Casein kinase 1 regulates connexin-43 gap junction assembly.

Phosphorylation of members of the connexin family of gap junction proteins has been correlated with gap junction assembly, but the mechanisms involved remain unclear. We have examined the role of casein kinase 1 (CK1) in connexin-43 (Cx43) gap junction assembly. Cellular co-immunoprecipitation experiments and in vitro CK1 phosphorylation reactions indicate that CK1 interacted with and phosphorylated Cx43, initially on serine(s) 325, 328, or 330. (32)P(i)-Metabolically labeled cells treated with CKI-7, a specific CK1 inhibitor, showed a reduction in Cx43 phosphorylation on site(s) that can be phosphorylated by CK1 in vitro. To examine CK1 function, normal rat kidney cells were treated with CKI-7, and Cx43 content was analyzed by Triton X-100 extraction, cell-surface biotinylation, and immunofluorescence. Western blot analysis indicated a slight increase in total Cx43, whereas gap junctional (Triton-insoluble) Cx43 decreased, and non-junctional plasma membrane Cx43 increased (as detected by cell surface biotinylation). Immunofluorescence experiments in the presence of CK1 inhibitor showed increases in Cx43 plasma membrane localization but not necessarily accumulation at cell-cell interfaces. Decreased gap junctional and phosphorylated Cx43 was also detected when cells were treated with IC261, a CK1 inhibitor specific for delta or epsilon isoforms. These data suggest CK1delta could regulate Cx43 gap junction assembly by directly phosphorylating Cx43.

Gap junctional intercellular communication facilitates direct communication among adjacent cells by allowing passage of molecules less than 1000 Daltons (1,2). Consisting of hundreds of intercellular channels, gap junctions are thought to be critically important in regulating embryonic development, excitable cell contraction, tissue homeostasis, and normal cell growth and differentiation (3,4). Gap junctions are composed of integral membrane proteins from the connexin gene family. Approximately 20 members have been cloned and characterized in humans (3). During intercellular channel formation, six connexin proteins oligomerize into a hemi-channel or connexon followed by connexon trafficking to the plasma membrane. The intact channel is formed when one hemi-channel docks with a second in an opposing cell. Once assembled, groups of these intercellular channels (termed gap junctional plaques) mediate the passage of amino acids, second messengers, and other metabolites between the connected cytoplasmic domains (1,2). The channels can be gated in response to various stimuli, including changes in voltage, pH, and connexin phosphorylation. Regulation of gap junctional communication could occur by controlling any one of the steps mentioned above; however, many of the regulatory mechanisms underlying these events remain elusive.
Connexin-43 (Cx43), 1 the most ubiquitously expressed connexin, has a relatively short half-life (1-5 h) compared with most integral membrane proteins (5)(6)(7)(8). This fast turnover rate could imply a high level of post-translational regulation. Indeed, Cx43 is differentially phosphorylated throughout its life cycle in homeostatic cells (6,8,9). Cellular Cx43 demonstrates multiple electrophoretic isoforms when analyzed by SDS-PAGE, including a faster migrating, non-phosphorylated (NP) form and at least two slower migrating forms commonly termed P1 and P2. Both P1 and P2 co-migrate with NP after alkaline phosphatase treatment, suggesting that phosphorylation is the primary covalent modification detected in SDS-PAGE analysis (9). Phosphoamino acid analysis indicates the majority of the phosphorylation events occur on serines (9 -11), although tyrosine phosphorylation has also been observed in the presence of activated pp60 src (12,13). Pulse-chase studies using brefeldin A indicate some Cx43 phosphorylation occurs before reaching the plasma membrane (14). In addition, studies investigating phosphorylation in normal rat kidney (NRK) cells show that Cx43 acquires resistance to Triton X-100 once it has been phosphorylated to the P2 form and assembled into gap junction plaques (8). Kinase activators, which can significantly increase levels of Cx43 phosphorylation, have allowed the linkage of activation and Cx43 phosphorylation at specific sites to channel closure. These approaches have implicated roles for protein kinase C and mitogen-activated protein kinase in Cx43 phosphorylation and acute gating of gap junction channels (15,16). Based on several studies, Cx43 undergoes multiple phosphorylation events because at least five phosphorylated serines have been detected on Cx43 isolated from unstimulated cells (17). These uncharacterized phosphorylation events have been correlated with changes in assembly, acquisition of Triton X-100 insolubility, and degradation of Cx43 gap junction channels and, hence, could play critical roles in regulating gap junctional communication. Here, we report evidence identifying casein kinase 1 (CK1) as the first kinase found to play a role in the processing of Cx43 and gap junction assembly in homeostatic NRK cells.
