Phenol as chaotropic solute can cause different kind of damage such as increase in a leakiness of membrane, enhance oxidative stress, and destabilize macromolecules due to the reduced water activity . Therefore, there are several cellular targets which can be disturbed by phenol. It is known that membrane permeabilizing effect of phenol as well as other aromatic compounds is reduced by rigidification of cell membrane, thus maintaining optimal cell membrane fluidity and permeability [3, 34]. Our flow cytometry analysis of phenol-exposed P. putida cultures demonstrated that phenol only slightly increased the amount of cells with PI permeable membrane indicating that cells quite well maintain their membrane homeostasis (Fig. 5). Instead, flow cytometry data indicated that the cell division step of the cell cycle is particularly sensitive to the toxic effect of phenol. This was concluded from the finding that the population structure changed essentially if growth medium contained high concentration of phenol. We observed that phenol caused accumulation of cells with higher DNA content indicating cell division arrest (Fig. 5). Phenol is considered to be toxic primarily because it easily dissolves in membrane compartments of cells, so impairing membrane integrity . Considering that cell division and membrane invagination need active synthesis of membrane components, it is understandable that this step is sensitive to membrane-active toxicant, and in this context, inactivation of cell division is highly adaptive for P. putida exposed to phenol. In accordance with our findings, literature data also suggest that cell division arrest may act as an adaptive mechanism to gain more time to repair phenol-caused membrane damage. For example, it has been shown by proteomic analysis that sub-lethal concentrations of phenol induce cell division inhibitor protein MinD in P. putida . It was also shown that cells of different bacterial species became bigger when grown in the presence of membrane-affecting toxicant . Authors suggested that bigger cell size reduces the relative surface of a cell and consequently reduces the attachable surface for toxic aromatic compound . However, our flow cytometry analysis showed that cell size (estimated by forward scatter) among populations with different DNA content (C1, C2 and C3+) did not change in response to phenol (data not shown). In all growth conditions the average size of cells with higher DNA content was obviously bigger than the size of cells with lower DNA content (data not shown). Therefore, our data indicate that phenol-caused accumulation of bigger cells occurs due to inhibition of cell division which helps to defend the most sensitive step of cell cycle against phenol toxicity.
In this study we disclosed several genetic factors that influence the phenol tolerance of P. putida. The finding that disturbance of intact TtgABC efflux machinery enhances phenol tolerance of P. putida is surprising because this pump contributes to toluene tolerance in P. putida strain DOT-T1E [28, 37]. So, our data revealed an opposite effect in case of phenol. In toluene tolerance the effect of TtgABC pump is obvious as it extrudes toluene , yet, its negative effect in phenol tolerance is not so easily understandable. Our results excluded the possibility that disruption of TtgABC pump can affect membrane permeability to phenol. Rather, flow cytometry data suggest that functionality of TtgABC pump may somehow affect cell division checkpoint. This is supported by the finding that phenol-exposed population of the ttgC mutant contained relatively less cells with higher DNA content than that of the wild-type, implying that in the ttgC-deficient strain the cell division is less inhibited by phenol than that in the ttgC-proficient strain. Interestingly, the MexAB-OprM pump, the TtgABC ortholog in P. aeruginosa, facilitates efflux of a quorum sensing molecule, N-(3-oxododecanoyl) homoserine lactone . The TtgABC homologue in Escherichia coli, AcrAB-TolC, is also involved in extrusion of quorum sensing signals and in regulation of population entering into stationary phase. Namely, it has been shown that acrAB-deficient strain can grow to higher cell density in stationary phase than the wild-type E. coli  indicating that its cell division is less inhibited by stationary phase factors. In case of P. putida, however, we found no evidence that inactivation of TtgABC pump could affect the growth of bacterial culture in stationary phase, as judged by optical density measurements (data not shown). Nevertheless, flow cytometry analysis of the phenol-exposed P. putida ttgC mutant revealed population structure indicative of more active cell division than that of the wild-type. However, at this stage of studies we cannot distinguish whether less arrested cell division is a reason for the increased phenol tolerance of the ttgC mutant or, vice versa, increased phenol tolerance results in less-inhibited cell division.
