SEARCH

SEARCH BY CITATION

Keywords:

  • AMPA receptor;
  • EPSC ;
  • GABA ;
  • glutamate;
  • neurotransmission

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  8. Supporting Information

Heterozygosity for missense mutations in Seipin, namely N88S and S90L, leads to a broad spectrum of motor neuropathy, while a number of loss-of-function mutations in Seipin are associated with the Berardinelli–Seip congenital generalized lipodystrophy type 2 (CGL2, BSCL2), a condition that is characterized by severe lipoatrophy, insulin resistance, and intellectual impairment. The mechanisms by which Seipin mutations lead to motor neuropathy, lipodystrophy, and insulin resistance, and the role Seipin plays in central nervous system (CNS) remain unknown. The goal of this study is to understand the functions of Seipin in the CNS using a loss-of-function approach, i.e. by knockdown (KD) of Seipin gene expression. Excitatory post-synaptic currents (EPSCs) were impaired in Seipin-KD neurons, while the inhibitory post-synaptic currents (IPSCs) remained unaffected. Expression of a shRNA-resistant human Seipin rescued the impairment of EPSC produced by Seipin KD. Furthermore, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-induced whole-cell currents were significantly reduced in Seipin KD neurons, which could be rescued by expression of a shRNA-resistant human Seipin. Fluorescent imaging and biochemical studies revealed reduced level of surface AMPA receptors, while no obvious ultrastructural changes in the pre-synapse were found. These data suggest that Seipin regulates excitatory synaptic function through a post-synaptic mechanism.

Abbreviations used
AMPA

α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid

BSCL2

Berardinelli–Seip congenital generalized lipodystrophy type 2

CGL

Congenital generalized lipodystrophy

CREB

cAMP response element-binding protein

EPSCs

excitatory post-synaptic currents

HRP

horseradish peroxidase

IPSCs

inhibitory post-synaptic currents

KD

Knockdown

PBS

phosphate buffered saline

Congenital generalized lipodystrophy (CGL), also known as Berardinelli–Seip syndrome, is a rare disorder characterized by loss of adipose tissue, severe insulin resistance, and hypertriglyceridemia (Agarwal and Garg 2006). CGL Patients with BSCL2 mutations have lower serum leptin levels, an earlier onset of diabetes, and higher prevalence of intellectual impairment (Van Maldergem et al. 2002; Agarwal et al. 2003).

BSCL2 encodes Seipin, a 398- or 462-amino acid endoplasmic reticulum (ER) protein with two transmembrane regions (Lundin et al. 2006). Two species of Seipin mRNA at approximately 2.4 and 1.8 kb are found in human, and are expressed in many tissues including brain, liver, skeletal muscle, kidney, pancreas, and testis (Magre et al. 2001). Genetic defects in the BSCL2 gene likely affect an early step in the development or differentiation of adipose tissue, resulting in generalized lipodystrophy (Payne et al. 2008; Cui et al. 2011). At the same time, human patients bearing BSCL2 gene mutations have higher prevalence of intellectual impairment and motor neuronal dysfunction compared with BSCL1 and normal individuals (Van Maldergem et al. 2002). However, the mechanisms underlying their pathogenesis, especially the neuronal defects, remain enigmatic.

Suppression of Seipin expression by small hairpin RNA (shRNA)-mediated Seipin knockdown (KD) results in down-regulation of the lipogenic enzymes 1-acylglycerol-3-phosphate O-acyltransferase 2 (AGPAT2, a gene implicated in CGL1) and diacylglycerol O-acyltransferase 2 (DGAT2) (Payne et al. 2008). It also interferes with the key adipogenic transcription factors peroxisome proliferator-activated receptor gamma (PPARγ) and CCAAT/enhancer-binding protein alpha (C/EBPα). These findings suggest that BSCL2 is upstream of AGPAT2 in the pathological process of lipodystrophy, and that loss of BSCL2 function interferes with the normal transcriptional cascade of adipogenesis during fat cell differentiation, leading to lipodystrophy (Payne et al. 2008; Chen et al. 2009).

Seipin is highly expressed in most regions of the central nervous system (CNS) including cerebellar cortex, cerebellum, and hypothalamus (Magre et al. 2001; Garfield et al. 2012). Considering its broad tissue distribution, it is possible that Seipin is a tissue-dependent and multi-functional protein. It has been speculated that the mutations in Seipin associated with CGL2 cause loss-of-function, whereas gain-of-function mutations in Seipin (N88S, S90L) lead to Seipinopathy (Agarwal and Garg 2004; Ito and Suzuki 2009). However, it remains unknown whether mutations in Seipin impact CNS function directly or the neurological deficit of CGL is secondary to lipodystrophy. As synaptic transmission mediates all brain related behavior including learning and memory, here we directly tested the function of Seipin in synaptic transmission by suppressing the expression of Seipin using shRNAs. We found that Seipin regulates excitatory synaptic transmission but not inhibitory synaptic transmission. Although Seipin is expressed throughout the neuronal cell body, we did not observe obvious pre-synaptic deficits. We found, however, a significant reduction in α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor-mediated whole-cell currents in Seipin KD neurons. Moreover, Seipin KD led to decreased AMPA receptor expression and surface AMPA receptor levels. Taken together, our results suggest that Seipin regulates synaptic transmission by modulating AMPA receptor levels.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  8. Supporting Information

Neuron cultures and whole-cell electrophysiological analysis

These procedures were performed essentially as previously described except that EPC10-2 (HEKA Electronic, Lambrecht, Germany) was used in our experiments (Maximov et al. 2007).

