TRPγ Channels Are Inhibited by cAMP and Contribute to Pacemaking in Neurosecretory Insect Neurons*

From a neuronal cDNA library of the cockroach Periplaneta americana we isolated a 3585-bp cDNA sequence encoding Periplaneta transient receptor potential γ (pTRPγ), a protein of 1194 amino acids showing 65% identity to the orthologous Drosophila channel protein dTRPγ. Heterologous expression of pTRPγ in HEK293 cells produced a constitutively active, non-selective cation channel with a Ca2+:Na+ permeability ratio of 2. In contrast to dTRPγ-mediated currents, pTRPγ currents were partially inhibited by 8-bromo-cAMP, and this effect was not mediated by protein kinase A (PKA) activation. pTRPγb, a truncated pTRPγ splice variant missing most of the C terminus, was insensitive to 8-bromo-cAMP. Thus, the critical cAMP-binding site seems to be located in the C-terminal part of pTRPγ, although there is no common cAMP-binding consensus sequence. While dTRPγ is only expressed in the photoreceptors, pTRPγ is expressed throughout the nervous system. In particular it is expressed in dorsal unpaired median (DUM) neurons. In these octopamine-releasing, neurosecretory cells a Ca2+ background current contributing to pacemaker activity was found to be up-regulated by the reduction of cAMP level. In addition, the Ca2+ background current was inhibited by LOE-908, 2-APB, and La3+, which similarly affected the pTRPγ current. We thus propose that the pTRPγ protein is involved in forming the channel passing the Ca2+ pakemaking background current in DUM neurons.

Proteins of the TRP 2 (transient receptor potential) family form cation-selective and Ca 2ϩ -permeable ion channels with multiple functions (1). The first TRP channel was described in Drosophila photoreceptors where mutants lacking this channel showed a transient receptor potential upon continuous light stimulation (2). In the fly photoreceptors two further TRP proteins, TRPL and TRP␥ (3), contribute to visual transduction (4). TRP proteins have six transmembrane segments, and most of them probably assemble to homo-or heterotetramers to form ion channels.
The superfamily of TRP channels has been grouped into several subfamilies (1,5). Members of the mammalian TRPC (canonical TRP) family show the highest homology with the Drosophila TRP/TRPL/TRP␥ variants. These channels appear to be receptor-operated in that they are activated by a variety of stimuli that lead to activation of phospholipase C (PLC).
The modulation of pacemaker conductances is a basic principle to adjust the neuronal activity to the physiological requirements. Certain peptide hormones affect ion channels involved in the regulation of pacemaking. Previous investigations on identified insect neurosecretory neurons, namely efferent dorsal unpaired median (DUM) neurons in the cockroach Periplaneta americana, have shown that a voltage-independent Ca 2ϩ background current providing for non-capacitative Ca 2ϩ entry (NCCE) changes the spike frequency of these cells. For example, up-regulation of this current by the Periplaneta adipokinetic hormone AKH I (pAKH I: pQ-V-N-F-S-P-N-W⅐NH 2 , also named neurohormone D (6)), accelerates spiking (7,8). A detailed analysis of the signal transduction mechanism, by which AKH I binding to the receptor leads to potentiation of Ca 2ϩ background current, revealed a complex pattern of events including activation of PLC but no apparent contribution from intracellular Ca 2ϩ stores (8). The final event of the signal transduction cascade was the down-regulation of the cAMP level, which directly enhanced the background current. Up-regulation of the cAMP level in DUM neurons attenuates the Ca 2ϩ background current and thus the spike frequency (9).
We were interested in identifying the molecular basis for the Ca 2ϩ background current in DUM neurons and thus looked for a channel that (i) may conduct a voltage-independent Ca 2ϩ current and (ii) that is sensitive to cAMP. Since TRP channels are known to fulfill the first criterion, we screened a Periplaneta cDNA library for putative members of the TRPC family. We succeded in finding a Periplaneta homolog to the Drosophila TRP␥ channel. In the present study, we report an unusual regulation of this TRP channel. We demonstrate that the Periplaneta but not the Drosophila TRP␥ channel is downregulated by cAMP. Based on an analysis of the pharmacological profile of the Periplaneta TRP␥ channel we propose that it is involved in forming the channel that conducts the Ca 2ϩ background current in DUM neurons.
