Plasmid information

DNA fragments encoding rat wild-type CSPα1 (pCMV-Csp-IRES2-egfp), its phosphodeficient mutant (pCMV-Csp-S10A-IRES2-egfp), or its phosphomimetic mutants (pCMV-Csp-S10D-IRES2-egfp; pCMV-Csp-S10E-IRES2-egfp) were obtained from our previous study (Chiang et al., Reference Chiang, Hsiao, Yang, Lin, Lu and Wang2014). To target the expression specifically to SACs, these DNA fragments were subcloned into pmGluR2-IRES2-egfp (designated Ctrl, hereinafter) (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012; Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019) using BglII and NotI, yielding pmGluR2-Csp-IRES2-egfp (designated CSP-WT, hereinafter), pmGluR2-Csp-S10A-IRES2-egfp (designated CSP-S10A, hereinafter), pmGluR2-Csp-S10D-IRES2-egfp (designated CSP-S10D, hereinafter), and pmGluR2-Csp-S10E-IRES2-egfp (designated CSP-S10E, hereinafter) (Figs. 2, 3A, 3B, and 4). To verify the ectopic gene expression after ex vivo electroporation (Fig. 3D and 3E), CSP and its phosphodeficient mutant were constructed into the pmGluR2-HA-IRES2-egfp vector (designated HA-Ctrl) (Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019), yielding pmGluR2-HA-Csp-IRES2-egfp (designated HA-CSP-WT) and pmGluR2-HA-Csp-S10A-IRES2-egfp (designated HA-CSP-S10A) in Fig. 3D and 3E.

Fig. 1. CSPα1 is expressed in developing rat SACs during the first postnatal week. (A) Immunostaining of CSP (green), choline acetyltransferase (ChAT, the SAC marker; red), or DAPI (blue) in retinal cross-sections from P2 and P6 rats. The colocalization signals were shown in yellow. GCL, ganglion cell layer; IPL, inner plexiform layer; NBL, neuroblast layer. A9 and A10, Magnification of white boxes in A7 and A8. Scale bars, 50 (A1 – A8) and 5 μm (A9 and A10). (B) Immunostaining of CSP (green) and ChAT (red) in the IPL from the P2 whole-amount retina (1.5-μm z-section). Scale bar, 50 μm. (C) Representative immunostaining of CSP (green), ChAT (red), or DAPI (blue) in a dissociated P4 rat SAC (1.5-μm z-section). BF, bright field. Scale bar, 7.5 μm. (D) The ratios were calculated for the cells displaying the CSP and/or ChAT immunoreactivity in an imaged region consisting of ~20 dissociated P4 retinal cells. Data were from four confocal images. CSP+/Total, the ratio for all dissociated cells with CSP immunoreactivity; ChAT+/Total, the ratio for all dissociated cells with ChAT immunoreactivity; CSP+/ChAT+, the ratio for all SACs with CSP immunoreactivity; ChAT+/CSP+, the ratio for all CSP-expressing cells as SACs. The cutoff intensity was 4.43 for CSP immunoreactivity and 1.9 for ChAT immunoreactivity. The average intensity was 37.15 ± 19.36 (mean ± s.d., n = 4) for CSP immunoreactivity and 7.84 ± 2.5 (mean ± s.d., n = 4) for ChAT immunoreactivity. (E) The endogenous mRNA levels of CSP isoforms (CSPα, CSPβ, and CSPγ) in intact P2 rat retinas. ***P < 0.0001 for CSPα versus CSPβ and CSPα versus CSPγ, One-way ANOVA with Student–Newman–Keuls post hoc test; n = 4 retinas from four pups. Inset panel, DNA gel electrophoresis after RT-qPCR experiments. (F) The endogenous mRNA levels of CSP isoforms (CSPα1, CSPβ, and CSPγ) in intact adult rat testes. ***P < 0.0001 for CSPα1 versus CSPβ and CSPβ versus CSPγ, One-way ANOVA with Student–Newman–Keuls post hoc test; n = 6 testes from three rats. For D–F, circles beside columns represent data from individual samples. For E and F, the mRNA levels normalized to the mean level of CSPβ and CSPα1, respectively. (G) DNA gel electrophoresis after RT-qPCR experiments validated the replicons based on these isoform-specific primers as positive control.

