Transient receptor potential canonical 4 (TRPC4) forms non-selective cation channels implicated in the regulation of diverse physiological functions. Previously, TRPC4 was shown to be activated by the Gi/o subgroup of heterotrimeric G-proteins involving Gαi/o, rather than Gβγ, subunits. Because the lifetime and availability of Gα-GTP are regulated by regulators of G-protein signalling (RGS) and Gαi/o-Loco (GoLoco) domain-containing proteins via their GTPase-activating protein (GAP) and guanine-nucleotide-dissociation inhibitor (GDI) functions respectively, we tested how RGS and GoLoco domain proteins affect TRPC4 currents activated via Gi/o-coupled receptors. Using whole-cell patch-clamp recordings, we show that both RGS and GoLoco proteins [RGS4, RGS6, RGS12, RGS14, LGN or activator of G-protein signalling 3 (AGS3)] suppress receptor-mediated TRPC4 activation without causing detectable basal current or altering surface expression of the channel protein. The inhibitory effects are dependent on the GAP and GoLoco domains and facilitated by enhancing membrane targeting of the GoLoco protein AGS3. In addition, RGS, but not GoLoco, proteins accelerate desensitization of receptor-activation evoked TRPC4 currents. The inhibitory effects of RGS and GoLoco domains are additive and are most prominent with RGS12 and RGS14, which contain both RGS and GoLoco domains. Our data support the notion that the Gα, but not Gβγ, arm of the Gi/o signalling is involved in TRPC4 activation and unveil new roles for RGS and GoLoco domain proteins in fine-tuning TRPC4 activities. The versatile and diverse functions of RGS and GoLoco proteins in regulating G-protein signalling may underlie the complexity of receptor-operated TRPC4 activation in various cell types under different conditions.
Transient receptor potential canonical (TRPC) channels are receptor-operated Ca2+-permeable cation channels involved in many physiological processes (see reviews in [1,2]). Among the seven TRPCs (TRPC1–TRPC7), TRPC4 has been implicated to function in neurons, smooth muscles, endothelium and cancer. These include contributions to epileptiform burst firing in brain neurons and seizure-induced neurodegeneration [3,4], synaptic transmission [5,6], contractility regulation of intestinal smooth muscle [7,8], microvascular permeability  and renal cancer proliferation [10,11]. These activities are believed to be related to Ca2+ and Na+ influx mediated by TRPC4 channels, which triggers Ca2+ signalling and membrane depolarization. To achieve strong control of the cellular function, the TRPC4 channels are tightly regulated through multiple levels of cross-talk among signalling networks .
Generally, the activation of TRPC channels is thought to be triggered by the stimulation of the phospholipase C (PLC) pathway via either the Gq/11 subgroup of heterotrimeric G-proteins or receptor tyrosine kinases (RTKs) [13,14]. However, for TRPC4 and TRPC5, the Gi/o subgroup of G-proteins also plays an important role in channel activation [15–18]. In particular, the activation of TRPC4 depends on activated Gαi/o subunits , which are usually produced through stimulation of a subset of G-protein-coupled receptors (GPCRs), known as Gi/o-coupled receptors. To fully activate TRPC4, the stimulation by Gi/o proteins also needs to coincide with Ca2+-dependent activation of PLCδ1, which forms a positive-feedback loop, allowing sustained TRPC4 activity .
However, it was not clear to what extent the relative activities of Gi/o proteins and their sustainability affect TRPC4 channel function during continued GPCR stimulation. GPCRs may be considered as guanine-nucleotide-exchange factors (GEFs) that promote the release of GDP from heterotrimeric G-proteins in exchange for binding of GTP. This causes the dissociation of the heterotrimer into GTP-bound Gα and free Gβγ subunits. Each GPCR type has its own subset of preferred G-proteins with specificity set typically by the Gα subunits. For example, M2 muscarinic receptor (M2R) and μ-opioid receptor (μOR) are coupled to Gi/o, whereas M3 muscarinic receptor (M3R) is coupled to Gq/11. Both Gα-GTP and free Gβγ dimers act as signal transducers in cell signalling through effector coupling. The termination of G-protein signalling is determined by the intrinsic GTPase activity of the Gα subunit, which hydrolyses GTP into GDP, allowing the Gα to reassociate with the Gβγ subunits.
The intrinsic GTPase activity of Gα can be accelerated by GTPase-activating proteins (GAPs), such as regulator of G-protein signalling (RGS) proteins. A family of more than 30 genes encoding RGS proteins has been identified . Through GAP activities, the RGS proteins help to switch off G-protein signalling and would therefore be expected to accelerate the deactivation kinetics of downstream effectors and decrease their activities.
