Primary cell culture

As we previously described [4], cortical neurons were prepared from 19-day-old Wistar rat embryos. Pregnant female Wistar rats were obtained from the Animal Care Facility of the Mossakowski Medical Research Institute of the Polish Academy of Sciences (Warsaw, Poland). Animal care was in accordance with the Directive of the European Communities Council (86/609/EEC). Briefly, embryonic rat cortices were dissected, collected, rinsed in ice-cold Hanks’ Balanced Salt Solution (Sigma) that was supplemented with 11 mM HEPES buffer solution (Gibco) and 100 U/ml/100 µg/ml penicillin/streptomycin solution (Gibco), and treated with trypsin (Gibco) for 37 min at 37 °C. The tissue was then rinsed and dissociated by pipetting. For the co-immunoprecipitation (Co-IP) and biotinylation assays, neurons were seeded in 100 mm Biocoat poly-D-lysine (PDL) Cellware dishes (Corning) at a density of 10 × 106 cells/dish. For isolation of the synaptosomal fraction and the preparation of cell lysates, neurons were seeded on PDL-precoated 6-well plates (Corning) at a density of 1.2 × 106 cells/well. For the enzyme-linked immunosorbent assay (ELISA), neurons were seeded at a density of 3 × 105 on 24-well plates that were precoated with PDL (Corning). For immunofluorescence and whole-cell patch clamp, neurons were seeded at a density of 2.2 × 105 per 13 mm glass coverslip that was coated with a mixture of laminin (1.25 µg/ml; Roche, Mannheim, Germany) and PDL (37.5 µg/ml; Sigma-Aldrich, St. Louis, MO, USA) in 24-well plates. Neurons were grown in Neurobasal medium (Gibco, Paisley, UK) that was supplemented with 2% B27 (Gibco), 0.5 mM glutamine (Sigma), 12.5 µM glutamate (Sigma), and penicillin/streptomycin solution (Gibco) at 37 °C in a humidified atmosphere with 5% CO2/95% air. On day in vitro 3 (DIV3), half of the medium was replaced with a glutamate-free growth medium that contained the non-neuronal cell proliferation inhibitor CultureOne Supplement (Gibco). The experiments were performed after DIV10.

Lentiviral production and neuronal transduction

HEK 293 T/17 cells (American Type Culture Collection) were grown in Dulbecco’s Modified Eagle Medium (DMEM) that was supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin solution (Gibco) at 37 °C in a humid environment with 5% CO2/95% air. The viruses were prepared by the calcium phosphate transfection method. Three days after transfection, supernatants were collected, filtered through 0.45 μm membranes, concentrated in Vivaspin 100 kDa units (Sartorius) in a swing-out rotor at 3000 × g, aliquoted, and stored at −80 °C until needed.

Four commercially available Rat 29-mer target-specific short-hairpin RNA (shRNA) constructs (shStim1 C, shStim1 D, shStim2 C, and shStim2 D) in pLenti-green fluorescent protein vectors (GFP; Origene, Rockville, MD, USA) were used to knockdown STIM proteins. The targeting sequences for STIMs were as follows: GGATAATGGCTCTATTGGTGAGGAGACAG (TL707032C, shStim1 C), CTTCCAATGGTAGCCATCGGCTGATTGAG (TL707032D, shStim1 D), TTAGCCAGAAGCAGTAGTTTATGCCGCTC (TL704348C, shStim2 C), AGTCTGGAAGCACTTCAGACAATACATAA (TL704348D, shStim2 D). Scramble shRNA cassette in pLenti-GFP (scrRNA; Origene) was used as a control. The non-effective control sequence was 5' GCACTACCAGAGCTAACTCAGATAGTACT 3'. Neurons were transduced on DIV5 to knockdown STIM, and the experiments were performed 5 days after transduction (DIV10).

Experimental conditions

Short-term neuronal overactivation was induced by 50 µM NMDA (Alomone) and 100 µM glycine (Sigma) for 15 min at 37 °C [51] in modified extracellular solution that was composed of 10 mM HEPES, 1.3 mM CaCl2, 110 mM Na2SO4, 5 mM Cs2SO4, 0.75 mM Na2HPO4, and 10 mM glucose (pH 7.4); [42]. To prevent cell damage during NMDAR overactivation, K+ and Cl− were replaced by Cs+ and SO42−, respectively [52].

