Morphology and Physiology of the Ascidian Nervous Systems and the Effectors
Neurobiology in ascidians has made many advances. Ascidians have offered natural advantages to researchers, including fecundity, structural simplicity, invariant morphology, and fast and stereotyped developmental processes. The researchers have also accumulated on this animal a great deal of knowledge, genomic resources, and modern genetic techniques. A recent connectomic analysis has shown an ultimately resolved image of the larval nervous system, whereas recent applications of live imaging and optogenetics have clarified the functional organization of the juvenile nervous system. Progress in resources and techniques have provided convincing ways to deepen what we have wanted to know about the nervous systems of ascidians. Here, the research history and the current views regarding ascidian nervous systems are summarized.
KeywordsSimple brain Motor ganglion Muscle Motor control Development of membrane excitability Chordate evolution Motor pattern generation Metamorphosis Connectome
Nervous systems of ascidians have attracted much interest. In ascidians, two distinct nervous networks develop in tadpole larvae and in adults. Both of them are relatively simple, offering us a cellular-based understanding of the working principles of the systems. Also, both the larval and adult nervous systems have molecular developmental bases that are comparable to those of nervous systems in vertebrates, including humans.
Here, I summarize various aspects of research history in this field and current views on ascidian nervous systems. The field is now integrating knowledge from classic and recent sophisticated neuroanatomy, embryology, neurophysiology, and new technologies, including transgenics. Ascidian neurons, especially those of larvae, are derived from defined cell lineages through an invariable cleavage pattern, and we can therefore reproducibly identify most of the neurons. The lower redundancy of genes in the genome and of neurons in the nervous systems of the larva can offer the chance to tightly connect the operational logics of the genes, neurons, and whole body. Triggering metamorphosis is also a function of the larval nervous system and thus a target of the neurobiology in this animal. The adult central nervous system represents an example of wholly regenerative tissues; on the other hand, its structural correspondence can be pursued to our brain (e.g., Dufour et al. 2006). I hope that this review is of help for the future understanding of such appealing ascidian nervous systems. This research field is now advancing and expanding rapidly; readers can refer to recent excellent reviews on structural, molecular developmental, and evolutionary perspectives of the nervous systems of ascidians (e.g., Satoh 2014, 2016; Hudson 2016).
16.2 Larval Nervous System
Ascidian larvae sense light, presumably gravity and mechanical stimuli, and possibly water pressure and temperature, and they generate swimming patterns in response to those sensory stimuli (Grave 1920; Mast 1921; Crisp and Ghobashy 1971; Kajiwara and Yoshida 1985; Svane and Young 1989; Nakagawa et al. 1999; Tsuda et al. 2003a, b; McHenry and Strother 2003; Sakurai et al. 2004; Zega et al. 2006; Nishino et al. 2011). The nervous system of ascidian tadpole-shaped larvae is composed of a few hundred neurons and a smaller number of non-neuronal cells, mostly ependymal cells lining the central cavity of the sensory vesicle and the caudal nerve cord. Their number, relative positions of subtypes, and connections among them are believed to be mostly invariant among individuals in a species-specific manner. Recently, those of a single larva of Ciona intestinalis were completely identified by means of 3D reconstruction of serial images of electron-microscopic ultrathin sections (ssEM studies) (Ryan et al. 2016). Also, developmental origins of many of the cells in nervous systems have been identified through careful observations and extensive labeling studies (Nicol and Meinertzhagen 1988a, 1988b; Okada et al. 1997; Cole and Meinertzhagen 2004; Taniguchi and Nishida 2004; Stolfi and Levine 2011; Nishitsuji et al. 2012; Oonuma et al. 2016). Based on these fascinating research foundations, the developmental process of the neurons identified can be reproducibly examined and confirmed for the “same” type of neurons over individuals (e.g., Imai et al. 2009; Stolfi and Levine 2011; Stolfi et al. 2015; Ohtsuka et al. 2014; Esposito et al. 2017). These characteristics indicate the advantage of this nervous system as a research model comparable to that of the nematode Caenorhabditis elegans, for which the developmental origins and connections of all the neurons have already been uncovered (e.g., Sulston et al. 1983; White et al. 1986).
Differing from the nervous system of C. elegans, on the other hand, the ascidian larval nervous system can be directly compared with those of vertebrates (e.g., Meinertzhagen and Okamura 2001). The tunicate clade including ascidians constitutes a sister group of vertebrates. The larvae of ascidians are tadpole-shaped and can freely swim in water as vertebrate fish do, despite their ultimately simplified body (Nishino et al. 2011). Studying the swimming mechanisms of ascidian larvae should provide insights into the ancient way of swimming that originated before the evolution of vertebrates.
Ryan et al. (2016) reported that the Ciona larval nervous system is composed of about 180 neurons in the CNS and about 28 neurons in the PNS. They identified 45 neuron subtypes according to the positions, shapes, and connections. Over the past 30 years, Meinertzhagen and colleagues have succeeded in clarifying the architecture and building-up process of the Ciona larval nervous system by means of scanning and transmission electron microscopy, digital reconstruction of serial-section images, and consecutive nuclear staining (Nicol and Meinertzhagen 1988a, 1988b, 1991; Cole and Meinertzhagen 2004; Ryan et al. 2016, 2017). They also performed mosaic labeling of larval neurons to determine the morphologies of neuron subtypes by electroporation-mediated transgenesis of the green fluorescent protein (GFP) gene under control of the synaptotagmin gene promotor/enhancer, as shown in Fig. 16.1 (Okada et al. 2001, 2002; Imai and Meinertzhagen 2007a, 2007b).
