Calcium feedback mechanisms regulate oscillatory activity of a TRP-like Ca²⁺ conductance in C. elegans intestinal cells

Abstract

Inositol-1,4,5-trisphosphate (IP₃)-dependent Ca²⁺ oscillations in Caenorhabditis elegans intestinal epithelial cells regulate the nematode defecation cycle. The role of plasma membrane ion channels in intestinal cell oscillatory Ca²⁺ signalling is unknown. We have shown previously that cultured intestinal cells express a Ca²⁺-selective conductance, I_ORCa, that is biophysically similar to TRPM7 currents. I_ORCa activates slowly and stabilizes when cells are patch clamped with pipette solutions containing 10 mm BAPTA and free Ca²⁺ concentrations of ∼17 nm. However, when BAPTA concentration is lowered to 1 mm, I_ORCa oscillates. Oscillations in channel activity induced simultaneous oscillations in cytoplasmic Ca²⁺ levels. Removal of extracellular Ca²⁺ inhibited I_ORCa oscillations, whereas readdition of Ca²⁺ to the bath caused a rapid and transient reactivation of the current. Experimental manoeuvres that elevated intracellular Ca²⁺ blocked current oscillations. Elevation of intracellular Ca²⁺ in the presence of 10 mm BAPTA to block I_ORCa oscillations led to a dose-dependent increase in the rate of current activation. At intracellular Ca²⁺ concentrations of 250 nm, current activation was transient. Patch pipette solutions buffered with 1–4 mm of either BAPTA or EGTA gave rise to similar patterns of I_ORCa oscillations. We conclude that changes in Ca²⁺ concentration close to the intracellular opening of the channel pore regulate channel activity. Low concentrations of Ca²⁺ activate the channel. As Ca²⁺ enters and accumulates near the pore mouth, channel activity is inhibited. Oscillating plasma membrane Ca²⁺ entry may play a role in generating intracellular Ca²⁺ oscillations that regulate the C. elegans defecation rhythm.

Fluctuating intracellular Ca²⁺ levels control numerous, diverse cellular processes including fertilization, differentiation, gene expression, exocytosis, secretion, cell motility and contraction, cell proliferation and programmed cell death (Berridge et al. 2000). Calcium changes can be categorized broadly as occurring globally or within localized regions of the cell, and they can occur either as single transients or in a repetitive, oscillatory fashion. Oscillating Ca²⁺ signals contain frequency, amplitude and spatial information that direct cells to carry out specific tasks (Berridge, 1997; De Koninck & Schulman, 1998; Li et al. 1998).

Although studied extensively, the mechanisms leading to the generation of intracellular Ca²⁺ oscillations are incompletely understood. The nematode Caenorhabditis elegans provides numerous experimental advantages for the functional analysis of genes and genetic pathways involved in biological processes such as Ca²⁺ signalling (Barr, 2003; Strange, 2003). C. elegans has a short lifespan, is genetically tractable and has a fully sequenced and well-annotated genome. It is also relatively easy and economical to manipulate gene expression in C. elegans by transgenesis and RNA interference.

Calcium signalling is essential for regulating the C. elegans defecation cycle, an ultradian rhythm with a periodicity of 45–50 s. The defecation cycle consists of the sequential contraction of the posterior body wall muscles, anterior body wall muscles and enteric muscles every 45–50 s when the animals are feeding (Iwasaki & Thomas, 1997). Posterior body wall muscle contraction is regulated by inositol-1,4,5-trisphosphate (IP₃)-dependent Ca²⁺ oscillations in intestinal epithelial cells (Dal Santo et al. 1999).

We are using a combination of patch clamp electrophysiology, behavioural assays, intracellular Ca²⁺ imaging and forward and reverse genetic screening in order to define the molecular mechanisms of oscillatory Ca²⁺ signalling in C. elegans intestinal epithelial cells. Using the whole-cell patch clamp technique, we demonstrated recently that cultured C. elegans intestinal epithelial cells express two highly Ca²⁺-selective conductances (Estevez et al. 2003). One of these, I_ORCa (outwardly rectifying calcium channel current), displays strong outward rectification, inhibition by intracellular Mg²⁺, and insensitivity to intracellular Ca²⁺ store depletion. I_ORCa resembles the Mg²⁺-inhibited current (MIC), Mg²⁺-nucleotide regulated metal ion current (MagNUM), and TRPM7 currents studied in mammalian cells (Nadler et al. 2001; Hermosura et al. 2002; Prakriya & Lewis, 2002; Kozak et al. 2002).

We report here that I_ORCa oscillates when cultured C. elegans intestinal cells are patch clamped with pipette solutions containing low concentrations of BAPTA or EGTA. Oscillating channel activity gives rise to oscillating intracellular Ca²⁺ levels. Our results suggest that Ca²⁺ concentration changes close to the intracellular opening of the channel pore regulate outwardly rectifying calcium (ORCa) channel activity. Calcium entry initially activates ORCa channels. As Ca²⁺ accumulates at the pore mouth, channel activity is inhibited. Oscillating plasma membrane Ca²⁺ entry may play a role in the generation of intracellular Ca²⁺ oscillations that drive the C. elegans defecation rhythm.

Methods

C. elegans strains

Wild-type N2 and JR1838 (wIs84) worm strains were maintained at 20–25°C utilizing standard methods (Brenner, 1974). JR1838 expresses green fluorescent protein (GFP) driven by an intestinal-specific promoter (elt-2::GFP) and was used to identify intestinal cells in culture (Estevez et al. 2003).

