On the voltage-dependent Ca2+ block of serotonin 5-HT3 receptors: a critical role of intracellular phosphates
- 1Swammerdam Institute for Life Sciences, Center for NeuroScience, University of Amsterdam, the Netherlands
- Corresponding author J.A. van Hooft: Swammerdam Institute for Life Sciences, Center for NeuroScience, University of Amsterdam, PO Box 94084, NL-1090 GB Amsterdam, the Netherlands. Email: j.a.vanhooft{at}uva.nl
Abstract
Natively expressed serotonin 5-HT3 receptors typically possess a negative-slope conductance region in their I–V curve, due to a voltage-dependent block by external Ca2+ ions. However, in almost all studies performed with heterologously expressed 5-HT3 receptors, this feature was not observed. Here we show that mere addition of ATP to the pipette solution is sufficient to reliably observe a voltage-dependent block in homomeric (h5-HT3A) and heteromeric (h5-HT3AB) receptors expressed in HEK293 cells. A similar block was observed with a plethora of molecules containing a phosphate moiety, thus excluding a role of phosphorylation. A substitution of three arginines in the intracellular vestibule of 5-HT3A with their counterpart residues from the 5-HT3B subunit (RRR-QDA) was previously shown to dramatically increase single channel conductance. We find this mutant to have a linear I–V curve that is unaffected by the presence of ATP, with a fractional Ca2+ current (Pf%) that is reduced (1.8 ± 0.2%) compared to that of the homomeric receptor (4.1 ± 0.2%), and similar to that of the heteromeric form (2.0 ± 0.3%). Moreover, whereas ATP decreased the Pf% of the homomeric receptor, this was not observed with the RRR-QDA mutant. Finally, ATP was found to be critical for voltage-dependent channel block also in hippocampal interneurons that natively express 5-HT3 receptors. Taken together, our results indicate a novel mechanism by which ATP, and similar molecules, modulate 5-HT3 receptors via interactions with the intracellular vestibule of the receptor.
The serotonin 5-HT3 receptor is a ligand-gated ion channel belonging to the family of Cys-loop transmitter-gated ion channels. Two subunits have been cloned: the compulsory 5-HT3A subunit, which forms a homomeric ion channel (Maricq et al. 1991), and the accessory 5-HT3B subunit, which enables the formation of a heteromeric ion channel (Davies et al. 1999). In addition, several candidate genes encoding accessory 5-HT3 receptor subunits have been identified in the human, but not in the rodent genome (Niesler et al. 2003, 2007). Both the homomeric and the heteromeric 5-HT3 receptor are non-selective cation channels. However, the 5-HT3B subunit confers a few striking properties to the heteromeric receptor: the single channel conductance of the heteromeric ion channel is ∼30-fold higher, and the permeability to Ca2+ ions is ∼2-fold lower than that of the homomeric receptor (Davies et al. 1999; Peters et al. 2005). It has been shown that three arginine residues (R432, R436 and R440), located in the intracellular vestibule between transmembrane domains 3 and 4 of the h5-HT3A subunit play a critical role in determining the single channel conductance: mutation of these arginine residues into the corresponding amino acids of the h5-HT3B subunit (RRR-QDA) resulted in a functional homomeric 5-HT3 receptor with a single channel conductance similar to that of the heteromeric receptor (Kelley et al. 2003).
The 5-HT3 receptor is expressed in the central nervous system (CNS), particularly in GABAergic interneurons (Morales & Bloom, 1997; Chameau & van Hooft, 2006). One of the signature features of 5-HT3 receptor-mediated ion currents in the CNS is the presence of a region of negative-slope conductance in the I–V curve that reflects a voltage-dependent block by external Ca2+ ions (Kawa, 1994; McMahon & Kauer, 1997; Roerig et al. 1997; van Hooft & Wadman, 2003), analogous to the voltage-dependent block of NMDA receptors by Mg2+ ions. However, the structural/molecular determinants of the voltage-dependent Ca2+ block of 5-HT3 receptors are not known. A major hindrance in understanding this phenomenon is due to the apparent absence of the voltage-dependent Ca2+ block in heterologous expression systems; so far, a voltage-dependent Ca2+ block was reported only in the initial study on the cloning of the mouse 5-HT3A subunit (Maricq et al. 1991), whereas in subsequent studies on the cloned 5-HT3 receptor subunits in heterologous expression systems an inward rectifying or linear I–V curve was reported (Yakel et al. 1993; Hussy et al. 1994; Gill et al. 1995; Brown et al. 1998; Davies et al. 1999; Dubin et al. 1999; Gunthorpe et al. 2000; Hanna et al. 2000; Hapfelmeier et al. 2003; Hu & Lovinger, 2005; Hu et al. 2006).
