N-Arachidonoyl l-Serine, a Putative Endocannabinoid, Alters the Activation of N-Type Ca²⁺ Channels in Sympathetic Neurons

Abstract

The effect of N-arachidonoyl l-serine (ARA-S), a recently discovered lipoamino acid found in the CNS, on N-type Ca²⁺ channels of rat sympathetic ganglion neurons was determined using whole cell patch clamp. Application of ARA-S produced a rapid and reversible augmentation of Ca²⁺ current that was voltage dependent and resulted from a hyperpolarizing shift in the activation curve. ARA-S did not influence G protein modulation of Ca²⁺ channels and appeared to act independently of G-protein-coupled receptors. These findings provide a foundation for investigating possible roles for ARA-S in nervous system function.

INTRODUCTION

Endocannabinoids comprise a group of endogenously produced lipid-derived molecules that bind to and affect CB₁ and CB₂ G-protein-coupled cannabinoid receptors. The best-characterized endocannabinoids, 2-arachidonoyl glycerol (2-AG) and N-arachidonoyl ethanolamine (anandamide), act as agonists at both CB₁ and CB₂ receptors and are implicated in a variety of physiological responses (Pacher et al. 2006). In addition to receptor-mediated effects, endocannabinoids interact directly with voltage- and ligand-gated ion channels producing alterations in gating usually manifested as channel inhibition (Oz 2006). Recently, a novel endocannabinoid-like compound, N-arachidonoyl l-serine (ARA-S), was identified in extracts of bovine brain and shown to produce endothelium-dependent vasodilation of rat mesenteric arteries (Millman et al. 2006). Synthetic ARA-S has minimal activity at traditional endocannabinoid targets (CB₁R, CB₂R, and TRPV1 channels) but appears to act via a novel, but as yet unidentified, G-protein-coupled non-CB₁/CB₂ cannabinoid receptor. ARA-S is also reported to be an antagonist at a non-CB₁/CB₂ cannabinoid receptor found in neutrophils (McHugh et al. 2007). While screening GPR35 (Guo et al. 2008), an orphan G-protein-coupled receptor, as a potential target of ARA-S, we found that ARA-S produced receptor-independent effects on the activation of N-type Ca²⁺ channels (Ca_V2.2) in rat sympathetic neurons leading to an enhancement of current especially at hyperpolarized potentials.

METHODS

Cell isolation and electrophysiology

Superior cervical ganglion (SCG) neurons from male Wistar rats (150–300 g) were enzymatically dissociated and placed in short-term (<24 h) culture as described previously (Ikeda 2004). For some experiments, the neurons were incubated overnight in tissue culture media containing 500 ng/ml Bordetella pertussis toxin (PTX; List Biological Laboratories, Campbell, CA) as previously described (Guo and Ikeda 2004). Rats were killed by decapitation after anesthesia with CO₂ as approved by the Institutional Animal Care and Use Committee. Neurons were voltage-clamped using the whole cell patch-clamp technique as described previously (Guo and Ikeda 2004). I_Ca tail currents were filtered at 10 kHz prior to digitization at 50 kHz. Series resistance was electronically compensated ≥80%. Experiments were carried out at room temperature (22–26°C).

Solutions and chemicals

The external recording solution contained (in mM) 140 methanesulphonic acid, 145 tetraethylammonium hydroxide, 10 HEPES, 10 glucose, and 10 CaCl₂ and 0.0003 tetrodotoxin (Alomone Labs, Jerusalem, Israel), pH 7.4 with TEA-OH. For tail current experiments, the CaCl₂ was reduced to 5 mM. The pipette solution contained (in mM) 120 N-methyl-d-glucamine, 20 tetraethylammonium hydroxide, 11 EGTA, 10 HEPES, 10 sucrose, 10 HCl, 1 CaCl₂, 4 MgATP, 0.3 Na₂GTP, and 14 Tris creatine phosphate, pH 7.2 with methanesulphonic acid. The osmolalities of the bath and pipette solutions were adjusted with sucrose to 325 and 300 mOsmol/kg, respectively. N-arachidonoyl-l-alanine (ARA-A), N-arachidonoyl-l-glycine (ARA-G), N-arachidonoyl-l-Serine, and N-arachidonoyl-dopamine (ARA-DA) were purchased from Cayman Chemical (Ann Arbor, MI) as ethanol stock solutions (25–140 mM) and dissolved directly into the recording solution on the day of the experiment. Ethanol was <=0.1% in all solutions and at this concentration produced no discernable effect on I_Ca. Drugs were applied by positioning the outlet tube (200 μm ID) of a custom-designed gravity-fed microperfusion system ∼100 μm from the cell body.

