ABSTRACT

All three subtypes of inositol 1,4,5-trisphosphate receptor (IP3R) are intracellular Ca2+ channels that are co-regulated by IP3 and Ca2+. This allows IP3Rs to evoke regenerative Ca2+ signals, the smallest of which are Ca2+ puffs that reflect the coordinated opening of a few clustered IP3Rs. We use total internal reflection microscopy (TIRF) microscopy to record Ca2+ signals in HEK cells expressing all three IP3R subtypes or a single native subtype. Ca2+ puffs are less frequent in cells expressing one IP3R subtype, commensurate with them expressing fewer IP3Rs than wild-type cells. However, all three IP3R subtypes generate broadly similar Ca2+ puffs with similar numbers of IP3Rs contributing to each. This suggests that IP3R clusters may be assembled by conserved mechanisms that generate similarly sized clusters across different IP3R expression levels. The Ca2+ puffs evoked by IP3R2 had slower kinetics and more prolonged durations, which may be due to IP3 binding with greater affinity to IP3R2. We conclude that Ca2+ puffs are the building blocks for the Ca2+ signals evoked by all IP3Rs.

INTRODUCTION

Many receptors evoke Ca2+ signals by stimulating phospholipase C (Berridge, 2016). Inositol 1,4,5-trisphosphate (IP3) then evokes Ca2+ release from the endoplasmic reticulum by stimulating IP3 receptors (IP3Rs), which are widely expressed Ca2+ channels. IP3 binding to IP3R initiates its activation by promoting the Ca2+ binding that leads to channel opening (Prole and Taylor, 2016). This dual regulation allows IP3Rs to evoke regenerative Ca2+ signals, the smallest of which are Ca2+ puffs (Smith and Parker, 2009). These arise from the nearly simultaneous opening of a few clustered IP3Rs as Ca2+ released by one IP3R ignites the activity of its neighbours. Ca2+ puffs are thought to be fundamental building blocks of IP3-evoked Ca2+ signalling. They require assembly of IP3Rs into small clusters and their anchoring at sites close to the plasma membrane (Smith et al., 2009a) where additional signals prime or ‘license’ IP3Rs to respond (Thillaiappan et al., 2017).

Vertebrate genomes encode three IP3R subtypes (IP3R1, IP3R2, IP3R3). Each forms a Ca2+ channel that is co-regulated by IP3 and Ca2+, but the subtypes differ in their expression patterns (Taylor et al., 1999), affinities for IP3 (Iwai et al., 2007) and modulation by additional signals (Prole and Taylor, 2016). The differences are consistent with evidence that implicates IP3R subtypes in specific cellular responses (Miyakawa et al., 1999; Wei et al., 2009), the distinctive phenotypes of mice lacking single subtypes (Futatsugi et al., 2005) and with different diseases arising from mutations in different subtypes (Hisatsune and Mikoshiba, 2017). However, many of these associations might be more related to the predominance of IP3R subtypes in different tissues than to fundamental differences in the behaviour of IP3R subtypes. Ca2+ puffs have been reported for cells in which different IP3R subtypes predominate (Keebler and Taylor, 2017), but it is unknown whether all three IP3R subtypes can generate Ca2+ puffs.

Here, we use total internal reflection fluorescence microscopy (TIRFM) to record Ca2+ puffs evoked by photolysis of caged-IP3 (ci-IP3) in human embryonic kidney 293 (HEK) cells expressing one or all three IP3R subtypes. We demonstrate that all three IP3R subtypes generate Ca2+ puffs with broadly similar properties. We conclude that Ca2+ puffs are the fundamental building blocks for Ca2+ signals evoked by all IP3Rs.

RESULTS AND DISCUSSION

Cells with single IP3R subtypes generate Ca2+ puffs

We used wild-type (WT) HEK cells, which express three IP3R subtypes (Fig. 1), and HEK cells where CRISPR/Cas9 had been used to generate cells expressing single subtypes (Alzayady et al., 2016). Identical stimuli were delivered to each cell by photolysis of ci-IP3, which uniformly releases the active, but slowly degraded, analogue of IP3, iIP3 (d-2,3-O-isopropylidine-IP3) (Dakin and Li, 2007). We confirmed, by recording Ca2+ signals evoked by photolysis of caged-Ca2+, that Ca2+ buffering was similar in each cell line (Fig. S1A,B).

Fig. 1.

Expression of IP3R subtypes in HEK cells. (A) Western blots using IP3R subtype-selective antibodies (Ab1, Ab2, Ab3) or a common antibody (AbC). Lanes were loaded with 20, 40, 80 or 120 μg protein for WT cells, and with 80 or 120 μg for other cell lines. Molecular mass markers (kDa) are shown. Arrowheads indicate IP3R bands used for quantification. (B–D) Summary of results (in %, mean±s.e.m., n=5) show relative expression of IP3R subtypes in WT cells (B), determined by calibrating bands identified by Ab1, Ab2 or Ab3 to bands identified by AbC in cell lines expressing single IP3R subtypes; IP3R expression (in %) in cell lines expressing single subtypes relative to the same IP3R subtype in WT cells (C); and total number of IP3Rs (in %) expressed in each cell line relative to number in WT (D), by comparing intensities of protein bands recognised by AbC. In C, the asterisk (*) denotes values where 95% confidence interval does not include 100%. No significant differences (P<0.05) were observed between values in D.

