Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Aging-Induced Proteostatic Changes in the Rat Hippocampus Identify ARP3, NEB2 and BRAG2 as a Molecular Circuitry for Cognitive Impairment

  • Philipp Ottis,

    Affiliation Department of Neuropathology, Heinrich Heine University of Düsseldorf, Düsseldorf, Germany

  • Bianca Topic,

    Affiliation Center for Behavioral Neuroscience, Department Experimental Psychology, Heinrich Heine University of Düsseldorf, Düsseldorf, Germany

  • Maarten Loos,

    Affiliations Department of Molecular and Cellular Neurobiology, Faculty of Earth and Life Sciences, Center for Neurogenomics and Cognitive Research, Neuroscience Campus Amsterdam, VU University, Amsterdam, The Netherlands, Synaptologics B.V., Amsterdam, The Netherlands

  • Ka Wan Li,

    Affiliation Department of Molecular and Cellular Neurobiology, Faculty of Earth and Life Sciences, Center for Neurogenomics and Cognitive Research, Neuroscience Campus Amsterdam, VU University, Amsterdam, The Netherlands

  • Angelica de Souza,

    Affiliation Center for Behavioral Neuroscience, Department Experimental Psychology, Heinrich Heine University of Düsseldorf, Düsseldorf, Germany

  • Daniela Schulz,

    Current address: Department of Neurobiology and Behavior, Stony Brook University, Stony Brook, New York, USA

    Affiliation Center for Behavioral Neuroscience, Department Experimental Psychology, Heinrich Heine University of Düsseldorf, Düsseldorf, Germany

  • August B. Smit,

    Affiliation Department of Molecular and Cellular Neurobiology, Faculty of Earth and Life Sciences, Center for Neurogenomics and Cognitive Research, Neuroscience Campus Amsterdam, VU University, Amsterdam, The Netherlands

  • Joseph P. Huston,

    Affiliation Center for Behavioral Neuroscience, Department Experimental Psychology, Heinrich Heine University of Düsseldorf, Düsseldorf, Germany

  • Carsten Korth

    ckorth@uni-duesseldorf.de

    Affiliation Department of Neuropathology, Heinrich Heine University of Düsseldorf, Düsseldorf, Germany

Aging-Induced Proteostatic Changes in the Rat Hippocampus Identify ARP3, NEB2 and BRAG2 as a Molecular Circuitry for Cognitive Impairment

  • Philipp Ottis, 
  • Bianca Topic, 
  • Maarten Loos, 
  • Ka Wan Li, 
  • Angelica de Souza, 
  • Daniela Schulz, 
  • August B. Smit, 
  • Joseph P. Huston, 
  • Carsten Korth
PLOS
x

Correction

9 Oct 2013: Ottis P, Topic B, Loos M, Li KW, de Souza Silva MA, et al. (2013) Correction: Aging-Induced Proteostatic Changes in the Rat Hippocampus Identify ARP3, NEB2 and BRAG2 as a Molecular Circuitry for Cognitive Impairment. PLOS ONE 8(10): 10.1371/annotation/2546f4c2-b07f-4450-8d3c-f3ee81127b4a. https://doi.org/10.1371/annotation/2546f4c2-b07f-4450-8d3c-f3ee81127b4a View correction

Abstract

Disturbed proteostasis as a particular phenotype of the aging organism has been advanced in C. elegans experiments and is also conceived to underlie neurodegenerative diseases in humans. Here, we investigated whether particular changes in non-disease related proteostasis can be identified in the aged mammalian brain, and whether a particular signature of aberrant proteostasis is related to behavioral performance of learning and memory. Young (adult, n = 30) and aged (2 years, n = 50) Wistar rats were tested in the Morris Water Maze (MWM) to distinguish superior and inferior performers. For both young and old rats, the best and worst performers in the MWM were selected and the insoluble proteome, termed aggregome, was purified from the hippocampus as evidence for aberrant proteostasis. Quantitative proteomics (iTRAQ) was performed. The aged inferior performers were considered as a model for spontaneous, age-associated cognitive impairment. Whereas variability of the insoluble proteome increased with age, absolute changes in the levels of insoluble proteins were small compared to the findings in the whole C. elegans insoluble proteome. However, we identified proteins with aberrant proteostasis in aging. For the cognitively impaired rats, we identified a changed molecular circuitry of proteins selectively involved in F-actin remodeling, synapse building and long-term depression: actin related protein 3 (ARP3), neurabin II (NEB2) and IQ motif and SEC7 domain-containing protein 1 (BRAG2). We demonstrate that aberrant proteostasis is a specific phenotype of brain aging in mammals. We identify a distinct molecular circuitry where changes in proteostasis are characteristic for poor learning and memory performance in the wild type, aged rat. Our findings 1. establish the search for aberrant proteostasis as a successful strategy to identify neuronal dysfunction in deficient cognitive behavior, 2. reveal a previously unknown functional network of proteins (ARP3, NEB2, BRAG2) involved in age-associated cognitive dysfunction.

Introduction

The unraveling of the specific 'pathophysiology' of natural, non disease-associated brain aging is only emerging. Whereas general principles of cellular aging like telomere shortening [1], mitochondrial dysfunction leading to increased intracellular oxidative stress [2], or the involvement of insulin/IGF-1 (insulin-like growth factor 1)-like signaling [3] are well established, the specific molecular features of cellular aging in post-mitotic neurons of the brain are still not well understood.

Changes in protein homeostasis (proteostasis), i.e. the orderly life cycle of synthesis and degradation of proteins, have been described for the aged mammalian brain in terms of gene expression [4], epigenetic changes [5], and protein composition (reviewed by VanGuilder and Freeman in 2011 [6]). Proteomic changes comparing aged and young rodents primarily have been assigned to cellular processes such as glucose metabolism [7,8,9,10,11,12], signal transduction [7,8,9,10,11,13], oxidative stress [9,13], and cytostructure regulation [8,12]. Changes in the expression of proteins that are involved in synaptic processes appear to be more specific to changes in cognition rather than aging [14,15,16,17]. Peter Douglas and Andrew Dillin reviewed the potential effects of age-associated proteostasis changes on neuronal health [18].

Studies on the nematode Caenorhabditis elegans [19,20] revealed increases in the overall content of insoluble proteins with age in the whole organism. Subsequent experiments demonstrated that RNAi knockdown of some of the identified insoluble proteins increased the worms’ lifespan [19]. This suggested that a decreased clearance of insoluble proteins may contribute to age-associated pathophysiology.

These findings also indicated that, at least in C. elegans, the mechanisms for quality control in proteostasis undergo an age-associated decline independent of any disease. Consequently, the analysis of specific proteins accumulating as a result of clearance dysfunction may reveal insights into the cellular mechanisms of neuronal aging, and provide potential targets for therapeutic intervention. In humans, one phenotype associated with brain aging is mild cognitive impairment (MCI) [21,22]. MCI is defined as a decrease in cognitive abilities in elderly subjects that is clearly discernible but not yet interfering with tasks of daily life [22]. As such, this condition often precedes Alzheimer’s disease [23,24].

As humans and animals age, individual differences become apparent across various behavioral domains [25,26]. While some aged subjects maintain performance levels comparable to that of young ones, termed successful or healthy aging [27,28], a fraction of aged individuals show impaired performances. This spontaneously arising increase in variability, found in aged cohorts of outbred rat strains, has been extensively studied in relation to learning and memory in aged rodents (e.g. [29,30,31,32]).

Here, the Morris water maze (MWM) represents a widely used task to assess individual differences in aging-related decrements in spatial learning and memory, with the spatial performance in this task being dependent on the functional integrity of the hippocampus [33].

In this study we aimed to answer 1. whether aging-associated disturbances in proteostasis, reflected by the segregation of distinct proteins into the insoluble proteome, can be observed in the rat brain, and 2. whether such segregation of certain insoluble proteins is functionally related to changes in learning and memory in aged rats.

Using quantitative proteomics, we demonstrate an age-dependent change in hippocampal proteostasis, indicated by compositional changes in the insoluble proteome, termed aggregome. Focussing on changes specific for the aged rats showing impaired performance compared to their age-matched superior performers, we identify a molecular circuitry related to synaptic plasticity.

Materials and Methods

Animals

The present study continues a report of results presented in Schulz et al. (2007) [34] using a subset of the same animals. Drug-naive adult (n = 30; 3 months) and aged (n = 50, 24 months) male outbred Wistar rats were obtained from the animal facilities of the University of Düsseldorf and were maintained under a reversed light-dark cycle (lights went on at 7: 00 pm for 12 h). They were housed in standard Macrolon cages of Type IV in groups of 2-3 old or 5 young animals per cage, and had free access to water and standard laboratory chow (Ssniff Spezialdiät). Over a period of three months the animals were behaviorally characterized by assessing their performance in the open field test, black-white box, elevated plus maze as well as learning and extinction trials in the Morris water maze (MWM). Behavioral testing was conducted during the dark period between 09:00 am and 06:00 pm and took place every 48 h (see also Schulz et al., 2007 [34]). All experiments were carried out in accordance with and approved by the German Animal Protection Law (Bezirksregierung Düsseldorf), as well as National and European Regulations.

Morris water maze (MWM)

The procedure, experimental design and water maze apparatus have been described in detail elsewhere [34]. Briefly, the water maze consisted of a black circular swimming pool made of polyethylene that was filled with water (20 ± 1 °C) up to a depth of 30 cm. The diameter of the pool was 185 cm. For the cued version of the water maze a 0.5 cm diameter metal peg (height: 22 cm) with black and white stripes was fixed onto the circular platform (18 cm in diameter, 1.5 cm under the water surface level) with a clip and tagged with vanilla aroma. Within one day, rats were released into the water maze for four successive trials with the platform cued. If a rat did not escape onto the platform within 1.5 min, it was gently guided to it by the experimenter. Two days later, the animals were trained in the hidden platform place version of the water maze for 6 days with two training trials per day (one in the morning and one in the afternoon). During this phase, the platform was fixed 1.5 cm below the water surface in the center of one quadrant of four equally large virtual quadrants of the maze. The platform location was randomly varied between all rats, but was maintained in a fixed location for a given rat during each task. A trial ended either when a rat escaped onto the hidden platform, or after 2.5 min had elapsed. After each training trial in the water maze, the animals were dried under a red-light heating lamp, before being returned to their home cages.

