Cyclic ADP-ribose increases Ca²⁺ removal in smooth muscle

Ca²⁺ release from the sarcoplasmic reticulum (SR) via the ryanodine receptor (RyR) regulates both general cellular activities such as metabolism and more specific events such as smooth muscle contraction. RyR opening is itself promoted by Ca²⁺ (Endo et al., 1970; Fabiato, 1983) and decreased by FK506 binding proteins (FKBPs) (Brillantes et al., 1994; Xin et al., 2002). Among other substances, proposed to modulate RyR activity and release Ca²⁺ from the SR, cADPR has received considerable attention. Evidence supports such a regulatory role for cADPR in several cell types including sea urchin eggs, macrophages, neurones, myocytes and pancreatic β-cells (Galione and Sethi, 1996; Petersen and Cancela, 1999; Lee, 2001). The wide distribution of enzymes for its synthesis (ADP-ribosyl cyclase) and degradation (cADPR hydrolase) has been cited in support of this activity (Lee, 2001). Further evidence is the observation that cADPR might also act as a second messenger in the response to acetylcholine (ACh), cholecystokinin and oestrogen (Prakash et al., 1998; Petersen and Cancela, 1999; Fukushi et al., 2001; Leite et al., 2002). For example, in smooth muscle, cADPR increased and a cADPR inhibitor (8-amino-cADPR) blocked the amplitude and frequency of ACh-induced Ca²⁺ oscillations (Prakash et al., 1998). cADPR also released Ca²⁺ from the SR in both renal and coronary artery cells (Kannan et al., 1996; Li et al., 2000) and longitudinal intestinal smooth muscle (Kuemmerle and Makhlouf, 1995) and increased RyR activity in lipid bilayers (Li et al., 2001) though extreme conditions e.g. low ATP concentrations and high [Ca²⁺] might be necessary (Sitsapesan et al., 1994). Together, these results suggest that cADPR might play an important role in the regulation of RyR activity to influence Ca²⁺ signalling.

The mechanism of RyR activation by cADPR is, however, disputed. In contrast to the above support for its direct activation of the receptor, other evidence suggests that cADPR might activate RyR indirectly via intermediaries such as FKBPs (FKBP12.6) (Noguchi et al., 1997) or the SR Ca²⁺ pump (SERCA) (Lukyanenko et al., 2001). FKBPs by themselves might be important in Ca²⁺ signalling by regulating release channels. In the absence of FKBP12.6 (which suppresses RyR activity), cardiac excitation-contraction coupling was facilitated and displayed an increased Ca²⁺-induced Ca²⁺ release (CICR) (Xin et al., 2002). However, the role of FKBPs themselves in Ca²⁺ signalling is controversial. The ability of FKBPs (12 kDa or 12.6 kDa) to bind to those RyR isoforms (RyR2, RyR3) present in smooth muscle was evident in some (Bultynck et al., 2001a,b) but not in all (Carmody et al., 2001) studies. Removal of FKBPs from RyR2 might have either no effect (Timerman et al., 1996) or promote channel activity (Marx et al., 2000; Tang et al., 2002); removal of the binding proteins from RyR3 has either no effect (Copello et al., 1999) or prolongs channel opening (Bielefeldt et al., 1997). Given the uncertainty surrounding the ability of FKBPs themselves to regulate RyR, it is not clear whether or not cADPR modulates RyR activity by binding to the proposed intermediary (FKBP12.6).

Another proposed intermediary is SERCA. cADPR stimulates SERCA activity to raise the SR Ca²⁺ content and to open RyR by lumenal Ca²⁺ regulation of the channel (Lukyanenko et al., 2001). Evidence for this was the observation that, although cADPR and caffeine each increased Ca²⁺ spark frequency in cardiac myocytes, the onset of cADPR's effect was significantly later (>2 minutes) than that of caffeine. This delay was attributed to the time required for the SR Ca²⁺ content to be increased following SERCA activation by cADPR (Lukyanenko et al., 2001). This proposal could explain the frequently observed pronounced delay in the onset of cADPR activity (Cui et al., 1999; Aarhus et al., 1995; Li et al., 2000).

However, despite the evidence that supports a role for cADPR in Ca²⁺ release by modulating RyR activity either directly or indirectly, controversy persists regarding receptor involvement. This dispute has been sustained by the apparent failure of cADPR itself to contract tracheal and intestinal smooth muscle (Iizuka et al., 1998; Kuemmerle and Makhlouf, 1995), to modify the contraction induced by carbachol in tracheal smooth muscle (Iizuka et al., 1998), or indeed to activate RyR at all (Copello et al., 2001; Lukyanenko et al., 2001). Our interest in the mode of action of cADPR, in the light of this controversy, has led us to re-examine its effects and mechanisms on smooth muscle, in order to assess its role in Ca²⁺ signalling. We chose a cell type with RyR known to contribute to its physiological responses – colonic myocytes (McCarron et al., 2002). The investigation was particularly concerned to establish whether, in smooth muscle, cADPR regulated the release of Ca²⁺ via RyR and, if so, whether receptor involvement took place directly or indirectly via intermediaries such as FKBPs. To avoid changes in Ca²⁺influx and therefore in [Ca²⁺]_c brought about by changes in membrane potential (e.g. Petersen and Cancela, 1999), single voltage-clamped myocytes were used. Photolysis of caged cADPR was used to increase cADPR concentrations rapidly and directly.

This study has shown that cADPR failed to increase either bulk average or subsarcolemma [Ca²⁺] in smooth muscle, nor did it induce CICR. By contrast, the RyR activator caffeine altered both bulk average and subsarcolemma [Ca²⁺] and induced CICR. FKBP12.6, a proposed receptor for cAPDR (Noguchi et al., 1997), was absent from the myocytes. The introduction of recombinant FKBP12.6 as substitute reduced SR Ca²⁺ release but, in its presence, cADPR remained ineffective. In contrast to the absence of any direct or indirect effects on RyR, cADPR accelerated and a cADPR antagonist slowed the rate of Ca²⁺ removal from the cytoplasm. Thus, in an intact cell with RyR, cADPR might not affect Ca²⁺ release directly but might modulate [Ca²⁺]_c via Ca²⁺ removal.

