The recently discovered family of regulators of G protein signaling (RGS) accelerates the intrinsic GTPase activity of certain Gα subunits, thereby terminating G protein signaling. Particularly high mRNA levels of one family member, RGS3, are found in the adult kidney. To establish the temporal and spatial renal expression pattern of RGS3, a polyclonal antiserum was raised against the COOH terminus of RGS3. Staining of mouse renal tissue at different gestational stages revealed high levels of RGS3 within the developing and mature tubular epithelial cells. We tested whether RGS3 can modulate tubular migration, an important aspect of tubular development, in response to G protein-mediated signaling. Several mouse intermedullary collecting duct (mIMCD-3) cell lines were generated that expressed RGS3 under the control of an inducible promoter. Lysophosphatidic acid (LPA) is a potent chemoattractant that mediates its effects through heterotrimeric G proteins. We found that induction of RGS3 significantly reduced LPA-mediated cell migration in RGS3-expressing mIMCD-3 clones, whereas chemotaxis induced by hepatocyte growth factor remained unaffected by RGS3. Our findings suggest that RGS3 modulates tubular functions during renal development and in the adult kidney.
- renal tubular epithelial cells
- kidney development
- G protein
- regulators of G protein signaling
rgs proteins belong to a gene family with structural and functional homologies sharing a region of ∼125 amino acids, the regulator of G protein signaling (RGS) domain, which accelerates the intrinsic GTPase activity of certain Gα subunits (reviewed in Refs.3, 19, 20, 32, 33). Members of this protein family have been shown to regulate pheromone signaling in yeast and a variety of G protein-linked signaling pathways in mammalian cells. RGS proteins stabilize activated Gα subunits in a conformation that favors the release of bound GTP (4, 57). In addition, RGS proteins appear to antagonize G protein signaling through a direct interaction with Gα and Gβ subunits (30,54, 61). Currently, ∼20 different RGS proteins are known (reviewed in Ref. 3). RGS proteins are widely expressed and can be found in virtually all organs (6, 15, 16, 22, 29, 34, 51, 52). Only a few family members display a restricted expression pattern; for example, RGS7 is found predominantly in the brain and lung (28, 34), RGS8 is found almost exclusively in the brain (28, 49), and RGS9 is expressed only in the retina and in the brain (28, 29). Although RGS3 is expressed in several organs, including heart, skeletal muscle, lung, liver, and brain (7, 22, 28, 62), RGS3 exceeds the expression of other RGS family members in the kidney (22, 34, 62). In contrast, RGS2, RGS4, RGS5, RGS7, RGS9, RET-RGS1, RGS12, RGS14, and RGS16 are absent or only expressed at low levels in the kidney in comparison with other organs (10, 22, 25, 29, 34, 51, 53, 62).
RGS3 encodes a protein of 519 amino acids with a relative molecular mass of 57 kDa (22) and has been implicated in the regulation of several signaling cascades. Expression of RGS3 impairs interleukin-8 receptor and lysophosphatidic acid (LPA)-mediated ERK1/ERK2 activation (9, 22), modulates the G protein-gated inwardly rectifying potassium channel (21), and inhibits the Gqα-coupled production of inositol phosphate (9, 41) as well as the Gsα-coupled production of cAMP (9), although the mechanism by which RGS3 impairs Gsα-coupled signaling remains unclear. Intrigued by the high expression levels of RGS3 in the kidney, we decided to analyze the spatial and temporal expression of RGS3 during renal development. A polyclonal antiserum raised against the COOH-terminal 100 amino acids of RGS3 revealed a predominantly tubular expression pattern during renal embryogenesis. To examine the effect of RGS3 on tubular epithelial cell motility, a mouse intermedullary collecting duct (mIMCD-3) cell line was generated that expressed RGS3 under the control of an inducible promoter. With the use of this model system, we demonstrated that RGS3 modulates tubular cell migration in response to heterotrimeric G protein signaling.