Seipin knockdown by lentivirus-mediated short-hairpin RNAs

Duplex siRNA oligonucleotides, wh342 (TCG ACC CGT AGA ACT CTA CTC TGA CTT TCA AGA GAA GTC AGA GTA GAG TTC TAC TTT TTT GGA AAT) and wh343 (CTA GAT TTC CAA AAA AGT AGA ACT CTA CTC TGA CTT CTC TTG AAA GTC AGA GTA GAG TTC TAC GGG) target against the coding region from nt 541 downstream of the start codon of mouse BSCL2 (GenBank accession number NM_008144). Wh324 (TCG ACC CGC TGA AGC AGA AGT TTA TGT TCA AGA GAC ATA AAC TTC TGC TTC AGC TTT TTT GGA AAT) and wh325 (CTA GAT TTC CAA AAA AGC TGA AGC AGA AGT TTA TGT CTC TTG AAC ATA AAC TTC TGC TTC AGC GGG) were used as a non-specific control. These two synthesized complementary DNA oligonucleotides were annealed and inserted immediately downstream of the H1 promoter of pFHUUIG vector (Fig. 3a), which was derived from FUGW (Lois et al. 2002). The human BSCL2 was inserted after Ub promoter with BamHI/EcoRI for the rescue experiment. All vesicular stomatitis virus (VSV)-G pseudotyped lentiviral vector stocks were produced by Effectin transfection of human embryonic kidney 293 T cells. Briefly, human embryonic kidney 293 T cells were cultured in Dulbecco's modified Eagle's medium (GIBCO Invitrogen, Carlsbad, CA, USA) containing 10% fetal bovine serum (HyClone, Logan, UT, USA), 100 units of penicillin, and 100 μg/mL streptomycin. The cells were cotransfected with appropriate amounts of vector plasmid, the HIV-1 lentiviral packaging constructs pRSVREV and pMDLg/pRRE (Dull et al. 1998), and the VSV-G expression plasmid pHCMVG (Yee et al. 1994). The viruses were collected from the culture supernatants on days 2 and 3 post-transfection. Neurons were infected with the lentivirus at DIV 5 and used for experiments at DIV 14–16.

Immunocytochemistry and image analysis

Neurons were fixed at DIV 14–16, with a solution containing 4% formaldehyde (Polyscience Inc., Warrington, PA, USA) and 4% sucrose (Sigma, St. Louis, Missouri, USA) in phosphate buffered saline (PBS) for 30 min at ∼23°C. Cells were washed three times with PBS, blocked for 1 h with 0.1% Triton X-100 and 5% goat serum in PBS, incubated with primary antibodies over night at 4°C, and subjected to secondary antibody application (Alexa 488-, 546-, or 633-conjugated) for 60 min. Antibodies used in the study include monoclonal against synaptophysin (Synaptic System, Göttingen, Germany; 1 : 200), mouse antibody against MAP2 (Abcam, Cambridge, UK; ab3096, 1 : 200), and secondary antibodies (Molecular Probes, Eugene, OR, USA. 1 : 400). Images were acquired on a confocal laser scanning microscope with 405, 488, 561, and 638 nm laser lines (A1R; Nikon, Tokyo, Japan). Image quantification was performed using National Institutes of Health's ImageJ software (v 1.44, Bethesda, MD, USA). Measurements were performed on total fluorescence using the JACOP plugin (Bolte and Cordelieres 2006).

Transmission electron microscopy

Samples for cultured neurons were fixed in 2.5% glutaraldehyde in PBS at 4°C for 1 h before osmication with 1% osmium tetroxide, pH 7.4 for 1 h. Subsequently, the samples were dehydrated through an ascending series of ethanol at ∼23°C before infiltration with acetone and resin followed by final embedding in resin which was allowed to polymerize at 60°C for 24 h. Samples were cut by an ultra-microtome (Leica, Wetzlar, Germany), mounted on formvar-coated copper grids and doubly stained with uranyl acetate and lead citrate. The grids were viewed in a JEOL 1010 transmission electron microscope (Jeol USA, Peabody, MA, USA).

