1. 0 INTRODUCTION During the course of evolution of single cells to multicellular organisms, new genes and functions were developed to assure coordinated regulation of cellular processes within these more complex entities. In multi-cellular organisms a delicate balance of regulatory and integrative mechanisms has to be achieved in order to maintain a relatively stable internal environment in the face of variations in the external conditions. This state of equilibrium, termed homeostasis, is controlled by three major communication processes: 1) extracellular communication, 2) intracellular communi-cation, and 3) intercellular communication. The integrated control of these three communication routes permits an organism to proliferate, to differentiate, to undergo apoptosis (programmed cell death) and to adapt / respond to stimuli (Trosko and Ruch, 1998). It has been suggested that the gain in ability of more complex organisms to regulate cell growth, differentiation and apoptosis is directly related to the appearance of intercellular communication.
This notion is supported by several studies that have shown that this type of communication is severely impaired in cancers, which often exhibit behavior characteristic of a single-cellular organism. Processes typical to these organisms such as uncontrolled cell proliferation, inability to terminally differentiate and an altered apoptosis rate are often found in cancerous cells (Trosko and Ruch, 1998). Besides growth control, intercellular communication plays an essential role in a variety of physiological processes including electrical coupling, tissue response to hormones, and regulation of embryonic development. The diversity and the vital character of these functions for the survival of an organism clearly demonstrate the importance of understanding the mechanisms involved in the control of intercellular signaling. Intercellular communication is mediated through the gap junction communication channel, which has become one of the focal points of the research in the field with a particular emphasis on the regulatory aspects of gap-junctional intercellular communication (GJIC). The insight gained from these studies can lead to the development of new therapeutic approaches to a variety of patho-physiological states caused by abnormal intercellular communication.
1. 1 The Gap Junction Communication Channel 1. 1. 1 Gap Junctions Gap junctions are sites between adjoining cells that allow for direct intercellular passage of molecules smaller than 1 kDa. These channels are relatively non-specific and the observed selectivity is based mainly on molecular size, allowing the passive diffusion of small molecules and ions such as water, cAMP, sugars, nucleotides, amino acids, fatty acids, small peptides and drugs, but blocking the transfer of proteins, complex lipids, polysaccharides, nucleic acids and other large molecules. Recent studies have challenged the apparent non-specificity of GJs hinting at more complex regulatory mechanisms behind the control of GJ permeability to certain compounds.
Gap junctions are found in almost all multicellular organisms of the animal kingdom, both invertebrates and vertebrates, and in almost all tissues with the exception of a few terminally differentiated cells such as erythrocytes, lymphocytes, skeletal muscle and adipocytes (Kumar and Gilula, 1996). An integral membrane protein called a connexin is the main structural unit of the gap junction channel (Fig. 1, A). Connexins comprise a multi-gene family with at least twenty mammalian connexin genes discovered to date. The expression of a particular connexin type is strictly correlated with the cell type and tissue in which it is found. Gap junctions have a relatively simple molecular organization: they display a hierarchy of assembly with six connexin subunits forming an integral complex called a connexon (Fig.
1, B). The connexons from the plasma membrane of the neighboring cells align and interact to form the gap junction channel (Fig. 1 C). In turn, gap junctions cluster in specific regions resulting in the formation of gap junction plaques (Fig. 1 D). Figure 1.
The Structure of a Gap Junction. Schematic model of gap junction showing the arrangement of six connexin subunits to form a connexon, which contains the central channel connecting the cytoplasm of the two adjoining cells. All connexins appear to have the same topology in the plasma membrane with the polypeptide spanning the lipid bilayer four times, with both N- and C- termini located in the cytoplasm. One of the four transmembrane domains, M 3, is amphipatic in nature and contributes to the lining of the channel. The two extracellular loops (E 1 and E 2) (Fig. 1 A) negotiate the docking of the two opposing connexons by facilitating the initial interactions between them.
