Encoding of contextual fear memory in hippocampal–amygdala circuit

In contextual fear conditioning, experimental subjects learn to associate a neutral context with an aversive stimulus and display fear responses to a context that predicts danger. Although the hippocampal–amygdala pathway has been implicated in the retrieval of contextual fear memory, the mechanism by which fear memory is encoded in this circuit has not been investigated. Here, we show that activity in the ventral CA1 (vCA1) hippocampal projections to the basal amygdala (BA), paired with aversive stimuli, contributes to encoding conditioned fear memory. Contextual fear conditioning induced selective strengthening of a subset of vCA1–BA synapses, which was prevented under anisomycin-induced retrograde amnesia. Moreover, a subpopulation of BA neurons receives stronger monosynaptic inputs from context-responding vCA1 neurons, whose activity was required for contextual fear learning and synaptic potentiation in the vCA1–BA pathway. Our study suggests that synaptic strengthening of vCA1 inputs conveying contextual information to a subset of BA neurons contributes to encoding adaptive fear memory for the threat-predictive context.

(b) Mice were exposed to a novel context (red bars) after tamoxifen injection (Tam) to label vCA1 neurons.
(d) Images showing the auditory cortex with no eYFP-labeled soma (left) and the basal amygdala with eYFP+ axons of labeled vCA1 neurons (right). Red, Nissl stain.
(e)-(f) Experimental setup for (g). After tamoxifen injection, mice were exposed to Context A to label with tdTomato vCA1 neurons active in Context A as in Fig. 3e. Mice were then exposed to Context A (A-A group) or Context B (A-B group) before brain fixation for c-Fos immunostaining (c-Fos IHC). (b) A week after virus injection surgery, mice in Label 1x group received a tamoxifen injection and exposed to Context A as in Fig. 3b. Mice in Label 2x group underwent one additional context labeling session, whereas mice in Label 3x group underwent two additional labeling sessions with a 1-week interval as in Fig. 3h.
Recording experiments were performed three weeks after the first labeling session.
(c) EPSC traces in the Label 1x group. Each EPSC trace indicates synaptic responses recorded in individual BA neurons (n = 19 neurons). EPSCs were induced by photostimulation of the same intensity (20.0 mW/mm 2 ) and recorded in randomly selected BA neurons.

(f)
Comparison of the average EPSC amplitude between groups (** p = 0.008, *** p =0.00008, two-sided Kruskal-Wallis multiple comparisons). Open circles indicate the peak amplitude of EPSC recorded in each BA neuron. n = 19 cells for Label 1x and Label 2x groups. n = 20 cells for Label 3x group.  (c) Trace representing the average of 50 current traces induced by hyperpolarizing voltage pulses as in (b) and recorded in a putative principal neuron in the BA (eYFP-, inset; scale bar, 10 µm).
(d) Scatter plot of membrane capacitance and input resistance calculated in eYFP-labeled GABAergic interneurons (red, 22 cells) and putative principal neurons (PN) examined in Fig. 4 and 5 (black, 349 cells). As membrane capacitance was less than 100 pF (dotted vertical line) in all the interneurons examined, recorded neurons with a membrane capacitance larger than 100 pF were regarded as principal neurons.
Source data are provided as a Source Data file.  Error bars represent the SEM. Source data are provided as a Source Data file.

Supplementary Figure 10. Comparison of synaptic efficacy in Context A vCA1 inputs to CeA pathway
(a) Experimental setup (b)-(e). AAV-DIO-ChR2-eYFP was injected into the vCA1 in Fos-CreER T2 mice. EPSCs in Context A vCA1 inputs to the CeA were recorded and compared between groups.
(b) Left: mice were exposed to Context A to label vCA1 neurons active in the context as in Fig. 4b. After labeling, mice in the FC group were trained for discriminative fear in Context A on Days 1-5 as in Supplementary Figure  1b. Mice in the NS control group were exposed to the contexts as in the FC group but did not receive the US.
(c) Quantification of freezing behavior in the FC (6 mice) and NS groups (5 mice) on Day 5. (e) Left: traces of EPSCs induced by paired photostimulations of nonspecific vCA1 axons with a 50 ms interval. n = 14 cells from 3 mice for each group. Right: there was no significant difference in the paired-pulse ratio between groups (main effect of groups, p = 0.10; group × intensity interaction, p = 0.67, two-way ANOVA).
Error bars represent the SEM. Source data are provided as a Source Data file.

Supplementary Figure 12. Comparison of excitatory and inhibitory synaptic responses in Context A vCA1 inputs to BA fear neurons versus other BA neurons (a) Experimental setup for (b)-(h)
. AAV-DIO-ChR2 was injected into the vCA1 in Fos-CreER T2 x ROSA-LSL-tdTomato mice.
(b) Left: mice were exposed to Context A for ChR2 expression in vCA1 neurons active in Context A. On Day 1, the mice were fear conditioned in Context A for tdTomato (tdT) expression in BA fear neurons. On Days 2-5, mice were trained for discriminative fear in Context A as in Fig. 5b. Right: quantification of freezing behavior in Contexts (Ctx) A and B on Day 5. n = 5 mice.
(c) Traces of AMPA receptor-mediated EPSCs induced by photostimulation of Context A vCA1 inputs (20.0 mW/mm 2 ; 1 ms duration; blue vertical bars) and recorded in tdT-and tdT+ BA neurons at holding potentials (Vh) of −80 mV. EPSCs were recorded in the presence of SR-95531.