Cell viability was estimated by measuring lactate dehydrogenase that was released to the extracellular solution using the CytoTox-ONE Homogenous Membrane Integrity Assay (Promega, Madison, WI, USA). Lactate dehydrogenase that was released by shStim- or scrRNA-transduced neurons was normalized to wild-type cells. Lactate dehydrogenase that was released by NMDA- and glycine-treated neurons was normalized to control cells.

Total protein isolation

To prepare cell lysates, scrRNA- or shStim-transduced neurons were centrifuged at 1500 × g for 5 min at 4 °C. The pellet was lysed in buffer (pH 7.5), that contained 50 mM Tris, 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate (DOC), 1% NP-40, 1 mM phenylmethylsulfonyl fluoride, and complete mini EDTA-free Protease Inhibitor Cocktail (Roche), for 30 min on ice and centrifuged at 17,000 × g for 15 min at 4 °C. The supernatant was collected, and total protein concentration was measured using Bradford reagent (Sigma). Laemmli buffer was added, and the samples were boiled and resolved by SDS-PAGE.

Western blot

Equal amounts of proteins were analyzed by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to a nitrocellulose membrane (Amersham), and stained for total protein (Ponceau S staining). Membranes were blocked with 5% nonfat dried milk in TBS with 0.1% Tween 20 (TBST) and then incubated with the following primary antibodies: mouse anti-GluN1 (Thermo Fisher, Catalog No. 32-0500, 1:200), rabbit anti-GluN2A (Alomone, Catalog No. AGC-002, 1:200), rabbit anti-GluN2B (Proteintech, Catalog No. 21920-1-AP, 1:200), rabbit anti-STIM1 (Proteintech, Catalog No. 11565-1-AP, 1:200), rabbit anti-STIM2 (Proteintech, Catalog No. 21192-1-AP, 1:200), mouse anti-EEA1 (BD, Catalog No. GT10811, 1:200), rabbit anti-post-synaptic density 95 (PSD95; Alomone, Catalog No. APZ-009; Cell Signaling, Catalog No. 2507S, 1:200), rabbit anti-N-cadherin (Proteintech, Catalog No. 22018-1-AP, 1:2000), rabbit anti-α-tubulin (Proteintech, Catalog No. 11224-1-AP, 1: 200 000), rabbit anti-Na+/K+-ATPase (Proteintech, Catalog No. 14418-1-AP, 1:1000), rabbit anti-calnexin (Proteintech, Catalog No. 10427-2-AP, 1:1000), and rabbit anti-LDH (Proteintech, Catalog No. 14824-1-AP, 1:1000). The specificity of ProteinTech anti-STIM1 and anti-STIM2 antibodies has been verified earlier several times in [53, 54] with the use of STIM1 siRNA or in [55, 56] with the use of STIM2 siRNA. Specific band for STIM1 and STIM2 was detected at approximately 90 kDa and 100 kDa, respectively, as observed by other authors using the same antibody from ProteinTech [53, 57,58,59]. The blot confirming the specificity of the antibody using shRNA presented in Fig. S1A, shows that shStim1 does not reduce STIM2 and shStim2 does not reduce STIM1. Membranes were then incubated with anti-mouse (Sigma-Aldrich, 1:4000) or anti-rabbit (Sigma-Aldrich, 1:5000) horseradish peroxidase-conjugated secondary antibodies. Bound antibodies were visualized by the chemiluminescence detection reagent Amersham ECL (Cytiva). If necessary, the membranes were stripped and reprobed. The samples were imaged using the Fusion FX imaging system (Vilber Lourmat, Marne-la-Vallée, France) and quantified using Image Studio Lite 5.2 software. The intensity of bands was normalized to total protein densities that were acquired by Ponceau S staining, corresponding to the same lane [60, 61], and quantified using ImageJ software with the Gel analyzer feature (National Institutes of Health, Bethesda, MD, USA) or to the intensity of α-tubulin (cell lysate) or N-cadherin (surface proteins) bands.

Synaptic protein isolation

Synaptosomal fractions from wild-type neurons or scrRNA- or shStim-transduced neurons were isolated using Syn-PER Synaptic Protein Extraction Reagent (Thermo Scientific, Catalog No. 87793). Neurons were centrifugated at 1200 × g for 10 min at 4 °C to dispose of cellular debris. The supernatant was then centrifuged at 15,000 × g for 20 min at 4 °C, and the pellet was suspended in Syn-PER reagent. This procedure yielded synaptosomes that contained pre- and post-synaptic proteins.

