Stenotrophomonas maltophilia SeITE02, a New Bacterial Strain Suitable for Bioremediation of Selenite-Contaminated Environmental Matrices

Abstract

Biochemical and proteomic tools have been utilized for investigating the mechanism of action of a new Stenotrophomonas maltophilia strain (SeITE02), a gammaproteobacterium capable of resistance to high concentrations of selenite [SeO₃²⁻, Se(IV)], reducing it to nontoxic elemental selenium under aerobic conditions; this strain was previously isolated from a selenite-contaminated mining soil. Biochemical analysis demonstrated that (i) nitrite reductase does not seem to take part in the process of selenite reduction by the bacterial strain SeITE02, although its involvement in this process had been hypothesized in other cases; (ii) nitrite strongly interferes with selenite removal when the two oxyanions (NO₂⁻ and SeO₃²⁻) are simultaneously present, suggesting that the two reduction/detoxification pathways share a common enzymatic step, probably at the level of cellular transport; (iii) in vitro, selenite reduction does not take place in the membrane or periplasmic fractions but only in the cytoplasm, where maximum activity is exhibited at pH 6.0 in the presence of NADPH; and (iv) glutathione is involved in the selenite reduction mechanism, since inhibition of its synthesis leads to a considerable delay in the onset of reduction. As far as the proteomic findings are concerned, the evidence was reached that 0.2 mM selenite and 16 mM nitrite, when added to the culture medium, caused a significant modulation (ca. 10%, i.e., 96 and 85 protein zones, respectively) of the total proteins visualized in the respective two-dimensional maps. These spots were identified by mass spectrometry analysis and were found to belong to the following functional classes: nucleotide synthesis and metabolism, damaged-protein catabolism, protein and amino acid metabolism, and carbohydrate metabolism along with DNA-related proteins and proteins involved in cell division, oxidative stress, and cell wall synthesis.

In trace amounts, selenium (Se) is a metalloid element essential for life in many organisms from bacteria to mammals (27). Nevertheless, selenium has a very narrow margin of safety. In fact, the span between the minimum (i.e., deficiency) and maximum (i.e., toxicity) tolerable levels of this element within different organisms is often a matter of parts per million or even less (5). Therefore, selenium contamination represents a real health concern, not only for humans but also for livestock and wildlife. Although selenium can enter the environment as a consequence of natural erosion and leaching mechanisms affecting seleniferous parental rocks, other sources of contamination for soil, freshwater, groundwater, and sediments include, among others, coal, gold, silver, nickel, and phosphate mining; metal smelting; municipal landfills; oil transport, refining, and utilization; combustion of fossil fuels in power plants; production of pigments; pharmaceutical preparations; glass manufacturing; and repeated irrigation of selenium-rich agricultural soils (4).

Se can occur in a variety of oxidation states {selenate [Se(VI)], selenite [Se(IV)], elemental selenium [Se(0)], and selenide [Se(−II)]}, whose toxicities are strictly related to their degrees of water solubility and hence to their bioavailabilities (2). Selenite (SeO₃²⁻) and selenate (SeO₄²⁻) are the two soluble species of selenium that can be found mostly in aerobic habitats. Both oxyanions are toxic and tend to bioaccumulate (8). Se in the form of Se(IV) is more toxic to most organisms than Se in the form of Se(VI), while Se(0) is insoluble and thus unavailable to biological systems.

Microbial activity is thought to be the primary means by which soluble, chemically active forms of Se can be reduced to the elemental state and thus precipitated (29). Since elemental Se is insoluble, this means that it has less of a chance of moving up through the soil into the food chain or down through the soil into the groundwater (19). In the last 2 decades, among the various Se species, selenite has attracted a great deal of attention as a potential substrate for microbial reduction due to its high toxicity. Evidence has been collected for microbial selenite reduction in both anaerobic and aerobic conditions. Various enzymatic systems, such as nitrite reductase, sulfite reductase, and glutathione (GSH) reductase (GR), have been proposed for or suspected to be involved in the reduction of selenite in bacteria, especially in regard to dissimilatory reduction. On the other hand, evidence has also accumulated for the microbial reduction of selenite and selenate in aerobic conditions. Nevertheless, this reduction mechanism, generally referred to as the “detoxification mechanism,” is poorly understood. The role of a periplasmic dissimilatory nitrite reductase in Se(IV) reduction has been hypothesized for Thauera selenatis, where mutants lacking this enzymatic activity were shown not to be able to reduce either nitrite or selenite (6). Starting from the consideration that dissimilatory nitrite reductase activity had been identified for purple nonsulfur bacteria such as Rhodopseudomonas palustris (22), Rhodobacter sphaeroides (24), and Rhodobacter capsulatus (26), Kessi (16) recently investigated the possible reciprocal influence between nitrite dissimilation and Se(IV) reduction in Rhodospirillum rubrum and R. capsulatus. While in R. rubrum the reduction pathway of Se(IV) was eventually completely separated from that of nitrite, in R. capsulatus both nitrite and selenite were metabolized simultaneously, with these oxyanions interfering at the level of the transport system rather than at the level of the reductase. However, selenite reductase activity seems to be constitutively expressed in both R. capsulatus and R. rubrum. On the other hand, nitrite reduction is supposed not to be constitutively expressed in these bacteria, with the nitrite reductase lacking any selenite reductase activity. Also, GSH has been implicated in bacterial Se(IV) reduction, assuming a mechanism originally described by Ganther (11) for Se(IV) reduction in mammalian systems, based on the formation of selenodiglutathione (GS-Se-SG), which is then reduced to Se(0) by GR. Since GSH occurs at mM concentrations in a wide range of bacteria (including cyanobacteria and alpha-, beta-, and gammaproteobacteria) (20) and strongly reacts with Se(IV) (17), it seems to be a possible ingredient of both dissimilatory and aerobic reduction of Se(IV) in microbes able to synthesize this metabolite. The proof that Se(IV) can induce GR in R. sphaeroides and Escherichia coli (3) strengthens this hypothesis. In the present work, microbiological, biochemical, and proteomic approaches were used to investigate the mechanisms responsible for Se(IV) reduction in Stenotrophomonas maltophilia SeITE02, a new strain of gammaproteobacteria that is resistant to high concentrations of Se(IV) thanks to its capability of reducing this oxyanion to nontoxic elemental selenium in oxic conditions. These features, with particular reference to the capability of reducing selenite in completely aerobic conditions, make this strain worthy of possible exploitation in bioremediation. Very little is known about the detoxification mechanisms of selenite reduction in aerobic conditions, while more information exists on the respiratory pathways of SeO₃²⁻ in anoxic environments. Therefore, the aim of this work is to shed new light on the biological mechanisms actuated by S. maltophilia SeITE02 during aerobic selenite transformation to Se(0). The strain SeITE02 considered in this study had been previously isolated from a selenium-contaminated mining soil in Italy (7).

