Government through the National Institutes of Health under grant number CA- 4865409 and the United States Government may own certain rights in this invention.

BACKGROUND OF THE INVENTION Vaccinia virus is a member of the poxvirus family of DNA viruses.

Poxviruses including vaccinia virus are extensively used as expression vectors since the recombinant viruses are relatively easy to isolate, have a wide host range, and can accommodate large amounts of DNA.

The vaccinia virus genome contains nonessential regions into which exogenous DNA can be incorporated. Exogenous DNA can be inserted into the vaccinia virus genome by well-known methods of homologous recombination. The resulting recombinant vaccinia viruses are useful as vaccines and anticancer agents.

The use of vaccinia virus recombinants as expression vectors and particularly as vaccines and anticancer agents raises safety considerations associated with introducing live recombinant viruses into the environment. Virulence of vaccinia virus recombinants in a variety of host systems has been attenuated by the deletion or inactivation of certain vaccinia virus genes that are nonessential for virus growth.

However, there remains a need in the art for the development of vectors that have reduced pathogenicity while maintaining desirable properties of wild-type virus, such as host range, and active protein synthesis of a desired gene product.

SUMMARY OF THE INVENTION The present invention provides a recombinant vaccinia virus from which the region encoding an N-terminal portion of the E3L gene product has been deleted. In a preferred embodiment, the region encoding the N-terminal 83 amino acids of the E3L gene product has been deleted. The recombinant vaccinia virus of the invention may further comprise exogenous DNA.

The present invention further provides a composition comprising the recombinant vaccinia virus of the invention and a carrier.

DETAILED DESCRIPTION OF THE INVENTION The vaccinia virus E3L gene codes for double-stranded RNA binding proteins, and has been shown to be necessary for the vaccinia virus interferon-resistant phenotype. The E3L gene product of the vaccinia virus is a 190 amino acid polypeptide. Amino acids 118 to 190 have been implicated in dsRNA binding, as disclosed by Kibler et al. (1997) J. Virol. 71: 1992, incorporated herein by reference.

The present invention provides a recombinant vaccinia virus from which the region of the viral genome encoding an N-terminal portion of the E3L gene product has been deleted. An N-terminal portion of the E3L gene product is defined herein as including at least amino acids 1 through 37 of the E3L gene product. Amino acid numbering as used herein is adopted from Goebel et al. (1990) Virologv 179: 247-66,577-63, the disclosure of which is incorporated herein by reference. An N- terminal portion of the E3L gene product as defined herein may encompass the region from the N-terminus (amino acid 1) up to and including amino acid 117. Accordingly, a region encoding at least 37, and as many as 117, contiguous N-terminal amino acids of the E3L gene product is deleted from the recombinant vaccinia virus of the present invention.

In a preferred embodiment, the region of the viral genome encoding the N-terminal 83 amino acids of the E3L gene product has been deleted. In this preferred

embodiment, the recombinant vaccinia virus of the present invention contains a nucleic acid fragment encoding amino acids 84-190 of the E3L gene product instead of the gene encoding amino acids 1-190 of the E3L gene product at the E3L locus of vaccinia virus.

The recombinant vaccinia virus may further contain exogenous, i. e., nonvaccinia virus, DNA. Exogenous DNA may encode any desired product, including for example, an antigen, an anticancer agent, or a marker or reporter gene product. The recombinant vaccinia virus may further have deletions or inactivations of nonessential virus-encoded gene functions. Nonessential gene functions are those which are not required for viral replication in a host cell.

The recombinant vaccinia virus of the present invention may be constructed by methods known in the art, and preferably by homologous recombination. Standard homologous recombination techniques utilize transfection with DNA fragments or plasmids containing sequences homologous to viral DNA, and infection with wild-type or recombinant vaccinia virus, to achieve recombination in infected cells. Conventional marker rescue techniques may be used to identify recombinant vaccinia virus. Representative methods for production of recombinant vaccinia virus by homologous recombination are disclosed by Piccini et al. (1987) Methods in Enzvmologv 153: 545, the disclosure of which is incorporated herein by reference.

For example, the recombinant vaccinia virus of a preferred embodiment of the present invention may be constructed by infecting host cells with vaccinia virus from which the E3L gene has been deleted, and transfecting the host cells with a plasmid containing a nucleic acid encoding amino acids 84-190 of the E3L gene product flanked by sequences homologous to the left and right arms that flank the vaccinia virus E3L gene. The vaccinia virus used for preparing the recombinant vaccinia virus of the invention may be a naturally occurring or engineered strain. Strains useful as human and veterinary vaccines are particularly preferred and are well-known and commercially available. Such strains include Wyeth, Lister, WR, and engineered deletion mutants of Copenhagen such as those disclosed in U. S. Patent 5,762,938, which is incorporated herein by reference.

