WO1992012731A1

WO1992012731A1 - Glycosylated gag antigens and antibodies which react with live cells expressing those antigens

Info

Publication number: WO1992012731A1
Application number: PCT/US1992/000528
Authority: WO
Inventors: Pinter Abraham; Furong Shang; William J. Honnen; Kathy Revesz
Original assignee: Public Health Research Institute Of The City Of New York, Inc.
Priority date: 1991-01-22
Filing date: 1992-01-22
Publication date: 1992-08-06
Also published as: AU1250492A

Abstract

A glycoprotein expressed on the surface of HIV infected cells and which contains gag amino acid sequences has been isolated. Monoclonal antibodies to the glycoprotein have also been isolated and have a variety of uses.

Description

GLYCOSYLATED GAG ANTIGENS AND ANTIBODIES WHICH REACT WITH LIVE CELLS EXPRESSING THOSE ANTIGENS

Field of the Invention

This invention is in the field of HIV related antigens and certain antibodies to those antigens.

Background of the Invention

The gag gene of HIV encodes the major structural proteins of the viral core, and is expressed initially as a 55 d polyprotein (Pr55) which is myristoylated at its amino terminus. During the process of viral maturation, this is proteolytically cleaved by the virus-encoded protease into four separate domains; the matrix protein pl7, which corresponds to the amino terminal domain of Pr55 and which carries the myristic acid substitution, the capsid protein p24, and ribonucleoproteins p6 and p9. Occasionally, via a frame-shifting mechanism, a larger product consisting of both gag and pol sequences is formed, which is believed to be the precursor for the mature pol gene products, protease, reverse transcriptase, and integrase. The myristic acid moieties of both Pr55 and pl7 are believed to be associated with the lipid bilayer, thereby targeting the gag protein to the interior of the virus membrane. This is consistent with immunoelectron microscopy and molecular modeling studies, which support the association of pl7 with the inner surface of viral membranes. There have been several reports that antisera and monoclonal antibodies against pl7 can neutralize viral infectivity, suggesting that at least some epitopes of pl7 may be exposed on the surface of the intact virion. There is, however, no evidence for the expression of gag proteins on the surfaces of virions in other retrovirus systems. In some retroviruses, an alternative pathway exists for the expression of the gag gene which results in the formation of a large, membrane- associated glycoprotein. For the murine leukemia virus, these molecules exist in the form of two cell-surface glycoprotein of approximately

80,000-90,000 daltons, which are secreted from infected cells as two proteolytic fragments of 55 and 40 kd. The cell surface forms of these molecules synthesized by endogenous HuLVs have been shown to correspond to the Gross cell surface antigen (GCSA) , which functions as an efficient target for antibody- and cell-mediated cytotoxicity reactions. The mechanism of expression of the glycosylated proteins of MuLV has been shown to involve initiation of translation of the gag gene at a CUG initiation codon, located 264 nucleotides upstream of the normal gag initiation site. This results in the introduction of a novel N-terminal domain which includes several hydrophobic regions which may function as leader sequences for membrane insertion. Analyses of deletion mutants of MuLV have shown that this molecule is not essential for replication of the virus in tissue culture, but may facilitate infection. Until now, no membrane associated glycoprotein which contains gag encoded sequences has been known to exist for HIV.

SUMMARY OF THE INVENTION

Monoclonal antibodies have been discovered which are specific for the HIV-1 matrix protein, P17³⁵³, and which also react with glycoproteins expressed on the surface of live, HIV infected cells. We have also isolated the glycoproteins which are expressed on the surface of HIV infected cells and which are immunologically cross reactive with pl7. One such glycoprotein has a molecular weight of about 150,000 as measured by SDS-PAGE. Another has a molecular weight of about 90,000 as measured by SDS-PAGE. Fragments of these glycoproteins have also been isolated.

Monoclonal antibodies are included which are specific for a glycoprotein which is expressed on the surface of HIV infected cells, which glycoprotein contains gag sequences and has a molecular weight of about 150,000 as measured by SDS-PAGE, and which monoclonal antibodies react with live HIV infected cells. Also included in the invention are monoclonal antibodies whose immunological binding is competitively inhibited by novel monoclonal antibodies Gllgl or Gllh3.

Also included are monoclonal antibodies whose immunological binding is competitively inhibited by novel monoclonal antibodies D3b3 or B4F8.

Diagnostics and therapeutics using these novel materials are included as well. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 shows the results of a western blot analysis of monoclonal antibodies tested with HIV lysate. Fig. 2 shows the results of an experiment to show specificity of anti-pl7 antibodies for ³⁵cys- labeled viral proteins.

Fig. 3 shows the results of epitope mapping of monoclonal antibodies to pl7. Fig. 4 shows results of an experiment to show the detection of cell-surface staining of HIV-1-infected cells by mAbs against ll³-*-¹ by fluorescence activated cell sorting analysis.

