Discontinuities in the DNA synthesized by an avian retrovirus

The unintegrated linear DNA synthesized in cells infected by Rous sarcoma virus is a predominantly double-stranded structure in which most of the minus-strand DNA, complementary to the viral RNA genome, is genome sized, whereas the plus-strand DNA is present as subgenomic fragments. We previously reported the application of benzoylated naphthoylated DEAE-cellulose chromatography to demonstrate that of the linear viral DNA species synthesized in quail embryo fibroblasts infected with Rous sarcoma virus greater than 99.5% contain single-stranded regions and these regions are predominantly composed of plus-strand DNA sequences (T. W. Hsu and J. M. Taylor, J. Virol. 44:47-53, 1982). We now present the following additional findings. (i) There were on the average 3.5 single-stranded regions per linear viral DNA, and these single-stranded regions could occur at many locations. (ii) With a probe to the long terminal repeat, we detected, in addition to a heterogeneous size distribution of subgenomic plus-strand DNA species, at least three prominent discrete size classes. Each of these discrete species had its own specific initiation site, but all had the same termination site. Such species were analogous to those reported by Kung et al. (J. Virol. 37: 127-138, 1981). (iii) These discrete size classes of plus-strand DNA were present not only on the major size class of linear DNA but also on a heterogeneous of slower-sedimenting species, which we have called immature linears. Our interpretation is that we have thus detected several additional sites for the initiation of plus-strand DNA. (iv) The 340-base plus-strand strong-stop DNA was only found associated with the immature linears. (v) From a size and hybridization comparison of these discrete size classes of plus-strand DNA with minus-strand DNA species, as synthesized in the endogenous reaction of melittin-disrupted virions, it was found that the putative additional initiation sites for plus-strand DNA synthesis corresponded to many of the pause sites in the synthesis of minus-strand DNA.     

Structure of a replication intermediate in the synthesis of Rous sarcoma virus DNA in vivo

Intermediates in the synthesis of Rous sarcoma virus DNA in vivo contain a short second strand of DNA (plus strong-stop DNA) synthesized by using the region near the 5' end of the first (minus) strand of DNA as the template. In this report, we show that the 3' end of plus strong-stop DNA is extended about 15 to 20 nucleotides beyond the 5' end of the minus-strand DNA template, probably copying a portion of the tRNATrp molecule that serves as primer for synthesis of the minus strand of DNA. The extra sequences present in plus strong-stop DNA may play a central role in the generation of the long terminal repeat present in mature forms of viral DNA.     

RNA-directed DNA polymerase of Rous sarcoma virus: initiation of synthesis with 70 S viral RNA as template

DNA synthesis by the RNA-directed DNA polymerase of Rous sarcoma virus with 70 S viral RNA as template initiates by the covalent attachment of dAMP to the 3′ terminal adenosine of an RNA molecule. Initiation continues throughout the course of a 90-minute enzymatic reaction, and chain propagation occurs on most if not all of the dAMP residues attached to primer RNA. The nature of the primer molecules was established in two ways. First, the RNA was tagged by attachment of radioactive mono- and oligodeoxynucleotides. Second, primers were isolated directly from their covalent complexes with nascent DNA. The results of both procedures indicate that DNA synthesis initiates on the 3′ termini of 4 S RNA molecules hydrogen-bonded to 70 S RNA. Purified primer RNA has a nucleotide composition (G + C = 64%) different from that (G + C = 60%) of other 4 S RNAs found hydrogen-bonded to the 70 S RNA of Rous sarcoma virus.     

Appearance of virus-specific DNA in mammalian cells following transformation by Rous sarcoma virus

To detect Rous sarcoma virus-specific DNA in mammalian cells, we have measured the capacity of unlabeled cell DNA to accelerate the reassociation of labeled double-stranded DNA synthesized by the Rous sarcoma virus RNA directed DNA polymerase. Two populations of double-stranded polymerase products are identified by their reassociation kinetics and represent approximately 5% and 30% of the viral 70 S RNA genome. Using two strains of Rous sarcoma virus and four lines of transformed mammalian cells, we found two copies of DNA homologous to both DNA populations in Rous sarcoma virustransformed rat and mouse cells, but not in normal cells. The Rous sarcoma viruslike DNA can be demonstrated in the non-repeated fraction of transformed cell DNA and in nuclear DNA. The results are supported by evidence that the techniques employed detect the formation of extensive well-matched duplexes of cell DNA and viral polymerase products.     