MATERIALS AND METHODS
Plasmids and reagents. The cDNA of the full-length human RGS3 (kindly provided by K. M. Druey) was fused to a Flag-His tag sequence MLDYKDDDDKHHHHHHHHH and subcloned into CDM8 (F.RGS3). A fragment containing the COOH-terminal 300 bp was generated by PCR and cloned into pMAL-c3 (New England Biolabs, Beverly, MA) to produce the maltose-binding protein (MBP) fusion protein MBP-RGS3421–520. For the generation of an inducible, stable cell line, RGS3 followed by the encephalomyocarditis virus (EMCV) internal ribosome entry site (EMCV-IRES) (kindly provided by P. Sarnow) and the green fluorescent protein (gfp) (kindly provided by B. Seed) were subcloned into the expression vector pIND (Invitrogen, San Diego, CA). A control pIND vector was generated containing the EMCV-IRES followed by gfp, but lacking RGS3. LPA, epidermal growth factor (EGF), angiotensin II, bradykinin, and wortmannin were obtained from Sigma (St. Louis, MO). Pertussis toxin (PTX) was obtained from GIBCO, Life Technologies (Gaithersburg, MD), and L. G. Cantley kindly provided hepatocyte growth factor (HGF). LPA was dissolved in chloroform:methanol:acetic acid (90:5:5) and further diluted in DMEM. HGF was dissolved in DMSO. Solvents alone were used in control experiments at the appropriate final concentrations.
Generation and affinity purification of an antiserum specific for RGS3. Gel-purified MBP-RGS3421–520 was used to immunize rabbits (Cocalico Biologicals, Reamstown, PA) with a standard immunization protocol. For affinity purification of the rabbit antiserum, recombinant MBP-RGS3421–520 was coupled to AminoLink Plus (Pierce, Rockford, IL), and RGS3-specific antiserum was eluted from the AminoLink column with 0.1 M glycine-HCl (pH 3.0) and dialyzed against PBS. The concentration of the affinity-purified RGS3 antiserum was determined by ELISA.
Western blot analysis. Western blot analysis was performed to confirm the specificity of the RGS3 antiserum. Human embryonic kidney cells (HEK 293T) were transiently transfected with plasmids encoding Flag-tagged hRGS3, hRGS7, or vector (pCDM8) with the calcium phosphate method. After 24 h of incubation, cells were harvested and lysed in sample buffer. The samples were fractionated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane (NEN Biological Products, Boston, MA). The membrane was incubated with affinity-purified anti-hRGS3 antiserum, followed by horseradish perioxidase-coupled anti-rabbit IgG (Amersham, Life Science, Arlington Heights, IN) and ECL (SuperSignal, Pierce). Blocking experiments were performed with recombinant MBP-RGS3421–520. The expression of hRGS3 in the inducible mIMCD-3 cell lines was determined at 24 and 56 h after induction with 1 μM muristerone A.
Immunohistochemistry. Mouse (Swiss Webster) fetal kidneys (days E14–E18, newborn and adult) were fixed in 4% paraformaldehyde (Sigma) in phosphate buffer (pH 7.4) for 30–90 min and incubated in 30% sucrose (Acros Organics, NJ) in phosphate buffer (pH 7.4) overnight at 4°C. The tissue was embedded in OCT compound (VWR Scientific, Bridgeport, NJ), and 8-μm sections were prepared at −24°C. The sections were initially incubated in TBS, pH 7.6, and endogenous peroxidase was blocked with 1% H2O2(Sigma) in methanol (Fisher Scientific, Pittsburgh, PA) for 30 min, rinsed in deionized H2O, followed by two washes in TBS for 5 min. For antigen retrieval, sections were heated to 90°C in ACCU-TUF (Accurate Chemicals, Westbury, NY) for 10 min, cooled for 10 min, and washed again. Heat-inactivated normal goat serum (Sigma) was used at 5% in TBS, pH 7.6, for 30 min to block unspecific bindings. Tissue sections were incubated with anti-RGS3 or preimmune rabbit IgG at a concentration of 10 or 20 μg/ml in TBS, pH 7.6, containing 1% normal goat serum for 60 min at room temperature or overnight at 4°C. After two rinses in TBS on an orbital shaker for 8 min, the sections were incubated with an anti-rabbit IgG-biotinylated secondary antibody (Vector Laboratories, Burlingame, CA) at a 1/200 dilution in 50 mM Tris, pH 7.6, for 30 min. Regular ABC Reagent (Vector Laboratories) was added to slides for 30 min. Staining was developed with diaminobenzidine (Research Genetics, Huntsville, AL), and counterstaining was performed with 1× hematoxylin (Harris Formula, Surgipath, Richmond, IL). Sections were dehydrated in 75–100% ethanol, saturated with xylene (Fisher Scientific), and mounted with xylene-based Cytoseal (Stephen’s Scientific, Riverdale, NJ). The same procedure was used to detect RGS3 in the inducible stable cell lines at 24 or 48 h after induction.
