Protein phosphatase 1 regulates reactive oxygen species-dependent degradation of histone deacetylase 5 by intermittent hypoxia

Abstract

We recently reported pheochromocytoma 12 (PC12) cells and rats subjected to intermittent hypoxia (IH), a hallmark manifestation of obstructive sleep apnea (OSA), exhibit reduced histone deacetylase activity and HDAC5 protein. Our study further suggested that posttranslational modifications rather than transcriptional mechanism(s) mediate IH-induced HDAC5 degradation. These observations prompted our current study to investigate the mechanism(s) underlying HDAC5 degradation by IH in PC12 cell cultures. IH-induced HDAC5 degradation was blocked by an antioxidant, and reactive oxygen species (ROS) mimetics decreased HDAC5 protein, suggesting that ROS mediates HDAC5 degradation by IH. NADPH oxidases (NOX) 2 and 4 were identified as sources of ROS that mediate the effects of IH. HDAC5 degradation during IH was associated with dephosphorylation of HDAC5 at serine²⁵⁹, and this response was blocked by a NOX inhibitor, suggesting that ROS-dependent dephosphorylation mediates HDAC5 degradation. IH-induced dephosphorylation of HDCA5 was inhibited by calyculin A, an inhibitor of protein phosphatase (PP)-1 and -2, or by the overexpression of nuclear inhibitor of PP1 (NIPP1). HDAC5 dephosphorylation by IH lead to augmented hypoxia-inducible factor (HIF)-1α protein and an increase in its transcriptional activity. These data suggest that PP1-dependent dephosphorylation of S²⁵⁹ destabilizes HDAC5 protein in response to IH, resulting in HIF-1α stabilization and transcriptional activity. Our findings highlight hither to unexplored role of protein phosphatases, especially PP1 in regulating HDAC5 protein, which is an upstream activator of HIF-1 signaling by IH.

INTRODUCTION

Histone acetylation is a major epigenetic regulator of gene expression. Acetylation status of histones is dictated by a balance between histone acetylases (HATs) and deacetylases (HDACs) (1). HDACs have been implicated in different pathologies (2, 3). For instance, HDAC5, a member of the class IIa HDAC family, functions as a transcriptional repressor in cardiac hypertrophy, skeletal muscle differentiation, and angiogenesis (4–6). Obstructive sleep apnea (OSA) is a highly prevalent breathing disorder, which is characterized by brief repetitive stoppage of respiration during sleep (7). Intermittent hypoxia (IH) is a hallmark manifestation of OSA (8). We recently reported that pheochromocytoma (PC-12) cells and adult rats treated with IH, patterned after blood O₂ profiles during OSA, manifest reduced HDAC enzyme activity and HDAC3 and HDAC5 proteins (9). The IH-induced reduction of HDAC5, but not HDAC3 protein, was associated with increased lysine acetylation of the hypoxia-inducible factor (HIF)-1α subunit, elevated HIF-1α protein, and HIF-1-dependent transcription (9). It was further shown that HIF-1 transcriptional activation contributes to sympathetic activation and hypertension in rodent models of IH (9). However, the mechanism(s) underlying HDAC5 degradation by IH are not known.

Our earlier study showed that IH had no effect on HDAC5 mRNA levels, indicating that posttranslational modification(s), rather than transcriptional mechanisms, contribute to HDAC5 protein degradation by IH (9). Altered phosphorylation of highly conserved serine residues (S²⁵⁹, S²⁷⁹, and S⁴⁹⁸) in the N-terminal domain of HDAC5 is a key posttranslational mechanism(s) regulating HDAC5 activity and stabilization (10, 11). Protein kinases involved in the phosphorylation of HDAC5 include calcium/calmodulin-dependent protein kinases (CAMKs) (12, 13), protein kinase D (PKD) (14–16), protein kinase C (PKC) (17), and protein kinase A (PKA) (18). IH is a potent activator of CAMK, PKA, and PKC (19–21). Reactive oxygen species (ROS) activate protein kinases (21, 22). ROS signaling, especially involving H₂O₂, mediates cellular and systemic responses to IH (23). Given that IH increases ROS levels, we tested the hypothesis that ROS, by activating either CAMK or other protein kinases, mediate reduced HDAC5 protein and activity by IH in PC12 cell cultures.

