SRC3 acetylates calmodulin in the mouse brain to regulate synaptic plasticity and fear learning

Abstract

Protein acetylation is a reversible posttranslational modification, which is regulated by lysine acetyltransferase (KAT) and lysine deacetyltransferase (KDAC). Although protein acetylation has been shown to regulate synaptic plasticity, this was mainly for histone protein acetylation. The function and regulation of nonhistone protein acetylation in synaptic plasticity and learning remain largely unknown. Calmodulin (CaM), a ubiquitous Ca²⁺ sensor, plays critical roles in synaptic plasticity such as long-term potentiation (LTP). During LTP induction, activation of NMDA receptor triggers Ca²⁺ influx, and the Ca²⁺ binds with CaM and activates calcium/calmodulin-dependent protein kinase IIα (CaMKIIα). In our previous study, we demonstrated that acetylation of CaM was important for synaptic plasticity and fear learning in mice. However, the KAT responsible for CaM acetylation is currently unknown. Here, following an HEK293 cell-based screen of candidate KATs, steroid receptor coactivator 3 (SRC3) is identified as the most active KAT for CaM. We further demonstrate that SRC3 interacts with and acetylates CaM in a Ca²⁺ and NMDA receptor-dependent manner. We also show that pharmacological inhibition or genetic downregulation of SRC3 impairs CaM acetylation, synaptic plasticity, and contextual fear learning in mice. Moreover, the effects of SRC3 inhibition on synaptic plasticity and fear learning could be rescued by 3KQ-CaM, a mutant form of CaM, which mimics acetylation. Together, these observations demonstrate that SRC3 acetylates CaM and regulates synaptic plasticity and learning in mice.

Keywords: SRC3; calmodulin; contextual fear learning; hippocampus; protein acetylation; synaptic plasticity.

Figure 1

Subcellular distribution of KAT proteins and acetylation of CaM by SRC3 in HEK293 cells.A, the family and representative proteins of KAT. B, expression of Histone H3 (H3), α-tubulin, and synapsin 1 (SYN1) in the nuclear (N), cytoplasmic (S2), and P2 fractions. The bottom image is the Ponceau staining. C, distribution of different KAT proteins in the nuclear (N), cytoplasmic (S2) and P2 fraction of mouse forebrain. Histone H3 (H3), α-tubulin, and Synapsin1 (SYN1) are markers for the N, S2, and P2 fractions. D, heat map of KAT protein levels in the nuclear (N), cytoplasmic (S2), and P2 fractions of mouse forebrain. Original western blots were shown in panel C. E, overexpression of SRC3 induces acetylation of CaM in HEK293 cells. The total lysates from HEK293 cells treated with 1 μM TSA or transfected with different KAT plasmids for 36 h were probed with the indicated Abs. F, expression of Flag-tagged KAT constructs in HEK 293 cells. The total lysates of HEK293 cells transfected with different Flag-tagged KAT constructs were probed with anti-Flag antibodies. The arrows indicate the main bands for different KAT proteins. G, SRC3 has a higher capacity to acetylate CaM than other KATs. Shown were acetylated CaM levels normalized by the intensity of Flag signal in panel F. Data were represented as mean ± SD. NS, not significant, ∗∗p = 0.0024, ∗∗∗p < 0.0001, compared with control, one-way ANOVA, n = 6, data were normalized by control.

