PKCμ promotes keratinocyte cell migration through Cx43 phosphorylation-mediated suppression of intercellular communication

Summary Downregulation of intercellular communication through suppression of gap junctional conductance is necessary during wound healing. Connexin 43 (Cx43), a prominent gap junction protein in skin, is downregulated following wounding to restrict communication between keratinocytes. Previous studies found that PKCμ, a novel PKC isozyme, regulates efficient cutaneous wound healing. However, the molecular mechanism by which PKCμ regulates wound healing remains unknown. We have identified that PKCμ suppresses intercellular communication and enhances cell migration in an in vitro wound healing model by regulating Cx43 containing gap junctions. PKCμ can directly interact with and phosphorylate Cx43 at S368, which leads to Cx43 internalization and downregulation. Finally, utilizing phosphomimetic and non-phosphorylatable S368 substitutions and gap junction inhibitors, we confirmed that PKCμ regulates intercellular communication and in vitro wound healing by controlling Cx43-S368 phosphorylation. These results define PKCμ as a critical regulator of Cx43 phosphorylation to control cell migration and wound healing in keratinocytes.


INTRODUCTION
Wound healing is an intricate physiological response to an injury that encompasses four phases; hemostasis, inflammation, proliferation, and tissue remodeling. 1 Complications in wound healing can arise at any stage and lead to chronic, or non-healing, wounds that have been well documented in wounds associated with diabetic foot ulcers, venous ulcers, surgery, and the aging population. 2 Downregulation of intercellular communication between various cell types is essential for efficient wound healing.Direct communication between cells is mediated in part through gap junctions composed of connexin proteins which allow for the exchange of small ions and secondary metabolites between adjacent cells. 3Connexin 43 (Cx43) is the most abundantly expressed connexin isoform in the stratum basale and spinosum layers of the skin epidermis. 4][10] Similarly, heterozygous Cx43 knockout mice display accelerated wound closure due to enhanced proliferation and re-epithelialization. 11 Cx43 function is tightly regulated by multiple phosphorylation events in its intracellular carboxy-terminal domain. 12Phosphorylation of Cx43 at Serine-368 reduces channel conductance and promotes the degradation of Cx43 plaques through lysosomal and proteasomal degradation pathways, thereby reducing intercellular signaling. 13,14Cx43-S368 phosphorylation is also highly regulated during wound healing.The unwounded human epidermal layer contains an even distribution of Cx43 with low levels of S368 phosphorylation.In contrast, phosphorylated Cx43-S368 is significantly increased in basal keratinocytes 24 h post wounding.A key role for Cx43-S368 has been emphasized by the fact that treatment with Cx43 specific peptide inhibitor a-CT-1 accelerates wound healing in vivo. 15It has also been reported that a-CT-1 enhances the phosphorylation of Cx43-S368 in injury models including scratch wounded cultured cells and in vivo models of ventricular injury. 16 variety of kinases have been shown to target Cx43, including members of the Protein Kinase C (PKC) family which phosphorylate S368 on the C-terminal tail of Cx43. 14,17PKCs are a superfamily of serine/threonine kinases that play a vital role in multiple cell functions such as proliferation, migration, and apoptosis. 18There are three classes of PKCs based on their secondary messenger requirements.Classical PKCs (a, b I , b II , and g) require diacylglycerols (DAGs) and calcium ions for activation whereas novel PKC isozymes (d, ε, h, m, and q) only require DAGs.0][21] Novel PKC isozymes PKCd and PKCε can PMA dramatically suppressed intercellular communication, which was substantially blocked when PKCm was inhibited by CRT (Figures 2A and  2B).CRT alone did not influence intercellular communication in the SL/DT assay (Figures S2A and S2B).Furthermore, the PMA-induced suppression of intercellular communication was also reversed when HaCaT cells were depleted of PKCm by shRNA, as PKCm depleted cells displayed high levels of dye migration even in the presence of PMA (Figures 2C and 2D).These data suggest that PKCm is an important mediator of gap junction-mediated intercellular communication in response to PMA.Phosphorylation of Cx43 on S368 reduces channel conductance and signals Cx43 plaque internalization and degradation. 14Therefore, we next assessed whether suppressing PKCm activity influences Cx43-S368 phosphorylation status.When HaCaT cells were treated with PMA, the phosphorylation level of Cx43 was dramatically enhanced, which was reversed by the PKCm inhibitor CRT in a dose-dependent manner (Figures 2E and 2F).Furthermore, the decrease in Cx43-S368 phosphorylation status due to CRT treatment mirrored that of PKCm-S916, indicating a direct correlation between PKCm activation status and Cx43-S368 phosphorylation (Figures 2E, 2F, and S2C).Similarly, HaCaT cells depleted of PKCm also showed reduced levels of pCx43-S368 compared to control shRNA treated cells in the presence of PMA, which again mirrored the decrease in PMA induced PKCm activation status (Figures 2G and 2H).Similar results were also observed in mouse embryonic fibroblasts (MEFs) that were either treated with the PKCm inhibitor CRT or depleted of PKCm with shRNA, suggesting a potentially wider relevance of this signaling pathway (Figures S2D and S2E).To determine if downregulation of Cx43, another consequence of S368 phosphorylation, was influenced by PKCm inhibition, HaCaT cells were treated in the presence of PMA with or without CRT along with cycloheximide to inhibit de novo protein synthesis. 42Monitoring the loss of Cx43 following cycloheximide, we found that PMA treatment enhanced degradation of Cx43, whereas treating with CRT blocked the PMA-induced degradation of Cx43 (Figures 2I and 2J).Taken together, these results suggest that Cx43-S368 phosphorylation and degradation are regulated by PKCm.

