Inhibition of TRP3 channels by lanthanides: block from the cytosolic side of the plasma membrane

The lanthanide ions La(3+) and Gd(3+) block Ca(2+)-permeable cation channels and have been used as important tools to characterize channels of the transient receptor potential (TRP) family. However, widely different concentrations of La(3+) and Gd(3+) have reportedly been required for block of TRP3 channels in various expression systems. The present study provides a possible explanation for this discrepancy. After overexpression of TRP3 in Chinese hamster ovary cells, whole-cell currents through TRP3 were reversibly inhibited by La(3+) with an EC(50) of 4 microm. For comparison, the organic blocker SKF96365 required an EC(50) of 8 microm. Gd(3+) blocked with an EC(50) of 0.1 microm, but this block was slow in onset and was not reversible after wash-out. When the two lanthanides were added to the cytosolic side of inside-out patches, block was achieved with considerably lower concentrations (EC(50) for La(3+), 0.02 microm; EC(50) for Gd(3+), 0.02 microm). Uptake of La(3+) into the cytosol of Chinese hamster ovary cells was demonstrated with intracellular fura-2. We conclude that lanthanides block TRP3 more potently from the cytosolic than from the extracellular side of the plasma membrane and that uptake of lanthanides will largely affect the apparent EC(50) values after extracellular application.


Summary
The lanthanide ions La 3+ and Gd 3+ block Ca 2+ -permeable cation channels and have been used as important tools to characterize channels of the TRP family. However, widely different concentrations of La 3+ and Gd 3+ have reportedly been required for block of TRP3 channels in various expression systems. The present study provides a possible explanation for this discrepancy. After overexpression of TRP3 in CHO cells, whole-cell currents through TRP3 were reversibly inhibited by La 3+ with an EC 50 of 4 µM. For comparison, the organic blocker SKF96365 required an EC 50 of 8 µM. Gd 3+ blocked with an EC 50 of 0.1 µM but this block was slow in onset and was not reversible after wash-out. When the two lanthanides were added to the cytosolic side of inside-out patches, block was achieved with considerably lower concentrations

Introduction
Elucidation of the mechanisms responsible for receptor-mediated Ca 2+ influx in electrically non-excitable cells remains a continuous challenge. Although it is generally accepted that in many cells, the predominant signal for Ca 2+ influx is the depletion of intracellular calcium stores (1), the underlying mechanisms are not known in detail. Furthermore, the molecular structure of the Ca 2+ channels permitting Ca 2+ entry across the plasma membrane has not been clarified. Proteins of the recently discovered TRP 1 family may be an essential part of these channels because antisense constructs of several TRP cDNAs inhibit store-operated Ca 2+ influx (2)(3)(4)(5)(6). Heterologous expression of several members of the TRP family leads to the appearance of Ca 2+ -permeable cation channels (7) but these exhibit properties not congruent with those of channels mediating store-operated Ca 2+ influx, particularly the Ca 2+ -selective I CRAC channels (8;9).
Moreover, overexpression of corresponding TRP orthologues in different cell types by different research groups resulted in ion currents that obeyed different regulatory principles. For example, TRP3 was initially characterized as a constitutively active, store-independent, Ca 2+ -regulated channel (10) but was also described as being regulated by diacylglycerol (DAG) (11). Other studies suggest a store-dependent mechanism for the activation of TRP3 (2;12;13). Additionally, an important regulation of TRP3 occurs through its interaction with the InsP 3 receptor (14;15). This interaction occurs at a defined region of TRP3 localized within the C-terminal tail that extends into the cytosol (16). A recent report (17) on this interaction indicates that overexpressed TRP3 channels are not activated by store-emptying alone.
In light of these discrepant results on the regulation of TRP3, a detailed functional and biophysical characterization would be desirable for any study on heterologously expressed TRP family members, to clarify whether identical channels are expressed in the respective expression systems. In general, important tools for the characterization of ion channels are channel blockers.
