Spontaneous purinergic neurotransmission in the mouse urinary bladder

Spontaneous purinergic neurotransmission was characterized in the mouse urinary bladder, a model for the pathological or ageing human bladder. Intracellular electrophysiological recording from smooth muscle cells of the detrusor muscle revealed spontaneous depolarizations, distinguishable from spontaneous action potentials (sAPs) by their amplitude (< 40 mV) and insensitivity to the L-type Ca2+ channel blocker nifedipine (1 μm) (100 ± 29%). Spontaneous depolarizations were abolished by the P2X1 receptor antagonist NF449 (10 μm) (frequency 8.5 ± 8.5% of controls), insensitive to the muscarinic acetylcholine receptor antagonist atropine (1 μm) (103.4 ± 3.0%), and became more frequent in latrotoxin (LTX; 1 nm) (438 ± 95%), suggesting that they are spontaneous excitatory junction potentials (sEJPs). Such sEJPs were correlated, in amplitude and timing, with focal Ca2+ transients in smooth muscle cells (measured using confocal microscopy), suggesting a common origin: ATP binding to P2X1 receptors. sAPs were abolished by NF449, insensitive to atropine (126 ± 39%) and increased in frequency by LTX (930 ± 450%) suggesting a neurogenic, purinergic origin, in common with sEJPs. By comparing the kinetics of sAPs and sEJPs, we demonstrated that sAPs occur when sufficient cation influx through P2X1 receptors triggers L-type Ca2+ channels; the first peak of the differentiated rising phase of depolarizations – attributed to the influx of cations through the P2X1 receptor – is of larger amplitude for sAPs (2248 mV s−1) than sEJPs (439 mV s−1). Surprisingly, sAPs in the mouse urinary bladder, unlike those from other species, are triggered by stochastic ATP release from parasympathetic nerve terminals rather than being myogenic.

In most mammals, excitation of the detrusor muscle of the urinary bladder is achieved through the parasympathetic activation of both purinergic and cholinergic neurotransmission, leading to voiding of the bladder. The brief, initial phase of contraction is atropine insensitive (Henderson & Roepke, 1934) and purinergic (Burnstock et al. 1978); the second phase, of longer duration, is cholinergic (Krell et al. 1981). The balance between purinergic and cholinergic neurotransmission observed experimentally varies widely. This variation is often attributed to the following three factors. (i) Species used. Detrusor contraction is almost entirely controlled by purinergic neurotransmission in cat (Nergårdh, 1981) and rabbit (Sibley, 1984), whereas cholinergic neurotransmission dominates in the Old World primates (Craggs & Stephenson, 1985). (ii) Age. For example, there is an increase with age in the contribution of purinergic This paper has online Supplemental material. neurotransmission to bladder voiding in the human (Yoshida et al. 2001) and rat (Saito et al. 1991). (iii) Experimental protocol. Single pulse or low frequency, repetitive transmural electrical stimulation evokes an atropine-insensitive contraction (Ambache & Zar, 1970); high frequency repetitive stimulation evokes excitation that comprises a much larger proportion of cholinergic than purinergic neurotransmission (Sibley, 1984).
Normal voiding of the human bladder is attributed to the actions of acetylcholine as the primary excitatory neurotransmitter. The administration of the muscarinic cholinergic receptor antagonist atropine to healthy subjects resulted in difficulty in initiating, and pain during, micturition (Cullumbine et al. 1955) and in some in vitro studies, atropine was shown to abolish electrically evoked contraction (Sjörgen et al. 1982;Sibley, 1984). There are other in vitro studies, however, which suggest a role for purinergic neurotransmission in normal voiding; Cowan & Daniel (1983), for example, demonstrated J Physiol 586.23 that atropine abolished only 50% of stimulated contractions, and others (e.g. Elmér et al. 1986;Hoyle et al. 1989;Luheshi & Zar, 1990) have shown atropine-resistant contractions, albeit comprising a smaller proportion of excitatory neurotransmission than is attributed to the actions of acetylcholine. The wide variation in reported contributions of purinergic neurotransmission to bladder voiding is partly explained by the methodological difficulty of evoking neural stimulation without also evoking potentially masking direct muscle stimulation (Hoyle et al. 1989). It is, however, generally agreed that purinergic neurotransmission is up-regulated in bladder dysfunction; Sjörgen et al. (1982) demonstrated, for example, that while atropine abolished contractions in preparations from normal bladders, preparations from pathological bladders exhibited contractions that were up to 50% atropine resistant.
It has been demonstrated that ATP, released from parasympathetic nerve terminals, binds to P2X receptors (Burnstock & Kennedy, 1985;Howson et al. 1988) on urinary bladder smooth muscle cells (SMCs), producing an excitatory junction potential (EJP) that triggers the opening of L-type Ca 2+ channels and the subsequent initiation of an action potential (Hoyle & Burnstock, 1985;Howson et al. 1988). Heppner et al. (2005) recently identified purinergic focal Ca 2+ transients, corresponding to the influx of Ca 2+ through P2X receptors upon ATP binding in rat urinary bladder smooth muscle.
In summary, although purines may not be the major excitatory neurotransmitters responsible for human bladder voiding, their up-regulated role in pathological states and with ageing makes purinergic neurotransmission a key area of study. The aims of the present study are to: (i) use simultaneous electrophysiological recording of membrane potential and high-speed imaging of Ca 2+ fluorescence to determine whether a relationship exists between spontaneous purinergic Ca 2+ transients and spontaneous depolarizations in mouse urinary bladder smooth muscle, and (ii) to characterize spontaneous depolarizations and, in particular, to determine the mechanisms underlying their generation.

