Figure 1.
SORCS3 localizes to the post-synaptic density of hippocampal neurons.
(A) Hippocampal extracts of wild types mice were subjected to subcellular fractionation as described in the method section. Western blot analysis identifies SORCS3 in crude synaptosomal membranes (P2), light membrane fraction (P3), synaptic plasma membranes (SPM), as well as in the post-synaptic densities (PSD) I and II. The receptor is not seen in the synaptic vesicle-enriched fraction (LP2). Detection of PSD95 and synaptophysin served as markers of post-synaptic densities and synaptic vesicles, respectively. H: Hippocampal homogenate. (B) Total mouse brain extracts were subjected to affinity purification on resin coupled with synthetic peptides encompassing 14 amino acids of the cytoplasmic tails of SORCS3 (CS3-CT) or SORCS2 (CS2-CT) as detailed in the method section. Two proteins purified from CS3-CT but not the CS2-CT column as shown by SDS-PAGE and staining with Coomassie. These proteins were identified as PSD93 and PSD95 by mass spectrometry. (C) COS7 cells were transiently transfected with constructs encoding PSD95 together with either full-length murine SORCS3 or a receptor variant lacking the PDZ domain binding site (CS3CTSV). SORCS3 (lane 3) but not CS3CTSV (lane 4) co-immunoprecipitated with an anti-PSD95 antiserum (panel IP-PSD95). Panel Input (lanes 1 and 2) represents the cell lysate tested for PSD95 and SORCS3 prior to immunoprecipitation.
Figure 2.
Generation and histological analysis of a SORCS3-deficient mouse model.
(A) For targeted disruption of the murine Sorcs3 locus, a targeting vector was constructed introducing the neomycin phosphotransferase gene driven by the phosphoglycerate kinase promoter (PGK-neoR) into intron 2 of the Sorcs3 locus. The PGK-neoR cassette was flanked by FRT sites (open triangles). In addition, the vector introduced two loxP recombination sites (closed triangles) 5’ and 3’ of exon 1, respectively. Following standard homologous recombination in embryonic stem cells and blastocyst injections, mice carrying the modified gene through the germ line were bred with the flp deleter strain to remove PGK-neoR (targeted allele). For gene inactivation, mice were crossed with cre deleter mice to excise exon 1 that encodes the start codon and signal and pro-peptides of Sorcs3 (deleted allele). (B) RT-PCR analysis on brain tissue documents complete loss of transcripts encoding exon 1 in Sorcs3-/- animals as compared to Sorcs3+/+ controls. Minor amounts of an aberrant transcript encompassing exons 13-15 are seen. (C) Successful ablation of SORCS3 protein expression was confirmed by Western blot analysis of extracts from hippocampus (Hip), cortex (Ctx), and cerebellum (Cer) detecting SORCS3 in wild type mice (+/+), but in animals homozygous for the disrupted allele (-/-). Detection of Na+/K+ ATPase α-1 subunit (Na/K) served as loading control. (D) Histological sections stained with Nissl from hippocampi of Sorcs3+/+ and Sorcs3-/- mice. (E) Western blot analysis of the indicated proteins in the PSD fraction of hippocampi from Sorcs3+/+ and Sorcs3-/- mice. NR1/2B, NMDA glutamate receptor subunit 1/2B; GluR1/2, AMPA glutamate receptor subunit 1/2; mGlur5, metabotropic glutamate receptor type 5; p75NTR, nerve growth factor receptor; TrkB, neurotrophic tyrosine kinase, receptor, type 2.
Figure 3.
Long-term potentiation is normal in SORCS3-deficient mice.
Field recordings in the stratum radiatum in the CA1 of the hippocampus after stimulation of Schaffer collaterals in slices of wild type and SORCS3-deficient mice. (A) Exemplary traces depicting fEPSPs before (black line) and after tetanic stimulation (red line) using a high-frequency 100 Hz protocol in slices from wild type and SORCS3-deficient mice. Scale bar: 0.4 mV/4 ms. (B) Averaged slopes of fEPSPs in wild type and in SORCS3-deficient slices before and after high-frequency stimulation (arrow). Data are given as mean ± SEM (+/+: 157.5 ± 11.0%; -/-: 142.5 ± 8.7%, p>0.05).
Figure 4.
Long-term depression is impaired in SORCS3-deficient mice.