PKCm colocalizes with Cx43
Next, we conducted immunofluorescence staining to determine the subcellular localization of PKCm and Cx43 in the presence or absence of PMA.We observed that PKCm and Cx43 colocalized near the plasma membrane (Figure S3A).Furthermore, upon PMA stimulation, both PKCm and Cx43 appeared to be internalized from the plasma membrane and continued to colocalize in the cytosol (Figure S3A).Following a time course of PMA treatment, we found that PKCm and Cx43 showed colocalization in untreated cells at the membrane and further colocalization in internalized regions in the cytosol (Figure 3A).Interestingly, at 30 min post PMA treatment, the two signals appeared to be separated, with PKCm remaining closer to the interface of adjacent cells than Cx43.These results indicate that while PKCm and Cx43 function at similar locations in the cell and are trafficked from the plasma membrane similarly following PMA stimulation, the dynamics of internalization of the two proteins may differ, suggesting that the interaction between PKCm and Cx43 may be transient and that the two proteins likely do not form a long-lasting stable complex.Furthermore, when human skin samples were either left unwounded or subjected to a punch biopsy wound, we observed that phosphorylation of PKCm-S916 increased 3.6-fold in the epidermis (Figures 3B and 3C).In addition, in these human skin biopsies, pCx43-S368 appears to be internalized into the cytosol and localized in a perinuclear fashion, similar to the localization of Cx43 in HaCaT cells following treatment with PMA (Figure 3A).Taken together, these data support the notion that stimulation of human keratinocyte-derived cells or human skin tissue, either through PMA treatment or wounding, promotes PKCm-S916 phosphorylation and subsequent Cx43-S368 phosphorylation and internalization, to suppress intercellular communication to promote wound healing.

PKCm interacts with, and phosphorylates, Cx43 at Serine-368 in vitro
Prior studies have shown that PKCd and PKCε isozymes interact with, and phosphorylate, Cx43 at S368. 13,43 The C-terminal tail of Cx43 consists of a PKC consensus sequence (RXSSR) corresponding to the amino acid sequence RASSR in which the first serine residue is S368. 44Since we have shown that PKCm inhibition or depletion reduces Cx43-S368 phosphorylation, we examined whether PKCm directly interacts with Cx43.To this end, we purified glutathione S-transferase (GST) fused to the amino-terminus of full-length Cx43 (GST-Cx43), as well as unfused GST (serving as a negative control) (Figures S3B and S3C).Amino-terminal hemagglutinin (HA) epitope-tagged PKCm (HA-PKCm) was immunoprecipitated with HA antibody conjugated to agarose from transfected HEK 293T cells.Interestingly, HA-PKCm transfected 293T cells had a higher basal level of Cx43-S368 phosphorylation when compared to cells transfected with the control vector (Figure S3D), corroborating our finding that Cx43-S368 phosphorylation is controlled, in part, by PKCm.Utilizing immunoprecipitated HA-PKCm that remained bound to agarose beads, we performed an in vitro pull-down assay in the presence of GST or GST-Cx43.We found that GST-Cx43 bound HA-PKCm, whereas untagged GST did not (Figure 4A), indicating that PKCm is capable of directly interacting with Cx43 in vitro.
PKCm targets multiple substrates for phosphorylation to regulate downstream cellular functions. 45While our studies above indicate that PKCm can regulate Cx43-S368 phosphorylation in cells, to examine whether PKCm can directly phosphorylate Cx43-S368, we conducted an  in vitro kinase assay with HA-PKCm and GST-Cx43 in the presence or absence of ATP.First, we observed a phosphorylated S368 signal in GST-Cx43 when incubated with HA-PKCm in an ATP-dependent manner (Figure 4B).Furthermore, when we purified HA-PKCm from PMA treated cells, we observed an enhanced ability of HA-PKCm to phosphorylate GST-Cx43 on S368 (Figure 4B), suggesting that purified HA-PKCm from PMA treated cells had a higher activation status.7][48] Each of these HA-PKCm proteins were immunoprecipitated and utilized in an in vitro kinase assay.While we found that wild-type HA-PKCm was efficient at phosphorylating GST-Cx43, catalytically inactive HA-PKCm K612W was unable to carry out this activity (Figure 4C).Overall, these data suggest that PKCm can directly bind to and phosphorylate Cx43 at S368.

