EP0752000A1

EP0752000A1 - Packaging-deficient lentiviruses

Info

Publication number: EP0752000A1
Application number: EP95912349A
Authority: EP
Inventors: Andrew Michael Lindsay Lever; Geoffrey Philip Dept. of Medicine HARRISON; Eric The Univ. of Alabama at Birmingham HUNTER
Original assignee: Syngenix Ltd
Current assignee: Syngenix Ltd
Priority date: 1994-03-24
Filing date: 1995-03-24
Publication date: 1997-01-08
Also published as: JPH09510361A; WO1995025806A2; WO1995025806A3

Abstract

Packaging-deficient lentiviruses are described. In particular, a vector is capable of producing proteins of a virus selected from MVV (Maedi-Visna Virus) and HIV-2, but not of packaging the viral RNA.

Description

PACKAGING-DEFICIENT LENTIVIRUSES Field of the Invention

This invention relates to lentiviruses, recombinant DNA technology and novel vectors for gene transfer. Background to the Invention

All lentiviruses (a subgroup of retroviruses) encapsidate their RNA genome as a di er into the virion particle. The genomic RNA species is the full-length unspliced message which is also capable of translation, encoding the gag and pol genes of the virus. Lentiviruses are able to specifically encapsidate their genomic RNA, because unique cis-acting signals that are present within this RNA enable differentiation from other viral RNA species. A number of RNA regions in retroviruses are reviewed by Schlesinger et al (1994), Seminars in Virology ,5:39-49, and have been identified as containing packaging signals that enable encapsidation (packaging) of genomic RNA into a virion particle. The cis-acting messages involved are usually in the form of sequence and structural motifs and are found in the 5' coding regions close to the splice donor (from deletion mutagenesis studies) .

However the actual RNA protein interactions involved in packaging are poorly understood. The evidence suggests that an interaction of a RNA motif or motifs, with a component of the gag or gag /pol polyproteins is likely.

Apart from a four base motif (GACG) identified by Konings et al (1992) J. Virol j66:632-640, there is no sequence conservation between the identified packaging signal regions. This motif is often found in regions of lentiviral RNA (e.g. MPMV, SMRV-H) that are predicted to be unfolded. Therefore, whilst this motif may have an accessory function, the evidence so far suggests that it is not a major packaging signal. The packaging regions of various retroviruses have undergone further secondary structure RNA modelling using biochemical analysis and free energy minimisation studies in conjunction with phylogenetic studies. Harrison and Lever (1992) J. Virol 16:4144-4153, report a study involving the HIV-1 lentivirus, showing RNA encapsidation to be far more complex in this particular retroviruβ subgroup than in other retroviruses such as avian and murine viruses. Apparently, packaging signals both within and outside of the 5' leader sequence are required for encapsidation.

Once the lentiviral packaging signals have been identified and characterised, the prospect of therapeutic, prophylactic and/or diagnostic applications against lentiviral infections is tenable.

It would be desirable to generate vaccines against lentiviral infections. Traditional methodologies of vaccine production (heat, formalin, ethylene-immine) often lack efficacy as these treatments alter the antigenic nature of the material being presented to the host immune system.

A packaging-defective lentivirus can be generated, by inactivating the known packaging signals for use as a vaccine. This has been achieved in HIV-1 by Richardson et al (1993) J. Virol. 67:3997-4005; see also WO-A-9317118 and the commonly-owned copending US Patent Application Serial No. 08/295,737, filed August 26, 1994, the contents of which are incorporated herein by reference. The non- infectious virion protein coat (containing antigenic epitopes) would then be presented to the host immune system to confer protective immunity without risk of infection, as the lentivirus lacks the viral RNA that gives rise to pathogenicity. It is envisaged that 'subunit' type vaccines could be designed on the basis of further study of the antigenic epitopes to which the immune response is directed.

Lentiviral infections persist in cells of the macrophage/monocyte series and, in this immunologically privileged site, they are difficult to eradicate. Furthermore, pharmacological approaches to ameliorating lentiviral infection have proved difficult, due to the close relationship between virus and host processes. Even inhibitors of the virus-specific enzyme reverse transcriptase have demonstrable effects on host cell polymerases. Furthermore, lentiviruses can achieve a latent state in host cells, requiring lifelong therapeutic treatment which causes significant problems of drug toxicity due to accumulation.

Therefore heterologous RNA, in which a lentiviral packaging signal is incorporated, could be packaged into its corresponding virion particle (instead of the normal viral RNA) . This heterologous RNA may encode a gene which could then be specifically targeted to cells, using the hybrid virion, which are susceptible to a specific lentiviral infection.