MATERIALS AND METHODS

Chemicals.

All the chemicals employed in this study were purchased from Sigma-Aldrich (Steinheim, Germany). Nutrient broth and bacteriological agar were furnished by Oxoid Italia Spa (Garbagnate Milanese, Italy). Modified porcine trypsin (sequencing grade) was obtained from Promega (Madison, WI). DC protein assay, the fluorescent stain Sypro Ruby, the immobilized pH 3 to 10 gradient strips, and the electrophoresis-related reagents were provided by Bio-Rad Laboratories (Hercules, CA). Complete protease inhibitor was furnished by Roche Diagnostics (Basel, Switzerland).

Microbiological techniques.

All microbiological experiments were carried out in 250-ml Erlenmeyer flasks containing 100 ml of growth medium incubated at 28°C on an orbital shaker (180 rpm). Each flask was inoculated with aliquots from stationary-phase cultures of SeITE02 strain in such a way as to reach a final optical density of 0.01. Both microbial growth and SeO₃²⁻ content in the medium were measured. Culture samples collected at different times during the experiment were analyzed both for microbial growth and for residual selenite in the medium. Bacterial growth was monitored by CFU counts on agarized nutrient broth, whereas residual selenite in the medium was determined spectrophotometrically as described in “Selenite content determination” below.

(i) Microbiological growth substrates.

The following culture media were modified from DM (defined minimal medium), described by Frassinetti et al. (10), and sterilized by temperature treatment (121°C for 15 min): DM-A {Na₂HPO₄ [2.2 g/liter], KH₂PO₄ [0.8 g/liter], yeast extract [1.0 g/liter], vitamin solution, Wolfe solution} and DM-B {Na₂HPO₄ [2.2 g/liter], KH₂PO₄ [0.8 g/liter], (NH₄)₂SO₄ [2.475 g/liter], yeast extract [1.0 g/liter], vitamin solution, Wolfe solution}. DM-A and DM-B were supplemented with 0.1% (w/vol) glucose as the carbon source. Vitamin solution, Wolfe mineral solution, and carbon substrates were sterilized by filtration with cellulose acetate filters (0.2 μm; Millipore).

(ii) Determination of the 50% lethal dose for nitrite.

Strain SeITE02 was inoculated in 100 ml of DM-A supplemented or not with an increasing sodium nitrite concentration (final concentrations were 2.9, 7.2, 11.6, 18.7, 29.0, and 72.5 mM). Bacterial growth was checked after 48 h of incubation. Each analysis was performed in triplicate.

(iii) Evaluation of nitrite preinduction effect.

In order to evaluate the effect of nitrite preinduction on the ability of bacterial cells to grow in presence of selenite while reducing it to elemental selenium, strain SeITE02 was inoculated in 100 ml of DM-A supplemented or not with 10 mM sodium nitrite. After a 24-h incubation, 25 ml of each culture was collected and transferred to 100 ml (final volume) of fresh medium. This procedure was repeated twice. Eventually, both the control and the nitrite-preinduced bacterial cells were inoculated in DM-A amended with 0.5 mM sodium selenite.

(iv) Evaluation of bacterial growth and selenite reduction in cultures with both NO₂⁻ and SeO₃²⁻ added.

The effect of the simultaneous presence of nitrite and selenite on the viability and Se(IV)-reducing activity of strain SeITE02 was investigated. SeITE02 was inoculated in 100-ml portions of DM-A containing either selenite only (0.2 mM final concentration), both nitrite (18.7 mM) and selenite (0.2 mM), or neither nitrite nor selenite (control).

(v) Cultures under nitrite uptake suppression.

Stationary-phase preinocula were inoculated with 100 ml of DM-A containing selenite (0.2 mM final concentration) and supplemented or not with the nitrite uptake inhibitor 2,4-dinitrophenol (DNP) (25), which was added at three final concentrations of 0.1, 0.5, and 1.0 mM.

(vi) Cultures under GSH synthesis suppression.

SeITE02 was inoculated in 100 ml of DM-A supplemented or not with S-n-butyl homocysteine sulfoximine (BSO) (0.5, 1.0, and 3.0 mM final concentrations) (16). Each culture was then supplemented with selenite (0.2 mM final concentration), which was added either at the beginning of the bacterial growth or during exponential growth (after 18 h of incubation) or the stationary phase (42 h of incubation).

(vii) Preparation of bacterial cells for enzymatic assays and proteomic studies.

Stationary-phase preinocula were infused with 400 ml of DM-A supplemented or not with 16 mM sodium nitrite or 0.2 mM sodium selenite. Incubation took place until the stationary phase was reached; after that, the bacterial cells were recovered by centrifugation (10,000 × g, 10 min) and further processed.

Separation of the subcellular protein fractions.

