CA2245646A1 - Synthetic hiv genes - Google Patents

Abstract

Synthetic DNA molecules encoding HIV genes and modifications of HIV genes are provided. The codons of the synthetic molecules use codons preferred by the projected host cell. The synthetic molecules may be used as a polynucleotide vaccine which provides effective immunoprophylaxis against HIV infection through neutralizing antibody and cell-mediated immunity.

Description

CA 02245646 l99X-08-06 W O 97/31115 PCT~US97/02294 TITLE OF THE rNVENTION
SYNTHETIC HIV GENES

C~OSS-REFERENCE TO RELATED APPLICATIONS
This is a nonprovisional application related to U.S. Serial No. 60/012,0g2, filed February 22, 1996.

STATEMENT REGARDING ~EDE~ALLY-SPONSOKED R&D
Not applicable.
REFERENCE TO MICROFICHE APPENDI~
Not applicable.

FIELD OF THE lNVENTION
HIV Vaccines.

BACKGROUND OF THE INVENTION
1 . HIV Infection:
Human Immunodeficiency Virus-l (HIV-l) is the 20 etiological agent of acquired h~lm~n immune deficiency syndrome (AIDS) and related disorders. HIV-l is an RNA virus of the Retroviridae family and exhibits the 5'LTR-gag-pol-enl~-LTR3' organization of all retroviruses. ln addition, HIV-1 comprises a handful of genes with regulatory or unknown functions, including the tat and 25 ~ ev genes. The env gene encodes the viral envelope glycoprotein that is translated as a 160-kilodalton (kDa) precursor ~gpl60) and then cleaved by a cellular protease to yield the external 1 20-kDa envelope ~ glycoprotein (gpl20) and the transmembrane 41-kDa envelope glycoprotein (gp41). Gpl20 and gp41 remain associated and are 30 displayed on the viral particles and the surface of HIV-infected cells.
Gp 120 binds to the CD4 receptor present on the surface of helper T-lymphocytes, macrophages and other target cells. After gpl20 binds to CD4, gp41 mediates the fusion event responsible ~or virus entry.

CA 0224~646 1998-08-06 W O97/31115 PCT~US97/02294 Infection begins when gpl20 on the viral particle bind,s to the CD4 receptor on the surface of T4 lymphocytes or other target cells. The bound virus merge.s with the target cell and reverse transcribes its RNA genome into the double-stranded DNA of the cell.
S The viral DNA is incorporated into the genetic material in the cell's nucleus, where the viral DNA directs ~e production of new viral RNA, viral proteins, and new virus particles. The new particles bud from ~he target cell membrane and infect other cells.
Destruction of T4 Iymphocytes, which are critical to 10 immllne defen,se, is a major cause of the progressive immune dysfunction that is the h~lim~rk of HIV infection. The los,s of target cells seriously impairs the body's ability to fight most invaders, but it has a particularly severe impact on the defenses against viruses, fungi, parasites and certain bacteria, including mycobacteria.
HIV-1 kills the cells it infects by replicating, budding from them and ~l~m~ging the cell membrane. HIV-l may kill target cells indirectly by means of the viral gpl20 that i,s displayed on an infected cell's surface. Since the CD4 receptor on T cells has a strong affinity for gpl20, healthy cells expressing CD4 receptor can bind to gpl20 and 20 fuse ~,vith infected cells to form a syncytium. A syncytium cannot survive.
HIV-l can also elicit normal cellular immune defenses against infected cells. With or without the help of antibodies, cytotoxic defensive cells can destroy an infected cell that displays viral proteins on 2~ its surface. Finally, free gpl20 may circulate in the blood of individuals infected with HIV-l. The free protein may bind to the CD4 receptor of uninfected cells, m~kin~ them appear to be infecled and evoking an immune response.
Infection with HIV-1 is almost always fatal, and at present 30 there are no cures for HIV-I infection. Effective vaccines for prevention of HIV- I infection are not yet available. Becau.se of the danger of reversion or infection, live attenuated virus probably cannot be used as a vaccine. Most subunit vaccine approaches have not been CA 0224F,646 1998-08-06 ~ successful at preventing HIV infection. Treatments for H~ infection, while prolonging the lives of some infected persons, have serious side effects. There is thus a great need for effective treatments and vaccines to combat this lethal infection.

2. Vaccines Vaccination is an effective form of disease prevention and has proven successful against several types of viral infection.
Determinirlg ways to present HIV- 1 antigens to the human immune system in order to evoke protective humoral and cellular immunity, is a difficult task. To date, attempts to generate an effective HIV vaccine have not been successful. In AIDS patients, free virus is present in low levels only. Tr~n~micsion of HIV-1 is enhanced by cell-to-cell interaction via fusion and syncytia formation. Hence, antibodies generated against free virus or viral subunits are generally ineffective in elimin~ting virus-infected cells.
Vaccines exploit the body's ability to "remember" an antigen. After fir~t encounters with a given antigen the immune system generates cells that retain an immunological memory of the antigen for an individllal's lifetime. ~ubse4uent exposure to the antigen stimulates the immune response and results in elimin~tion or inactivation of the pathogen.
The immune system deals with pathogens in two ways: by humoral and by cell-mediated responses. ln the humoral response lymphocytes generate specific antibodies that bind to the antigen thus inactivating the pathogen. The cell-mediated response involves cytotoxic and helper T Iymphocytes that specifically attack and destroy ~ infected cells.
Vaccine development with HIV-1 virus presents problems ~ 30 because HIV- 1 infects some of the same cells the vaccine needs to- activate in the immune system (i.e., T4 Iymphocytes). It would be advantageous to develop a vaccine which inactivates HIV before impairment of the immllne system occurs. A particularly suitable type CA 0224~646 1998-08-06 W O 97/31115 PCTrUS97/02294 of HIV vaccine would generate an anti-HIV immune response which recognizes HIV variants and which works in HIV-positive individuals who are at the beginning of their infection.
A major challenge to the development of vaccines a~ainst 5 viru.ses, particularly those with a high rate of mutation such as the human immunodeficiency virus, again.st which elicitation of neutr~li7in~
and protective imm~lne responses is desirable, is the diversity of the viral envelope proteins among different viral isolates or strains.
Because cytotoxic T-lymphocyte.s (CTLs) in both mice and humans are 10 capable of recognizing epitopes derived from conserved interna~ viral proteins, and are thought to be important in the immune response against viruses, efforts have been directed towards the development o~
CTL vaccines capable of providing heterologous protection against different viral strains.
It is known that CD8+ CTLs kill virally-infected cells when their T cell receptors recognize viral peptides associated with MHC clas.s I molecules. The viral peptides are derived from endogenously synthesized viral proteins, regardless of the protein's location or function within the virus. Thus, by recognition of epitopes from 20 con.served viral proteins, CTLs may provide cros,s-strain protection.
Peptides capable of associating with MHC class I for CTL recognition originate from proteins that are present in or pass through the cytoplasm or endoplasmic reticulum. In general, exogenous proteins, which enter the endosomal processing pathway (as in the case of 25 antigens presented by MHC class II molecules), are not effective at generating CD8~ CTL responses.
Most efforts to generate CTL responses have used replicating vectors to produce the protein antigen within the cell or they have focused upon the introduction of peptides into the cytosok These 30 approaches have limit~tions that may reduce their utility as vaccines.
Retroviral vectors have restrictions on the size and structure of polypeptides that can be expressed as fusion protein,s while maintaining the ability of the reco~nbinant virus to replicate, and the effectivene.s.s of CA 0224~646 1998-08-06 W O 97/31115 PCTrUS97/02294 vectors such as vaccinia for subsequent immllni7~tions may be compromised by immune responses against the vectors themselves.
Also, viral vectors and modified pathogens have inherent risks that may hinder their use in humans. Furthermore, the selection of peptide 5 epitopes to be presented is dependent upon the structure of an individual's M~C antigens and, therefore, peptide vaccines may have limited e~fectiveness due to the diversity of MHC haplotypes in outbred populations.

I0 3. DNA Vaccines Benvenis~y, N., and Reshef, L. [PNAS 83, 9551 -9555, (19~6)] showed that CaC12 precipitated DNA introduced into mice intraperitoneally (i.p.), intravenously (i.v.) or intramuscularly (i.m.) could be expressed. The i.m. injection of DNA expression vectors 15 without CaC12 treatment in mice resulted in the uptake of DNA by the muscle cells and expression of the protein encoded by the DNA . The plasmids were m~int~ined episomally and did not replicate.
Subsequently, persistent expre.ssion has been observed after i.m.
injection in skeletal muscle of rats, ~lsh and primates, and cardiac 20 mu,scle oi~ rats. The technique of using nucleic acid,s as therapeutic agents was reported in WO90/11092 (4 October 1990), in which naked polynucleotides were used to vaccinate vertebrates.
It is not necessary for the success of the method that immunization be intramuscular. The introduction of gold 25 microprojectiles coated with DNA encoding bovine growth hormone (BC~H) into the skin of mice resulted in production of anti-BGH
antibodies in the mice. A jet injector has been used to transfect skin, muscle, fat, and mammary ti,ssues of living ~nim~l.s. Various methods for introducing nucleic have been reviewed. Intravenous injection of a 30 DNA:cationic liposome complex in mice was shown by Zhu et al., rScience 261 :209-211 (9 July 1993) to result in systemic expression of a cloned transgene. Ulmer et al., ~Science 259:1745-1749, (1993)]
reported on the heterologous protection against influenza virus infection CA 0224~646 1998-08-06 by intramuscular inJection of DNA encoding influenza virus proteins.
The need for specific therapeutic and prophylactic agents capable of eliciting desired immune responses against pathogens and tumor antigens is met by the instant invention. Of particular importance in this therapeutic approach is the ability to induce T-cell immune responses which can prevent infections or disease caused even by virus strains which are heterologous to the strain from which the antigen gene was obtained. This is of particular concern when dealing with HIV as this virus has been recognized to mutate rapidly and many virulent isolates have been identified ~see, for example, LaRosa et al., Science 249:932-935 (1990), identifying 245 separate HIV isolates~. In response to this recognized diversity, researchers have attempted to generate CTI~s based on peptide immunization. Thus, T:~k:~h~hi et al., [~cience 255:333-336 ~1992)] reported on the induction of broadly cross-reactive cytotoxic T cells recognizing an HIV envelope (gp 160) determin~nt. However, those workers recognized the difficulty in achieving a truly cross-reactive CTL response and suggested that there is a dichotomy between the priming or restimulation of T cells, which is very stringent, and the elicitation of e~fector function, including cytotoxicity, from already stimulated CTLs.
Wang et al. reported on elicitation of immune responses in mice against HIV by intramuscular inoculation with a cloned, genomic (unspliced~ HIV gene. However, the level of immllne responses achieved in these studies was very low. In addition, the Wang et al., DNA construct utilized an essentially genomic piece of HIV encoding contiguous Tat/R~V-gpl60-Tat/REV coding sequences. As is described in detail below, this is a suboptimal system for obtaining high-level expression of the gp 160. It also is potentially dangerous because expression of Tat contributes to the progression of Kaposi's Sarcorna.
3Q WO 93/17706 describes a method for vaccinating an animal against a virus, wherein carrier particles were coated with a gene construct and the coated particles are accelerated into cells of an animal.
In regard to HIV, essentially the entire genome, minus the long te~nina~

CA 0224~646 1998-08-06 repeats, was proposed to be used. That method represents substantial risks for reclpients. ~t is generâlly believed ula~ cons~Fucts o~ H~r should contain less than about 50% of the HlV genome to ensure safety of the vaccine; this ensures that enzymatic moieties and viral regulatory proteins, many of which have unknown or poorly understood functions have been eliminated. Thus, a number of problems remain if a useful hllm:~n HIV vaccine is to emerge from the gene-delivery technology.
The instant invention contemplates any of the known method,s for introducing polynucleotides into living tissue to induce expression of proteins. However, this invention provides a novel immunogen for introducing HIV and other proteins into the antigen processing pathway to efficiently generate HIV-specific CTLs and antibodies. The pharmaceutical is effective as a vaccine to induce both cellular and humoral anti-HIV and HIV neutr~li7ing immune responses.
In the instant invention, the problems noted above are addressed and .solved by the provision of polynucleotide immunogens which, when introduced into an ~nim:~l, direct the efficient expression of HIV
proteins and epitopes without the attendant risks associated with those methods. The imrnune responses thus generated are effective at recognizing HIV, at inhibiting replication of HIV, at identifying and killing cells infected with HIV, and are cross-reactive against many HIV
strains.

4. Codon Usa~e and Codon Context The codon pairings of organisms are highly nonrandom, and differ from organism to organism. This information is used to construct and express altered or synthetic genes having desired levels of translational efficiency, to determine which regions in a genome are protein coding regions, to introduce translationâl pause sites into ~ 30 heterologous genes, and to ascertain relationship or ancestral origin of - nucleotide sequences The expression of foreign heterologous genes in transformed organisms is now commonplace. A large number of CA 0224~646 1998-08-06 _ g _ m~mm~lian genes, including, for example, murine and hllm~n genes, have been successfully inserted into single celled org~nisms. Standard techniques in this regard include introduction of the foreign gene to be expressed into a vector such as a plasmid or a phage and utilizing that 5 vector to insert the gene into an organism. The native promoters for such genes are commonly replaced with strong promoters compatible with the host into which the gene i.s inserted. Protein sequencing machinery permits elucidation of the amino acid sequences of even minute quantities of native protein. From these amino acid sequences, DNA sequences coding for those proteins can be inferred. DNA
synthesis is al,so a rapidly developing art, and synthetic genes corresponding to those inferred DNA sequences can be readily constructed.
Despite the burgeoning knowledge of expres.sion system~s and recombinant DNA, significant obstacles remain when one attempts to express a foreign or synthetic gene in an organism. Many native, active proteins, for example, are glycosylated in a manner different from that which occurs when they are expressed in a foreign host. For this reason, eukaryotic host,s ,such as yeast may be preferred to bacterial hosts for expressing many m~mm~lian genes. The glycosylation problem is the subject of continuing research.
Another problem i,s more poorly understood. Often translation of a synthetic gene, even when coupled with a strong promoter, proceeds much less efficiently than would be expected. The same is frequently true of exogenous genes foreign to the expression organism. Even when the gene is transcribed in a sufficiently efficient manner that recoverable quantities of the translation product are produced, the protein is often inactive or otherwise different in properties from the native protein.
It i.s recognized that the latter problem i,s commonly due to differences in protein folding in various org~ni~ms. The solution to this problem has been elusive, and the mechanisms controlling protein folding are poorly under,stood.