RT-PCR Analysis of Periplaneta Tissues-A nested RT-PCR approach was developed for testing ganglion-specific expression and single-cell analysis of isolated DUM neurons. The following intronspanning PCR primers were designed: Sc-F1, 5Ј-cagaaagaacgtcggctcat-3Ј; Sc-F2, 5Ј-caactgcccactgtagcaga-3Ј; Sc-B1, 5Ј-gggactgttgctgtggctat-3Ј; Sc-B2, 5Ј-cagagtcttccgacctgctc-3Ј. For single-cell RT-PCR analysis of pTRP␥ expression the cytosol of single DUM neurons was harvested with patch pipettes and directly transferred to the RT reaction mixture. RT reaction was performed with either Superscript II reverse transcriptase (Invitrogen) and ganglion RNA or with Sensiscript reverse transcriptase (Qiagen) and isolated cytosol of single DUM neurons as described by the manufacturer, respectively. The two nested PCR reactions were performed with AmpliTaq Gold polymerase (Applied Biosystems) with initial 6 min at 94°C enzyme activation followed by 40 cycles 30 s at 94°C; 30 s at 55°C; 1 min at 72°C. PCR products of the second PCR reaction were visualized on a 2% agarose gel. As a positive control we amplified part of a ubiquitously expressed actin gene from P. americana.
Construction of a pTRP␥ pIRES2-EGFP Expression Vector (HEK293 Expression)-Full-length pTRP␥ was amplified using Advantage TaqII DNA polymerase mixture (Clontech, Palo Alto, CA) and ganglion cDNA using the following three primers: FL-for1, 5Ј-aaaggcctcgagatgagtgacaattatgacgtg-3Ј; FL-for2, 5Ј-acagcatcctcgagatgatggaggaggagaacgtc-3Ј; and FL-rev, 5Ј-tctgcgctcgagttatagccagcctgctgacat-3Ј. PCR was performed for 35 cycles with 30 s at 94°C, 1 min at 64°C (or 58°C during initial two cycles), and 12 min at 68°C. Finally, PCR products were cloned into the pIRES2-EGFP vector using XhoI sites incorporated into primer sequences (shown in italics). Among these clones a novel splice variant pTRP␥b was found, harboring a novel exon and introducing a stop codon resulting in a shorter C-terminaly truncated splice variant.
Production and Specificity of pTRP␥ Antibodies-Two pTRP␥ antisera were raised in rabbits against the synthetic peptides (1) QPVSGH-NMSAGW and (2) QNKSRNG representing two different C-terminal motifs. The peptides were cross-linked to keyhole limpet hemocyanin (KLH) by means of glutaraldehyde. The peptides, the peptide-KLH conjugates, and the antiserum (1) were produced by Sigma Genosys (Cambridge, UK). The specificity of antiserum (1) was tested by Western blot (Fig. 2B). For immunocytochemical control of antiserum (2) we used the preimmune serum (Fig. 3, B and E), and we performed experiments without the primary antibody (data not shown). In both control experiments there was no staining.
Western Blot-Freshly isolated P. americana thoracic and abdominal ganglia from five animals were homogenized in 100 l of lysis buffer (50 mM Tris, 150 mM NaCl, 5 mM EDTA, 1 mM EGTA, 1% Triton X-100, 1 tablet Complete protease inhibitor mixture (Roche Diagnostics) per 50 ml). After electrophoretical separation of 20 g protein per lane (SDS-PAGE) and blotting on polyvinylidene difluoride transfer membranes (Millipore), standard wash and detection procedures were used. Anti-pTRP␥ antibody (1) was 1:2000 diluted, and peroxidase-coupled secondary antibodies (anti-rabbit IgG or anti-mouse IgG) were applied in 1:10,000 dilution, followed by chemoluminescence detection with the ECL Plus kit (Amersham Biosciences). For control we used pTRP␥transfected and mock-transfected HEK293 cells. 2 days after transfection about 1 ϫ 10 5 cells were harvested and immediately lysed in 200 ml of lysis buffer, and 10 l were used for SDS-PAGE. To exclude unspecific bands due to antibodies directed against KLH, control blots were performed with antiserum (5 l) that was preincubated with 100 g of KLH protein in 100 l of phosphate-buffered saline, pH 7.4). This procedure prevented the Western blot detection of KLH protein by the serum but did not affect the 129 kDa signal in Periplanea tissues or in transfected HEK293 cells.
