Origin of the minor glycoproteins of murine leukemia viruses.

Polyacrylamide gel electrophoretic analysis and immunoprecipitation were used to study glycoproteins from purified Rauscher murine leukemia virus (R-MuLV) and from AKR thymic lymphoblastoid cell membranes. In addition to gp70, a minor glycoprotein of approximately 52,000 daltons (gp52) was demonstrated in purified R-MuLV preparations, which was antigenically related to gp70. Analysis of R-MuLV glycopeptides obtained after exhaustive Pronase digestion showed that gp70 has at least two different glycopeptide size classes with molecular weights of 5,100 and 2,900, respectively. gp52, however, contained only a single glycopeptide size class of approximately 5,100 daltons, indicating that the two glycoproteins contain distinct carbohydrate components. Trypsin treatment of R-MuLV converted gp70 into a product with a molecular mass of approximately 52,000 daltons as well as a 45,000-dalton minor product, with little effect on virus infectivity. Similarly, trypsin treatment of 125I-labeled glycoproteins derived from AKR mouse lymphoblastoid cell membranes generated fragments antigenically related to gp70 and similar in size to those obtained by trypsin treatment of R-MuLV. In both cases, the appearance of cleavage products was accompanied by a decrease in gp70 during trypsin treatment. The occurrence of glycosylated components antigenically related to gp70 in AKR membrane glycoprotein preparations and in purified R-MuLV preparations which were similar to those generated by trypsin treatment supports the concept that these minor components arise from proteolytic cleavage of gp70.     

Glycosylation and intracellular transport of membrane glycoproteins encoded by murine leukemia viruses. Inhibition by amino acid analogues and by tunicamycin

Addition of asparagine-linked oligosaccharides to nascent murine leukemia virus (MuLV)-encoded membrane glycoproteins was inhibited either completely by tunicamycin or specifically at Asn-X-Thr glycosylation sites by incorporation of the threonine analogue beta-hydroxynorvaline. In conditions of partial analogue substitution, a series of subglycosylated components is formed which are related by a constant apparent Mr difference when assayed by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. The total number of asparagine-linked oligosaccharides is then estimated by dividing the measured apparent Mr of one oligosaccharide into the total apparent Mr difference between the complete glycoprotein and the polypeptide chain that is synthesized in cells incubated with tunicamycin. Correct results were obtained using glycoproteins with known numbers of oligosaccharides. Our analyses indicate that the gp70 membrane envelope glycoproteins of certain ecotropic MuLVs contain seven oligosaccharides, whereas the GIX+ antigen-containing variant gp70 contains one fewer Asn-X-Thr-linked oligosaccharide. The membrane glycoprotein encoded by the gag gene of Friend MuLV contains only one asparagine-linked oligosaccharide. Similarly, the gp55 membrane glycoprotein encoded by Friend erythroleukemia virus contains four asparagine-linked oligosaccharides. Pulse-chase and cell surface iodination analyses indicate that MuLV membrane envelope glycoprotein processing by partial proteolysis and transport to the cell surface can be efficiently blocked by structural perturbations caused by incorporation of different amino acid analogues or by loss of oligosaccharides. Our data also suggest that loss of oligosaccharides may expose new antigenic sites in viral membrane glycoproteins and increase their susceptibility to intracellular proteolysis.     

Tunicamycin inhibits glycosylation of precursor polyprotein encoded by env gene of Rauscher murine leukemia virus.

The molecular weight of the precursor polyprotein to the envelope proteins of Rauscher murine leukemia virus is reduced from 85,000 to 68,000 daltons when synthesized in the presence of tunicamycin, a specific inhibitor of the synthesis of oligosaccharides that attach to glycoproteins via asparagine residues. The unglycosylated precursor protein (Pr68^$\text{env}$) is synthesized at a rate comparable to that of the normal carbohydrate-containing envelope precursor (gPr85^$\text{env}$). Pr68^$\text{env}$ is not proteolytically processed and remains undegraded in the cell. Thus, most if not all of the carbohydrate content of gPr85^$\text{env}$ is N-linked, and glycosylation appears to be necessary for normal processing of precursor proteins into viral particles.     

