Glutathione S-Transferase P Influences Redox and Migration Pathways in Bone Marrow

To interrogate why redox homeostasis and glutathione S-transferase P (GSTP) are important in regulating bone marrow cell proliferation and migration, we isolated crude bone marrow, lineage negative and bone marrow derived-dendritic cells (BMDDCs) from both wild type (WT) and knockout (Gstp1/p2 −/−) mice. Comparison of the two strains showed distinct thiol expression patterns. WT had higher baseline and reactive oxygen species-induced levels of S-glutathionylated proteins, some of which (sarco-endoplasmic reticulum Ca2+-ATPase) regulate Ca2+ fluxes and subsequently influence proliferation and migration. Redox status is also a crucial determinant in the regulation of the chemokine system. CXCL12 chemotactic response was stronger in WT cells, with commensurate alterations in plasma membrane polarization/permeability and intracellular calcium fluxes; activities of the downstream kinases, ERK and Akt were also higher in WT. In addition, expression levels of the chemokine receptor CXCR4 and its associated phosphatase, SHP-2, were higher in WT. Inhibition of CXCR4 or SHP2 decreased the extent of CXCL12-induced migration in WT BMDDCs. The differential surface densities of CXCR4, SHP-2 and inositol trisphosphate receptor in WT and Gstp1/p2 −/− cells correlated with the differential CXCR4 functional activities, as measured by the extent of chemokine-induced directional migration and differences in intracellular signaling. These observed differences contribute to our understanding of how genetic ablation of GSTP causes higher levels of myeloproliferation and migration.


Introduction
The bone marrow produces all the differentiated hematopoietic cells for peripheral blood. This tissue is extremely sensitive to alterations in redox homeostasis, as proliferation and differentiation are influenced by physiological changes in a number of factors that are sensitive to reactive oxygen species (ROS; [1,2]). ROS and the oxidation/reduction of thiols have key roles in cell signaling events that regulate a variety of biological functions [3][4][5][6][7]. Reduced glutathione (GSH) is the most abundant non-protein thiol at cellular concentrations that range from 0.1 to 10 mM [8]. Glutathione S-transferase P (GSTP) is one of a family of GST isozymes and has functions as a catalytic enzyme, protein chaperone, kinase regulator and in regulating the forward reaction of protein S-glutathionylation [1,9]. This post-translational modification occurs in certain protein clusters that have roles in events regulating cell proliferation [10]. It is reasonable to speculate that the difference in ROS levels in myeloid progenitor and quiescent hematopoietic stem cells (HSCs) may act in intracellular signaling events that drive HSC differentiation and that modulation of redox-sensitive cysteines through S-glutathionylation may have a key role in these events.
As early as 1953 a role for cysteines and thiols in bone marrow cell proliferation was established [11]. Various drugs that disrupt thiol homeostasis have been shown to exert a redox-based influence on components of bone marrow proliferation. For example, N-acetyl cysteine has been used in the management of patients with HIV, enabling more robust immune responses through T helper cells [12]; NOV-002 (a glutathione disulfide mimetic) enhances marrow recovery following immunosuppressive drugs in both rodents [13] and humans [14]; TLK199, [cglutamyl-S-(benzyl)-cysteinyl-R-(-) phenyl glycine diethyl ester] (Telintra) as a GSH peptidomimetic inhibitor of GSTP stimulates myeloproliferation in both rodents [15] and man [16,17]. In mice, Telintra increases all peripheral blood cell lineages in wild type mice as compared to GSTP-deficient mice. Even in the absence of Telintra, GSTP-null animals have increased hematopoietic progenitor cell (HPC) numbers, myeloid cell differentiation and proliferation [18]. Telintra has shown positive results (decreased requirements for red blood cell, platelet and growth factor support) in ongoing Phase 2 clinical trials for patients with myelodysplastic syndrome [16,17], a stem cell disorder characterized by ineffective blood cell production and an increased risk for transformation to acute leukemia.
Bone marrow is a relatively hypoxic tissue (1% to 2% O 2 ) [19], but within the three-dimensional marrow compartment, selfrenewing HSCs, HPCs and mature blood cells are able to migrate in a site/niche-specific fashion regulated by factors such as O 2 and Ca 2+ gradients. At any given time, approximately 75% of HSCs are in a quiescent phase of the cell cycle [20]. At the bone-bone marrow interface (osteoblastic niche), the microenvironment favors HSC quiescence, while closer to blood vessels (vascular niche), proliferation and differentiation is more likely [21][22][23][24][25]. Osteoclast and osteoblast-mediated bone remodeling results in an increased extracellular Ca 2+ in the endosteum and Ca 2+ gradient between osteoblastic and vascular niches, enabling HSCs to sense and migrate appropriately [26]. Adhesive molecules, cytokines and chemokine signaling determine population and niche characteristics. The chemokine CXCL12 plays an essential role in retaining and maintaining HSCs in bone marrow and depletion of a related cytokine, CXCR4, increases HSCs in the peripheral blood [27,28]. The interplay between ROS and thiol balance/gradients is critical to myeloproliferation and/or migration, as the redox status can be regulated by shifts of thiol-disulfide equilibrium [2]. Since pharmaceutical inhibition of GSTP has translational applications in myeloproliferation, the present studies were designed to address how genetic ablation of GSTP impacts bone marrow cell redox parameters and influences downstream events that contribute to proliferation and migration in this tissue.

