Phosphoproteomic Analysis of Mouse Thymoma Cells Treated With Tributyltin Oxide: TBTO Affects Proliferation and Energy Sensing Pathways

We report the results of phosphoproteomic analysis of mouse thymoma cells treated with tributyltin oxide (TBTO), an immunotoxic compound. After cell lysis, phosphoproteins were isolated using Phosphoprotein Puriﬁcation Kit, separated by SDS-PAGE and subsequently digested with trypsin. Phosphopeptides were enriched employing titanium dioxide, and the obtained fractions were analyzed by nano-LC-MS/MS. A total of 160 phosphoproteins and 328 phosphorylation sites were identiﬁed in thymoma cells. Among the differentially phosphorylated proteins identiﬁed in TBTO-treated cells were key enzymes, which catalyze rate-limiting steps in pathways that are sensitive to cellular energy status. These proteins included acetyl-CoA carboxylase isoform 1, which catalyzes the rate-limiting step of fatty acid synthesis. Another enzyme was glutamine: fructose-6-phosphate amidotransferase, GFAT1, the ﬁrst and rate-limiting enzyme for the hexoamine synthesis pathway. a multicomplex enzyme that catalyzes the rate-limiting step of aerobic oxidation of fuel carbohydrates, was identiﬁed in both TBTO-treated and control cells; however, phosphorylation at residue S293, known to inhibit PDH activity, was identiﬁed only in control cells. A lower expression level of ribosomal protein S6 kinase 1, a downstream kinase of the mammalian target of rapamycin signaling pathway implicated in protein synthesis through phosphorylation of 40 ribosomal S6, was observed in the treated cells. Giant kinases like AMP-activated protein kinase (AMPK) and cAMP-dependent protein kinase (PKAR1A), which are known to mediate the phosphorylation of these enzymes, were identiﬁed in TBTO-treated cells. Downregulation of proteins, such as MAPK, matrin-3 and ribonucleotide reductase, subunit RRM2, which are implicated in cell proliferation, was also observed in TBTO-treated cells. Together, the results show that TBTO affects proliferation and energy sensor pathways.

Tributyltin oxide (TBTO), an organotin compound, has been used as a molluscicide, agent in wood preservation, and as antifouling in marine paints, industrial water systems, and textiles (Elsabbagh et al., 2002). Human exposure mainly occurs through the food chain by consumption of dietary marine products (Kannan et al., 1996). TBTO is primarily characterized by its immunotoxicity, which is manifested through both specific and nonspecific immune suppressive effects (Aluoch et al., 2006;Van Loveren et al., 1990;Vos et al., 1990). In rodents, TBTO causes atrophy of the thymus gland .
In addition to its immunotoxic effect, TBTO has adverse effects on both vertebrate and invertebrate endocrine systems (Iguchi et al., 2007) and, similar to various other organotins (Costa, 1985;Doctor et al., 1982; and references therein), has been shown to be a potent neurotoxicant; impaired learning ability and memory loss were observed in rats treated with this compound (Elsabbagh et al., 2002). Furthermore, it has been shown that TBTO inhibits protein synthesis (Osman et al., 2009;Raffray et al., 1992) and, like other-related organotin compounds, disrupts mitochondrial energy metabolism (Penninks et al., 1983;Raffray et al., 1992;Soracco and Pope, 1983;Veiga et al., 1996), and blocks mitosis and cytokinesis, disturbing spindle formation in Chinese hamster V79 cells (Jensen et al., 1991). In accord with the latter observation, triaryl-and trialkyltin compounds have been shown to inhibit in vitro polymerization of rat brain tubulin (Tan et al., 1978). Although numerous studies have been undertaken to understand the mechanisms of TBTO toxicity, including messenger RNA (mRNA) microarray investigations (Baken et al., 2006(Baken et al., , 2007Katika et al., 2011), the target molecules of TBTO toxicity still remain unclear. The EL-4 thymoma cell line was shown to possess an immature thymus cell phenotype (Tanaka et al., 1987) and has been used as a model for mouse thymocytes in mechanistic studies (El-Darahali et al., 2005;Lee et al., 2008). Recently, we have reported proteomic profiling of mouse thymoma cells treated with TBTO and found that this compound altered the expression levels of 12 proteins, including prothymosin alpha (ProtTa) (Osman et al., 2009), an essential protein in cell proliferation (Evasfieva et al., 2000;Sburlati et al., 1991). Based on these results, we proposed that the downregulation of this immune-modulating protein could account for the previously reported antiproliferative effect of TBTO (Osman et al., 2009).