Detection of pan-neural or subtype-specific molecular markers using antibodies and in situ hybridization has also enhanced our knowledge of the larval nervous system. The whole image of the larval nervous system was first visualized by expression of a sodium channel gene (Okamura et al. 1994) and later by a monoclonal antibody that specifically labels neurons (Takamura 1998). Many molecular markers specific for neuronal subtypes have been utilized to analyze the network. The markers include genes for specific transcription factors, signaling molecules, vesicular transporters or synthetic enzymes for specific neurotransmitters, receptors for those, and functional molecules for specific sensory processing (Takamura et al. 2002; Moret et al. 2005b; Horie et al. 2005, 2008a, 2008b; Zega et al. 2008; Imai et al. 2009; Horie et al. 2010; Nishino et al. 2010, 2011; Stolfi and Levine 2011; Stolfi et al. 2015; Razy-Krajka et al. 2012).
Structural overviews of the above-stated four regions of the ascidian larval nervous system are given below.
16.2.1 Peripheral Nervous System
The peripheral nervous system of the ascidian larva includes papillar neurons in adhesive papillae, epidermal neurons (ENs) that each project a basal axon and an apical process into the larval tunic, and bipolar tail neurons (BTNs) with long bipolar neurites along the CNC and a cell body outside the CNC (Fig. 16.1) (Torrence and Cloney 1982, 1983; Katz 1983; Takamura 1998; Imai and Meinertzhagen 2007b; Horie et al. 2008a; Stolfi et al. 2015).
In general, ascidian larvae have three adhesive papillae: two dorsal papillae (called P1 on the right and P2 on the left) and one ventral papilla (P3). Imai and Meinertzhagen (2007b) proposed that each papilla harbors eight or more neuronal cells called anchor cells. The anchor cells extend an axon terminating on RTENs or on the posterior portion of the BV. They express vesicular glutamic acid transporter (VGluT), a reliable marker for neurons having glutamate for chemical transmissions, while acetylcholinesterase historeactivity was also reported in the papillae (Imai and Meinertzhagen 2007b; Horie et al. 2008a). The anchor cells are thought to function as both mechanosensors and chemosensors, although physiological studies have not yet proved the cues convincingly.
Epidermal neurons are classified into several subpopulations according to their position. The first subpopulation includes six or seven pairs of rostral trunk ENs (RTENs) regularly aligned on the dorsal and lateral surface of the trunk anterior to the BV in Ciona (Takamura 1998; Yokoyama et al. 2014). The second subpopulation consists of anterior and posterior apical trunk ENs (four ATENa and four ATENp cells) at the dorsal side of the anterior and posterior BV respectively. The third one includes dorsal and ventral caudal ENs (DCENs and VCENs) and an EN at the tail tip (TTN) (Yokoyama et al. 2014). DCENs and VCENs form eight sets and six sets of neighboring pairs respectively, on average, and are serially aligned along the dorsal and ventral midlines of the tail (Crowther and Whittaker 1994; Yokoyama et al. 2014). A similar distribution pattern of ENs has been reported in a distantly related ascidian species, Halocynthia, while VCENs are scarce in this species (Ohtsuka et al. 2001). These ENs each extend a long apical process to form a mesh-like network in the larval tunic, an extracellularly built, flat, fin-like structure (Crowther and Whittaker 1994; Ohtsuka et al. 2001; Pasini et al. 2006; Terakubo et al. 2010; Yokoyama et al. 2014). Although these ENs seem to express VGluT as anchor cells do (Horie et al. 2008a; Pasini et al. 2012), the physiological properties of the ENs have not yet been clarified. Functions as mechanosensors and/or chemosensors have been proposed, but the adequate stimuli remain to be elucidated.
Bipolar tail neurons are recently identified neurons in the larval tail (Imai and Meinertzhagen 2007b; Coric et al. 2008; Stolfi et al. 2015). They are derived from the posterior margin of the neural plate and migrate anteriorly during the tailbud stage. They extend anterior and posterior processes. The posterior (distal) process almost reaches the tip of the tail and often turns ventrally, whereas the anterior (proximal) one terminates on the posterior BV (Fig. 16.1) (Imai and Meinertzhagen 2007b; Ryan et al. 2017). The somata of BTNs are located in the middle of the larval tail, and they are associated with, but outside, the CNC (Stolfi et al. 2015). These neurons express an acid-sensing ion channel gene that encodes a possible molecular mediator for nociception and mechanosensation (Coric et al. 2008; Stolfi et al. 2015). As BTNs seemingly correspond to what were immunolabeled against γ-aminobutyric acid (GABA) in the Ciona tail (Brown et al. 2005), they are regarded as being GABAergic. Accordingly, BTNs express genes for glutamic acid decarboxylase (GAD), which is the enzyme to synthesize GABA (Zega et al. 2008; Stolfi et al. 2015), and for a homolog of vesicular inhibitory amino acid transporter (VIAAT), a factor with the activity of incorporating GABA or glycine into the synaptic vesicles (Horie et al. 2010). It should also be noted that some of the cells that Horie et al. (2008a) reported as being glutamatergic appear to share morphological traits with BTNs, and thus glutamatergic BTN-like cells may also be present.