Fig. 2. SAC-specific expression of CSP-S10A reduces the frequency of spontaneous Ca²⁺ transients. (A) Molecular perturbation in P1-P2 rat SACs was performed by the mGluR2 promoter-driven expression, including Ctrl (pmGluR2-IRES2-egfp), CSP-WT, CSP-S10A, CSP-S10D, and CSP-S10E. Representative spontaneous Ca²⁺ transients after SAC-specific expression were shown in the 10-min live imaging. (B) Ca²⁺ transient frequency. **P = 0.0015, One-way ANOVA with Student–Newman–Keuls post hoc test. #P = 0.0305 for CSP-S10A versus Ctrl; no significance for any other groups versus Ctrl (P = 0.26 for CSP-WT vs. Ctrl; P = 0.28 for CSP-S10D vs. Ctrl; P = 0.18 for CSP-S10E vs. Ctrl), two-tailed Student’s unpaired t-test. (C) The intercellular correlation of spontaneous Ca²⁺ transients. n.s., no significance (P = 0.07), Repeated measures ANOVA with Tukey post hoc test. (D) A single Ca²⁺ transient. Top, the starting point (black arrows) and the end point (gray arrows) were defined by the criteria in section “Event definition.” The baseline (the gray solid line); The RMS noise (the gray dash line). Bottom, the Ca²⁺ transient duration and amplitude (ranging from the baseline to the peak, the red line). (E) Ca²⁺ transient duration. n.s., no significance, One-way ANOVA with Student–Newman–Keuls post hoc test (P = 0.81) or two-tailed Student’s unpaired t-test (P = 0.36 for CSP-WT vs. Ctrl; P = 0.29 for CSP-S10A vs. Ctrl; P = 0.66 for CSP-S10D vs. Ctrl; P = 0.21 for CSP-S10E vs. Ctrl). (F) Ca²⁺ transient amplitude. n.s., no significance, Kruskal–Wallis test with Dunn post hoc test (P = 0.82), Mann–Whitney test (P = 0.53 for CSP-WT vs. Ctrl), or two-tailed Student’s unpaired t-test (P = 0.30 for CSP-S10A vs. Ctrl; P = 0.53 for CSP-S10D vs. Ctrl; P = 0.17 for CSP-S10E vs. Ctrl). For B, C, E, and F, Data were acquired from 440 cells, 44 imaged regions, 24 retinal explants, and 10 pups for Ctrl; 480 cells, 48 imaged regions, 24 retinal explants, and 8 pups for CSP-WT; 467 cells, 47 imaged regions, 26 retinal explants, and 9 pups for CSP-S10A; 470 cells, 47 imaged regions, 25 retinal explants, and 8 pups for CSP-S10D; 480 cells, 48 imaged regions, 24 retinal explants, and 9 pups for CSP-S10E. Circles beside columns represent data from individual retinal explants.

Fig. 3. PKA-mediated phosphorylation is reduced by expressing CSP-S10A in developing SACs. (A) Immunostaining of ChAT (red) or CSP (green) in dissociated SACs expressing Ctrl (pmGluR2-IRES2-egfp), CSP-WT, or CSP-S10A (1.5-μm z-section). Scale bars, 5 μm. (B) Quantification of the CSP immunoreactivity in transfected SACs. ^##P = 0.0057 versus Ctrl, Mann–Whitney test. ^#P = 0.011 versus Ctrl, two-tailed Student’s unpaired t-test. n.s., no significance (P = 0.83), Mann–Whitney test. n = 23 SACs for each transfection group. (C) The relative mRNA levels of CSPα1 in transfected retinas expressing Ctrl, CSP-WT, or CSP-S10A. ^###P < 0.0001 versus Ctrl; n.s., no significance (P = 0.73), Mann–Whitney test. n = 21, 22, and 21 transfected retinas for Ctrl, CSP-WT, and CSP-S10A, respectively. (D) Immunostaining of ChAT (magenta), HA (green), or phospho-PKA substrate (red) in dissociated SACs expressing HA-Ctrl (pmGluR2-HA-IRES2-egfp), HA-CSP-WT, or HA-CSP-S10A (1.5-μm z-section). Scale bars, 5 μm. (E) Quantification of the phospho-PKA substrate immunoreactivity in transfected SACs. ***P = 0.0002, two-tailed Student’s unpaired t-test. ^#P = 0.03 for HA-CSP-S10A versus HA-Ctrl; no significance (P = 0.14) for HA-CSP-WT versus HA-Ctrl, Mann–Whitney test. n = 39, 42, and 38 SACs for HA-Ctrl, HA-CSP-WT, and HA-CSP-S10A, respectively. For B, C, and E, circles beside columns indicate the data from individual SACs.

