SEARCH

SEARCH BY CITATION

Keywords:

  • detergent-resistant membranes;
  • membrane rafts;
  • post-synaptic density;
  • PSD ;
  • type I synapse;
  • type II synapse

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgments and conflict of interest disclosure
  8. References
  9. Supporting Information
Thumbnail image of graphical abstract

We systematically investigated the purification process of post-synaptic density (PSD) and post-synaptic membrane rafts (PSRs) from the rat forebrain synaptic plasma membranes by examining the components and the structures of the materials obtained after the treatment of synaptic plasma membranes with TX-100, n-octyl β-d-glucoside (OG) or 3-([3-cholamidopropyl]dimethylammonio)-2-hydroxy-1-propanesulfonate (CHAPSO). These three detergents exhibited distinct separation profiles for the synaptic subdomains. Type I and type II PSD proteins displayed mutually exclusive distribution. After TX-100 treatment, type I PSD was recovered in two fractions: a pellet and an insoluble fraction 8, which contained partially broken PSD-PSR complexes. Conventional PSD was suggested to be a mixture of these two PSD pools and did not contain type II PSD. An association of type I PSD with PSRs was identified in the TX-100 treatment, and those with type II PSD in the OG and CHAPSO treatments. An association of GABA receptors with gephyrin was easily dissociated. OG at a high concentration solubilized the type I PSD proteins. CHAPSO treatment resulted in a variety of distinct fractions, which contained certain novel structures. Two different pools of GluA, either PSD or possibly raft-associated, were identified in the OG and CHAPSO treatments. These results are useful in advancing our understanding of the structural organization of synapses at the molecular level.

We systematically investigated the purification process of post-synaptic density (PSD) and synaptic membrane rafts by examining the structures obtained after treatment of the SPMs with TX-100, n-octyl β-d-glucoside or CHAPSO. Differential distribution of type I and type II PSD, synaptic membrane rafts, and other novel subdomains in the SPM give clues to understand the structural organization of synapses at the molecular level.

Abbreviations used
CaMKII

Ca2+/calmodulin-dependent protein kinase II

CaMKIIα

α subunit of CaMKII

CHAPSO (also abbreviated as CH in the fraction names)

3-([3-cholamidopropyl]dimethylammonio)-2-hydroxy-1-propanesulfonate

cvPSD

conventional PSD

DRM

detergent-resistant membrane

GABAR

GABA receptor

GluA2/3

glutamate receptor subunits 2/3

GM1 ganglioside

monosialotetrahexosylganglioside

IAA

iodoacetamide

IS

insoluble fraction

MβCD

methyl-beta-cyclodextrin

OG

n-octyl β-d-glucoside

ppt

pellet

PSD

post-synaptic density

SDG

sucrose density gradient

SPM

synaptic plasma membrane

S

soluble fraction

TX

TX-100

VDAC

voltage-dependent anion selective channel protein

 

Abbreviated fraction names are composed of the detergent concentration, detergent name and fraction number

The method of purifying post-synaptic density (PSD), which comprises mainly type I excitatory synapses, was established in the late 1970s. Analyses of the PSD components have served as an initial step for understanding the synaptic transmission and synaptic plasticity. In 1997, the concept of lipid rafts, later also called membrane rafts (Pike 2004), was proposed (Simons and Ikonen 1997). Both PSD and membrane rafts are detergent-insoluble at 4°C, although isolated rafts are artificial aggregates of individual membrane raft domains in vivo. In synaptic regions, the two types of Triton X-100 (TX-100)-insoluble materials, PSD and membrane rafts, were separable by sucrose density gradient (SDG) ultracentrifugation, owing to the difference in their density (Suzuki et al. 2001). However, we found PSD–membrane raft complexes in detergent-resistant membranes (DRMs), after the treatment of synaptic plasma membranes (SPMs) with a relatively low concentration of TX-100 (Suzuki et al. 2011). We also found that n-octyl β-d-glucoside (OG) treatment of SPMs completely separated the PSD and DRMs, and produced PSD-free DRMs (Liu et al. 2013). These findings suggest that solubilization of the synaptic subcomponents associated with SPMs varies depending on the conditions of detergent treatment, such as the detergent types and detergent concentration or detergent–protein ratio. Furthermore, brain synaptic fractions, such as synaptosomes and the SPMs, which are sources for the isolation of DRMs and PSDs, are heterogeneous, containing both type I and type II synapses. However, systematic analyses isolating and identifying the various synaptic subdomains, such as the DRMs and different types of PSDs, have not yet been reported, and the relationship and interaction between these synaptic subdomains as well as the organization of the functional synaptic architecture at the molecular level are not well known. In fact, there is only limited information on the isolation of type II PSD.

In this study, we systematically investigated different types of PSDs and DRMs (isolated raft domains) by monitoring the distribution of type I and type II PSD markers on SDG after the treatment of SPMs with various concentrations of three different detergents: TX-100, OG and 3-([3-cholamidopropyl]dimethylammonio)-2-hydroxy-1-propanesulfonate (CHAPSO). Clear differences in the isolation of synaptic subdomains were found among the three. The results afford important insight into the synaptic architecture at the molecular level.

