Abstract

Proteomic study of the synapse has generated an extensive list of molecular components, revealing one of the most complex functional systems currently known to cell biology. While fundamental to neural information processing, behaviour and disease, the molecular organisation of the synapse and its relation to higher-level function has yet to be clearly understood. Neurotransmitter receptor complexes, such as the N-methyl-D-aspartate receptor complex (NRC/MASC), are major components of the synaptic proteome. We have recently completed a detailed study of MASC, its functional organisation and involvement in behaviour and disease. This pointed to simple design principles underlying synaptic organisation. Drawing together the results of proteomic and analytical study, we sketch out a model for synaptic functional organisation.

INTRODUCTION

The human brain, with an estimated 1015 synapses interlinking a network of some 1012 neurons into an integrated neuronal circuit, is a prime example of one of the most complex biological entities: the mammalian brain. The cellular complexity of the brain, primarily observed with the light microscope, has in recent years been extended to the molecular level through the identification of molecular components of the synapse. Proteomic studies involving the large-scale isolation and identification of proteins have described ∼103 proteins in mammalian synapses. The human genome with ∼109 nucleotides and ∼104 genes encodes not only the components of the synapse and brain, but also their organization.

Evidence of an evolutionary origin for brain organization has been clear since the 19th century following comparative anatomical studies of the ontogeny and phylogeny of many nervous systems. There are not only distinct anatomical subregions within the brain (for example, hippocampus, cortex and cerebellum), but also a common pattern to their embryological origins and segmentation. In recent years, some of these anatomical features have been placed on a molecular footing with, for example, the identification of homoeotic genes regulating segmentation and body plan and molecular signalling pathways underlying self organization of cell fate in neuronal precursors.

We have studied the composition and organization of the synapse proteome for several reasons:

  • It is the most important structure for communication between nerve cells.

  • The neurotransmitter receptors and signal transduction machinery within the synapse respond to patterns of electrical activity and instigate biochemical changes in the nerve cell and in so doing modify the brain in response to behavioural experience.

  • Synapse proteins correspond to many human disease genes and drug targets for therapeutics that modulate cognitive illnesses.

  • Synapse proteomic studies have compiled a first draft of the protein composition of the synapse, revealing an unexpectedly high degree of molecular complexity.

In this article, we overview our study of the organization of the synapse proteome with an emphasis on relating form to function.

Functional regulation at the molecular level

The ability of the synapse to regulate its own properties (synaptic plasticity) lies at the heart of current efforts to elucidate the molecular basis of memory, learning, behaviour and mental illness. Control of molecular function occurs through the regulation of abundance, localization and functional activity.

Molecular abundance influences function in a fundamental way, determining the range of potential interactions and constraining their rate of occurrence. Different classes of molecule fluctuate over different time scales, encoding different types of information. While protein abundance and baseline ion and small molecule concentrations reflect relatively stable synaptic properties, the fluctuation of ion and small molecule concentrations and the rapid re-organization and post-translational modification of proteins may be utilized for signal transduction.

Localization tunes the functional behaviour of a set of molecules, narrowing the range of possible interactions and modulating their rate. This creates specialized functional subregions, such as the synapse itself. Protein–protein binding is perhaps the most precise and versatile form of localization. By linking together specific sets of proteins, either by direct interaction or through the mediation of scaffolding molecules, complex functional units may be created. These may in turn be tethered to specific locations through binding to cytoskeletal or membrane proteins. Physical interactions are highly structured. Distinct binding domains possess their own characteristic ligands, interaction with which may be dynamically regulated.

Through the dynamic balance of functional activity, focused by localization, synaptic properties are maintained, altered and information is processed. The regulation of molecular activity by reversible phosphorylation is central to many signalling pathways and known to play an important role in synaptic function and disease. Phosphorylation directly regulates channel activity, enzymatic activity and physical interactions, which in turn regulate kinase function and the construction of phosphorylation cascades.

