The immediate-early gene Arc is rapidly expressed in response to neuronal activity and is thought to be involved in mechanisms of synaptic plasticity. The function of Arc in these processes remains unknown. The present study demonstrates that during an in vitro neural correlate of eyeblink classical conditioning, there is a rapid and transient increase in levels of Arc protein that require activation of N-methyl-d-aspartate receptors. In the early phase of conditioning during conditioned response (CR) acquisition, there is significantly greater colocalization of Arc protein and GluR4-containing AMPA receptors at synaptic sites, however, colocalization of Arc and GluR4 was not observed after later stages of conditioning during CR expression. There was also significantly enhanced coimmunoprecipitation of Arc with GluR4 subunits and actin early in conditioning but not of Arc with NR1 subunits, and these associations declined to control levels in later stages of conditioning. These data suggest a role for Arc protein in the synaptic delivery of GluR4-containing AMPA receptors by interactions with cytoskeletal protein complexes during the acquisition phase of in vitro classical conditioning.
Immediate-early genes (IEGs) are rapidly expressed in response to neuronal activity (Morgan et al. 1987; Saffen et al. 1988). Arc, activity-regulated cytoskeleton-associated protein also known as Arg 3.1, belongs to the group of effector IEGs that directly effect neuronal function, whereas other IEGs encode various transcription factors (c-fos, Egr-1 also known as zif268, c-jun) and indirectly effect neuronal activity by altering gene expression (Guzowski 2002; Lanahan and Worley 1998). It is likely that locally translated Arc protein is involved in neuronal plasticity processes at activated synapses. Electrical stimulation of the perforant-path projections to the dentate gyrus of the hippocampus causes newly synthesized Arc mRNA to transport and accumulate selectively in activated dendritic segments (Steward and Worley 2001; Steward et al. 1998). Arc mRNA is induced by calcium and activity of protein kinase A (PKA) or mitogen-activated protein kinase (MAPK) (Waltereit et al. 2001), and dendritic targeting requires activation of N-methyl-d-aspartate (NMDA) receptors (Steward and Worley 2001). Also in the hippocampus, Arc expression is induced after spatial learning tasks with a maximal increase in mRNA levels after 30 min of training, suggesting that Arc expression is related to task-relevant encoding processes (Guzowski et al. 1999, 2001). Intrahippocampally infused antisense oligodeoxynucleotides to inhibit Arc protein expression has been found to impair consolidation of long-term memory, supporting the hypothesis that Arc expression is induced by learning and is important for memory consolidation (Guzowski et al. 2000).
Biochemical studies have shown that Arc protein has regions of homology with the cytoskeletal protein α-spectrin and coprecipitates with F-actin (Lyford et al. 1995). Arc interacts with newly polymerized microtubules which implicates it in dendritic remodeling by interacting with cytoskeletal proteins (Fujimoto et al. 2004). Proteomic characterization with mass spectrometry and immunoblotting of NMDA receptor multiprotein complexes indicate strong interactions with Arc protein in mouse brain (Husi et al. 2000). While the behavioral data on learning and these protein-protein interactions are suggestive, the role of Arc in synaptic plasticity is unknown. The question as to whether it is involved directly in interactions with other proteins enriched in the postsynaptic density (PSD) such as neurotransmitter receptors and signal transduction molecules remains unanswered. By comparison, other IEG protein products appear to directly contribute to synaptic plasticity by regulating the structure or localization of neurotransmitter receptors. For instance, Homer protein binds to group 1 metabotropic glutamate receptors (mGluRs) and modulates mGluR-induced intracellular calcium release by linking mGluRs with inositol trisphosphate receptors (Tu et al. 1998; Xiao et al. 1998). Another IEG, Narp, mediates clustering of AMPA-type glutamate receptors at excitatory synapses (O’Brien et al. 2002; Xu et al. 2003).
The classically conditioned eyeblink response is one of the most extensively studied models for understanding the cellular mechanisms that underlie learning and memory (Bloedel and Bracha 1995; Christian and Thompson 2003). However, few studies have exploited this model to study the role of IEGs in this form of associative learning (Carrive et al. 1997; Irwin et al. 1992), and none have examined Arc. Using an in vitro model of classical conditioning, we have shown that GluR4-containing AMPA receptors are targeted to synaptic sites during conditioning by an NMDA receptor-dependent mechanism (Keifer 2001, 2003; Keifer and Clark 2003; Mokin and Keifer 2004). Electrical stimulation of the auditory nerve by a conditioned stimulus (CS) paired with an unconditioned stimulus (US) applied to the trigeminal nerve results in a neural analog of a conditioned eyeblink response recorded in the ipsilateral abducens nerve of an isolated brain stem preparation (Keifer 1993; Keifer et al. 1995). In the present study, we used this in vitro model of abducens classical conditioning to study the expression of Arc protein at two time points representing an early stage of conditioning during conditioned response (CR) acquisition and a later stage during CR expression. Immunocytochemical and Western blot analysis revealed that Arc protein was rapidly and transiently expressed during conditioning. Blockade of conditioning by the NMDA receptor antagonist 2-amino-5-phosphonopentanoic acid (AP-5) resulted in a decline of Arc to control values, suggesting that protein expression requires NMDA receptor activation. Furthermore, triple-label immunocytochemical studies of Arc, GluR4, and synaptophysin using confocal imaging of abducens motor neurons showed significant colocalization of these proteins during early stages of conditioning but not at later stages when Arc was no longer associated with GluR4 and synaptophysin. Finally, coimmunoprecipitation studies indicated that Arc has its strongest interaction with GluR4 and actin in the early stages of conditioning. Taken together, these data suggest a role for Arc protein in the synaptic delivery of GluR4-containing AMPA receptors by interactions with cytoskeletal protein complexes during the acquisition phase of in vitro classical conditioning.