Synaptosomal preparations and immunoprecipitation

Mice were anaesthetized by intraperitoneal injection of pentobarbital sodium and decapitated. The brains were rapidly removed and immersed in ice-cold Krebs solution (in mM: 116 NaCl; 1.8 KCl; 1 MgSO4; 1.2 KH2PO4; 25 NaHCO3; 2 CaCl2; 10 d-glucose) and homogenized in 0.32 M sucrose and 10 mM HEPES (pH 7.4). The homogenate was spun at 800 g for 5 min and supernatant collected and spun at 20 000 g for 20 min. The pellet was resuspended in 0.32 M sucrose, 10 mM HEPES and spun again (20 min, 20 000 g). The pellet thus formed was resuspended in Krebs + HEPES solution (in mM: 127 NaCl; 3.73 KCl; 1.18 MgSO4; 1.18 KH2PO4; 1.8 CaCl2; 11 d-glucose; 20 HEPES; pH 7.4). All procedures were carried out at 4°C. The extracts were pre-cleared by incubating with protein A Sepharose beads for 6–8 h at 4°C.

GluR2, Seipin antibodies, and rabbit IgG (3 μg) bound to protein A-Sepharose beads were each incubated overnight with 250 μL synaptosome extracts in 250 μL IP buffer (150 mM NaCl, 20 mM Tris-HCL, 2 mM EDTA, 2 mM dithiothreitol, 1% Triton X-100, pH 7.4 plus protease inhibitors) at 4°C for 6–8 h. Beads were washed twice in buffer A (identical to IP buffer except that it contains 0.5% Triton X-100) and twice in buffer B (identical to IP buffer except that it contains 0.1% Triton X-100) before being re-suspended in sodium dodecyl sulfate sample buffer. Immunoprecipitated proteins were separated on sodium dodecyl sulfate–polyacrylamide gel electrophoresis and analyzed by immunoblot using respective antibodies.

Cell-ELISA assays

Cell-ELISA assays (colorimetric assays) were carried out as recommended in the In-Cell ELISA Colorimetric Detection Kit (The Thermo Scientific Pierce, Rockford, IL, USA).

Neurons were culture in 24-well plates, and fixed in 4% paraformaldehyde for 20 min. Cells were incubated with a monoclonal antibody against the N-terminal extracellular domain of the GluR2 receptor for labeling the receptors on the cell surface under non-permeabilized conditions, or the entire receptor pool under permeabilized conditions (with 0.2% Triton X-100 in PBS for 10 min), horseradish peroxidase (HRP)-conjugated secondary antibody (1 : 400) was then used. The cells were washed three times with PBS, and incubated with 0.5 mL of HRP substrate for 5 min. The absorbance at 450 nm (A450) was read on a spectrophotometer. The plate was washed twice with 200 μL/well of ultrapure water and 100 μL/well Janus Green Whole-Cell Stain was added for 5 min at ∼23°C. After washing three to five times with 200 μL/well of ultrapure water, the Elution Buffer was added and the plate was incubated for 10 min at ∼23°C. The absorbance at 615 nm (A615) was then measured. Four groups were averaged from each experimental condition, and the A450 values were normalized by A615 values from corresponding wells.

Biotinylation of cell-surface proteins

A Cell-Surface Protein Isolation Kit (Thermo Scientific Pierce, Rockford, IL, USA) was used for biotinylation of proteins according to the manufacturer's protocol. Neurons were cultured in 60-mm cell culture plates for about 2 weeks. For the AMPA receptor plasma membrane insertion experiment, the neurons were treated with vehicle control (perfusion bath solution) or solution with glycine (200 μM) for 5 min at ∼23°C. The neurons were washed twice with ice-cold PBS and incubated with EZ-link Sulfo-NHS-SS-Biotin for 30 min at 4°C followed by quenching solution (150 μL/plate). Cells were then lysed with 250 μL of lysis buffer with a protease inhibitor mixture. After centrifugation at 10 000 g for 2 min at 4°C, 50 μL cell lysate was taken as total protein. The cell lysate was precipitated with Immobilized NeutrAvidin Gel. The bound proteins were released by sample buffer (62.5 mM Tris-HCl, pH 6.8, 1% sodium dodecyl sulfate, and 10% glycerol) containing 50 mM dithiothreitol. The biotinylated cell surface protein was subjected to 10% Tris-glycine gels and transferred to nitrocellulose for chemiluminescence detection and densitometric quantification.

Miscellaneous

All cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). The human BSCL2 was cloned from ATCC 328876 with primers (5′-GC GGATCC ATG TCTACAGAAAAGGTA GACC-3′, 5′-GG GAATTC TC A GGAACTAGAGCAGGTG -3′).

