Electrical excitability involving characteristic sets of voltage-dependent ion channels is an essential property of neurons working as the elements of a highly efficient biological information processor, the brain. Thus the development of electrical excitability or expression of voltage-dependent ion channels is one of the fundamental aspects of neural differentiation. In addition, because various muscular cells, many endocrine cells (236), and recently so-called nonexcitable cells, such as glial cells (5), are all known to express voltage-dependent ion channels at certain stages of cellular differentiation, general mechanisms specifying cell types may also be understood by analyzing expression processes of ion channels in differentiating cells.
Membrane differentiation is one of the key mechanisms in any type of cellular differentiation by which individual cells can be distinguished from each other through the recognition by their cell surfaces at various stages of development. Because so many molecular species have recently been identified in the plasma membrane, each type of cells is likely to be specified by a characteristic combination of functionally significant molecular species expressed on their membranes. One such example is the gene family of various voltage-dependent ion channels, which are distinctly identified by both molecular biological and electrophysiological methods (96). Thus, because of their molecular and functional distinctions, ionic channels can be useful markers for cellular differentiation not only in neural tissues but in other tissues as well.
Neurogenesis in vertebrate embryos is known to be initiated by the commitment of dorsal ectoderm under inductive influence of mesoderm into neural plate formation at the beginning of gastrulation (84, 130, 288). Although ectodermal cells forming the neural plate differentiate into neuroepithelial cells, the neural plate transforms into the neural tube as a result of morphogenesis. Neuroepithelial cells further differentiate into various kinds of neuronal and glial cells under the influence of morphogenic fields formed with neural and surrounding tissues (107, 120). For example, motoneurons in the vertebrate spinal cord are induced by diffusible factors from the notochord situated closely underneath the ventral side of the cord (327), and type 2 astrocytes in the rat optic tract are known to be derived from O-2A precursor cells by the inductive influence of growth factor, ciliary neuron growth factor (CNGF), released from neighboring cells (158).
Cellular differentiation is generally determined by cytoplasmic factors successively transferred from the ancestral oocyte through antecedent parent cells, or induced by factors that are produced by neighboring embryonic cells or delivered by diffusion from a more distant part of the embryo. The former is known as autonomous differentiation, and the latter as inductive differentiation (48, 49). Thus morphogenesis, which is a result of migration or segregation of cells committed to various extents, establishes in turn the field of cell-cell interaction provoking new commitment of embryonic cells to specified differentiation pathways or the segregation of cytoplasmic factors into various embryonic compartments. These morphogenetic cycles, setting as both result and cause of cellular differentiation, occur repeatedly until all embryonic cells have finally differentiated. This introduces difficulties to the analysis of the mechanism of cellular differentiation during embryogenesis in vivo. Therefore, in this review, we confine the discussion to results from simple differentiating systems in which autonomous and inductive processes are experimentally separable. In these cases, cellular differentiation can be analyzed independently of a clearly definable morphogenetic field.
The fact that the larva of the ascidian, a protochordate, has a simplified body organization that is essentially identical to that of the vertebrates was discovered in the mid 19th century by Kowalevsky (139). Since then, the ascidian embryo has been considered to be a prototype of the vertebrate embryo (Fig. 1, D and E). On the other hand, at the beginning of this century, the cell lineage studies by Conklin (Fig. 1C) revealed a fate map on the ascidian egg and provided the classical example of a mosaic egg (30, 31).
The hypothesis that cytoplasmic factors buried in the matrix of egg cytoplasm can be segregated into various presumptive regions by cleavage and commit these regions to various cell fates has been invoked to explain autonomous determination during the development of mosaic eggs (47). The existence of cytoplasmic factors in the ascidian embryo was ingeniously proven by Whittaker (318). Whittaker (318) showed that acetylcholinesterase, one of the characteristics in striated muscle fibers of the ascidian tadpole, is actually expressed in the blastomeres containing presumptive muscular regions when cleavage of the embryo is arrested early in development, at the 8- to 64-cell stages. This report has been further supported by the analysis of other characteristics performed in our experiments, such as various kinds of electrical excitability (299). Distinctive forms of excitability appear in specific cleavage-arrested blastomeres as the result of expression of Na+, K+, and Ca2+ channels in their membranes.
However, even in this ascidian embryo, a classical example of mosaic development, Rose (255) and Reverberi and Minganti (246) reported that the embryo from which presumptive notochordal blastomeres were excised, leaving the presumptive neural blastomeres intact at the eight-cell stage, cannot differentiate neural tissues, such as brain vesicle, otolith, and eye. They suggested that inductive effects from the presumptive notochordal regions are required to generate neural tissues in the ascidian embryo, just as in the vertebrate neurogenesis (247).
On the basis of these classical observations, we first identified autonomous differentiation of respective blastomeres isolated from the cleavage-arrested 1- to 16-cell ascidian embryo (225) and then examined the alteration of developmental fates by contact with the blastomere of another type (Fig. 2) to demonstrate the inductive differentiation of neural characteristics represented by electrical excitability (224). We found that the isolated anterior animal blastomere from the eight-cell cleavage-arrested embryo, which is known to include presumptive brain vesicle region, autonomously develops long-lasting Ca2+-dependent action potentials. It was then found that the action potential was derived from the expression of slowly inactivating Ca2+ channels characteristic of epidermal differentiation of the ascidian larva (99, 225). However, when it is cultured in contact with the anterior vegetal blastomere, which is known to include presumptive notochordal region, and raised as a contacted two-cell system, the same anterior animal blastomere now develops a typical Na+ spike by expression of Na+ channels and tetraethylammonium (TEA)-sensitive delayed rectifier K+ channels (Fig. 2) (223, 224).
In this review, we focus particularly on the early developmental processes of various ion channels in neuronal and other excitable membranes observed in this extraordinarily simple interacting two-cell system and compare these results with those in other significant and definable systems for neural differentiation.
B. Cell Lineage of Ascidian Embryos
Classical cell lineage studies on ascidian embryos by Conklin (31) early in this century revealed that each blastomere is uniquely tissue specified at the 64-cell stage. Deletion of a part of an embryo resulted in an imperfect embryo lacking the tissues involved in those presumptive regions, indicating that the ascidian embryo is a typical example of mosaic eggs with a fairly discrete fate map (30, 32).