Information processing and subsequent changes to synaptic function are the result of extensive dynamic feedback between molecular abundance, localization and functional activity. In our search for organizational structure, we first briefly review information from proteomic studies concerning the functional range of the synapse and the relative importance of functions within that repertoire. We then turn to the organization of synaptic proteins into functional complexes, focusing on those associated with the post-synaptic response to glutamate.

The post-synaptic proteome: functional repertoire

Several large-scale studies have now been made of the post-synaptic proteome, generally focusing on the post-synaptic density (PSD—a protein-dense layer underlying the post-synaptic membrane) and its components [1–8]. These studies were recently subjected to a combined analysis [8], which we now review. In total, 1124 PSD proteins have been identified, drawn from a number of functional classes (Table 1). These classes not only reflect specialization for intercellular signalling but also suggest a degree of autonomy unknown in other sub-cellular structures. The PSD possesses the molecular machinery of protein synthesis, degradation and transport, as well as containing the components of metabolic pathways. When compared to the proteome as a whole, the PSD was found to be enriched with protein domains associated with Ca2+- and GTP-dependent signalling, membrane localization, scaffolding and phosphorylation. These would appear to be the primary functional processes underlying synaptic signalling and its organization. Significantly, enrichment was greatest for SH3 and PDZ protein interaction domains that allow kinases to be bound into functional complexes.

Table 1:

Composition of PSD and protein complexes. Relative abundance of functional protein classes in the PSD and associated protein complexes. Data collated from [2, 9]

 %PSD %MASC %mGluR5 %AMPA 
Channels and receptors 26 56 
MAGUKs/adaptors/scaffolders 11 12 
Kinases 12 
Protein phosphatases 
G-proteins and modulators 10 12 
Signalling molecules and enzymes 25 21 11 
Transcription and translation 11 
Cell adhesion and cytoskeletal 14 19 18 33 
Synaptic vesicles/protein transport 14 12 18 11 
Uncharacterized/novel 10 
Other 
Total number of proteins 1124 186 66 
 %PSD %MASC %mGluR5 %AMPA 
Channels and receptors 26 56 
MAGUKs/adaptors/scaffolders 11 12 
Kinases 12 
Protein phosphatases 
G-proteins and modulators 10 12 
Signalling molecules and enzymes 25 21 11 
Transcription and translation 11 
Cell adhesion and cytoskeletal 14 19 18 33 
Synaptic vesicles/protein transport 14 12 18 11 
Uncharacterized/novel 10 
Other 
Total number of proteins 1124 186 66 

Protein complexes: molecular machines

The excitatory neurotransmitter glutamate activates post-synaptic receptors that can broadly be categorized into those transmitting the electrical depolarization (AMPA receptors) and those activating signalling and plasticity mechanisms (ionotropic NMDA receptor, metabotropic mGluR receptors). Proteomic profiling of glutamate receptors has identified and characterized several associated protein complexes embedded within the PSD [2, 8–11]. These fall into three classes: small AMPA complexes of nine proteins; mGluR5 complexes of 66 proteins (including NMDA receptor subunit NR2A); and large (2–3 MDa) complexes of 186 proteins, containing NMDA receptors, mGluR receptors, MAGUK proteins (SH3/PDZ domain scaffolders) and associated molecules, referred to here as NRC/MASC (NMDA receptor complex/MAGUK-associated signalling complex). We have recently completed a detailed investigation of MASC [9]. This has started to elucidate the relationships between molecular complexity, signal transduction, electrophysiology and disease, suggesting general principles underlying the organization of protein complexes.

We first subjected MASC proteins to systematic annotation, involving extensive literature searching and manual data curation. To investigate the involvement of proteins in biochemical pathways, they were annotated for structure and function. Enrichment of functional domains in MASC matched that of the PSD, with additional enrichment for Ca2+ binding domains. The functional roles of MASC proteins reflected the presence of diverse signalling pathways, suggesting their co-ordination within the complex.

