In this study, we characterized the signalling pathway activated by acetylcholine that encodes Ca2+ oscillations in rat duodenum myocytes. These oscillations were observed in intact myocytes after removal of external Ca2+, in permeabilized cells after abolition of the membrane potential and in the presence of heparin (an inhibitor of inositol 1,4,5-trisphosphate receptors) but were inhibited by ryanodine, indicating that they are dependent on Ca2+ release from intracellular stores through ryanodine receptors. Ca2+ oscillations were selectively inhibited by methoctramine (a M2 muscarinic receptor antagonist). The M2 muscarinic receptor-activated Ca2+ oscillations were inhibited by 8-bromo cyclic adenosine diphosphoribose and inhibitors of adenosine diphosphoribosyl cyclase (ZnCl2 and anti-CD38 antibody). Stimulation of ADP-ribosyl cyclase activity by acetylcholine was evaluated in permeabilized cells by measuring the production of cyclic guanosine diphosphoribose (a fluorescent compound), which resulted from the cyclization of nicotinamide guanine dinucleotide. As duodenum myocytes expressed the three subtypes of ryanodine receptors, an antisense strategy revealed that the ryanodine receptor subtype 2 alone was required to initiate the Ca2+ oscillations induced by acetylcholine and also by cyclic adenosine diphosphoribose and rapamycin (a compound that induced uncoupling between 12/12.6 kDa FK506-binding proteins and ryanodine receptors). Inhibition of cyclic adenosine diphosphoribose-induced Ca2+ oscillations, after rapamycin treatment, confirmed that both compounds interacted with the ryanodine receptor subtype 2. Our findings show for the first time that the M2 muscarinic receptor activation triggered Ca2+ oscillations in duodenum myocytes by activation of the cyclic adenosine diphosphoribose/FK506-binding protein/ryanodine receptor subtype 2 signalling pathway.
Ca2+ signalling is the most `universal and versatile' signal transduction mechanism, mediating the intracellular effects of environmental signals into specific cellular responses (Berridge et al., 2000). The specificity of the Ca2+ signal depends on Ca2+ channel activation by stimulation of G-protein coupled receptors (GPCR). Ryanodine receptors (RYR) form a Ca2+ channel family of the sarcoplasmic reticulum (SR) that support release of stored Ca2+. Three RYR subtypes (RYR1, RYR2 and RYR3) are encoded by three different genes (Fill and Copello, 2002). Four RYR monomers associate to form functional homo- or heterotetrameric channels (Xiao et al., 2002). RYR1 is predominantly expressed in skeletal muscle and is physically coupled to voltage-dependent Ca2+ channel Cav1.1. RYR2 is predominantly expressed in cardiac myocytes and is activated by Ca2+ influx through the Cav1.2 channel via the Ca2+-induced Ca2+ release (CICR) process. RYR3 is ubiquitously expressed and is also activated by CICR (Fill and Copello, 2002). In smooth muscles, two mechanisms of RYR activation by GPCR have been described. Firstly, RYRs are responsible for the CICR mechanism. In this case, the trigger for RYR activation is either a Ca2+ influx (Morel et al., 1996; Arnaudeau et al., 1997) or an inositol 1,4,5-trisphosphate (InsP3)-dependent Ca2+ release (Boittin et al., 1999). When the three RYRs are co-expressed, the release of stored Ca2+ is supported by both RYR1 and RYR2 (Coussin et al., 2000). When only RYR2 and RYR3 are co-expressed, Ca2+ release is reported to depend on RYR2 activation (Ji et al., 2004). Also, RYRs can be activated by cyclic adenosine diphosphate ribose (cADPR) synthesized by ADP-ribosyl cyclase (Kuemmerle and Makhlouf, 1995; Lee, 2004). The receptor for cADPR has not yet been identified but it has been suggested that RYRs, FKBP (12 or 12.6 kDa FK506-binding protein, associated with RYRs), InsP3Rs or Ca2+ pumps could fulfil this role (for a review, see Guse, 2004). The functional interactions of FKBP with RYRs make it an important cADPR target. FKBP12 or FKBP12.6 physically stabilize the coordinated gating of RYRs but reduce the open probability of the channel (Kaftan et al., 1996). Conversely, FKBP removal increases open probability and enhances RYR-dependent Ca2+ signals (McCall et al., 1996). Furthermore, coupling between muscarinic receptors and the cADPR pathway has been proposed to occur in several cell lines (Morita et al., 1997; Higashida et al., 1997).
In smooth muscle, it is generally accepted that acetylcholine (ACh) activates Ca2+ release by M3 muscarinic receptor stimulation and an InsP3-dependent mechanism (Thomas and Ehlert, 1994; Morel et al., 1997), although two studies have suggested that M1 or M2 muscarinic receptors may be involved in Ca2+ signalling (Ge et al., 2003; White et al., 2003). In the present study, we show that ACh induces Ca2+ oscillations by stimulation of RYR2 through a cADPR/M2 muscarinic receptor pathway. Both cADPR and ACh-induced Ca2+ oscillations also require FKBP12.6 interaction with RYR2.
