Redox Biology 41 (2021) 101885
A ailable online 1 Feb ua y 2021
2213-2317/© 2021 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Resea ch Pape
Selec i e modula ion by PARP-1 o HIF-1
α
- ec ui men o ch oma in du ing
hypoxia is equi ed o umo adap a ion o hypoxic condi ions
Juan Manuel Ma í
a
, Angel Ga cia-Diaz
a
, Daniel Delgado-Bellido
a
, F ancisco O’Valle
b
,
A iannys Gonz´
alez-Flo es
a
, Onin za Ca le a is
c
, Jos´
e Manuel Rod íguez-Va gas
d
,
Jean Ch is ophe Am´
e
d
, F ançoise Dan ze
d
, Geo ge L. King
e
, Klaudia Dziedzic
c
, Edu ne Be a
c
,
E. de ´
Ala a
, A.T. Ama al
, Es e M. Hammond
g
, F. Ja ie Oli e
a
,
*
a
Ins i u e o Pa asi ology and Biomedicine L´
opez-Ney a, CSIC, and CIBERONC, 18100, G anada, Spain
b
Pa hology Depa men , School o Medicine, IBIMER, CIBM, Uni e si y o G anada, Spain and Biosani a y Resea ch Ins i u e (IBS. GRANADA), Uni e si y o G anada,
G anada, Spain
c
CIC BioGUNE, Pa que Tecnol´
ogico de Bizkaia- Ed. 801A, 48160, De io, Spain, CIBERONC
d
Poly(ADP- ibosyl)a ion and Genome In eg i y, Labo a oi e D’Excellence Medalis, UMR7242, Cen e Na ional de La Reche che Scien i ique/Uni e si ´
e de S asbou g,
Ins i u de Reche che de L’Ecole de Bio echnologie de S asbou g, Boule a d S. B an , BP10413, 67412, Illki ch, F ance
e
Sec ion o Vascula Cell Biology and Complica ions, Dianne Nunnally Hoppes Labo a o y o Diabe es Complica ions, Joslin Diabe es Cen e , Ha a d Medical School,
Bos on, MA, USA
Ins i u e o Biomedicine o Se illa (IBiS), Vi gen Del Rocio Uni e si y Hospi al/CSIC/Uni e si y o Se illa/CIBERONC, Se ille, Spain
g
Ox o d Ins i u e o Radia ion Oncology, Depa men o Oncology, Uni e si y o Ox o d, Ox o d, UK
ARTICLE INFO
Keywo ds:
Hypoxia
PARP-1
PARyla ion
ChIP-seq
Tumo mic oen i onmen
ABSTRACT
Backg ound: The adap a ion o hypoxia is mainly con olled by he HIF ansc ip ion ac o s. Inc eased exp es-
sion/ac i i y o HIF-1
α
co ela es wi h poo p ognosis in cance pa ien s. PARP-1 inhibi o s a e used in he clinic
o ea BRCAness b eas /o a ian cance and ha e been shown o egula e he hypoxic esponse; he e o e, hei
use could be expanded.
Me hods: In his wo k by in eg a ing molecula /cell biology app oaches, genome-wide ChIP-seq, and pa ien
samples, we elucida e he ex en o which PARP-1 exe s con ol o e HIF-1- egula ed genes.
Resul s: In human melanoma, PARP-1 and HIF-1
α
exp ession a e s ongly associa ed. In esponse o a hypoxic
challenge poly(ADP- ibose) (PAR) is syn hesized, HIF-1
α
is pos - ansc ip ionally modi ied (PTM) and s abilized
by PARyla ion a speci ic K/R esidues loca ed a i s C- e minus. Using an unbiased ChIP-seq app oach we
demons a e ha PARP-1 dic a es hypoxia-dependen HIF- ec ui men o ch oma in in a ange o HIF- egula ed
genes while analysis o HIF-binding mo i s (RCGTG) e eals a es ic ion on he ecogni ion o hypoxia
esponsi e elemen s in he absence o PARP-1. Consequen ly, he cells a e poo ly adap ed o hypoxia, showing a
educed i ness du ing hypoxic induc ion.
Conclusions: These da a cha ac e ize he ine- uning egula ion by PARP-1/PARyla ion o HIF ac i a ion and
sugges ha PARP inhibi o s migh ha e he apeu ic po en ial agains cance ypes displaying HIF-1
α
o e -
ac i a ion.
1. In oduc ion
Hypoxia is a common e en du ing umo de elopmen consequence
o accele a ed umo g ow h. When he mass exceeds a olume o a ew
mm
3
, egions o low oxygen concen a ion occu in he inne pa s o he
umo . Unde his si ua ion cells mus modi y hei me abolism o cope
wi h his new en i onmen al con ex . The adap a ion o he hypoxic
si ua ion in ol es he exp ession o hund eds o genes implica ed in he
main enance o cellula su i al h ough me abolic adap a ion. These
adap a ions include new essels o ma ion [1], glycolysis ac i a ion [2],
cance s em cells (CSCs) egula ion [3] and e en umo exosome p o-
duc ion [4]. All hese changes acili a e cell su i al, umo g ow h,
mig a ion, and me as asis [5–7]. The hypoxic esponse is associa ed
wi h poo o e all su i al, lowe disease- ee su i al, and diminished
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (F.J. Oli e ).
Con en s lis s a ailable a ScienceDi ec
Redox Biology
jou nal homepage: www.else ie .com/loca e/ edox
h ps://doi.o g/10.1016/j. edox.2021.101885
Recei ed 23 Decembe 2020; Recei ed in e ised o m 27 Janua y 2021; Accep ed 28 Janua y 2021
Redox Biology 41 (2021) 101885
2
loco- egional con ol [8–10].
This ansc ip ional induc ion is media ed by he hypoxia-inducible
ac o s (HIFs). This amily o ansc ip ion ac o s a e ac i e as an
alpha/be a he e odime which binds o nucleo ides sequences, being he
mos common known as HRE (5′-RCGTG-3’) o hypoxia esponse
elemen , localized in he p omo e s o he hypoxic exp essed genes,
inducing hei ansc ip ion [11].
The HIF amily is composed o 1 be a chain (HIF-1β) and 3 alpha
subuni s (HIF−1
α
, HIF-2
α
and HIF-3
α
). The ine- uning o he HIF
esponse depends on he alpha chain, which induces he exp ession o
di e en gene ic pa e ns, allowing he cell o espond e icien ly o
di e en hypoxic in ensi ies and du a ions. The hypoxic esponse is
egula ed unde well oxygena ed condi ions ia HIF-
α
deg ada ion.
P olyl hyd oxylase domain p o eins (PHDs) hyd oxyla e p oline esi-
dues on he alpha subuni s speci ically du ing no moxia. This hyd ox-
yla ion is ecognized by he on Hippel-Lindau (VHL) E3 ligase, causing
he ubiqui ina ion o he alpha subuni s and hei subsequen deg ada-
ion ia he p o easome [12]. Al hough his is he majo mechanism
con olling HIF s abili y and ac i i y, many o he PTM including phos-
pho yla ion, ace yla ion, SUMOyla ion and hyd oxyla ion ha e been
desc ibed as modula o s o HIF induc ion/ac i i y [13]. The C- e minus
domain o HIF-1
α
has gained inc easing a en ion as a egula o y si e o
HIF-1
α
ac i a ion. This domain unde goes hyd oxyla ion o an aspa a-
gine by FIH ( ac o inhibi ing HIF), educing he ec ui men o
co-ac i a o s such as he p300 and hus educing he exp ession o
hypoxic genes in di e en cance ypes [14,15].
