhancer )(bal to Xhol region of pING2121b, an expression plasmid identical to pING2108b (Liu, A.Y., et al . , J. Immunology 139:3521 (1987)) except that the L6 V _| _ region (Liu, A.Y., et al . , Proc. Nat! . Acad. Sci. USA 84:3439 (1987)) was used in its construction instead of the 2H7 V[_ region. The resulting plasmid was pING2122.

The mouse Abelson virus LTR was obtained from pel in2 (provided by Dr. Owen Witte, UCLA). pelin2 contains the pl20 viral 3' LTR (Reddy, E.P., et al . , Proc. Natl . Acad. Sci. USA 80: 3623 (1983)) except that the Bglll site at viral position 4623 has been modified by insertion of the EcoRI oligonucleotide linker GGATTCC. The 0.8 kb EcoRI to Kpnl fragment of pelin2 containing the pl20 3' LTR promoter was inserted into Kp_nl plus EcoRI cut pUC18. The LTR was excised as an EcoRI to Sail fragment and ligated to EcoRI plus Sail cut pING2122, resulting in a plasmid where the LTR promoter is adjacent to the L6 L chain gene (pING2126). An Xhol to Sail fragment containing SV40 16S splice donor and acceptor sites was excised from pUC12/pLl (Robinson, R.R. et al . , supra) and inserted into the Sail site of pING2126, screening for the

UTE SHEET

orientation where the splice donor was between the LTR and the L chain gene (pING2133).The polyadenylation/transcription termination region of the kappa expression vector was also modified. The first step was the HindUI digestion and religation of plasmid pING2121a, which is identical to pING2108a (Liu, A.Y., et al . , supra) except that the L6 V[_ was used in its construction instead of the 2H7 Vj_, to form pING2121a-deltaH. The 1.1 kb BgHI to BamHI fragment of mouse genomic DNA distal to the polyadenylation site (Xu, M., et al., J. Biol. Chem. 261:3838 (1986) was isolated from pS107A (provided by Dr. Randolph Wall, UCLA) and inserted into the BamHI site of pING2121a-deltaH, screening for the orientation homologous to the native gene. The 3.3 kb Bglll to SstI fragment containing this modified 3' region was ligated to the 5.2 kb Bglll to SstI fragment of pING2121b to form pING1703. The Bglll to Sail fragment of pING1703 with the modified 3' region and chimeric kappa coding sequence was ligated to the large Bglll to Sail fragment of pING2133, resulting in the 9.1 kb kappa expression vector pING1712 shown in Figure 4.

The Abelson LTR promoter was also used in the chimeric H chain expression vector pING1714. pING2111 (Robinson, R.R., e_t al. , supra) was modified by the insertion of an Aatll oligonucleotide linker at the Xbal site, followed by Aatll cleavage and religation to form pING1707. The Aatll to Sail fragment containing the Abelson LTR promoter was excised from pING2133 and ligated to the large Aatll to Sail fragment of pING1707 to form pING1711. The H chain enhancer was deleted from pING1711 by EcoRI digestion, T4 polymerase treatment, ligation to Aatll ol gonucleotide linker, and cleavage and religation with Aatll to form the 7.7kb expression vector PING1714.

A similar plasmid, pING2227, contains two additional regulatory elements, the IgH enhancer and the human genomic IgG polyadenylation sequence. pING2227 is identical to pING1712 in the region from Bglll to Sail containing the IgH

SUBSTITUTE SHEET

enhancer the Abelson LTR promoter, and the 16S slice donor and acceptor sites. The human genomic IgG 3' end sequence was ligated as an 1185 bp XmaHI DNA fragment into an XmaHI site located 6 bp past the termination codon for the H chain gene in pING1714. The 1300 bp XmaHI fragment containing the genomic gamma 3' end was isolated from a derivative of pHG3A (Ellison, et al . , Nucl . Acids Res. 10:4071 (1982)).

4. Construction of Chimerl H and L Chain Expression Plasmid The cDNA clone containing the 2E12 H chain, pClG-12, was adapted for mammalian expression by introducing convenient restriction endonuclease sites by site directed mutagenesis

(Kramer, W., et aj_., Nucl. Acids Res. 11:9441) into appropriate M13 subclones, Figure 3. Oligonucleotides were synthesized on a Cyclone DNA synthesizer, New Brunswick Scientific Co., and purified by acrylamide gel electrophore- si s . The J region mutagenesi s primer 5'-GGCTGAGGAGACGGTGACCGTGG-3' was used to insert a fistEII site into the M13 subclone pClGSDM, and the oligonucleotide δ'-GAGGTCCTGTCGACTTAGTAACTGT-3' was used to insert a Sail restriction site into pClGBstEII upstream of the initiation codon ATG. The restriction fragment containing the 2E12 V _| _ _| region bounded by SalI and BstEII was then cloned into the expression vector pING2227.

The cDNA clone containing the 2E12 L chain, pClK-7, was adapted for mammalian expression in a similar way, Figure 4. The J region mutagenesis primer δ'-GTTTTATTTCAAGCTTGGTCC-3' was used to insert a HindlH site into the M13 subclone p C l K S DM , a n d t h e o l i g o n u c l e o t i d e δ'-TGAGAACTTGGTCGACAGAGTCCGGCG-3' was used to insert a Sail restriction site into pClKHIII upstream of the initiation codon ATG. The restriction fragment containing the 2E12 V _| _ region bounded by Sail and Hindlll was then cloned into the expression vector pING1712.

δ. Stable Transfection of Mouse Lymphoid Cells for the Production of Chimeric Antibody

The cell line Sp2/0 (American Type Culture Collection #CRL1581) was grown in Dulbecco's Modified Eagle Medium plus 4.5 g/1 glucose (DMEM, Gibco) plus 10% fetal bovine serum. Media were supplemented with glutamine/penici11in/streptomycin (Irvine Scientific, Irvine, California). The electroporation method of Potter, H., et al ., (Proc. Nat! . Acad. Sci., USA, 81:7161 (1984)) was used. After transfection, cells were allowed to recover in complete DMEM for 24 hours, and then seeded at 1-3 x 10^ cells per well in 96-well culture plates in the presence of selective medium. G418 (GIBCO) selection was at 0.8 mg/ml, and ycophenolic acid (Calbiochem) was at 6 μg/ml plus 0.25 mg/ml xanthine. The electroporation technique gave a transfection frequency of 1-10 x 10 ^" ^ for the Sp2/0 cells.

The Chimerl L chain expression plasmid pING3110 was linearized by digestion with £yul restriction endonuclease and transfected into Sp2/0 cells, giving mycophenolic acid resistant clones which were screened for L chain synthesis. The best producer after outgrowth and subsequent subcloning, was transfected with PyuI-linearized pING3112, the expression plasmid containing the Chimerl H chain gene. After selection with G418, the clone producing the most L plus H chain, Sp2/0-3310 11E10 + 3312 5C11.4C9, secreted antibody at approximately δ /_g/ml .

6. Purification of Chimerl Antibody Secreted in Tissue Culture

Sp2/0-3110 11E10.1B2 + 3112 5C11.4C9 cells were grown in culture medium HB101 (Hana Biologies) + 1% Fetal Bovine Serum, supplemented with lOmM HEPES, lx Glutamine-Pen-Strep (Irvine Scientific #9316). The spent medium was centrifuged at about δ,000 xg for 20 minutes. The antibody level was measured by ELISA. Approximately 23L of cell culture supernatant was

ITUTE SHEET

- 2δ - concentrated 14-fold over a S10Y30 cartridge using DC-10 concentrator (Amicon Corp). Supernatant containing about 14 mg of antibody was loaded onto a 2 ml Protein A-column (Pharmacia) in O.lδ M NaCl, 10 mM sodium phosphate buffer pH 8.4. The Chimerl antibody was eluted with a step pH gradient (pH 5.6 and 3.5). Fractions containing antibody (75% yield) were combined and concentrated 22.5-fold by ultrafiltration (YM30 membrane, stirred cell, Amicon Corp.) diluted 25-fold with PBS, reconcentrated 10-fold, diluted 10-fold with PBS, and finally reconcentrated 5-fold. The antibody was stored in 1.0 ml aliquots at -20°C.

7. Study Performed on the Chimerl Antibody

A test was performed with Chimerl to show that it retained the antigen binding characteristics of the mouse 2E12 antibody. The test demonstrated that both 2E12 and Chimerl recognized the same viral antigen. Commercially available HIV Western Strips (Dupont or equivalent) were incubated with either Chimerl (2 g/ml) or 2E12 (10-20 jug/ml). Strips incubated with 2E12 mouse antibody were incubated with rabbit anti-mouse antibody followed by Protein A-gold followed by silver enhancement (BioRad). Strips incubated with Chimerl were incubated with Protein A-gold directly. Specific binding to a viral antigen at approximately 120 kD molecular weight was detected with each antibody.

