Membrane lipid modulations remove divalent open channel block from TRPL and NMDA channels

Open channel block is a process in which ions bound to the inside of a channel pore block the flow of ions through that channel. Repulsion of the blocking ions by depolarization is a known mechanism of open channel block removal. For the N-methyl-D-aspartate (NMDA) channel, this mechanism is necessary for channel activation and is involved in neuronal plasticity. Several types of Transient Receptor Potential (TRP) channels, including the Drosophila TRP and TRP-Like (TRPL) channels, also exhibit open channel block. Therefore, removal of open channel block is necessary for the production of the physiological response to light. As there is no membrane depolarization before the light response develops, it is not clear how the open channel block is removed, an essential step for the production of a robust light response under physiological conditions. Here we present a novel mechanism to alleviate open channel block in the absence of depolarization by membrane lipid modulations. The results of this study show open channel block removal by membrane lipid modulations in both TRPL and NMDA channels of the photoreceptor cells and CA1 hippocampal neurons, respectively. Removal of open channel block is characterized by an increase in the passage-rate of the blocking cations through the channel pore. We propose that the profound effect of membrane lipid modulations on open channel block alleviation, allows the productions of a robust current in response to light in the absence of depolarization.

Open channel block (OCB) is a process in which ions have access to intra-channel binding sites inside the pore of an ion channel and block the flow of ions through that channel. The known mechanism of OCB removal is repulsion of the blocking ion by depolarization. For cyclic nucleotide gated channels, divalent OCB greatly improves the signal to noise ratio (Kaupp and Seifert, 2002). For the NMDA channel, OCB by Mg²⁺ and its removal by depolarization (Mayer et al., 1984) are crucial for the function of this channel as a coincident detector in the brain (Nowak et al., 1984). Accordingly, the ligand-gated NMDA channel cannot open by application of its ligands in resting membrane potential due to OCB, and depolarization is required to remove this block (Kandel, 2000). The Drosophila light activated channels TRPL and TRP also undergo divalent OCB (TRPL- (Parnas et al., 2007), TRP- (Hardie and Mojet, 1995)). Since there is no membrane depolarization to remove the divalent OCB before activation of the TRPL and TRP channels, it is not clear how the robust (>4nA) light induced current develops at light onset. This fundamental difficulty in understanding Drosophila phototransduction has been ignored but needs to be resolved.

The TRP superfamily is conserved throughout evolution and plays important roles in signal transduction in many cells types (Clapham, 2003;Dhaka et al., 2006;Hardie, 2007;Jordt et al., 2003;Minke and Cook, 2002;Montell, 2005;Voets et al., 2004). Experimental evidence has suggested that several members of the TRP family (e.g. TRPC2 (Lucas et al., 2003), TRPC6 (Spassova et al., 2006), TRPV3 (Hu et al., 2006), TRPM6 (Topala et al., 2007) and TRPM7 (Nadler et al., 2001) undergo OCB. However, the physiological mechanism underlying the alleviation of OCB in TRP channels is still unknown. Lipids were found to activate or modulate TRP and NMDA channels. Accordingly, diacylglycerol (DAG) activates TRPC2 (Lucas et al., 2003), TRPC3 (Hofmann et al., 1999), TRPC6 (Hofmann et al., 1999) and TRPC7 (Okada et al., 1999). Polyunsaturated fatty acids (PUFA) activate the Drosophila TRP and TRPL (Chyb et al., 1999), the mammalien TRPV3 (Hu et al., 2006)) and the NMDA channels (Miller et al., 1992). However, the underlying mechanism of the lipids action is not clear.

In the present study, we show that the application of lipids removes divalent OCB without depolarization from both TRPL and NMDA channels, by facilitating the passage rate of the blocking cations in the channels’ pore. We suggest that the effect of lipids is indirect and operates by modulating interactions between membrane lipids and the channels. We propose that the profound effect of membrane lipid modulations on open channel block alleviation, allows the productions of a robust current in response to light in the absence of depolarization.

Materials and Methods

Drosophila

Fly stocks

White-eyed Drosophila of the following strains were used: trp^P343 and norpA^p24;;trp^p343. Flies were raised at 24°C in a 12 h light/dark cycle. For the experiments of Fig. 2C flies were raised in darkness to prevent light induced degeneration of the norpA^p24;;trp^p343 eyes.

Figure 2. The action of LA is not mediated via PLC.

A. PLC activity was monitored by using S2 cells expressing the Drosophila muscarinic receptor (DM1) and eGFP-PH, which binds to PIP₂ and IP₃. Application of 50μM LA, under conditions which activated the channels, did not elicit any change in the eGFP-PH distribution, as monitored by confocal images of the GFP fluorescence (LA) relative to control. Application of carbachol (CCH) elicited a robust translocation of eGFP-PH to the cytosol, which was reversed by CCH removal (wash), thus indicating activation of PLC (n=6). The relative fluorescence intensity at a cross section of the cell (marked by line) is also presented below the confocal images. The time course of the fluorescence changes measured in the cytosol is presented on the right.

B. Western blot analysis of heads homogenate of dark raised WT and norpA^p24;;trp^p343 double mutant flies. Head membrane was extracted with SDS buffer and subjected to Western blot analysis with antibodies specific for the Drosophila proteins TRP, NORPA and Dmoesin as indicated. No TRP and NORPA proteins were detected in the norpA^p24;;trp^p343 double mutant.

C. Left, Representative I-V curves measured by whole cell recordings from photoreceptors of the norpA^p24;;trp^p343 mutant lacking PLC and the TRP channel (see Fig. 2B). In darkness the TRPL channels are closed (red curve 1). The effect of the 60μM LA was not altered by the absence of PLC and a linear I-V curve was obtained (black curve 4). The effect of LA at positive membrane potentials preceded that of negative membrane potentials (Curves 2 and 3, blue and purple respectively) in a similar manner to the results of Figs. 1B. and S2. Application of 10mM Ca²⁺ restored the outward rectifying I-V curve (green curve 5). Inset, enlargement of curves 1, 2 and 5, demonstrating more clearly the outward rectification. Right, the effect of LA on the current is presented at 90mV (black dots) and -90mV (dark red dots) as a function of time. The numbers correspond to the curves presented in Fig. 2C left (n=4).

Light Stimulation

A xenon high-pressure lamp (PTI, LPS 220, operating at 50W, Lawrenceville, NJ) was used and the light stimuli were delivered to the ommatidia by means of epiillumination via an objective lens (in situ). The intensity of the orange light (Schott OG 590 edge filter) at the specimen, with no intervening neutral density filters, was 13 mW/cm² and it was attenuated by neutral density filters in log scale.

Cell Culture

Schneider S2 cells were grown in 25 cm² flasks, at 25 °C in Schneider medium (Beit Haemek biological industries, Israel) supplemented with 10 % fetal bovine serum and 1 % pen-strep. TRPL-eGFP (TRPL accession number NM_165694) channels were stably expressed and NMDA channels (accession numbers NR1 — NM_017010 and NR2 — NM_012574), were transiently expressed using Escort IV (Sigma).

Electrophysiology

For S2 cells, cells were seeded on polylysine coated plates at a confluence of 25 %, 24-72 hours prior to the experiment. 24 hours prior to the experiment, 500μM CuSO₄ was added to the medium to induce expression of the channels. For Drosophila ommatidia, dissociated Drosophila ommatidia were prepared from newly emerged flies (<1 h after eclosion) and whole-cell, patch-clamp recordings were performed as previously described (Peretz et al., 1994). Whole-cell and single-channel currents are recorded at room temperature using borosilicate patch pipettes of 5-8 MΩ and an Axopatch 1D (Molecular Devices, Inc.) voltage-clamp amplifier. For single channel recordings pipettes were coated with Dow Corning sylgard. For S2 cells, voltage-clamp pulses were generated and data captured using a Digidata 1322A interfaced to a computer running the pClamp 9.2 software (Axon Instruments, Inc.). For Drosophila ommatidia, Digidata 1200 and pClamp 8.0 software (Molecular Devices, Inc.) were used. Currents were filtered using the 8-pole low pass Bessel filter of the patch-clamp amplifier at 10 kHz and sampled at 50 kHz (single channel recordings) or filtered at 5 kHz and sampled at 20 kHz (whole cell recordings). To measure I-V curves with minimal distortions, only cells with low (<10 MΩ) series resistance were used and the series resistance was compensated by ∼80 %.

