ORIGINAL ARTICLE

Interaction between PARP-1 and HIF-2a in the hypoxic responseA Gonzalez-Flores1,6, R Aguilar-Quesada1,6, E Siles2, S Pozo3, MI Rodrıguez-Lara1, L Lopez –Jimenez1, M Lopez-Rodrıguez1,A Peralta-Leal1, D Villar4, D Martın-Oliva5, L del Peso4, E Berra3,6 and FJ Oliver1,6

Hypoxia-inducible factors (HIFs) mediate the transcriptional adaptation of hypoxic cells. The extensive transcriptional programmregulated by HIFs involves the induction of genes controlling angiogenesis, cellular metabolism, cell growth, metastasis, apoptosis,extracellular matrix remodeling and others. HIF is a heterodimer of HIF-a and HIF-b subunits. In addition to HIF-1a, HIF-2a hasevolved as an isoform that contributes differently to the hypoxic adaptation by performing non-redundant functions. Poly (ADP-ribose) polymerase-1 (PARP-1) is a nuclear protein involved in the control of DNA repair and gene transcription by modulatingchromatin structure and acting as part of gene-specific enhancer/promoter-binding complexes. Previous results have shown thatPARP-1 regulates HIF-1 activity. In this study, we focused on the cross-talk between HIF-2a and PARP-1. By using differentapproaches to suppress PARP-1, we show that HIF-2a mRNA expression, protein levels and HIF-2-dependent gene expression, suchas ANGPTL4 and erythropoietin (EPO), are regulated by PARP-1. This regulation occurs at both the transcriptional and post-trancriptional level. We also show a complex formation between HIF-2a with PARP-1. This complex is sensitive to PARP inhibitionand seems to protect against the von Hippel–Lindau-dependent HIF-2a degradation. Finally, we show that parp-1� /� mice displaya significant reduction in the circulating hypoxia-induced EPO levels, number of red cells and hemoglobin concentration.Altogether, these results reveal a complex functional interaction between PARP-1 and the HIF system and suggest that PARP-1is involved in the fine tuning of the HIF-mediated hypoxic response in vivo.

Oncogene advance online publication, 4 March 2013; doi:10.1038/onc.2013.9

Keywords: hypoxia; PARP-1; HIF-2

INTRODUCTIONWhen oxygen (O2) availability decreases (hypoxia), the transcrip-tion of specific genes engaged in adaptive mechanisms such asglucose transport, glycolysis, erythropoiesis or angiogenesis isquickly increased.1–4 The hypoxia-inducible factor (HIF) mediatesthe upregulation of many of these hypoxia-regulated genes. HIF isa heterodimer of a- and b-subunits5 that binds to target genesthrough cis-acting hypoxia-response elements.6 To date, three HIF-a isoforms have been described, with HIF-1a and HIF-2a being thebest characterized.

The HIF-a subunits are regulated by O2 availability, whereasHIF-b remains largely unaffected by changes in O2 levels. Undernormoxic conditions, HIF-a subunits are constitutively translatedand degraded by the proteasome. HIF-a is indeed rapidlyrecognized by the product of the von Hippel–Lindau (pVHL)tumor-suppressor gene,7 that targets the protein for degradation.pVHL recognition of HIF-a is dependent on hydroxylation ofspecific proline residues8,9 by a family of prolyl hydroxylasedomain-containing proteins (PHDs).10,11 Three different PHDs(PHD1, PHD2 and PHD3) acting as critical O2 sensors have beencharacterized in mammals. Under hypoxic conditions,hydroxylation is inhibited resulting in increased HIF-a stability;HIF-a translocates into the nucleus and heterodimerizes with theHIF-b subunit and thus form the active HIF complex. Moreover,HIF-a is hydroxylated on an asparagine residue by factor-inhibitingHIF, which is subjected to similar O2-dependent regulatory

mechanisms as the PHDs and suppresses HIF transcriptionalactivity under normoxic conditions by blocking its association withthe coactivator p300/CBP.12

The HIF-2a subunit, also known as endothelial PAS domainprotein-1, shares 48% amino-acid sequence identity with HIF-1a.13

Despite these similarities, targeted disruption of the HIF-1a orHIF-2a gene in mice results in strikingly different phenotypes,demonstrating non-redundancy of their functions duringembryonic development.14–17 Furthermore, HIF-2a knock-in didnot rescue a HIF-1a null phenotype.18 Thus, it is accepted thatHIF-1 and HIF-2 contribute differently to the hypoxic adaptationby regulating the expression of overlapping but non-identicaltarget genes.19 Transcriptional activation by HIF-2 regulates sometarget genes in common with HIF-1, as well as other genes thatare uniquely regulated by HIF-2. For example, HIF-2 regulatesexpression of erythropoietin (EPO),20 induction of PHD2 isexclusive of HIF-1, whereas HIF-1 and HIF-2 induce PHD3.21

HIF-1 and HIF-2 show even opposite effects on cell proliferationthrough c-MYC regulation.22,23

Some mechanisms have been proposed to account for HIF-1and HIF-2 specificities.24,25 In the present work, we studied theimpact of poly (ADP-ribose) polymerase-1 (PARP-1) on HIF-2-dependent response to hypoxia and the regulation of HIF-2signaling by PARP-1. PARP-1 is an abundant nuclear enzyme thatuses NADþ as a substrate to catalyze the covalent attachment ofADP-ribose units on nuclear acceptor proteins or on PARP-1 itself.

