https://orcid.org/0000-0001-9499-0946
Figures
Abstract
Ethanol is a fermentation product widely used as a fuel and chemical precursor in various applications. However, its accumulation imposes severe stress on the microbial producer, leading to significant production losses. To address this, improving a strain’s ethanol tolerance is considered an effective strategy to enhance production. In our previous research, we conducted an adaptive evolution experiment with Escherichia coli growing under gradually increasing concentrations of ethanol, which gave rise to multiple hypertolerant populations. Based on the genomic mutational data, we demonstrated in this work that adaptive alleles in the EnvZ-OmpR two-component system drive the development of ethanol tolerance in E. coli. Specifically, when a single leucine was substituted for a proline residue within the periplasmic domain using CRISPR, the mutated EnvZ osmosensor caused a significant increase in ethanol tolerance. Through promoter fusion assays, we showed that this particular mutation stabilizes EnvZ in a kinase-dominating state, which reprograms signal transduction involving its cognate OmpR response regulator. Whole-genome proteomics analysis revealed that this altered signaling pathway predominantly maintains outer membrane stability by upregulating global porin levels and attenuating ferric uptake and metabolism in the tolerant envZ*L116P mutant. Moreover, we demonstrated that the hypertolerant envZ*L116P allele also promotes ethanol productivity in fermentation, providing valuable insights for enhancing industrial ethanol production.
Author summary
Ethanol is a versatile chemical with many applications, but producing it in high quantities remains a challenge. This is because Escherichia coli, a candidate ethanol production strain, is naturally sensitive to this short-chain alcohol, especially when levels are gradually accumulating during fermentation. To resolve this bottleneck, we have investigated how E. coli can acquire tolerance to its own toxic fermentation product. Our research indicated that a single amino acid substitution in EnvZ- a key sensor protein that normally protects E. coli against extreme osmotic stress- is sufficient to confer ethanol tolerance. Further analysis revealed that the mutation perturbs the EnvZ-mediated signaling cascade, which, in turn, changes the transporter composition in the outer membrane and attenuates the cell’s ferric metabolism. These adaptations enable E. coli to survive under high-ethanol conditions, thereby promoting its ethanol production efficiency. This discovery provides a suitable strategy to increase ethanol titers in industrial settings using fermentation.
Citation: Schalck T, Katyal M, De Graeve S, Roba L, Baldoma JV, Swings T, et al. (2025) Reprogramming the EnvZ-OmpR two-component system confers ethanol tolerance in Escherichia coli by stabilizing the outer membrane and altering ferric homeostasis. PLoS Genet 21(12): e1011707. https://doi.org/10.1371/journal.pgen.1011707
Editor: Morten Kjos, Norwegian University of Life Sciences: Norges miljo- og biovitenskapelige universitet, NORWAY
Received: April 29, 2025; Accepted: November 4, 2025; Published: December 17, 2025
Copyright: © 2025 Schalck et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The raw data, additional data, and scripts that support the findings of this study are publicly available on GitHub through the URL: https://github.com/Michielslab/EnvZ_OmpR_supplementaryData.
Funding: Fonds Wetenschappelijk Onderzoek (https://fwo.be/en/) (1S14417N to T.S., 11PNI24N to M.K., 12O1922N to B.VdB., 12O1917N to B.VdB, G0B0420N to J.M., and G0C4322N to J.M.), Fonds Wetenschappelijk Onderzoek - Strategisch Basisch Onderzoek (https://www.fwo.be/en/support-programmes/all-calls/senior-researchersresearch-teams/sbo-strategic-basic-research-projects/) (S001421N to J.M.), Fonds Wetenschappelijk Onderzoek - EOS Research Project (https://fwo.be/en/support-programmes/all-calls/senior-researchersresearch-teams/eos-research-project/) (40007496 to J.M.) grants and internal KUL (928 C16/17/006 to J.M.) and VIB funding to J.M. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. T.S., M.K., and B.VdB received salary from their respective funders.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The global climate crisis today is mainly attributed to the increased use of fossil fuels and petrol-derived products and the associated increase in high greenhouse gas (GHG) emissions [1,2]. To limit the widespread impact of global warming and reduce GHG emission-related health issues, producing biofuels from renewable resources (e.g., agricultural residues or energy crops) instead of extracting crude oil from oil reservoirs is considered the most adequate strategy [3–5]. Particularly, bioethanol has become an established petroleum alternative in the US, Brazil, and the EU as transportation fuel [6–8] but also as a precursor for bulk chemicals such as ethylene oxide [6,7].
Although microbial-based bioethanol production is a cost-effective method, maximizing product titers often remains challenging as the microbial producer (e.g., Saccharomyces cerevisiae, Zymomonas mobilis, or Escherichia coli [8–11]) experiences stress from high concentrations of (lignocellulosic) biomass-derived inhibitors and end-product (e.g., ethanol) [12–15]. Due to its amphiphilic character, ethanol can interact with the microbial plasma membrane, resulting in severe membrane disorders and disruption of the proton motive force and ion/nutrient fluxes [16–20]. Furthermore, this small, two-carbon alcohol may cross the membrane and, as a result, impede transcription and translation, damage protein structures and DNA (whether or not due to accumulation of reactive oxygen species) [20–25].
Previous research has indicated that adaptation to ethanol stress is a complex process in which a combination of cell envelope-adapting, ROS-scavenging, protein-refolding, and energy-restoring pathways are involved [20,23,25–29]. Here, we focus on a crucial adaptation mechanism that conserves the outer membrane (OM) structure under lethal ethanol stress. Based on the mutational dataset of a previously conducted evolution experiment under continuous ethanol stress, the amino acid substitution, L116P, in the central EnvZ histidine kinase osmosensor was found to confer high ethanol tolerance in E. coli [30,31]. We show that this particular amino acid alteration inside the periplasmic domain (PD) disturbs the native kinase/phosphatase activity. As a result, this evolved envZ allele reprograms the expression pattern of downstream-regulated genes and, consequently, represses the enterobactin biosynthesis and transport systems, involved in ferric uptake, and adjusts the porin level within the outer membrane to survive lethal ethanol concentrations. Furthermore, we demonstrated that this EnvZ/OmpR-involved tolerance mechanism is also a suitable strategy to accelerate ethanol production parameters of E. coli in fermentation settings.
Results
The EnvZ-OmpR two-component system underlies adaptation to lethal ethanol stress
We have previously evolved 16 E. coli populations (HT1-HT16) by gradually increasing the imposed ethanol concentration to obtain high-tolerant E. coli lines that can eventually grow in 8.5% ethanol (S1 Fig) [30,31]. To prioritize the most relevant tolerance-conferring adaptation mechanisms in E. coli, we here first conducted a Gene Ontology (GO) enrichment analysis on the evolution dataset of Swings et al. (2017) (S1 Data) [30]. We specifically focused our analysis on the 2,560 genes, exhibiting at least one mutation across the 16 parallel-evolved populations at each sequenced time point. This Gene Ontology (GO) enrichment analysis revealed that, throughout evolution, E. coli predominantly acquired mutations in genes that encode membrane-associated proteins with a role in transport or kinase-mediated signal transduction (S2 Fig; S1 Table). We further focused on the relevance of the E. coli Two-Component Systems (TCSs) in ethanol adaptation since the cellular localization and function of these bacterial signaling pathways correspond to the enriched GO terms (S2 Fig; S2, S3 and S4 Tables). More specifically, a bacterial TCS is composed of a membrane-associated sensor that, depending on the environmental conditions, (de)phosphorylates its cognate response regulator to induce or repress expression of downstream-regulated genes [32–34] (Fig 1B). Interestingly, mutations in the sensor and response regulator moiety of 26 distinct TCSs were frequently observed among all ethanol-evolved E. coli populations at some stage of adaption, especially in the EnvZ-OmpR, DpiBA, RstBA, BaeSR, and AtoSC signal transduction pathways (S3 Fig).
(A) The EnvZ-OmpR signal-transduction system was appointed the most relevant TCS in ethanol tolerance based on the priority score. This score was defined as the product of the mutation frequency (or coverage) across all evolved populations and the number of membrane-associated transporters within the corresponding regulon. (B) Schematic representation of the EnvZ-OmpR TCS. From N-to C-terminus, the EnvZ osmosensor consists of five principal domains: the transmembrane (TM), periplasmic, the HAMP (which is shared among Histidine kinases, Adenylate cyclases, Methyl-accepting proteins, and Phosphatases), the Dimeric, Helical domain containing the central H243 (encircled H) that is phosphorylated (DHp) and the C-terminal catalytic domain that binds ATP (CA) [33]. A hyperosmotic shock induces the kinase (K) activity of EnvZ which involves auto-trans-phosphorylation of H243 and phosphotransfer of the phosphate (encircled P) towards the D55 residue of OmpR [34–36]. Contrary, a hypoosmotic condition stimulates EnvZ’s phosphatase activity which removes the phosphate from OmpR [39,40]. The L116P amino acid substitution is depicted as a black, orange highlighted star, located at the periplasmic domain.
Based on the frequency at which mutations emerged in the sensor and response regulator moieties (S3 Fig) and the number of membrane-associated transporters within the TCS’s regulon (S4 Fig), an overall priority score was finally assigned to each TCS (Fig 1A). The EnvZ-OmpR signaling pathway was the highest-ranked evolutionary target. Therefore, we decided to investigate the relevance of this TCS in ethanol tolerance and focused on the impact of the most frequently occurring envZ allele, called L116P. The native gene product, EnvZ, acts as a histidine kinase/phosphatase osmosensor that tunes the phosphorylation state of its cognate response regulator OmpR, according to the environmental osmolarity [32,34–38]. In turn, the degree of phosphorylation determines the DNA binding affinity of OmpR and, thus, defines the expression pattern of all downstream-regulated genes involved in the response towards extreme osmotic stresses [39,40]. Given the emergence of the L116P substitution in different parallel populations and the pivotal role of EnvZ in membrane protein regulation, we anticipated that the mutant EnvZ*L116P osmosensor may confer ethanol tolerance by remodeling the cell envelope structure as a result of altered regulation of membrane-associated (transporter) proteins.
The L116P mutation in envZ is adaptive and improves ethanol tolerance in vivo
To verify whether the L116P allele of interest is sufficient to improve ethanol tolerance, the mutation was reconstructed in the wild-type (WT) background. The growth kinetics and cell survival of envZ*L116P were subsequently assessed in the presence of 5% ethanol. This concentration completely inhibits the growth of the WT BW25113 strain and also represents the ethanol level at which the evolution experiment was originally initiated [30]. Additionally, strains lacking either envZ or ompR were also subjected to 5% ethanol stress (Fig 2). Strikingly, the envZ*L116P mutant is able to grow in a 5% ethanol-enriched medium and survives an otherwise lethal ethanol stress. In contrast, disrupting EnvZ and OmpR results in a strong decrease in both microbial growth (Fig 2A; S2 Data) and survival (Fig 2B; S3 Data). Hence, OmpR mediates the tolerance-conferring property of the mutant EnvZ*L116P osmosensor since deleting this response regulator completely abolishes the growth and survival advantage of the gain-of-function L116P allele.
(A) The splines-fitted growth curves for the BW25113 WT, the envZ*L116P mutant, and EnvZ-OmpR deletion mutants (ΔenvZ and ΔompR) under a 12h exposure to 5 (v/v)% ethanol. Based on the QurvE-fitted growth curves [41], the optical density (OD595) at the plateau phase and starting point (t0) were extracted (S5 Fig). The OD595 increment at the plateau phase vs. the start (t0) for each strain was considered to extract the statistics. P-values were obtained from one-way ANOVA with a Dunnett’s posthoc test. Error bars represent the 95% confidence intervals (n = 4). (B) Survival fraction of the corresponding E. coli strains. Data points indicate the mean and error bars represent the 95% confidence intervals (n = 4). Levels of significance are indicated as follows: *, P ≤ 0.05, **, P ≤ 0.01, ***, P ≤ 0.001, ****, P ≤ 0.0001 and are derived from a Generalized Linear Interaction Model with Time [h] as a (discrete) factor using the WT strain as the baseline, paired with a Dunnett’s post-hoc test. The P-values of the ΔompR strains are identical and highlighted in red.
Kinase-promoting envZ variants confer ethanol tolerance
Given the central role of EnvZ in osmoregulation and tolerance [42], we anticipated that the L116P substitution resulted in a reprogrammed bifunctional kinase:phosphatase activity of EnvZ and, therefore, caused the improved tolerance phenotype. To compare the kinase:phosphatase activity of the mutant with the WT EnvZ sensor, we recorded the expression pattern of ompC (S4 Data) and ompF (S5 Data), encoding the major outer membrane porins, using a fluorescence-coupled promoter fusion assay [43–45]. Previous research has indicated that ompC is induced or repressed and ompF repressed or induced when OmpR-P levels are high or low, respectively (Fig 3C) [39].
(A) (Top) When the kinase-promoting L116P and T247R are present in the EnvZ osmosensor, ompC (blue) is upregulated and ompF (grey) is repressed, compared to the WT. In contrast, the kinase-defective EnvZ*F390L induces neither ompC nor ompF. (Bottom) Schematic representation of the EnvZ domain structure (PD, periplasmic domain; H-box, the amino acid sequences surrounding the central H243; EnvZc, catalytic C-terminal moiety encompassing the DHp and CA domains, the other references match the nomenclature in Fig 1) together with the ethanol-evolved L116P mutation (black star) and literature-derived T247R (red star) and F390L (blue star) substitutions [46–48]. Levels of significance are derived from a (Generalized) Linear Mixed Effects Regression model with Dunnett’s post-hoc test (with WT as the reference). (B) The impact of reprogramming the enzyme activities of EnvZ on ethanol tolerance, expressed in terms of survival. The envZ*L116P and envZ*T247R mutants that show elevated ompC levels (i.e., display high kinase:phosphatase ratios), can cope with ethanol stress significantly better than envZ*F390L or envZWT that do not share the increased ompC expression feature. Levels of significance are derived from a Generalized Linear Mixed Effects Regression model followed by a Dunnett’s post-hoc test (with WT as reference). (C) The anticipated OmpR-P levels and OM porin expression profile in E. coli as a result of changing EnvZ kinase:phosphatase activities in response to the environmental osmolarity. The large pore size OmpF porin predominates at low osmolarity, corresponding to decreased kinase:phosphatase ratios and hence low OmpR-P levels. High osmolarity, in contrast, induces high kinase:phosphatase ratios, increases the OmpR-P concentration, and favors ompC expression, encoding an OM porin with a small diffusion pore [49]. The log10-transformed ompC:ompF expression ratio, reflecting the EnvZ kinase:phosphatase bifunctional activity, in response to hyperosmotic (D), mimicked using high PEG6000 concentrations, and ethanol (E) stress. Expression intensities of each porin in each strain (WT, grey and envZ*L116P, black) were expressed in terms of FITC fluorescence values as measured through flow cytometry. Levels of significance are derived from a t-test on the slope of a Linear Regression Model. Levels of significance:*, P ≤ 0.05, **, P ≤ 0.01, ***, P ≤ 0.001, ****, P ≤ 0.0001. Raw flow cytometry histograms are provided in S8 Fig.
