HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Bautista, D. M. Articles by Lewis, R. S. Search for Related Content PubMed PubMed Citation Articles by Bautista, D. M. Articles by Lewis, R. S.

Journal of Physiology (2002), 541.3, pp. 877-894
© Copyright 2002 The Physiological Society
DOI: 10.1113/jphysiol.2001.016154

Enhancement of calcium signalling dynamics and stability by delayed modulation of the plasma-membrane calcium-ATPase in human T cells

Diana M. Bautista, Markus Hoth and Richard S. Lewis

	ABSTRACT

Top Abstract Introduction Methods Results Discussion References

In addition to its homeostatic role of maintaining low resting levels of intracellular calcium ([Ca²⁺]_i), the plasma-membrane calcium-ATPase (PMCA) may actively contribute to the generation of complex Ca²⁺ signals. We have investigated the role of the PMCA in shaping Ca²⁺ signals in Jurkat human leukaemic T cells using single-cell voltage-clamp and calcium-imaging techniques. Crosslinking the T-cell receptor with the monoclonal antibody OKT3 induces a biphasic elevation in [Ca²⁺]_i consisting of a rapid overshoot to a level > 1 µM, followed by a slow decay to a plateau of ~0.5 µM. A similar overshoot was triggered by a constant level of Ca²⁺ influx through calcium-release-activated Ca²⁺ (CRAC) channels in thapsigargin-treated cells, due to a delayed increase in the rate of Ca²⁺ clearance by the PMCA. Following a rise in [Ca²⁺]_i, PMCA activity increased in two phases: a rapid increase followed by a further calcium-dependent increase of up to approximately fivefold over 10-60 s, termed modulation. After the return of [Ca²⁺]_i to baseline levels, the PMCA recovered slowly from modulation ( ~4 min), effectively retaining a 'memory' of the previous [Ca²⁺]_i elevation. Using a Michaelis-Menten model with appropriate corrections for cytoplasmic Ca²⁺ buffering, we found that modulation extended the dynamic range of PMCA activity by increasing both the maximal pump rate and Ca²⁺ sensitivity (reduction of K_M). A simple flux model shows how pump modulation and its reversal produce the initial overshoot of the biphasic [Ca²⁺]_i response. The modulation of PMCA activity enhanced the stability of Ca²⁺ signalling by adjusting the efflux rate to match influx through CRAC channels, even at high [Ca²⁺]_i levels that saturate the transport sites and would otherwise render the cell defenceless against additional Ca²⁺ influx. At the same time, the delay in modulation enables small Ca²⁺ fluxes to transiently elevate [Ca²⁺]_i, thus enhancing Ca²⁺ signalling dynamics.

(Received 27 December 2001; accepted after revision 27 March 2002)
Corresponding author R. Lewis: Beckman Center B-111A, Stanford University School of Medicine, Stanford, CA 94305, USA. Email: rslewis{at}stanford.edu

	INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

Intracellular Ca²⁺ signalling adopts a wide variety of dynamic patterns, ranging from spikes to oscillations to sustained plateaus (Berridge & Dupont, 1994). Recent studies have shown that cells are capable of extracting specific information from both the frequency and amplitude of Ca²⁺ signals, and therefore that the complexity of the signals may serve to increase the amount of information that can be encoded via calcium-dependent pathways. Complex Ca²⁺ patterns regulate the motility of growth cones in developing neurones and the twitch properties in muscle (Spitzer et al. 2000; Torgan & Daniels, 2001). In lymphocytes, the spike, plateau and oscillating Ca²⁺ signals activate distinct transcriptional pathways involved in cell activation (Dolmetsch et al. 1997, 1998). By showing that information can be encoded in the complexity of Ca²⁺ signals, these examples underscore the need to understand the processes that generate these response patterns.

Ca²⁺ dynamics are generated by the interplay of Ca²⁺ influx and efflux pathways. In many types of non-excitable cell, Ca²⁺ comes from the endoplasmic reticulum (ER) through inositol 1,4,5-trisphosphate (IP₃) receptors and ryanodine receptors, while Ca²⁺ from the extracellular space enters through store-operated Ca²⁺ channels in the plasma membrane that open in response to the depletion of intracellular Ca²⁺ stores (Berridge, 1995; Parekh & Penner, 1997; Lewis, 2001). Although historically, Ca²⁺ entry mechanisms have been given the most attention in terms of shaping Ca²⁺ signals, pathways for Ca²⁺ clearance are becoming increasingly well recognized as powerful influences. The plasma-membrane Ca²⁺-ATPase (PMCA), sarcoplasmic and ER Ca²⁺-ATPases (SERCA), Na⁺-Ca²⁺ exchange, mitochondria and intracellular buffers can each have characteristic effects on the time course and amplitude of cellular Ca²⁺ signals, depending upon the cell type examined.

Several features of the PMCA gives it a great potential for augmenting the complexity of intracellular Ca²⁺ signals. Over 30 different isoforms with different Ca²⁺ transport properties can be produced through alternative splicing of four PMCA genes. Many of the isoforms differ with respect to their basic Ca²⁺ transport properties and are expressed in unique combinations in different cells (Guerini et al. 1998; Caride et al. 2001; Strehler & Zacharias, 2001). Moreover, PMCA activity can be modulated in vitro by phosphorylation and by calcium-calmodulin in an isoform-specific manner. The degree and kinetics of modulation differ for the different isoforms, such that each could in principle shape the time course of Ca²⁺ signals in different ways (Stauffer et al. 1995; Penniston & Enyedi, 1998; Caride et al. 2001; Strehler & Zacharias, 2001). Slow PMCA modulation has been described in several cell types in vivo (erythrocytes, neutrophils and endothelial cells), where it is thought to affect the generation of biphasic Ca²⁺ responses, spikes and oscillations (Scharff et al. 1983; Scharff & Foder, 1994; Madge et al. 1997; Klishin et al. 1998; Snitsarev & Taylor, 1999).

In the study presented here, we have examined the role of the PMCA in shaping Ca²⁺ responses in Jurkat T cells. We have found that the PMCA is the primary extrusion mechanism in these cells and that its activity is slowly modulated by changes in [Ca²⁺]_i. This modulation allows the pump to adapt to increases in calcium-release-activated Ca²⁺ (CRAC) channel activity, thereby enabling the cell to maintain stable [Ca²⁺]_i elevations even at levels that saturate pump transport sites. In addition, the slow time course of the pump modulation and its reversal play a major role in producing the initial overshoot of the biphasic [Ca²⁺]_i response, and enables the pump to act as a high-pass filter with memory. Thus, in addition to its homeostatic role, the PMCA actively contributes to the complexity of Ca²⁺ signalling in T cells. Some of these results have been reported in abstract form (Bautista et al. 1998).

	METHODS

Top Abstract Introduction Methods Results Discussion References

Cells and solutions

Experiments were performed with Jurkat E6-1, a human leukaemic T-cell line (ATCC 1378; American Type Culture Collection, Rockville, MA, USA). Cells were grown in medium consisting of RPMI 1640 (Mediatech, Herndon, VA, USA) supplemented with 10 % fetal calf serum (Atlanta Biologicals, Atlanta, GA, USA), 1 mM L-glutamine, 50 U ml^-1 penicillin and 50 µg ml^-1 streptomycin (Mediatech). Cells were maintained in log-phase growth at 37 °C and 6 % CO₂. Extracellular Ringer solution contained (mM): 155 NaCl, 4.5 KCl, 2 CaCl₂, 1 MgCl₂, 10 D-glucose and 5 Hepes (pH 7.4 with NaOH). In the calcium-free solution, 2 mM MgCl₂ + 1 mM EGTA were substituted for CaCl₂. The sodium-free solution contained 155 mM N-methyl-D-glucamine-Cl or 155 mM LiCl in place of NaCl. The pipette solution for perforated-patch recording contained (mM): 115 caesium aspartate, 1 CaCl₂, 5 MgCl₂, 10 NaCl and 10 Hepes (pH 7.2 with CsOH), plus 300 µg ml^-1 amphotericin B (Sigma Chemical, St Louis, MO, USA). Thapsigargin (LC Biochemicals, Woburn, MA, USA) was diluted from a 1 mM stock solution in DMSO, and oligomycin (Sigma) was prepared from a 2 mM DMSO stock solution. Antimycin A1 (Sigma) was diluted from a 2 mM stock in 100 % ethanol. OKT3 ascites was generously provided by I. A. Graef and G. R. Crabtree (Stanford University) and was diluted 1:100 in Ringer solution.

