Ca2+ influx through voltage-gated R-type (CaV2.3) Ca2+ channels is important for hormone and neurotransmitter secretion and other cellular events. Previous studies have shown that CaV2.3 is both inhibited and stimulated through signaling mechanisms coupled to muscarinic ACh receptors. We previously demonstrated that muscarinic stimulation of CaV2.3 is blocked by regulator of G protein signaling (RGS) 2. Here we investigated whether muscarinic inhibition of CaV2.3 is antagonized by RGS3. RGS3 is particularly interesting because it contains a lengthy (∼380 residue) amino-terminal domain of uncertain physiological function. CaV2.3, M2 muscarinic ACh receptors (M2R), and various deletion mutants of RGS3, including its native isoform RGS3T, were expressed in HEK293 cells, and agonist-dependent inhibition of CaV2.3 was quantified using whole cell patch-clamp recordings. Full-length RGS3, RGS3T, and the core domain of RGS3 were equally effective in antagonizing inhibition of CaV2.3 through M2R. These results identify RGS3 and RGS3T as potential physiological regulators of R-type Ca2+ channels. Furthermore, they suggest that the signaling activity of RGS3 is unaffected by its extended amino-terminal domain. Confocal microscopy was used to examine the intracellular locations of four RGS3-enhanced green fluorescent protein fusion proteins. The RGS3 core domain was uniformly distributed throughout both cytoplasm and nucleus. By contrast, full-length RGS3, RGS3T, and the amino-terminal domain of RGS3 were restricted to the cytoplasm. These observations suggest that the amino terminus of RGS3 may serve to confine it to the cytoplasmic compartment where it can interact with cell surface receptors, heterotrimeric G proteins, and other signaling proteins.
- calcium channels
- regulator of G protein signaling proteins
- muscarinic acetylcholine receptors
- enhanced green fluorescent protein-fusion proteins
- voltage-gated R-type calcium channels
native r-type Ca2+ channels have been pharmacologically defined by their resistance to antagonists of other high voltage-activated (L-, N-, and P/Q-type) Ca2+ channels and by their sensitivity to block by SNX-482, a peptide component of spider venom (33, 43, 60). Solid evidence indicates that at least some native R-type Ca2+ channels contain CaV2.3 as their pore-forming and voltage-sensing subunit (26, 40, 48, 53). Recent studies have identified a wide range of physiological roles for CaV2.3 and native R-type currents, including their participation in neurotransmitter and hormone secretion, synaptic plasticity, spatial memory, fear behavior, pain perception, morphine analgesia and tolerance, blood glucose regulation, and cerebellar function (1, 5, 15, 23, 25, 26, 28, 37, 39, 45, 54, 56, 58). Previous studies have also shown that CaV2.3 and native R-type currents are modulated by G protein-coupled signaling pathways (2, 14, 22, 30, 31, 36, 38, 49, 50, 57, 59). Furthermore, the modulation of CaV2.3 by muscarinic ACh receptors is attenuated by RGS2 (30), a regulator of G protein signaling protein known to antagonize signaling by Gqα and Giα proteins (17, 21). A closely related RGS protein, RGS3, also efficiently blunts signaling by Gqα and Giα (47), but whether RGS3 can influence modulation of CaV2.3 has not been previously investigated. RGS3 is particularly interesting because it contains a lengthy (∼380 residue) amino-terminal domain of uncertain cellular function (12).
In the present study, we used electrophysiological methods to investigate whether RGS3 and its shorter splice variant RGS3T (8) can attenuate inhibition of CaV2.3 through M2 muscarinic ACh receptors (M2R). We also used confocal microscopy to examine the subcellular distributions of several RGS3 constructs. Finally, we used site-directed mutagenesis to identify structural regions of RGS3 that determine its subcellular localization. Our findings demonstrate that RGS3 and RGS3T effectively antagonize inhibition of CaV2.3 through M2R. These results identify RGS3 and RGS3T as potential physiological regulators of native R-type Ca2+ channels. Our analysis also indicates that the amino terminus of RGS3 does not interfere with its signaling efficacy. Intriguingly, our mutagenesis study/confocal analysis reveals that the amino terminus of RGS3 plays a critical role in restricting this protein to the cytoplasmic compartment. Together these findings extend our understanding of the cellular biology of RGS3 and its potential role in modulating voltage-dependent Ca2+ influx.