Recent reports link CK1, a constitutively active kinase, to cell cycle progression (18), nuclear protein translocation (19), intracellular protein trafficking (20), and cell morphogenesis (21). Ubiquitously expressed in many cell types, several CK1 isoforms including ␣, ␤, ␥ 1-3 , ␦, and ⑀ have been identified and localized to different tissues and subcellular locations. Although CK1s are reported to be constitutively active, the CK1␦ and -⑀ isotypes have autoregulatory domains (22,23). CK1␦ and -⑀ both phosphorylate p53 and are thought to be involved in regulating DNA repair and chromosomal segregation (24). Inhibition of CK1␦ and -⑀ activity has been shown to trigger mitotic checkpoint control leading to mitotic arrest (25). Increased CK1␦ transcription has been correlated with brain tissue affected by Alzheimer's disease (26). CK1⑀ activity has also been implicated in Wnt signal transduction and appears to affect Wnt/␤-catenin-mediated gene regulation (27,28). Thus, CK1 activity plays a number of diverse biological roles and is critical for regulation of many cellular events.
In this study, we have investigated the role of CK1 activity on Cx43 gap junction assembly and function in homeostatic NRK cells. Our data suggest that Cx43 is a direct substrate for CK1, most likely the ␦ isoform. In addition, investigation of Cx43 trafficking after inhibition of CK1 function implies a role for CK1 activity in governing assembly of Cx43 gap junction channels.
To determine the sites of CK1␦ phosphorylation, HisCx43CT was phosphorylated by purified CK1␦ as described above. Phosphorylated HisCx43CT was digested overnight with sequencing grade trypsin at 37°C (Promega, Madison WI). The tryptic fragments were separated with a Vydac C18 column (218TP54) on a Hewlett Packard 1050 LC using a linear gradient of acetonitrile from 0 to 5% over 7 min, 5 to 40% over 25 min, and 40 to 99% over 17 min (with 0.06% trifluoroacetic acid throughout). Fractions were taken every min (up to 42 min) and Cerenkov-counted. Phosphorylated fractions (as determined by radioactivity incorporation) were concentrated in a Savant Speed Vac rotary evaporator and analyzed at the Fred Hutchinson Cancer Research Center Mass Spectrometry Facility by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry.
HisCx43CT/cell lysate phosphorylation analysis was done according to a standard published procedure (30) with minor modifications. Briefly, NRK cells were lysed on ice in lysis buffer (100 mM NaCl, 50 mM NaF, 1 mM Na 3 VO 4 , 0.25% Triton X-100, 50 mM HEPES, pH 7.4, supplemented with 2 mM phenylmethylsulfonyl fluoride, and 1ϫ Complete protease inhibitors; Roche Molecular Biochemicals). Lysates were clarified by centrifugation at 4°C, and 5 l of lysate was mixed in duplicate with HisCx43CT bound to Talon Affinity Chromatography Resin (Clontech, Palo Alto, CA). Phosphorylation reactions were performed by adding 5 l of reaction mixture containing 5 mM MgCl 2 , 50 M ATP, [␥-32 P]ATP (ϳ1 Ci) in 20 mM Tris-Cl, pH 7.5, at 25-30°C for 30 min. To analyze phosphorylation mediated by CK1 in these experiments, we added the CK1␦ and -⑀ inhibitor, IC261 (2.5 M), to one of the pair of duplicate reactions. Samples were separated by SDS-PAGE and analyzed by autoradiography.
