Cell Physiology


We characterized the role of guanine nucleotide dissociation inhibitor (GDI) in RhoA/Rho-kinase-mediated Ca2+ sensitization of smooth muscle. Endogenous contents (∼2–4 μM) of RhoA and RhoGDI were near stoichiometric, whereas a supraphysiological GDI concentration was required to relax Ca2+ sensitization of force by GTP and guanosine 5′-O-(3-thiotriphosphate) (GTPγS). GDI also inhibited Ca2+ sensitization by GTP · G14V RhoA, by α-adrenergic and muscarinic agonists, and extracted RhoA from membranes. GTPγS translocated Rho-kinase to a Triton X-114-extractable membrane fraction. GTP · G14V RhoA complexed with GDI also induced Ca2+ sensitization, probably through in vivo dissociation of GTP · RhoA from the complex, because it was reversed by addition of excess GDI. GDI did not inhibit Ca2+ sensitization by phorbol ester. Constitutively active Cdc42 and Rac1 inhibited Ca2+ sensitization by GTP · G14V RhoA. We conclude that 1) the most likely in vivo function of GDI is to prevent perpetual “recycling” of GDP · RhoA to GTP · RhoA; 2) nucleotide exchange (GTP for GDP) on complexed GDP · RhoA/GDI can precede translocation of RhoA to the membrane; 3) activation of Rho-kinase exposes a hydrophobic domain; and 4) Cdc42 and Rac1 can inhibit Ca2+ sensitization by activated GTP · RhoA.

  • calcium sensitization
  • Cdc42
  • Rac
  • RhoGDI
  • Y-27632
  • smooth muscle

the small GTPase RhoA modulates, through its effector(s), the Ca2+sensitivity of smooth muscle contraction and nonmuscle cell motility, mediating these, and other, important physiological mechanisms (24, 28, 37, 41, 57, 60; reviewed in Refs. 26, 35,51, 55, 63). Smooth muscle and nonmuscle myosin II ATPases are physiologically activated by actin upon phosphorylation of the serine 19 residue of their regulatory light chain (RLC) by a Ca2+-calmodulin-activated myosin light chain kinase (MLCK) and are inactivated by dephosphorylation by a Ca2+-independent myosin light chain phosphatase (MLCP; reviewed in Refs. 20 and 54). Therefore, inhibition of MLCP can increase phosphorylation of the RLC of smooth and nonmuscle myosin II and, consequently, force and motility, without a necessary increase in intracellular Ca2+ concentration ([Ca2+]i) (55, 56; reviewed in Refs.20 and 54). Such inhibition of MLCP can be effected by activation of certain G protein-coupled receptors and, in permeabilized preparations, by guanosine 5′-O-(3-thiotriphosphate) (GTPγS) (30, 53) through activation of RhoA (Ref.16 and references therein) and its effector, Rho-kinase (and some other serine/threonine kinases; seediscussion). Rho-kinase phosphorylates the regulatory subunit of the trimeric MLCP and inhibits its catalytic activity, causing an increase in RLC phosphorylation and contraction (11, 27, 28; reviewed in Refs. 20 and 55).

Guanine nucleotide dissociation inhibitor (RhoGDI; henceforth referred to as GDI) is an important component of signaling through Rho subfamily proteins. Known functions of GDI include 1) inhibition of both guanine nucleotide dissociation from and hydrolysis by Rho proteins; 2) complexation with Rho proteins that are hydrophobic, as the result of their prenylated (geranylgeranylated) COOH termini, thereby maintaining a large fraction of these proteins as soluble cytosolic and inactive Rho (Rac, Cdc42)/GDI complexes; and3) extraction of Rho protein from cell membranes. GDI can complex with both the GTP- and GDP-bound forms of the Rho subfamily proteins (RhoA, Cdc42, Rac1) (18, 21, 42, 46) and interacts with them through their highly conserved GDI binding surfaces (36). When activated, either indirectly, by agonists acting on surface membrane receptors, or more directly, by GTPγS, in permeabilized smooth muscle, RhoA dissociates from its cytosolic complex with GDI and translocates to the surface membrane (13,14). Activation of RhoA by upstream trimeric G proteins (Gα13) is mediated by guanine nucleotide exchange factors (GEF) such as p115 RhoGEF (19), a member of the GEF family containing Dbl and pleckstrin homology domains (reviewed in Ref.8). The cytosolic RhoA/GDI complex contains GDP as the bound nucleotide, whereas activation of most Ras family, including Rho subfamily, proteins requires exchange to the GTP-bound form (reviewed in Ref. 3). It has not been previously determined whether, in cells, nucleotide exchange (replacement of GDP by GTP) precedes translocation and occurs in the cytosol or follows dissociation of Rho from GDI and occurs at the cell membrane. Similarly, although considerable information is available about the properties of GDI in solution (reviewed in Ref. 43), little is known about details of its physiological mechanism of action and function in cells.

We have used highly purified recombinant RhoA/GDI complexes containing either GTPγS, GTP, or GDP to determine whether it is sufficient to exchange nucleotide on the complex to enable it to Ca2+sensitize smooth muscle. We also have determined the exogenous concentrations of GDI required for inhibiting Ca2+sensitization by GTP or its nonhydrolyzable analog, GTPγS, and compared it with endogenous cellular RhoA and GDI contents. We present an experimental model indicating that prior formation of the GTP · RhoA/GDI complex is sufficient for subsequent translocation of GTP · Rho to the membrane and activation of Ca2+ sensitization. We further suggest that the physiological regulatory role of GDI is to prevent the recycling of non-GDI-complexed GDP · RhoA (and possibly other Rho family members) generated by GTP hydrolysis to active GTP · RhoA, rather than to extract GTP-bound RhoA from the membrane, although such extraction is possible at unphysiologically high GDI-to-RhoA stoichiometry. We also found that the time course of the response to GDI indicates close temporal coupling of RhoA to the downstream mediators of Ca2+ sensitization.

A preliminary report of some of these findings has been published (45a).


Smooth muscle preparations.

For measurement of isometric tension, small strips (200 μm wide, 3 mm long) of rabbit portal vein and ileum longitudinal smooth muscle were dissected, and isometric tension was measured as published previously (29, 31). Muscle strips were permeabilized by incubation with β-escin (75 μM) for 35 min at 24°C, conditions that retain agonist-induced Ca2+-sensitization pathways. Force was expressed as a percentage of the maximal Ca2+-induced contraction obtained in permeabilized tissue at the end of the experiment. Agonists were generally added to the strips at supramaximal concentrations to minimize diffusional delays.

