Activation of Protein Serine/Threonine Phosphatase PP2Cα Efficiently Prevents Liver Fibrosis

Background Over-activation of TGFβ signaling pathway and uncontrolled cell proliferation of hepatic stellate cells (HSCs) play pivotal roles in liver fibrogenesis, while the protein serine/threonine phosphatase PP2Cα was reported to negatively regulate TGFβ signaling pathway and cell cycle. Our study aimed to investigate the role of PP2Cα in liver fibrogenesis. Methodology/Principal Findings The effects of PP2Cα activation on liver fibrosis were investigated in human HSCs and primary rat HSCs in vitro using western blotting, real-time PCR, nuclear translocation, cell viability and cell cycle analyses. The antifibrogenic effects in carbon tetrachloride (CCl4)- and bile duct ligation (BDL)-induced mice in vivo were assessed using biochemical, histological and immunohistochemical analyses. The results demonstrated that activation of PP2Cα by overexpression or the new discovered small molecular activator NPLC0393 terminated TGFβ-Smad3 and TGFβ-p38 signaling pathways, induced cell cycle arrest in HSCs and decreased α-smooth muscle actin (α-SMA) expression, collagen deposition and hepatic hydroxyproline (HYP) level in CCl4- and BDL-induced mice. Conclusions/Significance Our findings suggested that PP2Cα activation might be an attractive new strategy for treating liver fibrosis while the small molecular activator NPLC0393 might represent a lead compound for antifibrogenic drug development. Moreover, our study might provide the first evidence for the role of PP2C family members in the fibrotic disease.


Introduction
Liver fibrosis is a major public health threat causing portal hypertension, liver failure, and risk of hepatocellular carcinoma. Hepatic stellate cells (HSCs) play critical roles in liver fibrogenesis. Once intoxicated by stimuli, quiescent HSCs could transdifferentiate into activated HSCs which secrete some proinflammatory and profibrogenic cytokines such as tumor necrosis factor alpha (TNFa) and transforming growth factor beta (TGFb), leading to over-accumulation of extracellular matrix (ECM) and altered matrix degradation. Meanwhile, these cytokines further activate HSCs and enhance their proliferation and survival, thus exacerbating fibrogenesis [1]. Recently, emerging strategies against liver fibrosis have been proposed, such as selective antagonization of CB1 cannabinoid receptor [2], targeting 5-hydroxytryptamine (5-HT) class of receptors [3], inhibition of Toll-like receptor 4 (TLR4) [4], and activation of STAT1 [5], etc. However, the efficient strategies are still lacking due to the complicated pathogenesis associated with this disease [6].
Protein serine/threonine phosphatases (PS/TPs) dephosphorylate phosphoserine/phosphothreonine-containing proteins and comprise three structurally distinct families: phosphoprotein phosphatases (PPPs), metal-dependent protein phosphatases (PPMs), and the aspartate-based phosphatases represented by FCP/SCP (TFIIF-associating component of RNA polymerase II CTD phosphatase/small CTD phosphatase). Protein phosphatase 2C, which belongs to PPM family, is a structurally and functionally distinct group of enzymes that currently contain about 22 different family members. The members of this family are distinguished by their monomeric property and dependency on Mg 2+ and Mn 2+ [7]. It should be noted that except the oncoprotein PP2Cd (also known as Wip1) [8,9], all the other members from this family have been identified as tumor suppressors based on their inhibition of cell growth and cellular stress signaling [10,11].
Protein phosphatase 2C alpha (PP2Ca; EC 3.1.3.16), the most extensively characterized member of PP2C family, plays an important role in TGFb, cell growth, stress and inflammation signaling [10,[12][13][14][15]. PP2Ca was reported to dephosphorylate Smad2/Smad3 to block TGFb signaling pathway [12], activate p53 and dephosphorylate Cdk2/Cdk6 to induce cell cycle arrest [13,14], inhibit p38 and JNK signaling pathways to prevent stress [16] and dephosphorylate IKappa B kinase b (IKKb) to prevent inflammatory response [15]. Recently, the potential role of PP2Ca in tumorigenesis has been revealed [11], whereas its function in the fibrotic disease still remains unknown. The current study therefore aimed to investigate the role of PP2Ca in liver fibrosis by assessing the effects on TGFb signaling pathway and cell cycle of HSCs and ECM expression in mouse models. Our findings suggest that PP2Ca activation might be a promising new strategy for the treatment of liver fibrosis.

