Alterations in the Proteome of Developing Neocortical Synaptosomes in the Absence of MET Signaling Revealed by Comparative Proteomics

Alterations in the expression of genes encoding proteins involved in synapse formation, maturation, and function are a hallmark of many neurodevelopmental and psychiatric disorders. For example, there is reduced neocortical expression of the MET receptor tyrosine kinase (MET) transcript and protein in Autism Spectrum Disorder (ASD) and Rett syndrome. Preclinical in vivo and in vitro models manipulating MET signaling reveal that the receptor modulates excitatory synapse development and maturation in select forebrain circuits. The molecular adaptations underlying the altered synaptic development remain unknown. We performed a comparative mass spectrometry analysis of synaptosomes generated from the neocortex of wild type and Met null mice during the peak of synaptogenesis (postnatal day 14; data are available from ProteomeXchange with identifier PXD033204). The analyses revealed broad disruption of the developing synaptic proteome in the absence of MET, consistent with the localization of MET protein in pre- and postsynaptic compartments, including proteins associated with the neocortical synaptic MET interactome and those encoded by syndromic and ASD risk genes. In addition to an overrepresentation of altered proteins associated with the SNARE complex, multiple proteins in the ubiquitin-proteasome system and associated with the synaptic vesicle, as well as proteins that regulate actin filament organization and synaptic vesicle exocytosis/endocytosis, were disrupted. Taken together, the proteomic changes are consistent with structural and functional changes observed following alterations in MET signaling. We hypothesize that the molecular adaptations following Met deletion may reflect a general mechanism that produces circuit-specific molecular changes due to loss or reduction of synaptic signaling proteins.


Abstract
Alterations in the expression of genes encoding proteins involved in synapse formation, maturation, and function are a hallmark of many neurodevelopmental and psychiatric disorders. For example, there is reduced neocortical expression of the MET receptor tyrosine kinase (MET) transcript and protein in Autism Spectrum Disorder (ASD) and Rett syndrome. Preclinical in vivo and in vitro models manipulating MET signaling reveal that the receptor modulates excitatory synapse development and maturation in select forebrain circuits. The molecular adaptations underlying the altered synaptic development remain unknown. We performed a comparative mass spectrometry analysis of synaptosomes generated from the neocortex of wild type and Met null mice during the peak of synaptogenesis (postnatal day 14; data are available from ProteomeXchange with identifier PXD033204). The analyses revealed broad disruption of the developing synaptic proteome in the absence of MET, consistent with the localization of MET protein in pre-and postsynaptic compartments, including proteins associated with the neocortical synaptic MET interactome and those encoded by syndromic and ASD risk genes. In addition to an overrepresentation of altered proteins associated with the SNARE complex, multiple proteins in the ubiquitin-proteasome system and associated with the synaptic vesicle, as well as proteins that regulate actin filament organization and synaptic vesicle exocytosis/ endocytosis, were disrupted. Taken together, the proteomic changes are consistent with structural and functional changes observed following alterations in MET signaling. We hypothesize that the molecular adaptations following Met deletion may reflect a general mechanism that produces circuitspecific molecular changes due to loss or reduction of synaptic signaling proteins.

Introduction
Disruption of synaptic function represents a key point of convergence in the pathology underlying many neurodevelopmental and psychiatric disorders. Consistent with this, pathway analyses of postmortem human brain have revealed enrichment in altered expression of genes that encode proteins involved in synapse formation, maturation, and function (e.g., [1][2][3][4][5][6][7][8][9]), including the MET receptor tyrosine kinase (MET). Specifically, there is reduced expression of MET transcript and/or protein in the postmortem temporal lobe of Autism Spectrum Disorder and Rett syndrome cases compared to age-and sex-matched control subjects [10][11][12][13]. While not a causal gene for either disorder, the altered expression has functional implications given the role of the receptor in synapse development.
