Welcome to our new website

Nervous control of ciliary beating by Cl-, Ca2+ and calmodulin in Tritonia diomedea
Owen M. Woodward, A. O. Dennis Willows


In vertebrates, motile cilia line airways, oviducts and ventricles. Invertebrate cilia often control feeding, swimming and crawling, or gliding. Yet control and coordination of ciliary beating remains poorly understood. Evidence from the nudibranch mollusc, Tritonia diomedea, suggests that locomotory ciliated epithelial cells may be under direct electrical control. Here we report that depolarization of ciliated pedal epithelial (CPE) cells increases ciliary beating frequency (CBF), and elicits CBF increases similar to those caused by dopamine and the neuropeptide, TPep-NLS. Further, four CBF stimulants (zero external Cl-, depolarization, dopamine and TPep-NLS) depend on a common mode of action, viz. Ca2+ influx, possibly through voltage-gated Ca2+ channels, and can be blocked by nifedipine. Ca2+ influx alone, however, does not provide all the internal Ca2+ necessary for CBF change. Ryanodine receptor (RyR) channel-gated internal stores are also necessary for CBF excitation. Caffeine can stimulate CBF and is sensitive to the presence of the RyR blocker dantrolene. Dantrolene also reduces CBF excitation induced by dopamine and TPep-NLS. Finally, W-7 and calmidazolium both block CBF excitation by caffeine and dopamine, and W-7 is effective at blocking TPep-NLS excitation. The effects of calmidazolium and W-7 suggest a role for Ca2+-calmodulin in regulating CBF, either directly or via Ca2+-calmodulin dependent kinases or phosphodiesterases. From these results we hypothesize dopamine and TPep-NLS induce depolarization-driven Ca2+ influx and Ca2+ release from internal stores that activates Ca2+-calmodulin, thereby increasing CBF.


The control mechanisms for ciliary beating remain unclear. This is despite the fact that beating cilia have numerous and critical roles in most organisms ranging from locomotion in ciliates (Eckert, 1972; Noguchi et al., 1991; Pernberg and Machemer, 1995), marine invertebrate larvae (Mackie et al., 1969; Mackie et al., 1976; Wada et al., 1997) and nudibranchs (Audesirk, 1978; Willows et al., 1997) to the chordate development of left-right symmetry, circulation of cerebral spinal fluid spinal in brain ventricles, clearing of mucus in airways and movement of eggs along fallopian tubes (Ibanez-Tallon et al., 2003; Snell et al., 2004).

Ciliary beating may be under hormonal control (Verdugo, 1980; Korngreen and Priel, 1996; Korngreen et al., 1998; Lieb et al., 2002; Barrera et al., 2004) or nervous control via dopamine and neuropeptides (Willows et al., 1997; Woodward and Willows, 2006), serotonin (Goldberg et al., 1994; Christopher et al., 1996; Willows et al., 1997; Nguyen et al., 2001; Doran et al., 2004), acetylcholine (ACh) (Salathe and Bookman, 1999; Zagoory et al., 2001; Zagoory et al., 2002) or depolarization (Aiello and Guideri, 1964; Mackie et al., 1969; Mackie et al., 1976; Murakami and Takahashi, 1975). Work permitting any distinction between these control mechanisms is sparse and this may be due, in part, to technical challenges presented by vertebrate models. Other systems such as Tritonia diomedea, in which both ciliated epithelia and neurons are accessible, present an opportunity to study both neural and transduction pathways controlling beating of individual ciliated cells as well as coordinating beating across epithelia.

