Gap Junctional Communication in Morphogenesis

Abstract

Gap junctions permit the direct passage of small molecules from the cytosol of one cell to that of its neighbor, and thus form a system of cell-cell communication that exists alongside familiar secretion/receptor signaling. Because of the rich potential for regulation of junctional conductance, and directional and molecular gating (specificity), gap junctional communication (GJC) plays a crucial role in many aspects of normal tissue physiology. However, the most exciting role for GJC is in the regulation of information flow that takes place during embryonic development, regeneration, and tumor progression. The molecular mechanisms by which GJC establishes local and long-range instructive morphogenetic cues are just beginning to be understood. This review summarizes the current knowledge of the involvement of GJC in the patterning of both vertebrate and invertebrate systems and discusses in detail several morphogenetic systems in which the properties of this signaling have been molecularly characterized. One model consistent with existing data in the fields of vertebrate left-right patterning and anterior-posterior polarity in flatworm regeneration postulates electrophoretically-guided movement of small molecule morphogens through long-range GJC paths. The discovery of mechanisms controlling embryonic and regenerative GJC-mediated signaling, and identification of the downstream targets of GJC-permeable molecules, represent exciting next areas of research in this fascinating field.

Introduction

The mechanisms that allow an embryo to reliably self-assemble the complex morphology, physiology, and behavior appropriate to its species represent one of the most fundamental and fascinating areas of research in modern science. Pattern is acquired simultaneously on three orthogonal axes, and on several logs of scale, from the molecular (e.g., cytoskeletal elements) to the organismic (graded antero-posterior (AP), dorso-ventral (DV), and left-right (LR) axial specification). Though some animals undergo a mosaic mode of development, most embryos show impressive abilities to regulate in the face of environmental or genetic perturbations. The events underlying this generation and maintenance of form require a complex web of information flow between cell and tissue subsystems during development. With the advent of multi-cellularity, the cell membrane has become a key nexus for regulatory information flow. Receptor-mediated signal exchange via secreted messenger molecules has been studied extensively. However, another important system of signaling exists: the direct cell-cell exchange of small molecules through gap junctions.

The cell biology of gap junctions has been described in a number of excellent recent reviews (Falk, 2000; Goodenough et al., 1996); the reader is also referred to the classic (Loewenstein, 1981) for a historical and detailed perspective of some of the definitive studies on gap junction properties. This extremely versatile system for communication allows for rapid synchronization among cells in a tissue and the passage of signals, both of which can be regulated at many levels. Indeed cells often up-regulate gap junctional communication when they need to share ions with their neighbors (Aslanidi et al., 1991; Larre et al., 2006). Thus, GJC is a perfect conduit for information flow during development, which depends on the ability of cells and tissues to communicate. The converse however is also paramount - embryos contain independent compartments (Rela and Szczupak, 2004; Sutor and Hagerty, 2005) that must remain isolated for proper morphology to result.

Gap junctions have been studied for years using biophysical methods (such as electron microscopy and X-ray diffraction), but some of the most exciting information on gap junctions has come more recently, from the use of molecular embryology techniques to probe the role of GJC in physiological processes (Goodenough and Musil, 1993; Lo, 1999). A number of human syndromes have been identified as mutations in gap junction genes (Maestrini et al., 1999; Richard et al., 2002), and transgenic mice are beginning to allow the molecular dissection of GJC function in these contexts (Krutovskikh and Yamasaki, 2000). For example, GJC is now known to be a general mechanism for achieving rapid syncitial communication within a tissue. Contexts include the spread of electric waves in cardiac tissue (Kimura et al., 1995; Severs, 1999) and the brain (Budd and Lipton, 1998; Momose-Sato et al., 2004; Momose-Sato et al., 2005), Ca⁺⁺ transients controlling interkinetic nuclear movement in retinal cells (Pearson et al., 2005b), and the spread of signals through gland cells to synchronize hormonal action and secretion (Meda, 1996).

Molecularly, gap junctions can consist of proteins from a number of gene families. Vertebrate gap junctions are commonly assumed to consist of connexins (Cruciani and Mikalsen, 2006; Sohl and Willecke, 2003; Sohl and Willecke, 2004). In contrast, invertebrate gap junctions are made of innexins, a family formerly called OPUS (Barnes, 1994). The ability of innexins to form functional gap junction channels has recently been demonstrated directly for a number of innexins (Landesman et al., 1999; Phelan et al., 1998; Stebbings et al., 2000). Innexins show no sequence homology to connexins but have the same topology, including four transmembrane domains (Phelan, 2005; Stebbings et al., 2002). Interestingly, this simple division does not reflect the complexity that is now becoming uncovered. First, it is becoming increasingly clear that proteins from the innexin family exist in vertebrates, prompting their proposed renaming to “pannexins” (Bao et al., 2004; Baranova et al., 2004; Bruzzone et al., 2005; Bruzzone et al., 2003; Connors and Long, 2004; Ray et al., 2005; Sohl et al., 2005). Indeed, members of these genes have also been identified in viruses (Turnbull and Webb, 2002; Turnbull et al., 2005), as well as being expressed in tissues known to rely on GJC (Dvoriantchikova et al., 2006; Ray et al., 2005; Sohl et al., 2005). Second, although connexin genes have not been demonstrated in lower invertebrate genomes, connexins from invertebrate chordates have been described (Sasakura et al., 2003; White et al., 2004).