The behavioral analysis during the acquisition trials for each rat comprised the distance to platform (cm) and the time spent within the platform quadrant (PQ, expressed as percentage of total trial duration) as well as the swimming speed (cm/sec), which were automatically recorded via the EthoVision tracking software (Version 3, Noldus, Wageningen, The Netherlands).

Since 11 animals (10 aged and 1 adult) exhibited obvious signs of physical weakness (such as body tumors, eye infection) during the course of experimental testing, their data were excluded from the behavioral analyses, resulting in aged: n = 40 and adults: n = 29.

Clustering of animals according to their learning abilities

In the present report, we examined whether subgroups of superior and inferior learners in the water maze also exhibited differences in the insoluble hippocampal proteome. For classification into superior vs. inferior learners, we calculated the mean distance to the platform over all hidden platform place-learning trials for each rat to establish an overall score of learning performance. In resemblance to other studies on individual differences in learning and memory [30,35,36] the animals were ranked according to their overall score. This was done separately for each age group. Of each age group, 8 of the best performers were assigned to the "superior" group and 8 of the worst performers to the "inferior" group. However, the 15% of animals on each extreme side of the median were excluded in order to omit animals suggestive of any possibly unidentified disease or motoric disabilities, and also in order to reduce sampling bias.

Statistical analysis of behavioral data

Data are presented as mean ± standard error of the mean (SEM). For the water maze, the mean curve level A0, assessed as total mean distance to platform, as well as the time spent in the platform quadrant (expressed as percentage of trial duration) was taken as an index of the average performance, and the linear trend component a1, describing the slope of the curve, was calculated as an estimate of the rate of behavioral change over the course of training [37]. Three-way repeated measures analyses of variance (ANOVAs) with pairwise multiple comparisons using Bonferroni adjustment were conducted for statistical analysis, with ‘age’ and ‘learning performance’ (superior and inferior learners) as between-groups factors and days as the repeated measures factor for the water maze acquisition (computed on mean of trials per day) data. For the cued version of the water maze, trials were used as the repeated measures factor. When appropriate, t-tests for independent groups were carried out using the A0 and a1 values to determine differences between superior and inferior learners within each age group and also superior and inferior learners between the age groups. The level of significance was set to p ≤ 0.05.

Tissue and protein extraction

From a total of 32 animals comprising four groups (8 ‘adult inferior’, 8 ‘adult superior’, 8 ‘aged inferior’, and 8 ‘aged superior’) hippocampi were dissected, homogenized and aliquoted. Each rat hippocampus was processed individually in a blinded approach and in random order. One aliquot each, representing 20 mg of tissue, subjected to purification of detergent-insoluble proteins according to a protocol modified from Leliveld et al., 2008 [38] and Ottis et al., 2011 [39]. Briefly, the homogenates, supplemented with 2 mM phenylmethylsulfonyl fluoride (PMSF) and 1 x cOmplete, EDTA-free Protease Inhibitor Cocktail (Roche Applied Science, Mannheim, Germany), were incubated overnight at 4 °C in the presence of DNase I to degrade all DNA present in the sample. The next day, the aggregome was purified via three subsequent ultracentrifugation steps, each at 100.000 x g. Two subsequent centrifugation runs were carried out in a high-density sucrose buffer (1.1 M and 1.6 M), followed by one step in high salt buffer (1.5 M NaCl). Apart from the high salt treatment, all steps were carried out in the presence of 1.0% nonidet-P40 (NP-40) and 0.2% N-lauroylsarkosine (sarkosyl). Subsequently, the resulting pellet was washed twice in HEPES-buffer (50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, pH 7.5) to remove salts and detergents incompatible with iTRAQ-experiments. All centrifugation steps were carried out using a TLA-55 rotor and 1.5 mL ultracentrifugation tubes (Beckman Coulter, Krefeld, Germany).

Labeling for quantitative mass spectrometry

The pellets were dried in a speedvac and proteins were then denatured in 30 µL of 0.5 M triethyl ammonium bicarbonate (TEAB), 6 M urea, 0.8% RapiGest SF Surfactant (Waters Corporation, Milford, MA) and subjected to vigorous shaking at 25 °C for 10 min. After addition of 2 µL reducing reagent (supplied with the iTRAQ Reagent-Multiplex Buffer Kit; AB SCIEX, Darmstadt, Germany), samples were incubated at 27 °C for 2 hours with alternating shaking and pause intervals of 10 sec and 1 min, respectively. Subsequently, the reduced, free cysteines were blocked by addition of 1 µL of cysteine-blocking reagent (AB SCIEX, Darmstadt, Germany) and incubation with shaking at 25 °C for 10 min. For tryptic digestion, hydrolyzed, sequencing grade modified trypsin (Promega, Mannheim, Germany) was re-solubilized in 0.5 M TEAB to 0.4 mg/mL and 10 µL were added to each protein sample, followed by incubation at 37 °C overnight. Next day, 80 µL of HPLC-grade 2-propanol were added to each of the digests and tubes were vortexed briefly.

The denatured, reduced, blocked and digested peptides were labeled using 8-plex isobaric tagging for relative and absolute quantification (iTRAQ) [40]. To enable comparison of all 32 samples in 5 separate 8-plex analyses, 5/32 (19 µL) of each sample were pooled, split in 5 equal aliquots of 104 µL each, and were treated alongside the other samples to serve as internal standards of all 5 experimental sets (each standard labeled with iTRAQ-reagent 121).

After addition of 1 u iTRAQ-reagent per sample, tubes were incubated at 25 °C, under constant agitation, for 2 h. Following this, labeled samples of each 8-plex experimental set were pooled and centrifuged at 18.000 x g for 5 min. Supernatants were transferred to fresh tubes and pH was adjusted to 3.0-3.5 using HPLC-grade 5% trifluoroacetic acid (TFA). Tubes were incubated under constant agitation at 25 °C for 30 min before being dried in a speedvac overnight.

Peptide separation and mass spectrometry

Labeled samples were prepared for mass spectrometric analysis as described previously [41]. Briefly, the 5 mixtures were subjected to two-dimensional liquid chromatography (LC). Multiple LC fractions of iTRAQ labeled peptides were captured, mixed with matrix and every two consecutive LC fractions deposited as 192 spots on a single MALDI plate. Mass spectrometry (MS/MS) was performed to identify peptides and quantify the iTRAQ signal using an ABI 4800 proteomics analyzer (Applied Biosystems). This procedure was repeated 5 times for all the biological independent tissue samples.

Protein identification and statistical analyses

MS/MS spectra were searched against rat database using GPS Explorer (ABI) and Mascot (MatrixScience) with trypsin specificity and fixed iTRAQ modifications at lysine residues and N-termini of the peptides. Mass tolerance was 100 ppm for precursor ions and 0.5 Da for fragment ions; missed cleavage was allowed. For each MS/MS spectrum, a single peptide hit with the highest Mascot score in the Swissprot database (version 11/2011) was considered for further analysis. If a spectrum could not be annotated using Swissprot database, a second Mascot search was performed in the larger but more redundant NCBI database (version 11/28/2011). Next, the precursor protein sequences of all peptides from all 5 sets of samples were retrieved from the respective databases. NCBI sequences sharing more than 90% similarity over 85% of the sequence length with a Swissprot sequence were clustered as a single protein. Peptides that matched the sequence of multiple proteins were not removed from analyses. Proteins were included in analyses if at least 2 peptides had been identified with Mascot confidence > 95%, in addition to the criterion of at least 3 reliably quantified peptides (iTRAQ signature peak area above 500) in each of the 5 sets of samples. Individual peak areas of each iTRAQ signature peak of each peptide were log2-transformed, normalized to the average of all peak areas of the respective iTRAQ signature peak and mean centered. Within each set of samples, the abundance of a protein was calculated as the average of these mean centered iTRAQ values of multiple peptides, yielding 8 independent measurements of protein abundance for each of the 4 treatments.

Significance of treatment effects was evaluated using Student’s t-tests. To address the problem of multiple testing, resulting p-values were converted into q-values [42] giving an estimate of the false discovery rate (FDR) for each statistical test.

Data analyses

Gene-Ontology (GO) – Analyses were performed using the online DAVID Bioinformatics Resources 6.7 tool [43,44] with a subset of rat hippocampal gene expression, provided by the Gene Expression Atlas of the European Bioinformatics institute [45,46], as background. P-values stated were calculated by the DAVID tool and were corrected according to Bonferroni.

F-actin precipitation assay

The assay was adapted from Cenni et al., 2003 [47]. Briefly, human NLF neuroblastoma cells, grown to confluency and supplemented (30 min before lysis) with either DMSO only, 1 µM, or 5 µM of Mycalolide B (Santa Cruz Biotechnology, Inc., Heidelberg, Germany), an F-actin depolymerizing substance [48]. Cells were harvested from a 10 cm culture dish by trypsination and, subsequently, were washed twice with PBS. Then 0.4 mL of lysis- and F-actin stabilization buffer (50 mM Pipes, pH 6.9; 50 mM NaCl; 5 mM MgCl2; 5 mM EGTA; 5% [v/v] glycerol; 0.1% NP-40; 0.1% Triton X-100; 1% Tween 20; 1 x cOmplete, EDTA-free Protease Inhibitor Cocktail) was added to the pelleted cells and cells were lysed by pipetting repeatedly through a 200 µL pipette-tip. After incubating the lysate for 10 min at 37 °C, a 100 µL aliquot was centrifuged at 350 x g for 5 min to pellet any cell debris. The supernatant was then transferred into a 1.5 mL ultracentrifuge tube and was subjected to centrifugation at 100.000 x g for 60 min. The supernatant containing soluble proteins and unpolymerized G-actin was carefully removed and the pellet was re-solubilized in the same volume (100 µL). Both fractions were supplemented with SDS-loading buffer and were subjected to SDS-PAGE and subsequent immuno-blotting using an α-actin antibody (A2066; Sigma-Aldrich, Munich, Germany) and an α-ARP3 antibody (ab49671; Abcam, Cambridge, UK).

Results

Behavioral tests

Wild-type, adult and aged, male Wistar rats were characterized for their cognitive functions in the Morris-Water-Maze (MWM). For the cued version of the water maze, repeated measures ANOVA revealed a significant decrease in the distance covered to the cued platform over the four trials (F3,84 = 12.725, p ≤ 0.001), indicative of learning, but failed to reveal significant main effects for ‘age’ (F1,28 = 0.436, p = 0.515) or ‘learning performance’ (F1,25 = 0.018, p = 0.893; data not shown). These results indicate that sensory and motor deficits did not affect the groups differentially. Thus, group differences detected in the hidden platform task are rather attributable to differences in special learning capacities, than to sensory and motor deficits [35,49].