The purpose of the present study was to extend our knowledge on the effect of TBTO on protein expression by investigating whether TBTO modifies the phosphorylation state of the proteome of treated thymoma cells as compared with that of untreated cells. Phosphorylation is probably the most studied posttranslational modification since most cellular processes utilize reversible phosphorylation for regulation, e.g., metabolism, signaling pathways, cell cycle, differentiation, or as a code for protein degradation. However, the identification of phosphoproteins by mass spectrometry poses challenges mainly because of low abundance of phosphoproteins within cells compared with nonmodified proteins, low stoichiometry of phosphorylation, ionic suppression, and difficulties in locating phosphorylation sites as well as the intrinsic dynamic nature of phosphorylation in cells (Mann et al., 2000). Therefore, in the present study, we used the following approach: Upon lysis of cells and extraction of proteins, first, we enriched the phosphoprotein fractions by using phosphoprotein purification columns, followed by 1D SDS-PAGE. Proteins were then digested in gel with trypsin. Second, we used titanium dioxide for enriching phosphopeptides prior to the analysis of peptide fractions by nano-LC-MS/MS analysis. The results showed that the treatment of thymoma cell line with TBTO causes differentially phosphorylated proteins and that this immunotoxic compound especially affects the phosphorylation state of key enzymes that catalyze rate-limiting steps of energy sensing pathways and proteins involved in cell proliferation.
Cell culture and treatment with TBTO. EL-4 cells were cultured as described before (Osman et al., 2009). Each treated sample or control contained 5 3 10 7 cells in a volume of 20 ml. A 1mM TBTO stock solution was prepared in ethanol, and 10 ll of this solution was added to each one of the cell suspension samples so that the final concentration of TBTO was 0.5lM, whereas 10 ll of ethanol without TBTO was added to the control cells. Two independent replicate experiments for both treated sample and control were performed for the proteomics approach and two other replicates for immunoprecipitation of matrin-3 and Western blot (WB) analysis. Subsequently, cells were incubated at 37°C for 6 h. Cell viability was evaluated by trypan blue dye exclusion and ranged 85-90% in all experiments, which was in agreement with previous reports (Osman et al., 2009;Raffray and Cohen, 1991). TBTO concentrations higher than 1lM and incubation times longer than 6 h led to apoptosis of cells (Osman et al., 2009;Raffray and Cohen, 1991). TBTO has a lipophilic character and, like other tributyltin compounds, is rapidly taken up and distributed in tissues (WHO, 1990).
Protein extraction and phosphoprotein enrichment procedure. After exposure, cell suspensions were centrifuged at 1200 rpm for 10 min at 4°C and the pellets were washed in 20 ml 10mM Tris-HCl, pH 7 and centrifuged again. This procedure was repeated twice. The cell pellet of each sample or control was then suspended in 5 ml of phosphoprotein lysis buffer prepared following the instructions of Phosphoprotein Purification Kit (Qiagen). This volume of the lysis buffer contained, in addition to the detergent CHAPS and benzonase (10 ll), one tablet of protease inhibitor and phosphatase inhibitor cocktail 1 (10 ll). The cell suspension was incubated on ice for 30 min and vortexed briefly and gently every 10 min. After that, the cell lysate was centrifuged at 10,000 rpm for 30 min at 4°C. The supernatant was then taken and protein concentration determined according to Bradford (Bradford, 1976). A total amount of 2.5 mg protein, at the concentration of 0.1 mg/ml, was loaded on the affinity column, which was preequilibrated with lysis buffer as recommended by the manufacturer. The chromatography was performed at room temperature. After the column was washed with the remaining lysis buffer, the enriched bound fractions of phosphoproteins were eluted using the phosphoprotein elution buffer, accompanying the kit. These enriched fractions were combined and the protein concentration was determined. The combined fractions were dialyzed to remove CHAPS and concentrated using Amicon Ultra 0.5 ml, 3 K cut off. The ultrafiltrate was reduced to approximately 50 ll and used for electrophoresis.
Gel electrophoresis. Electrophoresis of the samples was performed essentially as described in detail before (Osman et al., 2009). Briefly, 80 lg of protein was loaded per well of a 10% NU-PAGE-Tris-gel and run at 160 V for 90 min. After that, the gel was rinsed three times with deionized water (5 min each time), followed by staining it with Coomassie (Simply Blue Safe Stain) for 2 h. The gel lanes of sample and control were horizontally cut into 14 fractions. Each gel slice was cut into pieces and put into a 1.5 ml Eppendorf tube.
In-gel digestion. The fractions were destained, dehydrated, reduced with dithiothreitol, and alkylated with iodoacetamide as described before (Osman et al., 2009). The fractions were digested with trypsin and incubated for 17 h. After that, to each fraction, 30 ll of ammonium bicarbonate buffer was added and incubated for additional 2 h. The supernatant of each fraction was put into an Eppendorf tube and subsequently 25 ll of 50% ethanol containing 5% formic acid was added to each fraction, mixed, and pooled it to the previous one. This process was repeated once. The fractions were freeze-dried by Speed-Vac, suspended in 60 ll acetonitrile, and again freeze-dried. Finally, the peptide fractions were suspended in 50 ll solution of 0.5% formic acid/2% acetonitrile and used for phosphopeptide enrichment.