(d) Comparison of the average amplitude of AMPAR EPSCs recorded in tdT-(17 cells) and tdT+ neurons (17 cells, two-sided unpaired t-test) as in (c).
(e) Traces of NMDA receptor-mediated EPSCs induced by photostimulation of Context A vCA1 inputs (20.0 mW/mm 2 ; 1 ms duration) and recorded in tdT-and tdT+ BA neurons at Vh = +40 mV. EPSCs were recorded in the presence of SR-95531.
(g) Traces of inhibitory postsynaptic currents (IPSC) in the vCA1-BA feed-forward inhibitory circuit. IPSCs were evoked with photostimulation of Context A vCA1 inputs (20.0 mW/mm 2 ; 1 ms duration) and recorded in tdT+ and tdT-BA neurons at Vh = 0 mV. IPSCs were recorded in the absence of SR-95531. (b) Left: after tamoxifen injection, mice were exposed to Context A for 12 minutes without the unconditioned stimulus (US, shock) three times at 1-week intervals to induce ChR2 and tdT expression in vCA1 and BA neurons active in Context A, respectively. On Days 2-5, mice were exposed to Contexts (Ctx) A and B without the US. Right: quantification of freezing responses on Day 5. n = 4 mice.
(c) Traces of EPSCs recorded in tdT-and tdT+ BA neurons. EPSCs were induced and recorded as in Fig. 4d.
(e) Quantification of difference in the AMPA/NMDA ratio between tdT+ and tdT-neurons in mice with context exposure alone as in (b) (13 pairs) versus mice fear conditioned in Context A in Fig. 5a-g (18 pairs) (two-sided unpaired t-test).
Error bars represent the SEM. Source data are provided as a Source Data file. Fig. 6a-g. Mice received three footshocks (0.5 mA, 2 s duration) in Context A 14 hours after tamoxifen administration. Immediately after fear conditioning, mice received an intraperitoneal injection of anisomycin (150 mg/kg body weight, ANI) or saline (SAL) and returned to their home cages. Mice received 3 more injections of anisomycin (50 mg/kg body weight) or saline at 2-hour intervals. Eighteen hours after the last injection, mice were tested for freezing behavior in Context A.

(b)
Diagram showing a behavioral protocol for (c)-(d). Mice received three footshocks in Context A. Immediately after fear conditioning, mice received an intraperitoneal injection of anisomycin (150 mg/kg body weight, ANI) or saline (SAL) and returned to their home cages. Ninety minutes after fear conditioning, the brain tissue was fixed for c-Fos immunohistochemistry. (b) Mice were given tamoxifen (Tam) injection and exposed to Context A twice with a 1-week interval to induce ChR2 expression in vCA1 neurons active in Context A. After a week, the mice received the third tamoxifen injection. Fourteen hours after tamoxifen administration, the mice received an intraperitoneal injection of MK-801 or saline (SAL) and fear conditioned in Context A 30 minutes later, which induced tdTomato (tdT) expression in BA neurons active during contextual fear conditioning.
(c) Comparison of freezing behavior in Context A 24 hours after fear conditioning. Two-sided unpaired t-test. n = 6 mice per group. (c) Comparison of mCherry+ BA neuron density in Fos-CreER T2 versus Arc-CreER T2 mice. p = 0.07 (two-sided unpaired t-test.

(d)
Experimental setup for (f)-(g). AAV-DIO-mCherry was injected into the vCA1 in Fos-CreER T2 mice. After an intraperitoneal injection of tamoxifen, mice were exposed to novel Context A three times for 12 minutes each to label vCA1 neurons active in Context A as in Fig. 3b.
(e) Left: experimental setup for (f)-(g). AAV-DIO-mCherry was injected into the vCA1 in Arc-CreER T2 mice. After surgery, mice were exposed to novel Context A once for 12 minutes and received an intraperitoneal injection of 4-hydroxytamoxifen (4-OHT) or vehicle 10 minutes later. Right: microscopic images showing mCherry expression (red) in a population of vCA1 neurons in mice that received 4-OHT injection after context exposure but not in mice that received vehicle injection. Blue, Nissl stain.

(d) Comparison of the density of TVA-G-GFP-labeled BA neurons between the FC and HC groups. Two-sided unpaired t-test.
Error bars represent the SEM. Source data are provided as a Source Data file.

