Supporting Online Material
Materials and Methods
Generation of transgenic mice. For construction of the fos-tTA transgene (S1), the tTA gene was inserted into a plasmid carrying artificial introns and a polyadenylation signal from SV40 (S2) and then subcloned into the AccI site of the first exon of the mouse c-fos gene (S3). To generate the tetO-GFP-GluR1 transgene, EGFP-GluR1 was similarly flanked by introns and an SV40 polyadenylation signal and placed downstream of the tetO promoter element of plasmid pMM400 (S2). The transgenes were excised from vectors and were purified by agarose gel electroelution and Elutip minicolumn. Purified DNA fragments were microinjected (2ng/μl) into BALB/cByJ and C57BL/6J F2 hybrid embryos and implanted into pseudopregnant females. The fos-tTA construct was co-injected with a fos-nlsGFP transgene (c-fos promoter driven EGFP with nuclear localizing signals) and segregated as a single transgene. Genotypes were determined by PCR with genomic DNA isolated from tails using the following primers: htTaF23 (5’-ACCTGGACATGCTGTGATAA-3’), htTaB22 (5’-TGCTCCCATTCATCAGTTCC-3’), MM400F1 (5’-ATAGAAGACACCGGGACCGAT-3’), and GFPR2 (5’-TTCAGCTCGATGCGGTTCAC-3’). The PCR reaction was run at 94 °C for 30sec, 55 °C for 30sec, and 72 °C for 1min for 35 cycles. All transgenic mice were maintained as heterozygotes and were backcrossed to C57BL/6J mice. All mice were kept on a Dox diet (40mg/kg) beginning from the time of weaning (3~4 weeks old) and used for experiments at 11~14 weeks old. Only males were used for experiments.
Fear conditioning. Mice were individually housed for 1 week prior to the experiments. For fear conditioning, mice were placed in a novel rectangle chamber with a grid-floor. After a 3 min baseline period, 4 tone-shock pairings were presented. Each pairing consisted of a 20 sec 85 dB 2800 Hz tone simultaneously ending with a 2 sec 0.75 mA shock. There was a 1 min interval between each pairing. The mice remained in the chamber for 1 min after the last shock before being returned to the home cage. The total duration of training was 500 sec. Freezing behavior was scored and analyzed automatically by a video-based system (S1). For the “context-only” condition, the mice were placed in the same fear conditioning chamber for 500 sec and received the same tones but did not receive foot shock.
Unpaired shock training. An immediate shock alone does not induce c-fos expression in the hippocampus, even if mice are exposed to a new context for the same amount of time as fear conditioning training (data not shown, S4, S5). In order to induce expression of c-fos promoter driven GFP-GluR1 in the hippocampus, mice were first placed in a novel chamber (context A: square plastic chamber surrounded with white walls, plastic floor with sani-chips) for 500 sec without foot-shocks. After the context A exposure, they were returned to their home cages. Fifteen min later, they received 4 shocks (2 sec 0.75 mA/shock) immediately after being placed in the conditioning chamber (context B: rectangle chamber, plexiglass front and back with black-and-white cross-stripes pattern and aluminium side walls, grid-floor). There was a 4 sec interval between shocks. After the final shock, they were returned to their home cages.
Contextual memory extinction training. Mice received 3 extinction trials in a day. Mice were exposed to the chamber where fear conditioning was conducted for 30 min for each trial without shocks. After each training, they were returned to their home cages. There was a 2 hour interval between trials.
Immunoblot. Hippocampal lysates (15μg of total protein) were separated by SDS–PAGE and transferred onto Hybond-P membrane. Membranes were blocked with 5% skim milk in TBST for 1 hour and probed with rabbit anti-GluR1 polyclonal antibody or mouse anti-GluR2 monoclonal antibody, followed by peroxidase-conjugated AffiniPure goat anti-rabbit IgG or peroxidase-conjugated AffiniPure donkey anti-mouse IgG. Signals were developed using the ECL PLUS system and exposed to Hyperfilm ECL. Immunoreactive intensities were analyzed with ImageJ software.