There have even been suggestions of connexin-like proteins in plants (Meiners et al., 1991; Yahalom et al., 1991), although this is quite controversial (Mushegian and Koonin, 1993). Moreover, in hydra, an antibody generated to a vertebrate connexin appears to modulate GJC-dependent morphogenesis (Fraser et al., 1987). Hydra is known by ultrastructural methods to contain gap junctions (Wood and Kuda, 1980), which have been proposed to serve as the conduit for positional information during regeneration (Wakeford, 1979) because they form between graft and host tissue during inhibition of head regeneration by grafted head tissue. The functional data in hydra using a connexin-based reagent suggests the possibility that connexin-like motifs remain to be discovered in other invertebrate systems. Finally, there is ductin - a VATPase H⁺ pump component that appears to form a hexamer trans-membrane channels on its own in both vertebrate and invertebrate systems (Bohrmann and Lammel, 1998; Finbow et al., 1992; Finbow et al., 1991; Finbow et al., 1995; Finbow and Pitts, 1998; Inoue et al., 1999; Saito et al., 1998; Skinner et al., 1999). Ductin (like connexins) has four transmembrane domains, exists in plants and most metazoan animals, and has approximately 85% identity between insects and human. The role of ductin in endogenous GJC is currently controversial and remains an important area for investigation (Bruzzone and Goodenough, 1995; Finbow et al., 1994; Finbow and Pitts, 1993).

What sorts of signals endogenously traverse gap junctions during patterning? In general, a variety of small molecules and metabolites are thought to permeate. cAMP (Bedner et al., 2003; Burnside and Collas, 2002; Webb et al., 2002), ATP (Bao et al., 2004; Pearson et al., 2005a), and Ca⁺⁺ (Blomstrand et al., 1999; Paemeleire et al., 2000; Toyofuku et al., 1998) have been investigated in a number of physiological contexts, but endogenous morphogenetic roles for these signals have not yet been defined in the context of GJC. Interestingly, several recent reports have demonstrated that molecules much larger than the normal ≈1 kDa size cutoff can penetrate through gap junctions under some circumstances (Brooks and Woodruff, 2004; Valiunas et al., 2005b). These include siRNA’s and calmodulin; one possibility is that the crucial parameter is shape, not overall size, and that long, thin molecules may be able to traverse gap junction channels. The alignment of such molecules to facilitate GJC-mediated transfer has been suggested to be performed by an endogenous electric field (Woodruff, 2005). The issues of endogenous GJC-permeable signals and the roles of electric fields in regulating movement of signals through GJC in patterning systems will be examined in detail below. However, it is crucial that junctional selectivity can result in radically-different permeabilities of gap junctions to different types of molecules (Bevans et al., 1998; Goldberg et al., 1999; Goldberg et al., 2002; Nicholson et al., 2000). This may have important consequences for the morphogenetic system, and means that studies of this process in vivo may be strongly dependent upon the experimental probe used. For example, at 7.5 days, the mouse embryo was found to be subdivided into at least nine GJC compartments with respect to Lucifer Yellow (LY) transfer, but only two domains with respect to ionic coupling (Kalimi and Lo, 1989). In the loach (Misgurnus fossilis), LY, fluorescein, and DAPI showed consistent differences in their ability to transfer between tissues during mesoderm induction and patterning (Bozhkova and Rozanova, 2000). Moreover, the chemical selectivity of the gap junctions connecting early embryonic cells changes appreciably during loach development (Bozhkova, 1998), suggesting that embryonic patterning can utilize regulation of not only the amount of GJC but also of the various types of molecules being passed through the gap junctions. These results clearly underscore the need to explore GJC with a variety of different probes to truly understand what paths are available for endogenous small molecules and current in the embryo. Moreover, the determinants of GJC selectivity, including PKC-dependent phosphorylation and specific connexin make-up of the gap junctions (Ayad et al., 2006; Ek-Vitorin et al., 2006; Weber et al., 2004), represent important upstream regulatory factors that must be characterized during embryogenesis.

Before considering the role of GJC in complex pattern formation, it is important to be aware of gap junctional signaling’s role in two special cases: control of stem cell behavior and the misregulation of pattern that occurs during neoplastic growth (Pierce, 1983; Pierce et al., 1986; Ruiz i Altaba et al., 2004). One of the most tantalizing roles for GJC is in the processes that distinguish normal tissue from tumor cells (Krutovskikh and Yamasaki, 1997; Li and Herlyn, 2000; Loewenstein and Rose, 1992; Mesnil et al., 2005; Omori et al., 2001; Yamasaki et al., 1995). It has been shown that normal tissue generally possesses a much higher degree of GJC than tumor tissue, and a loss of GJC accompanies early steps in neoplastic transformation (Pitts et al., 1988). Moreover, a neoplastic phenotype can be induced in cell culture by ectopic closing of gap junctions using pharmacological agents or dominant-negative constructs (Omori and Yamasaki, 1998). Most interestingly, neoplastic characteristics can be suppressed by ectopic induction of GJC in tumor tissue (Hellmann et al., 1999; Mehta et al., 1991; Rose et al., 1993). More recently it has been shown that non-transformed cells can effectively normalize the conductance of gap junctions expressed by adjacent tumor cells (Valiunas et al., 2005a). These data are consistent with a role of GJC in mediating the information flow that is necessary for coordinated tissue activity and morphogenesis, which is lost in tumors (Pierce, 1983; Pierce et al., 1986; Rubin, 1985). However, in some cases, the epistasis of these events is still under investigation, since for example, data from cervical cancer progression suggests that loss of GJIC is perhaps a consequence of, rather than being causal to, epithelial dysplasia (Aasen et al., 2003).