For co-precipitation of CK1 and Cx43 proteins, NRK cells were rinsed in PBS and lysed on ice in RIPA buffer (25 mM Tris-HCl, 100 mM NaCl, 10 mM EDTA, 50 mM NaF, 500 M Na 3 VO 4 , 0.25% Triton X-100, 0.02% NaN 3 , 2 mM phenylmethylsulfonyl fluoride, and 1ϫ Complete protease inhibitors). Cx43 was immunoprecipitated by mixing NRK cell lysates with Cx43 (Cx43CT2) antibody cross-linked to protein A for 2 h at 4°C. After several washes in PBS or RIPA buffer, Cx43 (along with interacting proteins) was eluted in 2ϫ Laemmli sample buffer and separated by SDS-PAGE on 12% polyacrylamide gels. Protein was transferred to nitrocellulose, the membrane was blocked, and antibodies were incubated as previously indicated (11). Primary and secondary antibodies utilized included mouse and rabbit anti-zona occludens 1 (anti-ZO-1) (Zymed Laboratories Inc., San Francisco, CA), rabbit anti-Cx43 (C6219, Sigma), and peroxidase-conjugated donkey anti-mouse or mouse antirabbit secondary antibodies (Jackson Immunoresearch Laboratories, Inc., West Grove, PA). Signal was visualized with SuperSignal West Pico or Femto chemiluminescent substrate (Pierce) followed by exposure to Kodak Biomax MR film.
For IC261 characterization, CK1␦ and CK1␣ were immunoprecipitated from NRK RIPA buffer cell lysates (generated as described above). After several washes in RIPA buffer (no SDS), immunoprecipitates were autophosphorylated in reaction buffer containing 5 mM MgCl 2 , 50 M ATP, [␥-32 P]ATP (ϳ1 Ci) in 50 mM Tris-Cl, pH 7.5 at 25-30°C for 30 min in the presence or absence of 2.5 M IC261. Samples were separated on 12% polyacrylamide gels, and the gels were dried and analyzed by autoradiography. 32 P i Metabolic Labeling of NRK Cultures-Cell cultures were metabolically labeled as previously described (17,31). Briefly, cells were labeled for 4 h at 37°C in the presence or absence of CK1 inhibitors CKI-7 (50 or 100 M) or IC261 (2.5 M). After cell labeling, cells were washed 3 times in cold medium and solubilized in cold RIPA buffer containing 0.6% SDS and 1% Triton X-100. Cells were harvested, and the DNA was sheared by drawing the lysate through a 26-gauge needle. After centrifugation, immunoprecipitation of Cx43 was performed as described above. Samples were boiled and analyzed on 10% polyacrylamide gels. After staining with Coomassie Blue, phosphorylated bands were detected by autoradiography of the wet gel. The band representing Cx43 was diced, washed, and dehydrated after a previously published protocol (32). Gels were rehydrated in 50 mM ammonium bicarbonate, and in-gel digestion was performed with the addition of 200 ng of modified trypsin (Promega) and 2 g of nonphosphorylated recombinant Cx43 fusion protein, HisCx43CT. After elution from the gel pieces, the peptides were separated by reverse-phase HPLC, fractions were collected every minute and Cerenkov-counted, concentrated, and analyzed by MALDI-TOF mass spectrometry.
Immunofluorescence-NRK cells seeded onto glass coverslips were treated with CKI-7 for 2 or 4 h at 37°C, washed twice in PBS, and fixed in 2% formaldehyde in 0.1 M sucrose, 0.1 M cacodylate buffer, pH 7.2, for 20 min. Cells were permeabilized in 0.1% Triton X-100 in PBS for 10 min. After blocking for 1 h with 1% bovine serum albumin in PBS, permeabilized cells were incubated for 1 h with anti-Cx43 antibody (C6219, Sigma) and anti-ZO-1 antibody (33-9100, Zymed Laboratories Inc.) diluted in blocking solution. After several PBS washes, the cultures were incubated with Alexa 594 anti-rabbit secondary antibody and fluorescein isothiocyanate-conjugated donkey anti-mouse secondary antibody diluted in blocking solution for 30 -60 min followed by several washes in PBS. The coverslips were mounted onto slides with DABCO antifade medium (25 mg/ml of 1,4-diazobicyclo-(2,2,2)octane (Sigma) diluted in Spectroglycerol (Kodak) and 10% PBS, pH 8.6) and viewed with a Nikon Diaphot TE300 fluorescence microscope equipped with a 40ϫ (1.3 numerical aperture) objective and a Princeton Instruments digital camera driven by an attached PC and Metamorph imaging software.