Recombination plasmids may be made by standard methods known in the art. The nucleic acid sequences of the vaccinia virus E3L gene and the left and right flanking arms are well-known in the art, and may be found for example, in Earl et al. (1993) in Genetic Maps: locus macs of complex gnomes, O'Brien, ed., Cold Spring Harbor Laboratory Press, 1.157 the disclosure of which is incorporated by reference, and Goebel et al. (1990), supra. The amino acid numbering used herein is adopted from Goebel et al. (1990), supra. The vaccinia virus used for recombination may contain other deletions, inactivations, or exogenous DNA as described hereinabove.

Following infection and transfection, recombinants can be identified by selection for the presence or absence of markers on the vaccinia virus and plasmid.

Recombinant vaccinia virus may be extracted from the host cells by standard methods, for example by rounds of freezing and thawing.

The resulting recombinant vaccinia virus may be further modified by homologous recombination to provide other deletions, inactivations, or to insert exogenous DNA.

It has been discovered in accordance with the present invention that a recombinant vaccinia virus having a deletion of the DNA encoding a N-terminal portion of the E3L gene product, and preferably amino acids 1-83 of the E3L gene product, maintains viral replication, protein synthesis, inteferon-resistance and cell tropism that is indistinguishable from wild-type virus, but has remarkably reduced pathogenicity in mice relative to wild-type vaccinia virus of the same strain.

The present invention further provides a composition comprising the recombinant vaccinia virus of the invention and a carrier. The term carrier as used herein includes any and all solvents, diluents, dispersion media, antibacterial and antifungal agents, microcapsules, liposomes, cationic lipid carriers, isotonic and absorption delaying agents, and the like.

The recombinant vaccinia viruses and compositions of the present invention may be used as expression vectors in vitro for the production of recombinant gene products, or as delivery systems for gene products, as human or veterinary vaccines, or anticancer agents. Such utilities for recombinant vaccinia viruses are known in the art, and disclosed for example by Moss (1996)"Poxviridae:

The Viruses and Their Replication"in Virology, Fields et al., eds., Lippincott-Raven, Philadelphia, pp. 2637-2671, incorporated herein by reference.

All references cited herein are incorporated in their entirety.

The following nonlimiting examples serve to further illustrate the invention.

Example 1 Construction of Recombinant Vaccinia Virus The plasmid pMPE3AGPT (described by Kibler et al. (1997) J. Virol.

71: 1992, incorporated herein by reference) was used for recombining a truncated E3L gene into the E3L locus of the WR strain of vaccinia virus. The recombination plasmid pMPE3AGPT is a derivative of pBSIISK (Stratagene, La Jolla, CA) that has had the p-galactosidase sequences deleted, and that contains sequences homologous to the left and right arms flanking the vaccinia virus E3L gene, but that lacks the E3L gene itself. The recombination plasmid contains the E. coli gpt gene outside the E3L flanking arms and thus allows for selection of transfected cells by treatment with mycophenolic acid (MPA).

The Aat II (blunt-ended) Sal I fragment of E3L was subcloned into the pGEM3-5T vector (described by Chang et al. (1993) Virolosv 194: 537, the disclosure of which is incorporated by reference) and subsequently cloned into the pMPE3AGPT recombination plasmid using Bam HI and Hind III restriction sites. The E3L fragment encodes amino acids 84-190 of the E3L gene product as numbered by Goebel et al (1990), supra, and has a deletion of the DNA encoding the N-terminal amino acids 1- 83. The plasmid resulting from the cloning of the E3L fragment into pMPE3AGPT is designated pMP-A83N.

In vivo recombination was performed in baby hamster kidney (BHK) cells. Subconfluent BHK cells were simultaneously infected with the WR strain of vaccinia virus deleted of the E3L gene (WRAE3L) at a multiplicity of infection (MOI) of 5 and transfected with l, ug of pMP-1v83N using Lipofectace (Gibco BRL).

WRAE3L was prepared by replacing the E3L gene from the WR strain of vaccinia

virus with the lacZ gene, by homologous recombination with pMPE3AGPT in which the lacZ gene was inserted between the E3L flanking arms.

Thirty hours post infection, the cells were harvested and recombinant virus was subjected to selection as follows. Virus was extracted from infected/transfected cells by three rounds of freezing and thawing and used to infect confluent BHK cells that had been pretreated for six hours with MPA selection medium (Modified Eagle Medium (MEM) containing 10% fetal bovine serum (FBS), 10, ug/ml mycophenolic acid, 250, ug/ml xanthine, 15/g/ml hypoxanthine). Following infection, cells were overlayed with MPA selection medium. At 24-72 hours post infection, plaques were visible and dishes were overlayed with MPA selection medium containing 0.5% molten agarose and 0.4, ug/ml X-gal (5-bromo-4-chloro-3- indolyl-p-galactoside). Blue plaques were isolated four to six hours after X-gal overlay. Two more rounds of MPA selection were performed on the isolated blue plaques.

