M Phase-specific Expression and Phosphorylation-dependent Ubiquitination of the ClC-2 Channel*

Cl− channel activities vary during the cell cycle and are thought to play various roles including regulation of cell volume. We have shown previously that ClC-2 channels are directly phosphorylated and functionally regulated by the M phase-specific cyclin-dependent kinase p34cdc2/cyclin B. We investigate here to determine whether the expression levels of ClC-2 channel protein vary during the cell cycle. Immunoblot and immunocytochemical analyses of cells cycle-synchronized by serum depletion/replenishment reveal that ClC-2 channel protein is expressed predominantly at M phase in cells with two nuclei and a clear constriction ring, whereas RNA blot analysis shows that ClC-2 mRNA expression does not change during the cell cycle. Ubiquitin assays reveal that the ClC-2 channels are ubiquitinated at M phase, whereas the magnitude of ubiquitination is suppressed by incubation with olomoucine, an inhibitor of p34cdc2/cyclin B, and it is almost completely abolished in ClC-2 channels having an S632A mutation, which cannot be phosphorylated by p34cdc2/cyclin B, indicating that ubiquitination of ClC-2 channels requires phosphorylation by M phase-specific p34cdc2/cyclin B. Regulation at the post-transcriptional level, including phosphorylation-dependent ubiquitination, may contribute to M phase-specific expression of ClC-2 channels. Cell cycle-dependent regulation of expression at the protein level in addition to the regulation of function suggests that the ClC-2 channel plays a physiological role in the cell cycle.

Cl Ϫ channel activity varies during the cell cycle and is thought to play several roles in the progression of the cell cycle including regulation of membrane potential and cell volume (1,2). At M phase, activation of inwardly rectifying Cl Ϫ currents has been reported in two species, Caenorhabditis elegans oocytes (3) and ascidian embryos (4,5). The biophysical properties of the Cl Ϫ channels activated at M phase in these species (3)(4)(5)(6) strongly resemble those of the ClC-2 channel (7-10); both are activated by hyperpolarization, show inward rectification and no inactivation, and are activated by cell swelling. Accord-ingly, mammalian ClC-2 channels may well be involved in progression of the cell cycle. Recently, ClC-2-deficient mice were established that exhibit male infertility due to failure of germ cell meiosis, suggesting that the ClC-2 channel plays an important role in the transepithelial transport of Sertoli cells to maintain ionic homeostasis in seminiferous tubules (11). However, the function of the ClC-2 channels ubiquitously expressed in mammalian cells remains unclear.
Recently, we have shown that the ClC-2 channel is the target of regulation by the M phase-specific cyclin-dependent kinase p34 cdc2 /cyclin B; the C terminus of ClC-2 is directly phosphorylated by p34 cdc2 /cyclin B, and channel activities are inhibited by phosphorylation by p34 cdc2 /cyclin B when ClC-2 channels are expressed in Xenopus oocytes (12). In the present study, we have attempted to determine whether the expression levels of ClC-2 channel protein also vary during the cell cycle. Our data indicate that the expression of ClC-2 channel protein is M phase-dependent; it is highest in dividing cells at M phase and thereafter decreases rapidly. Ubiquitin assay showed that ClC-2 channel protein is ubiquitinated at M phase and that ubiquitination is dependent on phosphorylation by p34 cdc2 /cyclin B. Accordingly, the rapid decrease in ClC-2 channel expression after cell division is likely to be due at least in part to ubiquitination and subsequent degradation of the ClC-2 channel protein. The M phase-specific expression and the phosphorylation of ClC-2 channel protein suggest a physiological role of ClC-2 channels in the cell cycle.
Cell Cycle Synchronization-Exponentially growing NRK-49F or NIH3T3 cells were introduced into quiescence phase (G 0 /G 1 ) by serum deprivation for 30 -40 h, when the cells were released from G 0 /G 1 phase by the addition of 10% fetal calf serum (FCS) for NRK-49F cells and 15% FCS for NIH3T3 cells (16,17). At the appropriate times after serum addition, the cells were collected, and cellular DNA content was determined by staining with propidium iodide and measuring fluorescence by fluorescence-activated cell sorter (EPICS Elite ESP; Beckman Coulter).