- CA 0224~646 1998-08-06 W O 97/31115 PCTrUS97/02294 The problems related to translational efficiency are believed to be related to codon context effects. The protein coding regions of genes in all organisms are subject to a wide variety of functional constraints, some of which depend on the requirement for 5 encoding a properly functioning protein, as well as appropriate translational start and stop signals. However, several features of protein coding regions have been discerned which are not readily understood in terms of these constraints. Two important classe,s of such features are those involving codon usage and codon context.
~0 It is known that codon l~tili7~tion is highly biased and varies considerably between different org~ni~m~. Codon usage patterns have been shown to be related to the relative abundance of tRNA
isoacceptors. Genes encoding proteins of high versus low abund~nce show differences in their codon preferences. The possibility that biases in codon usage alter peptide elongation rates has been widely discussed.
While differences in codon use are associated with differences in translation rates, direct effects of codon choice on translation have been difficult to demonstrate. Other proposed constraints on codon usage patterns include maximi7in~ the fidelity of translation and optimizing the kinetic efficiency of protein synthesis.
Apart from the non-random use of codons, considerable evidence has accumulated that codon/anticodon recognition is influenced by sequences outside the codon itself, a phenomenon termed "codon context." There exists a strong influence of nearby nucleotides on the efficiency of suppression of nonsense codons as well as missense codons.
Clearly, the abundance of suppressor activity in natural bacterial populations, as well as the use of "terrnination" codons to encode selenocysteine and phosphoserine require that tern~in~tion be context-dependent. Similar context effects have been shown to influence the - 30 fidelity of translation, as well as the efficiency of translation initiation.
Statistical analyses of protein coding regions of E. coli have demonstrate another manifestation of "codon contex~." The presence of a particular codon at one position strongly influences the frequency of -CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 occurrence of certain nucleotides in neighboring codons, and these context constraints differ markedly for genes expressed at high versu.s low levels. Although the context effect has been recognized, the predictive value of the statistical rules relating to preferred nucleotides 5 adjacent to codons is relatively low. This has limited the utility of such nucleotide preference data for selecting codons to effect desired levels of translational efficiency.
The advent of automated nucleotide sequencing e~uipment has made available large quantities of se~luence data for a wide variety 10 of organisms. Understanding those data presents substantial difficulties.
For example, it is irnportant to identify the coding regions of the genome in order to relate the genetic sequence data to protein sequences. In addition, the ancestry of the genome of certain org:~ni~ms is of substantial interest. It is known that genomes of some organisms I5 are of mixed ancestry. Some se~uences that are viral in origin are now stably incorporated into the genome of eukaryotic organisms. The viral sequences themselves may have originated in another substantially unrelated species. An understanding of the ancestry of a gene can be important in drawing proper analogies between related genes and their 20 translation products in other org~ni~m.s.
There is a need for a better understanding of codon context effects on translation, and for a method for determining the appropriate codons for any desired translational effect. There is also a need for a method for identifying coding regions of the genome from nucleotide 25 sequence data. There is also a need for a method for controlling protein folding and for insuring that a foreign gene will ~old appropriately when expressed in a host. Genes altered or constructed in accordance with desired translational efficiencies would be of significant worth.
Another aspect of the practice of recombinant DNA
30 techniques for the expression by microor~ni~m~ of proteins of industrial and pharmaceutical interest is the phenomenon of "codon preference". While it was earlier noted that the existing machinery for gene expression is genetically transformed host cells will "operate" to CA 0224~646 1998-08-06 W O 97/31115 rCT~US97/02294 .

- construct a given desired product, levels of expression attained in a microorganism can be subject to wide variation, depending in part on specific alternative forms of the amino acid-specifying genetic code present in an inserted exogenous gene. A "triplet" codon of four 5 possible nucleotide bases can exist in 6~ variant forms. That these forms provide the message for only 20 different amino acids (as well as transcription initiation and termination) means that some amino acids can be coded for by more than one codon. Indeed, some amino acids have as many as six "redundant", alternative codons while some others 10 have a single, required codon. For reasons not completely understood, alternative codons are not at all uniformly present in the endogenous DNA of differing types of cells and there appears to exist avariable natural hierarchy or "preference" for certain codons in certain types of cells.
As one example, the amino acid leucine is specified by any of six DNA codons including CTA, CTC, CTG, CTT, TTA, and TTG
(which correspond, respectively, to the mRNA codons, CUA, CUC, CUG, CUU, UUA and UUG). Exhaustive analysis of genome codon frequencies for microorganisms has revealed endogenous DNA of }3.
20 coli most commonly contains the CTG leucine-specifying codon, while the DNA of yeasts and slime molds most commonly includes a TTA
leucine-specifying codon. In view of this hierarchy, it is generally held that the likelihood of obtaining high levels of expression of a leucine-rich polypeptide by an E. coli host will depend to some extent on the 25 frequency of codon use. For example, a gene rich in lTA codons will in all probability be poorly expressed in E. coli, whereas a CTG rich gene will probably highly express the polypeptide. Similarly, when - yeast cells are the projected transforrnation host cells for expression of a leucine-rich polypeptide, a preferred codon for use in an inserted DNA
30 would be TTA.
The implications of codon preference phenomena on recombinant DNA techniques are manifest, and the phenomenon may serve to explain many prior failures to achieve high expression levels of CA 0224~646 1998-08-06 exogenous genes in successful~y transformed host organisms-a less "preferred" codon may be repeatedly present in the inserted gene and the host cell machinery for expression may not operate as efficiently.
This phenomenon suggests that synthetic genes which have been 5 designed to include a projected host cell's preferred codons provide a preferred form of foreign genetic material for practice of recombinant DNA techniques.

5. Protein Traffickin~
The diversity of function that typifies eukarycate cells depends upon the structural differentiation of their membrane boundaries. To generate and m~int~in these structures, proteins must be transported from their site of synthesis in the endoplasmic reticulum to predetermined destinations throughout the cell. This requires that the 1~ trafficking proteins display sorting signals that are recognized by the molecular machinery responsible for route selection located at the access points to the main trafficking pathways. Sorting decisions for most proteins need to be made only once ase they traverse their biosynthetic pathways since their final destination, the cellular iocation 20 at which they perform their function, becomes their permanent residence.
Maintenance of intracellular integrity depends in part on the selective sorting and accurate transport of proteins to their correct destinations. Over the past few years the dissection of the molecular 25 machinery for targeting and localization of proteins has been studied vigorously. Defined sequence motifs have been identified on proteins which can act as 'address labels'. A number of sorting signals have been found associated with the cytoplasmic domains of membrane proteins.

Synthetic DNA molecules encoding HIV env and modifications of HIV env are provided. The codons of the synthetic molecules include the projected host cell's preferred codons. The CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 ~ synthetic molecules provide preferred forms of foreign genetic material. The synthetic molecules may be used as a polynucleotide vaccine which provides effective immunoprophylaxis against HIV
infection through neutralizing antibody and cell-mediated immunity.
5 This invention provides polynucleotides which, when directly introduced into a vertebrate in vivo, including m~mm~ls such as primates and humans, induce the expression of encoded proteins within the ~nim~l 10 ~RIEF DESC~R~PTION OF THE DRAWINGS
Figure 1 shows Hrv env cassette-based expression strategies.
Figure 2 shows DNA vaccine mediated ant~-gpl20 responses.
Figure 3 shows anti-gpl20 ELISA titers of murine DNA
vaccinee sera.
Figure 4 shows the relative expression of gpl20 after HIV
env PNV cell culture transfection.
Figure 5 shows the mean anti-gpl20 ELISA re.sponses following tPA-gpl43/optA vs. optB DNA vaccination.
Figure 6 shows the neutralization of HIV by murine DNA
vaccinee sera.
Figure 7 shows HIV neutralization by sera from murine HIV env DNA vaccinees.
Figure 8 is an immunoblot analysis of optimized HIV env DNA constructs.

DETAI~ED DESCRIPTION OF THE INVENTION
Synthetic DNA molecules encoding HIV env and synthetic - 30 DNA molecules encoding modified forms of HIV env are provided.
The codons of the synthetic molecules are designed so as to use the codons preferred by the projected host cell. The synthetic molecules may be used as a polynucleotide vaccine which provides effective CA 0224~646 1998-08-06 WO 97/31115 PCT~US97/02294 immunoprophylaxis against HIV infection through neutralizing antibody and cell-mediated immllnity. The synthetic molecules may be used as an immunogenic composition. This invention provides polynucleotides which, when directly introduced into a vertebrate in vivo, including 5 m~mm~ls ,such as primate,s and hllm~ns, induce the expression of encoded proteins within the ~nimAI.
As used herein, a polynucleotide is a combination of nucleic acids which contain regulatory elements such that upon introduction into a living, vertebrate cell, the polynucleotide is able to direct cellular 10 m~chinery to produce translation products encoded by the nucleic acids comprising the polynucleotide. In one embodiment of the invention, the polynucleotide is a polydeoxyribonucleic acid comprising at least one HIV gene operatively linked to a transcriptional promoter. In another embodiment of the invention, the polynucleotide vaccine (PNV) 15 comprises polyribonucleic acid encoding at least one HIV gene which is amenable to translation by the eukaryotic cellular machinery (ribosomes, tRNAs, and other translation factors). Where the protein encoded by the polynucleotide i,s one which does not normally occur in that animal except in pathological conditions, (i.e., a heterologous 20 protein) such as proteins associa~ed with human ~inmunodeficiency virus, (HIV), the etiologic agent of acquired immune deficiency syndrome, (AIDS), the ~n1m~s' immllne system is activated to launch a protective immune response. Because these exogenous proteins are produced by the ~nim~l~s' tissues, the expressed proteins are processed 25 by the major histocompatibility system, MHC, in a fashion analogous to when an actual infection with the related organism (HIV~ occurs. The result i,s induction of immune responses against the cognate pathogen.
The polynucleotides of the instant disclosure, when introduced into a biological ,system, induce the expression of HIV
30 proteins and epitopes and the production of a specific in}mune re,sponse.
The induced antibody response i,s specific for the expressed HIV protein and neutralizes HIV. ~n addition, cytotoxic T-lymphocytes (C~TL) which speci*cally recognize and destroy HIV infected cells are induced.

CA 0224~646 1998-08-06 .

The instant disclosure provides methods for using a polynucleotide which, upon introduction into m~mm~ n tissue, induces the expression in a single cell, in vivo, of discrete gene products. The di,sclosure provides a solution which does not require multiple S manipulations of REV dependent HIV genes to obtain R~V-independent genes. The REV-independent expression system described herein is useful in its own right and is a system for demonstrating the expression in a single cell in vivo of a single desired gene-product.
Because many of the applications of the instant invention 10 apply to antiviral vaccination, the polynucleotides are frequently referred to as polynucleotide vaccine(s), or PNV. This is not to say that additional utilities of these polynucleotides, in immune stimulation and in antitumor therapeutics, are considered to be outside the scope of the invention.
In one embodiment, a gene encoding an HIV gene product is incorporated in an expression vector. The vector contains a transcriptional promoter recognized by an eukaryotic RNA polymerase, and a transcriptional terminator at the end of the HIV gene coding se~uence. In a preferred embodiment, the promoter is the cytomegalovirus promoter with the intron A sequence (C~MV-intA), although those skilled in the art will recognize that any of a number of other known promoter.s such as the strong immllnoglobulin, or other eukaryotic gene promoters may be used. A preferred transcriptional terminator is the bovine growth hormone terminator. The combination 2~ of C~MVintA-BGH terminator is particularly preferred.
To assist in preparation of the polynucleotides in prokaryotic cells, an antibiotic resistance marker may be included in the - expression vector; the marker may be under transcriptional control of a prokaryotic promoter so that expression of the marker does not occur - 30 in eukaryotic cells. Ampicillin-resistance genes, neomycin-resistance genes and other pharmaceutically acceptable antibiotic resistance markers may be used. To aid in the production of high levels of the polynucleotide by fermentation in prokaryotic organisms, it may be CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 advantageous for the vector to contain a prokaryotic origin of rep~ication and be of high copy number. A number of commercially-available prokaryotic cloning vectors provide these benefits. It is desirable to remove non-essential DNA sequences and to ensure that the S vectors are not able to replicate in eukaryotic cells. This minimi~e,s the risk of integration of polynucleotide vaccine sequences into the recipients' genome. Tissue-specific promoters or enhancers may be used whenever it is desirable to limit expression of the polynucleotide to a particular tissue type.
In one embodiment, the expression vector pnRSV is used, wherein the Rous Sarcoma Virus (RSV) long terminal repeat (LTR) is - used as the promoter. In another embodiment, V 1, a mutated pBR322 vector into which the CMV promoter and the BGH transcriptional terminator were cloned is used. In another embodiment, the elements 15 of Vl and pUC19 have been been combined to produce an expression vector named VlJ. Into VlJ or another desirable expres.sion vector is cloned an HIV gene, such as gp 120, gp4 1, gp 160, gag, pol, env, or any other HIV gene which can induce anti-lIIV imml-ne responses. In another emobodiment, the ampicil~in resistance gene is removed from 20 VlJ and replaced with a neomycin resistance gene, to generate VlJ-neo into dif~erent HIV genes have been cloned for use according to this invention. In another embodiment, the vector is VlJns, which is the same as VlJneo except that a unique Sfil restriction site has been engineered into the single Kpn 1 site at position 21 14 of V 1 J-neo The 25 incidence of Sfil sites in human genomic DNA is very low (aproximately I site per 100,000 bases). Thus, this vector allows careful monitoring for expression vector integration into host DNA, simply by Sfil digestion of extracted genomic DNA. In a further refinement, the vector is VIR. In this vector, as much non-e.ssential 30 DNA as possible was "trimmed" from the vector to produce a highly compact vector. This vector is a derivative of VlJns. This vector allows larger inserts to be used, with less concern that undesirable sequences are encoded and optimizes uptake by cells.

CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/~2294 - One embodiment of this invention incorporates genes encoding HIV gp 160, gpl20, gag and other gene products from laboratory adapted strain.s of HIV such as SF2, mB or MN. Those skilled in the art will recognize that the use of genes from HIV-2 strains 5 having analogous function to the genes from HIV-1 would be expected to generate immune responses analagous to those described herein for HIV-l constructs. The cloning and manipulation methods for obtaining these genes are known to those slcilled in the art.
It is recognized that elicitation of immune responses against 10 laboratory adapted strains of HIV may llOt be adequate to provide neutralization of primary field isolates of HIV. Thus, in another embodiment of this invention, genes from virulent, primary field isolates of HIV are incorporated in the polynucleotide immunogen.
This is accomplished by preparing cDNA copies of the viral genes and 15 then subcloning the individual genes into the polynucleotide immunogen. Sequences for many genes of many HIV strains are now publicly available on GENBANK and such primary, field i~solates of HIV are available from the National Institute of Allergy and Infectious Diseases (NLAID) which has contracted with Quality ~io}ogical, Inc., [7581 Lindbergh Drive, Gaithersburg, Maryland 20~79] to make these strains available. Such strains are also available from the World Health Org~ni7~tion (WHO) [Network for HIV Isolation and Characterization, Vaccine Development Unit, Office of Research, Global Programme on AIDS, CH-1211 Geneva 27, Switzerland]. From this work those skilled in the art will recognize that one of the utilities of the instant invention is to provide a system for in vivo as well a.s in vitro testing and analysis so that a correlation of HIV sequence diversity with serology of H~V
neutralization, as well as other parameters can be made. Incorporation of genes from primary isolates of HIV strains provides an immunogen - 30 which induces immune response.s against clinical isolates of the viru~s and thus meets a need as yet unmet in the field. Furthermore, as the virulent isolates change, the immunogen may be modified to reflect new sequences a.s necessary.

