Introduction

Benzene is a noxious hydrophobic substance that is sufficiently soluble to be bioavailable, is highly inhibitory to the majority of microbial species, can bioaccumulate along food chains, and poses a significant risk to animal and human health (Lamm and Grunwald 2006; Kelly et al. 2007; Radniecki et al. 2008; Hamada et al. 2009; Littlejohns et al. 2010). Hydrophobic substances, including benzene, are environmentally ubiquitous and also occur in microbe-driven experimental and industrial processes/systems. These include hydrocarbons derived from plants, oil, agrochemicals (e.g., pesticides), landfill leachate, industrial solvents, and biotechnological substrates and products (Keane et al. 2008; Koopman et al. 2010; Timmis 2010; Womack et al. 2010; Escapa et al. 2012; Cray et al. 2015a). Hydrocarbons—including a considerable number of the substances collectively known as volatile organic compounds—are well known as cellular toxicants or stressors of microbial cells (e.g., Velázquez et al. 2005; Cray et al. 2013a, 2015a; Suryawanshi et al. 2015); those which partition into the hydrophobic domains of cellular macromolecules and membranes tend to disorder macromolecular systems (Sikkema et al. 1995; Bhaganna et al. 2010).Footnote 1

The various activities of solutes, such as water activity, Hofmeister effects, ionic strength, and chaotropicity, can determine the biotic windows of microbial species, the composition of microbial communities, and the extent of the microbial biosphere (Rummel et al. 2014; Ball and Hallsworth 2015; Santos et al. 2015; Stevenson et al. 2015a, b). A number of studies have demonstrated that interactions between solutes with different/opposing chemical activities can redefine, and even extend, the growth windows for microbial cells in saline systems (see Hallsworth et al. 2007; Oren and Hallsworth 2014; Yakimov et al. 2015). Additional studies, of cellular and enzyme activity, provided evidence that protectants against cellular water stress (i.e., compatible solutes such as glycerol) reverse the inhibitory action of soluble chaotropes, such as ethanol, and hydrophobic stressors, such as toluene and benzene (Hallsworth 1998; Joo et al. 2000; Hallsworth et al. 2003a; Bhaganna et al. 2010; Trevisol et al. 2011; Bell et al. 2013; Alves et al. 2015).

Whereas whole-cell studies and macromolecular assays provide evidence for a protective effect of compatible solutes against hydrocarbon stressors including benzene (see above), there is a paucity of information on genome-mediated responses in relation to this aspect of cellular biology. Indeed, relatively little is known about interactions between diverse stress parameters and solute activities in relation to windows for microbial metabolism and cell division (Chin et al. 2010; Harrison et al. 2013), hence the recent interest in the physical and chemical parameters that constrain and/or permit life processes (e.g., Stevenson and Hallsworth 2014; Harrison et al. 2015; Lievens et al. 2015). The role of compatible solutes in microbial ecology, stress biology, and biotechnology has been relatively well studied,Footnote 2 yet aspects of the biophysics and ecophysiology of glycerol remain enigmatic. The aims of the current study were to: (1) see whether microbial stress induced by benzene (which is a paradigm for hydrophobic stressors) and (2) characterize the response of cells to benzene stress versus benzene + glycerol, using proteomic profiling of Pseudomonas putida. This soil bacterium was used as a model system because: it has a well-characterized proteome; hydrocarbons are central to its ecology; chaotropicity-induced stresses were first characterized using this species (Hallsworth et al. 2003b; Bhaganna et al. 2010); of the key roles that pseudomonads play in bioremediation, industrial biocatalysis, and biological control (Timmis 2010; Hassan et al. 2011; Cray et al. 2013a); and, for the strain selected for study (KT2440), it is unable to degrade benzene.

