Selective Inhibition of Cav3.3 T-type Calcium Channels by Gαq/11-coupled Muscarinic Acetylcholine Receptors*

T-type calcium channels play critical roles in controlling neuronal excitability, including the generation of complex spiking patterns and the modulation of synaptic plasticity, although the mechanisms and extent to which T-type Ca2+ channels are modulated by G-protein-coupled receptors (GPCRs) remain largely unexplored. To examine specific interactions between T-type Ca2+ channel subtypes and muscarinic acetylcholine receptors (mAChRS), the Cav3.1 (α1G), Cav3.2 (α1H), and Cav3.3 (α) T-type Ca2+1Ichannels were co-expressed with the M1 Gαq/11-coupled mAChR. Perforated patch recordings demonstrate that activation of M1 receptors has a strong inhibitory effect on Cav3.3 T-type Ca2+ currents but either no effect or a moderate stimulating effect on Cav3.1 and Cav3.2 peak current amplitudes. This differential modulation was observed for both rat and human T-type Ca2+ channel variants. The inhibition of Cav3.3 channels by M1 receptors is reversible, use-independent, and associated with a concomitant increase in inactivation kinetics. Loss-of-function experiments with genetically encoded antagonists of Gα and Gβγ proteins and gain-of-function experiments with genetically encoded Gα subtypes indicate that M1 receptor-mediated inhibition of Cav3.3 occurs through Gαq/11. This is supported by experiments showing that activation of the M3 and M5 Gαq/11-coupled mAChRs also causes inhibition of Cav3.3 currents, although Gαi-coupled mAChRs (M2 and M4) have no effect. Examining Cav3.1-Cav3.3 chimeric channels demonstrates that two distinct regions of the Cav3.3 channel are necessary and sufficient for complete M1 receptor-mediated channel inhibition and represent novel sites not previously implicated in T-type channel modulation.

out the central and peripheral nervous systems. In thalamic reticular and relay neurons, T-type channels are involved in rhythmic rebound burst firing and spindle waves associated with slow-wave sleep (1)(2)(3)(4)(5). Studies on knock-out mice and a rat model of absence epilepsy indicate that altering T-type activity within thalamic cells can contribute to pathological conditions such as sleep disorders and epilepsy (1)(2)(3)(4)(5). Certain human epilepsies appear to be associated with T-type Ca 2ϩ channel point mutations conferring channel gain-of-function phenotype (6 -9). T-type channels also play crucial roles in dendritic integration and Ca 2ϩ spiking in hippocampal pyramidal cells (10,11). Within the olfactory bulb, T-type channels are implicated in modulating Ca 2ϩ transients and synaptic release at dendrodendritic synapses (12,13). In the periphery, antisense oligonucleotides and pharmacological approaches have implicated T-type channels in contributing to both acute and chronic nociceptive behaviors (14,15).
Previous studies have identified three main subtypes of T-type Ca 2ϩ channel ␣ 1 subunits (Cav3.1/␣ 1G , Cav3.2/␣ 1H and Cav3.3/␣ 1I ) and characterized their voltage-dependent, kinetic, and pharmacological properties (16 -21). Cav3.1 and Cav3.2 channels display "typical" T-type properties, including relatively small conductance, fast activation and inactivation kinetics, and slow deactivation kinetics, whereas Cav3.3 channels uniquely display a larger conductance, much slower activation and inactivation kinetics, as well as faster deactivation kinetics (17,19). Some of the distinct biophysical properties associated with Cav3.3 T-type currents have been observed in certain populations of native T-type currents (4,17,19,22). The biophysical differences between the T-type channels likely enable them to differentially shape and modulate firing patterns, with the more slowly inactivating Cav3.3 currents able to produce longer bursts of spikes and tonic firing patterns (17,23,24).
Although the basic properties of both cloned and native T-type channels have now been largely characterized, there remains relatively little information concerning their modulation by GPCR 2 -linked pathways. Neurotransmitters such as acetylcholine have been shown to either attenuate or stimulate low threshold Ca 2ϩ currents depending on the type of native cells examined, and sometimes multiple forms of modulation can be observed within the same cell type (25)(26)(27)(28)(29). Multiple T-type Ca 2ϩ channel subtypes are expressed in most native cells (30,31), although pharmacological tools with the specificity needed to separate these currents have not been generated. In this regard, the description of the modulation of specific T-type Ca 2ϩ channels in heterologous systems will provide insights crucial toward further investigations within native systems. This approach is also well suited for GPCR studies as most neurotransmitters activate multiple receptor subtypes in neurons.
Within thalamic reticular, hippocampal pyramidal, and olfactory granule cells, there is evidence for the expression of both T-type Ca 2ϩ channels and G␣ q/11 -coupled muscarinic acetylcholine receptors (mAChRs) (25, 30 -36). As both T-type Ca 2ϩ currents and mAChRs have been independently shown to play important physiological roles within these cell types, their functional coupling could be relevant to a number of neuronal processes. Here we studied the modulatory effects of mAChRs on the three main subtypes of low threshold T-type Ca 2ϩ channels expressed in the mammalian nervous system. We found the selective modulation of Cav3.3 Ca 2ϩ channels by G␣ q/11coupled mAChRs and combined pharmacological, genetic, and chimeric channel approaches to examine the G-protein-mediated pathway and structural regions responsible for the distinct Cav3.3 signaling characteristics.
