TC-PTP directly interacts with connexin43 to regulate gap junction intercellular communication

ABSTRACT

Protein kinases have long been reported to regulate connexins; however, little is known about the involvement of phosphatases in the modulation of intercellular communication through gap junctions and the subsequent downstream effects on cellular processes. Here, we identify an interaction between the T-cell protein tyrosine phosphatase (TC-PTP, officially known as PTPN2) and the carboxyl terminus of connexin43 (Cx43, officially known as GJA1). Two cell lines, normal rat kidney (NRK) cells endogenously expressing Cx43 and an NRK-derived cell line expressing v-Src with temperature-sensitive activity, were used to demonstrate that EGF and v-Src stimulation, respectively, induced TC-PTP to colocalize with Cx43 at the plasma membrane. Cell biology experiments using phospho-specific antibodies and biophysical assays demonstrated that the interaction is direct and that TC-PTP dephosphorylates Cx43 residues Y247 and Y265, but does not affect v-Src. Transfection of TC-PTP also indirectly led to the dephosphorylation of Cx43 S368, by inactivating PKCα and PKCδ, with no effect on the phosphorylation of S279 and S282 (MAPK-dependent phosphorylation sites). Dephosphorylation maintained Cx43 gap junctions at the plaque and partially reversed the channel closure caused by v-Src-mediated phosphorylation of Cx43. Understanding dephosphorylation, along with the well-documented roles of Cx43 phosphorylation, might eventually lead to methods to modulate the regulation of gap junction channels, with potential benefits for human health.

INTRODUCTION

Gap junctions are cell-to-cell channels that enable the direct cytoplasmic exchange of ions and low-molecular-mass metabolites between adjacent cells. They provide a pathway for propagating and/or amplifying the signal transduction cascades triggered by cytokines, growth factors and other cell signaling molecules involved in growth regulation and development. Gap junction channels are formed by the apposition of connexons, each formed by six connexin proteins, from adjacent cells. Although the 21 human connexin isoforms share significant sequence similarity, the major divergence in primary structures occurs in the cytoplasmic domains (Söhl and Willecke, 2004). The 43-kDa isoform, connexin43 (Cx43, also known as GJA1) is the most ubiquitously expressed connexin isoform and the most well studied in terms of structure, function and regulation (Laird, 2006).

Cx43 channels can be regulated by a variety of molecules and physiological conditions (e.g. Ca²⁺, pH and intercellular voltage), including through phosphorylation (Solan and Lampe, 2005; Solan and Lampe, 2007). Cx43 is differentially phosphorylated at various times throughout its life cycle, which includes trafficking, assembly/disassembly, degradation and channel gating (Solan and Lampe, 2007). Unfortunately, an understanding of the mechanism(s) by which Cx43 phosphorylation alters channel function is lacking. The negative charge of the phosphate could affect the permeability of ions through the pore, alter the structure of the transmembrane α-helices to influence pore size or modify the binding affinities of molecular partners involved in Cx43 regulation. Notably, if phosphorylation modifies protein interactions to affect the kinetics of channel assembly/disassembly or degradation, cell-to-cell communication will also be altered. Numerous protein kinases have been identified that directly phosphorylate Cx43, including those that phosphorylate tyrosine residues (Laird, 2005; Moreno, 2005). Tyrosine kinases have long been reported by many investigators to disrupt gap junction intercellular communication (GJIC) – well-studied cases involve c- and v-Src-induced phosphorylation of Cx43 (Azarnia et al., 1988; Filson et al., 1990; Kanemitsu et al., 1997; Kurata and Lau, 1994; Loo et al., 1995).

Phosphorylation is a primary means of mediating the signal transduction events that control cellular processes, and a highly regulated dynamic interplay exists between protein kinases and phosphatases. Although kinases play established roles in the life cycle of a connexin, substantially less information is known about the involvement of phosphatases. Their importance was initially evident from studies using pervanadate, a tyrosine phosphatase inhibitor, which was found to decrease GJIC and enhance Cx43 tyrosine phosphorylation (Postma et al., 1998). In 2003, two potential Cx43 tyrosine phosphatases were identified – vascular protein tyrosine phosphatase 1 (identified by mass spectroscopy) and receptor protein tyrosine phosphatase μ (identified by immunoprecipitation); however, direct interactions were not verified, and direct roles in cell-to-cell communication were not established (Giepmans et al., 2003; Singh and Lampe, 2003). In this study, we identified Cx43 as a novel substrate of the T-cell protein tyrosine phosphatase (TC-PTP, also known as PTPN2).

TC-PTP was cloned from a peripheral human T-cell cDNA library and, despite its name, is ubiquitously expressed in all tissues and at all stages of mammalian development (Cool et al., 1989). Alternative splicing in the C-terminus gives rise to a 48-kDa form (the minor form), localized at the endoplasmic reticulum, and a 45-kDa form (the major form) that is targeted to the nucleus (Cool et al., 1989). Mice lacking the gene encoding TC-PTP die within 5 weeks of birth from defects in hematopoiesis and immune function (You-Ten et al., 1997). Results obtained from the use of heterozygous TC-PTP^+/− mice implicated TC-PTP as an important regulator of inflammatory cytokine signaling and in the pathophysiology associated with inflammatory bowel disease (Hassan et al., 2010). The identities of several substrates, such as the insulin and epidermal growth factor (EGF) receptors (Galic et al., 2003; Tiganis et al., 1998), as well as cytoplasmic JAK and STAT proteins (Lu et al., 2007; Simoncic et al., 2002; ten Hoeve et al., 2002), suggest that TC-PTP might play a role in oncogenesis. We provide evidence of a relationship between Cx43 tyrosine phosphorylation by v-Src and dephosphorylation by TC-PTP. TC-PTP increased the stability of gap junctions at the plasma membrane and partially reversed the closure of Cx43 gap junction channels caused by v-Src phosphorylation.

