Connexins as Substrates for Protein Kinases and Phosphoprotein Phosphatases

Jose L. Vega1,2, Elizabeth Baeza1, Viviana M. Berthoud3 and Juan C. Saez1,4 1Departamento de Fisiologia, Pontificia Universidad Catolica de Chile, Santiago, 2Laboratorio de Fisiologia Experimental (EPhyL), Universidad de Antofagasta, Antofagasta, Chile, 3Department of Pediatrics, Section of Hematology, Oncology, University of Chicago, IL, 4Instituto Milenio, Centro Interdisciplinario de Neurociencias de Valparaiso, Valparaiso, 3USA, 1,2,4Chile


Introduction 1.1 Mammalian connexins
Connexins are protein subunits expressed by cordates that form gap junction channels (GJCs) and hemichannels (HCs) (Goodenough, 1974;Makowski et al., 1977).A GJC is formed by the head-to-head docking of two HCs, each contributed by one of the two contacting cells (Meşe et al., 2007).Each HC is an oligomeric assembly of six identical (homomeric) or six different (heteromeric) Cx subunits (Sáez et al., 2005).GJCs and HCs subserve different functions; while GJCs communicate the cytoplasm of contacting cells, HCs provide a pathway for communication between the intracellular and extracellular compartments (Bruzzone and Dermietzel, 2006).Although both types of channels are permeable to ions and small molecules, GJCs and HCs composed of the same Cx subtype are likely to present differences in permeability and regulatory properties (Sáez et al., 2003;Meşe et al., 2007;Sáez et al., 2010).
The family of connexin genes has 20 members in the mouse genome and 21 members in the human genome (Eiberger et al., 2001;Willecke et al., 2002;Söhl and Willecke, 2003;2004).Most Cx genes have a similar structure and contain the protein coding region as a single exon (Willecke et al., 2002;Söhl and Willecke, 2003;2004;Pfenniger et al., 2011).Cxs were initially denoted according to the tissue of origin or the apparent size of a polypeptide as determined by SDS-PAGE.Shortly thereafter, it became clear that such designations were inappropriate, because many of these proteins are expressed in more than one tissue (Beyer et al., 1987) and their apparent molecular mass may vary with electrophoresis conditions (Green et al., 1988).Therefore, a standard nomenclature was developed to distinguish members of this family.The current nomenclature uses the abbreviated symbol "Cx" (for connexin) followed by a suffix that indicates the molecular mass of the Cx amino acid sequence (in kDa) predicted from its cDNA.In some cases, a prefix is added to indicate the species of origin.Hydropathicity plots of the Cx amino acid sequences have been used to www.intechopen.comProtein Kinases 200 predict their membrane topology.These analyses predicted the presence of four hydrophobic domains, three hydrophilic cytoplasmic domains (the amino and carboxyl termini and an intracellular loop) and two extracellular loops (Heynkes et al., 1986;Paul, 1986;Beyer et al., 1987).This topology was supported by experiments that studied the binding of site-specific antibodies and protease sensitive sites (Zimmer et al., 1987;Hertzberg et al., 1988;Milks et al., 1988;Yancey et al., 1989;Zhang and Nicholson, 1994;Quist et al., 2000).The cytoplasmic loop and the carboxyl terminus vary extensively in length and amino acid composition and probably contain most of the regulatory sites of GJCs and HCs.

Mammalian protein kinases and phosphoprotein phosphatases
Most Cxs contain putative phosphorylation sites (Lampe and Lau, 2004).As with all phosphoproteins, their phosphorylation state will depend on the activities of protein kinases and phosphoprotein phosphatases.Mammalian cells express several different types of protein kinases and phosphoprotein phosphatases with more than 500 putative kinase genes in the human and mouse genome (Manning et al., 2002;Caenepeel et al., 2004).Protein kinases and phosphoprotein phosphatases have been subdivided according to their substrate specificities, activators, cofactors and/or amino acid sequence homology.It would be beyond the scope of this chapter to attempt to review them here and thus, we will briefly summarize the characteristics of the kinases and phosphatases that have most frequently been studied as possible effectors of the phosphorylation state of connexins.