Different protein fractions (namely, cytoplasmic, periplasmic, and membrane proteins) were separated for localizing the enzyme(s) responsible for selenite reduction at a subcellular level. The bacterial cells recovered by centrifugation were washed twice with 400 ml of a physiological solution (0.9% NaCl); after that the cells were centrifuged again and subjected to periplasmic protein solubilization according to the method of Osborn and Munson (21). The spheroplasts were pelleted by centrifuging at 25,000 × g for 20 min, and then they were suspended in 10 ml of a solution containing 50 mM NaCl and one tablet of Complete protease inhibitor, while the supernatant with the periplasmic protein fraction was recovered, filtered (0.2-μm filter; Millipore), and stored at −20°C. Spheroplast disruption was achieved by sonication, with the samples being kept in ice. After sonication, the solution was centrifuged at 200,000 × g for 75 min. The soluble cytoplasmic proteins present in the supernatant were recovered, filtered, and stored at −20°C, while the membrane fragments, visible as a brown pellet at the bottom of the centrifuge tubes, were solubilized in 10 ml of a 50 mM phosphate buffer solution (pH 7.4) containing 0.5% Triton X-100, frozen, and stored at −20°C.

Spectrophotometric assays. (i) Selenite content determination.

Selenite concentrations in culture media or reaction mixtures was determined by measuring the absorbance at 377 nm of the Se-2,3-diaminonaphtalene complex in cyclohexane according to the procedures of Kessi et al. (18). Calibration curves were constructed by using 0, 50, 100, 150, and 200 nmol of selenite dissolved in the same medium as used in the reaction vials.

(ii) Nitrite content determination.

Nitrite was quantified according to the work of Greenberg et al. (13). The samples were centrifuged to remove the particles in suspension (in the case of cell cultures), and then the supernatant was diluted to final nitrite concentrations in the range 0.1 to 1.0 μg/ml. Eight hundred microliters of each supernatant was taken and added with 200 μl of a solution obtained by mixing 1 volume of the NEDA solution [5.6 mg/ml of N-(1-naphthyl)-ethylenediamine dihydrochloride in Milli-Q H₂O] with 10 volumes of the sulfanilamide solution (8 mg/ml sulfanilamide in 1.8 M phosphoric acid). The reactants were ice-cold, and the reaction took place on ice for 5 min before the absorbance at 545 nm was read. The calibration curve was prepared with the following nitrite concentrations: 0.2, 0.4, 0.6, 0.8, and 1.0 μg/ml.

(iii) Protein content determination.

Protein concentration was determined by means of the DC protein assay (Bio-Rad, Hercules, CA) as recommended by the manufacturer.

Two-dimensional map analysis.

Bacterial cytoplasm proteins were solubilized in a strongly denaturant buffer containing 7 M urea, 2 M thiourea, 3% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), 40 mM Tris, and 5 mM tributyl phosphine. Cysteine alkylation was performed by adding 10 mM free acrylamide in the presence of 5 mM fresh tributyl phosphine; after that, 0.5% carrier ampholytes (pH 3 to 10) and traces of bromophenol blue were added just before the samples were loaded. For each sample, five large-size (18- by 20-cm) replica maps were prepared (900 μg of total proteins, 18-cm-long nonlinear immobilized pH 3 to 10 gradient strips). Focusing took place until reaching a value of 75,000 V × h (10,000-V maximum), while the second dimension (sodium dodecyl sulfate-polyacrylamide gel electrophoresis) was performed on polyacrylamide gradient gels (10 to 20% T), and staining was with the Sypro Ruby fluorescent stain. Finally, a three-way comparative analysis was performed simultaneously with the 15 maps of the three samples, considering the proteins extracted from the cells grown in simple DM-A as the reference for the other “treated” samples (DM-A plus 0.2 mM selenite and DM-A plus 16 mM nitrite). The spots of interest (i.e., those modulated by at least twofold and that passed the t test with P values of <0.01) were analyzed by tandem mass spectrometry (MS-MS).

MS-MS analysis.

The protein spots were excised from stained gels and subjected to in-gel trypsin digestion according to the work of Shevchenko et al. (28). The recovered peptide mixtures were then separated by using a nanoflow high-performance liquid chromatography system (Ultimate/Switchos/Famos; LC Packings, Amsterdam, The Netherlands) coupled with a high-capacity ion trap, HCTplus (Bruker-Daltonik, Germany). Protein identification was performed by searching in the National Center for Biotechnology Information nonredundant database (NCBInr) by using the Mascot program (http://www.matrixscience.com). For positive identification, the score of the result had to be over the significance threshold (P < 0.05).

Statistical data analysis.

A one-way analysis of variance (ANOVA) was used to treat the data, and significant differences were obtained for F values of <0.05.

RESULTS AND DISCUSSION

Effect of nitrite on bacterial growth and selenite reduction in vivo.