To evaluate the role of the complex in synaptic function, behaviour and disease, an array of phenotypic data was collated from the scientific literature. This comprised of physiological data obtained from rodent studies, where mutations or drugs that specifically interfere with a given protein were tested for their effects on synapse electrophysiology or behaviour. Reports on the involvement of specific molecules in human diseases were also collated. Almost a quarter of MASC proteins were known to be essential for normal synaptic plasticity (without the function of these proteins, synaptic plasticity was impaired), with approximately the same number involved in rodent behaviour. Nearly a third of MASC proteins were implicated in human mental illness. Of these, more than twice as many were linked to cognitive (schizophrenia, mental retardation) than to affective disorders (bipolar disorder, depression). While the role of NMDA receptors in plasticity and learning was widely appreciated, it was not obvious that so many other proteins involved in synaptic plasticity, behaviour and disease would be brought into close association by the complex.

Functional, phenotypic and phylogenetic annotations were examined for significant molecular overlap, uncovering an extremely high degree of commonality between synaptic plasticity, behaviour and disease. Of the human diseases, schizophrenia in particular, possessed a highly significant overlap with both synaptic plasticity and behaviour. Linking these higher-level processes to molecular function, ionotropic signalling via the NMDA receptor was strongly associated with synaptic plasticity, behaviour and cognitive disorders (primarily schizophrenia) while G-protein-dependent signalling showed a contrasting association with affective disorders.

Overall, our analysis suggests that MASC is central to the post-synaptic processing of information encoded in neural activity, orchestrating cellular responses underlying synaptic plasticity. The complex as a whole appears to be responsible for the induction of synaptic plasticity, in particular hippocampal LTP/LTD, with MASC containing a large number of proteins implicated through study of CA3–CA1 synapses. We also see marked evidence of a common molecular foundation to synaptic plasticity, rodent behaviour and human mental illness—the complex being closely associated with schizophrenia.

We also investigated the organization of MASC proteins, looking at their assembly into functional complexes via protein–protein interactions. Curating high-quality interaction data from the literature we generated a network representation of the complex, in which proteins were represented as nodes and interactions as edges linking pairs of nodes. Examining its structure, we found that the average number of interactions separating any two proteins was very low, allowing for the rapid integration of information and coordination of responses. This suggests that functional roles are distributed over sets, or clusters of proteins within MASC. Subsequent clustering revealed that the network possessed a modular structure (Figure 1) with a clear functional logic (Figure 2). Clusters of proteins around ionotropic and metabotropic glutamate receptors (modules 1 and 2) form the primary sites for signal reception. These clusters may directly regulate effector mechanisms such as retrograde signalling (ionotropic) and vesicular trafficking (metabotropic, G-protein-dependent), although whether this represents the trafficking of other post-synaptic components (e.g. AMPA receptors) or the metabotropic signalling machinery itself is unclear. The primary signals from these clusters are modulated by other sources of information, both internal (e.g. small input/processing module 10, Figure 1) and external to the complex. The main body of proteins (module 3) integrates these disparate sources, co-ordinating common effector pathways (via modules 4, 5 and others) and providing regulatory feedback onto input modules. Intermediary output clusters 4 and 5 signal to more restricted, overlapping sets of effector mechanisms, while small clusters tend to be specific to individual functional processes. This integrated model fits well with classical reductionist studies of molecular plasticity, where output clusters such as the ERK pathway (module 4) are well known as important effectors of glutamate receptor mediated synaptic plasticity [13]. Mouse genetic studies have shown that distinct cognitive subprocesses (e.g. strategy choice, perception, learning) can be separated by mutations of different genes in NRC/MASC [14]; Cuthbert et al., in preparation). To a certain extent this is reflected in modularity of the MASC network: signal reception modules 1 and 2 are specialized for different streams of information, transmitted by distinct mechanisms (Ca2+- and GTP-dependent signalling), that map onto cognitive and affective disorders/processing, respectively.