General
Mice of the C57BL/6 strain, of either sex, weighing 18-30 g, were killed by head concussion followed by cervical dislocation. Efforts were made to minimize the number of animals used and their suffering; all experiments were carried out in accordance with the UK Animals (Scientific Procedures) Act 1986 and European Communities Council Directive 86/09/EEC. The urinary bladder was removed and its ventral wall was opened longitudinally from the bladder neck (posterior) to the top of the dome (anterior). The urothelium was removed by careful dissection using iris scissors; care was taken not to stretch or damage the underlying detrusor muscle. Tissue strips, which contained a few bundles of smooth muscle, 6-8 mm long and 1-2 mm wide, were cut along the craniocaudal axis of the dorsal surface of the detrusor. Strips were pinned out on a Sylgard (Dow Corning Corporation, Midland, MI, USA)-lined plate at the bottom of a shallow chamber (volume, approximately 1 ml), which was mounted on the stage of an upright microscope. Preparations were superfused with warmed (35 • C) physiological saline solution (PSS) (composition, mM: NaCl, 120; KCl, 5.9; MgCl 2 , 1.2; CaCl 2 2.5; NaHCO 3 , 15.5; NaH 2 PO 4 , 1.2 and glucose, 11.5; gassed with 95% O 2 and 5% CO 2 ) at a constant flow rate (100 ml h −1 ), maintaining a pH of 7.2-7.3 (Hashitani & Brading, 2003). Preparations, when pinned, were allowed to equilibrate (i.e. to accommodate under tension) for at least 30 min before electrophysiological recording or Ca 2+ imaging.

Electrophysiological recordings
Individual bladder smooth muscle cells in muscle bundles were impaled with glass capillary microelectrodes, filled with 0.5 M KCl (tip resistance, 100-300 M ). Membrane potential changes were recorded using a high input impedance amplifier (Axoclamp-2B, Axon Instruments, Inc., Sunnyvale, CA, USA) and displayed on a digital oscilloscope (DSO 1602, Gould, Ilford, UK). Membrane potential changes were digitized using PowerLab/4SP (ADInstruments, Chalgrove, UK) at 4 kHz and stored on computer for later analysis.
Recordings were made from a minimum of six cells, for at least 5 min each, in the absence of drugs ('control' measurements). One long, continuous recording was then made, which included a control period (typically 10 min), and then a period (up to 1 h) in the presence of the drug. Recordings were subsequently made from a further six cells (minimum), for at least 5 min each. Subsequent analysis of the effect of the superfused reagents compared the median frequency for all cells recorded in the control period with the median frequency of events in the presence of the reagent, in a paired comparison.
A series of 200 frames was captured at 13.5 Hz to generate one image set. Such sets were acquired once every minute. Ten sets were generated for each region of the preparation, with at least three regions sampled per drug treatment and/or preparation.
Microelectrodes used for simultaneous confocal recordings (tip resistances of 200-300 M ) were filled with a potassium salt of Oregon Green BAPTA-1 (100 μM in 0.5 M potassium acetate; weight 10 kDa; Invitrogen). The microelectrode tip (or part of it) was always within the field of view of the confocal image.
Image analysis was performed using plug-ins for ImageJ (http://rsb.info.nih.gov/ij/download.html) that were custom-written by Dr R. J. Amos to detect focal increases in fluorescence of the Ca 2+ indicator. Data were exported to Excel (Microsoft, Redmond, WA, USA) for formatting and then to Spike 2 (Cambridge Electronic Design, Cambridge, UK) for analysis in conjunction with electrophysiological recordings.