Downregulation of G protein-mediated Ca2+ sensitization was carried out as described previously (15). Briefly, rabbit portal vein strips permeabilized with α-toxin were incubated with pCa 6.7 (control) or in the same solution containing GTPγS (50 μM, 5–6 h). After incubation, the strips were washed in Ca2+-free solution twice for a total of ∼5 min and then further permeabilized with β-escin (50 μM, 20 min) in Ca2+-free solution to allow penetration of proteins. G14V RhoA was added after the response to pCa 6.5 reached steady state to determine whether it could rescue the downregulated Ca2+sensitization.

Statistical analysis.

Results are expressed as means ± SE obtained from nexperiments. Statistical analysis was performed using paired or unpaired Student's t-test where appropriate.

Western blot protocols.

Tissue homogenates and recombinant Rho protein/GDI complex standards were submitted to SDS-PAGE, and the proteins were transferred to polyvinylidene difluoride (PVDF) membranes. The membranes were blocked with 5% nonfat dry milk in phosphate-buffered saline containing 0.05% Tween 20 (PBST) for 1 h and then incubated with primary antibody for 1 h at 37°C. The following dilutions of primary antibodies were used: anti-RhoA (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:3,000, anti-GDI (Santa Cruz) at 1:5,000, anti-Rac1 (Santa Cruz) at 1:500, anti-Rho-kinase (Transduction Laboratories, Lexington, KY) at 1:500, anti-Ras (Transduction Laboratories) at 1:500, and anti-PAK (Santa Cruz) at 1:2,000. The blots were washed in PBST, incubated in secondary antibody for 1 h at 37°C, and detected with enhanced chemiluminescence (ECL; Amersham). The protein signals were quantitated by densitometry using a Bio-Rad GS-670 imaging densitometer.

Quantitation of Rho proteins and GDI in rabbit smooth muscle.

Rabbit portal vein and the longitudinal smooth muscle layer of ileum were carefully dissected, and excess connective tissue was removed. The tissue samples were weighed and homogenized in buffer containing 1% sodium dodecyl sulfate (SDS) and 10 mM dithiothreitol (DTT). The homogenized tissues were centrifuged at 800 g for 5 min at 4°C to sediment nonsolubilized material. Protein concentrations were measured with a modified Lowry protein assay (Bio-Rad) with bovine serum albumin (BSA) standards. A serial dilution of each homogenate was used for quantitation, over the linear range, on Western blots utilizing as standards known quantities of purified recombinant Rho proteins or GDI that were transferred on the same PVDF membranes. Comparisons of RhoA, GDI, and Rho-kinase contents in intact and β-escin-permeabilized muscles were done on size-matched strips of muscle.

Coimmunoprecipitation of RhoA with GDI from smooth muscle cytosol.

The longitudinal layer of ileum smooth muscle was homogenized in ice-cold homogenization buffer [25 mM Tris · HCl, pH 7.4, 5 mM MgCl2, 150 mM NaCl, and protease inhibitor cocktail (P-8340; Sigma, St. Louis, MO)] and centrifuged at 200,000g for 20 min at 4°C. The supernatant was collected as the cytosolic fraction, and the protein concentration was determined. Cytosolic proteins (100 μg) were precleared with protein A-agarose (Santa Cruz) and then incubated with 2 μg of an anti-GDI antibody that recognizes the NH2 terminus of GDI (A-20 anti-GDI; Santa Cruz) for 2 h at room temperature. Protein A-agarose was added and incubated for an additional 2 h at room temperature. The immunoprecipitate was collected by centrifugation at 2,500 gfor 5 min at 4°C and washed twice with homogenization buffer. The supernatant and immunoprecipitate were made up in Laemmli sample buffer, and a RhoA Western blot was performed to determine whether the RhoA coimmunoprecipitated with GDI.

Determination of the subcellular distribution of RhoA and Rho-kinase.

A minimum of 10 pooled rabbit portal vein strips per point (dimensions given in Smooth muscle preparation) were used for determining the subcellular distribution of RhoA and Rho-kinase. Tissues were permeabilized with α-toxin as previously described (15), relaxed in Ca2+-free solution, and stimulated with 50 μM GTPγS for 1 or 20 min; they were then homogenized in ice-cold NaCl buffer (10 mM Tris · HCl, pH 7.5, 5 mM MgCl2, 2 mM EDTA, and 100 mM NaCl), with the following protease inhibitors and reducing agent: 1 mM 4-(2-aminoethyl) benzonesulfonyl fluoride, 20 μg/ml leupeptin, 20 μg/ml aprotinin, and 1 mM DTT. The homogenate was centrifuged at 800 g, and the supernatant was centrifuged at 100,000 g for 30 min at 4°C (Optima TLX ultracentrifuge, TLA 120.1 rotor; Beckman Instruments); this supernatant was collected as the cytosolic fraction and the 100,000 g pellet as the membrane fraction. RhoA and Rho-kinase were detected by Western blotting. As a control, 90% of total tissue lactic dehydrogenase (a cytosolic marker) was in the cytosolic fraction.

Phase separation by Triton X-114.

Because Rho-kinase in membrane fractions might be associated with cytoskeletal proteins that may cosediment with membrane lipids, we analyzed the distribution of Rho-kinase in membrane fractions using Triton X-114. This zwitterionic detergent separates proteins containing hydrophobic domains from hydrophilic proteins (5). For this purpose, α-toxin-permeabilized portal vein tissues were incubated with or without 50 μM GTPγS in pCa 6.5 solution (20 min) and then homogenized in a buffer containing 50 mM Tris · HCl, pH 7.5, 5 mM MgCl2, 2 mM EDTA, 1 mM DTT, 150 mM NaCl, and protease inhibitors. The membrane fraction (100,000 gpellet, as described above) was resuspended in the same buffer containing 2% Triton X-114 (Sigma) and kept on ice for 30 min with occasional mixing. The samples were then warmed to 37°C for 5 min, and the micelles formed were pelleted by centrifugation at 37°C (5 min, 800 g). The upper fraction containing hydrosoluble proteins and the pellet containing the hydrophobic proteins were separated, mixed with 2× sample buffer, run on SDS-PAGE, and transferred to PVDF membranes. Membranes were immunoblotted for RhoA, Rho-kinase, and GDI and detected by chemiluminescence.

Extraction of RhoA by GDI from smooth muscle membranes.

