HDAC3 Serine 424 Phospho-mimic and Phospho-null Mutants Bidirectionally Modulate Long-Term Memory Formation and Synaptic Plasticity in the Adult and Aging Mouse Brain

Abstract

Long-term memory (LTM) formation is negatively regulated by histone deacetylase 3 (HDAC3), a transcriptional repressor. Emerging evidence suggests that posttranslational phosphorylation of HDAC3 at its serine 424 (S424) residue is critical for its deacetylase activity in transcription. However, it remains unknown if HDAC3 S424 phosphorylation regulates the ability of HDAC3 to modulate LTM formation. To examine the functionality of S424, we expressed an HDAC3-S424D phospho-mimic mutant (constitutively active form) or an HDAC3-S424A phospho-null mutant (phospho-dead form) in the dorsal hippocampus of mice. We assessed the functional consequence of these mutants on LTM formation and long-term potentiation (LTP) in young adult male mice. We also assessed whether the HDAC3-S424A mutant could ameliorate age-related deficits in LTM and LTP in aging male and female mice. Results demonstrate that young adult male mice expressing the HDAC3-S424D phospho-mimic mutant in the dorsal hippocampus exhibit significantly impaired LTM and LTP. In contrast, the HDAC3-S424A phospho-null mutant expressed in the hippocampus of young adult male mice enabled the transformation of subthreshold learning into robust LTM and enhanced LTP. Similarly, expression of the HDAC3-S424A mutant enabled LTM formation and enhanced LTP in aging male and aging female mice. Overall, these findings demonstrate that HDAC3 S424 is a pivotal residue that has the ability to bidirectionally regulate synaptic plasticity and LTM formation in the adult and aging brain.

Keywords: HDAC3; aging; epigenetics; memory; phosphorylation; synaptic plasticity.

Figure 1.

HDAC3 S424 phospho-mutants. A, Key domains of endogenous HDAC3 that influence its deacetylase activity. Green-highlighted regions represent the conserved catalytic domain of HDAC3 (∼a.a. 3–316). Orange-highlighted regions represent nonconserved domains of HDAC3 (∼a.a. 316–428) B, A hypothesized schematic of endogenous HDAC3 S424 phosphorylation dynamics on LTM formation. C–E, Representative Immunofluorescence staining against V5 tag. V5 in green. DAPI in blue. Bottom panels are magnified images of the top panel. C, HDAC3-S424D-V5 expression in the dorsal hippocampus. D, HDAC3-S424A-V5 expression in dorsal hippocampus. E, EV-V5 expression in the dorsal hippocampus (control).

Figure 2.

HDAC3-S424D expression impairs OLM performance and LTP in young adult mice. A, Experimental design. The HDAC3-S424D virus or an EV control virus was infused into the dorsal hippocampus of a cohort of 3-month-old mice (n = 16). B, Habituation over 6 d (2-way repeated–measure ANOVA; main effect of day, F_(5,80) = 24.36; p < 0.0001; no group differences in distance traveled; group, F_(1,16) = 1.479; p = 0.24; interaction, F_(5,80) = 1.208; p = 0.31). Sidak post hoc test: **p < 0.01; ***p < 0.001; ****p < 0.0001 compared with Day 1 of habituation. C, Total exploration time of both groups on training day (two-tailed, unpaired t test, t₁₄ = 0.0204; p = 0.98) D, Distribution of DI scores on training day across groups (two-tailed, unpaired t test; t₍₁₄₎ = 0.1134; p = 0.91). E, Total exploration time of both groups on test day (two-tailed, unpaired t test; t₍₁₄₎ = 0.5100; p = 0.62). F, OLM performance on test day. HDAC3-S424D mice exhibit impaired OLM performance (two-tailed, unpaired t test; t₍₁₄₎ = 2.298; *p < 0.05) compared with control mice. Both groups of mice had a significant increase in DI scores on test day than on training day (control mice, two-tailed, paired t test; t₍₆₎ = 2.956; ^#p < 0.05; HDAC3-S424D mice; two-tailed, paired t test; t₍₈₎ = 4.408; ^##p < 0.01) G, Time course of LTP expressed as the fEPSP slope as a percentage of the baseline is impaired in HDAC3-S424D slices compared with control slices (n = 24). Representative traces collected during the baseline (black line) and 50–60 min post-TBS (red line). Scale bar, 1 mv/5 ms. H, Mean potentiation 50–60 min post-TBS (two-tailed, unpaired t test; t₍₂₂₎ = 6.319; ****p < 0.0001). I–K, Baseline synaptic transmission: (I) paired-pulse facilitation (2-way ANOVA; no significant interaction between group × stimulus; F_(2,44) = 1.576; p = 0.22). J, Input/output curve. K, Slopes of the input/output curve (two-tailed, unpaired t test; t₍₂₂₎ = 0.4500; p = 0.65).