Freshwater pond turtles, Pseudemys scripta elegans (n = 83), obtained from commercial suppliers were anesthetized by hypothermia and decapitated. Protocols involving the use of animals complied with the guidelines of the National Institutes of Health and the Institutional Animal Care and Use Committee. The brain stem was transected at the levels of the trochlear and glossopharyngeal nerves and the cerebellum was removed as described previously (Anderson and Keifer 1999). Therefore this preparation consisted only of the pons with the cerebellar circuitry and the red nucleus removed. The brain stem was bathed in physiological saline (2–4 ml/min) containing (in mM) 100 NaCl, 6 KCl, 40 NaHCO3, 2.6 CaCl2, 1.6 MgCl2, and 20 glucose, which was oxygenated with 95% O2-5% CO2 and maintained at room temperature (22–24°C) (Anderson and Keifer 1997, 1999). Suction electrodes were used for stimulation and recording of cranial nerves. The US was an approximately two-times-threshold single-shock stimulus applied to the trigeminal nerve; the CS was a subthreshold 100 Hz, 1 s train stimulus applied to the ipsilateral posterior root of the eighth nerve that was an average of 73% of the threshold amplitude required to produce activity in the abducens nerve (Anderson and Keifer 1997, 1999; Keifer 2001; Keifer et al. 1995). The latter nerve will be referred to as the auditory nerve as it carries predominantly auditory fibers. Neural activity was recorded from the ipsilateral abducens nerve which projects to the extraocular muscles controlling movements of the eye, nictitating membrane, and eyelid. The CS-US interval used in this study was 20 ms, which is defined as the time between the offset of the CS and the onset of the US. This interval was found to be optimal for conditioning (Keifer 2001). Intertrial interval between the paired stimuli was 30 s. A pairing session consisted of 50 CS-US presentations followed by a 30 min rest period in which there was no stimulation (Keifer et al. 1995). The abducens nerve response is a population burst discharge of abducens motor neuron pools in response to stimulation. As motor neurons are recruited to discharge in response to stronger stimulation, the level of discharge recorded extracellularly increases. Conditioned responses were defined as abducens nerve activity that occurred during the CS and exceeded the amplitude of double the baseline recording level. The total number of population burst responses out of 50 per session that exceeded this baseline level were recorded and are reported as percent of CRs for each pairing session. Conditioned preparations that received paired stimuli and pseudoconditioned controls that received unpaired stimuli in random order were presented with either two or five pairing sessions. At the end of the conditioning experiments, preparations for immunocytochemistry were immersion fixed in cold 3 or 0.5% (the latter for triple-labeling studies) paraformaldehyde, whereas those for Western blot analysis and coimmunoprecipitation were frozen in liquid nitrogen and stored at −70°C.
The NMDA receptor antagonist AP-5 (100 μM; Tocris Cookson, St. Louis, MO) was dissolved in physiological saline and perfused through the bath. Preparations underwent the conditioning procedure in normal physiological saline for one pairing session during which time CRs were exhibited. AP-5 was perfused through the bath at the beginning of the rest period 30 min prior to the second pairing session and remained in the bath for the second pairing session. Following this, preparations were fixed or frozen.
