Cloning and characterization of a novel mammalian PP2C isozyme.

PP2C is a structurally diversified protein phosphatase family with a wide range of functions in cellular signal transduction. A novel PP2C subtype, designated PP2Cdelta, was identified from a rat cDNA clone, which encodes a protein of 392 amino acid residues. While PP2Cdelta shares approximately 30% sequence identity in its catalytic domain with the mammalian PP2C, it lacks a 90-residue carboxyl-terminal sequence conserved in mammalian PP2C. Northern blot analysis showed that PP2Cdelta is widely expressed in rat tissues. The transcription of the PP2Cdelta gene was activated in response to stress, such as the addition of ethanol to the culture medium or UV irradiation of cells. Recombinant PP2Cdelta purified from bacteria exhibited a potent Mn2+-dependent serine/threonine phosphatase activity. Unlike other members of the PP2C family, the activity of PP2Cdelta was inhibited, rather than stimulated, by Mg2+. Transfection with PP2Cdelta resulted in inhibition of cell growth, precluding generation of stable 293 or CHO transfectants. Using a modified tetracycline-regulated PP2Cdelta-GFP dicistronic expression cassette, it was revealed that overexpression of PP2Cdelta blocked cell cycle progression and arrested cells at early S phase, resulting in inhibition of DNA synthesis and leading to cell death. These results suggest that PP2Cdelta plays a role in regulation of cell cycle progression via dephosphorylation of its substrates whose appropriate phosphorylation states might be crucial for cell proliferation.

The reversible phosphorylation of proteins is critical in the regulation of cellular signal transduction pathways. The state of protein phosphorylation in vivo is controlled by the relative activity of two families of enzymes with opposing actions, namely, protein kinases and protein phosphatases (1,2). The protein phosphatases that selectively dephosphorylate either tyrosine or serine-threonine residues are structurally and functionally diverse enzymes that can be divided into three distinct families: PTP, PPP, and PPM (1,3). The PTP family, including both receptor-like and cytosolic phosphatases, specifically dephosphorylates tyrosine residues of proteins. A subfamily of PTPs, dual specificity phosphatases, is capable of dephosphorylating both tyrosine and serine-threonine residues. The PTPs are characterized by the presence of a conserved catalytic domain of approximately 200 -300 residues with the unique active-site sequence motif HCXXGXXRS(T) (4 -6). While the catalytic domains of receptor-like and cytosolic PTPs share greater than 30% sequence identity, the sequence homology between the dual specificity phosphatases and other subfamilies of PTPs is less than 5%. However, the overall structural fold in the x-ray structures resolved from these different subfamilies of PTPs displays a striking similarity (4,5), indicating that a common architecture is required for dephosphorylation of tyrosine. The PPP family comprises three subtypes of phosphatases including PP1, PP2A, and PP2B (3), which specifically dephosphorylate serine/threonine residues (1,7). Moreover, the dephosphorylation activity of the PPP family can be inhibited by either protein inhibitors (inhibitors 1 and 2) or the tumor promoter okadaic acid (1,8). The PPP family shows extreme evolutionary sequence conservation of up to 43% homology in their catalytic domain, implying that the entire primary structure of this family is important for their physiological functions (9 -12). The PPM protein phosphatases belong to a large and diversified family of protein phosphatases expressed in both eukaryotes and prokaryotes and function to dephosphorylate serine and threonine residues as does the PPP family. PP2C relates to the main enzyme subtype of PPM, including mammalian PP2C␣, PP2C␤, and PP2C␥ (13)(14)(15)(16)(17), Arabidopsis ABI1 (18), Arabidopsis KAPP-1 (19), pyruvate dehydrogenase phosphatase (20), and Bacillus subtilis spoIIE phosphatase (21). The dephosphorylation activity of PP2C absolutely requires