Rabbit bladder was dissected from connective tissue and homogenized in 25 mM Tris, pH 8.0, 100 mM NaCl, 5 mM MgCl2, 100 μM phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin, and 2 μM pepstatin A with a polytron and centrifuged for 5 min at 800g to pellet unbroken cells and cellular debris. For nucleotide exchange, the supernatant was removed and incubated at room temperature for 45 min with 10 mM EDTA and 500 μM GTPγS. The membrane fraction was pelleted at 240,000 g for 1 h and resuspended in 25 mM Tris, pH 8.0, 100 mM NaCl, 5 mM MgCl2, 10 mM EDTA, and 100 mM GTPγS, and a Bio-Rad protein assay was performed. Mg2+ was added to a concentration of 13 mM, and aliquots of 100 μg of membrane protein were incubated with increasing concentrations of recombinant His6 GDI in a final volume of 60 μl for 10 min at room temperature (46). Quantitative RhoA Western blots using known quantities of purified recombinant RhoA/GDI complex as a standard were performed to quantitate the RhoA in the assay. The membranes were pelleted by ultracentrifugation, washed, and resuspended in 2× Laemmli sample buffer. Western blots were used to determine the amount of RhoA, Ras, and Rho-kinase in the membrane fractions.

In a second approach, to further verify that GTPγS was bound to RhoA extracted from membranes of activated smooth muscle, α-toxin-permeabilized portal vein tissues were stimulated for 20 min with 50 μM GTPγS in the presence of 500 μCi [35S]GTPγS (NEN) to label activated RhoA. After homogenization of the tissues, the membranes were isolated, resuspended in homogenization buffer, and incubated or not with 30 μM GDI for 20 min at room temperature. The membranes were rapidly centrifuged at 200,000 g for 10 min, and the supernatants were immediately applied to a mini-Q column (Smart System; Pharmacia). The proteins were eluted over a period of 30 min with a NaCl gradient (0–0.8 M) in 25 mM Tris buffer (pH 7.5), 5 mM MgCl2, and 1 mM DTT. The fractions were counted for radioactivity and immunoblotted for RhoA and GDI.

Expression and purification of recombinant His6 GDI.

An amino-terminal His6 affinity tag was added by ligating the GDI cDNA into a modified pET22B vector with a replaced polylinker from vector pFastBac-HTa (Life Technologies), the resultant plasmid was used to transform Escherichia coli strain BL21(DE3), and His6 GDI was expressed and purified as previously described (52). The purified GDI was concentrated to 30–40 mg/ml with a MICROSEP microconcentrator (Pall Filtron). Protein concentrations were determined using the Bradford protein assay (Bio-Rad) with BSA standards. Because imidazole interfered with force measurements in β-escin-permeabilized smooth muscle, the GDI was buffer-exchanged into PIPES buffer (30 mM 1,4-piperazinediethanesulfonic acid, pH 7.1, 5 mM magnesium methanesulfonate, and 165 mM potassium methanesulfonate) using a Sephadex G-25 superfine column (Pharmacia Biotech) and then reconcentrated to 30–40 mg/ml. This second concentrator filtrate was used as a buffer control for force measurement in β-escin-permeabilized smooth muscle.

Expression and purification of recombinant His6 G14V RhoA/FLAG-GDI complex.

Construction of human Rho protein and GDI expression plasmids for yeast were previously described (46, 47). Coexpression of His6-prenylated G14V RhoA and FLAG-GDI inSaccharomyces cerevisiae and subsequent His6G14V RhoA/FLAG-GDI complex purification from the yeast cytosolic fraction were previously described in detail (46). Coexpression yields milligram quantities of pure complex with stoichiometrically bound nucleotide and circumvents the low yield and handling difficulties of singly expressed prenylated RhoA. The purified complex was GDP-to-GTP nucleotide-exchanged using a standard nucleotide exchange protocol (46) and then buffer-exchanged into PIPES buffer using a Sephadex G-50 centrifuge column (45). HPLC analysis as previously described (46) revealed that 50–70% of the complex bound GTP, with the remainder binding GDP.

As a control for testing heterodimer stability after GDP-to-GTP nucleotide exchange and buffer exchange into PIPES buffer, 500 μg of GTP-bound complex were run on a Superdex-75 16/60 gel-filtration column (Pharmacia Biotech) equilibrated with PIPES buffer and eluted in this same buffer at 0.4 ml/min. The protein eluted at the same time as a GDP · RhoA/GDI complex control, and both eluted 8 min earlier than control His6 GDI. A silver-stained SDS-PAGE gel of the GTP · RhoA/GDI elution peak revealed equimolar amounts of RhoA and GDI. Therefore, the heterodimer remained complexed after nucleotide exchange to GTP and buffer exchange into PIPES buffer (data not shown).

His6 RhoA/FLAG-GDI was similarly expressed and purified and then used as a standard for the quantitation of RhoA and GDI in rabbit portal vein and ileum.

Purification of posttranslationally modified His6G14V RhoA, His6 G12V Cdc42, and His6 Q61L Rac1 from yeast membranes.

SY1 yeast was transformed with YEpPGAL-His6 G14V RhoA-tPMA1. The yeast were cultured with galactose induction of Rho protein expression and lysed, and the cytosol and membranes were fractionated as previously described (46). The membrane fractions from an 18 L culture were resuspended in Tris buffer (25 mM Tris base, pH 8.0, 100 mM NaCl, and 5 mM MgCl2) with 50 μM GTP. The membrane was solubilized with 1.2% CHAPS {3-[(3-chloramidopropyl) dimethylammonio]-1-propanesulfonate} with a detergent-to-protein ratio of 5:1 for 1 h at 4°C. The solubilized membrane fraction was then ultracentrifuged at 240,000 g for 1 h at 4°C to sediment nonsolubilized membrane fraction components. The supernatant was diluted with Tris buffer to reduce the CHAPS concentration to 0.3%, and 50 μM GTP was added. The sample was incubated with 5 ml of pre-equilibrated TALON (Clontech Laboratories, Palo Alto, CA) metal chelate affinity resin for 45 min at room temperature. The resin was packed into a chromatography column and washed extensively, first with resuspension buffer with 0.3% CHAPS and then with Tris buffer with 5 mM imidazole. The bound protein was eluted with Tris buffer with 50 μM GTP, 100 mM imidazole, and 0.1% CHAPS and then concentrated in a 15-ml Amicon (10 MWCO) centrifuge concentrator to <1 ml. The purified constitutively active GTP · Rho proteins were buffer-exchanged into PIPES buffer and reconcentrated, and the GTP · Rho protein concentration was quantitated by determination of bound nucleotide. The protein was used for determining its effects on force developed by the β-escin-permeabilized rabbit smooth muscle.


Quantitation of RhoA and GDI in rabbit ileum and portal vein smooth muscle.