Materials and methods

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

Materials

TX-100, iodoacetamide (IAA), and ImmunoStar LD were purchased from WAKO Pure Chemical Industries. Ltd. (Osaka, Japan); OG was from Dojindo Laboratories (Kumamoto, Japan); CHAPSO was from Sigma-Aldrich (St Louis, MO, USA); anti-GABA receptor (GABAR) Aα1 and Aα2 were from AVIVA Systems Biology (San Diego, CA, USA); anti-chapsyn-110/PSD-93 antibody was from Thermo Scientific (Rockford, IL, USA); anti-Ca2+/calmodulin-dependent protein kinase II (CaMKII) was from BD Transduction Laboratories (Lexington, KY, USA); anti-HSP70 (ΚΟ4), which recognizes both p72 Hsp70 and p73 Hsc70, was produced as described previously (Suzuki et al. 1999). Affinity-purified anti-gephyrin, anti-synArfGEF/BRAG3, and anti-collybistin antibodies were produced in a rabbit using full-length Glutathione S-transferase (GST)-fused mouse gephyrin (Accession no. NM_145965.2), N-terminal 293 a.a. of rat synArfGEF (Accession no. AB057643) and GST-fused C-terminal 426–515 a.a. of mouse collybistin (ARHGEF9) (Accession no. BC141385), respectively. The anti-CaMKII antibody (BD Transduction Laboratories) recognizes both α and β subunits, but the reactivity to the β subunit is stronger than α subunit. The sources of the other antibodies were described previously (Suzuki et al. 2008).

Preparation of SPMs, PSDs, and DRMs

Animals were handled according to the Regulations for Animal Experimentation of Shinshu University, which were approved by the Committee for Animal Experiments of Shinshu University. The animal protocol was approved by the Committee for Animal Experiments of Shinshu University (Approval Number 240066). Based on the national regulations and guidelines, all experimental procedures were reviewed by the Committee for Animal Experiments and finally approved by the president of Shinshu University. SPMs were prepared from Wistar rats (male, 6 weeks old, specific pathogen-free) (Japan SLC, Inc., Hamamatsu, Japan), essentially as described previously (Suzuki 2011). To prevent artificial oxidation during the preparation of SPMs, buffers A and B were supplemented with 2 mM IAA unless stated otherwise. Purified SPMs were stored unfrozen in buffers containing 50% glycerol at −30°C. Longer storage at −80°C did not cause any deterioration of the samples in this experiment. DRMs were isolated from SPMs (500 μg protein/1.75 mL) after treatment with various detergents in 20 mM Tris–HCl (pH 7.4) containing 150 mM NaCl and 1 mM EDTA (TNE buffer), and SDG centrifugation was performed as described previously (Du et al. 2006; Suzuki et al. 2011); All detergent treatments were supplemented with protein inhibitor mixtures (p8340; Sigma-Aldrich) at 1/200 dilution along with IAA (2 mM). In some cases, after the detergent treatment and prior to the SDG centrifugation, samples were separated into the soluble and pellet fractions by ultracentrifugation at 87 700 gmax for 48 min. The soluble fractions and pellets that were hand-homogenized in 1.75 mL of TNE buffer were mixed with an equal volume of TNE containing 80% sucrose, and poured into centrifuge tubes. The solutions were overlaid with TNE buffers containing 30% sucrose and then 5% sucrose, and centrifuged at 256 000 gmax for 30 h. Eleven fractions were collected and the pellets (fraction 12) were suspended with 955 μL of TNE buffer.

Analyses of protein, GM1 ganglioside, and western blotting

For protein profiling, each SDG fraction (20 μL unless stated otherwise) was separated via sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and the gels were stained with silver. Detection of monosialotetrahexosylganglioside (GM1 ganglioside) was described previously (Suzuki et al. 2011). Western blotting was carried out using a chemiluminescent substrate and visualized with a CCD video camera system (Densitograph Lumino-CCD, ATTO Corporation, Tokyo, Japan).

Other methods

Electron microscopy was carried out as described previously (Suzuki et al. 2011). Nano-W (Methylamine Tungstate) (Nanoprobes, Inc., Yaphank, NY, USA) was used for negative staining. The protein bands stained with silver were cut, de-stained, reduced with dithiothreitol, alkylated with IAA and digested in-gel with trypsin. The resultant peptide mixtures were extracted and analyzed by mass spectrometry using rat UniProtKB database and IDENTITYE, which consists of nanoACQUITY, Xevo QTof MS and ProteinLynx Global SERVER (PLGS) 2.5.2. (Nihon Waters K.K., Tokyo, Japan).

Results

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

We investigated the distribution profile of PSD proteins on the SDG after the treatment of SPMs with varying concentrations of the three different detergents, TX-100, OG or CHAPSO, which are nonionic or zwitterionic detergents. A summary flowchart of this study is provided in Figure S1. TX and OG, but not CHAPSO, have been used for PSD purification (Suzuki 2011). The detergent concentration was changed gradually to observe the process of solubilization. The distribution profile of both type I and type II PSD proteins on a SDG was examined by western blotting. We examined GABAR Aα1, GABAR Aα2, gephyrin, and collybistin as type II PSD protein markers, and α-amino-3-hydroxy-5-methylisoxazole-4-propionate-type glutamate receptors (GluRs) [glutamate receptor subunits 1 (GluA1) and 2/3 (GluA2/3)], NMDA-type GluR 1 (GluN1) and 2B (GluN2B), PSD-95, chapsyn-110/PSD-93 and CaMKII, as representative type I PSD protein markers, respectively. SynArfGEF and HSP70 were also used as markers for both type I and type II PSDs.