Results
Increased DNA synthesis in Gstp1/p2 2/2 bone marrow cell populations Cell proliferation was assessed by BrdU incorporation. In the presence of stem/progenitor cell growth factors (stem cell factor, thrombopoietin and Flt3L), BrdU incorporation was ,10% higher in the knockout (Gstp1/p2 2/2 ) lineage-negative (Lin(2)) cells, compared to wild type (WT) Lin(2) cells. Gstp1/p2 2/2 bone marrow derived-dendritic cells (BMDDCs) had a 65% higher DNA synthesis rate than WT cells in response to GM-CSF stimulation. For both Lin(2) cells and BMDDCs, significant differences were observed between WT and Gstp1/p2 2/2 mice (Fig. 1). These results support our previous publication that ablation of GSTP either genetically or pharmacologically results in the over-production of lymphoid, erythroid and myeloid lineage cell lineages as well as platelets [26].
The impact of GSTP expression on the redox state of the intracellular protein thiol levels was further examined (Fig. 3A). Significantly higher levels of reduced intracellular protein thiols were observed in Gstp1/p2 2/2 Lin(2) cells and BMDDCs when compared to WT cells.
We also measured the in situ levels of reduced and oxidized glutathione (GSH and GSSG) in bone marrow populations derived from WT and Gstp1/p2 2/2 mice. As shown in Fig. 3B, compared to WT cells, levels of reduced GSH were higher, whereas levels of oxidized GSSG were lower in Gstp1/p2 2/2 crude BMCs, Lin(2) cells or BMDDCs. Significant differences were only observed in GSH levels in Gstp1/p2 2/2 crude BMCs and BMDDCs when compared to WT cells. In addition, a MALDI-MS bone imaging method was developed to visualize GSH and GSSG simultaneously in bone marrow tissues without any labeling (Fig. 3C). This method generated a multidimensional spatial expression map of the biomolecules directly from a tissue section. As seen in the differential ion intensities of GSH and GSSG, the bone marrow distribution results were consistent with the biochemical measurements showing that differences in qualitative distribution of GSH and GSSG were present between the WT and Gstp1/p2 2/2 samples.
Altered intracellular calcium dynamics in Gstp1/p2 2/2 bone marrow cell populations Fig. 4 shows typical calcium responses of Lin(2) cells and BMDDCs after stimulation with either A23187 or Thapsigargin (ThG). A23187 is an ionophore that forms lipid-soluble complexes with divalent metal cations, increasing specific permeability of membranes to Ca 2+ . A23187 induced a rapid increase in [Ca 2+ ] i , followed by a sustained high level of [Ca 2+ ] i in Lin(2) cells. Alternatively, only a transient increase in [Ca 2+ ] i was observed in BMDDCs, with no obvious differences between WT and Gstp1/ p2 2/2 cells (Fig. 4A). Inhibition of the sarco-endoplasmic reticulum Ca 2+ -ATPase (SERCA) pump by ThG is a commonly used method for manipulating calcium stores. ThG-induced rises in [Ca 2+ ] i reflect the passive leak of Ca 2+ from the endoplasmic reticulum (ER) following SERCA inhibition [33,34]. In the absence of extracellular Ca 2+ , ThG induced a rapid and transient elevation of [Ca 2+ ] i in both Lin(2) cells and BMDDCs. Interestingly, Gstp1/p2 2/2 Lin(2) cells and BMDDCs showed significantly lower Ca 2+ responses to ThG compared to WT cells (Fig. 4B), implying that SERCA activity may be different between WT and Gstp1/p2 2/2 cells. Supporting this conclusion, S-glutathionylation increases SERCA2 activity [35] and GSTP promotes S-glutathionylation [32]. Therefore, we investigated whether GSTP influenced the S-glutathionylation level of SERCA2 in BMDDCs. To detect S-glutathionylated SERCA2 under control conditions, immunoprecipitation with the anti-GSH antibody followed by immunoblotting with the anti-SERCA2 antibodies was used. As shown in Fig. 4C, compared with Gstp1/ p2 2/2 BMDDCs, WT cells have significant higher levels of Sglutathionylated SERCA2 under control conditions. There were no quantitative differences in SERCA2 mRNA levels or protein expression levels between WT and Gstp1/p2 2/2 cells (Fig. 4D).
Altered CXCL12/CXCR4 signaling in Gstp1/p2 2/2 bone marrow cell populations CXCL12 binds primarily to CXCR4 and initiates divergent signaling pathways downstream of ligand binding, resulting in a number of important cellular responses including chemotaxis, mobilization of intracellular calcium and activation of ERK1/2 and Akt kinases. CXCL12-mediated chemotaxis is mediated at least in part by activation of PI3 kinase/Akt and ERK1/2. Calcium mobilization by CXCL12 is achieved via phospholipase C beta activation and formation of inositol trisphosphate (IP3) and diacylglycerol. IP3/IP3 receptor (IP3R) signaling triggers the opening of the Ca 2+ channel on the surface of ER, and thus release of Ca 2+ into the cytoplasm [36,37]. We compared the results of GSTP ablation in Lin(2) cells (Fig. S1) and BMDDCs (Fig. 5) on chemotaxis, calcium mobilization, plasma membrane potential and intracellular signaling (Akt and ERK activation) after CXCL12 treatment. Both Lin(2) cells and BMDDCs migrated toward CXCL12. The number of Lin(2) cells and BMDDCs that migrated in response to CXCL12 was significantly higher than those not exposed to CXCL12 as a chemoattractant (control). This CXCL12-dependent migration was stronger in WT Lin(2) cells and BMDDCs than in Gstp1/p2 2/2 cells ( Fig. S1A and Fig. 5A).
