RACK1 regulates angiotensin II-induced contractions of SHR preglomerular vascular smooth muscle cells

Abstract

The preglomerular microcirculation of spontaneously hypertensive rats (SHR) is hypersensitive to angiotensin (ANG) II, and studies have shown that this is likely due to enhanced coincident signaling between G protein subunits α_q (Gα_q; released by ANG II) and βγ (Gβγ; released by G_i-coupled receptors) to active phospholipase C (PLC). Here we investigated the molecular basis for the enhanced coincident signaling between Gβγ and Gα_q in SHR preglomerular vascular smooth muscle cells (PGVSMCs). Because receptor for activated C kinase 1 (RACK1; a scaffolding protein) organizes interactions between Gβγ, Gα_q, and PLC, we included RACK1 in this investigation. Cell fractionation studies demonstrated increased levels of membrane (but not cytosolic) Gβ, Gα_q, PLCβ₃, and RACK1 in SHR PGVSMCs compared with Wistar-Kyoto rat PGVSMCs. In SHR PGVSMCs, coimmunoprecipitation demonstrated RACK1 binding to Gβ and PLCβ₃, but only at cell membranes. Pertussis toxin (which blocks Gβγ) and U73122 (which blocks PLC) reduced membrane RACK1; however, RACK1 knockdown (shRNA) did not affect membrane levels of Gβ, Gα_q, or PLCβ₃. In a novel gel contraction assay, RACK1 knockdown in SHR PGVSMCs attenuated contractions to ANG II and abrogated the ability of neuropeptide Y (which signals via Gβγ) to enhance ANG II-induced contractions. We conclude that in SHR PGVSMCs the enlarged pool of Gβγ and PLCβ₃ recruits RACK1 to membranes and RACK1 then organizes signaling. Consequently, knockdown of RACK1 prevents coincident signaling between ANG II and the G_i pathway. This is the first study to implicate RACK1 in vascular smooth muscle cell contraction and suggests that RACK1 inhibitors could be effective cardiovascular drugs.

primary hypertension is multifactorial in nature; however, increased renal vascular resistance due to functional and anatomic changes in the renal microvasculature is essential in its pathogenesis (2, 21). Spontaneously hypertensive rats (SHR), a well-studied model of primary hypertension, exhibit enhanced renal vasoconstriction in response to angiotensin II (ANG II) compared with normotensive Wistar-Kyoto rats (WKY) (22, 26, 29, 32, 33, 36, 50). This may be attributed to signaling cross talk, or coincident signaling, between the G_q and G_i pathways, converging on phospholipase C (PLC) (18, 22–26, 31, 45). For instance, although ANG II signals via G_q, infusion of a G_i agonist further enhances ANG II-mediated renal vasoconstriction in SHR, but not WKY (18, 25). Moreover, inhibition of G_i signaling with pertussis toxin (PTx) normalizes renovascular responses to ANG II and lowers blood pressure in SHR (18, 22, 26, 31). Finally, inhibition of PLC and signal transduction pathways downstream of PLC inhibits the exaggerated renovascular response to ANG II in SHR kidneys (23–25). While these findings set a foundation for coincident signaling in the context of enhanced renal vasoconstriction, mechanisms governing this phenomenon remain poorly understood.

Past studies suggest exaggerated vasoconstrictive (2, 21, 30, 32, 50) and structural remodeling (30) responses to ANG II in preglomerular blood vessels of SHR compared with WKY. Therefore, it is likely that preglomerular blood vessels are responsible for the overall enhanced renovascular response to ANG II in SHR. Because preglomerular vascular smooth muscle cells (PGVSMCs) are the primary regulators of preglomerular resistance, it is likely that PGVSMCs mediate the enhanced preglomerular and overall renovascular responses to ANG II in SHR.

The purpose of the present study was to investigate the molecular basis for the enhanced coincident signaling between G protein βγ (Gβγ) and Gα_q in SHR PGVSMCs, with particular emphasis on the possible role of receptor for activated C kinase 1 (RACK1). RACK1 is a highly conserved, ubiquitous, and versatile scaffolding protein that binds and organizes complexes of signal transducers, including Gβγ and PLC (1, 4, 10). This is noteworthy, because coincident signaling involving G_q- and G_i-coupled receptors, which release Gα_q and Gβγ, respectively, converges on PLCβ₃ to enhance PLC activity in a synergistic manner (41, 45). Because RACK1 interacts with Gβγ and PLC, it may well be that RACK1 is critical for this type of coincident signaling. These facts compel us to hypothesize that there may be differences in localization of RACK1 or its binding partners in SHR PGVSMCs and that this may explain why coincident signaling involving Gα_q, Gβγ, and PLC is increased in SHR compared with WKY.

METHODS

Animals.

Adult (14- to 16-wk-old) male normotensive WKY and SHR were obtained from Charles River Laboratories (Wilmington, MA). The Institutional Animal Care and Use Committee approved all procedures. The investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (8th ed., 2011).

Peptides and signaling inhibitors.

ANG II and neuropeptide Y_1–36 (NPY) were obtained from American Peptide Company (Sunnyvale, CA), and PTx and U73122 were obtained from Tocris Bioscience (Minneapolis, MN). Because we found that NPY is easily denatured, NPY solutions were made fresh each day and were not subjected to freezing or vigorous shaking.

Isolation and culture of PGVSMCs.