In this study, we re-examined this issue and found that a region of negative-slope conductance in the I–V curve of cloned h5-HT3 receptors expressed in HEK293 cells can be reconstituted by the mere inclusion of adenosine triphosphate (ATP) in the pipette solution. This effect does not depend on phosphorylation of the 5-HT3 receptor, as a plethora of molecules containing a phosphate moiety, including non-hydrolysable analogues of ATP, mimic the effect of ATP. In addition, by directly measuring the fractional calcium currents of the h5-HT3 receptor we find the Ca2+ permeability of the receptor to be modulated by intracellular phosphates. Furthermore, we show that the intracellular vestibule is critically involved in the actions of ATP. Finally, we show that 5-HT3 receptors native to rat hippocampal interneurons are similarly regulated by intracellular phosphates.
Methods
Cell culture
Human embryonic kidney 293 (HEK293) cells were maintained in minimum essential medium (MEM) supplemented with 10% fetal calf serum, 2 mm glutamine, and 100 μg ml−1 penicilline–streptomycin at 37°C in a humidified atmosphere containing 5% CO2. Cells were passaged weekly and medium was refreshed every 2–3 days. Cells were plated on 12 mm coverslips and were transiently transfected with a vector encoding the human 5-HT3A subunit or the RRR-QDA mutant (Kelley et al. 2003) using the calcium phosphate precipitation method. For expression of heteromeric 5HT3 receptors, vectors encoding h5-HT3A and h5-HT3B were cotransfected at a 1: 1 ratio. A vector encoding EGFP (EGFP-N1, Clontech) was cotransfected at a ratio of 1: 5 to permit detection of transfected cells by standard epifluorescence. All human cDNAs were provided by Dr J. A. Peters (University of Dundee, UK).
Electrophysiology
Whole-cell patch-clamp recordings from HEK293 cells were preformed 1–4 days post-transfection. A coverslip containing the transfected cells was placed in the recording chamber of a Nikon Diaphot inverted microscope and cells were continuously superfused with extracellular solution containing (in mm): 135 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 5 Hepes (pH 7.3 with NaOH). Pipettes were pulled from borosilicate glass and had a resistance of 2–5 m when filled with pipette solution containing (in mm): 105 potassium gluconate, 30 KCl, 0.5 CaCl2, 5 MgCl2, 5 EGTA, 10 Hepes (pH 7.3 with KOH). Phosphate-containing chemicals were added to the pipette solution when indicated, and MgCl2 was omitted when using Mg2+ salts. Series resistance was < 10 m and was compensated for at least 70%. Serotonin (5-HT, 100 μm) was applied using a SF-77B Perfusion Fast-Step device (Warner Instruments LLC, Hamden, CT, USA) with intervals of at least 3–4 min in order to allow for recovery from desensitization. Currents were recorded (sampling frequency 5 kHz) using a PC (Atari TT030)-controlled Axopatch 200A amplifier (Axon Instruments/MDS Analytical Technologies, Union City, CA, USA) and custom-made software. All voltages were offline corrected for liquid junction potential (13 mV). All chemicals were obtained from Sigma.
Preparation of rat hippocampal slices and recording of 5-HT3 receptor-mediated currents from hippocampal interneurons were performed as previously described (van Hooft & Wadman, 2003). In short, male Wistar rats (P14–P16) were decapitated without prior anaesthesia, and parasaggital slices (250 μm) of the hippocampus were cut. Interneurons in stratum radiatum of the hippocampal CA1 area were visualized using infrared differential interference contrast videomicroscopy. 5-HT3 receptor-mediated ion currents were evoked by pressure ejection of 5-HT in the vicinity of the soma, and currents were recorded using whole-cell voltage clamp. Experiments were conducted according to the ethics committee guidelines of the University of Amsterdam.