Figure 1:

Network cluster analysis. Clustering of the largest connected component of the MASC network identified 13 clusters. Significant overlap with functional and phenotypic annotations is indicated for the largest clusters, 1–3. Clusters 4 and 5 both correspond to MAPK signalling pathways regulating various functional processes (briefly summarized). All five are followed by a brief descriptive phrase indicating their general functional role. Functional roles suggested by composition and interactions are indicated for the remaining clusters. Graphical representation of network produced using BioLayout [12]. Reprinted with permission from Molecular Systems Biology.

Figure 1:

Network cluster analysis. Clustering of the largest connected component of the MASC network identified 13 clusters. Significant overlap with functional and phenotypic annotations is indicated for the largest clusters, 1–3. Clusters 4 and 5 both correspond to MAPK signalling pathways regulating various functional processes (briefly summarized). All five are followed by a brief descriptive phrase indicating their general functional role. Functional roles suggested by composition and interactions are indicated for the remaining clusters. Graphical representation of network produced using BioLayout [12]. Reprinted with permission from Molecular Systems Biology.

Figure 2:

Modular structure and functional organization within MASC. MASC proteins are clustered into modules with well-defined functional roles. The size of each module reflects the number of proteins it contains. Primary signal reception modules (1 and 2) are formed around ionotropic and metabotropic receptors. These inputs are integrated within a large signal-processing module (3) responsible for overall co-ordination of functional processes. Other sources of input (‘other receptors’) may feed into this module directly, or through smaller input/processing modules (such as cluster 10, Figure 1). Note that within this general structure, individual modules may play multiple functional roles (e.g. regulation of effector mechanisms by input modules). In this way, information processing and regulation of effector pathways are distributed over multiple modules. The general principles underlying functional organization within MASC are apparent in the co-ordinated regulation of common downstream effector pathways: a single, large module (3) is responsible for overall co-ordination; several intermediate modules (4, 5 …) regulate overlapping sets of pathways; while numerous small modules are specific to individual effector responses. Note that this is not a simple feed-forward mechanism, rather a dynamic balance between multiple functional processes. The resulting synchronization of multiple cell-biological processes induces synaptic plasticity, manifest at a higher level of neurological function through behavioural learning. Numbering of the 5 largest clusters reflects that of Figure 1, as do the interactions between them (solid black lines). Internal/external modulation of MASC function and the regulation of effector mechanisms are denoted by dashed lines. The line between clusters 4 and 5 denotes the fact that other interactions (e.g. phosphorylation) play an important role in MASC function. Reprinted with permission from Molecular Systems Biology.

Figure 2:

Modular structure and functional organization within MASC. MASC proteins are clustered into modules with well-defined functional roles. The size of each module reflects the number of proteins it contains. Primary signal reception modules (1 and 2) are formed around ionotropic and metabotropic receptors. These inputs are integrated within a large signal-processing module (3) responsible for overall co-ordination of functional processes. Other sources of input (‘other receptors’) may feed into this module directly, or through smaller input/processing modules (such as cluster 10, Figure 1). Note that within this general structure, individual modules may play multiple functional roles (e.g. regulation of effector mechanisms by input modules). In this way, information processing and regulation of effector pathways are distributed over multiple modules. The general principles underlying functional organization within MASC are apparent in the co-ordinated regulation of common downstream effector pathways: a single, large module (3) is responsible for overall co-ordination; several intermediate modules (4, 5 …) regulate overlapping sets of pathways; while numerous small modules are specific to individual effector responses. Note that this is not a simple feed-forward mechanism, rather a dynamic balance between multiple functional processes. The resulting synchronization of multiple cell-biological processes induces synaptic plasticity, manifest at a higher level of neurological function through behavioural learning. Numbering of the 5 largest clusters reflects that of Figure 1, as do the interactions between them (solid black lines). Internal/external modulation of MASC function and the regulation of effector mechanisms are denoted by dashed lines. The line between clusters 4 and 5 denotes the fact that other interactions (e.g. phosphorylation) play an important role in MASC function. Reprinted with permission from Molecular Systems Biology.