Figure 2

Acetylation of CaM by SRC3 in a calcium-dependent manner.A, Flag-SRC3 purified from HEK293 cells could acetylate GST-WT-CaM. 3KR mutation diminished CaM acetylation, but not SRC3-CaM interaction. Flag-SRC3 was immunoprecipitated by anti-Flag-conjugated beads and then incubated with GST-tagged WT or 3KR-CaM for in vitro acetylation assay (top two lanes). In another experiment, Flag-SRC3 was immunoprecipitated by anti-Flag-conjugated beads and then incubated with GST-tagged WT or 3KR-CaM for pull-down assay (bottom two lanes). B, quantification of Ac-K/CaM in panel A. Data were represented as mean ± SD. ∗∗∗p < 0.0001, t test, n = 6, data were normalized to WT-CaM. C, quantification of CaM-SRC3 interaction as CaM/SRC3 in panel A. Data were represented as mean ± SD. NS, not significant, t test, n = 6, data were normalized to WT-CaM. D, CaM acetylation at K22, 95, and 116 by SRC3. GST-CaM was subjected to in vitro acetylation assay with Flag-SRC3 purified from HEK293 cells overexpressing Flag-SRC3, and the reaction was probed with site-specific anti-CaM antibodies. Note that the site-specific antibodies against Ac-CaM cannot recognize the CaM proteins with the K-R mutation. E, Stoichiometry levels of K22, 95, and 116 acetylated by SRC3 in vitro from panel D. Data were represented as mean ± SD. n = 6. F, the KAT domain instead of RID of SRC3 could acetylate CaM. His-tagged SRC3 HAT domain and RID were purified from bacteria and used for in vitro acetylation assay with GST-CaM. G, purified CBP KAT domain cannot acetylate GST-CaM. H, acetylation of CaM by the KAT domain of SRC3 is Ca²⁺-dependent. All proteins were purified from bacteria. Acetylation was performed in the presence of 0.1 mM Ca²⁺ or 1 mM EGTA.

Figure 3

CaM-SRC3 interaction in a Ca²⁺and NMDAR-dependent manner.A, increased CaM-SRC3 interaction by Ca²⁺. His-SRC3-KAT and GST-CaM were purified from bacteria and their interaction was investigated by GST pull-down during which acetylation could not occur. B, quantification of CaM-SRC3 interaction as His/GST in panel A. Data were represented as mean ± SD. ∗∗∗p < 0.0001, one-way ANOVA, n = 6, data were normalized to 0.01 mM Ca²⁺. C, the binding curve of GST-CaM and His-SRC3-KAT under different Ca²⁺ concentrations. The K_d of Ca²⁺ for the interaction between GST-CaM and His-SRC3-KAT is 2.734 μM, n = 3. Data were represented as mean ± SD. D, CaM-SRC3 interaction in HEK293 cells depends on Ca²⁺. The immunoprecipitation (IP) of Flag-SRC3 was probed with anti-CaM antibodies. The CaM-SRC3 interaction was abolished when the IP buffer contained 1 mM EGTA. E, increased CaM-SRC3 interaction by cLTP, which is dependent on NMDAR. The CaM-SRC3 interaction was assessed by coimmunoprecipitation (co-IP) from total lysate of hippocampal slices with different treatment. The co-IP was performed in RIPA buffer without EGTA or detergent. F, quantification of CaM-SRC3 interaction as CaM/SRC3 in panel E. Data were represented as mean ± SD. ∗∗∗p < 0.0001, one-way ANOVA, n = 6, data were normalized to control.