Results
Activation of PP2Ca inhibited TGFb-Smad3 and TGFb-p38 signaling pathways in HSCs Since Smad3 was regarded as the main mediator of TGFbinduced fibrotic response [12,17], we first assessed the impact of PP2Ca on TGFb-induced Smad3 phosphorylation in human hepatic stellate cell line LX-2 cells. As shown in Figure 1A, TGFb stimulated Smad3 phosphorylation, while the stimulation was obviously decreased after PP2Ca overexpression and slightly enhanced with PP2Ca knock-down by shPP2Ca494. Similarly, the TGFb-induced Smad2 phosphorylation was reduced with PP2Ca overexpression and mildly increased with PP2Ca knockdown. Considering that p38 was also reported to mediate TGFbinduced fibrotic effects [16,18], we examined the effect of PP2Ca on TGFb-induced p38 phosphorylation. The result revealed that TGFb stimulated the phosphorylation of p38 and this stimulation could be regulated by PP2Ca overexpression or knock-down ( Figure 1A).
We next studied whether PP2Ca affected TGFb-induced collagen transcription that was reported to be up-regulated by Smad3 and p38 phosphorylations [17,18]. Consistently, the results indicated that TGFb increased a1(I) procollagen mRNA transcription, whereas PP2Ca overexpression aborted the stimulatory effect of TGFb while PP2Ca knock down enhanced it ( Figure 1B). These findings demonstrated that overexpression of PP2Ca suppressed TGFb-Smad3 and TGFb-p38 signaling pathways in HSCs.

Activation of PP2Ca induced cell cycle arrest in HSCs
The regulation of PP2Ca on cell cycle in several cell lines was reported previously [11,13,14]. Consistent with these reports, our work demonstrated that overexpression of PP2Ca inhibited LX-2 cell viability in a dose-dependent manner ( Figure 1C).
Cdk2, an important regulatory protein of G1-S transition, was reported to mediate PP2Ca induced cell cycle arrest [13]. Therefore, to verify whether the cell viability loss was due to cell cycle arrest induced by PP2Ca, we examined Cdk2 phosphorylation. The result ( Figure 1D) indicated that overexpression of PP2Ca attenuated Cdk2 phosphorylation, while knock down of PP2Ca enhanced it. These results suggested that PP2Ca induced cell cycle arrest in LX-2 cells through down-regulating Cdk2 phosphorylation.

Identification of NPLC0393 as a small molecular PP2Ca activator
To further verify the therapeutic potential of PP2Ca, we identified a small molecular PP2Ca activator, NPLC0393, through a reconstituted in vitro PP2Ca phosphatase assay ( Figure 2A) [19]. The result revealed that NPLC0393 dosedependently increased PP2Ca activity with an EC 50 value of 6.72 mM using pNPP as substrate ( Figure 2B). Additionally, we further confirmed the enhancement of PP2Ca activity by NPLC0393 using the phosphopeptide substrate FLRTpSCG, which is derived from AMP-activated protein kinase and was previously reported to be a good substrate for PP2Ca [14]. The result indicated that NPLC0393 also increased PP2Ca activity in a dose-dependent manner, with an EC 50 value of 6.43 mM ( Figure 2C).
Subsequently, we confirmed the direct binding of NPLC0393 to PP2Ca through surface plasmon resonance (SPR) technology based assay. The dissociation equilibrium constant (K D ) was thus determined as 19.2 mM ( Figure 2D). In addition, the isothermal titration calorimetry (ITC) was also applied to analyze the stoichiometry and thermodynamics of NPLC0393/PP2Ca interaction by titrating NPLC0393 to PP2Ca ( Figure 2E). The results revealed that the stoichiometric ratio was 1.0560.03, implying that a single molecule of NPLC0393 could interact with one molecule of PP2Ca. Furthermore, the determined K D was approximately 14.7 mM, similar to the SPR result. Notably, the change in Gibbs' free energy (DG) resulting from NPLC0393/ PP2Ca interaction was driven primarily by a favorable entropy (TDS, 5.93 kcal/mol), compared with the enthalpy (DH, 20.669 kcal/mol), suggesting that NPLC0393/PP2Ca binding was mainly mediated by the increase of the buried surface area rather than the polar interactions ( Figure 2E).
To assess the targeting specificity of NPLC0393, we evaluated the effects of NPLC0393 on two representative mammalian Ser/ Thr phosphatases (PP1 and PP2A) and one typical Tyrosine phosphatase (PTP1B). The results in Table 1 thereby indicated that NPLC0393 had no obvious activities against these three tested phosphatases, further suggesting its good specificity against PP2Ca.
Collectively, our results demonstrated that NPLC0393 as a specific small molecular activator of PP2Ca might be used as a potential probe to elucidate the biological significance of PP2Ca in relevant diseases.

