Narayan, R.; Le, C. C.; Khurana, J. K.; Nieto, V.; Olson, C. A.; Turnbaugh, P. J.; Balskus, E. P.

Microorganisms in the human gut influence the efficacy and metabolism of host-targeted small molecule therapeutics, including the frontline Parkinson's disease drug levodopa (L-dopa). Previous work has identified a mechanism-based inhibitor of gut bacterial decarboxylases that degrade L-dopa, -fluoromethyltyrosine (AFMT). However, early experiments with AFMT in rodent models suggested undesirable in vivo metabolism by host tyrosine hydroxylase, producing a metabolite likely to worsen Parkinson's phenotypes and prevent application as an L-dopa co-treatment. Here, we demonstrate oxidation of AFMT in vitro by recombinant human tyrosine hydroxylase. We then develop AFMT analogs that retain activity against bacterial decarboxylases but have reduced susceptibility to host hydroxylation. Suitable arenes for inhibitor design were identified using assays with commercially available noncanonical amino acids, which revealed aryl difluorination as a promising modification. Difluoroaryl AFMT derivatives are less prone to degradation by tyrosine hydroxylase in vitro yet still inhibit L-dopa metabolism by bacterial decarboxylases. This work exemplifies how substrate reactivity can streamline design of mechanism-based enzyme inhibitors, as well as how constraints posed by the host can be incorporated during development of microbiome-targeted therapeutics. The compounds reported here are promising starting points for future studies in animal models and further exploration of gut bacterial effects on L-dopa treatment efficacy.