TNT (purity, 99%) was purchased from ChemService (West Chester, PA, United States).

Yarrowia lipolytica was grown aerobically at 30°C for 1 day on Sabouraud agar medium containing (per liter) glucose 10 g, peptone 10 g, yeast extract 5 g, NaCl 0.25 g, and agar 20 g. Yeast cells were harvested, washed with 16 mM phosphate buffer (pH 7.0) and added into 250 mL Erlenmeyer flasks containing 50 mL of synthetic medium. The synthetic medium was composed of 28 mM glucose, 7.6 mM (NH₄)₂SO₄, 2 mM MgSO₄ and buffered with 16 mM K-Na-phosphate buffer to pH 7.0. The initial cell concentration was adjusted to an optical density (600 nm, A₆₀₀) of 1.0, and growth was measured using a Lambda 35 spectrophotometer (Perkin Elmer) with cell-free (filtered) culture medium as reference. TNT was added as ethanol stock solution to a final concentration 440 μM, and flasks were incubated at 30°C with shaking speed 150 rpm. TNT-free control experiments contained pure ethanol (1.33 mL absolute ethanol into 50 mL medium). Samples of yeast cells were taken during culture growth at time points corresponding maximum formation of analytically measured TNT metabolites. After centrifugation, yeast biomass was washed with 16 mM phosphate buffer twice, frozen in liquid nitrogen, lyophilized, and stored at -80°C. These samples were used for proteins analysis. All experiments were set up in two biological repetitions. To exclude proteins of the first step of TNT transformation, Boolean algorithm for binary data sets of TNT-treated and TNT-untreated yeast proteins at pH 4.6 and pH 6.6 was applied.

TNT and its biotransformation products were detected with a HPLC (Thermo Scientific^TM Dionex^TM UltiMate^TM 3000) equipped with autoinjector, diode array detector, the column oven, and Supelcosil (C-8) column (150 by 4.6 mm; particle size, 5 μm), as described previously (Borch and Gerlach, 2004; Khilyas et al., 2013). Nitrates and nitrites were detected using ion chromatograph 761 Compact IC (Metrohm, Sweden) equipped with a Metrosep A SUPP 5 column (150/4.0; particle size, 5 μm).

TE-buffer (EDTA, pH 8.0, Tris base) containing 1% Protease Inhibitor Mix HP (Serva, Heidelberg, Germany) was added to the yeas biomass. Yeast cells were disrupted by glass beads using a Potter-Elvehejem homogenizer at 4°C (Wheaton, Millville, NJ, United States). After centrifugation (14000 g for 30 min) supernatant containing cytoplasmic proteins was collected. The pellet including membrane-associated proteins was resuspended in an ice-cold solution (8 M urea and 2 M thiourea). Delipidation of supernatant and pellet was performed by methanol-chloroform method (Wessel and Flügge, 1984). Protein quantification was carried out by 2-D Quant kit (GE Healthcare, United States). Cytosolic or membrane-associated proteins (150 μg) were subjected to gel strips (pH 7–11 NL, 13 cm) in sample buffer [4% CHAPS, 30 mM DTT, 20 mM Tris-base, 2% IPG buffer, 1% bromophenol blue (added immediately before use)] and maintained 24 h at room temperature for rehydration. Then, strips were placed in a horizontal running tray for IEF and covered with mineral oil to prevent dehydration during electrophoresis (Wenge et al., 2008). Electrophoresis was performed at 0–100 V (1 mA) for 5 h, 100–3500 V (1 mA) for 6 h and 3500 V (1 mA) for 6 h at 20°C. Strips were washed with 1% DTT for 15 min, with 4% iodoacetamide for 15 min and three times with 50 mM Tris buffer (pH 6.8). Equilibrated strips were transferred on 12.5% polyacrylamide SDS gel (18 cm × 20 cm) using tweezers. Electrophoresis conditions per one gel were 15 mA (600 V, 50 W) for 15 min and 110 mA (600 V, 50 W) for 6 h. Densitometric analyses were done with the PDQuest software (Bio-Rad)after Coomassie-staining.

Coomassie-stained protein spots (n = 404) were excised with the ExQuest Spot Cutter (BioRad) and transferred into 96-wells plates. Digestion of protein was done using liquid handling system (Microstarlet, Hamiltonrobotics, Martinsried, Germany). Gel plugs were washed 10 min with 150 μl 50% acetonitrile (ACN), dehydrated 2 min with 150 μl 100% ACN, swelled 5 min with 150 μl 50 mM NH₄HCO₃, dehydrated 2 min with 150 μl 100% ACN. The proteins were digested for 120 min at 45°C in 25 mM NH₄HCO₃ containing 0.5 mg trypsin (sequencing grade, Promega, Mannheim, Germany). Peptides were stabilized by 10 μL 0.1% TFA and stored at 21°C until use (Wenge et al., 2008).

