Chorismate aminations: partial purification of Escherichia coli PABA synthase and mechanistic comparison with...

Chorismate is converted by regiospecific amination/aromatization sequences to o-aminobenzoate and p-aminobenzoate (PABA) by anthranilate synthase (AS) and PABA synthase (PABS), respectively. We report here the first partial purification of the large subunit of Escherichia coli PABA synthase, previously reported to be quantitatively inactivated in purification attempts. The subunit encoded by the pabB gene was overexpressed from a T7 promoter and purified 9-fold to 25-30% homogeneity. The pabB subunit appears unusually sensitive to inactivation by glycerol so this cosolvent is contraindicated. The Km for chorismate is 42 microM in the ammonia-dependent conversion to PABA, and we estimate a turnover number of 2.6 min-1. A variety of chorismate analogues have been prepared and examined. Of these compounds, cycloheptadienyl analogue 11 has been found to be the most potent inhibitor of Serratia marcescens anthranilate synthase (Ki = 30 microM for an RS mixture) and of the E. coli pabB subunit of PABA synthase (Ki = 226 microM). Modifications in the substituents at C-3 [enolpyruyl ether, (R)- or (S)-lactyl ether, glycolyl ether] or C-4 (O-methyl) of chorismate lead to alternate substrates. The Vmax values for (R)- and (S)-lactyl ethers are down 10-20-fold for each enzyme, and V/K analyses show the (S)-lactyl chorismate analogue to be preferred by 12/1 over (R)-lactyl for anthranilate synthase while a 3/1 preference was observed for (R)-/(S)-lactyl analogues by PABA synthase. The glycolyl ether analogue of chorismate shows 15% Vmax vs. chorismate for anthranilate synthase but is actually a faster substrate (140%) than chorismate with PABA synthase, suggesting the elimination/aromatization step from an aminocyclohexadienyl species may be rate limiting with AS but not with PABS. Indeed, studies with (R)-lactyl analogue 14 and anthranilate synthase led to accumulation of an intermediate, isolable by high-performance liquid chromatography and characterized by NMR and UV-visible spectroscopy as 6-amino-5-[(1-carboxyethyl)oxy]-1,3-cyclohexadiene-1-carboxylic acid (17). This is the anticipated intermediate predicted by our previous work with conversion of synthetic trans-6-amino-5-[(1-carboxyethenyl)oxy]-1,3-cyclohexadiene-1-carbo xylic acid (2) to anthranilate by the enzyme. Compound 17 is quantitatively converted to anthranilate on reincubation with enzyme, but at a 1.3-10-fold lower Vmax than starting lactyl substrate 14 under the conditions investigated; the basis for this kinetic variation is not yet determined.