Rabbit polyclonal antibody against vGLUT 1, mouse monoclonal antibodies against vGAT, and Bassoon were purchased from Synaptic Systems GmbH. Rabbit antibodies against GluR2 (ab20673, Abcam) and mouse monoclonal antibody against MAP2 (ab3096, Abcam) were from Abcam. Mouse antibody against GluR2 (MAB397, Millipore Corporation, Billerica, MA, USA) was from Millipore. Anti-Seipin polyclonal antibodies were raised by injecting purified GST-mSeipin (285–443) fusion protein into rabbits, and produced by the Biological Resource Centre (BRC) of Biomedical Sciences Institutes (Singapore).

Statistical analysis

Data were shown as means ± SEMs. Student's t-test was performed for two groups and one-way anova for > two groups, with values of p < 0.05 considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  8. Supporting Information

Expression pattern of Seipin in nervous system

Previous studies have shown high mRNA expression of BSCL2 in the brain (Magre et al. 2001; Garfield et al. 2012). We thus investigated protein expression of Seipin in the brain to understand its possible role in the CNS. Western blots showed that Seipin was strongly expressed in different brain regions including the cortex, hippocampus, hypothalamus, and cerebellum (Fig. 1a). Immunofluorescence staining confirmed a broad range and high level expression of Seipin in cortex, caudate-putamen, and cerebellar cortex (Fig. 1b). Seipin showed clear colocalization with the synaptic vesicle marker, synaptophysin in the cortex and hippocampus (Fig. 1c and d). In cultured cortical neurons, Seipin was colocalized with ER marker calnexin (Figure S1a), confirming Seipin as an ER membrane protein (Windpassinger et al. 2004). Consistent with the notion that ER is present in both dentrites and axons (Ramirez and Couve 2011), we found that Seipin was colocalized with the dendritic marker MAP2 (Figure S1b), and synaptic vesicle marker synaptophysin (Figure S1c), but not with post-synaptic protein PSD95 (Figure S1d).

image

Figure 1. Distribution of Seipin expression in brain (a) Expression of Seipin in brain of 3-month old mouse. (b) Immunofluorescence staining of Seipin in a sagittal section of mouse brain. Scale bar = 500 μm. (c) Cortical expression of Seipin (green), synaptophysin (red). (d) Hippocampal expression of Seipin (green), synaptophysin (red). Scale bar = 100 μm.

Download figure to PowerPoint

KD of Seipin does not affect synapse formation

To define the role of Seipin in synaptic neurotransmission, we adopted the loss of function approach by using a shRNA-mediated suppression of gene expression. By screening of several shRNAs from lentiviral expression, we identified one shRNA that effectively suppressed Seipin expression. Western blots showed significant reduction in Seipin expression in both total cell lysates and membrane fractions of neurons after Seipin KD (Fig. 2a and b).

image

Figure 2. Normal excitatory and inhibitory synapse distribution by Seipin KD (a) Total and membrane expression of Seipin in control and Seipin KD neurons. (b) Normalized expression of total and membrane expression of Seipin in control and Seipin KD neurons. n = 4, = 0.03 and 0.02 for total and membrane expression, respectively. Control (Ctrl), Seipin KD (KD). All data are shown as means ± SEMs. t-test, *< 0.05. (c) vGLUT (green) and Bassoon (red) immunostaining for scrambled shRNA control and Seipin KD neurons, Scale bar = 10 μm. (d) vGAT (red) and pre-synaptic marker synaptophysin (green) immunostaining for scrambled shRNA control and Seipin KD neurons. Scale bar = 10 μm.

Download figure to PowerPoint

To determine whether Seipin KD affects synapse formation, we analyzed both excitatory (Fig. 2c) and inhibitory (Fig. 2d) synapse distribution in scrambled shRNA controls (Fig. 2c, d upper panel) and in Seipin KD neurons (Fig. 2c, d lower panel) at DIV 14. Antibodies to vGLUT and vGAT were used to immunolabel excitatory and inhibitory synapses, and the pre-synaptic markers Bassoon or synaptophysin antibodies was used to immunolabel total synapses, respectively. The colocalization analysis for excitatory synapses in Fig. 2c (green signal - vGLUT positive puncta, red signal - Bassoon positive puncta) for control and Seipin KD neurons were not significantly different. The Pearson's coefficient (PC) (r) of the vGLUT-labeled (green channel image) and Bassoon labeled axon terminals (red channel image) were 0.53 and 0.50 for control and Seipin KD neurons, respectively (Figure S3ai, iii). The distributions of PC against the thresholds were almost the same for control and Seipin KD neurons (Figure S3aii, iv). Similar results were observed after immunostaining for inhibitory synapses (Figure S3b). There was no significant difference for the vGAT colocalization with synaptophysin for control and Seipin KD neurons (Fig. 2d, Figure S3b). Together, these results show that suppression of Seipin expression has no effect on either excitatory or inhibitory synapse formation.