Treatment of SPMs with TX-100

Untreated SPMs were distributed to fraction 8. TX treatment left relatively intact DRMs in fractions 5 and 6 when using the concentration of 0.05% and 0.15%, judging from the restricted distribution of the DRM markers, flotillin-1, and GM1 ganglioside (Figs 1 and 2). The DRMs obtained after 0.15% TX-100 treatment were relatively undamaged, as reported previously (Suzuki et al. 2011). Treatment with TX-100 at doses higher than 0.15% dispersed these DRM markers to heavier fractions and reduced the protein content in the DRM fractions, which indicates the disruption of DRM integrity. Fractions 4–6 are typical DRM fractions, and fractions 8–11 are a mixture of soluble proteins (see the bovine serum albumin distribution in Fig. 2a) and the cytoskeleton. The peak fractions of DRM after treatment with 0.05% and 0.15% TX-100 was found to be 6 in this study carried out in the presence of IAA, whereas the peaks were 5 in our previous study (Suzuki et al. 2011), where the purification and treatment of SPMs were carried out in the absence of IAA. Therefore, we tested the effects of IAA (2 mM) during the purification and treatment of SPMs on SDG (Figure S2). It was confirmed that the shift of the peaks was because of the presence of IAA, which probably prevented cross-linking and conformational changes in the structures present in the DRMs.

image

Figure 1. Protein profiles on sucrose density gradient (SDG) after the treatment of synaptic plasma membranes (SPMs) with various concentrations of TX-100. SPMs were treated with various concentrations of TX-100 and the proteins subjected to SDG, with 12 fractions collected. (a) SDG profiles without prior ultracentrifugation. (b) SDG profiles with prior ultracentrifugation. The proteins in each fraction were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and visualized by silver staining. The fraction numbers and detergent concentrations are shown on top and on the left of each blot, respectively. Fractions 5–7 after 0.05% and 0.15% TX-100 treatment are shown in rectangles for easy tracing of the relatively unbroken TX-100-DRMs. In this and subsequent Figures, the numbers to the left of gel images indicate molecular weight in kDa.

Download figure to PowerPoint

image

Figure 2. Distribution profiles of the post-synaptic density (PSD) proteins on sucrose density gradient (SDG) after the treatment of synaptic plasma membranes (SPMs) with various concentrations of TX-100. SDG fractions were collected after the treatment of SPMs with various concentrations of TX-100 and the proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). (a) SDG profiles without prior ultracentrifugation. (b) SDG profiles with prior ultracentrifugation. The distribution patterns of typical type I and type II PSD proteins were analyzed by western blotting. Flotillin-1 and GM1 were used as detergent-resistant membrane microdomain (DRM) markers. Bovine serum albumin (BSA) (50 μg) was treated with 1% TX-100 followed by SDG analysis and silver staining. The fraction numbers and detergent concentrations are shown on top and on the left of each blot, respectively. Fractions 4–6 after 0.05%-0.15% TX-100 treatment are shown in rectangles for easy identification of the relatively unbroken TX-100-DRMs. Aα1 and Aα2 refer to GABA receptor (GABAR) Aα1 and Aα2, respectively. Flot1, geph, colly, and chap refer to flotilin-1, gephyrin, collybistin, and chapsyn-110/PSD-93, respectively. HSP70 is a protein family that contains Hsp70 (p72) and Hsc70 (p73) (Suzuki et al. 1999).

Download figure to PowerPoint

To discriminate between soluble proteins and insoluble cytoskeletal proteins, both of which were co-localized in the same fractions 8–11 under the centrifugation conditions used (Fig. 2a), these two classes of proteins were separated into soluble (S) and insoluble (IS) fractions by ultracentrifugation prior to the SDG centrifugation step (Fig. 1b). By this prior separation of the soluble and insoluble proteins, it was found that a lot of the proteins had accumulated in the insoluble fraction 8 after 1.0% TX-100 treatment (1% TX-IS-8, similar designations to follow throughout). Much of the protein was lost from TX-IS-8 after 5% TX-100 treatment. Part of it may have been solubilized and the rest, possibly type I PSD proteins, may have shifted into the IS-12 fraction. The lack of proteins in the IS-12 may have been caused by heavy compaction and aggregation of the PSDs during the first centrifugation, which prevented proteins from penetration into the separation gel, even after boiling the sample in the SDS-PAGE sample buffer. This explanation was supported by the electron microscopic observation of a number of PSDs in this fraction (data not shown). The apparent lack of proteins in the 0.15% TX-100-DRMs after prior centrifugation (0.15% TX-IS-6) may be because of the progress of solubilization during the additional centrifugation and shift of the protein complexes into fraction 7.

The western blotting in Fig. 2 indicates the distribution pattern of various PSD proteins on the sucrose gradient after the treatment of SPMs with TX-100. There were no typical type II PSD proteins in the DRMs (fractions 4–6) at any concentration of TX-100. In contrast, typical type I PSD proteins were distributed in the DRMs after treatment with 0.05–1.0% of TX-100, as expected from our previous finding of the presence of PSD-raft complexes in the TX-DRM (Suzuki et al. 2011). Most of the type I PSD proteins were recovered in the PSD-containing pellets (Suzuki et al. 2011) after treatment with > 0.5% TX-100. Unexpectedly, a portion of GluA1 and GluA2/3 was solubilized and their recovery from the PSD-containing pellets was low after treatment with 1.0 and 5.0% TX-100. The distribution profiles of synArfGEF and HSP70 appeared to be a mixture of type-I and type II PSD proteins.

In the western blotting of the S and ISs followed by SDG (Fig. 2b), all of the type II PSD marker proteins were completely solubilized into fractions 8–11 even with 0.05% TX-100. In contrast, most of the type I PSD markers remained insoluble, even after treatment with a higher concentration of TX-100. A portion of the GluAs and CaMKII was solubilized. Insoluble type I PSD proteins after 1% TX-100 treatment peaked in fraction 8 (1% TX-IS-8). Flotillin-1 was also associated with this fraction, while approximately the same amount was solubilized. The IS-12 fraction contained no protein bands in the gel, probably because of heavy PSD compaction and aggregation.