CA 0224~646 1998-08-06 W O9713111~ PCT~US97/02~94 To keep the terminology consistent, the following convention is followed herein for describing polynucleotide immllnogen constructs: "Vector name-HIV strain-gene-additional elements". Thu,s, a construct wherein the gpl60 gene of the MN strain is cloned into the 5 expression vector VlJneo, the name it is given herein is: "VlJneo-MN-gp 160". The additional elements that are added to the construct are described in further detail below. As the etiologic strain of the vlrus changes, the precise gene which is optimal for incorporation in the pharmaceutical may be changed. However, as i,s demonstrated below, I0 because CTL responses are induced which are capable of protecting against heterologous strains, the strain variability is less critical in the immunogen and vaccines of this invention, as compared with the whole virus or subunit polypeptide based vaccines. In addition, because the pharmaceutical is easily manipulated to insert a new gene, this is an 15 adjustment which is easily made by the standard techniques of molecular biology.
The term "promoter" as used herein refers to a recognition site on a DNA strand to which the RNA polymerase binds. The promoter forms an initiation complex with RNA polymerase to initiate 20 and dr~ve transcriptional activity. The complex can be modified by activating sequences termed "enhancers" or inhibiting sequence~s termed "silencers. "
The term "leader" as used herein refers to a DNA sequence at the S' end of a structural gene which is transcribed along with the 25 gene. The leader usually results in the protein having an N-terminal peptide extension sometimes called a pro-se4uence. For proteins destined for either secretion to the extracellular medium or a membrane, this signal sequence, which is generally hydrophobic, direct,s the protein into endoplasmic reticulum from which it is discharged to 30 the appropriate destination.
The term "intron" as used herein refers to a section of DNA occurring in the middle of a gene which does not code for an amino acid in the gene product. The precursor RNA of the intron is CA 0224~646 1998-08-06 W O 97/31115 PCTrUS97/02294 -- excised and is therefore not transcribed into mRNA nor translated into protein.
The terrn "cassette" refers to the sequence of the present invention which contains the nucleic acid secluence which is to be S expressed. The cassette is similar in concept to a cassette tape. Each cassette will have its own sequence. Thus by interch~nging the cassette the vector will express a different sequence. Because of the restrictions sites at the 5' and 3' ends, the cassette can be easily inserted, removed or replaced with another cassette.
The term "3' untranslated region" or "3' UTR" refers to the sequence at the 3' end of a structural gene which is usually transcribed with the gene. This 3' UTR region usually contains the poly A sequence. Although the 3' UTR is transcribed from the DNA it is excised before translation into the protein.
The term "Non-Coding Region" or "NCR" refers to the region which is contiguous to the 3' UTR region of the structural gene.
The NCR region contains a transcriptional termination signal.
The term "restriction site" refers to a sequence specific cleavage site of restriction endonucleases.
The term "vector" refers to some means by which DNA
fragments can be introduced into a host organism or host tissue. There are various types of vectors including plasmid, bacteriophages and cosmids.
The term "effective amount" means sufficient PNV is injected to produce the adequate levels of the polypeptide. One skilled in the art recognizes that this level may vary.
To provide a description of the instant invention, the - following background on HIV is provided. The human immunodeficiency virus has a ribonucleic acid (RNA) genome, the ~ 30 structure of which is represented in Figure 1. This RNA genome mustbe reverse transcribed according to methods known in the art in order to produce a cDNA copy for cloning and manipulation according to the methods taught herein. At each end of the genome is a long terminal CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 repeat which acts as a promoter. Between these termini, the genome encodes, in various reading frames, gag-pol-env as the major gene products: ga~ is the group specific antigen; pol is the reverse transcriptase, or polymerase; also encoded by this region, in an alternate 5 reading frame, is the viral protease which is responsible ~or post-translational processing, for example, of gpl60 into gpl20 and gp41;
env is the envelope protein; vif is the virion infectivity factor; REV is the regulator of virion protein expression; neg is the negative regulatory factor; vpu is the virion productivity factor "u"; tat is the 10 trans-activator of transcription; vpr is the viral protein r. The function of each of these elements has been described.
- In one embodiment of this invention, a gene encoding an HIV or SIV protein is directly linked to a transcriptional promoter.
l'he enl~ gene encodes a large, membrane bound protein, gpl60, which is post-translationally modified to gp41 and gpl20. The gpl20 gene may be placed under the control of the cytomegalovirus promoter for expression. ~Iowever, gpl20 is not membrane bound and therefore, upon expression, it may be secreted from the cell. As HIV tends to remain dormant in infected cells, it i.s desirable that imnnune responses directed at cell-bound HIV epitopes also be generated. Additionally, it is desireable that a vaccine produce membrane bound, oligomeric ENV
antigen similar in structure to that produced by viral infection in order to generate the most efficacious antibody responses for viral neutralization. This goal is accomplished herein by expression in vivo of the cell-membrane associated epitope, gp l 60, to prime the immune system. However, expression of gp160 i,s repressed in the absence of REV due to non-export from the nucleus of non-spliced gene.~i. For an understanding of this system, the life cycle of HIV must be described in further detail.
In the life cycle of HIV, upon infection of a host cell, HIV
RNA genome is reverse-transcribed into a proviral DNA which integrates into host genomic DNA as a single transcriptional unit. The LT~ provides the promoter which transcribes HIV genes from the 5' to CA 0224~646 1998-08-06 W O 97/3111~ PCTrUS97/02294 -- 3' direction (gag, pol, env), to form an unspliced transcript of the entire genome. The unspliced transcript functions as the mRNA from which gag and pol are translated, while limited splicing must occur for translation of env encoded genes. For the regulatory gene product REV
S to be expressed, more than one splicing event must occur because in the genomic setting, REV and env, as is shown in figure 1, overlap. In order for transcription of env to occur, REV transcription must stop, and vice versa. In addition, the presence of R~V is required for export of unspliced RNA from the nucleus. For REV to function in this 10 manner, however, a REV responsive element (RRE) must be present on the transcript [Malim et al., Nature 338:254-257 (19~9)].
In the polynucleotide vaccine of this invention, the obligatory splicing of certain HIV genes i.s eliminated by providing fill!y spliced genes (i.e.: the provision of a complete open reading frame for 15 the desired gene product without the need for switches in the reading frame or elimin~tion of noncoding regions; those of ordinary skill in the art would recognize that when splicing a particular gene, there is some latitude in the precise sequence that results; however so lo~g as a functional coding sequence is obtained, this is acceptable). Thus, in one 20 embodiment, the entire coding sequence for gpl 60 is spliced such that no intermittent expression of each gene product is required.
The dual humoral and cellular immune responses generated according to this invention are particularly significant to inhibiting HIV
infection, given the propensity of HIV to mutate within the population, 25 as well as in infected individuals. In order to formulate an effective protective vaccine for HIV it is desirable to generate both a multivalent antibody response for example to gpl60 (env is approximately 80%
- conserved across various HIV-I, clade B strains, which are the prevalent strains in US human populations), the principal neutralization - 30 target on HIV, as well as cytotoxic T cells reactive to the conserved portions of gpl60 and, internal viral proteins encoded by gag. We have made an HIV vaccine comprising gpl60 genes selected from common laboratory strains; from predominz~nt, primary viral isolates found CA 0224~646 l99X-08-06 within the infected population; from mutated gpl 60s designed to umnask cross-strain, neutralizing antibody epitopes; and from other representative H~V genes such as the gag andp~l genes (~95%
conserved across H~V isolates.
S Virtually all HTV seropositive patients who have not advanced towards an immunodeficient state harbor anti-gag CTLs while about 60% of the,~e patient~ show cross-strain, gp160-specific ~TLs.
The amount of HIV ,specific CTLs found in infected individuaLs that have progres~ed on to the disease state known as AIDS, however, is 10 much lower, demonstrating the significance of our findings that we can induce cross-strain (~TL responses.
Immune responses induced by our env andgag polynucleotide vaccine constructs are demonstrated in mice and primate.s. Monitorirlg antibody production to env in mice allows 15 confirmation that a given construct is suitably immunogenic, i.e., a high proportion of vaccinated ~nim~ls show an antibody response. Mice also provide the most facile animal model suitable for testing CTL induction by our constructs and are therefore used to evaluate whether a particular construct is able to generate such activity. Monkeys (African 20 green, rhesus, chimpanzees) provide additional species including primate,s for antibody evaluation in larger, non-rodent ;Inim~ls. These species are also preferred to mice for antisera neutralization as.says due to high levels of endogenous neutralizing activities against retroviru,se,s observed in mouse sera. These data demonstrate that suf~icient 25 immllnogenicity is engendered by our vaccines to achieve protection in experiments in a chimpanzee/HIVIIrg challenge model based upon known protective level.s of neutralizing antibodie.s for this system.
However, the currently emerging and increasingly accepted definition of protection in the scientific community is moving away from so-called 30 "sterilizing immunity", which indicates complete protection from HIV
infection, to prevention of disease A number of correlates of this goal include reduced blood viral titer, as measured either by HIV reverse transcriptase activity~ by infectivity of samples of serum, by ELISA

CA 0224~646 1998-08-06 W O 97/3111~ PCT~US97/02294 - assay of p24 or other HIV antigen concentration in blood, increased CD4+ T-cell concentration, and by extended survival rates ~see, for example, Cohen, J., Science 262:1820-1821, 1993, for a discussion of the evolving definition of anti-HIV vaccine efficacy]. The imm~lnogens 5 of the instant invention also generate neutralizing immune responses against infectious (clinical, primary field) i.solates of HIV.

Immunolo~y A. Antibody ~esponses to env.
1. gpl60 and gpl20. An ELISA assay is used to determine whether vaccine vectors expressing either secreted gpl20 or membrane-bound gp 160 are efficacious for production of env-specific antibodies.
Initial in vitro characterization of env expression by our vaccination vectors is provided bu imm~lnoblot analysi,s of gpl60 transfected cell lysates. These data confirm and quantitate gpl60 expression using anti-gp41 and anti-gpl20 monoclonal antibodies to visualize transfectant cell gp 160 expression. In one embodiment of this invention, gpl 60 is preferred to gpl20 for the following reasons: (1) an initial gpl20 vector gave inconsistent immunogenicity in mice and was very poorly or non-responsive in African green Monkeys; (2) gpl60 contributes additional neutralizing antibody as well as CTL epitopes by providing the addition of approximately 190 amino acid residues due to the inclusion of gp41; (3) gpl60 expression is more similar to viral env with respect to tetramer assembly and overall conformation, which may provide oligomer-dependent neutralization epitopes; and (4) we find that, like the success of membrane-bound, influenza HA constructs for producing neutralizing antibody responses in mice, ferrets, and - nonhuman primates anti-gpl60 antibody generation is superior to anti-gpl20 antibody generation. Selection of which type of enl~, or whether - 30 a cocktail of env subfragments, is preferred is determined by the experiments outlined below.

CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 2. Presence and Breadth of Neutralizing Activity.
ELISA positive antisera from monkeys is tested and shown to neutralize both homologous and heterologous HIV strains.
3. V3 vs. non-V3 Neutralizing Antibodies. A major 5 goal for env PNVs is to generate broadly neutralizing antibodies. It has now been shown that antibodies directed against V3 loops are very strain specific, and the serology of this response has been used to define strains.
a. Non-V3 neutralizing antibodies appear to 10 primarily recognize discontinuous, structural epitopes within gpl20 which are responsible for CD4 binding. Antibodies to this domain are polyclonal and more broadly cross-neutralizing probably due to restraints on mutations imposed by the need for the virus to bind its cellular ligand. An ln vitro assay is used to test ~or blocking gpl 20 15 binding to CD4 immobilized on 96 well plates by sera from immunized ~nim~ls. A second in vitro assay detects direct antibody binding to synthetic peptide,s representing selected V3 domains immobilized on plastic. These assays are compatible for antisera from any of the anirnal types used in our studies and define the types of neutralizing antibodie.s 20 our vaccines have generated as well as provide an in vitro correlate to virus neutralization.
b. gp41 harbors at least one major neutralization detetmin~nt, corresponding to the highly conserved linear epitope recognized by the broadly neutralizing 2F5 monoclonal antibody 25 (commercially available from Viral Testing Systems Corp., Texas Commerce Tower, 600 Travis Street, Suite 4750, Houston, TX 77002-3005(USA), or Waldheim Pharmazeutika GmbH, Boltzmangasse 11, A-1091 Wien, Austria), as well as other potential sites including the well-conserved "fusion peptide" domain located at the N-terminus of gp41.
30 Besides the detection of antibodies directed against gp41 by immunoblot as described above, an in vitro assay test is used for antibodies which bind to synthetic peptides representing these domain.s immobilized on plastic.

CA 0224~646 1998-08-06 W O 97/31115 PCTrUS97/02294 - 4. Maturation of the Antibody Response. In HIV
seropositive patients, the neutralizing antibody responses progress from chiefly anti-V3 to include more broadly neutralizing antibodies comprising the structural gpl20 domain epitopes described above (#3), 5 including gp41 epitopes. These types of antibody responses are monitored over the course of both time and subsequent vaccinations.

B. T ~ell Reactivities A~ain,st env and gag.
1. Generation of CTL Responses. Viral proteins which 10 are synthesized within cells give rise to MHC I-restricted CTL
responses. Each of these proteins elicit,s CTL in seropositive patients.
Our vaccines also are able to elicit CTL in mice. The immunogenetics of mouse strains are conducive to such studies, as demonstrated with influenza NP. Several epitopes have been defined for the HIV proteins 15 env, ~EV, nef and gag in Balb/c mice, thus facilitating in vitro CTL
culture and cytotoxicity assays. It is advantageous to use syngeneic tumor lines, such as the murine mastocytoma P815, transfected with these genes to provide targets for CTL as well as for in vitro antigen specific restimulation. Methods for defining immunogens capable of 20 eliciting MHC class I-restricted cytotoxic T lymphocytes are known. A
peptide encompassing amino acids 152-176 was also found to induce HIV neutralizing antibodies], and these methods may be used to identify immunogenic epitopes for inclusion in the PNV of this invention.
Alternatively, the entire gene encoding gp160, gpl20, protease, or,~a,~,~
25 could be used. As used herein, T-cell effector function is associated with mature T-cell phenotype, for example, cytotoxicity, cytokine secretion for B-cell activation, and/or recruitment or stimulation of macrophages and neutrophils.
2. Measurement of T~ Activities. Spleen cell cultures ~ 3~ derived from vaccinated ~nim~l~ are tested for recall to specific antigen.s by addition of either recombinant protein or peptide epitopes.
Activation of T cells by such antigens, presented by accompanying splenic antigen presenting cells, APCs, is monitored by proliferation of these cultures or by cytokine production. The pattern of cytokine production also allow.s classification of TH response as type I or type 2.
Because dominant TH2 responses appear to correlate with the exclusion of cellular imrnunity in immunocompromised seropositive patients, it is 5 possible to define the type of response engendered by a given PNV in patients, permitting manipulation of the resulting immune responses.
3. Delayed Type Hypersensitivity (DTH). DTH to viral antigen after i.d. injection is indicative of cellular, primarily MHC ~I-restricted, immunity. 33ecause of the commercial availability of 10 recombinant HTV proteins and synthetic peptides for known epitopes, DT~ responses are easily determined in vaccinated vertebrates using these reagents, thus providing an additional in vivo correlate for inducing cellular immunity.

1 5 Protection Based upon the above immunologic studies, it is predictable that our vaccines are effective in vertebrates againsts challenge by virulent HIV. These studies are accomplished in an HIVIIIg/chimpanzee challenge model after sufficient vaccination of 20 these ~nim~ with a PNV construct, or a cocktail of PNV constructs comprised of gpl60IIIg, gagIIIg~ nefIIIB and REVIIIg. The IIIB
strain is useful in this regard as the chimpanzee titer of lethal dose.s of this strain has been established. However, the same studies are envisioned using any strain of HIV and the epitopes specific to or 25 heterologous to the given strain. A second vaccination/challenge model, in addition to chimpanzees, i~s the scid-hu PBl, mouse. This model allows testing of the human Iymphocyte immune system and our vaccine with subsequent H~V challenge in a mouse host. This system is advantageous as it is easily adapted to use with any HIV strain and it 30 provides evidence of protection against multiple strains of primary field isolates of HIV. A third challenge model utilizes hybrid HIV/SIV
viruses (SHIV), some of which have been shown to infect rhesus monkeys and lead to immunodeficiency disease resulting in death [see CA 0224~646 1998-08-06 ~ Li, J., et al., 3. AID~ 5:639-6~6, 1992]. Vaccination of rhesus with our polynucleotide vaccine constructs is protective against subsequent challenge with lethal doses of SHIV.

S PNV Construct Summary HIV and other genes are ligated into an expression vector which has been optimized for polynucleotide vaccinations. Essentially all extraneou.s DNA is removed, leaving the essential elements of transcriptional promoter, immunogenic epitopes, transcriptional 10 terminator, bacterial origin of replication and antibiotic resistance gene.
Expression of HIV late genes such as env and gag is REV-dependent and requires that the REV response element (RRE~) be present on the viral gene transcript. A secreted form of gpl20 Gan be generated in the absence of REV by substitution of the gpl20 leader peptide with 15 a heterologous leader such as from tPA (tissue-type plasminogen activator), and preferably by a leader peptide such as is found in highly expressed m~mm~ n proteins such as immunoglobulin leader peptides.
We have inserted a tPA-gpl20 chimeric gene into VlJns which e~ficiently expresses secreted gpl20 in transfected cells (RD, a human 20 rhabdomyosarcoma line). Monocistronic gpl60 does not produce any protein upon transfection without the addition of a REV expression vector.