Data analysis and statistics

Nonlinear least-squares curve fitting was performed using a Marquardt-Levenberg algorithm from Igor Pro version 6.02A (WaveMetrics, Lake Oswego, OR). Statistical comparisons, as indicated in the text, were determined with Prism 4 version 4.0c (GraphPad Software, San Diego, CA). P < 0.05 was considered significant. Summary data are presented as means ± SE.

RESULTS

We tested the effect of ARA-A, ARA-G (Huang et al. 2001), ARA-S, and ARA-DA (Huang et al. 2002) application on Ca²⁺ currents (I_Ca) recorded from dissociated adult rat sympathetic neurons using whole cell patch clamp. Under these recording conditions, I_Ca arises primarily from ω-conotoxin GVIA-sensitive N-type Ca²⁺ channels (Ikeda 1991). Currents were evoked with a 25-ms test pulse to −10 mV from a holding potential of −80 mV. Application of ARA-A (10 μM) or ARA-S (10 μM) produced a rapid and reversible enhancement of I_Ca amplitude (Fig. 1 A) without overt modification of I_Ca kinetics (Fig. 1B). Conversely, application of ARA-DA (10 μM) to the same neuron produced minimal effects. A second series of sequential drug applications produced similar effects indicating that the minuscule ARA-DA response did not result from tachyphylaxis. ARA-G (10 μM) was tested in a separate set of experiments and produced results similar to those observed following ARA-S application (Fig. 1C). Comparison of the average change in I_Ca amplitude (Fig. 1C) revealed that ARA-S produced the largest increase (132 ± 14%, n = 14) and thus further experiments focused on this compound.

FIG. 1.

Voltage-dependent enhancement of Ca²⁺ currents (I_Ca) by N-arachidonoyl l-serine and related compounds. Data are from whole cell patch-clamp recordings of rat sympathetic neurons obtained at room temperature (22–26°C). A: time course of I_Ca amplitude during exposure to 10 μM of N-arachidonoyl l-alanine (ARA-A, □), N-arachidonoyl l-serine (ARA-S, ▪), or N-arachidonoyl dopamine (ARA-DA, ). B: I_Ca traces for the time course shown in A were evoked with 25-ms test pulses to +10 mV from a holding potential of −80 mV. Superimposed traces in the absence or presence of drug as indicated. - - -, the 0 current level. C: mean ± SE increase in I_Ca amplitude produced by application of ARA-A, ARA-S, N-arachidonoyl l-glycine (ARA-G), and ARA-DA. With the exception of ARA-S vs. ARA-G, all means differ significantly (P < 0.001) from each other as determined from 1-way ANOVA followed by Newman-Keuls post hoc test. D: effect of ARA-S (10 μM) on I_Ca at different test potentials. I_Ca was evoked from a holding potential of −80 mV with 70-ms pulses to the indicated potential in the absence or presence (•) of ARA-S. E: current-voltage (I-V) curves were obtained in the absence (○) or presence of ARA-S (10 μM, •) and then normalized to the I_Ca amplitude at +10 mV in the absence of ARA-S. The data represent the means ± SE for 10 neurons. F: concentration-response curve for ARA-S. Three concentrations of ARA-S were applied to each neuron (0.1, 1.0, and 10 or 0.3, 3.0, and 30 μM ARA-S). I_Ca was allowed to recover between applications.