Arc immunocytochemistry and image analysis
To determine immunoreactivity of abducens motor neurons for Arc, four groups of preparations were examined, those that were conditioned for two (n = 6) or five (n = 5) pairing sessions and those that were presented with unpaired pseudoconditioning stimuli in random order for two (n = 6) or five (n = 5) sessions. Tissue sections were cut at 30 μm and preincubated in 10% normal goat serum for 1 h followed by incubation in primary antibody overnight at 4°C with gentle shaking. The primary antibody was a polyclonal raised in rabbit that recognizes Arc (sc-15325, Santa Cruz Biotechnology, Santa Cruz, CA; 1:1,000). Following the primary antibody, sections were incubated with a Cy3-conjugated goat anti-rabbit IgG secondary antibody for 2 h (Jackson ImmunoResearch, West Grove, PA; 1:100). The specificity of the Arc primary antibody was tested by performing Western blots on turtle and rat brain tissue (see results). The secondary antibodies were tested for their specificity by omission of the primary antibody or incubation with an incompatible secondary as described by Keifer and Carr (2000). To analyze immunoreactivity of the abducens motor nuclei to Arc, digital images of immunofluorescence were obtained using a Zeiss Axioskop microscope equipped with an Axiocam color camera. Image analysis was performed by using Adobe Photoshop software where the average luminosity level ± SD of the pixels in the selected region of interest was calculated (Keifer 2001; Keifer et al. 2003; Lindahl and Keifer 2004; Meyer et al. 2004). Each selected field encompassed the cell bodies of the principal or accessory abducens motor nuclei ipsilateral to the side that was conditioned and values of intensity of label were obtained for each tissue section analyzed. Background values were taken from unlabeled areas in the same section and were subtracted from signal in the abducens nuclei for every section analyzed to control for variations in staining. Generally, 6–10 sections were analyzed and averaged for each case. StatView software (SAS, Cary, NC) was used for statistical analysis using ANOVA and values are reported as means ± SE.
Triple-labeling studies and confocal microscopy
For triple-labeling studies of Arc-GluR4-synaptophysin or Arc-NR1-synaptophysin, five groups of preparations were examined, those that were presented with paired stimuli and were conditioned for two (GluR4, n = 5; NR1, n = 4) or five (GluR4, n = 5; NR1, n = 3) pairing sessions, those that were presented with unpaired pseudoconditioning stimuli for two (GluR4, n = 4; NR1, n = 4) or five (GluR4, n = 4; NR1, n = 3) sessions, and preparations that were conditioned and treated with AP-5 in the second pairing session (GluR4, n = 3). Tissue sections were prepared as described in the preceding text. The primary antibodies used were a polyclonal raised in rabbit that recognizes Arc (sc-15325, Santa Cruz Biotechnology; 1:1,000), a polyclonal raised in goat that recognizes the GluR4 subunit of the AMPA receptor (sc-7614, Santa Cruz; Biotechnology, 1:100), a polyclonal raised in goat that recognizes NR1 subunits of the NMDA receptor (sc-1467, Santa Cruz Biotechnology; 1:1,000), and a monoclonal antibody raised in mouse that recognizes synaptophysin (S-5768, Sigma; 1:1,000). After incubation in the primary antibodies overnight, sections were incubated with the appropriate Cy2-, Cy3-, or Cy5-conjugated secondary antibodies for 2 h (1:100). Images of triple-label staining of abducens motor neurons were obtained using an Olympus Fluoview 500 laser scanning confocal microscope. In a previous study using tract tracing combined with intracellular dye fills, Keifer and Mokin (2004) found that most of the trigeminal and auditory nerve terminals were on the soma and proximal dendrites of abducens motor neurons. Therefore these sites were targeted for analysis in the present study. Tissue samples were scanned using a ×60 1.4 NA oil-immersion objective with triple excitation using a 488 nm argon laser, a 543 nm HeNe laser, and a 633 nm HeNe laser. Punctate staining of at least twofold greater intensity than the diffuse fluorescence of the cell background was quantified (Mokin and Keifer 2004). Arc protein (Cy5, blue fluorescence), glutamate receptor subunits (Cy3, red), and the presynaptic marker synaptophysin (Cy2, green) were determined to be colocolized if blue, red and green pixels were adjacent to one another or overlapping (appearing as white pixels). The mean number of puncta for each protein or combination of proteins was determined per unit area of the region of interest in the image. The area of the selected field was measured using National Institutes of Health Image (http://rsb.info.nih.gov/nih-image). StatView software was used for statistical analysis and data are presented as means ± SE.
Western blot analysis
For Western blotting, brain stem preparations underwent conditioning for two (n = 5) or five (n = 5) sessions or pseudoconditioning for two (n = 5) or five (n = 5) sessions. A third group was treated with AP-5 in the second pairing session (n = 5). Tissue was homogenized and homogenates centrifuged at 1,000 g for 10 min at 4°C and supernatants recentrifuged at 20,000 g for 20 min. Pellets were resuspended in ice-cold HPLC grade water, spun at 7,600 g, resuspended again in HPLC grade water and clarified by centrifugation at 48,000 g for 20 min at 4°C. Final pellets were resuspended in 50 mM HEPES buffer (pH 7.4), alloquoted, and stored at −70°C. Protein sample concentrates were solubilized in 1× SDS/β-mercaptoethanol and boiled for 3 min prior to separation by 8% SDS-PAGE. After electrophoresis, membranes were blocked with 5% nonfat dry milk in PBS/0.1% Tween-20 for 4 h at 4°C. The membranes were incubated with primary antibodies used for immunocytochemistry (1:500, except for NR1, which was 1:1,000) overnight in PBS/0.1% Tween-20/0.1%BSA at 4°C, washed, and incubated with in HRP-conjugated goat anti-rabbit IgG (1:100,000) for 2 h at room temperature. Loading controls were performed using primary antibodies to actin (Chemicon, Temecula, CA; 1:500). Proteins were detected using the ECL-Plus chemiluminescence system (Amerisham Pharamcia, Piscataway, NJ). Immunoreactive signals were captured on Kodak X-omatic AR film and quantified by computer-assisted densitometry.
