Modulation of long-term memory for object recognition via HDAC inhibition

Abstract

Histone acetylation is a chromatin modification critically involved in gene regulation during many neural processes. The enzymes that regulate levels of histone acetylation are histone acetyltransferases (HATs), which activate gene expression and histone deacetylases (HDACs), that repress gene expression. Acetylation together with other histone and DNA modifications regulate transcription profiles for specific cellular functions. Our previous research has demonstrated a pivotal role for cyclicAMP response element binding protein (CREB)-binding protein (CBP), a histone acetyltransferase, in long-term memory for novel object recognition (NOR). In fact, every genetically modifiedCbp mutant mouse characterized thus far exhibits impaired long-term memory for NOR. These results suggest that long-term memory for NOR is especially sensitive to alterations in CBP activity. Thus, in the current study, we examined the role of HDACs in memory for NOR. We found that inducing a histone hyperacetylated state via HDAC inhibition transforms a learning event that would not normally result in long-term memory into an event that is now remembered long-term. We have also found that HDAC inhibition generates a type of long-term memory that persists beyond a point at which normal memory for NOR fails. This result is particularly interesting because one alluring aspect of examining the role of chromatin modifications in modulating transcription required for long-term memory processes is that these modifications may provide potentially stable epigenetic markers in the service of activating and/or maintaining transcriptional processes.

In the past 5 years, chromatin modification has been identified as a pivotal molecular mechanism underlying certain forms of synaptic plasticity and memory. One of the best studied chromatin modifications is histone acetylation, which modulates histone-DNA interactions and provides recruitment sites for additional chromatin regulatory proteins (reviewed in ref. 1). The enzymes that regulate levels of histone acetylation are histone acetyltransferases (HATs), which generally activate gene expression and histone deacetylases (HDACs), which generally repress gene expression (2). Acetylation together with other histone and DNA modifications regulate transcription profiles for specific cellular functions. Recently, HAT enzymes, such as cyclicAMP response element binding protein (CREB)-binding protein (CBP) and HDACs, have been shown to be essential components of the molecular mechanisms underlying memory formation.

By using genetically modified Cbp mutant mice, we and others have shown that CBP is necessary for specific forms of hippocampal long-term potentiation (LTP), hippocampus-dependent long-term memory, and long-term memory for object recognition (3–8). Interestingly, all of the different types of genetically modified Cbp mutant mice studied to date exhibit deficits in long-term memory for object recognition (3–7); reviewed in ref. 1. This evidence suggests that brain regions required for long-term memory for object recognition (9–16) may be particularly sensitive to alterations in CBP activity and histone acetylation. The results from Cbp mutant mice with regard to long-term memory for object recognition suggest that this type of memory may be well suited for studying the role of histone modifying enzymes in memory formation. Because CBP HAT activity is opposed by HDAC activity, we examined the role of HDACs as potential memory suppressor genes involved modulating molecular mechanisms required for long-term memory for object recognition in this study.

Previously, we demonstrated that blocking HDAC activity with nonspecific HDAC inhibitors, such as trichostatin A (TSA) or sodium butyrate (NaBut), enhances synaptic plasticity and memory, suggesting that HDACs may actually serve to return chromatin to a repressive state and silence transcription required for long-term memory formation (17, 18). In the current study, we show that HDAC inhibition can transform a learning event that does not normally lead to long-term memory for object recognition into a long-lasting form of memory. Moreover, HDAC inhibition during memory consolidation generates a form of long-term memory that persists beyond the point at which normal memory fails. Together, these results suggest HDACs may serve as critical memory suppressor genes and show that HDAC inhibition may generate more persistent forms of long-term memory, which has great therapeutic and translational value.

Results

Identification of Behavioral Parameters Affecting Long-Term Memory for Novel Object Recognition.