Titanium dioxide enrichment of phosphopeptides. Phosphopeptides were enriched from peptide mixture obtained as described above using titanium dioxide MonoTip Tio pipette tips. Essentially, the operation protocol of the manufacturer was followed with slight modification. The MonoTip Tio was preconditioned with 100% acetonitrile, and this process was repeated twice. TBTO AFFECTS STRESS SENSOR PATHWAYS 85 Next, the tip was conditioned with 0.2M phosphate buffer, pH 7.0 and equilibrated with 50% aqueous acetonitrile solution containing 0.1% formic acid, repeating this step three times. After that, peptide fraction was drawn and ejected. This adsorption process was repeated 25 cycles per fraction, after which the tip was rinsed with 30% aqueous acetonitrile containing 0.1% formic acid and 0.1M KCl. The rinsing process was repeated six times per fraction. Finally, the tip was desalted with the equilibration buffer, drawing and ejecting three times, followed by elution with 2% aqueous ammonia. The desalting step did not affect negatively the efficiency of the enrichment method. The eluted fractions were freeze-dried, suspended with 60 ll of 100% acetonitrile per fraction, and again freeze-dried. After that, each fraction was suspended with 30 ll of 1% formic acid/2% acetonitrile aqueous solution. The fractions were kept at À20°C until analysis.
Nano-LC/MS/MS. Peptide fractions were analyzed by a nano-LC-MS/MS using an Agilent 11000 Series LC set up (vacuum degasser, autosampler, and one high pressure-mixing binary pump without static mixer) coupled to an LCQ Deca Quadruple Ion Trap mass spectrometer (Thermo Finnigan, San Jose, CA) as previously described by Meiring et al. (2002). The conditions of elution were essentially as described in Osman et al. (2009). Briefly, we loaded 5 ll of peptide solution to a trap column (Aqua C18 [Phenomenex]; l ¼ 15 mm-100 lm ID, packed in house) at 5 l/min of 100% solvent A (0.1M acetic acid). Following a decrease of the flow to 150 nl/min by a splitter, peptides were transferred to Reprosil C18 RP column (Dr Maisch GmbH [l ¼ 20 cm, 50 lm ID]) with a linear gradient from 0 to 100% solvent B (0.1M acetic acid in 80% acetonitrile) in 40 min. Solvent B was maintained at 100% for 5 min, followed by decreasing solvent B to 0% in 0.1 min, washing, and reequilibrating the system at 100% A for 10 min. The LCQ operated in the data-dependent mode using a positive ion mode. Full MS spectra from m/z 400 to 2000 were acquired followed by a tandem spectra of the three most intense precursor ions present in the MS scan. Other instrument parameters have been previously described (Osman et al., 2009).
Database searches. All MS/MS data were analyzed using SEQUEST engine (Bioworks 3.3 version) (Yates et al., 1995) and Proteome Discoverer (version 1.1) against the IPI mouse 3.69 fasta database. A mass tolerance of 1.4 for the precursor and 1 Da for mass fragments were applied. Trypsin was used for protein digestion, allowing two missed cleavages, static modification for cysteine (carbamidomethylation) and variable phosphorylation of Ser/Thr/Tyr and methionine oxidation. The following filtering threshold criteria were used: a Delta Cn (Dcn) was set to 0.08, Sequest X corr versus charge state (X corr ¼ 2 for z ¼ 1, X corr ¼ 1.5 for 2, and X corr ¼ 2.5 for z ¼ 3). The exposure experiments for the proteomics were performed in duplicate and each treated or control sample was analyzed twice (technical duplicate). Phosphoproteins and the corresponding phosphopeptides identified in the control samples and their technical duplicates were combined in a list; the same procedure was applied to the treated samples. The comparison of these lists allowed determining the common phosphoproteins and the corresponding phosphopeptides identified in both cells as well as those detected only in either of them. The relevance of the differentially phosphorylated proteins was evaluated mainly on the number of differentially phosphorylated peptides identified and the known role of the phosphorylation sites, in addition to their potential role in explaining what is already known about TBTO effects.
Preparation of total cell extracts for immunoprecipitation and immunoblot. Total cell extracts used for WB analyses and immunoprecipitation of matrin-3 were prepared by using cell pellets obtained as the described above. Each sample was suspended in 1 ml of RIPA lysis buffer containing 10 ll of PMSF, 10 ll of protease inhibitor cocktail, l0 ll of sodium orthovanadate, phosphatase inhibitor cocktail 1 (l0 ll), and 1mM final concentration of EDTA. For immunoprecipitation of matrin-3, sometimes (see below), EDTA was not included in the lysis buffer. Cell suspensions were incubated on ice for approximately 30 min and subsequently disrupted by repeated aspiration through a 21 gauge needle, transferred to an Eppendorf, and centrifuged at 12,000 3 g for 10 min. The supernatants were used for protein concentration determination, divided into aliquots, and stored at À20°C.