The cell lineage in the ascidian embryo was classically established by tracing segregation of endogenous colored granules distributed differentially in various regions of cytoplasm in eggs of some ascidian species, such as Styla partia (31). For example, orange/yellow-colored granules that are inferred to be mitochondial aggregates are initially located within the presumptive muscular regions in the posterior vegetal quadrant of the egg and finally segregated into muscle cells in the tail of the tadpole larva (245). With five distinct types of colored granules, neural, muscular, mesenchymal, ectodermal, and endodermal regions can be traced distinctly (31).
Modernized cell-tracing techniques with horseradish peroxidase (HRP) reveal the precise cell lineage of the Halocynthia embryo up to the 110-cell stage at which point all blastomeres are committed to give rise to single tissue types during subsequent development (206, 208, 209). The results confirm most of the classical studies of Conklin (31); however, some important details concerning the caudal muscular lineage and neural lineage in dorsal part of the neural tube in the tail have been revised (206). The muscular lineage has been also studied by fluorescent probes for the mitochondial membrane, such as rhodamine 123 or merocyanine 540, confirming the revised lineage revealed by HRP injection (333).
Early ascidian cleavages are principally symmetrical across the mid-plane through the vegetal-animal axis, which is clearly indicated by the existence of two polar bodies at the animal pole (31, 261, 262).
Morphological studies of cleavage in early ascidian embryos are described with optical microscopic observations by Conklin (31). On Halocynthia roretzi embryos, Satoh and co-workers (261, 263) and Nishida (205) precisely described the observations with scanning microscope until the late gastrula stages.
At the eight-cell stage, four relatively smaller animal blastomeres are located slightly anteriorly against four larger vegetal blastomeres, and thus both anteroposterior and vegetal-animal axes become apparent (Fig. 2). Until the 32-cell stage, all cell divisions are synchronized, and after this stage, cell division in the vegetal hemisphere becomes relatively asynchronous (31, 205).
At the 64-cell stage, the embryo becomes disklike in shape, being compressed along vegetal-animal axis. At the 110-cell stage, the vegetal half starts to hollow slightly, indicating initiation of the morphogenic movement of gastrulation.
During gastrulation, the group of animal cells most anteriorly located comes up dorsally and overlays the anterior vegetal group of cells that is in the presumptive notochordal region and responsible for neural inducing activity. The mesodermal region responsible for muscular tissue and mesenchyme is enclosed inside the blastopore and becomes located posterolaterally and ventrally. Thus the locations of various morphogenic regions, such as ventral ectoderm, neural plate, dorsal mesoderm, myogenic region, and endoderm, in this ascidian gastrula, show close similarity to those of vertebrate and particularly amphibian gastrulas (31, 139).
Morphogenic movement after gastrulation was reexamined by tracing cell lineage of Halocynthia embryos with injected HRP (206) or in Ciona embryos by direct observation with Nomarski optics (201, 202).
The cell-cleavage pattern and cell maps of the nervous system have been described in great detail, reviewed in Satoh (262). The cell lineage of all cells in the neural plate of Ciona intestinalis was especially determined in detail (201). The neural plate has been shown to be derived from 10 cells at the 64-cell stage. Of these, four derive from vegetal (dorsal) hemisphere and the remaining six cells from a row at the anterior animal border. After the 64-cell stage, three generations advance from 7th to 10th during gastrulation, and there are 75 neural plate cells at the start of neurulation in C. intestinalis. This number is extremely small when compared with other vertebrates, such as Xenopus embryos. By the time the neural tube becomes completely internalized, 170 cells of 10th to 12th generation exist in the tube, and only one further division or at most two are required to complete the neural cell lineage (201).
The nervous tissue in the ascidian tadpole has been recently studied in detail by serial section of epoxy-embedded materials (203). Characteristic features of a brain or a sensory vesicles, a visceral ganglia, and the neural tube in the tail have been described by optical and electron microscopic studies, and their homology to the tubular nervous system characteristic of vertebrates has been established (75, 126, 139).
Putative sensory cells in the adhesion processes at the rostral extreme possibly send axons through the ventral side of the brain vesicle to the visceral ganglion. The brain or sensory vesicle is an anterior counterpart of neural tube. The dorsoanterior wall of the vesicle consists of a thin monolayer of neuroepithelial cells, and the ventral and posterior walls include a single mechanosensory cell of the otolith (52, 55, 305, 306) and several photocells associated with a single pigment cell (4, 55), respectively. They send their axons to the visceral ganglion located just posterior to the sensory vesicle.
Phototransduction potentials recorded from ascidian tadpole photocells show hyperpolarizing responses to light-on stimuli, just as in the case of vertebrate photocells (71). It is of interest in terms of vertebrate evolution that the photoresponse is hyperpolarizing in contrast to the depolarizing response observed in the invertebrates, such as crustacean photocells. Morphologically, ascidian photocells are regarded as modified ciliary cells, just as in the case of rod or cone cells of the vertebrate retina (52, 55, 71).
The visceral ganglion is believed to include several motoneurons that are located ventrally and send axons down to the tail along the neural tube (203). Quite recently, the lineage of motoneurons was clarified in Halocynthia larvae, and the derivation from the anterovegetal blastomeres (A-line cells) in the eight-cell embryos was established, as described later (222). These motoneurons may be responsible for the rhythmic activity of six lines of serially connected striated muscles in the tail. Neuromuscular junctions on the striated muscle fibers have been proven to be cholinergic just as in vertebrate skeletal muscle cells (221). There are a number of sensory cells in the epidermal cell layers in the tail of the larva (305).
Among the cellular populations in neural tissue, two-thirds of cells in the sensory vesicle and the visceral ganglion of the trunk and most cells in the neural tube of the tail are not differentiated neurons but are neuroepithelial or ependymoglial cells with cilia protruding into the lumen of the tube (38, 126). Thus the total number of mature neurons in the central nervous system (CNS) is only ~70 in the larva of C. intestinalis (203). The same structure of the nervous system has been described in various species, including both solitary and colonial ascidian larvae. The precise number of neurons in the CNS of the larva has not been reported, except for those in the CNS identified morphologically in the case described above (203).
After the metamorphosis of the larva, the sensory vesicle, visceral ganglion, and neural tube all degenerate. The undifferentiated cell mass of the cerebral ganglion situated anterodorsally to the sensory vesicle becomes the hypophysis and esophageal ganglion in adult ascidians, regulating contraction of the muscular wall (11).
