Pacemaker channels in mouse thalamocortical neurones are regulated by distinct pathways of cAMP synthesis

Abstract

A crucial aspect of pacemaker current (I_h) function is the regulation by cyclic nucleotides. To assess the endogenous mechanisms controlling cAMP levels in the vicinity of pacemaker channels, I_h regulation by G-protein-coupled neurotransmitter receptors was studied in mouse thalamocortical neurones. Activation of β-adrenergic receptors with (−)-isoproterenol (Iso) led to a small steady enhancement of I_h amplitude, whereas activation of GABA_B receptors with (±)-Baclofen (Bac) reduced I_h, consistent with an up- and down-regulation of basal cAMP levels, respectively. In contrast, a transient (τ_decay, ∼200 s), supralinear up-regulation of I_h was observed upon coapplication of Iso and Bac that was larger than that observed with Iso alone. This up-regulation appeared to involve a cAMP synthesis pathway distinct from that recruited by Iso, as it was associated with a reversible acceleration in I_h activation kinetics and an occlusion of modulation by photolytically released cAMP, yet showed an 11 mV as opposed to a 6 mV positive shift in the activation curve and an at least seven-fold increase in duration. GABA, in the presence of the GABA_A antagonist picrotoxin, mimicked, whereas N-ethylmaleimide, an inhibitor of G_i-proteins, blocked the up-regulation, supporting a requirement for GABA_B receptor activation in the potentiation. Activation of synaptic GABA_B responses via stimulation of inhibitory afferents from the nucleus reticularis potentiated Iso-induced increments in I_h, suggesting that synaptically located receptors couple positively to cAMP synthesis induced by β-adrenergic receptors. These findings indicate that distinct pathways of cAMP synthesis target the pacemaker current and the recruitment of these may be controlled by GABAergic activity within thalamic networks.

The autonomous beating of the heart and a considerable number of rhythmic activities in the brain are controlled by the hyperpolarization-activated cation currents I_h, also termed pacemaker currents (for review, see Pape, 1996; Lüthi & McCormick, 1998; Santoro & Tibbs, 1999; Robinson & Siegelbaum, 2003). Pacemaker channels are gated upon hyperpolarization and generate a depolarizing drive back towards threshold, thereby facilitating the next rhythmic firing episode. Pacemaker currents are enhanced when intracellular concentrations of cAMP are increased (for review, see Santoro & Tibbs, 1999; Kaupp & Seifert, 2001; Robinson & Siegelbaum, 2003). The cAMP-mediated augmentation of cardiac I_h is crucial for the accelerating effects of sympathetic activity on the heartbeat (Brown et al. 1979). Moreover, cAMP-mediated enhancement of I_h in the brain contributes to control the slow periodicities in neuronal network activities related to sleep and epilepsy (Bal & McCormick, 1996; Lüthi & McCormick, 1999).

Pacemaker channels are composed of subunits from the family of the hyperpolarization-activated cation non-selective (HCN) channels. In addition to a voltage sensor contained within the six transmembrane segments, HCN channels possess a C-terminal cyclic nucleotide-binding domain (for review, see Santoro & Tibbs, 1999; Kaupp & Seifert, 2001; Wainger et al. 2001), which has a high selectivity for cAMP (Kaupp & Seifert, 2001). This modular structure provides the molecular basis for the dual gating of HCN channels by both voltage and cyclic nucleotides. Direct cAMP-dependent modulation constitutes a major regulatory pathway of native pacemaker channels as well, as evident from the similarity in the concentration-dependence and kinetics of cAMP-modulation to cloned channels (DiFrancesco & Tortora, 1991; Ludwig et al. 1998; Lüthi & McCormick, 1999; Seifert et al. 1999).

In spite of this important role of cAMP in the direct regulation of pacemaker channel function, little is known about the strength and the type of cAMP signals generated in the vicinity of pacemaker channels in intact cells. Here, we have addressed the diversity of cAMP signalling by studying the regulation of I_h in mouse thalamocortical (TC) cells, which shows a high sensitivity to cAMP (Lüthi & McCormick, 1999; Seifert et al. 1999). In TC neurones, multiple neurotransmitter receptor systems coupled both positively and negatively to cAMP synthesis (via G_s- and G_i-proteins, respectively) control the electrophysiological activities related to sleep and arousal (for review, see McCormick & Bal, 1997). We find that I_h is steadily up- and down-regulated by neurotransmitter receptors coupled positively or negatively to cAMP synthesis, as previously described (for review, see Pape, 1996). However, the most vigorous modulation of I_h is observed upon coactivation of G_s- and G_i/o-coupled receptors, which produces a supralinear cAMP signal. These data demonstrate that, in mouse TC neurones, I_h is modulated by several cAMP signals differing both in strength and time course. This suggests that the pacemaker channels may be surrounded by distinct cAMP synthesis pathways, perhaps incorporating distinct adenylyl cyclases (ACs), that are recruited according to the type and timing of neurotransmitter stimuli.

Methods

Slice preparation

Mice of either sex between 16 and 21 days were anaesthetized by intraperitoneal injection of 90 mg kg⁻¹ ketamine and 21 mg kg⁻¹ xylazine and decapitated according to the Guidelines of the Veterinary Institute of the Canton Basel-Stadt. Coronal slices containing the dorsal lateral geniculate nucleus and the ventrobasal nucleus were prepared on a vibratome (VT1000S, Leica, Germany) in an ice-cold oxygenated solution containing (in mM): 63 NaCl, 107 sucrose, 2.5 KCl, 1.25 NaH₂PO₄, 26 NaHCO₃, 0.5 CaCl₂, 7 MgCl₂, 18 dextrose. The slices were allowed to recover for 5 min in a home-made interface-type chamber at 35.0 °C in the cutting solution, before being transferred to a sucrose-free solution containing 126 mM NaCl instead. After an additional 30 min, slices were incubated at room temperature for 1–2 h before recordings commenced.