Brain stem preparations (n = 15) were homogenized in lysis buffer containing (in mM) 50 HEPES, 150 NaCl, 1 EDTA, 2.5 EGTA, 1 DDT, 1 NaF, 0.1 Na3SO4, 0.1 PMSF, 10% glycerol, and 0.1% Tween 20, at pH = 7.5. Protein samples (100 μg each) were incubated with either GluR4, NR1 or Arc primary antibodies (2 μg antibody/protein sample, same antibodies as used for Western blotting) or nonspecific IgGs (Santa Cruz) at 4°C for 8 h. Protein G-Agarose beads (Calbiochem) were rinsed with buffer (50 mM HEPES, 150 mM NaCl, pH = 7.5) three times, and 20 μl beads slurry was added to protein samples and incubated at 4°C for 2 h. After incubation, beads were washed three times with lysis buffer, solubilized in 2× SDS-β -mercaptoethanol and boiled for 5 min to elute the proteins. Eluates were separated by 6% SDS-PAGE, then transferred to PVDF membranes. The membranes were blocked with 5% nonfat dry milk, incubated with primary antibodies to Arc, actin or synaptophysin overnight, washed and incubated in HRP-conjugated secondary antibodies. Immunoreactivity was visualized by chemiluminescence.
All antibodies used in the present study were tested for their specificity by Western blot analysis of turtle and rat brain tissue (Fig. 1). Antibodies to Arc protein, GluR4 AMPA receptor subunits, synaptophysin, NR1 NMDA receptor subunits, and actin, all showed a single band at the appropriate molecular weight for both turtle and rat tissue (Fig. 1, A–E). These data indicate that the antibodies used in the present study maintained their specificity in brain tissue from turtles as they do in rat brain.
Arc immunoreactivity of abducens motor neurons is rapidly and transiently increased during conditioning
Immunoreactivity of the abducens motor neurons for Arc protein expression in conditioned and pseudoconditioned preparations was examined, and these data are illustrated in Fig. 2. Both the conditioned and pseudoconditioned groups were examined at two time points: in the early stages of conditioning after two pairing sessions (or for a total time period of 80 min after the onset of conditioning) and in the later stages of conditioning after five pairing sessions (after 4 h). The pseudoconditioned preparations were presented with the same number of stimuli as the conditioned group except that the CS and US were unpaired and presented in random order. A summary of the conditioning data for this series of experiments is shown in Fig. 2, A–C. A representative abducens nerve recording from a preparation that exhibited conditioning is shown in Fig. 2A. The recording shows an initial short latency burst discharge characteristic of abducens CRs in this in vitro brain stem preparation in response to an auditory nerve CS (Fig. 2A, arrow) followed by the trigeminal nerve evoked UR (the US is indicated by the •). Acquisition curves of abducens nerve CRs for preparations that were conditioned or pseudoconditioned are summarized in Fig. 2B for those stimulated for two sessions and in Fig. 2C for those stimulated for five sessions. Preparations that demonstrated conditioning after two pairing sessions showed a mean of 2 ± 6% CRs in the first session and 29 ± 25% CRs in the second session (• Fig. 2B). Those conditioned for five sessions showed an average of 61 ± 7% CRs in sessions two through five (Fig. 2C). The control preparations that were pseudoconditioned for two and five sessions all demonstrated a mean of 0 ± 0% CRs (○; Fig. 2, B and C). After the conditioning experiments, preparations were placed in fixative and quantitative analysis of the immunoreactivity of abducens motor neurons for Arc protein was performed. The photomicrographs in Fig. 2D illustrate immunoreactivity of accessory abducens nucleus neurons for Arc from a preparation that was pseudoconditioned for two pairing sessions and one that was conditioned for two sessions. Pseudoconditioned preparations showed light cytoplasmic label and very little labeling of dendrites. In contrast, conditioning resulted in increased intensity of cytoplasmic label and considerably enhanced labeling of the dendrites. Quantitative data are summarized in Fig. 2E. Findings for the principal and accessory abducens nucleus were not significantly different from one another, and therefore these data were combined. The intensity of Arc staining of the abducens nuclei was significantly increased after two pairing sessions in preparations that demonstrated conditioning compared with those that were either naive, or not stimulated, or pseudoconditioned (P < 0.001). Moreover, after two sessions of conditioning, there was a significant correlation between the number of abducens nerve CRs and enhanced Arc immunoreactivity of abducens motor neurons (R2 = 0.35, P < 0.01). However, there were no significant differences in the level of Arc immunoreactivity in preparations that were conditioned for five pairing sessions compared with those that were naive or pseudoconditioned for five sessions (P = 0.23). These conditioning-related changes in immunoreactivity of the abducens nucleus to Arc contrast with those obtained for another cell group that is part of the pontine abducens eyeblink circuit, the principal sensory trigeminal nucleus (Zhu and Keifer 2004). While these neurons were immunopositive for Arc, there were no significant changes in the immunoreactivity of this nucleus after conditioning compared with pseudoconditioned preparations (P = 0.10; data not shown). Therefore our findings are specific to abducens motor neurons and show that immunoreactivity to Arc is rapidly and transiently increased during in vitro conditioning.