metal cations, Mn 2ϩ or Mg 2ϩ , but its activity is not sensitive to the tumor promoter okadaic acid or other inhibitors of the PPP family (1). Molecular cloning has defined PP2Cs as distinct protein phosphatases, since they constitute a separate gene family and are monomeric enzymes (1,17,(22)(23)(24)(25)(26)(27). The substantial sequence disparities of PP2C enzymes from various organisms may indicate a functional diversity of this phosphatase subfamily. Indeed, the list of potential physiological functions proposed for PP2C spans a wide range (1,17,27,28). It is now well established that certain members of PP2C play a role in reversing protein kinase cascades activated by stress (29 -32). Recent research further revealed that PP2C may be involved in cell apoptosis (33), gene expression, and other cellular functions (34 -36). In addition, it was recently reported that a growth factor-inducible PP2C, FIN13, can negatively regulate cell cycle progress (37). Since the processes of both cellular signal transduction and stress responses are very complicated, many unknown protein phosphatases are expected to be involved in these processes. Defining new phosphatases involved in these processes should facilitate our understanding of the related intracellular events. In the present report, we describe the identification of a novel protein phosphatase that, when overexpressed, can inhibit cell cycle progression. Since this novel protein phosphatase shares an appreciable sequence homology with the PP2C subfamily in the catalytic domain but is distinguished from other members in the sequence of its noncatalytic domain and in its reversed inhibitory sensitivity to Mg 2ϩ concentration, it is therefore designated as PP2C␦ (1).

EXPERIMENTAL PROCEDURES
Materials-Dulbecco's modified Eagle's medium, Ham's F-12 medium, fetal bovine serum, doxycycline, penicillin, and streptomycin were from Life Technologies, Inc. The human embryonic kidney cell line 293 was obtained from American Type Culture Collection. The reverse tetracycline-controlled transactivator and pTR5-DC-GFP plasmid DNAs were kindly provided by Dr. Massie (Biotechnology Research Institute, National Research Council of Canada). The Gene Amp PCR system type 9600, model 377 DNA sequencer, and DNA sequencing kit were obtained from Perkin-Elmer. The Geneclean II kit was purchased from Bio 101. PCR TM II vector was obtained from Invitrogen. Rat brain stem-spinal Lambda ZAP II cDNA library was purchased from Stratagene. Oligo(dT)-cellulose, the first strand cDNA synthesis kit, and Ready-To-Go DNA labeling kit were products of Amersham Pharmacia Biotech. [  cDNA Synthesis and PCR Screening-Total rat hypothalamic RNA was extracted from male Sprague-Dawley CD rats and purified as described previously (38). The cDNA for PCR was synthesized by using a cDNA synthesis kit following the manufacturer's instructions. Degenerate primers were designed from known PP2C sequences. PCR was carried out as follows: cDNA denaturation at 95°C for 10 min and then at 94, 40, and 72°C for 30 s each for 25 cycles. The amplified PCR products were separated on an 1% agarose gel, and DNA bands between 500 and 800 bp were excised, purified, and subcloned into the pCR TM II vector. The plasmid DNA was transformed into DH5␣ cells, and white colonies were screened for inserts by PCR amplification using the T7 and SP6 primers. The individual PCR products were characterized by restriction analysis using HaeIII. Representative PCR products were analyzed by DNA sequencing.
Screening of the Novel Protein Phosphatase Gene-A clone, number 6, containing 550 bp was found to share high homology with the PP2C family. A probe, based on this clone, was labeled by using a Ready-To-Go kit and used to screen a rat brain stem-spinal cord Lambda ZAPII cDNA library as described before (39).
Northern Blot Analysis-A 32 P-labeled probe representing the entire coding sequence of the PP2C␦ cDNA was used for Northern blot analysis using a rat multiple tissue membrane in which each lane contains approximately 2 g of poly(A) ϩ RNA from the tissues.