The Poly(ADP- ibose) polyme ases (PARP) p o eins a e a amily o
in acellula enzymes cha ac e ized by he p esence o a domain
e e ed as “PARP signa u e”. Subs i u ions on his domain make only
he membe s PARP-1, 2 and anky ases able o syn hesize a polyme o
poly(ADP-Ribose) o PAR, while he es o he amily is inac i e o
gene a es mono(ADP-Ribose) o MAR [16]. Bo h PAR and MAR equi e
NAD
+
o be consumed as a subs a e in an ATP-dependen eac ion [17].
The inal p oduc o he PARP enzyma ic ac i i y (mono o poly
(ADP- ibose)) is hen co alen ly bound on o Glu, Asp, Lys and Se [18].
PARPs hemsel es can unde go his modi ica ion (au o modi ica ion) o
hey can modi y o he p o eins (he e o modi ica ion). These changes a e
conside ed a ype o e e sible PTM and a e igh ly egula ed by en-
zymes ha clea e he linkage be ween ADP- ibose uni s and he ac-
cep o s (e.g. ADP- ibosyl-accep o hyd olases (ARH3, Mac oD1,
Mac oD2) o poly(ADP- ibose) glycohyd olases (PARG)).
Th ough PARyla ion, PARPs egula e a wide a ay o cellula p o-
cesses, including genomic ins abili y and su i al [19], ch oma in o -
ganiza ion [20], p o ein deg ada ion ia p o easome [21], RNA
me abolism [22] and cell dea h [23,24].
PARP-1 was he i s desc ibed and mos ac i e membe o he
amily, gene a ing up o he 90% o he polyme obse ed in he cell
[25]. I is known p ima ily o i s impo an ole du ing DNA epai .
Owing o his p ope y, he inhibi ion o PARP ac i i y is ecognized as a
po en ial he apeu ic s a egy o enhance he cy o oxic ac ion o
an i-cance d ugs o adio he apy, and o he ea men o cance s wi h
speci ic DNA epai de ec s [26,27]. PARP-1 has also been desc ibed as a
egula o o he hypoxic esponse [28–30]; howe e , he signal linking
he selec i e egula ion ia PARP-1 o HIF-dependen genes is no
known. In his s udy we desc ibe a new pa hway demons a ing ha
ea ly du ing he esponse o hypoxia, inc eased poly(ADP- ibose) ac-
i i y leads o PARyla ion o HIF-1
α
a speci ic esidues loca ed a i s
C- e minus domain. This PARyla ion di ec s HIF-1
α
o a selec i e sub-
g oup o genes (mos ly excluding glucose me abolism- ela ed genes)
sugges ing a hie a chy o a oid o e lap in he egula ion in
HIF-dependen gene exp ession and o p o ec c i ical su i al unc ions
du ing hypoxia.
2. Ma e ials and me hods
2.1. Cul u es and g ow h medium
HEK 293T cells we e cul u ed using Dulbecco’s Modi ied Eagle me-
dium low glucose supplemen ed wi h L-glu amine 4 mM, MEM non-
essen ial amino acids 0.1 mM, 10% hea -inac i a ed e al bo ine
se um (FBS) plus penicillin (50 IU/ml) and s ep omycin (50 mg/ml).
HepG2 cells we e cul u ed using Dulbecco’s Modi ied Eagle medium
high glucose supplemen ed wi h 10% hea inac i a ed FBS plus peni-
cillin (50 IU/ml) and s ep omycin (50 mg/ml). COS cells we e cul u ed
in Dulbecco’s Modi ied Eagle medium low glucose supplemen ed wi h
10% o hea inac i a ed FBS plus penicillin (50 IU/ml) and s ep omycin
(50 mg/ml). All cells we e g own using a egula incuba o a 37 ◦C in a
humidi ied 21% O
2
and 5% CO
2
a mosphe e. Hela cells we e cul u ed on
D-MEM (high glucose), 10% hea inac i a ed FBS, 0.1 mM MEM
Nonessen ial Amino Acids (NEAA), 2 mM L-glu amine, 1% Pen-S ep.
C8161 we e cul u ed on RPM11640 (GIBCO) supplemen ed wi h 10%
hea inac i a ed FBS, glu amine (0.8 Mg/ml), and gen amicin (10 ng/
ml). MUM2b cells we e g own on Eagle’s minimum essen ial medium
(EMEM) (BioWhi ake , Walke s ille, MD) supplemen ed wi h 10%
hea -inac i a ed FBS (Fishe , On a io, ON, Canada), 1% nonessen ial-
amino acids (NEAA-Mix u e, 100; BioWhi ake ) 2 mmol/L L-glu a-
mine, and penicillin/s ep omycin.
2.2. T ea men s
The PARP inhibi o PJ34 was pu chased om Enzo Li e Sciences
(San Diego, CA, USA) and was used a a concen a ion o 10
μ
M, 90 min
be o e hypoxia. The PARP inhibi o olapa ib was used a 5
μ
M, 90 min
be o e hypoxia and pu chased om Selleckchem (Hous on, USA). The
ROS inhibi o MPG-2, also e e ed as N-(2-Me cap op opionyl) glycine,
was pu chased om Sigma-Ald ich (San Louis. Missou i, USA) and was
used a 300
μ
M 90 min be o e he exposu e o di e en imes o hypoxia.
Mi ochond ial an ioxidan Mi oTempo was pu chased om Sigma-
Ald ich (San Louis. Missou i, USA) and used a a inal concen a ion o
20
μ
M. Cells we e p e ea ed o 120 min and hen exposed o hypoxia.
2.3. Hypoxia
Hypoxic incuba ion was achie ed using a sealed hypoxic wo ks a ion
(Ruskin, B idgend, UK). Cells we e exposed o 1% O
2
and 5% CO
2
a
37 ◦C. Hypoxic du a ion was 4 h i a di e en du a ion is no indica ed.
2.4. Wes e n Blo
Fo Wes e n Blo analysis, cells we e pla ed in six-well pla es a a
densi y o 2.5 ×10
5
, lysed using TRE bu e and hen he whole cell
ex ac was sonica ed, esuspended, and boiled o 5 min in modi ied
Laemli cha ge bu e (250 mM T is-HCl (pH 75), glyce ol 20%, SDS
10%, 1,4 M o me cap oe hanol and 1% blue b omophenol). Samples
we e ans e ed o a ni ocellulose memb ane h ough we ans e
(Ame sham Biosciences). P o eins we e hen isualized using he ECL
sys em (Ame sham Biosciences) a e using speci ic an ibodies o : HIF-
1
α
(Be hyl), GST an ibodies (Be hyl), Tubulin (Sigma-Ald ich), Ac in
(Sigma-Ald ich), Myc- ag (Cell Signaling Technology), Poly(ADP-
ibose) (T e igen), PARP-1 (Enzo), RPA (Cell Signaling), p-RPA
(Be hyl), P53 (San a C uz), p-P53 (Millipo e), H2AX (Millipo e), p-
H2AX (Millipo e), His one H1 (San a C uz), Laminin B (Abcam).