RESULTS

TC-PTP directly interacts with and dephosphorylates the Cx43 carboxyl-terminal domain

The identification of proteins that interact with connexins has been invaluable in understanding the regulation of GJIC. Based upon the well-characterized ability of Src-mediated phosphorylation to inhibit GJIC, we sought to identify whether a tyrosine phosphatase could reverse this effect. Interestingly, a correlation exists between the activation of Src by stress conditions (e.g. oxidation stress) (Zhougang and Schnellmann, 2004) or mitogens (e.g. insulin and EGF) (Goi et al., 2000; Li et al., 2010; Rosenzweig et al., 2004) and the redistribution of the tyrosine phosphatase TC-PTP to the plasma membrane (Fukushima et al., 2010; Lam et al., 2001; Tiganis et al., 1998). This redistribution takes the form of TC-PTP diffusion from the nucleus and/or prevention of the (re)entry of the protein into the nucleus. To identify whether a direct interaction exists between TC-PTP and the Cx43 carboxyl-terminal (Cx43CT) domain, nuclear magnetic resonance (NMR) titration experiments were performed with purified TC-PTP catalytic domain (TC-PTP_1–314) and Cx43CT. NMR is an ideal method to detect protein–protein interactions because chemical shifts are sensitive to the environment and even small changes in structure and/or dynamics can influence the chemical shift of an amino acid. Different concentrations of unlabeled TC-PTP_1–314 were titrated into ¹⁵N-labeled Cx43CT (residues 236–382) and ¹⁵N-heteronuclear single quantum coherence (HSQC) spectra were acquired. The ¹⁵N-HSQC is a two-dimensional experiment in which each residue (except proline) gives one chemical shift that corresponds to the amide group. Therefore, it is necessary to know the Cx43CT resonance assignment in order to determine which residues are affected by the interaction. Our lab has previously published the Cx43CT ¹⁵N-HSQC assignment (Grosely et al., 2013); however, the spectrum was previously collected in phosphate-buffered saline (PBS), a buffer that caused TC-PTP_1–314 to precipitate at elevated concentrations, unlike Tris-NaCl buffer. To assign the Cx43CT domain in Tris-NaCl buffer at pH 7.5, Cx43CT in PBS was initially titrated from pH 5.8 to pH 7.5, and then PBS was titrated to Tris-NaCl buffer at pH 7.5. The optimal buffer to visualize the peaks and maintain the solubility of TC-PTP_1–314 was found to be a mixture of 25% PBS and 75% Tris-NaCl buffer at pH 7.5 (a sampling of the reassignments is shown in Fig. 1A). Titration of TC-PTP_1–314 caused a subset of Cx43CT residues to broaden beyond detection. These are highlighted on the Cx43CT sequence shown in Fig. 1B. The strongly affected Cx43CT residues G261–T275 included Y265 and Y267. Also affected were areas that included Y247, Y286, Y301 and Y313. The decrease in signal intensity caused by increasing TC-PTP_1–314 concentrations was fit according to the nonlinear least-square method. The binding affinity (K_D) was determined to be 350±80 µM (±s.e.m.). A glutathione S-transferase (GST) pulldown experiment was performed to validate the NMR results. Purified GST or GST–TC-PTP_1–314 bound to glutathione–agarose were incubated with purified Cx43CT, and immunoblotting using an antibody against Cx43CT identified that GST–TC-PTP_1–314, but not GST, directly interacted with the Cx43CT (Fig. 1C).

Fig. 1.

Characterizing the Cx43CT residues affected by the direct interaction with TC-PTP. (A) Overlaid ¹⁵N-HSQC spectra of the ¹⁵N-Cx43CT (residues 236–382; 30 µM) in the presence of different concentrations (60 µM–540 µM) of unlabeled TC-PTP_1–314. The cross-peak color changes according to the concentration ratio. A subset of the affected peaks is labeled. ppm, parts per million. (B) The amino acid sequence of Cx43CT. Residues strongly affected by the addition of TC-PTP_1–314 are highlighted in green (residues broadened beyond detection at 1∶12 molar ratio) and those that are less affected are highlighted in yellow (residues broadened beyond detection at 1∶18 molar ratio). In red are the Cx43 residues Y247 and Y265 that are phosphorylated by Src. The remaining four tyrosine residues have been underlined. (C) Purified GST or GST–TC-PTP_1–314 on glutathione–agarose beads were incubated with (+) or without (−) purified Cx43CT_236–382. After washing, the anti-GST antibody (upper panel) detected both GST alone (26 kDa) and GST–TC-PTP_1–314 (58 kDa); however, the anti-Cx43CT antibody (lower panel) only detected Cx43CT when incubated with GST–TC-PTP_1–314.

Of the six Cx43CT tyrosine residues affected by TC-PTP_1–314, Y265 and Y247 are sites that are known to be phosphorylated by c-Src and v-Src, leading to the closure of Cx43 gap junction channels (Kanemitsu et al., 1997; Lin et al., 2001). To identify whether TC-PTP dephosphorylates these residues, an in vitro phosphatase assay was conducted, in which peptides containing phosphorylated Y247 (pY247) or pY265 were incubated with TC-PTP_1–314. By following the protocol for the Malachite Green assay (Millipore), we observed an increase in the amount of inorganic phosphate production, indicating that TC-PTP dephosphorylates Cx43 on pY247 and pY265 (Fig. 2). The rate of dephosphorylation was different at the two sites, with pY265 being more efficiently dephosphorylated by TC-PTP than was pY247. This observation is consistent with the kinetic rate constant data (K_m and k_cat), which indicate that TC-PTP is more efficient in dephosphorylating pY265 than in dephosphorylating pY247 (Table 1).

Fig. 2.

TC-PTP dephosphorylates Cx43 residues pY247 and pY265 in vitro. Plot of the Malachite Green assay shows the timecourse of dephosphorylation of Cx43 phospho-peptides (600 nM) containing pY247 or pY265 by the TC-PTP catalytic domain (0.6 nM). Data were recorded based on readings at OD₆₅₀. Pos. Ctrl, positive control; Neg. Ctrl, negative control. Data show the mean±s.e.m. (three independent experiments).

Table 1. Kinetic constants for dephosphorylation of pY265 and pY247 by TC-PTP_1–314.