1Institute of Parasitology and Biomedicine Lopez-Neyra, CSIC, Granada, Spain; 2Department of Experimental Biology, University of Jaen, Jaen, Spain; 3CIC bioGUNE, ParqueTecnologico de Bizkaia, Derio, Spain; 4Institute of Biomedical Research Alberto Sols, CSIC, Faculty of Medicine, Autonomous University of Madrid, Madrid, Spain and 5Departmentof Cellular Biology, Faculty of Sciences, University of Granada, Granada, Spain. Correspondence: Dr E Berra, CIC bioGUNE, Parque Tecnologico de Bizkaia-Ed. 801A, Derio 48160,Spain or Professor F Oliver, Department of Immunology and Cell Biology, CSIC, Avda Conocimiento s/n, Armilla, Granada 18100, Spain.E-mail: eberra@cicbiogune.es or joliver@ipb.csic.es6These authors contributed equally to this work.Received 31 May 2012; revised 22 October 2012; accepted 7 December 2012

Oncogene (2013), 1–8& 2013 Macmillan Publishers Limited All rights reserved 0950-9232/13

www.nature.com/onc

The resulting PAR alters the functional properties of the differenttarget proteins, with consequences in a number of biologicalfunctions.26 Initially, it was assumed that PARP-1 was primarilyinvolved in DNA repair, but additional studies have providedevidence that PARP-1 acts also as a component of enhancer/promoter regulatory complexes.27,28

We found that PARP-1 interacts with HIF-2a, and PARPinhibition affected complex formation and HIF-2a protein stability.Interestingly, PARP-1 was also found to be present at the HIF-2apromoter. PARP-1 depletion strongly decreased HIF-2a mRNA and,consequently, protein levels, as well as the transcription of HIF-2-dependent genes. Accordingly, hypoxia-induced EPO synthesis,mainly regulated by HIF-2, as well as the number of red cells andhemoglobin concentration, was also decreased at systemic level inPARP-1-knockout mice. Therefore, we conclude that PARP-1ablation/inhibition exerts multiple regulatory actions on theHIF-2-mediated hypoxic response at transcriptional and post-transcriptional level.

RESULTSPARP-1 impacts on hypoxia-dependent HIF-2a accumulationPARP-1 has been shown to have an impact on HIF-1 activation indifferent settings,29,30 however, the interaction between PARP-1and HIF-2a remains unknown. To evaluate the effect of PARP-1 onHIF-2a levels, we measured the accumulation of HIF-2a uponhypoxia in mouse embryonic fibroblasts (MEFs) derived from wild-type and knocknout PARP-1 mice (Figure 1a). Hypoxic induction ofHIF-2a was impaired in parp-1� /� MEFs. Nuclear accumulation ofHIF-2a was also prevented in PARP-1-knockout cells upon hypoxia(Supplementary Figure S1). To ascertain that the above effect isspecific of PARP-1 ablation, we knocked down PARP-1 by small

interfering RNA (siRNA). Indeed, silencing of PARP-1 affectedhypoxic induction of HIF-2a protein levels (Figure 1b). Validationof our siRNAs targeting PARP-1, by transient transfection of thehuman embryonic kidney cells HEK 293T, is presented inSupplementary Figure S2. Furthermore, treatment of the cellswith PJ34, a PARP-1 inhibitor, decreased the expression levels ofFlag-HIF-2a in normoxia as well as in hypoxia (Figure 1c). We alsoconfirmed the previous results with the endogenous protein(Figure 1d). To test whether PARP-1 was affecting HIF-2a stability,the human renal carcinoma-derived cells RCC4 (±pVHL) wereincubated in the absence or presence of the PARP-1 inhibitor, PJ34(Figure 1e). In RCC4 cells, HIF-2a protein levels did not change inresponse to PJ34 either in normoxia or in hypoxia. In contrast,PARP-1 inhibition decreased the barely detectable expression ofHIF-2a in normoxia and completely abolished hypoxic inductionof the protein upon hypoxia in pVHL-restored RCC4 cells. All theseresults strongly suggested that PARP-1 has an impacts on pVHL-mediated HIF-2a stabilization and accumulation. Furthermore,PARP-1 activity is required in this process.

PARP-1 controls HIF-2a mRNA levelsHIF-2a mRNA levels also decreased drastically in immortalizedparp-1� /� MEFs compared with parp-1þ /þ cells both undernormoxic and hypoxic conditions (Figure 2a). Similarly, silencing ofPARP-1 affected basal HIF-2a mRNA levels (Figure 2b), suggestingthat transcriptional regulation of HIF-2a was impaired in theabsence of PARP-1 under both normoxic and hypoxic conditions.However, PARP-1 inhibition does not affect HIF-2a mRNA levels,contrary to what happens at the protein levels. Therefore, PARP-1seems to impact on HIF-2a expression at the mRNA and proteinlevels through different mechanisms.

N

parp-1 +/+ parp-1 -/-

HIF-2α

β-Actin

PARP-1

H N H

a b

HIF-2α

PARP-1

β-Actin

Control PARP-1siRNA:

- - + - +

dc

Em

pty

Vec

tor

Flag

N HPJ34 (10μM)

Flag-HIF2α

HIF-2α

- + - +

N H

PJ34 (10μM)

HIF-2α

- + - +N H

PJ34 (10μM) - + - +N H

RCC4 RCC4 + pVHLe

β-Actin

β-Actin

β-Actin

Figure 1. PARP-1 is required for the expression of HIF-2a protein. (a) parp-1þ /þ and parp-1� /� immortalized MEFs. (b) siRNA of PARP-1 inHEK293T. (c) COS cells transiently transfected with pcDNA3-Flag-HIF-2a and incubated for 2 h in the absence (� ) or presence (þ ) of PJ34(10mM) and then in N or H for 16 h. (d) HEK293T cells incubated for 2 h in the absence (� ) or presence (þ ) of PJ34 (10mM) and then in N or Hfor 16 h. (e) RCC4±pVHL cells incubated for 2 h in the absence (� ) or presence (þ ) of PJ34 (10 mM) and then in N or H for 16 h.

