Retroviral integrase: conserved sequence information as a guide to understanding structure and function of the retroviral and bacterial IS3 DD(35)E transposases

Abstract

The retroelement integrases (IN) are one of the three enzymes required for the retroviral life cycle. Much attention has been focused on retroviruses as a result of the AIDS epidemic and IN became a target for the rational development of anti-AIDS drug therapies. However two things that would help this effort, the mechanism of the chemical reaction catalysed by IN and the structure were initially unknown. This review shows how sequence alignments helped design and interpret experiments which showed that the conserved DDE residues were catalytically required, and which showed the catalytic function of three acidic residues in a motif here called the DD(35)E motif, in turn leading to the solution of its structure by xray crystallography; and finding the DNA binding function of the C-terminus in turn leading to the solution of its structure by NMR. Sequence alignments showed that IN belonged to a greater family of transposases found in retroelements and IS3 bacterial insertion sequences, whose function was to transpose a piece of dsDNA with the consensus sequence CA at the 3′-ends. IN was composed of three separate motifs. The central region of about 110 residues which very similar in sequence and which shared 3 conserved acidic groups (D,D and E). The 2nd D and the E were always separated by 35 residues, and together with the first D, were used to name this family the DD(35)E transposases. It was proposed that the DD(35)E region was an autonomously folding domain; had the catalytic site which recognised the CA 3′-ends; and that these 3 residues were involved in DNA bond breaking/making and divalent metal ion binding. The N-terminal part of IN contained a conserved HHCC motif, presumably a Zn finger motif, which was not found in the other DD(35)E transposases. Aside from the HHCC residues, this part of the enzymes did not align well suggesting that if the structure was conserved in the family, then the binding of the Zn atom would have to be the overriding factor. The C-terminus was very variable not able to be aligned within the DD(35)E transposase family, except for very closely related enzymes. The C-terminus for another transposase of similar size (γ–δres, Abdel-Meguid et al., Proc. Nat. Acad. Sci. USA 81, (1984) 2001–2005) was known to bind DNA. However no DNA binding motifs such as helix–turn–helix were found by computer assisted searches in the C-terminus of the IN family and no function was able to be assigned to it. It seemed likely that IN was a 3 domain protein and it was reasonable to explore the function of IN in domains in the same manner as had be pioneered with res.