Patch-clamp photometry
Fractional Ca2+ currents were determined using the method described by Egan & Khakh (2004; see also Supplemental material). Cells were loaded with 2 mm Fura-2 via the pipette solution. Fluorescence upon excitation at 380 nm from an Hg bulb source was measured using a Nikon Fluor 40× oil objective and a photomultiplier tube (TILL Photonics, Gräfelfing, Germany). An adjustable pinhole was used to limit the area of recording to the proximity of the cell in order to minimize background fluorescence. All fluorescence signals were expressed as ‘bead units’ (BU), determined each day from the average fluorescence signal obtained from 10 separate 6 μm Inspeck® Green beads (Molecular Probes). One hundred micromolar 5-HT was applied for 10 s and the F380 signal was acquired throughout this period, starting from 1 s before the application of 5-HT. At the end of each recording session the membrane was perforated by an electric stimulus (the ‘zap’ mode of the amplifier) and the subsequent maximum influx of Ca2+ (maximum decline in fluorescence at 380 nm) was recorded to ensure that Fura-2 was not saturated due to the influx of Ca2+. Control experiments using non-transfected cells revealed that application of 5-HT did not result in an ion current nor a change in intracellular Ca2+ levels (not shown).
Data analysis
I–V curves were determined using a ramp protocol (unless mentioned otherwise) in which the cell was first held at a potential
of +47 mV for 1 s, followed by a ramp from +47 mV to −153 mV during 500 ms. The protocol was repeated in the presence of 5-HT,
and the currents recorded in the absence of 5-HT were subtracted from those obtained in its presence (Fig. 1A). I–V curves were subsequently normalized to the amplitude at –30 mV. Control experiments using non-transfected HEK293 cells revealed
that application of the voltage ramp per se did not evoke any endogeneous currents (not shown). Fractional Ca2+ currents (Pf%) were calculated according to:
(1)where QCa is the charge carried by Ca2+ and Qt is the total charge entering the cell upon 5-HT application. Qt was calculated as the integral of the 5-HT-induced ion current during the first 3 s after application of 5-HT. QCa was determined according to:
(2)where F380 is the decrease of the F380 signal during the first 3 s after application of 5-HT (see also Supplemental material). Fmax was calculated from experiments in which Ca2+ was the only permeant ion (see Supplemental material) and amounted to 0.0047 ± 0.0002 BU pC−1 (n = 27). Offline analysis was performed using IGOR Pro software (Wavemetrics, Lake Oswego, OR, USA). All data are presented
as means ± s.e.m. Student's t test was used for comparison, and a P-value of 0.05 was used to indicate a significant difference.
Results
Intracellular ATP confers voltage-dependent Ca2+ block on heterologously expressed h5-HT3A receptors
We recorded whole-cell currents from HEK293 cells transiently expressing the h5-HT3A subunit and I–V curves were determined using a voltage-ramp protocol (Fig. 1A). When recordings were performed 5–10 min (or later) after reaching the whole-cell configuration, the I–V curve of the 5-HT-induced ion current was inward-rectifying (n = 15), similar to the I–V curves usually reported for the heterologously expressed 5-HT3 receptor. Interestingly, we found that when the recordings were performed shortly (< 2 min) after establishing the whole-cell configuration, a region of negative-slope conductance in the I–V curve was evident (Fig. 1B). An intuitive explanation for the time-dependent disappearance of the region of negative-slope conductance in the I–V curve would be wash-out of an intracellular factor upon reaching the whole-cell configuration. Indeed, experiments in which wash-out was prevented by using perforated-patch clamp with gramicidin, showed that the region of negative-slope conductance persisted (not shown).
We found that mere inclusion of ATP (5 mm) in the pipette solution is sufficient to maintain this region of negative-slope conductance in the I–V curve of the 5-HT3 receptor. This effect persisted throughout the recording session, which lasted at least 20 min, and in some cases longer than 60 min (Fig. 1C). Inclusion of ATP in the pipette solution did not affect the resting membrane conductance (101 ± 43 MΩ (n = 17) and 109 ± 39 MΩ (n = 13) in the presence and absence of ATP, respectively).