Rats were anesthetized with thiobutabarbital (Inactin, 100 mg/kg ip), and batches of PGVSMCs were isolated from kidneys of two to three rats and cultured using our previously described protocol (39). Briefly, iron oxide particles (5% suspension in DMEM) were infused into the suprarenal aorta of anesthetized rats. The kidneys were harvested and decapsulated, and the cortical tissue was obtained and placed in supplemented DMEM containing antibiotics, fungicidals, and HEPES. After the cortical tissue was minced, it was dispersed using a wire mesh to separate the microvessel-containing fraction from surrounding tissue. The microvessel fraction was repeatedly washed with ice-cold supplemented DMEM; a magnet was used to retain the iron-laden vessels after each wash. The microvessel-containing fraction was gently digested with type I collagenase at 37°C for 30 min and then passed through a 20-gauge hypodermic needle to shear off glomeruli. The arteriolar fraction retained after sieving through an 80-μm mesh was suspended in DMEM supplemented with 20% fetal calf serum, plated, and incubated at 37°C in 5% CO₂-95% air and 98% humidity. The medium was changed every day. In our experience, if PGVSMCs seed the culture plate, they do so in ~5-7 days (and proliferate slowly); however, if fibroblasts seed the culture plate, they do so in ~1-2 days (and proliferate rapidly). Therefore, if cells were detected before day 5, the presumption was that the cells were fibroblasts and the plate was discarded and the process was repeated anew. When cells seeded the plate after day 5 and the morphology of the proliferating cells was diagnostic of vascular smooth muscle cells (VSMCs, i.e., “hill-and-valley”), the cells were allowed to grow to confluence. As an additional precaution against contamination by fibroblasts, the PGVSMCs were repeatedly subjected to selective plating using the technique described by Aviv and co-workers (3) to eliminate residual fibroblast contamination. Batching of cells during isolation was necessary to obtain sufficient numbers of cells to conduct an experiment. Since SHR and WKY are inbred strains, this approach did not generate undue heterogeneity. The cells were passaged by trypsinization and used between passages 2 and 3 for the experiments described below. Animals were euthanized by exsanguination, and the diaphragm was punctured while the animals were anesthetized with Inactin.

In addition to the precautions described above to avoid fibroblast contamination, the purity and morphology of PGVSMCs were determined by immunofluorescence imaging for α-smooth muscle cell actin (catalog no. sc-53142, Santa Cruz Biotechnology, Santa Cruz, Ca) and vimentin (mAb 5741, Cell Signaling Technology, Danvers, MA). Immunofluorescence imaging was accomplished using our previously described method (49). Although α-smooth muscle cell actin is expressed in VSMCs, but not fibroblasts, there is no unique marker for fibroblasts. However, although vimentin is expressed in both fibroblasts and VSMCs (16), in fibroblasts and VSMCs in the proliferative state, vimentin is expressed throughout the cytoplasm, whereas in VSMCs in the contractile state, vimentin is expressed predominantly in the perinuclear region (51). Figure 1 illustrates the staining pattern for SHR PGVSMCs and, for comparison, SHR fibroblasts (obtained from SHR hearts, i.e., a rich source of fibroblasts). As is readily apparent, PGVSMCs stained diffusely for α-smooth muscle cell actin (red); moreover, all nuclei [as detected by blue 4′,6-diaminido-2-phenylindole (DAPI) staining] were surrounded by α-smooth muscle cell actin, suggesting that these were, indeed, pure PGVSMCs. In contrast, SHR fibroblasts demonstrated little, if any, staining for α-smooth muscle cell actin. With regard to vimentin staining (green), fibroblasts demonstrated diffuse staining throughout the cytoplasm, spreading to the cell periphery. In contrast, in SHR PGVSMCs, vimentin staining was mostly perinuclear (i.e., surrounding the DAPI staining).

Fig. 1.

A primary antibody against α-smooth muscle cell actin (α-SMA; red) or vimentin (green) was used to generate immunofluorescence images of spontaneously hypertensive rat (SHR) preglomerular vascular smooth muscle cells (PGVSMCs) and fibroblasts. Nuclei were imaged with 4',6-diamidino-2-phenylindole (DAPI; blue).

Collagen gel contraction experiments.

WKY or SHR PGVSMCs were plated into T-75 flasks and cultured in DMEM/F-12 medium containing 10% fetal bovine serum (FBS), penicillin (20 U/ml), and streptomycin (20 μg/ml) at 37°C with 5% CO₂. For experiments involving RACK1 knockdown, PGVSMCs were transduced with lentivirus-expressing short-hairpin RNA (shRNA) before collagen gel contraction (see below). At confluence (~7 × 10⁶ cells), PGVSMCs were trypsinized and resuspended in a collagen solution provided with the cell contraction assay (Cell Biolabs, San Diego, CA) to 6 × 10⁵ cells per milliliter of collagen gel solution. The cell-collagen suspension (0.5 ml) was plated into each well of a 12-well dish (3 × 10⁵ cells/well) and incubated at 37°C for 1 h to solidify. Complete culture medium (2 ml) was added on top of the gels and incubated under standard tissue culture conditions for 24 h. Thereafter, the growth medium was replaced with serum-free medium for an additional 24 h. In experiments involving signaling inhibitors, PTx was added 24 h before gel release, and U73122 was added 30 min before gel release. The serum-free medium with or without signaling inhibitors was then replaced with serum-free medium containing 1) no agonist, 2) ANG II (100 nmol/l), 3) NPY (100 nmol/l), or 4) ANG II (100 nmol/l) + NPY (100 nmol/l). PTx (100 ng/ml) or U73122 (10 μmol/l) was also added to each of the four treatment solutions in the respective experiments. To initiate contraction, the gels were gently released from the sides and bottom of the culture dish (i.e., were floating free of attachments to the dish) with a sterile spatula and imaged at time 0 (before gel release) and at 3, 6, 24, 48, 72, 96, and 120 h postrelease.