Regulation of HIF-2 by PARP-1A Gonzalez-Flores et al

2

Oncogene (2013), 1 – 8 & 2013 Macmillan Publishers Limited

HIF-2a form a complex with PARP-1 that increases during hypoxiaTo investigate the molecular mechanisms of PARP-1-mediatedregulation of HIF-2a protein accumulation, we evaluated whetherboth proteins physically interact, as previously shown for PARP-1and HIF-1a.30 We immunoprecipitated HIF-2a from RCC4 whole-cell extracts and tested for the presence of PARP-1 by western blotanalysis. HIF-2a formed a complex with PARP-1, which wasincreased upon hypoxia (Figure 3a). Moreover, the presence of thePARP-1 inhibitor, PJ34, resulted in a strong decrease of thecomplex upon hypoxia. Immunoprecipitation with anti-PARP-1was also able to pull down HIF-2a (data not shown). Similarbinding was also observed for overexpressed Flag-HIF-2a withendogenous PARP-1 protein, and this complex was dependent onPARP activity (Figure 3b). g-irradiation of the cells to activate PARPdid not further increase PARP-1 and HIF-2a complex formation(Figure 3c).

PARP-1 controls HIF-2a transcriptionTo explore whether PARP-1 was implicated in the regulation ofHIF-2a transcription, we performed chromatin immunoprecipita-tion analysis. PARP-1 binds to the HIF-2a promotor as it did to awell-known binding site within the inducible nitric oxide synthasepromoter and not to an intronic region of HIF-2a used as apositive and negative control, respectively (Figure 4a). Further-more, PARP-1 did not bind to the HIF-1a promoter, suggesting aspecific mechanism in the activation of HIF-2 gene expression byPARP-1. PARP-1 binding to the HIF-2a promoter was confirmed byquantitative PCR (Figure 4b). These results suggest that PARP-1specifically binds to the HIF-2a promoter and may regulate HIF-2aexpression at transcriptional level.

HIF-2 signaling and PARP-1We next addressed whether PARP-1 or its inhibition affectedHIF-2-dependent gene expression. Using a hypoxic array, weidentified ANGPTL4 as the most differentially expressed genebetween parp-1þ /þ and parp-1� /� cells upon hypoxia (data notshown). Interestingly, ANGPTL4 is induced by hypoxia and hasbeen described as a HIF-dependent gene (Imamura et al., 2009).

We validated these data by analyzing the expression of differentHIF-target genes in MEFs derived from wild-type and knocknoutPARP-1 mice by quantitative PCR. Hypoxic induction of ANGPTL4and PHD3 was strongly abolished in parp-1� /� MEFs, whereasBNIP-3 or PHD2 expression was not affected (Figure 5a; data notshown). To further confirm the specific role of PARP-1 in thecontrol of HIF-2-dependent target genes, we silenced PARP-1, HIF-2a or HIF-1a and analyzed the expression of ANGPTL4 and BNIP-3,which expression is HIF-1- and HIF-2-dependent or exclusively HIF-1-dependent, respectively (Figure 5b). Silencing of PARP-1reduced significantly the hypoxic induction of ANGPTL4 thatwas indeed completely blocked following HIF-2a silencing,whereas BNIP-3 expression was not affected by PARP-1 nor HIF-2a silencing. Thus, knocking down of PARP-1 significantly reducedhypoxia-induced ANGPTL4 expression. Moroever, hypoxia wasunable to induce ANGPTL4 following HIF-2a silencing, whereasablation of HIF-1a still allows hypoxia-dependent ANGPTL4induction. In contrast, in the case of BNIP-3 (which is a well-known and specific HIF-1 target), HIF-1a silencing completelyprevented hypoxia-induced BNIP-3 expression, whereas neitherHIF-2a or PARP-1 silencing affected hypoxia-induced BNIP-3expression. Furthermore, reintroduction of PARP-1 into theparp-1� /� MEFs restored ANGPTL4 mRNA (Figure 5c) and proteinlevels (data not shown). Finally, supporting our previous data,PARP inhibition, using PJ34, also affected the hypoxia-inducedexpression of HIF-2-dependent genes such as ANGPTL4 and PHD3(Figure 5d).

PARP-1 contributes to EPO regulation in miceTo evaluate the physiological relevance of our finding, we decidedto study the regulation of EPO expression, which is mainly drivenby HIF-2 at the transcriptional level. First, we measured EPO mRNAupon hypoxia in parp-1þ /þ and parp-1� /� MEFs (Figure 6a). Asexpected, hypoxia-induced expression of EPO was significantlyreduced in parp-1-deficient cells. Similar inhibition was measuredwhen we inhibited PARP-1 activity by using PJ34 (Figure 6b).

We next moved to study EPO in vivo regulation and measuredcirculating EPO levels in wild-type and PARP-1-knockout mice

b

Control

PARP-1siRNA:

***

0.00

0.25

0.50

0.75

1.00H

IF-2

α m

RN

Aex

pre

ssio

n

c

1.00

2.00

0.00

HIF

-2α

mR

NA

exp

ress

ion

- +PJ34 (10μM)

a

N

***

H N H

**0.00

0.25

0.50

0.75

1.00

1.25

HIF

-2α

mR

NA

exp

ress

ion

parp-1 +/+parp-1 -/-

0.00

0.25

0.50

0.75

1.00

1.25

HIF

-1α

mR

NA

exp

ress

ion

N H N H

Figure 2. PARP-1 is required for the expression of HIF-2a mRNA. (a) parp-1þ /þ - and parp-1� /� -immortalized MEFs. (b) siRNA of PARP-1 inHEK293T. (c) HEK293T cells transiently incubated for 2 h in the absence (� ) or presence (þ ) of PJ34 (10 mM).