The overall aim of this study was to examine the role of histone-modifying enzymes in the formation of object recognition memory and to determine how altering those enzymes changes memory formation. Therefore, we first examined what parameters are critical for establishing long-term memory for novel object recognition (NOR). We first assessed the effect of training duration and habituation duration on memory formation for NOR. We examined 3 different groups. Group 1 received habituation and a 10-min training session. Group 2 received habituation and a 3-min training session. Group 3 received no habituation and a 10-min training session. The percentage of time spent exploring the objects during training did not significantly differ between training conditions (Fig. 1A). All 3 groups were given a 24-h retention test. A 1-way ANOVA showed that the effect of training was significant [F_{(2, 27)} = 6.27, P < 0.01]. Post-hoc analysis using the Student-Newman-Keuls test (α = 0.05) indicated that Group 1 had a significantly higher discrimination index (DI = 48.1 ± 10.0%, n = 10) than both Groups 2 (DI = 13.2 ± 8.7%, n = 10) and 3 (DI = 1.7 ± 10.2%, n = 10); no other differences were statistically significant (Fig. 1B). These results demonstrate that a training duration of 3 min is not sufficient for animals to form a long-term memory for NOR, and nonhabituated mice are unable to form a memory for the familiar object even with a 10-min exposure to the familiar object. Also, we have found that a 10-min training period is sufficient to generate long-term memory for the familiar object only when the mice are habituated to the context.

Fig. 1.

Effect of training period (3 min versus 10 min) and habituation to context on long-term memory for the familiar object. (A) Two groups of mice were habituated and then exposed to the objects for either 10 min or 3 min. A third group received no habituation followed by a 10-min training period. All 3 groups exhibited equal exploration times for each object. (B) During the 24-h retention test, animals receiving 10 min of training after 3 days of habituation (n = 10) displayed a significant preference for the novel object, whereas those that received 3 min of training (n = 10) or 10 min of training with no habituation (n = 10) showed no significant preference (P > 0.05). (C) During a 90-min retention test, animals receiving 10 min of training (n = 10) displayed a significant preference for the novel object, whereas those that received 3 min of training (n = 10) showed no significant preference.

To test training duration on short-term memory, a retention test was given 90 min after training to a different set of mice. Mice receiving a 10-min training session exhibit significant memory for the familiar object as demonstrated by the discrimination index [DI = 42.7 ± 6.2%; t₍₁₀₎ = 5.87; P < 0.001; Fig. 1C]. In contrast, mice receiving a 3-min training session did not exhibit memory for the familiar object. These results establish that a 10-min training period is sufficient to generate short- and long-term memory for the familiar object, but that a 3-min training period does not result in either short- or long-term memory. Thus, we next examined how increasing histone acetylation via HDAC inhibition affects memory for NOR.

HDAC Inhibition Transforms a Learning Event that Would Not Normally Result in Long-Term Memory into an Event That Is Now Remembered Long-Term.

In a previous study, we demonstrated that HDAC inhibition could transform a transient transcription-independent form of E-LTP into a long-lasting robust transcription-dependent form of LTP (18). However, at the behavioral level, whether HDAC inhibition can transform a learning event that would normally not result in long-term memory into a form of memory that is long-lasting remains to be shown. Thus, we next examined the effect of HDAC inhibition on memory after a 3-min training period. Mice were subject to habituation followed by a 3-min training period. Immediately after training, mice were administered either vehicle or sodium butyrate (NaBut) and given a retention test 24 h later. We and others have shown that a single systemic injection of NaBut enhances memory (17, 19) associated with increases in histone acetylation in the brain (20). As shown in Fig. 2A, NaBut-treated mice exhibited significantly increased memory for the familiar object (NaBut, DI = 36.9 ± 7.5%; n = 8; t = 2.38, P < 0.05) compared with vehicle controls (vehicle, DI = 12.6 ± 6.9%; n = 8). Similar results were obtained using a 0.6-g/kg dose of NaBut (supporting information (SI) Fig. S1). These results show that HDAC inhibition can enhance memory for the familiar object and transform what is learned by just a 3-min training period (that does not lead to long-term memory by itself) into an event leading to long-term memory.