Fractionation of PGVSMCs into cytosolic and membrane isolates.

Membrane and cytosolic fractions were isolated from WKY or SHR PGVSMCs using the Mem-PER Plus Membrane Protein Extraction Kit (Thermo Scientific, Rockford, IL). PGVSMCs were plated into T-75 flasks and maintained in DMEM/F-12 medium containing 10% FBS, penicillin (20 U/ml), and streptomycin (20 μg/ml) under standard culture conditions. For experiments involving signaling inhibitors, PGVSMCs were treated for 24 h before fractionation in serum-free medium with or without PTx (100 ng/ml) or 30 min before fractionation in serum-free medium with or without U73122 (10 μmol/l). For experiments involving RACK1 knockdown, PGVSMCs were transduced with lentivirus-expressing shRNA before fractionation (see below). At confluence (~7 × 10⁶ cells), cells were trypsinized and washed twice with cell wash solution. One-fifth of the cells were removed, centrifuged, and lysed with RIPA buffer (Thermo Scientific) for total cell lysate. The remaining cells were centrifuged, and the supernatant was replaced with permeabilization buffer, vortexed, and incubated at 4°C for 10 min with constant mixing. Permeabilized cells were centrifuged for 15 min at 16,000 g. Supernatant containing cytosolic proteins was collected. The pellet was resuspended in solubilization buffer and incubated at 4°C for 30 min with constant mixing. Solubilized membranes were centrifuged for 15 min at 16,000 g, and supernatant containing membrane proteins was collected. Samples were stored at −80°C until further use.

Western blot analysis of RACK1, Gβ, Gα_q, and PLCβ₃ expression.

Protein concentration in total, cytosolic, and membrane isolates was determined by bicinchoninic acid (BCA) assay (Pierce Biotechnology, Rockford, IL). Then the isolates were diluted with 4× Laemmli sample buffer (Bio-Rad, Hercules, CA) containing 2-mercaptoethanol (Bio-Rad), boiled for 10 min, and cooled on ice for 5 min. Each sample (6 μg of protein per sample) was subjected to SDS-polyacrylamide gel electrophoresis. Samples were loaded along with Precision Plus Protein Kaleidoscope Prestained Standard (Bio-Rad) onto 15-well 4–20% Mini-PROTEAN TGX precast gels (Bio-Rad). Electrophoresis was carried out in premixed Tris-glycine-SDS running buffer (Bio-Rad), and proteins were transferred onto 0.45-μm Amersham Hybond PVDF membranes (GE Healthcare Bio-Sciences, Marlborough, MA). Membranes were blocked in TBS-Tween 20 containing 5% nonfat dry milk for 60 min at room temperature and then incubated overnight at 4°C with anti-RACK1 B-3 mouse monoclonal antibody (1:3,000 dilution; Santa Cruz Biotechnology), anti-Gβ T-20 rabbit polyclonal antibody (1:500 dilution; Santa Cruz Biotechnology), anti-Gα_q E-17 rabbit polyclonal antibody (1:500 dilution; Santa Cruz Biotechnology), anti-PLCβ₃ C-20 rabbit polyclonal antibody (1:750 dilution; Santa Cruz Biotechnology), or anti-β-actin R-22 rabbit polyclonal antibody (1:500 dilution; Santa Cruz Biotechnology). Membranes were washed, incubated with horseradish peroxidase-conjugated goat anti-mouse or horseradish peroxidase-conjugated goat anti-rabbit antibody (Thermo Scientific) for 1 h at room temperature, and developed onto X-ray film using chemiluminescence-based ECL 2 Western Blotting Substrate (Thermo Scientific). Antibodies were removed from the membranes with Restore PLUS Western blot stripping buffer (Thermo Scientific), and the process beginning at blocking was repeated for the remaining primary antibodies. The RACK1 B-3 monoclonal antibody was validated in this study and in our previous study (53) by Western blotting, which demonstrated B-3 antibody knockdown of RACK1 protein in cells treated with RACK1 shRNA. The other antibodies have been frequently used by others [anti-Gβ T-20 rabbit polyclonal antibody by Spiegelberg and Hamm (47), anti-Gα_q E-17 rabbit polyclonal antibody by Zhang et al. (52), and anti-PLCβ₃ C-20 rabbit polyclonal antibody by Diaz Añel (8)].

Coimmunoprecipitation of RACK1 with Gβ, Gα_q, or PLCβ₃.

Coimmunoprecipitation was performed using the Pierce Classic Magnetic IP/Co-IP kit (Thermo Scientific). Prior to immunoprecipitation, SHR PGVSMCs were fractionated into cytosolic and membrane isolates (see above). Protein (~6 μg) from each isolate was set aside for the input. The rest (~640 μg) of the cytosolic and membrane isolates were incubated with or without anti-RACK1 B-3 mouse monoclonal antibody (Santa Cruz Biotechnology) to an antibody-to-isolate ratio of 2:500 (μg:μg) overnight at 4°C with constant mixing to form the immune complex. Protein A/G magnetic bead slurry (35 μl) was washed with IP lysis/wash buffer and collected on a magnetic stand. The antibody-isolate mixture was added to the prewashed magnetic beads and incubated for 1 h at room temperature with constant mixing. Beads were collected on a magnetic stand and washed three times with 500 μl of immunoprecipitation lysis/wash buffer and, finally, with 500 μl of nanopure water. Supernatant was removed, and 100 μl of 1× Laemmli Sample Buffer (Bio-Rad) containing 2-mercaptoethanol (Bio-Rad) were added to the beads, which were incubated at room temperature for 10 min. Supernatants containing immunoprecipitated proteins were collected and stored at −20°C until further use. Western blot analysis for RACK1, Gβ, Gα_q, and PLCβ₃ expression in both input and immunoprecipitation samples was performed as described above.