There is an abundance of in vivo and in vitro studies demonstrating that MET modulates excitatory synapse development and maturation, including visual cortex critical period timing, in select circuits of the forebrain [14][15][16][17][18][19][20][21]. In the developing primate and rodent forebrain, MET/Met is enriched in subpopulations of excitatory projection neurons in specific regions of the neocortex, the CA1 hippocampus subfield and a limited number of subcortical structures, with peak expression corresponding to periods of dendritic growth and synapse formation [22][23][24][25]. MET protein is enriched in the neuropil of these regions during development, accumulating at pre-and postsynaptic compartments [26]. When activated by its only known ligand, hepatocyte growth factor, MET interacts with select synapse-associated proteins and engages specific signaling cascades to mediate synaptic vesicle clustering and synapse formation in neocortical neurons in vitro [14,20,27]. In mice in which Met is deleted in cells arising from the dorsal pallium or its expression is extended temporally via a transgene in neocortical and hippocampal excitatory neurons, alterations occur in developing hippocampal and neocortical circuitry. Functionally, excitatory synapses mature precociously in the absence of Met [19] but exhibit a lengthened period of plasticity when the normal temporal window of MET signaling is extended later [15,17]. Structurally, alterations in pyramidal neuron dendritic arbor and spine head volume are observed [18,19,28]. Thus far, studies of molecular adaptations to altered MET expression have focused on candidate postsynaptic proteins, revealing altered developmental expression of select glutamate receptor subunits [17,19]. The full breadth of altered protein expression at the developing neocortical synapse in the context of altered MET signaling has not been determined.

Methods and Materials
Mice Conditional Met null mice (Met fx/fx /nestin cre ) were generated and genotyped as previously described [50]. Breeding lines were backcrossed onto the C57BL/6J background for more than 10 generations and subsequently at least every fifth generation using mice purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were housed in ventilated racks in a climate-controlled vivarium under at 13:11-h light/dark cycle, with food and water provided ad libitum. All mouse procedures were approved by the Institutional Animal Care and Use Committee at the Children's Hospital Los Angeles and conformed to NIH guidelines.

Generation of Synaptosomes
Unless otherwise noted, all procedures were performed at 4°C and all chemicals are from Sigma (St Louis, MO). Synaptosomes were generated and prepared for proteomics as previously described [26,51], with minor modifications. Briefly, wild type and null mice were anesthetized with saturated isoflurane vapors (Covetrus, Portland, ME) between 8:00 and 11:00 a.m. on postnatal day (P) 14, during the period of peak synaptogenesis [52][53][54][55], decapitated, and the brain removed. The neocortex was dissected, snap frozen, and stored at −80°C until use. Synaptosomes were generated from two independent wild type and null neocortical samples on the same day, with each sample representing tissue pooled from 2 male mice. Pooled tissue was homogenized in 320 mM sucrose, then centrifuged at 1,000 g for 10 min to remove nuclei. The sucrose concentration of the resulting supernatant was adjusted to 1.25 M, overlaid with 1 M sucrose, and centrifuged at 100,000 g for 3 h. The synaptosome fraction, located at the sucrose interface, was removed and precipitated overnight with ice-cold acetone at −20°C. Precipitated samples were centrifuged at 14,000 g for 10 min, dried at room temperature, and solubilized in 1% PPS

Proteomics Screen