Immunoprecipitation of matrin-3. Control cell lysates extracted in lysis buffer with and without EDTA, as described above, were used for the immunoprecipitation experiments. In each case, 500 lg protein was incubated with goat anti-matrin-3 antibody (2 lg) for 2 h at 4°C, after which 20 ll of A/G Plus agarose suspension was added to the lysate and incubated overnight at 4°C. The immunoprecipitates were collected by centrifugation at 2500 rpm for 5 min at 4°C. Pellets were washed four times with 1 ml of 10mM Tris-HCl, pH 7, containing 0.15M NaCl, each time repeating the centrifugation step. After the final wash, the pellet was suspended in 40 ll of 13 electrophoresis buffer. Samples were boiled for 3 min, centrifuged at 2500 rpm, and the supernatants used for WB analysis. Five microliters of each sample was subjected to SDS-PAGE.
WB analysis. Both the enriched phosphoproteins and the total cell extracts obtained as described above were used for immunoblotting. In the case of the enriched phosphoproteins, equal amounts of protein (60 lg per lane) from TBTO-treated and nontreated cells were separated by SDS-PAGE, followed by electrotransfer of the proteins using Hybond-P PVDF membrane at 25 V for 90 min. For the total cell extracts, the same amount of protein as the enriched phosphoproteins was used for one of the proteins (matrin-3) and 20 lg of protein per lane for RPS6k1. The membrane was blocked by using nonfat milk powder in PBS containing 0.2% Tween and probed with goat anti-matrin-3 antibody diluted in (1:400) or with rabbit anti-p70 S6 kinase (Thr421/Ser424, corresponding to Thr444/Ser447 in mouse sequence) antibody diluted in (1:400). Detection was performed by using donkey anti-goat IgG-HRP (1: 2000) and goat anti-rabbit IgG-HRP (1:2000), respectively, and ECL Western blotting detection reagents (Amersham Pharmacia). The antibody of the loading control was diluted in 1:1000.

Validation and Procedure of Enrichment of the Phosphoprotein Purification Affinity Column
The effectiveness of the phosphoprotein purification affinity column was evaluated by loading onto the affinity column a mixture of known phosphorylated proteins (a-and b-casein), a partially phosphorylated protein, ovalbumin, and a nonphosphoprotein, BSA. We conditioned the column according to the manufacturer's instructions. The flow-through fraction, wash, and the eluate were collected and run on 1D SDS-PAGE. The phosphoproteins aand b-casein and a part of ovalbumin were detected in the eluate, whereas BSA and the remaining part of ovalbumin were detected in the flow-through and wash fractions (data not shown). This result indicates the ability of the phosphoprotein affinity column to enrich phosphoproteins. Next, we applied our cell extracts, containing 2.5 mg protein at the concentration of 0.1 mg/ml, as recommended by the manufacturer, to the affinity column. Approximately 10% of the loaded protein was bound to the column (Fig. 1). The insert shows a typical elution profile of the enriched phosphoprotein fractions. Fractions 2-6 were combined, though most of the bound proteins were found in the third fraction. After separation of the isolated phosphoproteins by SDS-PAGE, followed by in-gel digestion with trypsin, we enriched phosphopeptides using titanium dioxide prior to nano-LC-MS/MS analysis of the samples.

Classification of the Phosphoproteins Identified
A total of 160 phosphoproteins and 328 phosphorylation sites were identified from thymoma cells. Examples of CID 86 OSMAN AND VAN LOVEREN spectra from two of the identified phosphopeptides, NRNp-SNVVPYDFNR from isoform 2 of leukocyte common antigen and RGGGSGGGDEpSEGEEVDED from transcriptional activator protein Pur-beta, are reported in Figure 2. The CID spectra show the matched b and y ions as well as the neutral loss of phosphate group from parent ions, thus allowing the identification of these phosphopeptides.
Additional examples of MS/MS of identified phosphopeptides are reported in Supplementary table 1S. Of these proteins 52 were common to both TBTO-treated and non-treated cells (Table 1), whereas equal number of proteins (54) were detected in TBTOtreated (Table 2) and nontreated cells (Table 3). These data (160 phosphoproteins and the 328 sites) were compared with the public database of Phosphosite (http://www.phosphosite.org/). Of the phosphoproteins and sites identified, 3 proteins and 13 sites were not reported in this public database and hence are marked with asterisk (Tables 1-3). The three proteins are an oxidoreductase of the cytochrome P450 family (CYP4x1), G-proteincoupled receptor 45 (GPR45), a signal transducer, and a protease isoform 2 of astacin-like metalloendopeptidase (ASTL) (Tables 2  and 3). Examples of representative MS/MS spectra of validated phosphopeptides are reported in Supplementary table 1S. The identified phosphorylation sites were distributed as follows: approximately 74% occurred on serine, 18% on threonine, and 8% on tyrosine (Supplementary fig. 2SA). Figure 3A reports the classification of the total identified phosphoproteins on the basis of their biological function. The majority of the identified proteins belonged to DNA-and RNA-binding proteins, which, combined with transcription factors, represented 31% of the total (Fig. 3A). Cytoskeletal proteins were the next abundant group (17%). Furthermore, proteins involved in signal pathways, metabolism, DNA replication and repair, transport, proteasome and ubiquitin-conjugating system, and chaperons were also identified ( Fig. 3A). Upon comparison of the proteins identified in TBTO-treated cells ( Fig. 3B and Table 2) with those detected in the controls (Supplementary fig. 2SB and Table 3), proteins involved in cell cycle, proliferation, and translation were relatively more representative in the controls than in the treated cells. In contrast, proteins involved in signal inhibitory effect and in apoptosis were found in TBTO-treated cells. For instance, both caspase 3 (CASP3) and programmed cell death (PDCD5), which are involved in apoptotic cell processes, were identified only in TBTO-treated cells (Table 2). CD5, a cell surface glycoprotein, which has been implicated in the negative regulation of TCRmediated growth responses in thymocytes (Tarakhovsky et al., 1995), was detected only in TBTO-treated cells ( Table 2).