Production of lentivirus-expressing RACK1 shRNA.

See Zhu et al. (53) for a complete description of our method for generating lentivirus-expressing shRNA.

Knockdown of RACK1.

SHR PGVSMCs were plated into T-75 flasks and maintained in DMEM/F-12 medium containing 10% FBS, penicillin (20 U/ml), and streptomycin (20 μg/ml) under standard culture conditions. Cells reached 60–70% confluence on the following day, and the growth medium was replaced with fresh medium containing 8 μg/ml Polybrene (Santa Cruz Biotechnology) to increase lentivirus transduction efficiency. RACK1 shRNA lentivirus (5 μl) or nontargeting control shRNA lentivirus (5 μl) was pipetted directly into the medium, which was incubated overnight at 37°C. Medium containing lentivirus was replaced with fresh medium and incubated overnight at 37°C. Green fluorescent protein was observed under a fluorescence microscope 48 h posttransduction, confirming expression of lentiviral vector elements. To select for only transduced SHR PGVSMCs, medium was replaced with fresh medium containing puromycin (5 μg/ml) every 24 h for 2 days. Puromycin medium was then replaced with fresh growth medium to allow for recovery and incubated for 2 days at 37°C. Thereafter, the cells were harvested for collagen gel contraction experiments or fractionation experiments (see above).

Statistical analysis.

Single comparisons were conducted with unpaired Student’s t-test. Interactions between ANG II, NPY, and RACK1 shRNA on contractile responses of PGVSMCs were analyzed using three- and two-factor analysis of variance (ANOVA), with specific contrasts using Fisher’s least significant difference test. The criterion for significance was P < 0.05. Values are means ± SE.

RESULTS

Intracellular localization of RACK1, Gβ, Gα_q, and PLCβ₃ in WKY and SHR PGVSMCs.

Because RACK1 is known to interact with Gβγ and PLC and because both Gα_q and Gβγ converge on PLCβ₃, we hypothesized greater membrane localization of these signal transducers in SHR PGVSMCs. G protein β- and γ-subunits exist as a tight, inseparable complex in physiological settings; therefore, for our measurements of the β-subunit, we assumed that it represents the βγ complex. Analysis of RACK1, Gβ, Gα_q, and PLCβ₃ expression in total, cytosolic, and membrane isolates from WKY and SHR PGVSMCs revealed greater RACK1 (P = 0.0040), Gβ (P = 0.0090), Gα_q (P = 0.0463), and PLCβ₃ (P = 0.0033) in the membranes of SHR than WKY PGVSMCs (Fig. 2). Expression of these four proteins in total and cytosolic isolates was not significantly different between WKY and SHR PGVSMCs (Fig. 2). Enhanced membrane localization of Gβγ, Gα_q, and PLCβ₃ represents a plausible molecular mechanism for enhanced coincident signaling in SHR PGVSMCs, considering this is precisely where activity appears to occur. Nevertheless, RACK1 is likely involved by virtue of known interactions with Gβγ and PLC.

Fig. 2.

Membrane distribution of receptor for activated C kinase 1 (RACK1), G protein subunits (Gβ and Gα_q), and phospholipase Cβ₃ (PLCβ₃) is increased in SHR PGVSMCs. Subcellular [cytosolic (Cyto) and membrane (Mem)] fractions of PGVSMCs from Wistar-Kyoto rats (WKY) and SHR were processed on the same gel with equal amounts of protein loaded. A–D: densitometry measurements for RACK1, Gβ, Gα_q, and PLCβ₃. P values compare expression in WKY vs. SHR PGVSMCs (by unpaired Student’s t-test). Values are means ± SE for 3 experiments from different batches of cells in which each batch was isolated from kidneys from 2–3 animals.

Immunoprecipitation with RACK1 antibody to coimmunoprecipitate Gβ, Gα_q, and PLCβ₃ in SHR PGVSMCs.

To test specific interactions between RACK1 and these contractile signal transducers, we coimmunoprecipitated RACK1 with Gβ, Gα_q, and PLCβ₃ in cytosolic and membrane isolates from SHR PGVSMCs. The very low level of expression of RACK1 in WKY PGVSMC membranes precluded comparable experiments in WKY PGVSMCs. G protein subunits are found almost exclusively at the membrane, with very little detectable in the cytosol, and are unlikely to interact with cytosolic RACK1. Moreover, other studies have not shown an interaction between RACK1 and Gα_q. Therefore, we hypothesized that RACK1 would only coimmunoprecipitate with Gβ and PLCβ₃ in the membrane and, possibly, with PLCβ₃ in the cytosol. Analysis of Gβ, Gα_q, and PLCβ₃ expression following RACK1 pull-down confirmed that RACK1 binds Gβγ and PLCβ₃, but not Gα_q, in membrane isolates (Fig. 3); however, RACK1 did not bind Gβγ, Gα_q, or PLCβ₃ in cytosolic isolates (Fig. 3). Importantly, there was little to no detection of Gβ, Gα_q, and PLCβ₃ in samples in which immunoprecipitation was performed without RACK1 antibody (Fig. 3). These data yield conclusive evidence that, in SHR PGVSMCs, RACK1 interacts with Gβγ and PLCβ₃, but only at the membrane of SHR PGVSMCs. The absence of RACK1 binding to Gβ or PLCβ₃ in the cytosol implies that Gβγ or PLCβ₃ already present at the membrane is recruiting RACK1. As a scaffolding protein, RACK1 may be enhancing signaling efficiency between membrane-localized Gβγ and PLCβ₃ to potentiate contractile responses in SHR PGVSMCs and, thus, augmenting renovascular responses in SHR kidneys.