16.2.2 Brain Vesicle
Ryan et al. (2016) reported that there are 143 neuronal cell bodies in the BV. The BV has a cavity inside, and a melanin-containing otolith protrudes from the ventromedial wall into the cavity. A single layer or a few layers of cells, mostly neurons, encircle the cavity; a single ocellus melanocyte and lens cells reside in the right side of the BV wall and directly face the cavity. These melanocytes have been proposed to be a phylogenetic counterpart of vertebrate neural crest cells (Abitua et al. 2012). Thirty or more photoreceptor cells (approximately 20 type I and 10 type II) are located in close vicinity to the ocellus pigment cell. The outer segments of type I photoreceptors are covered by the pigment cup, whereas those of type II face the cavity of the BV (Horie et al. 2008b). These photoreceptors are ciliary, not rhabdomeric, and express a type of opsin (Opsin-1 in Ciona intestinalis) and arrestin (Taniguchi and Nishida 2004; Horie et al. 2008b; Oonuma et al. 2016). Another group of six or seven photoreceptors expressing Opsin-1, type III photoreceptors, is located on the ventral wall of the BV and projects bulbous outer segments into the ventricle (Horie et al. 2008b; Ryan et al. 2016). Possible differences in the functions of these photoreceptor groups have not been determined. The melanin-containing otolith cell is associated with the fine processes of a pair of antennal mechanosensory neurons that are adjacent to the type III photoreceptors (Torrence 1986; Sakurai et al. 2004; Ryan et al. 2016). Some experiments, including laser ablation, suggested that the otolith and the associated antennal neurons constitute a gravity sensor of the larva (Tsuda et al. 2003a; Sakurai et al. 2004). Other than these, hypothetical pressure sensors called coronet cells also reside in the left wall of the BV (Eakin and Kuda 1971; Horie et al. 2008b; Ryan et al. 2016), although the result of an experimental trial to determine whether Ciona larvae can sense water pressure was negative (Tsuda et al. 2003a). About half of the coronet cells express a gene encoding tyrosine hydroxylase (TH), which catalyzes the formation of L-dihydroxyphenylalanine (L-DOPA) from tyrosine to be a rate-limiting factor to generate a catecholamine neurotransmitter, dopamine (DA) (Moret et al. 2005b). Ventrolateral TH-positive cells in the BV, including several coronet cells, do have DA immunoreactivity and expression of a gene for another enzyme, aromatic amino acid decarboxylase (AADC), which derives DA from L-DOPA (Moret et al. 2005b; Razy-Krajka et al. 2012). Pharmacological analyses in Ciona larvae have suggested that these DA cells might negatively regulate swimming activities (Razy-Krajka et al. 2012).
The photoreceptor cells and mechanoreceptive antennal neurons appear to express VGluT (Horie et al. 2008a). These neurons mostly send their axons to the posterior portion of the BV (Horie et al. 2008a, b), and connectome analysis has revealed that there are several interneuron groups that relay inputs from the sensory cells, including photoreceptors, antennae, epidermal sensory neurons, and BTNs (Stolfi et al. 2015; Ryan et al. 2016). One of the posterior BV neurons called eminens2 extends a descending axon to mediate sensory signaling from ENs to the MG, mainly to ddNs (see below), to possibly evoke a specific locomotor output (Ryan et al. 2017). These observations suggest that the posterior portion of the BV is the processing center of sensory inputs. In fact, a representative type of glutamate receptor is expressed in this region and is involved in the development of sensory cells and in the proper progression of metamorphosis (Hirai et al. 2017). Some subunits for the acetylcholine (ACh) receptor and GABA receptor are also known to be expressed in this posterior region of the BV (Zega et al. 2008; Nishino et al. 2011; our unpublished observation).
16.2.3 Motor Ganglion
According to Ryan et al. (2016), the MG of Ciona intestinalis contains 25 neurons including a pair of descending decussating neurons (ddNs) (Ryan et al. 2017). Results of transgenic studies using fluorescent protein genes suggested that this pair of ddNs might be vesicular acetylcholine transporter (VAChT)-positive, a reliable marker for cholinergic neurons, and paired-class homeobox gene Dmbx-positive (Takamura et al. 2010; Stolfi and Levine 2011). Recently, Ryan et al. (2017) showed that the network around this pair of interneurons is quite reminiscent of the Mauthner system, which is well known in relation to the startle response in vertebrate fish and tadpoles (Fetcho 1991; Eaton et al. 2001).
Neuron types other than motor neurons have also been reported (Imai and Meinertzhagen 2007a; Ryan et al. 2017). These neurons include three pairs of MG interneurons (MGINs) and seven dorsally located ascending MG neurons (AMGs or contrapelo cells). MGINs have descending axons and have been proposed to mediate or suppress the signaling relay from ddNs to MNs (Ryan et al. 2017). AMGs may include two populations that extend presumed cholinergic and GABAergic axons to the BV respectively (Brown et al. 2005; Takamura et al. 2010). Some of the AMGs, including the cholinergic ones, have been proposed to transmit mechanosensory signals from the epidermal sensory neurons to ddNs (Takamura et al. 2010; Ryan et al. 2017).