In separate experiments, HEK 293H cells were co-transfected with M1 and wild type (WT) or chimeric human Cav3.1 or Cav3.3 channels using standard Ca 2ϩ phosphate transfection with 2 g of total cDNA/dish, 0.15 to 0.4 g of channel cDNA/dish, and 0.2 g of M1 cDNA/dish. In a subset of these experiments involving co-transfection of WT Cav3.3 and M1, either 200 M di-C8 PI(4,5)P 2 (Echelon Biosciences Inc., Salt Lake City, UT) or 50 g/ml PI(4,5)P 2 IgG 2b antibody (ϳ1:30 dilution) (Assay Designs, Ann Arbor, MI) was included in the internal solution to explore the role of PI(4,5)P 2 signaling. As the PI(4,5)P 2 antibody was supplied in a phosphate-buffered saline solution containing 10% calf serum and 0.05% sodium azide, the control Cav3.3 ϩ M1 cells were recorded in an internal solution containing a 1:30 dilution of phosphate-buffered saline with 10% fetal bovine serum and 0.05% sodium azide. Electrophysiological recordings for all experiments were performed 24 -48 h after transfection. Transiently transfected cells were selected for CD8 or pEGFP expression using either adherence of Dynabeads (Dynal, Great Neck, NY) or fluorescence of EGFP under UV light.
Electrophysiological Recordings and Analysis-Macroscopic currents were recorded using the perforated patch clamp technique to reduce current rundown and to preserve cytoplasmic signaling pathways. The external recording solution contained (in mM) 2 CaCl 2 , 1 MgCl 2 , 10 HEPES, 40 tetraethylammonium chloride, 92 CsCl, 10 glucose, pH 7.4, and the internal pipette solution contained (in mM) 120 Cs ϩ methanesulfonate, 11 EGTA, 10 HEPES, 2 MgCl 2 , 75-100 M ␤-escin, pH 7.2. For these perforated patch recordings, experimental recording did not begin until the series resistance was below 20 megohms and constant, as measured by amplifier compensation. Whole-cell recordings were used for the transiently transfected WT or chimeric human Cav3.1 and Cav3.3 channel experiments as well as the G␣ transfection experiments. The internal solution for these recordings contained (in mM) 120 Cs ϩ methanesulfonate, 11 EGTA, 10 HEPES, 2 MgCl 2 , 4 Mg-ATP, 0.3 sodium GTP. Macroscopic currents were recorded using Axopatch 200A and 200B amplifiers (Axon Instruments, Foster City, CA), controlled and monitored with Pentium 4 personal computers running pClamp software version 9 (Axon Instruments). Patch pipettes (borosilicate glass BF150-86-10; Sutter Instruments, Novato, CA) were pulled using a Sutter P-87 puller and polished with a Narishige (Tokyo, Japan) microforge, with typical resistances of 3-6 megohms when filled with internal solution. The bath was connected to the ground via a 3 M KCl agar bridge.
Data were low-pass filtered at 2 kHz using the built-in Bessel filter of the amplifier, with sampling at 10 kHz. The amplifier was also used for capacitance and series resistance compensation between 70 and 85% on every cell. Leak subtraction of capacitance and leakage current was performed on-line using a P/5 protocol or else performed with Clampfit (Axon Instruments) during off-line analysis. Figures and fittings utilized the software program Microcal Origin (version 7.5, Northampton, MA). All recordings were performed at room temperature (20 -22°C).
The voltage dependence of activation for Cav3.1, Cav3.2, and Cav3.3 currents was measured by a series of 100 -220-ms depolarizing pulses applied from a holding potential of Ϫ110 mV to membrane potentials from Ϫ80 to ϩ10 mV, increasing at 5-mV increments, with 2 s between pulses. The potential that elicited peak currents ("peak potential" ranging from Ϫ45 to Ϫ25 mV) was obtained from this protocol and used in subsequent protocols. Series resistance was also monitored with a 5-ms depolarizing pulse to Ϫ105 mV immediately before the test pulse to ensure that this variable was relatively constant, and any changes in peak current levels were not because of significant changes in series resistance. Effects of saturating concentrations of mAChR agonist (1 mM CCh) on stable T-type currents were then investigated using steps to peak potential every 5 s (0.2 Hz) from a holding potential of Ϫ110 mV. These depolarizing steps were 80 ms in duration for Cav3.1 and Cav3.2 and 200 ms in duration for Cav3.3. The Ϫ140-mV prepulse protocol for Cav3.3 included a 1-s prepulse to Ϫ140 mV to remove any accumulated channel inactivation. To quantify the percent of channel inhibition, stimulation, or washout during CCh or control solution perfusion, the peak current magnitude at equilibrium was averaged (2-5 values). When distinct effects were observed (i.e. stimulation versus no effect of M1 on Cav3.1 currents), all cells displaying a Ͼ10% modulating effect with a clear exponential time course were grouped into one group, while the rest of the cells were grouped into the "no effect" group.
Current-voltage relationships were fitted with the modified Boltzmann where V m is the test potential; V 0.5a is the half-activation potential; E rev is the extrapolated reversal potential; G max is the maximum slope conductance, and k a reflects the slope of the activation curve. Data from CCh concentration-response studies were fitted with the equation, y ϭ (( where A 1 is initial amplitude (ϭ0) and A 2 is final block value; x 0 is IC 50 (concentration causing 50% inhibition of currents), and P gives a measure of the steepness of the curve. The activation and inactivation rates during steps to peak potential were well described by single exponential curves to give act and inact values, respectively. Statistical significance was tested with Student's t tests with significance being determined at a confidence interval of p Ͻ 0.02.