Fig. 3.

Gβ and PLCβ₃ coimmunoprecipitation with RACK1 in membrane isolates of SHR PGVSMCs. Immunoprecipitation was conducted with an antibody to RACK1 or nonimmune serum. Immunoprecipitates, along with 6 μg of input protein from cytosolic and membrane fractions, were analyzed by Western blotting with antibodies to RACK1, Gβ, Gα_q, or PLCβ₃. Results are from experiments with different batches of cells; each batch was isolated from kidneys from 2–3 animals. IB, immunoblot.

Effects of G_i inhibition, PLC inhibition, and RACK1 knockdown on intracellular localization of RACK1, Gβ, Gα_q, and PLCβ₃ in SHR PGVSMCs.

If Gβγ and/or PLCβ₃ is recruiting RACK1 from the cytosol to the membrane, then inhibition of Gβγ or PLCβ₃ should subsequently decrease RACK1 expression in membrane isolates from SHR PGVSMCs. Furthermore, our hypothesis that Gβγ or PLCβ₃ is inducing membrane localization of RACK1, and not vice versa, suggests that inhibition of RACK1 should not affect intracellular localization of Gβγ, PLCβ₃, or Gα_q. Indeed, Gβγ inhibition with PTx decreased membrane expression of Gβ (P = 0.0590, near significant) and RACK1 (P = 0.0180) and did not affect localization of PLCβ₃ or Gα_q (Fig. 4). Moreover, PLC inhibition with U73122 decreased membrane expression of PLCβ₃ (P = 0.0017) and RACK1 (P = 0.0201) and did not affect localization of Gβ (Fig. 5). U73122 treatment also slightly enhanced Gα_q expression in total and membrane fractions, which likely represents a minor compensatory response. Overall, these findings indicate that Gβ and PLCβ₃ individually contribute to recruitment of RACK1 to the membrane. RACK1 shRNA significantly reduced RACK1 (P = 0.0002) in the membrane compared with nontargeting control shRNA and did not alter subcellular distribution of Gβ, Gα_q, or PLCβ₃, as predicted (Fig. 6). This further supports our hypothesis that Gβγ and PLCβ₃ induce RACK1 translocation to the membrane in SHR PGVSMCs. The results described above lend evidence to a signaling paradigm where Gβγ and PLCβ₃ recruit RACK1 to the membrane, which subsequently enhances interaction efficiency between the two. Scaffolding proteins are known to bind multiple members of a particular signaling pathway and organize its components in a manner that regulates downstream signaling. In our case, RACK1 likely mediates the G_i pathway by scaffolding Gβγ and PLCβ₃ to potentiate contractile signaling.

Fig. 4.

Pertussis toxin (PTx) decreases membrane distribution of RACK1 in SHR PGVSMCs. SHR PGVSMCs were incubated for 24 h with serum-free medium or PTx (100 ng/ml). Subcellular fractions were processed on the same gel with equal amounts of protein loaded. A–D: densitometry measurements for RACK1, Gβ, Gα_q, and PLCβ₃. P values (by unpaired Student’s t-test) compare controls with PTx-treated cells. Values are means ± SE for 3 experiments from different batches of cells in which each batch was isolated from kidneys from 2–3 animals.

Fig. 5.

U73122 decreases membrane distribution of RACK1 in SHR PGVSMCs. Cells were incubated for 30 min with serum-free medium or U73122 (10 μmol/l). Subcellular fractions were processed on the same gel with equal amounts of protein loaded. A–D: densitometry measurements (normalized to β-actin) for RACK1, Gβ, Gα_q, and PLCβ₃. P values (by unpaired Student’s t-test) compare controls with U73122-treated cells. Values are means ± SE for 3 experiments with different batches of cells; each batch was isolated from kidneys from 2–3 animals.

Fig. 6.

RACK1 knockdown does not affect membrane distribution of Gβ, Gα_q, or PLCβ₃ in SHR PGVSMCs. Cells were transduced with RACK1 shRNA lentivirus or nontargeting (NT) shRNA lentivirus, puromycin-selected (5 μg/ml) for 2 days, and allowed to recover in complete growth medium for an additional 2 days before fractionation. Subcellular fractions were processed on the same gel with equal amounts of protein loaded. A–D: densitometry measurements for RACK1, Gβ, Gα_q, and PLCβ₃. P values (by unpaired Student’s t-test) compare samples from nontargeting vs. RACK1 shRNA-treated cells. Values are means ± SE for 4 experiments from different batches of cells; each batch was isolated from kidneys from 2–3 animals.

Effects of RACK1 knockdown on the ANG II-NPY interaction on contraction of SHR PGVSMCs.