Caged cADPR, fluo-3 AM ester and fura-2-conjugated dextran were purchased from Molecular Probes (Cambridge Bioscience, Cambridge, UK); caged inositol (1,4,5) triphosphate [Ins(1,4,5)P₃] trisodium salt, cADPR, thapsigargin and fluo-3 penta-ammonium salt from Calbiochem-Novabiochem (Nottingham, UK). FK506 was gifted by Fujisawa (München, Germany). Coomassie Plus Protein Assay Reagent was purchased from Perbio (Cheshire, UK). All other reagents were from Sigma (Poole, UK).

Single myocytes were enzymatically dissociated (McCarron and Muir, 1999) from colonic muscle from male guinea pigs (550-750 g), humanely killed by cervical dislocation and exsanguination in accordance with the guidelines of the UK Animal (Scientific Procedures) Act (1986).

The extracellular solution contained 80 mM sodium glutamate, 40 mM NaCl, 20 mM tetraethylammonium chloride (TEA), 1.1 mM MgCl₂, 3 mM CaCl₂, 10 mM Hepes and 30 mM glucose (pH 7.4 with NaOH). The Na⁺-free bathing solution was the same with LiCl replacing Na⁺ salts (pH 7.4 with LiOH). The pipette solution contained 85 mM Cs₂SO₄, 20 mM CsCl, 1 mM MgCl₂, 30 mM Hepes, 3 mM MgATP, 2.5 mM pyruvic acid, 2.5 mM malic acid, 1 mM NaH₂PO₄, 5 mM creatine phosphate, 0.5 mM guanosine phosphate, 0.1 mM fluo-3 penta-ammonium salt and either 0.025 mM caged inositol trisphosphate (InsP₃) trisodium salt or 0.05 mM or 0.5 mM caged cyclic adenosine 5′-diphosphate ribose (cADPR). Where spontaneous transient outward currents (STOCs) were measured, the extracellular solution contained 60 mM NaCl and 4.7 mM KCl; TEA was omitted. In the pipette solution, Cs₂SO₄ and CsCl were replaced by 105 mM KCl. In controls, to determine whether FKBP12.6 could enter the cell via the patch pipette, caged compounds and fluo-3 were omitted and 0.035 mM fura-2-conjugated dextran potassium salt (10 kDa; a fluorescence indicator of comparable weight to that of FKBP12.6) included to estimate the time course of entry of FKBP12.6. Cell fluorescence from the fura-2-conjugated dextran following excitation at the fura-2 isosbestic point (360 nm; the [Ca²⁺]-independent wavelength) increased to 50% of the final value in 673±175 seconds. These results establish the time course by which the high molecular weight compounds (such as recombinant FKBP12.6) could enter the cell via the patch pipette.

Whole cell currents were amplified by an Axopatch 1D (Axon Instruments, Union City, CA), low pass filtered at 500 Hz (eight-pole Bessel filter; Frequency Devices, Haverhill, MA) and digitally sampled at 1.5 kHz using a Digidata interface, pCLAMP software (version 6.0.1, Axon Instruments) and Axotape (Axon Instruments) and stored on a PC for analysis.

[Ca²⁺]_c was measured using either a microfluorimeter system (McCarron et al., 1999; Flynn et al., 2001) or a wide-field digital imaging system. Briefly, single cells were illuminated at 488 nm (bandpass 9 nm) from a monochromator and imaged through a Nikon 40×, 1.3 numerical aperture, oil-immersion objective. Ca²⁺ images were captured using an intensified, cooled, frame-transfer CCD camera (Pentamax Gen IV) in virtual chip mode with WinView32 (Roper Scientific, Marlow, Buckinghamshire, UK). Images (160×160 pixels), with a pixel size of 563×563 nm at the cell, were acquired at 100 frames second^–1. Imaging and electrophysiology data were synchronized on pClamp by recording a transistor-transistor logic (TTL) output from the CCD camera, reporting the camera's readout status, with the electrophysiology data. In imaging experiments, cells were incubated with 1 μM fluo-3 AM ester and 10 μM wortmannin (to minimize contraction) for 30 minutes (20±2°C) and fluo-3 omitted from the pipette solution.

Fluorescence signals were expressed as ratios (F/F₀ or ΔF/F₀) of fluorescence counts (F) relative to baseline (control) values (taken as 1) before stimulation (F₀). Where the fluo-3 signals were converted into [Ca²⁺]_c (nM), the K_d for fluo-3 was taken as 390 nM, F_min was assumed to be 0 and F_max was determined from in vivo calibrations (Cannell et al., 1994). Image data analyses were performed using the program Metamorph (Roper Scientific, Marlow, Buckinghamshire, UK). Experiments on the colonic cells were carried out at either room temperature (20±2°C) or at 35±2°C, as indicated in the text; those on sea urchin eggs at 16°C.

A rise in [Ca²⁺]_c was obtained (1) by depolarizing the membrane from a holding potential of –70 mV to either 0 mV or +10 mV, (2) using 10 mM caffeine [applied by pressure ejection for 3 seconds using a PV820 Pneumatic Picopump (WPI, Hertfordshire, UK)] or (3) by the photorelease of caged InsP₃. Submaximal responses to maximal concentrations of caffeine (10 mM) were obtained by increasing the distance of the ejection pipette from the cell. Quantitative differences between maximal and submaximal responses to caffeine were confirmed in each experiment. In the case of STOCs the membrane potential was held over the range –30 mV to 0 mV.