In summary, the neural tissue of the ascidian tadpole larva is the simplest CNS in an ancestral form of vertebrate CNS and its development provides a good model for vertebrate neurogenesis. As for the ascidian neurogenesis as a prototype of vertebrates, we hope the reader will also refer to a previous review by Okamura et al. (230).
C. Development of Excitability in
Cleavage-Arrested Embryos
Differentiation without cell cleavage was first described by Lillie (157) for Chaetopterus eggs treated with concentrations of K+ greater than 100 mM in seawater, which causes activation without ensuring division. The cilia that characterize the ectoderm of trochophore larvae of Chaetopterus pergamentaceus are observed on the surface of cleavage-arrested, one-cell embryos at the same developmental time as in intact embryos.
It was suggested that segregation of cytoplasmic components, such as ectoplasm and yolk-containing endoplasm, are produced equally in cleavage-arrested embryos, and thus differentiation occurs without cellular compartmentalization (57, 58, 157). The increased intracellular concentration of Ca2+ due to sustained depolarization by K+ may be responsible for activation of Chaetopterus eggs, but is not necessary for differentiation without cleavage after activation (17).
In ascidian embryos, Whittaker (318) achieved cleavage arrest with the cytokinesis inhibitor cytochalasin B and demonstrated the expression of tissue-specific enzymes in the blastomeres, including respective presumptive regions (318). His experiment is considered to provide evidence for the existence of cytoplasmic factors in tissue determination (318).
We intended to utilize cleavage-arrested embryos for analysis of membrane differentiation, especially the development of excitability, which consists of sequential expression of various ion channels, since large cells without complex cellular processes are ideal preparations for analysis of ion channel currents with voltage clamp. When Halocynthia embryos are cleavage arrested at the 16-cell stage and cultured in seawater containing cytochalasin B until control embryos hatch, the blastomere membranes reveal electrical excitability, generating Na+ and Ca2+ spikes (99, 299).
The Ca2+ spikes found in the presumptive muscular blastomeres in the cleavage-arrested embryo are identical to those in muscle cells previously observed in the control larva (186). Calcium spikes in muscle cells have been confirmed by whole cell recording techniques applied to dissociated muscle cells from intact young larvae of other ascidian species (275). From these studies, it seems clear that electrical excitability represented by combinations of ion channels can be used to identify types of differentiation in distinct blastomeres in cleavage-arrested ascidian embryos.
The observation that Na+ spikes appeared only in the presumptive neural region in cleavage-arrested embryos led us to consider Na+ spikes as neural and Ca2+ spikes as muscular markers in the ascidian larva, respectively.
However, further studies on the cleavage-arrested embryo revealed that Ca2+-dependent action potentials appeared also in ectodermal blastomeres destined to become epidermal tissues (97, 99). Voltage-clamp studies of ectodermal blastomeres reveal that the ectodermal blastomeres show slowly inactivating Ca2+ channels without concomitant delayed rectifier K+ channels, in contrast to the case of muscularly differentiated blastomeres. Ectodermal blastomeres generating Ca2+ action potentials also bind monoclonal antibodies directed to the larval epidermis (98, 223). Thus it is concluded that the long-lasting Ca2+ action potentials in these blastomeres are different from the sharp Ca2+ spikes in muscle cells of the intact tadpole larva and characteristic of larval epidermis.
The long-lasting action potential in the epidermis has been reported in other ascidian larvae by Mackie and Bone (164), and action potentials propagating over the body surface are observed in amphibian epidermal cells at the hatching stage of tadpoles (253, 260).
Tables 1 and 2 illustrate the correspondence of developmental fate with the differentiation of specific classes of ion channels (99). It is clear that each blastomere has a tendency to express a particular tissue type, dependent on time of cleavage arrest and location of the blastomere in the early embryo. When the same blastomere shows more than two types of differentiation, the phenotype of each blastomere seems not to be mosaic of multiple types of characteristics but to have selected one of the possible types. The exclusivity principle in differentiation has been discussed by Weiss (314).
Interesting patterns of differentiation are obtained by arresting cleavage before the first division, in which situation all cytoplasmic factors, if any, exist in one cellular compartment. It is expected that all types of tissue differentiation may occur after culturing in cytochalasin B-containing seawater. However, Crowther and Whittaker (37) have reported that all tissue types appeared to be expressed in the one-cell embryo in terms of structures visualized by electron microscopy and have suggested that multiple expression is in accordance with mosaic development of ascidian eggs. In the case of Halocynthia one-cell embryos, epidermal characteristics are dominantly expressed in terms of ion channel expression, whereas other characteristics are barely expressed (98, 225). The binding of antiepidermal antibodies also reveals exclusivity of epidermal differentiation (177, 211, 212), and there is evidence that other types of differentiation, such as muscular characteristics, are suppressed (212, 225). In Caenorhabditis elegans embryos, it is also known that the cytochalasin B-treated, one-cell embryos express always epidermal characters, whereas descendant cells in an untreated embryo give rise to other tissue types (35).
In contrast, it has been reported that a fraction of cleavage-arrested, one-cell Ciona embryos evidently express the muscular characteristics as assessed by binding of antimuscular antibodies (212). However, it is still possible that muscular and epidermal characteristics are essentially mutually exclusive; exclusivity mechanisms may be present even in case of Ciona, although the suppressive mechanism is not always complete and leaves only a dominant tendency to either muscular or epidermal characteristics. Actually, in the cleavage-arrested, one-cell Ciona embryo, the two muscle characteristics acetylcholinesterase and tropomyosin-containing filaments tend to appear simultaneously (36). Thus separate sets of muscle characteristics may be regulated by the same regulatory mechanism, probably through a single multigene regulatory factor, such as MyoD (313), as suggested by Crowther et al. (36), whereas genes necessary for other tissues may be suppressed as groups.
Thus suppression mechanisms must be an important feature of cellular differentiation. Because a type of electrical excitability that is produced by a specific combination of ion channels can be equivalent to a set of differentiation markers that must be regulated as a group of genes, cleavage-arrested blastomeres examined by electrophysiological methods provide unique opportunities to analyze intracellular mechanisms of determination by selective suppression.