Fig. 5 shows the results of immunofluorescent visualization of the binding of mAb Gllh3 to live (right panels) and acetone-fixed (left panels) HIV-infected cells (top panels) .

Fig. 6 shows the results of an experiment testing the competition of binding of biotinylated mAb Gllh3 to rpl7-coated plates.

Fig. 7 shows the results of an experiment carrying out radioimmunoprecipitation of ³Hglu- labeled lysates of HIV-infected H9 cells.

Fig. 8 shows further results of an experiment carrying out radioimmunoprecipitation of Hglu- labeled supernatants of HIV-infected H9 cells.

Fig. 9 shows the results of an experiment to analyze radioiodinated cell surface components recognized by anti-HIV antibodies. Fig. 10 shows the results of a study of the neutralization of HTLV-IIIb infection by mAbs tested.

Fig. 11 shows Table I, a listing of the reactivities of mAbs which were obtained. Fig. 12 shows Table II, a listing of the pl7 peptides used for epitope mapping of the mAbs which were obtained.

DETAILED DESCRIPTION OF THE INVENTION

The glycosylated gag antigens of the present invention are recognized by certain antibodies which are specific for the HIV gag proteins. In particular, they have been found to react preferentially with two novel monoclonal anti-pl7 antibodies. Those antibodies were found to stain live cells, and the presence of the glycoproteins recognized by these antibodies on the cell surface was confirmed, as shown in the experimental section below. This reaction of the novel anti-pl7 antibodies with live cells was unexpected, since available data indicates that pl7 is located on the interior of the infected cell's plasma membrane. The glycosylated gag antigens which have molecular weights of about 150,000 and about 90,000 are resolved as closely spaced doublets on SDS gels, suggesting that they exist in two slightly different forms. These antigens are distinct from env gene products.

In addition to being found on the cell surface, the glycoproteins of the invention are released in various sizes into supernatant medium from infected cells. Immunoprecipitation analysis shows that the glycosylated gag products are released from infected cells in a form similar in size to the full-length cell-associated molecules, as well as in smaller forms which may be proteolytic fragments. All substantially pure versions of the glycosylated product, whether found at the infected cell's surface or released from infected cells, are included in applicant's invention. Applicants do not wish to be bound by any theory of the mechanism for the generation of the glycosylated gag antigens. Nevertheless, while examination of the HIV gag gene does not identify an obvious sequence 5' to the gag initiation codon which can serve an analogous function to the 5' leader sequence of the glycosylated gag product of MuLV, such an initiation codon may exist. Whereas AUG is the predominant initiation codon in prokaryotic and eukaryotic systems, precedents exist for initiation at GUG, UUG, AUU, and ACG in E. coli, and ACG and UAG in eukaryotic systems in addition to the CUG initiation codon utilized for the MuLV glycosylated gag protein. Again, while not wishing to be bound by any theory of the invention, the possibility also exists that the generation of the glycosylated gag product of HIV requires a splicing event. Another feature which distinguishes these HIV molecules from surface gag encoded glycoproteins of MuLV is their large size. This may suggest that the HIV molecules contain significant regions of pol sequences in addition to the gag regions identified by immunological reactivities, or this may be a reflection of extensive glycosylation. The lack of recognition of these proteins by a hyperimmune serum prepared against recombinant reverse transcriptase may indicate that the pol region of the molecule is either absent or is modified, perhaps by glycosylation. Glycosylated gag antigens of HIV-1, HIV-2, and SIV (Simian Immunodeficiency Virus) are encompassed by the invention. One of the monoclonal antibodies which recognizes these proteins on the cell surface (Gllgl) of HIV-1 infected cells has been found to also react with such proteins on cells infected with both HIV-2 and SIV isolates. It can be assumed that those antibodies will react with the glycosylated gag antigens from those infected cells.

The glycosylated gag antigens may be obtained from well known HIV infected cell lines, such as H-9 and others by using procedures well known to one skilled in the art. The glycosylated gag proteins can be purified on an immunoaffinity chromatography column prepared with antibodies of the invention. Those antibodies can either be Gllgl or Gllh3 or other antibodies of the invention obtained by one skilled in the art. Other conventional methods may also be used to purify the glycosylated gag proteins as well, such as ion-exchange chromatography, lectin chromatography and sizing chromatography.

The invention includes monoclonal antibodies which react with the glycosylated gag antigen glycoprotein having a molecular weight of about 150,000 as measured by SDS-PAGE, which also reacts with live HIV infected cells. Whether a monoclonal antibody reacts with live HIV infected cells can be determined by fluorescence activated cell sorting. An antibody is considered to react, according to the invention, where at least about 50% of the infected cells are shown to react in the procedure. Two cell lines which produce mAbs according to the invention and described herein as Gllgl and Gllh3 have been deposited at the A.T.C.C. and have been assigned accession numbers HB10669 and HB10668 respectively. MAbs from these cell lines can be used for screening to identify cell lines producing mAbs specific for the same epitope, to purify glycosylated gag antigens, as well as for other uses, such as for assays, described herein. These cultures have been deposited to exemplify the invention, but one skilled in the art can obtain the monoclonal antibodies according to this invention independently.

