- Email: [email protected]
Modulation of synaptic function by cGMP and cGMP-gated cation channels
Neurochemistry International 45 (2004) 875–884
Modulation of synaptic function by cGMP and cGMP-gated cation channels Colin J. Barnstable∗ , Ji-Ye Wei, Ming-Hu Han Department of Ophthalmology and Visual Science, Yale University School of Medicine, 330 Cedar Street, New Haven, CT 06520-8061, USA Available online 15 June 2004
Abstract Cyclic nucleotide-gated cation channels have been studied intensively in the primary sensory neurons of the visual and olfactory systems. Using both anatomical and physiological methods we have shown that they have a much more widespread distribution in the nervous system. In many retinal ganglion cells cGMP, but not cAMP, activates a non-selective conductance that has many of the properties of CNG channels. As many neurons also contain cGMP-dependent protein kinases (PKGs), we have used a variety of cGMP analogues to distinguish the actions of cGMP. Sp-8-Br-PET-cGMPS is a potent non-hydrolyzable cGMP analogue that is an agonist of PKG. We found that Sp-8-Br-PET-cGMPS acts as a competitive inhibitor of at least the rod CNG channel. Rp-8-Br-cGMPS has shown the opposite effects, namely as an agonist of the rod CNG channel and an inhibitor of PKG. In dissociated cell cultures and slices of rodent visual cortex cGMP had multiple rapid and reversible effects on transmission at glutamatergic synapses. Extracellular application of 8-Br-cGMP or Sp-8-Br-PET-cGMPS reduced stimulus evoked EPSPs in cortical slices. In cortical cultures both analogs reduced the frequency of spontaneous EPSCs, but not their amplitude. The effects on both EPSPs and EPSCs were presynaptic. The effects on evoked EPSPs may be due, in part, to reduced calcium influx through voltage-gated calcium channels. The effects on spontaneous EPSCs may be due, in part, to modulation of calcium fluxes through internal stores. Similar modulations of synaptic transmission have been found at gabaergic synapses. On postsynaptic cells, PKG activation produced a dramatic enhancement of the responses to applied NMDA. No effects were detected on applied AMPA/kainate or GABA. Together the results suggest that cGMP may use multiple mechanisms to modulate synaptic efficacy and that its actions may include regulating synaptic plasticity and the relative strength of excitatory and inhibitory drive through neural pathways. © 2004 Elsevier Ltd. All rights reserved. Keywords: Synaptic function; Modulate synaptic efficacy; cGMP-gated cation channel; cGMP-dependent protein kinase; Glutamatergic synapses; GABAergic synapses
1. Introduction Cyclic nucleotide (cGMP and cAMP) modulation of central synapses has been investigated for decades. Earlier reports focused on the effects of cAMP and cGMP on rapid synaptic modulation (see review by Bloom, 1976 and Bloom, 1979). Later studies concentrated on cGMP or cAMP-induced long-term synaptic potentiation (LTP) or depression (LTD) that is associated with learning and memory (Zhuo et al., 1994; Lev-Ram et al., 1997). It has been suggested that cGMP and cAMP play opposite regulatory roles in excitatory central synapses (Bloom, 1979). Using extracellular recordings and multibarrel ionphoresis in whole animals it was shown that cGMP caused excitation of a majority of identified pyramidal tract neu∗
Corresponding author. Tel.: +1-203-785-4282; fax: +1-203-785-7401. E-mail address: [email protected] (C.J. Barnstable).
0197-0186/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuint.2004.03.018
rons but cAMP had depressive effects on the pyramidal neurons in the mammalian motor cortex (Stone et al., 1975; Stone and Taylor, 1977). In mammalian visual cortex, excitation induced by cGMP in some pyramidal cells has been detected by extracellular recording in brain slices (Cudeiro et al., 1997; Cudeiro and Rivadulla, 1999). However, the above experiments were performed in the whole animal or brain slices that contain polysynaptic networks with both excitatory and inhibitory synaptic connections. This raises the question of whether the facilitation of excitatory synapses by cGMP in the brain slice and/or whole animal is due to direct effects of cGMP on glutamatergic synapses, or through inhibiting adjacent GABAergic neurons. It was originally thought that the cAMP- and cGMPdependent protein kinase families were the primary effectors of these cyclic nucleotides in a variety of physiological processes (Kuo and Greengard, 1969; Krebs and Beavo, 1979; Walter, 1989; Hofmann et al., 1992; Butt et al., 1993;
876
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
Francis and Corbin, 1994; Wang and Robinson, 1997). It now appears, however, that the other major biological effector of cyclic nucleotide action is a class of cyclic nucleotide-gated (CNG) cation channels in the cell membrane (Yau and Baylor, 1989; Kaupp, 1991; Barnstable, 1993; Yau, 1994; Kaupp, 1995; Zimmermann, 1995; Biel et al., 1995; Zagotta and Siegelbaum, 1996; Finn et al., 1996; Zufall et al., 1997). CNG channels were first described in rod photoreceptors (Fesenko et al., 1985; Yau and Nakatani, 1985). Since then, other CNG channels have been identified as responsible for cone photoreceptor transduction (Cobbs et al., 1985; Haynes and Yau, 1985) and olfactory sensory transduction (Nakamura and Gold, 1987). In the work reviewed below we have shown first that CNG channels are widely distributed in the mammalian Central Nervous System (CNS) and second, that cGMP is an important modulator of both glutamatergic and gabaergic synapses in cerebral cortex. 2. CNG channels are widely distributed in the mammalian CNS Our first indication that CNG channels might be expressed in cells other than primary sensory neurons came from in situ hydridization experiments in which labeling was consistently seen over neurons of the inner retina, particularly retinal ganglion cells (Ahmad et al., 1990). Subsequent patch clamp recording from retinal ganglion cells in culture confirmed the presence of a conductance that was activated by cGMP but not cAMP (Ahmad et al., 1994). This conductance was non-selective for sodium or potassium, could be blocked by cadmium and was decreased in the presence of calcium, all properties expected for a CNG channel. We also found that the conductance could be activated by application of nitric oxide donors, suggesting that the activity of the CNG channels might be coupled to Nitric Oxide (NO) signaling pathways. Of even more interest was the observation
that the conductance could be activated by the application of phosphodiesterase inhibitors. This suggests that cGMP production and hydrolysis within retinal ganglion cells, and possibly other CNS neurons, is continuous and that levels of cGMP can be rapidly modulated by altering either synthesis or breakdown. Based on these results, we proposed a model whereby the activity of CNG channels might regulate a number of processes at synaptic terminals (Fig. 1). Reducing the resting membrane potential and increasing calcium concentrations within the terminal would both be expected to increase transmitter release at the terminal and thus act in a facilitatory manner. Although these results were generally accepted as representing events in cultured retinal ganglion cells, it was only more recently that data have been obtained to support the idea that CNG channels might operate in this way in vivo (Kawa and Sterling, 2002). In subsequent work we examined the expression of CNG channels in other regions of the CNS. In hippocampus we obtained clear evidence by both in situ hybridization and patch clamp recording for CNG channel expression in pyramidal neurons of CA1 through CA3 layers and the granule cells of the dentate gyrus (Leinders-Zufall et al., 1995; Kingston et al., 1996). Work from other groups has also established that CNG channels are expressed in the mammalian CNS (el-Husseini et al., 1995; Bradley et al., 1997). Further studies using PCR identified expression of at least the olfactory type of CNG channel expressed in many regions of the rodent CNS (Kingston et al., 1999). Similarly, in situ hybridization studies using visual cortex have detected the expression of CNG channel RNA in many layers (Samanata Roy and Barnstable, 1999). There is an interesting correlation between the spatial expression of CNG channels in mammalian CNS neurons and that of elements of the NO system. This led us to propose that CNG channels might be one of the major effectors of a NO transmitter system. CNG channels have a number of advantages as effectors. First, they are not voltage dependent and thus their activity is independent of other aspects of
Fig. 1. Schematic representation of pathways that might regulate the activity of the cGMP-gated channel in retinal ganglion cells. cGMP synthesis could be stimulated by NO adjacent amcrine cells. Activation of cGMP-gated channels will increase Ca2+ influx and enhance Ca2+ driven processes. At the same time, G-protein coupled receptors activated by neurotransmitters from bipolar and amacrine cell terminals may regulate the activity of one or more PDEs, which control hydrolysis cGMP and thus the activity of the cGMP-gated channels. Modified from Ahmad et al. (1994).
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
neuron activity. Second, they do not show inactivation and can thus provide a continuous response to signals such as NO. Third, they are calcium permeable and thus their activity can readily be coupled to many other cellular processes by modulating cytoplasmic calcium levels.
