AMPA Receptor Complex Dynamics in Time and Space
2014; Cell Press; Volume: 84; Issue: 1 Linguagem: Inglês
10.1016/j.neuron.2014.09.018
ISSN1097-4199
Autores Tópico(s)Mass Spectrometry Techniques and Applications
ResumoUsing a comprehensive proteomic approach, Schwenk et al. (2014), in this issue of Neuron, resolve the differential composition of AMPA receptor complexes in brain regions and through development. This work reveals a specificity in AMPA receptor complex assembly that is dynamic in both space and time. Using a comprehensive proteomic approach, Schwenk et al. (2014), in this issue of Neuron, resolve the differential composition of AMPA receptor complexes in brain regions and through development. This work reveals a specificity in AMPA receptor complex assembly that is dynamic in both space and time. Proteomic approaches have transformed our understanding of how proteins interact to form functional units in cells. In particular, the characterization of multiprotein complexes by affinity purification and mass spectrometry has been incredibly useful to understand how molecular functions are coordinated and regulated by sets of gene products. Multiprotein complexes can range from the relatively small with few components, to molecular machines such as the proteasome, to larger assemblies or supercomplexes within the postsynaptic density. Purification of membrane-associated protein complexes makes the definition of the complex rather dependent on the stringency of the isolation methods, particularly the use of detergents, which can certainly influence the recovery of native complexes. However, this approach has been successfully used to characterize many important protein complexes in the brain, particularly those associated with a variety of ion channels such as NMDA (Husi et al., 2000Husi H. Ward M.A. Choudhary J.S. Blackstock W.P. Grant S.G. Nat. Neurosci. 2000; 3: 661-669Crossref PubMed Scopus (1025) Google Scholar, Collins et al., 2006Collins M.O. Husi H. Yu L. Brandon J.M. Anderson C.N. Blackstock W.P. Choudhary J.S. Grant S.G. J. Neurochem. 2006; 97: 16-23Crossref PubMed Scopus (343) Google Scholar) and AMPA receptors (Kang et al., 2012Kang M.G. Nuriya M. Guo Y. Martindale K.D. Lee D.Z. Huganir R.L. J. Biol. Chem. 2012; 287: 28632-28645Crossref PubMed Scopus (24) Google Scholar, Schwenk et al., 2012Schwenk J. Harmel N. Brechet A. Zolles G. Berkefeld H. Müller C.S. Bildl W. Baehrens D. Hüber B. Kulik A. et al.Neuron. 2012; 74: 621-633Abstract Full Text Full Text PDF PubMed Scopus (300) Google Scholar, Schwenk et al., 2009Schwenk J. Harmel N. Zolles G. Bildl W. Kulik A. Heimrich B. Chisaka O. Jonas P. Schulte U. Fakler B. Klöcker N. Science. 2009; 323: 1313-1319Crossref PubMed Scopus (302) Google Scholar, von Engelhardt et al., 2010von Engelhardt J. Mack V. Sprengel R. Kavenstock N. Li K.W. Stern-Bach Y. Smit A.B. Seeburg P.H. Monyer H. Science. 2010; 327: 1518-1522Crossref PubMed Scopus (211) Google Scholar). NMDA receptors associate with a large number of proteins that can directly regulate ion channel gating or regulate the trafficking and stability of receptors in membrane subdomains. Together with scaffolding proteins and other signaling molecules, these protein assembles orchestrate many aspects of signal transduction in glutamatergic synapses. In fact, such complexes appear to be hot spots for neurological diseases with numerous genes associated with disorders, such as schizophrenia, and the biochemical definition of these sets of proteins as complexes is proving useful for overlaying data from human genetics studies, allowing connections between rare mutations to be made (Purcell et al., 2014Purcell S.M. Moran J.L. Fromer M. Ruderfer D. Solovieff N. Roussos P. O’Dushlaine C. Chambert K. Bergen S.E. Kähler A. et al.Nature. 2014; 506: 185-190Crossref PubMed Scopus (1004) Google Scholar). Compared to NMDA receptors, AMPA receptors appear to form more discrete protein complexes which are composed of the pore-forming subunits and members of at least three families of auxiliary subunits which together compose the inner core of the complex. Over the years, a growing list of proteins have been shown to interact with AMPA receptors, but it was only recently that comprehensive proteomic analyses of AMPA receptor complexes were reported. Schwenk et al. generated a comprehensive blueprint for an average AMPA receptor (AMPAR) complex purified from whole brain tissue using a combination of multiepitope affinity purification of complexes, native gel electrophoresis, and quantitative mass spectrometry (Schwenk et al., 2012Schwenk J. Harmel N. Brechet A. Zolles G. Berkefeld H. Müller C.S. Bildl W. Baehrens D. Hüber B. Kulik A. et al.Neuron. 