Abstract
Glycine receptors (GlyRs) are the major mediators of fast synaptic inhibition in the adult human spinal cord and brainstem. Hereditary mutations to GlyRs can lead to the rare, but potentially fatal, neuromotor disorder hyperekplexia. Most mutations located in the large intracellular domain (TM3-4 loop) of the GlyRα1 impair surface expression levels of the receptors. The novel GLRA1 mutation P366L, located in the TM3-4 loop, showed normal surface expression but reduced chloride currents, and accelerated whole-cell desensitization observed in whole-cell recordings. At the single-channel level, we observed reduced unitary conductance accompanied by spontaneous opening events in the absence of extracellular glycine. Using peptide microarrays and tandem MS-based analysis methods, we show that the proline-rich stretch surrounding P366 mediates binding to syndapin I, an F-BAR domain protein involved in membrane remodeling. The disruption of the noncanonical Src homology 3 recognition motif by P366L reduces syndapin I binding. These data suggest that the GlyRα1 subunit interacts with intracellular binding partners and may therefore play a role in receptor trafficking or synaptic anchoring, a function thus far only ascribed to the GlyRβ subunit. Hence, the P366L GlyRα1 variant exhibits a unique set of properties that cumulatively affect GlyR functionality and thus might explain the neuropathological mechanism underlying hyperekplexia in the mutant carriers. P366L is the first dominant GLRA1 mutation identified within the GlyRα1 TM3-4 loop that affects GlyR physiology without altering protein expression at the whole-cell and surface levels.
SIGNIFICANCE STATEMENT We show that the intracellular domain of the inhibitory glycine receptor α1 subunit contributes to trafficking and synaptic anchoring. A proline-rich stretch in this receptor domain forms a noncanonical recognition motif important for the interaction with syndapin I (PACSIN1). The disruption of this motif, as present in a human patient with hyperekplexia led to impaired syndapin I binding. Functional analysis revealed that the altered proline-rich stretch determines several functional physiological parameters of the ion channel (e.g., faster whole-cell desensitization) reduced unitary conductance and spontaneous opening events. Thus, the proline-rich stretch from the glycine receptor α1 subunit represents a multifunctional intracellular protein motif.
Introduction
Neurotransmission in the human CNS essentially relies on the balance of excitation and inhibition at synaptic sites. Ligand-gated pentameric ion channels of the Cys-loop receptor (CLR) family, comprising excitatory serotonin [5-hydroxytryptamine (5-HT3A)] and nicotinic acetylcholine receptors, inhibitory GABAA receptors, and glycine receptors (GlyRs), regulate the sensitive balance by selective permeation of cation and anions, respectively (Lynch, 2004). These ion channels feature a common structural organization. Every subunit is composed of an extracellular ligand binding domain, followed by four α-helical transmembrane segments (TMs) that are connected by loops of various lengths and a short extracellular C terminus. While these elements share sequence similarities among different CLR family members, the large intracellular TM3-4 loop is considered to be the region with highest diversity and is part of the intracellular domain (ICD) of the receptor (Goyal et al., 2011). In prokaryotic CLR homologs, the TM3-4 loop is virtually absent (Bocquet et al., 2009; Hilf and Dutzler, 2009).
Synaptic GlyRs are predominantly expressed in the adult spinal cord and brainstem nuclei as heteropentameric proteins, where they act as key mediators of fast inhibition in motor reflex circuits (Yang et al., 2012; Patrizio et al., 2017). Human hyperekplexia (startle disease; Online Mendelian Inheritance in Man #149400) is often caused by hereditary mutations in GlyR subunit genes (GLRA1, GLRB). Hyperekplexia is characterized by exaggerated startle episodes in response to unexpected stimuli, generalized and epileptic hypertonia, as well as muscle stiffness (Bakker et al., 2006). Mutated genes associated with the disease in the order of affected frequency are the GLRA1, SLC6A5 (glycine transporter 2), GLRB, GPHN (gephyrin), and ARHGEF9 (collybistin; Harvey et al., 2004; Carta et al., 2012; Bode and Lynch, 2014). Mutations in GLRA1 can lead to disturbances in receptor trafficking and receptor function (Chung et al., 2010).
In this study, we analyzed a novel dominant GLRA1 mutation P366L located in the TM3-4 loop of the GlyRα1 subunit focusing on ion channel expression and function (P366L numbering refers to the mature peptide of the short GlyRα1 isoform without the splice cassette SPMLNLFQ in the TM3-4 loop). So far, several hyperekplexia-associated mutations have been identified within the TM3-4 loop, all of which significantly reduced GlyR surface expression (Villmann et al., 2009). Unfortunately, information about the structure of the TM3-4 loop of any CLR members is sparse. However, molecular modeling and circular dicroism spectroscopy described α-helical organizations close to TM3 and TM4 (Burgos et al., 2015). Furthermore, spectral signals of another secondary structure, a poly-proline helix type 2 (PPII) has been detected for the TM3-4 loop (Breitinger et al., 2004), forming a consensus sequence for binding to Src homology 3 (SH3) domains (Cascio et al., 2001). The PPII helix XXPXXP is highly conserved among all GlyR subunits with P366 being part of this structure. Notably, within the GlyRβ subunit, an overlapping region (435KPPPAKP441) was shown to mediate the direct intracellular interaction with syndapin I (also called PACSIN1), a membrane-shaping F-BAR domain protein involved in membrane trafficking, cytoskeletal dynamics, synapse formation, and GlyR organization (Del Pino et al., 2014).
Here, we investigated the pathophysiology of a novel hyperekplexia mutation localized within the PPII helix in the GlyRα1 subunit, which revealed novel interaction partners of the synaptic glycine receptors. In this study, we used electrophysiological and biochemical techniques to understand the molecular mechanisms by which the GlyR mutant variant α1P366L disrupted glycinergic signaling. We identified functional defects and an impaired syndapin I-mediated targeting that together likely contribute cumulatively to the hyperekplexia phenotype. Thereby, this study reveals a critical direct contribution of the intracellular PPII helix within the GlyRα1 subunit to both receptor function and targeting, a function that so far has only been attributed to the GlyRβ subunit.
Materials and Methods
Direct sequencing from genomic DNA.
Genomic DNA isolated from blood samples of male and female index patients was prepared according to the manufacturer protocol (DNeasy Blood and Tissue Kit, Qiagen). Specific intronic primers of the GLRA1 gene were used for PCR amplification of the appropriate exons followed by direct sequencing (LGC Genomics). These experiments were authorized by the Ethics Committee of the Medical Faculty of the University of Würzburg, Germany (reference no. 2019071702).
Site-directed mutagenesis.
cDNAs encoding the human GlyRα1 wild-type (α1wt) or mutated GlyRα1 [α1X, with X explaining the amino acid exchange of the mutant such as proline at position 366 into leucine (P366L)] were subcloned into the pRK5 expression vector. Mutations were introduced using site-directed mutagenesis. Primers (Thermo Fisher Scientific) contained the appropriate nucleotide exchanges. Mutated clones were sequenced to verify successful mutagenesis (Eurofins Genomics).
Animals.
CD1 (strain #022, Charles River Laboratories) and C57BL/6 (stock #664, The Jackson Laboratory) pregnant female mice were used to isolate embryonic hippocampal cultures. Brain tissue was isolated from adult male mice. All experiments were performed in accordance with the guidelines set by the European Union and approved by our institutional Animal Care Committee, the Utilization Committee, and the Regierung von Unterfranken (Würzburg, Germany). At embryonic day 15 (E15) pregnant Wistar rats were killed via CO2 inhalation, as approved by the University of Queensland Animal Ethics Committee, Australia (QBI/203/13/ARC).
Cell lines and transfection.
HEK293 cells (ATCC CRL-1573) were grown in Life Technologies Minimum Essential Medium (Thermo Fisher Scientific) supplemented with 10% fetal calf serum (FCS), glutaMAX (200 mm), sodium pyruvate (100 mm), and penicillin (10,000 U/ml)/streptomycin (10,000 µg/ml). The cells were incubated at 37°C and 5% CO2. HEK293 cells were transiently transfected with 1 µg of cDNA at 75% confluency of cells using a modified calcium-phosphate precipitation method. For visualization in electrophysiological recordings 1 µg of green fluorescent protein (GFP) cDNA was cotransfected. For pull-down experiments, HEK293 cells were transfected with GlyRα1 using Lipofectamine 2000 (Thermo Fisher Scientific).
Artificial synapse formation.
Primary cultures of embryonic spinal cord neurons were prepared as described previously (Zhang et al., 2015). Following dissection of the spinal cord, the tissue was trypsinized [10 ml of trypsin/EDTA (1 mg/ml) supplemented with 100 μl of DNase I, final concentration, 0.1 mg/ml] for 30 min at 37°C. One milliliter of FCS (final concentration, 10%) was added to stop the digestion. Cell were gently triturated using a fire-polished Pasteur pipette. A total of 80,000 cells were seeded on poly-d-lysine-coated coverslips with a diameter of 18 mm in Neurobasal medium supplemented with 2% B27 and 1% GlutaMax (Thermo Fisher Scientific). Neurons were maintained in culture for up to 4 weeks. One to 3 d before recordings, heterosynaptic cocultures were prepared by plating HEK293 cells 24 h post-transfection with either GlyRα1wt or GlyRα1P366L onto the primary neuronal cultures.
Hippocampal cultures.