Cookson (Ellisville, MO). (R p )-cAMP was obtained from BioMol International (Plymouth Meeting, PA). BAPTA-AM was obtained from Molecular Probes (Eugene, OR). Unless otherwise stated, all other drugs were obtained from Sigma. Drugs were dissolved in either distilled H 2 O or Me 2 SO, according to manufacturer's solubility data. The highest concentration of Me 2 SO in the recording solution did not exceed 0.1%, a concentration that did not detectably affect T-type properties (data not shown). Gravity-driven perfusion occurred at a rate of ϳ400 l/min, and the outputs of the manifold were placed within close proximity of the cell, resulting in the cell being bathed in new solutions with minimal delay (within 1 s).

Muscarinic M1 Receptors Selectively Inhibit Cav3.3 T-type
Ca 2ϩ Channels-To investigate the potential for T-type Ca 2ϩ channel modulation by mAChRs, we transiently transfected HEK cell lines stably expressing individual subtypes of recombinant rat brain T-type channels with the human muscarinic M1 receptor. Perforated patch recordings with ␤-escin demonstrated that activation of M1 with 1 mM CCh caused a rapid (Ͻ30 s) and robust inhibition of exogenously expressed rat brain Cav3.3 T-type channel peak currents (Ϫ45 Ϯ 2%, n ϭ 34) ( Fig. 1, C and F). Only a small subpopulation of stable Cav3.3 cells (Ͻ10%) was not affected by CCh application (likely representing cells untransfected with the M1 receptor).
Activation of M1 with 1 mM CCh had no significant effect (p Ͼ 0.05) on the voltage dependence of Cav3.3 currents but significantly increased both the rates of activation and inactivation (p Ͻ 0.001; Table 1). In contrast to the clear inhibition of Cav3.3 T-type currents, activation of M1 receptors with 1 mM CCh largely had no effect on the peak current amplitude of either rat brain Cav3.1(Ϫ2.1 Ϯ 2.0%, n ϭ 18) or Cav3.2 channels (Ϫ0.1 Ϯ 2.3%, n ϭ 17) (Fig. 1, A, B, D, and E). In a small subset of both Cav3.1 and Cav3.2 currents we noted a stimulation induced by M1 activation (Cav3.1 ϭ 35 Ϯ 12%, n ϭ 4; Cav3.2 ϭ 36 Ϯ 12%, n ϭ 5), with a slower time course to equilibrium of greater than 1 min (n ϭ 3 and n ϭ 4, respectively). For the prevalent null effect on Cav3.1 and Cav3.2 currents, 1 mM CCh application had no significant effect on channel activation and inactivation kinetics or the voltage dependence of activation (p Ͼ 0.05; Table 1). Different Cav3.3 T-type channel isoforms with distinct carboxyl termini have been identified from both the rat and human brain (17,19,21,38). To test whether inhibition of the Cav3.3 channel by M1 receptors was restricted to the rat brain short carboxyl-terminal isoform (17), we also examined the longer human Cav3.3 isoform (21) transiently co-transfected into HEK cells with the M1 receptor. Similar to that for the shorter rat brain isoform, application of 1 mM CCh resulted in significant inhibition of the human Cav3.3 peak current amplitude (Ϫ28 Ϯ 2%, n ϭ 15) and also significantly increased activation and inactivation kinetics (p Ͻ 0.001; Fig. 5A; Table 1). Additionally, similar to that for the rat Cav3.1 T-type channel, application of 1 mM CCh to HEK cells co-transfected with the human Cav3.1 channel and M1 receptor had no significant effect on peak current amplitude (Ϫ0.3 Ϯ 2.0%, n ϭ 9) or channel kinetics (for 100% cells tested; p Ͼ 0.05; Fig. 7B; Table 1). Overall, the differential modulation of T-type Ca 2ϩ channel subtypes mediated by M1 receptors was consistent across both rat and human recombinant T-type channels.
Muscarinic M1 Receptors Dose-dependently Modulate Cav3.3 Biophysical Properties-Perforated patch recordings on stable rat Cav3.3 cells transiently transfected with M1 receptors revealed that the CCh-induced inhibition of peak current levels was reversible over a time course of about 2 min (n ϭ 13; Fig. 2, A and B). As previously mentioned, activation of M1 receptors with 1 mM CCh caused a significant (p Ͻ 0.001) increase in inactivation kinetics (control, inact ϭ 86 Ϯ 6, n ϭ 27; 1 mM CCh, inact ϭ 31 Ϯ 2, n ϭ 27). Along with peak current inhibition, the CCh-induced increase in Cav3.3 inactivation kinetics was reversible (Fig. 2, A and C). Both the M1 receptor-induced inhibition of Cav3.3 peak currents and the increased inactivation rate would be predicted to reduce the total amount of Ca 2ϩ flowing through Cav3.3 T-type channels during a cellular depolarization. The effect of M1 receptor activation on total Ca 2ϩ influx was determined by integrating the area over Cav3.3 current traces during 200-ms depolarizing pulses to peak potential before and after 1 mM CCh application. Normalizing these Ca 2ϩ influx values to control levels showed a 77 Ϯ 2% (n ϭ 20) reduction in Ca 2ϩ influx mediated by M1 receptor activation (Fig. 2D).
Control experiments with mock transfections of an empty control vector or with a preincubated mAChR antagonist (atropine) demonstrated that the CCh-induced inhibition of Cav3.3 currents is mediated specifically via the transfected M1 receptor (Fig. 2E). Testing the effects of varying concentrations of CCh on stable Cav3.3 cells with transfected M1 receptors revealed that the inhibitory effect is dose-dependent (Fig. 2F). The IC 50 for inhibition of Cav3.3 currents by CCh ϭ 27 M, consistent with that reported for phosphatidylinositol hydrolysis triggered by M1 receptor activation in both HEK 293 and Chinese hamster ovary cells (39,40).