Immunocytochemistry-NIH3T3(rbClC-2-HA) cells were grown on cover glasses coated with collagen. Cells were fixed in 2% paraformaldehyde, permeabilized with 0.1% Triton X-100 for 20 min at room temperature, and incubated with high affinity anti-HA rat antibody (Roche Molecular Biochemicals) at a concentration of 0.1 g/ml followed by incubation with 1:500-diluted fluorescein isothiocyanate-conjugated rabbit anti-rat IgG. Cover glasses were mounted on a slide and observed under a confocal microscope (Fluoview FV300; Olympus).
Preparation of Crude Membrane Fraction and Immunoblot Analysis-Crude membrane fractions were prepared from untransfected NRK-49F cells or HEK293 cells transiently transfected with pcDNA3.1(ϩ)/rbClC-2-HA or pcDNA3.1(ϩ)/rbClC-2(S632A)-HA as described previously (19). Briefly, the cells were washed 3 times with phosphate-buffered saline (80 mM Na 2 HPO 4 , 20 mM NaH 2 PO 4 , and 100 mM NaCl (pH 7.4)), suspended in buffer A (50 mM Tris (pH 7.5), 1 mM EDTA, 1 mM phenylmethanesulfonyl fluoride, and 10 g/ml each aprotinin, leupeptin, and pepstatin), homogenized, and centrifuged at 600 ϫ g at 4°C for 10 min to remove cell debris and nuclear fraction, and the supernatant was centrifuged again at 10,000 ϫ g at 4°C for 1 h. The pellets were resuspended in 500 l of buffer A (crude membrane fraction) and stored at Ϫ80°C until immunoblot analysis. The crude membrane proteins (60 g for NRK-49F cells and 20 g for HEK293 cells) were boiled in SDS reducing sample buffer, electrophoresed on an SDS, 7% polyacrylamide gel, and transferred to a nitrocellulose membrane by electroblotting overnight at 4°C at 200 mA. The membrane was incubated with 1:300-diluted anti-ClC-2 rat antibody (Alomone Laboratories) for NRK-49F cells or 1:1000-diluted anti-HA rat antibody (Roche Molecular Biochemicals) for NIH3T3 cells followed by incubation with 1:800-diluted horseradish peroxidase-conjugated anti-rat IgG (Biosys), and the proteins were detected using an enhanced chemiluminescence system (Amersham Biosciences). Protein concentrations were determined using the bicinchoninic acid assay (Pierce). The protein expression level was quantified by measuring the immuno-density of the main band at about 100 kDa by NIH Image software.
Phosphorylation Assays in Cultured Cells-48 h after transfection of HEK293 cells in a 60-mm culture dish with 6 g of mock, pcDNA3.1(ϩ)rbClC-2-HA, or pcDNA3.1(ϩ)rbClC-2(S632A)-HA, the culture medium was replaced with 3 ml of phosphate-free Dulbecco's modified Eagle's medium (Invitrogen) containing 1 mCi of [ 32 P]orthophosphate (specific activity 900 ϳ 1100 mCi/mmol, PerkinElmer Life Sciences), and the cells were incubated at 37°C for 3 h. In experiments examining the effects of olomoucine, olomoucine at concentrations up to 100 M was added. Immunoprecipitation with anti-HA high affinity antibody was carried out as described above. Protein-antibody complexes were separated on two SDS, 7% polyacrylamide gels. One gel was subjected to immunoblotting with anti-HA antibody to check the level of ClC-2 protein expression in plasma membrane and the efficiency of immunoprecipitation. The density of immunoprecipitated protein blotted with anti-HA antibody was quantified using NIH Image software. The other gel was subjected to autoradiography, and the magnitude of phosphorylation was quantified using BAS1000 image analyzer (Fuji Photo Film Co. Ltd) and normalized to the density of immunoprecipitated protein.
Statistics-All measured values are presented as means Ϯ S.D. Analysis of variance was used to test for significance (p Ͻ 0.05). Fig. 1A shows a representative experiment of cell cycle synchronization in NRK-49F cells. The first peak delineates cells in G 0 /G 1 phase, the second peak delineates cells in G 2 /M phase, and the area between the two peaks delineates cells in S phase. Cells up to 10 h after serum addition are at G 0 /G 1 phase, 12-16 h are at S phase, 18 -22 h are at G 2 /M phase, and 24 h are at G 0 /G 1 phase again. To determine whether the expression of ClC-2 mRNA or protein varies during the cell cycle, RNA blot and immunoblot analysis were performed in cells at various time points after serum addition. The expression level of ClC-2 mRNA apparently did not change during the cell cycle (Fig.  1B). In contrast, the expression level of ClC-2 channel protein varied markedly during the cell cycle (Fig. 1C); an immunopositive band was barely detectable between 6 and 20 h after serum addition, but a strong immunopositive band at about 100 kDa, which is approximately the expected molecular size of the ClC-2 channel, was detected at 22 h. At 24 and 26 h, the intensity of the immunopositive band at 100 kDa was markedly diminished compared with that at 22 h.

