Cell cultureMaintenance of hES cells

Feeder-independent hES cells of line WA09/H9 (RRID: CVCL_9773; hPSCreg: WAe009-A) were obtained from WiCell and cultured on Matrigel-coated (15505739, Corning) dishes in mTeSR Plus medium (100-0276, StemCell Technologies). The medium was changed every other day and the cells passaged using ReLeaSR (05872, StemCell Technologies) every 3–5 days, depending on colony size. All cell cultures were maintained in a humidified incubator with 5% CO2 at 37 °C.

Maintenance of HEK cells and lentivirus production

Human embryonic kidney cells (HEK293T/17, American Type Culture Collection CRL-11268) were cultured at 37 °C with 5% CO2 in Dulbecco’s modified Eagle medium (DMEM)–GlutaMAX medium (31966047, Gibco) supplemented with 10% fetal bovine serum (FBS; F7524, Sigma). Medium was changed every 2 days and cells were split after reaching 70–80% confluence using trypsin-ethylenediaminetetraacetic acid (15400054, Gibco) or TrypLE (12605010, Gibco).

Lentiviruses were produced as described previously81, with slight modifications. HEK293 cells were seeded at 60% confluence and incubated 1 h before transfection with fresh medium supplemented with 25 μM chloroquine (C6628, Sigma). Cells were cotransfected using the calcium phosphate method with lentiviral helper plasmids as follows: 3.9 μg of pREV, 8.1 μg of pRRE, 6 μg of pVSVG and 12 μg of lentiviral vector DNA per 75 cm2 cell culture area. Medium was replaced again 2–3 h posttransfection. For constructs used on neurons, the medium was replaced with Neurobasal supplemented with 2% B27 (17504044, Gibco), GlutaMAX (35050061, Gibco) and 10 mM HEPES (15630080, Gibco). Lentiviruses were collected from the medium 40 h after transfection, pelleted by centrifugation at 1,500g for 10 min at 4 °C, aliquoted and frozen at −80 °C. For constructs used on ES cells to induce differentiation, medium replacement after transfection was done with fresh DMEM medium. Following collection and clearing, as described above, the lentiviral particles were pelleted by high-speed centrifugation (60,000g for 1.5 h), resuspended in MEM (51200046, Gibco) with 10 mM HEPES (100 μl per 30 ml of medium), aliquoted and snap frozen in liquid nitrogen.

Generation of iGluts

iGluts were generated from control (Ctrl1 and Ctrl2) and mutant (qKO1 and qKO2) ES cell clones according to previously described methods45. For each neuronal induction experiment, 250,000 hES cells were detached with Accutase (Gibco), plated on Matrigel-coated wells in mTeSR Plus containing Rho kinase inhibitor (Y27632, 1683, Axon Medchem, or Thiazovivin) and simultaneously transduced with lentiviruses FU–M2rtTA and Tet-O–Ngn2–puromycin. One day later (defined as DIV0), the medium was replaced with N2 medium (DMEM/F12; 11330032, Gibco), 1% N2 supplement (17502048, Gibco) 1% nonessential amino acids (11140050, Gibco), laminin (200 ng ml−1; 23017015, Thermo Fisher), brain-derived neurotrophic factor (BDNF) (10 ng ml−1; 450-02, Peprotech) and NT-3 (10 ng ml−1; 450-03, Peprotech) supplemented with doxycycline (2 μg ml−1, Alfa Aesar) to induce expression of Ngn2 and the puromycin resistance cassette. The following day, puromycin (1 mg ml−1) was added to the medium. After 48 h of selection, cells were detached with Accutase (A1110501, Gibco) and replated on Matrigel-coated coverslips along with mouse glia (typically at a density of 150,000 iGluts per 24-well) in B27 medium (Neurobasal-A (12349015, Gibco) supplemented with B27 (17504044, Gibco), GlutaMAX (35050061, Gibco) laminin, BDNF and NT-3). Half of the medium was replaced every second day for 8 days, with cytosine arabinoside (ara-C; C6645, Sigma) added to a working concentration of 2 μM to prevent glia overgrowth. Experimental lentiviral constructs (for example, to express liprin-α rescue constructs) were added to the medium on day 4. From DIV10, neuronal growth medium (Neurobasal-A supplemented with B27, GlutaMAX and 5% FBS (SH30071.03HI, Hyclone)) was washed in and used for partial medium replacements every 3–4 days until analysis, typically after 4–6 weeks in culture.