Some general points are worth making concerning the structure of the MASC network. Roughly speaking, the larger the cluster the greater its functional influence. This is clearly illustrated by the coordination of common effector mechanisms in Figure 2. The size (number of proteins) of a module also correlates with its degree of interaction with other clusters—while the functional behaviour of large, influential modules (1–3) appear closely coordinated, small functionally specific modules are fairly independent of each other (Figure 1). The final point is that function is in turns both modular (clusters possess well-defined functional roles) and distributed (e.g. multiple clusters may control a single effector pathway). Each module integrates a particular set of inputs (either external, or internally processed signals) and influences a particular set of functional processes (e.g. other clusters, effector pathways). Stated another way, each module reflects the correlation between a particular set of functions in response to a specific range of stimuli. This implies that different sets of signals (e.g. different patterns of action potentials) are processed by different sets of modules, and that the relative importance of each module varies according to the information being processed.

Network structure also has implications for the effect of molecular perturbations. Synaptic plasticity is surprisingly robust, with disruption by mutation or drugs only partially impairing rather than completely abolishing plasticity in most cases (see e.g. [15–19]). This can partially be explained by functional processes being distributed over sets of molecules and, at a higher level, modules. Another potential source of robustness lies in the general pattern of connectivity within the network, which was found to follow a power–law distribution at the level of both molecules and clusters. This reflects the presence of a few highly connected nodes (molecules/clusters) that mediate interaction within the less well-connected bulk of the network. Such networks are structurally robust to random deletions, but fragile to targeted removal of highly interacting nodes [20]. Given that the more interactions a node has the more likely it is to influence multiple functional processes (e.g. effector mechanisms), this naturally extends to a correlation between node connectivity and severity of functional effect on disruption. This leads to a surprising prediction: if MASC controls the induction of synaptic plasticity, then a correlation between protein connectivity and extent of functional influence entails a correlation between protein connectivity and quantitative perturbation of LTP/LTD on disruption. Collating quantitative data on the perturbation of LTP/LTD caused by disruption of individual MASC proteins, we were able to confirm this prediction.

Preliminary model of synaptic organization

Physical interactions clearly play a major role in shaping the functional behaviour of protein complexes and their susceptibility to genetic and pharmacological disruption. Our analysis suggests three simple organizational principles underlying network structure: the number of interactions separating any pair of elements (proteins or clusters) is low; the connectivity of elements follows an approximately power–law distribution; and the connectivity of each element is correlated with the extent of its functional influence. At each level of organization, these principles imply that there are a few highly interacting elements responsible for overall coordination and numerous sparsely connected elements specific to individual functional processes. In between these extremes lies a range of elements through which particular sets of functional processes are coordinated. This structure mirrors the organization of PSD proteins into functional complexes, ranging from the small and highly specific (AMPA) to the large, highly structured NRC/MASC complexes that coordinated multiple signalling pathways.

When drawn together, the elements of our discussion suggest the following rough outline of large-scale synaptic organization. Post-synaptic signalling seems to be organized around five key functional processes: Ca2+-dependent signalling; GTP-dependent signalling; membrane localization; protein complex formation and phosphorylation. The molecular machinery of Ca2+ and GTP signalling is shaped by the organization of proteins into functional complexes. These complexes are constructed around core integral membrane proteins (channels, receptors) regulated by vesicular trafficking, and other proteins whose membrane localization is primarily controlled by Ca2+/phospholipid interactions. Direct binding and scaffolding control the clustering of these proteins and their association with cytosolic molecules to form functional signalling complexes. These complexes probably consist of a stable core bound to integral membrane proteins and a more dynamically regulated shell. Directed by these complexes, phosphorylation and GTP-dependent signalling cascades regulate numerous effector mechanisms, including phosphorylation-dependent changes in scaffolding interactions that re-shape the functional complexes themselves.

FUTURE DIRECTIONS

The data and models we have discussed represent a homogenized synapse seen through the lens of glutamatergic signalling and its associated MAGUK scaffold. While of interest, they are not without their limitations and these should be addressed in future studies. The sensitivity and accuracy of proteomic isolation and detection methods must be carefully considered when attempting to extrapolate from such data. The identification and refinement of a complementary set of proteomic methods will be required to complete a detailed map of the synaptic proteome.