A. Oocyte Maturation, Fertilization, and Cytoplasmic Movement
When the oocyte membrane is situated at the starting position of differentiation or null point of development, several studies demonstrate the existence of electrical excitability similar to that observed in differentiated excitable membranes (87, 196). Thus, to study cellular differentiation in terms of electrical excitability, it is necessary to first understand various changes in ion channels of egg membranes occuring at the time of maturation or fertilization.
Both membrane and cytoplasm of the oocyte may undergo considerable modulation during maturation, after which the oocytes are arrested at the first or second metaphase of meiosis, depending on animal species, and ready for fertilization by sperm (68).
Upon fertilization, meiosis is mostly resumed, and the male and female pronuclei are fused (68). Developmental changes in the membrane ion channels begin at the time of fertilization. Attachment of sperm induces a long-lasting depolarization called fertilization or activation potential in various animal species, such as sea urchins, ascidians, and amphibians, possibly because of membrane permeability changes introduced by the explosive rise in intracellular Ca2+ concentration (26, 72, 269). After the fertilization potential is completed, the membrane is repolarized.
The characteristic Ca2+ waves and Ca2+ pulses observed correspond to various biological phenomena, such as resumption of meiosis followed by extrusion of polar bodies, ooplasmic segregation (286, 287), and mitosis followed by cell cleavage (316). Thus the increase in intracellular Ca2+ is one of the major causes of ion channel changes triggered by fertilization.
The primary drive of oocyte maturation has been recently clarified in terms of metaphase promoting factors (MPF) and cytostatic factors (CSF) arresting the mitotic process at metaphase, such as cdc2/cyclin B and CSF (27, 131, 316). In addition, involvement of intracellular Ca2+ in this maturation process has been suggested in starfish eggs and some mammalian eggs (54, 100, 127). Suppression of the changes in intracellular Ca2+ with Ca2+-chelating agents can arrest meiotic maturation at specific stages from germinal vesicle breakdown (GVBD) to metaphase of the second meiotic division in pig oocytes (127). An inositol 1,4,5-trisphosphate (IP3)-induced Ca2+ release mechanism develops during maturation of starfish and hamster oocytes (28, 64). Calcium release mechanisms become more and more sensitive to either mechanical or chemical agents. However, this may indicate that the oocyte is ready to increase its intracellular Ca2+ concentration at the time of fertilization (19, 114, 296).
In terms of electrical properties, the oocyte membrane becomes less permeable during maturation by losing K+-permeable channels, such as inward rectifier K+ channels or outward rectifier K+ channels (191, 192, 195), being more sensitive to depolarization (196). In starfish oocytes, the density of Ca2+ channels is known to increase during periods of oocyte growth before meiosis but remains relatively constant during meiotic maturation, this being contrasted to the case of K+ permeability. These changes may also derive from those in intracellular Ca2+ status during maturation. However, the essential requirement of an intracellular Ca2+ increase for maturation has been questioned and may be a result of hormonal application at least in starfish eggs (34, 319).
One of the significant results of the increase in intracellular Ca2+ at the fertilization is thought to be the production of changes in electrical potential of the egg membrane: the fertilization potential. However, the rapid rise of the earliest part of fertilization potential is not because of intracellular Ca2+ increase, but to Ca2+ spikes evoked by a transient depolarization at the time of sperm attachment: the sperm attachment potential (40, 46, 72). Calcium influx through open Ca2+ channels may contribute to the initial increase in intracelluar Ca2+, at least underneath plasma membrane; however, it is known that egg activation can proceed without the initial Ca2+ influx under voltage clamp (26). In almost all species of animal eggs, the existence of various types of Ca2+ channels has been confirmed.
The middle and later phases of the fertilization potential may be caused partly by a Ca2+-activated nonselective cationic conductance or by IP3-sensitive cation channels. In this phase, the coincidence between changes in potential and intracellular Ca2+ is significant in both sea urchin and ascidian eggs (18, 46). Calcium influx through these nonselective cation channels may contribute to the steady-state increase in intracellular Ca2+ (73). The depolarization produced by the fertilization potential, including both the early rapid phase and the long-lasting phase, is known to function as a fast mechanism of polyspermy block in various animal species, including echinoderms, Chaetopterus, ascidians, and amphibians (109).
Apart from the membrane properties, the most obvious intracellular Ca2+-dependent change in the case of ascidian eggs is cortical ooplasmic segregation (18, 286). The segregation is fairly important for cytoplasmic determinants of later cellular differentiation. After the initial, large, transient increase in intracellular Ca2+ concentration, exocytosis of cortical granules occurs in various other species of eggs, but not in the ascidian (334). The Ca2+ increase starts at the site of sperm attachment and propagates over the whole egg cortex as a Ca2+ wave, being followed by a wave of exocytosis in the sea urchin egg (308). In ascidian eggs, the exocytosis of cortical granules is rarely observed; instead, a cortical cytoplasmic movement follows Ca2+ waves triggered by the initial transient increase in intracellular Ca2+ (286). In this cytoplasmic movement, a remarkable contraction of the cortical layer produces a new distribution of cytoplasmic components in the egg. Remarkably, the mitochondria-rich layers are transferred first to the vegetal pole and then to the posterior equatorial region, referred to as phase I and phase II (119, 259, 266). In other species, more or less cytoplasmic movement may contribute to the segregation of ooplasmic factors after activation. The involvement of an increase in intracellular Ca2+ in exocytosis is suggested by analogy to the common secretory process (9, 115). Cytoplasmic movement generated by microfilament-related structures is also considered to be promoted by an actin-myosin interaction triggered by Ca2+ increase (266). In ascidian eggs, phase I is sensitive to cytochalasin B, indicating a microfilament-related process, whereas phase II is sensitive to colcemid, indicating a microtubule-dependent process (265, 266).
B. Ion Channels in Oocytes: Ca2+, Na+, and Inward and Outward Rectifier K+ Channels
The ion channels that contribute to potential changes in oocyte or egg membranes, as described above, have been extensively analyzed in various animal species, including ascidians, echinoderms, amphibians, and mice (87, 196). Voltage-dependent Ca2+ channels, voltage-dependent K+ channels including inward rectifier, and in some cases Cl
channels are known to exist in most immature oocytes. Voltage-dependent Na+ channels appear to be unique to ascidian oocytes (12, 227, 228), although tetrodotoxin (TTX)-sensitve Na+ channels have been recently reported in Xenopus oocytes (16).
