Rate, extent and concentration dependence of histamine-evoked Weibel–Palade body exocytosis determined from individual fusion events in human endothelial cells

Abstract

The rate, concentration dependence and extent of histamine-evoked Weibel–Palade body (WPB) exocytosis were investigated with time-resolved fluorescence microscopy in cultured human umbilical vein endothelial cells expressing WPB-targeted chimeras of enhanced green fluorescent protein (EGFP). Exocytosis of single WPBs was characterized by an increase in EGFP fluorescence, morphological changes and release of WPB contents. The fluorescence increase was due to a rise of intra-WPB pH from resting levels, estimated as pH 5.45 ± 0.26 (s.d., n = 144), to pH 7.40. It coincided with uptake of extracellular Alexa-647, indicating the formation of a fusion pore, prior to loss of fluorescent contents. Delays between the increase in intracellular free calcium ion concentration evoked by histamine and the first fusion event were 10.0 ± 4.42 s (n = 9 cells) at 0.3 μm histamine and 1.57 ± 0.21 s (n = 15 cells) at 100 μm histamine, indicating the existence of a slow process or processes in histamine-evoked WPB exocytosis. The maximum rates of exocytosis were 1.20 ± 0.16 WPB s⁻¹ (n = 9) at 0.3 μm and 3.66 ± 0.45 WPB s⁻¹ at 100 μm histamine (n = 15). These occurred 2–5 s after histamine addition and declined to lower rates with continued stimulation. The initial delays and maximal rate of exocytosis were unaffected by removal of external Ca²⁺ indicating that the initial burst of secretion is driven by Ca²⁺ release from internal stores, but sustained exocytosis required external Ca²⁺. Data were compared to exocytosis evoked by a maximal concentration of the strong secretagogue ionomycin (1 μm), for which there was a delay between calcium elevation and secretion of 1.67 ± 0.24 s (n = 6), and a peak fusion rate of ∼10 WPB s⁻¹.

Endothelial cells release the pro-coagulant protein von Willebrand factor (VWF), the cleaved propolypeptide of VWF (proregion) and the leucocyte adhesion molecule P-selectin from distinctive secretory organelles, the Weibel–Palade bodies (WPBs) (Weibel & Palade, 1964; Wagner, 1990). These proteins are important in regulating intravascular haemostasis and inflammation (Sadler, 1998; Andre et al. 2000; Hannah et al. 2005). VWF derived from WPBs that fuse with the apical endothelial membrane during vascular activation promotes platelet adhesion to the endothelial cell surface in vivo (Andre et al. 2000). Basolateral secretion of VWF from WPBs is thought to strengthen the attachment of endothelial cells to the basement membrane during vascular injury by interacting with extracellular matrix proteins and integrin receptors on the endothelial cell membrane (Wagner, 1990). The kinetics of WPB exocytosis and the intraorganelle conditions that promote the rapid expulsion of VWF onto the cell surface are important factors determining the efficiency of VWF action in these processes (Michaux et al. 2006).

A detailed analysis of the kinetics of secretagogue-evoked WPB exocytosis in intact endothelial cells has yet to be reported. WPB exocytosis is triggered by cell damage or by many extracellular signalling molecules and hormones that increase intracellular free calcium ion concentration ([Ca²⁺]_i) or intracellular cyclic adenosine monophosphate concentration (cAMP) (Carter et al. 1998; Vischer & Wollheim, 1998). Because VWF adheres strongly to the endothelial cell surface immediately following its release from WPBs (Hannah et al. 2005), biochemical assays of secreted, soluble VWF are unlikely to reflect the underlying kinetics or extent of WPB exocytosis. Assays based on cell surface P-selectin expression may provide more reliable data on the initial onset of WPB fusion (Hattori et al. 1989), but cannot report the overall time course or extent of the secretory response, and lack the sensitivity required for precise kinetic studies. A previous optical study of WPB exocytosis, using WPB targeted VWF-EGFP, was limited to a time resolution of about 1 min, too slow to resolve individual fusion events (Romani de Wit et al. 2003). High resolution membrane capacitance (C_m) measurements in single HUVECs (Zupancic et al. 2002) have revealed a distinct component of membrane capacitance C_m step amplitudes (4 fF to 9 fF) that were ascribed to WPB fusion, and that yielded a conservative estimate of the time course for WPB exocytosis that included a long delay and slow overall rate (Zupancic et al. 2002). Those studies used direct intracellular release of Ca²⁺ to elicit exocytosis, by flash-photolysis of Ca²⁺-DM-nitrophen, effectively by-passing hormone receptor and signal transduction cascades, and had potential complications in interpretation due to dialysis of cytosolic components. In addition, it should be pointed out that changes in membrane capacitance per se cannot provide direct information about the nature of the organelles that contribute to those changes.

Here we have used direct observation of WPBs expressing fusion proteins of enhanced green fluorescent protein (EGFP) and full length prepro-von Willebrand factor (VWF-EGFP) or the von Willebrand factor propolypeptide (proregion-EGFP) (Hannah et al. 2005) to determine, with subsecond time resolution, the rate, extent and concentration dependence of WPB exocytosis evoked by a physiological agonist, histamine, in intact single HUVECs at 37°C. These data were compared to that obtained with a maximally effective concentration of the receptor-independent agent ionomycin, a Ca²⁺–2H⁺ exchanger (Erdahl et al. 1994) and strong WPB secretogogue. The use of WPB-targeted EGFP fluorescence allowed the measurement of single granule fusion events without constraints due to whole cell dialysis by patch-pipettes, and for the analysis of WPB pH, proton buffering capacity and morphological changes that provide additional information relevant to the storage, processing and expulsion of VWF during WPB exocytosis.

Methods

Tissue culture and transfection

Primary human umbilical vein endothelial cells (HUVECs) were purchased from TCS Cellworks, Botolph, Claydon, UK and grown as previously described (Arribas & Cutler, 2000). HUVECs at passage 3 or 4 were nucleofected using 2–4 μg of expression vector DNA using the nucleofection device and buffers according to the manufacturer's instructions (Amaxa Gmbh, Cologne, Germany) and plated at confluent density in complete culture medium onto 35 mm diameter poly d-lysine coated glass bottomed culture dishes (MatTeK Corp. Ashland, USA) or 25 mm diameter uncoated glass coverslips for live cell imaging or onto gelatine-coated 9 mm diameter glass coverslips for immunocytochemistry. Cells were maintained without further medium changes until their use one to three days after nucleofection. The apparent transfection efficiencies, determined from the number of cells with detectable EGFP fluorescence and fluorescent WPBs varied between 5 and 20% for full length VWF-EGFP, 20–40% for proregion-EGFP.

Antibodies, reagents and immunocytochemistry

Rabbit polyclonal antihuman VWF was purchased from Dako Ltd (Ely, UK). Secondary antibodies coupled to fluorophores were purchased from Jackson Immunoresearch (USA). Full length VWF-EGFP and proregion-EGFP were made as previously described (Hannah et al. 2005). All other reagents were purchased from Sigma-Aldrich (Poole, UK) unless stated otherwise. Nucleofected HUVECs grown on gelatin-coated glass coverslips were processed for immunofluorescence as previously described (Hannah et al. 2005).

Epifluorescence imaging of WPB exocytosis in living cells

Nucleofected HUVECs were grown on poly d-lysine coated glass-bottomed chambers. Time-lapse images of EGFP fluorescence were recorded in release medium (RM; medium 199 supplemented with 0.2% BSA and 10 mm Hepes-NaOH, pH 7.2 at 24°C or 37°C, as indicated), on a Deltavision Imaging system (Applied Precision Inc., Seattle, WA, USA) using either an Olympus U-Plan-Apo ×100 objective, 1.35 NA, or Plan Apo ×60, 1.4 NA objective in conjunction with a Princeton Instruments Micromax air-cooled interline CCD camera. A 490 ± 20 nm excitation and 528 ± 38 nm emission filter was used in conjunction with a 4-bandpass standard Deltavision dichroic (Chroma; C13022-84100). Secretion was evoked by application of 100 μm histamine or 1 μm ionomycin. The Deltavision system was equipped with a z-stage controller and enclosed within a microscope incubator (Solent Scientific Ltd, Segensworth, UK) allowing 37°C temperature control. In experiments to determine the number of WPBs that underwent exocytosis in response to stimulation a z-series image stack (comprising 15–20 optical sections at 0.2 μm intervals through the cell) was obtained prior to a time-lapse sequence (obtained in a fixed optical plane). The numbers of WPBs prior to stimulation and those that underwent exocytosis were determined as described below in Data analysis and measurements.