Fig. 4. CSP-S10A dampens the strength of transmission from developing SACs. (A) Whole-cell patch-clamp recordings were performed in the RGCs proximal to transfected SACs expressing Ctrl (pmGluR2-IRES2-egfp), CSP-WT, or CSP-S10A. (B) The RGC was identified by quickly activated Na²⁺ currents (arrow) upon depolarizing pulses. (C) Action potentials were induced in a RGC by injecting various sizes of current pulses (−10 to +50 pA for 200 ms; ΔI = 20 pA). (D) The firing rate of action potentials in a RGC after SAC-specific expression. n.s., no significance (P = 0.15), Repeated measures ANOVA with Tukey post hoc test. Data were obtained from three to five RGCs, two retinal explants, and two pups. (E) The wave-associated postsynaptic currents (PSCs) were recorded in RGCs proximal to transfected SACs for 6 min. (F) PSC frequency. *P = 0.013, two-tailed Student’s unpaired t-test. ^##P = 0.0024 for CSP-S10A versus Ctrl; no significance (P = 0.24) for CSP-WT versus Ctrl, two-tailed Student’s unpaired t-test. (G) Amplitude, duration, and time to the peak (blue) in representative PSCs (by the criteria in section “Event definition”). Inset, the enlarged PSC event on a different scale from the left panel. The baseline (red); The peak line (green). (H) PSC amplitude. ^#P = 0.036 for CSP-S10A versus Ctrl; no significance (P = 0.44) for CSP-WT versus Ctrl; no significance (P = 0.06) for CSP-WT versus CSP-S10A, two-tailed Student’s unpaired t-test. (I) PSC duration. n.s., no significance (P = 0.79 for CSP-WT vs. Ctrl; P = 0.53 for CSP-S10A vs. Ctrl; P = 0.50 for CSP-WT vs. CSP-S10A), two-tailed Student’s unpaired t-test. (J) Time to peak PSC. n.s., no significance (P = 0.32 for CSP-WT vs. Ctrl; P = 0.07 for CSP-S10A vs. Ctrl; P = 0.17 for CSP-WT vs. CSP-S10A), Mann–Whitney test. (K) Slope to peak PSC. **P = 0.0029, two-tailed Student’s unpaired t-test. ^#P = 0.023 for CSP-S10A versus Ctrl; no significance (P = 0.19) for CSP-WT versus Ctrl, two-tailed Student’s unpaired t-test. (L) PSC integral. n.s., no significance (P = 0.69 for CSP-WT vs. Ctrl; P = 0.26 for CSP-S10A vs. Ctrl; P = 0.44 for CSP-WT vs. CSP-S10A), Mann–Whitney test. For (F) and (H–L), Data were acquired from seven RGCs, four retinal explants, and four pups for Ctrl; 14 RGCs, five retinal explants, and five pups for CSP-WT; seven RGCs, four retinal explants, and four pups for CSP-S10A. Circles beside columns represent data from individual retinas.

Animals

Postnatal (P1–P6) Sprague–Dawley (SD) rat pups with either sex were used in this study. All procedures were performed in accordance with protocols approved by the institutional animal care and use committees of National Taiwan University (NTU). The pups were bred from their own mothers (from BioLASCO, Taipei, Taiwan; ad libitum access to food and water) in the individually ventilated cages under well-controlled conditions (12:12 light/dark cycle with light on 7 AM; 22 ± 1°C). All rat pups were deeply anesthetized before decapitation with isoflurane to minimize suffering with all efforts.

