Differential modulation of the phases of a Ca2+ spike by the store Ca2+-ATPase in human umbilical vein endothelial cells

HOME

HELP

FEEDBACK

SUBSCRIPTIONS

Differential modulation of the phases of a Ca²⁺ spike by the store Ca²⁺-ATPase in human umbilical vein endothelial cells

Anthony J. Morgan and Ron Jacob

Vascular Biology Research Centre, Physiology Group, Biomedical Sciences Division, King's College London, London W8 7AH, UK

Received 14 May 1998; accepted after revision 30 July 1998.

ABSTRACT

Top
Abstract
Introduction
Methods
Results
Discussion
References

Histamine-stimulated cytosolic free Ca²⁺ ([Ca²⁺]_i) oscillations in human umbilical vein endothelial cells (HUVECs) comprise repetitive spikes generated by pulsatile release from stores. We have investigated the roles of the store Ca²⁺-ATPases in regulating both the upstroke and downstroke of a Ca²⁺ spike.

The sarco-endoplasmic reticulum Ca²⁺-ATPase (SERCA) inhibitor cyclopiazonic acid (CPA) dramatically affected oscillations whereas inhibition of the plasma membrane Ca²⁺-ATPase (PMCA) with La³⁺ had little effect. This and other evidence suggested that the downstroke of a spike is predominantly mediated by SERCA.

Artificial [Ca²⁺]_i spiking generated by repetitive pulsatile application of 0�3 µM histamine in Ca²⁺-free medium did not cause net loss of Ca²⁺ from the cell whereas repetitive pulsatile application of 1 and 10 µM histamine did, with the higher concentration being more effective. We conclude that there is an inverse relationship between stimulus intensity and relative SERCA activity.

For a Ca²⁺ transient, the initiation of release was suppressed by SERCA during either the lag phase or the interspike period (ISP) since: (i) the ISP was shortened by low CPA concentrations, (ii) higher concentrations of CPA stimulated an explosive Ca²⁺ release when applied during the ISP but not when applied in the absence of agonist, and (iii) CPA synchronized the initial Ca²⁺ response to a low histamine dose (even recruiting silent, histamine-unresponsive cells).

Two aspects of the regenerative upstroke of a spike were differently affected by SERCA inhibition: Ca²⁺ wave velocity was entirely unaffected by CPA whereas the local rate of rise was increased.

The [Ca²⁺]_i at which a Ca²⁺ spike terminated depended on SERCA since CPA dose dependently enhanced the peak [Ca²⁺]_i.

We conclude that SERCA plays a powerful and dynamic role in regulating [Ca²⁺]_i oscillations in HUVECs. SERCA differentially modulates the phases of Ca²⁺ release in addition to bringing about the falling phase of a Ca²⁺ spike.

INTRODUCTION

Top
Abstract
Introduction
Methods
Results
Discussion
References

In many cell types stimulation by agonist at submaximal concentrations causes oscillations of cytosolic free Ca²⁺ ([Ca²⁺]_i). These oscillations are generated by a periodic brief release of Ca²⁺ from stores; in many cases the frequency but not the amplitude of the [Ca²⁺]_i transients is modulated by the concentration of agonist (frequency-encoded signalling; Jacob et al. 1988; Petersen et al. 1994). Investigation of the mechanisms involved in these oscillations has tended to focus on the complex regulation of Ca²⁺ release whether via the inositol 1,4,5-trisphosphate (IP₃) or the ryanodine receptor (Petersen et al. 1994; Berridge, 1997). Less attention has been paid to the influence of the Ca²⁺-ATPases responsible for removing Ca²⁺ from the cytosol, with their role often being dismissed as a mere homeostatic mechanism which acts to buffer [Ca²⁺]_i.

There are two homologous families of Ca²⁺-ATPases: the plasma membrane Ca²⁺-ATPase (PMCA) which extrudes Ca²⁺ from the cell and the sarco-endoplasmic reticulum Ca²⁺-ATPase (SERCA) which resequesters Ca²⁺ into internal stores (Petersen et al. 1994). Both families consist of multiple isoforms derived from distinct genes and/or alternative splicing (Grover & Khan, 1992) although the functional consequences of these Ca²⁺-ATPase isozymes for a given cell type are largely unknown. This study focuses on the role of the SERCA family in regulating oscillations in human umbilical vein endothelial cells (HUVECs).

In endothelial cells with normal levels of [Na⁺]_i, Na⁺-Ca²⁺ exchange does not move significant amounts of Ca²⁺ across the cell membrane (Jacob et al. 1988; Sage et al. 1991) so that Ca²⁺-ATPases are responsible for maintaining basal [Ca²⁺]_i in unstimulated cells and for restoring it after stimulation. This 'homeostatic' role of the Ca²⁺-ATPases is an essential component of frequency-encoded signalling because of the need to effect periodic relaxation of the [Ca²⁺]_i spikes. However, the relative homeostatic roles of PMCA and SERCA may differ substantially according to cell type or species thus affecting the fate of released Ca²⁺. For example, SERCA activity predominates over PMCA activity in rat pancreatic acinar cells resulting in oscillations that are well maintained in Ca²⁺-free medium (Zhang & Muallem, 1992); in contrast, the PMCA activity of mouse pancreatic acinar cells predominates causing oscillations to run down quickly in Ca²⁺-free medium (Tepikin et al. 1992).

Ca²⁺ regulates its own movement at many locations such as the IP₃ receptor channel and the store-regulated Ca²⁺ entry channels, sometimes by both positive and negative feedback (Petersen et al. 1994; Berridge, 1995, 1997). It is quite likely that Ca²⁺ movement via Ca²⁺-ATPases affects such processes; this we term the 'modulatory' role of Ca²⁺-ATPases. A few reports have addressed the importance of SERCA activity affecting the release process itself during oscillations and waves (Petersen et al. 1993; Camacho & Lechleiter, 1993; Simpson & Russell, 1997). With the potential for SERCA to influence various facets of the spiking process, we have therefore undertaken a closer examination of its role in HUVECs. We consider each oscillatory spike in terms of distinct phases, namely the initiation of Ca²⁺ release, the regenerative upstroke, termination of release and restoration of basal [Ca²⁺]_i. We conclude that SERCA has a powerful homeostatic effect coupled to modulatory effects on the initiation, localized regeneration and termination of IP₃-induced release. By contrast, SERCA does not modulate Ca²⁺ wave propagation.

METHODS

Top
Abstract
Introduction
Methods
Results
Discussion
References

Cell culture

Cells isolated from human umbilical vein were cultured on glass coverslips and used at passage 1 or 2 as previously described (Jacob et al. 1988). Cells were maintained at 37�C in an atmosphere of 95 % air-5 % CO₂ in Medium 199 supplemented with 10 % fetal calf serum, 10 % newborn calf serum, 50 µg ml^-1 streptomycin, 50 i.u. ml^-1 penicillin, 2 mM L-glutamine and 20 µg ml^-1 endothelial cell growth factor (ECGF) prepared from bovine brain (Maciag & Weinstein, 1984). The HUVEC line, ECV304 (obtained from Dr K. Takahashi, Department of Biochemistry, National Defence Medical College, Saitama, Japan), passage 110-115, was cultured as described before (Morgan & Jacob, 1994) in the same medium as HUVECs but without newborn calf serum or ECGF. Experiments were performed on cells 2-6 days after plating.

Measurement of intracellular Ca²⁺ ([Ca²⁺]_i)

For most experiments, HUVECs were loaded with 1 µM fura-2 AM in Hepes-buffered Dulbecco's minimum essential medium plus 20 % fetal calf serum for 45 min at room temperature (18-22�C) (Morgan & Jacob, 1996). ECV304 cells were loaded similarly but with 2 µM fura-2 AM for 60 min (Morgan & Jacob, 1994). After loading, cells were maintained in a Hepes-buffered balanced salt solution (HBSS, 1 %) of the following composition (mM): 145 NaCl, 5 KCl, 1 MgCl₂, 1 CaCl₂, 10 Hepes, 10 glucose; plus 1 % (w/v) bovine serum albumin (BSA); pH 7�4. They were then transferred for measurement into the same solution (HBSS) but with a lower BSA concentration (0�1 %). Population experiments were performed using coverslips lodged diagonally in a cuvette placed in the thermostatically controlled (37�C) holder of a rotating wheel spectrophotometer (Cairn Research Ltd, Faversham, Kent, UK). Excitation was at 340 and 380 nm, with emission detected at 510 nm; data were collected at 2 Hz except where stated. For single cell experiments, coverslips were mounted on the thermostatically controlled stage (37�C) of an inverted epifluorescence microscope and superfused with HBSS by gravity feed. The microscope was linked to a Cairn spectrophotometer with the same excitation wavelengths as for the population measurement but emission was collected at > 470 nm (Morgan & Jacob, 1996). In some experiments, the rate of exchange of the bath superfusate was simultaneously monitored along with [Ca²⁺]_i by adding 20-30 nM fluorescein as a tracer to the HBSS in one of the perfusion lines and monitoring its appearance at the cell by additional excitation at 440 nm. [Ca²⁺]_i is reported as the 340 nm/380 nm ratio (R340/380) in most experiments, except where indicated when an in vitro calibration was used (Morgan & Jacob, 1994). Autofluorescence was estimated by addition of 2 mM Mn²⁺ plus either 1 µM ionomycin, or occasionally 10 µM histamine, in Ca²⁺-free medium (Morgan & Jacob, 1994). Ca²⁺-free media were prepared by omission of Ca²⁺ without addition of EGTA except where stated.

Single cell imaging and Ca²⁺ waves

Images were collected using an extended Isis M intensified CCD camera (Photonic Sciences, Robertsbridge, East Sussex, UK) and a Pixel Grabber board (Perceptics, Knoxville, TN, USA) hosted by a Macintosh Quadra 800 or an IMAXX-PCI board (Precision Digital Instruments, Redmond, WA, USA) hosted by a Macintosh 8500/120, both running Ionvision software (Improvision, Coventry, UK). Coverslips were mounted on the thermostatically controlled stage of a Zeiss Axiovert microscope equipped with a × 40 oil-immersion fluor achrostigmat objective (NA 1�3). Excitation was via an LEP dual filter wheel (Ludl, Hawthorne, NY, USA) equipped with 350 and 380 nm filters and a range of neutral density filters. Emission was measured > 400 nm. In those experiments analysing the global [Ca²⁺]_i signal from many individual cells at a time, a 350 nm/380 nm ratio image was acquired every 3 s. For Ca²⁺ wave analysis, HUVECs were loaded with 2-3 µM fura-2 AM for 45 min at room temperature and superfused at 30�C (minimizing dye loss) to optimize image quality over the course of the experiment. This loading regime gave similar results to those obtained at 37�C with cells that were less fura-2 loaded.