Results of the hidden platform task are summarized in Figure 1. Repeated measures ANOVA revealed a significant decrease in the mean distance covered to platform over the course of testing (F5,140 = 11.862, p ≤ 0.001), which significantly varied as a function of ‘learning performance’(F5,140 = 3.720, p = 0.003), but only marginally as a function of ‘age’ (F5,140 = 2.251, p = 0.053). Main effects were found for ‘age’ (F1,28 = 44.013, p ≤ 0.001) and ‘learning performance’ (F1,28 = 309.654, p ≤ 0.001). Also an interaction between ‘age’ and ‘learning performance’ became apparent (F1,28 = 5.727, p = 0.029).

thumbnail
Figure 1. Cognitive performance as distance to platform during hidden platform place learning in the Morris water maze.

Shown is the distance to the platform [cm] (+ SEM) for each acquisition day as averaged for adult superior (full triangles) and adult inferior (full circles) (A) as well as aged superior (open triangles) and aged inferior (open circles) (B) learners. C) Depicted is the mean curve level of the distance to platform (+ SEM) for the adult and aged rats as well as their subgroups adult superior, adult inferior, aged superior, and aged inferior learners (*, **, ***: p ≤ 0.05, 0.01, 0.001).

https://doi.org/10.1371/journal.pone.0075112.g001

Post-hoc analysis using t-tests showed that within each age group, the inferior performers, on average, moved longer distances to the hidden platform as compared to the superior performers (aged: t = -20.409; p ≤ 0.001; adult: t = -8.646; p ≤ 0.001). Furthermore, the aged inferior performers moved longer distances as compared to the adult inferior group (t = 4.725; p ≤ 0.001). Similarly, the aged superior group moved longer distances as compared to the adult superior group (t = -6.578; p ≤ 0.001). As to the slope (a1; data not depicted) over the hidden platform training trials, the aged inferior performers exhibited a stronger decrease in the distance to the hidden platform (and therefore a steeper slope a1) as compared to the aged superior group (t = 3.772; p = 0.002). However, such effects were not found when the adult inferior and the adult superior group were compared (t = 1.565; p = 0.140). Similarly, the aged superior performers did not differ from the adult superior performers in the slope a1 (t = -1.207; p = 0.247), but the aged inferior performers exhibited a steeper slope over the acquisition trials as compared to the respective adult performer group (t = -2.321; p = 0.036).

For the time spent in the platform quadrant (Figure 2), repeated measures ANOVA revealed a significant increase in the preference for the platform quadrant over the course of acquisition training (F5,140 = 4.688, p = 0.001), without any significant interaction effects (all F5,140 ≤ 1.612, all p ≥ 0.161). However, significant main effects were found for the factors ‘age’ (F1,28 = 18.131, p ≤ 0.001) and ‘learning performance’ (F1,28 = 17.315, p ≤ 0.001), indicating, that the groups also differed with respect to another direct measure of spatial learning, a preference for the reinforced platform quadrant. Post-hoc analysis using t-tests revealed that both superior performer groups exhibited a higher preference for the platform quadrant as compared to the respective inferior group of matched age (aged: t = 2.232, p = 0.036; adult: t = 4.649, p ≤ 0.001; Figure 2). Furthermore, adult inferior animals exhibited a stronger preference for the platform quadrant as compared to the aged inferior animals (t = -3.5, p = 0.004). Similarly, the adult superior rats showed a stronger preference for the platform quadrant as compared to the aged superior rats (t = -2.684, p = 0.018).

thumbnail
Figure 2. Cognitive performance place preference during hidden platform place learning in the Morris water maze.

Shown is the time spent in the platform quadrant expressed as percentage of trial duration (+ SEM) for each acquisition day displayed by adult superior (full triangles) and adult inferior (full circles) (A) as well as aged superior (open triangles) and aged inferior (open circles) (B) learners. C) Depicted is the mean curve level [%] of the place preference (+ SEM) for adult and aged rats as well as adult superior, adult inferior, aged superior, and aged inferior learners (*, **, ***: p ≤ 0.05, 0.01, 0.001).

https://doi.org/10.1371/journal.pone.0075112.g002

Repeated measures ANOVA revealed a significant decrease in speed of swimming over hidden platform place learning (F5,140 = 7.025, p ≤ 0.001), without any significant interaction effects (all F5,140 ≤ 2.046, all p ≥ 0.076). Furthermore, no significant main between-groups effects and interactions with the between-groups factors were revealed by ANOVA (all F1,28 ≤ 2.698, all p ≥ 0.112), ruling out the possibility that differences found during acquisition could be explained by differences in swimming speed.

Proteomic analysis

Brains from the 32 rats, each assigned to one of the four clusters ‘adult inferior’, ‘adult superior’, ‘aged inferior’, and ‘aged superior’ (Figure 3 A), were subjected to proteomic analysis. The hippocampus was selected as critical brain region since it is established to be essential for spatial memory [50] and is, therefore, an obvious brain region where molecular changes relating to impaired cognitive function, as assessed in MWM, could be present. For each of the four groups, hippocampal homogenates of the detergent-insoluble proteomes of 8 individual rats were purified to yield 32 aggregomes for quantitative mass-spectrometric analyses (Figure 3 B).

thumbnail
Figure 3. Grouping of animals and workflow.

The left panel (A) displays the applied grouping of the rats according to their age and learning abilities, resulting in four groups of 8 individuals, each. The right panel (B) depicts the workflow of the insoluble proteome purification from rat hippocampal homogenate (red: insoluble protein components; P: pellet; SN: supernatant).

https://doi.org/10.1371/journal.pone.0075112.g003

Behavioral performance-specific hippocampal aggregome

Only three proteins differed between aged superior and aged inferior learners in their segregation to the insoluble fraction (p < 0.05; Table 1). These proteins were identified as: Actin-related protein 3 (ARP3), spinophilin (neurabin-2, NEB2), and the IQ motif and SEC 7 domain-containing protein 1 (BRAG2). Most intriguingly, all three play crucial roles in synaptic plasticity, a phenomenon involved in memory formation [51,52,53] (Figure 4). We observed a decrease of insoluble ARP3 and NEB2, and an increase of insoluble BRAG2 in the aged inferior cohort, as compared to the group of aged superior animals.

Pearson Correlation
aged inferior vs. aged superiorA0-distance (MWM)a1-slope (MWM)
EntryProtein nameChangep-valueFDRrprp
ARP3_RATActin-related protein 30.00130.2156-0.720.0020.670.005
NEB2_RATNeurabin-20.00540.4662-0.620.0100.470.066
gi|109473862PREDICTED: IQ motif and Sec7 domain 1-like isoform 2 (Brag2)0.03100.93280.530.035-0.730.001
G3P_RATGlyceraldehyde-3-phosphate dehydrogenase0.05470.93280.450.080-0.550.027
MYH10_RATMyosin-100.05700.9328-0.490.0540.030.912
SHAN3_RATSH3 and multiple ankyrin repeat domains protein 30.07290.9328-0.350.1840.310.243
ACTB_RATActin, cytoplasmic 10.08410.9328-0.430.0960.380.147
DYHC1_RATCytoplasmic dynein 1 heavy chain 10.08610.93280.400.125-0.330.212
AGAP2_RATArf-GAP, GTPase, ANK repeat and PH domain-containing protein0.09650.93280.460.073-0.470.066
SYGP1_RATRas GTPase-activating protein SynGAP0.10080.93280.430.096-0.580.019

Table 1. Top 10 proteins altered comparing the insoluble proteome of aged inferior and aged superior rats.

Values in bold are significant. Arrow down: Lower abundance in aged inferior. Arrow up: Higher abundance in aged inferior. p: p-value. r: Pearson’s correlation coefficient. MWM: Morris Water Maze. A0-distance: Total distance swam to hidden platform. a1-slope: Slope of the A0 performance curve; learning pace.
CSV
Download CSV
thumbnail
Figure 4. Structural plasticity in NMDA receptor-mediated long-term depression.

Depicted is a scheme of NMDA receptor-induced LTD, bringing in context the presumably concerted actions of ARP3 – localized in the ARP2/3 complex; NEB2 – being activated via phosphorylation by CAMKII and acting on catabolic and anabolic processes on the filamentous actin network; and BRAG2 – binding to the AMPA receptor-GluA2 subunit upon its de-phosphorylation by the mGluR-activated protein tyrosine phosphatase (PTP) and inducing modulations of the actin cytoskeleton and internalization of AMPARs via its interaction with ARF6 [53,95]. Scheme based on Okamoto et al., 2009 [52], linking it to LTD [100] and including information concerning BRAG2 described by Scholz et al., 2010 [53] and summarized by Fitzjohn and Bashir, 2010 [95].

https://doi.org/10.1371/journal.pone.0075112.g004

To investigate a possible mechanism by which a higher level of a specific protein in the insoluble fraction of superior performers could be explained, we tested whether insolubility could be mediated by other proteins, more prone to sedimentation. Inspired by the joint role of the identified proteins ARP3, NEB2, and BRAG2 in synaptic plasticity, where polymerization of actin monomers to F-actin polymers plays a critical role, we performed F-actin precipitation assays and were able to validate a co-precipitation of ARP3 with F-actin (Figure 5). We could demonstrate ARP3 co-precipitation at a centrifugation speed of 100.000 x g with an intact F-actin network (Figure 5, DMSO control) but not after F-actin depolymerization by Mycalolide B (Figure 5, 5 µM Mycalolide B).

thumbnail
Figure 5. F-actin precipitation assay.