Retinal explant culture and transfection

Postnatal retinas were obtained from SD rat pups (P1–P2) and transfected by ex vivo electroporation (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012; Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019). Briefly, after decapitation, the retinas were isolated and cut into three pieces in dissection buffer [1 × HBSS (Gibco), 10 mM HEPES, and 0.35 g/L NaHCO₃, pH 7.35]. The retinal pieces (retinal explants) were attached onto nitrocellulose membranes (Millipore) with the GCL up. Retinal explants were incubated with serum-free culture medium-adult (SFCM-A) [Neurobasal-A (Gibco), 0.6% glucose, 2 mM l-glutamine (Sigma), 1 × B27 (Gibco), 10 mM HEPES, 1 mM sodium pyruvate (Gibco), 2.5 μg/ml insulin (Sigma), 100 μg/ml penicillin (Gibco), 100 units/ml streptomycin (Gibco), and 6 μM forskolin (Sigma)] at 35°C in a humidified atmosphere of 5% CO₂ and supplied with fresh culture medium daily. To perform transfection, retinal explants were incubated in dissection buffer containing plasmid DNA (200 ng DNA/μL) at RT for 10 min. The retinal explants were placed in the gap of horizontal platinum electrodes (4 mm) (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012) filled with 400 μL DNA-containing buffer and transfected by electroporation (27 V, 50 ms of pulse duration, 2 square pulses at 1-s interval; BTX ECM830, Harvard Apparatus). After electroporation, retinal explants were cultured in SFCM-A with forskolin and fed with fresh culture medium daily until 72 h post transfection for further experiments (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012; Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019).

Immunostaining

For immunostaining of retinal cross-sections, anesthetized postnatal pups (P2 or P6) were perfused with 1 × phosphate-buffered saline (PBS; 136.89 mM NaCl, 2.68 mM KCl, 10.14 mM Na₂HPO₄, and 1.76 mM KH₂PO₄, pH 7.4) and 4% paraformaldehyde (PFA). Isolated eyeballs were kept in 4% PFA at 4°C 30 min and in 30% sucrose at 4°C for overnight, followed by preservation in optimal cutting temperature (OCT) gel. Retinal cross-sections (16 μm) were prepared with a cryostat (Leica CM1850), placed on poly-lysine-coated slides, and blocked at RT for 1 h in donkey serum blocking solution [DBS; 3% donkey serum (Jackson Lab) and 0.5% Triton X-100 in 1 × PBS]. Retinal sections were incubated at 4°C overnight with the primary antibodies in 1% DBS [goat anti-choline acetyltransferase (ChAT; Millipore AB144P, 1:200) and rabbit anti-CSP (Millipore AB1576, 1:1000)], washed with PBS for 1 h, incubated at RT for 2 h with the secondary antibodies in 1% DBS [donkey-anti-goat IgG conjugated to Alexa Fluor 568 (Invitrogen, 1:400) and donkey-anti-rabbit IgG conjugated to Alexa Fluor 488 (Invitrogen, 1:400)], and further stained with 4′,6-diamidino-2-phenylindole (DAPI; Sigma) at RT for 15 min. The cross-sections on slides were sealed by coverslips with Fluoromount.

For immunostaining of whole-mount retinas, P2 whole-mount retinas were fixed in 4% PFA at RT for 30 min and then washed with 1 × PBS for 1 h. After fixation, the explants were blocked and permeabilized in DBS at RT for 1 h, and then incubated with the primary antibodies in 1% DBS (the same as described above) at 4°C for 2 days and washed with PBS for 1 h. The secondary antibodies in 1% DBS (the same as described above) were added to the explants at 4°C for 2 h. The explants were washed with PBS for 1 h. Finally, the explants were mounted on glass slides with Fluoromount and sealed with coverslips.

