HisA mutants with minor structural differences display major functional deviations

Even though enzymes tend to specialize on one reaction during evolution, enzyme promiscuity is an abundant phenomenon. The subject of this thesis is the Salmonella enterica N’-[(5’-phosphosoribosyl)-formimino]-5-aminoimidazole-4 carboxamide-ribonucleotide (ProFAR) isomerase (SeHisA), a (βα)8-barrel enzyme from the histidine biosynthesis that catalyzes one reaction on one substrate in one organism. In Actinobacteria HisA has evolved to a bifunctional enzyme called phosphoribosyl isomerase A (PriA): it is capable of catalyzing the reaction normally done by the N’-(5’-phosphoribosyl) anthranilate (PRA) isomerase (TrpF) as well. The functional plasticity of PriA is possible due to the common reaction mechanism of HisA and TrpF, called Amadori rearrangement. The Amadori rearrangement is an acid-base catalyzed isomerization reaction where the aminoaldose (ProFAR or PRA) is converted into the corresponding ketose (PRFAR or CdRP). A SeHisA mutant with a glutamine to arginine mutation in position 18 (Q18R) shows a detectable TrpF activity, whereas another mutant with a duplication of residues from 13 to 15 (dup13-15) loses its HisA activity and gains TrpF activity. My first aim was to improve the TrpF activity of the Q18R mutant. A G79S mutation was introduced inspired by PriA. Proteins were purified and crystallized. In order to gain complex structures with either the TrpF reaction product analogue reduced CdRP (rCdRP) or ProFAR, co-crystallization and soaking were done. ProFAR is not commercially available and had to be synthetized and purified. In vitro TrpF activities of the Q18R and Q18R/G79S mutants were measured. My second aim was to compare the Q18R mutant with the dup13-15 mutant, since there is very little structural difference between them, but they show high difference in catalytic activity. Mutants, which would bridge the functional gap between them, were designed and by using lambda red recombineering were introduced into a Salmonella typhimurium genome. In vivo growth rate was measured and relative fitness was calculated for each mutant in respect to their HisA and TrpF activity. HisA mutants Q18R and Q18R/G79S showed very poor TrpF activity in in vitro assays. No dissociation constants could be measured for either of the mutants using microscale thermophoresis, and a very low kcat/KM value (~2 s-1M-1) with a high error rate could be determined for Q18R/G79S. Complex structures of Q18R and Q18R/G79S mutants with ProFAR were solved at a 2.47 Å and a 1.78 Å, respectively, from soaked crystals. No structure with rCdRP was obtained. Growth rate measurements in comparison with a strain with wild type HisA and TrpF, gave striking results pointing out the important role of the residue in the position 16 when three residues are inserted after the arginine in position 18. A leucine in position 16 yielded wild type HisA activity (94%) and poor TrpF activity (0-5%), with a valine, no HisA activity (0%) and a moderate TrpF activity (38-46%) were found compared to the wild type. These results pointed out how small the barrier between a specialist, a promiscuous and a bifunctional enzyme can be. SeHisA, being a specialized enzyme can easily be modified in order to gain TrpF function, and as I have showed in the present study, a single methyl group (the difference between a leucine and a valine) can turn the activity of an enzyme inside out.