The intracellular pathways involved in transducing excitatory signals into increased ciliary beat frequency (CBF) have begun to be understood. A key step in increasing CBF in multicellular organisms or reversing directions, in e.g. paramecium (Eckert, 1972), is an increase in internal free Ca2+ (Verdugo, 1980; Villalon et al., 1989; Salathe and Bookman, 1995; Korngreen and Priel, 1996). Studies of CBF have yielded different possible mechanisms for increasing internal free Ca2+, including the phospholipase C (PLC) pathway (Christopher et al., 1999; Zagoory et al., 2001), both protein kinase C (PKC) (Gertsberg et al., 1997; Christopher et al., 1999; Barrera et al., 2004) and protein kinase A (PKA) (Braiman et al., 1998; Zagoory et al., 2002; Lieb et al., 2002), inositol 1,4,5-triphosphate (Barrera et al., 2004), nitrous oxide (Uzlander and Priel, 1999; Runer and Lindberg, 1999; Doran et al., 2003) and cAMP (Stommel and Stephens, 1985; Aiello, 1990). The next step in signal transduction after an increase in internal calcium may be direct binding of Ca2+ to the dynein motor (Salathe and Bookman, 1999) or Ca2+-calmodulin activation of kinases or phosphodiesterases (Zagoory et al., 2001), or possibly both. Unfortunately, little is known of possible electrical control of these pathways, though most early results were consistent with such a hypothesis (Aiello and Guideri, 1964; Mackie et al., 1969; Mackie et al., 1976; Murakami and Takahashi, 1975; Saimi et al., 1983a; Saimi et al., 1983b). This report provides evidence that CBF may be under electrical control and describes an intracellular pathway necessary for transduction.

Fig. 1.

Methods for recording ciliary beating frequency (CBF). (A) Pedal ciliated epithelial tissue is removed from the foot margin of Tritonia diomedea and dissected to small pieces with vigorously beating cilia, accessible for experimentation. (B) Phototransducer measurement of a video projection of beating cilia produces an oscillatory voltage output corresponding to the ciliary beating rate. (C) Using a Fast Fourier transform the dominant frequency from the voltage signal can be measured.

In an accompanying paper (Woodward and Willows, 2006), we showed that ciliated pedal epithelial (CPE) cells of Tritonia diomedea possess calcium dependent Cl- currents, and the direct suppression of these currents by dopamine or TPep-NLS leads to an increase in CBF; this result is mimicked by the Cl- channel blocker DIDS. However, the mode of action by which Cl- current blockage increases CBF is not understood. In the present study we test the hypothesis that CPE cells may possess voltage-gated Ca2+ channels, and that Ca2+ influx through voltage-gated channels is necessary for excitation of CBF. Further, we conclude that Ca2+ influx may be only a trigger for Ca2+ dependent Ca2+ release from intracellular stores via ryanodine receptor (RyR) channels. Finally, we find that Ca2+-calmodulin may be involved in CBF excitation through activation of Ca2+-calmodulin dependent kinases and phosphodiesterases. In sum we present a model for nervous system control of ciliary beating frequency, mediated by transmitter or neuropeptide release.

Materials and methods

Animals, explants and isolated cell preparations

Tritonia diomedea Bergh were trawled from Bellingham Bay, WA, USA, or collected using SCUBA in Puget Sound, WA, USA. They were maintained in 10°C seawater aquaria, and fed sea pens (Ptilosarcus gurneyi) either at Friday Harbor Laboratories (open circulation) or at the University of Washington, Department of Biology (recirculating aquaria).

For experiments, explants of the ciliated pedal epithelium were used. Pedal tissue (2-3 mm square) was dissected from the posterior two thirds of the≈ 20 cm long foot surface. Removal of small foot sections neither reduces the life span, nor alters eating or copulating behaviors. The pedal tissue was pinned out on a Sylgard™ coated Petri dish and the epithelial layer dissected away from the foot musculature. This produced sheets of tissue consisting primarily of epithelial cells. The sheets were further reduced to 100-500 cells (Fig. 1A). After dissection at room temperature, these explants were cooled to 10°C for> 30 min before experimentation. The explants were placed on a large coverslip (24 mm×60 mm), immobilized with a matrix of cotton fibers, and covered by a second smaller coverslip (22 mm×22 mm). Using two thin coverslips rather than glass slides increased the depth of field for the 40× objective and increased the heat transfer from the solution for more efficient cooling. Recovery of explants was indicated by beating cilia and absence of mucus secretion.

CPE cells were isolated from explants following published methods (Pavlova and Bakeeva, 1993): explants were dissected as finely as possible, then reduced to individual ciliated cells by drawing them in and out of a pipette repeatedly. The resulting cells were pipetted into 20 μl of seawater on a large coverslip. Cotton in the solution provided a substrate to which cells could stick, and a second coverslip was placed on top. Thus immobilized, cells and their beating cilia could be visualized and recorded for CBF measurements. Individual cells were considered viable if their cilia were still beating spontaneously.

