Submaximal stimulation of porcine endothelial cells causes focal Ca²⁺ elevation beneath the cell membrane

Materials

BODIPY FL-X ryanodine, Calcium Green-1 acetoxymethyl ester form (AM), Calcium Green-1 hexapotassium salt, Calcium Crimson, 3,3′-dihexyloxacarbocyanine iodine (DiOC6), 5-(and 6)-carboxy SNARF-1® AM, (4-trimethylammoniumphenyl)-6-phenyl-1,3,5-hexatriene p-toluenesulphonate (TMA-DPH), and sodium-binding benzofura isophthalate acetoxymethyl ester (SBFI AM) were obtained from Molecular Probes. Fura-2 AM and fura-2 pentapotassium salt were purchased from Lambda Fluorescence Technologies, Graz, Austria and FFP-18 was from Calbiochem, Vienna, Austria. Cell culture chemicals were obtained from Life Technologies, Vienna and fetal calf serum (FCS, premium quality) was from PAA Laboratories, Linz, Austria. Hepes acid was from Amresco, Vienna and buffer salts were from Merck, Darmstadt. All other materials were from Sigma.

Cell culture

Endothelial cells were isolated from porcine coronary arteries as described previously (Graier, Simecek & Sturek, 1995). Briefly, pigs were killed in the local slaughterhouse by electrical paralysis and were exsanguinated by cutting both carotae instantly followed by an extraction of the heart, which was put on ice and immediately shipped to the laboratory. Vessels were isolated within 30 min after animal death and were incubated at 37°C with Dulbecco's minimal essential medium (DMEM) containing 200 U ml⁻¹ collagenase (type II, Sigma) and trypsin inhibitor (soybean type I, Sigma; 1 mg ml⁻¹) plus dilutions of (v/v) 0.02 MEM essential amino acids (Gibco BRL), 0.01 MEM non-essential amino acids (Gibco BRL) and 0.01 MEM vitamins (Gibco BRL). Isolated cells were immediately used for experiments or cells were cultured up to passage 1 in Opti-MEM containing 3 % fetal calf serum. Cell culture purity was determined by the typical cobblestone morphology and absence of immunofluorescence detection of smooth muscle α-actin. Endothelial cell cultures were > 99 % pure. All experiments were performed with freshly isolated cells or with cells from first passage (∼10 days after isolation) or with cells from primary culture.

Measurement of inositol phosphates

Generation of inositol phosphates was measured as described previously (Graier, Schmidt & Kukovetz, 1991). Briefly, endothelial cells were cultured in six-well plastic plates and were loaded at 48 h with 5 μCi ml⁻¹ myo-[³H]inositol in DMEM. Before the experiment the monolayer was washed three times with Hepes buffer. Endothelial cells were stimulated with the agonist and experiments were stopped at times indicated by removal of the incubation buffer and addition of 1 ml 0.01 N HCl. Inositol 1,4,5-trisphosphate was separated using HPLC as described previously (Graier et al. 1991). As indicated formation of IP₃ was also measured using a commercially available radioactive binding assay (Biotrak, Amersham International).

Ca²⁺ measurement

Intracellular free Ca²⁺ concentration ([Ca²⁺]_i) was determined using Ca²⁺ dye fluorescence techniques for fura-2 (Graier, Simecek & Sturek, 1995), Calcium Green-1 and FFP-18. Different Ca²⁺-sensitive dyes were used for monitoring cytosolic and bulk free Ca²⁺ concentration ([Ca²⁺]_b; fura-2 and/or Calcium Green-1) and the subplasmalemmal Ca²⁺ concentration ([Ca²⁺]_sp; FFP-18). Before dye loading, cultured cells were harvested by treatment for 1 min with trypsin (Sigma; 0.02 % in phosphate-buffered solution plus 1 mmol l⁻¹ EDTA), centrifuged and resuspended in DMEM. Freshly isolated cells were centrifuged and resuspended in DMEM. For cell loading with the membrane-permeable dyes fura-2 AM and Calcium Green-1 AM, cells were incubated in the dark with 2 μmol l⁻¹ fura-2 AM and/or 2 μmol l⁻¹ Calcium Green-1 AM in DMEM. After 45 min at room temperature, cells were centrifuged, washed twice and resuspended in Hepes-buffered solution with or without 2.5 mmol l⁻¹ CaCl₂ containing (mmol l⁻¹): 145 NaCl, 5 KCl, 1 MgCl₂, 10 Hepes acid, pH 7.4. When low-Na⁺ solution was used, only 19 mmol l⁻¹ NaCl was added and 126 mmol l⁻¹ NaCl was replaced by 126 mmol l⁻¹ choline chloride while no other ion concentration was changed.

