Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Aug;142(4):504-511.
doi: 10.1111/jnc.14056. Epub 2017 Jul 25.

Hierarchical organization and genetically separable subfamilies of PSD95 postsynaptic supercomplexes

René A W Frank  1 , Fei Zhu  2 , Noboru H Komiyama  2 , Seth G N Grant  2
Affiliations

Hierarchical organization and genetically separable subfamilies of PSD95 postsynaptic supercomplexes

René A W Frank et al. J Neurochem. 2017 Aug.

Abstract

PSD95 is an abundant postsynaptic scaffold protein in glutamatergic synapses that assembles into supercomplexes composed of over 80 proteins including neurotransmitter receptors, ion channels and adhesion proteins. How these diverse constituents are organized into PSD95 supercomplexes in vivo is poorly understood. Here, we dissected the supercomplexes in mice combining endogenous gene-tagging, targeted mutations and quantitative biochemical assays. Generating compound heterozygous mice with two different gene-tags, one on each Psd95 allele, showed that each ~1.5 MDa PSD95-containing supercomplex contains on average two PSD95 molecules. Gene-tagging the endogenous GluN1 and PSD95 with identical Flag tags revealed N-methyl D-aspartic acid receptors (NMDARs) containing supercomplexes that represent only 3% of the total population of PSD95 supercomplexes, suggesting there are many other subtypes. To determine whether this extended population of different PSD95 supercomplexes use genetically defined mechanisms to specify their assembly, we tested the effect of five targeted mouse mutations on the assembly of known PSD95 interactors, Kir2.3, Arc, IQsec2/BRAG1 and Adam22. Unexpectedly, some mutations were highly selective, whereas others caused widespread disruption, indicating that PSD95 interacting proteins are organized hierarchically into distinct subfamilies of ~1.5 MDa supercomplexes, including a subpopulation of Kir2.3-NMDAR ion channel-channel supercomplexes. Kir2.3-NMDAR ion channel-channel supercomplexes were found to be anatomically restricted to particular brain regions. These data provide new insight into the mechanisms that govern the constituents of postsynaptic supercomplexes and the diversity of synapse types. Read the Editorial Highlight for this article on page 500. Cover Image for this issue: doi. 10.1111/jnc.13811.

Keywords: Ion channel supercomplexes; Kir2.3; NMDA receptor; PSD95; synapse diversity; synapse proteome.