KD of Seipin reduces the evoked excitatory post-synaptic current (EPSC)

We then examined Seipin's function in neurotransmission by electrophysiology. We examined excitatory synaptic transmission by analyzing pharmacologically isolated excitatory post-synaptic currents (EPSCs) after blocking inhibitory post-synaptic currents (IPSCs) mediated by GABAA receptors with picrotoxin (50 μM). Figure S2 illustrates the setup of electrophysiological experiments on infected neurons for the studies of synaptic vesicle released by applying local stimulation with bipolar electrodes placed near the post-synaptic neuron (Maximov et al. 2007, 2009). Postsynaptic neurons were held at −70 mV to detect EPSCs primarily mediated by AMPA receptors with the pipette solution containing 135 mM Cl- and 10 mM QX314 as a blocker of sodium currents in the post-synaptic cells.

To detect whether Seipin KD could alter basal synaptic properties of cultured neurons, whole-cell patch clamp was performed to record miniature EPSC (mEPSCs) in the presence of 1 μM TTX in control, Seipin KD, and Seipin KD neurons that express hSeipin (Fig. 3a). The frequency of the miniature EPSCs was significantly reduced by Seipin KD, and the reduction was reversed by human Seipin expression (2.10 ± 0.11 Hz for control, 1.32 ± 0.09 Hz for Seipin KD and 2.36 ± 0.10 Hz for Seipin rescue). The amplitudes of the miniature EPSCs were unaffected (11.57 ± 0.89 pA for control, 11.41 ± 0.91 pA for Seipin KD, and 11.75 ± 0.79 pA for Seipin rescue) (Fig. 3b–d). The cumulative frequency curves showed no significant difference between Seipin KD and control neurons (Fig. 3e).

image

Figure 3. Reduced excitatory post-synaptic currents (EPSCs) in Seipin KD neurons (a) Illustration of Seipin rescue strategy. See text for details. (b) mEPSC recording of control (Ctrl), Seipin KD (KD), and Seipin rescue (KD-WT) neurons. (c) Amplitude and (d) Frequency analysis of mEPSC. n = 18, 22, 20 for control (Ctrl), Seipin KD (KD), and Seipin KD-rescue (KD-WT), respectively. (e) mEPSC amplitude distribution and cumulative frequency of control (Ctrl), Seipin KD (KD), and Seipin rescue (KD-WT). (f) Evoked EPSC from neurons of control (Ctrl), Seipin KD (KD), Seipin KD-rescue by human Seipin (KD-WT). (g) Statistic analysis of amplitudes of evoked EPSC. n = 24, 17, 18 for control, Seipin KD and Seipin KD-rescue, respectively. (h) Sucrose-induced current in neurons of control (Ctrl), Seipin KD (KD), Seipin KD-rescue by human Seipin (KD-WT). (i) Statistic analysis of charge transfer for control (Ctrl), Seipin KD (KD), and Seipin KD-rescue (KD-WT) neurons. n = 20, 16, 15 for control, Seipin KD and Seipin KD-rescue, respectively. All data are shown as means ± SEMs, one way anova, *< 0.05, **< 0.01.

Download figure to PowerPoint

We also measured evoked EPSC of control, Seipin KD, and Seipin KD-rescue neurons, and found that the amplitude of evoked EPSC was significantly decreased by Seipin KD and that this reduction was rescued by human Seipin expression (1.19 ± 0.10 nA for control, 0.90 ± 0.06 nA for Seipin KD, and 1.55 ± 0.08 nA for Seipin rescue) (Fig. 3f and g). Next, we examined the charge integral induced by application of hypertonic buffer containing 0.5 M sucrose, as a measure of the readily releasable pool (RRP) of synaptic vesicles (Rosenmund and Stevens 1996). The charge integral was significantly reduced by Seipin KD, and this reduction was reversed by expression of human Seipin (0.47 ± 0.03 nC for control, 0.32 ± 0.04 nC for Seipin KD, and 0.61 ± 0.06 nC for Seipin KD-rescue neurons) (Fig. 3h and i).

Seipin KD does not affect inhibitory synaptic transmission

We also examined Seipin's function in inhibitory synaptic transmission by analyzing pharmacologically isolated GABAergic IPSCs in control and Seipin KD neurons. The excitatory AMPA receptors and N-Methyl-d-aspartate (NMDA) receptors were blocked by 6-cyano-7-nitroquinoxaline (CNQX) (20 μM) and 2-amino-5-phosphonopentanoic acid (APV) (50 μM). Miniature IPSCs (mIPSCs) were not affected by Seipin KD. No difference was observed for the amplitudes of mIPSC (38.87 ± 1.73 pA for control and 39.43 ± 2.22 pA for Seipin KD) or the frequency of mIPSC (0.86 ± 0.07 Hz for control and 0.91 ± 0.03 Hz for Seipin KD) (Figure S4a–c). Moreover, the amplitude distribution and cumulative frequency of the amplitudes of IPSCs were not different between Seipin KD neurons and controls (Figure S4d). Evoked IPSC (Figure S4e and f) and charge integral were measured (Figure S4g–h). KD of Seipin had no effect on the amplitudes of evoked IPSC (1.35 ± 0.12 nA for control; 1.26 ± 0.12 nA for Seipin KD) (Figure S4f) or the charge integral (1.60 ± 0.07 nC for control; 1.69 ± 0.09 nC for Seipin KD) (Figure S4h).