Materials and methods

Bacterial strain, media, culture conditions, and growth-rate determination

Pseudomonas putida strain KT2440 (DSMZ 6125) that has been well characterized at the genome-and proteome level (see Nelson et al. 2002; Heim et al. 2003; Ballerstedt et al. 2007) was obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany) and maintained in a liquid minimal mineral salt medium with glucose as the sole carbon substrate and NH4Cl as the sole nitrogen substrate (pH 7.4; for details, see Hartmans et al. 1989). All cultures were grown in 50 ml medium in serum bottles (100 ml, Supelco, Dorset, UK) sealed with aluminum caps with PTFE/butyl septa to avoid volatilization of added stressors, and incubated in a shaking incubator (New Brunswick Innova 44, set at 120 rpm) at 30 °C. Control cultures contained no added stressor or compatible solute. In a second batch of media, benzene was used to supplement media at a concentration (5.2 mM) that induced approximately 50 % growth-rate inhibition relative to the control (see Bhaganna et al. 2010). In the third batch of media, both benzene (5.2 mM) and glycerol (0.52 M) were used to supplement the medium. Glycerol was incorporated prior to autoclaving, whereas benzene was added after media were autoclaved and allowed to cool as described previously (Bhaganna et al. 2010). Whereas water potential is typically used to quantify water availability for wood-rotting fungi or other microbes which can be exposed to high matric potential (e.g. Kashangura et al. 2006), the water activities of culture media were determined for consistency with other studies of bacteria (e.g. Hallsworth et al. 2003b; Stevenson and Hallsworth, 2014; Stevenson et al. 2015a). The water activity values were: control medium, 0.999; benzene-supplemented medium, 0.999; and benzene + glycerol medium, 0.995, determined at 30 °C using a Novasina IC-II water activity machine fitted with an alcohol-resistant humidity sensor and alcohol filter—Novasina, Pfäffikon, Switzerland—as described previously (Hallsworth and Nomura 1999; Stevenson et al. 2015a). Inoculations were carried out through the caps’ septa using disposable syringes (1 ml) fitted with needles. All media were inoculated using cell suspensions from exponentially growing pre-cultures (0.4–0.6 ml) to give a turbidity reading (560 nm) at time zero between 0.1 and 0.15. Samples were taken at hourly or two-hourly intervals, and growth rates during the exponential growth phase were calculated from the slopes of the growth curves. All these experiments were carried out using three biological replicates.

Cell harvesting and protein extraction

Cells were harvested during exponential growth, once the turbidity had reached approximately 0.45, by centrifugation at 4 °C at 10,000 g for 15 min, and then resuspended in phosphate-buffered saline (PBS: 10 mM sodium phosphate buffer, pH 7.4, 138 mM NaCl; Hallsworth et al. 2003b). Harvesting, protein extraction, and proteome analysis were carried out using two technical replicates. Harvested cells were washed twice in PBS, resuspended in PBS buffer, aliquoted in microfuge tubes, pelleted by centrifugation, and then stored at −20 °C overnight as described previously (Hallsworth et al. 2003b). For protein extraction, a cell pellet was allowed to thaw on ice, reswelling solution added, and the suspension sonicated as described by Hallsworth et al. (2003b). The microfuge tubes were allowed to stand at room temperature (15 min), and cell debris and salts were removed as described previously (Hallsworth et al. 2003b; Bobadilla Fazzini et al. 2010). The resulting protein pellets were air-dried and resuspended in 300 ml reswelling solution, and the protein concentration was determined according to Bradford (1976).

Two-dimensional gel electrophoresis and identification of proteins

Two-dimensional gel electrophoresis was carried out according to Görg (1991). Each gel was loaded with proteins extracted from an equivalent cell biomass. Isoelectric focusing was based on the method used by Bobadilla Fazzini et al. (2013): Analytical determinations were made using the Bradford test (150 µg protein mixture) and made up to 300 µl using rehydration solution in the presence of ampholytes/under reducing conditions on ReadyStrip IPG strips (17 cm, pH 3–10). Passive rehydration (2 h, 20 °C) took place on the focusing tray following which samples were covered with silicon oil. Active rehydration was then performed (50 V for 12 h), and isoelectric focusing (10,000 V on Protean® IEF cell; Bio-Rad) was carried out to 75 kWh. Samples were then stored at −70 °C.

