Applying the PDR principle to AIDS
J. theor. Biol. (1988) 130, 469-480 Applying the PDR Principle to AIDS JOHN C. SANFORD Department of Horticultural Sciences, Cornell University, Hedrick Hall, Geneva, NY 14456, U.S.A. (Received 2 February 1987, and in revised form 9 October t987) The principle of pathogen-derived resistance (the PDR principle) has been put forward as a broadly-applicable conceptual tool for use in designing genes which will confer resistance to pathogens. This paper reveals an example of how the PDR principle may be applied in the field of human medicine. Specifically it is shown how the PDR principle can be employed in designing a series of genes which should be capable of protecting human blood cells from the retrovirus causing the AIDS disease. Prospects are discussed for using such genes in gene therapy treatment of people infected with this virus. Introduction The pathogen-derived resistance (PDR) principle (Sanford & Johnston, 1985), in its simplest form, can be stated as follows: "Nucleotide sequences derived from a pathogen can be used in the genetic modification of its host, such that the host becomes resistant to that pathogen." The concept of pathogen-derived resistance adds a new dimension to our under- standing of host-pathogen relations, and promises to be a useful tool in designing genes conferring resistance to pathogens. The genetic engineering of pathogen- derived resistance has already been demonstrated in several model viral systems. Sanford & Johnston (1985), analyzed the QB bacteriophage and predicted that the QB bacteriophage host (E. coil) could be made resistant to QB infection by at least three different PDR mechanisms. It was predicted that QB resistance could be engineered by: 1) cloning and expressing a QB regulatory protein (the coat protein) into the host; 2) subcloning a portion of the QB replicase gene, such that the host would express a protein fragment consisting of the RNA binding domain of the replicase; 3) expressing viral antisense RNA in the host (mRNA complementary to part of the viral genome). Two of these PDR mechanisms have already been proven valid. QB coat protein conditions high levels of resistance to QB infection when expressed in the host, with plaque number being reduced roughly 1000-fold (Grumet et aL, 1987a, b). A single antisense RNA construct had measurable but modest anti-viral effects in the QB system, seen mostly as reduced plaque size (Johnston & San ford, unpublished data). In the closely related SP bacteriophage system, Coleman et al. (1985), have demonstrated that certain antisense RNAs can have very significant anti-viral effects, as seen by reduced plaque number and plaque size. In addition, Abel et al. (1986) have shown that viral coat protein confers a degree of resistance to tobacco mosaic virus, when expressed in tobacco. Other recent examples of 469 0022-5193/88/040469+ t2 $03.00/0 O 1988 Academic Press Limited 470 J.C. SANFORD parasite-derived resistance include resistance to cucumber mosaic virus deriving from viral satellite eDNA (Baulcombe et al. 1986), and resistance to alfalfa mosaic virus deriving from the viral coat gene (Loesch-Fries et al. 1987). Taken collectively, such findings indicate that parasite-derived resistance (under the right circumstances) can be quite readily achieved, and can be realized through a variety of molecular mechanisms. Applying PDR to Medicine The general PDR concept was conceived and developed primarily for use in the field of agriculture, where genetic resistance is of great importance. In the field of human medicine, altering the genotype of the host has never been a credible method of fighting infectious disease. However, it is now believed that human somatic gene therapy will be possible in the relatively near future (Andersen, 1984), creating the prospect that PDR might be applied within the field of medicine. Two practical considerations would appear to limit PDR applications in medicine. First, the drastic and permanent nature of somatic gene therapy would seldom be justified. Where the human immune system failed to fight off infection, more orthodox medical treatments such as vaccination or chemotherapy would normally be practiced. Second, genetic protection would not be feasible in most cases, since most human tissues are not presently amenable to being genetically engineered. Emerging gene therapy techniques are only effective for the genetic modification of bone marrow and blood cells. Because of this limitation, the projected use of somatic gene therapy has generally been assumed to be limited to the correction of rare hereditary gene defects, where such defects center in bone marrow or blood cells (Andersen, 1984). Despite these limitations, there are certain persistent and life-threatening pathogens of the blood for which conventional defenses appear inadequate and where use of