Representative Construct Components Include (but are not restricted 25 to):
1 . tPA-gp 1 20MN;
3. gp 1 60IIIB;
- 10. ga*IIIB: for anti-gag CTL;
1 1. tPA-gp 1 2oIIIB;
12. gpl60 with structural mutations: Vl, V2, and/or V3 loop deletions or substitutions 20. Genes encoding antigens expressed by pathogens other than HIV, such as, but not limited to, influenza virus CA 0224~646 1998-08-06 W O 97/3111~ PCT~US97/02294 nucleoprotein, hemagglutinin, matrix, neuraminidase7 and other antigenic proteins; herpes simplex virus genes; hllman papillomavirus genes; tuberculosis antigens; hepatitis A, B, or C virus antigens.

The protective efficacy of polynucleotide HIV immunogens against subsequent viral challenge is demonstrated by immunization with the non-replicating plasmid DNA of this invention. This is advantageous since no infectious agent is involved, assembly of virus 10 particles is not required, and deterrninant selection is permitted.
Further~nore, because the sequence of gag and protease and several of the other viral gene products i.s conserved among various strains of HIV, protection against subsequent challenge by a virulent strain of HIV
that is homologous to, as well as strains heterologous to the strain from 1~ which the cloned gene is obtained, is enabled.
The i.m. injection of a DNA expression vector encoding gp 160 results in ~e generation of significant protective imrnunity against subsequent viral challenge. In particular, gpl60-specific antibodies and primary CTLs are produced. Immune responses directed 20 against conserved proteins can be effective despite the antigenic shift and drift of the variable envelope proteins. Because each of the HIV
gene products exhibit some degree of conservation, and because CTL
are generated in response to intracellular expression and MHC
processing, it is predictable that many virus genes give rise to responses 25 analogous to that achieved for gpl60. Thus, many of these genes have been cloned, as shown by the cloned and se4uenced junctions in the expression vector see below) such that these constructs are irnmunogenic agents in available form.
The invention offers a means to in~uce cross-strain 30 protective imrnunity without the need for self-replicating agents or adjuvants. In addition, immnni7~tion with the instant polynucleotides offers a number of other advantages. This approach to vaccination should be applicable to tumors as well as infectious agents, since the CA 0224~646 1998-08-06 W O 97131115 PCT~US97/02294 CD~+ CTL response is important for both pathophysiological processes [K. Tanaka et al., Annu. Rev. Immunol. 6, 359 (1988)]. Therefore, eliciting an immune response against a protein crucial to the transformation process may be an effective means of cancer protection 5 or immunotherapy. The generation of high titer antibodies against expressed proteins after injection of viral protein and human growth hormone DNA suggests that this is a facile and highly effective means of making antibody-based vaccines, either separately or in combination with cytotoxic T-lymphocyte vaccines targeted towards conserved 1 0 antigens.
The ease of producing and purifying DNA constructs compares favorably with traditional methods of protein purification, thus facilitating the generation of combination vaccines. Accordingly, multiple constructs, for example encoding gpl60, gpl20, gp41, or any 15 other HIV gene may be prepared, mixed and co-administered. Because protein expression is maintained following DNA injection, the persi~tence of B- and T-cell memory may be enhanced, thereby engendering long-lived humoral and cell-mediated immunity.
Standard techniclues of molecular biology for preparing 20 and purifying D~A con.structs enable the preparation of the DNA
immunogens of this invention. While standard techniques of molecular biology are therefore sufficient for the production of the products of this invention, the specific constructs disclosed herein provide novel polynucleotide immllnogens which surprisingly produce cross-strain and 25 primary H~V isolate neutralization, a result heretofore lln~tt~inable with standard inactivated whole virus or subunit protein vaccines.
The amount of expressible DNA or transcribed RNA to be introduced into a vaccine recipient will depend on the strength of the transcriptional and translational promoters used and on the 30 immunogenicity of the expressed gene product. In general, an immunologically or prophylactically effective dose of about l ng to I00 mg, and preferably about 10 ,ug to 300 ,ug is administered directly into muscle tissue. Subcutaneous injection, intradermal introduction, CA 0224~646 1998-08-06 W O97/31115 PCTrUS97/02294 impression through the skin, and other modes of al1mini~tration such as intraperitoneal, intravenous, or inh~l~tion delivery are also contemplated. It is also contemplated that booster vaccinations are to be provided. Following vaccination with HIV polynucleotide irrlmunogen, S boosting with ~IV protein immunogens such as gpl60, gpl20, and gag gene products is also contemplated. Parenteral ~rlministration, such as intravenous, intramuscular, subcutaneous or other means of ~-lmini~tration of interleukin-12 protein, concurrently with or subsequent to parenteral introduction of the PNV of this invention is 10 al.so advantageous.
The polynucleotide may be naked, that is, unassociated with any proteins, adjuvants or other agents which impact on the recipients' immllne sytem. In this case, it is desirable for the polycucleotide to be in a physiologically acceptable solution, such as, but not limited to, 15 sterile saline or sterile buffered saline. Alternatively, the DNA may be associated with liposomes, such as lecithin liposomes or other liposomes known in the art, as a DNA-liposome mixture, or the DNA may be associated with an adjuvant known in the art to boost ir~nune responses, such as a protein or other carrier. ~gents which assist in the cellular 20 uptake of DNA, such as, but not limited to, calcium ions, may also be used to advantage. These agents are generally referred to herein as transfection f~cilitating reagents and pharmaceutically acceptable carriers. Techniques for coating microprojectiles coated with polynucleotide are known in the art and are also useful in connection 25 with this invention.
The following examples are offered by way of illustration and are not intended to limit the invention in any manner.

~XAMPLE l 30 Materials descriptions Vectors pF411 and pF412: These vectors were subcloned from vector pSP62 which was conslructed in R. Gallo's lab. pSP62 is an available reagent from Biotech Research Laboratories, Inc. pSP62 has a CA 0224~646 l998-08-06 W O 97/31115 PCT~US97/02294 12.5 kb XbaI fragment of the HXB2 genome subcloned from lambda HXB2. SalI and Xba I digestion of pSP62 yields to HXB2 fragments:
5'-XbaI/SalI, 6.5 kb and 3'- SalI/XbaI, 6 kb. These inserts were subcloned into pUC 18 at SmaI and SalI sites yielding pF411 (5'-S XbaI/SalI) and pF412 (3'-XbaI/SalI). pF411 contains gag/pol and pF412 contains tat/rev/env/nef.

Repli~en rea~ents:
recombinant rev (IIIB), #RP1024-10 rec. gp l 20 (IIIB), #RP 1001 - 10 anti-r~v monoclonal antibody, ~P 1029- 10 anti-gpl20 mAB, #lC1, #RP1010-10 AIDS Research and Reference Rea~ent Program:
anti-gp41 mAB hybridoma, Chessie ~, #526 The strategies are designed to induce both cytotoxic T
Iymphocyte (CTL) and neutralizing antibody responses to HIV, principally directed at the HIV ga* (~95% conserved) and env (gpl60 or gpl20; 70-80% conserved) gene products. gp160 contains the only known neutralizing antibody epitopes on the HIV particle while the importance of anti-env and anti-ga~ CTL responses are highlighted by the known association of the onset of these cellular immunities with clearance of primary viremia following infection, which occurs prior to the appearance of neutr~li7ing antibodies, as well as a role for CTL in maintaining disease-free status. Because HIV is notorious for its genetic diversity, we hope to obtain greater breadth of neutralizing antibodies by including several representative env genes derived from clinical isolates and gp41 (~90% conserved), while the highly conserved gag gene should generate broad cross-strain CTL responses.

CA 0224~646 1998-08-06 W O 97/31115 P~l/U~7/~2294 E~AMPL~ 2 Heterologous Expression of HIV Late Gene Products HIV structural genes such as env and gag require expression of the HTV regulatory gene, rev, in order to efficiently produce full-length proteins. We have found that rev-dependent expression of gag yielded low levels of protein and that rev itself may be toxic to cells. Although we achieved relatively high levels of rev-dependent expression of gpl 60 in vitro this vaccine elicited low levels of antibodies to gpl60 following in vivo immlmi7.~tion with rcv/gpl60 10 DNA. This may result from known cytotoxic effects of rev as well as increased difficulty in obt~ining rev function in myotubules containing hundreds of nuclei (rev protein needs to be in the same nucleus as a rev-dependent transcript for gag or env protein e7cpression to occur).
However, it has been possible to obtain rev-independent expression 1~ using selected modifications of the env gene.gag. Evaluation of these plasmids for vaccine purposes is underway.
In general, our vaccines have uti}ized primarily ~IV (IIIB) env and ~ag genes for optimi7~tion of expression within our generalized vaccination vector, VlJns, which is comprised of a CMV immediate-20 early (IE~) promoter, B~H polyadenylation site, and a pUC backbone.Varying efficiencies, depending upon how large a gene segment is used (e.g., gpl20 vs. gpl60), of rev-independent expression may be achieved for e}~v by replacing its native secretory leader peptide with that from the tissue-specific pla.sminogen activator (tPA) gene and expressing the 2~ resulting chimeric gene behind the CMVI~ promoter with the CMV
intron A. tPA-gpl20 is an example of a secreted gp l 20 vector constructed in this fashion which functions well enough to elicit anti-gpl20 immune responses in vaccinated mice and monkeys.
Because of report.s that membrane-anchored proteins may 30 induce much more substantial (and perhaps more specific for HIV
neutralization) antibody responses compared to secreted proteins as well as to gain additional irnrnune epitopes, we prepared VlJns-tPA-gpl60 and VlJns-rev/~p 160. The tPA-gp 160 vector produced detectable CA 0224~646 1998-08-06 W O 97/3111S PCT~US97/02294 quantities of gpl 60 and gp 120, without the addition of rev, as shown by immllnoblot analysis of transfected cells, although levels of expression were much lower than that obtained for rev/gpl 60, a rev-dependent gpl60-expressing plasmid. Thi~s is probably because inhibitory regions (designated INS), which comCer rev dependence upon the gpl60 transcript, occur at multiple sites within gpl60 including at the COOH-terminus of gp41 (see schematic below). A vector was prepared for a COOH-terminally truncated fo~n of tPA-gpl60, tPA-gpl43, which was designed to increase the overall expression levels of env by elimination of these inhibitory sequences. The gpl43 vector also eliminates intracellular gp41 regions conf~ining peptide motifs (such a,s leu-leu) known to cause diversion of membrane proteins to the Iysosomes rather than the cell surface. Thus? g~l43 may be expected to increase both expression of the env protein (by decreasing rev-dependence) and the efficiency of transport of protein to the cell surface compared to full-length gpl60 where the.se proteins may be better able to elicit anti-gpl60 antibodies following DNA vaccination. tPA-gpl43 was further modified by extensive silent mutagenesis of the rev response element (RRE~) sequence (350 bp) to eliminate additional inhibitory sequences for expression. This construct, gpl43/mutRRE, was prepared in two forms: either elimin~ting (form A) or retaining (form B) proteolytic cleavage sites for gpl20/41. Both forms were prepared because of literature reports that vaccination of mice using uncleavable gpl60 expressed in vaccinia elicited much higher levels of antibodies to gpl60 than did cleavable forms.
A quantitative ELISA for gpl60/gpl20 expression in cell transfectants was developed to determine the relative expression - capabilities for these vectors. In vitro transfection of 293 cells followed by quantification of cell-associated vs. secreted/released gpl20 yielded the following results: (1) tPA-gpl60 expressed ~-lOX less gpl20 than rev/gpl60 with similar proportions retained intracellularly v,s.
trafficked to the cell surface; (2) tPA-gpl43 gave 3-6X greater secretion of gpl20 than rev/gpl60 with only low levels of cell-associated gpl43, confirming that the cytoplasmic tail of gpl60 causes intracellular retention of gpl60 which can be overcome by partial deletion of this sequence; and, ~3) tPA-gpl43/mutRRE A and B gave ~lOX greater expression leve}s of protein than did parental tPA-gpl43 S while elimin~tion of proteolytic processing was confirmed for form A.
Thus, our strategy to increase rev-independent expression has yielded stepwise increases in overall expression as well as redirecting membrane-anchored gpl43 to the cell surface away from Iyso,somes. It is important to note that it should be possible to in,sert gpl20 seyuences derived from various viral isolates within a vector cas.sette containing these modifications which reside either at the NH2-terminus (tPA leader) or COOH-terminus ~gp41), where few antigenic differences exist between different viral strains. In other words1 this is a generic construct which can easily be modified by inserting gpl20 derived from various primary viral isolates to obtain clinically relevant vaccines.
To apply these expression strategies to viruses that are relevant for vaccine purposes and confirm the generality of our approaches, we also prepared a tPA-gpl20 vector derived from a primary HIV isolate (cont:~ining the North American concensus V3 peptide loop; macrophage-tropic and nonsyncytia-inducing phenotypes).
This vector gave high expression/secretion of gpl20 with transfected 293 cells and elicited anti-gpl2Q antibodies in mice demonstrating that it was cloned in a functional form. Primary isolate gpl60 genes will also be used for expression in the same way as for gp l 60 derived from laboratory strains.

Immune Responses to HIV-l env Polynucleotide Vaccines:
African green (AGM) and Rhesus (RHM~ monkeys which received gpl 70 DNA vaccines showed low levels of neutralizing antibodies following 2-3 vaccinations, which could not be increased by additional vaccination. These results, as well as increasing awareness -CA 0224~646 1998-08-06 W O 97/31115 rCTrUS97/02294 , - within the HIV vaccine field that oligomeric gpl60 is probably a more relevant target antigen for eliciting neutralizing antibodies than gpl20 monomers (Moore and Ho, J. Virol. 67: 863 (1993)), have led us to focus upon obt~inin~ effective expression of gpl60-based vectors (see 5 above). Mice and AGM were also vaccinated with the primary isolate derived tPA-gpl20 vaccine. These ~nim~ls exhibited anti-V3 peptide (using homologous sequence) reciprocal endpoint antibody titers ranging from 500-5000, demonstrating that this vaccine design is functional for clinically relevant viral isolates.
The gp 160-based vaccines, rev-gp 160 and tP~-gp 160, failed to consistently elicit antibody responses in mice and nonhuman primates or yielded low antibody titers. Our initial results with the tPA-gpl43 plasmid yielded geometric mean titers > 103 in miGe :~nd AGM following two vaccinations. These data indicate that we have 15 signficantly improved the immunogenicity of gpl60-like vaccines by increasing expression levels. This construct, as well as the tPA-gpl43/mutRRE A and B vectors, will continue to be characterized for antibody responses, especially for virus neutralization.
Significantly, gpl20 DNA vaccination produced potent 20 helper T cell responses in all Iymphatic compartments tested (spleen, blood, inguinal, mesenteric, and iliac nodes) with THl-like cytokine secretion profiles (i.e., g-interferon and ~L-2 production with little or no IL-4). These cytokines generally promote strong cellular immunity and have been associated with maintenance of a disease-free state for 25 HlV-seropositive patients. Lymph nodes have been shown to be primary sites for HIV replication, harboring large reservoirs of virus even when virus cannot be readily detected in the blood. A vaccine which can elicit anti-HIV immune responses at a variety of lymph sites, such as we have shown with our DNA vaccine, may help prevent 30 successful colonization of the Iymphatics following initial infection.
As stated previously, we consider realization of the following obJectives to be essential to maximize our chances for succes,s with this program~ e~ -based vectors capable of generating CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 stronger neutralizing antibody responses in primates; (2) ga~,r and env vectors which elicit strong T-lymphocyte responses as characterized by CTL and helper effector functions in primates; (3) use of env and gag genes from clinically relevant ~IV-l strains in our vaccines and 5 characterization of the immunologic responses, especially neutralization of primary isolates, they elicit; (4) demonstration of protection in an ~nim~l challenge model such a.s chimpanzee/H~V (mB) or rhesus/~HlV
using appropriate optimized vaccines; and, (5) determination of the duration of imrnune responses appropriate to clinical use. Significant 10 progress has been made on the first three of these objectives and experiments are in progress to determine whether our recent vaccination constructs for gp 160 and ~ag will improve upon these initial results.