Loading of cell membrane-impermeable dyes by laser-generated stress wave

While fura-2 and Calcium Green-1 were available as membrane-permeable esters, reversible permeabilization of the cell membranes was necessary to load the cells with FFP-18. Endothelial cells were suspended in DMEM without Phenol Red at a concentration of approximately 0.9 × 10⁶ cells ml⁻¹. Under constant shaking, FFP-18 (from a 1 mmol l⁻¹ stock solution in 1:1 DMSO-H₂O₂, final concentration 2 μmol l⁻¹) and/or Calcium Green-1 (from a 2 mmol l⁻¹ stock solution in DMSO, final concentration 5 μmol l⁻¹) were added. The cell suspension (350 μl) was transferred to one microwell of a ninety-six-well plate (Falcone, Linz, Austria). The microwell was sealed with 1.25 mm thick aluminum foil which was in contact with the whole surface of the cell suspension covering. A laser-induced stress wave was generated by three to five shots of an Nd:YAG laser (70 mJ; Lumonics, Rugby, UK) to the metal foil, focused 0.5 cm behind the foil. Cells were equilibrated for 10-30 min in the dark at room temperature, centrifuged and resuspended in Hepes buffer (free [Ca²⁺] < 1 μmol l⁻¹). This procedure allows the loading of fura-2, Calcium Green-1 and FFP-18 into endothelial cells. Cell viability tests by Trypan Blue and the Schärfe cell counter (Casy-1®, Schärfe System Gmbh, Reutlingen, Germany) demonstrated 2 % dead cells after the laser pulse-wave loading procedure, while no changes were found in cell size and cell proliferation after re-seeding after laser pulse wave treatment. Membrane fluidity, measured by the fluorescence polarization technique using 1-TMA-DPH, was decreased by 27 % just after the laser pulse wave treatment and became normalized within 5 min (data not shown). Intracellular pH measured with SNARF-1 (laser pulse wave loading in the presence of 2 μmol l⁻¹ SNARF-1 as well as conventional loading using 10 μmol l⁻¹ SNARF-1 AM for 45 min followed by an in situ calibration of basal pH levels in the presence of 10 μmol l⁻¹ nigericin) was not affected by laser pulse wave treatment (data not shown). In contrast, the intracellular Na⁺ concentration was slightly increased by laser pulse wave treatment from 5.6 ± 1.9 mmol l⁻¹ (before laser pulse wave treatment using the conventional SBFI AM loading procedure; n = 4) to 7.6 ± 0.5 mmol l⁻¹ (after SBFI loading by laser pulse wave loading; n = 5, n.s. vs. prior laser pulse wave treatment). In situ calibration of intracellular [Na⁺] was performed by elevation of extracellular [Na⁺] in the presence of 2 μmol l⁻¹ gramicidin, 100 μmol l⁻¹ ouabain and 40 μmol l⁻¹ monensin. No differences in intracellular [Na⁺] were obtained when cells were loaded conventionally with SBFI AM before and after laser pulse wave treatment (data not shown).

In respect of the endothelial Ca²⁺ signalling, laser-pulse wave loading did not affect basal Ca²⁺ concentration even in the presence of extracellular Ca²⁺. The kinetics of the fura-2 and Calcium Green-1 signals monitored after conventional ‘AM’ loading is similar to that measured in cells loaded with free acids of these dyes by the laser-pulse loading procedure. When FFP-18-loaded cells were used, cells were equilibrated for 20 min and experiments were performed within 35 min after loading.

