Semiquantitative proteomic analysis of rat forebrain postsynaptic density fractions by mass spectrometry.

The postsynaptic density (PSD) of central excitatory synapses plays a key role in postsynaptic signal transduction and contains a high concentration of glutamate receptors and associated scaffold and signaling proteins. We report here a comprehensive analysis of purified PSD fractions by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS). We identified 374 different proteins that copurified with the PSD structure and discovered thirteen phosphorylated sites from eight proteins. These proteins were classified into numerous functional groups, implying that the signaling pathways in the PSD are complex and diverse. Furthermore, using quantitative mass spectrometry, we measured the molar concentration and relative stoichiometries of a number of glutamate receptor subunits and scaffold proteins in the postsynaptic density. Thus this proteomic study reveals crucial information about molecular abundance as well as molecular diversity in the PSD, and provides a basis for further studies on the molecular mechanisms of synaptic function and plasticity.


SUMMARY:
The postsynaptic density (PSD) of central excitatory synapses plays a key role in postsynaptic signal transduction and contains a high concentration of glutamate receptors and associated scaffold and signaling proteins. We report here a comprehensive analysis of purified PSD fractions by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS). We identified 374 different proteins that copurified with the PSD structure and discovered thirteen phosphorylated sites from eight proteins. These proteins were classified into numerous functional groups, implying that the signaling pathways in the PSD are complex and diverse. Furthermore, using quantitative mass spectrometry, we measured the molar concentration and relative stoichiometries of a number of glutamate receptor subunits and scaffold proteins in the postsynaptic density. Thus this proteomic study reveals crucial information about molecular abundance as well as molecular diversity in the PSD, and provides a basis for further studies on the molecular mechanisms of synaptic function and plasticity.
Proteomic Analysis of PSD INTRODUCTION: In excitatory synapses of the brain, the postsynaptic density (PSD) is a microdomain of the postsynaptic membrane specialized for signaling and plasticity (1)(2)(3). It contains neurotransmitter receptors (particularly for the transmitter glutamate), receptor-associated scaffold proteins, cytoskeletal elements and regulatory enzymes, which are assembled together in a disk-like structure, approximately 30-40 nm thick and a few hundred nm wide (4). Glutamate receptors in the PSD are organized into supramolecular signaling complexes by interacting with specific PDZ (PSD-95, Dlg, ZO-1 homology) domain-containing scaffolds (such as PSD-95/SAP90) and their associated proteins.
The molecular architecture of the PSD is specialized for signal transduction, but it is also highly modifiable to allow for strengthening and weakening of synaptic transmission (1)(2)(3). The function and composition of the PSD is dynamically regulated in response to neural activity, involving mechanisms such as protein phosphorylation (3), local translation (5), ubiquitination and degradation (6,7), and subcellular redistribution (8) In this study, we report the identification of 374 proteins in purified PSD fractions from rat forebrain. In addition to uncovering many previously unknown putative components of the The redundancy at the protein level was increased by the presence of different species origins and/or different names for the same proteins. In addition, one identified peptide might be shared within several proteins (e.g. protein isoforms). Thus we manually removed the redundant proteins and finally accepted 374 proteins (1,830 peptides) shown in supplementary Table S1 and Table S2. Among those peptides, there are 1,614 "unique peptides" whose amino acid sequences were found only in the assigned proteins, and 216 "shared peptides" whose sequences existed in one or more other proteins. We typically accepted proteins identified by at least one "unique peptide".

Functional categories of identified proteins in the PSD fraction
The identified proteins of the PSD fraction are shown in Table S1, grouped under functional categories in alphabetical order. To gauge the possible significance of the proteins identified, it is useful to estimate their relative abundance in the sample. Because of the high sensitivity of LC-MS/MS for protein identification, the actual molar quantity of the proteins detected may vary by several orders of magnitudes. The amplitude of MS signal detected for a specific peptide is affected by many factors, but especially its electrospray ionization efficiency, which is highly variable. Thus it is difficult to evaluate protein abundance directly from its peptide ion signal(s).
However, the number of times that peptides are identified for each protein is generally correlated with the abundance of the protein, if the protein size is normalized. Therefore a rough protein abundance index can be derived as described in the methods (36,37). This abundance index Proteomic Analysis of PSD ranges from 0.2 to 214.4 in the analysis of the PSD sample; the absolute number has little meaning, but relative values offer semi-quantitative information about relative abundance in the PSD.

