| Annotated protein: | Membrane-associated guanylate kinase, WW and PDZ domain-containing protein 2 (Atrophin-1-interacting protein 1) (AIP-1) (Membrane-associated guanylate kinase inverted 2) (MAGI-2) (Synaptic-scaffolding molecule) (S-SCAM). Gene symbol: MAGI2. Taxonomy: Rattus norvegicus (Rat). Uniprot ID: O88382 |
| antibody wiki: | |
| SynGO gene info: | SynGO data @ MAGI2 |
| Ontology domain: | Biological Process |
| SynGO term: | regulation of neurotransmitter receptor localization to postsynaptic specialization membrane (GO:0098696) |
| Synapse type(s): | hippocampus, glutamatergic |
| Annotated paper: | Danielson E, et al. "S-SCAM/MAGI-2 is an essential synaptic scaffolding molecule for the GluA2-containing maintenance pool of AMPA receptors" J Neurosci. 2012 May 16;32(20):6967-80 PMID:22593065 |
| Figure(s): | Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5 and Fig. 8C-F |
| Annotation description: | Fig. 1, MAGI2 overexpression in hippocampal neurons increase dendritic spines head width, the proportion of mushroom-type spines and maturation of dendritic spines. Literal: "Following S-SCAM expression, as shown in Figure 1C-E, there was a specific increase in spine head width by 24%... with no significant change in the spine length.... The cumulative plot of spine head width showed a uniform right shift, indicating that S-SCAM overexpression targeted all spines rather than a specific population (Fig. 1F; p < 0.001). Cumulative plot analysis of dendritic spine length did not show a significant difference between GFP control and S-SCAM overexpressed neurons (Fig. 1G; p = 0.35). The increase of spine head width after S-SCAM overexpression was accompanied by the increased proportion of mushroom-type spines (46% for GFP vs 62% for S-SCAM; p < 0.05) while decreasing the proportion of thin-type spines (37% for GFP vs 24% for S-SCAM; p < 0.05; Fig. 1H). There was no significant difference in stubby-type (16% GFP vs 12% S-SCAM; p = 0.33) and forked-type spines (1% GFP vs 2% S-SCAM; p = 0.17) following S-SCAM transfection. These results collectively indicate that increasing S-SCAM levels in hippocampal neurons promote the maturation of dendritic spines." Fig. 2, MAGI2 overexpression modify synaptic levels of some PSD proteins in excitatory synapses of hippocampal neurons Literal: "S-SCAM overexpression drastically increased GKAP puncta intensity at dendritic spines by 2.5-fold (Fig. 2A,B), when compared with nontransfected (Non-txf) neighboring neurons.... The increase of synaptic GKAP levels after S-SCAM overexpression was accompanied by a small decrease (14%) in GKAP puncta density ... (Fig. 2C), which was smaller than the decrease in spine density (~40% reduction; Fig. 1B). Further analyses indicated that S-SCAM overexpression increased the population of dendritic spines containing multiple GKAP puncta at the expense of single puncta spines ... (Fig. 2D,E). In contrast to GKAP, increased S-SCAM levels led to only a modest (1.2-fold) increase of PSD-95 intensity at dendritic spines ...(Fig. 2A,B). Surprisingly, despite the large increase in GKAP levels at dendritic spines, the levels of Shank, a GKAP-binding scaffolding protein, were reduced by 17% ... (Fig. 2A,B). Furthermore, the puncta densities of both PSD-95 and Shank were greatly reduced (by 44% for PSD-95; by 62% for Shank ...) (Fig. 2C), to a degree greater than the spine density reduction (by 30%; Fig. 1B).Together, these results demonstrate that increasing S-SCAM levels led to drastic changes in the protein composition of PSD scaffolding proteins." Fig. 3, MAGI2 overexpression specifically increases AMPA-type glutamate receptors at synapses. Literal: "We have found that increasing S-SCAM levels in hippocampal neurons led to a drastic increase in the amount of all three major GluA subunits (GluA1, GluA2, and GluA3) at dendritic spines (Fig. 3A,B). Surface staining of these proteins indicated that S-SCAM increased the surface expression of GluA1 (sGluA1) and GluA2 (sGluA2) at the dendritic spines as well (Fig. 3C,D).... However, the total expression levels of these proteins, as measured by the staining intensities in the soma and dendrites, did not change significantly (Fig. 3A; quantified in 3B). This suggests that the increase of AMPARs in the dendritic spine came at the expense of AMPARs localized outside dendritic spines. Since S-SCAM overexpression promoted the maturation of dendritic spines, we also examined the number of dendritic spines lacking sGluA2, which measures silent synapses indirectly. Remarkably, as shown in a Figure 3E, S-SCAM overexpression greatly reduced the percentage of the spines lacking sGluA2 ..., consistent with synapse maturation. Blocking NMDAR activity withAPV( 100miroM) did not prevent the S-SCAM-induced increase of sGluA2 levels (Fig. 3F,G...), indicating that the increase of AMPAR by S-SCAM overexpression is an activity-independent process. In contrast to AMPAR, GluN1 or GluN2B levels at dendritic spines were not significantly affected by S-SCAM overexpression (Fig. 3H, I), indicating that S-SCAM overexpression specifically increased AMPAR levels at synapses. Together, these data indicate that S-SCAM overexpression promoted