Direct Binding Between Orai1 and AC8 Mediates Dynamic Interplay Between Ca2+ and cAMP Signaling

Sci. Signal., 10 April 2012
Vol. 5, Issue 219, p. ra29
DOI: 10.1126/scisignal.2002299

Direct Binding Between Orai1 and AC8 Mediates Dynamic Interplay Between Ca2+ and cAMP Signaling

  1. Debbie Willoughby1,
  2. Katy L. Everett1,
  3. Michelle L. Halls1,*,
  4. Jonathan Pacheco2,
  5. Philipp Skroblin3,4,
  6. Luis Vaca2,
  7. Enno Klussmann4, and
  8. Dermot M. F. Cooper1,
  1. 1Department of Pharmacology, University of Cambridge, Cambridge CB2 1PD, UK.
  2. 2Departamento de Biología Celular, Instituto de Fisiología Celular, Universidad Nacional Autónoma de México, Ciudad Universitaria, Mexico 04510, Distrito Federal, México.
  3. 3Leibniz-Institut für Molekulare Pharmakologie, 13125 Berlin, Germany.
  4. 4Max-Delbrück-Centrum für Molekulare Medizin, 13125 Berlin, Germany.
  1. To whom correspondence should be addressed. E-mail: dmfc2{at}cam.ac.uk
  • * Present address: Drug Discovery Biology, Monash Institute of Pharmaceutical Sciences and Department of Pharmacology, Monash University, Parkville, Victoria 3052, Australia.

Abstract

The interplay between calcium ion (Ca2+) and cyclic adenosine monophosphate (cAMP) signaling underlies crucial aspects of cell homeostasis. The membrane-bound Ca2+-regulated adenylyl cyclases (ACs) are pivotal points of this integration. These enzymes display high selectivity for Ca2+ entry arising from the activation of store-operated Ca2+ (SOC) channels, and they have been proposed to functionally colocalize with SOC channels to reinforce crosstalk between the two signaling pathways. Using a multidisciplinary approach, we have identified a direct interaction between the amino termini of Ca2+-stimulated AC8 and Orai1, the pore component of SOC channels. High-resolution biosensors targeted to the AC8 and Orai1 microdomains revealed that this protein-protein interaction is responsible for coordinating subcellular changes in both Ca2+ and cAMP. The demonstration that Orai1 functions as an integral component of a highly organized signaling complex to coordinate Ca2+ and cAMP signals underscores how SOC channels can be recruited to maximize the efficiency of the interplay between these two ubiquitous signaling pathways.

Introduction

Signaling via Ca2+ in nonexcitable cells is typically initiated by the inositol 1,4,5-trisphosphate (IP3)–mediated release of Ca2+ from the endoplasmic reticulum (ER) Ca2+ stores in response to G protein (heterotrimeric guanosine triphosphate–binding protein)–coupled receptor stimulation of phospholipase C. Subsequent depletion of these stores triggers the entry of Ca2+ through store-operated Ca2+ (SOC) entry (SOCE) channels. SOCE, attributed to the activation of SOC channels within the plasma membrane, plays a critical role in the control of a number of cellular functions (1). Orai1, a 33-kD protein with four transmembrane domains found in numerous cell types, has been identified as the pore-forming subunit of SOC channels [including the well-characterized Ca2+ release–activated Ca2+ (CRAC) channel] (24). Clinically, mutations of the human Orai1 gene have been linked to severe immunodeficiency and various myopathies (5). The highly Ca2+-selective SOC channels, formed by a tetrameric assembly of Orai1 subunits (68), are activated by the ER Ca2+ sensor, stromal interaction molecule 1 (STIM1) (9, 10), upon ER Ca2+ store depletion and clustering of STIM1 at junctions between the ER and the plasma membrane (11, 12).

Of the nine membrane-bound adenylyl cyclases (ACs), the enzymes that convert adenosine triphosphate to cyclic adenosine monophosphate (cAMP), four isoforms respond to submicromolar Ca2+ concentrations in vitro (13). These Ca2+-sensitive ACs are found in various tissues and modulate such processes as cardiac rhythmicity, pulsatile hormone release, and hippocampal memory formation (1416). AC1 and AC8 are stimulated through an interaction with Ca2+-calmodulin (CaM) (1721), whereas AC5 and AC6 are directly inhibited by Ca2+ (22, 23). In cells, Ca2+-sensitive ACs show a marked selectivity for SOCE over other modes of increasing cytosolic Ca2+ (24, 25), even in excitable cells, and they reside in cholesterol- and sphingolipid-enriched domains of the plasma membrane known as lipid rafts, unlike their Ca2+-insensitive congeners (1, 26).

The role of Orai1 and STIM1 in mediating the selective regulation of ACs has been demonstrated in cultured human embryonic kidney (HEK) 293, pancreatic, and colonic cell lines, where their overexpression potentiates the Ca2+-dependent stimulation or inhibition of AC8 and AC6, respectively (27). STIM1-dependent Ca2+ store depletion has also been linked to AC activation, independently of any change in cytosolic Ca2+ (28). Data obtained with a genetically encoded Ca2+ sensor fused to AC8 (GCaMP2-AC8) indicated that the Ca2+-stimulable AC8 resides in a discrete Ca2+ microdomain that experiences rapid fluctuations in Ca2+ during SOCE, but is shielded from other modes of Ca2+ increase (such as Ca2+ release from the ER or ionomycin-mediated Ca2+ entry) (29). These findings support the existence of a proposed functional colocalization between AC8, Orai1, and STIM1 in the plasma membrane (25, 27).

Here, we used a multidisciplinary approach involving fluorescence resonance energy transfer (FRET), glutathione S-transferase (GST) pulldown, coimmunoprecipitation, and peptide array analyses to identify a direct protein-protein interaction between AC8 and Orai1. We used live-cell imaging with high-resolution AC-targeted Ca2+ and cAMP biosensors, together with small interfering RNA (siRNA) knockdown of Orai1, to demonstrate a role for Orai1 in generating dynamic changes in Ca2+ concentration within the AC8 microdomain to stimulate cAMP production. Comparisons were made with an N-terminally truncated form of AC8, 8M1, which was unable to bind Orai1 directly. Our data suggest that SOC channels can function as integral components of highly organized signaling complexes to facilitate the efficient and dynamic regulation of Ca2+-dependent cAMP signaling events.