In experiments aiming to assess evoked synaptic transmission (Fig. 3), the protocol for generation of iGluts was slightly different. Specifically, cells from each clone were further separated into two groups. In group 1, cells were infected with pFU–M2rtTA, pTet-O–Ngn2–puromycin and with lentiviruses expressing Channelrhodopsin oChiEF fused to tdTomato (termed here ChR–tdTomato)82. In group 2, cells were infected with pFU–M2rtTA, pTet-O–Ngn2–puromycin and lentiviruses to express nuclear-localized GFP (nGFP). Four days later, cells from groups 1 and 2 were washed three times with phosphate-buffered saline (PBS) to remove any lentivirus trace, detached and mixed at a ratio of 80%/20% (80% with ChR and 20% with nGFP), reseeded on Matrigel-coated coverslips along with mouse glia and cultured as described above. To record evoked synaptic transmission GFP + TdTomato cells were patched in whole-cell voltage clamp configuration and the presynaptic inputs onto patched cells activated with brief (5–10 ms) pulses of blue light (488 nm) using a light-emitting diode.

Generation of iGABAs

iGABAs were generated according to published protocols46. hES cells were treated with Accutase (Sigma), then plated and immediately infected lenti-rtTA, lenti-Ascl1, and exposed to doxycycline 1 day later to drive expression of Acsl1 and Dlx2. Two days later, puromycin and hygromycin (H3274, Sigma) were added to the medium during 24 h for selection. After four additional days of hygromycin selection, remaining cells were detached with Accutase and replated on Matrigel-coated coverslips along with mouse glia. Half of the medium was then changed every second day for 8 days and 2.5% FBS was added to support astrocyte viability. After DIV10, induced GABAergic neurons were cultured in B27/Neurobasal medium containing GlutaMAX (Gibco), 5% FBS and 10 ng ml−1 BDNF until performing analysis.

Generation of induced astrocytes

Induced astrocytes were generated following previously published methods47. Briefly, control and mutant ES cells were treated with Accutase (Sigma) and then seeded on Matrigel-coated 24-well plates at a density of 90,000 cells per well. Cells were maintained in mTeSR Plus medium supplemented with Y27632 (Axon Medchem). Cells were then transduced with lentiviruses FU–M2rtTA, Tet-O–Sox9–puromycin and Tet-O–Nfib–hygromycin and kept in DMEM/F12 medium containing 10% FBS, 1% N2 supplement and 1% GlutaMAX (expansion medium). One day later, 2.5 μg ml−1 of doxycycline (D9891, Sigma) was added to the medium to drive expression of Sox9 and NF1B. Two days later, 1.25 μg ml−1 of puromycin and 200 μg ml−1 of hygromycin were added to the medium for selection. From day 3 onward, cells were kept in expansion medium, with the gradual addition of fibroblast growth factor (FGF) medium (Neurobasal-A medium supplemented with 2% B27, 1% nonessential amino acids, 1% GlutaMAX (all from Gibco) and 1% FBS (Sigma)), 8 ng ml−1 of FGF (100-18, Peprotech), 5 ng ml−1 of ciliary neurotrophic factor (450-13, Peprotech) and 10 ng ml−1 of BMP4 (120-05, Peprotech), with 2.5 μg ml−1 of doxycycline and 200 μg ml−1 of hygromycin, until expansion medium was completely replaced with FGF medium (also containing 2.5 μg ml−1 of doxycycline). Finally, on day 10, medium was replaced with B27-supplemented final medium (Neurobasal-A medium, 2% B27, 1% GlutaMAX and 5% FBS) containing 2.5 μg ml−1 of doxycycline. At day 21, induced astrocytes were detached and seeded along with induced glutamatergic neurons derived from Ctrl or qKO hES cells on Matrigel-coated coverslips.

Mouse glia cell isolation

Primary mouse glial cell culture was performed essentially as described previously83. Briefly, cortices from 0.5–2.5-day-old wild-type C57BL/6 mice of both sexes, housed under standard conditions in a 12/12 h light–dark cycle with food and water ad libitum, were dissected, pooled and triturated using a fire-polished Pasteur pipette followed by passage through a cell strainer. Cells were plated in flasks (two cortices/1× T75) precoated with poly-l-lysine (5 mg ml−1;P1274, Sigma) in DMEM supplemented with 10% FBS (Sigma). Upon reaching confluence, the glial cells were dissociated by trypsinization and reseeded twice to remove potential trace amounts of mouse neurons before the glia cell cultures were used for coculture with induced neurons. Animal procedures were approved by the Swedish Board of Agriculture, the Robert Koch Institute (Germany) and the ‘Regierungsprasidium’ Karlsruhe (Germany).