Proteomics procedures were performed at the Mass Spectrometry Research Center Proteomics Core at the Vanderbilt University School of Medicine. Two 4-plex isobaric tag for relative and absolute quantitation (iTRAQ) experiments, using independent samples, were performed. Each experiment included two independent wild type and null samples. Mass spectrometry methods and data analyses were performed similarly to those described previously [51,56], with some minor modifications. One unit of labeling reagent (reporter tags: 114, 115, 116, 117) was used for 25 μg protein. MudPIT analysis was performed using either a Q Exactive or Q Exactive Plus mass spectrometer. Samples analyzed on a Q Exactive were analyzed similar to Voss et al. [56] 2015 with a few differences. The Q Exactive instrument was operated in datadependent mode acquiring HCD MS/MS scans after each MS1 scan on the 20 most abundant ions using an MS2 target of 1 × 10 5 ions. The HCD-normalized collision energy was set to 30, and dynamic exclusion was set to 30 s. For experiments conducted on a QE Plus instrument, the instrument was operated in datadependent mode acquiring HCD MS/MS scans on the 15 most abundant ions using an MS1 ion target of 3 × 10 6 ions and an MS2 target of 1 × 10 5 ions. For experiments on both the QE Plus, a Dionex Ultimate 3,000 nano LC and autosampler were used, and peptides were gradient-eluted from the reverse analytical column at a flow rate of 350 nL/min. For the peptides from the first 11 strong cation exchange (SCX) fractions, the reverse phase gradient consisted of 2-50% solvent B (0.1% formic acid in acetonitrile) in 83 min, followed by a 10 min equilibration at 2% solvent B. For the last 2 SCX-eluted peptide fractions, the peptides were eluted from the reverse phase analytical column using a gradient of 2-98% solvent B in 83 min, followed by a 10 min equilibration at 2% solvent B. Peptide/protein identifications and quantitative analysis were performed using Spectrum Mill (Agilent Technologies, Santa Clara, CA) as described previously. MS/MS spectra were searched against a subset of the UniProt KB protein database (www.uniprot. org) containing Mus musculus proteins. Autovalidation procedures in Spectrum Mill were used to filter the data rigorously to <1% false discovery rates at the protein and peptide levels. For each group in each experiment, log2 protein ratios were fit to a normal distribution using nonlinear (least squares) regression. The calculated mean derived from the Gaussian fit was used to normalize individual log2 ratios for each quantified protein.
The mass spectrometry proteomics data were deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD033204 [57]. In each experiment, one wild-type sample was randomly designated as the comparator sample. The quantitative ratios for each protein in the two null and remaining wild-type sample were calculated to determine adaptations in the synaptosomal proteome in the absence of MET signaling. To combine data from the two independent experiments, the ratios for each null sample were then normalized to the wild-type ratio and a dataset comprising proteins exhibiting a mean fold difference >1.25 or <0.8 was generated. For each protein, one-sample t tests were used to compare the ratios to one (no difference in expression between wild type and null) and filtered for false discoveries (false discovery rate, <0.05) using the Benjamini-Hochberg procedure, where the original p value was compared to the Benjamini-Hochberg critical value.
Protein identifiers were converted to Entrez Gene identifiers using DAVID. Genes corresponding to the proteins detected in the developing neocortical synaptosomes were analyzed for (1) proteins detected in a published mouse developing neocortical synaptosome proteome [32], (2) genes annotated in Synaptic Gene Ontologies (SynGO release 1.1; https://www.syngoportal.org/) [58], (3) members of the neocortical MET synaptic interactome defined by Xie et al. [27] 2016, and 4) proteins encoded by ASD risk genes defined by Simons Foundation Autism Research Initiative Gene (SFARI Genesyndromic or category 2 and above; downloaded August 28, 2021). Gene ontology (GO) enrichment analyses were performed using the database for Annotation, Visualization, and Integrated Discovery (DAVID, version 2021) [59,60] with default settings, and SynGO against the brain expressed background setting and with first and second level terms as labels. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed using DAVID alone.

