Estrogen receptor α promotes Cav1.2 ubiquitination and degradation in neuronal cells and in APP/PS1 mice

Abstract Cav1.2 is the pore‐forming subunit of L‐type voltage‐gated calcium channel (LTCC) that plays an important role in calcium overload and cell death in Alzheimer's disease. LTCC activity can be regulated by estrogen, a sex steroid hormone that is neuroprotective. Here, we investigated the potential mechanisms in estrogen‐mediated regulation of Cav1.2 protein. We found that in cultured primary neurons, 17β‐estradiol (E2) reduced Cav1.2 protein through estrogen receptor α (ERα). This effect was offset by a proteasomal inhibitor MG132, indicating that ubiquitin–proteasome system was involved. Consistently, the ubiquitin (UB) mutant at lysine 29 (K29R) or the K29‐deubiquitinating enzyme TRAF‐binding protein domain (TRABID) attenuated the effect of ERα on Cav1.2. We further identified that the E3 ligase Mdm2 (double minute 2 protein) and the PEST sequence in Cav1.2 protein played a role, as Mdm2 overexpression and the membrane‐permeable PEST peptides prevented ERα‐mediated Cav1.2 reduction, and Mdm2 overexpression led to the reduced Cav1.2 protein and the increased colocalization of Cav1.2 with ubiquitin in cortical neurons in vivo. In ovariectomized (OVX) APP/PS1 mice, administration of ERα agonist PPT reduced cerebral Cav1.2 protein, increased Cav1.2 ubiquitination, and improved cognitive performances. Taken together, ERα‐induced Cav1.2 degradation involved K29‐linked UB chains and the E3 ligase Mdm2, which might play a role in cognitive improvement in OVX APP/PS1 mice.


| INTRODUC TI ON
Cav1.2 is the pore-forming subunit of L-type voltage-gated calcium channels (LTCC) and accounts for approximately 70% of LTCC in the brain (Zamponi, Striessnig, Koschak, & Dolphin, 2015). Cav1.2 is mainly located at extra-and postsynaptic sites and plays a critical role in intracellular calcium transients (Tippens et al., 2008). During aging, LTCC activity is increased, and calcium influx via LTCC may be the preferred source for calcium-induced calcium release (CICR), which plays a necessary role in aging-related biomarkers (Thibault, Gant, & Landfield, 2007). Evidence has suggested that dysfunction of LTCC and calcium overload contribute to cell death and the pathophysiology of Alzheimer's disease (AD) (Bezprozvanny & Mattson, 2008). Accordingly, LTCC inhibitors have exhibited therapeutic potential in AD (Nimmrich & Eckert, 2013).
Among the multiple mechanisms underlying estrogen action in the brain (Nilsson et al., 2001), the involvement of LTCC-mediated calcium dynamics has been documented (Vega-Vela et al., 2016). Brief exposure (90 s) of 17β-estradiol (E2) induces rapid increase in LTCC currents, which is mediated by direct interaction of E2 with LTCC subunit (Sarkar et al., 2008). Interestingly, estrogen also induces rapid release of calcium from intracellular store within minutes in isolated neurons, possibly by mechanism of CICR (Beyer & Raab, 1998;Thibault et al., 2007). On the other hand, long-term estrogens can inhibit LTCC in neuronal cells. For instance, E2 (25 hr) inhibits the high-voltage-activated calcium channel currents, which are associated with LTCC (Kumar & Foster, 2002). In glutamate-or high potassium-primed neurons, E2 (5 min ~ 25 hr) reduces calcium entry and cell death through LTCC (Kurata, Takebayashi, Kagaya, Morinobu, & Yamawaki, 2001;Sribnick, Re, Ray, Woodward, & Banik, 2009). Chronic E2 replacement prevents the age-related increase of LTCC currents in the hippocampus of ovariectomized (OVX) rats (Brewer et al., 2009). Therefore, identifying the molecular mechanisms underlying estrogen regulation of Cav1.2 may aid to the understanding and treatment of AD.
In the current study, we found that E2 reduced Cav1.2 protein through estrogen receptor α (ERα). This effect involved lysine 29linked ubiquitin chains and the E3 ligase Mdm2 (double minute 2 protein). In OVX APP/PS1 mice, systematic administration of E2 and ERα agonist PPT (propylpyrazoletriol) led to the reduced Cav1.2 protein and the enhanced Cav1.2 ubiquitination in the brain, which were accompanied by the improved cognitive functions.
Cav1.2 protein level was decreased at 6 hr and remained to be decreased for up to 48 hr in primary neurons incubated with E2. This effect seemed to be dependent on estrogen receptors (ERs), as a nonselective ER antagonist fulvestrant (ICI182780, ICI, 1 μM) diminished the effect of E2 on Cav1.2 ( Figure 1c). Similar results were found by immunofluorescent studies (Figure 1d). To further assess whether E2 may affect Cav1.2 function, the voltage-gated calcium currents (I Ca 2+ density) were measured using whole-cell patch recordings. As shown in Figure 1e, the peak I Ca 2+ density (pA/pF) was significantly decreased by E2 treatment (24 hr ERα or ERβ mRNA showed that ERα shRNA2 and ERβ shRNA3 were most effective in reducing ERα or ERβ protein, respectively ( Figure   S1). As expected, ERα shRNA2 significantly increased Cav1.2, which attenuated E2-induced reduction of Cav1.2 ( Figure 1h). In contrast, ERβ shRNA3 did not alter the basal Cav1.2 and failed to further attenuate Cav1.2 reduction by E2 (Figure 1i). These results indicated that ERα mediated E2 reduction of Cav1.2.
Moreover, ERα agonist PPT also failed to alter Cav1.2 phosphorylation. And the level of the calcineurin (PP2B, CN), the phosphatase involved in Cav1.2 dephosphorylation (Oliveria, Dell'Acqua, & Sather, 2007), was not altered ( Figure S2C and D). Although PPT enhanced CN association with Cav1.2, CN inhibitor FK506 did not prevent PPT effect on Cav1.2 in primary neurons ( Figure S2E and F). Thus, it was also unlikely that the phosphorylation-related mechanisms are involved in ERα-mediated Cav1.2 reduction.