Data collection

[Ca²⁺]_i was monitored every 0.25 s in monolayer or suspended cells in thermostatically controlled spectrofluorometers (Hitachi F-4500, Hitachi F-2000 and Perkin Elmer LS50-B/FFA). In single-cell recordings using fura-2, cells were excited alternately every 50 ms with 360 and 380 nm light and emission light was detected at 510 nm using the photon counting technique as previously described (Graier et al. 1995). Counts were converted to analog by an optical processor circuit (Sturek, Smith & Stehno-Bittle, 1991) and wavelengths were registered by a PC running AxoBASIC 1.0 (Axon Instruments). Changes in Ca²⁺ concentration were monitored by measuring the ratio of 340/380 nm excitation and 510 nm emission for fura-2, the fluorescence intensity at 506 nm excitation and 534 nm emission for Calcium Green-1 and the ratio 335/375 nm excitation and 490/502 nm emission for FFP-18. Simultaneous recordings of fura-2 and Calcium Green-1 or FFP-18 and Calcium Green-1 were performed using a four-wavelength spectrofluorometer where each fluorescence was recorded every 0.8 s.

Data acquisition

Due to the failure of the calibration techniques reported by many groups (Sturek et al. 1991; Graier et al. 1995; Wang, Lau, Li, Yoshikawa & van Breemen, 1995), [Ca²⁺]_i in each experiment is expressed as the ratio 340/380 nm excitation at 510 nm emission (F₃₄₀/F₃₈₀) for fura-2, the ratio 335/375 nm excitation at 490/502 nm emission (F₃₃₅/F₃₇₅) for FFP-18 and the fluorescence at 506 nm excitation and 534 nm emission (F534) for Calcium Green-1. Within the instrumental setups used, there are differences in fluorescence intensity (I, where I =εI_oQYK) based on variations in excitation light intensity (I_o) and instrumental constants (K, photomultiplier sensitivity, etc.), while ε(molar extinction coefficient) and the quantum yield of the dyes (QY) are constant. Thus, all results demonstrated in a given figure represent data collected on the same instrumental setup and can be compared, while caution is necessary when comparing the results within separate figures where different instrumental setups were used.

Confocal microscopy

A point scanning Bio-Rad MRC-600 with a krypton-argon dual laser permitted simultaneous confocal imaging with the two excitation wavelengths, 488 and 568 nm. After dye loading (10⁻⁷ to 10⁻⁶ mol l⁻¹ for 10 min at 37°C in the dark), cells were washed twice and intracellular dyes were excited at 488 nm and the emitted fluorescence of BODIPY FL-X ryanodine (loaded in the presence of 1 μmol l⁻¹ for 10-15 min), BODIPY FL thapsigargin (loaded in the presence of 0.5-1 μmol l⁻¹ for 10 min) and DiOC6 (loaded in the presence of 250 nmol l⁻¹ for 10-15 min) was monitored at 510 nm emission by one photomuliplier, while the transmitted light image was monitored in a separate channel by a second photomultiplier. Images were collected at intervals of 1 μm on the z-axis set by a software-controlled stage focus stepper motor. The confocal pinhole aperture was set to give a 0.5 μm thick section. The confocality of the system was verified by studying fluorescent microspheres with a diameter of 1 μm (Molecular Probes).

Deconvolution microscopy

The deconvolution microscope provided in combination with a scientific-grade CCD camera, three-dimensional images using conventional lamp illumination based on computer algorithms to calculate the out-of-focus light by a three-dimensional point spread function for each microscope objective lens (Waters & Brown, 1996). Dyes were excited using 488 nm excitation and emission was monitored at 510 nm using a liquid-cooled CCD camera (Photometrics, Inc., Tuscon, AZ, USA) on an inverted microscope (Nikon Diaphote Inc.). Images were collected at 1 μm intervals in the z-axis set by a stage focus stepper motor (Ludl Ltd, Hawthorne, NY, USA) controlled by the Ludl-Z-stage control box connected with Micro Tome deconvolution software (VayTek, Inc., Fairfield, IA, USA). Confocal images were calculated using Micro Tome deconvolution software using an advanced constrained iterative deconvolution algorithm.