Materials and Methods
The investigation conforms to the European Community and French guiding principles in the care and use of animals. Authorization to perform animal experiments (A-33-063-003) was obtained from the Préfecture de la Gironde (France). Rats (100-120 g) were killed by cervical dislocation. The longitudinal layer of duodenum smooth muscle was cut into several pieces and incubated for 10 minutes in low Ca2+ (40 μM) physiological solution (Hanks' balanced salt solution), and then 0.8 mg/ml collagenase (EC 22.214.171.124), 0.2 mg/ml Pronase E (EC 126.96.36.199), and 1 mg/ml bovine serum albumin were added at 37°C for 20 minutes. After this time, the solution was removed, and pieces of duodenum were incubated again in a fresh enzyme solution at 37°C for 20 minutes. Tissues were placed in an enzyme-free solution and triturated using a fire-polished Pasteur pipette to release cells. Cells were seeded at a density of 103 cells/mm2 on glass slides imprinted with squares for localization of injected cells. Cells were maintained in short-term primary culture in M199 medium containing 5% foetal calf serum, 20 U/ml penicillin and 20 μg/ml streptomycin; they were kept in an incubator in 95% air and 5% CO2 at 37°C. The myocytes were used within 4 to 30 hours or were cultured in this medium for 2-4 days. The physiological solution contained 130 mM NaCl, 5.6 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 11 mM glucose and 10 mM Hepes, pH 7.4 with NaOH. Ca2+-free solution was prepared by omitting CaCl2 and adding 0.5 mM EGTA.
To permeabilize myocytes, physiological solution was replaced by a solution containing 140 mM KCl, 20 mM Hepes, 0.5 mM MgCl2 and 10 μg/ml saponin, pH 7.4 with NaOH. When 100 μg/ml saponin was used, the solution was supplemented with 100 μM ATP. Cell permeabilization was estimated by application of tetramethylrhodamine isothiocyanate (TRITC)-coupled phalloidin. In cells incubated in the presence of 100 and 10 μg/ml saponin, the TRITC fluorescence increased to a maximum value in 1.0±0.5 minute (n=20) and 6.2±1.4 minutes (n=20), respectively, and remained steady in the continuous presence of saponin, whereas no TRITC fluorescence was detected in intact cells for 20 minutes (data not shown). Both caffeine- and ACh-induced Ca2+ responses similar in amplitude and kinetics were obtained in this solution for 15-20 minutes, suggesting that the leakage of fluo-4 out of the cells was not significant during this time period.
Total RNA was extracted from duodenum myocytes using the RNA preparation kit from Epicentre (Madison, WI, USA), following the instructions of the supplier. The RNA concentration was determined at DO260nm with an Eppendorf biophotometer (Eppendorf, Le Pecq, France). The reverse transcription reaction was performed on 50 ng of RNA using Sensiscript-RT kit (Qiagen, Hilden, Germany). PCR conditions were described previously (Morel et al., 2004). PCR were performed with a thermal cycler (Eppendorf). Amplification products were separated by electrophoresis (2% agarose gel) and visualized by ethidium bromide staining. Gels were photographed with EDAS120 and analysed with KDS1D 2.0 software (Kodak Digital Science, Paris, France). Sense (s) and antisense (as) primer pairs specific for RYRl, RYR2 and RYR3 were described previously (Coussin et al., 2000).
Microinjection of oligonucleotides
Phosphorothioate antisense oligonucleotides (denoted with the prefix `as') used in the present study were described previously (Coussin et al., 2000). Briefly, oligonucleotides were designed on the known cloned sequences deposited in the GenBank database with DNAstar (Lasergene software). Oligonucleotides were injected into the nuclei of myocytes by a manual injection system with femtotips II (Eppendorf) as previously described (Macrez-Lepretre et al., 1997). Myocytes were then cultured for 2 to 4 days in culture medium. For physiological experiments, the glass slides were transferred into the perfusion chamber.
Cytosolic Ca2+ measurements
Cells were incubated in physiological solution containing 2 μM fluo-4 acetoxymethylester (AM) for 20 minutes at 37°C. The cells were then washed and allowed to cleave the dye to the active fluo-4 compound for 10 minutes. Images were acquired using the image series or line-scan mode of a confocal Bio-Rad MRC 1024ES (Bio-Rad, Paris, France) connected to a Nikon Diaphot microscope. Excitation light was delivered by a 25 mW argon ion laser (Ion Laser Technology, Salt Lake City, UT, USA) through a Nikon Plan Apo x60, 1.4 NA objective lens. Fluo-4 was excited at 488 nm, and emitted fluorescence was filtered and measured at 540±30 nm. At the setting used to detect fluo-4 fluorescence, the resolution of the microscope was 0.4×0.4×1.5 μm (x-, y- and z-axis). Image series consist of images of the same confocal section of the cell taken at 1.2 second intervals. To analyse variation of fluorescence, regions of interest (ROI), used for each frame of the series, were drawn around each myocyte. The fluorescence value was divided by the fluorescence of the first frame (baseline) and reported as F/F0. Image processing and analysis were performed using Lasersharp 2000 (Bio-Rad) software and IDL software (RSI, Boulder, CO, USA), respectively. ACh, cADPR and caffeine were applied by pressure ejection from a glass pipette for the period indicated in the figures. All experiments were carried out at 26±1°C.