NPLC0393 inhibited TGFb-Smad3 and TGFb-p38 signaling pathways in HSCs
The effects of NPLC0393 on TGFb-Smads and TGFb-p38 signaling pathways were assessed in LX-2 cells and primary rat hepatic stellate cells (HSCs). The results indicated that NPLC0393 decreased Smad3 phosphorylation in both time-and dosedependent manners ( Figure 3A), and the TGFb-induced Smad3 and p38 phosphorylations were also reduced by NPLC0393 treatment ( Figure 3B). Moreover, NPLC0393 inhibited Smad3 nuclear localization ( Figure 3C), which was reported to depend on its phosphorylation [12]. Additionally, it should be pointed out that NPLC0393 rendered no evident influence on basal or TGFb-induced Smad2 phosphorylation ( Figure 3A,B). Finally, NPLC0393 decreased basal and TGFb-induced a1(I) procollagen mRNA expression ( Figure 4A,B). Furthermore, NPLC0393 failed to exert the above effects in PP2Ca stable knock-down cells (shPP2Ca cells) ( Figure 4C,D), thus confirming that these effects of NPLC0393 were mediated by PP2Ca. Altogether, these findings indicated that treatment of NPLC0393 could block TGFb-Smad3 and TGFb-p38 signaling pathways through inhibiting Smad3 and p38 phosphorylations and Smad3 nuclear localization.

NPLC0393 induced cell cycle arrest in HSCs
The impact of NPLC0393 on cell cycle was also examined in LX-2 and primary rat HSCs cells. As indicated in Figure 5A and B, NPLC0393 decreased HSCs cell viability in a time-and dosedependent manner as assessed by MTT assay. To figure out whether such cell viability reduction was due to cell cycle arrest, we next carried out flow cytometry analysis. The results demonstrated that 48h incubation of NPLC0393 dose-dependently induced G1 phase arrest in LX-2 cells ( Figure 5C,D). Analysis of cell cycle regulatory proteins revealed that NPLC0393 decreased phosphorylation of Cdk2 in LX-2 cells ( Figure 5E, left). Considering that Platelet-Derived Growth Factor (PDGF) could stimulate cell proliferation by increasing Cdk2 phosphorylation [20], we also examined the effect of NPLC0393 on PDGF-induced p-Cdk2 level in LX-2 cells. The result displayed that NPLC0393 obviously inhibited PDGF-induced Cdk2 phosphorylation in a dose-dependent manner ( Figure 5E, right). Moreover, the effects of NPLC0393 on cell cycle were subsequently studied in shPP2Ca cells. The results showed that NPLC0393 failed to decrease cell viability ( Figure 5F) and Cdk2 phosphorylation ( Figure 5G) in shPP2Ca cells compared with control cells. Therefore, these findings indicated that activation of PP2Ca by NPLC0393 induced cell cycle arrest in HSCs.