the PDR principle might be both feasible and justified. Most noteworthy of these diseases is the acquired immune deficiency syndrome (AIDS) (Wong-Staal & Gallo, 1985; Weiss et aL, 1986). This disease is caused by a retrovirus which has been called HTLV III, LAV, ARV, or most recently, HIV (Coffin et al., 1986). The development of an unorthodox and "permanent" treatment for the AIDS disease would appear highly justifiable, given the scope of the epidemic (Curran et al., 1985), the lethal nature of the disease, the permanent nature of the provirus (Weiss, 1985, Folks et al., 1986), and the questionable efficacy of more orthodox treatments. The development of anti-HIV genes and an effective gene therapy treatment for AIDS would appear to be feasible given the proven effectiveness of the PDR principle (Sanford & Johnston, 1985; Coleman et al., 1985; Abel et al. 1986; Baulcombe et al., 1986, Grumet et al., 1987a, b; Loesch-Fries et al., 1987), the well-characterized genetics of the HIV virus (Wong-Staal & Gallo, 1985), the apparent localization of the virus in cell types that can be genetically altered (Wong-Staal & Gallo, 1985; Barnes, 1986; Gartner et al., 1986), and the theoretical stability of PDR in the face of new viral strains (Sanford & Johnston, 1985). Perhaps the greatest uncertainty regarding the potential efficacy of a gene therapy cure for AIDS involves the localization of HIV within cells that can be genetically altered. PDR PR INCIPLE AND A IDS 471 While it is clear that HIV infects T4 lymphocytes and other cells in the bloodstream and macrophages in the brain, it is possible that HIV may also infect brain neuron cells (Navia et al., 1986a,b; Johnson & McArthur, 1986). Such cells are not presently amenable to genetic modification, therefore they are not themselves subject to gene therapy. However, by curing blood cell infection, possible brain cell infection can be prevented where it has not yet occurred. If brains cells are actually infected, it might be hoped that the curing of blood cell infection and restoration of the immune system might reverse such brain infection, or at least stop its spread. Numerous viral target sites and diverse mechanisms for implementing the PDR principle become evident upon analyzing the genomic structure and the reproductive mechanisms of HIV. The basic viral functions to be blocked by anti-HIV genes are reverse transcription, proviral transcription and translation, and assembly and export of viral particles. A series of anti-HIV mechanisms blocking these processes will be discussed in the following sections. In the interest of simplicity, anti-viral gene products will be emphasized, rather than genes. (The basic structure at the actual genes can largely be deduced from their products.) In all cases, the virus' own sequences and machinery will be turned against itself, which is the essence of the PDR principle. Gene Products Designed to Block Reverse Transcription All retroviruses, including HIV, must undergo a complex process of reverse transcription within a host cell before that cell can successfully be infected. Therefore, blocking reverse transcription is a logical first line of defense against HIV and other retroviruses. BLOCKING THE FOUR HYBRID IZAT ION STEPS OF REVERSE TRANSCRIPT ION There are four critical steps in the reverse transcription of HIV, which require nucleic acid hybridization (Gilboa et aL, 1979). If any of these hybridizations are blocked, the infection process will be aborted. Each of these hybridization steps can theoretically be blocked by pre-hybridization of the critical sites in the viral genome to complementary molecules coded for by genetically-engineered "resist- ance" genes in the host cell. The general process of using DNA or RNA which will hybridize and block viral sequences has been termed "hybridization interference" (Zamecnik et aL, 1986). In the present case hybridization interference will be described as a means for blocking reverse transcription, but it can also be used to block mRNA translation and viral packaging. Some of these mechanisms have been discussed previously by Zamecnik et al. (Stephenson & Zamecnik, 1978; Zamecnik & Stephenson, 1978; Zamecnik et al., 1986), but not from the perspective of identifying coding sequences for use in gene therapy. BLOCKING PRIMER BINDING A retrovirus exists as a single-stranded RNA at the time of infection. The first hybridization step which is critical for reverse transcription involves the annealing 472 J.C. SANFORD of a primer complex (lysine tRNA from the host, complexed to the viral reverse transcriptase enzyme) to the primer binding site (PBS) in the central portion of the viral genome. Host-encoded RNA complementary to this general region and extend- ing beyond it (anti-PBS) should compete with the primer complex for this site, and once annealed to the site, should effectively block