Vulnerability of HIF1α and HIF2α to damage by proteotoxic stressors

Abstract

Transcription factors HIF1 and HIF2 are central regulators of physiological responses to hypoxia and important for normal functioning of tissue stem cells and maintenance of healthy microvasculature. Even modest decreases in HIF activity exert detrimental effects in tissues although it is unclear what factors can directly impair HIF functions. We hypothesized that the presence of functionally important, large intrinsically disordered regions in HIFα subunits of HIF1/2 could make them structurally vulnerable to protein-damaging conditions. We found that common protein-damaging agents such as endogenous/exogenous aldehydes (formaldehyde, acetaldehyde), moderate heat shock and the environmental toxicant cadmium cause inactivation of HIF1 and HIF2 due to structural damage to HIFα subunits. Aldehydes triggered a rapid and selective depletion of HIF1α and HIF2α, which occurred as a result of enhanced binding of Pro-hydroxylated/VHL-ubiquitinated HIFα by 26S proteasomes. In the absence of proteasomal degradation, aldehyde-damaged HIF1 and HIF2 were transactivation defective and HIFα subunits became insoluble/denatured when their VHL-mediated ubiquitination was blocked. Protein damage by heat shock and cadmium resulted in the insolubility of Pro-nonhydroxylated HIFα. Thus, VHL-dependent ubiquitination of damaged HIFα also acts as means to maintain their solubility, permitting capture by proteasomes. The observed control of HIFα stability at the point of proteasome binding may extend to several posttranslational modifications that occur in the conformationally flexible regions of these proteins. Our findings revealed vulnerability of HIF1 and HIF2 to direct inactivation by protein-damaging agents, which helps understand their tissue injury mechanisms and favorable responses of hypoxic tumors to hyperthermia.

Keywords: Cadmium; Formaldehyde; HIF1A; HIF2A; Heat shock; Proteotoxicity.

Figure 1.. FA-induced loss of HIF1α and HIF2α proteins.

Except for panel C, immunoblots were run with soluble extracts. Tubulin and ribosomal protein L7a were used as loading controls. (A) Immunoblots of HIF1α and HIF2α in H460 and IMR90 cells treated with FA for 1 h. Cells were maintained at 5% O₂ during FA treatments. Poly-Ub: polyubiquitinated HIF1α. (B) Loss of HIFα in WI38 normal human lung fibroblasts and normal human neonatal keratinocytes after 1 h FA treatments. WI38 cells were kept at 5% O₂ and keratinocytes at 20% O₂ during FA exposures. (C) Immunoblots of HIF1α and HIF2α in whole cell lysates (2% SDS lysis) of H460 and IMR90 cells treated with FA for 1 h. (D) Time-course of HIFα depletion and HSF1 phosphorylation by FA in IMR90 and H460 cells. Cells were kept at 5% O₂ for 3 h before the addition of FA. (E) HIF1α and HIF2α levels in H460 cells at different recovery time after FA removal (200 μM FA for 1 h, 5% O₂ during and after FA exposures). (F) Cytotoxicity of FA treatments for 30 and 60 min. Cell viability was measured after 72 h recovery in FA-free media. Cells were kept at 5% O₂ during and after FA exposures. Data are means±SD, n=3. (G) mRNA levels of HIF1A and HIF2A in H460 cells treated with FA for 30 and 60 min. Data are means±SD (n=3), ns - not significant changes. (H) Representative immunoblot of HIF1β in H460 cells treated with FA for 30 min. (I) Relative protein expression of HIF1β and four transcription factors in H460 cells treated with FA for 30 min. Band intensities were analyzed by ImageJ. Data are means±SD, n=4. (J) Representative immunoblots of four transcription factors and phospho-HSF1 in cells treated as in panel I.

Figure 2.. Testing for protein degradation pathways involved in FA-induced HIFα loss.

Cells were maintained at 5% O₂ during treatments. All westerns were run with whole cell lysates. (A) Immunoblots of IMR90 and H460 cells pretreated with the LMP7 inhibitor ONX-0914 (LMP7i) for 2 h prior to the addition of 200 μM FA for 1 h. (B) HIFα levels in IMR90 and (C) H460 cells pretreated with autophagy inhibitors NH₄Cl (20 mM) or bafilomycin (50 nM) for 4 h prior to the addition of 200 μM FA for 1 h. (D) Immunoblots of cells pretreated with the proteasome inhibitors MG132 or (E) MG115 for 1 h before the addition of 200 μM FA for 1 h. (F) Immunoblots of IMR90 cells pretreated with the RPN13 inhibitor RA190 for 1 h prior to the addition of 200 μM FA.

Figure 3.. VHL- and O₂-dependence of HIFα depletion by FA.