Epifluorescence imaging of changes in [Ca²⁺]_i and WPB exocytosis in living cells

Nucleofected HUVECs were grown on 24 mm square uncoated glass coverslips in growth medium. For Fura-2 loading the cells were transferred into Hanks' solution (140 mm NaCl, 5 mm KCl, 1.8 mm CaCl₂, 2 mm MgCl₂, 10 mm glucose, 10 mm Hepes-NaOH, pH 7.40) supplemented with 2.5 μm Fura-2/AM (Molecular Probes, Eugene, OR, USA) diluted from a stock solution of 5 mm Fura-2/AM in DMSO with 2.0% Pluoronic F-127. Cells were incubated at room temperature in the dark for 20 min before being transferred into fresh Hanks' solution and further maintained in the dark at room temperature until use. Under these loading conditions the kinetics of histamine (0.3, 3 or 30 μm)-evoked exocytosis was identical to that of sham-loaded cells (data not shown). For experiments, cells were transferred to the stage of an Olympus IX71 inverted fluorescence microscope where they where maintained at between 36 and 37°C using a stage heater, objective heater and in-line solution heater (Harvard apparatus). The imaging system comprised a monochromator (Optoscan, Cairn Research Faversham, UK) for fast (∼1 ms) wavelength switching, a U-Plan Apo ×100, 1.35 NA objective and an Ixon EMCCD camera (Andor Belfast, Northern Ireland). Wavelength switching was synchronized with image capture using WinFluor software (Dr John Dempster, Strathclyde University). Fura-2 was illuminated sequentially with 355 ± 5 nm and 380 ± 5 nm light and EGFP with 470 ± 10 nm light. Excitation light was reflected onto the specimen using a 500DCXR dichrochic mirror (Chroma Rockingham, USA) and the emitted light passed through a 535 ± 50 nm emission filter (Fura-2 and EGFP emit in essentially the same region of the visible spectrum) onto the chip of the EMCCD camera. The EMCCD camera was operated in frame transfer mode at full gain and cooled to – 65°C. Full frame (512 by 512 pixels) images were acquired at 30 frames s⁻¹ giving ∼10 fura-2 ratios and ∼10 EGFP images per second. Histamine or ionomycin were applied to the cell by pressure injection from a glass micropipette positioned close to the cell (puffer application). The distance between the pipette and cell (typically 10–15 μm) and the compressed air pressure (2 p.s.i.) for application were adjusted such that control solutions did not elicit a change in [Ca²⁺]_i due to fluid shear effects. To estimate the time course for drug arrival at the cell during puffer application, experiments were carried out in which 7 μm Alexa-647 was included in the puffer pipette solution. These experiments were carried out as part of studies to determine the effect of fura-2 loading on secretion, specifically in control experiments in non-fura-2 loaded cells expressing proregion-EGFP and stimulated with histamine (0.3, 3 or 30 μm). The cells were excited sequentially with 488 ± 20 nm (EGFP) and 565 ± 20 nm (Alexa-647) light using an EGFP/DsRed filter cube (Chroma; part number 51019) containing a dual band-pass dichroic mirror and dual band emission filter (Chroma; part number 51019 bs and 51019 m, respectively), allowing sequential capture of EGFP images and Alexa-647 images at 30 frames s⁻¹. Activation of the puffer controller was monitored by a +1 V analog telegraph output recorded simultaneously with image capture within the acquisition program WinFluor (Dr John Dempster; Strathclyde University). The time course for exchange of solution at the cell was determined from the Alexa-647 fluorescence change in a rectangular region of interest (ROI) placed over the cell.

Confocal imaging of WPB exocytosis

HUVECs were nucleofected with proregion-EGFP and plated onto poly d-lysine coated glass-bottomed chambers. Forty-eight hours after nucleofection, cell culture medium was changed for Hanks' solution (140 mm NaCl, 5 mm KCl, 1.8 mm CaCl₂, 2 mm MgCl₂, 10 mm glucose, 10 mm Hepes-NaOH, pH 7.40) containing 7 μm Alexa-647 (Molecular Probes) at room temperature (22–24°C). Secretion was stimulated by addition of 1 μm ionomycin. Images were collected using a Leica laser scanning spectral confocal microscope (TCS SP2) equipped with either a ×63, 1.32 NA or a ×100, 1.40 NA objective (Leica). EGFP and Alexa-647 were simultaneously excited with 488 nm and 561 nm lasers and emitted light was separated by an acousto-optical tunable filter (498–550 nm; EGFP, 600–850 nm; Alexa-647) onto two photomultiplier channels. Confocal images were acquired at ∼7.4 frames s⁻¹.

Imaging of WPB exocytosis by total internal reflection fluorescence microscopy

Nucleofected HUVECs were plated onto uncoated 25 mm glass coverslips in growth medium and transferred into Hanks' solution for total internal reflection fluorescence microscopy (TIRFM) microscopy. TIRFM experiments used an objective based TIRFM system built on a Zeiss Axiovert S100 microscope equipped with a Fluor ×100, 1.45 NA objective (Zeiss), and with a custom optical arrangement to produce evanescent field excitation as previously described (Mashanov et al. 2003, 2004). Time-lapse images of VWF-EGFP- or proregion-EGFP-labelled WPBs in live HUVECs were acquired approximately every 94 ms (∼10 frames s⁻¹) using custom written software (Mashanov et al. 2003) and 488 nm excitation light (Protera 15mW 488 nm laser, Novalux Inc., Sunnyvale, CA, USA) was attenuated by a 0.85 neutral density filter to reduce photobleaching. The entire microscope system and recording chamber was enclosed within a microscope incubator (Solent Scientific Ltd, Segensworth, UK) and maintained at 37°C.

In vitro and in vivo pH titration of EGFP fluorescence

In vitro experiments using purified EGFP (2 μm; Clontech) were preformed using a SPEX fluoromax fluorimeter at room temperature (24°C). EGFP was dissolved in buffer comprising 120 mm NaCl, 5 mm NaCl, 0.5 mm CaCl₂, 0.5 mm MgSO₄, containing 10 mm Mes, 10 mm Mops and 10 mm citrate and adjusted to different pH values (8.0–4.0) with NaOH as described in Kneen et al. (1998). The EGFP fluorescence, normalized to that at pH 8.0, was plotted against the pH of the solution to generate a calibration curve. In vivo calibrations were carried out at room temperature (24°C) as described by Wu et al. (2001) using cell-permeabilizing calibration solutions described in Kneen et al. (1998), comprising 120 mm KCl, 20 mm NaCl, 1 mm MgSO₄, 20 mm Hepes-NaOH, 0.5 mm CaCl₂, 10 mm glucose, supplemented with 5 μm nigericin and 5 μm monensin, titrated to different pH values (range 8.0–4.0) with KOH. HUVECs expressing VWF-EGFP or proregion-EGFP 48 h post-nucleofection were used because at this time point, in the majority of cells, EGFP was confined predominantly within WPBs, with little or no fluorescence signal from newly synthesized EGFP in the endoplasmic reticulum (ER) or Golgi apparatus (GA). Single HUVECs were imaged using a Deltavision Imaging system as described above. Excitation light was attenuated using a 50% neutral density filter and shuttered between image acquisitions (at 60 s intervals) to limit photobleaching of EGFP fluorescence. After the initial base-line fluorescence had been established, calibration solutions were perfused sequentially onto the cell allowing sufficient time for fluorescence intensity to stabilize at each pH. The mean whole-cell fluorescence normalized to that at pH 8.0 (corresponding to F_max for EGFP) was plotted against the pH of the solution to generate a calibration curve (see Tompkins et al. 2002). To derive the in vitro pK_a for purified EGFP and the in vivo (intra-WPB) pK_a values for EGFP fused to VWF or proregion, data were fitted by a modified Hill equation (Kneen et al. 1998) describing the dependence of EGFP fluorescence F on the pH of the calibration solution:

(1)

where F_max is the maximal fluorescence of EGFP (at pH 8.0) n_H is the Hill coefficient and pK_a is the pH corresponding to half-maximal fluorescence.