the initiation of reverse transcrip- tion from its proper starting point. The effectiveness of this block will depend on the relative numbers of primer complexes vs. anti-PBS molecules and their access to the site. BLOCKING THE "F IRST JUMP" OF REVERSE TRANSCRIPT ION The second nucleic acid hybridization required for successful HIV reverse tran- scription involves the "first jump" of the reverse transcriptase to a new template. Reverse transcription proceeds toward the 5' end of the virus, continuously degrading the RNA template as the cDNA is synthesized. When the enzyme reaches the 5' end of the RNA template, it apparently stops and backs up, leaving the single- stranded 3' end of the cDNA exposed. Because there is a redundant 98 base pair sequence at both ends of the RNA genome, which is called the R-sequence, the first 98 base pairs of the cDNA is complementary to the last 98 base pairs of the RNA genome. This allows nucleic acid hybridization, and circularization of the RNA-DNA complex. This is necessary for reverse transcription to continue. RNA which is either complementary to the R-region (thereby annealing to the viral RNA), or equivalent to the R-region (thereby annealing to the cDNA), should block circularization of the complex, and the infection is aborted. BLOCKING IN IT IAT ION OF PLUS-STRAND DNA SYNTHESIS The third nucleic acid hybridization required for successful reverse transcription involves initiation of plus-strand DNA synthesis. As reverse transcription proceeds past the R-region where the "first jump" has occurred, a region known as the "polypurine region" is transcribed into DNA. This DNA region is where plus-strand synthesis is initiated, beginning with the annealing of an RNA primer at this site. The reverse transcriptase enzyme recognizes this RNA/DNA complex, and causes a nick at a specific site in the primer, from where DNA synthesis begins (Resnik et aL, 1984; Smith et aL, 1984). Plus-strand DNA synthesis then proceeds in the opposite direction as reverse transcription, back trhough the R-region, making a complement of the original reverse transcription RNA primer. The hybridization of a primer to the polypurine region, and the subsequent initiation of DNA synthesis at the correct point, can theoretically be blocked by complementary RNA molecules spanning and extending beyond this region, but lacking perfect homology to the natural primer at the key nick recognition site. Even if such molecles act non- specifically as initiators of plus-strand synthesis, the DNA terminus will be shifted, eliminating the terminal repeats of the double-stranded DNA virus, which are essential for viral insertion into the host chromosome (Panganiban & Temen, 1983; Panganiban, 1985). PDR PR INCIPLE AND A IDS 473 BLOCKING THE " 'SECOND JUMP" OF REVERSE TRANSCRIPT ION The fourth nucleic acid hybridization required for successful HIV reverse tran- scription involves the "'second jump" of the reverse transcriptase to a new template. After the "first jump" has occurred, reverse transcription continues along the new template (from the 3' end of the viral genome), back towards the PBS site where the process first began. The RNA primer is displaced from the PBS region of the viral genome either by plus-strand synthesis or by reverse transcription, resulting in a break in the circular complex. Reverse transcription proceeds through the PBS region and at the end of the molecule it stops and backs up for a second time, leaving a single-stranded DNA complement of the PBS region exposed. Plus-strand DNA synthesis has produced a terminus complementary to this. Annealing of these ends results in re-circularization and allows for the continuation of the synthesis of viral DNA in both directions. This fourth nucleic acid hybridization can be blocked either by RNA complementary to the PBS region, or RNA equivalent to it. An RNA molecule complementary to the PBS region, as already discussed, also has the potential to interfere with the initiation of the reverse transcription process. BLOCKING REVERSE TRANSCRIPTION WITH FALSE PRIMERS AND FALSE TEMPLATES The reverse transcription process can in theory be sidetracked through the use of false priming at incorrect sites, and through the use of false templates which should simultaneously "disarm" primer complexes while producing anti-viral cDNA. Reverse transcription must begin in a specific place within the viral genome in order for the resulting cDNA to be infectious and have the correct termini. The initiation point for this process is controlled by the primer complex. This complex consists of a reverse transcriptase molecule bound to a tRNA molecule. The tRNA molecule, which is supplied by the host cell, has its 3' tail fully exposed when it is complexed to the enzyme, such that this tail can hybridize to the primer binding site (PBS) of the virus. In the case of HIV, it is