By tracking ompC and ompF expression levels, we demonstrated that the envZ*L116P mutant, in stress-free medium, stimulates ompC and strongly represses ompF with respect to the WT (Figs 3A, S6; S6 and S7 Data). Hence, the EnvZ*L116P osmosensor has an altered kinase:phosphatase balance in favor of the kinase activity that is reminiscent of the hyperosmotic, kinase-dominant state. Furthermore, mutants such as ΔenvZ and ΔompR that are hypersensitive towards ethanol and are dysfunctional in EnvZ-OmpR signal transduction, express both ompC and ompF barely (S6C Fig). Altogether, our findings suggest that a high kinase:phosphatase ratio and the resulting, reprogrammed signal transduction pathway underlies ethanol tolerance. To further corroborate this, we studied two envZ alleles that have been examined before in terms of their impact on the kinase:phosphatase equilibrium but have not been evaluated yet with regard to ethanol tolerance (Fig 3A). The first one, T247R, is located in the DHp-associated H box, nearby the catalytic H243, and is known to favor EnvZ’s kinase state, similar to the L116P allele [46,47]. In contrast, the CA domain-located F390L substitution renders EnvZ kinase-deficient and simultaneously increases its phosphatase activity [48]. While the envZ*T247R mutant, similar to envZ*L116P, displays improved tolerance, the envZ*F390L strain is equally hypersensitive to ethanol as the ancestral WT (Fig 3B; S8 Data).
Next, we subjected the ethanol-tolerant envZ*L116P mutant to ethanol and hyperosmolarity to assess whether the evolved EnvZ*L116P osmosensor is ‘locked’ in a constitutive high kinase state irrespectively of the environmental stressor and, thus, whether the ethanol-adapted system is still responsive to stresses. To impose a hyperosmotic stress, we mixed polyethylene glycol (PEG6000, S7 Fig) as osmolyte into the growth medium to, ultimately, force EnvZ into its high kinase:low phosphatase state (Fig 3C; S9 and S10 Data). Consequently, the latter should be reflected in a high ompC:ompF expression ratio. In response to hyperosmotic stress, the L116P allele displays native osmosensing ability as ompC expression is triggered slightly and ompF expression is repressed strongly in accordance with the PEG6000 concentration, similar to the WT EnvZ (Figs 3D, S8A, S8B, S8E, and S8F). As a result, the ompC:ompF ratio, albeit consistently higher in the envZ*L116P mutant, increases as would be expected in response to hyperosmotic stress (Fig 3D). In contrast, ethanol stress did not evoke any change in biochemical activity, neither in the WT nor in the ethanol-evolved osmosensor (Figs 3E, S8C, S8D, S8G, and S8H). The latter suggests that the EnvZ osmosensor, regardless of the L116P substitution, is inherently insensitive to ethanol. Therefore, cells can only benefit from improved ethanol tolerance when they acquire the L116P allele that forces the EnvZ osmosensor into a constitutive, ethanol-independent high K:P state.
The outer membrane porin OmpC contributes to ethanol tolerance in the envZ*L116P mutant
The EnvZ-OmpR-tandem acts as a signal transduction cascade in which the OmpR response regulator ultimately adjusts the expression pattern of multiple genes that belong to the OmpR regulon, according to the kinase:phosphatase balance of EnvZ [50]. Importantly, eliminating ompR turns both the WT and envZ*L116P mutant strains hypersensitive to ethanol, indicating that at least one OmpR downstream-regulated gene should be involved in the ethanol tolerance phenotype. Therefore, we first evaluated the role of OmpC and OmpF in ethanol tolerance since the expression of both porins was significantly altered in the envZ*L116P mutant (Fig 3A). In general, inactivating ompC drastically compromised growth (Figs 4A, 4B; S9 and S11 Data) and survival (Fig 4D, 4E and S12 Data) under ethanol stress, although the effect was less pronounced compared to the inactivation of ompR (Figs 4C, S10, S11, S12, S13 and S14). However, deleting ompC did not fully abolish the envZ*L116P-associated tolerance phenotype since the survival of envZ*L116P ΔompC mutant was still higher than the WT reference strain with ompC after 24h of ethanol exposure (P < 0.0001). In contrast to eliminating ompC, removing ompF has no impact on ethanol tolerance (Fig 4A and 4B). Alongside the survival-based tolerance assay (S13 Data), we reached the same conclusions on ethanol tolerance or sensitivity when the cell shape was microscopically monitored over time in a differentially-labeled, coculturing ethanol exposure experiment (S15 Fig; S14, S15, S16 and S17 Data). While ethanol-induced morphological deformations were limited in envZ*L116P and envZ*L116P ΔompF (S15A, S15B, S15C, S15I, S1J and S15K Fig), the WT, ΔompC, and envZ*L116P ΔompC strains exhibited elongated cell shapes (S15E, S15F and S15G Fig), reflecting their hypersensitivity to ethanol stress. Moreover, in the ΔompC strains, subpopulations emerged that were unable to stably express their fluorescent label when exposed to prolonged ethanol stress, suggesting a deficit in gene expression (S16 Fig). Hence, OmpC is an important member of the OmpR regulon involved in ethanol tolerance.
The growth rate of the WT ancestor (A) or envZ*L116P (B) lacking either the ompC or ompF porin genes. (-) indicates either the WT or tolerant mutant with intact porins. Growth rates were determined from OD-curves (at 595 nm) as shown in S9 Fig and statistically compared using a one-way ANOVA with Dunnett’s post-hoc test (in which either the WT, A, or envZ*L116P, B, serve as a reference). (C) Quantitative summary representing the mean spot area and whiteness intensity of four independent assays including eight different strains. Bacteria were diluted up to 106-fold and spotted on LB agar containing 6% ethanol. The results of the 106 dilution were excluded from this picture since barely any bacterial growth could be detected. Each bacterial spot was quantitatively analyzed in ImageJ (see Materials & Methods). The original images are displayed in S10, S11, S12, S13 and S14 Figs. The effect of deleting ompC or ompF on the survival under ethanol stress, measured by CFU enumeration, of the WT (D) or the envZ*L116P strain (E). Statistics of (D) and (E) are derived from a Generalized Linear and Linear Model (based on the AIC outcome), respectively, paired with a Dunnett’s post-hoc test. (F-H) Microscopy image samples of a mixed WT (green) - envZ* (magenta) (F), envZ*L116P ΔompC (green) – ΔompC (magenta), and (H) envZ*L116P ΔompF (green) – ΔenvZ*L116P (magenta) populations, exposed to 5% ethanol for 8h. The white line at the bottom right of each image represents the 5 µm reference scale bar. A detailed analysis of the microscopy images can be found in S15 Fig.
OmpF rescues the ethanol sensitivity of a ΔompC deletion by reducing the ethanol-induced permeabilization
Our results show that E. coli cells lacking OmpC exhibit a lower survival (Fig 4) and suffer from severe morphological defects under ethanol stress (S15 Fig). In contrast, OmpF was not identified as a crucial factor for improved survival under ethanol stress neither in the WT nor in the envZ*L116P mutant, even though this porin shares high sequence similarity with OmpC and is also implicated in the osmotic stress response [51]. Compared to OmpC, the diffusion pore of OmpF is flanked by a lower number of charged residues increasing its pore size [44,51]. Therefore, we reasoned that the tolerance difference between the ΔompC and ΔompF deletion strains may be attributed either to distinctive, inherent properties of OmpC and OmpF (e.g., pore size) or because their absence causes the global outer membrane porin (OMP) content to change. The latter is more relevant in case of OmpC since we already provided evidence that the expression of ompC is at least equally high as that of ompF in absence of stress and even higher when the host is exposed to ethanol or hyperosmotic stress (Fig 3A, 3D and 3E). If the global porin level in the OM is indeed critical, compensating for the lack of ompC in ΔompC deletion mutants by providing an additional copy of ompF should rescue the ethanol strains under ethanol stress. Therefore, we replaced the original ompC gene in the WT and envZ*L116P strains with an identical, second ompF gene copy (called ompF’) to create two new strains, designated ompF’/ompF and envZ*L116P ompF’/ompF respectively. Importantly, the original ompC promoter and the 5’ UTR were preserved ensuring that ompF’ was regulated in the same manner as the ompC gene it replaced. Indeed, LC-MS/MS-based quantification of the porin abundances between the envZ*L116P and envZ*L116P ompF’/ompF mutants revealed that the global porin level was identical. However, while the predominant porin in envZ*L116P was clearly OmpC (with very few OmpF), the outer membrane in the envZ*L116P ompF’/ompF mutant solely consisted of OmpF (Fig 5A). Similarly, we also examined whether other less related outer membrane proteins, including OmpA, LamB, or TolC, could also compensate for the loss of OmpC in ethanol tolerance by replacing the ompC coding sequence with each of the respective genes.
(A) The relative ion intensities (or counts), defined as the normalized label-free quantification (LFQ) values of each porin (for each replicate), as a proxy for relative porin abundance. Yellow represents OmpC whereas purple indicates the OmpF level. Statistical interference is derived from one-way ANOVA combined with a Tukey multiple comparisons post-hoc test. (B) The envZ*L116P with the additional ompF’ copy is equally tolerant as the envZ*L116P mutant, as opposed to envZ*L116P ΔompC. Levels of significance are obtained from a Generalized Linear Model paired with a Dunnett’s post-hoc test using envZ*L116P as the reference (n = 4). Error bars represent the 95% confidence intervals. (C) Correlation between the N-Phenylnaphthalen-1-amine (NPN) uptake, which is a measure for outer membrane permeability, at the 5 vs. 0 (v/v)% ethanol condition. The points represent the mean values while the crosshairs indicate the 95% confidence intervals (in x-, and y-direction). The dotted line in the background denotes a Linear Regression Model together with the grey ribbon reflecting the 95% confidence interval. Levels of significance: *, P ≤ 0.05, **, P ≤ 0.01, ***, P ≤ 0.001, ****, P ≤ 0.0001.
In terms of outer membrane protein abundance, the total OmpC and/or OmpF porin level in strains carrying the EnvZ*L116P mutant sensor was also higher compared to the WT (a one-way ANOVA with Tukey’s post-hoc test, P = 0.04), indicating that its superior tolerance is linked to a global increase in the major outer membrane protein, OmpC and OmpF, content. When ompC was replaced with an ompF copy, the survival of the envZ*L116P mutant under 5% ethanol stress was not compromised, while just knocking out ompC severely diminished ethanol tolerance (Figs 5B, S17; and S18 Data). However, substituting ompC for any of the alternative OmpA and LamB porins or the TolC efflux channel failed to restore the superior tolerance phenotype of the envZ*L116P mutant when ompC was deleted (S18 Fig; S19 Data). Since outer membrane proteins, such as the β-barrel outer membrane porins OmpC and OmpF (OMPs), contribute to the structural integrity of the outer membrane in Gram-negatives, we anticipate that altering their composition has a profound impact on the cell envelope physiology and properties [52–54]. Because ethanol is known to cause cell envelope defects, reflected in the severe cell shape deformations observed in S15 Fig, we argued that a cell’s outer membrane (OM) permeability and stability, as dictated by its OMP abundance, defines its tolerance or susceptibility to ethanol stress. To corroborate this, we monitored the uptake of N-phenylnaphthalen-1-amine (NPN) as an indicator for ethanol-induced disruption, and hence permeabilization, of the OM. This dye exclusively turns fluorescent when it crosses a permeabilized OM and binds to the inner, cytoplasmic membrane [55,56]. As a proof-of-concept, the fluorescence signal intensity of the NPN dye, representing OM permeability, is dependent on the applied concentration of ethanol or polymyxyin B, a well-known antibiotic that causes OM damage (S19 Fig). This assay revealed that the OM permeability under 5% ethanol exposure globally correlates approximately with a strain’s intrinsic OM permeability (without any ethanol) (Fig 6C; S20 Data). Importantly, OM permeability as a measure for membrane integrity or stability was directly linked to a strain’s genotype. Most noticeably, deleting ompR severely impairs OM permeability and the same is also valid for ompC, although the effect is less pronounced. Providing an additional ompF gene copy restores the OM permeability (i.e., integrity) of the ΔompC strains back to the baseline OM permeability of WT and envZ*L116P, respectively (Fig 5C).
Effect of deleting ompC, tolC, and lpcA on ethanol tolerance in the WT (A) and envZ*L116P mutant (B) strains. Statistics are inferred from a Generalized Linear Model with a Dunnett’s post-hoc test (n = 4, error bars represent 95% confidence interval) using either the WT or envZ* as the reference. Levels of significance: *, P ≤ 0.05, **, P ≤ 0.01, ***, P ≤ 0.001, ****, P ≤ 0.0001. (C) The impact of deleting ompC, tolC, and lpcA on the OM permeability under 5% ethanol stress (measured as NPN uptake). Statistics are derived from a Mixed Effects Linear model linked with a Dunnett’s post-hoc test with either the WT or envZ*L116P as reference. (D) Tolerance, expressed as mean survival fraction, vs. the mean NPN uptake under 5% ethanol exposure for 5h (n = 3). The correlation parameters are derived from a Spearman correlation test. The grey infill represents the 95% confidence interval of the black correlation curve.
In conclusion, rather than composition, the total OmpC-OmpF outer membrane porin abundance determines ethanol tolerance, as insufficient porins affect OM properties, including permeability, making the host more ethanol-sensitive.