Three-dimensional reconstruction and quantitative analysis of focal data

The three-dimensional reconstruction for data obtained with the confocal and deconvolution microscope was conducted with VoxBlast (VayTek, Inc.), a volume rendering program that utilizes a back-propagation-weighted average algorithm. Out-of-focus fluorescence was minimized for accurate three-dimensional reconstruction of the slices by testing the system on 1 μm microspheres. Refractive index mismatch between the aqueous cell and bathing solution and the oil immersion objective which stretches images in the z-axis nearly threefold (Visser & Oud, 1994) was corrected. For quantitative analysis of central vs. peripheral distribution of the dyes, only images with clearly defined plasmalemmal borders at the middle depth of the cell were used. Central vs. peripheral distribution ratios have provided an excellent quantitative index for the distribution of fluorophores (Liou & Morgan, 1994). Plasmalemma was excellently delineated using extracellular membrane-impermeable Ca²⁺ crimson (Sturek, 1995). Transmitted light images were binarized to facilitate the use of a semiautomatic edge tracking function to define the plasmalemma of the cell (Image-Pro plus, Media Cybernetics, Silver Spring, MD, USA). The peripheral area was then defined by eroding the plasmalemmal border by 3 pixels, corresponding to a distance of 0.6 μm from the plasmalemma towards the centre of the cell. To calculate the peripheral/central distribution ratio the density (pixel intensity/area) was used to normalize for different areas.

Statistics

Data points represent the means ±s.e.m. Analysis of variance (ANOVA) was used for data evaluation and statistical significance of differences between means was estimated by Scheffe's F test. The level of significance was defined as P < 0.05.

Sodium fluoride increases inositol 1,4,5-trisphosphate level without any effect on bulk Ca²⁺ concentration

Stimulation of endothelial cells with sodium fluoride (10 mmol l⁻¹) yielded an increase in inositol 1,4,5-trisphosphate (IP₃) levels to 339 ± 17 % of the initial level within 7 min of stimulation (Fig. 1A; n = 6). Lowering the extracellular Na⁺ concentration to 19 mmol l⁻¹ did not affect NaF-induced IP₃ formation (data not shown). The kinetics of this NaF-stimulated IP₃ elevation significantly differed from those observed for bradykinin (1 μmol l⁻¹ bradykinin), which yielded rapid but transient elevation of IP₃ within 0.5 min to 556 ± 42 % of the initial IP₃ content, followed by a decline to basal IP₃ levels within 5 min after stimulation (Fig. 1A; n = 6). In terms of the formation of IP₃, within 30 s 1 nmol l⁻¹ bradykinin was equally as effective as 10 mmol l⁻¹ NaF (Fig. 1B, n = 6-8).

As expected, bradykinin-evoked IP₃ formation resulted in a detectable elevation of free bulk Ca²⁺ concentration ([Ca²⁺]_b) in a concentration range between 10 nmol l⁻¹ and 1 μmol l⁻¹ measured with fura-2 and Calcium Green-1 (e.g. 100 nmol l⁻¹ measured with Calcium Green-1; Fig. 2A). Surprisingly, stimulation of single endothelial cells with 10 mmol l⁻¹ NaF failed to increase free bulk Ca²⁺ concentration ([Ca²⁺]_b) measured by either fura-2 or Calcium Green-1 in the absence of extracellular Ca²⁺ (n.s. vs. baseline; n = 24; Fig. 2B). Sodium fluoride activated capacitative Ca²⁺ entry in single endothelial cells, indicated by an increase in [Ca²⁺]_b in response to the addition of 2.5 mmol l⁻¹ CaCl₂, while there was no increase detectable without prior stimulation with NaF (Fig. 2B) or bradykinin (data not shown). Similar to NaF, submaximal concentrations of bradykinin (i.e. 1 nmol l⁻¹) failed to elevate [Ca²⁺]_b but initiated capacitative Ca²⁺/Mn²⁺ entry indicated by a significantly greater quenching of fura-2 fluorescence in response to the addition of 2.5 mmol l⁻¹ Mn²⁺ in cells pre-stimulated with 1 nmol l⁻¹ bradykinin compared with the effect of Mn²⁺ addition to resting cells (-12.3 ± 2.6 % within 3 min and -5.2 ± 1.9 % within 3 min, respectively; n = 6; P < 0.05).