Results

FRET reveals an association between AC8 and Orai1

We used FRET to assess direct, or indirect, interactions between fluorescently tagged AC8 and Orai1 constructs in plasma membrane regions of live HEK293 cells. Analysis of FRET images, corrected for bleedthrough, with micro-FRET (see Materials and Methods) produced a normalized corrected FRET (NFRETc) value of 6.5 ± 0.6 (× 105) in cells expressing yellow fluorescent protein (YFP)–tagged AC8 (YFP-AC8) and cyan fluorescent protein (CFP)–tagged Orai1 (Orai1-CFP), consistent with a protein-protein interaction between AC8 and Orai1 (Fig. 1A). We observed a similar degree of FRET [5.9 ± 0.6 (× 105)] when Orai1-CFP was coexpressed with 8Tm1-YFP-Tm2 [a construct containing the full-length N terminus and transmembrane domains of AC8 but lacking much of its catalytic C1 and C2 domains, with YFP sandwiched between AC8 residues 397 and 654 (30), see fig. S1]. The observation of FRET between Orai1-CFP and the catalytically inactive 8Tm1-YFP-Tm2 construct suggests that the interaction of Orai1 with AC8 is unaffected by cAMP signaling. In contrast, FRET between Orai1-CFP and YFP-8M1 [a fluorescently tagged N-terminally truncated form of AC8 that lacks the first 106 residues] was markedly reduced [1.1 ± 0.2 (× 105); P < 0.0001 compared to FRET between Orai1-CFP and YFP-AC8]. This loss of FRET signal suggests that the N terminus of AC8, which is critical for AC8 regulation by SOCE (31), was also required for AC8 and Orai1 interaction.

Fig. 1

FRET and LG-TIRFM imaging of the interaction between Orai1 and AC8. (A) HEK293 cells coexpressing CFP-tagged (green) and YFP-tagged (red) constructs for AC8, Orai1, and STIM1 (see fig. S1A for depictions of the constructs). Overlay images show colocalization of all construct pairings. Pseudocolor images represent corrected FRET (FRETc) in the presence of 1 μM thapsigargin (Tg). Bar graphs indicate normalized FRET efficiencies (NFRETc) for each pairing ± Tg treatment. Bars represent means ± SEM; n values for cells analyzed in parentheses. ***P < 0.001. n.s., not significant. (B) LG-TIRFM images from cells coexpressing YFP-AC8 or YFP-8M1 and Orai1-CFP, colocalization in yellow (overlay). FRET efficiency shown in pseudocolor with bar graph of mean ± SEM values for all cells analyzed. ***P < 0.001. (C) Representative LG-TIRFM images of 21 cells (from three independent transfections) exposed to Tg ± methyl-β-cyclodextrin (MβCD) for 15 min. Pixel colocalization for the AC8-Orai1 FRET signal (% FRET) and lipid raft marker, cholera toxin B subunit (CT-B), is shown in correlation plots. The numbers of pixels in different areas of the plot are represented by a pseudocolor scale, with yellow representing high pixel number.

The degree of FRET between Orai1 and all AC8 constructs was unchanged after store depletion with the ER Ca2+ uptake inhibitor thapsigargin (Tg) (Fig. 1A). This observation implies that the association of Orai1 with AC8 is constitutive and independent of the activation of Orai1 that is evoked by Ca2+ store depletion and the formation of Orai1-STIM1 puncta (11, 12). We observed a robust FRET signal between Orai1-CFP and STIM1-YFP after store depletion [12.5 ± 1.1 (× 105); P < 0.0001 compared to without Tg]. FRET analysis revealed some colocalization between fluorescently tagged AC8 and STIM1 constructs after Tg-induced store depletion; however, the FRET signal was negligible compared to that between Orai1-CFP and STIM1-YFP (Fig. 2, A, C, and D). Coexpression of fluorescently tagged AC8 and STIM1 with a nonfluorescent Orai1 construct (Orai1-myc) enhanced the association of STIM1 with AC8 during Tg-induced SOCE [3.0 ± 1.0 (× 105); Fig. 2B], although FRET was significantly less than that between AC8 and Orai1 (Fig. 1A; P < 0.001). These data suggest that AC8 associates with the pore-forming Orai1 and not with its binding partner, STIM1.

Fig. 2

FRET analysis reveals limited interaction between AC8 and STIM1. (A) HEK293 cells coexpressing either YFP-AC8 and STIM1-CFP or CFP-AC8 and STIM1-YFP. Overlay images show colocalization of construct pairings, and pseudocolor images represent corrected FRET (FRETc) images. Images were collected in either the absence or the presence of 1 μM thapsigargin (Tg). (B) Coexpression of fluorescently tagged AC8 and STIM1 with Orai1-myc enhanced the association of STIM1 with AC8 during Tg-induced SOCE. (C) A clear FRET signal is detected between Orai1-CFP and STIM1-YFP as a consequence of 1 μM Tg treatment. (D) Bar chart of normalized FRET efficiencies (NFRETc) for each pairing ± Tg treatment. Bars represent means ± SEM, n values for number of cells analyzed in parentheses. ANOVA analysis showed no significant FRET between AC8 and STIM1; however, one-tailed t tests comparing the effects of Tg treatment on AC8 and STIM1 association indicated a modest degree of FRET when coexpressed with Orai1-myc. ^P < 0.05 compared to without Tg.

High-resolution light guide–based total internal reflection fluorescence microscopy (LG-TIRFM) detects fluorescence events (such as FRET) that occur within or near (30 to 100 nm) the cell surface and is well suited to the study of lipid raft dynamics in living cells (32). Consistent with our micro-FRET data, LG-TIRFM revealed substantial FRET between YFP-AC8 and Orai1-CFP, with a FRET efficiency of 20.28 ± 0.85 (Fig. 1B). In contrast, the FRET efficiency between Orai1-CFP and the N-terminally truncated AC construct YFP-8M1 was only 5.13 ± 0.60 (P < 0.0001 compared to FRET between Orai1-CFP and YFP-AC8). Distribution of the YFP-AC8 and Orai1-CFP FRET signal was nonuniform within the plasma membrane after Tg stimulation (Fig. 1C) and showed a high degree of colocalization with a fluorescently tagged form of cholera toxin B subunit (CT-B), a marker for lipid rafts that binds to GM1 gangliosides (32, 33). Furthermore, disruption of lipid rafts by extraction of cholesterol with methyl-β-cyclodextrin (MβCD) markedly reduced the Orai1-AC8 FRET signal (Fig. 1C), an observation that supports previous evidence that the integrity of lipid rafts is crucial to activation of AC8 by SOCE (27, 31) and indicates that the association of AC8 with Orai1 occurs in these domains. As reported previously (34), FRET between Orai1 and CT-B in wild-type HEK293 cells increased markedly after store depletion (fig. S2).