Figs. S1B and Fig. 5B show the dynamics of intracellular calcium in Lin(2) cells and BMDDCs after stimulation with CXCL12. CXCL12 induced a rapid and transient elevation of [Ca 2+ ] i in both Lin(2) cells and BMDDCs through IP3/IP3R signaling. Interestingly, similar to the results obtained with ThGinduced calcium release from ER stores through SERCA inhibition, Gstp1/p2 2/2 Lin(2) cells and BMDDCs showed significantly different Ca 2+ responses to CXCL12 as compared to WT cells. Stronger intracellular calcium oscillations were observed in WT Lin(2) cells and BMDDCs than in Gstp1/p2 2/2 cells following CXCL12. The response in WT cells was characterized by a relatively rapid rate of rise and robust peak change in [Ca 2+ ] i , whereas the response in Gstp1/p2 2/2 cells was characterized by a relatively slow rate of rise and small peak change in [Ca 2+ ] i .
[Ca 2+ ] i affects ion permeability and membrane potential and so we determined whether different Ca 2+ responses between WT and Gstp1/p2 2/2 cells were also reflected by different potential responses under CXCL12 stimulation. Plasma membrane potential was determined using the slow-response potential-sensitive dye, bis-(1,3-diethylthiobarbituric acid)trimethine oxonol (DiS-BAC2(3)). The anionic dye can enter depolarized cells where it binds to intracellular proteins or membranes and exhibits enhanced fluorescence. Increased depolarization results in additional influx of the anionic dye and thus, an increase in fluorescence. Conversely, hyperpolarization is indicated by a decrease in fluorescence. Fig. S1C and Fig. 5C show the dynamics of plasma membrane potential in Lin(2) cells and BMDDCs following stimulation with CXCL12. Following CXCL12 stimulation, depolarization of plasma membranes was detected in both WT and Gstp1/p2 2/2 Lin(2) cells, whereas delayed, amplified and reversible plasma membrane depolarization was observed in Gstp1/p2 2/2 cells (Fig. S1C). For BMDDCs, CXCL12 caused membrane hyperpolarization in both WT and Gstp1/p2 2/2 cells, and this hyperpolarization was stable during incubation. The effects of CXCL12 on membrane potential were slightly more pronounced in WT BMDDCs (Fig. 5C).
Next we assessed whether the different chemotactic responses of WT and Gstp1/p2 2/2 cells following CXCL12 stimulation were also reflected at other steps of intracellular signaling. As shown in Fig. 5D, treatment of BMDDCs with CXCL12 induced a rapid phosphorylation and activation of Akt and ERK in both WT and Gstp1/p2 2/2 cells. Compared to Gstp1/p2 2/2 cells, more intense phosphorylation of both Akt and ERK was observed in WT cells after CXCL12 stimulation. At 15 mins, pERK levels did drop compared to pAkt, perhaps reflecting the slightly different roles that these kinases play in regulating proliferative pathways.
To find a possible explanation for the differential chemotactic responsiveness and intracellular signaling capacity between WT and Gstp1/p2 2/2 cells, we determined chemokine CXCL12 receptor CXCR4, IP3 receptor IP3R mRNA levels as well as protein expression ( Fig. 5E and F). Both WT and Gstp1/p2 2/2 BMDDCs expressed CXCR4 and IP3R, with significantly higher mRNA and protein expression levels of CXCR4 and IP3R3 in WT than in Gstp1/p2 2/2 cells, likely linked to the different redox status of those cells. In addition, existing evidence implies that expression of the thiol active phosphatase SHP-2 is important in regulating CXCR4 signaling [38,39] and our data indicate that SHP-2 expression was lower in Gstp1/p2 2/2 than in WT cells (Fig. 5E). Furthermore, we analyzed the effect of SHP-2 specific inhibitor PHPS1 [40] and CXCR4 specific antagonist AMD3100 on CXCL12-induced migration of BMDDCs. PHPS1 or AMD3100 treatment significantly reduced the CXCL12-induced migration in both WT and Gstp1/p2 2/2 BMDDCs, although it had a greater effect on the WT cells (Fig. 5G). Taken together, the differential surface density of CXCR4 and SHP-2 in WT and Gstp1/p2 2/2 cells may correlate with the differential CXCR4 functional activity measured by the extent of CXCL12-induced cellular responses in hematopoietic cells.