Construction of Expression Vector and Purification of Recombinant PP2C␦-The coding sequence of PP2C␦ was amplified by PCR from the original PP2C␦ clone (number 25) using a 5Ј-primer containing a BamHI restriction site (AATGGATCCATGGACCTATTCGGGGACT) and a 3Ј T3 primer. The PCR product was digested by BamHI and HindIII and cloned into a modified pET-3c (40) expression vector in which a sequence encoding for 6 histidines was immediately followed by the PP2C␦ coding sequence. The construct was confirmed by DNA sequencing.
The expressed PP2C␦ fusion protein was purified by using a metal chelate affinity chromatography (MCAC) as described before (41,42). Briefly, the lysate containing PP2C␦ protein was incubated with 0.3 ml of Ni 2ϩ -NTA agarose and eluted by MCAC-40 buffer after washing in the buffers of MCAC-0 and MCAC-10, respectively. The PP2C␦-containing fractions were identified on 9% SDS-PAGE after staining with Coomassie Blue.
Preparation of 32 P-Labeled Myelin Basic Protein (MyBP)-The phosphorylation of MyBP was performed by using a protein serine/threonine phosphatase assay system kit. After labeling with [␥-32 P]ATP, the final concentration of labeled MyBP was calibrated to 0.3 g/l.
Production of Anti-PP2C␦ Serum from Rabbits-Anti-PP2C␦ antibodies are generated by injection of the purified PP2C␦ into rabbits following a protocol as described before (43).
Protein Phosphatase Activity Assay-The dephosphorylation activity assay was performed in 30 l of assay buffer containing 1 g of purified PP2C␦ protein and 1.2 g of 32 P-labeled MyBP in the presence of 1 mM of Mg 2ϩ , Mn 2ϩ , or Ca 2ϩ . After a 10-min incubation at 30°C, the reaction was terminated by addition of 200 l of 20% trichloroacetic acid. The supernatant was used to determine released 32 P. The effect of Mn 2ϩ on phosphatase activity was performed in the same assay buffer but with a gradient concentration ranging from 0 to 20 mM. Phosphatase activity was displayed as pmol/min calculated from labeled substrate MyBP.
Metal Cation Competition Assay-The Mg 2ϩ competition assay was carried out by increasing the concentration of Mg 2ϩ from 0 to 16 mM in the assay buffer containing 1 g of PP2C␦ protein, 1.2 g of [ 32 P]MyBP, and 1 mM of Mn 2ϩ .
Expression of PP2C␦ Gene in Response to Stress in CHO Cells-CHO cells were maintained at 90% confluence in Ham's F-12 medium supplied with 10% fetal bovine serum. The stress environments of cells were produced by treatment for 3 min with either ice-cold water or hot (43°C) water or irradiation of cells with UV light (460 mm). For ethanol treatment, cells were incubated in 50 mM ethanol. The cells were harvested after 90 min of incubation. Poly(A) RNAs were purified by using oligo(dT)-cellulose, and 3 g of each purified RNA was used for Northern blot analysis as described before (39). The hybridization signal was quantified and normalized to the human ␤-actin by computerized image analysis (44).
Expression of PP2C␦ in 293 Cells-The coding sequence of PP2C␦ was amplified by PCR and cloned into the HindIII and XhoI sites of pcDNA3. The expression of PP2C␦ is under the control of the cytomeg- 1 The abbreviations used are: NTA, nitrilotriacetic acid; FACS, fluorescence-activated cell sorting; GFP, green fluorescent protein; IPTG, isopropyl-␤-D-thiogalactopyranoside; MyBP, myelin basic protein; PCR, polymerase chain reaction; MCAC, metal chelate affinity chromatography; CHO, Chinese hamster ovary; bp, base pair(s). alovirus promoter. 293 cells were transfected with the constructed plasmid and selected for neomycin-resistant cells (45).