^*All data were collected at pH 7.5, 25°C. ^#Notably, the kinetic constants determined for the positive control (DADEpYL) by TC-PTP_1–314 are consistent with those reported for the same peptide with the PTP1B catalytic domain, the sequence of which is very similar to that of TC-PTP_1–314 (Peters et al., 2000).

TC-PTP dephosphorylates Cx43 in NRK cells

To determine whether TC-PTP-mediated dephosphorylation of pY265 and pY247 occurs in cells, we initially tested whether endogenous Cx43 and TC-PTP colocalize in normal rat kidney epithelial (NRK) cells. Immunostaining in the absence of EGF showed that TC-PTP localizes in the nucleus and Cx43 localizes at the plasma membrane (Fig. 3A, −EGF). Within 1 h after EGF treatment, TC-PTP was recruited to the plasma membrane and partially colocalized with Cx43 (Fig. 3A, +EGF). Co-immunoprecipitation (co-IP) using a TC-PTP-specific antibody that interacts with the C-terminal domain (anti-TC-PTP-CT), followed by western blot analysis, demonstrated that TC-PTP and Cx43 are present in the same complex (Fig. 3B). Numerous studies have shown that differential phosphorylation of Cx43 results in the existence of multiple electrophoretic isoforms – a fast-migrating isoform (P0) and multiple slower-migrating isoforms (P1 and P2) (Crow et al., 1990; Matesic et al., 1994; Solan and Lampe, 2005). TC-PTP has a preference for the P2 isoform in NRK cells.

Fig. 3.

EGF induces TC-PTP to colocalize with Cx43 in NRK cells. (A) Cellular localization of endogenous Cx43 and TC-PTP in NRK cells with or without human recombinant EGF (50 ng/ml; 1 h, Peprotech, Rocky Hill, NJ), as visualized by using immunofluorescence (green, Cx43; blue, DAPI-stained DNA; red, TC-PTP). White arrows show colocalization of Cx43 and TC-PTP (yellow). White dashed boxes enclose the area enlarged to the right. Scale bars: 20 µm. Colocalization of Cx43 and TC-PTP was analyzed based on ten images from three independent experiments. Manders method was used to measure green signal (Cx43) coincident with red signal (TC-PTP) over the total intensity of green signal. Data show the mean±s.e.m. (B) Lysates of NRK cells were immunoprecipitated (IP) with anti-TC-PTP-NT, anti-TC-PTP-CT or IgG and then blotted (IB) for Cx43 and TC-PTP. The input is NRK cell lysate. The Cx43 mobility shifts (P0, P1 and P2) are labeled.

TC-PTP can interact with protein partners through both the N-terminal catalytic domain and the C-terminal inhibitory and nuclear-targeting domain (Iversen et al., 2002). Dephosphorylation of Cx43 would therefore depend on the interaction with the TC-PTP catalytic domain. Co-IP using an anti-TC-CT antibody pulled down Cx43 (Fig. 3B). However, an antibody against the TC-PTP N-terminal catalytic domain (anti-TC-PTP-NT) was unable to pull down Cx43, suggesting that the epitope was blocked owing to the interaction of the catalytic domain with Cx43 – results consistent with the in vitro binding experiments. Based on these observations, the catalytic domain alone (TC-PTP_1–314, a fragment that lacks the nuclear localization domain and is therefore localized in the cytoplasm) was transfected into NRK cells to test for TC-PTP-mediated dephosphorylation of Cx43. Immunostaining demonstrates that TC-PTP_1–314 colocalizes with Cx43 on the plasma membrane with or without EGF treatment (Fig. 4A). Next, Cx43 phospho-specific antibodies were used to test whether TC-PTP could decrease Cx43 tyrosine phosphorylation levels in cells (Fig. 4B). The amount of Cx43 pY265 and pY247 was decreased in the TC-PTP_1–314-transfected group compared with the pcDNA 3.1 vector transfection group. Of note, the basal level of Cx43 tyrosine phosphorylation observed in the NRK cells (Fig. 4B, −TC-PTP_1–314) is significantly greater than what has been observed in other cell lines (Kanemitsu et al., 1997; Lampe et al., 1998; Lau et al., 1992; Lidington et al., 2002). Although, to the best of our knowledge, the basal level of Cx43 tyrosine phosphorylation in NRK cells has not been previously reported, a basal level of Cx43 tyrosine phosphorylation has been observed in a derivative of the NRK cell line [LA-25 cells (Solan and Lampe, 2014), discussed further below]. The differences between cell lines might be the result of higher sensitivity of the specific antibodies used in this study and in Solan and Lampe (2014) and/or TC-PTP might be somewhat less active in NRK (and LA-25) cells under the given conditions. Taken together, these studies indicate that TC-PTP and Cx43 exist in the same complex in cells and that TC-PTP can cause a decrease in the levels of Cx43 tyrosine phosphorylation.

Fig. 4.

TC-PTP causes Cx43 dephosphorylation in NRK cells. (A) Immunofluorescence of NRK cells transfected with the catalytically active cytoplasmic TC-PTP domain (TC-PTP_1–314) with or without EGF (50 ng/ml; 1 h). Green, Cx43; blue, DAPI-stained DNA; red, TC-PTP. White arrows show colocalization of Cx43 and TC-PTP (yellow). White dashed boxes enclose the area enlarged to the right. Scale bars: 20 µm. Colocalization of Cx43 and TC-PTP was analyzed based on eight images from three independent experiments. Data show the mean±s.e.m. (B) Western blot using Cx43 Y247 and Y265 phospho-specific antibodies and antibodies against Cx43 and TC-PTP in NRK cells. The relative protein levels were quantified by analyzing scanned blots using ImageJ software. The data are representative of three independent experiments. NT, N-terminus; wt, wild-type.