Hippocampal neurons were prepared from embryonic CD1 mouse embryos at stage E18. Mice were killed, and hippocampi were dissected. Hippocampi were trypsinized using 10 ml of trypsin/EDTA (1 mg/ml) and 100 μl of DNase I (final concentration, 0.1 mg/ml) for 30 min at 37°C. Trypsinization was stopped with 1 ml of FCS (final concentration, 10%). Hippocampal neurons were triturated three times. Supernatants with cells were centrifuged at 800 rpm for 10 min. Cells were counted and plated on poly-d-lysine-coated coverslips using a density of 150,000 cells/3 cm dish. Neurons were grown in Life Technologies Neurobasal medium (Thermo Fisher Scientific) containing 5 ml of l-glutamine (200 mm) and Life Technologies B27 supplement (Thermo Fisher Scientific) with an exchange of 50% medium after 7 d in culture. Neurons older than day in vitro 13 (DIV13) were used for experiments.
Lentiviral vector.
cDNAs encoding the human GlyRα1 wild-type (α1wt) or GlyRα1P366L were cloned into a modified form of the lentiviral transfer vector pFUGW (Lois et al., 2002), resulting in hGlyRα1IRESGFP and hGlyRα1P366L-IRESGFP (in which “h” signifies human). To monitor transduced cells, all vectors express GFP under control of an internal ribosomal entry site 2 (IRES2). Lentiviral particles were produced in HEK293T cells. For this, the lentiviral expression vectors were cotransfected with a pseudotyping vector (pMD2.G), expressing VSV-G (vesicular stomatitis virus G) protein, and the packaging vector pCMVΔR8.91 (Zufferey et al., 1998) using Lipofectamine 2000 (Thermo Fisher Scientific). Lentiviral particles were separated from the supernatant by ultracentrifugation and stored at −80°C as follows (in mm): 50 Tris–HCl, pH 7.8, 130 NaCl,10 KCl, 5 MgCl2. The number of infectious particles was determined using serial dilutions of the viral vectors on HEK293T cells. For transduction of hippocampal neurons, a multiplicity of infection of 20 was used. Neuronal transduction was performed during cell seeding. Hippocampal neurons were analyzed at DIV14.
Immunocytochemical stainings: HEK293 cells.
To label surface receptors only, live stain experiments on transfected HEK293 cells were performed. Cells were incubated with the GlyRα1-specific antibody mAb2b (residues 1–10 of the mature GlyRα1; catalog #146111, Synaptic Systems). After primary antibody incubation, cells were fixed with 4% paraformaldehyde (PFA) and 4% sucrose in phosphate-buffered saline (PBS). In contrast to surface GlyRα1, whole-cell GlyRs were stained following fixation, blocking and permeabilization (0.1% Triton X-100, 5% NGS in PBS) with the pan-α antibody mAb4a. MAb4a recognizes a denatured N-terminal epitope at fixed cells (residues 96–105, catalog #146011, Synaptic Systems). Cells were incubated with secondary antibody goat-anti-mouse Cy3 (1:500; Dianova) for 1 h. After washing with PBS cells were incubated with DAPI (1:5000) for 5 min and mounted with Mowiol.
Hippocampal neurons.
At DIV15, live cell stainings of surface GlyRs were performed using the α1-specific mAb2b antibody for 2 h at 4°C. After 20 min of fixation (4% PFA, 4% sucrose), cells were blocked and permeabilized with 5% goat serum and 0.1% Triton X-100 for 30 min. Endogenous syndapin I was stained overnight with the rabbit anti-syndapin I antibody (Qualmann et al., 1999) at 4°C. Subsequently, cells were washed and incubated with secondary antibodies goat-anti-mouse Cy3 and goat-anti-rabbit Cy5 (Dianova) for 2 h at room temperature. Following DAPI staining (1:5000), cover slips were mounted with Mowiol.
Data analysis for quantification of syndapin I was performed using the software Openview described in the study by Tsuriel et al. (2006). The signal of the stained GlyRα1 (mAb2b) was used as a mask. Syndapin I signal intensities were measured within mAb2b clusters and separated into whole-cell, soma, and neurite signals (three independent experiments; 4–7 z-projections were analyzed).
Biotinylation of cell surface protein.
Forty-eight hours post-transfection, cell surface biotinylation experiments were performed from transfected HEK293 cells, as described previously (Atak et al., 2015). Following incubation of the transfected cells with biotin and an additional quenching step, cells were lysed, and membrane proteins solubilized in Tris-buffered saline (TBS: 50 mm Tris, 150 mm NaCl, pH 8.0) containing 0.1% Triton X-100, 0.1 mm PMSF, 2 mm EDTA, and cOmplete, EDTA-free protease inhibitor cocktail (Roche). Following centrifugation, supernatants were taken as whole-cell fraction. The supernatant was further supplied with streptavidin beads to bind surface-exposed labeled proteins. The bead fraction therefore represents thesurface pool. Whole-cell and surface fractions were separated on 11% SDS-PAGE gels. Western blots were labeled with mAb4A for GlyRαreceptor protein and a pan-cadherin antibody (Cell Signaling Technology). Cadherin was used as a membrane marker protein.
Immunoprecipitation.
HEK293 cells were grown on 10 cm dishes as described above and used 48 h after transfection. Cells were washed once with PBS and cell lysis was performed with TBS with 1% TritonX-100 and cOmplete EDTA-free protease inhibitor cocktail (Roche). Fifty microliters of 10% DOC solution [10% (w/v) sodium-desoxycholate in 20 mm Tris/HCl, pH 9.0] was added to 600 µl of cell lysate and rotated for 10 min at 22°C. Probes were centrifuged for 1 min at 13,000 × g, and the supernatant was used as input for coimmunoprecipitation experiments. The protein solution (450 µl) was supplemented with mAb2b antibody (1:200; Synaptic Systems) and 50 µl of protein A-Sepharose beads/protein G-agarose mix (1:1; GE Healthcare), and agitated overnight at 4°C. Following a short centrifugation step, the supernatant was removed. The beads were washed three times with NET buffer (160 mm NaCl, 5 mm EDTA, 1% TritonX-100 in 50 mm Tris HCl, pH 8). Proteins were eluted from beads by incubating in 30 µl of 2× SDS sample buffer at 95°C for 7 min. Proteins were separated on 11% SDS gels and stained after Western blotting with rabbit anti-syndapin I antibody (catalog #196002, Synaptic Systems) and afterward with mouse anti-GlyR pan-α antibody (catalog #146011, Synaptic Systems) for input detection and precipitation control of the GlyR.
Electrophysiological whole-cell recordings.
Whole-cell patch-clamp recordings were used to measure maximal currents (Imax) from transfected HEK293 cells 48 h post-transfection. Signals were amplified with an EPC 9 amplifier controlled by Pulse Software (HEKA). One millimolar glycine was applied using a U-tube system, which enabled bathing the suspended cell in a laminar flow of solution with a time resolution for equilibration of 10–30 ms. The external buffer consisted of the following (in mm): 137 NaCl, 5.4 KCl, 1.8 CaCl2, 1 MgCl2, and 5 HEPES, with pH adjusted to 7.4 with NaOH. The internal buffer was as follows (in mm): 120 CsCl, 20 N(Et)4Cl, 1 CaCl2, 2 MgCl2, 11 EGTA, and 10 HEPES, with pH adjusted to 7.2 with CsOH. Recording pipettes were fabricated from borosilicate capillaries (World Precision Instruments) with an open resistance of 6–8 MΩ. Cells were held at −60 mV, and recordings were performed at room temperature (∼22°C).
Concentration–response curves were constructed from the peak current amplitudes obtained with at least seven appropriately spaced concentrations in the range 3–1000 μm glycine. Using a nonlinear algorithm (Microcal Origin Software), concentration–response data were first analyzed using the following Hill equation Iglycine/Isat = [glycine]nHill/[glycine]nH + EC50nH, where Iglycine refers to the current amplitude at a given glycine concentration, Isat is the current amplitude at saturating concentrations of glycine, EC50 is the glycine concentration producing half-maximal current responses, and nH is the Hill coefficient.
For desensitization analysis, whole-cell current traces were transferred to Microcal Origin (Microcal Origin Software). The equation of a single exponential function plus a constant was used to analyze the decay currents in presence of 1 mm glycine, as follows:
Iobs represents the observed total current amplitude, I1 the fraction of current desensitizing with the time constant τ1, and Iconst is the amplitude of the nondesensitizing current fraction. A single exponential decay plus a constant current term were sufficient to describe desensitization behavior. The desensitizing current fractions were calculated by fitting the curves from the peak current of the trace to the 2 s time point where glycine application ended.
Estimated data were compared using the one-sample t test (GraphPad Prism, GraphPad Software) followed by Bonferroni post hoc test, with a probability of error of p < 0.05 considered significant.
Electrophysiological single-channel recordings.
To minimize the formation of homomeric receptors, heteromeric GlyRs were expressed by cotransfecting the cDNAs encoding the α1 and β subunits at an α1/β ratio of 1:100. All recordings were performed at room temperature (22 ± 1°C) at a clamped potential of −70 mV, unless otherwise indicated, in an outside-out patch configuration. The intracellular solution was composed of the following (in mm): 145 CsCl, 2 CaCl2, 2 MgCl2, 10 HEPES, and 10 EGTA, adjusted to pH 7.4 with CsOH. The extracellular solution was composed of the following (in mm): 140 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES, and 10 d-glucose, adjusted to pH 7.4 with NaOH. The single-channel conductance (γ) was determined by constructing current–voltage (I–V) plots from data obtained at clamped voltages of ±70, ±35, ±15, and 0 mV. The conductance was calculated using the following:
where i is the unitary (single-channel) current, Vhold is the holding potential (−70 mV), Vljp is the liquid junction potential, and Vrev is the reversal potential. Vljp was calculated to be 4.7 mV, and Vrev was estimated directly from the I–V plots. Single-channel currents were recorded using an Axon 200B amplifier (Molecular Devices), filtered (−3 dB, four-pole Bessel) at 5 kHz and sampled at 50 kHz. Currents were filtered offline to 5 kHz for making figures.
