We re-examined the generally accepted concept that replication of Rous sarcoma virus (RSV) requires host DNA synthesis. We used terminally differentiated chicken myotubes as the host because chromosomal DNA replication is completely abolished by the natural differentiation process. Southern blot analysis detected unintegrated viral DNA in both the nucleus and cytoplasm of infected myotubes. This indicated that reverse transcription of the infecting viral RNA and transport of the newly synthesized viral DNA into the nucleus proceeded normally in myotubes. However, restriction enzyme digestion of high Mr DNA prepared from infected myotubes produced none of the fragments specific for RSV, indicating that the viral DNA had failed to integrate into the myotube chromosomal DNA. In these infected myotubes, viral RNA was detected by in situ hybridization. Northern blot analysis showed the presence of all three RSV mRNAs (38S, 28S and 21S). The amount of these viral RNAs in infected myotubes was comparable with that found in infected fibroblasts. We conclude that host DNA synthesis is required for RSV integration, but, in contrast to the generally accepted concept, viral DNA integration is not an absolute requirement for transcription of the RSV genome.

We studied Rous sarcoma virus (RSV) protein synthesis in RSV-infected, terminally differentiated chicken myotubes ('late-infected' myotubes), in which no viral DNA integration takes place but all three viral mRNAs (38S, 28S and 21 S) are transcribed normally. With the use of specific anti-RSV protein antisera, we found that only the viral gag and pol proteins were synthesized at levels similar to those synthesized in RSVtransformed fibroblasts; the synthesis of env and v-src proteins was significantly reduced in these infected myotubes. We concluded that the viral RNA transcribed from the unintegrated RSV DNA was functional but that genes at the 3' end of the RSV genome were translated at a lower level. By contrast, when mononucleated replicating chicken myoblasts were infected with a mutant (tsNY68) carrying a temperature-sensitive v-src gene and maintained at the nonpermissive temperature for this gene, they developed into myotubes with viral DNA integrated in their chromosomal DNA. These 'early-infected' myotubes expressed all four viral genes (gag, pol, env and v-src) at a level similar to that in infected fibroblasts. This result ruled out the possible presence of specific factor(s) in myotubes that preferentially inhibit the 3' genes of RSV, and suggested other translational control(s) of viral gene expression in late-infected myotubes.