Our present results call for a review of the existing hypotheses explaining the involvement of M2 and M3 muscarinic receptors in mICAT generation in visceral smooth muscles. When PI-PLC was inhibited by U-73122 in the presence of weak [Ca2+]i buffering, mICAT inhibition was accompanied by a substantial reduction of [Ca2+]i (e.g., Figure 3a), consistent with the classical views that the M3/Gq/11/PLC link is only important in raising [Ca2+]i, which in turn potentiates mICAT due to the strong Ca2+ dependence of these cationic channels (for a review, see Bolton et al., 1999). One of the reported nonspecific actions of relatively high concentrations of U-73122 (e.g., 5 μM) is to increase [Ca2+]i and InsP3 levels (Smallridge et al., 1992). Indeed, occasionally, [Ca2+]i increased slightly following U-73122 application (e.g., Figures 2d and 3b; probably nonspecific action, see above), but mICAT was still invariably suppressed. Thus, these processes are not necessarily related. These observations prompted us to perform the tests at strongly buffered [Ca2+]i, whereby any Ca2+-dependent changes of the cationic channel gating could be prevented. Under these conditions, U-73122 also abolished mICAT (Figures 4, 5, 8), thus suggesting that a link via PLC upstream of IICR is crucially important in opening cationic channels.
Recently, it was found in cardiac myocytes that IK,ACh could be inhibited by U-73122 irrespective of G-protein activation, for example, channel activity induced by Na+ ions was also suppressed (Cho et al., 2001). As there was no evidence of a direct channel block, the authors postulated the interference with PIP2-channel interaction as the most likely mechanism. The role of Gβγ and Na+ in stabilizing the interactions of PIP2 with the KACh channel to generate IK,ACh is well established, but at present no data are available for a similar mechanism of cationic channel opening, and thus the role of PIP2 in mediating the inhibitory effect of U-73122 on mICAT remains to be established. Notably, the inhibition of KACh channel was voltage-independent and the IC50 values for both blockers were similar and significantly lower (0.12–0.16 μM; Cho et al., 2001) than the values reported for PLC inhibition by U-73122 (1–4 μM;Bleasdale et al., 1990; Smith et al., 1990; Thompson et al., 1991). The IC50 value for mICAT inhibition (0.5–0.6 μM; Figure 4d) is closer to that expected to produce a significant inhibition of PI-PLC. Moreover, it was nearly the same for both carbachol- and GTPγS-evoked currents.
Identical changes of mICAT biophysical properties (Figure 5b, d), closely similar kinetics (Figure 4a, b) and concentration dependence of the inhibitory action of U-73122 (Figure 4d) on both carbachol- and GTPγS-evoked currents are several lines of evidence, which taken together strongly suggest that the inhibitor blocks both currents via the same process. Since its inactive counterpart, U-73343, did not affect GTPγS-induced current, the site of U-73122 action is likely to be PLC, and since GTPγS activates G-proteins directly, our findings are in agreement with the postulated site of U-73122 action at the interface of the G-protein/PLC system (Bleasdale et al., 1990; Thompson et al., 1991; Yule & Williams, 1992). Moreover, the specificity of the action of U-73122 was further confirmed by the findings that inhibition of the muscarinic receptors by an M3-selective antagonist p-F-HHSiD at low, 10−7M, concentration mimicked its action (Figure 9a).
If U-73122 acted simply as a channel blocker, its action would have to be strongly voltage-dependent, increasing with membrane hyperpolarization (Figure 5) in a manner indistinguishable from the intrinsic voltage dependence of the cationic conductance (since the Boltzman slope of the activation curve remained unchanged when mICAT was suppressed). Also, such presumed channel blockade should have the properties similar to those of mICAT desensitization (Zholos & Bolton, 1996), but exactly opposite to those seen when agonist concentration is increased (Zholos & Bolton, 1994) or GTPγS acts to activate mICAT (Figure 5a). Thus, we consider that a channel block by U-73122 is a highly unlikely possibility. Nevertheless, mICAT inhibition by U-73343 raised some doubts and therefore further tests were performed, which showed that it could act as a competitive antagonist of the muscarinic receptors (Figure 6b). This effect must be added to the growing list of other actions of this drug. This action seems a general one, since in the mouse urinary bladder carbachol-evoked contraction was also competitively antagonized by U-73343 (Dr Jörg Wegener, Institut für Pharmakologie und Toxikologie, München, Germany; personal communication).