Solutions and chemicals

For the CBF experiments we used filtered seawater (FSW) (Millipore, 0.22μ M filter), artificial seawater (ASW), 30 mmol l-1 K+ seawater, zero Ca2+ seawater, and zero Cl- seawater. The ASW contained (in mmol l-1): 400 NaCl, 10 KCl, 10 CaCl2, 50 MgCl2, 10 Hepes, pH 8.0. 30 mmol l-1 K+ seawater contained: 380 NaCl, 30 KCl, 10 CaCl2, 50 MgCl2, 10 Hepes, pH 8.0. Zero Ca2+ seawater contained: 400 NaCl, 10 KCl, 60 MgCl2, 10 Hepes, 5 mmol l-1 EGTA, pH 8.0. Zero Cl- seawater contained: 400 sodium gluconate, 10 potassium gluconate, 10 calcium gluconate, 50 magnesium gluconate, 10 Hepes, pH 8.0.

The total volume of solution on the slide was ≈200 μl and solutions were changed for the CBF experiments using a 100 μl pipette. The solutions were delivered at their final concentrations on one side of the coverslip, then an equal amount was drawn off the other side wicking the solution over the explants or individual cells; a total of 500 μl was perfused for each solution change. Nifedipine (Sigma-Aldrich Corp., St Louis, MO, USA) was made up in ethanol in a 50 mmol l-1 stock solution, which was diluted in seawater to a final concentration of 50 μmol l-1 (final solution contained 0.1% ethanol). Xestospongin C (Sigma-Aldrich Corp.), calmidazolium (Calbiochem, San Diego, CA, USA) and dantrolene (Sigma-Aldrich Corp.) were all prepared from a stock solution in DMSO, then diluted to final concentrations in seawater, with a final DMSO concentration always <0.1%. The final concentrations for the inhibitors nifedipine, calmidazolium, W-7, and xestospongin C were chosen from the literature, taking into account the proven maximal effective dose and the possible non-specific effects of the drugs at high concentrations. For none of the treatments did we use a concentration lower than half of that proven maximally effective in molluscs.

Ciliary beat frequency experiments

Slides containing either explants or isolated cells were viewed with a Nikon inverted microscope scope (Nikon USA) at 40× magnification. The slide rested on a stage cooled to 10-12°C by a Peltier device. Beating cilia were recorded using an Elmo CCD camera #TSN401A (Plainview, NY, USA) scanning at 59.94 Hz. The video was then projected on a monitor. A Photonic Sensor (fiber optic device that measures small changes in light intensity near a small probe) was placed in front of the video projected beating cilia and transduced the CBF into an oscillating voltage (Fig. 1B). The voltage signal was amplified 10× and digitally filtered (bandpass 5-30 Hz) to reduce the raster scan signal from the video monitor. The resulting signal was acquired with a DASH-4U (Astro-Med Inc., West Warwick, RI, USA) digital oscilloscope, and a Fast Fourier Transform analysis used to recover the dominant frequency of voltage oscillations (Fig. 1C). We validated these measurements by using a QuickTime (Apple Computer Inc., Cupertino, CA, USA) video of an artificial `cilium' beating at known rates (2-16 Hz). CBF was sampled just after solution changes and then at 3 min intervals. For blocking experiments with dantrolene, xestospongin C and calmidazolium, explants were incubated in the blocking agent for 60, 30 or 30 min, respectively, prior to experimentation. All experiments ended with washout of the applied drug or drugs, and only one experiment was performed on tissue (in the form of a single explant, or single isolated ciliated cell) from each animal.

Data analysis

All graphs were generated in Sigma Plot 2000 6.10 software (SPSS Inc., Chicago, IL, USA) or in Microsoft Excel software (Microsoft Corporation, Redmon, WA, USA) and all statistical analyses were made using Microsoft Excel software. For most experiments comparisons were made between the mean basal rate of CBF in either FSW or the incubating inhibitor and the maximal change in CBF after treatment, unless otherwise noted. This comparison was performed using a paired Student's t-test. Populations were compared using a two-tailed Student's t-test. Means are reported ± s.e.m., with N values noted.