Western Blot
Western blot analyses were used to validate select proteins in the dataset using synaptosomes generated from a separate cohort of mice. Synaptosomes were prepared as described for iTRAQ analysis, except that the acetone-precipitated samples were solubilized in 5% SDS in Tris-HCl (pH 7.2). SDS-PAGE and Western blotting were performed as described previously [26], with proteins <60 kD resolved with 12% gels and those >60 kD with 7.5% gels. Unless noted otherwise, all incubations/washes were at room temperature. Prior to immunoblotting, membranes were cut at a desirable range of protein mass, based on Precision Plus All Blue Prestained Protein Standards (Bio-Rad, Hercules CA, online suppl. Fig. 1; for all online suppl. material, see www.karger.com/doi/10.1159/000529981). Images of the cut membranes, including the molecular weight standards, are provided in online supplementary Figure S1. Blots were first incubated for 1 h in blocking buffer comprising 4% Nonfat Dry Milk (Cell Signaling Technologies, Danvers MA) in Trisbuffered saline (TBS), pH 7.2. Blots were then incubated overnight at 4°C with one of the following primary antibodies diluted in 0.05% Tween 20 in blocking buffer: rabbit anti-VAMP2/synaptobrevin 2 (Synaptic Systems Cat# 104 202, Cedarlane Laboratories, Burlington NC; 1:2,000), rabbit anti-Trim3 (ThermoFisher Cat# 28392-1-AP, 1:2,000), rabbit anti-Cdc42ep4 (LifeSpan BioSciences Cat# LS-C804922, Seattle, WA, 1:1,000), rabbit anti-Galpha(z) (Cell Signaling Technologies Cat # 3904, 1:1,000), and mouse anti-Syntaxin 1A (Synaptic Systems Cat# 110111, 1:2,000). Following 8 × 5-min washes in TBS, blots were incubated for 1 h in species appropriate horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch: mouse -Cat# 71-035-10; rabbit -Cat# 715-035-152, West Grove PA) diluted 1:5,000 in 0.05% Tween 20 in blocking buffer. Following 8 × 5-min washes in TBS, immunoreactive bands were visualized with Pierce SuperSignal West Dura Chemiluminescent substrate (Thermo Scientific). Image capture and densitometric analysis was performed using a CC camera coupled to a UVP BioImaging System using VisionWorks LS Image Acquisition and Analysis software (v. 7.0.1, VisionWorks). The density of each band was measured, a global background subtraction applied, and the resulting value normalized to the wild-type mean. A Hedge's g, appropriate for small sample sizes, was calculated as an estimate of effect size. A protein was considered validated if it met the same mean fold change criterion as for the iTRAQ analyses (<0.8 or >1.25), with the same direction of change and with a medium (>0.5) to large (>1.0) effect size.

Characterization of P14 Neocortical Synaptosomes
This study used a discovery-based MS approach to profile synaptosomes isolated from the mouse neocortex at the peaks of synaptogenesis and MET expression (P14). Western blot analyses demonstrated enrichment of the presynaptic marker, synaptophysin, and depletion of the nuclear marker, lamin B, in isolated synaptosomes compared to the crude homogenate and the cytoplasm fraction ( Fig. 1a; online suppl. Fig. S2). We identified 3,616 unique proteins across the two iTRAQ experimentseach experiment comprising two independent wild type and Met −/− biological replicateswith 2,149 proteins detected in both experiments ( Fig. 1b; online suppl. Table S1). Only proteins detected in both experiments were considered for further analysis (online suppl. Table S2). Approximately 80% of proteins detected in our P14 neocortical synaptosomes were also detected in cortical synaptosomes isolated from 3-to 8-week-old male C57BL/6J mice using a 4-plex iTRAQ approach (Fig. 1c; online suppl. Table S2) [32], demonstrating reproducibility across laboratories when using the same approach. We next performed enrichment analyses for GO cellular component (GO-CC) using genome-wide (DAVID) and synapse-specific (SynGO) tools. Analyses using DAVID revealed an enrichment in proteins associated with the synapse, with 6 of the top 20 terms (based on p values) representing synaptic components; other enriched terms included those representing organelles and structural elements commonly observed at the synapse, including mitochondria, cytoskeleton, and endosomes (Fig. 1d). Five hundred and thirty proteins were annotated in SynGO, with enrichment in both pre-and post-synaptic components (Fig. 1e). The complete set of enriched terms, and the proteins associated with each term, is listed in online supplementary Tables S3 (DAVID) and S4 (SynGO).
We next determined adaptive molecular changes in neocortical synaptosomes in the absence of MET signaling. The distribution of protein ratios across the four samples was overlapping but not identical (Fig. 2a). Based on average fold change across the four samples, 509 proteins met a fold-change criteria of either <0.8 (227 proteins) or >1.25 (282 proteins; Fig. 2b; online suppl. Table S5). Of these, 389 were significantly altered in the absence of MET signaling, 315 of which exceeded FDR <0.05 (online suppl. Table S6). Because this is a discovery study, we used all 389 proteins in the subsequent analyses.