Figure 4

Impaired CaM acetylation and hippocampal LTP by SRC3 inhibition.A, increased Ac-CaM, p-CaMKIIα, and p-GluR1 after cLTP stimulation were attenuated by 1 μM SI-2. Total lysate of hippocampal slices after different treatment was probed for Ac-CaM, CaM, p-CaMKIIα, CaMKIIα, p-GluR1, and GluR1. B–D, quantification of Ac-CaM/CaM (B), p-CaMKIIα/CaMKIIα (C), and p-GluR1/GluR1 (D) in panel A. Data were represented as mean ± SD. ∗∗∗p < 0.0001, two-way ANOVA followed by Tukey’s multiple comparisons test, n = 6, data were normalized to control. E, diagram showing field EPSP recording at SC-CA1 synapses. WT hippocampal slices were treated with vehicle or SI-2. F, normalized fEPSP amplitudes were plotted every 1 min for hippocampal slices treated with vehicle or different concentrations (0.1, 0.3, 1 μM) of SI-2, which was applied during the period indicated by the bar. G and H, SI-2 attenuated PTP (G) and LTP (H) in a dose-dependent manner. Data in panel F were quantified. Data were represented as mean ± SD. NS, not significant, ∗p = 0.0187, ∗∗p = 0.0062, ∗∗∗p = 0.001, compared with Veh, one-way ANOVA, n = 11 slices from five mice for Veh, n = 9 slices from four mice for other groups. Veh: vehicle. I, SI-2 (1 μM) did not alter input–output (I/O) curves at SC-CA1 synapses. Data were represented as mean ± SD. F (1,16) = 3.736, p = 0.0711, two-way-ANOVA, n = 9 slices from four mice for each group. J, SI-2 (1 μM) did not affect paired pulse facilitation (PPF) at SC-CA1 synapses. Data were represented as mean ± SD. F (1,22) = 1.975, p = 0.1738, two-way-ANOVA, n = 12 slices from four mice for each group. K, normalized fEPSP amplitudes were plotted every 1 min for hippocampal slices treated with vehicle or 1 μM SI-2, which was applied during the period indicated by the bar. L and M, SI-2 (1 μM) inhibited LTP induction but not maintenance. Data of PTP (L) and LTP (M) in panel K were quantified. Data were represented as mean ± SD. NS, not significant, ∗∗p = 0.0093, ∗∗∗p = 0.0004, compared with Veh, one-way ANOVA, n = 8 slices from four mice for each group. Veh: vehicle, Pre: before TBS, Post: after TBS.

Figure 5

SRC3 regulates hippocampal LTP through acetylation of CaM.A, diagram showing whole-cell recording of eEPSC in CA1 pyramidal neurons of mouse hippocampus. Recording pipettes were infused with GST-tagged WT or 3KQ-CaM proteins. B, the input resistance of CA1 pyramidal neurons was not affected by SI-2 (1 μM) treatment. Data were represented as mean ± SD. NS, not significant, compared with WT-CaM, one-way ANOVA, n = 10 cells from ten mice for WT-CaM and SI-2 + 3KQ-CaM, n = 6 cells from six mice for SI-2 + WT-CaM. C, normalized eEPSC amplitudes were plotted every 1 min for CA1 pyramidal neurons infused with 100 nM GST-WT-CaM or GST-3KQ-CaM proteins with or without SI-2 treatment. D and E, 3KQ-CaM rescued the deficit of PTP (D) and LTP (E) after SRC3 inhibition. Data in panel C were quantified. Data were represented as mean ± SD. NS, not significant, ∗p = 0.0346, ∗∗p = 0.0036, compared with GST-WT-CaM, one-way ANOVA, n = 10 cells from ten mice for WT-CaM and SI-2 + 3KQ-CaM, n = 6 cells from six mice for SI-2 + WT-CaM.

Figure 6

Attenuated CaM acetylation and hippocampal LTP by SRC3 downregulation.A, top, schematic diagram of AAV-Src3 gRNA-EGFP. U6, human U6 Polymerase III promoter, CMV, human cytomegalovirus promoter, bottom, diagram showing the stereotaxic injection of AAV-Src3 gRNA-EGFP into the hippocampus of Rosa26-Cas9 knockin mice. B, experimental design. C, expression of EGFP in the hippocampus 3 weeks after AAV injection. Scale bar, 400 μm. D, reduced protein levels of SRC3 in Rosa26-Cas9 hippocampus injected with Src3 gRNA AAV. The hippocampal lysates were probed with the indicated Abs. E, quantification of SRC3/α-tubulin in panel D. Data were represented as mean ± SD. ∗∗∗p < 0.0001, t test, n = 6, data were normalized to control gRNA. F, increased Ac-CaM, p-CaMKIIα after cLTP stimulation were attenuated by SRC3 knockdown (KD). Total lysate of hippocampal slices after different treatment were probed for Ac-CaM, CaM, p-CaMKIIα, and CaMKIIα. G and H, quantification of Ac-CaM/CaM (G) and p-CaMKIIα/CaMKIIα (H) in panel F. Data were represented as mean ± SD. NS, not significant, ∗∗∗p < 0.0001, two-way ANOVA followed by Tukey’s multiple comparisons test, n = 6, data were normalized to control. I, diagram showing field EPSP recording at SC-CA1 synapses. Rosa26-Cas9 mice were injected with AAV expressing control or Src3 gRNA and EGFP. J, normalized fEPSP amplitudes were plotted every 1 min for Rosa26-Cas9 hippocampal slices injected with indicated AAV. K and L, reduced PTP (K) and LTP (L) in mice injected with Src3 gRNA AAV. Data in panel J were quantified. Data were represented as mean ± SD. ∗∗∗p < 0.0001, t test, n = 8 slices from four mice for each group.