Inhibition of Cav3.3 T-type Ca 2؉ Channels by mAChRs
In contrast to the partial effect of G␤␥ signaling antagonists, co-expression of the regulator of G-protein signaling 2 (RGS2), an effector antagonist for G␣ q/11 (44), completely prevented the M1 receptor-induced inhibition of Cav3.3 currents for all cells examined. In perforated patch recordings of Cav3.3 cells cotransfected with M1 and RGS2, application of 1 mM CCh either had no effect (Fig. 3, C and D; 1% Ϯ 5%, n ϭ 7) or caused a stimulation of Cav3.3 currents (30% Ϯ 9%, n ϭ 5). RGS2 has been thoroughly characterized and shown to be a selective GTPase-activating protein for G␣ q / G␣ 11 , but not for other G␣ proteins (44), and is an effector antagonist that does not block the G␤␥-mediated inhibition of R-type Ca 2ϩ channels (45).
Constitutively Active G␣ q/11 Proteins Modulate Cav3.3 T-type Ca 2ϩ Channels-To further test whether active G␣ q/11 G-proteins are sufficient to produce inhibition of Cav3.3 currents, stable Cav3.3 cells were transiently transfected with various constitutively active G␣ subunit constructs. These constructs contain missense mutations that confer constitutive activity by reducing GTPase activity. If G␣ q/11 is the downstream signal of M1 receptor activation mediating the effects on Cav3.3 currents, then it is hypothesized that activation of the co-expressed G␣ q or G␣ 11 mutants by dialysis of GTP would cause a reduction in current amplitude and an increase in inactivation kinetics. Similar to a study that analyzed inhibition of KCNQ2/ KCNQ3 channels by G␣ q/11 (46), we used constitutively active G␣ q (G␣ q-Q209L ) and G␣ 11 (G␣ 11-Q209L ) mutants to test for the hypothesized effect and a constitutively active G␣ protein (G␣ 13-Q226L ) that does not couple to the same downstream effectors (PLC) as a negative control. We also performed controls wherein empty vectors were transfected. By comparing traces 30 s after forming the whole-cell configuration with traces 2 min after whole-cell in Fig. 4, A-D, we found that dialysis of the cell with the GTP-containing pipette internal solution caused both a significant reduction in peak current levels and an increase in inactivation kinetics only for the G␣ q-Q209L and G␣ 11-Q209L transfections. The ratio of peak current levels at 2 min divided by the peak current levels at 30 s was significantly reduced (p Ͻ 0.001) for G␣ q-Q209L (n ϭ 16) and G␣ 11-Q209L (n ϭ 15) compared with the control transfection (n ϭ 18), whereas the G␣ 13-Q226L (n ϭ 17) transfection current ratio was not significantly altered ((p Ͼ 0.05), Fig. 4E). The rates of inactivation ( inact ) were determined during depolarizing steps from Ϫ110 to Ϫ30 mV for all transfection types. The inact was significantly faster (p Ͻ 0.001) for G␣ q-Q209L and G␣ 11-Q209L compared with control transfections, whereas the inact was not significantly different for G␣ 13-Q226L (p Ͼ 0.02; Fig. 4F).
G␣ q/11 -coupled Muscarinic Receptors Selectively Inhibit Cav3.3 Channels-If inhibition of Cav3.3 T-type Ca 2ϩ channels by M1 receptors is primarily dependent on G␣ q/11 signaling, then all G␣ q /G␣ 11 -coupled mAChRs should similarly inhibit Cav3.3 currents, whereas G␣ i -coupled mAChRs should have no effect. Indeed, activation of co-expressed G␣ i -coupled M2 and M4 receptors with 1 mM CCh had no effect on Cav3.3 current amplitude (M2 ϭ Ϫ4 Ϯ 2%, n ϭ 11; M4 ϭ Ϫ4 Ϯ 3%, n ϭ 8) or kinetics (Fig. 6, A, C, E, and G; Table 1). In contrast, upon transfection of either the G␣ q /G␣ 11 -coupled M3 or M5 receptor subtypes into stable Cav3.3 cells, perforated patch recordings revealed a significant CCh-mediated inhibition (M3 ϭ Ϫ25 Ϯ 3%, n ϭ 10; M5 ϭ Ϫ31 Ϯ 3%, n ϭ 10) as well as a concomitant increase in both activation and inactivation kinetics (Fig. 6, B, D, F, and H; Table 1). Overall, experiments with genetically encoded antagonists of G␣ q/11 (RGS2) and G␤␥ (MAS-GRK3ct) and genetically encoded G␣ subtypes, as well as inhibition experiments with various mAChRs, all support the assertion that inhibition of Cav3.3 channels by mAChRs specifically occurs through G␣ q/11 . .3 cells transfected with various control or G␣ plasmids during depolarizing steps from Ϫ110 to Ϫ30 mV. Traces were obtained 30 s (black) and 2 min (gray) after the whole-cell conformation was formed, using an internal solution that contained 4 mM ATP and 0.3 mM GTP. The stable Cav3.3 cells were mock-transfected with empty plasmid (A) or transfected with the constitutively active forms (lack of GTPase activity) of G␣ proteins as follows: G␣ 13-Q226L (B), G␣ q-Q209L (C), G␣ 11-Q209L (D). E, G␣ q-Q209L and G␣ 11-Q209L cause a time-dependent reduction in Cav3.3 current magnitude. The peak current levels at 2 min were divided by the peak current levels at 30 s to determine the level of inhibition because of internal solution dialysis for the various types of transfected Cav3.3 cells, as described above. The Cav3.3 currents co-transfected with G␣ q-Q209L and G␣ 11-Q209L had a significant (p Ͻ 0.001) reduction in current ratio