TC-PTP dephosphorylates Cx43 residues pY247 and pY265 in LA-25 cells

NRK cells containing a temperature-sensitive v-Src (known as LA-25 cells) are commonly used in the gap junction field to characterize Cx43 regulation by v-Src (Solan and Lampe, 2008; Zhou et al., 1999). v-Src is active in this cell line at the permissive temperature (35°C) and not at the non-permissive temperature (40°C). It is worth mentioning that temperature alone does not affect gap junction communication in NRK cells (Atkinson et al., 1981). Here, the LA-25 cells were used to characterize the interplay between TC-PTP and v-Src in regulating the tyrosine phosphorylation levels of Cx43. To begin with, immunostaining data indicated that active v-Src at 35°C colocalized with Cx43 at the plasma membrane (Fig. 5A). Immunostaining data then confirmed that, after activation of v-Src, endogenous TC-PTP colocalized with Cx43 at the plasma membrane (Fig. 5B). These data suggest that TC-PTP could mitigate the effect of active v-Src in order to maintain cell-to-cell communication.

Fig. 5.

v-Src induces TC-PTP to colocalize with Cx43 in LA-25 cells. Immunostaining studies show that (A) active Src (Src pY416, red) and (B) TC-PTP (red) colocalize with Cx43 at the permissive temperature (35°C) in LA-25 cells. Green, Cx43; blue, DAPI-stained DNA. White arrows show colocalization of Cx43 and active Src or TC-PTP (yellow). White dashed boxes enclose the area enlarged to the right. Scale bars: 50 µm (A), 20 µm (B). Colocalization of Cx43 and active Src or TC-PTP was analyzed based on 12 images from three independent experiments. The data show the mean±s.e.m.

To test this possibility, tyrosine phosphorylation levels were evaluated in the LA-25 cells using phospho-specific antibodies against Cx43 pY265 and pY247 (Fig. 6). Cells cultured at the permissive temperature of 35°C showed increased phosphorylation of both Y265 and Y247 compared with that of cells cultured at 40°C. However, if transfected with TC-PTP_1–314 prior to v-Src activation, the phosphorylation levels of Y265 and Y247 decreased significantly. Additionally, at 35°C, TC-PTP_1–314 did not dephosphorylate v-Src (Src pY416), which supports the notion that the decrease in Cx43 pY247 and pY265 in the TC-PTP_1–314-transfected group is caused by TC-PTP and not by decreased kinase activity of v-Src. Because MAPK and PKC have been reported to phosphorylate Cx43, leading to the downregulation of Cx43-dependent GJIC in response to active v-Src in LA-25 cells [as well as in other cell lines (Mitra et al., 2012; Solan and Lampe, 2008)], we investigated whether TC-PTP indirectly affected serine phosphorylation. Cx43 serine phosphorylation levels were evaluated using phospho-serine-specific antibodies against pS279/282 (MAPK-dependent phosphorylation sites) and pS368 (PKC-dependent phosphorylation site). pS279/282 and pS368 levels were increased at 35°C as expected; interestingly, expression of TC-PTP_1–314 indirectly led to decreased levels of pS368 but not pS279/282 at 35°C. S368 is known to be phosphorylated by different PKC isoforms (α, δ and ε) (Cone et al., 2014; Mitra et al., 2012), and PKCα and PKCδ, in particular, can be upregulated by v-Src expression (Zang et al., 1995). To address whether TC-PTP affects the activation of PKC, we initially identified that LA-25 cells express PKCα and PKCδ (Fig. 6B) but not PKCε (data not shown). Next, a well-established method of measuring membrane-associated PKC isoforms was used, because PKC isoforms translocate from the cytosol to the cell membrane when activated (Borner et al., 1992; Ha and Exton, 1993; Saxon et al., 1994). The level of activated PKCα and PKCδ was increased at 35°C compared with that observed at 40°C (from 16.3±2% to 25.7±3% for PKCα; from 13.0±2% to 36.7±0.3% for PKCδ; ±s.e.m.). Transfection of TC-PTP_1–314 caused a decrease in the level of both membrane-associated PKCα (14.3±4%) and PKCδ (16.7±5%), indicating that TC-PTP can inhibit the activation of PKCα and PKCδ by v-Src (Fig. 6B).

Fig. 6.

TC-PTP dephosphorylates Cx43 residues pY247 and pY265 in LA-25 cells. (A) Western blot of TC-PTP [both wild-type (wt) and transfected TC-PTP_1–314 (NT) forms], active Src (Src pY416), total Src, pY265, pY247, pS279/282, pS368 and total Cx43 (as determined by blotting with Cx43NT1) levels from LA-25 cells at 40°C and 35°C (12 h) with or without transfection of TC-PTP_1–314. Relative protein levels were quantified by analyzing scanned blots using ImageJ software, with normalization of protein expression to the −TC-PTP_1–314 lane at 35°C lane (value set arbitrarily as 1). The data are representative of three independent experiments. (B) Activation of PKC isoforms α and δ in LA-25 cells at 40°C and 35°C with or without transfection of TC-PTP_1–314. PKC isoforms α and δ in cytosolic (C) and membrane (M) fractions were analyzed by western blotting. Na⁺/K⁺ ATPase was blotted as a marker of the membrane fraction; GAPDH was blotted as a marker of the cytosolic fraction. Relative protein levels were quantified by analyzing scanned blots from three independent experiments using ImageJ software. Data show the mean±s.e.m.; *P<0.05. (C) LA-25 cells treated with scramble RNA or TC-PTP siRNA were incubated at 35°C for 12 h. Cx43 phosphorylation levels were analyzed by western blotting, and relative protein levels were quantified using ImageJ, and normalized to the expression in the scramble-RNA-treated sample (arbitrarily set as 1).

To determine the specificity of TC-PTP in decreasing the level of Cx43 tyrosine phosphorylation, siRNA against TC-PTP was used to knock down endogenous TC-PTP in LA-25 cells. The colocalization of TC-PTP with Cx43 was investigated at 35°C; at this temperature, endogenous TC-PTP was found to colocalize with Cx43 at the cell membrane (Fig. 5B). In the TC-PTP siRNA group, both pY265 and pY247 increased twofold (even with a decrease in the total levels of Cx43) compared with their levels in the scramble siRNA group. The decrease in Cx43 expression correlated with the decrease observed for pS279/282, indicating that downregulation of TC-PTP does not affect MAPK-mediated phosphorylation. By contrast, the relative decrease in Cx43 was greater than that of pS368, indicating an upregulation of PKC activity in the absence of TC-PTP. This result is consistent with the downregulation of PKCα and PKCδ following transfection of TC-PTP_1–314 (Fig. 6B). Next, we addressed the functional significance of Cx43 dephosphorylation by TC-PTP.