The effect of ARA-S at different potentials was examined by evoking I_Ca with 70-ms test pulses over the range −60 to +80 mV from a holding potential of −80 mV in the presence or absence of 10 μM ARA-S (Fig. 1D). Under control conditions, I_Ca became apparent around −30 mV, reached a maximum amplitude near +10 mV, and thereafter declined asymptotically toward baseline. In the presence of ARA-S, I_Ca was augmented several-fold at negative test potentials with lesser increases at more depolarized potentials. At very depolarized test potentials (> +40 mV), I_Ca amplitude was minimally affected, suggesting that a shift in Ca²⁺ channel activation (Fig. 1E) rather than an increase in the number of available channels was responsible for the action of ARA-S. A concentration-response curve for ARA-S is illustrated in Fig. 1F. Three concentrations of ARA-S (0.1, 1.0, and 10 or 0.3, 3.0, and 30 μM) were applied to neurons allowing time to recover between each concentration. Potentiation of I_Ca amplitude was evident starting at 3 μM and increased monotonically without evidence of saturation ≤30 μM, the highest concentration examined.

To investigate shifts in channel activation, Ca²⁺ channel tail current amplitudes were determined in the absence or presence of ARA-S (10 μM). Tail currents result from the deactivation of channels on return to a hyperpolarized potential following a sojourn at a step potential that produces channel activation (Fig. 2 A). Because tail currents are measured at a constant potential (here −40 mV), driving force, which is nonlinear for large ionic gradients such as Ca²⁺, remains constant and thus I_Ca amplitude can be equated with conductance. To facilitate accurate tail current measurement, analog filter bandwidth (−3 dB) and digital sampling rate were increased to 10 and 50 kHz, respectively. External [Ca²⁺] was decreased to 5 mM, which served to decrease tail current amplitude and hence the effects of residual uncompensated series resistance. Under these conditions, tail current decays were well fit (following a 100-μs delay to allow uncompensated capacitive transients to settle) by a single-exponential function (Fig. 2B, —) with τ of ∼0.5 ms. Activation curves were plotted as tail current amplitude, normalized to maximum amplitude (e.g., +80-mV step potential) in the absence of drug, versus step potential (Fig. 2C) and fit (—) with a two-component modified Boltzmann equation

where I_tail is the normalized tail current amplitude and V_step is the step potential. I_n, V_hn, and k_n are the fractional amplitude, half activation potential, and slope factor, receptively for each component. Under control conditions, Ca²⁺ channel activation curves from adult rat SCG neurons are composed of two components—a result of tonic G protein modulation (Ikeda 1991). In the presence of ARA-S, the activation was shifted toward more hyperpolarized potentials, retained a two-component profile, and achieved a similar maximum amplitude. Analysis of the activation parameters (Fig. 2E) revealed significant decreases in V_h1, V_h2, and k₂ but no change in the fractional contribution of each component. The hyperpolarizing shift in the activation curve account for the voltage-dependent I_Ca increases as shown in Fig. 2D. Both the tail and step I_Ca were affected to a much greater extent by ARA-S at hyperpolarized voltages. The coincidence of the step and tail I_Ca data in Fig. 2D provide evidence for the fidelity of the tail current recordings.

FIG. 2.

ARA-S shifts I_Ca activation curves toward hyperpolarized potentials. A: tail current voltage protocol (bottom) and representative traces, recorded at a step potential of −10 mV, obtained in the absence or presence of ARA-S (10 μM). External [Ca²⁺] was decreased to 5 mM to reduce current amplitude thereby reducing series resistance error. B: tail current decay illustrated on an expanded time scale (same traces as in A). The initial 100 μs following the step pulse have been blanked. —, nonlinear least-square fits to single-exponential function. C: activation curves composed of mean ± SE (n = 5) normalized tail current amplitudes vs. step potential. Tail current amplitudes were determined from the exponential fit of the tail current decay phase. Amplitudes obtained in the absence (○) or presence (•) of ARA-S were normalized to the maximum amplitude in the absence of drug. —, the best fit of a 2-component Boltzmann equation (see text) to the data using nonlinear least-squares regression. D: voltage dependence of ARA-S effects illustrated by plotting the ratio of I_Ca amplitude in the presence of ARA-S to the control condition for step (□) and tail currents (▪) vs. step potential. E: table of mean ± SE Boltzmann equation parameters. A paired t-test was used for statistical comparison between the control and drug condition. ***, P < 0.001; *, P < 0.05.

