Recombinants between endogenous and exogenous avian tumor viruses: role of the C region and other portions of the genome in the control of replication and transformation.

Endogenous retroviruses of chickens are closely related to exogenous viruses isolated from spontaneous tumors in the same species, yet differ in a number of important characteristics, including the ability to transform cells in culture, ability to cause sarcomas or leukemias, host range, and growth rate in cell culture. To correlate these differences with specific sequence differences between the two viral genomes, the genome RNA of transforming subgroup E recombinants between the Prague strain of Rous sarcoma virus, subgroup B (Pr-RSV-B), and the endogenous Rous-associated virus-0 (RAV-0), Subgroup E, and seven nontransforming subgroup E recombinants between the transformation-defective mutant of Pr-RSV-B and RAV-0 was examined by oligonucleotide fingerprinting. The pattern of inheritance among the recombinant viruses of regions of the genome in which Pr-RSV-B and RAV-0 differ allowed us to draw the following conclusions. (i) Nonselected parts of the genome were, with a few exceptions, inherited by the recombinant virus progeny randomly from either parent, with no obvious linkage between neighboring sequences. (ii) A small region in the Pr-RSV-B genome which maps in the 5' region was found in all transforming but only some of the nontransforming recombinants, suggesting that it plays a role in the control of the expression of transformation. (iii) A region of the Pr-RSV-B genome which maps between env and src was similarly linked to the src gene and may be either part of the structural gene for src or a control sequence regulating the expression of src. (iv) The C region at the extreme 3' end of the virus genome which is closely related in all the exogenous avian retroviruses but distinctly different in the endogenous viruses is the major determinant responsible for the differences in growth rate between RAV-0 and Pr-RSV-B. This latter observation allowed us to redefine the C region as a genetic locus, c, with two alleles cn (in RAV-0) and cx (in exogenous viruses).     

Molecular basis of host range variation in avian retroviruses.

Previous genetic analysis has localized the region of the Rous sarcoma virus (RSV) env gene responsible for host range specificity to that encoding the middle one-third of gp85. To better understand the host range determinants, the relevant regions of the genomes of infectious molecular clones of the transformation-defective Prague strain of RSV, subgroup B (Pr-RSV-B) and Rous-associated virus 0 (RAV-0) (subgroup E) were sequenced and compared with the sequence of Pr-RSV-C. This comparative analysis identified two variable regions of low amino acid sequence homology flanked by highly conserved amino acid sequences. The first variable region (hr1) begins at base 5654 in the Pr-RSV-C sequence and encodes 32 amino acids. The second variable region (hr2) begins at base 5846 and encodes 27 amino acids. To test the role of the variable regions in host range specificity, we determined the sequence of this region of the env gene of NTRE-4, a recombinant virus between Pr-RSV-B and RAV-0 which exhibits an extended host range. This analysis revealed that the recombinant subgroup-encoding region of NTRE-4 is composed of 200 bases of RAV-0 sequence, including hr2, flanked by sequences which are otherwise of Pr-RSV-B origin. This study indicates that hr1 and hr2 are the domains of gp85 responsible for host range determination in avian retroviruses.     

Nucleotide sequence of the 3' end of MCF 247 murine leukemia virus

We isolated DNA clones of MCF 247, a leukemogenic, recombinant type C virus obtained from the thymus of an AKR mouse. We determined the nucleotide sequence of the viral long terminal repeat (LTR) and the 3' end of env, and we compared the sequences to corresponding sequences of the genome of Akv virus, the putative ecotropic parent of MCF 247. By analogy with Moloney leukemia virus, we identified the amino terminus of Prp15E, the C-terminal proteolytic cleavage product of env and precursor to mature virion p15E. In MCF 247 the presumptive Prp15E is encoded by a 603-nucleotide open reading frame. The majority of this sequence is identical to that of Akv. However, a recombination event near the 3' end of the Prp15E-coding region introduces nonecotropic sequences into MCF 247, and these extend to the 3' end through the U3 portion of the LTR. The U3 regions of Akv and MCF 247 are about 83% homologous. The R and U5 regions of the LTR of MCF 247 and Akv are identical. Large RNase T1-resistant oligonucleotides analyzed previously in numerous ecotropic and MCF viral genomes were located within the Akv and MCF 247 DNA sequences. The resulting precise T1 oligonucleotide maps of the 3' ends of MCF viral genomes reveal that the biologically defined, leukemogenic class I MCFs isolated from thymic neoplasms of inbred mice all share the sequence pattern seen in MCF 247, a representative of this group; they possess recombinant Prp15E genes and derive U3 from their nonecotropic parents.     