Western blot analysis

Two flies’ heads were used for each lane of the Western blots. Proteins were extracted with 1× SDS-PAGE buffer (2% SDS, 100mM DTT, 10 glycerol in 65 mM Tris-HCl [pH 6.8]) and subjected to SDS-PAGE using 8% polyacrylamide gels (Midget System, Amersham Pharmacia Biotech, Germany). Protein levels were detected using anti-TRP (monoclonal, MAb83F6, a gift of S. Benzer), anti-PLC (polyclonal, a gift of A. Huber) and anti-Dmoesin (polyclonal, a gift of F. Payre) antibodies.

Confocal imaging

For measurements of eGFP-tagged PH domain, optical sections of S2 cells were visualized using the Fluoview confocal microscope (model 300 IX70; Olympus) using Olympus UplanF1 60x/0.9 water objective. Optical sections were recorded from the middle of the cell.

Solutions

For S2 cells, standard extracellular solution contained in mM: 150 NaCl, 5 KCl, 4 MgCl₂, 10 TES, 25 proline, 5 alanine and 0.5 EGTA. Standard intracellular solution contained in mM: 130 CsCl, 10 TES, 2 MgCl₂, 4 Mg-ATP, 0.4 Na-GTP and 15 TEA. For Drosophila ommatidia, the extracellular solution contained in mM: 120 NaCl, 5 KCl, 4 MgSO₄, 10 TES, 25 proline, 5 alanine (1.5mM Ca²⁺ was added when light stimulation was used, to allow PLC activity). The intracellular solution, which blocked K⁺ channels, contained in mM: 120 CsCl, 10 TES, 2 MgSO₄, 4 Mg-ATP, 0.4 Na-GTP, 1 β-NAD and 15 TEA. All solutions were titrated to pH 7.15. Cells were perfused via BPS-8 valve control system (Scientific Instruments) at a rate of ∼30 chambers per min. Chemicals were applied via the perfusion system.

Data analyses

Data was analyzed and plotted using pClamp 9.2 software (Axon Instruments, Inc.) and Sigma Plot 8.02 software (Systat software, Inc.). Kinetic analysis, single-channel event identification and burst analysis were performed as previously described (Parnas et al., 2007). For event histograms, only events longer than twice T_r were used (Colquhoun and Sigworth, 1995). The single channel current amplitude was calculated relative to the closed state base line (as implemented in the pClamp 9.2 softwere). Accordingly an automatic correction for baseline drifts is given and the average current of the closed state remains zero. Therefore, the events histograms present only the channel openings which are already corrected for baseline drifts.

Statistical analysis

For statistical analysis, student’s t-test was used. All Error bars are standard error of the mean.

Rat brain slices

Electrophysiology

For whole cell recordings from hippocampal brain slices we used the experimental procedure as previously described (Crepel et al., 1997). Briefly, Sabra male rats (Harlan, Jerusalem), 28–32 days old, were anaesthetized and euthanized. The brain was rapidly dissected and coronal hippocampal slices (350 μm) were cut on a vibratome VT-1000 (Leica, Germany) in ice-cold artificial cerebrospinal fluid (aCSF), containing in mM: 126 NaCl, 1.6 KCl, 2.4 CaCl₂, 1.2 NaH₂PO₄, 1.2 MgCl₂, 18 NaHCO₃ and 11 glucose, saturated with 95% O₂ and 5% CO₂). Slices were allowed to recover for at least 1-2 hrs in aCSF at 32°C and were then transferred to a chamber with continuous flow (2.5 ml/min) of aCSF that contained 10 μM CNQX (Sigma) and 100 μM picrotoxin (Sigma) to block AMPAR EPSCs and GABA_AR IPSCSs, respectively. Cells were visualized using an upright microscope with infrared illumination. Whole-cell voltage-clamp recordings were performed using Multiclamp 700B amplifier (Molecular Devices). Electrodes pulled from glass capillaries (2–4 MΩ resistance) were filled with a solution (pH 7.2–7.4) containing in mM: 117 CsCH₃SO₃, 20 HEPES, 0.4 EGTA, 2.8 NaCl, 5 N(CH₂CH₃)₄Cl, 2.5 Mg-ATP, 0.25 Mg-GTP. A bipolar stimulating electrode was placed in the Schaffer collateral/commissural afferents within the CA1 region and recording were made from pyramidal hippocampal neurons in the CA1. Afferents were stimulated at 0.1Hz to evoke excitatory postsynaptic currents (EPSCs). EPSCs were filtered at 2kHz, digitized at 5–10 kHz and recorded using Igor Pro software (Wavemetrics, Lake Oswego, OR). Current-voltage (I-V) curves of NMDAR EPSCs were generated by recordings at holding potentials that ranged from -70mV to +40mV, at 10mV increments. Linoleic acid was applied through the patch pipette by means of diffusional exchange.

Results

Linoleic acid (LA) removes divalent OCB from heterologously expressed and native TRPL channels from the photoreceptor cell

Under physiological conditions at steady state, the divalent OCB of both TRPL (Fig. 1B) and TRP (Fig. S2A) channels is manifested in the current-voltage relationships (I-V curves), when the channels are activated either by light (pink I-V curves) or by low concentration of LA (blue I-V curves). These I-V curves revealed strong outward rectification implying that these channels are nearly closed in the physiological voltage range of -60 to 10 mV. Robust openings were only observed at a non-physiological positive voltage range (Raghu et al., 2000). In contrast to the minor currents observed in the I-V curves at -60 mV, whole cell recordings from isolated ommatidia of the trp^P343 (Fig 1B, inset) and trpl³⁰² null mutants (Fig. S2A, inset, which express only the TRPL and TRP channels, respectively, (Scott et al., 1997)) revealed robust transient light induced current (LIC, >4 nA) at the onset of the response to a light pulse at -60 mV membrane potential (Figs. 1B, inset and S2A, inset). OCB is a very fast and efficient process (Hille, 1992) and every activated channel is expected to be immediately blocked upon opening. Since there is no membrane depolarization before TRPL and TRP channels open to remove the divalent OCB observed in the I-V curves, it is not clear how these robust LICs develop in response to light (Figs 1B and S2A, insets). The apparent contradiction between the results of the I-V curve and the LIC (Figs. 1 and S2A), presents a fundamental difficulty in understanding Drosophila phototransduction, which needs to be resolved.

Figure 1. Linoleic acid removed OCB from TRPL channels.

A. Left, Representative I-V curves measured from S2 cells by whole cell patch clamp recordings using voltage ramps from -100mV to 100mV in 1s. The typical outward rectification of the TRPL channels (red curve 1) was modified to a linear I-V curve after application of 40μM LA (black curve 4). The effect at positive membrane potentials preceded that of negative membrane potentials (Curves 2 and 3, blue and purple respectively, see also 1A right). Application of 5mM Ca²⁺ restored the outward rectifying I-V curve (green curve 5). Right, The current values at 90mV (black dots) and -90mV (dark red dots) are presented as a function of time. The numbers correspond to the curves presented in Fig. 1A left (n=15).

