The present invention is directed to a vector comprising a hybrid SIV/HIV genome. This vector can be used to establish an animal model for studying the HIV-1 virus.

Human immunodeficiency virus type 1 (HIV-1) and, to a lesser extent, human immunodeficiency virus type 2 (HIV-2) are etiologic agents of acquired immune deficiency syndrome (AIDS) in humans [Barre-Sinoussi, F. , et al.. Science 120:868-871 (1983); Gallo .C., et al.. Science 224.500-503 (1984); Clavel, F. , et al., AIDS 1:135-140 (1987)]. These viruses are related to simian immunodeficiency viruses that infect feral populations of sooty mangabeys, African green monkeys, and mandrills [Desrosiers, R.C., et al., Ann. Rev. Immunol. jB:557-558 (1990)]. A simian immunodeficiency virus (SIV _mac ) capable of infecting and inducing an AIDS-like disease in macaques is closely related to HIV-2 and SIV _smτn [Letvin N.L., et al.. Science 230:71-73 (1985); Daniel M.D., et al.. Science 228:1201-1204 (1985)].

The primate immunodeficiency viruses establish persistent infections in their hosts even in the face of an antiviral immune response. Part of this ability may reside in the capacity of these viruses to tightly regulate

expression of the viral proteins, as evidenced by the presence of four conserved regulatory genes in all members of this group of retroviruses.

In addition to the gag, pro, pol and env genes typical of retroviruses, these viruses contain vif. tat, rev, and nef genes [Haseltine W. , et al., Raven Press (1990)]. The tat protein stimulates the viral LTR to express viral RNA [Arya S., et al., Science 229:69-73 (1985); Sodroski J., et al., Science 229:74-77 (1985)] while the rev protein promotes the nuclear egress of viral messenger RNA's encoding the structural gene products [Emerman M. , et al., Cell 57:1155-1165 (1989); Malim M. , et al. , Nature 138:254-257 (1989)]. Both tat and rev genes are essential for viral replication [Dayton A., et al., Cell 44:941-947 (1986); Fisher A.G., et al., Nature 320:367-371 (1986); Sodroski J., et al., Nature 321:412-417 (1986)]. The vif and nef genes, although dispensable for virus replication in some tissue culture settings, are well-conserved [Sodroski J., et al., Science 231:1549-1553 (1986); Fisher A.G., et al., Science 237:888-893 (1987); Strebel K. , et al., Nature 328:728-730 (1987); Kestler H.W. , et al., Cell 6J5:651-662 (1991)]. Depending upon the particular primate immunodeficiency virus, vpx. vpr and/or vpu genes are present in the proviral DNA [Desrosier R.C., et al., Ann. Rev. Immunol. 8:557-578 (1990); Haseltine W. , Raven Press (1990)] . These genes are also dispensable for virus replication in tissue culture. The vpx and vpr proteins are incorporated into virions and are believed to play a positive role in the early phase of the virus life cycle [Cohen E.A. , et al., JAIDS 3:11-18 (1990); Yu, X-F, et al., J. Virol. 64:5688-5693 (1990) ; Henderson .E. , et al. ,

Science 241:199-201 (1988); Hu, W. , et al., Virology 173:624-630 (1989); Kappes J.C., et al., Virology 184:197-209 (1991); Hattori N. , et al., Proc. Natl. Acad. Sci. U.S.A. 87:8080-8084 (1990)]. The vpu gene is found - only in HIV-1 and encodes a 15-20 kD protein, depending upon the virus isolate [Terwilliger E.F., et al., Proc. ^~ Natl. Acad. Sci. U.S.A. 86:5163-5167 (1989); Cohen E.A. , et al., Nature 344:532-534 (1988); Strebel K. , et al., Science 241:1221-1223; (1988); Klimkait T., et al., J. Virol. 64:621-629 (1990)]. The vpu protein is associated with the host cell membranes and facilitates the redistribution of viral proteins from inside the infected cell to free virion particles [Terwilliger E.F., et al., Proc. Natl. Acad. Sci. U.S.A. 86:5163-5167 (1989); Cohen E.A. , et al. , Nature 344:532-534 (1988); Strebel K. , et al., Science 241:1221-1223; (1988); Klimkait T. , et al., J. Virol. 64:621-629 (1990); Strebel K. , et al, J. Virol. 61:3784-3791 (1989)]. Thus, the major function of the vpu product is to modulate virus release, although other effects of vpu on envelope glycoprotein or CD4 steady state levels have been observed [Willey R. , et al., J. Virol. 6i£:226-234 (1992); Kimura T. and Karn J., personal communication] . The in vivo function of the VPU protein is unknown.

The persistence of primate immunodeficiency virus infection is also made possible by the particular features of the viral envelope glycoproteins. The viral glycoproteins are synthesized as a 160 Kd precursor, which is cleaved intracellularly to yield the gpl20 exterior envelope glycoprotein and the gp41 transmembrane glycoprotein [Allan J.S., et al., Science £21:1091-1093

(1985); Robey W.G., et al., Science 228:593-595 (1985)]. The gpl20 glycoprotein binds the CD4 receptor, following which the gpl20 and gp41 glycoproteins in concert contribute to the membrane fusion process [Klatzmann D., et al., Nature 312:767-768 (1984); Dalgleish A.G. , et al., Nature 312:763-767 (1984); Helseth E. , et al., J. Virol. 64:2416-2420 (1990)]. The latter process mediates both virus entry and viral cytopathic effect, which consists of multinucleated giant cell (syncytium) formation and single cell lysis [Sodroski J., et al., Nature 322:470-474 (1986); Lifson J.D., et al., Nature 323:725-728 (1986); Kowalski K., et al., J. Virol. 65:281-291 (1991)]. The exterior envelope glycoproteins of these viruses are heavily glycosylated and contain regions of hyper-variability, most of which are thought to consist of disulfide-linked loops exposed to the exterior of the protein [Leonard C. , et al., J. Biol. Chem. 265:10373-10382 (1990)]. In the case of HIV-1, most of the neutralizing antibody response elicited early in the course of infection is directed against the third variable (V3) loop of the gpl20 glycoprotein [Nara P., et al., Proc. Quatreime Collogue des Cent Gardes (Girard, Valette, eds, Paris: Pasteur Vaccins) pp. 203-215 (1989)]. These antibodies inhibit some aspect of the membrane fusion process [Skinner M. , et al., J. Virol. £2:4195-4200 (1988); Linsley P., et al., J. Virol. 62:3695-3702 (1988)]. Neutralization is generally - strain-restricted due to variation in the V3 region, but some antibodies recognize better conserved elements near the tip of the loop [Ohno T. , et al., Proc. Natl. Acad. Sci. U.S.A. 88:10726-10729 (1991); Matthews T., et al., Proc. Natl. Acad. Sci. U.S.A. 83:9709-9713 (1986); Javaherian K. , et al., Science 250:1590-1593 (1990)]. The anti-V3 loop antibodies are protective against intravenous

challenge by homologous HIV-1 [Berman P., et al. , Nature 345:622-625 (1990); Emini E. , et al., Nature 355:728-730 (1992)]. Later in the course of HIV-1 infection, antibodies that neutralize a broader range of HIV-1 isolates are generated [Weiss R.A. , et al., Nature 124:572-575 (1986); Profy A. , et al., J. Immunol. 144:4641-4647 (1990); Berkower I., et al., J. EXP. Med. 170:1681-1695 (1989)]. These antibodies recognize discontinuous epitopes near the CD4 binding site of gpl20 and block the binding of gpl20 to CD4 [Ho D. , et al., J. Virol. 65:489-493 (1991); Kang C-Y, et al. , Proc. Natl. Acad. Sci. U.S.A. 88:6171-6175 (1991); Steimer K.S., et al., Science 254:105-108 (1991)]. The epitopes for some of these antibodies have been mapped by extensive mutagenesis, and depend upon amino acids located in all five conserved gpl20 regions [Thali M. , et al., J. Virol. £5. 18-6193 (1991); Thali M. , et al., Discontinuous, conserved neutralization epitopes overlapping the CD4 binding region of the HIV-1 gpl20 envelope glycoprotein, J. Virol. j6_6:5635-5641] . These neutralizing antibodies do not keep virus replication in check indefinitely, probably because of virus variation and selection of neutralization-resistant viruses and because of immunosuppression and compromised ability of the host to respond to novel epitopes [Nara P., et al., J. Virol. 64:3779-3791 (1990); Gegerfelt A., et al., Virology 185:162-168 (1991); Arendrup M. , et al., JAIDS 5:303-307 (1992)]. It is critical to both immunotherapeutic and vaccine efforts th _* .. understanding of lizing antibodies and virv - iriation as a meant scape is achieved.

Hosts infected with the primate immunodeficiency viruses also generate cellular antiviral immune responses [Walker, B. , et al., Nature 328:345-348 (1987); Plata, F. , et al., Nature 328:348-351 (1987); Roup, R. , et al., Research in Immunology 140:92-95 (1989); Miller, M.D., et al., J. Immunol. 144:122-128 (1990)]. In particular, CD8-positive, MHC class I-restricted cytotoxic T lymphocytes (CTL) directed against a number of viral structural and regulatory proteins have been identified in HIV-1 infected humans and SIV-infected monkeys [Walker, B. , et al., Nature 328:supra: Plata, F. , et al., Nature 328. supra: Koup, R. , et al., Research in Immunology 140. supra: Miller, M.D. , et al., J. Immunol. 144. supra! . These CTL are capable of lysing autologous target cells expressing viral proteins and are likely to constitute an important immune mechanism for suppressing established viral infections. Acute retroviral infection is often accompanied by an intitial period of viremia, which subsequently abates in most infected hosts [Coombs, R. , et al., New Eng Journal of Medicine 321:1626-1631 (1989); Ho, D., et al., New Eng. Journ. of Med. 321:1621-1625 (1989); Piatak, M. , et al. , Science 259:1749-1752 (1993)]. This abatement has been temporally ^" associated with the development of a virus-specific CTL response in murine leukemia virus-infected mice, SIV-infected macaques and HIV-1-infected humans [Blank, K. , personal communication; Yasutomi, Y. , et al., J. Virol. 67:1707-1711 (1993); Koup, R. , et al., personal communication]. In the former case, specific depletion of CD8-positive T lymphocytes results in a persistence of the initial viremic episode, strongly suggesting a necessary role for CTL in suppression of the initial wave of infection [Blank, K. , supra1.