Fig. 1.

Expression of IP3R subtypes in HEK cells. (A) Western blots using IP3R subtype-selective antibodies (Ab1, Ab2, Ab3) or a common antibody (AbC). Lanes were loaded with 20, 40, 80 or 120 μg protein for WT cells, and with 80 or 120 μg for other cell lines. Molecular mass markers (kDa) are shown. Arrowheads indicate IP3R bands used for quantification. (B–D) Summary of results (in %, mean±s.e.m., n=5) show relative expression of IP3R subtypes in WT cells (B), determined by calibrating bands identified by Ab1, Ab2 or Ab3 to bands identified by AbC in cell lines expressing single IP3R subtypes; IP3R expression (in %) in cell lines expressing single subtypes relative to the same IP3R subtype in WT cells (C); and total number of IP3Rs (in %) expressed in each cell line relative to number in WT (D), by comparing intensities of protein bands recognised by AbC. In C, the asterisk (*) denotes values where 95% confidence interval does not include 100%. No significant differences (P<0.05) were observed between values in D.

Antibodies selective for each IP3R subtype (Ab1–3) confirmed that WT HEK cells express three IP3R subtypes (IP3R3≥IP3R1>IP3R2; Fig. 1A,B). This profile differs from previous analyses using immunoblots (Wojcikiewicz, 1995) or QPCR (Tovey et al., 2008), which suggested IP3R1 to be the minor subtype. The differences might be due to the methods used or the source of HEK cells – our cell line is the one from which the gene-edited cells were derived. Within each cell line expressing a single IP3R subtype, its expression level was lower than in WT cells (Fig. 1A,C). By using an antibody that recognises a sequence conserved in all IP3R subtypes (AbC), we established that overall IP3R expression levels were similar in each cell line expressing a single IP3R subtype (Fig. 1D), but lower (∼32%) than in WT cells (Fig. 1D).

In WT cells, but not in HEK cells without IP3Rs (Fig. S1C), photolysis of ci-IP3 evoked Ca2+ puffs after a short latency. The Ca2+ signal then propagated after a few seconds to produce a global increase in cytosolic free Ca2+ concentration ([Ca2+]c) (Fig. 2A–D). These responses are similar to those observed in other cell types (see Keebler and Taylor, 2017 and references therein). However, and consistent with our analyses of HEK and HeLa cells (Keebler and Taylor, 2017; Thillaiappan et al., 2017), the sites at which Ca2+ puffs occurred in the 40 s after photo-release of iIP3 (33.8±2.7 sites, n=15 cells; Fig. 2I) were about ten-times more abundant than reported in studies using different cell types (Nakamura et al., 2012; Smith and Parker, 2009; Thomas et al., 2000). Our detection of more Ca2+ release sites is not due to mis-identification of Ca2+ puffs but probably arises from improved detection through automated analysis and longer effective recordings (Keebler and Taylor, 2017).

Fig. 2.

Ca2+ puffs evoked by photolysis of ci-IP3. TIRFM images from part of a WT HEK cell show fluorescence before and after photolysis of ci-IP3. The first frame after the UV flash has a red border. Images were collected at 5.32-ms intervals, with every 20th frame shown. Arrowheads show Ca2+ puffs identified by Flika image processing. (B) Fluorescence (F/F0) recorded before and after the UV flash shows Ca2+ puffs before the Ca2+ signal propagates globally (recording area 1.76×1.76 μm). (C,D) Time courses of Ca2+ puffs in WT cells (C) and fluorescence profiles at the peak of each puff (D). (E) Latencies to first detected Ca2+ puff. (F) Cells (in %) that showed a global increase in [Ca2+]c within the 40-s recording. (G–I) Frequency of Ca2+ puffs (G), total numbers of Ca2+ puffs (H) and release sites (I) detected during 40-s recording. In E and G-I, results show individual mean values from 15 cells, and the mean±s.e.m. of those values. *P<0.05, ***P<0.001, relative to WT. The colour code used in E–I is explained in F.

Fig. 2.

Ca2+ puffs evoked by photolysis of ci-IP3. TIRFM images from part of a WT HEK cell show fluorescence before and after photolysis of ci-IP3. The first frame after the UV flash has a red border. Images were collected at 5.32-ms intervals, with every 20th frame shown. Arrowheads show Ca2+ puffs identified by Flika image processing. (B) Fluorescence (F/F0) recorded before and after the UV flash shows Ca2+ puffs before the Ca2+ signal propagates globally (recording area 1.76×1.76 μm). (C,D) Time courses of Ca2+ puffs in WT cells (C) and fluorescence profiles at the peak of each puff (D). (E) Latencies to first detected Ca2+ puff. (F) Cells (in %) that showed a global increase in [Ca2+]c within the 40-s recording. (G–I) Frequency of Ca2+ puffs (G), total numbers of Ca2+ puffs (H) and release sites (I) detected during 40-s recording. In E and G-I, results show individual mean values from 15 cells, and the mean±s.e.m. of those values. *P<0.05, ***P<0.001, relative to WT. The colour code used in E–I is explained in F.

