Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice iScience . 2026 Mar 13;29(4):115317. doi: 10.1016/j.isci.2026.115317 Search in PMC Search in PubMed View in NLM Catalog Add to search Response dynamics of discrete subiculum→retrosplenial cortex projections underlying trace fear conditioning Thomas E Bassett Thomas E Bassett 1 Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY, USA Find articles by Thomas E Bassett 1 , Yi-Zhi Wang Yi-Zhi Wang 3 Department of Neurology, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA Find articles by Yi-Zhi Wang 3 , Elizabeth M Wood Elizabeth M Wood 1 Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY, USA Find articles by Elizabeth M Wood 1 , Hui Zhang Hui Zhang 1 Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY, USA Find articles by Hui Zhang 1 , Zorica Petrovic Zorica Petrovic 1 Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY, USA Find articles by Zorica Petrovic 1 , Vivien Prifti Vivien Prifti 1 Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY, USA Find articles by Vivien Prifti 1 , Vladimir Jovasevic Vladimir Jovasevic 2 Department of Pharmacology, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA Find articles by Vladimir Jovasevic 2 , Naoki Yamawaki Naoki Yamawaki 4 Department of Biomedicine, Aarhus University, Aarhus, Denmark 5 PROMEMO, Aarhus University, Aarhus, Denmark 6 DANDRITE, Aarhus University, Aarhus, Denmark Find articles by Naoki Yamawaki 4, 5, 6 , Lynn Ren Lynn Ren 7 Department of Psychiatry and Behavioral Sciences, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA Find articles by Lynn Ren 7 , Natalia Khalatyan Natalia Khalatyan 3 Department of Neurology, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA Find articles by Natalia Khalatyan 3 , Viktoriya Grayson Viktoriya Grayson 7 Department of Psychiatry and Behavioral Sciences, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA Find articles by Viktoriya Grayson 7 , Jeffrey N Savas Jeffrey N Savas 3 Department of Neurology, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA Find articles by Jeffrey N Savas 3 , Jelena Radulovic Jelena Radulovic 1 Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY, USA 4 Department of Biomedicine, Aarhus University, Aarhus, Denmark 5 PROMEMO, Aarhus University, Aarhus, Denmark 6 DANDRITE, Aarhus University, Aarhus, Denmark Find articles by Jelena Radulovic 1, 4, 5, 6, ∗ , Ana Cicvaric Ana Cicvaric 1 Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY, USA Find articles by Ana Cicvaric 1, 8, ∗∗ Author information Article notes Copyright and License information 1 Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, Bronx, NY, USA 2 Department of Pharmacology, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA 3 Department of Neurology, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA 4 Department of Biomedicine, Aarhus University, Aarhus, Denmark 5 PROMEMO, Aarhus University, Aarhus, Denmark 6 DANDRITE, Aarhus University, Aarhus, Denmark 7 Department of Psychiatry and Behavioral Sciences, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA ∗ Corresponding author [email protected] ∗∗ Corresponding author [email protected] 8 Lead contact Received 2025 Jul 24; Revised 2025 Oct 20; Accepted 2026 Mar 6; Collection date 2026 Apr 17. © 2026 The Author(s) This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). PMC Copyright notice PMCID: PMC13068534 PMID: 41971990 Previous version available: This article is based on a previously available preprint posted on bioRxiv on June 15, 2025: " Response dynamics of discrete subiculum→retrosplenial cortex projections underlying trace fear conditioning ". Summary Associating events separated in time depends on the CA1, subiculum (SUB), and retrosplenial cortex (RSP). The degree to which their connectivity and underlying circuit mechanisms contribute to the association of such temporally discontiguous events is not known. Here we showed, using trace fear conditioning (TFC), wherein mice learn to associate tone and shock separated by a temporal trace, that molecularly distinct excitatory VGluT1 + and VGluT2 + SUB→RSP projections subserve the associative and temporal components of TFC. During trace memory formation, VGluT2 + SUB→RSP projections showed increased and decreased bulk calcium activity at tone and trace onset, respectively, an activity pattern that was re-established during memory recall. Such pattern was not observed after pseudoconditioning, suggesting that learning-specific patterns for the associative and temporal components of TFC emerge at the level of SUB→RSP synapses before being presented to the RSP. Our findings establish a circuit mechanism for representing complex temporal information in episodic memory. Subject areas: neuroscience, omics, sensory neuroscience Graphical abstract Open in a new tab Highlights • VGluT1 + and VGluT2 + hippocampal outputs support distinct aspects of trace fear • VGluT2 + SUB→RSP afferents exhibit phase-specific dynamics during trace fear • VGluT2 + response dynamics is specific to trace fear learning Neuroscience; Omics; Sensory neuroscience Introduction Episodic memory is defined as a form of consciously recallable long-term memory encompassing details of personally experienced events, including their temporal, spatial, and relational properties. 1 , 2 Beyond simply situating a memory in time, the temporal component of episodic memory holds information regarding the order in which events happened, their durations, and the time intervals between them. 3 Collectively, these temporal properties allow individuals to integrate discrete events separated in time into a cohesive and meaningful memory. The capacity to bind discontinuous events can be tested in rodent models with paradigms such as trace fear conditioning (TFC), in which animals learn to associate a tone and foot-shock separated by a temporal gap. 4 , 5 , 6 TFC requires coordinated activity across multiple components of both hippocampal complex and cerebral cortex to support proper memory formation and recall. 4 , 7 Suppressing activity of the dorsal hippocampal CA1 subfield significantly impairs both conditioning and recall in TFC, 8 , 9 , 10 while the entorhinal cortex contributes to TFC through NMDAR-mediated glutamatergic projections into the CA1. 4 , 11 , 12 Silencing the dorsal CA1 projections to the dorsal subiculum (SUB) during the temporal gap impairs TFC memory formation 13 and projections from the SUB back into the EC are necessary for TFC memory recall, thus forming a critical circuit within the hippocampal formation underpinning TFC. 10 , 13 In the cortex, the retrosplenial cortex (RSP) contributes significantly to TFC memory formation and extinction, 14 , 15 , 16 and plays a key role in episodic memory as a temporal processing hub. 17 However, while the individual functions of the hippocampal formation and the RSP are well-established, if and how these regions communicate during TFC remains unexplored. Our previous work demonstrates that the hippocampal complex interfaces with the RSP via inhibitory projection from the CA1 18 and glutamatergic projections originating from the SUB, comprising of vesicular glutamate transporter 1 (VGluT1 + ) and vesicular glutamate transporter 2 (VGluT2 + ) positive pyramidal neurons, 19 , 20 with different contributions to the formation of associative context memories. 18 , 19 Using TFC in combination with chemogenetic inhibition, we found that while SUB→RSP VGluT1 + and VGluT2 + projections redundantly encoded the aversive tone-shock association, only VGluT2 + projections were required for the processing of the temporal trace. Using fiber photometry, we showed that SUB→RSP VGluT2 + afferents exhibited biphasic calcium transients, with an early calcium peak occurring at tone onset, followed by a pronounced dip during the trace period. This pattern emerged during training, reappeared during recall, and was absent in the pseudoconditioned control group, demonstrating that VGluT2 + SUB→RSP activity patterns specifically encoded the temporal gap. Results VGluT2 + SUB→RSP afferents are specifically required for processing a temporal trace separating a cue and a shock In delay fear conditioning (DFC), mice are exposed to a presentation of a tone terminating with footshock, whereas TFC introduces a temporal gap (trace) between them. In TFC, unlike DFC, the activation of dorsal hippocampus (DH) 21 or RSP 15 is necessary for the association of tone and shock. To test whether the communication between these two regions is specifically necessary for the formation of the tone-shock association in TFC, we chemogenetically silenced all SUB→RSP excitatory projections using designer receptors exclusively activated by designer drugs (DREADD) during training in both TFC and DFC. We injected the DH of C57BL/6N male mice with Cre-independent inhibitory DREADD AAV8- hM4D (Gi) -mCherry virus and 6 weeks later microinfused the RSP with clozapine-N-oxide (CNO) through bilateral cannulas to locally block presynaptic release from RSP hM4D (Gi) -expressing axon terminals without modifying spiking activity of the cell soma ( Figure 1 A). 22 Our chemogenetic manipulations targeted the SUB→RSP projections, which are the major source of DH afferents to RSP ( Figures S1 A and S1B). 19 In TFC-trained mice, pre-training CNO infusions significantly impaired freezing during the tone and trace at tests, compared with vehicle control ( Figure 1 B), suggesting that SUB→RSP projections were necessary for the formation of tone-shock association. On the other hand, CNO injection did not affect freezing in DFC-trained mice expressing Cre-independent inhibitory DREADD AAV8-hM4 (Gi) -mCherry ( Figure 1 C). As expected, CNO infusions also had no effect on freezing mice expressing the control virus AAV8-mCherry ( Figure S2 A). Additionally, at test, DFC trained mice did not freeze post tone, unlike mice trained in TFC paradigm, indicating that one-trial tone-shock pairing allows for clear differentiation between TFC and DFC at the behavioral level ( Figure S2 B). Together, these findings showed that SUB→RSP projections are necessary for TFC, but not DFC. Figure 1. Open in a new tab VGluT2 + SUB→RSP afferents are specifically required for processing a temporal trace separating a cue and a shock (A, D, and G) Experimental design of behavioral tasks. Diagrams depict virus infusion sites in DH and cannula placements for CNO injections in RSP (top and right), and viral expression in RSP and DH (bottom and right). (B) When compared to vehicle, CNO injections 30 min before TFC training, impaired freezing at test in response to both the tone and trace periods in mice receiving AAV-hM4D (Gi) ( n = 9–10; two-way ANOVA with repeated measures; factor: treatment, p = 0.0098, F (1, 17) = 8.444, factor: phase, p < 0.0001, F (2, 34) = 67.41, factor: trial × phase, p = 0.0084, F (2, 34) = 5.522). (C) CNO injection before training in delay fear conditioning (DFC) did not affect freezing to tone or post-tone when compared to vehicle-injected mice ( n = 9–10; two-way ANOVA with repeated measures; factor: treatment, p = 0.6738, F (1, 17) = 0.1835, factor: phase, p < 0.0001, F (2, 34) = 97.84, factor: trial × phase, p = 0.1764, F (2, 34) = 1.827). (E) In VGluT2-Cre mice, CNO injection 30 min before training significantly impaired freezing during the trace but not tone compared to vehicle group ( n = 10; two-way ANOVA with repeated measures; factor: treatment, p = 0.0132, F (1, 18) = 7.565, factor: phase, p < 0.0001, F (2, 36) = 41.47, factor: trial × phase, p = 0.0001, F (2, 36) = 12.00. (F) In VGluT1-Cre mice, terminal silencing with CNO did not affect freezing compared to vehicle controls ( n = 9; two-way ANOVA with repeated measures; factor: treatment, p = 0.0931, F (1, 16) = 3.189, factor: phase, p < 0.0001, F (2, 32) = 13.78, factor: trial × phase, p = 0.9968, F (2, 32) = 0.003253) when compared to vehicle controls. (H) Injections of CNO ( n = 10) before trace-light conditioning (TLC) training did not affect freezing during either the tone or light periods in mice expressing inhibitory DREADD only in VGluT2 + SUB→RSP projections ( n = 10; two-way ANOVA with repeated measures; factor: treatment, p = 0.2577, F (1, 18) = 1.366, factor: phase, p < 0.0001, F (2, 36) = 52.27, factor: trial × phase, p = 0.5714, F (2, 36) = 0.5685). (I) Injections of CNO before TLC training significantly impaired freezing to both tone and light when compared to vehicle in mice expressing inhibitory DREADD in all SUB→RSP projections compared to vehicle injected group ( n = 5–6; two-way ANOVA with repeated measures; factor: treatment, p = 0.0087, F (1, 9) = 11.12, factor: phase, p < 0.0001, F (2, 18) = 19.01, factor: trial × phase, p = 0.3846, F (2, 18) = 1.008. Data presented as mean ± SEM ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001, ∗∗∗∗ p < 0.0001; NS, not significant; WT, wild-type. All scale bars, 250 μm. Our previous work demonstrated that the two excitatory SUB→RSP layer 3 projections differentially contribute to formation and persistence of recent and remote context memories. 