Immunocytochemistry-Adult P. americana L. were taken from laboratory colonies at the University of Jena (maintained at 27°C under a 12 h light/12 h dark regime). Animals were anesthetized by cooling under crushed ice and decapitated. Brains and the sixth abdominal ganglia were dissected from animals of both sexes under ice-chilled Tris-HCl saline and transferred into freshly prepared fixative: 4% formaldehyde in 0.1 M Millonigs buffer (pH ϭ 7.3-7.4) overnight at room temperature. The preparations were washed in Tris-HCl buffer (146 mM NaCl, 50 mM Tris-OH, pH ϭ 7.4, for 12 h at 4°C on a shaker. 40-m-thick sections of agar-embedded tissues were cut on a vibratome (Technical Products, St. Louis, MO). The washed free-floating vibratome sections were then incubated in rabbit anti-TRP polyclonal antibody (cf. above) diluted 1:1000 in a solution of Tris-HCl buffer containing 2% normal goat serum, 0.25% Triton-X, 3% skim milk powder, and 0.25% bovine serum albumin (MPB). Subsequently, the tissue was washed in MPB overnight and incubated in the Cy3-tagged secondary antibody (Jackson Immu-noResearch Laboratories, Inc., West Grove, PA) for 3 h at a dilution of 1:800 in MPB. The vibratome sections were mounted and embedded in Mowiol. For the photographic documentation we used the bright-field optics on a Zeiss Axiophot microscope, a Hamamatsu digital camera C4742-95, and the software OpenLab (Improvision Ltd., Coventry, England). HEK293 cells growing on coverslips were treated as the vibratome sections.
Ion currents in HEK293 cells and in DUM neurons (isolated according to Ref. 7) were measured at room temperature using whole-cell patch clamp with appropriate compensation of series resistance and of capacitive and leakage currents. Some experiments were performed in the inside-out configuration. Pipettes having resistances of 2-4 M⍀ (HEK293 cells) or 0.5-0.8 M⍀ (DUM neurons) were pulled from borosilicate capillaries. Current measurements and data acquisition were performed using an EPC9 patch clamp amplifier controlled by PULSE software (HEKA Elektronik, Lambrecht, Germany).
HEK293 Cells-The pipette solution contained (in mM) 140 CsCl, 1 Mg-ATP, 10 EGTA, 10 HEPES (pH ϭ 7.3), and the bath solution contained 140 CsCl, 15 glucose, 15 HEPES (pH ϭ 7.4). For inside-out measurements the pipette contained the bath solution and the bath solution was exchanged by the pipette solution after establishing the seal. To cAMP Inhibits TRP␥ Channels evaluate the permeability of pTRP␥ channel to Na ϩ and Ca 2ϩ the bath solution contained 140 NaCl, 15 glucose, 15 HEPES, and 140 N-methyl-D-glucamine-Cl, 5 CaCl 2 , 10 glucose, 15 HEPES, respectively. The permeability ratios were calculated according to Equations 1 and 2.
where P ion are the permeabilities, [ion] i and [ion] o are the internal and external concentration of the respective ion, and V rev is the reversal potential (11).
To rule out an effect on endogenous HEK293 channels (12), 8-Br-cAMP and all other substances used below were first tested in mocktransfected cells (n ϭ 5-8). In no case we found any significant effect indicating that all reported effects of the agents relied on the expressed TRP␥ channels. On the other hand, it cannot be excluded that heterologous expression of TRP proteins in HEK293 cells may lead to the formation of new TRP heteromultimers including endogenous TRPCs (13) or to up-regulation of TRPC expression. As outlined under "Discussion," this is rather unlikely in the case of the TRP␥ proteins used in this study.
Data Analysis-Results were given as means Ϯ S.E. of mean (S.E.), n ϭ number of cells. The evaluation of statistical significance of differences was performed with Student's t test. For data analysis the software IgorPro (WaveMetrics, Lake Oswego, OR) and Prism 4 (Graph Pad Software, San Diego, CA) were used.

RESULTS
Sequence Analysis of P. americana TRP␥-Using a combined degenerated PCR and RACE protocol, we were able to isolate the full-length sequence of the P. americana pTRP␥ channel. The 3585-bp cDNA sequence (GenBank TM accession number AY387857) is coding for a protein of 1154 amino acids showing all typical features of a TRP-like channel: four ankyrin-like repeats at the N terminus, a coiled-coil domain, six transmembrane regions, a TRP-pore sequence typical for a nonspecific cation channel, and a long C terminus harboring several conserved TRP-specific sequences. The pTRP␥ sequence is 66% identical and 75% homologous to the orthologous D. melanogaster sequence (dTRP␥), whereas sequence similarity to paralogous Drosophila sequences (dTRP and dTRPL) is significantly lower and spans only the very conserved core region of ϳ900 amino acids of the protein, with ϳ50% identity and 65% similarity. An alignment of pTRP␥ and dTRP␥ indicating several conserved protein features is shown in Fig. 1. The C terminus is less conserved between Periplaneta and Drosophila TRP␥ as seen in various gaps in the alignment and in a C-terminal extension of pTRP␥. A comparison with the Drosophila genomic sequence (3) indicated that alternative splicing only in one case might be responsible for these variations, i.e. the pTRP␥ sequence lacked one dTRP␥ exon (indicated in Fig. 1). In addition, when amplifying the full-length pTRP␥ sequence, a short alternatively spliced variant, pTRP␥b, was discovered. This splice variant harbored the homologous dTRP␥ exon missing in the full-length form, but it introduced a stop codon thereafter, resulting in a truncated pTRP␥ version with only 791 amino acids, i.e. lacking most of the C-terminal sequence. In addition, a short variable region was identified in the C terminus, resulting in either deletion of a single amino acid (alanine 827) or in an expansion by four residues (GNTS) preceding alanine 827. The mechanism of this sequence variation is unclear but may represent alternative usage of exon-intron splice sites or polymorphisms.