Western blot of the supernatant (SN) and insoluble pellet (P) fraction of cell lysates subjected to an F-actin precipitation assay. Visible are immunostained bands of actin and ARP3 as designated. Cells were pre-treated with 1 µM or 5 µM of the F-actin de-polymerizing agent Mycalolide B, or with the respective amount of the solvent DMSO only. Upon treatment with 5 µM Mycalolide B, actin (F-actin) and ARP3 simultaneously disappear from the pellet fraction indicating that ARP3’s insolubility is dependent on F-actin.

https://doi.org/10.1371/journal.pone.0075112.g005

In addition to the detected changes in levels of ARP3, NEB2, and BRAG2, we also found correlations between these 3 proteins and the respective cognitive performances of the individual aged rats, as assessed by total mean swimming distance (A0) to the hidden platform in the MWM (Table 1, Figure 6 B, E, H) corroborating their critical role in maintaining cognitive performance. ARP3 and NEB2 accumulation correlated negatively with A0, contrary to the concentration of insoluble BRAG2. Furthermore, ARP3 and BRAG2 showed correlations to the learning rate a1 of the individual aged animals (Table 1, Figure 6 C, F, I), with ARP3 displaying a positive correlation and BRAG2 being negatively correlated with a1. For NEB2, the calculated (positive) correlation exceeded the significance threshold (p = 0.066).

thumbnail
Figure 6. Relative abundance of insoluble ARP3, NEB2, and BRAG2 comparing aged superior and aged inferior learners.

Displayed are the log2-transformed relative abundances of each of the three proteins as determined for all 16 aged rats by iTRAQ mass spectrometry. Graphs A, D, and G show a group-wise comparison of aged superior and the aged inferior rats. Panels B, E, and H display the log2 protein values plotted against the rats’ performance in the MWM assessed as mean total distance A0 [cm]. Drawn in graphs C, F, and I are the log2 protein values plotted against the rats’ performance in the MWM as the learning pace, assessed as a1-slope over the 6 trial days. Besides the Pearson correlation values (r) and the respective p-value, graphs B, C, E, F, H, and I also display the line of linear regression along with its 95%-confidence interval.

https://doi.org/10.1371/journal.pone.0075112.g006

The amount of ARP3 protein in the detergent-insoluble fractions of aged rats also correlated positively with the amount of NEB2 protein (Figure 7 A) identified in the same fractions. ARP3 showed a negative correlation with BRAG2 in the aged cohort (Figure 7 B), consistent with the inverse accumulation of BRAG2 and its diverse role in the post-synapse. A similar correlation was found for all aged and adult rats combined (r = -0.377, p = 0.034; data not shown).

thumbnail
Figure 7. Protein inter-correlations.

Displayed are the log2-transformed relative abundances of ARP3, NEB2, and BRAG2 plotted against one another to visualize possible correlative changes. Graphs show NEB2-values plotted against ARP3-values (A), BRAG2 against ARP3 (B), and BRAG2 against NEB2 (C). Besides the Pearson correlation values (r) and the respective p-value, graphs also display the line of linear regression along with its 95%-confidence interval.

https://doi.org/10.1371/journal.pone.0075112.g007

Age-specific hippocampal aggregome

Comparing the aged and adult rat cohort, we identified 52 insoluble proteins that differed as a function of age (p < 0.05, FDR < 11.7%; Table 2). Amongst those proteins enriched significantly in the insoluble fraction of the aged (p < 0.014; FDR < 6.7%), the cluster of microtubule-associated proteins (dynactin subunit 1, cytoplasmic dynein 1 heavy chain 1, microtubule-associated protein 2, tubulin alpha-4A chain) was represented most prominently in aged compared to adult rats with a 20-fold enrichment (p = 0.035), as determined by gene ontology (GO) clustering analyses. Decreases in the insoluble fraction (p ≤ 0.014) on the other hand, were found for members of the post-synaptic density, PSD, (disks large homolog 2, glutamate [NMDA] receptor subunit epsilon-1, SH3 and multiple ankyrin repeat domains protein 1-3), that were enriched by 45-fold (p = 9.7 x 10-5). The additional identified GO-cluster of general ‘cytoskeleton’ (especially actin) associated proteins, extended the PSD-cluster by the proteins actin-related protein 2/3 complex subunit 2, protein bassoon and SRC kinase signaling inhibitor 1. These proteins of the ‘cytoskeleton’ GO-cluster were found 8-fold enriched among the set of decreased proteins (p = 2.1 x 10-4).

aged vs. adult
EntryProtein nameChangep-valueFDR
HPLN2_RATHyaluronan and proteoglycan link protein 22.6 x 10-92.3 x 10-7
GFAP_RATGlial fibrillary acidic protein2.0 x 10-60.00012
MOBP_RATMyelin-associated oligodendrocyte basic protein5.5 x 10-60.00023
DCTN1_RATDynactin subunit 12.8 x 10-50.00089
BSN_RATProtein bassoon4.9 x 10-50.00126
SHAN3_RATSH3 and multiple ankyrin repeat domains protein 38.6 x 10-50.00184
HOME1_RATHomer protein homolog 10.000220.00406
SNIP_RATSRC kinase signaling inhibitor 10.000260.00414
AGAP2_RATArf-GAP, GTPase, ANK repeat and PH domain-containing protein 20.000620.00881
gi|293342552PREDICTED: collagen, type IV, alpha 20.001110.01414
SHAN1_RATSH3 and multiple ankyrin repeat domains protein 10.001220.01418
ERC2_RATERC protein 20.001690.01760
CN37_RAT2', 3'-cyclic-nucleotide 3'-phosphodiesterase0.001790.01760
TBA4A_RATTubulin alpha-4A chain0.002470.02255
GLNA_RATGlutamine synthetase0.003260.02778
DLG2_RATDisks large homolog 20.003770.03014
G3P_RATGlyceraldehyde-3-phosphate dehydrogenase0.004550.03426
BAIP2_RATBrain-specific angiogenesis inhibitor 1-associated protein 20.006750.04665
SHAN2_RATSH3 and multiple ankyrin repeat domains protein 20.006930.04665
DYHC1_RATCytoplasmic dynein 1 heavy chain 10.008120.04889
gi|157818467Heat shock 70kDa protein 12A0.008130.04889
gi|293345780PREDICTED: similar to Gene model 9960.008410.04889
SYN1_RATSynapsin-10.011250.06007
MTAP2_RATMicrotubule-associated protein 20.011270.06007
IDH3B_RATIsocitrate dehydrogenase [NAD] subunit beta, mitochondrial0.013960.06635
NMDE1_RATGlutamate [NMDA] receptor subunit epsilon-10.013970.06635
ARPC2_RATActin-related protein 2/3 complex subunit 20.014000.06635
DLGP2_RATDisks large-associated protein 20.016140.07203
DPYL2_RATDihydropyrimidinase-related protein 20.017070.07203
gi|157823479Prickle homolog 20.017310.07203
gi|33563266NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 40.017450.07203
DLGP3_RATDisks large-associated protein 30.018350.07289
NEB2_RATNeurabin-20.019370.07289
NCAN_RATNeurocan core protein0.019920.07289
EF2_RATElongation factor 20.019940.07289
ATPA_RATATP synthase subunit alpha, mitochondrial0.020840.07365
HPLN1_RATHyaluronan and proteoglycan link protein 10.021300.07365
MYO1D_RATMyosin-Id0.023080.07773
GELS_RATGelsolin0.026840.08348
TBB2C_RATTubulin beta-2C chain0.027210.08348
RL15_RAT60S ribosomal protein L150.027400.08348
DYN1_RATDynamin-10.025440.08348
gi|5031595Actin related protein 2/3 complex, subunit 4 [Mus musculus]0.029120.08667
AT1A1_RATSodium/potassium-transporting ATPase subunit alpha-10.030960.09005
RL7A_RAT60S ribosomal protein L7a0.035150.09773
NMDE2_RATGlutamate [NMDA] receptor subunit epsilon-20.035350.09773
BEGIN_RATBrain-enriched guanylate kinase-associated protein0.035890.09773
NSF_RATVesicle-fusing ATPase0.040100.10547
gi|149066065rCG599840.040390.10547
TBB3_RATTubulin beta-3 chain0.043070.11022
DLG4_RATDisks large homolog 40.047260.11634
KCC2A_RATCalcium/calmodulin-dependent protein kinase type II alpha chain0.047280.11634

Table 2. Proteins significantly altered comparing the insoluble proteome of aged and adult rat hippocampi.

Arrow down: Lower abundance in the aged. Arrow up: Higher abundance in the aged. FDR: False discovery rate.
CSV
Download CSV

Discussion and Conclusions

In this study, we demonstrate in a quantitative proteomics (iTRAQ) approach that the analysis of changes in the insoluble proteome of the hippocampus identifies molecular components correlating with functional systems data, here the animal’s age and cognitive performance in a learning and memory task (MWM).

Our study is the first proteome-wide approach describing insoluble protein composition in the aged mammalian brain and relates to previous studies performed in the nematode Caenorhabditis elegans where changes in protein solubility in the course of aging were detected [19,20]. In those studies, detergent-insoluble proteins were purified from young and old, pooled whole C. elegans, and protein components of the aggregomes were quantified using iTRAQ. Compared to the studies of David et al. and Reis-Rodrigues et al. in C. elegans where changes above 10-fold were reported [19,20], the range of the detected differences between groups in our study was much smaller. One explanation for this may be the presence of different protein degradation mechanisms in mammalian cells as compared to nematodes, allowing for a more efficient clearing of protein aggregates [54], or the specific focus on a brain subregion in our study, as opposed to using the whole organism. In addition, there are important technical differences between the C. elegans studies and our investigations. In our study on the mammalian brain, we chose an approach with high statistical power (8 vs. 8 and 16 vs. 16 biological replicates for ‘cognition in aged’ and ‘aging’, respectively) compared to smaller numbers of biological replicates (n = 2-4) used in the investigations on whole C. elegans [19,20]. We noted that, despite normalization to input tissue-weight before purification, similar to previous studies, strong variances in total detected protein levels were detected between single animals after purification, exceeding the differences in protein levels between the analyzed groups. Therefore, for the present investigation, we had to adopt a global normalization method prior to statistical testing. Finally, the aggregome purification protocols applied in the C. elegans studies were distinct from ours and changes in the use of detergents, centrifugation speed, and a final extraction with formic acid, may result in differences in the set of proteins identified in the aggregome. Compared to David et al. [20], we used higher centrifugation speed (100.000 x g instead of 20.000 x g), applied less ionic detergent (0.2% sarkosyl instead of 1% SDS/SDO), and performed two subsequent lipid and low molecular weight protein extractions in high-density buffer (1.1 and 1.6 M sucrose instead of 1.0 M). Furthermore, we used 1.5 M NaCl instead of 0.75 M in our high ionic strength buffer, and did not apply a final extraction in 70% formic acid, that was introduced by David and co-workers to exclude their nematode cuticular debris from the analyses. Reis-Rodriguez and colleagues [19] pre-cleared their samples by centrifugation at 3.000 x g and reduced the purification protocol to three subsequent washes with 1% SDS and centrifugation steps at 16.000 x g, followed by a final extraction with 70% formic acid.