2.5. T ans ec ion assay
Cells we e cul u ed in six-well pla es a a densi y o 2.5 ×10
5
cells
pe well. 24 h la e hey we e ans ec ed using he eagen je PRIME
(Polyplus, Illki ch, F ance) using he concen a ions indica ed by he
manu ac u e o he solu ions and plasmids o in e es . Cells we e
J.M. Ma í e al.
Redox Biology 41 (2021) 101885
3
cul u ed o 24 h be o e being exposed o he eac i e. Fou hou s la e
he medium was e eshed and one day la e cells we e lysed using TR3
bu e . Fo pull down and immunop ecipi a ion assays, cells we e pla ed
in p100 pla es a 1.5 10
6
concen a ion. Then he je PRIME p o ocol was
comple ed in he same way ollowing Polyplus ins uc ions.
2.6. Pull down assay
Cells we e co- ans ec ed wi h Myc-HIF1
α
C- e and di e en GST-
PARP-1 domains. A e ollowing he je PRIME p o ocol hey we e
ha es ed in 20 nM T is-HCl (pH 7.5), 400 mM NaCl, 20% glyce ol, 5
mM DTT, 0,5 mM pe abloc and p o ease inhibi o s (Comple e Mini;
Roche, Mannheim, Ge many). Lysa es we e clea ed by cen i uga ion
and incuba ed o 2 h wi h glu a hione-sepha ose 4B (Sigma, S Louis,
MO, USA) Beads we e washed h ee imes wi h 20 nM T is–HCl (pH 7.5),
150–500 nm NaCl, 0.1% NP-40 and p o ease inhibi o s. All samples
we e esuspended and boiled o 5 min in modi ied Laemli cha ge bu e
(250 mM T is-HCl (pH 7.5), glyce ol 20%, SDS 10%, 1,4 M me cap-
oe hanol and 1% blue b omophenol). Then samples we e analyzed by
wes e n blo . Blo s we e subsequen ly incuba ed wi h he an i-GST
an ibody.
2.7. I adia ion
Cell cul u es we e i adia ed in an Xs ahl RS225 cabine using issue
lasks a oom empe a u e. A ol age o 195 kV X- ays was used a a
dose a e o 1.6 Gy/min un il a inal dose o 5Gy was accumula ed on he
sample.
2.8. Immunop ecipi a ion
Fo IP, 2x10
6
cells pe condi ion we e incuba ed. A e unde going
he di e en ea men s, hey we e collec ed and exposed o ice o 20
min on a lysis bu e (1% NP-40, 1 mM EGTA (pH 8.0), 100 mM NaCl, 1
mM sodium o ho anada e, 20 mM T is-Cl (pH 7.6), 10 mM sodium
py ophospha e, 10 mM NaF, 500 mM phenylme hyl sul onyl luo ide
and cock ail o p o eases inhibi o s (Comple e Mini, Roche, Zu ich,
Swi ze land)). Lysa es we e incuba ed wi h P o ein A Sepha ose o
educe non-speci ic signaling. Then, he p e-clea ed supe na an s we e
isola ed and incuba ed wi h he an ibodies o in e es plus P o ein A
Sepha ose o e nigh a 4 ◦C. The ex ac s we e washed i e imes wi h
saline bu e (100 mM NaCl, 20 mM T is-HCl (pH 8.0), 0,5% NP-40, 1
mM EDTA, and p o ease inhibi o s). The immunop ecipi a ed com-
plexes we e isola ed by cen i uga ion and boiled o 5 min in modi ied
Laemli cha ge bu e (250 mM T is-HCl (pH 7.5), glyce ol 20%, SDS
10%, 1,4 M me cap oe hanol and 1% blue b omophenol). Then, samples
we e analyzed by wes e n blo .
2.9. Immunos aining
On 6-well pla es, s e ilized co e slips we e placed, hen 2.5 ×10
5
cells we e seeded on op o hem. A e g owing o e nigh cells we e
exposed o di e en ea men s. They we e hen washed gen ly wi h PBS
1x and ixed using 3% pa a o maldehyde du ing 15 min a oom em-
pe a u e. Then, hey we e washed using PBS 1x and pe meabilized wi h
0.25% T i on o 10 min. They we e washed again wi h PBS 1x and
blocked wi h BSA 2% du ing 1 h a oom empe a u e. Samples we e
incuba ed wi h BSA 2% using an i-FLAG (Sigma Ald ich) o an i-53BP1
(Milipo e) (1:100) o 45 min a 37 ◦C. A seconda y an ibody, an i- abbi
o an i-mouse, was incuba ed wi h BSA 2% o 20 min a 37 ◦C. Cells
we e washed wi h PBS 1x and hen sealed wi h Vec ashield an i ade
moun ing medium +DAPI. Images we e ob ained using con ocal mi-
c oscopy (SP5 Con ocal Leica Mic oscope). De ec ing he an i- abbi
signal a 488 nm and he an i-mouse a 647 nm.
2.10. Real ime q-PCR
Cells we e ha es ed and he RNA isola ed ollowing he ins uc ions
p o ided o he RNeasy mini ki (Qiagen, Hilden, Ge many). Re e se
T ansc ip ion was pe o med using he iSc ip Re e se T ansc ip ion
Supe mix o RT-qPCR (Bio-Rad, He cules, USA). Real ime PCR was
pe o med on a CFX96 he mocycle using he iTaq™ Uni e sal SYBR®
G een Supe mix (Bio-Rad, He cules, USA). The p ime s we e syn he-
sized by he “PCR P ime Design” se ice (Sigma-Ald ic, Sain Louis,
USA). 36B4 was used as housekeeping gene. The lis o p ime s goes as
ollow:
h HIF1
α
Fw CTGCAACATGGAAGGTATTGCA
h HIF1
α
R TACCCACACTGAGGTTGGTTACTG
h CAIX Fw TAAGCAGCTCCACACCCTCT
h CAIX R TCTCATCTGCACAAGGAACG
h LDHA Fw TGGGAGTTCACCCATTAAGC
h LDHA R AGCACTCTCAACCACCTGCT
h ANGPTL4 Fw CGTACCCTTCTCCACTTGGG
h ANGPTL4 R GCTCTTGGCGCAGTTCTTG
h GLUT-1 Fw CAGTTTGGCTACAACACTGGAGT
h GLUT-1 R ATAGCGGTGGACCCATGTCT
h VEGFA Fw GGGCAGAATCATCACGAAGT
h VEGFA R TGGTGATGTTGGACTCCTCA
h DDIT4 Fw: GACAGCAGCAACAGTGGCTTC
h DDIT4 R : CCACGCTATGGCAGCTCTTGC
h UBE2M Fw: CCTGCCCAAGACGTGTGATA
h UBE2M R : CCCTGGCCCACCTTAAAACT
h GAPDH Fw: ACAGTCAGCCGCATCTTCTT
h GAPDH R : ACGACCAAATCCGTTGACTC
2.11. ROS de e mina ion
To measu e ROS, DCFA-Cellula Reac i e Oxygen Species De ec ion
Assay Ki (AbCam) was used. Cells we e seeded a a concen a ion o 2.5
×10
4
pe poin o 96 wells, 24 h la e hey we e s ained using 25
μ
M o
DCFDA o 45 min. Nex , he DCFDA was washed and cells we e exposed
o he di e en ea men s. ROS we e measu ed by luo escence a Ex
485 ±20 nm Em 535 ±25 nm using a TECAN pla e eade In ini e PRO
200.
