Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
1 o 17
MOLECULAR BIOLOGY
HMCES co up s eplica ion o k s abili y du ing base
excision epai in homologous
ecombina ion–de icien cells
Ma ía José Peña- Gómez1,2, Yaiza Rod íguez- Ma ín1,2, Ma a del Rio Oli a1,2,
Yodha a Wijeseka a Han hi3,4, Sa a Be ada5,6, Raimundo F ei e7,8,9, Jean Y es Masson5,6,
José Ca los Reyes1, Vincenzo Cos anzo3,4, I án V. Rosado1,2*
Apu inic/apy imidinic (AP) si es and single- s and b eaks a ising om base excision epai (BER) du ing he misin-
co po a ion o damaged nucleobases may hinde eplica ion o k s abili y in homologous ecombina ion–de icien
(HRD) cells. A empla ed AP si es, c oss- links be ween he DNA and 5- hyd oxyme hylcy osine binding, emb y-
onic s em cell–speci ic (HMCES) egula e eplica ion o k speed, a oiding cy o oxic double- s and b eaks. While
he ole o HMCES a he empla e DNA s and is well s udied, i s e ec s on nascen DNA a e no . We p o ide
e idence ha HMCES–DNA- p o ein c oss- links (DPCs) a e de imen al o he BER- media ed emo al o 5- hyd o
xyme hyl- 2′- deoxycy idine (5hmdC)–de i ed 5- hyd oxyme hyl- 2′- deoxyu idine om eplica ion o ks. HRD cells
ha e heigh ened HMCES- DPCs, which inc ease u he upon 5hmdC exposu e, sugges ing ha HMCES binds bo h
spon aneous and 5hmdC- induced AP si es. HMCES deple ion subs an ially supp esses 5hmdC- media ed eplica ion
o k de ec s, ch omosomal abe a ions, and cell dea h in HRD cells. This e eals ha HMCES- DPCs a e a sou ce o
BER- ini ia ed single- s anded DNA gaps, which indica es ha endogenous DPCs con ibu e o genomic ins abili y
in HRD umo s.
INTRODUCTION
Chemical o physical modi ica ions o DNA bases, in addi ion o
u acil (and i s de i a i e o ms) misinco po a ion, a e pe haps he
majo sou ce o DNA lesions in cells. These ha m ul DNA lesions
appea a an es ima ed equency o 1×105 o 5×105 cell−1 day−1
(1). Du ing he epai o hose lesions, apu inic/apy imidinic (AP)
si es a e o med by DNA glycosylase ac i i y. To a oid hei dele e i-
ous consequences due o hei in insic chemical ins abili y, AP si es
a e apidly con e ed o single- s and b eaks (SSBs) by ei he he AP
lyase o he AP endonuclease (APEX) ac i i ies o ce ain class o
nucleases. The e o e, he epai o base al e a ions, AP si es, and
SSBs is pe haps he mos impo an p ocess esponding o oxida i e
and alkyla ing base damages (2,3). Mo e in iguingly is how hese
damaged bases a e epai ed by base excision epai (BER) in he con-
ex o eplica ion o ks. This is o u mos ele ance as AP si es and
SSBs a e po en eplica ion o k–blocking lesions, which, i le un e-
pai ed, hinde eplica ion o k s abili y. Mo eo e , he pe sis ence o
single- s anded DNA (ssDNA) gaps behind eplica ion o ks is ele-
an in he clinic o e adica e homologous ecombina ion–de icien
(HRD) umo s, as hey a e e icien ly a ge ed by poly(adenosine
5′- diphospha e– ibose) polyme ase (PARP) inhibi o s. Fo hese ea-
sons, unde s anding ssDNA gap o ma ion caused by BER ac i i y is
o pa amoun in e es in he con ex o HRD, as many ac o s in ol ed
in HR [b eas cance gene 1 (BRCA1), b eas cance gene 2 (BRCA2),
adia ion sensi i e p o ein 51 (RAD51), anconi anemia g oup D2
(FANCD2), e c.] a e also c ucial o eplica ion o k main enance.