Video microscope measurements of [Ca²⁺]_i

Cells were loaded with 1 µM fura-2/AM (Molecular Probes, Eugene, OR, USA) at 22-25 °C for 30 min in culture medium, washed with fresh medium, and attached to poly-L-lysine-treated coverslip chambers on the stage of a Zeiss Axiovert 35 microscope. Several minutes prior to imaging, cells were washed with Ringer solution. Cells were illuminated using a xenon light source and filter wheel (Lambda LS and Lambda-10, Sutter Instruments, Novato, CA, USA) for 132 ms, alternately at 350 nm and 380 nm (bandpass filters from Chroma Technology, Brattleboro, VT, USA) through a 40 Zeiss Achrostigmat objective (NA 1.3). Fluorescence emission at > 480 nm (longpass filter from Chroma Technology) was captured with an intensified CCD camera (Hamamatsu, Bridgewater, NJ, USA) and was digitized, background corrected and analysed with a VideoProbe imaging system (ETM Systems, Irvine, CA, USA), as described previously (Dolmetsch & Lewis, 1994). Background-corrected 350/380 ratio images were collected every 5 s. [Ca²⁺]_i was determined from the relationship.

[Ca²⁺]_i = K*(R - R_min)/(R_max - R),

where R is the F₃₅₀/F₃₈₀ ratio, R_min and R_max are the ratios at 0 Ca²⁺ and saturating Ca²⁺ (10 mM), respectively, and K* is the apparent dissociation constant (Almers & Neher, 1985). Values of K*, R_min and R_max were measured in situ in Jurkat cells, as described previously (Lewis & Cahalan, 1989). All experiments were conducted at 22-25 °C.

Perforated-patch recording with simultaneous [Ca²⁺]_i measurements

Cells were loaded with 1 µM indo-1/AM in culture medium at 22-25 °C for 25 min, washed and attached to coverslip chambers on the stage of a Nikon Diaphot TMD microscope. All experiments were conducted at 22-25 °C. Cells were illuminated at 360 nm (360/25 filter; Chroma Technology) for 40-80 ms every 0.5-4.0 s through a 40 Nikon Fluor objective (NA 1.3), and the fluorescence emissions at 405 nm and 485 nm (405/25 and 485/25 filters, respectively; Chroma Technology) were collected simultaneously with two photomultipliers (HC124-02, Hamamatsu), and averaged. The background-corrected 405/485 ratio was calculated every 0.5-4.0 s. [Ca²⁺]_i was estimated using the equation listed above. K*, R_min and R_max were measured in situ in Jurkat cells, as described previously (Zweifach & Lewis, 1995).

Recording pipettes were pulled from 100 µl capillaries (VWR Scientific, South Plainfield, NJ, USA), coated with Sylgard (Dow Corning, Midland, MI, USA), and fire-polished to resistances of 2-4 M. Membrane currents were recorded using an Axopatch 200 amplifier (Axon Instruments, Union City, CA, USA), filtered at 2 kHz and digitized at a rate of 5 kHz. Command potentials and data collection were controlled by a Power Macintosh G3 computer (Apple, Cupertino, CA, USA) driving an ITC-16 interface (Instrutech, Great Neck, NY, USA), using custom-designed software extensions to Igor Pro (Wavemetrics, Lake Oswego, OR, USA). All command potentials were corrected for a measured liquid junction potential of -12 mV existing between the perforated-patch pipette solution and 2 mM Ca²⁺ Ringer solution. Access resistance in perforated-patch experiments varied from 4 to 6 M. CaCl₂ (1 mM) was included in the pipette solution so that accidental rupture of the patch could be easily detected by saturation of the indo-1 signal.

In the experiments reported in Fig. 2 and Fig. 10, calcium-release-activated Ca²⁺ current (I_CRAC) was measured at different holding potentials in the following way. The whole-cell current was averaged for 100-200 ms at the stated holding potential every 0.5-4.0 s, and 100 ms voltage ramps from -100 or -120 mV to +50 mV were applied every 12 s to verify the stability of I_CRAC and the leak current (I_leak). I_leak was measured in response to voltage ramps in calcium-free Ringer solution at the beginning and end of each experiment. Only experiments in which I_leak remained constant were analysed. I_CRAC was then determined at a given holding potential by subtracting the value of I_leak at that potential (obtained from the ramp currents) from the total average current.

ATP measurements

Intracellular ATP concentrations were measured using a luciferin- luciferase ATP assay kit (Calbiochem, La Jolla, CA, USA). Briefly, a suspension of 10⁶ cells ml^-1 was lysed in ATP-releasing agent, incubated with an excess of D-luciferin, and immediately monitored for peak luminescence in a Berthold luminometer. ATP standards were used to convert luminescence to intracellular ATP concentration, assuming an average cell volume of 2 pl, based on the average capacitance of Jurkat cells.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Delayed activation of Ca²⁺ clearance contributes to biphasic [Ca²⁺]_i responses

Stimulation of fura-2-loaded Jurkat T cells with anti-CD3 monoclonal antibody (OKT3) elicited a biphasic increase in [Ca²⁺]_i after a delay of 60 ± 4 s (mean ± S.E.M., n = 307 cells; Fig. 1A). The response consisted of a rapid rise in [Ca²⁺]_i to a peak of 1 µM, followed by a slower decline to a plateau of ~0.5 µM. This type of biphasic response is typical of many cell types following stimulation leading to the production of IP₃. In most cases, including T cells, the initial spike has been attributed to the release of Ca²⁺ from intracellular stores by IP₃, while the plateau is thought to reflect sustained Ca²⁺ influx through store-operated Ca²⁺ channels in the plasma membrane following store depletion (Putney, 1997).

		View larger version [in this window] [in a new window]
Figure 1. T cell receptor stimulation evokes a biphasic rise in [Ca2+]i A, Jurkat cells were treated with OKT3 (1:100 ascites) in 2 mM Ca2+ Ringer solution to release Ca2+ from intracellular Ca2+ stores and activate Ca2+ influx through calcium-release-activated Ca2+ (CRAC) channels. A biphasic [Ca2+]i response was observed in the population average (n = 307 cells) as well as in individual cells (not shown). B, the biphasic response occurred even after Ca2+ release from stores was complete. Cells were treated with 1:100 OKT3 in Ca2+-free Ringer solution to completely empty the inositol 1,4,5-trisphosphate (IP3)-sensitive stores. Subsequent readdition of 2 mM Ca2+ evoked a biphasic [Ca2+]i rise due to influx through open CRAC channels. Average of 277 cells.

However, Ca²⁺ release from intracellular stores does not appear to be the primary cause of the large Ca²⁺ overshoot observed in T cells. As shown in Fig. 1B, treatment with OKT3 in the absence of extracellular Ca²⁺ (Ca) elicits a small release transient that is only a fraction of the size of the overshoot seen in the presence of Ca. Interestingly, despite the cessation of Ca²⁺ release from stores, reapplication of 2 mM Ca causes a large [Ca²⁺]_i overshoot, indicating that the biphasic response depends upon Ca²⁺ entry. Several mechanisms could in principle be involved, including calcium-induced Ca²⁺ release (CICR), which may be triggered by Ca²⁺ influx, a gradual decline of Ca²⁺ entry, or delayed removal of cytosolic Ca²⁺ by intracellular organelles or Ca²⁺ pumps in the plasma membrane. Each of these possibilities is considered in turn below.

Two results argue against a significant role for CICR in generating the Ca²⁺ overshoot. First, application of 4 µM ionomycin to cells pretreated with OKT3 failed to release any additional increase in [Ca²⁺]_i, demonstrating that intracellular stores were fully depleted by OKT3 in calcium-free Ringer solution (Fig. 2A, left panel). Similar results were obtained following treatment with 1 µM thapsigargin (TG; data not shown), a potent inhibitor of SERCA-type Ca²⁺-ATPases (Thastrup et al. 1989). These results do not rule out the possibility that in the presence of OKT3, Ca²⁺ influx might load a store that could subsequently release Ca²⁺ through CICR. However, release of Ca²⁺ was not observed following a brief application of Ca²⁺ to OKT3-treated cells (Fig. 2A, right panel) that brought [Ca²⁺]_i to a level similar to the plateau shown in Fig. 1. Thus, store refilling followed by Ca²⁺ release cannot account for the overshoot.

		View larger version [in this window] [in a new window]
Figure 2. Ca2+ clearance shapes the biphasic response A, calcium-induced Ca2+ release (CICR) was tested as a possible contributor to the biphasic Ca2+ response. Cells were treated with OKT3 as in Fig. 1B. Following Ca2+ release from stores, application of 4 µM ionomycin caused a minimal increase in [Ca2+]i, indicating that stores had indeed been emptied by OKT3 (left graph). Average of 271 cells. In the right-hand graph, extracellular Ca2+ (Ca) was applied briefly to OKT3-treated cells to elevate [Ca2+]i to the plateau level seen in Fig. 1B. [Ca2+]i did not continue to increase after removal of Ca, showing that the overshoot in Fig. 1B was not due to CICR. Average of 186 cells. B, constant Ca2+ influx evoked biphasic [Ca2+]i elevation in a voltage-clamped cell. Thapsigargin (TG; 1 µM) was applied to an indo-1-loaded Jurkat cell bathed in 2 mM Ca2+ Ringer solution to deplete Ca2+ stores and open CRAC channels. In this perforated-patch recording, shifting the holding potential from +30 mV to -100 mV induced Ca2+ influx through open CRAC channels, resulting in a rapid increase in [Ca2+]i to a peak level followed by a slow decay to a steady-state plateau. The inset shows a leak-subtracted ramp current measured 8 s after hyperpolarization, illustrating the inward rectification and positive reversal potential characteristics of CRAC.