Voltage-clamp fluorometry.
Oocytes from female Xenopus laevis (Xenopus Express) were prepared as previously described (Pless et al., 2007) and injected with 10 ng of mRNA. The oocytes were then incubated at 18°C for 3 − 5 d in ND96 solution containing in the following (mm): 96 NaCl, 2 KCl, 1 MgCl2, 1.8 CaCl2, 5 HEPES, 0.6 theophylline, 2.5 pyruvic acid, and 50 μg/ml gentamycin, pH 7.4. Frog animal experiments were approved by University of Queensland Animal Ethics Committee (Approval no. QBI/059/13/ARC/NHMRC). Rhodamine methanethiosulfonate (MTSR) and 2-((5(6)-tetramethylrhodamine) carboxylamino) ethyl methanethiosulfonate (MTS-TAMRA; both from Toronto Research Chemicals) were dissolved in dimethylsulfoxide (DMSO) and stored at −20°C. On the day of recording, oocytes were incubated for 30 s in 10 μm MTSR or MTS-TAMRA dissolved in ice-cold ND96. Oocytes were then thoroughly washed and stored in ND96 for up to 6 h on ice before recording. Oocytes were placed in a recording chamber on an inverted microscope. The microscope was equipped with a high-Q tetramethylrhodamine isothiocyanate filter set (Chroma Technology, a Plan Fluor 40× objective lens (Nikon Instruments), and a Hamamatsu H7360-03 photomultiplier (Hamamatsu Photonics) coupled to a PMT400R photomultiplier subsystem (Ionoptix). A 150 W halogen lamp was used as light source. Cells were maintained at −40 mV by conventional two-electrode voltage clamp, and currents were recorded with a Gene Clamp 500B amplifier (Molecular Devices). Current and fluorescence traces were acquired at 200 Hz via a Digidata 1322A interface using Clampex 9.2 software (Molecular Devices). For analysis and display, fluorescence signals were digitally filtered at 1 − 2 Hz with an eight-pole Bessel filter. Results are expressed as the mean ± SEM of three or more independent experiments.
Preparation of whole-brain lysates.
After cervical dislocation, whole brains from adult male/female C57BL/6 wild-type mice were removed from the scull and rapidly cooled to 4°C. Brain tissue was homogenized with 1 ml lysate buffer per 200 mg using a pistol homogenizer with eight strokes at 900 rpm. Lysate buffer contained the following (mm): 20 HEPES, 100 KCH3COOH, 40 KCl, 5EGTA, 5 MgCl2, 5 DTT, 1 PMSF, 1% Triton X-100, and cOmplete EDTA-free protease inhibitor cocktail (Roche). The homogenate was centrifuged at 10,000 × g and 4°C for 15 min. The supernatant was aliquoted and flash frozen in liquid nitrogen, and stored at −80°C.
Peptide synthesis.
Peptides were synthesized as C-terminal amides in 50 µmol scale using a Life Technologies PAL-PEG-PS resin (Thermo Fisher Scientific) with a specific loading of 0.18 mmol/g. Automated microwave-assisted solid-phase peptide synthesis (SPPS) was conducted in a Liberty Blue microwave peptide synthesizer (CEM) with a 30 ml Teflon reaction vessel using peptide reagent from Sigma-Aldrich Chemie and Iris Biotech. Fluorenylmethoxycarbonyl protecting group (Fmoc)-protected amino acids were dissolved (0.2 m) in dimethylformamide (DMF). The activator solution was N,N′-diisopropylcarbodiimide (DIC; 0.5 m) in DMF with diisopropylethylamine (DIPEA; 0.01%), and ethyl 2-cyano-2-(hydroxyimino) acetate (Oxyma; 1 m) in DMF with DIPEA (0.01%). Microwave assisted synthesis was conducted based on the following protocol: deprotection (150 W, 15 s, 75°C, followed by 30 W, 60 s, 90°C), coupling (190 W, 180 s, 75°C, followed by 36 W,120 s, 90°C). Double coupling by Oxyma/DIC/amino acid in the relation (1:1:1 equivalence) in fivefold excess over the resin loading achieved peptide elongation. Peptide cleavage from the solid support was conducted by treatment of the resin for 2 h with a cleavage mixture containing trifluoroacetic acid (TFA)/triisopropylsilane (TIPS)/H2O/ethanedithiol/thioanisol (90:2.5:2.5:2,5:2,5, 3 ml). Peptides were precipitated in diethyl ether (−20°C) spun down and washed two more times. The pellets were dissolved in 50% H2O, 50% ACN, and 0.1% TFA, and purified by reverse-phase HPLC (Waters) with a ZORBAX 300SB-C18 PrepHT column (21.2 × 250 mm, 7 mm; Agilent Technologies) using a gradient of 20–50% acetonitrile over 25 min. Fractions (10 ml) were collected. Pure fractions were identified based on liquid chromatography (LC)-MS, and the purity of the pooled peptide fractions was determined by ultra-performance LC after lyophilization on a freeze dryer.
Peptide microarray synthesis.
For screening in microarray format, peptides were synthesized using a ResPep SL synthesis robot equipped with a Celluspot synthesis module and printed using a Slide spotter (Intavis AG). All peptides were synthesized in parallel on acid-soluble Fmoc-β-alanine etherified cellulose disks (area 0.12 cm2, loading 1.0 μmol/cm2; Intavis AG). Synthesis was based on Standart Fmoc-based SPPS using reagents from Sigma-Aldrich Chemie and Iris Biotech. Amino acids were coupled by adding two times 1.2 µl Oxyma/DIC/amino acid in the relation (1.1:1.5:1.0). Residual unreacted binding sites were capped by the addition of 4 µl of 10% acetic anhydride solution in DMF. Peptide deprotection was performed using 150 μl per disk of a side chain deprotection solution (TFA/TIPS/water/DCM: 80%, 3%, 5%, 12%) for 2 h at room temperature. Disks were then solubilized overnight in 250 μl of cellulose solvation solution (TFA/trifluoromethansulfonic acid/TIPS/water: 88.5%, 4%, 2.5%, 5%). Seven hundred fifty microliters of diethyl ether (−20°C) was added to the dissolved cellulose−peptide conjugates. The mixture was briefly agitated and precipitated at −20°C for 1 h. Precipitated conjugates were pelleted by centrifugation at 2000 × g. The pellet was additionally washed twice with fresh cold ether. After the final washing step, residual ether was evaporated and 500 μl of DMSO was added to resolve the cellulose–peptide conjugates. The cellulose−peptide conjugate stock solutions were stored at −20°C. Fifty microliters of peptide stock solutions were transferred to a 384-well plate, and mixed 5:1 with SSC buffer (150 μm NaCl, 15 μm Na3C6H5O7, pH 7.0). Thirty nanoliters of each peptide was transferred to each position on coated glass slides with the slide spotter (Intavis AG).
Peptide array analysis.
After drying overnight, the arrays were washed and equilibrated with TBS, pH 7.6. Syndapin I full length, syndapin I lacking the SH3 domain (sdpIΔSH3), as well as whole-brain lysate were incubated on the arrays at 4°C for 1 h. Protein binding to the spotted peptides was detected by chemiluminescence after Western analysis with antibodies. The resulting dot blots were analyzed using the array Analyze software (Active Motif), which defines the error range of each dataset by comparing the intensities of each peptide duplicate on the analyzed array. For following calculations, we only included peptide array datasets where >90% of the peptide spots were within a <5% error margin. The spot densities of the accepted datasets were then normalized against wild-type intensity, as indicated. Three peptide array duplicates, resulting in eight datasets of 8 × 240 peptides were used to calculate the average spot densities and the rms deviations.
Covalent immobilization of peptides.
Purified cysteine containing peptides were coupled covalently to Ultra Link iodoacetyl resin (Thermo Fisher Scientific). Peptides were dissolved in coupling buffer (50 mm Tris, 5 mm EDTA, pH 8.5) at a concentration of 1 mm and incubated for 2 h at room temperature (∼22°C) with washed and equilibrated UltraLink beads. After removing excess peptides, the unreacted iodoacetyl groups on the UltraLink beads were quenched with 1 mm cysteine for 2 h. The resin was washed three times and stored at 4°C.
Protein pulldown from brain lysate.
Incubation, washing and elution of the resin was executed in 400 µl micro-spin columns (Thermo Fisher Scientific). Whole-brain lysate from adult C57BL/6 mice (male) was incubated on the immobilized peptides for 1 h at 4°C. The resin was washed five times with lysate buffer. Bound proteins were eluted with Laemmli buffer containing 10% SDS at 90°C for 3 min. The eluate was applied to gel electrophoresis and after in-solution digestion analyzed via nanoLC-MS/MS (tandem mass spectrometry).
In-solution digestion.
