Small- and intermediate-conductance Ca2+-activated K+ channels directly control agonist-evoked nitric oxide synthesis in human vascular endothelial cells

Jian-Zhong Sheng, Andrew P. Braun


The contribution of small-conductance (SKCa) and intermediate-conductance Ca2+-activated K+ (IKCa) channels to the generation of nitric oxide (NO) by Ca2+-mobilizing stimuli was investigated in human umbilical vein endothelial cells (HUVECs) by combining single-cell microfluorimetry with perforated patch-clamp recordings to monitor agonist-evoked NO synthesis, cytosolic Ca2+ transients, and membrane hyperpolarization in real time. ATP or histamine evoked reproducible elevations in NO synthesis and cytosolic Ca2+, as judged by 4-amino-5-methylamino-2′,7′-difluorofluorescein (DAF-FM) and fluo-3 fluorescence, respectively, that were tightly associated with membrane hyperpolarizations. Whereas evoked NO synthesis was unaffected by either tetraethylammonium (10 mmol/l) or BaCl2 (50 μmol/l) + ouabain (100 μmol/l), depleting intracellular Ca2+ stores by thapsigargin or removing external Ca2+ inhibited NO production, as did exposure to high (80 mmol/l) external KCl. Importantly, apamin and charybdotoxin (ChTx)/ triarylmethane (TRAM)-34, selective blockers SKCa and IKCa channels, respectively, abolished both stimulated NO synthesis and membrane hyperpolarization and decreased evoked Ca2+ transients. Apamin and TRAM-34 also inhibited an agonist-induced outwardly rectifying current characteristic of SKCa and IKCa channels. Under voltage-clamp control, we further observed that the magnitude of agonist-induced NO production varied directly with the degree of membrane hyperpolarization. Mechanistically, our data indicate that SKCa and IKCa channel-mediated hyperpolarization represents a critical early event in agonist-evoked NO production by regulating the influx of Ca2+ responsible for endothelial NO synthase activation. Moreover, it appears that the primary role of agonist-induced release of intracellular Ca2+ stores is to trigger the opening of both KCa channels along with Ca2+ entry channels at the plasma membrane. Finally, the observed inhibition of stimulated NO synthesis by apamin and ChTx/TRAM-34 demonstrates that SKCa and IKCa channels are essential for NO-mediated vasorelaxation.

  • calcium
  • endothelium
  • hyperpolarization
  • small-conductance calcium-activated potassium channel
  • intermediate-conductance calcium-activated potassium channel channel

the vascular endothelium exerts precise control over the contractile state of the vessel wall through the stimulated synthesis and release of both constrictor and dilatory substances. Stimulus-induced vasodilation appears to occur principally via cellular mechanisms that involve the synthesis and release of nitric oxide (NO) or EDRF (20) and prostacyclin along with a non-NO, non-prostanoid EDHF, whose identity remains a matter of debate. As recently discussed (5, 40), potential candidates for EDHF include K+, cytochrome P-450 metabolites of arachidonic acid, hydrogen peroxide, and C-type naturetic peptide as well as electrical coupling between cells via myoendothelial gap junctions. As the EDHF-type response appears to display differing pharmacological sensitivity in a variety of vascular preparations and species, it is possible that more than one cellular mechanism may contribute to this phenomenon.

In endothelial cells (ECs), a number of vasodilatory agonists, such as acetylcholine, bradykinin, and ATP, elevate cytosolic free Ca2+ as a result of intracellular release and external entry and further evoke a hyperpolarization of membrane potential (45). Several studies (14, 36, 37) have further demonstrated that these events were closely associated with the release of EDRF from isolated ECs. Collectively, such key observations led to the hypothesis that membrane hyperpolarization contributed to the agonist-induced production of EDRF by increasing the electrical driving force for Ca2+ influx. Observations of agonist-evoked changes in cytosolic free Ca2+ and membrane hyperpolarization in ECs from a number of vascular beds and species have indicated that these events are widespread (45) and are thus likely to be of physiological importance in endothelial function.

Through the use of electrophysiological recordings, pharmacological agents, and the detection of mRNA species, it is now evident that ECs isolated from various sources express Ca2+-activated K+ (KCa) channels, which are capable of producing membrane hyperpolarization in response to elevations of cytosolic [Ca2+]. The channel types commonly observed are small-conductance (SKCa) and intermediate-conductance KCa (IKCa) channels, whereas the expression of the large-conductance KCa (BKCa) channel appears to be more variable (1, 45). In isolated ECs, pharmacological inhibitors of SKCa channels [such as apamin (59)] and IKCa channels [i.e., charybdotoxin (ChTx) (59) and triarylmethane-34 (TRAM-34) (61)] have been shown to block agonist-induced K+ currents (4, 8, 30, 39, 49), and, in intact arteries, such blockers have been further reported to inhibit selectively non-NO/non-prostanoid or EDHF-induced relaxations (5, 12, 15, 17, 25).