The microbial detoxification mechanisms exploited by bacteria in the attempt to resist to high concentrations of selenite are still poorly understood. In particular, great attention has been given to dissimilatory reduction, although there is evidence for the microbial reduction of selenite and selenate in aerobic conditions as well. Because in some cases periplasmic dissimilatory nitrite reductase has been proposed to induce selenite reduction (6), possible interactions between selenite and nitrite reduction pathways were taken into account in this study. For this purpose, the nitrite concentration was determined so as to obtain the same growth inhibition effect as observed with cells grown in the presence of 0.2 mM selenite, even though the concentrations of the two toxic compounds differed greatly (0.2 mM selenite and 18.7 mM nitrite) and were far from representative of a realistic environmental scenario. In fact, our objective was not to investigate the effect of nitrite and selenite on S. maltophilia cells in vivo or their interaction but rather merely to shed some light on the selenite reduction mechanism activated by such a bacterial strain. Moreover, the use of equally toxic concentrations of selenite and nitrite allowed us to perform a comparative proteomic analysis aimed at discovering the detoxification mechanisms used by S. maltophilia strain SeITE02 to cope with the two oxyanions. The effect of the presence of nitrite during the Se(IV) reduction process was investigated either in living-cell systems (in vivo experiments) or with enzyme extract solutions (in vitro experiments). When we tried to treat cells with nitrite prior to incubation with selenite, no appreciable differences in growth rates and reduction kinetics could be observed between preinduced and noninduced cells (not shown). This suggests that the detoxification mechanism is independent of nitrite reduction. In fact, the nitrite reduction system is usually inducible; thus, if it were directly involved in Se(IV) reduction, adaptation to nitrite prior to incubation with Se(IV) should confer some advantage to the bacterial cells. This hypothesis is also corroborated by the fact that the strain SeITE02, when cultivated in a medium with 0.5 mM Se(IV) added, presents the same lag time as observed for cultures not supplemented with selenite (7). However, when liquid cultures were grown in presence of both nitrite and selenite, the different nitrogen sources greatly affected either microbial counts at the stationary phase or Se(IV) reduction efficiency (Fig. 1). Actually, cell growth quite unexpectedly reached the highest value only in the presence of yeast extract as the nitrogen source. On the other hand, both nitrite and ammonium had only moderate effects on bacterial growth kinetics, while they consistently influenced the amounts of cells at the stationary phase. In this regard, ammonium performed worst, allowing for a total cell number 10-fold lower (1 logarithmic point) than that seen for the control. However, the most impressive difference was recorded in Se(IV) reduction kinetics, which was really rapid in the case of yeast extract and ammonia, where 90% of the Se(IV) initially present had already been reduced after a 20-h incubation (i.e., during the exponential growth phase). Conversely, when nitrogen was added in the form of nitrite, Se(IV) reduction started at the end of the exponential growth phase and proceeded slowly, with only 20% of SeO₃²⁻ transformed after 50 h. These results are in agreement with data from the work of Kessi (16) pertaining to R. capsulatus, which performs Se(IV) reduction during the exponential growth phase, without any growth delay due to the need for adaptation in the presence of nitrite. In R. capsulatus as well, moreover, nitrite greatly affects Se(IV) reduction during exponential growth phase. A possible interpretation of this behavior is that the reduction pathways of nitrite and selenite may share one or more enzymes, most probably at the level of transport. In order to verify whether nitrite delayed the selenite reduction onset by competition at the transport level, an experiment was performed with cells grown in the presence of Se(IV) along with the metabolic inhibitor DNP, which has been reported to efficiently inhibit nitrite uptake in other organisms (12, 15). As shown in Fig. 2, Se(IV) reduction in the presence of DNP was very limited, reaching a total conversion of about 20% of the initial amount of selenite after 24 h, while in the remaining incubation time no further conversion of Se(IV) to elemental selenium took place. Furthermore, the inhibitory effect was not related to DNP concentration, at least at the doses tested (0.1 to 1.0 mM). This suggests that the lowest concentration was enough to achieve a saturation effect. Also, nitrite uptake was efficiently inhibited in presence of 1 mM DNP, with the anion concentration decreasing slowly during the first 24 h (30% conversion) and then increasing again during the remaining period of incubation, probably because of a slow degradation of DNP, which could have released some free nitrate into the medium. In fact, although no information is available on the ability of SeITE02 to metabolize nitroaromatic compounds, various Pseudomonas species have been reported to be able to degrade such substances in aerobic conditions with the release of nitrite into the culture medium (9, 14). Such results, obtained with a bacterial growth comparable to that observed for cultures with no DNP (data not shown), signify that Se(IV) is likely to enter the cells via the nitrite absorption pathway, as already proposed by Schloemer and Garrett (25). Additional in vitro experiments have been performed in order to better understand the role of other nitrite-related enzymes (e.g., nitrite reductase).

FIG. 1.

Growth rates (CFU) and selenite reduction kinetics of cells cultured in DM-A added with three different nitrogen sources: sole yeast extract (YE), YE plus ammonium (YE + NH₄⁺), and YE plus nitrite (YE + NO₂⁻). Vertical bars indicate the standard deviations as determined by measurements in triplicate. Significant differences among growth curves were found after the 21st hour of incubation (ANOVA; F < 0.05).

FIG. 2.

Selenite reduction and nitrite conversion in the presence of DNP (0.1, 0.5, and 1.0 mM). Significant differences (ANOVA; F < 0.05) were found for each curve as a function of time. No significant difference was found for the various incubation times between the 0.1 mM DNP curve and the 0.5 mM DNP curve (selenite reduction) and between the 1.0 mM DNP curves (selenite and nitrite reduction).

The role of GSH in selenite reduction in vivo.

Nonprotein thiols, such as thioredoxin and GSH, are widely known for playing a fundamental role in cell protection from toxic oxidizing compounds. GSH has been proposed to drive Se(IV) reduction in bacterial species as well, according to the reaction scheme hypothesized by Kessi and Hanselmann (17). With this reaction mechanism, GSH functions as the main electron donor, reacting with selenite and forming selenodiglutathione. This is then converted into selenopersulfide of GSH by GR. Finally, the last intermediate dismutates into reduced GSH and elemental selenium. For evaluating whether strain SeITE02 relies on a detoxification mechanism based on reduced thiols, bacterial cells were grown in the presence of BSO, a well-known inhibitor of GSH synthesis also used by Kessi for evaluating the influence of GSH level on bacterial growth and Se(IV) reduction rates seen for other species (16), while Se(IV) reduction was monitored through quantification of the selenite remaining in the culture medium. As clearly shown in Fig. 3, BSO inhibited Se(IV) reduction to a great extent when selenite was added at the beginning of the experiment (Fig. 3A), and the inhibitory effect was directly proportional to BSO concentration, which, when present in sufficient amounts (3 mM), almost completely prevented Se(IV) reduction during the lag and exponential growth phases (see Table 1 for inhibition percentage data). Interestingly, no difference between control and cultures amended with BSO could be observed when selenite was supplied during the exponential growth (18 h; Fig. 3B) or during the stationary phase (42 h; Fig. 3C). Therefore, the reduction pathway proposed by Kessi and Hanselmann seems additionally to apply well to strain SeITE02. These findings allow us to draw some conclusions as follow. (i) GSH is clearly involved in the Se(IV) reduction mechanism of the bacterial strain here discussed. (ii) Different mechanisms are likely to coexist. In fact, BSO has no effect when cultures are added with selenite either during exponential phase or in stationary phase. Moreover, when Se(IV) is present from the very beginning (lag phase), it is successfully reduced once the cells have entered the stationary phase. (iii) As a corollary, it is likely that GSH plays its role during the very first steps of the cell response to the toxic oxyanion (lag and exponential growth); that accounts for a total conversion of about 90% of the Se(IV) initially present (Fig. 1).