Some of the differentially phosphorylated proteins identified in this study appear to be relevant for the mechanism (s) of TBTO toxicity.

Identification of Critical Phosphorylation Sites on Key Enzymes
Among the differentially phosphorylated proteins identified in the treated cells but not in control cells were key enzymes that catalyze rate-limiting steps in biosynthetic pathways. One of these enzymes was acetyl-CoA carboxylase (ACC) isoform 1, which catalyzes the rate-limiting step of fatty acid synthesis (Ha et al., 1994). Phosphorylation sites at residues S79, S77, and S76 were identified for this protein (Table 2). Another enzyme that was identified in TBTO-treated cells only was glutamine: fructose-6-phosphate amidotransferase 1 (GFPT1), which catalyzes the rate-limiting step in hexoamine pathway (Eguchi et al., 2009). Peptides with phosphorylation sites at T260 and T261 were identified for this enzyme (Table 2).
A key regulatory enzyme that was identified in both TBTOtreated and control cells was the multicomplex protein pyruvate dehydrogenase (PDH) ( Table 1). This enzyme links glycolysis to the tricarboxylic acid cycle by catalyzing the irreversible oxidative decarboxylation of pyruvate to form CO 2 , NADH, and acetyl-CoA. Although only one phosphorylation site at the residue S295, common to both cells, was identified, additional phosphorylation sites (S293, T231, Y289, and Y301) were identified in the control cells (Table 1). This enzyme's activity is regulated via phosphorylation (Linn et al., 1969). Phosphorylation at the specific sites of S232, S293, and S300 are known to inhibit the enzyme activity (Rardin et al., 2009 and references therein). One of these phosphorylation sites, S293, was identified in control cells. Moreover, p70RPS6k1, a downstream effecter of mammalian target of rapamycin (mTOR) pathway and a major kinase for 40 ribosomal S6, whose phosphorylation is often associated with increase in protein synthesis and cell growth (Fingar and Blenis, 2004;Zhou et al., 2010; and references therein), was identified in both control and TBTO-treated cells (Table 1). One phosphorylated site, at residue S447, was detected in TBTO-treated cells (Table 1), whereas in addition to this residue, two more phosphorylated sites at T444 and S452 were identified for this kinase in control cells (Table 1). Because RPS6k1 phosphorylation is used as a readout of mTOR activity (Hay and

88
OSMAN AND VAN LOVEREN Sonenberg, 2004), we evaluated the level of the phosphorylated protein in both cells using p70s6k (Thr421/Ser424) antibody, corresponding to Thr444 and Ser447 in mouse sequence. WB analysis of both enriched phosphoproteins and total cell extracts confirmed the decrease in the intensity signal of the band corresponding to Thr444/Ser447 of the treated cells as compared with that of the control (Fig. 4A). In addition to these major enzymes, proteins implicated in cell proliferation and growth were affected by TBTO exposure.