Immunoprecipitation. Hippocampi of the GFP-GluR1c-fos Tg mice were prepared 24 hours after fear conditioning. Hippocampi of wild-type and Tg mice were homogenized in lysis buffer (20mM HEPES, pH 7.5, 150mM NaCl, 1% Triton X-100, protease inhibitors cocktail) and incubated for 1 hour at 4 °C. The homogenate was cleared by centrifugation for 10 min at 15,000 g. The supernatant was pre-cleared with protein A agarose, then incubated with rabbit anti-GFP antibody attached to protein A agarose for 2 hours at 4 °C. The precipitated agarose was washed with lysis buffer 5 times and processed for immunoblotting with mouse anti-GluR2 monoclonal antibody.
Immunohistochemistry. Sagittal brain slices were cut at a thickness of 100 μm using a vibratome and fixed with cold 4% paraformaldehyde in PBS for 1 hour. Free-floating slices were permeabilized with 0.15% of TritonX-100 in 5% BSA/PBS at room temperature for 30 min, then incubated with rabbit anti-GFP antibody at 4 °C overnight, rinsed with PBS, and thereafter incubated with goat anti-rabbit Alexa 488 secondary antibody at 4 °C overnight. For visualization of nuclei, slices were counterstained with TO-PRO-3 and mounted in Slowfade Light antifade mounting medium. For surface GFP-GluR1 staining, slices were not permeabilized.
DiI labeling. Micropipettes were coated with DiI (1,1'-dioctadecyl-3,3,3'3'-tetramethylindocarbocyanine perchlorate) dissolved in ethanol at 10 mg/ml. The tip of DiI-coated micropipettes was inserted at several positions in the CA1 pyramidal cell layer of fixed slices using a micromanipulator. Slices were then placed in PBS for 3~4 days at 4 °C to allow the DiI to spread throughout the dendritic arbor.
Acquisition and analysis of spine images. Dendrites in stratum oriens (basal dendrites) and stratum radiatum (apical dendrites) were randomly selected for imaging if they were strongly double-labeled with GFP immunoreactive signal and DiI fluorescence. Primary dendrites were excluded from the analysis. Selected dendrites were imaged using a confocal microscope by sequential illumination with Argon (488 nm) and HeNe (543 nm) laser. Under the condition of laser power and PMT we used, intrinsic GFP fluorescence was too dim to detect. Images were acquired as a series of optical z-sections at 0.4 μm increments using an oil immersion 63× objective (NA 1.4) with a zoom of 2. Each optical section was scanned 3 times and Kalman filtering was employed to reduce noise. The stack images were converted to merged TIFF images of green channel (GFP immunoreactivity) and red channel (DiI fluorescence) for every focal plane using ImageJ software. They were exported to Adobe Photoshop for analysis. GFP-positive spines were defined as the existence of overlapping GFP signals on spines, as marked by yellow in the confocal images. GFP signals were often observed at the head of spines as a punctate signal. Spine morphology and density was analyzed on two-dimensional projection images constructed using a maximum intensity method. Spines were classified according to previously described criteria (S6): the “thin” type has a long neck and a small head; the “stubby” type has a large head but does not have a neck; the “mushroom” type has a large head with a neck. For spine density measurement, all protrusions along 30 μm dendritic segments were counted as spines (For apical dendrites, 58, 53, 55, 56, and 61 segments were examined for UP24h, CT24h, FC24h, FC72h, and FC72h+EXT, respectively. For basal, 55, 53, 52, 51, and 68 segments were examined for UP24h, CT24h, FC24h, FC72h, and FC72h+EXT, respectively). Analysis was performed blind with respect to the experimental conditions.
fig. S1. Induction and trafficking of newly synthesized GFP-GluR1 in the hippocampus of GFP-GluR1c-fos Tg mice. Mice were removed from Dox for 4 days and fear conditioned to examine the earliest time point for GFP-GluR1 induction in the soma and the trafficking rate of induced GFP-GluR1 in dendrites of the CA1 field of the dorsal hippocampus. Mice were examined at 1, 2, and 6 hours following training. Brain slices were fixed, pearmeabilized with 0.15% TritonX-100, then stained with anti-GFP antibody (green) and TO-PRO-3 nuclear marker (blue). Images were acquired using a confocal microscope. At 1 hour after stimulus (FC 1h), GFP-GluR1 expression was slight in the soma and dendritic layer, which is similar to home cage controls (home). At 2 hours (FC 2h), prominent GFP-GluR1 expression was detected in the soma but the dendritic expression was slight and limited to the proximal region. At 6 hours (FC 6h), GFP-GluR1 expression was detected in the soma and throughout the dendritic layer. These results suggest that the GFP-GluR1 is expressed by 2 hours in the soma in response to behavioral stimuli and transported to the distal dendrites within at least 6 hours in CA1 hippocampal neurons of GFP-GluR1c-fos Tg mice.