An understanding of the correlates of protective immunity at both humoral and cellular levels is critical to the development of an effective HIV-1 vaccine. A major tool in vaccine efforts is an animal model in which variables related to immunization protocols, adjuvant choice, virus challenge strain, state of the virus (cell-free versus cell-associated) and mode of challenge (mucosal versus intravenous) can be assessed. In the HIV-1 system, many of these variables have not been suitably addressed due to the scarcity and expense of the chimpanzees required. The chimpanzee model is of limited utility for immunotherapeutic efforts since the level of HIV-1 replication is variable and, to date, little evidence of disease induction exists [Fultz P., et al., Science 226:549-552 (1986)]. HIV-1 infection of mice, rabbits and SCID-hu mice has been reported [McCune J., et al., Science 247:564-566 (1990); Mosier D. , et al., Science 251:791-794 (1991); Filice G., et al., Nature 335:366-369 (1988)], but the applicability of these models for vaccine and immunotherapeutic efforts remains questionable.

The most useful current model for HIV infection is the infection of macaques with various strains of ^SIV smm' ^SIV mac ^{or HIV} - ² [Desrosier R.C. et al. , Ann. Rev. Immunol. 8:557-578 (1990); Letvin N.L. et al. , Science 230:71-73 (1985)]. AIDS-like diseases have been induced using infectious molecular clones of SIV _mac , and recently an animal-passaged jlIV-2 has been reported to induce disease in cynomolgus monkeys [Biberfeld G., et al., Sixieme Colloσue des Cent Gardes (eds., Girard, Valette, Paris, Pasteur Vaccins), pp. 225-229 (1991)]. The availability of these monkeys has allowed many more

experimental variables relevant to vaccine development or genetic variation to be tested, relative to the HIV-1/chimpanzee model. Although the contribution of human cell components to protection has recently clouded interpretation of some of the vaccine results using the SIV/macaque model [Scott E.J. et al., Nature 353:393 (1991)], it does appear that priming with vaccinia/gpl60g _jV followed by boosting with purified SIV gpl20 can confer protection against the homologous virus [Hu S.L. , et al., Science 255:456-459 (1992)]. Also, the in vivo role of the nef gene of SIV in contributing to virus replication was demonstrated using this system [Kestler H.W., et al., Cells 65:651-662 (1991)]. While relevant to primate immunodeficiency viruses in general and relevant to HIV-2 in particular, it has limitations with respect to HIV-1, specifically.

As a greater understanding of the molecular details of the antiviral immune response to the primate immunodeficiency viruses emerges, it is becoming increasingly clear that significant differences exist between the envelope glycoproteins of HIV-1 and those of the HIV-2/SIV _smm viruses. While general organizational features of these glycoproteins are conserved among the primate immunodeficiency viruses, structural differences exist in the number of disulfide bonds, the particular amino acids near or within the CD4 binding region or gp41 amino terminal fusion peptide, and the composition of the hypervariable loops [Myers G., et al., "Human retroviruses and AIDS" (1991)]. These structural differences translate into significant differences in the immune responses generated to the HIV-1 and HIV-2/SIV _smm glycoproteins.

Both major groups of neutralizing antibodies generated by HIV-1 do not apparently correspond to those elicited by the HIV-2/SIV _smm viruses. The V3 loop of HIV-1 is exposed on the native glycoprotein, efficiently: elicits neutralizing antibodies and can determine the ability of HIV-1 to enter primary monocytes/macrophages [Ohno T. , et al. , Proc. Natl. Acad. Sci. U.S.A. 88:10726-10725 (1991); Matthews T., et al., Proc. Natl. Acad. Sci. U.S.A. 83:9709-9713 (1986); Javaherian K. , et al., Science 250:1590-1593 (1990); Hwang S.S., et al., Science 253:71-74 (1991)]. By contrast, the analogous region of SI _mac is not variable, does not appear to be well-exposed on the native glycoprotein, is not an efficient target for neutralizing antibodies, and does not correspond to the region of SIV env implicated in primary macrophage tropism [Burns D.P.W., et al., J. Virol. 65:1843-1854 (1991); Overbaugh J., et al., J. Virol. 61:7025-7031 (1991); Scott Putney, personal communication; Mori K., et al., J. Virol, in the press (1992)]. In contrast to the broadly neutralizing antibodies of HIV-1, many of which block gpl20-CD4 binding and recognize discontinuous epitopes spanning most of the conserved regions [Ho D., et al., J. Virol. 65:489-493 (1991); Kang C-Y, et al., Proc. Natl. Acad. Sci. U.S.A. 88:6171-6175 (1991); Steimer K.S., et al., Science 254:105-108 (1991); Thali M., et al., J. Virol. ?5:6188-6193 (1991); Thali M. , et al., J. Virol. 66. supra: Posner M. , et al., J. Immunol. 146:4325-4332 (1991); Tilley S.A., et al., Res Virol. 142:247-259 (1991)], the major conserved neutralization epitope of SIV _mac is confined to a 179 amino acid carboxyl fragment of gpl20 [Javaherian K. , et al. , Sixieme Colloαue des Cent Gardes (eds., Girard, Valette, Paris, Pasteur Vaccins), pplδl-164 (1991)]. Although the latter

epitope is discontinuous, it does not overlap the CD4 binding site, since ternary complexes of gpl20, soluble CD4, and neutralizing antibody can be formed [Javaherian K. , et al., Sixieme Collogue des Cent Gardes (eds., Girard, Valette, Paris, Pasteur Vaccins) , pplT.1-164 (1991)]. The existence of differences between the conserved neutralization epitopes of the HIV-1 and HIV-2/SIV _smm groups of viruses is further emphasized by the observation that broadly neutralizing monoclonal antibodies directed against these viruses do not cross-neutralize members of the other group [Weiss R.A. et al., Nature 324:572-575 (1986); Ho D., et al. , J. Virol. 65:489-493 (1991); Posner M. , et al., J. Immunol. 146:4325-5332 (1991); Tilley S.A., et al., Res. Virol. 142:247-259 (1991)]. Thus, we cannot expect that the details of vaccine or immunotherapeutic formulaton will be_directly extrapolatable from the HIV-2/SIV _smm models to HIV-1. This, plus the absence of certain genes like vpu from the SIV genome, illustrate the need to establish new animal models that allow assessment of the in vivo consequences of variation within HIV-1-specific genes. Preferably, these models should involve both infection by the virus and disease induction.

HIV-1 has been reported to infect Macaca nemestrina after inoculation with cell-associated virus [Katze J., et al., personal communication, UCLA Symposium of HIV and Related Viruses, Keystone, CO (1992)]. However, to date, no disease has been reported to be observed in the infected macaques, six months after inoculation. Although the HIV-1/M-nemestrima system would have HIV-1 specific genes, it is limited to one macaque species and has not been reported to yield disease. Furthermore, macaque nemestrima

is not widely used in primate centers in this country, so it would take considerable time, effort and expense to be able to use such species.

Shibata, R. , et al. reported preparing a chimeric virus containing HIV-1 tat, rev and env genes in a SIV provirus. J. Virol. 65:3514-3522 (1991). This SIV provirus does not contain functional vpr and nef genes. Indeed, it was reported that SIV vpr and SIV nef are not necessary for viral replication and infection of tissue cultured cells by these authors. These chimeric viruses were reported to replicate in macaque peripheral blood mononuclear cells (PBMC) . The Shibeta et al. chimeric virus has been claimed to Infect macaques, but the level of virus replication was very low and the infection did not persist beyond two months (Hayami, personal communication).

It would be desirable to have a vector containing HIV-1 genes which produces a virus that could be used to infect a number of animals in addition to humans and chimpanzees in order to be able to develop an animal system for studying the disease. It would also be useful if such a system was set up so that different HIV strains could readily be studied. It would also be useful if an animal model could be established so that antibody protection, virus variation and virus infection could be studied.

Still further, it would be useful to be able to use such a system for the preparation and/or screening of vaccines, therapeutics and modes of administration.

It is yet another objective of the present invention to

prepare a vector into which the different envelope genes of the various HIV-1 strains can be inserted, which can then be used to infect animal models in order to prepare vaccines, prepare therapeutics and follow the evolution and differentiation of envelope glycoprotein In vivo.

All of these uses require a virus that replicates efficiently and achieves a high titer in several monkey species. Such a system is currently unavailable.