3. Pharmacological isolation of cGMP effector pathways One of the problems studying the effects of cGMP on neurons is that it can activate protein kinases as well as CNG channels. To allow each pathway of cGMP action to be studied separately, we have investigated a series of agonists and antagonists. Some molecules that interact with CNG channels were first characterized because of actions on other molecules. For example, LY83583 (6-anilino-5.8-quinolinequinone), first characterized as a guanylyl cyclase inhibitor, can block olfactory CNG channels with high potency, which means that it cannot be used to study the role of guanylyl cyclase on CNG channel activity (Leinders-Zufall and Zufall, 1995). W-7 [N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide], a calmodulin (CaM) inhibitor is also a potent inhibitor of olfactory CNG channel independent of its action at the CaM-binding site localized on the cytoplasmic side of the CNG channel (Kleene, 1994). We have investigated the function of a series of substituted cGMP derivatives. It was shown many years ago that 8-Br-cGMP was a good agonist of rod cGMP-gated channels with 10-fold higher potency, increased membrane-permeability and reduced velocity of PDE hydrolysis (Zimmermann, 1995). Wei et al. (1996) identified the first competitive antagonist of the photoreceptor CNG channel that is also an activator of PKG. Sp-8-Br-PET-cGMPS is both
877
membrane-permeant and PDE-resistant and can help elucidate the physiological roles of CNG channels in intact cells (Fig. 2). The properties of -phenyl-1-, N2-ethenosubstituted cGMP analogs are probably due to modification of three key sites that generate high binding affnity but eliminate the capacity for activation. The related compound, Rp-8-Br-PET-cGMPS, is an antagonist of both CNG channels and PKG. Furthermore, a cGMP-analog, Rp-8-Br-cGMPS, has shown the opposite effects on the rod CNG channel and PKG, acting as an agonist of the rod CNG channel and an inhibitor of PKG (Wei et al., 1998). The results suggest that antagonists of the rod CNG channel and PKG are based on different structural features of the ligands. The most important determinants for the rod CNG channel are in the guanine ring, whereas for the kinase they include the isomeric position of the thiophosphate S atom (Rp versus Sp). Interestingly, the olfactory CNG channel shares a similar ligand property with protein kinases, since the ligand Rp-cGMPS can activate the rod CNG channel and is an antagonist of the olfactory CNG channel (Kramer and Tibbs, 1996). The authors made chimeric photoreceptor and olfactory CNG channels by replacing the C-terminal domain, which contains the cyclic nucleotide binding site, and confirmed that this domain determines ligand selectivity and is coupled to channel openings.
4. cGMP modulates cortical glutamatergic synapses at presynaptic terminals To gain an overall impression of the effects of cGMP on cortical neurons we studied pyramidal neurons of visual cortex in brain slices from 2- to 4-week-old rats using
Fig. 2. PET-substituted cGMP analogs interact with the binding domains of rod CNG channels. The syn- and anti-conformations of the analogs have the substituted guanine ring in different orientations relative to the sugar residue. Residue D597 cannot interact with modified N1.N2 sites of Sp-/Rp-8-Br-PET-cGMPS. However, the 8-Br substitution still increases the affinities of binding of these two analogs. Modified from Wei et al. (1998).
878
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
standard whole-cell patch-clamp recording techniques (Wei et al., 2002). Neurons were included in the study if they had resting membrane potentials of −60 to −75 mV and could generate a train of overshooting action potentials in response to intracellular depolarizing current pulses. We focused on the effects of cGMP on excitatory synapses in layers V and VI because these layers are abundant in [3 H] cGMP binding and its related NOS staining (Aoki et al., 1997; Bladen et al., 1996). The membrane-permeable cGMP analog 8-Br-cGMP (100 M) reduced EPSPs recorded in layer V/VI of visual cortex. The EPSP produced in a layer V cell by stimulation of layer II/III cells was reduced by 40% after application of 100 M 8-Br-cGMP for 3 min. A similar result was obtained in layer VI cells. The average reduction by 8-Br-cGMP was 42 ± 10% (P < 0.001, n = 5) in layer V and 36 ± 7% (P < 0.001, n = 6) in layer VI (Fig. 3B). After washout, the peak EPSPs in the two cells recovered. The mean change in
membrane potential induced by 8-Br-cGMP itself was only 0.38 ± 2 mV (n = 20). To investigate whether CNG channels or PKGs were responsible for the cGMP effect, and to determine whether the action of cGMP was pre- or post-synaptic, we bath applied 50–100 M Sp-8-Br-PET-cGMPS. After 4 min of bath application of Sp-8-Br-PET-cGMPS, the EPSP was reduced by 42% (Fig. 3C, left side). Five cells were recorded in layer VI with a reduction of 37% (Fig. 3D, left side). This is the same as the reduction seen with 8-Br-cGMP, indicating that essentially all the effect of cGMP is exerted through the activation of PKG. In the same series of experiments a membrane-impermeable specific PKG inhibitory peptide was infused for 5–10 min into the postsynaptic cell. Under these conditions, the EPSPs were still reduced by extracellular iontophoresis or perfusion of Sp-8-Br-PET-cGMP (Fig. 3C, right side). In a group of four cells recorded under
Fig. 3. cGMP and PKG inhibit evoked EPSPs in slices of visual cortex. (A) Reduction of evoked EPSPs by 8-Br-cGMP in layersV (left side) and VI (right side). EPSPs were evoked by stimulation of layers II/III for recordings in layer V and VI. The membrane potentials are shown in the left columns of each figure. (B) Histograms show the average reduction of EPSPs by 8-Br-cGMP in layers V and VI, and error bars show the standard deviation (n = 5 for layer V and n = 6 for layer VI). (C) The specific PKG activator Sp-8-Br-PET-cGMPS shows a similar reduction of EPSPs (left side). When a PKG peptide inhibitor was infused via the recording electrode to block the postsynaptic effects of PKG, perfusion of Sp-8-Br-PET-cGMPS still produced a reduction of the EPSP evoked in slices of visual cortex (right side). (D) Histograms illustrate the average reduction of EPSPs by Sp-8-Br-PET-cGMPS, with or without PKG peptide inhibitor postsynaptically (n = 5 for without PKG peptide inhibitor and n = 4 for with PKG peptide inhibitor). Adapted from Wei et al. (2002).
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
these conditions, we observed a reduction of 29 ± 8% by Sp-8-Br-PET-cGMP (Fig. 3D, right side). This result suggests that the action of cGMP was on the presynaptic input to the recorded cells. A presynaptic action of cGMP was also observed when cultured cortical neurons were studied (Wei et al., 2002). High density cultures of cortical neurons form abundant synaptic interactions that can be detected in single electrode voltage-clamp recordings as spontaneous EPSCs. Bath application of DNQX (20 M) together with AP5 (100 M) blocked the sEPSCs, thus confirming that the spontaneous synaptic activity studied in these experiments was glutamatergic. In observations of mEPSCs, application of 1 mM 8-Br-cGMP or 100 M Sp-8-Br-PET-cGMPS with or without 400 nM TTX for 1–2 min substantially inhibited glutamate mediated spontaneous activity. In a set of 18 neurons 8-Br-cGMP reduced the frequency of mEPSCs by 33 ± 9% and Sp-8-Br-PET-cGMPS by 41 ± 11% of the controls (P < 0.01, paired t-test; n = 18) (data not shown). On the other hand these compounds did not change the amplitude of the mEPSCs (99 ± 3% for 8-Br-cGMP and 100 ± 1% for Sp-8-Br-PET-cGMPS). This indicates a presynaptic mechanism of cGMP acting through PKG.
5. cGMP modulates cortical gabaergic synapses at presynaptic terminals To test the effects of cGMP on gabaergic synapses we grew cortical neurons in low density cultures on collagen islands that promoted the formation of autapses, synaptic contacts by a cell on itself. A cell showing the network of processes by calcein staining and the distribution of synaptic vesicles by SVP38 antibody staining is shown in Fig. 4A. Focal perfusion of 0.1 mM 8-Br-cGMP for 30 s elicited up to 99% abolition of the inward currents without inducing any direct membrane conductance change and sodium current change (Fig. 4B 8-Br-cGMP trace). After washout the response recovered in <5 min (Fig. 4B, washout trace). This suggests that this neuron temporally became “silent” when perfused with 8-Br-cGMP. The evoked current was reversibly blocked by 100 M picrotoxin (Fig. 4C), indicating the GABAergic origin of the inhibitory postsynaptic currents (IPSCs). This current was not affected by 100 M APV/20 M CNQX (Fig. 4E). Both sodium and evoked currents were reversibly blocked by 400 nM TTX (Fig. 4D), confirming this was a synaptic current. Similar results were obtained in 26 other neurons for which bicuculline and/or picrotoxin at 100 M totally blocked the IPSCs. IPSCs were inhibited about 72.46% by 100 M 8-Br-cGMP (n = 6), 90.25% by 100 M Sp-8-Br-PET-cGMPS (n = 6) and only minimally (14.77%) by 100 M 8-Br-cAMP. This suggests that the inhibition was due to the activation of PKG and that activation of PKA did not induce the same response.