Inasmuch as RACK1 binds to Gβγ and PLCβ₃ and organizes a signaling complex, we hypothesized that knockdown of RACK1 should abrogate the ability of G_i-coupled agonists such as NPY to enhance the contractile response to ANG II (a Gα_q-coupled agonist). To measure the long-term contractions of PGVSMCs, we developed a collagen gel contraction assay. Although this approach has been successfully used to measure contractility of smooth muscle cells from airways and aortas (11, 37), to our knowledge this approach has not been applied to microvascular smooth muscle cells such as PGVSMCs. We examined gel areas over the course of 5 days after gel release from the sides and bottom of the well in the absence and presence of ANG II (100 nmol/l), NPY (100 nmol/l), or ANG II + NPY. These experiments were performed in cells treated with nontargeting shRNA and cells treated with RACK1 shRNA. As the PGVSMCs contracted, the gel area gradually decreased, and the contractile response at each time point was calculated simply as percent reduction in gel area. To obtain an integrated measure of the sustained contractile response, for each sample we plotted time vs. percent reduction in gel area and calculated the area under the contraction curve using GraphPad Prism (GraphPad Software, San Diego, CA). As shown in Fig. 7, in PGVSMCs treated with nontargeting or RACK1 shRNA, ANG II significantly increased long-term contractions of SHR PGVSMCs. However, the response to ANG II was attenuated in PGVSMCs treated with RACK1 shRNA. Also depicted in Fig. 7, in PGVSMCs treated with nontargeting or RACK1 shRNA, NPY did not affect long-term contractions of SHR PGVSMCs. However, in PGVSMCs treated with nontargeting shRNA, NPY augmented the contractile response to ANG II. In contrast, in PGVSMCs treated with RACK1 shRNA, NPY did not enhance contractile responses to ANG II. To test the statistical significance of these results, we first analyzed the data using a three-factor (RACK1 shRNA, ANG II, and NPY) ANOVA. This analysis (Fig. 7) returned a significant P value (0.0128) for the three-way interaction among RACK1 shRNA, ANG II, and NPY. This suggested that the ability of NPY to enhance responses to ANG II was reduced by RACK1 shRNA and justified the separate analysis of the nontargeting and RACK1 shRNA groups using a two-factor (ANG II vs. NPY) ANOVA. For the nontargeting shRNA groups, this analysis returned a significant P value (0.0371) for the two-way interaction between NPY and ANG II. However, for the RACK1 shRNA groups, this analysis did not reveal an interaction between NPY and ANG II. These statistical results confirm the visual impressions shown in Fig. 7. We also performed gel contraction experiments in WKY PGVSMCs; however, even in naïve WKY PGVSMCs, NPY was unable to enhance responses to ANG II. Therefore, we did not test the effects of RACK1 knockdown on the NPY-ANG II interaction, since there was no interaction to inhibit.

Fig. 7.

RACK1 knockdown attenuates contractile responses to angiotensin II (ANG II) and abolishes contractile interaction between ANG II and neuropeptide Y_1–36 (NPY). SHR PGVSMCs were incorporated into a collagen gel, and contraction of the gel upon release was quantified as the area under the contractile curve (AUCC). Some SHR PGVSMCs were pretreated with a nontargeting shRNA, and others were treated with RACK1 shRNA. Cells were also treated with vehicle (Basal), ANG II (100 nmol/l), NPY (100 nmol/l), or NPY + ANG II. Results were analyzed initially using 3-factor ANOVA. Because 3-way interaction was highly significant, data were disaggregated and analyzed separately by 2-factor ANOVA. These analyses revealed a significant NPY-ANG II interaction in PGVSMCs treated with nontargeting shRNA, but not in PGVSMCs treated with targeting shRNA. S, statistically significant for specific contrasts (Fisher’s least significance difference tests). Values are means ± SE for 3 experiments with different batches of cells; each batch was isolated from kidneys from 2–3 animals.

DISCUSSION

The pathophysiology of hypertension in SHR likely involves an enhanced renovascular sensitivity to ANG II (see Ref. 32 for summary of evidence). Because the mechanistic details underlying the augmented renal sensitivity to ANG II in SHR remain unclear, our overall goal is to better understand why the SHR renal vasculature is more responsive to ANG II.

ANG II can act through two receptor subtypes: AT₁R and AT₂R. Selective inhibition of both receptors demonstrates that renal vasoconstrictor responses to ANG II are mediated primarily by the AT₁R. This concept is supported by other studies showing that the AT₁R is the predominant receptor subtype expressed in both human and rat kidneys (32, 44). While the AT₁R is primarily known for its vasoconstrictive properties through G_q-mediated signaling, it also couples to G_i (20). It is well established that G_i signaling enhances G_q-mediated PLC activity, intracellular Ca²⁺ increase, and protein kinase C (PKC) activity, all of which are crucial to regulating contraction (45). Both Gβγ and Gα_q can activate PLCβ₃ in a synergistic and cooperative manner (41). Philip et al. observed that the PLCβ₃ response to Gβγ and Gα_q was 10-fold greater than that predicted for the sum of activities elicited by either subunit alone (41). Like all heterotrimeric G proteins, G_q family heterotrimers can release Gβγ; however, only G_i family heterotrimers are expressed and release Gβγ in sufficient quantities for signaling events (7, 42). As such, G_q-G_i coincident signaling involves synergistic activation of PLCβ₃ by Gα_q from the G_q heterotrimer and Gβγ from the G_i heterotrimer.

Our prior studies involving the G_i inhibitor PTx and the G_i agonist UK-14,304 demonstrated the importance of G_i signaling in ANG II-mediated renal vasoconstriction. PTx treatment normalized renovascular responses to ANG II in SHR, whereas UK-14,304 further enhanced renovascular responses to ANG II (18, 25, 26). However, neither PTx nor UK-14,304 affected ANG II-mediated responses in WKY. Consistent with the hypothesis that G_q-G_i coincident signaling converges on PLCβ₃, pretreatment with the PLC inhibitor U73122 almost completely blocked ANG II- and UK-14,304-induced changes in perfusion pressure in isolated SHR kidneys (23). While these data show that enhanced renovascular responses to ANG II in SHR are likely due to G_q-G_i coincident signaling, why only SHR, but not WKY, are sensitive to the effects of G_i signaling on ANG II-mediated renal vasoconstriction remained unclear.