To assess the concentrations of physiologically relevant, endogenous proteins involved in Ca2+ sensitization, we determined the amounts of RhoA and GDI in intact portal vein and ileum longitudinal smooth muscle by quantitative Western blotting using serial dilution of known concentrations of purified recombinant Rho and GDI proteins to generate a standard curve. The quantity of RhoA and GDI in the tissues, interpolated from the standard curves, were 151 ± 17 and 147 ± 12 pg/μg total proteins for RhoA and 390 ± 140 and 254 ± 64 pg/ml total proteins for GDI in portal vein and ileum, respectively (n = 3). Equivalent cellular concentrations were calculated (1.8 and 1.7 μM for RhoA and 4.2 and 2.7 μM for GDI in portal vein and ileum, respectively) on the basis of the assumption that 80% of tissue weight is water and that RhoA and GDI molecular weights are 21.4 and 23.2 kDa, respectively.

We also determined the amount of endogenous GDI, RhoA, Rho-kinase, and p21-activated kinase (PAK) lost from the β-escin-permeabilized smooth muscles used for functional studies. Quantitation of these proteins was based on the protein content of strips of identical dimensions from β-escin-permeabilized muscle and intact muscle, rather than on total protein, which could change with permeabilization. The contents of RhoA, GDI, Rho-kinase, and PAK in β-escin-permeabilized portal vein strips were 69 ± 2, 63 ± 5, 95 ± 3, and 92 ± 18.7% of the contents of the respective intact (nonpermeabilized) strips (n = 4), indicating that with the permeabilization protocol used in the present study (75 μM β-escin, 35 min at 24°C), only a modest amount of endogenous RhoA and GDI was lost over the ∼45 min following the beginning of permeabilization, whereas losses of Rho-kinase and PAK were negligible.

Translocation of Rho-kinase to the membrane fraction and Triton X-114 partition after stimulation with GTPγS in α-toxin-permeabilized rabbit portal vein smooth muscle.

After stimulation with GTPγS, Rho-kinase translocated from the cytosolic to membrane fractions (Fig.1 A) with a significant amount in the hydrophobic Triton X-114-soluble fraction (Fig. 1 B). As a control, RhoA was immunoblotted on the same PVDF membrane used for detecting Rho-kinase. As reported previously (14), the majority of RhoA in unstimulated smooth muscles is cytosolic and decreased within 1 min after stimulation with GTPγS, while it increased in the hydrophobic Triton X-114-extractable membrane fraction (Fig. 1 B, pellet). GDI, as also found previously, was recovered in the cytosolic fraction, and its distribution was not affected by the presence of nucleotide. This result also indicated that membrane fractions did not contain detectable amounts of cytosoluble proteins.

Fig. 1.

Guanosine 5′-O-(3-thiotriphosphate) (GTPγS)-induced translocation of Rho-kinase. Rabbit portal vein smooth muscle strips permeabilized with α-toxin were incubated in submaximally activating [Ca2+] (pCa 6.7) for 10 min, and then GTPγS (50 μM) was added. A: the strips were homogenized in ice-cold homogenization buffer before (control) or 1, 5, and 20 min after GTPγS was added. Separation of the cytosolic and particulate fractions and detection of Rho-kinase were carried out as described in methods. Results are means ± SE for 4–6 experiments *P < 0.05, **P < 0.01.B: α-toxin-permeabilized portal vein tissues were stimulated or not with 50 μM GTPγS and then homogenized. Membrane (Par) and cytosolic (Cyt) fractions were separated by ultracentrifugation, and the hydrophobic proteins were extracted from the particulate fraction by resuspending the pellet in homogenization buffer containing Triton X-114 (2%). Phase separation was performed at 37°C for 5 min, and low-speed centrifugation was used to separate the micelle pellet (P). The cytosolic fraction and both the pellet (P) and supernatant (SN) of the Triton X-114 extraction were immunoblotted for Rho-kinase (ROKα), RhoA, and guanine nucleotide dissociation inhibitor (GDI). Note that nearly all RhoA and a significant amount of Rho-kinase are translocated by GTPγS into the Triton X-114-extractable fraction (P). These results are representative of 3 independent experiments.

GDI extracts RhoA, but not Rho-kinase or Ras, from membrane fraction and complexes all detectable cytosolic RhoA.

The concentration of RhoA in GTPγS-stimulated rabbit bladder membrane preparations used for the GDI extraction assay was 70–80 nM. After the membranes were incubated in buffer containing submicromolar Mg2+ with excess GTPγS to nucleotide exchange the RhoA, addition of recombinant GDI extracted RhoA from the membranes in a concentration-dependent manner. GDI (0.50 μM) extracted approximately half of the RhoA in the membrane fraction, and increasing GDI concentration above 5 μM caused no significant further extraction. Thus extraction of GTPγS · RhoA from the membrane fraction required at least an ∼10-fold molar excess of GDI. A small amount of RhoA could not be extracted from the membranes by even 50 μM GDI. GDI did not extract Rho-kinase or Ras from the membranes (n= 3; Fig. 2 A). All the detectable cytosolic RhoA was coimmunoprecipitated with GDI from rabbit smooth muscle cytosol fraction with a rabbit polyclonal IgG antibody that recognizes the NH2 terminus of GDI (amino acids 2–21) as shown in Fig. 2 B.

Fig. 2.

GDI extracts RhoA, but not Rho-kinase or Ras, from membranes, and all detectable cytosolic RhoA is complexed with GDI.A: whole cell lysate was incubated with 500 μM GTPγS, and rabbit bladder membrane fractions were prepared as described inmethods. The membranes were incubated in submicromolar Mg2+ with 100 μM GTPγS to nucleotide exchange the RhoA. Recombinant GDI extracted the RhoA from the membranes in a concentration-dependent manner. GDI did not extract Ras or Rho-kinase (ROK) from the membranes (n = 3). B: Western blot of RhoA detected. Lane 1, total cytosolic extract;lane 2, supernatant of sample precleared with protein A;lane 3, pellet of sample precleared with protein A;lane 4, supernatant of sample immunoprecipitated with anti-GDI antibody (see methods) and protein A; lane 5, pellet of sample immunoprecipitated with anti-GDI and protein A. Note that all RhoA present in smooth muscle cytosol was immunoprecipitated with GDI, indicating that there is little or no free RhoA in the cytosol.

In permeabilized smooth muscle stimulated with [35S]GTPγS, GDI also significantly reduced the total RhoA content of the membrane fraction; this was accompanied by a concomitant increase of RhoA in the supernatant compared with control. When the supernatants obtained from the control and GDI-treated membranes were fractionated on a mini-Q column (Fig.3 A), the fraction that contained GDI (fraction 9) also showed the presence of RhoA (Fig. 3 B) as well as an increase in radioactivity (Fig.3 A, inset), while in controls, RhoA was not detected in the supernatant.