Photolysis of ci-IP3 also evoked Ca2+ puffs in HEK cells expressing single IP3R subtypes. In these cell lines, all of which were exposed to the same stimulus, the latency before the first Ca2+ puff was longer than in WT cells, Ca2+ puffs were less frequent, we detected fewer Ca2+-release sites and it was rare for Ca2+ signals to propagate into a global increase in [Ca2+]c (Fig. 2E–I).

Ca2+ puffs initiate at fewer sites in cells expressing only one IP3R subtype

In HEK WT cells (Keebler and Taylor, 2017), as in other cell types (Smith et al., 2009b; Thillaiappan et al., 2017), Ca2+ release sites remain at fixed locations for many minutes. The opportunity to identify Ca2+ release sites depends on the number of Ca2+ puffs: more puffs provide more opportunities to find sites. Ca2+ puffs were frequent in WT cells, but were detected for only a few seconds before the Ca2+ signal invaded the cell (Fig. 2B,F). In the other HEK cell lines used by us, Ca2+ puffs were less frequent (Fig. 2G) but the useful recording interval was longer because Ca2+ signals rarely became global (Fig. 2F, Fig. S1D). Fortuitously, the average number of Ca2+ puffs detected during the entire recording interval was not significantly different between the four cell lines, but we detected significantly more Ca2+ release sites in WT HEK cells (Fig. 2H,I).

Another approach to identifying the total number of release sites is to recognise that, as the recording proceeds, each successive Ca2+ puff has a diminished opportunity to reveal a new site. For cells expressing single IP3R subtypes, we divided the 40-s recording into 10-s intervals, in each of which a similar number of Ca2+ puffs was detected (Fig. S2A). The number of new sites identified in each successive interval declined mono-exponentially with half-times (t1/2=3.76–5.53 s) (Fig. S3A) that were fast enough to ensure that, within the 40-s recording interval (>4 half-lives), we should identify >94% of all Ca2+ release sites. The analysis of WT cells used a shorter time-bin (2 s) to accommodate the increased frequency of Ca2+ puffs (Fig. 2G) but yielded similar results. The typical effective recording period (∼10 s) for WT cells was about four-times longer than the half-time for the mono-exponential decay of the discovery of new Ca2+ release sites (t1/2=2.01 s), indicating that we should again detect >93% of sites (Fig. S3B). Hence, for all cell lines our analyses are likely to detect most Ca2+ release sites. The fewer sites identified in cells with single IP3R subtypes are, therefore, not due to inadequate sampling. We conclude that the cell lines with single IP3R subtypes express about one-third the total number of IP3Rs of WT cells (Fig. 1D) and about half the number of Ca2+-release sites (Fig. 2I).

Ca2+ puffs evoked by IP3R2 are more long-lasting

There is considerable variation in the shape and duration of Ca2+ puffs (Fig. 2C,D, Fig. S1E) (Smith and Parker, 2009). For Ca2+ puffs in WT HEK cells, the fluorescence signals typically rose rapidly to a peak before more slowly decaying back to baseline (Fig. 2C). To secure sufficient data for quantitative analysis, we analysed Ca2+ puffs for as long as practicable within the 40-s recording. For WT cells this entailed recording frequent Ca2+ puffs for the relatively short interval (typically ∼10 s) before the Ca2+ signal invaded the cell, whereas – for the other cell lines – it entailed recording less frequent Ca2+ puffs for longer periods (40 s) (Fig. 2F,G). To validate comparisons of WT cells with the other cell lines, we showed there were no systematic changes in Ca2+ puff frequency across the analysis intervals (Fig. S2A).

We compared mean rise times, decay times and durations of the Ca2+ puffs in the four cell lines (Fig. 3A–C) and their frequency distributions (Fig. S4). The rise times (∼20 ms), decay times (∼30 ms) and durations at half-maximal amplitude (∼30 ms) were similar in WT cells and in cells expressing only IP3R1 or IP3R3. However, the rise and decay times (44.6±4.7 ms and 53.1±5.0 ms, respectively) were significantly longer in cells expressing only IP3R2, as was the mean duration of Ca2+ puffs (62.6±5.6 ms) (Fig. 3). The frequency distributions revealed that the more-prolonged events in cells expressing IP3R2 were largely attributable to an extended tail in the distribution: very long events rarely seen in other cell lines were more frequent in IP3R2-expressing cells (Fig. S4). We conclude that all three IP3R subtypes generate Ca2+ puffs with broadly similar characteristics, although Ca2+ puffs evoked by IP3R2 typically last about twice as long as those in cells expressing IP3R1, IP3R3 or all three subtypes.

Fig. 3.

Properties of Ca2+ puffs in cells expressing a single IP3R subtype. (A–C) Mean rise time (A) – i.e. time for fluorescence to increase from 20% to 100% of peak, decay time (B) – i.e. time to fall from 100% to 20%, and mean duration at half-maximal amplitude (C) of Ca2+ puffs evoked by photolysis of ci-IP3. Results show mean values for each of 15 cells for each cell line (∼1000 puffs per cell line were analysed), and mean±s.e.m. of these values. *P<0.05, ***P<0.001, relative to WT.