There is compelling evidence of the importance of the M2-Gi/Go link in mICAT generation. Previous studies have shown that Pertussis toxin treatment abolishes this current in ileal and gastric cells (Inoue & Isenberg, 1990a; Komori et al., 1992; Pucovsky et al., 1998; Rhee et al., 2000). These results suggested that the M2 receptor subtype was involved, as it couples preferentially to Gi/Go proteins, and this coupling is selectively blocked by Pertussis toxin. Later on, using antibodies, the G proteins involved in the signal transduction were pinpointed as Go in gastric and ileal cells (Kim et al., 1998a; Yan et al., 2003; in the latter study, the role of Gβγ subunit was also excluded) or both Gi and Go in tracheal myocytes (Wang et al., 1997). Parallel pharmacological studies have consistently established the importance of the M2 receptor subtype (Bolton & Zholos, 1997; Kang et al., 1997; Wang et al., 1997; Zholos & Bolton, 1997; Komori et al., 1998).
However, in recent years, it also became increasingly evident that the M3 subtype stimulation is also very important for mICAT generation. Although this is agreed in general, the underlying molecular mechanisms remain unknown, while the particulars of the M3-mediated effect on mICAT differ. Thus, in tracheal cells, stimulation of M2 receptors alone was insufficient to activate mICAT, unless there was a parallel rise of [Ca2+]i associated with the M3 receptor activation (Wang et al., 1997). In terms of the underlying mechanisms, such M3 receptor involvement has the most straightforward interpretation: Ca2+ release via the M3/Gq/11/PLC/InsP3 pathway potentiates cationic channel opening, which is a prominent feature of these channels (Inoue & Isenberg, 1990b; Pacaud & Bolton, 1991). In guinea-pig gastric cells, when Ca2+ was present in the external solution and [Ca2+]i was not buffered, inhibition of the M3-subtype reduced the sensitivity of the M2-mediated response to the agonist, without affecting the maximal response (Rhee et al., 2000). Finally, and perhaps most strikingly, when [Ca2+]i was ‘clamped’ at 100 nM in the absence of external Ca2+ (that is under conditions when any change in [Ca2+]i was prevented), the M3-subtype inhibition still strongly suppressed the maximal response without affecting significantly the sensitivity to the agonist (Bolton & Zholos, 1997; Zholos & Bolton, 1997) and, under the same conditions, significant correlation of muscarinic agonist potencies between the M2/mICAT and M3/PLC/InsP3 systems was found (Okamoto et al., 2002).
Thus, based on these observations, we have previously proposed that some ‘permissive’ mechanism is associated with the M3-subtype stimulation, which operates independently of [Ca2+]i elevation. It is interesting that, resembling the nonsurmountable action of U-73122 (Figure 6a), we have earlier found that M3 receptor antagonists also acted to suppress the mICAT response noncompetitively. It should be noted that M3-selective antagonists even at high concentrations produced less than complete inhibition of mICAT, while PLC inhibition abolished the response. Also, we have recently found that Gαq/Gα11 antibodies, which are expected to inhibit the M3/PLC system (though complete inhibition using the antibodies is difficult to achieve), did not affect mICAT in ileal myocytes (Yan et al., 2003). These apparently contradictory results can be explained on the assumption that the G-proteins associated with the M2 subtype can also couple to PI-PLC, as shown in many tissues including longitudinal muscle from the small intestine of the guinea-pig, where Pertussis toxin treatment reduced inositol phospholipid hydrolysis by about 50% (Prestwich & Bolton, 1995). Thus, crosslinkage of the M2 and M3 receptor subtypes to the same G-proteins may explain the above contradictions; when PLC is inhibited by U—73122, nearly complete inhibition of mICAT results.