879
6. cGMP potentiates postsynaptic NMDA responses in visual cortical cells To determine whether the observed effects of cGMP might also have had a postsynaptic component, we observed the effects on responses induced by agonists of specific glutamate receptors in the presence of 0.5–1 M TTX. We iontophoresed NMDA before, during and after iontophoresis of 8-Br-cGMP (20 mM) or Sp-8-Br-PET-cGMPS (5 mM) using multibarrel electrodes near the recorded neurons in visual cortex slices. The NMDA-mediated response was dramatically potentiated by a prior application of 8-Br-cGMP for <1 min by iontophoresis (Fig. 5A). This effect of 8-Br-cGMP was reversed after 2–3 min. To confirm that the cGMP-induced potentiation of NMDA was PKG-dependent, Sp-8-Br-PET-cGMPS was applied and a significant enhancement of the NMDA response was also seen (Fig. 5B). Similar results were seen on eight cells. Overall, the NMDA response was potentiated by a factor of 5.1 ± 3 (P < 0.004) by 8-Br-cGMP and by a factor of 1.9 ± 0.3 (P < 0.002) by Sp-8-Br-PET-cGMPS. With application of 8-Br-cGMP, the membrane potentials showed essentially no change (0.15 ± 1.6 mV (n = 22)). The cGMP analogues also did not change the input resistance of the cells (mean change 1.0 ± 8.9 M (n = 5)). As a further confirmation of the role of PKG in NMDA receptor response changes caused by 8-Br-cGMP and Sp-8-Br-PET-cGMPS, we put the specific PKG inhibitory pseudosubstrate peptide (100 M) in the recording electrode. Under these conditions, iontophoretic application of 100 M Sp-8-Br-PET-cGMPS did not enhance the NMDA response (Fig. 5C). Two other neurons produced the same results. Together these results indicate that cGMP can activate PKG in postsynaptic neurons and that one of the consequences of this is an enhancement in the magnitude of the NMDA receptor response. Because of the complexity of circuitry in the slices we carried out a more detailed analysis of the effects of PKG stimulation using dissociated cells in culture. Similar results were obtained with cultured cortical neurons. After application of 1 mM 8-Br-cGMP for 1 min, the NMDA response increased from 27 pA to 50 pA. After washout, the NMDA response returned to control levels. Application of 100 M Sp-8-Br-PET-cGMPS gave a similar potentiation in the same neuron. However, good recovery was harder to achieve than with 8-Br-cGMP. 12 neurons showed significantly enhanced NMDA responses by 8-Br-cGMP and Sp-8-Br-PET-cGMPS, with increases of 31 ± 6% (P < 0.01) and 37 ± 8% (P < 0.005), respectively. To test whether the effects of cGMP we observed were due to a cross reactivity with a cAMP-dependent protein kinase we also applied cAMP to some of the cells showing a cGMP response. In 10 cells application of 1 mM 8-Br-cAMP gave an NMDA response that increased by 5.2 ± 3%. This
880
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
Fig. 4. cGMP silences GABAergic autapses in single visual cortical neuronal culture. (A) Phase contrast image of a typical visual cortical interneuron cultured for 7 days. (B) In this neuron, evoked inward current was abolished by 0.1 mM 8-Br-cGMP. After washout, the current was recovered. (C) The 0.1 mM picrotoxin blocked this inward current and it also recovered after washout. (D) TTX abolished the IPSCs. (E) The 20 M CNQX/100 mM APV, the glutamate receptors antagonists had no any effect on this inward current.
indicates that the effects of cGMP on the NMDA response are due to activation of a cGMP-dependent kinase rather than a cAMP-dependent kinase. In contrast to NMDA, responses to the non-NMDA glutamate receptor agonists kainate and AMPA were not potentiated by cGMP and PKG activators. NMDA responses
were potentiated by 8-Br-cGMP, but in the same neuron 8-Br-cGMP had no visible effect on AMPA-induced transient and sustained responses. Similarly, 8-Br-cGMP did not change properties of the kainate responses in different neurons. 8-Br-cGMP and Sp-8-Br-PET-cGMPS treated cells showed responses that were 101 ± 1% (n = 23) for kainate
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
881
Fig. 5. cGMP and PKG potentiate postsynaptic NMDA responses in slices of visual cortex. (A) Potentiation of NMDA response by 8-Br-cGMP using multi-barrel iontophoresis. A prior iontophoresis of 8-Br-cGMP for 20 s enhanced the NMDA response, and this returned to control levels after washout (first line). All the base line membrane potentials showed no change during the experiment. The second line in (A) showed that membrane conductance as injections of currents into the recording cell was not changed by 8-Br-cGMP. (B) Sp-8-Br-PET-cGMPS potentiated the NMDA response in another neuron. (C) When PKG peptide inhibitor was infused into the recorded cell, the NMDA response showed little change even when 8-Br-cGMP was applied. Adapted from Wei et al. (2002).
and 96 ± 10% (n = 8) for AMPA as compared with control responses. Every cortical neuron we tested showed an enhancement of the NMDA receptor response induced by either 8-Br-cGMP or Sp-8-Br-PET-cGMPS. When a cell was recorded with 100 M PKG inhibitory peptide in the electrode no enhancement of the NMDA response by PKG activators was obtained. We obtained 98 ± 3% of control responses in seven cells treated with 8-Br-cGMP and 97 ± 7% of control responses in 4 cells treated with Sp-8-Br-PET-cGMPS. These results confirm the findings made in the cortical slices and verify that the inhibitory peptide diffused adequately throughout the cell.
7. cGMP and PKG inhibit whole-cell calcium currents Previous studies in cardiac cells have suggested that cGMP can modulate activity through phosphorylation of calcium channels mediated by PKG (Jiang et al., 2000). This has not been shown for any central neuron and so we investigated whether inhibition of calcium currents may underlie the inhibition of synaptic transmitter release by cGMP analogs.
The recorded neurons were held at −70 mV, and a command pulse (70 mV, 40 ms) was applied to evoke a current through voltage-gated calcium channels. Both 8-Br-cGMP and Sp-8-Br-PET-cGMP were found to inhibit the whole cell calcium currents (IBa ) in the visual cortex neurons. A typical example of the effect of PKG agonists on the neuron is shown in Fig. 6A. Upper traces show that 100 M 8-Br-cGMP substantially suppressed IBa , and that the suppression was reversible. Middle traces from the same cell show that 100 M Sp-8-Br-PET-cGMPS also reversibly inhibited IBa . In contrast, in the same neuron, 1 mM 8-Br-cAMP had no visible effect on IBa (lower traces). I–V curves of IBa from other neurons before and after application of 8-Br-cGMP or Sp-8-Br-PET-cGMPS are illustrated in Fig. 6B. The shape of the I–V curve was not changed by either 8-Br-cGMP or Sp-8-Br-PET-cGMPS. The average amount of reduction of IBa by 8-Br-cGMP and Sp-8-Br-PET-cGMPS was similar (32 ± 5 and 33 ± 4% respectively, P < 0.005, n = 8). Assuming that cGMP and PKG exert similar actions in the terminals as in the perikarya, cGMP and PKG inhibition of presynaptic neurotransmitter release may be mediated by inhibition of calcium channels in the presynaptic nerve terminal.
882
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
Fig. 6. Reduction of whole cell calcium current by cGMP in cultured visual cortical cells. (A) 8-Br-cGMP (upper trace) and Sp-8-Br-PET-cGMPS (middle trace) reduced the whole cell IBa but 8-Br-cAMP (lower trace) showed no effect in the same neuron. (B) I–V curves of whole cell calcium current in control condition, after application of 8-Br-cGMP and after washout (left side). Similar curves for application of Sp-8-Br-PET-cGMPS are shown on the right side. Adapted from Wei et al. (2002).
8. Discussion In the work described above, we have focused on rapid and transient synaptic events modulated by cGMP. At present we do not know whether longer term changes can also be mediated by cGMP, either independent of the events described above or as a consequence of them. We have shown that activation of CNG channels can occur in many CNS neurons. These channels may be one of the main routes for calcium entry into neurons that is voltage-independent. Activation of CNG channels can increase neurotransmitter release and may be an important part of several forms of activity-dependent plasticity (reviewed in Zufall et al., 1997). On the other hand, our studies in cortical slices and
cortical neurons suggests that the major actions of cGMP at synapses are mediated by PKG. The action at presynaptic terminals seems to be to reduce both evoked and spontaneous activity. Much of the reduction in evoked activity is likely to be due to decreases in the activity of the voltage-gated calcium channel, thus limiting the amount of calcium available to promote transmitter release. The reduction in spontaneous activity is probably due to other actions of PKG. We have previously shown that the calcium used for spontaneous release is derived from internal stores rather than the extracellular environment (Han et al., 2001). It is possible that PKG also affects the activity of the IP3 receptors controlling calcium release from the internal stores.
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
At postsynaptic sites we observed an increase in NMDA receptor activity mediated by cGMP, with no detectable changes in AMPA/kainate receptors. Such changes might affect the balance of responses to glutamate and facilitate the generation of NMDA-receptor mediated plasticity. We have not yet fully resolved the actions of cGMP on gaba receptors. We have observed inhibition of gaba receptors in some cells and no effects on other cells (data not shown). This suggests that the ability of cGMP to affect gaba receptors may depend on some other variable that we do not understand at present. Nevertheless, the results do suggest that under some circumstances cGMP can reduce inhibitory input and thus make cells more excitable. It is likely that NO is one of the major regulators of cGMP concentration in the CNS. Whether NO acts as a retrograde transmitter or is acting to provide more lateral signals among cells is not yet known. The other pathway by which cGMP can be regulated is through the activity of specific phosphodiesterases. The activity of some of these can be influenced by cAMP, thus providing an important point of interaction between the two cyclic nucleotides. Other regulators of PDE activity have yet to be elucidated. Overall, our results suggest that cGMP is an important modulator of synaptic function in a variety of CNS neurons. We have outlined some of the pathways by which it can act, but others remain to be identified. The role of cGMP, and NO, in longer term synaptic plasticity also needs further work. Finally, there is increasing evidence that cGMP plays an important role in axonal pathfinding during embryonic development. If sprouting and regeneration use similar pathways, it is essential that we investigate the role of cGMP in neuronal injury and repair.
Acknowledgements Work from this laboratory has been supported by grants from the NIH and by the David Woods Kemper Memorial Foundation and the Foresight Foundation. We thank Nigel, Daw, Gordon Shepherd and Frank Zufall for many stimulating discussions. References Ahmad, I., Leinders-Zufall, T., Kocsis, J.D., Shepherd, G.M., Zufall, F., Barnstable, C.J., 1994. Retinal ganglion cells express a cGMP-gated cation conductance activatable by nitric oxide donors. Neuron 12, 155– 165. Ahmad, I., Redmond, L.J., Barnstable, C.J., 1990. Developmental and tissue-specific expression of the rod photoreceptor cGMP-gated ion channel gene. Biochem. Biophys. Res. Commun. 173, 463–470. Aoki, C., Rhee, J., Lubin, M., Dawson, T.M., 1997. NMDA-R1 subunit of the cerebral cortex co-localizes with neuronal nitric oxide synthase at pre- and post-synaptic sites and in spines. Brain Res. 750, 25–40. Barnstable, C.J., Cyclic nucleotide-gated nonselective cation channels: a multifunctional gene family. In: Siemen, D., Hescheler, J.-J. (Eds.), Nonselective Cation Channels: Pharmacology, Physiology and Biophysics, Birkhauser. Basel 66 (1993) 121–133.