B. Left, Representative I-V curves measured as in A from mutant Drosophila ommatidia that express only TRPL channels (trp^P343). In darkness the TRPL channels were closed (red curve 1). After application of 60μM LA, a linear I-V curve was obtained (black curve 3). The effect of LA at positive membrane potentials preceded the effect at negative membrane potentials in a similar manner to expressed channels (blue curve 2). Application of 10mM Ca²⁺ blocked the TRPL channel (green curve 4, left), ruling out the possibility that the linear I-V curve was due to leak current. Right, the effect of LA on the current is presented as a function of time at 90mV (black dots) and -90mV (dark red dots). The numbers correspond to the curves presented in Fig. 1B left (n=5). The I-V curve of the LIC is presented in pink (the maximal light intensity was attenuated by 2 log units). The inset shows the LIC of the trp^P343 mutant in response to the above light intensity.

A possible explanation for this contradiction is that light and LA have a dual role: i) to activate the channels and ii) to remove OCB. A convenient preparation to isolate channel activation from removal of OCB is heterologously expressed TRPL channels in Schneider 2 (S2) cells.

In Drosophila photoreceptor cells, the TRPL and TRP channels are closed in the dark and cannot be activated by depolarization (see Fig 1B, red curve 1). In contrast, TRPL channels heterologously expressed in S2 cells are constitutively active (Fig. 1A, red curve 1), but otherwise have the same properties as the native channels of the photoreceptor cell (Hardie et al., 1997). Importantly, the constitutively active TRPL channels still reveal OCB, as manifested by a strong outward rectification (Fig. 1A, red) and very little single channel activity (Fig. 5A, left) at negative membrane potentials (Parnas et al., 2007). This property of the expressed TRPL channels enables us to isolate OCB removal by lipids from channel activation. While functional TRPL channels can be easily expressed, functional expression of the TRP channel could not be obtained (Minke and Parnas, 2006). For this reason, the study of the mechanism of OCB removal without depolarization was carried out on heterologously expressed TRPL channels and the main findings were confirmed in the native system of the photoreceptor cells.

Figure 5. LA increased the passage rate of cations through the channel pore.

A. Inside-out patch clamp recording from S2 cells expressing TRPL in symmetrical Na⁺ solution (see Materials and Methods).

Representative single channel recordings at ±100mV membrane potentials (top) and a representative voltage ramp from -100mV to 100mV (bottom) are shown (n=5).

B. The paradigm of A was repeated after application of 10μM LA (data recorded from the same cell as in A). An increase in channel activity was observed especially at negative membrane potentials (n=5).

C. Histogram showing the apparent open probability of the TRPL channel which was calculated from the step traces of A and B. At -100mV membrane potential, the measured open probability was 0.004, while at 100mV membrane potential, the measured open probability was 0.1 (i.e. a 25 fold larger at positive membrane voltage). After application of 10μM LA, the measured open probability increased to 0.11 and 0.17 at negative and positive membrane voltages, respectively. Thus, at 100mV there was only a 1.5 fold increase in the open probability in the presence of LA (n=5).

D. LA increased the unitary channel conductance. top, Inside-out patch clamp recordings in standard bath solution at 100mV holding potential revealed TRPL channel activity before (red trace) and after (black trace) application of 10μM LA. bottom, Amplitude histogram of the channel openings before (red dots) and after (black dots) LA application. The measurements showed that LA increased the single channel current from 5.5pA to 7.1pA (p-value<0.001, n=7).

E. Whole cell recordings from S2 cells expressing the TRPL channel under bi-ionic conditions (the extracellular solution contained DE-amine (a large monovalent organic cation), while the intracellular solution contained Cs⁺ (see scheme). The experiment was performed under divalent free conditions. I-V curve measurements before and after application of 40μM LA, are presented. No change in E_rev was observed. Inset, The I-V curve of E is presented in a magnified current scale (n=6).

F. I-V curves were measured by whole cell recordings from S2 cells expressing the TRPL channel under bi-ionic conditions with either DE-amine or Na⁺ as the extracellular cation (Cs⁺ was the intracellular cation, i.e. in divalent free conditions). No change in E_rev between the two conditions was observed. Inset, The I-V curve of F is presented in a magnified current scale (n≥6).

G. Single channel recordings from excised inside-out patch clamp measurement at 100mV membrane potential when Na⁺ (130mM) was the sole ion (top). When Na⁺ was replaced by DE-amine (130mM) the single channel current was largely reduced (bottom) (n=4).

OCB is characterized by an outwardly rectifying I-V curve. Removal of this block results in linearization of the I-V curve. Therefore, we first examined whether application of LA results in linearization of the TRPL I-V curve in the presence of divalent cations. To this end, we measured I-V curves every 5s from TRPL channels expressed in S2 cells. Prior to application of LA, outward rectification of the constitutively active TRPL channels was observed with 4mM Mg²⁺ (Fig. 1A left, red curve 1) indicating OCB by Mg²⁺. Application of LA (40μM) gradually caused linearization of the I-V curve (Fig. 1A, black curve 4). Since LA could not be washed out, we applied a high (5mM) concentration of Ca²⁺, which restored the outwardly rectifying I-V curve by shifting the equilibrium of the divalent cations-binding sites at the channel pore towards the bound states (Hille, 1992) (Fig. 1A left, green curve 5. For statistical analysis see Supplemental material, Fig. S1A). To confirm that the effects seen in Fig. 1A are specific, control experiments were conducted with the following results: i. Application of up to 100μM LA did not show any effect on S2 cells, in untransfected S2 cells (Supplemental material, Fig. S1A). ii. In the absence of LA, no increase of channel openings was observed (Supplemental material, Fig. S1B). iii. Application of the water soluble isoform of LA (Na-LA), which does not form micelles in aqueous solution, yielded the same results as LA (data not shown). In addition, application of other lipids, such as oleic acid and stearoylarachidonylglycerol, an analog of the PLC product DAG, revealed effects similar to those produced by LA (data not shown).

To examine whether the application of LA also results in linearization of the I-V curve of the native TRPL and TRP channels in Drosophila photoreceptor cells, we performed whole cell recordings from isolated ommatidia of the trp^P343 and trpl³⁰² null mutants, which express only the TRPL and TRP channels, respectively (Scott et al., 1997). We carried out the same protocol as that in Fig. 1A and obtained similar results from the native (Fig. 1B) and the heterologously expressed TRPL channels (Fig. 1A). For the native TRP channels, a full linearization of the I-V curve was not observed, but the outwardly rectifying I-V curve was changed into a dually rectifying curve, similar to that observed under divalent free conditions (Fig. S2A, (Hardie and Minke, 1994)).

The results of Fig. 1 are consistent with the notion that in active TRPL and TRP channels, both application of LA at resting membrane potential or depolarization remove OCB by divalent cations (Parnas et al., 2007).

The effect of LA is not mediated via activation of PLC

It was previously suggested that the effects of LA on the TRPL channel is relayed via PLC (Estacion et al., 2001;Minke and Parnas, 2006). To examine this notion, we tested the effect of LA on PLC activity in S2 cells. It is well known that activation of PLC results in the hydrolysis of phosphatidylinositol 4,5 bis phosphate (PIP₂) into DAG and Inositoltrisphosphate (IP₃, (Berridge and Irvine, 1989)). In order to measure PLC activity in situ, we co-expressed in S2 cells the Drosophila muscarinic receptor DM1 (which activates PLC via a Gq-protein (Hardie et al., 1997)) along with eGFP-tagged pleckstrin homology (PH) domain from PLC-δ1 (eGFP-PH, which binds to PIP₂ and IP₃ (Balla and Varnai, 2002;Suh et al., 2006). The latter was expressed to directly show PIP₂ hydrolysis indicating PLC activity, because of its high affinity binding to both PIP₂ and IP₃. When PLC is inactivated eGFP-PH is located at the plasma membrane bound to PIP₂. Upon activation of PLC, eGFP-PH translocates from the plasma membrane to the cytosol due to the hydrolysis of PIP₂ and the production of IP₃ (Balla and Varnai, 2002;Suh et al., 2006).