Each loop contains a characteristic arrangement of three cysteine residues that are thought to enhance the rigidity of these extracellular segments necessary for the docking of the two connexons and formation of a gap junction. The intracellular loop between the transmembrane domains M 2 and M 3 as well as the C-terminus exhibit the greatest variability among different connexins and therefore are believed to be important for regulation of the protein (Kumar and Gilula, 1996). The X-ray crystallographic study of a recombinant gap junction membrane channel lacking its C-terminus has revealed that the outer diameter of a GJ within a membrane has a diameter of ~ 70 A with the intercellular portion of the channel narrowing to ~ 50 A (Unger et al. , 1999). The vertical cross-section of the interior of the channel disclosed that its diameter narrows down from ~ 40 A to ~ 15 A within its transmembrane region in proceeding from the cytoplasmic to the extracellular portion of the bilayer.
The channel widens again to ~ 25 A within its extracellular section. Taking into account the contributions from the amino acids side chains, the narrowest diameter of the GJ is about 5 A (Unger et al. , 1999). 1. 1. 2 Assembly and Degradation Similarly to other integral membrane proteins, “connexins are synthesized at the endoplasmic reticulum (ER) membrane in a typical process that involves a signal recognition particle, an integral signal anchor sequence, docking of the nascent-chain / ribosome complex to the translocon, and co translational integration of the connexin polypeptides into the ER membrane” (Falk, 2000).
The final topological organization of the connexins is attained at the level of their ER membrane integration (Falk et al. , 1994; Ahmad et al. , 1999; Diez et al. , 1999).
The exact location at which connexin oligomerization takes place has not been clarified. Several studies have reported that connexins are assembled into oligomers within specialized region of the ER (George et al. , 1998, 1999), either the intermediate compartment or ER-Golgi Intermediate Compartment (Schweizer et al. , 1990).
However, other studies in NRK and CHO cell lines suggested that oligomerization of endogenously expressed Cx 43 takes place after exit from the ER in late Golgi membranes (Musil and Goodenough, 1993). These observations strongly suggest that the process is likely cell – type specific. Connexins are trafficked to the plasma membrane by a consecutive series of vesicle fusion and budding from the ER through the Golgi apparatus by following the general secretory pathway (Musil and Goodenough, 1991). The process of targeting connexons to the plasma membrane is not currently understood but two general models have been proposed. The first involves the direct transport of the connexon to the GJ site whereas the other entails an indirect transport of the connexon by lateral movement within the plasma membrane towards the gap junction site following its insertion at a distant location (Falk, 2000).
Once localized into the plasma membrane, the connexons from the neighboring cells dock with each other via interactions of their connexin subunit extracellular loop domains forming a complete GJ channel. GJ channels in turn aggregate to form GJ plaques. The formation of plaques is assisted by calcium-dependent cell-adhesion molecules such as E-cadherin’s (Laird et al. , 1995; Laird 1996). The cytoskeletal proteins ZO-1 and ? -spec trin have also been reported to play a role in this process (Toyofuku et al. , 1998).
The degradation and turnover of gap junctions are the least understood aspects of connexin processing within the cells. The turnover of connexins is relatively fast compared to other integral membrane proteins with the reported half-lives in the range of 1 – 3 h for Cx 43 and Cx 45 in cultured cells or tissues (Beardsley et al. , 1998; Saf fitz et al. , 2000) contrasted with 17 – 100 h for integral membrane proteins measured in hepatocytes (Chu and Doyle, 1985). The current evidence suggests that gap junctions are removed from the membrane by internalization of entire channels within large fragments of GJ plaques (Severs et al. , 1989).
The subsequent degradation process involves both lysosomal as well as proteasomal pathways (Laing and Beyer, 1995; Laing et al. , 1997; Laird, 1996; Rahman et al. , 1993; Musil et al. , 2000).
1. 1. 3 Gap Junctions and Their Role in Human Diseases Gap junctions play an essential role in a number of diverse physiological processes. Charcot-Marie-Tooth (CMTX) disease was the first pathology associated with a connexin.