compared with the control transfection, whereas the G␣ 13-Q226L transfection caused no significant change (p Ͼ 0.05). F, rate of inactivation ( inact ) was determined during depolarizing steps from Ϫ110 to Ϫ30 mV for all transfection types. The inact was significantly (p Ͻ 0.001) faster for G␣ q-Q209L and G␣ 11-Q209L compared with control transfections, whereas the inact was not significantly (p Ͼ 0.02) different for G␣ 13-Q226L . All data points correspond to mean Ϯ S.E. * indicates significance at p Ͻ 0.001 compared with transfected control. (42,50). Chimeric T-type Ca 2ϩ channels between human Cav3.1 and human Cav3.3 were generated to determine the molecular regions of the Cav3.3 channel involved in the M1 receptor-mediated inhibition (Fig. 7). The Cav3.1 and Cav3.3 full-length channels were initially divided into four approximately equal portions, and chimeric channels were constructed using restriction enzyme digestion and religation (see Ref. 37). The four channel portions were named as follows: region 1 ϭ amino terminus ϩ domain I; region 2 ϭ domain I-II linker, domain II ϩ the first 39 -63 amino acids of the domain II-III linker; region 3 ϭ remainder of the domain II-III linker ϩ domain III; and region 4 ϭ the domain III-IV linker, domain IV ϩ the carboxyl terminus. Chimeric channel names were assigned based on whether the chimera contained Cav3.1 (G) or Cav3.3 (I) sequence in each of the four regions described (e.g. the chimeric Cav3.3 channel that contained region 2 from Cav3.1 is called IGII).

Two Distinct Cav3.3 Channel Regions Are Involved in M1-mediated Inhibition-Most modulation of Ca 2ϩ channels by intracellular signaling pathways involves physical interactions between various effectors and cytoplasmic channel domains
Co-expression of M1 receptors with chimeric GIII and IIGI T-type channels both showed a similar degree of M1 receptor-mediated peak current inhibition compared with that of the inhibition of the WT Cav3.3 channel (IIII) (Fig.  7, A and E). In contrast, when the IGII chimera was co-transfected with M1 receptors, application of 1 mM CCh resulted in a significantly attenuated degree of inhibition (Ϫ5.6 Ϯ 2.1%, n ϭ 11, (p Ͻ 0.001)) compared with the wild type IIII channel (Ϫ26.9 Ϯ 2.3%, n ϭ 9; Fig.  7E). Interestingly, although the chimeric IGII channels exhibited lowered M1 receptor-mediated inhibition, they still possessed significantly increased inactivation kinetics (p Ͻ 0.001; Table 2). Finally, although the IIIG chimeric channels showed similar degree of M1 receptor-mediated peak current inhibition compared with the wild type IIII, the rate of inhibition was notably slower (not shown). The changes in the rate of inhibition for IIIG and the significant decrease in the amount of inhibition for IGII suggested that both regions 2 and 4 might be involved in the M1-induced inhibition of Cav3.3 channels. To explore this, a double chimera (IGIG) was co-transfected into HEK cells with M1 receptors. Fig. 7, C and E, shows that the inhibiting effect of 1 mM CCh application on peak current amplitude was completely abolished for the IGIG chimera (0.9 Ϯ 2.5%, n ϭ 8). Activation of M1 with 1 mM CCh still caused a significant increase (p Ͻ 0.001) in the inactivation kinetics of IGIG, but the inact decreased by less than 25% for IGIG, and it decreased by 40 -65% for all the single chimeric and wild type Cav3.3 channels ( Table 2). The chimeric channel loss-of-function experiments indicate that both regions 2 and 4 are involved in the M1-mediated inhibition of current amplitude and increase in inactivation kinetics of Cav3.3 currents. In gain-of-function experiments, substitution of either region 2 (GIGG) or region 4 (GGGI) into the Cav3.1 channel resulted in 1 mM CCh-induced inhibition (GIGG ϭ Ϫ14.3 Ϯ 0.8%, n ϭ 7; GGGI ϭ Ϫ9.1 Ϯ 2.6%, n ϭ 9) that was significantly different (p Ͻ 0.001 and p Ͻ 0.02, respectively) when compared with GGGG (Ϫ0.3 Ϯ 2.2%, n ϭ 9; Fig. 7, B and E; Table 2). In contrast, inclusion of either region 1 or region 3 of Cav3.3 into Cav3.1 resulted in no significant change (p Ͼ 0.05) in M1-mediated inhibition when compared with GGGG (Fig. 7E). Although both GIGG and GGGI were inhibited by M1, the level of inhibition was significantly lower (p Ͻ 0.001) than the inhibition of IIII by M1 (Fig. 7E). When the effect of 1 mM CCh application on GIGI current amplitude was tested, M1 activation was found to produce a significant level of  5%, n ϭ 15). B, dialyzing cells with 50 g/ml PI(4,5)P 2 antibody for 10ϩ min had no effect on the inhibition of human Cav3.3 currents by M1 receptor activation (Ϫ30.8 Ϯ 3.9%, n ϭ 7). C, dialyzing cells with 200 M di-C8 PI(4,5)P 2 for 5ϩ min had no effect on the inhibition of human Cav3.3 currents by M1 receptor activation (Ϫ29.0 Ϯ 1.8%, n ϭ 6). D, bar graph showing that attenuating PI(4,5)P 2 signaling with either PI(4,5)P 2 antibodies or di-C8 PI(4,5)P 2 had no significant effect (p Ͼ 0.05) on M1-mediated inhibition compared with whole-cell recordings from control human Cav3.3 ϩ M1 cells.