Depolarization induces increase in CBF

In an accompanying paper (Woodward and Willows, 2006) we present evidence for voltage-gated Ca2+ channels in CPE cells, and hypothesize that the blockage of Cl- channels by dopamine and TPep-NLS leads to depolarization and increases in CBF. Influx of extracellular Ca2+ through voltage-gated Ca2+ channels has also been proposed as a method for increasing [Ca2+]in and CBF (Christopher et al., 1996). In addition, a voltage-gated Ca2+ channel could link Ca2+ entry and membrane voltage, thereby providing a mechanism for electrical control of CBF. To test whether depolarization and Ca2+ influx through voltage-gated channels is necessary for CBF excitation, we applied 30 mmol l-1 K+ seawater to CPE explants to depolarize the cells. Seawater with 30 mmol l-1 K+ increased CBF 134.7±25.0% above basal levels (N=5; P≤0.006) (Fig. 2A). After 3 min in 30 mmol l-1 K+ the excitation declined to 92.7±30.9% above basal levels (N=5; P≤0.04), and after 6 min of 30 mmol l-1 K+ exposure the CBF declined to 52.0±27.8% above basal levels (N=5), no longer significantly different from seawater controls (P≤0.13).

We repeated the 30 mmol l-1 K+ experiments, but in a zero Ca2+ seawater bath, to learn if the depolarization-induced increase in CBF is dependent on the influx of extracellular Ca2+ (Fig. 2A). Initial exposure to 30 mmol l-1 K+ in the zero Ca2+ bath only increased CBF 5.3±11.4% (N=6), a significant decrease from 30 mmol l-1 K+ alone (P≤0.0007). Interestingly, the zero Ca2+ seawater alone produced a marginal increase in CBF compared to controls, though only significantly different at one of three time points (viz. after 6 min, P≤0.02). Finally, we used the specific L-type Ca2+ channel blocker, nifedipine, to identify the Ca2+ channels in CPE cells. 50 μmol l-1 nifedipine initially caused an increase in CBF (N=5; P≤0.02); however, this maximal CBF excitation is reduced compared to excitation caused by 30 mmol l-1 K+ seawater alone (N=5; P≤0.006) (Fig. 2A). To confirm that the 30 mmol l-1 K+ seawater acts directly, we tested individual isolated CPE cells. Under these circumstances, 30 mmol l-1 K+ seawater caused an 83.2±17.3% increase (N=5; P≤0.009) in CBF above basal levels (Fig. 2B). These results confirm that depolarization is acting directly on ciliated cells to cause CBF excitation and that depolarization mimics the excitation mediated by dopamine and TPep-NLS.

Fig. 2.

Depolarization in high external K+ causes CBF excitation. (A) 30 mmol l-1 K+ bath (filled circles) significantly increases CBF as compared to ASW controls (N=5, P≤0.0003) in CPE cell explants. After 3 min, 30 mmol l-1 K+ is still excitatory (P≤0.02) but less so. After 6 min, 30 mmol l-1 K+ is no longer excitatory. 30 mmol l-1 K+ seawater was not excitatory in a zero Ca2+ bath (open squares, N=6), and only transiently so (P≤0.02) in a bath containing 50 μmol l-1 of the Ca2+ channel blocker nifedipine (triangles, N=5). (B) 30 mmol l-1 K+ significantly excites CBF in isolated, individual CPE cells (N=5, P≤0.007). *Significant change from mean basal levels.

Fig. 3.

Ca2+ influx is necessary for increases in CBF. (A) Zero Cl- bath (filled circles) excites CBF (N=5, P≤0.001), and the effect does not diminish over time. The excitatory effect is abolished (N=5, P≤0.10 compared to control) when 50 μmol l-1 nifedipine (Nif; open squares) is included in the bath solution. (B) Excitation induced by 100 μmol l-1 dopamine (DA) and 10 μmol l-1 TPep-NLS (TPep) is also dependent on Ca2+ influx. Zero external Ca2+ (N=5, P≤0.008) and 50 μmol l-1 nifedipine (N=5, P≤0.003) both significantly reduce dopamine excitation. Similarly, zero external Ca2+ (N=5, P≤0.00009) and 50 μmol l-1 nifedipine (N=6, P≤0.01) also significantly reduce TPep-NLS excitation. *Significant change from mean basal levels (A) or significant reduction from stimulated levels (B).

Ca2+ influx is necessary for increases in CBF

Though Ca2+ influx may be necessary for depolarization-induced increases in CBF, it may not be required for other CBF excitatory signals. To confirm the role of Ca2+ influx for all excitation we used nifedipine to block Ca2+ in the presence of a number of known CBF stimulants. Zero external Cl- bath solution increased CBF 117.1±17.4%, a significant increase (Fig. 3A) (N=5; P≤0.001) that mimics the increases seen with Cl- channel blockers (Woodward and Willows, 2006). However, the excitatory effect of a Cl--free bath solution is markedly reduced in the presence of 50 μmol l-1 nifedipine (N=5; P≤0.0009) (Fig. 3A), so much so that the residual excitation is indistinguishable from controls (P≤0.10).