No ultrastructural changes in Seipin KD synapses

The reduction in EPSC and the RRP by Seipin KD raises the possibility that the number of vesicles docked at the active zone may be decreased. To test this possibility, we examined the ultrastructure of synapses in cultured neurons by electron microscopy. Excitatory and inhibitory synapses were analyzed after classifying them into ‘asymmetric’ and ‘symmetric’ synapses, a method that has been shown to reliably distinguish between excitatory and inhibitory synapses (Colonnier 1968). Analysis of asymmetric synapses in the cultured cortical neurons showed that Seipin KD exhibited no difference in number of vesicles that are either close to the active zone (9.69 ± 0.38 for control and 10.03 ± 0.80 for Seipin KD within 100 nm distance from the synaptic cleft) or docked at the active zone (3.0 ± 0.22 for control and 3.13 ± 0.30 for Seipin KD) (Fig. 4a and b). The post-synaptic density of the synapses in Seipin KD neurons was not significantly different from controls (20.68 ± 1.31 nm for control and 21.06 ± 2.37 nm for Seipin KD). Moreover, no difference was found in the synaptic vesicle size (26.04 ± 0.58 nm for control and 25.46 ± 1.46 nm for Seipin KD). Analyses of symmetric synapses in the cultured cortical neurons also showed no significant difference between Seipin KD neurons and controls (data not shown). Seipin was previously shown to regulate fusion of lipid droplets (Fei et al. 2008). The finding that synaptic vesicle size was not affected in Seipin KD neurons is consistent with the proposed cell-autonomous functions of Seipin in different cell types (Tian et al. 2011).

image

Figure 4. Reduced α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) current in Seipin KD neurons (a) Ultrastructure analysis of synaptic morphology in cultured cortical neurons. EM images for control (left panel) and Seipin KD neurons (right panel). T- Axon terminal, D - Dendrite, * - synapse. (b) Analysis of asymmetric synapses for control and Seipin KD neurons. There was no significant difference for vesicle number (within 100 nm from synaptic cleft), docked vesicle number (within 200 nm length of cleft), post-synaptic density and vesicle size. n = 24, 26 for control and Seipin-KD, respectively. (c) Dose dependent AMPA current densities in control and Seipin KD neurons. (d) Normalized AMPA current density (I) with maximum current density (Imax) of control and Seipin KD neurons (n = 8 and 10 for control and Seipin KD, respectively). (e) Recording of AMPA current induced by application of 100 μM AMPA in control (black) and Seipin KD neurons (faint red). The current was blocked by 20 μM CNQX, gray line. Voltage was clamped at −70 mV. (f) Statistic analysis of the AMPA current density by Seipin KD under the application of 100 μM AMPA. n = 18, 15, 15 for control, Seipin-KD, and Seipin KD-rescue, respectively. All data are shown as means ± SEMs, one way anova, *< 0.05.

Download figure to PowerPoint

Seipin KD reduces post-synaptic AMPA current and glutamate receptor expression

To examine possible changes in the post-synaptic expression of receptors by Seipin KD, the dose dependent AMPA current was tested in control and Seipin KD neurons. Decreased AMPA current density was observed after Seipin KD (Fig. 4c), but normalized current densities showed no significant difference for the dose dependent analysis (Fig. 4d). Post-synaptic AMPA currents with 100 μM AMPA application (Fig. 4e–f) and GABA currents with 100 μM GABA application (Figure S5a–b) were shown and analyzed. The AMPA currents can be suppressed by extracellular application of the AMPA receptor blocker CNQX (20 μM) (Fig. 4e) and GABA currents can be suppressed by the GABA receptor blocker bicuculline (20 μM) (data not shown). AMPA current density was significantly decreased by Seipin KD, and the reduction was rescued by human Seipin expression (0.090 ± 0.010 nA/pF for control, 0.070 ± 0.007 nA/pF for Seipin KD and 0.095 ± 0.007 nA/pF for Seipin KD-rescue) (Fig. 4e–f), but GABA current density was not affected (0.199 ± 0.010 nA/pF for control and 0.196 ± 0.010 nA/pF for Seipin KD) (Figure S5a and b).