Receptors and channels
A cardinal feature of the PSD is the high density of glutamate receptors. In our PSD preparation, all classes of glutamate receptors were identified by mass spectrometry: NMDA receptors (subunits NR1, NR2A, NR2B), AMPA receptors (GluR1, GluR2, GluR3), and metabotropic glutamate receptors (mGluR1, mGluR3, mGluR6) (Table S1). Based on the abundance index, the level of NR2B in the PSD is higher than NR2A (Table S1). Among AMPA receptor subunits, GluR2 appeared to be the most abundant (Table S1). GluR4 was not detected, consistent with the predominant expression of this subunit early in development. Kainate receptor subunits were weakly detected in the purified NMDA receptor complex by immunoblotting (17), but were not picked up in our MS screen, possibly because of low abundance. Similarly, we identified only one peptide from type I metabotropic glutamate receptors (mGluR1), consistent with the fact that these postsynaptic receptors reside largely outside of the PSD (38).
Several ion channel subunits were identified in the PSD fraction. In addition to L-type voltage-dependent calcium channels, we found the alpha-1e (pore forming) subunit of R-type voltage-dependent calcium channel, which has been invoked as the major voltage-sensitive calcium channel in dendritic spines (39). We did not detect the more abundant P/Q type (alpha-Proteomic Analysis of PSD 1A) and N-type (alpha-1B) calcium channels, which are believed to be primarily presynaptic.
Subunits of GABAB receptors (GABAR1 and GABAR2), but not GABAA receptors, were detected in the PSD preparation (Table S1). The non-detection of GABAA receptors by MS would argue against significant contamination of our PSD preparation with inhibitory synapses; however, we did find one peptide corresponding to the glycine receptor-binding protein gephyrin, which is concentrated in inhibitory synapses.

Scaffolds
Studies of the molecular organization of the PSD have enlightened our current understanding of PDZ domain-containing scaffold proteins and their role in the clustering of receptors and the assembly of signaling complexes (40). Members of the PSD-95 membrane-associated guanylate kinase (MAGUK) family were found abundantly in the PSD, as reflected in the frequent identification of peptides derived from PSD-95, PSD-93/chapsyn-110, and SAP102 (Table S1).
According to the abundance index, there appear to be higher levels of PSD-95 in the PSD than chapsyn-110/PSD-93 or SAP102 (Table S1). Moreover, PSD-95 family molecules seem to be much more abundant than NMDA receptors by this semi-quantitative analysis, in agreement with absolute measurements (see below). PSD-95 binds to GKAP/SAPAP, which in turn interacts with Shank (also known as ProSAP) and Homer (also known as Vesl) (3,40,41).
In contrast, we detected in the PSD fraction a paucity of scaffold proteins associated with AMPA-type glutamate receptors. Neither SAP97 (a PSD-95 family protein that binds GluR1 (43) nor GRIP/ABP (multi-PDZ proteins that binds directly to GluR2/3) (44,45) were detected by the LC-MS/MS analysis. A small number of peptides were identified corresponding to the liprin-α family of scaffold proteins that are implicated in AMPA receptor trafficking (46) and synapse development (47,48). It is possible that AMPA receptor-associated proteins are easily extracted from the PSD by detergents, or alternatively, their interaction with AMPA receptors is more involved in trafficking rather than postsynaptic anchoring (49,50).

GTPases and regulators
The MS analysis of the PSD found a striking diversity of proteins involved in small GTPase signaling. The monomeric GTPases identified included members of the Ras family (K-ras, Rap-1a, Ral-a, ras-related c3 botulinum toxin substrate homolog), Rho family (Rac2, Rho-G), and PIKE-L (a protein containing a Ras-like GTPase domain) ( Table S1). The small G-proteins were generally identified based on a single peptide hit, suggesting they are not abundant components of the PSD. However, this is perhaps not surprising given that small GTPases are generally recruited to specialized membrane microdomains (such as synapses) upon activation. More abundantly represented in the PSD than the small GTPases themselves were the guanine nucleotide exchange factors (GEFs) and GTPases activating proteins (GAPs) that regulate the activity of small G-proteins. Particularly prominent was SynGAP, a PSD-95associated RasGAP that was identified in earlier studies (17,51,52). We also identified a known PSD-enriched RapGAP (SPAR) (53), two ArfGAPa (GIT1, PIKE-L) (54-56), and a RacGEF known to regulate spine morphogenesis (Kalirin) (57). In addition, we detected two putative RhoGAPs (KIAA0097 and KIAA1688), two RhoGEF (clone:6230420N16 and zizimin1), a novel putative RapGAP (KIAA0440), a cAMP-regulated RapGEF (cAMP-GEFII), two putative ArfGEFs (KIAA0522 and KIAA0763), and a GAP for Rab3 (KIAA0856), which were previously not known to be present in the PSD. Despite the relative abundance of RasGAPs, we did not pick up any known RasGEFs in our MS screen. In addition to the small GTPases, heterotrimeric G-proteins were found in the PSD, with the G(o) alpha subunit being particularly abundant (Table S1).