Regulation of HIF-2 by PARP-1A Gonzalez-Flores et al

3

& 2013 Macmillan Publishers Limited Oncogene (2013), 1 – 8

exposed to hypoxia (Figure 6c). EPO levels were reduced inparp-1� /� mice compared with the wild-type animals, and thus,the in vivo data confirmed our in cellulo studies. Two hematolo-gical parameters, such as the number of red blood cells andthe hemoglobin concentration, were accordingly reduced inparp-1� /� mice after exposure to hypoxia (Figure 6d and e). Inaddition, hypoxia-induced EPO mRNA declined in parp-1� /� micein both kidney and liver, which are the main sites of EPO synthesis,whereas in brain, EPO mRNA levels were not affected(Supplementary Figure S3). These results agree with previousfinding showing the predominant role of HIF-2 in maintaining EPO

and iron homeostasis.31 Therefore, functionally, parp-1� /� micereflect the defective HIF-2 activation, suggesting that PARP-1might have an important function in the regulation of HIF-2-dependent physiological tuning.

DISCUSIONThere are still many open questions about the specific role ofPARP-1 in the hypoxic response as well as the partners of thehypoxic signaling implicated in the interaction with PARP-1. InFigure 7, we summarized the current findings showing the

PARP-1

a

IP HIF2α

INPUT

- + - +

N HPJ34 (10μM)

HIF-2α

PARP-1

HIF-2α

***

- + - +

PA

RP

-1 a

sso

ciat

edto

HIF

-2α

**

0

1

2

3

PJ34 (10μM)

N HE

mp

ty V

ecto

r

Flag

N HPJ34 (10μM)

Flag-HIF2α

b

IP Flag

INPUT

-

PARP-1

- + - +

c

PARP-1

Flag

PARP-1

PJ34 (10μM)

Gy (10) - - + +

- - + -

Em

pty

Vec

tor

Flag-HIF2α

IP Flag

INPUT

Figure 3. HIF-2a form a complex with PARP-1 dependent on PAR P-1 activity. (a) HIF-2a immunoprecipitation (IP) was done in RCC4 cellsincubated for 2 h in the absence (� ) or presence (þ ) of PJ34 (10mM) and then in N or H for 16 h. Co-IP experiments were repeated three times,and the histograms show the average of the three independent experiment. (b) COS cells transiently transfected with pcDNA3-Flag-HIF-2aand incubated for 2 h in the absence (� ) or presence (þ ) of PJ34 (10 mM) and then in N or H for 16 h. (c) HEK293T cells were transfected asbefore and subjected to 10Gy irradiation; co-IP of transfected protein was performed in untreated controls or after irradiation.

miNOS promoter [-978, -710]

mHIF-1α promoter [-700, +127]

mHIF-2α intron [+15921, +16320]

mHIF-2α promoter [-96, +283]

a

par

p-1

-/-

par

p-1

+/+

par

p-1

-/-

par

p-1

+/+

par

p-1

-/-

par

p-1

+/+

no

ne

Input IgG PARP-1 bparp-1 +/+

parp-1 -/-

0

5

10

15

20

25

30

Imm

un

op

reci

pit

ated

Ch

rom

atin

(% f

rom

inp

ut)

Figure 4. PARP-1 regulates HIF-2a expression at the transcriptional level. (a, b) Chromatin immunoprecipitation analysis was performed byusing PARP-1 antibody in normoxic parp-1þ /þ and parp-1� /� immortalized MEFs.

Regulation of HIF-2 by PARP-1A Gonzalez-Flores et al

4

Oncogene (2013), 1 – 8 & 2013 Macmillan Publishers Limited

intricate actions of PARP-1 (the protein and its poly(ADP-riboseactivity) at different levels in the HIF-2-mediated hypoxicresponse: (1) PARP-1, independently of its catalytic activity, bindsto the HIF-2a promoter, and depletion of PARP-1, eithergenetically or by transient gene silencing, decreases HIF-2a mRNAlevels; (2) PARP-1 also binds to HIF-2a at the protein level andPARP-1/HIF-2a association, which is dependent on PARP-1poly(ADP-ribose) activity, seems indeed to protect HIF-2a againstpVHL-mediated destabilization. (3) Therefore, PARP-1 controls thehypoxic induction of HIF-2 target genes; whether this is a directand/or an indirect effect (mediated through the impact onHIF-2a mRNA and/or protein levels) should require furthercharacterization.

Both HIF-2a and HIF-1a are expressed at high levels in a varietyof human tumors and tumor cell lines, although the relativecontribution of each protein to tumor initiation and progression isa rather contentious area. For reasons that remain unclear, inneoplastic epithelial cells of clear cell renal carcinomas, the normalpredominance of HIF-1a expression in non-neoplastic renaltubules is strikingly altered in favor of HIF-2a expression.32

Furthermore, genetic manipulation of clear cell renal carcinomacells indicates that activation of HIF-2a but not HIF-1a promotestumor growth.33–36 Hemangioblastomas, highly vascularizedtumors of the central nervous system, are the most frequent

manifestation of the autosomal-dominant-inherited VHL disease,and they show HIF-2a upregulation,37 which determineneuroblastoma aggressiveness.38 In view of the previousexamples, HIF-2a could represent an important target in cancertherapy, and it would be very interesting to rely on a tool tospecifically target HIF-2a and to counteract tumor development.Based on our results, PARP-1 appears as a good candidate at leastin certain tumor models where inhibition of PARP activity mightcounteract HIF-1a and HIF-2a accumulation. Nonetheless, furtherstudies are necessary to elucidate the molecular mechanismunderlying the specific impact of PARP-1 on HIF-2a.