The fine balance of proliferation and differentiation must also be maintained in stem cell populations. Gap junctions are beginning to be uncovered as an element of this process (Sheardown and Hooper, 1992; Tazuke et al., 2002; Trosko et al., 2000; Trosko, 2005; Wong et al., 2004). Interestingly, neural stem cells exhibit a unique signature based on GJC and ion transporters, and GJC based on C×43 and C×45 is essential for their survival and proliferation (Cai et al., 2004). For example, the gap junction protein Zero Population Growth (zpg) is required for germ cell differentiation in Drosophila ovary. In the absence of ZPG, the stem cell daughter destined to differentiate instead dies (Gilboa et al., 2003). Germ line stem cells differentiate upon losing contact with their niche, which likely involves GJC; as an example, it has been proposed that gap junction-mediated cAMP signaling between blastomeres and somatic cells results in changes in somatic cell gene expression (Burnside and Collas, 2002).

Overview of GJC in developmental and regenerative morphogenesis

The first step towards formulation of developmental models is knowledge of exactly what GJC paths exist among which cells in a given patterning system. Comprehensive expression analyses of gap junction genes now exist in several species (de Boer and van der Heyden, 2005; Katbamna et al., 2004; Nogi and Levin, 2005; Stebbings et al., 2002). Because of the many levels of regulation to which gap junctions are subject, the presence of connexin mRNA in a tissue is no guarantee of the existence of gap junction complexes or of functional cell-cell connectivity. Thus, some studies have addressed the localization of connexin protein by immunohistochemistry. For example, in the rabbit, C×43 and C×32 are expressed in the blastoderm at streak stages (Liptau and Viebahn, 1999). The rich regulation to which mature gap junctions are subject, and the potential for compensation among co-expressed GJC proteins, make it necessary to ascertain functional GJC in addition to mRNA and protein localization data. Direct dye coupling assays have demonstrated compartments of communication in a number of embryonic systems, including a very early observation in the squid embryo (Potter et al., 1966), the detection of ionic coupling among cells of the Triturus (Ito and Loewenstein, 1969), an investigation of electrotonic GJC in the chick blastoderm (Sheridan, 1966), and evidence for preferentially directional and pH-dependent GJC-mediated transfer in the early Xenopus embryo (Guthrie, 1984; Guthrie et al., 1988; Nagajski et al., 1989; Turin and Warner, 1980). It has also been shown using these techniques that the wing imaginal disk in Drosophila is subdivided into a number of communication compartments during differentiation (Weir and Lo, 1984).

The regulation of GJC paths by endogenous factors in embryos is key to formulating models of their role in patterning. For example, using the scrape-loading/dye transfer technique, it was shown that chick limbs exhibit a gradient of GJC along the AP axis (Coelho and Kosher, 1991). The highest GJC was observed in cells adjacent to the zone of polarizing activity, while no GJC was present at the opposite end of the limb bud. Mesenchymal tissues in the middle of the limb had an intermediate level of dye coupling and it has been hypothesized that polarizing region cells communicate to anterior mesenchyme cells via gap junctions (Allen et al., 1990). These patterns appear to be directed by FGF-4 (Makarenkova et al., 1997), while asymmetries in gap junctional coupling in the early frog embryo (Nagajski et al., 1989; Turin and Warner, 1980) appear to be controlled by members of the wnt pathway (Olson et al., 1991; Olson and Moon, 1992). Unidirectional junctions are thought to form between C×32 and C×43 (Robinson et al., 1993; Xin and Bloomfield, 1997), potentially allowing an embryo to establish one-way signaling paths. Together with pH, membrane voltage, and other asymmetries, it is clear that the potential for incredibly complex and specific signal transfer exists through GJC.

Mechanistic investigations of GJC in patterning have been fairly few (Levin, 2001; Lo, 1996; Sheardown and Hooper, 1992; Warner, 1999), but a number of early functional experiments indicated that endogenous patterns of embryonic GJC have important morphogenetic roles. Introduction of antibodies raised to specific portions of connexin proteins in mouse embryos resulted in developmental defects (Becker et al., 1995). In Xenopus, microinjection of antibodies has been reported to disrupt axial patterning (Warner et al., 1984); the treated embryos contained differentiated mesodermal derivatives such as notochord and muscle tissue and Warner et al. argued that GJC has a role in pattern formation per se, rather than in the induction of specific cell types. These are summarized in Table 1; the most data from mammalian systems are available for the limb, heart, gastrulation, and neurulation.

Table 1.