Fig. 3.

Properties of Ca2+ puffs in cells expressing a single IP3R subtype. (A–C) Mean rise time (A) – i.e. time for fluorescence to increase from 20% to 100% of peak, decay time (B) – i.e. time to fall from 100% to 20%, and mean duration at half-maximal amplitude (C) of Ca2+ puffs evoked by photolysis of ci-IP3. Results show mean values for each of 15 cells for each cell line (∼1000 puffs per cell line were analysed), and mean±s.e.m. of these values. *P<0.05, ***P<0.001, relative to WT.

A similar number of IP3Rs contributes to Ca2+ puffs for each IP3R subtype

During the falling phase of a Ca2+ puff, it was sometimes possible to observe stepwise closures of individual IP3Rs (Fig. 4A, Fig. S1E) (Smith and Parker, 2009). The amplitude of these unitary fluorescence steps was indistinguishable in the four cell lines (Fig. 4B), suggesting that each IP3R subtype is similarly distributed within the TIRF field. The results are also consistent with evidence that IP3R subtypes have a similar cation conductance (Vais et al., 2010). We used the average mean amplitude of the unitary fluorescence step from all cells (ΔF=0.101±0.002, n=40) and the peak amplitude of each Ca2+ puff, to estimate the number of IP3Rs contributing to a Ca2+ puff (Smith and Parker, 2009). The results show that most Ca2+ puffs involve fewer than seven IP3Rs (Fig. 4C, Fig. S5D). In WT cells, the mean number of IP3Rs per puff was 2.77±0.10 (Fig. 4C,D). Our results are similar to those from SH-SY5Y cells, for which 50% of IP3Rs within a cluster are suggested to open during a typical Ca2+ puff (Smith et al., 2009a). Collectively, these observations are consistent with analyses of endogenous IP3Rs in HeLa cells, suggesting that IP3R clusters include about eight IP3Rs (Thillaiappan et al., 2017). The mean peak amplitudes of Ca2+ puffs were indistinguishable in WT cells and cells expressing only IP3R3. The amplitudes in cells expressing IP3R1 and IP3R2 were different from those in WT cells (Fig. 4C,D), but the differences were modest (<18%), with an estimated 2.28±0.11 and 3.19±0.13 IP3Rs per puff for cells with IP3R1 and IP3R2, respectively (Fig. 4D). We conclude that, despite differences in IP3R expression between WT cells and cells expressing a single IP3R subtype (Fig. 1D), Ca2+ puffs are generated by a similar number of IP3Rs in each cell line (Fig. 4).

Fig. 4.

Amplitudes of Ca2+ puffs evoked by different IP3R subtypes. (A) Examples of Ca2+ puffs from WT cells show discrete fluorescence steps in falling phase. Dashed lines show unitary steps. (B) Summary of results (mean±s.e.m. n=10 puffs from different cells) show the amplitude of unitary fluorescence steps during the falling phase of Ca2+ puffs. Values do not differ significantly between cell lines (P>0.05). (C) Frequency distributions (mean±s.e.m., n=15 cells) show the estimated number of active IP3Rs within each Ca2+ puff. The dashed line indicates the mean value for WT cells. (D) Summary of results (mean±s.e.m., n=15 cells) show the ΔF values from which the number of active IP3Rs within each Ca2+ puff was calculated. *P<0.05, relative to WT.

Fig. 4.

Amplitudes of Ca2+ puffs evoked by different IP3R subtypes. (A) Examples of Ca2+ puffs from WT cells show discrete fluorescence steps in falling phase. Dashed lines show unitary steps. (B) Summary of results (mean±s.e.m. n=10 puffs from different cells) show the amplitude of unitary fluorescence steps during the falling phase of Ca2+ puffs. Values do not differ significantly between cell lines (P>0.05). (C) Frequency distributions (mean±s.e.m., n=15 cells) show the estimated number of active IP3Rs within each Ca2+ puff. The dashed line indicates the mean value for WT cells. (D) Summary of results (mean±s.e.m., n=15 cells) show the ΔF values from which the number of active IP3Rs within each Ca2+ puff was calculated. *P<0.05, relative to WT.

We considered whether Ca2+ puffs occurring in the most-active cells differed in their properties. There were no clear differences in the frequency distribution for rise time, decay time or Ca2+ puff duration, or in the estimated number of IP3Rs per puff for the more-active and less-active cells (Fig. S5). We also compared the mean peak amplitude of Ca2+ puffs evoked in each cell with mean rise and fall times. Again, there was no significant correlation between puff amplitude and kinetics; the only exception was a positive correlation between amplitude and decay time for cells expressing IP3R2 (Fig. S6).