NPLC0393 attenuated liver fibrogenesis in vivo
To further investigate the anti-liver fibrosis potential of PP2Ca, two different mouse models bearing liver fibrosis were treated with the PP2Ca activator NPLC0393. Compared with the vehicle group, treatment of NPLC0393 (2.5 mg/kg) rendered no obvious influence on the serum alanine transaminase (ALT), aspartate transaminase (AST) levels or the liver histology, implying that NPLC0393 was little toxic in vivo (data not shown). As shown in Figure 6A and B, 2.5 mg/kg of NPLC0393 administration decreased a-SMA expression in both CCl 4 and BDL-induced liver fibrosis mice. In addition, Masson staining of collagen indicated that NPLC0393 reduced the fibrosis area in both models ( Figure 7A). The CCl 4 and BDL-induced a1(I) procollagen mRNA levels were also decreased in the NPLC0393-treated mice ( Figure 7B). Moreover, NPLC0393 administration declined the ECM marker, hydroxyproline (HYP) content in the two kinds of liver fibrosis mice ( Figure 7C). It should be also noted that NPLC0393 decreased the ALT and AST levels in both CCl 4 and BDL-induced liver fibrosis mice, suggestive of its protective function in liver injury (data not shown). Taken together, all these results thus suggested that NPLC0393 as a PP2Ca activator could significantly attenuate liver fibrogenesis in both CCl 4 -and BDLinduced liver fibrosis mice.
Here, we demonstrated that PP2Ca activation could terminate TGFb signaling pathway and simultaneously induce cell cycle arrest in HSCs, leading to significant anti-fibrogenic effects both in vitro and in vivo, although we could not exclude the possibility that the anti-fibrotic effects of PP2Ca activation might be also mediated by reduction of stress and inflammatory response, which is however beyond our current study.
The crucial role of TGFb signaling in liver fibrogenesis has been widely recognized [1,23]. Several anti-TGFb signaling pathwaytargeted strategies were recently proved effective, such as inhibition of latent TGFb activation or prevention of TGFb binding to its receptor [24]. These strategies, however, mainly involved large molecular inhibitors (e.g. monoclonal antibodies and antisense oligonucleotides) against TGFb receptor which might block the systemic immunosuppressive effects of TGFb [24,25]. The current anti-fibrogenic reports concerning small molecular inhibitors of TGFb signaling are only restricted to the inhibitors of TGFb type I receptor kinase [26][27][28]. In our work, we determined that the natural product NPLC0393 as a specific small molecular PP2Ca activator could efficiently alleviate liver fibrosis. Therefore, our work is expected to provide new insights into the understanding of TGFb signaling inhibition-based antiliver fibrogenesis research, while the discovered small molecular PP2Ca activator NPLC0393 might be used as a potential lead compound for anti-liver fibrotic drug discovery.
Interestingly, although Smad2 and Smad3 were both shown to be dephosphorylated by PP2Ca [12], our study revealed that NPLC0393 only selectively dephosphorylated Smad3 without altering Smad2 phosphorylation. Based on the different roles of Smad3 and Smad2 in TGFb signaling [29,30] and the fact that Smad3, but not Smad2, mediates the liver fibrosis response [17], we thereby propose that NPLC0393 might supply a promising interest in the treatment of liver fibrosis with high specificity, although the detailed mechanism of such specificity needs to be further investigated. Additionally, consistent with the previous report [16], we uncovered