Although a number of reports have highlighted a role for SKCa and IKCa channels in the phenomenon of EDHF-mediated vasorelaxation (for reviews, see Refs. 5 and 40), few studies have specifically examined the direct contribution of these same channels in agonist-evoked NO production. This has been largely due to the difficulty of simultaneously monitoring NO synthesis and functional responses (e.g., vasodilation, membrane hyperpolarization, and cytosolic Ca2+ transients) in a single preparation. Most recently, however, Stankevicius et al. (56) showed that blockade of SKCa and IKCa channels by apamin and ChTx, respectively, interfered with acetylcholine-induced NO production and vasorelaxation in rat mesenteric arteries, thereby establishing a functional role for SKCa and IKCa channels in NO-mediated vasodilation. In the present study, we sought to define mechanistically the functional role(s) of SKCa and IKCa channels in agonist-induced NO production. To do so, we utilized highly selective SKCa and IKCa channel inhibitors in combination with single-cell microfluorimetry and patch-clamp electrophysiology to examine directly agonist-induced NO synthesis, changes in cytosolic free [Ca2+], and membrane hyperpolarizations in single human vascular ECs. Using this strategy, we acquired well-resolved temporal and spatial data that reveal novel insights into the cellular mechanism underlying stimulated NO synthesis by Ca2+-mobilizing agonists and define the critical role of SKCa and IKCa channels in this process.


Cell culture and fluorescence measurements.

The EC line EA.hy926 (16), derived from the human umbilical vein [human umbilical vein ECs (HUVECs)], was cultured and loaded with the membrane-permeable forms of the fluorescent dyes 4-amino-5-methylamino-2′,7′-difluorofluorescein (DAF-FM) or fluo-3, as recently described (54). In our isolated EC preparations, agonist-stimulated increases in DAF-FM fluorescence were abolished in the presence of the NO synthase (NOS) inhibitor N-nitro-l-arginine methyl ester (0.1 mmol/l), consistent with the reported specificity of this fluorescent reporter (31). Fluorescence measurements were performed in a ∼0.3-ml bath chamber mounted on the stage of a Nikon TE300 inverted microscope equipped with a 75-W xenon arc lamp and SFX-1 microfluorimeter. Both DAF-FM and fluo-3 fluorescence signals were measured using excitation and emission band-pass filters centered on 488 and 520 nm, respectively; data were acquired using AxoScope software and analyzed with pCLAMP 7 and SigmaPlot software suites. As the fluorescent intensity of the triazole- or NO-bound form of DAF-FM originating from a single cell was typically quite modest, the strong excitation light needed to observe reliable fluorescent signals often resulted in some photobleaching of the NO-modified form of DAF-FM during continuous cell illumination. Exposure of the cell to intermittent illumination through the use of a timer-driven, optic shutter reduced but did not completely eliminate the photobleaching of NO-modified DAF-FM. A manually controlled diaphragm was used to restrict the region of light collection to the cell of interest.


Voltage- and current-clamp measurements were performed using perforated patch-clamp methodology in combination with an Axopatch 200B amplifier, Digidata 1200B analog-to-digital interface, and Clampex 7 software. Electrical signals recorded under current clamp and voltage clamp were typically sampled at 1 Hz and 5 KHz, respectively. Borosilicate glass micropipettes (2–4 MΩ tip resistance) were first briefly dipped into standard filling solution [final concentration (in mmol/l) 100 K-aspartate, 30 KCl, 1 MgCl2, 2 Na2-ATP, and 10 HEPES (pH 7.2) with 1 mol/l KOH] and then back filled with the same filling solution containing nystatin (50 mg/l final concentration). The bath solution for both fluorescence and electrophysiological recordings contained (in mmol/l) 135 NaCl, 5 KCl, 1 MgCl2, 1.5 CaCl2, and 10 HEPES (pH 7.4) with 1 mol/l NaOH. The high-KCl bath solution was prepared by an equimolar substitution of NaCl with KCl; for the Ca2+-free solution, CaCl2 was omitted and replaced by 2 mM EGTA. Cells in the bath chamber were constantly superfused at ∼1 ml/min, and solution changes were performed by gravity flow from a series of elevated solution reservoirs using manually controlled solenoid valves. All fluorescence and electrophysiological recordings were performed at 35°C.