883
Biel, M., Zong, X., Hofmann, F., 1995. Molecular diversity of cyclic nucleotide-gated cation channels. Naunyn-Schmiedeberg’s Arch. Pharmacol. 353, 1–10. Bladen, C., Loewen, D., Vincent, S.R., 1996. Autoradiographic localization of [3 H]-cyclic GMP binding sites in the rat brain. J. Chem. Neuroanant. 10, 287–293. Bloom, F.E., 1976. The role of cyclic nucleotides in central synaptic function. Adv. Biochem. Psychopharmacol. 15, 273–282. Bloom, F.E., 1979. Cyclic nucleotides in central synaptic function. Fed. Proc. 38, 2203–2207. Bradley, J., Zhang, Y., Bakin, R., Lester, H.A., Ronnett, G.V., Zinn, K., 1997. Functional expression of the heteromeric “olfactory” cyclic nucleotide-gated channel in the hippocampus: a potential effector of synaptic plasticity in brain neurons. J. Neurosci. 17, 1993–2005. Butt, E., Geiger, J., Jarchau, T., Lohmann, S.M., Walter, U., 1993. The cGMP-dependent protein kinase-gene, protein, and function. Neurochem. Res. 18, 27–42. Cobbs, W.H., Barkdoll III, A.E., Pugh Jr., E.N., 1985. Cyclic GMP increase photocurrent and light sensitivity of retinal cone. Nature 317, 64–66. Cudeiro, J., Rivadulla, C., 1999. Sight and insight-on the physiologicalrole of nitric oxide in the visual cortex. TINS 22, 109–116. Cudeiro, J., Rivadulla, C., Rodriguez, R., Grieve, K.L., Martinez-Conde, S., Acuna, C., 1997. Actions of compounds manipulating the nitric oxide system in the cat primary visual cortex. J. Physiol. 504, 467–478. el-Husseini, A.E.-D., Bladen, C., Vincent, S.R., 1995. Expression of the olfactory cyclic nucleotide gated channel (CNG1) in the rat brain. Neuroreport 6, 1459–1463. Fesenko, E.E., Kolesnikov, S.S., Lyubarsky, A.L., 1985. Induction by cGMP of cationic conductance in plasma membrane of retinal rod outer segment. Nature 313, 310–313. Finn, J.T., Grunwald, M.E., Yau, K.W., 1996. Cyclic nucleotide-gated ion channels: an extended family with diverse functions. Annu. Rev. Physiol. 58, 395–426. Francis, S.H., Corbin, J.D., 1994. Progress in understanding the mechanism and function of cyclic GMP-dependent protein kinase. Adv. Pharmacol. 26, 115–170. Han, M-H., Kawasaki, A., Wei, J-Y., Barnstable, C.J., 2001. Miniature postsynaptic currents depend on Ca2+ released from internal stores via PLC/IP3 pathway. Neuroreports 12, 2203–2207. Haynes, L.W., Yau, K.W., 1985. Cyclic GMP-sensitive conductance in outer segment membrane of catfish cones. Nature 317, 61–64. Hofmann, F., Dostmann, W., Keilbach, A., Landgraf, W., Ruth, P., 1992. Structure and physiological role of cGMP-dependent protein kinase. Biochem. Biophys. Acta. Mol. Cell Res. 135, 51–60. Jiang, L.H., Gawler, D.J., Hodson, N., Milligan, C.J., Pearson, H.A., Porter, V., Wray, D., 2000. Regulation of cloned cardiac L-type calcium channels by cGMP-dependent protein kinase. J. Biol. Chem. 275, 6135–6143. Kaupp, U.B., 1991. The cyclic nucleotide gated channels of vertebrate photoreceptors and olfactory epithelium. TINS 14, 150–157. Kaupp, U.B., 1995. Family of cyclic nucleotide gated ion channels. Curr. Opin. Neurobiol. 5, 434–442. Kawa, F., Sterling, P., 2002. cGMP modulates spike responses of retinal ganglion cells via a cGMP-gated current. Vis. Neurosci. 19, 373–380. Kingston, P.A., Zufall, F., Barnstable, C.J., 1999. Widespread expression of olfactory cyclic nucleotide-gated channelgenes in rat brain: implications for neuronal signalling. Synapse 32, 1–12. Kingston, P.A., Zufall, F., Barnstable, C.J., 1996. Rat hippocampal neurons express genes for both rod retinal and olfactory cyclic nucleotide-gated channels: novel targets for cAMP/cGMP function. Proc. Natl. Acad. Sci. U.S.A. 93, 10440–10445. Kleene, S.J., 1994. Inhibition of olfactory cyclic nucleotide-activated current by calmodulin antagonist. Br. J. Pharmacol. 111, 469–472. Kramer, R.H., Tibbs, G.R., 1996. Antagonists of cyclic nucleotide-gated channels and molecular mapping of their site of action. J. Neurosci. 16, 1285–1293.
884
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
Krebs, E.G., Beavo, J.A., 1979. Phosphorylation-dephosphorylation of enzymes. Annu. Rev. Biochem. 48, 923–959. Kuo, J.F., Greengard, P., 1969. Cyclic nucleotide-dependent protein kinases. IV. Widespread occurrence of adenosine 3 ,5 -monophosphatedependent protein kinase in various tissues and phyla of the animal kingdom. Proc. Natl. Acad. Sci. U.S.A. 64, 1349–1355. Leinders-Zufall, T., Zufall, F., 1995. Block of cyclic nucleotide-gated channels in salamander olfactory receptor neurons by the guanylyl cyclase inhibitor LY83583. J. Neurophysiol. 74, 2759–2762. Leinders-Zufall, T., Rosenboom, H., Barnstable, C.J., Shepherd, G.M., Zufall, F., 1995. A calcium-permeable cGMP-activated cation conductance in hippocampal neurons. Neuroreport 6, 1761–1765. Lev-Ram, V., Jiang, T., Wood, J., Lawrence, D.S., Tsien, R.Y., 1997. Synergies and coincidence requirements between NO, cGMP, and Ca2+ in the induction of cerebellar long-term depression. Neuron 18, 1025– 1038. Nakamura, T., Gold, G.H., 1987. A cyclic nucleotide-gated conductance in olfactory receptor cilia. Nature 325, 442–444. Samanata Roy, D., Barnstable, C.J., 1999. Temporal and spatial pattern of expression of cyclic nucleotide-gated channels in developing rat visual cortex. Cerebral Cortex 9, 340–347. Stone, T.W., Taylor, D.A., 1977. Microiontophoretic studies of the effects of cylic nucleotides on excitability of neurones in the rat cerebral cortex. J. Physiol. 266, 523–543. Stone, T.W., Taylor, D.A., Bloom, F.E., 1975. Cyclic AMP and cyclic GMP may mediate opposite neuronal responses in the rat cerebral cortex. Science 187, 845–846. Walter, U., 1989. Physiological role of cGMP and cGMP-dependent protein kinase in the cardiovascular system. Rev. Physiol. Biochem. Pharmacol. 113, 41–87. Wang, X., Robinson, P.J., 1997. Cyclic GMP-dependent protein kinase and cellular signaling in the nervous system. J. Neurochem. 68, 443–456.
Wei, J-Y., Cohen, E.D., Genieser, H-G., Barnstable, C.J., 1998. Substituted cGMP analogs can act as selective agonists of the rod photoreceptor cGMP-gated cation channels. J. Mol. Neurosci. 10, 53–64. Wei, J-Y., Jin, X., Cohen, E.D., Daw, N.W., Barnstable, C.J., 2002. cGMP-induced presynaptic depression and postsynaptic facilitation at glutamatergic synapses in visual cortex. Brain Res. 927, 42–54. Wei, J-Y., Roy, D.S., Leconte, L., Barnstable, C.J., 1998. Molecular and pharmacological analysis of cyclic nucleotide-gated channel function in the central nervous system. Prog. Neurobiol. 56, 37–64. Wei, J.-Y., Cohen, E.D., Yan, H.-G., Genieser, Y.-Y., Barnstable, C.J., 1996. Identification of competitive antagonists of the rod photoreceptor cGMP-gated cation channel: -phenyl-1, N2 -etheno substituted cGMP analogues as probes of the cGMP-binding site. Biochemistry 35, 16815–16823. Yau, K.-W., Baylor, D.A., 1989. Cyclic GMP-activated conductance of retinal photoreceptor cells. Annu. Rev. Neurosci. 12, 289–327. Yau, K.-W., 1994. Cyclic nucleotide-gated channel: an expanding new family of ion channels. Proc. Natl. Acad. Sci. U.S.A. 91, 3481– 3483. Yau, K.-W., Nakatani, K., 1985. Light-suppressible, cyclic GMP-sensitive conductance in the plasma membrane of truncated rod outer segment. Nature 317, 252–255. Zagotta, W.N., Siegelbaum, S.A., 1996. Structure and function of cyclic nucleotide-gated channels. Annu. Rev. Neurosci. 19, 235–263. Zhuo, M., Hu, Y., Shultz, C., Kandel, E.R., Hawkins, R.D., 1994. Role of guanylyl cyclase and cGMP-dependent protein kinase in long-term potentiation. Nature 368, 635–639. Zimmermann, A.L., 1995. Cyclic nucleotide gated channels. Curr. Opin. Neurobiol. 5, 296–303. Zufall, F., Shepherd, G.M., Barnstable, C.J., 1997. Cyclic nucleotide-gated channels as regulators of CNS development and plasticity. Curr. Opin. Neurobiol. 7, 404–412.
