N2A (ATCC) cells were cultured in DMEM GlutaMAX (Sigma-Aldrich) supplemented with 10% [v/v] FBS (Biowest) and 1% [v/v] penicillin/streptomycin (Gibco) at 37°C. The N2A cells were seeded on poly-D-lysine (Thermo Fisher)-coated coverslips. The following day, the cells were treated with differentiation medium [Neurobasal medium 1 g/l glucose (Gibco) + 1X B-27 (Gibco) + 1X sodium pyruvate (Gibco) + 1X glutamine (Gibco) + 1% [v/v] penicillin/streptomycin (Gibco)] and transfected with siRNA against either control (siScr; Dharmacon) or INPP5K (siInpp5k; Thermo Fisher) according to the manufacturer's protocol using Lipofectamine 2000 (Invitrogen). After 4 days of differentiation, the cells were fixed with 100% ice-cold methanol, permeabilized with 0.25% TritonX100, blocked in 5% normal donkey serum (NDS), and stained overnight with primary antibodies at 4°C, followed by incubation with the corresponding secondary antibodies or streptavidin coupled to fluorophores (Invitrogen). The following primary antibodies/lectins were used: biotin WGA (Biozol) 1:50, rabbit anti-INPP5K (Proteintech) 1:100, and mouse anti-PDI (Enzo) 1:300. Nuclei were stained with DAPI (10 μg/ml, Invitrogen). Images were taken with a Keyence microscope (BZ-X800E). Brightfield pictures were used for protrusion analysis, while fluorescence pictures were used for ER and glycosylation analysis. The experiments were performed three times. N2A cell differentiation was morphologically evaluated by measuring the length and numbers of dendrite-like protrusions of first-order, second-order, and third-order protrusions. The length and number of protrusions were measured by tracing protrusions manually with the tool “segmented line” in ImageJ. The length was measured in micrometers with the option “measure.” All single cells from at least five to seven images per condition and experiment were traced manually.

For measuring the total protrusion numbers of individual N2A cells, all traced/visible protrusions per cell were counted for all single cells in at least five to seven images per condition and experiment, and the mean per cell was calculated for each condition.

For measuring the total protrusion length per cell, the length of all traced/visible projections per cell was measured for all single cells from at least five to seven images per condition and experiment, and the mean per cell was calculated for each condition.

To determine the mean protrusion length per order, the length of all traced/visible protrusions per respective protrusion order was measured for each cell and each condition.

N2A (ATCC) cells were cultured in DMEM GlutaMAX (Sigma-Aldrich) supplemented with 10% [v/v] FBS (Biowest) and 1% [v/v] penicillin/streptomycin (Gibco) at 37°C. The N2A cells were seeded in 10-cm cell culture dishes (Greiner). The following day, the cells were treated with differentiation medium [Neurobasal medium 1 g/L glucose (Gibco) + 1X B-27 (Gibco) + 1X sodium pyruvate (Gibco) + 1X glutamine (Gibco) + 1% [v/v] penicillin/streptomycin (Gibco)] and at the same time transfected with siRNA against either control (siScr; Dharmacon) or INPP5K (siInpp5k; Thermo Fisher) according to the manufacturer's protocol using Lipofectamine 2000 (Invitrogen). After 3 days of differentiation, the cells were harvested and lysed in RIPA buffer [50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% [v/v] NP-40, 1% [w/v] sodium deoxycholate, 0.1% [w/v] SDS, 1 mM EDTA] and a complete protease inhibitor (Roche). Cell homogenates were centrifuged at 10,000 g, and the protein concentration in the supernatant was determined using the BCA assay kit (Thermo Fisher). Samples were stored at −20°C until further use. The experiments were performed three times.

N2A (ATCC) cells were cultured in DMEM GlutaMAX (Sigma-Aldrich) supplemented with 10% [v/v] FBS (Biowest), 1% [v/v] penicillin/streptomycin (Gibco) at 37°C. The N2A cells were seeded on poly-D-lysine (Thermo Fisher) coated coverslips. The following day, the cells were treated with differentiation medium [Neurobasal medium 1 g/L glucose (Gibco) + 1X B-27 (Gibco) + 1X sodium pyruvate (Gibco) + 1X glutamine (Gibco) + 1% [v/v] penicillin/streptomycin (Gibco)] and allowed to differentiate for 3 days. Then, the cells were transfected with siRNA against either control (siScr; Dharmacon) or INPP5K (siInpp5k; Thermo Fisher) according to the manufacturer's protocol using Lipofectamine 2000 (Invitrogen). After 4 days of differentiation, cells were fixed and processed as described above. The experiments were performed three times.