In a second test, Chimerl and 2E12 were compared in an ELISA test. The Genetics Systems (Seattle, WA) HIV-1 ELISA kit was used as specified by the manufacturer except that the second antibody used for detection recognized either the C region of mouse (2E12) or human (Chimerl) antibody. Results showing that the antibody of this invention and 2E12 react similarly are shown in Table 1.

E SHEET

TABLE 1.

Second

Antibody Concentration ^A 4δO ^J Antibody Specificity

1 Absorbance measured at 4δ0 nm.

² Genetic Systems Positive and Negative Human Control

EXAMPLE 2

A Chimeric Mouse-Human Fab Fragment Produced in E. coli and Specific for an HIV Envelope Protein

Bacteria are suited for production of chimeric antibodies expressed from mammalian cDNA since entire coding sequences can be expressed from well characterized promoters. Escherichia coli is one of many useful bacterial species for

SUBSTITUTE SHEET

production of foreign proteins since a wealth of genetic information is available for optimization of its gene expression. E. coli can be used for production of foreign proteins internally or for secretion of proteins out of the cytoplasm, where they most often accumulate in the periplasmic space (Gray, et al.. Gene 39:247 (1985); Oka, et al., Proc. Natl. Acad. Sci. USA 82:7212 (1985)). Secretion from the L coli cytoplasm has been observed for many proteins and requires a signal sequence. Proteins produced internally in bacteria are often not folded properly (Schoner, et aj_. ,

BioTechnology 3:151 (1985)) . Protein secreted from bacteria, however, is often folded properly and assumes native secondary and tertiary structures (Hsiung, et aj_. ,

BioTechnology 4:991 (1986).

A Fab molecule consists of two nonidentical protein chains linked by a single disulfide bridge. These two chains are the intact antibody L chain and the V, J, and Cμl portions of the antibody heavy chain, Fd. The proper cDNA clones for the Chimerl L and Fd genes have already been identified. In this example, these cDNA clones were organized into a single bacterial operon (a dicistronic message) as gene fusions to the pectate lyase (pelB) gene leader sequence from Erwinia carotovora (Lei, et al . , J. Bacteriol. 169:4379 (1987) and expressed from a strong regulated promoter. The result is a system for the simultaneous expression of two protein chains in E. coli, and the secretion of immunologically active, properly assembled Fab of Chimerl antibody. The following sections detail the secretion of Chimerl Fab from E. coli.

1. Assembly of the pelB Leader Sequence Cassette

Erwinia carotovora (EC) codes for several pectate lyases

(polygalacturonic acid trans-eli inase) (Lei, et al., Gene

35:63 (1985)) . Three pectate lyase genes have been cloned, and the DNA sequence of these genes has been determined. When cloned into E. coli under the control of a strong promoter,

T TE SHEET

the pelB gene is expressed and large quantities of pectate lyase accumulate in the periplasmic space and culture supernatant. The pelB signal sequence functions efficiently in E. coli and was used as a secretion signal for antibody genes in this example. (Other signal sequences would also be useful for this application.) The nucleotide sequence surrounding the signal sequence of the pelB gene is published (Lei, et al ., supra).

The pelB signal sequence contains a HaelH restriction site at amino acid 22, adjacent to the signal peptidase cleavage site: ala-ala. Plasmid pSS1004 (Lei, et al . , supra) containing the pelB gene in pUC8 (Vierra et al., Gene 19:259 (1982)) was digested with HaelH and EcoRI. This DNA was ligated with an eight base pair SstI linker to Ssp_I and EcoRI cut pBR322. The resulting plasmid contained a 300 bp fragment which included the 22 amino acid leader sequence of pelB and about 230 bp of upstream E. caratovora DNA. This plasmid, pING173, contains an insert that upon digestion with SstI and treatment with T4 DNA polymerase can be ligated directly to a DNA fragment flanked by the first amino acid of a mature coding sequence for any gene to generate a protein fusion containing a functional bacterial leader sequence in frame with the incoming gene. The SstI to EcoRI restriction fragment in pING173 was cloned into pUC18 (Yanich-Perron, et al.. Gene 33:103 (1985)) to generate pRR17δ, which contains the pelB leader and adjacent upstream non-coding sequence (including a ribosome binding site) downstream of the lac promoter. Plasmid pINGlδOO, derived from pRR17δ, contains only the region from the -48 of the pelB gene to an Xhol site downstream of the pelB leader, and includes the SstI site at the junction.

SUBSTITUTE SHEET

2. In Vitro Mutagenesi s

Site-directed in vitro mutagenesis was performed as described by Kramer, et aj.. _. supra, to introduce a BsmI restriction site into the 2E12 L chain cDNA sequence (Figure 1) at the junction of the leader peptide and mature coding region with the ol igonucl eotide primer δ'-TCACAATCTCCGCATTCACTCCAGAGAT-3' .

A Kj)nl site was similarly introduced at the junction of the leader peptide and mature coding region of the 2E12 H chai n with the ol igonucl eotide primer δ'-GTTGCACCTGGTACCGGACACCTGTAG-3' (Figure 2).

3. Preparation of L Chain for Bacterial Expression

The Chimerl V _[ _ region containing a BsmI restriction site at the signal sequence processing site and a unique HindHI site in the J region of pClKHB served as the starting point for bacterial expression. The plasmid pClKHB was cut with BsmI. treated with T4 polymerase, and digested with HindlH (Figure 5A). The 320 bp fragment containing the V _| _ region was purified and ligated to pINGlδOO that was cut with SstI, treated with T4 polymerase, and cut with Xhol (Figure δB) along with a HindlH to Xhol restriction fragment containing the human C _κ plus lδ nucleotides of 3' genomic DNA from pING3102 (Figure δC). The resulting plasmid that contained a pelB: :Chimerl L chain fusion was sequenced to determine that the proper in-frame fusion was formed. This plasmid was called pING3111.

4. Preparation of Fd for Bacterial Expression

The intact Chimerl chimeric Fd gene containing a Kpnl restriction site at the signal sequence processing site and a BstEII restriction site in the J region in pClGBK served as the starting point for bacterial expression. The plasmid pClGBK was cut with Kp_nl, treated with T4 polymerase, and digested with BstEII (Figure 5D). The approximately 335 bp

TE SHEET

fragment containing the Fd V region was purified and ligated to pINGlδOO that was cut with SstI, treated with T4 polymerase, and cut with Xhol (Figure 5E) along with the human CHI region (which had been previously generated by introducing a stop codon in hinge, Robinson, R.R., et al . , supra) from pF3D (Figure 5F). The resulting plasmid that contained a pelB: :Chimerl Fd fusion was sequenced to determine that the proper in-frame fusion was formed. This plasmid was called pING311δ.

5. Multicistronic Expression System for L Chain and Fd Gene For production of bacterially-derived Fab, both L chain and Fd need to be produced simultaneously within the cell. Using the plasmids constructed with each of these genes separately, a series of expression vectors were constructed that contain both genes aligned so that transcription from a single promoter will specify both genes. This was done in a way that minimized the noncoding DNA between the two genes. Each gene has a ribosome binding site needed for translation initiation and the identical DNA sequence from -48 to the pelB leader::antibody gene junction.

Plasmid pING31U was cut with Sfihl, treated with T4 polymerase, cut with EcoRI, and the vector fragment was purified (Figure 5G). Similarly, pING311δ was cut with Xhol, treated with T4 polymerase, cut with EcoRI and the fragment containing the Fd gene was purified (Figure 5H). These two purified DNA fragments were ligated to produce pING3116, which contained the two Chimerl gene fusions linked in close proximity. The two-gene cistron was placed under the control of the araB promoter in pING3104. Plasmid pING3116 was cut with S£hl, treated with T4 polymerase, cut with Xhol, and the fragment containing the Fd and K genes was purified (Figure 51). This DNA fragment was ligated to the vector fragment from pING3104 that had been cut with EcoRI, treated with T4 polymerase, and cut with Xhol (Figure 5J), generating

UBSTITUTE SHEET

pING3119. The unique BstEII fragment in the Fd C region was shown by restriction analysis to be in the incorrect orientation; when the Fd gene was assembled, it assembled improperly. To correct this, pING3119 was cut with BstEII and religated to generate the final Chimerl Fab vector pING3127 (Figure 5K). This vector contains all the necessary features for expression of Chimerl Fab in E. coli.