lysine tRNA which is involved in the primer complex. It should be possible to create false primers by taking the natural lysine tRNA sequence and modifying it so that the "tail" (the last 18 base pairs at the 3' end) are complementary to some part of the HIV genome different from the PBS site. Such false primers might then complex with the reverse transcrip- tase enzyme and initiate reverse transcription from a new site in the virus. The result would be the progressive degradation of the viral RNA template and the production of an abortive cDNA with incorrect termini. In addition to using false primers, false templates could be used to interfere with viral reverse transcription. As previously stated there are less than 100 mature primer complexes which are carried into the cell by the retroviral particle. It would be highly desirable if these molecules could be made unavailable to the virus. One way of doing this would be through the use of false templates, which would contain PBS sequences and would bind and "disarm" the mature primer complexes. Once disassociated, functional reverse transcriptase/tRNA complexes could presumably not be reformed without the maturing process associated with viral packaging and 474 J~ c. SANFORD release. The effectiveness of false templates could be enhanced by making their 5' ends homologous to various target sites in the virus. Consequently, the false template would be reverse transcribed by the reverse transcriptase, to produce cDNA molecules which could have further anti-viral effects. The idea of employing the virus' own reverse transcriptase to help the host produce more anti-viral molecules from mRNA templates helps illustrate how the virus' own sequences and machinery can be turned against itself. Gene Products Designed to Block Proviral Expression All retroviruses, including HIV, once inserted into the host chromosome, must have their genes translated into viral proteins which function properly. It has already been shown that if viral proteins are not abundant, HIV cannot efficiently propagate to other cells and is not cytopathic to the infected host cell (Dayton, et aL, 1986; Fisher et al., 1986). Therefore, blocking proviral gene expression is the second logical line of defense against retroviruses. DERIV ING REPRESSOR PROTEINS FROM THE TAT AND ART PROTEINS It was previously proposed by Sanford & Johnston (1985) that a DNA- or RNA-binding viral protein could be modified to function as a highly specific viral repressor. Since then, it has been shown that functional DNA- or RNA-binding domains of proteins can be subcloned, and that by subcloning such domains, gene inducers or activators can be converted into gene repressors (Keegan et al., 1986; Smith et aL, 1984). One of the key RNA-binding HIV regulatory proteins is the TAT protein, which has been shown to be essential to the replication and cytopathic effect of HIV (Keegan et aL, 1986; Smith et aL, 1984). TAT was originally believed to bind to the 5' end of proviral DNA, and be involved in transcriptional activation (Sodroski et aL, 1985; Rosen et al., 1985). More recently, evidence has been provided that it binds to the 5' end of viral mRNA, and that it is involved in translational activation (Rosen et al., 1986). In either case, it is an essential activator protein which recognizes and binds to the nucleic acid sequence corresponding to the 5' end of the R-region of the virus. It appears that the RNA-binding domain of the TAT protein occurs in the second half of the molecule (Sodroski, et al., 1985). By engineering a gene coding for a protein which includes the RNA-binding domain, but is defective in other TAT domains, it is reasonable to expect that the gene product should bind to the TAT binding site, but should fail to activate it. Such a molecule should act as a viral repressor by blocking the binding of the functional TAT protein. Alternatively, the TAT coding region could be subjected to directed mutagenesis or systematic alteration in a way that preserved its RNA-binding capacity but abolished its activator properties. It is likely that a TAT-derived repressor would be a particularly effective anti-viral molecule, given TAT's pivotal role in HIV regulation, and given that relatively large amounts of the modified protein could be produced constituitively in the cell. Recently, a second trans-activating gene (ART) has been shown to be important in HIV translation (Sodroski, et al., 1986), and the putative RNA-binding domain PDR PR INCIPLE AND A IDS 475 of the protein has been identified. Presumably this activator gene could similarly be converted into a repressor. BLOCKING V IRAL mRNA SPL IC ING AND TRANSLAT ION BY ANT ISENSE RNA The validity of using antisense RNA to block protein translation has been well established (Izant & Weintroub, 1985; Peska, et al., 1984; Rosenberg, et aL, 1985). The concept of using this