Alternative splicing controls teneurin-latrophilin interaction and synapse specificity by a shape-shifting mechanism.

Cell culture High-Five insect cells

(Trichoplusia ni, female, ovarian. Thermo Fisher, B85502 ) cultured in Insect-Xpress medium (Lonza, 04351Q) supplemented with 10 μg/mL gentamicin at 27 °C were used for production of recombinant proteins. HEK293T mammalian cells (ATCC, CRL-3216) were used for cell-surface expression assays and flow cytometry binding assays and were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, 11965092) supplemented with 10% FBS (Sigma-Aldrich, F0926) at 37 °C in 5% CO 2.

Cloning and expression in insect cells

TEN2 splice variant Lasso (UniProt: Q9NT68 -2) and LPHN (LPHN1, UniProt: O88917 ; LPHN3, UniProt: Q9HAR2 ) constructs were cloned into a pAcGP67a vector and expressed in High-Five insect cells using the baculovirus expression system. Sf9 cells (Thermo Fisher, 12659017) were co-transfected with the linearized baculovirus DNA (Expression Systems, 91-002) and the constructed plasmid using the Cellfectin II (Thermo Fisher, 10362100) transfection reagent. Baculovirus was amplified in Sf9 cells in SF900-III medium containing 10% (v/v) FBS (Sigma-Aldrich, F0926). Large-scale protein expression was performed by infection of High Five cells (Thermo Fisher, B85502 ) in Insect-XPRESS medium (Lonza, 12-730Q) medium at a cell density of 2.0 × 10 6 cells/ml for 72 h at 27 °C. For the structural studies, TEN2 ECRΔ1 (residues T727-R2648) and LPNH3 ECR (residues S21-V866) were cloned with carboxyl-terminal 6XHis-tags separately and co-expressed in High-Five insect cells. The following primers were used for amplification of High Five cells expressed human TEN2 ECRΔ1: F: 5′-CATTCTGCCTTTGCGGCGGATCCCACTTCCTGTGCTGATAACAAGGATAATGAG-3′ and R: 5′-GGATCAGATCTGCAGCTTAGTGATGGTGATGGTGATGCCTCTTTCCCATCTCATTCTGTC-3′. The following primers were used for amplification of High Five cells expressed human LPHN3 ECR: F: 5′-CATTCTGCCTTTGCGGCGGATCCCTCCCGCGCACCCATTCC-3′ and R: 5′-GGATCAGATCTGCAGCTTAGTGATGGTGATGGTGATGCACGTCCAGCAGCAGATCGTG-3′. Seventy-two hours after viral infection, the medium containing secreted glycosylated proteins was collected and centrifuged at 900 g for 15 min at room temperature. The supernatant was transferred into a beaker and mixed with (final concentrations): 50 mM Tris pH 8.0, 5 mM CaCl 2 and 1 mM NiCl 2 and stirred for 30 min. After centrifugation at 8000 g for 30 min, the clarified supernatant was incubated with nickel-nitrilotriacetic agarose resin (QIAGEN) for 3 h at room temperature. The resin was collected with a glass Buchner funnel and rinsed with HBS buffer containing 20 mM imidazole, then transferred to a poly-prep chromatography column (Bio-rad). The protein was eluted with HBS buffer containing 200 mM imidazole and run on size-exclusion chromatography (Superdex 200 10/300 GL; Superose 6 Increase 10/300 columns; GE Healthcare), purified in a final buffer comprised of 10 mM Tris pH 8.5, 150 mM NaCl. For the flow cytometry binding assays, LPHN1 Lec (residues S26-Y131) and LPHN3 Lec (residues S21-Y126) were cloned with carboxyl-terminal 6XHis-AVI-tags and captured on nickel-nitrilotriacetic resin as described above. Following a wash with HBS buffer containing 20 mM imidazole, final concentrations of 50 mM Bicine pH 8.3, 100 mM NaCl, 10 mM Mg-acetate, 10 mM ATP, 0.5 mM biotin and 5 mM BirA were added to the resin, which was then rotated for 2 h at 27 °C. After removing residual BirA and ATP by washing with HBS buffer containing 20 mM imidazole, the biotinylated lectin was eluted with HBS buffer containing 200 mM imidazole. Purified protein was applied to size-exclusion chromatography. The efficiency of biotinylation was assessed using a streptavidin bead pulldown assay. Cloning and expression in mammalian cells Full-length TEN2 (residues M1-R2648) construct and TEN2 mutants (DHR mutant: D1737N, H1738T, R1739T; LR mutant: L1990N, R1992T) bearing HA-tag (inserted between K405/E406) and carboxyl-terminal FLAG-tag were cloned into a pcDNA3.1 vector for cell-surface expression assays and flow cytometry binding assays in HEK293T cells. The following primers were used for amplification of HEK cells expressed human TEN2 ECR: F: 5′-GGATGACGACGATAAAGGCGGTAAGCTTAGCCCACCTCTC-3′ and R: 5′-TTACTTATCGTCGTCATCCTTGTAATCCCTCTTTCCCATCTCATTCTGTCTT-3′. The following primers were used for amplification of HEK cells expressed human TEN2 ΔTox: F: 5′-GGATGACGACGATAAAGGCGGTAAGCTTAGCCCACCTCTC-3′ and R: 5′-TTACTTATCGTCGTCATCCTTGTAATCTTCATAGGGAGGAGGCACGAAATACAT-3′. The following primers were used for amplification of HEK cells expressed human TEN2 ΔToxΔBarrel: F: 5′-GGATGACGACGATAAAGGCGGTAAGCTTAGCCCACCTCTC-3′ and R: 5′- TTACTTATCGTCGTCATCCTTGTAATCGAAGGCATTAAGAACAGGCTTGTTC-3′. TEN2 DHR mutants were generated using a standard two-step PCR-based strategy with primers: F: 5′-ATTCGGACTGAAAAGATCTATGATAACACCACGAAGTTCACCCTGAGGATCATTTATG-3′ and R: 5′-CATAAATGATCCTCAGGGTGAACTTCGTGGTGTTATCATAGATCTTTTCAGTCCGAAT-3′. TEN2 LR mutants were generated using a standard two-step PCR-based strategy with primers: F: 5′-AGTGAGACTCCCCTCCCCGTTGACAACTACACCTATGATGAGATTTCTGGCAAGGTG-3′ and R: 5′-CACCTTGCCAGAAATCTCATCATAGGTGTAGTTGTCAACGGGGAGGGGAGTCTCACT-3′.

