An α7-related nicotinic acetylcholine receptor mediates the ciliary arrest response in pharyngeal gill slits of Ciona Free

Cilia, hair-like organelles projecting from cells, exhibit movement that is often coordinated to make propagating waves through an array of cilia and even over neighboring cells, called metachronal waves (Gray, 1930). Metachronal waves propel circumferential fluid to exert a number of physiological activities such as swimming, circulation, protection, ingestion, and the transport of cells, including germ cells. Ciliary movement patterns may change in response to physiological conditions, which are mostly under the control of nervous systems (e.g. Carter, 1926; Kinosita and Murakami, 1967). The most drastic response is the arrest of metachronal waves, called ciliary arrest, which is widely observed in the animal kingdom (MacGinitie, 1939; Mackie, 1970; Murakami and Takahashi, 1975; Satir et al., 1976). The signaling process for arrest in cilia on the gills of marine filter feeders, such as bivalves and ascidians, is similar to that for muscle contraction or the exocytosis of neurotransmitters: neural transmission stimulates excitation of the postsynaptic membrane, thereby increasing cytosolic Ca²⁺ (Takahashi et al., 1973; Mackie et al., 1974; Murakami and Takahashi, 1975; Saimi et al., 1983; Bergles and Tamm, 1992). However, the molecular mechanisms involved in controlling ciliary arrest have not been elucidated to the same extent as those for muscle contraction and secretion.

Ascidians belong to the phylum Chordata (or superphylum Chordata; Satoh et al., 2014) and constitute one of the most closely related invertebrate groups to vertebrates. Ciona is an ascidian research model that has provided extensive information on genomic architecture as well as body architectures throughout its development (Dehal et al., 2002; Satoh, 2013); the neuroanatomical traits of the larvae and juveniles have been examined in detail (e.g. Hozumi et al., 2015; Ryan et al., 2016, 2017, 2018; Osugi et al., 2017; Nishino, 2018). The pharynx in juvenile and adult Ciona has several to thousands (depending on the growth stage) of perforations, called gill slits or stigmata. Seawater inhaled from the incurrent pore, the mouth, flows into the pharyngeal cavity, and then through stigmata to the atrium, while food particles are captured in the pharynx and conveyed into the alimentary canal (e.g. Petersen, 2007; Nakashima et al., 2018; Nakayama et al., 2019). This water flow is in general propelled by the metachronal waves formed by cilia on stigmatal cells in ascidians (Takahashi et al., 1973; Mackie et al., 1974; Bergles and Tamm, 1992). Previous ultrastructural studies on other ascidians, Corella and Chelyosoma, reported that each stigmatal cell was laterally narrow, an array of multiple cilia lined its apex, and a stigmatal wall comprised several rows of these flattened multiciliated cells (Mackie et al., 1974; Arkett et al., 1989). The flow rate through stigmata is regulated and the arrest of this flow is mediated by neural transmission (Mackie et al., 1974; Arkett, 1987; Petersen et al., 1999; Cima and Franchi, 2016). A gentle touch on body parts susceptible to mechanical stimuli elicits ciliary arrest on stigmata, which persists for several seconds and occurs concomitantly with ‘squirting’ (see Movie 1). The pharynx of Ciona may be an ideal model for clarifying the molecular mechanisms responsible for the neural control of the ciliary arrest response.

Previous physiological analyses showed that the ciliated cells of ascidian stigmata depolarize to make action potentials (APs), which stimulate increases in cytosolic Ca²⁺ concentrations at ciliary arrest (Takahashi et al., 1973; Arkett, 1987; Bergles and Tamm, 1992). APs have been recorded in the stigmatal ciliated cells of adult Corella and Chelyosoma (Mackie et al., 1974; Arkett, 1987), and were shown to be induced by the electrical stimulation of a sensory nerve in Chelyosoma, in which the response was blocked by the local administration of d-tubocurarine (d-TC, a representative nAChR antagonist) to stigmata (Arkett, 1987). Nerve fibers from the cerebral ganglion in Chelyosoma and Botryllus have been shown to terminate on the ciliated cells of stigmata, and these endings were stained by acetylcholinesterase (AChE) histochemistry (Mackie et al., 1974; Arkett et al., 1989; Zaniolo et al., 2002; Mackie and Burighel, 2005). Hozumi et al. (2015) recently reported that transgenic Ciona juveniles, which express channelrhodopsin-2 in some cholinergic central neurons that extend axons to stigmata, exhibited ciliary arrest in response to blue light irradiation. These findings demonstrated that the ciliomotor innervations of central neurons are cholinergic, and that neurociliary transmission stimulates the membranes of ciliated cells to elicit ciliary arrest.

To expand on the findings of these studies, we attempted to elucidate the molecular and physiological bases of ciliary arrest using Ciona pharyngeal gills. AChE histochemistry was applied to this species in order to visualize the pattern of cholinergic inputs on Ciona stigmatal cells. Fresh preparations of adult pharyngeal gills in Ciona were established, which enabled us to examine the molecular signaling for ciliary arrest. Using this preparation, we examined the direct effects of agonists of nicotinic ACh receptors (nAChRs) on ciliary arrest, and the results obtained demonstrated the dose-dependent duration of arrest to these agonists. By analyzing gene expression patterns, we identified nAChR-A7/8-1 (hereinafter referred to as A7/8-1), a cognate in Ciona of vertebrate α7-class nAChR genes, as a candidate cholinergic receptor expressed on Ciona stigmatal cells. The pharmacological traits of exogenously expressed A7/8-1 on Xenopus oocytes corresponded to agonist-induced ciliary arrest responses in vivo, thereby supporting the role of A7/8-1 as an essential mediator of neurociliary transmission. We also pharmacologically examined the involvement of downstream factors of A7/8-1 in sustaining the arrested status for seconds. Phospholipase C (PLC) is proposed to be a downstream regulator of A7/8-1.

