Introduction

Under normal physiological conditions extracellular pH (pHo) in healthy tissues is maintained within a narrow range between 7.3 and 7.4, while intracellular pH (pHi) is kept between 7.1–7.21. In exercising muscle both extracellular and intracellular pH can drop as low as 6.9 and 6.7 respectively2. In cancerous tumours and wounded tissue, pH variations from the normal values can be even more extreme3,4,5,6. Dysregulated pH is one of the hallmarks of cancer progression, whereas pH of a wound can be used as a predictor of wound healing outcomes4,5,6,7. Activity of almost ubiquitously expressed Ca2+ release activated Ca2+ (CRAC) channels, formed by Orai1 and STIM1 proteins, has been shown to strongly depend on both extracellular and intracellular pH8,9,10. Considering that CRAC channels have an important role in regulation of immune response, skeletal muscle function and in cancer progression, their dependence on pH is likely to have some physiological significance.

Extracellular acidification inhibits, whereas alkalinisation increases CRAC current (ICRAC) amplitude with pKa of about 8. This has been consistently shown in several publications using both, heterologous expression of Orai1 and STIM1 and cells expressing endogenous ICRAC 8,9,10,11. The dependence of ICRAC on extracellular pH in the presence of extracellular Ca2+ is mediated by E106 in the Orai1 pore, with some contribution from nearby D110 and D112 in the first extracellular loop8, 9. In the absence of Ca2+, Na+ permeability seems to be affected by protonation of E190 residue in the Orai1 pore10. The mechanism of ICRAC dependence on intracellular pH is less well understood. Intracellular acidification has been shown to inhibit both, endogenous ICRAC in different types of cells, and ICRAC mediated by Orai1 and STIM1 heterologously expressed in HEK293 cells8, 10, 12. However, intracellular alkalinisation strongly enhanced the amplitude of Orai1/STIM1 mediated ICRAC in some, but not all studies10, whereas the amplitude of endogenous ICRAC in RBL cells and Jurkat T lymphocytes was not affected by pHi rise from 7.4 to 8.48, 12.

Theoretically, there are two main mechanisms that may mediate the dependence of ICRAC function on pHi. Protonation/deprotonation of specific residues in Orai1 may affect conductance through the Orai1 pore, and/or protonation/deprotonation of specific residues in Orai1 and/or STIM1 may affect their interaction. The evidence obtained thus far cannot exclude either of these possibilities. Intracellular acidification has been shown to functionally uncouple STIM1 and Orai1 without causing a complete dissociation of STIM1/Orai1 complex, suggesting that pHi affects STIM1/Orai1 interaction13. Furthermore, mutation of His155 in the intracellular loop of Orai1 to phenylalanine (H155F) abolishes the effect of alkalinisation on ICRAC and diminishes ICRAC inhibition caused by acidification of pHi 10. The Orai1 region containing H155 has previously been implicated in fast Ca2+-dependent inactivation (FCDI) of ICRAC, and therefore may be involved in STIM1/Orai1 interactions14.

In this work, we investigated pHi dependence of ICRAC mediated by WT Orai1 and WT or mutated STIM1 ectopically expressed in HEK293 cells at two Orai1:STIM1 expression ratios. Using cells expressing WT Orai1 and WT STIM1 we confirmed that ICRAC amplitude strongly depends on pHi and showed that intracellular acidification introduces a strong re-activation component in ICRAC traces recorded in response to voltage steps between −80 mV and −140 mV. (In this study, term “re-activation” is used exclusively in relation to slow increase of ICRAC amplitude during voltage steps from 0 mV to potentials between −80 and −140 mV, and is opposite of FCDI.) As shown previously, this ICRAC re-activation could also be observed at normal pHi, but only in the cells that were transfected with higher amounts of Orai1 cDNA, relative to STIM19, 15, 16. To investigate whether there is any overlap between the mechanisms that regulate dependence of ICRAC on pHi and pHo, we used E106D Orai1 mutant. Glutamate 106 in the selectivity centre of Orai1 pore was previously shown to mediate ICRAC dependence on pHo 9. In further search for the potential protonation sites that may be responsible for ICRAC dependence on pHi, we evaluated EE482/483AA and DD475/476AA double mutations within STIM1 inactivation domain (IDSTIM). Considering that IDSTIM is highly negatively charged and is indispensable for Ca2+ dependent inactivation of ICRAC 17, 18, it is logical to hypothesise that it is involved in pHi sensitivity of CRAC channel.