19 Thus, we next sought to determine, using selective chemogenetic silencing of VGluT1 + or VGluT2 + SUB→RSP terminals, whether this functional difference is extended to TFC. We injected the DH of male mice expressing Cre recombinase under control of either VGlut1 (VGluT1-Cre mice) 23 or VGluT2 (VGlut2-Cre mice) 24 promoter with Cre-dependent inhibitory DREADD AAV8-DIO-hM4D (Gi) -mCherry virus and 6 weeks later locally microinfused RSP with CNO through bilateral cannula before TFC training ( Figure 1 D). Silencing VGluT2 + terminals during training specifically impaired freezing during trace compared with control group ( Figure 1 E), while silencing VGluT1 + terminals during training had no effect on freezing ( Figure 1 F). Silencing SUB→RSP projections selectively during testing had the same effect ( Figure S1 C), indicating that exclusively activity of VGluT2 + terminals is required both for formation and recall of memory during trace. On the other hand, context memories ( Figure S1 D) required activity of VGluT1 + terminals, consistent with our previous findings. 19 We excluded a sex effect by replicating these findings in female animals ( Figures S1 E and S1F). To assess if the freezing impairment during trace caused by silencing VGluT2 + SUB→RSP terminals disrupted encoding of the temporal gap or event sequences, we used tone-light conditioning (TLC) paradigm. TLC is conceptually similar to TFC, except that the temporal gap separating two intermittent events in TFC is replaced by another neutral stimulus (flashing light) in TLC. Therefore, in TLC mice are presented with a sequence of three events (tone-light-shock) rather than two events separated in time as in TFC (tone-trace-shock). Following training, we quantified freezing in both TFC and TLC conditioned mice; first exposing them to a tone-trace, and then the following day to a tone-light sequence. After both TFC and TLC, mice acquired fear conditioning to the tone, but freezing during the trace period was significantly higher in mice trained in TFC, whereas freezing during the light period was significantly higher in mice trained in TLC, suggesting that mice formed different associations (tone-trace-shock vs. tone-light-shock) in these paradigms ( Figure S2 C). Next, we injected either VGluT2-Cre mice with Cre-dependent or C57BL/6N Cre-independent inhibitory DREADD and 6 weeks later, locally injected CNO in RSP to silence all excitatory or specifically VGluT2 + SUB→RSP afferents, respectively ( Figure 1 G). Inhibition of VGluT2 + SUB→RSP afferents did not affect freezing during TLC tests ( Figure 1 H) but inhibiting all excitatory SUB→RSP terminals significantly impaired freezing during both tone and light presentation compared with controls ( Figure 1 I). This indicated that the trace freezing deficits induced by silencing VGluT2 + SUB→RSP projections were not a consequence of impaired processing of stimuli sequences but likely reflected their inability to encode the temporal gap between tone and shock. Characterization of VGluT1 + and VGluT2 + neurons in the SUB→RSP circuit Previous studies have shown that throughout the brain, VGluT1 and VGluT2 exhibit largely complementary expression, 25 with neurons of the cerebellum, cortex, and hippocampus primarily expressing VGluT1 positive, and subcortical brain areas mostly expressing VGluT2. 26 To identify the neuronal origins of these projections, we used genetically modified VGluT1-Cre or VGlut2-Cre mice injected with Cre-dependent color flipping retrograde reporter virus to label these discrete populations. Using this approach, where Cre + population express GFP, and Cre − population express TdTomato ( Figure 2 A), we found that CA1 and the SUB area have both VGluT2 (Cre + ) and putative VGluT1 (Cre − ) populations of neurons and that VGluT2 neurons are predominantly located in CA1 the deep layer, whereas in the SUB VGluT2 and putative VGluT1 neurons are spread heterogeneously ( Figure 1 B). Retrograde viral injections of Cre-dependent color flipping reporter virus either in SUB or RSP, revealed that both VGluT1 and VGluT2 neurons of the CA1 project to the SUB and that VGluT1 and VGluT2 neurons in the SUB project to the RSP ( Figures 2 C and 2D). Analysis of the online database Hippseq revealed that VGluT2 expression was limited to several SUB clusters (fibronectin 1 [Fn1], Ly6g6e, and S100b, Figure S5 ). 27 , 28 Previous studies show that Fn1 is enriched in distal SUB, which provides main input to RSP, 29 indicating that the VGluT2 + cells targeted in our study are likely Fn1 positive ( Figure S5 ). Figure 2. Open in a new tab Characterization of VGluT1 + and VGluT2 + neurons in the SUB→RSP circuit (A) Schematic of Cre-Switch viral vector (pAAV-Ef1a-DO_DIO-TdTomato_EGFP-WPRE). TdTomato is expressed in Cre-negative cells, while in Cre-positive cells, Cre recombination inverts and excises the TdTomato cassette, leading to EGFP expression. (B) Cre + (VGluT2 + , green) and Cre − (red, presumably VGluT1 + ) hippocampal neurons visualized after injection of the Cre-dependent color flipping nuclear reporter (pAAV8-EF1a-Nuc-flox(mCherry)-EGFP) into the DH. (C) Retrograde Cre-dependent “color flipping” reporter (pAAV-Ef1a-DO_DIO-TdTomato_EGFP-WPRE) was injected into RSP of VGluT1-Cre or VGluT2-Cre mice. Red and green signals in the SUB show that both VGluT1 + and VGluT2 + population are projecting to the RSP. (D) Retrograde Cre-dependent “color flipping” reporter was injected into SUB of VGluT1-Cre or VGluT2-Cre mice. Green and red signals in DH and SUB indicate presence of both VGluT1 and VGluT2 populations of neurons in both areas as well as presence of VGluT1 + and VGluT2 + CA1→SUB projections. (E) Left, schematic of proximity biotinylation assay. Right, injection sites for preBirA∗ in DH and corresponding GFP signal in DH and RSP. (F) Right, cytoBirA ∗ was injected into DH of naive VGluT1-Cre and VGluT2-Cre mice, biotinylated proteins were pulled down, and quantified from RSP. Left and up, KEGG terms of differentially expressed proteins in VGluT1 + or VGluT2 + biotinylated terminals. Bottom and left, after injection of the virus expressing preBirA ∗ , the levels of biotinylated proteins from VGluT1 RSC were dissimilar, as indicated by lack of correlation, suggesting differences in their presynaptic proteomes. (G) Main cell populations identified using known neuronal and non-neuronal cell markers. (H) VGluT1 and VGluT2 expression across cell types reveals largely non-overlapping populations. All scale bars, 250 μm. To examine the proteins expressed in the SUB→RSP terminals, we used in vivo proximity biotinylation in combination with tandem mass spectrometry (MS)-based proteomic analysis 30 , 31 ( Figure S3 A; see STAR Methods ). Similar to the tracing approaches, EGFP co-expressed by preBirA ∗ AAV constructs localized in RSP layer 3 ( Figure 2 E). To investigate the protein diversity of VGluT1 and VGluT2 presynaptic terminals, we employed tandem mass tag (TMT)-based quantitative MS to directly compare their proteomes. Specifically, VGlut1-Cre and VGlut2-Cre mice were injected with either AAV-FLEx-preBirA ∗ or AAV-FLEx-cytoBirA ∗ . Proteins exhibiting a preBirA ∗ to cytoBirA ∗ ratio grater then 1.5 were considered enriched and thus classified as presynaptic terminal proteins. We mined the VGluT1 + and VGluT2 + SUB→RSP terminal proteomic datasets using the online Database for Annotation, Visualization, and Integrated Discovery (DAVID). As indicated by Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, the proteomes of VGluT1 + and VGluT2 + RSP terminals differed from one another at baseline conditions ( Figure S3 B). VGluT2 + terminal proteome showed significantly enriched KEGG terms involved in neurotransmitter release and synaptic plasticity (endocytosis, SNARE interactions in vesicular transport, ribosome, long-term potentiation, and axon guidance), and various neuromodulatory signaling and co-transmitter phenotype (endocannabinoid, GABAergic, dopaminergic, and cholinergic synapse), 25 activity-dependent modulation and involvement in learning and memory processes ( Figure S3 B). Overall, differentially expressed proteins in VGluT2 biotinylated terminals were glycoproteins, membrane, receptor, and transmembrane proteins involved in signaling (preBirA ∗ /cytoBirA ∗ >1.5-fold), while the greatest changes (preBirA ∗ /cytoBirA ∗ >1.5-fold) in VGluT1 terminals were in proteins involved with exosomes, acetylation, cytoskeletal dynamics, chaperone function, and protein folding ( Figure 2 F). We correlated the levels of biotinylated proteins pulled down from VGluT1 + and VGluT2 + terminals to control for specificity of the proteomic differences. Samples from mice injected with preBirA ∗ ( Figure 2 F, left and down) were highly divergent ( R 2 = 0.03), reflecting substantial differences in their presynaptic terminal proteomes. In contrast, in mice injected with the control virus cytoBirA ∗ ( Figure S3 C), protein abundances showed a strong positive correlation ( R 2 = 0.82), indicating similar protein profiles due to non-specific biotinylation of cytoplasmatic proteins. Together these data shows that VGluT1 + and VGluT2 + SUB→RSP have distinct presynaptic proteomic signatures, indicating a specialized role in synaptic function. VGlut2 + proteome suggest a more modulatory and dynamic role, while the VGluT1 + terminals show profile associated with structural support and protein synthesis. To further explore VGluT1 + and VGluT2 + neuronal and non-neuronal hippocampal populations, we used our previously published dataset for which, we dissected the DH and used 10× genomics platform to perform single-nucleus RNA sequencing (snRNA-seq). 