Supplementary Figure 1. Screening for monoclonal antibodies against GluA1 by immunoblotting.

Transcription

1 Supplementary Figure 1 Screening for monoclonal antibodies against GluA1 by immunoblotting. Hippocampal extract was subjected to western blotting with the hybridoma supernatants of candidate monoclonal antibodies against GluA1. Positive antibodies are labeled with an asterisk. First upper left lane was blotted with an anti- GluA1 monoclonal antibody (Millipore, C3T, raised against the GluA1 cytoplasmic tail: anti-glua1-ct). Second upper left lane was blotted with an anti-gfp monoclonal antibody (anti-gfp). To check the weak positive samples, we also examined Z9131 and Z9141 in staining shown in figure 1c. To test the negative samples, we also examined Z in figure 1c.

2 Supplementary Figure 2 Basic conditions of the CALI experiment with Z9139-eosin in vitro and effect of NASPM. a. Power and time dependence of light irradiation on the CALI effect with Z9139-eosin in vitro. Glutamate puff-induced responses were obtained in GluA1 expressing CHO cells with the indicated combinations of time and power for CALI. The data for 0 mw and 7.5 mw/cm 2, 1 min were the same as in Figure 2a. F(4,31)=5.618, *p<0.001, **p<0.05 and p=0.296 in 2.5mW/cm 2-1min vs 0mW (one-way ANOVA followed by post hoc analysis with the Fisher s PLSD test). n=5, 7, 6, 8 and 8 cells for 0mW, 2.5mW-1min, 5mW-1min, 7.5mW-1min and 7.5mW-0.5min, respectively. b. A saturating concentration of NASPM reduced the synaptic AMPA receptor-mediated currents in hippocampal primary cultures ~50%. AMPA receptor-mediated synaptic currents after NASPM application at different concentrations; values are the ratio of the current after to before NASPM application. Note that the effect of a saturating concentration (20 M) of NASPM on synaptic AMPA receptor currents was examined in the experiment in Figure 3d. F(5,49)=9.805, *p<0.001, **p<0.01 and p=0.273 in 0.01 vs 0 M (one-way ANOVA followed by post hoc analysis with the Fisher s PLSD test). n=11, 7, 8, 11, 9 and 9 cells for 0 M, 0.01 M, 0.1 M, 0.5 M, 5 M and 20 M, respectively. n.s. indicates no significance.

3 Supplementary Figure 3 Endosomal trafficking is not affected by CALI with Z9139-eosin. Fluorescence images of Alexa488-transferrin before and after recycling for 30 min with or without CALI using Z9139- eosin. The relative fluorescence intensity was examined before and after recycling in the dendrites and cell body (n=30 regions of interest [ROIs]). ROIs were obtained from 3 independent experiments. Scale bar indicates 5 m. t(52.769)=1.002, p=0.321 (unpaired two-tailed t-test). n.s. indicates no significance.

4 Supplementary Figure 4 Basic properties of the IA task-related experiment. a. Expression of GluA1 cytoplasmic tail in the hippocampus prevents the acquisition of IA memory. Latency period for re-entering the dark box of IA-trained mice expressing GFP or GFP-ctail (cytoplasmic terminus of GluA1: a peptide that blocks synaptic GluA1 delivery) in the CA1 region of the dorsal hippocampus. The latency period for re-entering the dark box was significantly shorter in the GFP-ctail-expressing animals than in the GFP-expressing ones. *p<0.001 (Mann-Whitney U test). n=9 and 7 animals for GFP and c-tail, respectively. b. Membrane properties of the examined neurons (n=14 cells shown in Figure 4b) in slices with or without IA conditioning. t(26)=1.127, p =0.270 in Cm, t(26)=0.028, p =0.781 in Rm and t(26)=0.139, p =0.891 in Ra (unpaired two-tailed t-test). n.s. indicates no significance.

5 Supplementary Figure 5 Infusion of Z9139 into the hippocampus in vivo. a. (Left) Bright field image of a brain slice including the hippocampus obtained from an animal in which a cannula was implanted. (Right) Location of the cannula labeled with DiI. b. Distribution of injected Alexa546-labeled Z9139 in the hippocampus of mice in which a cannula was implanted. To visualize the Z9139 antibody, it was labeled with Alexa546. Z9139-Alexa546 was bilaterally injected at the same stereotaxic coordinates. After 7.5 hours (the same duration as in the in vivo CALI experiment), the mice were fixed by perfusion, and the brain was removed and sliced (100 m) on a microtome. The coordinates m are the distance from the anterior end of the hippocampus. The intensity of the Alexa-546 signal was plotted along the CA1 region (yellow dashed line) in the right panels. Scale bar indicates 200 m. The Z9139 antibody diffused approximately 400 m laterally and 200 m sagittally in the CA1 region. c. Quantitative analysis of the alexa546-labeled Z9139 diffusion using the slices at 1000 and 1100 m obtained from 4 animals. d. Location of whole-cell recordings after in vivo CALI. Right panel is an enlarged view of the red box in the left panel. Yellow box indicates the position of the light cannula, determined by the injury in the cortex. We performed whole-cell patch clamp experiments (Figures 4e, 5b) within the white box, which was 150 m bilaterally, from just under the cannula (asterisk) along the CA1 region. Scale bars are 1 mm (left panel) and 150 m (right panel).