In control conditions before application of LA, strong eGFP fluorescence was observed at the plasma membrane, indicating localization of PIP₂ in the plasma membrane (Fig. 2A control). Application of 50μM LA, which is sufficient to activate TRPL channels (Fig. 1A), did not change eGFP-PH distribution (Fig. 2A, LA), indicating that PLC was not activated. Application of carbachol (CCH), an activator of DM1, which was used as a positive control, elicited a robust translocation of eGFP-PH to the cytosol, indicating activation of PLC (Fig. 2A CCH). The translocation of eGFP-PH was reversible after washing the CCH from the cells (Fig. 2A, wash, see also 2A, right). Together, figures 1A and 2A show that LA activates the TRPL channels without activation of PLC as monitored by hydrolysis of PIP₂.

To examine if the above conclusion is also valid for the native TRPL channels in the photoreceptor cells, we constructed the double mutant norpA^P24; trp^P343 (Fig. 2B) and repeated the experiments described in Fig. 1B. The norpA^P24; trp^P343 double mutant lacks the TRP channels and virtually lacks the eye PLC (NORPA, (Bloomquist et al., 1988), Fig. 2B). Results very similar to those of Fig. 1B were obtained following the application of LA to the norpA^P24; trp^P343 mutant (Fig. 2C), indicating that both TRPL activation by LA and removal of OCB by LA do not require PLC.

Together, the experiments described in Fig 2 ruled out the possibility that LA action is mediated via PLC and suggest a direct effect of LA on the channels in both experimental systems (Hardie et al., 2003)).

A reduction of OCB by activation of PLC in both expressed and native TRPL channels

We next examined whether removal of OCB by exogenous application of LA is pertinent to alleviation of OCB under more physiological conditions. Although the effect of LA is not mediated through activation of PLC (Fig. 2), it has been well established that PLC activation hydrolyses PIP₂ and produces DAG and PUFAs (Leung et al., 2008). In addition, under physiological conditions PLC is necessary for TRPL activation (Cook et al., 2000). Since OCB removal is required for production of a robust TRPL and TRP-mediated current under physiological conditions, it was interesting to examine whether activation of PLC alleviates OCB in a manner similar to that of LA action in these channels. To monitor PLC activity we used the experimental tools described in Fig. 2 and co-expressed TRPL channel together with the Drosophila muscarinic receptor DM1 and eGFP-PH in S2 cells ((Balla and Varnai, 2002;Suh et al., 2006) and see Fig. 2). Under control conditions, the plasma membrane is strongly marked by eGFP fluorescence due to binding of eGFP-PH to PIP₂ in the plasma membrane (Fig. 3A, control). Upon application of CCH, eGFP-PH translocated to the cell body (Fig. 3A, CCH). This translocation was reversible with the removal of CCH (Fig. 3A, wash, see also 2A, right). Thus, figure 3A demonstrates reversible activation of PLC by CCH as monitored by hydrolysis of PIP₂. Application of CCH to S2 cells co-expressing DM1, TRPL and eGFP-PH caused a near linearization of the TRPL I-V curve (Fig. 3B, left). When the DM1 receptor was not expressed, application of CCH did not induce any inward current (Fig. 3B, right) or translocation of eGFP-PH (data not shown). In order to quantify the effect of CCH on the TRPL outward rectification, we compared the current ratio (I_80mV/I_-80mV) after application of CCH to the control. The current ratio was reduced from 5.26±0.74 to 1.85±0.19 (n=6), demonstrating a significant decrease in the outward rectification following activation of PLC. This result demonstrates that OCB of TRPL can be reduced by receptor-activation of PLC.

Figure 3. A reduction of OCB by activation of PLC.

A, B: Activation of the DM1 receptor by CCH activated PLC and removed divalent OCB in S2 cells.

A. A representative series of confocal images of S2 cells co-expressing eGFP-tagged PH domain, TRPL channels and DM1. In control conditions, the plasma membrane is strongly marked by eGFP-PH fluorescence due to its binding to PIP₂ (control). The relative fluorescence intensity at a cross section of the cell (marked by line) is also presented below the confocal images. Application of CCH to the bathing solution, in a concentration that activated the TRPL channels (B, 100μM CCH) induced movement of the eGFP-tagged PH domain to the cell body (CCH), thus indicating the activation of PLC and hydrolysis of PIP₂. Subsequent wash of CCH from the bathing solution resulted in reversible marking of the plasma membrane with eGFP-tagged PH domain (wash, n=6). The time course of the fluorescence changes measured in the cytosol is presented on the right.

B. Left, whole cell recordings from S2 cells as in A. Application of 100μM CCH to S2 cells expressing both TRPL and DM1 increased the inward currents relative to the outward current and a nearly linear I-V curve was observed (black curve). Right, control S2 cells expressing the TRPL channel and the eGFP-PH domain but not the DM1 receptor show no effect of carbachol (n=6).

C. Illumination reduced the outward rectification of the native TRPL channels. Whole cell recordings from isolated ommatidia of the trp^p343 null mutant that expresses only the TRPL channels. The membrane potential of the photoreceptor cell was held at -80 mV and a constant light pulse was applied, which elicited an inward current through the TRPL channels. Then, at various time points during the rise time of the LIC, the membrane voltage was stepped to 80 mV and two current values at -80mV and 80mV were measured. The traces show an example of early and late time points.

D. A summery of experiments as presented in C. The current ratio at 80mV and - 80mV was calculated and presented as a function of time from the beginning of the LIC (from the time that the LIC reached 50pA). It is seen that as the LIC progressed the ratio became closer to 1, indicating a reduction in the OCB.

E. A histogram of the results presented in D. The first bin (0-50ms) significantly differs from the rest of the bins (p-value<0.001, n≥5).

We next examined whether light stimulation, in addition to native channel activation, also removes OCB. In Drosophila photoreceptor cells, it is well known that light activates the TRPL and TRP channels via PLC (Bloomquist et al., 1988;Devary et al., 1987). Therefore, we examined whether light activation of PLC reduces the outward rectification of the native TRPL and TRP channels. To this end we performed whole cell recordings from isolated ommatidia of the trp^P343 and trpl³⁰² null mutants, which express only the TRPL and TRP channels, respectively (Scott et al., 1997). We had to modify our experimental protocol to accommodate the physiological properties of the native channels. Since the rising phase of the light response to medium light intensity is very fast (<100ms Fig. 3C, (Peretz et al., 1994)), voltage ramps could not be used to measure I-V curves. Therefore, the following protocol was applied: The membrane potential of the photoreceptor cells was held at - 80 mV and a constant light pulse was applied, which elicited an inward current through TRPL (Fig. 3C) and TRP (Fig. S2B, C) channels. Then, at various time points during the rise time of the light induced current, the membrane voltage was stepped to 80 mV and two current values at -80mV and 80mV were measured (Figs. 3C and S2B, C). The traces presented in figure 3C, D show examples of early and late time points of the TRPL current. The current ratio (I_80mV/I_-80mV) at the early time point was 5.01, while at the late time point the current ratio was reduced to 2.12, showing a reduction in the outward rectification (Fig. 3C). These experiments are summarized in figures 3D, E and S2B, C, which plot the current ratio as a function of time during the LIC. These results show that light activation of the native TRPL and TRP channels reduces the outward rectification of these channels during the rise time of the light response. This suggests that light removes OCB from these channels in timescale of <50 ms and hence allows the production of a robust LIC in the presence of divalent cations (Figs. 1 and S2A).