MATERIALS AND METHODSCell cultureImmortalized MEFs derived from both wild-type and knockout PARP-1mice were grown as previously described.39 RCC4±pVHL and COS cellswere grown in Dulbecco’s Modified Eagle medium supplemented with10% fetal bovine serum plus penicillin (50 IU/ml) and streptomycin (50mg/ml).Medium used for HEK 293T cells was Dulbecco’s Modified Eagle mediumlow glucose supplemented with 10% fetal bovine serum, L-glutamine 4 mM,MEM non-essential amino acids 0.1 mM and antibiotics (Gibco, Paisley, UK).For normoxic conditions, cells were incubated in a regular incubator (21%O2, 5% CO2). Hypoxic conditions (1% O2, 5% CO2) were reached byincubation of cells in a sealed hypoxic workstation (Ruskinn, Bridgend, UK).

PHD3

**

0

5

10

15

20m

RN

A e

xpre

ssio

n(f

old

ind

uct

ion

)BNIP3

0

5

10

15

20

a parp-1 +/+

parp-1 -/-

*

ANGPTL4

0

1

2

3

4

5

N H N H N H N HN H N H

d

10

20

30

40

50

60

70

- + - +PJ34 10μM

***

PHD3

mR

NA

exp

ress

ion

(fo

ld in

du

ctio

n)

NH

0

5

10

15

***

ANGPTL4

- + - +

1

2

3

4

5 **

ko

ko P

AR

P

PARP-1

α-tubulin

mR

NA

exp

ress

ion

(fo

ld in

du

ctio

n)

ANGPTL4c

- +PARP-1

ANGPTL4

0.00

1.25

2.50

3.75

5.00

6.25

7.50

***

BNIP3

**

N H N H N H N H

Control PARP-1siRNA: HIF-2α HIF-1α

mR

NA

exp

ress

ion

(fo

ld in

du

ctio

n)

b

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

**

mR

NA

exp

ress

ion

(fo

ld in

du

ctio

n)

Figure 5. HIF-2-dependent gene expression is modulated by PARP-1. (a) parp-1þ /þ - and parp-1� /� -immortalized MEFs. (b) siRNA of PARP-1,HIF-1a and HIF-2a in HEK293T. (c) PARP-1 expression restored in parp1� /� immortalized MEFS. (d) HEK293T cells incubated for 2 h in theabsence (� ) or presence (þ ) of PJ34 (10mM) and then in N or H for 16 h.

Regulation of HIF-2 by PARP-1A Gonzalez-Flores et al

5

& 2013 Macmillan Publishers Limited Oncogene (2013), 1 – 8

Reagents and antibodiesThe PARP inhibitor, PJ34 (N- (6-Oxo-5,6-dihydro-phenanthridin-2-yl)-(N,N-dimethylacetamide.HCl), was purchased from Enzo Life Sciences (SanDiego, CA, USA) and it was used at a concentration of 10mM. For westernblot analysis, we used polyclonal anti-HIF-1a from Bethyl (Montgomery,TX, USA), polyclonal anti-HIF-2a from Novus Biologicals, monoclonal anti-PARP-1 (C2-10) from Alexis Biochemicals (San Diego, CA, USA) andmonoclonal anti-b-actin from Sigma (St Louis, MO, USA). For co-immunoprecipitation assays, Protein A Sepharose-monoclonal anti-FLAGfrom Sigma and polyclonal anti-HIF-2a from Novus Biologicals were used.For chromatin immunoprecipitation analysis, polyclonal anti-PARP-1 fromAlexis Biochemicals was used.

Co-immunoprecipitationCells were treated as indicated in the figure legends. To prepare cellextracts, cells were washed twice in phosphate-buffered saline, thencollected with lysis buffer (1% NP-40, 100 mM NaCl, 20 mM Tris-Cl pH 7.6,10 mM sodium pyrophosphate, 1 mM EGTA pH8.0, 10 mM NaF, 1 mM

sodium orthovanadate, 500mM phenylmethylsulfonyl fluoride andcocktail of proteases inhibitors (Complete Mini, Roche, Zurich, Switzerland))and incubated on ice for 20 min. Cell lysates were first incubated withProtein A Sepharose. The precleared supernatants were incubatedwith anti-FLAG mAb, anti-PARP-1 pAb or anti-HIF-2a pAb antibodiesplus Protein A Sepharose overnight at 4 1C. Immunoprecipitates werewashed three times with phosphate-buffered saline. The immunoprecipi-tated complexes were boiled in SDS sample buffer, separated by SDS–PAGE on a 7.5% gel and then transferred to a polyvinylidene difluoridemembrane.

Small interfering RNA preparation and transient transfectionPARP-1 siRNA was chemically synthesized and annealed by Eurogentec(Liege, Belgium); and the control and HIF-2a siRNAs by Ambion. The PARP-1 siRNA (GenBank accession No. NP_001609) corresponds to the codingregion 421–441 relative to the start codon. The HIF-2a siRNA (GenBankaccession No. NP_001421) corresponds to the coding region 1111–1131relative to the start codon. The siRNA used as an irrelevant control hasbeen previously described.40 HIF-1a siRNA was purchased from Santa CruzBiotechnology (Santa Cruz, CA, USA). Lipofectamine 2000 reagent

(Invitrogen, Paisley, UK) was used for HEK 293T cells transfection andJetPei reagent (PolyPlus Transfection, Illkirch, France) was used in COS cells.All transfection reagents were used according to the manufacturer’srecommendations. Cells were analyzed 48 h after transfection.