mechanism in the genetic engineering of virus resistance was put forward by Sanford & Johnston (1985), as part of the PDR principle, and the concept has already been found to condition resistance to viruses in model bacteriophage systems by Coleman et al. (1985), and Johnston & Sanford (unpub- lished). Although any part of the mRNA might be attached by antisense RNA, the most logical and effective points of attack would seem to be at the 5' leader region, the initiation codon regions, and RNA splicing regions (Coleman et al., 1985). The complement of the R-region, which has already been described as a molecule capable of blocking reverse transcription, is also one of the most promising molecules for blocking protein translation. This molecule should hybridize to the first 98 base pairs of the viral mRNA, thereby blocking recognition and binding to ribosomes. In addition, such hybridization should block the attachment of the TAT protein, which binds in this region of the mRNA and which is essential for activating efficient translation. Initiation codons and mRNA splice sites also appear to be important target sites to block, using antisense RNA. Because splicing involves RNA folding and single- strand sequence recognition, antisense RNA should be particularly effective in blocking such splicing events. HIV has numerous splice sites, four of which are essential for translation of the TAT protein (Arya et al., 1985). These four sites would appear to be especially promising targets for antisense RNA, particularly the first two, which are close enough to key translation initiation codons that single small RNAs would easily cover both the splice site and the initiation codon. The primary disadvantage of the antisense approach to blocking viral translation is that a very substantial numerical advantage is required by the interfering molecule over the mRNA. This may be very difficult to achieve where the transcriptional promoter of the virus is strong. Mechanisms for blocking HIV protein translation, whether they involve antisense RNA or repressor proteins, should be enhanced by a negative feedback mechanism associated with the TAT activator system. As viral protein synthesis is inhibited, the amount of TAT protein will also drop off, which will tend to further inactivate viral translation. Where a TAT-derived repressor is used, the modified repressor protein will become proportionately more abundant as the TAT is reduced, allowing the repressor to more effectively compete for the TAT binding site and might shut the viral system down completely. Gene Products Blocking Viral Packaging and Export Assuming a cell has been successfully infected by HIV, and that the viral machinery is still not fully shut down, a final line of defense still exists. This is to block viral 476 J . c . SANFORD packaging and export, such that the infection does not spread to other uninfected cells. Such packaging involves folding of the retroviral RNA genome based on internal homologies (Darlix et al., 1980), binding of viral protein to specific sequences of the viral genome (Darlix et al., 1982), and hybridization of viral genome pairs within complementary regions, to form infectious dimer structures (Darlix & Spahr, 1982). Most of the anti-HIV gene products discussed thus far can be expected to interfere with such packaging, including anti-R, anti-PBS and anti-splice RNA molecules and the TAT-derived repressor protein. In fact, if anti-R or anti-PBS hybridize to the viral genome prior to packaging, and the viral particle somehow is still packaged, subsequent reverse transcription will be pre-blocked, and the particle should not be infectious. Of particular interest in terms of blocking viral packaging are the regions involved with protein binding and dimer formation. In the Rous sarcome retrovirus, these regions are found just upstream of the GAG coding region (Darlix & Spahr, 1982; Dickson et aL, 1984). In HIV, this same region also contains the first splice site and the GAG initiation codon (Arya et al., 1985). Therefore, RNA complementary to this general region might be particularly effective in blocking viral packaging as well as blocking viral translation. Discussion A series of putative anti-HIV molecules have been designed using the PDR principle. These molecules, their corresponding HIV sequences, and their modes of action are summarized in Table 1. From a genome of only seven genes, at least 13 prospective anti-viral sequences can be derived, representing a diverse array of PDR mechanisms. FACTORS AFFECT ING THE EFFECT IVENESS OF ANTI -H IV GENES Some anti-HIV molecules are expected to be only of marginal utility, such as the homolog of the PBS-region or false templates, and have been described largely for theoretical interest. Other anti-HIV molecules are expected to have more pronounced effects, especially where they combine multiple