Flow cytometry

HEK293T cells were cultured in 6-well plates and were transfected 2 μg cDNA using LipoD293 transfection reagent. Cells at 50-60% confluence were transiently transfected as follows: 2 µg cDNA was diluted in 50 µl serum-free DMEM, and 3 µl LipoD293 transfection reagent (SignaGen, SL100668) was diluted with 47 µl serum-free DMEM. The diluted LipoD293 was added to the diluted cDNA and incubated for 10 min. Then, the transfection mixture was added dropwise to each well. The cells were detached using citric saline solution (50 mM sodium citrate, 135 mM KCl) after 48 h incubation and washed with PBS + 2% BSA. To test TEN2 WT and mutant cell-surface expression, cells were stained with a primary antibody mixture: mouse anti-FLAG M2 (Sigma, F3165) 1:1000 and rabbit anti-HA (Life Technologies, 715500) 1:1000 for 30 min at room temperature. After wash with PBS + 2% BSA, cells were stained with a secondary antibody mixture: donkey anti-mouse Alexa Fluor 488 (nvitrogen, A21202 ) 1:3000 and goat anti-rabbit Alexa Fluor 647 (Invitrogen, A32733 ) 1:3000 for 30 min. After washing, cell pellets were resuspended in PBS + 2% BSA immediately before flow cytometry data acquisition (Accuri C6 flow cytometer, 10000 events measured) after washing. Acquired data were analyzed using the FlowJo analysis software (FlowJo LLC). For the binding assays, His-Avi-tagged Lec was captured on nickel-nitrilotriacetic resin and purified as described above. Biotinylated Lec was tetramerized and fluorescently labeled through incubation with NeutrAvidin DyLight 488 (Thermo, 22832) on ice for 20 min. Cultured cells expressing HA-tagged TEN2 were detached and then washed as described above. Next, the cells were stained with rabbit anti-HA 1:1000 antibody and, following two wash cycles, stained with goat anti-rabbit Alexa Fluor 647 antibody in the presence of the 100 nM NAV488 labeled Lec mixture. The following primers were used for amplification of His-Avi-tagged human LPHN1 Lec: F: 5′-CGGCGGCGCATTCTGCCTTTGCGGCGAGCCGGGCTGGACTCCCATTTGG-3′ and R: 5′-CTTCTGAGCCTCGAAAATATCATTAAGACCGCGGTAAGGGACACAGTCGTACTGC-3′. The following primers were used for amplification of His-Avi-tagged human LPHN3 Lec: F: 5′-GGCGGCGCATTCTGCCTTTGCGGCGTCCCGCGCACCCATTCCTATGGCCG-3′ and R: 5′-TTCTGAGCCTCGAAAATATCATTAAGACCGCGATATGGCACGCACTCGTACTGCACT-3′.