Expression Analysis of pTRP␥-The expression of pTRP␥ channels in neuronal tissue was tested by a nested RT-PCR approach. An intronspanning 355-bp PCR fragment was amplified from Periplaneta brain as well as from all thoracic and abdominal ganglia indicating a widespread expression in cockroach neuronal tissue. In addition, in four out of ten investigated DUM neurons a pTRP␥ signal was found on single-cell level as well ( Fig. 2A). A Western blot analysis of neuronal tissues showed a distinct band with the expected molecular mass of 129 kDa in abdominal and thoracic ganglia with very low background staining (Fig.  2B). A protein of the same size was detected by the same antiserum (1) in transiently transfected HEK293 cells expressing pTRP␥ but not in mock-transfected cells.
Distribution of pTRP␥ in the Nervous System-Control immunostainings were performed with both pTRP␥ antibodies in HEK293 cells transfected with pTRP␥ DNA. 24 h after transfection, HEK293 cells showed significant immunofluorescence (Fig. 2, C and E), which was missing in non-transfected cells (Fig. 2, D and F). Immunostainings within the Periplaneta nervous system revealed a frequent but distinct expression of pTRP␥ both in brain (Fig. 3, A-H) and abdominal ganglia (Fig. 3, J-L). Both antibodies, which were designed to recognize different C-terminal pTRP␥ motifs, were seen to stain the same structures (Fig. 3, A, C, D, and F). Generally less intense immunofluorescence was seen in neuronal somata ranging from virtually absent to clearly pronounced staining as shown in Fig. 3, J and K, for the DUM neurons. Within the brain, immunostaining was found in the protocerebrum, in the deutocerebrum, and in the optic lobes. In the protocerebrum longitudinal and commissural fibers are stained. Within the central complex there is strong immunoreactivity in fibers around the upper (Fig. 3, D and F) and the lower (Fig. 3D) part of central body. Interestingly, the arborization areas within the central body do not show any immunoreactivity. This might indicate that pTRP␥ is expressed in fibers exept for their terminal branches including synaptic areas (Fig. 3, D and F). Similarly, there was clear staining around the olfactory glomeruli in the antennal lobe but not at all within the glomeruli where olfactory signals are processed via manifold synaptic contacts (Fig. 3G). The antennocerebral tract containing the projection neuron fibers was extensively stained (Fig. 3H). Besides the regions of olfactory information processing there is prominent staining in the optic lobes, particularly within the medulla (Fig. 3I). In the abdominal ganglia the somata of DUM neurons (Fig. 3, J and K) and those of a few smaller, not identified neurons (data not shown) appear stained. As in the brain, some fiber tracts are prominently stained while others such as the giant axons lack any staining (Fig. 3, J and L).
cAMP Down-regulates pTRP␥ but Not dTRP␥ Currents-Expression of pTRP␥ in HEK293 cells produced a constitutively active non-selective cation conductance, which was not found in mock-transfected cells.
The pTRP␥ current showed outward rectification, even with symmetrical ion concentrations (Fig. 4, A and B). The ratio of ion permeabilities was estimated from the dependence of the reversal potential on the composition of bath versus pipette solutions. With 5 mM Ca 2ϩ and 140 mM Na ϩ in the bath versus 140 mM Cs ϩ in the pipette the reversal potential was Ϫ16.0 Ϯ 0.9 mV (n ϭ 5) and Ϫ0.4 Ϯ 1.1 mV (n ϭ 6), respectively. This yields, according to Equations 1 and 2, a permeability cAMP Inhibits TRP␥ Channels FEBRUARY 10, 2006 • VOLUME 281 • NUMBER 6 ratio of P Na :P Cs :P Ca ϭ 1:1.01:2.05, which is very similar to the corresponding ratio reported for dTRP␥ (3). HEK293 cells expressing pTRP␥ were thus exposed to sustained influx of ions including Ca 2ϩ and showed considerably shortened survival time in culture compared with mock-transfected cells. 24 h after transfection the ratio of green fluorescing cells to non-fluorescing cells was 0.19 Ϯ 0.08 (n ϭ 200). Among cAMP Inhibits TRP␥ Channels the cells expressing pTRP␥ the morphology of 26 Ϯ 7 (n ϭ 38) cells was typical for HEK293 cells; the remaining cells appeared spherical. 48 h after transfection only 9 Ϯ 3% of cells (n ϭ 200) showed green fluorescence, and all these cells were of spherical shape.