The above findings indicate that decreased EPSCs could be caused by the reduction in post-synaptic glutamate receptor expression. To test this hypothesis, we analyzed the expression level of glutamate receptor in cultured neurons. Seipin KD neurons showed significantly reduced expression level of Seipin (0.41 ± 0.04), GluR1 (0.57 ± 0.04), and GluR2 (0.68 ± 0.04, control = 1) (Fig. 5a and b). Moreover, quantitative PCR (Q-PCR) showed that Seipin, stearoyl-CoA desaturase (SCD), fatty acid synthase (FAS), low density lipoprotein receptor (LDLR), sterol regulatory element binding protein 2 (SREBP2), SREBP1c, cAMP response element-binding protein (CREB), GluR1, and GluR2 mRNA levels were significantly decreased by Seipin KD compared with controls (Fig. 5c).

image

Figure 5. Reduced expression of GluR1 and GluR2 by Seipin KD (a) western blotting for protein expression for control and Seipin KD neurons. (b) Statistic analysis of normalized expression level of proteins affected by Seipin KD. n = 4. Data were normalized to β-actin expression levels. (c) mRNA levels revealed by Q-PCR. n = 3. All data are shown as means ± SEMs. t-test, *< 0.05, **< 0.01.

Download figure to PowerPoint

We confirmed the results of quantitative PCR (qPCR) by immunofluorescence staining in Seipin KD neurons. Cell surface GluR2 immunostaining showed significant reduction in Seipin KD neurons, and this reduction was rescued by human Seipin expression (Fig. 6a). Colorimetric assay (Man et al. 2000) revealed marked reduction in surface and total GluR2 expression level (Fig. 6b and c), but the surface/total ratio of the GluR2 expression was not changed (Fig. 6d). To verify whether Seipin KD suppresses AMPA receptor membrane trafficking, AMPA receptor insertion was analyzed in infected neurons after glycine stimulation (Park et al. 2004). Western blots showed decreased surface and total GluR2 expression level by Seipin KD (Fig. 6e), but the membrane inserting ability of AMPA receptor, represented by the ratio of surface receptor/total receptor was not decreased by Seipin KD (Figure S5c). We next determined whether Seipin has any direct connection with post-synaptic components. Seipin antibody did not immunoprecipitate AMPA receptors (Figure S5d). Thus, the decreased expression level of AMPA receptor by Seipin KD was not induced by direct interaction between Seipin and AMPA receptors. Also, no direct interactions were detected between Seipin and the post-synaptic marker PSD95 (Figure S1d).

image

Figure 6. Reduced GluR2 surface expression level in Seipin KD neurons (a) Upper panel, immunostaining for scrambled shRNA infected neurons (green, GFP), cell surface GluR2 (red), and total GluR2 (blue). Middle panel, immunostaining for Seipin KD neurons. Lower panel, immunostaining for Seipin KD-rescue (KD-WT) neurons. Scale Bar = 10 μm. (b–d) Colorimetric assay for surface and total GluR2 expression level in neurons. (b) Normalized total GluR2 expression level. (c) Normalized surface GluR2 expression level. (d) Normalized surface/Total ratio of GluR2 expression level. (e) GluR2 Inserting into cell membrane by glycine stimulation (200 μM) and isolated by surface biotinylation assays. Western blot for total and surface expression level of GluR2 and Seipin in scrambled control, Seipin KD and Seipin rescue neurons. All data are shown as means ± SEMs, one way anova, *< 0.05.

Download figure to PowerPoint

Seipin functions through a post-synaptic mechanism

As Seipin showed apparent colocalization with pre-synaptic markers (Fig. 1c and d), we tested whether Seipin had a possible pre-synaptic function by analyzing paired-pulse facilitation. Seipin KD led to reduced AMPA currents, but had no effect on paired-pulse facilitation (Figure S6a and b), suggesting that synaptic release probability is normal in Seipin KD neurons. We also examined EPSCs mediated by NMDARs or AMPARs. Seipin KD did not affect NMDAR-mediated EPSC (Figure S6c and d). Consistent with reduced AMPA currents, the AMPA/NMDA ratio was significantly decreased (Figure S6e). These results are consistent with the notion that Seipin regulates neurotransmission via a post-synaptic mechanism, although a pre-synaptic role cannot be ruled out.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  8. Supporting Information

Seipin was first identified as a candidate for CGL2 (Agarwal and Garg 2004). Individuals homozygous for null mutations in Seipin have severe lipoatrophy and intellectual impairment but not abnormality of motor neurons (Agarwal and Garg 2003, 2004; Fu et al. 2004). In contrast, lipodystrophy and metabolic disturbances have not been reported in Seipinopathies associated with dominant hereditary motor neuron diseases (Irobi et al. 2004; Auer-Grumbach et al. 2005). Mutations in the N-glycosylation site of Seipin are associated with the motor neuron diseases and result in accumulation of unfolded protein in the ER. This leads to activation of the unfolded protein response (UPR) and cell death, suggesting that these diseases are associated with ER stress (Ito and Suzuki 2009).