Cloning of plasmid constructsLentiviral rescue constructs

Human full-length complementary DNA clones for PPFIA1 (HsCD00460680) and PPFIA3 (HsCD00341187) were obtained from the Harvard PlasmID repository, and PPFIA2 (HsCD00877565) and PPFIA4 (HsCD00946340) from DNASU. Full-length cDNA and truncation mutants were PCR-amplified using PrimeSTAR (R010A, Takara) and gel purified using the QIAEX II DNA purification kit (20051, Qiagen). Using the HiFi DNA assembly mix (E2621S, NEB), the amplicons were inserted in a lentiviral vector (‘pFU-’) downstream of the ubiquitin promoter and an N-terminal enhanced (E)GFP fusion (amplified from pEGFP–N1; Clontech). Correct clones were verified by Sanger sequencing and amplified using the Midiprep Plus kit (12945, Qiagen). Point mutants were generated using the QuickChange Site-Directed Mutagenesis kit (210518, Agilent Technologies).

Other constructs

All constructs were cloned by Gibson assembly as described above. pFU–Venus–ELKS1 and pFU–mScarlet–ELKS1 were cloned by fusing the cDNA of human ELKS1 (HsCD00860679; DNASU) downstream of mVenus or mScarlet, respectively. pFS–HA–PTPRS was generated from cDNA of human PTPRS (short isoform lacking meA, meB and FN4-7; NM_130853) with an intracellular myc-tag. The HA tag was placed in the N-terminus by replacing the endogenous signal peptide with that of Ig-kappa, followed by an HA tag. pCMV–LRRTM2–GFP was cloned by inserting the cDNA of LRRTM2 (HsCD00419164; PlasmID repository at Harvard Medical School) in the vector pEGFP_N1. pCMV–IL1RAPL1–GFP was cloned by inserting the cDNA of human IL1RAPL1 (HsCD00082647; DNASU) in the vector pEGFP_N1. pCMV–NGL3–GFP was cloned by inserting the cDNA of rat NGL3/Lrrc4b in the vector pEGFP_N1. pCMV–Nlgn1–Cherry and pCMV–TrkC–Cherry were generated using the insert of corresponding GFP-tagged constructs. The Nlgn1 construct contains the rat cDNA lacking the A and B splice inserts. For a summary of plasmids used in this study, see Supplementary Table 2.

Gene editing of PPFIA1–4 sgRNA design and cloning

Exons to target were selected on the basis of the following criteria: (1) presence in all transcripts and (2) preferably containing a noninteger number of codons such that its full deletion would be expected to cause a frame shift. The design of single guide (sg)RNA sequences was aided by the CHOPCHOP design tool (v3)84. The following sgRNA sequences for PPFIA1 and 2 were cloned into SpCas9(BB)-2A–GFP (PX458) and sgRNAs for PPFIA3 and PPFIA4 in LentiCRISPRv2 plasmids, as described previously85,86, with protospacer adjacent motif sequences in bold:

PPFIA1 (exon 17): 5′-GTGCAGCCGGTCTAACCGAA GGG

PPFIA2 (exon 20_1): 5′-TGTTGGCACTACCAAGCCCG AGG

PPFIA2 (exon 20_2): 5′-TCTTCAATAGGACGTTTGTT TGG

PPFIA3 (exon 11): 5′-TAAGCGGCTGTCCGAGACGG TGG

PPFIA4 (exon 16): 5′-AGCGCGTCCCCACCACTCAG CGG

Gene editing of hES cells

Two liprin-α genes per electroporation experiment were simultaneously targeted by combining Cas9- and sgRNA-encoding plasmids containing either puromycin resistance (LentiCRISPRv2) or GFP (PX458) as selection markers. Cells at ~80% confluency were treated with 2 μM thiazovivin for 2 h before transfection and detached with Accutase. Per transfection, 500,000 cells were resuspended in solution P3 (V4XP-3032, Lonza), mixed with 1.5 μg of each plasmid and electroporated in a 16-Nucelocuvete strip using the 4D-Nucleofector system (Lonza) set at program CA-137. Immediately after completion of the pulse, cells were resuspended in 100 μl of equilibrated mTeSR Plus with thiazovivin and plated on Matrigel-coated 6-well plates. Cells were lifted 15 h posttransfection and GFP-positive cells sorted using a FACSAria III Flow Cytometer (BD) equipped with an automated cell deposition unit, using a 100 µm nozzle at 20 psi. Around 25,000 cells were sorted in bulk and plated on 2× wells of a 6-well plate. The medium was changed the next day and a 24 h period of puromycin selection (at 1 mg ml−1) was started (48 h posttransfection). Colonies were collected 1 week later for screening of mutant clones. A single clone with unambiguous null alleles in PPFIA1 and PPFIA4 was isolated and used for further targeting. After the simultaneous editing of PPFIA2 and PPFA3, no clone with bi-allelic disruption of PPFIA2 could be obtained and the cells were thus subjected to a final round of editing with a new sgRNA toward the same PPFIA2 exon. This resulted in the isolation of clones qKO1 and qKO2.