16.2.4 Caudal Nerve Cord
It had once been assumed that the CNC does not contain any neurons and is merely composed of ependymal cells forming dorsal, ventral, and lateral walls facing the neural tube canal. We now know, however, that the dorsal nerve tube in the tail is not merely a scaffold for the motor nerves, but contains a countable number, nine or more, of neurons (Imai and Meinertzhagen 2007a; Horie et al. 2010; Takamura et al. 2010; Ryan et al. 2016, 2017). Two pairs of anterior caudal interneurons, ACINs, of Ciona larvae have anteriorly extending commissural axons and express VIAAT (Horie et al. 2010). Although the gene for the GABA synthesizing enzyme, GAD, does not appear to be expressed in this anterior caudal region (Zega et al. 2008), cells immunolabeled with a glycine antibody were found there (Nishino et al. 2010), and thus the ACINs are presumably glycinergic. These are derived from two siblings of A11.116 descendants and develop under the control of SoxB1 expression (Nishitsuji et al. 2012). Although the Ciona larva subjected to the ssEM study by Ryan et al. (2016, 2017) did not have two pairs of them, these ACINs have been proposed to receive a signal from ipsilateral excitatory axons at their cell body and to inhibit contralateral motor axons by releasing the inhibitory transmitter (Horie et al. 2010; Ryan et al. 2017). The motor neurons, and possibly other MG-descending neurons, express a glycine receptor gene, and the reciprocal inhibition between the left and right sides represents a network basis for alternating tail beats (Horie et al. 2010; Nishino et al. 2010).
Pairs of midtail neurons, MTNs, with fusiform somata, are located posterior to ACINs and scattered along the CNC. Ryan et al. (2016, 2017) identified four of them by ssEM up to the anterior half of the tail, but there seem to be more of them posteriorly. Six pairs visualized by VAChT gene expression seem to correspond to MTNs (Horie et al. 2010; Takamura et al. 2010). Postsynaptic nAChR clusters on the dorsal muscle cells also seem to be formed in association with this cell type (Nishino et al. 2011), and these MTNs are therefore categorized as motor neurons (Ryan et al. 2017). Other than these, Ryan et al. (2016, 2017) found some interneurons in the anterior CNC, called posterior MG interneurons (PMGNs). BTNs have recently been identified in the tail, as mentioned above; their somata are located out of CNC and are a component of the PNS (Stolfi et al. 2015).
16.3 Larval Muscular System
The number of larval tail muscle cells is generally small, although the larvae of some species have hundreds of muscle cells (Jeffery and Swalla 1992). Ciona and Halocynthia larvae have 18 and 21 muscle cells respectively, on either side of the tail (Nishida and Satoh 1985; Nishino et al. 2011). The number and arrangement of muscle cells are left–right symmetrical and invariant in a species-specific manner (Fig. 16.2). The muscle cells in each side form a single-layered sheet and are arranged in three rows in Ciona: dorsal D1-7, middle M1-4, ventral V1-6, and posterior-most terminal cell in a total of 18 cells (Fig. 16.2) (Nishino et al. 2011). Despite this structural invariance and simplicity, the larval muscle band is not a mere gathering of muscle cells, but constitutes a system, as do our skeletal muscles and heart. The dynamics of the muscle bands is variable and graded. High-speed video imaging revealed that the tail of a Ciona larva produces repetitive antero-posteriorly-propagating and left–right alternating waves with magnitudes that vary stochastically and in response to light–dark conditions (Nishino et al. 2011). This implies that the muscle band represents a “soft” actuator that functions not in an all-or-none fashion, but in a graded manner. Although mature muscle cells can show regenerative Ca2+-based action potentials, Ca2+ spikes, in experimental conditions, the action potential is not often seen in direct electrophysiological recordings (Ohmori and Sasaki 1977; Bone 1992; Nishino et al. 2011). The muscle cells on either side adhere to each other to form gap junctional communication (Bone 1992). This gap junctional communication considerably decreases the input resistance, making it difficult to evoke action potentials (Bone 1992). This property of the muscle band of tending not to make action potentials seems to provide a physiological basis for the graded dynamics.
Ascidian larval muscle responds to the neurotransmitter acetylcholine (ACh) (Ohmori and Sasaki 1977). We examined expression patterns in the larvae of all of the nicotinic ACh receptor (nAChR) subunit genes encoded in the Ciona genome, and we found that three subunit genes, called A1, B2/4, and BGDE3, and a gene for a clustering protein of muscle nAChR, rapsyn, are expressed in the muscle band (Fig. 16.2) (Nishino et al. 2011). At least two of these genes, A1 and rapsyn, are expressed more strongly in dorsal cells than in ventral or middle cells, whereas the expression of B2/4 and BGDE3 genes appears to be uniform among all of the muscle cells (see Fig. S4 in Nishino et al. 2011). Transgenic expression of fluorescent protein-tagged nAChR subunits and rapsyn protein allowed us to detect postsynaptic nAChR clusters in D1-7, M1, and posterior terminal muscle cells (Fig. 16.2) (Nishino et al. 2011; our unpublished observation). D1-7, M1, and probably posterior terminal muscle cells seem to receive ACh released from the axons of motor neurons. On the other hand, M2-4 and V1-6 cells harbor well-developed myofibrils and receive the activation signals via gap junctions. Because of the low input resistance of the muscle band, endplate potentials elicited by local nAChR activation do not evoke an action potential, but propagate over the surface of the muscle band with attenuation. The voltage-gated CaV1 channels and Ca2+-induced Ca2+ release mechanisms of ryanodine receptors on the intercellular Ca2+ stores, both residing in the muscle band, would work as amplification components of the signal derived from the endoplate potential (Nakajo et al. 1999; Nishino et al. 2011). Considering different projection patterns of motor axons of MN1–5 and MTNs, spatiotemporally regulated ACh inputs into the muscle bands may achieve graded and propagating patterns, neither all-or-none nor simultaneous patterns, of flexions on the muscle band (Nishino et al. 2011; Ryan et al. 2017). Halocynthia larvae are known to have a different number, three pairs, of motor neurons and a different pattern of projections of their axons; the middle motoneuron pair called Moto-b innervates the anterior-most ventral muscle cell (Okada et al. 2002). These morphological variations may explain the differences between the swimming performance of the larvae of this species and that of the larvae of Ciona. The glycine receptor gene is expressed in the muscle bands of Ciona (Nishino et al. 2010). Although it is unknown whether the gene product functions there, inhibitory inputs into the muscle band through this receptor channel may further support the fine regulation of graded flexions.