Generation of stable mIMCD-3 cell lines. The mIMCD-3 (48) was cultured in DMEM (Bio-Whittaker, Walkersville, MD) supplemented with 8% calf serum (Sigma). To avoid adverse effects by RGS3 during cell propagation and phenotypic changes during selection, an inducible gene expression system for mammalian cells was utilized (Ecdysone System, Invitrogen, San Diego, CA). The pVgRXR vector directing the expression of the heterodimeric ecdysone-retinoid receptor (VgEcR/RXR) was transfected into mIMCD3 cells, followed by a selection in zeocin at 25 μg/ml. Positive clones were identified with a pIND reporter plasmid that mediates the expression of gfp. Expression of GFP was induced by muristerone A (1 μM), a steroid analog that binds to the heterodimeric nuclear VgEcR receptor and mediates the recognition of a modified ecdysone-responsive element. A second transfection introduced the pIND vector directing the expression of RGS3-EMCV-IRES-gfp, or the control pIND plasmid expressing EMCV-IRES-gfp, and was followed by a double selection with zeocin and geneticin (400 μg/ml; GIBCO, Life Technologies). Positive clones were identified by green fluorescence after induction with 1 μM muristerone A.
Chemotaxis assays. The chemotaxis assay was essentially performed as described (17). A modified Boyden chamber (Neuro Probe, Cabin John, MD) with 48 single wells was used to test for gradient-directed cell movement activated by different chemoattractants or control medium (DMEM). Polycarbonate membranes (8-μm pore size; Poretics, Livermore, CA) were coated with rat tail collagen type I (Upstate Biotechnology) and washed in PBS. DMEM containing HGF, LPA, FBS, angiotensin II, bradykinin, EGF, or solvent controls was added to the bottom chamber and overlaid with the membrane. Cells (2 × 104) were added to the upper chamber. Incubation was carried out at 37°C for 4 h. The membranes were then removed, fixed, and stained with Diff-Quik (Baxter Healthcare, Miami, FL). Cells that had passed through the pores of the membrane were counted by light microscopy. The mean value ± SD of each condition (n = 6–12) was determined per square millimeter. PTX or wortmannin was added to the cell culture 1 h before assay and was added at the same concentration to both the upper and the lower chamber during the assay.
Proliferation assays. The mIMCD-3 cell lines were seeded at a density of 12,500 cell/well in 12-well plates and incubated overnight in DMEM containing LPA, EGF, angiotensin II, bradykinin, or HGF or serum at concentrations as indicated. Cells were pulsed with 1 μCi of [3H]thymidine (NEN) per each well. After 24 h, cells were washed with cold PBS, air-dried, and lysed with 200 μl of 1.5 N sodium hydroxide for 30–60 min at 37°C. Samples of 100 μl were added to 5 ml ScintiVerse (Fisher, Fair Lawn, NJ), and counts per minute were measured with a Tri-Carb Liquid Scintillation Analyzer 2200CA (Packard, Meridawn, CT). PTX or wortmannin was added for 18 h before cell harvesting as indicated. In selected experiments, [3H]thymidine incorporation was followed for up to 5 days.
Cell viability. Cell viability was assessed with trypan blue exclusion. Whereas higher pertussis concentrations exerted a toxic effect, >95% of cells remained viable at PTX concentrations <150 ng/ml. No toxic effects were observed for wortmannin at 10 nM.
