Role of Rho and Rho kinase in the activation of volume-regulated anion channels in bovine endothelial cells


Corresponding author B. Nilius: Katholieke Universiteit Leuven, Laboratorium voor Fysiologie, Campus Gasthuisberg O&N, B-3000 Leuven, Belgium. Email:


  • 1We have studied the modulation of volume-regulated anion channels (VRACs) by the small GTPase Rho and by one of its targets, Rho kinase, in calf pulmonary artery endothelial (CPAE) cells.
  • 2RT-PCR and immunoblot analysis showed that both RhoA and Rho kinase are expressed in CPAE cells.
  • 3I Cl,swell, the chloride current through VRACs, was activated by challenging CPAE cells with a 25 % hypotonic extracellular solution (HTS) or by intracellular perfusion with a pipette solution containing 100 μM GTPγS.
  • 4Pretreatment of CPAE cells with the Clostridium C2IN-C3 fusion toxin, which inactivates Rho by ADP ribosylation, significantly impaired the activation of ICl,swell in response to the HTS. The current density at +100 mV was 49 ± 13 pA pF−1 (n= 17) in pretreated cells compared with 172 ± 17 pA pF−1 (n= 21) in control cells.
  • 5The volume-independent activation of ICl,swell by intracellular perfusion with GTPγS was also impaired in C2IN-C3-pretreated cells (31 ± 7 pA pF−1, n= 11) compared with non-treated cells (132 ± 21 pA pF−1, n= 15).
  • 6Activation of ICl,swell was pertussis toxin (PTX) insensitive.
  • 7Y-27632, a blocker of Rho kinase, inhibited ICl,swell and delayed its activation.
  • 8Inhibition of Rho and of Rho kinase by the above-described treatments did not affect the extent of cell swelling in response to HTS.
  • 9These experiments provide strong evidence that the Rho-Rho kinase pathway is involved in the VRAC activation cascade.

In most mammalian cells, cell swelling activates a Cl current (ICl,swell) through volume-regulated anion channels (VRACs). These channels are involved in cell volume regulation, electrogenesis, control of electrochemical gradients for ion channels and transporters and possibly in cell proliferation and differentiation (Strange et al. 1996; Nilius et al. 1996a, 1997a; Okada, 1997; Kirk, 1997; Kirk & Strange, 1998). The nature of the cellular volume sensor, as well as the transduction apparatus that controls the gating of VRACs, is unknown. The cell swelling-induced activation of VRACs requires one or more tyrosine phosphorylation steps, since inhibitors of protein tyrosine kinases inhibit ICl,swell whereas inhibitors of protein tyrosine phosphatases potentiate ICl,swell (Tilly et al. 1993; Voets et al. 1998). It was recently shown that p56lck, a member of the src protein tyrosine kinase family, mediates activation of ICl,swell in lymphocytes (Lepple-Wienhues et al. 1998). We have previously shown that ICl,swell is also activated by reducing the intracellular ionic strength (Γi) (Nilius et al. 1998) or by intracellular perfusion with GTPγS (Voets et al. 1998). How reduced Γi or intracellular GTPγS activate VRACs is unknown, but similar to the swelling-induced activation one or more tyrosine phosphorylation events are involved. Furthermore, the Γi- and GTPγS-activated Cl currents are still sensitive to changes in cell volume, as indicated by the blocking effect of cell shrinkage induced by extracellular hypertonicity (Voets et al. 1998; Nilius et al. 1998). The GTPγS effect suggests a crucial role of GTP-binding proteins in the activation of VRACs.