FIG. 3.

Selenite reduction in the presence of BSO (0.5, 1.0, or 3.0 mM). Selenite was added at different growth stages: lag phase (0 h) (A), exponential phase (18 h) (B), or stationary phase (42 h) (C). Residual amounts of selenite are expressed as percentages of the initial amounts. Vertical bars indicate the standard deviations, as determined by measurements in triplicate. Significant differences (ANOVA; F < 0.05) were found among 3.0 mM BSO and the other curves between the 4th and the 48th hour of incubation. CTRL, control.

TABLE 1.

Inhibition of selenite reduction in the presence of BSO

BSO concn (mM)	% Inhibition of selenite reduction kinetics at indicated time (h)a
	4	24	36	48
0.5	2.9	6.0	5.5	9.7
1.0	15.9	18.6	4.5	0.4
3.0	20.5	31.8	34.0	35.0

^aAverage data from triplicate measurements (ANOVA; F < 0.05).

Selenite reduction in vitro: subcellular localization, optimal reaction conditions, and role of GSH.

As previously reported elsewhere, S. maltophilia strain SeITE02 grown in the presence of selenite confers an intense coloration to the culture medium due to the formation of red colloidal particles. These consist of elemental selenium and accumulate either inside the cells (cytoplasm) or outside in the medium (7). On the other hand, the selenium particles have never been found in the periplasmic space. Such evidence leaves doubts about the site of selenite reduction (widely supposed to be occurring at the periplasmic level) as well as the nature and origin of the extracellular particles. Therefore subcellular-compartment fractionation was carried out in order to perform in vitro enzymatic assays aimed to the localization of selenite reduction activity. All three protein fractions were incubated with Se(IV) solutions buffered at various pH values (5.0, 6.0, 6.5, 7.0, and 8.0) in the presence of NADH as the electron donor. For each sample, three negative controls were prepared in order to evaluate the single contribution of each component (protein extract, selenite, and electron donor) to the appearance of red color due to the reduction of selenite to elemental selenium. The experiment revealed that only the cytoplasm fraction was capable of reducing selenite. By comparing the reduction kinetics slopes at different pH values, we could establish the optimal conditions (pH 6.0 equals 100% activity) and the relative activities at nonoptimal pH values, expressed as percentages of the maximum one (Table 2). Other electron donors were also evaluated. Cytoplasm proteins were incubated with NADH, NADPH, reduced ascorbate, or reduced hydroquinone as electron donors in the presence of selenite at pH 6.0. The formation of elemental selenium was followed by reading the absorbance at 415 nm. In this case, the best electron donor was hydroquinone, followed by NADPH, NADH, and ascorbate (Table 2). Although hydroquinone seemed to be the best donor, NADPH was preferred, since in the presence of hydroquinone, nonspecific, nonenzymatic conversion of selenite also occurred. On the other hand, NADPH allowed for a selenite conversion twice as rapid as that observed for NADH without giving abiotic selenite reduction in the absence of the enzymatic component. The fact that NADPH is preferred to NADH by the enzyme(s) involved in selenite reduction supports the hypothesis that GSH and GR play an important role. In fact, NADPH is the preferential electron donor of GR, which, according to the mechanism proposed by Kessi and Hanselmann (17), would restore the pool of reduced GSH by oxidizing NADPH and reducing selenodiglutathione. In this regard, we have tried to add reduced GSH to the reaction environment for the purpose of verifying whether it had any effect on the Se(IV) reduction kinetics. For this purpose, we used three different cytoplasm protein extracts, deriving from cultures grown in the presence (0.2 or 0.5 mM) or in the absence of Se(IV). Moreover, three negative controls, which were either without GSH (−G), without proteins (−P), or with neither of these (−GP), were used. Figure 4 shows the reduction kinetics of the six systems over a total incubation time of 40 min. Whereas when both the proteins and GSH were missing, only a modest reduction occurred (10% of the initial amount) over the total reaction time, the other two controls (−G and −P) showed a more consistent Se(IV) reduction in the first 10 min (30%), although afterwards the conversion did not proceed any further. This can be explained if we consider that the first reduction step between GSH and selenite takes place spontaneously, without the need for enzymatic catalysis. However, once the GSH pool has been converted, it cannot be restored by GR, and the reaction stops. On the other hand, in the case of complete reaction mixtures with proteins from cultures without Se(IV) or with 0.2 mM Se(IV), the reaction developed more rapidly at the beginning (65% and 50% conversion, respectively) and then continued slowly until reaching a total conversion of about 90% of the initial selenite at the end of the experiment. Therefore, in the presence of GSH, selenite conversion proceeded much more rapidly (minutes) than in the case in which only the electron donor was present (hours). On the other hand, the proteins extracted from cultures grown in presence of 0.5 mM Se(IV) behaved like those for the −G and −P controls, suggesting that high concentrations of Se(IV) during cell growth suppress this mechanism rather than inducing it.

TABLE 2.

Selenite reduction activity in cytoplasm fraction as a function of pH and electron donor^a

pH	% Activity vs. pH	% Activity for indicated electron donor
		Hydroquinone	NADPH	NADH	Ascorbate
5.0	0
6.0	100	100	57.0	33.0	10.6
6.5	32.8
7.0	0
8.0	0

^aActivities are expressed as a function of pH values and electron donors as percentages of the maximum values (for pH 6.0 and hydroquinone; shown in bold). Average data from triplicate measurements are shown (ANOVA; F < 0.05).