PubMed Disclaimer

Figures

Figure 1
Figure 1
Quantification of the GluN1/PSD95 relative abundance in vivo gave a 17 : 1 molar ratio of PSD95 over GluN1. (a) GluN1 and PSD95 were detected by sulfate–polyacrylamide gel electrophoresis (SDS‐PAGE) (lower panel) and triplicate dot blot (DB) Flag immunoblot (upper panel) of Glun1 TAP / TAP, Psd95 TAP / TAP mouse forebrains respectively. The TAP‐GluN1 and PSD95‐TAP SDS‐PAGE bands are similar to their counterparts in singularly TAP‐tagged mice (Psd95 TAP / TAP and Glun1 TAP / TAP), which indicates equal loading. (b) The molar ratio of Flag from Psd95 TAP / TAP and Glun1 TAP / TAP was quantified densitometrically using a dilution series, in which Psd95 TAP / TAP forebrain extracts were diluted with that of wildtype. Densitometry of dilution series indicated TAP‐PSD95 was 17 ± 3‐fold (mean ± SD) more concentrated than TAP‐GluN1.
Figure 2
Figure 2
Quantification of the in vivo stoichiometry of PSD95 in each 1.5‐PSD95 using compound heterozygous knockin tags (Psd95 TAP / GFP). (a) Schematic showing the gene structure of the last protein coding exon (exon 20) of Psd95 in Psd95 TAP / GFP hybrid mutant mice. Equal expression of both alleles is expected in each cell because Psd95 is on an autosomal chromosome (Chr 11). (b) Schematic showing the expected partitioning of TAP‐tagged PSD95 between captured (resin) and flow‐through (unbound) in a GFP immunoprecipitation (IP) from Psd95 TAP / GFP hybrid mutant forebrain extract. Only PSD95‐TAP assemblies containing at least one PSD95‐GFP will be captured and the expected distribution of PSD95‐TAP and PSD95‐GFP subunits between captured and flow‐through samples is depicted for different homo‐oligomeric states of 1.5‐PSD95 containing on average: 1 (monomer), 2 (dimer), 3 (trimer) or 4 (tetramer) PSD95 molecules. This partitioning, indicated as the percentage split for PSD95‐TAP captured and in the flow‐through, is dependent on the stoichiometry of PSD95 molecules in each complex. Green and cyan ellipses correspond to PSD95‐GFP and PSD95‐TAP subunits, respectively. For each oligomeric state, all possible assemblies containing different combinations of GFP‐ and TAP‐tagged PDS95 are shown. (c) Dilution series of Psd95 TAP / GFP forebrain extract into that of wildtype indicated sensitivity of quantification. These data show the dynamic range of sulfate–polyacrylamide gel electrophoresis (SDS‐PAGE) Flag immunoblot detection. (d) PSD95‐GFP was immunoprecipitated (IP) from Psd95 TAP / GFP and negative control (Psd95 TAP / WT) forebrain extract supernatant (Input) with GFP antibody. Top panel, SDS‐PAGE GFP immunoblot of IP shows near complete immunoprecipitation of all PSD95‐GFP (Captured) from Psd95 TAP / GFP hybrid double mutant forebrain extracts. Second panel, Flag immunoblot shows half PSD95‐TAP co‐precipitated with PSD95‐GFP, the remaining unbound PSD95‐TAP was detected in the flow‐through. Lower panel, Flag immunoblot of control IP (Psd95 TAP / WT) shows no PSD95‐TAP was captured in the absence of GFP‐tagged PSD95. The total protein loaded for SDS immunoblot was normalized across Input, Flow‐through and Captured lanes by supplementing with non‐tagged (wildtype)samples. Representative data from triplicate experiments shown. (e) Densitometric immunoblot quantification of PSD95‐GFP (GFP) in GFP IP. The band intensities from triplicate samples were measured and normalized to that of input. Error bars indicate 1 SD. These data show essentially all the PSD95‐GFP was captured by GFP IP. (f) Densitometric immunoblot quantification of PSD95‐TAP (Flag) in GFP IP (g, middle panel). The band intensities from triplicate samples were measured and normalized to that of input. Error bars indicate 1 SD. These data show half the PSD95‐TAP was co‐captured by PSD95‐GFP immunoprecipitation. Thus, each 1.5‐PSD95 supercomplex contains on average a dimer of PSD95 molecules. (g) 1.5‐PSD95 each containing two molecules of PSD95 were isolated in two sequential steps: Flag immunoaffinity purification (‘IAP’) followed by GFP immunoprecipitation (‘IP’) from Psd95 TAP/GFP hybrid double mutant mice. Psd95 TAP/WT mice were used as a negative control. PSD93 and GluN1 were detected by SDS‐PAGE immunoblot. These data show that the subset of 1.5‐PSD95 containing PSD93 and GluN1 each also contain a dimer of PSD95.
Figure 3
Figure 3
Characterization of supercomplex subtypes. (a) Mutant mouse screen of synaptic supercomplexes. BNP immunoblots of forebrain extracts from five mutant mouse lines show genetic dependencies for the assembly of 1.5‐Kir2.3, 1.5‐Arc, 1.5‐IQsec2 and 1.5‐Adam22. Each panel contains duplicates of wildtype (WT) (lane 1,2, duplicates), Psd95 −/− (lane 3,4), Psd93 −/− (lane 5,6), Glun2b 2A( CTR )/2A( CTR ) (lane 7,8), Glun2a 2B( CTR )/2B( CTR ) (lane 9,10), GluN2a del CTD. Immunoblotting antibody is indicated below each panel (IB). Arrow indicates 1.5 MDa bands. Molecular weight in MDa shown on right. Representative results from triplicate experiments shown. (b) Table summarizing data in Fig. 3a and recently published (Frank et al. 2016) showing the effect of different mutations (columns) on distinct components of 1.5‐PSD95 supercomplexes (rows). X, denotes assembly of supercomplex was blocked by the mutation. ‐, denotes assembly of the supercomplex was not blocked by the