To con i m he ROS induc ion obse ed du ing hypoxia we used a
non DCFDA- ela ed me hod o ROS de ec ion based on he molecule
DHE (Dihyd oe hidium). Following he ins uc ions p o ided by he
DHE Assay Ki (Abcam) 2.5 ×10
4
we e seeded pe poin on a 96 well.
A e 24 h hey we e washed using PBS 1X and hen exposed o DHE 5
μ
M o 90 min a 37 ◦C in he da k. The DHE was hen emo ed and he
expe imen s pe o med. Using he TECAN pla e eade In ini e PRO 200,
DHE luo escence was de ec ed a a Ex 550 ±10 nm and Em 590 ±10
nm.
2.12.
32
P-PAR syn hesis o he in i o PARyla ion assay
Pu i ied
32
P-PAR was p epa ed by incuba ing 20
μ
g o pu i ied
human PARP-1 (Alexis) in a mix u e con aining 50 mM T is-HCl (pH
8.0), 4 mM MgCl
2
, 0.15 M NaCl, 1 mM DTT, 0.5. DNase I ea ed-DNA,
0.1 mM BSA, 0.4 mM NAD
+
, 0.5 Ci [
32
P]-NAD
+
(800 Ci/mmol, 5 Pe -
kinElme ) in a o al 500
μ
l eac ion. PARP-1 and he mix u e we e
incuba ed o 1 h a 32 ◦C. Then, 2
μ
l o 10 mg/ml DNAse I (Roche) and
2
μ
l o 1 M CaCl
2
we e added and incuba ed o 1 h a 37 ◦C. Following
his s ep, 250
μ
l H
2
O, 5
μ
l 20% SDS and 20
μ
l o 10 mg/ml P o einase K
we e added o he mix and incuba ed o 4 h a 37 ◦C. P o eins we e hen
ex ac ed using a phenol/CHCl
3
ea men : 500
μ
l phenol/CHCl
3
we e
added, o exed and cen i uga ed o 5 min a 4 ◦C. The uppe phase
was eco e ed, and he PAR p ecipi a ed by adding 100
μ
l o 3 M po-
assium ace a e and 1 ml isop opyl alcohol. The samples we e hen le
o p ecipi a e a −20 ◦C o e nigh , hen cen i uged in low binding
J.M. Ma í e al.
Redox Biology 41 (2021) 101885
4
Eppendo ubes a 14,000 pm a 4 ◦C 1 h. Ca e ully he liquid was
emo ed, and he p ecipi a e was washed wi h 1 ml 80% e hanol. The
PAR was le hen o d y a oom empe a u e o 1 h. Finally, he PAR
was esuspended in 400
μ
l TE bu e .
2.13. PAR binding assay on pep ides
Bio inyla ed pep ides we e o de ed a Gensc ip . D ied pep ides we e
esuspended in H
2
O o 3
μ
g/
μ
l 3
μ
g o pep ides we e bound o 15
μ
l
slu y o s ep a idin-magne ic beads (Millipo e) on ice o 1 h wi h
equen agi a ion using 1.5 ml Eppendo ubes. Beads we e washed
wice in 1 ml PBS 1x. The beads-pep ides complexes we e incuba ed
wi h 50,000 cpm
32
P-PAR in 200
μ
l PBS 1x o 1 h on ice wi h agi a ion.
The beads we e hen washed h ee imes wi h 1 ml PBS 1x, and he ubes
we e changed a e each wash o ensu e ha no adioac i i y was bound
non-speci ically on o he plas ic o he ubes. A e he las wash, he
supe na an was emo ed almos o d y he beads and he adioac i i y
bound o he beads was e alua ed by Ce enko coun ing using a Packa d
coun e .
2.14. CRISPR-Cas9 cells
The p oduc ion o HEK 293T cells PARP-1 KO was pe o med hanks
o he use o CRISPR-Cas9 echnology. Di e en sgRNAs we e selec ed
using he “Zhang Lab Op imized CRISPR design ool” and hen cloned
in o he pL-CRISPR.EFS.GFP ob ained om Addgene (#57818). Cells
we e ans ec ed using Je P ime eagen s, hen he GeneA Genomic
Clea age De ec ion Ki (In i ogen) was used o alida e he sgRNAs
ollowing he manu ac u e ’s ins uc ions. The guides numbe 2 and 4
we e selec ed o , allelic dis up ion.
Guide 2: GAGTCGAGTACGCCAAGAGC
Guide 4: GCATCCCCAAGGACTCGCTC
These guides we e ans ec ed again, he GFP posi i e cells we e
isola ed and a clone selec ion was pe o med a e single-cell sepa a ion
h ough low cy ome y. Finally, PARP-1 KO s a us was con i med by
Sange sequencing and Wes e n Blo .
2.15. ChIP-seq
The ch oma in immunop ecipi a ion was pe o med using he Sim-
pleChIP enzyma ic Ch oma in IP Ki wi h magne ic beads #9003 (Cell
SignalingTechnology, Dan e s, USA). Following he p o ide in-
s uc ions, 2 ×10
7
cells we e seeded and 5
μ
g o HIF−1
α
an ibody
ab2185 (AbCam, Camb idge, UK) we e used pe condi ion. To es IP
e iciency, a quan i a i e PCR was pe o med o he hypoxic-induced
gen EGLN3. A e wa d, he Pe cen Inpu was calcula ed o measu e
his exp ession as a pe cen o he global inpu ch oma in.
2.16. ChIP-seq sequencing, da a alignmen , and peak calling
Immunop ecipi a ed DNA was p ocessed o sequencing using s an-
da d p o ocols and sequenced on an Illumina Nex seq 500 wi h 75-bp
single end eads. Two biological eplica es o each ChIP in hypoxic
condi ions, and one biological eplica e o he emaining no moxic
samples we e sequenced ob aining 21,064,518 as he a e age ead
numbe . Quali y assessmen and samples alignmen we e pe o med
using he miARma-Seq pipeline [31,32]. In de ail, as qc [33] we e
applied o ga he he o e all sequence quali y and possible adap e
accumula ion. Then, all samples we e aligned o he e e ence Human
GRCH38 genome om he GENCODE po al, e sion 28. All samples
we e aligned using Bu ows-Wheele Aligne so wa e [34]. PCR du-
plica es we e emo ed using sam ools ma kdup wi h op ions –s – [35].