6. Production of Chimerl Fab in Bacteria

Expression of Chimerl Fab from pING3127 in E. coli is under the inducible control of the araB promoter. Upon arabinose induction, Fab secreted into the growth medium increases more than 10-fold. The E. coli strain harboring pING3127 was cultured at 32 ^β C in 10L of minimal medium, supplemented with 1.7% glycerol as the carbon source. Dissolved oxygen was maintained at 20% of saturation by addition of a 5X concentrate of glycerol-supplemented minimal medium. The culture was induced at 0D5gø= 0 with 0.05% arabinose for greater than 16 hours in a 14L fer enter (Chemap). Fab was detected in the fermentation broth by antibody sandwich ELISA using rabbit anti-human K chain antiserum as the solid phase agent to bind Fab, followed by detection with monoclonal anti-human Fd and goat anti-mouse IgG peroxidase conjugate.

About 7L of a culture supernatant was concentrated to 2 liters using a S10Y10 cartridge (DC10 concentrator, Amicon Corp.). The concentrate was passed through a 500g DEAE cellulose type DE52, Whatman) column pre-equilibrated with lOmM sodium phosphate at pH 8.0. Sufficient 0.2M monosodium phosphate was added to adjust pH to 6.8, and the sample was concentrated over a YM10 membrane (Stirred Cell 2000, Amicon). The sample was then diluted with sufficient water and reconcentrated to 200 ml to give a conductivity of 1.1 mS/cm. The total amount of protein was estimated by a colorimetric assay, and the sample was loaded onto a carboxy ethylcellulose

HEET

type (CM52, Whatman) column at a ratio of 10 mg total protein per lg CM52 (pre-equilibrated with lOmM sodium phosphate buffer at pH 6.8). The CM52 column was eluted with a linear gradient of increasing NaCl concentration (0-0.IN) in the same phosphate buffer. The fractions containing Fab (assessed by enzyme immunoassay) were further analyzed by SDS-PAGE and then pooled. The combined Fab fractions were concentrated over a YM10 membrane to an Fab concentration of about 1 mg/ml , and stored frozen.

The Chimerl Fab purified from E. coli has identical molecular weight properties as other Fab molecules produced from E. coli, as assessed by SDS gel electrophoresis. The bacterially-produced Chimerl Fab is correctly assembled as a kappa plus Fd chain dimer because of its positive reaction in the ELISA detecting molecules with both Kappa and Fd determinants, and both chains are properly folded because the Fab reacts with gpl20 on Western immunoblot test strips.

7. Study Performed on Chimerl Fab

Chimerl Fab was tested along with 2E12 mouse antibody and Chimerl antibody for binding to antigen fixed to commercially available HIV Western immunoblot test strips (Dupont or equivalent). Chimerl Fab was incubated with the test strip, followed by interaction with goat anti-human Fab antibody, and then with rabbit anti-goat IgG. Detection of bound antibody was with Protein A-gold followed by silver enhancement (BioRad). Chimerl Fab specifically recognized the same size antigen on the test strip as did Chimerl whole antibody, specifically recognizing an antigen of about 120 kD.

TITUTE SHEET

EXAMPLE 3

A Chimeric Mouse-Human Fab Fragment Produced in Yeast and Specific for an HIV Envelope Protein

Yeast cells are capable of expressing and secreting foreign proteins. In this example, yeast served as the host for the production of mouse-human chimeric Fab.

1. Yeast Strains and Growth Conditions

Saccharomyces cerevisiae strain PS6 (ura3 leu2 MATa) was developed at INGENE and used as a host for yeast transformations performed as described by Ito, et aj_. , « .

Bacteriol. 163:163-168 (1983). Yeast transformants were selected on SD agar (2% glucose, 0.67% yeast nitrogen base, 2% agar) and grown in SD broth buffered with 50mM sodium succinate, pH 5.δ.

2. Construction of Yeast Expression Plasmids Containing Antibody Genes

The gene sequences encoding the mature form of the VL region of 2E12 and containing a HindlH site in the J region (as described in Example 1) and a BsmI site introduced at the signal sequence processing site was fused to the human C _κ region. Plasmid pCIKHB was cut with BsmI _. treated with T4 polymerase, and cut with HindlH. The V region fragment was ligated along with the human C _[ _ region prepared from pING3102 as a HindlH to Xhol fragment into pING1149. The plasmid pING1149 contains the gene sequence encoding the yeast invertase signal sequence (Taussig, R. et al . , Nucl .Acids Res. 11:1943-1964 (1983)) under control of the yeast PGK promoter (Hitzeman, R.A., et al . , Nucl. Acids Res. 10:7791-780 (1982)). The resultant plasmid pING3114 (Figure 6A) contains the PGK promoter (P), along with the invertase signal sequence (S), fused to the Chimerl coding sequence. As the result of this fusion, the gene sequence encoding the mature form of the Chimerl L chain was fused in frame to the gene sequence encoding the yeast invertase signal sequence. The PGK

promoter - invertase signal sequence - chimeric L chain (V,C _K ) fusion was cloned into a partial 2 micron circle (2μ), ura3 yeast expression vector containing the PGK polyadenylation signal (Tm) to generate pING3118 (Figure 6C) . The gene sequence encoding the mature form of the Vμ region of Chimerl containing a BstEII site in the J region (as described in Example 1) and a Kpnl site introduced at the signal sequence processing site from pCIGBK was fused along with the human C^l region (which had been previously generated by introducing a stop codon in hinge, Robinson, R.R., et al ., supra) from pF3D into pING1149 that had been cut with PstI, treated with T4 polymerase and cut with Xhol. This generated plasmids pING3117 and pING3137. Plasmid pING3117 had the correct fusion junction between the invertase leader sequence and the mature V region coding sequence, but a BstEII fragment from the Cμl region of pF3D was in the incorrect orientation. Plasmid pING3137 had improperly fused the invertase leader to the Chimerl Vμ region, but had the C region BstEII fragment in the correct orientation. A BamHI to PstI fragment from pING3117 containing the PGK promoter, invertase leader and a portion of the Chimerl V region was ligated along with a PstI to Xhol fragment from pING3137 which contained the remainder of the Chimerl V region and the O-C^l portion into a partial 2 micron circle (2μ) expression vector containing the PGK polyadenylation signal (Tm) to generate pING3138, Figure 6D.

A single yeast expression vector containing both the chimeric L chain and Fd chain genes and their respective expression signals was constructed from pING3118 and pING3138. This final vector, pING3142, Figure 6E, contains a portion of 2 micron circle (oriY, REP3) and the two selectable markers 1eu2d and ura3.

3. Yeast Secretion of Chimerl Fab

The plasmid pING3142 was transformed into S. cerevisiae PS6 and the transformants were grown in broth under selective

SUBSTITUTE SHEET

- 3δ - conditions as described above. The culture supernatants were assayed by ELISA and contained Fab levels of approximately 900 ng/ml . This material can be purified from fermentation broth and is expected to have properties identical to those of Fab produced as described in Example 2; yeast Chimerl Fab is expected to have the same binding specificity as Fab produced in E. coli.

EXAMPLE 4

A Chimeric Mouse-Human Immunoglobulin (Chimer2) Produced in Mammalian Cells and Specific for an HIV gag Protein

The mouse mAb 2G12 (described as anti-p24 mAb in: Yoshihara, P. et al . , Proc. 4th Int'l. Conf. on AIDS, June, 1988, p237; and Marcus-Sekura, C.J. et al .. Biochim. Biophvs. Acta 949:213-223 (1988)) was derived from a hybridoma fusion utilizing spleen cells from a BALB/c mouse immunized with purified gag gene product. The Sp2/0 myeloma line was used as the fusion partner. Clone 2G12 produced immunoglobulin of the IgGl subclass. MAb 2G12 was reactive against cloned gag gene product by ELISA analysis. In addition, 2G12 was capable of immunofluorescent staining HIV-infected cells. Western blot analysis of the 2G12 mAb against viral lysates demonstrated predominant bands developed at 55, 45, 39, and 24 kD.

1. RNA Purification and cDNA Library Construction

One liter of 2G12 hybridoma cells at approximately 1 x 10^ cells/ml were collected by centrifugation and washed in 100 ml of PBS. RNA was prepared, a cDNA library was constructed in pBR322, and L and H chain cDNA clones were identified as described in Example 1.

The V _| _ and Vμ fragments from the full length cDNA clones, pC2K-14 and pC2G-6 respectively, were inserted into M13 bacteriophage vectors for nucleotide sequence analysis. The complete nucleotide sequences of the V region of these clones were determined (Figures 7 and 8) by the dideoxy chain

TE SHEET

termination method. These sequences predict V region amino acid compositions that agree well with the observed compositions, and predict peptide sequences which have been verified by direct amino acid sequencing of portions of the V regions.

The nucleotide sequences of the cDNA clones show that they are immunoglobulin V region clones as they contain amino acid residues diagnostic of V domains (Kabat, et al . , supra).