Figure 7

Importance of SRC3 in contextual fear learning.A, increased Ac-CaM, p-CaMKIIα, and p-GluR1 after contextual fear conditioning were attenuated by SI-2. Total lysates of hippocampus from different groups of mice were probed for Ac-CaM, CaM, p-CaMKIIα, CaMKIIα, p-GluR1, and GluR1. B–D, quantification of Ac-CaM/CaM (B), p-CaMKIIα/CaMKIIα (C), and p-GluR1/GluR1 (D) in panel A. Data were represented as mean ± SD. ∗∗∗p = 0.0003 for column A and B, ∗∗∗p = 0.0005 for column B and D in panel B, ∗∗p = 0.0085, ∗p = 0.013 in panel C, ∗p = 0.0118, ∗∗p = 0.0094 in panel D, two-way ANOVA followed by Tukey’s multiple comparisons test, n = 3, data were normalized to control. E, reduced freezing during paired tests by SI-2 treatment 1 h before training. SI-2 had no effect on freezing when it was delivered 1 h before paired tests. Data were represented as mean ± SD. NS, not significant, ∗∗∗p = 0.0007, two-way ANOVA followed by Dunnett’s multiple comparisons test, n = 16 for group a (control), n = 16 for group b (formation), n =12 for group c (retrieval). Top, schematic diagram of behavioral tests. F, reduced freezing during paired tests in Rosa26-Cas9 mice injected with AAV expressing Src3 gRNA. Data were represented as mean ± SD. ∗∗∗p = 0.0008, two-way ANOVA followed by Sidak’s multiple comparisons test, n = 14 for each group. Top, schematic diagram of behavioral tests. G, 3KQ-CaM partially rescued fear learning deficits by SI-2. The mice receiving lentivirus expressing CaMKD+WT-CaM or CaMKD+3KQ-CaM were treated with SI-2 1 h before training. Data were represented as mean ± SD. NS, not significant, ∗∗∗p < 0.0001, two-way ANOVA followed by Dunnett’s multiple comparisons test, compared with control, n = 13 for each group. Top, schematic diagram of behavioral tests.

Figure 8

No effects of SRC3 inhibition or downregulation on locomotion.A, equal distance traveled during the first 30 min in open field for mice treated with vehicle or SI-2. Data were represented as mean ± SD. Genotype F (1,22) = 0.3879, p = 0.5398, two-way ANOVA, n = 12 for each group. B, similar time staying in the center and margin of the open field during the first 5 min for mice treated with vehicle or SI-2. Data were represented as mean ± SD. NS, not significant, p = 0.5613, two-way ANOVA followed by Sidak’s multiple comparisons test, n = 12 for each group. C, equal distance traveled during the first 30 min in open field for Rosa26-Cas9 mice receiving hippocampal injection of the indicated AAV. Data were represented as mean ± SD. NS, not significant, Genotype F (1,22) = 0.9887, p = 0.3309, two-way ANOVA, n = 12 for each group. D, similar time staying in the center and margin of the open field during the first 5 min for Rosa26-Cas9 mice receiving hippocampal injection of the indicated AAV. Data were represented as mean ± SD. NS, not significant, p = 0.9811, two-way ANOVA followed by Sidak’s multiple comparisons test, n = 12 for each group.