Adult wild-type pacific Ciona intestinalis Linnaeus 1767 (type A), the species recently proposed to be defined as Ciona robusta Hoshino and Tokioka 1967 by Brunetti et al. (2015), were obtained from the Misaki Marine Biological Station, University of Tokyo (Miura, Japan) or Maizuru Fisheries Research Station, Kyoto University (Maizuru, Japan) via the National BioResource Project (NBRP), Japan. Animals were maintained for weeks in laboratory tanks containing artificial seawater (ASW; Rei-Sea Marine, Iwaki, or Marine Art BR, Osaka Yakken), which was gently agitated using a polyvinyl chloride paddle attached to a synchronous motor (15 rpm; Nidec Servo) under constant light and room temperature (RT, 18–19°C). Some adults from NBRP were directly used in physiological and histochemical analyses, while others were used for the collection of fertilized eggs to prepare juveniles. Mature eggs and sperm were collected surgically from the gonadal ducts of multiple adults, and eggs were inseminated in ASW (Marine Art BR). Embryos were placed in polystyrene dishes containing ASW and allowed to develop in an air-conditioned room (18–19°C) or cool incubator at 18°C. Larvae hatched at approximately 17 h post-fertilization. Larvae underwent metamorphosis naturally on plastic dishes, and juveniles were fed for approximately 2–3 weeks with a diatom (Chaetoceros calcitrans) that was cultured in the laboratory. Juveniles corresponding to the 2nd ascidian stage (stage 8) defined by Chiba et al. (2004) were harvested after 3 days of starvation for whole-mount in situ hybridization.

The pharyngeal gills of Ciona adults were fixed with 4% paraformaldehyde in buffer containing 0.5 mol l⁻¹ NaCl and 0.1 mol l⁻¹ MOPS (pH 7.5) at 4°C overnight. Fixed specimens were washed several times with phosphate-buffered saline containing 0.1% Tween-20 (PBST). After washing, they were soaked in blocking solution [0.5% w/v blocking reagent (Roche) in PBST] at 4°C for 1 h. Specimens were treated with a mouse monoclonal anti-acetylated tubulin antibody (Sigma-Aldrich T7451, 1:1000 in blocking solution) at 4°C overnight. They were washed with PBST, and then treated with Alexa Fluor 488-Phalloidin (Thermo Fisher, 1:1000) and with a polyclonal goat anti-mouse immunoglobulin G antibody conjugated to Alexa Fluor 594 (Thermo Fisher, 1:1000) in blocking solution at 4°C for 6 h. After several washes with PBST, specimens were stained in combination with Hoechst 33342 (Dojindo, 10 µmol l⁻¹ in PBST) and then observed using laser confocal microscopy (FV1000-D mounted on IX81, Olympus).

AChE histochemical staining is a method that detects cholinergic nerves or their terminals. We performed AChE histochemical staining to visualize ciliomotor nerves using the method originally described by Karnovsky and Roots (1964). Adult gills were fixed using the same fixative as above at 4°C for 1 h, and were then washed with PBST. Specimens were soaked in a reaction mixture containing 0.2 mg ml⁻¹ ACh iodide, 65 mmol l⁻¹ sodium acetate, 3 mmol l⁻¹ copper sulfate, 0.5 mmol l⁻¹ potassium ferrocyanide and 5 mmol l⁻¹ sodium citrate (pH 5.5) at RT for several hours. After thorough staining, specimens were washed with PBST and observed under a light microscope (OPTIPHOT, Nikon). After observations, specimens were dehydrated by an ethanol series followed by xylene, and were then embedded in paraffin. Paraffin blocks were sectioned at a thickness of 8 µm, counterstained with Hematoxylin and Eosin, and observed under the light microscope.

To investigate the effects of ACh on ciliary movement at stigmata, we planned to directly inject ACh and other agonists into the blood sinus between the gill slits. The tunic was peeled off of an adult Ciona and cut open from the tip of the oral siphon to the vicinity of the heart along the endostyle (Fig. 1A,B). The specimen was laid on a Petri dish covered with a silicon-rubber pad (SYLGARD 184, Dow Corning) with the inside of the branchial sac facing the silicon pad. Connective tissues and body wall muscles were dissected to disclose the pharyngeal gills, and the specimen was then set by pinning connective tissues on the pad with needles (Fig. 1C). Four to eight stigmata were aligned and surrounded by the transverse and longitudinal vessels (Fig. 1D). This fresh preparation was washed gently with ASW and then soaked in ASW or ASW containing an antagonist [d-tubocurarine (d-TC), α-bungarotoxin (α-BTX), atropine, YM-254890, U-73122 or BIM-1]. Ciliary movement on stigmata was observed under a light microscope (ECLIPSE E600FN and Plan Fluor 10× or Plan Fluor 40× W, Nikon) and recorded with a video camera (HDR-CX420, Sony) or high-speed video camera (VCC-1000, Digimo, Japan) mounted on the microscope. Each specimen was used within 2–4 h of being dissected. Sequential images captured by the high-speed camera for 0.9–4.0 s (512×464 pixels recorded at 250 frame s⁻¹ or 512×232 pixels recorded at 500 frame s⁻¹, with no enhancement in grayscale color) were imported into Image J (v.1.52a), and changes in 8-bit grayscale brightness (0–255) at ‘points of interest’ were analyzed through each image sequence. The ciliary beating frequency was calculated from the duration between one peak to another peak of the brightness change and the number of cycles between these two peaks (average of 4 or more cycles).

A glass needle was prepared from a borosilicate glass capillary (GD-1, Narishige) using a micropipette puller (P-87, Sutter Instrument), and filled with an agonist solution, either of ACh, (–)-nicotine (Nic), carbachol (Car) or choline, dissolved in ASW. The needle was sealed with back-filled mineral oil, and then it was connected to a water-pressure syringe and set on a micromanipulator (MX-4, Narishige). The needle tip was placed in the blood sinus between stigmata, and the agonist solution was injected manually (Fig. 1E). The injection of agonists elicited ciliary arrest (see Results). Effective strokes of stigmatal cilia are directed from the pharyngeal cavity side to the atrial cavity side. At ciliary arrest, metachronal waves stop and cilia point inside the pharynx, exceeding the initial position of an effective stroke (Takahashi et al., 1973; Mackie et al., 1974; Arkett, 1987; Bergles and Tamm, 1992). After seconds, cilia change their posture from the arrested (directing inside) state to a ‘quiescent’ or ‘inactive’ state, in which cilia stand parallel with the plane of the stigmatal aperture (Takahashi et al., 1973). Ciliary beating to revive metachronal waves starts after cilia take this posture of quiescence. We herein quantified the duration of ciliary arrest evoked by the agonist administration as the period (seconds) from the time when cilia arrested to the time when more than half of arrested cilia assumed the quiescent posture. Although ‘spontaneously occurring’ arrest was observed in fresh preparations of the pharynx, it was sufficiently rare and, thus, the coincidence of this ‘spontaneous’ and our ‘agonist-induced’ response was disregarded.