MALDI-TOF MS was performed using an UltraflexI TOF/TOF mass spectrometer (Bruker Daltonics, Bremen, Germany) equipped with a nitrogen laser and a LIFT-MS/MS facility. The instrument was operated in the positive-ion reflectron mode using 2,5-dihydroxybenzoic acid (Sigma) in 50% ACN/1% phosphoric acid as matrix. Acquired sum spectra consist of 200–400 single spectra. For data processing and instrument control the Compass 1.1 software package consisting of FlexControl 2.4, FlexAnalysis 2.4, and ProteinScape 3.0 was used.

Proteins were identified by MASCOT peptide mass-fingerprint search¹ using the MSDB database. The search was restricted to Yarrowia with a mass tolerance of 100 ppm and carbamidomethylation of cysteine as global modification and oxidation of methionine as variable modification.

After 3 h of cultivation in TNT-free medium, yeast cells were grown to 1.7 units (cell density of 600 nm), whereas in TNT medium cells reached the density of 1.3 units after 5 h of cultivation only (Supplementary Figure S1A). In view of the fact that the biosynthesis for a spectrum of synthesized proteins depends on pH, we sampled the yeast biomass at different time points (3 and 5 h), but at equal levels of pH = 6.5. HPLC data showed that the TNT transformation through aromatic ring reduction led to a maximum accumulation of TNT-monohydride complexes (H^--TNT) after 5 h of yeast cultivation. The maximum concentrations 3-H^--TNT and 1-H^--TNT were 289 and 17 μM, accordingly (Supplementary Figure S1B). At the same time, the TNT transformation via nitro group reduction resulted in the accumulation of 26 μM of 2-hydroxylamino-4,6-dinitrotoluene (2-HADNT), 46 μM of 4-hydroxylamino-2,6-dinitrotoluene (4-HADNT) and 43 μM of nitrites (Supplementary Figure S1C). Thus, the yeast cells grown for 3 h without TNT and the cells grown for 5 h with TNT were used for protein analysis. We relate these time point to the first step of TNT transformation, which leads to the accumulation of TNT-monohydride complexes.

After 12 h of incubation, the pH of culture medium decreased to 4.6. Under these conditions, the TNT-monohydride complexes transformed into TNT-dihydride complexes (3,5-2H^--TNT⋅H⁺), and the concentration of 2-HADNT and 4-HADNT increased to 49 and 96 μM, respectively. The appearance of 2,4-DNT (11 μM) during 3,5-2H^--TNT⋅H⁺ destruction was associated with a release of 133 μM NO2−, which was partially oxidized to NO3− and accompanied by NO formation as it was shown previously (Ziganshin et al., 2010; Khilyas et al., 2013). The time-point of 12 h was chosen for the second step of protein analysis.

The continued cultivation of Y. lipolytica for 24 h resulted in a pH shift to a strong acidic range (pH 3.1-3.6) and the intensive destruction of TNT-mono- and dihydride complexes. However, HADNTs, 4-ADNT, and 2,4-DNT, as well as NO3−. As a product of NO2− oxidation could be still found in growth medium (Supplementary Figures S1B,C). This time point (24 h) was chosen for the last step of protein analysis.

Next, the membrane and cytosolic proteins were analyzed to determine the quantitative and qualitative changes in the protein profiling of Y. lipolytica based on the maximum accumulation of TNT-monohydride complexes after 5 h of cultivation. Approximately, 1300–2000 protein spots were detected on each gel: eight gels correspond to TNT-free cultivation (four for membrane proteins and four for cytosolic proteins) and eight gels were obtained based on the cultivation with TNT. For twenty revealed protein spots, the difference (p < 0.05) between TNT-treated and TNT-untreated yeast was more than fivefold. All these proteins were up-regulated in the presence of TNT. The influence of pH alteration in medium was excluded (Supplementary Table S1).

Figure 1A shows the intracellular distribution of 20 up-regulated proteins in the compartments of yeasts and their role in the cellular processes. A significant number of the up-regulated proteins could be assigned to cytosolic (45% of the total proteins) and mitochondrial (35% of the total proteins) compartments of yeast cells; 15% of all up-regulated proteins were related to plasma membrane proteins. Five percent of other proteins were associated with proteasomes, endoplasmic reticulum, and microsomes, respectively. About 5% of proteins could not be identified. More detailed information is presented in Supplementary Table S1. The cytosolic and membrane proteins from Y. lipolytica were involved in the metabolic process (40%), redox process (23%), electron-proton transport (7%), ATP synthesis (7%), biosynthetic process (4%), catabolic process (4%), fatty acid metabolic process (3%), RNA splicing (3%), propanoate metabolism (3%), and glutathione metabolism (3%). The functions performed by 3% of proteins remained unknown (Figure 1B).