Cell-aggregation assays HEK293T cells

(ATCC) were grown to 90% confluence in a T-75 flask. Cells were trypsinized with 3 mL 0.05% trypsin-EDTA (Gibco, 25300-054) and resuspended to 10 mL with DMEM/10% FBS/1% Penicillin–Streptomycin media (Complete DMEM). Three-hundred µL of the cell suspension was added to each well of a 6-well plate containing 3 mL of Complete DMEM media and incubated overnight at 37 °C. Cells in each well were then co-transfected with 2 µg of either pCMV (empty vector) + pEmerald, pCMV LPHN3 + pEmerald, pCMV (empty vector) + pCMV dsRed, or dsRed and the indicated TEN2 construct using the Calcium Phosphate method. All cDNAs were encoded in the pCMV5 or pcDNA3 vector and driven by the CMV promoter. Three days after transfection, the media was aspirated and cells were gently washed with 2 mL of PBS. Cells were resuspended by adding 1 mL of Resuspension Solution (PBS containing 1 mM EGTA) and then incubated for 5 min at 37 °C. Fifteen µL of 1 mg/20 µL DNAse (Sigma, D5025) was then added to each well and cells were triturated by pipetting up-and-down (16 times) in each well to resuspend cells off the plate bottom and create single-cell suspensions. Cells were then transferred to a new Eppendorf tube and another 15 µL of DNAse solution was added to each sample. Cells were mixed in 1:1 ratio by adding 70 µL of pCMV (empty vector) + pEmerald or LPHN3 + pEmerald with 70 µL of pCMV (empty vector) + pCMV-dsRed or TEN2 Construct + dsRed in a new Eppendorf that contained 360 µL of Incubation Solution (DMEM containing 50 mM HEPES-NaOH pH 7.4, 10% FBS, 10 mM CaCl 2 and 10 mM MgCl 2 ) for a final volume of 500 µL. The mixture was triturated and the entire volume was transferred to one well in a non-coated 12-well plate (Costar, 3737). Images were taken immediately (time = 0) using a Leica Fluorescent DMIL LED Microscope with a 10x objective. Cells were then placed on a shaking incubator at 125 rpm at 37 °C for 20 min and imaged again (time = 20). Aggregation index at time = 20 was calculated using ImageJ, measuring the percentage of signal/frame occupied by cells forming complexes of two or more cells relative to the total signal of the frame.