C. elegans embryonic cell culture

Embryonic cells were prepared by treating synchronized adult nematodes with an alkaline hypochlorite solution (0.5 m NaOH and 1% NaOCl) for 5 min (Lewis & Fleming, 1995). Eggs released by this treatment were pelleted by centrifugation and then washed three times with egg buffer containing 118 mm NaCl, 48 mm KCl, 2 mm CaCl₂, 2 mm MgCl₂ and 25 mm Hepes (pH 7.3, 345 mOsm) (Edgar, 1995). Carcasses were separated from washed eggs by density centrifugation in 30% sucrose. The egg layer was removed by pipette and washed once with egg buffer and then pelleted. Eggshells were removed by resuspending pelleted eggs in egg buffer containing 1–2.5 U ml⁻¹ of chitinase for 45–90 min at room temperature. After digestion of the eggshell, the suspension was gently pipetted up and down several times to dissociate the cells. Cells were washed twice with L-15 cell culture medium (Life Technologies, Grand Island, NY, USA) containing 10% FBS (Hyclone, Logan, UT), 50 U ml⁻¹ penicillin and 50 μg ml⁻¹ streptomycin and adjusted to 345 mmol kg⁻¹ H₂O with sucrose.

Dissociated embryo cells were filtered through a sterile 5 μm Durapore syringe filter (Millipore Corporation, Bedford, MA, USA) to remove undissociated embryos and newly hatched larvae. Filtered cells were plated on 12 mm diameter glass cover slips coated with 0.5 mg ml⁻¹ peanut lectin agglutinin. Cultures were maintained at 24°C in a humidified incubator in L-15 cell culture medium.

Patch-clamp recordings

Coverslips with cultured embryo cells were placed in the bottom of a bath chamber (model R-26G; Warner Instrument Corp., Hamden, CT, USA) that was mounted onto the stage of a Nikon TE300 inverted microscope. Cells were visualized by fluorescence and video-enhanced differential interference contrast (DIC) microscopy.

Patch electrodes were pulled from soft glass capillary tubes (PG10165-4, World Precision Instruments, Sarasota, FL, USA) that had been silanized with dimethyl-dichloro silane. The standard pipette solution for whole-cell recordings from intestinal cells contained: 147 mm NaGluconate, 0.05 mm CaCl₂, 1 mm MgCl₂, 1 mm BAPTA, 10 mm Hepes, 2 mm Na₂ATP, 0.5 mm Na₂GTP, pH 7.2 (adjusted with CsOH), 325 mmol kg⁻¹ H₂O (adjusted with sucrose). The standard bath solution contained (mm): 145 NaCl, 1 CaCl₂, 5 MgCl₂, 10 Hepes, 20 glucose, pH 7.2 (adjusted with NaOH), 340–345 mmol kg⁻¹ H₂O (adjusted with sucrose). Free Mg²⁺ and Ca²⁺ levels in the various solutions used were calculated using MaxChelator software WINMAXC v.2.4 (http://www.stanford.edu/~cpatton/maxc.html). Nominally Ca²⁺-free bath solution contained no added Ca²⁺ and 1 mm EGTA. Nominally Mg²⁺-free bath solution contained no added Mg²⁺ and 1 mm EDTA with free Ca²⁺ levels maintained at 1 mm by addition of extra Ca²⁺.

Experiments were also performed using bath and pipette solutions with more physiologically relevant ion concentrations. Physiological bath solution contained (mm): 145 NaCl, 5 KCl, 1 CaCl₂, 5 MgCl₂, 10 Hepes, 20 glucose, pH 7.2 (adjusted with NaOH), 340–345 mmol kg⁻¹ H₂O (adjusted with sucrose). The physiological pipette solution contained (mm): 125 KGluconate, 18 KCl, 4 NaCl, 50 nm CaCl₂, 2.5 MgATP, 10 Hepes, pH 7.2 (adjusted with CsOH), 325 mmol kg⁻¹ H₂O (adjusted with sucrose). Calculated free Mg²⁺ and free Ca²⁺ concentrations were 471 μm and 16 nm, respectively.

Whole-cell currents were recorded using an Axopatch 200B (Axon Instruments, Union City, CA, USA) patch clamp amplifier. Command voltage generation, digitization, and data analysis were carried out on a 1.6 GHz Pentium computer (Dimension 4400; Dell Computer Corp.) using a Digidata 1322A AD/DA interface with pClamp 8.2 and Clampfit 8.2 software (Axon Instruments). Electrical connections to the amplifier were made using Ag/AgCl wires and 3 m KCl/agar bridges. Patch-clamp electrodes had DC resistances of 4–7 MΩ. Cells were used only if the series resistance was no more than 2.5-fold greater than the pipette resistance and the reversal potential was within ±5 mV of the mean value of +26 mV.

Unless otherwise noted, currents were elicited by ramping membrane voltage from −80 to +80 mV for 1 s every 5 s. The holding potential was set at 0 mV. Currents were filtered at 5 kHz and digitized at 20 kHz. It should be noted that the voltage ramp frequency used in these studies was, in some recordings, too slow relative to oscillatory current changes to permit accurate estimate of current amplitude and rates of increase and decrease. Thus, the values reported in this paper may be slight underestimates of actual values. These underestimates do not change the conclusions of this study.

During recording of whole-cell current oscillations, we occasionally observed changes in current rectification and a hyperpolarizing shift in reversal potential to values more negative than −40 mV (see Results). The appearance of these changes was unpredictable and occurred during either the upswing or downswing of the oscillation. In addition, the changes were always transient lasting <5–15 s. Because of the hyperpolarizing shift in reversal potential (E_rev), which was not consistent with the activity of ORCa channels, we eliminated the traces from our recordings and data analysis. Doing so changed none of the conclusions put forward in this paper.