fig. S2. Metabolic stability of GFP-GluR1 in the hippocampus of GFP-GluR1CaMKIIα Tg mice. (A) Mice in which GFP-GluR1 expression was driven broadly in the hippocampus by a CaMKIIα-tTA transgene (S2) were fed a regular diet without Dox for more than 2 months to induce maximal expression of GFP-GluR1. They were then switched to a Dox diet (6g/kg chow for the first 2 days and then to 40mg/kg) for the period of days indicated to suppress new expression of GFP-GluR1. Sagittal brain slices were cut at a thickness of 100μm and fixed with cold 4% paraformaldehyde in PBS for 1 hour. Images of GFP fluorescence in the dorsal hippocampus were acquired using a fluorescent microscope with the same exposure time. (B) Relative GFP fluorescent intensities (%) in the hippocampus over time after Dox treatment. Data are represented as mean ± s.e.m. (n = 2 for each condition).
fig. S3. Contextual memory retention test after fear conditioning and unpaired-shock training. (A) Behavioral protocol. GFP-GluR1c-fos Tg animals used in the memory test are different animals that are analyzed for GFP-GluR1+ spines in because of the absence of memory test in the experiments of . Mice (n = 7 females, 20~27 weeks old) were trained with fear conditioning in context B (rectangle chamber, plexiglass front and back with black-and-white cross-stripes pattern and aluminium side walls, grid-floor), then returned to the context B for 3 min for a contextual memory test at 24 hours after training (FC). For unpaired shock training (UP), mice (n = 7 females, 20~27 weeks old) were first exposed to context A (square plastic chamber surrounded with white walls, plastic floor with sani-chips) for 500 sec without shocks, then received immediate shocks in context B. Twenty-four hours after training, they were returned to the context A, then B for 3 min for memory tests. (B) Percentage of time spent freezing. The UP group showed low levels of freezing for both context A and B, indicating that the unpaired-shock training caused less long-term fear memory associated with context B where shocks were given and context A where c-fos promoter driven GFP-GluR1 expression was induced. Data are represented as mean ± s.e.m.
fig. S4. Visualization of surface GFP-GluR1 in the hippocampal slice. (A) Predicted membrane topology of GFP-GluR1. (B) For surface GFP-GluR1 labeling, sagittal brain slices of GFP-GluR1CaMKIIα Tg mice were fixed with 4% PFA on ice for 1 hour. They were then incubated with rabbit anti-GFP polyclonal antibody, rinsed with PBS, and thereafter incubated with anti-rabbit Alexa594 antibody (red). To detect total GFP-GluR1, the slices were subsequently permeabilized with 0.15% Triton X-100 for 30min, and labeled with rabbit anti-GFP polyclonal antibody and anti-rabbit Alexa488 antibody (green). Confocal hippocampal images were acquired using a confocal microscope. Note that the lack of cytoplasmic labeling in the CA1 pyramidal cells before permeabilization (red), indicating the method labels the plasma membrane surface GFP-GluR1 but not internal receptor. Scale bar, 30μm. (C) To verify the ability of fixed tissue to exclude antiserum until permeabilized, brain slices of GFP-GluR1c-fos Tg mice (FC 24h) were stained with anti-GluR1 polyclonal antibody to a C-terminal epitope with and without permeabilization (0.15% Triton X-100 for 30min). Confocal hippocampal images were acquired with identical condition. Scale bar, 200μm.
S1. L. G. Reijmers, B. L. Perkins, N. Matsuo, M. Mayford, Science 317, 1230−1233 (2007).
S2. M. Mayford et al., Science 274, 1678−1683 (1996).
S3. K. Schilling, D. Luk, J. I. Morgan, T. Curran, Proc Natl Acad Sci U S A 88, 5665−5669 (1991).
S4. S. Milanovic et al., Brain Res 784, 37−47 (1998).
S5. N. C. Huff et al., J Neurosci 26, 1616−1623 (2006).
S6. K. M. Harris, F. E. Jensen, B. Tsao, J Neurosci 12, 2685−2705 (1992).