We next compared the protein components in the PSD protein-enriched fractions with those of conventional PSD (cvPSD), which was purified from rat forebrains by the ‘short’ method of the Siekevitz group (Suzuki 2011) (Fig. 3). The protein components of 1% TX-IS-8 were similar to TX-12 PSD in being enriched with CaMKIIα and actin. The protein profiles of these PSD protein-enriched fractions, TX-IS-8 and TX-12, were nearly the same as cvPSD except for proteins in the 30–45 kDa range, although the content of these proteins in the cvPSD varied (data not shown). These proteins were identified by mass spectrometry as creatine kinase (CK) (Q5BJT9), voltage-dependent anion selective channel protein (VDAC) 1 and 3 (Q9Z2L0 and Q9R1Z0, respectively) and ADP/ATP translocase (AAT) (Q09073, Q05962, and Q80WBI8), as shown in Fig. 3(a). The western blotting in Fig. 3(b) shows the content of typical PSD proteins in the TX-12 and cvPSD. Type II PSD marker proteins were absent in both TX-12 and cvPSD, even under greatly enhanced immunostaining, and the contents of GluA1, GluNs, and CaMKIIα/β were approximately the same in the two fractions. There were differences in the content of flotillin-1, GluA2/3, PSD-95, chapsyn-110/PSD-93, and HSP70. In particular, GluA2/3 and PSD-95 were extremely concentrated in cvPSD, although they were more abundant in TX-12 than SPMs. Chapsyn-110/PSD-93 was also more highly concentrated in the cvPSD than TX-12, but the difference was not substantial compared with PSD-95.

image

Figure 3. Comparison of the various post-synaptic density (PSD)-enriched fractions. (a) Protein composition. Proteins in 1% TX-IS-8, 1% TX-12, and 5% TX-12 (20 μL each) and 0.55 μg protein of cvPSD were separated on sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and stained with silver. The major proteins in the cvPSD are indicated on the right. CK, VDAC, and AAT refer to creatine kinase, voltage-dependent anion selective channel protein and ADP/ATP translocase, respectively. (b) Western blotting. Content of the typical PSD proteins, both type I and type II, in the various PSD-enriched fractions were examined by western blotting. In the blots marked with asterisks, 0.15% TX-10 was used as a positive control (P. C.) instead of synaptic plasma membranes (SPMs), because protein bands were not detected because of the small amount in the SPMs. PSD-95 bands were visualized under two different exposure and contrast conditions using the same blots. PSD-95(A) shows the extent of the enrichment of the protein in the 1% TX12 and cvPSD compared with SPMs, while PSD-95(B) exhibits different content between the 1% TX12 and cvPSD. The signals in PSD-95(B) are in a linear range. (c) Silver staining of proteins in the western blots. Each lane contains the same amount of protein, which was verified by densitometry. The protein abbreviations are the same as those in Fig. 2. MW refers to molecular weight.

Download figure to PowerPoint

Electron microscopic observation of TX-DRMs, particularly the presence of the PSD-post-synaptic membrane rafts (PSR) complexes in the 0.15% TX-DRMs, was reported in our previous studies (Suzuki et al. 2011). In this study, fractions that were not previously reported were examined. Judging from the protein profiles on the SDG, the IS-8~11 obtained after treatment with a very low concentration of TX-100 was expected to contain structures that were easily released from the SPMs. The 0.05% TX-IS-8~10 contained horseshoe-like (200–250 nm in width) and ring structures (approx. 100 nm in diameter) (double arrows and arrowheads, respectively, in Fig. 4a). The origin of these structures is not known. The horseshoe-like structure, although similar, appeared to not be type I PSDs, because there was a vacant space in the center and an absence of PSD proteins in the 0.05% TX-IS-8~10 (Fig. 2b). The presence of the PSD structure in the 1% TX-IS-8, which was expected from protein profile (Fig. 1b) and western blotting (Fig. 2b), was confirmed by electron microscopy (Fig. 4b). The structure contained in this fraction may be made up of partially broken PSD–PSR complexes, since it contained flotillin-1 (Fig. 2b) and membrane components (asterisks in Fig. 4b). The PSD structures contained in fraction 12 are almost free of raft domains, judging from the result that this fraction lacked both membranes (Suzuki et al. 2011; Figure S3a) and flotillin-1 (Fig. 2). The flotillin-1 in the 1% TX-IS-8 may be solubilized in the 5% TX-100 (Fig. 2a and b) and the PSD separated from flotillin-1 may be precipitated in the pellet.

image

Figure 4. Electron microscopic observation of the fractions obtained by TX-100 treatments of the synaptic plasma membranes (SPMs). Typical micrographs of 0.05% TX-IS-8~10 (a) and1% TX-IS-8 (b) are shown. Some of the horseshoe-like structures, post-synaptic density (PSD)s, ring-like structures and membrane structures are indicated by the double arrows, arrows, arrowheads and asterisks, respectively. The insert in (b) is a magnified micrograph of the membrane-containing structure inside a rectangle.

Download figure to PowerPoint

Treatment of the SPMs with OG

Figure 5 shows the protein profiles obtained after treatment with OG. Intact OG-DRMs were distributed mostly between fractions 4 and 6, and 5.0% OG treatment destroyed the DRMs, judging from the distribution of GM1 ganglioside. Unexpectedly, these OG-DRMs were recovered in the soluble fraction (Fig. 5b). The protein amount in the PSD-containing pellet fraction was highest in the 0.75% OG treatment, and was relatively decreased in the 1.0% and 5.0% OG treatments. In contrast to the treatment with TX-100, OG solubilized most of the SPM proteins except for actin, as seen in OG-IS-11 (Fig. 5b). The lack of proteins in the pellet of insoluble material (OG-IS-12) may be because of heavy compaction and aggregation of PSDs during the first ultracentrifugation.

image

Figure 5. Protein profiles on sucrose density gradient (SDG) after the treatment of synaptic plasma membranes (SPMs) with varied concentrations of n-octyl β-d-glucoside (OG). (a) SDG profiles without prior ultracentrifugation. (b) SDG profiles with prior ultracentrifugation. The methods and descriptions are the same as those in Fig. 1 legend except for the usage of OG. Fractions 4–6 after 0.75% and 1.0% OG treatment are shown in rectangles for easy identification of the relatively unbroken OG-detergent-resistant membrane microdomains (DRMs). The arrowheads in (b) indicate actin, identified by mass spectrometry.