For example, monoclonal antibodies of the invention can be obtained by immunizing Fisher rats with HIV viral lysate, as described below, fusing immune spleen cells with appropriate myeloma cell lines to obtain hybridoma cell clones » producing monoclonal antibodies to HIV, and screening those clones using ELISA assays for antibodies to HIV generally, and for antibodies to pl7. Commercial assays for antibodies to HIV are available using lysate as a capture antigen on a solid phase. Recombinant pl7 is available for use in a standard format ELISA to detect anti-pl7 antibodies. Further methods for determining which clones obtained produce antibodies of the invention are described below, including using Western blot to determine whether the antibodies are to pl7, and using fluorescent cell sorting techniques to determine which antibodies react with the live infected cell surface. Similarly, to determine whether cultures are producing mAbs to the same epitope as one of the antibodies of the invention, such as Gllgl or Gllh3, competition assays may be used to determine if the binding of Gllgl of Gllh3 to antigen is inhibited. MAbs which compete would be specific for the same or similar epitope. Competition assays are well known in the art.

Alternately, the animal can be immunized with infected cells, or with purified glycosylated gag antigen of the invention, and the antibodies obtained screened as described above.

Other animals may be similarly immunized and monoclonal antibodies of the invention obtained. Antibodies of the invention may also be obtained from humans infected with HIV. The antibodies of the invention may be advantageously used in an assay for determining the number of infected live cells in a sample, since the antigen detected is at the cell surface. Such an assay is useful to determine the course of HIV infection. Similar assays using antibodies to gpl20 antigens expressed at the infected cell's surface are less accurate because the antibodies to gpl20 bind to circulating gpl20 which may be bound to a CD-4 receptor of an uninfected cell, thereby erroneously increasing the count of infected cells. The glycoprotein discovered by us is not known to bind to any uninfected cell receptor, and this problem associated with using antibodies to gpl20 is avoided. Also, while antibodies to gpl20 tend to be specific for highly variable epitopes, antibodies Gllgl and Gllh3 discovered by applicants have been found to be specific for epitopes which are conserved among HIV strains. Antibody Gllgl was found to react with HIV-1 strains ARV-2, Illb, MN, and SF-2, as well as with LAV-2 and SIV-1. Antibody Gllh3 was found to react with Illb, MN, SF-2 and ARV-2, but not with LAV-2 and SIV-1. Antibodies Gllh3 and Gllgl were characterized as being of the type IgG_2a.

An assay for determining the presence of the antigen which the mAbs of the invention bind to can be performed using, for example, a sandwich format, well known in the art per se. A solid phase is coated with one antibody of the invention, such as Gllgl, the sample is added, and then biotin labeled mAb of the invention such as Gllh3 is added. Following a wash, enzyme labeled avidin would then be added as well as enzyme substrate. Numerous specific protocols for sandwich assays are well known in the art.

The materials of the invention may be used for other types of assays as well, such as competitive inhibition assays.

Again, while not wishing to be bound by any theory of the invention, the presence of glycosylated gag antigen may have significance in the natural development of antiviral antibodies, and protective immune responses in infected patients. Antibodies against both gag and pol gene products are produced in high titers early in infection, and evidence has been presented for a correlation between the loss of antibodies against both pl7 and p24 and progression to AIDS. It is possible that the cell surface molecules play a role in the generation of such antibodies. Furthermore, the glycosylated gag antigens may act as targets for such protective immune functions as antibody-dependent cellular cytotoxicity (ADCC) and antibody-dependent complement-mediated cytotoxicity (ACC) . MAbs against the glycosylated gag antigens may effectively be administered in order to mediate those functions.

The glycoprotein molecules discovered are especially important as targets for immunotoxin therapy, using the mAbs of the invention, since the molecules are expressed at the infected cell surface. Such a therapeutic agent may be prepared by covalent attachment of a toxin such as ricin A or pokeweed antiviral protein to the mAbs. It has been demonstrated that such anti-HIV-1 Abs-toxins (immunotoxins) are capable of specifically killing HIV-1 infected cells in vitro. An advantage of using a mAb of the invention over an antibody to env is that the antigen of the invention is not known to bind to uninfected cells, as are env antigens. Therefore the mAbs of this invention will not kill uninfected cells as will mAbs to env.

These mAbs can also be used to kill HIV-1 infected cells via protective immune response for which the mAbs act as a target. It is believed advantageous to administer the mAbs of the invention in conjunction with other anti-viral therapies, such as AZT, to slow the progress of HIV-1 induced disease.