Various systems for the introduction of foreign DNA into host cells have been described. Carrier particles and their use for gene therapy are the subject of, inter alia, WO-A-9204916 and WO-A-9211846. Carriers and methods described there may be suitable for the purposes of the present invention. Summary of the Invention

The present invention is based at least in part on the investigation of a stable secondary structure model for the 5' leader region of the prototype D-type retrovirus Mason Pfizer Monkey Virus, that is presented 3' of the primer- binding site. Using biochemical probing of RNA from this region in association with free energy minimisation, a stem-loop structure in the region has been identified, which from other studies has been shown to be important for genomic RNA encapsidation. The structure involves a highly stable stem of 5 G:C pairs terminating in a heptaloop. Comparison with a predicted structure for Squirrel Monkey Retrovirus with MPMV, shows a structure with an identical stem and a common ACC motif in the loop. Free energy studies of secondary structure in nine other retroviruses predict stem loops which have similar GAYC motifs.

More specifically, therefore, the present invention utilises the consensus sequence GAYC as a unique packaging signal when enclosed in a stem-loop with a positioning 5' to the gag initiation codon and where this motif is not involved in another known interaction, for example, the complementary sequence of the tRNA primer binding site in the leader sequence. If that motif is already base-pairing with another RNA such as the tRNA primer, then it is unlikely to be involved in packaging; however, if it is uninvolved in any known interaction and is in the appropriate position, it appears that the motif is involved. According to the present invention, therefore, a vector is capable of producing a lentivirus protein, but not of packaging lentivirus RNA. For this purpose, it may have a deletion of or in, or a non-functionalised version of, the GAYC motif. According to another aspect of the invention, one or more oligonucleotides including the GAYC motif, or a complementary sequence, i.e. acting as a mimetic of the binding site based on the information presented herein, are used to compete with viral proteins that interact with the RNA. The competition serves to prevent packaging of the viral RNA.

Alternatively, a novel recombinant molecule for use as a vector may be generated. This may comprise nucleotides corresponding to the packaging nucleotides of a lentivirus, a heterologous gene and, flanking the packaging nucleotides and the heterologous gene, sequences including the .GAYC motif sufficient for packaging, reverse transcription and integration of the vector into target cells, to enable expression of the heterologous gene. This aspect of the invention is based on the realisation that lentiviruses are the only retroviruses able to integrate their DNA into non-dividing cells, increasing the potential target population for gene transfer. The property seems to reside in the matrix protein which probably has a nuclear localisation signal and may transport ONA through the nuclear membrane. Incorporation of this protein into particles will enhance nuclear transport of introduced genetic sequences. Integration may require an integrase enzyme. Alternatively, if the DNA includes suitable homologous sequences it may recombine with chromosomal DNA by homologous recombination. This process would both excise any unwanted gene and insert the correct one.

In one specific embodiment, the targeted gene could encode anti-sense expressing gene specifically targeted at an MW or HIV-2 sequence or at a cellular sequence involved in MW or HIV-2 replication. The lentiviral inhibitory gene or genes (multiple antiviral genes could be contained in one or multiple vectors delivered to a cell) might be expressed constitutively in the cell. Alternatively, the gene might be engineered such that is under the control of MW or HIV-2 regulatory genes, e.g. tat gene, whereby it would remain transcriptionally silent until activated by an incoming lentivirus.

In another aspect of the present invention, a recombinant DNA molecule comprises a gene which is functionally dependent on a metal, e.g. under the control of a non-physiological metal-dependent promoter. This is based in part on the realisation that the control of expression of genes may be affected by upstream promoter sequences. Some of these are metal-dependent, e.g. the MT2 promoter; transcription can then be activated by the appropriate cation. Further, a metal may cause activation of an enzyme that controls expression.

In particular, the information herein can be applied to any of the subject matter claimed in WO-A-9317118 and said commonly-owned copending US Patent Application.

Lentiviruses to which this invention relate include MPMV, and others. Description of the Invention

Extensive and comprehensive free energy predictions have been made, from the sequences of MPMV, SRV-1 and 2, SMRV-H, and Mouse Mammary Tumour Virus (MMTV) , from the PBS to 300 nucleotides into the gag open reading frame. The structures of this region in 8 other more distantly related retroviruses were also examined. EMBL accession codes of all sequences referred to in the text are shown in Table 1. Free energy minimisation studies were carried out on a UNIX by folding lengths of sequences ranging from 100 to 300 bases at intervals of 30 bases from bases 700 to 1050. The programmes employed were MFold adapted for GCG (Genetics computer group University of Wisconsin) [Zucker (1989) Science 244:48-52. and Jaguar et al (1989) PNAS USA 8-6:7706-7710], with the graphical presentation of Squiggles [Osterburg and Sommer (1991) Computer Progs in Biomed. 12:101-109] in the GCG programme Plotfold. The MFold programme presents suboptimal foldings within 5 to 10% of the calculated minimum free energy. Biologically functional foldings may be found in suboptimal predictions, because the folding rules and energy parameters are not accurately known. Free energy calculations on the MPMV sequence repeatedly predicted the following stem loops in the following regions: 776 to 791, 837 to 862, 878 to 899, 902 to 946.

