Lanthanides Potentiate TRPC5 Currents by an Action at Extracellular Sites Close to the Pore Mouth*

Mammalian members of the classical transient receptor potential channel (TRPC) subfamily (TRPC1–7) are Ca2+-permeable cation channels involved in receptor-mediated increases in intracellular Ca2+. Unlike most other TRP-related channels, which are inhibited by La3+ and Gd3+, currents through TRPC4 and TRPC5 are potentiated by La3+. Because these differential effects of lanthanides on TRPC subtypes may be useful for clarifying the role of different TRPCs in native tissues, we characterized the potentiating effect in detail and localized the molecular determinants of potentiation by mutagenesis. Whole cell currents through TRPC5 were reversibly potentiated by micromolar concentrations of La3+or Gd3+, whereas millimolar concentrations were inhibitory. By comparison, TRPC6 was blocked to a similar extent by La3+ or Gd3+ at micromolar concentrations and showed no potentiation. Dual effects of lanthanides on TRPC5 were also observed in outside-out patches. Even at micromolar concentrations, the single channel conductance was reduced by La3+, but reduction in conductance was accompanied by a dramatic increase in channel open probability, leading to larger integral currents. Neutralization of the negatively charged amino acids Glu543and Glu595/Glu598, situated close to the extracellular mouth of the channel pore, resulted in a loss of potentiation, and, for Glu595/Glu598 in a modification of channel inhibition. We conclude that in the micromolar range, the lanthanide ions La3+ and Gd3+ have opposite effects on whole cell currents through TRPC5 and TRPC6 channels. The potentiation of TRPC4 and TRPC5 by micromolar La3+ at extracellular sites close to the pore mouth is a promising tool for identifying the involvement of these isoforms in receptor-operated cation conductances of native cells.


From the Institut fü r Pharmakologie, Freie Universitä t Berlin, Thielallee 69-73, Berlin 14195, Germany
Mammalian members of the classical transient receptor potential channel (TRPC) subfamily (TRPC1-7) are Ca 2؉ -permeable cation channels involved in receptormediated increases in intracellular Ca 2؉ . Unlike most other TRP-related channels, which are inhibited by La 3؉ and Gd 3؉ , currents through TRPC4 and TRPC5 are potentiated by La 3؉ . Because these differential effects of lanthanides on TRPC subtypes may be useful for clarifying the role of different TRPCs in native tissues, we characterized the potentiating effect in detail and localized the molecular determinants of potentiation by mutagenesis. Whole cell currents through TRPC5 were reversibly potentiated by micromolar concentrations of La 3؉ or Gd 3؉ , whereas millimolar concentrations were inhibitory. By comparison, TRPC6 was blocked to a similar extent by La 3؉ or Gd 3؉ at micromolar concentrations and showed no potentiation. Dual effects of lanthanides on TRPC5 were also observed in outside-out patches. Even at micromolar concentrations, the single channel conductance was reduced by La 3؉ , but reduction in conductance was accompanied by a dramatic increase in channel open probability, leading to larger integral currents. Neutralization of the negatively charged amino acids Glu 543 and Glu 595 /Glu 598 , situated close to the extracellular mouth of the channel pore, resulted in a loss of potentiation, and, for Glu 595 /Glu 598 in a modification of channel inhibition. We conclude that in the micromolar range, the lanthanide ions La 3؉ and Gd 3؉ have opposite effects on whole cell currents through TRPC5 and TRPC6 channels. The potentiation of TRPC4 and TRPC5 by micromolar La 3؉ at extracellular sites close to the pore mouth is a promising tool for identifying the involvement of these isoforms in receptor-operated cation conductances of native cells.
Mammalian isoforms of the classical transient receptor potential channel (TRPC) 1 subfamily, TRPC1-7, are likely candidates for cation channels mediating phospholipase C-depend-ent, receptor-operated Ca 2ϩ influx (for reviews, see Refs. [1][2][3][4][5]. The properties and activation mechanisms of TRPCs have been studied extensively in heterologous expression systems. By contrast, relatively little information is available on their role in native cells. Two recent studies (6,7) report that endogenous receptor-stimulated cation currents in vascular smooth muscle cells show properties identical to those described for heterologously expressed TRPC6. Therefore, precise knowledge of the properties of heterologously expressed TRPC channels may be essential to evaluate the involvement of TRPC proteins in receptor-operated cation conductances in native cells.
Structurally, TRP channels, like many other cation channels, have been proposed to have six transmembrane segments (S1-S6), intracellular N and C termini, and a pore-forming reentrant loop between S5 and S6. Based on amino acid sequence similarity, the mammalian members of the TRPC subfamily can be subdivided into four groups (4,8): TRPC1 (group 1), TRPC2 (group 2), TRPC3/6/7 (group 3), and TRPC4/5 (group 4). This subdivision is also supported by functional data. One of the major functional criteria is the mechanism of channel activation. From studies in heterologous expression systems, it is undisputed that receptor-mediated stimulation of phospholipase C is a key event in the activation of all TRPC isoforms. Evidence indicates that currents mediated by TRPC3, TRPC6, or TRPC7 can be activated by diacylglycerol (DAG) independently of protein kinase C (9 -12), although there is some controversy regarding the physiological significance of this stimulation (13). By contrast, TRPC4 and TRPC5 are not activated by DAG, and some evidence indicates that unidentified components of the phospholipase C pathway other than DAG or store depletion activate the channels (14 -16). Other evidence supports a store-dependent activation mechanism (17)(18)(19)(20). TRPC1 has been reported to be a store-dependent channel in some studies (21)(22)(23), but doubts have been raised as to whether TRPC1 expression results in the formation of functional plasma membrane channels in mammalian cells (16,24,25). The few data available support a store-operated activation mechanism for TRPC2 (26,27), but, here again, other evidence suggests that TRPC2 does not form functional channels in all tissues (28). Data from our group, confirmed in several independent laboratories, indicate that several biophysical features are also characteristic of certain groups of TRPCs. Thus, TRPC3-7 have a characteristic doubly rectifying, or S-shaped current-voltage relation (9, 14 -16, 29 -31). Furthermore, at the single channel level, even though the amplitudes of single channel events are similar, the openings of TRPC3 and TRPC6 are very brief (9,(32)(33)(34)(35) compared with those of TRPC4 and TRPC5 (15,16,31,36).