Fig. 3.

Extraction of membrane-associated GTPγS-RhoA by GDI. α-Toxin-permeabilized portal vein tissues were stimulated with 50 μM GTPγS (20 min) in the presence of [35S]GTPγS and then homogenized. The membrane fraction was separated and resuspended in buffer containing 30 μM GDI (see methods). After 20 min, the membranes were sedimented and the supernatant was passed on a mini-Q anion exchange column and eluted with a NaCl gradient (0–80%). A: chromatogram obtained with the supernatant of membrane fraction in control and GDI-treated samples.Inset: radioactivity detected in fraction 9 from control (C) and GDI-treated samples (n = 3, *P < 0.05). B: the indicated peak fractions were screened by Western blot for GDI and RhoA. The fraction containing GDI also contained RhoA, with [35S]GTPγS demonstrating that GDI is capable of extracting membrane-associated GTPγS-RhoA.

Inhibition by GDI and Y-27632 of Ca2+sensitization induced by constitutively active recombinant prenylated GTP · G14V RhoA.

The constitutively active mutant GTP · G14V RhoA expressed in yeast caused significant contraction of β-escin-permeabilized rabbit portal vein strips at constant free [Ca2+], as was previously seen with GTP · G14V RhoA expressed in a baculovirus/Sf9 system (16).

Recombinant GDI (20 μM) relaxed GTP · G14V RhoA-induced Ca2+ sensitization from 19.5 ± 3.1% (n = 5) to 2.1 ± 2.3% of maximal force (n = 5, P < 0.001; Fig.4, B and C). In time-matched controls, the filtrate of the final concentration step after PIPES buffer exchange of GDI (see methods) caused no detectable relaxation. The half-time of GDI-induced relaxation was 1.00 ± 0.22 min.

Fig. 4.

Relaxation of GTP · G14V RhoA-induced Ca2+ sensitization by the Rho-kinase inhibitor Y-27632 and by GDI. A: yeast-expressed, constitutively active mutant GTP · G14V RhoA (1 μM) caused significant further increase in the steady-state contraction induced by submaximal Ca2+(pCa 6.5) in β-escin (75 μM, 35 min)-permeabilized rabbit portal vein. The GTP · G14V RhoA-induced contraction was completely relaxed by Y-27632 (10 μM, n = 5). B: the GTP · G14V RhoA-induced contraction was also relaxed by GDI (20 μM, n = 3). C: summary of the experiments in A and B. The 0 force level represents the force at pCa 6.5. ***P < 0.001.

GTP · G14V RhoA-induced contractions were also relaxed by a selective Rho-kinase inhibitor, Y-27632, which inhibits agonist-induced Ca2+ sensitization in smooth muscle (12, 60), from control 19.5 ± 3.1% (n = 3) to −0.7 ± 1.2% of maximal force (n = 3, P < 0.01; Fig. 4, A and C). This result verifies the untested, although reasonable, assumption that recombinant GTP-bound RhoA induced Ca2+ sensitization of smooth muscle by activating Rho-kinase. We have previously shown that Ca2+sensitization of force by GTP · G14V RhoA is accompanied by increased, and relaxation of Ca2+-sensitized force by decreased, myosin RLC phosphorylation (12, 16).

GDI inhibits phenylephrine- and carbachol-induced, but not phorbol ester-induced, Ca2+ sensitization of rabbit portal vein and ileum smooth muscle.

Phenylephrine (PE; 100 μM) plus GTP (10 μM) induced Ca2+ sensitization of force (17.4 ± 4.8%,n = 4) in β-escin-permeabilized rabbit portal vein smooth muscle strips, and GDI (100 μM) relaxed the steady state of PE- plus GTP-induced contraction (at constant Ca2+) by 76 ± 11.2% (n = 4; Fig.5). In paired control strips, PE plus GTP caused a similar magnitude of contraction (20 ± 5.2%,n = 4) that was not significantly relaxed by the filtrate from the last concentration step of the GDI purification (6.3 ± 5% relaxation, n = 4; Fig. 5). The half-time of GDI-induced relaxation in these experiments was 1.80 ± 0.16 min.

Fig. 5.

Relaxation by GDI of the phenylephrine (PE)- plus GTP-induced, but not the phorbol 12,13-dibutyrate (PDBu)-induced, Ca2+ sensitization. A: β-escin-permeabilized rabbit portal vein smooth muscle was contracted by 10 μM GTP and 100 μM PE in the presence of submaximal concentration of Ca2+(pCa 6.4). GDI (100 μM; left) relaxed the GTP- plus PE-induced contraction, whereas the control, GDI concentrator filtrate (right), did not cause significant relaxation. B: GDI (100 μM) caused no significant relaxation of the PDBu-induced contraction. C: summary of the experiments in Aand B (n = 4–5). ***P < 0.001.

Because the Ca2+-sensitization protocol is carried out on muscles partially contracted with Ca2+ (pCa 6.4), GDI could have relaxed either the Ca2+-induced and/or the PE- plus GTP-induced contractions. Indeed, supramaximal concentrations of GDI (100 μM) relaxed submaximal Ca2+ (pCa 6.3)-induced contractions by 27 ± 8.7% (n = 4), in keeping with our previous finding that the Rho-kinase inhibitor Y-27632 partially relaxed submaximal Ca2+ contractions due to some constitutive Rho-kinase activity of these preparations (12). After GDI-induced relaxation of the submaximal Ca2+ contraction, the addition of GTP plus PE produced force that was significantly less (19 ± 2.2%, n= 5, P < 0.001) than that in the absence of GDI (29 ± 0.6%), indicating significant inhibition of the PE- plus GTP-induced Ca2+ sensitization.

To determine whether the inhibitory effect of GDI was specific for the α1-adrenergic receptor signaling pathway, we also determined its effect on Ca2+ sensitization in rabbit ileum smooth muscle by the muscarinic agonist carbachol. Preincubation with GDI (100 μM) also inhibited carbachol (100 μM)- plus GTP-induced Ca2+ sensitization of force from control 22 ± 4.2% (n = 7) to 14 ± 4.9% (n = 7,P < 0.01).

Phorbol 12,13-dibutyrate (PDBu; 1 μM)-induced Ca2+sensitization was not significantly affected by preincubation with 100 μM GDI. The PDBu (1 μM)-induced contraction was 15 ± 4.5% (n = 5) in control and 18 ± 4.7% (n = 5, P > 0.05) of the maximal Ca2+-induced contraction in the presence of 100 μM GDI. When GDI (100 μM) was added at the plateau of PDBu-induced Ca2+ sensitization (Fig. 5), its small relaxant effect was comparable to its effect on pCa 6.4-induced tension alone.