Modulation of synaptic function by cGMP and cGMP-gated cation channels Colin J. Barnstable∗ , Ji-Ye Wei, Ming-Hu Han Department of Ophthalmology and Visual Science, Yale University School of Medicine, 330 Cedar Street, New Haven, CT 06520-8061, USA Available online 15 June 2004
Abstract Cyclic nucleotide-gated cation channels have been studied intensively in the primary sensory neurons of the visual and olfactory systems. Using both anatomical and physiological methods we have shown that they have a much more widespread distribution in the nervous system. In many retinal ganglion cells cGMP, but not cAMP, activates a non-selective conductance that has many of the properties of CNG channels. As many neurons also contain cGMP-dependent protein kinases (PKGs), we have used a variety of cGMP analogues to distinguish the actions of cGMP. Sp-8-Br-PET-cGMPS is a potent non-hydrolyzable cGMP analogue that is an agonist of PKG. We found that Sp-8-Br-PET-cGMPS acts as a competitive inhibitor of at least the rod CNG channel. Rp-8-Br-cGMPS has shown the opposite effects, namely as an agonist of the rod CNG channel and an inhibitor of PKG. In dissociated cell cultures and slices of rodent visual cortex cGMP had multiple rapid and reversible effects on transmission at glutamatergic synapses. Extracellular application of 8-Br-cGMP or Sp-8-Br-PET-cGMPS reduced stimulus evoked EPSPs in cortical slices. In cortical cultures both analogs reduced the frequency of spontaneous EPSCs, but not their amplitude. The effects on both EPSPs and EPSCs were presynaptic. The effects on evoked EPSPs may be due, in part, to reduced calcium influx through voltage-gated calcium channels. The effects on spontaneous EPSCs may be due, in part, to modulation of calcium fluxes through internal stores. Similar modulations of synaptic transmission have been found at gabaergic synapses. On postsynaptic cells, PKG activation produced a dramatic enhancement of the responses to applied NMDA. No effects were detected on applied AMPA/kainate or GABA. Together the results suggest that cGMP may use multiple mechanisms to modulate synaptic efficacy and that its actions may include regulating synaptic plasticity and the relative strength of excitatory and inhibitory drive through neural pathways. © 2004 Elsevier Ltd. All rights reserved. Keywords: Synaptic function; Modulate synaptic efficacy; cGMP-gated cation channel; cGMP-dependent protein kinase; Glutamatergic synapses; GABAergic synapses
1. Introduction Cyclic nucleotide (cGMP and cAMP) modulation of central synapses has been investigated for decades. Earlier reports focused on the effects of cAMP and cGMP on rapid synaptic modulation (see review by Bloom, 1976 and Bloom, 1979). Later studies concentrated on cGMP or cAMP-induced long-term synaptic potentiation (LTP) or depression (LTD) that is associated with learning and memory (Zhuo et al., 1994; Lev-Ram et al., 1997). It has been suggested that cGMP and cAMP play opposite regulatory roles in excitatory central synapses (Bloom, 1979). Using extracellular recordings and multibarrel ionphoresis in whole animals it was shown that cGMP caused excitation of a majority of identified pyramidal tract neu∗
Corresponding author. Tel.: +1-203-785-4282; fax: +1-203-785-7401. E-mail address: [email protected] (C.J. Barnstable).
0197-0186/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuint.2004.03.018
rons but cAMP had depressive effects on the pyramidal neurons in the mammalian motor cortex (Stone et al., 1975; Stone and Taylor, 1977). In mammalian visual cortex, excitation induced by cGMP in some pyramidal cells has been detected by extracellular recording in brain slices (Cudeiro et al., 1997; Cudeiro and Rivadulla, 1999). However, the above experiments were performed in the whole animal or brain slices that contain polysynaptic networks with both excitatory and inhibitory synaptic connections. This raises the question of whether the facilitation of excitatory synapses by cGMP in the brain slice and/or whole animal is due to direct effects of cGMP on glutamatergic synapses, or through inhibiting adjacent GABAergic neurons. It was originally thought that the cAMP- and cGMPdependent protein kinase families were the primary effectors of these cyclic nucleotides in a variety of physiological processes (Kuo and Greengard, 1969; Krebs and Beavo, 1979; Walter, 1989; Hofmann et al., 1992; Butt et al., 1993;
876
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
Francis and Corbin, 1994; Wang and Robinson, 1997). It now appears, however, that the other major biological effector of cyclic nucleotide action is a class of cyclic nucleotide-gated (CNG) cation channels in the cell membrane (Yau and Baylor, 1989; Kaupp, 1991; Barnstable, 1993; Yau, 1994; Kaupp, 1995; Zimmermann, 1995; Biel et al., 1995; Zagotta and Siegelbaum, 1996; Finn et al., 1996; Zufall et al., 1997). CNG channels were first described in rod photoreceptors (Fesenko et al., 1985; Yau and Nakatani, 1985). Since then, other CNG channels have been identified as responsible for cone photoreceptor transduction (Cobbs et al., 1985; Haynes and Yau, 1985) and olfactory sensory transduction (Nakamura and Gold, 1987). In the work reviewed below we have shown first that CNG channels are widely distributed in the mammalian Central Nervous System (CNS) and second, that cGMP is an important modulator of both glutamatergic and gabaergic synapses in cerebral cortex. 2. CNG channels are widely distributed in the mammalian CNS Our first indication that CNG channels might be expressed in cells other than primary sensory neurons came from in situ hydridization experiments in which labeling was consistently seen over neurons of the inner retina, particularly retinal ganglion cells (Ahmad et al., 1990). Subsequent patch clamp recording from retinal ganglion cells in culture confirmed the presence of a conductance that was activated by cGMP but not cAMP (Ahmad et al., 1994). This conductance was non-selective for sodium or potassium, could be blocked by cadmium and was decreased in the presence of calcium, all properties expected for a CNG channel. We also found that the conductance could be activated by application of nitric oxide donors, suggesting that the activity of the CNG channels might be coupled to Nitric Oxide (NO) signaling pathways. Of even more interest was the observation
that the conductance could be activated by the application of phosphodiesterase inhibitors. This suggests that cGMP production and hydrolysis within retinal ganglion cells, and possibly other CNS neurons, is continuous and that levels of cGMP can be rapidly modulated by altering either synthesis or breakdown. Based on these results, we proposed a model whereby the activity of CNG channels might regulate a number of processes at synaptic terminals (Fig. 1). Reducing the resting membrane potential and increasing calcium concentrations within the terminal would both be expected to increase transmitter release at the terminal and thus act in a facilitatory manner. Although these results were generally accepted as representing events in cultured retinal ganglion cells, it was only more recently that data have been obtained to support the idea that CNG channels might operate in this way in vivo (Kawa and Sterling, 2002). In subsequent work we examined the expression of CNG channels in other regions of the CNS. In hippocampus we obtained clear evidence by both in situ hybridization and patch clamp recording for CNG channel expression in pyramidal neurons of CA1 through CA3 layers and the granule cells of the dentate gyrus (Leinders-Zufall et al., 1995; Kingston et al., 1996). Work from other groups has also established that CNG channels are expressed in the mammalian CNS (el-Husseini et al., 1995; Bradley et al., 1997). Further studies using PCR identified expression of at least the olfactory type of CNG channel expressed in many regions of the rodent CNS (Kingston et al., 1999). Similarly, in situ hybridization studies using visual cortex have detected the expression of CNG channel RNA in many layers (Samanata Roy and Barnstable, 1999). There is an interesting correlation between the spatial expression of CNG channels in mammalian CNS neurons and that of elements of the NO system. This led us to propose that CNG channels might be one of the major effectors of a NO transmitter system. CNG channels have a number of advantages as effectors. First, they are not voltage dependent and thus their activity is independent of other aspects of
Fig. 1. Schematic representation of pathways that might regulate the activity of the cGMP-gated channel in retinal ganglion cells. cGMP synthesis could be stimulated by NO adjacent amcrine cells. Activation of cGMP-gated channels will increase Ca2+ influx and enhance Ca2+ driven processes. At the same time, G-protein coupled receptors activated by neurotransmitters from bipolar and amacrine cell terminals may regulate the activity of one or more PDEs, which control hydrolysis cGMP and thus the activity of the cGMP-gated channels. Modified from Ahmad et al. (1994).
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
neuron activity. Second, they do not show inactivation and can thus provide a continuous response to signals such as NO. Third, they are calcium permeable and thus their activity can readily be coupled to many other cellular processes by modulating cytoplasmic calcium levels.