RESULTS
To further examine cell cycle-specific expression of ClC-2 channel protein, the ClC-2 channel was immunostained at various stages of the cell cycle in NIH3T3 cells stably expressing rbClC-2-HA (NIH3T3(rbClC-2-HA)). Fig. 2A shows representative differential interference contrast images (upper panels) and immunofluorescence images (lower panels) at 18, 22, and 26 h after serum addition. Although the differential interference contrast image shows the presence of many cells at Because the ubiquitination-26 S proteasome pathway is frequently utilized to degrade proteins with short half-lives (less than 2 h) (21,22), including several cell cycle-related proteins (23), ubiquitination could well be involved in the rapid decrease of ClC-2 channel protein after cell division. To determine whether this is the case, NIH3T3(rbClC-2-HA) cells treated with 26 S proteasome inhibitor zLLLal (10 M) (24, 25) for 2 h was immunoprecipitated with anti-HA antibody followed by immunoblotting with anti-ubiquitin antibody. A smear pattern was observed at molecular weights greater than about 100 kDa (the third lane in Fig. 3A). In controls, the only weak smear was observed in NIH3T3(rbClC-2-HA) cells without treatment with zLLLal (second lane), and there was no smear in NIH3T3 cells transfected with mock (first lane). These findings indicate that ClC-2 channels are polyubiquitinated. To determine whether ubiquitination of the ClC-2 channels varies during the cell cycle, we performed ubiquitin assay in cells collected at various time points after serum addition. The magnitude of ubiquitination was greater when zLLLal was added between 21-23 h and cells were collected at 23 h rather than at other time points (Fig. 3C).
We found in previous experiments using in vitro and cell-free phosphorylation assays that the ClC-2 channel is phosphoryl- ated by the M phase-specific cyclin-dependent kinase p34 cdc2 / cyclin B at Ser-632 in the C terminus (12). Because the tryptophan-tryptophan (WW) domain of Nedd4, an E3 ubiquitin ligase, binds to cdc25 only when it is phosphorylated at the exit of the mitotic phase (26,27), phosphorylation by p34 cdc2 /cyclin B might well link the ClC-2 channels with an E3 ubiquitin ligase and ubiquitination. We first determined whether or not ClC-2 channels are phosphorylated in vivo. In vivo phosphorylation assay revealed incorporation of 32 P at about 100 kDa in HEK293 cells transfected with rbClC-2-HA but not in mocktransfected HEK293 cells (Fig. 4A). 32 P incorporation was markedly diminished in HEK293 cells transfected with rbClC-2(S632A)-HA, a mutant in which Ser-632 in the consensus sequence of phosphorylation by p34 cdc2 /cyclin B is replaced with Ala (Fig. 4A). Incubation with olomoucine, a cyclin-dependent kinase inhibitor (28), also diminished 32 P incorporation (Fig. 4B). To investigate the relationship between ubiquitination of ClC-2 channel and its phosphorylation by p34 cdc2 / cyclin B, we performed a ubiquitin assay in the presence of olomoucine and in cells transfected with rbClC-2(S632A). The magnitude of ubiquitination was significantly suppressed by olomoucine and was almost completely abolished in cells transfected with rbClC-2(S632A) (Fig. 5).
We examined HEK293 cells transiently transfected with pcDNA3.1(ϩ)/rbClC-2-HA or pcDNA3.1(ϩ)/rbClC-2(S632A)-HA to determine if zLLLal or olomoucine affected the steady-state level of ClC-2 channel expression. Compared with control conditions (first lane in Fig. 6), incubation with zLLLal (10 M) for 6 h (second lane in Fig. 6, A and B) or olomoucine (10 M) for 6 h (third lane in Fig. 6, A and B) significantly increased the expression level of ClC-2 channel protein. In NIH3T3 cells transfected with rbClC-2 containing the S632A mutation, the level of ClC-2 channel expression also was significantly increased (fourth lanes in Fig. 6, A and B) compared with control conditions.