To measure Ca²⁺ wave velocity, images were captured at full video rate (25 Hz) onto videotape at 380 nm excitation. Data were digitized and then analysed using NIH Image and custom written macros. The stages of analysis were: (1) ratio all images against the initial pre-stimulus image (F/F₀) to reduce variation in intensity due to cell thickness; this produces an image sequence that shows relative changes but is no longer an absolute measure of [Ca²⁺]_i; (2) visually assess the direction of wave propagation and define a rectangular box (typically 25 pixels in width, where 1 pixel equv 0�45 µm) with its long axis aligned with the direction of wave propagation and spanning the cell; (3) for each F/F₀ image, average across the width of the rectangular box to produce a line plot of average intensity along the axis down which the wave is travelling; (4) stack these line plots to produce an image with distance along one axis and time along the other; (5) use the automatic thresholding to create a binary image showing the progression of the wavefront with time; (6) fit a regression line to the wavefront, the gradient of which yields the wave velocity.

Assessment of the degree of cell synchronization

Imaging experiments revealed cell-to-cell variation in the latency of the response to histamine. We quantified this by calculating the S.D. of the times at which the responses were initiated in each of the cells: the more synchronous the responses, the smaller the S.D. The latencies are likely to follow an offset Poisson distribution in which case the S.D. would reflect the mean latency.

Results are expressed as means � S.E.M., with significance tested by Student's t test, which was two-tailed and paired where appropriate. P < 0�05 was considered significant.

Materials

Fura-2 AM was obtained from Molecular Probes (Eugene, OR, USA) and Calbiochem (Nottingham, UK) and fura-2 (K⁺ salt) from Molecular Probes. Cyclopiazonic acid (CPA), thapsigargin and ionomycin were from Calbiochem. LaCl₃ and histamine were from Sigma. All tissue culture reagents were from Gibco. All other reagents were of analytical grade.

RESULTS

Top
Abstract
Introduction
Methods
Results
Discussion
References

PMCA inhibition fails to inhibit oscillations

Ca²⁺ oscillations can occur in HUVECs in the absence of extracellular Ca²⁺ (Ca) with little detectable loss of Ca²⁺ from internal stores (Morgan & Jacob, 1996). This suggests that only a small fraction of Ca²⁺ is lost from the cell during a Ca²⁺ spike implying that resequestration of released Ca²⁺ by SERCA activity predominates over extrusion of released Ca²⁺ by PMCA. We addressed this issue more directly using several approaches starting with the effect of 1 mM La³⁺ (an inhibitor of all Ca²⁺ movements across the plasma membrane including PMCA; Kwan et al. 1991) on oscillations. Experiments with La³⁺ and the associated controls were performed in nominally Ca²⁺-free medium.

To establish that La³⁺ could inhibit the PMCA and 'seal' the cell to Ca²⁺ we investigated the effect of 1 mM La³⁺ on the ability of a 15 s exposure to a supramaximal concentration (50 µM) of histamine to empty stores in Ca²⁺-free medium, assessed by the response to a second exposure to 50 µM histamine 100 s later (Fig. 1). In the absence of La³⁺ (Fig. 1A) the second histamine response was only 31 � 10 % (n = 5) of the first response, indicating that the first exposure had substantially depleted the stores. In the presence of La³⁺ (Fig. 1B) the second response was 105 � 5 % of the first response (n = 5, P < 0�001 vs. -La³⁺) indicating that the released Ca²⁺ had been resequestered into agonist-sensitive stores rather than extruded from the cell. These results also indicate that released Ca²⁺ is not substantially taken up into an agonist-insensitive internal compartment. The slight decline in R340/380 of the second response in Fig. 1B in the presence of La³⁺ was somewhat variable and may have been due to inhibition of release at high [Ca²⁺]_i resulting in reuptake into internal stores in some cells (preliminary data suggested this was prevented by CPA; not shown).

View larger version
[in this window]
[in a new window]

Figure 1. Effect of the Ca²⁺ extrusion inhibitor La³⁺ on net Ca²⁺ loss in single HUVECs
Single HUVECs were maximally stimulated for 15 s with 50 µM histamine (His) in Ca²⁺-free medium in either the absence (A) or presence (B) of 1 mM La³⁺. Following agonist washout and recovery, the state of the internal stores was probed 100 s later with the same concentration of histamine. Records are typical of 5 experiments using photometric recording. In this and Fig. 3 onwards, Ratio refers to R_340/380.

Oscillations in Ca²⁺-free medium run down in frequency even though the amount of stored Ca²⁺ is not substantially depleted (Morgan & Jacob, 1996). This run-down could be due to a marginal loss of Ca²⁺ in which case inhibition of PMCA should prevent the run-down by preventing the loss of Ca²⁺ from the cell. We examined whether oscillations occurred at all in the presence of La³⁺ and if they did, whether the rate of run-down was affected by La³⁺. Cells were stimulated with 0�3 µM histamine for 10 min under three conditions: in Ca²⁺-containing HBSS as a control (where cells show little run-down; Morgan & Jacob, 1996), in Ca²⁺-free HBSS (where cells show marked run-down; Morgan & Jacob, 1996) and in Ca²⁺-free HBSS together with 1 mM La³⁺ to inhibit extrusion. Figure 2B shows that oscillations ran down in the absence of Ca. Consistent with a predominant SERCA activity during oscillations, oscillations were observed in cells in the presence of La³⁺ (Fig. 2C). However, they were only observed in 53 % (74/140) of cells in the presence of La³⁺ compared with 76 % (72/95) in 1 mM Ca and 86 % (114/133) in Ca²⁺-free medium; this was because responses in the presence of La³⁺ were more heterogeneous with some cells showing a sustained [Ca²⁺]_i response.

View larger version
[in this window]
[in a new window]

Figure 2. Lack of effect of La³⁺ on Ca²⁺ oscillations
Single HUVECs were stimulated for 10 min with 0�3 µM histamine in medium containing 1 mM Ca²⁺_o (A), nominally Ca²⁺-free medium (B), or Ca²⁺-free medium supplemented with 1 mM La³⁺ (C). Store size was probed at the end of the run by stimulating with 10 µM histamine in the absence of Ca²⁺_o. La³⁺ had little effect upon oscillations compared with the Ca²⁺-free medium. Traces are representative of at least 48 cells using imaging. Ratio refers to R_350/380.

The rate of run-down was analysed by plotting the interspike periods (ISPs) against the times of the spikes at the end of each ISP and calculating the slope for each individual cell (dISP/dt; Morgan & Jacob, 1996); a positive value indicates run-down. In Ca²⁺-free medium oscillations ran down (dISP/dt = 0�15 � 0�01, n = 99, P < 0�002 vs. Ca²⁺-containing medium; previously reported by Morgan & Jacob, 1996). Interestingly, when PMCA was inhibited by 1 mM La³⁺ in Ca²⁺-free HBSS, spiking frequency ran down at the same rate as in the absence of La³⁺ (dISP/dt = 0�14 � 0�03, n = 48, P > 0�6 vs. Ca²⁺-free medium) (Fig. 2). The lack of effect of La³⁺ on the rate of run-down in Ca²⁺-free medium again suggests that PMCA is not the primary Ca²⁺-ATPase responsible for returning [Ca²⁺]_i to basal during oscillations.

At the end of each experiment cells were challenged with 10 µM histamine. The response was more sustained in La³⁺-treated cells compared with controls confirming PMCA inhibition. La³⁺ did not act internally since entry of La³⁺ is readily detected by its high-affinity binding to fura-2 (Haugland, 1992) and this was not seen even in Ca²⁺-free medium after maximal store depletion (Fig. 1; n = 7, P > 0�1). Also consistent with La³⁺ acting extracellularly and upon the PMCA was the rapid cessation of the maintained [Ca²⁺]_i response to histamine when La³⁺ was removed (Fig. 1B).

Repetitive application of histamine shows no reduction of the response at low concentrations

We also investigated the role of SERCA vs. PMCA activity during low dose stimulation using a sequence of ten repetitive 30 s pulses of 0�3 µM histamine in Ca²⁺-free medium (cf. Cheek et al. 1994; Bootman et al. 1994). We assessed the degree of store emptying by examining the decline in the [Ca²⁺]_i responses; maximal histamine responses were also measured before and after the sequence of pulses as another measure of store depletion. This protocol reduces the extent of any agonist receptor desensitization (unlikely since HUVEC histamine receptors desensitize over a matter of hours; McCreath et al. 1994), but, more importantly, is likely to be a more efficient stimulus than continuous application by minimizing inactivation of release at the IP₃ receptor, as argued for caffeine at the ryanodine receptor (Cheek et al. 1994). Also, where submaximal agonist stimulation selectively releases Ca²⁺ from a subpopulation of stores which is characterized by heightened sensitivity to IP₃ (Morgan & Jacob, 1996), this protocol selectively probes this subpopulation and so is more likely to reveal store depletion. In contrast to similar experiments in other cell types (Cheek et al. 1994; Bootman et al. 1994) pulses of 0�3 µM histamine did not cause any reduction in spike height even after ten pulses (Fig. 3A and C) and a total period of stimulation (300 s) for which a continuous application showed run-down in spiking frequency (Fig. 2B; Morgan & Jacob, 1996). Moreover, the stores remained full at the end of the run since a subsequent 10 µM histamine spike was 95 � 6 % (n = 14, P > 0�3) of a test response at the start of the experiment. We again conclude that during 0�3 µM histamine stimulation SERCA resequesters the released Ca²⁺ and keeps the stores full.