For immunostaining of dissociated retinal neurons, isolated retinal neurons were acquired from P2 or P4 pups (Grozdanov et al., Reference Grozdanov, Muller, Sengottuvel, Leibinger and Fischer2010), cultured on the coated coverslips, fixed with 4% PFA at RT for 20 min, and washed with 1 × PBS for 20 min. After fixation, the cells were permeabilized with 0.1% Triton X-100 for 10 min and blocked in 3% DBS with 0.1% Triton X-100 at RT for 1 h. After blocking, the neurons were incubated at 4°C overnight with the primary antibodies [goat anti-ChAT and rabbit anti-CSP; goat anti-ChAT, rabbit anti-phospho-PKA substrate (Cell Signaling 9624, 1:200), and mouse anti-HA (Covance MMS-101P, 1:800)], washed with PBS for 1 h, incubated at RT for 2 h with the secondary antibodies in 1% DBS [donkey-anti-goat IgG conjugated to Alexa Fluor 647 (Invitrogen, 1:400), donkey-anti-rabbit IgG conjugated to Dylight 549 (Jackson Immuno Research, 1:400), and donkey-anti-mouse IgG conjugated to Alexa Fluor 488 (Invitrogen, 1:400)], and further stained with DAPI at RT for 15 min. Finally, the neurons on coverslips were mounted on slides with Fluoromount.

Images were acquired by laser-scanning confocal microscopy (Leica TCS SP5 spectral) in z-series, consisting of one plane (1.5-μm thickness) for whole-mount retinas/dissociated cells and 8–11 planes for retinal cross-sections (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012; Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019). For quantification of immunoreactivity, images were imported into MetaMorph software to quantify the changes in fluorescence intensity for each cell, with subtraction and normalization by the same-sized background (ΔF/F) (Arndt-Jovin et al., Reference Arndt-Jovin, Robert-Nicoud, Kaufman and Jovin1985; Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019; Webster et al., Reference Webster, Tworig, Caval-Holme, Morgans and Feller2020). The levels of fluorescence intensity were normalized to the mean level of the Ctrl group (Fig. 3B and 3E). Data were further analyzed by Excel and Origin 8.

Reverse-transcriptase quantitative PCR

The RNA sample from whole-mount retinas or testes was extracted by TRIzol reagent (Invitrogen). Briefly, the samples were homogenized by adding 1 ml TRIzol reagent and phase-separated by 0.2 ml chloroform. After centrifugation at 12,000 g at 4°C for 15 min, the RNA was extracted in the upper aqueous phase (~0.6 ml) and then precipitated with 0.5 ml isopropanol. The RNA-containing solution was centrifuged at 12,000 g at 4°C for 10 min, and the supernatant was discarded. The RNA pellets were washed once with 75% ethanol, followed by centrifugation at 12,000 g at 4°C for 10 min. After complete removal of the supernatant, the RNA pellets were air-dried at RT for 20 min upon transparency, dissolved in 20 μL of diethyl pyrocarbonate (DEPC)-treated water, and incubated at 60°C for 10 min. The RNA samples were stored at −80°C for further experiments. The cDNAs were synthesized from RNAs using the ProtoScript II First Strand cDNA Synthesis kit (New England BioLabs). Reverse-transcriptase quantitative PCR (RT-qPCR) was performed on the cDNA samples using the LabStar SYBR qPCR kit (TAIGEN Bioscience Corporation). The primers specific for the transcripts of CSPα1, CSPα2, CSPβ, CSPγ, or β-actin were obtained from our previous study (Chiang et al., Reference Chiang, Hsiao, Yang, Lin, Lu and Wang2014). To detect the ectopic HA-CSP (WT or S10A) expression, the forward primer was designed to target the HA-tag (5′-CCA TGT ACC CAT ACG ATG TTC CAG-3′, underlined) and the reverse primer was designed to target CSPα1 (5′-TCT GCA GCC TCT GGG TTA TC-3′), yielding the replicon length as 200 base pairs. The SYBR fluorescence data were collected during the extension step of each cycle (totally 40 cycles of 95°C for 3 s, 60°C for 45 s, and 72°C for 30 s) and transformed into the cycle threshold (Ct) values using the qPCR machine (Qiagen Rotor-Gene Q) and supplemental software (Qiagen Rotor-Gene Series Software 1.7). The ΔCt was obtained by subtracting the Ct of the reference gene (β-actin) from the Ct of the target gene. The ΔΔCt was obtained by subtracting the average of the ΔCt of CSPβ (Fig. 1E), CSPα1 (Fig. 1F), or Ctrl (Fig. 3C) from the ΔCt of the other target gene or transfection group. The relative expression levels of the target genes were calculated as 2^(−ΔΔCt).