MATERIALS AND METHODS
Cell culture and transfection.
Human embryonic kidney (HEK293) cells were obtained from the American Type Culture Collection (Manassas, VA) and maintained at 37°C in a humidified atmosphere containing 5% CO2. The culture medium contained 90% DMEM (GIBCO-Invitrogen, Grand Island, NY), 10% FBS (GIBCO-Invitrogen), 100 U/ml penicillin, and 100 μg/ml streptomycin. One time per week, cells of low passage number (<20) were replated at low density (∼20–30% coverage) on 35-mm culture dishes and transfected within 3–5 days using CaPO4 precipitation (CellPhect Kit; Amersham Biosciences, Buckinghamshire, UK). The transfection mixture included expression plasmids encoding CaV2.3 (formerly α1E), α2δ, and β3 calcium channel subunits (each at 1.0 μg/dish) and separate plasmids that encoded the M2R (0.05 μg/dish) and enhanced green fluorescent protein (EGFP; 0.1 μg/dish). RGS3-expressing cells were additionally transfected with the aforementioned plasmids (minus EGFP) plus one of the following fusion proteins: EGFP-RGS3, EGFP-RGS3T, RGS3NT-EGFP, or EGFP-RGS3ΔNT (0.5 μg/dish). Later (1 day), transfected cells were briefly trypsinized and replated at low density on 12-mm round glass cover slips. Electrophysiological recordings and confocal observations were performed 24–36 h later. Successfully transfected cells were visually identified by their green fluorescence under ultraviolet illumination. Only green cells were used for confocal or electrophysiological experiments.
Expression plasmids and mutant construction.
Rabbit brain CaV2.3 (α1E; GenBank accession no. X67856) was in pcDNA3.1+ (Invitrogen, Carlsbad, CA). Rat brain α2δ (M86621) was in pMT2 (Genetics Institute, Cambridge, MA). Rabbit brain β3 (X64300) was in pcDNA3 (Invitrogen). Human M2 receptor (X15264) was in pRK5. Jellyfish EGFP (U55763) was in pEGFP (Clontech, Cambridge, UK). Human RGS3 (AF006610), RGS3T (U27655), and the deletion mutant RGS3ΔNT were in pEGFP-C3 (Clontech). The amino-terminal domain of RGS3 (RGS3NT) was constructed using high-fidelity PCR to amplify the appropriate region of RGS3. The amplicon was ligated in pEGFP-N2 and fully sequenced to confirm conservation of the reading frame and absence of unintended mutations. RGS3NT-EGFP corresponds to the initial 313 amino acids of RGS3 fused in-frame (and connected by an 18-residue linker) to EGFP. The conserved core domain of RGS3 (RGS3ΔNT) was constructed using high-fidelity PCR to amplify the appropriate region of RGS3T. The amplicon was ligated into pEGFP-C3 and fully sequenced to confirm conservation of the reading frame and absence of unintended mutations. EGFP-RGS3ΔNT corresponds to EGFP fused in-frame (and connected by a 6-residue linker) to amino acids 378–519 of RGS3.