It has been shown before that the region of negative-slope conductance in the I–V curve of 5-HT3 receptors in the CNS is due to a voltage-dependent block by external Ca2+ ions (Kawa, 1994; McMahon & Kauer, 1997; van Hooft & Wadman, 2003). This is also the case for the h5-HT3A receptor expressed in HEK293 cells: when we omitted Ca2+ from the external solution we observed an inward-rectifying I–V curve, even in the presence of intracellular ATP. As reported previously (van Hooft & Wadman, 2003), the shape of the I–V curve was similar when recorded using the ramp protocol or when determined by the amplitudes of the 5-HT-evoked ion current at different holding potentials (Fig. 1E). Taken together, our results demonstrate that a region of negative-slope conductance in the I–V curve of the 5-HT3 receptor can be reliably reproduced in a heterologous expression system, requiring the presence of both intracellular ATP and extracellular Ca2+ ions.
A plethora of phosphate-containing molecules mimic the effect of ATP on the 5-HT3A receptor
The observation that intracellular ATP is required for the regulation of I–V curve of the 5-HT3A receptor raises the question of whether this effect is specifically mediated by ATP, and of whether ATP interacts directly with the ion channel, or via an indirect mechanism such as phosphorylation. We found that GTP, cytidine 5′-triphosphate (CTP) and uridine 5′-triphosphate (UTP) mimic the effect of ATP on the I–V curve (Fig. 2A). Inclusion of inositol 1,4,5-trisphosphate (IP3) in the pipette solution did not result in a clear region of negative-slope conductance but rather in an intermediate I–V curve that had reduced conductance at hyperpolarized membrane potentials as compared to control (Fig. 2A). In the presence of ATPγS, a non-hydrolysable analogue of ATP, the I–V curve displayed a region of negative-slope conductance similar to that in the presence of ATP (Fig. 2B), suggesting that phosphorylation does not play a role in the maintenance of the negative-slope conductance region. Including adenylyl imidophosphate (AMP-PNP), another non-hydrolysable analogue of ATP, in the pipette solution resulted in an intermediate I–V curve that does not clearly possess a region of negative-slope conductance but has reduced inward rectification as compared to control (Fig. 2B). When ADP was included in the pipette solution, a region of negative-slope conductance was still evident in the I–V curve (Fig. 2C), indicating that the γ phosphate is not critical, and further supporting the notion that phosphorylation does not play a role in this form of regulation. In the presence of either AMP or adenosine, the I–V curve was inward-rectifying and similar to control (Fig. 2C). Inclusion of triphosphate (PPPi), pyrophosphate (PPi) or orthophosphate (Pi) in the pipette solution was also sufficient to facilitate a region of negative-slope conductance (Fig. 2D). The voltage-dependent block in the presence of intracellular PPi and Pi appeared at more depolarized membrane potentials and caused a more robust inhibition compared to the block observed in the presence of ATP. Taken together, the results indicate that the regulation of the I–V curve of the 5-HT3A receptor is not mediated through phosphorylation, and that the main regulating determinant is a phosphate group, while the nucleoside plays a negligible role in this regulation.
Intracellular levels of ATP, as well as levels of inorganic phosphates, can fluctuate between different cells and in different metabolic states (Erecinska et al. 1977; Fulceri et al. 1993; Glinn et al. 1997, 1998; Tomasetti et al. 2001). Therefore, we tested whether the effects of ATP and inorganic phosphates on the I–V curve of the 5-HT3A receptor are concentration dependent. Increasing the levels of intracellular Pi resulted in a more robust inhibition of current, which started at more depolarized membrane potentials (Fig. 3A). At concentrations above 1 mm, the block occurred at more depolarized potentials but no change was observed in the maximal block, whereas in concentrations below 1 mm Pi the maximal block was reduced (Fig. 3B). A similar concentration dependence of the maximal block was observed with ATP (Fig. 3B).