The results of Figs. 1-3 demonstrate that exogenous application of lipids removes divalent open channel block without depolarization from constitutively active expressed TRPL channels. The results also show that LA initially activated the native TRPL and TRP channels and then removed their OCB without depolarization, suggesting that LA has a dual role of both activation and OCB removal. The relevance of OCB removal by LA to the physiological response to light was demonstrated at two levels: i) PLC activation removed OCB from expressed TRPL channels in a manner similar to exogenous application of lipids and ii) light also alleviated OCB of native TRP and TRPL in <50 ms and allows production of robust transient currents despite the existence of OCB mechanism. Since in photoreceptor cells depolarization or Ca²⁺ removal in the dark do not open the TRP and TRPL channels, it is clear that removal of OCB by itself is not sufficient to activate these channels. Rather, OCB removal is a necessary but not sufficient step for light activation (as for the NMDA channel). The above observations led us to examine the mechanism underlying OCB alleviation by lipids without depolarization.

LA reduces the blocking efficiency of divalent cations

In addition to linearization of the I-V curve by LA, Fig. 1A revealed that the increase in currents was faster at positive membrane potentials than at negative ones (Fig. 1A left, blue (2) and purple (3) curves). To better illustrate this non-symmetrical effect of LA, we present the temporal behavior of the I-V curves by plotting the current values at ±90mV as a function of time (Fig. 1A right, the numbers correspond to the I-V curves seen in Fig. 1A, left). It has been shown that depolarization reduces the blocking efficiency of divalent cations on the TRPL channel (Parnas et al., 2007). The observation that LA had a faster effect at positive than at negative membrane potentials raises the possibility that LA acts in synergism with depolarization to reduce the blocking efficiency of divalent cations. To test this notion, we examined the magnitude of OCB as a function of Mg²⁺ concentration. In the absence of LA the blocking efficiency of Mg²⁺ is voltage dependent (Fig. 4A). In the presence of LA, at negative membrane potentials, a significantly higher concentration of Mg²⁺ was needed to reach 50% inhibition, while the blocking efficiency of Mg²⁺ became voltage independent (Fig. 4B).

Figure 4. LA reduced the blocking efficiency of divalent cations on the TRPL channels.

A. Measurements of normalized whole cell currents (I/I_max) in response to voltage ramps, at the indicated voltages, as a function of Mg²⁺ concentration. The graphs present measurements at three holding potentials, -90mV (black curves), 50mV (red curves) and 90mV (green curves), before application of LA. Relatively low concentrations of Mg²⁺ were required to block the current at negative membrane potential, as compared to positive membrane potentials (n=5).

B. The paradigm of A was repeated after application of LA. Only a small difference between curves, which were measured at positive and negative membrane potentials, was observed. This indicates that LA removed the voltage dependence of the Mg²⁺ block. Accordingly, EC₅₀ at -90mV was 0.8±0.01mM in the control (A) and it was shifted to 9.35±2.4mM following LA application. The effect of LA on the EC₅₀ value decreased with the increase in depolarization. At 50mV, the control EC₅₀ was 7.9±0.8mM (A) and it was shifted to 13.8±3.6mM following LA application (n=5).

Thus, Figure 4 shows that LA reduced the blocking efficiency of divalent cations in a similar manner to that of positive membrane potentials. The reduced blocking efficiency of divalent cations, especially at negative membrane potential, abolishes the outward rectification of the TRPL channels until a linear I-V curve is obtained. This occurs despite the presence of divalent cations.

LA increases cations passage rate through the channel pore

Since LA reduces the blocking efficiency of divalent cations, we examined whether LA exerts this effect on the channel pore where the block occurs (Chung et al. 2005). To this end we carried out single channel recordings of TRPL expressed in S2 cells. We first demonstrated that linearization of the I-V curve by LA is also observed in single channel measurements. Inside-out patch clamp recordings (4mM Mg²⁺) showed that prior to LA application there is a higher open probability at positive membrane potentials than at negative ones (Fig. 5A, and see (Parnas et al., 2007)). After application of 10μM LA, a significant increase of the channels’ open probability was observed, mainly at the negative membrane potentials, and the open probability was almost independent of membrane potential (Fig. 5B, C).

The effect of LA on the channel pore was examined by measuring the single channel current. These measurements showed that LA increased the single channel current from 5.5pA to 7.1pA at 100mV membrane potential (Fig. 5D, p-value < 0.001). For a reliable single channel current analysis, only opening events longer than 2T_r were used (Colquhoun and Sigworth, 1995) (T_r is the rise-time of channel opening, see Methods).

The observed increase in single channel current could arise from an increase in the permeability of the channel. To measure ionic permeability ratios, we used the procedure that was previously described (Reuss et al., 1997). Accordingly, we measured the reversal potential under bi-ionic conditions in which Cs⁺ was the only cation included in the pipette solution. The bath contained one of a variety of monovalent cations (see scheme in Fig. 5E). Under these conditions, the permeability can be derived from the bi-ionic reversal potential (E_rev). Using this approach, we examined the effect of LA on E_rev of three monovalent organic cations of increased diameters: diethylamine (DE-amine), triethylamine (TriE-amine) and tetraethylamine (TE-amine). Monovalent organic cations were used for two reasons: i. they allow a systematic stepwise increase of the monovalent cationic diameter. ii. The permeability of inorganic monovalent cations is very fast and changes in their permeability might be too small to be detected. We observed the following differences among the permeabilities of the various monovalent organic cations: TE-amine did not pass through the TRPL channel and showed no inward current (E_rev of -48.2±3.1, Supplemental material Fig. S3B), while DE-amine (E_rev of 8.8±1.9, Fig. 5E) and TriE-amine (E_rev of -8.5±1.9, Supplemental material Fig. S3A) generated a small inward current. Application of LA to the extracellular solution, which contained only DE-amine resulted in a large increase of the inward currents (Fig. 5E). LA application to the TriE-amine containing solution also increased the inward currents but to a lesser extent (Supplemental material Fig. S3A). Application of LA to a TE-amine containing solution did not induce any inward current (Supplemental material Fig. S3B). Importantly, the reversal potential during application of LA did not change in any of the experiments, thus indicating that no change in TRPL permeability was induced by LA, despite the large increase in current (Fig. 5E and Supplemental material Fig. S3). In addition, the data showed that in the absence of LA, all the organic cations that were examined produced outward rectification despite the lack of divalent cations. This indicates that large monovalent organic cations also block the open channel (see below, Fig. 5E and Supplemental material Fig. S3).

How can the effect of LA on the TRPL channels be explained, when no effect was found on the channel permeability? Hille and colleagues have previously investigated the mechanism underlying the passage of large organic cations through the K⁺ channel pore (Sanchez et al., 1986). They found that although K⁺ ions show a linear I-V curve, organic cations, that had similar permeability to that of K⁺ ions, revealed outwardly rectifying I-V curves. In addition, they found that the single channel current was dramatically reduced in the presence of the organic cations. These authors concluded that the organic cations blocked the open pore by a slow passage through the channel pore. According to this notion, linearization of the I-V curve is expected to occur when the rate of ion passage through the channel pore is increased. To examine whether DE-amine (before application of LA) produced the outward rectification by a slow passage through the pore (Fig. 5E), we examined the permeability and single channel current of DE-amine relative to those of Na⁺. Fig. 5F shows that DE-amine and Na⁺ (when Cs⁺ was the intracellular cation) showed the same E_rev, thus indicating that they have the same permeability. However, the single channel current was largely reduced by DE-amine (Fig. 5G). The similarity between the results of Fig. 5 and the previous study of Hille and colleagues on the K⁺ channels (Sanchez et al., 1986) suggests that the blocking effect of both large organic and divalent cations is by slow passage through the channel pore. The slow passage of the blocking cations produces only a small inward current by itself, while the large outward current is obtained by the small inorganic cation (Cs⁺) after repulsion of the blocking cation. Therefore, the effect of LA is by facilitating the passage of the blocking cation through the pore.