Cell Surface Biotinylation-Cell surface biotinylation was performed according to a previously published method (8) with some modifications. Briefly, parallel 90 -95% confluent NRK cultures were treated with CKI-7 for 2, 4, or 6 h or 0.1% Me 2 SO for 6 h at 37°C. Cells were rinsed twice in ice-cold PBS and labeled twice with EZ-Link NHS-LC-Biotin (0.5 mg/ml, Pierce) in PBS at 4°C for 10 min each. Biotinylated cells were rinsed 5 times in cold PBS containing 15 mM glycine to quench the biotinylation reaction. The third wash was for 5 min at 4°C. After lysis in 0.5 ml RIPA (supplemented with 1% SDS) buffer, cell lysates were clarified by centrifugation at 4°C. The supernatant was collected and mixed with 0.5 ml of 50 mM Tris-HCl, pH 7.8, buffer containing 50 mM NaF, 500 M Na 3 VO 4 , and protease inhibitors. Biotinylated proteins were precipitated via incubation with ImmunoPure immobilized avidin cross-linked agarose (Pierce) for 30 min at 4°C. After three washes in RIPA buffer (without SDS) and one in PBS, precipitated proteins were eluted in 1ϫ sample buffer containing 5% ␤-mercaptoethanol, boiled for 3 min, and analyzed by immunoblotting using Cx43 antibodies (Cx43CT1).
Microinjection-NRK cells were grown on 35-mm dishes to 70 -80% confluency and treated with 0.1% Me 2 SO or CKI-7 for 6 h in Opti-MEM reduced serum media (Invitrogen). Me 2 SO-and CKI-7-treated donor cells were microinjected with Lucifer yellow (1 mg/ml dissolved in 0.15 M LiCl) and allowed to transfer dye for 3 min. Digital images were collected at identical camera settings on the microscope described above, and the number of recipient cells for both conditions was quantified in a blind manner. Data were analyzed statistically using linear regression modeling.

RESULTS
Cx43 Is Phosphorylated by CK1␦ in Vitro and in Cell Culture-Several studies have shown that Cx43 is phosphorylated on multiple different serine residues throughout its life cycle in homeostatic cells (12,33,34). Many kinases including CK1 are thought to remain constitutively active in cells and could phosphorylate Cx43 in the absence of exogenous kinase stimulation. To determine whether Cx43 was a substrate for CK1, we performed in vitro kinase reactions using purified CK1␦ to phosphorylate HisCx43CT (His 6 attached to amino acids 236 -382 of Cx43). After trypsin digestion, peptides were separated by reverse-phase HPLC. Fractions were collected every minute and counted by Cerenkov counting. As shown in Fig. 1, top panel, purified CK1␦ readily phosphorylated HisCx43CT, producing major phosphorylated peptides with elution times of 23, 29, and 35-36 min. MALDI-TOF mass spectrometry analysis indicated that fractions 23 and 29 contain peptides with relative molecular masses (M r ) of 1895 and 2777, respectively. These molecular masses are consistent with the predicted masses for trypsin-digested Cx43 nonphosphorylated peptides Gln-304 -Arg-319 and Met-320 -Lys-345, respectively. Calculation of the extent of phosphorylation indicated that it was between 0.1 and 0.2 mol of phosphate per mol of HisCx43CT. Therefore, we detected primarily the nonphosphorylated species by mass spectrometry. Although phosphorylation can cause slightly earlier elution of short peptides via reverse phase HPLC, our previous experience with these longer Cx43 peptides indicates that this could cause a shift of at most one fraction (31). We could not identify any Cx43 tryptic fragments in fractions 33-42. These fractions probably contain partially digested, phosphorylated HisCx43CT or autophosphorylated CK1␦.