FIG. 4.

In vitro selenite reduction in the presence of cytoplasmic protein extracts, NADPH, and GSH. The complete reduction systems contained cytoplasm protein extracts from cultures either without selenite (no Se^IV) or with 0.2 or 0.5 mM selenite. The negative controls were without the protein factor (−P), without GSH (−G), or with neither of them (−GP). Residual selenite (initial amount, 0.2 mM) was determined directly by spectrophotometric quantification. The vertical bars indicate the standard deviations as determined by measurements in triplicate. Significant differences (ANOVA; F < 0.05) were found among complete reaction systems with 0 and 0.2 mM selenite and the other samples (negative controls and 0.5 mM selenite). The negative control, −GP, also significantly differed from all the other samples during the entire assessment.

Global two-dimensional analysis of the proteomic response to nitrite and selenite.

Global two-dimensional analysis has been used for obtaining a comprehensive picture of the proteomic response of S. maltophilia strain SeITE02 to the two oxyanions here investigated. Since Se(IV) reduction was observed exclusively in the presence of the cytoplasm fraction, and since the prefractionation of complex samples is highly recommended prior to any analysis, we decided to focus our attention on this protein fraction for evaluating the selenite and nitrite global effects on the proteome of S. maltophilia. We have considered these two treatments because previous analysis had suggested that some connection exists between the two detoxifying pathways. Three cell cultures were prepared in simple DM-A (our reference) or in media infused with either 0.2 mM selenite or 16 mM nitrite (final concentrations). For a direct comparison of all the protein spots detected in the various samples, all the maps were processed in a unique matched set, which allowed for the detection of approximately 900 protein spots. Selenite was found to significantly modulate 96 protein spots (23 induced, 28 repressed, 17 positively modulated, and 28 negatively modulated by a factor of at least 2), while nitrite caused expression variations in 85 different spots (62 induced, 8 repressed, 8 positively modulated, and 7 negatively modulated by a factor of at least 2). Although nearly equal amounts of proteins were modulated by the two treatments (about 10% of the total spots visualized on the gels), only a few of them were found to be common elements of the SeITE02 responses to selenite and nitrite. Figure 5 shows the master map that summarizes the comparison between the control and the two treated samples, where the modulated spots are highlighted by different symbols and standard spot protein (SSP) numbers, the latter also being listed in Table 3, which reports the protein identities as obtained by MS-MS analysis. Although only a few of the proteins identified presented the same modulation trend in the two treatments, they could be grouped into functional classes which revealed the existence of some common defense mechanisms activated by our strain in response to both of the oxyanions here investigated. Such functional classes are reported in Table 4, with the most interesting ones, i.e., damaged-protein catabolism, DNA metabolism and cell division, and oxidative stress response, highlighted in bold. The fact that various functional classes are modulated in response to either nitrite or selenite was quite expected, as when organisms are challenged with environmental conditions highly different from those they are adapted to, the proteins modulated are numerous and belong either to specific response mechanisms or to general biochemical pathways. Among all the proteins listed in Table 3, the following ones are considered to be particularly interesting.

FIG. 5.

Master map of the three-way comparison among cells not treated (control) and those treated with 0.2 mM selenite or 16 mM nitrite. Symbols: squares, on in selenite; crosses, on in nitrite; triangles, up in selenite; circles, up in nitrite; plusses, off in selenite; diamonds, off in nitrite; inverted triangles, down in selenite. All the spots here analyzed were subjected to t testing with P values of <0.01.

TABLE 3.