The focused strips were then transferred to 1.5-mm-thick gradient sodium dodecyl sulfate (12 %, w/v)–polyacrylamide (15 %, w/v) gels and developed using the IsoDalt system (Amersham Pharmacia Biotech) overnight. The gels were treated with trichloroacetic acid (10 % w/v), stained with Coomassie brilliant blue G-250 (0.1 % w/v) overnight, and then destained using distilled water (Figures S1–S3; Bobadilla Fazzini et al. 2013). The gel images were scanned using a molecular imager GS-800 calibrated densitometer (Bio-Rad), and treatment gels were compared with control gels using Z3 v3.0.7 image-analysis software (Compugen, San Jose, CA) for protein differential expression analysis (Figure S4). Protein identification was carried out using protein spots with differential expression under the experimental treatments; visual analysis confirmed that protein spots on replicate gels did not differ and so spots from one replicate in each case were used for protein identification (Hathout et al. 2002; Hallsworth et al. 2003b). The two replicates from each condition showed high reproducibility by software comparison (our earlier studies also confirm the consistency of biological replicates; Bobadilla Fazzini et al. 2010). For comparative analysis, the relative intensity of each pair of corresponding spots was measured and the data displayed as differential expression values. The data were then checked manually to exclude any error in matching pairs and to judge the significance of the differential analysis by zooming in the spot of interest and checking its incidence in the other set of gels (Hathout et al. 2002). Only matching spots were excised and analyzed. Differentially expressed proteins were selected using a fold-change cutoff: Increases of ≥2.5 were indicative of up-regulation and fold changes of ≤0.5 were indicative of down-regulation. Protein spots (Figure S5) were cut out of gels manually, destained, digested, eluted, and desalted as described by Bobadilla Fazzini et al. (2013): Spots were destained and digested overnight using sequence-grade modified trypsin (Promega); the peptides were eluted/desalted using ZipTip® (Millipore). Samples were loaded along with α-cyano-4-hydroxycinnamic acid matrix for mass spectrometer analysis and analyzed using an Ultraflex II ToF (Bruker Daltonics). Spectra were analyzed using FlexAnalysis 2.0 and Biotools 2.2 software (Bruker Daltonics) and searching carried out on the NCBInr database using Profound version 4.10.5 (Proteometrics). For ESI Q-ToF analysis, 3 µl of sample was directly analyzed after Zip-Tip® elution in a Micromass Q-ToF micro TM mass spectrometer (Waters Corporation) and sequencing analysis carried out using MassLynx Mass Spectrometry Software 4.0 (Waters Corporation). Sequence similarity searching against protein databases was performed using FASTA (European Bioinformatics Institute, Cambridge, UK, at http://www.ebi.ac.uk/fasta33/).

The peptide mass fingerprints obtained were identified using a P. putida (KT2440) protein database that was constructed in conjunction with a genome sequencing study (Nelson et al. 2002). Proteins were classified according to their presumed primary metabolic function as described by Hallsworth et al. (2003b), Volkers et al. (2006), and Bhaganna et al. (2010).