Kinases/Phosphatases and Regulators
CaMKII is well-known to be highly enriched in the PSD, and the alpha and beta isoforms were abundantly identified (Table S1), in agreement with previous MS studies of PSD and NMDA receptor complex (14,17). We also detected CaMKII delta subunit and gamma. Protein kinase C Proteomic Analysis of PSD (PKC)-gamma is another abundant kinase identified in the PSD, consistent with the brainspecific expression of this subtype (58). However, there were notable "absences" from our list of identified kinases in the PSD. We did not find the MAP kinases ERK1/ERK2, or any tyrosine kinases (e.g. Src, PYK2), although they had been detected in the NMDA receptor complex by immunoblotting (17). Protein kinase A was also not identified, although we detected its anchoring protein AKAP150 repetitively (by 4 peptides). Given the lack of detectable MAP kinases, it was notable to find peptides corresponding to MINK-2 and TNIK in the PSD. These kinases belong to the Ste20 family of ser/thr kinases that typically function in MAP kinase cascades. In addition to protein kinases, several protein phosphatases were found in the PSD: PP1-alpha, PP1-gamma, r-PTP-zeta, and receptor-linked form P1 protein tyrosine phosphatase. Calcineurin/PP2B was not detected in the PSD by MS. In addition to the protein kinases/phosphatases, we found several lipid kinases and phosphatases involved in phosphoinositide metabolism (Table S1).

Cell adhesion
A surprising variety of cell-cell and cell-matrix interaction molecules were identified in PSD fraction. Cadherins are localized at synapses and believed to be important for adhesion of preand postsynaptic membranes and for synapse maturation (59). We detected cadherins-2, -10 and -13 in the PSD, of which cadherin-2 (better known as N-cadherin) was the most abundantly represented. A number of adhesion proteins of the immunoglobulin superfamily were identified Proteomic Analysis of PSD in the PSD, including NCAM-140, NCAM-L1, kilon, neurotrimin, OBCAM and contactin. We also detected extracellular matrix proteins (tenascin-R/janusin and tenascin-C), and proteoglycans (versican V0, brain enriched hyaluronan binding protein, chondroitin sulfate proteoglycan, neurocan core protein, proteoglycan link protein). Another abundant protein identified was densin-180, a well-established component of the PSD and a putative synaptic adhesion molecule (10). Rather surprisingly, the gap junction protein connexin 43 was robustly detected in the PSD. The diversity of PSD proteins involved in cell junction assembly is consistent with the idea that the synapse is a heterogeneous and sophisticated adhesive structure formed between presynaptic and postsynaptic membranes.

Actin cytoskeleton
As expected, the PSD was found to contain a large number of cytoskeletal proteins in great abundance, particularly those of the actin-based cytoskeleton. The largest number of peptide hits occurred for the non-erythrocyte alpha and beta spectrins, though erythrocyte beta-spectrin was also abundant. It has been reported that filamentous actin (F-actin) consisting of beta-and gamma-isoforms of actin is highly concentrated in dendritic spines (60). In our study, we found alpha-actin, beta-actin, and gamma actin in the PSD. Numerous actin binding or actin regulatory proteins were identified by LC-MS/MS, including ABLIM; adducin (alpha and beta); adducin-like protein; afadin; alpha-actinin-1, -2, and -4; ankyrin 1, 2 and 3; components of the Arp2/3 complex; F-actin capping proteins; cofilin; coronin (1b and 1c); cortactin; drebrin; We also identified many septin family proteins in purified PSDs, including septin, septin 2, septin 6, septin 7 and CDCrel-1A. Septins are self-polymerizing GTP-binding proteins that and function in cytokinesis and actin organization (61,62). The presence of a septin in the PSD was previously reported and confirmed by immunostaining (14); however, the functional Proteomic Analysis of PSD significance of septins in the PSD are unknown.