Receptor-binding domain of murine leukemia virus envelope glycoproteins.

The surface glycoprotein (SU) of murine leukemia viruses (MuLVs) comprises two domains connected by a proline-rich hinge. The interaction of MuLV particles with subgroup-specific cell surface receptors depends primarily on two variable regions (VRA and VRB) located in the amino-terminal domain. To delineate the minimal receptor-binding domains, we examined the capacity of soluble envelope fragments to compete with the entry of virus particles. Amphotropic, ecotropic, polytropic, and xenotropic truncated SUs were produced by inserting stop codons in the env gene of the 4070A, Friend, MCF247 and NZB MuLVs, respectively. These fragments, as well as full-length envelope glycoproteins, were stably expressed in cells bearing the corresponding receptor. Synthesis, posttranslational modifications, transport, and secretion of the env gene products were monitored by immunoprecipitation. Cells expressing the modified SUs or naive cells preincubated with SU-containing conditioned media were infected with different pseudotypes of a retroviral vector carrying a beta-galactosidase marker gene. Reduction of cell susceptibility to infection in the presence of SU was used as a measure of receptor occupancy. The results indicated that the amphotropic and ecotropic envelope amino-terminal domains contain all of the determinants required for receptor binding. In contrast, additional sequences in the proline-rich region were needed for efficient interaction of the polytropic and xenotropic amino-terminal domains with the receptors.     

Receptor choice determinants in the envelope glycoproteins of amphotropic, xenotropic, and polytropic murine leukemia viruses.

The envelope glycoproteins (SU) of mammalian type C retroviruses possess an amino-terminal domain of about 200 residues, which is involved in binding a cell surface receptor. In this domain, highly conserved amino acid sequences are interrupted by two segments of variable length and sequence, VRA and VRB. We have studied the role of these variable regions in receptor recognition and binding by constructing chimeric molecules in which portions of the amino-terminal domains from amphotropic (4070A), xenotropic (NZB), and polytropic (MCF 247) murine leukemia virus SU proteins were permuted. These chimeras, which exchanged either one or two variable regions, were expressed at the surface of replication-defective viral particles by a pseudotyping assay. Wild-type or recombinant env genes were transfected into a cell line producing Moloney murine leukemia virus particles devoid of envelope glycoproteins in which a retrovirus vector genome carrying an Escherichia coli lacZ gene was packaged. The host range and sensitivity to interference of pseudotyped virions were assayed, and we observed which permutations resulted in receptor switch or loss of function. Our results indicate that the determinants of receptor choice are found within the just 120 amino acids of SU proteins. Downstream sequences contribute to the stabilization of the receptor-specific structure.     

A nonstructural gag-encoded glycoprotein precursor is necessary for efficient spreading and pathogenesis of murine leukemia viruses.

In addition to the Gag-Pol and Env precursors whose translation initiates at AUG codons, murine, feline, and simian type C oncoviruses also express glycosylated Gag-Pol precursors (glycoGag), glycoGag translation is initiated at CUG codons located upstream of the Gag AUG initiation codon. In contrast to Gag, glycoGag is translocated into the endoplasmic reticulum and is absent from virions. Since glycoGag has been described to be dispensable ex vivo, we investigated the in vivo effects of a glycoGag- mutation in the Friend murine leukemia virus (F-MuLV). F-MuLV induces severe early hemolytic anemia and subsequent erythroleukemia within 2 months after inoculation of newborn mice. We obtained a glycoGag- F-MuLV, strain H5, by inserting an octanucleotide linker downstream of the CUG codon leading to the reading of a stop codon in all reading frames upstream of the Gag AUG. F-MuLV H5 did not induce severe early hemolytic anemia, and latency of erythroleukemia was significantly increased most likely because of an approximately 1-week delay in the in vivo spreading. Accordingly, induction of recombinant polytropic viruses was also significantly delayed. Close examination of ex vivo spreading kinetics also showed a slower dissemination of F-MuLV H5. Western blot (immunoblot) performed after inoculation of newborn mice with this glycoGag- virus indicated the emergence of new glycoGag+ viruses. PCR analyses with F-MuLV-specific primers demonstrated in vivo pseudoreversions restoring the glycoGag reading frame. Our results demonstrated that glycoGag expression is positively selected and essential for full spreading and pathogenic abilities.     