METHODS

Exposure of PC12 Cells to IH

PC12 cells (RRID:CVCL_0481) (original clone from L Green, Columbia University, NY) (24) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 5% fetal bovine serum (FBS), 10% horse serum under 10% CO₂, and 90% room air (20% O₂) at 37°C (19). Experiments were performed on cells serum starved overnight in serum-free DMEM medium. Cells were exposed to 60 cycles of IH [each cycle consisting of 1.5% O₂ for 15 s (at nadir) followed by 20% O₂ for 5 min at 37°C as described previously (25)]. Ambient O₂ levels in the IH chamber were monitored by an O₂ analyzer (Alpha Omega Instruments, Huston, TX).

Measurement of HDAC Activity

HDAC activity was measured in nuclear lysates using Epigenase HDAC colorimetric activity kit (sensitivity 0.5 ng of purified enzyme) (Epigentek, Farmington, NY). Briefly, 5 μg of nuclear extract was added to the plate coated with acetylated histone HDAC substrate. Active HDAC binds and deacetylates H3, which is recognized with a specific antibody and measured by reading the absorbance at 450 nm. The activity is directly proportional to the absorbance measured and expressed as nanogram/minute/milligram protein.

Immunoblot Assays

PC12 cell extracts were prepared in lysis buffer (phosphate buffer, pH 7.4 containing 150 mM NaCl, 1% triton X-100, 1% sodium deoxycholate, 0.1% SDS, 5 mM EDTA, and protease inhibitor cocktail). Nuclear extracts of PC12 cells were prepared using nuclear extraction kit (Active Motif, Carlsbad, CA). Briefly, cells (1 × 10⁶) were homogenized in 200 µL of hypotonic buffer provided in the kit and centrifuged at 850 g for 10 min at 4°C. The pellet was resuspended in 200 µL of hypotonic buffer and small sample is checked under the microscope to verify that cells have been efficiently lysed and that nuclei have been released. The suspension is centrifuged at 14,000 g for 30 s. The nuclear pellet was resuspended in 40 µL of complete lysis buffer provided. Cell and nuclear lysates were analyzed by polyacrylamide-SDS gel electrophoresis. Immunoblots were probed with primary antibodies followed by corresponding horseradish peroxidase (HRP)-conjugated secondary antibody detected by Clarity Western ECL substrate kit (Bio-Rad, Hercules, CA). Immunoblots were scanned and quantified using an Odyssey Fc imaging system (LI-COR, Lincoln, NE).

Transient Transfections and HRE Reporter Gene Assay

PC12 cells were transiently transfected with overexpression plasmids using TransIT-2020 transfection reagent (Mirus Bio LLC, Madison, WI). Plasmids HDAC5, pHDAC5S259, pNIPP1, pRSV-CAMKII, and pRSV-CAMKII (1-290) were purchased from Addgene (Watertown, MA). Briefly, cells were plated in 60-mm tissue culture plates at a density of 2 × 10⁶ cells/plate in serum-containing growth medium. After 24 h, DNA-TransIT reagent in the ratio of 1:3 was added. For genetic silencing with siRNA, cells were transfected with NOX2 siRNA (sc-61838), NOX4 siRNA (sc-61887), or nontargeting control siRNA (sc-37007) from Santacruz using TransIT-X2 transfection reagent (# MIR6000, Mirus Bio LLC, Madison, WI). Cells were serum starved after 36 h and exposed to IH. To analyze HIF-1 transcriptional activity, cells were cotransfected with two reporter plasmids: p2.1, which contains firefly luciferase coding sequences downstream of a basal SV40 promoter and a 68 bp hypoxia response element (HRE) from the human Eno1 gene (26), and Renilla luciferase control reporter vector (pRL). The Bright-Glo Luciferase Assay System (Promega, Madison, WI) was used to measure luciferase activity in cell lysates and normalized to protein content measured by Bio-Rad protein assay kit (Bio-Rad, Richmond, CA).