The substantially pure glycosylated gag antigens of this invention can be used to prepare a therapeutic agent to protect against HIV-1. The effectiveness of such a therapeutic may be attributed to the stimulation of protective immune response initiated by antibodies produced in response to the antigen.

Antibodies Gllgl and Gllh3 are not known to have virus-neutralizing activity, but use of the antibodies of the invention to neutralize viruses cannot be ruled out.

The invention also includes moieties having the same function as monoclonal antibodies, such as Fab fragments, which bind the same epitope as the monoclonal antibodies of the invention.

Techniques for producing such fragments are known to one skilled in the art [Parham, P. (1986) In Cellular Immunology 4th Ed. (ed. D.M Weir) , Blackwell Scientific Publications, California]. Transformed cell lines producing recombinant monoclonal antibodies are included within the scope of the invention, as are the antibodies produced thereby. Such recombinant monoclonal antibodies may use the gene sequence encoding the antibodies discovered to produce chimeric antibodies having sequences of human antibodies. The invention also includes monoclonal antibodies produced in human cell lines, or otherwise. Such human antibodies can be produced from clones produced from sera of an infected patient using the screening techniques described above. The invention also includes kits for determining an antigen for which the mAbs of the invention are specific. For example such a kit may comprise a mAb of the invention, a solid phase on which is coated a mAb of the invention, and means for detecting the formation of a complex among the mAbs of the invention, and a glycosylated gag antigen found from a serum sample. The practice of such a sandwich type immunoassay is well known to one skilled in the art.

Monoclonal antibodies which were isolated by us, while not reacting with glycosylated gag antigen of live cells, are nevertheless important as antibodies to pl7. These antibodies, D3b3 and B4f8 are specific for conserved epitopes among pl7 antigens, and have high affinity for pl7. They are especially useful in immunoassays, though other uses will be apparent to those skilled in the art. D3b3 and B4f8 have been deposited at the A.T.C.C. and have been assigned accession numbers HB10670 and HB10671 respectively. Antibody D3b3 was found to react with peptide 7 shown in Table I. Antibody B4f8 was found to react with peptide 6 in Table I. Those two peptides can be used in order to immunize animals in order to produce mAbs having similar specificities to those of D3f3 and B4f8, as well as to screen clones produced thereby.

The present invention is further described herein below. These examples are for illustration and are not intended to limit the invention.

Examples

Example 1

Isolation and characterization of anti-p!7 monoclonal antibodies. Fisher rats were immunized subcutaneously with disrupted preparations of HTLV-IIIb (50 ug/ml) obtained from Organon Teknika prepared as emulsions with complete Freund's adjuvant, followed by boosters with an equivalent amount of virus in incomplete Freund's adjuvant at monthly intervals. Once a high titer of antibodies against HIV proteins were obtained, which required three or four immunizations, spleen cells of the immunized animals were fused with SP2/0 myeloma cells, or, in the case of mAb 10A-B4F8, with YB2/0, and supernatants of hybridoma clones were screened by ELISA on plates coated with the HIV lysate used as the immunogen. Positive wells were recloned to obtain stable monoclonal producing hybridoma lines.

Eight monoclonal antibodies against pl7 were isolated from two separate fusions of spleen cells of rats immunized with solubilized HTLV-IIIb lysate (Table I) . Screening was initially performed by ELISA assays against the same viral lysate used as the immunogen, and specificity was determined by Western blot assays, using strips prepared with lysates of the same strain of HIV-1. The results are shown in Fig. 1. The following samples were analyzed against HIV-1 viral lysate: A- human immune serum; B- mAb Gllh3; C- mAb Gllgl; D- mAb D3b3; E- mAb B4f8; F- mAb D12g2; G- mAb D7hll; H- mAb A8g2; I- mAb Hlc7; J- mAb E910 (anti-p24) ; K-control (no antibody) . The human immune serum recognized a number of viral proteins, including components of 160 and 120 kd, reverse transcriptase (p66 and p51) , the gag polyprotein (p55) , gp41, integrase (p31) , and gag proteins p24 and pl7. We have previously shown that the 160 and 120 kd bands consist mostly of gp41 trinters and tetramers. Eight of the monoclonal antibodies reacted strongly with the pl7 band, and these antibodies also recognized the gag precursor protein, p55. Several antibodies also reacted with a larger component of approximately 80-90 kd, which may represent a processing intermediate of the gag-pol polyprotein. One antibody recognized p24 (lane J) . No bands were seen in a control sample in which the first antibody was omitted.

The strain specificity of these antibodies was determined by immunofluorescence assays of the reactivity of the antibodies with cells infected by various strains of HIV. All of the antibodies reacted with the Illb SF-2 and MN strains of HIV-1. Antibody Gllgl also cross-reacted with cells infected with the ROD strain of HIV-2 as well as with the BK28 strain of SIV, and thus appears to recognize an interspecies-specific epitope.