In our previous study, where we showed that the colR-deficient P. putida is less tolerant to phenol than its parental strain, we argued that membrane permeability of the colR mutant to phenol may be increased . However, results of the current study suggest that the phenol entry into the colR-deficient strain is not increased. The latter was supported by the assay which measured the ability of glucose-grown cells to survive in the presence of 50 mM phenol. Unexpectedly, no differences in cell survival between the wild-type and the colR-deficient strain were recorded after phenol-shock, indicating similar membrane permeability to phenol in the colR-deficient and the wild-type cells. As phenol is known to cause membrane permeabilization  we therefore tested whether population of phenol-exposed colR-deficient strain could contain more cells with PI permeable membrane. However, as judged by flow cytometry analysis of gluconate-grown bacteria, also the membrane permeabilizing effect of phenol is similar to the wild-type and the colR mutant (Fig. 5). Thus, other reasons than enhanced phenol entry or increased membrane permeability should underlie behind the lowered phenol tolerance of the colR mutant.
Interestingly, population analysis at single cell level revealed that compared to the wild-type, phenol more efficiently enhanced the relative amount of subpopulations with higher DNA content in case of the colR mutant, suggesting that cell division of the colR mutant is more sensitive to phenol inhibition than that of the wild-type (Fig. 5). However, it is hard to distinguish whether it occurs due to lowered phenol tolerance or reflects some sort of specific response. Surprisingly, our current study demonstrates that phenol sensitivity of the colR-deficient strain drastically depends on whether bacteria are growing or not - no effect of the ColRS system on phenol tolerance of P. putida could be detected under conditions of starvation (Fig. 3C). Thus, our data imply that state of metabolic dormancy prevents phenol from hitting its target in the colR-deficient cells. We have previously shown that ColR regulates several membrane proteins and is involved in avoidance of several membrane-related disorders [8, 10, 12]. Therefore it is reasonable to suppose that absence of ColR specifically impairs synthesis or turnover of membrane components and this leads to the reduced phenol tolerance in case of actively growing bacteria. However, in starving cells synthesis reactions are down-regulated and that may cut off the effect of ColR deficiency on phenol tolerance. Such scenario would also explain why differences in survival between the wild-type and the colR-deficient strain disappear under growth-permitting conditions at elevated phenol concentrations (Fig. 3A). Eventually, high phenol concentration will totally inhibit biosynthetic processes necessary for cell growth and division, thereby eliminating the target of phenol action in the colR mutant.
In addition to increased phenol stress, the colR mutant experiences serious glucose-specific stress resulting in cell lysis . Importantly, the presence of phenol strongly enhances glucose-dependent cell lysis of the colR mutant as well as proportion of cells with PI-permeable membrane (Fig. 3 and 5). This raises an interesting question about interconnections between phenol- and glucose-caused stresses experienced by the colR-deficient P. putida. It has been shown by Santos and co-workers that phenol induces expression of proteins involved in cell envelope biosynthesis. Namely, LpxC (UDP-3-O-acyl N-acetylglucosamine deacetylase) and MurA (UDP-N-acetylglucosamine enolpyruvyl transferase) are induced by phenol in a concentration-dependent manner . LpxC and MurA are involved in lipopolysaccharide and peptidoglycane biosynthesis, respectively, suggesting that adaptation to phenol involves higher need for synthesis of cell envelope components. As both pathways use UDP-N-acetylglucosamine, this suggests also enhancement of nucleotide sugar metabolism in response to phenol stress. Considering that lysis of the colR-mutant strictly depends on carbon source, the enhancement of glucose-dependent cell lysis by phenol could occur through its dual effect on cell metabolism and membrane homeostasis. Our data suggest that although phenol can significantly enhance the glucose-induced stress in case of the colR-deficient strain, nevertheless, the phenol- and glucose-caused stresses are not directly coupled. This was concluded from the cell lysis and membrane permeability measurement data (Fig. 2 and 5) showing that the increased phenol tolerance of the colR-deficient strain acquired by the disruption of the ttgC gene cannot alleviate the effect of phenol as a facilitator of glucose-dependent autolysis of the colR mutant. Our data rather suggest that in the glucose-growing colR-deficient strain phenol can activate or inhibit also signals not directly related to its toxicity. It is possible that some kinds of cell growth or division signals are misread in the presence of phenol in the colR mutant, which eventually leads to the cell lysis. In that case phenol could act as a signal, leading to the cell death, rather than being killing factor itself. Our further experiments will hopefully clarify whether phenol- and glucose-caused stresses originate from the same defect of the colR mutant or they are caused by different reasons.