LPS biosynthesis and the TolC OM channel are implicated in EnvZ*L116P-mediated ethanol tolerance
We noticed that E. coli strains lacking OmpR are more vulnerable to ethanol stress than those with a single ompC deletion (Figs 2 and 4). To identify additional genes within the OmpR regulon (S6 Table) that confer tolerance, we pursued two complementary strategies: phenotyping knock-out mutants and comparing the proteomic profiles of the tolerant envZ*L116P mutant and the WT strain (as discussed in the next section). First, we constructed a collection of deletion mutants in the envZ*L116P background, each lacking a single known target of the OmpR regulon, and assessed the ethanol tolerance of each mutant (S21 Data; S1 Text).
The first approach identified two gene deletions that significantly impaired the ethanol tolerance phenotype of envZ*L116P. Deletion of lpcA, which encodes the sedoheptulose 7-phosphate isomerase lipopolysaccharide (LPS)-biosynthesis enzyme, or tolC, a component of the AcrAB-TolC efflux pump-associated outer membrane channel, turn the envZ*L116P mutant susceptible to ethanol stress (Figs 6A, 6B, S20; and S22 Data). Remarkably, eliminating the AcrA and AcrB subunits of the AcrAB-TolC tripartite extrusion system or their corresponding repressor, AcrR, does not affect the superior ethanol tolerance phenotype of envZ*L116P (S20 Fig).
Since all identified hits within the OmpR regulon are either an intrinsic component of the OM (tolC, and ompC) or are involved in LPS biogenesis (lpcA), we also studied the impact of deleting these genes on OM permeability (S23 Data). We observed that permeability was significantly affected in both the WT and envZ*L116P mutant when lpcA or tolC were deleted (Fig 6C). In addition, strains with a higher OM permeability, due to defects in OM integrity, also tend to score lower in terms of ethanol tolerance (Fig 6D; S24 Data). Although it should be noted that, despite the lower cell envelope integrity (Fig 4) and tolerance of the WT, its permeability—as measured by NPN uptake—is not higher than that of envZ*L116P (Fig 6C). This implies that destabilizing cell envelope integrity as a result of ethanol exposure is not always linked with an increase in NPN measured outer membrane permeability.
Reduced enterobactin biosynthesis and ferric transport underlie the superior tolerance of the envZ*L116P mutant
Aside from investigating the importance of OmpR regulon members for the ethanol tolerance phenotype, we expanded our scope to identify additional tolerance-conferring candidates using whole-genome shotgun proteomics. Gene Ontology and KEGG pathway enrichment analysis of the proteomes revealed that proteins related to iron and siderophore transport (FhuA, Fiu, FecABE, and OmpF), and enterobactin biosynthesis (EntACEF) are underrepresented in the tolerant envZ*L116P and envZ*L116P ompF’/ompF mutants compared to the WT strain (S21, S22, S23, S24 Figs; S25 Data). Interestingly, this observation is largely consistent with previous work, in which the corresponding authors had shown that kinase dominant envZ mutants negatively impact ferric transport [57,58]. However, while Gerken et al. (2020) [58] demonstrated that another kinase-dominant envZ*R397L represses ferric uptake but activates ferrous import, our proteomics data only provide evidence for repression of the ferric transporter and enterobactin-mediated ferric uptake system.
To confirm the causality of the ferric uptake-involved gene set for the EnvZ*L116P-associated tolerance phenotype, we tested the survival of the corresponding deletion mutants under 5% ethanol stress (S26 Data). We hypothesized that knocking out these specific genes would bring the tolerance level closer to, or even match, that of envZ*L116P since siderophore biosynthesis and ferric transport gene clusters are repressed in this mutant. The subsequent survival assay (Fig 7C) revealed that deleting almost all genes within the enterobactin biosynthesis cluster, except for entC, along with the OM enterobactin transporter (fepA) enhances ethanol tolerance of the sensitive WT strain. In addition, knocking out fecB, encoding the periplasmic substrate-binding component of the ferric citrate ABC transporter, also proved beneficial for ethanol tolerance (Fig 7C).
Summary figures are produced from the ShinyGO application [59] (https://bioinformatics.sdstate.edu/go/). (C) The survival of the ion metabolism deletion mutants under 5% ethanol exposure. Colored dots represent the mean of the strains that significantly display a higher tolerance than the WT strain (grey). Black dots represent mutants that show the same survival as the WT. Error bars (n = 4) represent the 95% confidence intervals and the asterisks highlight the collective P-value of all tolerant strains as derived from a Generalized Linear Mixed Effects Model with a Dunnett’s post-hoc test.
Next, we investigated the relevance of ferric homeostasis and uptake for the ethanol tolerance phenotype. Because of the established link between ethanol toxicity and ROS stress [22,60], we hypothesized that reduced uptake of ferric, as a result of impaired enterobactin (siderophore) biosynthesis (cf. Ent-operon) or decreased expression of dedicated ferric-associated transporters (cf. FecB and FepA), may explain the superior tolerance phenotype of the envZ*L116P tolerant strain. To assess the level of ROS stress experienced by the WT and envZ*L116P strains under 5% ethanol stress, the fluorescence signal intensity from two distinct promoter fusion plasmids was monitored using flow cytometry [61,62]. The soxS promoter responds to superoxide-induced stress, triggered by 2 mM of paraquat (PQ) (S25A Fig), and the dps promoter senses hydrogen peroxide (S25B Fig). When cells are exposed to 5% ethanol, the alcohol significantly induces expression of the soxS-linked GFP signal, albeit to a lesser extent than 2 mM PQ, while expression of dps is not triggered (S27, S28 Data).
When comparing the relative induction levels in the sensitive WT and tolerant envZ*L116P strain, we could not detect any difference in soxS or dps expression profile under ethanol exposure, neither under ROS stress. To revisit the link between ROS stress and ethanol tolerance, the WT and “ferric metabolism attenuated”, tolerant mutants were subjected to a range of PQ and hydrogen peroxide concentrations to determine their resistance profiles towards these ROS (S25C and S25D Fig). This assay revealed that the more ethanol-tolerant mutants, including envZ*L116P, could not systematically cope with oxidative stress better than the WT reference strain. Hence, despite its causal relationship with ethanol tolerance, reduced iron import does not seem to reduce the ROS stress under ethanol tolerance, which is believed to be a hallmark of ethanol tolerance, or provide any protection to ROS stress for the host.
The tolerance-improving, L116P evolved allele also enhances ethanol production
Finally, we aimed to explore the potential of utilizing the envZ*L116P-conferred ethanol tolerance to improve a strain’s ethanol production characteristics. To assess the industrial potential of the L116P allele, the glucose consumption and ethanol production parameters of the envZ*L116P mutant were tracked during an eight-day serial, fed-batch fermentation experiment (S29 Data) and compared to those of the WT. When fed on glucose, the envZ*L116P produced ethanol faster than the ethanol-susceptible, WT ancestor over a 180h period (Figs 8, S26; and S30 Data). Hence, our production results indicate that an improved tolerance phenotype, due to the L116P allele, also stimulates the production rate (+35%), even at ethanol concentrations below the 5% toxicity limit (with final ethanol levels reaching 3.5-4.0 v/v%). Hence, this also suggests that bacterial production performance is even more sensitive to increasing ethanol concentrations than cellular survival.
(A) The absolute amount of ethanol (in g) in the WT (black dots, solid line) and envZ*L116P strains (red diamonds, dotted line) over time. The production data are fitted using a four-parametric, sigmoidal Gompertz equation to extract the production rate value. Error bars indicate the 95% confidence interval and the ribbons surrounding the curves highlight the 95% confidence interval of the Gompertz curves as derived from a Monte-Carlo simulation. (B) Ethanol production rates [in g/h] in the WT and envZ*L116P tolerant mutant. The production rates were statistically compared using a two-sided t-test (n = 6) and the level of significance is indicated as: *, P ≤ 0.05, **, P ≤ 0.01, ***, P ≤ 0.001, ****, P ≤ 0.0001.
The L116P evolved allele solely provides tolerance to ethanol and not to other n-alkanols
Finally, we assessed whether the tolerance-conferring effect of the L116P allele was restricted to ethanol (Fig 9A; S31 Data) or whether its gain-of-function extends beyond other medium-chain, more hydrophobic alcohols (S32 and S33 Data). Therefore, both the WT and the envZ*L116P mutant were also subjected to n-propanol (3.4%, Fig 9B) and n-butanol (1.7%, Fig 9C). While L116P did not influence E. coli’s tolerance to propanol (killing about 90% of the populations in 24h), it did surprisingly render E. coli more sensitive to n-butanol. This indicates that L116P cannot be simply transferred to improve E. coli’s tolerance to other (commercially relevant) alcohols.
Data points indicate the mean and error bars represent the 95% confidence intervals (n = 3).
Discussion
Exposure to high concentrations of ethanol imposes significant stress on microorganisms, leading to cellular intoxication and, ultimately, cell death. Hence, ethanol toxicity is often responsible for reducing microbial ethanol production in industrial settings, especially when the alcohol concentration accumulates over time. Therefore, improving the intrinsic tolerance of microbial production strains is an effective strategy to mitigate the adverse impact of ethanol and ensure continuous ethanol production.
Based on the mutational data from the evolutionary experiment of Swings et al. (2017), we identified an ethanol-tolerance improving mechanism involving a single leucine to proline substitution in EnvZ that significantly altered the balance of kinase-phosphatase activity of the enzyme. This activity is crucial for transducing osmotic stimuli towards OmpR and for engaging an osmotic response [34]. We have prioritized this system because both the sensor and the response regulator were frequently mutated in distinct, parallel-evolved populations. Moreover, the EnvZ-OmpR signal transduction pathway plays a pivotal role in the regulation of multiple membrane-associated transporters, which, according to the GO enrichment performed in this study, are required for E. coli to acquire ethanol tolerance. We assessed the role of the most recurrent adaptive mutation in EnvZ, L116P, and demonstrated that this single amino acid substitution turns E. coli more tolerant towards ethanol. The hypertolerant phenotype as a result of the L116P substitution is attributed to an improved outer membrane stability, due to an increase in total OmpC and OmpF levels, in combination with a reduction in enterobactin biosynthesis and ferric uptake.
Regarding the effect of the periplasmic LL116P substitution, we established that this adaptive mutation, alongside T247R, shifts the enzymatic balance of EnvZ in favor of kinase activity. To the best of our knowledge, the majority of the mutations that affect the kinase or phosphatase reactions have been identified in the catalytic, C-terminal domain of EnvZ [38,48,63,64]. Interestingly, Yaku et al. (1997) were the only authors who discovered Leu residues within the periplasmic domain that upon substitution with Pro residues distort the kinase:phosphatase balance in favor of the kinase activity [65]. Hence, structural rearrangements in the periplasmic domain, as a result of amino acid substitutions, may be propagated towards the cytoplasmic domain causing perturbation of helical backbone stabilization, an altered spatial orientation of H243 relative to other key residues (such as D244), and ultimately a switch in kinase:phosphatase behavior [35].
Clearly, the observed improvement in tolerance is not merely the result of an adapted signal transduction pathway but results from the differential expression of downstream-regulated genes involved in the stress response. First, we established that porins (OmpC and OmpF) have a profound impact on the structural integrity of the outer membrane as adjusting their expression level, through EnvZ, mitigates severe cell envelope stress [66]. Eliminating ompC reduces E. coli’s survival under ethanol stress, as evidenced by the appearance of morphological defects that become more pronounced with prolonged alcohol exposure. Indeed, Zhang et al. (2018) previously suggested a role of OmpC in ethanol tolerance in E. coli. Moreover, Sen et al. (2023) reported that alcohols cause a strong reduction in outer membrane OmpC levels, explaining the toxic effect of these chemicals on E. coli [67]. While we observed that eliminating ompC sensitizes cells, we show here that the diminished ethanol tolerance can be compensated for by introducing an additional genomic copy of ompF, but not ompA, tolC, or lamB. This suggests that not the OmpC-OmpF porin composition, but rather the global OmpC-OmpF porin abundance, defines a cell’s ethanol tolerance. Importantly, this rule applies only to the OmpC and OmpF porins, indicating that these outer membrane proteins are functionally similar in providing ethanol tolerance, whereas the other candidates—OmpA, TolC, and LamB—are functionally distinct. We hypothesize that OmpC and OmpF engage in specific physical interactions with the lipids of the outer membrane, stabilizing the cell envelope, and that these interactions cannot be simply replicated by any other outer membrane protein. Moreover, the total porin content in the envZ*L116P evolved strain is higher than that of the sensitive WT strain, indicating that a higher OM porin content reinforces the cell envelope against ethanol exposure. In contrast, lowering the total porin content in the OM by deleting ompC does clearly affect the tolerance of the envZ*L116P mutant, but does not completely abolish it. Hence, the beneficial effect of envZ*L116P is not solely attributed to an increase in collective OmpC, and OmpF porin levels. Therefore, we investigated the importance of other downstream-regulated genes of OmpR, resulting in the identification of TolC and LpcA as key determinants for ethanol tolerance. Interestingly, unlike ompC, eliminating these new targets does completely annihilate the hypertolerant phenotype of envZ*L116P. Together with AcrAB, the TolC OM channel constitutes a tripartite efflux pump that is well known to cause antibiotic resistance by expelling antimicrobial drugs [68,69]. Previously, enhanced expression of tolC has been shown to confer tolerance to cyclohexane [70]. Here, we established that TolC fulfills a crucial function in a cell’s resilience to ethanol toxicity, a polar alcohol. While it is well established that overexpressing efflux pumps helps to confer tolerance to terpenoids (such as α-pinene or limonene) [71], our data do not provide evidence that the poor tolerance of tolC deletion strains is explained by the lack of any ethanol efflux system. Indeed, eliminating tolC directly increases OM permeability dramatically, suggesting severe impairment of cell envelope integrity, while knocking out acrB does not affect the tolerance phenotype at all. Moreover, simple alcohols such as ethanol are not recognized as substrates by AcrB, indicating that solvent extrusion via this efflux pump is not relevant to ethanol tolerance [72]. The second identified OmpR-regulated gene, called lpcA, encodes a sedoheptulose-7-phosphate isomerase that is involved in LPS synthesis. Woodruff et al. (2013) previously reported that overexpression of lpcA improves growth under ethanol stress, while Reyes et al. (2013) demonstrated that lpcA is upregulated in n-butanol-adapted E. coli cells [73,74]. Interestingly, none of the other OmpR-regulated genes influenced the tolerance level of the envZ*L116P mutant, although some of them were previously linked to ethanol tolerance, including the cad acid stress-resistant operon and the osmolyte synthesis repressor BetI [27,28]. Hence, our study demonstrates that EnvZ-OmpR-mediated adaptation to ethanol stress primarily targets OM stabilization, with the LPS synthesis enzymes LpcA and OM-embedded OmpC porin and TolC channel playing a key role. Eliminating each of these membrane-stabilizing genes did affect the stability and permeability of the OM. This observation is in agreement with previous reports highlighting the importance of interactions between LPS and outer membrane proteins in maintaining cell shape and envelope integrity under both normal physiological and stressful conditions [53,54,75]. Deleting lpcA results in incomplete LPS structures that may be unable to participate in cation-mediated intermolecular interactions at the OM, potentially leading to (partial) collapse of this cellular component [76,77]. Hence, weakening of the OM structure, caused by a reduction of key OMPs or inconsistent LPS biosynthesis, is expected to increase the susceptibility of E. coli to ethanol.