Peak calling was pe o med using MACS2 [36] wi h op ion – BAM –g hs
–s 75 –B –SPMR –nomodel.
All ChIP-Seq aw da a a e a ailable in he Gene Exp ession Omnibus
(GEO) eposi o y: GEO accession GSE144189.
2.17. Immunohis ochemis y
Fou -mic ome e - hick issue sec ions om pa a in blocks (mela-
noma in si u n =11 and me as asis n =12) we e dewaxed in xylene and
ehyd a ed in a se ies o g aded alcohols. Sec ions we e imme sed in 3%
H
2
O
2
aqueous solu ion o 30 min o exhaus endogenous pe oxidase
ac i i y, hen co e ed wi h 1% blocking eagen (Roche, Mannheim,
Ge many) in 0.05% Tween 20-PBS, o block nonspeci ic binding si es.
An igen e ie al was pe o med using a p essu e cooke and EDTA
bu e (pH 8.0). Sec ions we e incuba ed wi h p ima y an ibodies du ing
1 h o an i-PARP1 (#BML-SA250-0050, Enzo Li e Sciences, Fa m-
ingdale, NY) (1/300) o o e nigh a 4 ◦C o an i- HIF-1
α
(#A300-286A,
polyclonal an i–HIF–1
α
(Be hyl, Mon gome y, TX, USA) (1/100).
Pe oxidase-labelled seconda y an ibodies and 3,3-diaminobenzidine
we e applied o de elop immuno eac i i y, acco ding o manu ac-
u e ’s p o ocol (EnVision; Dako, Glos up, Denma k). Slides we e hen
coun e s ained wi h hema oxylin and moun ed in DPX (BDH Labo a-
o ies, Poole, UK). Sec ions in which p ima y an ibody was omi ed we e
used as nega i e con ols. Immunos aining was e alua ed by an expe-
ienced pa hologis . The pe cen age o immunos ained umo cells was
sco ed as ollows: 0, nega i e; <19%, weak: and >20%, posi i e.
Rep esen a i e images we e acqui ed in a mic oscope (Olympus BX-61).
2.18. Glycolysis assay
The glycolysis assay was pe o med ollowing he ins uc ions p o-
ided o he p oduc “Glycolysis assay” ab197244 (Abcam). This assay
measu ed ex acellula acidi ica ion due o lac a e p oduc ion du ing
glycolysis. As ins uc ed, cells (8 ×10
4
) we e cul u ed in 96-well pla es
o e nigh and hen pu ged om CO
2
on a CO
2
- ee incuba o a 37 ◦C o
3 h. Then 150
μ
l o he Respi a ion bu e we e added in combina ion wi h
10
μ
l o he Glycolysis Assay Reagen , which con ained a cell impe me-
able, pH-sensi i e luo opho e. Then cells we e exposed o hypoxia and
a e his s ep, pla es we e aken o a TECAN luo escen spec oscope,
whe e luo escence (Ex 380 ±40 nm Em 620 ±10 nm) was measu ed
e e y 5 min o 1 h.
2.19. O
2
consump ion assay
The assay was pe o med ollowing he ins uc ions p o ided by he
supplie o he p oduc “Ex acellula Oxygen Consump ion Assay”
ab197242 (Abcam). This assay ga e a eal- ime kine ic analysis o he
oxygen consump ion, in o ming o cellula espi a ion a e and mi o-
chond ial ac i i y. Cells we e cul u ed a a concen a ion o 8 ×10
4
pe
condi ion and le o g ow o e nigh . The nex day we e added 10
μ
l o
he Ex acellula oxygen consump ion eagen , which con ained a luo-
escen dye ha was quenched in he p esence o oxygen. Then cells
we e exposed o hypoxia. Be o e oxygen de e mina ion, 100
μ
l o he
High sensi i i y mine al oil we e added on op o he assay medium,
limi ing oxygen di usion in o he medium. On his con ex mi ochon-
d ial ac i i y educed oxygen concen a ion, educing he dye quench-
ing. O
2
consump ion was measu ed by luo escen mic oscopy on a
TECAN luo escen spec oscope (Ex 380 ±20 nm Em 650 ±20 nm)
e e y 5 min o 1 h.
3. Resul s
3.1. HIF-1
α
s abili y depends on PARP-1 PARyla ion ac i i y
P e ious esul s om di e en g oups (including ou s) [29,30,37]
ha e shown a connec ion be ween PARP-1 and HIF-1
α
ac i a ion.
Howe e , he mechanism connec ing bo h p o eins, he ex en o which
PARP-1 de e mines HIF-dependen gene ec ui men / ansc ip ion, and
J.M. Ma í e al.
Redox Biology 41 (2021) 101885
5
he ele ance in cance pa ien s, has no been explo ed ye . In o de o
assess he clinical ele ance o he associa ion be ween HIF-1
α
and
PARP-1, using he TCGA da abase we analyzed he co ela ion in gene
exp ession in me as a ic [38], ac al [39] and u eal melanoma, whe e
hypoxia plays an impo an ole du ing umo de elopmen [40]. Using
cBiopo al [41,42] o que y his associa ion we ound a s a is ically
signi ican posi i e co ela ion be ween PARP-1 and HIF-1
α
gene
exp ession (Fig. 1a). To u he suppo he p e ious da a, we cha ac-
e ized he exp ession o bo h p o eins in biopsies om in si u melanoma
pa ien s using a issue mic o a ay. PARP-1 and HIF-1
α
exp ession was
co-inciden in consecu i e sec ions and o e exp essed when compa ed
o he adjacen heal hy issue (Fig. 1b). The eliance o HIF-1
α
on
PARP-1 du ing ea ly hypoxia was es ed in HeLa cells, me as a ic
(C8161), and u eal (MUM2B) melanoma. PARP inhibi ion wi h olapa ib
led o HIF-1
α
deple ion in all cases (Fig. 1c). To es i his e ec was
speci ically PARP-1-dependen , we con i med ha PARP-1 silencing
caused he same deple ion on HIF-1
α
accumula ion (Fig. 1d). To
disc imina e i HIF-1
α
s abili y depended on PARP-1 i sel o on i s
PARyla ion ac i i y, we obse ed ha a e PARG silencing HIF-1
α
was
induced in pa allel wi h polyme accumula ion (mainly syn hesized by
PARP-1) (Fig. 1e). These esul s sugges ed ha HIF-1
α
s abili y was
a ec ed no only by he p esence PARP-1, bu also by PARP ac i i y.