This progressive neuropathy results from myelin disruption and axonal degeneration of peripheral nerves and is caused by a mutation in Cx 32 (Martin et al. , 2000). This connexin has a wide tissue distribution and is one of the major GJ proteins in liver as well as in oligodendrocytes. The fact that a mutation in Cx 32 results in only a mild neuropathy suggests that connexins can compensate for the loss of other connexins in different tissues. Other diseases associated with mutations in various connexins include dominant and recessive hearing loss (Cx 26, Cx 31, Cx 30), dominant epidermal disease (Cx 31, Cx 30. 3, Cx 30), dominant skin disease caused by abnormalities in keratin ization, cataracts (Cx 46, Cx 50), and abnormal cardiac development (Cx 43) (Table 1) (Bennett, 1994; Kelsell et al.
, 2001). Aberrant regulation of GJIC has been implicated in several types of neoplasia (Andrade-Rozen tal et al. , 2000). The majority of tumors express less connexin, have fewer gap junctions and exhibit a decrease in GJIC (Ces en-Cummings et al. , 1998; Yamazaki, 1990). Table 1.
Human diseases associated with mutations in connexin genes (adapted from Kelsell et al. , 2001) Connexin (gene) Human disorder Cx 32 (GJB 1) X-linked Charcot-Marie-Tooth Disease: neuropathy often associated with hearing loss Cx 26 (GJB 2) Dominant and recessive non-syndromic moderate-profound sensorineural hearing lossDominant epidermal disease (palmoplantar keratoderma) and Vohwinkel’s syndrome Cx 31 (GJB 3) Recessive non-syndromic moderate-profound sensorineural hearing lossDominant non-syndromic high frequency hearing lossDominant skin disease (Erythrokeratoderma variabilis) Dominant sensorineural hearing loss and neuropathy Cx 30. 3 (GJB 4) Dominant skin disease (Erythrokeratoderma variabilis) Cx 30 (GJB 6) Dominant hearing lossClouston’s hidrotic ectoderm al dysplasia: skin disease (palm-plantar keratoderma), hair loss, nail defects and often mental deficiency Cx 43 (GJA 1) Association with visceroatrial heterotaxy Cx 46 (GJA 3) Cx 50 (GJA 8) Dominant zon ular pulver ant cataract 1. 2 REGULATION OF GAP JUNCTIONS 1. 2.
1 General Aspects of GJ Regulation GJ regulation is a complex process that can be achieved by a broad range of mechanisms that can generally be categorized into two major groups. The first category includes all the processes that affect the formation and levels of GJs starting with Cx gene expression through translation, oligomerization into connexons, transport to the membrane, assembly into GJ and plaques, gating and finally degradation. The impaired GJIC caused by down-regulation of Cx in certain types of neoplasia is illustrative of this type of regulatory mechanisms. The second group includes all the modes of regulation that do not alter the cellular levels of GJ but affect its functionality, which is directly related to the fluctuation of the channel between the ‘open’ and ‘closed’s tates.
This GJ gating can be effected by a variety of exogenous and endogenous factors including extracellular soluble factors (mitogen’s, hormones, anesthetics and drugs) and biomolecules (oncogenes, growth factors, tumor promoters) (Berthoud et al. , 2000; Evans and Boitano, 2001). These agents can exert their GJIC modulatory action either directly by introducing structural modifications to the channels, or indirectly for example by perturbing the lipidic environment in the vicinity of the channel and thus affecting its conformation (Hossain and Boynton, 2000). The majority of GJIC modulators act by initiating complex signaling pathways leading to the activation of various kinases, phosphatases and adaptor proteins. 1. 2.