Blocking Ca2+ influx also reduces excitation by dopamine and TPep-NLS. Nifedipine significantly reduced dopamine (N=5; P≤0.003) and TPep-NLS (N=6; P≤0.01) induced CBF excitation (Fig. 3B). Like nifedipine, removal of external Ca2+ also significantly reduced the excitatory effects of TPep-NLS (N=5; P≤0.00009) and dopamine (N=5; P≤0.008) (Fig. 3B). Unlike nifedipine, and especially in the case of dopamine, the effects of zero external Ca2+ were highly variable. In some instances, zero Ca2+ had little or no effect, while in others it resulted in a complete abolition of excitation, yielding a large standard error (±2.54 Hz, corresponding to ±35.6% change from control).

Fig. 4.

Ca2+ release from stores contributes to the excitation of CBF. (A) Caffeine stimulated CBF. Both 1 mmol l-1 (open circles, N=5, P≤0.001) and 10 mmol l-1 (filled squares, N=6, P≤0.0003) caffeine increases CBF, however, the effects of 1 mmol l-1 caffeine decline over time. (B) Caffeine (Caff) excitation is reduced by the ryanodine receptor blocker dantrolene (Dant). 10 mmol l-1 Caffeine (N=6) induced CBF excitation is significantly reduced in 50 μmol l-1 dantrolene (N=7, P≤0.00004). 100 μmol l-1 dopamine and 10 μmol l-1 TPep-NLS excitation is also reduced by 50 μmol l-1 dantrolene (N=6, P≤0.0002; N=6, P≤0.00009, respectively). *Significant change from mean basal levels.

Ca2+ release from stores contributes to excitation of CBF

Increases in [Ca2+]in are well established as a prerequisite for CBF excitation (Verdugo, 1980; Villalon et al., 1989; Salathe and Bookman, 1995; Korngreen and Priel, 1996), but the source of the Ca2+, e.g. from stores or influx, remains unclear. Having already established that Ca2+ influx occurs during CBF stimulation we also wanted to explore the possible contribution of endoplasmic reticulum (ER) stores to [Ca2+]in. We used caffeine, known at high concentrations to activate selectively the ryanodine receptor (RyR) channel located in the ER membrane and control Ca2+ release from stores. Caffeine applied at 1 mmol l-1 (Fig. 4A) increased CBF 150.3±18.9% above basal levels (N=5; P≤0.001); however, the excitatory effect waned over time, decreasing to a 64.5±13.1% change by 6 min. However repetition of the experiment with 10 mmol l-1 caffeine (Fig. 4A) also produced CBF excitation (215.5±24.1% increase above basal levels, N=6; P≤0.0003) and these effects did not diminish after 6 min. We chose therefore to use 10 mmol l-1 caffeine in all subsequent experiments. We also found that 50 μmol l-1 nifedipine had little or no effect (N=5; P≤0.16) on the action of caffeine, a result consistent with the hypothesis that internal sources of Ca2+ alone are sufficient to produce CBF excitation.

To confirm the specificity of caffeine action on RyR channels we used the cell permeable RyR channel blocker, dantrolene (Blackwell and Alkon, 1999; Kawai et al., 2004). The presence of dantrolene significantly reduced excitation caused by 10 mmol l-1 caffeine (N=7; P≤0.00004) (Fig. 4B), confirming that caffeine is working specifically to produce Ca2+ release via RyR channels. Dantrolene also significantly reduced excitation induced by dopamine (N=6; P≤0.0002) and TPep-NLS (N=6; P≤0.00009) (Fig. 4B), further supporting the hypothesis that Ca2+ release from stores is required for CBF excitation in addition to Ca2+ influx.