32 An unsupervised algorithm run on the snRNA-seq dataset identified 30 clusters ( Figure S4 A). Using canonical markers and existing hippocampal databases (see STAR Methods ), we identified CA1 neurons (3 clusters), SUB neurons (1 cluster), CA3 neurons (2 clusters), dentate gyrus granule cells (DGGC) (4 clusters), interneurons (2 clusters), and different non-neuronal cells (8 clusters; Figures 2 G and S4 A). snRNA-seq data showed that VGluT1 and VGluT2 were present in non-overlapping cellular populations ( Figures 2 H and S4 B), which concurs with our tracing experiments. A small subset of cells in the CA1.1 and SUB clusters was both VGluT1 and VGluT2 positive, a population also seen in DH of mice injected with Cre-dependent color flipping nuclear reporter ( Figure 2 C, right, indicated with yellow arrow). Overall, we demonstrate that VGluT1 phenotype is most prevalent in the majority of hippocampus, with exception of CA1 and SUB, which contain mostly non-overlapping VGluT1 + and VGluT2 + populations. Both CA1 and SUB outputs are VGluT1 + and VGluT2 + ; however, in the CA1 region VGluT1 + and VGluT2 + populations are organized into deep and superficial layers, while the SUB shows a more heterogeneous distribution of VGluT1 + and VGluT2 + populations ( Figure 2 D). VGluT2 + SUB→RSP afferents have distinct response dynamics dependent on TFC phase To assess activity of VGluT2 + SUB→RSP afferents during TFC, we used fiber photometry to measure the populations’ activity through changes in Ca 2+ activity. We injected mice with Cre-dependent calcium indicator AAV-DIO-GCaMP7s in DH of VGluT2-Cre mice, implanted with a fiber targeting the SUB→RSP layer 3 projections and recorded the fluorescence signal changes—as an indicator of bulk calcium activity—across 3 trials of training and testing days ( Figure 3 A). Figure 3. Open in a new tab VGluT2 + SUB→RSP afferents have distinct response dynamics dependent on TFC phase (A) Top, representative image of viral expression and fiber track, together with the diagram indicating position of injection and fiber placement. Bottom, experimental design for fiber photometry recordings. Scale bars, 500 μm (main image) and 100 μm (inset). (B and C) Average GCaMP fluorescence traces during 3 CS-US presentations on training day (B) and during 3 CS presentations on test day (C) recorded from RSP VGluT2 + terminals of projections originating in SUB. Arrows indicate significant increases and decreases of signal. (D) Increase in average Z score following shock exposure ( n = 9; two-way ANOVA with repeated measures; factor: trial, p = 0.2195, F (2,32) = 1.591, factor: phase, p < 0.0001, F (1, 16) = 97.98, factor: trial × phase, p = 0.1259, F (2, 32) = 2.212). (E) Average Z score at tone onset and during trace on training (left) and testing (right) day. Training tone vs. trace ( n = 9; two-way ANOVA with repeated measures; factor: trial, p = 0.0689, F (1.927, 30.84) = 2.951, factor: phase, p = 0.0021, F (1, 16) = 13.40, factor: trial × phase, p = 0.0258, F (2, 32) = 4.107). Cue test tone vs. trace ( n = 9; two-way ANOVA with repeated measures; factor: trial, p = 0.0355, F (1.910, 30.56) = 3.795, factor: phase, p = 0.0029, F (1, 16) = 12.37, factor: trial × phase, p = 0.0165, F (2, 32) = 4.679). Data presented as mean ± s.e.m. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001, ∗∗∗∗ p < 0.0001; NS, not significant. In chemogenetic inhibition, we used single-trial training and testing to check for acute effects of VGluT2 + SUB→RSP projection silencing on TFC, whereas in fiber photometry experiments, we used multi-trial approach to capture how neural dynamic evolves across trials. Consistent with previous studies, we found that every presentation of a foot-shock during training elicited a large, sustained increase in the GCaMP fluorescence ( Figures 3 B–3D and S6 C). 9 We compared average Z score for signals obtained from the RSP terminals originating in the SUB across different phases of training and found that fluorescent signal increased during first 15 s of the tone followed by a dip during trace with repeated tone-shock pairing ( Figure 3 E). Area under the curve (AUC) showed a similar pattern of signal dynamics, where onset of tone caused a significant increase in the GCaMP signal at the second and third presentation of tone, that was followed by the negative-going responses (“dips”) during trace. Dip in GCaMP signal during trace was significantly amplified at third tone presentation ( Figure S6 A). During the cue test, average Z scores showed differences in bulk calci micrometer activity during the first 15 s of the tone and 15 s of trace at first and last trial ( Figures 3 C and 3E). On the testing day, a significant increase of the GCaMP signal at the tone onset was elicited at the first presentation of the tone, followed by a moderate negative-going response increase ( Figures 3 C and 3E). While the signal increase at the tone onset was only present at the first tone presentation, dip during trace period re-emerged at third tone presentation ( Figures 3 C and 3E). AUC comparison showed similar signal dynamics ( Figure S6 B). The negative correlation between GCaMP signal and freezing indicates that fluctuations in VGluT2 + SUB→RSP projection activity co-varied with freezing behavior across tone and trace periods during both training and testing ( Figures S6 D and S6F). Animals associated the tone and the trace as indicated by significantly different levels of freezing across test phases ( Figure S6 E). Together these results show that VGluT2 + SUB→RSP projection response dynamic is characterized by overshoot at tone onset followed by the signal dip during trace formed during memory acquisition and replayed at first memory recall. Changes in response dynamics of VGluT2 + SUB→RSP afferents are specific to learning of tone-shock association To exclude the possibility that the observed response dynamics is a non-associative effect of tone-induced arousal rather than consequence of associative learning, we recorded GCaMP signals from VGluT2 + SUB→RSP afferents in mice undergoing pseudoconditioning ( Figures 4 A and 4B). TFC and pseudoconditioning differ by a key feature, timing between shock and tone; in pseudoconditioning shock and tone are presented temporally unpaired, preventing the formation of a predictive relationship between two stimuli. 33 Thus, if the response dynamic is a consequence of the learned tone-trace-shock association, it will not appear at tone presentation during pseudo fear conditioning. At training, all three shock presentations elicited a response, indicating that shock response during training is not specific for tone-shock association ( Figure 4 C). Tone presentation both during training and testing did not cause changes in GCaMP fluorescent signal suggesting that the pattern of activity observed during TFC is a consequence of tone-trace-shock association ( Figures 4 D and 4E). Animals did not show significantly different levels of freezing across test phases, or correlation between GCaMP signal and freezing, indicating that they did not form an association between the tone and the shock ( Figures S7 A and S7B). Taken together, these data suggest that differences in activity dynamic of VGluT2 + SUB→RSP afferents encode association of the tone and shock that are separated by a time gap. Figure 4. Open in a new tab Changes in response dynamics of VGluT2 + SUB→RSP afferents are specific to learning of tone-shock association (A) Experimental diagram detailing GCaMP7s AAV injection in the DH and fiber placement in the RSP, enabling recording of calcium signals from SUB→RSP terminals with representative image of viral expression and fiber track. Scale bars, 500 μm (main image) and 200 μm (inset). (B) Schematic of the unpaired fear conditioning paradigm. (C) Top, averaged GCaMP fluorescence traces during 3 unpaired shock presentations on training day. Bottom, significant differences were found in average Z score between pre shock and post shock period ( n = 11; two-way ANOVA with repeated measures; factor: trial, p = 0.0199, F (2, 40) = 4.324, factor: phase, p < 0.0001, F (1, 20) = 48.23, factor: trial × phase, p = 0.1750, F (2, 40) = 1.8210). (D) Top, averaged GCaMP fluorescence traces during 3 unpaired tone presentations on training day. Bottom, no significant differences were found in average Z score between first half of the tone and trace ( n = 11; two-way ANOVA with repeated measures; factor: trial, p = 0.0988, F (2, 40) = 2.454, factor: phase, p = 0.0068, F (1, 20) = 9.090, factor: trial × phase, p = 0.9558, F (2, 40) = 0.0453). (E) Top, averaged GCaMP fluorescence traces during 3 unpaired tone presentations during the cue test. Bottom, no significant differences were found in average Z score between first half of the tone and trace ( n = 11; two-way ANOVA with repeated measures; factor: trial, p = 0.1915, F (2, 40) = 1.723, factor: phase, p = 0.0018, F (1, 20) = 12.92, factor: trial × phase, p = 0.9272, F (2, 40) = 0.0758). Data presented as mean ± s.e.m. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001, ∗∗∗∗ p < 0.0001; NS, not significant. Discussion We demonstrated that VGluT2 + SUB→RSP projections significantly contribute to the associative and temporal components of temporal associative memories, presented to the RSP as an integrated pattern of activity acquired during training and reinstated at retrieval. This was indicated by the observations that concurrent inhibition of VGluT2 + and VGlutT1 + SUB→RSP projections impaired tone-shock associations whereas inhibition of the VGluT2 + SUB→RSP projections selectively impaired freezing during the trace interval. During TFC, VGluT2 + SUB→RSP projections acquired pattern of bulk calcium activity involving a transient increase of the calcium signal at tone onset, followed by suppression during the trace interval. Such activity patterns were not seen in the CA1 population activity, as reported previously, 9 suggesting that integration of associative and temporal components of TFC occurred at the SUB or even at the level of the synaptic terminals. We previously demonstrated that the formation of associative contextual memories predominantly depends on the VGluT1 + SUB→RSP projections while VGluT2 + projections contributed to memory persistence. 19 Here, we found a similar functional divergence but with respect to the encoding of the temporal trace, which solely depended on the activity of VGluT2 + SUB→RSP projections. This was indicated by the effects of chemogenetic inhibition of these projections, which specifically impaired freezing during the trace period, but not during the tone, or sequences of tone-light-shock pairings lacking a temporal gap. Silencing VGluT2 + SUB→RSP projections significantly impaired, but did not abolish, freezing during trace period. This residual effect may stem from incomplete silencing of SUB→RSP projections, given the fact that they extend broadly along anterior-posterior axis, or alternatively may reflect the capacity of parallel circuits to compensate for the loss of this pathway. Early postnatal excitatory circuits rely on VGluT2 activity, but, as hippocampal, cortical, and cerebellar circuits mature, neurons switch from VGluT2 to VGluT1 phenotype. 34 The CA1, the SUB, and a couple of other brain areas are unique in that their neurons express both vesicular transporters. While several findings suggested that VGluT1 and VGluT2 confer different functional properties to excitatory projections, such as different synaptic vesicle dynamics and glutamate release probability, 35 , 36 it has remained unclear whether they stem from discrete or overlapping neuronal populations. Here we showed, using both viral and snRNA-seq approaches, that only a small subset of CA1 and SUB neurons contains both transporters, whereas most are positive for either VGluT1 or VGluT2. In the CA1, we show that VGluT2 + neurons were predominantly located in the developmentally older, 37 deep layer, suggesting that these neurons, rather than switching to VGluT1, retained their early developmental VGluT2 phenotype. The segregation of VGluT1 and VGluT2 across deep and superficial layers could contribute to their robust differences in firing rates and burst frequency, 38 as well as in their unique contributions to spatial and temporal associative memories. The VGluT neuronal phenotypes in the distal SUB, the primary source of RSP projections, appeared less segregated. Based on our analysis of publicly available databases showing co-expression of VGluT2 and Fn1, 27 it is likely that our manipulations primarily targeted VGluT2 + Fn1 + neurons of the distal SUB. In addition to their different cellular origins, excitatory SUB→RSP projections also showed significant differences in their synaptic proteomes, with VGluT2 + presynaptic terminals showing enrichment in neuromodulatory and SNARE-mediated vesicular transport proteomes relative to VGluT1 + synapses. The overall greater proteomic variability of VGluT2 + terminals may be particularly adept at modulating cortical networks in response to dynamic sensory inputs (e.g., TFC versus contextual fear conditioning), thereby fine-tuning cortical circuits for specific behavioral outcomes. It is currently believed that temporal associations are encoded through three interacting processes: trace holding (recent sensory input), temporal expectation (prediction of forthcoming stimuli), and explicit timing (estimation of stimulus onset relative to prior cues). 