Rho (or p21rho) is a monomeric 21 kDa GTPase of which there are three mammalian isoforms (RhoA, RhoB and RhoC) and which, together with Rac, Cdc42, RhoD, RhoE, RhoG and TC10, form the Rho GTPase family (Mackay & Hall, 1998). Rho proteins are involved in the organization of the actin cytoskeleton, in the formation of stress fibres, in exo/endocytosis and in formation of focal adhesions (Ridley, 1996; Symons, 1996). It has recently been shown that inactivation of Rho by pretreatment with the Clostridium botulinum C3 exoenzyme greatly reduced the swelling-induced efflux of iodide in human Intestine 407 cells (Tilly et al. 1996). A signalling cascade whereby activation of Rho results in the activation of p125FAK (a protein tyrosine kinase localized to focal adhesion contacts) and phosphatidylinositol-3-kinase was suggested to be critically involved in the activation of VRACs (Tilly et al. 1996). However, several alternative signalling pathways downstream of Rho have been described, among which are protein kinases that are activated by binding to GTP-bound Rho (Ridley, 1996; Tapon & Hall, 1997). A subclass of these Rho-activated protein kinases is formed by Rho kinase which is a serine/threonine protein kinase with structural homology to the myotonic dystrophy kinase (Matsui et al. 1996). Two distinct isoforms (Rho Kinase/ROKα/ROCK-II and p160ROCK/ROKβ/ROCK-I) with approximately 90 % identity in their kinase domain have been identified (Leung et al. 1996; Narumiya et al. 1997). Rho-dependent activation of Rho kinase promotes the formation of focal adhesions and actin stress fibres in fibroblasts and HeLa cells and increases the Ca2+ sensitivity of the contractile apparatus in smooth muscle (Narumiya et al. 1997).

In this study we have investigated the role of the Rho-Rho kinase signalling cascade in the activation of ICl,swell in calf pulmonary artery endothelial (CPAE) cells. We show that inhibiting Rho or Rho kinase results in an impaired activation of ICl,swell by either cell swelling or GTPγS. We therefore conclude that in CPAE cells the Rho-Rho kinase pathway modulates the gating process of VRACs.



The use of the Clostridium limosum C3 exoenzyme (an ADP-ribosyltransferase acting on Rho) is hampered by the difficulty of introducing C3 into the cell (Barth et al. 1998). We have therefore used the fusion toxin C2IN-C3. It consists of the N-terminal part of the C2I component of the Clostridium botulinum toxin C2 fused to the entire C3 exoenzyme. Clostridium botulinum C2 toxin is a heterodimer consisting of the enzymatic component C2I (an ADP-ribosyltransferase acting on actin) and a membrane-binding protein C2II responsible for cellular uptake of C2. Cellular uptake of C2I requires its N-terminal part (C2IN) which interacts with C2II and thereby allows internalization of C2I (Just et al. 1992; Barth et al. 1998). Similarly, cellular uptake of the C2IN-C3 fusion toxin requires the presence of C2II. The C2IN-C3 fusion toxin (5 μg) and C2IIa (10 μg) were dissolved in 50 μl water each. One micromole of both solutions (e.g. 100 ng C2IN-C3 fusion toxin and 200 ng C2IIa) were added to 1 ml culture medium (Barth et al. 1998). Incubation lasted between 4 and 24 h. The Rho kinase inhibitor Y-27632 ((+)-(R)-trans-4-(1-aminoethyl)-N-(4-pyridyl)) was kindly provided by Yoshitomi Pharmaceutical Industries, Ltd, Osaka. A 10 mM stock solution in water was diluted to give a final concentration of 10, 50 or 100 μM. The non-specific effects of C2IN-C3, C2IIa or Y-27632 on CPAE cells were tested by checking for the inward rectifier K+ channel which was identical in control and treated cells. GTPγS (lithium salt) was purchased from Sigma. For some experiments cells were pretreated with 500 ng ml−1 pertussis toxin (PTX, Sigma) for 16 h at 37°C to test for a possible involvement of PTX-sensitive G proteins.

Endothelial cell culture

Calf pulmonary artery endothelial (CPAE) cells were purchased from the American Tissue Culture Collection (CCL-209) at passage 16 and used between passages 22 and 27. They were grown in Dulbecco's modified Eagle's medium (Gibco) containing 20 % fetal calf serum, 2 mM L-glutamine, 100 U ml−1 penicillin and 100 μg ml−1 streptomycin, maintained at 37°C in a fully humidified atmosphere of 10 % CO2 in air, and passaged by brief exposure to 0.5 g l−1 trypsin in a Ca2+- and Mg2+-free solution. Only non-confluent and non-clustered cells were used for current measurements.