The extent to which the complement of synaptic molecules varies between neuron types (and even between synapses in a single neuron) is currently unknown. Large-scale programmes characterizing the gene expression profiles of identified neuron types (e.g. GENSAT) will eventually provide a list of molecules that may be present in synapses of a specific neuron and further lead to an understanding of the transcriptional regulation of the synapse and its development. Refocusing the ‘lens’ through the use of different isolation techniques, in particular different molecular ‘hooks’ for co-immuoprecipitation studies (e.g. specific receptor/channel subunits), will result in different complexes being identified.

From an evolutionary perspective, the analysis of the molecular mechanics of the synapse is at an early stage. How subtle sequence changes interact over time to cause functional changes within complexes is hard to determine, yet their functional significance is likely to be critical to our understanding of cognition and its disorders. Such changes may alter molecular abundance, localization and functional activity, resulting in substantial differences in network architecture and its functional properties.

In the near term, there are some obvious and tractable studies to be done to refine even these relatively simple static network models. For the post-synaptic proteome as a whole, interaction mapping is highly incomplete and almost entirely unverified. Extensive and high quality datasets are essential for these studies. Ideally these data would come from multiple species, where the effects of sequence level events occurring in evolutionarily orthologous proteins can be related to possible changes in protein characteristics.

Ultimately, synaptic signalling will only be properly understood through the construction of dynamic models. Of fundamental importance to these models will be the characterization of phosphorylation-dependent changes in protein binding and functional activity.

Stock should also be taken of the extensive literature on synapse biology that already exists. The mapping of electrophysiological, behavioural and human disease data onto NRC/MASC complexes has given insight into their functional organization—a process that should be extended to the post-synaptic proteome as a whole. Although in its infancy, the publication of more structured behavioural, LTP and disease data (and its deposition into central databases) will alleviate current reliance on text mining and expert curation to produce these datasets.

Key Points

  • The organization of interactions within mammalian post-synaptic NRC/MASC multi-protein complexes follows simple design principles, whereby distributed functional processing is embodied in a modular network.

  • This network architecture appears central to the post-synaptic processing of information encoded in neural activity and orchestration of cellular responses underlying synaptic plasticity.

  • Extended to the synapse proteome as a whole, this suggests a new model for understanding the organization of molecular complexity and its relationship to synapse physiology, behaviour and disease.

  • Enrichment of the postsynaptic proteome highlights five key functional processes underlying synaptic signalling and its organization.

  • These processes suggest that the dynamic restructuring of protein complexes plays an important computational role.

Acknowledgements

A.J.P. was supported by the Medical Research Council (UK) through a Special Research Training Fellowship in Bioinformatics. S.G.N.G. and J.D.A. were supported by the Wellcome Trust Genes to Cognition programme. Thanks to Ms J.V. Turner for editorial assistance.

References

Walikonis
RS
Jensen
ON
Mann
M
, et al.  . 
Identification of proteins in the postsynaptic density fraction by mass spectrometry
J Neurosci
 , 
2000
, vol. 
20
 (pg. 
4069
-
80
)
Husi
H
Grant
SG
Isolation of 2000-kDa complexes of N-methyl-D-aspartate receptor and postsynaptic density 95 from mouse brain
J Neurochem
 , 
2001
, vol. 
77
 (pg. 
281
-
91
)
Yamauchi
T
Molecular constituents and phosphorylation-dependent regulation of the post-synaptic density
Mass Spectrom Rev
 , 
2002
, vol. 
21
 (pg. 
266
-
86
)
Yoshimura
Y
Yamauchi
Y
Shinkawa
T
, et al.  . 
Molecular constituents of the post-synaptic density fraction revealed by proteomic analysis using multidimensional liquid chromatography-tandem mass spectrometry
J Neurochem
 , 
2004
, vol. 
88
 (pg. 
759
-
68
)
Jordan
BA
Fernholz
BD
Boussac
M
, et al.  . 
Identification and verification of novel rodent postsynaptic density proteins
Mol Cell Proteomics
 , 
2004
, vol. 
3
 (pg. 
857
-
71
)
Peng
J
Kim
MJ
Cheng
D
, et al.  . 
Semiquantitative proteomic analysis of rat forebrain postsynaptic density fractions by mass spectrometry
J Biol Chem
 , 
2004
, vol. 
279
 (pg. 
21003
-
11
)
Li
KW
Hornshaw
MP
Van Der Schors
RC
, et al.  . 
Proteomic analysis of rat brain postsynaptic density. Implications of the diverse protein functional groups for the integration of synaptic physiology
J Biol Chem
 , 
2004
, vol. 
279
 (pg. 
987
-
1002
)
Collins
MO
Husi
H
Yu
L
, et al.  . 
Molecular characterization and comparison of the components and multi-protein complexes in the postsynaptic density
J Neurochemistry
 , 
2005
 