Proteins identified by liquid chromatography-MS-MS^a

SSP no.	Modulation trend for treatment with:		Coverage (identified peptide[s])	Mascot score	Mr	pI (observed)	Protein name [homologous organism]
	Sel	NO2
0203	On		1	56	25.5	3.40	Purine nucleoside phosphorylase [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21112337)
0508		Up	4	150	38.1	4.10	Conserved hypothetical protein [Xanthomonas axonopodis pv. citri strain 306] (gi 21108564)
0720	On		2	80	66.7	4.07	Catalase [Xanthomonas axonopodispv. citri strain 306] (gi 21107359)
2113	On		1	60	21.6	5.26	ATP-dependent Clp protease proteolytic subunit [Xanthomonas campestrispv. campestris strain ATCC 33913] (gi 21112001)
2209	Up		1	60	24.9	4.89	Citrate synthase [Xanthomonas campestris pv. vesicatoria strain 85-10] (gi 78037490)
2212		On	1	60	23.7	5.02	Cxytidylate kinase [Xanthomonas axonopodis pv. citri strain 306] (gi 21108544)
2316	Up		2	68	30.6	4.88	Glutamate-cysteine ligase precursor [Xanthomonas campestrispv. campestris strain ATCC 33913] (gi 21114650)
2518	Off		4	241	42.7	4.84	Enolase [Xanthomonas axonopodis pv. citri strain 306] (gi 21107915)
2624		On	3	118	53.5	5.11	Elongation factor G [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21107104)
3115	Off		2	60	20.4	5.63	Hypothetical protein AcidDRAFT_7986 [Solibacter usitatus Ellin6076] (gi 67933900)
3218		On	3	187	24.1	5.58	Elongation factor Tu [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21111917)
3720	Off		2	76	69.2	5.53	Glycyl-tRNA synthetase beta chain [Xanthomonas axonopodispv. citri strain 306] (gi 21110642)
4018		On	1	62	14.0	6.36	Glutathione synthetase [Xanthomonas axonopodispv. citri strain 306] (gi 21109427)
4216	On		2	116	24.7	6.27	Phenol hydroxylase [Xanthomonas campestrispv. campestris strain 8004] (gi 77761325)
4516		On	5	339	38.7	6.25	Acyl-CoA dehydrogenase [Xanthomonas axonopodis pv. citri strain 306] (gi 21107744)
			2	93			Oxidoreductase, zinc binding [Pseudomonas syringae pv. tomato strain DC3000] (gi 28869885)
			1	64			Xaa-Pro aminopeptidase [Hahella chejuensis KCTC 2396] (gi 83643941)
4622		On	3	143	46.0	6.16	Kynurenine 3-monooxygenase [Xanthomonas axonopodis pv. citri strain 306] (gi 21107785)
4717	Up		2	70	58.2	6.27	Polynucleotide phosphorylase [Xanthomonas axonopodis pv. citri strain 306] (gi 21108963)
5010	Up		1	40	15.8	6.46	Conserved hypothetical protein [Xanthomonas axonopodis pv. citri strain 306] (gi 21106892)
5213	On		1	75	25.5	6.67	Enoyl-CoA hydratase [Xanthomonas axonopodis pv. citri strain 306] (gi 21107472)
5318	On	On	2	84	30.1	6.50	Exodeoxyribonuclease III [Legionella pneumophilasubsp.pneumophilastrain Philadelphia 1] (gi 21106384)
5320	On		1	55	33.3	6.62	2-Dehydro-3-deoxyphosphooctonate aldolase [Xanthomonas axonopodis pv. citri strain 306] (gi 21107913)
5326		On	5	302	33.2	6.50	Putative glycerol-3-phosphate dehydrogenase [Xanthomonas campestris pv. vesicatoria strain 85-10] (gi 78034193)
5419	On		3	171	32.9	6.84	2-Dehydro-3-deoxyphosphooctonate aldolase [Xanthomonas axonopodis pv. citri strain 306] (gi 21107913)
5627	Off		3	134	46.2	7.07	Putative phosphoglucomutase [Shigella flexneri 2a strain 301] (gi 24053647)
5636		On	2	98	47.0	6.48	Glycerol kinase (ATP:glycerol 3-phosphotransferase) [Xylella fastidiosa] (gi 24636898)
6219	Off		1	57	25.2	7.01	Hypothetical protein [Nocardia farcinica IFM 10152] (gi 54017403)
6312	Up		1	70	29.8	7.02	Esterase [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21114640)
6421	On		1	71	36.7	7.52	UDP-N-acetylmuramate-alanine ligase [Xanthomonas axonopodis pv. citri strain 306] (gi 21106901)
6516	On		2	99	40.2	7.14	Acetoacetyl-CoA thiolase [Xanthomonas axonopodis pv. citri strain 306] (gi 21107510)
6521		On	6	290	38.7	7.44	Twitching motility protein [Xanthomonas axonopodis pv. citri strain 306] (gi 21109227)
6607	Off		2	88	46.9	7.23	Conserved hypothetical protein [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21112741)
6610		On	2	144	53.9	7.48	Topoisomerase IV subunit B [Xanthomonas axonopodis pv. citri strain 306] (gi 21107909)
6611		On	7	275	54.6	7.41	Peptide chain release factor 3 [Xanthomonas axonopodis pv. citri strain 306] (gi 21109358)
6709		On	3	105	58.8	7.10	Amidophosphoribosyltransferase [Xylella fastidiosa 9a5c] (gi 9107049)
6711		On	2	126	58.8	6.97	Amidophosphoribosyltransferase [Xanthomonas axonopodis pv. citri strain 306] (gi 21107173)
7315		On	1	52	28.7	7.65	Inorganic polyphosphate/ATP-NAD kinase [Xanthomonas campestris pv. vesicatoria strain 85-10] (gi 78035622)
7316		On	1	45	28.5	7.76	Extragenic suppressor protein [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21113417)
7317		On	1	62	30.9	7.81	Bifunctional purine biosynthesis protein [Xanthomonas axonopodis pv. citri strain 306] (gi 21106603)
7532	Off		1	61	42.9	8.00	Conserved hypothetical protein [Pseudomonas fluorescens Pf-5] (gi 68347656)
7534	On		3	120	38.7	7.85	DNA topoisomerase I [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21115033)
7538		Up	1	45	39.7	8.10	Sugar ABC transporter ATP-binding protein [Xanthomonas axonopodis pv. citri strain 306] (gi 21108297)
7539		On	2	96	44.3	8.05	Conserved hypothetical protein [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21112741)
7604	Up		2	86	56.8	7.58	Conserved hypothetical protein [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21115308)
7710		On	3	120	60.0	7.50	RNA polymerase beta subunit [Xanthomonas axonopodis pv. citri strain 306] (gi 21107099)
8015		On	1	7270	16.2	8.84	ATP synthase alpha chain [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21111547)
			2				Transcription termination factor NusB [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21111708)
8119		On	3	118	19.0	8.40	Septum site-determining protein [Xanthomonas oryzae pv. oryzae MAFF 311018] (gi 84368934)
8319		On	3	118	34.9	8.35	Inositol-5-monophosphate dehydrogenase [Xylella fastidiosa Temecula1] (gi 28199329)
8322		On	4	20263	35.0	9.20	Diadenosine tetraphosphatase [Xylella fastidiosa 9a5c] (gi 9107285)
			1				Regulatory protein, LysR:LysR, substrate binding [Ralstonia metallidurans CH34] (gi 68556888)
8410	Off		2	82	33.7	8.31	GTP binding protein [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21112324)
8528		On	2	118	39.6	9.11	TolB protein [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21114229)
8615	Up		1	63	56.8	7.88	Conserved hypothetical protein [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21115308)
8623		Up	4	194	46.5	8.73	Chromosomal replication initiator [Xanthomonas axonopodis pv. citri strain 306] (gi 21106039)
8722	Off		1	56	63.2	8.34	Exodeoxyribonuclease V alpha chain [Xanthomonas axonopodispv. citri strain 306] (gi 21110777)
8732		On	4	193	55.5	8.19	GTP-binding protein [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21113105)
9119	Up		1	50	18.0	9.48	Single-stranded DNA binding protein [Xylella fastidiosa9a5c] (gi 9106400)
9307	Off		4	171	36.5	8.46	Transcriptional regulator [Xanthomonas axonopodis pv. citri strain 306] (gi 21107845)
9311		Up	1	58	30.3	8.37	Shikimate kinase [Pelodictyon phaeoclathratiforme BU-1] (gi 68550241)
9314		On	2	55	29.8	9.45	Chromosomal replication initiator [Xanthomonas axonopodis pv. citri strain 306] (gi 21106039)
9405	Up		2	82	37.9	9.51	Dihydrolipoamide acetyltranferase [Xanthomonas axonopodis pv. citri strain 306] (gi 21110045)
9503	Off		2	59	43.8	8.77	Trehalose-6-phosphate synthase [Xanthomonas campestris pv. campestris strain ATCC 33913] (gi 21114302)
9506		On	2	88	40.7	9.67	6-Phosphofructokinase [Xanthomonas axonopodis pv. citri strain 306] (gi 21109796)