Electrophysiological recordings

Whole-cell voltage-clamp recordings were obtained from visually identified TC neurones in the dorsal lateral geniculate and the ventrobasal nucleus of the thalamus (BX51WI microscope, Olympus, Germany) at 33.5–35.0 °C. No difference in cAMP-dependent modulation of I_h was found for neurones in these two nuclei and the data were pooled. The bath solution contained (in mM): 126 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 26 NaHCO₃, 1.5 CaCl₂, 2 MgCl₂, 1.5 BaCl₂, 18 dextrose, 1.7 L(+)ascorbic acid. Unless stated otherwise, 1.5 mM Ba²⁺ ions were present to prevent activation of K⁺ currents by (±)-Baclofen (Bac). In some experiments, tetrodotoxin (0.5 μM) was included to reduce spontaneous synaptic activity. Patch pipettes (2.5–3.5 MΩ, WPI, Sarasota, FL) were filled with (in mM): 130 KGluconate, 10 HEPES, 10 KCl, 2 K₂-ATP, 0.2 Na-GTP, 10 Phosphocreatine, 2 MgCl₂, pH 7.25, 290 mOsm. GTP was freshly added daily from 100-fold concentrated stock solutions. This solution yielded a liquid junction potential of 12 mV that was taken into account for all voltages. Recordings yielded series resistances between 5 and 15 MΩ that were electronically compensated by 40–70% and the capacitive transient monitored in parallel with the I_h responses. When the series resistance was <9 MΩ, no compensation was applied, but the stability of the capacitive transient was checked before each voltage step. Data were discarded if the capacitive transient changed by >20% of the original amplitude.

The h-current was activated in whole-cell voltage-clamp mode by applying 5 s hyperpolarizing voltage commands from a holding potential of −62 mV to a test potential of −92 mV, the half-maximal activation voltage, at interstimulus intervals of 12 s. Recordings were selected when current amplitudes reached levels of at least −100 pA. This voltage protocol ensured that, on the one hand, >85% of the steady-state current amplitude at this potential could be activated and the current deactivated completely upon return to −62 mV. On the other hand, current amplitudes could be measured frequently enough to monitor the temporal development of current modulation. To follow the time course of transient up-regulation of I_h induced by focal application of neurotransmitter agonists or by flash photolysis of caged cAMP, I_h activation was limited to 2.5 s.

Bath application of neurotransmitter receptor agonists was limited to one per experiment due to the incomplete wash-out of the effects on I_h. (−)-Isoproterenol (Iso) and 8-bromo-cAMP (8Br-cAMP) solutions were prepared fresh daily from frozen stocks. Drugs were applied in the bath (4 ml min⁻¹) or via pressure application through a patch pipette placed in the vicinity of the cell recorded from. As the efficacy of the large number of GABA_B antagonists on Bac-induced cAMP formation is characterized incompletely (Cunningham & Enna, 1996; Knight & Bowery, 1996), the action of antagonists on Bac effects or on those induced by synaptically activated GABA_B receptors on the Iso-induced potentiation of I_h was not systematically evaluated. Data were collected through an Axopatch200B amplifier (Axon Instruments, Foster City, CA), digitized at 1 kHz, and analysed off-line using PClamp8.0 software. Monoexponential time constants were analysed by fitting the first 1.5 s of the I_h transient elicited at −92 mV (V_0.5) using the Chebychev fitting routine, while leaving away the initial lag in the onset of activation (∼120 ms). Activation curves (Figs 1 and 3) were fitted to the Boltzmann function, with I/I_max= (1 + exp [(V−V_0.5)/s])⁻¹, with V_0.5 the voltage for half-maximal activation, and s the slope factor. Normalization of tail current amplitudes was always done with respect to the maximal tail current observed under control conditions. Origin software (Version 4.1) was used for the fits to the data presented in Fig. 5. Data are presented as mean ±s.e.m. Paired or unpaired t-tests as appropriate were used for statistical analysis and a value of P < 0.05 was considered statistically significant.

Figure 1. Iso and Bac modulate I_h in a manner consistent with the coupling of β-adrenergic and GABA_B receptors to adenylyl cyclase.

A, bath application of Iso (500 nM) induced a small steady enhancement of I_h amplitude to 128.0 ± 4.3% of control (n= 6, P < 0.001). Inset shows an overlay of I_h activated during a voltage step from −62 to −92 mV in control and during Iso application at steady-state. B, activation curve of I_h in the absence (○) and in the presence (•) of Iso. Activation curves were constructed from tail current analysis and normalized to the maximal current under control conditions. This yielded V_0.5=−94.4 ± 0.6 mV in control and V_0.5=−87.6 ± 1.0 mV in Iso, respectively (n= 4, P < 0.05). C, bath application of a saturating concentration of Bac induced a steady reduction in I_h amplitude (•) to 74.6 ± 3.3% (n= 15, P < 0.02). This effect was not associated with a decrease in the input resistance of the neurone (□), but rather a small increase to 111.6 ± 9.0% of control (n= 7, P > 0.05), probably due to decreased I_h amplitude. Inset shows an overlay of I_h activated during a voltage step from −62 to −92 mV in control and during Bac application in steady-state. D, evaluation of the relative tail current amplitudes of I_h shows that Bac induced a leftward shift in V_0.5 of the activation curve from −92.0 ± 0.3 to −97.0 ± 0.4 mV (n= 6, P < 0.01) with no change in the maximal conductance. E, Left, Bac effects on I_h were occluded when 5–10 μM 8Br-cAMP were present in the pipette solution. Right, pooled data illustrating the Bac-induced decrease in I_h in control (74.6 ± 3.3% of control amplitude, n= 15, P < 0.02) and with 8Br-cAMP present in the pipette (93.8 ± 3.1% of control, n= 6, P > 0.05). F, pooled data illustrating the effects of bath application of IBMX (filled columns) on the amplitude of I_h in control (117.8 ± 8.8%, n= 10, P < 0.001) and during preceding exposure to Bac (105.1 ± 5.3% of control, n= 9, *P < 0.05 compared to values in control) and of uncaging cAMP (open columns) on the amplitude of I_h in control (131.1 ± 5.8% of control, n= 4, P < 0.05) and during preceding exposure to Bac (133.7 ± 9.1% of control, n= 4, P > 0.05 compared to values in control). In E and F, step voltages from −62 to −92 mV were used to evoke I_h.

Figure 3. The potentiation of I_h by coapplication of Iso and Bac is associated with a strong positive shift in the activation curve with no change in maximal conductance.

A, top, family of h-currents during control (left) and in the presence of Iso and Bac (right). Note the pronounced increase in current amplitude at intermediately hyperpolarized potentials. A, bottom, graph of h-current amplitudes as a function of test potential in control (○) and in the presence of Iso and Bac (•). Maximal current amplitudes are not changed by Iso and Bac (−1701 ± 181 pA in control versus−1786 ± 207 pA in Iso + Bac, n= 6, P >0.05). B, activation curves in control (○) and in the presence of Iso and Bac (•). All tail currents were normalized with respect to the maximal tail current under control conditions. Co-applied Iso and Bac produced an 11 mV positive shift in the activation curve (from −95.7 ± 0.7 to −84.6 ± 0.9 mV, n= 6, P < 0.05) with no change in the maximal activation of the current.