Next, we examined cellular Cx43 phosphorylation in the presence of CK1 inhibitors. NRK cells were metabolically labeled with [ 32 P]orthophosphate in the presence (CKI-7) or absence (CON) of 100 M CKI-7. Phosphorylated Cx43 was immunoprecipitated and separated by SDS-PAGE, and the wet gel was analyzed by autoradiography and Cerenkov counting. Reduced phosphorylation was observed with Cx43 isolated from CKI-7-treated cells (Fig. 1, bottom panel, inset, CKI-7, cpm ϭ 5256) as compared with control (CON, cpm ϭ 9442). To determine the potential sites of Cx43 phosphorylation, in-gel trypsin-digested Cx43 was extracted from gel pieces and analyzed by HPLC separation using the same protocol described for in vitro phosphorylated HisCx43CT. All samples were spiked with unphosphorylated HisCx43CT to facilitate identification, because tryptic fragments of cellular Cx43 were not isolated in high enough abundance in these experiments to be detected by MALDI-TOF mass spectrometry. As shown in Fig.  1, bottom panel, phosphorylated peptides appear with elution times corresponding to 9, 14, 23, 28, 29, and 33-34 min. The fractions were concentrated and analyzed by MALDI-TOF. Similar to the observations made with CK1␦-phosphorylated HisCx43CT, we identified peptides including Gln-304 -Arg-319 (23 min, M r 1895) and Met-320 -Lys-345 (29 min, M r 2777). We could also identify Ala-371-Ile-382 (M r 1355) co-eluting with the radioactivity at 28 min and Val-347-Arg-366 (M r 2144) associated with the 33-34 peak. At least three of the six peptides showed a reduction in phosphorylation in the presence of CKI-7. Because peptides in fractions 28 and 29 were not easily separated by HPLC, it is difficult to quantify if there was any reduction in Ala-371-Ile-382 or the extent of the reduction in Met-320 -Lys-345 phosphorylation in response to CK1 inhibition. It is possible some of the decrease in the phosphorylation of Met-320 -Lys-345 (fraction 29) was hidden under the signal for Ala-371-Ile-382 (fraction 28). Consistently, in vitro CK1␦ reactions indicated that Cx43 Met-320 -Lys-345 was a more efficient substrate than Cx43 Ala-371-Ile-382 ( Fig. 1 and 2B). In agreement with TenBroek et al. (31), fractions 33 and 34 contained the highest level of cellular phosphorylation and were identified as Cx43 amino acids Val-347-Arg-366. Val-347-Arg-366, however, does not appear to be a direct substrate for CK1␦ as indicated by our in vitro experiments. Because phosphorylation of Cx43 to the P2, Triton-insoluble form is probably hierarchical, phosphorylation of serines 306, 314, 325, 328, or 330 may be required before phosphorylation at Ser-364 (see "Discussion"). We were unable to detect any Cx43 peptides in fraction 14 by mass spectrometry. Peptide co-migration studies have indicated that this fraction probably contains peptide 259 -264. 2 It is possible that earlier, less hydrophobic tryptic peptides eluting in our profile are not easily detected by mass spectrometry. To ensure we used a CKI-7 concentration specific for CK1, we immunoprecipitated Cx43 from cells metabolically labeled in the presence or absence of 50 M CKI-7. We also observed a reduction in phosphorylation signal (cpm for control ϭ 3221 versus 2452 for CKI-7). Together these data suggest that Cx43 is a substrate for CK1 in NRK cells.
Full-length Cx43 Interacts with CK1 in Homeostatic Cells-To examine if full-length Cx43 interacted with CK1 in NRK cells, Cx43 was immunoprecipitated from NRK cells using protein A-cross-linked Cx43CT2 antibody (␣Cx43). Immunoprecipitations were performed in the absence (␣Cx43) or presence (AbϩAg) of the peptide antigen (amino acids 360 -382 of Cx43) used to make Cx43CT2 antibody. Samples were analyzed by immunoblotting for CK1␦, Cx43, and ZO-1. The presence of ZO-1, a protein known to interact with Cx43 in many cell types (35,36), was assayed as a positive control. As shown in Fig. 2C, both proteins, CK1␦ and ZO-1, co-precipitated with immunoprecipitated Cx43. None of the tested proteins, Cx43, CK1, or ZO-1, was detected in immunoprecipitations performed in the presence of the peptide antigen or with protein A only. These data suggest that Cx43 is capable of interacting with cellular CK1␦ and that this interaction may occur in intact NRK cells.
Inhibition of CK1 Affects Cx43 Plasma Membrane Localization in NRK Cells-To determine the effects of CK1 inhibition on Cx43 localization, NRK cells were treated with CKI-7 for 2 or 4 h and analyzed by immunofluorescence. In Fig. 3, Cx43 staining appears as small punctate spots at the plasma membrane and cell-cell interfaces at 0 h (Fig. 3, 0 h CKI-7), which is typical of Cx43 gap junction staining. After 2 h of treatment, Cx43 staining appears more continuous in some areas (Fig. 3,  2 h CKI-7, arrows). More strikingly, after 4 h of CKI-7 treatment (Fig. 3, 4 h CKI-7), Cx43 plasma membrane staining appears to be more intense and continuous. Interestingly, cellcell contact did not appear to be necessary to observe continu-2 C. D. Cooper and P. D. Lampe, unpublished observation. Where indicated, the Cx43 peptide antigen (Ag, representing amino acids 360 -382) used to make Cx43CT2 was added to reactions with antibody (AbϩAg), or the antibody used for immunoprecipitation was omitted (Prot A) as controls.