AC8 and Orai1 interact via their N termini

To determine whether AC8 bound directly to Orai1 and to identify the regions of AC8 involved in this interaction, we tagged cytosolic fragments of AC8 with GST and used them to pull down endogenous Orai1 from HEK293 cell lysates (Fig. 3A and fig. S3). We examined three cytosolic regions of AC8 that are highly divergent in length and sequence between AC isoforms (13) as possible Orai1 binding sites: the N terminus, the C1 domain, and the C2 domain (fig. S1B). In agreement with our FRET data, GST pulldowns identified an association between Orai1 and GST-AC8 1–179, a tagged form of the full-length N terminus of AC8 (Fig. 3A, lane 3; P < 0.001 compared to GST alone), whereas the tagged C1 and C2 domains failed to associate with Orai1 (Fig. 3A, lanes 6 to 8). GST pulldowns showed no association between Orai1 and AC8 fragments containing only residues 1 to 77 or 73 to 179, which suggested that Orai1 binding spans both halves of the N terminus of AC8 or requires an intact secondary structure in this region (Fig. 3A, lanes 4 and 5). The interaction between Orai1 and GST-AC8 1–179 seen in Ca2+-free conditions persisted in the presence of Ca2+-CaM, which binds to an amphipathic helix of AC8 between residues 34 and 51 (20, 35). GST-AC8 1–179 bound to STIM1 in Ca2+-free conditions (Fig. 3B, P < 0.001 compared to GST alone), although this latter interaction was lost in the presence of Ca2+-CaM (which could be a consequence of either the recruitment of CaM to the N terminus of AC8 or the altered conformation of STIM1 induced by Ca2+ binding to its C-terminal EF hand). Affinity coimmunoprecipitation with hemagglutinin (HA)–tagged AC8 confirmed that both Orai1 and STIM1 interacted with full-length AC8 (AC8-HA) (Fig. 3, C and D) but not with the N-terminally truncated form (8M1-HA) (Fig. 3, C and D). These data support the hypothesis that Orai1 associates with the N terminus of AC8 and suggest that the AC8 N terminus also associates with STIM1 under these experimental conditions to form a multimolecular signaling complex.

Fig. 3

Binding between the N termini of Orai1 and AC8. (A) GST pulldowns from HEK293 cell lysate assess interaction of cytosolic regions of AC8 with endogenous Orai1 in the absence or presence of added Ca2+ (CaM). Exemplar blots of five similar experiments are shown. Densitometry plots represent means ± SEM, normalized to the intensity of GST bands (see fig. S3A). ***P < 0.001 versus GST alone. (B) GST pulldowns to assess AC8 binding to endogenous STIM1 in four repeat experiments. The arrow indicates authentic STIM1. (C and D) Immunoprecipitation by AC8-HA, but not 8M1-HA, of endogenous Orai1 (C) or endogenous STIM1 (D). Densitometry plots represent means ± SEM from five repeats normalized to untransfected control cells expressing similar amounts of HA-tagged protein (see fig. S3B). **P < 0.01; *P < 0.05 compared to HEK293. ^P < 0.05 compared to AC8-HA. (E) Orai1 knockdown at 48 to 72 hours after expression of Orai1-selective siRNA or scrambled siRNA (control). Bar chart quantifies the knockdown as means ± SEM from four repeats. *P < 0.05 compared to scrambled siRNA. (F) GST pulldown of Orai1 or STIM1 in cell lysates from HEK293 cells with the full-length AC8 N terminus (GST-AC8 1–179). The association of GST-AC8 1–179 with Orai1, but not STIM1, is lost after siRNA knockdown of Orai1. Densitometry plots represent means ± SEM from four repeat experiments, normalized to GST alone. ***P < 0.001; *P < 0.05 compared to GST alone. ^^P < 0.01 compared to scrambled controls.

Orai1-selective siRNA reduced Orai1 abundance in HEK293 cells by 68 ± 10% compared to that in control cells treated with scrambled siRNA (Fig. 3E). This was accompanied by a significant decrease in the pulldown of Orai1 by GST-AC8 1–179 (Fig. 3F), confirming the selectivity of the interaction between Orai1 and the N terminus of AC8. The association of AC8 with STIM1 was retained (Fig. 3F), suggesting that the interaction between STIM1 and the N terminus of AC8 (Fig. 3, B and D) is unlikely to be an indirect consequence of the ability of both proteins to bind Orai1. Moreover, Orai1 knockdown did not affect STIM1 abundance (fig. S4).

We used peptide array analysis to confirm that the interaction between AC8 and Orai1 was direct and to identify the region of Orai1 that was responsible for any interaction with the N terminus of AC8. A library of overlapping peptides (25-mers) from the cytoplasmic regions of Orai1 (residues 1 to 87, 141 to 173, and 256 to 301) was immobilized on cellulose membranes and probed for interactions with GST alone, GST-AC8 1–179, or GST-AC8 1–77 (Fig. 4A). Positive interactions (represented by dark spots) were seen for peptide spots A2 to A6, which bound to GST-AC8 1–179 and to a lesser extent to GST-AC8 1–77. Analysis of peptide spot sequences (fig. S5) identified Orai1 amino acids Gly26, Ser27, Arg28, Arg29, and Ser30 as the likely site of the AC8-Orai1 interaction. These five amino acids constitute a polybasic arginine-rich domain within the N terminus of Orai1 that is absent in Orai2 and Orai3 orthologs (36), suggesting that AC8 might associate selectively with Orai1 but not Orai2 and Orai3.

Fig. 4

Peptide array identifies an interaction between the N terminus of AC8 and an arginine-rich region within the N terminus of Orai1. (A) Peptide array data for overlapping sequences containing 25 residues from the cytosolic regions of Orai1 probed for interaction with GST-tagged AC8 N-terminal peptides. Protein binding detected with anti-GST and secondary streptavidin-HRP (POD)-conjugated anti-rabbit antibodies. Positive interactions were seen for peptide spots A2 to A6 containing Gly-Ser-Arg-Arg-Ser (GSRRS, residues 26 to 30) of the N terminus of Orai1. Data are representative of three experiments. (B) Double alanine substitutions introduced into the arginine-rich region of spot A4 (A) reduced binding of GST-tagged AC8 N-terminal peptides. Data are representative of five experiments. Sequences for all spots tested are provided in fig. S5. (C) A representative GST pulldown of wild-type or arginine-mutated MBP-Orai1 N terminus (residues 1 to 60) in cell lysate from HEK293 cells with the full-length AC8 N terminus (GST-AC8 1–179). The association of GST-AC8 1–179 with MBP-tagged Orai1 wild-type N terminus (MBP-Orai1 WT) is lost in the mutant form of Orai1 N terminus (MBP-Orai1 mutant) containing alanine substitutions of arginine residues 28, 29, 31, 32, and 33. Densitometry plots represent means ± SEM from four repeat experiments, normalized to input. *P < 0.05 compared to MBP alone; #P < 0.05 compared to MBP-Orai1 WT.