Conclusions

We show, for the first time, that all three IP3R subtypes generate Ca2+ puffs with broadly similar properties. Ca2+ puffs have hitherto been observed only in cells that express all three IP3R subtypes, albeit with one subtype usually predominating. Ca2+ puffs originate at fixed intracellular sites, at each of which a loose cluster of a few IP3Rs is anchored near the plasma membrane, where an additional signal licenses the cluster to respond to IP3 (Smith et al., 2009a; Thillaiappan et al., 2017). Neither the scaffold that assembles clusters nor the signal that licenses them to respond has been identified. WT HEK cells expressed three-times more IP3Rs than HEK cells with single IP3R subtypes (Fig. 1D), yet each cell line generated Ca2+ puffs in which the unitary steps during the falling phase and number of active IP3Rs contributing to each Ca2+ puff were similar (Fig. 4B-D, Figs S1E, S5D). These results suggest that, first, for each subtype, IP3Rs are similarly distributed across the ∼90-nm depth of the TIRF field, consistent with each coming close to the plasma membrane. Second, across a range of IP3R expression levels, IP3Rs are assembled into similarly sized clusters, but with more clusters in cells with more IP3Rs (Fig. 2I, Fig. S3). We conclude that the clusters of IP3Rs that generate Ca2+ puffs are probably not random associations. Instead, IP3Rs appear to be assembled through mechanisms that are conserved for all IP3R subtypes, and which ensure that similarly sized clusters are assembled irrespective of IP3R expression levels.

The similar properties of Ca2+ puffs evoked by cells expressing all IP3R subtypes or by those expressing only a single subtype are striking. However, Ca2+ puffs evoked by IP3R2 were more long-lasting than those evoked by other subtypes. We have not determined the mechanisms responsible for the slower kinetics of Ca2+ puffs evoked by IP3R2. One possibility is that the greater affinity of IP3 for IP3R2 (Iwai et al., 2007) slows IP3 dissociation and delays channel closures. In Xenopus oocytes, however, Ca2+ puffs evoked by the high-affinity IP3 analogue adenophostin A were briefer than those evoked by IP3 (Marchant and Parker, 1998). Furthermore, in our analyses, the frequency of Ca2+ puffs evoked by identical stimuli in cell lines expressing IP3R1, IP3R2 or IP3R3 were indistinguishable (Fig. 2G), which seems difficult to reconcile with the idea that IP3Rs have substantially different affinities in situ. It might be that the mechanism that allows IP3R clusters to respond overrides any underlying differences in the IP3 affinity of IP3R subtypes.

We provide the first demonstration that each IP3R subtype generates Ca2+ puffs with broadly similar properties, leading us to suggest that the small, loose IP3R clusters from which Ca2+ puffs originate are the building blocks for the Ca2+ signals evoked by all IP3Rs.

MATERIALS AND METHODS

Materials

Cal520-AM was from Stratech Scientific (Suffolk, UK). Bovine serum albumin (BSA) was from Europa Bio-Products (Ely, UK). Gibco TrypLE Express, NP-EGTA-AM (caged EGTA, O-nitrophenyl-EGTA-AM) and Dulbecco's modified Eagle's medium (DMEM)/Ham's F-12 (50:50) supplemented with GlutaMAX were from ThermoFisher (Waltham, MA). Pluronic acid F-127, dimethyl sulfoxide (DMSO), poly-l-lysine and foetal bovine serum were from Sigma-Aldrich (Poole, Dorset, UK). EGTA-AM was from Merck Millipore (Darmstadt, Germany). Caged-IP3-PM [ci-IP3-PM, d-2,3-O-isopropylidine-6-O-(2-nitro-4,5-dimethoxy)benzyl-myo-inositol 1,4,5-trisphosphate-hexakis (proprionoxymethyl)ester] was from SiChem GmbH (Bremen, Germany). Custom-made rabbit anti-peptide antibodies against IP3R2 (Ab2, GFLGSNTPHENHHMPPH, 1:1000) and to a sequence conserved in all three IP3R subtypes (AbC, PMRYSAKQKFWKA, 1:500) were provided by Poccono Rabbit Farm and Laboratory Inc. (Canadensis, PA). Additional primary antibodies were: IP3R1 (Ab1, rabbit, #3763, Cell Signaling Technology, Leiden, The Netherlands, 1:1000), IP3R3 (Ab3, mouse, #610312, BD Biosciences, Wokingham, UK, 1:1000) and β-actin (mouse, #3700, Cell Signaling Technology, 1:1000). Secondary antibodies conjugated to horseradish peroxidase (anti-mouse, sc-516102; anti-rabbit, sc-2357, 1:5000) were from Insight Biotechnology Ltd (Wembley, UK).

Cell culture

WT HEK cells, from which the gene-edited lines were generated, were provided by Dr D. Yule (University of Rochester, NY). HEK cells in which CRISPR/Cas9 had been used to generate lines expressing a single IP3R subtype or devoid of all IP3Rs (HEK-KO) (Alzayady et al., 2016) were from Kerafast (Boston, MA). We confirmed that all cells were mycoplasma-free. Cells were maintained in DMEM/Ham's F-12 supplemented with GlutaMAX and 10% foetal bovine serum in humidified air at 37°C with 5% CO2, and passaged when they reached ∼80% confluence. For imaging, cells (5×104/dish) were grown on poly-l-lysine-coated (10 µg/ml) glass-bottomed dishes (P35GC-1.5-14-C, MatTek, Ashland, MA). They were used after 48 h, when they were sub-confluent.