Results

Intracellular pH affects CRAC channel gating

To investigate ICRAC dependence on pHi, the currents were recorded using pipette solutions with pHi adjusted to 6.3, 7.3 or 8.3. Previously it has been shown that the current amplitude, fast Ca2+ dependent inactivation (FCDI), re-activation, potentiation by 2-APB, and selectivity of CRAC channels for divalent cations strongly depend on the relative amounts of Orai1 and STIM1 proteins in the cell15, 16. It is possible that other properties of ICRAC, including pH dependence, are also influenced by the Orai1:STIM1 expression ratios. Therefore, we investigated the effects of pHi on ICRAC at two transfection conditions. To achieve different expression ratios, HEK293T cells were transfected with Orai1- and STIM1-containing plasmids at either 1:4 or 1:1 molar ratios. For both transfection conditions, the amplitude of ICRAC exhibited strong dependence on pHi. ICRAC was smaller at pHi 6.3 and larger at pHi 8.3, compared to pHi 7.3 (Fig. 1a). Consistent with previous publications, at pHi 7.3 and 6.3 cells transfected with higher relative amount of STIM1 (1 Orai1: 4 STIM1 ratio) produced larger ICRAC, compared to cells transfected with equal amounts of STIM1 and Orai1 (1 Orai1: 1 STIM1 ratio; Fig. 1a). However, the effect of the relative expression ratio on the current amplitude, at least within the employed transfection range (1:4 and 1:1), was absent when pHi was raised to 8.3 (Fig. 1a). Majority of cells at pHi 7.3 produced ICRAC with noticeable FCDI at potentials between −80 and −140 mV when transfected with 1 Orai1: 4 STIM1 ratio (Fig. 1b,i). Some re-activation was also evident with longer pulses (Fig. 1b,ii). It was observed that raising pHi to 8.3 resulted in elimination of visible signs of the re-activation component, even with longer pulses. However, the extent of FCDI of the currents recorded in response to 200 ms pulses at pHi 8.3 was also reduced, compared to pHi 7.3, and the time course of inactivation was significantly slower (c.f. Fig. 1ci and bi). In contrast, lowering pHi to 6.3 produced ICRAC with pronounced re-activation at potentials between −80 and −140 mV and no visible FCDI (Fig. 1d).

Figure 1
figure 1

The dependence of ICRAC amplitude and kinetics on pHi. (a) Current density at −100 mV, obtained from instantaneous I–V plots in response to 100 ms ramps between −120 and 120 mV recorded after a complete development of ICRAC, at indicated pHi and Orai1:STIM1 transfection ratios. The amplitude of ICRAC at pHi 7.3 was significantly different from the amplitudes at pHi 6.3 and 7.3 at both transfection ratios (P < 0.001; unpaired t-test). The amplitudes of ICRAC at 1Orai1:1STIM1 transfection ratio was significantly smaller that the amplitudes at 1Orai1:4STIM1 ratio at pHi 6.3 and 7.3 (P < 0.03; unpaired t-test), but not at pHi 8.3. (b,c and d) The examples of WT ICRAC traces at pHi 7.3, 8.3 and 6.3, correspondingly. Currents were recorded in HEK293T cells transfected with Orai1 and STIM1 at 1:4 ratio, in response to 200 ms (i) and 600 ms (ii) voltage steps from 0 mV holding potential to −120 mV.

Full size image

To compare ICRAC Ca2+ dependent gating (FCDI and re-activation) under different conditions and a range of membrane potentials, we used the amplitudes of tail-currents obtained at −100 mV after voltage steps between −140 and +80 mV, normalised to the amplitude of the tail current after a step to +80 mV (see Methods)9. The resulting data were used to construct apparent P o curves9. ICRAC that exhibited FCDI and little or no re-activation produced apparent P o data that could be fitted with a standard Boltzmann equation (eq. 1), whereas ICRAC with pronounced re-activation exhibited bell-shaped P o curves which could not be fitted with a single Boltzmann function (Fig. 2a). At the 1Orai1:4STIM1 transfection ratio FCDI was more pronounced at pHi 7.3 than at pHi 8.3. At pHi 6.3, the P o curve was bell-shaped with a maximum at −20 mV, which was expected, considering the presence of re-activation. However, despite the apparent absence of FCDI in current traces recorded at pHi 6.3 (Fig. 1d), the FCDI was still present, and the extent of it, relative to the maximum P o, was similar to that of ICRAC recorded at pHi 7.3 (Fig. 2a).