To photolyse caged InsP₃ or caged cADPR the output of a xenon flash lamp (Rapp Optoelektronik, Hamburg, Germany) was passed through a UG-5 filter to select ultraviolet light and merged into the excitation light path of the microscope using a quartz bifurcated fibre optic bundle (McCarron and Muir, 1999). The concentration of the caged, non-photolysed compound given in brackets in the text refers to that in the pipette rather than to that released in the cytosol by photolysis.

Owing to their variability in amplitude, duration and frequency, STOC activity was summarized by analysing the area under the curve (AUC) of a randomly selected 5-second or 100-second period of STOC discharge under control and test conditions (expressed as picocoulombs, pC). For calculation of the probability of a STOC occurring (P_STOC), the threshold for STOC activity was set at two and a half times the baseline noise value.

The increase in [Ca²⁺]_c, as indicated by fluo-3, is termed the `measured' Ca<sup>2+</sup>. The `calculated' [Ca²⁺]_c was determined from Ca²⁺ current amplitude (I_Ca) measurement using the following equation: ∫ – I_Cadt/2FV, where ∫ – I_Cadt is total charge entry, F the Faraday constant and V the cell volume. I_Ca was integrated by measuring the area under the curve with reference to the current level obtained after complete block of I_Ca with 1 mM cadmium chloride. The cell volume was 2 picolitres (Bradley et al., 2002).

The rate of Ca²⁺ decline (d[Ca²⁺]_c/dt) was measured as a function of [Ca²⁺]_c. Owing to noise inherent in fluorescence [Ca²⁺]_c measurements, the decline was smoothed by fitting polynomials (fourth to seventh order, selected by the highest r² value, where r² is the coefficient of determination and represents the strength of linearity in a given relationship. It is usually taken as a measure of goodness of fit) to the data obtained. The derivative was obtained by averaging the slopes of two adjacent data points.

Freshly isolated guinea pig brain and colon in 3 ml buffer [0.9% NaCl, 0.3 M sucrose, pH 7.2, protease inhibitor cocktail (Roche Diagnostics, East Sussex, UK)] per gram of tissue were homogenized using an Ultraturrax three times for 15 seconds each at 4°C. The homogenate's protein content was assayed (using Coomassie Plus Protein Assay Reagent) and preparations resuspended at 4 mg ml^–1 total protein. In some experiments, homogenates were treated with either 20 μM FK506 or 50 μM cADPR for 30 minutes at 37°C, followed by centrifugation at 54,000 g for 15 minutes at 4°C. Pellets were resuspended at 2.5 mg ml^–1 in 4× Laemmli buffer (NuPage LDS sample buffer; Invitrogen, Grönigen, The Netherlands) with 5 mM dithiothreitol and assayed for FKBP12 by immunoblotting.

SDS-PAGE and immunoblotting were performed (Currie and Smith, 1999) using the NuPAGE system (Invitrogen, Grönigen, The Netherlands) with 12% Bis-Tris gels and MES running buffer for FKBP12 detection, using recombinant FKBP12 and/or FKBP12.6 as controls. Blots were incubated with either rabbit anti-FKBP12/12.6 antiserum (SA168) or goat polyclonal anti-FKBP12 (c-19; Santa Cruz Biotechnology, CA), each diluted 1:500. Membranes were then washed and incubated with either goat anti-rabbit IgG coupled to horseradish peroxidase (HRP) or donkey anti-goat coupled to HRP, respectively, both at 1:10,000 dilution. Blots were developed using the ECL detection system (Amersham, Buckinghamshire, UK).

FKBP12 protein abundance was quantified by scanning developed films on a calibrated GS-170 scanner (BioRad, Hertfordshire, UK). Films contained known amounts of total protein and recombinant FKBP12 as a normalization standard. Densitometry volumetric analysis of individual bands was carried out using Quantity One software (BioRad, Hertfordshire, UK).

Sea urchins (Psammechinus; University Marine Biological Station, Isle of Cumbrae, UK) were placed in artificial sea water at 16°C that contained 430 mM NaCl, 27 mM MgCl₂, 28 mM MgSO₄, 10 mM CaCl₂, 10 mM KCl, 2.5 mM NaHCO₃, 1 mM EDTA (pH adjusted to 8 with NaOH). Eggs were released by intracoelomic injection of 0.5 M KCl, passed three to four times through nylon mesh (pore size 130 μm, 47% open area; Plastok, Birkenhead, UK), placed on poly-L-lysine-coated (0.02 mg ml^–1) glass coverslips and immersed in artificial seawater. 250 μM fluo-3 and 1 mM caged cADPR in buffer [20 mM Pipes, 500 mM KCl, 0.1 mM EGTA (pH 6.7 with KOH)] were microinjected into the eggs (Gillot and Whitaker, 1994). The injected volume was estimated from bubbles of air created by pressure ejection into a chamber of hydrated oil [immersion oil type DF (Nikon, Kingston-upon-Thames, UK)] from a method modified from that of Moore et al. (Moore et al., 1990). (The volume of the bubble was estimated by measuring its radius and assuming it to be spherical.) The maximum injected volume was calculated to be ∼2% of the egg volume, so the final concentration of caged cADPR inside the egg was ∼22 μM, comparable to that used in the smooth muscle cells. Experiments were performed at 16°C.

Results are expressed as means±SEM of n cells and a Student's t test (paired and unpaired as appropriate) applied to test and control conditions; P<0.05 was considered to be significant. Microsoft^® Excel^® statistical software was used with either t tests or one-way ANOVA with post hoc t test.