Example 2 Specificity of anti-p!7 antibodies for ³⁵cys- labeled viral proteins.

The specificities of the various anti-sera used in these experiments were shown by radioimmunoprecipitation of viral lysates prepared from HTLV-IIIb-infected H9 cells which were labeled with ³⁵cys for 24 hours (Fig. 2) .

H9 cells infected with HTLV-IIIb were labeled with ³⁵cys 24 hour supernatants were adjusted to 0.5% NP-40 and 0.5M NaCl, and immunoprecipitated with the following antibodies: human immune serum (lane A) , rat hyperim une serum against recombinant pl7 (lane B) , mAb B4f8 (lane C) , mAb D3b3 (lane D) , mAb Gllgl (lane E) , mAb Gllh3 (lane F) , or rabbit anti-rat Ig serum (lane G) . Immunoprecipitates were collected with fixed staph A, and analyzed by SDS-PAGE. Since the rat monoclonal antibodies did not bind to protein A, rabbit anti-rat Ig serum (1:100 dilution) was added to these samples prior to incubation with the staph A.

The human immune serum recognized a number of viral proteins, including gpl20, p24, and pl7. The anti-pl7 antibodies reacted strongly with pl7, and also immunoprecipitated several larger bands of 40 kd and 55 kd, which were also recognized by the human serum. These apparently correspond to the gag precursor protein and its intermediate proteolytic products. A control immunoprecipi- tation with rabbit anti-rat Ig but with no rat monoclonal antibodies did not recognize any bands, demonstrating the specificity of the anti-sera tested.

Example 3

Epitope mapping of the monoclonal antibodies obtained to p!7. The epitope specificities of the monoclonal antibodies against pl7 were investigated by examining the abilities of these antibodies to bind to synthetic peptides corresponding to various regions of the pl7 sequence. ELISA assays were used to measure binding of the antibodies against plates coated with rpl7 and peptides 1-13 listed in Table II. Data is plotted as OD 405 for each peptide.

The ELISA assay was performed as follows. Titertek immunoassay plates (Flow-ICN Inc.) were coated with the series of overlapping peptides representing the complete HTLV-IIIb pl7 sequence. Solutions of the peptides at 1 ug/ml in 15 mM Na₂C0₃- 35 mM NaHC0₃, pH 9.6, were added to individual wells and incubated overnight at room temp. The wells were then blocked with 2% BSA and washed with PBS + 0.05% Tween 20. Antibodies at 1 ug/well in PBS were incubated with the peptide- coated wells for 1 hour at room temp., and after washing, the wells were incubated with alkaline phosphatase-conjugated rabbit anti-rat Ig for 30 min at room temp. Bound antibody was quantitated by adding p-nitrophenyl phosphate in diethanolamine buffer, and after incubation for 30 min at room temp. Absorbance was read at 405 nm on a Titertek Multiskan Plus ELISA reader. Peptides 1-13 were purchased from the Peptide Laboratory Inc., Berkeley CA.

In addition to those 13 peptides, an additional peptide, (labelled 9A in Table 2 and 86-115 in Fig. 3) which overlapped with regions of smaller peptides 9, 10, and 11 was also tested.

All of the antibodies bound strongly to rpl7. Antibody B4f8 bound to peptide 6, and antibody D3b3 reacted with peptide 7. Antibodies Gllgl and Gllh3 bound well to rpl7, but they did not bind to any of the peptides tested. A similar lack of reactivity for these two antibodies was also observed when a different set of larger peptides was tested (data not shown) . This suggests that they recognize a discontinuous sequence or a conformational structure dependent on a larger region of the protein. Example 4

Detection of cell-surface staining of

HIV-1-infected cells bv mAbs against 17^gag.

This experiment demonstrated the reactivity of two rat monoclonal antibodies against the pll³¹¹-- protein of HIV-1 with a conserved epitope which is present on the surface of live HIV-infected cells. Live H9 cells infected with HTLV-IIIb were incubated with anti-pl7 mAbs Gllgl, Gllh3 or B4f8 diluted into complete medium. After washing the cells were treated with FITC-labeled rabbit anti- rat antibody, washed, and analyzed by cell sorting, using a Coulter fluorescent-activated cell sorter. The samples were calibrated with a standard preparation of fixed blood cells and the percentage of live cells stained with the antibodies quantitated.

The above analysis, as shown by the results in Figure 4, demonstrated that antibodies Gllgl and Gllh3 were bound to the surfaces of the live cells, while antibody B4f8 was not. Control cells which were fixed with acetone prior to addition of the antibodies were stained by all three antibodies (results not shown) .