Cell-surface-binding assays HEK293T cells

(ATCC) were grown to 90% confluence in a T-75 flask. Cells were trypsinized with 3 mL 0.05% trypsin-EDTA (Gibco, 25300-054) and resuspended to 10 mL of DMEM + 10% FBS + 1% Penicillin–Streptomycin (complete DMEM) media. Fifty µL of cell suspension was added to each well of a 24-well plate that contained a Matrigel-coated coverslip and 1 mL complete DMEM and incubated overnight. Cells were then co-transfected with 1 µg of either empty pCMV, wild-type Teneurin 2 or the indicated mutant Teneurin construct and 1 µg of pEmerald using the Calcium Phosphate method and incubated for 2 days at 37 °C. Transfection media was gently removed and 500 µL of chilled DMEM containing 250 µM of purified, Avi-fusion, biotinylated, rat LPHN1 Lec or human LPHN3 Lec was added to each well. Plates were wrapped in foil and incubated overnight at 4 °C to reduce endocytosis, with gentle shaking. This was performed essentially as described in 55 . Cells were gently washed 2x using 1 mL of PBS and fixed with 300 µL of ice-cold 4% PFA/4%sucrose/PBS. Plates were wrapped in foil and incubated for 20 min at 4 °C during the fixation. Cells were gently washed 3× using 1 mL of room temperature PBS and blocked with 300 µL of 5% BSA (Sigma, 10735086001)/PBS (blocking buffer) for 1 h at room temperature. Bound biotinylated Lec was detected by immunofluorescence using 300 µL per well of Streptavidin (AlexaFluor-555 conjugated, Invitrogen, S21381 , at 1:10,000 dilution) diluted into blocking buffer for 1 h at room temperature. Cells were gently washed 3× with 1 mL of PBS. Cells were re-blocked, and HA-tagged, surface Teneurins were detected by adding 300 µL of rabbit anti-HA antibody (Cell Signaling Technologies, 3724) at 1:1,000 dilution in blocking buffer. Cells were gently washed 3× with 1 mL PBS. Goat anti-rabbit secondary antibodies (Alexa-Fluor 633 conjugated, Invitrogen) and DAPI (Sigma, 10236276001) staining was done for 30 min at 1:10,000 and 1:5,000, respectively, in blocking buffer, followed by 3× gentle washes with 1 mL of PBS. Coverslips were mounted onto slides (UltraClear microscope slides Denville Scientific, M1021) in mounting media (Fluoromount-G, Southern Biotech, 010020). Images were acquired using a Nikon A1 Eclipse Ti2 confocal microscope with a ×60 oil-immersion objective, operated by NIS-Elements AR acquisition softw×are. The same confocal acquisition settings were applied to all samples of the experiment. Collected z-stacks at a 0.4 µm z-step size were analyzed blindly using Nikon Elements Analysis software. Co-localization was calculated using the Pearson’s correlation coefficient of Lec-Streptavidin-555 to Teneurin-HA-633 emission.

Artificial synapse formation assay

HEK293T cells were transfected with the expression vectors of the cell-adhesion molecules. 24 h later, HEK293T cells were co-cultured with cultured cortical neurons (DIV16) from P0 mice. After 24 h, cells were fixed with 4% PFA and immunostained with rabbit anti-Flag (Sigma; 1:1000 both) together with mouse anti-PSD95 (Sysy, 124011, 1:500) or mouse anti-GABAA α2 (Sysy; 224211, 1:500) respectively. Images were collected with a Nikon A1 confocal microscope using a ×60 objective. A human NPR mutant (1-118 aa of the full-length protein) which comprises an A domain containing low-complexity sequences is used as the negative control in the artificial synapse formation assays. The signals of the synaptic markers that were recruited to the surface of the HEK293T cells were quantified using Image J. Normalized values equal the fluorescent intensity of the synaptic marker that was examined (GABAα2/PSD95) / the fluorescent intensity of the Flag-tagged protein expressed in the HEK cells (TENs /Nrn1β).

Cryo-EM data acquisition 2.5 μl purified human TEN2

ECRΔ1 and human LPHN3 ECR complex (0.22 mg/ml) was applied on glow-discharged holey carbon grids (Quantifoil R1.2/1.3, 300 mesh), and vitrified using a Vitrobot Mark IV (FEI Company). The specimen was visualized using a Titan Krios electron microscope (FEI) operating at 300 kV and equipped with a K3 direct electron detector (Gatan, Inc.). Images were recorded with a nominal magnification of ×81,000 in super-resolution counting mode, corresponding to a pixel size of 0.54 Å on the specimen level. To maximize data collection speed while keeping image aberrations minimal, image shift was used as imaging strategy using one data position per hole with four holes targeted in one template with one focus positio. In total, 4967 images with defocus values in the range of −1.0 to −2.5 μm were recorded using a dose rate of 14.6 electrons/Å 2 /s. The total exposure time was set to 4.2 s with frames recorded every 0.105 s, resulting in an accumulated dose of about 60.1 electrons per Å 2 and a total of 40 frames per movie stack.