Statistical analysis. Data are presented as means ± SE and represent at least three independent experiments; statistical significance was determined by the Student’st test.
RGS3 is expressed in the embryonal and adult kidney. Because RGS3 mRNA is expressed at significant levels in the adult kidney (22), we hypothesized that a more precise spatial and temporal definition of the renal RGS3 expression would provide the first clues for its putative function in the kidney. A COOH-terminal fragment of RGS3 fused to the MBP was used to generate an antiserum in rabbits. The specificity of the affinity-purified antiserum was confirmed by Western blot analysis of HEK 293T cells, transiently transfected with constructs expressing RGS3, RGS7, or pCDM8 vector. As demonstrated in Fig.1 A, the anti-MBP-RGS3 antiserum specifically recognized RGS3, but not, for example, RGS7. RGS3 displays an electrophoretic mobility of ∼83 kDa, significantly higher than its predicted molecular mass, which is potentially due to a high percentage of charged amino acids (>30%) and/or extensive posttranslational modification. Pretreatment of the antiserum with MBP-RGS3 completely abolished the detection of RGS3 (Fig. 1 B). With the use of this affinity-purified antiserum, embryonal mouse kidneys at embryonaldays E14 toE18, newborn and adult mouse kidney, were analyzed by immunohistochemistry. RGS3 was detectable in the mouse embryonal kidney as early as E14. Staining was distinct and initially limited to the apical plasma membrane and the brush border of developing tubules and S-shape bodies (Fig.2 A). At day E16 (Fig.2 B), RGS3 expression intensified and a more lateral and basolateral staining of tubular structures became apparent. This staining pattern, involving all tubular structures at later gestational stages, was detectable atday E18 (Fig.2 C), in newborn tissue (Fig.2 D), and in the adult kidney (Fig.2 E). No significant mesangial staining was detectable at later stages of embryonal development or in the adult kidney.
Generation of an RGS3-expressing mIMCD-3 cell line. The predominantly tubular expression pattern of RGS3 prompted us to generate a tubular epithelial cell line that expresses RGS3. To avoid effects that a negative regulator of G protein signaling, such as RGS3, might exhibit on cell growth and cellular phenotype, the inducible ecdysone system was utilized to express RGS3 in a mIMCD-3 cell line in response to muristerone A. Control cell lines lacking RGS3 were generated in a similar fashion. Screening for positive clones was facilitated by inserting thegfp downstream of EMCV-IRES in tandem with the RGS3 cDNA. After induction, ∼95% of cells of an individual RGS3 (Fig. 3,A andB) or control clone (Fig. 3,C andD) became positive for GFP and ∼100% of cells stained positive for RGS3 ∼48 h after induction (Fig. 3, E andF). There was virtually no staining in clones before induction (Fig. 3 E) or in control clones (data not shown). Western blot analysis confirmed the expression of RGS3 protein in the inducible mIMCD-3 cell clones. A specific band for RGS3 was detectable at 24 and 56 h after induction (Fig. 4). RGS3 was not detectable in uninduced cell lines or in cell lines lacking the RGS3 expression cassette.
