Mechanisms for the generation of src-deletion mutants and recovered sarcoma viruses: identification of viral sequences involved in src deletions and in recombination with c-src sequences

Genetic polymorphism of the nsp2 gene in North American type--porcine reproductive and respiratory syndrome virus

We determined the complete nucleotide sequence of EDRD-1, a Japanese strain of the North American type-Porcine reproductive and respiratory syndrome virus (PRRSV), and identified a novel 117-base deletion and 108-base insertion previously reported in the nsp2 gene of the SP strain, which contains the largest genome among PRRSV strains. Based on genetic analysis of the partial nsp2 gene in 30 additional Japanese isolates and 50 strains from various countries, we classified North American-type PRRSVs into three nsp2-types, represented by EDRD-1, which contains the 117-base deletion and 108-base insertion; prototypic VR-2332, which does not contain the deletion and insertion; and SP, which contains only the 108-base insertion. The three nsp2-types were phylogenetically separated, suggesting that these structural changes only occurred at earlier stages of viral evolution. In the nsp2 genes, we identified an additional 19 deletions ranging from 3 to 378 bases and 2 insertions of 3 and 21 bases which were not common within each nsp2-type, suggesting that these changes occurred at later stages of viral evolution. In addition, our results suggest that the three nsp2-types can be rapidly differentiated by RT-PCR using their polymorphisms as natural tags.

Genomic stability of murine leukemia viruses containing insertions at the Env-3' untranslated region boundary

Retroviruses containing inserts of exogenous sequences frequently eliminate the inserted sequences upon spread in susceptible cells. We have constructed replication-competent murine leukemia virus (MLV) vectors containing internal ribosome entry site (IRES)-transgene cassettes at the env-3' untranslated region boundary in order to examine the effects of insert sequence and size on the loss of inserts during viral replication. A virus containing an insertion of 1.6 kb replicated with greatly attenuated kinetics relative to wild-type virus and lost the inserted sequences in a single infection cycle. In contrast, MLVs containing inserts of 1.15 to 1.30 kb replicated with kinetics only slightly attenuated compared to wild-type MLV and exhibited much greater stability, maintaining their genomic integrity over multiple serial infection cycles. Eventually, multiple species of deletion mutants were detected simultaneously in later infection cycles; once detected, these variants rapidly dominated the population and thereafter appeared to be maintained at a relative equilibrium. Sequence analysis of these variants identified preferred sites of recombination in the parental viruses, including both short direct repeats and inverted repeats. One instance of insert deletion through recombination with an endogenous retrovirus was also observed. When specific sequences involved in these recombination events were eliminated, deletion variants still arose with the same kinetics upon virus passage and by apparently similar mechanisms, although at different locations in the vectors. Our results suggest that while lengthened, insert-containing genomes can be maintained over multiple replication cycles, preferential deletions resulting in loss of the inserted sequences confer a strong selective advantage.

Molecular and biochemical bases for activation of the transforming potential of the proto-oncogene c-ros