A second possible mechanism for the Ca²⁺ overshoot is a slow decline in Ca²⁺ entry, perhaps resulting from inactivation of CRAC channels or membrane depolarization, which reduces the driving force for Ca²⁺ entry. To test for these possibilities, we monitored changes in [Ca²⁺]_i caused by a step increase in I_CRAC in cells under voltage-clamp conditions. As shown in Fig. 2B, hyperpolarization of a TG-treated cell from +30 mV to a fixed potential of -100 mV evoked a constant increase in inward current. The current was identified as I_CRAC based on several characteristic properties, including an absolute dependence on Ca, inward rectification, lack of easily detectable current noise, and a lack of outward current at potentials up to ~+50 mV (Parekh & Penner, 1997; Lewis, 2001). Even though I_CRAC increased to a constant level, a biphasic [Ca²⁺]_i increase was produced. This result demonstrates that the overshooting Ca²⁺ response does not require changes in Ca²⁺ influx through CRAC channels, but instead must reflect a delayed increase in the Ca²⁺ clearance rate.

Contributions of mitochondria and ER to Ca²⁺ clearance

To determine the source of the delayed increase in Ca²⁺ clearance, we first examined the contribution of different known Ca²⁺ clearance mechanisms to the recovery from small and large increases in [Ca²⁺]_i. Following store depletion with OKT3 in 0 Ca, Ca²⁺ influx was stimulated by exposure to 2 mM Ca²⁺ for either 15 s or 40 s (Fig. 3A). In each case, we estimated the rates of Ca²⁺ clearance by fitting a single- or double-exponential curve to the decline of [Ca²⁺]_i immediately following Ca removal.

		View larger version [in this window] [in a new window]
Figure 3. Complex Ca2+ clearance kinetics in Jurkat cells A, recovery from 15- and 40-s periods of Ca2+ influx in OKT3-treated cells. Ca2+ (2 mM) was applied as indicated by the bars. The recovery of [Ca2+]i following a 15-s period of Ca2+ entry followed an exponential time course. Addition of Ca2+ for 40 s evoked larger transients that decayed to baseline with an approximately biexponential time course in single cells as well as in the population average from 122 cells shown here. B, oscillations during [Ca2+]i recovery in single cells. Recovery from the second Ca2+ application in A is shown for four cells, superimposed on the average response from A. Of the cell population, 12 % (15/122 cells) responded in this way.

In the experiment shown in Fig. 3A, a 15-s application of 2 mM Ca elicited a rapid but modest rise of [Ca²⁺]_i (peak values of 620 ± 27 nM, n = 122 cells). Recovery of [Ca²⁺]_i to baseline followed an exponential time course in single cells ( = 36 ± 5 s) and the population average (Fig. 3A). Subsequent addition of Ca for 40 s evoked larger transients (1.63 ± 0.04 µM) that decayed to baseline with a time course approximated by the sum of two exponentials (₁ = 10.4 ± 0.7 s and ₂ = 96 ± 7 s). Thus, increasing the duration of capacitative Ca²⁺ entry (CCE) from 15 to 40 s reduced the initial time constant of recovery (₁) by approximately fourfold, indicating a calcium-dependent increase in the clearance rate. As shown in Fig. 3B, low-amplitude [Ca²⁺]_i oscillations occurred during the slow phase of the [Ca²⁺]_i decay in 12 % of the cells (15/122 cells). The complexity of the [Ca²⁺]_i decay kinetics may reflect the activity of multiple Ca²⁺ clearance pathways. Below we examine several of these, including sequestration and release by mitochondria, uptake into ER stores by SERCA-type Ca²⁺-ATPases, and Ca²⁺ export across the plasma membrane by PMCA and Na⁺-Ca²⁺ exchange.

Mitochondria have a large capacity to sequester and release Ca²⁺, which can significantly influence the time course of Ca²⁺ recovery in many cells (Thayer & Miller, 1990; Friel & Tsien, 1994; Hoth et al. 1997; Babcock & Hille, 1998). We tested for a role of mitochondria by inhibiting mitochondrial Ca²⁺ uptake with 2 µM antimycin A1 + 1 µM oligomycin (Fig. 4B). The Ca²⁺ clearance kinetics following a 15-s application of Ca were not significantly altered by the inhibitors ( = 36.7 ± 0.7 s, n = 255). However, the inhibitors changed the recovery time course after the 40-s Ca²⁺ pulse to a single exponential with a time constant that was intermediate between the fast and slow time constants measured under control conditions ( = 20.2 ± 0.5 s, n = 255 cells; cf. Fig. 4A). These results are consistent with evidence from T cells and other cells showing that mitochondria accelerate the early phase of recovery by sequestering Ca²⁺, and create a second slower phase by releasing this stored Ca²⁺ after [Ca²⁺]_i falls below ~400 nM (Friel & Tsien, 1994; Hoth et al. 1997; Colegrove et al. 2000).

		View larger version [in this window] [in a new window]
Figure 4. Contributions of plasma-membrane Ca2+-ATPase (PMCA), sarcoplasmic and endoplasmic reticulum Ca2+-ATPases (SERCA) and mitochondria to Ca2+ clearance Cells were stimulated with OKT3 or TG in calcium-free Ringer solution with or without mitochondrial inhibitors, followed by brief applications of 2 mM Ca. In each panel, the average [Ca2+]i of 122-255 cells is shown, and the cartoons at the right illustrate the major Ca2+ flux pathways that are functional during recovery in zero Ca. A, clearance in OKT3-stimulated cells. Data reproduced from Fig. 3A. B, mitochondrial contribution to Ca2+ clearance in cells with functional SERCA. Inhibiting mitochondria with 2 µM antimycin A1 + 1 µM oligomycin (anti/oligo) elicited a monoexponential decay following both 15-s and 40-s applications of Ca2+. The decay time course following the 40-s period of capacitative Ca2+ entry (CCE) was intermediate between the fast and slow time constants seen in A. C, contribution of SERCA to Ca2+ clearance. Full inhibition of SERCA by 1 µM TG had no effect on clearance after a short period of Ca2+ entry. Following 40 s of CCE, the slow component of decay was accelerated by ~30 %. D, effect of simultaneous inhibition of SERCA and mitochondria. With SERCA- and mitochondria-mediated Ca2+ uptake blocked, the acceleration of Ca2+ clearance during the second Ca2+ application must arise from enhanced efflux across the plasma membrane.

SERCA are known to influence T-cell Ca²⁺ signalling by controlling store refilling and therefore CRAC channel activity, but the degree to which ER Ca²⁺ uptake directly contributes to Ca²⁺ clearance in T cells is not known (Lewis, 2001). The role of SERCA-type Ca²⁺ pumps in clearing cytosolic Ca²⁺ was assessed using TG. Cells were treated with 1 µM TG in calcium-free Ringer solution to completely deplete stores (Fig. 4C). The time course of [Ca²⁺]_i recovery following a subsequent 15-s application of Ca ( = 40.2 ± 0.9 s, n = 240 cells) was not significantly different from that of OKT3-treated cells with functional SERCA (Fig. 4A). Thus, at least under conditions of strong T-cell receptor (TCR) stimulation (and presumably high concentrations of IP₃), it appears that SERCA plays no detectable role in Ca²⁺ clearance at low [Ca²⁺]_i. However, SERCA inhibition does affect clearance after 40-s applications of Ca, although in a surprising way. As in the control cells, the recovery time course of TG-treated cells is biexponential; however, TG treatment significantly accelerated the slow component (₁ = 6.9 ± 0.7 s and ₂ = 63.4 ± 6.6 s, n = 240 cells), while eliminating the small [Ca²⁺]_i oscillations normally seen during the slow phase of the decay (compare Fig. 4C with Fig. 4A). Thus, contrary to expectation, SERCA-mediated ER store refilling actually retards the removal of Ca²⁺ from the cytosol following prolonged periods of Ca²⁺ influx, by slowing the final phase of clearance.

To understand further the paradoxical role of SERCA activity in slowing Ca²⁺ clearance, we inhibited simultaneously both SERCA and mitochondria-mediated [Ca²⁺]_i uptake (Fig. 4D). Inhibition of SERCA had no effect on clearance in cells lacking functional mitochondria (i.e. treated with antimycin A1 + oligomycin; compare Fig. 4B and 4D). This suggests that SERCA activity does not influence Ca²⁺ clearance unless mitochondria are able to sequester and/or release Ca²⁺. Inhibition of either SERCA or mitochondria eliminated the [Ca²⁺]_i oscillations associated with the slow phase of clearance, demonstrating that the oscillations are produced by the combined action of both organelles. One plausible interpretation of these results is that mitochondria release Ca²⁺ that is then taken up by the ER and eventually released through IP₃ receptors, and that this cycling of Ca²⁺ between the two organelles slows the final phase of extrusion across the plasma membrane (see Discussion).