Splice acceptor site for the env message of Moloney murine leukemia virus.

We report the isolation and sequence of a cDNA clone containing part of the env message of the Moloney murine leukemia virus (MoMuLV). This clone was derived from a rat thymic lymphoma induced by MoMuLV. The AG acceptor site employed in this message is located at position 5490 in the MoMuLV genome. This splice site is detectable at the cDNA level by the creation of a novel SacI restriction site not present in the viral genome. In the -1 to -40 region, this AG acceptor site is preceded by four conserved heptanucleotides (PyXPyTPuAPy) that may function as acceptors for removal of the 5' end of the intron.     

Structure of the murine leukemia virus envelope glycoprotein precursor.

The glycosylated env gene precurosr (Pr80env) of Moloney murine leukemia virus has been isolated by selective immunoprecipitation. Use of the drug tunicamycin to inhibit nascent glycosylation or specific cleavage with endoglycosidase H demonstrated that the precursor contained an apoprotein with a molecular weight of 60,000. The finished virion glycoprotein (gp70) was largely resistant to the action of endoglycosidase H. Chromatography of the glycopeptides of Pr80env in conjunction with endoglycosidase H digestion studies suggested that the precursor contained two distinct major glycosylation sites. Analysis of partial proteolytic cleavage fragments of Pr80env before and after endoglycosidase H treatment placed the two glycosylation sites within a 30,000-dalton region of the apoprotein sequence. Kinetic experiments showed that carbohydrate processing as well as proteolytic cleavage are late steps in the maturation of Pr80env.     

Genomes of murine leukemia viruses isolated from wild mice.

The genomes of murine leukemia viruses (MuLV) isolated from wild mice have been studied. Detailed restriction endonuclease maps of the 8.8-kilobase (kb) unintegrated linear viral DNAs were derived for five ecotropic and five amphotropic MuLV's from California field mice, for Friend MuLV, and for one ecotropic and one xenotropic MuLV from Mus musculus castaneus. In general, the California MuLV's were similar in their leftward 6 kb (corresponding to the leftward long terminal repeat [LTR], gag, and pol) and rightward 1 kb (7.8 to 8.8 kb, corresponding to p15E and the rightward LTR). For the region spanning 6.0 to 7.7 kb (which includes the sequences that encode gp70) the amphotropic MuLV's shared few enzyme sites with the ecotropic MuLV's, although the California ecotropic MuLV's were highly related to each other in this region, as were the amphotropic MuLV's. Cross-hybridization studies between amphotropic and California ecotropic MuLV DNAs indicated that they were not homologous in the region 6.3 to 7.6 kb; the California ecotropic viral DNAs cross-hybridized in this region to AKR ecotropic MuLV. When the California viral DNAs were compared with AKR ecotropic viral DNA, many differences in enzyme sites were noted throughout the genome. The U3 regions of the wild mouse LTRs showed partial homology to this region in AKR MuLV. The LTR of Moloney MuLV was highly related to that of the California MuLV's, whereas the LTR of Friend MuLV appeared to be a recombinant between the two types of LTRs. The M. musculus castaneus isolates were most closely related to ecotropic and xenotropic MuLV's isolated from inbred mice. One amphotropic MuLV DNA was cloned from supercoiled viral DNA at its unique EcoRI site in pBR322. Viral DNAs with one and two LTRs were isolated. After digestion with EcoRI, DNAs of both types were infectious. It is concluded that ecotropic and amphotropic MuLV's differ primarily in the region which encodes gp70.     

Role of intron-contained sequences in formation of moloney murine leukemia virus env mRNA.