DISCUSSION
In the present study, we examined ClC-2 channel expression levels during the cell cycle. We find that ClC-2 channel protein is expressed predominantly in dividing cells at M phase, that ClC-2 channel protein is ubiquitinated strongly at M phase, and that this ubiquitination is dependent on phosphorylation of the channel by p34 cdc2 /cyclin B, which leads to proteasomal degradation of ClC-2 channel protein.
Immunoblotting and immunocytochemistry show that ClC-2 channel protein is expressed predominantly in dividing cells. Because RNA blot analysis revealed no appreciable alteration in mRNA expression level during the cell cycle, M phase-dominant expression of ClC-2 channel protein is most likely regulated at the post-transcriptional level rather than at the transcriptional level. ClC-2 channel protein expression rapidly decreased immediately after division of the cell, suggesting active degradation of ClC-2 channel protein at this point in the cell cycle. Several proteins containing a PEST sequence of enriched Pro, Glu, Ser, and Thr flanked by positively charged amino acids, particularly in the C terminus, have short intracellular half-lives (less than 2 h) (29,30). They appear to be important as signals for ubiquitination (30,31); for example, phosphorylation of the G 1 cyclin in yeast within the PEST region leads to recognition by an E3 ubiquitin ligase (31). Analysis using PEST-FIND software found the five PEST sequences to be clustered in the C terminus of ClC-2 channels, and ubiquitin assays showed ClC-2 channels to be polyubiquitinated. Because ubiquitination of ClC-2 channels is highest around the exit of M phase (21-23 h), it may well be responsible for the rapid disappearance of the ClC-2 channels after cell division. The mechanisms responsible in up-regulation of ClC-2 channel protein expression at M phase, on the other hand, are not known. Although in many cell cycle regulators such as the cyclins, cyclic protein expression occurs due to a burst of transcription that precedes the regulated degradation, cyclic expression of some proteins occurs due to modulation of protein synthesis (32,33). The latter may be the case in the ClC-2 channel, because there is no noticeable regulation of ClC-2 channel transcription (Fig. 1B). To clarify the mechanism of regulated protein synthesis of the ClC-2 channel requires further investigation and is beyond the scope of the present study. However, it is interesting that cell cycle-dependent phosphorylation of the translational repressor binding protein (4E-BP1) by p34 cdc2 /cyclin B was recently found to cause its dissociation from the eukaryotic initiation factor 4F (eIF-4F), switching on mRNA translation at M phase (34).
We found previously using in vitro and cell-free phosphorylation assays that ClC-2 channels are phosphorylated by the M phase-specific cyclin-dependent kinase p34 cdc2 /cyclin B at Ser-632 in the C terminus (12). We show in the present study that ClC-2 channels are phosphorylated in vivo. Application of olomoucine or the presence of the S632A mutation markedly but not completely diminished the magnitude of phosphorylation, suggesting that p34 cdc2 /cyclin B is at least in part responsible for phosphorylation of the ClC-2 channel at 632 Ser and also that the ClC-2 channel is phosphorylated by some other kinase at a residue other than Ser-632. This finding is consistent with the previous report that the ClC-2 channel expressed in Xenopus oocytes is functionally regulated by protein kinase A (9). In contrast, ClC-2 channel ubiquitination was almost completely abolished by the replacement of Ser-632 to Ala and was strongly inhibited by olomoucine. These data indicate that ubiquitination of the ClC-2 channel is dependent on phosphorylation of the channel by p34 cdc2 /cyclin B.
In addition to being a signal of proteasomal degradation, ubiquitination is now recognized to be a signal in a broad spectrum of events, including endocytosis, intranuclear trafficking, viral budding, and endosome trafficking (35). In the present study, inhibition of a 26 S proteasome with zLLLal or of phosphorylation by p34 cdc2 /cyclin B with olomoucine or insertion of the S632A mutation raised the steady-state level of ClC-2 channel protein, indicating that this ubiquitination participates in proteasomal degradation.
We showed previously that ClC-2 channels are functionally regulated by M phase-specific phosphorylation (12). ClC-2 channel currents expressed in Xenopus oocytes are inhibited by phosphorylation by p34 cdc2 /cyclin B. Accordingly, p34 cdc2 /cyclin B may well be important both in regulation of ClC-2 channel function and in expression of channel protein in a cell cycle-dependent manner. Further study is required to clarify the physiological role of the ClC-2 channel in the cell cycle.