Ca²⁺ export in Jurkat T cells occurs solely via the PMCA

Two Ca²⁺ extrusion mechanisms have been proposed to operate in Jurkat T cells: the Na⁺-Ca²⁺ exchanger and the PMCA (Balasubramanyam et al. 1993, 1994 Donnadieu & Trautmann, 1993; Balasubramanyam & Gardner, 1995; Gardner & Balasubramanyam, 1996). To examine the roles of each in Ca²⁺ clearance, experiments were performed in the presence of TG and antimycin A1 + oligomycin to eliminate any contribution of SERCA or mitochondria.

We tested for the presence of Na⁺-Ca²⁺ exchange by removing extracellular Na⁺ during recovery from a brief [Ca²⁺]_i rise. Following store depletion with TG, Ca²⁺ influx was stimulated twice for 60 s; extracellular Na⁺ was present during the first period of recovery, and was replaced with N-methyl-D-glucamine (NMDG) during the second. In some cells we observed a decrease in the peak magnitude of the Ca²⁺ responses over time. Since upregulation of the clearance rate is dependent upon [Ca²⁺]_i (Fig. 4D), rundown of the peak [Ca²⁺]_i would be expected to diminish the clearance rate regardless of whether or not a Na⁺-Ca²⁺ exchange was present. Therefore, only cells having stable peak [Ca²⁺]_i values (to within 100 nM) were analysed. As shown in Fig. 5A, the average clearance kinetics in NMDG ( = 28 ± 9 s, n = 302 cells) were identical to those observed in the presence of Na⁺ ( = 30 ± 8 s). Single-cell analysis shows that removal of Na⁺ does not discernibly affect Ca²⁺ extrusion over a wide range of peak [Ca²⁺]_i values from 200 nM to 2 µM (Fig. 5B). Furthermore, Na⁺ removal did not alter the Ca²⁺ clearance rates measured after ATP depletion or in the presence of La³⁺ to block PMCA activity (see below). Therefore, the evidence suggests that Na⁺-Ca²⁺ exchange does not contribute significantly to Ca²⁺ clearance in Jurkat T cells.

		View larger version [in this window] [in a new window]
Figure 5. Na+-Ca2+ exchange does not contribute to Ca2+ clearance Ca2+ extrusion was not affected by the replacement of extracellular Na+ with N-methyl-D-glucamine (NMDG). A, cells were stimulated as in Fig. 4D (TG + anti/oligo). CCE was induced for 60 s and clearance rates were measured in the presence and absence of Na+, as indicated. B, extrusion rates vs. peak [Ca2+]i in single cells from the experiment shown in A.

Three independent methods were used to assess the role of the PMCA in Ca²⁺ extrusion. First, Ca²⁺ clearance was measured in cells depleted of ATP. Inhibition of mitochondrial respiration and ATP synthesis using 2 µM antimycin A1 + 2 µM oligomycin alone did not affect ATP levels. However, antimycin A1 and oligomycin in combination with 10 mM 2-deoxy-D-glucose to inhibit glycolysis was effective; after 60 min at 37 °C, ATP levels in Jurkat cells were reduced from 1.5 mM to 3 µM, a decrease of ~99 % (see Methods). ATP depletion greatly diminished the rate of Ca²⁺ entry via CRAC channels, consistent with previous studies of thymocytes and rat basophil leukaemia cells (Mohr & Fewtrell, 1990; Gamberucci et al. 1994; Marriott & Mason, 1995). Thus, to elevate [Ca²⁺]_i in ATP-depleted cells we applied 2 µM ionomycin rather than TG. ATP depletion slowed the rate of Ca²⁺ clearance to variable extents in single cells; effects ranged from a slight reduction to a total lack of extrusion, as shown by the examples given in Fig. 6A. Exponential curves were fitted to the [Ca²⁺]_i recovery data from single cells, and the results are summarized in the cumulative histogram in Fig. 6B. We estimated the overall effect of ATP depletion on clearance by calculating the mean time constant from single cells in the population. ATP depletion lengthened the mean time constant for Ca²⁺ removal from 33 ± 12 s to 1032 ± 287 s (n = 288), suggesting a 97 % inhibition of Ca²⁺ extrusion. These results are consistent with a dominant role of Ca²⁺-ATPases in exporting Ca²⁺ across the plasma membrane.

		View larger version [in this window] [in a new window]
Figure 6. The role of the PMCA in Ca2+ clearance from Jurkat cells A, effect of ATP depletion on Ca2+ clearance kinetics in single cells. Control cells were treated with 2 µM ionomycin to generate a large Ca2+ transient, followed by calcium-free Ringer solution. In ATP-depleted cells (see text) Ca2+ clearance was inhibited to varying extents, as demonstrated by these single-cell responses. B, effects of ATP depletion at the population level (288 cells). The cumulative histogram shows the fraction of cells with particular clearance time constant ; the right-most point includes cells with 1000 s. The dotted line shows that clearance has been inhibited by 95 % in > half of the cells. C, La3+ (2 mM) inhibited Ca2+ clearance by > 90 % in a population of 257 cells. D, effect of carboxyeosin on Ca2+ clearance. Because carboxyeosin is fluorescent at fura-2 wavelengths, experiments were conducted on single indo-1-loaded cells following a 30-min preincubation with 20 µM carboxyeosin. Following treatment with carboxyeosin, the recovery from a CCE transient was reduced by a factor of approximately three (average of three cells).

To test further the role of the PMCA, we applied La³⁺ and carboxyeosin, two reagents that have been reported to inhibit PMCA activity in other cells (Carafoli, 1991; Gatto et al. 1995). When applied at a concentration of 2 mM, La³⁺ inhibited clearance by ~90 % (Fig. 6C). Preincubation with carboxyeosin (20 µM) inhibited clearance by 64 % and raised the resting [Ca²⁺]_i, as would be expected from inhibition of resting Ca²⁺ efflux without inhibition of CRAC channels (Fig. 6D). A higher dose of carboxyeosin (50 µM) did not further inhibit efflux, suggesting that the incomplete inhibition we observed is not concentration limited.

Taken together, these results suggest that the PMCA is the primary Ca²⁺ extrusion mechanism in the Jurkat cell plasma membrane, and that it is solely responsible for Ca²⁺ clearance when SERCA and mitochondrial uptake are inhibited. Therefore, we conclude that the increased clearance rate observed after prolonged Ca²⁺ entry (see Fig. 4D) is caused directly by calcium-dependent modulation of PMCA activity.

Calcium- and time-dependent modulation of the PMCA

To measure the speed with which PMCA modulation develops, we quantified the rate of PMCA-mediated clearance after stimulating Ca²⁺ influx for different periods of time. CCE was triggered in cells pretreated with antimycin + oligomycin + TG by exposure to 2 mM [Ca²⁺]_o for 10-300 s (Fig. 7A). After each period of Ca²⁺ exposure, clearance followed a single exponential time course, and the time constant declined with increasing duration of CCE. The time constant of Ca²⁺ removal approached a minimum value of 18 s after 60 s of CCE, representing a three- to fourfold increase in PMCA activity (Fig. 7B). It should be noted that because recovery follows an approximately exponential time course, the change in reflects the change in pump rate at any given value of [Ca²⁺]_i; in other words, the decrease in indicates that the pump rate is increased at all levels of [Ca²⁺]_i. Thus, PMCA modulation can be viewed as separate from the calcium-dependent activity that results from Ca²⁺ binding to the transport sites.

		View larger version [in this window] [in a new window]
Figure 7. Calcium- and time-dependent modulation of PMCA activity The experimental conditions described in Fig. 4D were also used here. A, Ca2+ influx was induced in TG-treated cells for varying periods of time (10-300 s) by changing the duration of exposure to 2 mM Ca2+ Ringer solution (dashed arrows). Single-exponential fits are superimposed on each recovery phase. B, PMCA activity increased with the duration of CCE. For each duration, the average (± S.E.M.) was obtained from ~500 single-cell measurements in 3 experiments, as described in A.

The kinetics of reversal of PMCA modulation were measured using the three-pulse protocol illustrated in Fig. 8A. A brief 20-s application of Ca²⁺ (the 'reference pulse') was given first to evoke a small [Ca²⁺]_i rise with a slow subsequent decay, reflecting the unmodulated state. Thereafter, a 90-s elevation of [Ca²⁺]_i (the 'modulating pulse') was applied to cause substantial PMCA modulation, reducing the time constant of clearance. A variable recovery period in zero-calcium Ringer solution was then followed by a third brief 20-s Ca²⁺ pulse (the 'test pulse'). The degree of recovery from modulation was determined by comparing the decay of the test pulse response to that of the reference. The results are summarized in the normalized recovery curve shown in Fig. 8B. After termination of Ca²⁺ entry, recovery from PMCA modulation occurred with a roughly exponential time course ( = 240 s). Thus, the reversal of modulation is much slower than its induction and is 5-10 times slower than the time course of [Ca²⁺]_i recovery.