Figure 5. The presence of Bac changes the time course of the cAMP transient induced by Iso.

A, data from a single experiment, illustrating the time course of I_h amplitudes following focal exposure to Iso before and after bath application of Bac. Bac alone reduced I_h in this cell by 15%. Selected I_h recordings are presented in the inset (1–5). B, averaged decay time course of the potentiation of I_h by local application of Iso alone (○) and in the presence of Bac (•). Lines show the non-linear least square fit of a monoexponential curve to the data, yielding a time constant τ= 241 ± 30 s for Iso + Bac versusτ= 32 ± 6 s for Iso (n= 4, P < 0.005). C and D, same experiment as in A and B, but with Bac applied at 0.8 μM. Note that the decay time course was further decelerated by low concentrations of Bac, such that monoexponential fitting was not possible.

Electrical stimulation of nucleus reticularis (nRt) afferents

Electrical stimulation of afferents from the nRt was achieved via bipolar tungsten electrodes (115 μm spacing, Frederick Haer & Co., Bowdoinham, ME) positioned within the nRt cell layer and exposed to constant current pulses (300–700 μA, 100 μs). To isolate GABA_B receptor-mediated responses, the bathing solution contained dl-2-amino-5-phosphonopentanoic acid (APV, 100 μM), 2,3-dioxo-6-nitro-1,2,3,4-tetrahydrobenzo[f]quinoxaline-7-sulphonamide disodium salt (NBQX, 10 μM) to block glutamatergic receptors and picrotoxin (100 μM) to block GABA_A receptors. In these experiments, Ba²⁺ was omitted from the extracellular solution until a synaptic GABA_B response was identified.

Flash photolysis of caged cAMP

For flash photolysis, caged cAMP [(P¹-(2-nitrophenyl)ethyl ester, 100 μM] was added to the patch solution from a 100-fold concentrated stock solution in dimethylsulfoxide. Flashes were applied via a UV-lamp attached to the epifluorescence pathway of the microscope and discharged via the capacitive discharges of the FlashMic (Rapp Optoelectronics, Germany), set at 4 V. Trains of 10 flashes at 10 Hz produced maximal responses and were used for the occlusion experiments (Fig. 4). UV-light applied to cells free of caged cAMP induced no change in I_h (n= 2). The magnitude of the response to photolysed cAMP was not significantly different for at least three flashes delivered at 1–2 min intervals (n= 3, P > 0.05), indicating that similar amounts of cAMP could be released at least three times within one cell.

Figure 4. The potentiation of I_h by coapplication of Iso and Bac leads to a reversible occlusion of I_h modulation via photolytically released cAMP.

A, data from two separate experiments showing the response to uncaged cAMP (photolysis was initiated at the time depicted by the arrow). In the experiment presented on the left, cAMP was photolytically released during control (•) and during the peak of the potentiation induced by Iso and Bac (○). In the experiment presented on the right, cAMP was photolytically released at the peak of the potentiation (○) and after its full decay (•). To facilitate comparison of the extent of the potentiation during the different periods, data were normalized to the average of the three data points preceding flash application. B, pooled data illustrating the increase in I_h amplitude following flash photolysis of caged cAMP during control (128.9 ± 6.3% of basal I_h amplitude, n= 8, P < 0.005), at the peak of the potentiation (106.6 ± 2.5% of I_h amplitude preceding the flash, n= 5, P >0.05), and after its full decay (123.6 ± 4.1% of basal I_h amplitude, n= 11, P >0.05versus increase before application of Iso and Bac). C, histogram of the time constants of activation of I_h during exposure to Forsk (825 ± 114versus561 ± 86 ms, n= 4, *P < 0.05) and during the response to coapplication of Iso and Bac (780 ± 57 ms in control; 533 ± 23 ms at the peak, n= 6, *P < 0.01; 723 ± 82 ms during recovery, P > 0.05 compared to control).

Pharmacological agents used in this study were purchased from SIGMA (St Louis, MO), except for caged cAMP (Calbiochem, Germany), tetrodotoxin (Latoxan, France), CGP54626 and NBQX (Tocris, UK).

Results

Regulation of I_h by G_s- and G_i/o-coupled neurotransmitter receptors in mouse TC cells

We initially confirmed the regulation of I_h by cAMP following activation of GPCRs the coupling of which to cAMP synthesis is well documented. We used the selective agonist Iso to activate β-adrenergic receptors, known to lead to stimulation of cAMP synthesis (Bloom et al. 1975; Madison & Nicoll, 1986; McCormick & Pape, 1990) and Bac to stimulate GABA_B receptors, which inhibit cAMP production (Wojcik & Neff, 1984; Knight & Bowery, 1996). Both β-adrenergic and GABA_B receptors are expressed in TC neurones (Rainbow et al. 1984; Princivalle et al. 2001; Kulik et al. 2002) and contribute to the control of intrinsic and synaptic processes related to thalamic function during sleep and arousal (for review, see McCormick & Bal, 1997; Huguenard, 1998).

In agreement with previous reports, bath application of Iso (0.5 μM), a selective β-adrenergic agonist, induced a gradual enhancement of the amplitude of I_h to 128.0 ± 4.3% of control (n= 6, P < 0.001; Fig. 1A). This enhancement showed a shallow dose dependence with a maximal value reached at 135.8 ± 5.7% of control amplitude for 5 μM Iso, and a half-maximal effect around 1 nM (116.2 ± 5.3% of control). In the continuous presence of Iso, the enhancement of I_h was maintained (<10% decay) for at least 5 min following start of the bath application. The Iso-induced increase in I_h amplitude corresponded to a positive shift in the activation curve of I_h from −94.4 ± 0.6 to −87.6 ± 1.0 mV with no change in maximal conductance (n= 4, P < 0.05; Fig. 1B; see Methods), similar to values reported previously (McCormick & Pape, 1990). Monoexponential fitting of current traces at −92 mV revealed time constants of 543 ± 97 ms in Iso compared to 726 ± 120 ms in control (n= 6, P < 0.02). The Iso-induced enhancement of current amplitude, the shift in the activation curve and the acceleration of the activation time course typically reflect increased binding of h-channels to cAMP (for review, see Pape, 1996; Wainger et al. 2001).