Fig. 2.

HDAC inhibition facilitates long-term memory, measured at 24 h, after 3 min of object recognition training. (A) During a 24-h retention test, mice that received 3 min of training followed by an i.p. administration of NaBut (n = 8) displayed significantly enhanced preference for the novel object compared with vehicle-treated mice (n = 8). (B) During a 90-min retention test, mice that received NaBut (n = 8) immediately after 3 min of training showed no preference for the novel object compared with vehicle-treated mice (n = 8). (C) Mice that did not receive habituation to the context before a 10-min training period were unable to form a long-term memory for the familiar object. On the training day, mice were exposed to the experimental apparatus with 2 identical objects for 10 min. Both groups showed similar exploration times for each object. During the 24-h retention test, mice treated with NaBut (n = 8) or vehicle (n = 8) did not display a significant preference for the novel object.

We next examined the effect of HDAC inhibition on short-term memory. Mice were subject to habituation, received a 3-min training period, immediately followed with vehicle or NaBut treatment and tested 90 min later. Fig. 2B shows that mice receiving NaBut (NaBut, DI = 16.2 ± 4.5%; n = 8; t = 0.095, P = 0.98) exhibit similar memory for NOR as vehicle-treated animals (vehicle, DI = 15.6 ± 4.2%; n = 8). Together, these findings suggest that the 3-min training is sufficient to initiate molecular mechanism that can be “captured” by HDAC inhibition and lead to long-term memory, but that short-term memory processes are not affected.

As shown in Fig. 1B, mice subject to a 10-min training period without habituation exhibit no long-term memory for object recognition. Because these animals tend to explore the context rather than the objects, they do not distinguish between novel and familiar objects. Thus, as predicted, mice receiving a 10-min training period without habituation still do not exhibit memory for the familiar object even when treated with an HDAC inhibitor (vehicle, DI = 13.0 ± 12.5%; n = 8; NaBut, DI = 8.3 ± 11.9%; n = 8; t = 0.27, P = 0.79; Fig. 2C). These results serve as a control to demonstrate that HDAC inhibition does not simply increase a performance variable that confounds interpretation of our data.

HDAC Inhibition Generates a Type of Long-Term Memory That Persists Beyond a Point at Which Normal Memory for NOR Fails.

We next determined whether HDAC inhibition-dependent memory could persist over long retention intervals. Again, mice were subject to habituation, 3 min of training, followed immediately by an injection of vehicle or NaBut. To measure persistence, a retention test was given 7 days after training. Mice treated with NaBut (NaBut, DI = 39.4 ± 7.6%; n = 9; t = 2.49, P < 0.01) showed significantly better memory for the familiar object than vehicle controls (vehicle, DI = 16.0 ± 5.5%; n = 8; Fig. 3A). These results demonstrate that HDAC inhibition is able to induce a form of memory that persists beyond a typical 24-h memory retention test.

Fig. 3.

HDAC inhibition generates a type of long-term memory that persists at least 7 days. (A) During the 7-day retention trial, animals receiving 3 min of training followed by an i.p. administration of NaBut (n = 9) displayed a significant preference for the novel object compared with vehicle-treated controls (n = 8). (B) To determine the effect of NaBut on retrieval, animals receiving a 10-min training period were injected with i.p. NaBut 1 h before a 7-day retention test. We examined 2 different doses of NaBut (normal dose: 1.2 g/kg, n = 9; low dose: 0.6 g/kg, n = 11; vehicle, n = 9). Neither dose of NaBut affected behavior of mice in the retrieval test compared with vehicle-treated mice. (C) To demonstrate that NaBut was active during the retention test in Fig. 3B, we retested the same mice 24 h later using a different novel object paired with the former novel object in the 7-day retention test. NaBut-treated mice (low dose, n = 11; normal dose, n = 9) exhibited significantly enhanced preference for the novel object compared with vehicle-treated mice (n = 9).