A second Ca2+ release channel is common in the ER membrane, an inositol 1,4,5-triphosphate (IP3) mediated Ca2+ release channel. We investigated the possible involvement of IP3 mediated Ca2+ release with the potent membrane permeable IP3 mediated Ca2+ release blocker, xestospongin C, that has proven effective in marine organisms (Yazaki et al., 2004). The inclusion of xestospongin C with FSW slightly lowered the basal beating rate (P≤0.01) (Fig. 5A), but did not prevent a large increase in CBF with the addition of dopamine. Dopamine's excitatory effect with or without xestospongin C did not diminish over time. In fact, a comparison of the percent increase in CBF induced by dopamine (198.9±16.2%; N=5) (Fig. 5B) revealed that xestospongin C (223.3±12.5%; N=5) had no affect on excitation (P≤0.27). Similarly TPep-NLS induced increase (126.2±9.2%; N=10) was unaffected by the presence of xestospongin C (158.9±11.9%; N=5; P≤0.06) (Fig. 5B). These results suggest the IP3 pathway is not utilized for dopamine or TPep-NLS induced CBF excitation.

Calmodulin dependent processes necessary for CBF excitation

Finally, we wanted to investigate the pathways involved downstream of the [Ca2+]in increase. There is no consensus as to the relationship between second messenger Ca2+ and the increases in dynein motor activity and subsequent global increases in CBF. It was proposed that Ca2+ might act directly on the axoneme in epithelial cells (Salathe and Bookman, 1999); however, numerous other intermediaries remain possibilities. Calmodulin is one option. Calmodulin is known, once bound with Ca2+, to interact directly with the axoneme (Stommel and Stephens, 1985), and also to activate Ca2+-calmodulin kinases that can phosphorylate targets in the axoneme including beta tubulin (Fabczak et al., 1999). We studied the potential role of calmodulin in the excitatory response to neurotransmitters and neuropeptides by using a number of calmodulin antagonists that work by inhibiting Ca2+-calmodulin dependent kinases and phosphodiesterases, viz. W-7 and calmidazolium. To begin, we tested the calmodulin antagonists on the excitation caused by caffeine.

Fig. 5.

IP3 gated internal stores do not contribute to CBF excitation pathway. (A) CBF excitation caused by 100 μmol l-1 dopamine (DA) in CBF explants either bathed in FSW (open squares, N=5) or SW including 1 μmol l-1 of the IP3 inhibitor xestospongin C (Xesto; filled circles, N=5). (B) Comparisons of percent change in CBF with and without 1 μmol l-1 xestospongin C show it has no effect on dopamine or TPep-NLS (TPep) induced CBF excitation. 100 μmol l-1 Dopamine (N=5, P≤0.27) and 10μ mol l-1 TPep-NLS (N=5, P≤0.06) induced CBF excitation in a bath containing 1 μmol l-1 xestospongin C was indistinguishable from excitation induced by either alone.

Fig. 6.

Calmodulin dependent processes necessary for CBF excitation. Calmodulin antagonists W-7 (50 μmol l-1) and calmidazolium (Calmid; 5μ mol l-1) both significantly reduced 10 mmol l-1 caffeine (Caff) excitation (W-7: N=5, P≤0.003; calmidazolium: N=4, P≤0.006) and 100 μmol l-1 dopamine (DA) excitation (W-7: N=5, P≤0.007; calmidazolium: N=5, P≤0.001). W-7 also reduced 10 μmol l-1 TPep-NLS (TPep) excitation (N=5, P≤0.0001), however, calmidazolium did not (N=3, P≤0.17). *Significant reduction from stimulated levels.

Caffeine presumably works upstream of calmodulin and so should be affected by an antagonist. W-7 (N=5; P≤0.003) and calmidazolium (N=4; P≤0.006) both significantly reduce the caffeine-induced excitation (Fig. 6). This result is consistent with the placement of calmodulin activity after Ca2+ release from stores, not during the Ca2+ influx, though it could be active at both times. Further, both W-7 (N=5; P≤0.007) and calmidazolium (N=5; P≤0.001) reduced dopamine-induced excitation as well (Fig. 6). W-7 also proved effective at reducing TPep-NLS excitation of CBF (N=5; P≤0.0001), however, calmidazolium had no effect (N=3; P≤0.17). The failure of calmidazolium to produce an effect is the only experimental result that suggests a difference between the dopamine and TPep-NLS excitatory pathways. The reduction in CBF caused by W-7 and calmidazolium suggest both the involvement of calmodulin in the CBF control pathway, but also that Ca2+ acts via one of its phosphodiesterase or kinase targets, not directly on the axoneme.