Lowering extracellular Na⁺ concentration unmasks Ca²⁺ release initiated by NaF or low concentrations of bradykinin

Prevention of endothelial Na⁺-Ca²⁺ exchange activity by lowering the extracellular Na⁺ concentration from 145 to 19 mmol l⁻¹ unmasked NaF-induced intracellular Ca²⁺ release. While at a physiological extracellular Na⁺ concentration (i.e. 145 mmol l⁻¹) no elevation of [Ca²⁺]_b was detectable in response to 10 mmol l⁻¹ NaF, in 19 mmol l⁻¹ Na⁺-containing buffer 10 mmol l⁻¹ NaF significantly increased [Ca²⁺]_b (P < 0.01; n = 19; Fig. 3A). Moreover, NaF-initiated capacitative Ca²⁺ entry was increased by 93 % in low-Na⁺ buffer (P < 0.01; n = 19; Fig. 3A). The correlation between the NaF-evoked increase in [Ca²⁺]_b and the extracellular Na⁺ concentration is shown in Fig. 3B. The correlation was described by a polynomial function:

(r = 0.985). Similar to the data obtained with NaF, intracellular Ca²⁺ release in response to submaximal concentrations of bradykinin (i.e. 1 nmol l⁻¹) was unmasked in low extracellular Na⁺ conditions. While in 145 mmol l⁻¹ Na⁺ solution, [Ca²⁺]_b did not change in response to 1 nmol l⁻¹ bradykinin (0.69 ± 0.01 ratio units in the absence and 0.70 ± 0.01 ratio units in the presence of 1 nmol l⁻¹ bradykinin) in the nominal absence of extracellular Ca²⁺, in low-Na⁺ buffer (i.e. 19 mmol l⁻¹ Na⁺) 1 nmol l⁻¹ bradykinin increased [Ca²⁺]_b from 0.71 ± 0.01 to 0.88 ± 0.09 ratio units (P < 0.05; n = 9; data not shown). As shown in Fig. 3C, lowering extracellular Ca²⁺ significantly enhanced increases in [Ca²⁺]_b in the nominal absence of extracellular Ca²⁺ in response to stimulation with bradykinin at concentrations from 1 nmol l⁻¹ to 1 μmol l⁻¹.

Lowering extracellular Na⁺ concentration from 145 to 19 mmol l⁻¹ enhanced the potency of NaF to deplete the bradykinin-releasable Ca²⁺ pool. In cells pre-stimulated with NaF in 19 mmol l⁻¹ Na⁺ solution, the effect of 100 nmol l⁻¹ bradykinin of elevating [Ca²⁺]_b was reduced by 47 % compared with the effect on cells pre-stimulated with NaF in 145 mmol l⁻¹ solution (P < 0.01; n = 16; Fig. 4A). Co-incubation with 25 μmol l⁻¹ ryanodine diminished the observed elevation in [Ca²⁺]_b in response to NaF (10 mmol l⁻¹) under low extracellular Na⁺ conditions (P < 0.05; n = 4) and prevented depletion of the bradykinin-releasable pool(s) (P > 0.05; n = 4; Fig. 4B).

Prevention of endothelial Na⁺-Ca²⁺ exchange during NaF activation attenuates the loss of cellular Ca²⁺ and translocates Ca²⁺ to the mitochondria

In low-Na⁺ solution (19 mmol l⁻¹), the amount of intracellular Ca²⁺ stored within the cells and releasable by ionomycin (5 μmol l⁻¹) was increased by 50 % (P > 0.01; n = 10; Fig. 5A) compared with cells stimulated with NaF in normal extracellular Na⁺ solution (i.e. 145 mmol l⁻¹; n = 12). This increase was similar in extent to the augmentation of the mitochondrial Ca²⁺ pool releasable with carbonyl cyanide p-(trifluormethoxy) phenylhydrazone (FCCP; P < 0.01; Fig. 5B) in cells pre-stimulated with NaF in low-Na⁺ solution (19 mmol l⁻¹; n = 7) compared with experiments and physiological Na⁺ concentration (i.e. 145 mmol l⁻¹; n = 8). In agreement with these data obtained with FCCP, mitochondrial Ca²⁺ release in response to the ATP synthase inhibitor oligomycin (10 μmol l⁻¹) was also increased when cells were pre-stimulated with 10 mmol l⁻¹ NaF in solution containing 19 mmol l⁻¹ extracellular Na⁺ instead of 145 mmol l⁻¹ (Fig. 5B; n = 4).