Gαi/o-Loco (GoLoco) motif containing proteins, on the other hand, act as guanine-nucleotide-dissociation inhibitors (GDIs) of Gα subunits , which interrupt the GDP dissociation from Gα and in turn prevent G-protein activation by GPCRs or GEFs. The GoLoco motifs specifically act at Gi/o proteins, locking Gαi/o in the inactive GDP-bound form but releasing Gβγ at the same time. This results in an inhibition of Gαi/o-mediated signalling but an enhancement of Gβγ-mediated functions [21,22]. The mammalian GoLoco motif-containing proteins consist of a diverse group of distantly related members sharing one or more 19-amino-acid GoLoco motifs, including in group 1 the R12 subfamily of RGS proteins (RGS12 and RGS14), in group 2 G-protein signalling modulators 1 and 2 (GPSM1 and GPSM2), formerly known as activator of G-protein signalling 3 (AGS3) and LGN respectively, and in group 3, GPSM3 and GPSM4, also known as G18 and Purkinje cell protein-2 (Pcp-2) respectively [21,22].
Previously, G-protein-activated inwardly rectifying K+ (GIRK) channels have been shown to be modulated by RGS and GoLoco proteins [23–26]. These studies confirmed some predictions expected from regulation by the Gβγ dimer, such that, although the basal GIRK current was increased by LGN due to enhanced free Gβγ release, deactivation of the stimulated GIRK current was accelerated by RGS4 because of its GAP function. However, the studies also uncovered some unexpected effects. For example, RGS4 also accelerated the agonist-induced GIRK channel activation and increased the current amplitude [23,24]; the GoLoco motifs progressively reduced the responses of the channel to repeated agonist stimulation .
Because TRPC4 channels are activated by Gi/o signalling, the modulation of these G-proteins by RGS and GoLoco proteins will probably affect the activation process of TRPC4 channels. In the present study, we examined how several RGS and GoLoco domain proteins affect Gi/o-mediated activation of TRPC4 heterologously expressed in human embryonic kidney (HEK)293 cells. We show that TRPC4 currents activated through stimulation of Gi/o-coupled receptors are suppressed by the expression of Gαi/o-coupled RGS and GoLoco proteins. The GAP and GoLoco domains are responsible for the inhibitory actions of these proteins. On the other hand, these proteins do not alter the surface expression of TRPC4 proteins or TRPC4 currents elicited by its direct agonist, englerin A [10,11], indicating that RGS and GoLoco proteins specifically regulate Gi/o-mediated TRPC4 function.
MATERIALS AND METHODS
cDNA constructs and mutagenesis
cDNAs for human RGS4, RGS6, RGS12, RGS14, LGN, AGS3 and 3x-HA–M2R (HA is haemagglutinin) were purchased from Missouri S&T cDNA Resource Center (http://www.cdna.org) and were placed in pcDNA3.1+ (Invitrogen). Point mutations of human RGS4 and RGS14 were introduced using the QuikChange site-directed mutagenesis kit (Stratagene). The mutations were verified by DNA sequencing. The AGS3-short (AGS3sh) and myristoylatable AGS3 short form (Myr-AGS3sh) expression constructs were provided by Dr J.B. Blumer and Dr S. Lanier (Medical University of South Carolina, Charleston, SC, U.S.A.). cDNAs for mouse TRPC4β in pEGFPN1 (Clontech) and rat M2R or μOR in pIREShyg2 (Clontech) were as described previously [16,18].
Cell culture and transient transfection
HEK293 cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% (v/v) FBS and 2 mM L-glutamine at 37°C in a humidity-controlled incubator with 5% CO2. The stable HEK293 cell line expressing μOR  was maintained in the above medium supplemented with 100 μg/ml hygromycin B (Calbiochem). All cell culture reagents were purchased from Invitrogen and Sigma–Aldrich. For transient transfection, cells were seeded in wells of a 12-well plate and allowed to grow overnight. The following day, the transfection was carried out using polyethyleneimine (PEI) and a total of 0.5 μg/well cDNA as recently described . For co-expression with μOR, the TRPC4β/RGS (or GoLoco) cDNA ratio was 1:1.5 and the transfection was performed on the stable μOR-expressing cells. For co-expression with M2R, the TRPC4β/M2R/RGS (or GoLoco) cDNA ratio was 1:1:1 and transfection was performed on wild-type HEK293 cells. Electrophysiological recordings were performed between 24 h and 36 h after transfection.
After trypsinization, cells were transferred to a recording chamber on the stage of an inverted fluorescence microscope (TE200, Nikon) and allowed to attach to the glass coverslip at the bottom of the chamber for at least 10 min prior to patch-clamp recording. Transfected cells were identified by the green fluorescence of TRPC4β–GFP. Whole-cell voltage clamp recordings were made using pipettes pulled from standard wall borosilicate tubing with filament (Sutter Instrument) to a tip resistance of 3–6 MΩ when filled with the intracellular solution containing 140 mM CsCl, 0.5 mM EGTA, 0.2 mM Tris/GTP, 3 mM Mg-ATP and 10 mM HEPES, with the pH adjusted to 7.3 using CsOH. The standard or physiologically relevant external solution [PSS (physiologically relevant external solution) or normal tyrode's] contained 135 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM glucose and 10 mM HEPES, with the pH adjusted to 7.4 using NaOH. The Cs+-rich external solution was prepared by replacing NaCl and KCl of the PSS with equimolar CsCl and the pH was adjusted to 7.4 also using CsOH.