Ca2+ sensitization of force by the recombinant GTP · G14V RhoA/GDI complex.

His6 G14V RhoA and FLAG-GDI were coexpressed in yeast and purified as a complex from yeast cytosol (46). Over 90% of complex purified from this expression system is GDP bound (46). Therefore, the protein was nucleotide-exchanged to the GTP-bound form by a 6-h incubation at room temperature in the presence of 7.5 mM GTP and submicromolar Mg2+, resulting in 50–70% of complex being GTP bound (46). Concentrations of 1, 3, and 10 μM of the GTP-bound complex of G14V RhoA and GDI caused 16.5 ± 4.6% (n = 3), 25.6 ± 2.7% (n = 3), and 19.0 ± 1.3% (n = 5) of the maximal Ca2+-induced contraction in β-escin-permeabilized rabbit portal vein, respectively. The control, 10 μM GDP-bound G14V RhoA/GDI complex, caused only 8.9 ± 4.3% (n = 3) of the maximal Ca2+-induced contraction under identical conditions. Similarly to the contraction induced by GTP · G14V RhoA alone, the 3 μM GTP · G14V RhoA/GDI complex-induced contraction was completely relaxed (Fig. 6) by the Rho-kinase inhibitor Y-27632 (3 μM), from 25.6 ± 2.7% (n = 3) to 3.0 ± 1.0% (n = 3,P < 0.01). Contractions induced by 3 μM GTP · G14V RhoA/GDI complex were also completely relaxed (Fig.6), from 22.2 ± 6.7% (n = 3) to −2.2 ± 2.6% (n = 3, P < 0.01), by excess free GDI (20 μM).

Fig. 6.

Ca2+ sensitization of force induced by the complex of RhoA and GDI coexpressed in yeast and its relaxation by a Rho-kinase inhibitor, Y-27632, and by GDI. The His6 G14V RhoA and FLAG-GDI were coexpressed in yeast, purified, and loaded with GTP, as described in methods. This GTP · G14V RhoA/GDI complex (3 μM) caused Ca2+ sensitization of force in β-escin (75 μM, 35 min)-permeabilized rabbit portal vein. The Ca2+ sensitization of force induced by the complex was completely relaxed by Y-27632 (3 μM, n = 3) and GDI (20 μM, n = 3). The 0 force level represents the force at pCa 6.5. ***P < 0.001.

Comparison of inhibition by GDI of GTP- and GTPγS-induced Ca2+ sensitization.

To determine whether GDI-induced inhibition of the Rho/Rho-kinase pathway was affected by hydrolysis of GTP bound to RhoA, we compared the activity of GDI in the presence of a nonhydrolyzable nucleotide analog (GTPγS) with that in the presence of (hydrolyzable) GTP. The concentrations of GTP and GTPγS were chosen to yield equal levels of Ca2+-sensitized force, 39% of maximal pCa 4.5-induced force. The results (Fig. 7) suggest that GDI was, albeit modestly, more effective in inhibiting the effects of GTP than those of GTPγS. The concentrations of GDI that resulted in 50% relaxation of 10 μM GTP- and 0.1 μM GTPγS-induced Ca2+-sensitization were 0.56 ± 0.20 and 0.89 ± 0.11 μM, respectively (n = 4). The half-times of 10 μM GDI-induced relaxation of GTP (10 μM)- and GTPγS (0.1 μM)-induced force (Fig. 7 B) were 3.21 ± 0.31 and 4.15 ± 0.35 min, respectively (n = 5).

Fig. 7.

Relaxation by GDI of GTP- and GTPγS-induced Ca2+ sensitization. β-Escin-permeabilized rabbit portal vein smooth muscle strips were contracted by 10 μM GTP (○) or 0.1 μM GTPγS (●) in the presence of a submaximal concentration of Ca2+, pCa 6.5. Force was equivalent for the 2 nucleotide concentrations chosen. Increasing concentrations of GDI (A) or, for measurements of the time course of GDI-induced relaxation, 10 μM GDI (B) were added to the strips. Results are expressed as the percentage of nucleotide contraction obtained in the presence of pCa 6.5 (A) or the percentage of maximum relaxation (B). Note that the experimental design using EC50 values for GTP and GTPγS (0.1 μM vs. 10.0 μM) effectively negated the effect of GTP hydrolysis, favoring the GTP Rho state.

Inability of GTP · G14V RhoA to reconstitute GTPγS-induced Ca2+ sensitization of force in GTPγS-downregulated rabbit portal vein.

Prolonged incubation of α-toxin-permeabilized smooth muscle with GTPγS causes, by a yet unknown molecular mechanism, loss of the Ca2+-sensitizing response to GTPγS (downregulation) but not that to PDBu (15). To determine whether this could be due to sequestration of endogenous RhoA, we attempted to rescue the downregulation with constitutively active, recombinant RhoA. However, GTP · G14V RhoA (1 μM) caused no significant contraction (1.2 ± 1.2%, n = 5) in downregulated smooth muscle, and it did not recover the response to GTPγS (Fig.8), indicating that downregulation was not simply due to sequestration or loss of function of endogenous RhoA. In time-matched control strips, 1 μM GTP · G14V RhoA caused 12.8 ± 2.2% (n = 5) of the maximal Ca2+-induced contraction, which was not significantly different from the contraction induced by GTP · G14V RhoA in freshly β-escin-permeabilized tissue (Fig.8 C). Addition of the phosphatase inhibitor microcystin (MC) produced a large, rapid increase in force, indicative of normal catalytic phosphatase activity in GTPγS-downregulated muscles.

Fig. 8.

Inability of GTP · G14V RhoA to reconstitute GTPγS-induced Ca2+ sensitization of force in GTPγS-downregulated rabbit portal vein. GTPγS-induced Ca2+ sensitization was downregulated by prolonged incubation with GTPγS (50 μM, 6 h) in α-toxin-permeabilized rabbit portal vein, as described in methods. The GTPγS-downregulated strip (A) and time-matched control strip (B) were then permeabilized with β-escin (50 μM, 20 min) to create larger pores in the membrane to introduce the 21-kDa RhoA protein. GTP · G14V RhoA (1 μM) did not cause significant contraction and did not recover the response to subsequently added GTPγS in the GTPγS-downregulated strip (A). In the time-matched control (B), GTP · G14V RhoA caused significant contraction and GTPγS induced further force development, as found previously in fresh β-escin-permeabilized rabbit portal vein (15). Microcystin (MC), a phosphatase inhibitor, caused a similar amplitude of contraction in the GTPγS-downregulated and time-matched control strips. C: summary of the experiments in A andB; n = 5 for each group. ***P < 0.001.