RGS3 inhibits LPA-activated cell movement in mIMCD-3 cells. An important aspect of tubulogenesis requires the coordinated cellular movement of developing tubular epithelial cells, and a number of hormones and growth factors have been implicated to control this process. We analyzed the effect of RGS3 on the chemotactic behavior of wild-type, control, and RGS3-expressing mIMCD-3 cells exposed to angiotensin II, bradykinin, LPA, EGF, HGF, and serum. All cell lines exhibited a pronounced chemotactic response to LPA, EGF, HGF, and serum but not to different concentrations of angiotensin II or bradykinin (Fig.5 A). The LPA-mediated chemotaxis was concentration dependent, and a maximal response was found at normal LPA serum concentration ranging from 5 to 20 μM (Fig. 5 B); levels over 60 μM inhibited chemotaxis because of toxicity. The migratory responses of two different control mIMCD-3 cell lines (c16 and c18) lacking the RGS3 transgene are shown in Fig. 5 C. Induction of these clones with muristerone A for 48 h had no effect on the chemotaxis in response to serum or LPA. However, the induction of RGS3 expression in the RGS3-expressing mIMCD-3 cell lines (VI7 and VI21) by muristerone A resulted in a significant inhibition of migration in response to LPA (Fig.5 D) (P < 0.001); in a total of 16 experiments, RGS3 reduced the LPA-mediated migration by 44–58%. In addition, a less pronounced but reproducible reduction in migration was also observed in response to the serum; RGS3 expression reduced the number of migrated mIMCD-3 cells by 12–25% (P < 0.05). In comparison, PTX at 100 ng/ml, a commonly used dose (10, 11) that did not affect the viability of the migrating mIMCD-3 cells as assessed by trypan blue exclusion, inhibited both serum and LPA-dependent migration by 83 and 80%, respectively, but had no significant effect on the HGF-mediated chemotaxis (Fig. 5 E). Because PTX impaired LPA-mediated chemotaxis in a dose-dependent manner (Fig.5 F), it is likely that the effects of LPA on mIMCD-3 cell movement are Giα linked. Interestingly, at an intermediate concentration of 40 ng/ml PTX that typically exerts ∼50% inhibition of LPA-dependent migration (Fig.5 F), PTX did not augment the inhibitory effects of RGS3 on LPA- and serum-dependent chemotaxis (Fig.5 G), supporting the notion that both PTX and RGS3 inhibit identical, Giα-dependent signaling.
Effect of RGS3 on DNA synthesis in mIMCD-3 cells. Although RGS3 clearly reduced the migratory responses of mIMCD-3 cells after activation of G protein-linked signaling, we were able to demonstrate that this effect was not due to an unspecific cytopathic effect. We analyzed the proliferative responses of RGS3-expressing cell lines triggered by serum, LPA, or HGF. Exponential incorporation of [3H]thymidine was observed by wild-type mIMCD-3 cells (data not shown), the control mIMCD-3 cells, and RGS3-expressing mIMCD-3 cells over a period of 2–5 days at normal serum concentrations (Fig.6 A) or reduced serum concentrations (Fig.6 B). Whereas the proliferation of mIMCD-3 cells was not affected by LPA at concentrations between 5 and 40 μM (Fig.6 C), HGF mediated a dose-dependent proliferative response of mIMCD-3 cells (Fig.6 C). Consistent with previous reports (6), this proliferative response was inhibited by wortmannin, but not by PTX (Fig. 6 D) or the induction of RGS3 expression (Fig.6 E).
Most RGS proteins belong to a family of GTPase-activating proteins that limit signaling through seven membrane-spanning receptors by binding and inactivating Giα subunits (reviewed in Refs. 3, 19, 20, 32, 33). The expression of RGS3 in the fetal kidney suggests that this protein may regulate important signaling events during renal development. Several signaling cascades, including the prostaglandin, endothelin, and renin-angiotensin system, are linked to seven membrane receptors and their heterotrimeric G proteins and have been implicated in certain aspects of renal development. For example, tubular abnormalities have been demonstrated after inhibition of the renin-angiotensin system by angiotensin-converting enzyme inhibitors (37), and the temporal and spatial expression of kallikrein and bradykinin B2 receptors in the differentiating epithelium of the distal nephron, the site of kinin formation, supports the hypothesis that kinins modulate the maturation of certain nephron segments (23, 24). The importance of prostaglandin signaling during renal development is underlined by the severe renal dysgenesis observed in cyclooxygenase-2-deficient mice (18, 40).