Voltage commands were given and currents were recorded using a MultiClamp 700A amplifier, coupled to Digidata 1350A, and operated using the pCLAMP software (v.9) (all from Molecular Devices). Currents were continuously recorded at 5 kHz with the cell held at −60 mV. Voltage ramps from +100 to −100 mV over a period of 500 ms were applied from the holding potential of −60 mV every 10 s to examine the current–voltage (I–V) relationship of the currents. Carbamoylcholine (carbachol; CCh) was purchased from Sigma–Aldrich, [D-Ala2, N-MePhe4, Gly-ol]-enkephalin (DAMGO) was from Bachem Chemicals and englerin A was from Cerilliant. The drugs were diluted to the final desired concentrations in the Cs+-rich external solution and applied using a gravity-fed continuous whole-chamber perfusion system. All electrophysiological recordings were performed at room temperature (22–24°C). Data analyses were made using pCLAMP v.10.3 and Origin software v.75 (Microcal).
Western blotting and surface biotinylation
Transfected cells were washed with PBS and then incubated in 0.5 mg/ml sulfo-succinimidyl-6-(biotinamido) hexanoate (sulfo-NHS-LC-biotin) (Pierce) in PBS for 30 min on ice. Free biotin was quenched by the addition of 100 mM glycine in PBS, after which cell lysates were prepared by passing the cell suspension in a lysis buffer [50 mM HEPES, pH 7.4, 150 mM NaCl, 2 mM EDTA, 2 mM MgCl2, Complete™ protease inhibitor mixture tablet (Roche Applied Science) and 1% Triton X-100] through a 26-gauge needle 10–20 times. The lysates were centrifuged at 13300 g for 15 min at 4°C to remove any insoluble material, and protein concentrations of the supernatants were estimated by the absorbance at 280 nm determined using a NanoDrop-1000 instrument (Thermo Scientific). To isolate biotinylated (surface-expressed) proteins, 40 μl of a 50% slurry of NeutrAvidin beads (Pierce) was added to cell lysates containing 500 μg of proteins. After incubation for 1 h at room temperature with continuous rotation, the mixtures were centrifuged at 325 g at room temperature for 2 min in a microcentrifuge and supernatants were discarded. The beads were then washed three times with ice-cold 0.5% Triton X-100 in PBS. For TRPC4–GFP, the washed beads were extracted in 4× SDS/PAGE sample buffer (1× contains 62.5 mM Tris/HCl, 2.1% SDS, 5% 2-mercaptoethanol and 13.1% glycerol, pH 6.8) with heating at 60°C for 2 min. For HA–M2R, the washed beads were incubated in 2× SDS/PAGE sample buffer, supplemented with 50 mM DTT, at room temperature for 90 min. Aliquots of total cell lysates were also treated in similar fashion for determination of total TRPC4–GFP and HA–M2R respectively. The treated samples were then analysed by SDS/PAGE (8% gel) and probed by anti-GFP (Invitrogen, A11122; 1:10000 dilution) and anti-HA (1:100 dilution, Roche, 11867423001) antibodies for TRPC4–GFP and HA–M2R respectively for Western blotting.
Data presentation and statistical analysis
All data are expressed as means ± S.E.M. Statistical significance was determined using unpaired Student's t tests or ANOVA. P values of less than 0.05 are considered statistically significant.
Co-expression of RGS proteins inhibits TRPC4 currents
To test how RGS proteins affect TRPC4 currents, we co-expressed TRPC4–GFP with M2R together with a selected RGS protein in HEK293 cells. TRPC4 currents in response to activation of M2R using a muscarinic receptor agonist, CCh, were recorded by the whole-cell voltage clamp technique. In these cells, an endogenous Gq/11-coupled muscarinic receptor type, probably M3R, is also present and has been shown to facilitate Gi/o-mediated TRPC4 activation . To further help the development of TRPC4 currents, a Cs+-based internal solution was used throughout and the bath was replaced with a Cs+-rich external solution soon after the establishment of whole-cell configuration in the Na+-based normal Tyrode's solution before agonist application. Under these conditions, CCh (100 μM) evoked a large inward current at −60 mV, which typically reached a peak in less than 20 s and then slowly desensitized (Figure 1A). The I–V relationship exhibited an ‘S-shaped’ curve (Figure 1C) with inward currents at negative potentials typically larger than the outward currents at positive potentials and a ‘flat’ or ‘negative slope’ region between 5 and 40 mV, which was probably caused by Mg2+ block as shown for TRPC5 channels . This I–V relationship is typical for homomeric TRPC4 and TRPC5 when these channels are maximally activated [13,29].