Figure 2
figure 2

The dependence of WT ICRAC apparent P o on pHi. (a) Apparent P o curves were obtained using tail currents as described in Methods. HEK293T cells were transfected with Orai1 and STIM1 at 1:4 ratio and ICRAC was recorded at pHi 6.3 (clear triangles, n = 6), 7.3 (filled circles, n = 16) and 8.3 (clear circles, n = 7). (b) The comparison of P o curves obtained at Orai1:STIM1 transfection ratios 1:1 (filled circles, n = 9) and 1:4 (clear circles, n = 16) at pHi 7.3. (c) Apparent P o curves obtained from tail currents recorded in cells transfected with Orai1 and STIM1 at 1:1 ratio. ICRAC was recorded at pHi 6.3 (clear triangles, n = 4), 7.3 (filled circles, n = 9) and 8.3 (clear circles, n = 8). Data on panels (a,b and c) were obtained with pipette solution containing EGTA. (d) The effect of BAPTA on P o. Apparent P o curves were obtained using cells transfected with Orai1 and STIM1 at 1:4 transfection ratio and BAPTA in the pipette solution at pHi 7.3 (clear circles) and 6.3 (clear squares). For comparison, P o curves obtained using cells transfected with Orai1 and STIM1 at either 1:4 (clear triangles) or 1:1 (filled circles) transfection ratio and EGTA in the pipette solution at pHi 8.3 are shown.

Full size image

Larger ICRAC amplitude at alkaline pHi is due to pH dependence of EGTA

Phenomenologically, the effect of intracellular acidification on the ICRAC kinetics and P o (Figs 1 and 2) was similar to the effect of increasing Orai1 expression relative to STIM116. ICRAC recorded at pHi 7.3 in the cells transfected with 1Orai1:1STIM1 ratio showed strong re-activation during voltage steps from 0 mV to −120 mV and produced bell-shaped P o curve, which looked similar to the P o curve obtained at pHi 6.3 with 1 Orai1: 4 STIM1 transfection ratio (c.f. Fig. 2a and b). Lowering pHi to 6.3 in cells transfected with 1 Orai1: 1 STIM1 ratio further increased the re-activation (Fig. 2c). In contrast, rising pHi to 8.3 virtually eliminated current re-activation (Fig. 2c). The apparent P o curves obtained at pHi 8.3 in cells transfected with 1:1 and 1:4 Orai1:STIM1 ratios were almost identical between two transfection conditions (Fig. 2d). The observed changes in the kinetics and the extent of ICRAC FCDI induced by raising pHi to 8.3 (Fig. 1c) are similar to those caused by replacing EGTA with BAPTA at pHi 7.39, 19, 20. Due to its’ ability to bind Ca2+ faster than EGTA, BAPTA is believed to reduce Ca2+ concentration at the intracellular mouth of CRAC channels, thus slowing down and reducing FCDI19, 20. The apparent P o curve obtained at pHi 7.3 using cells transfected with 1 Orai1: 4 STIM1 ratio and BAPTA as Ca2+ buffer, was virtually identical to P o curves obtained at pHi 8.3 and EGTA in the pipette solution (Fig. 2d). Using BAPTA in the internal solution instead of EGTA with pHi 6.3 also decreased ICRAC re-activation at negative potentials and therefore reduced positive apparent P o (Fig. 2d).

To investigate whether intracellular Ca2+ buffer contributes to the dependence of ICRAC on pHi, we used extracellular application of 30 mM NH4Cl, which is known to alkalinise pHi 13, 21. Application of NH4Cl to the bath, when EGTA was used as Ca2+ buffer in the pipette solution, drastically increased the ICRAC amplitude (Fig. 3a,b) and caused inhibition of both ICRAC FCDI and re-activation (Fig. 3c), in agreement with the results obtained using pipette solution with EGTA and pHi 8.3 (Figs 1a and 2a,c). In contrast, application of NH4Cl when BAPTA was used in the pipette solution instead of EGTA, had very little effect on ICRAC amplitude (Fig. 3a,b).

Figure 3
figure 3

ICRAC potentiation by intracellular alkalinisation depends on intracellular Ca2+ buffer. (a,b) The effect of 30 mM NH4Cl application to the bath on the amplitude of ICRAC. Each point on panel a represents ICRAC amplitude at −100 mV obtained from 100 ms voltage ramps between −120 and 120 mV, applied every 2 s. WT ICRAC was recorded in cells transfected with Orai1 and STIM1 at 1:4 ratio using either 10 mM BAPTA (clear circles) or 10 mM EGTA (filled circles) in the pipette solution (n = 5 for each condition), pHi 7.3. (c) Apparent P o curves obtained before (filled circles) and after (clear circles) application of NH4Cl to the bath (n = 5).