Resting intra-WPB (pH_g) and proton buffering capacity (β) of individual WPBs in unstimulated HUVECs

The proton buffering capacity (β) of individual WPBs was determined using the ammonium chloride pulse technique (Roos & Boron, 1981), utilizing intra-WPB EGFP fused to proregion as the pH indicator. Experiments were carried out at room temperature (24°C). HUVECs expressing proregion-EGFP, 48 h post-nucleofection, were imaged using an Ultraview Spinning Disk confocal microscope (Perkin-Elmer), equipped with an Olympus U-Plan Apo ×100, 1.35 NA objective. Excitation light (488 nm) was from an argon ion laser and emitted light (525 ± 50 nm) was detected by a Hamamatsu ORCA-ER CCD camera. Cells expressing well-defined fluorescent WPBs were chosen and the cells maintained in a control solution comprising 140 mm NaCl, 5 mm KCl, 10 mm Hepes-NaOH, 1.8 mm CaCl₂, 1 mm MgCl₂, 10 mm glucose, pH 7.4. The control solution was changed for one containing 0.5, 1.25 or 2.5 mm NH₄Cl (adjusted to pH 7.40 with NaOH), and the EGFP fluorescence of individual WPBs was monitored and allowed to stabilize before the solution was exchanged for a pH 8.0 calibration solution containing 5 μm nigericin and 5 μm monensin (described above). Buffering capacity β was defined as:

(Roos & Boron, 1981). Ammonium ([NH₄⁺]) load of WPBs was calculated assuming complete transparency of cell and WPB membranes to ammonia (NH₃) and impermeability to NH₄⁺. At equilibrium the WPB ammonium load within the cell and organelles is given by:

(2)

where pH_g is the resting intra-WPB pH in control conditions and pH_o is the extracellular pH (7.40). With designated as the steady state intra-WPB pH in NH₄Cl solutions and the pH change (ΔpH) inside the WPB as , the value of β can be estimated from:

(3)

Calculation of the WPB ammonium load requires knowledge of pH_g.pH_g and pH_gNH4 were calculated from:

(4)

using values for pK_a (5.84) and n_H (0.74) derived from the in vivo pH titration of WPB EGFP fluorescence (see Results). F_max is the maximal fluorescence of EGFP in the pH 8.0 calibration solution. Data normalized to F_max was fitted by eqn (4) where F represents the steady-state WPB EGFP fluorescence either in control solutions (such that pH = pH_g, e.g. see Fig. 3) or immediately after exposure to NH₄Cl solutions (such that pH =, e.g. see Fig. 3).

Figure 3.

Intra-WPB pH and proton buffering capacity A, the total fluorescence (normalized to that at pH 8.0) of an individual proregion-EGFP-containing WPB in a live cell exposed first to a control solution (F_rest), second to solution containing NH₄Cl (2.5 mm, pH 7.40 ()) and finally to a pH 8.0 calibration solution containing 5 μm nigericin and 5 μm monensin (F_max). Images of the WPB taken at times indicated by the asterisks are shown above; scale bar is 2 μm. B, the intra-WPB pH (pH_g) of the WPB in A calculated using eqn (4) (see Methods) taking values for the pK_a of intra-WPB EGFP of 5.84 and n_H of 0.74 (see Results). C, a plot of ammonium load versusΔpH for individual WPBs exposed to either 0.5 (^), 1.25 (•) or 2.5 mm (Δ) NH₄Cl, respectively. Large grey circles correspond to the mean values for ammonium load (±s.d.) and ΔpH (±s.d.) at each concentration of NH₄Cl and the continuous line shows a regression fit to the mean data yielding a mean value for β of 55 mm pH unit⁻¹.

Data analysis and measurements

Image analysis was carried out using custom-written software (GMview; Mashanov et al. 2003) and custom written plugins and macros in ImageJ software (http://rsb.info.nih.gov/ij/). Confocal images were filtered with 3-pixel radius mean ImageJ filter. The mean background subtracted fluorescence was measured in minimal oval regions of interest including the whole WPB. Dataset plotting, fitting and analysis were performed in Microsoft Excel and Origin 7 (OriginLab Corp., Northampton, MA, USA). Results are expressed as means ±s.d. unless indicated otherwise. Statistical differences (at 95% confidence limit) between population means were determined using a non-paired two-way t test. The proportion of WPBs that underwent exocytosis was determined in epifluorescence experiments. Using the Deltavision system a z-series image stack of a cell was taken prior to the start of the experiment and exported to ImageJ and analysed within the Pointpicker plugin (http://rsb.info.nih.gov/ij/plugins/index.html). The total number of WPBs within the cell was counted manually. The number of WPBs that under went exocytosis was determined (in ImageJ) from the time-lapse images. Using the Ca²⁺ and EGFP imaging system a 20 frame averaged image, taken prior to stimulation, was exported to ImageJ and analysed within the Pointpicker plugin. The total number of WPBs within the cell was counted manually twice and the mean value taken. The numbers of WPBs that underwent exocytosis were determined. Changes in [Ca²⁺]_i were calculated as the ratio of fluorescence at 355 nm divided by that at 380 nm within the WinFluor program. Briefly, at each wavelength, 355 nm or 380 nm, background was measured, using the same ROI placed at a point in which no cell was present. Mean background values from this ROI were then subtracted from the fluorescence values obtained from a similar ROI placed on the cell. The ratio of the resulting background-subtracted fluorescence signals at 355 nm and 380 nm was taken to represent the change in [Ca²⁺]_i. The time of the initial rise of [Ca²⁺]_i evoked by agonist was determined as the time point at which the fura-2 fluorescence ratio rose above +3 s.d. of the mean resting fluorescence ratio (see Fig. 5). The s.d. of the resting fura-2 fluorescence ratio was calculated from a 10 s stretch of data prior to agonist application.

Figure 5.

Delays in histamine-evoked increases in [Ca²⁺]_i and exocytosis A illustrates how the time of secretogogue application to the cell was determined and defines the delays D₁and D₂. Aa, the +1 V analog output from the micropipette solution puffer controller that signals activation of the device. The point of activation was used to define time = 0 s for subsequent analysis of delay times. Ab, the time course for arrival of puffer solution over the cell. The continuous trace represents the mean time course (normalized to the peak steady-state fluorescence level, n = 15 cells) for the fluorescence increase due to Alexa-647 in the puffer solution, in a ROI placed over the cell. Dotted lines represent s.e.m. The time for solution arrival at the cell was taken arbitrarily as the mean time taken to reach 50% of the steady-state Alexa-647 signal, and is indicated on the record by the vertical dashed line projecting down to the time axis in Ac. Ac, a representative record of the increase in [Ca²⁺]_i evoked by 3 μm histamine (continuous trace) and the time for the first fusion event detected in this cell (•). The dash–dot line corresponds to +3 s.d. of the mean resting Fura-2 fluorescence ratio, determined from a 10 s time period prior to agonist application. Elevation of the Fura-2 fluorescence ratio above this threshold was used to determined the point of increase in [Ca²⁺]_i. The delays between application and the evoked increase in [Ca²⁺]_i (D₁) and between the increase in [Ca²⁺]_i and the first fusion event (D₂) are indicated by the horizontal bars.