GIGI inhibition (Ϫ25.1 Ϯ 2.4%, n ϭ 11; p Ͻ 0.001) compared with GGGG that was not significantly (p Ͼ 0.05) different from the inhibition of IIII (Fig. 7, A, B, D, and E). Application of 1 mM CCh also significantly increased (p Ͻ 0.001) the rate of inactivation for the GIGI Cav3.1 chimera but not for GIGG, GGGI, or the other Cav3.1 single chimeras (Table 2). Overall, the combined substitution of regions 2 and 4 from the Cav3.3 channel into the Cav3.1 channel completely restores M1-induced inhibition together with the associated increase in channel inactivation kinetics.

DISCUSSION
In this study we systematically explored the effects of activated muscarinic GPCRs on the three main T-type Ca 2ϩ channel isoforms expressed in the mammalian nervous system, and we report for the first time the differential modulation between a G-protein signaling pathway and Cav3.3 T-type Ca 2ϩ channels. Most studies on T-type Ca 2ϩ channel modulation have involved the Cav3.2 (␣ 1H ) subtype, revealing specific modulatory responses to G␤ 2 ␥, CAMKII, and redox modulation that are not observed for the Cav3.1 and Cav3.3 T-type Ca 2ϩ channel isoforms (42,51,52). The exclusive inhibition of Cav3.3 channels by G␣ q/11 -coupled mAChRs is the first example of specific GPCR-mediated modulation of a T-type Ca 2ϩ channel subtype other than for Cav3.2.
Differential Effects of mAChRs on T-type Ca 2ϩ Channel Isoforms-Examination of the literature shows that activation of mAChRs can result in multiple effects on native T-type Ca 2ϩ currents, including causing stimulation (26,27,53), inhibition (28), or having no effect (54). Given the heterogeneous nature of native low threshold Ca 2ϩ currents, without investigating interactions between specific mAChR gene products and specific T-type Ca 2ϩ channel isoforms, the published differences in modulation FIGURE 6. Inhibition of Cav3.3 currents occurs specifically through G␣ q/11 -coupled muscarinic receptors. A and C, representative perforated patch current traces during depolarizing pulses from Ϫ110 to Ϫ40 mV showing no effect on Cav3.3 currents by the G␣ i -coupled M2 and M4 receptors, respectively. Traces during control perfusion and perfusion of 1 mM CCh are indistinguishable, and CCh application had no significant effect on channel kinetics (Table 1). E and G, averaged time course of normalized peak current levels during perfusion of control recording solution (2 mM Ca 2ϩ ) followed by 1 mM CCh for Cav3.3 (ϩM2/M4) currents. Perfusion of CCh had no effect on Cav3.3 currents for M2 (E; Ϫ4 Ϯ 2%, n ϭ 11) and M4 (G; Ϫ4 Ϯ 3%, n ϭ 8) receptors. B and D, representative perforated patch current traces during depolarizing pulses from Ϫ110 to Ϫ40 mV showing inhibition of Cav3.3 currents by the G␣ q/11 -coupled M3 and M5 receptors, respectively. Activating the receptors with 1 mM CCh also significantly (p Ͻ 0.02) increased channel kinetics (Table 1). F and H, averaged time course of normalized peak current levels during perfusion of control recording solution (2 mM Ca 2ϩ ) followed by 1 mM CCh for Cav3.3 plus either M3 or M5 receptors. Perfusion of CCh caused a 25 Ϯ 3% decrease (n ϭ 10) of Cav3.3 currents with co-transfected M3 receptors (F) and a 31 Ϯ 3% decrease (n ϭ 10) of Cav3.3 currents with co-transfected M5 receptors. Only a small number of cells (n ϭ 2 for both M3 and M5) were not inhibited by CCh. All data points correspond to mean Ϯ S.E.

Inhibition of Cav3.3 T-type Ca 2؉ Channels by mAChRs
are nearly impossible to interpret. Our results using exogenous expression of cloned T-type Ca 2ϩ channels indicates that M1 receptor activation has a robust inhibitory effect on Cav3.3 currents and has either no effect or a small stimulation on both Cav3.1 and Cav3.2 currents. Similarly, experiments examining native Cav3.2 Ca 2ϩ channels in NIH3T3 cells transiently transfected with mAChRs demonstrated that M1 receptor activation had either no effect or a stimulatory effect if a PKC inhibitor was applied (27). Active G␤ 2 ␥ subunits have been shown to specifically inhibit Cav3.2 currents (42,55), and the lack of inhibition of Cav3.2 channels by M1 receptors in our study is likely because of the absence of any functional coupling between M1 receptors and G␤ 2 proteins (56). We also found that all G␣ q/11coupled mAChR subtypes (M1, M3, and M5) cause attenuation of Cav3.3 currents, whereas G␣ i -coupled M2 and M4 receptors had no effect on Cav3.3 currents. Thus it is likely that any stimulation of T-type Ca 2ϩ currents by mAChRs in native systems does not involve Cav3.3 channels. Experiments testing the effects of recombinant M2-M5 receptors on the Cav3.2 and Cav3.1 Ca 2ϩ channel isoforms in a heterologous system are required to further facilitate the possibility of interactions between these T-type channels and mAChRs.