We repeated this protocol with a sequence of pulses at a higher but still submaximal histamine concentration of 1 µM. The initial response to 1 µM histamine was greater than that to 0�3 µM, but only by a small amount (percentage of response to 10 µM histamine, 0�3 µM: 28 � 2 %, n = 15; 1 µM: 39 � 4 %, n = 13; P < 0�05). In contrast to the response to 0�3 µM histamine, the response to 1 µM showed a progressive reduction in height (Fig. 3B and C) which correlated with substantial depletion of Ca²⁺ stores assessed by the subsequent response to 10 µM histamine (percentage of the first 10 µM histamine response: 31 � 4 %, n = 11; P < 0�001 vs. 0�3 µM). The increased depletion could have been due to the spike width being greater for 1 µM than for 0�3 µM histamine, presumably because of a reduced latency at the higher concentration. We therefore reduced the stimulation time to 20 s with 1 µM histamine, and even when spike duration was well-matched for the two concentrations (mean spike width: 0�3 µM, 20 � 1 s; 1 µM (30 s pulse), 25 � 1 s; 1 µM (20 s pulse), 21 � 0 s; n = 3), a more substantial depletion was evoked by a sequence of pulses of 1 µM histamine (Fig. 3). These data show that the balance between SERCA and PMCA activity depends on the agonist concentration.

View larger version
[in this window]
[in a new window]

Figure 3. Relationship between stimulus intensity and pool depletion in single HUVECs
A, after a brief test of a maximal histamine response in 1 mM Ca²⁺_o, cells were stimulated with a sequence of 10 30 s pulses of 0�3 µM histamine in nominally Ca²⁺-free medium. Pool depletion was minimal as determined by the constant pulse response as well as the magnitude of a subsequent response to maximal histamine in Ca²⁺-free medium. B, similar experiment to A using 1 µM histamine. C, summary of pulse data in A and B normalized to the first submaximal response where 20 and 30 s refer to the different pulse times tested with 1 µM histamine (the interpulse interval was increased from 50 to 60 s when a 20 s pulse was used). Each record is representative of at least 11 experiments using the photometric system. Data are expressed as means � S.E.M.

SERCA inhibition synergistically facilitates store emptying

To complement the PMCA inhibition experiments we used the selective, but chemically dissimilar, SERCA inhibitors CPA and thapsigargin to block reuptake into endothelial cell Ca²⁺ stores (Gericke et al. 1993; Morgan & Jacob, 1994; Morgan & Jacob, 1996). In initial experiments, we examined the ability of SERCA inhibitors to release Ca²⁺. Table 1 shows the dose-response relationship of the two agents with respect to either the change in [Ca²⁺]_i ( $Delta$ [Ca²⁺]_i) or the maximal rate of rise observed in populations of HUVECs. In the light of these results, we used thapsigargin at 200 nM and CPA at 30 µM when a maximal inhibition of SERCA was required, concentrations similar to those used by others (Jackson et al. 1988; Demauraux et al. 1992; Gericke et al. 1993). We used CPA in most experiments since it was less problematic in terms of binding to the perfusion tubing, and has a lower incidence of reported side effects than thapsigargin (Nelson et al. 1994; Palmer et al. 1994; Du et al. 1996). The kinetics of store emptying were far slower with CPA than with histamine, both in terms of the initial rate of rise and the rate of emptying of the stores assessed by a subsequent response to 10 µM histamine (Fig. 4); at least a 5 min exposure to CPA was required to empty the agonist-sensitive stores. As expected, 30 µM CPA was able to block the ability of stores to refill (data not shown).

View larger version
[in this window]
[in a new window]

Figure 4. Slow rate of internal pool depletion with the SERCA inhibitor cyclopiazonic acid (CPA)
Following a test pulse with 10 µM histamine in 1 mM Ca²⁺_o, CPA (30 µM) was applied for 100 (dashed trace), 200 (grey trace) or 300 s (black trace) in nominally Ca²⁺-free medium to single HUVECs. The state of the internal stores was assessed after these periods with 10 µM histamine. $Delta$ [Ca²⁺]_i was 112 � 25, 88 � 16 and 23 � 6 % of the initial test histamine response, respectively (n = 3; 100 and 200 s: P > 0�6 vs. test response; 300 s: P < 0�005; paired two-tailed t test).

Table 1. Dose-dependent effect of cyclopiazonic acid and thapsigargin on [Ca²⁺]_i in populations of HUVECs

		$Delta$ [Ca²⁺]_i (nM)	Rate of rise (nM s^-1)
Cyclopiazonic acid	0�1 µM	23 � 3	0�09 � 0�01
	1 µM	53 � 9	0�36 � 0�04
	10 µM	300 � 6	9�17 � 1�19
	30 µM	427 � 33	18�48 � 4�30
Thapsigargin	0�002 µM	63 � 20	0�29 � 0�05
	0�02 µM	320 � 25	8�20 � 0�34
	0�2 µM	410 � 35	12�00 � 0�52
	2 µM	417 � 9	14�03 � 1�47

$Delta$ [Ca²⁺]_i refers to the difference between the basal [Ca²⁺]_i and either the peak response or the value 150 s after addition if the peak had not already been reached over this time. The rate of rise has any preceding basal value subtracted. Results are means � S.E.M. of 3 experiments.

Having defined the properties of the inhibitors, we determined the effect of SERCA inhibition on agonist-stimulated Ca²⁺ release. As a first step, we investigated the effect of SERCA inhibition on the histamine dose-response curve, for which we used less labour-intensive cell populations. Because of the large numbers of cells required, we used the HUVEC-derived cell line ECV304, which shares many features with HUVECs (Morgan & Jacob, 1994). Different concentrations of histamine were applied to cells with or without SERCA inhibition in nominally Ca²⁺-free HBSS (supplemented with 0�5 mM EGTA), followed by 100 µM histamine to determine the residual pool size (Fig. 5). As expected, histamine dose dependently emptied the internal stores, with a reciprocal relationship between the initial peak response and the residual pool size (Fig. 5).

View larger version
[in this window]
[in a new window]

Figure 5. Effect of maximal SERCA inhibition upon the histamine dose-response curve
Populations of ECV304 cells were stimulated in Ca²⁺-free medium (0�5 mM EGTA) with different concentrations (x µM) of histamine as indicated next to each trace. Residual pool size was subsequently determined by addition of 100 µM histamine. For the first stimulation, histamine was simultaneously added with either vehicle (0�1 % DMSO) or 30 µM CPA. A, fluorescence (R_340/380) traces illustrating the relationship between first stimulus and pool depletion in control cells. B, same as A in the presence of CPA. For clarity, the second 100 µM histamine responses have been aligned when there were small differences in the first stimulation period (hence trace breaks). The second response to 100 µM histamine was negligible after the initial 100 µM response in A or after the initial responses to 10 and 100 µM histamine in B and were therefore omitted. C, summary of the peak [Ca²⁺]_i response to different histamine concentrations in the presence or absence of CPA as a percentage of the response to 100 µM histamine alone. D, the inverse relationship between the histamine concentration used for the first stimulus and residual pool size (determined with 100 µM histamine). In both C and D, curves were fitted by eye. Data collection rate was 5 Hz. Data are expressed as means � S.E.M. of 3-4 experiments.

Addition of either 30 µM CPA or 200 nM thapsigargin at the same time as histamine potentiated the depletion of internal stores as measured by a subsequent response to 100 µM histamine. Figure 5 shows the data for CPA but similar results were observed when thapsigargin was used (n = 3-4). The potentiation was more than an additive effect of the separate abilities of the inhibitor and histamine to release Ca²⁺ since the inhibitors alone did not significantly reduce the response to 100 µM histamine over this time period: residual pool size (expressed as a percentage of the response to 100 µM histamine alone) was 69 � 10 % (n = 4, P > 0�1) after CPA alone and 81 � 15 % (n = 3, P > 0�4) after thapsigargin alone (similar to the time course of emptying seen in HUVECs; Fig. 4). The potentiation of store emptying by SERCA inhibitors may be related to a larger initial response or to the abolition of resequestration during the slower falling phase of the transient. However, there was a marginally greater effect upon the residual pool size than the amplitude of the initial response (cf. Fig. 5C and D) although this difference was more marked for single cells (see below). It therefore appeared that the sensitivity of stores to agonist (as defined by either the amplitude of the first response or subsequent pool depletion) is partly set by SERCA activity.

Population studies can present a misleading picture; in particular, the potentiation of the peak response observed above could result simply from an increased synchronization of the individual cellular responses (see below) rather than an increase in the response of individual cells. We therefore repeated key experiments in single HUVECs. As with populations of ECV304 cells, 0�3 µM histamine only partially discharged stores since a subsequent application of 10 µM histamine could still release substantial amounts of Ca²⁺ (histamine $Delta$ R340/380, 0�3 µM: 1�05 � 0�12; subsequent 10 µM: 2�43 � 0�38; n = 11; Fig. 6A) whereas in the presence of 30 µM CPA, 0�3 µM histamine stimulated a larger peak response that almost completely emptied stores (0�3 µM: 1�88 � 0�13; subsequent 10 µM: 0�52 � 0�20; n = 11, P < 0�001 for both measurements vs. -CPA; Fig. 6A). In accordance with Fig. 5, CPA alone caused only a small depletion of stores over a similar time course (subsequent 10 µM histamine response: 2�35 � 0�24, n = 7), confirming the synergistic emptying of stores by agonist and CPA. Curiously, the rate constant for decay of the 0�3 µM histamine response was unaffected by CPA (control: 0�089 � 0�022 s^-1; + CPA: 0�088 � 0�012 s^-1, n = 11), although the small oscillations often superimposed on the falling phase were not seen in the presence of CPA. These results, particularly in single cells, are consistent with SERCA influencing graded release since SERCA inhibition greatly potentiated the fraction of the internal stores mobilized by submaximal histamine doses. This appears to have been mainly by abolition of the resequestration of released Ca²⁺ rather than direct modulation of the release process itself since the potentiation of store depletion in single cells was considerably larger than the potentiation of the initial response.

View larger version
[in this window]
[in a new window]

Figure 6. CPA potentiates both IP₃-dependent and IP₃-independent store emptying
Single HUVECs were stimulated in Ca²⁺-free medium with either 0�3 µM histamine (A) or 0�4 µM ionomycin (B). In each panel, traces depict experiments performed in the presence (grey traces) or absence (black traces) of 30 µM CPA. The residual pool size was determined by subsequent addition of 10 µM histamine. Simultaneous application of CPA resulted in a rapid and complete pool depletion. Traces are typical of 7-11 experiments.

