Common and unique features of T antigens encoded by the polyomavirus group

Although 12 different members of the polyomavirus group have now been identified, only SV40 and PyV have been studied extensively. Whereas each member of the group shows a restricted host range, viruses infecting species from birds to humans have been reported. Although little is known concerning the biology of natural infections in the wild, it is apparent that these viruses exhibit various cell-type tropisms. Some viruses, such as LPV (B lymphocytes) or KV (pulmonary endothelium), are tightly restricted to specific cell types, while others, such as PyV, infect a variety of tissues in the animal. Despite these differences, all polyomaviruses share a common strategy of productive infection, expressing T antigens which act both on cellular targets, preparing cellular metabolism for supporting optimal viral replication, and then on targets within the viral genome, to regulate viral DNA replication, transcription, and assembly. Presumably, this common replication strategy restricts the degree to which the sequences of these viruses can diverge. Thus, sequence motifs conserved among these different viruses may indicate key structural elements essential for biochemical function. In this article I have compared the sequences of all polyomavirus-encoded large and small T antigens sequenced to date. This has led to the following conclusions and speculations. (i) Comparison of the domain organization of different large T antigens reveals that these proteins fall into two structural classes. Members of the SV40 class, which include SV40, JCV, BKV, and SA12, possess a carboxyl-terminal domain, which in SV40 has been shown to be dispensable for viral DNA replication but essential for virion assembly. The PyV class lacks the carboxyl-terminal domain and carries additional amino acids within the amino-terminal domain. When total amino acid identity is examined, members of the SV40 class show the highest degree of conservation (65 to 85%), while sequence identity among the remaining viruses varies from 18 to 55%. (ii) The DNA binding domains of most large T antigens are closely related, with amino acid identities ranging from 35 to 86%. Several residues within this domain are invariant among all T antigens. All of these viruses have multiple copies of the consensus T-antigen-binding pentanucleotide (GAGGC) in their ori region, suggesting that all T antigens recognize this sequence. The single exception is the large T antigen encoded by the avian virus BFDV. The putative DNA binding domain of this protein shows little or no sequence relation to that of other T antigens. Furthermore, the GAGGC motif is not found in the ori region of this virus.(ABSTRACT TRUNCATED AT 400 WORDS)