| PPT treatment reduced Cav1.2 protein in OVX APP/PS1 mice
As shown in Figure 5a These results indicated that ERα activation promoted the ubiquitination and degradation of Cav1.2 in OVX APP/PS1 mice.

| PPT treatment attenuated cognitive decline in APP/PS1 mice
We next assessed the effect of E2 or ER agonists on spatial and associative learning and memory performances in OVX APP/PS1 mice.
In the hidden platform test, the escape latency and total traveling distance in E2-or PPT-treated OVX APP/PS1 (AD + E2 or AD + PPT) mice were significantly shorter than that in saline-treated OVX APP/ PS1 (AD) mice beginning on the third day (Figure 6a-c). In the probe trial when the platform was removed, the passing times crossing over

| D ISCUSS I ON
The efficacy of estrogen-based hormone therapy (HT) in AD is not conclusive. Some studies reveal that HT may reduce the risk of AD in postmenopausal women (Zandi et al., 2002), whereas others find that HT does not reduce dementia or even has an adverse effect (Shumaker et al., 2004). One explanation is that compared with the natural E2, the conjugated equine estrogen used in clinics does not easily diffuse into the brain (Lan, Zhao, & Li, 2015;Steingold, Cefalu, Pardridge, Judd, & Chaudhuri, 1986). Another explanation is that the initiation of the therapy may have been delayed (Sherwin, 2005). Given that LTCC blockers attenuate AD-associated pathology, E2-induced reduction of Cav1.2 in our study favors the beneficial role in AD.
Evidence has suggested that UPS is involved in Cav1. The seven lysine residues in UB play distinct roles in cellular function (Swatek & Komander, 2016). While K29-linked chains and K29-specific deubiquitinating enzyme TRABID are involved in proteasome regulation (Kim et al., 2011;Kristariyanto et al., 2015), K6linked UB chains are associated with mitophagy but with no known functions in UPS (Swatek & Komander, 2016 Figure 2e). Interestingly, only K29 seems to be critical in ERα regulation of Cav1.2, as PPT-induced reduction of Cav1.2 is diminished only in K29R but not in K6R-overexpressing cells. In line with this, the deubiquitinating enzyme TRABID that selectively acts on K29 in mammalian cells also diminishes PPT effect (Figure 2i).
In physiological conditions, both ERα and ERβ seem to be involved in spatial memory enhancement (Walf, Paris, & Frye, 2009).
We propose possible mechanisms through which ERα reduces