Unfortunately, there are no specific inhibitors known for any particular member of the TRP family. A widely used compound is SKF96365, but it is not specific because it inhibits Ca 2+ entry channels at similar concentrations and effectivities as it inhibits other channels such as Cland cation channels (18). Whether SKF96365 can be used to discriminate between various Ca 2+ channels or between channels appearing after overexpression of members of the TRP family has not been determined in detail.
In the absence of specific organic blockers, many researchers have resorted to the ions of the lanthanides gadolinium (Gd 3+ ) and lanthanum (La 3+ ). They block a wide range of Ca 2+permeable channels; however, the sensitivity to Gd 3+ and La 3+ may vary between different channels such that the sensitivity to lanthanides has been used as part of the characterization of overexpressed TRP channels. In the case of TRP3, however, discrepancies in this sensitivity have been reported. Zhu et al. required 250 µM La 3+ for an inhibition by 30-40% of Ca 2+ entry through TRP3 channels expressed in COS-M6 cells; a complete block was achieved with 1 mM (2). The same authors induced a complete inhibition with 150 µM La 3+ when TRP3 was expressed in HEK293 cells (19). An EC 50 of 24 µM was estimated for the inhibition of TRP3 by La 3+ in cultured bovine pulmonary endothelial cells (20), in line with a report that 50 µM abolished TRP3 currents in porcine aortic endothelial cells (21). Considerably lower concentrations (10 µM) induced a complete inhibition of TRP3 in COS-1 cells (12). Gd 3+ inhibited TRP3-related Ca 2+ entry completely in HEK293 cells at a concentration of 200 µM, but 10 µM in the bath could be used to discriminate Ca 2+ entry pathways endogenous in these cells from that attributable to TRP3 which was still apparent in the presence of 10 µM Gd 3+ in the bath (19).
When we tested the effects of the two lanthanides on TRP3 expressed in CHO cells, we were surprised to notice that the concentrations required for block were markedly lower than reported anywhere else in the literature. Although the effects of lanthanides are attributed to an action strictly confined to the extracellular side of the plasma membrane (22), we examined block by Gd 3+ and La 3+ on the cytosolic side of TRP3 channels in inside-out patches. Concentrations were effective that were even lower than those required during extracellular application.
Furthermore, entry of La 3+ was directly demonstrated in fura-2 loaded cells. We propose that lanthanides may act on TRP3 from the intracellular side and that the extracellular concentrations at which lanthanides inhibit Ca 2+ entry may be largely influenced by their cell-specific uptake rates. to 300±10 mosm/kg using mannitol; the osmolarity-difference between bath-and pipettesolution was kept below 5 mosm/kg. 1-Oleoyl-2-acetyl-sn-glycerol (OAG) was first dissolved in DMSO and then added to the bath solution. The final concentration of OAG in the bath was 100 µM, the concentration of DMSO was 1%. SKF96365, GdCl 3 , or LaCl 3 were added to the bath solution at concentrations as indicated. The holding potential was set to -60 mV. The membrane current was recorded with a HEKA EPC-9 patch-clamp amplifier using HEKA's PULSE software (HEKA Elektronik, Lambrecht/Pfalz, Germany). In whole-cell mode, currents were filtered at 1 kHz, in inside-out mode at 3 kHz; the sampling rates were set appropriately.

Fura-2 measurements
Measurements were performed using a digital imaging system (T.I.L.L. Photonics, Germany). Fura-2 was excited with light of the wavelength λ ex =360 nm, emission was measured at λ em =510 nm. For in-vitro experiments, fura-2 salt (Calbiochem) was dissolved in nominally Ca-and Mg-free PBS ([fura-2]=20 µM). For in-vivo experiments, CHO cells were loaded with fura-2 by incubation in fura-2/AM (Calbiochem) as described (24); during measurements, the cells were kept in nominally Ca-and Mg-free PBS. Additionally, spectra of fura-2 in the presence of various concentrations of La 3+ were obtained in a RF-5001PC spectrofluorophotometer (Shimadzu).