Example 5

Cell surface staining

The ability of the two monoclonal antibodies, Gllgl and Gllh3 to react with live cells was also demonstrated by immunofluorescent microscopy, showing that the pl7 epitopes with which they react are at the cell surface.

Intact, HTLV-IIIb-infected H9 cells were mixed with the monoclonal antibodies in complete growth medium, with care taken at all steps in the procedure to prevent dehydration of the cells, and then allowed to attach to multiwell slides coated with polylysine (4 ug/well) for 30 min. at 37°. The slides were then washed three times with PBS prior to fixation with acetone, and bound antibodies visualized by treatment with FITC-conjugated rabbit anti-rat antibody. Antibodies against known internal antigens do not stain cells under these conditions.

Under these conditions, antibodies Gllgl and Gllh3 stained the cells. The results are shown in Figure 5. The staining pattern observed appeared to be localized to surface membranes, in contrast to a more generalized cytoplasmic staining observed for cells previously permeabilized with acetone. None of these antibodies stained uninfected H9 cells, whether the antibodies were added before or after acetone fixation. The anti-pl7 antibodies other than Gllgl and Gllh3 and antibodies against other gag proteins did not react with live cells (data not shown) , although all of these antibodies efficiently stained infected cells which had been fixed with acetone prior to addition of the antibodies.

Example 6

Binding Competition Assay

In order to obtain further information about the epitopes recognized by the latter two antibodies, binding competition studies were performed. The two antibodies were biotinylated, and the ability of each of the anti-pl7 monoclonal antibodies to compete with the binding of these antibodies to rpl7 was determined. As shown in Fig. 6, antibody Gllh3 competed well with antibody Gllgl, but did not compete appreciably with the other antibodies even at the highest concentration tested. Similar results were obtained when biotinylated Gllgl was competed by the other antibodies (data not shown) . These results indicate that the epitopes recognized by Gllgl and Gllh3 are proximally located in the pl7 molecule, and are separated from the epitopes recognized by the other antibodies tested.

Example 7

Radioimmunoprecipitation of Hglu- labeled lysates of HIV-infected H9 cells.

This experiment demonstrated that the components recognized by the anti-pl7 antibodies corresponded to glycosylated products of the HIV gag gene. Cells infected with the HTLV-IIIb strain of

HIV were labeled with ³H-labeled glucosamine (lanes B-J) for 24 hours, and cell lysates immunoprecipitated with the following antibodies: human immune serum (lanes A and B) ; chimp antι-gpl20 serum (lane C) ; rat ant -p66 serum (lane D) ; rat anti-PG2 (lane E) ; rat anti-pl7 serum (lane F) ; mAb B4f8 (lane G) ; mAb D3b3 (lane H) ; mAb Gllgl (lane I) ; and mAb Gllh3 (lane J) . Lane A contained an immunoprecipitate of a supernatant of infected cells labeled with ³⁵cys which was included as a marker. PG2 is a recombinant protein which contains most of the p24 sequence in addition to some pl5 sequences (Centocor) . Lanes B and C were exposed to X-ray film for 1 day, lane J for 4 days, and lanes D-I for 21 days. The results are shown in Fig. 7.

Human immune serum and an anti-gpl60 serum recognized predominantly gpl60, the env protein precursor, and gpl20, the viral SU (surface) protein. Antibodies Gllgl (lane I) and Gllh3 (lane J) reacted with a complex of proteins, consisting predominantly of a closely-spaced doublet of approximately 150 kd along with a more minor doublet of approximately 90-100 kd. These proteins were also recognized, although considerably less well, by two other monoclonal antibodies against pl7 (lanes G and H) and by hyperimmune rat sera prepared against recombinant pl7 (lane F) and p24 (lane E) . None of the proteins were recognized by a hyperimmune rat serum prepared against recombinant reverse transcriptase (lane D) or when the first antibody was omitted.

Example 8

Radioimmunoprecipitation of Hglu- labeled supernatants of HIV-infected H9 cells.

This experiment demonstrated that the glycosylated gag proteins of HIV recognized by the anti-pl7 antibodies are released from infected cells.

The presence of the glycosylated gag antigens in supernatant medium of infected cells was assayed by immunoprecipitation. The results are shown in Fig. 8.

The antibodies used in this experiment were: human immune serum (lane A) ; chimp anti gpl20 serum (lane B) ; rat anti-PG2 serum (lane C) ; mAb B4f8 (lane D) ; mAb Gllgl (lane E) ; mAb Gllh3 (lane F) , and rabbit anti-rat Ig serum (lane G) . Lanes A and B were exposed for 5 days, and lanes C-G for 21 days. The human immune serum reacted with gpl60, gpl20, and gp41, as well as gpl20 degradation products of approximately 80 and 43 kd. The antiserum against recombinant p24, and all three anti-pl7 monoclonal antibodies precipitated a 150 kd protein, as well as several smaller products of approximately 100, 85, and 42 kd.