Figure 3.

HDAC3-S424A expression enhances OLM performance and LTP in young adult mice. A, Experimental design. HDAC3-S424A or an EV control virus was infused into the dorsal hippocampus of a cohort of 3-month-old mice (n = 20). A subthreshold version of the OLM task (3 min of training) was performed 2 weeks after mice recovered. B, Habituation over 6 d (2-way repeated–measure ANOVA; main effect of day, F_(5,90) = 50.64; p < 0.0001; no group differences in distance traveled; group, F_(1,18) = 0.5867; p = 0.45; interaction, F_(5,90) = 0.5985; p = 0.70). Sidak post hoc test: ****p < 0.0001 compared with Day 1 of habituation. C, Total exploration time of both groups on training day (two-tailed, unpaired t test; t₍₁₈₎ = 0.5894; p = 0.56). D, Distribution of DI scores on training day across groups (two-tailed, unpaired t test; t₍₁₈₎ = 0.6256; p = 0.54). E, Total exploration time of both groups on test day (two-tailed, unpaired t test, t₍₁₈₎ = 0.3184; p = 0.75). F, OLM performance on test day. HDAC3-S424A mice exhibited enhanced OLM performance compared with controls (two-tailed, unpaired t test; t₍₁₈₎ = 2.224; *p < 0.05). Both groups of mice had a significant increase in DI scores on test day than on training day (control mice, two-tailed, paired t test; t₍₈₎ = 3.019; ^#p < 0.05; HDAC3-S424A mice, two-tailed, paired t test; t₍₁₀₎ = 6.836; ^####p < 0.0001). G, Time course of LTP expressed as the fEPSP slope as a percentage of the baseline is enhanced in HDAC3-S424A slices compared with control slices (n = 14). Representative traces collected during the baseline (black line) and 50–60 min post-TBS (red line). Scale bar, 1 mv/5 ms. H, Mean potentiation 50–60 min post-TBS (two-tailed, unpaired t test; t₍₁₂₎ = 7.444; ****p < 0.0001). I–K, Baseline synaptic transmission: (I) paired-pulse facilitation (2-way ANOVA; no significant interaction between group × stimulus; F_(2,24) = 1.751; p = 0.20). J, Input/output curve. K, Slopes of the input/output curve (two-tailed, unpaired t test; t₍₁₂₎ = 0.6754; p = 0.51).

Figure 4.