Next, we introduced single or double alanine substitutions into the peptide sequence used for spot A4 and determined the effect of these substitutions on AC8 N terminus binding. These data suggested that the arginines in this region made a major contribution to Orai1 interaction with the N terminus of AC8 (Fig. 4B and fig. S5). Although single alanine substitutions were without effect, substitution of Arg28 and Arg29 prevented binding of GST-AC8 1–179 and GST-AC8 1–77 (Fig. 4B, spot E4). Substitutions just downstream of this region, Arg31 and Arg32, Arg32 and Arg33, or Arg33 and Ser34 (represented by peptide spots E7 to E9), also led to a loss of AC8 binding.

We further assessed the importance of these arginine residues to the AC8-Orai1 interaction in pull-down assays, using GST-AC8 1–179 with either maltose binding protein (MBP)–tagged Orai1 N terminus (residues 1 to 60) or an MBP-tagged Orai1 N terminus in which alanines were substituted for Arg28, Arg29, Arg31, Arg32, and Arg33. Consistent with the peptide array data, the MBP-Orai1 wild-type N terminus bound to GST-AC8 1–179 (Fig. 4C, lane 2), and this association was lost when alanines were substituted for the five arginine residues (Fig. 4C, lane 3, P < 0.05 compared to wild-type Orai1 N terminus).

N-terminal truncation of Orai1 prevents its interaction with AC8

Truncation of the first 49 residues of the N terminus of Orai1—including the “arginine-rich domain”—produces a form of Orai1 that can still mediate functional SOCE (37). We generated CFP-tagged N-terminally truncated Orai1 constructs to determine whether removal of the residues identified as potential AC8 binding sites by our peptide array analysis and N-terminal AC8 and Orai1 pulldowns (Fig. 4) prevented FRET between Orai1 and AC8 (Fig. 5). All Orai1 truncations expressed well in HEK293 cells and showed good colocalization with AC8 at the plasma membrane (Fig. 5A). However, FRET was significantly reduced in Orai1 constructs lacking the first 40 (Orai1Δ40-CFP) or 60 residues (Orai1Δ60-CFP) (Fig. 5, A and B) (P < 0.001 compared to wild-type Orai1-CFP); thus, our data are consistent with Arg28 and Arg29 (and potentially Arg31, Arg32, Arg33, and Ser34) being important for interaction with AC8. In contrast, an Orai1 construct lacking just the first 20 N-terminal residues (Orai1Δ20-CFP) displayed FRET with AC8 comparable to that seen with wild-type Orai1 (Fig. 5, A and B). Consistent with the truncated Orai1 constructs retaining functionality, the tagged mutant Orai1 proteins all exhibited substantial FRET with STIM1-YFP after store depletion with 1 μM Tg (Fig. 5, A and B).

Fig. 5

Truncation of the N terminus of Orai1 reduces FRET between fluorescently tagged AC8 and Orai1 constructs. (A) Left: example images of HEK293 cells coexpressing CFP-tagged Orai1 constructs (green) and YFP-AC8 (red) ± 1 μM thapsigargin (Tg). Overlay images reveal colocalization of AC8 with all Orai1 constructs. Pseudocolor images, representing corrected FRET (FRETc), suggest interaction with AC8 and WT (full-length) Orai1, or Orai1 lacking the first 20 N-terminal residues (Orai1Δ20). FRET is reduced between AC8 and Orai1 N-terminal truncations lacking the arginine-rich domain (Orai1Δ40 and Orai1Δ60). Right: All CFP-tagged Orai1 truncation mutants interacted with STIM1-YFP after Tg treatment. (B) Normalized FRET efficiencies (NFRETc) for each pairing ± Tg treatment. Bars represent means ± SEM; n values range from 12 to 42, with data for wild-type Orai1 also plotted in Fig. 1A but used here for comparison purposes. ***P < 0.001. n.s., not significant. (C) Cartoons depicting tagged constructs used for FRET analysis of AC8 or STIM1 interaction with N-terminally truncated Orai1.

Orai1 dictates Ca2+ changes within the “AC8 microdomain”

To investigate the physiological consequences of a direct association between the SOC channel subunit Orai1 and the Ca2+-regulated AC8, we compared local Ca2+ changes within the immediate vicinities of AC8 and its N-terminally truncated form, 8M1 (which did not bind Orai1). We fused the green fluorescent protein (GFP)–based Ca2+ sensor GCaMP2 to the N terminus of AC8 or 8M1 to create targeted Ca2+ sensors that could monitor Ca2+ changes within each respective microdomain. Both AC8 and 8M1 are highly sensitive to stimulation by Ca2+ in vitro (20, 31) and reside in lipid rafts (31); however, 8M1 is insensitive to SOCE in the intact cell (20, 31). Both targeted sensors, GCaMP2-AC8 and GCaMP2-8M1, localized to the plasma membrane of HEK293 cells (Fig. 6A); however, they reported different Ca2+ profiles within their respective microenvironments. GCaMP2-AC8 reported a large rapid Ca2+ change during SOCE but showed poor sensitivity to Ca2+ signals arising from IP3-induced Ca2+ release from the ER (Fig. 6, B and C) (29). This preferential response to SOCE is apparent in a scatter plot showing the peak response of each cell tested to carbachol (CCh)–induced Ca2+ release from the ER in the absence of external Ca2+ and the subsequent SOCE after readdition of Ca2+ to the bath (Fig. 6D). In contrast, GCaMP2-8M1 reported a robust increase in Ca2+ during Ca2+ mobilization from the ER and a smaller, slower response to SOCE (Fig. 6, B, C, and E). Time to peak SOCE was 43.3 ± 1.8 s for GCaMP2-8M1 compared to 16.4 ± 2.0 s for GCaMP2-AC8 (P < 0.0001). These data suggest that interaction with Orai1 places AC8 in a specific cellular microdomain that experiences rapid changes in Ca2+ concentration during SOCE. Removal of the N terminus to produce 8M1, which largely precludes binding to Orai1, places the AC in a functionally distinct region of the plasma membrane.

Fig. 6

Orai1-dependent changes in Ca2+ concentration within AC8 and 8M1 microdomains. (A) Plasma membrane localization of GCaMP2-AC8 and GCaMP2-8M1 sensors in HEK293 cells. Scale bars, 20 μm. (B and C) Average changes in Ca2+ concentration detected with GCaMP2-AC8 (black) and GCaMP2-8M1 (blue) during 300 μM CCh–induced Ca2+ release (B) and subsequent SOCE (C). (D and E) Scatter plots of the peak responses of GCaMP2-AC8 and GCaMP2-8M1 to Ca2+ mobilization from the ER versus SOCE in all cells tested (n = 80 and 120, respectively). (F) Loss of SOCE after Orai1 knockdown monitored with cytosolic GCaMP2. Traces represent average responses from 26 and 21 cells for scrambled control and Orai1 siRNA, respectively. (G and H) Effects of Orai1 knockdown (blue) on CCh-induced Ca2+ signals measured by GCaMP2-AC8 (n = 36 for control, n = 16 for Orai1 siRNA) or GCaMP2-8M1 (n = 28 for control and Orai1 siRNA). (I) Effects of Orai1 knockdown on the peak Ca2+ increases detected by AC-targeted GCaMP2 sensors. *P < 0.05; **P < 0.01; ***P < 0.001 compared to scrambled controls. (J and K) Scatter plots showing the effects of Orai1 knockdown on GCaMP2-AC8 and GCaMP2-8M1 detection of Ca2+ mobilization from the ER versus SOCE in individual cells.