Proteins Associated With Cell Proliferation Were
Downregulated in TBTO-Treated Cells Proteins that are implicated in cell proliferation and growth were identified in control cells. Based on the number of phosphorylated peptides identified, these proteins appeared to be downregulated in TBTO-treated cells. One of these proteins was MAPK1, known also as extracellular signalregulated kinase, which has essential roles in cell proliferation, growth, and differentiation (Warren et al., 2009). Phosphorylation sites at T179, T183, T188, and Y185 were identified for this protein in control cells (Table 3). This protein was not detected in TBTO-treated cells. Another protein identified only in control cells and considered as a marker for cell proliferation was ribonucleotide reductase M2. The enzyme ribonucleotide reductase catalyzes the conversion of ribonucleotides to deoxyribonucleotides for DNA synthesis. The enzyme consists of two subunits: M1 and M2. The activity of M1 is stable during the whole cell cycle, whereas the activity of M2 is transient and has been shown to be present in late G1/early S phase of the cell cycle (Engström et al., 1985;Heidel et al., 2007). A phosphorylation site at S20 was identified for this protein (Table 3). Finally, another protein that seemed to be downregulated in TBTO cells was matrin-3 (Table 1). This nuclear matrix protein has been implicated in replication, transcription, and RNA processing (Erazo et al., 2011;Valencia et al., 2007;and references therein). The phosphorylation site at the residue S598 was identified in both TBTO-treated and nontreated cells (Table 1). But additional phosphorylated sites at the residues S596, Y597, S188, Y526, and S157 were identified in the control cells (Table 1). To investigate whether this differential phosphorylation state of matrin-3 might be attributed to differences in the expression level of the protein, we performed immunoblot analysis of both enriched phosphoproteins and total cell extracts. Several bands were observed in both treated and nontreated cell extracts. The intensities of these bands were more marked in the enriched phosphoproteins (Fig. 4B). However, in the enriched phosphoproteins, only one band corresponding to the apparent molecular weight of matrin-3 (ca. 126 kDa) was observed in control extract but not in TBTO-treated sample (Fig. 4B), confirming the proteomic data. When total cell extracts were used, a band with much weaker intensity, with apparent MW of 126 kDa, was observed in the treated crude extract as compared with that of the control (Fig.   4B), again confirming the downregulation of matrin-3 in the treated cells. Similar migration position was reported in several studies for matrin-3, obtained from different mammalian sources, including mouse cells (Alexandra et al., 1990;Nakasu and Berezney, 1991). The bands with the weak signal intensity observed in the control extract with migration positions relatively close to the band with molecular weight 126 kDa (Fig. 4B) are most likely differentially phosphorylated matrin-3 isoforms. Differentially phosphorylated matrin-3 isoforms have been previously reported in rat liver cells (Giordano et al., 2005), in rat neuron cells (Hibino et al., 2006), and in human fetal brain (Bernert et al., 2002). In contrast, the bands with the lower molecular weights might be degradation products of matrin-3. WB analysis of immunoprecipitated matrin-3 from control lysates, prepared in RIPA lysis buffer with or without EDTA showed, in addition to a band of strong signal intensity at high molecular weight, corresponding to matrin-3, the presence of two additional bands with apparent molecular weights of 51 and 25 kDa, which were especially marked in signal intensity in the lysates extracted with the buffer without EDTA (Fig. 4C).
Finally, because it was previously shown that organotin compounds disrupt microtubule assembly causing spindle disturbance in V79 Chinese hamster cells (Jensen et al., 1991), it is worth to note that, among the cytoskeleton proteins identified in the treated cells, there were two proteins belonging to the family tubulins, both of beta chain TUBB4 and TUBB5 ( Table 2). The phosphorylation sites at residues S335 and T55 were identified for these proteins, respectively ( Table 2). Polymerization of ato b-tubulins leads to the formation microtubules, which carry out a variety of functions in the cell, including cell division, where they form mitotic spindles, structures that segregate faithfully sister chromatids. A number of kinases are known to mediate phosphorylation of tubulin subunits, some of which inhibit the polymerization reaction (Macrae, 1997). Also, a microtubule-binding protein, collapsin response mediator protein-2 (DPYSL2 or CRMP-2) was identified in the treated cells but not in the control cells (Table 2). A number of phosphorylation sites, including T514, were identified for this protein ( Table 2). The nonphosphorylated form of this protein has been shown to promote axon outgrowth via microtubule assembly, whereas the phosphorylated form, especially at position T514, weakens the binding of DPYSL2 to tubulins, thereby hindering axon outgrowth (Yoshimura et al., 2005). Therefore, this phosphoprotein might contribute to the reported organotin-mediated disruption of microtubule assembly.