Field recordings of the stratum radiatum in the CA1 of the hippocampus after stimulation of Schaffer collaterals in slices of wild-type and in SORCS3-deficient mice (A-B) A low-frequency 1 Hz protocol was applied to induce NMDA receptor-dependent long-term depression (LTD). (A) Representative traces depicting fEPSPs before (black line) and after low-frequency stimulation (red line) in wild type and receptor-deficient mice. (B) Averaged fEPSP slopes before and after low-frequency stimulation (arrow) in slices of wild type and SORCS3 deficient animals (+/+: 83.6 ± 4.2%; -/-: 99.2 ± 4.9%, p<0.05). (C-D) A low-frequency 1 Hz paired-pulse protocol was used to elicit mGluR-dependent LTD. (C) Exemplary traces depicting fEPSPs before (black line) and after low-frequency paired-pulse stimulation (red line) in wild type and receptor-deficient mice. (D) Averaged fEPSP slopes before and after low-frequency paired-pulse stimulation (arrow) are given (+/+: 78.3 ± 8.3%; -/-: 98.1 ± 4.4%, p < 0.05). Scale bars in A and C: 0.4 mV/4 ms. Data in B and D are given as mean ± SEM.
Figure 5.
Normal paired-pulse facilitation in SORCS3-deficient mice.
(A) Representative traces of fEPSPs after paired pulses with increasing interpulse intervals are depicted for the indicated genotypes and conditions. Scale bars: 0.3 mV/20 ms. (B) Paired-pulse facilitation (PPF) was calculated as the ratio of the second fEPSP slope to the first fEPSP slope and plotted at different interstimulus intervals. No significant differences (p>0.05) at any interstimulus intervals were seen comparing Sorcs3+/+ and Sorcs3-/- mice. Application of the GABAB receptor agonist baclofen increased the PPF ratio equally in mice of both genotypes. Note that y-axis starts at 100%. Data are given as mean ± SEM.
Figure 6.
Normal behavior of SORCS3-deficient mice in the open field paradigm.
Locomotion of Sorcs3+/+ and Sorcs3-/- mice in an open field test run for 90 min on day 1 (panel A) and day 3 (panel B) of the test (n=3 per genotype). The distance traveled over a period of 5 minutes was averaged for each time interval (mean ± SD).
Figure 7.
Sorcs3-/- mice display defects in spatial learning and memory in the Barnes maze.
(A) Representative track plots illustrate the deficiency of individual Sorcs3-/- animals to locate the target hole on day 4 of the trial as compared to Sorcs3+/+ controls. (B) Wild type mice (Sorcs3+/+, n=7) display a steep learning curve as illustrated by the decreasing time required to enter the target hole during the four test days. In contrast, Sorcs3-/- mice (n=15) fail to acquire spatial memory. Statistical analysis was performed by a two-way ANOVA documenting a significant effect of genotype on learning performance: F=5.11, p<0.5. Subsequently, post hoc analyses were performed at individual time points using the two-tailed Student’s t-test for independent samples at each time point (*p<0.05). (C) Sorcs3-/- mice show significantly more nose poke errors on the fourth test day compared to wild-type controls. Two-tailed Student’s t-test was used for testing independent samples at day 4 (*p<0.05).
Figure 8.
Lack of SORCS3 results in increased fear extinction in mice.
(A) Mice were exposed to an inhibitory avoidance experiment as described in the method section. SORCS3-deficient mice display a more rapid extinction of fear over four consecutive days albeit at an intact acquisition of fear memory (shock). In contrast, wild type mice show no reduction in fear memory during the course of the experiment. Statistical analysis was performed by two-way ANOVA documenting a significant effect of genotype on learning performance: F=5.57, p<0.5. Subsequently, post hoc analyses were performed at individual time points using the two-tailed Student’s t-test for independent samples (*p<0.05, **p<0.01). (B) SORCS3-deficient and wild type mice display similar remote fear memory 16 days after the foot shock.
Figure 9.
SORCS3 interacts with PICK1 and affects its localization at the post-synaptic density.
(A) Lysates of HEK293 cells stably expressing SORCS3 were mixed with recombinant GST or GST-PICK1 as described in the method section. Pull-down experiments with glutathione sepharose beads (lanes 3-4) recovered SORCS3 from lysates containing GST-PICK1 (lane 4) but to a significantly lesser extent from GST-containing lysates (lane 3). Panel Input (lanes 1-2) documents the presence of SORCS3 in both samples prior to pull-down. (B) Immunodetection of PICK1 on extracts of hippocampus (Hip), cortex (Ctx), and cerebellum (Cer) documents equal levels of the protein in wild type (+/+) and SORCS3-deficient mice (-/-). (C) Representative Western blot analysis of synaptosomal preparations of wild type and Sorcs3-/- mouse hippocampi for PICK1 documents reduces levels of PICK1 in the SORCS3-deficient PSD as compared to the wild type control. Fractions were further probed against PSD95 as a control for accuracy of fractionation (absent in the synaptic vesicle preparation, LP2) and of equal loading. LS1, input supernatant for synaptic vesicle fraction; LP2, synaptic vesicle preparation; SPM, synaptic plasma membrane; PSDI and PSDII, postsynaptic densities. (D) Densitometric measurement of PICK1 levels in PSDI and PSDII fractions (as shown in panel B) from four independent experiments (13 mice per genotype). Intensities for PICK1 from PSDI and PSDII were combined and normalized against PSD95 for each experiment.