Immunostaining of RYRs
Myocytes were washed with PBS, fixed with 4% (vol/vol) formaldehyde and 0.05% glutaraldehyde for 10 minutes at room temperature, and permeabilized in PBS containing 3% FCS and 1 mg/ml saponin for 20 minutes. Cells were incubated with PBS, saponin (1 mg/ml) and anti-RYR antibodies overnight at 4°C. Then, cells were washed (4× 5 minutes) and incubated with the appropriate secondary Alexa Fluor 488 antibody for 45 minutes at room temperature. After washing in PBS, cells were mounted in Vectashield (AbCys, Paris, France). Images of the stained cells were obtained with a confocal microscope, and fluorescence was estimated by grey level analysis using IDL software (RSI) in 0.5 μm confocal sections. On each cell, fluorescence measurement was acquired from a z-series analysis (20±5 sections) using Lasersharp software (Bio-Rad) and expressed by unit volume. Cells were compared by keeping acquisition parameters (such as grey scale, exposure time, iris aperture, gain and laser power) constant.
ADP-ribosyl cyclase activity measurement
To evaluate the activity of ADP-ribosyl cyclase, we used the ability of the cyclase to catalyse the cyclization of nicotinamide guanine dinucleotide (NGD+) in cyclic GDP ribose (cGDPR), a fluorescent compound. NGD+ was excited at 300 nm and emitted fluorescence was filtered and measured at 420 nm (Graeff et al., 1994). Permeabilized cells were incubated in the presence of 100 μM NGD+ and the emitted fluorescence was collected with a CoolSnap HQ charge-coupled device camera (Roper Scientific, Evry, France). Images were acquired as described by LeBlanc et al. (LeBlanc et al., 2004). The signal was processed by correcting each fluorescence image for background fluorescence. Averaged frames were collected every 0.5 seconds. To analyse variation of fluorescence, the sum of the fluorescence of each myocyte was integrated and this value of fluorescence (F) was divided by the fluorescence of the first image (F0) and reported as F/F0. Measurements were made at 26°C±1°C.
Chemicals and drugs
Fluo-4 acetoxymethylester (Fluo4-AM) was from Teflab (Austin, TX, USA). Alexa Fluor-labelled secondary antibodies were from Molecular Probes (Leiden, The Netherlands). Caffeine was from Merck (Nogent sur Marne, France). Ryanodine was from Calbiochem (Meudon, France). Medium M199, streptomycin, penicillin and collagenase were from Invitrogen (Cergy Pontoise, France). Anti-CD38 antibody was from Santa Cruz Biotechnology (Santa Cruz, CA, USA). 4-diphenylacetoxy-N-(2-chloroethyl)-piperidine hydrochloride (4DAMP) and methoctramine were from RBI (Natik, MA, USA). All primers and phosphorothioate antisense oligonucleotides were synthesized by and purchased from Eurogentec (Seraing, Belgium). cADPR was from Amersham Biosciences (Orsay, France). All other chemicals were from Sigma. The rabbit anti-RYR3-specific antibody was directed against the deduced amino acid sequence, residues 4326-4336 (11 amino acids), of rabbit RYR3 (Jeyakumar et al., 1998). The rabbit anti-RYR2-specific antibody was directed against the deduced amino acid sequence, residues 1344-1365 (22 amino acids), of rabbit RYR2 (Jeyakumar et al., 2001). The rabbit anti-RYR1-specific antibody was directed against the deduced amino acid sequence, residues 4476-4486 (11 amino acids), of rabbit RYR1 (Jeyakumar et al., 2002).
Data are expressed as means ± s.e.m.; n represents the number of tested cells. Significance was tested by means of Student's t-test. P values <0.05 were considered as significant.