Conditioning-induced increase in Arc protein expression and NMDA receptor dependency
Western blot analysis for Arc protein was performed on five groups of brain stem preparations, those that were conditioned for either two or five pairing sessions, those that were pseudoconditioned for two or five sessions, and those that underwent blockade of abducens nerve CRs by bath application of the NMDA receptor antagonist AP-5 in the second pairing session. These results are summarized in Fig. 3. Acquisition curves for the five groups are shown in Fig. 3, A–C. Preparations conditioned for two pairing sessions obtained a mean of 56 ± 38% CRs in the second session (Fig. 3A). Preparations conditioned for five sessions showed an asymptote of CRs in sessions three through five averaging 56 ± 26% CRs (Fig. 3B). Pseudoconditioned preparations all showed 0 ± 0% CRs. The final group of preparations was first conditioned for one pairing session to a mean of 18 ± 14% CRs. AP-5 (100 μM) was then applied to the bath which resulted in blockade of the CRs to a mean of 1 ± 2% in session two (Fig. 3C). Western blots were performed on all experimental groups, and these findings are summarized in Fig. 3D. A statistically significant increase of 229% in Arc protein from preparations that showed conditioning after two sessions compared with those that were pseudoconditioned for two sessions was observed (P < 0.0001). Representative bands for Arc are shown as are actin-loading controls. Preparations in which abducens nerve CRs were blocked with AP-5 showed levels of Arc protein similar to pseudoconditioned preparations suggesting that NMDA receptor activation is required for enhanced expression of Arc protein during conditioning. Preparations that were conditioned for five pairing sessions resulted in no significant increase in Arc protein compared with pseudoconditioned controls after five sessions and these values were not different from the pseudoconditioned or AP-5-treated groups after two sessions. Therefore there was a significant and transient increase in Arc protein expression during the early stages of conditioning that is sensitive to the NMDA receptor antagonist AP-5.
Arc protein is transiently colocalized with GluR4-containing AMPA receptors at synaptic sites during conditioning
Triple-label immunocytochemistry allowed us to investigate the relationship between Arc protein and GluR4 and NR1 glutamate receptor subunits at synaptic sites on abducens motor neurons using confocal microscopy. A previous study indicated that GluR4-containing AMPA receptor localization at synaptic sites was significantly enhanced after conditioning, whereas the synaptic localization of NR1-containing NMDA receptors was stable (Mokin and Keifer 2004). To investigate this pattern of conditioning-related subunit trafficking further and its relationship with Arc protein, a total of 193 abducens motor neurons triple labeled for Arc-GluR4-synaptophysin and 118 neurons triple labeled for Arc-NR1-synaptophysin were quantitatively analyzed for colocalization of punctate staining. Acquisition curves for conditioned and pseudoconditioned preparations from this data set were similar to those shown in Fig. 3, A–C. These findings are illustrated by the confocal images in Fig. 4 and quantitative data are shown in Figs. 5 and 6. In all preparations, Arc protein (blue), GluR4 subunits (red) and synaptophysin (green) appeared as punctate staining localized to the soma and dendritic branches of abducens motor neurons. Staining for all three proteins significantly increased in preparations that showed conditioning after two pairing sessions (Fig. 5, A–C; P < 0.0001). This finding is illustrated by the increased punctate staining of a neuron from a preparation that underwent conditioning for two pairing sessions shown in Fig. 4B compared with a neuron from a pseudoconditioned preparation in Fig. 4A. Moreover, colocalization of Arc with GluR4 (magenta pixels in Fig. 4B, arrows) was significantly increased (Fig. 5D; P < 0.0001), as was the colocalization of GluR4 with synaptophysin (yellow pixels in Fig. 4B, open arrowheads; Fig. 5E; P < 0.0001), indicating enhanced localization of GluR4-containing AMPA receptors and Arc protein at synaptic sites after two pairing sessions. Consequently, there was significantly increased triple label colocalization of Arc-GluR4-synaptophysin after conditioning for two pairing sessions (white pixels in Fig. 4B, solid arrowheads; Fig. 5F; P < 0.0001). By the fifth pairing session of conditioning, GluR4 colocalization with synaptophysin was significantly increased above the level observed after two pairing sessions (yellow pixels in Fig. 4C; Fig. 5E; P = 0.0003) corresponding with the increase in abducens nerve CRs. However, the colocalization of Arc with GluR4 subunits returned to control values by the fifth pairing session (Fig. 5D).