DISCUSSION
The study of organotin toxicity has been a subject of intense investigation for decades, yet, the molecular mechanisms responsible for the toxicity of these compounds currently    (Aldridge, 1958;Iguchi et al., 2007); nor is it clear how, for instance, TBTO affects the expression of so many genes and cellular processes (Baken et al., 2006(Baken et al., , 2007. The results of this study showed that treatment of thymoma cells with the model immunotoxic compound (TBTO) caused differential phosphorylations in the proteome of the treated cells compared with that of the control. However, not all differentially phosphorylated proteins were discussed because the currently unknown roles of their modified sites make it difficult to decipher their biological meaning. Among the differentially phosphorylated proteins were key enzymes that catalyze rate-limiting steps of major pathways that are sensitive to cellular energy status. These enzymes included PDH, ACC, and glutamine: fructose-6-phosphate amidotransferase 1 (GFPT1) as well as ribosomal S6 kinase 1 (RPS6K1), the latter is a downstream effecter of mTOR pathway (see Fig. 5). PDH complex plays a central role in glucose oxidation and controls the rate-limiting step that forms acetyl-CoA, which is delivered to tricarboxylic acid cycle. PDH is regulated via phosphorylation, and three specific phosphorylation sites (S232, S293, and S300) are known to inhibit its activity (Rardin et al., 2009). Though PDH was identified in both treated and nontreated cells, one of the critical phosphorylation sites, S293 that inhibits the enzyme activity was identified only in control cells (Table 1). This suggests that PDH is more active in the treated cells. Another differentially phosphorylated enzyme that was identified in the treated cells was ACC isoform 1. This enzyme catalyzes the irreversible carboxylation of acetyl-CoA to form malonyl-CoA, a metabolite used for fatty acid synthesis. The identification of two critical phosphorylation sites S77 and S79 (Table 2), known to inhibit ACC activity (Ha et al., 1994), indicates that this enzyme is inactive in TBTO-treated cells. This would allow acetyl-CoA to be delivered to tricarboxylic acid cycle rather than for fatty acid Note. The asterisk indicates that the phosphoprotein and/or the phosphorylation site was not reported in the public database of Phosphoproteinsite (http:// www.phosphosite.org). Peptides indicated in the black color were identified in both TBTO-treated and nontreated cells; peptides indicated in light blue were identified only in TBTO-treated cells, whereas those in red were identified only in control cells. synthesis. An earlier study showed that cAMP-dependent protein kinase (PRKAR1A) mediates the phosphorylation of S77, whereas AMP-activated protein kinase (PRKAA1) phosphorylates S79 (Ha et al., 1994). Both kinases were identified in the TBTO-treated cells (Table 2). Two other biosynthetic pathways affected by the TBTO exposure are hexoamine and mTOR pathways (Fig. 5). Glutamine: fructose-6-phosphate amidotransferase 1 is the enzyme responsible for the catalysis of the first and rate-limiting step reaction in the hexoamine pathway. A previous study showed that cells depleted of glucose resulted in the inhibition of this enzyme's activity via phosphorylation (Eguchi et al., 2009). This phosphorylation is mediated by AMPK (Eguchi et al., 2009). Ribosomal S6 kinase 1, a downstream effector of mTOR pathway and a target of AMPK, was downregulated in TBTOtreated cells. RPSK1 has been shown to phosphorylate 40 ribosomal S6, which is believed to regulate synthesis of proteins associated with translation machinery by facilitating the translation of the transcripts corresponding to these proteins. These transcripts are characterized by the presence of 5#-terminal pyrimidine tract and are consequently referred to as Top mRNAs (Fingar and Blenis, 2004;Zhou et al., 2010). This downregulation of RPS6K1 in the treated cells could explain the reported inhibition of protein synthesis by TBTO (Osman et al., 2009;Raffray et al., 1992).

OSMAN AND VAN LOVEREN
Since several studies showed that TBTO severely depleted ATP levels in various biological systems (Raffray et al., 1992;Soracco and Pope, 1983;Veiga et al., 1996), it is not surprising   (Corton et al., 1994), inhibits ATP-consuming anabolic pathways, while activating ATP-generating catabolic pathways in TBTOtreated cells. In relation to this energy homeostasis, it is of interest to note that a recent finding showed TBTO to be a potent inhibitor of fish peroxisome proliferators-activated receptor a/b (Colliar et al., 2011), transcription factors involved in energy homeostasis. PPARa controls the level of malonyl-CoA by inducing the expression of malonyl-CoA decarboxylase, which degrades malonyl-CoA, whereas malonyl-CoA in turn controls the rate of fatty oxidation (Campbell et al., 2002). On one hand, this observation supports the notion that organotin compounds affect energy metabolism; on the other hand, the inhibition of ACC by phosphorylation (present work) already limits the level of malonyl-CoA and therefore an increase in the expression of malonyl decarboxylase would probably be futile.