^a“Interesting” proteins are highlighted in bold. CoA, coenzyme A.

TABLE 4.

Functional protein classes^a

Functional class	Selenite modulations	Nitrite modulations
Nucleotide biosynthesis and metabolism	Purine nucleoside phosphorylase; polynucleotide phosphorylase; GTP binding protein	Amidophosphoribosyltransferase; cytidylate kinase; bifunctional purine biosynthesis protein; inositol-5-monophosphate dehydrogenase; inorganic polyphosphate/ATP-NAD kinase; ATP synthase alpha chain; diadenosine tetraphosphatase
Protein catabolism	ATP-dependent Clp protease proteolytic subunit; prolyl oligopeptidase
Protein and amino acid metabolism	Glycyl-tRNA synthetase beta chain	Elongation factor G; elongation factor Tu; peptide chain release factor 3; kynurenine 3-monooxygenase; Xaa-Pro aminopeptidase; Shikimate kinase
Lipid metabolism	Enoyl-CoA hydratase; esterase; acetoacetyl-CoA thiolase	Acyl-CoA dehydrogenase
Sugar metabolism	Citrate synthase; dihydrolipoamide acetyltranferase; enolase; putative phosphoglucomutase; trehalose-6-phosphate synthase	Putative glycerol-3-phosphate dehydrogenase; glycerol kinase; 6-phosphofructokinase; sugar ABC transporter ATP-binding protein; HPr kinase/phosphatase
DNA-related proteins and cell division	DNA topoisomerase I; exodeoxyribonuclease III; exodeoxyribonuclease V alpha chain; single-stranded DNA binding protein	RNA polymerase beta subunit; transcription termination factor NusB; regulatory protein, LysR:LysR, substrate binding; exodeoxyribonuclease III; topoisomerase IV subunit B; septum site-determining protein; chromosomal replication initiator
Oxidative stress-related proteins	Catalase; glutamate-cysteine ligase precursor	Glutathione synthetase
Cell wall synthesis	2-Dehydro-3-deoxyphosphooctonate aldolase; UDP-N-acetylmuramate-alanine ligase

^a“Interesting” classes are highlighted in bold. CoA, coenzyme A.

SSP 0720 (induced by selenite): catalase.

This enzyme converts hydrogen peroxide to oxygen and water, thus constituting a central enzyme in the oxidative stress response. Because H₂O₂ is a natural by-product of the abiotic reduction of selenite catalyzed by GSH (17), the fact that this enzyme is expressed only during growth in the presence of selenite strongly supports the hypothesis that GSH takes part in the first steps of selenite reduction in vivo.

SSP 2316 (positively modulated by selenite): glutamate-cysteine ligase precursor.

Along with the enzyme GSH synthetase, this enzyme drives the biosynthetic pathway of glutathione in most of the eukaryotes, cyanobacteria, and purple bacteria (30).

SSP 4018 (induced by nitrite): GSH synthetase.

GSH (the most abundant nonprotein thiol compound present in a cell) and its related enzymes (among which are glutamate-cysteine ligase, GSH synthase, GR, and GSH S-transferase) constitute an extremely important protection mechanism against oxidant agents in both eukaryotic and prokaryotic cells. GSH has been reported to provide high resistance to acidic environments and osmotic and oxidative stress (23).

Conclusions.

Bacteria can be efficaciously exploited for the bioremediation of polluted natural matrices (e.g., soils, groundwater, and sediments) and also in some cases for the biorestoration of handworks and precious artworks (1). As concerns the bacterium here discussed, previous findings (7) suggested that S. maltophilia strain SeITE02 could be potentially usable for the bioremediation of selenite-contaminated environments. It is actually capable of growing in aerobic conditions in the presence of elevated concentrations (up to 50 mM) of this oxyanion, meanwhile reducing such a pollutant to nontoxic insoluble elemental Se. In the present work, attempts have been made to clarify the biochemical mechanism(s) responsible for such a peculiar activity. Evidence has been gained that the SeITE02 strain of Stenotrophomonas maltophilia is greatly affected by nitrite in terms of its capacity for selenite reduction. This can be interpreted as a possible competition between selenite and nitrite for the same transport mechanism, which tends to exclude Se(IV) from the proper intracellular recognition sites.

Moreover, the global proteomics analysis, despite providing no further details on the specific mechanism responsible for selenite reduction in S. maltophilia SeITE02, has revealed that both nitrite and selenite probably exert their toxicities by generating reactive oxygen species during their conversion into less dangerous or harmless compounds. In fact, various enzymes involved in the oxidative stress response and in the processes of protein and DNA reparation have been found to be newly expressed, or at least positively modulated, upon cell treatment with one of the two oxyanions. There is also further evidence that GSH actually plays a role in the first steps of selenite reduction in vivo. In fact, selenite seems to strongly induce the expression of at least two different GSH-related enzymes (glutamate-cysteine ligase and GSH synthetase), which are responsible for the biosynthesis of this compound in prokaryotes.