Co-immunoprecipitation

For the Co-IP of proteins, neurons were collected and lysed for 1.5 h. Lysates were centrifuged at 15,000 × g for 15 min at 4 °C, and the supernatant was collected and mixed with Protein G-Agarose beads (Roche). Precleared lysates (700 µg) were incubated overnight with protein G-Agarose that was preincubated for 3 h with the following antibodies: rabbit anti-GluN2A (Alomone, Catalog No. AGC-002) or rabbit anti-GluN2B (Proteintech, Catalog No. 21920-1-AP). Rabbit anti-immunoglobulin G (IgG) antibody (Merck, Catalog No. 12-370 and 12-371) was used as a negative control. The precipitated proteins were eluted and separated by SDS-PAGE.

Cell surface protein quantification assays

In wild-type neurons, surface proteins were isolated using the cell-surface protein isolation kit (Abcam, Catalog No. ab206998). Briefly, surface proteins were biotinylated with a solution of Sulfo-NHS-SS-Biotin for 30 min at 4 °C. Thereafter, a quenching solution was added to bond excess biotin. Cells were centrifuged, and the pellet was washed with Tris-buffered saline (TBS) and lysed. On average, we yielded 400 µg of proteins during one experiment in 400 µl of the lysis buffer. According to supplier protocol, 50 µl was left as „lysate” and 350 µl was incubated with 150 µl of streptavidin beads for 2 h with end-over-end mixing at 4 °C. In subsequent steps of the kit protocol, it is not possible to determine the protein concentration due to the buffer formulation nor to estimate what percentage of protein was collected by the beads. The surface proteins were eluted from the beads using 60 µl phosphate-buffered saline (PBS) with the strong reducing agent 0.1 M dithiothreitol to break the Sulfo-NHS-SS-biotin disulfide bonds, releasing the isolated proteins into solution and subjecting them to initial degradation and solubility. Laemmli buffer was then added in the presence of SDS for complete denaturation. Two gels were loaded with 25 µl of both the "lysate" and "surface protein" fractions. Approximately, the amount of protein in the eluted surface fraction is 2.2 times lower than in the lysates, as shown by Ponceau staining of the membrane (Fig. S2A).

Due to material limitations, in experiments on primary neurons in culture that were transduced with scrRNA or shStim (1n = cerebral cortex from 30 rat embryos), surface proteins were analyzed by a quantitative cell-based ELISA [62]. Neurons were fixed in 4% paraformaldehyde and 4% sucrose in PBS. The permeabilization step was omitted, because only the NMDAR surface was measured. Nonspecific binding was then blocked by 5% bovine serum albumin in PBS. Subsequently, cells were incubated with primary antibodies for the extracellular epitope of GluN1 (guinea pig, Alomone, Catalog No. AGP-046, 1:100), GluN2A (mouse, Abcam, Catalog No. ab240884, 1:100), or GluN2B (rabbit, Alomone, Catalog No. AGC-003, 1:100) in blocking solution for 2 h. Cells were then incubated with horseradish peroxidase-conjugated anti-guinea pig, anti-mouse, or anti-rabbit secondary antibody (Sigma-Aldrich, 1:2000) for 1 h. Afterward, 1-Step Turbo TMB-ELISA Substrate Solution (Thermo Fisher, Catalog No. 34022) was added, incubated for 30 min in the dark, and terminated by 1 M H2SO4. Absorption was measured at 450 nm using a Tecan INFINITE M1000 PRO plate reader with dedicated software (Tecan Group, Mannedorf, Switzerland). For normalization, the staining of cell nuclei with Hoechst 33,342 (Thermo Fisher, 1:10,000, 5 min) was used, the fluorescence of which was measured at an excitation wavelength of 358 nm and emission wavelength of 461 nm.