In the absence of more specific pharmacological tools, the lanthanides lanthanum (La 3ϩ ) and gadolinium (Gd 3ϩ ) are commonly used blockers of nonselective cation channels and other Ca 2ϩ -permeable channels. Interestingly, recent studies have reported that 100 M La 3ϩ has potentiating effects on mouse, rat, and human TRPC4 and mouse TRPC5 (15,16,31). However, the actions of La 3ϩ on TRPC4 and TRPC5 have not been characterized further. By contrast, other TRPCs are inhibited by micromolar concentrations of La 3ϩ or Gd 3ϩ (6,29,(37)(38)(39)(40).
Because the different effects of lanthanides are potentially an important distinguishing feature of the group 4 TRPC channels, we characterized the effect of these ions on TRPC5 in detail and compared them with those on TRPC6, a member of group 3. For this study we chose the rat TRPC6B slice variant (12), which lacks 54 amino acids at the distal N terminus compared with rat TRPC6A, and has not previously been characterized electrophysiologically. Unlike TRPC6A, TRPC6B has been reported to be activated by agonist application but not by 1-oleoyl-2-acetyl-sn-glycerol (OAG) (12). In whole cell patch clamp recordings, we found that TRPC5 was bimodally modulated by lanthanides, with potentiation at micromolar concentrations being succeeded by inhibition at millimolar concentrations. In contrast, TRPC6 was inhibited by micromolar concentrations and showed no potentiation. At the single channel level, the effects of La 3ϩ on TRPC5 are complex, affecting the single channel conductance, the mean open time, and the frequency of channel openings. By site-specific neutralization of extracellular negatively charged amino acids, we have identified two sites, close to the pore mouth, that are involved in potentiation of TRPC5 by La 3ϩ .

EXPERIMENTAL PROCEDURES
Molecular Biology and Stable Transfection-The isolation of TRPC5 from mouse brain total RNA has been described previously (15). For cloning of TRPC6, total RNA was prepared from rat brain or A7r5 smooth muscle cells using a TriZol reagent (Invitrogen) according to the standard protocol. For cDNA synthesis, 1 g of total RNA was reverse transcribed according to the protocol provided by the manufacturer using 200 units of Superscript II reverse transcriptase (Invitrogen) and 5 pmol of the primer 5Ј-CCAGTGAGCAGCAGAGTGACGAGGACTCG-AGCTCAAGCTTTTTTTTTTTTTTTTT. TRPC6 was amplified by 30 cycles of PCR using an annealing temperature 57°C for 15 s then extension for 210 s at 72°C with Expand-HF polymerase (Roche Molecular Biochemicals). The primer set used for amplification was 5Ј-CCG-GTACCGCCCTTATGAGCCGGGGTAATGAAAACAGAC (sense) and 5Ј-CCGGATCCCTATCTGCGGCTTTCCTCTTGTTT (antisense). The PCR products were subcloned into the pCR2.1 vector (Invitrogen) and the sequences confirmed by DNA sequencing of both strands (ABI-Prism, PerkinElmer Life Sciences). The rat TRPC6 characterized in this study corresponds to the sequence of rTRPC6B published by Zhang and Saffen (12) (GenBank™ accession number AB051213) with the exception of two amino acid exchanges: M757I and S767F.
For the generation of a stably transfected TRPC6 cell line, the KpnI/ BamHI fragment of the cloned cDNA was ligated into the tetracyclineinducible eukaryotic expression vector pcDNA4/TO and transfected into T-REx-293 cells (both from Invitrogen) using the transfection reagent FuGENE 6 (Roche Molecular Biochemicals). Clonal selection was performed according to the manufacturer's protocol with 250 g/ml Zeocin (Invitrogen), and 36 clones were functionally tested for TRPC6 expression with electrophysiological methods (measuring the current response to AlF 4 Ϫ infusion). Two positive clones were selected and used for further analysis.
Point mutations in TRPC5 C-terminally fused to yellow fluorescent protein (YFP) were introduced using the QuikChange site-directed mutagenesis kit (Stratagene) and appropriate primer sets. Sequences of the mutants were confirmed by DNA sequencing.
Cell Culture and Transient Transfection-Human embryonic kidney (HEK293) cells (ATCC, Manassas, VA) were maintained according to the supplier's recommendations. For transient transfection, cells were seeded in 35-mm culture dishes. The following day, 0.5-2 g/dish of pcDNA3 vector containing the cDNA for TRPC5, TRPC5-YFP, or point mutants of TRPC5-YFP was mixed with 100 ng/dish of the rat histamine H 1 receptor (in pcDNA3) and, in the case of TRPC5, 50 -100 ng/dish of pEGFP-C1 (Clontech), and transfected into the cells using the transfection reagent FuGENE 6 (Roche Molecular Biochemicals) ac-cording to the manufacturer's protocol. After 18 -24 h, the cells were trypsinized and seeded onto glass coverslips.
