Hypothetical three-dimensional models for the entire polymerase domain of HIV-1 reverse transcriptase (HIV RT) and conserved regions of HSV-1 DNA polymerase (HSV pol) were created, primarily from literature data on mutations and principles of protein structure, and compared with those of E. coli DNA polymerase I (E. coli pol I). The corresponding parts, performing similar functions, were found to be analogous, not homologous, in structure with different β topologies and sequential arrangement. The polymerase domain of HSV pol is shown to form an anti-parallel β-sheet with α-helices, but with a topology different from that of the Klenow fragment of E. coli pol I. The main part of the polymerase domain of HIV RT is made up of a basically parallel β-sheet and α-helices with a topology similar to the nucleotide-binding p21 ras proteins. The putative functions of some conserved or invariant amino acids in the three polymerase families are discussed.
Genetic evidence for two protein domains and a potential new activity in bacteriophage T4 DNA polymerase.