Chemicals were purchased from Sigma-Aldrich (St. Louis, MO) and were of ACS grade or higher. DAF-FM diacetate and Fluo-3 AM were obtained from Molecular Probes (Eugene, OR). TRAM-34 was kindly provided by Dr. Heike Wulff (UC Davis).


In HUVECs loaded with the NO-sensitive fluorescent dye DAF-FM diacetate (31), histamine, or the purinergic agonist ATP evoked reproducible increases in cellular fluorescence under control conditions (Fig. 1A); however, agonist-evoked increases in fluorescence were inhibited in the presence of apamin and ChTx, blockers of SKCa and IKCa channels (59), respectively (Fig. 1, B and C). Brief exposure of dye-loaded cells to the direct NO donor sodium nitroprusside (SNP) at the end of each experiment demonstrated that apamin and ChTx did not interfere with DAF-FM activation and fluorescence in loaded cells. Exposure of cells to either apamin or ChTx alone produced only partial (20–40%) inhibition of agonist-evoked increases in DAF-FM fluorescence (data not shown). These observations thus demonstrate that both apamin and ChTx act directly on the endothelium to block agonist-evoked NO production. In contrast to the observed inhibitory effects of apamin + ChTx, agonist-induced NO production was unaffected by a bath application of tetraethylammonium (TEA; 10 mmol/l), which would be expected to block BKCa channels along with some types of voltage-gated K+ channels (i.e., Kv1) (Fig. 2A) (23, 44). We also observed only a very modest (∼10%) inhibition by TEA of agonist-evoked Ca2+ transients in single fluo-3-loaded HUVECs (see Supplemental Fig. 1).1 Collectively, these data are consistent with the rather low-affinity block by external TEA of both native and recombinant IKCa channels (IC50 value: 8–10 mmol/l) (2, 27, 35) along with the SKCa2 and SKCa3 channel isoforms detected in the vascular endothelium (IC50 values: ∼3 and ∼9 mmol/l, respectively) (4, 43). Similar to TEA, we also observed that evoked NO production was unaltered in the presence of 50 μmol/l BaCl2 and 100 μmol/l ouabain, which block inwardly rectifying K+ (Kir) channels (23) and Na+-K+-ATPase (29), respectively (Fig. 2B). Taken together, these data suggest that BKCa, Kv1, and Kir channels, along with Na+-K+-ATPase, do not functionally contribute to agonist-induced NO production in single HUVECs.

Fig. 1.

Agonist-stimulated nitric oxide (NO) production is inhibited by apamin (Apa) and charybdotoxin (ChTx). A: fluorescence tracing recorded from a single human umbilical vein endothelial cell (HUVEC) loaded with 4-amino-5-methylamino-2′,7′-difluorofluorescein (DAF-FM) dye. Exposure of the cell to either ATP (10 μmol/l) or histamine (10 μmol/l) is indicated by the horizontal bars above the tracing. The break in the recording indicates the ∼20-min control incubation period prior to reapplication of ATP and histamine. In B, a single HUVEC was exposed to 1 μmol/l Apa and 0.1 μmol/l ChTx for ∼20 min prior to a second application of ATP and histamine. The addition of sodium nitroprusside (SNP; 10 μmol/l) at the end of each experiment is denoted by the arrow beneath the fluorescence tracings. Changes in cellular DAF-FM fluorescence (ΔF/F0) in response to the first (control) and second agonist applications, in either the absence (incubation only) or presence of Apa and ChTx, are shown in C. Data are presented as means ± SE of fluorescence tracings recorded from 4–6 individual cells for each agonist under each condition.

Fig. 2.

Exposure to either tetraethylammonium (TEA) or BaCl2 and ouabain does not prevent agonist-induced NO production. Stimulation of single DAF-FM-loaded HUVECs by either ATP (10 μM) or histamine (10 μM) produced characteristic increases in cellular fluorescence. Following the addition of either 10 mM TEA (A) or 50 μM BaCl2 and 100 μM ouabain (B), the same cell was reexposed to first ATP and then histamine, as indicated by the horizontal bars. The histogram in C quantifies the agonist-induced DAF-FM fluorescence signals (ΔF/F0) for both ATP and histamine in the absence and presence of either TEA or BaCl2 and ouabain. Data are presented as means ± SE of fluorescence tracings from 4 individual cells for each agonist under each condition.