Ascidian oocytes are shown to have a characteristic N-shaped nonlinear current-voltage (I-V) relation produced by inward K+ rectifier channels and outward K+ rectifier channels (66, 189, 190, 219). The nonlinear I-V relation due to the inward K+ rectifier is also observed in most other marine eggs, including echinoderms (89, 90, 187, 194), annelids (88), and coelenterates (92).
In starfish oocytes, the changes in expression of inward rectifier K+ and A-current channels have been analyzed in relation to maturation induced by a starfish hormone, one-methyl adenine (188, 192, 195). Major changes during maturation are the reduction of outwardly rectifying A-current channels and of inward rectifier K+ channels, resulting in a high membrane resistance in the potential range from 
70 to 0 mV. This higher resistance is effective in producing a larger depolarization triggered by sperm attachment (192), i.e., a larger fertilization potential. Because depolarization by the fertilization potential functions as a fast polyspermy blocker (109), these changes during maturation are favorable for monospermic fertilization. Immature oocytes are easily polyspermically fertilized, since they do not generate a depolarization to the level of the fertilization potential of mature eggs (181, 183).
A current-type channels are also found in various oocytes, such as ascidians (228), starfish (192), and mice (227); however, recent studies reveal that the "A current" in mouse eggs may be an outward current through T-type Ca2+ channels (239). In Xenopus or frog oocytes, outward currents through various K+ channels have been identified. Channels activated by transmitters and peptides exist mostly in follicular cells connected to oocytes by gap junction (176), but some of these channels have been identified as delayed rectifier K+ channels on the oocytes themselves (162, 237). The inward K+ rectifier has not been found in Xenopus, frog, or mouse oocytes.
Calcium channels exist, similar to those in excitable membranes, and Ca2+ spike potentials can be evoked by membrane depolarization above a critical potential of 50 to 40 mV in most of oocyte membranes, including those of ascidians (12, 229), echinoderms (91, 153), mice (227, 239, 332), Xenopus (15), annelids (88), and coelenterates (92). Most Ca2+ channels in oocytes exhibit fast and voltage-dependent inactivation. However, slowly inactivating channels frequently exist simultaneously (14, 42, 61). The density of Ca2+ channels increases during oocyte maturation before meiotic maturation or GVBD in starfish, that is, during the phase of oocyte growth (191). In ascidian oocytes, changes in the density of Na+ or Ca2+ channels during the maturation phase have not been examined.
Extracellular Ca2+ is known to be effective in increasing the frequency of intracellular Ca2+ oscillation and is necessary for meiotic resumption after egg activation in ascidian (72, 171), murine (101, 182, 273, 304), and amphibian oocytes (69). The effects may be mediated by Ca2+ release from cortical Ca2+ stores of egg endoplasmic reticulum stimulated by Ca2+ influx (182). The channels mediating the influx have been inferred to be Ca2+-induced or IP3-mediated nonselective cation channels (331), but further analysis may be required (182). The involvement of voltage-dependent Ca2+ channels has not been excluded.
Sodium channels in ascidian oocytes, which induce a Na+ and Ca2+ mixed spike in the rapidly rising phase of the ascidian fertilization potential, are known to be relatively permeable to Ca2+ (229). It is highly probable that the Ca2+ or Na+ channels that are capable of transferring divalent cations are partly responsible for the Ca2+ influx that regulates the early phase of intracellular Ca2+ release and accordingly contribute to the resumption of meiosis or the initiation of mitosis (316). Increases in Ca2+ or Na+ currents may also be effective in generating the rapid rise in fertilization potential that is essential for the fast mechanism of polyspermy block in various oocytes, such as ascidians (73), sea urchins (108), Urechis (74), Xenopus (78), and nemerteans (135). In both hamster and mouse oocytes, Na+ channels have not been reported; however, Na+ current can flow through egg Ca2+ channels when the extracellular Ca2+ concentration is reduced below 1 µM (227, 239, 332).
In amphibian oocytes, Ca2+-dependent Cl channels producing inward current exist in addition to Ca2+ and/or Na+ cationic channels (136, 175, 238); they are increased during meiotic maturation before GVBD but tend to decrease after GVBD (309). In the environment of a natural external medium of pond water, the current through Cl channels is inwardly directed, and the channels in association with some K+ channels are responsible for fertilization currents in amphibian oocytes (112). In lamply oocytes, Ca2+-dependent Cl current has been detected during fertilization, whereas in other fish eggs, such as those of Medaka, the Cl current has not been identified as a component of fertilization current (137, 215).
At fertilization, densities of various ion channels undergo enhancement or reduction due to changes in the intracellular environment caused by activation, often due to increased intracellular Ca2+ and/or protein kinase C activation (296). As described above, the intracellular Ca2+ concentration inside fertilized eggs usually increases in two phases: an early, large, and transient increase within a few minutes after sperm attachment and a later, relatively small and slowly oscillatory change during a few tens of minutes after fertilization in the eggs of ascidians, echinoderms, amphibians, and murines (19, 171, 182, 285, 296). These changes are now considered to be due mainly to Ca2+ release from IP3 receptor- and/or ryanodine receptor-regulated Ca2+ stores (9, 10). The increase in steady intracellular free Ca2+ concentration may be due to increased Ca2+ influx through various cation channels that are either voltage dependent or Ca2+ and/or IP3 dependent (238, 331).
Correspondingly, the changes in membrane ion channel densities can be considered in two phases. In ascidian eggs, both early and later changes have been analyzed in detail by voltage clamp during activation with sperm or with Ca2+ ionophores (12, 18, 40, 72, 140, 298).
The holding current at 70 mV shows a large, transient, inwardly directed increase during the early phase and a continuously relatively high level of inward current during the later phase, which produces the fertilization potential under constant-current condition (140). These two phases of changes in ionic channels may correspond to the changes in intracellular Ca2+ concentration analyzed in other ascidian species, described above (18, 171, 285). The channels responsible for these holding currents, fertilization currents, in ascidian eggs may correspond to the nonselective cation channels activated by sperm attachment (39, 40) during the early phase or by intracellular Ca2+ release (140, 298) during the later phase. Similar cation channels are also known to produce fertilization potentials in urechis, nemertean, and sea urchin eggs (46, 110, 134, 295).