Cells were maintained at 20% O₂. Whole cell lysates were analyzed by western blotting. (A) Immunoblots of IMR90 cells pretreated with the VHL inhibitor VHL298 for 1 h prior to the addition of 300 μM FA for 1 h. (B) HIFα stability in VHL-null and VHL-complemented cells. RCC4 cells complemented with VHL-expressing (VHL+) or empty vector (VHL−/−) were treated with 200 μM FA for 15-60 min. (C) Stability of exogenous HIF1α in VHL−/− and VHL+ RCC4 cells. Cells were transfected with a vector expressing HA-tagged HIF1α and treated with 200 μM FA for 1 h. (D) Stability of HIF1α and HIF2α proteins in PHD-inhibited H460 or (E) IMR90 cells. Cells were pretreated with 2 mM DMOG, 300 μM CoCl₂, or 400 μM NiCl₂ for 6 h prior to the addition of 200 μM FA for 1 h. (F) Immunoblots for HA-tagged HIF1α in 293T cells transfected with plasmids expressing HA-HIF1α or HA-HIF1α P402A/P564A double mutant and treated with 200 μM FA for 1 h. (G) Protein levels of VHL and PHD2 in H460 and IMR90 cells treated with FA for 1 h. (H) Proteasome binding of Pro564-hydroxylated HIF1α in H460 cells pretreated with the RPN13 inhibitor RA190 for 1 h and then treated with 300 μM FA for 30 min. Proteasomes were separated by native PAGE. (I) Standard SDS-PAGE immunoblot for total cytosolic Pro564-hydroxylated HIF1α in cells treated as in panel H. (J) Proteasome binding of Pro564-hydroxylated HIF1α in cells pretreated with MG132 for 1 h and then treated with 300 μM FA for 30 min. Proteasomes were separated by native PAGE. (K) Standard SDS-PAGE immunoblot for total cytosolic Pro564-hydroxylated HIF1α in cells treated as in panel J.

Figure 4.. Testing effects of inhibitors of signaling processes and activity of HIFs.

H460 cells were treated with inhibitors for 1 h before the addition of FA. Treatments were carried out in 20% O₂ except for some experiments in panel E. Whole cell lysates were used for all westerns. (A) HIF1α depletion in cells treated with 200 μM FA in the presence of inhibitors of various signaling pathways. Kinase inhibitors: JNK (JNKi) - 10 μM SP600125, p38 (p38i) - 10 μM SB203580, ERK (ERKi) - 10 μM FR180204, CDK9 (CDK9i) - 10 μM inhibitor II, CDK4/6 (CDK4/6i) - 1 μM PD0332991, CDK1 (CDK1i) - 10 μM RO-3306, ATR (ATRi) - 10 μM VE821, ATM (ATMi) - 10 μM KU55933, DNAPK (DNAPK-i) - 3 μM NU7441, AMPK (AMPKi) - 1 μM Compound C, PLK1/3 (PLK1/3i) - 0.5 μM Gw843682x, GSKα/β (GSK3α/βi) - 1 uM Chir99021; protein deacetylation inhibitor - 3 μM Trichostatin A (TSA). (B) HIF1α depletion and recovery in cells with inactivated SUMO E1 activity (SUMOi – 1 μM TAK-981). (C) Immunostaining for HIF1α in cells treated for 6 h with 2 mM DMOG alone or in the presence of 200 μM FA. (D) RT-qPCR for HIF1-inducible genes in cells treated for 6 h with 2 mM DMOG in the presence of FA. Means±SD, n=3, **- p<0.01, ***-p<0.001. (E) HIF2-dependence of NDRG1 and PFKFB3 upregulation (6 h in 2% O₂, si-ns: nonspecific siRNA, si-HIF2A: targeted siRNA). Means±SD, n=3, **- p<0.01. Western blot shows a specific knockdown of HIF2α by HIF2A-targeted siRNA. (F) Suppression of HIF2-dependent gene upregulation by DMOG in the presence of FA (conditions as in panel D).

Figure 5.. Impact of FA and VHL on the formation of insoluble HIFα proteins.

All experiments were done at 20% O₂. HIFα westerns were performed with the insoluble fraction of cells. (A) Insoluble HIF1α and HIF2α in H460 cells treated with 0 or 200 μM FA (1 h) with and without preincubation with 0-250 μM VHL inhibitor VH298 (VHLi) and/or 10 μM MG132 for 1 h. (B) Insolubility of HIFα proteins in FA-treated H460 cells (300 μM, 30 min) preincubated with the NEDD8-activating enzyme inhibitor MLN4924 (NEDD8i, 1 μM for 1 h) and/or 10 μM MG132 (30 min). (C) Formation of insoluble HIF1α by FA in IMR90 cells with VHL knockdown by siRNA. Cells were pretreated with 10 μM MG132 for 1 h before the addition of FA for 1 h.

Figure 6.. Instability and loss of solubility of HIFα proteins under different proteotoxic conditions.

All experiments were done in 20% O₂. (A) Loss of HIFα proteins in H460 cells treated with acetaldehyde (AcA, 1 h). p-HSF1: Ser326-phosphorylated HSF1. (B) Resistance of HIFα proteins to depletion by acetaldehyde in the presence of inhibitors of O₂-dependent ubiquitination (2 mM DMOG, 0.2 mM VH298). Inhibitors were added 1 h before acetaldehyde. (C) Accumulation of insoluble HIFα proteins during heat shock in H460 cells. Equal percentages (by volume) of each fraction were analyzed by westerns. (D) Loss of HIFα solubility in normal human keratinocytes after a brief hyperthermia (43°C, 10 min). (E) Accumulation of insoluble HIFα in H460 cells treated with cadmium for 2 h. Equal percentages (by volume) of each fraction were used for westerns. (F) Detection of the transcription factors p53 and E2F1 in the soluble and insoluble fractions of H460 cells treated with cadmium as in panel E. (G) Formation of insoluble HIFα and p53 in H460 cells treated with Cd(II) for 6 h.