Large-bore patch pipettes were pulled from 100-μl borosilicate glass micropipettes (World Precision Instruments, Sarasota, FL) and filled with an intracellular solution containing (in mM) 155 CsCl, 10 Cs2-EGTA, 4 MgATP, 0.32 Li-GTP, and 10 HEPES, with pH adjusted to 7.4 with CsOH. Aliquots of pipette solution were stored at −80°C, kept on ice after thawing, and filtered at 0.22 μm immediately before use. Filled pipettes had direct current resistances of 1.0–1.5 MΩ. The bath solution contained (in mM) 140 NaCl, 40 CaCl2, 2 KCl, and 10 HEPES, with pH adjusted to 7.4 with NaOH. After forming a gigaohm seal in the cell-attached configuration, residual pipette capacitance was compensated using the negative capacitance compensation circuit of the Axopatch 200B patch-clamp amplifier (Axon Instruments, Union City, CA). No corrections were made for liquid junction potentials. Ca2+ currents were recorded in the whole cell, ruptured-patch mode. The steady holding potential was −90 mV. Test depolarizations were delivered at 0.1–0.2 Hz; stimulation rate was adjusted for each cell to avoid significant cumulative inactivation. Macroscopic Ca2+ currents were filtered at 5–10 kHz using the built-in Bessel filter (4-pole low-pass) of the amplifier and sampled at 10–40 kHz using a Digidata 1200 analog-to-digital board (Axon Instruments) installed in a personal computer. The pCLAMP software programs Clampex and Clampfit (version 9.2; Axon Instruments) were used for data acquisition and analysis, respectively. Figures 1–4, data fits, and statistical comparisons were performed using the software program Origin (version 6.0 and 7.5; Microcal, Northampton, MA). Linear cell capacitance (C) was determined by integrating the area under the whole cell capacity transient, which was evoked by clamping from −90 to −80 mV with the whole cell capacitance compensation circuit of the amplifier turned off. The average value of C was 16.5 ± 0.7 pF (n = 160 cells). Series resistance (RS) was calculated as τ × (1/C), where τ is the time constant for decay of the whole cell capacity transient. τ and RS were minimized for each cell by using the series resistance compensation circuit of the amplifier. The average values of compensated τ and RS were 42.0 ± 2.0 μs and 2.8 ± 0.1 MΩ, respectively (n = 160 cells). Ca2+ currents were typically evoked by step depolarizations from −90 to +30 mV. The average maximal current, measured at the time of peak inward current for each cell, was 1,544 ± 175 pA, and the corresponding average maximal voltage error was 3.7 ± 0.3 mV (n = 160 cells). The direct current resistance of the whole cell configuration was typically >1 GΩ. Ca2+ currents were corrected for linear capacitance and leakage currents using -P/6 or -P/4 subtraction. All experiments were performed at room temperature (20–23°C).
EGFP-tagged proteins were visualized in living HEK293 cells. Microscopy was performed using a Leica Confocal System model TCS SP2 microscope equipped with a Kr/Ar laser. EGFP fluorescence was examined with a FITC filter and a ×63 water immersion lens. Image analysis was performed using Interactive Data Language (Research Systems, Boulder, CO) software. Briefly, fluorescence intensity profiles corresponding to the equatorial plane in each analyzed cell were constructed by tracing multiple radii from the previously determined geometric center. Each radial fluorescence intensity profile was normalized to its maximum intensity. Next, the normalized fluorescence intensities from all radii of one cell (>200 radii) were averaged to obtain the mean fluorescence profile for each cell. Finally, the fluorescence data from each group of cells were averaged, and the mean radial fluorescence profile was determined. To eliminate any possibility of saturation in the reported confocal images, we determined the percentage of saturated pixels exhibited in each cell image [(no. of saturated pixels/total no. of pixels of the cell image, excluding the background pixels) × 100]. Such analysis indicates that most of the images (79 from 117 total cell images) showed no saturated pixels at all. Moreover, in images that exhibited some saturated pixels (38 from 117 total cell images), the mean percent of saturated pixels was very low (0.17 ± 0.03; minimum value: 0.02% and maximum value: 1.28%).
Results are reported as means ± SE. Means were compared using a one-tailed, unpaired Student’s t-test or one-way ANOVA, as indicated. The Bonferroni correction was applied to successive t-tests following ANOVA (16). Statistical significance was set at P < 0.05. For multiple comparisons, the experimental design corresponds to generalized random blocks, where the blocking criterion was the transfection.
RGS3 attenuates inhibition of CaV2.3 through M2 receptors.
We previously demonstrated that CaV2.3 is both inhibited and stimulated through M2R (31). This dual modulation of CaV2.3 reflects independent signaling events attributable to pertussis toxin (PTX)-sensitive G proteins that mediate inhibition and PTX-insensitive G proteins that produce stimulation of CaV2.3 (31). In the present study, we tested whether RGS3 could attenuate M2 receptor-initiated inhibition of CaV2.3. We focused exclusively on channel inhibition because this form of modulation is considerably less variable than stimulation of CaV2.3 through M2R. Because we used relatively brief (1 min or less) applications of receptor agonist [carbachol (CCh)], the current in most cells exhibited mainly inhibition. Cells in which the expressed CaV2.3 current displayed significant stimulation (>10%) were excluded from analysis.