In some experiments, intracellular Ca²⁺ levels were monitored simultaneously during patch-clamp recordings. For these studies, cells cultured from wild-type N2 worms were patch clamped with the standard pipette solution containing 30 μm fluo-4 (pentapotassium salt). Images were captured with a high resolution, cooled Micromax CCD-1300 camera (Princeton Instruments, Tucson, AZ, USA). Data acquisition and analysis were preformed with MetaFluor software (Universal Imaging Corporation, Downingtown, PA, USA).

Chemicals

BAPTA, thapsigargin and fluo-4 were purchased from Molecular Probes (Eugene, OR, USA) and IP₃ was purchased from Calbiochem (La Jolla, CA, USA). All other chemicals were obtained from Sigma (St Louis, MO, USA). Thapsigargin and fluo-4 were dissolved in DMSO as concentrated stocks and added to pipette solutions at final DMSO concentrations of 0.1% and 0.6%, respectively.

Statistical analyses

Data are presented as means ±s.e.m. Unless otherwise indicated, statistical significance was determined using a Mann–Whitney test where P-values of < 0.05 were taken to indicate statistical significance.

Results

ORCa channel activity oscillates when intestinal cells are patch clamped with solutions containing low concentrations of Ca²⁺ buffers

As shown previously (Estevez et al. 2003), an outwardly rectifying Ca²⁺ current, I_ORCa, activates slowly after whole-cell access is achieved when intestinal cells are patch clamped with a pipette solution buffered with 10 mm BAPTA (∼17 nm free Ca²⁺; Fig. 1A) or 10 mm EGTA (Estevez et al. 2003). Full current activation typically occurs after 3–4 min and the current remains stable or runs down slightly over recording periods of up to 13 min. Mean ±s.e.m. peak whole-cell current in cells buffered with 10 mm BAPTA (measured at +80 mV) was 80 ± 14 pA pF⁻¹ (n = 10).

Figure 1. Whole-cell current activation in cells patch clamped with 1 mm or 10 mm BAPTA.

Whole-cell currents were elicited by ramping membrane voltage from −80 to +80 mV for 1 s every 5 s. Holding potential was 0 mV. A, activation of whole-cell current in a cell patch clamped with 10 mm BAPTA (17 nm free Ca²⁺). Time 0 is when whole-cell access is achieved. B, current activation in a cell patch clamped with 1 mm BAPTA (11 nm free Ca²⁺). C, current–voltage (I–V) relationship of whole-cell current at the time point denoted with * in B. D, inhibition of whole-cell current oscillations by La³⁺; 100 μm La³⁺ was added to the bath for the time period indicated by the black bars. Oscillations are immediately re-initiated when La³⁺ is washed out.

When intestinal cells were patch clamped with a pipette solution buffered with 1 mm BAPTA (∼11 nm free Ca²⁺), stable whole-cell currents were not observed. Instead, whole-cell current underwent transient and repetitive activation and inactivation (Fig. 1B). Current oscillations were observed in 123 of 132 cells. The mean ± s.e.m. oscillation frequency was 0.52 ± 0.03 oscillations min⁻¹ (n = 123) during a mean recording period of 8.3 ± 0.5 min. Non-oscillating cells showed two patterns of current activity. In six cells, current levels were initially high and slowly declined over time. The mean ±s.e.m. whole-cell current (measured at +80 mV) at the beginning of the recording was 94 ± 32 pA pF⁻¹ and declined to 31 ± 9 pA pF⁻¹ after 4 min. The other three cells initially expressed low levels of current that slowly increased over time. Mean ±s.e.m. initial whole-cell current was 15 ± 6 pA pF⁻¹ and increased to 160 ± 60 pA pF⁻¹ over a 4 min recording period.

The timing of current oscillations was variable from cell to cell. Time intervals between peak current amplitudes ranged from 15 to 380 s. Oscillations in most cells exhibited no obvious rhythmicity. We calculated the mean coefficient of variance (CV), which is the standard deviation expressed as percentage of the mean, for the times between peak current amplitude in 84 cells. In the majority of cells (n = 72) oscillations were arrhythmic, displaying a mean ±s.e.m. CV of 46 ± 3%. However, a small percentage of cells (n = 12) displayed rhythmic oscillatory activity with CVs ranging from 3% to 9%.

We characterized the properties of the oscillating current to determine if it was due to the activity of the ORCa channel. No K⁺ was present in the solutions used in these studies ruling out the contribution of K⁺ currents to the oscillations. A small outwardly directed Na⁺ gradient (cell Na⁺= 152 mm; bath Na⁺= 145 mm) exists, which would give rise to an E_rev of −1 mV if oscillations were due to the activity of a non-selective cation channel. The mean ±s.e.m.E_rev of the whole-cell current during oscillations was +26 ± 1 mV (n = 32). The mean E_rev for ORCa channels described by us previously under identical ionic conditions was +21 mV (Estevez et al. 2003).

Occasionally during the recording of current oscillations, we saw changes in current rectification and a hyperpolarizing shift in reversal potential to voltages more negative than −40 mV. These changes were always transient, lasted <5–15 s, and occurred during either the upswing or downswing of the oscillation. The predicted Cl⁻E_rev is −110 mV, suggesting that the transient current might due to activation of a Ca²⁺-dependent anion channel. Because the current was transient and its appearance unpredictable, detailed characterization was not possible. Nevertheless, we examined the effect of bath Cl⁻ removal on current oscillations. In four cells recorded from under control conditions for 60–70 s, we detected 2 oscillations cell⁻¹ (mean ±s.e.m. oscillation frequency = 1.7 ± 0.2 oscillations min⁻¹; n = 4). Current oscillations continued in all four cells when bath Cl⁻ was replaced with gluconate. A total of 2–10 current oscillations were observed over 165–300 s recording periods in Cl⁻-free bath solution. The mean ±s.e.m. oscillation frequency in the absence of extracellular Cl⁻ was 1.3 ± 0.3 oscillations min⁻¹ (n = 4), which was not significantly (P > 0.6) different from that observed under control conditions.