Full-length cDNAs for all nAChR subunits and a rapsyn homolog in C. intestinalis (C. robusta) were previously obtained and reported (GenBank accession no. AB539786–AB539796; Nishino et al., 2011). These cDNA clones were linearized, and the antisense probe for each was prepared using a digoxigenin labeling mix (Roche) and T3 or T7 RNA polymerase by standard methods. Synthesized RNA probes were fragmented into ∼300 bp by an alkali treatment.

Whole-mount in situ hybridization was performed using a previously described protocol with minor modifications (Ogasawara et al., 2002). Juveniles were relaxed with menthol-saturated ASW, and then fixed with 4% paraformaldehyde in 0.5 mol l⁻¹ NaCl and 0.1 mol l⁻¹ MOPS (pH 7.5) at 4°C for 12 h. Specimens were stored in 80% ethanol at −30°C and rehydrated in PBST when being prepared for use. To remove the tunic, specimens were treated with 10 µg ml⁻¹ proteinase K in PBST at RT for 40 minutes (min) and shaken by gentle pipetting. After washing with PBST, specimens were postfixed with 4% paraformaldehyde in PBST (pH 7.5) at RT for 1 h. After thorough washes with PBST, juvenile specimens were set into the InSitu Chip (ALOKA), and prehybridized with hybridization buffer containing 50% formamide, 5× SSC, 50× Denhardt's solution, 0.1 mg ml⁻¹ yeast tRNA and 0.1% Tween-20. To avoid sudden changes in the densities of solutions, the following procedure was repeated ten times: (i) removing 20 μl of the solution soaking specimens (200 μl), (ii) adding 20 μl hybridization buffer to the tube, (iii) mixing gently, and (iv) leaving specimens to stand for 5 min, which allowed the concentration of hybridization buffer to gradually increase. After another whole exchange of solution with fresh hybridization buffer, specimens were maintained in this buffer at 42°C for 2 h. Juvenile specimens were then soaked in hybridization buffer containing probes at 42°C for 16 h. Specimens were washed at 50°C for 20 min each using a series of washing buffers containing 50% formamide, 0.1% Tween-20 and 4×, 2×, 1× or 0.1× SSC. Thereafter, washing buffer was gradually replaced with Solution A (0.5 mol l⁻¹ NaCl, 5 mmol l⁻¹ EDTA, 10 mmol l⁻¹ Tris-HCl, pH 8.0, and 0.1% Tween-20), and excess probes were digested with RNase A (20 μg ml⁻¹) in Solution A at 37°C for 30 min. The gradual replacement of solution into washing buffer containing 1× SSC and 0.1% Tween-20 allowed the digested probes to be removed. Specimens were soaked in blocking solution [0.5% w/v blocking reagent (Roche) in PBST] and then in the presence of 1:2000 alkaline phosphatase-conjugated anti-digoxigenin antibody (Fab fragment, Roche) in blocking solution at 4°C overnight. After thorough washing of non-reacted antibodies using PBST, the probe was detected with NBT/BCIP (Roche). After a few PBST washes, stained specimens were observed under a differential interference contrast (DIC) microscope (BX51, Olympus).

The coding sequences of A7/8-1 were amplified from previously reported full-length cDNA (GenBank accession no. AB539794) (Nishino et al., 2011) by PCR and were subcloned into the XhoI-NotI site of pSD64TF (a gift from Dr Terrance Snutch, University of British Columbia, Canada). After linearization of the plasmid with BamHI, cRNA was synthesized according to the procedures provided in the manufacturer's kit (mMessage mMachine, Thermo Fisher). We similarly prepared cRNAs of nAChR-A1, -B2/4, and -BGDE3 and a mutant of the latter named nAChR-BGDE3-E261GQ (see Nishino et al., 2011) and utilized them as references (see Fig. S1).

Xenopus laevis oocytes were prepared according to a standard method (Goldin, 1992). Experiments were performed as previously described (Nishino et al., 2011) and in accordance with the official guidelines for animal care. cRNA encoding A7/8-1 or mixtures of others were injected into defolliculated oocytes (50 nl, 0.02–0.07 µg µl⁻¹ for each subunit). After an incubation at RT for 2–4 days in ND96⁺⁺ solution (96 mmol l⁻¹ NaCl, 2 mmol l⁻¹ KCl, 2 mmol l⁻¹ CaCl₂, 1.8 mmol l⁻¹ MgCl₂, 5 mmol l⁻¹ HEPES, pH 7.5, 550 μg ml⁻¹ sodium pyruvate, 10 U ml⁻¹ penicillin, and 10 µg ml⁻¹ streptomycin), oocytes were placed in a cylindrical chamber with an approximate volume of 1 ml and equipped with a perfusion system. Agonist-induced currents were measured by a conventional two-electrode voltage clamp using a ‘bath-clamp’ amplifier (OC-725C, Warner Instruments). In most experiments, 50 nl of 25 mmol l⁻¹ BAPTA [1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid] was injected into oocytes at least 15 min before recording to block intrinsic Ca²⁺-induced currents. Microelectrodes were filled with 3 mol l⁻¹ KCl and their resistance was set to 0.1–0.8 MΩ in normal external solution, OR2 (82.5 mmol l⁻¹ NaCl, 2 mmol l⁻¹ KCl, 1 mmol l⁻¹ MgCl₂ and 5 mmol l⁻¹ HEPES, pH 7.5) or ND96 (96 mmol l⁻¹ NaCl, 2 mmol l⁻¹ KCl, 2 mmol l⁻¹ CaCl₂, 1.8 mmol l⁻¹ MgCl₂, and 5 mmol l⁻¹ HEPES, pH 7.5) containing 2 µmol l⁻¹ atropine. The oocyte was clamped at −40 mV, and 2.5 µl of the agonist solution (ACh, Nic, or Car) was puffed onto the cell surface. The agonists added were sometimes removed by perfusion of the bath solution, which was driven by a peristaltic pump, with the corresponding external solution. Agonist-induced currents were monitored with an analog-to-digital/digital-to-analog converter (ITC-16) and software (Pulse, HEKA Electronik) running on a Windows PC. Traces were stored and then analyzed offline using Pulse/Pulsefit (HEKA), CUT (Evergreen, Japan), IGOR Pro (WaveMetrics), R (v.3.5.0) and Excel (Microsoft).