		View larger version [in this window] [in a new window]
Figure 8. Reversal kinetics of PMCA modulation A, a three-pulse protocol was used to measure the time course of recovery from PMCA modulation. Experimental conditions were identical to those in Fig. 4D. An initial brief Ca2+ application established the activity of the unmodulated PMCA (ref), and a subsequent longer application induced modulation, as shown by the decreased of clearance (mod). A third brief test pulse was given at varying intervals after the modulating pulse, and the time constant of clearance (test) was used to monitor recovery from modulation. The upper trace shows that PMCA modulation persisted after 180 s, while the lower trace shows that reversal of modulation was complete after 500 s. B, kinetics of recovery from modulation. The extent of recovery was calculated as (test - mod)/(ref - mod), and each data point is an average of ~1000 cells from a minimum of 3 experiments. A single exponential curve with a of 240 s is superimposed on the data.

PMCA modulation changes both the K_M and maximal rate of extrusion (V_max)

To characterize further the changes in PMCA activity that underlie the modulation process, we measured the [Ca²⁺]_i dependence of pump rates in the resting and modulated states. The [Ca²⁺]_i dependence of the unmodulated PMCA was determined in response to [Ca²⁺]_i transients of varying amplitudes, effected by hyperpolarizing the membrane for 2-7 s from the holding potential (+30 mV) to potentials of -20 to -100 mV (Fig. 9A). To minimize modulation during these responses, only the initial slope of the [Ca²⁺]_i recovery (measured within 9 s of initiating the [Ca²⁺]_i rise) was measured. The time course of modulation shown in Fig. 7 predicts about 10 % modulation within 7 s during the largest responses seen in Fig. 9. Thus, a plot of the initial slope (-d[Ca²⁺]_i/dt) against [Ca²⁺]_i reflects the Ca²⁺ dependence of mostly unmodulated PMCAs (Fig. 9C).

		View larger version [in this window] [in a new window]
Figure 9. PMCA modulation changes the Ca2+ dependence of the clearance rate Perforated-patch recording and [Ca2+]i measurements were made from an indo-1-loaded cell treated with 1 µM TG. A, measurement of the Ca2+ dependence of clearance by the unmodulated PMCA. [Ca2+]i transients of varying amplitudes were generated by hyperpolarization for 2-7 s from the holding potential of +30 mV to potentials of -50 to -120 mV (bars). Fits to the initial slope of the [Ca2+]i decay are superimposed and are plotted in C (). B, Ca2+ clearance by the modulated PMCA. A steady-state level of modulation was produced by hyperpolarization to -120 mV for 80 s. Upon return to +30 mV, [Ca2+]i decayed back to baseline. The derivative of the [Ca2+]i decay is plotted against the corresponding [Ca2+]i in C (). C, PMCA modulation enhanced clearance at all levels of [Ca2+]i. Michaelis-Menten relationships (see text) were fitted to the d[Ca2+]i/dt vs. [Ca2+]i data from A and B, yielding estimates of KM = 480 nM, Vmax = 29 nM s-1, nH = 2 (unmodulated) and KM = 430 nM, Vmax = 50 nM s-1, nH = 2 (modulated). Because the effects of cytosolic Ca2+ buffering are not included here, these fits provide only an empirical description of the Ca2+ dependence of the clearance rate, and do not describe PMCA activity itself (compare with Fig. 11).

To measure the Ca²⁺ dependence of the modulated PMCAs, the same cell was hyperpolarized to -100 mV for 80 s, triggering a typical biphasic Ca²⁺ response. Upon repolarization to +30 mV, [Ca²⁺]_i decayed back to baseline (Fig. 9B). Modulation is expected to reach a steady state during the pulse, and given the slow reversal rate (Fig. 8) should decline by 10 % during the time it takes [Ca²⁺]_i to fall to baseline. Therefore, a plot of the derivative of the entire [Ca²⁺]_i decay against the corresponding [Ca²⁺]_i reveals the Ca²⁺ dependence of the modulated PMCA (Fig. 9C).

The results plotted in Fig. 9C show that as expected, the rate of Ca²⁺ clearance increases with [Ca²⁺]_i, and that modulation of the PMCA further increases this rate over the entire range of [Ca²⁺]_i values. Similar results were obtained in two other cells. A Michaelis-Menten model can be used to describe this behaviour:

(1)

where V is the rate of Ca²⁺ extrusion, V_max is the maximal rate, K_M is the [Ca²⁺]_i at half of V_max, and n_H is the Hill coefficient describing the cooperativity of the process. By assuming that d[Ca²⁺]_i/dt represents the pump rate, the Michaelis-Menten model has been used to describe Ca²⁺ clearance data in other cells (Sedova & Blatter, 1999), and in fact the model fitted our Jurkat data well. The curves shown in Fig. 9C represent K_M = 480 nM, V_max = 29 nM s^-1 and n_H = 2 for the unmodulated PMCA, and K_M = 430 nM, V_max = 50 nM s^-1 and n_H = 2 for the modulated pump. However, such analysis can be highly misleading, as the d[Ca²⁺]_i/dt values do not reflect pumping alone, but rather the combination of pumping and Ca²⁺ redistribution between intracellular buffers and the cytosol. For this reason, we measured the intracellular Ca²⁺ buffering capacity and used this information to convert Ca²⁺ clearance rates to PMCA flux rates.

Intracellular buffers slow the rate at which [Ca²⁺]_i changes in proportion to the fraction of Ca²⁺ that is bound, according to the relationship:

(2)

where J_Ca is the Ca²⁺ flux rate (units of mol s^-1) and is given by:

= v, (3)

where v is the accessible cell volume (units of litres), and is the total Ca²⁺ binding capacity of endogenous and exogenous buffers in the cytosol. has units of litres and can be considered to be the 'equivalent cell volume'. is defined as the ratio of the total Ca²⁺ entering or leaving the cell ([Ca²⁺]_TOT) to the change in free cytosolic Ca²⁺, [Ca²⁺]_i (Neher, 1995):

(4)

Equation (2) provides a means of measuring over a range of [Ca²⁺]_i in indo-1-loaded Jurkat cells. When I_CRAC is changed abruptly by an amount I_CRAC, the change in the total Ca²⁺ flux rate causing a change in d[Ca²⁺]_i/dt is given by:

(5)

where F is Faraday's constant. We calculated from eqn (5), based on the measured changes in I_CRAC and d[Ca²⁺]_i/dt in response to brief voltage-clamp hyperpolarizations (Fig. 10A and B). With suitably brief hyperpolarizations, the change in [Ca²⁺]_i is small enough that and PMCA activity are approximately constant. This condition was satisfied by applying 1-2 s voltage steps to -100 mV first from a holding potential of +30 mV (to measure at the resting [Ca²⁺]_i level of 50-150 nM; Fig. 10A) and thereafter from a holding potential of -80 mV (to measure at a high [Ca²⁺]_i level of approximately 1 µM; Fig. 10B). Calculated values of (expressed in terms of equivalent volume) for the cell shown in Fig. 9 are plotted against [Ca²⁺]_i in Fig. 10C.

		View larger version [in this window] [in a new window]
Figure 10. The Ca2+ dependence of cytosolic Ca2+ buffering in indo-1-loaded cells Measurements from the same cell as used for Fig. 9. A, perturbations of ICRAC and d[Ca2+]i/dt at low [Ca2+]i. d[Ca2+]i/dt and ICRAC were increased by a brief hyperpolarization (+30 mV to -50 mV). ICRAC was measured after subtraction of the leak current collected in zero-calcium Ringer solution at the appropriate voltage (+30 or -50 mV) as described (see Methods). The initial slope of d[Ca2+]i/dt was determined as shown. B, perturbations of ICRAC and d[Ca2+]i/dt at high [Ca2+]i. The cell was hyperpolarized from +30 mV to -80 mV to drive [Ca2+]i to a high plateau. Then, brief steps from -80 to -120 mV were applied and changes in ICRAC and d[Ca2+]i/dt were measured. C, the effective cell volume, , a measure of total buffering capacity () scaled by cell volume, as a function of [Ca2+]i. was calculated at each value of [Ca2+]i using eqn (5), and ICRAC and d[Ca2+]i/dt values were measured as described in A and B. The continuous curve is the best fit of eqns (6) and (7) (see text for parameter values). The dashed curve indicates the estimated contribution of intracellular indo-1 alone to cytosolic buffering and effective cell volume (calculated after setting cyto to 0 in eqn (6)).