Download figure to PowerPoint

Western blot analyses of the OG-treated samples (Fig. 6) revealed a PSD protein distribution pattern that was different from that after TX-100 treatment. 0.75% OG treatment retained portions of the type II PSD markers (GABAR Aα1, Aα2, and collybistin) in the DRM fractions. After treatment with OG higher than 0.75%, all of these proteins were distributed to fractions 8–11, in which they were present in the soluble fractions (Fig. 6b). They were not present in the type I PSD-containing pellets.

image

Figure 6. Distribution profiles of the post-synaptic density (PSD) proteins on sucrose density gradient (SDG) after the treatment of synaptic plasma membranes (SPMs) with varied concentrations of n-octyl β-d-glucoside (OG). (a) SDG profiles without prior ultracentrifugation. (b) SDG profiles with prior ultracentrifugation. The methods and descriptions are the same as in Fig. 2 legend except for the usage of OG. Fractions 4–6 after 0.5%, 0.75% and 1.0% OG treatment are shown inside rectangles for easy identification of the relatively unbroken OG-detergent-resistant membrane microdomains (DRMs).

Download figure to PowerPoint

The extent of the recovery of type I PSD proteins in fraction 12 varied among the PSD proteins. The recovery of the type I PSD scaffolding proteins was lower than after the 0.75 and 5.0% OG treatments by unknown reason. Interestingly, the distribution pattern was different between the GluAs and the GluNs: GluA1 and GluA2/3 were distributed in both the 0.75% and 1.0% OG-DRMs and OG-pellets, whereas GluN1 and 2B were not present in the DRMs. Recovery of various GluRs in the pellet was highest after 0.75% OG treatment and decreased when the OG concentration was increased further. Type I PSD scaffolding protein PSD-95, chapsyn-110/PSD-93 and CaMKII were not detected in OG-DRMs, similar to the GluNs. They were present in the pellet after treatment with OG greater than 0.75%. The distribution profile of synArfGEF and HSP70 was a mixture of type I and type II PSD proteins. The solubilization of type I PSD proteins by relatively higher concentration of OG was confirmed and is shown in the soluble fraction in Fig. 6b. The capacity to solubilize type I PSD proteins is an interesting characteristic of OG.

In contrast to the type I PSD proteins, portions of the type II PSD proteins, such as GABAR Aα1, Aα2, and collybistin, were distributed to the OG-DRM fractions, typically in 0.75% OG-5. Gephyrin, a scaffolding protein for type II PSD (Choquet and Triller 2003), was not co-distributed to the OG-DRM fractions. Pre-treatment of SPMs with methyl-β-cyclodextrin (MβCD), a cholesterol-depleting agent (Christian et al. 1997), disrupted the DRM localization of these type II PSD proteins as well as raft markers (Figure S4a and c).

The OG-DRM and OG-pellet fractions were examined under electron microscopy and representative micrographs are shown in Figure S3b and c. OG-DRMs contained mostly membrane sacks with a smooth surface, as reported in our previous work (Liu et al. 2013). However, a few membranes were associated with small structures which were either an aggregation of small vesicles or non-vesicular/non-membranous structures (arrowheads in Figure S3b). OG-pellets had abundant PSD structures and an absence of membrane structures, as reported previously (Liu et al. 2013) (see also Figure S3c).

Treatment of SPMs with CHAPSO

CHAPSO is a zwitterionic detergent. The presence of both positive and negative charges in a molecule renders the molecule neutrally charged overall and functions as a mild detergent at neutral pH. It has been reported that CHAPSO treatment of brain tissue leaves the DRMs, which are highly cholesterol enriched and comprised of different components from TX-100-DRMs (Williamson et al. 2010), a characteristic which led us to test CHAPSO in this study.

CHAPSO treatment resulted in DRMs, although they displayed a relatively wider distribution pattern on the SDG fractions, judging from the distribution of GM1 ganglioside (Fig. 7a). However, the SDG of the CHAPSO-S and IS materials revealed a restricted localization of DRMs in fraction 5 (1% CH-IS-5), or 4 and 5 (2% CH-IS-4, 5) (Fig. 7b). Thus, CHAPSO-DRMs were recovered in a soluble fraction, similar to OG-DRMs.

image

Figure 7. Protein profiles on sucrose density gradient (SDG) after the treatment of synaptic plasma membranes (SPMs) with varied concentrations of 3-([3-cholamidopropyl]dimethylammonio)-2-hydroxy-1-propanesulfonate (CHAPSO). (a) SDG profiles without prior ultracentrifugation. (b) SDG profiles with prior ultracentrifugation. The methods and descriptions are the same as in Fig. 1 legend except for the usage of CHAPSO. Fractions 4–6, 5/6, or 4/5 after treatments with 0.5–2.0% CHAPSO are shown in rectangles for easy identification of the relatively unbroken CHAPSO-detergent-resistant membrane microdomains (DRMs). The arrowhead in (b) indicates actin, identified by mass spectrometry.

Download figure to PowerPoint

Treatment with CHAPSO resulted in a distinct IS-11. This is in evident contrast with 1% TX-IS-8. The contents of CH-IS-11 were inversely decreased in parallel with the concentration of CHAPSO. The 45-kDa protein, which was identified to be actin by mass spectrometry, was selectively retained in the 5% CH-IS-11, as seen in 0.75% and 1% OG-IS-11 (arrowhead in Fig. 7b).