2 Regulation of Gap Junctions by Phosphorylation of Connexins Most connexins (including Cx 31, Cx 32, Cx 37, Cx 40, Cx 43, Cx 45, Cx 46, Cx 50, and Cx 56) contain protein kinase “consensus phosphorylation sequences” and have been demonstrated to be phosphoproteins (Lampe and Lau, 2000). The majority of the phosphorylation events occur on the C-terminal cytoplasmic tail, which has been shown to be important for regulation (Fig. 2). It is interesting to note that the Cx 43 truncated mutant that lacks a C-terminal retains its ability to form functional channels in Xenopus oocytes or SKHep 1 cells, albeit with altered permeability and electrophysiological properties compared with those of a wild type Cx 43 channel (Fishman et al. , 1991; Dunham et al. , 1992).
A Cx 32 mutant that is truncated at its C-terminal has been linked to CMTX disease (R abadan-Diehl et al. , 1994). The above examples emphasize the important regulatory role of the C-terminal domain in channel gating and possibly in the rates of connexin transport, assembly and turnover. A variety of protein kinase modulators have been applied to investigate the effects of phosphorylation on gap junction function, revealing that phosphorylation events are connexin and cell-type specific. However, some general trends have been observed across different connexins. For example, compounds that increase levels of intracellular cAMP tend to enhance GJIC.
This observation has been attributed to increased rates of connexin transcription, increased levels of connexin RNA or protein (Darrow et al. , 1996), changes in connexin phosphorylation status (Traub et al. , 1987), as well as changes in Cx 43 transport (Atkinson et al. , 1995) and GJ conductance (Spray et al.
, 1991; Jong sma et al. , 2000). On the other hand, agents that activate PKC such as phorbol esters generally diminish GJIC. Staurosporine, a protein kinase inhibitor, has been shown to increase GJIC in SKHep cells expressing Cx 43, whereas okadaic Figure 2.
A schematic representation of the primary structure of rat Cx 43. Phosphorylation sites in Cx 43 targeted by V-Src, MAP kinase and PKC are indicated. (Borrowed from Lampe et al. , 2000) acid, a phosphatase inhibitor, has had the opposite effect on GJIC (Moreno et al. , 1994). These studies taken together demonstrate that phosphorylation is an important regulatory mechanism for GJ function.
However, the signaling pathways responsible for these observations have not yet been fully deciphered. 1. 2. 3 Phosphorylation of Cx 43 Cx 43 is the most common connexin expressed across a broad range of cell lines and tissues. The phosphorylation of Cx 43 can have both an inhibitory and stimulatory effect on GJIC. The primary sites of Cx 43 phosphorylation are multiple serine residues (21 residues) on its C-terminal domain (Fig.
2). The extent of phosphorylation on threonine residues is significantly smaller (Kanemitsu et al. , 1997). Cx 43 phosphorylation on tyrosine has been reported only in cells expressing oncoproteins such as v-Src or v-Fps (Kurta and Lau, 1994).
The protein kinases PKC, MAP kinase and pp 60 src were shown to phosphorylate Cx 43 (Lau et al. , 2000). Many other protein kinases could also be involved in Cx 43 phosphorylation. Cx 43 extracted from untreated mammalian cells and separated by SDS-PAGE can generally be resolved into several bands corresponding to its different isoforms. Typically observed are a nonphosphorylated (NP ~ 42 kD) form that migrates faster on the gel and two slower bands that correspond to the most common phospho isoforms (P 1 and P 2, ~44 kD and ~46 kDa respectively), which have been shown to be phospho-related mainly on serine residues (Musil and Goodenough, 1991). However, neither the protein kinase (s) isoforms responsible for these phosphorylations nor their target sites in untreated cells have been identified.
1. 2. 3. 1 Phosphorylation of Cx 43 During Its Life Cycle Phosphorylation events have been associated with the localization of Cx 43 to the plasma membrane as well as with aggregation of connexons into GJ plaques, pointing to the possible role of this post-translational modification in the regulation of Cx 43 transport and / or the assembly or disassembly of GJs (Musil and Goodenough, 1991).