In addition to the OmpR-targeted deletion method, we applied an unbiased proteomics-driven approach that identified ferric transport and homeostasis as key factors in the envZ*L116P-associated tolerance phenotype. Reduced expression of enterobactin biosynthesis genes and ferric transporters underlies the envZ*L116P-associated tolerance. When these primary targets were knocked out in the sensitive WT strain, tolerance was improved. The underlying reason why reduced ferric uptake would contribute to ethanol tolerance remains unclear, although it does not seem to be related to reduced ROS stress under ethanol exposure
Recently, Machas et al. (2021) discovered that ompR is also implicated in the transcriptional response of E. coli under exposure to styrene [78]. The latter suggests that modifying the EnvZ-OmpR signal transduction, by substituting key residues in the osmosensor, for instance, may be a suitable strategy to extend microbial tolerance towards a broad range of industrially-relevant hydrophilic (e.g., alcohols) or hydrophobic (e.g., styrene) solvents and chemicals. However, the tolerance-conferring L116P mutation in envZ specifically enhances the host’s resilience to ethanol—the stressor under which this adaptive allele was isolated during an adaptive laboratory evolution experiment—and does not confer cross-tolerance to other, more hydrophobic alcohols such as n-propanol and n-butanol. Nonetheless, the EnvZ*L116P mutant sensor still holds promising industrial relevance as the resulting improvement in ethanol tolerance is also linked to superior fermentation production kinetics. This proof-of-concept demonstrates that tolerance engineering is relevant for microbial production, making sustainable, microbe-based manufacturing more attractive.
Materials and methods
Note: All R scripts (in quarto format) and Fiji (imageJ) scripts that were used to statistically process the data and quantitatively analyze pictures from spot assays are provided at our GitHub page.
Prioritization of the EnvZ-OmpR TCS from the mutation dataset
A mutational dataset was extracted from the study of Swings et al. (2017) to identify the most predominant pathways that conferred ethanol tolerance in the ethanol-adapted E. coli populations [30,31]. This dataset is provided as S1 Data. The entire analysis workflow is described in the R script, Fig 1-IdentifyEthanolTolerancePathways.qmd. First, a mutation matrix was constructed for each of the 16 evolved E. coli populations (labeled from HT1–16) indicating the absence (0) or presence (1) of a certain mutation within the population. These matrices enabled summing the binary matrices to obtain the total number of mutations within a gene across all 16 evolved populations at any ethanol level, called the global pattern. Hence, if a gene was mutated multiple times, all evolved mutations within the same gene contributed to the global pattern. To retrieve the most relevant mutations (among the > 2500 unique mutations), we arbitrarily set the “mean occurrence” threshold at 1, meaning that a gene should carry (on average) one mutation to be retained for follow-up GO enrichment analysis. In the script, the corresponding entrez ID is linked to a specific locus or gene name since the GO enrichment tool, GOfuncR (https://www.bioconductor.org/packages/release/bioc/html/GOfuncR.html), requires this input format. Linking the entrez ID to the locus or gene names was performed with the ‘bitr’ function in the clusterprofiler package (https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html) and the readily available “org.EcK12.e.g.,db” (https://www.bioconductor.org/packages/release/data/annotation/html/org.EcK12.eg.db.html) genome-wide annotation. The script behind these operations can be found in a separate quarto file: Fig 1-ExpandEntrezID.qmd. Thereafter, the enriched GO terms from the subset of genes that exceeded the mean occurrence threshold were determined using the online ShinyGO application (http://bioinformatics.sdstate.edu/go/) and, in parallel, the GOfuncR Bioconductor package. A graphical representation of the ShinyGO output is provided as supplementary figure (S2 Fig), while the output of GOfuncR is summarized in an Excel file, named Fig 1-GO_summary.xlsx (at the GitHub repository, OutputFiles folder). Both GO enrichment approaches prioritized similar GO terms, related to signaling, kinase (protein phosphorylation) activity, nucleotide binding, and predominantly restricted to the cell envelope. Therefore, we focused the analysis pipeline on two-component systems since these signal transduction pathways are commonly associated with these GO terms. To study these TCSs in more detail, we first composed a list of all known TCSs in E. coli based on the information provided in the KEGG database (https://www.genome.jp/kegg-bin/show_pathway?ko02020, S2 Table). Then, a corrected frequency score ( was assigned to each TCS-associated mutation (
) based on its occurrence (
) throughout all parallel-evolved populations and the sequencing coverage (
, i.e., frequency, in the population-wide WGS data:
in which represents the total number of parallel-evolved populations (16). These results are provided in Fig 1-frequencyTableTCS.xlsx (at the GitHub repository, OutputFiles folder). Next, the corrected frequency score was transformed into a global frequency score
by summing the previously-calculated frequency scores of all mutations (
) within the TCS of interest to the number of ethanol levels considered (
=7).
A graphical representation of the (global) frequency scores for each TCS is provided in S3 Fig. Since the ShinyGO also identified membrane-related transporters to be important for ethanol tolerance, we looked into the regulons of the TCSs to prioritize those that have the highest impact on the expression of membrane-related transporters. Therefore, a list of all downstream-regulated genes for each response regulator was retrieved from the RegulonDB database (https://regulondb.ccg.unam.mx/datasets, S4 Table). Next, all downstream-regulated genes were GO annotated using the AnnotationDbi package to count for every TCS the number of genes encoding membrane-associated transporters () using the keywords: “membrane” as cellular component and “Transport” as biological process. Finally, we assigned an overall priority score (
for each TCS, which is defined as the product of the global frequency scores (
) and the number of membrane-related transporters (
):
Strains and culture conditions
The E. coli strain BW25113 served as the reference, WT strain in all experiments. In case deletion mutants were included in tests, these strains were directly retrieved from the Keio library when the effect was studied in presence of WT envZ (Table 1) [79]. Otherwise, the FRT-KmR-FRT cassette, replacing the gene of interest, was recombined in the envZ*L116P mutant background to examine the combined effect of the L116P substitution and the gene deletion on ethanol tolerance (method discussed in the next section) (Table 1). All strains were grown overnight in Lysogeny broth (LB) in an orbital shaker at 200 rpm and 37°C. As an exception, strains harboring the heat-sensitive CRISPR-FRT or pKD46 plasmids were grown at 30°C (Table 2).
Construction of deletion mutants
A recombination method based on Datsenko and Wanner [82] was applied to transfer the FRT-KmR-FRT cassette, which replaces the gene of interest in a specific Keio library mutant, towards the envZ*L116P mutant. Therefore, the KmR-cassette was PCR amplified making sure that 200–500 bp homology overhangs were included at both sides. The Q5 PCR mix was composed according to the supplier’s protocol (New England Biolabs) and the (standard desalted) primers were ordered at Integrated DNA Technologies (Belgium) (primer sequences are listed in Table 3). Afterward, the product was checked on a 0.7 (w/v)% agarose gel, PCR purified using the DNA Clean & Concentrator-5 (Zymo Research) and, finally, quantified using the NanoDrop ND-1000. To facilitate recombination, the pKD46 plasmid that expresses the λ-Red recombination system was introduced into the envZ*L116P mutant strain (Table 2). This plasmid was extracted from an E. coli TOP10 culture using the NucleoSpin Plasmid kit (Macherey-Nagel), following the supplier’s protocol. Next, around 100 ng pKD46 was introduced into the envZ*L116P mutant by means of chemical transformation as described by Green and Rogers (2013) [86]. When the transformation was successful, cells were incubated overnight at 30°C in the presence of 100 µg/mL ampicillin (Ap100) diluted in fresh LB with Ap100 and 0.2% L-arabinose to induce expression of λ-red genes. After 5h, cells were washed four times in 10% glycerol prior to electroporation with 200 ng PCR-amplified and purified FRT-KmR-FRT oligo (BioRad Pulser Xcell, 1 mm cuvette, 1.8 kV, 5 ms). Afterward, electroporated cells were recovered for 1h at 30°C and plated out on selective LB agar supplemented with 40 µg/mL kanamycin (Km40) and Ap100 for overnight incubation at 30°C. The next day, integration of the cassette at the intended locus was verified by PCR and a single correct clone was streaked on Km40-rich LB agar to cure the pKD46 plasmid at 42°C. After overnight incubation, loss of pKD46 was checked by spotting a couple of colonies on Ap100. Finally, the KmR-cassette was removed using the pCP20 plasmid which was introduced by electroporation (Table 2) [79]. Then, a couple of colonies were spotted on agar, supplemented with Km40, to verify removal of the KmR-cassette. Similarly to pKD46, pCP20 was cured of those clones that appear to be Km-sensitive at 42°C and loss of the pCP20 plasmid was again checked on Ap100.
Reconstruction of the L116P, F390L, and T247R alleles in the WT BW25113 E. coli strain
The L116P mutation was introduced in the BW25113 strain, according to the CRISPR-FRT method as described in Swings et al. (2018) [80]. Briefly, the rescue oligo was retrieved from the HT11 evolved E. coli population, by amplifying the envZ locus including 500 bp up-and downstream from the mutated evnZ gene to facilitate recombination [30,31]. Thereafter, this rescue oligo was electroporated into the ΔenvZ::FRT-KmR-FRT Keio clone harboring the pCas9 and pKDsgRNA-FRT plasmid (Table 2) [80]. In contrast to L116P, the F390L and T247R alleles were not readily available in our ethanol-evolved E. coli collection. Hence, these rescue oligos needed to be produced de novo using Splicing by Overlap Extension PCR (SOE-PCR) [87]. Therefore, additional primers were designed in which the overlapping 20–25 nt tails encompass the envZ codon to be modified. First, the fragments upstream and downstream of the desired substitution were individually amplified using the Q5 polymerase protocol (NEB). After purification, both fragments were combined in equimolar concentration (10 nM) in a 25 µL Q5 reaction mixture without additional primers. In this PCR step, 10 cycli were performed in which the PCR products were allowed to dimerize at an optimal annealing temperature, dictated by the Tm calculator tool (https://tmcalculator.neb.com/#!/main). In the second round of PCR, 25 µL Q5 mix, including the outermost primers, was supplemented to the reaction mixture and 30 cycles were run to enrich the dimerized, full-size PCR product. When the SOE PCR was completed, the product was purified using the Wizard SV Gel and PCR Clean-Up System (Promega). This gel-purified product (ca. 200 ng) was again electroporated into ΔenvZ::FRT-KmR-FRT strain to finish the CRISPR-FRT protocol [80]. Once these mutants were confirmed to be Km-sensitive, the envZ gene was amplified and sent for Sanger sequencing to confirm the presence of the desired SNP (Macrogen sequencing service). Finally, both CRISPR plasmids were either cured using an elevated growth temperature (42°C), in case of the pKDsgRNA-XXX plasmids, or using a Cas9-aided approach to remove the pCas9 plasmid according to Reisch and Prather (2015) and Swings et al. (2018) [80,81].
Exchanging ompC for ompF, ompA, tolC, or lamB in the WT or envZ*L116P strains
Similarly to the previous section, the CRISPR-FRT protocol was implemented here either in the ΔompC::KmR or envZ*L116P ΔompC::KmR, both carrying the pCas9cr4 and pKDsgRNA-FRT plasmids. Each of the rescue oligos for ompF, tolC, ompA, and lamB was prepared using SOE-PCR with the primer pairs summarized in Table 3. Restoring the ompC deletion (using the method as explained in the previous section) introduced a second copy of either ompF, tolC, ompA, or lamB under control of the original ompC promoter.
Growth and survival assays under alcohol stress
Strains were grown overnight at 37°C in 5 mL glass tubes and, the next day, the optical density at 595 nm (OD595) was calibrated to 0.2 (ThermoFisher, Genesys 10S UV-Vis spectrophotometer) for every individual strain. Next, 2.5 mL of the calibrated cell suspension was added to each alcohol-rich flask (with screw caps to prevent evaporation) and incubated at 37°C, corresponding to a cell density of 106 (which is about a 100-fold dilution). Every hour, the OD595 was recorded (in the first 12h interval) to retrieve the growth dynamics and survival was quantified by means of dilution plating at fixed time points (0, 5, 10, and 24h or 0, 1, 3, 5, 10, and 24h). OD-based growth curves were collected by subtracting the OD595 value of 5% ethanol-rich LB from a culture’s OD595 at any given time point. Growth parameters (including growth rate, lag time, and initial and final density) were retrieved from a spline fit, using QurvE, onto the log-transformed OD-data [41]. Finally, either the fold-increase in OD or the growth rate of the different strains was statistically compared to the WT using a one-way ANOVA with Dunnett’s post-hoc test (using the WT E. coli or envZ*L116P as a reference, significance level = 0.05). In addition, a strain’s survival fraction over 24h to 5% ethanol (routinely) or either 3.4% propanol or 1.7% butanol exposure was calculated as the ratio of the CFU count at a certain time point vs. the initial cell count. Cells were plated using the EddyJet2 spiral plater (IUL instruments) and the cell counts were automatically determined using the Flash & Grow device (IUL instruments). The survival of each mutant over the entire incubation period was compared to the reference strain using (Generalized) linear (mixed-effects) models (Time [h] x strain) with a Dunnett’s post-hoc test. Multiple models were considered and the model with the lowest Akaike Information Criterion (AIC) was selected as the most appropriate one.
Competition tolerance assay
Construction of differentially labeled strains.