Immunofluorescence and image acquisition

For immunofluorescent staining, neurons were fixed in ice-cold 4% paraformaldehyde and 4% sucrose, permeabilized with 0.05% saponin (Sigma) in PBS, and blocked with 2% normal goat serum in PBS. The cells were then incubated at room temperature for 2 h with the following antibodies: rabbit anti-GluN1 (Abcam, Catalog No. ab17345, 1:50), mouse IgG2a anti-GluN1 (Thermo Fisher, Catalog No. 32-0500, 1:50), rabbit anti-GluN2A (Alomone, Catalog No. AGC-002, 1:50), mouse IgG2a anti-GluN2A (Abcam, Catalog No. ab240884, 1:50), rabbit anti-GluN2B (Alomone, Catalog No. AGC-003, 1:50), mouse IgG anti-GluN2B (Abcam, Catalog No. ab93610, 1:50), rabbit anti-STIM1 (Proteintech, Catalog No. 11656-1-AP, 1:50), rabbit anti-STIM2 (Alomone, Catalog No. ACC-064, 1:50), mouse IgG anti-EEA1 (BD, Catalog No. GT10811, 1:50), and chicken anti-MAP2 (Thermo Fisher, Catalog No. PA1-16,751, 1:500) diluted in blocking solution with 0.05% saponin. Immunoreactivity was detected with anti-mouse IgG Alexa Fluor 488-, anti-mouse IgG2a Alexa Fluor 488-, anti-chicken Alexa Fluor 568-, or anti-rabbit Alexa Fluor 647-conjugated secondary antibodies (Thermo Fisher). Nuclei were stained with Hoechst 33,342. Coverslips were mounted on slides with ProLong Gold Antifade Mountant (Thermo Fisher). For the detection of surface proteins, the permeabilization step was omitted, and incubation with antibodies without detergent was performed.

To label surface NMDAR subunits in STIM1/STIM2-silenced or control (scrRNA) neurons [63], cells were fixed and incubated for 2 h at room temperature with rabbit antibody directed against the N-terminal extracellular epitope of GluN1 (Alomone, Catalog No. AGC-001, 1:50), GluN2A (Alomone, Catalog No. AGC-002, 1:50), or GluN2B (Alomone, Catalog No. AGC-003, 1:50). After washing, neurons were incubated with anti-rabbit IgG Alexa Fluor 647 secondary antibody (Thermo Fisher, 1:500) at room temperature for 45 min. Neurons were then permeabilized, blocked, and incubated with the same primary rabbit antibody against GluN1, GluN2A, or GluN2B and mouse antibody against EEA1 (BD, Catalog No. GT10811, 1:50). Immunoreactivity was visualized using a combination of anti-rabbit and anti-mouse secondary antibodies conjugated to Alexa Fluor 405 (Thermo Fisher) and Alexa Fluor 568 (Thermo Fisher), respectively.

Immunostained cells were viewed under a Zeiss LSM780 Axio Observer confocal microscope (Carl Zeiss AG, Oberkochen, Germany). Images were acquired using a 63 × Alpha Plan-Apochromat oil immersion objective. Image resolution was 1024 × 1024 pixels. The laser power and detector gain were adjusted to obtain the maximum signal without oversaturation and cross-fluorescence with minimum background signal. Images were processed with ZEN 3.0 (blue edition) and deconvoluted using the ImageJ software (National Institutes of Health, Bethesda, MD, USA). Manders’ co-localization coefficient (M1) was used to estimate the level of protein co-localization [64].

Whole-cell patch-clamp recording

NMDAR currents were recorded in the whole-cell configuration of the patch-clamp technique at room temperature. Series resistance was 3–13 MΩ (constant during recording, controlled before and after recording). During recordings, neurons were voltage-clamped at −60 mV and submerged into the solution stream that was supplied by a manually controlled multi-barreled perfusion system (~ 1 ml/min). The intracellular solution comprised 140 mM CsF, 5 mM BAPTA, 1 mM CaCl2, 4 mM MgCl2, 10 mM HEPES, and 2 mM Na2ATP (pH 7.3). The extracellular solution contained 140 mM NaCl, 1.3 mM CaCl2, 5 mM KCl, 25 mM HEPES, 33 mM glucose, and 500 nM tetrodotoxin (pH 7.35; [51]). To prevent desensitization of the NMDA channels, the currents were evoked by 4 subsequent 4-s applications of 50 µM NMDA and 100 µM glycine, separated by 5-min intervals (total experiment time 15 min). Each of these four records is recorded from one cell. Shown are recordings of currents in one cell obtained after subsequent doses of NMDA and glycine. To block NMDAR endocytosis, the glycine site antagonist MDL29951 (5 µM) was used. The compounds were dissolved in extracellular solution and applied to the recording chamber via a perfusion system. NMDAR currents were recorded with an Axopatch 1-D amplifier. Data were digitized using DigiData1200A, filtered (2 kHz), and acquired using pClamp10.6 software.

Statistical analysis

The statistical analysis was performed using Prism 10 software (GraphPad, San Diego, CA, USA). The results are expressed as the mean ± standard error of the mean (SEM) of three-to-six independent experiments. Statistical significance was determined by unpaired t test, or by one- or two-way Anova + Tukey's post hoc test for multiple comparisons, as indicated in figure legend. Values of p < 0.05 were considered statistically significant.