SERCA plays a homeostatic role independently of IP₃

The different effects of SERCA inhibition on the peak response and store emptying could arise from a varying imbalance between the homeostatic and modulatory roles of SERCA during different phases of the [Ca²⁺]_i response. In order to gain an insight into the relative contribution of these roles, we assessed the homeostatic role of SERCA in response to Ca²⁺ released by ionophore, independently of IP₃ receptor activation; we have previously demonstrated that ionomycin and histamine discharge Ca²⁺ from the same stores (Morgan & Jacob, 1996).

Figure 6B shows that for HUVECs in Ca²⁺-free medium 0�4 µM ionomycin alone only partially depleted the internal stores since 10 µM histamine could stimulate a further release ( $Delta$ R340/380; ionomycin: 0�81 � 0�11; subsequent 10 µM histamine: 1�37 � 0�28; n = 7) (Morgan & Jacob, 1996). Since ionophore-mediated Ca²⁺ release is unlikely to involve the IP₃ receptor, this partial discharge was probably due to a balance between mobilization and SERCA-mediated reuptake. This was confirmed by repeating the experiment in the presence of 30 µM CPA; the combination of ionomycin and CPA rapidly and completely emptied the internal stores as 10 µM histamine had no further effect ( $Delta$ R340/380; ionomycin: 1�68 � 0�12; 10 µM histamine: 0�04 � 0�01; n = 7, P < 0�001 vs. control). Moreover, both the rate of rise and fall of the response were enhanced, again consistent with the ionomycin-induced [Ca²⁺]_i increase being limited by SERCA activity (percentage of control, rate of rise: 268 � 28 %; rate constant for recovery: 250 � 30 %; n = 7, P < 0�001 vs. control). The effect of SERCA activity on IP₃-independent Ca²⁺ discharge provides more direct evidence that the effect upon agonist responses is not simply at the level of release, i.e.these data (together with the earlier discrepancy between peak and emptying) point to a powerful homeostatic role for SERCA.

SERCA and its influence on Ca²⁺ release, its modulatory role

Having suggested a homeostatic role for SERCA, we turned to its involvement in regulating the release phase, its modulatory role. The earlier results showing the potentiation of the histamine-induced Ca²⁺ peak already pointed to an involvement in the regulation of release, but we wanted to assess in more detail which component(s) of the release process was influenced (i.e.initiation, the regenerative phase or termination).

Our working hypothesis was that during activation of the IP₃ receptor, SERCA activity could be sufficiently powerful to dissipate localized [Ca²⁺]_i increases (Dupont & Swillens, 1996). Consequently, SERCA inhibition might allow these Ca²⁺ gradients to build up and, once a threshold was reached, promote release from neighbouring IP₃ receptors via a form of Ca²⁺-induced Ca²⁺ release (CICR) dependent on the positive feedback of Ca²⁺ on Ca²⁺ release via the IP₃ receptor (Petersen et al. 1994; Berridge, 1997). Summation of localized release would give a net regenerative Ca²⁺ wave (Bootman et al. 1997; Berridge, 1997).

SERCA inhibition promotes an explosive release of Ca²⁺

Co-application of a SERCA inhibitor and histamine had profound consequences on Ca²⁺ handling (Figs 5 and 6) but it was difficult to dissect the underlying mechanisms so we separated the two by adding CPA (or thapsigargin) after the histamine. First, in HUVEC populations, we added a submaximal concentration of histamine (0�3 µM) in Ca²⁺-free medium (plus 0�25 mM EGTA) to partially deplete stores and, when [Ca²⁺]_i had returned to basal values, then added CPA or thapsigargin (in the continued presence of histamine; Fig. 7). In the absence of histamine, SERCA inhibitors caused a rise in [Ca²⁺]_i that was commensurate with the unmasking of a leak pathway (cf. Fig. 4 and Table 1). By contrast, in the presence of histamine both inhibitors initiated a 5- to 8-fold faster rate of rise (Table 2) even though the initial [Ca²⁺]_i was similar when the SERCA inhibitor was added. This potentiation could have been due to CICR and/or the unmasking of the true release rate repressed by SERCA.

View larger version
[in this window]
[in a new window]

Figure 7. SERCA inhibition results in a rapid release of Ca²⁺ when applied in the presence of histamine
HUVEC populations were stimulated in Ca²⁺-free medium (0�25 mM EGTA) with either 30 µM CPA (A) or 200 nM thapsigargin (Tg; B) alone, resulting in a slow elevation of [Ca²⁺]_i (grey traces). This rate of rise was considerably slower than that induced by submaximal 0�3 µM histamine (initial responses; black traces). However, if CPA or thapsigargin was added after the initial histamine transient returned to basal levels, a rapid rise in [Ca²⁺]_i was observed (black traces). Data collection rate was 5 Hz. Traces are representative of 6-7 experiments, as summarized in Table 2.

Table 2. Effect of the presence of histamine on the responses to SERCA inhibitors in HUVEC populations

	Agent alone			Agent + histamine
	$Delta$ [Ca²⁺]_i (µM)	Rate of rise (nM s^-1)	n	$Delta$ [Ca²⁺]_i (µM)	Rate of rise (nM s^-1)	n
30 µM cyclopiazonic acid	0�21 � 0�01	6 � 1	6	0�37 � 0�04	46 � 6	7
0�2 µM thapsigargin	0�27 � 0�01	12 � 1	6	0�46 � 0�03	55 � 4	6
0�3 µM histamine	0�34 � 0�02	73 � 8	20	-	-	-

All experiments were carried out in Ca²⁺-free medium. Either SERCA inhibitors were added alone (Agent alone), or in the presence of 0�3 µM histamine (Agent + histamine). For details, see Fig. 7. The continued presence of histamine potentiated the kinetics of the response to CPA and thapsigargin. Data collection rate was 5 Hz. Data represent means � S.E.M. of n experiments.

We extended these separate addition studies to single HUVECs where recordings were made in one of two ways: either large numbers of cells were measured simultaneously using imaging, or the response kinetics were measured using the superior temporal resolution of the photometric system (2-5 Hz). Figure 8 shows illustrative photometric traces. The degree of SERCA inhibition determined the effect on oscillations: at very low concentrations of CPA, the oscillatory frequency could be enhanced in some cells, even in Ca²⁺-free medium (Fig. 8C), whereas modest to maximal CPA concentrations stimulated a regenerative Ca²⁺ release without the subsequent train of oscillatory spikes (Fig. 8D and E). Note also that the amplitude of the spike increased with CPA concentration (Table 3). Furthermore, to strengthen the idea that CICR is involved in the CPA response, ionomycin was added instead of the SERCA inhibitor. As with CPA, the effect of ionomycin depended on its concentration: at 30 nM (the lower end of the dose- response curve in ECV304 cells; Morgan & Jacob, 1994) oscillations were reversibly stimulated, whereas 300 nM evoked a single regenerative Ca²⁺ spike. Both results implicate Ca²⁺ positive feedback mechanisms in the CPA effect.

View larger version
[in this window]
[in a new window]

Figure 8. Effect of applying different concentrations of ionomycin or CPA during Ca²⁺ oscillations in Ca²⁺-free medium
Single HUVECs were stimulated to oscillate with the indicated concentration of histamine. Either extracellular Ca²⁺ (1 mM) was briefly applied as shown (A and C-E), or was otherwise absent throughout the entire experiment (B). Different concentrations of ionomycin (30 or 300 nM) or CPA (0�3-30 µM) were applied during the ISP and the residual pool size was usually tested at the end of the run with 10 µM histamine. Low doses of ionomycin and CPA reversibly increased the oscillatory frequency. All traces shown are photometric recordings, typical of at least 3 experiments. The break in the trace in E represents removal of artefacts due to bubbles.

As expected, there was an inverse relationship between the amplitude of the response to CPA applied during the ISP and the residual pool size (Table 3; cf. Fig. 5). However, the CPA dose-response relationship was different for each effect. Whereas there was little difference between the responses to 3 and 30 µM CPA, there was a 4-fold difference in the subsequent 10 µM histamine response. That is, there was a discrepancy between the magnitude of the initial release phase and the consequences for pool emptying, qualitatively similar to co-addition experiments (Figs 5 and 6). Furthermore, ionomycin could not fully mimic 30 µM CPA in either respect.

Similar experiments repeated at higher temporal resolution revealed that addition of CPA during the ISP evoked a spike whose rate of rise depended on the CPA concentration (Table 3). [Ca²⁺]_i rose with the same kinetics as the previous oscillatory spike at low CPA concentrations but more quickly at maximal SERCA blockade with 30 µM CPA; however, the increase in kinetics was proportional to the increase in the magnitude of the response so that the time to peak remained constant (see below). In contrast, increasing the ionomycin concentration from 30 to 300 nM had no significant effect upon oscillatory spike kinetics (Table 3).

Table 3. Effect of different concentrations of CPA and ionomycin on single cells when applied during the interspike period of histamine-induced oscillations

		Imaging			Photometric
Agent		$Delta$ Initial spike (%)	Residual pool size (%)	n	$Delta$ Initial spike (%)	Rate of rise (%)	n
Cyclopiazonic acid	0�3 µM	110 � 6	510 � 35	34	110 � 4 *	76 � 14	6
	3 µM	215 � 11 ***	407 � 24	76	155 � 33	120 � 21	5
	30 µM	254 � 15 ***	107 � 13	55	237 � 32 ***	280 � 45 ***	17
Ionomycin	30 nM	n.d.	n.d.	-	108 � 10	121 � 19	3
	300 nM	n.d.	n.d.	-	108 � 10	138 � 19	9

Columns on the left indicate experiments performed using imaging (

0�3 Hz collection rate) concentrating on the magnitude of responses; after cells were stimulated with 0�3 µM histamine for 4 min in Ca²⁺-free medium, CPA was applied for 2 min and then pool size was probed with 10 µM histamine. Columns on the right refer to similar experiments using the photometric system with a higher temporal resolution (2-5 Hz). Data represent means � S.E.M. of n cells where both kinetic and amplitude parameters are expressed as a percentage of the preceding true oscillatory spikes. n.d., not determined. * P < 0�05, *** P < 0�005, vs. oscillatory spike; paired two-tailed t test.