HDAC3-S424A expression ameliorates deficits in OLM performance and LTP in aging male mice. A, Experimental design. HDAC3-S424A or an EV control virus was infused into the dorsal hippocampus of a cohort of aging male mice (18–20 months; n = 26). A subthreshold version of the OLM task (10 min of training) was adapted for the aging mice and performed 2 weeks after recovery. B, Habituation over 6 d (2-way repeated–measure ANOVA; main effect of day, F_(5,120) = 49.90; p < 0.0001; no group differences in distance traveled; group, F_(1,24) = 0.0001; p = 0.99; interaction, F_(5,120) = 0.4334; p = 0.82). Sidak post hoc test: ****p < 0.0001 compared with Day 1 of habituation. C, Total exploration time of both groups on training day (two-tailed, unpaired t test; t₍₂₄₎ = 0.1985; p = 0.84). D, Distribution of DI scores on training day across groups (two-tailed, unpaired t test; t₍₂₄₎ = 0.0193; p = 0.98). E, Total exploration time of both groups on test day (two-tailed, unpaired t test; t₍₂₄₎ = 0.4046; p = 0.69). F, OLM performance on test day. Aging male HDAC3-S424A mice exhibited enhanced OLM performance compared with controls (two-tailed, unpaired t test; t₍₂₄₎ = 3.053; **p < 0.01). Only the aging male HDAC3-S424A mice had a significant increase in DI scores on test day than on training day (aged male control mice; two-tailed, paired t test; t₍₁₀₎ = 0.8817; p = 0.40; aged male HDAC3-S424A mice, two-tailed, paired t test; t₍₁₄₎ = 6.191; ^####p < 0.0001). G, Time course of LTP expressed as the fEPSP slope as a percentage of the baseline is enhanced in HDAC3-S424A slices compared with control slices (n = 20). Representative traces collected during the baseline (black line) and 50–60 min post-TBS (red line). Scale bar, 1 mv/5 ms. H, Mean potentiation 2 min post-TBS (two-tailed, unpaired t test; t₍₁₈₎ = 2.383; *p < 0.05). I, Mean potentiation 50–60 min post-TBS (two-tailed, unpaired t test; t₍₁₈₎ = 7.155; ****p < 0.0001). J–L, Baseline synaptic transmission: (J) paired-pulse facilitation (2-way ANOVA, significant group × stimulus interaction; F_(6,80) = 2.431; *p < 0.05). K, Input/output curve. L, Slopes of the input/output curve (two-tailed, unpaired t test; t₍₁₈₎ = 0.9197; p = 0.36).

Figure 5.

HDAC3-S424A ameliorates deficits in OLM performance and LTP in aging female mice. A, Experimental design. The HDAC3-S424A or an EV control virus was infused into the dorsal hippocampus of a cohort of aging female mice (18 months; n = 19). A subthreshold version of the OLM task (10 min of training) was adapted for the aging mice and performed 2 weeks after recovery. B, Habituation over 6 d (2-way repeated–measure ANOVA; main effect of day, F_(5,85) = 31.10; p < 0.0001; no group differences in distance traveled; group, F_(1,17) = 0.4585; p = 0.51; interaction, F_(5,85) = 0.1340; p = 0.98). Sidak post hoc test: ****p < 0.0001 compared with Day 1 of habituation. C, Total exploration time of both groups on training day (two-tailed, unpaired t test; t₍₁₇₎ = 0.5196; p = 0.61). D, Distribution of DI scores on training day across groups (two-tailed, unpaired t test; t₍₁₇₎ = 0.4208; p = 0.68). E, Total exploration time of both groups on test day (two-tailed, unpaired t test; t₍₁₇₎ = 0.8951; p = 0.38). F, OLM performance on test day. Aging female HDAC3-S424A mice exhibited enhanced OLM performance compared with controls (two-tailed, unpaired t test; t₍₁₇₎= 2.679; *p < 0.05). Only the aging female HDAC3-S424A mice had a significant increase in DI scores on test day than on training day (aged female control mice, two-tailed, paired t test; t₍₉₎ = 0.4449; p = 0.67; aged female HDAC3-S424A mice: two-tailed, paired t test; t₍₈₎ = 5.025; ^##p < 0.01). G, Time course of LTP expressed as the fEPSP slope as a percentage of the baseline is enhanced in HDAC3-S424A slices compared with control slices (n = 24). Representative traces collected during the baseline (black line) and 50–60 min post-TBS (red line). Scale bar, 1 mv/5 ms. H, Mean potentiation 2 min post-TBS (two-tailed, unpaired t test; t₍₂₂₎ = 1.057; p = 0.30). I, Mean potentiation 50–60 min post-TBS (two-tailed, unpaired t test; t₍₂₂₎ = 8.385; ****p < 0.0001). J–L, Baseline synaptic transmission: (J) paired-pulse facilitation (2-way ANOVA, no significant group × stimulus interaction; F_(2,44) = 0.1384; p = 0.87). K, Input/output curve. L, Slopes of the input/output curve (two-tailed, unpaired t test; t₍₂₂₎ = 1.952; p = 0.06).

