Sprague Dawley rats (3–6 months old, weight 400–900 g, in good health) of both sexes (n = 68 total) were assigned randomly to groups. Sex assignment was balanced in each group. Rats were singly housed in polycarbonate cages on a 12-h light/dark cycle with ad libitum access to food and water. All procedures were approved by the Memorial University Institutional Animal Care Committee and carried out in compliance with the guidelines of the Canadian Council on Animal Care.

Context A consisted of a custom-made olfactometer for air and odorant delivery attached to a plexiglass chamber that was positioned atop an electrified grid, connected to a shock generator/scrambler (Muromachi Kikai Model SGS-003DX). Polyvinyl carbonate bottles were used for each odor and connected to the olfactometer by C-flex tubing. Evacuation tubing with a fan was attached to the top lid to promote odor removal. Background white noise of 60 dB was played in the behavioral rooms during experiments. For auditory conditioning experiments, a pure tone (2 kHz, 80 dB) was played using computer speakers connected to a laptop which were placed on opposite walls outside of the conditioning chamber so the animals could not see them.

Context B was a plexiglass chamber covered in a checkerboard pattern in a separate behavioral room from Context A. Odor was delivered by soaking filter paper in odor and placing small pieces inside of fenestrated 15 mL conical tubes adhered to the corners.

In all cases the context was thoroughly cleaned with 70% ethanol between exposures and residual odor was extruded for at least 15 min before the next exposure.

Terpinene was diluted with mineral oil (6.63%) so that delivery would emit a vapor-phase partial pressure of 1 Pascal (Devore et al., 2013).

For standard habituation (Hsd), rats were placed in Context A for one 30 min session each on two consecutive days immediately prior to training. To avoid any negative association to the experimenter, rats were handled for 5 min each before entrance to and upon exit of Context A. For overnight habituation (Hon), rats were placed in Context A with hydrogel and food pellets and remained inside overnight followed by training the next morning in Context A.

Rats were trained individually with four separate exposures to either odor (O only) or odor and shock (O/S pair) at 5, 15, 20, and 30 min during a 30 min training session in Context A. Odorant (terpinene) was delivered for 1 min at each time point and co-terminated with a shock (0.5 mA for 1 s). “O only” animals were exposed to the odor in the same manner without being shocked. Animals were returned to their home cages immediately following the final odor exposure.

The day after training (recall day) all rats were individually exposed to terpinene (CS1) in Context A for 5 min while recording freezing behavior.

On day 1, control (context only during 2nd order conditioning; C only) and experimental (odor and context paired; O/C pair) rats were first order conditioned individually in Context A with four separate shocks (0.5 mA for 1 s) at 5, 15, 20, and 30 min during a 30 min training, giving Context A the value of conditioned stimulus 1 (CS1). Animals were returned to their home cages immediately following the final shock in Context A.

On day 2, “O/C pair” rats were second order conditioned individually by exposing them to terpinene for 5 min continuously in Context A such that an association was formed between the odor and the context, giving odor the value of conditioned stimulus 2 (CS2). “C only” rats were simply placed in Context A for 5 min. Animals were returned to their home cages immediately following training.

On day 3 (recall day) all rats were individually placed in novel Context B and exposed to terpinene (CS2) for 5 min while recording freezing behavior.

On the day proceeding habituation to Context A, experimental (tone and shock paired; T/S pair) rats were trained individually with four separate exposures to tone and shock at 5, 15, 20, and 30 min during a 30 min training session. Tone (2 kHz, 80 dB) was delivered for 1 min at each time point and co-terminated with a shock (0.5 mA for 1 s). Animals were returned to their home cages immediately following the last shock. Control rats (tone and shock unpaired; T/S unpair) were placed in Context A and given 4 shocks at 1, 3, 4, and 6 min. After 30 min, the 2 kHz tone was presented for 5 min.

The day after training (recall day) all rats were individually placed in Context A for 5 min, then exposed to the 2 kHz tone (CS) for 5 min while recording freezing behavior.

Following the final test, rats were returned to home cages undisturbed for 90 min to limit cFos expression unrelated to the behavioral experiment. Rats then received an i.p. injection of pentobarbital (200 mg/kg), and were transcardially perfused with 0.9% ice cold saline followed by 4% ice cold paraformaldehyde (PFA). Brains were carefully removed and placed in 4% PFA in glass vials for up to 1 week then transferred to PBS until sectioned.