Full size image

Is there any overlap between mechanisms regulating ICRAC dependence on pHo and pHi?

Previous investigations have shown that the amplitude of native ICRAC in different cell types and ICRAC mediated by heterologously expressed Orai1 and STIM1 strongly depends on extracellular pH8,9,10, 12. Superficially, the dependence of ICRAC amplitude on pHo looks similar to its dependence on pHi 8,9,10, 12. However, possible reasons for similarities between pHi and pHo effects on ICRAC have not been yet considered. Could changing pHo affect pHi in patch clamping experiments? To investigate this question, we used cells transfected with Orai1 and STIM1 at 1:1 molar ratio, which showed a very pronounced re-activation at negative potentials (Fig. 4a). Raising pHi to 8.3 eliminates ICRAC re-activation at negative potentials (Fig. 2b), and if raising pHo results in a rise of pHi, one would expect a reduction of current re-activation. The results show that increasing pHo from 7.4 to 8.3 does not reduce ICRAC re-activation and has no effect on the P o curve (Fig. 4). Therefore, it can be safely concluded that pHi in these patch clamping experiments is not affected by changes in pHo.

Figure 4
figure 4

Extracellular alkalinisation has no effect on ICRAC kinetics. (a,b) Examples of ICRAC traces recorded in response to −120 mV steps in the bath solution of pHo 7.4 and after changing pHo to 8.3, correspondingly. (c) Apparent P o curves at pHo 7.4 (filled circles) and after changing pHo to 8.3 (clear circles) (n = 4). HEK293T cells were transfected with Orai1 and STIM1 plasmids at 1:1 molar ratio. pH of the pipette solution was 7.3.

Full size image

One of the main residues responsible for ICRAC dependence on pHo is Glu 106 in the Orai1 pore9. Could pHi affect protonation of Glu 106 in the Orai1 pore? Despite the observation that ICRAC kinetics is unaffected by pHo, it is possible that ICRAC amplitude dependence on pHi and pHo is mediated by the same protonatable site in the pore. To investigate this possibility, we used an E106D Orai1 mutant. Previous studies have shown that the E106D Orai1 differs from WT Orai1 in several respects9. Firstly, it is less selective for Ca2+ and supports a significant Na+ conductance. Secondly, while E106D-mediated ICRAC exhibits strong inactivation at negative potentials that looks similar to FCDI of WT ICRAC (Fig. 5a, cf. Fig. 1b), it has a different underlying mechanism. The inactivation of E106D-mediated ICRAC during steps to negative potentials is caused by Ca2+ block of Na+ permeation through the pore; it does not require interaction with IDSTIM, and it is not affected by BAPTA or Orai1:STIM1 transfection ratios9. Finally, and importantly for this investigation, the Ca2+ dependent block of Na+ permeation through the E106D pore is strongly pHo dependent, whereas the peak amplitude of ICRAC mediated by E106D Orai1 is not influenced by pHo 9. Changing the pipette solution pH revealed that the amplitude of E106D-mediated ICRAC was pHi-dependent – the current was strongly inhibited by pHi 6.3 and enhanced by pHi 8.3, similarly to WT ICRAC (Fig. 5c, cf. Fig. 1a). E106D-mediated ICRAC recorded in the absence of Na+ in the bath solution, when Ca2+ was the only permeating cation, also showed pHi dependence of the amplitude similar to that of WT ICRAC (Fig. 5d). However, the kinetics and the extent of Ca2+ dependent block of Na+ permeation through E106D Orai1 was not appreciably affected by pHi (Fig. 5b, cf. Figs 5a and 1d). If Asp 106 could be protonated from the intracellular side at low pHi, one would expect the changes in E106D-mediated ICRAC to be similar to those induced by low pHo 9, which was not the case. These results demonstrate that pHi and pHo affect ICRAC through different mechanisms, and that Glu 106 which is located in the Orai1 pore does not mediate the pHi-dependence of ICRAC amplitude.

Figure 5
figure 5

pHi dependence of E106D-mediated ICRAC. (a,b) E106D Orai1 current traces were recorded in response to voltage steps ranging from −120 mV to 80 mV, in 20 mV increments, after a pre-pulse to 80 mV at pHi 7.3, and pHi 6.3, correspondingly. (c,d) The dependence of E106D Orai1 mediated ICRAC amplitude on pHi in the control bath solution (c), and in a bath solution with 140 mM NaCl replaced with 140 mM NMDGCl (d). HEK293T cells were transfected with E106D Orai1 and STIM1 plasmids at 1:4 molar ratio.