In single smooth muscle cells, [Ca²⁺]_c (expressed as the fluorescence ratio F/F₀) was increased by depolarization (–70 mV to +10 mV, ΔF/F₀=1.3±0.2, n=9, P<0.001; I_Ca=-152 ± 49 pA, Fig. 1Aa) and by photolysis of caged 25 μM InsP₃ (ΔF/F₀=1.5±0.1, n=9, P<0.001; Fig. 1Ab) (McCarron et al., 2000). By contrast, photolysis of 50 μM caged cADPR did not increase [Ca²⁺]_c (ΔF/F₀=0.1±0.0, n=9, P>0.05; Fig. 1Bb), although depolarization (–70 mV to +10 mV, ΔF/F₀=0.8±0.2, n=9, P<0.01, I_Ca=–109±38 pA; Fig. 1Ba) and 10 mM caffeine (ΔF/F₀=1.8±0.5, n=9, P<0.01; Fig. 1Bc) were each effective. Photolysis of a tenfold increase in the concentration to 500 μM cADPR was no more effective (ΔF/F₀=0.0±0.01, n=4, P>0.05), although again depolarization (–70 mV to 0 mV, ΔF/F₀=0.8±0.2, n=4, P<0.05) and 10 mM caffeine (ΔF/F₀=1.4±0.3, n=4, P<0.05) each increased F/F₀ (data not shown). As a control, to ensure the efficacy of the caged cADPR, the ability of the photolysed compound to release Ca²⁺ in sea urchin eggs (a cell type in which cADPR's ability to modulate RyR is agreed) (Galione et al., 1991; Lee et al., 1993) was confirmed (Fig. 2; ΔF/F₀=3.1±0.9, n=10, P<0.01; time to peak 14±3 seconds).

The ability of cADPR to release Ca²⁺ from the SR is dependent on calmodulin (Lee et al., 1994). To determine whether the inability of cADPR to release Ca²⁺ from the SR in colonic myocytes was due to a low cytosolic concentration of calmodulin, the contractile response (also dependent on calmodulin) (Stull et al., 1988) following depolarization or photolysis of caged cADPR was measured. 50 μM cADPR again failed to increase [Ca²⁺]_c (Fig. 3C; n=9, P>0.05) or contract the myocytes (Fig. 3Di,Ei; change in control cell length (Δl/l₀) =0.0±0.0 after cADPR, n=18, P>0.05). Depolarization (–70 mV to +10 mV; Fig. 3A) activated I_Ca (Fig. 3B; –201±36 pA), increased [Ca²⁺]_c (Fig. 3C; ΔF/F₀=1.4±0.1, n=9, P<0.05) and evoked contraction (Fig. 3Dii,Eii; l/l₀=1.0±0.0; l/l₀=0.7±0.0 8 seconds after the depolarization, n=18, P<0.05). The cells subsequently relaxed over a period of minutes to approximately their original length (l/l₀=0.9±0.02, Fig. 3Dii,Eii).

In a subsequent series of experiments (not shown) 1 μM calmodulin was introduced into the cells via the patch pipette filling solution but cADPR remained ineffective. In these experiments, resting [Ca²⁺]_c was 1.1±0.1 F/F₀ and remained at 1.1±0.1 F/F₀ after photolysis of cADPR (n=8). The [Ca²⁺]_c increase to depolarization to +10 mV (from –70 mV; ΔF/F₀=1.1±0.3) remained unchanged by prior photolysis of cADPR (ΔF/F₀=1.2±0.3; n=6, P>0.05).

Although cADPR failed to activate RyR, it might have sensitized them to Ca²⁺, which itself activates RyR to induce CICR. Therefore, the effect of cADPR on depolarization-evoked (–70 mV to +10 mV) increases in [Ca²⁺]_c was examined. 50 μM cADPR released simultaneously with the onset of depolarization or 0.5 seconds, 1 second, 2 seconds, 3 seconds, 4 seconds, 5 seconds, 10 seconds or 15 seconds beforehand failed to alter significantly either the depolarization-evoked rise in [Ca²⁺]_c (Fig. 4; Table 1) or the amplitude of I_Ca (Table 1). A tenfold increase in the concentration of caged cADPR to 500 μM was no more effective (Table 2), nor were the [Ca²⁺]_c increases in response to depolarization to –20 mV (from –70 mV; ΔF/F₀=0.7±0.2) altered by prior photolysis of cADPR (ΔF/F₀=0.7±0.2; n=6, P>0.05).

By contrast, CICR was induced by a recognized activator of RyR. Caffeine (500-750 μM, added to the bath 3 minutes beforehand) increased the Ca²⁺ transient evoked by depolarization (–70 mV to +10 mV) to 168±40% of control (172±35 nM before and 237±47 nM with caffeine, n=10, P<0.01; Fig. 5A). Upon washout of the drug, the depolarization-evoked response returned towards control 126±8% (177±43 nM, n=6). That caffeine indeed induced CICR following depolarization was confirmed when the relationship between the amount of Ca²⁺ entering the cell [i.e. the `calculated' Ca<sup>2+</sup> increase from the integral of I<sub>Ca</sub> (see Materials and Methods)] and the `measured' increase in [Ca²⁺]_c (as demonstrated by the change in fluorescence of fluo-3) were compared. 100 milliseconds after the start of the depolarization, the ratio of `calculated' to `measured' [Ca²⁺]_c was 424±143 without caffeine (Fig. 5Di) and 163±51 in caffeine (Fig. 5Dii; n=10, P<0.05). Following washout, the ratio returned towards control (266±74, Fig. 5Diii). The decreased relationship (see above) between the `calculated' and `measured' Ca²⁺ increases, in caffeine, suggests that a release of Ca²⁺ occurred from the SR subsequent to I_Ca (i.e. CICR).

The ability of cADPR to release Ca²⁺ in cardiac myocytes is reportedly temperature dependent (Iino et al., 1997). Accordingly, the effects of cADPR in the present study were re-examined at 35±2°C. cADPR again failed to change [Ca²⁺]_c (ΔF/F₀=0.0±0.0, n=7, P>0.05). Increases in [Ca²⁺]_c were, however, produced each by depolarization (–70 mV to +10 mV; ΔF/F₀=1.7±0.4, n=7, P<0.05) and by caffeine (ΔF/F₀=1.7±0.7, n=7, P<0.05) at 35°C, but these depolarization-evoked increases were unaffected by cADPR when it was photolysed at different time points beforehand (Table 3).