3.2. Hypoxia-induced ROS leads o PARP ac i a ion and HIF-1
α
accumula ion
To unde s and HIF-1
α
and PARP-1 in e ac ion, we ocused in he
signals a ising du ing ea ly hypoxia (4 h, 1% O
2
). This is he case o
eac i e oxygen species (ROS), ha ha e been desc ibed o inc ease e y
ea ly a e acu e hypoxia [43]; ROS a e a amily o molecules known o
igge he poly (ADP- ibose) ac i i y [44]. A e di e en minu es o
hypoxia ROS, we e measu ed using wo di e en p obes: DCFA (Fig. 2a)
and DHE (Figu e S1b). In bo h cases, ROS peaked du ing he i s hou
and hei induc ion was p e en ed using he gene al ROS sca enge
MPG-2 (N-(2-me cap op opionyl) glycine) (Fig. 2a and S1a). Du ing
ea ly hypoxia PAR le els we e apidly up- egula ed in pa allel wi h
HIF-1
α
accumula ion. The inhibi ion o PAR syn hesis comple ely p e-
en ed PAR accumula ion, consequen ly educing HIF-1
α
s abiliza ion
(Fig. 2b and S1c). Cells ea ed wi h he ROS sca enge also displayed
dec eased PAR syn hesis and educed HIF-1
α
accumula ion (Fig. 2c and
Fig. 1. a Co ela ion o gene In i o s udy o PARP1 and HIF-1
α
exp ession in melanoma pa ien s using he publicly a ailable da abase cBiopo al; op panel,
me as a ic melanoma (CM Spea man: 0.50 (p =8.722e-3)Pea son:0.42 (p =0.0309)); middle panel, u eal melanoma (UM Spea man: 0.30 (p =6.295e-3), Pea son:
0.33 (p =2.655e-3)); lowe panel, ac al melanoma AM Spea man: 0.55 (p =4.636e-4)Pea son: 0.59 (p =1.516e-4)). b Tissue a ay on 3 melanoma pa ien s,
immunochemis y analysis is pe o med o obse e PARP-1 and HIF1
α
exp ession. Consecu i e sec ions a e shown. c Th ee di e en cell lines (MUM2B, C8161 and
Hela) show a educ ion on HIF-1
α
a e PARP inhibi ion using Olapa ib 5
μ
M du ing hypoxia 4 h. d T ansien silencing o PARP-1 on HEK 293T cells shows an
impai men on HIF-1
α
accumula ion du ing ea ly (4 h) hypoxia. e In Hela cells polyme accumula ion is induced on PARG silenced cell. A e 4 h o hypoxia HIF-1
α
is
mo e s able on he polyme enhanced con ex .
J.M. Ma í e al.

Redox Biology 41 (2021) 101885
6
S1d), un eiling he impo ance o ROS induc ion o PARP-1 ac i a ion,
and subsequen ly, o HIF-1
α
s abiliza ion. Using he mi ochond ial ROS
sca enge Mi oTempo we we e unable o educe he ROS induc ion
du ing hypoxia, implying ha maybe o he ROS sou ces (non--
mi ochond ial) we e in ol ed (Figu e S1e).
As o al HIF-1
α
le els we e dec eased a e PARP inhibi ion, we
asked i HIF-1
α
ansc ip ional ac i i y was also down- egula ed. This
was indeed he case, as e lec ed by he exp ession o i e di e en genes
known o be HIF-1
α
a ge s: CAIX, ANGPTLA4, GLUT1, VEGF and LDH
(Fig. 2d–h). A signi ican down- egula ion was ound o all es ed genes
a e ea ing wi h ei he ROS sca enging o PARP inhibi ion, wi h
excep ion o LDHA a e 24 h o hypoxia ollowing PJ34 ea men . Non-
signi ican al e a ions on HIF-1
α
mRNA le els we e desc ibed (Fig. 2i).
These da a s eng hen he idea o a igh link be ween hypoxia, ROS
p oduc ion, PARyla ion induc ion, and HIF-1
α
s abiliza ion and
ac i a ion.
3.3. PARP-1 in e ac s wi h he C- e minus domain o HIF-1
α
egula ing
he s abili y o he p o ein
I has been p e iously shown ha PARP-1 and HIF-1
α
o m a com-
plex a e ciclopi ox olamine ea men (hypoxia mime ic) and in
esponse o Eps ein Ba i us (EBV) in ec ion [28,29,37]. Howe e , his
in e ac ion and he domains media ing i s o ma ion ha e no been
add essed in he hypoxic con ex , whe e HIF-1
α
pe o ms i s majo ac-
i i y. To u he dissec he in luence o PARP-1 in HIF−1
α
s abili y
du ing no moxia and hypoxia, we gene a ed a double mu an HIF-1
α
o
he PHD si es P
402
and P
564
. These p olines we e changed in o alanines
using si e-di ec ed mu agenesis on wo cons uc s spanning he HIF-1
α
ull leng h (DML: 1–826), and a unca ed HIF-1
α
lacking he C- e minus
domain (DMS: 1–657). In addi ion, we gene a ed a C- e minus agmen
o HIF-1
α
(C- e : 630–826) ha lacks si es o PHD egula ion (Fig. 3a).
pVHL o e exp ession a ec ed he s abili y o he wild ype ull leng h
and wild ype sho HIF-1
α
, bu no o he C- e domain (Fig. 3b, c and
d). As expec ed, when he PHD si es we e mu a ed, he DML and DMS
became insensi i e o pVHL o e -exp ession. The e o e, bo h cons uc s
we e s able du ing no moxia (Fig. 3e and ). We hen exposed cells
ans ec ed wi h he di e en cons uc s o PARP inhibi ion, his p e-
en ed he accumula ion o he DML and C- e bu did no a ec he
s abili y o he DMS cons uc (Fig. 3g). These esul s showed ha he
C- e minus (p esen in bo h he C- e and he DML domain) was
esponsible o his egula ion. Mo eo e , GST-pull down assay using
di e en PARP-1 agmen s (Fig. 3h) demons a ed ha he C- e o
HIF-1
α
in e ac ed wi h he PARP-1 domain D, which is known o be
esponsible o p o ein-p o ein in e ac ions (Fig. 3i).
3.4. HIF-1
α
is modi ied by poly(ADP- ibose) a i s C- e minus
Ha ing demons a ed ha HIF-1
α
and PARP-1 do in e ac , we asked
i HIF-1
α
is a subs a e o PARP-1-media ed PARyla ion du ing hypoxia.
Using a co-IP app oach wi h an i HIF-1
α
as ba e, we ound ha HIF-1
α
is
PARyla ed in no moxia and ha his modi ica ion inc eased du ing
hypoxia (Fig. 4a). Co-IP wi h an i-PARP-1 and HIF-1
α
con i med he
p esence o bo h p o eins o ming a complex ha inc eased du ing
Fig. 2. Hypoxia induces ROS and PAR accumula ion, bo h e en s necessa y o HIF-1
α
s abili y and ac i i y. a ROS measu emen du ing inc easing imes o hypoxia
( om 10 o 240 min) on HEK 293T cells wi h and wi hou p e ea ing wi h he ROS sca enge MPG2 a 300
μ
M b,c Wes e n blo showing PARyla ion le els and HIF-
1
α
accumula ion on HEK 293T du ing he same inc easing imes o hypoxia, HIF-1
α
accumula ion is impai ed when he cells we e p e ea ed wi h PJ34 a 10
μ
M o
MPG2 300
μ
M espec i ely. d-i Gene exp ession o di e en hypoxic- ela ed genes. Du ing no moxia and hypoxia (2, 4 and 24 h). Cells we e p e ea ed wi h PJ34 10
μ
M and MPG-2 a 300
μ
M. In bo h cases HIF-1
α
ansc ip ional ac i i y is impai ed.