Conversely, alterations in the amplitude of cAMP-sensitive currents following activation of G_i/o-coupled neurotransmitter receptors have been associated with an inhibition of either basal or forskolin-stimulated AC activity (Pape, 1992; Ingram & Williams, 1994; Gerber & Gähwiler, 1994; Svoboda & Lupica, 1998). However, a decrease in I_h has also been attributed to a shunting action of K⁺ currents activated by G_i/o-coupled GABA_B receptors (Watts et al. 1996). In our experiments, the continuous presence of 1.5 mM Ba²⁺ blocked outward currents activated by Bac [172 ± 14 pA in the absence (n= 4); −6 ± 7 pA in the presence (n= 18) of Ba²⁺; P < 0.001; Pape, 1992; Sodickson & Bean, 1996], while I_h amplitude remained unchanged (<10% decrease, n= 11). The subsequent application of Bac at a concentration leading to maximal activation of K⁺ currents (80 μM; Sodickson & Bean, 1996) produced a persistent decrease in the amplitude of I_h to 74.6 ± 3.3% of the original value (n= 15, P < 0.02), without a decrease in input resistance (Fig. 1C). In contrast, no significant decrease in I_h was found when Bac was applied at 0.8 μM, close to the threshold for activation of K⁺ currents (111.5 ± 5.9% of control, n= 7, P >0.05; Sodickson & Bean, 1996). The Bac-induced decrease was associated with a negative shift in the activation curve of I_h from −92.0 ± 0.3 to −97.0 ± 0.4 mV (n= 6, P < 0.01) with no change in the maximal conductance of the current (Fig. 1D). Moreover, the monoexponential time constants of activation of I_h were increased from 669 ± 31 to 797 ± 35 ms (n= 8, P < 0.01; see Methods), consistent with a reduction of basal cAMP levels surrounding h-channels. In support of this possibility, inclusion of a saturating concentration of a non-hydrolysable analogue of cAMP in the patch pipette, 8Br-cAMP (5−10 μM), prevented the Bac-induced reduction of I_h amplitude (93.8 ± 3.1% of control amplitude, n= 6, P >0.05; Fig. 1E), indicating an occlusion of Bac-mediated inhibitory effects by cAMP previously bound to h-channels. Furthermore, we studied the effects of 3-isobutyl-1-methyl-xanthine (IBMX, 100 μM), a general phosphodiesterase (PDE) inhibitor, to prevent hydrolysis of cAMP molecules. Bath application of IBMX for 4–6 min induced a steady increase in the amplitude of I_h to 117.8 ± 8.8% (n= 10, P < 0.001; Fig. 1F), revealing a basal production of cAMP that is normally counteracted by the hydrolytic activity of PDEs. In nine cells exposed to Bac, this IBMX-induced enhancement was limited to, on average, 105.1 ± 5.3% of control amplitude (P < 0.05 compared to effects obtained with IBMX alone; Fig. 1F), suggesting that basal cAMP synthesis in the vicinity of h-channels was reduced. To exclude Bac-induced decreases in channel sensitivity to cAMP, we studied the effects of photolytic release of caged cAMP on I_h before and after application of Bac. Maximally three flashes were applied per experiment (see Methods). When cAMP was applied photolytically, it induced an increase of I_h amplitude to 131.1 ± 5.8% of control (n= 4, P < 0.05; Fig. 1F). Following application of Bac (80 μM), the increase amounted to 133.7 ± 9.1% (n= 4, P > 0.05; Fig. 1F), indicating that the sensitivity of I_h to exogenously applied cAMP was not affected by Bac. Taken together, our data strongly suggest that Bac leads to an inhibition of on-going AC activity in a concentration range covering that of GABA_B-mediated activation of K⁺ currents (Sodickson & Bean, 1996). Activated GABA_B receptors thus reduce cAMP levels and provoke an unbinding of cAMP tonically bound to h-channels.

In summary, pharmacological activation of G_s- and G_i/o -coupled GPCRs on TC neurones led to time-invariant changes in I_h that are consistent with an accumulation and a decrease of available cAMP in the vicinity of h-channels, respectively.

Co-activation of G_s- and G_i/o-coupled neurotransmitter receptors strongly up-regulates I_h

There is evidence that G_i/o-coupled receptors can strongly modulate cAMP-dependent regulation of ionic currents via G_s-coupled receptors in intact cells, either in an antagonistic (Hartzell, 1988; Pape, 1992; Gerber & Gähwiler, 1994) or in a potentiating fashion (Andrade, 1993; Pedarzani & Storm, 1996). However, the direction and physiological role of cross-talk between these two types of GPCRs has never been examined in the TC system, in which β-adrenergically mediated cAMP turnover is intimately involved in cellular activities related to states of arousal (Cirelli et al. 1996; McCormick & Bal, 1997). The expected time course of change in I_h amplitude, assuming an antagonistic effect of Bac on the stimulatory action of Iso, is demonstrated in Fig. 2B (thick line). In this case, I_h amplitude should deviate by <10% from control at all time points. Superimposed are the experimentally measured values of I_h amplitude. In sharp contrast to the expected time course, the amplitude of I_h was strongly enhanced during coapplication of Iso and Bac. As illustrated in the example in Fig. 2A, I_h amplitude approximately doubled (from −400 to −800 pA) upon coapplication of Iso and Bac, but then gradually decayed back to control with a monoexponential time constant of 164 s. In six cells, coapplication of Iso and Bac yielded a 152.6 ± 9.6% enhancement of I_h at the peak (range 132–222%) which decayed with τ= 281 ± 80 s (measured in n= 5 of 6 cells). During the coapplication, holding current levels showed no significant change (73 ± 21 pA before versus63 ± 24 pA, n= 6, P > 0.05; Fig. 2B; lower panel). The input resistance, measured from the instantaneous current response during the hyperpolarizing step, decreased by ∼30% (from 302 ± 56 MΩ to 212 ± 38 MΩ, n= 6, P < 0.005; Fig. 2B, middle panel) at the peak of I_h enhancement, and was restored to control before I_h enhancement had fully decayed (272 s after the start of agonist exposure, I_h amplitude at this point: 129.4 ± 4.2%, n= 6, P < 0.01). Therefore, the increase in I_h could not be explained by alterations in passive cellular properties, and the decrease in input resistance probably resulted from a contribution of strongly enhanced I_h to instantaneous current amplitudes. The results from the coapplication suggested that Iso and Bac effects on cAMP synthesis displayed a marked divergence from linear summation during an initial period, whereas they cancelled each other during a delayed steady-state phase.