Functional Effects of M1 Receptor Activation on Cav3.3 Currents-Activation of M1 receptors dramatically altered Cav3.3 currents by both reversibly attenuating peak current levels and increasing the rate of inactivation, resulting in a significant reduction in the influx of Ca 2ϩ . The relationship between these effects was explored using both structural channel chimeras and classical gating property studies. In chimeric studies (see below), the activation of M1 receptors primarily caused an increase in inactivation kinetics of the IGII chimera and, conversely, primarily a decrease in peak current levels for the GIGG chimera. Both this isolation of the two specific M1 receptor-mediated effects and the gating results discussed below suggest that the effects of M1 on current amplitude and inactivation kinetics are complementary but distinct phenomena. For gating studies, reduction of Cav3.3 current magnitude by M1 receptor activation was equally robust when the Cav3.3 channels were held in various states including: 1) during a prolonged hyperpolarization with no test depolarizations (chan-nels mostly in closed state); 2) after a strong hyperpolarizing prepulse to Ϫ140 mV; and 3) during 200-ms test depolarizations to peak potential at 0.2 and 0.5 Hz. Combining this lack of use dependence with the observed reduction in peak current amplitude and the increase in activation and inactivation kinetics indicates that all states of the Cav3.3 channel are subject to modulation by M1 receptor activation. The acceleration of Cav3.3 channel kinetics by M1 receptor activation also supports the hypothesis that modulation affects channel biophysical properties and not channel density via internalization, which has recently been shown to occur for the voltage-independent, GPCR-mediated inhibition of N-type Ca 2ϩ channels on a relatively fast time scale (57). Physiologically, the combined decrease in Cav3.3 peak currents and the increased activation and inactivation kinetics would be predicted to alter neuronal firing patterns and perhaps eliminate rhythmic oscillations (23,58). In support of this notion, the concomitant reduction in peak current and increase in inactivation kinetics of Cav3.3 currents triggered by anandamide have been shown to completely eliminate the sustained, rhythmic Cav3.3 current during an action potential voltage clamp experiment with an oscillating thalamic waveform (59).
Signal Transduction Pathway of M1 Receptor-mediated Cav3.3 Inhibition-Use of genetically encoded antagonists of G␤␥ (MAS-GRK3ct and G␣ t ) and G␣ q/11 (RGS2) demonstrated that G␤␥ may partially contribute to the M1-mediated inhibition of Cav3.3 currents, whereas G␣ q/11 is absolutely required for complete inhibition. The potential involvement of both G␣ q/11 and G␤␥ in a nonclassical, voltage-independent mechanism of Ca 2ϩ channel inhibition by mAChRs has been described previously for HVA Ca 2ϩ channels. In rat superior cervical ganglion sympathetic neurons, application of a muscarinic agonist causes the voltage-independent inhibition of endogenous N-type Ca 2ϩ channels that is abolished by co-expression of RGS2, G␣ t , or MAS-GRK3ct and exhibits a time course similar to the Cav3.3 inhibition reported here (41). As G␤␥ is a cofactor for PLC␤ activity, a possible explanation is that sequestering G␤␥ reduces PLC␤ activity (60). Although G␤␥ may potentiate the inhibitory effect of M1 receptor activation, transfection of constitutively active G␣ q/11 mutants into Middle, effect of activating M1 receptors with 1 mM CCh on the normalized peak current levels of chimeric channel types shown to the left. Inclusion of Cav3.1 sequence at regions II and IV (C) eliminated M1-mediated inhibition and attenuated the effect on inactivation kinetics, whereas inclusion of Cav3.3 sequence at regions II and IV (D) restored M1-mediated inhibition to a level that was not significantly (p Ͼ 0.05) different from IIII inhibition levels (see Table 1). Right, insets include chimeric whole-cell current traces during depolarizing pulses from Ϫ110 mV to peak potential before (line arrow) and after (block arrow) application of 1 mM CCh. Traces are representative of the various chimeras in terms of activation and inactivation kinetics as well as magnitude of inhibition. For inset scale bars, x ϭ 50 ms and y ϭ 100 pA. E, histogram where GIII, IGII, IIGI, IIIG, and IGIG inhibition values were statistically compared with the IIII control, and IGGG, GIGG, GGIG, and GGGI values were compared with the GGGG control, and the GIGI value was compared with both the IIII and GGGG controls. * indicates significant difference (Ͻ0.02) compared with IIII inhibition, and ** indicates a significant difference (p Ͻ 0.02) compared with GGGG modulation levels. All data points correspond to mean Ϯ S.E. IIII  IGII  IIIG  IGIG  GGGG  GIGG  GGGI  GIGI Control Ϫ inact (ms) 117 Ϯ 6, n ϭ 10 86 Ϯ 7, n ϭ 11 103 Ϯ 11, n ϭ 10 62 Ϯ 2, n ϭ 8 14Ϯ 1, n ϭ 9 20 Ϯ 1, n ϭ 9 38 Ϯ 3, n ϭ 8 58 Ϯ 2, n ϭ 10 1 mM CCh Ϫ inact (ms) 41 Ϯ 5, n ϭ 10 a 43 Ϯ 3, n ϭ 11 a 58 Ϯ 4, n ϭ 10 b 47 Ϯ 3, n ϭ 8 b 13 Ϯ 1, n ϭ 9 17 Ϯ 2, n ϭ 7 32 Ϯ 2, n ϭ 8 38 Ϯ 1, n ϭ 10 a a p Ͻ 0.001. b p Ͻ 0.02.