TC-PTP increases GJIC and stability of the gap junction plaque

To determine whether TC-PTP-mediated dephosphorylation of Cx43 affects cell-to-cell communication, the junctional transfer of the fluorescent tracer Lucifer Yellow (molecular mass 443 Da) was measured in a scrape-loading assay using LA-25 cells (Fig. 7A). Gap junctions were functional at 40°C, and the number of fluorescent cells was significantly reduced at 35°C from 115.2±5 to 8.6±3 (Fig. 7B ±s.e.m.). By contrast, transfection of TC-PTP_1–314 increased the number of fluorescent cells at 35°C, albeit not to the levels observed without active v-Src [from 8.6±3 to 56.4±3 (with active v-Src) versus 115.2±5 (without active v-Src)]. These findings indicate that the inhibition of Cx43 GJIC caused by v-Src can be partially rescued by TC-PTP. A plausible explanation for the fact that the rescue of GJIC was only partial is that the levels of MAPK-mediated Cx43 phosphorylation were unaffected by TC-PTP_1-314 (Fig. 6). Using the MAPK-specific inhibitor U0126, the number of fluorescent (non-TC-PTP_1–314-transfected) cells at 35°C displayed a greater increase than that observed with TC-PTP_1–314 transfection alone (88±5), but was still reduced compared with the number observed at 40°C. Finally, we used the PKC-specific inhibitor BIM to test the contribution of PKC phosphorylation to channel closure. BIM treatment also resulted in an increased number of fluorescent cells (43±5) compared with that of the control (DMSO), but not compared with the number observed following TC-PTP_1–314 transfection or treatment with the MAPK-specific inhibitor U0126.

Fig. 7.

Scrape-loading assay shows that TC-PTP increases GJIC in LA-25 cells. (A) Cells grown on glass coverslips were scrape-loaded with Lucifer Yellow (LY, green) and Texas-Red-conjugated dextran (red). The cell nucleus was stained with DAPI to show cell confluence. Scale bars: 100 µm. (B) Quantification of fluorescent cells shows the effect of TC-PTP_1–314 and serine kinase inhibitors on gap junctional permeability. The number of fluorescent cells at 40°C is significantly higher than at 35°C (P<0.0001). At 35°C, transfection of TC-PTP_1–314 or addition of the MAPK inhibitor U0126 (50 µM) or PKC inhibitor BIM (0.1 µM) significantly increased the number of fluorescent cells compared with that observed for the control (DMSO) at 35°C (P<0.0001). Data show the mean±s.e.m.; three independent experiments.

Cx43 gap junction channels are localized in detergent-insoluble junctional plaques (Musil and Goodenough, 1991; Sirnes et al., 2008). To test the possibility that the rescue of gap junctional coupling by TC-PTP was caused by an increase in the amount of Cx43 at junctional plaques, we assessed solubility in Triton X-100 detergent. Detergent extraction followed by in situ fluorescence and western blotting showed that active v-Src in LA-25 cells at 35°C decreased the detergent-insoluble fraction of Cx43 (Fig. 8A,B). Quantification of the data revealed that the insoluble to insoluble+soluble ratio [I/(S+I)] was 79.6±8% at 40°C (Fig. 8C). However, the number of assembled gap junctions significantly decreased at 35°C to 45.1±10%. The effect of v-Src was substantially blocked by the transfection of TC-PTP_1–314. TC-PTP_1–314 increased the amount of Cx43 that was resistant to Triton X-100 [I/(S+I) = 72.9±8%] at 35°C, suggesting that TC-PTP contributes to the stability of Cx43 at the junctional plaque.

Fig. 8.

Triton X-100 solubility assay shows that TC-PTP stabilizes Cx43 at the gap junction plaque in LA-25 cells. (A) Cx43 was extracted in situ with 1% Triton X-100 and cells were subsequently immunostained for Cx43 (green). Blue, DNA stained with DAPI. Scale bars: 10 µm. (B) Equal amounts of the total protein fraction (T), the Triton X-100 soluble fraction (S) and the insoluble fraction (I) were subjected to SDS-PAGE and blotted with antibody against Cx43. (C) Protein levels were quantified to determine the insoluble to insoluble+soluble ratio [I/(I+S)]. I/(I+S) at 40°C is significantly higher than that at 35°C (P<0.001). At 35°C, I/(I+S) in the TC-PTP_1–314-transfected group is significantly different from the empty-vector-transfected control (P<0.01). Data show the mean±s.e.m.

DISCUSSION

Here, we identified that TC-PTP is the first tyrosine phosphatase, to the best of our knowledge, that directly interacts with Cx43 to modulate GJIC. TC-PTP-mediated dephosphorylation of the Cx43 tyrosine residues Y265 and Y247, as well as indirect serine dephosphorylation at S368, increased both GJIC and the stability of gap junction channels at the plaque. Our studies are consistent with prior observations that TC-PTP is a negative regulator of cytokine signaling, as EGF stimulation of NRK cells caused TC-PTP to colocalize with Cx43 at the plasma membrane. In addition, v-Src expression alone was able to mimic the effects of cytokine stimulation. How cytokines and/or v-Src induce TC-PTP redistribution is not clear. TC-PTP has been reported to exit from the nucleus by passive diffusion after cellular stress, although it is also possible that cytoplasmic accumulation of TC-PTP occurs through the inhibition of nuclear import (Lam et al., 2001).

