Purification of DNA complementary to the env gene of avian sarcoma virus and analysis of relationships among the env genes of avian leukosis-sarcoma viruses

Mechanisms of avian retroviral host range extension

Alpharetroviruses provide a useful system for the study of the molecular mechanisms of host range and receptor interaction. These viruses can be divided into subgroups based on diverse receptor usage due to variability within the two host range determining regions, hr1 and hr2, in their envelope glycoprotein SU (gp85). In previous work, our laboratory described selection from a subgroup B avian sarcoma-leukosis virus of an extended-host-range variant (LT/SI) with two adjacent amino acid substitutions in hr1. This virus retains its ability to use the subgroup BD receptor but can also infect QT6/BD cells, which bear a related subgroup E receptor (R. A. Taplitz and J. M. Coffin, J. Virol 71:7814-7819, 1997). Here, we report further analysis of this unusual variant. First, one (L154S) of the two substitutions is sufficient for host range extension, while the other (T155I) does not alter host range. Second, these mutations extend host range to non-avian cell types, including human, dog, cat, mouse, rat, and hamster. Third, interference experiments imply that the mutants interact efficiently with the subgroup BD receptor and possibly the related subgroup E receptor, but they have another means of entry that is not dependent on these interactions. Fourth, binding studies indicate that the mutant SU proteins retain the ability to interact as monomers with subgroup BD and BDE receptors but only bind the subgroup E receptor in the context of an Env trimer. Further, the mutant SU proteins bind well to chicken cells but do not bind any better than wild-type subgroup B to QT6 or human cells, even though the corresponding viruses are capable of infecting these cells.

Genetic determinants of neoplastic diseases induced by a subgroup F avian leukosis virus

Two subgroup F avian leukosis viruses, ring-necked pheasant virus (RPV) and RAV-61, were previously shown to induce a high incidence of a fatal proliferative disorder in the lungs of infected chickens. These lung lesions, termed angiosarcomas, appear rapidly (4 to 5 weeks after infection), show no evidence of proto-oncogene activation by proviral integration, and are not induced by avian leukosis viruses belonging to other subgroups. To identify the viral sequences responsible for induction of these tumors, we constructed recombinant viruses by exchanging genomic segments of molecularly cloned RPV with those of a subgroup A leukosis virus, UR2AV. The ability to induce rapid lung tumors segregated only with the env sequences of RPV; the long terminal repeat of RPV was not required. However, recombinants carrying both env and long terminal repeat sequences of RPV induced lung tumors with a shorter latency. In several cases, recombinant viruses exhibited pathogenic properties differing from those of either parental virus. Recombinants carrying the gag-pol region of RPV and the env gene of UR2AV induced a high incidence of a muscle lesion termed infiltrative intramuscular fibromatosis. One recombinant, EU-8, which carries the gag-pol and LTR sequences of RPV, and the env gene of UR2AV, induced lymphoid leukosis after an unusually short latent period. The median time of death from lymphoid leukosis was 6 to 7 weeks after infection with EU-8 compared with approximately 5 months for UR2AV.