Two fluorescent markers expressed under a strong, constitutive, artificial promoter were inserted at the lacZ locus in order to discriminate the competing strains. The green fluorescent sfGFP (super-folder GFP) label, was under the control of a synthetic, constitutive promoter (iGEM part BBa_K314100, designed by T. Huber and J. Spiegel (2010)) on the plasmid pSB1C3 (gift from Nicholas Coleman, University of Sydney) [88,89]. The RFP (red fluorescent protein), linked to a synthetic (Biofab) promoter, was derived from the pUltraRFP-KM plasmid [90,91]. Both promoter-fluorescent constructs were amplified from their corresponding plasmid and integrated into the E. coli strain of interest at the lacZ locus by means of SOE-PCR and homologous recombination (Table 1). Once the fluorescent constructs were successfully integrated, the KmR marker was eliminated using the pCP20 plasmid as described by Cherepanov and Wackernagel (1995) (Table 2) [83].
Setup of the differentially-labeled, pairwise tolerance assay.
For the survival-based competition assay, the fluorescently-labeled strains were individually grown overnight in LB and OD595-calibrated to 0.2. Then, 2.5 mL of each fluorescent strain was mixed into 50 mL 5% LB medium. For determining the survival fraction, we used the same routine survival assay, but colonies were counted under the Illumatool Tunable Lighting System LT-9500 (Lightools Research) to discriminate the red and green fluorophores. For microscopy-based analysis, samples were collected every 2h (until 8h after ethanol was added to the flasks) and imaged using a Nikon Ti-E inverted microscope equipped with a Qi2 CMOS camera and 100x objective at the phase-contrast (exposure: 100 ms) and fluorescent (green: 30 ms exposure and red: 900 ms) channels. The obtained multi-channel images were processed using the MicrobeJ plugin (https://brunlab.com/research/microbej/) in Fiji to extract the morphological cell parameters: cell area [µm²] and circularity [92]. In addition, each cell was characterized by fluorescence to retrieve its corresponding genotype. For statistics on morphological data, the mutant groups were first tested for normality using the Shapiro-Wilk test in R [93]. Since the datasets appeared not normally distributed, the non-parametric Mann-Whitney test was applied at each time point [94]. Finally, the P-values were adjusted for multiple comparison testing using the Holm correction [95].
Robot-assisted high-throughput tolerance assay.
The OmpR deletion collection was created using pKD46-mediated homologous recombination to transfer the Km-cassette from the corresponding Keio clone to the envZ*L116 mutant strain, as described earlier. The resulting E. coli mutants and the envZ*L116P strain (Table 1) were, after calibration of cell density, treated with ethanol using the same protocol as in the standard growth and survival assay. At the initial time point (t0), 20 µL was sampled from each individual Erlenmeyer flask and transferred to a 96-well plate to prepare a 3-fold dilution series using the Opentrons OT-2 liquid handling robot. Then, the robot spotted 5 µL of each dilution onto a rectangular agar plate (Nunc Omnitray, Thermo Scientific). The droplets were allowed to dry before the plate was incubated overnight at 37°C. At 12h, the spot quantification procedure was repeated. The next day, the individual colonies within each spot were counted and the survival of each deletion mutant was expressed relative to the envZ*L116P reference strain that did not miss any of the genes that belong to the OmpR regulon. To identify gene deletions that were responsible for an increase/decrease in survival, we defined a linear mixed effects model (with the screened batch ID as random factor) and compared it to the envZ*L116P reference strain using a Dunnett’s post-hoc test.
Agar spot assay.
After ON incubation at 37°C, each strain was OD595 calibrated to 1. Then, a 10-fold dilution series (ranging from 100 until 10-6) was prepared for each strain. Finally, 5 µL droplets were sampled from each dilution and spotted on 5 (v/v)% ethanol-rich LB agar (120 mm x 120 mm square plates, Greiner Bio-one). The agar plates were incubated at 37°C for 24h and, the next day, a picture was taken from both plates using the Canon EOS 1300D camera. This procedure was repeated four times independently. This procedure was repeated four times independently.
Afterward, the spot area and intensity were quantified using three dedicated Fiji-based scripts that are supplemented separately [96]. Briefly, a grid was superimposed onto the entire picture using the script ImageSplit.ijm. Using the grid as a guide, the entire picture was split into multiple sub-images, containing exactly one bacterial spot. These sub-images were stored separately for more detailed analysis in terms of spot area and pixel intensity. In case of spot area quantification, the sub-images were processed using either the SpotAnalyzer_globalTreshold.ijm script in which a particle analyzer calculated the spot area (with a surface threshold of 100 px²). In case of pixel intensity, the histogramCalculator.ijm script extracted a histogram for each sub-image in which the mean pixel intensity was included. Based on these results, the mean area and pixel intensity was determined for each bacterial spot within the entire image. Finally, a Fiji-generated image was created based on the mean spot area and intensity parameters to visually summarize the results of the four experimental repeats by means of the spotSimulator_Area&Grey_whiteBckgrd.ijm script.
Fluorescent reporter expression assay.
To estimate the bifunctional activity of the EnvZ-OmpR TCS, the promoter fusion, proposed by Zaslaver et al. (2006) [84], was applied. The pUA66 PompC-gfpmut2 and pUA66 PlacZ-gfpmut2 constructs were directly available from the Uri Alon library (https://www.weizmann.ac.il/mcb/UriAlon/download/downloadable-data), whereas the pUA139 PompF-gfpmut2 was constructed ourselves (Table 2). Therefore, the ompF promoter (PompF) was amplified from the BW25113 genome with primers including either the XhoI or BamHI recognition sites. This promoter was inserted in front of the gfpmut2 on the pUA139 plasmid, following the restriction-ligation protocol described by Zaslaver et al. (2006). The ligated construct was transformed into chemically competent E. coli TOP10 cells. Afterward, correct integration of PompF was verified by colony PCR and sequencing. Once all three constructs were available, these three kinds of plasmids were introduced into the strains of interest and the resulting transformants were inoculated in Km-rich 5 mL LB tubes at 37°C for further testing. The next day, cells were OD595 calibrated to 0.2 and 2.5 mL of the cell suspension was added to each flask, filled with 50 mL 2-fold diluted and Km40-supplemented LB. We preferred to use 2-fold diluted growth broth to minimize background fluorescence. In case of the PlacZ-gfpmut2 constructs, 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) was additionally mixed into the 50 mL flasks. Every hour (up to 10h), a 150 µL sample was withdrawn from the flasks to record the fluorescent intensity (excitation: 480nm and emission: 510 nm) and OD595 in 96 well, F-bottom black microplates (Greiner Bio-one) using the Synergy Mx (Biotek) reader. The relative fluorescence intensity, , at time point (t) was expressed as:
in which represents the optical density as measured at 595 nm and
, and
represent cell sample and background, respectively. The ompC and ompF expression in the WT and envZ*L116P E. coli strains at t = 10h was quantified in relationship to the
value from the lacZ control plasmid in each strain by dividing the
linked to ompC or ompF by the
associated to lacZ.
To measure the response of the EnvZ-OmpR TCS on osmotic stress, induced using PEG6000, or ethanol stress, the WT and envZ̈L116P were cultured in 400 µL 2-fold diluted LB in a 96-deepwell plate setup (U-bottom Nunc plates, Thermo Scientific) that was shaken on an incubation platform (Heidolph Titramax 1000, 1200 rpm, 37°C) for 3h. PEG6000 was added in concentrations of 5 or 10 (w/v)% and ethanol in 5 and 10 (v/v)%. After incubation, cells were collected by centrifugation and resuspended in 400 µL phosphate-buffered saline (PBS) (137 mM sodium chloride, 2.7 mM potassium chloride, 10 mM disodium hydrogen phosphate, and 1.8 mM monopotassium phosphate, pH 7.4). If needed, the PBS-dissolved cells were further diluted in PBS to reach a cell density of approx. 106 CFU/mL and the samples were analyzed using the CytoFLEX S flow cytometer (Beckman-Coulter). Cells were discriminated from background signals using forward (FSC) and side scatter (SSC) gating and fluorescence intensity of the GFPmut2 marker was recorded at the FITC channel. The fluorescent histograms were retrieved from the Beckman-Coulter CytExpert v2.3 software and the histograms of the different conditions per strain were overlaid. For statistical comparisons, we constructed a linear model with the ompC:ompF expression ratio, expressed as , vs. the stress condition (w/v% PEG6000 or v/v% ethanol), in which
and
represents the mean fluorescence intensity at the FITC channel for ompC or ompF, respectively. When the slope was significantly different from 0 (using a t-test), we considered the EnvZ osmosensor to respond to the stressor.
Determination of OM permeability using the N-phenylnaphthalen-1-amine (NPN) uptake assay.
Strains were incubated ON in LB tubes at 37°C in an orbital shaker (Table 1). The next day, a 12 mL bacterial stock of an OD595 0.5 was prepared in LB and three glas s tubes were filled with 3 mL calibrated cell culture. To validate that OM permeabilization due to polymyxin B (PxB) treatment or ethanol could be detected, the cell cultures were incubated in presence of 0, 3, 5, 8, 10, and 15% ethanol or 0, 0.25 (1xMIC), 1.25 (5xMIC), or 2.5 µg/mL (10xMIC) PxB. Once the assay was confirmed to be suitable to quantify OM permeability, OM permeability of the strains was assayed in 0 or 5% ethanol. After 5h incubation, each of the strains was again calibrated to an OD595 of 0.5 in 5 mM HEPES buffer (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, Sigma-Aldrich, pH = 7.2), according to Defraine et al. (2018) [84]. Afterward, 150 µL of HEPES-dissolved cells were mixed with 50 µL 40 µM N-phenylnaphthalen-1-amine (NPN, TCI Europe, dissolved in 5 mM HEPES) and the fluorescence (λex = 504, λem = 523 nm) and absorbance (OD595) were recorded using the Synergy Mx multimode reader.
The NPN uptake value of each strain under ethanol stress was defined as the NPN-derived fluorescence intensity vs. the optical density of the HEPES-resuspended E. coli culture at 595 nm. Based on these values, two statistical tests were conducted. First, a linear model was built to correlate the NPN uptake under 5% ethanol with the inherent NPN uptake values of the same strain. For ompC, tolC, and lpcA, the impact of deleting the target genes was corroborated with a Mixed Effects Linear Model (background x deletion) linked to a Dunnett’s post-hoc test with either the WT or envZ*L116P strain as reference. Finally, the relation between survival and NPN uptake was demonstrated using a Spearman correlation test.
Whole-genome proteomics.
Sample preparation.
Overnight bacterial cultures were diluted 100-fold and inoculated in 50 mL LB for ca. 3.5h until an OD595 of 0.8 was reached. For each strain, 15 mL culture was collected and washed three times in PBS. Then, the cell pellets were frozen in liquid nitrogen for shipment to the VIB proteomics core (Ghent, Belgium). In the VIB proteomics core facility, the cell pellets were homogenized in 100 µl lysis buffer containing 5% sodium dodecyl sulfate (SDS) and 50 mM triethylammonium bicarbonate (TEAB), pH 8.5. Next, the resulting lysate was transferred to a 96-well PIXUL plate and sonicated with a PIXUL Multisample sonicator (Active Motif) for 5 minutes with default settings (Pulse 50 cycles, PRF 1 kHz, Burst Rate 20 Hz). After centrifugation of the samples for 15 minutes at maximum speed at room temperature (RT) to remove insoluble components, the protein concentration was measured by bicinchoninic acid (BCA) assay (Thermo Scientific) and from each sample 100 µg of protein was isolated to continue the protocol. Proteins were reduced and alkylated by addition of 10 mM Tris(2-carboxyethyl)phosphine hydrochloride and 40 mM chloroacetamide and incubation for 10 minutes at 95°C in the dark. Phosphoric acid was added to a final concentration of 1.2% and, subsequently, samples were diluted 7-fold with binding buffer containing 90% methanol in 100 mM TEAB, pH 7.55. The samples were loaded on the 96-well S-Trap plate (Protifi), placed on top of a deepwell plate, and centrifuged for 2 min at 1,500 x g at RT. After protein binding, the S-trap plate was washed three times by adding 200 µl binding buffer and centrifugation for 2 min at 1,500 x g at RT. A new deepwell receiver plate was placed below the 96-well S-Trap plate and 50 mM TEAB containing trypsin (1/100, w/w) was added for digestion overnight at 37°C. Using centrifugation for 2 min at 1,500 x g, peptides were eluted in three times, first with 80 µL 50 mM TEAB, then with 80 µL 0.2% formic acid (FA) in water and finally with 80 µL 0.2% FA in water/acetonitrile (ACN) (50/50, v/v). Eluted peptides were dried completely by vacuum centrifugation. Samples were dissolved in 100 µL 0.1% TFA in water/ACN (98:2, v/v) and desalted on a reversed phase (RP) C18 OMIX tip (Agilent). The tip was first washed 3 times with 100 µL pre-wash buffer (0.1% TFA in water/ACN (20:80, v/v)) and pre-equilibrated 5 times with 100 µL wash buffer (0.1% TFA in water) before the sample was loaded on the tip. After peptide binding, the tip was washed 3 times with 100 µL of wash buffer and peptides were eluted twice with 100 µL elution buffer (0.1% TFA in water/ACN (40:60, v/v)). The combined elutions were transferred to HPLC inserts and dried in a vacuum concentrator.
LC-MS/MS analysis.
Peptides were re-dissolved in 20 µL loading solvent A (0.1% trifluoroacetic acid in water/acetonitrile (ACN) (99.5:0.5, v/v)) of which 2 µL of the sample was injected for LC-MS/MS analysis on an Ultimate 3000 Pro Flow nanoLC system in-line connected to a Q Exactive HF mass spectrometer (Thermo). Trapping was performed at 20 μL/min for 2 min in loading solvent A on a 5 mm trapping column (Thermo Scientific, 300 μm internal diameter (I.D.), 5 μm beads). The peptides were separated on a 250 mm Aurora Ultimate, 1.7 µm C18, 75 µm inner diameter (Ionopticks) kept at a constant temperature of 45°C. Peptides were eluted by a non-linear gradient starting at 0.5% MS solvent B reaching 26% MS solvent B (0.1% FA in acetonitrile) in 75 min, 44% MS solvent B in 95 min, 56% MS solvent B in 100 minutes followed by a 5-minute wash at 56% MS solvent B and re-equilibration with MS solvent A (0.1% FA in water). The mass spectrometer was operated in data-independent mode, automatically switching between MS and MS/MS acquisition. Full-scan MS spectra ranging from 375-1500 m/z with a target value of 5E6, a maximum fill time of 50 ms and a resolution at of 60,000 were followed by 30 quadrupole isolations with a precursor isolation width of 10 m/z for HCD fragmentation at an NCE of 30% after filling the trap at a target value of 3E6 for maximum injection time of 45 ms. MS2 spectra were acquired at a resolution of 15,000 at 200 m/z in the Orbitrap analyser without multiplexing. The isolation intervals ranging from 400– 900 m/z were created with the Skyline software tool. The polydimethylcyclosiloxane background ion at 445.120028 Da was used for internal calibration (lock mass) and QCloud [97,98] was used to control instrument longitudinal performance during the project.