In the present study we sought to determine a molecular cause for the disparate G_q-G_i coincident signaling in the renal vasculature of SHR, leading to an enhanced vasoconstrictive interaction between ANG II and G_i-coupled agonists. RACK1 is a highly conserved member of the tryptophan-aspartate (WD) repeat family and is characterized by a seven-bladed β-propeller structure (1). It was first described by Mochly-Rosen et al., who demonstrated the existence of intracellular receptor proteins that bound activated PKC and named them receptors for activated C kinase (RACKs) (38). However, emerging evidence from the past several decades has cemented the role of RACK1 as a scaffolding protein that brings proteins and their binding partners in close proximity to enhance signaling efficiency (1, 46). Our initial work in studying RACK1 in PGVSMCs was with respect to its role in cell proliferation. We implicated RACK1 as a positive regulator of PGVSMC proliferation, likely by its interaction with the inositol-1,4,5-trisphosphate receptor and enhancement of Ca²⁺ release from the sarcoplasmic reticulum (5). Furthermore, we found that Y₁ receptor (G_i-coupled) agonists, such as NPY and peptide YY, stimulated SHR more than WKY PGVSMC proliferation, likely due to enhanced membrane localization of RACK1 and subsequent organization of the Gβγ/PLC/PKC pathway (6, 27). Because Ca²⁺ is a crucial element in smooth muscle contraction and because G_i signaling is required to augment ANG II-mediated renal vasoconstriction in SHR, we hypothesized that RACK1 enhances not only proliferation, but also contraction, of SHR PGVSMCs.

While VSMCs have long been thought to adopt one of two phenotypes, contractile (high contractility, low proliferation) or synthetic (low contractility, high proliferation), it is now well recognized that the two phenotypes are not mutually exclusive (40, 43). Moreover, several studies have shown significant heterogeneity in VSMC populations in a given blood vessel (17, 19). Considering that increased vascular remodeling and vascular tone are features common to hypertensive disease, we surmise that SHR PGVSMCs can demonstrate increased proliferation and enhanced contractile G_q-G_i coincident signaling and that RACK1 regulates both (15). This should come as no surprise, given the remarkably large number of proteins known to interact with RACK1 (1, 46). Specifically, we hypothesized that differences in RACK1 localization between WKY and SHR PGVSMCs could, in part, explain the sensitivity of SHR to the effects of G_i signaling on ANG II-induced renovascular responses.

In the present study, fractionation of both WKY and SHR PGVSMCs demonstrated significantly enhanced RACK1 localization to the membrane of SHR PGVSMCs. However, Gβ, Gα_q, and PLCβ₃ were also enhanced in SHR PGVSMC membranes. RACK1 shares significant sequence homology with Gβ, another member of the WD40 repeat family, and WD40 proteins often form homo- or heterodimers with each other (4). One study demonstrated that the RACK1-Gβ heterodimer has the ability to stabilize signaling complexes at the membrane of rat whole brain homogenates (48). Another found that RACK1 binds PLC, an important substrate of Gβ, but no study has shown a RACK1-Gα_q interaction (10). Therefore, the significance of membrane RACK1 localization in SHR PGVSMCs is likely to scaffold and enhance interaction efficiency between the large pool of Gβγ and PLCβ₃ already present there and, subsequently, augment contractile responses from within the G_i arm of the G_q-G_i coincident signaling paradigm.

Our coimmunoprecipitation results confirmed RACK1 binding to both Gβ and PLCβ₃ in the membrane fraction of SHR PGVSMCs. Lack of binding in the cytosolic fraction suggests that 1) RACK1 interacts with Gβγ and PLCβ₃ only at the cell membrane and 2) membrane-localized Gβγ and PLCβ₃ recruit RACK1 from the cytosol, rather than vice versa. Evidence to support the latter was shown with fractionation experiments in SHR PGVSMCs pretreated with PTx or U73122, where inhibition of Gβγ or PLCβ₃ reduced RACK1 translocation to the membrane. It is conceivable that Gβγ, a membrane-anchored protein with little to no cytosolic presence, may be recruiting PLCβ₃ first and then RACK1. This would coincide with findings that both PTx and U73122 inhibit RACK1 membrane localization, but this is unlikely, as PTx did not affect PLCβ₃ distribution. Thus, Gβγ and PLCβ₃ individually contribute to RACK1 translocation to the cell membrane. Importantly, RACK1 knockdown with shRNA did not affect Gβ or PLCβ₃ localization, further supporting these concepts. Because Gβ and PLCβ₃ levels are significantly greater in membrane isolates from SHR than WKY PGVSMCs, it follows that RACK1 would also be increased. Why Gβγ and PLCβ₃ are enhanced in SHR PGVSMC membranes is beyond the scope of the current study; however, there is likely a genetic component, as distribution of these proteins is nearly identical between SHR PGVSMCs and SHR aortic VSMCs, as we found in preliminary studies, and likely permeates beyond VSMC types, as RACK1 is also enhanced in membranes of SHR cardiac fibroblasts (53).

In the collagen gel contraction experiments, we selected NPY to stimulate the G_i pathway because 1) NPY is a cotransmitter in the renal sympathetic nerves and is, therefore, directly released onto PGVSMCs (9), 2) G_i-coupled Y₁ receptors are, indeed, expressed in preglomerular microvessels (13), 3) exogenous NPY augments ANG II-induced renal vasoconstriction in SHR, but not WKY, kidneys (13), and 4) endogenous NPY released from renal sympathetic nerves augments ANG II-induced renal vasoconstriction in SHR, but not WKY, kidneys (12). Together, these published findings indicate that NPY represents an important endogenous regulator of ANG II-induced renovascular responses.