Chemicals and Antibodies

Xanthine, xanthine oxidase, tert-butyl hydroperoxide (TBH), allopurinol, KN93, bisindolylmaleimide I (BIS), PK1 (14-22) amide, phosphatase inhibitor cocktail (PIC), p-bromotetramisole oxalate (BTO), and cantharidin were purchased from Sigma. MnTmPyP and VAS2870 were purchased from ENZO, and calyculin A was purchased from Cell Signaling. Stock solutions were prepared fresh before the experiment, and the cells were preincubated 30 min before and during IH exposure with either drug or vehicle. The following antibodies were used to detect protein expression levels by immunoblots: HIF-1α (1/1,000 dilution; NB100-123, RRID:AB_2232953, Novus Biologicals, Centennial, CO), TBP (1/1,000 dilution; Tata binding protein; #ab51841, RRID:AB_945758, Abcam), tubulin (1/10,000; #T6199, RRID:AB_477583, Sigma, St. Louis, MO); and HDAC5 (# 20458, RRID:AB_2713973, 1/500 dilution), HDAC5S259 (#3443, RRID:AB_2118723, 1/500 dilution), PP1α/γ (1/100 dilution #2582, RRID:AB_330822, PP2A-A (#2041, RRID:AB_2168121, 1/1,000 dilution), PP2A-B #2290, RRID:AB_659890, 1/1,000 dilution), PP2A-C (#2259, RRID:AB_561239, 1/1,000 dilution), and green fluorescent protein (GFP) (#2955, RRID:AB_1196614, 1/1,000 dilution) were from Cell Signaling Technology (Danvers, MA).

Statistical Analysis

Data are expressed as means ± SE from 3 to 5 independent cell culture experiments. Statistical analysis was performed by analysis of variance (ANOVA). The Mann–Whitney test was used for analysis of normalized data. P values <0.05 were considered significant.

RESULTS

ROS Mediate Reduced HDAC 5 Protein by IH

IH increases ROS generation in rodents (27–30) and cell cultures (8) as well as in humans experiencing IH as a consequence of OSA (31–34). Given that ROS are potent posttranslational regulators of proteins (22), we hypothesized that ROS mediate degradation of HDAC5 protein by IH. If, our hypothesis is correct, then, scavenging ROS with an antioxidant should prevent IH-induced reduction of HDAC5 protein. To assess this possibility, PC12 cells exposed to 60 cycles of IH (IH₆₀) were treated with MnTmPyP (Mn(III)tetrakis(1-methyl-4-pyridyl) porphyrin), a membrane-permeable antioxidant. As reported previously (9), IH₆₀ reduced HDAC5 protein, and this response was absent in the presence of 50 μM MnTmPyP (Fig. 1A). Conversely, treating control room air cultured cells with xanthine (250 ug) and increasing concentrations of xanthine oxide (XO, 0.01–0.1 U/mL), which increase ROS or with tert-butyl hydroperoxide (TBH, 10–100 uM), an established ROS mimetic, decreased HDAC5 protein akin to IH (Fig. 1, B and C). These results demonstrate that ROS contribute to HDAC5 protein degradation by IH.

Figure 1.

ROS contribute to HDAC5 protein degradation by IH. A, top: HDAC5 protein expression in room air (N) and IH₆₀ exposed cells treated with vehicle or MnTmPyP (50 µM). A, bottom: Quantitative analysis of the blots using Image studio by Odyssey Fc imaging system (LI-COR, Lincoln, NE) (n = 4). HDAC protein expression was normalized to tubulin and expressed relative to vehicle treated control cells exposed to room air (N). B and C: representative immunoblots (top) and quantitative analysis of the blots (bottom) of HDAC5 protein expression in PC12 cells treated with different concentrations of ROS mimetics [xanthine/xanthine oxidase (XO)] and tertiary butyl hydroperoxide (TBH) (means ± SE; n = 4). **P ≤ 0.05; ns = not significant (P > 0.05) as determined by Mann–Whitney test. ROS, reactive oxygen species.

NOX-Derived ROS Contribute to the Effects of IH on HDAC5

We next determined the source of ROS which mediate the effects of IH on HDAC5. Our earlier studies identified xanthine oxidase (XO) and NADPH oxidase (NOX) as major sources of ROS generation by IH (35–38). To determine which of these two oxidases contribute to HDAC5 degradation by IH, cells were treated with 50 µM allopurinol, an established inhibitor of XO, or 20 µM VAS2870, a pan inhibitor of NOX isoforms, and were then exposed to IH. VAS2870, but not allopurinol prevented HDAC5 degradation by IH (Fig. 2, A and B).

Figure 2.