Vectors For Vaccine Production A. VlJneo EXPRESSION VE~TOR:
It was necessary to remove the ampr gene used for antibiotic selection of bacteria harboring VlJ because ampicillin may 20 not be used in large-scale fermenters. The ampr gene from the pUC
backbone o~ VIJ was removed by digestion with SspI and Eaml 105I
restriction enzymes. The remaining plasmid was purified by agarose gel electrophoresis, blunt-ended with T4 DNA polymerase, and then treated with calf intestinal alkaline phosphatase. The commercially 25 available kanr gene, derived from transposon 903 and contained within the pUC4K plasmid7 wa.s excised using the PstI restriction enzyme, purified by agarose gel electrophoresis, and blunt-ended with T4 DNA
polymerase. This fragment was ligated with the VlJ backbone and plasmids with the kanr gene in either orientation were derived which 30 were designated as VlJneo #'s 1 and 3. Each of these plasrnids was confirmed by restriction enzyme digestion analysis, DNA sequencing of the junction regions, and was shown to produce simil~r 4uantities of plasmid as ~llJ. Expression of heterologous gene products was also CA 0224~646 1998-08-06 W O 97/31115 PCTrUS97/~2294 - comparable to VlJ for these VlJneo vectors. We arbitrarily selected V~Jneo#3, referred to as VlJneo hereafter, which contains the kanr gene in the same orientation as the ampr gene in VlJ as the expre.ssion construct.

B. VIJns E~pression Vector:
An Sfi I site was added to VlJneo to facilitate integration studies. A commercially available 13 base pair Sfi I linker (New England BioLabs) was added at the Kpn I site within the BGH sequence 10 of the vector. VlJneo was linearized with Kpn I, gel purified, blunted by T4 DNA polymerase, and ligated to the blunt Sfi I linker. Clonal isolates were chosen by restriction mapping and verified by sequencing through the linker. The new vector was designated VlJns. Expression of heterologous genes in VlJns (with Sfi I) was comparable to 1~ expression of the same genes in VlJneo (with Kpn I).

C. V lJns-tPA:
In order to provide an heterologous leader peptide sequence to secreted and/or membrane proteins, VlJn was modified to 20 include the human tissue-specific plasminogen activator (tPA) leader.
Two synthetic complementary oligomers were annealed and then ligated into VlJn which had been BglII digested. The sense and antisense oligomers were 5'-GATC ACC ATG GAT GCA ATG AAG AGA GGG
CTC TGC TGT GTG CTG CTG CTG TGT GGA GCA GTC TTC GTT
25 TCG CCC AGC GA-3', SEQ.ID:18:, and 5'-GAT CTC GCT GGG CGA
AAC GAA GAC TGC TCC ACA CAG CAG CAG CAC ACA GCA
GAG CCC TCT CTT CAT TGC ATC CAT GGT-3'. The Kozak sequence is underlined in the sense oligomer. These oligomers have overhanging bases compatible for ligation to BglII-cleaved sequences.
30 After ligation the upstream BglII site is destroyed while the downstream BglII is retained for subsequent ligations. Both the junction sites as well as the entire tPA leader sequence were verified by DNA sequencing.
Additionally, in order to conform with our consensus optimized vector W O97/31115 PCT~US97/02294 VlJns (=VlJneo with an SfiI site), an SfiI restriction site was placed at the KpnI site within the BGH terminator region of VlJn-tPA by blunting the Kpnl site with T4 DNA polymerase followed by ligation wi~h an SfiI linker (catalogue #1138, New England Biolabs). This modification was verified by restriction digestion and agarose gel electrophoresis .

EXAMPLl~ S
I. HIV env Vaccine Con,structs:
10 Vaccines Producing Secreted env-derived Antigen (gpl20 and ~pl40):
Expression of the REV -dependent enl~ gene as gpl20 was conducted a.s follows: gpl20 was PCR-cloned from the MN strain of Hl~ with either the native leader peptide se(luence ~V 1 Jns-gp 120), or as a fusion with the tissue-plasminogen activator (tPA) leader peptide 15 replacing the native leader peptide (V lJns-tPA-gp I 20). tPA-gp 120 expression has been shown to be REV-independent rB.S. Chapman et al., Nuc. Acids Res. 19, 3979 (1991); it should be noted that other leader sequences would provide a similar function in rendering the gpl20 gene REV independent]. This was accomplished by preparing the following 20 gpl20 constructs lltili7ing the above described vectors:

~p 120 Vaccine Con.structs:
A. VlJn.s-tPA-HlVM~ gpl20:
HIVMN gp 120 gene (Medimmune) was PCR-amplified using oligomers designed to remove the first 30 amino acids of the peptide leader sequence and to facilitate cloning into VlJns-tPA creating a chimeric protein consisting of the tPA leader peptide followed by the rem~ining gpl20 se(luence following amino acid residue 30. This 30 design allows for REV -independent gpl20 expression and secretion of soluble gpl20 from cells harboring this plasmid. The sense and antisense PCR oligomers used were 5'-CCC CGG ATC CTG ATC
ACA GAA AAA TTG TGGGTC ACA GTC-3', and 5'-C CCC AGG

CA 0224~646 1998-08-06 W O 97/31115 PCTrUS97/02294 .

- AATC CAC CTG TTA GCG CTT TTC TCT CTG CAC CAC TCT
TCT C-3'. The translation stop codon is underlined. These oligomers contain BamHI restriction enzyme sites at either end of the translation open reading frame with a BclI site located 3' to the BamH~ of the sense 5 oligomer. The PCR product was sequentially digested with BclI
followed by BamHI and ligated into VlJns-tPA which had been BglII
digested followed by calf intestinal ~Ik~line phosphata.se treatment. The resulting vector was sequenced to confirm inframe fusion between the tPA leader and gpl20 coding sequence, and gpl20 expression and 10 secretion was verified by immunoblot analysis of transfected RB cells.

B. V lJns-tpA-H~vTl!TR gp l 20:
This vector is analogous to I.A. except that the HIV IIIB
strain was used for gpl20 sequence. The sense and anti~ense PCR
15 oligomers used were: S'-GGT ACA TGA TCA CA GAA AAA TTG
TGG GTC ACA GTC-3', and S'-CCA CAT TGA TCA GAT ATC TTA
TCT TTT TTC TCT CTG CAC CAC TCT TC-3', respectively. These oligomers provide BclI sites at either end of the insert as well as an EcoRV just upstream of the BclI site at the 3'-end. The S'-terminal BclI
20 site allows ligation into the BglII site of VlJns-tPA to create a chimeric tPA-gpl20 gene encoding the tPA leader sequence and gpl20 without its native leader sequence. Ligation products were verified by restriction digestion and DNA sequencing.

gpl40 Vaccine Con.structs:
These constructs was prepared by PCR similarly as tPA-gp 120 with the tPA leader in place of the native leader, but designed to produce secreted antigen by tern~in:~ting the gene imrnediately NH2-terminal of the transmembrane peptide (projected carboxyterminal amino acid sequence = NH2-.. TNWLWYIK-COOH). Unlike the gpl20-producing constructs, gpl40 constructs should produce oligomeric antigen and retain ~nown gp41-contained antibody W O 97/31115 PCT~US97/02294 neutrali~ation epitopes such as E~LDKWA defined by the 2F5 monoclonal antibody.
Constructs were prepared in two forms (A or B~ depending upon whether the gp I 60 proteolytic cleavage sites at the junction of S gpl20 and gp41 were retained (B) or elimin:~.ted (A) by appropriate amino acid substitutions as described by Kieny et al., (Prot. Eng. 2:
219-255 (19~8)) (wild type sequence = NH2-...KAKRRVVQREKR...COOH and the mutated sequence = NH2-...KAONHVVQNE~Q...COOH with mutated amino acids underlined).
A. VIJns-tPA-gp 140/mutRRE-A/SRV-1 3'-UTR (based on HIV-~IIIB~:
This construct wa.s obtained by PCR using the following sense and antisense PCR oligomers: 5'-CT GAA AGA CCA GCA ACT
15 CCT AGG GAAT TTG GGG TTG CTC TGG-3', SEQ.ID: :, and 5'-CGC AGG GGA GGT GGT CTA GAT ATC 1~ TTA TTT TAT
ATA CCA CAG CCA ATT TGT TAT G-3' to obtain an AvrII/EcoRV
segment from vector IVB (containing the optimized RRE-A segment).
The 3'-UTR, prepared as a synthetic gene segment, that is derived from 20 the Simian Retrovirus-1 (SRV-l, see below) was inserted into an Srfl restriction enzyme site introduced immediately 3'- of the gpl40 open reading frame. This UTR sequence has been described previously as facilitating rev-independent expression of HIV env and ga*.

25 B. VlJns-tPA-gpl40/mutR~E-B/SRV-1 3'-UT~ (based on HIV-lIIIB~:
This construct is similar to IIA excep- rhat the env proteolytic cleavage sites have been retained by us~lg construct IVC as starting material.
C. VIJns-tPA-~p l40/opt30-A (based on HIV-I mR~:
This construct was derived from IVB by AvrII and SrfI
restriction enzyme digestion followed by ligation of a synthetic DNA

CA 0224~646 1998-08-06 W O 97/31115 ~1/U~7/02294 ~ segment corresponding to gp30 but comprised of optimal codons for translation (see gp32-opt below). The gp30-opt DNA was obtained from gp32-opt by PCR amplification using the following sense and anti-sense oligomers: S'-GGT ACA ~CT AGG CAT CTG GGG CTG CTC
S TGG-3'; and, S'-CCA CAT GAT ATC G CCC GGG C TTA TTA TTT
GAT GTA CCA CAG CCA GTT GGT GAT G-3', respectively. This DNA segment was digested with AvrII and EcoRV restriction enzyme.s and ligated into V lJns-tPA-gp 143/opt32-A (IVD) that had been digested with AvrII and SrfI to remove the corresponding DNA segment. The 10 resulting products were verified by DNA sequencing of ligation junctions and immunoblot analysis.

D. VlJns-tPA-~pl40/opt30-B (based on HTv-lTTTR):
This construct is similar to IIG except that the env 15 proteolytic cleavage sites have been retained.

E. V 1 Jns-tPA-~p I 40/opt all-A:
The env gene of this construct is comprised completely o~
optimal codons. The constant regions (Cl, C5, gp32) are those 20 described in IVB,D,H with an additional synthetic DNA segment corresponding to variable regions 1-5 is inserted using a synthetic DNA
segment comprised of optimal codons for translation (see example below based on HIV- 1 MN V 1 -V5).

25 F. V 1 Jns-tPA-~p I 40/opt all-B:
This construct i.s similar to IIE except that the env proteolytic cleavage sites have been retained.
.

G. VlJns-tPA-gF)140/opt all-A (non-IIIB strains):
This construct is similar to IIE above except that env amino acid se~uences from strains other than IlIB are used to determine optimum codon useage throughout the variable (V 1 -V5) regions.

W O 97/31115 PCT~US97/02294 H. VlJn,s-tPA-~pl40/opt all-B (non-IIIB strains):
This construct is similar to ~IG except that the env proteolytic cleavage sites have been retained.

gpl60 Vaccine Constructs:
Constructs were prepared in two forms (A or B) depending upon whether ~e gpl6û proteolytic cleavage sites as described above.

10 A. VlJns-rev/env:
This vector is a variation of the one described in section D
above except that the entire tat coding region in exon I is deleted up to the beginnin~ of the REV open reading frame. VlJns-gpl60IIIg (see section A. above) was digested with PstI and KpnI restriction enzymes 15 to remove the 5'-region of the gpl60 gene. PCR amplification was used to obtaill a DNA segment encoding the firs~REV exon up to the KpnI
site in gpl60 from the HXB2 genomic clone. The sense and antisense PCR oligomers were S'-GGT ACA CTG CAG TCA CCG TCC T ATG
GCA GGA AGA AGC GGA GAC-3l, and S'-CCA CAT CA GGT ACC
20 CCA TAA TAG ACT GTG ACC-3', respectively. These oligomer~
provide PstI and KpnI restriction enzyme sites at the 5'- and 3'- termini of the DNA fragment, respectively. The resulting DNA was digested with PstI and KpnI, purified from an agarose electrophoretic gel, and liga~ed with VlJns-gpl60(PstVKpnI). The resulting plasmid was 25 verified by restriction enzyme digestion.

B . V lJns-~p 160:
HIV~IIb gpl60 was cloned by PCR amplification from plasmid pF412 which contains the 3'-terminal half of the HIvIIIb 30 genome derived from HIVIIIb clone HXB2. The PCR sense and anti,sense oligomer.s were 5'-GGT ACA TGA TCA ACC ATG AGA
GTG AAG GAG AAA TAT CAG C-3'" and 5'-CCA CAT TGA TCA
GAT ATC CCC ATC TTA TAG CAA AAT CCT TTC C-3'"

CA 0224~646 1998-08-06 W O 97131115 PCT~US97/02294 .

respectively. The Kozak sequence and translation stop codon are underlined. These oligomers provide BclI restriction enzyme sites outside of the translation open reading frame at both ends of the env gene. (Bcl~-digested sites are compatible for ligation with BglII-digested 5 sites with subsequent loss of sensitivity to both restriction enzymes.
BclI was chosen for PCR-cloning gpl 60 because this gene contains internal BglII and a.s well as BarnHI sites). The antisense oligomer also inserts an EcoRV site just prior to the BclI site a,s described above for other PCR-derived genes. The amplified gpl60 gene was agarose gel-10 purified, digested with BclI, and ligated to VlJns which had beendigested with BglII and treated with calf intestinal ~lk~qline phosphatase.
The cloned ~ene was about 2.6 kb in size and each junction of gp 160 with VlJns was confirmed by DNA sequencing.

15 C. VlJns-tPA-~pl60 (based on HIv-lTT~R):
This vector is similar to Example l(C) above, except that the full-length gpl60, without the native leader sequence, wa,s obtained by PCR. The sense oligomer was the same as used in I.C. and the antisense oligomer was 5'-CCA CAT TGA TCA GAT ATC CCC ATC
20 TTA TAG CAA AAT CCT TTC C-3'. These oligomers provide BclI
sites at either end of the insert as well as an EcoRV just upstream of the BclI site at the 3'-end. The 5'-terminal BclI site allows ligation into the BglII site of VlJns-tPA to create a chimeric tPA-gpl60 gene encoding the tPA leader sequence and gp 160 without its native leader sequence.
25 Ligation products were verified by restriction digestion and DNA
sequencing.
~ D. V lJns-tPA-~p 160/opt C l/opt41 -A (ba,sed on HIV- 1 TTTR~:
This construct was based on IVH, having a complete ~ 30 optimized codon segment for C5 and gp41, rather than gp32, with an additional optimized codon segment (see below) replacing Cl at the amino te~ninus of gpl20 following the tPA leader. The new C1 segment was joined to the rem~ining gpl43 segment via SOE PCR using CA 02245646 l99X-08-06 W O 97/3111~ PCT~US97/02294 the following oligomers for PCR to synthesize the joined Cl/143 segment: 5'-CCT GTG TGT GAG TTT AAA C TGC ACT GAT TTG
AAG AAT GAT ACT AAT AC-3'. The resulting gpl43 gene contains optimal codon useage except for Vl-V5 regions and has a unique PmeI
restriction enzyme site placed at the junction of C1 and Vl for insertion of variable regions ~rom other HIV genes.
E. V lJns-tPA-~p I 60/opt CI /opt41 -B (based on HIV-l rTTR):
This construct is similar to IIID except that the env proteolytic cleavage sites have been retained.