Motor proteins
An interesting class of proteins uncovered in the PSD was the molecular motors. Actin-based motors were most heavily represented, with the most abundant being myosin Va, and myosin heavy chain non-muscle type b. The PSD fraction also contained an atypical myosin with a PDZ domain in its N-terminal region (MysPDZ) (63). The only microtubule-based motor found was dynein (both heavy and light chains were detected).

Membrane trafficking
Consistent with dynamic membrane flux in the postsynaptic compartment, numerous proteins involved in trafficking were identified in the PSD. Molecules that mediate receptor endocytosis were especially abundantly represented: e.g. clathrin heavy chain; subunits of the clathrin adaptor complexes AP2 and AP3; and the GTPases dynamin-1 and -2. Many proteins involved in vesicle fusion were also identified, such as N-ethylmaleimide sensitive factor (NSF), synaptotagmin I, syntaxin 1, SNAP25, Munc-18, and Munc-13 (Table S1).

Translation
Long-term synaptic plasticity requires new protein synthesis, and mounting evidence points to local translation of specific mRNAs at postsynaptic sites (5). In the PSD fraction, LC-MS/MS Proteomic Analysis of PSD analysis identified multiple ribosomal proteins and several factors that regulate translation (e.g. elongation factor 1-alpha, elongation factor Tu). In this regard, it is interestingly that the putative RNA-binding protein Fragile X-related protein 1 was also detected in the PSD. Obviously, it is possible that the presence of translation-related proteins is due to non-specific contamination of the PSD (e.g. with ribosomes). However, it is now widely accepted that protein synthesis can occur in the postsynaptic compartment, so it is not implausible that the translational machinery could be physically linked to the PSD.
Presynaptic proteins are differentially extracted with Triton X-100 at pH 8.1 during biochemical purification of synaptic junctions (64). Although we used double Triton X-100 extraction at pH 8.1 to prepare our PSDs, we nevertheless detected well-known presynaptic cytomatrix proteins (piccolo and bassoon), and proteins associated with synaptic vesicles or exocytosis (synapsin I and II, synaptotagmin I and 7, SNAP-25, syntaxin 1) in our samples (see also previous section).
Their presence in the PSD fraction may reflect the high sensitivity of the MS technique and "contamination" of the PSD preparation by presynaptic structures. This would not be surprising, since it is believed that pre-and postsynaptic specializations are bridged by strong adhesive protein interactions (64). On the other hand, it cannot be excluded that some of these proteins also play a role postsynaptically in membrane fusion processes. Of additional note in this regard, several alpha-latrotoxin receptors, putative mediators of cell-matrix and cell-cell interactions in synaptic junctions, were found in the PSD fraction. These include alpha-latrotoxin receptor 3; Proteomic Analysis of PSD CL1AA; and neurexins.

Identification of phosphorylated proteins
The MS/MS spectrum can be utilized not only to identify proteins, but also to determine the precise amino acid residues affected by post-translational modifications such as phosphorylation. Phosphorylation of a residue introduces a change in mass associated with a phosphate molecule (80 Da) that can be distinguished during database searching. In this case, we found 13 phosphorylation sites present in 8 proteins (Table S3). Six sites are modified by proline-directed kinases. These phosphorylation events may be relevant to physiological functions of the proteins. For example, neuromodulin (also known as GAP-43 and B-50) plays a critical role in axonal growth and synapse development and plasticity, and is regulated by PKC phosphorylation (65).