Thus it appears that the glycosylated gag antigen is released from infected cells in a form similar in size to the full-length cell-associated molecules, as well as in smaller forms.

Example 9

Analysis of cell surface components recognized by anti-HIV antibodies. This experiment demonstrated that the glycosylated gag antigens discovered were in fact localized on the cell surface. Intact HIV- infected cells were labeled by lactoperoxidase- catalyzed radioiodination, and lysates were immunoprecipita ed by the same series of antibodies as used above. The results are shown in Fig. 9.

Cell surface components of intact HIV-infected H9 cells were labeled by lactoperoxidase-catalyzed radioiodination. Cell lysates (lanes B-G) were prepared from the infected cells, and were then immunoprecipitated with: human immune serum (lanes A and B) ; anti-gp41 mAb 50/69 (13); Mab B4f8 (lane D) ; mAb Gllh3 (lane F) ; rabbit anti-rat Ig serum (lane G) . Lane A contained an immunoprecipitate of ³⁵cys-labeled supernatants of HIV-infected cells included as a marker. The immune human serum resolved bands of approximately 160 kd, 120 kd, 80 kd, and 41 kd. Significant labeling of p24 or other core proteins was not seen, demonstrating that the labeled cells were intact, and that labeling was in fact specific for cell surface proteins. Human immune serum recognized predominately the 120 kd band, with only minor labeling of gpl60 and gp41 (lane B) . An anti-gp41 monoclonal antibody recognized the gpl60 and gp41 bands, as expected (lane C) . Monoclonal antibody B4f8 did not precipitate any bands (lane D) , nor did a control sample containing only rabbit anti-rat Ig serum (lane G) . Antibody Gllh3 (lane F) reacted with a number of bands, including molecules similar in size to the 150 kd doublet detected by ³H-glucosamine labeling, together with a number of smaller components, ranging in size from 90 kd to 40 kd. A relatively small number of counts were detected at the position of pl7, indicating that pl7 is not efficiently exposed at the cell surface, if at all, in comparison with the glycosylated gag antigens.

These results show that the large glycoproteins which are recognized by the two anti-pl7 antibodies, Gllgl and Gllh3, are in fact exposed at the cell surface, and that these molecules account for the reactivity of those antibodies with live HIV-infected cells. Example 10

Virus neutralizations

Previous reports have described antisera and monoclonal antibodies directed against HIV pl7 which possess virus-neutralizing activity. The ability of the Gllgl and Gllh3 antibodies to neutralize HTLV-IIIb was determined by a quantitative fluorescent focus assay in which viral infection was measured after a single round of infection.

Supernatant medium of HTLV-IIIb-infected cells was filtered through a 0.8 micron Nalgene syringe filter, and diluted with fresh culture medium (RPMI1640 + Pen-Strep + 10% fetal bovine serum) to a concentration of approximately 4 x 10⁴ infectious units of virus/ml (as determined by limiting dilution assays) . 50 ul of virus was preincubated for 30 min at room temp with an equal volume of purified antibody diluted into fresh medium, and then added to 100 ul of medium containing approximately 100,000 uninfected H9 cells in wells of 96-well tissue culture plates. After incubation for 20-24 hours at 37°C, 35 ul of the cells suspension were plated on separate wells of polylysine-coated slides for approximately 15 minutes, washed by successive dipping in two baths of PBS and one of distilled water, and fixed by incubation in acetone for eight minutes. Cells were tested for infection by staining sequentially with a monospecific rat anti-nef serum, followed by staining with FITC-conjugated rabbit anti-rat serum (Zymed) . The slides were then counterstained with Evans blue, and the extent of infection was quantitated by counting immuno- fluorescent cells versus total counterstained cells using a Nikon Diaphot microscope equipped for epifluorescence. A minimum of 2,000 cells per well were counted, and each sample was assayed in duplicate. As controls, a pl7 antibody negative for intact cells (D3b3) and a human monoclonal antibody specific for the CD4-binding site of gpl20 were also assayed. The results are shown in Fig. 10. Whereas the anti-gpl20 antibody reduced infection by greater than 80% at concentrations below 1 ug/ml, and greater than 95% at 10 ug/ml, significant neutralization was not seen for either of the anti-pl7 antibodies even at a concentration of 100 ug/ml. This suggests either that the two anti-pl7 antibodies may not bind to the surface of infectious virions, or that if they do, such binding is not detrimental to functions of the virus required for infection.

Claims

We claim:

1. A substantially pure glycoprotein which is expressed on the surface of HIV infected cells, which glycoprotein contains gag amino acid sequences.