In general, TNT primarily induced the biosynthesis of oxidoreductases and transferases of Y. lipolytica. Three membrane isoforms of NADH flavin oxidoreductases/NADH oxidases demonstrated a higher up-regulation (by 16–60 times) (Figure 1C). Five different isoforms of retinal dehydrogenases 2, belonging to the NAD(P)+-dependent aldehyde dehydrogenase superfamily (ALDH-SF), also increased by eight times. GST and disulfide isomerase/thioredoxin showed a 15- and 10-fold increase, respectively (Figure 1C). Thus, the first step of TNT transformation was associated with an increased level of yeast proteins catalyzing metabolic redox reactions.

The second step of TNT transformation was characterized by the accumulation of TNT-dihydride complexes. In total, 102 up-regulated proteins in cytosolic and membrane fractions were revealed, that is, about five times higher compared with the first step of transformation (Supplementary Table S2). In addition, the number of mitochondrial proteins grew from 35 to 48% compared with the first step. At the same time, the number of cytosolic proteins insignificantly reduced (41.2%) (Figures 1A, 2A). Others proteins were localized in peroxisomes (4.9%), microsomes and cell wall (3.9% each), plasma membrane, endoplasmic reticulum and vacuoles (2% of each), proteasomes and nucleus (1% of each), while the localization of 4.9% of the proteins remained unknown (Figure 2A).

Figure 2B illustrates the functional classification of up-regulated intracellular Y. lipolytica proteins which can be attributed to the accumulation of TNT-dihydride complexes at pH 4.6. The identified proteins participated in metabolic (29%) and redox (17%) processes, and the tricarboxylic acid (TCA) cycle (10%), as well as in biosynthetic (9%) and catabolic (6%) processes (Figure 2B and Supplementary Table S2). In the second step, we also obtained the insignificant amount of proteins that were not up-regulated during the first stage of TNT transformation. These proteins participated in protein biogenesis (5%), biogenesis of RNA (5%), ATP synthesis (2%), glycolysis (2%), gluconeogenesis (2%), glyoxylate cycle (2%), propanoate metabolism (1%), electron-proton transport (1%), and other processes (8%) (Figure 2B and Supplementary Table S2).

Interestingly, the level of NADH flavin oxidoreductases/NADH oxidases was much lower than in the first step of TNT transformation (corresponding to a sixfold increase only) (Figure 2C). Six isoforms of retinal dehydrogenase 2 were up-regulated 45 times. At this stage only, two isoforms of acetaldehyde dehydrogenases were up-regulated. The isoform 2 of NADH flavin oxidoreductase was not detected in the TNT-free proteome at both stages of TNT transformation, whereas the NADPH:quinone reductase was for the first time detected at the second stage (Figure 2C). The GST and disulfide isomerase/thioredoxin were up-regulated, although the level of up-regulation was lower compared with that achieved in the first step of TNT transformation (by 4.5 times only) (Figure 2C). Additionally, the level of glutathione transferase protein showed a 33-fold increase (Figure 2C). Finally, we established an eightfold up-regulation of thioredoxin reductase (Figure 2C).

After 12 h of TNT transformation, we observed an increase in the level of proteins involved in the regulation of RNA biogenesis (transcription and translation elongation) and protein biogenesis (protein folding, modification, and proteolysis) (Supplementary Table S2). Several TNT-induced heat shock proteins (HSP), including HSP 60 (GroEL), HSP78 and piso0_004415 (ATPases associated with a wide variety of cellular activities), ATP-dependent molecular chaperone HSC82, and chaperone protein (DnaK) were identified (Supplementary Table S2). Furthermore, saccharopepsin and carboxypeptidase C, which have a specific location in vacuoles and perform multiple beneficial functions inside the yeast cells, such as proteolysis, were identified (Supplementary Table S2).

The amount of unknown and hypothetical proteins activated at the second stage of TNT-transformation was significantly increased compared with the stage of accumulating the TNT-monohydride complexes (Supplementary Tables S1, S2). Thus, the second step of TNT transformation was characterized by the sustained increase in yeast proteins participating in metabolic redox reactions, proteolysis, and stress-response.

