Figure 1.
Gephyrin is palmitoylated in vivo.
(a) Mouse brain lysates were fractionated via ultracentrifugation and immunoblotted with gephyrin and PSD-95–specific antibodies. (b) Representative immunoblots of enriched synaptosomes after extraction with Triton X-100 or Saponin, overlayed with a discontinuous sucrose gradient and ultracentrifugation. Ten fractions were taken from top to bottom. (c) Gephyrin palmitoylation detected by metabolic labeling with 17-ODYA followed by click chemistry and affinity purification. (d) Basal palmitoylation of gephyrin in primary hippocampal neurons demonstrated by the ABE assay. Replacing hydroxylamine (HA) with a Tris-buffer (−HA) served as an internal negative control to proof efficient blockade of thiols. Palmitoylation of gephyrin was blocked with 50 µM 2-BP for 14 h. (e) Gephyrin palmitoylation status in the cytosolic (S2) compared to synaptosomal (P2) fraction analyzed by the ABE assay. Lanes of the top panel have been spliced from the same membrane. Volumes of the S2/P2 fractions have been adjusted to similar amounts of gephyrin (right panel). Experiment was performed with two individual synaptosomal preparations with essentially the same results.
Figure 2.
Palmitoylation of gephyrin is essential for clustering at GABAergic synapses.
(a) Immunostaining of gephyrin (red) and VGAT (green) in primary hippocampal neurons. Insets are taken from 20 µm dendritic segments. Arrows point to nonsynaptic gephyrin clusters upon incubation with 2-BP. Scale bar, 20 µm. (b–d) Histograms showing quantitative comparisons of (b) gephyrin colocalization with VGAT, (c) gephyrin cluster size, and (d) VGAT puncta area in control and 2-BP–incubated cultures. (e) Quantification of percentage of palmitoylated proteins in mouse brains. Immunoblots with increasing volumes of ABE-assay–processed protein lysates were used to create a band-intensity standard curve for ABE-assay eluates of gephyrin and PSD-95. The experiment has been performed with three brains. (f′) Streptavidin-HRP immunoblot served as a specificity and accuracy control of the assay. L, lysate; +HA, samples processed with hydroxylamine; +HAB, +HA supernatant after Neutravidin beads sedimentation to prove efficient affinity-purification of biotinoylated proteins; −HA, internal control of the ABE assay without hydroxylamine.
Figure 3.
Palmitoylation on Cys212 and Cys284 is essential for gephyrin localization and clustering in neurons.
(a) Scheme of gephyrin domain architecture with surface-exposed and Palm-CSS 3.0–predicted cysteine residues. (b) Immunoprecipitation of ABE-assay–processed gephyrin cysteine-to-serine mutants expressed in HEK293 cells. Streptavidin-HRP shows palmitoylation status of individual variants, whereas GFP-immunoblot was used as the loading control. Experiment was repeated three times with similar results. (c) SDS-PAGE of ABE-assay–processed and Ni-NTA affinity-purified gephyrin and LC-MS/MS analysis of the protein bands. (d) Quantitative comparison of cluster size of GFP-tagged WT and mutant gephyrin variants in primary hippocampal cultures. (e) Total intensity of WT gephyrin or 212,284Geph clusters (f) Representative dendrites of gephyrin-GFP and 212,284Geph. Scale bar, 5 µm. All quantifications are means ± SEM (***p<0.001 using Student's t test; NS, not significant). At least three independent cultures were used per experiment.
Figure 4.
Morphological analysis of palmitoylation-deficient gephyrin variants in primary hippocampal neurons.
(a) Representative examples are illustrated for WT gephyrin-GFP and 212,284Geph. Scale bar, 20 µm. (a′) Insets demonstrate postsynaptic clustering of gephyrin by colocalization with VGAT (red). 212,284Geph clusters show postsynaptic (arrowheads) and nonsynaptic (arrows) localization. (b, c) Quantitative analysis of (b) gephyrin cluster numbers and (c) VGAT puncta in gephyrin-GFP and 212,284Geph-transfected primary hippocampal neurons. (d) Colocalization of gephyrin-GFP and 212,284Geph clusters with GABAAR α2 immunofluorescence. Arrows show 212,284Geph clusters that are not localized at GABAergic synapses. Scale bar, 5 µm. (e) Quantitative analysis of cluster colocalized with α2-containing GABAARs. (f) Representative immunoblot of subcellular fractionation of 14 DIV cultured hippocampal neurons after incubation with 50 µM 2-BP or the solvent. (g–h) Quantitative line-scan analyses of fluorescence intensities ranging from individual clusters towards the cytosol as indicated by white arrows. Scale bar, 5 µm. (h′) Histogram shows quantification of cluster/cytoplasm intensity rations in gephyrin-GFP compared to 212,284Geph-transfected dendrites. All data are means ± SEM (***p<0.001 using Student's t test; NS, not significant). At least three independent cultures were used for the quantifications.
Figure 5.