39 According to prevailing models, sustained neural representations of the conditioned stimulus (tone in the case of TFC) across the trace interval are required for the association of temporally discontiguous events. 40 Single cell analyses of the activity of DH and SUB neurons have not provided conclusive evidence on learning-related activity patterns coding for the different stimuli and the temporal gap between them. Deep-brain calcium imaging identified a subgroup of neurons, both in the CA1 and the SUB, that maintained tone-evoked activity during the trace period at memory acquisition and consolidated this plasticity after learning in CA1 (but not SUB); becoming increasingly active in anticipation of shock, during the tone-trace interval, an activity correlated with later fear retrieval. Based on these findings, it was suggested that the SUB temporarily supports stimulus maintenance, while CA1 stably encodes this information for long-term memory formation. 13 Importantly, the study used AAV (serotype 9) driven by the CaMKIIα promoter for expression of the calcium sensor, which is known to target both excitatory pyramidal neurons and subsets of GABAergic interneurons, 41 suggesting that some of the calcium signals could potentially reflect activity of local SUB interneurons. Other studies showed more nuanced responses of CA1 neurons. Two-photon imaging has identified a sparse group of cue-activated CA1 neurons displaying stochastic dynamics across trials, suggesting that trace is not bridged by persistent activity but rather widespread shifts in neuronal ensemble activity, 42 whereas single unit recordings showed increased firing only at the first presentation of the tone. 43 Our analysis of bulk synaptic activity, revealed by calcium responses in VGluT2 + SUB→RSP synapses, revealed a different activity pattern: an increase in fluorescent signal at tone onset, followed by a pronounced signal dip during the trace interval. These responses were acquired during training, reconstructed during the first memory test, and not found after pseudo conditioning, demonstrating their specificity for temporal associative learning as well as temporal coding. Current views assume that the association between two events separated in time is simply strengthened by sustained excitatory firing during trace, with inhibition serving merely as a dampening force. 39 , 40 Our findings indicate that, rather than simply reflecting a drop in excitatory drive, this dip may instead mark a temporally structured inhibition that regulates how and when subicular output modulates RSP activity required for temporal coding. 17 Feedforward inhibition of SUB VGluT2 + neurons is a likely mechanism for the decrease of their activity during the temporal interval within which associative learning can take place; however, it is unclear at this time whether such inhibition is provided by local microcircuits or from long-range inhibitory projections marked by neuropeptide Y (NPY), somatostatin (SOM), and muscarinic acetylcholine receptor type 2 (M2R). 44 These pathways could temporally inhibit SUB output. The transient reduction in VGluT2 + SUB→RSP activity during the trace interval likely represents one component of a temporal coding process that links the conditioned cue to the temporally discontiguous aversive outcome. Our data do not determine whether the transient dip simply marks trace onset or contributes to representing its duration, leaving unsolved whether VGluT2 + SUB→RSP projections are involved in encoding the timing or persistence of conditioned response. Findings from trace eyeblink conditioning show that hippocampal lesions abolish adaptive timing, producing conditioned eyelid closures that peak during cue rather than at expected aversive stimulus. 45 , 46 Similarly, the classic work by Gabriel et al. demonstrated that hippocampal output modulates RSP and anterior thalamic activity during learning, and that SUB lesions disrupt the triphasic CS-evoked responses and disrupt conditioned response timing. 47 Although TFC is not ideally suited for precise analyses of the timing response, our results extend this framework by identifying the VGluT2 + SUB projections as a likely pathway through which the hippocampal output supports temporal precision in associative learning. The reproducible dip we observed during both acquisition and recall suggest a stable, experience-dependent motif that re-emerges as contextual representations consolidate. These results support a model in which VGluT2 + SUB→RSP pathway coordinates hippocampal-cortical dynamics to maintain temporal expectancy and sustain defensive behavior across the trace interval. Rather than merely marking the onset of the trace, this suppression may reflect a transition in network state that organizes RSP activity throughout the trace interval and interacts with sustained hippocampal or entorhinal signals encoding for elapsed time. Trial-by-trial changes in pattern could reflect rapid recalibration of temporal coding as the animal updates its memory representations upon cue re-exposure. The consistent, time-locked inhibition we observe at VGluT2 + terminals may provide an internal reference point for maintaining expectancy once the sensory input ceases. Building on these observations, the Ca 2+ signal showed a brief increase at tone onset during the first test trial. This likely reflects a learned, sensory-evoked response to the abrupt appearance of the conditioned stimulus. This onset peak was followed by re-emergence of the dip during trace interval. The negative correlations between GCaMP signal and freezing observed in both the tone and the trace indicate that lower activity of VGluT2 + SUB→RSP terminals is associated with higher freezing levels. This suggests that suppression of this pathway accompanies the expression of conditioned fear, regardless of the trial phase. Rather than reflecting a simple sensory-evoked excitation, this biphasic patter appears to encode state transitions. The decrease in activity marks the shift from cue processing to sustained defensive immobility. This inverse relationship between activity and freezing supports the reduced hippocampal-RSP drive and helps stabilize the freezing state during the stimulus-free trace interval. Interestingly, recordings of bulk calcium signals from the CA1 and SUB, apart from shock responses, did not show cue- or trace-specific patterns during TFC 9 as found in the VGluT2 + SUB→RSP projections. There are several possible explanations for the differences in the activity patterns detected in SUB inputs and outputs. It is possible that synaptic population activity reveals TFC-specific patterns better than neuronal population activity. The predictive activity pattern at VGluT2 + SUB→RSP terminals implied the presence of circuit-level computation within the SUB. This observation aligns with prior findings that designate the SUB as a critical hub for transforming and routing hippocampal output. 48 Indeed, large-scale optogenetically targeted electrophysiological recordings show that dorsal SUB neurons conveyed comparable or greater information per unit of time compared to CA1 neurons, and depending on information type, SUB conveyed information out either uniformly or selectively to specific projection targets, with their firing precisely controlled by theta oscillations and SWR. 48 Such mechanism could enable flexible modulation of downstream cortical networks in a manner consistent with dynamic memory demands. Alternatively, TFC-specific activity patterns in VGluT2 + SUB→RSP terminals may not be directly correlated with SUB neuronal activity, 13 but instead generated through plasticity at discrete SUB→RSP synapses. There is an emerging support for the view that separate mechanisms contribute to the processing of different components of human episodic memories, with different developmental trajectories for memories of facts, space, and time. 49 , 50 The functional differentiation of circuits identified in our work might thus be applicable to human memory processes from early development throughout adulthood. Moreover, our findings suggest that temporal coding in TFC may emerge not from excitatory persistence but from a circuit-level interplay of excitation and well-timed inhibition, coordinated along the hippocampus–SUB–RSP axis. Future studies are required to delineate the sources of inhibitory inputs to the SUB and to define how its output signals are integrated within downstream target areas. Limitations of the study The current study recorded bulk calcium signal activity because our main interest was in the synaptic activity patterns conveyed from SUB to RSC. Such approach could mask the detection of activity of individual neurons that projects to RSC, which may exhibit finer, more specific patterns of activity, overlooking transient or subtle dynamics that could provide additional insight into the underlying processes of temporal memory consolidation. Resource availability Lead contact Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Ana Cicvaric ( [email protected] ). Materials availability This study did not generate new unique reagents. Data and code availability • All results of synaptic proteomes analysis have been deposited through an interactive portal ( https://proteome-phenome-atlas.com/ ). The mass spectrometry (MS) data presented in this study have been deposited in the Mass Spectrometry Interactive Virtual Environment (MassIVE) under identifier MSV000098056 ( https://massive.ucsd.edu/ProteoSAFe/dataset.jsp?task=aa8f8aa9910b400b8382992f39700ba1 ) and ProteomeXchange under identifier PXD064462 ( https://proteomecentral.proteomexchange.org/cgi/GetDataset?ID=PXD064462 ). • To generate snRNA-seq graphs, this paper analyzes existing, publicly available data, accessible at NCBI Gene Expression Omnibus database under accession GSE254780 and source code available at https://github.com/RadulovicLab/Nature-2024 . • All other data reported in this paper and any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. All other software and analytical methods used in this study are publicly available, as listed in the key resources table . Acknowledgments This work was funded by NIMH grants MH108837 and MH078064 to J.R.; J 4271 FWF (Austrian Science Fund) to A.C.; NARSAD Young Investigator Grant to Y.-Z.W.; Individual Biomedical Research Award from The Hartwell Foundation and NIH S10OD032464 to J.N.S. We thank Genetics and Computational Genomics Cores at Albert Einstein College of Medicine, especially Junya Zhang, for their help in data analysis. Illustrations of nonscientific data were created with BioRender.com through an institutional license. Author contributions J.R. and A.C. designed the study; A.C., T.E.B., and J.R. wrote the original draft; Y.-Z.W., V.J., N.K., and J.N.S. performed proteomic experiments; E.M.W. and H.Z. performed snRNA-seq experiments; A.C. and T.E.B. performed fiber photometry experiments; N.Y. and A.C. performed tracing experiments; L.R., V.G., J.R., V.P., Z.P., and A.C. performed behavior tests and data analysis; Y.