Reconstruction of an Inducible, Dicistronic Expression Vector of pTR5-PP2C␦-GFP-The BamHI-HindIII fragment containing the entire open reading frame of the PP2C␦ cDNA was blunted and cloned into the blunted BglII site of a modified tetracycline-inducible dicistronic expression cassette of pTR5-DC-GFP vector wherein the two ribosome binding sites are placed in the same transcript (46). In the resultant construct, pTR5-PP2C␦-GFP, the PP2C␦ sequence preceded the GFP gene.
Cell Death and DNA Content Analysis-293 cells, which stably express the reverse tetracycline-controlled transactivator (47), were transfected with 20 g of pTR5-PP2C␦-GFP DNA by the calcium phosphate method as described previously (48). Culture medium was changed at 16 h after transfection, and doxycycline (a tetracycline derivative) was added at a final concentration of 1 g/ml for induction.
For cell death analysis, the transfected cells were harvested at 1, 2, 4, and 6 days after induction and analyzed by fluorescence-activated cell sorting (FACS). For cell cycle analysis, the transfected cells were harvested at 45 h after induction and stained with Hoechst 33342 at a final concentration of 10 M for 60 min. DNA content was analyzed by FACS using multicycle software (Phoenix Flow System, San Diego, CA).
DNA Synthesis Study-24 h after induction with doxycycline, cells transfected with pTR5-PP2C␦-GFP were separated by FACS, and approximately 20,000 cells/well were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum for 5 days for recovery. [ 3 H]Thymidine (3 Ci) was added to each well after 16 h of starvation and then incubated for an additional 20 h. Cells were treated with 0.1 N NaOH, and incorporated radioactivity was measured after washing and fixing with 10% trichloroacetic acid (49).
Western Blotting Analysis-Western blotting was performed as described previously (49,50). Briefly, 293 cells transfected with PP2C␦ containing plasmid were harvested at 24 h after induction with doxycycline and incubated in a lysis buffer consisting of 20 mM Tris-HCl, 2 mM MnCl 2 , 0.1 mM phenylmethysulfonyl fluoride, and 0.03% BRIJ 35 for 20 min on ice. The lysate was centrifuged at 4°C, and 8 l of supernatant was subjected to a 9% SDS-polyacrylamide gel electrophoresis, transferred to Hybond-C membrane, and immunoblotted with anti-PP2C␦ serum (number 63) at a final concentration of 1:2500. The specific band for PP2C␦ was detected by enhanced chemiluminescence.

Sequence Analysis and Expression of PP2C␦ in Rat
Tissues-A PCR fragment (550 bp) sharing a sequence homology with the PP2C family was obtained from rat hypothalamic cDNA by using degenerate primers designed from PP2C se- quences. This fragment was used as a probe to screen a rat brain stem-spinal cord cDNA library. Five positive clones were obtained, all possessing identical overlapping sequences, except one that had an insertion of 141-bp intron-like sequence located at nucleotide 1119 of the cDNA. The longest clone contains 1318 bp with a single open reading frame encoding for 392 amino acids (Fig. 1).
Comparison of the cDNA sequence with those in GenBank TM revealed that the deduced protein sequence shares a significant homology over 285 amino acid residues in the catalytic domain of the PP2C family: approximately 32% identity to the rat isoforms of PP2C␣ and PP2C␤ or 27% to FIN13, respectively (Figs. 2, A and B). Thus, this new phosphatase is designated as PP2C␦. The N-terminal sequence of PP2C␦, consisting of 76 amino residues, shares no homology with any known proteins. In addition, this novel phosphatase lacks the carboxyl-terminal 90-residue sequence (Fig. 2B), characteristic of mammalian PP2C (3,(51)(52)(53).
The expression of PP2C␦ in rat tissues was assessed by Northern blot analysis. A 32 P-labeled cDNA probe corresponding to the entire coding region of PP2C␦ was hybridized to a commercial membrane of multiple rat tissues. A band of approximately 1.4 kilobases was readily detected in most rat tissues studied. The level of transcription of the PP2C␦ gene was variable from tissue to tissue, with the highest expression observed in kidney, liver, and muscle (Fig. 3).