Unexpectedly, protein was detected in fraction 12 from 1.0 to 5.0% CH-S, in which the protein components were the same. This was also observed in 1.0% and 5.0% TX-S-12 and 1.0% OG-S-12, but was difficult to observe without enhanced silver staining. The precipitation of proteins in the S-12 fraction may be because of the low value of the Svedberg constant for the structures containing these proteins.

The protein components of the CHAPSO-DRMs, CHAPSO-PSDs, CHAPSO-S, and CHAPSO-IS fractions were further investigated using enhanced silver staining (Fig. 8). Some of the proteins in the CHAPSO-DRMs were identified by mass spectrometry (Table S1). The CHAPSO-DRMs exhibited a protein composition distinct from the TX-100- and OG-DRMs, although it was somewhat similar to that of the OG-DRMs compared to the TX-DRMs. In particular, the contactin 1, VDAC, and 31-kDa protein(s) were enriched and CaMKIIα was absent in the former (Fig. 8a). 31-kDa protein(s) were abundant in CHAPSO-PSD, similar to OG-PSD (Fig. 8b). The CHAPSO-PSD components were also present in the 1% CH-S-12, but their contents were relatively low except for the 31-kDa protein(s) (Fig. 8c). A high content of actin in the 5% CH-IS-11 was confirmed in the enhanced silver staining using an increased amount of 5% CH-IS-11 (twofold the amount applied in Fig. 7b) (Fig. 8c).

image

Figure 8. Protein components of the 3-([3-cholamidopropyl]dimethylammonio)-2-hydroxy-1-propanesulfonate (CHAPSO) treatment-derived fractions. Proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and silver-stained. (a) Protein composition of the typical detergent-resistant membrane microdomain (DRM) fractions after CHAPSO treatment (1% CH-S-5 and 2% CH-S-4) and comparison with the typical DRMs obtained by treatments with other detergents [0.15% TX-6 and 1% n-octyl β-d-glucoside (OG)-S-4]. Staining conditions were adjusted lane to lane to allow the protein bands in each lane to be clearly seen. The proteins in 1% CH-S-5 were identified by mass spectrometry (see Table S1 for protein identification). Protein identification in the TX-DRMs and OG-DRMs (No. 1 to 15) are based on a previous study (see table 1 in Liu et al. 2013). The bands marked with protein names and (+) contain the indicated protein in the bands. α(−) indicates that the band did not contain the α subunit of CaMKII (CaMKIIα). (b) Comparison of the protein components of typical post-synaptic density (PSD)s prepared with TX-100, OG and CHAPSO (1% TX-12, 0.75% OG-12, and 2% CH-12, respectively). (c) Protein composition of the CHAPSO treatment-derived fractions other than DRMs and PSD proteins (CH-S-12, CH-IS-11). The proteins indicated in (b) and (c) were identified by mass spectrometry of the proteins in the TX-PSD. cont1, A, T, OxDH, chap, α-IN, α, β and VDAC refer to contactin 1, actin, tubulin, oxoglutarate dehydrogenase, chapsyn-110/PSD-93. α-internexin, CaMKIIα, CaMKIIβ and the voltage-dependent anion channel, respectively. The 31-kDa protein contains three major proteins (adenosine diphosphate/adenosine triphosphate (ADP/ATP) translocase, prohibitin, and 14-3-3 protein).

Download figure to PowerPoint

The CHAPSO treatment-derived fractions were also analyzed by western blotting (Fig. 9). The distribution patterns of the type II PSD proteins were similar to those after OG treatment, portions of them localizing in the DRM fractions after 0.75% and 1.0% CHAPSO treatment (Fig. 9a). The type II PSD markers that were widely distributed in fractions 8–11 were recovered in the soluble fraction (Fig. 9b). The distribution patterns of the type I PSD proteins were different from those after the TX-100 and OG treatments. In particular, a small number were localized in the DRMs (fraction 5) after 2% CHAPSO treatment, although this number was very small (Fig. 9a). Most of the type I PSD proteins were localized in fractions 7 and 8 after the treatment at and below 1%, but also in fraction 11 and the pellet after the treatment at 2.0 and 5.0% (Fig. 9a). Unexpectedly, many of the type I PSD proteins were more abundant in fraction 11 than fraction 12 (Fig. 9a). Flotillin-1, a typical raft marker, was also highly concentrated in fraction 11 at 2.0% and 5.0% CHAPSO (Fig. 9a). However, it was distributed in the soluble fraction and was not associated with the type I PSD proteins (Fig. 9b). GluAs exhibited two separate pools: one was associated with CHAPSO-DRMs, which contained a portion of the type II PSD proteins, whereas the other was associated with the type I PSD proteins. A portion of type II PSD proteins, such as GABAR Aα1, Aα2, and collybistin, but not gephyrin, were distributed in the 0.75% and 1% CH-5. These DRM distribution patterns were disrupted by pre-treatment of SPMs with MβCD, together with a disruption of raft markers (Figure S4).

image

Figure 9. Distribution profiles of the post-synaptic density (PSD) proteins on sucrose density gradient (SDG) after the treatment of synaptic plasma membranes (SPMs) with varied concentrations of 3-([3-cholamidop-ropyl]dimethylammonio)-2-hydroxy-1-propanesulfonate (CHAPSO). (a) SDG profiles without prior ultracentrifugation. (b) SDG profiles with prior ultracentrifugation. The methods are the same as in Fig. 2 legend except for the usage of CHAPSO. Fractions 4–6 after treatments with 0.5% to 2.0% CHAPSO are shown inside rectangles for easy identification of the relatively unbroken CHAPSO-detergent-resistant membrane microdomains (DRMs).