ROS generated by NADPH oxidases (NOX2 and NOX4) regulate HDAC5 expression in response to IH. A: representative immunoblot showing HDAC5 protein expression and tubulin as loading control in cell lysates from room air (N) and IH₆₀ exposed PC12 cells treated with allopurinol (XO inhibitor; 50 µM) or VAS2870 (NOX inhibitor; 20 µM). B: quantitative analysis of the blots (means ± SE; n = 3) using Image Studio by Odyssey Fc imaging system. C: HDAC5 protein expression in lysates from PC12 cells treated with siRNA targeted to either NOX2 or NOX4 or scrambled siRNA and exposed to IH₆₀ or normoxia (N) (n = 3). HDAC5, NOX2, and NOX4 proteins were analyzed by the WES System according to the manufacturer’s instructions. D: quantitative analysis of the blots. Data were analyzed by the integrated Compass software (protein simple). Briefly, relative amount of each protein was calculated by peak area of protein/peak area of tubulin (loading control) from the chromatogram and normalized to normoxic controls **P ≤ 0.05; ns = not significant (P > 0.05) as determined by Mann–Whitney test. ROS, reactive oxygen species.

Among several NOX isoforms, NOX2 and NOX4 were shown to be major sources of ROS generation by IH (35, 37–40). However, NOX inhibitor VAS2870, cannot distinguish between NOX2 and NOX4. Therefore, we employed a genetic approach to assess the relative contribution of NOX2 and NOX4 to HDAC5 reduction by IH. Cells were transfected with silencing RNA (siRNA) targeted either to NOX2 or NOX4 and then were treated with IH. Cells treated with scrambled RNA served as controls. NOX2 and NOX4 proteins were determined using capillary electrophoresis by WES approach. IH-treated cells showed increased NOX2 and NOX4 protein abundance in scrambled RNA-treated cells but not in cells treated with siRNA targeted either to NOX2 or NOX4 (Fig. 2, C and D). The HDAC5 degradation seen during IH was absent in cells treated with either NOX2 or NOX4 siRNA compared with control scrambled RNA-treated cells (Fig. 2, C and D). These findings suggest that ROS generated by both NOX2 and NOX4 contribute to reduced HDAC5 protein expression by IH.

Dephosphorylation at S²⁵⁹ Leads to HDAC5 Protein Degradation by IH

We then sought to determine the mechanism by which ROS decrease HDAC5 protein. Altered phosphorylation is an important mechanism regulating HDAC5 activity and stability (10, 11). CAMK-dependent HDAC5 phosphorylation increases ubiquitination and proteasomal degradation (6, 41). IH activates CAMKII (19). Therefore, we tested whether CAMK contributes to HDAC5 degradation by IH. PC12 cells were treated with KN93 (5 µM), a CAMK inhibitor. KN93 decreased HDAC5 protein both in control and IH-treated cells (Supplemental Fig. S1A). Ectopic expression of constitutively active form of CAMKII (CAMKII 290) or WT CAMKII, had no effect on IH-induced HDAC5 degradation (Supplemental Fig. S1B). These results indicated that HDAC5 degradation by IH does not involve phosphorylation by CAMKII. In addition to CAMKII, both PKA and PKC can phosphorylate HDAC5 (15, 17). IH activates PKC and PKA through ROS signaling (20, 21). However, treating cells with either BIS (10 µM), a PKC inhibitor, or with 14–22 amide (1.25 µM), a PKA inhibitor, did not prevent HDAC5 degradation by IH (Supplemental Fig. S1C).

Earlier studies reported that altered phosphorylation of S²⁵⁹, S²⁷⁹, and S⁴⁹⁸ residues regulates HDAC5 protein stability (10, 11). We determined the effect of IH on S²⁵⁹ phosphorylation. IH₆₀ decreased HDAC5^S259 phosphorylation (pHDAC5^S259), and this response was blocked by NOX inhibitor VAS2870 (Fig. 3A). To further elucidate the significance of S²⁵⁹ phosphorylation, cells were transfected with plasmids encoding either a wild-type HDAC5 protein or mutant HDAC5 protein in which S²⁵⁹ was substituted with alanine. Under room air conditions, HDAC5 and pHDAC5S²⁵⁹ protein abundance was 8- to 10-fold greater in cells with wild-type ectopic expression than in control cells. IH was ineffective in reducing the overexpressed HDAC5 protein and S²⁵⁹ phosphorylation (Fig. 3B). It is likely that the high abundance of HDAC5 protein exceeds the dephosphorylation capability of IH. On the other hand, cells expressing the mutant HDAC5^S259A showed reduced HDAC5 as well as reduced pHDAC5^S259A both in control and IH-treated cells (Fig. 3B). These results indicate that HDAC5 protein degradation seen during IH is associated with reduced phosphorylation of HDAC5 at residue S²⁵⁹.