Summary of GJC involved in patterning events (only those not discussed in detail in the main text)

Morphogenetic role	References
Growth limitation	(Fujimoto et al., 2004; Jursnich et al., 1990; Roger et al., 2004; Rong et al., 2002; Tanaka and Grossman, 2001)
Signals between oocytes and follicle/nurse cells	(Bohrmann and Haas-Assenbaum, 1993; Goodenough et al., 1999; Juneja et al., 1999)
Embryonic compaction and cell adhesion	(Becker et al., 1992; Lee et al., 1987; Paul et al., 1995)
Cardiac patterning	(Ewart et al., 1997; Kirchhoff et al., 2000; Kumai et al., 2000; Ya et al., 1998)
Neural crest behavior	(Bannerman et al., 2000; Huang et al., 1998a; Lo et al., 1997; Lo et al., 1999; Sullivan et al., 1998; Waldo et al., 1999)
Limb patterning	(Allen et al., 1990; Ganfornina et al., 1999; Laird et al., 1992; Makarenkova et al., 1997; Makarenkova and Patel, 1999; Meyer et al., 1997)
Muscle/somite patterning	(Araya et al., 2005; Araya et al., 2003; Berthoud et al., 2004; De Boer et al., 2005; Mege et al., 1994; Starich et al., 1996; Todman et al., 1999; Wang, 1989)
Bone patterning	(Bani-Yaghoub et al., 2000; Chatterjee et al., 2003; Chen et al., 2004; Coleman et al., 2003; Coleman and Tuan, 2003; Davy et al., 2006; Guillotin et al., 2004; Lecanda et al., 2000; Pizard et al., 2005)
Gastrulation	(Bohrmann and Lammel, 1998; Kidder and Winterhager, 2001; Tepass and Hartenstein, 1994; Yancey et al., 1992)
Skin patterning	(Arita et al., 2004; Di et al., 2005; Richard, 2001; Risek et al., 1992; Serras et al., 1993; Watanabe et al., 2006)

In invertebrate systems, C. elegans and Drosophila have been paramount in providing examples of GJC roles. Numerous defects were observed in a C. elegans innexin-3 mutant, including rupture of hypodermis, failure of elongation, and failure of pharynx to attach to anterior (Starich et al., 2003). Likewise, epithelial organization and polarity of the embryonic epidermis is dependent on heteromerization of innexins 2 and 3 in Drosophila (Bauer et al., 2004; Lehmann et al., 2006). Establishment of the proventriculus requires Innexin-2, which is a target of Wingless signaling (Bauer et al., 2002; Bauer et al., 2001). This work is particularly interesting because it links GJC to a canonical signaling pathway. Wnt signaling is likewise upstream of GJC patterns in vertebrate systems (Ai et al., 2000; Olson et al., 1991; Olson and Moon, 1992; van der Heyden et al., 1998), suggesting possible conservation of signaling modules involving GJC. Interestingly, overexpression of C×43 provides a partial recovery of the formation of the cerebellum in Wnt-1 knock-out mice (Melloy et al., 2005), strengthening the link to the Wnt pathway and suggesting that establishing ectopic domains of cell:cell communication may be a useful modality to control certain patterning defects.

Gene targeting studies of connexins in mice have provided evidence that gap junctions are important not only for the function of the mature heart (Hagendorff and Plum, 2000; Severs, 2001; Verheijck et al., 2001), but also for septation (Kirchhoff et al., 2000), and morphogenesis of the outflow tract (Gu et al., 2003a) and endocardial cushions (Kumai et al., 2000). Many of these effects appear mediated by effects on the behavior of neural crest (Ewart et al., 1997; Huang et al., 1998a; Huang et al., 1998b; Li et al., 2002; Lo et al., 1997; Lo et al., 1999; Sullivan et al., 1998; Waldo et al., 1999; Waller et al., 2000; Xu et al., 2001), although some have argued for a role of C×43 in suppressing cellular delamination from the neural tube (Wei et al., 2004). Because of the increasing data that connexins or ductin can have functions independent of their GJC activity (Huss et al., 2002; Jiang and Gu, 2005; Kalra et al., 2006), it is important to keep in mind that genetic loss-of-function studies do not guarantee that any particular phenotype is in fact due to reduction of GJC per se. For example, in contrast to models based on GJC-permeable morphogens, Wei et al. propose that in the case of cardiac cell effects, C×43 protein directly interacts with a number of structural membrane proteins (such as β-catenin, cytoskeletal proteins, and cadherins), leading to the activation of pathways that control adhesion, motility, and cell shape (Wei et al., 2004).

How might gap junction-dependent signals control pattern formation? One mechanism is through guidance of cell differentiation (Araya et al., 2005; Araya et al., 2003; Bani-Yaghoub et al., 1999; Gu et al., 2003b; Hirschi et al., 2003; Zhang et al., 2002). Another is through the regulation of proliferation (Paraguassu-Braga et al., 2003; Pearson et al., 2005a). This is likely to mediate the stunting effects of connexin loss-of-function on craniofacial morphogenesis in the chick (McGonnell et al., 2001). Importantly, the ability of connexins to regulate progression through the cell cycle may also function through direct roles of connexins in the nucleus not dependent on GJC (Dang et al., 2003; Doble et al., 2004) and this possibility needs to be considered in genetic work on gap junction proteins that does not assay GJC directly. Finally, GJC seems to control migration in some contexts (Huang et al., 1998a; Kjaer et al., 2004; Lecanda et al., 2000; Minkoff et al., 1997; Oviedo-Orta et al., 2002).

A final but very important mechanistic role for GJC is to establish patterns of isopotential, iso-pH, and homogenous [K⁺] cell fields (Contreras et al., 1995; Fitzharris and Baltz, 2006; Potapova et al., 1990; Sherman and Rinzel, 1991). This epigenetic “prepattern” overlaid upon the embryonic morphology, working in concert with the molecular fields defined by gene expression domains, is now recognized to be an important control element in embryonic patterning. Though the field is too broad to be reviewed here, the control of endogenous bioelectric signals by GJC paths allows gap junction-permeable molecules to ultimately direct cell migration, proliferation, and differentiation (Levin, 2003a; McCaig et al., 2005; Nuccitelli, 2003a; Nuccitelli, 2003b; Robinson and Messerli, 2003).