Resolution of the in vivo recombination occurs when the MPA selection medium is removed, resulting in either recovery of the original virus, WRAE3L (containing lacZ in the E3L locus) or a recombinant virus containing the A83N deletion of E3L in the E3L locus. MPA-resistant blue plaques were used to infect untreated rabbit kidney RK13 cells. At 24-48 hours post infection, dishes were overlayed with MEM medium containing 0.5% molten agarose and 0.4, ug/ml X-gal.

Both blue and clear plaques were visible. Blue plaques indicate resolution of WRAE3L with lacZ in the E3L locus. Clear plaques indicate resolution of virus containing the A83N deletion of E3L in the E3L locus.

Two more rounds of infections with clear plaques were performed to purify plaques containing WRA83N. Recombinant virus was amplifie in RK13 cells.

Nucleic acid sequencing was used to confirm that the A83N fragment of E3L of plasmid pMP-A83N was present in the recombinant virus. Viral DNA was extracted from cells infected by WRA83N. Infected cells were freeze-thawed three times, followed by a thirty second sonication. Cell debris was removed by centrifugation at 700 x g for ten minutes. Nucleic acid was obtained by

phenol/chloroform extraction of the supernatant, and PCR was performed using primers to the E3L flanking arms. The PCR reaction products were subjected to agarose gel electrophoresis. DNA was extracted from the band of interest and DNA sequencing was performed. The A83N fragment was identified by sequence comparison to the plasmid DNA sequence of pMP-A83N.

Example 2 Host Range and Interferon Resistance of WR, WRA83N and WRA3L Wild-type vaccinia virus of the WR strain (WR) and variants WRAE3L and WRA83N as described in Example 1 were assessed for interferon resistance as follows.

RK13 cells were set down in six well tissue culture dishes at 70-80% confluency. Cells were treated with varying concentrations of rabbit interferon alpha (0-1000 U/ml) for sixteen hours prior to infection. Cells were infected with approximately 100 plaque forming units (pfu) of WR, WRAE3L or WRA83N virus.

Dishes were stained with crystal violet 24 hours post infection and plaques were counted.

WRAE3L exhibited interferon sensitivity (as measured by plaque reduction) at a concentration of 10 Units/ml of interferon, whereas WRA83N and WR did not exhibit interferon sensitivity at 10 or 100 Units/ml, but both showed plaque reduction at a concentration of 1000 Units/ml.

The foregoing results indicate that WRAE3L is sensitive to the effects of interferon, and that WRA83N, like WR, exhibits an interferon-resistant phenotype.

WR, WRAE3L and WRA83N were assayed for host range as follows.

Six-well tissue culture dishes containing RK13 cells or HeLa cells were set down simultaneously at 70-80% confluency. Both cell types were infected with equal dilutions of virus, and 24-48 hours post infection cells were stained with crystal violet and plaques were counted for each cell type. A comparison was made by determining the efficiency of plaquing (number of plaques in HeLa cells divided by number of

plaques in RK13 cells) for each virus. The efficiencies of plaquing were: WR: 0.98; WRA83N: 1.06; WRAE3L: <0.01.

These results indicated that WRAE3L has a restricted host range in that it cannot replicate in HeLa cells but exhibits nearly wild-type replication in RK13 cells. WRA83N, like wild-type WR, replicates in RK13 cells and HeLa cells.

The foregoing results show that WR and WRA83N are identical with respect to host range and interferon resistance in the cultured cells evaluated, whereas WRAE3L is sensitive to interferon and has a restricted host range.

Example 3 Virulence ofWR. WRAE3L and WRA83N Virus (WR, WRAE3L or WRA83N) was amplified by infection of RK13 cells until 100% CPE (cytopathic effect) was observed. Cells were scraped and resuspended in 1 mM Tris, pH 8.8. Amplified virus was freeze-thawed three times to release virus from cells. Debris was removed by centrifugation at 700 x g for 10 min.

Supernatant was used for mouse infections. Various dilutions of virus in 1 mM Tris, pH 8.8 were used in the experiment to determine LD50.

Three to four week old c57bl6 mice were anesthetized by intrafemoral injection of a cocktail of ketamine, acepromazine, and xylazine. Mice were subsequently infected with 10, ul of virus or a dilution of virus intranasally using a pipetman and gel loading tip. Mice were then replaced in their cages and observed daily for pathogenesis and death.

Intranasal inoculation with WR resulted in death at 104 pfu, whereas no pathogenesis could be detected with WRAE3L at the highest dose. For inoculation with WRA83N, 107 pfu was required for death, indicating that the amino-terminus of E3L is an important determinant for virus virulence.