Figure 3.

Dephosphorylation of S²⁵⁹ contributes to IH-induced HDAC5 downregulation. A, top: HDAC5 and phosphorylated S²⁵⁹ HDAC5 (pHDAC5^S259) protein expression with tubulin as a loading control in lysates from room air (N) and IH-exposed cells treated with vehicle or VAS2870 (n = 3). A, bottom: Quantitative analysis of the blots (n = 3). B, top: HDAC5 and pHDAC5^S259 protein expression in lysates from room air (N) and IH₆₀-exposed cells transfected with plasmids coding for WT HDAC5 or HDAC5 in which the S²⁵⁹ was substituted with alanine or control vector plasmid (pCMV7). B, bottom: quantitative analysis of the blots (n = 3). Relative amounts of each protein were normalized to corresponding tubulin loading controls and expressed as ratio of phosphor-HDAC5 to total HDAC5 protein expression. **P ≤ 0.05; ns = not significant (P > 0.05) as determined by one-way ANOVA with post hoc Turkey HSD test. HSD, honestly significant difference.

Protein Phosphatases 1/2 Mediate HDAC5 Dephosphorylation by IH

Phosphatases catalyze dephosphorylation of phosphorylated proteins. To assess whether the decrease of HDAC5^S259 involve phosphatases, IH-exposed cells were treated with a phosphatase inhibitor cocktail (PIC) containing three inhibitors including: calyculin A, an inhibitor of protein phosphatases 1 (PP1) and 2 A (PP2A); cantharidin an inhibitor of PP2A; and (−)-p-bromolevamisole oxalate (BTO), an inhibitor of L-isoforms of alkaline phosphatases. PIC increased overall HDAC5 and pHDAC5^S259 proteins in control room air cells, and IH was ineffective at reducing the abundance of either of these proteins (Fig. 4, A and B), indicating the involvement of phosphatases in HDAC5 degradation by IH.

Figure 4.

IH induces protein phosphatase 1/2 (PP1/2)-dependent dephosphorylation of HDAC5^S259. A: HDAC5 and phosphorylated S²⁵⁹ HDAC5 (pHDAC5^S259) protein with tubulin as a loading control in lysates from room air (N) and IH₆₀-exposed cells treated with vehicle or phosphatase cocktail inhibitor (PIC) (1/100 dilution) (n = 3). B: quantitative analysis of the blots. C: HDAC5 and pHDAC5^S259 protein expression in PC12 cells treated with two different concentrations of p-bromolevamisole oxalate (BTO), cantharidin, and calyculin A (Cal A), individual components of PIC in room air (n = 3). D: HDAC5 and pHDAC5^S259 protein expression in lysates from room air (N) and IH₆₀-exposed cells treated with vehicle or calyculin A (10 nM). E: quantitative analysis of the blots (n = 3). **P ≤ 0.05; ns = not significant (P > 0.05) as determined by one-way ANOVA with post hoc Turkey HSD test. F: representative immunoblot showing the expression of PP1α/γ, PP2A-A, PP2A-B, and PP2A-C protein abundances in cell lysates treated with vehicle or calyculin A (10 nM) and exposed to normoxia or IH₆₀. HSD, honestly significant difference

To identify the specific phosphatase(s) contributing to dephosphorylation of HDAC^S259, cells cultured in room air were treated with different concentrations of either BTO, cantharidin, or calyculin A, and HDAC5 protein as well as S²⁵⁹ phosphorylation were subsequently analyzed. As shown in Fig. 4C, calyculin A increased the abundance of both HDAC5 and pHDAC5^S259 proteins whereas neither BTO nor cantharidin had any effect. Calyculin A also blocked IH-induced decrease of HDAC5 and pHDAC5^S259 proteins (Fig. 4, D and E).