Supplementary Figure 19. Contextual fear conditioning did not affect neuronal excitability or intrinsic membrane properties of context-specific vCA1 neurons.
(a) Left: experimental setup. After tamoxifen injection (Tam), Fos-CreER T2 × ROSA-LSL-tdTomato mice were exposed to Context A to label with tdTomato (tdT) vCA1 neurons active in Context A. Four days after vCA1 labeling, the mice were fear conditioned in Context A. The mice were tested for freezing behavior in Context A 24 hours later (Recall). Electrophysiological recordings (E-phys) were performed 2 days after memory recall test. Right: quantification of freezing behavior in Context A 24 hours after contextual fear conditioning. n = 3 mice.
(c) Representative traces of action potential (AP) firing induced by depolarizing current injection (500 ms long) and recorded in tdT-and tdT+ vCA1 neurons in current-clamp mode. Baseline membrane potential was adjusted to approximate -85 mV. tdT+ neurons were identified with red fluorescence (inset; scale bar, 10 µm).
(d) Summary plot of AP firing in tdT-and tdT+ vCA1 neurons. The average number of APs was plotted against injected currents. There was no significant difference in AP firings between tdT-and tdT+ vCA1 neurons (p = 0.67; repeated measures two-way ANOVA). n = 17 cells per group from 3 mice.
(e) Comparison of resting membrane potential (RMP, left) and input resistance (Rin, right) in tdT-and tdT+ vCA1 neurons. Input resistance was calculated as in Supplementary Figure 8. There was no significant difference in RMP (p = 0.94, or Rin between tdT-and tdT+ vCA1 neurons (p = 0.91, two-sided unpaired t-test). n = 18 cells per group from 3 mice.
Error bars represent the SEM. Source data are provided as a Source Data file. (b) Top: mice were exposed to Context A or B and injected with 4-OHT. After 7-11 days, mice were exposed to Context A, and the brain tissue was fixed 90 minutes later. Bottom: images showing vCA1 neurons labeled with mCherry (red) and shGFP (green). vCA1 neurons active during both the first and second context exposures expressed both mCherry and shGFP (circled).

Supplementary
(c) Comparison of the proportion of shGFP+ cells among all mCherry+ vCA1 neurons (* p < 0.05, two-sided paired t-test, n = 5 pairs of mice). Open circles indicate the proportion in each mouse, and closed circles are the average proportion.
Error bars represent the SEM. Source data are provided as a Source Data file.

Supplementary Figure 21. Vehicle injection did not affect contextual fear learning in mice with hM4Di expression in Context A vCA1 neurons.
(a) vCA1 neurons active in Context A expressed hM4Di-mCherry.
(b) Behavioral training and testing protocols for (c). Three weeks after labeling vCA1 neurons active in Context A with hM4Di-mCherry, mice received a vehicle injection 30 minutes before fear conditioning in Context A on Day 1 (vehicle 1) and Day 3 (vehicle 2). The mice were tested for fear memory in Context A on Days 2 and 4.
(c) Comparison of freezing behavior in Context A on Day 2 versus Day 4. There was no significant difference in freezing behavior on Day 2 versus Day 4 (p = 0.52, two-sided paired t-test; 9 mice), suggesting that the CNO effect in the hM4Di group in Fig. 9a-c was not due to the order of CNO and vehicle injections before fear conditioning.
Error bars indicate the SEM. Source data are provided as a Source Data file.

Supplementary Figure 22. Summary diagrams (a)
The context-shock pairing during contextual fear conditioning induces long-term synaptic potentiation (LTP) in the vCA1-BA pathway. A subset of vCA1 neurons is active in a specific context (e.g., Context A or Ctx A), whereas a neuronal population in the BA responds to the aversive stimulus (e.g., shock). During memory recall, vCA1 neurons responding to the threat-predictive context activate BA neurons more readily through the strengthened vCA1-BA pathway, inducing fear responses to the context. Fig. 4, contextual fear conditioning induces selective strengthening in vCA1 inputs that convey threat-predictive context (Ctx A) signals to the BA (red open circle, LTP).

(b) As demonstrated in
(c) As demonstrated in Fig. 5, 7, and 10a-g, contextual fear conditioning selectively strengthens synapses (red open circle, LTP) that consist of presynaptic vCA1 neurons active in the threat-predictive context (Ctx A, blue filled circle) and postsynaptic BA fear neurons active during fear conditioning (red filled circle).
(d) Synaptic potentiation in the vCA1-BA pathway is blocked under anisomycin-induced retrograde amnesia or MK-801-induced anterograde amnesia as demonstrated in Fig. 6a-g and Supplementary Figure 15.
(e) Silencing of context-specific vCA1 neurons during fear conditioning inhibits both synaptic potentiation in the vCA1-BA pathway and fear memory formation as shown in Fig. 10h-

Figure 8l
Repeated

Figure 3f
Paired tdTomato+ cells Context A-A vs Context A-B 5 0.42 0.689 (n = 6 pairs of mice)

Figure 3f
Paired c-Fos+ cells Context A-A vs Context A-B 5 0.68 0.528 (n = 6 pairs of mice)

Figure 3f
Paired Fos+ and tdT+ cells Context A-A vs Context A-B 5 3.65 0.015 (n = 6 pairs of mice)

Figure 3f
Paired Fos+ proportion Context A-A vs Context A-B 5 4.47 0.007 among tdT+ cells (n = 6 pairs of mice)