Phosphorylation has been demonstrated to occur as soon as 15 min after Cx 43 synthesis in cultured fibroblast cells, clearly showing that it can occur before the protein’s arrival at the membrane (Crow et al. , 1990). The fact that a nonphosphorylated form of Cx 43 can be detected at the membrane suggests that either these early phosphorylations are transient in nature or that dephosphorylation occurs at the membrane (Musil and Goodenough, 1991). Although Cx 43 does not have to be phosphorylated in order to be localized to the membrane, Cx 43 present in GJ plaques has been found to be phosphorylated almost exclusively to its P 2 form and selectively resistant to solubilization by Triton X-100 (Musil and Goodenough, 1991). However, it is not known if these phosphorylation (s) play a role in the formation and stabilization of GJs in plaques.
1. 2. 3. 2 Phosphorylation of Cx 43 by PKC and PKA The stimulation of PKC by the short exposure to phorbol ester TPA results in a marked decrease in GJIC, as indicated by a reduction in dye transfer in different cell types (Rived al and Ops ahl, 2001). PKC can phosphorylate Cx 43 on S 368 and S 372 in vitro (Saez et al. , 1997; Lampe et al.
, 2000). It has also been shown that S 368 is a major site of phosphorylation in vivo (Lampe et al. , 2000). The cells expressing the site-directed mutant S 368 A were resistant to dye transfer inhibition by TPA. Examination of the channel conductance in cells treated with TPA has further supported the notion that the PKC-mediated phosphorylation of S 368 in Cx 43 is involved in the mechanism of GJIC inhibition by TPA since the S 368 A mutant (unlike wild-type Cx 43) showed no decrease in the full open conductance state of GJ channels (Lampe et al. , 2000).
PKA does not appear to phosphorylate Cx 43 in vitro. However, compounds that stimulate an increase in cAMP levels generally enhance Cx 43 intracellular levels. Additionally, studies in which PKA or PKC were introduced into L 929 cells transfected with either wild type or S 364 P mutant Cx 43 cDNA have implied that both kinases may play a complex role in GJIC regulation (Britz-Cunningham et al. , 1995). 1. 2.
3. 3 Phosphorylation of Cx 43 by Tyrosine Protein Kinases Both EGF and PDGF receptor tyrosine kinase activation through ligand binding negatively affected GJIC causing a rapid inhibition of dye transfer and a significant increase in the phosphorylation of Cx 43 on serine residues (Kanemitsu and Lau, 1993; Hossain et al. , 1998). The effect of EGF receptor on Cx 43 GJIC was shown to rely on activation of MAP kinase but not to be related to PKC activity (Kanemitsu and Lau, 1993). MAP kinase seems to phosphorylate Cx 43 directly in EGF-treated cells by targeting Ser 255, Ser 279, and Ser 282 residues (Warn-Cramer et al. , 1996; 1998).
A triple mutant of these three residues formed functional GJs but was not affected by MAP kinase in response to EGF-stimulation (Warn-Cramer et al. , 1998). Evidence supporting the role of MAP kinase in the regulation of GJIC also comes from experiments in which purified MAP kinase was demonstrated to phosphorylate Cx 43 channels reconstituted into liposomes and to induce a decrease in their conductance (Kim et al. , 1999). The inhibition of GJIC by activation of the PDGF receptor involves more complex mechanisms and is less understood. The process appears to depend on both PKC and MAPK activity and may be regulated by other signaling pathways in addition to Cx 43 phosphorylation (Hossain et al.
, 1998; 1999 a; 1999 b). The potential sites of serine phosphorylation following stimulation of the PDGF receptor, the effective downstream serine protein kinase (s), and potential other regulatory protein (s) have not been determined. Activation of non-receptor cytoplasmic kinases, such as pp 60 src, potentiates phosphorylation of Cx 43 on tyrosine residues, resulting in a marked inhibitory effect on GJIC. The v-Src tyrosine kinase seems to phosphorylate Cx 43 directly in vitro and in vivo (Loo et al. , 1995). Cx 43 has been shown to interact directly with v-Src tyrosine kinase through the interactions between phospho tyrosine sites and proline-rich regions on Cx 43 and the SH 3 and SH 2 domains on v-Src respectively (Kanemitsu et al.