I_ORCa shows strong outward rectification, is inhibited by high intracellular Mg²⁺ concentration and is reversibly blocked by extracellular La³⁺ (Estevez et al. 2003). As shown in Fig. 1C, the oscillating current is strongly outwardly rectifying. No oscillations were detected in 9 out of 10 cells patch clamped with a pipette solution containing 6 mm free Mg²⁺ (data not shown). One high-Mg²⁺ cell showed a single oscillation over a 12.5 min recording period. Mean ±s.e.m. baseline current in high Mg²⁺ cells was 7 ± 1 pA pF⁻¹. Application of 100 μm La³⁺ rapidly and completely inhibited oscillations in all cells (n = 17; Fig. 1D). The effect of La³⁺ was reversible and oscillations were initiated immediately after washout bath in three of five cells tested (Fig. 1D). Based on these results, we conclude that intestinal cell current oscillations are due to repetitive activation and inactivation of ORCa channels.

The ORCa channel has biophysical properties that resemble those of TRPM7 (Estevez et al. 2003), which is permeable to extracellular Mg²⁺ (Nadler et al. 2001; Kozak et al. 2002; Monteilh-Zoller et al. 2002). Removal of extracellular Mg²⁺ (0 mm Mg²⁺ and 1 mm EDTA) had no effect on current oscillations. Four out of four cells tested showed current oscillations. The mean ±s.e.m. oscillation frequency was 0.54 ± 0.08 oscillations min⁻¹ (n = 4), which was not significantly (P > 0.9) different from that observed in control cells. These results demonstrate that Mg²⁺ influx and elevation of intracellular Mg²⁺ levels is not responsible for oscillatory channel activity.

Three types of I_ORCa oscillations were observed (Fig. 2). We characterized these oscillations by quantifying current activation and inactivation times. Activation time is defined as the time to peak current after the current begins to rise above baseline. Inactivation time is the time it takes for the peak current to decay back to baseline. Approximately 38% of oscillations exhibited slowly activating and rapidly inactivating currents (Fig. 2A). Mean ±s.e.m. activation and inactivation times were 104 ± 6 s and 22 ± 1 s (n = 161). Both current activation and inactivation were rapid for ∼50% of the oscillations and occurred over mean ±s.e.m. time periods of 25 ± 1 s and 22 ± 1 s (n = 214; Fig. 2B). The remaining oscillations exhibited rapidly activating and slowly inactivating currents (Fig. 2C). The mean ±s.e.m. times for current activation and inactivation were 20 ± 3 s and 91 ± 12 s (n = 47). Multiple types of oscillations were frequently observed in a single cell.

Figure 2. Characteristics of I_ORCa oscillations.

Three types of current oscillations are observed: A, slowly activating and rapidly inactivating oscillations; B, rapidly activating and inactivating oscillations and C, rapidly activating and slowly inactivating oscillations.

ORCa channel oscillations occur in the presence of physiologically relevant intracellular ion concentrations and changes in membrane potential

The data shown in Figs 1 and 2 were obtained by ramping membrane potential from −80 mV to +80 mV at 160 mV s⁻¹ every 5 s. While experimentally convenient, this voltage-clamp protocol is not physiologically relevant. It was therefore important to determine if oscillations in ORCa activity could be observed during membrane potential changes more similar to those observed in vivo.

Using microelectrode methods, Wei and Salkoff (A. D. Wei & L. Salkoff, Washington University School of Medicine) have shown that the basolateral membrane potential of the C. elegans intestine oscillates rhythmically from ∼−10 mV to ∼−40 mV with a period similar to that of the defecation cycle. This oscillating membrane potential was mimicked by voltage clamping intestinal cells at a holding potential of −10 mV for 4 s and then ramping voltage to −40 mV and back to −10 mV over a 47 s time period. As shown in Fig. 3A, inward current oscillations were observed under these voltage-clamp conditions. Peak currents were recorded at voltages of −23 to −37 mV. Current activation occurred during both the hyperpolarizing and depolarizing phases of the voltage ramp. These observations suggest that oscillations of ORCa activity may occur in intact cells during physiologically relevant fluctuations of membrane potential.

Figure 3. Current oscillations in the presence of physiologically relevant ion concentrations and changes in membrane potential.

A, effect of physiological membrane voltage oscillations on whole-cell current. Membrane potential was ramped from −10 mV to −40 mV and back to −10 mV over a 47 s time interval (as indicated in the top panel) to mimic normally occurring membrane potential fluctuations in intestinal cells (see Results). Current traces from four representative cells are shown. B, current oscillations are observed in cells patch clamped with pipette and bath solutions containing more physiologically relevant ion concentrations (see Methods).

We also examined the effect on current oscillations of patch pipette and bath solutions containing ion concentrations more similar to those expected in vivo. Figure 3B shows that current oscillations were observed when cells were patch clamped with solutions containing more physiologically relevant ion concentrations (see Methods). Oscillations were detected in 8 of 12 cells patch clamped under these conditions. The currents elicited with these solutions had lower amplitudes compared to those observed with our standard bath and pipette solutions (e.g. compare Figs 3B and 1B). This lower amplitude is probably due to the higher Mg²⁺ concentration of the intracellular solution. As we have shown previously, intracellular Mg²⁺ inhibits ORCa activity with an estimated K_½-value for free Mg²⁺ of 692 μm (Estevez et al. 2003).