Subgenomic mRNA in OK10 defective leukemia virus-transformed cells

OK10, a defective leukemia virus, is produced as a defective particle by so-called nonproducer transformed quail fibroblasts. OK10 defective viral particles contain an 8-kilobases (kb)-long genomic RNA, lack any detectable reverse transcriptase activity, and are not infectious. We studied the genetic content of OK10 RNA extracted from both virions and infected cells. As shown by RNA-cDNA hybridizations in stringent conditions, about 77% (6.4 kb) of the OK10 8.0kb RNA was related to avian leukosis viruses in the three structural genes gag, pol, and env, as well as in the c region. The remainder of the OK10 genome-encoding capacity (</=1.6 kb) was homologous to the MC29-specific transforming sequence myc(m) and therefore has been named myc(o). EcoRI restriction analysis of the OK10 integrated proviral DNA with different probes indicated the presence of only one provirus in the OK10 QB5 clone, which agreed with the gene order: 5'-gag-Deltapol-myc(o)-Deltaenv-c- 3'. Heteroduplex molecules formed between the viral OK10 8.0-kb RNA and the 6.8-kb SacI DNA fragment of the Prague A strain of Rous sarcoma virus confirmed that structure and indicated that the myc(o) sequence formed a continuous RNA stretch of 1.4 to 1.6 kb long between Deltapol and Deltaenv. We also examined the myc(o)-containing mRNA's transcribed in OK10-transformed cells. OK10-transformed quail fibroblasts (OK10 QB5) transcribed two mRNA species of 8.0 and 3.6 kb containing the myc(o) sequence. The genetic content of the 3.6-kb species made it a possible maturation product of the genome size 8-kb species by splicing out the gag and pol sequences. In OK10-transformed bone marrow cells (OK10 BM), a stable bone marrow-derived cell line producing OK10, the myc(o) sequence was found in four RNA species of 11.0, 8.0, 7.0, and 3.6 kb. Again, the genetic content of these mRNA's indicated that (i) the 3.6-kb species could be spliced out of the 8.0-kb-genome size mRNA and (ii) the 11.0-kb-long mRNA could represent a read-through of the OK10 provirus, the corresponding maturation product being, then, a 7.0-kb mRNA. The 7.0- and 3.6- kb mRNA's both contained the myc(o) sequence, but no sequences related to the gag or pol gene. In conclusion, whereas the myc sequences have been generally thought to be expressed through a gag-onc fusion protein, as for MC29 and CMII viruses, our experiments indicate that they could also be expressed as a non-gag-related product made from a subgenomic mRNA in the OK10-transformed cells.

Characterization of the oncogene (erb) of avian erythroblastosis virus and its cellular progenitor

Avian erythroblastosis virus (AEV) induces primarily erythroblastosis when injected intravenously into susceptible chickens. In vitro, the hematopoietic target cells for transformation are the erythroblasts. Occasional sarcomas are also induced by intramuscular injection, and chicken or quail fibroblasts can be transformed in vitro. The transforming capacity of AEV was shown to be associated with the presence of a unique nucleotide sequence denoted erb in its genomic RNA. Using a simplified procedure, we prepared radioactive complementary DNA (cDNAaev) representative of the erb sequence at a high yield. Using a cDNAaev excess liquid hybridization technique adapted to defective retroviruses, we determined the complexity of the erb sequence to be 3,700 +/- 370 nucleotides. AEV-transformed erythroblasts, as well as fibroblasts, contained two polyadenylated viral mRNA species of 30 and 23S in similar high abundance (50 to 500 copies per cell). Both species were efficiently packaged into the virions. AEV-transformed erythroblasts contained additional high-molecular-weight mRNA species hybridizing with cDNAaev and cDNA5' but not with cDNA made to the helper leukosis virus used (cDNArep). The nature and the role, if any, of these bands remain unclear. The erb sequence had its counterpart in normal cellular DNA of all higher vertebrate species tested, including humans and fish (1 to 2 copies per haploid genome in the nonrepetitive fraction of the DNA). These cellular sequences (c-erb) were transcribed at low levels (1 to 2 RNA copies per cell) in chicken and quail fibroblasts, in which the two alleged domains of AEV-specific sequences corresponding to the 75,000- and 40,000-molecular-weight proteins seemed to be conserved phylogenetically and transcribed at similar low rates.

Evidence for the common origin of viral and cellular sequences involved in sarcomagenic transformation

The src genes of six different strains of avian sarcoma virus (ASV) were compared with those of a series of newly isolated sarcoma viruses, termed "recovery avian sarcoma viruses" (rASV's). The rASV's were isolated recently from chicken and quail tumors induced by transformation-defective (td) deletion mutants of Schmidt-Ruppin Rous sarcoma virus. The RNase T1-resistant oligonucleotide maps were constructed for the RNA genomes of different strains of ASV and td mutants. The src-specific sequences, characterized by RNase T1-resistant oligonucleotides ranging from 9 to 19 nucleotides long, were defined as those mapping between approximately 600 and 2,800 nucleotides from the 3' polyadenylate end of individual sarcoma viral RNAs, and missing in the corresponding td viral RNAs. Our results revealed that 12 src-specific oligonucleotides were highly conserved among several strains of ASV, including the rASV's, whereas certain strains of ASV were found to contain one to three characteristic src-specific oligonucleotides. We previously presented evidence supporting the idea that most of the src-specific sequences present in rASV RNAs are derived from cellular genetic information. Our present data indicate that the src genes of rASV's are closely related to other known ASVs. We conclude that the src genes of different strains of ASV and the cellular sarc sequences are of common origin, although some divergence has occurred among different viral src genes and related cellular sequences.