Organization of shared and unshared sequences in the genomes of chicken endogenous and sarcoma viruses.

The genome of the genetically transmitted endogenous C type virus of chickens, RAV-O, is closely related to that of Rous sarcoma virus (RSV). Nevertheless, these viruses differ widely in oncogenicity and regulation by the host cell. Competitive hybridization analysis of ¹²⁵I-labeled genomic RNA demonstrated that the genome of RAV-O lacks about 35% of the sequences of nondefective RSV which formed hybrids with proviral DNA from RSV-infected cells, and that the genome of transformation-defective deletion mutants of RSV (td RSV) lacks about 15% of these sequences. Conversely, about 12% of the RAV-O sequences forming hybrids with normal chicken cell DNA were not detected in the sarcoma virus. A technique was developed to map the location of these unshared sequences by competitive hybridization. The deletion in the genome of td RSV was seen to begin at about 0.2 and to end at about 0.05 of the genome length from the 3′ end of sarcoma virus RNA, confirming the results of other laboratories using the method of mapping RNAase TI resistance of oligonucleotides. The 35% of RSV sequences missing and/or diverged in the genome of RAV-O were concentrated within 40% of the sarcoma virus genome from the 3′ end, and most of this large section did not appear to form hybrids with chicken DNA under the conditions of these experiments. A low level of hybrid formation was, however, detected between uninfected chicken cellular DNA and a small fraction of the nucleotides in the region of the td deletion. Analysis of RAV-O 3′ end fragments demonstrated that the genomic sequences of RAV-O missing in RSV were concentrated at the 3′ end of the endogenous viral genome. We conclude that the sequence differences between endogenous and sarcoma viruses are largely concentrated in specific regions of the viral genome.     

Molecular cloning of Snyder-Theilen feline leukemia and sarcoma viruses: comparative studies of feline sarcoma virus with its natural helper virus and with Moloney murine sarcoma virus

Extrachromosomal DNA obtained from mink cells acutely infected with the Snyder-Theilen (ST) strain of feline sarcoma virus (feline leukemia virus) [FeSV(FeLV)] was fractionated electrophoretically, and samples enriched for FeLV and FeSV linear intermediates were digested with EcoRI and cloned in lambda phage. Hybrid phages were isolated containing either FeSV or FeLV DNA "inserts" and were characterized by restriction enzyme analysis, R-looping with purified 26 to 32S viral RNA, and heteroduplex formation. The recombinant phages (designated lambda FeSV and lambda FeLV) contain all of the genetic information represented in FeSV and FeLV RNA genomes but lack one extended terminally redundant sequence of 750 bases which appears once at each end of parental linear DNA intermediates. Restriction enzyme and heteroduplex analyses confirmed that sequences unique to FeSV (src sequences) are located at the center of the FeSV genome and are approximately 1.5 kilobase pairs in length. With respect to the 5'-3' orientation of genes in viral RNA, the order of genes in the FeSV genome is 5'-gag-src-env-c region-3'; only 0.9 kilobase pairs of gag and 0.6 kilobase pairs of env-derived FeLV sequences are represented in ST FeSV. Heteroduplex analyses between lambda FeSV or lambda FeLV DNA and Moloney murine sarcoma virus DNA (strain m1) were performed under conditions of reduced stringency to demonstrate limited regions of base pair homology. Two such regions were identified: the first occurs at the extreme 5' end of the leukemia and both sarcoma viral genomes, whereas the second corresponds to a 5' segment of leukemia virus "env" sequences conserved in both sarcoma viruses. The latter sequences are localized at the 3' end of FeSV src and at the 5' end of murine sarcoma virus src and could possibly correspond to regions of helper virus genomes that are required for retroviral transforming functions.     

Genesis of Kirsten murine sarcoma virus: sequence analysis reveals recombination points and potential leukaemogenic determinant on parental leukaemia virus genome.

