Spatial analysis of key signaling proteins by high-content solid-phase cytometry in Hep3B cells treated with an inhibitor of Cdc25 dual-specificity phosphatases.

effects 2-(2-mercaptoethanol)-3-methyl-1,4-naphthoquinone, on the spatial regulation and activation kinetics of tyrosine phosphorylation-dependent signaling events using two methods: i) high-content, fluorescence-based, automated solid phase cytometry and ii) a novel cellular assay for Cdc25A activity in intact cells. Immunofluorescence studies demonstrated a concentration-dependent nuclear accumulation of phospho-Erk and phospho-p38, but not NFkB. Immunoblot Erk phosphorylation and to its the transcription factor Elk-1. Pretreatment of the by Compound of nuclear phospho-Erk, but not phospho-p38 accumulation, and protected cells of of Cdc25A in dephosphorylation of Erk that was reversed by 5. The data show that an inhibitor of Cdc25 increases Erk phosphorylation and nuclear accumulation and support the hypothesis that Cdc25A regulates Erk phosphorylation status. demonstrated that this inhibitor selectively activated dual-specificity phosphatase-dependent cellular events. Subsequent analyses using both genetic and pharmacological tools identified activation of the Erk pathway as the dominant component mediating Compound 5’s antiproliferative activity, and provided direct evidence that it could interfere with Cdc25A function in the cell. We propose that the combination of high-content, cell-based analyses coupled with a chemical complementation approach will be a powerful technique to identify cell active inhibitors of a variety of cellular targets.


Introduction
Approximately one third of mammalian proteins are thought to be posttranslationally modified by phosphorylation (1). The human genome contains hundreds of protein kinases (2), but the reversibility of the phosphorylation process suggests that phosphatases also play a major role in the regulation of protein phosphorylation. While the roles and cellular functions of kinases have been extensively studied, protein phosphatases have received much less attention. A long held view has been that phosphatases merely serve to reverse the actions of protein kinases. More recently has it been recognized that phosphatases may be as numerous and as tightly regulated as protein kinases, with as widely varying substrate specificities and signaling functions (3).
The preference of certain phosphatases for one phosphorylated hydroxy amino acid over others has resulted in the current classification of phosphatases as serine/threonine specific (STPase) 1 , tyrosine-specific (PTPases), and dual-specific (DSPases) phosphatases. While several highly potent and selective inhibitors of STPases have been isolated from natural sources, selective PTPase or DSPase inhibitors are still rare. 6 which further supported the hypothesis that Cdc25A regulates Erk phosphorylation status. 8 (Packard ViewPlate TM ), and allowed to attach overnight. Cells were treated for the times indicated with Compound 5 or IL-1α, fixed with 3.7% formaldehyde in PBS and permeabilized with PBS/Triton X-100. Cells were stained with antibodies against phospho-Erk, phospho-p38, phospho-JNK, or the 65 kDa subunit of NFkB, and washed with PBS/Tween20. Nuclei were stained with Hoechst 33342 fluorescent dye, and immunoreactive cells were visualized by AlexaFluor 488 secondary antibodies (Molecular Probes, Eugene, OR) using an XF100 filter set at excitation/emission wavelengths of 494/519 nm (Alexa 488), and 350/461 nm (Hoechst), respectively. Plates were analyzed by automated image analysis on the ArrayScan II system (Cellomics, Pittsburgh, PA) using the previously described cytoplasm-to-nuclear translocation algorithm (18). Control experiments omitting primary antibodies were performed each time to assess the amount of non-specific background staining.
Cell Fractionation and Western Blotting. Cytosolic and nuclear fractions were prepared using a slightly modified procedure as published by Schreiber et al. (19).