Calyculin A binds and inhibits both PP1 and PP2A catalytic subunits with similar IC₅₀ values (42). To determine which of the two phosphatases contribute to HDAC5 degradation, the effects of IH on PP1 and PP2A protein were determined in the presence and absence of calyculin A. PP1 isoforms are comprised of four subunits (α, γ1, γ2, and δ) (43), whereas PP2A is a heterotrimer consisting of a scaffolding A subunit (PP2A-A), a catalytic C subunit (PP2A-C) and a variable regulatory B subunit (PP2A-B) (44). IH by itself had no effect on PP1α/γ, PP2A-A, PP2A-B, and PP2A-C proteins, but calyculin A reduced PP1α/γ protein abundance but not PP2A-A, PP2A-B, and PP2A-C proteins in both control (room air treated) and IH-treated cells (Fig. 4F). These data suggest a role for PP1 in degradation of HDAC5 and dephosphorylation of S²⁵⁹ by IH.

Ectopic Expression of Endogenous Inhibitor of PP1 Blocks HDAC5 Degradation by IH

To further establish a role for PP1, cells were transfected with siRNA targeted to PP1α/γ subunits. However, siRNA silencing of PP1 affected cell viability under normoxia which deteriorated with IH treatment similar to higher concentrations (millimolar) of calyculin A treatment. Nuclear inhibitor of PP1 (NIPP1) is an endogenous inhibitor of PP1 (45). We hypothesized that IH activates PP1 by disassociating it from the NIPP1 complex, which then frees PP1 to dephosphorylate HDAC5. If this is valid, then inhibiting PP1 activity with ectopic expression of NIPP1 should block the HDAC5 dephosphorylation at S²⁵⁹ in IH-treated cells. To assess this possibility, cells were transfected with a GFP tagged NIPP1 plasmid (46) and exposed to IH. As shown in Fig. 5A and B, IH decreased HDAC5 protein and S²⁵⁹ phosphorylation, and these effects were absent in cells expressing GFP NIPP1 plasmid. Ectopic expression of NIPP1 reduced PP1α/γ protein in both control and IH-exposed cells (Fig. 5C).

Figure 5.

Ectopic expression of endogenous inhibitor of PP1 (NIPP1) blocks IH-induced HDAC5 downregulation. A: HDAC5 protein expression in lysates from room air (N) and IH₆₀-exposed cells transfected with either NIPP1-GFP or control vector GFP plasmids. B: quantitative analysis of the blots (n = 3). **P ≤ 0.05; ns = not significant (P > 0.05) as determined by one-way ANOVA with post hoc Turkey HSD test. C: representative immunoblot showing NIPP1, GFP, and PP1α/γ protein abundances in cell lysates expressing NIPP1 and exposed to room air (N) or IH₆₀. D: HDAC5, NIPP1, and PP1α/γ protein expression in nuclear and cytoplasmic lysates of PC12 exposed to room air (N) or IH₆₀ with tubulin and TBP (TATA box binding protein) as control loading marker proteins for cytosol and nuclear fractions, respectively. GFP, green fluorescent protein; HSD, honestly significant difference; NIPP1, nuclear inhibitor of PP1; PP1, protein phosphatase.

HDAC5 shuttles between the nucleus and cytoplasm in a phosphorylation-dependent manner (6, 41). Both NIPP1 and PP1 proteins function in nucleus. Therefore, HDAC5 protein distribution in nuclear and cytoplasmic fractions of cells was determined in IH-treated cells expressing NIPP1. HDAC5 as well as PP1α/γ proteins were localized primarily in nuclear fractions of cells treated with normoxia or IH as indicated by NIPP1 and nuclear loading control TBP (Fig. 5D). These findings demonstrate that inhibition of PP1 activity pharmacologically (calyculin A) or with ectopic expression of NIPP1 blocks HDAC5 dephosphorylation by IH in the nucleus.

Dephosphorylation of HDAC5 Contributes to HIF-1α Protein Stabilization by IH

We previously reported that HDAC5 degradation and ensuing decrease in enzyme activity in IH-treated cells contribute to HIF-1-dependent transcription by stabilizing HIF-1α protein through increased lysine acetylation (9). Given that we identified HDAC5 dephosphorylation mediates HDAC5 degradation by IH, we tested whether blockade of HDAC5 dephosphorylation prevents HIF-1α protein accumulation and HIF-1 transcription by IH. Cells were exposed to IH in presence of calyculin A, which blocked pHDAC5^S259. HIF-1α protein was absent in calyculin A-treated cells exposed to either room air or IH (Fig. 6A). IH-evoked HIF-1 transcription was also blocked by calyculin A, as evidenced by absence of HRE activation measured by luciferase activity in nuclear cell lysates (Fig. 6B). Likewise, ectopic expression of wild-type HDAC5 but not the mutant HDAC5^S259A prevented HIF-1α accumulation and HRE transactivation by IH (Fig. 6, C and D).