Misinco po a ion o damaged nucleo ides and depu ina ion o ad-
duc ed bases a e among he mos common spon aneous base lesions in
DNA (1,4–6), igge ing canonical BER (7,8). Du ing damaged base
emo al, AP si es gene a ed by DNA glycosylases [e.g., single-s and
selec i e mo no unc ional u acil DNA glycosylase (SMUG1) du ing
5- hyd oxyme hyl- 2′- deoxyu idine (5hmdU) excision] a e con e ed o
SSBs by APEX1. Upon DNA end p ocessing, DNA polyme ase β o λ
conduc s gap illing, lea ing nicked DNA eady o liga ion by he x- ay
epai c oss- complemen ing 1 (XRCC1)/DNA ligase 1 o ligase 3 com-
plexes (9–11). Despi e he “passing- he- ba on” model o BER, whe e
DNA in e media es a e passed along om one s ep o he nex , AP si es
o SSBs may pe sis and challenge eplica ion o k p og ession, elici ing
a eplica ion s ess (RS) esponse. Upon eplica ion o k s alling by
DNA in e media es, a axia elangiec asia and Rad3- ela ed (ATR) ac-
cele a es ec ui men and s abiliza ion o c i ical componen s o he
DNA damage esponse, such as he Fanconi anemia (FA) pa hway and
he HR ac o s FANCD2/FANCI-associa ed nuclease 1 (FAN1), s uc-
u e-speci ic endonuclease (SLX4), xe ode ma pigmen osum F (XPF),
BRCA1/2, o RAD51. To ul ill DNA syn hesis, he p imase- polyme ase
(PRIMPOL) ep imes ahead o bo h he s alled eplica ion o k and he
lesion o limi eplica ion o k ins abili y, hus dampening RS (12–14).
Howe e , PRIMPOL ac i i y gene a es po en ially cy o oxic ssDNA
gaps behind o ks (15–18), which a e subsequen ly illed in ei he by
REV1- POLζ anslesion syn hesis (TLS) polyme ases o by POLQ (19–
21). Se e al mechanisms ha e been ecen ly p oposed as majo con-
ibu o s o HRD cell le hali y. The uni ying ea u e among hem is he
1cen o Andaluz de Biología Molecula y Medicina Regene a i a (cABiMeR) Uni e -
sidad de Se illa- cSic- Uni e sidad Pablo de Ola ide, Se ille 41092, Spain. 2depa a-
men o de Gené ica, Facul ad de Biologia, Uni e sidad de Se illa, Se ille 41012,
Spain. 3iFOM, he AiRc ins i u e o Molecula Oncology, Milan, i aly. 4depa men
o Oncology and hema ology- Oncology, Uni e si y o Milan, Milan, i aly. 5Genome
S abili y labo a o y, chU de Québec Resea ch cen e , hdQ Pa ilion, Oncology di-
ision, 9 McMahon, Québec ci y, Qc G1R 3S3, canada. 6depa men o Molecula
Biology, Medical Biochemis y and Pa hology, la al Uni e si y cance Resea ch
cen e , Québec ci y, Qc G1 0A6, canada. 7Unidad de in es igación, hospi al Uni-
e si a io de cana ias, ins i u o de in es igación Sani a ia de cana ias (iiSc), la
laguna, San a c uz de ene i e, Spain. 8ins i u o de ecnologías Biomédicas, cen o
de in es igaciones Biomédicas de cana ias, Facul ad de Medicina, campus cien-
cias de la Salud, Uni e sidad de la laguna, San a c uz de ene i e, Spain. 9Uni e -
sidad Fe nando Pessoa cana ias, las Palmas de G an cana ia, Spain.
*co esponding au ho . email: i osado@ us. es
copy igh © 2025 he
Au ho s, some igh s
ese ed; exclusi e
licensee Ame ican
Associa ion o he
Ad ancemen o
Science. no claim o
o iginal U.S.
Go e nmen Wo ks.
dis ibu ed unde a
c ea i e commons
A ibu ion
noncomme cial
license 4.0 (cc BY- nc).
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
2 o 17
p esence o pe sis en ssDNA gaps a nascen DNA s ands, which co -
ela es wi h he exquisi e cy o oxici y o PARP inhibi o s (e.g., olapa ib)
obse ed in HRD cells (22,23). Howe e , he unde lying molecula
basis o his cy o oxici y s ill emains elusi e.
Recen ly, misinco po a ion o dU, 5′- chlo o- dU (5CldU), o
5hmdU has eme ged as a p e iously uniden i ied sou ce o eplica-
ion o k impai men (24–27). Upon misinco po a ion, pe sis ing
DNA epai in e media e s uc u es can be collided by he eplisome,
esul ing in BER- eplica ion con lic s (BRCs). These un esol ed
BRCs accoun o HRD cell dea h in he absence o p o icien eplica-
ion o k main enance o HR pa hways (24–27). In addi ion, modi-
ied DNA bases emaining unde ec able h oughou he cell cycle
induce an ATR- dependen RS and o k collapse esponse du ing he
nex cell cycle when p esen in empla e DNA (28,29). Whe eas
ssDNA gap accumula ion by BRCs du ing he nex cell cycle is well
documen ed (12,30–33), how ssDNA gaps a ise a nascen DNA
du ing he cu en cell cycle emains la gely unexplo ed (25,26).