Further support for our notion that LA increases the rate of passage of cations through the channels was obtained by analysis of the mean open time and bursting behavior of the channel according to a kinetic model of the TRPL channel. For a detailed explanation see mathematical model in the Supplemental material and Fig. S4.

All of the above results indicated that LA facilitates the flow rate of large cations through the TRPL channel, suggesting that this is the mechanism by which LA reduces the blocking efficiency of divalent cations and alleviate OCB.

Modulation of membrane properties alleviates OCB

Diverse lipids and lipophilic compounds are known to activate or modulate the activity of TRP (Chyb et al., 1999;Hu et al., 2006) and other channels (Casado and Ascher, 1998;Lundbaek et al., 2005;Patel et al., 2001). This raises the possibility that lipids do not affect the TRPL channel by direct and/or specific binding, but rather by affecting the properties of the plasma membrane at the channel-lipid boundary (Reaves and Wolstenholme, 2007). To examine this possibility, we employed procedures that are known to affect plasma membrane properties. Hypoosmotic solutions that cause cell swelling strongly affect the interaction of channels with the surrounding lipids of the plasma membrane (Hamill, 2006;Janmey and Kinnunen, 2006;Spassova et al., 2006). We therefore examined whether hypoosmotic solutions change the I-V curve. Two hypoosmotic solutions of 290 mOsm and 255 mOsm were used. Concomitantly with the increase in surface area (Fig. 6A), the inward current was reversibly enhanced and a more linear I-V curve was observed (Fig. 6B, for a statistical analysis and control see Supplemental material, Fig. S5A).

Figure 6. Removal of OCB by membrane lipid modulations.

A. Demonstration of the cell swelling by the hyposomotic solutions used in B. The plasma membrane is strongly marked by GFP fluorescence due to binding of eGFP-tagged PH domain to PIP₂ in the plasma membrane (see Fig. 2). The 255mOsm solution increased the surface area of the cells by ∼20% (corresponding to ∼125μm², n=4).

B. Hypoosmotic solutions had a similar effect as LA. S2 cells expressing the TRPL channel were bathed in an extracellular solution of 330 mOsm. Perfusion of 290 mOsm and 255 mOsm solutions (as indicated) enhanced the TRPL dependent whole cell current, mainly at negative membrane potentials, as indicated by current measurements at -90mV as a function of time (top) and by the series of I-V curves (bottom, n=4). S2 cells that did not express the TRPL channels showed no response to hypoosmotic solutions (Supplemental material, Fig S6).

C. Application of 30μg/ml polylysine via the patch pipette in whole cell recordings generated a nearly linear I-V curve which was reversed to outward rectification upon increase of extracellular Ca²⁺ concentration to 5mM (n=6).

D. Top, Inside-out patch clamp recordings revealed a robust increase of the open probability of the TRPL channels after exposure to 30μg/ml polylysine. Bottom, The change in open probability was accompanied by an increase in the amplitude of single channel current at 100mV membrane potential as illustrated by an amplitude histogram of the single channel current. The measurements showed that polylysine increased the single channel current from 5pA to 7.1pA (p-value<0.001, n=5).

E. The effect of 40μM LA on the TRPL dependent whole cell current was blocked by application of 10μM GsMTx-4 toxin, a specific inhibitor of mechanosensitive channels (Curves 2 and 3, blue and purple respectively). Washing the water soluble GsMTx-4 rapidly restored the outwardly rectifying I-V curve to a nearly linear I-V curve (black curve 4). Application of 2mM La³⁺ blocked the current (green curve 5, n=4).

An additional method to affect the membrane properties and hence the TRPL-membrane lipid boundary is to reduce the concentration of PIP₂ in the inner leaflet of the plasma membrane (Janmey and Kinnunen, 2006). Application of polylysine is known to sequester PIP₂ from the membrane and to affect the activity of mammalian TRP channels (Stein et al., 2006). Exposing the intracellular membrane to polylysine linearized the I-V curve of the TRPL channel (Fig. 6C), which was reversed into outward rectification by application of 5 mM Ca²⁺. Furthermore, application of polylysine in excised inside-out patch clamp experiments greatly increased the open probability of the TRPL channel (Fig. 6D, upper). This increase was accompanied by an increase in single channel current at 100mV membrane potential (from 5pA to 7.1 pA, p-value< 0.001), similar to that observed for application of LA (compare Figs. 5D and 6D, lower panels). This result suggests that both LA and polylysine exert a similar effect on the TRPL channel.

Another approach to test the above notion is to block the effect of lipids with a venom toxin. The tarantula peptide GsMTx-4 specifically blocks a range of stretch-activated channels, but not by specific interaction with the channel proteins themselves but rather by modification of the channel-lipid boundary (Hamill, 2006;Spassova et al., 2006;Suchyna et al. 2004). Figure 6E shows that the GsMTx-4 toxin did not block the constitutive activity of the expressed TRPL channel. However, in the presence of the toxin, the effect of LA was small and rapidly blocked. Washout of the toxin revealed the full extent of the effect of LA on the TRPL channel (for statistical analysis see Supplemental material, Fig. S5B). This result further supports the notion that LA affects the TRPL-membrane lipid boundary.

Although the results of each experiment (Fig. 6) can lead to several possible interpretations (Schnitzler et al. 2008;Spassova et al., 2006), the ensemble of all experiments is in favor of a mechanism operating by modulating the plasma membrane properties.

LA removes OCB from heterologously expressed and native post synaptic NMDA channels of hippocampal brain slices

The ionotropic glycine and glutamate activated NMDA channel plays a key role in the induction of many forms of synaptic plasticity (Kandel, 2000). The NMDA channel is a classical model for a channel that undergoes OCB by Mg²⁺, which is alleviated by depolarization (Mayer et al., 1984). However, unlike TRPL, the NMDA channel is not activated by LA in the absence of agonist (data not shown). Therefore, it was interesting to examine whether LA also alleviate the OCB of NMDA channels. In addition, it is well known that NMDA channel activation by ligand binding and OCB removal by depolarization are two distinct processes (Kandel, 2000). Therefore, the use of the NMDA channel further allows the isolation of the effect of LA on OCB from channel activation.

S2 cells, expressing the two subunits of the NMDA channel, NR1 and NR2B (Sucher et al., 1996), were activated by NMDA and glycine. In the presence of 4mM external Mg²⁺, an outwardly rectifying I-V curve was observed (Fig. 7A, left, red curve 1). Similar to the TRPL channel, addition of 40 μM LA enhanced first the outward current (Fig. 7A, left, blue curve 2) and then the inward current, reaching close to a linear I-V curve (Fig. 7A, left, black curve 3). To better illustrate the non-symmetrical effect of LA on the NMDA channel, we present the temporal behavior of the I-V curves at ±90mV and results similar to those found for the TRPL channel were observed (Fig. 7A, right). Increasing the external Mg²⁺ to 50mM blocked the inward current (Fig. 7A, left, green curve 4, for statistical analysis see Supplemental material Fig. S5C). In addition, no change in E_rev was observed (Fig. 7A) as was also true for TRPL channel. For the NMDA channel, it was shown that at positive membrane potentials there is no Mg²⁺ block (Perouansky and Yaari, 1993). Yet Fig. 7A shows that LA causes an increase of current at positive membrane potentials. This result is in agreement with previous studies, showing potentiation of the currents that are mediated by expressed NMDA channels following application of lipids, although the underlying mechanism is still unknown (Kloda et al., 2007;Miller et al., 1992). However, the linearization of the I-V curve (Fig. 7A) indicates, that in addition to the observed potentiation there is also removal of the Mg²⁺ OCB. In addition, we carried out single channel analysis of the NMDA channel. Similar to TRPL channels, application of LA (10 μM) increased the single channel current from 11pA to 13pA at 100mV membrane potential (p-value < 0.001, Fig. 7B). In addition, LA enhanced the open probability of the single channel activity (Fig. 7C).