Although cADPR failed to increase the sensitivity of RyR to Ca²⁺, it could have increased the sensitivity of RyR to caffeine. 50 μM caged cADPR was therefore photolysed 1 second before caffeine application. It was confirmed that consistent submaximal Ca²⁺ transients to repeated applications of caffeine (see Materials and Methods; ΔF/F₀=1.4±0.4, n=10) were significantly different (P<0.01) from maximal responses (ΔF/F₀=2.3±0.5, n=10). cADPR failed to increase the submaximal response to caffeine (ΔF/F₀=1.4±0.4, n=10, P>0.05; Fig. 6).

Furthermore, cADPR failed to increase [Ca²⁺]_c (ΔF/F₀=0.0±0.0, n=3, data not shown) in the presence of 500 μM caffeine [a concentration insufficient to increase (Ca²⁺)_c but that enhanced the sensitivity of RyR to Ca²⁺ (Fig. 5)]. Together, these results indicate that cADPR by itself failed to release Ca²⁺ from the SR or to modulate Ca²⁺ release when [Ca²⁺]_c was elevated by either depolarization or caffeine.

Failure to detect a cADPR-induced increase in [Ca²⁺]_c could have been caused by the use of global fluorescence measurements, which could have masked localized increases in [Ca²⁺]_c. Accordingly, the effect of cADPR on STOCs was examined (Fig. 7). STOCs arise by the activity of Ca²⁺-activated K⁺ channels on the sarcolemma that have been activated by the local subsarcolemma [Ca²⁺] increases from the SR via RyR (Benham and Bolton, 1986). STOCs thus provide an indirect measure of the local subsarcolemma [Ca²⁺] generated by RyR activity. The area under the curve of the STOC current trace (I_KCa) was calculated over a 5-second period and used as a measure of STOC activity in the presence and absence of cADPR. The probability of a STOC occurring (P_STOC) was also determined before and after photolysis of caged cADPR (Fig. 7vi). This analysis showed that photolysis of cADPR had no significant effect on the frequency or amplitude of STOCs (AUC before 199±82 pC and 184±57 pC after cADPR, n=8, P>0.05, Fig. 7i,ii) or the probability of a STOC occurring over a longer time period (∼2.5 minutes). Caffeine, by depleting the SR of Ca²⁺, significantly reduced both the amplitude and frequency of STOCs (AUC before 227±109 pC and –489±371 pC after caffeine, n=5, P<0.05, Fig. 7i,ii). STOCs recovered following caffeine removal (AUC 59±32 pC, P>0.05, n=4).

The inability of photolysed cADPR to modify the release of Ca²⁺ from the SR might be due to high endogenous levels of cADPR, which could have already maximally enhanced the sensitivity of RyR or another intracellular moiety, to Ca²⁺ (Walseth and Lee, 1993). If so the inhibition of endogenous cADPR by a selective antagonist (8-bromo cADPR) (Walseth and Lee, 1993) should decrease changes in [Ca²⁺]_c evoked by cADPR-sensitive mechanisms. In our study, 20 μM 8-bromo cADPR failed to decrease Ca²⁺ transients evoked by depolarization (ΔF/F₀=2.2±0.2 before and 2.1±0.3 in 8-bromo cADPR; P>0.05), InsP₃ (ΔF/F₀=3.2±0.2 before and 2.7±0.4 in 8-bromo cADPR; P>0.05) or caffeine (ΔF/F₀=3.0±0.3 before and 4.2±0.5 in 8-bromo cADPR, n=6 in all cases, P<0.05; data not shown).

Another reason for the lack of activity of cADPR could have been the absence of an appropriate binding site for the compound on RyR (i.e. FKBP12.6) (Noguchi et al., 1997; Tang et al., 2002). To confirm the presence of FKBP12.6, increasing protein loads from guinea-pig colon and brain (control) homogenates were examined by immunoblotting. Antiserum SA168 detected both FKBP12 and FKBP12.6, whereas the antibody c-19 was more sensitive to FKBP12 than to FKBP12.6. Strong bands corresponding to FKBP12/FKBP12.6 in brain (Fig. 8Ai,ii) were identified. In colon, however, they were absent or very weak (Fig. 8Ai) and more clearly detected using the antibody c-19 (Fig. 8Aii) than with the antiserum SA168, suggesting that little FKBP12.6 was present and that FKBP12 was the major FKBP.

If FKBP12.6 acted as the receptor for cADPR to modulate RyR opening (Noguchi et al., 1997; Tang et al., 2002), the absence of the protein (FKBP12.6), in colonic myocytes, might account for the lack of cADPR activity. To explore this possibility, recombinant FKBP12.6 was introduced into the myocytes to compensate for its absence (using the whole cell patch electrode) and the effect of cADPR on RyR mediated Ca²⁺ release, as assessed by the occurrence of STOCs, was re-examined (Fig. 8B). The frequency and amplitude of STOCs were significantly reduced by FKBP12.6 alone (AUC over a 100-second period was 12±4 nanocoulombs (nC) 100 seconds and 4±2 nC 700 seconds after the start of recording, n=5, P<0.05, Fig. 8B) compared with those in time-matched control experiments in which the STOC amplitude and frequency were not significantly different (AUC over 100-second period was 18±5 nC 100 seconds and 13±3 nC 700 seconds after the start of recording, n=7, P>0.05, Fig. 8C). cADPR in the presence of FKBP12.6 failed to alter STOCs (Fig. 8Bi, over a 5 second period AUC was 256±113 pC in control and 290±93 pC after cADPR) or [Ca²⁺]_c (Fig. 8Bii). These results suggest that, in the presence of FKBP12.6, cADPR remained ineffective in modulating Ca²⁺ release and raise the question of whether or not cADPR binds to FKBP12.6.