Discussion
The present studies were undertaken to determine why either genetic ablation or pharmacological inhibition of GSTP enhances myeloproliferation and migration, producing increased numbers of all committed cell lineages [15,18]. Existing literature implies that ROS can regulate certain bone marrow niches and influence proliferative status [1]. Consequently, we undertook a series of studies to determine how various bone marrow cell types from Gstp1/p2 2/2 mice differed from their WT counterparts. Because ROS conditions have direct influence on cellular redox homeostasis, a number of thiol-dependent pathways are implicated, and these downstream targets are influenced by the presence/absence of GSTP.
Our characterization of bone marrow cells identified a number of consistent differences between WT and Gstp1/p2 2/2 cells with respect to thiol balance. For example, since GSTP can catalyze the WT or Gstp1/p2 2/2 mice were exposed to various concentrations of H 2 O 2 (1, 5 or 10 mM) (left panel) or PABA/NO (0.5, 1 or 2 mM) (right panel), at 37uC for 15 min. Proteins were separated by non-reducing SDS-PAGE and protein S-glutathionylation evaluated by immunoblot with mouse anti-GSH antibodies. Even loading of protein was confirmed by probing with rabbit anti-Cdk2 antibodies. Goat anti-mouse and goat anti-rabbit fluorescent secondary antibodies were used and immunoblots imaged with a two-channel IR fluorescent Odyssey CLx imaging system. S-glutathionylation levels were quantified and normalized relative to Cdk2 loading for each concentration point. The ratio from no drug treatment was assigned a value of one and all other ratios are shown relative to this value. Bars represent the means (6SD) from three independent experiments. doi:10.1371/journal.pone.0107478.g002 forward reaction of S-glutathionylation [32,41], its absence diminished the capacity of cells to establish high levels of either general or specific protein S-glutathionylation ( Fig. 2 and Fig. 4C).
Presumably as a consequence of this, free protein thiols and GSH levels were higher in resting BMCs, BMDDCs and Lin(2) cells from Gstp1/p2 2/2 mice ( Fig. 3A and B). Moreover, in this report, we demonstrate for the first time that MALDI-MS bone imaging can be used for the simultaneous in situ visualization of both GSH and GSSG in sectioned bones with an intact bone marrow compartment (Fig. 3C). These results, while predominantly qualitative in nature, confirm the biochemical analyses that detail differences between GSH/GSSG in WT and Gstp1/p2 2/2 cell lineages and implicate how such differences might influence downstream events in proliferation, mobilization and/or migration of cells within the bone marrow compartment. Coincident with the S-glutathionylation differences between WT and Gstp1/ p2 2/2 cells, we performed a quantitative proteome analysis using cysteine reactive ICAT labeling with Orbitrap mass spectrometry and identified other thiol active proteins, the expression levels of which were different between Gstp1/p2 2/2 and WT BMDDCs. Within this group (Table 1) were a number of redox active enzymes (aldoketo-reductase family 1, peroxiredoxin 4, prolyl 4-hydroxylase, triose-phosphate isomerase 1, transitional endoplasmic reticulum ATPase); cytoskeletal proteins (vimentin); apoptotic proteins (apoptotic chromatin condensation inducer 1); calcium binding protein (S100 calcium binding protein A9); regulatory proteins (rho GTPase-activating protein 17); scaffold proteins (ras GTPase-activating-like protein IQGAP1) and signaling proteins (serine/threonine kinase 16). Quantitative expression differences could indicate a number of other ways that GSTP genotype could influence bone marrow cell function. In particular, although their specific role in marrow cell proliferation has yet to be addressed, the relevance of S100 and kinases seems self-evident. In addition, the altered expression of IQGAP1 is interesting in light of our recent data showing that sulfiredoxin (a redox active protein with roles in deglutathionylation) has sequence homology to IQGAP proteins and involved in controlling cell migration [42,43].
The redox status in cells is crucial in the regulation of the chemokine system, exemplified by the observations that antioxidants decrease chemokine receptor expression and chemotaxis, while H 2 O 2 or general GSH-depletion, increases chemokine receptor expression and chemotactic responses [44]. Accordingly, while both WT and Gstp1/p2 2/2 BMDDCs expressed CXCR4, the former had significantly higher mRNA and protein expression levels ( Fig. 5E and F)), implying a link to the different redox status of the cells. The thiol-dependent phosphatase SHP-2 regulates CXCR4 signaling [38,39]. Previous studies have shown that SHP-2 functions as an adaptor molecule, which can bind to several proteins and then transduce various proliferation/migration signals. CXCR4, SHP-2 and cbl collectively participate in the formation of a multimeric signaling complex and over-expression of SHP-2 increases CXCL12-induced chemotaxis, whereas phosphatase inhibitors significantly inhibit CXCL12-induced migration [38]. The SHP-2 specific inhibitor PHPS1 or CXCR4 specific antagonist AMD3100 significantly reduced CXCL12induced migration of WT BMDDCs, and had greater effect on the WT cells (Fig. 5G). Moreover, the phosphatase activity of SHP-2 is subject to regulation by S-glutathionylation in vitro [45] and the redox modifications could prove to be physiologically significant and pertinent to the observed differences between WT and Gstp1/ p2 2/2 bone marrow cells (where SHP-2 expression was lower). Taken together, the differential surface density of CXCR4 and SHP-2 in WT and Gstp1/p2 2/2 cells could be correlated with the differential CXCR4 functional activity measured by the extent of chemokine-induced directional migration and the intracellular signaling capacity (e.g. PI3K, MAPKs, calcium oscillation, etc.) ( Fig. S1 and Fig. 5A-D). It should be noted that CXCR4 is expressed in multiple cell types in the immune and central nervous systems, hematopoietic stem/progenitor cells, endothelial and epithelial cells and cancer cells. Its ligand, CXCL12, is expressed/ secreted in various tissues and organs (bone marrow, liver, lung, skin, skeletal muscle, brain, kidney and heart). CXCL12/CXCR4 signaling plays an important and unique role in the regulation of stem/progenitor cell trafficking, inflammation, embryo/organogenesis, tissue/organ regeneration, and tumor progression, angiogenesis, metastasis, and survival [37,46,47]. Thus, our observed link between GSTP and different levels/activities of CXCR4 and SHP-2 is likely functional.
The calcium ionophore A23187-caused a rapid influx of extracellular Ca 2+ into BMDDCs, a tightly regulated process that was quickly accompanied by restoration to basal levels (most likely through Ca 2+ influx into intracellular store(s)) (Fig. 4A). However, in Lin(2) cells, the influence of A23187 on Ca 2+ permeability and uptake is not regulated. ThG is a specific SERCA inhibitor [34] and caused a rapid increase and subsequent slow restoration of basal level of intracellular Ca 2+ only in WT Lin(2) cells (in the absence of extracellular Ca 2+ ). Treatment of BMDDCs with ThG showed similar kinetics, but with higher amplitude and a strong GSTP-dependent effect on intracellular Ca 2+ flux (Fig. 4B). Our data show equivalent level of SERCA2 expression and Sglutathionylation in BMDDCs ( Fig. 4C and D). S-glutathionylation of Cys674 (located in the cytosol-facing hinge domain) of SERCA provides a physiological, cGMP-independent mechanism of activating intracellular Ca 2+ fluxes. Over-oxidation of this cysteine (sulfinic/sulfonic acid) prevents its S-glutathionylation and impairs the dynamics of intracellular Ca 2+ flux [35]. The absence of GSTP in Gstp1/p2 2/2 BMDDCs results in decreased SERCA2 S-glutathionylation and consequent diminished activation (Fig. 4C). In contrast, GSTP mediates a faster increase and return of intracellular Ca 2+ to its basal levels in BMDDCs (Fig. 4B).
Amplitude of a CXCL12-mediated oscillation of intracellular Ca 2+ through its release from and influx into intracellular stores (E(S)R, mytochondria, etc) is proportional to GSTP level and is similar in Lin(2) cells (Fig. S1B). Both the IP 3 R channel and the plasmalemmal Ca 2+ -ATPase (PMCA) pump can be reversibly Sglutathionylated. IP3R channel activity is enhanced by glutathionylation, whereas PMCA pump activity is inhibited [48]. Our data are consistent with the hypothesis that GSTP can promote Sglutathionylation of the IP 3 R and PMCA causing Ca 2+ release from IP 3 -sensitive internal Ca 2+ stores and elevation of basal intracellular Ca 2+ levels -in the absence of extracellular Ca 2+ [49]. In Lin(2) cells, these effects cause an initial plasma membrane depolarization and GSTP-dependent additional depolarization/ hyperpolarization in the presence of extracellular Ca 2+ (Fig. S1C), most likely indicating a Ca 2+ -induced Ca 2+ -release effect. The amplitude of CXCL12 effects on mobilizing intracellular Ca 2+ in BMDDC is also GSTP dependent and similar to that in Lin (2) cells, but with minimal oscillations in the absence of extracellular Ca 2+ (Fig. 5B). CXCL12 can also cause a continuous, GSTPindependent hyperpolarization of plasma membranes in Table 1. List of selected proteins and their abundance ratios between WT and Gstp1/p2 2/2 BMDDCs. BMDDCs (in the presence of extracellular Ca 2+ , Fig. 5C). This conclusion is compatible with differences of plasma membranemediated Ca 2+ fluxes in Lin(2) cells and BMDDCs and would fit the model where CXCL12-mediated intracellular Ca 2+ dynamics and plasma membrane depolarization is linked with differences in GSTP mediated protein S-glutathionylation. Overall, genetic ablation of GSTP is causally linked with multiple events that contribute to the regulation of bone marrow cell proliferation and migration. A common link is the perturbation of redox homeostasis and those documented in this paper help to explain why Telintra, as an inhibitor of GSTP, has clinical activity as a small molecule myeloproliferative drug [16,17].