| Primary culture of cortical neurons
Cortical neurons were extracted from prenatal pups on embryonic day 18. Cells were placed on poly-L-lysine (0.1 mg/ml)-coated 6-well plates, at a density of 2 × 10 6 cells/ml for biochemical experiments and 0.5 × 10 6 cells/ml for molecular and immunohistochemistry experiments. Cultures were maintained in serum-free neurobasal (NB) medium (2% B27, 2 mM glutamine, 1% penicillin/streptomycin; Invitrogen) at 37°C with 5% CO 2 . Fifty percent of the medium was exchanged with fresh medium every other day for 2 weeks. On the day before experiment, culture medium was completely replaced by fresh medium in the presence of accurate dilution of chemicals.

| Western blotting and co-immunoprecipitation
Western blotting of Cav1.2 was performed as previously described (He et al., 2016).

| Whole-cell patch-clamp recording
Voltage-dependent calcium-mediated currents (ICa 2+ ) were recorded from cultured cortical neurons from 7 to 11 days in vitro (DIV), at room temperature using a MultiClamp 700B amplifier (Axon Instruments).

| In utero electroporation and related immunofluorescence
Briefly, in utero electroporation was performed on pregnant C57/ B6L mice on embryonic day 14.5 (E14.5) as previously described (Saito, 2006). The vector and Mdm2 plasmids were mixed with the Venus-GFP with a ratio of 3:1 at a final concentration of 1,000 ng/ μl. The mixture was injected into lateral ventricle. The electrode (CUY650P5, NEPA GENE) was positioned flanking the ventricular zones with the anode on the target side, and the embryos were then pulsed with a super electroporator NEPA21 type II (NEPA GENE) according to the following procedure: voltage = 33 V, pulse length = 50 ms, pulse interval = 950 ms, pulse number = 3.
The uterus was then returned to the abdominal cavity for further development. After birth, the brains were perfused with PBS on P0 and were fixed in 4% PFA (Sigma) overnight at 4°C, and then were replaced with 20% sucrose for 8 hr, followed by 30% sucrose at 4°C for an additional 8 hr. The brains were cut into coronal slices using a cryostat microtome (CM1950, Leica) and subjected to immunofluorescence staining. Slices were permeabilized and blocked in 10% goat serum, 3% BSA, and 0.3% Triton X-100 in PBS. Brain sections were then incubated with primary antibody (Cav1.2 1:100, Ubiquitin 1:100) for 1 hr at 37°C and then with the fluorescent-conjugated secondary antibodies (goat anti-mouse Alexa 647 1:100, goat anti-rabbit Cy3 1:100) for 1 hr at 37°C.
Confocal images were acquired and analyzed with the ZEISS system (LSM800). All animal studies were approved by the animal care committee of Chongqing Medical University. Image assay was performed using ImagePro Plus 6.0.
Animals were subcutaneously injected daily with vehicle, 17β-estradiol (E2, 30 μg/kg), PPT (1 mg/kg), and DPN (1 mg/kg) for two weeks. The dosage of E2 is thought to be within physiological range; and those of PPT and DPN are functional (Avtanski et al., 2014).

| Morris water maze and contextual and cued fear conditioning
The Morris water maze test included four platform trials per day for five consecutive days and a probe trial on the sixth day. Swimming activities (latency, distance, and strategy of search) were recorded by a video and analyzed by image analyzing software (ANY-maze; Stoelting). For contextual and cued fear conditioning, mice were tested in a 3-day paradigm. Behavior was recorded by video camera, and freezing data were measured using FreezeScan software.

| Statistical analyses
Data were presented as means ± SEM from at least three independent experiments. The statistical comparisons between two groups were tested using Student's t test. The comparisons among groups were tested using one-way or two-way ANOVA, and a post hoc pairwise comparison was used where it applied.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R ' S CO NTR I B UTI O N S
G-J Chen and Z Yan designed the research; Y-J Lai performed the research and analyzed the data; B-L Zhu, F Sun, D Luo, Y-L Ma, B-Luo, J Tang, M-J Xiong, L Liu, X-T Hu, L He, X-J Deng, J-H Zhang, and J Yang provided assistance with the research; G-J Chen and Y-J Lai wrote the paper. All authors read and approved the final manuscript.