2. Construction of Chimer2 H and L Chain Expression Plasmid The cDNA clone containing the 2G12 H chain, pC2G-6, was adapted for mammalian expression by introducing convenient restriction endonucleases sites by convenient cloning and site directed mutagenesis (Kramer, W., et aj_., supra) of appropriate M13 subclones, Figure 9. Oligonucleotides were synthesized on a Cyclone DNA synthesizer, New Brunswick Scientific Co., and purified by acrylamide gel electro- phoresis. An Ndel site, Figure 7, located approximately 45 bp upstream of the initiation codon ATG was used for introduction of a Sail restriction site 5' of the antibody V region. The J region mutagenesis primer δ'-GAGACGGTGACCAGAGTCCCT-3' was used to insert a BstEII site into the M13 subclone pSWO. The restriction fragment containing the 2G12 Vμ region bounded by Sail and BstEII was then cloned into the expression vector pING2240 (pING2240 is identical to pING2227 except it contains a different antibody V region).

The cDNA clone containing the 2G12 L chain, pC2K-14, was adapted for mammalian expression in a similar way, Figure 10. The J region mutagenesis primer δ'-GTTTGATTTCAAGCTTGGTGC-3' was used to insert a HindlH site into the M13 subclone pC2-K, and the oligonucleotide δ'-TGTCTGTGTGTCGACAGTGTGATGC-3' was used to insert a Sail restriction site into pSW6 upstream of the initiation codon ATG. The restriction fragment containing the 2G12 V[_ region bounded by Sail and HindlH in pSW7 was then cloned into the expression vector pING2216 (pING2240 is

SUBSTITUTE SHEET

identical to pING1712 except it contains a different antibody V region).

3. Stable Transfection of Mouse Lvmphoid Cells for the Production of Chimeric Antibody

The cell line Sp2/0 (American Type Culture Collection #CRL1581) was grown, transfected, and selected as described in Example 1, above. The Chimer2 L chain expression plasmid pING3005 was linearized as described in Example 1. The best producer after outgrowth and subsequent subcloning, was transfected with Pvul-linearized pING3004, the expression plasmid containing the Chimer2 H chain gene. After selection with G418, the clone producing the most L plus H chain, Sp2/0-3005 + 3004 3E9, secreted antibody at approximately 5 μg/ml.

4. Purification of Chimer2 Antibody Secreted in Tissue Culture

Sp2/0-3005 + 30043E9 cells were grown in culture medium HB101 (Hana Biologies) + 6% Fetal Bovine Serum, supplemented with 10 mM HEPES, lx Glutamine-Pen-Strep (Irvine Scientific #9316). The spent medium was centrifuged at about 14,000 xg for 20 minutes and the supernatant was filtered through a .4δu MiHipore nitrocellulose membrane filter and stored frozen. The antibody level was measured by ELISA. Approximately 11.3L of cell culture supernatant was concentrated 7-fold over a S10Y30 cartridge using DC-10 concentrator (Amicon corp). Supernatant containing about 60 mg of antibody was loaded onto a 2 ml Protein A-column (Pharmacia) in PBS pH 7.4 several times. The Chimer2 antibody was eluted with various pH gradients (pH 7-2) and eluted between pH 3.5-4.0. Fractions containing antibody (46% yield) were combined and concentrated 8δ-fold by ultrafiltration (YM30 membrane, stirred cell, Amicon Corp.) diluted 25-fold with PBS, reconcentrated

10-fold, diluted 10-fold with PBS, and finally reconcentrated 4-fold. The antibody was stored in 1.0 ml aliquots at -20 ^β C.

5. Studies Performed on the Chimer2 Antibody

A test was performed with Chimer2 to show that it retained the antigen binding characteristics of the mouse 2G12 antibody. The test demonstrated that both antibodies recognized the same viral antigen. Commercially available HIV Western Strips (Dupont or equivalent) were incubated with either Chimer2 (2 μg/ml) or 2G12 (10-20 μg/ml). Strips incubated with 2G12 mouse antibody were incubated with rabbit anti-mouse antibody followed by Protein A-gold followed by silver enhancement (BioRad). Strips incubated with Chimer2 were incubated with Protein A-gold directly. Specific binding to a viral antigen of approximately 24 kD was detected with each antibody.

In a second test, Chimer2 and 2G12 were compared in an ELISA test, as described in Example 1. Results showing that Chimer2 and 2G12 reacted similarly are shown in Table 1.

EXAMPLE 5

A Chimeric Mouse-Human Fab Fragment Produced in E. coli and Specific for an HIV gag Protein

Bacteria are suited for production of chimeric antibodies expressed from mammalian cDNA since entire coding sequences can be expressed from well characterized promoters. The following sections detail the secretion of Chimer2 Fab from E coli .

1. In Vitro Mutagenesis

Site directed in vitro mutagenesis was performed as described by Kramer, et al. , supra, to introduce a BsmI restriction site into the 2G12 L chain cDNA sequence (Figure 8) at the junction of the leader peptide and mature coding

SUBSTITUTE SHEET

region with the ol igonucl eotide primer 5'-ATCTGGATGTCAGCATTCGCACCTGTAAG-3' .

A BsmI site was similarly introduced at the junction of the leader peptide and mature coding region of the 2G12 H chai n with the ol igonucl eotide primer δ'-CTGGACCTCAGCATTCACACCTGCAGT-3' (Figure 7).

2. Preparation of L Chain for Bacterial Expression

The Chimer2 VL region containing a BsmI restriction site at the signal sequence processing site and a unique HindlH site in the J region of pSW8 served as the starting point for bacterial expression. The plasmid pSW8 was cut with BsmI, treated with T4 polymerase, and digested with HindlH (Figure 11A). The approximately 320 bp fragment containing the V _| _ region was purified and ligated to pINGlδOO that was cut with SstI, treated with T4 polymerase, and cut with Xhol (Figure 11B) along with a HindlH to Xhol restriction fragment containing the human C _κ plus lδ nucleotides of 3' genomic DNA from p3Q2 (Figure 11C). The resulting plasmid that contained a pelB::Chimer2 L chain fusion was sequenced to determine that the proper in-frame fusion was formed. This plasmid was called pSWIO-B.

3. Preparation of Fd for Bacterial Expression

The intact Chimer2 chimeric Fd gene containing a BsmI restriction site at the signal sequence processing site and a BstEII restriction site in the J region in pSW2 served as the starting point for bacterial expression. The plasmid pSW2 was cut with j3sml, treated with T4 polymerase, and digested with BstEII (Figure 11D). The approximately 336 bp fragment containing the Fd V region was purified and ligated to pINGlδOO that was cut with SstI, treated with T4 polymerase, and cut with Xhol (Figure HE) along with the human CHI region (which had been previously generated by introducing a stop codon in hinge, Robinson, R.R., et al., supra) from pF3D

ITUTE SHEET

(Figure 11F). The resulting plasmid that contained a βelB::Chimer2 Fd fusion was sequenced to determine that the proper in-frame fusion was formed. This plasmid was called pSW4-B.

4. Multicistronic Expression System for L Chain and Fd Gene For production of bacterially derived Fab, both L chain and Fd need to be produced simultaneously within the cell. Using the plasmids constructed with each of these genes separately, a series of expression vectors were constructed that contain both genes aligned so that transcription from a single promoter will specify both genes. This was done in a way that minimized the noncoding DNA between the two genes. Each gene has a ribosome binding site needed for translation initiation and the identical DNA sequence from -48 to the pelB leader::antibody gene junction.

Plasmid pSWIO-B was cut with Sj , treated with T4 polymerase, cut with EcoRI, and the vector fragment was purified (Figure 11G). Similarly, pSW4-B was cut with Xhol, treated with T4 polymerase, cut with EcoRI and the fragment containing the Fd gene was purified (Figure 11H). These two purified DNA fragments were ligated to produce pMBl, which contained the two Chimer2 gene fusions linked in close proximity. The two gene cistron was placed under the control of the araB promoter in pING3107. Plasmid pMBl was cut with SphI, treated with T4 polymerase, cut with Xhgl _> and the fragment containing the Fd and K genes was purified (Figure 111). This DNA fragment was ligated to the vector fragment from pING3107 that had been cut with EcoRI, treated with T4 polymerase, and cut with Xhol (Figure 11J), generating pING3211. This vector contains all the necessary features for expression of Chimer2 Fab in E. coli.

SUBSTITUTE SHEET

5. Production of Chimer2 Fab in Bacteria

Expression of Chimer2 Fab from pING3211 in E. coli is under the inducible control of the araB promoter. Upon arabinose induction, Fab secreted into the growth medium increases more than 10-fold. The E. coli strain harboring pING3211 was cultured at 32 ^β C in 10L of minimal medium, supplemented with 1.7% glycerol as the carbon source. Dissolved oxygen was maintained at 20% of saturation by adition of a 5X concentrate of glycerol-supplemented minimal medium. The culture was induced at ODsø _Q δO by addition of a δX concentrate of glycerol-supplemented minimal medium containing 20g arabinose in 3L of concentrate. The induction solution was added over a period of about 28 hours, then δX concentrate without minimal medium was added for about 4 hours. Fab was detected in the fermentation broth by ELISA.