Results

Proteomic profiling of benzene-stressed P. putida cells

As expected, bacterial cultures in benzene-supplemented media (5.2 mM benzene) were inhibited relative to the control cultures that had no benzene added (by 46 %, based on a comparison of exponential-phase growth rates, see Figs. 1, 2a). Analysis of the proteome profiles for benzene-stressed cells revealed changes in a total of 23 proteins, 15 of which were up-regulated and eight did not change significantly relative to those in the control; there were no down-regulated proteins (Table 1). These proteins were listed according to their primary metabolic function within two overall categories: those that are known descriptors for chaotropicity-mediated stress and those known to be general stress adaptations/responses (Table 1). There was a low level of general stress response: there was no apparent up-regulation of proteins involved in protein synthesis and there were only three proteins involved in energy metabolism that were up-regulated by ≥fourfold (Table 1). By contrast, there were increases (≥twofold) of ~10 proteins that, collectively, act as descriptors of chaotropicity-mediated stress. These include proteins associated with lipid metabolism and membrane composition, protein stabilization, compatible-solute transport and metabolism, and nucleic-acid protection/synthesis/repair, as well as those involved in responses to oxidative-stress (see Hallsworth 1998; Hallsworth et al. 2003b; Alves et al. 2015; Ball and Hallsworth 2015; Cray et al. 2015a). Furthermore, there was a tenfold to 15-fold up-regulation of five of these proteins (involved in compatible-solute transport and oxidative-stress response) relative to the control, see Fig. 2.

Fig. 1
figure 1

Growth curves of P. putida (30 °C) cultured in a defined liquid minimal mineral salt medium (see “Materials and methods” section) without any stressor (control; black line), supplemented with benzene (5.2 mM; red line), and benzene + glycerol (5.2 mM and 0.52 M, respectively; blue line). Stress tolerance assays were carried out in triplicate; standard deviations are shown

Full size image
Fig. 2
figure 2

Expression profiling of protein descriptors for chaotrope stress in exponential growth-phase P. putida cells cultured in media supplemented with a benzene + glycerol (5.2 mM and 0.52 M, respectively; 60 % growth-rate inhibition relative to the control—see dashed line) and b benzene (5.2 mM; 46 % growth-rate inhibition—see dashed line). The color coding denotes primary protein function as indicated on the x-axis; columns show fold changes relative to the control except for pale-red and pale-green columns which indicate proteins not present in control gels (plotted arbitrarily at the twofold level). All experiments were carried out in triplicate

Full size image
Table 1 Proteins from P. putida cells during exponential growth in media supplemented with benzenea
Full size table

Proteomic profiling of benzene + glycerol-grown cells

The high level of growth-rate inhibition observed in glycerol-protected, benzene-grown cultures, i.e., 60 % relative to the control (Figs. 1, 2a), was consistent with the intense stress-response of these cells. Among the 16 up-regulated proteins, there were 12 that were up-regulated between approximately four- and 70-fold that are involved primarily in protein synthesis and energy metabolism (Table 2). In contrast to the strong up-regulation of the protein descriptors for chaotrope stress in benzene-supplemented cultures (Table 1), glycerol-protected benzene-stressed cells only showed a weak level of chaotrope-stress response: despite the higher level of growth-rate inhibition, the up-regulation of chaotrope-stress descriptor proteins was limited to between ~twofold and fivefold, except for two proteins (DNA-binding stress protein PP1210 and amino acid ABC transporter, periplasmic amino acid-binding protein PputGB1 5004) that were up-regulated by ~sevenfold (Fig. 2a).

Table 2 Proteins from P. putida cells during exponential growth in media supplemented with benzene + glycerola
Full size table