F VlJn,s-tPA-~pl60/opt all-A (based on HIv-lTTTR~:
The env gene of this construct is comprised completetly of optimal codons as described above. The constant regions (C1, C5, gp32) are those described in IIID,~ which is used as a cassette (employed for all completely optimized gpl60s) while the variable regions, Vl-V5, are derived from a synthetic DNA segment comprised of optimal codons.

G. VlJns-tPA-~pl60/opt all-B:
This construct is similar to IlIF except that the env proteolytic cleavage sites have been retained.

H. V 1 Jns-tPA-~p I 60/opt all-A (non-III13 strains):
This construct is similar to ~IIF above except that env amino acid se4uences from strains other than IIIB were used to determine optimum codon useage throughout the variable (Vl-V5) regions.

I. VlJns-tPA-~p 1 60/opt all-B (non-l[IIB strains~:
This construct is similar to IIIH except that the en proteolytic cleavage sites have been retained.

CA 0224~646 1998-08-06 - EXAMPLl~ 9 gpl~3 Vaccine Constructs:
These construct,s were prepared by PCR ,similarly as other tPA-containing constructs described above (tPA-gpl20, tPA-gpl40, and tPA-gpl60), with the tPA leader in place of the native leader, but designed to produce COOH-terrnin~ted, membrane-bound env (projected intracellular amino acid sequence= NH2-NRVRQGYSP-COOH). This construct was designed with the purpose of combining the increased expression of env accompanying tPA introduction and minimi7ing the possibility that a transcript or peptide region corresponding to the intracellular portion of env might negatively impact expression or protein stability/transport to the cell surface.
Constructs were prepared in two forms (A or B) depending upon whether the gpl~0 proteolytic cleavage sites were removed or retained as described above. The residual gp41 fragment resulting from truncation to gpl43 is referred to as gp32.

A . V 1 Jns-tPA-~p 143:
This construct was prepared by PCR using pla.smid pF412 with the following sense and antisense PCR oligomers: 5'-G~T ACA
TGA TCA CA GAA AAA TTG TGG GTC ACA GTC-3', SEQ.ID: :, and 5'- CCA CAT TGA TCA G CCC GGG C TTA GGG T~A ATA
GCC CTG CCT CAC TCT GTT CAC-3'. The resulting DNA segment contains BclI restriction sites at either end for cloning into VlJns-tPA/BglII-digested with an SrfI site located immediately 3'- to the en open reading frame. Constructs were verified by DNA sequencing of ligation junction,s and immunoblot analysi.s of transfected cells.

B . V 1 Jns-tPA-g~ l 43/mutRRE-A:
This construct was based on TVA by excising the DNA
segment using the unique MunI restriction enzyme site and the downstream SrfI site described above. This segment corresponds to a portion of the gpl20 C5 domain and the entirety of gp32. A synthetic W O 97/31115 PCT~US97/02294 DNA segment corresponding to ~350 bp of the rev response element (RRE A) of gp 160,comprised of optimal codons for translation, was joined to the rem:~ining gp32 segment by splice overlap extension (SO~) PCR creating an AvrII restriction enzyme site at the junction of the two 5 segments (but no changes in amino acid sequence). These PC~ reaction.s were performed using the following sense and antisense PCR oligomers for generating the gp32-cont:~ining domain: 5'-CT GAA A~A CCA
GCA ACT C: CT AGG aAT TTa GGG TTG CTG TGG-3' and S'-CCA
CAT TGA TCA G CCC GGG C TTA GGG TGA ATA GCC CTG C~CT
10 C~AC TCT GTT CAC-3' (which was used as the antisense oligomer ~or IVA), respectively. The mutated RRE (mutRRE-A) segment was joined to the wild type sequence of gp32 by SOE PCR using the following sense oligomer, 5'-GGT ACA CAA TTG GAG GAG CGA GTT ATA
TAA ATA TAA G-3', and the antisense oligomer used to make the 15 gp32 segment. The resulting joined DNA segment was digested with MunI and SrfI restriction enzymes and ligated into the parent gpl43/MunI/Srfl digested plasmid. The resulting construct was verified by DNA sequencing of ligation and SOE PCR junctions and immunoblot analysis of transfected cells.
C. VlJns-tPA-~pl43/mutRRE-B:
This construct is similar to IVB except that the env proteolytic cleavage sites have been retained by using the mutRRE-B
synthetic gene segment in place of mutRR13-A.
2~
D. V I Jns-tPA-~p 143/opt32-A:
This construct was derived from IVB by AvrII and SrfI
restriction enzyme digestion followed by ligation of a synthetic DNA
segment corresponding to gp32 but comprised of optimal codons for 30 translation (see gp32 opt below). The resulting products were verified by DNA sequencing of ligation iunctions and immunoblot analysis.

CA 0224~646 1998-08-06 W O ~7/31115 PCT~US97/02294 E . V 1 Jns-tPA-gp 143/opt32-B:
This construct is similar to IVD except that the env proteolytic cleavage sites have been retained by using IVC as the initial plasmid.
s G. VlJns-tPA-~pl43/~RV-1 3'-UTR:
This construct is similar to IVA except that the 3'-UTR
derived from the Simian Retrovirus-l (SRV-1, see below) was inserted into the Srf~ restriction enzyme site introduced immediately 3'- of the 10 gpl43 open reading frame. This UTR sequence has been described previously as facilitating rev-independent expression of HIV env and gag.

H. VlJns-tPA-gpl43/opt Cl/opt32A:
This construct was based on IVD, having a complete optimized codon segment for C5 and gp32 with an additional optimized codon segment (see below) replacing C1 at the amino terminus of gpl20 following the tPA leader. The new Cl segment was joined to the rem~ining gpl43 segment via SOE PCR using the following oligomers 20 for PCR to synthesize the joined Cl/143 segment: 5'-CCT GTG TGT
GAG TTT AAA C TGC ACT GAT TTG AAG AAT GAT ACT AAT
AC-3'. The resulting gpl43 gene contains optimal codon useage except for Vl-V5 regions and has a unique PmeI restriction enzyme site placed at the junction of Cl and Vl for in.sertion of variable regions from 25 other HIV genes.

I. VlJns-tPA-gpl43/opt Cl/opt32B:
This construct is similar to IVE~ except that the env proteolytic cleavage sites have been retained.

J. VlJns-tPA-gpl43/opt all-A:
The em~ gene of this construct is comprised completely of optimal codons. The constant regions (Cl, C5, gp32) are those W O 97/31115 PCTnUS97/02294 described in 4B,D,H with an additional synthetic DNA segment corresponding to variable regions Vl-V5 is inserted using a synthetic DNA segment comprised of optimal codons for translation.

5 K. VlJns-tPA-~pl43/opt all-B:
This construct is similar to IVJ except that the env proteolytic cleavage sites have been retained.

L. V lJns-tPA-~p 143/opt all-A (non-IIIB ,strains):
This construct i.s similar to IIIG above except that env amino acid sequences from strains other than IIIB were u.sed to determine optimum codon u.seage throughout the variable (V l -VS) regions.

15 M. VlJns-tPA-~pl43/opt all-B (non-IIIB strains):
This construct is similar to IIIG above except that en amino acid sequences from strains other than IIIB were used to determine optimum codon useage throughout the variable (V I -V5 regions.
~ XAMPLE l() ~p 1 43/~lyB Vaccine Constructs:
These constructs were prepared by PCR simIlarly a,s other tPA-containing constructs described above (tPA-gpl20, tPA-gpl40, 25 tPA-gpl43 and tPA-gpl60~, with the tPA leader in place of the native leader, but designed to produce COOH-termin~ted, membrane-bound env as with gpl43. However, gpl43/glyB constructs differ from gpl43 in that of the six amino acids projected to comprise the intracellular peptide domain, the last 4 are the same those at the carboxyl terminus of 30 human glycophorin B (glyB) protein (projected intracellular amino acid sequence= NH2-NRLrKA-COOH with the underlined residues corresponding to glyB and "R" common to both em~ and glyB). This construct was designed with the purpose gaining additional enl~

W O 9713111S PCT~US97102294 .

- expression and directed targeting to the cell surface by completely elimin~ting any transcript or peptide region corresponding to the intracellular portion of env that might nega~ively impact expression or protein stability/transport to the cell surface by replacing this region 5 with a peptide sequence from an abundantly expressed protein (glyB) having a short cytoplasmic domain (intracellular amino acid sequence=
NH2-RRLIKA-COOH). Constructs were prepared in two forms (A or B) depending upon whether the gpl60 proteolytic cleavage sites were removed or retained as described above.
A . V I Jns-tPA-~p 143/opt32-A/~lyB:
- This construct is the same as IVD except that the following antisense PCR oligom.er was u.sed to rep!ace the i~trace!!u!ar peptide domain of gpl43 with that of glycophorin B as described above: 5'-15 CCA CAT GAT ATC G CCC GGG C TTA TTA GGC CTT GAT CAC~
CCG GTT CAC AAT GGA CAG CAC AGC-3'.

B . V l Jns-tPA-~p 143/opt32-B/~lyB:
This construct is similar to VA except that the env 20 proteolytic cleavage sites have been retained.

C. VlJn,s-tPA-~pl43/opt Cl/opt32-A/~lyB:
This construct is the same as VA except that the first constant region ~Cl) of gpl20 is replaced by optimal codons for 25 translation as with IVH.

D. VlJns-tPA-~pl43/opt Cl/opt32-B/~lyB:
This construct is similar to VC except that the env proteolytic cleavage sites have been retained.
E. VlJns-tPA-~pl43/opt all-A/~lyB:
The env gene of this construct is comprised completetly of optimal codons as described above.

W O 97/31115 PCT~US97/02294 F. VIJns-tPA-gpl43/opt all-B/~lyB:
This construct is similar to VE except that the en proteolytic cleavage sites have been retained.

G . V 1 Jns-tPA-gp I 43/opt all -A/glyB (non-IIIB strains):
This construct is similar to IIIG above except that env amino acid sequences frorn strains other than rIIB were u~ed to determine optimum codon useage throughout the variable (Vl-VS) 1 () regions.

H. VlJns-tPA-gpl43/opt all-B/glyB (non-IIIB strains~:
This construct is similar to VG except th~t the env proteolytic cleavage sites have been retained.
HIV en~ Vaccine Con,structs with Variable Loop Deletions:
These constructs may include all env forms listed above (gpl20, gpl40, gpl43, gpl60, gpl43/glyB) but have had variable loops within the gpl20 region deleted during preparation (e.g., Vl, V2, and/or V3).
20 The purpose of these modifications is to eliminate peptide segments which may occlude exposure of conserved neutralization epitopes such as the CD4 binding site. For example, the following oligomer was used in a PCR reaction to create a Vl/V2 deletion resulting in adjoining THE
Cl and ~2 segrnents: 5'-CTG ACC CCC CTG TGT GTG GGG GCT
25 GGC AGT TGT AAC ACC TCA GTC ATT ACA (~AG-3'.

E~AMPLE I 1 Design of Synthetic Gene Se~rnents for Increased env Gene Expre.ssion:
Gene segments were converted to sequences having 30 identical translated sequences (except where noted) but with alternative codon usage as defined by R. Lathe in a research article from J. Molec.
Biol. Vol. 183, pp. 1-12 (198~) entitled "Synthetic Oligonucleotide Probes Deduced from Amino Acid Secluence Data: Theoretical and CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 Practical Considerations". The methodology described below to increase r~?v-independent expression of HIV env gene segments was based on our hypothesis that the known inability to express this gene efficiently in m~mm~ n cells is a con.sequence of the overall transcript S composition. Thus, using alternative codons encoding the same protein sequence may remove the constraint,s on env expression in the absence of rev. Inspection of the codon usage within env revealed that a high percentage of codons were among those infrequently used by highly expressed human genes. The specific codon replacement method employed may be described as follows employing data from Lathe et al.:

1. Identify placement of codons for proper open reading frame.
2. Compare wild type codon for observed frequency of use by human gene.s (refer to Table 3 in Lathe et al.).
3. If codon is not the most comrnonly employed, replace it with an optimal codon for high expression based on data in Table 5.
4. Inspect the third nucleotide of the new codon and the first nucleotide of the adjacent codon immediately 3'- of the first. If a 5'-CG-3' pairing has been created by the new codon selection, replace it with the choice indicated in Table 5.
5. Repeat this procedure unti~ the entire gene segment has been replaced.

6. Inspect new gene sequence for undesired sequences generated by these codon replacements (e.g., "ATTTA" sequences, inadvertent creation of intron splice recognition sites, unwanted restriction enzyme sites, etc.) and substitute codons that elimin~te these ~ 30 se(~uences.

7. Assemble synthetic gene se~ment~s and test for improved expression.

WO 97/31115 PCT~US97/02294 These methods were used to create the following synthetic gene segments for HIV en~ creating a gene comprised entirely of optimal codon usage for expression: (i) gpl20-CI (opt~; (ii) Vl -VS (opt)~ (iii) RRE-A/B (mut or opt); and (iv) gp30 (opt) with percentages of codon S replacements/nucleotide substitution,s of 56/I 9, 73/26, 7~/2~s, and 61/25 obtained for each segment, respectively. Each of these .segments has been described in detail above with actual ,sequences listed below.

ap1 20-C1 (opt) This is a gpl20 constant region 1 (Cl) gene segment from the mature N-terminus to the beginning of Vl designed to have optimal codon usage for expression.

lS 1 TGATCACA~A GAAGCTGTGG GTGACAGTGT ATTATGGCGT GCCAGTCTGG

CA 0224~646 1998-08-06 WO 97131115 PCTrUS97/02294 - ~3 -~ MN Vl-V5 ~opt) This i,s a gene segment corresponding to the derived protein sequence for HIV MN Vl-VS (1066BP) having optima~ codon usage for 5 expre,ssion.

1 AGTTTA~ACT GCACAGACCT GAGGAACACC ACCAACACCA ACAACTCCAC

2~451 ATCAGGTCTG AGAACTTCAC AGACAATGCC AAGACCATCA TCGTGCACCT

CA 0224~646 1998-08-06 WO 97/31115 YCT~US97/02294 s lO01 AGGACACAGA CACCAACGAC ACCGA~ATCT TCAGGCCTGG GGGGGGGGAC

CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 RRF.Mut (A) This is a DNA segment corresponding to the rev response element (RRE) of HIV-l compri.sed of optimal codon usage for expression. The 5 "A" form also has removed the known proteolytic cleavage sites at the gpl20/gp~l junction by using the nucleotides indicated in boldface.

0 51 CCTGGGGGTG GCCCCAACAA AAGCT~-~rr~rGTGGTG CA~-~r~.AGC

101 ~CCAGGCCGT GGGCATTGGG GCCCTGTTTC TGGGCTTTCT GGGGGCTGCT

CA 02245646 l99X-08-06 W O 97/31115 PCTrUS97/02294 RR~.Mut (B) This is a DNA segment corresponding to the rev response element (RRE) of HIV-l comprised of optimal codon usage for expression. The S "B" form retains the known proteolytic cleavage sites at the gpl20/gp41 junction.

0 51 CCTGGGGGTG GCCCCAACAA AAGcT~r-~r-~r-~TGGTG ~-~r~ G~

101 ~r-~r-~-ccGT GGGCATTGGG GCCCTGTTTC TGGGCTTTCT GGGGGCTGCT

301 CAGGCCCGGG TG~T~7~GT CGAGCGCTAT CTGAAAGACC AGCAACTCCT

CA 0224~646 1998-08-06 WO 97t3111~ PCTrus97/02294 ~p32 (o~t) This is a gp32 gene segment from the AvrI~ site (starting immediately at the end of the RRE) to the end of gpl43 comprised of optimal codons S for expression.