Hep3B cells were plated in 100 mm tissue culture dishes, exposed to 10 µM Compound 5 for the indicated periods of time, and harvested by centrifugation. Cell pellets were resuspended in 200 µl of hypotonic buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 0.2 mM PMSF, 0.5 mM DTT, and 0.5% Nonidet-P-40), incubated on ice for 10 min, disrupted by repeated aspiration through a 20-gauge needle, and centrifuged at 2,500 x g for 15 min. The supernatant was collected as cytosolic extract. Nuclear pellets were resuspended in nuclear extraction buffer (20 mM HEPES, pH 7.9, 10% glycerol, 1.5 mM ice for 1 h, and centrifuged at 13,000 x g to collect the nuclear fraction. Solubilized proteins were resolved by 10% SDS-PAGE and transferred to polyvinylidene difluoride membranes (NEN, Boston, MA). Membranes were probed with anti-phospho-Erk, anti-Oct-1, and anti HSP-90 antibodies. Positive antibody reactions were visualized using peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA) and an enhanced chemiluminescence detection system (Renaissance, NEN, Boston, MA) according to manufacturer's instructions.
Erk activity assay. Erk activity in cytosolic and nuclear fractions was determined using a non-radioactive immunoprecipitation kit (Cell Signaling Technologies, Beverly, MA).
Briefly, 200 µg of nuclear or cytosolic proteins were incubated with 15 µl of agaroseconjugated anti-phospho-Erk antibody, and incubated overnight at 4 o C with gentle rocking. Immunoprecipitates were pelleted and washed twice with kinase buffer (25 mM Tris, pH 7.5, 5 mM β-glycerophosphate, 2 mM DTT, 0.1 mM Na 3 VO 4 , and 10 mM MgCl 2 ). Pellets were resuspended in 50 µl of kinase buffer supplemented with 200 µM ATP and 2 µg of Elk-1 GST fusion protein, and incubated for 30 min at 30 o C.
Immunoprecipitates were boiled in SDS-PAGE sample buffer and analyzed by Western blot using an anti-phospho-Elk-1 antibody.
Growth inhibition assay. The antiproliferative activity of Compound 5 in combination with the MEK inhibitor, U-0126, was measured by a previously described assay based on fluorimetric quantitation of total cellular DNA content using the fluorochrome Hoechst for three days with various concentrations of Compound 5 in the presence or absence of the MEK inhibitor, U-0126 (5 µM). Cells were lysed by repeated freeze-thawing and cellular DNA was quantitated as described (20). Three hours after transfection, the medium was replaced with complete growth medium and the cells were allowed to recover for 48 h. Cells were treated with 0-20 µM Compound 5 for 30 min and protein lysates were prepared and analyzed by SDS-PAGE and Western Blot analysis for phospho-Erk and Erk 2 levels as described above. For quantitation of protein expression levels, X-ray films were scanned on a Molecular Dynamics personal SI densitometer and analyzed using the ImageQuant software package (Ver. 4.1, Molecular Dynamics, Sunnyvale, CA).

A fluorescence-based high-content assay for phospho-Erk nuclear accumulation.
Compound 5 was previously found to induce the prolonged phosphorylation of tyrosines on a number of signaling proteins in the Erk cascade, including Erk1 and Erk2 (13,14).
We first asked whether this increase in tyrosine phosphorylation was associated with a change in phospho-Erk nuclear accumulation. Hep3B cells were either incubated with  resulted in a substantial increase in total phospho-Erk, with prominent nuclear accumulation (Fig. 1E). Overlay images ( Fig. 1C and F) illustrate the quantitation of cytoplasmic and nuclear phospho-Erk levels. Fluorescence-labeled cells were analyzed in two separate channels by the ArrayScan II and the cytoplasm-to-nuclear distribution determined by a previously described algorithm (18). Hoechst 33342 staining ( Fig. 1A and D) defined the nuclear area. Phospho-Erk fluorescence intensity within this nuclear area was referred to as "cytonuclear intensity". To assess the amount of fluorescently labeled phospho-Erk in the cytoplasm, a set of concentric rings spaced by two pixels was placed around the nuclear boundary. Phospho-Erk fluorescence intensity within the ring area was referred to as "cytoring intensity". Both cytonuclear and cytoring intensities were normalized to the total cytonuclear or cytoring area, and are expressed as average intensity per pixel. All cytoplasmic-to-nuclear difference values were calculated by subtracting the average cytoring intensity per pixel from the average cytonuclear intensity per pixel. Thus, an increase in the cytonuclear difference value is indicative of Erk activation through phosphorylation, translocation, or both.