I_ORCa oscillations are Ca²⁺ dependent

In order to exhibit oscillatory activity, ORCa channels must be regulated by mechanisms that function to both activate and inhibit the channel. Many types of Ca²⁺ channels are regulated by intracellular Ca²⁺ levels (Hardie & Minke, 1994; Schuhmann et al. 1997; Zimmer et al. 2000; Budde et al. 2002; Launay et al. 2002; Hofmann et al. 2003; Prawitt et al. 2003). Given these findings, we tested the hypothesis that intracellular Ca²⁺ regulates ORCa activity via both positive and negative feedback mechanisms.

Removal of extracellular Ca²⁺ completely inhibited current oscillations in nine of nine cells. When Ca²⁺ was added back to the bath, oscillations were reinitiated in eight of eight cells (Fig. 4A). Whereas current reactivation in one cell did not occur until 45 s after the switch back to Ca²⁺-containing bath, in the remaining seven cells current reactivation was rapid. Complete reactivation was observed within 6 ± 1 s. A similar pattern of rapid, transient current reactivation was observed when La³⁺ block was removed (discussed above; see Fig. 1D).

Figure 4. Effect of extracellular Ca²⁺ on whole-cell current oscillations.

A, removal of extracellular Ca²⁺ blocks I_ORCa oscillations. Note that I_ORCa is reactivated immediately upon return to the control bath solution. B, reactivation of I_ORCa upon return to the control Ca²⁺-containing bath solution is specifically dependent on Ca²⁺. I_ORCa is not reactivated upon switching from a Ca²⁺-free bath to a Ca²⁺-free bath containing 1 mm Ba²⁺. Note that readdition of Ca²⁺ causes an immediate reactivation of I_ORCa. Voltage clamp protocol was the same as described in Fig. 1 legend.

The rapid reactivation of current oscillations upon readdition of Ca²⁺ to the bath or removal of extracellular La³⁺ suggests that Ca²⁺ entry through ORCa channels initially stimulates channel activity. To further test this idea, we analysed the ability of extracellular Ba²⁺ to activate I_ORCa. Barium permeates many types of Ca²⁺ channels, but is unable to support Ca²⁺-dependent regulatory processes (Brehm & Eckert, 1978; Zweifach & Lewis, 1995; Lee et al. 2003). We first determined the permeability of Ba²⁺ through ORCa channels by measuring current reversal potentials during replacement of bath Na⁺ and Ca²⁺ with 130 mm NMDG⁺ and 10 mm Ba²⁺. Relative permeability with respect to Na⁺ was calculated using the Goldman–Hodgkin–Katz equation as previously described (Estevez et al. 2003). Barium permeated ORCa channels more readily than Na⁺, displaying a mean ±s.e.m. relative permeability (P_Ba/P_Na) of 14 ± 1 (n = 6). As shown in Fig. 4B, introduction of Ba²⁺ after exposure to Ca²⁺-free bath failed to stimulate I_ORCa oscillations. In eight of nine cells, oscillations resumed immediately upon switching back to a normal Ca²⁺-containing bath.

In addition to complete Ca²⁺ removal, we also monitored the effect of variations in extracellular Ca²⁺ concentration on current oscillations. The frequency of current oscillations was unaffected by raising bath Ca²⁺ concentration from 1 mm to 10 mm. Mean ±s.e.m. oscillation frequency was 0.51 ± 0.05 oscillations min⁻¹ (n = 29 cells) in 1 mm Ca²⁺ and 0.42 ± 0.15 (n = 7 cells) in 10 mm Ca²⁺. However, increasing bath Ca²⁺ concentration to 10 mm significantly (P < 0.0001) reduced peak current amplitude from 384 ± 26 pA pF⁻¹ (n = 79 oscillations) to 174 ± 13 pA pF⁻¹ (n = 19 oscillations). Reduced current amplitude in the presence of 10 mm Ca²⁺ may be due to increased Ca²⁺ entry with subsequent feedback inhibition of channel activity (discussed below). Alternatively, the effect of high extracellular Ca²⁺ could be due to interaction with an external Ca²⁺ inhibitory site on the channel.

As shown in Fig. 1A, the ORCa channel activates slowly after membrane rupture in cells dialysed with a 10 mm BAPTA-buffered pipette solution. The mechanism responsible for constitutive current activation is unknown, but could involve changes in intracellular Mg²⁺ concentration and/or the concentration of other regulatory factors during patch pipette dialysis (Nadler et al. 2001; Runnels et al. 2002; Estevez et al. 2003; Takezawa et al. 2004). We monitored the rate of current activation in cells patch clamped with solutions containing 17–500 nm free Ca²⁺ and 10 mm BAPTA. Increasing intracellular Ca²⁺ concentration increased the rate of current activation in a dose-dependent fashion (Fig. 5A). Low concentrations of Ca²⁺ resulted in stable current activation (Fig. 5B). However, in the presence of 250 nm Ca²⁺, current activation was not stable. Instead, the current reached a mean ±s.e.m. peak level of 234 ± 30 pA pF⁻¹ (n = 6) and then declined steadily (Fig. 5B). The mean ±s.e.m. inactivation observed was 77 ± 12% (n = 6). At higher intracellular Ca²⁺ concentrations, seals were unstable and could not be held for more than ∼30 s.

Figure 5. Effect of intracellular Ca²⁺ concentration on rate of whole-cell current activation and whole-cell current amplitude.