To characterize the Ca2+ dependence of agonist-evoked NO production, intracellular endoplasmic reticulum (ER) Ca2+ stores were disrupted by exposure to the sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA) pump inhibitor thapsigargin (TG). Following initial NO responses to ATP and histamine, rapid application of 0.5 μM TG abolished further increases in NO upon reexposure to either agonist (Fig. 3, A and C). However, brief application of SNP to TG-treated cells still produced large fluorescence signals, indicating that TG did not compromise the NO sensitivity of DAF-FM in the intact cell. As cells were typically exposed to TG immediately following agonist washout, intracellular Ca2+ stores may not have been sufficiently full to support a TG-evoked NO response. In separate experiments, however, we observed that an application of 0.5 μM TG alone produced a significant increase in DAF-FM fluorescence (data not shown), as recently reported (13). The contribution of external Ca2+ to agonist-evoked NO production was examined in DAF-FM-loaded cells by brief exposure to a bath solution containing 0 Ca2+-2 mM EGTA. Rapid removal of external Ca2+ abolished ATP- and histamine-induced NO production' however, reintroduction of the external solution containing 1.5 mmol/l Ca2+ completely restored agonist-evoked NO synthesis (Fig. 3, B and C). This latter finding is thus consistent with previous results showing that stimulated EDRF/NO production is inhibited in the absence of external Ca2+ (13, 26, 3234, 37).

Fig. 3.

Interfering with intracellular Ca2+ release or preventing external Ca2+ entry blocks agonist-induced NO production. A: fluorescence tracing recorded from a single DAF-FM-loaded HUVEC exposed to either 10 μmol/l ATP or 10 μmol/l histamine, as indicated by the horizontal bars above the tracing. Continuous application of 0.5 μmol/l thapsigargin (TG) is shown by the long horizontal bar. The arrow at the end of the tracing indicates the addition of the NO donor SNP (10 μmol/l). B: changes in DAF-FM fluorescence in a single dye-loaded HUVEC in response to either ATP or histamine. Exposure of the cell to an external saline solution containing 2 mmol/l EGTA and no added Ca2+ is indicated by the horizontal bar. Following washout of the EGTA-containing solution and return to physiological saline solution, the same cell was reexposed to both ATP and histamine. The effects of TG exposure or external Ca2+ removal on agonist-evoked DAF-FM fluorescence signals (ΔF/F0) are quantified in the histogram shown in C. Data are presented as means ± SE for each agonist; control responses to either agonist were recorded from 8–9 individual cells, whereas agonist-evoked responses in the presence of either TG or EGTA were taken from 4–5 cells under each condition.

While the above data indicate that intra- and extracellular Ca2+, along with SKCa and IKCa channels, play critical roles in agonist-stimulated NO synthesis, exactly how these components are interrelated temporally and mechanistically is unclear. To establish such relations, we carried out dual recordings of membrane potential and either cytosolic free Ca2+ or NO production using microfluorimetry in single patch-clamped HUVECs. In fluo-3-loaded cells, ATP and histamine evoked rapid elevations in cytosolic free Ca2+ and membrane hyperpolarizations that reversed upon agonist washout (Fig. 4). Following exposure of the same cell to apamin and the highly selective IKCa channel blocker TRAM-34 (61), agonist-evoked membrane hyperpolarizations were abolished and changes in fluo-3 fluorescence were significantly reduced, as shown in Fig. 4B. This latter finding is thus consistent with earlier studies demonstrating that agonist-evoked Ca2+ transients in isolated vascular ECs are decreased with membrane depolarization (9, 33, 51). In parallel experiments, ATP- and histamine-evoked increases in DAF-FM fluorescence were also found to be closely associated with membrane hyperpolarizations (Fig. 5). The addition of apamin and TRAM-34 to the same DAF-FM-loaded cell inhibited both of these responses upon reexposure to ATP and histamine. Interestingly, transient membrane depolarizations were typically observed upon agonist withdrawal (Figs. 4A and 5A), which may be due to Ca2+-dependent Cl channel activity, as previously reported (48).

Fig. 4.

Blockers of small-conductance (SKCa) and intermediate-conductance Ca2+-activated K+ (IKCa) channels interfere with agonist-stimulated increases in cytosolic free Ca2+ and transient membrane hyperpolarizations. Single HUVECs loaded with the Ca2+-sensitive dye fluo-3 were held under current clamp using a nystatin-permeabilized patch-clamp method. A: simultaneous and continuous recordings of fluo-3 fluorescence (top trace) and membrane potential (Vm; bottom trace) from a single cell. Brief exposure of the cell to either 10 μmol/l ATP or 10 μmol/l histamine is indicated by the bars above the fluorescence tracing. Continuous exposure of the cell to 1 μmol/l Apa and 1 μmol/l triarylmethane (TRAM)-34 is shown by the long horizontal bar. Agonist-stimulated changes in both fluo-3 fluorescence (ΔF/F0) and VmVm) in the absence and presence of Apa and TRAM-34 are quantified by the top and bottom axes, respectively, of the histogram shown in B. Data are presented as means ± SE of simultaneous recordings from 5 individual cells for each agonist.