Closure of Cx43 gap junction channels by v-Src was first observed over 30 years ago (Atkinson et al., 1981); however, the specific sites of Cx43 phosphorylation required for this process are still a matter of debate. Several studies implicate the activation of v-Src in both the direct phosphorylation of Cx43 tyrosines (Y247 and Y265) and the indirect phosphorylation of serines by MAPK (S279/S282) and PKC (S368) (Mitra et al., 2012; Solan and Lampe, 2008). We also show that v-Src increases the level of phosphorylation of Cx43 Y265, Y247, S279/282 and S368. Conversely, TC-PTP decreased the phosphorylation level of Y265 (direct), Y247 (direct) and S368 (indirect), but not that of S279/282. The indirect decrease in the level of pS368 was caused by the inactivation of PKCα and PKCδ. The fact that pS279/p282 levels were not affected is consistent with a previous study in which TC-PTP had no effect on MAPK signaling after EGF-induced activation (van Vliet et al., 2005). Of note, one mechanism by which EGF upregulates MAPK signaling is through c-Src activation (Faivre and Lange, 2007; Zhang et al., 2012). The use of kinase inhibitors identified that v-Src, MAPK and PKC are all involved in gap junction closure, although they decreased the level of communication in different ways. MAPK phosphorylation had the greatest effect on channel closure, and the observation that TC-PTP did not affect the phosphorylation level of S279/282 explains why TC-PTP only partially reversed the gap junction closure caused by v-Src.

Several hypotheses, which are not mutually exclusive, could describe the mechanism(s) by which phosphorylation regulates GJIC. First, the negative charge of the phosphate could affect the permeability of ions through the pore by electrostatic attraction or repulsion. In this study, the negatively charged Lucifer Yellow was used as probe for GJIC. The v-Src-dependent placement of at least 60 negative charges (five phosphorylation sites: two Src dependent, two MAPK dependent and one PKC dependent; phosphate charge of −2) within a connexon (six connexins) could contribute to the reduction in Lucifer Yellow transport. Second, phosphorylation can directly (and possibly indirectly, see Grosely et al., 2013) alter the binding affinities of proteins involved in Cx43 regulation. For example, the aromatic ring of Y247 is directly involved in microtubule binding, and Y247 phosphorylation inhibits this interaction (Saidi Brikci-Nigassa et al., 2012). Because microtubules play important roles in rapid connexin delivery to gap junction plaques (Shaw et al., 2007), pY247 would be detrimental to Cx43 intercellular communication. This is consistent with our results showing that dephosphorylation of pY247 by TC-PTP correlates with an increase in the number of Cx43-containing gap junctions at the plaque. Another recent study in HeLa cells identified two tyrosine-based AP-2-binding motifs in the Cx43 C-terminus (Y²⁶⁵AYF and Y²⁸⁶KLV), one of which is a Src-dependent phosphorylation site, that cooperate to mediate gap junction endocytosis (Fong et al., 2013). Mutation of the Y²⁸⁶KLV motif alone in a yeast-two-hybrid screen (no post-translational modification) using the Cx43CT as bait against the AP-2 μ2 subunit inhibited this interaction (Johnson et al., 2013). Taken together, these studies suggest that the phosphorylation of Y265 enables the interaction with AP-2 to mediate Cx43 endocytosis. This possibility is consistent with our finding that TC-PTP-mediated dephosphorylation of Y265 also correlates with an increase in the number of Cx43 gap junction channels at the plaque. Finally, phosphorylation of Cx43CT could alter the structure of the transmembrane α-helices to influence pore size and/or drive gating by a ‘ball and chain’ mechanism, similar to what has been shown for the pH gating of Cx43 (Ek-Vitorin et al., 1996; Morley et al., 1996).

Phosphorylation is a primary means of mediating signal transduction events that control cellular processes, and it is controlled by a highly regulated dynamic interplay between protein kinases and phosphatases. Many kinases are known to phosphorylate Cx43, with the predominant effect being a decrease in GJIC. Although kinases are known to play established roles in the life cycle of a connexin, substantially less is known about the involvement of phosphatases. The identification of a new tyrosine phosphatase, TC-PTP, which removes Src-mediated phosphorylation of Cx43, presents an opportunity to characterize the timeline of gap junction channel opening, thus furthering our understanding of gap junctions beyond channel closure, the process that has been under investigation in other studies. An understanding of both phosphorylation and dephosphorylation is necessary to provide a better foundation to allow the targeting of gap junction channels to benefit human health. For example, treatment of a mouse model of myocardial infarction with a c-Src inhibitor significantly increased Cx43 expression in scar border and distal ventricle, leading to improved conduction velocity and lower arrhythmic inducibility (Rutledge et al., 2014). Our data demonstrating that TC-PTP dephosphorylates Cx43 and increases the level of Cx43 on the cell membrane supports their conclusion, from a cellular perspective. Future research using an animal model to study the relevance of Cx43 dephosphorylation by TC-PTP might provide new avenues for disease therapy.

MATERIALS AND METHODS

Expression and purification of recombinant GST-tagged proteins

Rat Cx43CT_236–382 was expressed and purified as described previously (Duffy et al., 2002). TC-PTP_1–314 was cloned into the pGEX-6p-2 vector and overexpressed in the BL21 (DE3)-derived Rosetta strain (Novagen). A 20 ml overnight culture was diluted 1∶50 and grown at 37°C to an A₆₀₀ of 0.6, then induced with 0.5 mM isopropyl β-d-thiogalactoside for 20 h at 19°C. A total of 1 liter of cultured cells was lysed by using an EmulsiFlex-C3 in 25 ml of the following buffer: 25 mM Tris-HCl pH 7.5, 5 mM DTT, 0.1% Brij 35 and one protease inhibitor cocktail tablet (Roche). The cell lysate was incubated with glutathione–Sepharose beads (Genescript) for 2 h at 4°C, then the beads were incubated with Tris-HCl buffer (50 mM Tris-HCl, 2 mM ATP, 10 mM MgSO₄) for 10 min at 37°C to remove a contaminating chaperone protein. The beads were washed in Tris-HCl buffers containing high salt (300 mM NaCl) or detergent (1% Brij 35) and were then washed four times with 25 mM Tris-HCl (pH 7.5) and 5 mM DTT. Turbo 3C protease (Accelagen) was used to cleave TC-PTP_1–314 from the glutathione S-transferase tag at 4°C for 2 days.

NMR

All NMR data were acquired using a 600 MHz Varian INOVA NMR Spectrometer fitted with a cryo-probe at the NMR Facility of the University of Nebraska Medical Center. Gradient-enhanced two-dimensional ¹⁵N-HSQC experiments were performed to detect backbone amide bond resonances from the ¹⁵N-Cx43CT (30 µM) in the absence and presence of different concentrations of unlabeled TC-PTP_1–314. NMR spectra were processed using NMRPipe (Delaglio et al., 1995) and analyzed with NMRView (Johnson and Blevins, 1994). Binding affinities from the ¹⁵N-HSQC titration experiments were calculated by using Graphpad Prism 5 (GraphPad Software).