Proteins were reduced in 50 mm dithiothreitol (DTT) for 10 min at 90°C and alkylated with 120 mm iodoacetamide for 20 min at room temperature in the dark. Protein precipitation was performed according to the study by Wessel and Flügge (1984) with chloroform/methanol. Precipitated proteins were dissolved in 0.5% sodium deoxycholate (SDC; Sigma-Aldrich) in 100 mm ammonium bicarbonate. Digests were performed with 0.1 µg of trypsin overnight at 37°C. SDC was removed by extraction with ethyl acetate (Masuda et al., 2008). Peptides were dried in a vacuum concentrator (Concentrator 5301, Eppendorf) to remove remaining ethyl acetate. Peptides were desalted using C18 stage tips (Rappsilber et al., 2003). Each stage tip was prepared with three disks of C18 Empore SPE Disks (3 m) in a 200 µl pipette tip. Peptides were eluted with 60% acetonitrile in 0.3% formic acid, dried in a vacuum concentrator, and stored at −20°C. Peptides were dissolved in 2% acetonitrile/0.1% formic acid before nanoLC-MS/MS analysis.
nanoLC-MS/MS analysis.
nanoLC-MS/MS analyses were performed on an Orbitrap Fusion (Thermo Fisher Scientific) equipped with an EASY-Spray Ion Source and coupled to an EASY-nLC 1000 (Thermo Fisher Scientific). Peptides were loaded on a trapping column [2 cm × 75 µm inner diameter (i.d.); PepMap C18, 3 µm particles, 100 Å pore size] and separated on an EASY-Spray column (50 cm × 75 µm i.d.; PepMap C18, 2 µm particles, 100 Å pore size) with a 120 min linear gradient from 3% to 40% acetonitrile and 0.1% formic acid.
Both MS and MS/MS scans were acquired in the Orbitrap analyzer with a resolution of 60,000 for MS scans, and 15,000 for MS/MS scans. Higher-energy collisional dissociation with 35% normalized collision energy was applied. A top speed data-dependent MS/MS method with a fixed cycle time of 3 s was used. Dynamic exclusion was applied with a repeat count of 1 and an exclusion duration of 120 s. Singly charged precursors were excluded from selection. Minimum signal threshold for precursor selection was set to 50,000. Predictive automatic gain control (AGC) was used with AGC a target value of 5e5 for MS scans and 5e4 for MS/MS scans. EASY-IC was used for internal calibration.
MS data analysis.
Raw MS data files were analyzed with MaxQuant version 1.6.2.2 (Cox and Mann, 2008). A database search was performed with Andromeda, which is integrated in the used version of MaxQuant. The search was performed against the UniProt Mouse database. Additionally, a database containing common contaminants was used. The search was performed with tryptic cleavage specificity with three allowed miscleavages.
Protein identification was under control of the false discovery rate (FDR; <1% FDR on protein and peptide level). In addition to MaxQuant default settings, the search was performed against the following variable modifications: protein N-terminal acetylation; glutamine [Gln to pyro-Glu formation (N-terminal Gln); and oxidation (Met). Carbamidomethyl (Cys) was set as fixed modification. For protein quantitation, the label-free quantification (LFQ) intensities were used. Proteins with less than two identified razor/unique peptides were dismissed.
Further data analysis was performed using R scripts developed in-house. For the discrimination of unspecific enriched proteins, median LFQ intensities were calculated. Missing LFQ intensities were imputed with values close to the baseline. Data imputation was performed with values from a standard normal distribution with a mean of the 5% quantile of the combined log10-transformed LFQ intensities and an SD of 0.1. For the identification of significantly enriched proteins, boxplot outliers were identified in intensity bins of at least 300 proteins. Log2-transformed protein ratios of samples from different peptide pulldowns with values outside a 1.5× (significance 1) or 3× (significance 2) interquartile range, respectively, were considered as significantly enriched.
Experimental design and statistical analysis.
All experiments have been performed as at least three independent biological replicates (unless otherwise stated). The number of experiments or cells recorded for analyses is presented in the Results referring to the appropriate experiments.
Quantification of GlyR protein–mAb4A signals for each expressed variant were normalized to pan-cadherin signals and adjusted to GlyRα1 wild type. Calculated data were compared using ANOVA with a probability of error of p < 0.05 considered to be significant.
Quantification of data obtained from electrophysiological recordings: estimated data were compared using the unpaired two-tailed t test (GraphPad Prism, GraphPad Software or SigmaPlot Software) with a probability of error of p < 0.05 was considered to be significant. For statistical analysis of data comparing glycine-induced inward currents or desensitization behavior of GlyRα1wt with mutant GlyRs affected at position P366 or within the sequence 364NPPPAPSKSP373, data for each mutant were compared with GlyRα1wt of the same experiment.
Binding affinities of various GlyRα peptides to syndapin werevalidated using the two-tailed t test with p < 0.05 considered to be significant.
For statistical analysis of endogenous syndapin I and GlyRα1 following infection into hippocampal neurons, syndapin I signals were normalized to GlyRα1 input signals and adjusted to wild type. Calculated data were compared using ANOVA with a probability of error of p < 0.05 considered to be significant.
Results
Description of the pedigree
A novel hyperekplexia mutation GlyRα1P366L (in short α1P366L) was identified in the coding region of GLRA1 from a patient with symptoms characteristic of this neurologic disorder. Pedigree analysis revealed that α1P366L showed a dominant mode of inheritance (Fig. 1A). Proline 366 is located in the TM3-4 loop. Sequence alignment with other members of the CLR family indicates a high degree of conservation of this residue (Fig. 1B). While the transmembrane domains as well as the extracellular ligand-binding domain of GlyRα1 have been crystallized, little structural information about the TM3-4 loop is available (Fig. 1C). To investigate the consequences of the α1P366L mutation at the molecular, cellular, and functional level, we generated a set of α1 point mutations. The generated GlyRα1 variants differing in side chain volume, size, and polarity (α1P366L, α1P366C, α1P366R, α1P366T, α1P366W, and α1P366A). The corresponding α1 TM3-4 loop sequence 364NPPPAPSKSP373 not only contains the PXXPXR/K PPII helix structural element, but, additionally, displays a striking similarity with the corresponding GlyRβ subunit region (435KPPPAKP441), which mediates the interaction with intracellular syndapin I. To investigate the role of this structural element as well as a possible syndapin I binding of GlyRα1,we generated specific GlyRα1 sequence 364NPPPAPSKSP373 mutants, including exhaustive proline-to-alanine and proline-to-leucine exchanges (α1P373A, α1P1-4A, α11-5A, α1P1-5L with P365 = P1, P366 = P2, P367 = P3, P369 = P4, and P373 = P5;Fig. 1D).
Localization of the human P366L GlyRα1 variant. A, Family tree of the affected P366L patient (II/1). The dominant mode of inheritance demonstrates the mother (I/2) as the carrier of the P366L mutation, while the father (I/1) was unaffected. B, Sequence alignment of different human (Hs, Homo sapiens) members of the Cys-loop receptor family (5-HTR3A serotonin receptor subunit 3A; GABAARα1 GABAA receptor subunit α1; nAChRA1, nAChR subunit α1) showing the conservation of proline residues within the TM3-4 loop (black background). C, Cryo-EM structure of a single Danio rerio GlyRα1 subunit (Du et al., 2015). The ICD consists of the short TM1-2 loop and the large TM3-4 loop (dashed line) including the patient mutation α1P366L (red circle). D, Linear schematic representation of one GlyRα1 subunit with N-terminal ligand-binding domain, 4 α-helical transmembrane domains (TM1–TM4), and the short extracellular C terminus. TM1–TM3 are connected via short loop structures (top part). The large intracellular loop between TM3 and TM4 harbors a proline-rich stretch, including P366 that was investigated in this study. A series of point mutations and combined mutations were generated (bold, underlined) to analyze the influence of prolines 1–5 within the sequence 364NPPPAPSKSP373 on GlyRα1 function.
Proline 366 has no impact on surface and whole-cell expression levels of GlyRα1 protein
Previous in vitro studies on hyperekplexia mutants have demonstrated that other GlyRα1 TM3-4 loop variants affect GlyR trafficking and protein stability (Villmann et al., 2009; Chung et al., 2010). GlyRα1 constructs carrying an amino acid substitution at position 366 were analyzed for whole-cell receptor expression (cells were permeabilized and stained with pan-α antibody mAb4A) and surface-expressed GlyRα1 (cells were unpermeabilized, live staining with α1-specific antibody mAb2b) in HEK293 cells. No obvious change in the GlyRα1 expression pattern was detected for any of the constructs tested (Fig. 2A,B). The surface expression of the same constructs was further quantified using surface biotinylation, which allows discrimination between whole-cell and surface GlyRα1 protein. Whole-cell expression of constructs α1P366L, α1P366C, α1P366A, α1P366R, and α1P366W were comparable to wild-type α1wt (n = 3) levels, indicating that side chain volume and polarity at this position have no influence on receptor trafficking and expression (Fig. 2C–E). The surface expression of all α1P366X constructs displayed no significant reduction compared with α1wt (n = 3; Fig. 2C,E). Together, these results confirm that the patient mutation α1P366L does not affect receptor biogenesis in vitro.