To estimate the buffering capacity of the cell over the entire range of [Ca²⁺]_i values shown in Fig. 9C, we made the assumption that the total buffering capacity, , includes a component from endogenous cytoplasmic buffers (_cyto), and a component due to indo-1 (_indo):

= (_cyto + _indo)v, (6)

and that _cyto is constant (Neher, 1995). Buffering by indo-1 declines with increasing [Ca²⁺]_i according to:

(7)

where [indo-1]_TOT is the total concentration of intracellular indo-1 (free + bound), and K_D is its dissociation constant for Ca²⁺. Equations (6) and (7) were combined and fitted to the experimental data shown in Fig. 10C. The following parameter values provided the best fit: [indo-1] = 111 µM; K_D = 123 nM; v = 2 pl; _cyto = 80. While these values are actually quite reasonable (see Discussion), it should be noted that given the large number of free parameters, the fit is not unique; it merely provides an empirical description of as a function of [Ca²⁺]_i, which is necessary to extract Ca²⁺ flux rates from the measured values of d[Ca²⁺]_i/dt. It is noteworthy that intracellular indo-1 is expected to dominate the cytosolic buffering at [Ca²⁺]_i 0.4 µM (Fig. 10C, dashed curve), which has profound effects on the measurement of PMCA flux rates near the K_M (see below).

For the cell shown in Fig. 9 and Fig. 10, true PMCA flux rates were obtained by multiplying the measured d[Ca²⁺]_i/dt values during Ca²⁺ clearance by in a modified form of eqn (2):

(8)

Figure 11 displays the calculated pump flux as a function of [Ca²⁺]_i before and after PMCA modulation. From the best fits of a Michaelis-Menten model to each data set, we conclude that PMCA modulation increases V_max from 2.8 10⁶ to 4 10⁶ ions s^-1 and reduces K_M from 303 to 140 nM, with little change in the apparent cooperativity of extrusion (n_H from 1.8 to 2.1). A comparison of the curves with and without the correction for buffering (Fig. 11 and Fig. 9) illustrates dramatically the need to consider the effects of Ca²⁺ buffers in conducting this type of analysis.

		View larger version [in this window] [in a new window]
Figure 11. Modulation increases the Ca2+ sensitivity and dynamic range of PMCA activity PMCA velocity (VPMCA) as a function of [Ca2+]i is shown for the unmodulated () and modulated () conditions for the cell used for Figs 9 and 10. VPMCA at each level of [Ca2+]i was determined by scaling the d[Ca2+]i/dt measurements (Fig. 9C) by (Fig. 10C), as described in eqn (8). Fits of the Michaelis-Menten relationship are superimposed (parameter values given in the text).

Role of PMCA modulation in shaping Ca²⁺ responses

A central question is to what extent PMCA modulation helps to shape Ca²⁺ responses. The net flux of Ca²⁺ into or out of the cell combined with the effective cellular buffering determines the rate at which [Ca²⁺]_i changes. Thus, if [Ca²⁺]_i, the buffering capacity and all Ca²⁺ fluxes except the PMCA can be measured, it is possible to estimate the PMCA flux and hence its contribution to cellular Ca²⁺ dynamics. As shown earlier, in Jurkat cells treated with TG and antimycin A1 + oligomycin, the only source of Ca²⁺ is via I_CRAC and the only sink is the PMCA. Hence, the rate at which [Ca²⁺]_i changes is described by:

(9)

Eqn (9) can be solved for J_PMCA:

(10)

We applied this relationship to determine the contribution of the PMCA to the biphasic Ca²⁺ response shown in Fig. 2B. The [Ca²⁺]_i response and underlying constant I_CRAC from that experiment are reproduced in Fig. 12A and B. For each time point, the values of I_CRAC (Fig. 12B), d[Ca²⁺]_i/dt (from Fig. 12A), and as a function of [Ca²⁺]_i (average values from three cells) were substituted into eqn (10) to estimate the PMCA flux (Fig. 12C). It can be seen from this analysis that following the initial [Ca²⁺]_i rise in Fig. 12A, PMCA activity slowly increases until efflux is equal and opposite to influx, and consequently [Ca²⁺]_i reaches a maximum peak level. After this point, PMCA activity continues to rise, presumably due to the slow kinetics of modulation, causing efflux to exceed influx and [Ca²⁺]_i to drop. Even though [Ca²⁺]_i is falling, PMCA activity continues to rise for some time, although it eventually declines to a new steady state, at which it balances influx through CRAC channels, and a [Ca²⁺]_i plateau is attained.

		View larger version [in this window] [in a new window]
Figure 12. PMCA modulation enhances the dynamics and stability of signals produced by store-operated CRAC channels Data in A and B are reproduced from Fig. 2B, and predicted curves in C-E were calculated as described in the text. A and B, an overshooting Ca2+ response was evoked by a step increase in ICRAC. C, delayed PMCA modulation enhances the dynamics of [Ca2+]i signals. The calculated PMCA flux rises more slowly than the CRAC flux, until the two are equal at the peak of the [Ca2+]i response. Further increase in JPMCA causes [Ca2+]i to decline until JPMCA and JCRAC are equal and a plateau is attained. D, a comparison of JPMCA with the estimated flux in the absence of modulation (dotted line). The unmodulated PMCA flux was derived from eqn (1), with KM = 0.38 µM, nH = 1.8 and Vmax = 107 ions s-1. The comparison shows that modulation is responsible for the overshoot and boosts PMCA activity to match influx through CRAC channels and stabilize [Ca2+]i. E, PMCA modulation enhances the stability of [Ca2+]i signals. In the absence of PMCA modulation, constant ICRAC is predicted to cause [Ca2+]i to rise indefinitely because influx through CRAC channels exceeds the efflux afforded by the unmodulated pump (D).

To what extent are these slow changes in PMCA activity due to modulation? We addressed this question by using the Michaelis-Menten description of the unmodulated pump to predict the activity of the PMCA in the absence of modulation (shown by the dotted line in Fig. 12D). For this purpose, the [Ca²⁺]_i values shown in Fig. 12A were substituted into eqn (1), using average values for the Michaelis-Menten parameters obtained from three cells, and V_max was chosen to match the unmodulated pump flux to the total pump flux at early times, when the extent of modulation is low (see Fig. 12 legend). It can be seen that the predicted unmodulated PMCA flux reaches its maximum before [Ca²⁺]_i peaks, and is nearly constant for the remainder of the experiment; this saturation behaviour is in fact expected at [Ca²⁺]_i > 500 nM, given the estimated K_M of 300 nM. It is also apparent from Fig. 12D that the kinetics and amplitude of PMCA activity are largely a reflection of the modulation process.

To what extent does modulation of the PMCA contribute to stability of Ca²⁺ signals in the T cell? To address this question, we calculated the predicted time course of [Ca²⁺]_i in response to I_CRAC in the absence of modulation by substiuting the unmodulated pump flux for J_PMCA in eqn (9). The results of this simulation are shown in Fig. 12E. In the absence of modulation, [Ca²⁺]_i would continue to increase more or less monotonically to micromolar levels because J_CRAC is constant and exceeds J_PMCA, which is nearly constant due to saturation of the transport sites with Ca²⁺.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

In the study presented here, we provide the first evidence that the PMCA enhances the dynamics and stability of Ca²⁺ signals in Jurkat T cells. Following a rise in [Ca²⁺]_i, activation of extrusion via the PMCA is biphasic, consisting of a rapid increase followed by a slow enhancement or modulation. The initial activation of PMCA results from the direct binding of Ca²⁺ to transport sites. As with any enzyme and substrate, this binding confers a calcium dependence on the PMCA extrusion rate. The slow modulation of PMCA-mediated extrusion is a distinct process that alters both the Ca²⁺ sensitivity and the maximal rate of the PMCA, thereby increasing the extrusion rate at any given value of [Ca²⁺]_i. In addition to its slow development, PMCA modulation persists for several minutes after [Ca²⁺]_i returns to baseline. Thus, PMCA modulation acts as a high-pass filter with memory, shaping the Ca²⁺ signals generated through stimulation of the TCR. As the primary Ca²⁺ clearance mechanism, PMCA activity and modulation play a major role in determining the spatiotemporal profile of Ca²⁺ signals in T cells.

Mechanisms of Ca²⁺ clearance in T cells

Ca²⁺ clearance following a rise in [Ca²⁺]_i in T cells is quite complex, including contributions from the PMCA, mitochondria and SERCA. PMCA is the sole clearance mechanism at lower [Ca²⁺]_i values. PMCA and mitochondria primarily control Ca²⁺ recovery when [Ca²⁺]_i is high, with a small contribution from SERCA. Surprisingly, SERCA activity does not accelerate Ca²⁺ clearance in T cells. This contrasts with findings in many excitable cells, where SERCA-mediated refilling of Ca²⁺ stores is a major contributor to Ca²⁺ recovery. This difference may be attributable to the modest capacity of the ER Ca²⁺ store in T cells.