Analysis of correlations between focal Ca 2+ transients and spontaneous depolarizations
The peak amplitudes of focal Ca 2+ transients and spontaneous depolarizations were calculated using a custom-written script for Spike 2. The threshold for determining focal Ca 2+ transients varied between recordings in different SMCs. The threshold was adjusted until it provided specificity comparable to manual counting ( F/F range of 0.06-0.1).
Occasionally, > 1 coincident focal Ca 2+ transient was observed within the impaled cell. In this instance, the amplitudes of spontaneous depolarizations and focal Ca 2+ transients were not included in the subsequent analysis.
In order to estimate the time of initiation of focal Ca 2+ transients, experiments were performed in which Ca 2+ fluorescence was measured along the long axis of a SMC at a high acquisition rate (2 kHz) and the amplitudes and timing of focal Ca 2+ transients were correlated with the membrane potential of the imaged cell, simultaneously measured using the previously described electrophysiological recording technique. Ca 2+ fluorescence was measured in a 20 μm-wide window around each focal Ca 2+ transient for the duration of each acquisition period (approx. 2 s) using ImageJ. Using a custom-written script for R (http://www.r-project.org) a curve was then fitted, comprising: y = C o (the resting fluorescence), before the initiation of the focal Ca 2+ transient (i.e. t < t o ), and A is an arbritrary constant.
The time of initiation (t o ) was that giving the minimum residual sum of squares for this curve fit. This function arises from a simple model of Ca 2+ influx through P2X 1 receptors, with Ca 2+ then cleared by a first-order process (see Appendix).

Drug sources, preparation and application
Tetrodotoxin (Tocris, Bristol, UK), nifedipine (Bayer, Leverkusen, North Rhine-Westphalia, Germany) and atropine sulphate (VWR; Lutterworth, UK) were stored frozen (−20 • C) at a concentration of 10 mM in DMSO. NF449 (at 1 mM; Tocris, Bristol, UK) and α-latrotoxin (at 300 nM; Sigma-Aldrich) were stored frozen in distilled water. Drugs were diluted to their final concentration in PSS and applied to the preparation by switching the superfusate from standard PSS.

Statistical analysis
The normality and homogeneity of variance was tested prior to further statistical analysis using Kolmogorov-Smirnov and Levene's tests, respectively (SPSS version 11, SPSS Inc., Chicago, IL, USA). Unless otherwise stated, statistical comparisons are Student's two-tailed paired t test. The term n, used in the presentation of statistical analyses throughout, refers to the number of animals used.

Pharmacology of spontaneous depolarizations
Intracellular recording of smooth muscle cells in the mouse urinary bladder revealed both spontaneous depolarizations (sDeps) and spontaneous action potentials (sAPs), as previously described. sAPs can be distinguished from sDeps by their greater amplitude, the presence of an after-hyperpolarization or after-depolarization and their sensitivity to the voltage-dependent Ca 2+ channel (VDCC) antagonist nifedipine (Meng et al. 2008).
In order to determine the mechanisms underlying the generation of sDeps, the contributions of purinergic and cholinergic neurotransmission were explored through the application of specific antagonists. The purinergic P2X 1 antagonist NF449 (10 μM) greatly reduced the frequency of sDeps. The frequency at which sDeps were detected was reduced from 8.0 ± 3.6 min −1 in control cells to 0.7 ± 0.3 min −1 following 1 h of superfusion with NF449 (post-NF449, 8.5 ± 8.5% of controls; P < 0.01, n = 4; Fig. 1Aa and B), with the amplitude of the remaining events reduced to 75 ± 6% of controls (P < 0.05, n = 4). Superfusion of the P2X 1 -desensitizing agonist α,β-methylene ATP J Physiol 586.23