Cell culture

NRK and LA-25 cells (NRK cells expressing temperature-sensitive v-Src) were generous gifts from Paul Lampe (Fred Hutchinson Cancer Research Center). Both cell lines were grown in Dulbecco's modified Eagle's medium (DMEM; Hyclone, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS; Hyclone, Thermo Fisher Scientific) and antibiotics, under an atmosphere of humidified 5% CO₂. NRK cells were serum staved for 4 h in DMEM with 0.1% FBS before EGF treatment.

Antibodies and immunostaining

The following antibodies were used in this study: Cx43 monoclonal antibodies against amino acids 360–382 [Cx43CT1 and Cx43IF1, as described previously (Cooper and Lampe, 2002; Lampe et al., 2006)], rabbit anti-pY247, rabbit anti-pY265 and rabbit anti-pS279/282 [as described previously (Solan and Lampe, 2008), all generous gifts from Paul Lampe, Fred Hutchinson Cancer Research Center], Cx43 monoclonal antibody against residues 1–20 (Cx43NT1, Fred Hutchinson Cancer Research Center Hybridoma Development Facility), rabbit anti-pS368 (Millipore), TC-PTP antibodies against the N-terminus and C-terminus (SAB4200495 and SAB4200249, Sigma-Aldrich), non-specific phospho-tyrosine antibody (Abcam) and antibodies against total and active Src (pY416, Millipore).

Cells were immunostained as described previously (Mehta et al., 1991). Briefly, cells grown on coverslips to ∼60% confluence were fixed with 2% paraformaldehyde for 15 min. Cells were blocked for 30 min at room temperature in MPS buffer (1×PBS, 1% goat serum) containing 0.2% Triton X-100 for permeabilization. Next, cells were immunostained with the appropriate primary antibodies at room temperature for 1 h, followed by several washes in PBS containing 0.5% Tween. Secondary antibodies (Alexa-Fluor-594-conjugated goat anti-rabbit-IgG and/or Alexa-Fluor-488-conjugated goat anti-mouse-IgG) were applied for 1 h at room temperature. Images of immunostained cells were acquired with a Zeiss 510 Meta Confocal Laser Scanning Microscope using a 63×/1.4 NA objective with the appropriate filters. Colocalization was quantified using the Manders method in the ImageJ plugin JACoP, by which the Cx43 signal (green) coincident with a signal in the red channel over the total intensity of Cx43 was measured (Bolte and Cordelières, 2006).

Co-IP and western blotting

NRK cells were lysed in complete lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.5% sodium deoxycholate, 1% Triton X-100, 5 mM NaF and half a tablet of Roche complete protease inhibitor in 20 ml buffer), maintained on ice for 30 min, precleared with protein-G beads for 30 min at 4°C and then spun at 12,000 rpm 11,752 g for 15 min. Total protein was assessed using the bicinchoninic acid (BCA) protein assay kit (Pierce). A total of 2 mg of lysate was incubated with 2 µg of anti-TC-PTP-NT, anti-TC-PTP-CT or rabbit IgG (for 4 h at 4°C) and then incubated with 100 µl of protein-G–Sepharose (GE Healthcare) (overnight at 4°C). The Sepharose was washed four times with cold lysis buffer and the co-IP was analyzed by SDS-PAGE and western blotting. The anti-Cx43 antibody Cx43NT1 was used to detect Cx43 that co-immunoprecipitated with TC-PTP.

Protein levels were detected by SDS-PAGE and western blotting. One exception for detecting tyrosine phosphorylation was that 5 mM Na₃VO₄ was added to the blocking buffer and primer antibody buffer in order to minimize the loss of phosphorylation. Western blots were scanned and the data were quantified using ImageJ as described previously (Schneider et al., 2012).

GST pull downs

The GST pull-down assay was modified from a version published previously (Leykauf et al., 2006). Briefly, purified GST-tagged TC-PTP_1–314 and GST control protein were bound to glutathione–Sepharose beads in buffer 1 containing 25 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.1% Triton X-100, 1 mM DTT, 0.1 mM EDTA and half a tablet of complete protease inhibitor in 20 ml buffer. A total of 5 mg of purified Cx43CT was incubated with GST-tagged TC-PTP_1–314 or the GST control overnight at 4°C on a rotating wheel. Beads were washed five times with buffer 1 and bound proteins were eluted with SDS-PAGE sample buffer and were used for SDS-PAGE. Western blotting was used to analyze the GST pull-down results.

In vitro phosphatase assay

Cx43 phospho-peptides pY247 (KGKSDPpYHATSGA) and pY265 (GSQKpYAYFNG) were used in the Malachite Green assay (Millipore) to detect whether they are substrates of TC-PTP in vitro. Positive and negative controls were DADEpYL (as described in Peters et al., 2000) and ARKRIpYPP (as described in Ren et al., 2011), respectively. Peptides and enzyme were dissolved and diluted in reaction buffer [60 mM HEPES pH 7.2, 150 mM NaCl, 1 mM EDTA, 0.17 mM DTT, 0.83% glycerol, 0.017% bovine serum albumin (BSA) and 0.002% Brij 35] and temperature-equilibrated for 15 min at 25°C. 500 µM of each peptide was added to a 96-well plate, and then TC-PTP_1–314 was added to each peptide at different time-points. The final concentration of TC-PTP_1–314 was 0.6 nM. All reactions were terminated by adding 100 µl of Malachite Green solution and the samples were incubated for 15 min at room temperature to allow color development. The absorbance at 650 nm was read using a SpectraMax 190 spectrometer (Molecular Devices).