Immunoblotting

Cells were lysed in cold RIPA buffer (1 mM Tris HCl, 15 mM NaCl, 0.5 mM EDTA, 0.1% Triton X, pH 7.5) with protease inhibitors (cOmplete, EDTA-free protease inhibitor cocktail, Sigma-Aldrich). Proteins in the supernatant (900× g, 4 min, 4°C) were separated (NuPAGE 3-8% Tris-acetate protein gel, 1.0 mm, 12-well, ThermoFisher) and transferred to a polyvinyl difluoride (PVDF) membrane using an iBLOT gel-transfer system (ThermoFisher). Membranes were blocked in Tris-buffered saline (TBS: 50 mM Tris-HCl, 150 mM NaCl, pH 7.5, 0.2% Tween-20, 5% BSA (1 h, 20°C) and incubated (40 h, 4°C) with primary antibody. After three 5-min washes in TBS, membranes were incubated (1 h, 20°C) with secondary antibody and washed (3×5 min). Bands were visualised with ECL Primer Western blotting detection reagent (SLS, Nottingham, UK) and a Syngene PXi chemiluminescence detection system (Cambridge, UK). We confirmed that, within the range of loadings used for analysis, IP3R band intensities scaled linearly with the amount of protein loaded.

To quantify expression of IP3R subtypes in WT HEK cells, we used the cells expressing single subtypes to determine the ratio of band intensities for each IP3R subtype-selective antibody (Ab1-3) relative to the common antibody (e.g. Ab1:AbC for IP3R1). These ratios were used to convert intensities of IP3R subtype-selective bands in WT cells to AbC intensities, from which we estimated the relative expression of the three IP3R subtypes (Fig. 1B).

High-resolution imaging of Ca2+ signals

To record local Ca2+ signals, cells were loaded with EGTA to restrict Ca2+ diffusion (Parker and Smith, 2010), a Ca2+ indicator (Cal520) and ci-IP3. Cells were washed with HEPES-buffered saline (HBS; 135 mM NaCl, 5.9 mM KCl, 1.2 mM MgCl2, 1.5 mM CaCl2, 11.5 mM glucose, 11.6 mM HEPES pH 7.3) and then incubated (20°C in darkness) in HBS containing pluronic acid F-127 (0.02% w/v), Cal520-AM (5 µM) and ci-IP3-PM (1 µM). After 1 h, cells were washed and incubated in HBS containing pluronic acid F-127 and EGTA-AM (5 µM). After 45 min, the medium was replaced with HBS. After a further 30 min, the cells in HBS at 20°C, were used for imaging. The same method, but with NP-EGTA-AM (1 μM) replacing ci-IP3 PM, was used to load cells with caged EGTA (Fig. S1A,B).

An Olympus IX83 inverted microscope equipped with 100× oil-immersion TIRF objective (Olympus UApo N; numerical aperture, NA=1.49) was used for measurements of local Ca2+ signals by using total internal reflection fluorescence microscopy (TIRFM). Excitation (488 nm) was provided by a diode-pumped solid-state laser (150 mW, iLas Laser System, Cairn) and a band-pass filter (ET-405/488/561/643 quad band filter set, Chroma). The angle of the excitation beam was adjusted to achieve TIRF with a penetration depth of ∼90 nm. Emitted light was captured through a band-pass filter (Cairn Optospin, peak/bandwidth 525/50 nm) using an Andor iXon 897 EMCCD camera (image size: 120×120 pixels, 16 µm×16 µm per pixel; each image pixel had dimensions of 160 nm×160 nm). To achieve high temporal resolution, images were captured (188 Hz) from only a part of each cell (19.2 µm×19.2 µm, ∼25% of its TIRF footprint; see Fig. S2B,C) by streaming directly into RAM. After visualisation using Metamorph (Molecular Devices, Sunnyvale, CA), images were exported as tif files. Images were captured for 40 s before and after flash photolysis of ci-IP3.

Photolysis of ci-IP3 or NP-EGTA within the entire imaging field was achieved using a light-emitting diode (LED, 395 nm, Spectra Lumencor, Beaverton, OR), with the duration of exposure (50 ms, for ci-IP3; 300 or 1000 ms for NP-EGTA) determined by a shutter controlled within Metamorph. The conditions used for flash photolysis of ci-IP3 were optimised to allow identical stimulation of all cell lines, with each line providing sufficient Ca2+ puffs for quantitative analysis within a practicable recording interval (≤40 s).