Figure 1. Spontaneous depolarizations (sDeps) are purinergic in origin
A, comparing the effects of the superfusion of the P2X 1 antagonist NF449 (10 μM) (a), the P2X 1 -desensitizing agent α,β-methylene ATP (α,β-mATP; 1 μM) (b), the muscarinic acetylcholine receptor antagonist atropine (1 μM) (c), a blocker of Na + channels, tetrodotoxin (TTX, 1 μM) (d) and an antagonist of L-type Ca 2+ channels, nifedipine (1 μM) (e) on the membrane potential of SMCs. Note the longer time-scale of Ab. A depolarization and increase in the sAP frequency induced by α,β-mATP (Ab) is initially noted (consistent with its action as an agonist), before the membrane potential returns towards its resting level and the sAPs decline in frequency (presumably as the receptors desensitize). B, a comparison of the frequency of sDeps for each drug application to normalized control frequencies, with paired t test statistical comparisons (see text for details).
In order to determine whether blocking L-type Ca 2+ channels affected the frequency at which sDeps were detected, the effect of nifedipine was assessed. Nifedipine (1 μM) did not affect the frequency of sDeps, which was 6.6 ± 2.3 min −1 in control cells, and 6.5 ± 2.0 min −1 following nifedipine superfusion for 1 h (post-nifedipine, 100 ± 29% of controls; P = NS; Fig. 1Ae and B).

Pharmacology of spontaneous action potentials
Spontaneous action potentials (sAPs; of amplitude: 53.4 ± 2 mV) were observed in 38% of SMCs (i.e. 11 of 29, from n = 4) where they typically occurred at a low frequency (0.77 ± 0.20 min −1 , range: 0.2-2.6). Following superfusion of the P2X 1 antagonist NF449 (10 μM) for 1 h, sAPs were not observed in any SMC (i.e. 28 of 28, from n = 4). To confirm that NF449 had no direct effect on L-type Ca 2+ channels, SMC membrane properties were examined with a series of brief current injections. In control cells and those superfused with NF449 (10 μM) for 1 h, current injections ≥ +0.09 nA stimulated active currents consistent with normal opening of L-type Ca 2+ channels (see online Supplemental material, Fig. 1).

The relationship between sDeps and sAPs
To determine whether sDeps might intermittently initiate sAPs, their kinetics were compared. Both sDeps and sAPs occurred with variable rising phases, with times-to-peak ranging from 5 to 89 ms for sDeps (median: 30 ms, no. of events = 4755, e.g. Fig. 3A) and 11-49 ms for sAPs (median: 19 ms, no. of events = 369, e.g. Fig. 3B). In order to test the hypothesis that the more prolonged rise times of sDeps indicate a lower peak current after ATP binding to P2X 1 receptors, the rising phases of both sDeps and sAPs were differentiated and compared. The differentiated rising phases of sDeps usually contained one peak, which varied in amplitude with a positively skewed distribution from 0.4 to 4.5 V s −1 (Fig. 3A). By contrast, the differentiated rising phase of sAPs always contained two determinable peaks and the broad amplitude frequency distribution (0.6-6.3 V s −1 ) of the first peak is consistent with the highly variable time-to-peak of sAPs (Fig. 3B). Comparing the first peak of the differentiated rising phase of the depolarization revealed that the amplitude of the first derivative was much greater for sAPs (median: 2248 mV s −1 , range: 454-4913) than sDeps (median: 439 mV s −1 , range: 334-950) (P < 0.0001, n = 12, Wilcoxon signed rank test; Fig. 3C).
However, the peak inward current through P2X 1 receptors might provide a more sensitive measure of transmitter release than the amplitude of sDeps. To assess the inward current through P2X 1 receptors, the differentiated rising phases of only sDeps, pre-and post-superfusion with nifedipine (1 μM), were compared. The amplitude of the first derivative of the sDep increased from 237 mV s −1 in control cells (median amplitude; range: 182-359) to 268 mV s −1 (range: 196-422) following nifedipine superfusion (i.e. to 113 ± 3% of controls; Student's one-tailed paired t test, P < 0.05, n = 4) (data shown for one experiment, Fig. 4C). In control cells of these experiments, sAPs represented 12% (range: 3-21) of all depolarizations.