To examine the effect of inhibiting HDAC activity during the 7-day retention test, we delivered NaBut (normal dose, 1.2 g/kg; low dose, 0.6 g/kg) or vehicle 1 h before the 7-day test. In this experiment, mice received habituation and then a 10-min training session, which results in 24-h long-term memory. Seven days later, mice were administered vehicle or NaBut 1 h before the retention test. A repeated-measures ANOVA (between the 7-day test shown in Fig. 3B and the 24-h test a day later on the same mice shown in Fig. 3C) revealed a main effect of test [F_{(1, 26)} = 19.07, P < 0.01] and treatment [F_{(2, 26)} = 9.20, P < 0.01] with a significant interaction between test and treatment [F_{(2, 26)} = 15.14, P < 0.01]. Bonferroni post-hoc comparisons revealed no differences among groups during the 7-day retention test (Fig. 3B). The results in Fig. 3B indicate that HDAC inhibition before retention has no effect.

To demonstrate that the NaBut delivered 1 h before the 7-day retention test was active, we subjected the same set of mice to a subsequent retention test on day 8. We predicted that although HDAC inhibition did not affect retrieval, the animals are indeed learning something new during that retrieval test in the presence of HDAC inhibition, and thus NaBut-treated mice should exhibit enhanced preference for a novel object in a subsequent retention test on day 8. Indeed, we observed that mice treated with either a low dose (0.6 g/kg, DI = 40.6 ± 4.9%, n = 11; t₁₉ = 5.96, P < 0.01, Bonferroni post-hoc) or a normal dose (1.2 g/kg, DI = 44.1 ± 4.4%, n = 9; t₁₇ = 6.15, P < 0.01, Bonferroni post-hoc) showed significantly better memory for the familiar object than vehicle controls (vehicle, DI = −1.9 ± 5.6%, n = 9; Fig. 3C). Two results were obtained from experiments shown in Fig. 3. First, a 10-min training session is not sufficient to generate a form of long-term memory that persists 7 days. Second, NaBut delivered 1 h before the 7-day retention test had no effect. These results indicate that HDAC inhibition generates a form of long-term memory that persists up to at least 7 days.

To investigate the role of CBP in HDAC inhibition-dependent long-term memory formation, we examined Cbp knockin mice (Cbp^KIX/KIX) carrying a mutation in the KIX (CREB-binding) domain of CBP (21). We have found these mice to exhibit normal short-term memory for object recognition, but impaired long-term memory (7). To examine these mice in the current study, we needed to determine whether these mice also exhibit impairments under the training conditions used here. Fig. S2 shows that Cbp^KIX/KIX exhibit severe long-term memory impairments when given a 10-min training period followed by a 24-h retention test and that HDAC inhibition can ameliorate that memory impairment.

We next examined whether CBP was required for HDAC inhibition-dependent long-term memory that persists over a 7-day period (as shown in Fig. 3 in C57BL/6 mice). Cbp^KIX/KIX homozygous mice and wild-type Cbp^+/+ littermates were subject to habituation, a 10-min training period, and then immediately after given an injection of either vehicle or NaBut. In this experiment, the mice received a 7-day retention test. A 2-way ANOVA yielded no significant difference because of genotype and no significant interaction between genotype and treatment [Main effect of genotype F_{(1, 26)} = .016, P = 0.90; Genotype × Treatment interaction, F_{(1, 26)} = 1.27, P = 0.268]. However, there was a significant difference among treatment [F_{(1, 26)} = 27.32, P < 0.001]. One-way ANOVA revealed that neither Cbp^KIX/KIX homozygous mice treated with vehicle (Fig. 4A, DI = −4.5 ± 10.5%; n = 7), nor wild-type Cbp^+/+ littermates treated with vehicle (Fig. 4B, DI = 5.8 ± 7.7%; n = 7) exhibited persistent memory for the familiar object when tested 7 days after training [F_{(1, 12)} = .849, P = 0.373]. In contrast, both Cbp^KIX/KIX homozygous mice treated with NaBut (Fig. 4A, NaBut, DI = 47.8 ± 10.8%; n = 7) and wild-type Cbp^+/+ littermates treated with NaBut (Fig. 4B, NaBut, DI = 39.6 ± 8.5%; n = 9) performed significantly better than their respective vehicle controls [Cbp^KIX/KIX, NaBut, F_{(1, 12)} = 16.17, P < 0.001); Cbp^+/+, NaBut, F_{(1, 14)} = 10.60, P < 0.005], exhibiting long-term memory for the familiar object that persists up to 7 days, but were not significantly different from each other [F_{(1, 14)} = .48, P = 0.50]. These results indicate that CBP is not required for persistent HDAC inhibition-dependent memory. Further, they demonstrate that HDAC inhibition can induce a form of memory for object recognition that persists beyond the point at which normal memory fails.