Data analysis.
LC-MS/MS runs of all samples were searched together using the DiaNN algorithm (version 1.8.1), library free. Spectra were searched against the E. coli protein sequences in the Uniprot database (database release version of January 2024), containing 4,403 sequences (www.uniprot.org), supplemented with the universal protein contaminant database (database release version of 2023_02), containing 381 sequences [99]. Enzyme specificity was set as C-terminal to arginine and lysine, also allowing cleavage at proline bonds with a maximum of two missed cleavages. Variable modifications were set to oxidation of methionine residues and acetylation of protein N-termini while carbamidomethylation of the cysteine residues was set as fixed modifications. Mainly default settings were used, except for the addition of a 400–900 m/z precursor mass range filter and MS1 and MS2 mass tolerance was set to 15 and 20 ppm respectively. Further data analysis of the shotgun results was performed with an in-house script in the R programming language, version 4.2.2. Protein expression matrices were prepared as follows: the DIA-NN main report output table was filtered at a precursor and protein library q-value cut-off of 1% and only proteins identified by at least one proteotypic peptide were retained. After pivoting into a wide format, iBAQ intensity columns were then added to the matrix using the DIAgui’s R package (https://rdrr.io/github/mgerault/DIAgui/man/DIAgui.html) get_IBAQ function. PG.Max LFQ intensities were log2-transformed and replicate samples were grouped. Proteins with less than three valid values in at least one group were removed and missing values were imputed from a normal distribution centered around the detection limit (package DEP, [100]) leading to a list of 2,728 quantified proteins in the experiment, used for further data analysis. To compare protein abundance between pairs of sample groups (WT vs. envZ*L116P; envZ*L116P vs. envZ*L116P ompF’/ompF; WT vs. envZ*L116P ompF’/ompF sample groups), statistical testing for differences between two group means was performed, using the package limma [101]. Statistical significance for differential regulation was set to a false discovery rate (FDR) of <0.05 and fold change of >4- or <0.25-fold (|log2FC| = 2). Results are provided in S25 Data.
Significantly differentially expressed proteins in the envZ*L116P vs. the WT group were retained for further GO enrichment analysis. Significantly enriched GO terms were identified using three similar approaches based on the Bioconductor packages, GOfuncR and clusterProfiler, and the online-accessible ShinyGO application (https://bioinformatics.sdstate.edu/go/). All three approaches prioritized the same key GO terms. Next, each protein within a sample was normalized using the available log2.PG.MaxLFQ (log2-transformed maximal label-free quantification) values as:
This approach enables comparing the relative protein abundances, represented by their values, between the WT, envZ*L116P, and envZ*L116P ompF’/ompF strains using a one-way ANOVA with Tukey’s post-hoc test.
Quantification of the ROS response under ethanol, paraquat, and hydrogen peroxide stress.
Before the assay, the ROS reporter plasmids (pZE1-Pdps and pZE1-PsoxS) were introduced into the WT and envZ*L116P strains using heat shock transformation. Strains were cultured overnight in Ap100-rich LB medium and calibrated to OD595 of 0.2. In case of ethanol, the strains were diluted in fresh 5% ethanol-rich LB, just as in the growth and survival assay. When paraquat (PQ) or hydrogen peroxide (H2O2) was used, the same relative volume of calibrated cells was added to 5 mL LB, containing 2 mM PQ or H2O2. In the non-challenged condition, ethanol or the ROS molecules were replaced with demi-water. The fluorescence intensities of the promoter-linked GFP reporter were recorded using the CytoFLEX S flow cytometer at the start (t = 0h) and after 1h of exposure to ethanol, PQ, or H2O2. Then, we expressed the log2(induction) for each reporter plasmid construct and condition as:
in which , represents the FITC channel,
, the forward scatter,
, the challenged condition, and
, the non-challenged condition. The induction levels were pairwise compared using t-test statistics.
EC50 determination of strains for paraquat and hydrogen peroxide.
After overnight incubation, strains were calibrated to OD595 0.2 and 400-fold diluted in fresh LB. The cells were challenged to a 2-fold dilution gradient of H2O2 (ranging from 8 mM to 0.125 mM) or PQ (ranging from 10 to 0.156 mM). In each assay, non-treated controls were included as well. Each well of the 96-well plate was covered with silicone oil (Fisher scientific) to prevent evaporation and the plate was inserted in a Cerillo Stratus Microplate Reader. The OD595 was periodically recorded every 15 min for 48h at 37°C. The OD-data were processed using spline models in QurvE [41] and the growth rate was extracted. This growth parameter served as input for dose-response curve analysis (using the built-in drc functionality in QurvE) from which the EC50 values could be retrieved. The EC50-values of the different strains were statistically compared to the WT using a generalized mixed effects linear model with Dunnett’s post-hoc test.
Fermentation setup for ethanol production
To promote ethanol production in E. coli, the Cas9 sequence at pCas9 was replaced with the Z. mobilis pet operon genes (including pdcZm and adhBZm), resulting in the pEtOH plasmid. Therefore, the pCas9 plasmid was PCR linearized and the pdcZm - adhBZm operon was amplified. Both PCR products were equimolarly mixed for Gibson assembly using the NEBuilder HiFi DNA Assembly Master Mix. The resulting pEtOH construct was transformed into chemocompetent E. coli TOP10 cells, plasmid purified, and sent for sequencing (Macrogen). Finally, the pEtOH plasmid was introduced into the WT and envZ*L116P strains using heat-shock transformation.
Before fermentation, the pEtOH-carrying strains were inoculated in LB tubes, enriched with Cm30, and incubated overnight at 30°C. The next day, the ON cultures were diluted 100-fold into flaks with fresh Cm30-enriched LB medium. After 2h of incubation, the Z. mobilis ethanol-producing enzymes were induced with 100 ng/mL aTc for ca. 20h. Thereafter, cells were collected by centrifugation, the supernatant was discarded and the pellet was resuspended in 50 mL fermentation broth (Table 4, step I). Before the flasks were sealed off with an airlock system (Brouwland), a 1 mL sample was taken for HPLC analysis. During the next days, the addition of fermentation broth and sampling was repeated 3-times with 24h in between each cycle. After the final step, the fermentation process was allowed to continue at 30°C for a total period of 185h. This fermentation setup was repeated six times for each strain.
The ethanol and glucose concentration was measured using an Agilent HPLC 1200 series, equipped with a Bio-Rad Aminex HPX-87H column (temperature: 55 °C) and a refractive index detector (RID, temperature 35°C). For the analysis, a degassed 1 mM sulfuric (VWR) acid mobile phase with a flow rate of 0.6 mL/min was used. For each sample, the glucose and ethanol concentrations were inferred from calibration curves.
Finally, the cumulative amount of ethanol (in g), produced by the WT and envZ*L116P mutant, was fitted with the four-parametric sigmoidal Gompertz equation (NLS.G4) to determine the production rate [in h-1] in R. This fermentation parameter was then used to statistically compare the ethanol productivity of the envZ*L116P vs. the WT strain using a two-sided pairwise t-test.
Supporting information
S1 Fig. The course of evolutionary adaptation in 16 parallel-evolved strains (HT1-HT16).
https://doi.org/10.1371/journal.pgen.1011707.s001
(TIFF)
S2 Fig. Gene Ontology enrichment analysis representation of the (A) Cellular Component (B), Molecular Function, and (C) Biological Process terms.
https://doi.org/10.1371/journal.pgen.1011707.s002
(TIFF)
S3 Fig. The frequency score for each mutation in the sensor kinase or response regulator in 26 mutated E. coli TCSs.
The reader can find all individual graphs at https://github.com/Michielslab/EnvZ_OmpR_supplementaryData as svg files.
https://doi.org/10.1371/journal.pgen.1011707.s003
(TIFF)
S4 Fig. The total number of genes that are regulated by the corresponding TCS and the number of membrane-associated transporters within the regulon.
https://doi.org/10.1371/journal.pgen.1011707.s004
(TIFF)
S5 Fig. Optical density increase of the the BW25113 WT, the envZ*L116P mutant, and EnvZ-OmpR deletion mutants (ΔenvZ and ΔompR) over a 12h period under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s005
(TIFF)
S6 Fig. L116P changes kinase and phosphatase behavior of the EnvZ osmosensor.
https://doi.org/10.1371/journal.pgen.1011707.s006
(TIFF)
S8 Fig. Flow cytometry-based quantification of ompC and ompF expression in the WT and envZ*L116P strains under different stress conditions.
https://doi.org/10.1371/journal.pgen.1011707.s008
(TIFF)
S9 Fig. Eliminating outer membrane porins influences growth of the WT (A) and envZ*L116P (B) E. coli strains under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s009
(TIFF)
S10 Fig. Original image of four independent spot assays on 6% ethanol agar that were used to quantitatively determine spot area and grey intensity.
https://doi.org/10.1371/journal.pgen.1011707.s010
(TIFF)
S11 Fig. Original image of a spot assay on 6% ethanol agar (replicate 1).
https://doi.org/10.1371/journal.pgen.1011707.s011
(TIF)
S12 Fig. Original image of a spot assay on 6% ethanol agar (replicate 2).
https://doi.org/10.1371/journal.pgen.1011707.s012
(TIF)
S13 Fig. Original image of a spot assay on 6% ethanol agar (replicate 3).
https://doi.org/10.1371/journal.pgen.1011707.s013
(TIF)
S14 Fig. Original image of a spot assay on 6% ethanol agar (replicate 4).
https://doi.org/10.1371/journal.pgen.1011707.s014
(TIF)
S15 Fig. Summary of the morphological and survival data in the two strain competition experiments.
https://doi.org/10.1371/journal.pgen.1011707.s015
(TIFF)
S16 Fig. The composition of the fluorescence marker, used to label each of the two strains within a dual mixed-population, over time under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s016
(TIFF)
S17 Fig. The effect of exchanging ompC for an extra copy of ompF (ompF’) on ethanol tolerance as assayed by a bacterial agar spot assay.
https://doi.org/10.1371/journal.pgen.1011707.s017
(TIFF)
S18 Fig. The ethanol tolerance defect resulting from deleting ompC cannot be rescued by OmpA, TolC, or LamB.
https://doi.org/10.1371/journal.pgen.1011707.s018
(TIFF)
S19 Fig. OM permeabilization as a result of ethanol exposure or polymyxin treatment.
https://doi.org/10.1371/journal.pgen.1011707.s019
(TIFF)
S20 Fig. The impact of eliminating OmpR-regulated genes on the tolerance phenotype of envZ*L116P.
https://doi.org/10.1371/journal.pgen.1011707.s020
(TIFF)
S21 Fig. Comparison of the relative abundances of the differentially expressed proteins in envZ* between the WT, envZ*L116P, and envZ*L116P ompF’/ompF strains (panel A).
https://doi.org/10.1371/journal.pgen.1011707.s021
(TIFF)
S22 Fig. Comparison of the relative abundances of the differentially expressed proteins in envZ* between the WT, envZ*L116P, and envZ*L116P ompF’/ompF strains (panel B).
https://doi.org/10.1371/journal.pgen.1011707.s022
(TIFF)
S23 Fig. Comparison of the relative abundances of the differentially expressed proteins in envZ* between the WT, envZ*L116P, and envZ*L116P ompF’/ompF strains (panel C).
https://doi.org/10.1371/journal.pgen.1011707.s023
(TIFF)
S24 Fig. Comparison of the relative abundances of the differentially expressed proteins in envZ* between the WT, envZ*L116P, and envZ*L116P ompF’/ompF strains (panel D).
https://doi.org/10.1371/journal.pgen.1011707.s024
(TIFF)
S25 Fig. The superior ethanol tolerance of the envZ*L116P is not associated with diminished ROS stress (A, B) and increased resistance to ROS exposure (C, PQ; and D, H2O2).
https://doi.org/10.1371/journal.pgen.1011707.s025
(TIFF)
S26 Fig. Glucose consumption and ethanol production profiles in the WT and envZ*L116P strains.
https://doi.org/10.1371/journal.pgen.1011707.s026
(TIFF)
S1 Data. Summary of mutation dataset from the ALE experiment, conducted by Swings et al. [31].
https://doi.org/10.1371/journal.pgen.1011707.s027
(XLSX)
S2 Data. Growth dynamics of WT, envZ*L116P, ΔenvZ, ΔompR, and envZ*L116P ΔompR as measured by OD595 under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s028
(XLSX)
S3 Data. Survival of WT, envZ*L116P, ΔenvZ, ΔompR, and envZ*L116P ΔompR as measured by CFU counts under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s029
(XLSX)
S4 Data. ompC expression over time as measured by fluorescence intensity.
https://doi.org/10.1371/journal.pgen.1011707.s030
(XLSX)
S5 Data. ompF expression over time as measured by fluorescence intensity.
https://doi.org/10.1371/journal.pgen.1011707.s031
(XLSX)
S6 Data. ompC and ompF expression among all envZ* mutants as measured by fluorescence intensity.
https://doi.org/10.1371/journal.pgen.1011707.s032
(XLSX)
S7 Data. ompC and ompF expression in WT, envZ*L116P, ΔenvZ, ΔompR, and envZ*L116P ΔompR at 10h.
https://doi.org/10.1371/journal.pgen.1011707.s033
(XLSX)
S8 Data. Survival of all envZ* mutants as measured by CFU counts under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s034
(XLSX)
S9 Data. ompC:ompF expression ratio of WT and envZ*L116P under osmotic (PEG6000) and ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s035
(XLSX)
S10 Data. Growth dynamics of WT E. coli under increasing PEG6000 concentrations.
https://doi.org/10.1371/journal.pgen.1011707.s036
(XLSX)
S11 Data. Growth dynamics of WT, ΔompC, ΔompF, envZ*L116P, envZ*L116P ΔompC, and envZ*L116P ΔompF as measured by OD595 under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s037
(XLSX)
S12 Data. Survival of WT, ΔompC, ΔompF, envZ*L116P, envZ*L116P ΔompC, and envZ*L116P ΔompF as measured by CFU counts under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s038
(XLSX)
S13 Data. Survival data of each strain within the two-strain competition experiments as measured by CFU counts under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s039
(XLSX)
S14 Data. Cell morphology parameters retrieved from microbeJ-processed image analysis for competition experiment COMBO 1.