Pregnant mice or pups were sacrificed, and the brain tissues were dissected. The brain tissue was obtained from 3 E11.5 and 3 E18.5 embryos as well as from 3 postnatal day (P) 3 and 3 P18 pups. Tissue lysates were prepared as previously described (Franzka et al., 2022). Briefly, samples were homogenized with the Potter S tissue homogenizer (Sartorius) in RIPA buffer [50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% [v/v] NP-40, 1% [w/v] sodium deoxycholate, 0.1% [w/v] SDS, 1 mM EDTA] and a complete protease inhibitor (Roche). After sonication, homogenates were spun down at 16,900 g to remove nuclei and insoluble debris. Protein concentration in the supernatant was determined using the BCA assay kit (Thermo Fisher) and then stored at −80°C until further use.

Proteins isolated from three embryos/pups at the indicated developmental stages or from N2A cells (transfected with siRNA against either control or INPP5K) from three individual experiments were denatured at 90°C for 5 min in Laemmli buffer (4X Laemmli buffer: 50% glycerol, 5% SDS, 0.25% 1.5 M Tris pH 6.8, 30% β-mercaptoethanol, and 0.001% bromophenol blue, ddH₂O). After separation by SDS-PAGE, proteins were transferred onto PVDF membranes (Whatman). Membranes were blocked in 2% BSA and incubated with primary antibodies at appropriate dilutions overnight at 4°C. The following primary antibodies or lectins were used: rabbit anti-INPPK (Proteintech) 1:500, rabbit anti-GAPDH (Proteintech) 1:1,000, biotin WGA (Biozol) 1:300, biotin PNA (Bioworld) 1:300, biotin SNL (Biozol) 1:300, biotin RCAI (Biozol) 1:300, biotin Con A (Biozol) 1:300, biotin LCH (Ey Laboratories) 1:300, and biotin MAL (Biomol) 1:300. Primary antibodies or lectins were detected with HRP-conjugated secondary antibodies or HRP-conjugated streptavidin. Detection was performed using the Clarity Western ECL Substrate Kit (BioRad). The quantification of bands was performed using ImageJ.

Coomassie blue staining of PVDF membranes was performed as described previously (Franzka et al., 2021). For Coomassie blue staining of transferred proteins, PVDF membranes were fixed for 3 min (10% acetic acid, 40% EtOH), stained in Coomassie blue solution (0.1% Brilliant Blue R (Serva), 45% EtOH, 10% acetic acid) for 5 min, destained (10% acetic acid, 20–40% EtOH), rinsed in H₂O, and imaged.

Pregnant mice or pups were sacrificed, and whole embryos or brain tissues were dissected. Tissue was obtained from each of the three E11.5, E13.5, and E18.5 embryos, as well as from the three postnatal day (P) 3 and P18 pups. The tissues were frozen in TissueTek (company) on dry ice and cryo-sectioned. Sections were dried, fixed in 4% paraformaldehyde (PFA), permeabilized with 0.25% TritonX100, blocked in 5% normal goat serum (NGS), and stained overnight at 4°C with primary antibodies or lectins, followed by incubation with the corresponding secondary antibodies or streptavidin coupled to fluorophores (Invitrogen). The following primary antibodies or lectins were used: biotin WGA (Biozol) 1:100, rabbit anti-INPP5K (Proteintech) 1:100, and biotin PNA (Bioworld) 1:100. Nuclei were stained with DAPI (10 μg/ml, Invitrogen). Images were taken using a Keyence microscope BZ-X800E.

For statistical analysis, raw data were analyzed for normal distribution with the Kolmogorov–Smirnov test or by graphical analysis using the Box-Plot and QQ-Plot in GraphPad Prism 9. If appropriate, we either used a one-way ANOVA, a two-way ANOVA, or a two-tailed Student's t-test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001, and ns, not significant. For statistical analysis, we used GraphPad Prism 9. For all data, means with the standard error of the mean (SEM) and individual data points with the SEM are shown.