Sodium fluoride and a low concentration of bradykinin initiate subplasmalemmal Ca²⁺ release towards the cell membrane

Using the laser pulse-wave loading technique cells were loaded with cell membrane-impermeable fura-2 pentapotassium salt (Fig. 6A) and Calcium Green-1 hexapotassium salt (Fig. 6B) and the response to 100 nmol l⁻¹ bradykinin was measured. Both dyes showed identical kinetics of bulk Ca²⁺ signalling and the result did not differ from that obtained in cells conventionally loaded with the acetoxymethyl ester forms of the dyes without the laser pulse-wave loading procedure (e.g. Graier et al. 1995). In agreement with the result mentioned above, neither fura-2 nor Calcium Green-1 could detect elevation of [Ca²⁺]_b after stimulation with 10 mmol l⁻¹ NaF or 1 nmol l⁻¹ bradykinin (data not shown).

To investigate a possible elevation of subplasmalemmal Ca²⁺ concentration ([Ca²⁺]_sp) upon application of NaF or a low bradykinin concentration, cells were loaded with FFP-18 to detect Ca²⁺ elevations in the subplasmalemmal area. As shown in Fig. 6C, in the nominal absence of extracellular Ca²⁺ a low concentration of bradykinin (i.e. 1 nmol l⁻¹) yielded elevation of [Ca²⁺]_sp detected by an increase in the FFP-18 ratio (P < 0.01 vs. baseline; P < 0.01 vs. effect without laser pulse-stress wave loading procedure, LSWP; n = 5; Fig. 6C) and initiated capacitative Ca²⁺ entry in response to the addition of extracellular Ca²⁺ (P < 0.01; n = 5). To compare changes at the subplasmalemmal side with the perinuclear region, cells were loaded with Calcium Green-1 to measure the whole Ca²⁺ signal ([Ca²⁺]_b) and FFP-18 to detect Ca²⁺ elevation in the subplasmalemmal area ([Ca²⁺]_sp). In Calcium Green-1/FFP-18-loaded cells, 1 nmol l⁻¹ bradykinin yielded a rapid elevation of [Ca²⁺]_sp monitored by FFP-18, while it failed to increase [Ca²⁺]_b, indicated by no change in the Calcium Green-1 fluorescence (Fig. 6D). In agreement with these results, NaF (10 mmol l⁻¹) only affected FFP-18 fluorescence ([Ca²⁺]_sp), while no change in the Calcium Green-1 signal was detected (data not shown; n = 8). Similar to 1 nmol l⁻¹ bradykinin, NaF (10 mmol l⁻¹) increased [Ca²⁺]_sp from 1.35 ± 0.12 to 1.77 ± 0.15 ratio units (ratio F₃₅₅/F₃₇₅) within 1 min (n = 6, P < 0.05 vs. basal [Ca²⁺]_sp), while Calcium Green-1 fluorescence remained unchanged (724 ± 36 before and 745 ± 22 arbitrary fluorescence units 1 min after addition of NaF; n = 6, n.s.; data not shown).

Superficial endoplasmatic reticulum contributes to the hidden Ca²⁺ release initiated by NaF and low bradykinin concentration

The functional organization of endothelial endoplasmic reticulum (ER) was investigated using BODIPY FL-X ryanodine (Figs 7 and 8). The BODIPY FL-X probe in skinned cells which was widely distributed without significant accumulation within the cell given and the fluorescence detectable was less than 9 % of the fluorescence using the ryanodine conjugate (data not shown). Using laser scanning microscopy (Fig. 7) and deconvolution microscopy (Fig. 8) three-dimensional reconstruction of the ER could be performed without the need for cell fixation with, for example, paraformaldehyde. The transmitted light image (Fig. 7A), the corresponding confocal image in the x-y plane taken at approximately the middle depth of the cell (in the z-axis) and the composite image of Fig. 7B and C were used to define the distribution of BODIPY FL-X ryanodine staining in this endothelial cell. A three-dimensional reconstruction of thirteen optical sections taken at 1 μm steps in the z-axis shown in Fig. 7D and E indicates the ER distribution in endothelial cells. The correlation of ER distribution with the cell membrane is demonstrated in Fig. 7F. An identical ER distribution was found when ER was stained using DiOC6 (data not shown).