The relationship between spontaneous depolarizations and focal Ca 2+ transients
Loading detrusor smooth muscle with the Ca 2+ indicator Oregon Green BAPTA-1 AM revealed three types of spontaneous Ca 2+ transient (whole-cell Ca 2+ flashes, Ca 2+ waves and focal Ca 2+ transients), as reported by Meng et al. (2008). Simultaneous electrophysiology and confocal imaging of Ca 2+ indicator fluorescence was used to determine the relationship between spontaneous depolarizations (sDeps) of the membrane potential of SMCs, with spontaneous focal, transient increases in fluorescence. Focal Ca 2+ transients (Fig. 5A) were coincident with sDeps ( Fig. 5B and C).
The relationship between these two events was explored with higher temporal resolution (2 kHz), by correlating line-scans taken along the long axis of a SMC (Fig. 6A), with a simultaneous measure of the membrane potential, recorded with an intracellular electrode impaled in the imaged cell. A focal Ca 2+ transient occurred as a local event along the x-axis of the line-scan (Fig. 6B). A measure of the intensity of the Ca 2+ fluorescence for this region revealed a rapid rise in [Ca 2+ ] over a period of approximately 100 ms, and then a slow decline to baseline (Fig. 6C). For the same period of time and in the same cell, a sDep was observed that exhibits a rise-to-peak amplitude and time course of skewed amplitude distribution. B, sAPs (grey trace), which occurred with similarly varying waveforms as sDeps, reveal two peaks (arrows) when differentiated (black trace); the amplitude of the first peak of which has a broad amplitude distribution. Data shown for A and B are from a single, long-duration recording. Bin sizes for the amplitude distributions of first derivatives are 0.125 V s −1 . Note that the scale bars associated with the raw data of A and B are of differing magnitudes, and that the differentiated sAPs (B) are truncated at -1 V s −1 . C, the median amplitude of the first peak of the differentiated sAP is greater than the amplitude of the differentiated sDep. Data are median, quartiles and range (n = 12). See text for details of statistical comparison. decay that was much more rapid than the concurrent focal Ca 2+ transient (459 ms; median half-decay time: 568 ms, range 329-921; n = 3). Expanding the scale (Fig. 6D), the rise in [Ca 2+ ] starts approximately 8 ms after the initiation of the sDep. The latency of rise in [Ca 2+ ] relative to the initiation of a sDep was calculated for 12 events occurring in preparations from three animals. The median latency was 10.4 ms (range: 4.3-25.8 ms). The model fit to the Ca 2+ signal (e.g. Fig. 6D) is based on an exponentially declining calcium influx, which is then removed by a first-order process. The change in membrane potential is proportional to the number of P2X receptors open (i.e. the conductance change) and Figure 5. Spontaneous focal Ca 2+ transients are coincident with spontaneous depolarizations (sDeps) in mouse urinary bladder smooth muscle A, a region of a smooth muscle cell (SMC) in a mouse isolated urinary bladder (loaded with the Ca 2+ indicator Oregon Green 488 BAPTA-1 AM) during an intracellular recording. Images, acquired at 13.5 Hz, are 4 consecutive frames of a 200 frame series. A spontaneous focal Ca 2+ transient occurred on frame ii (arrowhead). B, intracellular recording of a period that includes the same four frames that compose A, showing simultaneous recordings of membrane potential (black line) and whole-cell fluorescence (grey dots). C, simultaneous Ca 2+ imaging and electrophysiology demonstrate the coincidence of focal Ca 2+ transients and sDeps during a longer period. J Physiol 586.23 hence to the rate of Ca 2+ influx (see Appendix). The sEJP fall time (median: 40 ms, interquartile range: 30-44 ms, calculated for 12 events which occurred in preparations from 3 animals) was similar to the estimate of the rate of Ca 2+ influx, 1/k 2 (median: 42 ms, 30-75 ms). There was no significant difference in the median time course of the sEJP and the modelled Ca 2+ influx (1/k 2 ) (P = NS, Mann-Whitney U test), although these two measures were not correlated on a pair-wise basis (Spearman rank correlation, ρ = −0.23, and P = NS).
In addition to the temporal correlation, there was also a correlation in the amplitudes of focal Ca 2+ transients and sDeps, both for large and small amplitude events (Pearson product moment correlation, ρ > 0.41 and P < 0.05 for each of n = 3 experiments; Fig. 7).
Whole-cell Ca 2+ flashes and Ca 2+ waves were also observed during simultaneous recordings, but both events occurred infrequently. Whole-cell Ca 2+ flashes were observed to be concurrent with sAPs but Ca 2+ waves did not correlate with any changes of the membrane potential of the impaled SMC.