The genome of Kirsten murine sarcoma virus was formed by recombination between Kirsten murine leukaemia virus sequences, and rat sequences derived from a retrovirus-like '30S' (VL30) genetic element encompassing the Kras oncogene. Using cloned DNAs we have determined the nucleotide sequences of the long terminal repeats and adjacent regions, extending across the points of recombination on the sarcoma and leukaemia virus genomes. Our results suggest that discrete regions of homology and other cryptic sequence features, may have constituted recombinational hot-spots involved in the genesis of the Kirsten murine sarcoma virus genome. We have also compared the sequence of the Kirsten murine leukaemia virus p15 env and adjacent long terminal repeat with the corresponding regions of the AKV and Gross A murine leukaemia virus genomes. This comparison has identified a leukaemogenic determinant in the U3 domain of the long terminal repeat, possibly within a enhancer-like sequence element.     

The Parodi-Irgens feline sarcoma virus and simian sarcoma virus have homologous oncogenes, but in different contexts of the viral genomes

We have identified the oncogene and the putative transforming protein of the Parodi-Irgens feline sarcoma virus (PI-FeSV). The PI-FeSV is defective and needs a helper virus for its replication. The v-onc sequences in the PI-FeSV were found to be related to the v-sis sequences of the simian sarcoma virus (SSV). PI-FeSV nonproducer cells express two viral RNAs, a 6.8-and a 3.3-kilobase RNA. The 6.8-kilobase RNA contains gag, sis, and env sequences but lacks the pol gene. The 3.3-kilobase RNA, on the other hand, contains only env sequences. We have detected one feline leukemia virus-related protein product in these cells, namely, a 76-kilodalton protein which contains determinants of the feline leukemia virus gag proteins p15 and p30. The v-sis sequences in the PI-FeSV have been located near the 5' end of the viral genome. Taken together, these results imply that the p76 protein contains both feline leukemia virus gag and sis sequences and probably is the transforming protein of this virus. In contrast, in SSV the sis sequences are located towards the 3' end of the viral genome, and the sis protein is thought to be expressed via a subgenomic RNA. PI-FeSV and SSV therefore use different schemes to express their onc-related sequences. The v-sis sequences in the PI-FeSV contain restriction sites which reflect the different origin of the v-sis sequences in the PI-FeSV and SSV. The homologous oncogenes of the PI-FeSV and SSV thus were transduced by two different retroviruses, feline leukemia virus and the simian sarcoma-associated virus, apparently from the genomes of different species.     

Inversion of a chicken ets-1 proto-oncogene segment in avian leukemia virus E26.

The segment of the avian leukemia virus E26 genome near the termination of the p135gag-myb-ets open reading frame contains an inversion of the chicken ets-1 sequence. The inversion contains at least 41 bp and may be as large as 46 bp. This results in the replacement of 13 amino acids of chicken ets-1, with 16 amino acids derived from reverse complement of the normal ets-1 coding strand or read-through into E26 env sequences. At least 13 of these codons are specified by the inverted ets sequences. This represents the first reported occurrence of inverted oncogene sequences in a natural retrovirus. The inverted ets sequences are immediately followed by sequences homologous to the Rous sarcoma virus Prague B env gene. Since the E26 env sequence is more closely related to subgroup B avian retroviruses than to avian retroviruses from subgroups A, C, D, or E, the progenitor of E26 was a virus belonging to avian retrovirus subgroup B.     

Cloning and nucleotide sequences of cDNAs spanning the splice junctions of Rous sarcoma virus mRNAs.

The cDNAs corresponding to the 5' ends of the mRNAs coding for the envelope protein precursor (gPr92env) of the B77 strain and the transforming protein (pp60src) of the Prague B strain of Rous sarcoma virus were cloned into pBR322, and the nucleotide sequences surrounding the splice junctions were determined. Both mRNAs are products of single splicing events from a common donor splice site at nucleotide 398 from the 5' end of the RNA to acceptor splice sites at nucleotides 5078 and 7054 for the env and src mRNAs, respectively. These results confirm and extend previous conclusions based on peptide mapping and single-strand nuclease mapping. Compared with the sequence of the Prague C genome RNA, the B77 strain contains a 6-nucleotide deletion in the sequence corresponding to the hydrophobic portion of the signal peptide of the envelope protein precursor.     

Physical map of the Kirsten sarcoma virus genome as determined by fingerprinting RNase T1-resistant oligonucleotides.