2012; 74: 621-633Abstract Full Text Full Text PDF PubMed Scopus (300) Google Scholar). At the same time, Shanks et al. reported a comparative, quantitative proteomic analysis of both AMPA and kainite receptor complexes (Shanks et al., 2012Shanks N.F. Savas J.N. Maruo T. Cais O. Hirao A. Oe S. Ghosh A. Noda Y. Greger I.H. Yates 3rd, J.R. Nakagawa T. Cell Rep. 2012; 1: 590-598Abstract Full Text Full Text PDF PubMed Scopus (136) Google Scholar). Both of these studies confirmed the presence of many known components, identified novel subunits of AMPAR complexes, and together generated valuable data sets for the community. The dynamic composition of multiprotein complexes is generally poorly characterized and is a dimension that must be explored in order to understand functional differences of complexes in different contexts. This is especially important for protein complexes that are differentially expressed across brain regions and cell types, such as the AMPAR complex. Furthermore, aspects of AMPA receptor function are variable in different brain regions, cell types, and individual synapses, a feature that is likely determined by heterogeneity of AMPAR complex composition and posttranslational regulation. While it may be possible to infer likely brain region specificity for some components of the AMPAR complex from gene expression atlases and immunohistochemistry data, this approach comes with usual caveats of incomplete coverage, incomplete correlation between RNA and protein expression levels, and potential crossreactivity of antibodies. Critically, the composition of the complex in the context of all constituent components and associated protein interactions cannot be accurately predicted. Therefore, to fully understand the composition of a protein complex and heterogeneity of a complex in different contexts, it would be very desirable to directly measure the levels of all components in an unbiased way. In this issue of Neuron, Schwenk et al. (2014) explore the diversity of AMPAR complexes across different brain regions and through development using functional proteomics (Figure 1 ). Using an approach that they have previously employed (Schwenk et al., 2012Schwenk J. Harmel N. Brechet A. Zolles G. Berkefeld H. Müller C.S. Bildl W. Baehrens D. Hüber B. Kulik A. et al.Neuron. 2012; 74: 621-633Abstract Full Text Full Text PDF PubMed Scopus (300) Google Scholar) which utilizes antibodies directed at all four pore-forming subunits, the authors were able to purify the entire complement of AMPAR complexes from brain tissue samples. This is particularly important for comparing composition of complexes in different brain regions in which different combinations of subunits predominate. AMPAR complexes were affinity purified from membrane lysates from dissected rat brain regions (olfactory bulb, cortex, striatum, thalamus, hippocampus, brainstem, and cerebellum) and analyzed by nanoscale liquid chromatography tandem mass spectrometry (LC-MS/MS). In this approach, protein samples are enzymatically digested into short peptides, chromatographically separated, and delivered to a mass spectrometer by electrospray ionization. Mass/charge measurements of the intact peptides and of their fragments obtained via gas phase dissociation permit their identification and quantification and inference of respective proteins and their relative abundance across samples. The analysis of complexes purified from different brain regions was restricted to the set of 34 robustly enriched components of the AMPAR complex from their previous characterization of complexes purified from whole brains (Schwenk et al., 2012Schwenk J. Harmel N. Brechet A. Zolles G. Berkefeld H. Müller C.S. Bildl W. Baehrens D. Hüber B. Kulik A. et al.Neuron. 2012; 74: 621-633Abstract Full Text Full Text PDF PubMed Scopus (300) Google Scholar). It should be noted that this approach may have missed proteins only detectable when purified from specific brain regions that were not identified in complexes purified from whole brains due to dynamic range issues. A library of peptides derived from AMPAR complex constituents were expressed as QconCAT proteins, which are concatenated peptide sequences made into artificial proteins to serve as peptide quantification standards (Beynon et al., 2005Beynon R.J. Doherty M.K. Pratt J.M. Gaskell S.J. Nat. Methods. 2005; 2: 587-589Crossref PubMed Scopus (393) Google Scholar). This allowed the molecular abundance of each protein to be calculated, which would indicate the average stoichiometry of a given component across the pool of AMPARs expressed in a brain region. This permitted the density of the pore-forming subunits to be calculated for each tissue, with the cortex containing 50% of all AMPARs and the hippocampus containing 40%. Importantly, the combinations of pore-forming subunits displayed considerable regional specificity particularly between the hippocampus and the cerebellum and a differential use of the auxiliary subunits CNIH2 and TARPs. Strikingly, all 34 components displayed considerable variations in their profile of association with the AMPAR complex, an observation which clearly demonstrates the utility of the approach but also the potential for functional heterogeneity of AMPAR complexes in different brain regions, especially when combinatorial effects are considered. In order to derive relationships between the amounts of AMPAR complex proteins and specific assemblies of pore-forming subunits, the abundance values were normalized to the total number of GluA tetramers in each brain region. This normalized abundance value can be considered as the average stoichiometry of each complex constituent or the probability of incorporation into the complex. This showed considerable diversity in the subunit composition for the pore-forming and auxiliary subunits. In the hippocampus TARP γ-8 and CNIH2 are present in the complex at similar amounts and are the predominant auxiliary subunits, while in the cerebellum, TARP γ-2 and TARP γ-7 dominate. Interestingly, the