Figure 2. Co-application of Iso and Bac induces a marked potentiation of I_h.

A, raw data showing the transient, strong increase in I_h upon bath application of Iso together with Bac at a concentration (80 μM) that reduced I_h when applied alone. Dotted lines are presented to facilitate comparison of instantaneous and h-current amplitudes between the four traces. The time course of the potentiation is illustrated at the bottom (•), together with the input resistance of the neurone (○). The data points deduced from the traces presented at the top (1–4) are indicated in the plot. B, pooled data from six cells illustrating the time course of the potentiation (top panel), of the input resistance (middle panel) and the holding current (lower panel). The thick line in the top panel depicts the linear sum of the effects of Iso and Bac illustrated in Fig. 1.

The effect of Iso and Bac on I_h showed properties distinct from that induced by Iso alone. Thus, the maximal enhancement of I_h obtained in the presence of Iso and Bac was markedly larger than that produced by Iso alone (n= 11 for Iso; n= 6 for Iso and Bac, P < 0.05). Furthermore, similar strong enhancements of I_h were found when 500 nM Iso were coapplied with low concentrations of Bac (0.8 μM; 182.3 ± 23.9%, n= 5, P < 0.05; cf. Fig. 5), which, when applied alone, did not reduce I_h (see above). The coapplication of Bac with Iso therefore induced a potentiating effect on I_h that differed with respect to strength and concentration dependence from that of Iso alone. This indicates the presence of a distinct regulatory pathway of I_h modulation induced by a synergistic action of Bac and Iso.

The synergistic effect is mediated by cAMP

The enhancement of I_h in the presence of Iso and Bac could be explained either by (1) a cAMP-dependent up-regulation of I_h, for example via an increased synthesis of cAMP triggered by the coapplication, or (2) alternative modulatory pathways targetting I_h that do not involve cAMP (see, e.g. Accili et al. 1997; Pan, 2003). We first determined the activation curve of I_h during the peak of the potentiation (Fig. 3A and B). The activation curve was shifted by 11.1 ± 1.7 mV (from −95.7 ± 0.7 to −84.6 ± 0.9 mV, n= 6, P < 0.05) towards more positive values (Fig. 3B), whereas the amplitude of the maximal currents elicited by voltage steps to −132 mV remained unchanged (−1701 ± 181 pA in the control versus−1786 ± 207 pA in Iso + Bac, n= 6, P > 0.05; Fig. 3A). Thus, the up-regulation could be described by a simple positive shift in the voltage dependence of I_h that was larger than that produced by Iso alone (P < 0.05) and consistent with a modulation of I_h by cyclic nucleotides. To further substantiate an involvement of cAMP, we used flash photolysis of caged cAMP to address the sensitivity of I_h to cyclic nucleotides at different time points during the modulation. When cAMP was photolytically released under control conditions, it produced an increase in amplitude to 128.9 ± 6.3% of control (n= 8, P < 0.005; Fig. 4A and B). In contrast, when caged cAMP was photolysed during the peak of the enhancement produced by Iso and Bac, the effect was reduced to 106.6 ± 2.5% of current amplitude preceding the flash (n= 5, P > 0.05; Fig. 4A and B), indicating that the response to cAMP was fully occluded when the action of Iso and Bac was maximal. In the continuous presence of Iso and Bac, the sensitivity to cAMP was restored upon complete decay of the enhancement, such that photolytically released cAMP increased I_h amplitude to 123.6 ± 4.1% of control (n= 11, P > 0.05versus flash-induced I_h increase before application of Iso and Bac; Fig. 4B). Thus, enhancement of I_h by coapplication of Iso and Bac occluded the response to cAMP, whereas its decay restored responsiveness. We also compared the properties of I_h at the peak of the enhancement with those induced following exposure to Forskolin (Forsk), a general AC activator. When bath-applied at a saturating concentration of 10 μM, Forsk produced a steady increase in I_h amplitude equaling 148.8 ± 5.3% of control (n= 4, P < 0.01), close to the enhancement produced by the coapplication of Iso and Bac (152.6 ± 9.6%, n= 6, P > 0.05). Monoexponential fitting of the time course of activation of I_h in the presence of Forsk yielded an acceleration of the time constant from 825 ± 114 to 561 ± 86 ms (n= 4, P < 0.05; Fig. 4C), reflecting increased cAMP binding to the channels. Similarly, the coapplication of Iso and Bac accelerated the time constant of I_h from 780 ± 57 ms to 533 ± 23 ms (n= 6, P < 0.01; Fig. 4C). Following the decay of the enhancement, however, the time constant recovered to 723 ± 82 ms (P > 0.05 compared to control). Using lower concentrations of Forsk (1 μM), we tested whether Bac could alter the sensitivity of interaction of h-channels with cAMP. Forsk alone enhanced I_h to 129.0 ± 2.8% of control (n= 4, P < 0.05), while the combination of Forsk and Bac yielded an enhancement of 114.0 ± 1.2% (n= 4, P < 0.05), slightly, but not significantly smaller than that observed with Forsk alone (P >0.05). Taken together, the alterations in the activation properties of I_h, the full occlusion of cAMP effects at the peak of the potentation, and the lack of a potentiating effect of Bac on the actions of Forsk are indicative of a mechanism dominated by cAMP that mediated the potentiation of I_h, probably via a stimulation of cAMP synthesis (see Discussion).