Induction of phospho-Erk and phospho-p38, but not NFkB, by Compound 5. We next examined whether Compound 5 caused selective nuclear Erk accumulation by comparing its effects to those of other signaling events, which have also been reported to be activated in a tyrosine phosphorylation-dependent manner, and are thought to mediate stress responses. Cells were treated for 30 min with either 10 µM Compound 5 or 25 ng/ml interleukin-1 alpha (IL-1α), immunostained with phospho-Erk, phospho-p38, phospho-JNK, or p65 NFkB antibodies, respectively and analyzed for differences in cytoplasmic-to-nuclear fluorescence intensity. A total of 100 cells were imaged in each well. Figure 2 shows that Compound 5 lead to a dramatic increase in nuclear accumulation of phospho-Erk and phospho-p38, but had only a moderate effect on phospho-JNK, and did not affect the nuclear accumulation of NFkB. IL-1α, in contrast, activated all three stress-response mediators (p38, JNK, and NFkB), but not Erk. Thus, the activity profile of Compound 5 was distinct from that of the cytokine IL-1α, suggesting that Compound 5 was not a general stress-inducing agent.

Kinetics of Erk and p38 activation by Compound 5.
Experiments with the stress inducer and phosphatase inhibitor, sodium arsenite, previously demonstrated that p38 and Erk were activated with different kinetics in a variety of cell lines (21). These authors also reported that Erk activation was abrogated by dominant negative forms of p38 and the p38 specific kinase inhibitor, SB-203580, suggesting an involvement of p38 in Erk activation. We thus examined concentration-dependence and kinetics of phospho-Erk and phospho-p38 activation in Hep3B cells. Figure 3 shows that maximal stimulation of both Erk and p38 was obtained at 10 µM Compound 5. Moreover, continuous exposure to 10 µM Compound 5 caused a progressively greater activation and nuclear accumulation with similar temporal characteristics ( Figure 3B). We have also found that the p38 inhibitor SB-203580 did not inhibit phospho-Erk nuclear accumulation (data not shown). These results suggest that Compound 5 acted differently than the nonspecific tyrosine phosphatase inhibitor sodium arsenite.

Irreversibility of Compound 5 action. Compound 5 is a sulfhydryl-arylating agent and
its sustained antiphosphatase activity has been ascribed to covalent modification of critical cysteine residues on dual-specific and tyrosine phosphatases (17). To test whether its effects were irreversible, we treated cells with Compound 5 for 5 or 10 min, followed by washout, and compared the magnitude of phospho-Erk and phospho-p38 accumulation to that obtained after a 30 min continuous exposure. Figure 4 shows that short pulses of Compound 5 resulted in substantial activation of both Erk and p38, consistent with a rapid and persistent inhibition of cellular phosphatases after compound removal.
Biochemical analysis confirms phospho-Erk nuclear accumulation. We next validated the results from the automated fluorescence-based analysis by conventional biochemical methods. Cells were treated with 10 µM Compound 5 for the indicated times, lysed, separated into cytosolic and nuclear fractions, and analyzed by Western blot using a phospho-Erk antibody ( Figure 5A). Untreated cells had almost no nuclear phospho-Erk, consistent with the whole cell images in Figure 1B. Within minutes, Compound 5 caused a time-dependent and sustained increase in nuclear phospho-Erk accumulation. In contrast, cytosolic phospho-Erk levels in control cells were higher than those in the nucleus and increased only after a longer exposure to Compound 5 (30 min, Figure 5A). The results from the immunoblot analysis thus confirmed those from the less arduous solid phase cytometry studies.