About 7 liters of a culture supernatant were concentrated, analyzed, and stored as in Example 2.

The Chimer2 Fab purified from E. coli has identical molecular weight properties as other Fab molecules produced from E. coli, as assessed by SDS gel electrophoresis. The bacterially-produced Chimer2 Fab is correctly assembled as a K plus Fd chain di er because of its positive reaction in the enzyme immunoassays detecting molecules with both K and Fd determinants, and because it reacts with core protein on commercially available test strips.

6. Study Performed on Chimer2 Fab

Chimer2 Fab was tested along with 2G12 mouse antibody and Chimer2 antibody for binding to antigen fixed to commercially available HIV Western test strips (Dupont or equivalent). Chimer2 Fab was incubated with the test strip, followed by interaction with goat anti-human Fab antibody, and then with rabbit anti-goat IgG. Detection of bound antibody was with Protein A-gold (BioRad) followed by silver enhancement. Chimer2 Fab specifically recognized the same size antigen on

SUBSTITUTE SHEET

the test strip as did Chimer2 whole antibody, specifically recognizing antigen of about 24 kD.

EXAMPLE 6

A Chimeric Mouse-Human Fab Fragment Produced in Yeast and Specific for an HIV gag Protein

Yeast cells are capable of expressing and secreting foreign proteins. In this example, yeast serve as a host for the production of mouse-human chimeric Fab. Yeast strains and growth conditions are as in Example 3.

1. Construction of Yeast Expression Plasmids Containing Antibody Genes

The gene sequences encoding the mature form of the VL region of 2G12 and containing a HindlH site in the J region (as described in Example 4) and a BsmI site introduced at the signal sequence processing site was fused to the human C _κ region. Plasmid pSW8 was cut with BsmI, treated with T4 polymerase, and cut with HindlH. The V region fragment was ligated along with the human C[_ region prepared from p3Q2 as a HindlH to Xhol fragment into pING1149. The plasmid pING1149 contains the gene sequence encoding the yeast invertase signal sequence (Taussig, R., et al . , supra) under control of the yeast PGK promoter (Hitzeman, R.A., et al., supra) . The resultant plasmid pSWll-Y (similar to Figure 6A) contains the PGK promoter (P) along with the invertase signal sequence (S) fused to the Chimer2 coding sequence. As the result of this fusion, the gene sequence encoding the mature form of the Chimer2 L chain was fused in frame to the gene sequence encoding the yeast invertase signal sequence. The PGK promoter - invertase signal sequence - chimeric L chain (V,C _K ) fusion was cloned into a partial 2 micron circle (2μ), ura3 yeast expression vector containing the PGK polyadenylation signal (Tm) to generate pSW13-Y (similar to Figure 6C). The gene sequence encoding the mature form of the V|_| region of

SUBSTITUTE SHEET

Chimer2 containing a BstEII site in the J region (as described in Example 4) and a BsmI site introduced at the signal sequence processing site from pSW2 was fused along with the human C^l region (which had been previously generated by introducing a stop codon in hinge, Robinson, R.R., et al . , supra) from pF3D into pING1149 that had been cut with Pst, treated with T4 polymerase and cut with Xhol. This generated plasmid pSWδ-Y (similar to Figure 6B). The PGK promoter - invertase signal sequence - chimeric Fd chain (V,CH1) fusion was cloned into a partial 2 micron circle (2μ), ura3 yeast expression vector containing the PGK polyadenylation signal (Tm) to generate pSW12-Y (similar to Figure 6D).

A single yeast expression vector containing both the chimeric L chain and Fd chain genes and their respective expression signals was constructed from pSW13-Y and pSW12-Y. This final vector, pING3208 (similar to Figure 6E), contains a portion of 2 micron circle (oriY, REP3) and the two selectable markers leu2d and ura3.

2. Yeast Secretion of Chimer2 Fab

The plasmid pING3208 was transformed into S. cerevisiae PS6 and the transformants were grown in broth under selective conditions as described above. The culture supernatants were assayed by ELISA and contained Fab levels of approximately 100 ng/ml . This material can be purified from fermentation broth and is expected to have properties identical to those of Fab produced as described in Example 3; yeast Chimer2 Fab is expected to have the same binding specificity as Fab produced in E. coli.

TE SHEET

EXAMPLE 7

A Chimeric Mouse-Human Immunoglobulin (Chimer4)

Produced in Mammalian Cel s and Specific for the HIV Reverse Transcriptase Protein

The mouse mAb 1C11 (developed by and commercially available from Epitope Inc., Beaverton OR) was derived from a hybridoma fusion utilizing spleen cells from a BALB/c mouse immunized with an HIV viral lysate. The Sp2/0 myeloma line was used as the fusion partner. Clone 1C11 produced immunoglobulin of the IgM subclass. MAb 1C11 was reactive with the cloned ggl gene product by Western blot analysis. In addition, 1C11 was capable of immunofluorescent staining of HIV-infected cells. Western blot analysis of the 1C11 mAb against viral lysates demonstrated predominant bands developed at δl and 6δ kD.

1. RNA Purification and cDNA Library Construction

One liter of 1C11 hybridoma cells at approximately 1 x 10^ cells/ml was collected by centrifugation and washed in 100 ml of PBS. RNA was prepared and a cDNA library was constructed in pBR322 as described in Example 1. cDNA libraries were screened by hybridization (Maniatis, T., supra) with ³² P-labeled; nick translated DNA fragments, i.e., for kappa clones with a mouse C _κ region probe and for H chain clones with a mouse IgM C region probe.

The VL and Vμ region fragments from the full length cDNA clones, pC4K-16 and pC4M-8 respectively, were inserted into M13 bacteriophage vectors for nucleotide sequence analysis. The nucleotide sequences of the V region of these clones were determined (Figures 12 and 13) by the dideoxy chain termination method. These sequences predict V region amino acid compositions that agree well with the observed compositions, and predict peptide sequences which have been verified by direct amino acid sequencing of portions of the V regions.

S TUT HEE

The nucleotide sequences of the cDNA clones show that they are immunoglobulin V region clones as they contain amino acid residues diagnostic of V domains (Kabat, et al .. supra).

2. Construction of Chimer4 H and L Chain Expression Plasmid The cDNA clone containing the 1C11 H chain, pC4M-8, was adapted for mammalian expression by a convenient cloning strategy. The DNA sequence of the 1C11 H chain gene is very similar to the previously cloned and chimerized anti-cancer AlO antibody gene. The region from at least 45 bp upstream of the initiation codon ATG, through the antibody leader sequence and to a BamHI restriction site in the V region at amino acids 16 and 17 contains identical DNA sequence. Previously, a Sal I restriction site was introduced into the AlO H chain cDNA sequence upstream of the ATG initiation codon by s i t e - d i re cted mutagenesi s with the primer 5ΑTGTCTGTGTCGACCACTGAAGAGA-3' .

The 1C11 H chain contains the J^3 sequence which contains a £stl site at its 3' end. The HindlH site (amino acids 3 and 4) and £stl site (J region) occurring in pC4M-8 (underlined and bold in Figure 12) were used to mobilize the the unique Chimer4 V region sequences adjacent to a Sail to HindlH DNA fragment from pTK7, which shared the δ' untranslated region, leader sequence, and the shared V region amino acid sequence. The resulting plasmid, pYZ124 contained the complete Chimer4 V|_| region, leader sequence and accompanying δ' untranslated sequence. The Sa.1 to PstI fragment from pYZ124 was cloned into the expression vector PING2227 to generate the Chimer4 H chain vector pING22δδ (Figure 14).

The cDNA clone containing the 1C11 L chain, pC4K-16, was adapted for mammalian expression taking advantage of site directed mutagenesis using the polymerase chain reaction (Horton, R.M., et al., 1989, Gene 77:611. A DNA fragment was amplified, using pC4K-16 as a substrate, which contained a

Sail restriction site upstream of the initiation codon ATG and a Bglll site in the J region with the oligonucleotides δ ' - G G C C G T C GA CTCA CCTGGACATGAT- 3 ' and 5'-AGCGCAGATCTCCAGCTTGGTGCC-3'. The DNA fragment was cut with Sail and Bglll and cloned into pING2016Egpt (Robinson, R.R., et al . , supra) to generate an in-frame V-J-C _κ chimeric gene fusion. Subsequent cloning to generate the final L chain vector, pING22S8 is as described in Figure lδ.

3. Stable Transfection of Mouse Lvmphoid Cells for the Production of Chimeric Antibody

The cell line Sp2/0 (American Type Culture Collection #CRLlδ81) was grown, transfected, and selected as described in Example 1, above.