(in press)
Pocklington
AJ
Cumiskey
M
Armstrong
JD
, et al.  . 
Neurotransmitter receptor proteomes form modular networks with distributed functionality underlying plasticity and behaviour
Molecular Systems Biology
 , 
2006
 
doi:10.1038/msb4100041
Husi
H
Ward
MA
Choudhary
JS
, et al.  . 
Proteomic analysis of NMDA receptor-adhesion protein signaling complexes
Nat Neurosci
 , 
2000
, vol. 
3
 (pg. 
661
-
9
)
Farr
CD
Gafken
PR
Norbeck
AD
, et al.  . 
Proteomic analysis of native metabotropic glutamate receptor 5 protein complexes reveals novel molecular constituents
J Neurochem
 , 
2004
, vol. 
91
 (pg. 
438
-
50
)
Enright
AJ
Ouzounis
CA
BioLayout – an automatic graph layout algorithm for similarity visualization
Bioinformatics
 , 
2001
, vol. 
17
 (pg. 
853
-
4
)
Sweatt
JD
Mitogen-activated protein kinases in synaptic plasticity and memory
Curr Opin Neurobiol
 , 
2004
, vol. 
14
 (pg. 
311
-
7
)
Migaud
M
Charlesworth
P
Dempster
M
, et al.  . 
Enhanced long-term potentiation and impaired learning in mice with mutant postsynaptic density-95 protein
Nature
 , 
1998
, vol. 
396
 (pg. 
433
-
9
)
Grant
SG
O'Dell
TJ
Karl
KA
, et al.  . 
Impaired long-term potentiation, spatial learning, and hippocampal development in fyn mutant mice
Science
 , 
1992
, vol. 
258
 (pg. 
1903
-
10
)
Watabe
AM
Zaki
PA
O'Dell
TJ
Coactivation of beta-adrenergic and cholinergic receptors enhances the induction of long-term potentiation and synergistically activates mitogen-activated protein kinase in the hippocampal CA1 region
J Neurosci
 , 
2000
, vol. 
20
 (pg. 
5924
-
31
)
Komiyama
NH
Watabe
AM
Carlisle
HJ
, et al.  . 
SynGAP regulates ERK/MAPK signaling, synaptic plasticity, and learning in the complex with postsynaptic density 95 and NMDA receptor
J Neurosci
 , 
2002
, vol. 
22
 (pg. 
9721
-
32
)
Opazo
P
Watabe
AM
Grant
SG
, et al.  . 
Phosphatidylinositol 3-kinase regulates the induction of long-term potentiation through extracellular signal-related kinase-independent mechanisms
J Neurosci
 , 
2003
, vol. 
23
 (pg. 
3679
-
88
)
Yasuda
H
Barth
AL
Stellwagen
D
, et al.  . 
A developmental switch in the signaling cascades for LTP induction
Nature Neuroscience
 , 
2003
, vol. 
6
 (pg. 
15
-
6
)
Albert
R
Jeong
H
Barabasi
AL
Error and attack tolerance of complex networks
Nature
 , 
2000
, vol. 
406
 (pg. 
378
-
82
)