The standard extracellular solution was a Krebs solution containing (mM): 150 NaCl, 6 KCl, 1 MgCl2, 1.5 CaCl2, 10 glucose, 10 Hepes (pH adjusted to 7.4 with NaOH). The osmolality of this solution, as measured with a vapour pressure osmometer (Wescor 5500, Schlag, Gladbach, Germany), was 320 ± 5 mosmol kg−1. At the beginning of the patch-clamp recordings, the Krebs solution was replaced by a Cs+ solution containing (mM): 105 NaCl, 6 CsCl, 1 MgCl2, 1.5 CaCl2, 10 glucose, 90 mannitol, 10 Hepes (pH adjusted to 7.4 with NaOH; 320 ± 5 mosmol kg−1). A 25 % (240 ± 5 mosmol kg−1) hypotonic solution was obtained by omitting 90 mM mannitol from this solution. Hypertonic solutions were obtained by addition of 100 mM mannitol, resulting in an osmolality of 410 ± 5 mosmol kg−1. In most experiments, a standard pipette solution was used, which contained (mM): 40 CsCl, 100 caesium aspartate, 1 MgCl2, 1.93 CaCl2, 5 EGTA, 4 Na2ATP, 10 Hepes (pH adjusted to 7.2 with CsOH; 290 mosmol kg−1). This solution was slightly hypotonic compared with the Krebs solution to avoid spontaneous activation of volume-sensitive Cl currents. The concentration of free Ca2+ in this solution was buffered at 100 nM, which is below the threshold for activation of Ca2+-activated Cl currents (Nilius et al. 1997B) but sufficient to fully activate volume-activated Cl currents in CPAE cells. GTPγS was included in the standard pipette solution at a concentration of 100 μM. After breaking in, cells were therefore continuously exposed to intracellular GTPγS.

Current measurements

CPAE cells were used 1-3 days after seeding on gelatine-coated coverslips (2000 cells mm−2). In the experimental chamber, the cells were continuously superfused with isosmotic Krebs solution at room temperature (22°C). Whole-cell membrane currents in ruptured patches were monitored with an EPC-9 (Heka Electronics, Lambrecht/Pfalz, Germany) patch-clamp amplifier. The following voltage protocol was used: from a holding potential of 0 mV, a step of 0.2 s duration to -100 mV was applied, followed by a 1.3 s linear voltage ramp to +100 mV, after which the potential was stepped back to the holding potential. This protocol was repeated every 15 s. Because of the occasional variability of ICl,swell, even under control conditions, probably due to different passages, we have always compared cells from the same batch and passage.

Capacitance measurements

The capacitance was routinely monitored from the capacitance-track output of the EPC-9 amplifier. These values were used to normalize the current amplitudes to cell capacitance.

Volume measurements

Cell volume was assessed in voltage-clamped cells simultaneously with the current. This method monitors the vertical shift of fluorescent beads deposited on the cell surface. The principles of this method have been described in detail elsewhere (Van Driessche et al. 1993). In short, cells seeded on a coverslip were incubated for about 30 min with Red Neutravidin-labelled microspheres (F-8775, Molecular Probes, 4 μl beads in 1 ml normal Krebs solution). The microspheres were visualized using the XF40/E filter set (Omega filter, Inc., Brattleboro, VT, USA) with a broad excitation band (maximal at 5460 nm) and a 600 nm long pass emission filter. Cell height (h), taken as a measure for cell volume, was determined as the vertical distance between the beads on the gelatine support of the cells and the beads on the cell surface (for details see Van Driessche et al. 1993). At the same time whole-cell current measurements were performed.