The transforming gene of avian sarcoma virus UR2, v-ros, encodes a receptor-like protein tyrosine kinase and differs from its proto-oncogene, c-ros, in its 5' truncation and fusion to viral gag, a three-amino-acid (aa) insertion in the transmembrane (TM) domain, and changes in the carboxyl region. To explore the basis for activation of the c-ros transforming potential, various c-ros retroviral vectors containing those changes were constructed and studied for their biological and biochemical properties. Ufcros codes for the full-length c-ros protein of 2,311 aa, Uppcros has 1,661-aa internal deletion in the extracellular domain, CCros contains the 3' c-ros cDNA fused 150 aa upstream of the TM domain to the UR2 gag, CVros is the same as CCros except that the 3' region is replaced by that of v-ros, and VCros is the same as CCros except that the 5' region is replaced by that of v-ros. The Ufcros, Uppcros, CCros, and CVros are inactive in transforming chicken embryo fibroblasts, whereas VCros is as potent as UR2 in cell-transforming and tumorigenic activities. Upon passages of CCros and CVros viruses, the additional extracellular sequence in comparison with that of v-ros was delected; concurrently, both viruses (named CC5d and CV5d, respectively) attained moderate transforming activity, albeit significantly lower than that of UR2 or VCros. The native c-ros protein has a very low protein tyrosine kinase activity, whereas the ppcros protein is constitutively activated in kinase activity. The inability of CCros and CVros to transform chicken embryo fibroblasts is consistent with the inefficient membrane association, instability, and low kinase activity of their encoded proteins. The CC5d and CV5d proteins are indistinguishable in kinase activity, membrane association, and stability from the v-ros protein. The reduced transforming potency of CC5d and CV5d proteins can be attributed only to their differential substrate interaction, notably the failure to phosphorylate a 88-kDa protein. We conclude that the 5' rather than the 3' modification of c-ros is essential for its oncogenic activation; the sequence upstream of the TM domain has a negative effect on the transforming activity of CCros and CVros and needs to be deleted to activate their biological activity.

Transforming variants of the avian myc-containing retrovirus FH3 arise prior to phenotypic selection

The avian retrovirus FH3, which encodes a Gag-Myc fusion protein, transforms chicken macrophages but not fibroblasts. However, passage of FH3 viral stock in fibroblasts leads to emergence of a virus capable of fibroblast transformation. This virus has not acquired myc mutations; instead, it carries internal gag deletions which confer the ability to transform fibroblasts. We now demonstrate that this and similar deletion variants emerge repeatedly during selection. Sequence analysis reveals direct repeats at or near deletion junctions, suggesting that errors during reverse transcription may be involved in genesis of these viruses, which are then positively selected in fibroblast culture. By using the polymerase chain reaction, we found that such variants preexisted in original stocks even before selection, although they could not be detected by focus assay.

Isolation of the avian transforming retrovirus, AS42, carrying the v-maf oncogene and initial characterization of its gene product

A novel avian transforming retrovirus was isolated from a chicken musculoaponeurotic fibrosarcoma. This virus (called AS42) induces tumors histopathologically indistinguishable from the original sarcoma after a long latent period when inoculated into newborn chickens. AS42 also exhibits a weak transforming activity when infected into chicken embryo fibroblasts (CEF). This virus is replication-defective and associated with a helper virus of subgroup A (called ASAV). An AS42-specific protein of about 100 kDa was immunoprecipitated from lysates of AS42-transformed CEF with antiserum directed against avian retrovirus virion proteins. Molecular analysis of the genomic structure of the AS42 virus has revealed that this 100-kDa protein represents a novel oncogene, v-maf of cellular origin, which is fused with a part of the viral gag gene (Nishizawa et al., Proc. Natl. Acad. Sci. USA 86, 7711-7715, 1989). Interestingly, some size variation was observed among the gag-maf fusion proteins found in individual clones of transformed CEF. Consistent with this observation, Southern blot analyses and nucleotide sequence determination of several independent isolates of proviral DNA indicated that this virus segregates multiple forms of deletion mutants, probably through homologous recombinations among the repetitive sequences present within the v-maf coding region.

Analysis of cDNAs of the proto-oncogene c-src: heterogeneity in 5' exons and possible mechanism for the genesis of the 3' end of v-src