To test whether RGS3 can modulate a complex tubular function, we examined the effects of RGS3 on LPA-mediated chemotaxis of tubular epithelial cells. The phospholipid LPA (1-acyl-glycerol-3-phosphate) activates a specific, seven transmembrane-spanning G protein-coupled receptor leading to remarkably divergent biological responses (reviewed in Refs. 26, 42). LPA triggers cytoskeletal changes through activation of Rho (35, 55, 60) that are accompanied by phosphorylation of FAK (12) and the recruitment of the cytoskeletal proteins talin, vinculin, and paxillin to the newly formed focal adhesions (1). LPA mediates Ras-dependent phosphorylation and activation of the mitogen-activated protein kinases ERK1/ERK2 and is mitogenic for several cells and tissues (8, 13, 14, 46, 59, 60). LPA functions as a growth and survival factor for tubular epithelial cells and glomerular mesangial cells and has been shown to regulate the function of Madin-Darby canine kidney and mesangial cells (2, 27, 31, 36). Finally, LPA induces migration of fibroblasts and epithelial cells that is, at least in part, mediated by a pertussis-sensitive G protein subunit (47, 50, 58). LPA is an intermediate product of the membrane phospholipid metabolism, but can also be produced by activated platelets and other tissues (reviewed in Ref. 36), and is present at a concentration of 5–20 μM in normal serum (38, 39, 42). LPA triggered the chemotaxis of mIMCD-3 cells in a dose-dependent fashion. We found that induction of RGS3 expression in this cell line resulted in a ∼50% inhibition of the LPA-mediated migration, consistent with the extent of other RGS3-mediated inhibitory events (9, 41). The RGS3-dependent effect on mIMCD-3 cell migration was highly specific for the G protein-mediated chemotaxis; other potent chemotactic agents such as HGF and EGF were not affected by the induction of RGS3 expression. Furthermore, RGS3 failed to affect DNA synthesis and proliferation of mIMCD3 cells, indicating that the effect of RGS3 on tubular cell migration was highly selective and not due to an unspecific cytopathic effect. Our results demonstrate that RGS proteins, and RGS3 in particular, may regulate G protein-dependent signaling during renal tubular development. Although abrogation of G protein signaling by deletion of certain Gα subunits has not revealed selective renal abnormalities (43-45), their ubiquitous distribution and participation in a broad range of cellular functions predicts an involvement of G proteins and their regulatory molecules during embryogenesis. Indeed, studies by Stow et al. (56) have demonstrated that Giα family members, susceptible to the activity of RGS3, are present in all tubular segments. Our study provides the first clues that RGS3 may exert a dual role, regulating complex cellular functions such as tubular cell migration during embryogenesis, whereas the regulation of more specialized functions, for example, the modulation of G protein-regulated ion channels, may predominate in the differentiated, mature tubular cell.
Bowman et al. (5) have recently demonstrated that RGS1, RGS3, and RGS4, but not RGS2, can reduce chemotaxis in response to activation of G protein-coupled receptors. Because most RGS family members bind and accelerate the intrinsic GTPase activity of Giα and Gqα subunits, it is not surprising that RGS proteins display overlapping functions. Although it is currently unknown how the specificity and activity of these proteins are controlled, several mechanisms have been proposed (reviewed in Refs. 3, 32). For example, Sst2p, the prototypic yeast RGS, is regulated at the transcriptional level (reviewed in Ref. 19), whereas a restricted temporal and spatial expression appears to curtail the function of other RGS proteins (11, 29). RGS3 is expressed in tubular epithelial cells during renal embryogenesis and in the adult kidney. It is therefore conceivable that this family member modulates the threshold of multiple Giα-dependent signaling events in tubular epithelial cells. Work remains to be done to determine whether other RGS proteins are present in tubular epithelial cells and to delineate their the specific functions in the kidney.
We are grateful to Kate Spokes and Lloyd G. Cantley for experimental advice.
Address for reprint requests and other correspondence: G. Walz, Renal Div., Dept. of Medicine, Beth Israel Deaconess Medical Center, 330 Brookline Ave., Boston, MA 02215 (E-mail:).
This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant RO1-DK-52897 (G. Walz) and a grant by the Polycystic Kidney Research Foundation (G. Walz). W. Grüning and F. Jochimsen were supported by the Deutsche Forschungsgemeinschaft (Gr 1612/1–1 and Jo 270/1–1). E. Kim was supported by National Institute of Mental Health Grant MH-01147.
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. §1734 solely to indicate this fact.
- Copyright © 1999 the American Physiological Society