Results

Increase in WPB EGFP fluorescence during exocytosis evoked by histamine or ionomycin coincides with the uptake of extracellular Alexa-647

We have previously shown that expression of VWF-EGFP or proregion-EGFP in HUVECs results in the formation of fluorescent WPBs that undergo exocytosis in response to stimulation with extracellular histamine (100 μm) or ionomycin (1 μm) (see Hannah et al. 2005). Fluorescence was observed in WPBs and in the ER, but a proportion of non-fluorescent WPBs (22.2 ± 16.1% of total WPBs, n = 20 proregion-EGFP expressing cells) were seen in cells fixed and stained for VWF. Presumably these were granules made before or following cessation of expression of either construct. Fluorescent WPBs varied in size from small approximately spherical structures (< 0.5 μm diameter) to large cigar-shaped organelles up to 4 μm in length with a similar distribution of lengths (not shown) to that observed for wild-type WPBs (e.g. Zupancic et al. 2002).

The earliest event detected during WPB exocytosis evoked by histamine (100 μm) or ionomycin (1 μm) was an increase in EGFP fluorescence followed immediately (< 100 ms), or with a short delay, by an abrupt change in WPB morphology (Figs 1A and 4 and also online Supplemental material movies 1 and 3). In the majority of cases (∼90%) these WPBs went on to release their fluorescent contents (see movies). In the remaining ∼10% of cases cigar-shaped WPBs were seen to collapse into approximately spherical, membrane-bound structures in which the VWF-EGFP or proregion-EGFP remained and was not released (e.g. see Hannah et al. 2005; Supplemental material movie 3). By use of simultaneous dual-colour confocal imaging it was found that WPBs rapidly (within the same 135 ms frame as the increase in EGFP fluorescence) accumulated the extracellular fluorescent marker Alexa-647 prior to release of fluorescent contents (90/90 observations; an example of which is shown in Fig. 1Ba and b, and also movie 2). Importantly, many WPBs (approximately half; see below for quantification) showed no change in EGFP fluorescence during stimulation with histamine or ionomycin. These WPBs exhibited no shape change, and in confocal studies were found not to accumulate Alexa-647, together indicating that they had not undergone plasma membrane fusion. These data support the idea that the increase in fluorescence upon addition of histamine or ionomycin in WPBs that exocytose is due to the formation of a fusion pore.

Figure 1.

EGFP fluorescence increases during WPB exocytosis A, a sequence of images of a single WPB containing VWF-EGFP (Supplemental material movie 1) showing the increase in fluorescence, change in morphology and subsequent release of VWF-EGFP during exocytosis. Times are indicated on each frame and the scale bar represents 2 μm. Ba shows a montage of images from a single, proregion-EGFP containing WPB taken from a time lapse movie of HUVEC exposed to 7 μm extracellular Alexa-647 (Aa) during stimulation with 1 μm ionomycin (movie 2). Images were acquired at 7.4 frames s⁻¹ on a Leica SP2 confocal microscope equipped with a 100×, 1.4 NA objective and the scale bars are 2 μm. The upper panel in Ba shows EGFP fluorescence, the middle panel shows fluorescence of Alexa-647, and the lower panel the merged images (EGFP, green; Alexa-647, red). Bb, the time course for the WPB-EGFP (black trace) and Alexa-647 (grey trace) fluorescence change during ionomycin evoked exocytosis for the WPB depicted in Ba. For comparison the fluorescence data in each case were scaled and offset to overlay one another.

Figure 4.

Stimulus-evoked changes in intra-WPB EGFP fluorescence and pH_g Aa, a montage of images of a single proregion-EGFP containing WPB taken from a time lapse movie (Supplemental material movie 3) of a HUVEC during stimulation with histamine (100 μm, 24°C, imaged by epifluorescence microscopy). The images correspond to the points between the arrows in Ac and the frame marked with an asterisk corresponds to the similarly marked data point in Ac. The images were acquired at 6.25 frames s⁻¹ and the scale bar is 2 μm. Ab, the mean fluorescence of the WPB in Aa during the exocytotic event. Ac, the region during which fluorescence increased on an expanded time scale. Resting and plateau fluorescence levels are marked by dashed lines. Ba, the distribution of pH_g for the combined data for individual mature WPBs in resting cells and for WPBs that underwent exocytosis (n = 248 WPBs). Bb, the distribution of pH_g for immature WPBs (n = 101 WPBs).

The WPB fusion-evoked EGFP fluorescence increase is due to an increase in intra-WPB pH from resting levels to pH 7.40

The fluorescence increase of EGFP containing WPBs may be used as a convenient marker of WPB–plasma membrane fusion for optical studies of the kinetics of exocytosis. The fluorescence of EGFP is known to be strongly pH sensitive (see below), and increases in fluorescence during exocytosis, monitored by GFP targeted to the lumen of secretory granules, have been attributed to a rise in intragranule pH from an initial acidic resting level to that of the extracellular solution (typically ∼pH 7.40) upon formation of a fusion pore (Miesenbock et al. 1998; Sankaranarayanan & Ryan, 2000; Gandhi & Stevens, 2003; Taraska et al. 2003; Tsuboi & Rutter, 2003; Fernandez-Alfonso & Ryan, 2004).

Titration of EGFP fluorescence in vitro and of WPB associated VWF- or proregion-EGFP in situ

Experiments were carried out to determine the intra-WPB pH (pH_g) and establish quantitatively if the increase in WPB EGFP fluorescence preceding exocytosis can be accounted for by a rise in pH_g to pH 7.40. Intra-organelle-targeted EGFP has been used previously to measure organelle pH (Kneen et al. 1998; Llopis et al. 1998; Tompkins et al. 2002) and was used here to determine the pH_g. The pH dependence of purified EGFP in vitro is shown by filled circles in Fig. 2A. Figure 2B shows an example of the pH titration of total proregion-EGFP fluorescence in situ by epifluorescence imaging within a HUVEC, permeabilized as described in Methods. The pH dependence of VWF-EGFP (open circles; n = 6 cells) and proregion-EGFP fluorescence (triangles; n = 4 cells) is summarized in Fig. 2A. In each case the data were fitted with a modified Hill equation, and yielded similar values for pK_a and n_H for purified EGFP, and EGFP fused to VWF or proregion. A fit to the combined data for VWF-EGFP and proregion-EGFP (not shown) gave a pK_a of 5.84 and n_H of 0.74. Under resting conditions (extracellular pH 7.40) the mean pH_g values reported by the whole cell VWF-EGFP (pH 5.68 ± 0.19, n = 6 cells) and proregion-EGFP (pH 5.55 ± 0.22, n = 4 cells) fluorescence were not significantly different. Although care was taken to ensure that the cells used expressed EGFP predominantly in WPBs, it is possible that the values of pH_g determined here in whole single cells might include a small component arising from EGFP in structures other than WPBs, such as the ER and GA. ER or GA are reported to have luminal pH values more alkaline than those of regulated secretory organelles (Llopis et al. 1998; Wu et al. 2001; Paroutis et al. 2004), and EGFP contained in such structures could lead to an over-estimation of the true pH_g.

Figure 2.

pH dependence of intra-WPB EGFP fluorescence A, summary of the relationship between EGFP fluorescence and pH in vitro (•), and for intra-WPB proregion-EGFP (^) or VWF-EGFP (triangles) fluorescence and calibration solution pH in situ in living HUVECS. The in vitro EGFP data points are from duplicate experiments. In all cases the fluorescence was normalized to the peak fluorescence at pH 8.0 and was fitted by a modified Hill equation (eqn (1) in Methods). Fitted values for pK_a and n_H of 5.88 and 0.9 (EGFP in vitro; continuous line), 5.86 and 0.72 (proregion-EGFP in situ; dotted line) and 5.82 and 0.75 (VWF-EGFP in situ; dotted line) were obtained. B, the total fluorescence of a proregion-EGFP expressing HUVEC, 48 h post nucleofection, during exposure to calibration solutions containing 5 μm nigericin and 5 μm monensin and titrated to the pH values indicated.