that PP2Ca overexpression or NPLC0393 treatment not only decreased the TGFb-induced Smad3 phosphorylation but also reduced the TGFb-induced p38 phosphorylation. Therefore, we assume that the decreased a1(I) collagen expression induced by PP2Ca and NPLC0393 might result from the inhibition of both TGFb-Smad3 and TGFb-p38 signaling pathways. Although TGFb1 transcription was reported to be Smad3-dependent [17], the undetectable decrease of TGFb1 mRNA expression in NPLC0393 treated liver fibrosis mice might be due to the other signaling pathways besides TGFb-Smad3, which are also involved in TGFb1 expression.
As indicated, apart from blocking TGFb signaling, reducing HSCs was also proved effective in preventing liver fibrogenesis [31,32]. Here, we determined that PP2Ca activation induced cell cycle arrest of HSCs through decreasing P-Cdk2, thus leading to the evident antifibrotic effects as evaluated in CCl 4 -and BDLinduced mouse models. By considering the well characterized antiproliferative effects of PP2C family members [22], we thus suggested that our findings might gain insights into their potential roles in the treatment of fibrotic diseases that are always associated with excessive proliferation of activated stellate cells.
To confirm the function of PP2Ca activation on liver fibrogenesis in mice, we carried out two mice models. One is toxic fibrosis model induced by CCl 4 and the other is biliary fibrosis model induced by BDL. These two models are mediated   by different mechanisms. The CCl 4 -induced liver fibrosis begins with inflammatory response which activates HSCs leading to the eventual accumulation of ECM, while the production of ECM in the BDL-induced model is not from inflammatory response which is not so evident in these mice [33,34]. Notably, our current study has revealed that activation of PP2Ca reduced a-SMA expression, collagen deposition and HYP level in both models, further suggesting that PP2Ca activation exhibited efficient antifibrogenic effects.
To date, quite few compounds targeting PP2Ca have ever been reported although the relevant catalytic mechanism and crystal structure regarding this phosphatase have been elucidated [19,35,36]. Considering the potent biological functions of PP2Ca, we randomly screened our in-house natural product library (,10,000 compounds) against the recombinant human PP2Ca phosphatase for identifying small molecular PP2Ca regulators (inhibitor or activator). The natural product NPLC0393 was thus determined as a specific PP2Ca activator. It should be also pointed out that the mRNA and protein levels of PP2Ca in HSCs were not affected by NPLC0393 (data not shown), further suggesting that NPLC0393 implemented its antifibrotic effects through enhancing PP2Ca enzymatic activity. NPLC0393 is a triterpene saponin extracted from Gynostemma pentaphyllum, which is widely used in the treatment of liver disease [37][38][39][40]. Our findings are thus expected to bring new insights into the potential pharmacological mechanism for this popular traditional herbal medicine, while NPLC0393 might represent a lead compound for antifibrogenic drug development.
Conclusively, our work has indicated that PP2Ca activation not only terminated TGFb-Smad3 and TGFb-p38 signaling pathways but also inhibited cell proliferation in hepatic stellate cells. The fact that PP2Ca activation by NPLC0393 remarkably prevented liver fibrogenesis in CCl 4 -and BDL-induced mice, has further confirmed that PP2Ca activation could be a promising strategy for treating liver fibrosis.