We first used GO-CC and KEGG pathway analyses to determine if adaptations to the synaptosomal proteome in the absence of MET signaling were enriched in specific cellular components and pathways. As noted above, the neocortical synaptosomal dataset is enriched for proteins at the synapse and in synaptosome-associated organelles. We therefore used the neocortical synaptosomal dataset, rather than whole genome (DAVID) or brain background (SynGO), to serve as the background comparison in these analyses. When all changed proteins were considered as a single dataset, DAVID returned one GO-CC term (integral component of the membrane) and three KEGG terms (Huntington disease, prion disease, and phagosome) ( Fig. 3a; online suppl. Table S7). Downregulated proteins returned four GO-CC terms (SNARE complex, extracellular region, membrane, integral component of membrane) and one KEGG term (SNARE interactions in vesicular transport) (Fig. 3b). The proteins associated with each term are listed in Table 1 (SNARE complex) or  online supplementary Table S7. Approximately, 17.5% (93/530) of the proteins annotated in SynGO exhibited altered expression in the absence of MET signaling. There were, however, no significantly overrepresented terms, but rather changes were observed broadly across the  Table S3. e 530 proteins were categorized in SynGO cellular component. Sunburst plot with color-coded enrichment significance for top-level and secondlevel GO: cellular component terms. The number of counts compared to the total number of hits for each term is indicated. The proteins associated with the significant GO terms can be found in online supplemental Table S4.   Table S7. c Sunburst plot with color-coded gene counts for top-level and second-level SynGO: cellular component terms that include proteins exhibiting differential expression in the absence of Met. No term was overrepresented. The number of counts compared to the total number of hits for each term is indicated. The proteins associated with each term can be found in online supplemental presynapse, the postsynapse, the synaptic cleft, and synaptic membrane ( Fig. 3c; online suppl. Table S8). In addition, multiple proteins in the ubiquitin-proteasome system (UPS) and associated with the synaptic vesicle, as well as proteins that regulate actin filament organization and synaptic vesicle exocytosis/endocytosis, were disrupted (Table 1).
Given the altered expression of Met/MET in ASD, we were interested in whether expression of other ASDassociated proteins was altered in the absence of MET signaling. We considered ASD-associated proteins to be those encoded by genes defined as category 1 (high confidence), category 2 (strong candidate), or category S (syndromic) by SFARI gene, totaling 546 proteins. We detected 160 ASD-associated proteins in our neocortical synaptosome dataset (Fig. 4c; online suppl. Table S2), representing a significant enrichment (χ 2 = 170.34, df = 1; p < 0.001) using the whole genome as the background comparison. Twenty-four ASD-related proteins detected in the neocortical synaptosomes were altered, with 15 upregulated and 9 downregulated (Fig. 4d). Although this did not represent a significant overrepresentation of proteins encoding ASD risk genes among proteins Actin filament organization Upregulated coronin-7; rho guanine nucleotide exchange factor 2; protein kinase C iota type Downregulated tropomodulin-2; tropomyosin alpha-1, -3 and -4 chain; protein BRICK1; protein kinase C and casein kinase substrate in neurons protein 1; paralemmin-2; MARCKS-related protein exhibiting altered expression (χ 2 = 0.302, df = 1; p = 0.582), it is consistent with MET being part of a larger network of synaptic signaling molecules that includes proteins encoded by syndromic and other high confidence genes. representatives of the MET interactome and proteins encoded by ASD risk genes. Four proteins met the average fold change criteria established for the iTRAQ analyses, with the same direction of change: VAMP2, 0.76 (Fig. 5b), STX1A, 0.79 (Fig. 5c), TRIM3, 1.36 (Fig. 5e), and GNAZ, 1.81 (Fig. 5f). One protein, however, did not meet this criterion: CDC42EP4, 0.90 (Fig. 5d). The 4 proteins that met criteria all exhibited a large effect size (Hedges g): VAMP2, 2.05; STX1A, 2.17; TRIM3, 1.25; GNAZ, 2.03.

