Yeast Evolves Resistance to Coal-Washing Chemical by Losing a Key Transcription Factor

When a chemical spill contaminated the water supply of Charleston, West Virginia in 2014, the culprit was 4-methylcyclohexane methanol, or MCHM, an alicyclic alcohol used as a synthetic hydrotrope in coal processing. The incident raised public health concerns and turned MCHM into a subject of intense laboratory scrutiny. The finding ties together threads of genome evolution, transcriptional regulation and metabolism into a coherent picture of how a cell copes with a synthetic molecule it would never encounter in nature.

MCHM is an odd molecule biologically. Unlike detergents, which act at low concentrations, hydrotropes must be present at higher levels to solubilize compounds. Endogenous hydrotropes such as ATP and RNA increase the solubility of organic molecules by driving liquid-liquid phase separation, forming membraneless condensates inside cells. MCHM, by contrast, is not readily metabolized, which makes it a useful model for asking what hydrotropes do to living systems. Earlier work showed that MCHM induces a G1 cell-cycle arrest and activates amino acid biosynthesis pathways, effectively creating a starvation signal even when nutrients are plentiful. Proteins with intrinsically disordered regions, which often drive phase separation, are especially sensitive to hydrotropes, and MCHM has been shown to disrupt zinc homeostasis, remodel plasma membrane lipids and trigger amino acid accumulation alongside activation of the environmental stress response.

They grew four biological replicates of the clinical isolate strain YJM789 in rich medium with and without 700 ppm crude MCHM, passaging the cultures every two days for six passages. Two single colonies from each replicate were then isolated, producing eight control strains and eight MCHM-treated strains. When tested on plates containing 1000 ppm MCHM, all eight treated strains showed increased resistance relative to the wild-type parent, while the eight control strains remained as sensitive as wild type. Resistant isolates were also passaged without the chemical to rule out epigenetic inheritance, confirming that the resistance had a genetic basis.

The team then sequenced all sixteen strains at roughly 25-fold coverage with Illumina paired-end reads. Because the standard laboratory reference genome differs substantially from YJM789, the researchers first built a higher-quality YJM789 reference using PacBio long-read sequencing, achieving 82-fold chromosomal coverage. This assembly identified 63,759 SNPs relative to the reference strain, similar to the roughly 60,000 previously reported, with about 90 percent of variants intergenic and nonsynonymous changes occurring three times less frequently than synonymous ones. Notably, both control and MCHM-evolved strains carried approximately 1,600 mutations across 777 genes, and the total number of variants did not differ between resistant and sensitive lines. Principal component analysis of variant content showed no clustering by resistance phenotype or lineage, meaning the raw mutation landscape alone could not distinguish the resistant strains from the controls.

One gene stood out. PDR3 was the only gene carrying nonsynonymous variants in all eight resistant strains, and it was absent from the control lines. PDR3 encodes a transcription factor that regulates the pleiotropic drug response, a yeast defense system built on ATP-binding cassette transporters such as Pdr5, Snq2, Pdr15 and Yor1, which pump chemically diverse toxins out of the cell. Pdr3 and its paralog Pdr1 bind pleiotropic drug response elements upstream of target genes, and different combinations of Pdr1 and Pdr3 dimers compete for binding sites to fine-tune drug-responsive expression. In the evolved strains, the PDR3 variants clustered in the N-terminal 400 amino acids, leaving the DNA-binding regions intact, and structural modeling suggested that truncated alleles could still produce functional proteins. Yet the decisive experiment was simpler: knocking out PDR3 entirely in the YJM789 parent produced resistance comparable to that of an evolved strain carrying a short C-terminal truncation, showing that loss of function, not gain of function, is what matters.

The resistance proved specific to MCHM. Neither the evolved strains nor the pdr3 knockout showed cross-resistance to vanillin, 4-NQO, beta-mercaptoethanol, cycloheximide or hydroxyurea. Resistance also depended on genetic background: it arose in YJM789 but not in the S96 strain of S288c background, despite both being passaged in MCHM, and four nonsynonymous SNPs distinguish the PDR3 alleles of the two strains. In hybrid diploids, the YJM789 allele conferred sensitivity while the S288c allele, or loss of both alleles, increased tolerance. Deleting PDR3 from the evolved strains themselves did not change their growth, further arguing that the evolved mutations mimic loss of function rather than adding a new activity. A pdr1 knockout was as resistant to MCHM as the pdr3 knockout, consistent with the absence of PDR1 mutations in the evolved lines and pointing to a shared pathway through the paralogous transcription factors.