Experiments conducted using the photometric system were also accompanied by controls; by analogy with Fig. 7, we applied ionomycin or CPA to confirm that they could not rapidly release Ca²⁺ on their own. As expected, the rate of rise of the spike evoked by 30 µM CPA applied during the ISP was 6 � 1-fold faster than that observed when added alone (n = 8 (alone), n = 17 (ISP); P < 0�002, unpaired two-tailed t test) and similarly the 300 nM ionomycin response was 8 � 2-fold faster when applied during the ISP compared with when added alone (n = 9 (alone), n = 12 (ISP); P < 0�01). In other words, the presence of histamine greatly potentiated the single cell responses to both CPA and ionomycin suggesting the involvement of regenerative Ca²⁺ release (cf. population studies; Fig. 7).

The photometric system also allowed us to monitor simultaneously the fura-2 signal and the bath exchange (with fluorescein, see Methods). The first discernible R340/380 rise stimulated by CPA during the ISP occurred when the bath exchange was 83 � 4 % complete (n = 13). In other words, the extracellular CPA concentration was unequivocally high before spike initiation.

Role of SERCA in modulating spatial Ca²⁺ signals

The spatial equivalent of a Ca²⁺ spike in a single cell is a Ca²⁺ wave (Petersen et al. 1994; Berridge, 1997). This has been demonstrated for HUVECs (Jacob, 1990) as for other cell types, where Ca²⁺ release starts in a trigger zone and then rapidly propagates across the entire cell (Jacob, 1990; Petersen et al. 1994; Berridge, 1997). The photometric recording of a regenerative spike in a single cell is a complex product of the [Ca²⁺]_i rise kinetics and the velocity of wave propagation (see below). To understand more fully the consequences of SERCA blockade for a spike, both facets must be examined separately. To do this we repeated some of the experiments above using high resolution imaging which allowed detailed subcellular characterization of the response. We compared (a) the initial response to histamine in the presence and absence of CPA (cf. Figs 5 and 6), and (b) a true oscillatory spike with the subsequent spike stimulated by application of CPA during the ISP (cf. Figs 7 and 8).

Just as photometric records revealed that the initial spike amplitude is proportional to histamine concentration (Figs 3 and 8), so the velocity of the corresponding wave increased with stimulus intensity (Fig. 9B, right-hand panel). This was not due to increased Ca²⁺ influx (Girard & Clapham, 1993) since Ca was removed 30 s before each stimulus. Despite the dose-dependent wave velocity, in more than 80 % of HUVECs the Ca²⁺ waves initiated from the same trigger zone after restimulation, irrespective of the histamine concentration.

To examine the effect of CPA we used a paired protocol (Fig. 9A) with the first stimulation being histamine alone and the second including 30 µM CPA; controls comprised two successive stimulations by histamine alone. Interestingly, despite the fact that the global spike amplitude was enhanced by CPA, there was no concomitant increase in wave velocity. Irrespective of the presence of CPA, Ca²⁺ waves evoked by 0�3 µM histamine propagated with a velocity of 30 µm s^-1 (Fig. 9B).

View larger version
[in this window]
[in a new window]

Figure 9. Influence of SERCA activity and agonist concentration on wave velocity and synchronization of the initial Ca²⁺ response
A, scheme representing general experimental protocol. After a 30 s washout of Ca²⁺_o, HUVECs were briefly (10-50 s) exposed to the relevant concentration of histamine and the response measured with video rate image capturing. Histamine was then removed, stores refilled for at least 2 min by brief exposure to 1 mM Ca²⁺_o and the process repeated under a different condition (i.e. 0�3 µM histamine � 30 µM CPA, or a different histamine concentration alone). Up to 3 successive stimuli were applied per run (always briefly and in Ca²⁺-free medium) where the order of addition of the different histamine concentrations was randomized across different runs (however, 0�3 µM histamine � CPA was always second). Wave velocity (B) and synchronization (C) were determined according to the Methods section. Results are means � S.E.M. from 33-76 cells (B) or 3-11 fields of cells (C). * P < 0�03 compared with 1st stimulation (paired t test).

Another important feature was that many cells that did not respond to 0�3 µM histamine alone subsequently gave a [Ca²⁺]_i wave when 0�3 µM histamine was added together with CPA (not shown). In other words, non-responders became responders when SERCA activity was blocked (Petersen et al. 1993). Closer scrutiny of the trigger zones also revealed that in a minority of cells a second trigger zone was unmasked by CPA at the opposite pole to the primary trigger zone. None of these effects were due to CPA inducing localized [Ca²⁺]_i waves per se since CPA by itself induced a slow, homogeneous rise in [Ca²⁺]_i even though 3 µM histamine stimulated waves in the same cell (n = 14, not shown).

We also assessed the degree of synchronization of the responses and the effect on this of agonist concentration and SERCA inhibition. The quantification of synchronicity is outlined in the Methods section. As expected, the cells were less synchronous at lower histamine concentrations (with a long and variable latency; Fig. 9C, right-hand panel), as indicated by the larger S.D. In support of a role for SERCA in modulating cell sensitivity to 0�3 µM histamine, CPA significantly decreased the S.D. of the response times (i.e. enhanced the synchronicity) compared with controls (Fig. 9C, left-hand panel). This increase in synchronicity presumably reflects increased sensitivity to histamine, consistent with the data above.

Having examined the effect of simultaneous addition of CPA and histamine on waves, we returned to the protocol where CPA was added during the ISP, for which we could directly compare a true oscillatory spike and the subsequent CPA-evoked spike (under conditions in which the CPA concentration is unequivocally high). It is worth pointing out that the initial Ca²⁺ spike in some non-excitable cell types such as HUVECs is often greater than subsequent spikes in terms of amplitude, width and kinetics (Carter & Ogden, 1994; Morgan & Jacob, 1996; Bootman & Berridge, 1996). Likewise, the wave associated with the first spike propagated more rapidly across the cell than subsequent waves (Fig. 10; wave velocity: initial spike, 27 � 3 µm s^-1; oscillatory spike, 15 � 2 µm s^-1; n = 53 cells; P < 0�001, paired t test). Nonetheless, CPA addition during the ISP still had no effect on wave velocity (14 � 2 µm s^-1; n = 53; P > 0�4) despite the potentiation of amplitude. In addition to their similar wave velocity, oscillatory and CPA-stimulated spikes also started in the same region in 92 % of cells (Fig. 10). Hence, together with the results of Fig. 9, we conclude that SERCA plays little role in regulating wave propagation. As an aside, there was a tendency for the initial wavefront to be better defined than for subsequent waves in some cells, reminiscent of the responses observed in HeLa cells (Bootman & Berridge, 1996).

View larger version
[in this window]
[in a new window]

Figure 10. A train of Ca²⁺ waves in the presence of submaximal histamine concentrations
Cells were stimulated in Ca²⁺-free medium with 0�3 µM histamine and images collected at full video rate following excitation at 380 nm. A sequence of F/F₀ images with their corresponding time (relative to the start of the video stack for each wave) of the same single cell are shown. Cool colours represent low [Ca²⁺]_i, warm colours a higher [Ca²⁺]_i where the colour scale has been optimized for the magnitude of each individual spike in A-C and therefore cannot be compared directly. All images were smoothed. The extreme right-hand panels show line plots for the respective waves generated using the region bound by the box in the first image in A. A, Ca²⁺ wave associated with the initial histamine response; B, Ca²⁺ wave corresponding to a true oscillatory spike 2 min 40 s later; C, wave stimulated by 30 µM CPA addition during the ISP 6 min after the initial response (cf. Fig. 8).

Having demonstrated that SERCA plays little role in regulating Ca²⁺ wave velocity in HUVECs, we turned our attention to the role of SERCA in modulating the kinetics of the rise in [Ca²⁺]_i. As already stated, a photometric recording of a spike over the entire cell is the net result of local rates of rise and the rate of wave propagation since it may take up to several seconds for the wave to cross the cell (see below). Therefore, to assess rates of rise in isolation, we recorded from as small a subcellular region as possible consistent with a reasonable signal-to-noise ratio. In this way, the wavefront passes almost instantaneously across this small region, so that the rise of the signal reflects only the local rate of rise.

Before dealing with the involvement of SERCA, we determined the relationship between the photometric records (global Ca²⁺ signals) and more localized Ca²⁺ release rates, allowing the kinetic data of Table 3 to be put into context. We examined the clearest records from the video rate F₃₈₀images previously used to determine the wave velocity and for each cell analysed either (a) the whole cell fluorescence ('global' signal) or (b) three subcellular regions of interest (ROIs) located in the trigger zone (ROI 1), central nucleus (ROI 2) and the opposite pole of the cell from the trigger zone (ROI 3). In the case of (b) our ROI size was typically a square of 2-3 µm in length for a cell of 40-60 µm diameter (meaning that a wavefront travelling at 30 µm s^-1 passes across this ROI in 66-100 ms). Figure 11A shows the global and local changes recorded simultaneously during a typical initial response to 0�3 µM histamine (note the F₃₈₀ data have been normalized and inverted, so that the absolute fluorescence changes have not been preserved). Taking the cell as a whole, the [Ca²⁺]_i rose relatively slowly, taking 1-2 s to peak. However, on a local level, the mean time to peak was only 51 % of the global value (range, 36-59 %). In other words, a photometric record depends not only on the absolute [Ca²⁺]_i change but is also influenced by the time taken for the wave to propagate across the cell.

We repeated this type of analysis for bona fide oscillations and for the spike induced by CPA application during the ISP. In all cases, the global time to peak was much longer than the local time to peak (P < 0�002, paired t test; summarized in Fig. 11B) underscoring the contribution of wave propagation to all photometric spike recordings. Certainly within ROIs 1 and 2, the time to peak was shorter for the first spike than for the subsequent oscillatory train. More importantly, the time for an oscillatory spike to peak was unaffected by SERCA blockade with CPA, either on a global or a local level, consistent with data shown in Table 3.

View larger version
[in this window]
[in a new window]

Figure 11. Comparison of global and local time to peak and their susceptibility to SERCA inhibition
A shows a typical comparison of the global and local time to peak during the first spike stimulated by 0�3 µM histamine in HUVECs. The traces are F₃₈₀ records which have been inverted and normalized to the start and end points for the regions depicted in the inset, where Global refers to the whole cell fluorescence, 1 is the trigger zone (ROI 1), 2 is the nucleus (ROI 2) and 3 is the end point (ROI 3). B shows a summary of the local and global time to peak observed during three types of spike seen in the presence of 0�3 µM histamine: the initial spike, an oscillatory spike and the spike stimulated by application of 30 µM CPA during the ISP. All experiments were performed in Ca²⁺-free medium. Results are means � S.E.M. of 10-21 cells.