Selection and screening of ES cells

Selected ES cells clones were initially screened for indels by PCR (HotStarTaq, Qiagen) followed by fragment analysis (‘IDAA’), essentially as described previously87. Selected clones were further analyzed by Sanger sequencing (Eurofins Genomics). For compound heterozygous clones, the PCR product was first cloned using the TOPO-TA kit (Thermo Fisher Scientific), to isolate allelic reads. Sanger traces were analyzed using Geneious Prime software and comparisons with the parental H9 line using the TIDE algorithm (v3.3.0; http://tide.nki.nl)88. The following primers were used: (PCR PPFIA1 flanking exon 17) F: 5′-ATGCCGACCATCAGCGAAG-3′; R: 5′-TCTCTTTCCACTCGTGCTTGG-3′; (PCR PPFIA2 flanking exon 20) F: 5′-GACTCACACTCTCCCTTCTTCC-3′; R: 5′-GTCTTCGATCCTTCTCAGCTTG-3′; (PCR PPFIA3 flanking exon 11) F: 5′-GACCTTGCCCGAGATAGAGG-3′; R: 5′-ACCACTGCCAGCCACATAG-3′; (PCR PPFIA4 flanking exon 16) F: 5′-CGGCATTGAGGGAAGAGTCT-3′; R: 5′-CACTGGGCAGGGTCATGA-3′.

CRISPR and AAV genome editing

To generate an HA-tagged NRXN1 knock-in line, we used the ‘AAV-cTr’ vector, previously described elsewhere54. A simplified protocol was used to produce adeno-associated virus (AAV) particles. HEK293T/17 cells were cotransfected using calcium phosphate with the plasmid and AAV (serotype DJ) helper plasmids. After transfection, the medium was replaced to mTeSR Plus and incubated for 72 h. AAV particles were collected from the cleared conditioned medium supernatant, washed and concentrated using 15 ml Centrifugal Filter Units (Amicon Ultra-4 100 kDa molecular weight cutoff; Merck).

CRISPR targeting with RNP complexes

For the formation of ribonucleoprotein complexes (RNP), a synthetic sgRNA targeting the 3′-UTR of NRXN1 (Integrated DNA Technologies) was incubated with Alt-R S.p. HiFi Cas9 Nuclease V3 (1081060, Integrated DNA Technologies) for 10 min at an equimolar sgRNA:Cas9 ratio in a concentration of 37 μM. The genomic sgRNA target sequence (with protospacer adjacent motif in bold) is 5′-TTGGGTTGGCTATAGAAAAG AGG. Briefly, 300,000 cells from control (Ctrl1) and mutant (qKO1) pretreated with thiazovivin were transfected with RNP complexes, as described above, and immediately infected with 4.5 μl of AAV supernatant expressing NRXN1–cTR-targeting vector as repair template. Targeted cells were selected with puromycin for 72 h and single-cell sorted by fluorescence-activated cell sorting for isolation of monoclonal lines.

Transfection of iGlut cells

Transfection of iGluts for analysis of axonal transport was performed at DIV7 by calcium phosphate. Medium was removed and kept aside in a replicate plate at 37 °C. Cells were briefly washed with MEM, and CaPO4 precipitates were applied for a 25 min incubation period. Precipitates were prepared as follows: 1 μg of plasmid DNA, 2 μl of 2 M CaCl2 and sterile water to a final volume of 15 μl were vortex-dropwise added to 15 μl of 2× HBS buffer pH 7.05 (274 mM NaCl, 1.4 mM Na2HPO4, 10 mM KCl, 15 mM D-glucose and 42 mM HEPES). Crystals were removed by two washes with 1× Hank’s balanced salt solution buffer (without CaCl2/MgCl2; Gibco) and one wash with MEM (Gibco) before returning the cells to the original conditioned medium.