To investigate whether cADPR binds to FKBP12.6, brain homogenates were used as a source of the protein, which was largely absent from colonic myocytes. FK506 binds to both FKBP12 and FKBP12.6 (Marks, 1996), and served as a positive control for the cADPR binding experiments. Densitometric analysis on FKBP12/FKBP12.6 in membrane fractions were performed on samples treated with either FK506 or cADPR. FK506 significantly reduced (by 45%) the FKBP12/FKBP12.6 detected in the membrane fraction (n=4; P<0.05) confirming the adequacy of detection of binding by the densitometric analysis. However, cADPR did not alter the FKBP12/FKBP12.6 detected (Fig. 8Di,ii, P>0.05, n=3), suggesting that cADPR does not bind significantly to FKBP12 or FKBP12.6.

The rate of Ca²⁺ removal from the cell increased with increasing [Ca²⁺]_c (i.e. the substrate), which is typical of many enzyme-substrate reactions. This rate was accelerated when free cADPR was introduced into the cell (using the access afforded by the whole cell electrode) over a [Ca²⁺]_c range of 275-450 nM (Fig. 9A). For example, at 350 nM [Ca²⁺]_c, the rate of decline was 80±6 nM second^–1 (control; n=25) and 120±16 nM second^–1 in 300 μM cADPR (n=7; P<0.01). In this series of experiments, the peak [Ca²⁺]_c values during the depolarization (1.2±0.1 ΔF/F₀ in control and 1.4±0.3 ΔF/F₀ in cADPR; P>0.05) were not different. However, 50 μM cADPR photolysed up to 15 seconds beforehand had no effect on the rate of Ca²⁺ removal following depolarization (–70 mV to 0 mV) at either 20±2°C or 35±2°C. At a higher concentration (500 μM), cADPR also failed to affect the rate of Ca²⁺ removal when photolysed up to 10 seconds before depolarization but significantly increased the rate when photolysed 15 seconds beforehand (P<0.05; Table 2), a time consistent with the delays inherent in cADPR effects (e.g. Cui et al., 1999); the reason for the delay is not understood.

The [Ca²⁺]_c range over which cADPR increased the rate of decline suggested an acceleration of plasma membrane Ca²⁺ pump activity (Bradley et al., 2002). To examine the effects of cADPR itself on the plasma membrane Ca²⁺ pump, the rates of Ca²⁺ removal in cADPR had subtracted from it, the control rate (i.e. with all removal mechanisms operative) minus those removal rates occurring after inhibiting mitochondria [by carbonyl cyanide m-chlorophenyl hydrazone (CCCP) and oligomycin together, each 5 μM], the SR Ca²⁺ pump (500 nM thapsigargin) and the Na⁺-Ca²⁺ exchanger (Na⁺-free bathing solution) to yield an inferred plasma membrane Ca²⁺ pump activity in cADPR (Fig. 9B). The inferred pump activity in cADPR was substantially increased. For example, at 300 nM [Ca²⁺]_c, the inferred plasma membrane Ca²⁺ pump rate was 47±8 nM second^–1 in the control, and this increased to 82±16 nM second^–1 in cADPR.

Although cADPR itself increased the rate of Ca²⁺ removal from the cytosol the antagonist 8-bromo cADPR (20 μM) slowed it significantly following photolysed InsP₃(Fig. 10Aa) or depolarization(Fig. 10Ab) evoked increases. For example, at 225 nM [Ca²⁺]_c (evoked by InsP₃) the rate of Ca²⁺ removal was 108±13 nM second^–1 in controls but was reduced by 8-bromo cADPR to 71±12 nM second^–1 (n=8, P<0.01). Similarly, following a depolarization-evoked increase, the rate of removal in controls was 39±7 nM second^–1 and 23±7 nM second^–1 in 8-bromo cADPR (n=8, P<0.01) at 150 nM [Ca²⁺]_c. In keeping with the observation of a slowed rate of Ca²⁺ removal, 8-bromo cADPR (20 μM) increased the resting steady-state [Ca²⁺]_c significantly (from 1.2±0.1 F/F₀ in control to 1.9±0.2 F/F₀ in 8-bromo cADPR, n=8, P<0.01), compatible with there being endogenous cADPR present, which presumably accelerated the rate of Ca²⁺ removal. I_Ca was reduced in the presence of the antagonist [from –95±13 pA in control to –61±13 pA in the presence of the antagonist (n=6; P<0.01)], again suggesting that removal is the major effect of cADPR. The increase in steady-state [Ca²⁺]_c (in the face of reduced entry) by 8-bromo cADPR suggests that Ca²⁺ efflux across the sarcolemma might be reduced. To study this, the effects of 8-bromo cADPR on the sarcolemma Ca²⁺ pump were examined when the other removal mechanisms (mitochondria, SR Ca²⁺ pump and Na⁺-Ca²⁺ exchanger) had been inhibited. CCCP, oligomycin, thapsigargin and a Na⁺-free bathing solution together significantly slowed the rate of Ca²⁺ removal (Fig. 10Ba,b). For example, at 250 nM [Ca²⁺]_c, the rate of decline was 67±9 nM second^–1 (control) and this slowed to 34±7 nM second^–1 in the presence of removal inhibitors (P<0.01). After 8-bromo cADPR addition (in the continued presence of the removal inhibitors), the baseline [Ca²⁺]_c increased (from 1.1±0.1 F/F₀ control to 2.0±0.2 F/F₀ in 8-bromo cADPR; P<0.01, n=6) and the rate of [Ca²⁺]_c removal slowed further (6±3 nM second^–1; Fig. 10Bb). In this series of experiments, the depolarization-evoked [Ca²⁺]_c increase was 1.6±0.2 ΔF/F₀ (n=8) in control and 1.4±0.3 ΔF/F₀ (n=8) in the presence of the Ca²⁺ removal inhibitors. In 8-bromo cADPR, the peak [Ca²⁺]_c achieved during depolarization was significantly reduced to 0.5±0.1 ΔF/F₀ (n=6), presumably because of the reduced I_Ca (–263±55 pA in control and –35±10 pA in 8-bromo cADPR with the Ca²⁺ removal inhibitors).

