An activity–resistance trade-off constrains enzyme evolution

Abstract

The presence of self-resistance genes in antibiotic-producing organisms poses a paradox: How can resistance evolve before the antibiotic exists, and how can an antibiotic producer arise without first evolving resistance? Here, we examine the evolutionary origins of self-resistance to mycophenolic acid (MPA), an inhibitor of inosine monophosphate dehydrogenase (IMPDH). The MPA biosynthetic gene cluster (BGC) includes a resistant IMPDH-B. Homologs of IMPDH-B occur not only in MPA producers but also in many nonproducing fungi, where remnants of the MPA BGC remain detectable. The phylogeny of IMPDH-B is incongruent with the fungal species tree, consistent with multiple horizontal gene transfer events between Aspergillus and Sordariomycetes. We characterized eleven extant IMPDH-Bs, five from MPA producers and six from nonproducers, along with seven resurrected ancestral enzymes (Anc1–Anc7). MPA resistance appeared between Anc2 and Anc3 and coincided with a loss of catalytic efficiency. Across both ancestral and extant enzymes, MPA resistance correlated strongly with reduced activity, revealing a robust activity–resistance trade-off that has persisted for millions of years. Unexpectedly, both the IMPDH-Bs and ancestral enzymes Anc3–Anc7 were also resistant to ribavirin-5′-monophosphate (RVP), an IMP-competitive inhibitor. Because MPA and RVP bind to similar enzyme conformations, the activity–resistance trade-off may reflect a design constraint imposed by the need to maintain resistance to multiple inhibitors. Intriguingly, although Anc1 and Anc2 are equally sensitive to MPA, Anc2 shows reduced susceptibility to RVP. This pattern suggests that preexisting resistance to another IMPDH inhibitor may have created a permissive background for the later evolution of MPA biosynthesis.