Fig. 5
Mechanistic insights based on non-observed potential CMPS products. a Retro-catalytic analyses of non-observed potential products showing proposed requisite precursors, i.e. (E)/(Z)-enolate intermediates and the 4-substituted-l-P5C derivative. A common feature in potential production of the non-observed products is the presence of an (E)-enolate and (4R)-4-methyl-l-P5C. b View from a CarB structure with the modelled (E)-propionyl-CoA enolate, resulting from the decarboxylation of (2R)-methylmalonyl-CoA25, and (4 R)-4-methyl-l-P5C. The model implies proximity between the methyl group of the (E)-enolate and that of (4R)-4-methyl-l-P5C (~2 Å) suggesting a steric clash. The combination of an (E)-enolate and (4R)-4-methyl-l-P5C may thus be disfavoured, consistent with the lack of the potential products in a; c model of CarB M108V, with the (E)/(Z)-propionyl-CoA enolate (the methyl group of the (E)-enolate is in pink for distinction), and l-P5C. The model implies proximity between the methyl of the (E)-enolate and the β-methyl of valine-108 (~1 Å); the distance between the methyl of the (4R)-4-methyl-l-P5C (not shown for clarity) and the β-methyl of the valine residue is modelled at ~3 Å. Both these interactions thus may manifest a steric clash. The CarB M108V/I variants may thus preferentially catalyse formation of t-CMP derivatives with the (4S,6R)-stereochemistry (Fig. 4, entries in a green box)