Materials and Methods
Mice C57BL/6 wild type mice were purchased from Jackson Laboratory (Bar Harbor, ME). Gstp1/p2 2/2 mice were generated as described earlier [50]. The mice were bred and kept in the

Primary cells and culture conditions
The femurs and tibias were harvested from WT or Gstp1/p2 2/2 mice immediately after cervical dislocation. Crude bone marrow cells were flushed from the bones into RPMI-1640 culture medium (HyClone, Logan, UT) using 26-gauge needles and 10-ml syringes, and filtered through a 40 mm nylon cell strainer (BD, Franklin Lakes, NJ) to prepare single-cell suspensions. Red blood cells were then lysed with ACK (Ammonium-Chloride-Potassium) lysis buffer (Gibco, Life Technologies, Carlsbad, CA).
Bone marrow derived-dendritic cells were generated according to previous reported procedures [51,52] with minor modifications. Murine BMCs (4-5610 5 /ml, 10 ml/plate) were plated in RPMI-1640 supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, 100 mg/ml streptomycin (all from Mediatech, Manassas, VA), and 20 ng/ml recombinant mouse GM-CSF (BioAb-Chem, Ladson, SC) (DC medium) into 100-mm culture dishes (Sarstedt, Newton, NC). Fresh DC medium was added on day 4 and was gently replaced by fresh DC medium containing 10 ng/  Representative immunoblots show phosphorylation of Akt and ERK in WT and Gstp1/p2 2/2 BMDDCs after stimulation with CXCL12 for the indicated periods. p-Akt or p-ERK levels were quantified and normalized relative to total Akt or ERK protein for each time point. The ratio (p-Akt:total Akt or p-ERK:total ERK) at 0 min CXCL12 exposure was assigned a value of one and all other ratios are shown relative to this value. Bars represent the means (6SD) from three independent experiments. (E, F) Impact of GSTP ablation on CXCR4, IP3R and SHP-2 levels in BMDDCs. Total protein levels were evaluated by immunoblots. Actin served as a loading control. Surface expression of CXCR4 was further analyzed by flow cytometry. Representative histograms are shown: CXCR4 expression in WT and Gstp1/p2 2/2 BMDDCs compared to isotype control showing the mean fluorescence intensity (MFI) and statistical summary of three independent measurements presented as means (6SD) with asterisks (*) indicating statistically significant differences between WT and Gstp1/p2 2/2 BMDDCs (p,0.05). (E). Relative gene expression levels were quantified by Real-Time RT-PCR. Bars represent the means (6SD) from three independent experiments (F). (G) Effect of SHP-2 specific inhibitor PHPS1 and CXCR4 specific antagonist AMD3100 on CXCL12-stimulated chemotaxis of BMDDCs. WT and Gstp1/p2 2/2 BMDDCs were either untreated (control) or pretreated with 10 mM PHPS1 or 1 mM AMD3100 for 2 h, and then cells were either left untreated or stimulated with 200 ng/ml CXCL12 for 6 h. Values are average percentages of specific chemotaxis (percentage of cells migrating to medium alone was subtracted from the percentage of cells migrating to medium with CXCL12) (6SD) from three independent experiments, with asterisks (*) indicating statistically significant differences between control and drug treatments (p,0.05). Chemotaxis of WT cells in the absence of inhibitor is considered 100% migration. doi:10.1371/journal.pone.0107478.g005 sion Medium (StemCell technologies, Vancouver, BC, Canada) supplemented with 100 ng/ml recombinant mouse stem cell factor (Sigma, Saint Louis, MO), thrombopoietin, and Flt3L (both from BioAbChem, Ladson, SC) (Lin(2) cell medium).