For experiments on excised patches, coverslips were coated with poly-L-lysine. All experiments were performed 2-3 days after transient transfection and, in the case of T-REx-r6 cells, 1-2 days after induction with 1 g/ml tetracycline (Roche Molecular Biochemicals).
Patch Clamp Recordings-Whole cell and single channel recordings were performed using an EPC-7 amplifier and Pulse software (HEKA, Lambrecht, Germany). Patch pipettes were made from borosilicate glass and had resistances of 3-6 megohms (whole cell recordings) or 6 -16 megohms (single channel recordings) when filled with the standard intracellular solutions.
Whole cell recordings were performed as described previously (15,31). To quantify current potentiation and current inhibition observed by bath application of lanthanides, we interpolated currents before and after application of lanthanides and normalized the values obtained in the presence of the lanthanides to the interpolated values. Interpolation was done to avoid errors arising from the fact that TRPC5 and TRPC6 currents decay with time. All current amplitudes were calculated as the difference between resting and histamine-, OAG-, or AlF 4 Ϫ -induced current levels. The concentration-response curve obtained for lanthanum inhibition of TRPC6 currents was fitted with Equation 1 relative amplitude ͑%͒ ϭ 100͑1 Ϫ 1/1 ϩ (IC 50 /͓La 3ϩ ͔͒ x ) (Eq. 1) to determine the IC 50 value. Values for the relative Ca 2ϩ permeability (P Ca /P Na ) were calculated from Equation 2, where V Ca and V Na are the reversal potentials in external solutions containing Ca 2ϩ and Na ϩ , respectively, and R, T, and F have their usual meanings.
For single channel recordings, the standard excised outside-out patch configuration (41) was used. After filtering at 10 kHz, single channel data were initially recorded onto digital audiotape (DAT, Biologic, Claix, France). For offline analysis, the single channel data were filtered at 1 kHz, subsequently digitized at 15 kHz, and analyzed with the pClamp6 software (Axon Instruments, Foster City, CA). Single channel amplitudes were obtained from events with open durations of more than 2 ms. In the case of TRPC5, channel activity was expressed as Values for t j were extracted from idealized traces generated with pClamp6.
The standard extracellular solution contained 140 mM NaCl, 5 mM CsCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM glucose, and 10 mM HEPES (pH 7.4 with NaOH). For NMDG ϩ solutions, Na ϩ and Cs ϩ were replaced by N-methyl-D-glucamine (NMDG), and Ca 2ϩ was omitted. In solutions with 20 mM CaCl 2 , the NaCl or NMDG ϩ concentration was reduced to 118 mM. The standard intracellular solution contained 110 mM cesium methanesulfonate, 25  Ϫ infusion, 1 l of 0.5 M NaF was mixed with 0.5 l of 3 mM AlCl 3 and diluted with 50 l of the pipette solution. The osmolarity of all solutions was between 290 and 310 mosmol/liter. All experiments were performed at room temperature (20 -25°C).
Fluorometric Measurements and Confocal Microscopy-Fluorometric measurements of Mn 2ϩ influx and confocal microscopy were performed as described previously (15,31).
Chemicals-Histamine, OAG, lanthanum (La 3ϩ ), and gadolinium (Gd 3ϩ ) were obtained from Sigma. Stock solutions were made in water or dimethyl sulfoxide (OAG) and diluted to final concentrations in the bath solutions.
Statistics-All data are given as the means Ϯ S.E. The statistical significance of differences between mean values was assessed using Student's t test. Differences were regarded as statistically significant for p Ͻ 0.05 and as highly statistically significant for p Ͻ 0.01.

Receptor-activated Currents through TRPC5 and TRPC6 -
We first performed whole cell voltage clamp recordings to compare currents mediated by TRPC5 with those mediated by TRPC6. The latter isoform was the rat TRPC6B splice variant, which has, to date, only been studied in fluorometric experiments (12) and has not been characterized electrophysiologically. As described previously (15), cells expressing TRPC5 and the histamine H 1 receptor displayed spontaneous channel activity. By contrast, no constitutive activity was observed at a holding potential of Ϫ60 mV in the T-REx-r6 cell line 24 -48 h after induction with tetracycline. Currents on break-in were in the range of Ϫ0.1 to Ϫ1.87 pA/picofarads (n ϭ 24), values not significantly different from those in control cells. Characteristic currents were activated by the application of 100 M histamine or infusion of AlF 4 Ϫ , a direct activator of G-proteins (44) in TRPC5- (Fig. 1A) or TRPC6-expressing cells but not in control cells. Furthermore, TRPC6-mediated inward currents could also be elicited by adding the membrane-permeable diacylglycerol OAG (100 M, n ϭ 11) (Fig. 1B) to the bath solution. This finding is in contrast to those of Zhang and Saffen (12), who detected receptor-but not OAG-induced Ba 2ϩ influx in COS cells expressing the species and splice variant used here.