To investigate whether superficial parts of the ER contribute to the hidden Ca²⁺ release in response to NaF and low concentrations of bradykinin, the ER organization was disturbed using nocodazole. As shown in Fig. 8, preincubation with nocodazole of cultured endothelial cells resulted in alteration of ER network and the collapse of the superficial ER towards the perinuclear region. Endothelial ER was labelled with BODIPY FL-X ryanodine and the distribution of both ER markers was determined in cells incubated for 16 h with the solvent alone (DMSO; Fig. 8A for the image and Fig. 8B for the deconvoluted image) and 10 μmol l⁻¹ nocodazole (Fig. 8C for the image and Fig. 8D for the deconvoluted image). The distribution of the ER markers within the cells was expressed as the ratio F_peripherial/F_perinuclear, where F_peripherial is peripherial fluorescence intensity and F_perinuclear is perinuclear fluorescence intensity, determined under control conditions and after nocodazole treatment. The ratio F_peripherial/ F_perinuclear of the ER marker was significantly decreased by the nocodazole treatment, from 2.77 to 0.82, indicating a greater than threefold increase of nuclear distribution after nocodazole treatment. Identical results were obtained using DiOC6, another well known marker for labelling ER (data not shown; Terasaki, Song, Wong, Weiss & Chen, 1984).

In cultured endothelial cells the collapse of the superficial ER caused by nocodazole unmasked intracellular Ca²⁺ release in response to NaF (Fig. 9A) and low concentrations of bradykinin (Fig. 9B). While in sham-treated cells [Ca²⁺]_b did slightly change in response to an addition of 10 mmol l⁻¹ NaF (n.s. vs. baseline; n = 17; Fig. 9A), NaF (10 mmol l⁻¹) yielded within 1 min a significant increase in [Ca²⁺]_b in cells pre-treated with nocodazole (P < 0.01 vs. sham-treated cells and P < 0.01 vs. baseline; n = 17; Fig. 9A). Hence, NaF-initiated apparent capacitative Ca²⁺ entry detected by an elevation of [Ca²⁺]_b was increased in cells pre-treated with nocodazole by 77 % (P < 0.01 vs. sham-treated cells; n = 22; Fig. 9A).

In agreement with the data obtained with NaF, 1 nmol l⁻¹ bradykinin slightly elevated [Ca²⁺]_b in sham-treated cells (n.s. vs. baseline; n = 22; Fig. 9B), while in nocodazole pre-treated cells the same concentrations of bradykinin yielded a marked increase in [Ca²⁺]_b (P < 0.01 vs. sham-treated cells and P < 0.01 vs. baseline; n = 22; Fig. 9B). Hence, the apparent capacitative Ca²⁺ entry initiated by bradykinin (1 nmol l⁻¹) detected by an elevation of [Ca²⁺]_b was increased in cells pre-treated with nocodazole by 336 % (P < 0.01 vs. sham-treated cells; n = 22; Fig. 9B). In contrast, nocodazole up to 100 μmol l⁻¹ had no acute effect on intracellular Ca²⁺ release and capacitative Ca²⁺ entry initiated by bradykinin, ATP and thapsigargin (data not shown). Moreover, IP₃ formation initiated by NaF (10 mmol l⁻¹) and bradykinin (1 and 100 nmol l⁻¹) remained unchanged in nocodazole-pre-treated cells (Table 1).

The collapse of the superficial ER also enhanced the effect of 100 nmol l⁻¹ bradykinin in elevating [Ca²⁺]_b in the nominal absence of extracellular Ca²⁺ and the capacitative Ca²⁺ entry (Fig. 10A), while the potency of thapsigargin (1 μmol l⁻¹) to elevate [Ca²⁺]_b in the nominal absence of extracellular Ca²⁺ and thapsigargin-stimulated capacitative Ca²⁺ entry remained unchanged after nocodazole treatment (Fig. 10B).