DISCUSSION

Top
Abstract
Introduction
Methods
Results
Discussion
References

Powerful homeostatic role of SERCA during the spike falling phase

Our data show that SERCA plays a powerful homeostatic role in HUVECs as suggested by others (Gericke et al. 1993). Firstly, internal stores remained full after continuous or repetitive application of 0�3 µM histamine in Ca²⁺-free solution indicating little net loss of Ca²⁺ via PMCA. Furthermore, inhibition of the PMCA with La³⁺ did not prevent spiking, underscoring its minor role in the restoration of basal [Ca²⁺]_i following a spike.

In addition, SERCA inhibitors facilitated store emptying by histamine in a synergistic manner (Figs 5, 6 and 8); the augmentation of store emptying was disproportionately larger than the modest augmentation of the peak response suggesting that the more important determinant of how much net Ca²⁺ is discharged during a transient is what happens during the slower, falling phase of a spike rather than the initial release phase. A similar synergism was seen between CPA and ionomycin (Fig. 6B) confirming a homeostatic role for SERCA independent of any potential modulatory effects on IP₃-mediated release. The slow kinetics of the ionomycin response were greatly enhanced by SERCA inhibition both on the upstroke and the downstroke (Fig. 6). Inhibition of resequestration can explain the potentiation of the upstroke, but not the downstroke, so the enhanced rate of fall must be due to other factors. For instance, the higher peak [Ca²⁺]_i in the presence of CPA might activate the PMCA (Camello et al. 1996) or SERCA inhibition might permit a higher local [Ca²⁺] at the PMCA if the two Ca²⁺-ATPases competed for the released Ca²⁺. These factors could also explain the surprising similarity of the falling phase of the response to 0�3 µM histamine in the presence and absence of CPA (Fig. 6).

Thus the falling phase of a Ca²⁺ spike is governed primarily by SERCA as in other cell types (Tse et al. 1994; Li et al. 1994). However, PMCA activity cannot be totally ignored since oscillations do run down in frequency in Ca²⁺-free medium and oscillations are observed in a smaller proportion of cells when PMCA is blocked by La³⁺.

Inverse relationship between homeostatic SERCA activity and agonist concentration

The relative contribution of the two Ca²⁺-ATPases to the falling phase of a spike depends on the stimulus strength in HUVECs as for pancreatic acinar cells (Zhang & Muallem, 1992). The contribution of SERCA relative to PMCA decreased at higher histamine concentrations; blocking PMCA with La³⁺ profoundly altered Ca²⁺ handling during maximal histamine stimulation (Fig. 1), whereas oscillatory responses to 0�3 µM histamine were affected to a lesser extent (Fig. 2). This correlates with a sequence of brief pulses of histamine in Ca²⁺-free solution having little effect on store size at 0�3 µM but causing a gradual depletion at 1 µM (Fig. 3), while a single pulse of 10 µM histamine substantially depleted stores (Fig. 1).

The reason for this inverse relationship between SERCA activity and histamine concentration is not clear. However, net reuptake into stores via SERCA is likely to be increasingly compromised as IP₃ receptor channel fluxes increase (Zhang & Muallem, 1992). Another possibility is that as the histamine concentration increases, stimulation of PMCA (perhaps by protein kinase C or Ca²⁺-calmodulin; Petersen et al. 1994; Camello et al. 1996) swamps stimulation of SERCA (perhaps by [Ca²⁺]_i and/or pool depletion; Petersen et al. 1994; Favre et al. 1996; Mogami et al. 1998).

Modulatory role of SERCA in regulating Ca²⁺ release

To facilitate discussion, we consider Ca²⁺ release as occurring in several distinct but contiguous phases, i.e. initiation, the regenerative phase and termination. Initiation is the earliest phase of Ca²⁺ mobilization where [Ca²⁺]_i rises to a critical level in a localized trigger region (Petersen et al. 1994; Berridge, 1997). If the local [Ca²⁺]_i (i.e. elementary event activity) exceeds a certain threshold (Bootman et al. 1997; Berridge, 1997), positive feedback occurs via CICR leading to the explosive regenerative phase of Ca²⁺ release and propagation of a wave across the rest of the cell. Priming of secondary stores may also contribute to a regenerative response in some systems (Petersen et al. 1994). Finally, termination of release is probably a consequence of the concerted action of several processes which inhibit release, e.g.inhibition of the IP₃ receptor by high cytosolic [Ca²⁺] (Zhang & Muallem, 1992; Petersen et al. 1994; Berridge, 1997), the fall in luminal Ca²⁺ content (Berridge, 1997) and other, Ca²⁺-independent, pathways of inhibition (Ilyin & Parker, 1994; Hajnóczky & Thomas, 1997).

Interaction of the IP₃ receptor with both luminal and cytosolic [Ca²⁺] at least in part provides the positive and negative feedbacks that are required for the generation of repetitive spiking (Petersen et al. 1994; Berridge, 1997). Given the influence of SERCA on both luminal and cytosolic [Ca²⁺], this means that SERCA could have a modulatory role on Ca²⁺ release itself. Indeed, Dupont & Swillens recently invoked the existence of discrete Ca²⁺ domains immediately around the cytosolic face of the IP₃ receptor that are regulated by neighbouring SERCA molecules and which in their model are central to determining the oscillatory period (Dupont & Swillens, 1996).

SERCA suppresses the initiation phase of release

The ability of low ionomycin concentrations (Fig. 8) to trigger regenerative spikes in the presence of histamine suggests the involvement of some form of CICR in initiating a Ca²⁺ spike. This is probably via the IP₃ receptor since we (Morgan & Jacob, 1996) and others (Bennett et al. 1996) find no evidence for ryanodine receptors in HUVECs and oscillations can be generated by IP₃ analogues independently of receptor-phospholipase C coupling (Morgan & Jacob, 1996).

Like ionomycin, SERCA inhibition initiated a rapid rise in [Ca²⁺]_i when applied during histamine stimulation (Figs 7 and 8). Indeed, even 0�3 µM CPA which itself released little Ca²⁺ (Table 1) potentiated repetitive spiking (i.e.it shortened the ISP) even in the absence of Ca (Fig. 8) (cf. Petersen et al. 1993). This action of CPA is probably due to the increased net release of Ca²⁺ that eventually results in CICR since it is mimicked by 30 nM ionomycin (Fig. 8). These results suggest that the initiation of an explosive release of Ca²⁺ via CICR is suppressed by SERCA during the ISP.

In addition to regulating the initiation phase of a true oscillatory spike, SERCA also plays a role in modulating initiation of the initial response to histamine (i.e. the latency, or lag phase of the first response). The latency of the response in HUVECs, like other cell types, depends on agonist concentration (Jacob et al. 1988) and may be governed by the time required to accumulate IP₃ (Chiavaroli et al. 1994; Wang et al. 1995) and/or reach a localized threshold [Ca²⁺]_i for propagation (Berridge, 1997). Cell-to-cell variation in latency (sensitivity) means that not all cells initiate at the same time, resulting in an asynchronous response. Simultaneous addition of CPA promoted a more synchronous histamine response (Fig. 9) reflecting a shortening of the latency (see Methods) and even recruited silent cells to give full-blown waves (cf. Petersen et al. 1993). Both results suggest that, during the lag phase, SERCA represses Ca²⁺ rises before the excitable threshold is reached, i.e. the initiation phase of the initial response is suppressed.

In spatial terms, the first sign of an increase in [Ca²⁺]_i was often the onset of rather diffuse fluctuations (Fig. 10) similar to those observed in HeLa cells (Bootman & Berridge, 1996); these were followed by co-ordinated polarized increases in [Ca²⁺]_i which in turn led to a wave (Fig. 10). This pattern was largely unaffected by CPA (Fig. 10) which suggests that SERCA activity is not involved in determining the initiation locus in HUVECs, unlike other cells (Lee et al. 1997).

A complex relationship between SERCA activity and the regenerative phase

The regenerative phase observed photometrically is a product of two phenomena: the local rates of rise (Bootman et al. 1997) and the propagation rate of the wave across the cell. Just as CICR is involved in initiation, the spike upstroke (the regenerative phase) also involves CICR (Petersen et al. 1994; Berridge, 1997). The local rate of rise and wave velocity are distinct but potentially interdependent parameters characteristic of the messenger that propagates across the cell (Allbritton & Meyer, 1993).

Although SERCA inhibition profoundly affected the initiation of a spike, it had no effect upon agonist-stimulated wave propagation. Neither the initial wave nor the subsequent slower oscillatory waves were affected by CPA, even though faster waves could be detected with higher histamine concentrations (Fig. 9). A lack of effect of SERCA on Ca²⁺ waves was also seen in Xenopus oocytes (Camacho & Lechleiter, 1993) although SERCA inhibition profoundly affected Ca²⁺ waves in epithelial cells (Lee et al. 1997). The reason for this variability is unknown. In addition to a possible cell-specific subcellular distribution of SERCA, susceptibility to SERCA inhibition may also hinge on whether [Ca²⁺]_i waves are generated by diffusion and subsequent amplification of Ca²⁺ or by diffusion of IP₃ (Allbritton & Meyer, 1993).

SERCA activity also affected the rate of rise of the spike. The local time to peak (i.e.the duration of Ca²⁺ release) was unaffected by SERCA inhibition (Fig. 11) and, since SERCA did not affect wave velocity, this resulted in the global time to peak also being unaffected (Fig. 11). However, because SERCA inhibition potentiated the global peak [Ca²⁺]_i it thereby potentiated the global rate of rise of [Ca²⁺]_i; since the [Ca²⁺]_i at the peak was relatively uniform across the cell the local peaks and rates of rise of [Ca²⁺]_i must also have been potentiated. Note that direct reliable calculations of these local parameters could not be made from the single wavelength imaging data. In summary, SERCA activity modulates Ca²⁺ release during the regenerative phase in the temporal, but not the spatial, domain, at least within the detection resolution of our measurement system.