This suggests the possible existence of crosstalk at the molecular level between AMPK and PPAR a/b transcription factors. Evidence for the existence of this interlinkage between AMPK and these transcription factors comes from recent observations that AMPK could directly or indirectly interact with the cofactor PGC-1a (peroxisome proliferators-activated receptor coactivator), affecting its transcriptional capacity (Jäger et al., 2007). AMPK was shown to phosphorylate PGC-1a both in vitro and in primary muscle cells, affecting positively the expression of mitochondrial genes (Campbell et al., 2002). In contrast, in liver AMPK repressed the induction of PGC-1a by preventing the nuclear import of TORC2, the transcriptional coactivator of CREB (Shaw et al., 2005). It should be noted that PGC-1a is a potent activator of gluconeogenic gene expression in the liver, whereas AMPK suppresses this anabolic pathway (Campbell et al., 2002). These observations underline the possible regulatory role of AMPK on the gene expressions

96
OSMAN AND VAN LOVEREN mediated by PGC-1a and might also account for the observed effect of TBTO on the PPAR a/b transcription factors (Colliar et al., 2011). Additionally, since AMPKinase has been implicated in the regulation of gene expression by phosphorylating directly a number of transcription factors and coactivators as well as components of the transcriptional machinery (Leff, 2003), this kinase may play a role in the observed effect on gene expression pattern induced by TBTO (Baken et al., 2006(Baken et al., , 2007. Furthermore, we identified differentially phosphorylated proteins that are implicated in cell proliferation and growth. One of these proteins is matrin-3, a multifunctional protein of the nuclear matrix. WB analysis confirmed the downregulation of matrin-3 in TBTO-treated cells (Fig. 4B). WB analyses of immunoprecipitates of matrin-3 prepared from control cell lysates in the presence or absence of EDTA allowed us to identify two degradation products of the protein, which were markedly increased when EDTA was omitted from the lysis buffer (Fig. 4C). This result suggests that the observed downregulation of matrin-3 might be due to accentuated degradation of the protein triggered by TBTO, probably, by altering calcium homeostasis. That organotin compounds affect Ca 2þ homeostasis has been documented (Chow et al., 1992;Corsini et al. 1998;Gennari et al., 2000). Furthermore, it has been previously shown that matrin-3 mediates neuronal death following NMDA activation of rat neurons and that this protein acts as a substrate of cAMP-dependent protein kinase, whose phosphorylations leads to degradation of matrin-3 (Giordano et al., 2005). A recent study has shown that matrin-3 is a downstream substrate of caspase 3 and caspase 8 and that matrin-3 forms a complex with calmodulin-binding protein in the presence of Ca 2þ (Valencia et al., 2007). Interestingly, both cAMP-dependent kinase and caspase 3 were identified in TBTO-treated cells (Table 2). Most probably, Ca 2þ perturbation destabilizes the interaction of matrin-3 with other proteins, like CaM, and this may render matrin-3 susceptible to degradation thus explaining why the band corresponding to matrin-3 was not detectable in the enriched phosphoproteins and was barely visible in the total extracts of TBTO-treated cells (Fig. 4B). Another phosphoprotein used as a marker of cell proliferation and identified in the control cells is ribonucleotide reductase, subunit RRM2. This protein has been shown to be present in late G1/early S phase of the cell cycle, contrary to the other subunit, RRM1, whose activity is stable throughout the cell cycle (Engström et al., 1985;Heidel et al., 2007). For this protein, the phosphorylation site S20 was identified (Table 3). Earlier study showed that RRM2 is phosphorylated at the S20 site by p34Cdc2 kinase (Chan et al., 1999). Using mutagenesis, the Wnt gene assay and RNA interference Tang et al. (2007) provided evidence that RRM2 may act as inhibitor of b-catenin, a downstream effecter of Wnt signaling pathway and that phosphorylation at S20 may relieve this inhibition. In other words, in addition to DNA synthesis, RRM2 upon phosphorylation at S20 decreases the threshold of b-catenin effect, hence increasing the latter's transcriptional effect on target genes. These results suggest that TBTO affects not only cell proliferation by inducing apoptosis as described before (Osman et al., 2009;Raffray et al., 1992) but also could, at lower concentrations, affect the rate of cell cycle. The observed downregulation of phosphoproteins involved in cell cycle and proliferation, like RPS6K1, might contribute to our understanding of how TBTO suppresses immune cell activation since T-cell activation and proliferation play a crucial role in immune responses.
In conclusion, our results provide evidence that TBTO, a model immunotoxic compound, affects important pathways that are sensitive to energy charge status by modifying the phosphorylation state of key enzymes that catalyze rate-limiting steps of these pathways. Important kinases like cAMPdependent protein kinase and AMPK, which are responsible for phosphorylating some of the sites identified in these enzymes, were also identified in the treated cells. Furthermore, proteins implicated in cell proliferation were found downregulated in TBTO-treated cells, lending support to the notion that TBTO, at higher concentrations, induces cells to apoptosis and at lower concentrations slows down the rate of cell cycle.

FIG. 5.
A schematic illustration of ATP-consuming anabolic pathways inhibited by TBTO exposure. Key enzymes, such as PDH, ACC, glutamine, F-6-P amidotransferase 1 and p70RPS6 kinase 1, were affected by the treatment of thymoma cells with TBTO. Some of the phosphorylations sites identified for these enzymes are mediated by 5#-AMP-dependent protein kinase (AMPK) (Ha et al., 1994;Eguchi et al., 2009). The sign t indicates that phosphorylation inhibits the indicated enzyme.
followed by immunoblotting (A) p70RPS6Kinase1 probed with p70SPS6K1 (Thr421/Ser424)-R antibody; (B) matrin-3 probed with anti-matrin-3 goat polyclonal antibody. The loading control in the total extracts was probed with rabbit polyclonal antibody against actin; (C) WB analysis of matrin-3 immunoprecipitated from control cell lysates extracted in lysis buffer with EDTA (lane þEDTA) or without EDTA (lane ÀEDTA). The products may also result from aspecific interactions of proteins or fragments of proteins with the antibodies. Details of sample preparations and Western blotting procedures are described in ''Materials and Methods.'' 98 OSMAN AND VAN LOVEREN