anti-viral mechanisms. For example, anti-R can be expected to interfere with reverse transcription, viral RNA translation, and, very likely, viral assembly and packaging. The complement of the PBS region should block two separate aspects of reverse transcription. Antisense RNA com- plementary to the general region surrounding the first splice site of the H IV genome can be expected to interfere with viral RNA splicing, viral RNA translation, and, very possibly, assembly and dimer formation. Three of the molecules described above have already been shown experimentally to have anti-viral activity. Zamecnik et al. (Stephenson & Zamecnik, 1978; Zamecnik & Stephenson, 1978; Zamecnik et aL, 1986) have studied the effects of synthetic oligonucleotides on retrovirai translation and replication. These studies were apparently aimed at identifying useful pharmaceuticals, as opposed to identifying coding sequences for use in gene therapy. Stephenson & Zamecnik (1978) showed that a synthetic 11 base pair complement of the R-region of Rous Sarcoma Virus inhibited in vitro translation of that virus. Zamecnik & Stephenson (1978) showed PDR PR INCIPLE AND A IDS 477 TABLE 1 PDR sequences for use in construction of prospective anti-HIV genes. Respective HIV target sites and modes of action are shown Anti-HIV Anti-HIV sequence name sequence a HIV target site b Modes of action 1. Anti-R 1-97 (minus strand) 3' R-region of viral RNA Block " lst jump" of reverse transcrip- and of mRNA. tion, TAT binding, and translation of mRNA. Block " lst jump" of reverse tran- scription. Block initiation of reverse transcrip- tion and "2nd jump". 2. R homolog 1-97 (plus strand) 3. Anti-PBS 170-210 (minus strand) 4. PBS homolog 182-199 (plus strand) 5. False primer lys tRNA (with 3' 18 bp substitution) 6. False template PBS homolog (with 5' false tail) 7. Polypurine 8630-8670 (plus strand) homolog 8. Anti-splice I 270-340 (minus strand) 9. Anti-splice 2 5340-5430 (minus strand) 10. Anti-splice 3 5610-5640 (minus strand) II. Anti-splice 4 7940-7970 (minus strand) 12. TAT repressor 5530-5593 (plus strand) 13. ART repressor 7956-8080 (plus strand) 3' R-region of minus- strand cDNA. PBS site of viral RNA and of plus-strand cDNA. 3' PBS region of minus- strand cDNA. Any non-PBS site of viral RNA. Primer complex and secondary site. Polypurine complement in minus-strand cDNA. Accepter site for 1st TAT splice, and GAG initiation codon of mRNA. Donor site for Ist TAT splice, and TAT initi- ation codon of mRNA. Acceptor site for second TAT splice of mRNA. Donor site for second TAT splice of mRNA. 5' end of mRNA. ART-binding site of mRNA. Block "2nd jump" of reverse tran- scription. Initiates reverse transcription at improper site. "'Disarm" primers, produce anti-viral cDNA. Block proper initiation of plus-strand DNA synthesis. Block splicing needed for TAT trans- lation, and initiation of GAG trans- lation. Block splicing needed for TAT trans- lation, and initiation of TAT trans- lation. Block mRNA splicing needed for TAT translation. Block of mRNA splicing needed for TAT translation. Block binding of TAT activator. Block binding of ART activator. The sequence to be transcribed (or translated) is indicated. Numbering based on HTLV3/LAV (Weiss, 1985). b For exaplantion of retroviral sites see Weiss (1985). See text for detailed explanations. Most anti-HIV genes should also interfere with viral packaging and export. that the same synthetic oligonucleotide could be added to cells in culture as an extracellular agent, and could inhibit multiplication of the virus. More recently, Zamecnik et al. (1986) have shown that synthetic oligonucleotides which are com- plementary to either the PBS site or one of the TAT splice sites of HIV can inhibit HIV multiplication within cultured cells, when added extracellularly. The equivalent molecules, when occurring as native gene products within the host cell, should have similar activity. A major disadvantage of using a hybridization interference defense strategy is that the interfering molecule must occur in large excess relative to the target molecule. 478 J . c . SANFORD As soon as this numerical advantage is lost, the virus can overhwelm the system, causing the resistance to break down. The viral protection derived from the synthetic oligonucleotides in the above experiments probably resulted from high intracellular concentrations of the oligonucleotide. Such high levels may be difficult to achieve or maintain based solely on gene transcription. A second major disadvantage with hybridization interference is that even when the interfering molecule occurs in high concentration, the target molecule is only partially blocked (Izant & Weintraub, 1985; Peska et aL, 1984, Rosenberg et aL, 1985). The degree of inhibition can be very modest, and is very seldom as high as 90%. Fortunately, parasite-derived resistance is not limited