The density of Na+ channels, which exist uniquely in some species of ascidian oocytes (12, 65, 220, 228, 234), doubles within 1 min after egg activation in the case of Halocynthia eggs (140); the density then decreases gradually almost to the previous level within 0.5 h (140).
Calcium channels with voltage-dependent inactivation exhibit decreases in density, reaching a minimum several minutes after fertilization, probably because of Ca2+-induced Ca2+ channel suppression (140).
Inward K+ rectifier channels reveal an abrupt, small, transient increase and then immediately decrease almost in parallel with Ca2+ channels (140). However, inward rectifier K+ channels gradually increase their density again after 10 min (140). This increase corresponds to the increase in permeability to K+ after the completion of the fertilization potential in echinoderms and amphibians.
All those changes may be explained primarily by the increase in intracelllular Ca2+ (298), but solid evidence is still lacking. In ascidian oocytes of Boltenia villosa, a decrease in Na+ channel density and slight increase in Ca2+ and inward rectifier K+ channel densities have been described after two cell stages (12); however, the continuous observation of voltage-dependent ion channels immediately after egg activation has not been reported.
In hamster oocytes, repetitive and transient increases in Ca2+-induced K+ current appear during the early phase after fertilization, in precise resister with oscillations of intracellular Ca2+ concentration (184). Evidence that Ca2+ oscillations are caused by Ca2+ release induced by IP3 (IICR) was obtained using blocking antibodies directed to the IP3 receptor (185). Similar changes in Ca2+-induced K+ current have been observed in mouse oocytes, but of much smaller amplitude than those of hamster oocytes (102). A fertilization potential as a polyspermic blocker is absent in the case of mouse eggs (113). However, the hyperpolarization induced by the K+ channels may be useful for increasing the driving force of Ca2+ influx.
In murine oocytes, fast inactivating Ca2+ channels are known to exist as described above (227, 239, 332). They increase in density almost twofold within 0.5 h after sperm attachment (328), which may correspond to the later phase of intracellular Ca2+ increase, and then decrease and disappear after the eight-cell stage (178).
In amphibian oocytes, as described above, Ca2+-dependent Cl channels increase in density as a result of the intracellular Ca2+ increase at fertilization, causing fertilization current (19, 111, 112, 136). Intracellular Ca2+ oscillations are reflected in the oscillatory changes in Cl currents and in fluctuations in the fertilization potential. Although the fertilization current appears to be carried by external cations in the case of many marine oocytes, which are fertilized in high ionic strength seawater, those of amphibians spawn in freshwater are carried by anions such as Cl, since Cl efflux causes an inward, depolarizing current. In Medaka eggs immersed in intravitellin fluid of relatively high ionic strength, the fertilization potential is considered to be due to cation influx of Na+ and Ca2+ (215).
The functional significance of changes in other ion channels than those used for fertilization potential, such as Na+ and Ca2+ channels, are not known. However, Ca2+ influx through Ca2+-permeable Na+ channels in ascidian eggs or that through Ca2+ channels in mouse eggs may contribute to the cortical cytoplasmic contraction or secretion of cortical exocytotic granules.
In summary, before initiation of embryonic development, the oocytes have already several kinds of ion channels, and relative amounts among various channels are changing depending on the stages of maturation, fertilization, and postfertilization. Some of the changes may be directly related to those changes in intracellular Ca2+ concentration.
A. Cleavage, Blastula, and Animal-Vegetal Axis
As a result of activation with sperm, the egg cell undergoes first mitosis after fusion of male and female pronuclei. The initiation of mitosis requires factors similar to those required for meiosis, such as the activated cdc2 and cyclin B complex, as described in section IIIA.
The increased intracellular Ca2+ concentration must play an important role on the initiation of mitosis (10, 172), probably through protein phosphorylation and dephosphorylation with calmodulin kinase II or Ca2+-dependent phasphatase IIa as shown in the case of resumption of meiosis in Xenopus eggs (161). Cell cleavage after mitosis includes formation of a contractile ring as well as new plasma membrane. The former may include actin microfilament interaction triggered by local Ca2+ increase (274, 284, 316), and the latter may induce new integration or disappearance of membrane ion channels.
Cell cleavage in Xenopus and zebrafish eggs occurs synchronously within an embryo until the midblastula stage without a detectable G1 period, indicating that there is no interruption of mitosis before going into the next cell cycle (121, 200). After the midblastula stage, the cell cycle becomes asynchronous; synchrony is maintained only within a domain that shares the same functional role in morphogenesis, as shown in the cases of zebrafish and Drosophila embryos (122).
Halocynthia embryos are synchronous until the 32-cell stage; after the 64-cell stage, the cells in the animal hemisphere exhibit a shorter cell cycle, as described by Nishida (205). After the cell cycle phase in the blastula becomes asynchronous, vertebrate embryos gastrulate in the first significant morphogenetic movement, which determines the location of presumptive areas for major organs in both anteroposterior and dorsoventral axis of the vertebrate body (68). The morphogenic movement is accomplished by the activity of a morphogenetic center, Spemann's organizer, which initiates gastrulation in vertebrate embryos (68, 288).
Although the animal-vegetal axis already exists in the unfertilized ascidian egg as well as in the amphibian egg, establishment of the dorsoventral axis in the ascidian embryo occurs after the first and second phases of cytoplasmic movement just after fertilization. The cytoplasmic movement shifts the myoplasm into the posterior end of the vegetal hemisphere, triggered by an increase in intracellular Ca2+ (6, 116, 286).
In the ascidian embryo, it is known that the two cells in the vegetal pole at the 64-cell stage are the pioneer cells to gastrulate (261). The presence of the gastrulation center at this vegetal pole in the ascidian embryo was recently confirmed by blocking the gastrulation at the 110-cell stage with ultraviolet radiation on the vegetal pole at the period of first cytoplasmic movement, just as in the amphibian embryo (117, 118).
B. Egg-Type Na+ and Ca2+ Channels
In this and the following sections, the changes in membrane ion channels are described during the blastula and early gastrula stages when cell cleavage occurs very rapidly in most animal oocyte species. In mature oocytes or eggs, there are several kinds of ion channels including Na+, Ca2+, inward rectifier K+, and delayed rectifier K+ channels. Among these ion channels, Na+ and some Ca2+ channels decrease in their densities during initial cleavage stages. In Halocynthia embryos, both Na+ and fast-inactivating Ca2+ channels decrease in density during the first 10 h at 10°C from fertilization to the early gastrula stage.