Triple-label studies of Arc-NR1-synaptophysin punctate staining confirmed previous findings (Mokin and Keifer 2004) that NR1 subunit number and synaptic localization did not change after conditioning. Neither the total number of puncta per unit area (Fig. 6A; P = 0.11) nor the colocalization of NR1 with synaptophysin (Fig. 6B; P = 0.42) were altered after two or five pairing sessions in conditioned compared with pseudoconditioned preparations. However, there was a significant increase in triple label colocalization of Arc-NR1-synaptophysin after two pairing sessions (Fig. 6C; P < 0.001). This is likely to be attributed to movement of Arc into synaptic sites.
AP-5 blocks conditioning-related increase in colocalization of Arc-GluR4-synaptophysin
We next examined the role of NMDA receptor activation in the transient colocalization of Arc with GluR4 at synaptic sites. Preparations treated with AP-5 demonstrated a total number of Arc, GluR4, and synaptophysin puncta similar to control values (Fig. 5, A–C). Also, AP-5 blocked the conditioning-related increase in colocalization of Arc with GluR4 subunits (Fig. 5D; P < 0.0001) and GluR4 with synaptophysin (Fig. 5E; P < 0.001), thereby reducing triple-label colocalization of these proteins (Fig. 5F; P < 0.0001). Taken together, the confocal imaging data reveal that early in conditioning during CR acquisition Arc is colocalized with GluR4 at synapses. However, this relationship is no longer apparent after five pairing sessions at which time GluR4 subunits remain synaptically localized during CR expression. These changes observed in localization of Arc and GluR4 require NMDA receptor activation as they are not observed after treatment with AP-5.
Conditioning-related coimmunoprecipitation of Arc with GluR4 and actin
To further examine possible interactions of Arc with GluR4 and NR1, we performed coimmunoprecipitation experiments from extracts of turtle brain stems pseudoconditioned for two sessions, conditioned for two sessions and conditioned for five sessions. To confirm the specificity of the coimmunoprecipitation experiments, immunoprecipitates were not observed after using nonspecific IgGs (Fig. 7, IgG lane). Coimmunoprecipitation of Arc and GluR4 AMPA receptor subunits or actin indicated a conditioning-related association between these proteins but not with NR1 NMDA receptor subunits. In pseudoconditioned preparations, weak association between Arc and GluR4 was detected as indicated by the presence of a sparse band for Arc that was precipitated with GluR4 antibody (Fig. 7A, pseudo 2 session). A greater degree of interaction between Arc and GluR4 was observed after two sessions of conditioning, as indicated by the denser Arc band compared with pseudoconditioned controls (Fig. 7, A and B; Cond 2 session). However, after five sessions of conditioning, the level of interaction between Arc and GluR4 declined to control levels (Cond 5 session). In contrast to the increased interaction of GluR4 with Arc early in conditioning, we did not observe any change in the interaction between NR1 and Arc in either of the conditioned groups compared with controls as shown by the similar bands across experimental conditions (Fig. 7C). Because the cytoskeletal protein actin is known to be associated with Arc (Lyford et al. 1995), we also performed coimmunoprecipitation of these two proteins to examine possible changes in their interaction during conditioning. There was a greater association of Arc with actin after two sessions of conditioning compared with pseudoconditioned preparations or those conditioned for five sessions (Fig. 7D), a pattern similar to findings for Arc and GluR4. As an additional control, we tested whether Arc immunoprecipitates with proteins abundant in the presynaptic terminal. Interactions between synaptophysin and Arc were examined (Fig. 7E) and failed to reveal any association between these proteins. Finally, immunoprecipitation for actin and consequent labeling with actin antibody was used as loading controls (Fig. 7F). Therefore the comimmunoprecipi-tation experiments confirm the confocal imaging studies that Arc and GluR4 interact in the early stages of conditioning during CR acquisition.
The present study demonstrates that during in vitro abducens classical conditioning there is a rapid and transient increase in levels of Arc protein that require activation of NMDA receptors. In the early phase of conditioning during CR acquisition, there is significantly greater colocalization of Arc protein and GluR4-containing AMPA receptors at synaptic sites, however, colocalization of Arc and GluR4 was not observed after later stages of conditioning during CR expression. There was also significantly enhanced coimmunoprecipitation of Arc with GluR4 subunits and actin early in conditioning but not of Arc with NR1 subunits, and these associations declined to control levels in later stages of conditioning. These data suggest that Arc protein, possibly through interactions with the cytoskeletal machinery present in the PSD, is involved in delivery of GluR4-containing AMPA receptors to synaptic sites during the acquisition phase of conditioning.