Live Ca²⁺ imaging

Transfected retinal explants were transferred into SFCM-A without forskolin overnight before imaging, and then incubated for 30–60 min in the Ca²⁺ indicator-containing medium (10 μM fura-2-AM, 0.02% pluronic acid, and 1% DMSO in the SFCM-A without forskolin). During image acquisition, retinal explants were perfused with artificial cerebrospinal fluid (ACSF) (in mM: 119 NaCl, 26.2 NaHCO₃, 2.5 KCl, 1.0 K₂HPO₄, 1.3 MgCl₂, 2.5 CaCl₂, and 11 D-glucose) bubbled with 95% O₂/5% CO₂ warmed to 30°C. Imaging experiments were performed under a 20 × water immersion objective (Olympus BX51WI) with excitation at 380 nm and emission at 510 nm. Fluorescence images were captured using a CCD camera (CoolSnap HQ2, PhotoMetrics) at 1-s intervals for 10 min (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012; Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019). Wave-associated Ca²⁺ transients were acquired from the fluorescence changes across 601 time frames, each with background subtraction using the MetaMorph software (Molecular Devices). To analyze the characteristics of spontaneous Ca²⁺ transients, the photobleached baseline was corrected using our Igor (WaveMetrics) procedure as previously reported (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012; Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019). The fluorescence value (F) in each Ca²⁺ transient was subtracted from the baseline (F ₀) and further divided by the baseline [(F − F ₀)/F ₀] to produce the trace of ΔF/F. Further, the written Igor procedure would automatically pick the peaks of Ca²⁺ transients (see the detail in section “Event definition”). All data of Ca²⁺ transients were averaged from each cell, averaged across 10 randomly selected cells out of one imaged region, and then averaged from two imaged regions out of one retinal explant. Finally, the mean data for each group were averaged from all retinal explants transfected with the same gene. Further analysis was performed by Excel and Origin 8 (OriginLab) (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012; Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019).

Spatial correlation of spontaneous Ca²⁺ transients was evaluated by the correlation index (CI) (Wong et al., Reference Wong, Meister and Shatz1993; Torborg et al., Reference Torborg, Hansen and Feller2005) according to the following equation:

$$ \mathrm{CI}\hskip0.35em =\hskip0.35em \frac{N_{\mathrm{ab}}\left(-\Delta t,+\Delta t\right)\times T}{N_a\left(0,T\right)\times {N}_b\left(0,T\right)\times 2\Delta t}. $$

N _ab is the transient number for which cell b exhibits in a time window ± Δt (3 s) from cell a. N_a and N_b are the total numbers of transients exhibited by cells a and b, respectively, during the total imaging time (T, 600 s). The averaged CI values were computed from the same distance group and plotted against the intercellular distance as previously described (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012; Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014).

Electrophysiology

Whole-cell patch-clamp recordings were performed on visualized RGCs (60 × water-immersion objective, Olympus) in transfected retinal explants with SAC-specific expression. During recordings, retinas were continuously perfused with oxygenated ACSF at 30°C as described previously (Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019). Borosilicate glass pipettes (WPI#PG52151-4) were pulled (Narishige PC-10) to a tip resistance of ~5.5 MΩ when filled with a pipette solution [98.3 K-gluconate, 1.7 KCl, 0.6 EGTA, 5 MgCl₂, 40 HEPES, 2 Na₂-ATP, 0.3 Na-GTP (in mM), pH 7.25 with KOH]. Recordings were made using an Axopatch 200B patch-clamp amplifier with Digidata 1440A interphase (Molecular Devices). Data were acquired and analyzed with the pClamp10 software (Molecular Devices). In whole-cell voltage-clamp recordings, the spontaneous wave-associated postsynaptic currents (PSCs) (filtered at 1 kHz and digitized at 5 kHz) were recorded at a holding potential of −72 mV (the liquid junction potential as −12 mV) (Fig. 4E–4L). The induced current responses were also measured using the protocols indicated elsewhere (Fig. 4B). Whole-cell current-clamp recordings were used to measure the action potential firings (filtered at 5 kHz and digitized at 10 kHz) (Fig. 4C and 4D). In successful recordings, gigaohm seals were obtained within 30 s, and the ratios of access resistance to input resistance were 5–15%. Data from wave-associated PSCs were averaged over all events from one cell. The final data for each group were averaged across all cells transfected with the same gene (Huang et al., Reference Huang, Hsiao, Kao, Chen, Chen, Chiang, Lee, Lu, Chern and Wang2014; Hsiao et al., Reference Hsiao, Shu, Chen, Yang, Chen, Hsu, Huang, Yang, Chen, Yu, Liou, Chiang, Huang, Cheng, Cheung, Lin, Lu and Wang2019).