A ramp pulse protocol (from −100 to 100 mV) was applied to oocytes expressing A7/8-1, and the current–voltage (I–V) relationship (the I–V curve) was examined by subtracting the current before puffing from that in the presence of ACh (10 or 100 μmol l⁻¹). To estimate relative Ca²⁺ permeability, reversal potential values were read from the I–V curves recorded in three types of external solutions with different Ca²⁺ concentrations (1.8, 4.0 and 10.0 mmol l⁻¹) as follows: 30 mmol l⁻¹ NaCl, 2 mmol l⁻¹ KCl, 1.8 mmol l⁻¹ CaCl₂, 1 mmol l⁻¹ MgCl₂, 5 mmol l⁻¹ HEPES, and 132 mmol l⁻¹ mannitol; 30 mmol l⁻¹ NaCl, 2 mmol l⁻¹ KCl, 4 mmol l⁻¹ CaCl₂, 1 mmol l⁻¹ MgCl₂, 5 mmol l⁻¹ HEPES, and 125.4 mmol l⁻¹ mannitol; 30 mmol l⁻¹ NaCl, 2 mmol l⁻¹ KCl, 10 mmol l⁻¹ CaCl₂, 1 mmol l⁻¹ MgCl₂, 5 mmol l⁻¹ HEPES, and 107.4 mmol l⁻¹ mannitol. Reversal potential values were plotted against corresponding external Ca²⁺ concentrations (see Fig. S1). These data were applied to the extended version of the Goldman–Hodgkin–Katz equation (Mayer and Westbrook, 1987) and approximate relative Ca²⁺ permeability value, P_Ca/P_Na, of the A7/8-1 channel was estimated. In the estimation, intracellular Na⁺, K⁺, and Ca²⁺ concentrations were set at 15, 80 and 0 mmol l⁻¹, respectively. In addition, we assumed that P_K equals P_Na and that ACh-induced Cl⁻ flux during the ramp tests on BAPTA-injected oocytes was negligible. We similarly prepared the cRNAs of nAChR-A1, -B2/4, -BGDE3 and a mutant of the latter named nAChR-BGDE3-E261GQ (see Nishino et al., 2011). A1,B2/4,BGDE3 is an identified nAChR that functions in Ciona larval muscle (Nishino et al., 2011) and A1,B2/4,BGDE3-E261GQ is known as its mutant that lacks Ca²⁺ permeability and inward rectification (Nishino et al., 2011). I–V curves of the oocytes expressing them were recorded in three types of external solutions as above, and the reversal potential values were utilized as references (see Fig. S1).

To clarify the cellular composition of the stigmata of Ciona adults, branchial sacs (pharyngeal walls) were isolated, fixed and labeled with the anti-acetylated tubulin antibody, Hoechst 33342 and Alexa Fluor 488-Phalloidin to highlight cilia, nuclei, and actin microfilaments, respectively (Fig. 2A,B). The anti-acetylated tubulin antibody clearly labeled cilia that aligned around the stigmatal apertures (Fig. 2A). The lateral walls parallel to the long axis of the slit-like aperture were composed of flat multiciliated stigmatal cells containing an elongated nucleus, herein referred to as stigmatal ‘lateral cells’ (Fig. 2A–C). Several of the lateral cells were regularly stacked and aligned to form the ciliated wall of stigmata (Fig. 2B,C). Lateral cells had relatively long cilia of ∼20 µm. By contrast, the distal ends of slits were occupied by a batch of tall cells with slight constriction of the cellular apex, which apparently lacked cilia (Fig. 2A–C). We refer to these cells as ‘distal tall cells’. Several layered multiciliated cells with a trapezial shape that were found between the regions of lateral cells and distal tall cells are herein referred to as stigmatal ‘trapezial cells' (Fig. 2A–C). Trapezial cells appeared to have shorter cilia than those of lateral cells.

To visualize possible cholinergic innervations on Ciona stigmatal cells, we performed AChE histochemical staining. In gill specimens of Ciona adults, 10 μmol l⁻¹ AChE activity was densely detected around trapezial cells, but was absent in and around distal tall cells (Fig. 2D). String-like fine structures were often stained around trapezial and lateral cells, some of which appeared to be innervating the basal side of trapezial cells and branching to form the nerve plexus in the sinus between the lateral cells of neighboring stigmata (Fig. 2D, arrows). Moderate staining was also found along the apical side (cilia side) of lateral cells (Fig. 2D). Cross-sections of stained stigmata showed dense historeactivity on the basal side (the sinus side, Fig. 2E, large arrowheads) and weaker historeactivity on the apical side of ciliated cells (Fig. 2E, small arrowheads). These results illustrate cholinergic innervations on stigmatal cells, especially dense innervations on trapezial cells, in Ciona.