The intracellular vestibule of the 5-HT3 receptor plays a critical role in the voltage-dependent Ca2+ block
Our results so far show that phosphate-containing molecules modulate the 5-HT3 receptor via a phosphorylation-independent mechanism, suggesting that these molecules might interact directly with the receptor. Whereas other cation-conducting members of the Cys-loop family of ligand-gated ion channels have a conserved negative ring of charge in their intracellular vestibule, the 5-HT3A receptor is unique in having a positive ring of charge instead. This positive ring of charge could potentially serve as an interface for interaction with negatively charged molecules such as a phosphate group. Unlike the 5-HT3A subunit, the 5-HT3B subunit does not possess such a positive ring of charge (Peters et al. 2005). Therefore, we coexpressed 5-HT3A and 5-HT3B subunits in HEK293 cells and recorded the I–V curve of 5-HT-induced currents in the presence and absence of intracellular ATP. In the absence of ATP, we found the I–V curve of the heteromeric form to be less inward-rectifying than the I–V curve of the homomeric receptor (n = 6, Fig. 4A and B), similar to previous studies reporting a slightly inward-rectifying or linear I–V curve (Davies et al. 1999; Dubin et al. 1999; Hanna et al. 2000). However, in the presence of intracellular ATP (5 mm), we observed a negative-slope region in the I–V curve that was similar to the one found for the homomeric form (n = 6; Fig. 4B).
Heteromeric 5-HT3 receptors consist of two 5-HT3A subunits and three 5-HT3B subunits (Barrera et al. 2005). Given the similar phenotype of the homomeric and heteromeric receptor with respect to their modulation by ATP, it is tempting to suggest that the positive charges in the intracellular vestibule contributed by two 5-HT3A subunits are sufficient to impose the region of negative-slope conductance in the I–V curve of the heteromeric receptor. A triple mutation that cancels this positive charge by replacing three arginines of the 5-HT3A with their corresponding residues in the 5-HT3B subunit (R432Q, R436D, R440A, denoted as RRR-QDA) was previously shown to result in a dramatic increase in single-channel conductance (Kelley et al. 2003). Interestingly, here we found that upon expression in HEK293 cells, the I–V curve of the RRR-QDA mutant was linear and did not show a region of negative-slope conductance in the presence of ATP (Fig. 4C). Thus we conclude that the three arginine residues present at the intracellular vestibule of the 5-HT3A subunit are critical for the regulation of the channel by ATP, and that the contribution of the arginine residues of two 5-HT3A subunits in the heteropentameric receptor is dominant for this phenotype.
ATP reduces the fractional Ca2+ current through h5-HT3 receptors
The QDA residues in the heteromeric receptor and in the RRR-QDA mutant are critically involved in the high single channel conductance (Kelley et al. 2003), but so far it has not been clear whether these residues also participate in determining ion selectivity. The Ca2+ permeability of the heteromeric receptor was previously reported to be lower than that of the homomeric receptor (PCa/PCs = 0.6 for the heteromer and 1.0 for the homomeric receptor; Brown et al. 1998; Davies et al. 1999). We determined the fractional Ca2+ currents (Pf%) of the h5-HT3 receptor channel using patch-clamp photometry (see Methods, Fig. 5A and Supplemental material). The Pf% value of h5-HT3A at −73 mV amounted to 4.1 ± 0.2% (n = 17, Fig. 5B), a value similar to the previously published Pf% value of the rat 5-HT3A receptor (4.7%, Egan & Khakh, 2004). Under the same conditions, the Pf% value of the heteromeric receptor amounted to 2.0 ± 0.3% (n = 9, Fig. 5B). Interestingly, the RRR-QDA mutant had a Pf% value similar to the heteromeric receptor (Pf% = 1.8 ± 0.2%, n = 10; Fig. 5B). This confirms the notion that the 5-HT3B subunit with the QDA residues is not only the major determinant for the single channel conductance of the 5-HT3 receptor, but is also involved in determining the ion selectivity of the channel.
When we included either ATP or Pi (5 mm) in the pipette solution, a significant reduction of 20–35% in Pf% was observed for the homomeric 5-HT3A receptor, with Pf% values of 3.2 ± 0.3% in the presence of ATP (n = 9; Fig. 5C) and 2.7 ± 0.3% in the presence of Pi (n = 9; not shown). Interestingly, the RRR-QDA mutant did not show a reduction of Pf% in the presence of ATP (Pf% = 2.4 ± 0.2%; n = 4; Fig. 5D). These results indicate that the effect of intracellular ATP on the 5-HT3A subunit, mediated by the RRR residues, is not limited to the voltage-dependent block of the receptor, but also results in reduced Ca2+ permeability.