Inhibition of intercellular communication accelerates wound healing in PKCm inhibited or depleted cells
To investigate whether PKCm regulates in vitro wound healing through Cx43-containing gap junctions, we blocked gap junction activity with the non-selective inhibitor Carbenoxolone (CBX) which is a glycyrrhetinic acid derivative that uncouples gap junctions. 49We first set out to determine whether the reversal of PMA-mediated suppression of intercellular communication by the PKCm inhibitor CRT was dependent on gap junction conductance.To this end, we cotreated HaCaT cells with PMA and CRT, in the absence or presence of CBX, followed by SL/DT assays.We observed that CBX inhibition dramatically reduced intercellular communication in cells treated with PMA and CRT (Figures 5A and  5B).Next, utilizing the same treatments, we assessed whether cell migration regulated by PMA and CRT in an in vitro wound healing assay would be similarly influenced by the gap junction inhibitor.While PMA treatment accelerated wound closure, which was blocked by CRT, we observed that treatment with CBX reversed the effect of CRT and enhanced wound healing to a similar extent as PMA treatment alone, suggesting that gap junction inhibition plays a key role in cell migration and wound healing regulated by PKCm (Figures 5C and 5D).In addition, we conducted SL/DT and in vitro wound healing assays in control and PKCm-depleted HaCaT cells treated with or without CBX.PMA dramatically reduced intercellular communication of HaCaT cells, which was blocked when PKCm was depleted (Figures 5E and S4A).Furthermore, CBX treatment completely reversed the effect of PKCm depletion on PMA-induced suppression of intercellular communication (Figures 5E  and S4A).Similarly, in vitro wound healing assays showed that gap junction inhibition with CBX increased the rate of wound healing in PKCm-depleted HaCaT cells (Figures 5F and S4B).Next, we assessed whether depleting Cx43 would block the effect of PKCm inhibition on intercellular communication.To this end, we generated HaCaT cells depleted of Cx43 using shRNA (Figure S5A).Control and Cx43depleted cells were then treated with PMA in the absence or presence of CRT, and intercellular communication was measured by SL/DT assay.We found that PMA suppressed intercellular communication, which was reversed by CRT in control cells (Figures 5G and 5H).However, Cx43depleted cells treated with PMA showed reduced intercellular communication similar to control cells treated with PMA, but CRT was unable to reverse the PMA-induced suppression of intercellular communication (Figures 5G and 5H).In an in vitro wound healing assay, we observed that basal cell migration was enhanced in Cx43-depleted cells compared to control cells and CRT was unable to reverse the effects of PMA treatment on cell migration (Figures 5I and 5J).Overall, these results indicated that PKCm regulation of cell migration in an in vitro wound healing assay is mediated by PKCm controlling intercellular communication through Cx43 containing gap junctions.