The presence of Bac transforms the time course of the cAMP signal induced by Iso

The potentiation induced by the coapplication of Iso and Bac suggests that a coincidental activation of β-adrenergic and GABA_B receptors recruited a powerful pathway of cAMP synthesis distinct from that targeted by the individual GPCRs. To further characterize this pathway, we investigated how the presence of Bac affected the time course of the response induced by a brief application of Iso. For this purpose, we combined focal application of Iso with bath application of Bac. When Iso (500 nM in a pressure ejection pipette) was applied during baseline recording, it produced a 132.5 ± 4.1% enhancement of I_h amplitude (n= 13, P < 0.001) that changed by <5% during subsequent applications in control (n= 3) and decayed with a time course of 32 ± 6 s. When Bac was applied at a saturating concentration of 80 μM, focal application of Iso produced a potentiation of 175.8 ± 10.0% of control amplitude (n= 5, P < 0.02; Fig. 5A). The potentiated response decayed with a time constant of 241 ± 30 s (measured in n= 4 cells; Fig. 5B), which was markedly slower than the control response (P < 0.005). When Bac was applied at 0.8 μM, focal application of Iso produced an enhancement equaling 190.6 ± 29.2% (n= 7, P < 0.02 compared to control responses which yielded 141.9 ± 11.0% increase of I_h; Fig. 5C). The decay of these responses was even further decelerated, with a remaining 152.7 ± 19.5% increase of I_h at 4 min after the application of Iso (Fig. 5D). Thus, the cAMP synthesis pathway requiring both Iso and Bac showed a distinct temporal profile of cAMP synthesis that depended on the strength of activation of Bac receptors. Interestingly, the time course of the potentiated response greatly outlasted the duration of the stimulus mediated by Iso alone. Thus, the presence of Bac allows the transformation of a transient positive input for cAMP synthesis into a more persistent cAMP signal.

Pharmacological properties of the receptors involved in the potentiation

We next verified whether γ-aminobutyric acid (GABA), the natural ligand for GABA_B receptors, could induce the potentiation of Iso effects by Bac. The potentiation was mimicked when Bac was replaced by GABA (1 mM) in the presence of the GABA_A receptor antagonist picrotoxin (100 μM) in 5 of 7 cells tested (145.8 ± 9.4% in control versus 183.6 ± 11.4% in GABA + Iso, P < 0.05; Fig. 6A and D), indicating that the endogenous agonist for GABA_B receptors could induce a potentiation of β-adrenergic responses. To further address the involvement of G_i-proteins in the potentiation of I_h, we used N-ethylmaleimide (NEM), a membrane-permeable inhibitor of pertussis-toxin sensitive G-proteins (Winslow et al. 1987; Shapiro et al. 1994; Hirono et al. 2001), to selectively interfere with G_i- but not G_s-proteins. Bath application of NEM for 2 min fully antagonized Ba²⁺-sensitive outward currents induced by Bac at –50 mV (data not shown), while it did not interfere with Iso-induced enhancements of I_h (92.2 ± 0.9% of responses without NEM, n= 4, P > 0.05), consistent with a selective inhibition of G_i-proteins. When Iso was locally applied in the presence of Bac (0.8 μM) and NEM, the potentiation was fully abolished (151.1 ± 4.0% in Iso versus 150.3 ± 6.8% enhancement in Iso and Bac, n= 4, P > 0.05; Fig. 6B). Moreover, the time constants of the decay of the cAMP transient in NEM were rapid, with a monoexponential decay of 90 ± 43 s (compared to 47 ± 16 s in control, P > 0.05; Fig. 6C). These data point to a requirement of G_i-proteins activated by GABA_B receptors in the potentiation and prolongation of the cAMP stimulation mediated by Iso.

Figure 6. Pharmacological characterization of the up-regulation.

A, the natural transmitter for GABA_B receptors, GABA, induced a potentiation of Iso responses in the presence of picrotoxin (100 μM). The graph illustrated the time course of I_h amplitudes in a single experiment. Selected I_h recordings are presented in the inset. B, the modulation of I_h by Iso remained unaltered in the presence of Bac (0.8 μM), when NEM (120 μM) was preapplied for 2 min. Time course of I_h amplitudes in a representative experiment. Selected I_h recordings are presented in the inset (1–5). C, average time course of decay of Iso-induced modulation of I_h before (○, τ= 47 ± 16 s, n= 4) and after (•, τ= 90 ± 43 s, n= 4, P >0.05) application of Bac in the presence of NEM. D, histogram summarizing the effects of different combinations of agonists for GPCRs. The responses to Iso were potentiated in the presence of GABA (1 mM) and picrotoxin (PTX, 100 μM) (183.6 ± 11.4%versus 145.8 ± 9.4% of control in 5 of 7 cells tested, *P < 0.05), but not in the presence of CPA (50 μM) (146.4 ± 16.4%versus143.0 ± 8.3% of control, n= 3, P > 0.05), an A1 agonist. DAMGO (1 μM), a μ-opioid receptor agonist, increased I_h in the absence of Iso to 148.0 ± 16.1% of control (n= 4, P < 0.02).

We then investigated whether activation of GPCRs other than baclofen-sensitive receptors also enhanced Iso responses. The adenosine A1 receptor agonist N⁶-cyclopentyladenosine (CPA, 50 μM) was tested because adenosine receptors are functionally expressed in TC neurones and share common mechanisms of action with GABA_B receptors in these cells (Pape, 1992). When applied in the bath, CPA did not affect the amplitude of control responses induced by puff application of Iso (146.4 ± 16.4% during control versus 143.0 ± 8.3% during CPA, n= 3, P > 0.05; Fig. 6D), suggesting that A1 receptors did not undergo synergistic interactions with β-adrenergic receptors. We then used [D-Ala-2, NMe-Phe-4, Gly-5-ol]-enkephalin (DAMGO, 1 μM) to activate G_i/o-coupled μ-opioid receptors that are widely expressed in TC neurones (Brunton & Charpak, 1998). Interestingly, application of DAMGO alone caused a rapid, transient increase in the amplitude of I_h to 148.0 ± 16.1% of control amplitude (n= 4, P < 0.02; Fig. 6D), which was associated with an acceleration in the monoexponential time constant from 803 ± 92 to 680 ± 99 ms (n= 4, P < 0.01), suggesting that activation of μ-opioid receptors alone coupled positively to cAMP production detected by I_h (see Discussion). Thus, distinct G_i/o-coupled neurotransmitters appear to couple differently to cAMP synthesis pathways in the vicinity of I_h, both when activated alone or in conjunction with β-adrenergic receptors.