Resting pH_g and proton-buffering capacity (β) of individual WPBs

We next determined both the resting pH_g and β of individual WPBs in live HUVECs using the ammonium chloride pulse technique (described in Methods). Figure 3A shows the fluorescence of an individual WPB in a cell exposed to a control solution (F_rest; solution pH 7.40), during exposure to a solution containing 2.5 mm NH₄Cl (; solution pH 7.40) and subsequently to the cell permeabilizing pH 8.0 calibration solution (F_max; pH 8.0, see Methods). In this example the mean resting pH_g in the presence of the control solutions, calculated using eqn (4); see Methods), was 5.49, and the intra-WPB pH in the presence of NH₄Cl (pH_gNH4) was 6.22 (Fig. 3B). From these data the ammonium load was 0.037 m, ΔpH was 0.74, and β was 52 mm pH⁻¹ unit. The ammonium load versusΔpH for individual WPBs exposed to 0.5 (open circles, n = 33), 1.25 (filled circles, n = 32) and 2.5 mm NH₄Cl (open triangles, n = 33), respectively, is summarized in Fig. 3C. Data at each concentration of NH₄Cl were obtained in four to seven cells, and only data that fell within the approximately linear region of the titration curve for intra-WPB pH versus EGFP fluorescence (pH 4.80–6.80; see Fig. 2A) were used. The mean value for β in resting WPBs was 55 mm (pH unit)⁻¹ (Fig. 3), but there was considerable variation in this parameter (see Discussion). The mean value for resting pH_g determined in these experiments was pH 5.45 ± 0.26 (n = 144) and was not significantly different (P = 0.12) from that determined in calibration experiments with whole-cell EGFP fluorescence (see above). The distribution of resting pH_g of individual immature WPBs was determined in cells 6.5–7 h hours post-nucleofection (Fig. 4Bb), the earliest time point at which we observed newly formed proregion-EGFP-containing WPBs in these experiments. The mean pH_g for this population of WPBs was pH 6.16 ± 0.45 (n = 101 WPBs from 10 cells), less acidic than older WPBs. At this time the majority of EGFP fluorescence was localized to ER. Making the assumption that the properties of EGFP were the same, the pH of the ER in these cells was 7.16 ± 0.10 (n = 11 cells).

The increase in intra-WPB EGFP fluorescence is due to a rise in pH_g to pH 7.40

Imaging WPB exocytosis with subsecond time resolution revealed that the increase in EGFP fluorescence reached a plateau during which the distinctive WPB morphology was maintained (see Fig. 4Aa,b and c, and also movie 3). The maximum fluorescence at the plateau was well within the dynamic range of the camera and did not represent pixel saturation. Normalizing the fluorescence increase to the peak (e.g. Fig. 4Ac) yielded a mean increase of 2.55 ± 0.68 (n = 105) for VWF-EGFP or 2.69 ± 0.50 (n = 54) for proregion-EGFP-containing WPBs during fusion at room temperature. These changes are similar to those values obtained for WPBs exposed to cell permeabilizing pH 8.0 calibration solutions (2.77 ± 0.74, n = 100) described above or from those values obtained for histamine or ionomycin evoked WPB fusion events observed at 37°C (Hannah et al. 2005). Thus a good estimate of the pH_g immediately prior to fusion can be obtained by assuming the maximum fluorescence was, at external pH 7.40, only 3% less than the value at pH 8.0 (F_max; see Fig. 2A). Taking values for the pK_a and n_H of EGFP as 5.84 and 0.74 (see above) and applying eqn (4) (see Methods) gives an estimate of the resting pH_g of 5.56 ± 0.29 (n = 104 WPBs), prior to fusion. This was not significantly different from pH_g obtained for unstimulated resting WPBs. The distribution of resting pH_g for the combined data for unstimulated WPBs and WPBs that underwent exocytosis is shown in Fig. 4Ba. To further test if the fluorescence increase was due to proton loss to the external solution through the fusion pore, cells were bathed acutely in solution of pH 5.8 and stimulated with 100 μm histamine. Under these conditions the increase in EGFP fluorescence on fusion was reduced to 1.68 ± 0.35 (n = 29 WPBs), a value close to that predicted from the measured pK_a of EGFP fused to proregion or VWF and a pH_g change from ∼5.5 to pH 5.8. These data indicate that the EGFP fluorescence increase seen following stimulation can be accounted for by a rise of resting pH_g to that of the extracellular solution and together with the uptake of Alexa-647 from the external solution is consistent with the formation of a fusion pore.

Time course, concentration dependence and extent of histamine-evoked WPB exocytosis

During histamine action WPBs will pass through a number of stages such as vesicle transport, priming and docking prior to final membrane fusion. Analysis of the time course of fusion will show if there are kinetically distinct steps in the secretory process and whether these can be modified by histamine concentration and other parameters. The time of the secretogogue application and the rise of Ca²⁺ are known, and the time of the final irreversible step in secretion is measured for each WPB. Thus, a plot of the intervals between stimulation and fusion will represent the kinetics of the underlying steps in secretion. Fluorescence microscopic observation shows that there are no specialized regions for secretion on the endothelial membrane, single fusion events are detected at sites over the cell surface, and the frequency of multiple compound fusion is very low (see below). Furthermore, under the conditions here the spontaneous rate of fusion was zero in eight cells monitored for 180–300 s. This suggests that lifetimes of individual WPBs are independent, and the zero spontaneous rate suggests that in the unstimulated cell WPBs are at a point in the process far from the final fusion event. The distribution of times from stimulation to fusion for all WPBs therefore provides information of the underlying kinetic steps analogous to that obtained in the first latency distributions applied to channel opening kinetics (Colquhoun & Hawkes, 1995). These assumptions hold in the conditions here but may not apply in cells with very high densities of WPBs, where molecular components may become limiting or where inter-WPB compound or cumulative fusion events may occur with a significant frequency.

The time of stimulation (histamine or ionomycin application) was determined from the fluorescence signal of Alexa-647 included in the agonist containing solution (see Methods) and was taken as the time point corresponding to the 50% rise of Alexa-647 fluorescence. This was ∼200 ms (2 frames) post-activation of the puffer pipette (vertical dotted line in Fig. 5). The time of WPB fusion was determined as the mid-time of each frame in which WPB-EGFP fluorescence increased due to fusion pore formation. Figure 5 illustrates the temporal relationship between histamine application, the evoked rise in [Ca²⁺]_i, shown by the fura-2 fluorescence ratio, and the first event of WPB fusion, and indicates the delays in Ca²⁺ release and exocytosis analysed here.

Rates and delays of histamine-evoke WPB exocytosis

At 0.1 μm histamine 12/14 cells studied exhibited a small increase in [Ca²⁺]_i (e.g. Fig. 6A upper panel) with a delay between application and the rise in [Ca²⁺]_i, D₁, of 6.26 ± 1.3 s (s.e.m., n = 12 cells). No WPB fusion events were detected at this histamine concentration. At 0.3 μm histamine, 21/23 cells responded with an increase in [Ca²⁺]_I; of these, nine cells underwent exocytosis. At concentrations > 0.3 μm all cells responded with both an increase in [Ca²⁺]_i and exocytosis, and such an experiment is shown in Supplemental material movie 4. Examples are shown in Fig. 6A of the histamine-evoked increase in [Ca²⁺]_i (upper panel) and histograms of the times of WPB fusion, measured from the rise in [Ca²⁺]_i, in individual cells exposed to either 0.1, 1.0, 10 or 100 μm histamine (histograms in lower panel). The time course for histamine action comprised a delay between application and the rise in [Ca²⁺]_i (D₁), of 2.75 ± 0.35 s (s.e.m., n = 21 cells) at 0.3 μm which shortened to 0.51 ± 0.03 s (s.e.m., n = 15 cells) at 100 μm histamine (Fig. 6B). Following the rise in [Ca²⁺]_i there was a further delay prior to the onset of exocytosis (D₂), of 10.0 ± 4.42 s (s.e.m., n = 9 cells) at 0.3 μm histamine and 1.57 ± 0.21 s (s.e.m., n = 15 cells) at 100 μm histamine (Fig. 6B). The overall time course for the WPB secretory response is summarized in Fig. 6C, which shows the times of WPB fusion events pooled from 15 to 20 cells exposed to histamine at the concentrations indicated. The rate of WPB fusion increased to a maximum between 2 and 5 s after the rise in [Ca²⁺]_i before declining to much lower levels in the continued presence of histamine. Approximately three-quarters of all exocytosed WPBs fused and released their contents within 30 s of stimulation, this number rising to ∼90% by 60 s post-stimulation. Exocytosis ceased after ∼500 s during prolonged stimulation (not shown). The maximal rate of WPB fusion increased from 1.2 ± 0.16 WPB s⁻¹ (mean, s.e.m., n = 9 cells) at 0.3 μm to 3.66 ± 0.45 WPB s⁻¹ (s.e.m., n = 15 cells) at 100 μm histamine (Fig. 6D). The highest rate of exocytosis recorded for histamine (30 μm) was 12 WPB s⁻¹. In TIRFM experiments the overall time course and maximal rate of WPB fusion, 2.68 ± 2.19 WPB s⁻¹ (range 0.36–9.27 WPB s⁻¹, n = 23 cells), evoked by 100 μm histamine was not significantly different from that determined for this concentration in epifluorescence experiments. Examples of TIRFM experiments are shown in movies 5 and 6 of supplemental materials.