Expression and Characterization of Recombinant PP2C␦ Produced in Escherichia coli-To characterize PP2C␦ protein in vitro, the coding sequence of the phosphatase was cloned into a modified pET-3c vector in which 6 histidine codons at the N terminus were fused in frame with PP2C␦ and expressed in bacteria. The recombinant PP2C␦, when induced with IPTG, accounted for approximately 10% of total bacterial proteins and migrated at an apparent molecular mass of 50 kDa (Fig. 4), slightly larger than the deduced mass (44 kDa) predicted from the amino acid sequence. The recombinant PP2C␦ was purified by Ni ϩ -NTA-agarose chromatography to near homogeneity (Fig. 4) and was subsequently utilized for immunization and for in vitro characterization. The phosphatase activity of the recombinant PP2C␦ was assayed by dephosphorylating a substrate, MyBP, which was 32 P-labeled on serine/threonine residues. Recombinant PP2C␦ displayed a potent Mn 2ϩ -dependent dephosphorylation activity (Fig. 5A) over a concentration range from 0 to 4 mM (Fig. 5B). It is distinguished from the other PP2Cs, since the dephosphorylation activity of PP2C␦ was not stimulated by Mg 2ϩ (Fig. 5A). In fact, Mg 2ϩ exerted a concentration-dependent inhibition of the PP2C␦ activity (Fig. 5C). For instance, the dephosphorylation activity of PP2C␦ was completely inhibited when the concentration of Mg 2ϩ was 8 times higher than that of Mn 2ϩ in the assay mixture (Fig. 5C). The phosphatase activity of PP2C␦ was not sensitive to okadaic acid (Fig. 5C).
Transcriptional Activation of the PP2C␦ Gene in Response to Stress-It has been shown that eukaryotic PP2C can antagonize to kinase cascade activated by stress. To examine whether PP2C␦ is involved in stress-mediated signaling in mammalian cells, CHO cells were subjected to various stress treatments. mRNAs were isolated from the stress-treated cells and analyzed by Northern blotting. As shown in Fig. 6, expression of the PP2C␦ gene was increased by 60% following ice-cold water incubation and 40% in hot water. Most significantly, the mRNA level of PP2C␦ increased by 2.5-fold following exposure to ethanol and 4.5-fold following UV irradiation.
The Effect of PTP2C␦ Expression on Cell Growth-To further investigate the function of PP2C␦ in vivo, we attempted to stably express the phosphatase in mammalian cells. PP2C␦ was cloned into the pcDNA3 vector, transfected into 293 cell line, and cultured in neomycin-containing medium. As a con-trol, the vector pcDNA3 alone was transfected in the cells. After 3 weeks of drug selection, few colonies (less than 10/dish) were obtained from the PP2C␦-transfected cells, while the control consistently had more than 50 colonies/dish. Moreover, Western blot analysis revealed that the expression level of PP2C␦ protein in drug-resistant cells transfected with the construct was similar to that in control cells (data not shown), suggesting that the detected PP2C␦ in these drug-resistant cells was contributed only by the endogenous phosphatase protein. Similar results were observed in COS-7 and CHO cells (data not shown). Repeated attempts to stably overexpress PTP2C␦ in mammalian cells were not successful, indicating that overexpression of PP2C␦ phosphatase might inhibit cell growth or cause cell death.