The Chimer4 L chain expression plasmid, pING22δ8, and the H chain expression plasmid, pING2255, were each linearized by digestion with Pyul restriction endonuclease and co- transfected into Sp2/0 cells, giving mycophenolic acid resistant clones which were screened for L + H chain synthesis and G418 resistance. The MPA- and G418-resistant clone producing the most L + H chain, Sp2/0-22δ8 + 22δδ 11F3.1D10, was used for production of Chimer4 for testing.

4. Purification and Testing of Chimer4 Antibody Secreted in Tissue Culture

Sp2/0-22δ8 + 22δδ 11F3.1D10 cells were grown in culture medium DMEM (GIBCO) supplemented with δ% fetal bovine serum,

HEPES buffer, glutamine, penicillin, and streptomycin (Irvine

Scientific, Irvine, CA). Chimer4 was purified in a similar manner to that described for Chimerl and Chimer2 in Examples 1 and 4 respectively. Purified Chimer4 antibody was tested to demonstrate that it retained the antigen binding characteristics of the mouse 1C11 antibody using commercially available HIV Western strips (Dupont or equivalent), as described above for Chimerl and Chimer2. Chimer4 and mouse

1C11 antibodies both recognized the same HIV reverse transcriptase antigen.

EXAMPLE 8

A Chimeric Mouse-Human Fab Fragment Produced in E. coli and Specific for HIV Reverse Transcriptase Protein

Bacteria are suited for production of chimeric antibodies expressed from mammalian cDNA since entire coding sequences can be expressed from well characterized promoters. The following sections detail the secretion of Chimer4 Fab from . . coli .

1. Preparation of Fd for Bacterial Expression

Since the Chimer4 H chain V region is homologous with the AlO H chain, use was made of the pelB::A10 H chain fusion in pTK15 for construction of a pelB: :Chimer4 Fd gene fusion. The V region (containing regions of sequence difference between AlO and Chimer4) were removed from pTK15 by cutting it with BamHI and Xhgl and purifying the vector. Sequences of interest remaining in pTKlδ include the pelB leader joined to the first 16 amino acids of the AlO first framework region (FR1). The V-J region (containing 10 bp of CHI) from pING22δδ was purified as a BamHI to Apal fragment and the CHI fragment (which had been previously generated by introducing a stop codon in hinge, Robinson, R.R., et al . , supra) from pQ16D was purified as an Ajjal to Xhgl fragment. A three-way ligation of these fragments generated pYZ117, Figure 16.

2. Preparation of L Chain for Bacterial Expression

The Chimer4 V|_ region was adapted from pC4K-16 for bacterial expression as shown in Figure 16. Two primers were used to obtain an oligonucleotide by polymerase chain reaction, δ ' -GATATCCAGATGACACAGACTACATCC-3' and 5'-AGCGCAGATCTCCAGCTTGGTGCC-3' which generated a DNA fragment

SHEET

that was blunt-ended at one end and contained a Bglll site in the J region at the other end. The DNA fragment was then cloned into pINGlδOO that was digested with SstI, treated with T4 polymerase and cut with BamHI, generating pING3314. Plasmid pING3314 contains the pelB leader fused in-frame with the Chimer4 V _κ -J _κ region.

3. Multicistronic Expression System for L Chain and Fd Gene For production of bacterially derived Fab, both L chain and Fd need to be produced simultaneously within the cell. Using the plasmids constructed with each of these genes separately, a series of expression vectors were constructed that contain both genes aligned so that transcription from a single promoter will specify both genes. This was done in a way that minimized the noncoding DNA between the two genes. Each gene has a ribosome binding site needed for translation initiation and the identical DNA sequence from -48 to the pelB leader: :antibody gene junction.

A two-gene vector containing the entire pelB: :Fd gene and the pe!B-V _κ -J _κ region was constructed by cutting pING3314 with SphI, treating it with T4 polymerase and cutting it with EcoRI and cloning into it the Fd gene from pYZ117 as an Xhgl _. T4 polymerase treated, EcoRI cut fragment, generating pING331δ.

The Fd and the K gene fusion from pING331δ was placed under the control of the araB promoter in pING3303. Plasmid pING331δ was cut with Sβjil, treated with T4 polymerase and cut with Xhgl. and the fragment containing the Fd and K genes was purified. The DNA fragment was ligated into the vector fragment of pING3303 that had been cut with PstI, treated with T4 polymerase and cut with Xhol, generating pING3316. The final two-gene expression vector, pING340δ, was constructed in a three way ligation with two fragments from pING3316 [Apal to Xhol (vector) and the Apal to HindlH (Fd/ ) and the HindlH to Xhgl fragment of pYZ12δ. [pYZ12δ contains the V _κ

SUBSTITUTE SHEET

region of Chimer4 fused to the C _κ region derived from pING2016Egpt]. These steps are outlined in Figure 16.

4. Production and Testing of Chimer4 Fab in Bacteria

Chimer4 Fab is produced and purified as described in Examples 2 and 5 for Chimerl and Chimer2 Fab. The purified Chimer4 Fab is tested for antigen binding as described in Examples 2 and δ. The Chimer4 Fab is expected to have the same binding specificity as the Fab fragment of the mouse 1C11 antibody.

EXAMPLE 9

A Chimeric Mouse-Human Fab Fragment Produced in Yeast and Specific for HIV Reverse Transcriptase Protein

Yeast cells are capable of expressing and secreting foreign proteins. In this example, yeast serve as a host for the production of mouse-human chimeric Fab. This Fab molecule is identical to the Fab produced in E. coli outlined in Example 8, except that a single amino acid change occurred at position 23 (from an isoleucine to a leucine). This was an unexpected result of DNA polymerization during site directed mutagenesis. This amino acid change is not expected to alter antigen binding ability. Yeast strains and growth conditions are as in Example 3.

1. Construction of Yeast Expression Plasmids Containing Antibody Genes

The gene sequences encoding the mature form of the V _| _ region of 1C11 was adapted for yeast expression as follows.

Sequential site directed mutagenesis of an M13 subclone of p C 4 K - 16 w i th t he J reg i o n p r i me rs δ'-CATCAGCCCGTTAGATCTCCAGCTTGG-3' and the leader mature primer δ'-CATCTGGATATCTGCAGTGGTACCTTGAA-3' generated pYZ122, which contained a PstI site at the leader mature junction and a

BgUI site in the J region. Plasmid pYZ122 was cut with PstI, treated with T4 polymerase and cut with HindlH (a site located in the V region; see Figure 13). The V region fragment was ligated along with the VJCK region prepared from pYZ125 (see Example 8) as a HindlH to Xhgl fragment into pING1149. The plasmid pING1149 contains the gene sequence encoding the yeast invertase signal sequence (Taussig, R. et al., supra) under control of the yeast PGK promoter (Hitzeman, R.A., et al . , supra) .

The resultant plasmid pING3157 (similar to Figure 6A) contains the PGK promoter (P), along with the invertase signal sequence (S), fused to the Chimer4 coding sequence. As the result of this fusion, the gene sequence encoding the mature form of the Chimer4 L chain was fused in-frame to the gene sequence encoding the yeast invertase signal sequence. The PGK promoter - invertase signal sequence - chimeric L chain (V,C _K ) fusion was cloned into a partial 2 micron circle (2μ), ura3 yeast expression vector containing the PGK poly¬ adenylation signal (Tm) to generate pING31δ8 (similar to Figure 6C) .

To construct an Fd yeast vector, advantage was taken of the similarity of DNA sequence between the lCU H chain and the AlO antibody as discussed in Examples 7 and 8. The mammalian H chain vector, pING222δ was cut with BamHI, incubated with calf intestinal alkaline phosphatase (CIAP), and cut with Ap_al. The plasmid pTK14, which contains the PGK promoter - invertase signal sequence-chi eric AlO Fd chain fusion was cut with BamHI, treated with CIAP and cut with Apal; in a separate reaction, pTK14 was cut with BamHI. The vector fragment from the former reaction and the PGK-invertase-V-region fragment from the latter were ligated to the BamHI to Agal V-J region fragment from pING22δδ to generate pYZ116. The PGK promoter - invertase signal sequence - chimeric Fd chain (V-J-Cμl) fusion (similar to Figure 6B) was cloned into a partial 2 micron circle (2μ), ura3 yeast

SUBSTITUTE SHEET

expression vector containing the PGK polyadenylation signal (Tm) to generate pYZllδ (similar to Figure 6D) .

A single yeast expression vector containing both the chimeric L chain and Fd chain genes and their respective expression signals was constructed from pYZ118 and pING3158. This final vector, pING31δ9 (similar to Figure 6E), contains a portion of 2 micron circle (oriY, REP3) and the two selectable markers leu2d and ura3.

2. Yeast Secretion of Chimer4 Fab

The plasmid pING31δ9 is transformed into S. cerevisiae PS6 and the transformants are grown in broth under selective conditions. The culture supernatants contain Fab. This material is purified from fermentation broth and is expected to have properties identical to those of Fab produced as described in Example 8; yeast Chimer4 Fab is expected to have the same binding specificity as Fab produced in E. coli.