To investigate the role of Orai1 in controlling the AC8 microdomain, we used siRNA to knock down endogenous Orai1. Orai1 knockdown was confirmed by the ~90% loss of peak SOCE detected by cytosolic (nontargeted) GCaMP2 (Fig. 6F) and the ~75% decrease in peak SOCE detected by GCaMP2-AC8 (Fig. 6, G and I; P < 0.0001 compared to scrambled controls). Scatter plot analysis confirmed the dependence on Orai1 of Ca2+ entry within the AC8 microdomain (Fig. 6J). Orai1 knockdown decreased the slower Ca2+ entry reported by GCaMP2-8M1 by 45% (Fig. 6, H, I, and K; P < 0.001 compared to scrambled siRNA). It is possible that residual Orai1 may contribute to Ca2+ entry within the vicinity of 8M1. Furthermore, 8M1 may be exposed to nonspecific Ca2+ entry after addition of 2 mM external Ca2+. Nevertheless, a large portion of both the AC8-localized and the 8M1-localized Ca2+ entry signals was due to Orai1-containing channels, even though the two sensors showed different temporal profiles for Ca2+ and differing abilities to bind Orai1. These findings support the hypothesis that the Orai1-AC8 interaction establishes robust Ca2+ events within the immediate vicinity of AC8 to promote Ca2+-stimulated cAMP production.

Orai1 affects the activity and subcellular targeting of AC8

Next, we used the FRET-based cAMP sensor Epac2-camps (38) to determine the effects of Orai1-dependent Ca2+ signaling on AC8 activity (Fig. 7). As predicted from the dependence of AC8 activity on SOCE (24) and our evidence for AC8-Orai1 binding (Figs. 1 and 3), Orai1 knockdown in HEK-AC8 cells was accompanied by ~70% decrease in SOCE-dependent cAMP accumulation (Fig. 7, A and D; P < 0.0001). In contrast, Orai1 knockdown was accompanied by increased cAMP accumulation in response to CCh-induced ER Ca2+ release in the absence of external Ca2+ (Fig. 7, A, C, and D). HEK293 cells, like many cell types, contain multiple AC isoforms (including AC2, AC3, AC6, and AC7) that could potentially contribute to cellular cAMP (39). Global cAMP measurements in wild-type HEKs revealed no Ca2+-evoked increases in cAMP accumulation, consistent with their dependence on AC8 activity (Fig. 7A). To more specifically examine the dependence of AC8 activity on Orai1, we expressed Epac2-camps–AC8 [in which Epac2-camps is tethered to functional AC8 to monitor cAMP within the AC8 microdomain; (40)]. When expressed in wild-type HEK293 cells, this AC8-targeted cAMP sensor reported patterns of response similar to those reported by the global Epac2-camps sensor in HEK-AC8 cells (Fig. 7B), indicating that AC8 was the only source of the globally detected cAMP.

Fig. 7

Contribution of Orai1 to SOCE-mediated AC8 activity assessed with an AC8-targeted cAMP sensor. (A) Real-time changes in cAMP concentration in HEK-AC8 cells monitored with Epac2-camps. Traces represent average effects of CCh-induced Ca2+ release and subsequent SOCE under control conditions (black) and after Orai1 knockdown (blue). Data from equivalent experiments on wild-type HEK293 cells are also shown (dashed lines). (B) Parallel experiments with Epac2-camps–AC8 to monitor cAMP accumulation within the AC8 microdomain. (C) Scatter plots showing peak changes in cAMP accumulation detected globally (circles) or within the AC8 microdomain (triangles) in individual cells. (D) Peak cAMP increase in arbitrary (arb) units in response to Ca2+ release and SOCE. n = 21 to 26. **P < 0.01; **P < 0.001. (E) Plasma membrane localization of YFP-AC8 and Epac2-camps–AC8 in HEK293 cells ~48 hours after cotransfection with either scrambled siRNA or Orai1-selective siRNA. Scale bars, 20 μm. (F) GCaMP2-AC8 measurements of changes in Ca2+ concentration within the AC8 microdomain during ionomycin (IM)–induced Ca2+ entry. n = 43 control cells; n = 37 Orai1 siRNA–treated cells. ***P < 0.001. Cells were pretreated with Tg (200 nM) and traces were corrected for any contribution from SOCE by subtraction of data from parallel experiments with no IM addition (fig. S6A).

Epac2-camps–AC8 measurements revealed that Orai1 knockdown decreased SOCE-induced AC8 activity by ~60% (Fig. 7D). The limited effectiveness of the Orai1 knockdown on subcellular analyses of SOCE-induced AC8 activity suggests that despite ~90% loss of global Ca2+ entry (Fig. 6F), any remaining Orai1 associated with and enhanced the activation of AC8. This is consistent with the greater residual Ca2+ entry observed in the AC8 microdomain after Orai1 knockdown (Fig. 6I). The enhanced activation of AC8 by ER Ca2+ release detected after Orai1 knockdown (Fig. 7, B and D) is consistent with the increase in CCh-evoked Ca2+ release detected within the vicinity of AC8 (Fig. 6I).

We hypothesized that when AC8 is unable to bind Orai1 it may delocalize to a different region of the cell. In this scenario, AC8 could function like 8M1, which does not bind Orai1 and shows a larger response to Ca2+ release from the ER. Confocal imaging revealed localization of both Epac2-camps–AC8 and YFP-AC8 at the plasma membrane, before and after Orai1 knockdown (Fig. 7E). To address the possible delocalization of AC8 to a different plasma membrane Ca2+ microdomain when it cannot bind Orai1, we examined the responsiveness of GCaMP2-AC8 to ionophore-mediated Ca2+ entry. Typically, AC8 is shielded from ionophore-induced Ca2+ entry, as assessed with GCaMP2-AC8 (29). However, Orai1 knockdown enabled the detection of substantial Ca2+ entry (corrected for any contribution from SOCE, fig. S6A) by GCaMP2-AC8 in response to ionomycin treatment (Fig. 7F). Control experiments using the nontargeted Ca2+ sensor GCaMP2 confirmed that Orai1 knockdown did not enhance the degree of nonspecific Ca2+ entry seen globally in HEK293 cells (fig. S6B). These data thus suggest that the interaction of AC8 with Orai1 not only plays an essential role in maximizing the regulation of AC8 by SOCE but may also contribute to the targeting of AC8 to discrete regions of the plasma membrane that are shielded from other Ca2+ events, such as IP3-induced Ca2+ release or ionophore-mediated Ca2+ entry.