Brains were serially sectioned coronally into 50 μm slices with a compresstome (PrecisionaryVF-210-0Z), starting from approximately −2.12 mm bregma, and collected in polyvinylpyrrolidone at 4°C for free-floating immunohistochemistry (IHC). Four to six brain sections spanning each region of the interest were selected from each animal for IHC.

Briefly, 50 μm sections were rinsed in Tris buffer and endogenous peroxidases were quenched by 1% H₂O₂ incubation. Sections were blocked in 10% normal goat serum with 0.1% Triton-x for 1 h prior to 4°C incubation in 1:10,000 cFos (Cell Signaling Technology, catalogue #2250S) for 2–3 days. Sections were rinsed in Tris buffer and incubated with 1:1000 biotinylated Goat anti-Rabbit IgG secondary antibody for 45 min at room temperature. The signal was amplified by an avidin-biotin peroxidase kit (ABC kit; Vector labs) for 2 h and then developed with an SG HRP substrate (SG grey; Vector labs). Sections were mounted onto slides, dried overnight under a fumehood, dehydrated by a series of ethanol and xylene, and coverslipped with Permount.

Images were taken in the lateral amygdala (LA), anterior basolateral amygdala (BLA), posterior piriform cortex (pPC), dorsal CA1 (dCA1), dorsal CA3 (dCA3), ventral CA1 (vCA1), ventral CA3 (vCA3) and auditory cortex layers II-III (Aud_II/III) and V (Aud_V), with an EVOS M5000 imaging system (Thermo Fisher Scientific) and a 10X objective with consistent brightness, exposure, and gain settings (see Supplementary Figure S1 for imaging locations). Four to six images covering the rostral to caudal range of each region were analyzed and data from two hemispheres were averaged for each animal. Cells were counted automatically using ImageJ software. Images were transformed to 8-bit, converted to black and white and thresholded manually by an experimenter who remained blind to the experimental conditions. The region of interest was selected, and “despeckle” and “watershed” functions were applied to further reduce background and separate incorrectly merged cells. The cells were counted using the “analyze particles” function with size set to 15-infinity μm² and circularity defined as 0.50–1.00. Cells/mm² was calculated by dividing the total number of positive cells by the total area.

For retrograde tracing experiments, Cholera Toxin subunit B (CTB) was infused bilaterally in the piriform cortex (200 nL; AP: −1.5 mm, ML: 5.6 mm bilateral, and DV: 8.7 mm from bregma) CTB-647 (0.5% w/v in phosphate buffer; Invitrogen) by a 32 g beveled 1 μL Hamilton syringe (Neuros 7001 KH) attached to a vertical infusion pump (Pump 11 Elite; Harvard Apparatus). Each infusion lasted 5 min, and remained in place for 5 min prior to syringe withdrawal. Rats were allowed 1 week for recovery before perfusion with 4% PFA. Brains were sectioned at 50 μm and cover-slipped with anti-fade fluorescence media with DAPI (abcam, catalogue # ab104139). CTB labeling in the range of DH and VH (AP 1.5 to 5.5) was examined using an EVOS M5000 system.

All statistics were performed with Origin 2022b software. Freezing data and cFos expression were first tested for normal distribution using Shapiro–Wilk test. Unpaired t-tests were used for two group comparisons if the data were distributed normally, and non-parametric Mann–Whitney tests were used for group data that were rejected in normality tests. For group data that did not have similar variance by F-tests, Welch correction was applied. Data are presented as mean ± SEM.