Figure 6.

Concentration dependence of histamine-evoked WPB exocytosis A, upper panel, representative records of the changes in fura-2 fluorescence-ratio (355 nm/380 nm) in single HUVECs during stimulation with 0.1, 1.0, 10 and 100 μm histamine as indicated. For comparison the fura-2 traces were offset so that the increase in [Ca²⁺]_i in each case occurred at time = 0. Lower panel, the frequency of WPB fusion (bin width 1 s) for each of the cells in the upper panel. There was no secretion at 0.1 μm histamine. B, summary of the delays between application and the rise in [Ca²⁺]_i (D₁; ^) and between the rise in [Ca²⁺]_i and exocytosis (D₂; •). C, the frequency of WPB fusion with time (determined from the rise in [Ca²⁺]_i) evoked by 1.0 μm (n = 20 cells, 313 WPBs, 1 s bin widths), 10 μm (n = 15 cells, 515 WPBs, 1 s bin widths) and 100 μm (n = 15 cells, 549 WPBs, 1 s bin widths) histamine. D, the maximal rate of WPB exocytosis for histamine over the concentration range indicated (mean ±s.e.m., n = 6 cells at 0.3 μm, and from between 16 and 24 cells at each of the higher concentrations).

Extent of WPB degranulation observed

To determine the percentage of EGFP-tagged WPBs that exocytosed (the percentage degranulation) cells were exposed to 0.1, 1.0 or 100 μm histamine. The extent of histamine-evoked degranulation increased from zero at 0.1 μm, to 19.63 ± 13.9% (n = 20 cells, range 0.6–50%) at 1.0 μm and to 47.8 ± 21.7% (n = 41 cells, range 7.5–90.0%) at 100 μm. The use of VWF-EGFP or proregion-EGFP to fluorescently label WPBs made no difference to the extent or kinetics of degranulation evoked by histamine or ionomycin, and so data from both constructs were pooled.

Extracellular Ca²⁺ dependence of histamine-evoked WPB exocytosis

Figure 7 shows the effect of removal of external Ca²⁺ on [Ca²⁺]_i and exocytosis evoked by histamine (30 μm). The upper panels show examples of the increase in [Ca²⁺]_i (continuous traces) and frequency of WPB fusion (histograms) in individual cells in the presence (Fig. 7A) or absence (Fig. 7B) of external Ca²⁺, and the lower panels show the times of fusion events pooled from 15 cells under each condition. In the absence of external Ca²⁺ the delays D₁ and D₂ and the maximal rates of WPB exocytosis were not significantly different from those in the presence of external Ca²⁺. However, following the initial burst of exocytosis the rate of WPB exocytosis declined more abruptly and ceased after approximately 30 s in the continued presence of histamine. Comparing data from equivalent numbers of cells showed a reduction of the total number of exocytotic events of almost 60% in the absence of external Ca²⁺ (Fig. 7 lower panels).

Figure 7.

Extracellular Ca²⁺ dependence of histamine-evoked WPB exocytosis Upper panels in A and B show examples of the changes in fura-2 fluorescence ratio (355 nm/380 nm) (continuous traces) and the frequency of WPB fusion (histograms, bin width 1 s) evoked by 30 μm histamine in the presence (A) or absence (B; plus 0.2 mm EGTA) of 1.8 mm external calcium in two single HUVECs. The fura-2 traces were offset so that the increase in [Ca²⁺]_i occurred at time = 0. The lower panels in A (n = 565 WPBs) and B (n = 232 WPBs) show the frequency of WPB fusion determined from the rise in [Ca²⁺]_i (1 s bin widths) for 15 cells in each case.

Time course and extent of ionomycin-evoked WPB exocytosis

Ionomycin is a strong secretogogue known to generate large increases in [Ca²⁺]_i in HUVECs (Morgan & Jacob, 1994). Ionomycin (1 μm) evoked a large, fast rising and more maintained increase in [Ca²⁺]_i (Fig. 8A continuous trace) with a mean delay between application and rise in [Ca²⁺]_i, D₁, of 0.1 ± 0.1 s (n = 6 cells) and delay between the rise in [Ca²⁺]_i and the first exocytotic event, D₂, of 1.62 ± 0.22 s (n = 6). In TIRFM and epifluorescence experiments the delays between application and the first exocytocic event were not significantly different. The frequency of WPB exocytosis in epifluorescence experiments was high during the first 1–3 s after the onset of exocytosis (histograms in Fig. 8A and B) with a mean maximal rate of 10.66 ± 6.59 WPB s⁻¹ (range 3.0–22 WPB s⁻¹, n = 6 cells), before declining over the next 10–20 s to a low level that was maintained in the continued presence of ionomycin. The time of fusion events pooled from six cells is shown in Fig. 5B to illustrate the overall time course for ionomycin-evoked exocytosis. In TIRFM experiments the maximal rate (10.01 ± 10.67 WPB s⁻¹, range 0.47–42.5 WPB s⁻¹, n = 28 cells) and overall time course of exocytosis determined for 1 μm ionomycin was not significantly different from that of epifluorescence experiments. Data combined from all epifluorescence experiments (Deltavision and monochromator based) showed that ionomycin released on average 66.5 ± 17.6% (n = 25 cells, range 16–94%) of fluorescent WPBs, a value significantly higher than that for 100 μm histamine. Thus the rates and magnitude of WPB release evoked by ionomycin are significantly greater than those seen at high histamine concentrations.

Figure 8.

Ionomycin-evoked WPB exocytosis A, the continuous trace is a representative record of the change in fura-2 fluorescence-ratio (355 nm/380 nm) in a single HUVEC during stimulation with 1.0 μm ionomycin. The fura-2 trace was offset so that the increase in [Ca²⁺]_i occurred at time = 0. The frequency of WPB fusion in this cell is shown as a histogram (bin width 1 s). B, summary of the frequency of WPB fusion with time, determined from the rise in [Ca²⁺]_i in 6 cells (447 WPBs, 1 s bin widths).

Discussion

EGFP fused to VWF or proregion as an intra-WPB pH indicator

The luminal pH of WPBs was estimated here from changes in fluorescence of EGFP fused to either VWF or proregion. The apparent pK_a values for VWF-EGFP and proregion-EGFP fluorescence analysed in situ at 24°C were the same (Fig. 2A), very similar to the in vitro value for EGFP determined here at 24°C (5.88), and similar to in vitro and in situ measurements of the pK_a for EGFP determined in other systems (room temperature; Kneen et al. 1998; Llopis et al. 1998; Tompkins et al. 2002). Although it cannot be ruled out that the properties of EGFP are attenuated by the microenvironment of the WPB, the similarities in the apparent pK_a for EGFP fused to VWF or proregion within WPBs, observed here, and for EGFP fused to different proteins in other cell types (Kneen et al. 1998; Llopis et al. 1998; Tompkins et al. 2002) indicate that effects of intraorganelle microenvironment have relatively little effect on the properties of EGFP. We conclude that VWF- or proregion-EGFP are useful tools for studying intra-WPB pH.