As reported previously for murine and human isoforms of TRPC5 and TRPC6 (9, 15), the I-V relation of currents mediated by both TRPC5 and TRPC6 displayed a characteristic doubly rectifying shape and reversal potentials close to 0 mV, indicative of poor cation selectivity ( Fig. 1, C and D). Although the I-V relations of currents mediated by TRPC5 and TRPC6 were similar, currents in TRPC6-expressing cells display a slightly stronger outward rectification.
Effects of Lanthanides on Currents Mediated by TRPC5 or TRPC6 -We compared the effects of lanthanides on currents mediated by TRPC5 and TRPC6. When 10 M La 3ϩ was applied to HEK293 cells coexpressing mTRPC5 and the histamine H 1 receptor during exposure to 100 M histamine, inward and outward currents were increased (Fig. 1A). Comparison of the I-V relations before and after application of La 3ϩ revealed that the potentiating effect of La 3ϩ was much more pronounced at negative potentials (Fig. 1C). The effectiveness of very low concentrations of La 3ϩ and the absence of a shift in reversal potential exclude the possibility that La 3ϩ acts as a charge carrier for the additional inward current. In control cells transfected only with the histamine H 1 receptor, neither application of histamine nor subsequent addition of La 3ϩ resulted in a current increase (n ϭ 6). In these cells, application of La 3ϩ decreased basal leak currents rather than causing potentiation. These data suggest that La 3ϩ affects currents carried by TRPC5. By contrast, when 100 M La 3ϩ was applied to TRPC6expressing T-REx-r6 cells stimulated with 100 M OAG, rapid inhibition of the inward current was observed at a holding potential of Ϫ60 mV (Fig. 1B). Inhibition was almost complete at all potentials tested (Fig. 1, B and D). For both TRPC5-and TRPC6-mediated currents, the effects of lanthanides were readily reversible on wash-out (Fig. 1, A and B).
The concentration dependence of the effects of lanthanides on TRPC5 has not been studied previously. We therefore tested the effect of various concentrations of La 3ϩ or Gd 3ϩ on TRPC5mediated currents and compared them with those on TRPC6 (Fig. 2, A and B). Because agonist-induced currents in T-REx-r6 cells were relatively short lived, infusion of AlF 4 Ϫ via the pipette was preferred. AlF 4 Ϫ -induced currents in T-REx-r6 cells were indistinguishable from agonist-induced currents with respect to the I-V relation and their sensitivity to lanthanides. However, the current response was slowed considerably, thus allowing the effects of different concentrations of lanthanides to be investigated in the same cell. The pooled concentration-response relationships for TRPC5-and TRPC6mediated currents are shown in Fig. 2, C and D. Relative inhibition or potentiation was calculated as described under "Experimental Procedures." The inhibitory effect of lanthanides on TRPC6 began at concentrations around 1 M, and application of 1 mM resulted in near complete inhibition. The concentration-inhibition relationship for La 3ϩ inhibition of TRPC6 was sigmoidal ( Fig. 2D) with an IC 50 value of 6.1 M. The effects of Gd 3ϩ on currents mediated by TRPC6 were examined at two concentrations, 10 M and 1 mM. The values for relative inhibition by Gd 3ϩ were similar to those obtained with La 3ϩ . As seen in Fig. 2, A and C, the effects of La 3ϩ and Gd 3ϩ on currents mediated by TRPC5 were more complex. Starting at a concentration of around 1 M, both La 3ϩ and Gd 3ϩ increased TRPC5-mediated currents in a concentrationdependent manner. The largest potentiation was observed between 10 M and 1 mM and resulted, on average, in a 3-fold increase in current. At millimolar concentrations, the potentiating effect was reduced. In 5 mM La 3ϩ , the mean current was less than the control value, indicative of inhibition. In contrast, in 5 mM Gd 3ϩ , the mean relative current amplitude was ϳ120%. It has to be noted, however, that the effects of 5 mM Gd 3ϩ on TRPC5-mediated currents were not uniform. Some cells responded to this concentration with potentiation of currents (n ϭ 2/5), whereas others responded with current inhibition (n ϭ 3/5). Similarly, dual effects of lanthanides were often observed during wash-in or wash-out of the ions at a concentration of 5 mM. On wash-in, inhibition was sometimes preceded by transient potentiation and on wash-out, removal of inhibition followed by transient potentiation (see inset in Fig.  2A). Thus, lanthanides inhibit TRPC6 but have dual effects on TRPC5. Low concentrations potentiate TRPC5 currents, but high concentrations are less effective and may result in inhibition.
Effect of Extracellular Ca 2ϩ on Potentiation by La 3ϩ -To investigate whether physiological cations can act at the same site as lanthanides, we tested the effect of Ca 2ϩ . Raising [Ca 2ϩ ] o from 2 to 20 mM resulted in a rapid increase in TRPC5 channel currents similar to that produced by micromolar La 3ϩ concentrations (n ϭ 5; data not shown). This similarity includes a stronger potentiation of currents at negative than at positive potentials. In addition to the rapid increase in inward current, there was a slower increase in both inward and outward currents which may reflect channel activation by an increase in [Ca 2ϩ ] i resulting from Ca 2ϩ entry through the channel (14, 15).
The decay of current upon returning [Ca 2ϩ ] o to 2 mM was slow compared with that after removal of La 3ϩ . In the presence of 20 mM Ca 2ϩ , potentiation by 10 M La 3ϩ was prevented or strongly reduced (n ϭ 5). These results indicate that Ca 2ϩ can compete with La 3ϩ for the same site.