Actions of synaptically activated GABA_B receptors on I_h modulation

In thalamic networks, synaptic activation of GABA_B receptors on TC neurones occurs during both sleep-related and during pathological hypersynchronous activity in vitro resembling generalized epilepsies (Blumenfeld & McCormick, 2000; Bal et al. 2000), and can result from a hyperexcitation of GABAergic afferents arising in the nRt. The effect of postsynaptically activated GABA_B receptors on cAMP metabolism in TC neurones is, however, unknown. To address the coupling of synaptically activated GABA_B receptors to cAMP-mediated modulation of I_h, we studied the effects of GABA_B receptor-mediated synaptic currents evoked via electrical stimulation in the nRt (see Methods). In the presence of glutamatergic and GABA_Aergic receptor antagonists (APV, 100 μM; NBQX, 10 μM; Picrotoxin, 100 μM), electrical stimulation evoked slow outward current responses that peaked at a delay of 80 ± 4 ms (range 59–100 ms) and reached amplitudes of 12 ± 3 pA (range 2–50 pA, n = 18; Fig. 7A), similar to responses previously described in rat (Ulrich & Huguenard, 1996). The highly selective GABA_B receptor antagonist CGP54626 (500 nM) was tested in four cells and completely blocked these outward currents, indicating that they were mediated by GABA_B receptors (Fig. 7A). Iso was then applied locally while concomitantly eliciting GABA_B responses (10 stimuli, 5 Hz in the presence of 1.5 mM Ba²⁺) and the modulation of I_h amplitude was monitored. Synaptic activation of GABA_B receptors during simultaneous local application of Iso induced a significant potentiation of I_h amplitude (I_h amplitude in Iso, 118.5 ± 4.3% of control, and I_h amplitude in Iso, with GABA_B receptors activated synaptically, 140.2 ± 12.0% of control, n = 7, P < 0.05; Fig. 7B and C). In contrast, synaptic stimulation alone induced a minor enhancement of I_h (101.7 ± 2.4% of control amplitude at 24 s after application of the stimulation; n= 4, P > 0.05; Fig. 7B) and these stimulation-dependent effects were subtracted from the responses obtained during concomitant stimulation and Iso application. Our results thus suggest that synaptically activated GABA_B receptors can contribute to the control of cAMP turnover in TC neurones, while a modulation of I_h via released compounds other than GABA appears to play a minor role.

Figure 7. Examination of the role of synaptic GABA_B receptors in the potentiation of I_h responses by Iso.

A, CGP 54626-sensitive outward currents elicited by stimulation of afferent nRt fibers via bipolar tungsten electrodes (300–700 μA, 100 μs). B, in a different cell, representative I_h responses were monitored under control conditions (ctrl) and after Iso was applied with (+ stim and + Iso) and without (+ Iso) concomitant activation of GABA_B receptors (10 stimuli at 5 Hz). Overlay shows control I_h, the current response to application of Iso alone, and the response following conjoint Iso application and electrical stimulation. During this experiment, Ba²⁺ ions (1.5 mM) were present to prevent activation of outward K⁺ currents. C, averaged data for seven experiments, indicating a significant increase in the Iso sensitivity of I_h amplitude after electrical stimulation (arrow, 118.5 ± 4.3% of control I_h amplitude during Iso application; 140.2 ± 12.0% of I_h amplitude during Iso application and coactivation of GABA_B receptors, n= 7, P < 0.05). (○, control responses; •, responses with GABA_B receptor activation.)

Discussion

Here we describe distinct types of GPCR-induced cAMP signals that modulate the pacemaker current of TC neurones. The strength and time course of modulation of I_h varied considerably depending on the pattern of activation of GPCRs. Steady increases or decreases in I_h amplitude occurred upon activation of single receptor types coupled positively or negatively to cAMP production. Thus, a basal cAMP turnover, which is pronounced in thalamus compared to other regions of the brain (Matsuoka et al. 1997; Ihnatovych et al. 2002), can be up- and down-regulated steadily under the tonic influence of neurotransmitter receptors. In contrast, coexposure to two agonists was integrated in a supralinear manner to produce a strong, transient increase in I_h, probably mediated by a pharmacologically and kinetically distinct cAMP synthesis pathway (see below). Thus, the pacemaker current can be exposed to cAMP signals originating from diverse sources, suggesting that pacemaker channels are surrounded by cAMP synthesis pathways with distinct molecular properties.

A crucial point in our study was to demonstrate that the dynamics of the modulation of I_h by agonists for GPCRs primarily reflected the time course of intracellular cAMP concentrations. While increases in I_h via Iso have been attributed to synthesis of cAMP in TC neurones (McCormick & Pape, 1990), Bac-mediated decreases of I_h were proposed to occur through a pathway independent of cAMP in several neuronal cell types (Jiang et al. 1993; Watts et al. 1996; Pape, 1996). In TC neurones, a highly cAMP-sensitive isoform of I_h is expressed (Seifert et al. 1999; Santoro et al. 2000) and basal cAMP synthesis rate in TC neurones is comparatively pronounced (Matsuoka et al. 1997; Ihnatovych et al. 2002), two factors advantageous for detecting decreases in cAMP via I_h. Indeed, the inhibition of I_h identified here showed characteristics typical for cAMP-dependent actions on h-channels, including a negative shift in the activation curve and decelerated activation kinetics (for review, see Pape, 1996; Wainger et al. 2001). In addition, a saturating cAMP concentration largely occluded Bac effects, the activity of ACs appeared reduced in the presence of Bac, and there was no measurable change in the sensitivity for cAMP in the presence of Bac. These findings therefore indicate that G_i/o-coupled GABA_B receptors, in addition to adenosine A1 receptors (Pape, 1992), primarily reduce I_h via inhibiting basal cAMP synthesis in TC neurones. However, a minor Bac-induced contribution to cAMP-independent modulation of I_h, for example via decreases in the concentration of cGMP (Fedele et al. 1997; Pape & Mager, 1992), can not be excluded.

A vigorous synthesis of cyclic AMP probably mediated the transient potentiation of I_h by Iso and Bac. First, the modulation of I_h was associated with a maximal shift in the activation curve of I_h, similar to that observed with exogenous addition of high cAMP concentrations intracellularly (McCormick & Pape, 1990; Lüthi & McCormick, 1999). Second, modulation of I_h, induced by photolysis of caged cAMP, was occluded during the peak of the potentiation induced by Iso and Bac, but fully reinstated following the decay of the potentiation. Third, reversible accelerations in the time course of activation of I_h, which are widely used hallmarks of cAMP-dependent actions on h-channels, paralleled the enhancement. Fourth, the characteristics of the potentiation could be mimicked by stimulation of endogenous ACs with Forsk. Fifth, Bac did not appear to alter the sensitivity of h-channels for cAMP generated in the presence of non-saturating concentrations of Forsk. Taken together, the enhancement of I_h by coapplication of Iso and Bac shows characteristics that are consistent with an exposure of h-channels to a powerful elevation of cAMP, which represents the most widely described pathway of pacemaker current modulation (for review, see Pape, 1996; Santoro & Tibbs, 1999; Kaupp & Seifert, 2001; Robinson & Siegelbaum, 2003). Alternate, more complex modulations of pacemaker channel function, such as decreases in the concentrations of cAMP required for channel gating, can, however, not be excluded at this point and would require experiments under cell-free conditions.