Transient increase in Arc protein during conditioning and NMDA receptor dependency
Many studies have shown that Arc protein expression in brain can be rapidly induced by various behavioral training paradigms and stimulation regimes. The hippocampus has been a primary site to study the characteristics and function of Arc protein expression because of its well-understood role in learning and the highly organized nature of its circuitry (Guzowski et al. 2000, 2001; Steward and Worley 2001; Steward et al. 1998). However, accumulating evidence suggests that other areas of the brain differentially express this protein in response to specific stimuli (Kelly and Deadwyler 2003). Using an in vitro model of classical conditioning, the present study found a significant increase in the immunoreactivity of abducens motor neurons for Arc and the amount of Arc protein after two pairing sessions or 80 min after the initiation of the conditioning protocol. These results are consistent with findings of others that Arc protein expression is rapid and can occur as early as within 30 min after synaptic stimulation (Donai et al. 2003; Guzowski et al. 2001; Steward and Worley 2001; Steward et al. 1998). Rapid expression of Arc is believed to be associated with local protein synthesis. That is, synapse-associated polyribosome complexes and mRNAs allow local translation in dendrites containing the target synapses (Steward and Levy 1982). In the later stages of conditioning, Arc protein rapidly declines to control levels after five pairing sessions or 4 h after conditioning onset. A similar pattern of rapid onset and transient expression in mRNA and protein are described for other IEGs such as Egr-1 and c-Fos (Guzowski et al. 2001). Our recent data show that Egr-1 protein expression rapidly increases early in conditioning similar to Arc but, in contrast to Arc, remains elevated during later phases of conditioning (Mokin and Keifer 2005). This finding indicates that Arc and Egr-1 are differentially expressed and may serve different functions in in vitro abducens classical conditioning. Significantly, the present study found that the increase in Arc protein requires activation of NMDA receptors. The addition of AP-5 to the bath prior to the second pairing session blocked CRs as well as the enhanced levels of Arc protein observed after conditioning. Previous studies have shown that both the induction and expression of abducens nerve CRs requires NMDA receptor activation as these processes are blocked by AP-5 (Keifer 2001; Keifer and Clark 2003). Consistent with our results for Arc, Steward and Worley (2001) have shown in hippocampus that Arc mRNA targeting depends on stimulation of NMDA receptors; in the presence of antagonist Arc, mRNA is diffusely distributed in dendrites rather than selectively localized to lamina containing activated synapses.
CR acquisition correlates with synaptic delivery of GluR4 and colocalization with Arc
Previously, we observed by confocal imaging of abducens motor neurons that GluR4-containing AMPA receptors were targeted to synaptic sites during in vitro classical conditioning while the synaptic localization of NR1 subunits was unchanged (Mokin and Keifer 2004). The synaptic localization of GluR4 subunits was shown to be dependent on the activation of NMDA receptors as colocalization with synaptophysin declined to pseudoconditioned control values after treatment with AP-5. However, that study examined receptor subunit localization only after five pairing sessions during the expression phase of conditioning and not earlier in conditioning during CR acquisition. Electrophysiological recordings show a significant increase in abducens nerve CR acquisition during the second pairing session. Thereafter the percent of CRs asymptotes to a mean of ∼60–70% CR expression that is typical for this preparation. Clearly significant cellular events that underlie abducens CR acquisition occur between the first and second pairing sessions. Due to the rapid and transient nature of Arc protein expression, the present imaging studies were performed during both the early and late phases of abducens conditioning. These findings revealed that in preparations conditioned for two pairing sessions colocalization of Arc, GluR4 subunits and synaptophysin was significantly increased over pseudoconditioned levels, indicating substantial delivery of both GluR4 and Arc to synaptic sites, corresponding with CR acquisition. Consistent with these findings, the interaction among GluR4, Arc and actin proteins as demonstrated by coimmunoprecipitation was greatest after two sessions of conditioning. In the later stages of conditioning, after five pairing sessions and during abducens CR expression, the synaptic localization of GluR4-containing AMPA receptors is further increased above levels observed after two sessions however, Arc colocalization with GluR4 declined to levels observed for pseudoconditioned preparations. In contrast to GluR4-containing AMPA receptors, there were no conditioning-related changes in NR1-containing NMDA receptor colocalization with synaptophysin observed by imaging. The increased levels in Arc-NR1-synaptophysin after two pairing sessions is likely to be due to the movement of Arc-GluR4 into synaptic sites. There were also no conditioning-related changes in NR1 coimmunoprecipitation with Arc, although there may be a linkage between the two proteins as shown by others (Husi et al. 2000). Consistent with the idea that the colocalization of GluR4 with Arc and the synaptic localization of GluR4 are related to the acquisition of abducens nerve CRs, such colocalization was not observed after application of AP-5, which inhibits CR acquisition.