Discussion
Purinergic neurotransmission in the detrusor muscle of the human urinary bladder is up-regulated in pathological states (Palea et al. 1993) and with ageing (Yoshida et al. 2001). The mouse is an ideal model for the study of purinergic neurotransmission due to the relative ease with which transgenic animals can be produced (e.g. Vial & Evans, 2000) and because purinergic neuro- transmission is a major component of detrusor excitation in this species (Holt et al. 1985;Meng et al. 2008). Thus, the aim of the current study was to characterize further spontaneous purinergic neurotransmission in the mouse urinary bladder, by first determining the origin of frequent 'spontaneous depolarizations' and spontaneous action potentials in smooth muscle cells (SMCs), and second to investigate their relationship to 'purinergic Ca 2+ transients' (Heppner et al. 2005) identified in rat urinary bladder smooth muscle cells.
Spontaneous depolarizations are neurogenic, purinergic spontaneous excitatory junction potentials ATP, released from nerve terminals, acts on P2X 1 receptors to cause frequent spontaneous depolarizations (sDeps) in the mouse urinary bladder. First, sDeps recorded using intracellular electrophysiology from SMCs were demonstrated to be purinergic in origin, consistent with the study of Meng et al. (2008). In the present study, sDeps were shown to be the result of ATP binding to P2X 1 receptors, as they were almost abolished by the specific P2X 1 receptor antagonist NF449 and the desensitizing agonist α,β-mATP (Fig. 1B). This conclusion is supported by the study of Vial & Evans (2000), in which ATP-induced inward currents were not present in bladder SMCs isolated from P2X 1 receptor-deficient mice. Also consistent with Meng et al. (2008) was the insensitivity of sDeps to atropine (1 μM; Fig. 1B). Second, the ATP that causes sDeps was shown to be released from nerve terminals. Following superfusion of α-latrotoxin (LTX; 1 nM), an agent which stimulates spontaneous neurotransmitter release (Ceccarelli & Hurlbut, 1980), the frequency of sDeps increased markedly ( Fig. 2A). LTX-stimulated sDeps were abolished by NF449 (Fig. 2C) but were insensitive to atropine (Fig. 2B), suggesting that LTX did not stimulate acetylcholine release that, in turn, stimulated ATP release from, for example, SMCs.
The possibility that sDeps could be generated by spontaneous transient inward currents ('STICs') seems unlikely given that the time course of sDeps (49 ± 26 ms; Meng et al. 2008) is an order of magnitude briefer than stretch-induced STICs in isolated SMCs (approx. 500 ms; Ji et al. 2002). The detection of the sDeps before the onset of local Ca 2+ transients (by about 10 ms), also argues against the Ca 2+ transient activating a local ion conductance to generate a depolarization.
It should also be noted that there is unlikely to be any influence from urothelium-derived ATP (produced, for example, in response to stretch; Burnstock, 2001) as experiments were performed on muscle strips following careful removal of the urothelium, and recordings were made from the serosal surface.
The observations that sDeps were (i) increased in frequency by LTX, (ii) unaffected by TTX, and (iii) that their time course is very different to that of STICs recorded from isolated myocytes, imply that sDeps are the result of spontaneous neurotransmitter release from nerve terminals and are therefore spontaneous excitatory junction potentials (sEJPs).

Spontaneous action potentials (sAPs) are neurogenic and purinergic
Spontaneous action potentials (sAPs) were demonstrated to have a similar origin to sEJPs, i.e. nerve terminalreleased ATP. sAPs, present in 38% of control cells sampled, were not observed in any cells following the superfusion of NF449 and were almost abolished following P2X 1 receptor desensitization with α,β-mATP. By contrast, atropine (1 μM) did not affect the frequency of sAPs. These findings suggest that sAP frequency at rest is driven by the action of ATP on P2X 1 receptors rather than by acetylcholine.
sAP frequency was not influenced by spontaneous activation of nerve branches or spontaneous nerve terminal action potentials, as the frequency of sAPs was similar in control cells (0.52 min −1 ; observed in 6 of 28 SMCs) to those superfused with TTX for 1 h (0.73 min −1 ; observed in 7 of 28 SMCs). In order to investigate whether sAPs result from nerve terminal-released ATP, experiments were performed in the presence of LTX. Superfusion of LTX (1 nM) for 10 min increased sAP frequency to 930 ± 450% of control.
During the present series of experiments, the sAPs had a broad range of rise times with a prominent after-hyperpolarization, similar to the pacemaker-type sAPs previously described (Meng et al. 2008). Very few action potentials with only an after-depolarization (previously described as spike-type) were observed; such a difference might have arisen because the shallow angle of incidence of the electrode used in present work might select more superficial cells than the more vertical electrode approach used by Meng, Young and Brading (J. S. Young, personal observation). The shallow angle of electrode approach is necessary for the simultaneous Ca 2+ -imaging experiments, so simultaneous recordings of a significant number of spike-type sAPs with Ca 2+ measurements presents an unconquered technical hurdle.