2. A substantially pure glycoprotein as in claim 1 which is bound by antibody Gllgl.

3. The substantially pure glycoprotein of claim 2 which has a molecular weight of about 150,000 as measured by SDS-PAGE.

4. The substantially pure glycoprotein of claim 2 which has a molecular weight of about 90,000 as measured by SDS-PAGE.

5. The substantially pure glycoprotein of claim 1 which is bound by antibody Gllh3.

6. The substantially pure glycoprotein of claim 2 having an amino acid sequence of an HIV-1 protein.

7. The substantially pure glycoprotein of claim 2 having an amino acid sequence of an HIV-2 protein.

8. A substantially pure fragment of the glycoprotein of claim 1, which fragment is released by the HIV infected cells.

9. The substantially pure fragment of claim 8 which is selected from the group consisting of fragments having a molecular weight of about 100,000, 85,000, and 42,000 kd.

10. A monoclonal antibody to a glycoprotein which is expressed on the surface of HIV infected cells, which glycoprotein contains gag amino acid sequences and has a molecular weight of about 150,000 as measured by SDS-PAGE, which monoclonal antibody reacts with live HIV infected cells.

11. The monoclonal antibody of claim 10 which reacts with HIV-1 infected cells.

12. The monoclonal antibody of claim 10 which reacts with HIV-2 infected cells.

13. A monoclonal antibody which competitively inhibits the binding of antibody Gllgl to a substantially pure glycoprotein which is expressed on the surface of HIV infected cells, which glycoprotein contains gag amino acid sequences and has a molecular weight of about 150,000 as measured on SDS-PAGE.

14. The monoclonal antibody of claim 13 which does not substantially bind to the peptides

A. MGARASVLSGGELDR B. GELDRWEKIRLRPGG

C. LRPGGKKKYKLKHIV

D. LKHIVWASRELERFAV

E. LERFAVNPGLLETSE

F. LETSEGCRQILGQLQ G. GQLQPSLQTGSEELRSL

H. EELRSLYNTVATLY

I. LYNTVATLYCVHQRI

J. YCVHQRIEIKDTKEALDKIEEEQNKSKKKA K. EIKDTKEALDKIEEE

L. EEEQNKSKKKA

M. SKKKAQQAAADTG

N. DTGHSSQVSQNY

15. The monoclonal antibody of claim 14 having the identifying characteristics of antibody Gllgl.

16. A monoclonal antibody which competitively inhibits the immunological binding of antibody Gllh3.

17. The monoclonal antibody of claim 16 which does not substantially bind to the peptides

A. MGARASVLSGGELDR B. GELDRWEKIRLRPGG

C. LRPGGKKKYKLKHIV

D. LKHIVWASRELERFAV

E. LERFAVNPGLLETSE

F. LETSEGCRQILGQLQ G. GQLQPSLQTGSEELRSL

H. EELRSLYNTVATLY

I. LYNTVATLYCVHQRI

J. YCVHQRIEIKDTKEALDKIEEEQNKSKKKA

K. EIKDTKEALDKIEEE L. EEEQNKSKKKA

M. SKKKAQQAAADTG

N. DTGHSSQVSQNY

18. The monoclonal antibody of claim 16 having the identifying characteristics of antibody Gllh3.

19. A therapeutic agent comprising the monoclonal antibody of claim 10 conjugated to a toxin.

20. A therapeutic agent comprising the monoclonal antibody of claim 13 conjugated to a toxin.

21. A therapeutic agent comprising the monoclonal antibody of claim 14 conjugated to a toxin.

22. An immunoassay comprising the steps of providing a monoclonal antibody to a glycoprotein which is expressed on the surface of HIV infected cells, which glycoprotein contains gag amino acid sequences and has a molecular weight of about 150,000 as measured by SDS-PAGE, which monoclonal antibody reacts with live HIV infected cells, and detecting the formation of a complex between the monoclonal antibody and antigen for which the monoclonal antibody is specific.

23. The immunoassay of claim 22 wherein the assay is an assay for determining the number of cells infected with HIV.

24. The immunoassay of claim 23 wherein the HIV is HIV-1.

25. The immunoassay of claim 23 wherein the HIV is HIV-2.

26. A kit for determining the presence of a glycoprotein which is expressed on the surface of, or secreted from, HIV infected cells, which glycoprotein contains gag amino acid sequences, which kit comprises the monoclonal antibodies of claim 10, and means for detecting the formation of a complex between the monoclonal antibodies and the glycoprotein.

27. A therapeutic agent comprising the substantially pure glycoprotein of claim 1 in combination with a pharmaceutically acceptable carrier.

28. A therapeutic agent comprising the substantially pure glycoprotein of claim 3 in combination with a pharmaceutically acceptable carrier.

29. A monoclonal antibody which competitively inhibits the immunological binding of antibody B4F8.

30. A monoclonal antibody which competitively inhibits the immunological binding of antibody

D3b3.

31. The monoclonal antibody of claim 29 having the identifying characteristics of antibody B4F8.

32. The monoclonal antibody of claim 30 having the identifying characteristics of D3b3.