Results
Expression of TRP3 in CHO cells resulted in markedly enhanced whole-cell cation currents, in comparison to those in control CHO cells. Currents reached a maximum right after obtaining the whole cell configuration and declined steadily over several minutes. Inward currents were mostly carried by Na + but also to a minor part by Ca 2+ and disappeared when extracellular Na + was substituted with the large impermeant cation NMDG (Fig. 1). These results are in agreement with our previous report on functional characterization of TRP3 (10).
[Place Fig. 1 here] To analyze the effects of lanthanides and SKF96365, we first determined the inhibition of currents by NMDG in each experiment. This defined the amplitude of cation currents.
Thereafter, the normal bath was restored. Then, the cells were exposed to various concentrations of the inhibitors. In some experiments, the inhibitors were washed out to test for reversibility of the block. To quantify inhibition in a situation when currents decline spontaneously, we performed a graphical extrapolation of the currents from time periods when no inhibitor was present to the time in the presence of the inhibitors, thereby obtaining current values expected to be present if inhibitors had been absent. Actual currents in the presence of inhibitors were divided by these extrapolated current values, yielding the relative inhibition.
The inhibition of TRP3 currents by La 3+ (Fig. 1A) and SKF96365 (Fig. 1B) was clearly distinct from the spontaneous current decline because it occurred much faster and was dependent on the concentration of the respective substance. An EC 50 value of 4 µM was assessed for La 3+ ( Fig. 2A). This inhibition could be reversed by wash-out of La 3+ (Fig. 1A). SKF96365 did not completely block the currents even at the maximal concentrations applied (30 µM); the EC 50 value was estimated to be 8 µM (Fig. 2B).

here]
Similarly as SKF96365 and La 3+ , Gd 3+ at concentrations between 1 and 30 µM induced a rapid inhibition of TRP3 currents (data not shown). This inhibition was complete since no further current depression was evoked by NMDG. Washout of Gd 3+ did not lead to a recovery of the currents. When lower concentrations (0.1-0.3 µM) of Gd 3+ were used (Fig. 1C), inhibition of the currents occurred slowly although still distinctly faster than the spontaneous decline. The absolute amplitude of the currents in the presence of these low concentrations was above that after substitution of extracellular cations with NMDG, indicating a partial inhibition by Gd 3+ .
The peculiar kinetics of the Gd 3+ effects could be explained if the action of the ion on TRP3 channels occurred from the cytosolic side of the membrane after a slow entry of Gd 3+ into the cell. To test this hypothesis directly, we performed experiments with inside-out patches in which the cytosolic side of the patches was exposed to Gd 3+ concentrations from 30 nM to 1 µM. As reported (10), TRP3 channels in inside-out patches from CHO cells are characterized by a spontaneous channel activity right after obtaining the inside-out configuration and a decline of the activity over 3-6 min. Since it has been reported that a "rigorous" wash abolishes channel activity (14), exchange of the bath was performed by a slow exchange of the bath volume over 15-20 s which did not cause rapid changes in channel activities.
[Place Fig. 3 here] Since channel activity was low in most patches and is difficult to quantify due to the short open times (mean open time ≤ 0.2 ms according to (10)), effects of higher Gd 3+ concentrations (i.e. 0.1 and 1 µM) were analyzed in patches stimulated with the membrane permeable DAG analogue OAG. OAG considerably stimulated TRP3 channels for about 2 min (Fig. 3A), in accordance with the report by Hofmann et al. (11). Addition of Gd 3+ at 1 µM (5 out of 5) and 0.1 µM (10 out of 10) to those patches abolished channel activity within a few seconds (Fig. 3B). Wash-out of Gd 3+ did not restore channel activity even in the continuous presence of OAG (not shown).