Quantitation of abundance of specific structural proteins and receptors
MS has been used in proteomics mainly for qualitative identification of proteins, but the power of this approach can be greatly increased if the quantity of protein can also be measured. counterparts will be detected as pairs of peptides with identical sequence differing only in mass.
In a mass spectrometer, both peptides are selected and fragmented to generate many ion pairs, one of which is monitored and quantified. The ratio between the intensities of the fragment ion pairs provides an accurate measure of the relative abundance of the endogenous versus "spiked" proteins in the original sample. If the mass of the isotopically-labeled peptide/protein is known, then the mass of the endogenous counterpart can be calculated.
We applied the AQUA method to measure the quantity of six proteins: GluR1 and GluR2 (the major AMPA receptor subunits), NMDAR1 (also known as NR1, the essential NMDA receptor subunit), and PSD-95, GKAP, Shank1 (three scaffold proteins that are linked together in the PSD) (3). 15 N-labeled recombinant proteins rather than peptides were used because they can also be used as controls for variation during trypsin digestion. The mass of the 15 N-labeled recombinant proteins was measured by titration of Commassie blue staining against known amounts of bovine serum albumin (BSA). Because Commassie blue dye interacts with basic residues in proteins, staining intensity of a protein depends on the number of lysine, histidine, arginine residues and the N-terminal amino group (66). Therefore protein staining intensity was adjusted according to the sum of its basic groups. To measure all six proteins, the PSD fraction (~20 µg) was spiked with these internal standards (4 picomoles each) and applied to a short SDS gel, which permit the removal of SDS, washing and efficient digestion. All proteins were digested into one highly complex peptide fraction. During the LC-MS/MS analysis, the targeted peptide pair was isolated and subjected to fragmentation by collision-induced dissociation. One pair of fragment ions was detected and quantified. In the case of PSD-95, a fragment ion y 6 pair from the peptide, SLENVLEINKR, was monitored ( Figure 4B). The endogenous fragment ions were shown as two major adjacent peaks with m/z of 772. 5 Figure 4A). In addition, we pre-examined the elution time of the peptide by analyzing the internal standard itself under the same elution gradient condition (data not shown). Comparing the two chromatograms, it was clear that both peptides were eluted around the 13.9 minute point, because noise peaks were readily distinguished from the real peak based on the peptide elution time ( Figure 4A) and the fragment ions ( Figure 4B). The ratio of the two peak areas (internal standard/endogenous) was about 2.8 and therefore the amount of PSD-95 was approximately 1.4 picomoles (4 picomoles/2.8) in the sample. The other five proteins were quantified in the similar manner. Consistent results were derived from two separate experiments, the measurements of abundance varying 2.5-15% from the mean values (Table 1).
Of the six proteins examined, PSD-95 is the most abundant (~1.6 pmol in 20 µg of PSD protein, ~0.8% of the total PSD protein). In molar terms, PSD-95 is 5-fold more abundant than NR1 ( Table 1). The PSD-95-related proteins chapsyn-110/PSD-93 and SAP102 can also bind NMDA receptors and they are approximately half as abundant as PSD-95 based on abundance index (Table S1). Therefore, assuming that there are two NR1 subunits per NMDA receptor (67), Proteomic Analysis of PSD proteins of the PSD-95 family are probably at least an order of magnitude in stoichiometric excess over NMDA receptor-channels. PSD-95 by itself is also ~4-fold and ~ 7-fold more abundant than GKAP1/SAPAP1 and Shank1, respectively (Table 1). Although other members of the GKAP and Shank families exist in the PSD, it seems likely that the PSD-95 family of scaffolds is in stoichiometric excess over GKAP and Shank scaffolds. As expected, AMPA receptors are less abundant in the PSD than NMDA receptors ( Table 1). The AQUA measurements indicate that GluR2 is more than 2-fold more abundant than GluR1, in keeping with the fact that GluR2 is present in GluR1/GluR2 as well as GluR2/GluR3 heteromers (68).
Proteomic Analysis of PSD DISCUSSION Our LC-MS/MS analysis detected virtually all the "core" proteins previously known to be biochemically enriched in the PSD. We failed to identify several proteins that were shown to be On the other hand, we identified numerous proteins that had not been previously associated with the PSD fraction, and examples of these are discussed in the results. Many of these unexpected proteins are intriguing candidates for being involved in synaptic morphology and/or signaling: e.g. regulators of Arf, Rac and Rap, and an atypical myosin containing a PDZ domain. However, it is critical to follow up the proteomics approach with more focused experiments to affirm whether these proteins are truly components of the PSD, and if so, what their physiological functions are. The systematic analysis of the distribution and function of previously unidentified PSD proteins is outside the scope of the present study. However, the large number of "new" components discovered and the diversity of their structures/activities imply that the PSD is even more complex than previously imagined from biochemical, yeast Proteomic Analysis of PSD two-hybrid and earlier MS studies. One important question is whether the diversity of proteins found in the PSD solely reflects the rich variety of components present in all PSDs, or whether the diversity is contributed by heterogeneous PSDs from different parts of the forebrain that contain distinct protein compositions.
A major concern for biochemically purified PSDs is contamination by subcellular structures that are unrelated to postsynaptic function. Such impurities would contribute artifactually to the observed multiplicity of proteins in the PSD. As discussed above, the "contamination" by presynaptic proteins is not surprising, given the intimate interactions that exist between pre-and postsynaptic membranes. However, the numerous "hits" on metabolic enzymes, mitochondrial proteins, myelin basic protein and glial fibrillary acidic protein (GFAP) certainly suggests contamination by housekeeping proteins, mitochondria, glial cytoskeleton and myelin sheaths, which are all abundant in the brain. Similar presumptive contaminants were also found in other mass spectrometry screens of the PSD (14,82). On the contrary, glycolytic enzymes such as glyceraldehyde-3-phosphate dehydrogenase have been localized to postsynaptic sites by immunocytochemistry and proposed to provide ATP to the PSD (83), so it is plausible that some of these presumptive "impurities" actually reflect specific association with the PSD. Impurities of the PSD isolation might have been reduced by use of stronger detergents such as N-lauryl sarcosinate (9), but strong detergent treatment could remove many proteins that are real and particularly interesting components of the PSD. Hence, we compromised by using the double Triton X-100 extraction method (4).