Inhibition of GTP · G14V RhoA-induced Ca2+ sensitization by GTP · G12V Cdc42 and GTP · Q61L Rac1.

To explore other G protein-coupled mechanisms that may cross talk with RhoA/Rho-kinase-mediated Ca2+ sensitization, we determined the effect of two other Rho subfamily GTPases, Rac and Cdc42. The constitutively active mutant GTP · G12V Cdc42 (10 μM) significantly inhibited subsequent GTP · G14V RhoA-induced contraction from 24% (n = 3, 1 μM) to 3.5% of the maximal Ca2+-induced contraction (n = 3,P < 0.01; Fig. 9) but, by itself, had no significant effect on force (3.1% relaxation in control with filtrate vs. 0.47% relaxation with Cdc42,n = 3).

Fig. 9.

Inhibition of GTP · G14V RhoA-induced Ca2+ sensitization by G12V Cdc42 and Q61L Rac1. β-Escin-permeabilized rabbit portal vein smooth muscle was partially contracted with Ca2+ (pCa 6.5) and incubated (20 min) with the constitutively active Rho family proteins. A: G12V Cdc42 (10 μM; filled bars) or its filtrate (open bars) **P < 0.01. B: Q61L Rac1 (10 μM; filled bars) or its filtrate (open bars). GTP · G14V RhoA (1 μM) then was or was not (control) added (n = 3–4), and finally maximal force in response to pCa 4.5 was increased. **P < 0.01, ***P < 0.001.

The constitutively active GTP · Q61L Rac1 (10 μM), like GTP · G12V Cdc42, also significantly inhibited the subsequent response to GTP · G14V RhoA from the control of 20 ± 1.4% (n = 4) to 6.2 ± 1.0% (n = 3, P < 0.001) and, by itself, caused a very small contraction (3.5% ± 1.3, n = 4) compared with the relaxation obtained with the control filtrate (1.25 ± 0.17,n = 4).


RhoA-coupled mechanisms are activated when GDP is replaced by GTP in the RhoA/GDI complex and GTP · RhoA is translocated to the plasma membrane (reviewed in Refs. 9, 43, and55). Our measurements of endogenous RhoA and its complex with GDI show a close-to-stoichiometric relationship (∼2–4 μM each) between endogenous RhoA and GDI, particularly when considering that some GDI is associated with other Rho subfamily proteins, Cdc42 and Rac1; thus there is not a large pool of excess GDI available in vivo for removing and sequestering membrane-bound GTP · RhoA by mass action.

Of the endogenous RhoA/GDI complex, only a small fraction (∼30%) was lost from tissues permeabilized with β-escin, which allows transmembrane permeation of up to 130-kDa proteins (22), including recombinant GTP · RhoA/GDI (Fig. 6). The limited loss of the small RhoA molecule (62) and the endogenous, compact heterodimer with GDI (36) suggests that the mobility of endogenous RhoA/GDI may be limited by its association in a much larger and/or anchored complex with other protein(s). This possibility is also supported by the fact that, although purified RhoA/GDI is readily ADP-ribosylated by the clostridial exoenzyme C3 (46), the unpurified endogenous complex is not (6,14, 46). The slow leak of GDI from smooth muscle (present study) contrasts with its reportedly complete loss from mast cells within 5 min of permeabilization with streptolysin-O (44), possibly due to differences between Rac/GDI (mast cells) and RhoA/GDI (smooth muscle), respectively, different methods of permeabilization, and/or that a fraction of poorly diffusible GDI remaining in mast cells was not measured (44).

GDI relaxed the contractions induced by recombinant GTP · G14V RhoA and agonists (Figs. 4 and 5), consistent with its ability to negatively regulate the Rho/Rho-kinase pathway (38; reviewed in Ref.43). The relaxant effect of Y-27632 (Fig. 4, Aand C), a selective Rho-kinase inhibitor (12,60), further confirmed that the GDI-inhibitable effects of G14V RhoA and agonists were mediated by Rho-kinase. GDI also extracted RhoA from smooth muscle, as from other, membranes (33).

Inhibition of Ca2+ sensitization fully activated by GTP or GTPγS required considerably more GDI than the endogenous concentration. The extraction of RhoA by GDI is not catalytic but requires one-to-one binding. Therefore, the physiological significance of the recombinant GDI required for extracting RhoA has to be assessed in light of the total ratio of GDI to RhoA molecules. The volume of permeabilized smooth muscle used in this study was ∼0.036 μl, and, assuming an upper limit of 4 μM endogenous RhoA, 60% of it translocated by GTPγS to the membrane is equivalent to 2.4 μM (or 0.086 pmol). The bath volume containing the 500 nM GDI (or 0.06 nmol) required for 50% relaxation of Ca2+-sensitized force (Fig.7) was 120 μl, providing a nearly 1,000-fold molar excess of GDI over RhoA. In contrast, the concentration of endogenous free GDI available in an intact cell is approximately the same as the amount of RhoA that dissociated from the complex; this GDI would be insufficient for removing, on a mole basis, active GTP · RhoA from the membrane and inactivating the Rho/Rho-kinase pathway.

Activated RhoA (GTP · RhoA) associates with the cell membrane (14); therefore, its removal by GDI presumably represents competition between membrane lipids and GDI for the geranylgeranylated COOH terminus of RhoA and another, positively charged surface that interacts with GDI (21, 36). Hence, although the in vitro affinities of GDI for, respectively, GDP- and GTP-bound Rho family proteins (affinity constant ∼30 nM for Cdc42H) (42) are similar, our present results in vivo, in conjunction with studies with liposomes (46), suggest that in cells GDI competes more effectively for GDP · RhoA than for GTP · RhoA. In the absence of GDI, the higher concentration of GTP than GDP in cells would cause GDP · RhoA to recycle to active GTP · RhoA, and Rho signaling would be perpetually “on” through regeneration of GTP from GDP and ATP. Therefore, we conclude that the most likely physiological role of GDI is not to terminate the activity of GTP · RhoA but to sequester GDP · RhoA after GTP hydrolysis.