Quantitative PCRAfter treatments, cells were washed with phosphate-buffered saline andcollected for total RNA extraction with RNeasy Mini Kit from Qiagen(Hilden, Germany). For mRNA expression analysis, 500 ng of total RNA fromeach sample was retro-transcribed to cDNA (MultiScribe Reverse Tran-scriptase; Applied Biosystems, Carlsbad, CA, USA) and amplified with thequantitative PCR MasterMix Plus for SYBR Green I (Eurogentec, Liege,Belgium). PCR amplifications were carried out in an ABI PRISM 7000Sequence Detection System (Applied Biosystems, Carlsbad, CA, USA) anddata were analyzed with ABI PRISM 7000 SDS software. For each sample,duplicate determinations were made, and the mRNA copy number wasnormalized for the amount of Arbp in MEFs, and Rplp in RCC4 and HEK293T cells. The primer sequences for quantitative PCR are available onrequest.

Western blottingCells were lysed and sonicated in Laemmli buffer after phosphate-bufferedsaline wash. The protein concentration was determined using the Lowryassay, and 60mg for HIF-2a or 40 mg for PARP-1 of whole-cell extracts wasresolved by SDS–PAGE (7.5%). Next, proteins were transferred onto anitrocellulose membrane (Amersham Biosciences, Piscataway, NJ, USA) forHIF-2a or polyvinylidene difluoride (Millipore, Billerica, MA, USA) for PARP-1. The same membranes were probed with anti-b-actin. Immunoreactivebands were visualized with the ECL system (Amersham Biosciences,Piscataway, NJ, USA).

In vivo experimentsMale C57/BL6 mice, wild-type or PARP-1 knockout,39 were exposed to anO2 concentration of 6% (or untreated) for different times as described inthe figure legend.

Acute hypobaric hypoxia was induced by using a slight modification of apreviously published procedure.29 Briefly, animals were introduced into a

aEPO

0.0

2.5

5.0

7.5

10.0

N H N Hm

RN

A e

xpre

ssio

n(f

old

ind

uct

ion

)

0***

- + - +PJ34 10μM

1

2

3

4

5

6

mR

NA

exp

ress

ion

(fo

ld in

du

ctio

n)

EPONH

b

parp-1+/+

parp-1-/-

5 days

d

100

200

300

EP

O (

mU

/ml)

0

4

8

12

RB

C (

x1012

/L) *

c

N H N H

*

0

parp-1+/+

parp-1-/-

Control0

4

8

12

16

Hb

(g

/dL

)

*

Control 5 days

e

Figure 6. Increase in EPO levels after hypoxia is impaired in parp-1� /� mice. (a) parp-1þ /þ - and parp-1� /� -immortalized MEFs. (b) HEK293Tcells incubated for 2 h in the absence (� ) or presence (þ ) of PJ34 (10 mM) and then in N or H for 16 h. (c) Serum EPO of parp-1þ /þ andparp-1� /� mice control (N, normoxia) or exposed upon hypoxia for 8 h. Red blood cells (RBCs) (d) and hemoglobin (Hb) concentration (e) ofparp-1þ /þ and parp-1� /� mice at day 5 following no treatment or exposition upon hypoxia for 8 h.

Regulation of HIF-2 by PARP-1A Gonzalez-Flores et al

6

Oncogene (2013), 1 – 8 & 2013 Macmillan Publishers Limited

hypobaric chamber in which the air pressure was controlled by means of acontinuous vacuum pump and an adjustable inflow valve. The chamberwas also provided with a manometer to check the experimental valuesduring the process. Hypoxia was induced by downregulation of theenvironmental pressure to 225 mm Hg, which results in a pO2 partialpressure of 48 mm Hg, around 30% of the sea level value (equivalent toapproximately 6% normobaric O2). These conditions simulated an altitudeof 31 000 ft (8100 m) and were maintained for 8 h. Ascent and descent rateswere kept o1000 ft/min. After the hypoxia period, a return to normobaricnormoxic conditions (pO2¼ 150 mm Hg in our setting) was attained inE30 min. Animals were either killed immediately after the hypobaricchamber was opened or kept at atmospheric pressure for 5 days and thenkilled. Animals kept in the chamber under normobaric normoxic conditionsserved as controls.

Chromatin immunoprecipitation analysisChromatin from normoxic parp-1þ /þ - and parp-1� /� -immortalized MEFswas fixed with formaldehyde and immunoprecipitated initially with controlIgGs and then overnight with anti-PARP-1 antibody following the protocoldescribed by Pescador et al.41 After DNA purification, the presence of theselected murine sequences was assessed by PCR amplification, and byquantitative PCR when it is indicated. The PCR products were resolved bygel electrophoresis and visualized by ethidium bromide staining.

Serum EPO and blood determinations in mice.Serum EPO was analyzed by enzyme immunoassay in an UnicellTMDxI-800 (Beckman Coulter, Barcelona, Spain). Red blood cell count andhemoglobin concentration determination were performed in a Hematol-ogy System Advia 2120 Analyzer (Siemens Diagnostics, Deerfield, IL, USA).

Statistical analysisData are expressed as means±s.e.m. Statistical comparisons between thedifferent indicated experimental groups were made by Student’s t-test,accepting Po0.05 as the level of significance.

CONFLICT OF INTERESTThe authors declare no conflict of interest.