Summary of Invention

We have now discovered a vector which consists essentially of a DNA sequence containing a 5' portion corresponding to a sufficient number of nucleotides to encode the following functional SIV or HIV-2 structural proteins, gag, pro, pol of SIV or HIV-2, and to encode as functional SIV regulatory proteins, vif, vpx. and vpr. and having a 5' SIV or HIV-2 LTR and a 3' portion corresponding to a sufficient number of nucleotides corresponding to an HIV-1 genome to encode as a functional HIV-1 structural protein, env. and as functional HIV regulatory proteins, tat and rev and as a functional SIV or HIV-2 regulatory protein, nef, and having a SIV or HIV-2 LTR. In a preferred embodiment, the 3' portion also contains a sufficient number of nucleotides to encode a functional HIV-1 vpu gene. Preferably, the nucleotide sequence of the vector contains sequences that correspond to the SIV or HIV-2 £a£ splice acceptor and/or the SIV or HIV-2 rev splice acceptor, but does not contain sequences corresponding to the HIV-1 tat splice acceptor. Preferably, the SIV genome corresponds to the SIV genome of

the strain SIV _mac , SIV _agm , SIV^. More preferably it corresponds to SIV .

The HIV genome can correspond to any of the known HIV-1 strains. In one preferred embodiment, the HIV segment corresponds to HIV genomes capable of encoding functional vpu proteins such as ELI, BH10, BRU, etc.

When the HIV or SIV strain corresponds to a genome not capable of encoding a function protein such as vpu of HXBc2 strain or nef of SIV _mac 239, one can modify the sequence in order to produce a sequence that will encode a functional protein. For example, with the strain HXBc2, one would modify the DNA sequence to insert an AUG codon immediately upstream and in proper reading frame with the vpu open reading frame at a nucleotide corresponding to immediately before HIV nucleotide 5541 or create a point mutation to generate such a sequence. This can readily be done by techniques well known in the art.

Brief Description of the Drawings

Figure 1 is a linear schematic showing the structure of one of the preferred hybrid vectors.

Figure 1A is a linear schematic of the genetic organization of the HIV-1 sequence and the SIV sequence.

Figure IB shows the details of the 5' SIV _mac HIV-l junction near the Sph I site (S) for a variety of hybrid vectors.

Figure 1C shows the details of the 3' HIV-l/SIV _mac 239 (nef open) junction near the Rsr II site (R) .

Figure ID is a linear schematic showing the structure of a different preferred hybrid vector.

Figure 2 is a graph showing the replication of these hybrid vectors in CEMxl74 lymphocytes.

Figure 3 shows viral protein production in cells infected by these vectors.

Figure 3A are autoradiographs of CEMxl74 cells infected with SIV _mac 239 virus or a virus produced by one of the vectors of the present invention from cynomolgus monkey PBMCs.

Figure 3B are autoradiographs showing CEMxl74 cells infected with virus from one of the vectors isolated from cynomolgus macaques at 2 weeks post inoculation.

Figure 4 is a graph showing neutralization of HIV _jj γ _jjc 2 by a SHIV-infected macaque.

Figure 5 is a graph showing neutralization of HIV^ by a SHIV-infected macaque.

Detailed Description of The Invention

We have now discovered vectors that will produce chimeric viruses containing HIV-1 components. As a result of transfecting a cell with these vectors, replication

competent viruses that are infectious in animal systems such as monkeys, mandrils, macaques, etc. can be produced.

The typical method of developing a vaccine to prevent infection by a virus has utilized suitable animal models. However, the two animal models most commonly used have serious deficiencies with respect to studying HIV-1. HIV-1 does not replicate to high titers in chimpanzees and infected chimpanzees do not develop immunodeficiency. [Alter, H., et al. Science 226:549-552 (1984); Fultz, P.N. , et al., J. Virol. 58:116-124 (1986); Fultz, P., et al., Science 226:549-552 (1986); Gajdusek, D.C., et al., Lancet 1:55-56 (1985); and Nara, P.L., et al., J. Virol. £1:3173-3180 (1987)]. Furthermore, trials in chimpanzees are limited to a few animals since the species is endangered and available chimpanzees and their care is expensive.

Rhesus and cynomolgus macaque monkeys infected with SIV such as the macaque strain of SIV SIV _mac ) produce high titers of virus and develop an AIDS-like syndrome [Daniel, M.D., et al., Science 228:1201-1204 (1985); Kestler, Science 248:1109-1112 (1990); Letvin, N.L., Science 230:71-73 (1985)]. However, differences exist in the immune response to SIV-1 and SIV _mac envelope glycoproteins, which represent the principal targets for protective immunity and the response to HIV-1 envelope glycoproteins. [Berman, P., et al., Nature 345:622-625 (1990); Emini, E. , et al. , Nature 355:728-730 (1992); Hu, S.L. , et al., Science 255:456-459 (1992)]. The major neutralizing antibodies in HIV-1 infected people are directed against two regions of the gpl20 envelope

glycoprotein. Antibodies against the HIV-1 third gpl20 variable (V3 region) have been reported to be protective [Emini, E. , et al. Nature, supra1. As the name implies, this region shows great sequence variation among the various HIV strains. In contrast, the corresponding region of the SIV _mac envelope glycoprotein does not exhibit sequence variation among isolates and is not a target for neutralizing antibodies in infected macaques. [Burns, D.P.W., et al., J. Virol. 65:1843-1854 (1991); Overbaugh, J., et al., J. Virol. 65:7025-7031 (1991)]. HIV-1 infected humans also exhibit a second group of neutralizing antibodies which are directed against a conserved discontinuous gpl20 region that binds the CD4 viral receptor. [Berkower, I., et al, J. Exp. Med. 170:1681-1695 (1989); Dalgleish A.G., et al. Nature 312:763-767 (1984); Haigwood, N. , et al., Vaccines 90:313-320 (1990); Ho, D., et al., J. Virol. 65:489-493 (1991); Kang, C.Y. , et al., Proc. Natl. Acad. Sci. U.S.A. 88:6171-6175 (1991); Klatzmann, D., et al. , Nature 312:767-768 (1984); McDougal, J.S., et al., J. Immunol. 137:2937-2944 (1986); Posner, M. , et al, J. Immunol. 146:4325-4332 (1991); Profy, A., et al., J. Immunol. 144:4641-4647 (1990); Steimer, K.S., Science 254:105-108 (1991); Tilley, S.A., Res. Virol. 142:247-259 (1991)].

However, this second group of antibodies recognize HIV-1 gpl20 regions that are distinct from those of the SIV gpl20 glycoproteins recognized by antibodies from infected macaques that neutralize multiple SIV strains. [Thali, M., et al., J. Virol. 65:6188-6193 (1991); Javaherian, K. , et al., pp. 161-4 in Sixieme Collooues des Cent Dardes (Eds., M. Girard and L. Valette, Paris, Pasteur

Vaccins)]. This difference between the antibodies that broadly neutralize HIV-1 and SIV strains is further stengthened by the finding that such antibodies do not cross-neutralize. [Weiss, R.A. , et al., Nature 324:572-575 (1986)].

We have found a vector which will produce a hybrid virus between HIV-1 and SIV (or HIV-2), which expresses HIV-1 envelope glycoproteins and is capable of replicating to high titers in animal systems such as monkeys, preferably macaque monkeys. As used herein the term SIV will also refer to HIV-2. SIV _smm , SIV _mac and HIV-2 have all been used to induce disease in monkeys.

The vector consists of a DNA sequence comprising the SIV LTRs and a sufficient number of nucleotides to encode functional SIV nef protein. The SIV sequences preferably correspond to SIV _agm , SIV^p, SIV _mac or HIV-2. More preferably the sequences correspond to SIV or HIV-2. Still more preferably, the sequences correspond to SIV _mac . The vector also comprises a sufficient number of nucleotides corresponding to the HIV-1 genome to encode functional HIV-1 tat, rev and env proteins. The vector also contains a sufficient number of nucleotides to encode functional HIV or SIV gag, pol. pro and vif and vpr proteins. Preferably, it also encodes functional vpu and/or vpx proteins. In one preferred embodiment the vector encodes functional HIV-1 gag, pol. pro, vif. vpr. vpu. env. rev and tat proteins, and functional SIV nef protein. In an alternative embodiment, the vector can encode ^" different combinations of HIV-1 gag, pol. pro, vif.

vpr. vpu. env. rev and tat proteins, and functional SIV nef protein.

In the preferred embodiment in which the vector encodes functional HIV-1 gag, pol, βro, vif, vpr. vpu. env. rev, and tat proteins, and functional SIV nef protein, there-are two junctions between the HIV-1 and SIV sequences. The 5' junction joins the SIV 5' LTR and the HIV-1 sequences immediately 5' to the tRNA primer binding site. The 3' junction joins the HIV-1 sequences immediately 3' to the HIV-1 env gene to the SIV sequences that include the nef gene and 3' LTR, and would be similar to that shown in Figure lC. Figure ID is a schematic showing this vector. In this Figure all the genes in the HIV-1 segment are active and will express functional protein.

In another preferred embodiment, the vector of the present invention comprises a DNA sequence corresponding to a sufficient number of nucleotides of the SIV genome to encode functional SIV gag, pro and pol structural proteins and functional vif. vpx. vpr and nef SIV regulatory proteins. The vector also contains a sufficient number of nucleotides corresponding to the SIV LTRs. These sequences are referred to as the SIV segment. The vector also contains a JDNA segment corresponding to a sufficient number of nucleotides of the HIV-1 genome to express a functional HIV-1 envelope glycoprotein and at least the tat and rev HIV-1 regulatory proteins. This segment is called the HIV-1 sequence. In a preferred embodiment, the HIV-1 sequence also contains a sufficient number of nucleotides corresponding to the HIV-1 genome to express a functional vpu regulatory protein.