Full size image

The effects of mutations in STIM inactivation domain on ICRAC dependence on pHi

One of the domains within STIM1/Orai1 complex critically important for FCDI is located on STIM1 between residues 470 and 491 (IDSTIM), C-terminal to CRAC activation domain (CAD)17, 18. Thus, neutralisation of Aspartate and Glutamate residues within a cluster of 7 negatively charged amino acids (475DDVDDMDEE483) in IDSTIM results in drastic changes in ICRAC FCDI18. It is possible that protonation/deprotonation of some of these residues contribute to ICRAC dependence on pHi. To investigate this possibility, we investigated pHi dependence of two double mutants of STIM1, DD475/476AA, which produced ICRAC with diminished FCDI (Fig. 6a,c), and EE482/483AA, which produced ICRAC with enhanced FCDI (Fig. 6b,c)17, 18. Using pipette solutions with pH adjusted to 6.3, 7.3 or 8.3 we found that the amplitude of ICRAC mediated by each of these STIM1 mutants co-expressed with WT Orai1 exhibited dependence on pHi similar to that of WT ICRAC (Fig. 6d; cf. Fig. 1a).

Figure 6
figure 6

The dependence of ICRAC amplitude, mediated by Orai1 and DD475/6AA and EE482/3AA STIM1 mutants, on pHi. (a,b) Examples of ICRAC traces recorded in response to 200 ms voltage steps from 0 mV to −120 mV in cells transfected with Orai1 and either DD475/6AA STIM1 (a) or EE482/3AA STIM1 (b) at pHi 7.3. (c) Apparent P o curves obtained at pHi 7.3 using cells transfected with Orai1 and DD475/6AA STIM1 (filled circles), Orai1 and WT STIM1 (clear circles), and Orai1 and EE482/3AA STIM1 (clear triangles), at 1:4 ratio. (d) IDSTIM mutants ICRAC amplitude was measured at −100 mV from the responses to 100 ms voltage ramps from −120 to 120 mV, at indicated pHi. HEK293T cells were transfected with Orai1 and DD475/6AA STIM1 or EE482/3AA STIM1 plasmids at 1:4 ratio.

Full size image

Next, we investigated the effects of DD475/476AA and EE482/483AA STIM1 mutations on the dependence of ICRAC kinetics on pHi. At the transfection ratio of 1 Orai1: 4 STIM1 the apparent P o for ICRAC mediated by DD475/476AA-STIM1 mutant showed a weaker dependence on pHi, compared to WT ICRAC (Fig. 7a; cf. Fig. 2a). Although pHi 8.3 reduced the re-activation component (Fig. 7a), as it did in WT ICRAC (Fig. 2a), pHi 6.2 failed to induce a significant change in the apparent P o of the Orai1/ DD475/476AA-STIM1 mediated current (Fig. 7a). Changing the transfection ratio to 1 Orai1: 1 DD475/476AA-STIM1 did not affect the apparent P o, or its dependence on pHi (Fig. 7b). However, we were unable to obtain the apparent P o curve at pHi 6.3 as the amplitude of the current was too small for a reliable extraction of the data.

Figure 7
figure 7

pHi has no effects on the apparent P o of ICRAC mediated by Orai1 and DD475/6AA and EE482/3AA STIM1. (a,b) Apparent P o curves obtained using cells transfected with Orai1 and DD475/6AA STIM1 at either 1:4 (a) or 1:1 (b) ratio and pHi 6.3 (clear triangles), 7.3 (filled circles) and 8.3 (clear circles). (c,d) Apparent P o curves obtained using cells transfected with Orai1 and EE482/3AA STIM1 at either 1:4 (a) or 1:1 (b) ratio and pHi 6.3 (clear triangles), 7.3 (filled circles) and 8.3 (clear circles).

Full size image

ICRAC mediated by Orai1 and EE482/483AA-STIM1 mutant also exhibited a weaker dependence of the kinetics on pHi, compared to WT CRAC (Fig. 7c,d). At the transfection ratio 1 Orai1: 4 EE482/483AA-STIM1, lowering pHi to 6.3 introduced a small re-activation component to the current (Fig. 7c). This can be seen on Po curve as deviation from simple Boltzmann distribution, whereas increasing pHi to 8.3 slightly reduced the extent of FCDI at negative potentials (Fig. 7c). At the transfection ratio 1:1, the changes in FCDI and re-activation induced by changes in pHi were more pronounced than at the ratio 1:4 (Fig. 7d, cf. Fig. 7c), however, these changes were significantly smaller than those induced by pHi changes in the WT ICRAC (Fig. 2a,c; cf. Fig. 7c,d). Overall, DD475/476AA and EE482/483AA STIM1 double mutations significantly diminished the dependence of ICRAC FCDI on pHi and the relative Orai1/STIM1 expression ratio, without affecting pHi dependence of ICRAC amplitude.