Phosphorylated Erk from Compound 5 treated cells is activated and phosphorylates
Elk-1. We then used the identical lysates from Compound 5 treated cells to investigate whether the observed Erk phosphorylation resulted in an increase in Erk kinase activity.
It is thought that upon phosphorylation by MEK1 and MEK2 in the cytosol, a fraction of Erk translocates to the nucleus, where it phosphorylates and activates transcription factors, such as c-fos, c-jun, and Elk-1 (22). To investigate whether phosphorylated Erk was functional in Compound 5-treated cells, we examined its ability to phosphorylate the transcription factor, Elk-1. Phospho-Erk was immunoprecipitated from Compound 5 treated and untreated cells and immunoprecipitates were subjected to an in vitro kinase assay using recombinant GST-Elk-1 fusion protein as a substrate. Assay mixtures were separated on SDS-PAGE and immunoblotted with an anti-phospho-Elk-1 antibody. Figure 5B shows that nuclear phospho-Erk had kinase activity and that its kinetics of activation correlated well with its phosphorylation status. Compound 5-induced nuclear phospho-Erk was thus functional and able to phosphorylate its physiological substrate, Elk-1.
MEK inhibition and nuclear translocation of phospho-Erk, phospho-p38 or phospho-JNK. We next investigated possible consequences of Erk or p38 activation by Compound 5. We first examined whether inhibition of MEK, the direct upstream activating kinase for Erk, would reduce phospho-Erk nuclear accumulation. Cells were pretreated with the MEK1/MEK2 inhibitor U-0126 (23) for 45 min, stimulated with Compound 5 for an additional 30 min in the presence of the inhibitor, and analyzed on the ArrayScan II for nuclear accumulation of phospho-Erk, phospho-p38, and phospho-JNK. Consistent with results from Figure 2, Compound 5 caused a robust increase in nuclear phospho-Erk and phospho-p38, but had only a partial effect on phospho-JNK ( Figure 6). Inclusion of 10 µM U-0126 caused almost complete inhibition of Compound 5-induced Erk activation, but, as expected, had little or no effect on p38 or JNK activation. These data suggest that MEK inhibition is sufficient to inhibit phospho-Erk nuclear accumulation by Compound 5. quantified by fluorimetry as previously described (17). Figure 7 shows that inclusion of the MEK inhibitor significantly reduced Compound 5-mediated cell growth inhibition.

The MEK inhibitor U-0126 protects cells from the antiproliferative effects of
This strongly suggests that activation of the Erk pathway is the major determinant in the antiproliferative effects of Compound 5. In contrast, p38 activation, which has been implicated in cell death in many cell types, did not appear to mediate growth inhibition of Hep3B cells by Compound 5 since in the presence of U-0126, cell growth continued despite high levels of nuclear phospho-p38 but depressed levels of phospho-Erk (see Figure 6).

The effects of Compound 5 on nuclear phospho-Erk accumulation are cell type dependent.
To determine whether the observed accumulation of phospho-Erk was specific for Hep3B cells, we examined the ability of Compound 5 to induce nuclear phospho-Erk accumulation in a variety of mammalian cell lines using the ArrayScan II.