Here, we determined the subcellular localization of Seipin in cultured cortical neurons and the distribution of Seipin in brain tissues. Seipin expression is detected in neurons of the cortex and hippocampus by immunohistochemistry. Suppression of Seipin expression in cultured cortical neurons decreased the expression level of CREB and AMPA receptors. By electrophysiological study, we found that Seipin KD reduced the amplitude of evoked EPSC and frequency of the mEPSC but not IPSCs, suggesting that Seipin plays an important role in neurotransmission. AMPA receptor mediated EPSCs are the major excitatory drive in the nervous system. The four AMPAR subunits, GluR1–4 assemble in different combinations to form tetrameric channels (Rosenmund et al. 1998). Most AMPARs are composed of GluR1-GluR2 or GluR2-GluR3 combinations. AMPARs are first assembled in the ER, and associate with the ER chaperones BiP and calnexin (Rubio and Wenthold 1999), GluR2 colocalizes extensively with BiP in the ER (Greger et al. 2002). Some AMPAR subunit mRNAs are dendritically localized, suggesting that local synthesis of AMPAR subunits regulates local receptor abundance and composition (Ju et al. 2004; Grooms et al. 2006). As Seipin is an ER protein, and Seipin KD reduced the total and surface expression level of GluR2, we tested whether Seipin could be another unidentified chaperone that regulates AMPA receptor expression and membrane insertion. Our data, however, suggest that Seipin does not affect AMPA trafficking from the intracellular pool to the plasma membrane pool.

It is not surprising that expression of lipogenesis genes such as SCD, FAS, LDLR, and SREBP were also reduced by Seipin KD, as BSCL2 gene was defined as a lipid dystrophy gene. SREBP transcription factors regulate expression of a number of genes involved in biosynthesis and uptake of cholesterol, saturated/ monounsaturated fatty acids, triglycerides, and phospholipids (Dietschy and Turley 2004; Nakamura et al. 2004). Reduction in expression levels of CREB and GluRs by Seipin KD suggests that the Seipin not only affects lipid biosynthesis but also the expression level of glutamate receptors. CREB regulates not only GluR2 promoter activity (Brene et al. 2000) but also several downstream transcription factors, such as c-fos, zif268, and fosB. The latter is a transcription factor that can up-regulate GluR2 (Kelz et al. 1999). Moreover, CREB regulates the expression of brain-derived neurotrophic factor (BDNF) (Finkbeiner et al. 1997) and the GluR1 AMPA receptor subunit (Borges and Dingledine 2001; Olson et al. 2005).

We noticed altered frequency, but not the amplitude of mEPSCs in Seipin KD neurons (Fig. 3c and d). Although this is often caused by a pre-synaptic origin of synaptic modifications, our data support that a post-synaptic defect may account for the above finding: (i) we found the total and surface AMPAR expression levels were decreased in seipin KD neurons (Figs 5, 6); (ii) NMDAR-mediated EPSC was not affected by seipin KD (Figure S6c and d); and (iii) neither excitatory nor inhibitory synapse numbers appears to be affected by seipin KD (Fig. 2c and d). One possible scenario is that the observed reduction in AMPAR expression may lead to reduced number of functional synapses with AMPARs without affecting the number of AMPARs at individual functional synapses.

Synaptic transmission underlies all brain related behaviors and a proper excitation-inhibition (E-I) is essential for normal neuronal function. E-I balances are determined by excitatory and inhibitory synaptic drive. Over-inhibition has been suggested to underlie certain forms of intellectual disability (Fernandez and Garner 2007; Garner and Wetmore 2012). Our results that a reduction in excitatory synaptic transmission (i.e. EPSCs) without obvious modification in IPSC after suppression of Seipin expression suggest that loss of Seipin function may lead to imbalance of E-I ratio, which may help explain the intellectual impairment aspect of BSCL2.

In summary, we report that Seipin specifically regulates EPSCs in neurons, possibly through a post-synaptic mechanism. Our study provides novel insights in understanding the biology of Seipin protein in the brain.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  8. Supporting Information

We thank Ms Y.J. Wu for support on this study, and Dr. Clement Khaw and SBIC-Nikon Imaging Center for support on confocal microscopy. This study was supported by intramural funding from Agency for Science, Technology and Research (A*STAR) Biomedical Research Council (W.H.); Robert Wood Johnson Foundation to CHINJ and Brain and Behavior Research Foundation (Z.P.P). The authors declare that there is no duality of interest associated with this manuscript.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  8. Supporting Information
FilenameFormatSizeDescription
jnc12099-sup-0001-FigS1-S6.pdfapplication/PDF8513K

Figure S1. Seipin expression in cultured cortical neurons.

Figure S2. Stimulation setup for electrophysiology experiments.

Figure S3. Synapse number analysis.

Figure S4. Normal amplitude and frequency of mIPSCs in Seipin KD neurons.

Figure S5. Postsynaptic response to GABA stimulation.

Figure S6. Presynaptic origin of the facilitation and NMDA dependent EPSC.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.