Initiation of translation of poliovirus RNA by ribosomal entry into an internal segment of the 742-nucleotide (nt)-long 5' nontranslated region involves transacting factors, including p57, a 57-kDa polypeptide which has been identified as the pyrimidine tract-binding protein (PTB). A UV cross-linking assay was used to compare the RNA-binding properties of the p57 present in various mammalian cytoplasmic extracts with those of purified murine p57 and recombinant human PTB. Three noncontiguous p57-binding sites were located within the poliovirus 5' nontranslated region, between nt 70 and 288, and 443 and 539 (domain V), and 630 and 730. With the same assay, a novel 34-kDa polypeptide was identified that bound nt 1 to 629 specifically. A single A-G substitution of nt 480 which attenuates poliovirus did not alter UV cross-linking of p57 to domain V. Although UV cross-linking of p57 to the internal ribosome entry site was specifically reduced by competition with poly(U) but not by competition with poly(C), poly(G), and poly(A) homoribopolymers, the presence of a polyuridine tract was not a sufficient determinant for binding of RNA to the p57 present in cytoplasmic extracts, nor was the polypyrimidine tract downstream of domain V necessary for binding to this site.

Infectious retrovirus particles are derived from structural polyproteins which are cleaved by the viral proteinase (PR) during virion morphogenesis. Besides cleaving viral polyproteins, which is essential for infectivity, PR of human immunodeficiency virus (HIV) also cleaves cellular proteins and PR expression causes a pronounced cytotoxic effect. Retroviral PRs are aspartic proteases and contain two copies of the triplet Asp-Thr-Gly in the active center with the threonine adjacent to the catalytic aspartic acid presumed to have an important structural role. We have changed this threonine in HIV type 1 PR to a serine. The purified mutant enzyme had an approximately 5-to 10-fold lower activity against HIV type 1 polyprotein and peptide substrates compared with the wild-type enzyme. It did not induce toxicity on bacterial expression and yielded significantly reduced cleavage of cytoskeletal proteins in vitro. Cleavage of vimentin in mutant-infected T-cell lines was also markedly reduced. Mutant virus did, however, elicit productive infection of several T-cell lines and of primary human lymphocytes with no significant difference in polyprotein cleavage and with similar infection kinetics and titer compared with wild-type virus. The discrepancy between reduced processing in vitro and normal virion maturation can be explained by the observation that reduced activity was due to an increase in K m which may not be relevant at the high substrate concentration in the virus particle. This mutation enables us therefore to dissociate the essential function of PR in viral maturation from its cytotoxic effect.