ous Cx43 staining (apparent single membranes indicated by arrows, Fig. 3, 2 h and 4 h CKI-7). Visualization of cell borders at each time point was aided by ZO-1 staining (Fig. 3, bottom  panels). Thus, more Cx43 is apparently localized to the plasma membrane following inhibition of CK1 activity.
Triton-insoluble Cx43-P2 Decreases in CK1-inhibited NRK Cells-Because inhibition of CK1 activity resulted in increased detection of Cx43 at the plasma membrane, we reasoned that CK1 could work to regulate gap junction assembly. To examine this hypothesis, NRK cells treated with CKI-7 for 2, 4, or 6 h were either directly solubilized in sample buffer to observe CKI-7 effects on total Cx43 (Fig. 4, Whole Cell panel) or were fractionated via Triton X-100 treatment into an insoluble fraction followed by immunoblot analysis for Cx43 (Fig. 4, Tritoninsoluble (Ins) panel). Inhibition of CK1 activity slightly increased the total Cx43 present in treated cells by 4 -6 h (Whole Cell panel). In contrast, the fraction of Cx43 that was Tritoninsoluble decreased as compared with control levels, most notably after 4 and 6 h of treatment with CKI-7. In agreement with several published studies, Triton-insoluble Cx43 was predominantly of the P2 form, which has been correlated with Cx43 localization within gap junction plaques (8). Whereas inhibition of CK1 activity increased the level of plasma membrane Cx43, these data suggest the protein was not localized to gap junctional plaques.
Cell Surface-biotinylated Cx43-NP Increases in CKI-7treated NRK Cells-Studies have reported the existence of non-junctional Cx43 in the form of hemi-channels within the plasma membrane of cells, including NRK cells (37,38). Our Triton X-100 fractionation studies indicated the increased Cx43 detected at the plasma membrane was not present in gap junctions. Because previous studies have shown that the Cx43 labeled with amine reactive, cell-impermeant biotinylation reagents was non-junctional (8,39), we used cell surface biotinylation to assay for an increase in non-junctional plasma membrane Cx43 after CKI-7 treatment. NRK cell surface proteins were biotinylated with NHS-LC-Biotin after treatment with CKI-7 for 0, 2, 4, or 6 h. After cell lysis, biotinylated proteins were precipitated with avidin-agarose and analyzed by SDS-PAGE and immunoblotting for Cx43. As shown in Fig. 4 (Cell Surface panel), more Cx43 was detected in NRK plasma membranes after a CKI-7 treatment time of 2 h as compared with control (0.1% Me 2 SO) levels (which was below the detection limit of our assay). The amount of non-junctional, plasma membrane Cx43 detected continued to increase up to the final time point of 6 h. This increase in non-junctional, plasma mem-brane-associated Cx43 correlated with the decrease observed in gap junctional Cx43 by Triton X-100 extraction. In agreement with previous results, biotinylated Cx43 was primarily of the NP form (8,39). These results suggest a role for CK1 in the assembly of gap junctional structures and that non-junctional, plasma membrane connexin may act as a source of protein in the assembly process.
Inhibition of CK1␦ and -⑀ with IC261 Reduces Cx43 Phosphorylation and Accumulation of Triton-insoluble Cx43-P2-To address CKI-7 specificity and the potential role of individual CK1 family members in Cx43 phosphorylation, a series of experiments were done using IC261, an inhibitor highly specific for the ␦ and ⑀ isoforms of CK1 (40). In Fig. 5A, CK1␣ and -␦ were immunoprecipitated and autophosphorylated in the presence (ϩ) or absence (Ϫ) of IC261. Only CK1␦ showed a decrease in phosphorylation, illustrating the specificity of IC261. Next, we investigated whether CK1␦ or -⑀ inhibition reduced Cx43 phosphorylation in NRK cellular lysates. Phosphorylation of HisCx43CT was dramatically reduced in the presence of IC261, indicating that CK1␦ or -⑀ present in cellular lysates can phosphorylate the C-terminal region of Cx43 (Fig. 5B). Similarly, when NRK cells were metabolically labeled with [ 32 P]orthophosphate in the presence of 2.5 M IC261, a reduction in Cx43 phosphorylation was observed when compared with control cells (Fig. 5C, cpm control ϭ 17392 versus IC261 ϭ 14757). Last, after IC261 treatment for 2, 4, or 6 h, Triton X-100-insoluble Cx43 was isolated from cells and analyzed by immunoblotting for Cx43 and ZO-1 (loading control). IC261 treatment led to a decrease in Triton-insoluble Cx43-P2 (Fig.  5D). Taken together, these data provide additional evidence for CK1-mediated regulation of Cx43 gap junction assembly and implicate the ␦ or ⑀ isoforms.