The Malachite Green assay was also used to measure the kinetic parameters of dephosphorylation (Peters et al., 2000). The reaction was performed in 96-well plates with a final volume of 25 µl. A standard curve of KH₂PO₄, used for calculating the release of inorganic phosphate, was determined on the same plate as the reaction samples. Different concentrations of peptides (substrate) were combined with 1.2 nM TC-PTP_1–314. The reactions were sampled at different time-points (0–15 min) to calculate the initial velocity (v) versus substrate. A Lineweaver-Burk double-reciprocal plot (rearranged from the Michaelis-Menten equation) was created, based on 1/v and 1/[substrate]. The kinetic parameters were determined by the linear equation from the Lineweaver-Burk double-reciprocal plot. All the reactions were conducted at pH 7.5, 25°C. The reported results were calculated from three independent experiments.

TC-PTP siRNA treatment

The oligonucleotides siRNA1 (5′-AACAGATACAGAGATGTAAGC-3′) and siRNA2 (5′-AAGATTGACAGACACCTAAAT-3′) were used to knock down TC-PTP in LA-25 cells as described previously (Galic et al., 2005). Oligonucleotide from the non-targeting pool (Dharmacon, Lafayette, CO) was used as a negative control. Lipofectamine RNAi MAX (Invitrogen) was used to transfect the oligonucleotide according to the manufacturer's protocol. Cells were incubated at 35°C for another 12 h after 36 h of siRNA treatment. Protein levels were detected by western blotting.

Subcellular fractionation

Analysis of subcellular colocalization of the PKC isoforms was performed as described previously (Saxon et al., 1994; Zang et al., 1995). Briefly, cells with different treatments from 60-mm Petri dishes were washed with ice-cold PBS and then suspended in 0.4 ml of digitonin buffer (0.5 mg/ml digitonin in 20 mM Tris-HCl pH 7.5, 140 mM NaCl, 25 mM KCl, 5 mM MgCl₂, 2 mM EDTA, 2 mM EGTA, one quarter of a protease inhibitor tablet and one phosphatase inhibitor tablet per 10 ml buffer). Cells were dispersed by pipetting, incubated on ice for 5 min, and the cell lysate was subsequently centrifuged at 100,000 g for 1 h. The supernatant was collected as the cytosolic fraction. The pellet was resuspended in the same volume of Triton X-100 buffer (1% Triton X-100 to substitute for the digitonin in the digitonin buffer) and incubated on a rotating wheel for 30 min at 4°C. The suspension was centrifuged as described above and the supernatant was collected as the membrane fraction.

Scrape-loading assay

Cells were scrape-loaded as described previously (Stauch et al., 2012). Briefly, LA-25 cells seeded on coverslips were transfected with TC-PTP_1–314 or pcDNA3.1 vector (mock group) for 24 h at 37°C and then incubated at 40°C for 1 h before the addition of the MAPK inhibitor (U0126, 50 µM) or PKC inhibitor (BIM, 0.1 µM) at 40°C for another 30 min. Next, cells were incubated at 35°C for 12 h. The medium with the inhibitor was changed every 3 h to ensure a consistent effect during the treatment. The confluence of cells was 100% on the coverslips. Cell culture medium from 100% confluent cells was removed and replaced with 1 ml of PBS containing 0.25% Lucifer Yellow and Texas-Red-conjugated fluorescent dextran (10 kDa, 1.5 mg/ml; fixable). Cells were scrape-loaded with a sterile scalpel by two longitudinal scratches and were then incubated at room temperature for 1 min. Cells were washed quickly three times with warm PBS (containing MgCl₂ and CaCl₂) or cell culture medium followed by incubating at 37°C for 5 min. After incubation, cells were washed twice with warm PBS and fixed with 3.7% buffered paraformaldehyde for 15 min. Autofluorescence was quenched by treatment with 0.1 M glycine for 15 min. Coverslips were mounted onto glass slides in a droplet of SlowFade (Invitrogen). The result was confirmed by repeating the experiment three times and, for each trial, four side-by-side images were captured. The method to quantify dye transfer was as described previously (Stauch et al., 2012). Briefly, cells containing Lucifer Yellow were counted, and the cells with dextran, which indicated the initially loaded cells, were excluded from the count.

Triton X-100 solubility

The Triton X-100 solubility assay was modified from a method described previously (Mitra et al., 2006). LA-25 cells grown in 10-cm Petri dishes were rinsed three times with PBS and scraped into 1 ml of lysis buffer (50 mM Tris-HCl pH 7.4, 1 mM EGTA, 1 mM EDTA, 1 mM PMSF, 100 mM NaCl, half a tablet of complete protease inhibitor per 25 ml of buffer, 5 mM NaF and 5 mM Na₃VO₄). Next, cells were sonicated for 10 s on ice. Protein quantification was performed using the BCA method. A total of 450 µl of cell lysates were added to 50 µl of 10% SDS, which was saved as total protein, or to 10% Triton X-100 (final concentration of 1% Triton X-100) and incubated at 4°C for 30 min. Lysates were then separated into cytosolic (supernatant, soluble) and membrane (pellet, insoluble) fractions by centrifugation at 100,000 g for 1 h at 4°C. The pellets were dissolved in 500 µl of dissolving buffer (70 mM Tris-HCl pH 6.8, 8 M urea, 2.5% SDS, 0.1 M DTT, 5 mM NaF, 5 mM Na₃VO₄ and half a tablet of complete protease inhibitor per 25 ml of buffer). Equal volumes of total lysate, Triton-X-100-soluble and Triton-X-100-insoluble portions were loaded onto a 12% gel, subjected to SDS-PAGE and immunoblotted with the anti-Cx43NT1 antibody.

For the detergent extraction in situ, LA-25 cells on coverslips were extracted in situ with 1% Triton X-100 buffer (1% Triton X-100, PBS pH 7.4, 1 mM CaCl₂, 1 mM MgCl₂ and protease inhibitor cocktail) for 30 min at 4°C, with gentle shaking every 10 min. The control cells were treated in the same way without 1% Triton X-100. The coverslips were fixed and immunostained as described previously (Johnson et al., 2013).

Statistical analysis

All data were analyzed by using GraphPad Prism 5.0 and were presented as the mean±s.e.m. Paired t-tests were used to compare differences between the experimental group and control or the two parallel experimental groups. P<0.05 was considered to be statistically significant.