The phenomenon of regeneration involves a profound example of pattern formation, as the system must not only build a complex structure, but unlike the case of embryonic development, must also know which components have been damaged and must be restored; in this case, the normal time-line of morphogenesis is disrupted and the system cannot rely on a pre-programmed sequence of events but must instead exchange signals to inform the various cells and tissues of the current morphogenetic state. An ideal example is provided by planaria - flatworms with three tissue layers, bilateral symmetry, and a centralized brain, possess remarkable powers of regeneration (Sanchez Alvarado, 2003; Sanchez Alvarado, 2004; Sanchez Alvarado et al., 2003; Slack, 2003; Tessmar-Raible and Arendt, 2003). After bisection, one of the edge blastemas regenerates a head, while the other forms a tail. The ability of previously-adjacent cells to adopt radically different fates highlights the necessity of long-range signaling allowing determination of position relative to, and the identity of, remaining tissue.

Following the hypothesis that GJC may underlie this signaling, Nogi et al. (Nogi and Levin, 2005) cloned and characterized the expression of the Innexin gene family during planarian regeneration. Planarian innexins fall into 3 groups according to both sequence and expression, and the concordance between expression-based and phylogenetic grouping suggests diversification of 3 ancestral innexin genes into the large family of planarian innexins. Innexin expression was detected throughout the animal, as well as specifically in regeneration blastemas, consistent with a role in long-range signaling relevant to specification of blastema positional identity as well as signaling among cells local to the regenerating tissue. Exposure to a GJC-blocking reagent that does not distinguish among gap junctions composed of different Innexin proteins resulted in bipolar (2-headed) animals, demonstrating the alteration of a high-level morphogenetic signal (“build a head”, as opposed to “build a tail” in the posterior-facing blastema).

Taken together, the expression data and the re-specification of the posterior blastema to an anteriorized fate by GJC loss-of-function suggest that innexin-based GJC mediates instructive signaling during regeneration in planaria. Because of the availability of a simple assay in a clear-cut example of GJC-dependent patterning, and the existence of large amounts of classical as well as molecular and genomic data, this system is very likely to allow important progress on the mechanisms utilized in this spatial regulation. Although the pharmacological GJC blockers, not being subject to compensation or redundancy effects, allowed the rapid identification of an important phenotype that might have been missed if single innexin genetic loss-of-function experiments were pursued, it is now crucial to utilize RNAi-based gene abrogation to determine precisely which innexins are responsible for mediating the anterior-posterior polarity signals, and use their expression pattern and biophysical properties to formulate testable mechanistic models of how large-scale polarity is imposed on tissue.

GJC and left-right asymmetry

A considerable amount of data on the function of GJC in vertebrate patterning, including evolutionary conservation in two species, is now available in the field of left-right (LR) axial patterning. The vertebrate body-plan is based on a bilaterally-symmetrical structure; however, the visceral organs and brain display marked and consistent asymmetries in their location or geometry with respect to the embryonic midline. Left-right asymmetry is a fascinating example of large-scale embryonic patterning and consistent laterality of vertebrates raises many deep theoretical issues linking molecular stereo-chemistry with multi-cellular pattern control (Burdine and Schier, 2000; Levin, 2005). Because no macroscopic force distinguishes right from left, a powerful paradigm has been proposed to leverage large-scale asymmetry from the chirality of sub-cellular components (Brown et al., 1991; Brown and Wolpert, 1990). In this class of models, some molecule or organelle with a fixed chirality is oriented with respect to the antero-posterior and dorso-ventral axes, and its chiral nature is thus able to nucleate asymmetric processes. Thus, the first developmental event that distinguishes left from right is generally thought to take place on a subcellular scale. However, it is now known that a pathway of multi-cellular fields of asymmetric gene expression directs the laterality of asymmetric organs (Levin, 1998; Yost, 2001). A mechanism must then exist to transduce subcellular signals to cell fields, and thus convert information about LR direction to that of a cell’s position with respect to the midline.

Levin and Mercola tested the hypothesis that GJC could play a role in the mechanism by which asymmetry at the level of a cell can be transduced into embryo-wide asymmetry of gene expression (Levin and Mercola, 1998a; Levin and Mercola, 1999). The initial hint of long-range signaling was provided by the observation that removing a distal lateral portion of the chick blastoderm causes ectopic gene induction in cells on the opposite side of the embryo (Fig. 1). Following initial observations on GJC in frog embryos (Guthrie et al., 1988; Warner et al., 1984), we pursued this question in embryos of Xenopus laevis. Using injections of a system of a small junctionally-permeable fluorescent dye together with a large molecule junctionally-impermeable dye (to mark the injected cell and rule out false positives due to cytoplasmic bridges and incomplete cell cleavage, Fig. 1G), we showed that there is indeed an asymmetry, with a zone of isolation across the ventral midline and good junctional coupling on the dorsal side (Fig. 2A). Misexpression of connexins lacking a regulatory region (C×26) can reveal induction of GJC in sectioned embryos (Fig. 1H), countering previous proposals that the changes in early frog embryo GJC, due to various treatments and previously reported by a number of labs (Brizuela et al., 2001; Guthrie, 1984; Guthrie and Gilula, 1989; Guthrie et al., 1988; Nagajski et al., 1989; Olson et al., 1991; Olson and Moon, 1992), were artifacts caused by lens flare effects from adjacent blastomeres and by pigmentation differences between dorsal and ventral Xenopus cells (Landesman et al., 2000).