Animals
C57/BL6 male mice at 8-week age were obtained from Shanghai SLAC Laboratory Animal Co. Ltd. The CCl 4 -induced liver fibrosis was generated by intraperitoneal injection of CCl 4 (0.5 ml/kg, diluted 1:10 in olive oil) twice weekly, alternating with an isovolumetric dose of 5% ethanol diluted in PBS 5 times per week [2]. NPLC0393 was dissolved in Tween-80 and intraperitoneal injected daily. Groups were as follows (n = 9): mice given olive oil and NPLC0393 (control); mice given CCl 4 , ethanol and Tween-80 (model); mice given CCl 4 , ethanol and treated with 2.5 mg/kg of NPLC0393 (NPLC0393). After 4 weeks, animals were starved overnight and executed 48 h after the last CCl 4 injection. The BDL-induced liver fibrosis was constructed by transecting the common bile duct between two ligations after midline laparotomy as described [2]. Groups were as follows (n = 9): mice receiving sham operation and Tween-80 (control); mice receiving BDL and Tween-80 (model); mice receiving BDL and treated with NPLC0393 (NPLC0393). Mice were sacrificed after 2 weeks. Liver samples were either fixed in buffered formalin or snap frozen in liquid nitrogen and stored at 280uC until use.

Histological and immunohistochemical analysis
Livers were fixed in 4% paraformaldehyde, embedded in paraffin and sectioned. Immunohistochemical staining of a-SMA was performed to quantify activated HSCs. Masson staining for collagen was used to quantify fibrosis area. The results were analyzed by Image-Pro Plus software (MediaCybernetics, France). Images of five fields were taken for each section with 9 mice in each group.   HSCs. (A,B) LX-2 cells and the isolated primary rat HSCs were treated with increasing concentrations of NPLC0393 for the indicated time points. MTT assay was performed to assess the effects of NPLC0393 on cell viability. The values were indicated as relative units normalized to the control. *P,0.05; **P,0.01; ***P # 0.001 compared with control group at the indicated time point. (C, D) LX-2 cells were exposed to increasing concentrations of NPLC0393 for 48 h. Then cells were harvested and the cell-cycle distribution was analyzed by Flow cytometry analysis. (E) LX-2 cells were treated as described in Figure 3C. Effect of NPLC0393 on Cdk2 phosphorylation was assessed by western blotting. For the PDGF-induced Cdk2 phosphorylation, LX-2 cells were cultured to confluence and growth-arrested for 24 h in DMEM with 10% FBS, and then for an additional 24 h treatment with NPLC0393 and PDGF (10 ng/ml) in DMEM plus with 0.2% FBS. (F) Control and shPP2Ca cells were treated with increasing concentrations of NPLC0393 for 24 h. Effects of NPLC0393 on cell viability in shPP2Ca cells and control cells were assessed by MTT assay. Significant difference of the reduction on cell viability by NPLC0393 in shPP2Ca cells versus that in control cells at indicated dose, *P,0.05, **P,0.01. (G) Cells were treated as described in Figure 3F and harvested for Western blotting. doi:10.1371/journal.pone.0014230.g005 Hepatic hydroxyproline determination Hepatic hydroxyproline content was measured using hydroxyproline detection kit (Jiancheng Institute of Biotechnology, Nanjing, China) according to the manufacturer's instruction. The results (mg/mg liver) were calculated according to the standard curve of hydroxyproline.
Primary HSCs isolation, cell lines, culture and treatment Primary HSCs were isolated from normal rat liver (male Sprague-Dawley rats, 400-450 g) as described [41]. Cells were cultured in Dulbecco's minimum essential medium (DMEM; GIBCO/Invitrogen) containing 10% fetal bovine serum (FBS; GIBCO/Invitrogen). All the experiments were performed using 6- day culture-activated HSCs whose activation was verified by a-SMA expression using western blotting. Human hepatic stellate cell line LX-2 [42] and HEK293T Phoenix-ampho retrovirus packaging cells (ATCC) were cultured in DMEM supplemented with 10% FBS in 5% CO2 at 37uC. TGFb and PDGF were from Sigma.
Establishment of stable LX-2 cell line expressing shPP2Ca pSRG vector and pSRG-shPP2Ca494 construct were transfected into 293T Phoenix-ampho retrovirus packaging cells. After 48 h, viral supernatant was collected, filtered, and supplemented with polybrene (8 mg/ml). LX-2 cells were infected with viral supernatant. At 48h post-infection, infected cells were selected with puromycin (3 mg/ml). After selection for 5 days, cells were collected and verified by western blotting [12].

Cell viability assay
Cell viability was evaluated using MTT (Sigma) assay as previously described [44].

Cell cycle analysis
Cell cycle was analyzed as previously described [44]. The samples were assayed with a FACS Calibur instrument and the data were analyzed with CellQuest 3.1 Software.

Nuclear translocation
Nuclear translocation was assessed by immunofluorescence experiment as described [25]. The images were taken by IN Cell Analyzer 1000 and the data were analyzed with Nuclear Translocation analysis module [45].
The PCR cycle was 95uC for 5 seconds, 58uC for 20 seconds and 72uC for 20 seconds.