Discussion

The identification of Orai1 and STIM1 as key molecular components of SOCE has advanced understanding of this mode of Ca2+ entry and its role in numerous physiological and pathophysiological processes (41). Many of the physiological effects of SOCE are mediated by the interplay between Ca2+ and other signaling pathways, most notably the cAMP pathway (14). The likely importance of these interactions is emphasized by the selectivity for and dependence of the Ca2+-regulated ACs on SOCE (27, 29), which can produce dynamic, coordinated changes in Ca2+ and cAMP (42). Here, we reveal that the long-observed functional dependence of Ca2+-stimulable AC8 on SOCE originates from a direct interaction between AC8 and Orai1, the pore-forming unit of SOC channels. Our data from FRET studies, GST pulldowns, and coimmunoprecipitation provide compelling evidence for an interaction between the cytosolic N termini of AC8 and Orai1. Although the FRET studies suggest a close interaction between AC8 and Orai1, they do not eliminate the possibility that the interaction is mediated by a third, scaffolding, partner. We tend to favor the hypothesis that there is a direct interaction between the two proteins, most particularly because of the peptide array data. We identified an arginine-rich domain within the N terminus of Orai1, recently proposed to play a role in SOC current reactivation (43), as the complementary site of Orai1 required to bind the N terminus of AC8. Substitution of a minimum of two arginine residues in this region prevented interaction with AC8 in peptide array analyses. Furthermore, alanine substitution of the five arginine residues in this region of Orai1 eliminated its interaction with an AC8 N-terminal peptide in pull-down assays. These findings explain the earlier observation that the N terminus of AC8 is critical for its regulation by SOCE (31) and identify the N terminus of Orai1 as the reciprocal site of this interaction. Removal of the first 49 residues of Orai1 does not limit channel function (37) or the ability of Orai1 to interact with STIM1, but it is likely to affect Ca2+ and cAMP interplay. Previous work on NCM460 cells speculated on a scaffolding role for the ER Ca2+ sensor, STIM1, to facilitate recruitment and activation of AC5 or AC6 independently of changes in cytosolic Ca2+, although no evidence of direct binding was found (28). GST pull-down and coimmunoprecipitation data presented here suggest that some form of interaction between AC8 and STIM1 exists, but this is not supported by our FRET data.

Targeted Ca2+ sensors that monitored local Ca2+ changes within the “AC8-Orai1 microdomain” and “8M1 (N-terminally truncated AC8) microdomain” revealed marked consequences of removing the N terminus of AC8, thereby preventing its association with Orai1. We found that AC8 and 8M1 resided in distinct Ca2+ microdomains at the plasma membrane. The targeting of AC8 to a “privileged” Ca2+ microdomain selectively exposed to rapid Ca2+ increases during SOCE (29) relied on the interaction between AC8 and Orai1. Either knockdown of Orai1 or removal of the N terminus of AC8 resulted in AC8 becoming less discriminating with respect to the source of Ca2+ signal it detected. However, it is possible that when overexpressed, a small percentage of the 8M1 resides in the same microdomain as full-length AC8; this could account for the initial peak in Ca2+ signal detected by GCaMP2-8M1 during SOCE (Fig. 6C). Previous studies have shown that both AC8 and 8M1 reside in cholesterol- and sphingolipid-enriched lipid raft regions of the plasma membrane (31). Our LG-TIRFM data indicated a selective association of AC8 with Orai1 in such regions. We speculate that the loss of association of AC8 with Orai1 could result in some AC8 “delocalizing” to a different region of the plasma membrane. This could be a subdomain of lipid rafts exposed to different Ca2+ signals. Lipid rafts exhibit a clear heterogeneity in live cells and have been reported to undergo continual dynamic rearrangement of their protein components (32, 44). It thus seems possible that the ability of AC8 to recruit, or not, Orai1 could enable its residence in rafts of differing protein composition.

We conclude that an intimate and selective association between AC8 and Orai1 has evolved that ensures dynamic, coordinated changes in the concentrations of Ca2+ and cAMP and integrated downstream consequences. Given the direct association of AC8 with signaling proteins such as protein phosphatase 2A (PP2A) and protein kinase A anchoring proteins (AKAPs) (human AKAP79 and its rodent ortholog AKAP150) (35, 45), it is likely that AC8 and Orai1, along with STIM1, can be components of larger signaling complexes. The ability of Orai1 to exist in intimate partnership with a target protein from a distinct Gs-coupled signaling pathway reveals a novel action for Orai1 and explains the long-observed selectivity of Ca2+-sensitive ACs for SOCE over other modes of elevating cytosolic Ca2+ (24). STIM1 can dictate whether Ca2+ entry occurs through Cav1.2 or SOC channels (46, 47). Thus, both components of the SOCE pair can play a role in ensuring direct communication with cAMP signaling pathways in physiological systems where all are expressed.

Materials and Methods

Cell culture and transfection

HEK293 cells were grown in minimum essential medium (MEM) supplemented with 10% (v/v) fetal bovine serum, penicillin (100 U/ml), streptomycin (100 μg/ml), and 2 mM l-glutamine and maintained at 37°C in 95% air and 5% CO2. Cells were plated onto either 100-mm dishes or 25-mm poly-l-lysine–coated coverslips and transfected with Lipofectamine 2000 (Invitrogen) following the manufacturer’s instructions. For siRNA experiments, SMARTpool oligonucleotides directed against Orai1 (Dharmacon, Thermo Scientific), or scrambled siRNA control oligos were transfected at a final concentration of 25 nM. Cells were used 48 to 72 hours after transfection.

Micro-FRET analysis of protein-protein interactions

All FRET images were collected with a Nikon Eclipse TE2000 inverted microscope equipped with a 40× oil-immersion objective [numerical aperture (NA) 1.3] connected to an iXon+ EMCCD camera (Andor) via an Optosplit (505DC) (Cairn Research) to separate CFP (470 nm) and YFP (535 nm) emission images. Images were acquired and analyzed according to the three-cube micro-FRET method as described previously (45). For quantification of the degree of sensitized FRET seen between different pairings of CFP- and YFP-tagged constructs, average intensities of the FRETc images were divided by the product of the CFP and YFP image intensities to produce normalized (NFRETc), values (48).