Next, we asked whether ARA-S affects I_Ca modulation by G-protein-coupled receptors. In SCG neurons, norepinephrine (NE) acts by binding to α₂-adrenergic receptors (Schofield 1990) thereby liberating Gβγ from the heterotrimeric G protein complex resulting in voltage-dependent modulation (Bean 1989; Ikeda 1996). This form of I_Ca modulation is detected using a double-pulse voltage protocol (Elmslie et al. 1990) (Fig. 3 B) in which two identical test pulses to +10 mV are separated by a depolarizing conditioning pulse to +80 mV. Facilitation, defined as the ratio of postpulse to prepulse I_Ca amplitude, provides a convenient index of Gβγ-mediated modulation (Fig. 3A). Increases in facilitation result from relief of Gβγ inhibition produced by the conditioning pulse. Application of NE produced I_Ca inhibition with a coincidental increase in facilitation that was similar in the presence of ARA-S. The mean inhibition of I_Ca, basal (i.e., in the absence of agonist) facilitation, and NE-induced facilitation were significantly different in the presence of ARA-S (Fig. 3C). However, the magnitude of the changes were small and possibly arose from effects of ARA-S on minor components of I_Ca that do not arise from N-type Ca²⁺ channels. Thus ARA-S does not appear to greatly influence either receptor-mediated or tonic modulation of N-type Ca²⁺ channels by Gβγ. The effects of ARA-S (10 μM) on neurons treated overnight with 500 ng/ml PTX were also examined. I_Ca potentiation following ARA-S application was not significantly altered by pretreatment with PTX (210 ± 4, n = 3 vs. 290 ± 31%, n = 9 for control and PTX-treated, respectively). Conversely, NE-mediated I_Ca inhibition was decreased (13 ± 6, n = 3 vs. 52 ± 3%, n = 9, for PTX and control, respectively) following PTX treatment providing evidence for toxin effectiveness. Thus activation of PTX-sensitive G proteins, namely G_i/o-containing heterotrimers, were not essential for I_Ca increases resulting from ARA-S application.

FIG. 3.

G-protein-mediated I_Ca modulation in the presence of ARA-S. A: time course of I_Ca amplitude during superfusion with norepinephrine (NE, 10 μM; □) and ARA-S (10 μM; ▪). I_Ca was evoked at 0.1 Hz with the voltage protocol shown in B. ○ and • (top), prepulse and postpulse I_Ca amplitudes, respectively. Facilitation (bottom, ▪) is the ratio of postpulse to prepulse current amplitudes. B: representative I_Ca traces from the time course shown in A. Numbers next to the traces refer to corresponding time points labeled in A. C, left: mean ± SE facilitation determined in the absence (□) or presence (▪) of ARA-S. Basal facilitation was determined prior to NE administration. Right: mean ± SE I_Ca inhibition produced by application of NE. Inhibition was determined from the prepulse current. Number of neurons used for all analyses indicated in parentheses. Means were compared with a paired Student's t-test. *, P < 0.05; **, P < 0.01.

DISCUSSION

Our data show that ARA-S produces a rapid and reversible augmentation of I_Ca in sympathetic neurons that is voltage dependent and results from a hyperpolarizing shift in the activation curve. At depolarized potentials, the same maximal conductance is attained in the presence of ARA-S arguing against a recruitment of covert channels or a change in the maximum probability of opening. ARA-S did not interfere with or contribute to G protein modulation of N-type Ca²⁺ channels. We thus conclude that I_Ca enhancement by ARA-S occurs independently of G-protein-coupled receptors and possibly results from direct interaction with the Ca_V2.2 channel or alteration of plasma membrane surface potential.