16.4 Excitability of Neurons and Muscle Cells
Independently of the research on the morphology of ascidian nervous systems, there has been important research history on electrophysiology that has focused on the excitability of ascidian larval cells. This research was pioneered by the past Dr. Kunitaro Takahashi (Moody and Okamura 2013). Neurons and muscle cells are excitable cells, and each cell type has a specific excitation pattern that is represented by the combined functions of Na+, Ca2+, K+, and other ion channels. The excitability of each cell type emerges in the course of development, and ascidian embryos have provided an ideal model for study. Takahashi et al. (1971) reported a beautiful sequence of four developmental stages in the changes in the membrane excitability of muscle cell lineage: a naïve state of excitability in the egg and cleavage stage; emergence of signs of excitability in the gastrula; occurrence in the tailbud stage of sustained all-or-none fashioned action potential (AP); and strengthening of currents to repolarize the membrane potential to sharpen the AP in hatchlings. As the muscle develops, the permeability of Na+ declines and mature muscle cells show Ca2+ spikes (Miyazaki et al. 1972). Takahashi and his colleagues developed a sophisticated schema in which neurons express specific Na+ currents and delayed K+ currents to elicit sharp Na+ spikes, muscle cells express specific Ca2+ currents and delayed K+ currents for Ca2+ spikes, and epidermal cells do not express delayed rectifier K+ currents and thus show a different sustained type of Ca2+ spikes (Hirano et al. 1984; Okado and Takahashi 1988; Simoncini et al. 1988; Takahashi and Okamura 1998; Nakajo et al. 1999). This cell-type-specific expression of membrane excitability can be seen and analyzed even in blastomeres that have been isolated manually and cleavage-arrested by cytochalasin B (Takahashi and Yoshii 1981). This allowed them to examine the differential expression of cell-type-specific membrane currents using a two-electrode voltage-clamp configuration (for a review, see Takahashi and Okamura 1998). The isolated and cleavage-arrested a-line blastomeres develop to express epidermis-type Ca2+ spikes, but the a-line blastomere isolated together with the A blastomere from eight-cell embryos and cleavage-arrested thereafter is differentiated to show Na+ spikes (Okado and Takahashi 1988, 1990a, b). This is a simple model of neural induction and it was later shown that the induction signal from the A-line can be replaced by treatment with serine proteases or basic fibroblast growth factor (Okado and Takahashi 1993; Takahashi and Okamura 1998).
The detailed kinetics of ion channels have been investigated using the voltage-clamp and patch-clamp techniques (Okamura and Shidara 1987, 1990a, b; Shidara and Okamura 1991; Okamura and Takahashi 1993). Okamura et al. (1994) succeeded in molecular cloning of the cDNA of a neurally expressed, voltage-gated Na+ channel named TuNa I from Halocynthia roretzi (Hr-NaV1). They examined the expression pattern of the gene and also knocked down the gene expression to show its involvement in the induced Na+ spikes in neural cells (see also Okada et al. 1997; Ono et al. 1999). On the other hand, establishment of Ca2+ spikes in the course of autonomous differentiation of muscle cells is now known on the basis of the expression of TuCa1 (Hr-CaV1) and its auxiliary subunit TuCaβ (Simoncini et al. 1988; Davis et al. 1995; Dallman et al. 1998; Nakajo et al. 1999; Dallman et al. 2000; Okagaki et al. 2001; Ohtsuka and Okamura 2007). K+ channels are also important for shaping APs in neurons and muscle cells, and expression patterns and functions of some of the K+ channel genes have also been studied in relation to the differentiation of excitabilities (Ono et al. 1999; Murata et al. 2001; Nakajo et al. 2003; Hill et al. 2008).
While this fruitful experimental system is now in danger because of a decrease in the number of researchers using electrophysiological techniques, the system can still provide an ideal research pipeline for understanding the developmental processes of differential membrane excitabilities. We now have a catalog of ion channel genes in Ciona (Okamura et al. 2005), and methods for the primary culture of neural cells and electrophysiological recording from the cultured cells have also been reported (Zanetti et al. 2007). It is expected that new fruits will be obtained by combining these research systems with genomic resources and with transgenic techniques.
16.5 Transgenic Techniques for Studying Larval Nervous and Muscular Systems
Transient transgenesis was first reported in ascidians in 1992, with recombinant plasmids being introduced into the fertilized eggs of Ciona savignyi and Halocynthia roretzi by microinjection (Hikosaka et al. 1992). In 1997, the introduction of transgenes into the fertilized eggs of Ciona intestinalis, in which microinjection has been considered relatively difficult, became possible by means of electroporation, at the same time as the use of the green fluorescent protein (GFP) gene in ascidians (Corbo et al. 1997). The use of GFP allowed living cells to be labeled, resulting in a great accumulation of information on neuron morphology and the structural basis of whole nervous systems. Okada and others isolated a synaptotagmin gene promotor/enhancer sequence from Halocynthia, and the construct of enhanced GFP (EGFP) placed downstream of this cis-regulatory sequence has been utilized for random labeling of neurons with chemical presynapses (Fig. 16.1) (Okada et al. 2001, 2002; Katsuyama et al. 2002; Imai and Meinertzhagen 2007a, b; Nishino et al. 2011). They also tried forced expression of a dominant-negative form of a voltage-gated K+ channel (DN-TuKV2) under control of the synaptotagmin cis-regulatory sequence by microinjecting this construct into A5.2 to change the swimming pattern of the transformants (Ono et al. 1999; Okada et al. 2002). This pioneering trial proved that transgenic manipulation of ion channel functions can affect locomotor patterns of the transgenic larva.
