Active site electrostatic effects modulate DBO inhibitor potency against OXA-48 variants

The OXA-48 class D {beta}-lactamase and its variants are clinically concerning through their carbapenemase and cephalosporinase activity. Effective inhibitors are therefore urgently needed. Diazabicyclooctanes (DBOs) are newly introduced non {beta}-lactam based inhibitors, but their potency is affected by mutations, with the more recently developed nacubactam consistently less potent than avibactam, the first DBO, towards OXA-48 variants. This difference is most pronounced for OXA-48 compared to its variants OXA-163 and -405, which contain mutations in the {beta}5-{beta}6 loop and preferentially hydrolyse cephalosporins over carbapenems. Here we use multiscale simulations to uncover the origin of these differences after determining that the second step of DBO acylation is rate-limiting. Using QM/MM MD simulations we show that nacubactam exhibits a consistently higher energy barrier than avibactam for covalent inhibition of OXA-48-like enzymes, a difference attributable to DBO-specific changes in the active site electric field. Simulations of non-covalent complexes further indicate that weaker binding, due to electrostatic repulsion between the positively charged nacubactam tail and Arg214 in the {beta}5-{beta}6 loop, partially explains the lower potency of nacubactam towards OXA-48, compared to OXA-163 and -405. Combining computational predictions of covalent reaction rates and non-covalent binding affinities accurately reproduces the relative order of experimental IC50 values, and predicts potency towards OXA-517, a recently discovered OXA-48 variant capable of hydrolysing many antibiotics. This work provides mechanistic insight into the inhibition of class D {beta}-lactamases and reveals the origins of differential DBO potency towards OXA-48 variants, highlighting how subtle differences between enzymes modulate inhibitory effects.