To probe for RNA secondary structure, in vitro transcribed RNAs were either digested with structure- specific enzymes or were modified with structure-specific biochemicals. The transcribed RNAs were generated from the 5' leader region of MPMV excised from the proviral clone pSHRMlδ [Rhee et al (1990) Eur. J. Bioch. 7:305-318] using the SphI site at position 567 and the Saσl site at 1561. The numbering is that of RESIVMPC (Genbank accession no. M12349) . This fragment was ligated into the same sites in the expression vector pGem-4Z (Promega, Southampton, UK) and RNA was transcribed from the T7 promoter. Biochemical and enzymatic probing was carried out essentially as described in Harrison and Lever (1992) . Briefly, after in vitro transcription, RNA samples were digested with DNAse 1, then were dissociated and reannealed by heating to 65^βC for 5 ins and cooling to 25^βC over 20 minutes. The RNAs were then divided into 2 μg aliquots and were modified with one of the following agents: Cobra venom RNase VI, (Pharmacia), hydroxymethylpsoralen (psoralen) (Sigma) , 2-keto-3-ethoxybutyraldehyde (kethoxal) (United States Biochemical) , RNAse Tl (Boehringer Mannheim) , and dimethyl sulphate (DMS) (Fluka) .

Photoreactions with psoralen were carried out under a 366 nm germicidal ultraviolet lamp. Samples were irradiated for 30 mins at room temperature in Eppendorf tubes at a distance of 20-30 mm from a Sylvania G875 tube light.

Modified RNAs were extracted with phenol/chloroform and 10 ng of a synthetic oligonucleotide primer was added in avian myeloblastosis virus (AMV) reverse transcriptase (RT) buffer (Pro ega) . Oligonucleotide primers which were used were: 5^,-TTG/CCC/CAT/ATC/CGA/GCG/C-3^, from bases 898 to 880. 5'-CCC/CGT/GTC/TTT/AAA/GCC/-3' from bases 937 to 939 and 5'-TAT/GGT/TCC/CTC/TTG/CGG/-3' from bases 1042 to 1025.

The RNAs were dissociated by heating the solution to 70^βC for 5 minutes. The solutions were then immediately stored on ice. It was found that the primer annealed when the solutions were incubated with the extension mix. Extension analyses were carried out by adding 1 unit of AMV RT (Promega) , a final cone of 1 mM each dATP, dTTP and dGTP and 1 μl [α³²P]dATP. These extension mixtures were incubated at 42°C for 1 hour. cDNAs were precipitated under ethanol, and redissolved in 1 x TE buffer. 1/lOth volume of formamide dye mix was added and approximately equal numbers of P counts were loaded onto 6% polyacrylamide-7M urea gels along with dideoxy-sequencing ladders [Sanger et al (1977) PNAS USA 79:5463-5467] generated using the same oligonucleotide primer. Cleavages and modifications caused by these agents make reverse transcriptase terminate or pause during DNA synthesis, which results in the generation of unique or more intense bands on autoradiographs. The position of these modifications can be determined by reading a dideoxy- sequencing ladder on the gel, which has been generated using the same oligonucleotide primer (at one base 3' to the RT termination or pause) . Pauses or stops in the cDNA synthesis give rise to bands one nucleotide shorter than the corresponding band in the dideoxysequencing ladder, since cDNA synthesis stops at the nucleotide immediately preceding the modified position or cleavage site. This is because the cDNA cannot complement the modified or cleaved nucleotide. Bands in the lanes of the untreated RNA form a reproducible pattern of stops [Shelness and Williams (1985) J. Biol. Chem. 260:8637-8646] where RT pauses or dissociates from the template, for reasons which are not yet understood [Bebenek et al (1989) J. Biol. Chem. 264:16948-19656. and Fry and Loeb (1992) PNAS USA 8£:763- 767].