Ca2+ oscillations evoked by acetylcholine depend on SR calcium release
In about 60% of rat duodenum myocytes, application of 1 μM ACh for 30 seconds induced a Ca2+ peak followed by regenerative oscillations (Fig. 1A,B). The mean amplitude of the first peak was estimated to be 5.3±0.2 ratio units (n=291) in control conditions (2 mM Ca2+ external solution). The amplitude of oscillations varied strongly but their frequency was reproducible independently of the duration of ACh application (5.2±0.3 oscillations/minute, n=291). In the non-oscillating cells, the Ca2+ responses were composed of a peak followed by a sustained phase during ACh application (not shown). The amplitude of Ca2+ responses (first Ca2+ peak) and the frequency of Ca2+ oscillations both depended on ACh concentration (1 μM ACh induced the maximal response). To test the role of membrane potential and Ca2+ influx in the generation of Ca2+ oscillations, we abolished the plasma membrane potential by permeabilization with 10 μg/ml saponin and studied the effects of external Ca2+ removal for 30 seconds in intact cells (Fig. 1). Under these experimental conditions, the parameters of ACh-induced Ca2+ oscillations (amplitude of the first peak and percentage of oscillating cells) were not significantly affected (Fig. 1B-D) and neither was the frequency of oscillations (intact cells superfused in Ca2+-free solution: 5.2±0.8 oscillations/minute, n=14; permeabilized cells: 4.3±0.9 oscillations/minute, n=16). These results indicate that variations of membrane potential and Ca2+ influx are not required for generation of Ca2+ oscillations, which may depend on intracellular stores.
To characterize the intracellular mechanisms that can drive Ca2+ oscillations evoked by ACh, the effects of different pharmacological substances were tested on the first Ca2+ peak amplitude, frequency of Ca2+ oscillations and percentage of oscillating intact cells. Inhibition of RYR by application of ryanodine (10 μM, 30 minutes before ACh application) (Boittin et al., 1999) induced a drastic inhibition of ACh-induced Ca2+ responses and practically suppressed Ca2+ oscillations (Fig. 1B). Ca2+ oscillations were detected in 64% of tested cells in control conditions and in only 5% of tested cells in presence of ryanodine. Depletion of the Ca2+ store can be modulated by thapsigargin (a SERCA inhibitor). The ability of thapsigargin to inhibit Ca2+ response has been previously evaluated (Morel et al., 2003) and it has been shown that thapsigargin, applied at 50 nM for 1 minute, partially depletes the Ca2+ store, since Ca2+ responses induced by 10 mM caffeine are decreased by approximately 50% (data not shown). Under these conditions, the first peak of ACh-induced Ca2+ responses was significantly decreased (Fig. 1B). The number of cells producing Ca2+ oscillations in response to ACh (8 of 19 cells tested) was not different from that obtained in control conditions. Thapsigargin decreased Ca2+ oscillation frequency but not in a significant manner (control: 6.4±0.3 oscillations/minute, n=19; thapsigargin: 5.2±0.6 oscillations/minute, n=8). However, a second application of ACh or caffeine, 5 minutes later, was ineffective, as expected if the Ca2+ store was completely depleted. In the same way, application of 1 μM thapsigargin for 1 minute, which totally inhibited caffeine-induced Ca2+ response, also inhibited ACh-induced Ca2+ response (data not shown). Application of 1 μM RU360 (an inhibitor of a mitochondria Ca2+ uniporter) did not modify the ACh-induced Ca2+ oscillations (Fig. 1B) indicating that mitochondria are not involved in the generation of Ca2+ oscillations. In permeabilized cells in the presence of 1 mg/ml heparin, an inhibitor of InsP R activation via InsP3, the amplitude of the first Ca2+ 3peak was reduced but the percentage of oscillating cells was not modified (Fig. 1D). Taken together, these results indicate that ACh-induced Ca2+ oscillations depend on Ca2+ release from the SR and that RYRs but not InsP3Rs are involved in these responses.
M2 and M3 muscarinic receptor-dependent transduction pathways
Duodenum myocytes express both M2 and M3 muscarinic receptors (Kuemmerle and Makhlouf, 1995; Morel et al., 1997). Application of 4DAMP, an inhibitor of M3 muscarinic receptors in the nanomolar range (0.1-10 nM), significantly decreased the amplitude of ACh-induced Ca2+ responses but did not modify the percentage of intact cells that showed Ca2+ oscillations (Fig. 1C). Application of methoctramine, a selective inhibitor of the M2 muscarinic receptor in the micromolar range (1-10 μM), decreased the frequency of Ca2+ oscillations before decreasing the amplitude of ACh-induced Ca2+ responses. In the presence of 10 μM methoctramine, only 27% of cells triggered Ca2+ oscillations during ACh application (55% in control conditions, Fig. 1C).
These results prompted us to identify the Ca2+ signalling pathways activated by each muscarinic receptor by using inhibitors of InsP3 and cADPR pathways on permeabilized cells. In the presence of heparin, application of 10 nM 4DAMP was unable to modify the heparin-resistant Ca2+ response whereas 10 μM methoctramine totally abolished it (Fig. 1D). These results suggest that M3 and M2 muscarinic receptor subtypes activate two different pathways and that only the M2 muscarinic receptor-activated pathway may generate Ca2+ oscillations.