With the given sustained Ca 2ϩ permeability of pTRP␥ in the heterologous expression system this channel meets one of the criteria for a channel involved in NCCE in DUM neurons. The most intriguing question was whether pTRP␥ would be inhibited by cAMP. To test this we used symmetric Cs ϩ solution to avoid Ca 2ϩ overload of cells. Application of the membrane-permeant cAMP analog 8-Br-cAMP (2 M) caused a significant reduction of the pTRP␥ current but no total block, since application of the TRP channel blocker 2-APB (100 M) in the presence of 8-Br-cAMP blocked the remaining current nearly completely (Fig. 4, A and C). The fraction of current resistant to 8-Br-cAMP was 0.63 Ϯ 0.06 at Ϫ100 mV and 0.67 Ϯ 0.04 at ϩ100 mV (n ϭ 16), i.e. the 8-Br-cAMP effect was not apparently voltage-dependent. The 8-Br-cAMP-induced inhibition of the pTRP␥ current started to develop within the first minute after application and saturated within a couple of minutes (Fig. 4B). After removal of 8-Br-cAMP the current recovered within 3-5 min (Fig. 6A). Taken together, 8-Br-cAMP seemed to act like a channel blocker. To support this more directly we performed experiments in the inside-out configuration. Bath application of cAMP again reduced the current, but even at a concentration of 1 mM more than 50% of current was resistant to cAMP (Fig. 4D). The dose-response curve of the cAMP effect shown in Fig. 4E is characterized by an IC 50 of 74 nM and a Hill coefficient of 0.83. At [cAMP] ϭ 10 M the effect saturates. Thus, the channel appears to be regulated in a physiological cAMP concentration range.
To our knowledge, no TRP channel has been found to be downregulated by cAMP. Therefore, we were interested to see whether Drosophila TRP␥ (dTRP␥), the only other presently known TRP␥ channel, was also sensitive to cAMP. When expressing dTRP␥ in HEK293 cells we obtained, similarly to pTRP␥, an outwardly rectifying current that was constitutively active. But this current was resistant to cAMP (Fig. 5,  A and B). Application of 8-Br-cAMP (2 M) only yielded a current reduction by 4% (n ϭ 8). For comparison, 2-APB (100 M) strongly reduced the dTRP␥ current (Fig. 5, A and B).
To characterize the selectivity of the putative cyclic nucleotide-binding site in the pTRP␥ channel, we applied the membrane-permeant cGMP analog 8-Br-cGMP (2 M). However, cGMP failed to affect the pTRP␥ and dTRP␥ currents (Fig. 5C). Thus, the down-regulation of TRP␥ channels is specific for cAMP.
The depressing action of cAMP on the NCCE channel in DUM neurons was a direct cAMP effect and not mediated by channel phosphorylation via PKA (8). To confirm that the down-regulation of pTRP␥ current directly relies on cAMP, we had to rule out a possible contribution of pTRP␥ channel phosphorylation by endogenous PKA. We thus tested the effect of 8-Br-cAMP in presence of the PKA inhibitors KT5720, Rp-cAMPS, and myristoylated PKA-inhibiting peptide 14 -22 amide (PKI). If the regulation of pTRP␥ was PKA-dependent, the presence of an inhibitor should abolish the effect of cAMP. There was, however, neither a direct effect of PKA inhibitors on pTRP␥ current (Fig. 6A) nor a significant change in the reducing effect of 8-Br-cAMP (Fig. 6B).