The relationship between sEJPs and sAPs
It is likely that the EJP is the basis of urinary bladder smooth muscle action potentials; weak electrical stimulation evokes only an EJP, while stronger stimulation evokes an action potential (e.g. Creed et al. 1983;Brading & Mostwin, 1989).
Both sEJPs and sAPs occur with highly variable times-to-peak, ranging from 5 to 89 ms for sEJPs and 11-49 ms for sAPs. The peak amplitude of the first-time derivative of the rising phase is proportional to the peak current into the cell (Young et al. 2007b), thus in order to test the hypothesis that sAPs arise when the P2X 1 receptor current is high, the rising phases of both sEJPs and sAPs were differentiated for comparison. For a sAP, the differentiated rising phase revealed two peaks; the first attributed to the influx of cations through the P2X receptor, and second, the influx of Ca 2+ ions through VDCCs (Fig. 3B). Thus comparing only the first peak of the differentiated rising phase of the depolarization revealed that the amplitude of the first derivative was much greater for sAPs than sEJPs. To further assess whether VDCC opening is correlated with the peak current into the cell through P2X 1 receptors, the differentiated rising phases of sEJPs were compared pre-and post-VDCC blockade. There was a significant increase in the amplitude of the first derivative following nifedipine (1 μM) superfusion. The magnitude of the increase in first-derivative amplitudes following nifedipine superfusion was, however, relatively small (13 ± 3%), but this is consistent with sAPs representing only a small proportion (12%) of the depolarizations observed.
An increase in the median first-derivative amplitude following VDCC blockade, taken together with a greater amplitude of the first derivative for sAPs than sDeps, suggests that sAPs are generated when a large-amplitude post-junctional response to spontaneous ATP release from nerve terminals depolarizes the smooth muscle membrane potential to the voltage threshold for L-type Ca 2+ channels.

Focal Ca 2+ transients are concurrent with sEJPs in mouse urinary bladder SMCs
Confocal imaging of Ca 2+ fluorescence identified, for the first time, frequent, spontaneous focal Ca 2+ transients in SMCs of the mouse urinary bladder (Fig. 5A). These events were similar in appearance and time course to purinergic Ca 2+ transients of SMCs of mouse vas deferens (Brain et al. 2002), rat mesenteric arteries (Lamont & Wier, 2002) and the rat urinary bladder (referred to there as 'localized Ca 2+ transients'; Heppner et al. 2005). To determine the relationship between the focal Ca 2+ transients of the mouse urinary bladder and sEJPs, simultaneous intracellular electrophysiology and confocal imaging of Ca 2+ fluorescence were performed. A temporal correlation between focal Ca 2+ transients and sEJPs ( Fig. 5B and C) was further examined at a high (2 kHz) resolution, which revealed a 10.4 ± 3.3 ms delay between the onset of a focal Ca 2+ transient and a sEJP recorded from the same SMC (Fig. 6D). This delay is of similar magnitude to the delay at neuroeffector junctions of the mouse vas deferens (6 ms; Brain et al. 2002) and rat urinary bladder (12-16 ms; Heppner et al. 2005). Binding of ATP to P2X 1 receptors generates an influx of Na + ions and Ca 2+ ions (Isenberg et al. 1992;Evans et al. 1996), but it is unlikely that these movements are genuinely separated by 10 ms in SMCs of the mouse urinary bladder. During the first few milliseconds, the Ca 2+ entering the cell is located close to the P2X receptors; the high local Ca 2+ concentration saturates the high-affinity Ca 2+ indicator so that the mean fluorescent signal underestimates the mean Ca 2+ concentration during the initial period of Ca 2+ entry. This non-linearity might account for at least part of the estimated delay. High endogenous Ca 2+ buffering close to the plasma membrane would also induce such an effect. It is also possible that calcium-induced calcium release (CICR) augments the Ca 2+ that enters the cell through the P2X 1 receptor upon ATP binding, causing the delay between the onsets of the sEJP and the focal Ca 2+ transient. CICR has been demonstrated to contribute to purinergic Ca 2+ transients in SMCs of the guinea-pig urinary bladder (Imaizumi et al. 1998;Hashitani et al. 2000) and mouse vas deferens (Brain et al. 2003), but plays little role in SMCs of the mesenteric arteries (Lamont & Wier, 2002) or urinary bladder (Heppner et al. 2005) of rat. The potential contribution by CICR to focal Ca 2+ transients in mouse urinary bladder SMCs awaits further experiments to inhibit CICR.