As described in section IIIB, Na+ channels show a characteristic transient enhancement during the first few minutes after fertilization (98, 140, 233, 298). Calcium channels disappear within 10-12 h after fertilization in Halocynthia embryos (140), whereas in Boltenia embryos, the total number of channels in the embryo remains constant until the eight-cell stage and disappears at the gastrula stage (12, 276). In Halocynthia and Boltenia embryos, Na+ channels are reduced in number on the whole embryonic membrane to one-fifth by the 8- to 32-cell stages. Among ascidian embryos, Ciona embryos lack Na+ channels (2, 14).
In amphibian embryos, the existence of Ca2+ channels was suggested by voltage-dependent activation of Ca2+-activated Cl channels, and the presence of endogenous Ca2+ current has been recently confirmed (15); developmental changes have not yet been examined during early development. In the mouse egg, a transient enhancement of Ca2+ channel densities during the first 2 h after fertilization, as described above, is followed by a decrease in Ca2+ current, and almost all Ca2+ current disappears until the eight-cell stage (178).
With the consideration that the Na+ channels in Halocynthia oocytes are permeable to Ca2+ (229), the intial transient increase of Na+ channels in ascidian oocytes or Ca2+ channels in mouse oocytes may be responsible for enhancing Ca2+ influx immediately after fertilization as suggested above. Oocytes kept in saline with a relatively high Ca2+ concentration, such as seawater or intercellular fluid of extraembryonic tissues, may have an ability to utilize Ca2+ influx for regulation of intracellular Ca2+ concentration to maintain short mitotic cycles or other functions during early development. The succeeding decrease may be necessary to establish a low level of intracellular Ca2+ after the initial phase so as not to disturb the progression of later development.
Quite recently, significant influences on excitabilty development produced by changing Ca2+ transient frequencies have been demonstrated in Xenopus embryonic spinal neurons (81, 82). The importance of intracellular Ca2+ dynamics has never been questioned in early developmental processes. However, further quantitative and time-sequential studies of Ca2+ influx and intracellular Ca2+ release, coordinated with analysis of changes in various ion channel densities, as actually carried in Xenopus spinal neurons (82), are still required to attain a complete understanding of their developmental roles in a simple differentiating system, such as ascidian blastomeres.
C. Various K+ Channels
In contrast to Ca2+ and Na+ channels, inward rectifier K+ channels increase their densities gradually during the blastula stage until 15 h in Halocynthia eggs as well as in those of other ascidians (12, 98, 235).
At 15 h of development or the midgastrula stages, the density of inward rectifier K+ channels in cleavage-arrested ectodermal blastomeres suddenly increases up to 20-30 times that in the unfertilized egg (235). In the intact Halocynthia gastrula embryo also, a marked increase in inward rectifier K+ current has been reported by Miyazaki et al. (190). The increase is found to be transcription dependent, since the injection of the transcriptional inhibitor 
-amanitin, a RNA polymerase II inhibitor (235), at the 32-cell stage blocks the increase of inward K+ rectifier (refer to Fig. 4), keeping the increase to less than 2-3 times that in the egg (235). This indicates that transcription for inward rectifier K+ channels may be initiated as early as the 32-cell stage, probably at the time of midblastula transition in the ascidian embryo. A minor fraction of the increase appears to be derived from maternal RNA, because a slight increase above the oocyte level still remains after ample injection of -amanitin.
View larger version (28K):
[in this window]
[in a new window]
| FIG. 4.
Differential expression of inward K+ rectifier channels between neurally induced and uninduced blastomeres. A: identification of inward rectifier K+ channes in an isolated, cleavage-arrested a4-2 blastomere at early neurula stage. Inward currents elicited by hyperpolarizing pulses to 95, 115, 135, 155, and 175 mV in artificial seawater with or without K+. B: comparison of development of inward rectifier K+ channels in neurally induced (protease treated; solid circles) and uninduced (no protease treatment; open circles) a4-2 blastomeres. Peak values of inward current evoked at 175 mV are plotted against developmental time. Timing of protease treatment is indicated by a bar. C: inhibition of inward rectifier K+ channel expression in neurally induced (solid circles) and uninduced cells (open circles) after injection of -amanitin. Relationship between levels expressed at 20-30 h of development, and timing of -amanitin injection is illustrated. On right side of graph, mean values and standard deviations for inward rectifier K+ channel current are shown for neurally induced (solid square) and uninduced blastomeres (open square) without injection of -amanitin. Period of induction by protease treatment is indicated by the column. [A-C from Okamura and Takahashi (235).]
|
|
This early zygotic expression of ion channels seems unique, because changes in other ion channels, such as Na+ and Ca2+ channels, during early development before 40 h (see Fig. 3) are not dependent on zygotic transcription in the Halocynthia embryo (98, 233, 235). Transcription for zygotic and lineage-specific expression of other ion channel genes, such as Na+, Ca2+, and delayed K+ channels, in ascidian embryos is known to be initiated after the midgastrula stage (231, 235, 276).
View larger version (38K):
[in this window]
[in a new window]
| FIG. 3.
Development of Na+ channels in neurally induced cleavage-arrested 8-cell blastomeres with voltage clamp and in intact tadpole larvae with in situ hybridization. A: macroscopic Na+ currents in neurally induced cleavage-arrested a4-2 blastomeres. Modes of currents decrease with prepulse depolarization at 3 stages. Type A is found in unfertilized eggs and in both induced and uninduced blastomeres at early stages until 40 h at 10°C (phase 1). Type C is expressed only in induced blastomeres, correponding to neuronal specific TuNa I gene (phases 2 and 3). Type B may be a precursor for neuron-specific Na+ channels and is seen only during initially rapidly increasing phase of expression (phase 2). B: peak amplitudes of Na+ channel currents in neurally induced blastomeres are plotted against developmental time at 9°C, on a semilogarithmic scale. Solid circles, H. aurantium embryos; open circles, Halocynthia roretzi embryos. Dashed line indicates development of Na+ channels in uninduced epidermally differentiated blastomeres. UF, Na+ currents at the stage of unfertilized eggs. C: control in situ hybridization of myosin mRNA. D and E: in situ hybridization of TuNa I mRNA in a tail bud and a young tadpole of H. aurantium. F: an enlargement of neurons stained in another young tadpole. Camera lucida images adjacent to in situ hybridization panels indicate previously mapped positions of neurons from Nicole and Meinertzhagen (203). Arrowheads indicate positions of epidermal neurons. EN, epidermal neuron; NH, neurohypophysis; No, a row of large notochordal cells; NT, position of neural tube; OC, ocellus; OT, otolith; PSV, postsensory vesicle; SV, sensory vesicle; UD, undefined neurons in head surface; VG, visceral ganglion. Bar, 100 µm. [A and B from Okamura and Shidara (223); C-F from Okamura et al. (231).]