Serine/threonine protein kinases
cAMP-and cGMP-dependent protein kinases (PKA and PKG, respectively) can be activated by increasing the concentration of the corresponding cyclic nucleotide (e.g., treatment with membrane permeable analogs of cAMP or cGMP such as 8-Bromo-cAMP and 8-Bromo-cGMP or forskolin, which activates adenylyl cyclase).The CAMKII isoenzymes are activated by binding of Ca 2+ /calmodulin but other protein binding partners can also regulate their activity (Griffith, 2004).Casein kinase I (CK1) is a family of monomeric serine/threonine kinases that are constitutively active.This family shows a strong preference for prephosphorylated substrates.Several inhibitors for members of this family have been described including CKI-7 and IC261 (Perez et al., 2011).Protein kinase C (PKC) has several isoforms that have been subdivided in three subtypes: conventional, novel and atypical.They differ in their activation by Ca 2+ , binding of diacylglycerol (DAG) and in their response to phorbol esters.Conventional PKCs bind Ca 2+ and DAG.Novel PKCs lack amino acids involved in Ca 2+ binding, but bind DAG.The catalytic activity of atypical PKCs is independent of Ca 2+ and DAG; these PKC isoforms do not bind phorbol esters (Newton, 1995).The phorbol ester tumor promoter, 12-O-tetradecanoylphorbol 13-acetate (TPA) and 1-oleoyl-2-acetyl-sn-glycerol (OAG), an analog of diacylglycerol have been commonly used as activators of PKC.MAPKs are subdivided in three subfamilies: the extracellular signalregulated kinases (ERKs), the c-Jun amino-terminal kinases (JNKs) and the p38 MAPKs.They are activated by protein kinase cascades [MKKK-MKK(or MEK for ERKs)-MAPK], although MKK-independent activation of p38 has been reported (Johnson and Lapadat, 2002).Finally, cyclin-dependent kinases (Cdks) constitute a family of serine/threonine kinases that regulate proliferation, differentiation, senescence and apoptosis.In post-mitotic neurons, all Cdks, with the exception of Cdk5, are silenced.

Tyrosine kinases
The tyrosine kinases can be divided in two groups: receptor tyrosine kinases (RPTKs; e.g, growth factor receptors, ephrin receptors) and non-receptor (cytoplasmic) tyrosine kinases (NRPTKs; e.g., Src, FAK, JAK).RPTKs can be further subdivided into 20 subfamilies and NRPTKs into 10 subfamilies.In the case of RPTKs, ligand-induced oligomerization and conformational changes result in tyrosine autophosphorylation of the receptor subunits which activates the catalytic activity and mediate the specific binding of cytoplasmic signaling proteins containing Src homology-2 (SH2) and protein tyrosine-binding domains.The NRPTK, c-Src, contains an SH2 domain through which it can bind to specific tyrosine autophosphorylation sites in ligand-stimulated RPTKs and mediate mitogenic signaling.c-Src can also be activated by binding to proline-rich sequences in target proteins through its SH3 domain or by dephophorylation of Tyr527 (Blume-Jensen and Hunter, 2001).The viral form of Src, v-Src, is constitutively active and oncogenic.It contains a shorter sequence at the carboxyl terminus that lacks Tyr527, which is required for inactivation.v-Src has been extensively studied in relation to connexins for its effects on gap junction function.

Serine/threonine phosphoprotein phosphatases
The phosphoserine/phosphothreonine protein phosphatases have been classified in three subfamilies (PPM, FCP and PPP).Members of the PPP (PP1, PP2A and PP2B) and PPM (PP2C) subfamilies which use a metal ion-catalyzed reaction account for most of the serine/phosphothreonine phosphatase activity in vivo (Barford et al., 1998).Several phosphatase inhibitors with different specificities are available including calyculin A (which inhibits PP1 and PP2A), cyclosporine A (an inhibitor of PP2B), FK506 (an inhibitor of PP2B) and okadaic acid (which inhibits PP1).

Phosphotyrosine phosphatases
The phosphotyrosine phosphatases (PTPs) have been classified in class I-IV based on the amino acid sequence of their catalytic domains (class I-III are cysteine-based PTPs and class IV are aspartic-based PTPs).The cysteine-based family can be subdivided in classical PTPs, dual-specificity PTPs, cdc25 PTPs, and low-molecular weight PTPs.Classical PTPs can be further subdivided into transmembrane receptor-like enzymes and intracellular nonreceptor PTPs.Eighty one of the 107 PTP genes in the human genome are active protein phosphatases (Alonso et al., 2004).