For our work, in order to include more potentially altered proteins in our post-hoc analyses, we decided, contingent upon statistical significance, to be less restrictive with regard to the FDRs (listed in Tables 1 and 2). For the proteins associated with ‘aging’, this resulted in the inclusion of proteins showing an FDR of 0.1163, whilst for the ‘cognition in aged’ analyses all three proteins identified as changed with p ≤ 0.05 were analyzed further.

Proteins identified in the insoluble fraction of superior vs. inferior performing aged rats

ARP3 and NEB2 proteins were identified at higher concentrations in the insoluble fraction of the aged superior performers as compared to their age-matched impaired, inferior performers. In the context of synaptic plasticity, a likely explanation is that with our biochemical fractionation we pull down significant amounts of dendritic F-actin (see also the likely increased pull-down of insoluble actin in superior performers; Table 1) and associated proteins like ARP3 (Figure 5). However, other reasons such as specific post-translational modifications, or secondary effects like a specific local increase in protein density with a change in facultative protein-protein interactions should also be considered. In contrast, BRAG2, being critical for the induction of LTD [53], accumulated in the aggregome of the aged inferior learners. The observed inverse correlation of BRAG2 with ARP3 (Figure 7 B) seems conclusive with its rather inverse role in synaptic plasticity.

The solubility changes of ARP3, NEB2, and BRAG2 detected in our proteomic analysis of aged rats are converging on an inter-related set of proteins involved in synaptic plasticity. This becomes even more intriguing, as these changes not only showed significance in the grouped comparison of aged superior and inferior performers, but, additionally, proved correlative for the cognitive performances (A0) of the individual animals (Figure 6 B, E, H), as well as – in the case of ARP3 and BRAG2 – for the individual learning rate a1 (Figure 6 C, I). Also inter-correlation of the solubility changes between NEB2 and ARP3, as well as between ARP3 and BRAG2 were observed (Figure 7). For ARP3, coprecipitation with immobilized NEB2 has been described [55].

So far, these proteins have not been reported in the context of changes in cognitive performance in analyses of whole synaptic protein fractions, i.e. comprising soluble and insoluble conformers [14,15,17,56,57,58]. Notably, in one study investigating differences in the whole hippocampal proteome of mice trained in different spatial memory tasks, Zheng and co-workers (2009) [16] found that ARP3, along with ARPC5 and F-actin-capping protein subunit beta, was differentially regulated in mice trained in different tasks. However, in their approach, changes were only observed comparing the different means of training, but did not correlate with any learning behavior within the trials.

The term synaptic plasticity describes structural and functional changes of dendritic spines and their postsynaptic densities (PSDs). These changes are observed following learning or the experimental induction of long-term potentiation (LTP) and long-term depression (LTD), respectively, and are believed to contribute to memory formation (reviewed by Lamprecht and LeDoux, 2004 [59]). LTP has been directly linked to performance in learning tasks such as the MWM [30], and thus, can be seen as a cellular manifestation of the observed rat behavior in learning and memory [60]. Basis of LTP and LTD are changes in the molecular composition of key proteins (e.g. AMPA receptors [61,62]) and cellular structures at synapses [52]. Both, LTP and LTD lead to the re-modeling of the actin cytoskeleton in dendritic spines [52,63]. Whereas LTP initializes the introduction of receptors into the postsynaptic membrane [64], its counter-player process LTD, induced by NMDA receptor (NMDAR) and metabotropic glutamate receptor (mGluR) activity [65], has been shown to evoke internalization of AMPA receptors (AMPARs) [64]. Like LTP, LTD is generally believed to play a crucial role in hippocampal memory formation [66].

The modulation of dendritic spine volume to mediate synaptic structural plasticity mainly involves the reorganization of the spine’s actin-cytoskeleton, which is mediated by signaling proteins such as Neurabin-2 (Spinophilin) and the Ca2+/calmodulin-dependent protein kinase, CAMKII [52] (Figure 4). This remodeling and the extension or degradation of the F-actin (filamentous actin) network in dendritic spines is crucial for synaptic plasticity and LTP/LTD maintenance [51,67,68], and there is a direct and crucial link of this to learning and memory formation [51,59,67,68,69,70,71].

ARP3.

Actin-related protein 3 (ARP3) was found elevated in the insoluble fraction of hippocampi from aged rats displaying superior cognitive abilities. ARP3 expression is crucial to embryonic viability past the blastocyst stage [72] and is part of a protein complex, which includes ARP2 and the five subunits ARPC1-5 [73,74,75,76,77]. The complex builds the branching points of F-actin filaments [76,78] and thereby mediates the formation of branched structures within the actin cytoskeleton network [76,79] (Figure 4). Showing high concentrations in dendritic spines [80], the ARP2/3 complex is responsible for the actin network organization in spine heads and disturbances in expression of its subunits results in impaired spine and synapse formation [81,82], and in changes of synapse activity [81].

NEB2.

The second protein identified as accumulating in the ‘aged superior’ aggregome, is spinophilin, also termed neurabin 2 (NEB2). It is a protein phosphatase I (PP-I) interacting and PP-I – activity modulating protein [83,84] that is primarily found in dendritic spines [83,85]. Intraspinal localization of NEB2 and its F-actin binding and bundling capacity were demonstrated to be modulated via its phosphorylation by the Ca2+/calmodulin dependent kinase II (CAMKII) [86] or the protein kinase A (PKA) [87], linking spinophilin action with its responsiveness to NMDA and adrenergic receptor activity [88,89] (Figure 4). The NEB2-mediated effect on F-actin organization within dendritic spines, thus, depends on Ca2+ as well as on cAMP signaling. It is, presumably, via modulation of the spinal F-actin network, that NEB2 modulates dendritic morphology [90] and, hence, has been found to be important for hippocampal integrity [90]. Notably, despite its effect on F-actin organization, mice, deficient in spinophilin, showed no altered LTP, but reduced LTD [90].

Previous experiments investigating a direct quantitative relation between NEB2, aging and cognitive abilities showed no positive results for the total and unfractionated protein levels [91,92]. These results, compared to our findings reported here, highlight the necessity to differentiate the solubility status of synaptic proteins for determining their function.

BRAG2.

The third protein identified in our study comparing aged inferior and superior rats, BRAG2, was found increased in the aggregome of aged inferior rats. First described by Someya et al. in 2001 [93], this protein features an IQ-like motive and a SEC7 domain, and acts as a guanine nucleotide-exchange protein for the ADP-ribosylation factor 6 (ARF6) and, hence, is also termed IQSEC1 or ARF-GEP100. Like the other two proteins found to be changed with cognitive ability in the aggregome of aged rats – ARP3 and NEB2 – BRAG2 has been implicated in the mechanism of actin-remodeling [94] and a direct effect on LTD-maintenance has been observed [53]. NMDAR- as well as mGluR-mediated LTD was found to rely upon BRAG2 expression [53] and a mechanism of BRAG2 binding to the GluA2 C-terminal part of AMPA receptors to induce their ARF6 mediated internalization [53,95] along with changes in the actin cytoskeleton [94], has been described (Figure 4).

In our proteomic results on BRAG2, one value (at 0.325) for the aged inferior cohort appeared to be standing out amongst the other values. Hence, a critical outlier-consideration was performed, where the value passed the ROUT and Grubb’s test for outliers with stringency set to 5%. Furthermore, no other data derived from this particular rat appeared to be irregular. Therefore, the value was included in the post-hoc analyses. Yet, even if excluded from the analysis, the comparison of aged inferior and aged superior rats in their amounts of insoluble BRAG2 still passed the significance threshold with p = 0.0449 (data not shown).

Proteins identified in the insoluble fraction of aged vs. adult rats

In the approach by David et al., an overall number of 461 insoluble proteins was found increased by more than two-fold in the aged samples [20], whereas Reis-Rodrigues et al. reported the finding of 203 insoluble proteins [19]. Elongation factor 2, heat shock protein 70, and glyceraldehyde-3-phosphate dehydrogenase showed high abundance for the aged samples both in the insoluble fractions for the nematode approaches [19,20] and in our rat hippocampal fraction. In addition, the work of David and co-workers shared the detection of up-regulated myosins with the present study [20]. The study conducted by Reis-Rodrigues et al. reported findings – similar to our results for rat hippocampi – of elevated levels of aggregated tubulin and mitochondrial ATP synthase subunit alpha in aged worms [19], with the latter protein component having been recognized as a common denominator in the aging-pigment lipofuscin [96,97]. The increase of 60S ribosomal protein in the insoluble fraction of the aged animals, observed in the nematode studies [19,20], however, opposes our findings of a decrease in aged rat brain aggregome.

Reis-Rodrigues and colleagues further reported the finding of significant extensions of lifespan of C. elegans upon RNAi-based knockdown for almost half of the genes tested and whose products were found to be elevated in the aged aggregome [19]. Amongst those were the elongation factor 2, implicated in the ribosomal translation elongation, and tubulin, a component of the microtubular cytoskeleton [19].

Notably, there is no large overlap in the proteome from the insoluble fraction used in this study, compared to the aged rat hippocampal synaptoproteome described by Van Guilder and colleagues [12]. This corroborates the expected specificity of proteins prone to aggregate with age and accounts for the solubility changes to not be mere mirroring effects of altered expression levels, but rather to reflect changes in the proteostasis maintaining cellular machinery. However, some commonalities could be described: Whereas dynamin-1, which appears to be downregulated in the aged rat synaptic proteome [12] and was also less abundant in the aged aggregome, other proteins rather showed an opposite trend. Synapsin-1, a phosphoprotein associated with synaptic vesicles [98], was reported to show elevated expression comparing aged and adult synaptoproteome [12], but was reduced in the aged aggregome. Heat shock protein 70, isocitrate dehydrogenase [NAD], NADH dehydrogenase (ubiquinone), and tubulins have been demonstrated to be less abundant in the soluble aged rat hippocampus synaptoproteome [12], whereas they showed elevated quantities in the detergent-insoluble fraction analyzed in this present study. This latter, inverse correlation may point towards proteins, that specifically become less soluble with age, e.g. by oxidative stress [99], and are therefore depleted of the soluble fraction analyzed by Van Guilder and colleagues [12].