ACKNOWLEDGEMENTSWe thank Dr Francoise Dantzer, CNRS, Strasbourg (France), for providing the pEGFP-PARP-1 construct. This work was supported by Ministerio de Ciencia e Innovacion(SAF2006-01094 and SAF2009-13281-C02-01), Fundacion La Caixa (BM06-219-0) andJunta de Andalucıa (P07-CTS-0239) to FJO; Ministerio de Educacion y Ciencia(SAF2007-64597 and SAF-2010-20067) and the BIZKAIA XEDE Program from theBizkaia County to EB.

HIF2αpVHL

HIF2αPROTEOLYSIS

HIF2αPARP1

AAAAAAA HIF2α mRNA

HIF2α PARP1

HIF targetHIF2α

PARPinhibitor

Promotes complex dissociation andpVHL-dependent HIF-2α degradation

Figure 7. Regulatory actions of PARP-1 and poly(ADP-ribose) in HIF-2-mediated hypoxic response. PARP-1 binds to de HIF-2a promoter andregulates HIF-2a (epas-1) transcription and HIF-2-dependent gene expression, including ANGPTL4 and EPO. HIF-2a and PARP-1 also form acomplex, which is stabilized by poly(ADP-ribose). PARP inhibition leads to complex dissociation and to pVHL-dependent HIF-2a degradation.

Regulation of HIF-2 by PARP-1A Gonzalez-Flores et al

7

& 2013 Macmillan Publishers Limited Oncogene (2013), 1 – 8

REFERENCES1 Semenza GL. Hypoxia-inducible factors in physiology and medicine. Cell 2012;

148: 399–408.2 Carmeliet P, Dor Y, Herbert JM, Fukumura D, Brusselmans K, Dewerchin M et al.

Role of HIF-1alpha in hypoxia-mediated apoptosis, cell proliferation and tumourangiogenesis. Nature 1998; 394: 485–490.

3 Hochachka PW, Buck LT, Doll CJ, Land SC. Unifying theory of hypoxia tolerance:molecular/metabolic defense and rescue mechanisms for surviving oxygen lack.Proc Natl Acad Sci USA 1996; 93: 9493–9498.

4 Benizri E, Ginouves A, Berra E. The magic of the hypoxia-signaling cascade. CellMol Life Sci 2008; 65: 1133–1149.

5 Wang GL, Jiang BH, Rue EA, Semenza GL. Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc Natl AcadSci USA 1995; 92: 5510–5514.

6 Pugh CW, Tan CC, Jones RW, Ratcliffe PJ. Functional analysis of an oxygen-regulated transcriptional enhancer lying 30 to the mouse erythropoietin gene.Proc Natl Acad Sci USA 1991; 88: 10553–10557.

7 Maxwell PH, Wiesener MS, Chang GW, Clifford SC, Vaux EC, Cockman ME et al. Thetumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature 1999; 399: 271–275.

8 Ivan M, Kondo K, Yang H, Kim W, Valiando J, Ohh M et al. HIFalpha targeted forVHL-mediated destruction by proline hydroxylation: implications for O2 sensing.Science 2001; 292: 464–468.

9 Jaakkola P, Mole DR, Tian YM, Wilson MI, Gielbert J, Gaskell SJ et al. Targeting ofHIF-alpha to the von Hippel-Lindau ubiquitylation complex by O2-regulatedprolyl hydroxylation. Science 2001; 292: 468–472.

10 Bruick RK, McKnight SL. A conserved family of prolyl-4-hydroxylases that modifyHIF. Science 2001; 294: 1337–1340.

11 Epstein AC, Gleadle JM, McNeill LA, Hewitson KS, O’Rourke J, Mole DR et al.C. elegans EGL-9 and mammalian homologs define a family of dioxygenases thatregulate HIF by prolyl hydroxylation. Cell 2001; 107: 43–54.

12 Lando D, Peet DJ, Gorman JJ, Whelan DA, Whitelaw ML, Bruick RK. FIH-1 is anasparaginyl hydroxylase enzyme that regulates the transcriptional activity ofhypoxia-inducible factor. Genes Dev 2002; 16: 1466–1471.

13 Tian H, McKnight SL, Russell DW. Endothelial PAS domain protein 1 (EPAS1), atranscription factor selectively expressed in endothelial cells. Genes Dev 1997; 11:72–82.

14 Ryan HE, Lo J, Johnson RS. HIF-1 alpha is required for solid tumor formation andembryonic vascularization. EMBO J 1998; 17: 3005–3015.

15 Iyer NV, Kotch LE, Agani F, Leung SW, Laughner E, Wenger RH et al. Cellular anddevelopmental control of O2 homeostasis by hypoxia-inducible factor 1 alpha.Genes Dev 1998; 12: 149–162.

16 Tian H, Hammer RE, Matsumoto AM, Russell DW, McKnight SL. The hypoxia-responsive transcription factor EPAS1 is essential for catecholamine homeostasisand protection against heart failure during embryonic development. Genes Dev1998; 12: 3320–3324.

17 Peng J, Zhang L, Drysdale L, Fong GH. The transcription factor EPAS-1/hypoxia-inducible factor 2alpha plays an important role in vascular remodeling. Proc NatlAcad Sci USA 2000; 97: 8386–8391.

18 Covello KL, Kehler J, Yu H, Gordan JD, Arsham AM, Hu CJ et al. HIF-2alpha reg-ulates Oct-4: effects of hypoxia on stem cell function, embryonic development,and tumor growth. Genes Dev 2006; 20: 557–570.

19 Schodel J, Oikonomopoulos S, Ragoussis J, Pugh CW, Ratcliffe PJ, Mole DR. High-resolution genome-wide mapping of HIF-binding sites by ChIP-seq. Blood 2011;117: e207–e217.