Expression of α1P366 variants in transfected HEK293 cells shows no differences in subcellular expression patterns. A, GlyR (red) was stained in permeabilized cells using the pan-GlyR antibody mAb4A. B, Live staining of α1 variants at the cellular surface using the α1-specific antibody mAb2b. White squares represent the region enlarged in magnification column. Scale bar, 50 µm. DAPI staining is shown in blue. C, Comparison of whole-cell (WC) and surface (SF) receptor expression observed from transfected HEK293 cells. Cadherin (135 kDa) was stained as a membrane marker control. All receptor constructs carrying an amino acid substitution at position P366 were detected by the GlyRα antibody mAb4A at 48 kDa. D, E, Quantification of surface (D) and whole-cell protein (E) of GlyRα1 constructs mutated at position α1P366 compared with α1wt. Note that no differences in protein expression of GlyRα1P366L (red bars) compared with α1wt (black bar) have been detected. Two-tailed t test; n.s., nonsignificant; dotted line marks the background signal of GFP-transfected cells. All values shown are the mean ± SEM.
GlyRα1P366L has minor effects on the physiological properties of α1-containing GlyRs
To determine the functional consequences of α1P366L, α1wt, and mutated variants at position 366 or at the other proline positions within the sequence 364NPPPAPSKSP373 (Fig. 1D) were transiently expressed in HEK293 cells as homomeric GlyRs and investigated using whole-cell patch-clamp electrophysiology. Since the mature GlyR isoform at the synaptic sites in mammalian spinal cord and brainstem is expressed as a combination of α1 and β subunits, we determined the influence of the α1P366L mutation in homomeric α1 as well as heteromeric α1β configurations (Fig. 3A). A reduction in Imax suggesting a decrease in glycine efficacy was observed for homomeric α1P366L and for heteromeric α1P366Lβ receptors compared with α1wtβ (homomeric α1P366L: Imax, 64 ± 9%; n = 21; t value = 2.277; df = 54; *p = 0.026; heteromeric α1P366Lβ: Imax, 61 ± 8%; n = 20; t value = 2.2229; df = 42; *p = 0.0312; Table 1).
Electrophysiological properties obtained from GlyRα1 mutants
Mutations of residue P366 lead to reduced maximal chloride currents and changes in the desensitization kinetics of the GlyRα1. A, Comparison of Imax on application of 1 mm glycine between α1wt (black) and α1P366L (red) in homomeric and heteromeric receptor configuration. B, Imax values of GlyRα1 constructs mutated at position P366 compared with α1wt (Table 1). C, Imax values of α1 constructs with different proline mutations within the sequence 364NPPPAPSKSP373. D, Representative traces of glycine-activated α1 constructs used for the calculation of Imax and the analysis of desensitizing current fractions. E, F, Comparison of fractions of desensitization currents obtained from GlyRα1 constructs mutated at position P366 (E) or within the 364NPPPAPSKSP373 motif (F). G, Dose–response relationship for homomeric (α1wt, black line; α1P366L, red dashed line) and heteromeric (α1wtβ, black dashed line; α1P366Lβ, red dotted line). Glycine concentrations ranged from 3 μm to 1 mm. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, n.s., nonsignificant. All values shown are the mean ± SEM.
With the exception of α1P366C (64 ± 14; n = 13; t value = 1.87; df = 46; p = 0.067), α1P366T (38 ± 8%; n = 12; t value = 3.214; df = 44; **p = 0.0024), α1P366R (32 ± 10%; n = 12; t value = 3.467; df = 44; **p = 0.0012), α1P366AW (46 ± 14%; n = 12; t value = 2.606; df = 44; *p = 0.012), and α1P366A (28 ± 4%; n = 17; t value = 4.590; df = 49; ****p < 0.0001) exhibited significantly reduced maximum inward currents in response to saturating concentrations (1 mm) of glycine (Fig. 3B,D, Table 1). This points toward a distinct influence of residue P366 on GlyRα1 glycine efficacy. According to surface expression analysis (Fig. 2), these observations were unlikely to result from reduced cell surface receptor expression. From all GlyR constructs with a modified PPII helix tested, the exchange of all proline residues within the sequence 364NPPPAPSKSP373 to leucines (α1P1-5L) and deletion of four prolines (α1ΔP1-4) led to significant reduction of glycine-gated currents (α1P1-5L: 51 ± 6% compared with α1wt; n = 26; F = 4.091; df = 39; *p = 0.016; α1ΔP1-4: 55 ± 13% compared with α1wt; n = 22; F = 4.105; df = 37; **p = 0.0044, respectively; Fig. 3C). Notably, replacing these prolines within the putative PPII structure by alanines had no effect on Imax (Fig. 3C).
Variations in the GlyRα1 TM3-4 loop sequences have previously been shown to influence whole-cell desensitization (Breitinger et al., 2002; Papke and Grosman, 2014; Langlhofer et al., 2015). For that reason, we next investigated the desensitization behavior of homomeric α1wt and α1 365–373 proline variants. Current traces (Fig. 3E) were analyzed to compare the differences in their desensitization kinetics. In the presence of 1 mm glycine, a fraction of 32 ± 3% of desensitizing inward currents was observed for α1wt (n = 25; Table 1). This value was taken as a reference and set to 100% to compare desensitization kinetics with mutant GlyRs. Mutants α1P366C (133 ± 26% desensitizing current fraction compared with α1wt; n = 9; t value = 2.040; df = 32; *p = 0.049), α1P366A (136 ± 12%; n = 17; t value = 2.697; df = 42; *p = 0.01) and α1P366L generated GlyRα1 currents that desensitized more rapidly (Fig. 3E, Table 1) compared with α1wt. For α1P366L, the fraction of desensitizing currents was averaged to 165 ± 18% (n = 15; t value = 3.804; df = 38; ***p = 0.0005) compared with α1wt, arguing for less chloride influx in the presence of extracellular glycine due to enhanced receptor desensitization. Mutating other prolines surrounding P366 (α1P373L, α1P1-4L, α1P1-4A, α1P1-5L, α1ΔP1-4) had no significant influence on channel kinetics (Fig. 3F).
To determine the glycine EC50, concentration–response experiments were conducted in transfected cells (Fig. 3G). Homomeric α1wt exhibited an EC50 of 94 ± 11 μm (n = 7). This value was not significantly different from that for α1P366L (106 ± 2 μm; n = 8). Similarly, α1P366Lβ (81 ± 17 μm; n = 8) showed glycine EC50 values comparable to those of α1wtβ (90 ± 4 μm; n = 9; Fig. 3G). These results indicate that the P366L mutation does have an impact on GlyRα1 efficacy, but not on potency.
Using an artificial synapse model, recordings of spontaneous glycinergic synaptic currents mediated by α1P366L GlyRs (n = 7) were compared with those mediated by α1wt GlyRs (n = 6). We observed no significant differences in synaptic current amplitudes, rise times, or decay times (two-tailed t test; amplitude: t = 0.5666, df = 9, p = 0.58; rise: t = 0.6520, df = 9, p = 0.53; decay: t = 2.021, df = 9, p = 0.074; Fig. 4A–C).
The P366L mutation does not alter mIPSCs. A–C, Properties of spontaneous mIPSCs mediated by the α1P366L mutation. A–C, Comparison of the mean amplitude (A), rise time (10–90%; (B), and decay time constants (C) for α1wt and the mutant α1P366L. Recordings were performed from artificial synapses between spinal cord neurons and transfected HEK293 cells expressing the indicated GlyRs.
The intracellular PPII helix has no impact on extracellular domain movements during ion channel opening
The observed lower net chloride flux for α1P366L compared with the α1wt could not be explained by reduced surface expression. Another possibility is that receptor gating dynamics have been altered by the P366L mutation. Glycine-induced dynamic changes in receptor configuration can be monitored using voltage-clamp fluorometry (VCF) techniques. By covalently attaching a fluorophore to the extracellular domain of the receptors, VCF enables the simultaneous recording of changes in current and fluorescence (Pless et al., 2007; Pless and Lynch, 2009). Following application of the ligand glycine, in addition to the observed chloride inward current a shift in fluorescence is observed, pointing to changes in molecule mobility. Two residues of GlyRα1 in the extracellular domain (N203 and R271; Fig. 5A) exhibit large changes in fluorescence in response to glycine application and thus have been determined as highly flexible in the process of receptor gating (Pless et al., 2007; Pless and Lynch, 2009). We constructed double mutants GlyRα1N203C/P366L and α1R271C/P366L in a vectorsuitable for the oocyte expression system. The application of saturating concentrations of glycine induced a chloride ioninflux and changes in fluorescence (Imax; α1N203C: 3.6 ± 0.7 µA; n = 14; α1N203C/P366L: 2.1 ± 0.6 µA; n = 14; two-tailed unpairedt test, t = 1.725, df = 27, p = 0.09; α1R271C: 1.3 ± 0.5 µA; n = 11; α1R271C/P366L: 1.8 ± 0.5 µA; n = 22; t = 0.5876; df = 33, p = 0.56; changes in fluorescence, ΔF; α1N203C: 53 ± 4%; n = 8; α1N203C/P366L: 43 ± 8%; n = 13; t = 0.912; df = 19; p = 0.37; α1R271C: 13 ± 2%; n = 8; α1R271C/P366L: 13 ± 1%; n = 14; t = 0.3674; df = 20; p = 0.71; Fig. 5B–E). There were no obvious differences between the α1N203C and α1N203C/P366L or α1R271C and α1R271C/P366L, arguing that the integrity of the intracellular PPII helix does not affect extracellular conformational changes that can be probed through residues α1N203C and α1R271C.