Identification of human PP2Ca activation by NPLC0393
NPLC0393 was isolated and purified as previously described [47]. Human PP2Ca was expressed in E. coli with C-terminal 6-His tag, and batch-purified using Ni-NTA resin according to the manufacturer's instruction (Qiagen). The assay was carried out in a reaction buffer containing 50 mM Tris-HCl, pH 7.0, 10 mM MnCl 2 [19]. NPLC0393 was dissolved in DMSO as a stock solution and diluted in reaction buffer to the final concentration. PP2Ca was diluted in reaction buffer as appropriate to 10 mg/ml, and reactions were started by the addition of 4 mM pNPP (Sigma), incubated with varying concentrations of NPLC0393 for 2 h at room temperature, and stopped with a solution containing 1 N NaOH. The effect of NPLC0393 on PP2Ca dephosphorylation of pNPP was determined by monitoring the absorbance change recorded at 410 nm, with 1% DMSO as a control.
With phosphopeptide FLRTpSCG (HD Biosciences; China) as the substrate [14], the reaction buffer containing 50 mM Tris-HCl, pH 7.0, 30 mM MgCl 2 was used. PP2Ca was diluted in reaction buffer as appropriate to 10 mg/ml and incubated with varying concentrations of NPLC0393 for 2 h at room temperature. Then the reaction was started with 500 mM FLRTpSCG for 30 min and terminated by adding 100 ml of malachite green/ ammonium molybdate reagent (upstate). Color development was allowed to proceed for 15 minutes at room temperature. Measurements were taken at 630 nm using microplate spectrophotometer (Bio-Rad). The effect of NPLC0393 on PP2Ca dephosphorylation of FLRTpSCG was determined by monitoring the absorbance change recorded at 630 nm, with 1% DMSO as control.
For selectivity assay, PP1 and PP2A were bought from Upstate. PTP1B was purified using Ni-NTA resin according to the manufacturer's instruction (Qiagen). The effects of NPLC0393 on these phosphatases dephosphorylation of pNPP were determined by monitoring the absorbance change recorded at 410 nm, with 1% DMSO as a control.

Surface plasmon resonance (SPR) technology based assay
The binding affinity of NPLC0393 to PP2Ca was evaluated by using a Biacore 3000 instrument (Biacore AB, Uppsala, Sweden). Immobilization of the purified PP2Ca to the hydrophilic carboxymethylated dextran matrix of the sensor chip CM5 (Biacore) was performed by the standard primary amine coupling reaction. PP2Ca (8.28 mg/mL in 10 mM sodium acetate, pH 4.2) was then injected over the surface until a desired immobilization level (6000 RU) was reached. Binding affinity measurements were carried out in a continuous flow of 20 ml/min HBS (10 mM HEPES, 150 mM NaCl and 0.005% (v/v) surfactant P20, pH 7.4) as the running buffer. NPLC0393 was diluted in the running buffer and automatically injected in a series of increasing concentrations. The binding responses were recorded continuously in resonance units (RU) at a frequency of 1 Hz as sensorgrams and presented as a function of time. Sensorgrams were processed by using automatic correction for nonspecific bulk refractive index effects. The dissociation equilibrium constant (K D ) was estimated by the 1:1 Langmuir binding fit model encoded in the Biacore analysis software.

Isothermal Titration Calorimetry (ITC) technology based assay
Binding of NPLC0393 to PP2Ca was also measured at 25uC by using a VP-ITC calorimeter (MicroCal, Northampton, MA). All the samples were dialyzed against ITC buffer (50 mM Tris-HCl, pH 7.0 and 1 mM Mn 2+ ) and degassed prior to titration. 1.43 ml of 50 mM PP2Ca was titrated by 300 ml of 500 mM PP2Ca using 30 injections. The heat of dilution of NPLC0393 was measured by titrating NPLC0393 into the ITC buffer and was subtracted for data analysis. Data were analyzed with Origin 7.0 software (MicroCal) using a single-site binding model.

Statistical analysis
All the experiments were repeated at least three times. Data were presented as mean 6 SD. Statistical analysis was performed using one-way ANOVA followed by Bonferroni's multiple comparison tests. p value of less than 0.05 was considered statistically significant.