PKCm-mediated suppression of intercellular communication and promotion of wound healing is through Cx43-S368 phosphorylation
While we have shown that PKCm suppresses intercellular communication and promotes wound healing in a manner that depends on functional gap junctions, and that this activity correlates with the regulation of Cx43-S368 phosphorylation, it remains unclear whether PKCm-mediated phosphorylation of Cx43 on S368 serves as the molecular basis for PKCm suppressing intercellular communication and promoting wound healing.To investigate whether PKCm regulates Cx43-S368 phosphorylation to suppress intercellular communication and promote wound healing, we reconstituted Cx43-depleted HaCaT cells (Figure S5A) with exogenously expressed wild-type Cx43 (WT) or Cx43 carrying phosphorylation mimetic or non-phosphorylatable versions of S368 (S368D or S368A, respectively) (Figure S5B).Cx43-depleted cells suppressed intercellular communication and promoted wound healing to a greater extent than shGFP control cells (Figures 5G-5J), which is in accordance with prior studies demonstrating a role for Cx43-mediated intercellular communication and wound healing. 50,51Cx43-depleted HaCaT cells reconstituted with Cx43 WT regained basal intercellular communication, wound healing capacity, and responsiveness to CRT treatment in both SL/DT and in vitro wound healing assays (Figures 6A, 6B, 6E, and 6F).Cx43-depleted HaCaT cells reconstituted with the phosphomimetic Cx43 S368D phenocopied Cx43-depleted cells, showed reduced basal intercellular communication along with enhanced wound healing, and were unaffected by CRT treatment (Figures 6A, 6C, 6E, and 6G).In contrast, reconstitution with non-phosphorylatable Cx43 S368A did not show reduced intercellular communication or enhanced wound healing in response to PMA treatment, consistent with an inability to be phosphorylated at S368 to downregulate Cx43 protein levels and reduce intercellular communication (Figures 6A,  6D, 6E, and 6H).In addition, cells reconstituted with Cx43 S368A also showed no further changes in intercellular communication and cell migration in response to CRT treatment (Figures 6A, 6D, 6E, and 6H).These results indicate that the ability of PMA to suppress intercellular communication and promote wound healing in a PKCm-dependent manner relies on Cx43 S368 phosphorylation.
a-CT1 enhances S368 phosphorylation and increases the rate of in vitro wound healing in the absence of PKCm CBX is a non-selective gap junction inhibitor that can inhibit multiple connexins expressed in keratinocytes such as Cx26 and Cx43. 52Therefore, we next utilized a Cx43 specific inhibitor termed alpha-carboxyl terminus 1 (a-CT1) which is a 25 amino acid peptide comprising the antennapedia internalization sequence and the carboxyl-terminal tail of Cx43. 53The mechanism of action of a-CT1 is that it inhibits the interaction between the tight junction associated protein Zonula Occludens 1 (ZO-1) and the C-terminus of Cx43. 54Previous studies have shown that treatment with a-CT1 in either mouse cardiac injury models or scratch wounded HeLa cells increases S368 phosphorylation of Cx43. 16ince our previous data demonstrated that Cx43-S368 phosphorylation is lower in PCKm inhibited HaCaT cells, we examined if a-CT1 can increase Cx43-S368 phosphorylation in an in vitro wound healing assay in HaCaT cells.In addition to a-CT1, we utilized the antennapedia internalization sequence alone (ANTP) and reverse inactive sequence of a-CT1 (a-CT1 RIS) as control peptides (Figure 7A).HaCaT cells were treated with PMA in the absence or presence of CRT with or without ANTP, a-CT1 RIS, or a-CT1, subjected to a scratch, and incubated for 16 h before preparing cell lysates.Western blotting assays showed that a-CT1 treatment increased phosphorylated Cx43-S368 in scratch wounded HaCaT cells treated with a combination of PMA and the PKCm inhibitor CRT (Figures 7B and 7C).Interestingly, total Cx43 protein levels were dramatically lower in the a-CT1 treated group similar to the cells treated with PMA alone, suggesting that PMA-dependent phosphorylation of Cx43-S368, which was suppressed by PKCm inhibition/depletion, was reversed by a-CT1 treatment.Interestingly, this loss of Cx43 over the 16-h time course, which is blocked by CRT, is consistent with analysis of Cx43 protein stability observed when PMA destabilizes Cx43 which is blocked by CRT (Figures 2I and 2J), further supporting the notion that PKCm is a critical regulator of Cx43 stability in keratinocytes.
Given that a-CT1 induced the phosphorylation of Cx43 S368 even in the presence of PMA/CRT, a-CT1 may promote Cx43 phosphorylation in a manner independent of PKCm.Next, we conducted SL/DT assays to determine the effect of a-CT1 on intercellular communication in PKCm inhibited cells.Treatment with control peptides (ANTP or a-CT1 RIS) did not influence the level of intercellular communication.However, a-CT1 treatment significantly reduced intercellular communication in HaCaT cells in which PKCm was inhibited by CRT (Figures 7D and  7E).We also carried out SL/DT assays in PKCm-depleted HaCaT cells and observed that a-CT1 treatment significantly reduced intercellular communication in PKCm depleted cells (Figures 7F-7H, S6A, and S6B).In addition to SL/DT assays, we carried out in vitro wound healing assays to determine whether a-CT1 can rescue the delayed wound healing observed in PKCm inhibited or depleted HaCaT cells.In assays using CRT to inhibit PKCm, a-CT1 treatment significantly increased the rate of cell migration and in vitro wound healing (Figures 8A and 8B).Similarly, a-CT1 treatment increased cell migration in PKCm-depleted HaCaT cells (Figures 8C-8F, S7A, and S7B).Given a-CT1 induced the phosphorylation of Cx43 S368 independent of PKCm and reversed the effects of PKCm inhibition or loss, these results indicate that PKCm regulates cell migration and in vitro wound healing by regulating the phosphorylation status of Cx43-S368 to control gap junction function and intercellular communication (Figure 7G).