High Concentrations of Intracellular BAPTA Do Not Reduce Potentiation by La 3ϩ -In general, the actions of lanthanides on ion channel behavior are considered to be strictly confined to the extracellular face of the plasma membrane. However, a recent report on lanthanide-induced inhibition of currents mediated by human TRPC3 transiently expressed in Chinese hamster ovary cells suggested intracellular regulatory actions for both La 3ϩ and Gd 3ϩ , although additional extracellular effects could not be excluded (39). If the effects of La 3ϩ on TRPC6-and TRPC5-mediated currents result from an intracellular action of the free cation they should depend on the trivalent cation buffering capacity of the intracellular solution. Thus, raising the buffer capacity of the intracellular solution by replacing 10 mM EGTA with 30 mM BAPTA, which binds both La 3ϩ and Gd 3ϩ more strongly than Ca 2ϩ , should reduce the lanthanide effects on TRPC5 and TRPC6. The free Ca 2ϩ concentration of both pipette solutions was kept at ϳ100 nM. As seen in Fig. 3, A and C, however, with BAPTA, 10 M La 3ϩ was still able to increase histamine-induced TRPC5 currents strongly. The relative increase of TRPC5-mediated currents was even more pronounced with BAPTA (7-fold, n ϭ 4) than with EGTA in the pipette solution. Interestingly, the amplitudes of basal and histamine-induced (Ϫ4.5 Ϯ 3.2 pA/picofarads, n ϭ 4, C m ϭ 21.9 Ϯ 5, p Ͻ 0.05) TRPC5 currents were greatly reduced in BAPTA-containing pipette solutions, confirming previous reports indicating that TRPC5-mediated currents are dependent on [Ca 2ϩ ] i (14,15). In the presence of intracellular BAPTA, the inhibitory effect of La 3ϩ on TRPC6 ( Fig. 3, B and D; inhibition by 89.5 Ϯ 2.2%, n ϭ 4) was similar to that in EGTA (96.8 Ϯ 1.5%, n ϭ 12). These results provide evidence for an extracellular site of action for the potentiating and inhibitory effects of lanthanides on TRPC5 and TRPC6, respectively.
Effects of La 3ϩ on Single Channel Currents through TRPC5-To investigate the mechanism involved in the dual effect of La 3ϩ on TRPC5, the single channel properties of TRPC5 were characterized in the outside-out patch configuration. Under the same conditions used for whole cell recordings, stimulation with 100 M histamine activated single channel events of around Ϫ2.5 pA (Fig. 4, A and C, and Table I) in TRPC5-expressing cells at a potential of Ϫ60 mV. To determine the potential dependence of the single channel current amplitude for TRPC5, the membrane potential was stepped from Ϫ100 to ϩ80 mV (Fig. 4C). Under the conditions used, patches tended to be less stable at positive potentials. The pooled single channel i-V relationship for TRPC5 (Fig. 4D) closely resembled that of whole cell currents, with a reversal potential close to 0 mV. At Ϫ60 mV, the chord conductance of TRPC5 was 41.3 Ϯ 1.1 picosiemens (n ϭ 12), and the mean open time of histamineinduced TRPC5 currents 7.5 ms (cf. Table I). As reported previously for human TRPC6 (9), single channel openings of rat TRPC6, elicited by application of histamine or OAG, were typically of very short duration, with more than 70% of the openings shorter than 0.5 ms. Hence, under the recording conditions used, most openings of TRPC6 channels could not be fully resolved. From openings longer than 2 ms, the amplitude of single channel events at a holding potential of Ϫ60 mV was Ϫ2.81 Ϯ 0.07 pA (n ϭ 8, Fig. 4B and Table I), a value that yielded a single channel chord conductance at Ϫ60 mV of 46.6 Ϯ 1.5 picosiemens (n ϭ 12).
When La 3ϩ was added to the bath solution at a concentration of 100 M, the open probability of TRPC5 was dramatically increased (Fig. 5A). At the same time, the single channel current amplitude was approximately halved compared with that recorded before the addition of La 3ϩ (Fig. 5A). An analysis of the La 3ϩ block of TRPC6 was precluded by the extremely short events, which also made estimates of open probability difficult. Channel activity was, however, abolished by the addition of 100 M La 3ϩ (Fig. 5B) and was associated with a decrease in the frequency of channel openings (Fig. 5C).
A reduction in the single channel current of TRPC5 upon application of 100 M La 3ϩ was observed at all potentials (Fig.  6, A and B), with no increase in open channel noise. This result indicates that La 3ϩ may produce a very fast, flickery block of the channel, not resolvable under our recording conditions, at a site outside the membrane electrical field. Inhibition by an allosteric mechanism cannot be excluded. The shape of the i-V relationship of TRPC5-mediated currents and the reversal potential of around 0 mV were preserved in the presence of La 3ϩ (Fig. 6B) (Table I). Because brief events (Ͻ0.5 ms, see "Experimental Procedures") were excluded from the analysis, and the proportion of these decreases with increasing La 3ϩ concentrations, the effect on mean open time will be underestimated. The combined effect of the increase in NP o and the decrease in single channel current was an ϳ10-fold increase in the total current through the patch (NP o ⅐i). Interestingly, the increase in patch current had a nearly identical dependence on the La 3ϩ concentration to the whole cell current. From the single channel data, the EC 50 for potentiation by La 3ϩ was around 3 M, whereas a 50% reduction of the single channel current occurred at a La 3ϩ concentration of 100 M. Taken together, the above data suggest that La 3ϩ exerts diverse effects on single channel activity, having opposite effects on channel conductance and open probability.