-Z.W., J.N.S., H.Z., E.M.W. and Z.P. assisted in manuscript revision. Declaration of interests The authors declare no competing interests. STAR★Methods Key resources table REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies Rabbit anti-mCherry Abcam Cat#ab167453; RRID: AB_2571870 Chicken anti-GFP Abcam Cat#ab13970; RRID: AB_300798 Biological samples Mouse brain samples This paper N/A Chemicals , peptides, and recombinant proteins CLOZAPINE N-OXIDE (CNO) Sigma Aldrich Cat#C0832 2-Methylbutane TCI chemicals Cat#M0167 Paraformaldehyde (PFA) Thermo Scientific Chemicals Cat#A11313 Biotin Sigma Cat#B4501 NeutrAvidin beads Thermo Fisher Scientific Cat #2901 TMT reagent Thermo Fisher Scientific Cat# 90111 Recombinant DNA ssAAV-9/2-hSyn1-chI-dlox-jGCaMP7s(rev)-dlox-WPRE-SV40p(A) VVF Zurich Cat#v407-9 AAV8-EF1a-Nuc-flox(mCherry)-EGFP Addgene Cat#112677-AAV8; RRID:Addgene_112677 AAVrg-EF1a-DO_DIO-TdTomato_EGFP-WPRE-pA Addgene Cat#37120-AAVrg; RRID:Addgene_37120 AAV8-hSyn-HA-hM4D(Gi)-mCherry Addgene Cat#44362-AAV8; RRID:Addgene_44362 AAV8-hSyn-DIO-hM4D(Gi)-mCherry Addgene Cat#50475-AAV8; RRID:Addgene_50475 AAVrh10-FLEx-preBirA∗ Savas lab/Packaged by ViroVek N/A AAV-FLEx-cytoBirA∗ Savas lab/Packaged by ViroVek N/A Experimental models: Organisms/strains Wild-type mice Envigo C57BL/6NHsd; IMSR_ENV:HSD-044 Vglut1-Cre mice Jackson Laboratory (Harris JA et al., 2014) Cat#037512; RRID:IMSR_JAX:037512 Vglut2-Cre mice Jackson Laboratory (Vong L et al., 2011) Cat#016963; RRID:IMSR_JAX:016963 Deposited data In vivo proximity biotinylation synaptic proteomes analysis This paper MSV000098056; PXD064462 Software and algorithms Prism 10.4.0 Graphpad https://www.graphpad.com MATLAB 2022a MathWorks https://www.mathworks.com R The R Project https://www.r-project.org Guppy Lehner Lab https://github.com/LernerLab/GuPPy Spyder MIT https://www.spyder-ide.org/ RStudio Posit PBC https://posit.co/products/open-source/rstudio/ DAVID NIH https://david.ncifcrf.gov IP2 Bruker, Eng et al., 1994 and Xu et al., 2014 http://www.integratedproteomics.com Video Freeze™ MedAssociates https://med-associates.com/product/videofreeze-video-fear-conditioning-software/ Open in a new tab Experimental and study participant model details We used male and female C57BL/6N mice, vGlut1-Cre 23 mice and vGlut2-Cre 24 mice, as described in detail recently. 19 Wild type C57BL6/J mice were purchased from Harlan, Indianapolis, IN. All Cre mouse lines were obtained from the Jackson Laboratory (Bar Harbor, ME). The VGluT1-Cre mouse line, also known as Slc17a7-IRES2-Cre or Vglut1-IRES2-Cre-D, was created by the Hongkui Zeng lab, Allen Institute for Brain Science, 23 the VGluT2-Cre knockin mice, also known as Slc17a6tm2(cre) and Lowl or VGlut2-ires-Cre, was generated as described previously. 24 Mice were bred, genotyped (using primers reported on the Jackson Laboratory website) and used for experiments at the age of 8 weeks. Typically, we obtained 4-6 litters/breeding cycle with 5-8 mice/litter with similar distribution of males and females in order to assess any sex specific differences. All litters were used for behavioral experiments and randomly allocated male mice were used for tracing, fiber photometry and proteomic studies. The mice were maintained under standard housing conditions (12 h/12 h light dark cycle with lights on at 7 a.m., temperature 20-22°C, humidity 30-60%) in our satellite behavioral facility. All animal procedures used in this study were approved by the Northwestern University’s Animal Care and Use Committee (protocols IS00002463 and IS00003359) and Albert Einstein’s Animal Care and Use Committee (protocols 00001289 and 00001268) in compliance with US National Institutes of Health standards. Method details Stereotaxic surgeries and infusions of viral vectors and drugs Mice were anesthetized with 1.2% tribromoethanol (vol/vol, Avertin) for viral vector intracranial infusion and cannula implantation. The Cre-dependent color-flipping retrograde reporter pAAV-Ef1a-DO_DIO-TdTomato_EGFP-WPRE-pA 51 (gift from Bernardo Sabatini, Addgene plasmid Cat# 37120) was injected unilaterally in RSP or SUB (RSP: 1.80 mm posterior, ±0.40 mm lateral, 1.00 mm ventral to bregma and SUB: 3.52 mm posterior, ±2.50 mm lateral, 1.85 mm ventral to bregma) and Cre-dependent color-flipping nuclear reporter AAV8-EF1a-Nuc-flox(mCherry)-EGFP 52 (gift from Brandon Harvey, Addgene, Cat # 112677) was injected in CA1 (1.80 mm posterior, ±1.00 mm lateral, 2.20 mm ventral to bregma). The viral vector carrying a construct coding for the Cre-independent inhibitory DREADD 53 (AAV8-hSyn-HA-hM4D(Gi)-mCherry, gift from Bryan Roth, Addgene, Cat # 50475) or Cre-dependent inhibitory DREADD 53 (AAV8-hSyn-DIO-hM4D(Gi)-mCherry, gift from Bryan Roth, Addgene, Cat # 44362) was bilaterally infused into the DH (1.80 mm posterior, ±1.00 mm lateral, 2.25 mm ventral to bregma). For Fiberphotometry experiments viral vector carrying a construct coding for the Cre-dependent GCaMP was injected to DH, (ssAAV-9/2-hSyn1-chI-dlox-jGCaMP7s(rev)-dlox-WPRE-SV40p(A), VVF Zürich, Cat # v407-9, DH: 1.80 mm posterior, ±1.00 mm lateral, 2.13 mm ventral to bregma). For behavioral experiments infusions were performed using an automatic microsyringe pump controller (Micro4-WPI) connected to a Hamilton microsyringe (Cat # 88400). The viral vectors were infused in a volume of 0.5 μL per site over 2 min, and syringes were left in place for 5 min prior to removal to allow for virus diffusion. Bilateral 26 gauge guide cannulas (Plastics One) were placed in RSP (1.8 mm posterior, ±0.4 mm lateral, 0.75 mm ventral to bregma). Mice were allowed 6 weeks for virus expression prior to behavioral testing. CNO (Sigma; 0.3 μg/mL; 0.20 μL per side, at a rate of 0.5 μL/min) was infused through the cannulas 30 min prior to either fear conditioning or memory retrieval testing. After the completion of behavioral testing, all brains were collected and cannula placements and virus spread were confirmed by immunohistochemical analysis using anti-mCherry antibodies (1:1000; Abcam, Cat # ab167453). For tracing and fiberphotometry experiments we used Microliter Neuros Syringe (Cat # 65460-02) with an automatic microsyringe pump controller (Micro4-WPI) to deliver unilaterally 0.2 μL of viral vectors over 2 min. Syringes were left in place for 5 min prior to removal to allow for virus diffusion. Optic fibers (MBF Bioscience, Cat # FOC-BF-200-125 with 200 um core diameter and NA=0.37) were implanted to record from SUB→RSP projections (1.80 mm posterior, ±0.30 mm lateral, 1.00 mm ventral to bregma). After the completion of experiments, all brains were collected and fiber placements and virus spread were confirmed by immunohistochemical analysis using anti-GFP antibodies (1:2000; Abcam, Cat # 13970). Immunohistochemistry and immunofluorescence Mice were anesthetized with an i.p. injection of 240 mg/kg Avertin and transcardially perfused with ice-cold 4% paraformaldehyde in phosphate buffer (pH 7.4, 150 mL per mouse). Brains were removed and post-fixed for 48 h in the same fixative and then immersed for 24 h each in 10%, 20% and 30% sucrose solution in phosphate buffer. Brains were frozen and 50 μm sections were cut for use in free-floating immunohistochemistry, as described previously. 54 Primary antibodies against mCherry (1:1000; Abcam AB167453 ) and GFP (1:2000, Abcam AB13970) were used and visualized with either diaminobenzidine (Sigma) or secondary antibodies obtained from Jackson ImmunoResearch (1:500 each, AlexaFluor® 594 Cat # 711-585-152, AlexaFluor® 488 Cat # 703-545-155). Sections were mounted using Vectashield (Vector) and observed with a confocal laser-scanning microscope (Olympus Fluoview FV10i) at 40×. Bright-field microscopy was used to visualize signals immunolabeled with choromogenic substrate diaminobenzidine (DAB, Sigma). Quantification of VGluT1 + and VGluT2 + presynaptic terminal proteomes by in vivo proximity biotinylation and TMT-MS First, we constructed a molecular probe by fusing the promiscuous biotin ligase BirA∗ to a presynaptic targeting motif (FLEx-preBirA∗). 30 , 31 , 55 We also included a T2a motif followed by GFP to visualize neurons expressing our probe and flanking FLEx elements to conditionally express the probe in a cell specific manner based on Cre expression. Once the probe is expressed in the brain, it will localize to presynaptic terminals and biotinylate nearby proteins with a radius of 10 nm. As a negative control, we removed the presynaptic targeting sequence from to generate a construct (FLEx-cytoBirA∗) that is not selectively targeted to any subcellular location. Finally, we packaged FLEx-preBirA∗ and FLEx-cytoBirA∗ into AAV viruses (packaged by ViroVek). We then sterotactically injected FLEx-preBirA∗ or FLEx-cytoBirA∗ AAVs into DH of VGluT1-Cre and VGluT2-Cre mice as described above and incubated the viruses for one month. Next, we administrated biotin (22.5 mg/kg subcutaneously, Sigma, Cat # B4501) to the mice daily for one week to induce biotinylation of presynaptic proteins. Then we sacrificed the mice and dissected the RSP regions. Dissected RSPs were homogenized in RIPA lysis buffer (50 mM Tris, 150 mM NaCl, 0.1% SDS, 1mM EDTA, 0.5% sodium deoxycholate, 1% Triton X-100, 1 x protease inhibitor cocktail (Thermo Fisher Scientific, Cat # 78443), 1 x phosphatase inhibitor (Thermo Fisher Scientific, Cat # 78420), pH 7.4) with an electronic homogenizer (Glas-Col, Cat # 099C-K54). Then excess 10% SDS solution was added into each sample to make the final SDS concentration to 1%. After sonication with a probe sonicator (Qsonica) for 3 x 1 min, RSP homogenates were solubilized at 4°C for 1 h with rotation. Insoluble components were removed by centrifuging at 13,000 x g for 30 mins. 400 μL of pre-washed NeutrAvidin beads (Thermo Fisher Scientific, Cat # 2901) were added into each sample and incubated at 4°C overnight with gentle rotation. We performed on-beads digestion based on previous reported protocol. 56 After overnight incubation with RSP homogenates, NeutrAvidin beads were rinsed for five times in one mL lysis buffer (6 M Guanidine, 50 mM HEPES, pH 8.5), then added one mL lysis buffer. Dithiothreitol (DTT, DOT Scientific Inc, Cat# DSD11000) was applied to a final concentration of 5 mM. After incubation at RT for 20 min, iodoacetamide (IAA, Sigma-Aldrich, Cat# I1149) was added to a final concentration of 15 mM and incubated for 20 min at room temperature in the dark. Excess IAA was quenched with DTT for 15 min. Samples were diluted with buffer (100 mM HEPES, pH 8.5, 1.5 M Guanidine), and digested for 3 h with Lys-C protease (1:100, ThermoFisher Scientific, Cat# 90307_3668048707) at 37°C. Trypsin (1:100, Promega, Cat# V5280) was then added for overnight incubation at 37°C with intensive agitation (1000 rpm). The next day, reaction was quenched by adding 1% trifluoroacetic acid (TFA, Fisher Scientific, O4902-100). The samples were desalted using HyperSep C18 Cartridges (Thermo Fisher Scientific, Cat# 60108-301) and vacuum centrifuged to dry. C18 column-desalted peptides were resuspended with 100 mM HEPES pH 8.5 and the concentrations were measured by micro BCA kit (Fisher Scientific, Cat# PI23235). For each sample, 25 μg of peptide labeled with TMT reagent (0.4 mg, dissolved in 40 μL anhydrous acetonitrile, Thermo Fisher Scientific, Cat# 90111) and made at a final concentration of 30% (v/v) acetonitrile (ACN). Following incubation at room temperature for 2 h with agitation, hydroxylamine (to a final concentration of 0.3% (v/v)) was added to quench the reaction for 15 min. Equal amounts of TMT-tagged samples were mixed. Combined sample was vacuum centrifuged to dryness, resuspended, and subjected to HyperSep C18 Cartridges. We used a high pH reverse-phase peptide fractionation kit (Thermo Fisher Scientific, Cat# 84868) to get eight fractions (5.0%, 10.0%, 12.5%, 15.0%, 17.5%, 20.0%, 22.5%, 25.0% and 50% of ACN in 0.1% triethylamine solution). The high pH peptide fractions were directly loaded into the autosampler for MS analysis without further desalting. 3 μg of each fraction or sample were auto-sampler loaded with a Thermo UltiMate 3000 HPLC pump onto a vented Acclaim Pepmap 100, 75 μm x 2 cm, nanoViper trap column coupled to a nanoViper analytical column (Thermo Fisher Scientific, Cat#: 164570, 3 μm, 100 Å, C18, 0.075 mm, 500 mm) with stainless steel emitter tip assembled on the Nanospray Flex Ion Source with a spray voltage of 2000 V. An Orbitrap Fusion (Thermo Fisher Scientific) was used to acquire all the MS spectral data. Buffer A contained 94.785% H 2 O with 5% ACN and 0.125% FA, and buffer B contained 99.875% ACN with 0.125% FA. The chromatographic run was for 4 h in total with the following profile: 0-7% for 7, 10% for 6, 25% for 160, 33% for 40, 50% for 7, 95% for 5 and again 95% for 15 mins receptively. We used a multiNotch MS3-based TMT method to analyze all the TMT samples. 