Perhaps the most important finding in the current study was that RACK1 knockdown both attenuated contractile responses of SHR PGVSMCs to ANG II and abolished the ability of NPY to augment ANG II-induced contractions. These results implicate RACK1 not only as a signaling component of ANG II-induced contractions, but also as a participant in the enhancement of ANG II-induced contractions via the G_i signaling pathway. This is consistent with our hypothesis that RACK1 amplifies signaling by directly binding Gβγ and PLCβ₃. Inasmuch as Gα_q directly binds to PLCβ_3, RACK1 represents the epicenter of a signaling complex (RACK1/Gβγ/PLCβ₃-Gα_q/PKC) that augments the interaction efficiency of four key signaling proteins (i.e., Gβγ, PLCβ₃, Gα_q, and PKC). This hypothesis predicts that RACK1 augments basal contractile responses to ANG II by organizing a complex of PLCβ₃-Gα_q/PKC and, when Gβγ is available, further augments ANG II-induced contraction by facilitating the interaction of Gβγ with the PLCβ₃-Gα_q/PKC complex. This explains why RACK1 knockdown attenuates responses to ANG II and prevents the augmentation of ANG II-induced responses by G_q-G_i coincident signaling.

In our collagen gel contraction experiments, NPY did not induce contractions. This is exactly what one would expect, because low concentrations of NPY do not contract the renal vasculature (12) and high concentrations only transiently contract the renal vasculature (14). So NPY does not contract (in a sustained fashion) the renal vasculature but does augment the effects of other vasoconstrictors, particularly ANG II.

Interestingly, in our collagen gel contraction experiments, we observed that the gel contracts under basal conditions, even in the absence of agonists. PGVSMCs in the preglomerular microcirculation respond to stretch by generating myogenic tone (35), and this myogenic response contributes to autoregulation of renal blood flow and attenuates barotrauma in the glomerular capillaries (35). It is likely that when PGVSMCs are incorporated into the gel, some PGVSMCs are stretched as the collagen gel attaches to the sides of the dish and the gel cross-links. This would generate myogenic tone, which results in basal contraction when the gel is released from the sides and bottom of the dish.

A caveat of the present study is that nontargeting shRNA also reduced RACK1 expression in PGVSMCs. This is likely due to the procedure (i.e., lentiviral infection with puromycin selection). Nonetheless, the comparisons reported here remain valid and meaningful, because RACK1 expression was decreased more by RACK1 shRNA than by nontargeting shRNA. However, the effects of RACK1 knockdown on contractions are likely even larger than revealed in this study, because the amount of RACK1 in cells treated with the nontargeting vector was reduced (thus attenuating the response in the control cells). RACK1 knockout is embryonically lethal; however, progress could be made by developing a smooth muscle-specific knockout in the SHR background.

In summary, PGVSMCs are likely responsible for the enhanced ability of G_i-coupled agonists to enhance ANG II-induced renal vasoconstriction previously observed in SHR kidneys. This is due to more efficient G_q-G_i coincident signaling mediated by crucial signal transducers, such as Gα_q, Gβγ, and PLCβ₃, which we found in greater amounts at the cell membrane of SHR PGVSMCs. While the convergence of Gα_q and Gβγ onto PLCβ₃ to potentiate contractile signaling has been established, a novel finding in the present study describes a membrane complex consisting of Gβγ, RACK1, and PLCβ₃. In scaffolding Gβγ and PLCβ₃, RACK1 increases the efficiency of G_q-G_i coincident signaling, leading to a greater ability of G_i agonists to enhance ANG II-induced contraction of SHR PGVSMCs. Therefore, enhanced membrane localization of RACK1 secondary to recruitment by the large pool of Gβγ and PLCβ₃ in SHR PGVSMC membranes can, in part, explain the respective sensitivity and insensitivity of SHR and WKY to G_i signaling on ANG II-mediated renovascular responses. Interestingly, effects of differential RACK1 localization on contraction are echoed in another study, which found that diminished RACK1 in membranes of aging rat hearts resulted in impaired contractile responses (28). Thus the importance of membrane-localized RACK1 in contraction is not limited to PGVSMCs. Our past studies implicated RACK1 in enhanced proliferation of SHR PGVSMCs, and the present study demonstrates that RACK1 augments G_q-G_i coincident signaling in SHR PGVSMCs, leading to a greater effect of G_i-coupled agonists on ANG II-induced contractions of SHR PGVSMCs. Because both proliferation and contraction of PGVSMCs would be expected to increase renal vascular resistance, impair renal hemodynamics, and raise systemic blood pressure, RACK1 may be a crucial element in the pathogenesis of primary hypertension. In fact, the role of RACK1 in hypertension may be multifaceted, because very recent findings show that RACK1 interacts with and serves as a coactivator of the mineralocorticoid receptor (34). Thus RACK1 is an attractive inhibitory target for cardiovascular and renal diseases, particularly in disease states in which membrane-localized RACK1 is elevated.

GRANTS

This work was supported by National Institutes of Health Grants DK-091190, HL-069846, DK-068575, HL-109002, and DK-079307. X. Zhu is a Howard Hughes Medical Institute Medical Research Fellow.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

X.Z. and E.K.J. conceived and designed research; X.Z. performed experiments; X.Z. and E.K.J. analyzed data; X.Z. and E.K.J. interpreted results of experiments; X.Z. and E.K.J. prepared figures; X.Z. and E.K.J. drafted manuscript; X.Z. and E.K.J. edited and revised manuscript; X.Z. and E.K.J. approved final version of manuscript.