The HIV genome can correspond to any of the known HIV-1 strains. Such HIV-1 strains include HXB2, ADA, MN, RF, SF-2, MAL, ELI, YU-2, CM ₂₃₅ , ZAM, U455, JRCSF, CDC42, JRFL, BAL, 89.6, NL43, SC, NDK, etc.„ The HIV genomes including SIV and HIV-1 have been extensively mapped for a variety of different strains. Thus, the skilled artisan can readily prepare nucleotide sequences that will contain a sufficient number of nucleotides to encode such functional proteins. Although there is strain to strain variation, both SIV (including HIV-2) and HIV-1 show a significant functional sequence homology, which phenomenon is well known to the person of ordinary skill in the a Thus, the skilled artisan can readily prepare sequences that will produce functional proteins. For example, although the HXBc2 provirus does not encode a functional vpu protein, it Is known that by inserting an AUG codon just upstream, and in frame with the vpu open reading frame, e.g., at HIV-1 HXBc2 nucleotide 5541, one can express a functional vpu protein [Cohen, E. , et al., Nature 334:532-534 (1988); Terwilliger, E. , et al., Proc. Natl. Acad. Sci. USA 86:5163-5167 (1989)]. Other alterations can be made to produce functional proteins. A functional structural protein -is one that when expressed assembles into a virion in conjunction with gag and performs a particular replicatlve function. A functional regulatory protein is one that will exhibit in vivo or in vitro a known functional property. For example, the tat protein stimulates viral LTR to express viral RNA. As used herein, the term corresponding to include conservative deletions, alterations and additions, e.g., coding for a change from one amino acid to another that will preserve

the function of the protein.

These sequences can be prepared by a variety of means well known to the skilled artisan. For example, one can use SIV proviruses and HIV proviruses ^{^} to generate the sequence. Another method involves the synthesis of the nucleotides based on known sequences. Preferably, the nucleotides that correspond to a sufficient number of nucleotides to encode a functional protein is the gene for the protein.

We have also found that it is preferable that the

> vectors do not contain nucleotide sequences corresponding to too many multiple splice acceptors. For example, it is preferred that the vector does not contain nucleotide sequences corresponding to the SIV tat splice acceptor, the

SIV rev splice acceptor and the HIV-1 tat splice acceptor.

Preferably, the vector does not have a sufficient number of sequences corresponding to the HIV-1 tat splice acceptor.

In one preferred embodiment, the vector can be derived by using an infectious SIV provirus such as SIV _mac 239 nef (virus) gag+, pro+. pol+ vif+. vpχ+. vpr+. tat+. rey+, env-f-. nef+) [Kestler, H. , et al. , Science 248:1109-1112 (1990); and Kestler, H.W. , et al., Cell 65:651-662 (1991), both of which are incorporated herein by reference], although other SIV strains can be used and, an HIV-1 provirus such as, BRU, ELI, Mai, HXBc2, BH10, BH5, ADA etc., for example, in the following discussion the HXBc2 strain (gag+. pro+. pol+ vif+, vpjr-, £at+, ∑ev+, vpju- , env+, nef-) [Fisher, et al., Nature 316:262-266 (1985), which is incorporated herein by reference] will be used as

exemplary, but any of the other HIV-1 strains can readily be used instead.

The vector contains a 5' portion and a 3' portion. The 5' portion of the vector has nucleotides corresponding to the SIV genome. Figure 1 is a linear schematic showing -the vector. The white boxes correspond to HIV-1 specific sequences, whereas the darkened correspond to SIV specific sequences. Genes that are defective in these strains, meaning they will not encode a functional protein, are denoted within astericks. The 5' SIV _mac HIV-l junction at the Sphl site (S) and the 3' HIV-l/SIV _mac junction at the RsrII site (R) are shown. The stippled 3' end of vpr of the hybrid vector represents sequences derived from the HIV-1 portion of the chimera that reconstitutes the SIV _fflac 239 vp . Thus, the 5' portion of the vector comprises a sequence corresponding to a 5' SIV-LTR, SIV gag gene, SIV pol gene, SIV vif gene, SIV vpx and a chimeric gene derived from the 5' and 3' portions that corresponds to the SIV vpr.

The 3' portion of the vector comprises ^" pat, rev and env sequence derived from HIV-1 isolates and the nef and 3' LTR sequence derived from SIV. Figure IB shows the details of a 5' SIV _mac HIV-l junction near the Sphl site (S) for a number of different vectors. The details of segments from SHIV-1 are represented by SEQ ID N0S:1, 2, 3 and 4: SHIV-2 are represented by SEQ ID N0S:1, 5 and 6; SHIV-3 are represented by SEQ ID N0S:7, 2, 3 and 4; and SHIV-4 are represented by SEQ ID N0S:7, 5 and 6. The positions of the splice acceptors (S. A.) for the SIV _mac tat and rev

messages and for the HIV-1 tat messages are shown above the figures, with the SIV _mac tat initiation codon, SIV _mac vpr stop codon and the HIV-1 tat initiation codon underlined and labeled with arrows below. The astericks denote sequence identity and the dot represents sequences not shown. The horizontal bars represent sequence deletions. The X marks the position of an Xba I site in 2 of the vectors designated SHIV-3 and SHIV-4. Figure 1C shows the details of the 3' HIV-l/SIV _mac 239 (nef open) junction near the Rsr II site (R) and is represented by SEQ ID NO:8. The stop codon for the HIV-1 env and the initiation codon for the STV _m nef are underlined and labeled with arrows.

In one embodiment the 5' proviral clones can be obtained from the known p239 SpSp 5' plasmid, which consists entirely of sequences from the SIV _mac 239 clone. Such sequences can be cloned into another plasmid. For example, the pBs (+) plasmid (Stratagene) modified to contain a unique Clal site in the polylinker region. 5' portions corresponding to the 5' portion of the vector can be readily prepared from other SIV strains using known plasmids by a similar methodology. This clone can be used to generate different vectors referred to herein as SHIV viruses. For example, one containing the SIV _mac 239 tat splice acceptor and tat initiation codon. One can use- known methods to eliminate or alter such sequences. For example, by use of sjLte-directed mutagenesis, you can create a modified clone in which the tat splice acceptor and tat initiation codon are modified as shown in Figure IB. See, particularly in Figure IB, the clones labeled

SHIV-3 and SHIV-4 (SEQ ID N0:7). Other alterations to the nucleic acid sequences to modify them to result in a "modified sequence" are permissible as long as they do not inactivate the protein produced by the sequences.

The 3' portion of these vectors preferably contains the HIV-1 tat gene, the HIV-1 rev gene, and the HIV-env gene. In one preferred embodiment, this portion also contains an HIV-1 vpu gene. The tat, rev and env genes can be derived from any HIV-1 strain, for example, HXBc2. The 3' portion also preferably contains the SIV nef gene and the SIV LTR. Preferably, the SIV LTR is the 3' SIV LTR.

The nef sequence, preferably the nef gene, encodes a functional nef protein. The sufficient number of nucleotides needed to encode such a functional protein can readily be determined by the person of ordinary skill in the art in light of the present disclosure. Thus, for example, when one uses, as the SIV sequence, DNA which corresponds to SIV _c 23 , the skilled artisan would know that one would have to modify the nef sequence to produce a functional nef protein. For example, one would use site-directed mutagenesis to change the 93rd codon of nef from a stop (TAA) to a, for example, Glu (GAA) codon, which will encode a functional nef protein. [Kestler, H.W. , et al., Cell 65:651-662 (1991)]. In vitro functional nef protein can be determined by its ability in primary lymphocytes to enhar replication at least two fold.

The junction between the HIV-1 sequence and the SIV _mac sequence in the 3' portion of the vector can be

formed by any methods of ligating the two segment together. For example, by using a Rsr II site, which can be created by means well known in the art, such as, for example, site-directed mutagenesis in both the HIV-1 and SIV sequence. For example, with HIV-1 HXBc2 strain and SIV _mac 239 (SEQ ID NO:8), Figure 1C shows details of a junction near the Rsr II site (R) .

As aforesaid, it is preferable that both the SIV tat splice acceptor sequence and the HIV tat splice acceptor sequences are not present. This can also be accomplished by means well known in the art. For example, using a unique Sphl site, which can be introduced by site-directed mutagenesis into the HIV region upstream of the HIV-1 tat gene, one can position the site such that the HIV-1 tat splice acceptor sequences will either be included or excluded from the 3' portion of the sequence. Vectors which include the HIV-1 tat splice acceptor are shown in Figure IB and referred to as SHIV-1 and SHIV-3 chimeric viruses, while the vectors which lack the HIV-1 tat splice acceptor were designated as SHIV-2 and SHIV-4 viruses.

It is also important that the vector encodes a functional vpr protein. Since the last few codons of the SIV _fflac vpr gene are located 3' to the natural Sph I site in the 5' proviral clone, these codons were supplied in the vector by modification of the 3' portion near the introduced Sph I site. Thus, the vpr reading frame is restored upon ligation of the 5' and 3' sequences at the Sph I site. See, Figure 1.

In one embodiment the vector can also contain a sufficient number of nucleotides encoding a protein other than from HIV-1 or SIV, for example, a marker protein such as CAT. However, because the inclusion of such additional DNA can slow down replication efficiency, its inclusion is dependent upon the particular use of the vector.

The vector can be transfected into cells by standard techniques. For example, the vector can be digested with Sph I and other restriction enzymes that recognize the flanking sequences, such as Cla I for the 5' portion and Xho I fcr the 3' section. The fragments containing the 5' and 3' sequences are ligated. The ligation reaction is mixed with cells in a standard mediu:- such as 1 milliliter of serum-free RPMI 1640 and DEAE-dextran. The cell-I suspension is then incubated at appropriate condition- , for example, 37°, for a sufficient time for transfection, for example, 1 hour. The cells are then washed and resuspended in the medium.

Virus production in these cells is monitored periodically, for example, every three to four days by a standard assay, such as reverse transcriptase assay. For example, as taught by Rho, H., et al. , Virology 112:355-360 (1981).