DISCUSSION
Efficient cutaneous wound healing requires suppression of intercellular communication which is achieved by increasing Cx43-S368 phosphorylation and downregulation of Cx43 containing gap junctions. 7On the other hand, delayed wound healing is accompanied by upregulation of the gap junction protein Cx43.Since gap junctional channel conductance and turnover rate of Cx43 are regulated by post-translational modifications of the Cx43 C-terminal tail, kinases that phosphorylate Cx43 may be viable targets for enhancing wound healing capacity, especially in clinical cases of chronic wounds. 12Previous studies have implicated novel PKC isozymes d and ε in wound healing in human and mouse fibroblasts, respectively. 26,27Both PKC isozymes can phosphorylate Cx43 at S368. 13,43 A recent study has discovered the role of another novel PKC isozyme termed PKCm in cutaneous wound healing in vivo. 28However, the mechanism by which PKCm regulates wound healing in the skin, and if it occurs through Cx43, is yet to be elucidated.
In the present study, we identified a previously unknown signaling pathway by which PKCm regulates cell migration and in vitro wound healing by controlling intercellular communication mediated by the gap junction protein Cx43.Our data revealed that PKCm regulates Cx43 by phosphorylating its C-terminal tail residue S368 (Figure 8G).Our in vitro studies suggested that PKCm can directly bind and phosphorylate Cx43.Furthermore, immunofluorescence followed by confocal microscopy and colocalization analysis suggested that PKCm and Cx43 co-localize at the plasma membrane.Interestingly, both PKCm and Cx43 internalize into the cytosol upon treatment with PMA and continue to colocalize in these internalized structures.The dynamics of internalization appear to be different between PKCm and Cx43, as PKCm either remains on the membrane or is closer to the adjacent cell than Cx43 at an intermediate time point (Figure 3A, 30 min).In addition, Cx43 proteins that do not co-localize with PKCm appear to form larger, more well-developed junctions, suggesting that the increased stability of Cx43 when not associated with PKCm may protect these channels from being internalized.Thus, future studies on the internalization dynamics of both these proteins are warranted.Our data also showed extensive PKCm staining outside of Cx43 positive regions indicating that while PKCm may colocalize with Cx43 to promote its phosphorylation and downregulation, there are likely other functions of PKCm outside of Cx43-containing gap junctions, and whether those other functions regulate intercellular communication and/or wound healing remains to be determined.In addition, we observed that in wounded human skin samples, PKCm-S916 becomes hyperphosphorylated (Figure 3).Interestingly, we observed a much more pronounced increase in PKCm-S916 phosphorylation compared to Cx43-S368.However, this may be due to the fact that we did not have control over when the wound would be administered in the human skin biopsy, and thus the tissue fixation may not have been carried out at an optimal time point to observe maximal PKCm-S916 or Cx43-S368 phosphorylation.Furthermore, the regulation of Cx43-S368 phosphorylation by PKCm promotes Cx43 degradation; thus, phosphorylated Cx43 may be internalized and degraded, thereby reducing the observed Cx43-S368 phosphorylation levels between native and wounded tissues.This idea is further supported by the observed perinuclear localization of pCx43-S368 in the cytosol, similar to internalization of Cx43 in HaCaT cells following PMA treatment (Figures 3A and 3B).In addition, while PMA is known to decrease intercellular communication and induce cell migration and wound healing, our results suggest that the molecular basis of PMA controlling intercellular communication in HaCaT keratinocytes is through the activation of PKCm-dependent Cx43-S368 phosphorylation to control gap junction function, and the ability of PMA to suppress intercellular communication and promote wound healing is completely blocked by PKCm inhibition or depletion (Figures 1 and 2).Furthermore, we used Cx43-depletion followed by reconstitution with Cx43 S368 phosphorylation mutants to demonstrate that the phosphorylation of S368 is required for the effects of PKCm.These findings are further supported by the observation that the ability of PKCm to regulate intercellular communication and cell migration can be reversed by either the pan-gap junction inhibitor Carbenoxolone or the Cx43-specific inhibitor (and an inducer of S368 phosphorylation) a-CT1, which reduced intercellular communication and accelerated in vitro wound healing when cells were treated with PMA and PKCm was inhibited or depleted.These studies demonstrate a key role for PKCm in the molecular pathways critical for regulating keratinocyte intercellular communication, cell migration, and in vitro wound healing.
PKCm is known to regulate cell proliferation and de-differentiation in cultured keratinocytes in conditions of low calcium switch. 55However, an in vivo study demonstrated that PKCm is dispensable for normal skin homeostasis including epidermal cell proliferation or differentiation. 28pon cellular injury such as wounding, epidermis specific PKCm knockout mice display delayed wound healing and reduced proliferation in keratinocytes. 28Our finding that PKCm inhibition or depletion delays wound healing in vitro is consistent with this in vivo study.While other PKC isoforms have been shown to regulate Cx43 phosphorylation at S368, the involvement of PKCm has not been well characterized, and was considered to have only a minor role. 29PKCε and PKCd phosphorylate Cx43-S368 in osteoblast cell lines and cardiomyocytes, respectively, leading to a reduction in unitary channel conductance. 13,43Until now, PKCm has not been implicated in regulating Cx43 phosphorylation or function, or intercellular communication.In fact, PKCm was thought to be of minor consequence to TPA-mediated downregulation of intercellular communication in rat R6 fibroblasts where continued TPA treatment led to a modest downregulation of PKCm unlike other PKC isozymes such as PKCa, PKCd, and PKCε. 29In contrast to rat fibroblasts, we find that PKCm plays a key role in maintaining intercellular communication in keratinocytes, suggesting that PKCm may play a preferential role in regulating Cx43 in a cellular or tissue context-dependent manner, such as in the skin.We further observed that PKCm can interact with and directly phosphorylate Cx43 at S368.The general consensus recognition sequence of PKCm is an LxRxxpS/T motif. 