Metabolic labeling with 32 P
The first reports that demonstrated a particular Cx to be a phosphoprotein using metabolic labeling with 32 P showed phosphorylation of Cx32 in hepatocytes (treated with phorbol esters, OAG, forskolin or cAMP analogs)( (Sáez et a., 1986;Takeda et al., 1989;Sáez et al., 1990) and phosphorylation of Cx43 in uninfected and Rous sarcoma virus (RSV)-transformed fibroblasts (Crow et al., 1990).Phosphoamino acid analysis indicated that hepatocyte Cx32 and Cx43 in uninfected fibroblasts were phosphorylated on seryl residues (Takeda et al., 1989;Crow et al., 1990;Sáez et al., 1990), but Cx43 was also phosphorylated in tyrosyl residues in RSV-transformed fibroblasts (Crow et al., 1990).Using metabolic labeling with 32 P, other studies described that EGF-induced phosphorylation of Cx43 on serine residues in T51B cells through activation of mitogen-activated protein kinase (MAPK) (Lau et al., 1992;Warn-Cramer et al., 1996), FGF-2 induced phosphorylation of Cx43 in cardiomyocytes (Doble et al., 1996), tyrosine phosphorylation of Cx43 in early passage hamster embryo fibroblast (Mikalsen et al., 1997), phosphorylation of Cx56 by PKC and Cx49 by casein kinase 1 (CK1) in lens fiber cells (Berthoud et al., 1997;Cheng and Louis, 1999).In some cases, the specific phosphorylation site has been identified in reconstituted connexons expressed in Xenopus laevis oocytes.Using this approach, it has been demonstrated that v-Src induces tyrosine phosphorylation of Cx43 but not Cx32 (Swenson et al., 1990), and that serine368 of Cx43 (but not serine372) is directly phosphorylated by PKC (Bao et al., 2004a;2004b).

www.intechopen.com
Connexins as Substrates for Protein Kinases and Phosphoprotein Phosphatases 203

In vitro phosphorylation
Another widely used approach to identify putative phosphorylation sites is in vitro phosphorylation assays.In this case, a polypeptide, fusion protein or synthetic peptide (corresponding to a fragment of the connexin that includes the putative phosphorylation site(s)) is incubated with a purified protein kinase in the presence of [ -32 P]ATP and its ability to be a substrate for that protein kinase is evaluated by the incorporation of 32 P. Sáez and collaborators (1990) also performed in vitro kinase assays using the catalytic subunits of PKA, PKC or CaMK II and purified gap junctions or synthetic peptides as substrates, and compared their two-dimensional pattern of phosphopeptides with those obtained from metabolically labeled cells.Using glutathione S-transferase (GST) fusion proteins of Cx56 containing the carboxyl terminus or the intracellular loop, in vitro phosphorylation of Cx56 by PKC and PKA have been demonstrated in serine118 (in the intracellular loop) and serine493 (in the carboxyl terminus) (Berthoud et al., 1997).
Phosphorylation of Cx43 is among the best characterized.Polypeptides, fusion proteins and several synthetic peptides containing putative phosphorylation sites within the carboxyl t e r m i n u s o f C x 4 3 h a v e b e e n u s e d t o c a r r y o u t in vitro phosphorylation and identify phosphorylation sites.These experiments have demonstrated that Cx43 is a substrate of p34 cdc2 kinase (cell division cycle 2 kinase also known as cyclin dependent kinase 1) which mediates phosphorylation of Cx43 on Ser255 and possibly Ser262 (Kanemitsu et al., 1998).Cx43 is also a substrate for PKC and PKA.Kinetic analyses of wild type and mutant (S364P and S365N) Cx43 peptides (containing amino acid residues 359-376) in vitro phosphorylated by PKA and PKC have suggested that phosphorylation of Ser364 may be required for subsequent phosphorylation by PKC (Shah et al., 2002).In vitro phosphorylation of Ser365, Ser368, Ser369, and Ser373 by PKA has been described using a His-tagged Cx43-CT (containing amino acid residues E227-I382) (Yogo et al., 2006).
Other studies have shown in vitro phosphorylation of perch Cx35 by PKA and mouse Cx36 by CaMKII using fusion proteins containing the carboxyl terminus or the intracellular loop (O'Brien et al., 2004;Ouyang et al., 2005;Alev et al., 2008).A polypeptide containing the polymorphic variants S319 and P319 of the carboxyl terminus of human Cx37 (amino acid residues 233-333) was in vitro phosphorylated by glycogen synthase kinase-3 (Morel et al., 2010).In vitro kinase assays have also been used to demonstrate that phosphorylation of Cx32 by PKC prevents its proteolysis by calpains (Elvira et al., 1993).
Analyses of two dimensional maps of mixes of tryptic phosphopeptides from a connexin immunoprecipitated after metabolic labeling and from a (poly)peptide after in vitro phosphorylation together with phosphopeptide sequencing have been used often to identify the phosphorylated sites of the immunoprecipitated connexin and changes in their phosphorylation state under different experimental conditions.