6 Supplementary Figure 6 in vivo CALI with Z9139-eosin under different conditions. a. Power and time dependence of the laser irradiation on the in vivo CALI effect. Latency for re-entering the dark box after the IA task in Z9139-eosin-treated animals with CALI was measured under the indicated conditions. The data for 0 mw and 60 mw, 2 min were same as in Figure 4c. *p<0.005, p=0.188 in 60mW-1min and p=0.854 in 30mW-2min vs 0mW (Kraskal-Wails test by post hoc analysis with Dunne s test). n=6, 8, 6 and 10 animals for 0mW, 60mW-1min, 30mW-2min and 60mW-2min, respectively. n.s. indicates no significance. b. The in vivo CALI effect with Z9139-eosin was retained for 1 week. The latency period for re-entering the dark box of IA-trained mice 1 week after treatment with or without CALI by Z9139-eosin. *p<0.05 (Mann-Whitney U test). n=11 and 15 animals for CALI- and CALI+, respectively. c. In vivo CALI with Z9139-eosin induced fear memory erasure in 8-week-old ICR mice. Eight-week-old ICR mice were subjected to in vivo CALI with Z9139-eosin. The experimental conditions were same as those used for 4-week-old ICR mice, shown in Figure 4c. *p<0.005 in Z9139 and p=0.852 in anti-myc (Mann-Whitney U test). n=6, 8, 7 nad 8 animals for Myc CALI-, Myc CALI+, Z9139 CALI- and Z9139 CALI+, respectively. n.s. indicates no significance.

7 Supplementary Figure 7 In vivo CALI with Z9139-eosin did not induce acute damage in memory-erased animals. Acute effect of in vivo CALI with Z9139-eosin was examined by re-conditioning 30 min after in vivo CALI. We reconditioned the mice whose IA memory we had erased by in vivo CALI with Z9139-eosin (IA test 1 & re-cond.) and tested them by IA test 2. Latencies for re-entering the dark box after the IA task in Z9139-eosin-treated animals without or with (IA test 1) CALI and in mice subjected to re-conditioning after Z9139-CALI (IA test 2). n=5 (CALI-) and n=8 (IA test1 and test2) animals. Data were analyzed by Kraskal-Wails test by post hoc analysis with Dunne s test. *p<0.01 vs CALI- and p=0.352 in IA test2 vs CALI-. n.s. indicates no significance.

8 Supplementary Figure 8 Specificity of in vivo CALI with Z9139-eosin for GluA1 homomeric receptors. a. GluA1 homomers were not detectable in synapses before learning. Responses at the hippocampal CA3-CA1 synapses in the acute brain slices obtained from IA-trained or untrained animals with or without NASPM. Average AMPA/NMDA ratio with or without NASPM before and after learning. t(18)=2.662, *p<0.05 in 1hr after learning and t(14)=0.667, p=0.515 in before learning (unpaired two-tailed t-test). n=8, 8, 9 and 11 cells for before learning DW, before learning NASPM, after learning DW and after learning NASPM, respectively. b. In vivo CALI with Z9139-eosin did not inactivate the GluA1 homomer-unrelated AMPA response. Responses at hippocampal CA3-CA1 synapses in the acute brain slices obtained from IA-untrained animals with or without in vivo CALI. Average AMPA/NMDA ratio with or without in vivo CALI. t(12)=0.388, p=0.705 (unpaired two-tailed t-test). n=8 cells, respectively. In Supplementary Figure 8a, the results of DW(before learning), NASPM(before learning), DW(after learning) and NASPM(after learning) were obtained from 6, 3, 5 and 3 animals, respectively. In Supplementary Figure 8b, the results of CALI- and CALI+ were obtained from 4 animals, respectively. n.s. indicates no significance.

9 Supplementary Figure 9 Re-learning of the IA task increased the synaptic delivery of homomeric GluA1. Synaptic responses at the hippocampal CA3-CA1 synapses in acute brain slices obtained from animals with or without reconditioning 0.5 h or 24 h after in vivo CALI. Average rectification index at hippocampal CA3-CA1 synapses with or without re-conditioning after in vivo CALI. t(14)=2.650, *p<0.05 in 24hr and t(15)=1.427, p=0.174 in 0.5hr (unpaired twotailed t-test). n=8, 9, 7 and 9 cells for 0.5hr re-cond-, 0.5hr re-cond+, 24hr re-cond- and 24hr re-cond+, respectively. The results of 0.5hr(re-cond-), 0.5hr(re-cond+), 24hr(re-cond-) and 24hr(re-cond+) were obtained from 3, 4, 4 and 3 animals, respectively. n.s. indicates no significance.

10 Supplementary Figure 10 Spatial specificity of in vivo CALI with Z9139-eosin. a. Expression of GluA1 cytoplasmic tail in the auditory cortex did not prevent the acquisition of IA memory. Latency period for re-entering the dark box of IA-trained mice expressing GFP or GFP-ctail in the auditory cortex. p= n=5 and 6 animals for GFP and c-tail, respectively. b. In vivo CALI with Z9139-eosin did not erase hippocampus-dependent IA fear memory. Latency period for re-entering the dark box of IA-trained mice with or without in vivo CALI. p= n=9 and 11 animals for CALI- and CALI+, respectively. Data were analyzed by Mann-Whitney U test. n.s. indicates no significance.

11 Supplementary Figure 11 Full-length of western blot by Z9139, anti-glua2, anti-glua3 and anti-actin antibody Full length immunoblots shown in Figure 1e. In GluA3 blots, asterisk indicates three lanes which was used in Figure 1e.

12 Supplementary Figure 12 Full-length of western blot by anti-arc and anti-gapdh antibodies. Full length immunoblots shown in Figure 4g. In GAPDH blots, asterisk indicates two lanes which was used in Figure 4g.