Immunocytochemistry and SIM imaging

Cultured iGluts were fixed with prewarmed paraformaldehyde (PFA) solution (4% PFA and 4% sucrose in PBS, pH 7.4) for 15 min at room temperature (RT). Then, cells were washed three times in PBS (10 min each) and permeabilized with 0,1% Triton X-100 in PBS for exactly 10 min at RT. Blocking was performed for 1 h in a blocking buffer (2% goat serum, 1% bovine serum albumin and 0,01% NaN3 in PBS). Primary antibodies diluted in the blocking buffer were applied overnight at 4 degrees inside a humid chamber. Cells were then washed three times with PBS and fluorescent-labeled secondary antibodies were incubated for 1 h at RT. Finally, cells were washed three times in PBS and once in double-distilled H2O and mounted in microscope slides using ProLong Gold mounting medium (Thermo Fisher Scientific). For PSD95 staining, immunofluorescence was performed with the following modifications: neurons were maintained in culture for 52–55 days and fixed in ice-cold methanol fixing solution (90% methanol, 10% 2-(N-morpholino)ethanesulfonic acid buffer: 100 mM 2-(N-morpholino)ethanesulfonic acid pH 6.9, 1 mM ethylene glycol tetraacetic acid (EGTA) and 1 mM MgCl2) at RT for 5 min. Cells were washed three times in PBS and incubated in blocking–permeabilizing solution (2% goat serum, 1% bovine serum albumin, 0.01% NaN3 and 0.1% Triton X-100 in PBS) for 30 min, before proceeding with staining. The following primary antibodies were used (for details, see Supplementary Table 3): MAP2 (Encor, 1:1000), pan-synapsin (Proteogenix, 1:1,000), PSD95 (NeuroMab, 1:100 and Addgene 1:100 for SIM experiments), RIM1/2 (SySy, 1:200), Munc13-1 (SySy, 1:200), SV2 (DSHB, 1:500), bassoon (Sigma, 1:200), RIMBP-2 (SySy, 1:200), synaptophysin-1 (SySy, 1:200), CASK (Neuromab, 1:200), ERC1/2 (SySy, 1:200), piccolo (SySy, 1:200), CaV2.1 (SySy, 1:200), Tuj1 (Biolegend, 1:1,000) and HA (Biolegend, 1:200). For analysis of synaptic markers, cells were imaged using either a Nikon Eclipse Ti2 or a Leica SP8 confocal microscope, in both cases using a 60×/numerical aperture (NA) 1.4 oil immersion objective. Images were acquired, processed and analyzed with the experimenter blinded to the sample genotype/condition using either NIS Elements software or a custom ImageJ macro, respectively.

SR-SIM imaging

Images were acquired using an Elyra 7 microscope lattice SIM (Zeiss) with a plan-apochromat 63×/1.4 NA oil objective, controlled via ZEN black (3.0 SR, v16.0.17.306). Each image consisted of three image channels sequentially acquired in the following order: PSD95, MAP2 and neurexin-1–HA, labeled with secondaries 568, 405 and 488, respectively (to avoid photobleaching of 405 channel signal by 488 nm excitation). PSD95 and neurexin-1 channels were coregistered by using a reference sample (multicolor beads of size ~100 nm, subdiffraction-limited registration accuracy). MAP2 channel was excited at 405 nm (2.0%, 30 µW) and emission collected via a dual-color emission filter (BP420-480 + BP495-550, exposure 150 ms), neurexin-1 was excited at 488 nm (2.0%, 80 µW) and emission collected via a dual-color emission filter (BP420-480 + BP495-550, exposure 200 ms) and PSD95 was excited at 561 nm (1.5 %, 63 µW) and emission collected via a dual-color emission filter (BP570-620 + LP655, exposure: 150 ms). To minimize image shifts between channels, a single dichroic filter was used with a quadruple bandpass design (LBF 405/488/561/642). The SIM gratings used were 27.5, 27.5 and 32 µm for 405, 488 and 561 nm excitation, respectively. Lattice SIM three-dimensional processing for each channel independently was done using ZEN Black 3.0 SR software (v.16.0.17.306, Zeiss). Z-stack reconstruction and nanocluster analysis was performed using ImageJ/Fiji, with PSD95 and NRXN1 nanocluster size and number quantified using the SynapseEM plugin and MATLAB script.

Heterologous synapse formation assay and receptor clustering assayHeterologous synapse formation

HEK293T cells were plated at a confluency of 60% and transfected with plasmids expressing fluorescent recombinant postsynaptic receptors (mCherry–Nlgn1, mCherry–TrkC, pVenus–Nlgn1, YFP–TrkC, EGFP–NGL3 and YFP–ILRAPL1) with the calcium phosphate method as described above. pmCherry–N1 or pEGFP–N1 (Clontech) was transfected as negative controls. After 24 h, transfected HEK cells were collected with 0.5 mM ethylenediaminetetraacetic acid (15575-020, Thermo Fisher) in Dulbecco’s phosphate-buffered saline (14190144, Gibco), passed through a 35 μm cell strainer and plated in coculture with iN cells (DIV17) at a density of 20,000 per coverslip (200 cells μl−1). After 2 days, 48 h after coculture with neurons, DIV19 (72 h posttransfection), cells were fixed with 4% PFA/4% sucrose solution for 15 min at RT. Immunolabeling of presynaptic components was performed with the following antibodies: rabbit anti-pan-synapsin (E028 or nc30-1, 1:1,000), rabbit anti-piccolo (Sysy, 1:200), rabbit anti-bassoon (Sysy, 1:200), mouse anti-SV2 (DSHB, 1:500) or mouse anti-synaptophysin (Sysy, 1:200). The signal of EGFP–liprin-α proteins was enhanced by an anti-GFP antibody (DSHB, 1:500). Species-specific AlexaFluor 405 (1:1,000), AlexaFluor 488 (1:1,000), AlexaFluor 568 (1:1,000) and AlexaFluor 633 (1:600) secondaries (Supplementary Table 3) were used and samples mounted using Prolong gold (P36930, Thermo Fisher Scientific). Images were collected with a Nikon Eclipse Ti2 confocal microscope using a 40×/NA 1.15 water immersion objective. Quantification of presynaptic specialization analysis was performed by NIS elements AR software (v.5.21.01, Nikon Instruments). Normalized values of recruitment signal were assessed by quantifying the binary area of the markers recruited onto the surface of HEK293T cells per total area of HEK293T cells expressing fluorescent-tagged postsynaptic receptors. Background correction for the 633 channel was employed using a constant for HEK293T artificial synapse formation assays. All images were acquired and analyzed with the experimenter blinded to the sample genotype/condition.