Olfactory threat conditioning was conducted following standard habituation (Figure 1A). Compared to “O only” rats, “O/S pair” rats froze for a significantly percentage of time (U = 0, Z = −3.86, p = 1.12E-4; Figure 1B) and expressed cFos in more cells in the anterior BLA (t_6.33 = −4.34, p = 0.004), the pPC (t₁₀ = −5.77, p = 4.18E-4), dCA1 (t₁₀ = −3.65, p = 0.0045), dCA3 (t_5.32 = −2.53, p = 0.049), vCA1 (t₈ = −10.57, p = 5.59E-6), and vCA3 (U = 0, Z = −2.51, p = 0.012; Figure 1C). The DH activation could indicate that either contextual cues were encoded to the threat memory trace or the DH participates in the recall of olfactory threat memories as part of the olfactory circuitry (Gourévitch et al., 2010). We probed these potential explanations by examining DH activation under two conditions: (1) intentional inclusion of contextual cues into the olfactory threat memory with second-order conditioning (SOC), and (2) uninterrupted context exposure through overnight habituation immediately followed by training. We used SOC to intentionally include contextual information in the odor threat memory (Figure 1D) as CS2 recall following SOC also activates CS1 representation areas in the brain (Hall, 1996). All rats were context conditioned (context CS1 + shock) with no prior habituation during phase I of the SOC. In phase II, the “O/C pair” groups were re-exposed to the context in the presence of an odor cue (CS2) to produce second-order threat conditioning, demonstrated by increased freezing behavior in response to the odor in a novel context the following day (t₁₂ = −4.19, p = 1.25E-3; Figure 1E). Subregional cFos expression in response to SOC threat memory recall (Figure 1F) was again enhanced in CA1 and CA3 of DH (CA1: t₁₂ = 3.56, p = 0.0039; CA3: U = 8, Z = −2.04, p = 0.041) and VH (CA1: t_7.83 = −4.94, p = 1.21E-4; CA3: t₁₂ = 5.80, p = 8.48E-5) in “O/C pair” rats, consistent with a role of DH in encoding of contextual cues. Finally, we minimized contextual coding by employing an overnight habituation protocol where the rats were kept in the shock chamber for 12 h followed by olfactory threat conditioning in the same context (Figure 1G). Again, “O/S pair” rats exhibited significantly more freezing than the “O only” group (t_4.21 = −12.06, p = 2.00E-4; Figure 1H). Intriguingly, when the contextual exposure was uninterrupted between habituation and conditioning, the DH was no longer activated in response to the odor threat (CA1: t₁₀ = −0.55, p = 0.60; CA3: t₁₀ = 0.049, p = 0.96), while VH activation was still observed (CA1: t₁₀ = −2.33, p = 0.042; CA3: t₁₀ = −2.59, p = 0.027; Figure 1I). Together, these experiments suggest that DH encodes contextual information, while VH is involved in olfactory threat conditioning regardless of context or DH activation. In both standard and overnight habituation protocols, pPC, BLA and VH are co-activated, consistent with the extensive mutual connections among these structures (Cenquizca and Swanson, 2007; McDonald and Mott, 2017). Projections from the VH to PC are demonstrated here by CTB retrograde tracing (Supplementary Figure S2).

We next tested if VH activation is involved in auditory threat conditioning. To mitigate cFos expression related to context, we used a design with equivalent exposures to shock, context, and sensory cues between “T/S paired” and “T/S unpaired” groups. Following Hsd (Figure 2A), “T/S paired” rats spent a higher percentage of time freezing than the “T/S unpaired” rats (U = 0, Z ₌ − 2.67, p = 0.007; Figure 2B) in response to the conditioned tone. Intriguingly, only cFos expression in the LA was significantly higher in the “T/S paired” group (t₉ = −2.48, p = 3.51E-2; Figure 2C), consistent with the role of the LA in auditory threat conditioning (Quirk et al., 1995; Nader et al., 2001). Because both groups of rats were shocked in the context in which they were tested, the opportunity for contextual conditioning was equal across groups, and cFos activation between groups is likely specific to the auditory cue. Even though the “T/S unpaired” rats showed a significantly lower percentage of freezing than the “T/S paired” rats, the freezing level was higher than the odor only rats with the same standard habituation method (Supplementary Figure S3), suggesting weak context conditioning occurred with these rats. Overnight habituation with auditory threat conditioning (Figure 2D) yielded similar results. “T/S paired” group spent more time freezing than the control group (t₁₀ = −10.32, p = 1.19E-6; Figure 2E) and increased cFos was observed in the LA (t₁₀ = −3.07, p = 1.18E-2; Figure 2F). Conversely, cFos expression did not differ between groups in all other regions in either of the habituation conditions (Figures 2C,F), suggesting that the hippocampus is not critically involved in the recall of auditory threat memory.