RT-PCR analysis

Poly-A+-enriched RNA from CPAE cells was prepared with the Oligotex direct kit (Qiagen, Hilden, Germany). Poly-A+ RNA (0.3 μg) was reversely transcribed in cDNA with MMLV reverse transcriptase and random primers (Ready To Go You Prime kit; Pharmacia, Uppsala, Sweden). Aliquots corresponding to 1/10 of the reverse transcription reaction mix were then amplified with Taq polymerase (Taq Core kit; Qiagen) according to the manufacturer's instructions. PCR primers for bovine RhoA (EMBL/GenBank accession number M27278) were TCAGAATTCGCCATGGCTGCCATCCGGAA and TCAGAATTCAAGCTTTCACAAGACAAGGCACC. PCR reactions for RhoA were carried out for 25 cycles: 1 cycle consisted of 1 min at 94°C, 1 min at 60°C, and 1 min at 72°C. PCR primers for bovine Rho kinase (EMBL/GenBank accession number U36909: corresponds to the bovine ROKα/ROCK-II homologue) were CAGATGAAGGCAGAAGAC and ACCACCAATCACATTCTC. Whether these primers can also amplify bovine ROKβ/ROCK-I cDNA is not known, since cDNAs for this bovine isoform have not yet been cloned. PCR reactions for Rho kinase were carried out for 25 cycles: 1 cycle consisted of 1 min at 94°C, 1 min at 54°C, and 1 min at 72°C. Negative control reactions were performed on poly-A+ RNA aliquots which had not been subjected to a reverse transcription reaction. PCR products for RhoA (600 nucleotides (nt)) and Rho kinase (580 nt) were analysed on a 1 % agarose gel. The identity of the fragments was confirmed by subcloning the PCR fragments and nucleotide sequence analysis on an automated ALFexpress system (Pharmacia).

Immunoblot analysis

CPAE cells (107) were lysed in a hypotonic buffer (1 ml) containing 25 mM Tris-HCl (pH 7.5), 20 mM NaCl, 2.5 mM EGTA, 0.5 % (v/v) Nonidet P-40 and a protease inhibitor cocktail (1 mM phenylmethylsulphonyl fluoride, 0.1 μg ml−1 leupeptin). The lysate was centrifuged and the supernatant stored at -20°C. Fifty micrograms of protein were separated on SDS-PAGE (7.5 % acrylamide for Rho kinase; 10 % polyacrylamide for RhoA) and transferred to a polyvinylidene difluoride (PVDF) microporous membrane (Immobilon, Millipore) by semi-dry electroblotting. RhoA was detected with a monoclonal antibody specific for RhoA (Santa Cruz No. sc-418, Santa Cruz, CA, USA) in a 1:2000 dilution. Bovine Rho kinase was detected with a rabbit polyclonal antiserum in a 1:500 dilution. Secondary antibodies were either anti-mouse (RhoA) or anti-rabbit conjugated to alkaline phosphatase. Immunoreactive bands were visualized with an enhanced chemifluorescence detection kit (Amersham) on a Storm 840 imager (Molecular Dynamics, Sunnyvale, CA, USA).

Statistical analysis

Pooled data are given as means ±s.e.m. Statistical significance was calculated at the 5 % level using Student's t test.


Expression of RhoA and Rho kinase in CPAE cells

We first checked whether CPAE cells express RhoA and the bovine Rho kinase isoform that was previously cloned from bovine brain and corresponds to the ROKα/ROCK-II isoform (Matsui et al. 1996). Figure 1 (A and C) shows that RT-PCR products of the expected length (RhoA: 600 nt; Rho kinase: 580 nt) could be amplified starting from CPAE poly-A+ RNA. Expression of RhoA and Rho kinase was confirmed at the protein level by immunoblot analysis using an anti-RhoA monoclonal antibody or an anti-Rho kinase polyclonal antiserum (Fig. 1A and D).

Figure 1.