Contrary to expectations, SERCA activity actually delays the return of [Ca²⁺]_i to baseline following a rise in [Ca²⁺]_i. The slowing of Ca²⁺ recovery is likely to result from the shuttling of Ca²⁺ between the ER and mitochondria. In the absence of a functioning SERCA, mitochondria have been shown to delay the return of [Ca²⁺]_i to baseline by slowly releasing sequestered Ca²⁺ (Hoth et al. 1997). Our observations suggest that under more physiological conditions, mitochondria interact with the ER, further slowing Ca²⁺ recovery.

A close proximal coupling of ER and mitochondria has been demonstrated to occur in several other cell types, which increases uptake of Ca²⁺ by mitochondria after its release through IP₃ receptors in the ER (Rizzuto et al. 1993; Hajnoczky et al. 1995). In addition, Ca²⁺ fluxes in the reverse direction, from the mitochondria to the ER, have also been inferred in some cells (Arnaudeau et al. 2001). Thus, one possible explanation for the SERCA-mediated delay in Ca²⁺ recovery, and the small [Ca²⁺]_i oscillations observed in some cells, is that Ca²⁺ that is released from the mitochondria is taken up by the ER, then released through IP₃ receptors in an oscillatory manner. This would predict possibly multiple rounds of shuttling of Ca²⁺ between these two stores until the PMCA transports the Ca²⁺ out of the cell. Taken together with evidence that mitochondria and CRAC channels also interact closely (Hoth et al. 1997, 2000), and the existence of a specialized Ca²⁺ store near CRAC channels that is coupled to their activation (Parekh et al. 1997; Huang & Putney, 1998; Broad et al. 1999), the emerging picture is one in which the ER, mitochondria and CRAC channels exist close to each other in a tightly regulated Ca²⁺ signalling network.

To examine directly the role played by PMCA in Ca²⁺ regulation, we measured Ca²⁺ recovery under conditions of ATP depletion or in the presence of the PMCA inhibitors La³⁺ and carboxyeosin. The variability of Ca²⁺ extrusion rates in ATP-depleted cells may be explained by varying degrees of ATP depletion. We do not believe it implies multiple Ca²⁺ clearance mechanisms, given that inhibition by La³⁺ and carboxyeosin was much more consistent from cell to cell. The complete block of clearance in some ATP-depleted cells, and the high degree of inhibition by La³⁺ further support the idea that the PMCA is the sole clearance mechanism. It should be noted that while inhibition by carboxyeosin was less complete, there is a precedent for incomplete block of PMCA in intact cells (Monteith et al. 1998). Although it is difficult to rule out absolutely an additional clearance mechanism, our results imply that it must be ATP dependent and cannot be the Na⁺-Ca²⁺ exchanger. Consistent with other studies in Jurkat T cells and the human T cell clone P28, we find no evidence of a functional Na⁺-Ca²⁺ exchanger (Donnadieu et al. 1992; Donnadieu & Trautmann, 1993). In vascular endothelial cells, both PMCA and Na⁺-Ca²⁺ exchange contribute to clearance, with the activity of one system compensating for changes in extrusion in the other (Sedova & Blatter, 1999). We could not detect any Na⁺-dependent Ca²⁺ extrusion in Jurkat T cells, even in the presence of PMCA inhibitors. Thus, it seems most likely that the incomplete inhibition of Ca²⁺ clearance by carboxyeosin results from the imperfect block of PMCA. In summary, our findings support the conclusion that the PMCA is the dominant Ca²⁺ extrusion mechanism in T cells.

PMCA modulation

Modulation augments PMCA activity by increasing both the Ca²⁺ sensitivity and maximal extrusion rate. To ascertain the Michaelis-Menten kinetic parameters of PMCA, we measured extrusion at a variety of different values of [Ca²⁺]_i, before and after modulation. Estimates of the Ca²⁺ fluxes through PMCA were derived from the rate of change of [Ca²⁺]_i, upon removal of [Ca²⁺]_o. A similar approach was used in endothelial cells to determine the Ca²⁺ dependence of PMCA activity and modulation, although the effects of intracellular buffering by the Ca²⁺ indicator were not considered (Sedova & Blatter, 1999). We have seen that because the K_D of indo-1 is near the resting [Ca²⁺]_i, the total buffering capacity of the cell changes markedly in the physiological range of [Ca²⁺]_i values (see Fig. 10). This in turn makes d[Ca²⁺]_i/dt a highly non-linear function of the Ca²⁺ flux rate, especially at [Ca²⁺]_i values near the K_D. The end result is that if buffering is not taken into account, fits of the Michaelis-Menten equation to the data grossly underestimate both the Ca²⁺ affinity (K_M) and maximal extrusion rates (V_max) of the PMCA.

The calcium-binding capacity of endogenous buffers and the exogenous buffer indo-1, as well as the cell volume, determines the effective cell volume, (eqn (6)). In fitting eqns (5), (6) and (7) to the data given in Fig. 10, we obtained fairly reasonable values for the underlying components of , including the endogenous buffering capacity, cell volume, and intracellular indo-1 concentration. The estimated cytosolic buffering capacity of ~80 in Jurkat T cells is similar to the buffering capacity of human T cells of 125 measured using a different method (Donnadieu et al. 1992). The estimate of cell volume (2 pl) is close to the value estimated from cell capacitance, and the estimated [indo-1] is within the range we found in measurements from other Jurkat cells. It is important to note that the small number of data points we could obtain from single cells limits our ability to constrain the parameters in fitting vs. [Ca²⁺]_i. However, regardless of the accuracy of determining these values, the fit to the data provides the necessary empirical description of how the effective buffering capacity varies with [Ca²⁺]_i.

We obtained true PMCA extrusion rates by applying to the plots of -d[Ca²⁺]_i/dt vs. [Ca²⁺]_i, according to eqn (8). The corrected velocity-substrate curves illustrate that the Ca²⁺ dependence of PMCA follows Michaelis-Menten enzyme kinetics, both before and after modulation. PMCA modulation increases V_max and decreases K_M. The small number of data points of the unmodulated PMCA activity limits the accuracy of the V_max and K_M values. Experimental difficulties limit our ability to measure the activity of the unmodulated PMCA over a large range of [Ca²⁺]_i values in single cells. First, in order to limit modulation, measurements must be made at short times, and the small size of I_CRAC limits the range of [Ca²⁺]_i values that can be attained. Second, full recovery from modulation takes ~10 min, limiting the number of Ca²⁺ transients that can be generated before the cell dies ( five per cell). The unmodulated extrusion rate vs. [Ca²⁺]_i data could not be well fitted if either V_max or K_M were constrained to be the values found for the modulated PMCA. Thus, modulation enhances both the affinity and the maximal rate of the PMCA.

Possible mechanisms of PMCA modulation

Several mechanisms could in principle contribute to the increase in PMCA-mediated extrusion upon Ca²⁺ influx. Insertion of new PMCA molecules through vesicle fusion would be expected to increase V_max, but not necessarily reduce K_M. Likewise, phosphorylation by protein kinase C and/or protein kinase A has been reported to increase PMCA activity primarily through an increase in V_max (Monteith & Roufogalis, 1995). Our findings of a decreased K_M in the modulated state are not easily reconciled with either of these two mechanisms. However, calmodulin (CaM) is known to increase the activity of PMCA in vitro by binding to and displacing an autoinhibitory region of the C-terminal tail (Carafoli, 1994; Caride et al. 1999). Several characteristics of CaM-induced PMCA activation are consistent with the modulation we have described in T cells. Calcium-CaM causes an increase in V_max and decrease in K_M, which occur with a time course roughly consistent with our observations in vivo, and acts upon PMCA 4b, the isoform most highly expressed in Jurkat T cells (Caride et al. 2001). Thus, it seems most likely that of the currently known mechanisms of PMCA modulation, the CaM binding mechanism is responsible. Additional experiments will be needed to distinguish between these possibilities.

The physiological impact of PMCA modulation

Delayed modulation of PMCA enhances the stability of Ca²⁺ signals in T cells. The unmodulated and modulated velocity-substrate curves saturate at ~400 nM (Fig. 11). At first glance it seems surprising that maximal transport would occur at such a low [Ca²⁺]_i because saturation would block further increases in Ca²⁺ extrusion, rendering the cell defenceless to additional Ca²⁺ influx (as shown in Fig. 12). In intact cells, several different mechanisms may limit the magnitude of Ca²⁺ responses, including calcium-dependent inactivation of I_CRAC, mitochondrial Ca²⁺ uptake and elevation of PMCA expression or other Ca²⁺ clearance pathways. While these mechanisms could in principle help prevent toxic Ca²⁺ loads, they would also limit the maximum magnitude of Ca²⁺ signals and would suppress brief events such as oscillatory Ca²⁺ spikes and transients. However, Ca²⁺ influx not only stimulates extrusion, but also increases PMCA activity to respond dynamically to a higher Ca²⁺ load. Thus, modulation stabilizes Ca²⁺ signals by matching the amount of efflux to that of influx through I_CRAC, such that a stable [Ca²⁺]_i plateau is attained and toxic levels of [Ca²⁺]_i are avoided.

