Analysis of mutant Moloney murine leukemia viruses containing linker insertion mutations in the 3' region of pol

Twelve linker insertion mutations have been constructed in the 3' part of the pol gene of Moloney murine leukemia virus. This region of the Moloney murine leukemia virus genome encodes IN or p46pol, which is required for integration of the retroviral DNA into the host cell chromosome. Viral proteins synthesized by these mutants were used to pseudotype a neo-containing retroviral vector. Ten of twelve linker insertion mutant pseudotypes were unable to generate stable proviruses in infected mouse cells, as measured by the formation of G418-resistant colonies. Two mutants mapping at the 3' terminus of the IN-encoding region were competent for the formation of stable vector proviruses (hundreds of G418-resistant colonies per mutant pseudotype-infected plate). Representative linker insertion mutants were also tested for the ability to synthesize viral unintegrated DNA in newly infected cells. All assayed mutants were capable of synthesizing all normal forms of viral unintegrated DNA. The structure of integrated vector proviruses generated by defective and nondefective linker insertion mutants was also analyzed. All replication-competent mutants generated normal proviruses, while the few obtainable proviruses generated by replication-defective mutants were sometimes aberrant in structure. These results argue strongly (and confirm previous data) that the IN-encoding region of pol does not play a significant role in DNA synthesis, but is absolutely required for the formation of normal proviral DNA.

Retroviral DNA integrated during infection by an integration-deficient mutant of murine leukemia virus is oligomeric

Mutants with lesions engineered into the 3' region of the pol gene of murine leukemia virus (MLV), the region that encodes p40pol, have allowed us to define a function that is required for the normal integration of retroviral DNA. Two such mutants, MLV-SF1, which bears a single missense mutation (L. A. Donehower and H. E. Varmus, Proc. Natl. Acad. Sci. USA 81:6461-6465, 1984), and MLV-SF2, which has a frameshift mutation at the same site, have been used to explore the consequences of integration deficiencies. Pseudotypes that were formed with viral proteins supplied by these mutants and a replication-defective genome that carried a selectable genetic marker induced colony formation at least 100-fold less efficiently than did pseudotypes that were formed with proteins encoded by wild-type helper virus. Most of the proviruses from the mutant pseudotypes were manifestly aberrant when they were analyzed by mapping with restriction enzymes. Rare proviruses that appeared to be normal by restriction enzyme mapping were molecularly cloned and subjected to more detailed study. Two proviruses resulting from infection with the MLV-SF1 pseudotype were identical or nearly identical to wild-type MI V proviruses. Two base pairs were missing from the ends of the long terminal repeats at the host-viral junctions, and 4- or 5-base-pair duplications of host DNA flanked the proviruses. Thus, the missense mutant appears to retain a low level of normal integration activity. Four proviruses from cells that were infected with the MLV-SF2 pseudotype were flanked on both sides by viral DNA of either helper or vector origin, with a single long terminal repeat at each end of the intact vector DNA and with irregular junctions of viral and host DNAs. These results are most simply explained by the nonspecific integration of dimeric or trimeric forms of viral DNA, which may arise during infection by replicative or recombinational mechanisms.

A mutant murine leukemia virus with a single missense codon in pol is defective in a function affecting integration

We have used site-directed mutagenesis of cloned Moloney murine leukemia virus (MuLV) DNA to define a function encoded in the 3' region of the viral pol gene and required for efficient integration of viral DNA. One mutant, MuLV-SF1, contained a single base substitution (C to T at base 4950) that resulted in an arginine to cysteine change in a region highly conserved among retroviruses. Mutant DNA, introduced into rat cells by cotransfection with a herpes simplex virus thymidine kinase gene (HSV tk), directed production of virus particles with reverse transcriptase activity. Infection of cells with these particles led to synthesis of full-length linear and circular forms of unintegrated viral DNA; however, integrated viral DNA was decreased at least by a factor of 10 when examined by DNA hybridization, and the mutant particles were less efficient then wild-type virus at establishing an infection by a factor of at least 300. Pseudotypes formed with the proteins of MuLV-SF1 and the genome of a replication defective marker MuLV, carrying the HSV tk gene, were less effective by at least a factor of 100 in producing tk+ colonies than pseudotypes formed with proteins encoded by wild-type virus. When the MuLV-SF1 pseudotypes did produce tk+ cells, most of the proviruses were integrated aberrantly. We conclude that the MuLV-SF1 pol gene is defective for a function that is required for normal integrative recombination and dissociable from DNA synthesis.

Further evidence for the protein coding potential of the mouse mammary tumor virus long terminal repeat: nucleotide sequence of an endogenous proviral long terminal repeat

The 3' half of an endogenous mouse mammary tumor virus from a C3H mouse was cloned in the Charon 4A vector phage. A comparison of the proviral clone with previously published endogenous mouse mammary tumor virus restriction maps identified it as endogenous unit II (J. Cohen and H. Varmus, Nature [London] 278:418-423, 1979), which is present in all inbred mouse strains derived from the original Bagg albino x DBA cross. The nucleotide sequence of the unit II long terminal redundancy (LTR) was determined and compared with the sequence previously determined for the exogenous C3H virus LTR (Donehower et al., J. Virol. 37:226-238, 1981). Virtually all sequence differences between the two LTRs were base substitutions. The total amount of sequence divergence was 6.6%. The large open reading frame reported previously in the exogenous LTR was preserved in the endogenous LTR. In addition, the pattern of sequence divergence was highly nonrandom with respect to the putative amino acid codons of the two open reading frames. Most of the base substitutions in this region resulted in silent or conservative amino acid codon changes. The nonrandom divergence pattern indicates that selective forces are operating on this segment of DNA and argues that the putative protein is functional in the life cycle of mouse mammary tumor virus. Possible roles for the protein and its mode of expression are discussed.

Regulatory and coding potential of the mouse mammary tumor virus long terminal redundancy

Molecular clones containing the 3' half of newly integrated mouse mammary tumor virus (MMTV) DNA with adjacent mouse cellular sequences were characterized. In addition, we cloned the long terminal redundancy joint from the unintegrated circular form of MMTV DNA. The entire nucleotide sequence of the integrated and part of the unintegrated terminal redundancy was determined; this allowed us to delineate the boundaries of the MMTV long terminal redundancy, which comprises 1,327 base pairs. The position of possible RNA polymerase II initiation and termination signals corresponded closely to the expected regions of viral RNA initiation and termination specified by current models. The MMTV long terminal redundancy also contained a large open reading frame with sufficient information for a protein of 198 amino acids. Initial comparison of flanking 3' cellular sequences from three independent integrated clones suggested there was no host sequence specificity in the MMTV integration event. However, specificity of integration with respect to viral sequences was precise.