TABLE 2 Effects of M1 receptor activation on chimeric T-type channel inactivation kinetics
stable Cav3.3 cells demonstrated that active G␣ q/11 alone is sufficient to induce the inhibition of Cav3.3 currents. In support of this notion, only G␣ q/11 -coupled mAChRs (M1, M3, and M5) inhibited Cav3.3 currents, whereas G␣ i -coupled M2 and M4 receptors that activate G␤␥ signaling have no effect on Cav3.3 currents. Unlike that reported for the attenuation of Cav3.2 channels by G␤ 2 ␥, this novel form of T-type Ca 2ϩ channel inhibition involves the G␣ q/11 subunit and also affects channel kinetics. This inhibitory mechanism for the Cav3.3 T-type isoform may be applicable to all G␣ q/11 -coupled receptors as we have also found a similar inhibition of Cav3.3 channels by mGluR 1a receptors (61). Pharmacological antagonists eliminated the potential involvement of various intracellular signals downstream of G␣ q/11 activation that may be involved in the inhibition of Cav3.3 by M1 receptor activation. Abolishing the activity of PKC, serine/threonine kinases (including cAMP-dependent protein kinase), tyrosine kinases, phosphatases, phosphoinositide 3-kinases, and intracellular Ca 2ϩ signaling all had no effect on inhibition. This profile of M1/G␣ q/11 -mediated Ca 2ϩ channel inhibition resistant to common antagonists of cytoplasmic signaling is not unique and has been reported for the inhibition of L-type channels by G␣ q -coupled M1/3/5 receptors in HEK cells (62). Like the inhibition of Cav3.3 via M1 receptors, this inhibition is voltage-independent, relatively slow kinetically ( on ϭ 13 s), and insensitive to antagonists of protein kinases and protein phosphatases (62).
A more recent explanation for the G␣ q/11 -mediated inhibition of ion channels, including voltage-gated K ϩ channels and HVA Ca 2ϩ channels, has emerged wherein channel activity is suppressed through the depletion of membrane PI(4,5)P 2 levels via PLC activity (46,63,64). In these studies, G␣ q/11 -mediated inhibition was shown to be inhibited via dialysis of synthetic PI(4,5)P 2 or a PI(4,5)P 2 -specific antibody into the cytoplasm. In our experiments, adding di-C8 PI(4,5)P 2 or the PI(4,5)P 2 antibody into the internal pipette solution and dialyzing cells for up to 25 min had no significant effect on M1-mediated inhibition of Cav3.3 channels, suggestive of another to-be-defined mechanism whereby G␣ q/11 signaling causes the inhibition of voltage-gated ion channels. Further biophysical and biochemical experiments are required to clarify the nature of the intracellular messengers and/or scaffolding proteins that can modulate Cav3.3 T-type Ca 2ϩ channels and also whether G␣ q/11 can interact directly with the channel through a novel mechanism.
G␣ q/11 -mediated Inhibition of Cav3.3 Involves Two Discrete Channel Regions-Replacing both regions 2 and 4 in the Cav3.3 channel with the corresponding Cav3.1 T-type Ca 2ϩ channel sequences abrogated both the M1 receptor-mediated peak current inhibition and concomitant increase in inactivation kinetics. Conversely, substituting regions 2 and 4 from Cav3.3 into Cav3.1 conferred M1 receptor-mediated inhibition and increased inactivation kinetics. These data suggest that regions 2 and 4 of the Cav3.3 channel are both necessary and sufficient to recapitulate M1 receptor-mediated channel modulation. Region 2 of the Cav3.1 and Cav3.3 sequence contains the highly divergent domain I-II linker, the highly conserved domain II, and 39 -63 amino acids of the domain II-III linker, and region 4 contains most of the III-IV linker, the highly conserved domain IV, and the highly divergent carboxyl terminus. Based on their putatively intracellular regions and their high divergence between the two T-type isoforms, the I-II linker, proximal region of the II-III linker, the III-IV linker, and the carboxyl terminus are all candidates for modulation sites within regions 2 and 4. Interestingly, the only identified sites of alternative splicing within the rat and human Cav3.3 channel occur both in the I-II linker and the carboxyl-terminal regions (38,65). The effects of these splicing variations on the biophysical properties (activation kinetics) of the human Cav3.3 channel are interdependent rather than additive, suggesting a possible direct interaction between the I-II linker and carboxyl terminus that affects channel kinetics (24). Both the human and rat Cav3.3 channels inhibited by M1 receptor activation in our study lack exon 9 located in the I-II linker, whereas both the rat and human Cav3.1 channels have a 10-amino acid insertion in this region in a manner similar to that for the ϩexon 9 Cav3.3 splice variant. Thus, several observations suggest that the I-II linker may be a target region in the inhibition of Cav3.3 by M1, and some evidence points to a possible role for the carboxyl terminus. However, as multiple structural determinants contribute to the slow inactivation kinetics of Cav3.3 compared with Cav3.1 in a nearly additive manner (37), and M1 activation dramatically speeds up Cav3.3 inactivation kinetics, it is also possible that multiple intracellular loci within regions 2 and 4 of the Cav3.3 channel may be involved in the M1-mediated effect.
In summary, we find that activation of known G␣ q/11 -coupled mAChRs results in the selective inhibition of Cav3.3 T-type Ca 2ϩ currents with a concomitant increase in inactivation kinetics. The G␣ q/11 -mediated signaling pathway appears to be mediated via two disparate regions of the Cav3.3 channel. Functional interactions between mAChRs and Cav3.3 Ca 2ϩ channels could potentially impact firing patterns of various cell types, including thalamic nRT cells. Biophysical and pharmacological evidence suggests that primarily Cav3.3 channels compose dendritic T-type currents in nRT cells (22), whereas immunostaining suggests the presence of M3 receptors in these cells (33). This raises the possibility that the inhibition of Cav3.3 T-type Ca 2ϩ channels by M3 receptors in the dendrites of nRT cells could be involved in cholinergic modulation of thalamic firing patterns.