Our results show that dopamine and TPep-NLS act by modulating Ca2+ influx and Ca2+ release from RyR gated stores, and by upregulating cilia motor proteins via the Ca2+-calmodulin pathway; thereby increasing CBF. The link between blockage of Cl- currents (Woodward and Willows, 2006) and Ca2+ influx is critical. We hypothesize that blockage of Cl- is linked to depolarization. Depolarization of cells in explants or isolated CPE cells mimics the increase in CBF caused by dopamine and TPep-NLS. Our findings differ from previous observations in other adult molluscs (Murakami and Takahashi, 1975; Saimi et al., 1983a; Saimi et al., 1983b), larval molluscs (Mackie et al., 1969; Mackie et al., 1976) and other marine invertebrates (Mackie et al., 1974). However, single cell organisms (e.g. Paramecium and Didinium nasutum) respond to depolarization by reversing cilia stroke direction and increasing beat frequency (Eckert, 1972; Pernberg and Machemer, 1995). We were unable to record changes in CPE membrane potentials using the whole cell patch clamp in current clamp due to the inherently leaky nature of the cells. We therefore relied on high K+ extracellular seawater to induce a transient depolarization of CPE cells.

We found that depolarization induced excitation is dependent on Ca2+ influx. Both zero external Ca2+ and nifedipine block depolarization induced CBF excitation, similar to other observations reported for serotinin (5-HT) induced CBF excitation in Helisoma embryos (Christopher et al., 1996). Our results are consistent with the hypothesis that CPE cells possess a small number of voltage-gated Ca2+ channels. This possibility separates molluscan CPE cells from the well-studied ciliated epithelium of vertebrate airways, which show no evidence of excitability, but nonetheless produce elevated plateaus of internal Ca2+ (Braiman et al., 2000). What appears to be similar is the necessity of Ca2+ influx for sustained CBF excitation. We found that excitation by dopamine, TPep-NLS, and high K+, are blocked by nifedipine and zero external Ca2+, indicating that Ca2+ influx is necessary for any CBF excitation. Though the excitable molluscan CPE cells are not similar to vertebrate epithelial cells possessing motile cilia, there are numerous examples of epithelial cells with non motile cilia that possess voltage-gated channels. The pigmented and non-pigmented ciliary epithelial cells of the vertebrate retina possess voltage-gated Ca2+ channels, which may aid in the spread of excitability across the retina (Farahbakhsh et al., 1994), and may occur also in kidney epithelial cells (Nauli et al., 2003).

Fig. 7.

Model of transmitter and neuropeptide action on CPE cells. Binding of dopamine (DA) or TPep-NLS to a receptor leads to a reduction in ICl(Ca), carried by Ca2+ dependent Cl- channels (CaCC) and I(Cl-)leak, also carried by CaCC channels. We hypothesize that the blockage of Cl- currents leads to a depolarization of the cell and the activation of voltage-gated Ca2+ channels (CavC) leading to an influx of Ca2+. This initial influx triggers the release of further Ca2+ from ryanodine receptor channel (RyR)-gated endoplasmic reticulum (ER) stores. The sharp rise of [Ca2+]in activates the Ca2+-calmodulin complex, which in turn, upregulates kinases and phosphodiesterases critical for increasing ciliary beating rate. Also present are voltage activated proton channels that play a significant role in pH maintenance and may activate after long depolarizations to aid in the repolarization of the membrane potential. Pluses indicate molecules or actions that increase current amplitude or beating rate, minuses represent inhibitory molecules or actions.

Finally, we also showed a direct link between Cl- flux and Ca2+ influx. Removing external Cl- excites CBF; however, the excitation is blocked by nifedipine. This provides an alternative to the hypothesis that simple activation of Ca2+ channels by transmitters causes Ca2+ influx in CPE cells. In addition, it establishes a novel involvement of Cl- channels in the regulation of motile ciliary beating control.

Ryanodine receptor (RyR) channel-gated internal Ca2+ stores contribute significantly to [Ca2+]in increases and CBF excitation. Vertebrate ciliary control also depends on Ca2+ stores, often through IP3-gated stores, not RyR (Braiman et al., 2000). Interestingly, RyR channel-gated stores have not been implicated previously in control of CBF. Our results do not exclude the possibility that both IP3- and RyR-gated stores are involved in CBF control. Xestospongin C has been used successfully to block IP3 activated functions widely in animals (Gafni et al., 1997; Barrera et al., 2004; Yazaki et al., 2004), although in our experiments we lacked a positive control for xestospongin C. We did find that xestospongin C depressed basal CBF rates, raising the possibility that the IP3 pathway contributes to both basal and excited CBF, but not to dopamine or TPep-NLS induced excitation.