Cells were patch clamped with a pipette solution containing 10 mm BAPTA and free Ca²⁺ concentrations of 17 nm, 100 nm, 250 nm, 500 nm. A, increasing intracellular Ca²⁺ concentration increases the rate of I_ORCa activation in a concentration-dependent manner. Initial rates of current activation were quantified by performing linear regression analysis on whole-cell currents measured during the first 30–130 s after obtaining whole-cell access. Values are means ±s.e.m. (n = 5–10). B, high levels of intracellular Ca²⁺ cause transient current activation. Under control conditions (17 nm free Ca²⁺), whole cell current activated relatively slowly and peak current amplitude remained stable. When cells were patch clamped with 250 nm free Ca²⁺ in the pipette, I_ORCa activates more rapidly. However, current levels did not remain stable and instead declined slowly over time.

The transient activation of current observed in cells patch clamped with 250 nm Ca²⁺ and 10 mm BAPTA suggests that elevated Ca²⁺ levels inhibit ORCa activity. To examine this possibility more directly, we patch clamped cells with pipette solutions buffered with 1 mm BAPTA and containing either 11 nm or 500 nm free Ca²⁺. With 11 nm free Ca²⁺ in the pipette, an average of 4 ± 0.5 current oscillations were observed in six of six cells over recording periods of 13–17 min. In cells dialysed with 500 nm Ca²⁺ and recorded from for 8–17 min, five of ten cells showed no oscillations (Fig. 6). The other five cells showed 1–2 current oscillations that were initiated within the first 5–120 s after obtaining whole-cell access. Oscillations observed during the first 2 min of whole-cell recording may reflect the time required to fully dialyse the cells with 500 nm Ca²⁺.

Figure 6. Effect of intracellular Ca²⁺ on whole-cell current oscillations.

A, I_ORCa oscillations are blocked in cells patch clamped with a pipette solution containing 1 mm BAPTA and 500 nm free Ca²⁺.

Oscillations in ORCa activity induce oscillating intracellular Ca²⁺ levels

As discussed in the Introduction, oscillating intracellular Ca²⁺ levels in the C. elegans intestinal epithelium are thought to regulate contraction of posterior body wall muscles that drive defecation. Oscillating Ca²⁺ entry could conceivably play a role in generating cytoplasmic Ca²⁺ oscillations. We therefore tested whether I_ORCa oscillations are associated with concomitant increases in intracellular Ca²⁺. Cells were patch clamped with a pipette solution containing 30 μm fluo-4 and intracellular Ca²⁺ levels and current activity were monitored simultaneously. In nine of nine cells, I_ORCa and intracellular Ca²⁺ levels oscillated concomitantly (Fig. 7A).

Figure 7. Oscillations in ORCa activity induce oscillating intracellular Ca²⁺ levels.

A, I_ORCa and Ca²⁺ oscillations occur concomitantly. Fluorescent images and whole-cell currents were recorded every 5 s. Fluo-4 fluorescence intensity changes are plotted as the ratio of F/F₀ where F₀ is the fluorescence intensity measured at time 0. B, I_ORCa oscillations are blocked by prolonged membrane hyperpolarization. Whole-cell current oscillations were detected by ramping membrane potential as described in Fig. 1 legend. After 2–4 current oscillations were observed, cells were held at −80 mV for 2–10 min (black bar). Prolonged hyperpolarization elevates intracellular Ca²⁺ levels and blocks current oscillations. C, depletion of intracellular Ca²⁽ stores by inclusion of 10 μm IP₃ and 1 μm thapsigargin in the patch pipette does not inhibit I_ORCa or intracellular Ca²⁺ oscillations. D, intracellular Ca²⁺ oscillations are inhibited by application of 100 μm La³⁺ (black bar) to the bath. Voltage clamp protocol was the same as described in Fig. 1 legend.

The simultaneous recordings of intracellular Ca²⁺ and current activity allowed us to further test the hypothesis that high intracellular Ca²⁺ levels inhibit I_ORCa oscillations. Intracellular Ca²⁺ levels should be increased by prolonged membrane hyperpolarization. We patch clamped cells and monitored ORCa activity by ramping membrane potential from −80 mV to +80 mV at 160 mV s⁻¹ every 5 s. After observing several current oscillations over a 2–8 min recording period, membrane potential was clamped at −80 mV for 2–10 min. As shown in Fig. 7B, prolonged hyperpolarization elevated intracellular Ca²⁺. In five of six cells, oscillations were inhibited completely by clamping membrane potential at −80 mV. When the ramping protocol was reinitiated, oscillations immediately resumed (Fig. 7B). In one cell, two current oscillations were observed during a 10 min recording period at −80 mV. Taken together, results shown in Figs 5B, 6 and 7B suggest that high intracellular Ca²⁺ concentration inhibits ORCa activity.

It is possible that intracellular Ca²⁺ changes we observed during ORCa oscillations are due to oscillating release of Ca²⁺ from intracellular stores, and that this in turn triggered current oscillations. To test this possibility, we depleted intracellular Ca²⁺ stores by patch clamping cells with a pipette solution containing 10 μm IP₃ and 1 μm thapsigargin (Estevez et al. 2003). As shown in Fig. 7C, IP₃ and thapsigargin caused a transient elevation of intracellular Ca²⁺, which is presumably due to store release. Depletion of intracellular stores, however, had no effect on current oscillations (Fig. 7C). Twenty-two of 24 cells examined exhibited normal current oscillations. The mean ±s.e.m. oscillation frequency in store-depleted cells was 0.42 ± 0.1 oscillations min⁻¹ (n = 22) and was not significantly different (P > 0.5) from that observed in control cells patch clamped with 0.1% DMSO (mean ±s.e.m. oscillation frequency = 0.47 ± 0.1 oscillations min⁻¹; n = 9).