In this study, we demonstrated age-associated changes of protein solubility in the mammalian hippocampus of the rat. We used quantitative proteomics to also differentiate inferior and superior performing aged rats in a functional assay of memory and identified three proteins ARP3, NEB2, and BRAG2 involved in synaptic plasticity and LTD as potential molecular correlates of the age-associated memory decline. We thereby demonstrated that quantitative proteomics of the aggregome is an appropriate method for identifying molecular components of behavior associated with memory/learning processes in a systems biology approach to studying the aged brain.

Author Contributions

Conceived and designed the experiments: CK PO JPH ABS BT ML AdS DS KWL. Performed the experiments: PO BT DS KWL AdS. Analyzed the data: CK PO JPH ABS BT ML AdS DS KWL. Contributed reagents/materials/analysis tools: CK PO JPH ABS BT ML AdS DS KWL. Wrote the manuscript: CK PO JPH ML ABS BT.

References

  1. 1. Blasco MA (2007) Telomere length, stem cells and aging. Nat Chem Biol 3: 640-649. doi:https://doi.org/10.1038/nchembio.2007.38. PubMed: 17876321.
  2. 2. Terman A, Kurz T, Navratil M, Arriaga EA, Brunk UT (2010) Mitochondrial turnover and aging of long-lived postmitotic cells: the mitochondrial-lysosomal axis theory of aging. Antioxid Redox Signal 12: 503-535. doi:https://doi.org/10.1089/ars.2009.2598. PubMed: 19650712.
  3. 3. O’Neill C, Kiely AP, Coakley MF, Manning S, Long-Smith CM (2010) Insulin and IGF-1 signalling: longevity, protein homoeostasis and Alzheimer’s disease. Biochem Soc Trans 40: 721-727.
  4. 4. Lu T, Pan Y, Kao SY, Li C, Kohane I et al. (2004) Gene regulation and DNA damage in the ageing human brain. Nature 429: 883-891. doi:https://doi.org/10.1038/nature02661. PubMed: 15190254.
  5. 5. Kosik KS, Rapp PR, Raz N, Small SA, Sweatt JD et al. (2012) Mechanisms of age-related cognitive change and targets for intervention: epigenetics. J Gerontol A Biol Sci Med Sci 67: 741-746. doi:https://doi.org/10.1093/gerona/gls110. PubMed: 22522509.
  6. 6. VanGuilder HD, Freeman WM (2011) The hippocampal neuroproteome with aging and cognitive decline: past progress and future directions. Front Aging Neurosci 3: 8. PubMed: 21647399.
  7. 7. Carrette O, Burkhard PR, Hochstrasser DF, Sanchez JC (2006) Age-related proteome analysis of the mouse brain: a 2-DE study. Proteomics 6: 4940-4949. doi:https://doi.org/10.1002/pmic.200600084. PubMed: 16912971.
  8. 8. Föcking M, Boersema PJ, O’Donoghue N, Lubec G, Pennington SR et al. (2006) 2-D DIGE as a quantitative tool for investigating the HUPO Brain Proteome Project mouse series. Proteomics 6: 4914-4931. doi:https://doi.org/10.1002/pmic.200600269. PubMed: 16927420.
  9. 9. Yang S, Liu T, Li S, Zhang X, Ding Q et al. (2008) Comparative proteomic analysis of brains of naturally aging mice. Neuroscience 154: 1107-1120. doi:https://doi.org/10.1016/j.neuroscience.2008.04.012. PubMed: 18495355.
  10. 10. Poon HF, Shepherd HM, Reed TT, Calabrese V, Stella AM et al. (2006) Proteomics analysis provides insight into caloric restriction mediated oxidation and expression of brain proteins associated with age-related impaired cellular processes: Mitochondrial dysfunction, glutamate dysregulation and impaired protein synthesis. Neurobiol Aging 27: 1020-1034. doi:https://doi.org/10.1016/j.neurobiolaging.2005.05.014. PubMed: 15996793.
  11. 11. Sato Y, Yamanaka H, Toda T, Shinohara Y, Endo T (2005) Comparison of hippocampal synaptosome proteins in young-adult and aged rats. Neurosci Lett 382: 22-26. doi:https://doi.org/10.1016/j.neulet.2005.02.053. PubMed: 15911115.
  12. 12. VanGuilder HD, Yan H, Farley JA, Sonntag WE, Freeman WM (2010) Aging alters the expression of neurotransmission-regulating proteins in the hippocampal synaptoproteome. J Neurochem 113: 1577-1588. PubMed: 20374424.
  13. 13. Weinreb O, Drigues N, Sagi Y, Reznick AZ, Amit T et al. (2007) The application of proteomics and genomics to the study of age-related neurodegeneration and neuroprotection. Antioxid Redox Signal 9: 169-179. doi:https://doi.org/10.1089/ars.2007.9.169. PubMed: 17115941.
  14. 14. McNair K, Davies CH, Cobb SR (2006) Plasticity-related regulation of the hippocampal proteome. Eur J Neurosci 23: 575-580. doi:https://doi.org/10.1111/j.1460-9568.2005.04542.x. PubMed: 16420465.
  15. 15. McNair K, Broad J, Riedel G, Davies CH, Cobb SR (2007) Global changes in the hippocampal proteome following exposure to an enriched environment. Neuroscience 145: 413-422. doi:https://doi.org/10.1016/j.neuroscience.2006.12.033. PubMed: 17261355.
  16. 16. Zheng JF, Patil SS, Chen WQ, An W, He JQ et al. (2009) Hippocampal protein levels related to spatial memory are different in the Barnes maze and in the multiple T-maze. J Proteome Res 8: 4479-4486. doi:https://doi.org/10.1021/pr9002596. PubMed: 19650667.
  17. 17. Chen WQ, Viidik A, Skalicky M, Höger H, Lubec G (2007) Hippocampal signaling cascades are modulated in voluntary and treadmill exercise rats. Electrophoresis 28: 4392-4400. doi:https://doi.org/10.1002/elps.200700336. PubMed: 17963288.
  18. 18. Douglas PM, Dillin A (2010) Protein homeostasis and aging in neurodegeneration. J Cell Biol 190: 719-729. doi:https://doi.org/10.1083/jcb.201005144. PubMed: 20819932.
  19. 19. Reis-Rodrigues P, Czerwieniec G, Peters TW, Evani US, Alavez S et al. (2012) Proteomic analysis of age-dependent changes in protein solubility identifies genes that modulate lifespan. Aging Cell 11: 120-127. doi:https://doi.org/10.1111/j.1474-9726.2011.00765.x. PubMed: 22103665.
  20. 20. David DC, Ollikainen N, Trinidad JC, Cary MP, Burlingame AL et al. (2010) Widespread protein aggregation as an inherent part of aging in C. elegans. PLOS Biol 8: e1000450.
  21. 21. Reisberg B, Ferris SH, de Leon MJ, Franssen ESE, Kluger A et al. (1988) Stage-specific behavioral, cognitive, and in vivo changes in community residing subjects with age-associated memory impairment and primary degenerative dementia of the Alzheimer type. Drug Dev Res 15: 101-114. doi:https://doi.org/10.1002/ddr.430150203.
  22. 22. Gauthier S, Reisberg B, Zaudig M, Petersen RC, Ritchie K et al. (2006) Mild cognitive impairment. Lancet 367: 1262-1270. doi:https://doi.org/10.1016/S0140-6736(06)68542-5. PubMed: 16631882.
  23. 23. Petersen RC (2004) Mild cognitive impairment as a diagnostic entity. J Intern Med 256: 183-194. doi:https://doi.org/10.1111/j.1365-2796.2004.01388.x. PubMed: 15324362.
  24. 24. Geslani DM, Tierney MC, Herrmann N, Szalai JP (2005) Mild cognitive impairment: an operational definition and its conversion rate to Alzheimer’s disease. Dement Geriatr Cogn Disord 19: 383-389. doi:https://doi.org/10.1159/000084709. PubMed: 15802914.
  25. 25. Gallagher M, Burwell RD (1989) Relationship of age-related decline across several behavioral domains. Neurobiol Aging 10: 691-708. doi:https://doi.org/10.1016/0197-4580(89)90006-7. PubMed: 2628781.
  26. 26. Rapp PR, Amaral DG (1992) Individual differences in the cognitive and neurobiological consequences of normal aging. Trends Neurosci 15: 340-345. doi:https://doi.org/10.1016/0166-2236(92)90051-9. PubMed: 1382333.
  27. 27. Lupien SJ, Wan N (2004) Successful ageing: from cell to self. Philos Trans R Soc Lond B Biol Sci 359: 1413-1426. doi:https://doi.org/10.1098/rstb.2004.1516. PubMed: 15347532.
  28. 28. Rowe JW, Kahn RL (1987) Human aging: usual and successful. Science 237: 143-149. doi:https://doi.org/10.1126/science.3299702. PubMed: 3299702.
  29. 29. Meijer OC, Topic B, Steenbergen PJ, Jocham G, Huston JP et al. (2005) Correlations between hypothalamus-pituitary-adrenal axis parameters depend on age and learning capacity. Endocrinology 146: 1372-1381. PubMed: 15564338.
  30. 30. Schulz D, Huston JP, Jezek K, Haas HL, Roth-Härer A et al. (2002) Water maze performance, exploratory activity, inhibitory avoidance and hippocampal plasticity in aged superior and inferior learners. Eur J Neurosci 16: 2175-2185. doi:https://doi.org/10.1046/j.1460-9568.2002.02282.x. PubMed: 12473085.
  31. 31. Schulz D, Sergeeva OA, Ianovskii E, Luhmann HJ, Haas HL et al. (2004) Behavioural parameters in aged rats are related to LTP and gene expression of ChAT and NMDA-NR2 subunits in the striatum. Eur J Neurosci 19: 1373-1383. doi:https://doi.org/10.1111/j.1460-9568.2004.03234.x. PubMed: 15016095.
  32. 32. Topic B, Willuhn I, Palomero-Gallagher N, Zilles K, Huston JP et al. (2007) Impaired maze performance in aged rats is accompanied by increased density of NMDA, 5-HT1A, and alpha-adrenoceptor binding in hippocampus. Hippocampus 17: 68-77. doi:https://doi.org/10.1002/hipo.20246. PubMed: 17111411.