The intracellular mutation α1P366L of the GlyR has not impact on changes of receptor conformation at the extracellular site. A, Single GlyRα1 subunit. Marked are residues N203, R271, and P366 are highlighted. B, Maximal current amplitudes of α1N203C (black bar) and α1N203C/P366L (red bar) and changes in fluorescence α1wt/α1N203C (gray bar) and α1P366L/α1N203 (red/gray striped; right graph). C, Same as in B but comparison of α1R271C and α1R271C/P366C. D, Representative traces of currents (black traces) and changes in fluorescence (red traces) induced by application of 30 mm glycine to oocytes expressing α1N203C or α1N203C/P366L. E, Same as in D but traces recorded from α1R271C and α1R271C/P366C are illustrated. All data represent the mean ± SEM. n.s., nonsignificant.
GlyRα1P366L exhibited altered single-channel properties
To further investigate functional changes of the α1P366L mutation, we analyzed currents mediated by single-receptor molecules focusing on heteromeric α1P366Lβ receptors. Single-channel recordings were used to determine the presence of spontaneous activity in the absence of glycine, the incidence of active periods of different amplitude levels, the I–V relationship, and conductance.
Our first experiments were aimed at determining whether the α1P366L mutation gave rise to spontaneous activity in heteromeric GlyRs. In the absence of glycine, outside-out patches expressing the GlyRα1P366Lβ in contrast to α1wtβ opened sporadically over a recording period of 5–10 min. Single-channel opening times were brief and lasting ∼2–3 ms (n = 3; Fig. 6A,B). This level of activity was more frequent than α1wt receptors under the same recording conditions (Fig. 6A), but was less frequent than another spontaneously active α1 mutation that gives rise to hyperekplexia, α1Q226E (paired Holm–Sidak test; α1wtβ compared with α1P366Lβ: t = 3.144, p = 0.02; α1P366Lβ compared with α1Q226Eβ: t = 5.363, p = 0.003; α1wtβ compared with α1Q226Eβ: t = 8.507, p < 0.001; Fig. 6C; Bode and Lynch, 2013). The rank order of opening probability was (n = 3) ∼0, <0.01, and < 0.03 for α1βwt, α1P366Lβ GlyR, and α1Q226Eβ GlyR, respectively (Scott et al., 2015). The more frequent spontaneous activity of α1P366Lβ is probably not due to differences in numbers of channels in each patch as the surface expression of α1P366Lβ was not changed compared with α1βwt receptors. Activated α1β GlyRs form discrete clusters of activity separated by long nonconducting periods that correspond to desensitization (Fig. 7A). Two main amplitude levels were apparent (Fig. 7A, top). The most frequent amplitude was 3.3 ± 0.1 pA (−70 mV, n = 5), which constituted >95% of the activity. A smaller amplitude (2.4 ± 0.1 pA) was also observed, but this was rare and usually occurred as a transition within a cluster (Fig. 7A, bottom). The I–V plot for α1wtβ receptors was near linear, producing a reversal potential of 5 mV (Fig. 7B). This enabled us to estimate the conductance of the main level according to Equation 2, which was 41.4 pS at −70 mV.
Spontaneous opening events in the absence of extracellular glycine. A, B, Single-channel recordings of heteromeric α1wtβ GlyRs (A) compared with α1P366Lβ (B) to investigate the spontaneous opening activity in the absence of glycine. Four independent traces for each condition are shown (black, α1wtβ; red, α1P366Lβ). C, Quantification of spontaneous open probabilities. The mutation α1Q226E was also analyzed since spontaneous openings for this mutant have been shown before (Bode et al., 2013). Statistical significance refers to ***p < 0.001, **p < 0.01, and *p < 0.05.
Proline 366 changes the unitary conductance. A, Single-channel activation clusters in a patch expressing at least three α1 wtβ GlyRs in the presence of 100 μm glycine. The most common (>95%) amplitude level was 3.3 pA (top). There was also a smaller amplitude of “2.4 pA”. Traces are from the same patch, but different sections of the recording. B, Sample recordings of single-channel activity of α1 wtβ at the indicated voltages and the accompanying group I–V (n = 5). The I–V was used to calculate a conductance of 41.4 pS. C, Single-channel activations of α1P366Lβ in the presence of 100 μm glycine. This channel opened to the following three levels: 3.0 pA (top trace), 2.1 pA (middle trace), and 1.0 pA (bottom trace). Activations to the largest level represented 75% of the overall activity. The top two traces are from the same patch, and the lower is from a different patch. Note, the middle and lower traces in A and C have a different time scale to show subconductance levels at higher temporal solution that were less frequent. D, Sample recordings of single-channel activity of mutant α1P366Lβ at the indicated voltages and the accompanying group I–V (n = 6). The I–V was used to calculate the corresponding conductances of 37.6, 26.3, and 12.5 pS. E, Single-channel conductance of α1wtβ and α1P366Lβ. Significance values, *p = 0.0102.
Single-channel activations of α1P366Lβ GlyRs also occurred in discrete clusters when exposed to 100 μm glycine (Fig. 7C, top). Three main amplitude levels were apparent, the largest one being 3.0 ± 0.6 pA (at −70 mV, n = 6). This amplitude was the most frequent, constituting 75 ± 4% of the activity. The other two levels were less frequent, being 17 ± 3% (2.1 ± 0.1 pA, n = 5) and 8 ± 2% (1.0 ± 0.1 pA, n = 5; Fig. 7C, bottom) of the overall activity. The I–V for the mutant receptor was also near linear over the voltage range tested and reversed at +5 mV (Fig. 7D). The corresponding conductance levels for the mutant receptors at −70 mV were 37.6, 26.3, and 12.5 pS, respectively. A direct comparison of single-channel conductance shows a reduction to ∼80% for α1P366Lβ receptors compared with the α1wtβ control (two-tailed t test: t = −3.238, df = 9, p = 0.0102; Fig. 7E).
GlyRα1P366L harbors a syndapin I binding site which encompasses P366
The spontaneous opening events of α1P366Lβ identified in single-channel recordings were less frequent than in α1Q226Eβ (Scott et al., 2015), suggesting that tonic activity alone might not explain the hyperekplexia phenotype observed. There might be other structural impacts of GlyRα1P366L in addition to the specified functional changes. Residues within the TM3-4 loop of GlyRs are involved in interactions with intracellular binding partners or are responsible for post-translational modifications of the GlyR (Langlhofer and Villmann, 2016). GlyRβ residues 398FSIVGSLPRDFELS411 mediate direct binding of gephyrin (Maric et al., 2014, 2017). Furthermore, the GlyRβ sequence (433NKKPPPAKPVIPTAA447) harbors a syndapin I binding site (Del Pino et al., 2014). Strikingly, the syndapin binding motif within the corresponding GlyRα1 region 364NPPPAPSKSP373 (Fig. 8A) contains the hyperekplexia mutation α1P366L studied here (Fig. 1B). We therefore investigated whether GlyRα1 also interacts with syndapin I. Syndapin I recognizes class I and II poly-proline-rich motifs (Teyra et al., 2017; Goretzki et al., 2018). The GlyRβ sequence (433NKKPPPAKPVIPTAA447) as well as the corresponding GlyRα1 region, with and without the hyperekplexia mutation studied here, were probed together with other reported syndapin I binding proteins (other known SH3 binding peptides containing class I and class II PxxP motifs) in microarray format for direct binding to syndapin I. The peptide NNNNTTNPPPAPSKS (GlyRα1 Mus musculus) was clearly bound to GST-syndapin I, as reported by syndapin I-HRP-coupled secondary antibody chemiluminescence (Fig. 8B). Introducing the α1P366L or α1P366A mutations into the samepeptide significantly reduced syndapin I binding (two-tailed unpaired t test: α1wt peptide NNNNTTNPPPAPSKS compared with α1P366L NNNNTTNPLPAPSKS: t = 107, df = 4, ****p < 0.0001; α1 wt peptide compared with α1P366A NNNNTTNPAPAPSKS: t = 167, df = 4, ****p < 0.0001; Fig. 8B).
Identification of short GlyRα1-derived peptides as syndapin I binding probes. A, B, Sequence alignment of known syndapin I binding partners [GlyRβ, SOS, and SHANK2; SOS1, son of sevenless homolog 1; SHANK2, SH3 and multiple ankyrin repeat domains protein 2-like (or PROSAP1); A] with corresponding binding intensities determined with the GST-syndapin I array (B). Syndapin I recognizes class I and II poly-proline-rich motifs (Teyra et al., 2017; Goretzki et al., 2018). Residues that satisfy the requirements of these recognition motifs are colored blue (charged basic), orange (proline), and red (cis-proline). The corresponding GlyRα1 sequences of the wild-type α1wt and the mutant α1P366L are shown (lower two lines). B, Peptide array-based alanine and leucine scan of the syndapin I binding site within the GlyRα1 subunit. Other known SH3 binding peptides containing class I and class II PxxP motifs were added as control peptides for the array (data not shown). Red residues refer to position 366 of the mature peptide. Syndapin I binding to peptides was detected by chemiluminescence of syndapin I-conjugated HRP. Shown are the relative averaged intensities of the detected syndapin spots of three peptide sets together with their SDs (error bars). Notably, a P366L or P366A mutation in the peptide 358NNNNTTNPPPAPSKS373 almost abolished binding to syndapin I (****p < 0.0001). C, Quantitative mass spectrometric analyses of pulldowns with the corresponding peptides derived from α1wt (362TTNPPPAPSKSPC374), α1P366L (362TTNPLPAPSKSPC374), and GlyRβ (433NKKPPPAKPVIPC445) subunits. Scatter plots show log2-transformed protein ratios (left graph, enrichment of pulldown using the GlyR β peptide compared with pulldown using the α1wt peptide; right graph, enrichment of pulldown using GlyRα1P366L peptide compared with α1wt peptide). Proteins that are similarly abundant in both eluates are close to zero log2(1) = 0 versus summed log10 LFQ protein intensities (highly abundant proteins show up at larger values). Significantly enriched proteins are classified as highly significant (red) and significant (green); proteins not enriched are shown in blue. The sizes of the dots correlate with the number of identified peptides.