From analysis of the large RNase T1-resistant oligonucleotides of Kirsten sarcoma virus (Ki-SV), a physical map of the virus genome was deduced. Kirsten murine leukemia virus (Ki-MuLV) sequences were detected in T1 oligonucleotides closest to the 3' end of the viral RNA and extended approximately 1,000 nucleotides into the genome. The rat genetic sequences started at this point and extended all the way to the very 5' end of the RNA molecules, where a small stretch of Ki-MuLV sequence was detected. By comparison of the fingerprints of Ki-SV RNA and the RNA of the endogenous rat src genetic sequences, it was found that more than 50% of the T1 oligonucleotides were similar between Ki-SV and the endogenous rat src RNA, suggesting an identical primary nucleotide sequence in over 50% of the viral genomes. The results indicate that Ki-SV arose by recombination between the 5' and 3' ends of Ki-MuLV and a large portion of the homologous sequences of the endogenous rat src RNA.     

The avian retrovirus env gene family: molecular analysis of host range and antigenic variants.

The nucleotide sequence of the env gp85-coding domain from two avian sarcoma and leukosis retrovirus isolates was determined to identify host range and antigenic determinants. The predicted amino acid sequence of gp85 from a subgroup D virus isolate of the Schmidt-Ruppin strain of Rous sarcoma virus was compared with the previously reported sequences of subgroup A, B, C, and E avian sarcoma and leukosis retroviruses. Subgroup D viruses are closely related to the subgroup B viruses but have an extended host range that includes the ability to penetrate certain mammalian cells. There are 27 amino acid differences shared between the subgroup D sequence and three subgroup B sequences. At 16 of these sites, the subgroup D sequence is identical to the sequence of one or more of the other subgroup viruses (A, C, and E). The remaining 11 sites are specific to subgroup D and show some clustering in the two large variable regions that are thought to be major determinants of host range. Biological analysis of recombinant viruses containing a dominant selectable marker confirmed the role of the gp85-coding domain in determining the host range of the subgroup D virus in the infection of mammalian cells. We also compared the sequence of the gp85-coding domain from two subgroup A viruses, Rous-associated virus type 1 and a subgroup A virus of the Schmidt-Ruppin strain of Rous sarcoma virus. The comparison revealed 24 nonconservative amino acid changes, of which 6 result in changes in potential glycosylation sites. The positions of 10 amino acid differences are coincident with the positions of 10 differences found between two subgroup B virus env gene sequences. These 10 sites identify seven domains in the sequence which may constitute determinants of type-specific antigenicity. Using a molecular recombinant, we demonstrated that type-specific neutralization of two subgroup A viruses was associated with the gp85-coding domain of the virus.     

Regulation of expression and chromosomal subunit conformation of avian retrovirus genomes.

We have investigated the copy number, chromosomal subunit conformation and regulation of expression of integrated avian retrovirus genomes. Our results indicate that there are approximately two copies of the endogenous viral genomes (RAV-O) per haploid cell genome in uninfected chick embryo fibroblasts (CEF) and red blood cells (RBC). The copy number and subunit conformation (as measured by DNAasel sensitivity) of the RAV-O genomes are independent of the level of expression of these viral DNA sequences. In cells isolated from embryos of the V+, gs-chf- and gs+chf+ phenotypes, approximately one of the two viral genomes is in a DNAase l-sensitive conformation. Upon infection with an exogenous Rous sarcoma virus (PR-RSV-C), one new viral genome is integrated per haploid CEF genome. The newly integrated RSV genome is completely sensitive to DNAase l, and the subunit conformation of the endogenous viral genomes is not altered by the integration of additional exogenous proviruses. Both the endogenous and newly integrated exogenous viral genomes are present in "nu-body" structures, and the selective sensitivity of these proviral DNA sequences to DNAase l is maintained in isolated nucleosomes. Our experiments revealing the DNAase l sensitivity of one of the two RAV-O genomes in gs-chf-CEF led us to reexamine the level of viral specific RNA in CEF of various GS genotypes. We find that GS/GS CEF contain approximately 100 copies of viral RNA per cell, gs/gs CEF contain no detectable viral RNA, and the heterozygote GS/gs CEF contain approximately 50 copies of viral specific RNA per cell. These results suggest that the GS gene controls production of RAV-O RNA sequences in CEF in a "cis" fashion. In RBCs, however, the expression of the RAV-O genome is independent of the GS gene, with both GS/GS and gs/gs RBCs containing roughly equivalent amounts of viral specific RNA. Our results suggest that the chromosomal structure of the endogenous viral genes is independent of the GS gene, and that the GS gene is cis-acting and tissue-specific.     