Discussion
The present study used a discovery-based proteomics approach to identify proteins that exhibit differential expression in developing neocortical synaptosomes in the absence of MET signaling. The study complements studies demonstrating the impact of Met deletion on disrupting structural and functional synaptic development, maturation, and physiological properties. The analysis revealed a broad disruption of the synaptic proteome during the peak of synaptogenesis, including changes in proteins associated with the neocortical synaptic MET interactome and those encoded by syndromic and ASD risk genes. In addition to the overrepresentation of proteins associated with the SNARE complex, multiple proteins in the ubiquitinproteasome system (UPS) and associated with synaptic vesicles, as well as proteins that regulate actin filament organization and the synaptic vesicle exocytosis/endocytosis, are also disrupted. These changes are consistent with the alterations in spine dynamics and size, and the precocious maturation of glutamatergic synapses observed in the absence of MET signaling [16,18,19,28].
There is an acceleration of the stereotypical developmental changes in N-methyl-D-aspartate (NMDA) and αamino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor subunit expression in the neocortex and hippocampus following reduced expression of Met [16,19]. These changes are due to altered trafficking of the subunits to the synaptic membrane rather than alterations in global expression [19]. We also do not detect changes in subunit expression in our whole synaptosome preparations, consistent with a previous study that reported no changes in total AMPA receptor expression but increased levels in the postsynaptic density in synaptosomal preparations in the context of synaptic plasticity [61]. We do, however, observe reduced expression of proteins with known roles in trafficking AMPA and NMDA receptors into and out of the PSD, namely, protein kinase C and casein kinase II substrate in neurons (PACSIN1), neuroplastin (NPTN), and calsytenin-1 (CLSTN1). Protein interacting with C-kinase 1 (PICK1), whose expression remains unchanged in the absence of MET signaling, regulates trafficking of AMPA receptors through its interaction with PACSIN1 [62] and knockdown of PACSIN1 alone is sufficient to reduce activity-dependent internalization of glutamate receptor 2 (GluR-2) in hippocampal neurons [62,63]. Homophilic binding of NPTN activates p38 MAPK and the subsequent internalization of glutamate receptor 1 (GluR-1) [64]. In Clstn1-null mice, the glutamate [NMDA] receptor subunit epsilon-2 (GluN2B) to glutamate [NMDA] receptor subunit epsilon-1 (GluN2A) ratio is disrupted in juvenile hippocampus [65]. Although these interactions have been examined primarily in the context of synaptic plasticity, it is likely that they perform similar functions during development and underlie in part the altered developmental trajectory of glutamate subunit expression in the synaptic membrane in conditional Met-null mice.
Similar to postsynaptic maturation, the development of the synaptic vesicle pool, including the emergence of appropriate patterns of exocytosis and endocytosis, follows a typical trajectory (reviewed in [66]), and it has been proposed that even "a small shift in the expression of key presynaptic proteins may be sufficient to alter circuit activity during development resulting in a shift in the critical period" [66]. The current study identified altered expression of synaptic vesicle proteins, as well as nonvesicle proteins that modulate synaptic vesicle exocytosis and endocytosis. This includes, for downregulated proteins, an overrepresentation of proteins associated with the SNARE complex that is crucial for synaptic vesicle exocytosis and synchronization of neurotransmitter release [67]. These findings are consistent with our previous biochemical and immunoelectron microscopic analyses demonstrating an association of MET with synaptic vesicles in the developing neocortex [26] and in vitro studies showing that HGF activation of MET increases clustering of synaptic vesicles at the active zone [20]. We further suggest, in line with the proposal put forth by Bonnycastle et al. [66], that the current data indicate there are presynaptic contributions to the circuit alterations observed following developmental disruptions in MET signaling. The ubiquitin proteasome system regulates protein turnover. Specifically, ubiquitins are attached to a substrate protein via multiple rounds of an enzyme cascade comprising an ubiquitin E1 activating enzyme, E2 conjugase, and E3 ligase, with the E3 ligase determining substrate specificity, creating a signal for degradation by the proteasome (reviewed in [68]). Based on the nature of the ubiquitin chain, protein ubiquitination can also regulate protein endocytosis and localization [69]. Multiple members of the UPS are altered in developing neocortical synaptosomes in the absence of MET signaling, including an E1 activating enzyme (ubiquitin-like modifier-activating enzyme 6) and several E3 ligases/components of an E3 ligase complex (RNF14, HECW1, HECW2, UBR4, UBE3A, TRIM3, cullin-1, cullin-5, SKP1, and DNA damage-binding protein 1), likely contributing to altered synaptic development and maturation. For example, the UPS plays a key role in synaptic plasticity, principally by influencing protein turnover and protein trafficking, including AMPA receptor trafficking [70][71][72][73]. The UPS also impacts expression levels of presynaptic proteins, including syntaxin 1 (reviewed in [74]), and in vitro evidence indicates that it may also influence synaptic vesicle cycling [75]. In these contexts, we suggest that the alterations in the UPS identified in the current study contribute to the accelerated maturation of glutamatergic synapses and altered expression of proteins associated with the synaptic vesicle cycle observed in the absence of MET.
