Ribosome-free Terminals of Rough ER Allow Formation of STIM1 Puncta and Segregation of STIM1 from IP3 Receptors

Summary Store-operated Ca2+ entry is a ubiquitous mechanism that prevents the depletion of endoplasmic reticulum (ER) calcium [1]. A reduction of ER calcium triggers translocation of STIM proteins, which serve as calcium sensors in the ER, to subplasmalemmal puncta where they interact with and activate Orai channels ([2–8]; reviewed in [9]). In pancreatic acinar cells, inositol 1,4,5-trisphosphate (IP3) receptors populate the apical part of the ER. Here, however, we observe that STIM1 translocates exclusively to the lateral and basal regions following ER Ca2+ loss. This finding is paradoxical because the basal and lateral regions of the acinar cells contain rough ER (RER); the size of the ribosomes that decorate RER is larger than the distance that can be spanned by a STIM-Orai complex [5, 10], and STIM1 function should therefore not be possible. We resolve this paradox and characterize ribosome-free terminals of the RER that form junctions between the reticulum and the plasma membrane in the basal and lateral regions of the acinar cells. Our findings indicate that different ER compartments specialize in different calcium-handling functions (Ca2+ release and Ca2+ reloading) and that any potential interference between Ca2+ release and Ca2+ influx is minimized by the spatial separation of the two processes.


Immunostaining
For endogenous STIM1 staining cells were fixed with 4% paraformaldehyde for 30 minutes. After permeabilisation with 0.1% Triton X-100 nonspecific antibody binding was blocked with 10% goat serum and 1% BSA for 60 minutes at room temperature. Cells were stained with anti-STIM1 primary antibody (1:100) for 2 hours and anti-rabbit Alexa488 conjugated secondary antibody for 30 minutes at room temperature. Cover slips were mounted using ProlongGold (Invitrogen, Paisley, UK) mounting medium.
We and others [1] found that methanol fixation was the best protocol for immunostaining of IP 3 receptors. However using this fixation technique we have not been able to visualise the translocation of endogenous STIM1 with either commercial or non-commercial STIM1 antibodies. To investigate the relative positioning of IP 3 receptors and STIM1 we had to use cells transfected with STIM1-EYFP and antibodies against EYFP together with IP 3 -R antibodies.
Cells were fixed in 100% cold methanol at -20°C for 10 minutes. Following blocking procedure (see above) cells were incubated with primary antibodies against the proteins mentioned in the main text. Orai 1 was stained with a non-commercial antibody obtained from Dr. Stefan Feske used in 1:100 dilution for 1 hour at room temperature. After secondary antibody staining for 20 minutes at room temperature cover slips were mounted on microscope slides with ProlongGold (Invitrogen, Paisley, UK). Confocal sections were spaced at 0.5µm in axial directions throughout the whole cell (from near cover slip section to the top of the cell). Alexa 488 dye was excited at 488nm and emitted light was collected at 500-540nm, Alexa 594 was excited at 543nm and a 580-620nm bandpass filter was used whilst Alexa 647 was excited at 633nm and emission was collected with a 640-700nm bandpass filter.

Sources of Common Chemicals and Reagents
Tris-buffered saline was from BioRad (Hercules, CA, USA). Cell culture reagents, trypsin inhibitor, TMRM, FM 4-64 were purchased from Invitrogen (Paisley, UK). Collagenase was from Worthington (Lorne Laboratories, Reading, UK). Thapsigargin was from Calbiochem (San Diego, CA, USA). Paraformaldehyde, glutaraldehyde, osmium tetroxide, uranyl acetate, lead citrate, epoxy resin, formvar and 100 and 200 mesh copper grids was from Agar Scientific (Stansted, UK). All other chemicals were from SIGMA (Gillingham, UK). Tables S1 and S2 represent measurements of minimum distances between ER and PM. A value for the minimum distance was measured for every junction. The data were then averaged.

Figure S1. Distribution of STIM1-EYFP and FM 4-64 in Live Unstimulated Pancreatic Acinar Cells
Under control conditions (i.e. cells not treated with TG or Ca 2+ -releasing agonists) STIM1-EYFP (green) fluorescence is not concentrated in puncta, instead it demonstrates a typical ER distribution, well-documented for this cell type, [2][3][4][5]. The fluorescence is higher in the basal region of the cells where an extensive network of the ER is located and is lower in the apical region where only some strands of ER penetrate into the granular (apical) region. The secretory granule region of the pancreatic acinar cell was reported to produce strong light scattering [6,7]. To image plasma membrane structures in this region we stained the cells with the lipophilic dye

Figure S3. Positioning of STIM1-EYFP (Green) with Respect to Orai1-mCh (Red) in Cells with Intact Ca 2+ Stores (Untreated with TG)
Scale bar corresponds to 10 µm. This figure shows the same cluster of pancreatic acinar cells as Figure 1D. (i) Shows confocal section nearest to the cover slip, (ii) 4µm from the cover slip towards the middle of the cells and (iii) the middle section (approximately 10µm from the cover slip). Note diffuse distribution of STIM1-EYFP and uniform staining of basolateral membrane with Orai1-mCh. Note the changes of distribution of both constructs from uniform to punctuate following the TG treatment (compare with Figure 1D). The apical region contains high level of Orai1-mCh (ii and iii of this figure). This region also contains IP 3 -Rs ( Figure 2) and we can not exclude that the two proteins could co-localise and / or interact in this region. We consider that the relationship of Orai1 and IP 3 -Rs in pancreatic acinar cells should be addressed in a separate study. The important conclusion from the experiments illustrated by this figure and by Figure 1D is that STIM1 and Orai1 transfected into acinar cells using adenoviral constructs change distribution upon depletion of Ca 2+ store and that co-clustering of STIM1 and Orai1 is observed in the basal and lateral, but not apical regions of the plasma membrane.  figure) was similar to that found using Orai1-mCh expressing live cells (see Figure 1D). The immunostaining was however too faint to investigate the clustering of endogenous Orai1.  Figure 3A.

Figure S6. STIM1-EYFP Translocation into Puncta upon Store Depletion with Various Agents
Formation of STIM1 puncta in pancreatic acinar cells was observed not only due to depletion of ER calcium with TG but also following application of a low concentration of ACh, a low (physiological) concentration of CCK, inhibition of energy production with Oligomycin (5µM) and Iodoacetate (2mM) as well as following treatment with the bile acid taurolithocholic acid 3sulfate (TLC-S) which is considered to be a putative trigger for an important disease, acute pancreatitis. Experiments summarised in this figure were conducted on live pancreatic acinar cells in the presence of 1mM Ca 2+ in the extracellular solution. Notably, the first STIM1 puncta induced by low concentrations of secretagogues developed 300-400s after the addition of CCK or ACh. This is a relatively short time with respect to the periods of physiological stimulation of pancreatic secretion. This and the low doses of physiological agonists that were effective suggest the involvement of store-operated calcium channels in stimulus-secretion coupling in pancreatic acinar cells.  Note that these control cells (not treated with TG, n = 14) show HRP-dependent precipitates in the ER strands (similarly to TG treated cells, Figure 4). This is in line with the study of Wu and colleagues from Richard Lewis laboratory [8].