A thermolabile mutant of adenovirus 5 resulting from a substitutional mutation in the protein VIII gene

The mutant adenoviruses H5sub304 and H5RI(r) were isolated sequentially from adenovirus 5 wild type by selection for the loss of EcoRI restriction endonuclease sites by Jones and Shenk (Cell 13:181-188, 1978). sub304 lacks the site of 84.0 map units (m.u.), and RI(r) lacks both that and the site at 75.9 m.u. A set of derivatives of RI(r) that lack the site at 75.9 m.u. accumulated virus more slowly at 38.8 or 39.5°C than those with the site present, as measured by low-multiplicity passage or single-step replication cycles, respectively. Since the EcoRI site at 75.9 m.u. is predicted to lie in the gene encoding the precursor to virion polypeptide VIII (pVIII), the failure to accumulate virus rapidly could lie either in some step in processing and assembly of virions or in an increased virion thermolability. The latter possibility was shown to be the case, as all strains mutated at the EcoRI 75.9 m.u. site were extremely thermolabile in vitro, even at 37°C. CsC equilibrium density centrifugation of heated crude stocks of RI(r) and sub304 demonstrated that loss of infectivity in RI(r) was accompanied by physical disruption of virions. Polyacrylamide gel electrophoresis of infected cell extracts or purified virions showed that pVIII of RI(r) had an apparent molecular weight that was slightly greater than that of sub304, and mature RI(r) and sub304 virions displayed polypeptide VIIIs which appeared to be of identical molecular weights. Nucleotide sequence analysis of RI(r) demonstrated that it contained a 9-base-pair (bp) substitution for 6 bp found in sub304, leading to a loss of the EcoRI site and a predicted insertion of a single amino acid. Comparison of the sequence of sub304 with the published sequence of adenovirus 2 revealed two changes, a single transversion at bp 1,722 and a bp deletion at 1,749, leading to the loss of a TaqI site. The predicted reading frame change would lead to a stop codon at bp 1,885. This raises the question of whether adenovirus 2 and adenovirus 5 use the same reading frame for pVIII.