to hybridization interference mechanisms. Thus far, the most dramatic examples of parasite-derived resistance have been where vital regulatory proteins have been employed (Grumet et aL, 1987a, b). The coat protein gene of QB has been found to condition extremely high levels of resistance to the phage, even when expressed in the host at low levels. The TAT-derived repressor would be a protein fragment analogous to the QB replicase- derived repressor (Sanford & Johnston, 1985). In theory, it should be the most potent anti-HIV gene product. As a translated gene product it should occur in much higher copy number than transcribed gene products, it should directly attack the "Achilles heel" of HIV, and it should benefit from a negative feedback mechanism which may inactivate the viral translation system completely. With the possible exception of the TAT-derived repressor, none of the gene products in Table 1 are expected to confer by themselves more than partial resistance to HIV. However, the imperfect or partial nature of such anti-viral activity does not mean that such activity will be ineffective in controlling the virus. Genes and coding sequences can be combined, and incremental or multi-component resistance is consistent with most natural forms of defense against pathogens. Natural defenses typically involve multiple lines of defense, each level of which is only partially effective. In combination, such defenses provide a very formidable barrier to that pathogen. To illustrate this point, if four different anti-HIV molecules were used in combination to block the four hybridization steps of reverse transcription and each step was blocked with 90% efficiency, then a given virus would only have one chance in 10 000 (0.14) of completing reverse transcription and successfully infecting the cell. The potential additivity of genetically engineered anti-viral genes has previously been demonstrated by Coleman et aL (1985), using three different anti- sense RNAs in the SP bacteriophage system. Lastly, even if resistance genes fail to actually eradicate HIV, it can be hoped that any remaining low-level infection would be asymptomatic and harmless to the patient. Given effective HIV resistance genes, it becomes important to ask if the resulting cellular resistance will be stable. It is known that HIV has an extremely variable genotype, and that multiple strains can arise in a given patient (Wong-Staal et aL, 1985, Hahn et aL, 1986; Benn et al., 1985). Such new strains can often escape previously formed immune responses, and might also circumvent resistance genes. In fact, it is well known that pathogens frequently override resistance genes by mutating to new virulent forms. Fortunately, this is not likely to occur in the case of PDR. The inherent stability of PDR has been discussed in broad terms elsewhere PDR PR INCIPLE AND AIDS 479 (Sanford & Johnston, 1985). In the case of HIV, the target sites in the virus genome such as the PBS-region, the TAT-region, and the R~-region are highly conserved. Simple point mutations in ~hese regions shou.ld not affect significantly hybridization or binding potential of anti-HIV molecules. Generally, only major rearrangements or deletions would prevent such hybridization or binding, which in these key regions would almost certainly prove deleterious or lethal to the virus. MOVING TOWARD A GENE THERAPY A gene therapy cure for AIDS appears justifiable and feasible and anti-HIV molecules have been designed, using the PDR principle, which alone or in certain combinations tentatively offer effective and stable HIV resistance. The actual implementation of a gene therapy cure would require a broad interdisciplinary research effort. Research would have to progress throughseveral levels, including: (1) construction of a series of anti-HIV genes; (2) testing of the effectiveness of such genes and-gene combinations in vitro; (3) testing the effectiveness of such genes with various promoters in vivo in animal model systems; (4) clinical trials in man. The complexities of human somatic gene therapy go far beyond the scope of this paper and have been reviewed elsewhere (Andersen, 1984). However, the prospect of a genetic cure for AIDS clearly has two important implications for gene therapy research which must be mentioned. First, a prospective genetic cure for AIDS increases the urgency for solving the technical problems standing in the way of human gene therapy. Second, emerging gene therapy approaches may have to be modified. It would appear that chimeric retroviral vectors containing HIV sequences would have a high probability of recombining with the natural HIV virus within infected patients undergoing treatment. This might increase the risk of creating new recombinant virulent viruses. Therefore, a gene therapy treatment for AIDS might require that anti-HIV genes only be delivered to patients using