The relative intensities of the peptide microarray bindingsignals validate the previously identified GlyRβ fragment as a high-affinity syndapin I interactor and at the same time indicate robust binding of the corresponding GlyRα1 region (Fig. 8A,B). Next, we wanted to further validate the in vitro discovered GlyRα1/syndapin I in ex vivo context. In the cell, binding of the native syndapin I may be further modulated by post-translational modifications and the presence of additional allosteric or directly competing endogenous ligands. Therefore, we used label-free quantitative MS/MS to explore the binding of syndapin I to the mapped GlyRβ and α1 regions as well as the P366L mutated α1 region in the context of their complete interactome. In label-free quantitative MS/MS, protein intensities (summed intensities of identified tryptic peptides of the corresponding protein) are compared among different LC-MS/MS runs.
First, we incubated whole-brain lysates from adult C57BL/6 mice (male) with resin-bound peptides corresponding to either α1wt or the α1P366L mutation or the corresponding GlyRβ region. Protein interactors from the lysate were eluted, digested, and identified through nanoLC-MS/MS (Fig. 8C). Mass spectrometric analysis revealed a clear enrichment of syndapin I in the interactome of GlyR subunit fragments (α1wt 362TTNPPPAPSKSPC374, α1P366L 362TTNPLPAPSKSPC374, and GlyRβ 433NKKPPPAKPVIPC445; Fig. 8C). Beyond syndapin I, the analysis also revealed enhanced binding of gephyrin, and endophilin (Fig. 8C), and the known interactor syndapin II to the GlyRβ fragment over the α1 fragment. We also used the label-free quantitative MS/MS to analyze and compare the pulldowns of α1wt versus α1P366L. Both fragments enriched syndapin I, but no significant change in the extent of enrichment could be resolved on mutation (Fig. 8C).
To further confirm the protein-protein interaction between the GlyRα1 subunit and syndapin I, we performed GST pull-down assays (Fig. 9A). GlyRα1 was successfully pulled down by GST-sdpISH3, but not GST alone. Moreover, transfected HEK293 cells with α1 or α1P366L together with either syndapin I-eGFP (fusion protein of syndapin I and eGFP; 80 kDa) or syndapin I-eGFP lacking the SH3 domain (sdpIΔSH3) were used for immunoprecipitation. The input images show the slight reduction in molecular weight of syndapin I lacking the SH3 domain compared with full-length syndapin I (Fig. 9B, top). The monoclonal antibody mAb2B successfully precipitated the GlyR [Fig. 9B, bottom (with lower image representing the precipitation control)]. Syndapin I-eGFP was coprecipitated together with the heteromeric α1β complexes and α1P366Lβ but also with α1 and the mutated α1 variant α1P366L when expressed alone. Additionally, syndapin I coprecipitation was almost abolished when syndapin I lacked the SH3 domain (Fig. 9B). Together, our data reveal binding of syndapin I to the mutated region within GlyRα1, indicating an interaction that so far has been attributed only to the GlyRβ subunit. The position of the mutation in the syndapin I binding motif suggests interference with syndapin binding. Direct interference of α1P366L with syndapin binding can be resolved in vitro using the recombinant protein but not based on lysates containing the endogenous protein.
The GlyRα1 subunit interacts with syndapin I. A, Pulldown of the GlyRα1 wild-type subunit using GST-syndapin I SH3 domain (GST-sdpISH3). Lysates from HEK293 cells transfected with GlyRα1wt were incubated with either GST alone or GST-tagged syndapin I SH3 domain (GST-sdpISH3). Aliquots of starting material (Input), unbound material (unbound), and precipitated proteins (pulldown) were analyzed for coprecipitated GlyRα1wt (m-α-GlyRα) or GST (rb-α-GST). Note that GlyRα1wt was successfully pulled down by GST-sdpISH3, but not by GST alone (marked by an asterisk, *). B, Immunoprecipitation of syndapin I with and without the SH3 domain (sdpISH3 and sdpIΔSH3) together with GlyRα variants (α1wt and α1P366L) with and without the β subunit. Input from transfected HEK293 cells are shown. Top, Input samples were stained against syndapin I, GlyR (mAb4a = pan-α), or GAPDH (as housekeeping control protein). UT refers to untransfected cells. Bottom, Immunoprecipitation was performed using the GlyRα1-specific antibody mAb2b. Precipitated syndapin I (sdpI) is demonstrated in the top panel. Both GlyRα1wt and GlyRα1P366L are able to precipitate sdpI SH3 and sdpIΔSH3, however, with less intensity. The bottom panel shows the precipitation control of the GlyR (stained by mAb4a = pan-α).
Interaction between syndapin I and the α1P366L mutant redistributes syndapin I to the cell soma
Our proteomic data demonstrate that syndapin I interacts with the SH3 consensus sequence of the GlyRα1 TM3-4 loop and that proline at position 366 contributes to this interaction. We speculated that mutant binding of syndapin I might affect normal syndapin I function in receptor trafficking and localization. Therefore, we explored the neuronal distribution of syndapin I in the presence of either α1 or α1P366L. Specifically, we stained endogenous neuronal syndapin I in virally infected primaryhippocampal cultures with constructs hGlyRα1IRESGFP or hGlyRα1P366L-IRESGFP, driving the low expression of either α1 or α1P366L (Fig. 10A). Cells were stained after infection for surface GlyRα1 and whole-cell syndapin I at DIV14 (n = 3; Fig. 10B–D). Uninfected hippocampal neurons lacked α1 expression (Fig. 10D). Strikingly, we observed a strong increase in syndapin I signal intensity in the soma of cells expressing the mutant α1P366L compared with α1wt and uninfected cells (four to seven z-stacks were used for analysis, with a two-tailed t test: t = 2.327, df = 9, *p = 0.045; Fig. 10E) that was accompanied by a decrease in the syndapin I signal intensity observed in neurites of α1P366L infected cells (t = 2.045, df = 9, p = 0.07). This result points out that α1P366L specifically alters endogenous syndapin I distribution by accumulating syndapin in the cell soma.
Expression of α1P366L in infected primary hippocampal neurons changes endogenous syndapin I distribution. A, Schematic overview of the lentiviral vector construct expressing GlyRα1 or GlyRα1P366L. The lentiviral vector is based on the self-inactivating (SIN) vector backbone FUGW (Lois et al., 2002). The human ubiquitin C (hUbiC) promoter drives expression of human GlyRα1 or the mutated variant GlyRα1P366L. An internal ribosomal entry site (IRES) is used for bicistronic expression of GFP. LTR, Long-terminal repeat; WPRE, Woodchuck Hepatitis Virus (WHP) Post-transcriptional Regulatory Element. B, C, Cell surface expressions of α1wt (B) and α1P366L (C) were determined in live stainings with the α1-specific antibody mAb2b recognizing a native epitope in the far N-terminal part of GlyRα1 (red). After permeabilization of cells, endogenous syndapin I was visualized with the rabbit anti-syndapin I antibody (green). Note that expression of the mutant α1P366L redistributed endogenous syndapin I to the cell soma. D, Expression of endogenous α1 GlyR and syndapin I in uninfected hippocampal primary neurons. E, Quantification of syndapin I signal intensity in hGlyRα1IRESGFP and hGlyRα1P366L-IRESGFP-positive cells revealed significant difference in somatic expression. Scale bar, 10 µm. DAPI (blue) staining marks DNA (nucleus). *p < 0.05.
Discussion
In this study, we identified a pattern of glycinergic dysfunction caused by the novel dominant startle disease mutation P366L, which is localized in the intracellular TM3-4 loop of the GlyRα1 subunit. Using this mutant, we uncovered a contribution of the α1 subunit to binding of the intracellular protein syndapin I, a function that has previously been previously assigned to the GlyRβ subunit only. Members of the syndapin family are SH3 domain-containing proteins involved in membrane protein trafficking, regulation, and interaction with actin cytoskeletal effectors (Qualmann and Kelly, 2000; Kessels and Qualmann, 2004; Dharmalingam et al., 2009; Koch et al., 2011; Schwintzer et al., 2011). The patient mutation α1P366L is localized in a poly-proline-rich region forming a PPII helix in the GlyRα1 intracellular TM3-4 loop able to interact with the SH3-domain of syndapin I.
Most GlyRα1 hyperekplexia mutations (α1R316X, α1A384P, α1D388A, and α1R392H) in the TM3-4 loop exhibit a compound heterozygous mode of inheritance. Mutant α1E375X represents a recessive mutation, and α1G342S is the only dominant inherited hyperekplexia mutation but with incomplete penetrance (Bode and Lynch, 2014). Recessive mutations have been shown to interfere with surface membrane integration of the GlyRs, whereas dominant mutations affect ion channel functional properties, including ion conductance, desensitization, and ligand binding (Villmann et al., 2009). Previous studies demonstrated highly reduced surface expression and complete loss of glycine-induced currents for most TM3-4 loop mutations in a homomeric receptor configuration (Chung et al., 2010; Bode et al., 2013; Schaefer et al., 2015).