Inducible Co-expression of PP2C␦ with GFP Revealed Its Inhibitory Effect on Cell Growth-To examine the effect of PTP2C␦ expression on cell growth in great detail, we employed a modified tetracycline-regulated expression vector, pTR5-PP2C␦-GFP, in which both PP2C␦ and GFP genes were constructed into a dicistronic cassette (pTR5-DC-GFP) with two ribosome binding sequences within the same transcript (46). 293 cells that stably express the transcriptional activator (reverse tetracycline-controlled transactivator) were transiently transfected with pTR5-PP2C␦-GFP and induced with doxycycline at 16 h post-transfection. After induction for 24 h, GFPpositive cells were separated by FACS. Western blot analysis showed that the amount of expressed PP2C␦ protein in GFPpositive cells increased about 10-fold when compared with the endogenous phosphatase level in control cells, indicating that both PP2C␦ and GFP were co-expressed in the cells (Fig. 7). Since the PP2C␦ gene is placed before the GFP gene in the dicistronic cassette, it is reasonable to use GFP expression as an indication of the expression of PP2C␦ in individual cells. Thus, the expression of PP2C␦ in cells can be examined by visualizing GFP protein under UV microscopy. Following 24-h induction, 20 -25% of cells were PP2C␦/GFPpositive as visualized under UV microscopy, and no noticeable morphological changes were observed at this time (Fig. 8A). However, most of the PP2C␦/GFP-positive cells gradually displayed an altered, rounded up cell morphology (Fig. 8, B and C) and finally detached from the plate in the following days. Weakly positive PP2C␦/GFP cells retained their normal morphology for an additional few days. Under microscopic examination, the number of PP2C␦/GFP-positive cells was dramatically decreased at day 4 of induction. At day 6 of induction, only a few weak PP2C␦/GFP-positive cells were attached on plates (data not shown). Control cells, which expressed GFP alone, retained a normal morphology and progressed to the usual proliferation over the corresponding days of induction (Fig.  8D).
It should be mentioned that the expression cassette for PP2C␦ in 293 cells was not tightly regulated by doxycycline due to a leaking expression of the activating factor, reverse tetra-cycline-controlled transactivator, in the system. Cells transiently transfected with pTR5-PP2C␦-GFP in the absence of inducer expressed PP2C␦ protein approximately 2-3 times more than the endogenous phosphatase (Fig. 7). These transfected cells remained as normal for approximately 2 weeks and detached finally. Hence, stably expressing PP2C␦ clones with a high level were not capable of being isolated in this inducible system. As a control, clones stably expressing GFP alone were readily established.
Overexpression of PP2C␦ Blocks Cell Cycle Progression and DNA Synthesis-The effect of overexpressed PP2C␦ protein on cell growth was further characterized by flow cytometric analysis. As above, 293 cells were transfected with pTR5-PP2C␦-GFP and induced at 16 h post-transfection. The PP2C␦/GFPpositive cells were recovered by FACS following induction. Cell death rate was calculated directly by FACS based on cell size. Following 2-day induction, the number of PP2C␦/GFP-positive cells dropped slightly from 22.5 to 17.3%, and the number of dead cells in the PP2C␦/GFP-positive cells accounted for 8.5%, which was not significantly higher than that of nontransfected cells (6.4%). However, the PP2C␦/GFP-positive cells substantially decreased over the next few days. For example, PP2C␦/ GFP-positive cells dropped to 6.1%, and their cell death rate was increased to 15.3% by day 4 of induction. At day 6 of induction, the rate of cell death of PP2C␦/GFP-positive cells reached 38%, leaving PP2C␦/GFP-positive cells accounting for only 2% of the total cell population (Table I). These results suggested that overexpression of PP2C␦ resulted in a deterioration of cell proliferation and induced cell death. To examine this hypothesis, DNA synthesis was assayed. The transiently transfected 293 cells were induced, and PP2C␦/GFP-positive cells were separated by using flow cytometry following 24-h induction. The PP2C␦/GFP-positive cells were then cultured for an additional 5 days. Examination by UV microscopy revealed that most PP2C␦/GFP-positive cells were not attached to the plate, and only a very small portion of weakly expressing PP2C␦/GFP cells were attached. DNA synthesis in these PP2C␦/GFP-positive cells revealed that whereas the rate of [ 3 H]thymidine incorporation in control cells reached 1700 cpm/ well within 20 h, the rate of [ 3 H]thymidine incorporation in the PP2C␦/GFP-positive cells was barely detectable (Fig. 9).