Event definition

For individual Ca²⁺ transients, the definitions of the duration and amplitude were presented in Fig. 2D. The written Igor procedure would automatically pick the peaks of Ca²⁺ transients, with their fluorescence intensity twofold greater than the root-mean-square (RMS) noise (about 0.25% ΔF/F) as reported previously (Chiang et al., Reference Chiang, Chen, Lu, Hsiao, Chang, Huang, Chang, Chang and Wang2012). The RMS noise was measured from the fluorescence trace starting with 30 s before the peak to 50 s following the peak. The starting point (x ₀, y ₀) of an event (Fig. 2D, black arrows) was defined by the time point where the first derivative was zero right before the peak (i.e., dy/dx = 0, where y was the fluorescence changes in %ΔF/F and x was the recording time in seconds). To define the end point (x′, y′) of an event (Fig. 2D, gray arrows), a line was first drawn to connect the time points where the fluorescence trace was back to the range within the RMS noise. The end point was defined by the time point with the minimal fluctuation of fluorescence (i.e., y′ − y ₀ = minimum). Finally, the Ca²⁺ transient duration was defined as the interval between the starting and end points. The Ca²⁺ transient amplitude was defined as the fluorescence change from the baseline to peak (shown as the red line in Fig. 2D).

For individual wave-associated PSCs, the definitions of amplitude, duration, and time to the peak were presented in Fig. 4G. The RMS noise was measured from the trace between 10 s prior to the event and 15 s following the event. The potential PSC events were picked with their current response 2.5-fold greater than the RMS noise. Further, to distinguish from the miniature synaptic events (minis), wave-associated PSCs were identified by the slow inward current (~s) with the large event integral (>3.5 pC). The amplitude of wave-associated PSCs was defined as the current change from the baseline (Fig. 4G, the red line) to the peak, that is, the maximal inward current that lies in the middle of the recording noise by taking the average across 10 data points (Fig. 4G, the green line). The PSC duration was defined as the interval between the trace left from (the starting point) and returned to (the ending point) within the RMS noise of the baseline. The time to the peak PSC was defined as the interval between the starting point and the peak of an event. The slope to peak PSC was individually calculated from the PSC amplitude divided by the time to peak, reflecting the transmission rate reaching to the maximal postsynaptic response. The PSC integral was calculated from the area of individual events, reflecting the amount of input signal received by postsynaptic cells.

Statistics

Data were presented as means with standard deviation (s.d.) (OriginLab). Statistical significance ≥3 groups (CSP-WT and three CSP phosphomutants) was evaluated by one-way ANOVA with Student–Newman–Keuls post hoc test for the parametric method or by Kruskal–Wallis test with Dunn post hoc test for the nonparametric method (*P < 0.05; **P < 0.01; ***P < 0.001). Statistical significance between control and the other transfection group was evaluated by two-tailed Student’s unpaired t-test as a parametric method or by Mann–Whitney method as a nonparametric method (^#P < 0.05; ^##P < 0.01; ^###P < 0.001 vs. Ctrl/HA-Ctrl). Repeated measures ANOVA with Tukey’s multiple comparison post hoc test was used to evaluate significant differences in the correlation index and RGC firing rate (InStat 3, GraphPad). n.s., no significance (P > 0.05).