Live imaging based on GFP has made much progress and several fluorescent fusion proteins are now also used in ascidians, including fusions with nucleus localization signal (NLS), histone H2B (localizing with chromosomes), MAP7 (microtubule-associated protein 7, ensconsin), LifeAct (a short peptide from Saccharomyces that binds to F-actin), pleckstrin homology (PH) domains (binding phosphatidylinositol lipids in cell membranes), EB3 (end binding 3 to label the plus end of microtubules), or Geminin and Cdt1 (specific cell-cycle regulators) to visualize subcellular structures, cytoskeletons, and phases in cell cycles (e.g., Ogura et al. 2011; Negishi and Yasuo 2015). Stolfi and Levine (2011) utilized a UNC-76::GFP fusion protein derived from GFP and a factor involved in axon extension in C. elegans, which improved visualization of neurites of zebrafish olfactory neurons (see Dynes and Ngai 1998). I myself tried EGFP fused with Ciona synaptobrevin for visualizing presynaptic vesicular accumulation, but the effect was not clear (Fig. 16.2) (Nishino et al. 2011). On the other hand, our trial to express fluorescent protein-tagged Ciona nAChR subunits and rapsyn in larval muscle succeeded in visualizing postsynaptic aggregates of the receptors (Fig. 16.2) (Nishino et al. 2011).
Transgenesis by electroporation established by Corbo et al. (1997) in C. intestinalis has pushed this ascidian species forward as an ideal model organism. This technique allows us to introduce transgenes into living cells without special skills. The basic procedure for electroporation of transgenes, however, requires dechorionation of eggs. Dechorionation is known to affect the development of embryos, especially the characteristics of left–right asymmetry. Nervous systems of the larva have many asymmetric traits, and dechorionation does indeed change them (Shimeld and Levin 2006; Nishide et al. 2012; Oonuma et al. 2016; Ryan et al. 2016). It is also known that tail elongation becomes distorted and that larvae that have developed without the chorion cannot swim normally (Nishide et al. 2012). To minimize these effects, transgenesis by microinjection without dechorionation was performed in eggs of C. intestinalis (Nishino et al. 2011; Oonuma et al. 2016).
Combining gene knock-down or knock-out and transgenic expression of the gene under consideration allows us to rescue the phenotypes to confirm the specificity of the knock-down/knock-out. Rescuing by an artificial fusion protein gene or a gene with an artificial mutation enables us to assess the effect(s) of the fusion or mutation respectively. We knocked down an indispensable subunit of muscle nAChR called BGDE3, and rescued the paralytic phenotype using wild-type BGDE3 and its mutant, which alters the electrophysiological properties of muscle nAChR (Nishino et al. 2011). In that experiment, the knocked-down phenotype caused by microinjection of a BGDE3-morpholino oligonucleotide was rescued by wild-type BGDE3 synthetic mRNA, the mutant mRNA, or the mutant gene expressed under the control of a muscle actin promoter. The larvae rescued by wild-type BGDE3 appeared to swim normally, whereas those rescued by the mutant showed altered patterns of movement (Nishino et al. 2011). This impairment is accounted for by a single mutation in the gene, and this kind of analysis can naturally be applied for assessing the various functions of paralogs or orthologs. We can now use tissue-/cell-type-specific genome editing (Sasaki et al. 2014; Stolfi et al. 2014; Treen et al. 2014) in addition to tissue-/cell-type-specific expression of a rescuing construct by means of well-designed transgenic constructs. Such approaches will lead to more reliable examination of a working hypothesis.
16.6 Neural Control of Metamorphosis
Attachment to the substrate via adhesive papillae transforms the swimming larva into a sessile adult. Internal signaling to trigger metamorphosis involves neural and/or neuroendocrine systems (Cloney 1982; Nakayama-Ishimura et al. 2009; Kamiya et al. 2014; Karaiskou et al. 2015; Matsunobu and Sasakura 2015). Although many substances that can induce metamorphosis have been described so far, the point of action of each remains unclear. The C. intestinalis larva ceases movement and regresses the tail in response to a hypothetical signal that emerges only after occurrence of competence for tail regression, approximately 12 h after hatching and after experience of continuous attachment via papillae for 28 min or more (Matsunobu and Sasakura 2015). This cessation of movement, also in addition to the emergence of competence and the memory of 28-min attachment, would reflect some unknown alterations engraved in the nervous or muscular systems in the larva. This would be one of research foci in which new transgenic approaches, including genome editing, would facilitate further cellular-/molecular-based resolution of this complicated series of phenomena in metamorphosis.
Metamorphosis, on the other hand, changes the entire nervous and muscular systems. The larval nervous and muscular systems degenerate, and those of the adult are rebuilt from undifferentiated cells that have been set aside in the larval body (Willey 1894; Hirano and Nishida 1997; Stolfi et al. 2010; Horie et al. 2011; Razy-Krajka et al. 2014). Horie et al. (2011) fully showed the ability of transgenic fluorescent proteins. Their work began by establishing a stable line, having a photoconvertible fluorescent protein, Kaede, in the nervous systems under the control of a β2-tubulin gene cis-regulatory sequence (Horie et al. 2011). Photoconversion of Kaede from green to red before metamorphosis enabled cells in the larval nervous system to be distinguished from the neural cells that differentiated after metamorphosis. They utilized several types of cell-type-specific cis-regulatory sequences to express Kaede in some neuron subtypes and in ependymal cells. Their analyses revealed that ependymal cells in the BV and neck region of the larval CNS are the cellular origin of the adult CNS, cerebral ganglion (CG), and that anteroposterior cellular organization of the adult CNS corresponds well with that of the larval CNS (Horie et al. 2011).