[Place Fig. 4 here] For a quantification of the concentration-dependence of Gd 3+ , unstimulated patches were used because the time course of OAG-induced stimulations was too variable to allow calculation of relative inhibitions. Each patch was consecutively exposed to two increasing Afterwards, a Gd 3+ -free bath was reestablished. Gd 3+ induced inhibitions of channel openings as fast as the bath exchange was performed, in a concentration-dependent manner (Fig. 4, 5A).
Again, wash-out of Gd 3+ failed to restore channel activity. The same result was obtained when patches were exposed to only one concentration (0.1 µM) of Gd 3+ for a short (20 s) time (data not shown). To test whether this observation indicates a long-lasting inhibition of TRP3 channels by Gd 3+ or, alternatively, may reflect spontaneous complete decline of channel activity, the same protocol was performed with La 3+ (0.03-1 µM) as inhibitor applied to the cytosolic side of inside-out patches (Fig. 6). There was a concentration-dependent reduction of TRP3 channel activity (Fig. 5B). In contrast to the experiments involving Gd 3+ , removal of La 3+ led to a restoration of channel activity that may be considered complete if the spontaneous inactivation of the channels is taken into consideration (Fig. 6). Estimated EC 50 values are 0.02 µM Gd 3+ and 0.02 µM La 3+ .
[Place Fig. 5 here] [Place Fig. 6 here] To demonstrate that La 3+ actually enters the cytosol of CHO cells, we performed experiments with fura-2. A shift in the excitation spectrum of fura-2 by lanthanum has been descibed (25), resulting in an increase in fluorescence at excitation wavelengths from 300 to 350 nm.
At 360 nm, we found that La 3+ led to a concentration-dependent decrease in the fluorescence of fura-2 in vitro (Fig. 7A). At shorter excitation wavelengths, the previously described increase in fluorescence was reproduced but we were unable to discriminate between the effects of La 3+ and those of contaminating Ca 2+ (Fig. 7B). Therefore, we used an excitation wavelength of 360 nm (the isosbestic wavelength for Ca 2+ in the absence of La 3+ ) to examine the effects of La 3+ on cells loaded with fura-2 after incubation with fura-2/AM. Addition of La 3+ to the bath evoked a rapid decrease in the fluorescence of the cells. This effect was reversible after wash of the cells (Fig. 7C). Thus, CHO cells are capable to take up and to extrude La 3+ at rates compatible with the observed block of TRP3 currents in whole-cell experiments. We did not detect any sizeable decrease of fura-2 fluorescence after addition of Gd 3+ in vitro, therefore a similar demonstration of Gd 3+ uptake into CHO cells was not possible.

Discussion
The important findings of this study are that the cations Gd 3+ and La 3+ inhibit TRP3 currents when applied from the extracellular as well as from the intracellular side of the cell membrane. Concentrations of either ion that evoked half-maximal inhibitions from outside were consistently found to inhibit TRP3 channels completely when added to the cytosolic side of inside-out patches. Intracellular block by lanthanides may be of particular relevance in CHO cells because entry of La 3+ was directly demonstrated. Therefore, the experimentally determined potency of extracellular Gd 3+ and La 3+ on TRP3 channels will be strongly influenced by the rate with which extracellular and intracellular concentrations reach an equilibrium. This may be of general importance in attempts to characterize TRP3 and possibly other TRP channels after heterologous expression. Discrepancies of results between various expression systems may simply reflect differences in the uptake and extrusion rate of Gd 3+ and La 3+ , rather than differences in the expressed gene products.
When comparing the extracellular effects of Gd 3+ with that of La 3+ , Gd 3+ -induced current inhibitions were much slower in onset. Gd 3+ block was slow even in comparison with SKF96365, which otherwise proved to be an unsuitable pharmacological tool because of its high EC 50 that precluded preparation of solutions which would evoke a complete inhibition of TRP3.