These infected cells can then be used to obtain virus which can be used to infect an animal. One can infect an animal, for example, a macaque, by standard techniques, such as inoculation intravenously with virus stock. Monkeys such as rhesus monkeys and macaques are the

preferred test animal. However, other mammals susceptable to SIV infection, preferably primates can be used, for example, baboons. One would also mock infect with a mock virus or SIV or HIV-1 to monitor disease progression. Following inoculation of the animal, such as a cynomolgus monkey, PMBCs can be isolated and cultured and the level of a marker protein, such as, for example, SIV gag p27 antigen in culture assayed by known me ns, such as that described by Miller, M.D., et al, J. Immunol. 144:122-128 (1990). The vectors described herein not only will infect the test animal, but also should result in the establishment of disease.

Another means of infecting monkeys with the vector is to inject the vector DNA intramuscularly into the animals. (Letvin et al., Nature)

These vectors create replication competent SIV-HIV-1 hybrid viruses that will express HIV-1 envelope glycoproteins as well as the HIV-1 regulatory proteins tat and rev in a variety of primate species, such as monkeys or apes. The rate of appearance of virus in the peripheral blood mononuclear cells of infected monkeys using the- present vector is comparable to that of the rate of infection with a pathogenic strain of SIV, such as SIV _mac 239. Based upon the results thus far obtained, it is expected that these viruses express functional vif. vpx. vpr and nef regulatory proteins of SIV and the tat and rev regulatory proteins of HIV-1. When an HIV-1 vpu gene, which expresses a functional vpu gene product is present, functional vpu protein is also expressed. Accordingly,

these results indicate that the restriction of HIV replication in monkeys, such as cynomolgus monkeys is not due to determinants in the tat, rev or envelope proteins.

Cynomolgus monkeys innoculated intravenously with a moderate dose, for example, about 7,000 TCIDc .of the above-described SHIV virus (vpu-) became infected with the virus. Based on the time course of virus production from CD8-depleted lectin-stimulated PBMCs, the efficiency of infection during the acute phase, which was initially about six weeks, was comparable to that seen for the pathogenic SIV _mac 239 isolates. Thereafter, the level of SHIV replication appears to decrease, which is similar to that seen in SIV-infected macaques having an antiviral immune response and in most HIV-I-infected humans. The virus then appears to enter a latency period with SHIV virus isolation being intermittent, but with virus isolated from some of the animals for over one year post-inoculation.

These animals raise an immune response to the proteins present in the chimeric virus. For example, the four SHIV-infected monkeys all raised antibodies which recognized HIV-1 envelope glycoproteins and the serum of all four monkeys neutralized HIV with an efficiency comparable to that of serum from HIV-1-infected humans. This indicates that these vectors can not only be used as an effective means of creating an animal model, but as an effective means of raising immunological response to the HIV components. For example, the vector should be able to be used to generate an immunological response in a recipient animal such as a human. Importantly, some

infected animals generate antibodies that neutralize HIV-1 isolates containing envelope glycoproteins quite divergent from that contained on the SHIV vector (See Figure 5). Thus, SHIV infection can generate more broadly neturalizing antibodies than can be generated by other gpl20-based immunogens tested to date. Accordingly, these vectors can be used as vaccine themself.

Because these vectors do not contain an entire HIV viral genome the use of such vector as a therapeutic to boost immune response in an HIV infected individual or as a vaccine to prevent infection should be safer than using whole inactivated or attenuated virus. Similarly, because these vectors produce a chimeric virus, rather than isolated antigens we expect that the immune generated will be stronger. Confirmatory of this is that some of the infected animals while not showing obvious signs of CD4 depletion have raised broadly neutralizing antibodies. Not only have these antibodies been difficult to raise otherwise, but they are typically seen in HIV-infected humans only very late in disease progression.

One would administer an effective amount of vector or SHIV virus to individuals to generate an immune response. Alternatively, one could use the vector to transfect lymphocytes ex vivo and administer an effective amount of the infected lymphocytes to the individual.

Additional animals have been infected with a SHIV (vpu+) vector, for example, cynomolgus monkeys and rhesus monkeys have been inoculated with a moderate dose of the

vpu positive SHIV virus. These animals have consistently been virus-isolation positive. The results, at this timepoint are comparable, and in fact somewhat better, than that seen with the isogenic vpu negative virus.

Furthermore, we believe that the high level of replication we have obtained with our vectors is a result of their containing functional vpr and nef gene products. Although Shibata, et al. has described a chimeric SIV _mac /HIV-l virus, that expresses the HIV-1 envelope glycoprotein, this virus is defective for both vpr and nef. J. Virol. 65:3514-3520 (1991). Shibata also teaches that neither vpr or nef is necessary for replication. We believe, however, that a virus produced according to the method of Shibata is not as efficient as that described here for both replication and infection in animal models. Preliminary reports indicate that the Shibata et al. virus establishes only a low level, transient infection in a percentage of inoculated monkeys [Sakuragi, S., et al., J. Gen. Virol. 73:2983-2987 (1992)].

It is preferable not to have an HIV-1 tat splice acceptor site in the vector. It appears that the presence of the HIV-1.tat splice acceptor near the 5' SIV _mac /HIV-l junction results in inefficient expression of viral genes.

The present vectors, which can create hybrid viruses, that can infect a wide range of primates such as monkeys, permit a wide variety of tests. For example, one can screen for the ability of vaccines to induce protective immune responses in monkeys to infection by the hybrid

virus which will permit a method for teaching the efficacy against viruses with HIV-1 envelope glycoproteins. This model can also be used to evaluate the ability of polyclonal and monoclonal antibodies to inhibit HIV-1 envelope function in animals, as well as to evaluate therapeutics designed to inhibit any of the HIV-1, tat, rev, env or vpu functions. These vectors permit the development of AIDS-like disease and further enhance the ability to allow dissection of the pathogenic potential of envelope glycoprotein variance, allowing assessment of therapeutic efficacy using clinical end points, which further allow evaluation of the ability of vaccine candidates to modify disease induction. This is addressed in more detail below.

Whether an animal has become infected with the hybrid virus can be determined by monitoring for signs of the disease by standard techniques, such as looking at clinical status. This can be done by standard means well known to the skilled artisan, for example, careful physical examinations on the inoculated animal at periodic intervals, e.g., bi-weekly, monthly, bimonthly. The animals can be monitored to see if there is any weight loss as well as development of lymphadenopathy and splenic and/or hepatic enlargement.

At periodic intervals blood can be drawn and the virus isolated and cultivated by standard means. Absolute peripheral blood lymphocytes subset counts can be assessed at such periodic intervals to determine onset of infection. Humoral response to the virus can also be

assessed. For example, one can determine antibody titer to the virus by, for example, indirect immunofluorescence and/or the presence of antibodies to the various proteins, such as anti-envelope and anti-core antibody response, for example, by radioimmuno-precipitation and gel electrophoresis.

As previously discussed, this model will permit the ability to more fully understand the role of variation within the HIV envelope glycoprotein. In HIV-1-infected humans most of the neutralizing antibodies elicited early in infection are directed against the V3 loop of the gpl20 glycoprotein. Neutralizing antibodies generated later in the course of infection are directed against more conserved epitopes among the HIV-1 isolates. For example, antibodies against the mostly discontinuous epitopes that overlap the CD4 binding site. The present model provides the ability to determine whether similar progression of immune responses will occur in animals infected w_ ⁺ :h the hybrid virus. In contrast to the case with humans, the monkeys will only be infected with one HIV-1 envelope strain. Thus, one can assess whether similar progression of immune responses occur in these animals by standard techniques such as collecting serum. In animals that generate antibodies to the HIV-1 envelope glycoprotein, the presence and titer of the neutralizing antibodies will be determined. One can use virus neutralization assays. One can also assess the strain restriction of neutralizing antibodies by using an env gene complementation assay. See for example, Helseth, E. , et al., J. Virol. 64:2416-2420 (1990). Such an assay will allow a precise estimate of the

activity of a serum or monoclonal antibody to neutralize a single round of virus entry into a variety of target cells. In one method, an env-defective HIV provirus encoding a marker such as the CAT enzyme will be co-transfected into a susceptible cell line, e.g. COS-1 cells, with a plasmid expressing the env gene of interest. One would use env genes from any of the various HIV-1 strains. Virons containing only mutant glycoproteins are then harvested from the COS-1 cell supernatants 48 hours after transfection, filtered, incubated with serum or antibody preparation and then placed on the target cells. CAT assays are done on the cells at a periodic time after infection, for example, two days. One can also prepare the claimed vector with different env sequences to determine the effect of different envelope glycoproteins on disease in animals. Thus, one could use a vector containing the env gene having a sequence of, for example, BRU, ELI, BH10, etc.

The env complementation assay can also be employed to address V3 loop specificity and whether there is any strain restriction of any of the observed neutralizing antibodies in infected animal serum. For example, neutralizing activity of the serum against, for instance, the parental HXBc2 envelope glycoproteins can be assessed. A replication-competent HXBc2 mutant envelope glycoprotein with a change of proline 313 at the tip of the V3 loop to serine can be tested in parrallel. By making changes such as this, one has been able to dramatically reduce the sensitivity of HXBc2 virus to neutralization by a variety of anti-V3 loop monoclonal antibodies. The ratio of the

ability of an animal serum to neutralize recombinant virus with the change in proline 313 to the ability to neutralize viruses with the parental HXBc2 envelope glycoprotein indicates the percentage of neutralizing activity in the serum that is directed against the tip of the V3 loop. A second approach that can be used to assess the effect of V3-directed antibodies to the neutralizing activity of infected animal serum, involves competition with peptides corresponding in sequence to the V3 loop. The competing V3 loop peptide must correspond in sequence to that of the envelope glycoprotein being utilized in the env complementation assay. Increasing concentrations of peptide, as well as a control peptide of scrambled sequence will be added to the animal serum prior to incubation with the recombinant virus. Peptides will also be incubated with neutralizing human monoclonal antibodies directed against epitopes outside of the V3 loop to permit the ability to control for non-specific effects of the peptide on the assay. The ratio of neutralizing activity in the presence of the highest concentration of peptide to the neutralizing activity observed in the absence of peptide represents the fraction of the patient activity -attributable to the V3 loop.