Proliferation assay
Cell proliferation was assessed using a BrdU Cell Proliferation assay kit (Cell Signaling Technology, Danvers, MA). In brief, WT or Gstp1/p2 2/2 Lin(2) cells or BMDDCs were seeded at 0.2610 6 cells/ml in Lin (2) cell medium or DC medium in a 96-well or 24-well plate and incubated for 48 h. BrdU solution (final concentration 10 mM) was then added and the cells kept at 37uC in the incubator for 4 h. The incorporation of BrdU into the DNA was measured by means of monoclonal anti-BrdU antibodies in a cellular ELISA format according to the manufacturer's protocol.

Cell treatment with H 2 O 2 Or PABA/NO
Wild type and Gstp1/p2 2/2 BMCs or BMDDCs (1610 6 cells/ ml) suspended in complete medium (RPMI-1640 supplemented with 10% FBS, 100 U/ml penicillin and 100 mg/ml streptomycin) were exposed to different concentrations of H 2 O 2 (1, 5 or 10 mM) or PABA/NO (0.5, 1 or 2 mM), at 37uC for 15 min as indicated in individual experiments. The reaction was terminated by immediate centrifugation at 500 g, 4uC for 5 min. Supernatants were removed and pellets washed once with ice-cold PBS and then solubilized by ice-cold lysis buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Triton, 1 mM EDTA, 1 mM EGTA, 5 mM NEM, plus a protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN)]. Cell lysates/supernatants were collected from the cells after spinning at 16,000 g for 10 min at 4uC and used for immunoblotting.

Cell treatment with CXCL12
Wild type and Gstp1/p2 2/2 BMDDCs were harvested, resuspended at a concentration of 1610 6 cells/ml in RPMI-1640 and kept in the incubator at 37uC for another 2 hours to reduce the basal activity of intracellular signaling pathways. Then BMDDCs were stimulated with chemokine, 200 ng/mL CXCL12 (BioAbChem) for various times (2-15 min). Stimulation was terminated by centrifugation at 500 g, 4uC for 5 min. Supernatants were removed and pellets washed once with ice-cold PBS and then solubilized by ice-cold lysis buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Triton, 1 mM EDTA, 1 mM EGTA, 40 mM b glycerophosphate, 5 mM sodium pyrophosphate, 5 mM sodium fluoride, 2 mM sodium orthovandate, plus a protease inhibitor cocktail]. Cell lysates/supernatants were collected from the cells after spinning at 16,000 g for 10 min at 4uC and used for immunoblotting.

Flow cytometry
Cell surface CXCR4 expression was determined by fluorescence-activated cell sorter (FACS) analysis using FITC-conjugated rat anti-mouse CXCR4 mAb or FITC-conjugated rat IgG2b (both from BD Pharmingen) as isotype control. A total of 0.5610 6 cells were incubated with primary antibodies in FACS buffer (2% FBS in PBS) for 30 min at room temperature, washed once in FACS buffer, resuspended in 250 ml of FACS buffer and analyzed on a BD FACSCalibur analytical flow cytometer and CellQuest Pro software (BD Pharmingen). After gating on CD11c (APCconjugated hamster anti-mouse CD11c mAb, Affymetrix, eBioscience, San Diego, CA) positive cells, 20,000 events per sample were analyzed.