Fig. 5.

Blockade of SKCa and IKCa channels prevents agonist-stimulated increases in NO production and transient membrane hyperpolarizations. Single HUVECs loaded with DAF-FM dye were held under current clamp using a nystatin-permeabilized patch clamp method. A: simultaneous and continuous recordings of DAF-FM fluorescence (top trace) and Vm (bottom trace) from a single cell. Brief exposure of the cell to either ATP (10 μmol/l) or histamine (10 μmol/l) is indicated by the bars above the upper tracing. Continuous exposure of the cell to 1 μmol/L Apa and 1 μmol/l TRAM-34 is shown by the long horizontal bar. Exposure of the cell to the NO donor SNP (10 μmol/l) at the end of the experiment is shown by the arrow beneath the Vm trace. Agonist-stimulated changes in both cellular DAF-FM fluorescence (ΔF/F0) and ΔVm in the absence and presence of Apa and TRAM-34 are quantified in top and bottom axes, respectively, of the histogram shown in B. Data are presented as means ± SE of recordings taken from 5 individual cells for each agonist.

A major benefit of carrying out such dual recordings of agonist-evoked membrane hyperpolarizations together with either Ca2+ or NO signals is that we were able to determine the precise temporal pattern among these three cellular events, which has not been previously described. In response to either ATP or histamine, increases in DAF-FM fluorescence lagged behind the onset of membrane hyperpolarization by 8–12 s (see Fig. 6A); in the case of TG-induced NO production, this delay was typically only 2–3 s (Table 1). In contrast, agonist-evoked increases in fluo-3 fluorescence typically preceded membrane hyperpolarization by 2–5 s (Fig. 6B). In agreement with our observations, Li et al. (62) have shown using simultaneous fura-2 and DAF-2 fluorescence measurements that the onset of agonist-evoked increases in cytosolic Ca2+ preceded NO production in the coronary artery endothelium. These data sets thus define a distinct temporal pattern in which an agonist-evoked increase in cytosolic [Ca2+] is followed by membrane hyperpolarization, which is then followed by NO synthesis.

Fig. 6.

Kinetic relations between agonist-stimulated changes in Vm and cytosolic free Ca2+ and NO production. Simultaneous recordings of Vm and either DAF-FM or fluo-3 fluorescence from a single HUVEC are shown in A and B, respectively. Temporal differences between the onsets of agonist-evoked hyperpolarization (denoted by the vertical dashed line marked t1) and stimulated increases in either the DAF-FM or fluo-3 fluorescence signals (denoted by the vertical dashed line marked t2) were quantified and are shown in Table 1.

View this table:
Table 1.

Temporal relations between the onset of stimulus-driven membrane hyperpolarization and increases in either DAF-FM or fluo-3 cellular fluorescence induced by ATP, histamine, or thapsigargin

To confirm that apamin and TRAM-34 were indeed acting to block endothelial KCa channels, whole cell membrane currents were recorded from single HUVECs stimulated by ATP in the absence and presence of apamin and TRAM-34. Under basal conditions, single HUVECs displayed a modest outwardly rectifying macroscopic current that typically reversed near −40 mV (Fig. 7). ATP increased the magnitude of the outward current and shifted the reversal potential to values near −70 mV, consistent with the activation of membrane K+ channels. This ATP-stimulated outward current was largely inhibited in the presence of apamin and TRAM-34 (Fig. 7, inset).

Fig. 7.

Sensitivity of ATP-activated membrane currents (I) to Apa and TRAM-34. Whole cell currents were recorded from single voltage-clamped HUVECs in response to a 50-ms voltage ramp (−100 to +50 mV) using a nystatin-perforated patch-clamp technique. Membrane currents were first recorded under basal conditions and then in response to 10 μM ATP, followed by 10 μM ATP + 1 μM TRAM-34 and 1 μM Apa. The mean percentage increases (±SE) in current magnitude above baseline evoked by ATP in the absence and presence of Apa and TRAM-34 are shown in the inset. Current tracings represent the averages of 3–4 single sweeps under each condition and are representative of 5 similar experiments.