LG-TIRFM analysis of FRET in lipid rafts

LG-TIRFM experiments were conducted with an Olympus IX81 inverted microscope equipped with an oil-immersion 100× Plan Apo objective (NA 1.45). The LG-TIRFM system (TIRF Technologies) was coupled to an argon-helium laser with three lines: 488, 543, and 633 nm (32). An Optosplit image splitter (Cairn Research) was positioned in front of the Andor iXon EMCCD camera to collect simultaneous images from all fluorophores as described previously (49). The CT-B conjugated to Alexa Fluor 594 (final concentration used, 5 μg/ml) was used in combination with YFP-AC8 and Orai1-CFP in FRET experiments. Cells were incubated for 5 min with CT-B to allow homogeneous labeling. In experiments with MβCD, cells were incubated for 15 min with 5 mM MβCD. A correlation between FRET efficiency and CT-B–decorated areas was obtained by mapping the FRET on the cell surface and determining the colocalization percentage with CT-B, as previously described (49). Pearson’s correlation coefficients were calculated with Huygens version 4 software (Scientific Volume Imaging). FRET efficiency was calculated with the sensitized emission method as previously described (50). Briefly, the amount of bleedthrough was determined by expressing the donor and acceptor separately. Both fluorescent proteins were excited with 488 and 543 lines, and channel crosstalk was measured. Before off line FRET determinations were conducted, all images were corrected for bleedthrough and background was subtracted with Huygens software. To determine the amount of basal FRET signal due to protein overexpression, we carried out control experiments in which the free form of CFP (donor) was expressed in combination with either YFP-AC8 or Orai-YFP, and FRET was calculated. This value gave FRET efficiency below 5% and was considered basal FRET.

Construction of YFP-8M1, GCaMP2-8M1, Orai1-CFP truncations, and MBP-Orai1 1–60

pEYFP-8M1 was produced by digesting pEGFP-8M1 (31) and pEYFP-C1 (Clontech) with Age I and Bgl II to remove the fluorophore from each plasmid. The YFP fragment was then ligated into the vector-8M1 fragment resulting in YFP-8M1. GCaMP2-8M1 was produced from pcDNA3.1zeo GCaMP2-AC8 (29) by use of the site-directed mutagenesis method to delete amino acids 1 to 106 from AC8 within the fusion protein. GCaMP2 and 8M1 were linked via the sequence GGSRGGG. Primer sequences were GGTTCTAGAGGAGGCGGGCCGGAACGCAGCGGGAGC (forward) and GCTCCCGCTGCGTTCCGGCCCGCCTCCTCTAGAACC (reverse). N-terminal truncations of Orai1-CFP were also generated by site-directed mutagenesis from pIRESneo Orai1-CFP. Primer sequences were as follows: Orai1Δ20-CFP, GATATCTGCGGCCGCATGGGCAGCACCACCAGCGGC (forward) and GCCGCTGGTGGTGCTGCCCATGCGGCCGCAGATATC (reverse); Orai1Δ40-CFP, GATATCTGCGGCCGCATGGGGGCCCCGCCACCG (forward) and CGGTGGCGGGGCCCCCATGCGGCCGCAGATATC (reverse); and Orai1Δ60-CFP, GATATCTGCGGCCGCATGTCCGAGGTGATGAGCCTC (forward) and GAGGCTCATCACCTCGGACATGCGGCCGCAGATATC (reverse). Residues 1 to 60 of Orai1 were amplified by polymerase chain reaction (PCR) and inserted into pMAL-p4G (New England Biolabs) between the Eco RI and the Xba I restriction sites. Primer sequences were GCGAATTCATGCATCCGGAGCCC (forward) and GCTCTAGACTAGTAACTCTGGCCGATCCAG (reverse). Residues Arg28, Arg29, Arg31, Arg32, and Arg33 were then mutated to Ala by site-directed mutagenesis. Primer sequences were CCACCAGCGGCAGCGCCGCGAGCGCCGCCGCCAGCGGGGACGGGG (forward) and CCCCGTCCCCGCTGGCGGCGGCGCTCGCGGCGCTGCCGCTGGTGG (reverse).

GST fragment purification and GST pull-down assays

Pull-down assays with HEK293 were carried out as described previously (45). Pulldowns were conducted in the presence of 100 μM EGTA or 10 μM Ca2+ to explore the influence of Ca2+-dependent endogenous CaM binding to the N terminus of AC8 (1820, 35) on the ability of the AC to interact with other proteins. MBP, MBP-Orai1 1–60, and MBP-Orai1 1–60 mutant were expressed in Rosetta2(DE3) cells for 2 hours at 37°C. Cells were lysed by sonication in phosphate-buffered saline containing 1 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, and 10 μM EDTA followed by addition of 0.2 or 1% Triton X-100, as indicated. Homogenates were then centrifuged at 26,000g for 15 min at 4°C, and the pellet was discarded. Supernatants were incubated with GST-AC8 1–179 for 6 hours at 4°C with rotation. Beads were then washed five times in sonication buffer containing 0.1% Triton X-100, and bound proteins were eluted by addition of an equivalent volume of 2× Laemmli buffer and boiling for 5 min before Western blot analysis. To elute GST-tagged proteins for peptide array analysis, we resuspended the washed protein–glutathione–Sepharose 4B resin in elution buffer [50 mM tris (pH 8) and 10 mM reduced glutathione] and rotated them for 10 min at room temperature. Supernatants were diluted in 50 mM tris (pH 8) and 150 mM NaCl and stored at 4°C.

Coimmunoprecipitation

For immunoprecipitation with anti-HA affinity–agarose beads (Roche), lysates from HEK293 cells expressing AC8-HA or 8M1-HA were prepared as described previously (45), in NP-40 solubilization buffer [50 mM tris (pH 7.4), 1 mM EDTA, 150 mM NaCl, 0.3% (v/v) NP-40, and protease inhibitors]. Cell lysates were rotated with 100 μl of prewashed bead slurry (50%) for 4 hours at 4°C. Beads were washed five times in NP-40 solubilization buffer. Bound proteins were eluted with NP-40 solubilization buffer supplemented with 1% (w/v) SDS. Laemmli buffer was added to the elution and incubated at 37°C for 30 min before Western blot analysis.

Western blot analysis

Proteins were resolved with 10% SDS–polyacrylamide gels. Separated proteins were transferred to a supported nitrocellulose membrane and blocked in tris-buffered saline (TBS) [20 mM tris (pH 7.5) and 150 mM NaCl] containing 5% (w/v) skimmed milk for 1 hour, followed by three 5-min washes in TBS supplemented with 0.05% (v/v) Tween-20 (TTBS). Membranes were incubated overnight at 4°C with anti-Orai1 antibody (1:200; Abcam), anti-STIM1 antibody (1:250; BD Biosciences), anti-MBP antibody (1:40,000; New England Biolabs), or anti-GST antibody (1:40,000; Sigma) in TTBS containing 1% (w/v) skimmed milk. After three 5-min washes in TTBS, membranes were incubated for 1 hour at room temperature with goat anti-mouse immunoglobulin G (IgG) conjugated to horseradish peroxidase (1:10,000 for anti-STIM1 and 1:20,000 for anti-MBP and anti-GST) or goat anti-rabbit IgG conjugated to horseradish peroxidase (1:20,000 for anti-Orai1) diluted in TTBS containing 5% (w/v) skimmed milk. Membranes were washed (three 5-min washes in TTBS, one 5-min wash in TBS) and immunoreactivity was detected with ECL Plus reagent (GE Healthcare) and exposure to film. Immunoreactive bands were quantified by densitometry with ImageJ.