Avian retroviruses that cause carcinoma and leukemia: identification of nucleotide sequences associated with pathogenicity

Avian myelocytomatosis virus (MC29V) is a retrovirus that transforms both fibroblasts and macrophages in culture and induces myelocytomatosis, carcinomas, and sarcomas in birds. Previous work identified a sequence of about 1,500 nucleotides (here denoted onc(MCV)) that apparently derived from a normal cellular sequence and that may encode the oncogenic capacity of MC29V. In an effort to further implicate onc(MCV) in tumorigenesis, we used molecular hybridization to examine the distribution of nucleotide sequences related to onc(MCV) among the genomes of various avian retroviruses. In addition, we characterized further the genetic composition of the remainder of the MC29V genome. Our work exploited the availability of radioactive DNAs (cDNA's) complementary to onc(MCV) (cDNA(MCV)) or to specific portions of the genome of avian sarcoma virus (ASV). We showed that genomic RNAs of avian erythroblastosis virus (AEV) and avian myeloblastosis virus (AMV) could not hybridize appreciably with cDNA(MCV). By contrast, cDNA(MCV) hybridized extensively (about 75%) and with essentially complete fidelity to the genome of Mill Hill 2 virus (MH2V), whose pathogenicity is very similar to that of MC29V, but different from that of AEV or AMV. Hybridization with the ASV cDNA's demonstrated that the MC29V genome includes about half of the ASV envelope protein gene and that the remainder of the MC29V genome is closely related to nucleotide sequences that are shared among the genomes of many avian leukosis and sarcoma viruses. We conclude that onc(MCV) probably specifies the unique set of pathogenicities displayed by MC29V and MH2V, whereas the oncogenic potentials of AEV and AMV are presumably encoded by a distinct nucleotide sequence unrelated to onc(MCV). The genomes of ASV, MC29V, and other avian oncoviruses thus share a set of common sequences, but apparently owe their various oncogenic potentials to unrelated transforming genes.

Integration of bovine leukemia virus DNA in the bovine genome

DNA preparations from circulating leukocytes, lymph node tumors, and spleens of three bovine leukemia virus-infected cattle were fractionated by Cs2SO4/3,6-bis(acetatomercurimethyl)dioxane density gradient centrifugation. Bovine leukemia virus proviral sequences were found in large GC-rich fragments having a buoyant density in CsCl close to 1.708 g/cm3. Provirus integration, therefore, does not take place at random locations in the host genome, but in a specific class of DNA segments. Hybridization of cDNA synthesized on viral RNA to EcoRI and Xba I restriction fragments of the DNA from infected cells showed that: (i) only one copy of proviral DNA is integrated per haploid genome; (ii) different restriction patterns were found in the proviral DNAs present in the genomes of different animals, providing evidence for the existence of several strains or mutants; and (iii) different integration sites for the proviral DNA were found in the genome of different animals and of different infected cells in the same animal. The latter finding strongly suggests a polyclonal origin of bovine leukemia virus-infected cells.

Heterogeneity of genetic loci in chickens: analysis of endogenous viral and nonviral genes by cleavage of DNA with restriction endonucleases

Restriction endonucleases can be used to define the structure and position of genetic loci for which specific molecular hybridization reagents are available. We have used this approach to compare 18 chicken embryos with respect to several cellular genes; endogenous viral DNA related to the replicative genes of avian sarcoma virus (ASV) or to RAV-O, an endogenous virus of chickens; and sequences related to the transforming (src) gene of ASV. Each cellular gene eas remarkably homogeneous within our test population. We found little or no variation in globin and ovomucoid genes; ovalbumin and transferrin (with one exception) showed variation which is probably allelic in nature. The endogenous viral DNA which has homology with RAV-O was found at several different positions in host DNA and its structure resembled that of proviruses acquired by experimental infection, with sequences from both ends of viral RNA repeated near both ends of viral DNA. Within the population of 18 chickens, one endogenous provirus was always present, whereas the several other proviruses were each found in only a few members of this group. However, screening of additional chickens identified individuals lacking the provirus common to the initial 18 animals surveyed; in at least one embryo no RAV-O-related DNA was detected. These findings suggest that the endogenous RAV-O-related sequences have entered the germ line by relatively recent infection and are still segregating in several contemporary chicken flocks. The sequences in the chicken genome which have homology with the src gene of ASV are invariant from bird to bird and in this sense resemble a cellular gene rather than a viral sequence.