Figure 7. LA removed Mg²⁺ OCB from the NMDA channel.

A. Left, LA affected the NMDA channel in a similar manner to its effect on the TRPL channel. Whole cell recordings from S2 cells expressing the two subunits of the NMDA channel, NR1 and NR2B. The channel was activated by combined application of NMDA (30μM) and glycine (30μM) (red curve 1). Application of 40μM LA resulted in a change of the I-V curve from outward rectification to almost linear curve (black curve 3). The effect of LA at positive membrane potentials preceded its effect at negative membrane potentials (blue curve 2). Application of 50mM Mg²⁺ restored the outward rectifying I-V curve (green curve 4). Right, the effect of LA is presented as a function of time, by presenting the current value at 90mV (black dots) and -90mV (dark red dots). The numbers correspond to the curves presented in Fig. 7A left (n=4).

B. top, Single channel activity of the NMDA channel before (red trace) and after (black trace) application of 10μM LA. Traces were obtained from inside out patch clamp recordings using standard bath solution at 100mV holding potential. bottom Amplitude histogram of the channel openings before (red dots) and after (black dots) LA application. The measurements showed that LA increased the single channel current from 11pA to 13pA (p-value<0.001, n=3).

C. Inside-out patch clamp recordings from S2 cells expressing the NMDA channel showed an increase in the open probability of the channels after application of 10μM LA (n=3).

To further examine the effect of LA on removal of OCB from NMDA channels under physiological conditions, we examined NMDA channel mediated currents that were elicited by afferent stimulation in hippocampal brain slices. To this end, whole cell recordings from CA1 pyramidal cells of rat brain slices were conducted. Electrical stimulations of the Schaffer collateral afferent neurons elicited pharmacologically isolated NMDA channels-mediated Excitatory Post Synaptic Currents (EPSCs) in the CA1 pyramidal cells (see Methods, (Crepel et al., 1997)). Figure 8A shows a series of NMDA channel mediated EPSCs at various membrane potentials in the presence of 1.2 mM extracellular Mg²⁺ ions, which blocked the EPSCs only at negative membrane potential (Fig. 8A, left). Strikingly, repeating the protocol of the control (Fig. 8A, left) in the presence of LA in the recording pipette induced a significant increase in the EPSCs at negative membrane potentials, despite the presence of the blocking Mg²⁺ ions (Fig. 8A, right). Figure 8B summarizes the data of experiments similar to that of Fig. 8A. The absolute values of the EPSCs’ peak without (control) and with LA are plotted as a function of membrane potential. In the control experiments a typical voltage dependent I-V curve of native NMDA channels was observed (Crepel et al., 1997) with little change up to -40 mV and with a full removal of Mg²⁺ OCB at -20 mV membrane potential and above, as manifested by a linear I-V curve. Importantly, in the presence of LA in the pipette, the negative slope of the I-V curve was significantly shifted to a more negative membrane potential. Furthermore, a significant increase of the inward current already at -50 mV was observed, with a full removal of Mg²⁺ OCB already at -30 mV membrane potential and above (Fig. 8B). In contrast to the expressed NMDA channels (Fig. 7A), there was no increase (relative to control) in the absolute currents at positive membrane potentials after application of LA. This result indicates that LA had no significant effect after a full removal of Mg²⁺ OCB, thus ruling out potentiation of NMDA channel currents by LA under physiological conditions.

Figure 8. LA removed Mg²⁺ OCB from the native NMDA channel in hippocampal CA1 pyramidal neurons.

A. Representative NMDA channel mediated EPSCs recorded at different holding potentials in the absence (control) and presence (in the pipette) of 30μM LA (LA) and in the presence of 100μM picrotoxin and 10μM CNQX in the extracellular solution. EPSCs were evoked orthodromically and recorded in a CA1 pyramidal cell layer using whole cell voltage-clamp.

B. A plot of the EPSCs’ peak amplitudes (not normalized) vs. membrane potentials for the control (empty circles, n=5) and LA (filled circles, n=7). Error bars are SEM.

Together, the results demonstrate a remarkable commonality between the TRPL and NMDA channels: both revealed removal of divalent OCB by lipids without depolarization.

Discussion

We have shown here a novel mechanism of OCB alleviation. The data suggest that OCB removal by LA arises from an increased flow rate of the blocking divalent cations through the channel pore. Modulation of the channel-membrane lipid interface might underlie the increase in the flow rate of the blocking cations and the removal of the OCB. We applied various methods to modify membrane lipid properties of the TRPL channel including: stretch of the plasma membrane by hypoosmotic solutions, sequestration of PIP₂ by polylysine and application of GsMTx-4 toxin and various lipids such as LA, OA and SAG. Although the results of each of these experiments can lead to several possible interpretations, the ensemble of all experiments is in favor of a mechanism operating by modulation of the plasma membrane properties, which affect channel-membrane lipid interactions. The results thus suggest that lipids do not affect the TRPL channel as second messengers but rather as modifiers of membrane lipid-channel interactions.

The TRPL and NMDA channels revealed striking similarity in the mechanism underlying their OCB and its removal. Both channels undergo OCB by divalent cations, which was removed by depolarization (Mayer et al., 1984;Parnas et al., 2007) and they were similarly affected by lipids. Accordingly, LA caused linearization of the I-V curve; the outwardly rectifying I-V curve could be restored upon increasing the concentration of the divalent cations; LA increased the single channel current and no change in E_rev was observed following application of LA. Taken together, these results suggest that a common mechanism underlies OCB removal by lipids from the TRPL and NMDA channels. For the NMDA channel, activation by ligands and removal of OCB are two distinct processes (Kandel, 2000). However, this is not the case for the TRPL channel, in which activation of PLC underlies both channel activation and removal of OCB. In the heterologously expressed TRPL, the channels are already in their active state and the negligible single channel activity at negative membrane potential (Fig. 5A) is due to OCB that can be removed by depolarization. In contrast, in the native photoreceptor cells the channels are closed in the dark and depolarization cannot open the channels (Fig. 1B, red). PLC activation by light or application of PUFAs is required to activate the native channels. However, divalent OCB is an efficient and fast process, which is expected to quickly block the activated channels and prevent current flow through the channel pore. Therefore, we suggest that activation of PLC (or application of PUFA) has a dual role: to activate the channel by a still unknown mechanism (Leung et al., 2008) and to remove OCB. The facts that expressed TRPL channels in the presence of divalent cations are not active at negative membrane potentials and that they can be activated by depolarization or membrane lipid modulations support the notion that activation and removal of OCB are two different processes. The present study highlights the need to remove OCB in order to allow current flow through the active channel. We show that PLC activation, which converts PIP₂ (with a large hydrophilic head-group) into DAG (with a small hydrophilic head-group, (Janmey and Kinnunen, 2006)) removes OCB and allows light induced current flow through the channels in the presence of ∼1.5 mM external Ca²⁺ and ∼4 mM external and internal Mg²⁺. Figures 3D-E and S3 show that at the onset of the LIC there is a strong OCB, which is reduced with the rise of the inward current at -80 mV. These data demonstrate alleviation of OCB by light under physiological conditions.