Figure 6.

Dephosphorylation of HDAC5^S259 is required for IH augmented HIF-1α protein and transcriptional activity. HIF-1α protein (A) and hypoxic response element (HRE)-transcriptional activity (B) in lysates from room air (N) and IH₆₀-exposed PC12 cells treated with vehicle or calyculin A (10 nM) (n =3). HIF-1α protein (C) and HRE activity (D) in cells expressing either wild-type (WT) HDAC5 or HDAC5 S259A or control vector plasmid (pCMV7) (n = 6). **P ≤ 0.05; ns = not significant (P > 0.05) as determined by one-way ANOVA with post hoc Turkey HSD test. HIF, hypoxia-inducible factor; HSD, honestly significant difference.

DISCUSSION

Our earlier study reported that PC12 cells treated with IH reduce HDAC5 protein and activity and these effects were associated with lysine acetylation of HIF-1α protein and HIF-1-dependent transcriptional activation of NOX4 (9). We further found that ROS generated by NOX4 activates the sympathetic nervous system and hypertension in IH rodents (9). These observations prompted the current study to investigate the mechanisms underlying HDAC5 protein degradation by IH in PC12 cell cultures.

Major findings of the current study are: 1) NADPH oxidases derived ROS mediate HDAC5 degradation by IH through reduced HDAC5 phosphorylation at S²⁵⁹; 2) protein phosphatase 1 (PP1) contributes to dephosphorylation of HDAC5^S259 involving NIPP1 an endogenous inhibitor of PP1; and 3) Blockade of HDAC5 de-phosphorylation prevented IH-induced HIF-1α protein accumulation and transcriptional activation. Our study demonstrates that phosphatases constitute another layer in HIF-1 transcriptional activity regulation independent of previously identified protein kinase mechanisms (19, 21).

ROS signaling is critical for cardiovascular pathology in patients with OSA (31–33) and rodents treated with IH (27–30). The following results demonstrate that ROS mediate the effects of IH on HDAC5 protein: ROS scavenger prevented HDAC5 protein degradation by IH and ROS mimetics, like IH, decreased HDAC5 protein in control cells treated with room air. The IH paradigm used in this study (15 s of hypoxia followed by 5 min of room air) is patterned after the duration of hypoxic episodes encountered in adult patients with OSA, which lasts on average no more than tens of seconds and induces a robust ROS generation in rodents (23). IH is also experienced by mountain climbers during brief ascents to high altitudes for preparing summit attempts (47). Unlike IH encountered with OSA, hypoxic episodes associated with summit attempts are of longer duration, lasting from several minutes to few hours. Whether IH experienced during summit preparation also affects HDAC protein through ROS signaling remains to be investigated. In addition, differential expression of HDAC members and variations in basal enzyme activity in diverse tissues/cells may account for specific HDAC roles in various biological processes in both physiological and pathophysiological states (2).

Pharmacological blockade or genetic silencing of NOX2 and NOX4 markedly attenuated HDAC5 degradation, by IH suggesting both NOX2 and NOX4 are the major sources of ROS mediating the effects of IH on HADC5 protein. NOX2 is an inducible NOX isoform that is localized primarily at the plasma membrane and is regulated by post-translational modifications of cytosolic factors such as p47phox, p67phox, and Rac1 (48). On the other hand, NOX4 is constitutively active, does not require cytosolic factors for its activation, and is found in intracellular membranes including the nuclear membrane (48). HDAC5 shuttles between the nucleus and cytoplasm in a calcium- and phosphorylation-dependent manner in response to stimuli (6, 41). The nuclear and cytosolic compartmentalization of NOX4 and NOX2 indicates a role for NADPH oxidases in regulating the nuclear export and import of HDAC5. ROS include superoxide anions, hydrogen peroxide, and hydroxyl radicals. Determining which of the species contribute to HDAC5 degradation by IH requires further studies.