|
|
Resting K+ conductances of other embryos appear to increase during blastula stages, such as in Xenopus, since the resting potential usually becomes more hyperpolarized and more sensitive to K+ as development advances from stage 6 to stage 9 in isolated and ouabain-treated Xenopus blastomeres (7). In amphibian embryos, the increase in K+ conductance associated with new membrane formation at the time of cell cleavage is well known (156, 312, 322).
In mouse embryos, voltage-independent K+ permeability increases after the eight-cell stage, and the measured resting potential becomes hyperpolarized (178). However, no other quantitative data are available for indication of the early development of inward K+ rectifier except in another ascidian embryo, Boltenia (12, 275). The significance of the early transcriptional dependence of the inward rectifier is discussed in section VD with special attention to neural differentiation in ascidian embryos.
Delayed rectifier K+ channels responsible for outward current above 0 mV in the early Halocynthia embryo are not yet clearly identified. There is no evidence that the density of outward rectifier channels increases or decreases during early development before lineage-specific expression of ion channels.
B. Two-Cell Induction System Constructed From Early Ascidian Embryos
In the ascidian embryo, the isolated anterior animal blastomere from the eight-cell cleavage-arrested embryo, including the presumptive brain vesicle region, autonomously develops a long-lasting action potential due to the expression of slowly inactivating Ca2+ channels characteristic for epidermal differentiation of the ascidian tadpole (225).
However, when cultured in contact with the anterior vegetal blastomere, including the presumptive notochordal region, and raised as the contacted two-cell system, the same anterior animal blastomere now develops typical Na+ spikes due to expression of Na+ channels and TEA-sensitive delayed rectifier K+ channels (223, 224, 232-234, 271).
Sequencing of cloned cDNAs reveals that these Na+ channels have close similarity in their primary structure to mammalian brain type II Na+ channels (231). Monoclonal antibodies directed to larval epidermal cells stain the blastomere only when cultured in isolation but show no binding when cultured as the two-cell induction system (224).
To ascertain that this inductive phenomenon of Na+ spikes is equivalent to neural induction in the ascidian embryo, we examined critical periods of Na+ spike induction by changing the stages at which cells were brought into contact (224). The effective contact for induction occurs from the 64-cell stage to the midgastrula stage, or from 10 to 12.5 h of development at 10°C, whereas the expression of Na+ spikes appears at much later stages, from 38 to 45 h of development (224).
The response-competent phase and inducer-active phase were determined separately. The responding anterior animal blastomere from various stages during 15 h from the eight-cell to the midneurula stage was contacted for 5 h with the inducing anterior vegetal blastomere from a fixed stage. Inversely, the responding blastomere from a fixed stage was contacted with inducing blastomere from various stages. The results indicate that the competent period of the presumptive neural blastomere is mainly responsible for the critical period for induction, whereas the inducer activity of the presumptive notochordal blastomere lasts much longer than the competent phase (224).
Because the critical period of neural induction in the amphibian embryos occurs at the early to midgastrula stage (240), it is highly probable that the induction of Na+ channels in the ascidian embryo is equivalent to the amphibian neural induction (223, 224).
Aside from the vertebrate embryo, neurogenesis in the Drosophila embryo also requires a kind of cell-cell interaction, but the major strategy is different from that of the vertebrate embryo in the following two aspects (20, 22): 1) neuroblasts are generated from the ventral ectoderm instead of the dorsal ectoderm in constrast to vertebrate embryos, and 2) the default differentiation of isolated competent ectoderm in the Drosophila is neural instead of epidermal. A group of almost 10 genes has been identified as neurogenic genes, having the same phenotype of suppressing further neuroblast formation around an antecedently differentiated neuroblast within the ventral neurogenic ectoderm (22). The Notch gene, one of the best known and well-analyzed neurogenic genes, is now known to code a ligand protein expressed on the surface of ectodermal cells and inferred to act through cell-cell interaction (21, 323).
In the case of the vertebrates, although the mechanisms of neurogenesis have been progressively elucidated as described above, many studies still remain to be carried out especially with respect to intracellular transduction after neuralizing signals. Thus, in ascidians, in which a smaller number of molecular cascades may underlie neural induction than in higher verteblates, the simple induction system potentially serves as an explicit model to fill in a gap between early cellular interactions and expression of neural functions, such as mRNA production for various ion channels. It is further of particular interest to study the cellular physiology during neural differentiation in the simple induction system, since its large cell size allows application of precise biophysical and cell biological techniques in addition to molecular biological techniques.
and Other Ion Channels
Top
References
In this review, we underscore the merits of using voltage-dependent ion channels as markers for neuronal differentiation from the early stages of uncommitted embryonic blastomeres. Furthermore, a fairly large part of the review is devoted to the descriptions of the establishment of a simple model system for neural induction derived from the cleavage-arrested eight-cell ascidian embryo by pairing a single ectodermal with a single vegetal blastomere as a competent and an inducer cell, respectively. The descriptions are focused particularly on the early developmental processes of various ion channels in neuronal and other excitable membranes observed in this extraordinarily simple system, and we compare these results with those in other significant and definable systems for neural differentiation. It is stressed that this simple system, for which most of the electronic and optical methods and various injection experiments are applicable, may be useful for future molecular physiological studies on the intracellular process of differentiation of the early embryonic cells. We have also highlighted the importance of suppressive mechanisms for cellular differentiation from the experimental results, such as epidermal commitment of the cleavage-arrested one-cell Halocynthia embryos or suppression of epidermal-specific transcription of inward rectifier channels by neural induction signals. It was suggested that reciprocal suppressive mechanisms at the transcriptional level may be one of the key processes for cellular differentiation, by which exclusivity of cell types is maintained.