To further characterize the gene structure of the proto-oncogene c-src and the mechanism for the genesis of the v-src sequence in Rous sarcoma virus, we have analyzed genomic and cDNA copies of the chicken c-src gene. From a cDNA library of chicken embryo fibroblasts, we isolated and sequenced several overlapping cDNA clones covering the full length of the 4-kb c-src mRNA. The cDNA sequence contains a 1.84-kb sequence downstream from the 1.6-kb pp60c-src coding region. An open reading frame of 217 amino acids, called sdr (src downstream region), was found 105 nucleotides from the termination codon for pp60c-src. Within the 3' noncoding region, a 39-bp sequence corresponding to the 3' end of the RSV v-src was detected 660 bases downstream of the pp60c-src termination codon. The presence of this sequence in the c-src mRNA exon supports a model involving an RNA intermediate during transduction of the c-src sequence. The 5' region of the c-src cDNA was determined by analyzing several cDNA clones generated by conventional cloning methods and by polymerase chain reaction. Sequences of these chicken embryo fibroblast clones plus two c-src cDNA clones isolated from a brain cDNA library show that there is considerable heterogeneity in sequences upstream from the c-src coding sequence. Within this region, which contains at least 300 nucleotides upstream of the translational initiation site in exon 2, there exist at least two exons in each cDNA which fall into five cDNA classes. Four unique 5' exon sequences, designated exons UE1, UE2, UEX, and UEY, were observed. All of them are spliced to the previously characterized c-src exons 1 and 2 with the exception of type 2 cDNA. In type 2, the exon 1 is spliced to a novel downstream exon, designated exon 1a, which maps in the region of the c-src DNA defined previously as intron 1. Exon UE1 is rich in G+C content and is mapped at 7.8 kb upstream from exon 1. This exon is also present in the two cDNA clones from the brain cDNA library. Exon UE2 is located at 8.5 kb upstream from exon 1. The precise locations of exons UEX and UEY have not been determined, but both are more than 12 kb upstream from exon 1. The existence and exon arrangements of these 5' cDNAs were further confirmed by RNase protection assays and polymerase chain reactions using specific primers. Our findings indicate that the heterogeneity in the 5' sequences of the c-src mRNAs results from differential splicing and perhaps use of distinct initiation sites. All of these RNAs have the potential of coding for pp60c-src, since their 5' exons are all eventually joined to exon 2.

Analysis of cDNAs of the proto-oncogene c-src: heterogeneity in 5' exons and possible mechanism for the genesis of the 3' end of v-src

To further characterize the gene structure of the proto-oncogene c-src and the mechanism for the genesis of the v-src sequence in Rous sarcoma virus, we have analyzed genomic and cDNA copies of the chicken c-src gene. From a cDNA library of chicken embryo fibroblasts, we isolated and sequenced several overlapping cDNA clones covering the full length of the 4-kb c-src mRNA. The cDNA sequence contains a 1.84-kb sequence downstream from the 1.6-kb pp60c-src coding region. An open reading frame of 217 amino acids, called sdr (src downstream region), was found 105 nucleotides from the termination codon for pp60c-src. Within the 3' noncoding region, a 39-bp sequence corresponding to the 3' end of the RSV v-src was detected 660 bases downstream of the pp60c-src termination codon. The presence of this sequence in the c-src mRNA exon supports a model involving an RNA intermediate during transduction of the c-src sequence. The 5' region of the c-src cDNA was determined by analyzing several cDNA clones generated by conventional cloning methods and by polymerase chain reaction. Sequences of these chicken embryo fibroblast clones plus two c-src cDNA clones isolated from a brain cDNA library show that there is considerable heterogeneity in sequences upstream from the c-src coding sequence. Within this region, which contains at least 300 nucleotides upstream of the translational initiation site in exon 2, there exist at least two exons in each cDNA which fall into five cDNA classes. Four unique 5' exon sequences, designated exons UE1, UE2, UEX, and UEY, were observed. All of them are spliced to the previously characterized c-src exons 1 and 2 with the exception of type 2 cDNA. In type 2, the exon 1 is spliced to a novel downstream exon, designated exon 1a, which maps in the region of the c-src DNA defined previously as intron 1. Exon UE1 is rich in G+C content and is mapped at 7.8 kb upstream from exon 1. This exon is also present in the two cDNA clones from the brain cDNA library. Exon UE2 is located at 8.5 kb upstream from exon 1. The precise locations of exons UEX and UEY have not been determined, but both are more than 12 kb upstream from exon 1. The existence and exon arrangements of these 5' cDNAs were further confirmed by RNase protection assays and polymerase chain reactions using specific primers. Our findings indicate that the heterogeneity in the 5' sequences of the c-src mRNAs results from differential splicing and perhaps use of distinct initiation sites. All of these RNAs have the potential of coding for pp60c-src, since their 5' exons are all eventually joined to exon 2.