Detection and analysis of Ca2+ puffs

We confirmed that the recording areas were similar for the four cell lines examined (Fig. S2C). Images were processed in Fiji (Schindelin et al., 2012). Images were corrected for background by subtraction of the fluorescence recorded from a region of interest (ROI) outside the cell. For automated analysis of Ca2+ puffs, the records for 2.7 s before and 40 s after photolysis were loaded into the image-processing program flika (Ellefsen et al., 2014) (https://github.com/flika-org/flika) that identifies clusters of pixels within which fluorescence changes exceed a critical threshold. From each recording, the last 500 frames (∼2.7 s) before photolysis were averaged to provide an F0 value and its standard deviation (s.d.) for each pixel. Image stacks of [(F/F0 )−1]/s.d. were then created and Gaussian-filtered, and pixels that exceeded a critical value (0.8 for our analysis) were located. A threshold-cluster algorithm (Rodriguez and Laio, 2014) then identified cluster centres by comparing them with adjacent bright pixels. Flika then identified the frames within which the event occurred and 100 frames either side of it. Fluorescence values of each pixel within a ROI (1.76×1.76 μm) centred on the brightest pixel were then averaged for each frame and a 2D Gaussian function was fitted to locate the centre of each Ca2+ puff. We confirmed, by manual inspection, the reliability and effectiveness of the automated identification of Ca2+ puffs. Flika then catalogued properties of the Ca2+ puffs, and assigned each Ca2+ puff to a Ca2+ release site. In our analysis, we considered Ca2+ puffs to arise at different sites when the centroids of the Ca2+ puffs were more than 0.96 μm apart.

To estimate the number of active IP3Rs contributing to a Ca2+ puff (N), we measured the unitary steps during the decay phase of 40 Ca2+ puffs (ten from each cell line, Fig. 4B). The average value of this unitary step (0.101±0.002) was used to estimate N from: N = (ΔF/0.101), where ΔF = Fpeak−Fpre. Fpeak and Fpre are, respectively, the F/F0 values determined at the peak of the Ca2+ puff and the average from the same region for ten frames before and after the puff.

Statistical analyses

Results are presented as means±s.e.m. unless stated otherwise, with n usually referring to the number of cells analysed. Statistical analyses used two-tailed Mann–Whitney tests or, for multiple comparison, the Kruskal–Wallis test with Dunn's correction for multiple comparisons (PRISM version 6, GraphPad, CA). Statistical significance is denoted by: *P<0.05, **P<0.01, ***P<0.001.

Acknowledgements

We thank David Yule (University of Rochester, NY) for generously making HEK cell lines lacking IP3Rs available, Kyle Ellefsen (University of California Irvine, CA) for advice and making the flika algorithm available, and Nagendra Babu Thillaiappan (University of Cambridge) for advice relating to microscopy.

Footnotes

Author contributions

Conceptualization: C.W.T.; Methodology: S.M., C.W.T.; Formal analysis: S.M., C.W.T.; Investigation: S.M., C.W.T.; Resources: S.M.; Data curation: S.M.; Writing - original draft: S.M., C.W.T.; Writing - review & editing: S.M., C.W.T.; Supervision: C.W.T.; Project administration: C.W.T.; Funding acquisition: C.W.T.

Funding

This work was supported by the Wellcome Trust [grant number 101844] and by a studentship to S.M. from the Biotechnology and Biological Sciences Research Council. Deposited in PMC for immediate release.