Figure 1.

Evidence for long-range GJC in chick and frog embryos determining left-right asymmetry

Chick embryos cultured as whole blastoderms (A) exhibit the normal left-sided expression of the markers Nodal (B) and Sonic hedgehog in Hensen’s node (C), indicating normal asymmetry. In contrast, when the distal left side of the embryo is cut off before culture (D), cells on the far right side begin to express Nodal (E) and Sonic hedgehog expression in the node is bilateral (F).

Panels A and D are schematics of st. 2 chick embryos (primitive streak stage); the region encompassed by the square indicates the part that was cultured. The purple stain in panels B-F is signal, indicating expression of the relevant gene marker via in situ hybridization. Red arrowheads indicate expression; white arrowheads indicate absence of expression.

In frog embryos, small molecule fluorescent probes are able to penetrate several cell diameters through gap junctions (G). Panel G shows four large blastomeres of the 16-cell embryo. The left-most cell (indicated by red arrow) was injected with a fluorescent small molecule tracer dye, which is seen spreading to all 4 of its neighbors in a clock-wise direction. Panel H shows, in section, induced GJC (via injection of C×26 mRNA) in early frog embryos. The transfer of Lucifer Yellow (< 1kD) but not rhodamine-labeled dextran (RLD, =10 kD) in section demonstrates true gap junctional communication and rules out cytoplasmic bridges and incomplete cleavages.

Figure 2.

Serotonin movement through gap junctions during LR patterning

(A) One model consistent with the data is that in Xenopus, the long-range GJC path around the ventral zone of isolation results in a LR-asymmetric distribution of a small molecule morphogen (yellow). (B) The same model can be superimposed on the chick, where connexin43 is expressed circumferentially around, but not in, the primitive streak. (C) Serotonin is initially homogenously distributed in the early frog embryo. Within the next few hours, serotonin appears as a circumferential gradient around the zone (D) and ultimately coalesces into a single cell on one side of the midline (E). If the gap junctions are blocked, serotonin is unable to move (F). The frog embryo stages shown are 16-32-cell stages in panels A,D,E,F, and 2-cell stage in panel C. The chick embryo in panel B is at the early primitive-streak stage.

The asymmetry in ability to transfer dye could be altered in predictable ways by chemical agents that are known to regulate junction permeability. Using dominant negative and constitutively-active connexin constructs, it was shown that interruption of the circumferential GJC path, as well as introduction of ectopic open gap junctions across the zone of isolation, both randomized early asymmetric gene expression and position of the asymmetric viscera in frog embryos (in the absence of other defects). Similar data were obtained in the chick - an embryo with very different gastrulation architecture - using antisense oligonucleotides and a variety of surgical manipulations to interrupt the GJC zone (Fig. 2B). Taken together, the results indicated that a zone of circumferential Connexin43 expression around the primitive streak (which does not express C×43) is required for normal laterality.

While much remains to be learned about the individual connexins functioning in the frog embryo and the permeabilities of the native gap junctions for different permeant molecules (Dicaprio et al., 1975; Landesman et al., 2000; Landesman et al., 2003), these experiments revealed a common system functioning in both species. Thus, a very similar picture emerged of the role of GJC in LR patterning in two embryonic systems. In both chicks and frogs, a circumferential large-scale pattern of GJC exists around a zone of isolation. The contiguity of this path is crucial for normal LR expression of asymmetric genes; interruption of this path by surgical, pharmacological, or molecular-biological methods specifically causes left-right randomization. Conversely, the zone of isolation is likewise crucial for LR patterning. Introducing GJC through this zone randomizes gene expression, probably because it can no longer serve as a barrier for accumulation of LR morphogens on either side of the midline.

The main features of this system appear to be the same in birds and amphibians; however, several differences have been found. Firstly, due to the different modes of gastrulation, the circumferential path is overlaid differently on the two embryos’ body-plans: it is perpendicular to the DV axis in chicks, but to the animal-vegetal axis in Xenopus. Also, the zone of isolation appears to be maintained at the mRNA level in chick (based on the exclusion of C×43 from the streak). In contrast, Xenopus appears to rely on a protein-level mechanism to exclude junctional transfer from the zone of isolation because drug agents that open existing gap junctions are effective in inducing GJC on the ventral side in Xenopus embryos.

The differences in GJC observed in chick are LR-symmetrical at the mRNA level. Likewise, the differences in GJC in Xenopus are along the dorso-ventral axis, downstream of the Wnt pathway (Olson et al., 1991; Olson and Moon, 1992). So, how may these asymmetries of GJC be translated into differential signaling to the left vs. right-sided cells? One model that is consistent with these data is a radial movement of small molecule morphogens through the GJC path, allowing them to accumulate on one side of the midline. If the net movement was biased (clockwise or counterclockwise), such sorting of a small LR morphogen through gap junctional paths would lead to the formation of a concentration gradient across the primitive streak, which could then induce asymmetric gene expression cascades in conventional ways (Levin and Mercola, 1998b).

The investigation of GJC in chick and frog asymmetry raised two main lines of inquiry. First, what controls the unidirectional (chiral) flow of LR information through the gap junctions? We have previously proposed (Levin, 2003b; Levin and Nascone, 1997) that asymmetries in electrical polarization among L vs. R cells (Levin et al., 2002) can result in an electrophoretic force that can drag certain small molecules through gap junctional paths in a preferentially-chiral direction.