A key clue to the mechanism came from Med15, a component of the tail subcomplex of the Mediator complex, which mediates interactions between transcription factors and RNA polymerase II. Med15 contains three intrinsically disordered regions of variable length, and earlier work had linked variation in Med15 to differences in MCHM response between yeast strains. Med15 is known to interact directly with Pdr3 and to form protein condensates with the transcription activator Gcn4. When the researchers generated double mutants lacking both MED15 and PDR3, the med15 pdr3 double mutant was as sensitive to MCHM as the parental strain, demonstrating that Med15 is required for pdr3-mediated resistance. The authors propose that inactivating Pdr3, Pdr1 or the related factor Yrr1 may free Med15 to function through an alternative mechanism that contributes to resistance, though whether transcription factors such as Pdr3 use liquid-liquid phase separation to regulate transcription remains an open question.

The metabolomic part of the story emerged from an unexpected genetic interaction. During strain construction, the researchers found that the lys2 auxotrophy carried by YJM789 increased sensitivity to MCHM compared with prototrophic strains, and that the pdr3 knockout did not alter growth when LYS2 was restored. This prompted a closer look at amino acid levels. Under untreated conditions, the metabolomes of wild type and the pdr3 knockout were not well separated, indicating that Pdr3 does not shape the metabolome in the absence of stress. Upon MCHM exposure, however, amino acid levels shifted strongly in wild-type cells but far less in cells lacking Pdr3, whose metabolic profile instead resembled untreated controls. In total, MCHM altered the levels of sixteen amino acids, including glutathione precursors and breakdown products. Cysteine was consistently depleted in treated strains, a significant observation because cysteine is required for glutathione synthesis, and glutathione and related intermediates were reduced in MCHM-exposed wild-type cells but not after PDR3 deletion.

Follow-up experiments reinforced the centrality of glutathione metabolism. Mutants defective in amino acid transport (gap1), ROS metabolism (sod1, sod2, ctt1) or glutathione production and recycling (glr1, gsh1, gsh2) were sensitive to MCHM, and the growth of these strains could be partially rescued by supplying glycine, glutathione, N-acetyl cysteine or glutamic acid. Ammonium phosphate restored growth in nearly all strains, though not in pdr5, gap1 or sod1 mutants, suggesting that excess nitrogen from most sources can largely bypass the metabolic bottleneck. Reexamination of the evolved genomes revealed that several resistant strains carried independent mutations in GEX2, which transports glutathione, and SAM3, which transports S-adenosylmethionine, the precursor of methionine and ultimately cysteine, the rate-limiting step in glutathione synthesis. Both gex2 and sam3 knockouts were sensitive to MCHM in the S288c background, and the recurrence of independent mutations in these transporters across resistant lines suggests they contributed to resistance rather than hitchhiking as passengers.

The broader context extends beyond yeast. Previous work showed that excess glutathione, but not ascorbic acid, rescued viability after MCHM exposure in mammalian HEK293T cells and in Xenopus embryos, indicating that the glutathione connection may be conserved. The authors propose that PDR3 indirectly regulates glutathione, enabling cells to withstand the metabolic stress caused by MCHM, and that loss-of-function mutations, being more common than gain-of-function events, offer a reproducible evolutionary route to resistance. Because earlier genetic screens had focused on sensitivity rather than resistance, the PDR3-driven phenotype had been missed until now. The study also leaves open questions: resistant strains showed reduced flocculation, MCHM inhibited invasive growth into agar, and the team could not evolve additional resistance in the naturally more tolerant S96 strain, which does not flocculate at all. What is clear is that a synthetic hydrotrope spilled into a river a decade ago has become a powerful probe of how eukaryotic cells rewire transcription and metabolism to survive chemical stress, and that sometimes the cheapest evolutionary solution is simply to switch a regulator off.

Subject of Research: Evolution of MCHM resistance in yeast through PDR3 loss-of-function mutations and altered glutathione metabolism

Article Title: Loss of PDR3 alters metabolome in response to MCHM, a synthetic hydrotrope

Article References: Momtareen, T., Ayers, M. C., Patriarcheas, D., McCarthy, L., Judge, D., Poziviak, S., Leombruno, G., Quinn, M., Wonsettler, N., Lowery, C., McCulloch, S., Dale, N., Jonas, F., & Gallagher, J. E. G. (2025). Loss of PDR3 alters metabolome in response to MCHM, a synthetic hydrotrope. Current Genetics, 71(1), Article 30. https://doi.org/10.1007/s00294-025-01333-w

Image Credits: AI Generated

DOI: 10.1007/s00294-025-01333-w

Keywords: MCHM, yeast, PDR3, in-lab evolution, pleiotropic drug response, glutathione, Med15, Mediator complex, metabolomics, hydrotrope, transcription factor, chemical resistance