Involvement of cADPR in ACh-induced Ca2+ oscillations
ACh has been reported to activate the cADPR pathway by binding to M2 muscarinic receptor (White et al., 2003) and it has been proposed that cADPR may induce Ca2+ release through RYR activation (reviewed by Guse, 2004). A selective and competitive inhibitor of cADPR binding sites, 8Br-cADPR, was applied (20 μM) 5 minutes before application of 1 μM ACh to permeabilized duodenum myocytes (Fig. 2A). In the presence of 8Br-cADPR, the amplitude of the first Ca2+ peak was significantly reduced while Ca2+ oscillations were practically suppressed (Fig. 2B). Application of methoctramine did not modify the 8Br-cADPR-resistant Ca2+ response (n=20) whereas 10 nM 4DAMP (Fig. 2B) and 1 mg/ml heparin (n=8) abolished the response. These results indicate that distinct signalling pathways are activated by ACh: the first one involves the M3 muscarinic receptor subtype and InsP3Rs and the second one involves the M2 muscarinic receptor subtype and cADPR.
To confirm the involvement of cADPR, ADP-ribosyl cyclase activity was inhibited in permeabilized cells using ZnCl2 (de Toledo et al., 2000), high concentrations of NGD+ and an anti-CD38 antibody (Sternfeld et al., 2003). Application of 2 mM ZnCl2 decreased significantly the amplitude of ACh-induced Ca2+ responses and suppressed Ca2+ oscillations (Fig. 2C). Application of 2 mM NGD+ (which competes with NAD+ at the substrate site of the enzyme) or 1 μg/ml anti-CD38 antibody (which inhibits ADP-ribosyl cyclase expressed in membranes) inhibited ACh-induced Ca2+ responses and strongly reduced the number of oscillating cells (Fig. 2C,D). Application of the inactivated anti-CD38 antibody (by heating at 90°C for 20 minutes) did not change the oscillating Ca2+ responses induced by ACh (Fig. 2D). Under these inhibiting conditions, the caffeine-induced Ca2+ responses were not affected, indicating that the Ca2+ store content was not modified (data not shown). These data suggest that cADPR may be the second messenger involved in ACh-induced Ca2+ oscillations.
ADP-ribosyl cyclase activity and cADPR-induced Ca2+ oscillations
As previously reported in biochemical studies on subcellular fractions (Graeff et al., 1994; Sternfeld et al., 2003), we used the cyclization of NGD+ into cGDPR (a fluorescent compound) to evaluate the ADP-ribosyl cyclase activity. In single permeabilized duodenum myocytes, application of 1 μM ACh for 30 seconds induced an increase in fluorescence detected at 420 nm (Fig. 3A) whereas application of a solution without ACh did not modify the fluorescence profile for 5 minutes. Both ADP-ribosyl cyclase inhibitors (ZnCl2 and the anti-CD38 antibody) and methoctramine (10 μM) inhibited the ACh-induced fluorescence signal whereas 4DAMP (1 nM) had no effect (Fig. 3B). Also the application of GTPγS (100 μM) induced an increase in fluorescence similar to that evoked by ACh application.
To confirm the key role of cADPR in ACh-induced Ca2+ oscillations, we studied the Ca2+ responses evoked by cADPR. In permeabilized duodenum myocytes, application of cADPR induced Ca2+ responses in a concentration-dependent manner. At 100 nM cADPR (Fig. 3C,D), the amplitude of Ca2+ responses was similar to that of ACh-induced Ca2+ responses (in the presence of cADPR, 4.1±0.4 ratio units, n=11 and in the presence of ACh, 4.5±0.4 ratio units, n=14) and Ca2+ oscillations were obtained in about 60% of tested cells with a frequency of 6.0±1.0 oscillations/minute (Fig. 3D). The application of cADPR to intact cells superfused in physiological solution (Fig. 3C) or in high-K+, Ca2+-free solution without saponin did not induce any Ca2+ response, indicating that the cADPR receptor was intracellular. In permeabilized cells, the Ca2+ release induced by cADPR was not significantly affected by heparin but was inhibited by 10 μM ryanodine or totally abolished by 20 μM 8Br-cADPR (Fig. 3D). Moreover, caffeine induced a significantly smaller Ca2+ response in the continuous presence of cADPR (3.73±0.31 ratio units, n=26) than in the absence of cADPR (6.90±0.52 ratio units, n=24) suggesting that cADPR induced a Ca2+ release from the caffeine-sensitive Ca2+ store. These data show that ACh stimulates the ADP-ribosyl cyclase activity via the M2 muscarinic receptor, and that cADPR alone may activate Ca2+ oscillations.