RGS4 suppresses TRPC4 currents activated via stimulation of M2R and accelerates current desensitization
As expected from the enhanced GAP activity, which shortens the lifespan of Gα-GTP, co-expression of RGS4 reduced CCh-evoked current density of TRPC4 by ∼70% (Figures 1B and 1D) and accelerated rate of current decline (desensitization) in the continued presence of CCh (Figures 1B and 1E). On the other hand, the co-expression of RGS4 did not significantly alter the surface or total expression of TRPC4 or M2R (Figures 1F–1I), indicating that the decreased current was due to a change in functional coupling. Furthermore, the functional effects were dependent on the type of RGS. RGS proteins have different preferences towards G-protein subtypes. For example, RGS2 prefers Gαq, whereas RGS4 mainly acts at Gαi/o subunits [30,31]. Consistent with the idea that TRPC4 activation is dependent on Gi/o rather than Gq/11 signalling, the co-expression of RGS2 did not significantly alter TRPC4 currents (Figure 1D).
To examine whether the GAP activity of RGS4 was critical for attenuating TRPC4 currents, we used two RGS4 mutants, N88S and L159F, which had been shown previously not to bind to Gαi1 and exhibit only 15% and 17% GAP activity respectively as compared with the wild-type RGS4 . We found that co-expression of RGS4N88S or RGS4L159F with TRPC4 did not alter the amplitude of TRPC4 currents (Figure 1D); however, both mutations appeared to slow down desensitization (Figure 1E), suggesting that they may also be dominant-negative, at least in the context of Gαi/o-mediated TRPC4 activation. These results support the idea that RGS proteins negatively affect TRPC4 channel activities through enhancing the GTPase activity of Gαi/o and switching off their signalling.
To ensure that the observed inhibitory effect was not due to the specific RGS protein (RGS4) or the muscarinic receptor, we also tested the effect of another Gi/o-selective RGS protein, RGS6, on TRPC4 currents evoked through stimulation of the Gi/o-coupled μOR. The use of μOR also allowed for Gi/o stimulation without a concomitant Gq/11 activation when using a μOR-specific agonist, DAMGO. This differs from the use of M2R because the muscarinic agonist CCh could also act at the endogenous Gq/11-coupled muscarinic receptors to cause simultaneous stimulation of both Gi/o and Gq/11 pathways in the M2R-expressing cells . Application of DAMGO (0.5 μM) in the Cs+-rich external solution to HEK293 cells that stably expressed μOR did not induce appreciable current (results not shown). However, transient expression of TRPC4–GFP in these cells allowed for robust current development in response to DAMGO (Figure 2A), with an I–V relationship similar to that seen in M2R/TRPC4 co-expressing cells (Figure 2C). The co-expression of RGS6 in these cells reduced DAMGO-evoked TRPC4 current by ∼80% (Figures 2B and 2D) and accelerated its desensitization in the continued presence of DAMGO (Figures 2B and 2E). Taken together, the above data demonstrate that Gαi/o-selective RGS proteins negatively modulate TRPC4 channel function by terminating Gαi/o signalling through their GAP activities and this may be a general effect unrelated to the receptors involved in Gi/o activation.
RGS6 suppresses TRPC4 currents activated via stimulation of μOR and accelerates current desensitization
Co-expression of GoLoco domain-containing proteins inhibits TRPC4 currents
To test how GoLoco proteins affect TRPC4 currents, we co-expressed LGN or AGS3 with TRPC4 and M2R in HEK293 cells. Both LGN and AGS3 contain four GoLoco motifs, which act as GDI on Gαi/o . The co-expression of LGN led to ∼56% reduction in CCh-induced TRPC4 currents via M2R (Figures 3A, 3B, 3D and 3E). Unlike RGS proteins, LGN did not significantly alter the rate of current desensitization (Figures 3A, 3B and 3F). Similarly, the co-expression of AGS3 also reduced TRPC4 current by ∼45% without affecting the rate of current desensitization (Figures 3C–3F). Importantly, neither protein significantly altered the surface or total expression level of TRPC4 or M2R (Figures 3G–3L), indicating again that the decreased current density was due to a change in functional coupling rather than maturation or trafficking of the expressed channel or receptor. Therefore, the GoLoco proteins also suppress TRPC4 currents, but they differ from the RGS proteins in that they do not affect the rate of current desensitization. This would be consistent with the role of GoLoco domain proteins in reducing the availability of Gαi/o.