In contrast to the results with extracellular Gd 3+ , Gd 3+ applied to the cytosolic side of insideout patches evoked its effects without any delay other than that attributable to the time required for the bath exchange. The block by Gd 3+ was irreversible over the observation times of this study. This may indicate a long-lasting binding of Gd 3+ to the TRP3 protein at low concentrations.
Our results suggest that whenever entry of Gd 3+ and La 3+ into cells occurs, it will certainly enhance blocking effects on TRP3. It may be asked whether inhibition of TRP3 currents by lanthanides is completely due to intracellular rather than extracellular action.
However, it cannot be deduced from our experiments whether entry is essentially required and to what extent block from the outside takes place. An experimental protocol to exclude extracellular effects would require outside-out patches. Unfortunately, we were unable to detect TRP3 channel activity in those patches. The reason for that failure may be the noise level, that was considerably higher than in inside-out patches. Under these circumstances, analysis of channel openings of TRP3 is impossible since the mean open time is extremely short. Even if experiments with outside-out patches were successful, they might not definitely rule out a strictly intracellular action of Gd 3+ , because transmembrane pathways for Gd 3+ may be present in isolated patches as well as in whole cell preparations. Taken together, we cannot exclude extracellular effects of Gd 3+ , but if they are present, they are certainly weaker than the intracellular ones.
In the case of La 3+ , the differences of equipotent extracellular and intracellular concentrations were more pronounced than in experiments with Gd 3+ . Thus, even a relatively small rate of La 3+ uptake may account for La 3+ block. At the same time, reversibility of the extracellular effects would be achieved by a modest extrusion rate.
Interestingly, lanthanides are likely to be chelated by Ca 2+ chelators such as EGTA and fura-2 (22), in spite of the fact that they differ from Ca 2+ in their valency. In experiments where La 3+ inhibited Ca 2+ entry at extracellular concentrations of 0.5 µM, addition of EGTA (10 µM) to the bath completely abolished this effect (26). A reported (27)  Effects of La 3+ and of Gd 3+ are frequently considered to be strictly confined to the extracellular space because cell membranes are believed to be impermeable for these ions (22).
However, there are exceptions reported for red blood cells (33) as well as myocardial cells (34). In our experiments, the fluorescence of intracellular fura-2 in CHO cells was decreased after addition of La 3+ to the bath, as was the fluorescence of fura-2 in vitro by La 3+ . As excitation wavelength in the experiments with cells, we chose the isosbestic wavelength of fura-2 for Ca 2+ in the absence of La 3+ . Therefore, any decrease in the intracellular Ca 2+ concentration that a block of Ca 2+ influx might induce would minimally affect the fluorescence of fura-2. Furthermore, the experiments were performed in nominally Ca-free bath (estimated Ca 2+ concentration 50 µM) in which block of Ca 2+ channels is expected to have no instant effects on intracellular Ca 2+ concentrations. Therefore, the experiments demonstrate that La 3+ rapidly enters CHO cells and is also quickly extruded from the cytosol after wash. Since these findings were obtained in wildtype CHO cells, transmembrane movements of La 3+ are not dependent on expression of TRP3.
The changes of fluorescence occurred as rapidly as block of TRP3 currents by La 3+ . Whereas these experiments provide information about the time over which cytosolic concentrations of La 3+ change, they do not allow a quantification of these concentrations. The cytosolic concentration of fura-2 would be important for the calculation of La 3+ -concentrations but is not known under our experimental conditions. Nevertheless, it is safe to conclude that La 3+ enters CHO cells in concentrations relevant for block of TRP3 channels.
One of the disturbing points of the literature on members of the TRP family are the apparent inconsistencies of results reported by different research groups. For studies that performed an overexpression of TRP proteins, this raises the question whether cation channels formed by the overexpressed gene products are identical. However, when channels are characterized by means of block with lanthanides, our present study emphasizes that conflicting results may be explained not by differences in the channel structure but by different transmembrane transport rates of lanthanides. We propose that any quantification of block by lanthanides should include a discrimination of extracellular and intracellular effects.