Expression of RhoA and Rho kinase in CPAE cells

A and C, RT-PCR was performed on CPAE cell poly-A+ RNA which was reversely transcribed in cDNA and subsequently amplified with primers specific for RhoA (A) or Rho kinase (C) as described in Methods. PCR products were analysed on a 1 % agarose gel. The specificity of the PCR products was verified by subcloning and sequence analysis. Lanes ‘+’ and ‘-’ refer to PCR reactions where the reverse transcription step was, respectively, included or excluded. Lane M is a DNA size standard. B and D, for Western blotting, 50 μg of total cellular protein was separated on SDS-PAGE (7.5 % acrylamide for Rho kinase or 10 % acrylamide for RhoA) and electroblotted on a PVDF membrane. Blots were probed with a monoclonal antibody against RhoA (B) or a polyclonal antiserum against Rho kinase (D). Lanes ‘+’ and ‘-’ correspond to blots where the primary antibody was, respectively, included or excluded. The migration pattern of protein molecular mass standards is shown on the left. The duplet in B is unlikely to represent different Rho isoforms since the monoclonal antibody is specific for RhoA. The position of the lower band is consistent with the molecular mass of RhoA (21.8 kDa); the upper band may represent a post-translational modification. The lower band in D migrating at ≈100 kDa probably represents a degradation product.

Inhibition of ICl,swell by Clostridium limosum C3 toxin

The C. limosum C3 exoenzyme inactivates Rho by ADP-ribosylation on asparagine 41, but it does not act on other members of the Rho superfamily such as Rac or Cdc42 (Just et al. 1992; Aktories, 1997). Since the C3 exoenzyme does not penetrate the cell, we used a fusion toxin C2IN-C3 in which the C3 exoenzyme was fused to the N-terminal part of C. botulinum C2I. CPAE cells were pre-incubated with C2IN-C3 and C2IIa, which is required for cellular uptake of the fusion protein (see Methods). Figure 2A shows a simultaneous measurement of whole-cell current and cell volume in a control CPAE cell, showing rapid cell swelling and activation of an outward current at +100 mV in response to a stimulation with hypotonic extracellular solution (HTS). Both cell swelling and current activation were completely reversible and could be reproduced several times. Current-voltage relations measured at the indicated time points show that the activated current, ICl,swell, was outwardly rectifying and reversed close to the equilibrium potential for Cl, ECl (Fig. 2A). We have previously described this current in more detail (see Nilius et al. 1997a). Activation of ICl,swell was drastically impaired in cells pretreated with C2IN-C3 and C2II (Fig. 2A). The amplitude of ICl,swell at +100 mV in response to HTS decreased from 172 ± 17 pA pF−1 (n= 21) in control cells to 49 ± 13 pA pF−1 (n= 17; P < 0.05) in C2IN-C3-C2II-pretreated cells. The volume measurements in Fig. 2A and C illustrate that the impaired activation of ICl,swell in the pretreated cells was not due to impaired cell swelling. In control cells, cell height (h) increased by 54 ± 5 % (n= 10) 100 s after application of HTS, versus 62 ± 16 % (n= 5) in C2IN-C3-C2II-pretreated cells. The toxin exerted an inhibitory effect on ICl,swell over the entire voltage range (Fig. 2A). The rate of activation of ICl,swell was also significantly reduced in the C2IN-C3-C2II-pretreated cells, and this reduction was even more prominent during a second application of HTS. The times for half-maximal activation of ICl,swell during the first and second hypotonic stimulus were, respectively, 71 ± 6 and 73 ± 5 s (n= 7) in control cells versus 105 ± 12 and 152 ± 19 s in the pretreated cells (n= 6,P < 0.05).

Figure 2.

Effect of inhibition of Rho on the HTS-induced activation of ICl,swell

A and C, simultaneous measurement of changes in cell height (Δh; trace) and in whole-cell current at +100 mV (○) during two applications of HTS (bars) in a control cell (A) and a cell treated with C2IN-C3-C2II (C). B and D, current-voltage relations obtained at time points a and b indicated in A and C.