20 Rankin EB, Biju MP, Liu Q, Unger TL, Rha J, Johnson RS et al. Hypoxia-induciblefactor-2 (HIF-2) regulates hepatic erythropoietin in vivo. J Clin Invest 2007; 117:1068–1077.

21 Aprelikova O, Wood M, Tackett S, Chandramouli GV, Barrett JC. Role of ETStranscription factors in the hypoxia-inducible factor-2 target gene selection.Cancer Res 2006; 66: 5641–5647.

22 Koshiji M, Kageyama Y, Pete EA, Horikawa I, Barrett JC, Huang LE. HIF-1alphainduces cell cycle arrest by functionally counteracting Myc. EMBO J 2004; 23:1949–1956.

23 Gordan JD, Bertout JA, Hu CJ, Diehl JA, Simon MC. HIF-2alpha promotes hypoxiccell proliferation by enhancing c-myc transcriptional activity. Cancer Cell 2007; 11:335–347.

24 Mastrogiannaki M, Matak P, Keith B, Simon MC, Vaulont S, Peyssonnaux C. HIF-2alpha, but not HIF-1alpha, promotes iron absorption in mice. J Clin Invest 2009;119: 1159–1166.

25 Loboda A, Jozkowicz A, Dulak J. HIF-1 versus HIF-2--is one more important thanthe other? Vascul Pharmacol 2012; 56: 245–251.

26 Schreiber V, Dantzer F, Ame JC, de Murcia G. Poly(ADP-ribose): novel functions foran old molecule. Nat Rev Mol Cell Biol 2006; 7: 517–528.

27 Aguilar-Quesada R, Munoz-Gamez JA, Martin-Oliva D, Peralta-Leal A, Quiles-Perez R,Rodriguez-Vargas JM et al. Modulation of transcription by PARP-1: consequences incarcinogenesis and inflammation. Curr Med Chem 2007; 14: 1179–1187.

28 Kraus WL, Lis JT. PARP goes transcription. Cell 2003; 113: 677–683.29 Martinez-Romero R, Canuelo A, Siles E, Oliver FJ, Martinez-Lara E. Nitric oxide

modulates hypoxia-inducible factor-1 and poly(ADP-ribose) polymerase-1 cross-talk in response to hypobaric hypoxia. J Appl Physiol 2012; 112: 816–823.

30 Elser M, Borsig L, Hassa PO, Erener S, Messner S, Valovka T et al. Poly(ADP-ribose)polymerase 1 promotes tumor cell survival by coactivating hypoxia-induciblefactor-1-dependent gene expression. Mol Cancer Res 2008; 6: 282–290.

31 Yoon D, Ponka P, Prchal JT. Hypoxia and hematopoiesis. Am J Physiol 2011; 300:C1215–C1222.

32 Rosenberger C, Mandriota S, Jurgensen JS, Wiesener MS, Horstrup JH, Frei U et al.Expression of hypoxia-inducible factor-1alpha and -2alpha in hypoxic andischemic rat kidneys. J Am Soc Nephrol 2002; 13: 1721–1732.

33 Maranchie JK, Vasselli JR, Riss J, Bonifacino JS, Linehan WM, Klausner RD. Thecontribution of VHL substrate binding and HIF1-alpha to the phenotype of VHLloss in renal cell carcinoma. Cancer Cell 2002; 1: 247–255.

34 Raval RR, Lau KW, Tran MG, Sowter HM, Mandriota SJ, Li JL et al. Contrastingproperties of hypoxia-inducible factor 1 (HIF-1) and HIF-2 in von Hippel-Lindau-associated renal cell carcinoma. Mol Cell Biol 2005; 25: 5675–5686.

35 Kondo K, Klco J, Nakamura E, Lechpammer M, Kaelin Jr WG. Inhibition of HIF isnecessary for tumor suppression by the von Hippel-Lindau protein. Cancer Cell2002; 1: 237–246.

36 Kondo K, Kim WY, Lechpammer M, Kaelin Jr. WG. Inhibition of HIF2alpha issufficient to suppress pVHL-defective tumor growth. PLoS Biol 2003; 1: E83.

37 Flamme I, Krieg M, Plate KH. Up-regulation of vascular endothelial growth factorin stromal cells of hemangioblastomas is correlated with up-regulation of thetranscription factor HRF/HIF-2alpha. Am J Pathol 1998; 153: 25–29.

38 Holmquist-Mengelbier L, Fredlund E, Lofstedt T, Noguera R, Navarro S, Nilsson Het al. Recruitment of HIF-1alpha and HIF-2alpha to common target genes is dif-ferentially regulated in neuroblastoma: HIF-2alpha promotes an aggressive pheno-type. Cancer Cell 2006; 10: 413–423.

39 de Murcia JM, Niedergang C, Trucco C, Ricoul M, Dutrillaux B, Mark M et al.Requirement of poly(ADP-ribose) polymerase in recovery from DNA damage inmice and in cells. Proc Natl Acad Sci USA 1997; 94: 7303–7307.

40 Berra E, Benizri E, Ginouves A, Volmat V, Roux D, Pouyssegur J. HIF prolyl-hydroxylase 2 is the key oxygen sensor setting low steady-state levels ofHIF-1alpha in normoxia. EMBO J. 2003; 22: 4082–4090.

41 Pescador N, Cuevas Y, Naranjo S, Alcaide M, Villar D, Landazuri MO et al. Identifi-cation of a functional hypoxia-responsive element that regulates the expression ofthe egl nine homologue 3 (egln3/phd3) gene. Biochem J 2005; 390(Pt 1): 189–197.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Regulation of HIF-2 by PARP-1A Gonzalez-Flores et al

8

Oncogene (2013), 1 – 8 & 2013 Macmillan Publishers Limited