Proteomic Analysis of PSD
Several MS studies of the PSD and the NMDA receptor complex have recently appeared (14,15,17,82). The most recent utilized MS in conjunction with 2DGE to identify most of their PSD proteins, with a smaller contribution from ICAT-based LC-MS/MS (82), resulting in identification of approximately 180 proteins. The increased sensitivity of our LC-MS/MS analysis allowed us to identify over twice as many proteins, providing a more comprehensive view of the PSD proteome. Perhaps more importantly, however, we measured directly for the first time the absolute abundance of several key components of the PSD, using the AQUA method.
Based on our measurements, PSD-95 is five-fold more abundant than NR1 in the PSD, and ten-fold more abundant than NMDA receptors (assuming two NR1 subunits per NMDA receptor complex (67). The stoichiometric excess could be greater than this, if we factor in the other members of the PSD-95 family (e.g. chapsyn-110/PSD-93 and SAP102) that can also bind to NR2 subunits of NMDA receptors (84,85), and which are approximately half as abundant as PSD-95 based on abundance index (Table S1). Our AQUA quantitation highlights the fact that PSD-95 family proteins, although first characterized as the binding partners for NMDA receptors, probably function as postsynaptic scaffolds for many other proteins, some of which could be more abundant than NMDA receptors. The great molar excess of PSD-95 family proteins over NMDA receptors could explain why mouse mutants in PSD-95 show apparently normal NMDA receptor localization at synapses (86), and why PSD-95 overexpression does not potentiate synaptic NMDA receptor responses (87).
Proteomic Analysis of PSD PSD-95 binds to GKAP, which in turn binds to Shank. These scaffold proteins are relatively abundant and form an "axis" of scaffolds connecting the membrane-proximal region to the cytoplasmic face of the PSD (16,88). In molar terms, we report here that PSD-95 is more abundant than GKAP/SAPAP1 (~4 fold) and Shank1 (~7-fold) in PSD fractions. The AQUA quantitation needs to be applied to other members of these scaffold protein families to obtain a comprehensive picture, but based on the current measurements and the abundance indexes of close family relatives, it seems that PSD-95 family proteins are in considerable molar excess over GKAP/SAPAPs. This discoordinate stoichiometry could reflect for instance that GKAP can bind to multiple molecules of PSD-95 (89). Alternatively, it could mean that the abundance of PSD-95, GKAP and Shank are independently regulated at postsynaptic sites, as has been suggested by studies of PSD protein ubiquitination (7). Quantitative measurements of other protein stoichiometries in PSD fractions will continue to shed light on the molecular organization of the PSD, and suggest functional hypotheses for further investigation.       Abundance Index was derived as described in the methods. #Unique is the number of peptides whose sequences are only found in the assigned protein.
#Shared is the number of peptides that are also found in one or more other identified proteins (e.g. family members)

peptides (374 proteins) identified in the PSD fraction
The list was sorted by functional categories and then by protein synonym. In the column of "peptides", the amino acid residues before and after identified peptides are shown and separated by dots. "#": oxidized methionine; "-" : N-or C-terminus Peptides shared by multiple proteins are shown repeatedly for their assigned proteins. tenascin-R, janusin R.CPT*DCS*S*RGLCVDGECVCEEPYTGEDCR.E The front and back amino acids of a tryptic phosphopeptide were also shown and separated by dots. The phosphorylation sites were indicated by asterisks.