Download figure to PowerPoint

The various fractions obtained after CHAPSO treatment were examined under an electron microscope. The CH-DRM fractions mostly comprised membranes. However, non-membranous materials were also evident in the magnified images (arrow in Fig. 10a). This was confirmed and more clearly seen in negative stained samples (Fig. 10b). The non-membranous materials were less abundant in the membranes obtained after treatment with 2% CHAPSO (Fig. 10b), in parallel with the decrease in the protein content (Fig. 7). The presence of PSDs was shown in the CHAPSO pellets, both CH-12 and CH-IS-12 (Fig. 10c, Figure S3d). Circular structures (~ 100 nm in diameter) were evident in the 5% CH-IS-12 (asterisks in Fig. 10c). The origin and structural importance are undetermined at present.

image

Figure 10. Electron microscopic observation of 3-([3-cholamidopropyl]dimethylammonio)-2-hydroxy-1-propanesulfonate (CHAPSO)-detergent-resistant membrane microdomains (DRMs) and CHAPSO-post-synaptic density (PSD). (a and b) Typical micro-graphs of CHAPSO-DRMs. The non-membranous material is indicated by the arrow in (a). The micrographs in (b) are the result of negative staining. (c) Typical micrographs of CHAPSO-IS-PSD. PSDs are packed in the preparations. The circular structures are indicated by asterisks. See also Figure S3d for CH-12, which was essentially the same as CH-IS-12. The fraction names and concentration of CHAPSO used are indicated on the top of each photograph.

Download figure to PowerPoint

CHAPSO treatment at various concentrations resulted in distinct fractions. These fractions, other than the CHAPSO-DRMs and PSDs, were also examined under electron microscopy. A cloudy band appeared in the 5% CH-S-11 region after SDG centrifugation of the CHAPSO-S fraction. This band fraction was packed with fine fiber meshwork-like structures (Fig. 11a). The protein composition of this band is unknown because a number of solubilized proteins other than the band components were also present (Fig. 7b). The structures in the 5% CH-S-12 were only observable by negative staining because of the small amounts. This fraction contained small (< 200 nm) structures that differed from PSDs and other known subcellular materials (Fig. 11b). The other distinct fraction was CH-IS-11. PSDs (arrows in Fig. 11c) and other structures, which may be structures associated with or surrounding PSDs, were present. The amount of non-PSD structures was reduced by the treatment with 2% CHAPSO compared with 1%.

image

Figure 11. Electron microscopic observation of the novel fractions derived from 3-([3-cholamidopropyl]dimethylammonio)-2-hydroxy-1-propanesulfonate (CHAPSO)-treated synaptic plasma membranes (SPMs). Typical micrographs of the 5% CH-S-11Band (a), as well as 5% CH-S-12 (b) and CH-IS-11 (c). The photographs shown in (b) are the result of negative staining. Scale bars in (b) 100 nm. The concentrations of CHAPSO used are indicated at the top of the micrographs in (c).

Download figure to PowerPoint

Discussion

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

Systematic examination on the solubilization of SPMs has not been done before. The method used in this study enables a simultaneous monitoring of all the components, which is indispensable to elucidate the relationship between the PSDs and PSRs, because they are dynamically associated in vivo (Suzuki et al. 2011; Liu et al. 2013). The usage of different types of detergent confers an advantage in this study. In fact, the detergents used in this study exhibited different solubilization processes for DRMs and PSDs. Examination at varied detergent concentrations was necessary to unveil the whole process of solubilization. PSDs and synaptic DRMs have been previously purified using detergents only at fixed concentrations (Besshoh et al. 2005; Gil et al. 2006; Matsuura et al. 2007; Williamson et al. 2010) except for a very few cases (Suzuki et al. 2011).

Isolation of the type I PSDs

This study showed that cvPSD, as well as the other types of PSD-enriched pellets (fraction 12) obtained in this study, does not contain type II inhibitory PSDs (Fig. 3), which has not been clearly demonstrated previously. Comparison of PSD-enriched fraction 1% TX-12 and the cvPSD prepared from forebrain synaptosomes were somewhat different, in particular, in terms of the Glu2/3 and PSD-95 content (Fig. 3b). The most plausible explanation is a difference in the conditions of retrieving the PSD proteins: the cvPSD is recovered in the interface between 1.5 and 2.1 sucrose after centrifugation at 201 800 gav for 2 h, whereas TX-12 is the only one of the PSD-containing three fractions (6, 8 and 12) obtained after ultracentrifugation at 256 000 gmax for 30 h on 5–40% sucrose. The cvPSD may thus be a mixture of fractions 8 and 12. This may explain why cvPSD contains membrane raft proteins, such as VDACs (Fig. 8) (Jordan et al. 2004; Li et al. 2004). Also, because of the presence of membrane components, possibly membrane rafts, it is necessary to wash the crude cvPSD retrieved immediately after SDG centrifugation with KCl/TX-100 for the final purified cvPSD (Cohen et al. 1977; Suzuki 2011).

Distribution and association of type II PSD proteins with membrane rafts

Type I and type II PSD proteins exhibited mostly a segregated pattern of distribution. Type II PSD proteins were easily solubilized in most cases, while type I PSD proteins tended to remain in the ISs (Figs 2, 6 and 9). A mixed distribution of type I and type II PSD proteins in HSP70 and synArfGEF is in good agreement with reports that they are present in both type I and type II PSDs, although presence of synArfGEF in the type I PSD is not confirmed immunocytochemically (Suzuki et al. 1999; Inaba et al. 2004; Fukaya et al. 2011; Machado et al. 2011).

OG and CHAPSO treatment produced a GM1 ganglioside- and flotillin-1-enriched light fraction (DRMs), as TX-100 did. However, the OG- and CHAPSO-DRMs were different from the TX-DRMs in that the former were recovered only in the supernatant (Figs 5b and 7b). This may be because of the small extent of the association of the membrane raft domains with proteinaceous structures in the former DRMs, as suggested by electron microscopic observation (Fig. 10a and Figure S3b). In other words, the protein:lipid ratio in the OG- and CHAPSO-DRMs appears to be smaller than that in the TX-100-DRMs.