57 , 58 , 59 The scan sequence began with an MS1 spectrum (Orbitrap analysis, resolution 120,000, 400-1400 Th, AGC target 2×10 5 , maximum injection time 200 ms). MS2 analysis, ‘Top speed’ (2 s), Collision-induced dissociation (CID, quadrupole ion trap analysis, AGC 4×10 3 , NCE 35, maximum injection time 150 ms). MS3 analysis, top ten precursors, fragmented by HCD prior to Orbitrap analysis (NCE 55, max AGC 5×10 4 , maximum injection time 250 ms, isolation specificity 0.5 Th, resolution 60,000). Protein identification/quantification and analysis were performed with Integrated Proteomics Pipeline - IP2 (Bruker, Madison, WI. http://www.integratedproteomics.com/ ) using ProLuCID, 60 , 61 DTASelect2, 62 , 63 Census and Quantitative Analysis. Spectrum raw files were extracted into MS1, MS2 and MS3 files using RawConverter ( http://fields.scripps.edu/downloads.php ). The tandem mass spectra were searched against UniProt mouse protein database (downloaded on 03-25-2014) 64 and matched to sequences using the ProLuCID/SEQUEST algorithm (ProLuCID version 3.1) with 5 ppm peptide mass tolerance for precursor ions and 600 ppm for fragment ions. The search space included all fully and half-tryptic peptide candidates within the mass tolerance window with no-miscleavage constraint, assembled, and filtered with DTASelect2 through IP2. To estimate peptide probabilities and false-discovery rates (FDR) accurately, we used a target/decoy database containing the reversed sequences of all the proteins appended to the target database. 65 Each protein identified was required to have a minimum of one peptide of minimal length of six amino acid residues; however, this peptide had to be an excellent match with an FDR < 1% and at least one excellent peptide match. After the peptide/spectrum matches were filtered, we estimated that the peptide FDRs were ≤ 1% for each sample analysis. Resulting protein lists include subset proteins to allow for consideration of all possible protein forms implicated by at least two given peptides identified from the complex protein mixtures. Then, we used Census and Quantitative Analysis in IP2 for protein quantification of TMT MS. experiments and protein quantification was determined by summing all TMT report ion counts. TMT MS data were normalized using with a build-in method in IP2. Spyder (MIT, Python 3.7, libraries, ‘pandas’, ‘numpy’, ‘scipy’, ‘statsmodels’ and ‘bioinfokit’) was used for data analyses. RStudio (version, 1.2.1335, packages, ‘tidyverse’, ‘pheatmap’) was used for data virtualization. The Database for Annotation, Visualization and Integrated Discovery (DAVID) ( https://david.ncifcrf.gov/ ) was used for protein functional annotation analysis. Single cell RNA sequencing The snRNA seq data shown in this publication was obtained by analyzing dataset previously deposited to NCBI's Gene Expression Omnibus by our group 32 and can be accessed using GEO Series accession number GSE254780 ( https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE254780 ). Briefly, single-cell RNA sequencing (scRNA-seq) data were initially analyzed using scTE with default settings to quantify transposable element (TE) gene expression. 66 The resulting .h5ad data objects were loaded into R (v4.3) and analyzed using the Seurat package (v5.0.3). 67 Within Seurat, cells were filtered to retain those with >500 UMI counts, >200 detected features, and <15% mitochondrial content. Genes expressed in fewer than 5 cells were also removed. Doublets were subsequently identified and removed using DoubletFinder. 68 The filtered datasets were then merged. Principal Component Analysis (PCA) was performed, and an optimal number of PCs was selected based on cumulative variance and elbow point heuristics. Batch effects between samples were corrected using Harmony integration on the PCA embeddings. 69 UMAP visualization and Leiden clustering (resolution 0.4) were performed on the Harmony-corrected dimensions. Cell type annotation was performed using scType with a custom brain-specific marker gene database ( Table S2 ). 70 Finally, differentially expressed genes (DEGs) between annotated cell types were identified using FindAllMarkers function from Seurat package, and these DEGs were further filtered to identify transposable element (TE)-derived transcripts based on a provided TE annotation file provided by the scTE package. The expression of VGluT2 ( Slc17a6 ) and VGlut1 ( Slc17a7 ) genes, was examined using FeaturePlot function from the Seurat package with the blend option set to TRUE, enabling the visualization of co-expression patterns between these genes. Data showing expression of VGluT2 ( Slc17a6 ) and VGlut1 ( Slc17a7 ) genes in the subicular clusters was generated using database previously published 27 and available for online analysis on Brain RNA-seq atlas ( https://scrnaseq.janelia.org/ ). Following online analysis with the available interactive tool, gene expression values were downloaded and plotted offline using GraphPad Prizm software. Trace fear conditioning Trace fear conditioning was performed in an automated system (TSE Systems) as described previously. 71 Briefly, mice were exposed for 120 s to a unfamiliar context (Context 1), followed by a 30 s of 10 kHz tone, 15 s temporal trace, and foot shock (2 s, 0.7 mA, constant current). To prevent scent-based cues from influencing behavior, chambers were cleaned thoroughly with 70% ethanol after every session. Mice were tested for memory retrieval 24 h later in a contextually distinct novel context (Context 2, context exposure duration 60 s, Tone duration 30 s and Trace duration 15 s). Chambers were cleaned thoroughly with 1% acetic acid after every session. Freezing was scored every 5 s during Context 2 and Tone exposure and every 3 s during Trace duration, and expressed as a percentage of the total number of freezing observations during which the mice were motionless. To control for the behavioral specificity of CNO effects, we also used delay fear conditioning (DFC) and tone-light-shock fear conditioning (TLC). DFC was performed as TFC except that shock was delivered immediately after termination of tone, thus there was no trace separating tone and shock. TLC was performed as TFC except that 500 ms light pulses were delivered during the 15 s trace. For fiberphotometry experiments involving trace fear conditioning, mice were placed in soundproof chambers (Med Associates Inc., St. Albans, VT) equipped with metal-grid flooring used to administer foot shocks. Each session began with a 120 s of context exposure (Context 1), after which the animals were presented with three auditory tones (30 s duration, 80 dB, 50 ms rise time). A brief foot shock (0.5 mA, 2 s) was delivered 15 s after each tone ended. Each Tone-Trace presentation was separated by a 60 s interval. To prevent scent-based cues from influencing behavior, chambers were cleaned thoroughly with 70% ethanol after every session. Memory recall was evaluated 24 h later in a contextually distinct environment (Context 2) featuring a flat white floor and a novel odor (1% acetic acid). Mice were placed in this altered chamber and, following a 120 s baseline period, the tone was played again for 30 s. Freezing responses were measured during the Baseline, Tone, and Trace period. For pseudoconditioning, at training mice were placed in Context 1 for 120 s, after which three brief foot shocks (0.5 mA, 2 s) were applied with an inter-shock-interval of 60 s. This was followed by three unpaired Tone representations (30 s duration, 80 dB, 50 ms rise time) with inter-tone-interval of 60 During test session, mice were exposed to the same protocol, except that the foot shocks were omitted. Behavior was recorded and quantified using Video Freeze® software (Med Associates Inc., Fairfax, VT), and independently validated by manual scoring (Context 2 and Tone exposure every 5 s and Trace every 3 s) by experimenters blinded to both treatment groups and experimental design through the use of coded identifiers. All behavioral experiments were performed between 10 am and 5 pm. Littermates were randomly assigned to the different treatment conditions. All behavioral tests and immunohistochemical analyses were performed by experimenters who were blind to genotypes and drug treatments. Fiber photometry We used fiberphotomerty technique to measure real time neuronal calcium transients in freely moving animals. 72 , 73 Data acquisition was performed using Neurophotometrics FP3002 system using 2 LEDs with emission of 470 nm (GCaMP wavelength) and 415 nm (isosbestic wavelength) coupled to a 200 μm 0.37 N.A. optical fiber (FOC-BF-200-125, Neurophotmetrics) with the light power at the tip constant across trials and testing days between 10-30 μW. The fluorescence signal was collected by the same optical fiber, filtered, and focused on a BlackFly CMOS camera. Intercalated samples were acquired at 60 FPS. Raw data were used as an input and visualized by Bonsai 74 and CropPolygon function was used to define ROIs. Synchronization with movement was provided by simultaneously triggering a TTL input from the fear conditioning system (MedAssociates Inc., St. Albans, VT). To minimize the patch cord autofluorescence, prior to the recording light at 470 nm was delivered for at least 15 h. Animals were handled and habituated in their home cage to optical fiber tethering for 3 days prior to behavioral tests. For data analysis we used GuPPy, a Python toolbox for fiberphotometry analysis. 75 The isosbestic wavelength records calcium-independent events such as motion artifacts and autofluorescence, and photobleaches at the similar rate as the calcium signal and was used as a control signal when calculating fluorescence signal changes from the baseline, using following equation ΔF/F = (F observed -F fitted )/F fitted . 76 High-pass filtered and transformed to z scores data (z score = (ΔF/F-μ ΔF/F ))/σ ΔF/F , where μ is mean and σ is standard deviation, where μ and σ are calculated across the whole session) was used to combine data across multiple animals and testing days. Ca 2+ activity associated with different test phases was assessed by aligning the ΔF/F signal to time 0 at each TTL timestamp and extracting a 90 s window, spanning 30 s before to 60 s after Tone onset. Data was binned to 15 s intervals and positive and negative areas under the curve (AUC) 77 were calculated for each detected peaks and average z score for each training and testing phase. Neural-behavioral correlation analysis Phase-specific neural-behavioral correlation were conducted by using tone and trace periods independently. In addition, TFC trained and Pseudoconditioned groups were processed as separate datasets. For each animal, mean z-scored ΔF/F values for tone and trace intervals were computed and paired with the corresponding freezing percentages for those same epochs. We used whole tone interval for correlation analysis. Previous studies have shown that freezing at tone stabilizes after initial orienting response, since early part of CS can include brief head movements or postural adjustments that introduce noise into trial-by-trial estimates. 