Cellular interactions of Arc protein
Previous studies suggest that Arc may be linked with protein aggregates of the cytoskeleton. Arc bears some sequence homology with the cytoskeletal protein α-spectrin and coprecipitates with F-actin (Lyford et al. 1995) and microtubules (Fujimoto et al. 2004). This is consistent with our findings that Arc coprecipitates with actin in turtle brain tissue. It has also been observed in hippocampal cultures that GFP-Arc colocalizes with punctate staining for PSD-95 (Fujimoto et al. 2004) and therefore may interact with a number of postsynaptic scaffolding proteins and receptors. Among these, Arc has been shown to be structurally linked to NMDA receptor complexes in mouse brain (Husi et al. 2000), consistent with our coimmunoprecipitation findings from turtle brain. Although in our preparation Arc coimmunoprecipitates with GluR4 as well as NR1 subunits, it is presently unknown whether Arc binds directly to either AMPA or NMDA receptor proteins. It seems more likely that these linkages are indirect through macromolecular complexes localized to the PSD. Such proteins that contribute to receptor trafficking may include SAP97, the AKAPs, ABP, and GRIP (Collingridge et al. 2004). Important linking proteins related to potential conditioning mechanisms are the 4.1 family of proteins that have been shown to bind the actin cytoskeleton and AMPA receptors. Coleman et al. (2003) have demonstrated a direct interaction between GluR4 at its C-terminal domain and 4.1 proteins that is required for surface expression of GluR4-containing AMPA receptors. Interactions between Arc and these proteins remain to be identified.
Role of Arc in synaptic delivery of GluR4-containing AMPA receptors during acquisition of conditioning
Activity-dependent synaptic delivery of GluR4 was previously demonstrated in early postnatal hippocampal slices (Zhu et al. 2000). Moreover, synaptic incorporation required PKA phosphorylation of GluR4 at Ser842 (Esteban et al. 2003; Gomes et al. 2004). After in vitro abducens classical conditioning, we observed selective upregulation and synaptic insertion of GluR4 AMPA receptor subunits in the abducens motor nuclei that correlated with the level of conditioning (Keifer 2001; Mokin and Keifer 2004). The present study extends our understanding of GluR4 subunit trafficking during in vitro conditioning by examining the role of Arc protein during two different time points in conditioning, early in conditioning during CR acquisition and later during CR expression. In pseudoconditioned or naive preparations, electrophysiological studies suggest that auditory nerve synapses that convey the CS directly to abducens motor neurons (Keifer and Mokin 2004) contain only NMDA receptors and are functionally silent (Keifer 2003). This aspect of our model requires validation by further electrophysiological and imaging studies. During CR acquisition, Arc and GluR4 protein expression is increased, and Arc and GluR4 become colocalized at synaptic sites. One possibility suggested by our immunoprecipitation data for actin is that the conditioning-related synaptic targeting of GluR4 subunits occurs through interactions with protein complexes of the cytoskeletal and scaffolding machinery in the PSD. These processes involve signal transduction mechanisms involving NMDA receptors (Keifer 2001; Keifer and Clark 2003; Mokin and Keifer 2004; present study) and activation of protein kinases including PKA (unpublished data). We postulate that by five sessions, when CR expression has reached asymptote, Arc has already delivered GluR4 to synapses, is no longer required, and declines. To gain further insight into AMPA receptor trafficking in in vitro abducens classical conditioning and the role of Arc, it may be instructive to examine localization of these proteins after CR extinction and reacquisition. If Arc is involved in the synaptic delivery of GluR4 during conditioning, then it would be expected that Arc and GluR4 synaptic colocalization parallel the expression of abducens CRs during these phases of conditioning. It is also possible that Arc is involved in synaptic removal of receptors as well as their delivery. The role of the IEGs like Arc in processes of learning and memory are just beginning to be elucidated. It is widely believed that consolidation of long-term memories involves protein synthesis for their stabilization. In the system under study here, both Arc and GluR4 protein are induced early in the acquisition phase of conditioning and are inserted into synapses. However, the expression of Arc is transient. Other cellular mechanisms, such as activation of protein kinase systems or other IEGs like Egr-1, may contribute to stabilization of synapses after the period of synaptic modification. It is unclear exactly when synapses that have undergone conditioning-related changes in this preparation become relatively stable, or if they do, but studies using application of protein synthesis inhibitors may provide some insight. The data here suggest that Arc has a role in the acquisition phase of conditioning and that other mechanisms may be involved in longer-term stabilization of synaptic plasticity mechanisms.
This work was supported by the National Institutes of Health Grants MH-58709, NS-051187, and P20 RR-015567, which is designated as a Center of Biomedical Research Excellence (COBRE) to J. Keifer. J. S. Lindahl was supported by NIH Grant K08 MH-64552.
We thank Dr. Frances Day for assistance with the confocal microscopy and Dr. Keith Miskimins for discussions of the coimmunoprecipitation experiments.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 by the American Physiological Society