Pharmacological modulation of phosphoprotein phosphatases
Changes in the phosphorylation state of Cxs can be induced by activating or inhibiting a specific intracellular phosphoprotein phosphatase.This type of approach allows identification of the protein phosphatases involved in the effects observed.

204
Using this approach, it has been demonstrated that treatment of V79 fibroblasts with several phosphoprotein phosphatase inhibitors (i.e., calyculin A, cyclosporin A or FK506) does not change the immunoblot pattern of Cx43 (Husøy et al., 1993;Cruciani et al., 1999).However, the dephosphorylation of immunoprecipitated Cx43 from TPA-exposed V79 cells is more efficiently reduced by PP2A than by PP1, PP2B or PP2C inhibitors (Cruciani et al., 1999).In WB-F344 cells, a rat liver epithelial cell line, calyculin A prevents the dephosphorylation of Cx43 induced by 18 -glycyrrhetinic acid (Guan et al., 1996).However, in primary cultures of astrocytes, calyculin A had little effect on hypoxia-induced Cx43 dephosphorylation; in this cell type, inhibition of PP2B with cyclosporin A or FK506 reduced Cx43 dephosphorylation after hypoxia (Li and Nagy, 2000).Calyculin A significantly retarded the loss of channel activity seen in ventricular myocytes in ATP-deprived conditions; conversely, stimulation of endogenous PP1 activity by treatment with p-nitrophenyl phosphate or 2,3-butanedione monoxime (a dephosphorylating chemical agent) induced a reversible interruption of cellto-cell communication (Duthe et al., 2000;2001).
The effect of okadaic acid on Cx43 also varies depending on cell type.It inhibits dephosphorylation of Cx43 in untreated and EGF-treated T5lB rat liver epithelial cells and prevents the dephosphorylation of Cx43 induced by 18 -glycyrrhetinic acid in WB-F344 rat liver epithelial cells (Lau et al., 1992;Guan et al., 1996).Okadaic acid also significantly retards the loss of gap junction channel activity seen in ventricular myocytes in ATPdeprived conditions (Duthe et al., 2000;2001).In other cell types, it has little or no effect on the immunoblot pattern of Cx43 (Berthoud et al., 1992;Husøy et al., 1993;Cruciani et al., 1999), and has little effect on hypoxia-induced Cx43 dephosphorylation in primary cultures of astrocytes (Li and Nagy, 2000).Altogether these results suggest the involvement of different protein phosphatases in the phosphorylation state of Cx43 in different cell types under various experimental conditions.