Expression of RYR subtypes
To investigate more precisely the function of RYRs in ACh-induced Ca2+ responses, expression of RYR subtypes was examined by RT-PCR and immunostaining using specific antibodies. RT-PCR revealed that RYR1, RYR2 and RYR3 are potentially expressed (Fig. 4A). Immunodetection of RYR subtypes was performed with specific anti-RYR1, anti-RYR2 and anti-RYR3 antibodies. Specificity and absence of cross-reactivity of these antibodies have been previously described (Jeyakumar et al., 1998; Jeyakumar et al., 2001; Jeyakumar et al., 2002). Immunodetection in confocal sections of duodenum myocytes of primary antibody binding sites was revealed with the anti-rabbit Alexa Fluor 488 secondary antibody and the specificity was attested by the use of available antigenic peptides. Non-specific fluorescence (NSF) was determined when specific anti-RYR subtype antibody was pre-incubated with its antigenic peptide 1 hour before application of the immunostaining protocol. When the cell fluorescence obtained with the anti-RYR subtype antibody was higher than NSF (twofold increase), the cell was considered to be immunopositive and specific fluorescence could be estimated. Fig. 4C illustrates typical immunostainings obtained in duodenum myocytes. Measurements of cell fluorescence (Fig. 4B) revealed a specific staining with the anti-RYR1, anti-RYR2 and anti-RYR3 antibodies.
Effects of RYR subtype inhibition on Ca2+ oscillations evoked by ACh, cADPR and rapamycin
As one of the main targets of cADPR is RYRs, we determined which RYR subtypes could be activated by cADPR in duodenum myocytes using an antisense strategy. The effects of the antisense oligonucleotides (as) were maximal 3 days after injection, as previously described (Coussin et al., 2000). We verified that 3 days after injection of asRYR1, asRYR2 or asRYR3 the immunostainings obtained with anti-RYR1-, anti-RYR2- or anti-RYR3-specific antibodies, respectively, were inhibited (Fig. 4D). Injection of asRYR1 did not modify significantly the immunostainings obtained with anti-RYR2 and anti-RYR3 antibodies. In the same way, asRYR2 and asRYR3 did not affect immunostainings of RYR1 and RYR3 or RYR2 and RYR1, respectively (not shown).
The Ca2+ responses evoked by 10 mM caffeine were strongly decreased 3 days after injection of 10 μM asRYR1 plus 10 μM asRYR2 (asRYR1+2; Fig. 5A). In asRYR1-injected cells and in asRYR2-injected cells, the caffeine-induced Ca2+ responses were significantly reduced whereas injection of asRYR3 potentiated the Ca2+ responses (Fig. 5A). In the same cells, ACh was tested 3 minutes after caffeine application. It can be noted that the amplitude of the ACh-induced Ca2+ responses in myocytes cultured for 3 days was similar to that obtained within 4-30 hours of culture (Fig. 1B,C). Injection of asRYR1+2, asRYR1 or asRYR2 strongly decreased the amplitude of ACh-induced Ca2+ responses whereas injection of asRYR3 induced higher Ca2+ responses (Fig. 5B). In contrast, the effects of RYR subtype inhibition on Ca2+ oscillations revealed that oscillations were never observed in asRYR1+2-injected cells whereas the proportion of oscillating cells was slightly affected in asRYR1-injected cells compared to asRYR2-injected cells (44% and 10%, respectively, versus 53% in non-injected cells). In asRYR3-injected cells, the proportion of oscillating cells reached 70% of tested cells. Similarly, the frequency of Ca2+ oscillations was slightly diminished by injection of asRYR1 (4.6±0.8 oscillations/minute), clearly inhibited by injection of asRYR2 (3.0±1.1 oscillations/minute) and increased by injection of asRYR3 (6.2±1.1 oscillations/minute) when compared to non-injected cells (5.2±1.0 oscillations/minute).
The cADPR-induced Ca2+ oscillations were differentially affected by inhibition of RYR subtypes. In asRYR1-injected cells, the amplitude of cADPR-induced Ca2+ responses was not significantly modified (Fig. 5C) and oscillations were obtained in 55% of tested cells. In contrast, the cADPR-induced Ca2+ responses were strongly decreased in both asRYR1+2 and asRYR2-injected cells (Fig. 5C) and Ca2+ oscillations were never observed. As seen with ACh and caffeine application, the cADPR-induced Ca2+ response was increased in asRYR3-injected cells (Fig. 5C) as well as the number of oscillating cells (65% of tested cells).
Several studies have reported that cADPR may modify the interactions between FKBP12/12.6 and RYRs (Noguchi et al., 1997; Tang et al., 2002). Therefore, we investigated the effects of 10 μM rapamycin, which induces uncoupling between FKBP12/12.6 and RYRs in permeabilized cells. The amplitude of rapamycin-induced Ca2+ oscillations was similar to that obtained with cADPR (Fig. 6A) but the number of oscillating cells was smaller (41% of tested cells). As observed with cADPR, both amplitude of rapamycin-induced Ca2+ responses and number of oscillating cells were not modified in asRYR1-injected cells, strongly decreased in both asRYR2- and asRYR1+2-injected cells and increased in asRYR3-injected cells (Fig. 6B,C). In the continuous presence of rapamycin (10 μM for 10 minutes), caffeine induced a reduced Ca2+ response whereas cADPR was unable to induce any response (Fig. 3D), suggesting that rapamycin and cADPR might act on the same protein, i.e. FKBP12/12.6.