Receptor clustering assay

HeLa cells (ATCC-CCL-2) were cultured in DMEM (Corning) supplemented with 10% FBS (Pan Biotech) and 50 U ml−1 penicillin and streptomycin. Transfections of indicated plasmids were performed with Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. One day after transfection, the cells were detached by trypsin treatment and subcultured onto ~20 μg ml−1 fibronectin (Millipore)-coated coverslips for additional 24 h. After fixation with 4% PFA, the cells were stained with indicated primary antibodies followed by fluorescent dye-conjugated secondary antibodies. Confocal images were acquired with a Nikon A1R confocal microscope. For immunofluorescence, primary antibodies against Flag (Sigma, 1:200 dilution) and HA (Cell Signaling, 1:200) were used. AlexaFluor 594-conjugated anti-mouse IgG or 647-conjugated anti-rabbit IgG was diluted 1:1,000. Images were analyzed using ImageJ.

Electrophysiological recordingsGeneral

On the day of recording, a coverslip containing induced neurons was placed in an RC-27 chamber (Sutter Instruments), mounted under a BX51 upright microscope (Olympus), equipped with differential interference contrast and fluorescent capabilities. Neurons were maintained at 26 ± 1 °C using a dual TC344B temperature control system (Sutter Instruments). Induced neurons were continuously perfused with oxygenated (95% O2/5% CO2) artificial cerebrospinal fluid solution containing (in mM): 125 NaCl, 2.5 KCl, 0.1 MgCl2, 4 CaCl2, 25 glucose, 1.25 NaH2PO4, 0.4 ascorbic acid, 3 myo-inositol, 2 Na-pyruvate and 25 NaHCO3, pH 7.4 and 315 mOsm. In a subset of experiments (Extended Data Figs. 1 and 2), the concentration of MgCl2 and CaCl2 was changed to 1 and 2 mM, respectively. Cells were approached and patched under differential interference contrast using 3.0 ± 0.5 MegaOhm glass pipettes (WPI), pulled with a PC10 puller (Narishige). Depending on the experimental configuration (see below), pipettes were filled with either voltage or current clamp internal solution containing (in mM) voltage clamp: 125 Cs–gluconate, 20 KCl, 4 MgATP, 10 Na–phosphocreatine, 0.3 GTP, 0.5 EGTA, 2 QX314 (HB1030, Hello Bio) and 10 HEPES–NaOH, pH 7.2, and current clamp 125 K-gluconate, 20 KCl, 10 HEPES, 0.5 EGTA, 4 ATP–magnesium, 0.3 GTP–sodium and 10 Na–phosphocreatine, osmolarity 312 mOsmol and pH 7.2, adjusted with KOH. For all experiments, a Multiclamp 700B amplifier (Axon Instruments) controlled by Clampex 10.1 and Digidata 1440 digitizer (Molecular Devices) were used. Detection and analysis of voltage and current clamp recordings were done with Clampfit 10.1 or with custom-written macros in Igor Pro 6.11. Electrophysiological recordings were done and analyzed with the experimenter blinded to the sample genotype/condition.

Current clamp recordings

In some experiments shown in Extended Data Fig. 1, whole-cell current clamp recordings from induced glutamatergic neurons were performed. In these experiments ~4 MΩ pipettes were used, and automatic bridge balance was performed after achieving whole-cell current clamp configuration. The membrane potential in all neurons was maintained at approximately −70 mV by injecting the appropriate feedback current into the cells. Current injections <50 pA were considered acceptable and those cells in which higher current injections were required were not included in the analysis.

Voltage clamp recordings

In most recordings (Figs. 3, 5 and 6), whole-cell voltage clamp was used. For recordings from induced glutamatergic neurons, membrane voltage was clamped at −70 mV, and miniature excitatory currents (recorded in the presence of 0.5 μM TTX; HB1035, Hello Bio) were detected as downward deflections. For recordings from induced GABAergic cells, membrane voltage was camped at 0 mV and inhibitory currents were recorded as upward deflections.

