Mechanism of imidazolium ionic liquids toxicity in Saccharomyces cerevisiae and rational engineering of a tolerant, xylose-fermenting strain

Abstract

Background: Imidazolium ionic liquids (IILs) underpin promising technologies that generate fermentable sugars from lignocellulose for future biorefineries. However, residual IILs are toxic to fermentative microbes such as Saccharomyces cerevisiae, making IIL-tolerance a key property for strain engineering. To enable rational engineering, we used chemical genomic profiling to understand the effects of IILs on S. cerevisiae.

Results: We found that IILs likely target mitochondria as their chemical genomic profiles closely resembled that of the mitochondrial membrane disrupting agent valinomycin. Further, several deletions of genes encoding mitochondrial proteins exhibited increased sensitivity to IIL. High-throughput chemical proteomics confirmed effects of IILs on mitochondrial protein levels. IILs induced abnormal mitochondrial morphology, as well as altered polarization of mitochondrial membrane potential similar to valinomycin. Deletion of the putative serine/threonine kinase PTK2 thought to activate the plasma-membrane proton efflux pump Pma1p conferred a significant IIL-fitness advantage. Conversely, overexpression of PMA1 conferred sensitivity to IILs, suggesting that hydrogen ion efflux may be coupled to influx of the toxic imidazolium cation. PTK2 deletion conferred resistance to multiple IILs, including [EMIM]Cl, [BMIM]Cl, and [EMIM]Ac. An engineered, xylose-converting ptk2∆ S. cerevisiae (Y133-IIL) strain consumed glucose and xylose faster and produced more ethanol in the presence of 1 % [BMIM]Cl than the wild-type PTK2 strain. We propose a model of IIL toxicity and resistance.

Conclusions: This work demonstrates the utility of chemical genomics-guided biodesign for development of superior microbial biocatalysts for the ever-changing landscape of fermentation inhibitors.

Fig. 1

Chemical genomic profiling of ionic liquids. For chemical genomic profiling a genome-wide set of deletion mutants are challenged with a specific compound or solvent control and grown as a pool for several generations. Mutant specific barcodes are then sequenced and compared to control conditions to determine mutants significantly responsive to the chemical stressor (chemical genetic interaction score), which are then used to predict mode of action and points for engineering tolerance

Fig. 2

Chemical genomic profiling of [EMIM]Cl reveals mitochondrial genes are highly sensitive. Of the top 20 most significantly sensitive deletion mutants grown aerobically in YPD with 10 µg/mL [EMIM]Cl, eight were annotated to the mitochondrion (a). We tested the individual sensitivities of the top two most significantly sensitive and resistant mutants compared to the control strain (b, c) using an eight point dose curve. Mutants of ARG2 and QCR2 had significantly lower growth in 0.5 % [EMIM] Cl compared to the WT, whereas mutants of PTK2 and SKY1 grew significantly better (d). (n = 3, Mean ± S.E)

Fig. 3

[EMIM]Cl treatment affects mitochondrial protein levels. Protein abundance and identity of yeast grown in the presence of [EMIM]Cl normalized against a solvent control demonstrates of the top 20 most depleted proteins, eight were annotated to the mitochondrial part. Among the most significantly (p < 0.01) more abundant proteins in the presence of [EMIM]Cl, two were specifically involved in calcium ion homeostasis (blue). (n = 3)

Fig. 4

Effects of [EMIM]Cl on respiration, mitochondrial structure, and membrane potential. Zones of inhibition caused by [EMIM]Cl on yeast grown on either glycerol or glucose (a). Dose dependent disappearance of yeast mitochondrial structure (tubular structures stained with SYTO18) in the presence of [EMIM]Cl (b). [EMIM]Cl treatment at sub lethal doses (0.25 %) causes increases DiOC₆(3) fluorescence, as does the ionophore valinomycin (c, d). The uncoupling agent antimycin is included as a positive control, and the tubulin poison benomyl is included as an inhibitor with a mode of action unrelated to the mitochondrion. DiOC₆(3) fluorescence of the PTK2 mutant when treated with [EMIM]Cl, [BMIM]Cl, valinomycin or control (d)

Fig. 5

The effect of IILs on cell growth in the background strain (Y133) or PTK2 mutant (Y133-IIL). IC₅₀ values were determined for each xylose fermenting yeast strain grown in YPD containing various concentrations of [EMIM]Cl (a), [BMIM]Cl (b), or [EMIM]Ac (c). In (d), Y133 was transformed with the indicated plasmids, and effects on IC₅₀ for [EMIM]Cl using the resulting transformants were assessed. To examine pH dependence on IIL toxicity, specific growth rates of the Y133 and Y133-IIL strains cultured in YPD media containing 1 % [EMIM]Cl at pH 5 or 6.5 (e). Mean ± S.E

Fig. 6

Growth (black), sugar consumption (glucose, green; xylose, blue), and ethanol production (red) of Y133-IIL (solid lines) vs Y133 (dashed lines) in YPXD media with 1 % [EMIM]Cl at under aerobic conditions at pH 6.5. (n = 3, Mean ± S.E, * p < 0.05)

Fig. 7

Final growth and metabolites analysis after of Y133 and Y133-IIL in the presence of [BMIM]Cl. Growth (a), glucose and xylose consumption (b, c) and ethanol production (d) after 72 h of culture under aerobic and anaerobic conditions at pH 6.5 or pH 5.0. (n = 3, except n = 2 for Y133 pH 6.5, Mean ± S.E, * p < 0.05)

Fig. 8

A model for IIL toxicity and resistance. We propose the model of imidazolium IIL toxicity. In the presence of IILs at near neutral pH (a), cells pump out protons via Pma1p, which is coupled with import of the [EMIM]⁺ cation that results in hyperpolerization of the mitochondrial membrane. PTK2 activates Pma1p via phosphorlaytion. Deletion of PTK2 alleviates this by reducing Pma1p activity, and thus [EMIM]Cl influx. The effects of mitochondrial perturbation are more acute in aerobic conditions (red stars vs yellow stars), where mitochondria are more active. At lower pH (b), [EMIM]Cl import is lessened, similar to the polyamine cation spermine, which is alone regulated by PTK2