Taken together, these results indicate that RYR2 is required for initiating Ca2+ oscillations in response to ACh, cADPR or rapamycin and suggest that decoupling of FKBP12.6 and RYR2 could be the molecular mechanism of these Ca2+ oscillations.
The results of the present study provide evidence for a pivotal role of RYR2 and cADPR in oscillating Ca2+ responses evoked by ACh in duodenum myocytes. The Ca2+ oscillation signalling pathway involves ACh binding to M2 muscarinic receptors, cADPR production and activation of RYR2 after uncoupling with FKBP12.6. RYR1 probably participates in the ACh-induced Ca2+ response but is not required for the generation of Ca2+ oscillations.
In different cell types, it has been proposed that Ca2+ oscillations can be supported by different Ca2+ channels or Ca2+ stores: InsP3Rs (Harootunian et al., 1991; Miyakawa et al., 1999; Morel et al., 2003) or RYRs (Morel et al., 1996) and mitochondria (Rizzuto et al., 2000). Activation of Ca2+ influx is another trigger for Ca2+ oscillations via membrane potential oscillations (Sneyd et al., 2004). Our results showed that in duodenum myocytes, Ca2+ influx was not needed to generate ACh-induced Ca2+ responses or to sustain Ca2+ oscillations since these responses were obtained in intact cells superfused in Ca2+-free EGTA-containing solution and in saponin-permeabilized cells. By pharmacological inhibition of a mitochondria Ca2+ uniporter, we also excluded the mitochondria as a Ca2+ buffer for these Ca2+ oscillations. Therefore, our results support the idea that the ACh-induced Ca2+ oscillations depend on a Ca2+ release from the sarcoplasmic reticulum. The distinct roles of RYRs and InsP3Rs in Ca2+ responses were illustrated by using heparin (an inhibitor of InsP3Rs), which decreased the first Ca2+ peak of ACh-induced responses without causing changes in Ca2+ oscillation generation and frequency, and ryanodine, which decreased both Ca2+ response amplitude, and generation and frequency of Ca2+ oscillations. RT-PCR and immunostainings indicate that InsP3R1 is the only InsP3R subtype expressed in rat duodenum myocytes (J.-L.M., unpublished data). In addition, expression of InsP3R1 alone is not sufficient to induce Ca2+ oscillations, as previously reported (Miyakawa et al., 1999; Morel et al., 2003). An important finding in this paper is that Ca2+ oscillations are totally dependent on RYR2 activation. Although involvement of RYRs has been proposed in Ca2+ oscillations, the RYR subtype responsible for Ca2+ oscillations has not been previously identified. In vascular myocytes, the function of RYR2 and RYR1 subtypes in Ca2+ release has been studied by using antisense oligonucleotides targeting the RYR subtypes (Coussin et al., 2000). With the same antisense strategy, we showed that, in duodenum myocytes, RYR2 was the trigger for ACh-induced Ca2+ oscillations whereas RYR1 participated in the Ca2+ response amplitude probably through a CICR mechanism.
Although activation of M2 muscarinic receptors is classically known to inhibit adenylyl cyclase activity (Peralta et al., 1988) or to modify membrane potential by inhibiting Ca2+-activated K+ channel (Kotlikoff et al., 1992), it has been recently proposed that M2 muscarinic receptors may induce Ca2+ signals by activation of the cADPR pathway (White et al., 2003) or by stimulation of a voltage-dependent Ca2+ channel (Cav1.2b) via the phosphatidylinositol 3-kinase/PKC pathway (Callaghan et al., 2004). Activation of the cADPR pathway by ACh in duodenum myocytes is shown by: (1) inhibition of Ca2+ oscillations by application of the cADPR competitive antagonist (8Br-cADPR), (2) inhibition of ACh-induced Ca2+ oscillations by inhibitors of ADP-ribosyl cyclase (ZnCl2, anti-CD38 antibody) and (3) detection of ADP-ribosyl cyclase activity by fluorescence experiments as the enzyme cyclizes NGD+ (non fluorescent) to produce cGDPR, a fluorescent compound (Graeff et al., 1994). This method has been used successfully in microsomes and cellular homogenates from bovine chromaffin cells (Morita et al., 1997), rat vascular myocytes (de Toledo et al., 2000) and human myometrium (Chini et al., 2002). We were able to detect ADP-ribosyl cyclase activity in single permeabilized cells under the same conditions that we used for Ca2+ measurements. ACh induced a fluorescence signal corresponding to an increase of ADP-ribosyl cyclase activity. With this method, we showed that ACh activated the cyclase by binding to the M2 muscarinic receptor and that this stimulatory effect was inhibited by ZnCl2 and the anti-CD38 antibody. These results are in agreement with data reporting an ADP-ribosyl cyclase activity in rat duodenum (de Toledo et al., 2000), and its activation by ACh in adrenal chromaffin cells (Morita et al., 1997).