Evoked currents

In these experiments (Fig. 3), we recorded from GFP+/ChR− neurons (see above) in voltage clamp at −70 mV holding potentials, while simultaneously activating presynaptic inputs to recorded neurons with a single, short (5–20 ms) pulse of blue light (488), generated via a CoolLED illumination system (pE-300) controlled by a transistor–transistor logic pulse (Extended Data Fig. 4b).

Sucrose responses

In these experiments (Extended Data Fig. 2, and Figs. 5 and 6), cells were maintained at −70 mV holding potentials (voltage clamp configuration) and stimulated 0.5 M sucrose solution for 5 s. Sucrose solution was delivered in the vicinity of recorded cells (20–30 μm away), using a low-resistance glass pipette (1.5 MΩ), connected to a custom pressure device (5 psi).

Time-lapse microscopyAxonal transport

iN cells were plated on 35 mm four-compartment CellView dishes (627870, Greiner) at a cell density of 80,000 cells cm−2 and transfected at DIV6 with plasmids encoding mCherry and SV2–GFP, as described above. Imaging to assess axonal transport of vesicles and active zone components was performed at DIV13 and 19. Image acquisition was performed using a Nikon Eclipse Ti2 confocal equipped with a humidity- and CO2-controlled incubation chamber at 37 °C with a 40×/NA 1.15 water immersion objective. Images were obtained in fast-scan mode with an ~30 Hz frame rate for a total of 300 s per field of view. The resulting time-lapse movies were median filtered and background subtracted using the ‘detect local maximum’ function in the NIS elements AR software (v5.21.01, Nikon Instruments). Kymographs were generated and analyzed using the Multi Kymograph plugin of Fiji/ImageJ (v2.3.0/1.53f). Only moving puncta were analyzed and quantified.

Fluorescence recovery after photobleaching experiments

HeLa cells were plated on 35 mm four-compartment CellView dishes (Greiner) and transfected with the indicated liprin-α3 constructs using TransIT-X2 (MIR6003, Mirus Bio). The next day, cells were treated with 2 μM phorbol 12-myristate 13-acetate (10 min before onset of imaging) and transferred to a Nikon Eclipse Ti2 confocal equipped with an humidity- and CO2-controlled incubation chamber at 28 °C. Images were obtained using a 40×/NA 1.15 water immersion objective at an ~1 Hz frame rate before and after photobleaching a small region of interest (ROI) containing a cytoplasmic condensate using the 405 nm laser at 100% power. The same ROIs were used to measure fluorescence over time, using NIS elements AR software. For colocalization experiments, HeLa cells cotransfected to express mScarlet-fused ELKS and the indicated liprin-α3 constructs were imaged live after treatment with 2 μM phorbol 12-myristate 13-acetate. Intensity profiles across representative images and the green–red Pearson correlation coefficient were analyzed using NIS elements AR software.

Western blot

Protein samples were extracted from iN cultures at DIV19–22 lysed in radioimmunoprecipitation assay buffer (50 mM Tris pH 8.0, 150 mM NaCl, 0.1% sodium dodecyl sulfate, 0.5% sodium deoxycholate and 1% Triton X-100) supplemented with phenylmethylsulfonyl fluoride (36978, Thermo Fisher) and Complete Proteinase Inhibitor Cocktail (11873580001, Merck) for 20 min. Lysates were centrifuged at 20,000g for 10 min at 4 °C and supernatants containing solubilized proteins collected. Protein samples (30 µg each) in Laemmli buffer, reduced with dithiothreitol (0.1 mM, final concentration) were heated to 96 °C for 5 min and separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis in precast tris-glycine TGX gels (Bio-Rad). Transfer to a nitrocellulose membrane (Amersham) was performed in Towbin transfer buffer (25 mM tris, 0.2 M glycine and 20% methanol). Membranes were blocked with 5% nonfat milk (Aplichem) for 1 h and primary antibodies were incubated overnight at 4 °C. After washing the membranes three times with TBS-T (20 mM tris pH 7.5, 137 mM NaCl and 0.05% Tween-20), species-specific 680RD- or 800CW-conjugated secondary antibodies (LI-COR, all at 1:10,000 dilution) were incubated in 1:1 TBS-T Odyssey Blocking (927-50000, LI-COR) for 1 h. Membranes were imaged using an Odyssey CLx or DLx system (LI-COR). Immunoblotted bands were quantified by densitometry using Image Studio 5.2 software (LI-COR). Loading controls on the same membrane were used to normalize data. For quantitative comparisons, each measurement was normalized to the average value per blot, with the median value of controls set to 1. The following primary antibodies were used (for details, see Supplementary Table 3): liprin-α1, -α2, -α3 or -α4 (all used at 1:200 dilution), PTPRS (MediMabs, 1:1,000), neurexin-1 (Millipore, 1:1,000), RIM1 (SySy, 1:1,000), ELKS1/2 (SySy, 1:1,000), Munc13 (SySy, 1:1,000), RIMBP-2 (SySy, 1:1,000), CASK (Neuromab, 1:1,000), Nlgn1 (Neuromab, 1:500), Homer1 (SySy, 1:1,000) syntaxin-1 (SySy, 1:1,000), synapsin-2 (Sigma, 1:1,000), PSD95 (Thermo Fisher Scientific, 1:500), Veli123 (SySy, 1:1,000), ERC1/2 (SySy, 1:1,000), Mint-1 (SySy, 1:1,000), Rab3a (SySy, 1:1,000), SNAP25 (Sigma, 1:2,000), β-actin (Sigma, 1:1,000), synaptotagmin-1 (SySy, 1:1,000), Tuj1 (BioLegend, 1:5,000) and GFP (Thermo Fisher Scientific, 1:1,000).