Discussion

Whereas the P. putida cells cultured in benzene (5.2 mM)-supplemented media were inhibited, their growth rate was nevertheless >50 % that of the control cultures, indicating that these cells were not close to the point of failure (for P. putida strain KT 2440, system failure occurs at ~9 mM benzene; Bhaganna et al. 2010).Footnote 3 This observation is consistent with the weakness of the stress response observed (Table 1). Proteins up-regulated in the presence of benzene (current study) are consistent with those up-regulated in the presence of chaotropic solutes such as ethanol, phenol, and benzyl alcohol (Hallsworth et al. 2003b). These included fadB1X (involved in lipid metabolism) and catalase/peroxidase HPI (removal of reactive oxygen species). Chaotropic substances induce oxidative-stress responses, regardless of microbial species, as a consequence of a breakdown in homeostasis (e.g., Alexandre et al. 2001; Russo et al. 2001; Tam et al. 2006; Wierckx et al. 2008; Ball and Hallsworth 2015; Cray et al. 2015a). Chaotropes cannot cause oxidative damage to cellular components directly, and it is likely that chaotrope-induced oxidative-stress results from disordering of cellular membranes and consequent disruption of electron transport processes that causes a release of free radicals (resulting in lipid peroxidation; Zager 1996; Russo et al. 2001; Hallsworth et al. 2003b; Albano 2006; Kavazis et al. 2009). In the present study, the katG-encoded, catalase/peroxidase HPI was up-regulated 15.1-fold in cells from the benzene-only treatment (Table 1), but not detected in cells treated with benzene + glycerol (Table 2). Up-regulation of this enzyme has previously been observed in P. putida cells exposed to phenol, pyruvate (Kurbatov et al. 2006), and the chlorophenoxy herbicides such as 2,4-dichlorophenoxyacetic acid, 2-(2,4-dichlorophenoxy) propanoic acid, and 2,4-dichlorophenol (Benndorf et al. 2006). The apparent lack of catalase/peroxidase HPI in P. putida cells from the benzene + glycerol treatment may be a consequence of the reduced level of benzene stress (see Bhaganna et al. 2010) and/or the ability of glycerol to scavenge reactive oxygen species (Smirnoff and Cumbes 1989).

General stress responses are characterized by an increased requirement for energy generation and an up-regulation of proteins involved in protein synthesis (Park et al. 2001; Hallsworth et al. 2003b; Veláquez et al. 2005; Domínguez-Cuevas et al. 2006). In the current study, P. putida cells up-regulated the sugar transport protein ABC transporter PP1015 under benzene stress (Table 1), but not in the benzene + glycerol treatment (Table 2). This is consistent with the increased energy required and increased sugar uptake observed in P. putida cells exposed to toluene stress (Segura et al. 2005). Segura et al. (2005) found additional evidence of the need to increase energy generation in response to toluene, as two enzymes of the tricarboxylic acid cycle were induced: dihydrolipoamide dehydrogenase (E3 component of 2-oxoglutarate dehydrogenase complex) and succinyl coenzyme A synthetase. Additional studies, of a transporter-deficient P. putida mutant, indicated that the ABC transporter is implicated in tolerance to toluene shock (at 1 %, v/v) (Kim et al. 1998); this study also demonstrated that an increased uptake of metabolites supports the high energy demands. F1F0–ATP synthase, up-regulated in benzene-stressed cells, is known to play roles in energy generation as well as proton translocation (Len et al. 2004). This protein may therefore be required by the cell in response to a breakdown in trans-bilayer proton gradients induced by the chaotropicity of benzene.

The periplasmic amino acid-binding protein PP3593 was up-regulated 2.2-fold in P. putida cells exposed to benzene (Table 1), but not those in the benzene + glycerol treatment (Table 2). Conversely, there was up-regulation of the periplasmic amino acid-binding proteins PputGB1 5004 and PputGB1_3991 in cells exposed to benzene + glycerol (Table 2), but not in the benzene-only treatment (Table 1). The general amino acid ABC transporter, periplasmic binding protein aapj, was up-regulated under both benzene and benzene + glycerol treatments (Tables 1, 2). However, the fold increase was notably larger under benzene stress (Table 1). Up-regulation of proteins involved in the transport and metabolism of amino acid has similarly been observed in cells of P. putida exposed to sublethal concentrations of phenol, pyruvate, toluene, tetracycline, and benzoate (Santos et al. 2004; Segura et al. 2005; Kurbatov et al. 2006; Yun et al. 2006, 2011). Enhanced uptake of amino acids via ABC transport proteins may relate to both accumulation of compatible solutes (see below) and the efflux of stressors/toxic metabolites in P. putida KT2440 (Nelson et al. 2002; Dos Santos et al. 2004).