One of the distinctive properties of Orai1/STIM1 mediated ICRAC is inhibition by high (over 100 µM) and potentiation by low (below 10 µM) concentrations of 2-APB, whereas application of intermediate concentrations of 2-APB (10–50 µM) cause transient potentiation of ICRAC followed by inhibition22. Previously we have shown that the extent of ICRAC potentiation by 2-APB depends on the relative expression levels of STIM1 and Orai116. The higher the expression of Orai1, relative to STIM1, the stronger the potentiation16. Here we investigated whether potentiation of ICRAC amplitude by 50 µM 2-APB is affected by pHi. The amplitude of ICRAC in cells transfected with WT STIM1 and Orai1 at 4:1 ratio increased more than 4-fold at acidic pHi of 6.3, but only 1.3-fold when pHi was raised to 8.3, compared to a potentiation of 2.5-fold at pH 7.3 (Fig. 8a). Despite the lack of pHi effect on FCDI and the apparent P o of ICRAC mediated by Orai1/EE482/483AA-STIM1, the dependence of 2-APB mediated potentiation of this mutant on pHi remained unchanged, compared to WT ICRAC (Fig. 8b).

Figure 8
figure 8

The pHi dependence of ICRAC potentiation by 2-APB. The Y-axis represents the ratio between the amplitudes of ICRAC recorded immediately after and before application of 50 µM 2-APB to the bath. The ICRAC amplitude was measured at −100 mV from the responses to 100 ms voltage ramps from −120 to 120 mV, applied every 2 seconds. HEK 293 T cells were transfected with WT STIM1 and Orai1 (a) or EE482/3AA STIM1and Orai1(b) at 4:1 ratio. pH of the pipette solution is indicated below the bars.

Full size image

Discussion

The key findings of this paper can be summarised as follows – (i) pHi regulates both, the amplitude of ICRAC and Ca2+ dependent gating of CRAC channels; (ii) increase in ICRAC amplitude in response to alkaline pHi in the presence of EGTA in the pipette solution is a result of pH dependence of the Ca2+ buffering properties of EGTA, not the CRAC channel itself; (iii) Glutamate 106 in the selectivity centre of Orai1 pore, which mediates ICRAC dependence on pHo, does not contribute to ICRAC dependence on pHi; (iv) negatively charged residues in IDSTIM domain play a role in pHi regulation of CRAC channel gating kinetics but not the amplitude of ICRAC. These data suggest that several mechanisms contribute to ICRAC regulation by pHi.

It has been shown previously that increasing the amounts of Orai1 relative to STIM1 results in a smaller ICRAC that exhibits re-activation at negative potentials which masks FCDI15, 16. The results presented here show that intracellular acidification has an effect on ICRAC similar to that of increasing the relative amounts of Orai1 (or decreasing the relative amounts of STIM1). Comparable changes in ICRAC kinetics and amplitude caused by intracellular acidification and increased Orai1:STIM1 ratio suggest that low pHi reduces the affinity of STIM1 binding to Orai1, likely due to protonation of specific residues, which is equivalent to a reduction of available STIM1. This notion is supported by previous observations that acidification of cytoplasm due to hypoxia reduces FRET between Orai1and STIM1 and inhibits ICRAC 13. In the study of Mancarella et al. (2012) the effect of hypoxia on ICRAC could be mimicked by application of extracellular propionate, which lowers pHi, and reversed by application of NH4Cl, which raises pHi 13. Intracellular acidification was shown to reduce FRET between STIM1-YFP and Orai1-CFP, but no change was observed in STIM1/Orai1 co-localisation in puncta13. These results suggested that pHi affects STIM1/Orai1 functional coupling leading to channel opening, but not the interactions that trap STIM1 and Orai1 in puncta13. The pH dependent changes in ICRAC kinetics reported here also point to the conclusion that intracellular acidification disrupts STIM1/Orai1 functional interactions.