RYRs, FKBP12/12.6, Ca2+ pumps (SERCA and plasma membrane Ca2+ pumps) and InsP3Rs have been reported to be the main cADPR targets in different cells types (Guse, 2004) and modulation by cADPR of Ca2+ oscillations has been shown in porcine tracheal muscle (Prakash et al., 1998). In duodenum myocytes, we showed that (1) cADPR induced Ca2+ oscillations similar to those evoked by ACh; (2) in the presence of cADPR, the amplitude of caffeine-induced Ca2+ responses was significantly decreased, suggesting that cADPR preferentially induced Ca2+ release; (3) cADPR-induced Ca2+ oscillations were not modified after inhibition of InsP3Rs by heparin, indicating no participation of InsP3Rs; (4) cADPR induced Ca2+ oscillations in permeabilized cells that were similar to those obtained in intact cells, excluding a role for the plasma membrane Ca2+ pumps; (5) partial inhibition of SERCA by thapsigargin did not affect the Ca2+ oscillation mechanism, suggesting that SERCA activation was not the trigger for Ca2+ oscillations. This is in contrast to previous data suggesting that cADPR may modulate SERCA activity. For example, cADPR has been reported to increase the rate of decline of Ca2+ responses following membrane depolarizations in colonic myocytes (Bradley et al., 2003) and both the frequency of Ca2+ sparks and amplitude of caffeine-induced Ca2+ responses in cardiomyocytes (Lukyanenko et al., 2001). In both cases, cADPR was unable to induce a Ca2+ signal by itself. Moreover, the cADPR effects on Ca2+ pumps are obtained with higher cADPR concentrations (1-10 μM) than those used to activate RYR-dependent Ca2+ release (1-100 nM), as reported by Guse (Guse, 2004). We also showed that (6) cADPR-induced Ca2+ responses were inhibited by ryanodine and the anti-RYR2 antisense oligonucleotide, suggesting that RYR2 could be the target of cADPR although the cADPR receptor has not been identified so far.
In pancreatic cells, various agonists stimulate cADPR production (Sternfeld et al., 2003) and cADPR may bind to FKBP12.6, a RYR-associated protein (Noguchi et al., 1997). cADPR and FK506, an activator of FKBP12/12.6, have similar effects, which are not additive (Ozawa, 2004). In arterial myocytes, activation of purified RYRs by cADPR has been shown to be inhibited by FK506 (Tang et al., 2002) and cADPR is unable to activate Ca2+ release in tracheal myocytes from FKBP12.6 knockout mice (Wang et al., 2004). In duodenum myocytes, we found that rapamycin (an inhibitor of FKBP12/12.6/RYR interactions, analogue to FK506) inhibited cADPR-induced Ca2+ oscillations, suggesting that the effects of cADPR on Ca2+ release involved FKBP. In smooth muscle, FKBP12.6 is associated with RYR2 but not with other RYR isoforms or InsP3Rs (Wang et al., 2004) and modulating interactions between FKBP12.6 and RYR2 have been reported (Tang et al., 2002). Using an antisense strategy, we showed that in duodenum myocytes, only the anti-RYR2 antisense oligonucleotide inhibited the rapamycin-induced Ca2+ oscillations, supporting the idea that the FKBP12 involved in the cADPR pathway is coupled to RYR2. In addition, in the continuous presence of rapamycin, cADPR was ineffective suggesting that both compounds interacted at the same protein level. Consequently, we can propose that in duodenum myocytes cADPR induces activation of RYR2 to produce Ca2+ oscillations through an uncoupling between RYR2 and FKBP12.6. RYR3 has also been proposed as a possible target for cADPR in T-lymphocytes and Jurkat cells (Kunerth et al., 2004). However, this possibility can be discarded in duodenum myocytes, since inhibition of RYR3 resulted in a stimulation of Ca2+ release in response to ACh, cADPR and rapamycin. An increase in Ca2+ spark frequency has been previously reported in vascular myocytes of RYR3 knockout mice (Löhn et al., 2001). The spliced variants described by Jiang et al. (Jiang et al., 2003) may offer an explanation for this observation since one of these spliced variant may heteromerize with the other RYR3 variants or RYR2 and inhibit their function. Therefore, in our experiments, inhibition of all the RYR3 variants may account for an increase in Ca2+ signals dependent on RYR2.
In conclusion, this study proposes for the first time a selective signalling pathway responsible for ACh-induced Ca2+ oscillations in rat duodenum myocytes. This pathway involves the M2 muscarinic receptor, resulting in activation of ADP-ribosyl cyclase and subsequent production of cADPR. Interactions between cADPR and FKBP12.6 may lead to activation of RYR2 and the secondary recruitment of RYR1 via a CICR mechanism. This signalling pathway may be important for the duodenum peristalsis, which is under vagal parasympathetic control.
This work was supported by grants from Centre National de la Recherche Scientifique and Centre National des Etudes Spatiales, France. N.F. is supported by a grant from Région Aquitaine. The authors thank C. LeBlanc for measurements of ADP-ribosyl cyclase activity and N. Biendon and J. L. Lavie for assistance.