Our findings also suggest that Ca2+-calmodulin is in the CBF control pathway. We place its role downstream of Ca2+ release from internal stores because W-7 and calmidazolium attenuate the observed caffeine excitation. Calmodulin has been implicated in many of the described CBF control pathways (Stommel and Stephens, 1985; Zagoory et al., 2001; Zagoory et al., 2002). Some researchers attribute W-7 and calmidazolium reduction of CBF excitation to non-specific chelating effects, instead believing that Ca2+ directly binds to ciliary beating mechanism (Salathe and Bookman, 1999). Others support the idea that Ca2+-calmodulin dependent kinases and phosphodiesterases influence dynein behavior (Zagoory et al., 2002). While our methods do not differentiate between these two possibilities, our findings do represent yet another CBF control system influenced by Ca2+-calmodulin. One result worthy of further investigation is the finding that while W-7 does inhibit TPep-NLS excitation, calmidazolium does not. One possibility is that a higher concentration of calmidazolium would be more effective. However we chose not to use the dose proven to have maximal effect on molluscan neurons (10 μmol l-1) (Onozuka and Watanabe, 1996) because at higher concentrations calmidazolium may produce non-specific effects (Sugita et al., 1999), rendering results at the higher concentrations questionable. It is also possible that TPep-NLS and dopamine are working through multiple mechanisms, each with a different response to calmidazolium.

In conclusion, our results demonstrate how dopamine and the neuropeptide TPep-NLS control CBF excitation in CPE cells (Fig. 7). Excitatory signals bind receptors that lead directly to a blockage of Ca2+ dependent Cl- currents, currents that also contribute to the resting leak current (Woodward and Willows, 2006). Blockage of an inward Cl- current contributing to Vrest may cause a depolarization of the cell membrane. Fluctuations in membrane potential trigger Ca2+ influx through Ca2+ channels sensitive to nifedipine. A small Ca2+ influx in turn triggers Ca2+ release from RyR-gated stores. The rise of [Ca2+]in activates calmodulin and it's dependent kinases and phosphodiesterases capable of interacting with the cilia motor proteins to increase CBF. We do not rule out multiple CBF excitation pathways. In most cases significant decreases in CBF excitation through pharmacological treatments did not abolish excitation altogether. This may represent either limited efficacy of the pharmacological agents, or alternatively, parallel pathways controlling Ca2+ influx/release and CBF. Further support for multiple pathways includes our observations that excitatory responses to dopamine and TPep-NLS are twofold. There is the easily quantifiable change in CBF, but also a noticeable induced coordination of all the cilia from a single explant. This induced coordination occurs in conjunction with even minor increases in CBF and is not inhibited by the calmodulin inhibitors. Could there be different pathways controlling coordination and CBF increase? We also provide evidence that IP3 may not be a significant player in CBF excitation, but that does not rule out a PLC dependent pathway using PKC, as found in Helisoma embryos (Christopher et al., 1999; Doran et al., 2004), nor the adenylate cyclase/cAMP/PKA pathway found to be necessary in Mytilus (Stommel and Stephens, 1985; Aiello, 1990) (preliminary findings suggest that cAMP plays no role in CPE cells; O. M. Woodwood, unpublished observation), nor other non-calmodulin dependent Ca2+ actions (Salathe and Bookman, 1999). Electrically induced excitation of CPE cells in Tritonia diomedea is consistent with their role as primary locomotory effectors, and like muscle, their RyR-gated stores and voltage-gated Ca2+ channels may be an adaptation permitting parallel CNS control.

List of abbreviations
artificial seawater
ciliary beat frequency
central nervous system
ciliated pedal epithelial cells
endoplasmic reticulum
filtered seawater
inositol 1,4,5-triphosphate
protein kinase A/C
phospholipase C
ryanodine receptor
Tritonia Pedal Peptide


This work was funded by NIH Fellowship 5 F31 NS047922-02 (O.M.W.). We appreciate numerous discussions with Drs R. Wyeth and S. Cain, and thank Drs. W. Moody and M. Bosma for reading earlier versions of the manuscript. A final thanks goes to Tatia Chay Woodward for all other possible types of support.


View Abstract