Intracellular ATP is necessary for voltage-dependent block of 5-HT3 receptors native to hippocampal interneurons
So far, our experiments focused on the voltage-dependent regulation of heterologously expressed 5-HT3 receptors. We demonstrated that the voltage-dependent block that was previously reported almost exclusively in native tissue can also be recapitulated in a heterologous expression system, and that it is dependent on the presence of intracellular ATP and other phosphate-containing compounds. To test whether ATP is also critical for the voltage-dependent block of natively expressed 5-HT3 receptors, we recorded 5-HT-evoked whole-cell currents from hippocampal CA1 stratum radiatum interneurons, a class of neurons known to express relatively high levels of 5-HT3 receptors, and subsequently determined the I–V curve. As shown in Fig. 6, the I–V curves in the presence and absence of intracellular ATP in interneurons corresponded to the I–V curves obtained for heterologously expressed 5-HT3, demonstrating a region of negative-slope conductance in the presence of ATP, and inward rectification in its absence.
Discussion
The main finding of the present study is that ATP (and similar phosphate-containing molecules) can directly and dynamically regulate the total ion flux and the specific permeability to Ca2+ of both cloned and native 5-HT3 receptors, in a voltage- and concentration-dependent manner. The presence of a region of negative-slope conductance in the I–V curve of 5-HT3 is well established in the CNS (Kawa, 1994; McMahon & Kauer, 1997; Roerig et al. 1997; van Hooft & Wadman, 2003) and was shown to result from external block by Ca2+ ions. However, so far, except for the original cloning study of the 5-HT3A subunit (Maricq et al. 1991), all other studies preformed in clonal cell lines or heterologously expressed 5-HT3 receptors have failed to observe such a negative-slope conductance region, reporting usually an inward-rectifying I–V curve (Yakel et al. 1993; Hussy et al. 1994; Gill et al. 1995; Brown et al. 1998; Davies et al. 1999; Dubin et al. 1999; Gunthorpe et al. 2000; Hanna et al. 2000; Hapfelmeier et al. 2003; Hu & Lovinger, 2005; Hu et al. 2006). Notably, all mentioned studies in HEK293 cells or clonal cell lines were done either in the absence of ATP from the intracellular milieu or the absence of extracellular Ca2+, and studies in Xenopus oocytes did not control for intracellular ATP or inorganic phosphates levels. Our data suggest that the voltage-dependent block of the receptor by external Ca2+ is a basic property of the receptor, mediated by intracellular phosphates. Furthermore, our data show for the first time that also heteromeric 5-HT3AB receptors can undergo a voltage-dependent calcium block.
Heterologously expressed 5-HT3 receptors are calcium-permeable ion channels (Hargreaves et al. 1994; Brown et al. 1998). Whereas most previous estimations of calcium permeability were derived from shifts of reversal potentials in different ionic conditions, our data are based on direct measurements of the calcium efflux through the receptor. The main advantages of this method are (1) that the measurement is performed under physiological conditions, and (2) that the fraction of the current carried by Ca2+ can be estimated independent of how the current is generated, thus being independent of Goldman–Hodgkin–Katz (GHK) assumptions (Schneggenburger, 1996; Egan & Khakh, 2004). Along these lines, we provide here for the first time Pf% values of the heteromeric 5-HT3AB, as well as of the human homomeric receptor. The lower Ca2+ permeability found in the heteromeric receptor is consistent with previous estimations of Ca2+ permeability that were based on ion substitution experiments, as previous studies reported PCa/PCs values of 1.0–1.1 for homomeric 5-HT3A receptors and 0.6 for the heteromeric receptor (Brown et al. 1998; Davies et al. 1999). Applying a derivation of the GHK current equation (Schneggenburger et al. 1993; Schneggenburger, 1996), these values translate into predicted Pf% values of ∼4.0% and ∼2.4% for the homomeric and heteromeric receptors, respectively, values that are in close proximity to the Pf% values reported here (Fig. 5B). The fact that the RRR-QDA mutant had a similar low Pf% value to that determined for the heteromeric 5-HT3AB receptor suggests that the QDA residues at the intracellular vestibule of the 5-HT3B subunit not only determine ion conductance (as previously demonstrated by Kelley et al. 2003) but also influence ion selectivity. In a previous study (van Hooft & Wadman, 2003) we employed a three-barrier–two-site (3B2S) model based on Eyring rate theory to describe the region of negative-slope conductance in the I–V curve of 5-HT3 receptors recorded from rat hippocampal interneurons. While the model predicted a reduction of calcium permeability upon channel blockade, as confirmed also by our experimental data reported here, the Pf% values predicted by this model were much lower (< 1%) than the experimental measured values. However, the 3B2S model assumes independent monovalent–divalent ion permeation, and similar permeabilities for all monovalent cations, which may not be the case for 5-HT3 receptors.