The R-U5-5'' leader sequence of neurovirulent wild mouse retrovirus contains an element controlling the incubation period of neurodegenerative disease.

The wild mouse ecotropic retrovirus CasBrE causes a spongiform neurodegenerative disease after neonatal inoculation, with an incubation period ranging from 2 to 12 months. We previously showed that introduction of long terminal repeat (LTR) and gag-pol sequences from a strain of Friend murine leukemia virus (FB29) resulted in a dramatic acceleration of the onset of the disease. The chimeric virus FrCasE, which consisted of the FB29 genome containing 3' pol and env sequences from the wild mouse virus, induced a highly predictable, lethal neurodegenerative disease with an incubation period of only 16 days. Here we report that the sequences which are primary determinants of the length of the incubation period are located in the 5' end of the viral genome between a KpnI site in the R region of the LTR and a PstI site immediately 5' of the start codon for pr65gag (R-U5-5' leader). This region contains the tRNA primer binding site, splice donor site for the subgenomic env mRNA, and the packaging sequence. Computer-assisted sequence analysis failed to find evidence of a consensus sequence for a DNA enhancer in this region. In addition, sequences within a region of the genome between a ClaI site at the 3' end of env to the KpnI site in the R region of the LTR (inclusive of U3) also influenced the incubation period of the disease, but the effect was distinctly weaker than that of the R-U5-5' leader sequence. This U3 effect, however, appeared to be independent of the number of direct repeats, since deletion of one of two duplicated 42-base repeats containing consensus sequences of nuclear-factor binding domains had no effect on the incubation period of the disease. On the basis of Southern blot analysis of total viral DNA in the tissues, the effect of these sequences on the incubation period appeared to be related to the level of virus replication in the central nervous system. All of the chimeric viruses analyzed, irrespective of neurovirulence, replicated to comparable levels in the spleen and induced comparable levels of viremia.     

Nucleotide sequence of the Syrian hamster intracisternal A-particle gene: close evolutionary relationship of type A particle gene to types B and D oncovirus genes.

We determined the complete nucleotide sequence of the intracisternal A-particle gene, IAP-H18, cloned from the normal Syrian hamster liver DNA. IAP-H18 was 7,951 base pairs in length with two identical long terminal repeats of 376 base pairs at both ends. On the coding strand, imperfect open reading frames corresponding to gag and pol of the retrovirus genome were observed, whereas many stop codons were present in the region corresponding to env. The putative H18 gag gene (809 amino acids) had a sequence homologous to the N-terminal half of the mouse mammary tumor virus gag gene and locally to the Rous sarcoma virus gag gene. The putative H18 pol gene (900 residues) was homologous to the Rous sarcoma virus pol gene almost throughout the entire region. Two conserved regions among the retrovirus pol genes have been reported. One presumably corresponds to the DNA polymerase and the RNase H domain, and the other corresponds to the DNA endonuclease domain of the multifunctional protein pol. By the comparison of the deduced amino acid sequences of the putative endonuclease domain of six representative oncovirus genomes, a phylogenetic tree of the oncovirus genomes was constructed, and the intracisternal A-particle (type A) genome was found to be more closely related to the mouse mammary tumor virus (type B) and squirrel monkey retrovirus (type D) genomes.     

Cloning and sequence analysis of genome from the Inner Mongolia strain of the endogenous betaretroviruses (enJSRV)

In order to amplify the complete genome of enJSRV from the strain of Inner Mongolia (enJSRV-NM), we used enJSRV-specific and JSRV-specific DNA probes in dot blot hybridization. Seven pairs of primers were designed based on Genbank sequences. Seven fragments were obtained by PCR and were cloned into the PMD19- T vectors. The recombinant plasmids were sequenced and analyzed. The results showed that the genome was 7 942 bp in length and contained four overlapping open reading frames corresponding to the gag, pro, pol and env genes as well as an additional open reading frame (orf-x) that overlaps the 3' end of the pol gene. The nucleotide acid sequences of the enJSRV-NM loci were compared with the sequences of South Africa enJS56A1 strain (Accession No. AF153615) and USA JSRV21 strain (Accession No. AF105220). The nucleotide acid identities were 99.2% and 92.3% respectively. Two zinc fingers were found in the NC region in the predicted amino acid sequence. However, the YXXM motif, which is a reliable molecular marker for the infectious exogenous virus, was not found in the TM region. It was found that the enJSRV-NM region was 90%-98% identical at the amino acid level to its exogenous infectious counterparts in most of the retroviral genome. This is the first nucleotide sequence of enJSRV reported in P.R China. The resource work has provided a wide range of information useful not only for expression genomics and annotation of genomic DNA sequence, but also for further research on the clinical diagnosis of OPA.     