mutation. (c) Subunit‐depletion of GluN2B and PSD95 removes 1.5‐Kir2.3. Extracts from Glun1 TAP / TAP /Psd95 GFP / GFP double knockin mice were subunit‐depleted with antibodies (shown in lanes), then separated on BNP for immunoblotting with Kir2.3 antibody to show 1.5‐Kir2.3. Lanes; Input, total extract; immunodepleting antibodies (lanes shown left to right), non‐specific IgG, GFP, GluN2A, GluN2B. Arrow indicates 1.5‐Kir2.3. Molecular weight in MDa shown on right. IB, immunoblotting antibody. (d) Subunit‐depletion of PSD95 removes all 1.5‐Arc. Extracts from Glun1 TAP / TAP /Psd95 GFP / GFP double knockin mice were subunit‐depleted with antibodies (shown in lanes) then separated on BNP for immunoblotting with Arc antibody to show 1.5‐Arc. Lanes; Input, total extract; immunodepleting antibodies (lanes shown left to right), non‐specific IgG, GFP, GluN2B. Arrow indicates 1.5‐Arc. Molecular weight in MDa shown on right. IB, immunoblotting antibody. (e) Subunit‐depletion of PSD95 removes all 1.5‐IQsec. Extracts from Glun1 TAP / TAP /Psd95 GFP / GFP double knockin mice were subunit‐depleted with antibodies (shown in lanes), then separated on BNP for immunoblotting with IQsec2 antibody to show 1.5‐IQsec2. (f) Subunit‐depletion of PSD95 does not remove all Adam22. Extracts from Glun1 TAP / TAP /Psd95 GFP / GFP double knockin mice were subunit‐depleted with antibodies (shown in lanes) then separated on BNP for immunoblotting with Adam22 antibody to show 1.5‐Adam22. (g) Schematic showing extended family tree of ~1.5 MDa supercomplexes that contain PSD95 and their relative abundance in the mouse forebrain. The 1.5‐PSD95 was divided into 1.5‐NR and 1.5‐Non‐NR subpopulations. PSD95 is 17‐fold more abundant than GluN1. Since ~50% N‐methyl D‐aspartic acid receptors (NMDARs) interact with PSD95 (Frank et al. 2016), 1.5‐Non‐NR is 34‐fold more abundant than 1.5‐NR (ratio indicated in blue). Each subpopulation was further subdivided into those containing Kir2.3, IQseq2, Adam22 and Arc. The distribution of 1.5‐Kir2.3, 1.5‐IQsec2, 1.5‐Adam22 and 1.5‐Arc (expressed as a ratio in blue) between 1.5‐NR and 1.5‐Non‐NR were estimated by densitometry of supercomplexes immunodepleted with GluN2B and PSD95‐EGFP respectively (see Fig. 3c–e).
Figure 4
Figure 4
Neuroanatomically restricted assembly of Kir2.3‐N‐methyl D‐aspartic acid receptors (NMDAR) ion channel‐channel supercomplexes. (a) Regional enrichment of Kir2.3 expression in mouse brain. Left, mouse brain sagittal sections (lower) stained with Kir2.3 antibodies and (upper) reference Nissl stain (Allen Brain Atlas). High expression in rostroventral midbrain and caudodorsal forebrain shown and boxed regions were dissected for TAP‐purification of NMDAR. White arrow indicates boxed region of piriform cortex used in high‐magnification images in Fig. 4b and c. Scale bar, 1 mm. Right, BNP immunoblots of TAP‐purified receptors from dissected brain regions of Glun1 TAP / TAP mice as indicated. Left immunoblot probed with Flag (TAP‐GluN1) and right immunoblot probed with Kir2.3 showing 1.5‐Kir2.3‐NMDAR ion channel‐channel supercomplex enriched in rostroventral midbrain. Filled arrow, 1.5‐Kir2.3/NR. Molecular weight in MDa shown on right. IB, immunoblotting antibody. (b) Kir2.3 localization in layers of piriform cortex of wildtype (WT) and Psd95 −/− mice, including layers 1a (L1a) and 2/3 (L2/3) of piriform cortex stained with antibodies to Kir2.3 (green) and nuclear stain (DAPI, blue). White boxes indicate regions further magnified in Fig. 4c. Scale bar, 6 μm. (c) Kir2.3 localization requires Psd95. Higher magnification of boxed regions in Fig. 4b shows synaptic localization of Kir2.3 is disrupted in Psd95−/− mice. Sections double‐stained with Kir2.3 (green, top) and pre‐synaptic marker synapsin1 (red, middle) antibodies and merged image (bottom). White arrowheads show large Kir2.3 aggregates in Psd95 −/− mice. Scale bar, 4 μm. Right, histograms quantifying changes in puncta size (upper graph) and density (lower graph) of piriform cortex Kir2.3 quantified from triplicate experiments of Psd95 −/− and WT sections. Error bar, 1 SD. **p ≤ 0.01; ***p ≤ 0.001.
See this image and copyright information in PMC

Comment in

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Broadhead M. J., Horrocks M. H., Zhu F. et al (2016) PSD95 nanoclusters are postsynaptic building blocks in hippocampus circuits. Sci. Rep. 6, 1–14. - PMC - PubMed
    1. Cheng D., Hoogenraad C. C., Rush J. et al (2006) Relative and absolute quantification of postsynaptic density proteome isolated from rat forebrain and cerebellum. Mol. Cell Proteomics 5, 1158–1170. - PubMed
    1. Christopherson K. S., Sweeney N. T., Craven S. E., Kang R., El‐Husseini A. E.‐D. and Bredt D. S. (2003) Lipid‐ and protein‐mediated multimerization of PSD‐95: implications for receptor clustering and assembly of synaptic protein networks. J. Cell Sci. 116, 3213–3219. - PubMed
    1. Emes R. D. and Grant S. G. N. (2012) Evolution of synapse complexity and diversity. Annu. Rev. Neurosci. 35, 111–131. - PubMed
    1. Emes R. D., Pocklington A. J., Anderson C. N. G. et al (2008) Evolutionary expansion and anatomical specialization of synapse proteome complexity. Nat. Neurosci. 11, 799–806. - PMC - PubMed

Publication types

Substances

Associated data

LinkOut - more resources