Genetic activation or inhibition of a protein kinase or phosphatase
In some studies, changes in the phosphorylation state of Cxs have been induced by genetic manipulation through chemical-induced mutagenesis of genomic DNA or transfection with mammalian expression vectors and/or infection with virus containing cDNAs coding for a protein of interest.These methods can be used to modify the kinase activity using cDNAs encoding active or dominant negative mutant forms of a specific kinase.Lampe et al. used the FT210 cell line which contains a temperature-sensitive mutant of p34 cdc2 /cyclin B kinase to demonstrate that the formation of the phosphoform of Cx43 present in mitotic cells was dependent on the activity of this kinase.
However, the two-dimensional tryptic phosphopeptide map of immunoprecipitated Cx43 from mitotic cells had many major and minor tryptic phosphopeptides that could not be attributed to direct p34 cdc2 /cyclin B kinase phosphorylation of the Cx43CT (Lampe et al., 1998).Doble et al. (2000) used transient tranfection and adenoviral infection of truncated or dominant-negative forms of PKC to demonstrate that this kinase is required for Cx43 phosphorylation in cardiomyocytes (Doble et al., 2000).
The mechanism by which v-Src affects Cx43 phosphorylation and function has been extensively explored.Several studies have shown that expression of v-Src in mammalian fibroblasts leads to phosphorylation of Cx43 in tyrosyl residues (Crow et al., 1990).Mutants of Cx43 and v-Src SH2 and SH3 domains have been used to demonstrate that the SH2 and www.intechopen.comConnexins as Substrates for Protein Kinases and Phosphoprotein Phosphatases 205 SH3 domains of v-Src interact with Cx43; the SH3 domain binds to a proline-rich motif and the SH2 domain binds to a phosphorylated tyrosyl residue in the carboxyl terminus of Cx43 (Kanemitsu et al., 1997).Two specific phosphorylation sites for v-Src have been identified in Cx43, Tyr247 and Tyr265, by stably re-expressing wild type or mutant Cx43 with v-Src in Cx43 knockout cells (Lin et al., 2001).Moreover, using a triple serine-to-alanine mutant at the MAPK sites (S255/279/282A) it has been shown that phosphorylation of Cx43 by MAPK is not required for v-Src-induced disruption of gap junctional intercellular communication (Lin et al., 2006).
Several studies have been carried out on cardiac cells.Phosphorylation of Cx43 in Ser262 regulates DNA synthesis in cardiomyocytes forming cell-cell contact (Doble et al., 2004).Expression of an activated mutant of mitogen-activated protein kinase kinase 7 (a JNKspecific upstream activator) in cultured cardiomyocytes and in the heart in vivo demonstrated that Cx43 expression is regulated by JNK, although this effect may not be mediated by direct phosphorylation of Cx43 (Petrich et al., 2002).Transgenic mice with cardiac-specific overexpression of a constitutively active form of calcineurin (a calcium-dependent serine/threonine phosphatase) showed differences in the distribution of Cx43 in the ventricles, and Cx43 was mainly present in the nonphosphorylated form (Chu et al., 2002).Overexpression of p21-activated kinase 1 (PAK1, an activator of PP2A) increased PP2A activity and induced dephosphorylation of Cx43 in rabbit myocytes and Cx43-overexpressing HEK293 cells (Ai et al., 2011).

Genetic modification of a phosphosite-specific mutant connexin
A more recent approach is the generation of connexin knock-in mice in which the coding region of the wild type protein is replaced by DNA encoding a phosphosite-specific mutant.The only available report to date using this approach showed that mice in which Cx43 was replaced by a Cx43 mutant at the CK1 sites in which serines 325/328/330 were replaced with phosphomimetic glutamic acids (S3E) were resistant to gap junction remodeling and less susceptible to the induction of arrhythmias.In contrast, mice in which a Cx43 mutant with serines 325/328/330 mutated to non-phosphorylatable alanines (S3A) was knocked-in in place of Cx43 had severe alterations in gap junction formation and function, and had a proarrhythmic phenotype (Remo et al., 2011).This report shows a mechanistic link between the phosphorylation state of Cx43 and arrhythmic susceptibility (Remo et al., 2011).

