A Two-Enzyme Adaptive Unit within Bacterial Folate Metabolism

Abstract

Enzyme function and evolution are influenced by the larger context of a metabolic pathway. Deleterious mutations or perturbations in one enzyme can often be compensated by mutations to others. We used comparative genomics and experiments to examine evolutionary interactions with the essential metabolic enzyme dihydrofolate reductase (DHFR). Analyses of synteny and co-occurrence across bacterial species indicate that DHFR is coupled to thymidylate synthase (TYMS) but relatively independent from the rest of folate metabolism. Using quantitative growth rate measurements and forward evolution in Escherichia coli, we demonstrate that the two enzymes adapt as a relatively independent unit in response to antibiotic stress. Metabolomic profiling revealed that TYMS activity must not exceed DHFR activity to prevent the depletion of reduced folates and the accumulation of the intermediate dihydrofolate. Comparative genomics analyses identified >200 gene pairs with similar statistical signatures of modular co-evolution, suggesting that cellular pathways may be decomposable into small adaptive units.

Keywords: DHFR; TYMS; adaptive unit; co-evolution; comparative genomics; dihydrofolate reductase; experimental evolution; folate metabolism; forward evolution; synteny; thymidylate synthase; trimethoprim.

Figure 1.. Biochemical and Statistical Representations of Folate Metabolism

(A) Biochemical pathway map of folate metabolism. Blue font indicates metabolites; abbreviated enzyme names are in black or gray text. Black text and lines correspond to enzymes annotated as highest-confidence interactions for DHFR (folA) in STRINGdb version 10.5. See Table S1 for a more complete description of each enzyme. (B and C) Heatmaps of evolutionary coupling between gene pairs in folate metabolism as evaluated by gene synteny (B) and co-occurrence (C), and indicated as a relative entropy, $D_{\mathit{ij}}^{\mathit{intra}}$. Enzyme names are indicated on the left and top of the matrix in black text, gene names are given at right in gray italics. In E. coli, a single gene (folD) encodes a bifunctional enzyme that catalyzes both the methylene tetrahydrofolate dehydrogenase (MTD) and methenyltetrahydrofolate cyclohydrolase (MTCH) reactions. See also Figure S1.

Figure 2.. CRISPRi-Based Measurements of Growth Rate Dependency for DHFR and TYMS

(A and B) Relative growth rates for CRISPRi knockdowns of folate genes paired with either DHFR (A) or TYMS (B). Growth rates are calculated relative to the fitness of a strain carrying an sgRNA with no target homology region (none). Gray bars indicate the growth rate effect of a single mutant; blue bars indicate the effect of the double mutant. Error bars correspond to a 95% confidence interval (CI) across at least three internal technical replicates (see also Method Details). The DHFR knockdown is significantly rescued by TYMS knockdown (p < 0.005 by Student’s t test), but not by other gene knockdown (p > 0.20). As a reference point, the absolute doubling time of the none strain in turbidostat monoculture is 0.83 ± 0.09 (95% CI) h. From this, we estimate that a relative growth rate of −0.2 in these mixed population measurements corresponds to a doubling time of approximately 1.8 h, and −0.4 is approximately zero growth rate (dead). To the right of each bar graph, we also plot a histogram of the growth rate effects for all knock downs of DHFR (A) or TYMS (B).

Figure 3.. Genotype and Phenotype of Selected Strains from Three Evolved Populations

(A–C) Ten single colonies (strains) were selected at the endpoint of each forward evolution condition in 5 μg/mL (A), 10 μg/mL (B), or 50 μg/mL (C) thymidine for genotyping and phenotyping (30 in total). The figure indicates the mutations observed in each strain sampled from that evolution condition. Genes that were mutated in two or fewer strains across all conditions are excluded, as are synonymous mutations (see Table S5 for sequencing statistics and Table S6 for a complete list of non-synonymous mutations). Gene names are labeled along the top edge of the map, with the corresponding residue or nucleotide change(s) denoted along the bottom. If a strain acquires any mutation in a particular gene, then the column section corresponding to that gene is shaded blue. All but four strains acquired mutations in both folA and thyA, encoding DHFR and TYMS, with the few exceptions lacking a folA mutation. A small red star indicates one strain with mutations in only DHFR and TYMS. To the right of each mutation map are trimethoprim (TMP) IC₅₀ and thymidine dependence measurements for each strain. Error bars represent SE over triplicate measurements. Evolved strains 4, 5, and 10 (50 μg/mL thymidine) grew very slowly at all concentrations of trimethoprim measured, and we were unable to determine an IC₅₀ by sigmoidal fit. Table S4 contains exact IC₅₀ values and errors. Thymidine dependence is represented as area under the log(OD₆₀₀) curve in 0 μg/mL thymidine over 10 h. Evolved strains are no longer viable in the absence of extracellular thymidine, indicating a loss-of-function mutation in TYMS. Three “reconstitution strains” featuring representative folA and thyA mutations recombined into a clean genetic background have been included in (A) and (C) for comparison (denoted R2–R4). Red dots in the rows of reconstituted genotypes R2–R4 indicate the IC₅₀ of a strain containing only the corresponding folA mutation.

Figure 4.. A Loss-of-Function Mutation in TYS Buffers Metabolic Changes from Decreased DHFR Activity

(A and B) Scatterplots of relative growth rate for DHFR mutants spanning a range of catalytic specificities (k_cat/K_m), and either a wild-type (WT, gray points) or catalytically dead (R166Q, red points) TYMS. Measurements were performed in M9 media supplemented with 0.2% amicase and either 5 μg/mL (A) or 50 μg/mL (B) thymidine. Error bars correspond to SE across triplicate measurements. (C) Liquid chromatography-mass spectrometry profiling of intracellular folate species. Rows reflect mutant DHFR-TYMS combinations, columns correspond to metabolites. Data represent the mean of three replicates; see also Figure S7 for associated errors. Each folate species can be modified by the addition of 1–5 glutamates. The color of each square denotes the log₂ abundance of each species relative to WT. (D) The corresponding doubling time for each mutant, as measured in batch culture (conditions identical to A).

Figure 5.. Genome-wide Analysis of Co-evolution in E. coli

(A and B) Enrichment of physical and metabolic interactions as a function of synteny coupling (A) or co-occurrence coupling (B). (C) A scatterplot of synteny-based coupling for all of the analyzed gene pairs. Each point represents a pair of Clusters of Orthologous Groups (COGs); coupling within the pair is shown on the x axis, and the strongest coupling outside the pair is shown on the y axis. Color-coding reflects annotations from the STRING database (physical interactions) or the KEGG database (metabolic pathways): green indicates binding, while pairs in dark blue or light blue are not annotated as physical interactions but are found in the same metabolic pathway. Dark blue gene pairs share a metabolic intermediate. The DHFR-TYMS pair is highlighted in red. See Table S8 for an annotated list of gene pairs below the diagonal. (D) Scatterplot of coupling by co-occurrence for all analyzed gene pairs (same format as C).