The finding that ATP (and other phosphate-containing molecules) imposes the voltage-dependent Ca2+ block on 5-HT3 receptors implies that the voltage-dependent block is not an embedded structural feature of the ion channel. However, the absence of the region of negative-slope conductance from the I–V curve of the RRR-QDA mutant shows that the three arginine residues at the intracellular vestibule are critically involved. The presence of these positively charged residues in the 5-HT3A subunit was previously suggested to create an unfavourable local charge distribution at the far cytoplasmic side of the ion channel, thereby interfering with ion conduction and resulting in a low single channel conductance (Kelley et al. 2003). Along these lines, we speculate that ATP directly interacts with those arginine residues to reduce the total ion flux. This notion is supported by the observations that (non-hydrolysable) analogues of ATP, as well as ADP, inorganic phosphates and other triphosphate nulceosides mimic the effect of ATP, thus excluding a phosphorylation-mediated mechanism and supporting a role for the common phosphate group. A direct interaction of ATP with K+ channels, and specifically with the Kir6 subunit, has been described extensively. The phosphate moiety of ATP was suggested to interact with at least two arginine residues at the cytoplasmic side of the Kir6 subunit, underlying the ATP-induced inhibition of the K+ current (for review see Nichols, 2006).
Voltage-dependent block was induced in low millimolar concentrations of ATP and similar molecules, and with 0.5 mm Pi (Figs 2 and 3). Previous studies in mammalian cells have determined the free cytosolic concentration of Pi to be in the low millimolar range (Veech et al. 1972; Guynn et al. 1974). Our results demonstrate that under physiological conditions, the voltage-dependent block of the 5-HT3 receptor by Ca2+ ions exists, but that its properties may be fine-tuned by alterations in intracellular phosphates concentrations. Cells may experience substantial shifts in intracellular [Pi] in different metabolic states, such as ischaemia (Erecinska et al. 1977; Fulceri et al. 1993; Tomasetti et al. 2001) and intracellular Pi was reported to have a neuroprotective role in the face of excitotoxic and oxidative insults (Glinn et al. 1998). This raises the intriguing possibility that in 5-HT3 receptor-expressing neurons, a rise in intracellular Pi may serve a protective role through a decrease of the 5-HT3 receptor-mediated Ca2+ influx, by altering both total conductance and ion selectivity. The full physiological implications of the phosphate-mediated regulation of 5-HT3 receptors await further studies.
Acknowledgements
We thank John Peters (University of Dundee, UK) for the generous gift of vectors encoding the h5-HT3A, h5-HT3B and h5-HT3A (RRR-QDA) subunits, and Pascal Chameau for his comments on the manuscript. Parts of this study were financially supported by a NUFFIC-Huygens scholarship to Y.N. and a fellowship of the Royal Netherlands Academy of Arts and Sciences to J.A.vH.
Footnotes
-
(Received 4 March 2008; accepted after revision 12 June 2008; first published online 12 June 2008)
-
This paper has online supplemental material.