Phosphospecific antibodies
Antibodies that recognize a specific phosphorylated (or dephosphorylated) site in a connexin have been developed.These have been extensively used to identify the state of phosphorylation of the phosphosite they recognize and to determine associated changes in connexin distribution in cells under different physiological and pathological conditions.Using this approach, it has been described that ischemic preconditioning prevents the changes in the phosphorylation state of Cx43 observed in a model of ischemia/reperfusion in pig hearts (Schulz et al., 2003).It has also been reported that PKC phosphorylates Cx43 in Ser368 (Solan et al., 2003), and that scratch wounding of primary human keratinocytes causes a PKC-dependent increase in phosphorylation at this site in cells adjacent to the scratch (Richards et al., 2004).Leykauf et al. used a specific antibody against PSer279-PSer282 of Cx43 to demonstrate that different phosphorylated forms of Cx43 coexist at the plasma membrane (Leykauf et al., 2003).Two antibodies recognizing the same phosphosites were used to show that EGF and activation of its receptor with quinones induce phosphorylation of Cx43 in these serine residues (Abdelmohsen et al., 2003;Leykauf et al., 2003).Using an antibody that specifically recognizes Cx43 phosphorylated at serines 325, 328 and/or 330 (PS325/328/330), Lampe and colleagues showed that while Cx43 relocalizes to the lateral edges in ischemic hearts, Cx43 phosphorylated at these residues remained mostly at the intercalated disk (Lampe et al., 2006).An antibody that recognizes dephosphorylated Ser364/Ser365 and binds preferentially to Golgi-localized Cx43 in cultured cells has been used to demonstrate conformational changes in Cx43 (Sosinsky et al., 2007).Other studies have described that phosphorylation of connexin 43 at Ser262 is associated with a cardiac injury-resistant state (Srisakuldee et al., 2009).
Phosphospecific antibodies have been used in combination with PKC or MEK inhibitors to determine the protein kinase pathway involved in the effects observed.Sirnes et al. reported that TPA induces phosphorylation of Ser255 and Ser262 of Cx43 in a MAPK-dependent manner (Sirnes et al., 2009).A MAPK-dependent phosphorylation of serines 255, 262 and 279/282 of Cx43 has also been demonstrated using phosphospecific antibodies and a MEK inhibitor in follicles exposed to luteinizing hormone (Norris et al., 2008).In MC3T3-E1 osteoblasts, treatment with fibroblast growth factor 2 induces a PKC -dependent increase in phosphorylation at Ser368 of Cx43 (Niger et al., 2010).Solan and Lampe used several anti-Cx43 phosphospecific antibodies that recognize Src, MAPK or PKC sites and LA-25 cells (which express a temperature-sensitive v-Src) grown at the permissive and non-permissive temperatures to show that distinct tyrosine and serine residues are phosphorylated in response to v-Src activity (Solan and Lampe, 2008).Li et al. used antibodies that specifically recognize PSer110 and PSer276 in Cx35 to demonstrate that the level of phosphorylation of these serines depends on PKA activity and regulates photoreceptor coupling in zebrafish retina (Li et al., 2009).

Mass spectrometry analyses
Another technique that has been used to identify putative phosphorylation sites is mass spectrometry (MS) analysis of connexins isolated from tissue or cultured cells or in vitro phosphorylated (poly)peptides.For this purpose, the immunoprecipitated/isolated connexin or in vitro phosphorylated polypeptide is digested with a protease or a mix of proteases, the sample is enriched in phosphopeptides and subjected to MS.This technique is highly sensitive and it does not require the use of radioactivity.
The first studies using this technique to identify phosphorylation sites in Cxs were reported several years ago (Cooper et al., 2000;Yin et al., 2000).Cooper et al. showed that in vitro phosphorylation of the carboxyl terminus of Cx43 with p34 cdc2 /cyclin B kinase resulted in phosphorylation of Ser255 using liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) (Cooper et al., 2000).Yin et al. demonstrated that lens Cx45.6 is phosphorylated in the chicken lens in vivo at Ser363 using nanoelectrospray and tandem mass spectrometry (Yin et al., 2000).
Several studies using mass spectrometry analysis have been performed on Cx43.Ser364 was identified as a phosphorylation site in Cx43 using matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF MS) and LC-MS/MS (TenBroek et al., 2001).MALDI-TOF MS in combination with metabolic labeling of normal rat kidney (NRK) epithelial cells (in the presence and absence of a casein kinase 1 inhibitor) and in vitro phosphorylation of Cx43CT fusion proteins with casein kinase 1 (CK1 ) have been used to determine that serines 325, 328 or 330 are potential sites of CK1 phosphorylation in these cells (Cooper and Lampe, 2002).Cameron et al. (2003) used MALDI-TOF MS to identify Ser255 of Cx43 as the preferred site for big MAPK 1 (BMK1)/ERK5 phosphorylation.This finding was further supported by the lack of phosphorylation of GST fusion proteins containing mutant carboxyl termini of Cx43 in which Ser255 had been mutated to alanine (S255A and S255A/S279A/ S282A).Axelsen et al. (2006) reported the time course of changes in phosphorylation of Cx43 immunopurified from perfused rat hearts under non-ischemic and ischemic conditions.These authors identified thirteen phosphorylation sites using MALDI MS and LC-MS/MS in non-ischemic conditions and detected site-specific changes Table 1.Techniques used for identification of connexins as substrates for protein kinases and phosphoprotein phosphatases.