Inhibition of CK1 Inhibits Dye Transfer in NRK Cells-To investigate the effects of CK1 inhibition on Cx43 gap junction function, we microinjected NRK cells in six individual experiments with Lucifer yellow after treatments with CKI-7 (n ϭ 95 total injections) or vehicle alone (n ϭ 94 total injections, 0.1% Me 2 SO) for 6 h. After 3 min of dye transfer, digital images were taken, and the number of recipient cells for each donor cell was quantified. The average number of recipient cells per injection in the inhibitor-treated group (11.5 cells) was less than those in the control group (15.2 cells). Furthermore, these differences were found to be statistically significant (p Ͻ 0.001) by linear regression analysis. These data indicate cells were less efficient at transferring dye when CK1 activity is inhibited. However, the 25% difference in the number of cells receiving dye is not extensive. The continued presence of gap junctions is consistent, however, with other data shown here. For example, the Triton extraction data indicate a decrease in gap junctional Cx43 after 6 h of CKI-7 and IC261 treatment, but a fraction of junctions remain assembled in a manner comparable with that observed for similar treatment times with brefeldin A (14,41). In addition, inhibition of CK1 activity by CKI-7 or IC261 might not be 100% complete, allowing some junctions to assemble even in the presence of the inhibitor. Finally, CK1-inhibited cells might adjust gap junction channel gating parameters to maximize existing channel conductivity. DISCUSSION Here we have presented data implicating a role for CK1 activity in the regulation of gap junction assembly in NRK cells. This conclusion is based on the following observations. 1) Cx43 is a substrate for purified CK1␦ in vitro. 2) Inhibition of CK1 activity results in reduced Cx43 phosphorylation in [ 32 P]orthophosphate metabolically labeled cells. 3) Immunoprecipitation reactions using Cx43 followed by SDS-PAGE and immunoblotting for CK1␦ suggest these proteins interact in communicating unstimulated NRK cells. 4) After treatment with CK1 inhibitor, we observe alterations in Cx43 localization, including increased plasma membrane associated Cx43. 5) Triton X-100 extraction and cell surface biotinylation experiments indicate that inhibition of CK1 activity leads to increased nonjunctional Cx43 in the plasma membrane. 6) Cells treated with CK1 inhibitor show a statistically significant reduction in their ability to transfer dye as measured by Lucifer yellow. Thus, we suggest that CK1, most likely the ␦ isoform, acts to regulate the timely assembly of Cx43 gap junction plaques in homeostatic cells, which could affect overall levels of communication.
Additional evidence presented here narrows down the poten-tial Cx43 sites for CK1 phosphorylation. The Cx43 CT tail contains 21 serines, some of which have already been shown to undergo phosphorylation by other kinases, including mitogenactivated protein kinase (16), protein kinase C (15), and Cdc2 kinase (11,42 Our metabolic labeling experiments also demonstrate a reduction in Cx43 phosphorylation on peptide Val-347-Arg-366 in the presence of CKI-7. Recently, a report investigating the mechanism of cAMP-up-regulated Cx43 gap junction assembly suggested that this phenomenon is dependent on phosphorylation at Ser-364 of Cx43, located within the Val-347-Arg-366 tryptic peptide (31). However, assembly under basal conditions was reported to occur in the presence and absence of Ser-364, indicating this region is not essential for basal levels of gap junction assembly, a function proposed here for CK1 activity. One explanation for the CKI-7-dependent decrease in Ser-364 phosphorylation observed in our experiments could be that Ser-364 phosphorylation requires a priming phosphorylation event mediated by CK1 phosphorylation. In other words, the lack of phosphorylation at Ser-364 might simply be a byproduct of reduced gap junction assembly and an associated reduction in P2 formation.