References
Figure 1. I–V properties of h5-HT3A receptors transiently expressed in HEK293 cells A, example of the ramp recording protocol used for determination of I–V curves. Currents recorded upon voltage-ramp application in the absence of agonist are subtracted from those recorded in the presence of 5-HT. The inset shows a magnification of the response to the ramp (black trace; dashed lines indicate the period of the ramp) with a superimposed current trace from the same cell without ramp (grey trace). B and C, representative I–V curves (normalized to −30 mV) for whole-cell measurements in 5-HT3A-expressing HEK293 cells in the absence (B) and the presence (C) of 5 mm intracellular ATP. The time of recording (since reaching the whole-cell configuration) is stated for each trace. In the absence of ATP, the negative-slope conductance region is evident only in the first minutes of recording. In the presence of 5 mm ATP, the negative-slope region is steady and hardly changes over time. D, averaged I–V curves obtained in the presence of 5 mm intracellular ATP and in the presence (n = 10) or absence (n = 3) of extracellular Ca2+ as indicated. The extracellular solution contained 1 mm Mg2+. E, comparison of the I–V curve as determined by the voltage-ramp protocol (black trace; grey traces represent the s.e.m., n = 9–15) to the I–V curve as determined from the peak amplitude of 5-HT3 receptor-mediated ion currents evoked at different holding potentials in the absence (open symbols) and presence (filled symbols) of 5 mm intracellular ATP (n = 2–7).
Figure 2. The region of negative-slope conductance in the I–V curve of h5-HT3A is regulated by phosphate-containing compounds A, average I–V curves of 5-HT3A receptors recorded in the presence of different triphosphate nucleosides (5 mm, except CTP: 4 mm) and IP3 (5 mm, n = 3–15). B, average I–V curves in the presence of the non-hydrolysable ATP analogues ATP.S (5 mm, n = 6) and AMP-PNP (5 mm, n = 7). C, average I–V curves in the presence of ATP, ADP, AMP and adenosine (all 5 mm, n = 3–15). D, average I–V curves obtained with ATP, PPPi, PPi and Pi (all 5 mm, n = 4–10). All I–V curves are normalized to their value at −30 mV, and were recorded > 10 min after establishing the whole-cell configuration.
Figure 3. The I–V curve of 5-HT3 receptors is modulated by ATP and Pi in a concentration-dependent manner A, average I–V curves (n = 3–9) in the presence of 25, 5, 1, 0.5 and 0.1 mm of intracellular Pi. B, percentage of relative channel block in the presence of Pi (left panel) or ATP (right panel) at different potentials, expressed as the amount of blocked current relative to control (in the absence of ATP or phosphate).
Figure 4. The negative-slope conductance region of the 5-HT3A receptor is critically dependent on arginine residues at the cytoplasmic loop A, average I–V curves of 5-HT3A receptors transiently expressed in HEK293 cells in the absence and presence of 5 mm ATP. The curves are identical to the ones presented in Fig. 1E and presented here for comparison. B, average I–V curves of heteromeric 5-HT3AB receptors in the absence and presence of 5 mm ATP (n = 6). C, average I–V curves of the 5-HT3A receptor mutant RRR-QDA, in the absence and presence of 5 mm ATP (n = 5).
Figure 5. Fractional calcium currents of h5-HT3 receptors A, representative trace of Pf% measurement. The black trace represents the 5-HT-induced whole-cell current, and the grey trace represents the Fura-2 signal (expressed in bead units) in response to Ca2+ entry. The continuous black line superimposed on the grey trace represents the integral of the whole-cell current (expressed in nanocoulombs). B, comparison of the Pf% values of homomeric 5-HT3A (n = 17), heteromeric 5-HT3AB (n = 9) and the RRR-QDA mutant (n = 10) in the absence of intracellular ATP. C, fractional calcium currents of the 5-HT3A receptor in the absence (n = 17) or presence (n = 9) of 5 mm ATP. D, fractional calcium currents of the RRR-QDA mutant, in the absence (n = 10) or presence (n = 4) of ATP in the pipette solution. All Pf% values were determined at −73 mV. Asterisks indicate significant differences (**P < 0.01, ***P < 0.001).
Figure 6. ATP- and voltage-dependent block of 5-HT3 receptors in rat hippocampal CA1 interneurons Average I–V curves from 5-HT induced whole-cell currents recorded in rat hippocampal CA1 interneurons in the presence (grey trace, filled symbols) and absence (black trace, open symbols) of intracellular ATP (2 mm). Traces represent I–V curves detemined using the ramp protocol (n = 4) and symbols represent the I–V curve as determined from the peak amplitude of 5-HT3 receptor-mediated ion currents evoked at different holding potentials (n = 2–4).