www.intechopen.com
Protein Kinases 208 in Cx43 phosphorylation during the course of ischemia.Phosphorylation of Ser365 has also been demonstrated in Cx43 immunoprecipitated from NRK cells using liquid chromatography coupled to electrospray ionization tandem mass spectrometry (LC/ESI MS/MS) (Solan et al., 2007).Fifteen putative phosphorylation sites on Cx43 have also been identified after in vitro phosphorylation of a GST fusion protein containing the Cx43CT with CaMK II by high-resolution mass spectrometry (Huang et al., 2011).
Post-translational modification by phosphate has also been identified by mass spectrometry in Cx26 and Cx32; Cx26 is phosphorylated in the intracellular loop and the second extracellular loop, and Cx32 is phosphorylated in the amino and carboxyl termini (Locke et al., 2006).Two studies have used mass spectrometry to identify phosphorylation sites in the bovine lens fiber connexins, Cx44 and Cx49.While phosphorylation sites were identified only on the carboxyl terminus of Cx44, phosphosites were identified in both the intracellular loop and carboxyl terminus of Cx49 (Shearer et al., 2008;Wang and Schey, 2009).

Luminescence resonance energy transfer
Another recent approach used to evaluate the functional effect(s) of phosphorylation of Cxs is the generation of hemichannels of known composition, stoichiometry that can be assessed by luminescence resonance energy transfer (LRET) (Bao et al., 2007).This method uses terbium ions (Tb 3+ ), which have a long lifetime emission as donor and fluorescein as acceptor.The technique is based on the detection of LRET between Cx43 subunits labeled with Tb 3+ and those labeled with fluorescein.The composition of the HCs can be determined based on the number of acceptor-labeled monomers per HC.Using HC of known composition, Bao and colleagues have determined that in a Cx43 HC all six subunits have to be phosphorylated by PKC at Ser368 to abolish sucrose permeability, although the HC pore still has a sizable diameter and allows permeation of smaller molecules (Bao et al., 2007).

Conclusions and future directions
In summary, connexins are substrates for various protein kinases and phosphoprotein phosphatases.Several of the phosphorylation sites have been identified, and the effect of phosphorylation at many of these sites on connexin channel activity has been studied.In some cases, pathophysiological conditions that alter their phosphorylation state have been reported.Although significant progress has been made in the area of connexin phosphorylation, there are many associated aspects that require further investigation.
A question that remains unanswered is whether all connexins are phosphoproteins.Does phosphorylation affect connexin channel function in all members?Does phosphorylation at a specific site induce consistent functional changes in gap junction channels and hemichannels?Or, can phosphorylation at a specific site induce changes in one channel type, and not in the other?.Because phosphorylation has been implicated in several steps of the connexin's life cycle, it is also important to determine which phosphorylation events are associated with proper trafficking to the plasma membrane, formation of gap junctional plaques or internalization and degradation.Are connexins sorted/targeted to different compartments depending on their cohort of phosphorylated sites?Where do these phosphorylation events take place?Since some hierarchy in the phosphorylation events has been shown for Cx43, it is interesting to know whether changes in phosphorylation are also associated with other post-translational modifications.Do these have a hierarchical sequence?Because connexins and changes in the activity of protein kinases/phosphoprotein phosphatases have been associated with disease, it would be important to know how the phosphorylation state of connexins is affected in disease.What are the intracellular signals and mechanisms of regulation of phosphorylation/dephosphorylation of connexins?What are the endogenous activators of the protein kinases/phosphoprotein phosphatases involved?Although the answers to some of these questions are known for some of the phosphorylation sites identified, especially in the case of Cx43, these questions have not been addressed for most connexins.