These vectors also provide a method for determining the specific epitopes of the envelope glycoproteins recognized by cytotoxic T-lymphocytes (CTL). In humans, single amino acid changes in the HIV-1 envelope glycoprotein can result in loss of recognition. However, with patient sera, it is not practical to generate a series of overlapping peptides representing an envelope glycoprotein of the predominant

HIV isolate to allow a meaningful mapping of the env epitopes recognized by CTL in that individual. Because these animals will be infected with a single envelope species, it is possible to assess the evolution and epitope specificity of HIV-1 env-specific CTL. For example, one can change the env sequence in the vector generating the virus that will be used to infect the animal and assess its effect.

Furthermore, since we are starting with a single virus, by periodically assaying the envelope glycoprotein as well as the infected RNA and DNA, for example, by polymerase chain reaction (PCR) , one can determine the occurrence, if any, of variation for virus in the animal. In one embodiment, virus is isolated periodically, for example, every month and tested. Tests can be, for example, the ability of serum from each time point to neutralize viruses taken at different times from the same animal. Where neutralization escape is observed, one can look at the sequence variation in the gene by the use of polymerase chain application. The results can be compared with samples of the gene obtained prior to seroconversion.

A method for evaluating the effect of the HIV-1 envelope glycoprotein and tropism is also possible with these vectors. For example, by employing two different envelope genes one can evaluate the effect of the different envelope genes both is vivo and in vitro. In one preferred embodiment, one can construct chimeric genes, such that one will obtain different gpl20 and gp41 ectodomain while retaining similar transmembrane and intracytoplasmic

tails. For example, the ADA envelope glycoproteins have been shown to infect primary human macrophages, but not to Infect or form syncytia with established T-cell lines. [Westervelt, P., et al. Proc. Natl. Acad. Sci. U.S.A. 88:3097-3101 (1991)]. Whereas the HXBc2 envelope glycoprotein are typical of those variants that allow efficient infection of established T-cell lines, induce the formation of synctytia, establish high multimeric affinity for CD4 and exhibit sensitivity to soluble CD4. [Sodroski, J. et al., Nature 322:470-474 (1986); Thali, M. et al., J. Virol. 65:5008-5012 (1991)]. By Inserting the Kpn I to Bsm I fragment of the ADA env gene into the analagous site of the HXBc2 env gene one will obtain a chimeric envelope sequence which will result in essentially the entire gpl20 and gp41 ectodomains derived from ADA and the transmembrane region and the intracytoplasmic tail derived from HXBc2. Using one of the vectors containing such a gene, hybrid virus can be produced which expresses this envelope glycoproteins. These proteins have been shown to retain the replicative and tropic phenotype associated with the complete ADA glycoprotein. These viruses can be tested in vitro for ability to replicate on PBMCs and primary macrophages of human and cynomolgus monkeys. Preferably, one would use primary macrophage derived from human peripheral blood. Preferably, the cynomolgus monkey macrophages can be derived from bronchoalveolar lavages and from bone marrow as described by Ringler, D. , et al., J. Med. Primatol. 18:217-226 (1989) and Watanabe, M. , et al., Nature 337:267-270 (1989). One will use standard means to determine whether the HIV ADA envelope glycoproteins allow entry into the cynomolgus monkey macrophages, e.g., tissue

culture observation for syncytria. In another embodiment animal models are used, one injects half the animal models with hybrid virus containing the ADA envelope chimeric genes and half with virus containing HXBc2 envelope genes. The monkeys are then evaluated to determine the differences between the two groups. For example, by looking at viral tropism, neutralizing antibody responses, envelope sequence drift, viral burden and pathogenicity. This can be done by standard techiques. One can also look at early and late antibody responses in these animals to determine whether the different envelope glycoproteins result in qualitative difference in anti-envelope antibody responses generated. Furthermore, PCR-based approaches, such as discussed above, can also be used to determine the evolution of env-sequence variation in these two groups of animals.

One can also use the animal model to assess the role of various regulatory proteins. For example, inserting functional vpu proteins into the HIV segment and comparing the replication of viruses that express vpu protein against those viruses that don't express such protein (Control virus) in vivo and In vitro. For example, the vpu positive and control virus can be analyzed in the animal models, for example, cynomolgus monkeys for relative rates of replication, potential differences in the ratio of cell-associated to cell-free virus in the peripheral blood and for pathogenicity.

One can also use the animal model to more fully understand the effect of the virus on various organs and body systems. For example, intracranial injections, and

screening animals permits the ability to study the effect of the virus on the central nervous system.

One sensitive indicator of a positive role of a gene in virus replication is the tendency to revert minimally altered but non-functional genes back to wild-type sequences. Accordingly, we will also use a vector which contains such a vpu gene, i.e. one where only a few nucleotide changes are required to allow expression of a fully functional vpu product. These changes can be determined by a variety of means, e.g., using vpu-antibody to look for vpu protein or by using PCR amplification. Any r Tted vpu sequences can then be recloned into infectious c eerie proviruses.

One can also screen potential compounds for their therapeutic effect on these viruses. Thus, one can administer potential vaccines to the animals and then try to inoculate with the hybrid viruses produced by these vectors, wait and determine whether infection occurs. This assessment can be made by any of the means discussed above. The standard technique would be to take a control group that is inoculated with a mock vaccine and a similar group that Is inoculated with the actual vaccine at various concentrations. Thereafter, one determines whether infection occurs.

Similarly, one can screen infected animals for therapeutic compounds. One inoculates the hybrid virus-infected animals with the test compound and a control (placebo) under standard conditions and screens to

determine whether or not the animals show any differences from the baseline control, which has been mock treated. One looks at the same criteria as one looked at before, such as virus titer, pathogenicity, weight loss or gain, antibody production, etc.

The drug to be tested, the vector or the virus when inoculated can be delivered by any of a number of means, for example, it can be administered by parenteral injection (intramuscular I.M.), intraperitoneal (I.P.), intravenous (I.V.), intracranial (I.C.) or subcutaneous (S.C.)), oral or other routes of administration well known in the art. Parenteral injection is typically preferred. For example, I.C. would be used when looking at drugs effecting the central nervous system. The amount to be tested or inoculated will usually be in the range of about 0.1 mg to about 10 mg/kg of body weight, referred to sometimes as effective amount. A desired dose is suitably administered as one or several sub-doses administered at appropriate intervals throughout the day, or other appropriate schedule.

The material can be administered in any means convenient, for example, it can be mixed with an inert carrier, such as sucrose, lactose or starch. It can be in the form of tablets, capsules and pills. For parental administration, it will typically be injected in a sterile aqueous or non-aqueous solution, suspensions or emulsion in association with a pharmaceutically-acceptable parenteral carrier.

In addition to using the animal models to test compounds and vaccines, one can also screen the adjuvants that are being used with, for example, the vaccines or therapeutic to determine the most appropriate adjuvant. This can be done in a manner similar to that described above for screening therapeutic compounds. In one preferred embodiment, one can test the desired therapeutic or vaccine with a variety of different adjuvants to see the different effects, if any, of the adjuvants.

The present invention is further illustrated by the following examples. These examples are provided to aid in the understanding of the invention and are not to be construed as a limitation, thereof.

Plasmid constructions

The vectors producing the hybrid viruses (sometimes also referred to as chimeric viruses) were constructed using the infectious, pathogenic SIV _mac 239 (ne -open) virus (gag+. pro+. pol+. vif+. vpx+. vpr+. tat+, rev+, env- . nef+) [Kestler, H. , et al., Science 248:1109-1112 (1990); Kestler III, H.W. , Cell 65:651-662 (1991)] and the HXBc2 HIV-1 virus (gag-t-. pro+. pol+. vif+, vpjr- , tat- rev+, vpu-. env+, nef-) [Fisher, A., et al., Nature 316:262-266 (1985)]. All four chimeric viruses (desimated SHIV) used in this study express the gag, pro, pol. -, if. vpr and nef proteins of SIV _mac 239 (nef open) and the tat. rev, and env proteins of HIV-1 (HXBc2) .

Each chimeric provirus clone was propagated in Ej. coli

using two plasmids, one containing the 5' half of the provirus and one containing the 3' half of the provirus.

The 5' proviral clones, derived from the p239 SpSp 5' plasmid [Kestler, H. , et al., Science 248.supra], consisted entirely of sequences from the SIV _ma( ,239 clone. The sequences from the 5' cellular flanking sequences to the unique Sph I site in the SIV _ffiac 239 genome were cloned into a pBS(+) plasmid (Stratagene) modified to contain a unique Cla I site in the polylinker region. This 5' clone, which was used to generate the SHIV-1 and SHIV-2 chimeric viruses, contains the SIV _ma( ,239 tat splice acceptor and tat iniation codon. Site-directed mutagenesis was used to create a modified 5' clone in which the STV_ma„c tat sp ^r lice acceptor and tat initiation codon were modified (Figure IB). The details of the 5' SIV _mac /HIV-l junction near the Sph I site (S) are shown for each of the SHIV chimeric viruses. The positions of the splice acceptors (S.A.) for the SI _mac tat and rev messages and for the HIV-1 tat message are shown above the figure, with the SIV _mac tat initiation codon, SIV _mac vpr stop codon and HIV-1 tat initiation codon underlined and labeled with arrows below. The astericks denote sequence identify and the dots -represent sequences not shown. The horizontal bars represent sequence deletions. The X marks the position of Xba I site in the SHIV-3 and SHIV-4 sequences (SEQ ID N0:7). This modified 5' clone was used to generate the SHIV-3 and SHIV-4 chimeric viruses.