How might ROS contribute to HDAC5 degradation by IH? ROS can affect protein phosphorylation either through activation of protein kinases or through protein phosphatases. Although IH activates protein kinases including CAMKII (19), inhibitors of either CAMKII or other protein kinases (e.g., PKC and PKA) were ineffective in blocking HDAC5 protein degradation by IH. On the other hand, following observations indicate ROS-dependent dephosphorylation by protein phosphatases contributes to HDAC5 protein degradation by IH: 1) IH reduced phosphorylation of HDAC5 at residue S²⁵⁹ and this effect was blocked by NOX inhibitor VAS2870; 2) cells expressing mutant HDAC5 with substitution of S²⁵⁹ with alanine decreased HDAC5 protein expression; and 3) phosphatase inhibitors blocked HDAC5 protein degradation by IH. Whether other serine phosphorylation sites including S²⁷⁹ and S⁴⁹⁸ also contribute to HDAC5 degradation by IH remains to be investigated.

Although considerable information is available on the role of protein kinases affecting HDAC5 stability, very little is known about the role of antagonistic phosphatases which counteract the effects of protein kinases. Ser/Thr phosphatases PP1 and PP2A have been implicated in the regulation of class II HDAC activities and function (10). Our results indicate phosphatase PP1α/γ is a regulator of HDAC5 protein stability in IH. A phosphatase inhibitor cocktail (PIC) or calyculin A, an inhibitor of PP1/PP2, increased phosphorylation of HDAC5 in both control and IH-treated cells. PP1 is the more widely expressed and highly conserved of the two phosphatases, which regulate a variety of cellular functions including transcription, pre-mRNA splicing, protein synthesis, and cell cycle progression in several cell types (43). On a subcellular level, all PP1 isoforms (α, β/δ, and γ) are abundant in the nucleus and form distinct multimeric holoenzyme complexes with a wide variety of regulatory proteins that assign substrate selectivity and direct subcellular localization of the catalytic subunit (49, 50). One such regulatory protein is nuclear inhibitor of PP1 (NIPP1) which has two PP1 binding regions, a central domain containing a so-called RVXF-motif, which represents a high-affinity binding motif, and a low-affinity binding region mapped to the NIPP1 C terminus (51). NIPP1 forms an inactive complex with PP1, both in vivo and in vitro, and prevents dephosphorylation of a number of phosphoproteins (51). Disruption of PP1 and NIPP1 complex in response to stress leads to PP1 activation (52). Because, the approach of silencing of PP1 with siRNA impaired cell viability, we chose ectopic expression of NIIP1 to delineate the selective role of PP1. Ectopic expression of NIPP1 blocked IH-induced HDAC5 de-phosphorylation and HDAC5 degradation, suggesting that IH may trigger PP1 activation by disruption of the PP1 interaction with NIPP1. However, further studies are needed to determine whether IH affects PPI-NIPP1 interaction either directly or indirectly.

Nonetheless, our results suggest a hither to uncharacterized role for PP1 in HIF-1 activation by IH, involving a concerted dephosphorylation of HDACs and the ensuing lysine acetylation of HIF-1 (Fig. 7). This is reminiscent of the previously described role of HDAC-PP1 complex in dephosphorylation and deacetylation of histone H3 in yeast and Ceanorhabditis elegans (53, 54).

Figure 7.

Schematic presentation of the role of HDAC5^S259 phosphorylation in IH-induced HIF-1 transcriptional activity by IH. IH activates PP1 by disruption of PP1 interaction with NIPP1 directly or indirectly via ROS-dependent mechanism. Activated PP1 dephosphorylates HDAC5 resulting in HDAC5 degradation and the subsequent increase in HIF-1α acetylation and transcription. HIF, hypoxia-inducible factor; NIPP1, nuclear inhibitor of PP1; PP1, protein phosphatase 1; ROS, reactive oxygen species.

Lysine acetylation of HIF-1α is physiologically relevant because HIF-1-transcriptional activation is a key molecular mechanism underlying heightened sympathetic nerve activation and the resulting hypertension by OSA/IH. Given the importance of protein phosphatases in chromatin-modifying activities, phosphatases may have broad implications in modulation of signal transduction pathways linked to numerous neurological and cardiovascular diseases.

SUPPLEMENTAL DATA

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.20017709.v1).

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant P01-HL044454.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.N. conceived and designed research; N.W. performed experiments; N.W. and J.N. analyzed data; N.W. and J.N. interpreted results of experiments; J.N. prepared figures; J.N. drafted manuscript; N.R.P. and J.N. edited and revised manuscript; N.W., N.R.P., and J.N. approved final version of manuscript.