Both the type I and type II PSDs appear to be associated with membrane raft domains. Interestingly, the association of type I PSDs with the raft domains was detected only after TX-100 treatment, not in the OG- or CHAPSO-treated SPMs. In contrast, the association of type II PSD proteins with the membrane rafts was identified after OG or CHAPSO treatment, but not after TX-100 treatment. This may be because of the difference in the properties of the detergents.

The codistribution of type II PSD proteins, such as GABARs and collybistin, RhoGEF for Cdc42 (Fritschy et al. 2008), with DRM, was observed only partly (0.75% OG-5, 0.75% CH-5 & 6, and 1% CH-5), while most of the type II PSD proteins were recovered in the soluble fractions (fractions 8–11). In contrast, gephyrin, a scaffolding protein for type II PSD (Choquet and Triller 2003), was always present in the soluble fractions, and did not colocalize with GABAR Aα1 or Aα2 in the DRMs. This is rather unexpected, because scaffolding proteins, in general, form a stable complex with synaptic receptors. This study suggests that the association of type II PSDs with membrane rafts may be weaker than that of type I PSDs and thus unstable in the presence of detergent, which may be related, at least partly, to the weak association of the GABARs with their scaffolding protein gephyrin. The weak association observed in this study may not be because of IAA, which prevents the formation of artificial disulfide linkages during preparation and detergent treatment of SPMs (Figure S2b).

MβCD treatment of SPMs prior to detergent treatment (OG and CHAPSO) eliminated these type II PSD proteins from DRMs (Figure S4). This result suggests a raft association of the type II PSD proteins. Redistribution of the type I PSD proteins to the peak raft marker fraction after MβCD pre-treatment and subsequent detergent treatment may be because of an artificial aggregation of membrane proteins rather than a linkage between type I and type II PSD structures, although this remains to be determined.

The idea that type II as well as type I PSDs (Suzuki et al. 2011) are associated with raft domains is supported by previous reports (Li et al. 2007; Suzuki and Yao 2014). However, as yet we have been unable to identify the type II PSD–PSR complex at the electron microscopic level, although an association of non-membranous structures with the membranes was observed in the OG- and CHAPSO-DRM fractions (Fig. 10a,b and Figure S3b). The thin structure of type II PSD may make identification of this complex difficult. Furthermore, type II PSD structures were disorganized, because gephyrin, a scaffolding protein for the type II inhibitory PSDs, was mostly solubilized.

Novel fractions containing unique subsynaptic structures

Figure S5 summarizes the characteristic fractions identified in this study. Among them, in particular, some of the CHAPSO treatment-derived fractions are distinct. CHAPSO produced DRMs in which the protein composition was different from that in the TX- and OG-DRMs (Fig. 8a). The origin and characteristics of the fine fiber meshwork-like structures found in the cloudy band produced in 5% CH-S-11 (Fig. 11a), and the structures identified in 5% CH-S-12 were not elucidated in this study. The structures and protein composition present in the CH-IS-11 changed depending on the concentration of CHAPSO. This fraction is one of the major sites in which PSD accumulates after treatment with CHAPSO. It is interesting that the type I PSD proteins were potently solubilized by CHAPSO and OG. This is in clear contrast with TX-100. These fractions await further investigation.

In the treatment with OG and CHAPSO, but not TX-100, the GluA distribution pattern exhibited a different profile compared with the other type I PSD proteins (Figs 6 and 9). Both the GluA1 and 2/3 subunits were distributed in the DRMs in a manner distinct from the other type I PSD proteins (0.75%, 1% OG) after treatment with OG (Fig. 6). A small portion of the GluA subunits was also distributed in the CHAPSO-DRMs along with the other type I PSD proteins (Fig. 9a). The separation of GluA from the other type I PSD proteins became clearer in the SDG analysis after prior separation into a soluble fraction and pellet (Fig. 9b). Thus, this study enabled a visualization of two pools of GluA in the synaptic region: PSD-associated and non-associated pools. The GluA transporting vesicles involved in the exocytosis and endocytosis of GluA in the spines (Newpher and Ehlers 2008; Tao-Cheng et al. 2011) may comprise the latter pool and contain membrane raft domains (Hou et al. 2008). The DRM fractions containing GluA, together with other CHAPSO treatment-derived fractions, are interesting targets for future analyses.

Conclusions

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

This study showed that synaptic DRMs as well as type I and type II PSDs were separated and solubilized in a different manner by TX-100, OG, and CHAPSO. The findings obtained in this study are helpful toward an understanding of the structural organization of PSDs and PSRs. This study also identified fractions with novel subsynaptic structures that may prove useful in future analyses of synaptic organization.

Acknowledgments and conflict of interest disclosure

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

This work was funded by a Shinshu University Grant. The authors have no conflicts of interest to declare.

All experiments were conducted in compliance with the ARRIVE guidelines.

References

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

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgments and conflict of interest disclosure
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
jnc12807-sup-0001-TabS1-FigS1-S5.pdfapplication/PDF8788K

Table S1. List of the major proteins contained in the 1% CHAPSO-DRMs (1% CH-S-5) derived from forebrain SPMs.

Figure S1. Summary of the systematic examination of the synaptic subdomains.

Figure S2. Effects of IAA on the detergent treatment of SPMs and subsequent SDG.

Figure S3. Electron microscopic observation of the fractions obtained by TX-12, OG-DRMs (typically OG-5), OG-12, and CH-12 prepared from the SPMs.

Figure S4. Effects of MβCD treatment on the protein distribution of type II PSD marker proteins in the DRM fractions.

Figure S5. Summary of fractions of interest.

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.