GoLoco domain proteins inhibit receptor-activated TRPC4 currents
Noticeably, the degree of suppression on current density appeared to be less with the co-expression of GoLoco proteins than with RGS proteins. Since AGS3 cycles between membrane bound and cytosolic locations , the low efficiency of AGS3 in inhibiting TRPC4 could be due to a low level of AGS3 proteins associated with the plasma membrane. To overcome this potential problem, we used a mutant ASG3 construct that contains a consensus myristoylation sequence at the N-terminus to promote membrane targeting . The shorter variant of AGS3, AGS3sh , which contains only three complete GoLoco motifs and lacks the tetratricopeptide repeat domains, was also used. Although co-expression of AGS3sh with TRPC4 and M2R in HEK293 cells inhibited CCh-induced TRPC4 current by ∼66% (Figures 3D and 3E), introduction of the myristoylation sequence to AGS3sh (Myr-AGS3sh) led to complete suppression of the current (Figures 3D and 3E). Similarly, Myr-AGS3sh also strongly inhibited DAMGO-evoked TRPC4 current by more than 97% (Figure 4). We interpret these as the improved membrane targeting of AGS3 stabilized the interaction between Gαi and AGS3 , which in turn prevented receptor-induced formation of Gαi-GTP and hence TRPC4 activation. As controls, neither AGS3sh nor Myr-AGS3sh affected the surface or total expression level of TRPC4 (Figures 3K and 3L).
Myr-AGS3sh completely inhibits TRPC4 currents evoked by μOR activation
Because the AGS3sh construct lacks the tetratricopeptide repeat domains, which are present in both LGN and AGS3-long [21,34], the above results also argue for the GoLoco motifs and their GDI activities to be responsible for the inhibitory effect on TRPC4 function. Notably, none of the GoLoco protein constructs, when co-expressed with TRPC4 and a Gi/o-coupled receptor, induced any constitutive TRPC4 current. Because GoLoco proteins cause Gβγ release from the heterotrimers without generation of Gαi-GTP, these data also support the previous conclusion that the Gαi/o arm rather than the Gβγ arm of the Gi/o signalling is responsible for TRPC4 activation .
Co-expression of RGS4 and LGN has an additive effect on suppressing TRPC4 currents
RGS and GoLoco proteins affect Gi/o via different mechanisms. However, functional interactions might be possible between the GAP and GDI activities, resulting in the suppression of one activity in the presence of the other. For example, agonist induced decrease in AGS3-Gαi1 interaction was abrogated by the co-expression of RGS4 . To test whether RGS-mediated inhibition of TRPC4 currents was altered by the co-existence of GoLoco proteins or vice versa, we compared DAMGO-evoked currents in the μOR stable cell line that transiently expressed TRPC4–GFP without or with either RGS4 or LGN individually or with both RGS4 and LGN together (Figure 5). When co-expressed individually with TRPC4 in μOR cells, RGS4 and LGN led to partial, but nonetheless significant, inhibition of DAMGO-induced TRPC4 currents via μOR (Figure 5E). As for CCh-evoked TRPC4 currents via M2R, only RGS4, but not LGN, accelerated desensitization of the DAMGO-evoked currents (Figure 5F). Co-expression of RGS4 and LGN together with TRPC4, however, caused stronger inhibition of the DAMGO-induced currents, from ∼72% and ∼63% for RGS4 and LGN alone respectively, to ∼90% when both were present (Figure 5E). These results suggest that the RGS protein and LGN probably acted separately on the G-proteins and the suppression of TRPC4 current was additive.
RGS4 and LGN additively inhibit TRPC4 currents evoked by μOR activation
Supporting the idea that RGS and GoLoco domain proteins act at G-proteins rather than the channel, we found that the co-expression of RGS4, LGN or AGS3 with TRPC4 did not affect the current evoked by the direct TRPC4/C5 agonist, englerin A (Figure 6). Therefore, the RGS and GoLoco domain proteins specifically suppress Gi/o-mediated TRPC4 activation. These data also demonstrate that the co-expression of these proteins did not alter the surface expression of TRPC4, consistent with the assessment by surface biotinylation followed by Western blotting.
RGS and GoLoco domain proteins do not affect TRPC4 currents activated by englerin A
RGS12 and RGS14 completely suppress TRPC4 currents via GAP and GDI activities
The R12 subfamily of RGS proteins (RGS12 and RGS14) contains both RGS and GoLoco [also known as G-protein regulatory (GPR)] motifs in a single polypeptide (Figure 7J). Based on the results with co-expression of RGS4 and LGN, the R12 RGS proteins should be able to inhibit TRPC4 current more effectively. Indeed, co-expression of RGS12 or RGS14 with TRPC4 and M2R completely abolished the CCh-induced TRPC4 currents (Figures 7A–7C and 7H), suggesting that having both the GAP and GDI activities can completely block TRPC4 channel function. Because RGS12 has a more complex domain organization than RGS14 (Figure 7J), we focused on RGS14 to examine whether both the RGS and GoLoco domains of these proteins are involved in the inhibition of TRPC4 activity.
