Dihydroxyacetone phosphate aldolase catalyzed synthesis of structurally diverse polyhydroxylated pyrrolidine derivatives and evaluation of their glycosidase inhibitory properties.

The chemoenzymatic synthesis of a collection of pyrrolidine-type iminosugars generated by the aldol addition of dihydroxyacetone phosphate (DHAP) to C-alpha-substituted N-Cbz-2-aminoaldehydes derivatives, catalyzed by DHAP aldolases is reported. L-fuculose-1-phosphate aldolase (FucA) and L-rhamnulose-1-phosphate aldolase (RhuA) from E. coli were used as biocatalysts to generate configurational diversity on the iminosugars. Alkyl linear substitutions at C-alpha were well tolerated by FucA catalyst (i.e., 40-70 % conversions to aldol adduct), whereas no product was observed with C-alpha-alkyl branched substitutions, except for dimethyl and benzyl substitutions (20 %). RhuA was the most versatile biocatalyst: C-alpha-alkyl linear groups gave the highest conversions to aldol adducts (60-99 %), while the C-alpha-alkyl branched ones gave moderate to good conversions (50-80 %), with the exception of dimethyl and benzyl substituents (20 %). FucA was the most stereoselective biocatalyst (90-100 % anti (3R,4R) adduct). RhuA was highly stereoselective with (S)-N-Cbz-2-aminoaldehydes (90-100 % syn (i.e., 3R,4S) adduct), whereas those with R configuration gave mixtures of anti/syn adducts. For iPr and iBu substituents, RhuA furnished the anti adduct (i.e., FucA stereochemistry) with high stereoselectivity. Molecular models of aldol products with iPr and iBu substituents and as complexes with the RhuA active site suggest that the anti adducts could be kinetically preferred, while the syn adducts would be the equilibrium products. The polyhydroxylated pyrrolidines generated were tested as inhibitors against seven glycosidases. Among them, good inhibitors of alpha-L-fucosidase (IC50=1-20 microM), moderate of alpha-L-rhamnosidase (IC50=7-150 microM), and weak of alpha-D-mannosidase (IC50=80-400 microM) were identified. The apparent inhibition constant values (Ki) were calculated for the most relevant inhibitors and computational docking studies were performed to understand both their binding capacity and the mode of interaction with the glycosidases.