The dimeric nature of the junction-resolving enzyme T7 endonuclease I.

Abstract

The 149 amino acid residue product of bacteriophage T7 gene 3, endonuclease I, has been shown to introduce cleavages into branched DNA species such as four-way junctions [l-31. In the presence of divalent metal ions such as magnesium, the four-way DNA junction folds into the stacked X-structure [4]. Recognition of DNA junctions by T7 endonuclease I and other junction resolving enzymes is structure-specific and involves a manipulation of the structure of the junction. This is observable as change in the global configuration of the junction arms in comparative gel electrophoresis [5-81. The junction resolving enzymes introduce cleavages into opposing strands within the junction, bringing about a productive resolution. T4 endonuclease VII [6] and yeast CCEl [7] have been shown to interact with four-way junctions as dimers. They also behave as weakly interacting dimers in free solution, exhibiting fast exchange between subunits. Mutants of T7 endonuclease I have been isolated that are completely inactive as nucleases while retaining their full structural selectivity of binding to junctions [8]. One of these mutants, E65K, has been expressed as an N-terminal fusion with protein A. The protein A portion of the fusion protein can be removed by digestion with the protease Factor Xa. The fusion protein and the fusion-released protein have different mobilties in gel electrophoresis when bound to radioactively-labelled junction and this has been exploited to investigate the quaternary structure of T7 endonuclease I. The fusion protein was removed from T7 endonuclease I E65K by digestion with increasing concentrations of Factor Xa and this material was bound to junction before being electrophoresed in a native polyacrylamide gel (figure 1). Three distinct species could be detected, suggesting that the protein binds to junction as a dimer. The slowest species corresponds to a homodimer of fusion protein bound to junction and the fastest species to a homodimer of unfused endonuclease I E65K bound to junction. The intermediate species is likely to be a complex of the DNA junction and a heterodimer of one cleaved and one uncleaved fusion protein. In a similar experiment, protein A-endonuclease I E65K and oligohistidine E65K were mixed for differing lengths of time, bound to junction and then electrophoresed in a native polyacrylamide gel. Only species corresponding to homodimers of the two fusion proteins were observed; no intermediate species corresponding to a heterodimer was observed even after extended incubation times. It is clear that, unlike T4 endonuclease VII and yeast CCEI, T7 endonuclease I forms a very strong dimer in free solution. It is possible to bring about an exchange of subunits by treatment with a high concentration of a denaturant. Equimolar amounts of protein A-endonuclease I E65K and oligohistidine-endonuclease I were mixed and denatured in 6M urea, which was then removed slowly by extensive dialysis. When this protein was incubated with junction and electrophoresed in a native polyacrylamide gel, an intermediate species corresponding to a heterodimer of the two fusion proteins could be obsewed. This material also retained its ability to cleave junctions -I