56For instance, PKCm phosphorylates Polycystin-2 (TRPP2) at S801 which conforms to the LxRxxpS/T motif. 57However, the amino acid sequence RASSR that encompasses Cx43-S368 does not fit to this putative consensus motif.There have been reports of several other PKCm substrates such as c-Jun and b-catenin that also do not fit this optimal consensus motif. 58,59t has been speculated that these proteins bind to PKCm at its docking site which facilitates the phosphorylation of sub-optimal motifs present in the substrates. 45Therefore, it is plausible to speculate that PKCm may bind to Cx43 and target S368 for phosphorylation through a similar mechanism.
Our colocalization studies revealed that PKCm is abundant in the plasma membrane where Cx43 gap junction plaques are present.Prolonged PMA treatment led to dispersal of both PKCm and Cx43 from the plasma membrane to the cytosol.In many cases, phosphorylation by PKCm induces interaction of its substrates with the scaffolding proteins such as those within the 14-3-3 family.For instance, 14-3-3 binds to phosphorylated RIN1 and sequesters it to the cytosol, preventing its function at the plasma membrane. 60,61Our results demonstrated that PMA-activated PKCm colocalizes with Cx43 in the cytosol, reducing the amount of Cx43 present at the plasma membrane.Cycloheximide chase assays further revealed that prolonged PMA treatment reduces total Cx43 levels.However, PKCm inhibition significantly reduced the rate of Cx43 degradation, suggesting that degradation of Cx43 in response to these stimuli requires PKCm.Interestingly, 14-3-3 has also been studied in relation to gap junction plaque internalization where 14-3-3t binding leads to Cx43 ubiquitination, internalization, and degradation. 62Therefore, our finding suggests that PKCm regulates Cx43 in a parallel manner to how PKCm targets other substrates like RIN1 and given 14-3-3 proteins often bind their targets in a phosphorylation-dependent manner, suggest that PKCm targeting Cx43 may provide that phospho-binding motif to allow its recognition by 14-3-3.Future studies are warranted to determine if 14-3-3s mediate the internalization of Cx43 following phosphorylation by PKCm.
4][65] In addition, identifying PKCm as a key regulator of Cx43-dependent pathways and downstream physiological functions as well as pathophysiology associated with Cx43 dysregulation opens the door to identifying upstream regulators of PKCm activation and function that may serve as therapeutic targets for wound healing as well as for additional disease settings.
Over the past few decades, gap junction inhibition has been a popular strategy to enhance wound healing, 66 leading to the assessment of various methods of targeting Cx43 inhibition, such as using antisense oligonucleotides that bind to Cx43 mRNA and connexin mimetic peptides that directly bind to Cx43, such as Gap27. 10,67We utilized two strategies to inhibit gap junctions in the present study.Carbenoxolone is a non-selective inhibitor of gap junctions that shows modest potency, and its therapeutic efficacy has been tested previously in patients with duodenal ulcers. 68However, CBX shows an unfavorable safety profile with several side effects, such as hypokalemia, which induces renal and neuromuscular damage with prolonged treatment. 68We observed that CBX substantially reduced intercellular communication and increased in vitro wound healing in PKCm inhibited HaCaT cells.The second strategy employed to inhibit gap junction activity was a-CT1, which inhibits the interaction of the Cx43 C-terminal tail with the scaffolding protein ZO-1.Previous studies have found that treatment with a-CT1 also promotes phosphorylation of Cx43-S368 in cardiac ischemia-reperfusion injury models in a PKCε-dependent manner. 69Phosphorylation of Cx43-S368 is associated with the reduction of Cx43 channels in the plasma membrane. 13,14Our observation that a-CT1 increases Cx43 phosphorylation at S368 and downregulates Cx43 in scratch wounded HaCaT cells is consistent with these findings.Moreover, reconstitution experiments with Cx43 phosphomimetic substitutions (S368D) suggested that the S368 residue in the Cx43 C-terminal tail is essential for PKCmmediated regulation of intercellular communication and wound healing.Furthermore, the degradation of Cx43 in response to PMA appears to be fully dependent on PKCm.Taken together, these results indicate that PKCm regulates intercellular communication and cellular migration by controlling gap junction biology.
Interestingly, our data revealed that a-CT1 reduced intercellular communication in HaCaT cells.Animal studies in mice and pigs have shown that a-CT1 increases the wound healing rate. 70,71Clinical trials with a-CT1 have determined that it improves scar visual appearance by 47% in patients who underwent laparoscopic surgery. 72Consistent with this finding, we observed that a-CT1 accelerated in vitro wound healing in PKCm-inhibited or depleted HaCaT cells, suggesting that Cx43-S368 phosphorylation may be necessary for PKCm-mediated control of intercellular communication and wound healing in keratinocytes.Therefore, this study has identified a potential PKCm pathway that could be included in future studies involving a-CT1 and Cx43 in the context of wound healing.
In conclusion, our study shows that PKCm controls cell migration in an in vitro wound healing model through Cx43-S368 phosphorylation to suppress intercellular communication, providing important insights into the role of PKCm in skin biology.Cx43 and its controlled downregulation in the wounded epidermis is essential for efficient wound healing, whereas chronic non-healing wounds have been shown to overexpress Cx43. 73In addition, the present study suggests that PKCm may play a key role in regulating Cx43 levels in keratinocytes to further promote wound repair warranting future studies assessing the regulation of Cx43 by PKCm in vivo.