Inhibition of ICRAC by low pHi has been demonstrated previously in several publications8, 10, 12. They all agree that ICRAC, both endogenous and mediated by ectopically expressed Orai1 and STIM1, is inhibited by approximately 70–90% at pHi of around 6, compared to pHi 7.38, 10, 12. In contrast, the effects of alkalinisation of pHi above 7.3 on ICRAC are inconsistent between different studies8, 10, 12. The results of the present work suggest that the reason for the discrepancy is likely to be due to the type of intracellular Ca2+ buffer used. Studies employing BAPTA in the pipette did not find much increase in ICRAC amplitude at higher pHi, whereas studies that used EGTA reported a significant potentiation of ICRAC amplitude by alkalinisation8, 10, 12. Calculations using Maxchelator (http://maxchelator.stanford.edu/) indicate that Ca2+ buffering capacity of EGTA is highly pH dependent, and raising pH by one unit increases EGTA binding affinity to Ca2+ two orders in magnitude, changing Kd from 1.28 × 10−7 M at pH 7.3 to 1.4 × 10−9 M at pH 8.3, whereas pH dependence of Ca2+ buffering by BAPTA is weak.

The observations reported here which show strong increase in ICRAC amplitude in response to NH4Cl application to the bath when EGTA is used in the pipette, and virtual absence of such effect when intracellular Ca2+ is buffered with BAPTA, suggest that the Ca2+ binding properties of EGTA play a significant part in ICRAC pHi dependence in the presence of EGTA, particularly, when pHi rises above 7.5. The increase in ICRAC amplitude at alkaline pHi is likely to be due to stronger and faster Ca2+ binding by EGTA, rather than increase in pHi per se. pH dependence of EGTA Ca2+ binding properties creates unwanted complications for the interpretation of the experimental results. However, many physiological intracellular Ca2+ buffers are likely to exhibit strong pH dependence, similarly to EGTA23, 24. This is supported by the observations that intracellular alkalinisation induced by application of NH4Cl to the bath in Ca2+ imaging experiments, when cells have endogenous intracellular Ca2+ buffering, potentiates store-operated Ca2+ entry in platelets and HT-29 cells21, 25. Therefore, results obtained using EGTA, rather than BAPTA, may have more physiological relevance. Much bigger amplitude of ICRAC activated by IP3 in the presence of BATPA in the pipette solution, compared to EGTA, was noticed very early on 19. However, the reason for this difference remains poorly understood.

The only residue that has been implicated in ICRAC dependence on pHi so far is His 155 in Orai110. H155F mutation in Orai1 was shown to abolish the increase of ICRAC amplitude in response to intracellular alkalinisation, but ICRAC mediated by H155F-Orai1 was still inhibited by about 60% at low pHi 10, which implies that His 155 is unlikely to be the only site that mediates ICRAC regulation by pHi. Data presented in this work indicate that Glut 106 in the Orai1 selectivity centre, which can be protonated from the extracellular side9, does not contribute to pHi dependence at all, which also suggests that Orai1 pore is not permeable to protons. Presence of seven negatively charged residues within IDSTIM and the fact that neutralisation of three of them, D476, D478, and D479, significantly reduced the FCDI, similarly to acidic pHi, made IDSTIM a good candidate for the pHi sensor of CRAC channel18. The results presented here indicate that IDSTIM is not involved in pHi dependence of the ICRAC amplitude, but mutations in IDSTIM affect pHi regulation of ICRAC Ca2+ dependent gating. The kinetics of ICRAC mediated by Orai1/EE482/483AA-STIM1 or Orai1/DD475/476AA-STIM1 was not appreciably affected by either acidic, or alkaline pHi. It is unlikely, however, that protonation/deprotonation of negatively charged resides in IDSTIM is responsible for the changes in ICRAC kinetics induced by the changes in pHi. Neutralisation of aspartates 482 and 483 increases FCDI, so protonation of these aspartates alone cannot be responsible for reduced FCDI at acidic pHi. It has been shown previously that neutralization of these aspartates together with glutamates in IDSTIM reduce FCDI18, i.e. the effect of neutralisation of glutamates overcomes the effect of neutralisation of aspartates. This suggests that if glutamates in the IDSTIM were protonated at acidic pHi, Orai1/EE482/483AA-STIM1 would display dependence of kinetics on pHi similar to that of WT ICRAC. However, this was not the case, which excludes IDSTIM as a direct pHi sensor.