Role of gag sequence in the biochemical properties and transforming activity of the avian sarcoma virus UR2-encoded gag-ros fusion protein

The transforming protein P68gag-ros of avian sarcoma virus UR2 is a transmembrane tyrosine protein kinase molecule with the gag portion protruding extracellularly. To investigate the role of the gag moiety in the biochemical properties and biological functions of the P68gag-ros fusion protein, retroviruses containing the ros coding sequence of UR2 were constructed and analyzed. The gag-free ros protein was expressed from one of the mutant retroviruses at a level 10 to 50% of that of the wild-type UR2. However, the gag-free ros-containing viruses were not able to either transform chicken embryo fibroblasts or induce tumors in chickens. The specific tyrosine protein kinase activity of gag-free ros protein is about 10- to 20-fold reduced as judged by in vitro autophosphorylation. The gag-free ros protein is still capable of associating with membrane fractions including the plasma membrane, indicating that sequences essential for recognition and binding membranes must be located within ros. Upon passages of the gag-free mutants, transforming and tumorigenic variants occasionally emerged. The variants were found to have regained the gag sequence fused to the 5' end of the ros, apparently via recombination with the helper virus or through intramolecular recombination between ros and upstream gag sequences in the same virus construct. All three variants analyzed code for gag-ros fusion protein larger than 68 kDa. The gag-ros recombination junction of one of the transforming variants was sequenced and found to consist of a p19-p10-p27-ros fusion sequence. We conclude that the gag sequence is essential for the transforming activity of P68gag-ros but is not important for its membrane association.

Transformation-defective mutants with 5' deletions of the src gene are frequently generated during replication of Rous sarcoma virus in established quail fibroblasts

Replication of Rous sarcoma virus (RSV) in avian fibroblasts leads to the generation of replication-competent variants that are defective for cell transformation (td virus). These td variants contain deletions affecting various portions of the v-src gene. We compared the rate of td virus production in Q3B cells, a quail cell line established by mutagen treatment, and in normal quail fibroblasts. Twenty-five days after infection with an RSV stock containing only transforming virions, Q3B cells harbor similar amounts of v-src-containing and v-src-deleted proviruses. However, these cells synthesize very low levels of p60v-src and generate large excess of td variants, as determined by biological assays. Unlike Q3B cells, normal quail fibroblasts infected with the same virus stock produce td variants only after multiple passages of undiluted virus on fresh cells. Restriction analysis showed that the td virus produced by Q3B cells is composed of two types of genomes: one lacking the entire v-src gene and the other carrying partial deletions of this gene predominantly located in the amino-terminal portion of the coding region of v-src. To study the mechanisms of these partial deletions, we molecularly cloned and sequenced the v-src genes of several td proviruses. We show that these mutants carry single or multiple v-src deletions of limited size, presumably generated by multiple mechanisms. Two deletions of 170 and 112 bp located in the 5' portion of v-src are frequently generated during RSV replication in Q3B cells and may represent preferential sites for v-src deletion in these cells.