In any biochemical and enzymatic study of RNA secondary structure, it should be noted that RNA in solution is in a dynamic state, and that alternative structures of the same regions exist simultaneously in a population of homologous molecules. Any individual RNA molecule may alternate between different conformations. This leads to some ambiguity in interpretation of biochemical data, as single stranded modifiers will be able to attack regions of RNA even if they are only transiently unpaired. A very stable double helix will however be much less vulnerable to single-stranded modification than a region which has no nearby complementary sequence.

Psoralen intercalates in base-paired regions, and forms covalent adducts with pyrimidines, notably uridines when irradiated with ultra-violet light at 366 nm [Youvan and Hearst (1982) Anal. Biochemistry 119:86-89_r and Bachellerie et al (1981) Nucleic Acids Res. 2:2207-2222]. Psoralen is found to intercalate particularly at the end of helices or in mismatches within helices [Leffers et al (1988) J. Mol. Biol. 104._/ 507-522]. Kethoxal modifies unpaired G residues, and DMS modifies A's and C's at positions which interfere with reverse transcription. RNAse Tl cleaves single-stranded RNA at G residues.

Conservation of the stem-loop, from 838 to 863, between MPMV, SRV-1 and SRV-2, was found. Together with the existence of an identical stem in SMRV and SMRV-H, this is strong evidence for its existence as a defined structure in vivo, and for its having an important function. Although being upstream from the SD, and thus not unique to the genomic RNA, the site is believed to be an important part of a packaging signal in MPMV.

Fig. 1 gives a comparison of predicted stem loops in the 5' leader regions of 11 different retroviruses. The murine structures which are boxed together were proposed by Tounekti et al (1992), J. Mol. Biol. 223.:205-220. The MoMuLV structure was first proposed by Alford et al (1991) , Virology 183:611-619. Sequence variations for SRV-l and 2 are indicated on the MPMV stem loop. The position of the gag AUG relative to the last C of the ACC motif is indicated for each retrovirus. In the 847-853 loop of the MPMV structure, there is a triplet ACC (849 to 851) which is conserved between MPMV and SMRV-H. In SRV-1 and SRV-2, the purine is conserved. Free energy predictions for B-type viruses and a series of distantly related retroviruses including gibbon-ape leukaemia virus (GALV) , [AKR murine leukaemia virus (AKV) , FBJ-murine osteosarcoma virus (FBJ-MSV) , Friend spleen focus-forming virus (SFFV) , FBR-MSV and caprine arthritis and encephalitis virus (CAEV) , show similar motifs in stem loop structures in analogous positions relative to the gag initiation codon (Figure 3) which conform to the general pattern GAYC (motif) . This is also found in bovine leukaemia virus (BLV) . The MPMV motif RCC is also seen in SNV and feline leukaemia virus (FeLV) at position 784. The stem and loop 878-898 containing the gag initiation codon are conserved between MPMV, SRV-1 and SRV-2 and SMRV and can also adopt a similar structure. The evidence strongly suggests that the GAYC motif may represent a common structural and sequence packaging motif in these viruses. Deletion mutations involving the region between the major splice site (nucleotides 303 to 304) and the gag initation codon (nucleotide 489) of the Maedi-Visna virus (MW) , which afflicts sheep, were constructed - see Figure 2. These mutations in the 5' untranslated region lead to a defect in the encapsidation of the genomic RNA of MW, without affecting production of MW proteins [nucleotide numbering as in Andresson et al (1993) Virology, 193:89- 105]. Thus the cis-acting replication defect observed is consistent with the deletion of a critical packaging signal for MW.

Figure 3 shows deletions in an analogous region between nucleotides 493 and 534 in the untranslated leader sequence in Human Immunodeficiency Virus type 2 (HIV-2) . These resulted in a similar packaging defective phenotype. The given nucleotide numbering is as in Guyader et al (1987) Nature 326:662-669.

A second packaging signal will be required in both of these viruses to enable encapsidation of heterologous RNA. This is located at the 3' end of the genome in a region adjacent to the σis-acting sequence through which the rev gene product acts. The rev responsive sequence will in part be different from the second packaging signal. The invention thus allows the construction of a replication-defective MW or HIV-2. It allows the production of viral particles devoid of viral nucleic acid. These could be used as vaccines. It allows the production of particles into which RNA containing the necessary packaging signals of MW or HIV-2 can be packaged. These latter RNAs can contain heterologous genes and thus MW and HIV-2 can be developed as gene vectors for delivery of genes to cells for which these viruses are tropic. The invention allows the use of the intrinsic property of these two viruses to integrate a DNA copy of the genetic material carried within the virus into a target cell even if that cell is not undergoing mitosis.