To examine the direct effects of nAChR agonists on the ciliary movements of Ciona stigmata, we established a method to open the pharyngeal wall of living adult Ciona (Fig. 1A–D). Using a micromanipulator, the tip of a glass needle filled with an agonist solution was placed in the blood sinus between stigmata, and the agonist was then applied to the basal side of ciliated cells (Fig. 1E). The direct application of 10 µmol l⁻¹ ACh reliably induced ciliary arrest responses in ciliated cells (Fig. 3A–E; Movie 2). The region at which arrest occurred was restricted to the vicinity of the sites of the ACh injection (Movie 2). The sequence of events after the ACh injection occurred in parallel with those observed in intact animals in response to the cutaneous stimulation (cf. Movie 1): (1) acute inclination of cilia with the cessation of beating (arrested), (2) standing up of cilia (quiescent), (3) re-initiation of beating, and (4) recovery of metachronal waves (Fig. 3A–E). Using a high-speed video camera, we compared the beat frequencies of cilia before the application of ACh and after the recovery of metachronal waves and confirmed that ciliary activity was restored to the level before arrest; from 9.6±2.0 Hz (before) to 9.4±1.5 Hz (after) (mean±s.d., n=6) (Fig. 3F).

We then recorded the duration of ciliary arrest evoked by the administration of agonist and found a clear dose dependency in the duration of ciliary arrest (Fig. 4A, Movie 2). Our trials to inject 0.3 μmol l⁻¹ ACh or ASW only evoked no ciliary arrest responses (Fig. 4A,B; Movie 2). By contrast, arrest responses were elicited by the agonists Car and Nic; the duration of arrest evoked by 100 µmol l⁻¹ of Car or Nic was slightly or markedly shorter, respectively, than that evoked by the same concentration of ACh (Fig. 4B). The administration of choline did not induce ciliary arrest at any of the concentrations tested (0.3–100 µmol l⁻¹; 17 trials in total). The addition of 100 µmol l⁻¹ d-TC to the bath solution (ASW) inhibited ciliary arrest elicited by 10 µmol l⁻¹ ACh (Fig. 4C). α-Bungarotoxin (α-BTX) was less effective at inhibiting ciliary arrest; a considerable number of trials of 10 µmol l⁻¹ ACh injection still evoked arrest, even at 10 µmol l⁻¹ α-BTX (Fig. 4C). Treatment with atropine (an inhibitor of muscarinic AChR) did not induce prominent changes (Fig. 4D). These results provide direct evidence that ACh and nicotinic, not muscarinic, receptors, which are less sensitive to α-BTX, mediate the ciliary arrest response in Ciona stigmata.

We attempted to clarify the molecular identity of the mediator on ciliated cells. Since we previously identified all nAChR subunit genes (10 in all) present in the Ciona genome (Okamura et al., 2005; Nishino et al., 2011), we performed whole-mount in situ hybridization on all 10 nAChR subunit genes. We found that only a single gene was expressed on juvenile stigmata, which encoded Ci-nAChR-A7/8-1 (herein referred to as A7/8-1) and that probes for the other nine genes showed no clear signal on stigmata (Fig. 5). The expression domains of these nine genes appeared to vary; the cerebral ganglion expressed Ci-nAChR-B2/4, -B3, -oB and -BGDE1 (Fig. 5C–F). Ci-nAChR-BGDE2 appeared to be expressed in body wall muscles (Fig. 5G), while the probes for Ci-nAChR-A3/6, -B3, -BGDE3 and -A7/8-2 showed relatively strong expression signals in the heart (Fig. 5B,D,H,J). The expression of the A7/8-1 gene was detected in the cerebral ganglion and possibly on the endostyle, in addition to the expression in stigmata (Fig. 5I). The expression signal of the A7/8-1 gene in stigmata was dense on trapezial cells, whereas neither tall nor lateral cells showed clear signals (Fig. 5K).

Our molecular phylogenetic analysis confirmed the relative kinship of A7/8-1 to the vertebrate α7/α8 clade of the nAChR gene family (Fig. 5L). This result also demonstrated that the sequence of A7/8-1, as well as that of another relative called A7/8-2, had considerably diverged from the other α7/α8-class nAChR subunits of invertebrates and vertebrates (Fig. 5L).

Vertebrate α7 receptors represent α-BTX-sensitive nAChRs, in which IC₅₀ to α-BTX is as low as ∼1 nmol l⁻¹ (e.g. Couturier et al., 1990). We examined the amino acid sequence of A7/8-1 around a residue that determines sensitivity to α-BTX (Barchan et al., 1995; Balass et al., 1997), which is close to the neighboring Cys pair that constitutes an essential part of the ACh-binding pocket (Fig. 5M, red). The nAChR α subunits, which are known to be highly sensitive to α-BTX, have an aromatic amino acid (Phe or Tyr) in the diagnostic site (Fig. 5M, colored cyan, arrowhead), whereas less-sensitive subunits have a non-aromatic amino acid (Fig. 5M) (Barchan et al., 1995; Balass et al., 1997). We found that this diagnostic residue of Ciona A7/8-1 was non-aromatic Lys (Fig. 5M, rectangle), similar to that in mammalian α3 or α4, which is resistant to α-BTX.

To establish whether the A7/8-1 subunit forms a functional nAChR channel, we introduced its synthetic cRNA into Xenopus oocytes. Under a two-electrode voltage clamp configuration, dose-dependent inward currents were elicited in response to the application of ACh onto oocytes (Fig. 6A,B). The vertebrate α7 nAChRs have also been reconstituted on exogenous expression systems including Xenopus oocytes, and their molecular properties have well been characterized so far; e.g. steep inward rectification and extremely high Ca²⁺ permeability (Couturier et al., 1990; Bertrand et al., 1993). We examined the current-voltage (I–V) relationship of ACh-induced currents through the A7/8-1 channel, and found that it exhibited a pattern of inward rectification (Fig. S1A). To estimate the relative Ca²⁺ permeability of A7/8-1, we analyzed reversal potential changes in extracellular solutions with three different Ca²⁺ concentrations using, as references, a previously studied Ciona larval muscle nAChR called A1,B2/4,BGDE3, which possesses relatively high Ca²⁺ permeability and steep inward rectification, and also its mutant A1,B2/4,BGDE3-E261GQ that exhibits almost no Ca²⁺ permeability and rectification (see Nishino et al., 2011). This comparative analysis demonstrated that the relative Ca²⁺ permeability of A7/8-1 was lower than that of Ciona larval muscle nAChR A1,B2/4,BGDE3 (P_Ca/P_Na≈2; Nishino et al., 2011), and, thus, markedly lower than that of vertebrate α7 nAChR (P_Ca/P_Na≈10; Bertrand et al., 1993) (Fig. S1).