To analyze the expression pattern of INPP5K during mouse development, we stained sagittal whole embryo or brain sections of different developmental stages with an antibody directed against INPP5K (Figure 1A, Supplementary Figures 2, 3). Prominent INPP5K signals were found in developing skeletal muscles and in the central nervous system (Figure 1A, Supplementary Figures 2, 3). In the brain, INPP5K expression was detected in various regions, especially in the hippocampus, the cortex, and Purkinje cells (Figure 1B). Excluding the unspecific binding of secondary antibodies, no signals were observed when the primary antibody was omitted (Supplementary Figure 4).

We next assessed the abundance of INPP5K in protein lysates from mouse brains isolated at different time points of embryonic development. Notably, INPP5K levels increased dramatically during embryonic as well as early postnatal (P) brain development (Figure 1C, Supplementary Figure 5).

Taken together, these findings suggest a pivotal role for INPP5K in the brain.

To assess whether INPP5K is relevant for cell differentiation, we transfected N2A cells with siRNAs directed against INPP5K (siInpp5k) or a scrambled control siRNA (siScr; Figure 2A, Supplementary Figure 5). We then serum-deprived the transfected N2A cells to induce neuronal-like differentiation and measured the numbers and length of first-, second-, and third-order protrusions 4 days later. While the number of protrusions did not differ from controls, the length of first-order dendrites was significantly decreased after the siRNA-mediated knockdown of INPP5K (Figure 2B, Supplementary Figure 1A). Dendrite numbers did not differ between conditions (Figure 2B, Supplementary Figure 1A).

To assess whether INPP5K is required for the maintenance of differentiated N2A cells, we induced differentiation for 4 days before we knocked down INPP5K. The number and length of the protrusions were evaluated 4 days later. While control cells maintained long primary protrusions, the length of primary protrusions was significantly reduced after siRNA-mediated knockdown of INPP5K in differentiated N2A cells (Figure 2C, Supplementary Figure 1B).

Patients with MDCCAID suffer from congenital muscular dystrophy with hypoglycosylation of dystroglycan (Osborn et al., 2017; Wiessner et al., 2017). To study whether the knockdown of INPP5K in N2A cells alters protein glycosylation, we isolated protein lysates of differentiated N2A cells transfected with either scRNA or siRNAs to knock down INPP5K upon the induction of differentiation. After separation by SDS-PAGE and blotting, membranes were incubated with different lectins to detect specific glycan structures. Upon knockdown of INPP5K, signals were diminished for wheat germ agglutinin (WGA), which detects N-acetylglucosamine residues, and for peanut agglutinin (PNA), which recognizes non-sialylated β(1–3)-linked galactose on N-acetylgalactosamine residues (Figure 3A, Supplementary Figure 5). No differences were found in non-sialylated/sialylated galactose/glucosamine on N-acetylglucosamine residues, as detected by ricinus communis agglutinin I (RCAI), sambucus nigra lectin (SNL) or maackia amurensis lectin (MAL). Similarly, no differences were found in fucosylated/mannose-carrying residues, as detected by the lens culinaris lectin (LCH) or concanavalin A (Con A; Figure 3A, Supplementary Figure 5).

We then fixed N2A cells either transfected with the scRNA or the siRNAs directed against INPP5K at the same time as the induction of differentiation and co-stained for N-acetylglucosamine residues and the luminal ER protein phosphodiesterase (PDI; Figures 3B, C). Consistent with our immunoblots, INPP5K signals were diminished upon knockdown of INPP5K. Notably, cells showing reduced INPP5K signals showed a prominent intracellular accumulation of WGA and a co-localization of PDI with WGA signals, suggesting that glycoproteins were retained within the ER (Figure 3B).

In summary, the knockdown of INPP5K in N2A cells affects protein glycosylation.

Because WGA and PNA reactivities were diminished in protein lysates of cells transfected with siRNAs to knock down INPP5K (Figure 3), we wondered about the developmental profile for N-acetylglucosamine residues detected by WGA and non-sialylated β(1–3)-linked galactose on N-acetylgalactosamine residues detected by peanut agglutinin (PNA).

We assessed the abundance of N-acetylglucosamine residues and non-sialylated β(1–3)-linked galactose on N-acetylgalactosamine residues in mouse brains isolated at different time points of embryonic and postnatal development. The respective protein lysates were separated by SDS-PAGE and blotted. For both glycan structures, we observed a strong increase between E11.5 and P18 (Figure 4, Supplementary Figures 2–5).