Interestingly, EE482/483AA-STIM1 significantly diminished the dependence of ICRAC kinetics not only on pHi, but also on the relative expression levels of STIM1 and Orai1. This could’ve been a result of saturating levels of expression of the mutant STIM1, compared to Orai1. If the expression levels of STIM1 are very high, moderate changes in the affinity of STIM1 biding to Orai1 due to changes in pHi, or moderate increase in Orai1 expression, are unlikely to have an appreciable effect on ICRAC kinetics. However, when the amounts of STIM1 are close to saturating, 2-APB does not potentiate ICRAC 16. Application of 2-APB to Orai1 EE482/483AA-STIM1 mediated ICRAC caused the same level of potentiation as in WT ICRAC at all intracellular pH tested. This indicates that expression levels of mutant STIM1 were not different from that of WT STIM1, and that pHi affected functional coupling of Orai1 with mutant STIM1 in the same way it has affected it’s functional coupling with WT STIM1. The lack of the dependence of Orai1 EE482/483AA-STIM1 ICRAC kinetics on the Orai1:STIM1 relative expression ratio and pHi suggests that the minimum number of this mutant STIM1 peptides which is needed to open Orai1 pore, is sufficient to support fully functional ICRAC FCDI.

In conclusion, the results presented here support the hypothesis that ICRAC inhibition by intracellular acidification is caused by disruption of functional coupling of STIM1 and Orai1, whereas the increase in ICRAC amplitude at alkaline pHi in the presence of EGTA is mainly due to increased Ca2+ buffering capacity of EGTA. Negatively charged IDSTIM is not a direct pHi sensor, but mutations neutralising negative charges in IDSTIM affect pHi dependence of ICRAC kinetics by changing the interaction between STIM1 and Orai1.

Methods

Cell culture and transfections

HEK-293T cells [human embryonic kidney-293 cells expressing the large T antigen of SV40 (simian virus 40)] (A.T.C.C. CRL 11268) were cultured at 37 °C in 5% (v/v) CO2 in air in DMEM (Dulbecco’s modified Eagle’s medium) supplemented with 100 μM nonessential amino acids, 2 mM L-glutamine and 10% fetal bovine serum9, 16. To co-express WT Orai1 with WT STIM1 or double STIM1 mutants (EE482/483AA and DD475/476AA), cells seeded on glass cover slips were transfected using Polyfect (Qiagen) transfection reagent according to the manufacturer’s instructions. The Orai1 and STIM1 (WT or mutant) plasmids were transfected at two Orai1:STIM1 molar ratios 1:1 and 1:49, 16. Plasmids containing EE482/483AA and DD475/476AA double STIM1 mutants were generously provided by Prof Richard Lewis (Stanford University, USA).

Patch clamping

Whole-cell patch clamping was performed at room temperature (23 °C) using a computer based patch-clamp amplifier (EPC-9, HEKA Elektronik) and PULSE software (HEKA Elektronik) as previously described9, 16. The control bath solution contained 140 mM NaCl, 4 mM CsCl, 10 mM CaCl2, 2 mM MgCl2 and 10 mM HEPES adjusted to pH 7.4 with NaOH. Depletion of intracellular Ca2+ stores was achieved using 20 μM Ins(3,4,5)P3 (Sigma) added to an internal solution containing 130 mM caesium glutamate, 10 mM CsCl, 5 mM MgCl2, 1 mM MgATP, 10 mM EGTA and either 10 mM MES adjusted to pH 6.3 with NaOH, or 10 mM HEPES adjusted to pH 7.3 or 8.3 with NaOH. Patch pipettes were pulled from borosilicate glass and fire polished to give a pipette resistance between 2 and 4 MΩ. Series resistance did not exceed 15 MΩ and was 50–70% compensated. Traces obtained before activation of ICRAC, or after its inhibition with 10 µM La3+ were used for leakage subtraction.

Data analysis

To obtain apparent (relative) open probability (P o) curves of CRAC channels, instantaneous tail currents recorded in response to voltage steps to −100 mV after test pulses between −140 and 80 mV, applied every 5 s in 20 mV increments, were normalised to the amplitude of the instantaneous tail current recorded after test pulse to 80 mV and plotted against corresponding test pulse voltage9. The length of the test pulses was set to 150 ms to make sure that both gating processes of ICRAC – inactivation and re-activation are captured in one protocol. Were possible, the data points were fitted with the Boltzmann distribution with an offset of the form:

$${P}_{{\rm{o}}}(V)={P}_{min}+(1{\textstyle \text{-}}{P}_{min})/(1+\exp (({V}_{1/2}-V)/k))$$
(1)

where P min is an offset, V is the membrane potential, V 1/2 is the half-maximal activation potential (V 1/2 corresponds to the inflexion point of the P o curve) and k is the slope factor. However, in many cases apparent P o data could not be fitted with Boltzmann distribution and the data points were fitted with a smooth curve using cubic spline procedure in Prizm 6 software.