Gephyrin is palmitoylated by DHHC-12.
(a) Individual HA-tagged DHHC proteins or GST-HA as control were co-expressed with gephyrin-GFP for 24 h in HEK293 cells and analyzed with the ABE assay. Omitting hydroxylamine demonstrated specify of the assay. (b) Gephyrin-GFP was co-expressed with individual HA-tagged DHHC enzymes or GST-HA as control in primary hippocampal neurons, and gephyrin cluster size was quantified by confocal laser scanning microscopy. In the histogram, DHHC enzymes are labeled according to [16]. (c) Representative images show phenotypical differences of gephyrin-GFP clusters in DHHC-12–expressing neurons compared to GST-expressing controls. Scale bar, 10 µm. (d) Gephyrin-GFP cluster density after 24 h expression of DHHC-12 compared to control neurons. (e–g) Quantification of gephyrin-GFP cluster size and density and representative image upon expression of dominant-negative DHHS-12–HA compared to mCherry-expressing control neurons. Scale bar, 5 µm. (h) Quantitative analysis of 212,284Geph cluster size in control mCherry and DHHC-12-HA–expressing neurons. (i) Expression of dhhc12 mRNA in cultured hippocampal neurons (h. neurons) was validated by PCR using cDNA prepared from neurons. A plasmid encoding the respective dhhc gene was used as control. (j) Primary hippocampal neurons were grown in medium containing cell-penetrating siRNAs to knock down expression of DHHC-12 or DHHC-16. Lysates were subjected to ABE, and the palmitoylation level of gephyrin, GABAAR γ2, and PSD-95 was analyzed by immunoblotting. We used 10% of the lysate to detect total levels of the respective proteins; scr., scrambled. (k) Quantification of three independent experiments shows gephyrin palmitoylation levels normalized to total protein (***p<0.001 using Student's t test).
Figure 6.
Electrophysiological analysis of GABAergic and glutamatergic mPSCs.
(a) Example traces showing mPSCs recorded from DHHC-12– and DHHS-12–expressing neurons (identified using bicistronic mCherry expression). AMPAR- and GABAAR-mediated mPSCs were distinguished according to fast (AMPAR) and slow (GABAAR) decay kinetics. (b) The zoomed region of a recording from a DHHC-12–expressing neuron illustrates exponential fits of the different decay time constant (gray, AMPAR, τ = 3.8 ms; red, GABAAR, τ = 19.2 ms). (c–f) Quantification of amplitude, frequency, rise times (10–90), and decay time constants of AMPAR- and GABAAR-mediated mPSCs in DHHC-12– or DHHS-12–expressing neurons. Number of quantified events or neurons are given in parentheses.
Figure 7.
Subcellular localization of the DHHC-12/gephyrin interaction.
(a) Representative image of a primary hippocampal neuron transfected with DHHC-12–HA (red) immunostained together with the Golgi-marker giantin (green). The higher magnification below shows Golgi localization of DHHC-12 in the cell body and in dendrites as vesicle-like structures. (b) PLA dots (red) demonstrate direct interaction of gephyrin-GFP and DHHC-12–HA in cultured neurons. PLA dots in the inset show gephyrin-GFP and DHHC-12–HA interaction outside gephyrin clusters. Scale bars, 20 µm. (c) Co-immunoprecipitation of gephyrin-myc from HEK293 lysates co-expressing HA-tagged DHHC enzymes as indicated with HA-tag–specific antibodies.
Figure 8.
Gephyrin palmitoylation is regulated by GABAAR-activity.
(a–e) Primary hippocampal neurons were cultured 14 DIV and exposed to 50 µM bicuculline (BIC) or GABA for 1 h. (b) Representative Western blot shows altered gephyrin palmitoylation in the presence of BIC and GABA. (c) Quantification of gephyrin palmitoylation in neurons treated as in (a). Palmitoylation was normalized to total levels of gephyrin (n = 3). (d) Representative immunoblot of subcellular fractionation of neurons after treatment as in (a). (e) Representative immunostaining of gephyrin and VGAT. Scale bar, 5 µm. Histogram showing quantification of gephyrin cluster size after exposure to bicuculline or GABA compared to control cultures (see a) from three independent cultures. All data are means ± SEM (**p<0.01 using Student's t test).
Figure 9.
Model for palmitoylation-mediated regulation of gephyrin clustering and plasticity.
Gephyrin, palmitoylated on Cys212 and 284, forms stable clusters at the postsynaptic membrane. Activity of GABAergic synapses promotes (de)palmitoylation of gephyrin by yet unknown mechanisms. Silencing of GABAergic transmission leads to gephyrin depalmitoylation and membrane release, ultimately decreasing the size of gephyrin clusters. The palmitoyl transferase DHHC-12, localized in the Golgi in neurons including presumed Golgi outposts in primary dendritic shafts, is the principle gephyrin-palmitoylating enzyme and allows dynamic (re)palmitoylation of gephyrin.