the accumulation and/or stabilization of AMPARs at dendritic spines." Fig. 4D-O, MAGI2 knockdown by RNAi reduces surface AMPARs and led to the loss of dendritic spines Litaral: "We first examined the effect of S-SCAM RNAi on GluA2, a major subunit of AMPARs in the hippocampal neurons. S-SCAM RNAi reduced both the intensity and number of sGluA2 puncta significantly (Fig. 4D-F). The reduction of sGluA2 intensity started to show as early as 1 d post-transfection of S-SCAM RNAi and reached 60% of control levels 3 d post-transfection (Fig. 4E). The reduction of sGluA2 puncta density became statistically significant 2 d post-transfection (Fig. 4F) and was reduced to 17% of control level after 3 d... The reduction in sGluA2 was accompanied by a significant increase in the proportion of dendritic spines lacking sGluA2 staining ... (Fig. 4G). These data clearly indicate that S-SCAM RNAi removes sGluA2 from dendritic spines. S-SCAM RNAi also had a dramatic effect on dendritic spines Fig. 4B,D; GFP channels). Knockdown of S-SCAM reduced the dendritic spine density (Fig. 4H) and the size of dendritic spines ... (Fig. 4I). Furthermore, S-SCAM knockdown increased the proportion of stubby spines at the expense of mushroom-type mature spines (Fig. 4J). These data suggest that losing S-SCAM from synapses promotes the collapse of mushroom-type dendritic spines to stubby ones, which are eliminated eventually. Consistent with the reduction in the number of dendritic spines, S-SCAM RNAi greatly reduced the puncta density of PSD-95 ... (Fig. 4K,L) and Bassoon ... (Fig. 4L), indicating the reduction of overall synapse numbers. To further confirm the specificity of S-SCAM RNAi effect, we performed a "rescue" experiment with S-SCAM RNAi-resistant S-SCAM .... When rescuing S-SCAM-r was coexpressed with S-SCAM RNAi, the amount of sGluR2 was increased to a level similar to S-SCAM overexpression (Fig. 4M,N) and the number of dendritic spines was similarly restored (Fig. 4O). Thus, the effect of S-SCAM RNAi on AMPARs and dendritic spines was specifically related to the loss of S-SCAM proteins and not caused by non specific effect of RNAi. Together, these data point out that S-SCAM is an essential scaffolding molecule for the stabilization/maintenance of AMPARs and dendritic spines. Fig. 5, MAGI2 specifically modulates AMPAR-mediated synaptic transmission. Litaral: "we measured AMPAR mEPSCs after transfecting dissociated hippocampal culture neurons with S-SCAM or S-SCAM RNAi constructs. Increasing S-SCAM levels enhanced AMPA mEPSC amplitudes ... (Fig. 5A-C), while S-SCAM RNAi reduces AMPA mEPSC amplitudes ... (Fig. 5A-C). We did not observe significant changes in the AMPA mEPSC frequency after S-SCAM overexpression ... despite the reduction in dendritic spine density. This is likely due to the "unsilencing" of silent synapses by acquiring AMPAR (Fig. 3E). In contrast, S-SCAM RNAi greatly reduced mEPSC frequency ... consistent with the reduction in synapse numbers after S-SCAM RNAi (Fig. 4H). Therefore, S-SCAM levels directly influence synaptic AMPAR levels. S-SCAM overexpression in CA1 pyramidal neurons of hippocampal slices cultured by sindbis virus infection drastically increased AMPAR-mediated synaptic transmission measured at -60 mV ... (Fig. 5E,F). In contrast, we did not detect significant changes in the NMDAR-mediated responses measured at +40 mV ... (Fig. 5E,G), indicating that S-SCAM overexpression did not affect NMDAR-mediated synaptic transmission significantly. Importantly, S-SCAM-infected neurons showed a significant increase in the AMPA/NMDA ratio ... (Fig. 5H), indicating that S-SCAM overexpression specifically increased the AMPA component of excitatory synaptic transmission. We did not find a significant difference in the PPR of S-SCAM-infected CA1 neurons from uninfected neighboring neurons (Fig. 5I, J ), indicating that S-SCAM overexpression did not change presynaptic function significantly." Fig. 8C-F MAGI2 overexpression blocks NMDA-induced AMPAR internalization and induction of hippocampal LTD. Literal: "Consistent with this idea, hippocampal CA1 neurons infected with S-SCAM-virus indeed showed no induction of LTD (Fig. 8C,D)... In contrast, neighboring uninfected CA1 neurons exhibited a pathway-specific robust and stable LTD. ... Inclusion of APV in the bathing medium completely abolished the LTD formation indicating that the LTD is NMDAR dependent. CA1 neurons transfected with S-SCAM RNAi showed normal LTD indistinguishable from Non-txf neighboring neurons (Fig. 8E,F). Together, these results suggest that S-SCAM is not involved in the regulation of LTD-forming AMPAR pool." |
| Evidence tracking, Biological System: | Cultured neurons |
| Evidence tracking, Protein Targeting: | RNAi / shRNA Over-expression Antibody (detection) |
| Evidence tracking, Experiment Assay: | Confocal |
| Annotator(s): | Chiara Verpelli (ORCID:0000-0003-2949-9725) Carlo Sala (ORCID:0000-0003-0662-9523) |
| Lab: | CNR Neuroscience Institute Milan and Dept. of Biotechnology and Translational Medicine, University of Milan, 20129 Milan, Italy |
| SynGO annotation ID: | 1750 |
| Dataset release (version): | 20231201 |
| View annotation as GO-CAM model: |  |