WO 97/31115 PCTnUS97/02294 SRV-1 CTE (A) This is a synthetic gene segment corresponding to a 3'-UTR from the Simian l~etrovirus-l genome. This DNA is placed in the following 5 orientation at the 3'-terminus of HIV genes to increase rev-independent expression.

Srfl EcoRV
5'-GCCC G~aC ~ATATc TA GACCACCTCC CCTGCGAGCT AAGCTGGACA

GCCAATGACG GGTAAGAGAG TGACATTTTT CACTAACCTA AGACAGGAGG

GCCGTCAGAG CTACTGCCTA ATCCAAAGAC GGGTAAAAGT GATAAA~ATG

TATCACTCCA ACCTAAGACA GGCGCAGCTT CCGAGGGATT TGTCGTCTGT

TTTATATATA TTTAAAAGGG TGACCTGTCC GGAGCCGTGC TGCCCGGATG

ATGTCTTGG ~A~A~C 9CCC GGGC -3' EcoRV Sr~I

CA 02245646 l998-08-06 W O 97/31115 PCT~US97/02294 SRV-l CTE (B) This synthetic gene segment is identical to SRV-1 CTE (A) shown above except that a single nucleotide mutation was used (indicated by boldface) 5 to elimin~te an ATTTA sequence. This sequence has been associated with increased mRNA turnover.

SrfI EcoRV
5'-~CCC GG~C ~-~TC TA GACCACCTCC CCTGCGAGCT AAGCTGGACA

GCCAATGACG GGTAAGAGAG TGACATTTTT CACTAACCTA AGACAGGAGG

GCCGTCAGAG CTACTGCCTA ATCCAAAGAC GGGTAAAAGT GATAAAAATG

TATGACTCCA ACCTAAGACA GGCGCAGCTT CCGAGGGATT TGTCGTCTGT

TTTATATATA TTBAAAAGGG TGACCTGTCC GGAGCCGTGC TGCCCGGATG

ATGTCTTGG ~-~ATC ~CCC GGG~
EcoRV SrfI

EXAMPLE~ I lL
ln Vitro ~pl20 Vaccine Expression:
In vitro expression was tested in transfected human rhabdomyosarcoma (RD) cells for these constructs. Quantitation of secreted tPA-gpl20 from transfected RD cells showed that VlJns-tPA-gp 120 vector produced secreted gp 120.

In Vivo gp120 Vaccination:
See figure 12 (mouse data):
vl3ns-tpA-~pl2ol\~N PNV-induced Class II MHC-restricted T lymphocyte ~pl20 .specific anti~en reactivities. Balb/c mice W O 97/31115 PCTrUS97/02294 which had been vaccinated two times with 200 ,ug VIJns-tPA-gpl20MN
were sacrificed and their spleens extracted for in vitro dete~ninations of helper T Iymphocyte reactivities to recombinant gpl20. T cell proliferation assays were performed with PBMC (peripheral blood mononuclear cells) using recombinant gp120IIIg (Repligen, catalogue #RP1016-20) at 5 ,ug/ml with 4 x 105 cells/ml. Basal levels of 3H-thymidine uptake by these cells were obtained by culturing the cells in media alone, while maximum proliferation was induced using ConA
stimulation at 2 ,ug/ml. ConA-induced reactivities peak at ~3 day.s and 10 were harvested at that time point with media control samples while antigen-treated samples, were harvested at 5 days with an additional media control. Vaccinated mice responses were compared with naive, age-matched syngeneic mice. C:~onA positive controls gave very high prolil~eration for both naive and immunized mice as expected. Very 15 strong helper T cell memory responses were obtained by gpl20 treatment in vaccinated mice while the naive mice did not re.spond (the threshold for specific reactivity is an stimulation index (SI) of ~3-4; SI
is calculated as the ratio of sample cpm/media cpm). Sl's of 65 and 14 were obtained for the vaccinated mice which compare~ with anti-gpl20 20 ELISA titers of 5643 and 11,900, respectively, for these mice.
Interestingly, for these two mice the higher responder for antibody gave significantly lower T cell reactivity than the mouse having the lower antibody titer. This experiment demonstrates that the secreted gpl20 vector efficiently activates helper T cells in vivo as well as generates 25 strong antiboby responses. ln addition, each of these immune re.sponses was determined using antigen which was heterologous compared to that encoded by the inoculation PNV (11~, vs. MN):

30 gp l 60 Vaccines In addition to secreted gpl20 constructs, we have prepared expression constructs for full-length, membrane-bound gpl 60. The rationales for a gpl60 construct, in addition to gpl20, are ~1) more CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 epitope.s are available both for both CTL stimulation as well as neutralizing antibody production including gp41, against which a potent HIV neutralizing monoclonal antibody ~2F5, see above) is directed; (2) a more native protein structure may be obtained relative to virus-5 produced gp 160; and, (3) the success of membrane-bound influenza HA
constructs for immllnogenicity [Ulmer et al., Science 259:1745-1749, 1993; Montgomery, D., et al., DNA and Cell Biol.. 12:777 783, 1993].
gpl~iO retains substantial REV dependence even with a heterologous leader peptide sequence so that further constructs were made to increase expression in the absence of REV.

Assay For Hiv Cytotoxic T-Lymphocytes:
The methods described in this section illustrate the assay as used for vaccinated mice. An essentially similar assay can be used with primates except that autologous B cell lines must be established for use as target cells for each animal. Thix can be accomplished for humans u.sing the l~pstein-Barr virus and for rhesus monkey using the herpes B
virus.
Peripheral blood mononuclear cells (PBMC) are derived from either freshly drawn blood or spleen using Ficoll-Hypaque centrifugation to separate erythrocytes from white blood cells. For mice, Iymph nodes may be used as well. Effecter CTLs may be prepared from the PBMC either by in vitro culture in IL-2 (20 U/ml) and concanavalin A (2!1g/ml) for 6-12 days or by using specific antigen using an equal number of irradiated antigen presenting cells. Specific antigen can consist of either synthetic peptides (9-15 amino acids usually) that are known eptitopes for CTL recognition for the MHC
haplotype of the ~nim~l~ used, or vaccinia virus constructs engineered to express appropriate antigen. Target cells may be either syngeneic or MHC haplotype-matched cell lines which have been treated to present appropriate antigen as described for in vitro stimulation of the CTLs.
For Balb/c mice the P18 peptide W O 97/31115 PCT~US97/02294 (ArgIleHisIleGlyProGlyArgAlaPheTyrThrThrLysAsn, SEQ.ID:51:, for HIV MN strain) can be used at 10 ,uM concentration to restimulate CTL
in vitro using irradiated syngeneic splenocytes and can be used to sensitize target cells during the cytotoxicity assay at 1-10 ,uM by 5 incubation at 37~C for about two hours prior to the assay. For these H-2d MHC haplotype mice, the murine mastocytoma cell line, P815, provides good target celIs. Antigen-sensitized target cells are loaded with Na5 1CrO4, which is released from the interior of the target cells upon killing by CTL, by incubation of targets for 1-2 hours at 37~C
lO (0.2 mCi for ~5 x 106 cells) followed by several washings of the target cells. CTL populations are mixed with target cells at valying ratios of effector~ to targets such as 100:1, 50:1, 25:1, etc., pelleted together, and incubated 4-6 hours at 37~C before harvest of the supernatants which are then assayed for release of radioactivity using a gamma counter.
15 Cytotoxicity is calculated as a percentage of total releasable counts from the target cells (obtained using 0.2% Triton X-100 treatment) from which spontaneous release from target cells has been subtracted.

20 A.ssay For Hiv Specific Antibodies:
ELISA were designed to detect antibodies generated against HIV using either specific recombinant protein or synthetic peptides as substrate antigens. 96 well microtiter plates were coated at 4~C
overnight with recombinant antigen at 2 ,ug/ml in PBS (phosphate 25 buffered saline) solution using 50 ,ul/well on a rocking plat~orm.
Antigens consisted of either recombinant protein (gp 120, rev: Repligen Corp.; gpl60, gp41: American Bio-Technologies, ~nc.) or synthetic peptide (V3 peptide corresponding to virus isolate sequences from IIIB, etc.: American Bio-Technologies, Inc.; gp4 1 epitope for monoclonal 30 antibody 2F5). Plates were rinsed four times using wash buffer (PBS/0.05% Tween 20) followed by addition of 200,uVwell of blocking buffer (1% Carnation milk solution in PBS/0.05% Tween-20) for l hr at room temperature with rocking. Pre-sera and immune sera were CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 diluted in blocking buffer at the desired range of dilutions and 100 ~1 added per well. Plates were incubated for 1 hr at room temperature with rocking and then washed four times with wash buffer. Secondary antibodies conjugated with hor.se radish peroxidase, (anti-rhesus Ig, 5 Southern Biotechnology Associate.,; anti- mouse and anti-rabbit Igs, Jackson Immuno Research) diluted 1:2000 in blocking buffer, were then added to each sample at 100 ,ul/well and incubated 1 hr at room temperature with rocking. Plates were washed 4 times with wash buffer and then developed by addition of 100 ~l/well of an o-phenylenediamine 10 (o-PD, Calbiochem) solution at l mg/ml in 100 mM citrate buffer at pH
4.~. Plates were read for absorbance at 450 nm both kinetically (first ten minutes of reaction) and at lO and 30 minute endpoints (Thermo-max microplate reader, Molecular Devices).

1~ FXAMPLE 15 Assay For Hiv Neutralizin~ Antibodies:
In vitro neutralization o~ HIV isolates assays using sera derived from vaccinated animals was performed as follows. Test .sera and pre-immune sera were heat inactivated at 56~c for 60 min before use. A titrated amount of HIV-l was added in 1:2 seria} dilution.s of test sera and incubated 60 min at room temperature before addition to 105 MT-4 human Iymphoid cells in 96 well microtiter plates. The virus/cell mixtures were incubated for 7 days at 37~C and assayed for virus-mediated killing of cells by staining cultures with tetrazolium dye.
2;S Neutralization of virus is observed by prevention of viru,s-mediated cell death.

Isolation Of Gene.s From Clinical Hiv I.solate~s:
HIV viral genes were cloned from infected P~3MC's which had been activated by ConA treatment. The preferred method for obtaining the viral genes was by PCR amplification from infected cellular genome using specific oligomers flanking the desired gene.s. A

:

W O 97/31115 PCT~US97/02294 second rnethod for obtaining viral genes was by purification of viral RNA from the supernatants of infected cells and preparing cDNA from this material with subsequent PC~R. This method was very analogous to that described above for cloning of the murine B7 gene except ~or the 5 PCR oligomers used and random hexamers used to make cDNA rather than specific priming oligomers.
Genomic DNA was purified from infected cell pellets by Iysis in STE solution (10 mM NaCl, 10 mM EDTA, 10 mM Tris-~CI, pH 8.0) to which Proteinase K and SDS were added to 0.1 mg/ml and 10 0.5% final concentrations, respectively. This mixture wa,s incubated overnight at 56~C and extracted with 0.5 volumes of phenol:chloroform:isoamyl alcohol (25:24:1). The aqueous phase was then precipitated by addition of sodium acetate to 0.3 M final concentration and two volume~ of cold ethanol. After pelleting the 1~ DNA from solution the DNA was resuspended in O.lX TE solution (lX
TE = 10 mM Tris-HCI, pH 8.0, I mM EDTA). At this point SDS was added to O.I % with 2 U of RNAse A with incubation for 30 minutes at 37~C. This solution was extracted with phenol/chloroform/isoamyl alcohol and then precipitated with ethanol as before. DNA was 20 suspended in 0.1 X TE and quantitated by measuring its ultraviolet absorbance at 260 nm. Samples were stored at -20~C until used for PC~.
PCR was performed using the Perkin-Elmer Cetus kit and procedure using the following sense and antisense oligomers for gpl60:
25 5'-GA AAG AGC AGA AGA CAG TGG CAA TGA -3', and 5'-GGG
CTT TGC TAA ATG GGT GGC AAG TGG CCC GGG C ATG TGG-3', respectively. These oligomers add an Srfl site at the 3'-terrninus of the resulting DNA fragment. PCR-derived segments are cloned into either the VlJns or VlR vaccination vectors and V3 regions as well as 30 ligation junction sites confirmed by DNA sequencing.

CA 0224~646 1998-08-06 W O 97/3111~ PCT~US97/~2294 E~AMPLE 17 T ~ell Proliferation Assays:
PBMCs are obtained and tested for recall responses to specific antigen as determined by proliferation within the PBMC
5 population. Proliferation is monitored using 3H-thymidine which is added to the cell cultures for the last 18-24 hours of incubation be~ore harvest. Cell harvesters retain isotope-containing DNA on filters if proliferation has occurred while quiescent cells do not incorporate the isotope which is not retained on the filter in free form. For either rodent or primate species 4 X 105 cells are plated in 96 well microtiter plates in a total of 200 !11 of complete media (RPMV10% fetal calf serum). Background proliferation responses are determined using PBMCs and media alone while nonspecific responses are generated by using lectins such as phytohaemagglutin ~PHA) or concanavalin A
15 (ConA) at 1- 5 ,ug/ml concentrations to serve as a positive control.
Specific antigen consists of either known peptide epitopes, purified protein, or inactivated virus. Antigen concentrations range from 1- 10 ,uM for peptides and I-10 ,ug/ml for protein. Lectin-induced proliferation peaks at 3-5 days of cell culture incubation while antigen-20 specific responses peak at 5-7 days. Specific proliferation occurs when radiation counts are obtained which are at least three-fold over the media background and is often given as a ratio to background, or Stimulation ~ndex (SI). HIV gpl60 is known to contain several peptides known to cause T cell proliferation of gp 1 60/gpl 20 immunized or HIV-25 infected individuals. The most commonly used of these are: Tl(LysGlnlleIleAsnMetTrpGlnGluValGlyLysAlaMetTyrAla,; T2 (HisGluAspIleIleSerLeuTrpAspGlnSerLeuLys); and, TH4 (AspArgValIleGluValValGlnGlyAalTyrArgAlaIleArg). These peptides have been demonstrated to stimulate proli~eration o~ PBMC ~rom 30 antigen-sensitized mice, nonhllm:~n primates, and humans.

W O 97/31115 PCT~US97/02294 EXAMPLE l~S
Vector V 1 R Preparation:
In an effort to continue to optimize our basic vaccination vector, we prepared a derivative of VlJns which was designated as 5 VlR. The purpose ~or this vector construction was to obtain a minimum-sized vaccine vector, i.e., without unnecessaly DNA
sequences, which still retained the overall optimized heterologous gene expression characteristics and high plasmid yields that VIJ and VlJns afford. We determined from the literature as well as by experiment 10 that (1) regions within the pUC backbone comprising the E. coli origin of replication could be removed without affecting plasmid yield from bacteria; (2) the 3'-region of the kanr gene following the kanamycin open reading frame could be removed if a bacterial terminator was inserted in its stead; and, (3) ~300 bp from the 3'- half of the BGH
15 termin~tor could be removed without affecting its regulatory function (following the original KpnI restriction enzyme site within the BGH
element).
VlR was constructed by using PCR to synthesize three segments of DNA from VlJns representing the CMVintA
20 promoter/BGH termin~tor, origin of replication, and k~n~mycin resistance elements, respectively. Restriction enzymes unique for each segment were added to each segment end using the PCR oligomer.s:
SspT and ~hoI for CMVintA/BGH; EcoRV and BamHI for the kan r gene, and, Bcll and SalI for the ori r These enzyme sites were chosen 25 because they allow directional ligation o~ each of the PCR-derived DNA
segments with subsequent loss of each site: EcoRV and SspI leave blunt-ended DNAs which are compatible for ligation while BamHI and BclI
leave complementary overhangs as do SalI and XhoI. After obtaining these segments by PCR each segment was digested with the appropriate 30 restriction enzymes indicated above and then ligated together in a single reaction mixture containing all th~ee DNA segments. The ~'-end of the ori r was designed to include the T2 rho independent terminator sequence that is normally found in this region so that it could provide W O 97131115 . PCT~US97102294 termination infonnation for the k~n~mycin resistance gene. The ligated product was confirmed by restriction enzyme digestion (>8 enzymes) as well as by DNA sequencing of the ligation junctions. DNA plasmid yields and heterologous expression using viral genes within VlR appear 5 similar to VlJns. The net reduction in vector size achieved was 1346 bp (VlJns = 4.86 kb; VlR = 3.52 kb), see figure 11, SEQ.ID:45:.