Rapid reversion of a deletion mutation in Moloney murine leukemia virus by recombination with a closely related endogenous provirus

During abortive infection of mouse cells, defective retroviruses carrying deletions in essential functions can recombine with endogenous retroviral sequences to form viable, replication-competent viruses. We have examined the reversion of a mutant Moloney murine leukemia virus with a deletion in the protease domain of the pol gene after infection of NIH/3T3 cells. In this system revertants arise quickly, only 2 weeks after infection. Analysis of DNA clones of the revertant viral genomes showed that they were derived by recombination with a long sequence of gag and pol exhibiting 95% sequence identity to Moloney virus. One such cloned recombinant was fully infectious, indicating that the repertoire of viral sequences in the NIH/3T3 genome must include substantial stretches of functional viral genes. Examination of the viral DNAs very early in the infection revealed the presence of defective genomes, formed by nonhomologous crossovers between the two parental sequences. We suggest that these may serve as intermediates in the eventual formation of the viable revertant genomes.

Transduction of proto-src sequences in tissue culture by a molecular clone of transformation-defective Rous sarcoma virus with an internal src deletion

The sporadic appearance of nondefective (nd) Rous sarcoma virus (RSV) from cells in tissue culture infected with a molecular clone of transformation-defective RSV was examined. Southern analysis of extrachromosomal, virus-specific DNA of three independent ndRSV isolates in each case indicated restoration of an isogenic src by homologous recombination with cellular proto-src. The frequency of transduction was estimated by fluctuation analysis to vary between one transduction per 0.4 x 10(7) to 1.6 x 10(7) infected cells.

Avian sarcoma viruses

Twelve independent isolates of avian sarcoma viruses (ASVs) can be divided into four groups according to the transforming genes harbored in the viral genomes. The first group is represented by viruses containing the transforming sequence, src, inserted in the viral genome as an independent gene; the other three groups of viruses contain transforming genes fps, yes or ros fused to various length of the truncated structural gene gag. These transforming sequences have been obtained by avian retroviruses from chicken cellular DNA by recombination. The src-containing viruses code for an independent polypeptide, p60src; and the representative fps, yes and ros-containing ASVs code for P140/130gag-fps, P90gag-yes and P68gag-ros fusion polypeptides respectively. All of these transforming proteins are associated with the tyrosine-specific protein kinase activity capable of autophosphorylation and phosphorylating certain foreign substrates. p60src and P68gag-ros are integral cellular membrane proteins and P140/130gag-fps and P90gag-yes are only loosely associated with the plasma membrane. Cells transformed by ASVs contain many newly phosphorylated proteins and in most cases have an elevated level of total phosphotyrosine. However, no definitive correlation between phosphorylation of a particular substrate and transformation has been established except that a marked increase of the tyrosine phosphorylation of a 34,000 to 37,000 dalton protein is observed in most ASV transformed cells. The kinase activity of ASV transforming proteins appears to be essential, but not sufficient for transformation. The N-terminal domain of p60src required for myristylation and membrane binding is also crucial for transformation. By contrast, the gag portion of the FSV P130gag-fps is dispensable for in vitro transformation and removal of it has only an attenuating effect on in vivo tumorigenicity. The products of cellular src, fps and yes proto-oncogenes have been identified and shown to also have tyrosine-specific protein kinase activity. The transforming potential of c-src and c-fps has been studied and shown that certain structural changes are necessary to convert them into transforming genes. Among the cellular proto-oncogenes related to the four ASV transforming genes, c-ros most likely codes for a growth factor receptor-like molecule. It is possible that the oncogene products of ASVs act through certain membrane receptor(s) or enzyme(s), such as protein kinase C, in the process of cell transformation.

