Channel matters: Overcoming diffusion bottlenecks via loop engineering of LinD for enhanced isoprene production

The selective chemical dehydration leading to CC double bond formation is a challenging reaction that harbors great potential for industrial applications. The cofactor independent bifunctional linalool dehydratase isomerase (LinD) from Castellaniella defragrans catalyzes the reversible dehydration of (S)-linalool to myrcene, as well as its isomerization to geraniol. We previously reported that LinD is able to convert the small alkenol 2-methyl-3-buten-2-ol to the valuable product isoprene. To foster the LinD-catalyzed production of isoprene in a novel recombinant E. coli whole-cell two-phase system, we targeted the active site and a flexible α-helix near the putative substrate channel via enzyme engineering. Interestingly, none of the active site variants exhibited an increased product formation. In contrast, saturation mutagenesis of the 10 amino acids forming the α-helix, identified the variants K103N, R104G, G107T and D112T, which exhibited a 1.73 ± 0.05, 1.56 ± 0.12, 2.08 ± 0.12 and 1.93 ± 0.06-fold increase in product formation compared to the wild-type enzyme, respectively. Notably, a combinatorial approach targeting these four variants led to decreased activity in all cases, compared to the corresponding single-point variants, indicating negative epistatic interactions. Thus, employing the most catalytically efficient single point variant G107T, which exhibited a 28-fold higher k_{cat (app)} compared to the wild-type, a total of 2.8 ± 0.2 mM isoprene was obtained utilizing the whole-cell two-phase system. Crystallographic analysis of G107T revealed only minor structural changes; however, molecular dynamic simulations uncovered striking conformational differences relative to the LinD wild-type, emphasizing the role of altered substrate channel in variant G107T.