Deoxyribonucleic Acid Polymerase of Rous Sarcoma Virus: Studies on the Mechanism of Double-Stranded Deoxyribonucleic Acid Synthesis

The deoxyribonucleic acid (DNA) polymerase of Rous sarcoma virus synthesizes both single- and double-stranded DNA, utilizing the ribonucleic acid (RNA) of the viral genome as the initial template. Results of pulse-chase experiments indicate that the single-stranded DNA serves as unconserved template and precursor for the synthesis of double-stranded DNA. The latter reaction is apparently initiated in association with the viral RNA and may involve a partially double-stranded intermediate form.     

RNA-DNA Covalent Bonds Between the RNA Primers and the DNA Products Formed by RNA Tumor Virus DNA Polymerase

Initiation of DNA synthesis by endogenous RNA primer molecules was studied with three different RNA tumor viruses. The influence of the method of virus disruption on the observed RNA-DNA bonds was ascertained. Ether disrupted virions of both murine leukemia virus (MuLV) and the B77 strain of avian sarcoma virus (B77 virus) have rC-dC and rA-dA covalent linkages between RNA primers and newly synthesized DNA. None of the 14 other possible bonds were formed. Ether-disrupted virions of avian myeloblastosis virus (AMV) have rU-dC and rA-dA linkages. In contrast, work reported herein and from other laboratories shows that Nonidet P-40 (NP-40)-disrupted virions of all three viruses have only the rA-dA junction. Studies with virus particles which were first disrupted with ether and then treated with NP-40 indicated that the detergent treatment disallowed the formation of the ribopyrimidine-dC internucleotide bond. The same transfers are found with AMV in the presence or absence of actinomycin D, where only single-stranded DNA is formed. This finding is consistent with the notion that virtually all of the significant primers have been recognized. In contrast to mature virions, transfer experiments with ether-disrupted early harvest (5 min) MuLV showed only the rC-dC bond; the rA-dA bond was absent. The short-time harvest contains a significantly higher proportion of infectious virions than 24-h harvests. Also, since the RNA from early harvest virus is appreciably more homogenous than the RNA of mature MuLV, it is concluded that the ribopyrimidine-dC linkage is the more significant initiation event from a biochemical standpoint.     

Molecular cloning and characterization of the RNA packaging-defective retrovirus SE21Q1b.

The nonconditional RNA packaging mutant SE21Q1b contains cis- and trans-acting defects which cause cellular mRNA, rather than viral genomic RNA, to be nonspecifically packaged into SE21Q1b viral particles. Using genomic libraries of the c-SE21Q1b quail cell line, we have been able to construct a molecular clone of the SE21Q1b provirus. Upon transfection into primary quail embryo fibroblasts, the SE21Q1b molecular clone is able to recapitulate the nonspecific RNA packaging phenotype of the c-SE21Q1b cell line. The RNA packaging phenotypes displayed by several SE21Q1b/avian sarcoma-leukemia virus hybrid provirus constructs have further indicated that sequences responsible for the altered RNA packaging phenotype of SE21Q1b are localized in the left third of the SE21Q1b proviral genome. DNA sequence analysis of this region has revealed that the 5' SE21Q1b deletion has removed 179 bp from the SE21Q1b left long terminal repeat and leader regions. Several differences were detected at the carboxyl terminus of the deduced SE21Q1b nucleocapsid protein sequence in comparison with that of Rous sarcoma virus PR-C. Results of site-directed oligonucleotide mutagenesis experiments indicate, however, that the presence of these residues in the nucleocapsid protein alone is not responsible for the decreased RNA packaging specificity of SE21Q1b.