There are two main implications of these findings. First, the ARA-S concentration (10 μM) used here falls well within the concentration range used to probe potential physiological roles of ARA-S and related endocannibinoids (Milman et al. 2006). Given the well-established role played by N-type Ca²⁺ channels in providing Ca²⁺ for synaptic transmission combined with the nonlinear relationship between [Ca²⁺] and transmitter release (Xu et al. 2007), one can easily envision changes in synaptic transmission produced by applying ARA-S. Thus receptor-independent effects, such as the one demonstrated here, will need to be considered before meaningful interpretations of ARA-S action are entertained. Although agents such as Bordetella pertussis toxin are useful for separating receptor-dependent (at least in terms of GPCRs using G_i/oα-containing heterotrimers) from receptor-independent effects, this ability degrades as the complexity of the system increases (e.g., in vivo experiments). Second, it is possible that direct actions of ARA-S on ion channels underlie physiological processes. For example, a related compound, anandamide, is proposed to influence physiological function by binding to CB₁R/CB₂R and subsequently activating downstream signaling cascades or directly activating TRPV channels without the aid of signaling intermediates (Smart et al. 2000; van der Stelt and Di Marzo 2005; van der Stelt et al. 2005). Although endocannabinoids and related lipid compounds (Bradshaw and Walker 2005) often have direct effects on ion channel function (Oz 2006), the effect of ARA-S on N-type Ca²⁺ channels is somewhat unique. Closely related compounds such as anandamide, 2-AG (Guo and Ikeda 2004) and ARA-DA (Fig. 1C) either have no effect or produced inhibition at similar concentrations, whereas both ARA-A and ARA-G were capable of augmenting I_Ca to varying degrees (Fig. 1C). From this series of compounds, the presence of a carboxylic acid group was common to the substances that enhanced I_Ca amplitude. We could not find literature values for the pK_a of ARA-S or related lipoamino acids thus the charge status of the carboxylic acid group at physiological pH is unclear. Given the pK_a of the carboxylic acid group in the free amino acids (2.1–2.4), it seems likely that ARA-S, ARA-G, and ARA-A are negatively charged at pH 7.4. Thus a possible explanation for the effects of ARA-S and related lipoamino acids is alteration of the membrane surface potential (Hille 2001) following incorporation of negative charges into the outer leaflet of the plasma membrane. The net result would be a negative shift in channel activation as seen following ARA-S application. It should be noted, however, that enhancement of L-type I_Ca in ventricular myocytes by long chain fatty acids occurred without shifting activation or inactivation along the voltage axis (Huang et al. 1992) thus arguing against this explanation as a universal mechanism for Ca²⁺ channel modulation by negatively charged lipophilic agents.

The actions of ARA-S are somewhat reminiscent of the effects of arachidonic acid on N-type Ca²⁺ channels in SCG neurons (Liu and Rittenhouse 2003; Liu et al. 2001). Although ARA-S, a conjugate of arachidonic acid and serine, likely breaks down to arachidonic acid, ARA-S lacks a Ca²⁺ channel inhibitory component characteristic of arachidonic acid effects. Therefore, it is unlikely that arachidonic acid mediates the effects of ARA-S although we cannot conclusively rule out this possibility. The stimulatory effects of ARA-S on N-type Ca²⁺ channels are strikingly similar to those observed for the fatty acid analogs palmitoyl coenzyme A and arachidonoyl coenzyme A (Barrett et al. 2001). Thus the lipid moiety in these compounds may share a common mechanism for Ca²⁺ channel stimulation that is conferred by the charge on the head group. At present, many details relevant to the biology of ARA-S, including synthetic and degradative pathways, partition coefficient, and sites of action have not been investigated. This void in our knowledge limits speculation as even fundamental facts, such as the relevant physiological concentration of ARA-S, remain unknown.

GRANTS

This work was supported by the intramural program of the National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism, Bethesda, MD.