The 3' proviral clones consisted of tat, rev and env sequences derived from the HXBc2 HIV-1 isolate and the nef and 3' LTR sequences derived from the SIV _mac 239 (nef

open) isolate [Kestler III, H.W., et al., Cell 65.supra] . In the SIV _mac 239 (nef open) variant, the 93rd codon of nef is changed from a stop (TAA) to a Glu (GAA) codon, allowing production of a functional nef protein. [Kestler III, H.W., et al., Cell 65.suora.

The HIV/SIV _mac junction in the 3' proviral clones was formed by ligating the HIV-1 and SIV _mac segments using the Rsr II site, which was created by site-directed mutagenesis in both the HIV-1 and SIV _mac 239 (nef open) sequences. See, Figure IC.

To allow efficient ligation of the 5' and 3' proviral halves, a unique Sph I site was introduced by site-directed mutagenesis into the HIV-1 region upstream of the HIV-1 tat gene. This Sph I site was positioned such that the HIV-1 tat splice acceptor sequences would be either included in or excluded from the 3' proviral clones. The 3' clone that included the HIV-1 tat splice acceptor was used to generate the SHIV-1 and SHIV-3 chimeric viruses, while the 3' clone lacking the HIV-1 tat splice acceptor was used to generate the SHIV-2 and SHIV-4 viruses (Figure lA) . The genetic organization of the HIV-1 (HXBc2) , SIV _mac 239 (nef open), or SHIV chimeric viruses is shown, with HIV-1 or SIV _mac -specific sequences designated as white or black boxes, respectively. Genes that are defective in the strains utilized are denoted with an asterick. The 5' SIV _mac HIV-l junction at the Sph I site (S) and the 3' HIV-l/SIV _mac junction at the Rsr II site (R) are shown. The stippled 3' end of vpr of the SHIV virus represents sequences derived from the HIV-1 portion of the chimera

that reconstitute the SIV _mac 239 vpr. Also since the last few codons of the SIV _mac vpr gene are located 3' to the natural Sph I site in the 5' proviral clone, these codons were supplied by modification of the 3' proviral clone near the introduced Sph I site. Thus, the vpr reading frame would be restored upon ligation of the 5'_ and 3' proviral clones at the Sph I site (Figure IB) .

Transfection of CEMxl74 Cells With Chimeric Proviruses

For transfection, 5 micrograms of the 5' and 3' proviral clones were-digested with Sph I and other restriction enzymes that recognize the flanking sequences (Cla I for the 5' proviral clone and Xho I for the 3' proviral clone). The fragments containing the 5' and 3' proviral sequences were ligated. The ligation reaction was then mixed with 3 X 10 ⁶ CEMX174 cells suspended in 1 ml of serum-free RPMI 1640 and 500 ug/ml~DEAE-dextran. The cell-DNA suspension was incubated at 37°C for one hour, after which the cells were washed with serum-free medium and resuspended in 10 ml RPMI 1640 with 10% fetal calf serum.

Reverse Transcriptase Assays

Virus production in transfected or infected cultures was monitored every 3-4 days by reverse transcriptase assays as described, using 1.5 ml of cell-free supernatant [Rho, H., et al., Virology 112:355-360 (1981)]. After removing supernatants for reverse transcriptase assay , cells-were resuspended in a sufficient amount of fresh

medium to maintain the cell density between 10^ and 10 cells/ml.

Infection of Cultured Monkey PBMCs

Typically, 2-4 X 10 ⁷ PBMCs were isolated from 15-30 ml whole blood from cynomolgus monkeys. Cells were isolated using Ficoll-Paque (Pharmacia) and resuspended in RPMI 1640 supplemented with 10% fetal calf serum and either phytohemagglutinin (PHA-C) (Boehringer-Mannheim) or concanavalin A (Con A, type IV, Sigma) at 5 ug/ml. Three ^~ to five days following PHA-C or Con A stimulation, the cells were washed and resuspended in RPMI 1640 with 10% fetal calf serum and 10 U/ml interleukin-2 (human recombinant, Boehringer-Mannheim). Two days later, PBMCs were in ^j ected with 1 X 10 reverse transcriptase units of virus c -.ived from transfected CEMxl74 cells. Three days __ after infection, PBMCs were washed and resuspended in fresh medium.

Reverse transcriptase measurements in cell supernatants were made on days 4, 6, 9 and 13 following infection.

Preparation Of Virus Stocks and TCID ₅₀ Determination

Virus stocks for animal inoculation were prepared in cynomolgus monkey PBMCs and frozen as cell-free supernatants without additives at -70°C. The virus titer was determined by incubation 100 ul of thawed stocks, either undiluted or as 10-fold serial dilutions, in quadruplicate with 1 X 10 ⁵ CEMxl74 cells in 1 ml of

medium. When cultures became confluent, cells were diluted 1/10. The wells were scored for the presence of syncytia after 2 weeks, and the TCIDJ in the virus stock calculated as described [Jawetz, E.^-et al. , pp. 371-385 in

Review of Medical Microbiology. 14th ed. Lange Medical Publications, Los Altos, CA] .

Immunoprecipitation of Infected Cultures

Approximately 2 X 10 ⁶ CEMxl74 cells were infected with HIV-1 (HXBc2 strain), SIV _mac 239 (nef open), or chimeric viruses. The cultures were labeled overnight with

35 S-cysteine 1-2 days prior to the peak of syncytium formation, and cell lysates were precipitated either with serum from an HIV-1 infected AIDS patient or from an

SIV _mac -infected rhesus macaque as described [Thali, M. , et al., J. Virol. 65:6188-6193 (1991)].

Inoculation of Cynomolgus Monkeys With Chimeric Virus

Two male and two female cynomolgus monkeys (M. fascicularis) were incoculated intravenously with 1 ml of virus stock containing 7 X 10 ³ TCID ₅₀ of the SHIV-4 chimeric virus.

Virus Isolation From Inoculated Cynomolgus Monkeys

At two and four weeks following inoculation of cynomolgus monkeys, CD8-depleted, Con A-stimulated PBMCs were cultured from each animal and the level of SIV _mac gag p27 antigen in culture supernatants assessed as described [Miller, M.D., et al., J.

Immunol. 144:122-128 (1990)]. Culture supernatants positive for viral antigen were used to infect CEMxl74 cells, which were labeled and used for immunoprecipitation as described above.

Chimeric Viruses

As discussed above, the sequences used for the construction were derived from the pHXBc2 DNA, a clone prepared from the IIIB strain of HIV-1 [Fisher, A., et al., Nature 116,supra] , and the p239 SpSp 5' and p239 SpE3'/ ^ne f-open plasmids derived from the SIV 239 strain of virus. Injection ^" of cynomolgus or rhesus monkeys with either purified SIV _mac 239 viral DNA or virus derived from this DNA has resulted in both high levels of viremia and an AIDS-like disease [Kestler, H. , et al., Science 248.surra: Kestler, III, H.W. , et al., Cell 65.supra: Letvin, N. , et al., Nature 349:573 (1991)].

Construction of the appropriate chimeric molecules was complicated by significant differences in the regulatory gpnes of the two viruses as well as the complex genetic organization of the primate immunodeficiency viruses [Desrosiers, R.C., et al., AIDS Res. Hum. Retro. 5:465-473 (1989); Guyader, M. , et al., Nature 326:662-669 (1987); Viglianti, G.A., et al., J. Virol. 62:4523-4532 (1988)]. Both HIV-1 and SIV _mac encode the regulatory genes vif. vpr. tat, rev and nef. The regulatory vpu is specified only by HIV-1 [Cohen, E.A. , et al., Nature 344:532-534 (1988); Klimkait, T. , et al., J. Virol. 64:621-629 (19"0) ; Strebel, K. , et al., J. Virol. 63:3784-3791 (1989); Strebel, K. , et al., Science 241:1221-1223 (1988);

Terwilliger, E.F., et al., Proc Nat'l Acad. Sci. U.S.A.

86:5163-5167 (1989); Willey, R. , et al. , J. Virol.

66:226-234 (1992)], whereas vpx is found only in HIV-2 or

SIV [Henderson, L.E., et al., Science 241:199-201 (1988) ; _

Hu, W., et al., Virology 173:624-630 (1989); Kappes, J.C., et al., Virology 184:197-209 (1991)]. By replacing the- tat. rev, and env sequences of SIV _mac 239 by the corresponding sequences of HXB2, the resultant virus contains the LTR gag, pol. vif. vpx. vpr. and nef of

SIVma„c and tat,' rev,' and env of HIV-1.

The initial chimeric virus made, designated SHIV-1 (SIV-HIV-chimeric virus-1), contains two tat splice acceptor sequences. The 5' tat splice acceptor sequence is of SIV _mac origin (SEQ ID N0:1) whereas the 3' tat acceptor sequence is derived from HIV-1 sequences (SEQ ID N0:3>. To minimize the possibility that the presence of two closely spaced splice acceptor sites might interfere with one another, derivatives of SHIV-1 were made that contain only the SIV _mac splice acceptor site (SHIV-2) (SEQ ID NOS: 1 and 6), only the HIV-1 splice acceptor site (SHIV-3) (SEQ ID NOS:7 and 3) or neither splice acceptor site (SHIV-4) (SEQ ID N0S:7 and 6) (Figure IB). In the virus that lacks both tat splice acceptors, it is believed that the SIV^ _ac rev acceptor substitutes for the tat acceptor.