Searching for a putative constituent of a neuronal NCCE channel we thus found the pTRP␥ channel, which displayed some of the expected properties including the unusual sensitivity to cAMP. For comparing the pharmacological profiles of TRP␥ channels from Periplaneta and Drosophila we tested, besides typical TRP channel blockers such as 2-APB and La 3ϩ (14) and SKF96365, which inhibits a variety of channels providing Ca 2ϩ influx (15,16), and LOE-908, which blocks NCCE in various preparations including DUM neurons (17,2). The pTRP␥ current was reduced by LOE-908 (10 M) by 26 Ϯ 5% (n ϭ 5), while dTRP␥ was insensitive (Fig. 7). 2-APB (100 M) reduced the pTRP␥ current by 71 Ϯ 6% (n ϭ 9) and the dTRP␥ current by 49 Ϯ 6% (n ϭ 7). SKF96365 (10 M) did not affect either of the currents, and La 3ϩ (1 mM) reduced the pTRP␥ current by 50 Ϯ 9% (n ϭ 5) and the dTRP␥ current by 23 Ϯ 9% (n ϭ 8). cAMP Inhibits TRP␥ Channels FEBRUARY 10, 2006 • VOLUME 281 • NUMBER 6 FIGURE 4. pTRP␥ current is attenuated by cAMP. pTRP␥ was expressed in HEK293 cells, and currents were measured in the whole-cell mode using symmetric CsCl solutions. A, families of currents obtained by 100-ms lasting voltage jumps from Ϫ80 to ϩ100 mV from a holding potential of Ϫ50 mV. The currents measured before (Control) and 5 min after application of 2 M 8-Br-cAMP are shown. Further application of 100 M 2-APB caused nearly completely current block after 5 min (2-APB). B, time course of the current reduction by 8-Br-cAMP. Currents were measured at ϩ100 mV and normalized to the current I Control that was obtained before application of 8-Br-cAMP (arrow). Data represent means Ϯ S.E. of n ϭ 5 cells. C, pTRP␥ currents obtained by voltage ramps from Ϫ100 to ϩ100 mV (in 400 ms). The pTRP␥ current shows outward rectification, the reversal potential is near 0 mV. 8-Br-cAMP (2 M, 5 min) caused a reduction of both inward and outward current. In the presence of 2-APB (5 min) only a residual current remained. D, pTRP␥ currents obtained as described for C in the inside-out configuration. cAMP (1 mM, 5 min) reduced the inward and outward current. E, concentration dependence of the cAMP-effect at ϩ100 mV.   The most marked difference of pTRP␥ and dTRP␥ is the cAMPmediated current inhibition in the Periplaneta variant suggesting that one may find a cAMP-binding site by comparing both channel sequences. Since there is no typical consensus cAMP-binding motif in the channel protein, we tested a pTRP␥ splice variant (pTRP␥b) lacking most of the C terminus (see Fig. 1). Expression of pTRP␥b yielded an outwardly rectifying current that was, however, insensitive to cAMP (Fig. 8, A and B). Therefore, the C terminus seems to play a critical role for cAMP binding and channel modulation. Interestingly, LOE-908 also failed to affect the pTRP␥b current, and 2-APB was less effective than in the case of pTRP␥ (Fig. 8C).
Does pTRP␥ Contribute to the NCCE Conductance in DUM Neurons? -Our immunocytochemical investigations have shown that DUM neurons express pTRP␥ channels, and the properties of the pTRP␥ current in the heterologous expression system are compatible with the suggestion that this channel may be involved in forming the NCCE channel in DUM neurons. To obtain further support for this hypothesis, we tested whether LOE-908, 2-APB, and La 3ϩ , which attenuated the pTRP␥ current in HEK293 cells (Fig. 7B), would also affect NCCE in DUM neurons. Using a bath solution containing only 2 mM Ca 2ϩ as charge carrier, LOE-908 (10 M) and 2-APB (100 M) reduced the resting current density at Ϫ90 mV by 0.43 Ϯ 0.15 pA/pF (n ϭ 5) and 0.54 Ϯ 0.09 pA/pF (n ϭ 5), respectively (Fig. 9A). This reduction is similar to the effect of 8-Br-cAMP (2 M) that caused a reduction by 0.42 Ϯ 0.06 pA/pF (n ϭ 7). This shows that the NCCE channel in DUM neurons, in addition to cAMP and LOE-908 (8), is sensitive to 2-APB. However, the relative effect of 2-APB appears to be weaker than on pTRP␥ channels expressed in HEK293 cells (Fig. 7A). While 2-APB produced a stronger reduction of the TRP current than cAMP and LOE-908, there was no significant FIGURE 5. dTRP␥ current is resistant to cAMP. dTRP␥ was expressed in HEK293 cells, and currents were measured as described in the legend to Fig. 4A. A, application of 2 M 8-Br-cAMP did not reduce the current, while application of 100 M 2-APB (5 min) caused strong current reduction (2-APB). B, dTRP␥ currents obtained by voltage ramps as described in the legend to Fig. 4B. The dTRP␥ current shows outward rectification, and the reversal potential is near 0 mV. The currents registered before (Control) and 5 min after 8-Br-cAMP application appear identical, by contrast to the diminished current measured in the presence 2-APB. C, effect of 8-Br-cAMP (cAMP, 2 M) and 8-Br-cGMP (cGMP, 2 M) on dTRP␥ and pTRP␥ currents measured at ϩ100 mV. Data represent means Ϯ S.E. of n ϭ 6 -10 cells.  cAMP Inhibits TRP␥ Channels FEBRUARY 10, 2006 • VOLUME 281 • NUMBER 6 difference in the effect of all three compounds in DUM neurons. To confirm that these findings in DUM neurons reflect inherent properties of NCCE channels, we repeated the experiments using 3 mM Sr 2ϩ as charge carrier to exclude Ca 2ϩ -dependent artifacts; the effects of cAMP, LOE-908, and 2-APB were not different either (Fig. 9A). With both charge carriers, La 3ϩ (1 mM) was the most potent blocker of resting current; it reduced the current density by 1.7 Ϯ 0.5 pA/pF (n ϭ 4) for Ca 2ϩ and by 1.6 Ϯ 0.4 pA/pF (n ϭ 7) for Sr 2ϩ . Taken together, the NCCE channel in DUM neurons is sensitive to those compounds that inhibit heterologously expressed pTRP␥ channels, but we find differences in the quantitative effect of some compounds in the two systems.