EXAMPLE 10

A Chimeric Mouse-Human Immunoglobulin (Chimerδ) Produced in Mammalian Cells and Specific for an HIV gag Protein

The mouse mAb 4D12 (described as anti-pl8 mAb in Yoshihara, P. et al . , Proc. 4th Int'l. Conf. on AIDS, June, 1988, p237) was derived from a hybridoma fusion utilizing spleen cells from a BALB/c mouse immunized with purified gag gene product. The Sp2/0 myeloma line was used as the fusion partner. Clone 4D12 produces immunoglobulin of the IgGl subclass. MAb 4D12 is reactive against the cloned gag gene product by ELISA analysis. In addition, 4D12 is capable of im unofluorescent staining of HIV-infected cells. Western blot analysis of the 4D12 mAb against viral lysates demonstrates predominant bands developed at 55, 45, and 39 kD. An additional weak band is developed at 18 kD depending upon the lysate used.

1. RNA Purification and cDNA Library Construction

One liter of 4D12 hybridoma cells at approximately 1 x 10^ cells/ml were collected by centrifugation and washed in 100 ml of PBS. RNA was prepared, a cDNA library was constructed in pBR322, and L and H chain cDNA clones were identified as described in Example I. The L and Vμ region fragments from the full length cDNA clones, pC5K-4 and pC5G-30 respectively, were inserted into M13 bacteriophage vectors for nucleotide sequence analysis. The complete nucleotide sequences of the V region of these clones were determined (Figures 17 and 18) by the dideoxy chain termination method. These sequences predict V region amino acid compositions that agree well with the observed compositions, and predict peptide sequences which have been verified by direct amino acid sequencing of portions of the V regions.

The nucleotide sequences of the cDNA clones show that they are immunoglobulin V region clones as they contain amino acid residues diagnostic of V domains (Kabat, et al . , supra).

2. Construction of Chimerδ H and L Chain Expression Plasmid The cDNA clone containing the 4D12 H chain, pCδG-30, was adapted for mammalian expression by introducing convenient restriction endonucleases sites by convenient cloning and site directed mutagenesis (Kramer, W., et al. , supra) of appropriate M13 subclones, Figure 19. Oligonucleotides were synthesized on a Cyclone DNA synthesizer, New Brunswick Scientific Co., and purified by acrylamide gel electro- phoresis. A Bell site, Figure 17, located approximately 4δ bp upstream of the initiation codon ATG was used for introduction of a Sail restriction site δ' of the antibody V region. The J region mutagenesis primer δ'-GAGACGGTGACCGAGGTTCCT-3' was used to insert a BstEII site into the M13 subclone pING3122. The restriction fragment containing the 4D12 V _| _ _| region bounded by Saj.1 and BstEII was then cloned into the expression vector

SUBSTITUTE HE

- 63 - pING2240 (pING2240 is identical to pING2227 except it contains a different antibody V region).

The cDNA clone containing the 4D12 L chain, pCδK-4, was adapted for mammalian expression in a similar way, Figure 20. The J region mutagenesis primer δ'-CAGCTCAAGCTTGGTCCC-3' used to insert a HindlH site into the M13 subclone pING3123. A Bell site, Figure 18, located approximately 30 bp upstream of the initiation codon ATG was used to introduce a Sail restriction site into pING312δ upstream of the initiation codon ATG. The restriction fragment containing the 4D12 L chain V region bounded by Sail and HindlH in pING3130 was then cloned into the expression vector pING1712 (pING2216 is identical to pING1712 except it contains a different antibody V region).

3. Stable Transfection of Mouse Lvmphoid Cells for the Production of Chimeric Antibody

The Chimerδ L chain expression plasmid, pING3132, and the

H chain expression plasmid, pING3126, were each linearized by digestion with Pyul restriction endonuclease and co- transfected into Sp2/0 cells, giving mycophenolic acid resistant clones which were screened for L + H chain synthesis and G418 resistance. The MPA- and G418-resistant clone producing the most L + H chain, Sp2/0-3132 + 3126 1G4, was used for production of Chimerδ for testing.

4. Purification and Testing of Chimerδ Antibody Secreted in Tissue Culture

Sp2/0-3132 + 3126 1G4 cells were grown in culture medium

DMEM (GIBCO) supplemented with 5% fetal bovine serum, HEPES buffer, glutamine, penicillin, and streptomycin (Irvine

Scientific, Irvine, CA) . Chimerδ was purified in a similar manner to that described for Chimer4 in Example 7. Purified

Chimerδ was tested to demonstrate that it retained the antigen binding characteristics of the mouse 4D12 antibody using commercially available HIV Western strips (Dupont or

ET

equivalent) as described for Chimerl, Chimer2, and Chimer4, above. Both the Chimerδ and mouse 4D12 antibodies recognized the same HIV gag antigen.

EXAMPLE 11

A Chimeric Mouse-Human Fab Produced in E. coli and Specific for an HIV gag Protein

Bacteria are suited for production of chimeric antibodies expressed from mammalian cDNA since entire coding sequences can be expressed from well characterized promoters. The following sections detail the secretion of Chimerδ Fab from E. coli.

1. In Vitro Mutagenesis

Site directed in vitro mutagenesis was performed as described by Kramer, et aj_. , supra, to introduce a PstI restriction site into the 4D12 L chain cDNA sequence (Figure 18) at the junction of the leader peptide and mature coding regi on wi th the ol igonucl eotide primer δ'-TCATCACAACATCTGCAGTGGTTTCCCGA-3' .

A Ajjal site was similarly introduced at the junction of the leader peptide and mature coding region of the 4D12 H chai n with the ol igonucl eotide primer δ'-TGGACCTGGGCCCGAACACCTGC-3' (Figure 17).

2. Preparation of L Chain for Bacterial Expression

The Chimerδ L chain V region containing a PstI restriction site at the signal sequence processing site and a unique HindlH site in the J region of pING3128 served as the starting point for bacterial expression. The plasmid pING3128 was cut with PstI, treated with T4 polymerase, and digested with HindlH (Figure 21A) . The 320 bp fragment containing the V|_ region was purified and ligated to pINGlδOO that was cut with SstI, treated with T4 polymerase, and cut with Xhol

(Figure 21B) along with a HindlH to Xhgl restriction fragment containing the human C _κ plus 15 nucleotides of 3' genomic DNA from pING3102 (Figure 21C). The resulting plasmid that contained a pelB: :Chimer5 L chain fusion was sequenced to determine that the proper in-frame fusion was formed. This plasmid was called pING3133.

3. Preparation of Fd for Bacterial Expression

The intact Chimerδ chimeric Fd gene containing a Apal restriction site at the signal sequence processing site and a BstEII restriction site in the J region in pING3129 served as the starting point for bacterial expression. The plasmid pING3129 was cut with Agal, treated with T4 polymerase, and digested with BstEII (Figure 21D). The approximately 33δ bp fragment containing the Fd V region was purified and ligated to pINGlδOO that was cut with SstI , treated with T4 polymerase, and cut with Xhgl (Figure 21E) along with the human CHI region (which had been previously generated by introducing a stop codon in hinge, Robinson, R.R., et al., supra) from pF3D (Figure 21F). The resulting plasmid that contained a pelB: :Chimerδ Fd fusion was sequenced to determine that the proper in-frame fusion was formed. This plasmid was called pING3131.

4. Multicistronic Expression System for L Chain and Fd Gene For production of bacterially derived Fab, both L chain and Fd need to be produced simultaneously within the cell. Using the plasmids constructed with each of these genes separately, a series of expression vectors were constructed that contain both genes aligned so that transcription from a single promoter will specify both genes. This was done in a way that minimized the noncoding DNA between the two genes. Each gene has a ribosome binding site needed for translation initiation and the identical DNA sequence from -48 to the pelB leader: :antibody gene junction.

TITUTE SHEET

- δ6 -

Plasmid pING3133 was cut with Sghl, treated with T4 polymerase, cut with EcoRI, and the vector fragment was purified (Figure 21G). Similarly, pING3131 was cut with Xhol, treated with T4 polymerase, cut with EcoRI and the fragment containing the Fd gene was purified (Figure 21H). These two purified DNA fragments were ligated to produce pING3136, which contained the two Chimerδ gene fusions linked in close proximity. The two gene cistron was placed under the control of the araB promoter in pING3107. Plasmid pING3136 was cut with Sgiil, treated with T4 polymerase, cut with Xhgl, and the fragment containing the Fd and K genes was purified (Figure 211). This DNA fragment was ligated to the vector fragment from pING3107 that had been cut with EcoRI, treated with T4 polymerase, and cut with Xhgl (Figure 21J), generating pING3139. This vector contains all the necessary features for expression of Chimerδ Fab in E. coli.