Second, what is the molecular nature of the small-molecule LR signals which are exchanged between cells on the L and R sides? The ideal candidate would be smaller than the size cut-off of gap junctions (< 1 kDa), be water-soluble (signaling molecules such as retinoic acid do not need gap junctions to move between cells), and be charged (to enable regulation of movement via ion pump-dependent voltage gradients that have been shown to be required for normal asymmetry (Adams et al., 2006; Levin et al., 2002). Serotonin fits these criteria, has been demonstrated to go through gap junctions in some contexts (Wolszon et al., 1994), and offers the benefits of a well-developed pharmacological tool-set (Gaster and King, 1997).

Serotonin is thus an ideal candidate for an early LR signal. Thus, taking advantage of the large number of well-characterized reagents available to test and characterize its role in left-right asymmetry in chick and frog embryos, serotonin’s role in asymmetry, prior to the formation of neurons, was recently probed in chick and frog (Fukumoto et al., 2005a; Fukumoto et al., 2005b). Using analysis of endogenous localization of serotonin and its receptors, as well as gain- and loss-of-function experiments using pharmacological serotonergic blockers and dominant-negative and wild-type expression constructs, it was shown that serotonergic signaling through receptor subtypes R3 and R4 was crucial for normal asymmetry. In early Xenopus embryos, serotonin was distributed in a striking radial pattern, eventually forming a gradient and coalescing into a single blastomere on one side of the midline by the 32/64-cell stage (Fig. 2C-E). While direct movement of serotonin through gap junctions in this system has not been demonstrated in vivo (serotonin is too small to be modified by fluorescent tags without significantly changing its properties), it was crucially shown that the ability of serotonin to localize asymmetrically through the GJC-connected blastomeres was dependent upon open gap junctions (Fig. 2F) and upon the function of the H,K-ATPase and V-ATPase ion pumps (Adams et al., 2006; Fukumoto et al., 2005b). Moreover, evidence was presented for a novel intracellular locus of serotonin activity - this is a requirement of the model since serotonin arriving within the target cells’ cytosol has to be able to activate receptor mechanisms there. This is an important general feature of GJC-dependent signaling since intracellular receptor mechanisms must be characterized for gap junction-dependent morphogens.

Models of Morphogen Movement through GJC paths

Taken together, these data suggested a model (Fig. 3) that provides a possible answer to the chirality of the morphogen movement through gap junctions in left-right patterning (Esser et al., 2006): that serotonin moves asymmetrically through the field of GJC-connected cells under an electrophoretic force provided by differential membrane voltages in cells at opposite ends of the circumferential cell field (Fukumoto et al., 2005b; Levin, 2003b; Levin et al., 2006). One unique aspect of this dataset is that most of the important properties are known or can be estimated quantitatively. This allowed a mathematical model to be formulated (Esser et al., 2006), which tested and supported the hypothesis that the known voltage difference across the GJ-coupled cell field can actually generate a significant LR morphogen gradient in the developmentally-available time, using an electrophoretic mechanism. The computer simulation used realistic values for the physical constants related to serotonin, cytoplasm, etc. to demonstrate that electrophoresis of small molecules across GJ-coupled cell fields (Fig. 3A) can give rise to significant gradients of the small molecule (Fig. 3B) under physiological voltage differences. Moreover, analysis of the model revealed that not only are chemical gradients formed across the entire GJ-coupled cell field, but also are present in a complex pattern across each individual cell (Fig. 3C), providing a potential mechanism by which cells in large GJ-coupled fields can derive directional information.

Figure 3.

A model of electrophoretic movement of morphogens through gap-junctional paths

(A) Schematically, a number of patterning systems can be visualized as a field of gap junctions terminating on a cell group at one end that establishes a strong polarization by ion exchange with the outside world. Gray level within the cells illustrate voltage gradient produced by pumps at the edge. (B) This results in an electrophoretic force that can, given realistic estimates of the physical parameters of cytoplasm and small molecules, result in a significant gradient of a small molecule across the cell field (yellow dots in panel A; this is now thought to be serotonin in the case of frog left-right patterning). (C) Importantly, mathematical analysis of the model illustrates that local gradients of GJC-permeable morphogens are thus set up within cells as well as across cell fields, potentially providing local directional cues.

This class of models makes a number of specific predictions. First, the influx of the LR morphogens into cells (as distinct from the well-understood serotonin receptor activation on the cell surface) has to control cell fate. This appears to be borne out in other systems, as the function of the SERT importer is involved in tumor growth (Nordenberg et al., 1999; Serafeim et al., 2002), as is serotonin (Dizeyi et al., 2004). An important aspect is the identification of gene transcription modulated by the intracellular arrival of GJC-dependent morphogens into cells. While these studies are just beginning in our lab in the context of the asymmetric gene cascade, others have begun to characterize, using transcriptome analysis, the genetic targets of gap junction-mediated signaling (Iacobas et al., 2005; Iacobas et al., 2004; Iacobas et al., 2003). In some cases, GJC modulates transcription by altering Sp1 and Sp3 transcription factors (Stains and Civitelli, 2005; Stains et al., 2003).

