INFORMATIONAL SUPPRESSION TO INVESTIGATE STRUCTURAL-FUNCTIONAL AND EVOLUTIONARY ASPECTS OF THE ERWINIA-CHRYSANTHEMI CELLULASE EGZ

The cellulase EGZ produced by the plant pathogen Erwinia chrysanthemi belongs to family 5 of the beta-glycohydrolases (also referred to as c ellulase family A), which contains over 40 members from Gram-negative and Gram-positive bacteria and fungi. Amber mutations were introduced into 16 codons of the celZ gene encoding EGZ. Targeted residues includ ed: (1) two Glu, two His and one Arg residue, strictly conserved throu ghout family 5; (2) one Arg and one His residue conserved in sub-famil y 5-2; and (3) one His and six Arg residues not conserved at all. Each amber allele was introduced into 13 Escherichia coli strains each car rying a different suppresser tRNA that inserts an amino acid at the mu tated position. In vivo stability of the mutated forms of EGZ and thei r cellulase activity were analysed as well as suppression efficiency F or some positions of particular interest, missense mutations were intr oduced into the celZ gene either to confirm the effect of the suppress or-mediated amino acid substitution or to broaden the spectrum of muta tions available. The substitution patterns of the two Glu positions we re interpretable in the light of the stereospecificity of the reaction catalysed by EGZ: Glu133 and Glu220 are proposed to act as a proton d onor and as a nucleophile, respectively, forming the glycosyl-enzyme i ntermediate. Substitution at His-occupied positions, including two non -conserved positions, yielded proteins affected in their catalytic act ivity but not their in vivo stability; In particular, evidence was obt ained for His at position 98 to be involved in interactions with the s ubstrate. The view that Arg residues are important in stabilizing prot eins was supported by the identification of three Arg residues, whose substitution yielded thermosensitive forms of EGZ. In addition, Pro su bstitutions of any of the six Arg residues altered protein stability i n vivo but the substitutions scored almost neutral for activity. Five positions, predicted to be within alpha-helices, were found to be susc eptible to Pro substitutions (but not to Ala) with respect to stabilit y in vivo. Overall, the systematic alteration of all His and Arg resid ues coupled with the simultaneous analysis of activity and in vivo sta bility allowed us to demonstrate that substitution matrices vary at ea ch position and for each biological property considered. Ideally, ther efore, substitution matrices used in sequence alignment procedures sho uld be reconsidered as position-specific and as property-specific. All owed amino acid substitutions at conserved and non-conserved positions were compared. with natural variability occurring in the set of 40 ce llulases present in family 5, thereby providing us with a direct compa rison between laboratory-induced and naturally occurring evolution and between experimentally and theoretically based substitution matrices.