We cannot say how SERCA modulates Ca²⁺ release rates, but one possibility is that SERCA may pull Ca²⁺ away from the IP₃ receptor and reduce positive feedback. SERCA inhibition would then potentiate release but might also be expected to moderate the inhibition of the IP₃ receptor by higher levels of cytosolic Ca²⁺ (Dupont & Swillens, 1996). However, the kinetics of inhibition of the IP₃ receptor by Ca²⁺ are probably slower than for activation (Li et al. 1994) so that the stimulatory effects of SERCA inhibition should dominate. Kinetic modelling of IP₃ receptor fluxes has suggested that the spike amplitude and duration of channel opening (time to peak) are inversely related (Bezprozvanny, 1994). Although this is borne out for graded responses to histamine (Fig. 11), inhibition of SERCA apparently disrupts this relationship suggesting the involvement of other factors such as Ca²⁺-independent inhibition of release (Ilyin & Parker, 1994; Hajnóczky & Thomas, 1997). In addition, SERCA inhibition may allow [Ca²⁺]_i to rise faster and higher as a result of inhibition of its homeostatic role.

SERCA modulates the termination phase of release

Oscillatory spikes are of smaller amplitude than the maximal Ca²⁺ response so that termination of spiking occurs before stores are fully discharged (graded release). Various mechanisms for limited release have been proposed such as all-or-none quantal release, inhibition by high [Ca²⁺]_i, a fall in luminal Ca²⁺ levels or inhibition by IP₃ (Morgan & Jacob, 1996; Berridge, 1997; Hajnóczky & Thomas, 1997). In HUVECs, SERCA activity is another factor affecting spike amplitude, probably by interacting with some of these elements. SERCA inhibition potentiated the peak response to histamine (Figs 5, 6 and 8) both locally and globally and irrespective of whether it was added with the histamine or afterwards during spiking. A key question is whether this is a phenomenon in its own right or a consequence of SERCA action during the preceding initiation and regenerative phases. Certainly, initiation and termination can be divorced since both 0�3 µM CPA and 30 nM ionomycin enhanced oscillatory frequency without significantly increasing the peak height or release kinetics. On the other hand, it is not clear whether [Ca²⁺]_i goes higher because the rate of rise is faster or vice versa. Add to this the complexity of summating elementary release events, and it should be clear that graded responses could also be modified if SERCA were to regulate the recruitment of such events.

SERCA differentially modulates the phases of a spike

Different phases of a spike are influenced by SERCA to different degrees as indicated by their sensitivities to CPA. Initiation was the most sensitive, since even the lowest concentration of CPA used (0�3 µM) promoted a regenerative release of Ca²⁺ during the ISP (Fig. 8). At the other extreme, the wave propagation rate was completely resistant to the effects of even a maximal CPA concentration (Fig. 9), whereas the temporal characteristic of the regenerative phase (and termination) required 3-30 µM CPA for a consistent effect (Fig. 8 and Table 3). On the other hand, the sensitivity of the homeostatic role of SERCA during the falling phase (reflected by the ability to potentiate depletion of the histamine-sensitive pool) was intermediate since even 3 µM CPA clearly potentiated emptying (Table 3). We conclude that SERCA activity is important for different phases of the spike to different degrees.

The role of SERCA in HUVEC Ca²⁺ signalling

Before summarizing the role of SERCA in HUVECs we note that its role is likely to vary according to cell type since its inhibition or overexpression produces variable results. We have already mentioned that in some cell types its homeostatic role is less powerful resulting in rapid run-down of spiking in Ca²⁺-free medium (Tepikin et al. 1992). This variability extends to its modulatory role. For example, in contrast to our observations in HUVECs, SERCA inhibition decreased and SERCA overexpression increased oscillatory frequency in oocytes (Camacho & Lechleiter, 1993). Similarly, inhibition decreased oscillatory frequency in fibroblasts (Rossi & Kao, 1997) and immediately stopped [Ca²⁺]_i spiking in chromaffin cells (without giving rise to a regenerative release of Ca²⁺) and failed to recruit silent cells to oscillate (D'Andrea et al. 1993). Submaximal SERCA inhibition in acinar cells diminished the peak Ca²⁺ oscillatory spike (Petersen et al. 1993). Maximal SERCA inhibition did not affect either the amplitude of myocyte Ca²⁺ sparks (Gómez et al. 1996) or the peak response of N1E-115 neuroblastoma cells to carbachol stimulation (Wang et al. 1995).

In HUVECs, SERCA plays a major homeostatic role, removing Ca²⁺ from the cytosol during the slow downstroke of a spike although this role is compromised as stimulus intensity increases. SERCA may also clear Ca²⁺ from the inhibitory sites of the IP₃ receptor allowing its recovery, although this is controversial (Hajnóczky & Thomas, 1997). However, our data show that SERCA also modulates the more rapid release phase since SERCA is involved in setting the sensitivity of the cell to agonist. It is likely that during the oscillatory interspike period when the IP₃ receptor fully recovers from its transient inactivation (Petersen et al. 1994), the low affinity conformation of the receptor prevails at basal [Ca²⁺]_i and the net efflux from stores will be small. If local [Ca²⁺]_i increases, it converts the IP₃ receptor to a high-affinity conformation, culminating in the regenerative release that is a spike. Our data suggest that powerful SERCA activity suppresses such initially small local Ca²⁺ rises during the ISP, thereby preventing the IP₃ receptor from 'prematurely' switching conformation which is eventually manifest as CICR.

This may partly explain why there is an extended delay between spikes in some cell types despite the fact that the IP₃ receptor may have fully recovered during the ISP (Petersen et al. 1994; Dupont & Swillens, 1996). Only when cell buffering and SERCA become saturated (or inhibited) would the local [Ca²⁺]_i rise sufficiently to induce CICR (Petersen et al. 1993; Dupont & Swillens, 1996). Indeed, a suppressive factor such as SERCA is crucial to systems featuring positive feedback resulting in a more discrete threshold of activation (Allbritton & Meyer, 1993). Thus SERCA may act as a filter that safeguards against firing in response to noise from low level fluctuations in hormone concentration; a regenerative release only occurs when the stimulus intensity is sufficient for IP₃-induced Ca²⁺ release to swamp SERCA and initiate a spike.

During the regenerative phase, SERCA has little effect in the spatial domain (the Ca²⁺ wave) whilst the local rate of release and the peak amplitude are somewhat suppressed by SERCA. Similarly (or perhaps because of these altered release kinetics), the amplitude of a Ca²⁺ spike is suppressed by SERCA. These effects upon regeneration and termination may well be via the ability of SERCA to pull Ca²⁺ away from the IP₃ receptor channel, tempering the positive Ca²⁺ feedback.

In the longer term, SERCA replenishes internal stores with Ca²⁺ and sustains their sensitivity to IP₃ (Morgan & Jacob, 1996). Together with proposals that SERCA can itself be regulated (Petersen et al. 1994; Favre et al. 1996; Mogami et al. 1998), it is clear that this enzyme is poised in a perfect position to influence oscillatory dynamics. One therefore needs to account for SERCA (and probably the co-operativity inherent to SERCA and the IP₃ receptor activity; Dupont & Swillens, 1996) in order to gain a full appreciation of the complex oscillatory cycle.