PCR oligomer sequences used to synthesize VlR ~restriction enzyme sites are underlined and identified in brackets following sequence):
(1) 5'-GGT ACA AAT ATT GG CTA TTG GCC ATT GCA TAC
G-3' [SspI], SEQ.ID:, (2) 5'-CCA CAT CTC GAG GAA CCG GGT CAA TTC TTC AGC
15 ACC-3' [XhoI], SEQ.ID::
(for CMVintA/BGH segment) (3) 5'-GGT ACA GAT ATC GGA AAG CCA CGT TGT GTC TCA
AAA TC-3'[EcoRV], SEQ.ID::
~4) 5'-CCA CAT GGA TCC G TAA TGC TCT GCC AGT GTT
ACA ACC-3' [BamHI], SEQ.ID::
(for k~n~mycin resistance gene segment) (5) 5'-GGT ACA TGA TCA CGT AGA AAA GAT CAA AGG ATC
TTC TTG-3'[BclI], SEQ.ID::, (6) 5'-CCA CAT GTC GAC CC GTA AAA AGG CCG CGT TGC
~ TGG-3' [SalI], SEQ.ID::
(for E. coli origin of replication~
Llgation junction.s were sequenced for VlR using the following ollgomers:

W O 97/31115 PCTr~S97/02294 5'-GAG CCA ATA TAA ATG TAC-3', SEQ.l~
[CMVintA/kanr junction]
AA TAG CAG GCA TGC-3', SEQ.ID:: ~BGH/ori junction]
5'-G CA~ GCA GCA GAT TAC-3', SEQ.ID:: Lori/l~anr junction]

Heterologous Expression of H~V Late Gene Products HIV structural genes such as env and gag require expression of the HIV regulatory gene, rev, in order to efflciently produce full-length proteins. We have found that rev-dependent expression of gag yielded low levels of protein and that rev itself may be toxic to cells. Although we achieved relatively high levels of rev-dependent expression of gpl60 in vitro this vaccine elicited low levels of antibodies to gpl60 following in vivo imm~ni7~tion with rev/gp~60 DNA. This may result from known cytotoxic effects of rev as well as increased difficulty in obtaining rev ~unction in myotubules containing hundreds of nuclei (rev protein needs to be in the same nucleus as a rev-dependent transcript in order for gag or env protein expression to occur). However, it has been possible to obtain rev-independent expression using selected modifications of the env gene.

1. re~-independent expre~ssion of env:
In general, our vaccines have utilized primarily HIV (IIIB) env and ,~ag genes for optimi7~tion of expression within our generalized vaccination vector, VlJns, which is comprised of a CMV immediate-early (IE) prornoter, a BGH-derived polyadenylation and transcriptional termination sequence, and a pUC backbone. Varying efficiencies, depending upon how large a gene segment is used (e.g., gpl20 vs.
gpl60), of rev-independent expression may be achieved for enl~ by replacing its native secretory leader peptide with that from the tissue-specific pl~minogen activator (tPA) gene and expressing the resulting CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 chimeric gene behind the CMVIE promoter with the CMV intron A.
tPA-gpl20 is an example of a secreted gpl20 vector constructed in this fashion which functions well enough to elicit anti-gpl20 immllne response,s in vaccinated mice and monkeys.
S Because of reports that membrane-anchored proteins rnay induce much more substantial (and perhaps more specific for HIV
neutralization) alltibody responses compared to secreted proteins as well as to gain additional epitopes7 we prepared VlJns-tPA-gpl60 and V lJns-~ ev/gp l 60. The tPA-gp l 60 vector produced detectable quantities of gpl60 and gpl20, without the addition of rev, as shown by immunoblot analysis of transfected cells, although levels of expression were much lower than that obtained for rev/gpl 60, a rel~-dependent gpl60-expressing plasmid. This is probably because inhibitory region,s, which confer rev dependence upon the gpl60 transcript, occur at multiple sites within gp160 including at the COOH-terminus of gp41. A
vector was prepared for a COOH-terminally truncated form of tPA-gpl~0 (tPA-gpl43) which was designed to increase the overall expression levels of env by elimin~tion of these inhibitory sequences.
The gpl43 vector also eliminate.s intracellular gp41 regions containing peptide motifs (such as Leu-Leu) known to cause diversion of membrane proteins to the Iysosomes rather than the cell surface. Thus, gpl43 may be expected to have increased levels of expression of the ~nv protein (by decreasing rev-dependence) and greater efficiency of transport of protein to the cell surface compared to full-length gpl60 25 where these proteins may be better able to elicit anti-gpl 60 antibodies following DNA vaccination. tPA-gpl43 was further modified by extensive silent mutagenesis of the rev response element (RRE) sequence (350 bp) to eliminate additional inhibitory sequences for expression.
This construct, gpl43/mutRRE, was prepared in two form,s: either 30 elimin~ting (form A) or retaining (form B) proteolytic cleavage sites for gpl20/41. Both forms were prepared because of literature reports that vaccination of mice using uncleavable gpl 60 expressed in vaccinia CA 0224~646 1998-08-06 W O 97/31115 PCT~US97102294 elicited much higher levels of antibodies to gpl60 than did cleavable forms.
A quantitative ELISA for gpl60/gpI20 expression in cell trans~ectants was developed to determine the relative expression S capabilities for these vectors. In vitro transfection of 293 cells followed by quantification of cell-associated vs. secreted/released gpl20 yielded the following results: (1) tPA-gpl60 expressed ~-lOX less gpl20 than ~cv/gpl60 with similar proportions retained intracellularly vs. released from the cell surface; (2) tPA-gp~43 gave 3-6X greater secretion of 10 gpl20 ~an re~/gpl60 with only low levels of cell-associated gpl43, confirming that the cytoplasmic tail of gpl60 causes intracellular retention of gpl60 which can be overcome by partial deletion of this se~uence; and, (3) tPA-gpl43/mutRRE A and B gave ~lOX greater expression levels of protein than did parental tPA-gpl43 while 15 e}imin~tion of proteolytic processing was confirmed for form A.
Thus, our strategy to increase rev-independent expres.sion has yielded stepwise increases in overall expression levels as well as redirecting membrane-anchored gpl43 to the cell surface away from Iysosomes. It is important to note that this is a generic construct into 20 which it should be possible to insert gpl20 sequences derived from various pr~mary viral isolates within a vector cassette cont~ining these modifications which reside either at the NH2-terminus (tPA leader) or COOH-terminus (gp41), where few antigenic differences exist between different viral strains.
2. Expression of ~pl20 derived from a clinical isolate:
To apply these expression strategies to viruses that are relevant for vaccine purposes and confirm the generality of our approaches, we also prepared a tPA-gpl20 vector derived ~rom a 30 primary HIV isolate (containing the North Arnerican concensus V3 peptide loop; macrophage-tropic and nonsyncytia-inducing phenotypes).
This vector gave high expression/secretion of gpl20 with transfected 293 cells and elicited anti-gpl20 antibodies in mice thus demonstrating CA 0224~646 1998-08-06 W O 97/31115 PCT~US97/02294 that it was cloned in a functional form. Primary isolate ~pl60 genes will also be used for expression in the same way as for gpl60 derived from laboratory strains.

B. Immune Responses to HTV-l env Polynucleotide Vaccines Effect of vaccination route on immune responses in mice:
While efforts to improve expression of gpl60 are ongoing, we have utilized the tPA-gpl20 DNA construct to assess immllne responses and ways to augment them. Intramuscular (i.m.) and intradermal (i.d.) vaccination routes were compared for this vector at 100, 10, and 1 ,~g doses in mice. Vaccination by either route elicited antibody responses (GMTs = 103-104) in all recipients following 2-3 vaccinations at all three dosage levels. Each route elicited similar anti-gpl20 antibody titers with clear dose-dependent responses. However, we observed 15 greater variability of responses for i.d. vaccination, particularly at the lower doses following the initial inoculation. Moreover, helper T-cell responses, as determined by antigen-specific in vitro proliferation and cytokine secretion, were higher following i.m. vaccination than i.d. We concluded that i.d. vaccination did not offer any advantages compared to 20 i.m. for this vaccine.

2. gp 120 DNA vaccine-mediated helper T cell immunity in mice:
gpl20 I~NA vaccination produced potent helper T-cell 25 responses in all lymphatic compartments tested (spleen, blood, inguinal, mesenteric, and iliac nodes) with TH 1 -like cytokine secretion profiles (i.e., g-interferon and IL-2 production with little or no IL-4). The,se cytokines generally promote strong cellular immunity and have been as.sociated with maintenance of a disease-free state for HTV-seropositive 30 patients. Lymph nodes have been shown to be primary sites for HIV
replication, harboring large reservoirs of virus even when virus cannot be readily detected in the blood. A vaccine which can elicit anti-HIV
immune responses at a variety of Iymph sites, such as we have shown W O 97/31115 PCT~US97/02294 with our DNA vaccine, may help prevent successful colonization of the lymphatics following initial infection.

3. env DNA vaccine-mediated antibody responses:
African green (AGM) and Rhesus (RHM) monkeys which received gpl20 DNA vaccines showed low levels of neutr~li7ing antibodies following 2-3 vaccinations, which could not be increa.sed by additional vaccination. These results, as well as increasing awareness within the HIV vaccine field that oligomeric gpl60 is probably a more 10 relevant target antigen for eliciting neutralizing antibodies than gpl20 monomer.s, have led us to focus upon obt~ining ei~fective e~pression of gpl60-based vectors (see above). Mice and AGM were also vaccinated with the primary isolate derived tPA-gpl20 vaccine. These animals exhibited anti-V3 peptide ~using homologous se~luence) reciprocal endpoint antibody titers ranging 500-~000, demonstrating that this vaccine design is functional for clinically relevant viral isolates.
The gp 160-based vaccines, rev-gp 160 and tPA-gp 160, failed to consistently elicit antibody responses in mice and nonhl~m~n primates or yielded low antibody titers. Our initial results with the 20 tPA-gpl43 plasmid yielded geometric mean titers (GMT) > 103 in mice and AGM following two vaccinations. These data indicate that we have signficantly improved the immunogenicity of gpl60-like vaccines by increasing expression levels and more efficient intracellular trafficking of env to the cell surface. This construct, as well as the tPA-25 gpl43/mutRRE A and B vectors, will continue to be characterized forantibody responses, especially for virus neutralization.

4. env DNA vaccine-mediated CTL responses in monke~s:
We continued to characterize CTL response.s of RHM that 30 had been vaccinated with gpI20 and gpl60/IRES/rel, DNA. All four monkeys that received this vaccine showed significant MHC Class I-restricted CTL activities (20-35% specific killing at an effector/target =
20) following two vaccinations. Following a fourth vaccination these W O 97/31115 PCT~US97/02294 activities increased to 50-60% killing under similar test conditions, indicating that additional vaccination boosted responses significantly.
The CTL activities have persisted for at least seven months subsequent to the final vaccination at about 50% of their peak levels indicating that 5 long-term memory had been established.

Claims (10)

WHAT IS CLAIMED:

1. A synthetic polynucleotide comprising a DNA
sequence encoding HIV env protein or a fragment thereof, the DNA
sequence comprising codons optimized for expression in a mammalian host.

2. The polynucleotide of Claim 1 which is selected from:
V1Jns-tPA-HIV MN gp 120;
V1Jns-tPA-HIVIV IIIB gp 120;
V1Jns-tPA-gp140/mutRRE-A/SRV-1 3'-UTR;
V1Jns-tPA-gp140/mutRRE-B/SRV-1 3'-UTR;
V1Jns-tPA-gp140/opt30-A;
V1Jns-tPA-gp140/opt30-B;
V1Jns-tPA-gp140/opt all-A;
V1Jns-tPA-gp140/opt all-B;
V1Jns-tPA-gp140/opt all-A;
V1Jns-tPA-gp140/opt all-B;
V1Jns-rev/env:;
V1Jns-gp160;
V1Jns-tPA-gp160;
V1Jns-tPA-gp160/opt C1/opt41-A;
V1Jns-tPA-gp160/opt C1/opt41-B;
V1Jns-tPA-gp160/opt all-A;
V1Jns-tPA-gp160/opt all-B;
V1Jns-tPA-gp160/opt all-A;
V1Jns-tPA-gp160/opt all-B;
V1Jns-tPA-gp143;
V1Jns-tPA-gp143/mutRRE-A;
V1Jns-tPA-gp143/mutRRE-B;
V1Jns-tPA-gp143/opt32-A;
V1Jns-tPA-gp143/opt32-B;

V1Jns-tPA-gp143/SRV-1 3'-UTR;
V1Jns-tPA-gp143/opt C1/opt32A;
V1Jns-tPA-gp143/opt C1/opt32B;
V1Jns-tPA-gp143/opt all-A;
V1Jns-tPA-gp143/opt all-B;
V1Jns-tPA-gp143/opt all-A;
V1Jns-tPA-gp143/opt all-B;
V1Jns-tPA-gp143/opt32-A/glyB;
V1Jns-tPA-gp143/opt32-B/glyB;
V1Jns-tPA-gp143/opt C1/opt32-A/glyB;
V1Jns-tPA-gp143/opt C1/opt32-B/glyB;
V1Jns-tPA-gp143/opt all-A/glyB;
V1Jns-tPA-gp143/opt all-B/glyB:
V1Jns-tPA-gp143/opt all-A/glyB;
V1Jns-tPA-gp143/opt all-B/glyB; and combinations thereof.

3. The polynucleotide of Claim 1 which induces anti-HIV neutralizing antibody, HIV specific T-cell immune responses, or protective immune responses upon introduction into vertebrate tissue, including human tissue in vivo, wherein the polynucleotide comprises a gene encoding an HIV gag, gag-protease, or env gene product,

4. A method for inducing immune responses in a vertebrate against HIV epitopes which comprises introducing between I
ng and 100 mg of the polynucleotide of Claim 1 into the tissue of the vertebrate.

5. A method for inducing immune responses against infection or disease caused by virulent strains of HIV which comprises introducing into the tissue of a vertebrate the polynucleotide of Claim 1.

6. The method of Claim 5 which further comprises administration of attenuated HIV, killed HIV, HIV env protein, HIV gag protein, HIV pol protein, and combinations thereof.

7. A vaccine against HIV infection which comprises the polynucleotide of Claim 1 and a pharmaceutically acceptable carrier.

8. A method for inducing anti-HIV immune responses in a primate which comprises introducing the polynucleotide of Claim 1 into the tissue of the primate and concurrently administering interleukin 12 parenterally.

9. A method of inducing an antigen presenting cell to stimulate cytotoxic and helper T-cell proliferation an effector functions including lymphokine secretion specific to HIV antigens which comprises exposing cells of a vertebrate in vivo to the polynucleotide of Claim 1.

10. A method of increasing expression of DNA encoding an HIV protein or a fragment thereof, comprising:
(a) identifying placement of codons for proper open reading frame;
(b) comparing wild type codons for observed frequency of use by human genes;
(c) replacing wild-type condons with codons optimized for high expressionof human genes; and (d) testing for improved expression.