RyR is important in Ca²⁺ signalling in several cell types, and modulators of RyR activity are of potential value in this process. In smooth muscle, RyR contributes to physiological functioning by increasing (via CICR) or lowering the bulk average [Ca²⁺]_c (Bolton et al., 1999; Sanders, 2001; Guerrero-Hernandez et al., 2002). The latter response occurs when the localized release of Ca²⁺ from RyR activates Ca²⁺-activated K⁺ channels, to generate STOCs, to hyperpolarize the sarcolemma and to reduce Ca²⁺ entry via voltage-dependent Ca²⁺ channels. Indeed, agonists that generate InsP₃ might evoke contraction by suppressing STOCs to promote Ca²⁺ entry via voltage-dependent Ca²⁺ channels (McCarron et al., 2002). RyR might thus play an important role in excitation-contraction coupling in colonic smooth muscle (McCarron et al., 2002).

Among substances proposed as RyR regulators, cADPR is of particular interest and has been found to be a potent mobilizer of intracellular Ca²⁺ in several invertebrate and vertebrate cell types (Lee, 2001). Yet there is a controversial literature pertaining to cADPR and Ca²⁺ release. On the one hand, several investigations have demonstrated that cADPR can increase Ca²⁺ release by a process involving RyR either directly (Galione and Sethi, 1996; Petersen and Cancela, 1999) or via intermediaries (Noguchi et al., 1997; Lee, 2001). On the other hand, other investigators have failed to repeat these observations (Nixon et al., 1994; Iizuka et al., 1998; Copello et al., 2001; Lukyanenko et al., 2001) and argued that RyR activity is not affected by cADPR, observations that might reflect tissue- or species-specific differences.

Against this background of controversy, the present results have provided evidence that cADPR does not directly increase Ca²⁺ release in colonic myocytes. This view is supported by several observations. First, as a positive control, caffeine (which activates RyR to release Ca²⁺) was used to establish the consequences of RyR activity on [Ca²⁺]_c. Three separate effects of caffeine on Ca²⁺ homeostasis were identified: (1) an increase in bulk average [Ca²⁺]_c; (2) an inhibition of STOC activity, consistent with modulation of local subsarcolemma [Ca²⁺]; (3) CICR occurred – there was an increased rise in bulk average [Ca²⁺]_c for a given time-integrated I_Ca in the presence of caffeine compared with controls. Each of these observations supports modulation of RyR by caffeine. In contrast to the action of caffeine, cADPR did not increase bulk average [Ca²⁺]_c, alter local subsarcolemma [Ca²⁺] or change the relationship between the rise in bulk average [Ca²⁺]_c and the time-integrated I_Ca. cADPR also failed to affect Ca²⁺ release from RyR evoked by caffeine. The lack of effect was unlikely to have arisen from any deterioration of cADPR itself or deficiencies in the apparatus or in the experimental design. The same stock of cADPR increased bulk average [Ca²⁺]_c in sea urchin eggs, under similar experimental conditions and apparatus to those used to study its effects on smooth muscle. Nor could the results have arisen as a consequence of high endogenous levels of cADPR saturating RyR; the cADPR antagonist 8-bromo-cADPR did not reduce depolarization-, caffeine- or InsP₃-evoked Ca²⁺ transients, suggesting that there is no synergistic involvement of cADPR in the Ca²⁺-releasing effects of these events. Together, these results suggest that, in intact single colonic myocytes [a cell type with RyR that contributes to the tissue's physiological responses (i.e. functional RyR)], cADPR does not act as a direct modulator of RyR activity.

Failure to demonstrate a direct modulation of release by cADPR is not unique to colonic smooth muscle. For example, none of Jurkat T-cells, RBL-2H3 cells or HEK-293 cells responded to cADPR (A. Fleig and R. Penner, personal communication). Nor did bronchial, intestinal (Iizuka et al., 1998), coronary (V. Ganitkevich, personal communication), cerebral artery (T. Kamishima and J. Quayle, personal communication; H. Knot, personal communication), aorta or vas deferens (Nixon et al., 1994) smooth muscle respond to cADPR. Negative findings with cADPR on cardiac myocytes have also been reported (Guo et al., 1996; Lukyanenko et al., 2001) (G. Smith, personal communication), in contrast to those mentioned in the introduction.

Some reconciliation of the differences in the results with cADPR in cardiac myocytes may be found in the different temperatures at which the experiments were conducted (Guo et al., 1996; Iino et al., 1997). At 36°C (Iino et al., 1997), but not at room temperature (Guo et al., 1996), cADPR was suggested to be effective in modulating Ca²⁺ release. However, cADPR failed to alter the activity of RyR1 and RyR2 derived from mammalian striated muscle at both room temperature and 37°C, as measured by their activity in planar bilayers or in SR microsomes, or the binding of [³²P]cADPR to the microsomes (Copello et al., 2001). Differences between the present findings in smooth muscle and those in the heart (Iino et al., 1997) are unlikely to be due to the effects of temperature. cADPR was ineffective in increasing bulk average [Ca²⁺]_c or inducing CICR at either 35°C or room temperature. The effectiveness of cADPR in raising [Ca²⁺]_c in sea urchin eggs at 16°C was confirmed in the present study.

An alternative reason for the failure of cADPR to modulate RyR activity in colonic myocytes is that, although RyR has been demonstrated to be a target for cADPR in some cell types, RyR itself might not, in others, respond directly to cADPR (Fruen et al., 1994; Copello et al., 2001; Lukyanenko et al., 2001), or its effects might be prevented by the normal ATP or glucose concentrations within cells (Sitsapesan et al., 1994; Cancela et al., 1998). In some tissues cADPR might release Ca²⁺ by a pathway which involves atypical RyR or does not involve RyR at all. For example, the release of Ca²⁺ by cADPR was not prevented by depletion of the caffeine-sensitive store or by conventional RyR antagonists, and was inhibited by the InsP₃ receptor blocker heparin (Morrissette et al., 1993; Thorn et al., 1994; Kannan et al., 1996; Lahouratate et al., 1997; Prakash et al., 2000). Indeed cADPR might interact with all isoforms of the InsP₃ receptor, stimulate InsP₃ binding to the type 1 InsP₃ receptor (Vanlingen et al., 2001) or inhibit InsP₃-mediated Ca²⁺ release (Missiaen et al., 1998).