Genome of avian myelocytomatosis virus MC29: analysis by heteroduplex mapping

The virion RNA of avian myelocytoma virus MC29 was hybridized to full genome length DNA of the Prague strain of Rous sarcoma virus and analyzed by heteroduplex mapping in the electron microscopy. The results show that MC29 specific sequences for which there are no homologous counterparts in the Rous sarcoma virus genome make up a contiguous stretch of RNA about 1.8 kilobases long. These sequences are located approximately in the middle of the genome, replacing the 3' half of the gag gene, the entire pol gene, and the 5' portion of the env gene, which are absent from MC29. This MC29 specific genetic substitution may contain information for the leukemogenic transformation of the host cell.

Mapping unintegrated avian sarcoma virus DNA: termini of linear DNA bear 300 nucleotides present once or twice in two species of circular DNA

Three major species of viral DNA have been observed in cells infected by retroviruses: a linear, double-stranded copy of a subunit of viral RNA; closed circular DNA; and proviral DNA inserted covalently into the genome of the host cell. We have studied the structures of the unintegrated forms of avian sarcoma virus (ASA) DNA using agarose gel electrophoresis in conjunction with restriction endonucleases and molecular hybridization techniques. The linear duplex DNA is approximately the same length as a subunit of viral RNA (approximately 10 kb) and it bears natural repeats of approximately 300 nucleotides at its termini. The repeats are composed of sequences derived from both the 3' and 5' termini of viral RNA in a manner suggesting that the viral DNA polymerase is transferred twice between templates. Thus the first end begins with a sequence from the 5' terminus of viral RNA and is permuted by about 100 nucleotides with respect to the 3' terminus of viral RNA; the linear DNA terminates with a sequence of about 200 nucleotides derived from the 3' end of viral RNA. We represent this structure, synthesized from right to left, as 3'5'-----3'5'. Two closed circular species of approximately monomeric size have been identified. The less abundant species contain all the sequences identified in linear DNA, including two copies in tandem of the 300 nucleotide 3'5' repeat. The major species lacks about 300 base pairs (bp) mapped to the region of the repeated sequence; thus it presumably contains only a single copy of that sequence. The strategies used to determine these structures involved the assignment of over 20 cleavage sites for restriction endonucleases on the physical maps of ASV DNA. Several strains of ASV were compared with respect to these sites, and the sites have been located in relation to deletions frequently observed in the env and src genes of ASV.

Nucleotide sequence relationships between the genomes of an endogenous and an exogenous avian tumor virus

We have used mapping of large T1 oligonucleotides to examine the genome of Rous-associated virus-O (RAV-O), an endogenous virus of chickens, and to compare it with that of Prague strain Rous sarcoma virus, subgroup B, (Pr-RSV-B), an exogenous sarcoma virus. To extend the sensitivity of such comparisons, we have developed a system of nucleic acid hybridization and hybridization-competition combined with fingerprinting. This method allows us to estimate the relative degree of relatedness of various portions of the viral genomes. From the results of this study, we have concluded that the genomes of Pr-RSV-B and RAV-O are related in the following way. The 5'-terminal half of the genomes (corresponding to the gag and pol regions) is virtually identical, with only scattered single nucleotide differences. This region is followed by a region comprising 25 to 30% of the genome (the env region) which contains substantial nucleotide sequence differences, most or all of which are due to single base changes. The env-coding region can be further subdivided into three regions: a more variable region probably containing sequences coding for subgroup specificity, flanked by relatively common sequences on each side. To the 3' side of the env region, the RAV-O genome contains a very short sequence not found in Pr-RSV-B, whereas the Pr-RSV-B genome contains a much longer unrelated sequence. The central portion of this sequence comprises the src gene as defined by transformation-defective mutants. Particularly striking is the absence, in the RAV-O genome, of any nucleotide sequence related to the "c region" found very near the 3' end of all exogenous tumor viruses. Both the Pr-RSV-B and RAV-O genomes contain the identical terminally redundant sequence of 21 nucleotides near each end of the genome.