Changes in WPB EGFP fluorescence and the formation of the WPB fusion pore

The first event detected during WPB exocytosis was an increase in intra-WPB EGFP fluorescence. In other secretory systems such increases in intragranule targeted EGFP fluorescence have been attributed to the formation of a fusion pore with subsequent loss of protons to the extracellular solution. This interpretation is supported here for single WPBs by the following: (1) there was a dependence of the increase in EGFP fluorescence on external pH; (2) there was an accumulation of the extracellular fluid phase marker Alexa-647 within WPBs during the hormone-evoked increase in EGFP fluorescence; (3) the time course of the alkalinization suggests a flux of protons through a fusion pore rather than a slow pH change that would result from gradual proton removal by a slow exchange process; (4) in large WPBs (> 2 μm in length), the fluorescence increase was seen to proceed as a wave along the length of the organelle (see movie 7 of supplemental materials), an observation consistent with the exit of protons to the exterior through a pore at one end of the WPB, and inconsistent with a gradual proton removal over the surface area of the organelle; and (5) in contrast to ammonium chloride addition, histamine and ionomycin alkalinized only those WPBs that subsequently changed their shape (of which ∼90% went on to released their contents, see above).

Function of the acidic WPB lumen

The low pH within WPBs may be important for multimerization and condensation of VWF (Wagner et al. 1986). In vitro studies have shown that VWF multimerization occurs optimally under prolonged exposure to acidic conditions (pH 5.8–5.4), suggesting that low pH affects VWF directly rather than through pH-dependent enzymes or sorting/packaging pathways (Mayadas & Wagner, 1989). Optimal conditions for protein condensation, aggregation and crystallization require high protein concentrations and correlate with an environmental pH close to the isoelectric point (pI) for the protein (Kantardjieff & Rupp, 2004). The predicted pI for mature VWF (aa764–2812) and proregion (aa23–761; calculated for human VWF using the pI tool at Expasy (http://us.expasy.org/tools/pi_tool.html), are pH 5.39 and pH 4.99, respectively. Although these predictions do not take into account post-translational modifications or protein folding within the WPBs that might affect the charge characteristics of VWF or proregion, they fall in the same range for pH_g reported here for both resting WPBs and WPBs that underwent exocytosis (and can thus be classed as mature secretory granules; MSG), and are similar to the apparent pI for plasma derived multimeric VWF (pH 5.9–5.7; Fulcher et al. 1983). In cells 24–48 h post-nucleofection a small number of WPBs had pH_g close to that of the GA (∼pH 6.5–6.2; Wu et al. 2001). Immature secretory granules (ISG) are reported to be less acidic than MSG (pH 6.3–5.7 versus pH 5.5–5.0; see Wu et al. 2001 and references therein) raising the possibility that a component of the less acidic organelles might be immature WPBs. The ability to recruitment Rab27a to the WPB membrane is thought to be property of mature WPBs (Hannah et al. 2003). At ∼5 h post-nucleofection < 5% of VWF-EGFP containing WPBs have Rab27a on their membranes, while > 80% had recruited Rab27a by 24 h post-nucleofection (Hannah et al. 2003). The mean pH_g of WPBs 6.5–7 h post-nucleofection was significantly less acidic than that of WPBs 24–48 h post-nucleofection, consistent with these organelles being ISG. Although the majority of immature WPBs had pH_g between pH 6.6 and 5.9, close to that of the GA, there was even at these early times a clear tail in the distribution toward values associated with mature WPBs suggesting that acidification can occur rapidly following organelle formation, contributing to conditions conducive for efficient VWF multimerization and condensation, and presumably membrane maturation.

Proton buffering capacity of WPBs

Experiments in which pH_g of mature WPBs were determined also yielded values for β, the ability of organelle matrix proteins to buffer protons from solution. VWF and its proregion constitute > 98% of the WPBs core proteins (Wagner, 1990; Sadler, 1998), both are highly negatively charged, and prime candidates as the major proton buffers of WPBs. The magnitude of β determined by the NH₄Cl pulse approach will depend on a number of factors, in particular pH_g and the properties (e.g. pK_a) and absolute amount of the buffer species (Roos & Boron, 1981). WPB pH_g, WPB size, and the numbers of rod shaped core structures of WPBs thought to reflect polymerized VWF vary considerably (Weibel & Palade, 1964; Elgjo et al. 1975; Kagawa & Fujimoto, 1987). This being the case it is not surprising that a large variability in β between individual WPBs was seen. The mean value for β of WPBs, 55 mm (pH unit)⁻¹, is similar to estimates of endosomal (46–50 mm (pH unit)⁻¹; Gekle & Silbernagl, 1995; Rybak et al. 1997) and lysosomal (46 mm (pH unit)⁻¹; Ishizaki et al. 2000) β, but higher than that reported for MSG of AtT20 cells (20 mm (pH unit)⁻¹; Wu et al. 2001), and for cytoplasm, ER and GA (6–40 mm (pH unit)⁻¹; Gekle & Silbernagl, 1995; Grabe & Oster, 2001; Wu et al. 2001).

Morphological changes in WPBs during exocytosis

An abrupt change in morphology of the WPB, from cigar-shaped to an approximately spherical shape, was seen following the increase in EGFP fluorescence due to fusion pore formation. Ultrastructural studies show that the major WPB core protein, VWF, is highly organized within mature WPBs, forming discrete rod-like structures (Wagner, 1990; Sadler, 1998), and the shape change seen here is likely to reflect a change in the organization of this protein. Although the mechanisms underlying the rapid collapse followed by expulsion of VWF (Hannah et al. 2005) from WPBs during exocytosis have not yet been elucidated, there may be parallels with the processes underlying the rapid expansion and explosive release of mucins during epithelial cell exocytosis (reviewed in Verdugo, 1990). The mucin family of glycoproteins are evolutionarily related to VWF (Desseyn et al. 2000), and like VWF, form a complex of large molecular weight polyionic chains that are highly condensed within their storage granule. A high intragranule [Ca²⁺] is thought to play an important role in maintaining mucin polymers in a condensed state by shielding the fixed anionic charges of the polymers (Verdugo et al. 1987a, (b). A similar shielding function has been suggested for Ca²⁺, H⁺, ATP, and catecholamines in a number of other secretory organelles (reviewed in Verdugo, 1990). Mucin granule fusion leads to the release of cationic shielding species and water entry, resulting in a rapid (subsecond) swelling and an explosive release of mucin onto the cell surface driven by the unshielded anionic charges of the untangled mucin chains (reviewed in Verdugo, 1990). Similarities in the structure and negative charge characteristics of mucins and VWF, suggest a similar process may also underlie the abrupt disordering of VWF and its subsequent explosive discharge onto the cell surface (see movies 8a and 8b, and Hannah et al. 2005). Although the nature of the shielding species in WPBs are not known, H⁺ and Ca²⁺ are likely candidates. Acute alkalinization of the WPB using pH 8.0 calibration solutions (e.g. Fig. 3A) or addition of > 20 mm NH₄Cl, which causes a rise of pH_g to > pH 7.40 (as judged by a 2.70 ± 0.73-fold increase in WPB EGFP fluorescence (n = 40 WPBs from 6 cells), does not result in an abrupt change in WPB morphology (see Supplemental material Fig. 1), but causes rounding up on a much slower time scale, typically many minutes to hours (see also Michaux et al. 2006). This suggests that neutralization of pH_g is not sufficient to account for the abrupt change in organization of VWF following agonist-evoked WPB fusion. Calcium ions are another possible candidate. In vitro, multimerization of VWF requires, in addition to a low pH, a high [Ca²⁺] (Mayadas & Wagner, 1989). It has been suggested that multimerization may occur most efficiently within WPBs, which in common with mucin granules are thought to contain high [Ca²⁺] (Verdugo et al. 1987a; Mayadas & Wagner, 1989).