Calcium ions are known to have diverse regulatory roles. Since TRP channels are non selective cation channels with a relatively large pore (Owsianik et al., 2005), it is important to regulate Ca²⁺ entry trough these channels under physiological conditions. The mechanism of OCB by slow passage is a useful mechanism to allow entry of these regulatory divalent cations, but in a controlled manner. The putative pore regions of the TRP and TRPL channel are composed of the following amino acids: ES(T)SQSLFWASFGM(L)VG(D)LDD(VS)FE(D)LS(A)GIK (where the putative selectivity filter is underlined and the amino acids of TRP are indicated in brackets where they differ from TRPL, (Owsianik et al., 2005)). Since OCB in the TRP/TRPL channels operates by slow passage of the blocking cation, it is reasonable to assume that the negatively charged amino acids (bold characters) in the pore region are involved in the OCB.

The kinetics of OCB removal by LA and by exogenous PLC activation in the heterologous expression system is rather slow (time scale of seconds, Fig.1). This slow kinetics is partially explained by the exogenous application of the different activators (LA and CCH). In contrast, the removal of OCB in the native photoreceptor cell is fast (time scale of ms, Fig. 3 D-F). This fast removal of OCB can be readily explained by the highly organized signaling compartment and the vast amounts of signaling molecules (e.g. PLC), which produce large quantities of lipids in a restricted area and in the ms range (Hardie and Raghu, 2001).

The current study also has important implications on the regulation of NMDA channels under physiological conditions. Previous studies have shown that the activity of NMDA channels was potentiated by PUFAs (Kloda et al., 2007;Miller et al., 1992) and by membrane stretching (Paoletti and Ascher, 1994) through alterations in the receptor’s lipid environment. However, these studies either showed a large NMDA dependent current at -70 mV membrane potential (Miller et al., 1992), which is not typical for the native channel, or did not present the I-V curve (Paoletti and Ascher, 1994). These observations led to the conclusion that arachidonic acid or stretch only potentiated the persistent NMDA channel current by a still unknown mechanism. In contrast, we show that LA induced a significant increase in the NMDA channels-mediated EPSCs at negative membrane potentials, despite the presence of the blocking Mg²⁺ ions and that it lacked any effect on the current at positive membrane potentials. Thus, these results indicate that LA action on NMDA channels is by the removal of divalent OCB without depolarization.

The well known function of the NMDA channel is to serve as the coincident detector of presynaptic and postsynaptic activity. This function is achieved through OCB removal by depolarization (Kandel, 2000). Accordingly, lipid-producing pathways that reduce OCB should modulate coincidence detection. Indeed, inhibition of phospholipase A2, which produces arachidonic acid has been shown to inhibit induction of long term potentiation (LTP) without specifying a mechanism, in CA1 hippocampal neurons (Massicotte et al., 1990). Together, the effects of lipids on OCB alleviation of different ion channels allow cross talk between channel activity of prime biological importance and lipid-producing pathways.

Supplementary Material

Supp1

Figure S1 Control experiments for Fig. 1A

A. S2 cells that did not express the TRPL channels revealed no effect of LA in whole cell recordings. The red and dark red curves were measured at membrane potentials of 90mV and -90mV respectively and showed no effect of 40μM LA. The gray and black curves show the currents measured at membrane potentials of 90mV and -90mV respectively, which were recorded from S2 cells expressing the TRPL channel for a comparison. Error bars are SEM (n=15).

B. No significant current were developed in S2 cells expressing the TRPL channel if LA was not applied. The blue and dark blue curves were measured at membrane potentials of 90mV and -90mV, respectively (n=6).

Supp2

Figure S2 Linoleic acid and illumination reduced the outward rectification of the native TRP channels in the presence of divalent cations

A. Left, I-V curves measured by whole cell patch clamp recordings from mutant Drosophila ommatidia that express only TRP channels (trpl³⁰²). In darkness the TRP channels were closed (red curve 1). After application of 25μM LA, a gradual increase of the inward current was obtained (black curve 3). The effect of LA at positive membrane potentials preceded the effect at negative membrane potentials (blue curve 2). Application of 10mM Ca²⁺ blocked the TRP channel and temporarily restored the outward rectifying I-V curve (green curve 4, left). Right, the effect of LA on the current is presented as a function of time at 90mV (black dots) and -90mV (dark red dots). The numbers correspond to the curves presented in Fig. S2A left. Together, these results indicated a reduction of the OCB from TRP channels by LA (n=4). The I-V curve of the LIC is presented in pink (the maximal light intensity was attenuated by 2 log units). The inset shows the LIC of the trpl³⁰² mutant in response to the above light intensity. The dual rectifying I-V curve in the presence of LA is due an intrinsic property of the TRP channel (Hardie and Minke, 1994) and OCB removal is manifested by the appearance of a large inward current.

B, C. Summery of whole cell recordings from isolated ommatidia of the trpl³⁰² null mutant that expresses only the TRP channels. The membrane potential of the photoreceptor cell was held at -80 mV and a constant light pulse was applied, which elicited an inward current through the TRP channels. Then, at various time points during the rise time of the LIC, the membrane voltage was stepped to 80 mV and two current values at -80mV and 80mV were measured.

B. A summery of the ratio of currents at 80mV and -80mV was calculated and presented as a function of time from the beginning of the LIC (from the time that the LIC reached 50pA). It is seen that as the LIC progressed the ratio became closer to 1, indicating a reduction in the OCB.

C. A histogram of the results presented in B. The first bin (0-50ms) significantly differs from the rest of the bins (p-value<0.001, n≥8).

Supp3

Figure S3 LA did not change the permeability of TRPL channels to large monovalent organic cations

Whole cell recordings from S2 cells expressing the TRPL channel under bi-ionic conditions with either Triethylamine (A) or Tetraethylamine (B), as the sole extracellular cation, while Cs⁺ was the intracellular cation. The presented I-V curves were measured before and after application of 40μM LA. No change in E_rev was observed. Insets, The I-V curves of A, B are presented in a magnified current scale (n≥4).

Supp4

Figure S4 LA reduces the dwell time of divalent cations in the channels pore of TRPL channels

A. A kinetic model of the TRPL channel. C, O and B are closed, open and blocked states respectively. O_L and O_S are long and short open states. The values of the rate constants are presented in (Parnas et al., 2007).

B. The histogram presents the mean open time of single channel activity, which was measured from inside out patch clamp recordings at -100mV membrane potential, before and after application of LA (the mean open time did not change also at 100mV, data not shown). τ_short and τ_long which determine the mean open time correspond to the closing time constants of the short and long open states respectively in the model presented in A (n=5, error bars are SEM).

C. The burst duration and number of events in the burst were measured from single channel activity of TRPL. Both parameters increased after application of LA thus indicating facilitated removal of the OCB by LA at the single channel level (n=6).

Supp5

Figure S5 Statistical analyses of the experiments

A. Hypoosmotic solutions mimicked the effect of LA. S2 cells expressing the TRPL channel (black dots) were bathed in an extracellular solution of 330 mOsm. Perfusion of 290 mOsm and 255 mOsm solutions (as indicated) enhanced TRPL dependent whole cell currents measured at -90mV membrane potential (n=4). S2 cells that did not express the TRPL channels (red dots) showed no response to hypoosmotic solutions (n=4). Error bars are SEM.

B. The effect of 40μM LA was blocked by application of 10μM GsMTx-4, a specific mechanosensitive channel inhibitor. Washing the water soluble GsMTx-4 rapidly enabled the LA effect. Application of 2mM La³⁺ blocked the current. Error bars are SEM (n=4).

C. LA affects the NMDA channel in a similar manner to the effect on the TRPL channel. Whole cell recordings from S2 cells expressing the two subunits of the NMDA channel. The channel was activated by combined application of NMDA (30μM) and glycine (30μM). The current values at 90mV (black dots) and - 90mV (dark red dots) are presented as a function of time. Application of 40μM LA increased the inward current relative to the outward current. In addition, the effect at positive membrane potentials preceded the effect at negative membrane potentials. Application of 50mM Mg²⁺ blocked the currents. Error bars are SEM (n=4).