3. Pharmacological isolation of cGMP effector pathways One of the problems studying the effects of cGMP on neurons is that it can activate protein kinases as well as CNG channels. To allow each pathway of cGMP action to be studied separately, we have investigated a series of agonists and antagonists. Some molecules that interact with CNG channels were first characterized because of actions on other molecules. For example, LY83583 (6-anilino-5.8-quinolinequinone), first characterized as a guanylyl cyclase inhibitor, can block olfactory CNG channels with high potency, which means that it cannot be used to study the role of guanylyl cyclase on CNG channel activity (Leinders-Zufall and Zufall, 1995). W-7 [N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide], a calmodulin (CaM) inhibitor is also a potent inhibitor of olfactory CNG channel independent of its action at the CaM-binding site localized on the cytoplasmic side of the CNG channel (Kleene, 1994). We have investigated the function of a series of substituted cGMP derivatives. It was shown many years ago that 8-Br-cGMP was a good agonist of rod cGMP-gated channels with 10-fold higher potency, increased membrane-permeability and reduced velocity of PDE hydrolysis (Zimmermann, 1995). Wei et al. (1996) identified the first competitive antagonist of the photoreceptor CNG channel that is also an activator of PKG. Sp-8-Br-PET-cGMPS is both
877
membrane-permeant and PDE-resistant and can help elucidate the physiological roles of CNG channels in intact cells (Fig. 2). The properties of -phenyl-1-, N2-ethenosubstituted cGMP analogs are probably due to modification of three key sites that generate high binding affnity but eliminate the capacity for activation. The related compound, Rp-8-Br-PET-cGMPS, is an antagonist of both CNG channels and PKG. Furthermore, a cGMP-analog, Rp-8-Br-cGMPS, has shown the opposite effects on the rod CNG channel and PKG, acting as an agonist of the rod CNG channel and an inhibitor of PKG (Wei et al., 1998). The results suggest that antagonists of the rod CNG channel and PKG are based on different structural features of the ligands. The most important determinants for the rod CNG channel are in the guanine ring, whereas for the kinase they include the isomeric position of the thiophosphate S atom (Rp versus Sp). Interestingly, the olfactory CNG channel shares a similar ligand property with protein kinases, since the ligand Rp-cGMPS can activate the rod CNG channel and is an antagonist of the olfactory CNG channel (Kramer and Tibbs, 1996). The authors made chimeric photoreceptor and olfactory CNG channels by replacing the C-terminal domain, which contains the cyclic nucleotide binding site, and confirmed that this domain determines ligand selectivity and is coupled to channel openings.
4. cGMP modulates cortical glutamatergic synapses at presynaptic terminals To gain an overall impression of the effects of cGMP on cortical neurons we studied pyramidal neurons of visual cortex in brain slices from 2- to 4-week-old rats using
Fig. 2. PET-substituted cGMP analogs interact with the binding domains of rod CNG channels. The syn- and anti-conformations of the analogs have the substituted guanine ring in different orientations relative to the sugar residue. Residue D597 cannot interact with modified N1.N2 sites of Sp-/Rp-8-Br-PET-cGMPS. However, the 8-Br substitution still increases the affinities of binding of these two analogs. Modified from Wei et al. (1998).
878
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
standard whole-cell patch-clamp recording techniques (Wei et al., 2002). Neurons were included in the study if they had resting membrane potentials of −60 to −75 mV and could generate a train of overshooting action potentials in response to intracellular depolarizing current pulses. We focused on the effects of cGMP on excitatory synapses in layers V and VI because these layers are abundant in [3 H] cGMP binding and its related NOS staining (Aoki et al., 1997; Bladen et al., 1996). The membrane-permeable cGMP analog 8-Br-cGMP (100 M) reduced EPSPs recorded in layer V/VI of visual cortex. The EPSP produced in a layer V cell by stimulation of layer II/III cells was reduced by 40% after application of 100 M 8-Br-cGMP for 3 min. A similar result was obtained in layer VI cells. The average reduction by 8-Br-cGMP was 42 ± 10% (P < 0.001, n = 5) in layer V and 36 ± 7% (P < 0.001, n = 6) in layer VI (Fig. 3B). After washout, the peak EPSPs in the two cells recovered. The mean change in
membrane potential induced by 8-Br-cGMP itself was only 0.38 ± 2 mV (n = 20). To investigate whether CNG channels or PKGs were responsible for the cGMP effect, and to determine whether the action of cGMP was pre- or post-synaptic, we bath applied 50–100 M Sp-8-Br-PET-cGMPS. After 4 min of bath application of Sp-8-Br-PET-cGMPS, the EPSP was reduced by 42% (Fig. 3C, left side). Five cells were recorded in layer VI with a reduction of 37% (Fig. 3D, left side). This is the same as the reduction seen with 8-Br-cGMP, indicating that essentially all the effect of cGMP is exerted through the activation of PKG. In the same series of experiments a membrane-impermeable specific PKG inhibitory peptide was infused for 5–10 min into the postsynaptic cell. Under these conditions, the EPSPs were still reduced by extracellular iontophoresis or perfusion of Sp-8-Br-PET-cGMP (Fig. 3C, right side). In a group of four cells recorded under
Fig. 3. cGMP and PKG inhibit evoked EPSPs in slices of visual cortex. (A) Reduction of evoked EPSPs by 8-Br-cGMP in layersV (left side) and VI (right side). EPSPs were evoked by stimulation of layers II/III for recordings in layer V and VI. The membrane potentials are shown in the left columns of each figure. (B) Histograms show the average reduction of EPSPs by 8-Br-cGMP in layers V and VI, and error bars show the standard deviation (n = 5 for layer V and n = 6 for layer VI). (C) The specific PKG activator Sp-8-Br-PET-cGMPS shows a similar reduction of EPSPs (left side). When a PKG peptide inhibitor was infused via the recording electrode to block the postsynaptic effects of PKG, perfusion of Sp-8-Br-PET-cGMPS still produced a reduction of the EPSP evoked in slices of visual cortex (right side). (D) Histograms illustrate the average reduction of EPSPs by Sp-8-Br-PET-cGMPS, with or without PKG peptide inhibitor postsynaptically (n = 5 for without PKG peptide inhibitor and n = 4 for with PKG peptide inhibitor). Adapted from Wei et al. (2002).
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
these conditions, we observed a reduction of 29 ± 8% by Sp-8-Br-PET-cGMP (Fig. 3D, right side). This result suggests that the action of cGMP was on the presynaptic input to the recorded cells. A presynaptic action of cGMP was also observed when cultured cortical neurons were studied (Wei et al., 2002). High density cultures of cortical neurons form abundant synaptic interactions that can be detected in single electrode voltage-clamp recordings as spontaneous EPSCs. Bath application of DNQX (20 M) together with AP5 (100 M) blocked the sEPSCs, thus confirming that the spontaneous synaptic activity studied in these experiments was glutamatergic. In observations of mEPSCs, application of 1 mM 8-Br-cGMP or 100 M Sp-8-Br-PET-cGMPS with or without 400 nM TTX for 1–2 min substantially inhibited glutamate mediated spontaneous activity. In a set of 18 neurons 8-Br-cGMP reduced the frequency of mEPSCs by 33 ± 9% and Sp-8-Br-PET-cGMPS by 41 ± 11% of the controls (P < 0.01, paired t-test; n = 18) (data not shown). On the other hand these compounds did not change the amplitude of the mEPSCs (99 ± 3% for 8-Br-cGMP and 100 ± 1% for Sp-8-Br-PET-cGMPS). This indicates a presynaptic mechanism of cGMP acting through PKG.
5. cGMP modulates cortical gabaergic synapses at presynaptic terminals To test the effects of cGMP on gabaergic synapses we grew cortical neurons in low density cultures on collagen islands that promoted the formation of autapses, synaptic contacts by a cell on itself. A cell showing the network of processes by calcein staining and the distribution of synaptic vesicles by SVP38 antibody staining is shown in Fig. 4A. Focal perfusion of 0.1 mM 8-Br-cGMP for 30 s elicited up to 99% abolition of the inward currents without inducing any direct membrane conductance change and sodium current change (Fig. 4B 8-Br-cGMP trace). After washout the response recovered in <5 min (Fig. 4B, washout trace). This suggests that this neuron temporally became “silent” when perfused with 8-Br-cGMP. The evoked current was reversibly blocked by 100 M picrotoxin (Fig. 4C), indicating the GABAergic origin of the inhibitory postsynaptic currents (IPSCs). This current was not affected by 100 M APV/20 M CNQX (Fig. 4E). Both sodium and evoked currents were reversibly blocked by 400 nM TTX (Fig. 4D), confirming this was a synaptic current. Similar results were obtained in 26 other neurons for which bicuculline and/or picrotoxin at 100 M totally blocked the IPSCs. IPSCs were inhibited about 72.46% by 100 M 8-Br-cGMP (n = 6), 90.25% by 100 M Sp-8-Br-PET-cGMPS (n = 6) and only minimally (14.77%) by 100 M 8-Br-cAMP. This suggests that the inhibition was due to the activation of PKG and that activation of PKA did not induce the same response.
879
6. cGMP potentiates postsynaptic NMDA responses in visual cortical cells To determine whether the observed effects of cGMP might also have had a postsynaptic component, we observed the effects on responses induced by agonists of specific glutamate receptors in the presence of 0.5–1 M TTX. We iontophoresed NMDA before, during and after iontophoresis of 8-Br-cGMP (20 mM) or Sp-8-Br-PET-cGMPS (5 mM) using multibarrel electrodes near the recorded neurons in visual cortex slices. The NMDA-mediated response was dramatically potentiated by a prior application of 8-Br-cGMP for <1 min by iontophoresis (Fig. 5A). This effect of 8-Br-cGMP was reversed after 2–3 min. To confirm that the cGMP-induced potentiation of NMDA was PKG-dependent, Sp-8-Br-PET-cGMPS was applied and a significant enhancement of the NMDA response was also seen (Fig. 5B). Similar results were seen on eight cells. Overall, the NMDA response was potentiated by a factor of 5.1 ± 3 (P < 0.004) by 8-Br-cGMP and by a factor of 1.9 ± 0.3 (P < 0.002) by Sp-8-Br-PET-cGMPS. With application of 8-Br-cGMP, the membrane potentials showed essentially no change (0.15 ± 1.6 mV (n = 22)). The cGMP analogues also did not change the input resistance of the cells (mean change 1.0 ± 8.9 M (n = 5)). As a further confirmation of the role of PKG in NMDA receptor response changes caused by 8-Br-cGMP and Sp-8-Br-PET-cGMPS, we put the specific PKG inhibitory pseudosubstrate peptide (100 M) in the recording electrode. Under these conditions, iontophoretic application of 100 M Sp-8-Br-PET-cGMPS did not enhance the NMDA response (Fig. 5C). Two other neurons produced the same results. Together these results indicate that cGMP can activate PKG in postsynaptic neurons and that one of the consequences of this is an enhancement in the magnitude of the NMDA receptor response. Because of the complexity of circuitry in the slices we carried out a more detailed analysis of the effects of PKG stimulation using dissociated cells in culture. Similar results were obtained with cultured cortical neurons. After application of 1 mM 8-Br-cGMP for 1 min, the NMDA response increased from 27 pA to 50 pA. After washout, the NMDA response returned to control levels. Application of 100 M Sp-8-Br-PET-cGMPS gave a similar potentiation in the same neuron. However, good recovery was harder to achieve than with 8-Br-cGMP. 12 neurons showed significantly enhanced NMDA responses by 8-Br-cGMP and Sp-8-Br-PET-cGMPS, with increases of 31 ± 6% (P < 0.01) and 37 ± 8% (P < 0.005), respectively. To test whether the effects of cGMP we observed were due to a cross reactivity with a cAMP-dependent protein kinase we also applied cAMP to some of the cells showing a cGMP response. In 10 cells application of 1 mM 8-Br-cAMP gave an NMDA response that increased by 5.2 ± 3%. This
880
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
Fig. 4. cGMP silences GABAergic autapses in single visual cortical neuronal culture. (A) Phase contrast image of a typical visual cortical interneuron cultured for 7 days. (B) In this neuron, evoked inward current was abolished by 0.1 mM 8-Br-cGMP. After washout, the current was recovered. (C) The 0.1 mM picrotoxin blocked this inward current and it also recovered after washout. (D) TTX abolished the IPSCs. (E) The 20 M CNQX/100 mM APV, the glutamate receptors antagonists had no any effect on this inward current.