Limitations of the study
Cx43-S368 is phosphorylated by several other PKC isozymes. 13,43

Figure 1 .
Figure 1.PKCm inhibition/reduction impairs wound healing in vitro (A) HaCaT cells treated with PMA (10 nM) in the absence or presence of CRT (1 mM CRT) for 16 h.Western blots of cell lysates probed for pPKCm-S916, PKCm, and tubulin.(B) Relative pPKCm-S916 levels in treated samples compared to DMSO control based on densitometric analysis from three replicate experiments as shown in (A).(C) Representative images of in vitro wound healing assay at 0 and 16 h on HaCaT cells treated with DMSO, PMA (10 nM), or PMA with CRT (1 mM CRT).Scale bar = 100 mm.(D) Percentage of wound area remaining after 16 h from three replicate experiments as shown in (C).(E) Representative images from immunofluorescence staining of pPKCm-S916 in scratch wounded HaCaT cells at 30-and 120-min post-wounding.Scale bar = 40 mm.(F) Relative pPKCm-S916 levels based on fluorescence intensity from three images from 30-and 120-min post-wounding as shown in (E).Relative levels are based on 10 random fields taken from cells adjacent to the wound and cells interior to the wound for each time point.(G) HaCaT cells transduced with control shRNA (shGFP) or two shRNAs directed against PKCm.Cell lysates were probed by western blotting for PKCm and tubulin.(H) Quantification of PKCm levels relative to shGFP from three replicate experiments as shown in (G).(I) Representative images of in vitro wound healing assay at 0 and 16 h on HaCaT cells transduced with control shRNA (shGFP) or two shRNAs directed against PKCm treated in the absence or presence of PMA (10 nM).Scale bar = 100 mm.(J) Percentage of wound area remaining after 16 h from three replicate experiments as shown in (I).All calculations are based on three replicates À/+ S.D. **p < 0.01, ***p < 0.001 (Student's t test).

Figure 2 .
Figure 2. PKCm regulates the level of intercellular communication and Cx43-S368 phosphorylation (A) Representative fluorescent microscopic images of SL/DT assay in HaCaT cells treated with CRT (5 mM) for 6 h followed by PMA (0.5 mM) for 30 min.Scale bar = 100 mm.(B) Quantification of dye migration area from three replicate experiments as shown in (A).(C) Representative fluorescent microscopic images of SL/DT assay in HaCaT cells transduced with control shRNA (shGFP) or two shRNAs directed against PKCm and treated with or without PMA (0.5 mM) for 30 min.Scale bar = 100 mm.(D) Quantification of dye migration area from three replicate experiments as shown in (C).(E) HaCaT cells treated with CRT at indicated concentrations for 6 h followed by a 30-min treatment with PMA (0.5 mM).Cell lysates were probed by western blotting for pPKCm-S916, PKCm, pCx43-S368, Cx43, and tubulin.(F) Quantification of pCx43-S368 in indicated treated conditions compared to DMSO based on densitometric analysis from three replicate experiments as shown in (E).(G) HaCaT cells transduced with control shRNA (shGFP) or two shRNAs directed against PKCm and treated with or without PMA (0.5 mM) for 30 min.Cell lysates were probed by western blotting for pPKCm-S916, PKCm, pCx43-S368, Cx43, and tubulin.(H) Quantification of pCx43-S368 in indicated treated conditions compared to DMSO based on densitometric analysis from three replicate experiments as shown in (G).(I) HaCaT cells were treated with Cycloheximide (100 mg/mL) in the absence or presence of PMA (0.5 mM) with or without CRT (5 mM).Whole-cell lysates were collected at indicated time points and treated with Lambda phosphatase and subsequently probed by western blotting for Cx43, PKCm, and tubulin.(J) Relative levels of total Cx43 for each treatment group compared to T = 0 h.All calculations are based on three replicates À/+ S.D. *p < 0.05, **p < 0.01, ***p < 0.001 (Student's t test).

Figure 3 .
Figure 3. PKCm and Cx43 colocalize in HaCaT cells (A) Representative images of immunofluorescence staining of PKCm (green) and Cx43 (red) in HaCaT cells treated with PMA at different time points.Scale bar = 10 mm.(B) Immunofluorescence images of intact and wounded human skin stained for pPKCm-S916, total PKCm, pCx43-S368, and total Cx43.Scale bar = 100 mm.(C) Relative pPKCm-S916 levels in intact and wounded skin based on immunofluorescence signal intensity of images shown in (B).Error bars represent S.D. ***p < 0.001 (Student's t test).

Figure 4 .
Figure 4. PKCm interacts with Cx43 and phosphorylates Cx43 at S368 (A) In vitro pull-down assay with HA-PKCm attached to HA antibody conjugated agarose beads were incubated with either GST or GST-Cx43 and subsequently washed to remove unbound material.Pulled down material was probed by western blotting for GST and HA.(B) In vitro kinase assay with HA-PKCm immunoprecipitated from untreated or PMA treated transfected HEK 293T cells incubated with bacterially expressed GST-Cx43 in the presence or absence of ATP (200 mM).Reactions were probed by western blotting for pCx43-S368, HA, and GST.(C) In vitro kinase assay as described in (B) with wild-type (WT) or catalytically inactive (K612W) HA-PKCm and GST-Cx43.Reactions were probed by western blotting for pCx43-S368, HA, and GST.

