Identification of putative active-site residues in the DNase domain of colicin E9 by random mutagenesis

We have used random mutagenesis to identify putative active-site residues in the C-terminal cytotoxic endonuclease domain of the bacterial toxin colicin E9. Six single-site mutations in the DNase domain were isolated which destroyed the toxic action of the colicin. DNA sequencing identified the mutations as Gly460Asp, Arg544Gly, Glu548Gly, Thr571Ile, His575Tyr and His579Tyr. All six wild-type residues are highly conserved in the DNase domains of both the E group colicins and the closely related pyocins. Site-directed mutagenesis was then used to substitute the wild-type amino acid residue at each of these positions for an alanine residue in order to distinguish important from unimportant sites. Two of the six alanine-mutant colicins (Gly460Ala and His579Ala) exhibited significant in vivo activity, unlike the original mutation of these residues, and were therefore not characterised further. The Thr571Ala mutant colicin, although not inactive, was significantly less active than the control. The other three alanine mutants (Arg544Ala, Glu548Ala and His575Ala remained completely inactive in the in vivo tests. Each 15 kDa alanine-mutant DNase domain was overexpressed and purified using a tandem-expression strategy which relies on the enzyme being able to bind to the natural inhibitor, Im9. Tryptophan emission spectra of the alanine mutants showed significant alterations in the emission maxima of all but the His575Ala mutant, suggesting changes in the tertiary structure of these mutant proteins. Activity measurements, using the spectrophotometric Kunitz assay, indicated that the Thr571Ala mutant was partially active as an endonuclease but the remaining alanine mutants were all completely inactive. All four mutant proteins, however, retained their ability to bind DNA in a gel shift assay, suggesting the mutations affect catalytic rather than substrate-binding residues. Searching the sequence databases for possible homology to other DNA-binding proteins revealed a significant match between residues 464 to 487 of the E9 DNase domain and helix IV of the POU domain of eukaryotic transcription factors. (C) 1996 Academic Press Limited