As the above data pointed to a critical role for membrane hyperpolarization in agonist-stimulated NO production, we exposed single DAF-FM-loaded HUVECs to a high KCl-containing bath solution to “clamp” the membrane voltage near 0 mV and prevent agonist-induced hyperpolarization. As expected, brief exposure to 80 mmol/l external K+ depolarized the endothelial membrane potential and abolished agonist-stimulated NO synthesis (Fig. 8). Upon washout and return to 5 mmol/l external K+, membrane potential recovered to the control level, and both ATP and histamine induced typical membrane hyperpolarizations that were associated with increases in DAF-FM fluorescence. If we further hypothesize that NO production evoked by Ca2+-mobilizing agonists is absolutely dependent on membrane hyperpolarization, it should then be possible to regulate agonist-evoked NO production in real time by directly controlling endothelial membrane potential. To test this possibility, single DAF-FM-loaded HUVECs were voltage clamped at potentials ranging from 0 to −80 mV, and ATP was then briefly applied at each potential. In agreement with the above hypothesis, the magnitude of agonist-evoked NO production was observed to increase at increasingly negative membrane potentials, reaching a maximum between −60 and −80 mV (Fig. 9). These data thus provide the first direct real-time demonstration that evoked membrane hyperpolarization acts as a critical, rate-limiting factor for stimulated NO production by Ca2+-mobilizing agonists in single vascular ECs.

Fig. 8.

Elevation of extracellular KCl prevents agonist-induced NO production. Single HUVECs loaded with DAF-FM dye were held under current clamp as described in Fig. 7. A: simultaneous and continuous recordings of DAF-FM fluorescence (top trace) and Vm (bottom trace) from a single cell. ATP or histamine was briefly applied to the cell, as indicated by the horizontal bars, in either the presence or absence of 80 mmol/l external KCl. Agonist-induced elevations in DAF-FM fluorescence (ΔF/F0) under both normal and elevated KCl conditions are quantified in B. Mean data (±SE) represent recordings from 4 individual cells for each agonist.

Fig. 9.

The magnitude of agonist-induced NO synthesis is directly dependent on Vm. Single HUVECs loaded with DAF-FM dye were voltage clamped via a nystatin-perforated patch and held at Vms ranging from 0 to −80 mV, as denoted by the step-wise series of horizontal bars (A). At each test potential, the voltage-clamped cell was briefly exposed to 10 μmol/l ATP as indicated by the horizontal bars. The data tracing beneath the stimulation protocol represent the continuous DAF-FM fluorescence signal simultaneously recorded from the single voltage-clamped cell at each Vm and in response to each application of ATP. B: quantification of the ATP-induced elevation in DAF-FM fluorescence (ΔF/F0) observed at each Vm. Results are expressed as means ± SE of fluorescence signals recorded from 5 voltage-clamped cells.


The release of NO from vascular ECs in response to vasorelaxant hormones, such as acetylcholine and histamine, is known to be affected by changes in intracellular and external Ca2+ levels along with membrane potential. To date, however, the precise mechanistic and temporal patterns linking these three cellular parameters have not been rigorously established by means of direct experimental measurements. In the present study, we utilized nystatin-perforated patch-clamp recordings in combination with single-cell microfluorimetry to monitor directly, in real time, agonist-stimulated membrane hyperpolarization, NO synthesis, and cytosolic Ca2+ transients in single HUVECs. In addition, we specifically addressed the contributions of endothelial SKCa and IKCa channels to these events by using the selective inhibitors apamin and ChTx/TRAM-34, respectively (59, 61). In doing so, this study provides novel experimental insights that define the mechanistic role of KCa channels in hormone-stimulated NO production at the level of a single EC.

Our fundamental observation that acute, agonist-evoked NO production is blocked by a combination of apamin and either ChTx or TRAM-34 in DAF-FM-loaded HUVECs (Figs. 1 and 5) reveals two novel and important insights. First, it implicates a critical role for endothelial SKCa and IKCa channels in evoked NO synthesis, which is consistent with functional data showing that apamin + ChTx interferes with agonist-stimulated, endothelium-dependent vasorelaxation (11, 56). Second, it strongly suggests that manipulations designed to decrease NO production by disrupting intracellular Ca2+ levels will interfere directly with the activation of SKCa and IKCa channels, which would be expected to mimic the toxin-induced inhibition of NO production shown in Fig. 1. In line with this latter point, we observed that prior depletion of ER stores by TG inhibited agonist-evoked NO production, whereas transient removal of external Ca2+ produced a similar effect (Fig. 2), as reported earlier by investigators using indirect measurements of EDRF release (33, 37). Although endothelial NOS (eNOS) itself may be sensitive to manipulations of intracellular free Ca2+, these observations are also consistent with earlier data showing that both intracellular release and external Ca2+ entry strongly influence the magnitude and/or duration of evoked Ca2+ transients (45) along with the duration of membrane hyperpolarization (3, 10) in stimulated ECs. In an elegant study using fluorescent probes to report simultaneous changes in cytosolic free Ca2+ and NO production, Isshiki et al. (26) demonstrated that agonist-evoked NO synthesis is strongly dependent on external Ca2+ entry and largely insensitive to Ca2+ released from intracellular stores in bovine aortic ECs. Taken together, these data highlight and contrast the functional roles of both intracellular Ca2+ release and Ca2+ entry for NO production evoked by Ca2+-mobilizing agonists.