The mechanism of transduction of proto-oncogene c-src by avian retroviruses

Chicken c-src sequences have been transduced by avian leukosis viruses (ALV) and by partial src-deletion (td) mutants of Rous sarcoma virus in several independent events. Analyses of the recombination junctions in the genomes of src-containing viruses and the c-src DNA have shed light on the mechanism of transduction, which involves at least two steps of recombination. The initial recombination between a viral genome and the 5' region of c-src appears to occur at the DNA level. This step does not require extensive homology and can be mediated by stretches of sequences with only partial homology. The 5' recombination junction can also be formed by splicing between viral and c-src sequences. The second recombination is presumed to occur between the transducing ALV or td viral RNA and the viral-c-src hybrid RNA molecule generated from the initial recombination. This step involving recombination at the 3' ends of those molecules restores the 3' viral sequences essential for replication to the viral-c-src hybrid molecule. High frequency of c-src transduction by partial td mutants suggests that the second recombination is greatly enhanced when there is sequence homology between the transducing virus and the 3' region of c-src. Incorporation of the c-src sequences into an ALV genome results in greatly elevated expression of the gene. However, increased expression of c-src alone is insufficient to activate its transforming potential. Structural changes in c-src are necessary to convert it into a transforming gene. The changes can be as small as single nucleotide changes resulting in single amino aid substitutions at certain positions. Mutations can occur rapidly during viral replication after c-src is incorporated into the viral genome. Therefore, it is most likely that transduction of c-src by ALV is followed by subsequent mutation and selection for the sarcomagenic virus. In the case of transduction by td viruses that retain certain src sequences, joining of these sequences with the transduced c-src apparently is sufficient to activate its transforming potential.

NH2-terminal sequences of two src proteins that cause aberrant transformation

Two isolates of recovered avian sarcoma viruses (rASVs), rASV157 and rASV1702, transform cells in culture, but have greatly reduced in vivo tumorigenicity. The src proteins of rASV157 and rASV1702 have alterations within their NH2 termini, are not myristoylated, and have an altered subcellular localization. We have molecularly cloned and determined the nucleotide sequences of the src genes of rASV157 and rASV1702. We found that their src proteins have unusual NH2 termini: the rASV157 src protein NH2 terminus consists of 30 amino acids of the env signal peptide attached to Ser-6 of the src sequence, while the rASV1702 src protein NH2 terminus consists of 45 amino acids of the env signal peptide attached to Ala-76 of the src sequence. Expression of recombinant Rous sarcoma virus constructs containing the molecularly cloned rASV src genes produced src proteins with the same properties as those of the parental viruses. Our results suggest that the NH2-terminal structures are responsible for many unusual properties of the mutant src proteins.

Transduction of c-src coding and intron sequences by a transformation-defective deletion mutant of Rous sarcoma virus

The mechanism of cellular src (c-src) transduction by a transformation-defective deletion mutant, td109, of Rous sarcoma virus was studied by sequence analysis of the recombinational junctions in three td109-derived recovered sarcoma viruses (rASVs). Our results show that two rASVs have been generated by recombination between td109 and c-src at the region between exons 1 and 2 defined previously. Significant homology between td109 and c-src sequences was present at the sites of recombination. The viral and c-src sequence junction of the third rASV was formed by splicing a cryptic donor site at the 5' region of env of td109 to exon 1 of c-src. Various lengths of c-src internal intron 1 sequences were incorporated into all three rASV genomes, which resulted from activation of potential splice donor and acceptor sites. The incorporated intron 1 sequences were absent in the c-src mRNA, excluding its being the precursor for recombination with td109 and implying that initial recombinations most likely took place at the DNA level. A potential splice acceptor site within the incorporated intron 1 sequences in two rASVs was activated and was used for the src mRNA synthesis in infected cells. The normal env mRNA splice acceptor site was used for src mRNA synthesis for the third rASV.

Two independent mutations are required for temperature-sensitive cell transformation by a Rous sarcoma virus temperature-sensitive mutant

We molecularly cloned the src coding region of tsNY68, a mutant of Rous sarcoma virus temperature sensitive (ts) for transformation, and constructed a series of ts wild-type recombinant src genes. DNA containing the hybrid genes was transfected into chicken cells together with viral vector DNA and helper viral DNA, and infectious transforming viruses were recovered. Characterization of these recombinant viruses indicated that at least two mutations are present in the 3' half of the mutant src gene, both of which are required for ts. Nucleotide sequence analysis revealed three differences in the deduced amino acid sequence compared with the parental virus. Two of these changes, a deletion of amino acids 352 to 354 and an amino acid substitution at position 461, are responsible for the ts phenotype.