J.M. Ma í e al.
Redox Biology 41 (2021) 101885
7
hypoxia and ha was des abilized in he p esence o PARP inhibi o s
(Fig. 4b). To u he deepen in he cha ac e is ics o his in e ac ion, co-
IP assays we e pe o med agains poly(ADP- ibose) in cells o e -
exp essing he HIF-1
α
domains C- e and DMS. PARyla ion was obse ed
in he endogenous HIF-1
α
p o ein and he C- e minus domain, and i was
educed a e PARP inhibi ion. No modi ica ion was obse ed on he
DMS domain (Fig. 4c).
To iden i y he speci ic amino acids unde going his modi ica ion, we
in e oga ed he C- e sequence o pu a i e PARylable si es. I has been
desc ibed ha amino acid sequences p esen ing epea ed lysine and
a ginine mo i s can be accep o s o poly(ADP- ibose) [45]. We iden i ied
wo sequences wi h simila cha ac e is ics on he C- e domain. Si e
di ec ed mu agenesis was pe o med o change hese lysins and a gi-
nines esidues in o non-PARylable alanines (Figu e S2a). Cells we e hen
ans ec ed wi h hese HIF-1
α
mu an s. Unexpec edly, he mu a ions
di ec ed HIF-1
α
o he cy osol (Figu e S2b). Fu he esea ch e ealed
ha hose po en ial PARylable si es we e loca ed in a seconda y bipa -
i e NLS desc ibed o HIF-1
α
[46]. We decided hen o pe o m an in
i o app oach whe e we obse ed how pu i ied HIF-1
α
was PARyla ed
by PARP-1 (Fig. 4d). To s udy he C- e domain, ou di e en
bio in-modi ied pep ides con aining he p e iously desc ibed sequences,
we e syn hesized on hei wild ype and mu an o m (Fig. 4e). Again, an
in i o PARyla ion assay was pe o med, showing ha bo h wild ype
pep ides we e modi ied wi h PAR; a e he alanine mu a ion, his PTM
was educed o he same le el han he nega i e con ol (Fig. 4 ).
3.5. PARP-1 condi ioned HIF-1
α
ec ui men o a ge p omo e s du ing
hypoxia
Once s udied he na u e o he in e ac ions be ween PARP-1 and HIF-
1
α
on he hypoxic con ex , we e alua ed he impac ha PARP-1 had on
he ec ui men o HIF-1
α
o he ch oma in. Fi s ly, we gene a ed a cell
line o HEK 293T PARP-1 knockou cells using CRISPR/Cas9. We
con i med ha PARP-1 was co ec ly emo ed and es ed he capaci y o
he cells o p oduce poly(ADP- ibose) du ing pe oxide-induced s ess
(Figu e S2c). We also de e mined ha , simila ly o PARPi, he absence o
PARP-1 diminished he amoun o HIF-1
α
accumula ed du ing hypoxia
(Figu e S2d).
Then, a ChIP-Seq analysis was pe o med du ing ea ly hypoxia (4
hou s) on he HEK 293T WT and PARP-1 KO cells. The ChIP-Seq
e ealed ha mos o he HIF-1
α
binding peaks de ec ed we e loca ed
wi hin he i s kilobase ups eam o ansla ion s a si es: 48,16% in
wild ype and 43,9% in PARP-1 KO (Figs. 5 and 6)) which includes he
5′-UTR and he p omo e egion o he genes. S a is ical analyses we e
comple ed o elimina e non-signi ican peaks and o compa e hem wi h
he inpu and he no moxic con ol signals. 123 peaks we e de ec ed on
he hypoxic wild ype cells and 68 in he PARP-1 KO cells. Compa ing
wi h a p e ious s udy by Sch¨
odel e al. [47] 400 high s ingency
HIF-1-binding si es we e ound. The di e ences wi h ou s udy could be
asc ibed o he di e en cell ype used in hei s udy (MCF7), and he
ac ha hey we e analysing he hypoxic esponse a e 16 h a 0,5% O
2
.
In ou case we ocused on he ea ly esponse (4 h) using a less limi ing
oxygen concen a ion (1% O
2
).
Fig. 3. HIF-1
α
in e ac s wi h PARP-1 a i s C- e minus domain egula ing i s s abili y. a HIF-1
α
cons uc s p esen ing he wild ype sequences o he HIF-1
α
w , sho
WT and C- e . The PHD esidues mu a ed a e p esen ed in blue, ob aining he DML and DMS. b- On HEK 293T wes e n blo s showing he down egula ion du ing
PHD o e exp ession o he HIF-1
α
and sho w . This educ ion is impai ed on he PHD insensi i e mu an s. g Wes e n blo s udy on HEK 293T o he h ee PHD-
insensi i e domains o HIF-1
α
, hei s abili y is compa ed du ing no moxia and hypoxia 4 h wi h PARP ac i a ed o inhibi ed using PJ34 10
μ
M h PARP-1 domains
p esen ing he ull p o ein, and sepa a ely, he DNA-binding, he au omodi ica ion and he ca aly ic domain. i pull-Down assay exposing he di e en domains o
PARP-1 o he C- e domain o HIF-1
α
. The C- e binds o PARP-1 ull p o ein and he au o modi ica ion domain.
J.M. Ma í e al.
Redox Biology 41 (2021) 101885
8
Du ing hypoxia, PARP-1 KO cells p esen ed 44,72% less HIF-1
α
binding si es and no a single new peak gained wi h espec o he wild
ype cells. A e g ouping hese genes acco ding o hei unc ions
(Fig. 5a), we ound ha he e ec o PARP-1 on HIF-1
α
associa ion o he
ch oma in was no andomly dis ibu ed, bu he e was a ange o hi-
e a chical educ ion depending on he gene on ology. While some
cellula ac i i ies emained almos unchanged ( his was he case o
glucose me abolism, angiogenesis, o DNA epai ), o he unc ions
almos disappea ed (mi ochond ial ac i i y, memb ane o ganiza ion
and cell cycle egula ion) (Fig. 5b). We hen compa ed he common
peaks ound among he op 30 binding si es epo ed in pe ious ChIP-
seq and ou esul s [47] and we ound ha 7 ou o 10 common genes
(ALDOA, ANKRD37, EGLN3, ENO1, GPI, PDK1, PFKFB3, PFKFB4, PKM,
RSBN1) a e in ol ed in glycolysis, con i ming ha he egula ion o
glucose me abolism i is likely o be a key pa ame e o be egula ed
du ing cell’s adap a ion o hypoxia.
In igued by his di e en ial loss we comple ed a MEME analysis
[48] on he sequences o which HIF-1
α
bound du ing hypoxia in he wild
ype and PARP-1 KO con ex (Fig. 5c). In e es ingly, HIF-1
α
p esen ed
high lexibili y binding o di e en sequences du ing he no mal hypoxic
induc ion, being HIF-1
α
peaks loca ed on long and a iable sequences.