In contrast, α1P366L expression levels were indistinguishable from α1wt protein. Recombinant expression of α1P366L as well as exchanges of proline to alanine, arginine, and threonine demonstrated that position 366 substantially affected glycine-gated chloride currents. These data suggest that side chain volume and polarity at residue 366 per se are not the main limiting factor affecting GlyR physiology. Moreover, the ∼25% reduction in maximal current responses detected in α1P366L is not a result of decreased membrane incorporation or ER retention as demonstrated for other intracellular loop mutations (Villmann et al., 2009; Bode et al., 2013). Instead, global conformational changes induced by the proline-to-leucine exchange are more likely to account for reduced inward currents.
Synaptic GlyRs are expressed as heteromeric protein complexes composed of α1 and β subunits representing the major adult isoform in the human spinal cord and brainstem nuclei (Patrizio et al., 2017). For mutations α1E375X, α1D388A, and α1R392H, a rescue of protein function by coexpression with the GlyRβ subunit has been described (Chung et al., 2010; Bode et al., 2013). Here, the presence of GlyRβ had no supportive effect on the glycine potency when coexpressed with the α1P366L. Furthermore, the glycine efficacy of the mutant GlyRα1P366L was unaltered in the presence and absence of the coexpressed β subunit.
The proline 366-to-leucine mutation seems not to influence the extracellular ligand binding or associated structural transitions. Our findings from voltage-clamp fluorometry recordings support this conclusion. Labeling of extracellular residues N203 or R271 (Pless and Lynch, 2009) showed no further changes in fluorescence when α1P366L was present.
Because of the location of proline 366 in a PPII helix, which may change its helical structure when proline is mutated to leucine, we speculated that α1P366L alters intracellular conformational rearrangements of the GlyR that occur in the processes of channel opening, desensitizing, and closing. Several in vitro studies demonstrated that small variations of the GlyRα1, GABAAR, or 5-HT3AR TM3-4 loop sequences affect desensitization kinetics (Hu et al., 2006; O'Toole and Jenkins, 2011; Langlhofer et al., 2015).
The analysis of the channel kinetics revealed that proline 366 mutations to leucine, alanine, and cysteine produced GlyRs with accelerated whole-cell desensitization. An enhanced fraction of desensitizing currents in other intracellular mutations α1P250T and α1A384P has previously been reported to account for disturbed glycinergic neurotransmission and human startle disease (Breitinger et al., 2001; Wang et al., 2018). In the case of P250T, the loss of an angular polypeptide structure was thought to destabilize the open channel conformation due to the close proximity to the pore lining TM2 (Saul et al., 1999). The pathophysiological importance of CLR desensitization is emphasized by several studies that connected changes in receptor kinetics to pathologic processes observed in different neurologic disorders (e.g., changes in nAChR desensitization rates associated with the development of frontal lobe epilepsy and congenital myasthenic syndrome; Sine et al., 2002). However, the observed increase in whole-cell desensitization of α1P366L is small compared with the channel-destabilizing hyperekplexia mutant α1P250T. The so far determined minor effects of the mutant α1P366L on receptor desensitization, the unchanged glycine sensitivity, and the unaltered surface expression also explain the lack of changes in the IPSCs obtained from artificial synapse recordings.
Our findings of reduced peak current in heteromeric α1P366Lβ receptors in whole-cell recordings were consistent with reduced conductance at the single-channel level. A significant decrease in single-channel conductance induced by α1P366L was observed similar to other mutations associated with hereditary hyperekplexia (Saul et al., 1999; Bode and Lynch, 2013). In respect of the spinal motor reflex arcs, it is accepted that reduction in glycinergic magnitudes disinhibit motoneurons and thereby enhance their excitability, leading to the increased frequency of muscle contractions.
Furthermore, we found α1P366Lβ receptors to open sporadically in the absence of extracellular glycine. The sporadic opening of α1P366Lβ is not due to a reduced number of receptors in the patch as surface expression of the mutant α1P366L was similar to that of α1wt receptors. Tonic GlyR activity has also been identified in other autosomal-dominant inherited hyperekplexia mutations α1Y128C, α1Q226E, α1V280M, and α1R414H (Chung et al., 2010; Bode and Lynch, 2013).
An increase in spontaneous activity can be easily explained by a concomitant decrease in glycine EC50, which is, however, not observed for α1P366Lβ. This is in line with α1Q226E not affecting the glycine EC50, whereas α1V280M significantly decreased and α1R414H slightly increased the EC50 for the agonist glycine (Bode et al., 2013). Studies on α1Q226E and α1V280M provided evidence using molecular modeling for channel widening at the cytoplasmic exit caused by these hyperekplexia mutations. Considering the distance of P366 from the transmembrane domain, it is more likely that α1P366L causes intracellular structural changes most likely of the PPII helix that influence spontaneous channel opening. Spontaneous opening events of α1P366Lβ were, however, less frequent than in α1Q226E, suggesting that tonic activity alone might not explain the hyperekplexia phenotype observed (Scott et al., 2015). Furthermore, the α1P366L mutation resulted in a net reduction of channel conductance. This was due to a smaller main conductance level in the mutant and the identification of low-conductance states that were not observed in the α1wt.
From our data, we cannot rule out that α1P366L disrupts the homomeric presynaptic GlyR function that regulates inhibitory transmitter release and thereby generates imbalances of inhibition and excitation in adjacent motoneurons (Turecek and Trussell, 2001; Xiong et al., 2014). Moreover, mutant α1P366L might undergo protein–protein interactions with presynaptic and extrasynaptic GABAARs impairing additionally GABAergic inhibition and, thus, contributing to the disease pathology of human startle disease (Zou et al., 2019).
The TM3-4 loops of both, GlyRα1 and β subunits contain proline-rich subdomains that have previously been demonstrated to create consensus sequences for the binding of SH3 domains (Cascio et al., 2001; Breitinger et al., 2004), and other PPII helix-binding partners. Recently, it has been pointed out that the dominant mode of ligand recognition of the SH3 domains involves interactions between poly-prolines and the PXXP-binding site, but noncanonical or multiple specificities are also common (Teyra et al., 2017). We found that syndapin I was bound to the GlyRα1 peptide 362TTNPPPAPSKSPC374, including the P366 locus as well as to the GlyRβ peptide 433NKKPPPAKPVIPC445. Our analysis clearly demonstrated the accumulation of syndapin I and endophilin in probes containing one of those peptides. These results are consistent with a study showing that syndapin I and endophilin I bind to overlapping proline-rich regions of dynamin I (Anggono and Robinson, 2007). The interaction between GlyRα1 and syndapin I was significantly diminished by the mutation of proline 366 to alanine or leucine at least in the peptide array approach. Similarly, the interaction of TRPV4 channels and syndapin three mediated by a proline-rich region with consecutive “PPP” residues was significantly weakened by an amino acid substitution of the second proline to either an leucine or an alanine (Goretzki et al., 2018).
The C-terminal SH3 domain of syndapin I is crucial for endocytosis, cytoskeletal remodeling, and postsynapse formation (Qualmann and Kelly, 2000; Dharmalingam et al., 2009; Schneider et al., 2014; Luo et al., 2016). Syndapin I is present at excitatory and inhibitory synapses. Syndapin I has been demonstrated to regulate AMPA receptor trafficking, internalization, degradation, and recycling processes, thus regulating synaptic strength of excitatory synapses (Quan and Robinson, 2013; Widagdo et al., 2016). Interestingly, syndapin I knock-out mice display generalized seizures, defects in excitatory and inhibitory synaptic transmission, impairments in vesicle formation and impairment of dynamin recruitment (Koch et al., 2011). Binding of syndapin I to the GlyRβ subunit has been shown with decreased GlyR cluster sizes and densities on syndapin I knockout (Del Pino et al., 2014). We observed a strong increase in syndapin I signal intensity in the soma of cells and a slight decrease in the neurites expressing the mutant α1P366L compared with α1wt. Our results suggest that the interaction between the mutated poly-proline subdomain in α1P366L and the SH3 domain of syndapin I changes the protein distribution in hippocampal neurons. Therefore, syndapin I might potentially act as a regulator for GlyR trafficking to neurite membranes. However, adistinct role of syndapin I in GlyR dynamics needs further investigation in an appropriate in vivo model. Together, we identified the GlyRα1subunit contributing to interactions with intracellular binding partners important for receptor trafficking, a function thus far only ascribed to the GlyRβ subunit.
We demonstrate that the patient GlyR mutation α1P366L exhibits multiple functional deficits of the inhibitory GlyRs, including decreased single-channel conductance, accelerated receptor desensitization kinetics, spontaneous channel opening, and reduced interaction with F-BAR protein syndapin I, hence providing experimental evidence explaining the neuromotor phenotype of the startle disease patient.
Footnotes
This work was supported by Deutsche Forschungsgemeinschaft Grants VI586 (C.V.), QU116/5, and DFG CRC166 (B.Q.). J.W.L. is supported by National Health and Medical Research Council Grant 1058542. G.L. was supported by the Graduate School of Life Sciences Würzburg as well as the Bayerische Forschungsstiftung. We thank Cord-Michael Becker for the assignment of blood samples from patients for research purposes. We also thank Nadine Vornberger for excellent technical assistance.
The authors declare no competing financial interests.
- Correspondence should be addressed to Carmen Villmann at Villmann_C{at}ukw.de