Figure 6.

HDAC3 phospho-mutants affect Per1 expression during memory consolidation for object location learning. A, Modified OLM experiment. For RT-qPCR experiments, we used a young adult male HDAC3-S424D cohort (3 months; n = 10), a young adult male HDAC3-S424A cohort (3 months; n = 9), an aged male HDAC3-S424A cohort (18 months; n = 10), and an aged female HDAC3-S424A cohort (18 months; n = 10). For ChIP-qPCR experiments, we used a separate young male HDAC3-S424D cohort (3 months; n = 10) and a separate young male HDAC3-S424A cohort (3 months; n = 9). For Western blot experiments, we used young wild-type (3 months; n = 10) and aged wild-type (17–18 months; n = 10) mice. B–E, RT-qPCR analysis of Per1 mRNA expression. B, Young adult males expressing HDAC3-S424D (two-tailed, unpaired t test, t₍₇₎ = 1.093; p = 0.31). C, Young adult males expressing HDAC3-S424A (two-tailed, unpaired t test; t₍₇₎ = 4.844; p < 0.01). D, Aged males expressing HDAC3-S424A (two-tailed, unpaired t test; t₍₈₎ = 4.876; p < 0.01). E, Aged females expressing HDAC3-S424A (two-tailed, unpaired t test; t₍₈₎ = 3.776; p < 0.01). F,G, ChIP-qPCR analysis of H3K27 acetylation (H3K27Ac IP) levels at the Per1 CRE promoter site. F, Young adult males expressing HDAC3-S424D (two-tailed, unpaired t test, t₍₈₎ = 0.3863; p = 0.70). G, Young adult males expressing HDAC3-S424A (two-tailed, unpaired t test; t₍₇₎ = 2.740; p < 0.05). H–L, Western blot analysis of endogenous levels of phosphorylated HDAC3 and total HDAC3 in young and aged male wild-type mice. H, Sample treated with λ-phosphatase abolishes phospho-HDAC3 band. I, Phosphorylated HDAC3 (two-tailed, unpaired t test; t₍₇₎ = 0.4466; p = 0.67) and total HDAC3 levels (two-tailed, unpaired t test; t₍₇₎ = 0.4792; p = 0.65) in home-cage versus OLM-trained young adult mice (3 months; n = 9). J, Phosphorylated HDAC3 (two-tailed, unpaired t test; t₍₈₎ = 0.1229; p = 0.91) and total HDAC3 levels (two-tailed, unpaired t test; t₍₈₎ = 0.4003; p = 0.70) in home-cage versus OLM-trained aged mice (17–18 months; n = 10). K, Phosphorylated HDAC3 (two-tailed, unpaired t test; t₍₇₎ = 2.185; p = 0.07) and total HDAC3 levels (two-tailed, unpaired t test; t₍₇₎ = 1.984; p = 0.09) in young (3 months) and aged home-cage (17–18 months) mice (n = 9). L, Phosphorylated HDAC3 (two-tailed, unpaired t test; t₍₇₎ = 0.3768; p = 0.72) and total HDAC3 levels (two-tailed, unpaired t test; t₍₇₎ = 0.4294; p = 0.68) in OLM-trained young (3 months) and aged (17–18 months) mice (n = 9). The “Aged OLM” sample in Lane 6 was excluded from analysis for being an outlier. Uncropped western blot images of H–L are provided in Extended Data Figures 6-1–6-3. HC, home-cage; OLM, object location memory; A, aged; Y, young.