In con as , in PARP-1 KO cells HIF-1
α
capaci y o bind o he DNA
became mo e igid and limi ed, loca ing in sho e and mo e
HRE-simila sequences.
In o de o unde s and he educed binding o HIF-1
α
o i s a ge
genes in he absence o PARP-1, we pe o med a mo e ex ensi e analysis
o he ChIP-Seq o obse e in de ail his new binding pa e n. When
s udying he ec ui men o HIF-1
α
conside ing he gene a qui ec u e we
obse ed how HIF-1
α
loca ed p ima ily in he p omo e egions o i s
a ge genes (Fig. 6a). Being his a ea he one ha p esen ed he s onge
educ ion in HIF-1
α
associa ion in PARP-1 KO cells.
We analyzed he da a conside ing he dis ance be ween HIF-1
α
binding si es and he closes TSS (measu ed in kilobases, o Kb) (Fig. 6b).
We obse ed how he mos equen accumula ion o HIF-1
α
ook place
wi hin he i s 1 Kb a ound he TSS. To s udy in de ail hese p omo e
egions, we ocused on he equency o HIF-1
α
de ec ion only 2 Kb
a ound he TSS (Fig. 6c). HIF-1
α
p esen ed a p edominan concen a ion
on he i s 1 Kb ci cling he TSS. In PARP-1 KO cells howe e , HIF-1
α
was educed mainly on ha egion, whe e as a ansc ip ion ac o i s
expec ed o pe o m i s main ac i i y.
Beyond HIF-1
α
ec ui men , i was in iguing he e ec ha PARP-1
abla ion had o e hypoxic gene exp ession when measu ed ia mRNA
Fig. 4. PARP-1 physically in e ac s and PARyla es HIF-1
α
a i s C-Te domain. a Immunop ecipi a ion o HIF-1
α
on HEK 293T cells unde going no moxia and
hypoxia 4 h. A complex is o m leading o PARyla ion. b Immunop ecipi a ion o HIF-1
α
and PARP-1 on HEK 293T cells exposed o no moxia o hypoxia 4 h. Du ing
PARP-1 inhibi ion using PJ34 10
μ
M he complex is des abilized. c Immunop ecipi a ion o PAR polyme on HEK 293T cells ans ec ed wi h he C- e o DMS domain
o HIF-1
α
du ing no moxia and ea ly hypoxia 4 h. Du ing PARP inhibi ion wi h PJ34 10
μ
M he complex obse ed be ween PARP-1 and he endogenous HIF-1
α
o he
C- e is los . d In i o PARyla ion assay o PARP-1 and HIF-1
α
. Coomasie s aining and
32
P-NAD-PAR au o adiog aphy is p esen ed. e Syn hesis o ou pep ides
loca ed on he HIF-1
α
C- e . Two WT and hei co esponden non PARylable analogs a e p esen ed. in i o PARyla ion assay pe o med in he ou pep ides. The
esul s a e p esen ed in coun s pe minu e.
J.M. Ma í e al.
Redox Biology 41 (2021) 101885
9
le els. No only hose genes ha p esen ed less HIF-1
α
ec ui men on
hei p omo e s showed educed mRNA le els, bu e en hose genes ha
p esen ed simila o e en s onge HIF-1
α
accumula ion on hei p o-
mo e s, displayed a educed mRNA exp ession (Figu e S3) showing ha
HIF-1
α
ec ui men pe se was no enough o gua an ee he exp ession o
i s a ge genes du ing PARP-1 abla ion. The possible mechanisms un-
de laying his egula ion will be add essed a he discussion sec ion.
3.6. PARP-1 inhibi ion/abla ion impac s cellula i ness du ing hypoxia
Once s udied in dep h he ole o PARP-1 o e di e en aspec s o
HIF-1
α
biology, and how i al e ed he exp ession o genes in ol ed in
hypoxic adap a ion. We e alua ed how PARP-1 inhibi ion o abla ion
could be a ec ing cellula i ness du ing hypoxia.
We e alua ed he s a us o s ess ou es like he DNA damage
esponse (DDR) and he eplica ion s ess, as well as he cells capaci y o
consume O
2
and he glycolysis s a us. To inish we obse ed he impac
o hese me abolic al e a ions on cellula p oli e a ion and mig a ion
capaci y.
The induc ion o he DDR in hypoxia has been p e iously epo ed in
condi ions close o anoxia (<0.1% O
2
) and i has been shown o include
eplica ion s ess ia p-RPA accumula ion [49–51]. We measu ed DNA
damage using he accumula ion o 53BP1 oci in immuno luo escence
and analyzed o he DDR and eplica ion s ess ma ke s ia wes e n blo .
In e es ingly, hypoxia ea men alone did no induce he ac i a ion o
he DDR. Howe e , an inc ease on 53BP1 oci was obse ed a e ola-
pa ib ea men (Figu e S4a and b) which was epo ed p e iously in
di e en se ings [52,53]. On he same di ec ion, wes e n blo analysis
o di e en p o eins in ol ed in he DDR showed no clea di e ence in
DDR o eplica i e s ess induc ion (Fig S4c). Indica ing ha he ROS
induced du ing hypoxia we e ac i a ing PARP-1 h ough a mechanism
independen o DNA damage.
Then we e alua ed he s a us o he cells ega ding oxygen con-
sump ion and glycolysis, key p ocesses du ing hypoxic adap a ion. The
s udy o oxygen consump ion (Fig. 7a) e ealed ha du ing no moxia
bo h HEK 293T WT and PARP-1 KO cells had ele a ed and simila ox-
ygen consump ion le els. Howe e , du ing hypoxia he HEK 293T
PARP-1 KO cells p esen ed only a pa ial educ ion in oxygen con-
sump ion compa ed wi h he obse ed on he HEK 293T WT cells. Gi en
he c ucial signi icance o he inhibi ion o he oxida i e phospho yla-
ion du ing hypoxia, his di e ence in he a e o oxygen consump ion
could esul in inc eased oxida i e s ess in he absence o PARP-1.
The s udy o he glycolysis pa hway on bo h cell lines showed low
le els o glucose consump ion in no moxia. Howe e , du ing hypoxia a
educed glycolysis induc ion was obse ed in he PARP-1 KO cells
compa ed o he WT cells (Fig. 7b). This esul could indica e he lack o
a p ope swi ch om he oxida i e o he glycoly ic pa hway. Howe e ,
his obse a ion is ha d o conclude because glycolysis was al eady
lowe on PARP-1 KO cells du ing no moxia.
We de e mined he impac o he al e a ions obse ed in hese
Fig. 5. Chip-seq analysis o HIF-1
α
capaci y o bind he p omo e s o i s a ge genes on WT s. PARP-1 KO cells. a ChIP-Seq pe o med o HIF-1
α
a e 4 h o hypoxia
on HEK 293T and PARP-1 K.O cells. The HIF-1
α
binding si es a e p esen ed conside ing he unc ion o he closes gene. b Rep esen a ion o he pe cen age o loss on
he di e en unc ions on he PARP-1 KO cells. c MEME analysis showing he mos common sequences whe e HIF-1
α
binds and hei E- alue.
J.M. Ma í e al.