 Solution

Y91F implicates the Tyr hydroxyl, while E57Q implicates the Glu carboxylic-acid chemistry. Replacing tyrosine by phenylalanine retains an aromatic side chain but removes its hydroxyl, supporting a role for Tyr91 as a hydrogen bond donor that polarizes the steroid C3 carbonyl and stabilizes an enolate-like intermediate or transition state. Its hydroxyl could participate in a proton relay, but inactivity alone does not establish that it supplies the final proton.

Replacing a glutamate residue by glutamine approximately preserves side-chain size while replacing a titratable carboxylic acid by an amide. Loss of activity, combined with the acidic activity range, supports protonated Glu57 as a carbonyl-activating acid and possible proton donor. The two residues can cooperate to stabilize oxygen charge during reduction. The inhibitor-binding structure supports local contacts; it is not a direct picture of a testosterone transition state, and an N–H contact of the aza inhibitor need not occur in testosterone.

To distinguish these chemical roles from indirect effects, compare mutant and wild-type abundance, membrane insertion and folding, cofactor binding, and steroid binding. Measure initial rates with matched enzyme amounts and include no-enzyme and inhibitor-treated wild-type controls. If binding survives but turnover fails, the evidence for a catalytic defect is stronger. The published study combines structure, mutagenesis and modeled testosterone placement, so the proposed residue roles should be distinguished from direct observation of proton trajectories.^[ref]