profile of certain components appears to be correlated: GluRA4 precludes assembly of TARP γ-8, and assembly of CNIH2 is determined by the sum of GluA1 and GluA2. In all brain regions except the cerebellum and brainstem, the data supported an inner core architecture of four inner core components which bind at two separate pairs of sites. In the cerebellum and brainstem, the number of inner core components was two, indicating a differential composition of the complex. This was investigated in more detail using antibody shift assays on native gels which allow the composition of intact complexes to be inferred by the observation of a mobility shift caused by antibody binding to the complex. In the hippocampus, almost all complexes contain several GluA1 and/or GluA2 subunits, CNIH2 and TARP γ-8 and/or TARP γ-2/3, whilst in the cerebellum these subunits are found in a small fraction of complexes. Using serial affinity purifications, it was demonstrated that in the cerebellum, the predominant pore-forming subunit GluA4 mostly assembles into complexes with GluA1, and these complexes lack inner core components, confirming the reduced amount of these subunits observed in the cerebellum. While the resolution of protein complexes to the level of brain regions is very useful, it is clear that this still represents an average AMPAR complex, and in order to assess the extent of averaging, the authors scaled down their proteomic approach to analyses complexes purified from 200 μm micropunches from three subregions (stratum radiatum of the hippocampal CA3 region, the molecular layer of the cerebellum, and the nucleus accumbens in the striatum) that would contain fewer cell types. Impressively, they were able to identify and quantify all AMPAR complex components from these minute tissue samples and demonstrated that, by and large, complexes isolated from intact brain regions were similar to those purified from subregions. A number of exceptions were identified for non-pore-forming subunits, and it is very likely that the number of these exceptions would increase with increased tissue/cellular resolution. A recurring problem with affinity purification-mass spectrometry approaches is that it is not possible to differentiate between coassembly of proteins within the same complex and coexpression of complexes with different components. The authors have attempted to disentangle this by Pearson correlation analysis of their data and propose that robustly correlated proteins likely represent coassembly. Highly correlated sets of subunits included PRRT1 and CNIH2 and GluA3, GSG1l and CKAMP44 and DLG4. Negatively correlated proteins represent subunits that disfavor coassembly and included TARPs γ-5 and γ-7, which are negatively correlated with GluA2 but are highly correlated with GluR4. Finally, the same proteomic approach was used to quantify the relative composition of AMPAR complexes in whole rat brains through a four-stage developmental time course (P0–P3, P7, P14, and p > 28). This revealed an intriguing set of profiles that varied over time, with evidence of transient association with the complex and inverse profiles in which components were replaced by other related proteins during development. The pore-forming subunits, however, remained relatively constant through the developmental time course, although a small antiparallel relationship between GluA2 and GluA4 and GluA1 and GluA3 was evident. The present study is clearly an impressive demonstration of the power of state-of-the-art functional proteomics, and data sets generated will be a very useful resource to the neuroscience community, particularly for in-depth follow-up studies to correlate differential composition with differential function of AMPA receptors in different brain regions and cell types. The observation of AMPAR complex heterogeneity in brain regions and in development prompts many questions. To what extent are AMPAR complexes heterogeneous in different cell types? The use of epitope-tagged receptors expressed in a cell-specific manner might allow this question to be addressed. Are AMPAR complexes different between synapses or even between specific assemblies in the same synapse? If we were to investigate posttranslational modifications of AMPAR receptor components, to what extent would this extend heterogeneity? Hints of this can be seen in the phosphorylation site profiles generated in the present study, but it is very likely the combinatorial effect of posttranslational modifications such as phosphorylation, ubiquitination, palmitoylation, glycosylation, and acetylation, among others, would extend the opportunity for AMPAR complex heterogeneity very significantly. Ultimately, measurement of compositional and modification state dynamics of AMPAR complexes in response to neuronal activity would allow the functional significance of heterogeneity to be investigated. Regional Diversity and Developmental Dynamics of the AMPA-Receptor Proteome in the Mammalian BrainSchwenk et al.NeuronSeptember 18, 2014In BriefAMPA receptors are macromolecular complexes assembled from a pool of more than 30 proteins. In this study, Schwenk et al. describe the spatial and temporal dynamics in protein composition of this main excitatory neurotransmitter receptor in the adult and developing mammalian brain. Full-Text PDF Open Archive
Referência(s)