Kinetics of histamine evoked-elevations of [Ca²⁺]_i and WPB exocytosis

Histamine is a potent WPB secretogogue that operates to elevate intracellular free calcium ion concentration ([Ca²⁺]_i) through a G-protein–phospholipase C coupled H1 receptor and production of the intracellular calcium-mobilizing second messenger inositol trisphosphate (InsP₃) (Hamilton & Sims, 1987; Brock & Capasso, 1988; Carter et al. 1998). Histamine-evoked elevations of [Ca²⁺]_i occurred with a delay, that reduced to ∼0.5 s at high concentrations (100 μm) presumably reflecting the time taken between receptor binding and the generation and action of intracellular InsP₃. Following the increase in [Ca²⁺]_i there was a further delay before WPB exocytosis occurred, on average ∼10 s at low histamine concentrations (0.3 μm) and ∼1.6 s at 100 μm histamine. This delay, together with the zero basal exocytosis in resting cells, indicate that a slow process in Ca²⁺-driven exocytosis must operate prior to final fusion. The nature of the slow process has yet to be determined. Despite causing an increase in [Ca²⁺]_i, 0.1 μm histamine and in many cases 0.3 μm histamine (12/20 cells) failed to evoke WPB exocytosis indicating that a threshold exists for Ca²⁺-driven WPB exocytosis. Evidence for such a threshold has come from biochemical studies in permeabilized HUVECs that indicated VWF secretion requires > 0.8 μm free Ca²⁺, and that secretion increases steeply in the range 1–20 μm free [Ca²⁺] (Scrutton & Pearson, 1989; Birch et al. 1992). Quantitative measurements of [Ca²⁺]_i in HUVECs, and endothelial cells derived from other vessels, using the low affinity Ca²⁺ indicator furaptra have shown that peak increases in [Ca²⁺]_i during hormone action occur in a similar range (2.5–12 μm free Ca²⁺; Carter & Ogden, 1994, 1997; Carter et al. 1998) and high [Ca²⁺]_i (5–10 μm) is required to evoke exocytosis in HUVECs, measured as changes in membrane capacitance (Carter et al. 1998; Zupancic et al. 2002).

A maximally effective concentration of the receptor-independent agonist ionomycin (1 μm) evoked an elevation of [Ca²⁺]_i with a shorter delay, consistent with a more direct action to elevate [Ca²⁺]_i (see below). Following the increase in [Ca²⁺]_i there was a further delay of ∼1.6 s prior to the onset of exocytosis, consistent with a slow step in Ca²⁺-activated WPB exocytosis. The high rate of WPB exocytosis evoked by ionomycin, indicates the existence of a significant reserve capacity in the secretory system, inaccessible to histamine even at high concentrations. This extra capacity may come into play during cell injury where [Ca²⁺]_i would rise to very high levels signalling the need for a rapid secretory response. Histamine caused a concentration-dependent increase in the extent of WPB degranulation, but failed to release as many WPBs as ionomycin. Differences in the delays, maximal rates and overall extent of degranulation for histamine and ionomycin are most likely due to differences in the magnitude of the elevation in [Ca²⁺]_i produced. In HUVECs the increase in [Ca²⁺]_i produced by 1 μm ionomycin is independent of InsP₃ production (Jaffe et al. 1987), and is thought to be due principally to a direct and unregulated action of this ionophore on the membrane of the internal Ca²⁺ store, rather than the plasma membrane (see Morgan & Jacob, 1994 and references therein). The resulting [Ca²⁺]_i elevation produced by 1 μm ionomycin is larger in amplitude than that produced by histamine (100 μm) (e.g. Figs 6A and 8; see also Morgan & Jacob, 1994).

Consistent with our previous findings (Zupancic et al. 2002), the maximal rates of WPB exocytosis varied considerably, but were significantly lower than that for Ca²⁺-driven exocytosis in neuroendocrine cells, where measured rates of 300–400 granules s⁻¹ have been reported (see Henkel & Almers, 1996 and references therein). The rates of WPB exocytosis were determined from EGFP containing organelles; however, because not all WPBs were fluorescently labelled, the absolute magnitude of the rates of exocytosis, in WPB s⁻¹, will be underestimated. The true rate of exocytosis can be estimated by knowing the mean proportion of WPBs that are fluorescently labelled; in these experiments this was ∼78%, and assuming that EGFP-labelled and non-labelled WPBs behave in the same way. Thus, for low histamine concentrations where the mean rate determined from fluorescent WPBs is low the corrected rate increases modestly from 1.2 WPB s⁻¹ to 1.5 WPB s⁻¹, at high histamine concentrations (100 μm) from ∼3.7 WPB s⁻¹ to ∼4.7 WPB s⁻¹, and for ionomycin from ∼10.7 WPB s⁻¹ to ∼13.7 WPB s⁻¹. The reasons why endothelial cells have a comparatively slow secretory mechanism are not clear but may reside in the nature and site of action of the secreted factors. Endocrine derived hormones typically act on disparate tissues throughout the body, and consequently may require a high rate of exocytosis in order to rapidly deliver sufficient material into the circulation to overcome dilution, sequestration and clearance effects. In contrast, acutely secreted VWF adheres efficiently to and acts primarily at the surface of the endothelial cell from which it is released and this, in conjunction with the large size of WPBs and highly condensed state of VWF, may allow a comparatively slow rate of vesicle fusion to operate and still efficiently delivery sufficient material to perform the local physiological functions required.

Data presented here show that ∼90% of exocytosed WPBs release their contents within 60 s after stimulation, and that the secretory response is essentially complete by 500 s in the continued presence of stimulus. This time course contrasts with those determined by biochemical assays for soluble VWF, a common assay for WPB exocytosis, that typically show a slow onset and protracted increase in VWF immunoreactivity lasting several tens of minutes during continued stimulation (Hamilton & Sims, 1987). The discrepancy is most likely due to a very slow dissociation of VWF from the surface of the endothelial cell into solution following its release by exocytosis (Hannah et al. 2005).

Removal of external Ca²⁺ had little effect on the initial kinetics of WPB exocytosis, consistent with the initial burst of secretion being due primarily to release of Ca²⁺ from internal stores. However, the frequency of fusion events declined more abruptly and the low frequency events seen during prolonged stimulation were abolished resulting in a more rapid termination and reduced extent of the secretory response, consistent with biochemical studies of secreted VWF (Hamilton & Sims, 1987).

Conclusions

The prompt onset of WPB exocytosis allows the endothelial cell to deliver adhesive and inflammatory molecules acutely to the cell surface following vascular challenge, essential for efficient regulation of haemostasis, for the recruitment of white cells to fight infection and for maintenance of the integrity of the endothelial lining. The observation that histamine, even at high concentrations, was unable to drive WPB exocytosis at its maximal rate (as indicated by ionomycin action) shows that a significant degree of reserve capacity exists that may come into play during cell injury, where [Ca²⁺]_i levels will rise to very high levels signalling the need for a faster and larger response. The efficient cell surface adhesion of VWF, required for its local action at the cell surface, may allow the endothelial cell to operate at comparatively low rates of vesicle fusion. The low WPB pH_g determined here is within the optimal range for VWF multimerization and condensation, processes that ultimately lead to the formation of an immobile (Kiskin et al. 2004) quasi-crystalline structure within the mature WPB (Hannah et al. 2002). A loss of cationic shielding species (e.g. H⁺) coupled with hydration of the WPB lumen following fusion is likely to underlie the collapse of quasi-crystalline VWF and the ejection of this adhesive protein onto the cell surface. This process may be important in dispersing VWF so as to maximize the surface area over which it acts.

Supplementary material

Online supplemental material for this paper can be accessed at:

http://jp.physoc.org/cgi/content/full/jphysiol.2007.132993/DC1

and

http://www.blackwell-synergy.com/doi/suppl/10.1113/jphysiol.2007.132993