78 , 79 Pearson’s correlation analyses were performed in Graphpad Prizm using built-in correlation function. Quantification and statistical analyses Statistical analyses were performed using Graphpad Prizm software and Matlab functions. For the behavioral studies, freezing data were analyzed for Treatment (CNO or Vehicle) and Test (repeated measure) as factors using two-way repeated measures ANOVA. For fiber photometry studies two-way repeated measures ANOVA was used to compare between phases during training and testing days (Baseline, Tone, Trace, Inter-Trial-Interval). Significant F values were followed by post hoc comparisons using Tukey test. For analyses of proteomic data, we used regression analyses and unpaired two-tailed Student’s t test. Homogeneity of variance was confirmed with Levene’s test for equality of variances. Statistical differences were considered significant for all P values < 0.05. Group sizes were determined using power analyses assuming a moderate effect size of 0.5. All key findings were replicated at least twice, and mostly three times, in different sets of mice (biological replicates). Only mice with correctly placed cannulas and fibers and robust virus expression in RSP terminals or injection site (> 70% of maximal expression determined by densitometry) were included in the analyses. Details of statistical analyses are found in figure legends. All data for the preparation of graphs and statistical analysis relevant data that support the conclusions are uploaded as source data. Published: March 13, 2026 Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.115317 . Contributor Information Jelena Radulovic, Email: [email protected]. Ana Cicvaric, Email: [email protected]. Supplemental information Document S1. Figures S1–S6 mmc1.pdf (5.8MB, pdf) Table S1. Significantly enriched gene annotation terms mmc2.xlsx (14.2KB, xlsx) Table S2. Brain-specific marker genes database used for scType cell type annotation mmc3.xlsx (56.4KB, xlsx) References 1. Tulving E. Episodic and semantic memory. Organization of memory. 1972;1:1. [ Google Scholar ] 2. Tulving E. Memory and consciousness. Canadian Psychology/Psychologie canadienne. 1985;26:1–12. [ Google Scholar ] 3. Eichenbaum H. Time cells in the hippocampus: a new dimension for mapping memories. Nat. Rev. Neurosci. 2014;15:732–744. doi: 10.1038/nrn3827. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Kitamura T. Driving and regulating temporal association learning coordinated by entorhinal-hippocampal network. Neurosci. Res. 2017;121:1–6. doi: 10.1016/j.neures.2017.04.005. [ DOI ] [ PubMed ] [ Google Scholar ] 5. Kim J.J., DeCola J.P., Landeira-Fernandez J., Fanselow M.S. N-methyl-D-aspartate receptor antagonist APV blocks acquisition but not expression of fear conditioning. Behav. Neurosci. 1991;105:126–133. doi: 10.1037//0735-7044.105.1.126. [ DOI ] [ PubMed ] [ Google Scholar ] 6. Misane I., Tovote P., Meyer M., Spiess J., Ogren S.O., Stiedl O. Time-dependent involvement of the dorsal hippocampus in trace fear conditioning in mice. Hippocampus. 2005;15:418–426. doi: 10.1002/hipo.20067. [ DOI ] [ PubMed ] [ Google Scholar ] 7. Gilmartin M.R., Miyawaki H., Helmstetter F.J., Diba K. Prefrontal activity links nonoverlapping events in memory. J. Neurosci. 2013;33:10910–10914. doi: 10.1523/JNEUROSCI.0144-13.2013. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Huerta P.T., Sun L.D., Wilson M.A., Tonegawa S. Formation of temporal memory requires NMDA receptors within CA1 pyramidal neurons. Neuron. 2000;25:473–480. doi: 10.1016/s0896-6273(00)80909-5. [ DOI ] [ PubMed ] [ Google Scholar ] 9. Puhger K., Crestani A.P., Diniz C.R.A.F., Wiltgen B.J. The hippocampus contributes to retroactive stimulus associations during trace fear conditioning. iScience. 2024;27 doi: 10.1016/j.isci.2024.109035. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Roy D.S., Kitamura T., Okuyama T., Ogawa S.K., Sun C., Obata Y., Yoshiki A., Tonegawa S. Distinct Neural Circuits for the Formation and Retrieval of Episodic Memories. Cell. 2017;170:1000–1012.e19. doi: 10.1016/j.cell.2017.07.013. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 11. Suh J., Rivest A.J., Nakashiba T., Tominaga T., Tonegawa S. Entorhinal cortex layer III input to the hippocampus is crucial for temporal association memory. Science. 2011;334:1415–1420. doi: 10.1126/science.1210125. [ DOI ] [ PubMed ] [ Google Scholar ] 12. Kitamura T., Pignatelli M., Suh J., Kohara K., Yoshiki A., Abe K., Tonegawa S. Island cells control temporal association memory. Science. 2014;343:896–901. doi: 10.1126/science.1244634. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Bai T., Zhan L., Zhang N., Lin F., Saur D., Xu C. Learning-prolonged maintenance of stimulus information in CA1 and subiculum during trace fear conditioning. Cell Rep. 2023;42 doi: 10.1016/j.celrep.2023.112853. [ DOI ] [ PubMed ] [ Google Scholar ] 14. Kwapis J.L., Jarome T.J., Lee J.L., Gilmartin M.R., Helmstetter F.J. Extinguishing trace fear engages the retrosplenial cortex rather than the amygdala. Neurobiol. Learn. Mem. 2014;113:41–54. doi: 10.1016/j.nlm.2013.09.007. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 15. Kwapis J.L., Jarome T.J., Lee J.L., Helmstetter F.J. The retrosplenial cortex is involved in the formation of memory for context and trace fear conditioning. Neurobiol. Learn. Mem. 2015;123:110–116. doi: 10.1016/j.nlm.2015.06.007. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 16. Opalka A.N., Wang D.V. Hippocampal efferents to retrosplenial cortex and lateral septum are required for memory acquisition. Learn. Mem. 2020;27:310–318. doi: 10.1101/lm.051797.120. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 17. Garvert A.C., Bieler M., Witoelar A., Vervaeke K. Area-specific encoding of temporal information in the neocortex. Cell Rep. 2025;44 doi: 10.1016/j.celrep.2025.115363. [ DOI ] [ PubMed ] [ Google Scholar ] 18. Yamawaki N., Li X., Lambot L., Ren L.Y., Radulovic J., Shepherd G.M.G. Long-range inhibitory intersection of a retrosplenial thalamocortical circuit by apical tuft-targeting CA1 neurons. Nat. Neurosci. 2019;22:618–626. doi: 10.1038/s41593-019-0355-x. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Yamawaki N., Corcoran K.A., Guedea A.L., Shepherd G.M.G., Radulovic J. Differential Contributions of Glutamatergic Hippocampal→Retrosplenial Cortical Projections to the Formation and Persistence of Context Memories. Cereb. Cortex. 2019;29:2728–2736. doi: 10.1093/cercor/bhy142. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Corcoran K.A., Yamawaki N., Leaderbrand K., Radulovic J. Role of retrosplenial cortex in processing stress-related context memories. Behav. Neurosci. 2018;132:388–395. doi: 10.1037/bne0000223. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. McEchron M.D., Bouwmeester H., Tseng W., Weiss C., Disterhoft J.F. Hippocampectomy disrupts auditory trace fear conditioning and contextual fear conditioning in the rat. Hippocampus. 1998;8:638–646. doi: 10.1002/(SICI)1098-1063(1998)8:6<638::AID-HIPO6>3.0.CO;2-Q. [ DOI ] [ PubMed ] [ Google Scholar ] 22. Stachniak T.J., Ghosh A., Sternson S.M. Chemogenetic synaptic silencing of neural circuits localizes a hypothalamus-->midbrain pathway for feeding behavior. Neuron. 2014;82:797–808. doi: 10.1016/j.neuron.2014.04.008. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Harris J.A., Hirokawa K.E., Sorensen S.A., Gu H., Mills M., Ng L.L., Bohn P., Mortrud M., Ouellette B., Kidney J., et al. Anatomical characterization of Cre driver mice for neural circuit mapping and manipulation. Front. Neural Circuits. 2014;8:76. doi: 10.3389/fncir.2014.00076. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Vong L., Ye C., Yang Z., Choi B., Chua S., Jr., Lowell B.B. Leptin action on GABAergic neurons prevents obesity and reduces inhibitory tone to POMC neurons. Neuron. 2011;71:142–154. doi: 10.1016/j.neuron.2011.05.028. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. El Mestikawy S., Wallén-Mackenzie A., Fortin G.M., Descarries L., Trudeau L.E. From glutamate co-release to vesicular synergy: vesicular glutamate transporters. Nat. Rev. Neurosci. 2011;12:204–216. doi: 10.1038/nrn2969. [ DOI ] [ PubMed ] [ Google Scholar ] 26. Fremeau R.T., Jr., Troyer M.D., Pahner I., Nygaard G.O., Tran C.H., Reimer R.J., Bellocchio E.E., Fortin D., Storm-Mathisen J., Edwards R.H. The expression of vesicular glutamate transporters defines two classes of excitatory synapse. Neuron. 2001;31:247–260. doi: 10.1016/s0896-6273(01)00344-0. [ DOI ] [ PubMed ] [ Google Scholar ] 27. Cembrowski M.S., Wang L., Lemire A.L., Copeland M., DiLisio S.F., Clements J., Spruston N. The subiculum is a patchwork of discrete subregions. eLife. 2018;7 doi: 10.7554/eLife.37701. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Cembrowski M.S., Wang L., Sugino K., Shields B.C., Spruston N. Hipposeq: a comprehensive RNA-seq database of gene expression in hippocampal principal neurons. eLife. 2016;5 doi: 10.7554/eLife.14997. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Cembrowski M.S., Phillips M.G., DiLisio S.F., Shields B.C., Winnubst J., Chandrashekar J., Bas E., Spruston N. Dissociable Structural and Functional Hippocampal Outputs via Distinct Subiculum Cell Classes. Cell. 2018;174:1036. doi: 10.1016/j.cell.2018.07.039. [ DOI ] [ PubMed ] [ Google Scholar ] 30. Wang Y.Z., Perez-Rosello T., Smukowski S.N., Surmeier D.J., Savas J.N. Neuron type-specific proteomics reveals distinct Shank3 proteoforms in iSPNs and dSPNs lead to striatal synaptopathy in Shank3B(-/-) mice. Mol. Psychiatry. 2024;29:2372–2388. doi: 10.1038/s41380-024-02493-w. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Kim J., Zhao T., Petralia R.S., Yu Y., Peng H., Myers E., Magee J.C. mGRASP enables mapping mammalian synaptic connectivity with light microscopy. Nat. Methods. 2011;9:96–102. doi: 10.1038/nmeth.1784. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Jovasevic V., Wood E.M., Cicvaric A., Zhang H., Petrovic Z., Carboncino A., Parker K.K., Bassett T.E., Moltesen M., Yamawaki N., et al. Formation of memory assemblies through the DNA-sensing TLR9 pathway. Nature. 2024;628:145–153. doi: 10.1038/s41586-024-07220-7. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 33. Curzon P., Rustay N.R., Browman K.E. In: Methods of Behavior Analysis in Neuroscience. 2nd ed. Buccafusco J.J., editor. CRC Press/Taylor & Francis; Boca Raton, FL: 2009. Cued and Contextual Fear Conditioning for Rodents. Chapter 2. [ PubMed ] [ Google Scholar ] 34. Berry C.T., Sceniak M.P., Zhou L., Sabo S.L. Developmental up-regulation of vesicular glutamate transporter-1 promotes neocortical presynaptic terminal development. PLoS One. 2012;7 doi: 10.1371/journal.pone.0050911. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Guillaud L., Dimitrov D., Takahashi T. Presynaptic morphology and vesicular composition determine vesicle dynamics in mouse central synapses. eLife. 2017;6 doi: 10.7554/eLife.24845. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Weston M.C., Nehring R.B., Wojcik S.M., Rosenmund C. Interplay between VGLUT isoforms and endophilin A1 regulates neurotransmitter release and short-term plasticity. Neuron. 2011;69:1147–1159. doi: 10.1016/j.neuron.2011.02.002. [ DOI ] [ PubMed ] [ Google Scholar ] 37. Cavalieri D., Angelova A., Islah A., Lopez C., Bocchio M., Bollmann Y., Baude A., Cossart R. CA1 pyramidal cell diversity is rooted in the time of neurogenesis. eLife. 2021;10 doi: 10.7554/eLife.69270. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 38. Mizuseki K., Diba K., Pastalkova E., Buzsáki G. Hippocampal CA1 pyramidal cells form functionally distinct sublayers. Nat. Neurosci. 2011;14:1174–1181. doi: 10.1038/nn.2894. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 39. Pilkiw M., Takehara-Nishiuchi K. Neural representations of time-linked memory. Neurobiol. Learn. Mem. 2018;153:57–70. doi: 10.1016/j.nlm.2018.03.024. [ DOI ] [ PubMed ] [ Google Scholar ] 40. Raybuck J.D., Lattal K.M. Bridging the interval: theory and neurobiology of trace conditioning. Behav. Processes. 2014;101:103–111. doi: 10.1016/j.beproc.2013.08.016. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. Veres J.M., Andrasi T., Nagy-Pal P., Hajos N. CaMKIIalpha Promoter-Controlled Circuit Manipulations Target Both Pyramidal Cells and Inhibitory Interneurons in Cortical Networks. eNeuro. 2023;10 doi: 10.1523/ENEURO.0070-23.2023. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Ahmed M.S., Priestley J.B., Castro A., Stefanini F., Solis Canales A.S., Balough E.M., Lavoie E., Mazzucato L., Fusi S., Losonczy A. Hippocampal Network Reorganization Underlies the Formation of a Temporal Association Memory. Neuron. 2020;107:283–291.e6. doi: 10.1016/j.neuron.2020.04.013. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 43. Gilmartin M.R., McEchron M.D. Single neurons in the dentate gyrus and CA1 of the hippocampus exhibit inverse patterns of encoding during trace fear conditioning. Behav. Neurosci. 2005;119:164–179. doi: 10.1037/0735-7044.119.1.164. [ DOI ] [ PubMed ] [ Google Scholar ] 44. Jinno S., Klausberger T., Marton L.F., Dalezios Y., Roberts J.D.B., Fuentealba P., Bushong E.A., Henze D., Buzsáki G., Somogyi P. Neuronal diversity in GABAergic long-range projections from the hippocampus. J. Neurosci. 2007;27:8790–8804. doi: 10.1523/JNEUROSCI.1847-07.2007. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 45. Solomon P.R., Vander Schaaf E.R., Thompson R.F., Weisz D.J. Hippocampus and trace conditioning of the rabbit’s classically conditioned nictitating membrane response. Behav. Neurosci. 1986;100:729–744. doi: 10.1037//0735-7044.100.5.729. [ DOI ] [ PubMed ] [ Google Scholar ] 46. Moyer J.R., Jr., Deyo R.A., Disterhoft J.F. Hippocampectomy disrupts trace eye-blink conditioning in rabbits. Behav. Neurosci. 1990;104:243–252. doi: 10.1037//0735-7044.104.2.243. [ DOI ] [ PubMed ] [ Google Scholar ] 47. Gabriel M., Sparenborg S.P., Stolar N. Hippocampal control of cingulate cortical and anterior thalamic information processing during learning in rabbits. Exp. Brain Res. 1987;67:131–152. doi: 10.1007/BF00269462. [ DOI ] [ PubMed ] [ Google Scholar ] 48. Kitanishi T., Umaba R., Mizuseki K. Robust information routing by dorsal subiculum neurons. Sci. Adv. 2021;7 doi: 10.1126/sciadv.abf1913. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 49. Picard L., Cousin S., Guillery-Girard B., Eustache F., Piolino P. How do the different components of episodic memory develop? Role of executive functions and short-term feature-binding abilities. Child Dev. 2012;83:1037–1050. doi: 10.1111/j.1467-8624.2012.01736.x. [ DOI ] [ PubMed ] [ Google Scholar ] 50. Lee J.K., Wendelken C., Bunge S.A., Ghetti S. A Time and Place for Everything: Developmental Differences in the Building Blocks of Episodic Memory. Child Dev. 2016;87:194–210. doi: 10.1111/cdev.12447. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 51. Saunders A., Johnson C.A., Sabatini B.L. Novel recombinant adeno-associated viruses for Cre activated and inactivated transgene expression in neurons. Front. Neural Circuits. 2012;6:47. doi: 10.3389/fncir.2012.00047. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 52. Bäck S., Necarsulmer J., Whitaker L.R., Coke L.M., Koivula P., Heathward E.J., Fortuno L.V., Zhang Y., Yeh C.G., Baldwin H.A., et al. Neuron-Specific Genome Modification in the Adult Rat Brain Using CRISPR-Cas9 Transgenic Rats. Neuron. 2019;102:105–119.e8. doi: 10.1016/j.neuron.2019.01.035. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 53. Krashes M.J., Koda S., Ye C., Rogan S.C., Adams A.C., Cusher D.S., Maratos-Flier E., Roth B.L., Lowell B.B. Rapid, reversible activation of AgRP neurons drives feeding behavior in mice. J. Clin. Investig. 2011;121:1424–1428. doi: 10.1172/JCI46229. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 54. Jovasevic V., Corcoran K.A., Leaderbrand K., Yamawaki N., Guedea A.L., Chen H.J., Shepherd G.M.G., Radulovic J. GABAergic mechanisms regulated by miR-33 encode state-dependent fear. Nat. Neurosci. 2015;18:1265–1271. doi: 10.1038/nn.4084. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 55. Wang Y.Z., Savas J.N. Uncovering Discrete Synaptic Proteomes to Understand Neurological Disorders. Proteomes. 2018;6:30. doi: 10.3390/proteomes6030030. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 56. Loh K.H., Stawski P.S., Draycott A.S., Udeshi N.D., Lehrman E.K., Wilton D.K., Svinkina T., Deerinck T.J., Ellisman M.H., Stevens B., et al. Proteomic Analysis of Unbounded Cellular Compartments: Synaptic Clefts. Cell. 2016;166:1295–1307.e21. doi: 10.1016/j.cell.2016.07.041. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 57. McAlister G.C., Nusinow D.P., Jedrychowski M.P., Wühr M., Huttlin E.L., Erickson B.K., Rad R., Haas W., Gygi S.P. MultiNotch MS3 enables accurate, sensitive, and multiplexed detection of differential expression across cancer cell line proteomes. Anal. Chem. 2014;86:7150–7158. doi: 10.1021/ac502040v. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 58. Ting L., Rad R., Gygi S.P., Haas W. MS3 eliminates ratio distortion in isobaric multiplexed quantitative proteomics. Nat. Methods. 2011;8:937–940. doi: 10.1038/nmeth.1714. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 59. Weekes M.P., Tomasec P., Huttlin E.L., Fielding C.A., Nusinow D., Stanton R.J., Wang E.C.Y., Aicheler R., Murrell I., Wilkinson G.W.G., et al. Quantitative temporal viromics: an approach to investigate host-pathogen interaction. Cell. 2014;157:1460–1472. doi: 10.1016/j.cell.2014.04.028. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 60. Eng J.K., McCormack A.L., Yates J.R. An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database. J. Am. Soc. Mass Spectrom. 1994;5:976–989. doi: 10.1016/1044-0305(94)80016-2. [ DOI ] [ PubMed ] [ Google Scholar ] 61. Xu T., Park S.K., Venable J.D., Wohlschlegel J.A., Diedrich J.K., Cociorva D., Lu B., Liao L., Hewel J., Han X., et al. ProLuCID: An improved SEQUEST-like algorithm with enhanced sensitivity and specificity. J. Proteomics. 2015;129:16–24. doi: 10.1016/j.jprot.2015.07.001. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 62. Cociorva D., D L.T., Yates J.R. Validation of tandem mass spectrometry database search results using DTASelect. Curr Protoc Bioinformatics. 2007;Chapter 13 doi: 10.1002/0471250953.bi1304s16. Unit 13.4. [ DOI ] [ PubMed ] [ Google Scholar ] 63. Tabb D.L., McDonald W.H., Yates J.R., 3rd DTASelect and Contrast: tools for assembling and comparing protein identifications from shotgun proteomics. J. Proteome Res. 2002;1:21–26. doi: 10.1021/pr015504q. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 64. UniProt Consortium UniProt: a hub for protein information. Nucleic Acids Res. 2015;43:D204–D212. doi: 10.1093/nar/gku989. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 65. Peng J., Elias J.E., Thoreen C.C., Licklider L.J., Gygi S.P. Evaluation of multidimensional chromatography coupled with tandem mass spectrometry (LC/LC-MS/MS) for large-scale protein analysis: the yeast proteome. J. Proteome Res. 2003;2:43–50. doi: 10.1021/pr025556v. [ DOI ] [ PubMed ] [ Google Scholar ] 66. He J., Babarinde I.A., Sun L., Xu S., Chen R., Shi J., Wei Y., Li Y., Ma G., Zhuang Q., et al. Identifying transposable element expression dynamics and heterogeneity during development at the single-cell level with a processing pipeline scTE. Nat. Commun. 2021;12:1456. doi: 10.1038/s41467-021-21808-x. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 67. Hao Y., Stuart T., Kowalski M.H., Choudhary S., Hoffman P., Hartman A., Srivastava A., Molla G., Madad S., Fernandez-Granda C., Satija R. Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat. Biotechnol. 2024;42:293–304. doi: 10.1038/s41587-023-01767-y. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 68. McGinnis C.S., Murrow L.M., Gartner Z.J. DoubletFinder: Doublet Detection in Single-Cell RNA Sequencing Data Using Artificial Nearest Neighbors. Cell Syst. 2019;8:329–337.e4. doi: 10.1016/j.cels.2019.03.003. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 69. Korsunsky I., Millard N., Fan J., Slowikowski K., Zhang F., Wei K., Baglaenko Y., Brenner M., Loh P.-R., Raychaudhuri S. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat. Methods. 2019;16:1289–1296. doi: 10.1038/s41592-019-0619-0. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 70. Ianevski A., Giri A.K., Aittokallio T. Fully-automated and ultra-fast cell-type identification using specific marker combinations from single-cell transcriptomic data. Nat. Commun. 2022;13:1246. doi: 10.1038/s41467-022-28803-w. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 71. Corcoran K.A., Donnan M.D., Tronson N.C., Guzmán Y.F., Gao C., Jovasevic V., Guedea A.L., Radulovic J. NMDA receptors in retrosplenial cortex are necessary for retrieval of recent and remote context fear memory. J. Neurosci. 2011;31:11655–11659. doi: 10.1523/JNEUROSCI.2107-11.2011. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 72. Gunaydin L.A., Grosenick L., Finkelstein J.C., Kauvar I.V., Fenno L.E., Adhikari A., Lammel S., Mirzabekov J.J., Airan R.D., Zalocusky K.A., et al. Natural neural projection dynamics underlying social behavior. Cell. 2014;157:1535–1551. doi: 10.1016/j.cell.2014.05.017. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 73. Martianova E., Aronson S., Proulx C.D. Multi-Fiber Photometry to Record Neural Activity in Freely-Moving Animals. J. Vis. Exp. 2019 doi: 10.3791/60278. [ DOI ] [ PubMed ] [ Google Scholar ] 74. Lopes G., Bonacchi N., Frazão J., Neto J.P., Atallah B.V., Soares S., Moreira L., Matias S., Itskov P.M., Correia P.A., et al. Bonsai: an event-based framework for processing and controlling data streams. Front. Neuroinform. 2015;9:7. doi: 10.3389/fninf.2015.00007. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 75. Sherathiya V.N., Schaid M.D., Seiler J.L., Lopez G.C., Lerner T.N. GuPPy, a Python toolbox for the analysis of fiber photometry data. Sci. Rep. 2021;11 doi: 10.1038/s41598-021-03626-9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 76. Lerner T.N., Shilyansky C., Davidson T.J., Evans K.E., Beier K.T., Zalocusky K.A., Crow A.K., Malenka R.C., Luo L., Tomer R., Deisseroth K. Intact-Brain Analyses Reveal Distinct Information Carried by SNc Dopamine Subcircuits. Cell. 2015;162:635–647. doi: 10.1016/j.cell.2015.07.014. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 77. Holly E.N., Davatolhagh M.F., Choi K., Alabi O.O., Vargas Cifuentes L., Fuccillo M.V. Striatal Low-Threshold Spiking Interneurons Regulate Goal-Directed Learning. Neuron. 2019;103:92–101.e6. doi: 10.1016/j.neuron.2019.04.016. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 78. Maren S. Neurobiology of Pavlovian fear conditioning. Annu. Rev. Neurosci. 2001;24:897–931. doi: 10.1146/annurev.neuro.24.1.897. [ DOI ] [ PubMed ] [ Google Scholar ] 79. Trott J.M., Hoffman A.N., Zhuravka I., Fanselow M.S. Conditional and unconditional components of aversively motivated freezing, flight and darting in mice. eLife. 2022;11 doi: 10.7554/eLife.75663. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Document S1. Figures S1–S6 mmc1.pdf (5.8MB, pdf) Table S1. Significantly enriched gene annotation terms mmc2.xlsx (14.2KB, xlsx) Table S2. Brain-specific marker genes database used for scType cell type annotation mmc3.xlsx (56.4KB, xlsx) Data Availability Statement • All results of synaptic proteomes analysis have been deposited through an interactive portal ( https://proteome-phenome-atlas.com/ ). The mass spectrometry (MS) data presented in this study have been deposited in the Mass Spectrometry Interactive Virtual Environment (MassIVE) under identifier MSV000098056 ( https://massive.ucsd.edu/ProteoSAFe/dataset.jsp?task=aa8f8aa9910b400b8382992f39700ba1 ) and ProteomeXchange under identifier PXD064462 ( https://proteomecentral.proteomexchange.org/cgi/GetDataset?ID=PXD064462 ). • To generate snRNA-seq graphs, this paper analyzes existing, publicly available data, accessible at NCBI Gene Expression Omnibus database under accession GSE254780 and source code available at https://github.com/RadulovicLab/Nature-2024 . • All other data reported in this paper and any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. All other software and analytical methods used in this study are publicly available, as listed in the key resources table . Articles from iScience are provided here courtesy of Elsevier ACTIONS View on publisher site PDF (12.9 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top