To further determine at which stage of the cell cycle progression was blocked, we investigated the effect of overexpressed PP2C␦ on the cell cycle distribution. 293 cells were transiently transfected with pTR5-DC-GFP for expressing GFP alone and with pTR5-PP2C␦-GFP for expression of both PP2C␦ and GFP. Cells were harvested at 45 h following induction, and GFPpositive cells were collected by FACS and stained with Hoechst 33342. The DNA content of the cells was analyzed by flow cytometry. As shown in Fig. 10, nontransfected cells (Fig. 10A) and cells expressing GFP only (Fig. 10C) displayed similar patterns of cell cycle distributions (Table II)  expressing both GFP and PP2C␦ exhibited a significant accumulation at the S phase, particularly early S phase, and concomitantly a significant decrease of cells in G 0 /G 1 as well as in G 2 /M phases ( Fig. 10B and Table II). A further increase of cell population in S phase and concomitantly an additional decrease of cells in G 2 /M phase were observed when cells were induced for a longer period of time, such as for 70 h (data not shown).

DISCUSSION
In the present study, we identify and characterize a novel protein phosphatase, PP2C␦. The PP2C␦ sequence can be divided into two distinct domains: an N-terminal domain of 76 amino acids and a catalytic domain encompassing the rest of the amino acid sequence. The N-terminal domain, which is absent in other mammalian PP2Cs and shares no homology with any known proteins, may enable PP2C␦ to specifically interact with other proteins, such as its physiological substrates. The catalytic domain of PP2C␦ shares approximately 30% sequence identity with other mammalian PP2Cs. However, PP2C␦ does not contain the 90-residue carboxyl-terminal sequence conserved in other members of the mammalian PP2C family (1,3). The overall structural features of PP2C␦ render it distinct from other mammalian PP2C members. Recombinant PP2C␦ expressed and purified from E. coli displayed a potent Mn 2ϩ -dependent dephosphorylating activity on serine/threonine residues. Strikingly, Mg 2ϩ , which stimulates the dephosphorylation activity of other PP2Cs (22,23,26,27), exerted a concentration-dependent inhibition of PP2C␦ activity. When the concentration of Mg 2ϩ was 8-fold higher than that of Mn 2ϩ , it completely inhibited the Mn 2ϩ -activated dephosphorylation activity of PP2C␦. The inhibitory effect of Mg 2ϩ on the activity of PP2C␦ is likely through competitive inhibition of Mn 2ϩ binding to conserved metal binding sites in the catalytic domain of PP2C␦.
Northern blot analysis showed that PP2C␦ is ubiquitously expressed in rat tissues, with high expression levels in muscle, liver, and kidney. As with many other ubiquitously expressed proteins, PP2C␦ may play an important role in determining the phosphorylation state of proteins participating in multiple signal transduction pathways in various tissues, rather than being restricted to a few target cell types as is the case with other mammalian PP2Cs. Transcriptional activation of either protein phosphatase or protein kinase genes has been suggested as an important signaling in response to stress (29,32). Recently, it has been reported that PP2C genes of yeast and plants can be transcriptionally activated to antagonize the protein kinase cascade stimulated by stress (29,30,32,33,54). In comparison with these eukaryotic PP2Cs, little is known about the biological function of mammalian PP2C in response to stress. Expression of the PP2C␦ gene in CHO cells in response to environmental stresses was therefore evaluated by Northern blot analysis. Following the treatments, it was found that expression of the PP2C␦ gene increased at least 40% in response to cold temperature and as much as 3.5-fold following UV irradiation. These results suggest that the expression of the PP2C␦ gene is positively regulated by these stresses. Further studies should include the transcriptional and translational regulation   Fig. 10. of the PP2C␦ gene and its correlation with protein kinases, such as stress-activated protein kinase/c-Jun N-terminal kinase, MSK1, and Krs, all of which have been suggested to be main stress-responsive molecules (55)(56)(57).