Studies of the stress metabolism of diverse microbial species demonstrate that tolerance of diverse cellular stresses is enhanced by up-regulation of proteins involved in protein stabilization (e.g., heat-shock proteins and chaperonins), import or synthesis and retention of compatible solutes, and by actively transporting protectants to the extracellular matrix in order to protect the exterior of the plasma membrane (Eleutherio et al. 1993; Hocking 1993; Albertyn et al. 1994; Kets et al. 1996a; Cray et al. 2015a). Whereas P. putida is known to produce high levels of protein stabilization proteins in response to chaotropic solutes and hydrophobic stressors, as do other bacterial species and eukaryotic microbes (Hallsworth et al. 2003b; Bhaganna et al. 2010; Cray et al. 2015a), this response was not so evident in strain KT2440 under benzene stress at the benzene concentrations and under the conditions used in the current study. Both amino acids and carbohydrates can act as compatible solutes in P. putida, e.g., glycine, betaine, glutamate, N-acetylglutaminylglutamine amide, glycerol, mannitol, and trehalose (see Kets et al. 1996b; Joo et al. 2000; Cray et al. 2013a). Glycerol, which is well known for its properties as a cryoprotectant, osmolyte, and protectant of macromolecular structures under solute-induced stresses (see Brown 1990; Hallsworth et al. 1998, 2003b; Hubalek 2003), has protective activity against benzene (though genome-mediated responses were not investigated; Bhaganna et al. 2010) hence the selection of glycerol as the focus of the current study.

At molar concentrations, glycerol can itself induce cellular inhibition, and this may relate to viscosity (see “Materials and methods” section; Chirife et al. 1984; Wyatt et al. 2015a, 2015 b), water activity reduction, and its own chaotropic activity (see Williams and Hallsworth 2009; Cray et al. 2013b; Alves et al. 2015). Benzene + glycerol-grown cells in the current study were more inhibited than those exposed to benzene alone (see Figs. 1, 2), but the relatively low concentration of glycerol used to supplemented culture media (i.e., 0.52 M) was an order of magnitude lower than would be needed for glycerol to behave in a chaotropic manner (Williams and Hallsworth 2009; Alves et al. 2015). Whereas a background level of chaotropicity-mediated benzene stress was evident in cultures from benzene + glycerol media, the overall level of these chaotrope-stress protein descriptors was 200–300 % lower than those observed in the benzene-stressed culture (see Fig. 2b).

The data obtained are consistent with the hypotheses that hydrocarbons induce cellular water stress, that the stress mechanism is mediated via chaotropicity, and that glycerol protects the cell’s macromolecular machinery via genome-wide responses/adaptations (see also Bhaganna et al. 2010; McCammick et al. 2010). Comparative analyses of the membrane lipid composition during growth with benzene and benzene + glycerol are now needed to shed further light on mechanisms of benzene tolerance. Evidence suggests that compatible solutes such as glycerol can mitigate competitive interactions between microbes which involve chaotropic antimicrobials, and reduce chaotropicity-induced stresses that occur during lignocellulosic and other fermentations (Bhaganna et al. 2010; Cray et al. 2013a, 2015a, 2015b; Alves et al. 2015). For instance, lignin breakdown products include chaotropic substances such as syringaldehyde and vanillin as well as hydrophobic stressors such as (cinnamaldehyde), see Cray et al. (2015a). Compatible solutes can mitigate stresses induced by such substances in yeast and fungi as well as bacteria (Bhaganna et al. 2010; Alves et al. 2015; Cray et al. 2015a). Further work is needed to elucidate the temporal sequence of molecular responses to hydrocarbon-induced water stress, apply biophysical techniques to quantify changes in the non-covalent interactions of macromolecular systems, investigate glycerol-mediated protection against hydrophobes in fungal systems (Cray et al. 2015a; Rangel et al. 2015b), and determine whether the kosmotropic and/or protectant activities of biologically relevant substances can be utilized to minimize hydrocarbon-induced stresses in microbial systems used for biocatalysis, biotransformation, bioremediation, and other industrial and environmental applications.