Creation of a packaging-defective MW or HIV-2 leads to the production of MW or HIV-2 virus particles which are non-infectious but which have all the correct MW or HIV-2 proteins assembled in their native form. They are therefore suitable for use as a vaccine. A similar approach might subsequently be used for other lentiviral disease of domestic animals such as Equine Infectious Anaemia virus and Caprine Arthritis-Encephalitis virus. Vaccine formulations may be of conventional nature. In one application of the present invention, the desired gene can be targeted to the appropriate cell, e.g. by using a particle or vector. A dose of the appropriate cation can be targeted to the same cell (particles) . The cation may be, for example, cadmium (which is rare, non- physiological, and toxic if in the wrong place at the wrong concentration) . The double targeting step provides a desirable degree of safety, cell-targeting specificity and also flexible control of transcription, by giving doses of cation-containing particle when desired. The foreign gene may be, for example, for a ribozyme. Ribozymes are RNA enzymes capable of cleaving RNA target sequences. They are potential antiviral agents. Their function may be to switch off expression of other genes, e.g. oncogenes. Ribozymes can be encoded on plasmids grown in bacteria. Bacterial growth may be made ribozyme-dependent, i.e. the ribozyme must be expressed to inhibit some toxic gene; the ribozymes will then evolve. Bacteria will be selected out in which the ribozymes are mutated to cleave the target sequence more efficiently. This is very rapid, since there can be many generations of bacteria per day. Ribozymes are divalent cation-dependent, probably for stability rather than at the active site. By including in the medium for growth high concentrations of a particular non-physiological cation, e.g. strontium, ribozymes may be evolved which work optimally or only in the presence of the cation. An antiviral/anti-oncogene ribozyme is then made, and into it is incorporated the cation-dependent sequence. The ribozyme may now be delivered to the target cell, followed by the strontium or other cation. The delivery of each may be achieved using particles, for example by a technique described in WO-A-9204916 or WO-A-9211846. The ribozyme cannot be activated in cells without the cation, and therefore will not inhibit physiological processes, e.g. a normal short burst of expression of the oncogene.

TABLE 1

EMBL

Abbreviation Full name accession no.

BaEV Baboon endogenous retrovirus X05470

BLV Bovine leukaemia virus K02120

CAEV Caprine arthritis encephalitis virus M33677

FeLV Feline leukaemia virus D00732

GALV Gibbon ape leukaemia virus K02989

MMTV Mouse mammary tumour virus M15122

MPMV Mason-Pfizer monkey virus Ml 2349

MoMuLV Moloney murine leukaemia virus J02255/6/7

MoMSV Moloney murine sarcoma virus J02266

SFFV Spleen focus forming virus K0021

SNV Spleen necrosis virus V01200

SMRV Squirrel monkey retrovirus M26927

SMRV-H Variant of SMRV M23385

SRV101 Simian retrovirus 1 M17561

SRV1LT Simian retrovirus 1 M17560

SRVRV1 Simian retrovirus 1 Ml 1841

SRV-2 Simian retrovirus 2 Ml 6605

Claims

I. A vector that is capable of producing proteins of a virus selected from MW (Maedi-Visna Virus) and HIV-2, but not of packaging the viral RNA.

2. A vector according to claim 1, which has a deletion corresponding to that between the major splice site and the gag initiation codon.

3. A vector that is capable of producing lentivirus protein and that has been modified by deletion or other non-functionalisation of a stem-loop structure comprising a GAYC motif or of 5 G:C pairs terminating in a loop containing an ACC motif.

4. A vector comprising nucleotides corresponding to the packaging nucleotides of a virus as defined in any preceding claim, a heterologous gene and, flanking the packaging nucleotides and the heterologous gene, sequences corresponding to those within and near the virus LTR sufficient for packaging, reverse transcription and integration of the vector into target cells, to enable expression of the heterologous gene.

5. A vector, for transferring a DNA sequence, that comprises a lentiviral DNA-integrating sequence and the DNA sequence to be transferred.

6. A vector according to claim 5, in which the integrating sequence encodes the matrix protein.

7. A vector according to claim 5 or claim 6, which also comprises a DNA sequence encoding an integrase enzyme.

8. A vector, according to claim 7, wherein the DNA sequence to be transferred is an inducible promoter sequence.

9. A vector according to claim 8, in which the inducible promotor sequence is activated by a metal.

10. A vector according to any of claims 4 to 9, which comprises the viral sequences as promoter for the heterologous gene.

II. A vector according to any of claims 4 to 10, wherein the heterologous gene is for a product useful in therapy.

12. A vector according to claim 11, for use in therapy by transfer of the foreign gene.

13. A vector according to any preceding claim, which comprises an undisrupted gag region.