The amplitude of ICl,swell in cells pre-incubated with C3 exoenzyme alone (300 ng (ml culture medium)−1) was 170 ± 29 pA pF−1 (n= 6), which is not significantly different from that in control cells (P > 0.05). This probably reflects the failure of the C3 exoenzyme to penetrate intact CPAE cells. Furthermore, pre-incubation of CPAE cells with C2II toxin alone did not affect activation of VRACs (same protocol as used for the fusion toxin). In five cells after 24 h incubation with C2II, the activated current was still 219 ± 29 pA pF−1 at +100 mV, which is not significantly different from all other controls. Additionally, pretreatment with PTX did not significantly inhibit the HTS-induced activation of ICl,swell: amplitudes at +100 mV were 211 ± 21 pA pF−1 (n= 5) in control cells and 193 ± 14 pA pF−1 (n= 5) in PTX-pretreated cells. Therefore, PTX-sensitive G proteins are most probably not involved in the activation of ICl,swell.

We have previously shown that ICl,swell can be transiently activated, in the absence of any hypotonic stimulus, by intracellular application of GTPγS (Voets et al. 1998). As illustrated in Fig. 3A, this transient activation of ICl,swell occurred without any cell swelling, and a subsequent stimulation with HTS was still able to activate ICl,swell. Activation of ICl,swell by GTPγS was significantly impaired in C2IN-C3-C2II-pretreated cells (Fig. 3A). Current amplitudes at +100 mV were 132 ± 21 pA pF−1 (n= 15) in control cells versus 31 ± 7 pA pF−1 (n= 11) in the pretreated cells. The shape of the current-voltage relation of the GTPγS- and HTS-induced currents was similar in control and pretreated cells (Fig. 3A and D). Pretreatment with PTX was again without effect: the amplitude of the GTPγS-activated current at +100 mV was 117 ± 8 pA pF−1 (n= 11) in control cells and 138 ± 14 pA pF−1 (n= 5) in PTX-pretreated cells.

Figure 3.

Effect of inhibition of Rho on the GTPγS-induced activation of ICl,swell

A, changes in cell height (Δh; trace) and activation of whole-cell current (○) in a control cell intracellularly perfused with 100 μM GTPγS and superfused with HTS (bar). B, current-voltage relations obtained at the time points a, b and c indicated in A. C and D, same as A and B but in a cell treated with C2IN-C3-C2II.

Modulation of VRACs by Y-27632, an inhibitor of Rho kinase

The results presented so far indicate that Rho is involved in the activation cascade of VRACs in endothelial cells, consistent with previous findings in Intestine 407 cells (Tilly et al. 1996). Tilly et al. presented evidence for the involvement of p125FAK and phosphatidylinositol-3-kinase in the Rho-dependent activation of VRACs in Intestine 407 cells. However, in CPAE cells, inhibition of phosphatidylinositol-3-kinase with wortmannin or activation of p125FAK with lysophosphatidic acid had no influence on ICl,swell (Szücs et al. 1996). We therefore investigated the role of Rho kinase, a serine/threonine protein kinase that is activated by Rho (Narumiya et al. 1997). Rho kinase can be inhibited by the pyridine derivative Y-27632. Y-27632 has recently been described as a selective inhibitor of Rho kinase with an IC50 value of 0.14 μM, but it does not discriminate between Rho kinase isoforms (Uehata et al. 1997). Figure 4A shows the inhibitory effect of the extracellular application of 10 μM Y-27632 after maximal activation of ICl,swell by HTS. A maximal inhibition of 20 ± 4 % (n= 7) was reached after 5-10 min. Pre-incubating the cells for 10 min with 10 μM Y-27632 (Fig. 4A) caused a pronounced delay of current activation and a much stronger inhibition of the current (62 ± 8 %, n= 6). The simultaneous volume measurement clearly shows that Y-27632 did not impair HTS-induced cell swelling (Fig. 4A). To investigate the time course of the inhibitory effect of Y-27632, we repeatedly activated ICl,swell in the presence of the blocker. Repetitive stimulation of control CPAE cells with HTS activates ICl,swell without significant run-down (Nilius et al. 1996B; see also Figs 2A and 4D). However, in the continuous presence of 10 μM Y-27632, we observed a slow but substantial decay of current responses, which could not be attributed to a reduced cell swelling (Fig. 4A). Half-maximal inhibition was reached within approximately 5 min of Y-27632 application (Fig. 4A). The relatively slow inhibitory effect of Y-27632 was not due to a slow penetration of the compound into the cell, as inclusion of 50 μM Y-27632 in the pipette induced a time-dependent inhibition of VRACs with a similar time course (not shown, n= 4).