Identification of Amino Acids Involved in Potentiation of TRPC5 by La 3ϩ -In an attempt to identify the site involved in potentiation of TRPC5 by La 3ϩ , we searched for negatively charged amino acids (Glu and Asp residues) in the putative extracellular loops of the TRPC5 protein. We identified 10 residues that are conserved in TRPC5 (Fig. 7A) and TRPC4 and generated six point mutants in which individual residues, and two in which pairs of close neighbors, were neutralized (Glu to Gln, or Asp to Asn). The C-terminally YFP-tagged mutants were expressed in HEK293 cells and their subcellular distribution and membrane targeting determined by confocal laser microscopy. Like wild type TRPC5 (TRPC5-wt) (15), the mutants displayed a clustered appearance in the plasma membrane and retention in a perinuclear compartment.
As an initial probe for channel function, we tested for the presence of histamine-induced Mn 2ϩ influx in fura-2AM-loaded cells. Five of the mutants (D392N, E404Q and E543Q, E570Q and E595Q/E598Q) displayed robust accelerations in Mn 2ϩ influx upon addition of histamine, E559Q only weak responses (data not shown). In contrast, E467Q/E470Q and E479Q did not respond. In whole cell recordings, the mutants that showed robust accelerations in Mn 2ϩ influx displayed histamine-activated currents indistinguishable from those of TRPC5-wt (n Ն 10 for each mutant). Addition of La 3ϩ resulted in increases in current in the mutants D392N, E404Q, and E570Q similar to those observed for TRPC5-wt (data not shown). In contrast, neutralization of negatively charged amino acids at two sites in the putative pore-forming loop, E543Q, just after transmembrane segment S5, and E595Q/E598Q, close to S6, led to a loss of La 3ϩ -induced potentiation (Figs. 7 and 8). Both of these mutants did, however, display inhibition. This inhibition was different for the two mutants. Both inward and outward currents mediated by the mutant E543Q were inhibited to a similar extent by La 3ϩ (Fig. 7, B and C). In contrast, inward currents mediated by the mutant channel E595Q/E598Q were inhibited to a larger extent than outward currents (Fig. 8, A and B).
We also tested whether potentiation by Ca 2ϩ was influenced in the mutants E543Q and E595Q/E598Q. Currents through the mutants did not show the rapid potentiation observed upon raising [Ca 2ϩ ] o from 2 to 20 mM for TRPC5-wt (data not shown). Both inward and outward currents mediated by  (5 mM E543Q were inhibited in 20 mM Ca 2ϩ (n ϭ 3, data not shown). In contrast, currents through E595Q/E598Q slowly increased to a maximum, then declined spontaneously in the continued presence of 20 mM Ca 2ϩ (n ϭ 3, data not shown). In these experiments, there was a more noticeable shift in the current reversal potential on raising [Ca 2ϩ ] o from 2 to 20 mM for E595Q/E598Q than for TRPC5-wt or E543Q. We therefore quantified the relative Ca 2ϩ permeability of E595Q/E598Q and compared it with that of TRPC5-wt. Current reversal potentials were first measured from voltage ramps in a nominally Ca 2ϩ -free, Na ϩ solution, then in a Na ϩ -free (NMDG ϩ ) solution containing 20 mM Ca 2ϩ . From the reversal potentials, we calculated values for P Ca /P Na of 1.83 Ϯ 0.18 (n ϭ 3) for TRPC5-wt, a value close to the 1.79 in our previous study (15), and 4.28 Ϯ 0.08 (n ϭ 4) for E595Q/E598Q. Thus, the mutants E543Q and E595Q/E598Q did not show rapid potentiation by La 3ϩ and Ca 2ϩ , further supporting a similar site of action. Furthermore, mutant E595Q/E598Q had a higher relative Ca 2ϩ permeability than TRPC5-wt. Effects of La 3ϩ on the Single Channel Properties of the Mutants E543Q and E595Q/E598Q-In outside-out patches, the single channel properties of the mutant E543Q were similar to those of TRPC5-wt (Fig. 7D). The mean single channel current at Ϫ60 mV was Ϫ2.63 Ϯ 0.10 pA (chord conductance, 43.8 picosiemens; n ϭ 4), a value not significantly different (p ϭ 0.28) from that of the wild type channel. It is noteworthy that very high levels of channel activity were observed in patches from mutant E543Q-expressing cells, necessitating the use of much smaller pipettes. Addition of 100 M La 3ϩ to the extracellular solution reduced the current amplitude to Ϫ1.47 Ϯ 0.16 pA (Fig. 7D), a decrease similar to that observed for TRPC5-wt. Like the inhibition of TRPC5-wt, the reduction in channel current was not accompanied by an increase in open channel noise. However, in contrast to the wild type channel, application of 100 M La 3ϩ did not result in an increase in NP o (0.50 Ϯ 0.21 and 0.53 Ϯ 0.14 (n ϭ 4; p ϭ 0.8) in control and 100 M La 3ϩ , respectively). The mutant E595Q/E598Q showed more drastic changes in its single channel properties (Fig. 8C). The single channel current of Ϫ2.98 Ϯ 0.03 pA at Ϫ60 mV (chord conductance, 49.7 picosiemens; n ϭ 9) was significantly (p Ͻ 0.01) larger than that for TRPC5-wt. Application of 100 M or 1 mM La 3ϩ resulted in weaker reductions in current amplitude than in TRPC5-wt (to Ϫ2.50 Ϯ 0.03 pA, n ϭ 7 and Ϫ1.25 Ϯ 0.07 pA, n ϭ 4, respectively) and a clear increase in open channel noise indicative of a slower, flickery channel block (Fig. 8C). Thus, at the single channel level, neither mutant showed potentiation. E543Q had a similar conductance and was inhibited by La 3ϩ in a manner similar to TRPC5-wt. In contrast, E595Q/ E598Q had a higher conductance than TRPC5-wt or E543Q, and inhibition by La 3ϩ was modified. DISCUSSION In the present study, we show that currents mediated by TRPC5 and TRPC6 are affected differently by the lanthanides La 3ϩ and Gd 3ϩ . Although whole cell currents through TRPC6 were inhibited concentration-dependently by lanthanides, those through TRPC5 were potentiated by low concentrations but inhibited by high concentrations. The dual effect of La 3ϩ on TRPC5 was also observed at the single channel level and involved a combination of an inhibitory effect on channel amplitude and an increase in channel open probability. By an analysis of point mutants, we identified two sites, close to the extracellular mouth of the pore, which are involved in La 3ϩand Ca 2ϩ -induced channel potentiation.