Focal Ca 2+ transients correlate in amplitude with concurrent sEJPs
In addition to temporal coupling, focal Ca 2+ transients and sEJPs were also correlated in amplitude throughout the amplitude range (Fig. 7). It has been suggested that spontaneous focal transients only arise from the release of the contents of a subpopulation of synaptic vesicles, giant dense-cored vesicles (Blair et al. 2003), rather than all vesicle types. The correlation of focal Ca 2+ transients with even low-amplitude sEJPs, also observed in another smooth muscle organ, the mouse vas deferens (Young et al. 2007a), argues against this hypothesis. As argued elsewhere (Young et al. 2007a), an amplitude correlation of focal Ca 2+ transients and sEJPs is contrary to the hypothesis that the variable amplitude of sEJPs results from the attenuation of depolarizations originating from distant release sites on neighbouring SMCs (Tomita, 1970;Purves, 1976). The hypothesis that the variable sEJP amplitudes represent the effects of ATP on only the impaled cell (Young et al. 2007a) is supported by the high input resistances of SMCs (388 ± 83 M ) in this tissue (Meng et al. 2008).
The observation that sEJP fall time (40 ms) was similar to the estimate of the rate of Ca 2+ influx, 1/k 2 (42 ms) argues that the focal Ca 2+ transients and sEJPs within a SMC are determined by a common mechanism -ion conductance through P2X 1 receptors.
As the focal Ca 2+ transients described in the present study were shown to be correlated in timing and amplitude with sEJPs, themselves abolished by the P2X 1 receptor antagonist NF449, it is likely that the focal Ca 2+ transients are purinergic in origin. It is unlikely that the focal Ca 2+ transients observed are sparks, as their time course (half-decay time, 568 ms) was similar to purinergic Ca 2+ transients of rat urinary bladder smooth muscle (112 ± 9 ms; Heppner et al. 2005) and much greater than that of sparks in urinary bladder smooth muscle (34 ± 5 ms, Heppner et al. 2005), mouse urinary bladder myocytes (69 ± 9 ms, Fritz et al. 2007) or vascular smooth muscle (range from 27 ms, Mironneau et al. 1996; to 61 ms, Jaggar et al. 1998).

Whole-cell Ca 2+ flashes correlate with sAPs
The present work confirms a correlation between whole-cell Ca 2+ flashes and sAPs. The whole-cell Ca 2+ flashes seem to represent the same type of event described as 'global Ca 2+ flashes' in rat urinary bladder smooth muscle cells (Heppner et al. 2005), as they have similar spatial and temporal characteristics, are abolished when L-type Ca 2+ channels are blocked, and decreased in frequency by P2X receptor inhibition or desensitization. In both cases these Ca 2+ transients arise as a consequence of smooth muscle action potentials, commonly initiated by the action of ATP on P2X receptors.

Conclusions
We have characterized spontaneous purinergic neurotransmission in mouse urinary bladder smooth muscle, demonstrating frequent sEJPs and sAPs that result from nerve-released ATP binding to P2X 1 receptors, sAPs occurring when the P2X 1 current is sufficient to trigger L-type Ca 2+ channel opening. sEJPs correlate in timing and amplitude with focal Ca 2+ transients, suggesting that both events result from the same quanta of ATP binding to P2X 1 receptors and providing further evidence for a variable effective post-junctional response to ATP binding in electrically well-coupled smooth muscle. The present study demonstrates that sAPs and sEJPs are triggered by the stochastic release of ATP from parasympathetic nerve terminals. J Physiol 586.23 Fitting this function to experimentally obtained line-scanning confocal microscopy data of smooth muscle Ca 2+ during a focal Ca 2+ transient allows k 2 to be calculated and compared with the known sEJP time constant.
The time at the peak Ca 2+ concentration (t max ; found when dC/dt = 0) is: The peak Ca 2+ concentration (C max ) is hence found (by substituting t max into eqn (2)) to be: So, for constants k 1 and k 2 , the peak Ca 2+ concentration (C max ) is proportional to the number of P2X 1 receptors opened at t = 0 (as this is proportional to A o ).