A cross-talk between Bac and Iso receptors in cAMP signalling could be induced at the level of the receptors, the G-proteins, the ACs and the PDEs. As application of Bac alone inhibited rather than enhanced I_h, an involvement of alternate GABA_B receptor induced modulatory pathways, such as Ca²⁺ release (Hirono et al. 2001) and an associated cAMP synthesis (Lüthi & McCormick, 1999) are unlikely to be involved in the synergism. Moreover, Bac failed to potentiate the action of low concentrations of Forsk, indicating that AC activity was not stimulated directly by Bac. Furthermore, inhibition of PDEs and decreasing cAMP degradation led to a weak augmentation of I_h that could not account for the strength of the potentiation. Therefore, the synergism appears to arise at a point upstream of cAMP synthesis. Accordingly, we were able to interfer with the potentiation by using NEM, an inhibitor of G_i-proteins that interferes with multiple GABA_B receptor-mediated effects on ionic currents (Sodickson & Bean, 1996; Hirono et al. 2001). Activation of G_i-proteins, probably induced by ligand-bound GABA_B receptors showing a high apparent affinity for Bac, appears thus to be a primary requirement for inducing a potentiation of cAMP synthesis, induced in conjunction with stimulation of G_s-proteins.

In biochemical assays of cAMP levels in neural tissue, activation of G_i/o-coupled neurotransmitter receptors, including GABA_B,α-adrenergic and μ-opioid receptors has been reported to potentiate cAMP accumulation induced by G_s-coupled neurotransmitter receptors by severalfold (Perkins & Moore, 1973; Sattin et al. 1975; Karbon & Enna, 1985; Makman et al. 1988). This paradoxical action of Bac is sensitive to pertussis toxin (Wojcik et al. 1989), and was proposed to include a Bac-induced strengthening of receptor coupling to AC (Scherer et al. 1989) and arachidonic acid metabolism (Duman et al. 1986; Schaad et al. 1989). However, a profile of modulation that could explain most directly the synergistic action of G_i/o-coupled receptors upstream of cAMP synthesis is presented by two types of AC isoforms, type II and IV. The activity of these enzymes is dramatically enhanced upon binding of βγ-subunits from G_i-proteins in the presence of α subunits from G_s-proteins (Tang & Gilman, 1991; for review, see Tang & Gilman, 1992; Anholt, 1994; Smit & Iyengar, 1998; Hanoune & Defer, 2001). These ACs may be involved in G_i/o-stimulated cAMP production in Xenopus oocytes (Uezono et al. 1997; Ulens & Tytgat, 2001), in ventricular myocytes (Belevych et al. 2001) in olfactory bulb (Olianas & Onali, 1999), and in cortex (Onali & Olianas, 2001), but their activity in intact neurones has not been addressed. In TC neurones, both type II and type IV AC are expressed, and type IV activity is up to 10-fold more pronounced than that of other ACs (Ihnatovych et al. 2002). This suggests that the marked cAMP production and the dependence on synergistic activation of both G_s- and G_i-proteins described here may be, at least in part, explained by the activation of this molecularly distinct AC in TC cells.

The activation of different types of G_i/o-coupled receptors showed variable capability to potentiate the action of Iso on I_h. Whereas A1 agonists failed to induce a potentiation, μ-opioid receptor activation induced a potentiation of I_h in the absence of β-adrenergic receptor activation. Thus, in mouse TC neurones, the coupling of G_i/o-coupled receptors to cAMP turnover detected by I_h, whether positive or negative, depends on the receptor type. Interestingly, in Xenopus oocytes, it was observed that μ-opioid receptors unexpectedly potentiate HCN2-mediated currents in ∼10% of the oocytes in a manner that is sensitive to AC blockers and limited by the availability of free α_s-subunits (Ulens & Tytgat, 2001). Thus, these GPCRs appear to preferentially couple to synthesis of cAMP that is detected by isoforms of HCN channels expressed in thalamus (Moosmang et al. 1999; Monteggia et al. 2000; Santoro et al. 2000).

Electrical stimulation in the nRt cell layer, which comprise the principal inhibitory afferents into the dorsal thalamus (for review, see Guillery et al. 1998; Crabtree, 1999), produced a GABA_B receptor-mediated postsynaptic response on TC neurones, with properties similar to those found in rat (Ulrich & Huguenard, 1996). To study the effect of synaptically activated GABA_B receptors on cAMP formation, the activation of K⁺ currents was prevented and stimulation applied repetitively within the frequency range of thalamic oscillations (McCormick & Bal, 1997). This protocol induced minor changes in I_h, indicating that GABA_B receptor activation was not strong enough to inhibit cAMP synthesis. Moreover, stimulation per se appeared not to induce release of other neurotransmitters that led to strong modulation of I_h. However, when stimulation occurred concomitantly with activation of β-adrenergic receptors, a marked potentiation of I_h amplitude was observed. Thus, the interaction of β-adrenergic receptors with GABA_B receptors to control cAMP synthesis extends to the synaptic level, although additional factors released upon stimulation that may interact synergistically with β-adrenergic receptors can not be excluded (see, e.g. Pedarzani & Storm, 1996). The present data therefore suggest that the GABAergic tone exerted by nRt cells may control the strength of cAMP synthesis induced by afferent neuromodulatory pathways. Interestingly, locus coeruleus neurones discharge synchronously with sleep-related EEG rhythms in the TC system, while they fire in isolation during states of waking (Aston-Jones & Bloom, 1981). Thus, activation of β-adrenergic receptors could take place over an increased level of activated GABA_B receptors during states of sleep and during the transition between sleeping and waking, perhaps associating these phases with intracellular cAMP signals distinct from those during waking. Norepinephrine plays an important role in the control of cAMP-dependent gene expression during states of arousal in the TC system (Cirelli et al. 1996; Cirelli & Tononi, 2000) and discrete temporal profiles of cAMP transients contribute to determine the patterns of gene expression (Bacskai et al. 1993; Kaang et al. 1993). The physiological role of various types of cAMP signals induced by GPCRs and via synergistic interactions between these could therefore extend into the determination of state-dependent patterns of gene expression.