Image processing and 3D reconstructions

Stack images were subjected to beam-induced motion correction using MotionCor2 56 . CTF parameters for each micrograph were determined by CTFFIND4 57 . Particle selection, two- and three-dimensional classifications were initially performed on a binned dataset with a pixel size of 4.32 Å using RELION-3 58 . In total, 4,475,958 particle projections were selected using automated particle picking and subjected to reference-free two-dimensional classification to discard false-positive particles or particles categorized in poorly defined classes, resulting in 3,307,148 particle projections for further processing. The initial 3D maximum-likelihood-based classification was performed on a binned dataset with a pixel size of 4.32 Å using the previously reported TEN2 structure 2 as the reference model. The detailed data processing flow is shown in Supplementary Figs. 2 and 3 . Briefly, for Tenurin–Letrophilin complex, 1,309,684 particles that showed well-defined density of Lec domain were selected after initial rounds of 3D classification. Then, two rounds of focused 3D classification with mask around Lec domain were performed without alignment. 3D refinement and post-processing was performed on the best class with clear features for the Lec domain (Supplementary Fig. 2 ). The final map for TEN2_Lec was resolved at 2.97 Å (Supplementary Fig. 2 ). To resolve domain 3, the same dataset was reprocessed with a total of 1,137,765 particles showing well-resolved domain 3. Two rounds of focused 3D classification with mask around domain 3 were performed, followed by 3D refinement and post-processing. The final map for TEN2_domain 3 was resolved at 3.07 Å (Supplementary Fig. 3 ). Reported resolutions are based on the gold-standard Fourier shell correlation (FSC) using the 0.143 criterion (Supplementary Fig. 1c ). All density maps were corrected for the modulation transfer function (MTF) of the K3 direct detector and then sharpened by applying a temperature factor that was estimated using post-processing in RELION-3. Local resolution was determined using ResMap 59 with half-reconstructions as input maps (Supplementary Fig. 1d ).

Model building and refinement

Model building was based on the structure of human TEN2 ECR (PDB: 6CMX ) and the Lec domain from human LPHN3 (PDB: 5AFB and 5FTT). The models were first docked into the EM density maps using Chimera 60 and then manually checked and adjusted residue-by-residue to fit the density using COOT 61 . The ECR of TEN2 was built based on that of chicken TEN2 (PDB: 6FB3 ) and manually adjusted to human sequence and splice form. Note that the Lec domain was not as well resolved as TEN2, so it was docked as a rigid body without fitting and manipulating the side chains. Both maps (TEN2/LPHN3 and TEN2 focusing on domain 3) were used for model building. There is a slight shift between the two maps from reconstruction, so they were aligned based on TEN-Lec before model building. The final model containing both ECR of TEN2 and Lec domain of LPHN3 was subjected to global refinement and minimization in real space using the phenix_real_space_refine module in Phenix 62 first against TEN2 domain 3 map while keeping the Lec domain as a rigid body, and then against TEN2-Lec map while keeping the domain 3 as a rigid body. FSC curves were calculated between the resulting model and either maps using Phenix M-triage (Supplementary Fig. 1 ). The final model statistics are provided in Table 1 . Nine N-linked glycosylation sites (on residues N1490, N1586, N1647, N1681, N1766, N1867, N2071, N2211, N2522.) and five disulfide bonds (C1394-C1402), (C1396-C1404), (C1106-C1109), (C1210-C1218), (C1277-C1330) were observed in TEN2.

Quantification and statistical analysis

Error bars in Figs. 5 , 6 , 7 , 8 and Supplementary Fig. 6 represent means ± SEM. Each measurement was repeated at least three times independently. Data were analyzed using software GraphPad Prism and ImageJ. Reporting summary Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Supplementary information Supplementary Information Peer Review File Reporting Summary