While A7/8-1 was not responsive to glycine (Gly) or norepinephrine (noradrenaline, NE) (Fig. 6B), this receptor channel responded to Car and Nic in dose-dependent manners. The A7/8-1 channel expressed in Xenopus oocytes appeared to be less sensitive to Car or Nic than to ACh (Fig. 6A–E). A7/8-1 responded to Car by ‘tonic’ persistent currents, similar to its response to ACh (Fig. 6A,C), while its response to Nic showed a ‘phasic’ curve (Fig. 6D). Mammalian α7 nAChR is known to be responsive to choline, whereas A7/8-1 expressed on Xenopus oocytes showed no evident currents in response to the choline application (up to 1 mmol l⁻¹) (Fig. 6B).

When oocytes expressing A7/8-1 were treated in advance with 10 μmol l⁻¹ d-TC, the currents evoked by 10 μmol l⁻¹ ACh markedly decreased (Fig. 6F,H). While a treatment with 1 μmol l⁻¹ d-TC did not exert evident effects, a treatment with 100 μmol l⁻¹ d-TC abolished 10 μmol l⁻¹ ACh-induced currents (Fig. 6F,H). The inhibitory effects of α-BTX were also examined. Although 10 μmol l⁻¹ ACh-induced currents through A7/8-1-expressing oocytes appeared to slow down to reach the peak amplitude, the current amplitudes themselves did not remarkably decrease, even in 10 μmol l⁻¹ α-BTX (Fig. 6G,H). These results demonstrate that exogenously expressed Ciona A7/8-1 forms a functional homomeric nAChR channel by itself, the properties of which differ in several respects from those in vertebrate α7 nAChR, but are consistent with those observed in vivo in Ciona pharyngeal gills.

Arkett (1987) showed that a class of central neurons in the cerebral ganglion of Chelyosoma exhibited spiking activity that was synchronized with APs on stigmatal ciliated cells. The duration of these APs recorded on ciliated cells was approximately 600 ms or shorter, while cilia arrested for seconds. This gap in duration implies the involvement of some signaling system(s) to sustain the arrested status. Bergles and Tamm (1992) reported that Ciona stigmata soaked in a solution containing a Ca²⁺ ionophore as well as a high concentration of Ca²⁺ (100 mmol l⁻¹) exhibited ciliary arrest that lasted for seconds, suggesting that sustained elevation of intracellular Ca²⁺ underlies occurrence of ciliary arrest. We found evident blockade of ciliary arrest when we treated a fresh preparation of pharyngeal gills in advance with 1–10 µmol l⁻¹ U-73122, a representative inhibitor of phospholipase C (PLC) that produces inositol 1,4,5-trisphosphate (IP₃) to stimulate a sustained increase in intracellular Ca²⁺ (Movie 3; Fig. 7). Treatment with 3 μmol l⁻¹ of the PLC inhibitor damped the responsiveness of the ciliated cells to 10 μmol l⁻¹ ACh (Fig. 7), which clearly blocked the acute inclination of cilia, but did not necessarily inhibit the cessation of beating (Movie 3). At 10 μmol l⁻¹, the PLC inhibitor almost completely blocked ciliary arrest (Fig. 7; Movie 3). This dose-dependent effect of U-73122 suggests the involvement of PLC in the signal relay leading to ciliary arrest. On the other hand, when we examined the effect of 10 μmol l⁻¹ BIM-1, an inhibitor of protein kinase C (PKC), a representative downstream Ca²⁺ signaling mediator, the duration of ciliary arrest was not different from that of the control (treated only with the solvent, DMSO) (Fig. 7). Also, a treatment with 10 µmol l⁻¹ YM-254890, an inhibitor of G protein-coupled receptor (GPCR) signaling (especially of Gq signaling that is a representative activator of PLC), did not effectively inhibit 10 μmol l⁻¹ ACh-induced ciliary arrest (Fig. 7), suggesting that the functional involvement of GPCR, including Gq-coupled muscarinic AChR, is negligible in this system (Fig. 7).

The present results, including histochemical staining, pharmacological analyses in vivo and electrophysiological tests in an exogenous expression system, provide evidence to show that A7/8-1 nAChR expressed in Ciona stigmata mediates neurociliary transmission to elicit ciliary arrest.

Previous histochemical and pharmacological experiments proposed that this neurociliary transmission in ascidian stigmata is cholinergic (Mackie et al., 1974; Arkett, 1987; Arkett et al., 1989). For instance, Arkett (1987) reported that APs on the branchial sac of Chelyosoma that trigger ciliary arrest were largely inhibited within a small region treated with d-TC (see also Mackie et al., 1974). In Ciona, a major experimental model of ascidians, the results from a recent optogenetic analysis suggested that the activation of central cholinergic neurons innervating stigmata induced ciliary arrest (Hozumi et al., 2015). We herein demonstrated that the stigmatal ciliated cells of Ciona were positive for AChE historeactivity, and that the direct administration of ACh, Car and Nic onto ciliated cells evoked persistent ciliary arrest in fresh preparations of the Ciona adult pharynx. Ciliary arrest induced by agonist injections was only observed in areas close to the site of injection, suggesting that the effects of agonists are not indirect, for example, via the stimulation of remote neurons or their neurites that innervate stigmata, but due to the direct activation of nAChRs on proximate ciliated cells. These results imply that cholinergic transmission via nAChRs leading to ciliary arrest in stigmata is a common feature in ascidians.