In a further attempt we asked to which extent the pTRP␥-blocking agents affect the NCCE potentiation. Arachidonic acid as well as its non-metabolizable analog ETYA were described to enhance NCCE in DUM cells (8). ETYA (10 M) produced an increase in resting current density at Ϫ90 mV by 1.07 Ϯ 0.26 pA/pF (n ϭ 7) (Fig. 9B). After preincubation of cells with 8-Br-cAMP (2 M), in the presence of LOE-908 (10 M) or 2-APB (100 M) the effect of ETYA was drastically reduced (Fig. 9B). Again, there was no significant difference in the blocking effect of these three compounds. These experiments led us to the conclusion that the three agents in the used concentration block at least 80% of the NCCE current in DUM neurons. This implies that the NCCE channel in DUM neurons is more sensitive to cAMP and LOE-908 than heterologously expressed pTRP␥ channels.

DISCUSSION
TRP proteins form a large and diverse family of ion channels with multiple functions (3). The first TRP channel was found in Drosophila photoreceptors where two further members of the family, TRPL and TRP␥, contribute to visual transduction. In adult Drosophila, the TRP␥ protein was identified only in the eyes, although the mRNA message was also found in the body (3). dTRP␥ appears highly enriched in photoreceptors where it preferentially forms heteromultimeric channels with TRPL. The eye-specific expression of dTRP␥ indicates that this channel protein in Drosophila is solely involved in phototransduction. By contrast, in adult Periplaneta we found the mRNA message of pTRP␥ throughout the nervous system, but we also detected the protein in neuronal cell bodies (Fig. 3). This indicates that the physiological role of TRP␥ in Periplaneta considerably extends that in Drosophila. Several attempts to amplify the two other TRPC homologs of Drosophila (TRP and TRPL) by a degenerated PCR strategy were unsuccessful, indicating that TRP␥ is probably the only TRPC-like channel with significant expression in the nervous system.
The immunocytochemical findings have indicated that the Periplaneta TRP␥ protein is frequently expressed within the nervous system. Most prominently, it occurs in nerve fibers involved in processing/conducting sensory, particularly olfactory and optic, information. The role  cAMP Inhibits TRP␥ Channels of pTRP␥ in these fibers remains elusive. In hippocampal neurons the insertion of TRPC5 channels in neurites terminates their outgrowth (18). An intriguing finding in the Periplaneta brain was the fact that branching regions of fibers do not express pTRP␥ (e.g. in the central body, Fig. 3D). Furthermore, we cannot exclude the possibility that pTRP␥ is also expressed in glial cells surrounding the neurites. This point needs to be analyzed in more detail, e.g. by using glial cell markers and by performing electron microscopy. The most important result of the immunocytochemical approach for the present study was to see that the pTRP␥ protein is located in the somata of DUM neurons.
Drosophila TRP, but also the mammalian counterparts that form the subfamily of canonical TRPs (TRPC), function as receptor-operated channels (3). Activation of PLC was found to be a key element in the signal transduction process leading to dTRP activation (19). On the other hand, heterologously expressed dTRPL and dTRP␥ proteins form constitutively active channels. However, coexpression of dTRPL with dTRP␥ produces regulated channels that can be activated by agonists for receptors that stimulate PLC (3).