Observations made by Musil and Goodenough (8,9) provide the first evidence that implicated phosphorylation in Cx43processing and gap junction assembly. Studies examining communication-competent and -deficient cell lines revealed that communication-deficient cells were incapable of phosphorylating Cx43 to the P2 form, and the communication-deficient phenotype was due to decreased gap junction assembly (8,9). In addition, Cx43 Triton X-100 insolubility was acquired only after Cx43 had been phosphorylated to the P2 form and had been assembled into gap junction plaques (8). These observations suggest a direct link between Cx43 phosphorylation and gap junction assembly. One intriguing possibility is that CK1 phosphorylation of Cx43 "tags" the Cx43 connexon for assembly, and without this marker, assembly of Cx43 gap junction structures is reduced. Our data showing co-precipitation of CK1␦ with Cx43 NP are consistent with an interaction during Cx43 trafficking to the gap junction (Fig. 2C). In addition, CK1␦ has been localized to the Golgi, vesicular transporting vesicles, and plasma membrane in yeast and mammalian cells (24,(43)(44)(45)(46), consistent with the idea of CK1␦-mediated phosphorylation occurring during Cx43 connexon trafficking or after arrival at the plasma membrane. Wherever CK1 phosphorylation occurs, our cell surface biotinylation results indicate that the consequences of CK1 inhibition appear to be an accumulation of non-junctional Cx43 in the plasma membrane (Fig. 4).
Because the extent of gap junction assembly is a balance between assembly and disassembly/degradation, CK1 activity could conceivably also have effects on gap junction degradation. If CK1 phosphorylation marked channels for degradation, we might expect to see an increase in the amount of Cx43 present in gap junctions, consistent with the observed increase in plasma membrane-associated Cx43 (Fig. 3) and total Cx43 (Fig.  4, Whole Cell). However, densitometric analysis of total Cx43 in 5 experiments indicated there was not a significant change in total Cx43 in the presence of CK1 inhibitors (1.03 Ϯ 0.20 ratio at 4 h of treatment). In addition, our Triton fractionation data indicate that a large portion of the Cx43 found at the plasma membrane was not junctional in cells treated with CK1 inhibitor, so increased Cx43 degradation does not seem to explain our results. Alternatively, CK1 phosphorylation could stabilize Cx43 in gap junctions, and consequently, inhibition of CK1 activity could lead to an increase in degradation of gap junctions resulting in a loss of P2 Cx43 and a possible increase in non-junctional plasma membrane Cx43 by an uncharacterized mechanism. Interestingly, the decreasing gap junctional Cx43 observed after IC261 treatment appears to be cell confluencydependent. Sub-confluent cells did not always show a decrease in gap junction assembly, whereas confluent cells consistently did after treatment with both CK1 inhibitors. Taken together, this lends further evidence that the effect of CK1 activity is most likely at the level of gap junction assembly.
Kinase inhibition and affinity chromatography data indicate that CKI-7 is a highly specific competitive inhibitor of CK1 activity (47), but no reports indicate it is specific for any one isotype. Recent findings indicate that Cx49 is a substrate for the ␣ isoform of CK1 isolated from sheep lens but that Cx43 is not (48,49). Although we did not detect CK1␣ in Cx43 coimmunoprecipitations, one of the other isoforms, CK1␤, -␥ 1-3 , or -⑀, that we could not test due to limited specific reagents could be involved in Cx43 phosphorylation in our cell-based assays. However, CK1␦ interacted with full-length Cx43 and phosphorylated GSTCx43CT, and specific inhibition of CK1␦/⑀ activity with IC261 reproduced the observed decrease in assembled Cx43. As shown in Fig. 5A, IC261 did not affect CK1␣ activity. Taken together, these data suggest CK1␦, not CK1␣, is involved in the regulation of Cx43 gap junction assembly.
In conclusion, we have found that CK1 plays a role in the assembly of gap junctions potentially through direct phosphorylation of Cx43. Our data are the first to characterize a kinase involved in promoting gap junction assembly under basal conditions. Because Cx43 appears to be phosphorylated on at least five different serines during its life cycle, these findings represent an initial contribution to understanding the regulation of gap junction communication in homeostatic tissue.