Whole cell Ca2+ currents were recorded from HEK293 cells coexpressing CaV2.3 and M2R (controls), or these channels and receptors plus RGS3. Figure 1A, top, shows a representative experiment using a control cell. As shown in the plot of current amplitude vs. time, 1-min application of CCh (1 μM) evoked a rapid and sustained inhibition of CaV2.3 current. By contrast, the same concentration of CCh produced significantly less inhibition of the current recorded from an RGS3 cell (Fig. 1A, bottom). On average, 1 μM CCh inhibited the current by 33.0 ± 1.8% (n = 12) in control cells and by 13.5 ± 1.4% (n = 12) in RGS3 cells (P < 0.001). As expected (29), RGS3 also accelerated the recovery of CaV2.3 from inhibition after washout of 1 μM CCh. The time required to reach the 70% of the initial current amplitude after washout of CCh was 13.3 ± 1.6 s for control cells (n = 7) and 3.8 ± 0.9 s in RGS3-expressing cells (n = 6; P < 0.001).
Figure 1B shows the average responses of control and RGS3 cells to various CCh concentrations. Saturating inhibition of CaV2.3 current was produced by ∼1 μM CCh in control cells, whereas a much higher concentration (>50 μM) was required in RGS3 cells. The fit of the average dose-response data to the Hill equation yielded an EC50 of 42 nM for controls and 870 nM for RGS3 cells. By contrast, the calculated Hill coefficients were fairly similar for control (1.1) and RGS3 (0.7) cells. These results clearly demonstrate that RGS3 shifts the dose-response relationship for M2 receptor-mediated inhibition of CaV2.3 currents to higher CCh concentrations.
RGS3 does not alter the voltage dependence of CaV2.3.
It was recently reported that RGS3 shifts the voltage dependence of N-type Ca2+ currents in chick neurons (51). We therefore investigated whether RGS3 could alter the voltage dependence of CaV2.3 currents in our expression system. The current-voltage (I-V) relationship of expressed CaV2.3 currents was measured by delivering a series of depolarizations from the steady holding potential (−90 mV) to test potentials ranging from −40 to +70 mV (Fig. 2A). The averaged, normalized I-V relationships are plotted in Fig. 2B. There were no significant differences in the I-V relationships of control vs. RGS3 cells. As illustrated in Fig. 2C, the voltage dependence of channel activation relationships were similarly indistinguishable; for example, the test potential for half-maximal activation of CaV2.3 current was 20.8 ± 3.8 mV in control cells (n = 9) and 17.7 ± 1.6 mV in RGS3 cells (n = 10; P > 0.05). As shown in Fig. 2D, the prepulse potential that produced half-maximal inactivation was also nearly identical in control (−41.3 ± 6.6 mV; n = 14) and RGS3 (−43.1 ± 8.1 mV; n = 11; P > 0.05) cells. Finally, rates of macroscopic activation, inactivation, and deactivation (measured from currents evoked by steps to +30 mV) were indistinguishable between control and RGS3 cells (data not shown). Altogether, these results indicate that RGS3 does not alter the voltage dependence of CaV2.3 under the conditions of our experiments.
Signaling efficiency of RGS3 is not reduced by its lengthy amino terminus.
RGS proteins increase the intrinsic rate of GTP hydrolysis by Gα subunits and/or act as effector antagonists to block interactions of Gα and Gβγ subunits with downstream effectors (3, 20, 41, 47, 55). The conserved RGS core domain is responsible for the GAP activity and effector antagonist function of RGS proteins (47). However, the extra-core domains of certain RGS proteins have been reported to influence their signaling effectiveness, either by altering the RGS protein’s conformation or its intracellular location (7, 9, 11, 18). To determine whether the amino terminus of RGS3 can influence its signaling efficacy, we compared the abilities of four different RGS3 proteins to attenuate muscarinic inhibition of CaV2.3. Each of these RGS3 constructs was fused in-frame to EGFP. Because we only recorded from green fluorescent cells, we were confident that every cell we examined expressed the transfected RGS3 construct.