R12 family RGS proteins completely inhibit M2R-mediated TRPC4 activation via RGS and GoLoco domains
We made three mutants of RGS14. E92A/N93A (RGS14EN) disrupts RGS domain and therefore interferes with its GAP activity ; Q516A/R517A (RGS14QR) interrupts the GoLoco-Gα interaction and hence eliminates the GDI activity ; E92A/N93A/Q516A/R517A (RGS14ENQR) combined mutations at both the RGS and GoLoco motifs and therefore is believed to be defective at both the GAP and GDI activities . Co-expression of RGS14EN or RGS14QR with TRPC4 and M2R led to partial suppression of CCh-evoked TRPC4 currents (Figures 7D, 7E and 7G), averaging to ∼66% and ∼77% inhibition for the RGS-null and GoLoco-null mutants respectively (Figure 7H). Consistent with GAP activity being involved in current desensitization, the RGS-null mutant, RGS14EN, markedly attenuated the desensitization of CCh-induced TRPC4 currents (Figures 7D and 7I). However, the GoLoco-null mutant, RGS14QR, with the RGS motif kept intact, significantly accelerated the desensitization (Figure 7I). Importantly, mutations at both RGS and GoLoco domains, RGS14ENQR, completely eliminated the inhibitory effect of RGS14 on CCh-evoked TRPC4 currents via M2R (Figures 7F–7H).
Similar to the CCh-evoked TRPC4 activation via M2R, the DAMGO-evoked activation of TRPC4 via μOR was also completely inhibited by the co-expression of wild-type RGS14 (Figure 8E). Mutation at the GoLoco domain, RGS14QR, partially rescued the suppressed current by ∼35% (Figures 8A, 8B, 8D and 8E), and the desensitization of the currents became faster than the control (Figure 8F). The co-expression of RGS14ENQR failed to decrease the DAMGO-evoked TRPC4 currents (Figures 8C–8E), neither did it affect the rate of current desensitization (Figure 8F). Therefore, the effects of the RGS and GoLoco domains of RGS14 on TRPC4 currents were mediated by Gαi/o proteins endogenously expressed in HEK293 cells rather than the specific receptor used for activating the G-proteins. Taken together, these results suggest that the RGS and GoLoco domains of RGS14 negatively regulate TRPC4 channel activities by acting at Gαi/o proteins.
RGS14 completely inhibits μOR-mediated TRPC4 activation
TRPC channels are commonly thought of as receptor-operated channels activated downstream of either Gq/11/PLCβ or RTK/PLCγ signalling [13,14]. Our recent data, however, suggest that TRPC4 exhibits a unique dependence on PLCδ1, which co-ordinates with Gi/o proteins to elicit TRPC4 currents . The Gq/11/PLCβ and RTK/PLCγ pathways appear to be dispensable, although they do facilitate the activation kinetics in a PLCδ1-dependent manner. On the other hand, the pertussis toxin-sensitive Gi/o proteins are absolutely required for TRPC4 activation . Our data from the present study reveal several important features of Gi/o-mediated TRPC4 activation and how these may be involved in the physiological and pathological functions of TRPC4 channels.
Gαi/o-dependence of TRPC4 function
Previously, TRPC4 was shown to physically interact with Gαi2 via a C-terminal site, suggesting a direct effect by the Gαi/o subunits on TRPC4 activation . Because the activity of Gαi/o proteins is suppressed by GAPs and GDIs, we reasoned that proteins containing these domains would probably have a negative impact on TRPC4 function. Indeed, we show that co-expression of either a RGS or a GoLoco domain protein with TRPC4 led to significant inhibition of TRPC4 currents evoked via stimulation of Gi/o-coupled receptors. The inhibition by RGS was due to its GAP domain as it also accelerated current desensitization, consistent with the increased GTPase activities (Figure 9), and was abolished by introducing mutations in the GAP domains of RGS proteins. On the other hand, the GoLoco domain proteins only caused current inhibition without affecting the desensitization kinetics, which is consistent with a GDI function that reduces Gαi/o availability. In particular, the GoLoco domain sequesters Gαi/o-GDP with concomitant production of free Gβγ dimers [20,21]. For Gβγ-regulated GIRK1/2, the co-expression of LGN increased the basal current , but for TRPC4, neither LGN nor AGS3 caused an increase in basal current, suggesting that the free Gβγ dimers produced by the expression of GoLoco domain proteins do not cause TRPC4 activation. Together with the results that RGS proteins accelerated current desensitization, the overall inhibitory effect of GoLoco proteins lends further support to the notion that the Gα subunits of the Gi/o proteins are responsible for stimulating TRPC4. In intestinal smooth muscle cells, the native muscarinic agonist-activated cation current, shown to be largely composed of TRPC4  and co-dependent on both M2R and M3R , was inhibited by intracellular dialysis of antibodies specific for Gαi3/Gαo or Gαo, but not for Gβ , further arguing for the role of Gαi/o rather than Gβγ in supporting also the native TRPC4 currents.