Immunoprecipitation of S-glutathionylated SERCA2
Five hundred micrograms of protein from WT or Gstp1/p2 2/2 BMDDCs lysates were pre-cleared by incubation with protein A/ G-agarose beads (Santa Cruz Biotechnology) for 1 hour at 4uC. After the removal of protein A/G-agarose beads by brief centrifugation, the samples were incubated with mouse monoclonal anti-GSH antibody for 2 hours at 4uC. The antibody-antigen complexes were immunoprecipitated by incubating with protein A/G-agarose beads overnight at 4uC. Non-specific bound proteins were removed by washing protein A/G-agarose beads once with ice-cold lysis buffer and twice with PBS. Bound immunocomplexes were solubilized in SDS-loading buffer and analyzed by subsequent SDS-PAGE, probing of the immunoblots with goat polyclonal anti-SERCA2 antibody (Santa Cruz Biotechnology). As a reagent control, cell lysates were incubated with isotype control IgG and subjected to the same procedures.

RNA isolation and Real-Time RT-PCR analysis
Total RNA was prepared using the Isolate II RNA mini kit (Bioline, Taunton, MA) and cDNA was then generated with the iScript cDNA synthesis kit (Bio-Rad) according to the manufacturers' protocols. Subsequently, quantification of gene expression was performed in duplicates using iQ SYBR Green supermix (Bio-Rad) with detection on an MyiQ Real-Time PCR System (Bio-Rad). The reaction cycles used were 95uC for 5 min, and then 40 cycles at 95uC for 15 s and 58uC for 1 min followed by melt curve analysis. The following primers were used: CXCR4, forward 59-TCAGTGGCTGACCTCCTCTT-39,

Measurement of GSH and GSSG levels
BMCs, Lin(2) cells or BMDDCs were harvested and cell lysates were prepared in the same way as described above. Reduced GSH was determined by a fluorometric method (Ex = 379 nm, Em = 513 nm), using Cayman Thiol Detection Assay Kits (Ann Arbor, Michigan) according to the manufacturer's protocol with few modifications. Proteins were precipitated from cell lysates by adding 5% trichloroacetic acid (Sigma-Aldrich, St. Louis, MO). After 10 min at 220uC, supernatants were collected after spinning at 16,000 g for 10 min at 4uC. GSH standards (0.0625 to 5 mM) were prepared by dilution in dH 2 O. The concentrations of GSH in cell lysates are expressed in nmol/mg protein. The determination of oxidized GSSG is based on its reduction by glutathione reductase using NADPH as a source of electrons and on measurement of NADPH fluorescence decrease at 37uC. The final concentrations of reagents were 2.66 mg/ml glutathione reductase and 50 mM NADPH in 1.5 ml 50 mM Tris-HCl pH 8.0 containing 1 mM EDTA and 10 mM sodium azide. The enzymatic reaction was monitored for 50 seconds before and until 300 seconds after the addition of each sample or GSSG standard with excitation at 340 nm and emission at 460 nm (resolution 0.1 sec). The concentrations of GSSG in cell lysates are expressed in nmol/mg protein. In both cases, the fluorescence was recorded by a QM-4 spectrofluorometer in standard kinetic mode as described above. GSH, GSSG, NADPH and glutathione reductase were all purchased from Sigma-Aldrich (St. Louis, MO).

ICAT labeling and nanohplc-Orbitrap Elite mass spectrometry
BMDDCs were harvested and the pellet was dissolved in ICAT denaturation buffer [100 mM Tris-HCl (pH 8.5), 4 M urea, 0.05% SDS, plus a protease inhibitor cocktail]. Protein (100 ug) was treated with 1.25 mM TCEP to reduce the cysteine residues and labeled with cleavable, cysteine reactive ICAT reagents according to the manufacturer's instructions (AbSciex, Foster city, CA). Protein from WT and Gstp1/p2 2/2 mouse was labeled with 'light (L)' and 'heavy (H)' ICAT reagents, respectively. ICAT labeling and nano-HPLC-Orbitrap Elite MS were performed at the Proteomics Core Facility of the Medical University of South Carolina. Protein identification was performed using Mascot and Sequest nodes within the Proteome Discoverer 1.3 platform (Thermo Scientific) against a RefSeq murine database (28925 entries, downloaded 03/05/12). Parameters for peptide identification were as follows: precursor mass tolerance of 100 ppm, fragment mass tolerance of 0.8 Da, one missed cleavage, dynamic modifications of cysteine with light or heavy labeled ICAT reagent, deamidation of asparagine and glutamine, and methionine oxidation. Only high confidence peptides with a false discovery rate (FDR) ,1% were included. The ICAT quantitation was performed using Proteome Discoverer 1.3 by calculating the relative intensities of the light and heavy ICAT labeled pairs. All ratios were normalized using the median ratio from all peptides identified.

MALDI-MS bone imaging
The femurs of mice were harvested and snap frozen in the vapor phase of liquid nitrogen for at least 10 min, and then stored at -80uC until analysis. A cryostat (Microm HM550, Thermo) was used to slice the frozen bones to a thickness of 10 mm. The bone slice was attached to double sided carbon tape (SPI supplies, West Chester, PA), then mounted onto a conductive indium tin oxide coated (ITO) slide (Bruker Daltonics, Billerica, MA). Slides were desiccated at room temperature for at least 15 min before matrix application. MALDI matrix (9-aminoacridine, 5 mg/ml in 70% ethanol) was applied to the bone slides using an ImagePrep spray station (Bruker Daltonics).
MALDI-MS analysis was performed using a Solarix 7T dual source ESI/MALDI Fourier transform ion cyclotron resonance (FTICR) mass spectrometer (Bruker Daltonics). Acquisition was set up using FlexControl 3.0 and FlexImaging 4.0 software (Bruker Daltonics). Glutathione metabolites were detected in negative ion mode analyzing the mass range m/z = 50-800 with a SmartBeam II laser operating at 1000 Hz, a laser spot size of 25 mm, and a raster width of 200 mm. For each laser spot, 500 spectra were averaged and all data was normalized using root means square. Images were generated using FlexImaging 4.0 software (Bruker Daltonics). Structural confirmation of GSH and GSSG was done by collision-induced fragmentation of the indicated ions.