We first compared the effects of RGS3 with its naturally occurring variant RGS3T (8), which corresponds to amino acids 314–519 of full-length RGS3 (Fig. 3C). RGS3T has been shown to function as a potent GAP for both Gi/o and Gq/11 proteins (8, 13, 42, 47). Interestingly, RGS3T was previously reported to antagonize Gqα-mediated signaling more effectively than full-length RGS3, suggesting that the lengthy amino terminus of RGS3 may interfere with its signaling activity or alter its intracellular location (47). However, in our experiments, RGS3 and RGS3T proved equally effective in attenuating muscarinic inhibition of CaV2.3 (Fig. 3). Thus, 1 μM CCh inhibited the Ca2+ current by 12.8 ± 1.6% in RGS3T cells (n = 7) and by an indistinguishable amount (15.0 ± 2.4%) in RGS3 cells (n = 7; P > 0.05). In parallel experiments with control cells, 1 μM CCh inhibited the current by 31.7 ± 2.2% (n = 7; P < 0.05; RGS3T vs. control). Similarly, our deletion mutant RGS3ΔNT, which encompasses amino acids 378–519 of RGS3 and therefore corresponds approximately to the conserved RGS core domain, was equally effective in attenuating muscarinic inhibition of CaV2.3. For example, 1 μM CCh inhibited the current by 16.0 ± 2.7% in RGS3ΔNT cells (n = 7), which is not different from the inhibition measured in RGS3T or RGS3 cells (P > 0.05). By contrast, the lengthy amino-terminal domain of RGS3 (construct RGS3NT, corresponding to amino acids 1–313 of RGS3) was completely without effect on M2R-mediated inhibition of CaV2.3. On average, 1 μM CCh inhibited the current by 28.1 ± 2.7% (n = 7) in RGS3NT cells, not different (P > 0.05) from inhibition in control cells. These data indicate that RGS3, RGS3T, and RGS3ΔNT are equally effective in antagonizing muscarinic inhibition of CaV2.3. In addition, these findings suggest that the amino-terminal domain of RGS3 (residues 1–313) does not reduce its signaling efficacy.
RGS3 is localized to the cytoplasm by virtue of its amino terminus.
Previous studies have reported that certain domains of RGS proteins can influence their subcellular distributions (9, 11, 27, 46). We therefore constructed deletion mutants of RGS3 to determine whether its lengthy amino terminus controls its intracellular location. It has been previously determined that an EGFP tag does not affect the intracellular distribution of RGS proteins, regardless of whether the tag is attached to the amino or carboxyl terminus of the RGS protein (9). Therefore, we attached EGFP to the amino termini of RGS3, RGS3T and RGS3ΔNT, and to the carboxyl terminus of RGS3NT, and used confocal microscopy to determine their intracellular distributions. Figure 4 shows characteristic images (Fig. 4A) and respective radial fluorescence profiles (Fig. 4, B and C) of cells expressing RGS3, RGS3T, RGS3NT, or RGS3ΔNT. Confocal images were obtained from living cells transfected with CaV2.3, M2R, and the corresponding fusion protein. Figure 4A shows the typical subcellular distribution patterns observed in each case. In control cells, unfused EGFP displayed a homogenous distribution throughout both cytoplasm and nucleus. By contrast, EGFP-RGS3, EGFP-RGS3T, and RGS3NT-EGFP were localized primarily within the cytoplasm. Importantly, these fusion proteins appeared mostly absent from the nucleus. It should be noted, however, that the intracellular distribution of EGFP-RGS3T was more variable than that of EGFP-RGS3 or RGS3NT-EGFP; this difference is discussed in more detail below (see discussion). Interestingly, EGFP-RGS3ΔNT exhibited a distribution pattern very similar to that of unfused EGFP, i.e., it was uniformly distributed throughout both cytoplasm and nucleus. This latter observation indicates that the RGS3 core domain can enter the nucleus even though it is attached to EGFP. By contrast, the amino-terminal domain of RGS3 appeared to be excluded from the nucleus. Taken altogether, these results suggest that one function of the RGS3 amino terminus is to localize this protein to the cytoplasmic compartment where it would potentially have access to cell surface receptors, G proteins, and other signaling proteins such as CaV2.3.
This study presents two main findings. The first is that RGS3 and its naturally occurring variant RGS3T can effectively attenuate inhibition of CaV2.3 channels through M2R. This finding identifies RGS3 and RGS3T as potential physiological regulators of CaV2.3, and suggests that these two RGS proteins may be involved in fine-tuning some of the physiological processes in which CaV2.3 participates. By reducing the ability of G protein-coupled receptors to modulate Ca2+ influx through CaV2.3, RGS3 proteins could strongly influence such diverse physiological processes as neuronal electrical excitability and insulin secretion. The second main finding is that RGS3 is restricted to the cytoplasmic compartment by virtue of its amino-terminal domain. This second finding provides new insight regarding the potential functional significance of the RGS3 amino terminus.