non-retroviral transformation systems and vectors. This work was carried out under a consulting contract with Greatbatch GenAid Ltd. Thanks is given to Betty Porterfield for her assistance in manuscript preparation. REFERENCES ABEL, P. P., NELSON, R. S., DE, B., HOFFMAN, N., ROGERS, S. G., FRALEY, R. T. & BEACHY, R. N. (1986). Science 232, 738-743. ANDERSEN, F. W. (1984). Science 226, 401-409. ARYA, S. K., GUO, C., JOSEPHS, S. F. & WONG-STAAL, F. (1985). Science 229, 69-73. BARNES, D. M. (1986). Science 232, 1091-1093. BAULCOMBE, D. C., SAUNDERS, G. R., BEVAN, M. W., MAYO, M. A. & HARRISON, B. D. (1986). Nature 321, 446. BENN, S., RUTLEGEE, R., FOLKS, T., GOLD, J., BAKER, L., MCCORMICK, J., FEORINO, P., PLOT, P., QUINN, Y. & MARTIN, M. (1985). Science 230, 949-952. COFFIN, J., HAASE, A., LEVY, J. A., MONTAGN1ER, L., OROSZLAN, S., TEICH, N., TEMIN, H., TOYOSHIMA, K., VARMUS, H., VOGT, P. & WEISS, R. (t986). Science 232, 697. 480 J . c . SANFORD COLEMAN, J., HIRASHIMA, A., INOKUCHI, Y., GREEN, P. J. & INOUYE, M. (1985). Nature (London) 315, 601-603. CURRAN, J. W., MORGAN, W, M., HARDY, A. M., JAFFE, H. W., DARROW, W. W., & DOWDLE, W. R. (1985). Science 229, 1352-1357. DARUX, J. L, SCHWAGER, M., SPAHR, P. F. BROMLEY, P. A. (1980). Nuc. Acids Res. 8, 3335-3354. DARLIX, J. L. & SPAHR, P. F. (1982). J. Mol. Biol. 1611, 147-161. DARLIX, U. L., ZUKER, M. & SPAHR, P. F. (1982). Nuc. Acids Res. 10, 5183-5196. DAYTON, A. l., SODROSKI, J. G., ROSEN, C. A., GOD, W. C. & HASELTINE, W. (1986). Cell 44, 941-947. DICKSON, C., EISENMAN, R., FAN, H., HUNTER, E. & TEICH, N. (1984). In: RNA Tumor Viruses. (Weiss, R., Teich, N., Varmus, H. Coffin, H., eds). New York: Cold Spring Harbor Laboratory, Cold Spring Harbor. pp. 513-648. FISHER, A. G., FEINBERG, M. B., JOSEPHS, S. F., HARPER, M. E., MARSELLE, L M., REYES, G., GONDA, M. A. & ALDOVINI, A. (1986). Nature 320, 367-371. FOLKS, T., POWELL, D. M., LIGHTFOOTE, M. M., BENN, S., MARTIN, M. A. & FAUCI, A. S. (1986). Science 23L 600-602. GARTNER, S., MARKOVITZ, P., MARKOVITZ, D. M., KAPLAN, M. H., GALLO, R. C. & PoPOVIC, M. (1986). Science 233, 215-218. GILBOA, E., MITRA, S. W., GOFF, S. & BALTIMORE, D. (1979). Cell 16, 93-100. GRUMET, R., SANFORD, J. C. & JOHNSTON, S. A. (1987a). In: Molecular Strategies for Crop Protection. (Vol. 48) p. 3. UCLA Symposium on Cellular and Molecular Biology. 48, 3. New York: A. R. Liss, Inc. GRUMET, R., SANFORD, J. C. & JOHNSTON, S. A. (1987b). Virology (in press). HAHN, B. H., SHAW, G. M., TAYLOR, M. E., REDFIELD, R. R., MARKHAM, P. D., SALAHUDDIN, S. Z., WONG-STAAL, F., GALLO, R. C., PARKS, E. S. & PARKS, P. W. (1986). Science 232, 1548-1553. lZANT, J. G. & WEINTRAUB, H. (1985). Science 229, 345-352. JOHNSON, R. T. & MCARTHUR, J. C. (1986). Trends Neurosci. 9, 9t. KEEGAN, L., GILL, G. & P'rASHNE, M. (1986). Science 231,699-704. LOESCH-FRIES, L. S., MERLO, D., ZINNEN, T., BURHOP, L., HILL, K., KRAHN, K., JARVIS, N., NELSON, S. & HALK, E. (1987). Embo. J. 6, 1845. NAVlA, B. A., JORDAN, B. D. & PRICE, R. W. (1986a). Ann. Neurol., in press. NAVlA, B. A., CHO, E. S., PETITO, C. K. & PRICE, R. W. (1986b). Ann. Neurol., in press. PANGANIBAN, A. T. & TEMEN, H. M. (1983). Nature (London) 306, 155-160. PANGAN1BAN, A. T. (1985). Cell 42, 5-6. PESKA, S., DAUGHERTY, B. L., JUNG, V., HOTTA, K. & PESKA, R. K. (1984). Proc. Natf Acad. Sci. U.S.A. 81, 7225-7528. RESNIK, R., OMER, C. A., & FARAS, A. J. (1984)..L Virol. 51, 813-821. ROSEN, C. A., SODROSKL J. G. & HASELTINE, W. (1985). Cell 41, 813-823. ROSEN, C. A., SODROSKI, J. G., GOH, W. C., DAYTON, A. I., LIPPKE, J. & HASELT1NE, W. (1986). Nature (London) 319, 555-559. ROSENBERG, N. B., PREISS, A., SEIFERT, E., JACKLE, H. & KNIPPLE, D. C. (1985). Nature (London) 313, 703-706. SANFORD, J. C. & JOHNSTON, S. A. (1985). J. Theor. Biol. 113, 395-405. SMITH, D. R., JACKSON, !. J. & BROWN, D. D. (1984). Cell 37, 645-652. SODROSKI, J. G., PATARCA, R., ROSEN, C., WANG-STAAL. F. & HASELTINE, W. (1985). Science 229, 74-77. SODROSKi, J., GOH, W. C., ROSEN, C., DAYTON, A., TERWILLIGER, E. & HASELTINE, W. (1986). Nature (London) 321,412-417. STEPHENSON, M. L. & ZAMECNIK, P. C. (1978). Proc. Natl Acad. Sci. U.S.A. 75, 285-288. WEISS, R., TEICH, N., VARMUS, H. & COFFIN, J. (1986). RNA Tumor Viruses. New York: Cold Spring Harbor Laboratory, Cold Spring Harbor. WEISS, R. (1985). In: RNA Tumor Viruses-2, eds. (Weiss, R., Teich, N., Varmus, H. & Coffin, J. eds). New York: Cold Spring Harbor Laboratory, Cold Spring Harbor. pp. 405-485. WONG-STAAL, F. & GALLO, R. C. (1985). Nature (London) 317, 395-403. WONG-STAAL, F., SHAW, G. M., HAHN, B. H., SALAHUDDtN, S. Z., POPOVIC, M., MARKHAM, P., REDFIELD, R. &. GALLO, R. C. (1985). Science 229, 759-762. ZAMECNIK, P. C. & STEPHENSON, M. L. (1978). Proc. Natl Acad. Sci. U.S.A. 75, 280-284. ZAMECNIK, P. C., GOODCHILD, J., TAGUCHI, Y. & SARIN, P. C. (1986). Proc. Nat lAcad. Sci. U.S.A. 83, 4143-4146.