Repeated attempts to establish stably expressing PP2C␦ cell lines in 293 or CHO cells were all unsuccessful, since clones selected in drug resistance showed no appreciable additional PP2C␦ when compared with levels of the endogenous protein.
These results suggest that PP2C␦ negatively regulates cell growth and/or that overexpression of PP2C␦ leads to cell death. To further investigate the effect of overexpression of PP2C␦ on cell growth, we used a modified tetracycline-inducible dicistronic cassette, pTR5-DC-GFP, which allows for the translation of the mRNA initiating from two ribosome binding sites within the same transcript (46). Induction of pTR5-PP2C␦-GFP-transfected 293 cells resulted in co-expression of both PP2C␦ and GFP proteins in individual cells, as shown by Western blot and the visualization of the GFP protein. After transfection and induction, over 20% of the cells expressed PP2C␦ as indicated by co-expressed GFP protein. However, these cells were rapidly lost after induction. Cells expressing GFP alone remained normal throughout the experiment. Similarly, DNA synthesis was negligible in PP2C␦/GFP-positive cells. Furthermore, analysis of DNA content revealed that overexpression of PP2C␦ causes cell cycle arrest in the S phase, particularly in early S phase, and concomitantly depletes cells in G 0 /G 1 and G 2 phases ( Fig.  10 and Table II). Taken together, these data suggest that overexpression of PP2C␦ negatively regulates cell growth by blocking the cell cycle progression in early S phase, resulting in the inhibition of DNA synthesis and leading to cell death, which caused the loss of cell attachment in cultured plates as observed.
The overexpression of a protein phosphatase may constitutively dephosphorylate its physiological target(s) resulting in cell cycle arrest. For example, phosphatase MKP1, when overexpressed in cells, dephosphorylates MAP kinase and blocks fibroblasts entering into the S phase (58,59). Overexpression or unregulated expression of PP2C␦ under these experimental conditions does not likely possess a direct effect on the physiological activity of the phosphatase in normal cells. A question thus arises as to the mechanism by which overexpressed PP2C␦ inhibits cell growth and leads to cell cycle arrest. One possibility is that overexpression of PP2C␦ may cause a widespread and nonspecific dephosphorylation of many proteins that are toxic to cells. Although this possibility cannot be completely ruled out, we expect that if this were indeed the case, the cell cycle would be arrested at any phase (37), not only in the early S phase. In addition, we observed that transfected cells with the inducible pTR5-PP2C␦-GFP cassette were not able to establish stable clones, suggesting that a low level of leaking expression of exogenous PP2C␦ is sufficient for inhibition of cell growth, leading to cell death. It is therefore conceivable that cell cycle arrest in cells overexpressing PP2C␦ may be caused by the unregulated dephosphorylation of its physiological substrates from which appropriate phosphorylations are required for normal cell cycle progression.
Interestingly, it has been recently reported that a growth factor-inducible serine/threonine phosphatase, FIN13, when overexpressed, exerted a similar inhibitory effect on cell growth, resulting in the accumulation of cells in G 1 and early S phase (37). Although PP2C␦ and FIN13 share only 27% sequence homology in their catalytic domains and display very different structural features and patterns of tissue expression, the two phosphatases may use similar mechanisms to inhibit cell growth through the dephosphorylation of their specific substrates, which, when appropriately phosphorylated, are es-sential for DNA synthesis and cell proliferation. At present, however, the precise function of both PP2C␦ and FIN13 in vivo is unclear. Defining physiological substrates for these two phosphatases will increase our understanding of their biological functions. It will be very interesting to determine whether PP2C␦ and FIN13 have some common physiological targets for the regulation of cell proliferation.