Developmental alterations in MET signaling disrupt activation of downstream signaling proteins, such as cdc42, Rac1, and cofilin-1, that influence the actin cytoskeleton [17,18], the major cytoskeletal component in spines [76]. This disruption has been hypothesized to underlie altered spine dynamics and maturation and spine head size observed following alterations in MET signaling [17]. The current analysis identified disrupted expression of proteins associated with actin filament organization in Met-null developing neocortical synaptosomes, including the downregulation of tropomodulin 2 (TMOD2), tropomyosin alpha-3 chain (Tpm3), tropomyosin alpha-1 chain (Tpm1), and tropomyosin alpha-4 chain (Tpm4), lending support to this hypothesis. Tropomyosins regulate the binding of actin-binding proteins and myosin motors to actin filaments [77], thus modulating actin filament remodeling required during spine formation and maturation. Tropomodulins cap the minus end of tropomyosin-coated actin, thus regulating the length of the actin filament [78]. Alterations in the expression of TMOD2 alter the number of dendritic spines and filopodia and influences spine head morphology in hippocampal neurons [79,80]. Further, deficits in learning and memory are observed following knockout of Tmod2 [81].
Finally, we performed analyses focused on the developing neocortical synaptic MET interactome and on proteins encoded by high confidence ASD and syndromic genes. First, we identified an overrepresentation of proteins from both groups, consistent with a role in synapse development/ function. Second, we identified changes in the expression of 18 members of the MET interactome and 24 proteins encoded by risk genes in the absence of MET, with one (TRIO) common to both groups. We note that we are measuring global protein expression in these analyses. As reported previously for glutamate receptor subunits, the trafficking of MET interactome or ASD/syndromic proteins, rather than total expression, may also be impacted. Analyses of the proteome of subsynaptic fractions in the absence of MET would address this possibility. We also note that changes in posttranslational modifications, including ubiquitination, phosphorylation, neddylation, and SUMOylation, would also impact protein function. Indeed, as noted previously, alterations in the phosphorylation status of cdc42, Rac1, and cofilin-1, rather than changes in expression levels, are seen following disruption of MET signaling [17,18]. The data here, together with other studies, indicate that a combination of altered protein expression, posttranslational modification, and intracellular trafficking likely contribute to disrupted synaptic development and maturation.
There are some limitations to the current study. The iTRAQ proteomics analysis does not readily detect low abundance proteins, and we cannot exclude the possibility that expression of these proteins is also impacted in the absence of MET signaling. In addition, the data represent the average response of all synapses in neocortex, including those that do not express MET. There are likely discrete responses in individual MET-expressing synapses within the neocortex that are unable to be detected using the current approach. Finally, we note that differences in the synaptosomal isolation procedure may influence the repertoire of protein changes identified in the synaptosome proteome in the absence of MET [82].
Although correlational, many of the changes in expression identified in the current study are consistent with structural/functional changes observed in vivo and in vitro in the absence of MET. Further, MET, while not itself encoded by a causal ASD gene, appears to be part of a larger network of synaptic signaling molecules that includes proteins encoded by syndromic and other high confidence genes. Alterations in MET expression and signaling during peak synaptogenesis in circuits in which the receptor is expressed may contribute long-term consequences on synaptic function and altered developmental outcomes.