Fig. 4.

HDAC inhibition can induce a form of memory for object recognition that persists beyond the point at which normal memory fails. (A) CBP^KIX/KIX mice that received 10 min of training immediately followed by an i.p. injection of NaBut (n = 7) displayed a significant preference for the novel object compared with vehicle-treated CBP^KIX/KIX mice (n = 7) during a 7-day retention test. (B) Wild-type CBP^+/+ mice that received 10 min of training immediately followed by an i.p. injection of NaBut (n = 9) displayed a significant preference for the novel object compared with vehicle-treated wild-type CBP^+/+ mice (n = 7) during a 7-day retention test.

Discussion

One of the most important results from these experiments is that HDAC inhibition can generate a form of long-term memory that persists beyond a point at which normal memory fails. An alluring aspect of examining chromatin modifications in regulating transcription required for long-term memory processes is that these modifications, in combination with DNA methylation, may provide transient and stable markers in the service of activating and/or maintaining transcriptional profiles underlying cellular functions (1). These transcription profiles in turn may play a role in the molecular mechanisms underlying neuronal changes that subserve persistent changes in behavior. Although much more work needs to be done to elucidate the precise mechanisms involved, our results show that modulating chromatin modification may generate a persistent form of long-term memory lasting beyond the point at which normal memory fails.

A second important finding from this study is that a learning event that does not lead to short-term or long-term memory can be transformed by HDAC inhibition into an event that does result in long-term memory. Conceptually, this finding is similar to what we observed in hippocampal slices examining the effect of HDAC inhibition on synaptic plasticity in a previous study. We have shown that HDAC inhibition enhances hippocampal LTP, allowing a protocol that usually only leads to a short-term form of LTP, E-LTP, to lead to a longer lasting form of transcription-dependent LTP (18). We now show at the behavioral level that a 3-min exposure to objects fails to induce long-term memory for a familiar object. However, when an animal receives a 3-min exposure to the objects and immediately afterward an injection of an HDAC inhibitor, the animal now forms long-term memory for the familiar object. Similar results were shown in a study by Fontan-Lozano et al. (20) using a 5-min exposure length. Thus, a common theme that has emerged from studies of HDAC inhibition and synaptic plasticity and memory formation is that an event that would normally lead to a transient or transcription-independent form of plasticity or memory can result in stable, long-lasting transcription-dependent plasticity or memory when paired with HDAC inhibition.

A unique result from our experiments is that the 3-min training protocol we used does not result in short-term memory by itself or when paired with post-training HDAC inhibition. This result may indicate that a 3-min exposure to objects is sufficient to begin rapid molecular mechanisms, but without additional exposure time (e.g., 10 min), these mechanisms do not result in short-term or long-term changes in behavior. However, HDAC inhibition may generate a chromatin state that is permissive to rapid molecular mechanisms engaging transcription-dependent pathways necessary for memory consolidation. This idea may have been predicted from genetically modified mice in which other enzymes affecting histone acetylation exhibit normal short-term memory, but impaired long-term memory (3, 5–8).