Replication Of Chimeric Viruses In Culture

The parenal SIV _mac 239 virus replicates well in the human CD4 ⁺ B T cell hybrid line CEMxl74 [Salter, R.D. , et

a ^τ . , Immunogenetics 21:235-246 (1985)]. CEMxl74 cells were '..ansfected with the parental SIV _mac 239 as well as SHIV recombinant DNAs. Virus replication was monitored by measurement of the amount of the virS.1 DNA polymerase (reverse transcriptase) released into the culture medium.

The data of Figure 2 shows that virus is produced from cultures that are transfected with all five DNAs. However, significant differences in the rate of appearance of reverse transcriptase in the medium was noted using different DNAs. Significant virus replication was evident by nine days post-transfection in cultures treated with either the parental SIV _ma( ,239 DNA (0) or the SHIV-2 ( ) or SHIV-4 DNAs ( ) (Figure 2A) . Detectable levels of reverse transcriptase were not present in the cultures transfected with SHIV-1 or SHIV-3 DNAs until day 13 post-transfection (Figure 2B) . The relative delay in appearance of virus in the supernatant of cultures transfected with SHIV-1 ( ) or SHIV-3 ( ) as compared to those transfected with SIV _m mac239,' SHIV-2 or SHIV-4 DNAs was observed in several independent experiments. Despite this reproducible delay, the rates of replication of all four chimeric viruses were indistinguishable when similar amounts of virus harvested from the supernatant fluids of the transfected cultures we.e used to reinfect CEMxl74 cells (data not shown) .

The ability of SHIV-2 and SHIV-4 viruses to initiate infection in primary peripheral blood mononuclear cells (PBMCs) derived from cynomolgus monkeys was examined. For these experiments the SIV _mac 239, SHIV-2 and SHIV-4

viruses harvested from the supernatant fluids of tranfected CEMxl74 cells were incubated with PHA-1 or Con A- activated monkey PBMCs. Three days after infection with these viruses, the PBMCs were washed and resuspended in fresh medium. Virus replication was measured by detection of reverse transcriptase activity in culture supernatant fluids.

Table 1.

Reverse Trascriptase Activity (cpm/1.5 ml x 10 ) In Supernatants of Cynomolgus Monkey PBMCs

Days After Infection

Virus , 4 6 9 13

SIV _mac 239 (nef-open) 33 45 30 61

The data of Table 1 show that all three viruses replicated well in cultures of PBMCs derived from cynomolgus monkeys. The rate of replication and amount of virus produced upon infection of the monkey PBMCs with either the SHIV-2 or SHIV-4 virus was similar to that obtained upon infection of the culture with SIV _mac 239.

Chimeric Nature Of The Recombinant Viruses

The SHIV chimeras produce gag and pol products of SIV _mac and env proteins of HIV-1. The viral gag proteins _ of HIV-1 and SIV _mac 239 can be distinguished by mobility differences on SDS-polyacrylamide gels, following - precipitation with sera from HIV-1 infected humans of SIV _mac -infec ed monkeys. Such sera contain antibodies that cross-react with gag but not with env proteins [Kanki, P., et al., Science 228:1199-1201 (1985)].

Viruses harvested from the supernatant fluids of infected PBMC cultures were used to infect CEMxl74 cells.

As controls,' CEMxl74 cells were infected with SIV„mac239

(nef open) and HIV-1 (HXBc2) viruses. The infected cells were labeled with 35S-cysteine, lysed, and the viral proteins precipitated with serum from an HIV-1-Infected

AIDS patient or serum from a SIV -infected macaque.

The precipitates were analyzed on SDS-polyacrylamide gels.

The data of Figure 3A show that, both the human and monkey sera recognize gag proteins of the parental HIV-1 and SIV viruses. CEMxl74 cells were infected with SIV _mac 239 (nef open) virus or SHIV-4 virus that had been produced from cynomolgus monkey PBMCs. In parallel, CEMxl74 cells were infected with HIV-1 (HXBc2 strain). Infected CEMxl74 cells and uninfected (Mock) controls were labeled, lysed, and precipitated either with HIV-1 positive human serum or serum from a SIV _mac -infected macaque. The position of the HIV-1 and SIV _mac -specific gag and env products are marked. The molecular weight markers shown

are 200, 96, 69, 46 and 30 kD. These proteins can be distinguished from one another by electrophoretic mobility of both the capsid proteins (HIV-1 p24 and SIV _fflac p27) and the gag precursor proteins (HIV-1 p55 and SIV _mac p58). The HIV-1 serum recognizes the gpl60 and gpl20 env glycoproteins present in CEMxl74 cells infected with HIV-1 but not the env proteins of cells infected with SIV _mac 239. The anti-SIV _mac serum recognizes the gpl60 and gpl30 env proteins present in cells infected with SIV _mac 239 but not with the HIV-1 virus.

In these experiments the gag proteins present in cells infected with the SHIV-4 virus exhibited the electrophoretic mobility characteristic of SIV _mac capsid proteins. The env proteins of these extracts were recognized by the anti-HIV-1 but not the anti-SIV _mac serum. The electrophoretic mobilities of the env proteins present in cells infected with the SHIV-4 virus corresponded to those expected for the envelope glycoproteins of HIV-1. These experiments confirm that the

SHIV-4 virus is chimeric and produces the gag proteins of

SIV_ma.„c and the env proteins of HIV-1.

Infection Of Cynomolgus Monkeys

SHIV-4 virus was grown in cynomolgus monkey PBMCs~as described above. The titer of virus produced in the PBMCs was determined using CEMxl74 cells as targets. An amount of virus equivalent to 7 X 10 ³ TCID ₅₀ units was injected intravenously into four cynomolgus monkeys that were seronegative for SIV _mac . At two and four weeks and

at the times indicated in Table 2 below post-infection PBMCs were isolated from the inoculated monkeys. The lymphocyte population was depleted for CD8 ⁺ T cells and activated with Con A as described previously [Miller, M.D. , et al, J. Immunol. 144:122-128 (1990) . ^' Virus was detectable by both p27 gag protein released into the culture fluid and by the formation of syncytia in activated PBMC cultures of all four monkeys at two and four weeks post-infection (data not shown).

TABLE 2

Animal Species Virus Isolation (days post-inoculation

0 16 41 97 134 146 181 237 265 385

128-97 M. fascicularis + +

129-01 M. fascicularis + + + +

132-99 M. fascicularis + +

133-55 M. fascicularis + + +

The culture fluid obtained from the activated PBMCs of the four monkeys was used to infect CEMxl74 cells with the SHIV-4 virus isolated therefrom (numbers 12897 and 13355) at two weeks post-inoculation. In parallel, CEMxl74 cells were infected with HIV-1 (HXBc2 strain) or SIV _mac 239 (nef open) viruses or mock-infected. The cells- were labeled with 35S-cysteine, lysed and precipitated either with anti-HIV-1 and anti-SIV_ma_„c serum as described above. The molecular weight markers shown are 200,96,69 and 46KD. The viruses isolated from all four animals encoded gag precursor proteins that exhibited a mobility identical to that of the SIV gag precursor protein, and encoded env proteins that were precipitated with HIV-1-positive but not SIV _c -positive serum (Figure 3B and data not shown) .

Generation of HIV-1-Neutralizing Antibodies in SHIV-Infected Monkeys

To measure the neutralizing activity present in the serum of the four cynomolgus monkeys infected with the vou-negative SHIV (HXBc2) chimera, serum was incubated (final concentration 1/20) with recombinant HIV-1 encoding chloramphenicol acetyltransferase (CAT) (see Helseth, E. , et al], J. Virol. 64:2416-2420 (1990)). Recombinant HIV-1 viruses containing either the HXBc2 envelope glycoproteins (correspnding to those on the SHIV virus) or the MN envelope glycoproteins were tested. CAT activity was measured in the target cells two days after incubating the recombinant virus with Jurkat lymphocytes. The results indicated that all four SHIV-infected animals produced virus-neutralizing antibodies. The time course of generation of HIV-1-neutralizing antibodies in one of the

infected animals is shown in Figures 4 and 5.

Infection of Macaoues with the vpu-positive SHIV virus

Vpu-positive SHIV viruses containing the HXBc2 HIV-1 envelope glycoprotein were prepared using the above-described virus wherein a start codon has been inserted immediately upstream and in proper reading frame wi*- the vpu open reading frame at a nucleotide corresponding to the point immediattely before HIV nucleotide 5541 using standard techniques. This vpu-positive SHIV virus was used to infect macaques by the methods described above. They were propagated in cynomologus monkey PBMCs and 7000 TCIDCQ of this virus preparation were inoculated intravenously into four cynomologus monkeys (M. fascicularis) (Table 3) and two rhesus monkeys (M. mulatta) (Table 4) . Viruses were isolated from CD8-depleted PBMCs of these animals, as shown in the following tables.

Table 3

Animal Species Virus Isolation (days post-inoculation)

Table 4

Animal Species Virus Isolation (days post-inoculation)

25 46 75

337-91 M. mulatta

421-90 M. mulatta

It Is evident that those skilled in the art, given the benefit of the foregoing disclosure, may make numerous modifications thereof and departures from the specific embodiments described herein, without departing from the inventive concepts and the present invention to be limited solely by the scope and spirit of the appended claims.