While the above findings imply critical roles for intra- and extracellular Ca2+ and SKCa and IKCa channels in stimulated NO synthesis, they do not establish a precise temporal relation among these three parameters. To define such a pattern, we performed simultaneous recordings of membrane potential and either fluo-3 or DAF-FM fluorescence in single HUVECs (Figs. 4 and 5). The results of this approach demonstrated that agonist-evoked increases in both cytosolic Ca2+ levels and NO synthesis were tightly associated with transient membrane hyperpolarizations, such that membrane hyperpolarization closely followed cytosolic Ca2+ elevations but preceded increases in agonist-evoked NO production (Fig. 6 and Table 1). This temporal pattern thus establishes membrane hyperpolarization as an essential intermediate step in stimulated NO production. Importantly, apamin and the highly selective IKCa blocker TRAM-34 (61) abolished both agonist-evoked membrane hyperpolarization and increases in NO synthesis and significantly reduced elevations in cytosolic free Ca2+. The modest Ca2+ transient remaining in the presence of apamin and TRAM-34 likely reflects the combination of Ca2+ release from intracellular stores and the residual entry of external Ca2+. Our observation that apamin and TRAM-34 inhibited an agonist-evoked, outwardly rectifying current in single HUVECs (Fig. 7) is further consistent with the activation of SKCa and IKCa channels in ECs and agrees with the reported presence of these channels in the vascular endothelium (4, 8, 22, 39, 45, 49, 50).

If agonist-evoked membrane hyperpolarization is truly an essential upstream event regulating NO synthesis, then preventing hyperpolarization by means other than blockade of SKCa and IKCa channels would also be expected to interfere with NO synthesis. As shown in Fig. 8, “clamping” membrane potential to ∼0 mV by a brief exposure to high external KCl blocked both agonist-induced membrane hyperpolarization and NO synthesis. This finding is thus consistent with the above prediction and provides a direct link between membrane potential and NO synthesis in a single EC. High external KCl has been reported previously to reduce EDRF release from populations of agonist-stimulated ECs (36), whereas Stankevicius et al. (56) recently demonstrated a similar inhibition of stimulated NO production by 80 mmol/l KCl in the rat mesenteric artery. Based on such data, we hypothesized that endothelial membrane hyperpolarization represents a critical, rate-limiting process regulating NO synthesis by Ca2+-mobilizing stimuli. By using voltage clamp to accurately control endothelial membrane potential in DAF-FM-loaded HUVECs, we observed that the magnitude of agonist-stimulated NO synthesis increased in a linear manner with the degree of membrane hyperpolarization between 0 and −80 mV (Fig. 9). This singular result thus establishes a direct quantitative relation between membrane potential and stimulated NO synthesis at the level of a single EC. In related experiments, it has already been reported that the amplitudes of agonist-evoked Ca2+ transients in isolated ECs are lower at more depolarized membrane potentials (6, 9, 33, 51, 60) and that such changes in Ca2+ transients appear to be linearly related to membrane voltage over the range of −80 to +40 mV (28, 53).

Mechanistically, our data suggest that NO synthesis evoked by Ca2+-mobilizing stimuli can be described by a pathway of discrete cellular events that feed forward in a positive manner (Fig. 10). While such a model incorporates many of the key findings reported in previous studies, it makes two important distinctions critical to stimulated NO production. First, agonist-mediated release of intracellular Ca2+ stores triggers not only the Ca2+-dependent activation of SKCa and IKCa channels (38, 52) but also initiates the entry of external Ca2+, which is primarily responsible for eNOS activation. Second, stimulus-evoked membrane hyperpolarization, mainly via SKCa and IKCa channels and possibly BKCa channels (45), is absolutely required for stimulated NO synthesis, likely due to its influence on external Ca2+ entry. It is noteworthy, however, that in intact arterial preparations, agonist-evoked elevations in EC cytosolic Ca2+ are reported to be unaltered in the presence of apamin and ChTx/TRAM-34 (21, 41, 56). Although the reason(s) behind such observations is unclear at present, it is possible that there may be unrecognized differences in the dynamics and/or detection of intracellular Ca2+ transients in isolated ECs versus cells present in an intact endothelial layer. As shown by our data (Fig. 3) and earlier results (26, 32, 37), stimulated Ca2+ entry, rather than agonist-induced Ca2+ release from ER stores, appears to be principally responsible for eNOS activation and further influences the duration of agonist-induced membrane hyperpolarization (3, 42). Our model thus distinguishes ER store Ca2+ release as the “trigger” that initiates both the opening of KCa channels and the opening of store-operated Ca2+ entry channels [i.e., transient receptor potential (TRP) channels] in the plasma membrane. The ensuing membrane hyperpolarization acts to support Ca2+ entry, which would be influenced by external [Ca2+] and the magnitude of the hyperpolarizing event. Based on the above observations, we can further suggest that these two sources of mobilized Ca2+ carry out distinct and largely noninterchangeable roles in the multistep process of agonist-evoked NO synthesis. Differences in the spatial distribution of Ca2+-sensitive molecules may further contribute to the greater efficiency of eNOS activation by Ca2+ entry compared with release from intracellular stores; for example, both eNOS and TRP channels are reported to be colocalized in membrane caveolae (19, 46).