We found Compound 5-induced nuclear phospho-Erk accumulation was not unique to  which is an upstream activator of Erk, is controlled by Cdc25A (16). Thus, we hypothesized that ectopic expression of Cdc25A might reduce Erk phosphorylation and provide a novel assay system to examine the acute actions of Compound 5 against intracellular Cdc25A. We selected HeLa cells as a model because in the absence of ectopic Cdc25A no nuclear phospho-Erk accumulation was seen with Compound 5 in these cells, possibly due to low endogenous Cdc25A activity. We predicted that this model would, therefore, have the lowest background and that any effect seen with a small molecule could be assigned to an action on the ectopically expressed Cdc25A. As illustrated in Figure 8A, ectopic Cdc25A expression reduced Erk phosphorylation by 50% (p<0.05, Figure 8B). This reduction in Erk phosphorylation absolutely required the intrinsic phosphatase activity of Cdc25A because a catalytically inactive Cdc25A (C430S) did not reduce Erk phosphorylation in these cells. We then asked whether Compound 5 was able to restore Erk phosphorylation after ectopic expression of wildtype Cdc25A. Cells transiently transfected with wild-type Cdc25A were allowed to recover for 48 h and, during the last 30 min of recovery, treated with vehicle or increasing concentrations of Compound 5. Figure 8B shows that Compound 5 gradually restored Erk phosphorylation to mock/control levels. Consistent with the inherent

Discussion
PTPases and DSPases play a major role in receptor-mediated signal transduction events. In contrast to their upstream activating kinases, tyrosine / threonine phosphorylated MAPKs translocate to the nucleus where they phosphorylate and activate their respective protein targets, which include several transcription factors. Very few studies have addressed spatial aspects of phosphorylation-dependent signaling events and to our knowledge, none have investigated small molecules that might perturb the subcellular localization of key signal transducers in the context of tyrosine phosphorylation. Here we have used a high-content, cell-based assay to evaluate the temporal and spatial dynamics of three key parallel signaling molecules in response to Compound 5, a synthetic K vitamin analog with in vitro antiphosphatase activity (12) and antiproliferative activity in a variety of cell lines (10). Using this novel methodology, we found that Compound 5 caused rapid and irreversible nuclear accumulation of phospho-Erk and phospho-p38.
The observed activation of Erk by a compound known to cause growth inhibition (11)(12)17) is somewhat surprising since brief activation of Erk is often associated with mitogenesis and survival. In contrast, JNK and p38 are thought to be mediators of stress responses and apoptosis (24). We considered the possibility that p38 activation might be a factor in the antiproliferative activity of Compound 5. In the presence of the MEK1 inhibitor U-0126, however, which inhibits Erk activation, Hep3B cells grew despite having high levels of phospho-p38. In contrast, we found that pretreatment of cells with U-0126 not only prevented Compound 5-induced Erk phosphorylation, but also protected cells from the growth inhibitory effects of Compound 5. These results strongly support our previous suggestion that prolonged activation of the Erk pathway is causally involved in the growth inhibitory effects of Compound 5 (13,17) . Moreover, our conclusion is in agreement with a growing body of data documenting an involvement of Erk in growth inhibition in neuronal cells (25), NIH3T3 cells (26), and MCF-7 cells (14). In addition, there is increasing evidence that p38 does not appear to exclusively mediate cytotoxicity, but can be cytoprotective under certain conditions (27)(28)(29).