EM

Neurons grown on glass coverslips were fixed in Karnovsky fixative (2.5% glutaraldehyde, 2% formaldehyde and 0.02% NaN3 in 0.05 M cacodylate buffer) at 37 °C for 25 min. The samples were subsequently washed five times with 0.1 M cacodylate buffer for a total of 1 h. The fourth change of buffer contained 50 mM glycine (blocking residual aldehydes from fixative). Staining was performed with 1% osmium and 1% potassium ferrocyanide in 0.1 M cacodylate buffer, for 20 min at RT and washed with water three times. Tertiary staining was made with 2% uranyl acetate for 20 min at RT and after three more washes in water, samples were dehydrated with ethanol (4 min each of 30%, 50%, 70%, 85%, 90% and four times with 100% ethanol). Next, samples were infiltrated in Agar 100 resin (AGR1140, Agar Scientific), through a series of increasing concentration of resin (15 min each of 25%, 50% and 75% and three times with 100%). Embedding was performed with resin–benzyldimethylamine (Electron Microscopy Sciences) in BEEM capsules (TAAB Laboratories Equipment Ltd). Samples were polymerized for 48 h at 60 °C, and subsequently sectioned at 70 nm and mounted on noncoated copper grids (mesh size 150). Before imaging, sections were contrasted with Raynold’s lead citrate for 5 min. Images were acquired using a Talos L120C transmission electron microscope (Thermo Scientific). Subsequent image analysis was performed using ImageJ/Fiji (v2.3.0/1.53f).

Analysis of synaptic vesicle counts

Synaptic vesicles were defined, for the purpose of this analysis, as all spherical vesicles with a diameter <68 nm within 1,000 nm of a PSD-like structure. The number of total synaptic vesicles per bouton, diameter of synaptic vesicles, PSD length and distance of each synaptic vesicle to the active zone were analyzed using the SynapseEM ImageJ plugin and a MATLAB script, as described in ref. 89.

Data analysis, statistics and reproducibility

Current and voltage clamp recordings were analyzed using Clampfit v10.2 (Molecular Devices) or written macros in Igor Pro v4.07 (WaveMetrics). Confocal images were handled and analyzed using NIS elements AR software (v5.21.01; Nikon Instruments), LASX (Leica) or ImageJ/Fiji (v2.3.0/1.53f) and numerical data processed in Excel (v16; Microsoft). EM and SIM images were analyzed using MATLAB (R2022a; MathWorks). Immunoblot images were handled and analyzed with Image Studio (v5.2; LI-COR). Sequence data were analyzed using Geneious Prime software (BioMatters). Stem cell work was performed in compliance with the German Stem Cell Act approved by the Robert Koch Institute.

Allocation (for example, the distribution of different experimental lentiviruses on separate coverslips, order of analysis and so on) was random. No statistical methods were used to predetermine sample sizes; the number of datapoints and independent repetitions was guided by previous studies 54,78,81,83. Results from multiple (typically three) independent experiments were performed, as indicated in figure legends and Supplementary Table 1, and the results were merged. Representative experiments were repeated at least once, except for screening PCRs (Extended Data Fig. 3e), the western blot confirming liprin-α1 deletion in human astrocytes (Extended Data Fig. 1c) and the liprin-α3 rescue condition for EM.

Summary data are shown as means ± s.e.m. Statistical analysis was performed using Prism 9 (GraphPad Software). Datasets were tested for normality (Gaussian distribution) using the D’Agostino Pearson test. For between-group comparisons, unpaired two-tailed t-tests were used if data distribution was normal, or two-tailed Mann–Whitney tests for non-Gaussian datasets. For multiple-group comparisons, statistical significance was determined by analysis of variance with Tukey’s or Holm–Šídák’s corrections for multiple comparisons, or Kruskal–Wallis followed by Dunn’s post hoc test for non-Gaussian datasets. ***P < 0.001, **P < 0.01 and *P < 0.05.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.