indicates that the effects of cGMP on the NMDA response are due to activation of a cGMP-dependent kinase rather than a cAMP-dependent kinase. In contrast to NMDA, responses to the non-NMDA glutamate receptor agonists kainate and AMPA were not potentiated by cGMP and PKG activators. NMDA responses
were potentiated by 8-Br-cGMP, but in the same neuron 8-Br-cGMP had no visible effect on AMPA-induced transient and sustained responses. Similarly, 8-Br-cGMP did not change properties of the kainate responses in different neurons. 8-Br-cGMP and Sp-8-Br-PET-cGMPS treated cells showed responses that were 101 ± 1% (n = 23) for kainate
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
881
Fig. 5. cGMP and PKG potentiate postsynaptic NMDA responses in slices of visual cortex. (A) Potentiation of NMDA response by 8-Br-cGMP using multi-barrel iontophoresis. A prior iontophoresis of 8-Br-cGMP for 20 s enhanced the NMDA response, and this returned to control levels after washout (first line). All the base line membrane potentials showed no change during the experiment. The second line in (A) showed that membrane conductance as injections of currents into the recording cell was not changed by 8-Br-cGMP. (B) Sp-8-Br-PET-cGMPS potentiated the NMDA response in another neuron. (C) When PKG peptide inhibitor was infused into the recorded cell, the NMDA response showed little change even when 8-Br-cGMP was applied. Adapted from Wei et al. (2002).
and 96 ± 10% (n = 8) for AMPA as compared with control responses. Every cortical neuron we tested showed an enhancement of the NMDA receptor response induced by either 8-Br-cGMP or Sp-8-Br-PET-cGMPS. When a cell was recorded with 100 M PKG inhibitory peptide in the electrode no enhancement of the NMDA response by PKG activators was obtained. We obtained 98 ± 3% of control responses in seven cells treated with 8-Br-cGMP and 97 ± 7% of control responses in 4 cells treated with Sp-8-Br-PET-cGMPS. These results confirm the findings made in the cortical slices and verify that the inhibitory peptide diffused adequately throughout the cell.
7. cGMP and PKG inhibit whole-cell calcium currents Previous studies in cardiac cells have suggested that cGMP can modulate activity through phosphorylation of calcium channels mediated by PKG (Jiang et al., 2000). This has not been shown for any central neuron and so we investigated whether inhibition of calcium currents may underlie the inhibition of synaptic transmitter release by cGMP analogs.
The recorded neurons were held at −70 mV, and a command pulse (70 mV, 40 ms) was applied to evoke a current through voltage-gated calcium channels. Both 8-Br-cGMP and Sp-8-Br-PET-cGMP were found to inhibit the whole cell calcium currents (IBa ) in the visual cortex neurons. A typical example of the effect of PKG agonists on the neuron is shown in Fig. 6A. Upper traces show that 100 M 8-Br-cGMP substantially suppressed IBa , and that the suppression was reversible. Middle traces from the same cell show that 100 M Sp-8-Br-PET-cGMPS also reversibly inhibited IBa . In contrast, in the same neuron, 1 mM 8-Br-cAMP had no visible effect on IBa (lower traces). I–V curves of IBa from other neurons before and after application of 8-Br-cGMP or Sp-8-Br-PET-cGMPS are illustrated in Fig. 6B. The shape of the I–V curve was not changed by either 8-Br-cGMP or Sp-8-Br-PET-cGMPS. The average amount of reduction of IBa by 8-Br-cGMP and Sp-8-Br-PET-cGMPS was similar (32 ± 5 and 33 ± 4% respectively, P < 0.005, n = 8). Assuming that cGMP and PKG exert similar actions in the terminals as in the perikarya, cGMP and PKG inhibition of presynaptic neurotransmitter release may be mediated by inhibition of calcium channels in the presynaptic nerve terminal.
882
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
Fig. 6. Reduction of whole cell calcium current by cGMP in cultured visual cortical cells. (A) 8-Br-cGMP (upper trace) and Sp-8-Br-PET-cGMPS (middle trace) reduced the whole cell IBa but 8-Br-cAMP (lower trace) showed no effect in the same neuron. (B) I–V curves of whole cell calcium current in control condition, after application of 8-Br-cGMP and after washout (left side). Similar curves for application of Sp-8-Br-PET-cGMPS are shown on the right side. Adapted from Wei et al. (2002).
8. Discussion In the work described above, we have focused on rapid and transient synaptic events modulated by cGMP. At present we do not know whether longer term changes can also be mediated by cGMP, either independent of the events described above or as a consequence of them. We have shown that activation of CNG channels can occur in many CNS neurons. These channels may be one of the main routes for calcium entry into neurons that is voltage-independent. Activation of CNG channels can increase neurotransmitter release and may be an important part of several forms of activity-dependent plasticity (reviewed in Zufall et al., 1997). On the other hand, our studies in cortical slices and
cortical neurons suggests that the major actions of cGMP at synapses are mediated by PKG. The action at presynaptic terminals seems to be to reduce both evoked and spontaneous activity. Much of the reduction in evoked activity is likely to be due to decreases in the activity of the voltage-gated calcium channel, thus limiting the amount of calcium available to promote transmitter release. The reduction in spontaneous activity is probably due to other actions of PKG. We have previously shown that the calcium used for spontaneous release is derived from internal stores rather than the extracellular environment (Han et al., 2001). It is possible that PKG also affects the activity of the IP3 receptors controlling calcium release from the internal stores.
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
At postsynaptic sites we observed an increase in NMDA receptor activity mediated by cGMP, with no detectable changes in AMPA/kainate receptors. Such changes might affect the balance of responses to glutamate and facilitate the generation of NMDA-receptor mediated plasticity. We have not yet fully resolved the actions of cGMP on gaba receptors. We have observed inhibition of gaba receptors in some cells and no effects on other cells (data not shown). This suggests that the ability of cGMP to affect gaba receptors may depend on some other variable that we do not understand at present. Nevertheless, the results do suggest that under some circumstances cGMP can reduce inhibitory input and thus make cells more excitable. It is likely that NO is one of the major regulators of cGMP concentration in the CNS. Whether NO acts as a retrograde transmitter or is acting to provide more lateral signals among cells is not yet known. The other pathway by which cGMP can be regulated is through the activity of specific phosphodiesterases. The activity of some of these can be influenced by cAMP, thus providing an important point of interaction between the two cyclic nucleotides. Other regulators of PDE activity have yet to be elucidated. Overall, our results suggest that cGMP is an important modulator of synaptic function in a variety of CNS neurons. We have outlined some of the pathways by which it can act, but others remain to be identified. The role of cGMP, and NO, in longer term synaptic plasticity also needs further work. Finally, there is increasing evidence that cGMP plays an important role in axonal pathfinding during embryonic development. If sprouting and regeneration use similar pathways, it is essential that we investigate the role of cGMP in neuronal injury and repair.
Acknowledgements Work from this laboratory has been supported by grants from the NIH and by the David Woods Kemper Memorial Foundation and the Foresight Foundation. We thank Nigel, Daw, Gordon Shepherd and Frank Zufall for many stimulating discussions. References Ahmad, I., Leinders-Zufall, T., Kocsis, J.D., Shepherd, G.M., Zufall, F., Barnstable, C.J., 1994. Retinal ganglion cells express a cGMP-gated cation conductance activatable by nitric oxide donors. Neuron 12, 155– 165. Ahmad, I., Redmond, L.J., Barnstable, C.J., 1990. Developmental and tissue-specific expression of the rod photoreceptor cGMP-gated ion channel gene. Biochem. Biophys. Res. Commun. 173, 463–470. Aoki, C., Rhee, J., Lubin, M., Dawson, T.M., 1997. NMDA-R1 subunit of the cerebral cortex co-localizes with neuronal nitric oxide synthase at pre- and post-synaptic sites and in spines. Brain Res. 750, 25–40. Barnstable, C.J., Cyclic nucleotide-gated nonselective cation channels: a multifunctional gene family. In: Siemen, D., Hescheler, J.-J. (Eds.), Nonselective Cation Channels: Pharmacology, Physiology and Biophysics, Birkhauser. Basel 66 (1993) 121–133.