16.7 Structural and Functional Analyses of Adult Nervous Systems
The organization and function of the adult CNS, cerebral ganglion (CG), and the adult PNS have been elucidated in several species by morphological and physiological analyses (Arkett 1987; Arkett et al. 1989; Koyama and Kusunoki 1993; Tsutsui and Oka 2000; Mackie and Burighel 2005). However, compared with those in the larva, our understanding of the juvenile/adult nervous systems has not been sufficient.
The adult nervous system consists of the CNS and PNS. The CG located between the two, oral and atrial, siphons of ascidians represents the CNS, associated with several branches of nerve tracts, a dorsal strand (dorsal cord), and a neural gland. Based on its relative position, the neural gland complex composed of the ciliated duct, neural gland, and dorsal strand had been discussed as being homologous to the vertebrate adenohypophysis, but recent research has suggested that it might not be appropriate to refer to the neural gland as a homologue of the adenohypophysis (for reviews, see Goodbody 1974; Gorbman 1995). There has been an accumulation of expression data on the neural gland complex of several kinds of genes for peptides and their receptor-like G protein-coupled receptors (Deyts et al. 2006; Matsubara et al. 2016). Another aspect of interest is that this CG and neural gland complex is wholly regenerative (Dahlberg et al. 2009, and see references therein). Dahlberg et al. (2009) utilized an enhancer-trap transgenic line called E15 that expresses GFP in neural cells (Awazu et al. 2007) to show fine cellular views of CG regeneration.
The CG is oval- or rod-shaped and has anterior–posterior (oral–atrial) polarity (Mackie and Burighel 2005; Hozumi et al. 2015). One pair, or more than one pair, of nerve tracts, the number being variable among species, extends from the anterior and posterior ends of the CG (Arkett et al. 1989; Koyama and Kusunoki 1993; Mackie and Burighel 2005). Some thin nerves protrude from the left and right lateral sides of the CG in Botryllus (Zaniolo et al. 2002). Ascidian adults possess specialized peripheral sensory apparatuses such as the coronal organ and cupular organ, which are known as hydrodynamic sensors. The coronal organ includes secondary sensory neurons that have stereovilli and a sensory cilium resembling vertebrate inner ear hair cells, and it is located in the velum and tentacles of the oral siphon (Mackie and Burighel 2005). The cupular organ has a gelatinous cupula and primary sensory cells that look like a lateral line neuromast of vertebrates, many of which are distributed over the surface of the atrium (Mackie and Burighel 2005; Ohta et al. 2010). In contrast, chemical sensation in ascidian adults is not well understood.
Anterior and posterior tracts of the CG include both sensory and motor nerves, whereas sensory nerves projecting to the oral siphon are from the anterior tracts and those projecting to the atrial siphon are from the posterior tracts (Hozumi et al. 2015). Most of the nerves innervating ciliated cells on stigmata, gill apertures, are within the posterior tract(s), called the visceral nerve (Arkett 1987; Arkett et al. 1989; Mackie and Burighel 2005). Most of the neurons in the CG are thought to be interneurons to relay and process sensory inputs, and others are motor neurons to innervate adult muscle cells and stigmatal ciliated cells (Mackie and Burighel 2005; Dufour et al. 2006; Hozumi et al. 2015). On the other hand, there are some sensory cells in the CG that respond to light (Tsutsui and Oka 2000). A considerable number of neurons (approximately 25% of cells in the CG of C. savignyi) are responsive to light on or light off stimuli with membrane depolarization or hyperpolarization, although it is not clear whether the recorded responses occurred directly on the cell or via the synapses (Tsutsui and Oka 2000).
Hozumi et al. (2015) effectively utilized stable transgenic lines and transiently transgenic animals of C. intestinalis, and revealed the neuronal organization of the juvenile CG. The juvenile nervous systems are mostly reorganized during metamorphosis and are distinguished from those in the larva, but they revealed that the organization in juveniles is comparable with that in the larva. Peripheral neurons express the VGluT gene, and central neurons expressing the same gene are mainly located in the anterior portion of the CG. Neurons expressing VIAAT are found only in the CG and their neurites are also within the CG. VAChT-positive neurons are mostly within the CG and extend long axons peripherally to innervate siphons, body-wall muscles, and stigmatal cells (Hozumi et al. 2015 and see also Dufour et al. 2006). Hozumi et al. (2015) further optogenetically confirmed the cholinergic regulation of the siphon and body-wall muscles and the cilia on stigmatal cells. They expressed channelrhodopsin-2 (Nagel et al. 2003; Zhang et al. 2007) fused in frame with yellow fluorescent protein (ChR2::YFP) mosaically in cholinergic neurons in the CG under the control of a VAChT gene cis-regulatory sequence. Some juveniles expressed ChR2::YFP in the CG neurons innervating the oral siphon, atrial siphon or body-wall muscle, and they contracted each of the targets in response to blue-light irradiation; others expressing ChR2::YFP in those innervating stigma evoked ciliary arrest upon irradiation (Hozumi et al. 2015). This indicated the possibility of using optogenetic techniques in this transparent organism that will enable us to examine, or manipulate, the function of genetically identified neurons and other excitable cells.