In addition to ensuring the stability of Ca²⁺ signals, delayed modulation of PMCA also contributes to the dynamics of Ca²⁺ signalling in T cells. The key to understanding this paradoxical dual function is that modulation is delayed. The slow kinetics of modulation onset and reversal help determine the magnitude and duration of Ca²⁺ spikes. A fuller understanding of how modulation contributes to Ca²⁺ signalling will require quantification of its dependence upon [Ca²⁺]_i and time. Thus far, the time course of PMCA modulation has been measured only under conditions where [Ca²⁺]_i is changing with time (Fig. 7); to achieve a more quantitative description of the process, rapid and stable step changes in [Ca²⁺]_i will be needed. Nevertheless, the results of the modelling shown in Fig. 12 show that the slow time course of pump modulation and its reversal play a major role in producing the initial overshoot of the biphasic [Ca²⁺]_i response, in effect enabling the pump to act as a high-pass filter with memory. This mechanism enables protection against large sustained increases in [Ca²⁺]_i, while maximizing the effect of small I_CRAC and transient release events. PMCA modulation has been shown to contribute to overshooting [Ca²⁺]_i responses in erythrocytes, neutrophils and endothelial cells (Scharff et al. 1983; Scharff & Foder, 1994; Madge et al. 1997; Klishin et al. 1998; Snitsarev & Taylor, 1999) and is therefore likely to serve similar functions in these and other cells that express PMCA 4b. Recent work suggests that the kinetics of PMCA modulation plays a broad role in tuning Ca²⁺ response characteristics for specific cell functions. Thus, cells engaged in rapid signalling, such as hair cells and cardiac cells, express an isoform with rapid modulation (PMCA 2a), while cells that generate slower signals, such as Jurkat, express a slowly modulated isoform (PMCA 4b; Caride et al. 2001).

It should be noted that our results do not rule out additional mechanisms for shaping Ca²⁺ signals in T cells. Experimental conditions in this study (voltage clamp, inhibition of SERCA and mitochondrial Ca²⁺ uptake) were chosen to allow us to isolate the contribution of PMCA modulation to Ca²⁺ signalling. Under physiological conditions, Ca²⁺ signals can be shaped by multiple mechanisms, including transient Ca²⁺ release from intracellular stores, slow deactivation of I_CRAC due to store refilling, slow inactivation of I_CRAC, and changes in membrane potential that influence the driving force for Ca²⁺ entry. A striking example is the [Ca²⁺]_i oscillations generated by delayed feedback between stores and CRAC channels in human T cells (Dolmetsch & Lewis, 1994). Thus, while PMCA modulation on its own is sufficient to generate a biphasic response, the spatiotemporal profiles of [Ca²⁺]_i responses will be determined in general by the concerted effects of influx through Ca²⁺ channels, Ca²⁺ uptake by intracellular organelles, Ca²⁺ binding to intracellular buffers and Ca²⁺ efflux through PMCA. Nevertheless, it is important to recognize that the ability of PMCA modulation to generate large overshoots does suggest that the commonly cited release and influx mechanism for biphasic responses needs to be carefully re-evaluated in many cells, particularly those that express PMCA 4b.

It is pertinent to ask whether the contributions of PMCA modulation to Ca²⁺ dynamics may have any particular significance for signal transduction. Recent studies have expanded our understanding of how the amplitude and kinetic signature of Ca²⁺ signals are decoded by cells to trigger specific responses. Ca²⁺ spikes, oscillations, and plateaus triggered by antigen help to coordinate activation of specific transcription factors and programs of gene expression in lymphocytes. B cells respond to antigen with a biphasic [Ca²⁺]_i elevation, consisting of a rapid rise followed by a slow decline to a low plateau. The spike and plateau phases activate distinct transcriptional pathways involved in B cell activation (Dolmetsch et al. 1997). Interestingly, B cells exposed to self antigen in vivo lose the Ca²⁺ spike component specifically, and this may contribute to the development or maintenance of immunological self-tolerance (Healy et al. 1997). During T cell activation by antigen, prolonged [Ca²⁺]_i elevation is required to drive the nuclear accumulation of nuclear factor of activated T cells (NFAT), a transcription factor that is required for the expression of interleukin-2 and other activation cytokines. [Ca²⁺]_i oscillations enhance the efficiency of signalling through NFAT, and the frequency of [Ca²⁺]_i oscillations contributes to signalling specificity by differentially activating distinct sets of calcium-dependent transcription factors (Dolmetsch et al. 1998). [Ca²⁺]_i must be closely regulated to ensure signals are large enough to activate signalling pathways while simultaneously preventing a potentially harmful overload of Ca²⁺.

PMCA modulation provides the cell with a unique, highly efficient clearance mechanism that promotes the generation of Ca²⁺ transients during brief periods of CCE, while retaining the ability to rapidly remove Ca²⁺ at high [Ca²⁺]_i levels, and thereby stabilize signals during prolonged CCE. In addition, the slow kinetics of pump modulation and its reversal play a major role in producing the overshoot characteristic of the biphasic [Ca²⁺]_i rise. Thus, in addition to its homeostatic role, the PMCA actively contributes to the complexity of Ca²⁺ signalling in T cells and constitutes an important part of a complex and highly interconnected Ca²⁺ signalling machine.

	REFERENCES

Top Abstract Introduction Methods Results Discussion References

ALMERS, W. & NEHER, E. (1985). The Ca signal from fura-2 loaded mast cells depends strongly on the method of dye-loading. FEBS Letters 192, 13-18	[Medline]
ARNAUDEAU, S., KELLEY, W. L., WALSH, J. V. JR & DEMAUREX, N. (2001). Mitochondria recycle Ca2+ to the endoplasmic reticulum and prevent the depletion of neighboring endoplasmic reticulum regions. Journal of Biological Chemistry 276, 29430-29439	[Abstract/Full Text]
BABCOCK, D. F. & HILLE, B. (1998). Mitochondrial oversight of cellular Ca2+ signaling. Current Opinion in Neurobiology 8, 398-404	[Medline]
BALASUBRAMANYAM, M. & GARDNER, J. P. (1995). Protein kinase C modulates cytosolic free calcium by stimulating calcium pump activity in Jurkat T cells. Cell Calcium 18, 526-541	[Medline]
BALASUBRAMANYAM, M., KIMURA, M., AVIV, A. & GARDNER, J. P. (1993). Kinetics of calcium transport across the lymphocyte plasma membrane. American Journal of Physiology 265, C321-327	[Medline]
BALASUBRAMANYAM, M., ROHOWSKY-KOCHAN, C., REEVES, J. P. & GARDNER, J. P. (1994). Na+/Ca2+ exchange-mediated calcium entry in human lymphocytes. Journal of Clinical Investigation 94, 2002-2008	[Medline]
BAUTISTA, D. M., HOTH, M. & LEWIS, R. S. (1998). Characterization of the plasma membrane calcium pump in T cells: modulation and memory. Journal of General Physiology 111, A57
BERRIDGE, M. J. (1995). Capacitative calcium entry. Biochemical Journal 312, 1-11	[Medline]
BERRIDGE, M. J. & DUPONT, G. (1994). Spatial and temporal signalling by calcium. Current Opinion in Cell Biology 6, 267-274	[Medline]
BROAD, L. M., ARMSTRONG, D. L. & PUTNEY, J. W. JR (1999). Role of the inositol 1,4,5-trisphosphate receptor in Ca2+ feedback inhibition of calcium release-activated calcium current (Icrac). Journal of Biological Chemistry 274, 32881-32888	[Abstract/Full Text]
CARAFOLI, E. (1991). Calcium pump of the plasma membrane. Physiological Reviews 71, 129-153	[Medline]
CARAFOLI, E. (1994). Biogenesis: plasma membrane calcium ATPase: 15 years of work on the purified enzyme. FASEB Journal 8, 993-1002	[Abstract]
CARIDE, A. J., ELWESS, N. L., VERMA, A. K., FILOTEO, A. G., ENYEDI, A., BAJZER, Z. & PENNISTON, J. T. (1999). The rate of activation by calmodulin of isoform 4 of the plasma membrane Ca2+ pump is slow and is changed by alternative splicing. Journal of Biological Chemistry 274, 35227-35232	[Abstract/Full Text]
CARIDE, A. J., FILOTEO, A. G., PENHEITER, A. R., PASZTY, K., ENYEDI, A. & PENNISTON, J. T. (2001). Delayed activation of the plasma membrane calcium pump by a sudden increase in Ca2+: fast pumps reside in fast cells. Cell Calcium 30, 49-57