Proviruses of avian sarcoma virus are terminally redundant, co-extensive with unintegrated linear DNA and integrated at many sites

We have analyzed the DNA from 15 clones of avian sarcoma virus (ASV)-transformed rat cells with restriction endonucleases and molecular hybridization techniques to determine the location and structure of proviral DNA. All twenty units of proviral DNA identified in these 15 clones appear to be inserted at different sites in host DNA. In each of the ten cases that could be sufficiently well mapped, entirely different regions of cellular DNA were involved. Thus ASV DNA can be accommodated at many positions in cellular DNA, but the existence of preferred sites has not been excluded. Six of the 15 clones carry only one normal provirus, two contain two normal proviruses, and seven harbor either one or two proviruses that appear anomalous in physical mapping tests. Both ends of at least 18 proviruses, however, were found to contain sequences specific to both the 3' and 5' termini of viral RNA. The organization of these terminally redundant sequences appeared identical to that of the 300 base pair (bp) repeats found at the ends of unintegrated linear DNA (Shank et al., 1978). Proviral DNA is therefore co-extensive, or nearly co-extensive, with unintegrated linear DNA and has a structure we denote as CELL DNA-3'5'----------3'5'-CELL DNA. Three of the four anomalous proviruses which were fully analyzed were deletion mutants lacking 25--65% of the genetic content of ASV; the fourth provirus had a novel site for cleavage by Eco RI but was otherwise normal. Tests for the biological competence of proviral DNA, based upon rescue of transforming virus after fusion with chicken cells, were generally consistent with the physical mapping studies.

Characterization of the env gene in avian oncoviruses by heteroduplex mapping

The genome of ring-necked pheasant virus, an avian oncovirus, is largely homologous to the genomes of chicken oncoviruses except for a specific nonhomology in env, the gene coding for the surface glycoprotein of the virion (J. Tal, D. J. Fujita, S. Kawai, H. E. Varmus, and J. M. Bishop, J. Virol. 21:497--505, 1977). We have used this nonhomology between ring-necked pheasant virus and chicken oncoviruses in electron microscopic studies of heteroduplex molecules. The env-specific region of nonhomology is 1.5 to 1.7 kilobases in length. Its 3' boundary is located 0.6 to 0.7 kilobases from the 3' end of the genome in transformation-defective viruses and 2.5 kilobases from the 3' end in nondefective avian sarcoma viruses. Comparison of several strains of avian oncoviruses shows that the 3' half of this env region is conserved, while the 5' half is more diverged. A small area at the very 3' end of env also shows divergence between different avian oncoviruses. We found no evidence for the presence of a previously unrecognized gene between env and src. An electrophoretic comparison of the glycoproteins from various avian oncoviruses shows that those of ring-necked pheasant virus and Chinese quail virus differ in molecular weight from the glycoproteins of the chicken oncoviruses.

env Gene of chicken RNA tumor viruses: extent of conservation in cellular and viral genomes

The env gene of avian sarcoma-leukosis viruses codes for envelope glycoproteins that determine viral host range, antigenic specificity, and interference patterns. We used molecular hybridization to analyze the natural distribution and possible origins of the nucleotide sequences that encode env; our work exploited the availability of radioactive DNA (cDNA(gp)) complementary to most or all of env. env sequences were detectable in the DNAs of chickens which synthesized an env gene product (chick helper factor positive) encoded by an endogenous viral gene and also in the DNAs of chickens which synthesized little or no env gene product (chick helper factor negative). env sequences were not detectable in DNAs from Japanese quail, ring-necked pheasant, golden pheasant, duck, squab, salmon sperm, or calf thymus. The detection of sequences closely related to viral env only in chicken DNA contrasts sharply with the demonstration that the transforming gene (src) of avian sarcoma viruses has readily detectable homologues in the DNAs of all avian species tested [D. Stehelin, H. E. Varmus, J. M. Bishop, and P. K. Vogt, Nature (London) 260: 170-173, 1976] and in the DNAs of other vertebrates (D. Spector, personal communication). Thermal denaturation studies on duplexes formed between cDNA(gp) and chicken DNA and also between cDNA(gp) and RNAs of subgroup A to E viruses derived from chickens indicated that these duplexes were well matched. In contrast, cDNA(gp) did not form stable hybrids with RNAs of viruses which were isolated from ring-necked and golden pheasants. We conclude that substantial portions of nucleotide sequences within the env genes of viruses of subgroups A to E are closely related and that these genes probably have a common, perhaps cellular, evolutionary origin.