16.8 Future Direction
Research into ascidian nervous systems has made much progress. Recent ssEM studies have revealed the precise number, types, and connections of neurons in a Ciona larva (Ryan et al. 2016, 2017). Techniques for transient and stable transgenesis, genome editing, and fluorescence imaging have been established (see other chapters in this book, and Sasakura et al. 2012). Optogenetics is also now being applied and has enabled manipulation of the physiological activity of living ascidians (Hozumi et al. 2015; Horie et al., personal communication). These resources and techniques represent a reliable map and vehicles to solve the remaining mysteries of the nervous systems in this model species.
Many mysteries do indeed remain. We do not yet know the possible variability of neuron networks among individuals. We do not know what happens in the network during the aging of larvae, while we do know for instance that light sensitivity of the Ciona larva and correspondingly the neurites of photoreceptors gradually develop after hatching (Tsuda et al. 2003b; Horie et al. 2005) and that DA in the BV increases as the larvae age (Moret et al. 2005b). We also do not know how the networks change in response to the irreversible activation of larval papillae (Matsunobu and Sasakura 2015), or to what degree the view of the CG structure in juveniles proposed by Hozumi et al. (2015) is applicable to that in mature adults (Dufour et al. 2006). We are not yet convinced of the full operational logics of larval neurons and muscles to facilitate effective swimming performance that can be modulated in response to sensory stimuli.
To answer these questions, we would be able to collect cis-regulatory sequences specific to each subtype for making it possible to drive specific transgenes in specific neuron subtypes. It would also be useful to know the neurotransmitter(s), the receptors, and voltage-gated ion channels that each subtype utilizes to function in the network. Furthermore, we would be able to uncover the mostly unknown developmental mechanisms for each neuron subtype to make proper connections to other subtypes, i.e., the mechanisms for axon pathfinding. Changing the properties and connections of specific neuron subtypes by utilizing DNA resources and transgenics that we have at present and in the future will provide a promising approach to solving operational logics of neuron networks. This is indeed possible in Ciona. The effectiveness of this approach would be greatly increased if we could overcome the remaining technical obstacles by establishing effective methods for managing living adults, including stable lines, controlling their repetitive reproduction, and by establishing methods to avoid artifacts caused by dechorionation, etc.
There are many ascidian and other tunicate species on earth. They have major or minor structural and functional differences in their sensory and motor systems, reflecting their lifestyles and patterns of adaptation to different environments. Berrill (1948) described intense variation in the presence and absence of the otolith, ocellus, and their combinatorial apparatus, photolith, in larvae of stolidobranch ascidians, for instance, but we do not yet know what the developmental causes are and to what degree such differences in the sensory apparatus affect the central neural network or their output; namely, the impacts on larval behavior and ecology. We are now in the era in which morphology can be integrated with any other fields, including developmental genetics, evolutionary biology, ecology, and physiology.
I am grateful for helpful comments by the editor, Dr. Yasunori Sasakura, which greatly improved the manuscript.
- Crisp DJ, Ghobashy AFAA (1971) Responses of the larvae of Diplosoma listerianum to light and gravity. In: Crisp DJ (ed) Fourth European marine biology symposium. Cambridge Press, Cambridge, pp 443–465Google Scholar
- Goodbody I (1974) The physiology of ascidians. Adv Mar Biol 12:1–149Google Scholar
- Matsubara S, Kawada T, Sakai T, Aoyama M, Osugi T, Shiraishi A, Satake H (2016) The significance of Ciona intestinalis as a stem organism in integrative studies of functional evolution of the chordate endocrine, neuroendocrine, and nervous systems. Gen Comp Endocrinol 227:101–108PubMedCrossRefGoogle Scholar
- Ohta N, Horie T, Satoh N, Sasakura Y (2010) Transposon-mediated enhancer detection reveals the location, morphology and development of the cupular organs, which are putative hydrodynamic sensors, in the ascidian Ciona intestinalis. Zool Sci 27:842–850Google Scholar
- Okamura Y, Nishino A, Murata Y, Nakajo K, Iwasaki H, Ohtsuka Y, Tanaka-Kunishima M, Takahashi N, Hara Y, Yoshida T, Nishida M, Okado H, Watari H, Meinertzhagen IA, Satoh N, Takahashi K, Satou Y, Okada Y, Mori Y (2005) Comprehensive analysis of the ascidian genome reveals novel insights into the molecular evolution of ion channel genes. Physiol Genomics 22:269–282PubMedCrossRefGoogle Scholar
- Satoh N (2014) Developmental genomics of ascidians. Wiley-BlackwellGoogle Scholar
- Satoh N (2016) Chordate origins and evolution. AcademicGoogle Scholar
- Svane I, Young CM (1989) The ecology and behavior of ascidian larvae. Oceanogr Mar Biol Rev 27:45–90Google Scholar
- Torrence SA, Cloney RA (1983) Ascidian larval nervous system: primary sensory neurons in adhesive papillae. Zoomorphology 102:111–123Google Scholar
- Zega G, Biggiogero M, Groppelli S, Candiani S, Oliveri D, Parodi M, Pestarino M, De Bernardi F, Pennati R (2008) Developmental expression of glutamic acid decarboxylase and of γ-aminobutyric acid type B receptors in the ascidian Ciona intestinalis. J Comp Neurol 506:489–505PubMedCrossRefGoogle Scholar