Our present results considerably extend the knowledge of the processes involved in this permissive action of M3 receptor activation. We tested several plausible mechanisms such as Ca2+ store depletion, as well as production of InsP3 and DAG as possible intermediates. However, none of these was found relevant, and thus the signal transduction is apparently limited to the upstream events such as PLC activation itself, or perhaps PIP2 hydrolysis on its own, irrespective of InsP3 or DAG accumulation. Interestingly, ATP inhibition of M current in sympathetic neurones is similarly PLC-dependent, but does not involve InsP3 or two important targets of diacylglycerol, PKC and Ras (Stemkowski et al., 2002). Notably, Ca2+ store depletion and OAG induced only tiny inward currents in ileal cells, which were about two orders of magnitude smaller than mICAT at maximally effective agonist concentration. In tracheal myocytes after Ca2+ store depletion, methacholine failed to activate the cationic current (Wang et al., 1997), while OAG did not induce any current and did not affect histamine- or methacholine-activated currents (Wang & Kotlikoff, 2000). In contrast, in portal vein myocytes, noradrenaline-evoked cationic current could be directly activated by OAG (Helliwell & Large, 1997; Albert & Large, 2001). The amplitude of this OAG-induced current in ileal and portal vein myocytes is similar, but since the maximal response to the agonist in ileal cells is much larger, this pathway cannot make any significant contribution to mICAT generation.
One of the most interesting and significant results of this study is the observation that PI-PLC inhibition can reverse the action of GTPγS in such a manner that the I–V relationships during mICAT development and during PLC inhibition were closely similar (Figure 5a). Formally, this could be interpreted as a voltage-dependent effect of the PLC blocker, since the inhibition at negative potentials began earlier and at lower U-73122 concentrations compared to the positive range (Figures 5a, c). However, in our previous studies, we established two principal mechanisms of mICAT activation: an increase in the maximal cationic conductance reflecting the increasing number of active channels and a negative shift of the activation curve, indicative of the changes in the gating properties of the activated channels (Zholos & Bolton, 1994). At that time, nothing was known about the convergence of several signalling pathways to open this channel. Thus, assuming that an activated G-protein directly gates this channel, such effects were explained by possible binding of several activated G-proteins to the same channel. Analysing the I–V relationships in a similar manner, we found that the process, which could appear as a voltage-dependent inhibition by U-73122, was in fact a positive shift of the cationic conductance activation curve (Figure 5b, d). These observations provide the first insight into the molecular mechanisms underlying the modulation of the voltage range of mICAT activation. Similar observations were made using an M3-selective antagonist p-F-HHSiD (Figure 9). It thus became evident that PLC activation not only determines the size of the response but also modulates the voltage sensitivity of cationic channels, and through this process it can strongly modulate the kinetics of mICAT (Figure 9).
The mechanisms of the interactions between PLC and cationic channel remain unclear. If indeed this cationic channel is a homo- or a heteromultimer of TRP proteins (Walker et al., 2001), then the problem of identifying the mechanism(s) of its opening has a much wider significance, as the mechanisms of activation and functions of TRP proteins remain generally unclear (Clapham et al., 2001). Interestingly, in the Drosophila phototransduction cascade, signalling depends on key elements which include TRP channel protein, PLCβ, PKC and calmodulin (Hardie & Raghu, 2001), and the same elements are important for mICAT generation is smooth muscles (Kim et al., 1995; 1998b; Ahn et al., 1997 and present study). It now appears that the signal transduction underlying mICAT is very complex, as signals from two different muscarinic receptor subtypes converge on the same channel protein, and each receptor subtype has a permissive role.