Fig. 10.

Model summarizing the relations between agonist-induced changes in cytosolic free Ca2+, Vm, and NO production in a single HUVEC. Hormonal activation of a G protein-coupled receptor causes the generation of inositol (1,4,5)-trisphosphate and the release of intracellular Ca2+ stores [endoplasmic reticulum (ER)] (step 1). Store release elevates cytosolic [Ca2+] (step 2) and further triggers store-operated channel (SOC)-mediated Ca2+ influx. Increased cytosolic [Ca2+] results in the activation of SKCa and IKCa channels (step 3), and the ensuing membrane hyperpolarization increases Ca2+ influx via SOCs (step 4). Enhanced Ca2+ influx is critical for stimulating adequate NO production by membrane-associated endothelial NO synthase (eNOS) (step 5) and further promotes KCa channel activity. Blockade of SKCa and IKCa channel-mediated membrane hyperpolarization by Apa and ChTx/TRAM-34, respectively, reduces store-operated Ca2+ influx, thereby preventing eNOS activation. CaM, calmodulin.

Finally, can we rationalize why membrane potential has such a dominant influence on Ca2+-dependent eNOS activation? Although membrane hyperpolarization would increase the electrical driving force for agonist-induced Ca2+ entry (6, 7, 9, 24, 36, 51), the degree of change in driving force (i.e., +160 mV under rest to +180 mV; see Fig. 4) based on Nernst-type calculations would be modest. More recently, however, the possibility that Ca2+-permeable TRP channels themselves may display voltage-dependent gating has been discussed (47). For the inwardly rectifying, Ca2+-permeable TRP channels in the endothelium (46), such voltage sensitivity would mean that TRP channel open probability and Ca2+ entry would be exponentially related to changes in membrane voltage. Current-voltage relations observed for agonist-evoked inward current through some Ca2+-permeable TRP channels (57) and native endothelial store-operated Ca2+ channels (18) appear consistent with such a possibility. As a result of such an exponential relationship, a modest hyperpolarization of −20 mV could produce a significantly larger increase in Ca2+ influx compared with a strictly linear relation predicted by a modest increase in the electrical driving force for external Ca2+ alone. Furthermore, the position along the voltage axis of the rectifying current-voltage relation for stimulated Ca2+ entry relative to the cell's resting potential may also influence the magnitude of hyperpolarization-induced Ca2+ influx.

In summary, our data indicate that the activation of SKCa and IKCa channels, leading to membrane hyperpolarization, represents an essential early event in the cellular pathway underlying agonist-stimulated NO production. Consistent with this conclusion, genetic knockout of either the endothelial SKCa3 or IKCa channels in mice gives rise to systemic hypertension and reduces hormone-induced, endothelium-dependent vasorelaxation (55, 58). Very recent observations demonstrating that apamin- and ChTx-sensitive KCa channels regulate acetylcholine-evoked NO production in the intact rat superior mesenteric artery (56) are further consistent with our data and strongly suggest that the mechanistic insights described in our study are relevant to the native vascular endothelium. Finally, the inhibition of agonist-evoked NO synthesis by apamin and ChTx/TRAM-34 observed in this study and by Stankevicius et al. (56) demonstrate a critical functional role for SKCa and IKCa channels in the cellular mechanisms underlying hormone-induced, NO-dependent vasorelaxation.


This work was supported by an operating grant from the Canadian Institutes of Health Research (to A. P. Braun). A senior research scholarship from the Alberta Heritage Foundation for Medical Research (to A. P. Braun) is gratefully acknowledged.


The authors thank Drs. Bill Cole, Rodger Loutzenhiser, and Pierre-Yves von der Weid (University of Calgary) and Drs. Michael Davis and Michael Hill (University of Missouri) for the insightful comments on this manuscript.


  • 1 Supplemental material for this article is available online at the American Journal of Physiology-Cell Physiology website.

  • 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.


View Abstract