Chemotaxis assay
Chemotaxis assays were performed in 24-well transwell chambers with polycarbonate membranes (5 mm pore size; Corning, NY) as described by Branham-O'Connor et al [52]. In brief, 600 ml of either chemotaxis buffer (RPMI-1640 with 200 ng/ml CXCL12) or as a control RPMI-1640 alone were placed in the lower chambers. Upper chambers were loaded with 100 ml cell suspensions of Lin(2) cells or BMDDCs at a concentration of 3610 6 cell/ml in RPMI 1640. Where indicated, BMDDCs were preincubated with 10 mM PHPS1 or 1 mM AMD3100 (both from Calbiochem) in serum-free RPMI for 2 h at 37uC prior to measuring chemotaxis. The complete chamber was kept at 37uC in the incubator for 6 or 16 h. After that, cells remaining in the upper chambers and cells that had migrated through the membrane to the lower chambers were counted with a Z1 Coulter Particle Counter (Beckman Coulter, Hialeah, FL). The percentage of migration was determined as migrated/total cells. To calculate the percentage of specific migration induced by chemokine CXCL12, the percentage of cells migrating to medium alone (control) was subtracted from the percentage of cells migrating to medium with CXCL12.

Intracellular calcium oscillations
Lin(2) cells or BMDDCs were harvested, re-suspended at a concentration of 1610 6 cell/ml and incubated with 5 mM Fluo-3-AM (Invitrogen, Life Technologies, Carlsbad, CA) at 37uC for 45 min in RPMI-1640 at 37uC in the dark. All subsequent manipulations were performed with the Fluo-3-AM labeled cells protected from light. The cells were washed three times with PBS to remove extracellular dye and then re-suspended at a concentration of 1610 6 cell/ml in PBS, containing 100 mM CaCl 2 (Ca 2+ -extracellular solution) or 2 mM EGTA (zero Ca 2+extracellular solution) immediately prior to use. The kinetics of intracellular free ionized calcium ([Ca 2+ ] i ) changes were measured using a QM-4 spectrofluorometer and standard kinetic mode with excitation at 506 nm and emission at 526 nm at 37uC under constant stirring. The emission of each sample was recorded for ,150 seconds before and until 500 seconds after the addition of 1 mM A23187 (0.1% DMSO); 1 mM ThG (0.1% DMSO) (both from Sigma-Aldrich, St. Louis, MO); 100 ng/ml CXCL12. A23187 (1 mM) and ThG (1 mM) stock solutions were prepared in DMSO, while CXCL12 (100 mg/ml) was reconstituted in PBS containing 0.1% BSA. All fluorescent measurements were corrected for fluorescence of unlabeled cells. Representative traces of three independent experiments were averaged and smoothed using standard Sigma-Plot 10.0 software (Systat Software, San Jose, CA).

Plasma membrane potential measurements
Plasma membrane potential was determined using the slowresponse potential-sensitive dye DiSBAC2(3) (Invitrogen, Life Technologies, Carlsbad, CA). Lin(2) cells or BMDDCs were harvested, washed three times with PBS, and re-suspended at a concentration of 0.5610 6 cell/ml in PBS with 100 mM CaCl 2 . Cell suspensions were incubated with 5 mM (final concentration) DiSBAC2(3) in a quartz cuvette under constant stirring at 37uC for 3 min in the dark. Then the kinetics of the emission at 560 nm (Ex = 530 nm) were monitored by a QM-4 spectrofluorometer. The emission of each sample was recorded for 100 seconds before and until 500 seconds after the addition of 100 ng/ml CXCL12. Representative traces of three independent experiments were averaged and smoothed using standard Sigma-Plot 10.0 software (Systat Software, San Jose, CA).

Statistical analysis
Student's t tests were used where P values,0.05 were regarded as statistically significant. Data were expressed as means 6 SD with n equal to the number of animals/group examined under each condition. Figure S1 Lin(2) cell responses to CXCL12. (A) Chemotaxis of Lin(2) cells to CXCL12. Wild type and Gstp1/p2 2/2 Lin(2) cells were either untreated (control) or stimulated with 200 ng/ml CXCL12 for 6 h. Values are average percentages of migration (6SD) from three independent experiments, with asterisks (*) indicating statistical significant differences between WT and Gstp1/p2 2/2 Lin(2) cells (p,0.05). To calculate the specific chemotaxis induced by CXCL12, the percentage of cells migrating to medium alone (control) was subtracted from the percentage of cells migrating to medium with CXCL12. (B, C) Intracellular calcium (B) and plasma membrane potential (C) dynamics in WT and Gstp1/p2 2/2 Lin(2) cells in response to CXCL12. The arrows indicate the addition of CXCL12. Data are representative traces of three independent experiments. (TIF)