The store depletion experiments indicate that Ca²⁺ oscillations in patch-clamped cells are not due to repetitive store release but instead result from plasma membrane entry, most likely via ORCa channels. To test this possibility directly, we inhibited ORCa by addition of 100 μm La³⁺ to the extracellular bath. As shown in Fig. 7D, La³⁺ inhibited both current and intracellular Ca²⁺ oscillations. Similar results were observed in four of four cells tested. Taken together, the results shown in Fig. 7 suggest that ORCa-mediated Ca²⁺ influx is sufficient to change intracellular Ca²⁺ concentration, and that oscillatory channel activity may contribute to oscillating intracellular Ca²⁺ levels.

ORCa activity is regulated by local changes in intracellular Ca²⁺

An important question is whether ORCa oscillations are induced by global or local changes in Ca²⁺ concentration. We addressed this by analysing the effects of BAPTA and EGTA on whole-cell current activity. BAPTA is a ‘fast’ buffer with more rapid Ca²⁺-binding kinetics than EGTA (Rintoul & Baimbridge, 2003), and is more effective at preventing Ca²⁺ concentration changes close to the intracellular pore opening of Ca²⁺ channels (Naraghi & Neher, 1997; Bauer, 2001). Therefore, differences in the effects of BAPTA and EGTA on a Ca²⁺-mediated regulatory process are indicative of the dependence of that process on local versus more global changes in Ca²⁺ levels (Gutnick et al. 1989; Adler et al. 1991; Hoth & Penner, 1993; Zweifach & Lewis, 1995; Fierro & Parekh, 1999).

When patch clamped with a pipette solution containing 1 mm EGTA, seven of ten cells showed current oscillations (Fig. 8). The mean ±s.e.m. oscillation frequency was 0.48 ± 0.06 oscillations min⁻¹ and was not statistically different from that observed with 1 mm BAPTA (P > 0.6). The kinetics of current activation and inactivation were also similar to those observed with BAPTA; 43% of the oscillations displayed slow activation and rapid inactivation (e.g. Fig. 2A), 47% displayed equally rapid current activation and inactivation (e.g. Figure 2B), and 11% displayed rapidly activating currents which inactivated slowly (e.g. Figure 2C). Using the Chi-square goodness of fit test, these proportions were not significantly different from those observed with 1 mm BAPTA (χ²= 0.418, 2 d.f.; P > 0.05).

Figure 8. Whole-cell current oscillations in a cell patch clamped with a pipette solution containing 1 mm EGTA and 11 nm free Ca²⁺.

Voltage clamp protocol was the same as described in Fig. 1 legend.

We also examined the effects of BAPTA and EGTA concentrations of 2–4 mm on current oscillations. Although increasing the concentration of buffer generally decreased the percentage of oscillating cells, no obvious differences between BAPTA- and EGTA-buffered cells were observed. In the presence of 4 mm BAPTA or 4 mm EGTA, 5 of 11 and 3 of 11 cells, respectively, exhibited current oscillations. The mean ±s.e.m. oscillation frequency was 0.26 ± 0.1 oscillations min⁻¹ for 4 mm BAPTA, and 0.97 ± 0.6 oscillations min⁻¹ for 4 mm EGTA. These values were not significantly different (P > 0.25). The similarity in the effect of BAPTA and EGTA on current oscillations suggests that regulatory changes in Ca²⁺ concentration are occurring in a region of the cell where both buffers are relatively ineffective. This is presumably very close to the intracellular mouth of the channel pore.

Discussion

We have demonstrated here that the TRP-like Ca²⁺ channel ORCa undergoes oscillatory changes in activity when cultured C. elegans intestinal cells are patch clamped with solutions containing low concentrations of Ca²⁺ buffers. Oscillations in channel activity occur in the presence of physiologically relevant intracellular ion concentrations and changes in membrane potential (Fig. 3), and lead to concomitant fluctuations in cytoplasmic Ca²⁺ levels (Fig. 7). Oscillatory channel activity is induced by changes in intracellular Ca²⁺ levels (Figs 4, 5, 6). We postulate that Ca²⁺ entry through the ORCa channel initially activates the channel and that channel activity is inhibited by intracellular Ca²⁺ accumulation.

Intracellular Ca²⁺ exerts both positive and negative regulatory effects on many cation channels, including voltage-gated and store-operated Ca²⁺ channels, and IP₃ receptor (IP₃R) channels (Iino, 1990; Zuhlke et al. 1999; Lee et al. 2000; Swatton & Taylor, 2002; Mak et al. 2003; Moreau et al. 2005). Several TRP cation channels are also dually regulated by intracellular Ca²⁺ (Hardie & Minke, 1994; Nilius et al. 2002; Prawitt et al. 2003; Strotmann et al. 2003; Agam et al. 2004; Lambers et al. 2004; Shi et al. 2004). For example, the rate and extent of TRPV4 activation are increased by intracellular Ca²⁺ entry in a calmodulin-dependent manner, while elevated Ca²⁺ levels function to feedback inhibit the channel (Strotmann et al. 2003). TRPM5 is activated by Ca²⁺ concentrations of 0.3–1 μm and inhibited at higher Ca²⁺ levels (Prawitt et al. 2003). Interestingly, TRPM5 activity is also sensitive to the rate of change of intracellular Ca²⁺ (Prawitt et al. 2003). TRPC6 and TRPC7 display biphasic regulation by Ca²⁺ (Shi et al. 2004). Calcium-dependent activation of TRPC6 and inactivation of TRPC7 are mediated by calmodulin (Shi et al. 2004).