Schematic diagram of RGS- and GoLoco-mediated inhibition of TRPC4 channels
Although Gαi/o proteins are abundantly expressed in the membrane of neurons , their functions are not completely understood. In addition to inhibiting adenylyl cyclases, Gi/o signalling has been implicated in stimulating GIRK, inhibiting voltage-gated Ca2+ channels and so on. However, many of these functions appear to be mediated by Gβγ , leaving few confirmed functions to the Gαi/o subunits. The activation of TRPC4 represents a novel function of Gαi/o, which, because of the cation permeability of TRPC4 channels, leads to membrane depolarization and hence excitation of the affected neuron. Therefore, through coupling to TRPC4, Gi/o signalling, and the active forms of Gαi/o in particular, can be linked to neuronal excitation, as opposed to the conventional view that Gi/o proteins were mainly associated with inhibitory neurological functions.
Inhibition by RGS proteins
The importance of RGS proteins in G-protein signalling has been increasingly recognized. The GAP function possessed by all RGS proteins serves to shorten the half-life of Gα-GTP and thereby terminate its actions on the effectors. This explains both the overall current reduction and the acceleration of desensitization by the RGS proteins. Importantly, RGS proteins only inhibited TRPC4 currents activated via receptor stimulation, but not those induced by the direct channel agonist englerin A (Figure 6), suggesting that these proteins exert their effect through G-proteins. Moreover, only the Gi/o-targeting RGS4 and RGS6 [31,42,43] were able to inhibit TRPC4 currents (Figures 1, 2 and 9). The Gq-targeting RGS2  did not have any effect (Figure 1D), further supporting the importance of Gi/o proteins in TRPC4 activation.
The effect of RGS proteins on TRPC4 current desensitization is similar to that on GIRK channels [23,34,44]. However, in addition to accelerating deactivation , RGS proteins also increased the activation kinetics of GIRK channels either without changing  or by strongly increasing [23,24] the maximal current amplitude. For TRPC4, we only observed a decreased current amplitude in response to receptor stimulation and no obvious effect on activation kinetics with the overexpression of RGS proteins. This could suggest a specific role of RGS on GIRK activation, perhaps through Gαo as previously shown . Alternatively, because GIRK channels are activated by Gβγ , the modulation of Gαi/o-GTP by RGS might only ‘indirectly’ affect these channels. Particularly, the time lapse between GTP hydrolysis and Gαi/o-GDP association with the free Gβγ, as well as the possible conformational difference between freshly formed Gαβγ trimers and older ones , could all contribute to the changes in GIRK currents in the presence of RGS proteins. By contrast, TRPC4 probably responds to Gαi/o-GTP directly, hence factors that affect heterotrimer reassociation and their ‘readiness’ for dissociation, as suggested previously , may have little impact on the channel activity.
Inhibition by GoLoco domain proteins
The GoLoco domains act as GDI to inhibit GDP release from Gα, but at the same time cause Gβγ dissociation from the heterotrimer, which results in a reduced availability of Gαi/o-GTP but an increased level of free Gβγ in the cell [20,21]. Therefore, the reduction in Gi/o-mediated activation of TRPC4 currents in the presence of the GoLoco proteins LGN and AGS3 is consistent with the Gαi/o-GTP-dependence of TRPC4 activation (Figure 9). The facilitation of this inhibition by membrane targeting of AGS3sh via myristoylation further suggests the dependence on Gαi/o in a membrane-delimited fashion. Importantly, the overexpression of GoLoco domain proteins did not alter cell-surface expression of TRPC4 as assessed biochemically by surface biotinylation and electrophysiologically by englerin A stimulation (Figure 6). The critical involvement of GoLoco motifs in this regulation was also demonstrated with the use of AGS3sh, which contains only the GoLoco motifs, and the Q516A/R717A mutant of RGS14 (RGS14QR), which has a disrupted GDI function (Figures 3, 7 and 8). Importantly, unlike RGS proteins, the GoLoco domain proteins did not alter desensitization kinetics of TRPC4, supporting the idea that they only reduce the availability, not the half-life, of Gαi/o-GTP.
The effects of GoLoco domain proteins on TRPC4 differ markedly from that on GIRK channels. Co-expression of LGN with GIRK1/2 increased the basal GIRK current but reduced the receptor-induced current, whereas suppression of LGN expression in neurons increased excitability under basal conditions . The change in the basal activity is consistent with the activation of GIRK currents by free Gβγ subunits. The decreased response to receptor stimulation was also anticipated from the sequestration of Gαi/o into the GDP-bound form, which will reduce the available heterotrimers to enter the G-protein cycle. Ironically, pipette dialysis of GoLoco domain peptides to cells during whole-cell recordings failed to alter the basal GIRK current and the initial response to receptor stimulation, but consistently dampened the receptor-induced GIRK current in response to repeated stimulation . This suggested that GoLoco domain proteins sequester Gα-GDP mainly from active G-protein cycles. In the co-expression studies, basal turnover of the G-proteins during culture probably allowed the expressed GoLoco proteins to bind to Gα-GDP subunits and render them unavailable to re-enter the cycle. This explains the reduction in receptor-evoked channel activation. The lack of effect of GoLoco domain proteins on TRPC4 basal current contrasts with their effect on GIRK, supporting the notion that Gβγ is not involved in TRPC4 activation.