It is now abundantly clear that large number of important gene networks are potentially controlled by GJC signaling and these represent excellent targets for future molecular dissection. For example, recent analysis links the expression of Syndecan-2 with connexin43 in mouse embryos (Iacobas et al., 2005), which is the first glimpse of conservation of regulatory circuits among the known roles of GJC and Syndecan-2 in lower vertebrates and mammals in establishment of left-right asymmetry (Fukumoto and Levin, 2005; Kramer et al., 2002; Kramer and Yost, 2002). Finally, the use of this mechanism in various patterning systems must include significant filtering and selectivity among the gap junctions, since tissues cannot afford all charged molecules to be moved in a particular direction - this force must be reasonably specific for the intended morphogens, and such developmental selectivity is supported by direct measurements (Bozhkova, 1998; Bozhkova and Rozanova, 2000; Goldberg et al., 2002; Suchyna et al., 1999).

An alternative set of models based on reaction-diffusion systems (Kondo, 2002; Takagi and Kaneko, 2005) are suggested by recent data on zebrafish pigmentation. The zebrafish leopard mutation gives rise to distinct pigmentation changes that are likely to reveal reaction-diffusion as the mechanism controlling pigment pattern (Asai et al., 1999). Interestingly, it has recently been discovered that the mutation is in connexin41.8 (Watanabe et al., 2006), suggesting that in this case of pigment system morphogenesis, the relevant signaling molecules may be moving through gap junctions. Theoretical discussion of this possibility has suggested the GJC-permeable molecules cAMP and ATP as the relevant Turing couple (Schiffmann, 1991).

The Future of GJC in Patterning

The characterization of GJC in embryonic patterning is currently at a very exciting stage; molecular tools now exist to probe every aspect of this type of intercellular information exchange in contexts where it is known to be important. It is expected that newly-developed confocal microscopy uncaging and FRET methodologies will greatly facilitate the study of endogenous GJC paths in embryos in vivo (Bedner et al., 2003; Braet et al., 2003; Cannell et al., 2004; Dakin et al., 2005; Di et al., 2005). Key future research areas include the characterization of factors which set up patterns of differential GJC in various embryonic tissues (at the transcriptional level, as well as at the level of controlling GJC states, such as by endogenous patterns of pH and voltage gradients (Calero et al., 1998; Ek-Vitorin et al., 1996; Francis et al., 1999; Gu et al., 2000; Morley et al., 1997), and the mapping of paths which exist in and between different tissues to molecules of various charges and sizes. It is already known that gap junctions are selectively permeable based on a complex function of molecular weight, size, net charge, shape, and interactions with specific connexins (Ek-Vitorin and Burt, 2005; Goldberg et al., 2004). Particularly interesting is the possible role of chemically-rectifying (unidirectional) gap junctions in embryogenesis (Robinson et al., 1993; Xin and Bloomfield, 1997; Zhang et al., 2004). A better understanding of endogenous GJC-permeable molecules will be crucial.

The zebrafish will likely emerge as an excellent embryonic system for studying GJC since the embryos are transparent, allowing easy dye-based analysis of GJC paths in vivo as well as use of more sophisticated pH and voltage-sensitive dyes for studying the spatial distribution of factors which control junctional gating. Expression of C×43.4 has already been followed in zebrafish using a GFP fusion construct (Essner et al., 1996). The amenability of the zebrafish to genetic screens may also be used to analyze mutants in junctional transfer, by designing lines containing fluorescent small molecules.

Experiments in mouse (and other systems) must, of necessity, be performed in the context of multiple connexin knock-out, since it is becoming increasingly clear that due to compensation and redundancy, single loss of gap junction genes can fail to reveal important phenotypes (Simon et al., 2004). Indeed, it is now known that ablation of Connexin43 can lead to up-regulation of a pannexin (Iacobas et al., 2006), suggesting that cell-cell communication can be restored in a regulative manner by alterations of transcription of GJC genes across families. Knock-in of dominant negative mutants with different specificities for endogenous connexin families is likely to reveal many important and novel roles (Beahm et al., 2006; Fiorini et al., 2004; Paul et al., 1995). Likewise, more sophisticated technologies allowing expression of dominant negative mutants restricted in space and time during development will allow the circumvention of embryonic lethal phenotypes and likely to lead to the discovery of novel patterning mechanisms (Bakirtzis et al., 2003; Becker et al., 1992). While we now have some understanding of downstream transcriptional changes that occur when gap junctional signaling is perturbed (Kim et al., 2005), microarray analysis of GJC-inhibited tissues may lead to the discovery of proximal early-response genes to GJC-permeable morphogens. Our lab is currently pursuing such efforts in the context of embryonic left-right patterning.

The ultimate proof of mechanistic understanding of a process is the ability to synthesize the elements, characterized quantitatively, back into a predictive computer model that can make and facilitate the testing of specific predictions. Theoretical analyses of GJC signaling have begun via mathematical models (Cooper, 1984; Cooper et al., 1989; Fortier and Bagna, 2006; Hofer et al., 2002; Nygren and Giles, 2000; Vogel and Weingart, 1998; Vogel and Weingart, 2002) and in future work it will be very important to apply these to the understanding of specific patterning processes in embryonic and regenerative morphogenesis (Esser et al., 2006). It is probable that the understanding of gap junctional signals will prove as important to basic developmental and evolutionary biology as has been the explosion in receptor-mediated signaling. The implications for the ability to control pattern formation are profound, and we are only beginning to glimpse the elusive, subtle, yet crucial molecular cross-talk between the interior machinery of cells.