Figure 4.

Effect of the Rho kinase inhibitor Y-27632 on ICl,swell

A, effect of the acute application of 10 μM Y-27632 on ICl,swell pre-activated by HTS. B, changes in cell height (Δh; trace) and whole-cell current (○) in a cell pre-incubated with 10 μM Y-27632 for the time period indicated by the bar. C, simultaneous measurement of changes in cell height ((Δh; trace) and in whole-cell current (○) during three applications of HTS (bars) in a cell with 10 μM Y-27632 for the indicated time period. D, normalized amplitude of ICl,swell at +100 mV induced by repetitive stimulations with HTS in control cells (○) and in the presence of 10 μM Y-27632 (•).

We also studied the effect of the Rho-kinase blocker Y-27632 on the GTPγS-activated VRACs. For this purpose, 100 μM GTPγS was included in the pipette together with 100 μM Y-27632. The GTPγS-induced current in the presence of Y-27632 (35 ± 6 pA pF−1, n= 7) was significantly smaller than that in the absence of the inhibitor (132 ± 21 pA pF−1, n= 15,P < 0.05).


Voltage-regulated anion channels (VRACs) are ubiquitously expressed in mammalian cells and play a pivotal role in various cell functions (for a detailed discussion see Strange & Jackson, 1995; Strange et al. 1996; Nilius et al. 1996a, 1997a; Okada, 1997; Kirk, 1997; Strange, 1998; Kirk & Strange, 1998). However, the gating mechanism of VRACs is still poorly understood. Several models have been proposed, such as direct mechanical activation in analogy with stretch-activated cation channels, changes of intracellular protein crowding, incorporation of channel proteins into the plasma membrane triggered by changes in cell volume, and more generally an involvement of several phosphorylation signalling cascades. A detailed overview of the possible gating mechanisms is given elsewhere (Nilius et al. 1997a).

Recently, evidence has been accumulating for an involvement of G proteins in the regulation of VRACs. Intracellular application of GTPγS has been shown to activate ICl,swell in several cell types (e.g. Doroshenko et al. 1991; Tilly, 1991; Nilius et al. 1994; Mitchell et al. 1997; Voets et al. 1998). This activation is transient and occurs without any change in cell volume. Moreover, activation of ICl,swell is significantly impaired in the presence of intracellular GDPβS (Doroshenko & Neher, 1992; Voets et al. 1998). However, these effects of GTPγS and GDPβS did not give any clue to the kind of G proteins that are involved. The reduction of the osmosensitive anion efflux in Intestine 407 cells induced by the Clostridium botulinum C3 toxin gave a strong initial indication for the involvement of the small GTPase Rho (Tilly et al. 1996). We have further investigated this possible involvement in endothelial cells. Our results show a 70-80 % inhibition of ICl,swell, activated either by HTS or by GTPγS, in cells in which Rho was inactivated by pretreatment with C2IN-C3-C2II. Since C3 exoenzymes ADP-ribosylate RhoA, RhoB and RhoC (Just et al. 1992; Aepfelbacher et al. 1997), our data do not allow us to discriminate which of these Rho isoform(s) is(are) involved in the VRAC activation cascade. However, the expression of RhoA in CPAE cells suggests that the observed effects can at least partially be ascribed to this isoform.