  33. 33. Gallagher M, Nicolle MM (1993) Animal models of normal aging: relationship between cognitive decline and markers in hippocampal circuitry. Behav Brain Res 57: 155-162. doi:https://doi.org/10.1016/0166-4328(93)90131-9. PubMed: 7906946.
  34. 34. Schulz D, Huston JP, Buddenberg T, Topic B (2007) "Despair" induced by extinction trials in the water maze: relationship with measures of anxiety in aged and adult rats. Neurobiol Learn Mem 87: 309-323. doi:https://doi.org/10.1016/j.nlm.2006.09.006. PubMed: 17079170.
  35. 35. Topic B, Dere E, Schulz D, de Souza Silva MA, Jocham G et al. (2005) Aged and adult rats compared in acquisition and extinction of escape from the water maze: focus on individual differences. Behav Neurosci 119: 127-144. doi:https://doi.org/10.1037/0735-7044.119.1.127. PubMed: 15727519.
  36. 36. Hasenöhrl RU, Söderstróm S, Mohammed AH, Ebendal T, Huston JP (1997) Reciprocal changes in expression of mRNA for nerve growth factor and its receptors TrkA and LNGFR in brain of aged rats in relation to maze learning deficits. Exp Brain Res 114: 205-213. doi:https://doi.org/10.1007/PL00005629. PubMed: 9166910.
  37. 37. Krauth J (1980) Nonparametric analysis of response curves. J Neurosci Methods 2: 239-252. doi:https://doi.org/10.1016/0165-0270(80)90013-8. PubMed: 7329095.
  38. 38. Leliveld SR, Bader V, Hendriks P, Prikulis I, Sajnani G et al. (2008) Insolubility of disrupted-in-schizophrenia 1 disrupts oligomer-dependent interactions with nuclear distribution element 1 and is associated with sporadic mental disease. J Neurosci 28: 3839-3845. doi:https://doi.org/10.1523/JNEUROSCI.5389-07.2008. PubMed: 18400883.
  39. 39. Ottis P, Bader V, Trossbach SV, Kretzschmar H, Michel M et al. (2011) Convergence of two independent mental disease genes on the protein level: recruitment of dysbindin to cell-invasive disrupted-in-schizophrenia 1 aggresomes. Biol Psychiatry 70: 604-610. doi:https://doi.org/10.1016/j.biopsych.2011.03.027. PubMed: 21531389.
  40. 40. Ross PL, Huang YN, Marchese JN, Williamson B, Parker K et al. (2004) Multiplexed protein quantitation in Saccharomyces cerevisiae using amine-reactive isobaric tagging reagents. Mol Cell Proteomics 3: 1154-1169. doi:https://doi.org/10.1074/mcp.M400129-MCP200. PubMed: 15385600.
  41. 41. Li KW, Miller S, Klychnikov O, Loos M, Stahl-Zeng J et al. (2007) Quantitative proteomics and protein network analysis of hippocampal synapses of CaMKIIalpha mutant mice. J Proteome Res 6: 3127-3133. doi:https://doi.org/10.1021/pr070086w. PubMed: 17625814.
  42. 42. Storey JD, Tibshirani R (2003) Statistical significance for genomewide studies. Proc Natl Acad Sci U S A 100: 9440-9445. doi:https://doi.org/10.1073/pnas.1530509100. PubMed: 12883005.
  43. 43. Huang da W, Sherman BT, Lempicki RA (2009) Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37: 1-13. doi:https://doi.org/10.1093/nar/gkp505. PubMed: 19033363.
  44. 44. Huang da W, Sherman BT, Lempicki RA (2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4: 44-57. PubMed: 19131956.
  45. 45. Kapushesky M, Emam I, Holloway E, Kurnosov P, Zorin A et al. (2010) Gene expression atlas at the European bioinformatics institute. Nucleic Acids Res 38: D690-D698. doi:https://doi.org/10.1093/nar/gkp936. PubMed: 19906730.
  46. 46. Kapushesky M, Adamusiak T, Burdett T, Culhane A, Farne A et al. (2012) Gene Expression Atlas update--a value-added database of microarray and sequencing-based functional genomics experiments. Nucleic Acids Res 40: D1077-D1081. doi:https://doi.org/10.1093/nar/gkr913. PubMed: 22064864.
  47. 47. Cenni V, Sirri A, Riccio M, Lattanzi G, Santi S et al. (2003) Targeting of the Akt/PKB kinase to the actin skeleton. Cell Mol Life Sci 60: 2710-2720. doi:https://doi.org/10.1007/s00018-003-3349-4. PubMed: 14685694.
  48. 48. Saito S, Watabe S, Ozaki H, Fusetani N, Karaki H (1994) Mycalolide B, a novel actin depolymerizing agent. J Biol Chem 269: 29710-29714. PubMed: 7961961.
  49. 49. Rapp PR, Rosenberg RA, Gallagher M (1987) An evaluation of spatial information processing in aged rats. Behav Neurosci 101: 3-12. doi:https://doi.org/10.1037/0735-7044.101.1.3. PubMed: 3828055.
  50. 50. O’Keefe J, Dostrovsky J (1971) The hippocampus as a spatial map. Preliminary evidence from unit activity in the freely-moving rat. Brain Res 34: 171-175. doi:https://doi.org/10.1016/0006-8993(71)90358-1. PubMed: 5124915.
  51. 51. Hotulainen P, Hoogenraad CC (2010) Actin in dendritic spines: connecting dynamics to function. J Cell Biol 189: 619-629. doi:https://doi.org/10.1083/jcb.201003008. PubMed: 20457765.
  52. 52. Okamoto K, Bosch M, Hayashi Y (2009) The roles of CaMKII and F-actin in the structural plasticity of dendritic spines: a potential molecular identity of a synaptic tag? Physiol (Bethesda) 24: 357-366. doi:https://doi.org/10.1152/physiol.00029.2009.
  53. 53. Scholz R, Berberich S, Rathgeber L, Kolleker A, Köhr G et al. (2010) AMPA receptor signaling through BRAG2 and Arf6 critical for long-term synaptic depression. Neuron 66: 768-780. doi:https://doi.org/10.1016/j.neuron.2010.05.003. PubMed: 20547133.
  54. 54. Mori K (2009) Signalling pathways in the unfolded protein response: development from yeast to mammals. J Biochem 146: 743-750. doi:https://doi.org/10.1093/jb/mvp166. PubMed: 19861400.
  55. 55. Barnes AP, Smith FD 3rd, VanDongen HM, VanDongen AM, Milgram SL (2004) The identification of a second actin-binding region in spinophilin/neurabin II. Brain Res. Mol Brain Res 124: 105-113. doi:https://doi.org/10.1016/j.molbrainres.2003.12.020. PubMed: 15135218.
  56. 56. Nelson TJ, Backlund PS Jr., Alkon DL (2004) Hippocampal protein-protein interactions in spatial memory. Hippocampus 14: 46-57. doi:https://doi.org/10.1002/hipo.10152. PubMed: 15058482.
  57. 57. Henninger N, Feldmann RE Jr., Fütterer CD, Schrempp C, Maurer MH et al. (2007) Spatial learning induces predominant downregulation of cytosolic proteins in the rat hippocampus. Genes Brain Behav 6: 128-140. doi:https://doi.org/10.1111/j.1601-183X.2006.00239.x. PubMed: 16643511.
  58. 58. Freeman WM, VanGuilder HD, Bennett C, Sonntag WE (2009) Cognitive performance and age-related changes in the hippocampal proteome. Neuroscience 159: 183-195. doi:https://doi.org/10.1016/j.neuroscience.2008.12.004. PubMed: 19135133.
  59. 59. Lamprecht R, LeDoux J (2004) Structural plasticity and memory. Nat Rev Neurosci 5: 45-54. doi:https://doi.org/10.1038/nrn1301. PubMed: 14708003.
  60. 60. Pastalkova E, Serrano P, Pinkhasova D, Wallace E, Fenton AA et al. (2006) Storage of spatial information by the maintenance mechanism of LTP. Science 313: 1141-1144. doi:https://doi.org/10.1126/science.1128657. PubMed: 16931766.
  61. 61. Hayashi Y, Shi S-H, Esteban JA, Piccini A, Poncer J-C et al. (2000) Driving AMPA receptors into synapses by LTP and CaMKII: requirement for GluR1 and PDZ domain interaction. Sci Signal 287: 2262–2267. PubMed: 10731148.
  62. 62. Shi S-H, Hayashi Y, Petralia RS, Zaman SH, Wenthold RJ et al. (1999) Rapid spine delivery and redistribution of AMPA receptors after synaptic NMDA receptor activation. Science 284: 1811-1816. doi:https://doi.org/10.1126/science.284.5421.1811. PubMed: 10364548.
  63. 63. Okamoto K, Nagai T, Miyawaki A, Hayashi Y (2004) Rapid and persistent modulation of actin dynamics regulates postsynaptic reorganization underlying bidirectional plasticity. Nat Neurosci 7: 1104-1112. doi:https://doi.org/10.1038/nn1311. PubMed: 15361876.
  64. 64. Kauer JA, Malenka RC (2007) Synaptic plasticity and addiction. Nat Rev Neurosci 8: 844-858. doi:https://doi.org/10.1038/nrn2234. PubMed: 17948030.
  65. 65. Malenka RC, Bear MF (2004) LTP and LTD: an embarrassment of riches. Neuron 44: 5-21. doi:https://doi.org/10.1016/j.neuron.2004.09.012. PubMed: 15450156.
  66. 66. Massey PV, Bashir ZI (2007) Long-term depression: multiple forms and implications for brain function. Trends Neurosci 30: 176-184. doi:https://doi.org/10.1016/j.tins.2007.02.005. PubMed: 17335914.
  67. 67. Krucker T, Siggins GR, Halpain S (2000) Dynamic actin filaments are required for stable long-term potentiation (LTP) in area CA1 of the hippocampus. Proc Natl Acad Sci USA 97: 6856-6861. doi:https://doi.org/10.1073/pnas.100139797. PubMed: 10823894.
  68. 68. Fukazawa Y, Saitoh Y, Ozawa F, Ohta Y, Mizuno K et al. (2003) Hippocampal LTP is accompanied by enhanced F-actin content within the dendritic spine that is essential for late LTP maintenance in vivo. Neuron 38: 447-460. doi:https://doi.org/10.1016/S0896-6273(03)00206-X. PubMed: 12741991.