https://doi.org/10.1371/journal.pgen.1011707.s040
(CSV)
S15 Data. Cell morphology parameters retrieved from microbeJ-processed image analysis for competition experiment COMBO 2.
https://doi.org/10.1371/journal.pgen.1011707.s041
(CSV)
S16 Data. Cell morphology parameters retrieved from microbeJ-processed image analysis for competition experiment COMBO 3.
https://doi.org/10.1371/journal.pgen.1011707.s042
(CSV)
S17 Data. Cell morphology parameters retrieved from microbeJ-processed image analysis for competition experiment COMBO 4.
https://doi.org/10.1371/journal.pgen.1011707.s043
(CSV)
S18 Data. Survival data of WT, envZ*L116P, envZ*L116P ΔompC, envZ*L116P ΔompF, and envZ*L116P ompF’/ompF as measured by CFU counts under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s044
(XLSX)
S19 Data. Survival data of WT, envZ*L116P, envZ*L116P ΔompC, envZ*L116P ΔompC::ompA, envZ*L116P ΔompC::tolC and envZ*L116P ΔompC::lamB as measured by CFU counts under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s045
(XLSX)
S20 Data. Permeability data of WT, ΔompC, ΔompF, ΔompR, ompF’/ompF, envZ*L116P, envZ*L116P ΔompC, envZ*L116P ΔompF, envZ*L116P ΔompR and envZ*L116P ompF’/ompF as measured by NPN uptake under 0 and 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s046
(XLSX)
S21 Data. CFU counts of the OmpR regulon deletion mutants in the envZ*L116P background exposed to 5% ethanol for 0 and 12h.
https://doi.org/10.1371/journal.pgen.1011707.s047
(XLSX)
S22 Data. Survival data of WT, envZ*L116P, ΔompC, envZ*L116P ΔompC, ΔtolC, envZ*L116P ΔtolC, ΔlpcA, and envZ*L116P ΔlpcA as measured by CFU counts under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s048
(XLSX)
S23 Data. Permeability data of WT, ΔompC, ΔompF, ΔompR, ompF’/ompF, ΔtolC, ΔlpcA, envZ*L116P, envZ*L116P ΔompC, envZ*L116P ΔompF, envZ*L116P ΔompR, envZ*L116P ompF’/ompF, envZ*L116P ΔtolC, and envZ*L116P ΔlpcA as measured by NPN uptake under 0 and 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s049
(XLSX)
S24 Data. Correlation between permeability data and survival to 5% ethanol (5h exposure).
https://doi.org/10.1371/journal.pgen.1011707.s050
(XLSX)
S25 Data. Differential expression analysis for proteomics data of WT, envZ*L116P, and envZ*L116P ompF’/ompF.
https://doi.org/10.1371/journal.pgen.1011707.s051
(XLSX)
S26 Data. Survival data of ΔentA, ΔentC, ΔentE, ΔentF, ΔfecA, ΔfecB, ΔfhuA, ΔfhuD, Δfiu, ΔfadL, ΔyddB, and ΔfepA vs. WT as measured by CFU counts under 5% ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s052
(XLSX)
S27 Data. Flow cytometry data of envZ*L116P vs. WT using the soxS promoter fusion reporter under ethanol or paraquat stress.
https://doi.org/10.1371/journal.pgen.1011707.s053
(XLSX)
S28 Data. Flow cytometry data of envZ*L116P vs. WT using the dps promoter fusion reporter under ethanol or hydrogen peroxide stress.
https://doi.org/10.1371/journal.pgen.1011707.s054
(XLSX)
S29 Data. Glucose consumption and ethanol production kinetics of envZ*L116P vs. WT [in g/L].
https://doi.org/10.1371/journal.pgen.1011707.s055
(XLSX)
S30 Data. Absolute ethanol production quantities of envZ*L116P vs. WT [in g].
https://doi.org/10.1371/journal.pgen.1011707.s056
(XLSX)
S31 Data. Survival fraction of envZ*L116P vs. WT under ethanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s057
(XLSX)
S32 Data. Survival fraction of envZ*L116P vs. WT under propanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s058
(XLSX)
S33 Data. Survival fraction of envZ*L116P vs. WT under butanol stress.
https://doi.org/10.1371/journal.pgen.1011707.s059
(XLSX)
S1 Table. Cross-referencing table between gene name and protein entrezID.
https://doi.org/10.1371/journal.pgen.1011707.s060
(XLSX)
S2 Table. Table of two-component systems (TCSs) in E. coli.
https://doi.org/10.1371/journal.pgen.1011707.s061
(CSV)
S3 Table. Cross-referencing table between Uniprot ID, gene name and gene locus.
https://doi.org/10.1371/journal.pgen.1011707.s062
(XLSX)
S4 Table. Downstream regulated genes for each transcription regulator (data retrieved from RegulonDB:
https://regulondb.ccg.unam.mx/).
https://doi.org/10.1371/journal.pgen.1011707.s063
(TSV)
S5 Table. Composition of strain combinations within the two-strain competition experiments.
https://doi.org/10.1371/journal.pgen.1011707.s064
(XLSX)
S6 Table. List of genes that belong to the OmpR regulon and were individually deleted in the envZ*L116P mutant.
https://doi.org/10.1371/journal.pgen.1011707.s065
(XLSX)
S1 Text. Strain labels for the OmpR regulon screening assay.
https://doi.org/10.1371/journal.pgen.1011707.s066
(TXT)
Acknowledgments
The authors would like to thank the VIB proteomics core for performing the LC-MS/MS analysis of the E. coli proteome and conducting the related data analysis. Furthermore, the authors acknowledge Dr. Philip Ruelens (Centre of Microbial and Plant Genetics, KU Leuven, Leuven, Belgium; and Center for Microbiology, VIB-KU Leuven, Leuven, Belgium) for his valuable advice on the statistical processing of the survival data.
References
- 1.
Allen MR. Global warming of 1.5°C: an IPCC special report on the impacts of global warming of 1.5°C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change. Intergovernmental Panel on Climate Change (IPCC). 2019.
- 2. Hoegh-Guldberg O, Jacob D, Taylor M, Guillén Bolaños T, Bindi M, Brown S, et al. The human imperative of stabilizing global climate change at 1.5°C. Science. 2019;365(6459):eaaw6974. pmid:31604209
- 3. Lelieveld J, Klingmüller K, Pozzer A, Burnett RT, Haines A, Ramanathan V. Effects of fossil fuel and total anthropogenic emission removal on public health and climate. Proc Natl Acad Sci U S A. 2019;116(15):7192–7. pmid:30910976
- 4. Lin B, Zhu J. The role of renewable energy technological innovation on climate change: Empirical evidence from China. Science of The Total Environment. 2019;659:1505–12.
- 5. Shindell D, Smith CJ. Climate and air-quality benefits of a realistic phase-out of fossil fuels. Nature. 2019;573(7774):408–11. pmid:31534245
- 6. Isikgor FH, Becer CR. Lignocellulosic biomass: a sustainable platform for the production of bio-based chemicals and polymers. Polym Chem. 2015;6(25):4497–559.
- 7. Mohsenzadeh A, Zamani A, Taherzadeh MJ. Bioethylene Production from Ethanol: A Review and Techno-economical Evaluation. ChemBioEng Reviews. 2017;4(2):75–91.
- 8. Dien BS, Cotta MA, Jeffries TW. Bacteria engineered for fuel ethanol production: current status. Appl Microbiol Biotechnol. 2003;63(3):258–66. pmid:13680206
- 9. Huffer S, Roche CM, Blanch HW, Clark DS. Escherichia coli for biofuel production: bridging the gap from promise to practice. Trends Biotechnol. 2012;30(10):538–45. pmid:22921756
- 10. Yang S, Fei Q, Zhang Y, Contreras LM, Utturkar SM, Brown SD, et al. Zymomonas mobilis as a model system for production of biofuels and biochemicals. Microb Biotechnol. 2016;9(6):699–717. pmid:27629544
- 11. Mohd Azhar SH, Abdulla R, Jambo SA, Marbawi H, Gansau JA, Mohd Faik AA, et al. Yeasts in sustainable bioethanol production: A review. Biochem Biophys Rep. 2017;10:52–61. pmid:29114570
- 12. Gibson BR, Lawrence SJ, Leclaire JPR, Powell CD, Smart KA. Yeast responses to stresses associated with industrial brewery handling. FEMS Microbiol Rev. 2007;31(5):535–69. pmid:17645521
- 13.
Taherzadeh MJ, Karimi K. Fermentation inhibitors in ethanol processes and different strategies to reduce their effects. Biofuels: Alternative Feedstocks and Conversion Processes. 2011. 287–311.
- 14. Mukhopadhyay A. Tolerance engineering in bacteria for the production of advanced biofuels and chemicals. Trends Microbiol. 2015;23(8):498–508. pmid:26024777
- 15. Ling H, Teo W, Chen B, Leong SSJ, Chang MW. Microbial tolerance engineering toward biochemical production: from lignocellulose to products. Curr Opin Biotechnol. 2014;29:99–106. pmid:24743028
- 16. Weber FJ, de Bont JA. Adaptation mechanisms of microorganisms to the toxic effects of organic solvents on membranes. Biochim Biophys Acta. 1996;1286(3):225–45. pmid:8982284
- 17. Ingram LO. Microbial tolerance to alcohols: role of the cell membrane. Trends in Biotechnology. 1986;4(2):40–4.
- 18. Cartwright CP, juroszek J-R, Beavan MJ, Ruby FMS, De Morais SMF, Rose AH. Ethanol Dissipates the Proton-motive Force across the Plasma Membrane of Saccharomyces cerevisiae. Microbiology. 1986;132(2):369–77.
- 19. Henderson CM, Block DE. Examining the role of membrane lipid composition in determining the ethanol tolerance of Saccharomyces cerevisiae. Appl Environ Microbiol. 2014;80(10):2966–72. pmid:24610851
- 20. Schalck T, Bergh BVd, Michiels J. Increasing Solvent Tolerance to Improve Microbial Production of Alcohols, Terpenoids and Aromatics. Microorganisms. 2021;9(2):249. pmid:33530454
- 21. Millar DG, Griffiths-Smith K, Algar E, Scopes RK. Activity and stability of glycolytic enzymes in the presence of ethanol. Biotechnol Lett. 1982;4:601–6.
- 22. Pérez-Gallardo RV, Briones LS, Díaz-Pérez AL, Gutiérrez S, Rodríguez-Zavala JS, Campos-García J. Reactive oxygen species production induced by ethanol in Saccharomyces cerevisiae increases because of a dysfunctional mitochondrial iron-sulfur cluster assembly system. FEMS Yeast Res. 2013;13(8):804–19. pmid:24028658
- 23. Cao H, Wei D, Yang Y, Shang Y, Li G, Zhou Y, et al. Systems-level understanding of ethanol-induced stresses and adaptation in E. coli. Sci Rep. 2017;7:44150. pmid:28300180
- 24. Haft RJF, Keating DH, Schwaegler T, Schwalbach MS, Vinokur J, Tremaine M, et al. Correcting direct effects of ethanol on translation and transcription machinery confers ethanol tolerance in bacteria. Proc Natl Acad Sci U S A. 2014;111(25):E2576-85. pmid:24927582
- 25. Stanley D, Bandara A, Fraser S, Chambers PJ, Stanley GA. The ethanol stress response and ethanol tolerance of Saccharomyces cerevisiae. J Appl Microbiol. 2010;109(1):13–24. pmid:20070446
- 26. Horinouchi T, Suzuki S, Hirasawa T, Ono N, Yomo T, Shimizu H, et al. Phenotypic convergence in bacterial adaptive evolution to ethanol stress. BMC Evol Biol. 2015;15:180. pmid:26334309
- 27. Gonzalez R, Tao H, Purvis JE, York SW, Shanmugam KT, Ingram LO. Gene array-based identification of changes that contribute to ethanol tolerance in ethanologenic Escherichia coli: comparison of KO11 (parent) to LY01 (resistant mutant). Biotechnol Prog. 2003;19(2):612–23. pmid:12675606
- 28. Goodarzi H, Bennett BD, Amini S, Reaves ML, Hottes AK, Rabinowitz JD, et al. Regulatory and metabolic rewiring during laboratory evolution of ethanol tolerance in E. coli. Mol Syst Biol. 2010;6:378. pmid:20531407
- 29. Ma M, Liu LZ. Quantitative transcription dynamic analysis reveals candidate genes and key regulators for ethanol tolerance in Saccharomyces cerevisiae. BMC Microbiol. 2010;10:169. pmid:20537179
- 30. Swings T, Van den Bergh B, Wuyts S, Oeyen E, Voordeckers K, Verstrepen KJ, et al. Adaptive tuning of mutation rates allows fast response to lethal stress in Escherichia coli. Elife. 2017;6:e22939. pmid:28460660
- 31. Swings T, Weytjens B, Schalck T, Bonte C, Verstraeten N, Michiels J, et al. Network-Based Identification of Adaptive Pathways in Evolved Ethanol-Tolerant Bacterial Populations. Mol Biol Evol. 2017;34(11):2927–43. pmid:28961727
- 32. Egger LA, Park H, Inouye M. Signal transduction via the histidyl-aspartyl phosphorelay. Genes Cells. 1997;2(3):167–84. pmid:9189755
- 33. Bhate MP, Molnar KS, Goulian M, DeGrado WF. Signal transduction in histidine kinases: insights from new structures. Structure. 2015;23(6):981–94. pmid:25982528
- 34. Casino P, Rubio V, Marina A. The mechanism of signal transduction by two-component systems. Curr Opin Struct Biol. 2010;20(6):763–71. pmid:20951027
- 35. Ghosh M, Wang LC, Huber RG, Gao Y, Morgan LK, Tulsian NK, et al. Engineering an Osmosensor by Pivotal Histidine Positioning within Disordered Helices. Structure. 2019;27(2):302-314.e4. pmid:30503779
- 36. Kenney LJ, Anand GS. EnvZ/OmpR Two-Component Signaling: An Archetype System That Can Function Noncanonically. EcoSal Plus. 2020;9(1):10.1128/ecosalplus.ESP-0001–2019. pmid:32003321
- 37. Zhu Y, Qin L, Yoshida T, Inouye M. Phosphatase activity of histidine kinase EnvZ without kinase catalytic domain. Proc Natl Acad Sci U S A. 2000;97(14):7808–13. pmid:10884412
- 38. Zhu Y, Qin L, Yoshida T, Inouye M. Phosphatase activity of histidine kinase EnvZ without kinase catalytic domain. Proc Natl Acad Sci U S A. 2000;97(14):7808–13. pmid:10884412
- 39. Yoshida T, Qin L, Egger LA, Inouye M. Transcription regulation of ompF and ompC by a single transcription factor, OmpR. J Biol Chem. 2006;281(25):17114–23. pmid:16618701
- 40. Qin L. The critical role of DNA in the equilibrium between OmpR and phosphorylated OmpR mediated by EnvZ in Escherichiacoli. Proceedings of the National Academy of Sciences. 2001;98(3):908–13.