, 1997). Phosphorylation of Y 265 and possibly Y 247 may play a role in the observed impairment of GJIC. The current body of evidence strongly supports the importance of the Y 265 residue in this process as a Y 265 F mutant was not able to associate with the activated pp 60 src kinase in Xenopus oocytes although it formed functional channels. Another model of GJIC disruption by v-Src kinase has been recently proposed based on findings that are in direct conflict with the above mentioned results.
In these studies, the activation of v-Src affected a Y 265 F mutant by decreasing junctional conductance in Xenopus oocytes (Zhou et al. , 1999). Additionally, the ability of v-Src to inhibit GJIC was eliminated by a treatment with the MEK inhibitor, PD 98059, implying that phosphorylation on a serine rather than a tyrosine residue can be of significance (Zhou et al. , 1999).
MAP kinase may be a potential enzyme through which v-Src exerts its action although further studies are needed to clarify the mechanisms involved (Zhou et al. , 1999). 1. 2. 3. 4 Dephosphorylation of Connexins by Protein Phosphatases The role of protein phosphatases in the regulation of gap junctions is only beginning to emerge.
The recent studies of selective inhibition of different classes of protein phosphatases in stimulated cells have highlighted the importance of these enzymes in the regulatory nature of the connexin phosphorylation-dephosphorylation cycle. Okadaic acid, an inhibitor of protein phosphatases PP 1 and PP 2 A, has been reported to stimulate the phosphorylation of Cx 43 in cultured rat myocytes, a process affecting GJ communication as indicated by an accompanying decrease in the unitary channel conductance (Moreno et al. , 1994). In another study the treatment of EGF-stimulated rat liver epithelial cells with okadaic acid slowed down the recommencement of GJIC and the dephosphorylation of Cx 43 (Lau et al. , 1992). Similarly, the usually observed dephosphorylation of Cx 43 upon treatment with 18-glycyrrhetinic acid was prevented in the presence of okadaic acid (Guan et al.
, 1997). The inhibitors of PP 2 B protein phosphatases exhibited properties similar to those of PP 1 and PP 2 A classes in that they retarded the restoration of GJIC and prevented the dephosphorylation of Cx 43 in cells treated with TPA (Cruciani et al. , 1999). However, only PP 2 B class inhibitors were able to induce the above effect in cells that were subjected to hypoxia (Li et al.
, 2000). Although the above evidence strongly supports the function of protein phosphatases in GJIC control, little is known about the exact identity of the enzymes involved. In addition, most of the studies to date have examined the impact of phosphatase inhibition in stimulated cells. The insight into the role of phosphatases in GJIC regulation in untreated cells is lacking mostly due to the fact that phosphatase inhibitors such as okadaic acid failed to alter GJIC in several unstimulated cell lines (Lau et al. , 1992; Guan et al. , 1997; Saez et al.
, 1993). This implies a more subtle level of control that protein phosphatases maintain on GJIC regulation in unstimulated cells. It is also reasonable to expect that the mechanisms of phosphatase function in the regulatory process are cell type specific. In summary, the following aspects of the role of the Cx phosphorylation/ dephosphorylation in the regulation of GJIC have been established (Lampe and Lau, 2000): 1.
After the arrival of a connexon at the plasma membrane, Cx phosphorylation is not always essential for the formation of a functional GJ channel, 2. phosphorylation of specific residues (Ser or Tyr) reduces GJIC, implicating the “gating” of GJ channels, 3. phosphorylation may facilitate protein-protein interactions important for Cx regulation, 4. phosphorylation is important for Cx processing, i.
e. trafficking, assembly and / or turnover, 5. dephosphorylation by protein phosphatases may be important for modulating the effects induced by phosphorylation.