By taking advantage of accumulated genomic information and previous findings on the genes encoding nAChR subunits in this species (Okamura et al., 2005; Nishino et al., 2011; Nishino, 2018), we herein identified A7/8-1 as the only nAChR subunit gene with detectable expression levels in stigmatal ciliated cells. Our analyses using the exogenous expression system of Xenopus oocytes revealed that A7/8-1 responded to ACh in a similar dose dependency to that in vivo (cf. Fig. 4A and Fig. 6B). The effectiveness of Car and Nic relative to ACh also appeared to be consistent between stigmatal ciliated cells in vivo and exogenously expressed A7/8-1. The response by ACh was inhibited by d-TC, but was not fully inhibited by α-BTX, with a similar dose dependency to that in vivo (cf. Fig. 4C and Fig. 6H). These results suggest that A7/8-1 is a mediator of the ciliomotor input from nervous systems to evoke ciliary arrest.

We confirmed that A7/8-1 is one of the cognates in Ciona of vertebrate α7/α8-class nAChRs. The molecular functions of α7 nAChR have been extensively examined in mammalian models and shown to be multifaceted and complex (Papke, 2014; Kalkman and Feuerbach, 2016). α7 nAChR is expressed not only in neurons (presynaptically as well as postsynaptically), immature neurons and glia in the central nervous system (CNS) but also in immune cells, including microglia, monocytes, dendritic cells, macrophages, and T and B lymphocytes (Kalkman and Feuerbach, 2016). Extensive studies reported its functional involvement in memory formation, addiction, arousal, neuroprotection and psychiatric disorders, including schizophrenia, autism, Parkinson's disease and Alzheimer's disease. In the periphery, α7 is involved in the regulation of inflammation (see review by Kalkman and Feuerbach, 2016).

To comprehend the multifaceted functions of α7, it is valuable to determine the role of the α7-related nAChRs of invertebrates, such as A7/8-1 in Ciona. Vertebrate homopentameric α7 nAChRs are responsive to EC₅₀≈10–100 µmol l⁻¹ of ACh, permeable to cations with very high selectivity to Ca²⁺ (P_Ca/P_Na≈10 or more), rapidly desensitized to ACh and other agonists, and blocked by 10 or less nmol l⁻¹ of α-BTX and >10 µmol l⁻¹ d-TC (e.g. Couturier et al., 1990; Bertrand et al., 1993; Séguéla et al., 1993; Jonsson et al., 2006). Our analyses of the I–V relationship of A7/8-1 showed that the current through this receptor channel was inwardly rectified, which is in accordance with the properties of vertebrate α7 (Bertrand et al., 1993; Séguéla et al., 1993). On the other hand, A7/8-1 exhibits some properties that differ from those of homomeric α7. A7/8-1 was less permeable to Ca²⁺ (estimated P_Ca/P_Na≈0.3, Fig. S1), and insensitive to choline as an agonist and to α-BTX as an antagonist. A7/8-1 did not appear so heavily desensitized to ACh or Car (Fig. 6), although pharmacological properties on the desensitization of receptor channels should be more strictly evaluated using devices for fast perfusion (e.g. Costa et al., 1994; Baburin et al., 2006; Papke, 2014). Assuming that A7/8-1 functions as a homopentamer, comparative analyses with α7 will be valuable to identify the amino acids responsible for these differences (see also Papke, 2014); e.g. chimeras may be constructed between A7/8-1 and α7 that exhibit ∼40% amino acid identity in their alignment. Insensitivity to α-BTX may be partly attributed to amino acid usage in the ACh-binding pocket, as indicated in Fig. 5. Relative Ca²⁺ permeability is also known to be largely affected by amino acid usage in the second transmembrane domain (TM2) that lines the channel pore (Bertrand et al., 1993; Fucile, 2004). Differences were observed in the alignment of α7/α8 and A7/8-1 TM2 sequences (Fig. S1), one of which corresponded to the α7 L254Q mutation reported by Bertrand et al. (1993) that markedly decreased Ca²⁺ permeability.

One important difference between our results in vivo from the Ciona pharynx and those from the Xenopus oocyte system is that dose dependency is represented by the magnitude of inward currents (μA) in voltage-clamped Xenopus oocytes, while that in vivo reflects the duration of ciliary arrest (seconds). This implies that we need to consider the mechanisms linking these different dose responses to the agonists.

Electrophysiological studies on the branchial sac of Chelyosoma by Arkett (1987) showed that an acute stimulation on a sensory nerve leading to ciliary arrest on stigmata elicited an AP on stigmatal ciliated cells. The duration of AP on ciliated cells was approximately 600 ms or less, while ciliary arrest lasted for seconds. Therefore, the dose-responsive effect of ACh demonstrated in vivo suggests a mechanism not in an all-or-none but a graded fashion, although ciliary arrest accompanies an AP. The dose-dependent amplification of the intracellular Ca²⁺ signal in response to A7/8-1 activation may be a potential mechanism for this, because a previous study reported that the application of an external solution containing a Ca²⁺ ionophore together with a high concentration (100 mmol l⁻¹) of Ca²⁺ to Ciona stigmata induced ciliary arrest that lasted for 5–10 s (Bergles and Tamm, 1992). This may also be supported by the present result showing that the different agonists that induced responses with different kinetics in Xenopus oocytes exerted similar effects on stigmata in vivo (cf. Fig. 4 and Fig. 6). We did indeed find that the inhibitor of PLC clearly blocked arrest with a similar dose dependency to that in mammalian systems involving PLC-dependent Ca²⁺ signaling (e.g. Bleasdale et al., 1990; Smith et al., 1990; Thompson et al., 1991; Yule and Williams, 1992). PLC is the enzyme responsible for producing IP₃, a second messenger to stimulate Ca²⁺ release from intracellular stores via the IP₃ receptor (e.g. Mikoshiba, 2015). The cascade involving PLC and IP₃ may be a reasonable signal amplifier to enable a persistent arrested state. We cannot conclude here how ligand binding to A7/8-1 activates PLC in stigmatal cells, while involvement of metabotropic signaling relay via GPCR including muscarinic AChR would be excluded. This may involve the direct stimulation of PLC by Ca²⁺ influx (Lomasney et al., 1999), at least partly through A7/8-1 itself, and also potentially through voltage-gated Ca²⁺ channels (VGCCs) that mediate AP on ciliated cells. Our newly established preparation of Ciona stigmata will contribute to the clarification of this issue.