Together, these studies present a complicated picture of cADPR's relationship with RyR. This led us to investigate the role of a proposed intermediary between cADPR and RyR. FKBP12.6, a protein reportedly linked to RyR2 and RyR3 in some studies, might act as a receptor for cADPR (Noguchi et al., 1997). FKBP12.6 might, by interacting with RyR, suppress Ca²⁺ release (Prestle et al., 2001; Xin et al., 2002) and the suppression might be relieved by cADPR. Yet FKBPs (12 and 12.6) are not present in all tissues. For example, in the present study [as in humans (Baughman et al., 1997)], FKBP12.6 was present in brain but not colonic myocytes. It might be argued that, given the lack of effect of cADPR on Ca²⁺ signalling, the absence of FKBP accords with the proposal that the protein was a necessary receptor for cADPR. However, this would be incorrect. After the introduction of recombinant FKBP12.6, to compensate for its absence in colonic myocytes, cADPR remained ineffective in modulating Ca²⁺ release from the SR and supports the view that cADPR does not modulate RyR activity via FKBP12.6. cADPR, furthermore, did not reduce the immunoblot signal detected by the antiserum to FKBPs in brain, whereas FK506 was effective, suggesting that cADPR might not bind to the FKBPs present (see also Bultynck et al., 2001b).

Another possible explanation for the lack of cADPR activity on Ca²⁺ release could have been a shortage of calmodulin (Lee et al., 1994), an essential intermediary for smooth muscle contraction (Stull et al., 1988). The present study showed this to be unlikely: repetitive contractile responses to depolarization-evoked increases in [Ca²⁺]_c, consistent with the presence of calmodulin in myocytes, were obtained routinely. cADPR, furthermore, remained ineffective following dialysis of calmodulin into the myocytes at concentrations similar to those which supported activity in other cells (Lee et al., 1994).

In contrast to the ineffectiveness of cADPR and its antagonist 8-bromo-cADPR on Ca²⁺ release, the compounds were surprisingly effective at modulating Ca²⁺ removal mechanisms – cADPR increased and 8-bromo-cADPR reduced the rate of Ca²⁺ removal. The antagonist also evoked a steady-state rise in [Ca²⁺]_c, an effect that requires an altered flux of the ion across the sarcolemma (because the SR has finite capacity). Because I_Ca – the major influx pathway – was inhibited, a decreased removal of Ca²⁺ across the sarcolemma by the antagonist might account for the rise in [Ca²⁺]_c. Neither a decrease in steady-state [Ca²⁺]_c nor an increase in SR Ca²⁺ content with cADPR was observed, presumably because of the labile nature of the compound after its photorelease. Pharmacological removal of mitochondria, SR Ca²⁺ pump and Na⁺-Ca²⁺ exchange activities, leaving the plasma membrane Ca²⁺ pump as the only operative Ca²⁺ removal mechanism, enabled 8-bromo-cADPR to increase baseline [Ca²⁺]_c and to slow the rate of Ca²⁺ removal substantially. This suggests that 8-bromo-cADPR inhibited the plasma membrane Ca²⁺ pump. Indeed, indirect evidence of involvement of the plasma membrane Ca²⁺ pump in cADPR's acceleration of [Ca²⁺]_c decline was the observations made of the inferred plasma membrane Ca²⁺ pump activity. The inferred plasma membrane Ca²⁺ pump activity in cADPR was obtained by subtracting from the rate of removal in cADPR the rate of removal in control (i.e. with all removal mechanisms operative) minus the rate obtained after Ca²⁺ removal by mitochondria, SR Ca²⁺ pump and Na⁺-Ca²⁺ exchange activity had been inhibited (i.e. leaving only plasma membrane Ca²⁺ pump operative; not shown). The measurement revealed a substantial increase in plasma membrane Ca²⁺ pump activity by cADPR. Interestingly in this connection, the cADPR binding site within cells has been identified as a 140 kDa protein (Walseth et al., 1993), a molecular mass similar to that of the plasma membrane Ca²⁺ pump (140 kDa) (Schatzmann, 1989).

By modulating Ca²⁺ pump activity on the SR (Lukyanenko et al., 2001) and plasma membrane (present study), the effects of cADPR and its antagonists on Ca²⁺ homeostasis and the responses to drugs and neurotransmitters would vary depending on the major Ca²⁺ source necessary for the response (i.e. SR or extracellular). For example, if the SR was the major source then the antagonist would reduce and the agonist (cADPR) enhance the response (each by modulating the SR Ca²⁺ content) (Lukyanenko et al., 2001); if external Ca²⁺ was the major source, the antagonist could enhance, have no effect on or reduce the response, depending on the balance of effects of the compound on influx (inhibited) and removal (inhibited). By modulating Ca²⁺ removal rather than RyR directly, cADPR may have important effects on Ca²⁺ regulation and provide an explanation for the diverse effects of cADPR and its antagonists on Ca²⁺ homeostasis.

The Wellcome Trust (060094/Z/00/Z) and the British Heart Foundation (PG/2001079) funded this work. We thank J. Craig for excellent technical assistance, J. Laurie for advice on the care of sea urchins, M. Whitaker and C. Leckie (Department of Physiological Sciences, University of Newcastle, Newcastle upon Tyne, UK) for assistance in the preparation and injection of eggs, T. Seidler for the antiserum SA168 and recombinant FKBP12 and FKBP12.6, G. Smith for helpful discussions, Fujisawa for the gift of FK506, and S. Chalmers for carefully reading the manuscript.