883
Biel, M., Zong, X., Hofmann, F., 1995. Molecular diversity of cyclic nucleotide-gated cation channels. Naunyn-Schmiedeberg’s Arch. Pharmacol. 353, 1–10. Bladen, C., Loewen, D., Vincent, S.R., 1996. Autoradiographic localization of [3 H]-cyclic GMP binding sites in the rat brain. J. Chem. Neuroanant. 10, 287–293. Bloom, F.E., 1976. The role of cyclic nucleotides in central synaptic function. Adv. Biochem. Psychopharmacol. 15, 273–282. Bloom, F.E., 1979. Cyclic nucleotides in central synaptic function. Fed. Proc. 38, 2203–2207. Bradley, J., Zhang, Y., Bakin, R., Lester, H.A., Ronnett, G.V., Zinn, K., 1997. Functional expression of the heteromeric “olfactory” cyclic nucleotide-gated channel in the hippocampus: a potential effector of synaptic plasticity in brain neurons. J. Neurosci. 17, 1993–2005. Butt, E., Geiger, J., Jarchau, T., Lohmann, S.M., Walter, U., 1993. The cGMP-dependent protein kinase-gene, protein, and function. Neurochem. Res. 18, 27–42. Cobbs, W.H., Barkdoll III, A.E., Pugh Jr., E.N., 1985. Cyclic GMP increase photocurrent and light sensitivity of retinal cone. Nature 317, 64–66. Cudeiro, J., Rivadulla, C., 1999. Sight and insight-on the physiologicalrole of nitric oxide in the visual cortex. TINS 22, 109–116. Cudeiro, J., Rivadulla, C., Rodriguez, R., Grieve, K.L., Martinez-Conde, S., Acuna, C., 1997. Actions of compounds manipulating the nitric oxide system in the cat primary visual cortex. J. Physiol. 504, 467–478. el-Husseini, A.E.-D., Bladen, C., Vincent, S.R., 1995. Expression of the olfactory cyclic nucleotide gated channel (CNG1) in the rat brain. Neuroreport 6, 1459–1463. Fesenko, E.E., Kolesnikov, S.S., Lyubarsky, A.L., 1985. Induction by cGMP of cationic conductance in plasma membrane of retinal rod outer segment. Nature 313, 310–313. Finn, J.T., Grunwald, M.E., Yau, K.W., 1996. Cyclic nucleotide-gated ion channels: an extended family with diverse functions. Annu. Rev. Physiol. 58, 395–426. Francis, S.H., Corbin, J.D., 1994. Progress in understanding the mechanism and function of cyclic GMP-dependent protein kinase. Adv. Pharmacol. 26, 115–170. Han, M-H., Kawasaki, A., Wei, J-Y., Barnstable, C.J., 2001. Miniature postsynaptic currents depend on Ca2+ released from internal stores via PLC/IP3 pathway. Neuroreports 12, 2203–2207. Haynes, L.W., Yau, K.W., 1985. Cyclic GMP-sensitive conductance in outer segment membrane of catfish cones. Nature 317, 61–64. Hofmann, F., Dostmann, W., Keilbach, A., Landgraf, W., Ruth, P., 1992. Structure and physiological role of cGMP-dependent protein kinase. Biochem. Biophys. Acta. Mol. Cell Res. 135, 51–60. Jiang, L.H., Gawler, D.J., Hodson, N., Milligan, C.J., Pearson, H.A., Porter, V., Wray, D., 2000. Regulation of cloned cardiac L-type calcium channels by cGMP-dependent protein kinase. J. Biol. Chem. 275, 6135–6143. Kaupp, U.B., 1991. The cyclic nucleotide gated channels of vertebrate photoreceptors and olfactory epithelium. TINS 14, 150–157. Kaupp, U.B., 1995. Family of cyclic nucleotide gated ion channels. Curr. Opin. Neurobiol. 5, 434–442. Kawa, F., Sterling, P., 2002. cGMP modulates spike responses of retinal ganglion cells via a cGMP-gated current. Vis. Neurosci. 19, 373–380. Kingston, P.A., Zufall, F., Barnstable, C.J., 1999. Widespread expression of olfactory cyclic nucleotide-gated channelgenes in rat brain: implications for neuronal signalling. Synapse 32, 1–12. Kingston, P.A., Zufall, F., Barnstable, C.J., 1996. Rat hippocampal neurons express genes for both rod retinal and olfactory cyclic nucleotide-gated channels: novel targets for cAMP/cGMP function. Proc. Natl. Acad. Sci. U.S.A. 93, 10440–10445. Kleene, S.J., 1994. Inhibition of olfactory cyclic nucleotide-activated current by calmodulin antagonist. Br. J. Pharmacol. 111, 469–472. Kramer, R.H., Tibbs, G.R., 1996. Antagonists of cyclic nucleotide-gated channels and molecular mapping of their site of action. J. Neurosci. 16, 1285–1293.
884
C.J. Barnstable et al. / Neurochemistry International 45 (2004) 875–884
Krebs, E.G., Beavo, J.A., 1979. Phosphorylation-dephosphorylation of enzymes. Annu. Rev. Biochem. 48, 923–959. Kuo, J.F., Greengard, P., 1969. Cyclic nucleotide-dependent protein kinases. IV. Widespread occurrence of adenosine 3 ,5 -monophosphatedependent protein kinase in various tissues and phyla of the animal kingdom. Proc. Natl. Acad. Sci. U.S.A. 64, 1349–1355. Leinders-Zufall, T., Zufall, F., 1995. Block of cyclic nucleotide-gated channels in salamander olfactory receptor neurons by the guanylyl cyclase inhibitor LY83583. J. Neurophysiol. 74, 2759–2762. Leinders-Zufall, T., Rosenboom, H., Barnstable, C.J., Shepherd, G.M., Zufall, F., 1995. A calcium-permeable cGMP-activated cation conductance in hippocampal neurons. Neuroreport 6, 1761–1765. Lev-Ram, V., Jiang, T., Wood, J., Lawrence, D.S., Tsien, R.Y., 1997. Synergies and coincidence requirements between NO, cGMP, and Ca2+ in the induction of cerebellar long-term depression. Neuron 18, 1025– 1038. Nakamura, T., Gold, G.H., 1987. A cyclic nucleotide-gated conductance in olfactory receptor cilia. Nature 325, 442–444. Samanata Roy, D., Barnstable, C.J., 1999. Temporal and spatial pattern of expression of cyclic nucleotide-gated channels in developing rat visual cortex. Cerebral Cortex 9, 340–347. Stone, T.W., Taylor, D.A., 1977. Microiontophoretic studies of the effects of cylic nucleotides on excitability of neurones in the rat cerebral cortex. J. Physiol. 266, 523–543. Stone, T.W., Taylor, D.A., Bloom, F.E., 1975. Cyclic AMP and cyclic GMP may mediate opposite neuronal responses in the rat cerebral cortex. Science 187, 845–846. Walter, U., 1989. Physiological role of cGMP and cGMP-dependent protein kinase in the cardiovascular system. Rev. Physiol. Biochem. Pharmacol. 113, 41–87. Wang, X., Robinson, P.J., 1997. Cyclic GMP-dependent protein kinase and cellular signaling in the nervous system. J. Neurochem. 68, 443–456.
Wei, J-Y., Cohen, E.D., Genieser, H-G., Barnstable, C.J., 1998. Substituted cGMP analogs can act as selective agonists of the rod photoreceptor cGMP-gated cation channels. J. Mol. Neurosci. 10, 53–64. Wei, J-Y., Jin, X., Cohen, E.D., Daw, N.W., Barnstable, C.J., 2002. cGMP-induced presynaptic depression and postsynaptic facilitation at glutamatergic synapses in visual cortex. Brain Res. 927, 42–54. Wei, J-Y., Roy, D.S., Leconte, L., Barnstable, C.J., 1998. Molecular and pharmacological analysis of cyclic nucleotide-gated channel function in the central nervous system. Prog. Neurobiol. 56, 37–64. Wei, J.-Y., Cohen, E.D., Yan, H.-G., Genieser, Y.-Y., Barnstable, C.J., 1996. Identification of competitive antagonists of the rod photoreceptor cGMP-gated cation channel: -phenyl-1, N2 -etheno substituted cGMP analogues as probes of the cGMP-binding site. Biochemistry 35, 16815–16823. Yau, K.-W., Baylor, D.A., 1989. Cyclic GMP-activated conductance of retinal photoreceptor cells. Annu. Rev. Neurosci. 12, 289–327. Yau, K.-W., 1994. Cyclic nucleotide-gated channel: an expanding new family of ion channels. Proc. Natl. Acad. Sci. U.S.A. 91, 3481– 3483. Yau, K.-W., Nakatani, K., 1985. Light-suppressible, cyclic GMP-sensitive conductance in the plasma membrane of truncated rod outer segment. Nature 317, 252–255. Zagotta, W.N., Siegelbaum, S.A., 1996. Structure and function of cyclic nucleotide-gated channels. Annu. Rev. Neurosci. 19, 235–263. Zhuo, M., Hu, Y., Shultz, C., Kandel, E.R., Hawkins, R.D., 1994. Role of guanylyl cyclase and cGMP-dependent protein kinase in long-term potentiation. Nature 368, 635–639. Zimmermann, A.L., 1995. Cyclic nucleotide gated channels. Curr. Opin. Neurobiol. 5, 296–303. Zufall, F., Shepherd, G.M., Barnstable, C.J., 1997. Cyclic nucleotide-gated channels as regulators of CNS development and plasticity. Curr. Opin. Neurobiol. 7, 404–412.