Much more work will be necessary to determine how exactly HDAC inhibition modulates memory. Presumably, HDAC inhibition is acting on memory consolidation. One advantage of using HDAC inhibitors to examine their effect on memory consolidation is that they can be delivered posttraining and their effects on histone acetylation are not detectable 24 h later when a typical long-term memory test is given (see also ref. 18). Thus, it is most likely that the HDAC inhibitor effects on memory are not because of effects on performance, which is a critical factor in studies examining memory enhancements/impairments (22–24). A key open question is what effect does HDAC inhibition have on gene expression required for memory consolidation? HDAC inhibition is thought to facilitate gene expression by inducing an open chromatin configuration. But does this form of open chromatin configuration increase the level of gene expression after an activity-dependent stimulus? Or does the open chromatin configuration help maintain expression of key genes involved in the generation of long-term memory? What is the overlap between genes normally turned on/off during memory consolidation and genes turned on/off during memory consolidation in the presence of HDAC inhibition? Answers to these questions will not only give us a better understanding of how HDAC inhibitors modulate memory, but perhaps critical insight into the regulation of gene expression required for memory formation.

In a previous study, we showed that the transformation of hippocampal E-LTP into a transcription-dependent form of LTP by HDAC inhibition depended on the interaction between CBP and cyclicAMP response-element binding protein (CREB), a transcription factor (18). To test whether the same is true for the enhancement of object-recognition memory by HDAC inhibition, we used mutant mice in which the interaction between CBP and CREB is disrupted. These mice carry mutations in three highly conserved residues (Tyr650Ala, Ala654Gln, and Tyr658Ala) within the CBP KIX domain (cbp ^KIX/KIX), which is where CBP interacts with phospho-CREB (21). Cbp ^KIX/KIX homozygous mutant mice exhibit impairments in long-term memory for contextual fear conditioning and NOR (7; this study). We initially predicted that CBP would also be required for HDAC inhibition to enhance memory for NOR. Contrary to what we expected, we found that HDAC inhibition was able to ameliorate memory impairments in cbp ^KIX/KIX mutant mice. There may be several possible explanations. First, this experiment examines the relationship between the CREB:CBP interaction and HDAC inhibition in an area of the brain other than the hippocampus. In our previous study, we only examined hippocampal LTP in the cbp ^KIX/KIX mutant mice (18). The object-recognition experiments we performed in this study may be hippocampus independent because we do not alter object location or the relationship between object and context, both of which have been shown to engage the hippocampus during object recognition (13, 25, 26). Thus, HDAC inhibition-dependent enhancement in hippocampal LTP may require CBP whereas HDAC inhibition-dependent modulation of long-term memory for object recognition may not. A second possibility is that our previous findings in hippocampal slices do not extend to memory processes at the behavioral level. For example, HDAC inhibition-dependent long-term memory processes could be engaged using systems level consolidation, which is not observable in hippocampal slices. Last, the CBP deficiency in cbp ^KIX/KIX mutant mice is not complete. In mouse embryonic fibroblasts from cbp ^KIX/KIX mutant mice there is still 30% of wild-type CBP transcriptional activity present. Thus, although cbp ^KIX/KIX mutant mice used in these experiments are homozygous knockins, the CBP activity is not completely abrogated. Future experiments will be necessary to fully understand the role of CBP and other HATs in the molecular mechanisms underlying the modulation of memory formation by HDAC inhibitors.

In summary we have found that HDAC inhibition can transform a learning event that does not lead to long-term memory into an event that does, which parallels what we have observed at the cellular level with regard to synaptic plasticity (18). We have also demonstrated that HDAC inhibition can generate a form of long-term memory that is persistent and lasts beyond the time at which normal memory for object recognition fails. Future studies will reveal additional critical components of chromatin modification mechanisms involved in memory processes such as the targets of CBP and individual HDACs, nonhistone acetylation, and interactions with DNA methylation, other histone modifications, and nucleosome remodeling.