Tilly et al. (1996) proposed p125FAK and phosphatidylinositol-3-kinase as downstream effectors of Rho in the activation of VRACs. As already pointed out above, these enzymes most probably do not significantly contribute to the activation of VRACs in CPAE cells (Szücs et al. 1996). Our present results point rather to Rho kinase as the downstream target of Rho, since ICl,swell was inhibited by ∼60 % after treatment with the Rho kinase inhibitor Y-27632. Further support for this hypothesis comes from the inhibition of the current activated by GTPγS. This current was reduced by ∼70 % in the presence 100 μM Y-27632 included in the pipette. It is at present not clear which Rho kinase isoform (ROKα/ROCK-II versus ROKβ/ROCK-I) is involved in the activation of VRACs. Indeed, Rho interacts with both isoforms (Leung et al. 1996; Ishizaki et al. 1996; Matsui et al. 1996) and the Y-27632 compound does not differentiate between the two isoforms (Uehata et al. 1997).

We can exclude a role for PTX-sensitive G proteins since PTX did not affect activation of VRACs. Similarly, the Rho-related GTPases Rac and Cdc42, as well as PAK (a protein kinase activated by Rac or Cdc42, see Ridley, 1996), are very unlikely to contribute since (i) C. limosum C3 exoenzyme does not act on Rac or Cdc42 (Just et al. 1992) and (ii) Y-27632 does not appreciably inhibit PAK (Uehata et al. 1997). To conclude, our data are consistent with the Rho-Rho kinase pathway being involved in the VRAC activation cascade, but the precise molecular identity of the Rho and Rho kinase isoforms remains to be established.

Whereas the involvement of protein tyrosine phosphorylation and of a Rho-dependent pathway in the regulation of VRACs seems to be established, it is at present far from clear how these enzymes fit into the activation mechanism for VRACs. There are two indications that the tyrosine phosphorylation step is closer to the channel than the Rho-dependent steps. Firstly, the effects of protein tyrosine kinase/phosphatase inhibitors on ICl,swell are significantly faster than the effects of Y-27632 and of GDPβS (Voets et al. 1998). Secondly and more directly, we have shown that the activation of ICl,swell by GTPγS can still be inhibited by tyrosine kinase inhibitors (Voets et al. 1998). However, a simple linear model (cell swelling → Rho-Rho kinase → protein tyrosine kinase → VRAC) seems unlikely, as it cannot explain how an extracellular hypertonic solution inhibits the GTPγS-induced ICl,swell (Voets et al. 1998). Moreover, inhibition of Rho or Rho kinase always caused a reduction of ICl,swell, but never a complete inhibition. An alternative model whereby Rho-Rho kinase exerts a permissive effect, e.g. by modulating the volume sensitivity of the activation of VRACs, might be more appropriate. The intriguing, but as yet unanswered, question remains of whether the Rho-Rho kinase effects on VRAC activation are related to its documented effects on the cytoskeleton such as formation of stress fibres and phosphorylation of myosin light chain (Narumiya et al. 1997) or even phosphorylation of ezrin/radixin/moesin (ERM) proteins, which are involved in actin filament-plasma membrane interaction (Matsui et al. 1998). Perturbing the F-actin microfilament system with phalloidin or cytochalasin B did not interfere with the swelling-induced activation of VRACs in endothelial cells (Oike et al. 1994). However, we have recently provided evidence that annexin II, a plasma membrane-associated, actin filament-binding protein plays a functional role in the activation of VRACs in CPAE cells, suggesting a role for the subplasma membrane cytoskeleton (Nilius et al. 1996B). Whether the Rho-Rho kinase effects on VRACs are exerted via the subplasma membrane cytoskeleton remains to be investigated.

In conclusion, we have provided strong evidence for an involvement of Rho-Rho kinase in the gating of VRACs in endothelium. However, a detailed picture of the activation of this functionally important channel is still elusive.


We are grateful to Mr. Y. Miura (Yoshitomi Pharmaceutical Industries, Ltd, Tokyo Laboratories, Saitama, Japan) for the kind gift of the Rho kinase inhibitor Y-27632. We thank D. Trouet and M. Kamouchi for helpful discussion. J.E. is a Research Associate and T.V. a post-doctoral researcher of the Flemish Fund for Scientific Research (F.W.O-Vlaanderen). This work was supported by a network grant of the European Community (Contract No. BMH4-CT96-0602).