Nucleotide sequences derived from pheasant DNA in the genome of recombinant avian leukosis viruses with subgroup F specificity

Recombination between viral and cellular genes can give rise to new strains of retroviruses. For example, Rous-associated virus 61 (RAV-61) is a recombinant between the Bryan high-titer strain of Rous sarcoma virus (RSV) and normal pheasant DNA. Nucleic acid hybridization techniques were used to study the genome of RAV-61 and another RAV with subgroup F specificity (RAV-F) obtained by passage of RSV-RAV-0 in cells from a ring-necked pheasant embryo. The nucleotide sequences acquired by these two independent isolates of RAV-F that were not shared with the parental virus comprised 20 to 25% of the RAV-F genomes and were indistinguishable by nucleic acid hybridization. (In addition, RAV-F genomes had another set of nucleotide sequences that were homologous to some pheasant nucleotide sequences and also were present in the parental viruses.) A specific complementary DNA, containing only nucleotide sequences complementary to those acquired by RAV-61 through recombination, was prepared. These nucleotide sequences were pheasant derived and were not present in the genomes of reticuloendotheliosis viruses, pheasant viruses, and avian leukosis-sarcoma viruses of subgroups A, B, C, D, and E. They were partially endogenous, however, to avian DNA other than pheasant. The fraction of these nucleotide sequences present in other avian DNAs generally paralleled the genetic relatedness of these avian species to pheasants. However, there was a high degree of homology between these pheasant nucleotide sequences and related nucleotide sequences in the DNA of normal chickens as indicated by the identical melting profiles of the respective hybrids.

Size and genetic content of viral RNAs in avian oncovirus-infected cells

Viral complementary DNA (cDNA) sequences corresponding to the gag, pol, env, src, and c regions of the Rous sarcoma virus genome were selected by hybridizing viral cDNA to RNA from viruses that lack the env or src gene or to polyadenylic acid [poly(A)]-containing RNA fragments of different lengths and isolating either hybridized or unhybridized DNA. The specificities, genetic complexities, and map locations of the selected cDNA's were shown to be in good agreement with the size and map locations of the corresponding viral genes. Analyses of virus-specific RNA, using the specific cDNA's as molecular probes, demonstrated that oncovirus-infected cells contained genome-length (30-40S) RNA plus either one or two species of subgenome-length viral RNA. The size and genetic content of these RNAs varied, depending on the genetic makeup of the infecting virus, but in each case the smaller RNAs contained only sequences located near the 3' end of the viral genome. Three RNA species were detected in Schmidt-Ruppin Rous sarcoma virus-infected cells: 39S (genome-length) RNA; 28S RNA, with an apparent sequence of env-src-c-poly(A); and 21S RNA, with an apparent sequence of src-c-poly(A). Cells infected with the Bryan high-titer strain of Rous sarcoma virus, which lacks the env gene, contained genome-length (35S) RNA and 21S src-specific RNA, but not the 28S RNA species. Leukosis virus-infected cells contained two detectable RNA species: 35S (genome-length) RNA and 21S RNA, with apparent sequence env-c-poly(A). Since gag and pol sequences were detected only in genome-length RNAs, it seems likely that the full-length transcripts function as mRNA for these two genes. The 28S and 21S RNAs could be the active messengers for the env and src genes. Analyses of sequence homologies among nucleic acids of different avian oncoviruses demonstrated substantial similarities within most of the genetic regions of these viruses. However, the "common" region of Rous-associated virus-0, an endogenous virus, was found to differ significantly from that of the other viruses tested.