- 41. Wirth NT, Funk J, Donati S, Nikel PI. QurvE: user-friendly software for the analysis of biological growth and fluorescence data. Nat Protoc. 2023;18(8):2401–3. pmid:37380826
- 42. Chakraborty S, Winardhi RS, Morgan LK, Yan J, Kenney LJ. Non-canonical activation of OmpR drives acid and osmotic stress responses in single bacterial cells. Nat Commun. 2017;8(1):1587. pmid:29138484
- 43. Zaslaver A, Bren A, Ronen M, Itzkovitz S, Kikoin I, Shavit S, et al. A comprehensive library of fluorescent transcriptional reporters for Escherichia coli. Nat Methods. 2006;3(8):623–8. pmid:16862137
- 44. Kumar A, Hajjar E, Ruggerone P, Ceccarelli M. Structural and dynamical properties of the porins OmpF and OmpC: insights from molecular simulations. J Phys Condens Matter. 2010;22(45):454125. pmid:21339612
- 45.
Walthers D, Go A, Kenney LJ. Regulation of porin gene expression by the two-component regulatory system EnvZ/OmpR. Bacterial and eukaryotic porins: Structure, function, mechanism. 2005. 1–24. https://doi.org/10.1002/3527603875ch.1
- 46. Dutta R, Yoshida T, Inouye M. The critical role of the conserved Thr247 residue in the functioning of the osmosensor EnvZ, a histidine Kinase/Phosphatase, in Escherichia coli. J Biol Chem. 2000;275(49):38645–53. pmid:10973966
- 47. Matsuyama S, Mizuno T, Mizushima S. Interaction between two regulatory proteins in osmoregulatory expression of ompF and ompC genes in Escherichia coli: a novel ompR mutation suppresses pleiotropic defects caused by an envZ mutation. J Bacteriol. 1986;168(3):1309–14. pmid:3536870
- 48. Hsing W, Russo FD, Bernd KK, Silhavy TJ. Mutations that alter the kinase and phosphatase activities of the two-component sensor EnvZ. J Bacteriol. 1998;180(17):4538–46. pmid:9721293
- 49. Russo FD, Silhavy TJ. EnvZ controls the concentration of phosphorylated OmpR to mediate osmoregulation of the porin genes. J Mol Biol. 1991;222(3):567–80. pmid:1660927
- 50. Shimada T, Takada H, Yamamoto K, Ishihama A. Expanded roles of two-component response regulator OmpR in Escherichia coli: genomic SELEX search for novel regulation targets. Genes Cells. 2015;20(11):915–31. pmid:26332955
- 51. Baslé A, Rummel G, Storici P, Rosenbusch JP, Schirmer T. Crystal structure of osmoporin OmpC from E. coli at 2.0 A. J Mol Biol. 2006;362(5):933–42. pmid:16949612
- 52. Maher C, Hassan KA. The Gram-negative permeability barrier: tipping the balance of the in and the out. mBio. 2023;14(6):e0120523. pmid:37861328
- 53. Mikheyeva IV, Sun J, Huang KC, Silhavy TJ. Mechanism of outer membrane destabilization by global reduction of protein content. Nat Commun. 2023;14(1):5715. pmid:37714857
- 54. Webby MN, Oluwole AO, Pedebos C, Inns PG, Olerinyova A, Prakaash D, et al. Lipids mediate supramolecular outer membrane protein assembly in bacteria. Sci Adv. 2022;8(44):eadc9566. pmid:36322653
- 55. Trimble MJ, Mlynarcik P, Kola M, Hancock REW. Polymyxin: Alternative Mechanisms of Action. Cold Spring Harb Perspect Med. 2016;6:22.
- 56. Hancock RE, Wong PG. Compounds which increase the permeability of the Pseudomonas aeruginosa outer membrane. Antimicrob Agents Chemother. 1984;26(1):48–52. pmid:6433788
- 57. Lundrigan M, Earhart CF. Reduction in three iron-regulated outer membrane proteins and protein a by the Escherichia coli K-12 perA mutation. J Bacteriol. 1981;146(2):804–7. pmid:6452446
- 58. Gerken H, Vuong P, Soparkar K, Misra R. Roles of the EnvZ/OmpR Two-Component System and Porins in Iron Acquisition in Escherichia coli. mBio. 2020;11(3):e01192-20. pmid:32576675
- 59. Ge SX, Jung D, Yao R. ShinyGO: a graphical gene-set enrichment tool for animals and plants. Bioinformatics. 2020;36(8):2628–9. pmid:31882993
- 60. Cheng Y, Du Z, Zhu H, Guo X, He X. Protective Effects of Arginine on Saccharomyces cerevisiae Against Ethanol Stress. Sci Rep. 2016;6:31311. pmid:27507154
- 61. Kohanski MA, Dwyer DJ, Hayete B, Lawrence CA, Collins JJ. A common mechanism of cellular death induced by bactericidal antibiotics. Cell. 2007;130(5):797–810. pmid:17803904
- 62. Dewachter L, Herpels P, Verstraeten N, Fauvart M, Michiels J. Reactive oxygen species do not contribute to ObgE*-mediated programmed cell death. Sci Rep. 2016;6:33723. pmid:27641546
- 63. Hsing W, Silhavy TJ. Function of conserved histidine-243 in phosphatase activity of EnvZ, the sensor for porin osmoregulation in Escherichia coli. J Bacteriol. 1997;179(11):3729–35. pmid:9171423
- 64. Park H, Inouye M. Mutational analysis of the linker region of EnvZ, an osmosensor in Escherichia coli. J Bacteriol. 1997;179(13):4382–90. pmid:9209057
- 65. Yaku H, Mizuno T. The membrane-located osmosensory kinase, EnvZ, that contains a leucine zipper-like motif functions as a dimer in Escherichia coli. FEBS Lett. 1997;417(3):409–13. pmid:9409762
- 66. Gerken H, Shetty D, Kern B, Kenney LJ, Misra R. Effects of pleiotropic ompR and envZ alleles of Escherichia coli on envelope stress and antibiotic sensitivity. J Bacteriol. 2024;206(6):e0017224. pmid:38809006
- 67. Sen O, Hinks J, Lin Q, Lin Q, Kjelleberg S, Rice SA, et al. Escherichia coli displays a conserved membrane proteomic response to a range of alcohols. Biotechnol Biofuels Bioprod. 2023;16(1):147. pmid:37789404
- 68. Poole K. Efflux pumps as antimicrobial resistance mechanisms. Ann Med. 2007;39(3):162–76. pmid:17457715
- 69. Wang Z, Fan G, Hryc CF, Blaza JN, Serysheva II, Schmid MF, et al. An allosteric transport mechanism for the AcrAB-TolC multidrug efflux pump. Elife. 2017;6:e24905. pmid:28355133
- 70. Aono R, Tsukagoshi N, Yamamoto M. Involvement of outer membrane protein TolC, a possible member of the mar-sox regulon, in maintenance and improvement of organic solvent tolerance of Escherichia coli K-12. J Bacteriol. 1998;180(4):938–44. pmid:9473050
- 71. Dunlop MJ, Dossani ZY, Szmidt HL, Chu HC, Lee TS, Keasling JD, et al. Engineering microbial biofuel tolerance and export using efflux pumps. Mol Syst Biol. 2011;7:487. pmid:21556065
- 72. Ankarloo J, Wikman S, Nicholls IA. Escherichia coli mar and acrAB mutants display no tolerance to simple alcohols. Int J Mol Sci. 2010;11(4):1403–12. pmid:20480026
- 73. Woodruff LBA, Pandhal J, Ow SY, Karimpour-Fard A, Weiss SJ, Wright PC, et al. Genome-scale identification and characterization of ethanol tolerance genes in Escherichia coli. Metab Eng. 2013;15:124–33. pmid:23164575
- 74. Reyes LH, Almario MP, Winkler J, Orozco MM, Kao KC. Visualizing evolution in real time to determine the molecular mechanisms of n-butanol tolerance in Escherichia coli. Metab Eng. 2012;14(5):579–90. pmid:22652227
- 75. Choi U, Lee C-R. Distinct Roles of Outer Membrane Porins in Antibiotic Resistance and Membrane Integrity in Escherichia coli. Front Microbiol. 2019;10:953. pmid:31114568
- 76. Labischinski H, Barnickel G, Bradaczek H, Naumann D, Rietschel ET, Giesbrecht P. High state of order of isolated bacterial lipopolysaccharide and its possible contribution to the permeation barrier property of the outer membrane. J Bacteriol. 1985;162(1):9–20. pmid:3980449
- 77. Snyder S, Kim D, McIntosh TJ. Lipopolysaccharide bilayer structure: effect of chemotype, core mutations, divalent cations, and temperature. Biochemistry. 1999;38(33):10758–67. pmid:10451371
- 78. Machas M, Kurgan G, Abed OA, Shapiro A, Wang X, Nielsen D. Characterizing Escherichia coli’s transcriptional response to different styrene exposure modes reveals novel toxicity and tolerance insights. J Ind Microbiol Biotechnol. 2021;48(1–2):kuab019. pmid:33640981
- 79. Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, et al. Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol. 2006;2:2006.0008. pmid:16738554
- 80. Swings T, Marciano DC, Atri B, Bosserman RE, Wang C, Leysen M, et al. CRISPR-FRT targets shared sites in a knock-out collection for off-the-shelf genome editing. Nat Commun. 2018;9(1):2231. pmid:29884781
- 81. Reisch CR, Prather KLJ. The no-SCAR (Scarless Cas9 Assisted Recombineering) system for genome editing in Escherichia coli. Sci Rep. 2015;5:15096. pmid:26463009
- 82. Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A. 2000;97(12):6640–5. pmid:10829079
- 83. Cherepanov PP, Wackernagel W. Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic-resistance determinant. Gene. 1995;158(1):9–14. pmid:7789817
- 84. Defraine V, Liebens V, Loos E, Swings T, Weytjens B, Fierro C, et al. 1-((2,4-Dichlorophenethyl)Amino)-3-Phenoxypropan-2-ol Kills Pseudomonas aeruginosa through Extensive Membrane Damage. Front Microbiol. 2018;9:129. pmid:29472905
- 85. Dwyer DJ, Belenky PA, Yang JH, MacDonald IC, Martell JD, Takahashi N, et al. Antibiotics induce redox-related physiological alterations as part of their lethality. Proc Natl Acad Sci U S A. 2014;111(20):E2100-9. pmid:24803433
- 86. Green R, Rogers EJ. Chemical transformation of E. coli. Methods Enzymol. 2013;529:329–36.
- 87. Luo W-G, Liu H-Z, Lin W-H, Kabir MH, Su Y. Simultaneous splicing of multiple DNA fragments in one PCR reaction. Biol Proced Online. 2013;15(1):9. pmid:24015676
- 88. Pédelacq J-D, Cabantous S, Tran T, Terwilliger TC, Waldo GS. Engineering and characterization of a superfolder green fluorescent protein. Nat Biotechnol. 2006;24(1):79–88. pmid:16369541
- 89. Beal J, Haddock-Angelli T, Baldwin G, Gershater M, Dwijayanti A, Storch M, et al. Quantification of bacterial fluorescence using independent calibrants. PLoS One. 2018;13(6):e0199432. pmid:29928012
- 90. Mutalik VK, Guimaraes JC, Cambray G, Lam C, Christoffersen MJ, Mai Q-A, et al. Precise and reliable gene expression via standard transcription and translation initiation elements. Nat Methods. 2013;10(4):354–60. pmid:23474465
- 91. Mavridou DAI, Gonzalez D, Clements A, Foster KR. The pUltra plasmid series: A robust and flexible tool for fluorescent labeling of Enterobacteria. Plasmid. 2016;87–88:65–71. pmid:27693407
- 92. Ducret A, Quardokus EM, Brun YV. MicrobeJ, a tool for high throughput bacterial cell detection and quantitative analysis. Nat Microbiol. 2016;1(7):16077. pmid:27572972
- 93. Shapiro SS, Wilk MB. An analysis of variance test for normality (complete samples). Biometrika. 1965;52(3–4):591–611.
- 94. Rosner B, Glynn RJ, Lee M-LT. Incorporation of clustering effects for the Wilcoxon rank sum test: a large-sample approach. Biometrics. 2003;59(4):1089–98. pmid:14969489
- 95. Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat. 1979;6:65–70.
- 96. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 2012;9(7):676–82. pmid:22743772
- 97. Chiva C, Olivella R, Borràs E, Espadas G, Pastor O, Solé A, et al. QCloud: A cloud-based quality control system for mass spectrometry-based proteomics laboratories. PLoS One. 2018;13(1):e0189209. pmid:29324744
- 98. Olivella R, Chiva C, Serret M, Mancera D, Cozzuto L, Hermoso A, et al. QCloud2: An Improved Cloud-based Quality-Control System for Mass-Spectrometry-based Proteomics Laboratories. J Proteome Res. 2021;20(4):2010–3. pmid:33724836
- 99. Frankenfield AM, Ni J, Ahmed M, Hao L. Protein Contaminants Matter: Building Universal Protein Contaminant Libraries for DDA and DIA Proteomics. J Proteome Res. 2022;21(9):2104–13. pmid:35793413
- 100. Zhang X, Smits AH, van Tilburg GB, Ovaa H, Huber W, Vermeulen M. Proteome-wide identification of ubiquitin interactions using UbIA-MS. Nat Protoc. 2018;13(3):530–50. pmid:29446774
- 101. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47. pmid:25605792