Materials and Methods

Subjects.

Male C57BL/6J mice obtained from The Jackson Laboratory were used in most experiments. The CBP^KIX/KIX homozygous knockin mice were generated as described in ref. 21. Briefly, the targeting vector for CBP contained the point mutations Tyr650Ala, Ala654Gln, and Tyr658Ala. The 3 mutations were introduced into the CBP locus of 129P2/OlaHsd-derived E14 embryonic stem cells by homologous recombination. Mice carrying the mutant allele of the KIX domain of CBP (designated CBP^KIX/KIX for homozygous knockin mice) have been bred and backcrossed in a heterozygous state on a C57BL/6 genetic background for 12 generations. Mice for experiments were generated from heterozygous matings, and wild-type littermates were used as controls. Mice were 8- to 10- weeks of age at the time of the experiment and had free access to food and water in their home cages. Lights were maintained on a 12-h light/12-h dark cycle, with all behavioral testing carried out during the light portion of the cycle. All experiments were conducted according to National Institutes of Health guidelines for animal care and use and were approved by the Institutional Animal Care and Use Committee of the University of California, Irvine. The investigator was blind to the genotype of the mice during behavioral testing.

Object Recognition.

The object recognition task consisted of a training phase and a testing phase. Before training, all mice were handled 1–2 min a day for 5 d and were habituated to the experimental apparatus 3 min a day for 3 consecutive days in the absence of objects. The experimental apparatus was a white rectangular open field (30 × 23 × 21.5 cm). During the training phase, mice were placed in the experimental apparatus with two identical objects (100-ml beakers, 1 inch circumference × 1.5 inch height; large blue Lego blocks, 1 × 1 × 2 inches) and were allowed to explore for either 3 or 10 min. The objects were thoroughly cleaned between trials to make sure no olfactory cues were present. Retention was tested at 90 min for short-term memory and 24 h for long-term memory. During these retention tests, mice explored the experimental apparatus for 5 min in the presence of 1 familiar and 1 novel object. The location of the object was counterbalanced so that one-half of the animals in each group saw the novel object on the left side of the apparatus, and the other half saw the novel object on the right side of the apparatus. A third object was used for the experiments in Fig. 3C (a small white light bulb, 1 inch circumference × 1.5 inch height).

All training and testing trials were videotaped and analyzed by individuals blind to the treatment condition and the genotype of subjects to determine the amount of time the mouse spent exploring the novel and familiar objects. A mouse was scored as exploring an object when its head was oriented toward the object within a distance of 1 cm or when the nose was touching the object. The relative exploration time was recorded and expressed by a discrimination index [D.I. = (t_novel − t_familiar)/(t_novel + t_familiar) × 100%]. Mean exploration times were calculated and the discrimination indexes between treatment groups were compared.

A different set of mice was used in each experiment unless otherwise stated. The only experiment in which the same set of mice were examined is in Fig. 3 B and C.

Delivery of HDAC Inhibitors.

For most of the experiments, mice received i.p. injections of 1.2 g/kg sodium butyrate (NaBut; Upstate) dissolved in distilled water or an equivalent volume of distilled water alone (vehicle) immediately after NOR training. We and others have used 1.2 g/kg NaBut in previous studies (17, 19). Similar results were obtained with 0.6 g/kg NaBut.

Data Analysis.

All NOR data were analyzed using 2-way ANOVAs to examine the interactions. Separate 1-way ANOVAs were used to make specific comparisons when interactions were observed. Student-Newman-Keuls posthoc tests were performed where appropriate. Simple planned comparisons were made using Student t tests with alpha levels held at 0.05. A P value within a bar in a given figure is derived from comparing testing and training, whereas a # is used to designate a P value < 0.05 comparing between treatment groups.