Peptide array analysis

The cytoplasmic parts of human Orai1 (amino acids 1 to 87, 141 to 173, and 256 to 301) were synthesized as overlapping peptide spots (25-mers, shift 5). Peptide SPOT arrays were generated by automated SPOT synthesis as described previously (5153). To prevent dimerization or oxidization to sulfoxides, we substituted cysteine 143 in the original protein sequence of Orai1 with serine in the respective peptide (B5). To assay the binding of GST or GST-tagged fusion proteins, we blocked peptide spot membranes for 1 hour in blocking buffer (3% bovine serum albumin in TTBS) before incubation with 0.1 μM protein (GST, GST-AC8 1–179, or GST-AC8 1–77) overnight at 4°C. Detection was carried out as described previously (54) with Immobilon Western Chemiluminescent HRP Substrate (Millipore).

GCaMP2-AC8 and GCaMP2-8M1 measurements

Ca2+ measurements by use of GCaMP2-based constructs were performed as described previously (29). Data were normalized to the maximal GCaMP2 signal seen upon the addition of 10 mM Ca2+ plus 5 μM ionomycin.

Epac2-camps and Epac2-camps–AC8 FRET measurements

Epac2-camps– or Epac2-camps–AC8–expressing HEK293 cells were imaged as described previously (45). Data were normalized to the maximal FRET change seen upon stimulation of saturating cAMP concentration with a cocktail of 10 μM isoprenaline + 10 μM prostaglandin E1 + 10 μM forskolin + 100 μM 3-isobutyl-1-methylxanthine (IBMX).

Confocal imaging

Live-cell images were obtained with an inverted Zeiss LSM 510 confocal microscope with an oil-immersion 63× Plan Apo objective (NA 1.4). GCaMP2-AC8 and GCaMP2-8M1 sensors were excited at 488 nm, and emission was collected with a 505- to 550-nm band-pass filter. Parallel images of the CellMask Deep Red plasma membrane stain were obtained by excitation at 633 nm, and emission was collected at >650 nm. Image overlays were generated with MetaMorph imaging software (Molecular Devices). For YFP-AC8 and Epac2-camps–AC8 images, excitation was at 488 nm, and emission was collected with a 505- to 550-nm band-pass filter.

Statistical analysis

Unless stated otherwise, data were analyzed by one-way analysis of variance (ANOVA) followed by Tukey or Bonferroni multiple comparison tests (GraphPad Prism, GraphPad Software Inc.). Data are presented as means ± SEM, with significance set at P < 0.05.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/5/219/ra29/DC1

Fig. S1. Constructs used to study AC8 and Orai1 interaction.

Fig. S2. Colocalization between Orai1 and cholera toxin B subunit is enhanced following thapsigargin treatment.

Fig. S3. Relative levels of GST-tagged AC8 fragments.

Fig. S4. Specificity of Orai1 and STIM1 antibodies and Orai1 knockdown.

Fig. S5. Spot sequences for peptide array data presented in Fig. 4.

Fig. S6. Isolation of the ionomycin-induced, nonspecific Ca2+ entry signal in GCaMP2-AC8– and global GCaMP2–expressing cells.

  • Received 23 June 2011.
  • Accepted 20 March 2012.

References and Notes

Acknowledgments: Orai1-CFP, STIM1-YFP, and STIM1-CFP were gifts from D. L. Gill (Temple University School of Medicine, Philadelphia). Funding: This work was supported by the Wellcome Trust (RG31760). D.M.F.C. is a Royal Society Wolfson Research Fellow. Author contributions: D.W. and D.M.F.C. conceived the study, designed the experiments, and wrote the manuscript. D.W. conducted the FRET experiments, the confocal imaging, and the single-cell GCaMP2 and Epac2-camps experiments. K.L.E. generated the GCaMP2-8M1, truncated Orai1-CFP constructs, and MBP-tagged Orai1 constructs and performed some GST pulldowns and coimmunoprecipitations. M.L.H. performed several GST pulldowns and coimmunoprecipitations. J.P. and L.V. performed and analyzed the LG-TIRFM imaging experiments. P.S. and E.K. performed the peptide array analysis. Competing interests: The authors declare that they have no competing interests.

Citation:

D. Willoughby, K. L. Everett, M. L. Halls, J. Pacheco, P. Skroblin, L. Vaca, E. Klussmann, and D. M. Cooper, Direct Binding Between Orai1 and AC8 Mediates Dynamic Interplay Between Ca2+ and cAMP Signaling. Sci. Signal. 5, ra29 (2012).

STIM2 enhances receptor-stimulated Ca2+ signaling by promoting recruitment of STIM1 to the endoplasmic reticulum-plasma membrane junctions
H. L. Ong, L. B. de Souza, C. Zheng, K. T. Cheng, X. Liu, C. M. Goldsmith, S. Feske, and I. S. Ambudkar
Sci Signal 8, ra3-ra3 (13 January 2015)

TRPC1 contributes to the Ca2+-dependent regulation of adenylate cyclases
D. Willoughby, H. L. Ong, L. B. De Souza, S. Wachten, I. S. Ambudkar, and D. M. F. Cooper
Biochem. J. 464, 73-84 (15 November 2014)

Single-molecule analysis of diffusion and trapping of STIM1 and Orai1 at endoplasmic reticulum-plasma membrane junctions
M. M. Wu, E. D. Covington, and R. S. Lewis
Mol. Biol. Cell 25, 3672-3685 (5 November 2014)

Adenylate cyclase-centred microdomains
D. M. F. Cooper, and V. G. Tabbasum
Biochem. J. 462, 199-213 (1 September 2014)

Adenylyl cyclase 6 mediates loading-induced bone adaptation in vivo
K. L. Lee, D. A. Hoey, M. Spasic, T. Tang, H. K. Hammond, and C. R. Jacobs
FASEB J. 28, 1157-1165 (1 March 2014)

A key phosphorylation site in AC8 mediates regulation of Ca2+-dependent cAMP dynamics by an AC8-AKAP79-PKA signalling complex
D. Willoughby, M. L. Halls, K. L. Everett, A. Ciruela, P. Skroblin, E. Klussmann, and D. M. F. Cooper
J. Cell Sci. 125, 5850-5859 (1 December 2012)

Science Signaling. ISSN 1937-9145 (online), 1945-0877 (print). Pre-2008: Science's STKE. ISSN 1525-8882