  69. 69. Ramachandran B, Frey JU (2009) Interfering with the actin network and its effect on long-term potentiation and synaptic tagging in hippocampal CA1 neurons in slices in vitro. J Neurosci 29: 12167-12173. doi:https://doi.org/10.1523/JNEUROSCI.2045-09.2009. PubMed: 19793974.
  70. 70. Soderling SH, Guire ES, Kaech S, White J, Zhang F et al. (2007) A WAVE-1 and WRP signaling complex regulates spine density, synaptic plasticity, and memory. J Neurosci 27: 355-365. doi:https://doi.org/10.1523/JNEUROSCI.3209-06.2006. PubMed: 17215396.
  71. 71. Grove M, Demyanenko G, Echarri A, Zipfel PA, Quiroz ME et al. (2004) ABI2-deficient mice exhibit defective cell migration, aberrant dendritic spine morphogenesis, and deficits in learning and memory. Mol Cell Biol 24: 10905-10922. doi:https://doi.org/10.1128/MCB.24.24.10905-10922.2004. PubMed: 15572692.
  72. 72. Vauti F, Prochnow BR, Freese E, Ramasamy SK, Ruiz P et al. (2007) Arp3 is required during preimplantation development of the mouse embryo. FEBS Lett 581: 5691-5697. doi:https://doi.org/10.1016/j.febslet.2007.11.031. PubMed: 18035060.
  73. 73. Machesky LM, Atkinson SJ, Ampe C, Vandekerckhove J, Pollard TD (1994) Purification of a cortical complex containing two unconventional actins from Acanthamoeba by affinity chromatography on profilin-agarose. J Cell Biol 127: 107-115. doi:https://doi.org/10.1083/jcb.127.1.107. PubMed: 7929556.
  74. 74. Machesky LM, Reeves E, Wientjes F, Mattheyse FJ, Grogan A et al. (1997) Mammalian actin-related protein 2/3 complex localizes to regions of lamellipodial protrusion and is composed of evolutionarily conserved proteins. Biochem J 328(1): 105-112. PubMed: 9359840.
  75. 75. Mullins RD, Stafford WF, Pollard TD (1997) Structure, subunit topology, and actin-binding activity of the Arp2/3 complex from Acanthamoeba. J Cell Biol 136: 331-343. doi:https://doi.org/10.1083/jcb.136.2.331. PubMed: 9015304.
  76. 76. Rouiller I, Xu XP, Amann KJ, Egile C, Nickell S et al. (2008) The structural basis of actin filament branching by the Arp2/3 complex. J Cell Biol 180: 887-895. doi:https://doi.org/10.1083/jcb.200709092. PubMed: 18316411.
  77. 77. Welch MD, DePace AH, Verma S, Iwamatsu A, Mitchison TJ (1997) The human Arp2/3 complex is composed of evolutionarily conserved subunits and is localized to cellular regions of dynamic actin filament assembly. J Cell Biol 138: 375-384. doi:https://doi.org/10.1083/jcb.138.2.375. PubMed: 9230079.
  78. 78. Svitkina TM, Borisy GG (1999) Arp2/3 complex and actin depolymerizing factor/cofilin in dendritic organization and treadmilling of actin filament array in lamellipodia. J Cell Biol 145: 1009-1026. doi:https://doi.org/10.1083/jcb.145.5.1009. PubMed: 10352018.
  79. 79. Mullins RD, Heuser JA, Pollard TD (1998) The interaction of Arp2/3 complex with actin: nucleation, high affinity pointed end capping, and formation of branching networks of filaments. Proc Natl Acad Sci U S A 95: 6181-6186. doi:https://doi.org/10.1073/pnas.95.11.6181. PubMed: 9600938.
  80. 80. Rácz B, Weinberg RJ (2008) Organization of the Arp2/3 complex in hippocampal spines. J Neurosci 28: 5654-5659. doi:https://doi.org/10.1523/JNEUROSCI.0756-08.2008. PubMed: 18509026.
  81. 81. Hotulainen P, Llano O, Smirnov S, Tanhuanpää K, Faix J et al. (2009) Defining mechanisms of actin polymerization and depolymerization during dendritic spine morphogenesis. J Cell Biol 185: 323-339. doi:https://doi.org/10.1083/jcb.200809046. PubMed: 19380880.
  82. 82. Wegner AM, Nebhan CA, Hu L, Majumdar D, Meier KM et al. (2008) N-wasp and the arp2/3 complex are critical regulators of actin in the development of dendritic spines and synapses. J Biol Chem 283: 15912-15920. doi:https://doi.org/10.1074/jbc.M801555200. PubMed: 18430734.
  83. 83. Allen PB, Ouimet CC, Greengard P (1997) Spinophilin, a novel protein phosphatase 1 binding protein localized to dendritic spines. Proc Natl Acad Sci U S A 94: 9956-9961. doi:https://doi.org/10.1073/pnas.94.18.9956. PubMed: 9275233.
  84. 84. Hsieh-Wilson LC, Allen PB, Watanabe T, Nairn AC, Greengard P (1999) Characterization of the neuronal targeting protein spinophilin and its interactions with protein phosphatase-1. Biochemistry 38: 4365-4373. doi:https://doi.org/10.1021/bi982900m. PubMed: 10194355.
  85. 85. Muly EC, Smith Y, Allen P, Greengard P (2004) Subcellular distribution of spinophilin immunolabeling in primate prefrontal cortex: localization to and within dendritic spines. J Comp Neurol 469: 185-197. doi:https://doi.org/10.1002/cne.11001. PubMed: 14694533.
  86. 86. Grossman SD, Futter M, Snyder GL, Allen PB, Nairn AC et al. (2004) Spinophilin is phosphorylated by Ca2+/calmodulin-dependent protein kinase II resulting in regulation of its binding to F-actin. J Neurochem 90: 317-324. doi:https://doi.org/10.1111/j.1471-4159.2004.02491.x. PubMed: 15228588.
  87. 87. Hsieh-Wilson LC, Benfenati F, Snyder GL, Allen PB, Nairn AC et al. (2003) Phosphorylation of spinophilin modulates its interaction with actin filaments. J Biol Chem 278: 1186-1194. doi:https://doi.org/10.1074/jbc.M205754200. PubMed: 12417592.
  88. 88. Uematsu K, Futter M, Hsieh-Wilson LC, Higashi H, Maeda H et al. (2005) Regulation of spinophilin Ser94 phosphorylation in neostriatal neurons involves both DARPP-32-dependent and independent pathways. J Neurochem 95: 1642-1652. doi:https://doi.org/10.1111/j.1471-4159.2005.03491.x. PubMed: 16300646.
  89. 89. Brady AE, Wang Q, Allen PB, Rizzo M, Greengard P et al. (2005) Alpha 2-adrenergic agonist enrichment of spinophilin at the cell surface involves beta gamma subunits of Gi proteins and is preferentially induced by the alpha 2A-subtype. Mol Pharmacol 67: 1690-1696. doi:https://doi.org/10.1124/mol.104.005215. PubMed: 15705742.
  90. 90. Feng J, Yan Z, Ferreira A, Tomizawa K, Liauw JA et al. (2000) Spinophilin regulates the formation and function of dendritic spines. Proc Natl Acad Sci U S A 97: 9287-9292. doi:https://doi.org/10.1073/pnas.97.16.9287. PubMed: 10922077.
  91. 91. Leuba G, Savioz A, Vernay A, Carnal B, Kraftsik R et al. (2008) Differential changes in synaptic proteins in the Alzheimer frontal cortex with marked increase in PSD-95 postsynaptic protein. J Alzheimers Dis 15: 139-151. PubMed: 18780974.
  92. 92. Calhoun ME, Fletcher BR, Yi S, Zentko DC, Gallagher M et al. (2008) Age-related spatial learning impairment is unrelated to spinophilin immunoreactive spine number and protein levels in rat hippocampus. Neurobiol Aging 29: 1256-1264. doi:https://doi.org/10.1016/j.neurobiolaging.2007.02.013. PubMed: 17353069.
  93. 93. Someya A, Sata M, Takeda K, Pacheco-Rodriguez G, Ferrans VJ et al. (2001) ARF-GEP(100), a guanine nucleotide-exchange protein for ADP-ribosylation factor 6. Proc Natl Acad Sci U S A 98: 2413-2418. doi:https://doi.org/10.1073/pnas.051634798. PubMed: 11226253.
  94. 94. Hiroi T, Someya A, Thompson W, Moss J, Vaughan M (2006) GEP100/BRAG2: activator of ADP-ribosylation factor 6 for regulation of cell adhesion and actin cytoskeleton via E-cadherin and alpha-catenin. Proc Natl Acad Sci U S A 103: 10672-10677. doi:https://doi.org/10.1073/pnas.0604091103. PubMed: 16807291.
  95. 95. Fitzjohn SM, Bashir ZI (2010) BRAGging about mechanisms of long-term depression. Neuron 66: 627-630. doi:https://doi.org/10.1016/j.neuron.2010.05.027. PubMed: 20547122.
  96. 96. Schutt F, Ueberle B, Schnölzer M, Holz FG, Kopitz J (2002) Proteome analysis of lipofuscin in human retinal pigment epithelial cells. FEBS Lett 528: 217-221. doi:https://doi.org/10.1016/S0014-5793(02)03312-4. PubMed: 12297308.
  97. 97. Ottis P, Koppe K, Onisko B, Dynin I, Arzberger T et al. (2012) Human and rat brain lipofuscin proteome. Proteomics 12: 2445-2454. doi:https://doi.org/10.1002/pmic.201100668. PubMed: 22707227.
  98. 98. Huttner WB, Schiebler W, Greengard P, De Camilli P (1983) Synapsin I (protein I), a nerve terminal-specific phosphoprotein. III. Its association with synaptic vesicles studied in a highly purified synaptic vesicle preparation. J Cell Biol 96: 1374-1388.
  99. 99. Guo J, Prokai-Tatrai K, Nguyen V, Rauniyar N, Ughy B et al. (2011) Protein targets for carbonylation by 4-hydroxy-2-nonenal in rat liver mitochondria. J Proteomics 74: 2370-2379. doi:https://doi.org/10.1016/j.jprot.2011.07.009. PubMed: 21801862.
  100. 100. Collingridge GL, Peineau S, Howland JG, Wang YT (2010) Long-term depression in the CNS. Nat Rev Neurosci 11: 459-473. doi:https://doi.org/10.1038/nrn2867. PubMed: 20559335.