RGS3 attenuates M2 receptor-mediated muscarinic inhibition of CaV2.3.
In a previous study, we showed that PTX-sensitive G proteins are responsible for the inhibition of CaV2.3 through M2 receptors (31). Here, we demonstrate that RGS3 attenuates muscarinic inhibition of CaV2.3 by shifting the dose-response relationship to higher agonist concentrations (Fig. 1B). This effect of RGS3 is consistent with previous reports that RGS3 antagonizes signaling by PTX-sensitive as well as by PTX-insensitive G proteins (47). In terms of functional significance, the RGS3-induced rightward shift in the dose-response curve means that CaV2.3 will be less sensitive to receptor-mediated modulation over a wide range of agonist concentrations. Thus, in cells where they are coexpressed, RGS3 should greatly reduce the ability of Giα-coupled receptors to inhibit CaV2.3. This action of RGS3 could have a significant impact on physiological processes involving Giα-coupled receptors and Ca2+ influx through CaV2.3 channels.
Human RGS3 does not alter the voltage dependence of rabbit CaV2.3.
Previously, Tosetti et al. (51) reported that long-term overexpression of chicken RGS3 proteins induced a significant shift, to more positive potentials, in the I-V relationship of endogenous N-type Ca2+ currents recorded from chick dorsal root ganglion neurons. In contrast, we found here that expression of human RGS3 in HEK293 cells has no detectable effect on the voltage dependence of CaV2.3 currents (Fig. 2). There are several potential explanations for our differing results. For example, Tosetti et al. (51) examined the effects of RGS3 on N-type currents that were presumably mediated by endogenous chicken CaV2.2 subunits, whereas we measured the effects of RGS3 on R-type currents mediated by rabbit CaV2.3 subunits. However, it should be noted that, in our previous study (29), RGS3T failed to change the voltage dependence of N-type currents produced by expression of rabbit CaV2.2 channels in HEK293 cells. Another difference is that the RGS3 proteins used by Tosetti et al. (RGS3s and RGS3ss; see Ref. 51) of 408 and 283 amino acids in length, respectively, were cloned from chicken tissues and were transfected using an adenovirus system. Furthermore, Tosetti et al. (51) recorded currents at 10 days posttransfection. In contrast, in our experiments, RGS3 (of 519 amino acids in length) was cloned from human tissues and was transfected using CaPO4 precipitation. Potentially, our methodology resulted in lower levels of RGS3 expression. Additionally, we recorded currents at ∼2 days posttransfection.
Amino terminus of RGS3 does not impair its signaling efficacy.
RGS3 belongs to the B/R4 family of RGS proteins (19, 44). Most members of this family, with the exception of RGS3 (70 kDa), are relatively small proteins (20–30 kDa). RGS3 is unusual because it contains a fairly lengthy (∼380 residue) amino-terminal domain of uncertain physiological significance (Fig. 3C and Refs. 12 and 35). In the present study, we found that full-length RGS3, RGS3T, and RGS3ΔNT, which corresponds approximately to the core domain of RGS3, were equally effective in attenuating muscarinic inhibition of CaV2.3 (Fig. 3B). Previously, we demonstrated that RGS3 and RGS3T are equally effective in attenuating muscarinic inhibition of N-type Ca2+ currents mediated by expression of CaV2.2 (29). These results suggest that signaling activity of the RGS3 core domain is unimpeded by the presence of the amino-terminal ∼380 residues. However, this interpretation may only be correct when RGS3 is expressed at relatively high concentrations. In this regard, Niu et al. (35) reported that the amino terminus of RGS3 contains a binding site for 14–3-3 proteins and that the interaction of RGS3 with 14–3-3 reduced the ability of RGS3 to interact with Gqα proteins, but only when RGS3 was not expressed at very high concentrations.
Amino terminus of RGS3 restricts it to the cytoplasmic compartment.