Fine-tuning TRPC4 function with RGS and GoLoco domain proteins
Our results indicate that, unless specifically membrane targeted, a GoLoco domain protein or a RGS protein with only RGS domain (e.g. RGS4 or RGS6) could only achieve partial inhibition on TRPC4 currents. The co-expression of both GoLoco and RGS proteins exhibited an additive effect (Figure 5). Therefore, it is interesting that RGS12 and RGS14, which contains both RGS and GoLoco domains, nearly completely suppressed receptor-induced TRPC4 activation (Figures 7 and 8). Selective disruption of the RGS and GoLoco domains of RGS14 partially abolished its inhibitory effect, resulting in similar levels of TRPC4 current suppression as when a RGS or GoLoco domain protein was expressed alone (Figures 7 and 8). More importantly, disruption of just the GoLoco domain (RGS14QR) allowed revealing of the RGS effect on current desensitization, reminiscent of the co-expression of RGS4 or RGS6 with the channel. By contrast, RGS14EN, which contains the GoLoco but not the RGS domain, behaved just like LGN or AGS3. Therefore, although both acting through Gαi/o subunits, the RGS and GoLoco domains modulate TRPC4 activity via separate mechanisms and the effects are additive. These results reveal the complexity of receptor-operated TRPC4 activation that allows for fine-tuning of the channel function in native systems, such as the central neurons. By varying the expression and subcellular localization of various RGS and GoLoco domain proteins, the vital functions regulated by TRPC4 channels in the central nervous system can be precisely tuned to fulfil their physiological roles.
Indeed, TRPC4 channels have been shown to play roles in synaptic transmission, neuronal excitability and neurodegeneration [3–6,45,46]. TRPC4 is expressed throughout CA1, CA2, CA3 and dentate gyrus areas of hippocampus . RGS and GoLoco proteins are also naturally present in hippocampal neurons where they may exert effects in fine-tuning TRPC4 function in regional and developmentally regulated fashions. For example, RGS14 expression occurs gradually during development, being largely undetectable until postnatal day 7 and reaching peak level only in adulthood . The period of no or low RGS14 expression coincides with the time of active dendritic growth and branching during early postnatal development. In adult brain, RGS14 is enriched in the CA2 and fasciola cinerea areas of hippocampus with sporadic presence also in CA1 [48,49]. The CA2 area is functionally different from other CA regions in that it typically lacks synaptic long-term potentiation and has a negative impact on hippocampus-related spatial learning and object recognition memory. All of these were reverted by the deletion of RGS14 gene . It would be interesting in future studies to test whether TRPC4 is involved in these functions. The CA2 pyramidal neurons are also more resistant to cell loss associated with temporal lobe epilepsy than CA1 and CA3 neurons [50,51]. This could be, at least in part, due to suppression of TRPC4 function by the high RGS14 levels in the CA2 neurons since TRPC4 is implicated in epilepsy-induced neuronal death [3,4]. Thus, the ability of RGS14 to fine-tune Gi/o-mediated TRPC4 activation and thereby regulate neuronal functions warrants further investigation.
In summary, we show that TRPC4 channels are negatively regulated by RGS and GoLoco domain proteins in line with their established roles in promoting GTPase activity and sequestering Gαi/o respectively. Our data support the notion that the Gα, but not the Gβγ, arm of Gi/o signalling is involved in TRPC4 activation. These findings reveal additional layers of complexity of TRPC4 channel modulation, which have significant implications in the regulation of neurons and other cell types where TRPC4 exerts functions.
Jae-Pyo Jeon, Insuk So and Michael Zhu designed the research; Jae-Pyo Jeon, Dhananjay Thakur and Jin-bin Tian performed the experiments; Jae-Pyo Jeon performed data analysis; Jae-Pyo Jeon and Michael Zhu wrote the paper.
We thank Dr Carmen Dessauer and Dr Stephen Lanier for suggestions on the study, Dr Joe B. Blumer and Dr Stephen Lanier for the AGS3sh and Myr-AGS3sh constructs, and Dr Guangwei Du for providing the PEI transfection reagent.
This work was supported by the National Institutes of Health (NIH) [grant numbers R01 NS092377 and R01 GM092759]; the American Heart Association Southwest Affiliate [grant numbers 15POST22630008 (to J.-P.J.) and 13PRE17200004 (to D.P.T.)]; and Graduate School of Biomedical Sciences, The University of Texas Health Science Center at Houston [Investing in Student Futures Scholarship (to D.P.T.)].
activator of G-protein signalling 3
carbachol or carbamoylcholine
[D-Ala2, N-MePhe4, -Gly-ol]-enkephalin
G-protein-activated inwardly rectifying K+
G-protein signalling modulator
human embryonic kidney
M2 muscarinic receptor
M3 muscarinic receptor
myristoylatable AGS3 short form
physiologically relevant external solution
regulator of G-protein signalling
receptor tyrosine kinase
transient receptor potential canonical