Cloning and Sequence Analysis of Genome from the Inner Mongolia Strain of the Endogenous Betaretroviruses （enJSRV）

In order to amplify the complete genome of enJSRV from the strain of Inner Mongolia （enJSRV-NM）, we used enJSRV-specific and JSRV-specific DNA probes in dot blot hybridization. Seven pairs of primers were designed based on Genbank sequences. Seven fragments were obtained by PCR and were cloned into the PMD19-T vectors. The recombinant plasmids were sequenced and analyzed. The results showed that the genome was 7 942 bp in length and contained four overlapping open reading frames corresponding to the gag, pro, pol and env genes as well as an additional open reading frame （orf-x） that overlaps the 3＇ end of the pol gene. The nucleotide acid sequences of the enJSRV-NM loci were compared with the sequences of South Africa enJS56A1 strain （Accession No.AF153615） and USA JSRV21 strain （Accession No. AF105220）. The nucleotide acid identities were 99.2% and 92.3% respectively. Two zinc fingers were found in the NC region in the predicted amino acid sequence. However, the YXXM motif, which is a reliable molecular marker for the infectious exogenous virus, was not found in the TM region. It was found that the enJSRV-NM region was 90%-98% identical at the amino acid level to its exogenous infectious counterparts in most of the retroviral genome. This is the first nucleotide sequence of enJSRV reported in P.R China. The resource work has provided a wide range of information useful not only for expression genomics and annotation of genomic DNA sequence, but also for further research on the clinical diagnosis of OPA.     

Genome of Reticuloendotheliosis Virus: Characterization by Use of Cloned Proviral DNA

Reticuloendotheliosis virus is an avian type C retrovirus that is capable of transforming fibroblasts and hematopoietic cells both in vivo and in vitro. This virus is highly related to the three other members of the reticuloendotheliosis virus group, including spleen necrosis virus, but it is apparently unrelated to the avian leukosis-sarcoma virus family. Previous studies have shown that it consists of a replication-competent helper virus (designated REV-A) and a defective component (designated REV) that is responsible for transformation. In this study we used restriction endonuclease mapping and heteroduplex analysis to characterize the proviral DNAs of REV-A and REV. Both producer and nonproducer transformed chicken spleen cells were used as sources of REV proviral DNA; this genome was mapped in detail, and fragments of it were cloned in lambdagtWES.lambdaB. The infected canine thymus line Cf2Th(REV-A) was used as a source of REV-A proviral DNA. The restriction maps and heteroduplexes of the REV and REV-A genomes showed that (proceeding from 5' to 3') (i) REV contains a large fraction of the REV-A gag gene (assuming a gene order of gag-pol-env and gene sizes similar to those of other type C viruses), for the two genomes are very similar over a distance of 2.1 kilobases beginning at their 5' termini; (ii) most or all of REV-A pol is deleted in REV; (iii) REV contains a 1.1 kilobase segment derived from the 3' end of REV-A pol or the 5' end of env or both; (iv) this env region in REV is followed by a 1.9-kilobase segment which is unrelated to REV-A; and (v) the helper-unrelated segment of REV extends essentially all of the way to the beginning of the 3' long terminal repeat. Therefore, like avian myeloblastosis virus but unlike the other avian acute leukemia viruses and most mammalian and avian sarcoma viruses, REV appears to be an env gene recombinant. We also found that the REV-specific segment is derived from avian DNA, for a cloned REV fragment was able to hybridize with the DNA from an uninfected chicken. Therefore, like the other acute transforming viruses, REV appears to be the product of recombination between a replication-competent virus and host DNA. Two other defective genomes in virus-producing chicken cells were also cloned and characterized. One was very similar to REV in its presumptive gag and env segments, but instead of a host-derived insertion it contained additional env sequences. The second was similar (but not identical) to the first in its gag and env regions and appeared to contain an additional 1-kilobase inversion of REV-A sequences.