RGS proteins contain a variety of structural motifs that could potentially influence their subcellular localizations (11, 19, 46). For example, nuclear localization sequences have been identified within RGS6, RGS7, RGS9–2, and RGS11 (7, 9); nuclear export sequences have been identified within RGS4 and RGS16 (7, 9); a cytoplasmic retention sequence associated with a cysteine-rich motif has been identified within RGS-GAIP, RGSZ, and Ret-RGS1 (9); and membrane-associating amphipathic α-helixes have been identified within the amino termini of all members of the B/R4 family of RGS proteins, Ret-RGS1, and RGS-GAIP (4, 10). However, the functional motifs present within the amino terminus of RGS3 remain incompletely understood. RGS3 exists as several different variants that have the same conserved RGS domain but different amino and/or carboxyl termini (24, 52). The amino terminus of RGS3 has been the focus of particular attention because of its relative length and complexity (12, 19, 24, 52). In the present study, we explored the function of this region by using confocal imaging to analyze the intracellular distributions of various RGS3-EGFP fusion proteins (Fig. 4). Our findings strongly suggest that the amino terminus of RGS3 plays a critical role in restricting RGS3 to the cytoplasmic compartment. This conclusion is supported by our observations that RGS3, RGS3T, and RGS3NT are found primarily within the cytoplasm, whereas RGS3ΔNT is distributed uniformly throughout both cytoplasm and nucleus.
Our confocal images reveal that RGS3T is located primarily within the cytoplasm and thus has a subcellular distribution basically similar to RGS3 (Fig. 4). However, it should be noted that, whereas RGS3 was unambiguously restricted to the cytoplasm, the distribution of RGS3T was somewhat less distinct. In fact, we did observe some cells in which RGS3T was also present within the nucleus. However, it should be emphasized that, unlike EGFP and RGS3ΔNT, RGS3T was never observed to be uniformly distributed throughout cytoplasm and nucleus; it was always present at a higher concentration in the cytoplasm, even in the cells where it exhibited some limited degree of nuclear localization. These observations suggest that RGS3T is retained within the cytoplasm but is not retained as effectively as RGS3. The finding that RGS3T can enter the nucleus is consistent with a previous report that RGS3T can trigger apoptosis (13).
Taken altogether, our results suggest that sequences within the amino terminus of RGS3 function to retain this protein within the cytoplasm. To identify interaction motifs within this domain, we performed an analysis of the first 378 residues of RGS3 using the ScanSite motif identification program (http://scansite.mit.edu/). This analysis identified two binding sites for 14–3-3 proteins (one at serine-264 and another at serine-365) and two binding sites for SH3 domain-containing proteins (one at proline-171 and another at proline-366). Previously, Niu et al. (35) reported that serine-264 of RGS3 binds to 14–3-3 proteins, which have an established function in retaining proteins within the cytoplasm (32). Niu et al. (35) also found that RGS3 bound at least two additional proteins besides 14–3-3, of molecular weight 100 and 130; however, these other binding partners of RGS3 were not identified. It seems reasonable to conclude that, based on their predominantly cytoplasmic distributions, both RGS3 and RGS3T have interactions that serve to retain them within the cytoplasm. Because RGS3 is longer, it probably has more such interaction sites and is thus more effectively retained within the cytoplasm.
Potential physiological significance.
Our present results identify RGS3 and RGS3T as potential physiological regulators of CaV2.3 channels. Recent studies have found that CaV2.3 is expressed in numerous cell types and is involved in a wide variety of physiological processes (1, 5, 15, 23, 25, 26, 28, 37, 39, 45, 54, 56, 58). RGS3 is also expressed in multiple tissues (12), but its physiological functions have not been extensively characterized. CaV2.3 is known to be coexpressed with RGS3 in specific tissues, including dorsal root ganglion neurons (45, 51) and striatum (34, 6). It is therefore possible that endogenous RGS3 proteins regulate native R-type channels within a physiological context, as suggested by the recent experiments of Tosetti et al. (51).
We have shown that RGS3 and RGS3T can attenuate muscarinic inhibition of CaV2.3 channels. We have also demonstrated that the amino terminus of RGS3 plays a significant role in restricting RGS3 to the cytoplasmic compartment, where it has the potential to interact with and influence signal transduction through G protein-coupled receptors. The new information provided by our work should help in elucidating the cellular functions and physiological significance of RGS3 proteins and CaV2.3 channels.
This work was supported by CONACyT (39865-Q and ER026) and UASLP (C03-FRC-06–7.8 and C06-FAI-03-4.7) grants to U. Meza and Muscular Dystrophy Association Grant MDA3663 to B. Adams. B. Adams was also supported by the Utah Agricultural Experimental Station. C. Toro-Castillo was the recipient of a CONACyT fellowship.
We thank Dr. Sergio Sanchez-Armass and Dr. Roger A. Bannister for comments on the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society