References

Alzayady
,
K. J.
,
Wang
,
L.
,
Chandrasekhar
,
R.
,
Wagner
,
L. E.
, II
,
Van Petegem
,
F.
and
Yule
,
D. I.
(
2016
).
Defining the stoichiometry of inositol 1,4,5-trisphosphate binding required to initiate Ca2+ release
.
Sci. Signal.
9
,
ra35
.
Berridge
,
M. J.
(
2016
).
The inositol trisphosphate/calcium signaling pathway in health and disease
.
Physiol. Rev.
96
,
1261
-
1296
.
Dakin
,
K.
and
Li
,
W.-H.
(
2007
).
Cell membrane permeable esters of d-myo-inositol 1,4,5-trisphosphate
.
Cell Calcium
42
,
291
-
301
.
Ellefsen
,
K. L.
,
Settle
,
B.
,
Parker
,
I.
and
Smith
,
I. F.
(
2014
).
An algorithm for automated detection, localization and measurement of local calcium signals from camera-based imaging
.
Cell Calcium
56
,
147
-
156
.
Futatsugi
,
A.
,
Nakamura
,
T.
,
Yamada
,
M. K.
,
Ebisui
,
E.
,
Nakamura
,
K.
,
Uchida
,
K.
,
Kitaguchi
,
T.
,
Takahashi-Iwanaga
,
H.
,
Noda
,
T.
,
Aruga
,
J.
, et al. 
(
2005
).
IP3 receptor types 2 and 3 mediate exocrine secretion underlying energy metabolism
.
Science
309
,
2232
-
2234
.
Hisatsune
,
C.
and
Mikoshiba
,
K.
(
2017
).
IP3 receptor mutations and brain diseases in human and rodents
.
J. Neurochem.
141
,
790
-
807
.
Iwai
,
M.
,
Michikawa
,
T.
,
Bosanac
,
I.
,
Ikura
,
M.
and
Mikoshiba
,
K.
(
2007
).
Molecular basis of the isoform-specific ligand-binding affinity of inositol 1,4,5-trisphosphate receptors
.
J. Biol. Chem.
282
,
12755
-
12764
.
Keebler
,
M. V.
and
Taylor
,
C. W.
(
2017
).
Endogenous signalling pathways and caged-IP3 evoke Ca2+ puffs at the same abundant immobile intracellular sites
.
J. Cell Sci.
130
,
3728
-
3739
.
Marchant
,
J. S.
and
Parker
,
I.
(
1998
).
Kinetics of elementary Ca2+ puffs evoked in Xenopus oocytes by different Ins(1,4,5)P3 receptor agonists
.
Biochem. J.
334
,
505
-
509
.
Miyakawa
,
T.
,
Maeda
,
A.
,
Yamazawa
,
T.
,
Hirose
,
K.
,
Kurosaki
,
T.
and
Iino
,
M.
(
1999
).
Encoding of Ca2+ signals by differential expression of IP3 receptor subtypes
.
EMBO J.
18
,
1303
-
1308
.
Nakamura
,
H.
,
Bannai
,
H.
,
Inoue
,
T.
,
Michikawa
,
T.
,
Sano
,
M.
and
Mikoshiba
,
K.
(
2012
).
Cooperative and stochastic calcium releases from multiple calcium puff sites generate calcium microdomains in intact Hela cells
.
J. Biol. Chem.
287
,
24563
-
24572
.
Parker
,
I.
and
Smith
,
I. F.
(
2010
).
Recording single-channel activity of inositol trisphosphate receptors in intact cells with a microscope, not a patch clamp
.
J. Gen. Physiol.
136
,
119
-
127
.
Prole
,
D. L.
and
Taylor
,
C. W.
(
2016
).
Inositol 1,4,5-trisphosphate receptors and their protein partners as signalling hubs
.
J. Physiol.
594
,
2849
-
2866
.
Rodriguez
,
A.
and
Laio
,
A.
(
2014
).
Machine learning. Clustering by fast search and find of density peaks
.
Science
344
,
1492
-
1496
.
Schindelin
,
J.
,
Arganda-Carreras
,
I.
,
Frise
,
E.
,
Kaynig
,
V.
,
Longair
,
M.
,
Pietzsch
,
T.
,
Preibisch
,
S.
,
Rueden
,
C.
,
Saalfeld
,
S.
,
Schmid
,
B.
, et al. 
(
2012
).
Fiji: an open-source platform for biological-image analysis
.
Nat. Methods
9
,
676
-
682
.
Smith
,
I. F.
and
Parker
,
I.
(
2009
).
Imaging the quantal substructure of single IP3R channel activity during Ca2+ puffs in intact mammalian cells
.
Proc. Natl. Acad. Sci. USA
106
,
6404
-
6409
.
Smith
,
I. F.
,
Wiltgen
,
S. M.
and
Parker
,
I.
(
2009a
).
Localization of puff sites adjacent to the plasma membrane: functional and spatial characterization of Ca2+ signaling in SH-SY5Y cells utilizing membrane-permeant caged IP3
.
Cell Calcium
45
,
65
-
76
.
Smith
,
I. F.
,
Wiltgen
,
S. M.
,
Shuai
,
J.
and
Parker
,
I.
(
2009b
).
Ca2+ puffs originate from preestablished stable clusters of inositol trisphosphate receptors
.
Sci. Signal.
2
,
ra77
.
Taylor
,
C. W.
,
Genazzani
,
A. A.
and
Morris
,
S. A.
(
1999
).
Expression of inositol trisphosphate receptors
.
Cell Calcium
26
,
237
-
251
.
Thillaiappan
,
N. B.
,
Chavda
,
A. P.
,
Tovey
,
S. C.
,
Prole
,
D. L.
and
Taylor
,
C. W.
(
2017
).
Ca2+ signals initiate at immobile IP3 receptors adjacent to ER-plasma membrane junctions
.
Nat. Commun.
8
,
1505
.
Thomas
,
D.
,
Lipp
,
P.
,
Tovey
,
S. C.
,
Berridge
,
M. J.
,
Li
,
W.
,
Tsien
,
R. Y.
and
Bootman
,
M. D.
(
2000
).
Microscopic properties of elementary Ca2+ release sites in non-excitable cells
.
Curr. Biol.
10
,
8
-
15
.
Tovey
,
S. C.
,
Dedos
,
S. G.
,
Taylor
,
E. J. A.
,
Church
,
J. E.
and
Taylor
,
C. W.
(
2008
).
Selective coupling of type 6 adenylyl cyclase with type 2 IP3 receptors mediates direct sensitization of IP3 receptors by cAMP
.
J. Cell Biol.
183
,
297
-
311
.
Vais
,
H.
,
Foskett
,
J. K.
and
Daniel Mak
,
D.-O.
(
2010
).
Unitary Ca2+ current through recombinant type 3 InsP3 receptor channels under physiological ionic conditions
.
J. Gen. Physiol.
136
,
687
-
700
.
Wei
,
C.
,
Wang
,
X.
,
Chen
,
M.
,
Ouyang
,
K.
,
Song
,
L.-S.
and
Cheng
,
H.
(
2009
).
Calcium flickers steer cell migration
.
Nature
457
,
901
-
905
.
Wojcikiewicz
,
R. J. H.
(
1995
).
Type I, II and III inositol 1,4,5-trisphosphate receptors are unequally susceptible to down-regulation and are expressed in markedly different proportions in different cell types
.
J. Biol. Chem.
270
,
11678
-
11683
.

Competing interests

The authors declare no competing or financial interests.

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.

Supplementary information