Formation of the Moloney murine leukemia virus envelope mRNA involves the removal of a 5,185-base pair-long intron. Deletion analysis of two Moloney murine leukemia virus-derived expression vectors revealed the existence of two short regions within the viral intron which are required for the efficient formation of the spliced RNA species. One region was present upstream from the 3' splice junction, extended at least 85 nucleotides beyond the splice site, and was not more than 165 nucleotides long. As yeast polymerase II introns, the Moloney murine leukemia virus intron contains the sequence 5'-TACTAAC-3' 15 nucleotides upstream from the 3' splice site. A second region located in the middle of the intron, within a 560-nucleotide-long sequence, was also essential for formation of the spliced RNA species. The efficient splicing of the env mRNA in the absence of expression of viral genes raises the possibility that similar mechanisms are used to remove introns of (some) cellular genes.     

A neutralizable epitope common to the envelope glycoproteins of ecotropic, polytropic, xenotropic, and amphotropic murine leukemia viruses.

An epitope common to all classes of murine leukemia viruses (MuLVs) was detected by reactivity of MuLVs with a rat monoclonal antibody (MAb) termed 83A25. The antibody is of the immunoglobulin G2a isotype and was derived after fusion of NS-1 myeloma cells with spleen cells from a Fischer rat immunized with a Friend polytropic MuLV. The antibody reacted with nearly all members of the ecotropic, polytropic, xenotropic, and amphotropic classes of MuLVs. Unreactive viruses were limited to the Friend ecotropic MuLV, Rauscher MuLV, and certain recombinant derivatives of Friend ecotropic MuLV. The presence of an epitope common to nearly all MuLVs facilitated a direct quantitative focal immunofluorescence assay for MuLVs, including the amphotropic MuLVs for which no direct assay has been previously available. Previously described MAbs which react with all classes of MuLVs have been limited to those which react with virion core or transmembrane proteins. In contrast, protein immunoblot and immunoprecipitation analyses established that the epitope reactive with MAb 83A25 resides in the envelope glycoproteins of the viruses. Structural comparisons of reactive and nonreactive Friend polytropic viruses localized the epitope near the carboxyl terminus of the glycoprotein. The epitope served as a target for neutralization of all classes of MuLV with MAb 83A25. The efficiency of neutralization varied with different MuLV isolates but did not correlate with MuLV interference groups.     

Characterization of intracellular precursor polyproteins of Moloney murine leukemia virus.

Intracellular Moloney murine leukemia viral precursor polyproteins were compared with mature viral proteins by immunoprecipitation and tryptic peptide mapping experiments. The results were consistent with precursor roles for Pr65gag, Pr200gag-pol, Pr135pol, and gPr83env. The glycosylated gag gene product gPr85gag, although containing sequences characteristic of all four core proteins plus additional sequences not found in Pr65gag, lacked a major tyrosine-containing p30 tryptic peptide, suggesting that gPr85gag is not processed to p30.     

Biochemical characterization of the amphotropic group of murine leukemia viruses.

The recently described amphotropic group of murine leukemia viruses constitutes a distinct biological group, differing from the ecotropic and xenotropic groups in host range, cross interference, and serological reactivity. Viruses of this group have been detected only in wild mice from certain areas in California. By using a [3H]DNA probe synthesized in an endogenous reaction from detergent-lysed amphotropic virus (strain 1504-A), it was demonstrated that the amphotropic murine leukemia viruses are distinct biochemically, in that 20% of the viral genome sequences are not shared by AKR-type ecotropic or nay of three types of xenotropic murine leukemia virus tested. A subset of these amphotropic unique sequences, comprising one half of them, is present in the genome of wild mouse ecotropic viruses and in Moloney and Rauscher viruses as well. Sequences homologous to the entire genome of 1504-A amphotropic virus are present in the cellular DNA of all eight inbred mouse strains tested, as well as in wild Mus in Asia, in amounts varying from three to six complete viral genomes per haploid cell genome. Evidence is presented that at least 20% of the DNA sequences in both mouse- and mink-grown murine leukemia virus probes are of host-cell origin.     

Titration of murine leukemia viruses with rat cell line RFL.

Normal rat embryo cell (RFL) from syncytia after infection with murine leukemia virus. The assay for counting the number of syncytium foci produced in RFL cells is a sensitive method for a direct infectivity assay of murine leukemia virus.