Inhibitory effects of different di-and trivalent cations, including La 3ϩ and Gd 3ϩ , have been described for most Ca 2ϩpermeable channels. Accordingly, current block by bath application of lanthanide ions has been reported for several members of the TRPC subfamily of TRP channels, e.g. for human and mouse TRPC3 (29,(37)(38)(39), for mouse TRPC6 (6,45), and for human TRPC7 (40). There is considerable variability in the IC 50 values obtained, with higher values in Ca 2ϩ imaging experiments than in electrophysiological recordings. The IC 50 value obtained for rat TRPC6 in the present study was in good agreement with the values obtained in whole cell patch clamp experiments for mouse TRPC6 (6) and human TRPC3 (39).
The pronounced increase in agonist-induced currents in TRPC5-expressing cells with micromolar La 3ϩ in the present study are in agreement with previous studies on this channel (15,16), and we have extended this observation to Gd 3ϩ , which is approximately equally effective. A further novel finding of the present study is that higher lanthanide concentrations (Ն1 mM) were less effective in potentiating the current and even reversibly inhibited currents carried by TRPC5. Millimolar concentrations of Ca 2ϩ also potentiated agonist-induced currents and prevented the effects of micromolar La 3ϩ , suggesting that physiological divalent cations may also bind at the same site.
Evidence for two different actions of lanthanides on the channel was supported by the effects of La 3ϩ on single channel currents in outside-out patches. La 3ϩ caused a concentrationdependent decrease in single channel current amplitude, while, at the same time, increasing the channel open probability (NP o ). Both effects were already observed at a concentration of 1 M. The concentration dependence of the increase in current in outside-out patches (NP o ⅐i) closely paralleled the increase in whole cell current, although the maximum potentiation was, on average, about 3-4-fold higher in outside-out patches than in whole cell experiments. Because of our inability to resolve currents at millimolar concentrations of La 3ϩ , it is not clear from the single channel data why potentiation declines and inhibition occurs. A decrease in single channel current is at least partly responsible for the decrease in whole cell current.
There are few reports of potentiating effects of La 3ϩ on ion channel currents, and more importantly, to our knowledge, there are no reports that describe dual effects on ion channel activity. Potentiating actions of 100 M La 3ϩ have been observed in whole cell recordings for mouse, rat, and human TRPC4-and mouse TRPC5-mediated currents (15,16,31) and for receptor-operated currents in cells coexpressing mouse TRPC1 and mouse TRPC5 (16). In the latter, heteromultimers of TRPC1 and TRPC5 are thought to be formed, which, compared with homomeric TRPC5, have a drastically reduced single channel current (Ϫ0.5 pA at a holding potential of Ϫ60 mV). The single channel current amplitude was not affected by the inclusion of La 3ϩ in the pipette solution. For native nonselective cation currents, there is one report of a potentiation of the native current (I cat ) in rat ileal smooth muscle cells by La 3ϩ , with an apparent K d of 190 M (46). From relaxation analysis, prolonged single channel mean open life times were suggested to be the main cause of the augmentative effect of La 3ϩ . Interestingly, in the mouse, TRPC4 is expressed in this tissue (47). With regard to heterologously expressed TRPC4 and TRPC5, it should be noted that some studies reported an inhibition by micromolar lanthanide concentrations (14,48).
The loss of the potentiating effects of La 3ϩ and Ca 2ϩ in mutants of two sites (Glu 543 and Glu 595 /Glu 598 ), which, according to models of TRP channel structure, are located opposite each other at the start and end of the pore-forming loop between S5 and S6, strongly supports an extracellular site of action. Importantly, identical amino acids are present in TRPC4 at the positions corresponding to Glu 543 and Glu 595 in TRPC5, but acid amino acids are not present at corresponding positions in TRPC3, TRPC6, and TRPC7. Larger variations in structure prevent an identification of corresponding residues in TRPC1. The differences between the TRPC isoforms provide an explanation for the specificity of the potentiating effect for TRPC4 and TRPC5. Interestingly, these sites are analogous to those in TRPV1 (VR1) which are involved in proton-mediated channel potentiation (Glu 600 ) and proton-mediated channel activation (Glu 648 ) (49) and can modulate sensitivity to the activator capsaicin (49,50). Indeed, at the latter site TRPC4, TRPC5, and TRPV1 have identical EFTE motifs. Because the distal steps leading to activation of this channel and the activation mechanism are not known, it is not clear how La 3ϩ or Ca 2ϩ binding to the extracellular sites results in current potentiation. By analogy to TRPV1, where neutralization of Glu 600 and Glu 648 leads to potentiation of the capsaicin sensitivity (49,50), it is tempting to speculate that La 3ϩ or Ca 2ϩ , by neutralizing the negative charges, potentiate the response of TRPC5 to its unknown activator.