Partial nucleotide sequence of Rous sarcoma virus-29 provides evidence that the original Rous sarcoma virus was replication defective

Rous sarcoma virus-29 (RSV-29) is the strain of RSV that has the least number of passages beyond its isolation from chicken tumor no. 1 among all current strains of RSV. Biological characterization indicated that it was replication defective. RNA analysis of nonproducer clones of RSV-29-infected chicken embryonic fibroblasts showed the presence of a subgenomic message of 2.6 kilobases containing src and a genomic RNA of 7.7 kilobases that contains gag, pol, and src, but not env. The src-containing EcoRI fragment of RSV-29 proviral DNA was molecularly cloned. Sequence analysis of the regions flanking src revealed that the env gene was completely deleted in RSV-29 and that the sequence across the deletion was exactly the same as the Bryan high-titer strain of RSV. The sequence immediately 3' to src in RSV-29 was closely related to that of the Prague strain of RSV. The fact that the strain of RSV which has the minimal number of passages beyond its isolation is replication defective supports the hypothesis of Lerner and Hanafusa (J. Virol. 49:549-556, 1984) that the original RSV is a defective transforming virus. This defective transforming virus is postulated to be the precursor to other defective RSVs like the Bryan high-titer strain and to nondefective RSVs like the Prague strain. The particular clone of RSV-29 that we studied also had a short stretch of sequence duplication at the 3' end of the pol gene, which was presumably created by an error of reverse transcription.

Unusual features of the leader sequence of Rous sarcoma virus packaging mutant TK15

TK15, a mutant derived from a temperature-sensitive mutant of Rous sarcoma virus (tsNY68), has extremely low infectivity although it has intact viral genes. Biological and biochemical analyses of the virus and virus-induced transformants showed that the mutant has a defect in the packaging of its own genomic RNA possibly owing to a deletion near the 5' end (S. Kawai and T. Koyama, J. Virol. 51:147-153, 1984; T. Koyama, F. Harada, and S. Kawai, J. Virol. 51:154-162, 1984). Nucleotide sequence analysis of the provirus DNA of the mutant revealed that the deletion extends from the 3' end of the primer binding site to 22 bases upstream of the gag initiation codon and also suggested that possible binding between an extra region of the primer molecule and the viral genome resulting from the deletion causes another defect in the replication of the TK15 genome. It was suggested that the deletion of 237 bases in TK15 was generated during reverse transcription of the genome by the skipping of a sequence between identical 13-base sequences present in the primer binding site and 35 to 22 bases upstream of the gag initiation codon of the parental virus.

Comparison of the transcriptional properties of the Friend and Moloney retrovirus long terminal repeats: importance of tandem duplications and of the core enhancer sequence

The effects of sequence differences within the long terminal repeats (LTRs) of various murine retroviruses on transcription are examined by linking genetically engineered recombinant LTRs to the protein coding region of the herpes simplex virus thymidine kinase (TK) gene and assaying TK expression in tissue culture fibroblasts. The Goldberg/Hogness box region and the enhancer region are examined independently. We find that the Friend and Moloney Goldberg/Hogness boxes (TATAAAA and AATAAAA, respectively) give similar results in this assay, whereas differences between the sequences in the enhancer region have a marked effect. Tandem duplications increase the transcription level of the LTR. A single nucleotide transition in the core enhancer sequence has as large an effect on transcription as the presence of a tandem duplication: thymidine in the fifth position of the core enhancer (TCTGTGGTAAG) leads to a much higher transcriptional activity than cytidine. The LTRs have been implicated by others in the oncogenic potential of murine retroviruses, which is perhaps dependent on the transcriptional properties endowed in part by the core enhancer and tandem duplications.