Regulation of ORCa activity by intracellular Ca²⁺ could be mediated by direct binding of Ca²⁺ to the channel and/or associated proteins, or by Ca²⁺-dependent activation of signalling components such as kinases and phosphatases. The precise mechanism by which Ca²⁺ both activates and inhibits the channel leading to regenerative current oscillations is unknown. However, possible mechanisms are suggested by studies on other channel types. For example, Ca²⁺ influx through P/Q-type and L-type Ca²⁺ channels enhances channel inactivation and induces facilitation via a calmodulin-dependent regulatory process (Zuhlke et al. 1999, 2000; Lee et al. 2000; DeMaria et al. 2001; Lee et al. 2003). Calmodulin consists of two Ca²⁺-binding lobes with different Ca²⁺ affinities (Ye et al. 2005). Binding of Ca²⁺ to a relatively high-affinity site on the carboxy terminus of calmodulin selectively induces P/Q-type channel facilitation, whereas binding to a lower affinity site on the amino terminus regulates channel inactivation (DeMaria et al. 2001). The differential blocking effects of BAPTA and EGTA on the two processes suggest that facilitation is regulated by rapid, local increases in Ca²⁺. In contrast, inactivation requires longer lasting and more global Ca²⁺ elevation (Lee et al. 2000).

The IP₃R channel also exhibits dual regulation by cytoplasmic Ca²⁺ (Iino, 1990; Bezprozvanny et al. 1991; Finch et al. 1991; Thrower et al. 2001). Low concentrations of Ca²⁺ activate the channel, whereas higher Ca²⁺ levels inhibit channel activity. Inhibitory and activating Ca²⁺-binding sites on the channel protein are thought to mediate Ca²⁺-dependent regulation (Swatton & Taylor, 2002; Mak et al. 2003). Modelling studies suggest that dual regulation by Ca²⁺ is sufficient to produce oscillating channel activity and oscillating intracellular Ca²⁺ levels (De Young & Keizer, 1992; Schuster et al. 2002).

It is conceivable that mechanisms similar to those described for P/Q channels and IP₃Rs regulate ORCa activity. For example, small amounts of basal ORCa activity could lead to local increases in Ca²⁺ concentration with subsequent Ca²⁺ binding to high-affinity regulatory sites on the channel and/or associated regulatory proteins. Channel activation triggered by binding to these sites would lead to large increases in both local and global Ca²⁺ concentration (e.g. Fig. 7). As local Ca²⁺ rises, binding to low-affinity regulatory sites could induce channel inhibition. Upon channel inactivation, Ca²⁺-buffering mechanisms and Ca²⁺ uptake into intracellular stores and extrusion across the plasma membrane are expected to reduce both local and global Ca²⁺ levels. Following a refractory period that could involve slow release of Ca²⁺ from low-affinity inhibitory sites, protein conformational changes and/or signal transduction mechanisms, the channel would be available for reactivation by Ca²⁺ binding to the putative high-affinity regulatory sites. While this scheme is clearly speculative, it represents a plausible hypothesis and starting point for additional studies.

While activation of ORCa channels leads to global changes in intracellular Ca²⁺ (Fig. 7), it is likely that the regulatory effects of Ca²⁺ occur very close to the intracellular side of the channel pore. We were unable to detect significant differences in channel oscillatory activity when cells were dialysed with BAPTA versus EGTA at buffer concentrations up to 4 mm. Differences in the effects of buffering with BAPTA versus EGTA on the behaviour of Ca²⁺-dependent physiological processes can be indicative of whether the regulatory effects of Ca²⁺ are occurring locally or globally (Gutnick et al. 1989; Adler et al. 1991; Hoth & Penner, 1993; Zweifach & Lewis, 1995; Borst & Sakmann, 1996; Fierro & Parekh, 1999). Mathematical modelling of Ca²⁺ concentration profiles in the membrane area surrounding a point source of Ca²⁺ entry suggests that BAPTA buffers effectively at distances >30 nm from the entry point, whereas EGTA is effective only at distances >100 nm (Gutnick et al. 1989; Naraghi & Neher, 1997; Bauer, 2001). Because BAPTA and EGTA at concentrations up to 4 mm support similar current oscillations, we suggest that the regulatory effects of Ca²⁺ are occurring <30 nm from the intracellular pore opening where neither buffer is capable of effectively preventing changes in Ca²⁺ levels.

The regulatory effects of Ca²⁺ on channel activity have important functional implications. For example, Ca²⁺-dependent inactivation of voltage-gated Ca²⁺ channels prevents excess Ca²⁺ entry and cell injury during channel opening (Chad, 1989). Positive and negative regulatory effects of intracellular Ca²⁺ on the IP₃ receptor (Iino, 1990; Bezprozvanny et al. 1991) play key roles in generating cytosolic Ca²⁺ oscillations and waves (Berridge, 1993; Kaftan et al. 1997). Calcium-dependent regulation of TRP channel activity probably regulates membrane voltage, Ca²⁺ influx, intracellular Ca²⁺ levels and Ca²⁺ oscillations (Launay et al. 2002, 2004; Hofmann et al. 2003; Prawitt et al. 2003).

IP₃-dependent Ca²⁺ oscillations in intestinal epithelial cells control the C. elegans defecation cycle (Dal Santo et al. 1999). The precise role played by intestinal cell plasma membrane Ca²⁺ channels in generating these oscillations remains unknown. However, plasma membrane Ca²⁺ entry is indispensable for maintaining oscillatory Ca²⁺ signalling in many different cell types (Scheenen et al. 1996; Kukuljan et al. 1997; Luo et al. 2001; Torihashi et al. 2002; Wu et al. 2002). If ORCa activity oscillates in the intact intestinal epithelium, fluctuating Ca²⁺ entry could contribute to the overall magnitude and duration of oscillatory intracellular Ca²⁺ changes. Calcium entry through ORCa channels may also function to activate IP₃ receptor Ca²⁺ channels and trigger intracellular store Ca²⁺ release. Experiments are currently underway to test these possibilities and to identify the genes encoding ORCa channels.