In Periplaneta DUM neurons we previously found a Ca 2ϩ -permeable conductance that was voltage-independent and constitutively active (8,20). This Ca 2ϩ -background conductance was up-regulated by the peptide hormone AKH I via PLC activation. It was further established that the regulation of this conductance was independent of intracellular stores, i.e. that it contributes to NCCE (8). One might speculate whether the channel responsible for the Ca 2ϩ -background conductance might belong to the TRP family. A detailed analysis of the AKH I-signal transduction process yielded a complex picture. The diacyglycerol produced by PLC activation is metabolized by diacyglycerol lipase to AA. This activates, by an unknown mechanism, NO-sensitive guanylyl cyclase. The resulting increase in cGMP level activates the phosphodiesterase 2, which in turn lowers the cAMP level (8). The final step in the up-regulation of Ca 2ϩ -background conductance was thus the reduction of an inhibition. Such kind of regulation is akin to that of ATP-sensitive K ϩ channels, which open upon a drop in ATP concentration (21).
Searching for the molecular substrate of the Ca 2ϩ -background conductance we found pTRP␥, a homolog of the Drosophila TRP␥. This protein is expressed in DUM cells, and we have seen that channels formed by pTRP␥ in HEK293 cells show some properties expected for a channel responsible for the Ca 2ϩ -background conductance in DUM neurons.
It has to be kept in mind that overexpression of TRP␥ proteins in HEK293 cells may lead to the formation of new TRP heteromultimers including endogenous TRPCs (13) or to up-regulation of TRPC expression. However, the difference in 8-Br-cAMP sensitivity of currents produced by expression of pTRP␥ and its C-terminally truncated version pTRP␥b (compare Fig. 4, C and D, with Fig. 8) clearly demonstrates that the overexpressed and not an endogeneous channel protein determines the channel properties. Moreover, expression of dTRP␥, the Drosophila homolog to pTRP␥, results in the generation of currents similar to those produced by pTRP␥ but with partially different pharmacological properties (cf. Fig. 7).
With respect to a putative role of pTRP␥ in forming Ca 2ϩ -background channels, the most important feature, besides its permeability to Ca 2ϩ , is the sensitivity to cAMP. Similarly important is the insensitivity of pTRP␥ to cGMP, since in DUM neurons the phosphodiesterase 2-mediated decrease in cAMP concentration requires an increase in cGMP concentration (8). Sensitivity of pTRP␥ to both cAMP and cGMP would thus not be compatible with the regulation of the NCCE channel in DUM neurons. Furthermore, both pTRP␥ and the back-ground conductance are inhibited by LOE-908, 2-APB, and by La 3ϩ . Taken together, our results are compatible with the hypothesis that pTRP␥ is involved in forming channels that conduct the Ca 2ϩ -permeable background conductance conferring NCCE in DUM neurons.
If our assumption was correct we had to propose a new functional role for a TRP channel, namely to act as a pacemaker channel. In DUM neurons the firing frequency is regulated by a variety of ion currents including the Ca 2ϩ -background current (22). Depression of this current, e.g. by FMRF-related peptides, decreases the firing rate (23), while potentiation, e.g. by AKH I, enhances the firing rate (8). Although AKH I modulates a set of DUM cell currents, i.e. Na ϩ and P/Q-type Ca 2ϩ current as well as Na ϩ -and Ca 2ϩ -dependent K ϩ currents, solely the up-regulation of the Ca 2ϩ -background current produces faster spiking (24). Furthermore, the Ca 2ϩ -background current in DUM neurons is involved in controlling the intracellular Ca 2ϩ concentration (20) and in filling of intracellular Ca 2ϩ stores (25). Up-regulation of this current can induce local as well as global Ca 2ϩ signals (8).
From the viewpoint of ion channel regulation the most intriguing result of the present study is the down-regulation of pTRP␥ current by cAMP. Although the binding site of cAMP on pTRP␥ is currently unknown, two conclusions can be drawn from sequence comparisons, pattern searches, and electrophysiological measurements of the short pTRP␥b splice variant. (i) The distal C-terminal 400 amino acids of pTRP␥ seem to contain the cAMP-binding site as the truncated pTRP␥b splice variant does not show any cAMP effect, and (ii) no consensus cAMP-binding sites were found in pTRP␥ indicating an unusual cAMP binding. Further work will be dedicated to analyze the cAMP effect on the level of single channels and to identify putative cAMPbinding sites in the channel protein. The latter will not exclude a search for N-terminal cAMP-binding sites, since it might also be possible that cAMP binds at the N terminus, while the distal C terminus is necessary to transduce the conformational change to channel closure.