High-Order Epistasis in Catalytic Power of Dihydrofolate Reductase Gives Rise to a Rugged Fitness Landscape in the Presence of Trimethoprim Selection

Abstract

Evolutionary fitness landscapes of several antibiotic target proteins have been comprehensively mapped showing strong high-order epistasis between mutations, but understanding these effects at the biochemical and structural levels remained open. Here, we carried out an extensive experimental and computational study to quantitatively understand the evolutionary dynamics of Escherichia coli dihydrofolate reductase (DHFR) enzyme in the presence of trimethoprim-induced selection. To facilitate this, we developed a new in vitro assay for rapidly characterizing DHFR steady-state kinetics. Biochemical and structural characterization of resistance-conferring mutations targeting a total of ten residues spanning the substrate binding pocket of DHFR revealed distinct changes in the catalytic efficiencies of mutated DHFR enzymes. Next, we measured biochemical parameters ( K m , K i , and k cat ) for a mutant library carrying all possible combinations of six resistance-conferring DHFR mutations and quantified epistatic interactions between them. We found that the high-order epistasis in catalytic power of DHFR ( k cat and K m ) creates a rugged fitness landscape under trimethoprim selection. Taken together, our data provide a concrete illustration of how epistatic coupling at the level of biochemical parameters can give rise to complex fitness landscapes, and suggest new strategies for developing mutant specific inhibitors.

Machine learning prediction of resistance to sub-inhibitory antimicrobial concentrations fromEscherichia coligenomes

“…Thus, lineages only carrying dfrI achieved 99% of the cell yield of those containing both dfrI and dfrV, as compared to 3% for lineages containing dfrV alone (Figure S5). This reflects a slight negative epistasis, consistent with the diminishing return of beneficial genes and variants reported in E. coli lab strains (44). The presence of chloramphenicol acetyltransferase gene, cat, improved the cell yield by 0, 5% an 12% of the total population range in low, moderate and high concentrations of chloramphenicol but only marginally reduced the cell doubling time (0.02% in the highest concentrations) (Figure 4).…”

Adaptation to mutational inactivation of an essential gene converges to an accessible suboptimal fitness peak

“…Such scenario is akin to antibiotic-mediated target inhibition, but distinct in that there is no selection for mutations that affect interactions with the drug, such as target modification and drug efflux mechanisms. For example, when treated with DHFR inhibitor trimethoprim, bacterial cells recurrently acquire resistance by high-fitness mutations near the active site of the target protein that decrease drug binding affinity, even at the expense of catalytic efficiency (Rodrigues et al, 2016; Tamer et al, 2019; Toprak et al, 2012). In contrast, a genetic perturbation in the same target gene provides the opportunity to study other potential unexplored mechanisms of adaptation in the absence of drug and evaluate their impact in the context of antibiotic resistance.…”

“…Here we study evolutionary adaptation upon functional inactivation of dihydrofolate reductase (DHFR), an essential E. coli enzyme. As the cellular target of the antibiotic trimethoprim, DHFR has been repeatedly observed to accumulate mutations leading to extremely high drug resistance levels in various experimental evolution studies (Rodrigues et al, 2016; Tamer et al, 2019; Toprak et al, 2012). Past efforts to link fitness effects of chromosomal variation in the folA locus encoding DHFR and their biophysical effects on DHFR protein allowed us to develop an accurate quantitative biophysical model of DHFR fitness landscape (Bershtein et al, 2015a; Bershtein et al, 2013; Bershtein et al, 2012; Bershtein et al, 2015b; Rodrigues et al, 2016).…”

A trimethoprim derivative impedes antibiotic resistance evolution

“…Strikingly, we found that evolution of TMP resistance in E. coli consistently proceeds through stepwise acquisition of multiple resistance-conferring mutations in the folA gene that encodes for DHFR (Fig. 1c ) 7 , 16 , 18 – 20 . Antibiotic sensitive E. coli populations evolve nearly four orders of magnitude higher TMP resistance by accumulating three to five mutations in folA .…”

“…Evolution of antibiotic resistance has been studied at the molecular level for decades with the ultimate goal of devising targeted therapies to impede the evolution of resistance. Targeting evolutionarily common resistance-conferring mutations was previously proposed as a promising strategy to impede evolution of resistance based on computer simulations 5 – 7 . However, to the best of our knowledge, there has been no biological validation of this strategy.…”