Figure 5 .
Figure 5. PKCm regulates wound healing through gap junction mediated intercellular communication (A) Representative fluorescent microscopic images of SL/DT assay in HaCaT cells treated in the absence or presence of CRT (1 mM CRT), with or without CBX (100 mM) for 6 h, followed by PMA (0.5 mM) for 30 min.Scale bar = 100 mm.(B) Quantification of dye migration area from three replicate experiments as shown in (A).(C) Representative images of in vitro wound healing assay in HaCaT cells treated with PMA (10 nM) in the presence or absence of CRT (1 mM), with or without CBX (50 mM) at 0 and 16 h.Scale bar = 100 mm.(D) Percentage of wound area remaining after 16 h from three replicate experiments in (C).(E) Quantification of dye migration area from SL/DT assays in HaCaT cells expressing shGFP, shPKCm A, or shPKCm B, and treated with or without CBX (100 mM) for 6 h followed by PMA (0.5 mM) for 30 min from three replicate experiments in (Figure S4A).(F) Percentage of wound area remaining after 16 h in HaCaT cells expressing with shGFP, shPKCm A, or shPKCm B, in the absence or presence of PMA (0.5 mM), with or without CBX (100 mM) for 16 h from three replicate experiments in (Figure S4B).(G) Representative fluorescent microscopic images of SL/DT assay in HaCaT cells expressing shGFP or shCx43 and treated in the absence or presence of CRT (1 mM) for 5 h with or without PMA (0.5 mM) for 30 min.Scale bar = 100 mm.(H) Quantification of dye migration area from three replicate experiments as shown in (G).(I) Representative images of in vitro wound healing assay of shGFP or shCx43 treated HaCaT cells treated with PMA (10 nM) in the presence or absence of CRT (1 mM) at 0 and 16 h.Scale bar = 100 mm.(J) Percentage of wound area remaining after 16 h from three replicate experiments shown in (I).All calculations are based on three replicates À/+ S.D. *p < 0.05, **p < 0.01, ***p < 0.001 (Student's t test).

Figure 7 .
Figure 7. Cx43 specific inhibitor a-CT1 reduces intercellular communication in PKCm inhibited/depleted HaCaT cells (A) Amino acid sequence of peptides a-CT1, reverse inactive peptide (a-CT1 RIS), and antennapedia control (ANTP).(B) HaCaT cells were scratched with a pipette tip followed by treatment with PMA (10 nM) with or without CRT (5 mM), in the absence or presence of ANTP, a-CT1 RIS, or a-CT1 peptides (each at 100 mM) for 16 h.Cell lysates were probed by western blotting for pCx43-S368 Cx43, pPKCm-S916, PKCm, and tubulin.(C) Relative pCx43-S368 levels in treated samples compared to DMSO control based on densitometric analysis from three replicate experiments as shown in (A).(D) Representative fluorescent microscopic images of SL/DT assay in HaCaT cells treatment with or without CRT (5 mM), in the absence or presence of ANTP, a-CT1 RIS, or a-CT1 peptides (each at 100 mM) for 6 h followed by PMA (10 nM) for 30 min.Scale bar = 100 mM.(E) Quantification of dye migration area from three replicate experiments as shown in (D).(F) Representative fluorescent microscopic images of SL/DT assay in HaCaT cells transduced with control shRNA (shGFP) or an shRNA directed against PKCm and in the absence or presence of ANTP, a-CT1 RIS, or a-CT1 peptides (each at 100 mM) for 6 h followed by PMA (10nM) for 30 min.Scale bar = 100 mm.(G and H) Quantification of dye migration area from three replicate experiments as shown in (F) for shGFP (G) or shPKCm.(H) All calculations are based on three replicates À/+ S.D. **p < 0.01, ***p < 0.001 (Student's t test).

Figure 8 .
Figure 8. Cx43 Specific inhibitor a-CT1 accelerates wound healing in PKCm inhibited/depleted HaCaT cells (A) Representative images of in vitro wound healing assay at 0 and 16 h in HaCaT cells treated with PMA (10 nM) with or without CRT (5 mM), in the absence or presence of ANTP, a-CT1 RIS, or a-CT1 peptides (each at 100 mM).Scale bar = 100 mm.(B) Percentage of wound area remaining after 16 h from three replicate experiments as shown in (A).(C-F) Representative images of in vitro wound healing assay at 0 and 16 h in HaCaT cells transduced with control shRNA (shGFP) (C) or an shRNA directed against PKCm (E) treated with PMA (10 nM) in the absence or presence of ANTP, a-CT1 RIS, or a-CT1 peptides (each at 100 mM).Scale bar = 100 mm.Percentage of wound area remaining after 16 h from three replicate experiments in cells transduced with shGFP (D) as shown in (C), or cells transduced with shPKCm (F) as shown in (E).(G) Schematic depicting PKCm regulating Cx43-S368 phosphorylation to suppress intercellular communication and promote wound healing.All calculations are based on three replicates À/+ S.D. **p < 0.01 (Student's t test).
However, it remains unclear how these other PKC isozymes may influence the interaction and targeting of Cx43 by PKCm.As previous studies have shown that PKCm itself is a target of DAG activated PKC isozymes, we (Continued on next page)