The third step of TNT transformation was characterized by the destruction of TNT-dihydride complexes. Specifically, we identified 31 up-regulated proteins in the cytosolic and membrane fractions. Figure 3A shows the intracellular distribution of up-regulated proteins. A significant number of the up-regulated proteins were localized in the cytosolic (41.9%) and mitochondrial (29%) compartments of yeast cells (Figure 3A). Other proteins were localized in microsomes (12.9%), plasma membrane (9.7%) and vacuoles (9.7%), peroxisome (6.5%), the perinuclear region (3.2%), lipid droplet (3.2%), eisosomes (3.2%) and glyoxysomes (3.2%) (Figure 3A).

Figure 3B illustrates the functional classification of up-regulated intracellular Y. lipolytica proteins that can be attributed to the destruction of TNT-dihydride complexes at pH 3.6. The identified proteins participated in metabolic (11%) and redox (29%) processes, TCA cycle, ATP synthesis, ionic transfer and proteolysis (8.6% of each), vesicle trafficking (5.7%) and cell wall biogenesis (5.7%) (Figure 3B and Supplementary Table S3).

The last stage of TNT transformation by Y. lipolytica led to the inhibition of metabolic processes and the activation of redox processes (Figure 3B). The average amount of proteins participating in energetic processes insignificantly decreased compared with the first and the second stages of TNT transformation, although the part of cellular processes occurring in plasma membrane, microsomes and vacuoles was higher (Figures 1A, 2A, 3A).

Figure 3C illustrates the induction of six retinal dehydrogenases 2, NADPH quinone reductase, and two isoforms of the NAD(P)+-dependent aldehyde dehydrogenase. The isoform 2 of retinal dehydrogenase 2 particularly increased in the third step of TNT transformation.

Furthermore, an increase in the level of sphingolipid long chain base-responsive protein located in eisosome nearly to plasma membrane was detected (Supplementary Table S3). In addition, we observed the overexpression of the potassium channel subunit, which was involved in secretory and endocytic vesicular trafficking pathways. Among HSPs, only HSP 60, or GroEL was found to be up-regulated (Supplementary Table S3).

The yeast cells that adapted to the TNT-induced stress reached an optical density similar to the one that could be observed for the TNT-free system at the end of the third stage of TNT transformation (Supplementary Figure S1A). Yeast cells performed the destruction of TNT-dihydride complexes by NADH flavin oxidoreductases/NADH oxidases, whereas a decrease in HADNTs and ADNTs was entailed by proteins of the NAD(P)+-dependent aldehyde dehydrogenase superfamily (ALDH-SF) (Figures 1C, 2C, 3C). Three isoforms of NADH flavin oxidoreductases/NADH oxidases, as well as disulfide isomerase/thioredoxin, thioredoxin reductase, and peroxiredoxin were found to be up-regulated at all stages of TNT transformation (Figures 1C, 2C, 3C). Generally, the predominant part of the up-regulated enzymes belonged to oxidoreductases. This fact points to the role of such enzymes in TNT transformation.

A significant part of the up-regulated enzymes besides NADH flavin oxidoreductases/NADH oxidases was related to transferases (Figure 4). An increase in the level of these enzymes reflects the activation of the second phase of xenobiotics detoxification. Active in the first phase of TNT transformation, oxidoreductases led to the formation of TNT metabolites, namely, HADNTs and ADNTs, which could not be fully utilized by Y. lipolytica (Ziganshin et al., 2007). Therefore, the conjugation of these compounds with cellular substrates such as glutathione was necessary to excrete the toxic substances. Indeed, we could observe a high level of transferases during the entire cycle of TNT transformation (Figure 4).

The TNT transformation process through the aromatic ring reduction and reduction of nitro groups by Y. lipolytica was closely associated with the generation of ROS. Catalase and SOD could be found at the up-regulated level during all stages of TNT transformation (Figure 4). It is important to note that the high level of catalase and SOD was observed in TNT-untreated cells, which is characteristic of yeast cell growth under aerobic conditions.

Previously, we demonstrated the extracellular generation of nitric oxide (NO) during TNT transformation by Y. lipolytica using ESR spectroscopy (Khilyas et al., 2013). ESR spectra confirmed that NO2− was converted into NO and NO3− following a decrease in pH below 4.5, which coincides with the biotransformation of 3H^--TNT to 3,5-2H^--TNT⋅H⁺ (Khilyas et al., 2013). In this study, we observed a unique induction of intracellular NOD of Y. lipolytica at the second and third stage of TNT transformation (Figure 4). NOD catalyzed the reaction of superoxide anion detoxification with the production of less toxic NO3− and was induced simultaneously to NO generation. However, minor amounts of NOD were detected in the TNT-untreated cells when pH of the medium reduced to 3.6 (Figure 4). The cellular generation of NO in the absence of TNT could be associated with NO-synthase activity which catalyzed the reaction between L-arginine, molecular oxygen and NAD(P)H (Xia et al., 2000; Ignarro et al., 2002).