The inability of Compound 5 to induce NFkB and, to a lesser extent, JNK, suggests specificity and that it is not a general stress-inducing stimulus. Both NFkB and JNK are activated by a variety of extracellular stimuli, such as oxidative stress or inflammatory cytokines. In addition, the broad PTPase inhibitors vanadate and pervanadate have been found to induce NFkB (30,31), providing further support for a unique and more specific action associated with Compound 5. We recently demonstrated that Compound 5 selectively inhibited members of the DSPase family with median inhibitory values of 4 µM for Cdc25B 2 and Cdc25A, while it was 10-fold less active against VHR, a prototype MKP, and 100-fold less active against PTP1B (12). Furthermore, Compound 5 caused a cell cycle arrest in both G1 and G2, which correlated with enhanced phosphorylation of the Cdc25 substrates Cdk1, Cdk2, and Cdk4, respectively (12). We suggested that the growth inhibitory properties of Compound 5 might be due to inhibition of the Cdc25 family, but in large part due to a lack of appropriate assays, there has been no direct evidence that Compound 5 inhibits Cdc25 phosphatases in the cell. To investigate whether Cdc25A could affect Erk phosphorylation status and be inhibited within cells by Compound 5, we devised a chemical complementation strategy based on earlier observations that Cdc25A associated with the Raf-1 oncoprotein (15). Functional evidence that Cdc25A regulates Raf-1 activity was obtained by Xia et al. (16), who showed that overexpression of Raf-1 together with wild-type Cdc25A reduced PDGFmediated Raf-1 tyrosine phosphorylation in NIH 3T3 cells. Raf-1 is one of the most important upstream activators of the Erk cascade (32). We thus hypothesized that Cdc25A overexpression would result in decreased Erk phosphorylation and that an inhibitor of Cdc25A would restore Erk phosphorylation to normal levels, by chemically complementing the loss-of-function phenotype caused by Cdc25A overexpression. To simplify the analysis, we chose HeLa cells, which did not respond to Compound 5 with increased nuclear phospho-Erk accumulation. By treating Cdc25A-overexpressing cells with concentrations of Compound 5 that did not cause Erk hyperphosphorylation under normal growth conditions, we were able to demonstrate that Compound 5 specifically inhibited the effects of the overexpressed Cdc25A protein on Erk phosphorylation. Thus, we have obtained, for the first time, evidence that Cdc25A regulates endogenous Erk phosphorylation status in whole cells, and that Compound 5 affected Cdc25A function in the cell.
Although the concentrations of Compound 5 required for inhibition of the MKP VHR in vitro are an order of magnitude higher than those for Cdc25A inhibition, it is possible that inhibition of MKPs by Compound 5 also contributes to Erk and p38 activation. A number of cytosolic and nuclear MKPs, which have overlapping substrate specificities, have been described. For example, the Erk isoforms are selectively inhibited by MKP-3, whereas M3/6 selectively dephosphorylates JNK (33). MKP-1 and 2 preferentially dephosphorylate JNK, but also have some activity toward p38 (34,35). More recently, a p38 specific phosphatase, MKP-5, has been reported (36). The prototype DSPase VHR, which seems to reside in the nucleus, dephosphorylates Erk (37), but its effect on other kinases has not been examined. The fact that Compound 5 only partially activated JNK suggests that it may have some selectivity. At this time, we do not have any information about whether Compound 5 has any specificity for the different MKPs, but this information should become available as we expanding our chemical complementation strategy to probe for cell-active inhibitors of MKPs.
In summary, using the ArrayScan II, we were able to quickly and quantitatively probe selective activation of tyrosine phosphorylation-dependent signal transduction events by a small molecule dual-specificity phosphatase inhibitor in intact cells. By performing fluorescence-based spatial analysis in a high-throughput compatible format, we demonstrated that this inhibitor selectively activated dual-specificity phosphatasedependent cellular events. Subsequent analyses using both genetic and pharmacological tools identified activation of the Erk pathway as the dominant component mediating Compound 5's antiproliferative activity, and provided direct evidence that it could interfere with Cdc25A function in the cell. We propose that the combination of highcontent, cell-based analyses coupled with a chemical complementation approach will be a powerful technique to identify cell active inhibitors of a variety of cellular targets.   Hep3B cells (4,000) were plated in each of the 96 wells of a darkwell plate, treated with Compound 5 or vehicle, and stained with anti-phospho-Erk, phospho-p38, phospho-JNK, and p65NFkB antibodies. A minimum of 100 cells per well were analyzed with the previously described (18) nuclear to cytoplasm translocation algorithm on the ArrayScan II (Cellomics, Pittsburgh, PA). Cytoplasm-to-nuclear difference values were calculated as described in the legend to Figure 1 and normalized to the maximum signal obtained (10 µM Compound 5 for Erk and p38, and 25 ng/ml IL-1α for JNK and NFkB). Data shown are the averages from quadruplicate wells ± S.D. and are from a single experiment that has been repeated at least two times with identical results.