REFERENCES

Top
Abstract
Introduction
Methods
Results
Discussion
References

Allbritton, N. L. & Meyer, T. (1993). Localized calcium spikes and propagating calcium waves. Cell Calcium 14, 691-697 [Medline]
Bennett, D. L., Cheek, T. R., Berridge, M. J., De Smedt, H., Parys, J. B., Missiaen, L. & Bootman, M. D. (1996). Expression and function of ryanodine receptors in nonexcitable cells. Journal of Biological Chemistry 271, 6356-6362 [Abstract/Full Text]
Berridge, M. J. (1995). Capacitative calcium entry. Biochemical Journal 312, 1-11 [Medline]
Berridge, M. J. (1997). Elementary and global aspects of calcium signalling. The Journal of Physiology 499, 291-306 [Medline]
Bezprozvanny, I. (1994). Theoretical analysis of calcium wave propagation based on inositol (1,4,5)-trisphosphate (InsP₃) receptor functional properties. Cell Calcium 16, 151-166 [Medline]
Bootman, M. D. & Berridge, M. J. (1996). Subcellular Ca²⁺ signals underlying waves and graded responses in HeLa cells. Current Biology 6, 855-865 [Medline]
Bootman, M. D., Berridge, M. J. & Lipp, P. (1997). Cooking with calcium: the recipes for composing global signals from elementary events. Cell 91, 367-373 [Medline]
Bootman, M. D., Cheek, T. R., Moreton, R. B., Bennett, D. L. & Berridge, M. J. (1994). Smoothly graded Ca²⁺ release from inositol 1,4,5-trisphosphate-sensitive Ca²⁺ stores. Journal of Biological Chemistry 269, 24783-24791 [Abstract]
Camacho, P. & Lechleiter, J. (1993). Increased frequency of Ca²⁺ waves in Xenopus laevis oocytes that express a Ca²⁺-ATPase. Science 260, 226-229 [Medline]
Camello, P., Gardner, J., Petersen, O. H. & Tepikin, A. V. (1996). Calcium dependence of calcium extrusion and calcium uptake in mouse pancreatic acinar cells. The Journal of Physiology 490, 585-593 [Abstract]
Carter, T. D. & Ogden, D. (1994). Acetylcholine-stimulated changes of membrane potential and intracellular Ca²⁺ concentration recorded in endothelial cells in situ in the isolated rat aorta. Pflügers Archiv 428, 476-484 [Medline]
Cheek, T. R., Berridge, M. J., Moreton, R. B., Stauderman, K. A., Murawsky, M. M. & Bootman, M. D. (1994). Quantal Ca²⁺ mobilization by ryanodine receptors is due to all-or-none release from functionally discrete intracellular stores. Biochemical Journal 301, 879-883 [Medline]
Chiavaroli, C., Bird, G. St J. & Putney, J. W. Jr (1994). Delayed 'all-or-none' activation of inositol 1,4,5-trisphosphate-dependent calcium signalling in single rat hepatocytes. Journal of Biological Chemistry 269, 25570-25575 [Abstract]
D'Andrea, P., Zacchetti, D., Meldolesi, J. & Grohovaz, F. (1993). Mechanism of [Ca²⁺]_i oscillations in rat chromaffin cells. Journal of Biological Chemistry 268, 15213-15220 [Abstract]
Demauraux, N., Lew, D. P. & Krause, K.-H. (1992). Cyclopiazonic acid depletes intracellular Ca²⁺ stores and activates an influx pathway for divalent cations in HL-60 cells. Journal of Biological Chemistry 267, 2318-2324. [Abstract]
Du, G.-G., Ashley, C. C. & Lea, T. J. (1996). Ca²⁺ effluxes from the sarcoplasmic reticulum vesicles of frog muscle: effects of cyclopiazonic acid and thapsigargin. Cell Calcium 20, 355-359 [Medline]
Dupont, G. & Swillens, S. (1996). Quantal release, incremental detection, long-period Ca²⁺ oscillations in a model based on regulatory Ca²⁺-binding sites along the permeation pathway. Biophysical Journal 71, 1714-1722 [Abstract]
Favre, C., Schrenzel, J., Jaquet, J., Lew, D. P. & Krause, K.-H. (1996). Highly supralinear feedback inhibition of Ca²⁺ uptake by the Ca²⁺ load of intracellular stores. Journal of Biological Chemistry 271, 14925-14930 [Abstract/Full Text]
Gericke, M., Droogmans, G. & Nilius, B. (1993). Thapsigargin discharges intracellular calcium stores and induces transmembrane currents in human endothelial cells. Pflügers Archiv 422, 552-557 [Medline]
Girard, S. & Clapham, D. (1993). Acceleration of intracellular calcium waves in Xenopus oocytes by calcium influx. Science 260, 229-232 [Medline]
Gómez, A. M., Cheng, H., Lederer, W. J. & Bers, D. M. (1996). Ca²⁺ diffusion and sarcoplasmic reticulum transport both contribute to [Ca²⁺]_i decline during Ca²⁺ sparks in rat ventricular myocytes. The Journal of Physiology 496, 575-581 [Abstract]
Grover, A. K. & Khan, I. (1992). Calcium pump isoforms: Diversity, selectivity and plasticity. Cell Calcium 13, 9-17 [Medline]
Hajnóczky, G. & Thomas, A. P. (1997). Minimal requirements for calcium oscillations driven by the IP₃ receptor. EMBO Journal 16, 3533-3543 [Abstract/Full Text]
Haugland, R. P. (1992). Handbook of Fluorescent Probes and Research Chemicals, 5th edn. Molecular Probes, Inc.
Ilyin, V. & Parker, I. (1994). Role of cytosolic Ca²⁺ in inhibition of InsP₃-evoked Ca²⁺ release in Xenopus oocytes. The Journal of Physiology 477, 503-509 [Abstract]
Jackson, T. R., Patterson, S. I., Thastrup, O. & Hanley, M. R. (1988). A novel tumour promoter, thapsigargin, transiently increases cytoplasmic free Ca²⁺ without generation of inositol phosphates in NG115-401L neuronal cells. Biochemical Journal 253, 81-86 [Medline]
Jacob, R. (1990). Imaging cytoplasmic free calcium in histamine stimulated endothelial cells and in fMet-Leu-Phe stimulated neutrophils. Cell Calcium 11, 241-249 [Medline]
Jacob, R., Merritt, J. E., Hallam, T. J. & Rink, T. J. (1988). Repetitive spikes in cytoplasmic calcium evoked by histamine in human endothelial cells. Nature 335, 40-45 [Medline]
Kwan, C. Y., Takemura, H., Obie, J. F., Thastrup, O. & Putney, J. W. Jr (1991). Effects of MeCh, thapsigargin, and La³⁺ on plasmalemmal and intracellular Ca²⁺ transport in lacrimal acinar cells. American Journal of Physiology 258, C10

Allbritton, N. L. & Meyer, T. (1993). Localized calcium spikes and propagating calcium waves. Cell Calcium 14, 691-697	[Medline]
Bennett, D. L., Cheek, T. R., Berridge, M. J., De Smedt, H., Parys, J. B., Missiaen, L. & Bootman, M. D. (1996). Expression and function of ryanodine receptors in nonexcitable cells. Journal of Biological Chemistry 271, 6356-6362	[Abstract/Full Text]
Berridge, M. J. (1995). Capacitative calcium entry. Biochemical Journal 312, 1-11	[Medline]
Berridge, M. J. (1997). Elementary and global aspects of calcium signalling. The Journal of Physiology 499, 291-306	[Medline]
Bezprozvanny, I. (1994). Theoretical analysis of calcium wave propagation based on inositol (1,4,5)-trisphosphate (InsP₃) receptor functional properties. Cell Calcium 16, 151-166	[Medline]
Bootman, M. D. & Berridge, M. J. (1996). Subcellular Ca²⁺ signals underlying waves and graded responses in HeLa cells. Current Biology 6, 855-865	[Medline]
Bootman, M. D., Berridge, M. J. & Lipp, P. (1997). Cooking with calcium: the recipes for composing global signals from elementary events. Cell 91, 367-373	[Medline]
Bootman, M. D., Cheek, T. R., Moreton, R. B., Bennett, D. L. & Berridge, M. J. (1994). Smoothly graded Ca²⁺ release from inositol 1,4,5-trisphosphate-sensitive Ca²⁺ stores. Journal of Biological Chemistry 269, 24783-24791	[Abstract]
Camacho, P. & Lechleiter, J. (1993). Increased frequency of Ca²⁺ waves in Xenopus laevis oocytes that express a Ca²⁺-ATPase. Science 260, 226-229	[Medline]
Camello, P., Gardner, J., Petersen, O. H. & Tepikin, A. V. (1996). Calcium dependence of calcium extrusion and calcium uptake in mouse pancreatic acinar cells. The Journal of Physiology 490, 585-593	[Abstract]
Carter, T. D. & Ogden, D. (1994). Acetylcholine-stimulated changes of membrane potential and intracellular Ca²⁺ concentration recorded in endothelial cells in situ in the isolated rat aorta. Pflügers Archiv 428, 476-484	[Medline]
Cheek, T. R., Berridge, M. J., Moreton, R. B., Stauderman, K. A., Murawsky, M. M. & Bootman, M. D. (1994). Quantal Ca²⁺ mobilization by ryanodine receptors is due to all-or-none release from functionally discrete intracellular stores. Biochemical Journal 301, 879-883	[Medline]
Chiavaroli, C., Bird, G. St J. & Putney, J. W. Jr (1994). Delayed 'all-or-none' activation of inositol 1,4,5-trisphosphate-dependent calcium signalling in single rat hepatocytes. Journal of Biological Chemistry 269, 25570-25575	[Abstract]
D'Andrea, P., Zacchetti, D., Meldolesi, J. & Grohovaz, F. (1993). Mechanism of [Ca²⁺]_i oscillations in rat chromaffin cells. Journal of Biological Chemistry 268, 15213-15220	[Abstract]
Demauraux, N., Lew, D. P. & Krause, K.-H. (1992). Cyclopiazonic acid depletes intracellular Ca²⁺ stores and activates an influx pathway for divalent cations in HL-60 cells. Journal of Biological Chemistry 267, 2318-2324.	[Abstract]
Du, G.-G., Ashley, C. C. & Lea, T. J. (1996). Ca²⁺ effluxes from the sarcoplasmic reticulum vesicles of frog muscle: effects of cyclopiazonic acid and thapsigargin. Cell Calcium 20, 355-359	[Medline]
Dupont, G. & Swillens, S. (1996). Quantal release, incremental detection, long-period Ca²⁺ oscillations in a model based on regulatory Ca²⁺-binding sites along the permeation pathway. Biophysical Journal 71, 1714-1722	[Abstract]
Favre, C., Schrenzel, J., Jaquet, J., Lew, D. P. & Krause, K.-H. (1996). Highly supralinear feedback inhibition of Ca²⁺ uptake by the Ca²⁺ load of intracellular stores. Journal of Biological Chemistry 271, 14925-14930	[Abstract/Full Text]
Gericke, M., Droogmans, G. & Nilius, B. (1993). Thapsigargin discharges intracellular calcium stores and induces transmembrane currents in human endothelial cells. Pflügers Archiv 422, 552-557	[Medline]
Girard, S. & Clapham, D. (1993). Acceleration of intracellular calcium waves in Xenopus oocytes by calcium influx. Science 260, 229-232	[Medline]
Gómez, A. M., Cheng, H., Lederer, W. J. & Bers, D. M. (1996). Ca²⁺ diffusion and sarcoplasmic reticulum transport both contribute to [Ca²⁺]_i decline during Ca²⁺ sparks in rat ventricular myocytes. The Journal of Physiology 496, 575-581	[Abstract]
Grover, A. K. & Khan, I. (1992). Calcium pump isoforms: Diversity, selectivity and plasticity. Cell Calcium 13, 9-17	[Medline]
Hajnóczky, G. & Thomas, A. P. (1997). Minimal requirements for calcium oscillations driven by the IP₃ receptor. EMBO Journal 16, 3533-3543	[Abstract/Full Text]
Haugland, R. P. (1992). Handbook of Fluorescent Probes and Research Chemicals, 5th edn. Molecular Probes, Inc.
Ilyin, V. & Parker, I. (1994). Role of cytosolic Ca²⁺ in inhibition of InsP₃-evoked Ca²⁺ release in Xenopus oocytes. The Journal of Physiology 477, 503-509	[Abstract]
Jackson, T. R., Patterson, S. I., Thastrup, O. & Hanley, M. R. (1988). A novel tumour promoter, thapsigargin, transiently increases cytoplasmic free Ca²⁺ without generation of inositol phosphates in NG115-401L neuronal cells. Biochemical Journal 253, 81-86	[Medline]
Jacob, R. (1990). Imaging cytoplasmic free calcium in histamine stimulated endothelial cells and in fMet-Leu-Phe stimulated neutrophils. Cell Calcium 11, 241-249	[Medline]
Jacob, R., Merritt, J. E., Hallam, T. J. & Rink, T. J. (1988). Repetitive spikes in cytoplasmic calcium evoked by histamine in human endothelial cells. Nature 335, 40-45	[Medline]
Kwan, C. Y., Takemura, H., Obie, J. F., Thastrup, O. & Putney, J. W. Jr (1991). Effects of MeCh, thapsigargin, and La³⁺ on plasmalemmal and intracellular Ca²⁺ transport in lacrimal acinar cells. American Journal of Physiology 258, C10