The TRPC5 mutation E595Q/E598Q also affected inhibition by La 3ϩ , whereas E543Q did not. For the wild type channel, the reduction in single channel current by La 3ϩ at all potentials without an increase in open channel noise is indicative of a fast block at a site outside the membrane electrical field. Similarly, inhibition of whole cell currents in the mutant E543Q, which lacked potentiation, was potential-independent. In contrast, the mutant E595Q/E598Q showed a slower flickery block at the single channel level and a clear potential dependence of whole cell current inhibition, with inward currents being more strongly reduced than outward currents. The loss of the fast block by mutation at the extracellular site E595Q/E598Q and the potential dependence of the block remaining after mutation indicate that in both cases La 3ϩ blocks the channel from the outside. The effect of this mutation on inhibition by La 3ϩ , the increase in single channel current, and the increase in P Ca /P Na suggest that this site lies close to, or in, the permeation pathway. Considering the change in channel inhibition by La 3ϩ , it is possible that the increase in single channel current in E595Q/E598Q results from a reduction in block by a physiological cation. By analogy to other channels with similar structure, the glutamates will form a negatively charged ring around the extracellular pore mouth, with, in tetramers, at least 12 negatively charged residues. These amino acids may act as "gatekeepers" controlling cation entry into the pore.
Further evidence that both potentiation of TRPC5 and inhi-bition of TRPC6 by lanthanides results from an extracellular action of La 3ϩ is provided by the presence of the effects in experiments with intracellular EGTA buffers and their persistence in the presence of higher concentrations of BAPTA. Both buffers have a very high affinity for lanthanides. Recently, Halaszovich et al. (39) suggested that La 3ϩ and Gd 3ϩ block human TRPC3 channels from the cytosolic side of the membrane and that different apparent IC 50 values might simply reflect different uptake rates for lanthanide ions in different cell types (see below). Our data support an extracellular site of action on TRPC6 and on TRPC5, although we cannot exclude additional intracellular effects. Because of the variability in results from different laboratories, the applicability of results from heterologous overexpression studies on TRPC channels to native channels has recently been questioned (e.g. 5). However, for TRPC6, at least, properties nearly identical to those observed after overexpression are seen for native channels in vascular smooth muscle cells (6,7). These properties include the doubly rectifying I-V relation, store depletion-independent activation, as well as stimulation by DAGs (13). For the other TRPC channels, too few data are available. Nonetheless, the elucidation and comparison of the properties of different heterologously expressed TRPC channels may be valuable tools to help clarify the role of TRPC channels in native cells. TRPC4 and TRPC5 (15,16,31) stand out from other TRPC channels and other nonselective cation channels in that they are strongly potentiated by La 3ϩ in the micromolar range. This effect is a functional property that can be assigned exclusively to group 4 TRPCs or heteromultimers containing group 4 TRPCs and may be an important distinguishing feature. Another characteristic property is the shape of the I-V relation. Currents from overexpressed, presumably homomeric, TRPC3-7 have a doubly rectifying I-V relation with a reversal potential close to 0 mV in physiological solutions. As also indicated in this study for TRPC6, the relative amplitude of outward compared with inward currents is larger for TRPC3 and TRPC6 than for TRPC4 and TRPC5. Interestingly, heteromers of TRPC1 and TRPC5 display a different shape of I-V relation, with a strong reduction in current at negative membrane potentials which results in a U-shaped I-V relation for inward currents (16). The activation of TRPC3, 6, and 7 by DAGs in a protein kinase C-independent manner is a characteristic feature of this group of channels, not shared with other TRPC channel groups (9,29). A previous report that the TRPC6B splice variant might be an exception to this rule in not being activated by DAGs (12) could not be confirmed in the present study. Within this group of channels, the effect of flufenamate can distinguish TRPC6 from TRPC3 and TRPC7. Flufenamate potentiates TRPC6 but inhibits TRPC3 and TRPC7 (6). Moreover, in the present study, the mean open time of TRPC5 and TRPC6 single channel events were found to be 7.5 ms and Ͻ1 ms, respectively. Similar results have been obtained in independent studies: brief opening events appear to be common for human TRPC3 (32)(33)(34)(35) and human TRPC6 (9) whereas mean open life times longer than 1 ms have been reported for mouse, human, and rat TRPC4 and mouse TRPC5 (15,16,31,36). No information is available on single channel properties of TRPC7. Hence, differential effects of lanthanides, activation by diacylglycerols, and single channel open times represent promising tools to distinguish contributions of the TRPC3/6/7 group from those of the TRPC4/5 group when endogenous phospholipase C-dependent cation currents are studied.
In conclusion, the potentiation of TRPC5 by micromolar La 3ϩ and Gd 3ϩ , a feature shared with TRPC4 but not with most other nonselective cation channels, involves interactions with negatively charged amino acids situated close to the extracellular pore mouth.