The relatively rapid time course of GDI-induced relaxation of Ca2+-sensitized force (Fig. 7 B) suggests that the activity of the downstream kinase(s) that Ca2+sensitizes by inhibiting smooth muscle myosin phosphatase (28) is closely temporally coupled to the presence of GTP · RhoA. This implies that activation of the effector kinase by RhoA is transient and that active phosphatase(s) in smooth muscle dephosphorylate the inhibitory site on the regulatory subunit of myosin phosphatase. The major effector kinase first identified as inhibiting MLCP is Rho-kinase (28), but other serine/threonine kinase(s) can also phosphorylate the inhibitory site of smooth muscle myosin phosphatase in vitro (11, 59). Because Rho-activated kinases also bind Rho (23, 34, 35, 39), negative regulation by GDI may also occur through removal of RhoA from its effector kinase(s) as well as from the plasma membrane.

We interpret the Ca2+-sensitizing effect of the GTP · G14V RhoA/GDI complex to result from its in vivo dissociation and the subsequent translocation of GTP · RhoA to the cell membrane, because GTP · RhoA, but not GDP · RhoA, spontaneously translocates from its complex with GDI to liposomes in vitro (46). [Should physiological translocation also require a membrane protein (4), then the results with liposomes (46) suggest that the protein is required for nucleotide exchange.] The Ca2+-sensitizing effect of GTP · G14V RhoA/GDI complex is not due to contamination by free GTP · G14V RhoA because the purified complex contains little or no detectable free RhoA (Ref.46, Fig. 4, lane 9; seemethods) and because similar concentrations of the complex and GTP · G14V RhoA caused similar forces (compare Fig. 4 with Fig. 6). It is also unlikely that the undissociated complex is active, like the undissociated Rac1/GDI that activates NADPH oxidase (1,7), because excess GDI inhibited Ca2+ sensitization induced by the GTP · G14V RhoA/GDI.

A significant portion of Rho-kinase (Fig. 1 A) and nearly all of the RhoA (present study and Ref. 14) translocated by GTPγS to the membrane fraction was Triton X-114 soluble (Fig.1 B). This finding indicates that activation of Rho-kinase not only translocates it to the membrane fraction, as also suggested by a recent immunofluorescence study (58), but also exposes its hydrophobic surface, consistent with an unfolding mechanism of Rho-kinase activation occurring at the cell membrane.

In addition to Rho-kinase (see above), other Ca2+-independent kinases can also phosphorylate the inhibitory site of the regulatory subunit of myosin phosphatase (Thr-695; Refs. 11 and 59). Furthermore, both Rho-kinase (32) and another Rho effector, protein kinase N (PKN) (17), can phosphorylate CPI17, a potent inhibitor, when phosphorylated, of myosin phosphatase (10). Regardless of the downstream effector(s), the inhibitory effect of GDI on agonist- and GTPγS-induced Ca2+ sensitization suggests that the most relevant kinases are RhoA effectors that converge to inhibit myosin II phosphatase.

Phorbol ester-induced Ca2+ sensitization mediated by protein kinase C(s) was not inhibited by GDI (present study, Fig. 5). This not only confirms but extends the conclusion, based on the use of a Rho-kinase inhibitor (12) and the effects of downregulation of, respectively, G protein-coupled and phorbol ester-induced Ca2+ sensitization (15, 25, 61), that the Rho/Rho-kinase pathway is independent of protein kinase C(s) stimulated by phorbol esters, although both converge on inhibiting myosin phosphatase. Our present results obtained with GDI rule out the involvement of RhoA in phorbol ester-induced Ca2+sensitization through not only Rho-kinase but also other Rho effectors, such as PKN.

Downregulation of G protein-coupled Ca2+ sensitization (15) did not decrease Rho-kinase content, and it could not be rescued by recombinant GTP · G14V RhoA. PDBu, as well as MC, an inhibitor of the catalytic subunit of myosin phosphatase, retain their Ca2+-sensitizing effect even when that of GTPγS is downregulated (15); therefore, the present results suggest that downregulation occurs downstream of RhoA through a yet-to-be-identified negative regulatory mechanism(s).

The two other constitutively active Rho-subfamily small GTPases, GTP · G12V Cdc42 (Fig. 9) and GTP · Q61L Rac1, significantly inhibited G14V RhoA-induced Ca2+sensitization of force. Negative regulation of Rho activity by constitutively active Rac1 and Cdc42 in NIH/3T3 cells has been ascribed to inhibition of RhoA activation: nucleotide exchange from GDP · RhoA to active GTP · RhoA (48); GTP · RhoA was estimated as the fraction of RhoA immunoprecipitated by a peptide containing the Rho-binding domain of rhotekin that interacts with GTP · RhoA but not with (inactive) GDP · RhoA (49). However, it is highly unlikely that such inhibition of activating nucleotide exchange (GTP for GDP) accounts for the inhibition of Ca2+ sensitization induced by constitutively active GTP · G14V RhoA (Fig. 9). It is possible that Rac1 and Cdc42 also interfere with the binding of GTP · RhoA to rhotekin and other effectors, including Rho-kinase. Other inhibitory mechanisms could involve PAK, a serine/threonine protein kinase effector of both Cdc42 and Rac1 (2) that can phosphorylate and inhibit MLCK (50). However, inhibition of MLCK by PAK should result in relaxation of the pCa 6.3-induced contraction independently of RhoA/Rho-kinase activity, and this was not the case. In any case, the effects of Cdc42 and Rac1 found in the present study suggest the existence of previously undescribed mechanism(s) through which these two GTPases that do not cause significant Ca2+sensitization can negatively regulate the Rho-kinase pathway, unlike the apoptosis RhoA pathway, which is not affected by either Rac1 or Cdc42 (40).


We are grateful to Akiko Yoshimura of Welfide Corporation for a generous gift of Y-27632. The human GDI cDNA was a gift from Dr. G. Bokoch, and the RhoA cDNA was a gift from Dr. A. Hall. We thank Barbara Nordin and Ann Folsom for preparation of the manuscript.


  • * M. C. Gong, I. Gorenne, and P. Read contributed equally to this work.

  • This work was supported by National Heart, Lung, and Blood Institute Grant HL-48807.

  • Present address of M. C. Gong: Department of Physiology, University of Kentucky School of Medicine, Lexington, KY 40536-0084.

  • Address for reprint requests and other correspondence: A. P. Somlyo, PO Box 800736, Charlottesville, VA 22908 (E-mail:aps2n{at}virginia.edu).

  • 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.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 45a.
  47. 46.
  48. 47.
  49. 48.
  50. 49.
  51. 50.
  52. 51.
  53. 52.
  54. 53.
  55. 54.
  56. 55.
  57. 56.
  58. 57.
  59. 58.
  60. 59.
  61. 60.
  62. 61.
  63. 62.
  64. 63.
View Abstract