The present results showed that the major factors involved in cholinergic transmission, AChE and A7/8-1, are concentrated on trapezial cells. The longitudinal walls of a stigma are mostly occupied by lateral cells, and the water current through a stigma is mainly powered by lateral cells, not trapezial cells. Lateral cells had long cilia, while trapezial cells and tall cells had short and no cilia, respectively (Fig. 2A–C) (Shimazaki et al., 2006). The absence of expression signal of A7/8-1 in lateral cells appears to be contradictory to the trigger for ciliary arrest being attributed to the function of this receptor in trapezial cells. One possibility is that lateral cells receive APs and/or second messengers, such as Ca²⁺ and IP₃, propagating laterally via gap junctional communications (GJCs). The presence of GJCs among stigmatal cells has already been reported (Mackie et al., 1974; Arkett et al., 1989; Lane et al., 1995). Assuming the presence of GJCs between trapezial cells and lateral cells, a signal from cholinergic central nerves is received by A7/8-1 on trapezial cells and elicits APs, and/or increases in cytosolic Ca²⁺ and IP₃, which propagate along lateral cells via GJCs, leading to ciliary arrest in stigmata. This is consistent with a previous finding in Chelyosoma showing that a treatment with octanol, a general inhibitor of GJCs, hampered the occurrence of ciliary arrest (Arkett, 1987). Trapezial cells may represent a focal site to sense and relay neurociliary signals in Ciona stigmata.

Previous findings on gene expression patterns in stigmata by Shimazaki et al. (2006) indicate other prospects. Transcripts encoding components of cilia, such as tubulins, meichroacidin (a radial spoke protein homolog), tektin isoforms and rootletin, were all expressed in region C, in which trapezial cells reside (Fig. 2C), while their expression signals were weak in lateral cells and absent in tall cells, reminiscent of the pattern for A7/8-1 (Fig. 5K). This result implies that the weak expression signal of transcripts in stigmata does not necessarily correspond to the scarceness of protein products because lateral cells harbor long and dense cilia. Shimazaki et al. (2006) also reported that the number of stigmata in growing Ciona increased by the fission of existing stigmata or formation of new stigmata from small circular perforations. Early perforations are encircled by tall cells at their birth. However, as Ciona grows, a pair of regions appear in the pro-stigma that express both types of genes that characterize trapezial cells or lateral cells. Then, regions B, C and D for tall, trapezial and lateral cells, respectively (Fig. 2C), become evident and independent of each other to show mutually exclusive gene expression patterns (Shimazaki et al., 2006). This finding indicates that lateral cells retain not only ciliary components, but also the A7/8-1 protein that had been translated in the early stage of stigma formation. Our result showing that AChE historeactivity was also detected on fine fibers around lateral cells and on the apical surface of lateral cells is not contradictory to this possibility; A7/8-1 is also present on lateral cells to receive ACh.

The frequency and direction of ciliary movements are regulated in many different organisms. In ciliates such as Paramecium, the avoidance response or reversal of ciliary effective beats may be elicited in response to a collision on the anterior surface, which is accompanied by the depolarization of the membrane potential and increases in intracellular Ca²⁺ (Naitoh and Eckert, 1969). Ciliary arrest is also a fundamental response that has been observed in a wide variety of organisms, from possible flagellar arrest in sponges and arrest in the colorful comb plates of ctenophores to the arrest of ciliary bands in invertebrate larvae, the gill lamellae of bivalves and the stigmatal cilia of ascidians (Carter, 1926; MacGinitie, 1939; Kinosita and Murakami, 1967; Mackie et al., 1974; Murakami and Takahashi, 1975; Satir et al., 1976; Moss and Tamm, 1986; Leys et al., 1999; Verasztó et al., 2017). In ciliated cells along the gill lamellae of bivalves, the AP and accompanying increases in intracellular Ca²⁺ cause the occurrence of ciliary arrest (Takahashi and Murakami, 1968; Walter and Satir, 1978; Saimi et al., 1983), which may involve a similar mechanism to that described here for ascidian stigmata (Takahashi et al., 1973; Mackie et al., 1974; Arkett, 1987; Bergles and Tamm, 1992; and the present study). A previous analysis revealed that the AChE histochemical reaction was detected in the ciliary spiracles of the pair of gill apertures of the appendicularian Oikopleura, which belongs to a separate class from ascidians in the tunicate (urochordate) clade (Nishino et al., 2000). It is also important to note that ciliary movement control in phototactic polychaete larvae is under cholinergic ciliomotor neurons innervating α9/α10-class nAChR-expressing ciliated cells on the ciliary band (Jékely et al., 2008; see also Fig. 5L).

The muscarinic regulation of ciliary beating has also been reported for cilia on the mammalian tracheal epithelium (Salathe et al., 1997; Klein et al., 2009). We did not observe any effects of the inhibition of muscarinic ACh receptors on ciliary arrest responses, but cannot exclude the possible involvement of metabotropic receptors in transducing neuronal signals in stigmata. For example, the rate of ciliary beating in Ciona pharyngeal gills may be negatively related to gut fullness, which is also under the control of the nervous systems (Petersen et al., 1999). This graded control of the beating frequency appears to be through different pathways from that proposed in the present study (see also Cima and Franchi, 2016). The preparation method of Ciona pharyngeal gill slits established here will help us to elucidate further control mechanisms of ciliary movement.

We thank Drs Satoe Aratake, Reiko Yoshida, Manabu Yoshida, Yutaka Satou and other staff at Misaki Marine Biological Station, Grad Schl of Sci, University of Tokyo and in Lab of Dev Genomics, Grad Schl of Sci, Kyoto University, who distribute Ciona under the National BioResource Project, AMED, Japan. Dr Kogiku Shiba (University of Tsukuba) contributed to data analyses and the editing of movies. We are grateful to Dr Kazuo Inaba (University of Tsukuba) for his encouragement. A7/8-1 was originally obtained by A.N. under the supervision of Dr Yasushi Okamura (Osaka University), who we also thank for critically reading the manuscript. We thank Dr Fumihito Ono (Osaka Medical College) for his discussions and critical reading of the manuscript. We are also grateful to anonymous reviewers for valuable comments that substantially improved the manuscript.