5. Production of Chimerδ Fab in Bacteria

Expression of Chimerδ Fab from pING3139 in E. coli is under the inducible control of the araB promoter. Upon arabinose induction, Fab secreted into the growth medium increased more than 10-fold. The E. coli strain harboring pING3139 was was cultured at 32 ^β C in 10L of minimal medium, supplemented with 1.7% glycerol as the carbon source. Dissolved oxygen was maintained at 20% of saturation by adition of a δX concentrate of glycerol-supplemented minimal medium. The culture was induced at OD5øø=δ0 by addition of a δX concentrate of glycerol-supplemented minimal medium containing δg arabinose in 3L of concentrate. The induction solution was added over a period of about 20 hours, then δX concentrate without minimal medium was added for about 4 hours. Fab was detected in the fermentation broth by ELISA.

About 7 liters of a culture supernatant was concentrated, analyzed, and stored as in Example 2.

SUBSTITUTE SHEET

The Chimerδ Fab purified from E. coli had identical molecular weight properties as other Fab molecules produced from E. coli, as assessed by SDS gel electrophoresis. The bacterially-produced Chimerδ Fab was correctly assembled as a K plus Fd chain dimer because of its positive reaction in the enzyme immunoassays detecting molecules with both / and Fd determinants, and because it reacted with core protein on commercially available test strips.

6. Study Performed on Chimerδ Fab

Chimerδ Fab was tested along with 4D12 mouse antibody and Chimerδ antibody for binding to antigen fixed to commercially available HIV Western test strips (Dupont or equivalent). Chimerδ Fab was incubated with the test strip, followed by interaction with goat anti-human Fab antibody, and then with rabbit anti-goat IgG. Detection of bound antibody was with Protein A-gold (BioRad) followed by silver enhancement. Chimerδ Fab specifically recognized the same size antigens of 55, 45, and 39 kD on the test strip as did Chimerδ whole antibody.

EXAMPLE 12

A Chimeric Mouse-Human Fab Fragment Produced in Yeast and Specific for an HIV gag Protein

Yeast cells are capable of expressing and secreting foreign proteins. In this example, yeast serve as a host for the production of mouse-human chimeric Fab. Yeast strains and growth conditions are as in Example 3.

1. Construction of Yeast Expression Plasmids Containing Antibody Genes

The gene sequences encoding the mature form of the V _| _ region of 4D12 and containing a HindlH site in the J region

(as described in Example 11) and a PstI site introduced at the signal sequence processing site was fused to the human C _κ

SUBSTITUTE SHEET

- 68 - region. Plasmid pING3128 was cut with PstI, treated with T4 polymerase, and cut with HindlH. The V region fragment was ligated along with the human C]_ region prepared from pING3102 as a HindlH to Xhgl fragment into pING1149. The plasmid pING1149 contains the gene sequence encoding the yeast invertase signal sequence (Taussig, R. et al . , supra) under control of the yeast PGK promoter, Hitzeman, R.A., et al . , supra) . The resultant plasmid pING3134, (similar to Figure 6A), contained the PGK promoter (P), along with the invertase signal sequence (S), fused to the Chimerδ coding sequence. As the result of this fusion, the gene sequence encoding the mature form of the Chimerδ L chain was fused in frame to the gene sequence encoding the yeast invertase signal sequence. The PGK promoter - invertase signal sequence - chimeric L chain (V,C _K ) fusion was cloned into a partial 2 micron circle (2μ), ura3 yeast expression vector containing the PGK polyadenylation signal (Tm) to generate pING3140 (similar to Figure 6C). The gene sequence encoding the mature form of the Vμ region of Chimerδ containing a BstEII site in the J region (as described in Example 11) and a Agal site introduced at the signal sequence processing site from pING3129 was fused along with the human C^l region (which had been previously generated by introducing a stop codon in hinge, Robinson, R.R., et al . , supra) from pF3D into pING1149 that had been cut with PstI, treated with T4 polymerase and cut with Xhgl. This generated plasmid pING3135. The PGK promoter - invertase signal sequence - chimeric Fd chain (V,C^1) fusion (similar to Figure 6B) was cloned into a partial 2 micron circle (2μ), ura3 yeast expression vector containing the PGK polyadenylation signal (Tm) to generate pING3141 (similar to Figure D).

A single yeast expression vector containing both the chimeric L chain and Fd chain genes and their respective expression signals was constructed from pING3140 and pING3141. This final vector, pING3143, similar to Figure 6E, contains a

SUBSTITUTE SHEET

portion of 2 micron circle (oriY, REP3) and the two selectable markers leu2d and ura3.

2. Yeast Secretion of Chimerδ Fab

The plasmid pING3143 is transformed into S. cerevisiae PS6 and the transformants are grown in broth under selective conditions as described above. Transformed yeast cultures secrete Chimer4 Fab into the culture medium, and this material is purified from fermentation broth and is expected to have properties identical to those of Fab produced as described in Example 3; yeast Chimerδ Fab is expected to have the same binding specificity as Fab produced in E. coli.

CONCLUSIONS

The examples presented above demonstrate a number of important qualities of the Chimerl, Chimer2, Chimer4, and Chimerδ antibodies and the genetically engineered fragments and derivatives of the invention. First, the chimeric antibodies, fragments and derivatives bind to the same antigens as the parent mouse mAbs. This is evidenced by direct binding of chimeric antibodies, fragments and derivatives to Western immunoblot test strips prepared with lysates of HIV.

To date there has been only very limited testing of antibody therapy for the treatment of HIV-infected individuals. Passive immunization of these patients with pooled high-titer human serum has resulted in some clinical improvement. Direct therapy with mouse mAbs is not likely to be feasible since mouse antibodies are known to elicit a strong immune response against themselves. The chimeric antibodies of this invention represent potentially important agents for treatment of HIV infection. Antibodies such as Chimerl, Chimer2, Chimer 4, and Chimer δ or their corresponding fragments and derivatives, either alone, as

E SHEET

immunoconjugates, or in combination with other agents may be advantageously used for HIV treatment.

The chimeric antibodies of this invention may also be used for any diagnostic purpose for which the similar mouse antibodies or their derivatives can be used. Furthermore, the chimeric antibodies, fragments and derivatives may be used as immobilized reagents for ex vivo adsorption of viruses, viral antigens, or virus-infected cells.

Chimeric antibody molecules, such as those of the present invention, and their derivatives, embody a combination of the advantageous characteristics of mouse and human mAbs. Like mouse mAbs, they can recognize and bind to viral antigens; however, unlike mouse mAbs, the "human specific" properties of chimeric antibodies lower the likelihood of an immune response to the antibodies, and result in prolonged survival in the circulation through reduced clearance. These properties have been observed when a chimeric antibody directed gainst a tumor marker was introduced in patients, LoBuglio, et al ., 1989, Proc. Natl. Acad. Sci. USA 86:4220-4224.

The human component of a chimeric antibody may enhance its ability to mediate target destruction, for example, of virally infected cells, in combination with human effector cells or complement. Moreover, using the methods disclosed in the present invention, any desired antibody isotype can be combined with any particular antigen combining site. This invention also enables the direct production of one or more domains of the antibody molecule in functionally active form.

SUBSTITUTE SHEET

DEPOSITS

Illustrative cell lines secreting Chimerl, Chimer2, Chimer4, and Chimerδ antibodies were deposited on October 25, 1989 at the ATCC, Rockville Maryland.

The E. coli strains deposited are as follows:

1. E. coli MC1061 transformed with pING3127 (G269) was granted ATCC Accession #68146.

2. E. coli MC1061 transformed with pING3211 (G270) was granted ATCC Accession #68147.

3. E. coli MC1061 transformed with pING3405 (G271) was granted ATCC Accession #68148.

4. E. coli MC1061 transformed with pING3139 (G272) was granted ATCC Accession #68149.

The transfected hypbrido a cell lines are as follows:

1. Sp/0 (pING3110 + pING3112) (C863) was granted ATCC Accession # HB10277.

2. Sp/0 (pING3005 + pING3004) (C850) was granted ATCC Accession # HB10276.

3. Sp/0 (pING2258 + pING22δδ) (C871) was granted ATCC Accession # HB10279.

4. Sp/0 (pING3132 + pING3126) (C864) was granted ATCC Accession # HB10278.

International Application No: PCT/

Form PCT/-IO-IJ (January tMI)

International Application No: PCT/

Form PCT/ROllM (J»nu»ιr 1HD

international Application No: PCT/

Form PCT/RO/1- (January 1M.)

International Application No: PCT/

Form PCT/F.O/1J4 (January 1M1)

International Application No: PCT/

International Application No: PCT/

International Application No: PCT/

Form PCT/RO'1-M [Jinui'T IMS)

International Application No: PCT/