Damaged DNA bases a e one o he main sou ces o AP si es. On
empla e DNA, AP si es a e exposed du ing DNA unwinding by Cdc45–
MCM–GINS (CMG) helicase and a e p o ec ed by he 5-hyd oxy-
me hylcy osine binding, emb yonic s em cell–speci ic (HMCES) p o ein
(34–39). Th ough i s in e ac ion wi h p oli e a ing cell nuclea an igen
(PCNA), HMCES associa es o eplica ion o ks and binds ssDNA o
shield AP si es (34,35). Th ough a hiazolidine bond be ween he Cys2
esidue and he aldehydic con o ma ion o he AP si e (34–36,40), HM-
CES a aches co alen ly o AP si es, o ming an HMCES–DNA-p o ein
c oss-links (DPC). HMCES-DPC o ma ion on empla e ssDNA de-
c eases eplica ion o k speed while a oiding APEX1- media ed cy o oxic
double- s and b eaks (DSBs) a eplica ion o ks (35). HMCES- DPCs a e
bypassed by a TLS s ep and subsequen ly emo ed om DNA by dis inc
mechanisms. Upon winding o DNA s ands, HMCES- DPC ca alyzes a
c oss- link sel - e e sal eac ion h ough i s Glu127 esidue, eleasing
HMCES om DNA (37,38). Al e na i ely, assis ed by he p o ein dena-
u a ion ac i i y o anconi anemia g oup J (FANCJ) (39), Sp T-like N-
e minal domain (SPTRN) o p o easome deg ades HMCES o allow
eplica ion o k esump ion (35,39). While he consequences o HMCES-
DPCs on empla e DNA a e ecen ly being add essed, he e ec s on
nascen DNA a e unknown. He e, we p o ide e idence o a de imen al
ole o HMCES- DPCs du ing nascen DNA syn hesis. HRD cells
display heigh ened ch oma in le els o HMCES, which inc ease upon
5-hyd oxyme hyl-2′- deoxycy idine (5hmdC) exposu e, sugges ing ha
HMCES binds o spon aneous and 5hmdU- induced AP si es. HMCES
loss escues eplica ion o k impai men , genomic ins abili y, and cy o-
oxic pheno ypes obse ed in HRD cells du ing 5hmdU misinco po a-
ion, indica ing ha HMCES is esponsible o he 5hmdU- media ed
gene ic ins abili y and le hali y o HRD cells. Gene ic deple ion o ssDNA
gap–gene a ing BER ac o s (e.g., SMUG1 o APEX1/2) la gely supp ess-
es hese pheno ypes, whe eas loss o ssDNA gap– illing ac o s PARP1
o XRCC1 exace ba es hem. Ou indings demons a e ha HMCES-
DPCs on nascen DNA a e esponsible o he eplica ion o k de ec s
obse ed in HRD- de icien cells and place endogenous DPCs on nascen
DNA as a no el sou ce o genomic ins abili y in HRD umo s.
RESULTS
Misinco po a ion o 5hmdU a ising om deamina ed
5hmdC causes genomic ins abili y in HRD cells
5hmdU can a ise om ei he he 5hmdC sal age pa hway o he cy o-
sine deme hyla ion p ocess (24,25,27). P e ious da a om ou g oup
and o he s ha e shown ha HRD cells lacking FANCD2, me hyl
me hanesul ona e and UV sensi i e p o ein 81 (MUS81), BRCA1, o
BRCA2 we e sensi i e o misinco po a ion o 5hmdU on genomic
DNA (25,27). To in es iga e he molecula mechanisms accoun ing
o 5hmdU- media ed geno oxici y, we examined whe he deple ion
o ac o s in ol ed in he py imidine sal age pa hway [e.g., deoxycy i-
dine kinase (DCK) o 2′- deoxycy idine 5′- monophospha e deami-
nase (DCTD)] had any e ec on he iabili y o 5hmdC- ea ed
Fancd2−/− mouse emb yonic ib oblas s (MEFs). Consis en wi h
p e ious esul s in human HRD cells (25,27), 5hmdC- media ed
Fancd2−/−
le hali y was la gely supp essed by knockdown o DCK o
DCTD, whe eas wild- ype cells emained unal e ed (Fig. 1A and ig.
S1), con i ming ha deamina ion o 5hmdC o 5hmdU media es
HRD cell cy o oxici y. As PARP apping by olapa ib exace ba ed
5hmdC- dependen Fancd2−/−
ch omosomal ins abili y and cell le-
hali y (25), we sough o de e mine le els o nuclea PARyla ion and
SSBs upon 5hmdC ea men . As expec ed, Fancd2−/−
e hynyl- 2′-
deoxyu idine–posi i e (EdU+) cells displayed a signi ican inc ease in
nuclea PARyla ion (Fig. 1B), also con i med in EdU un es ic ed
cells ( ig. S2A), while i emained ela i ely unchanged in wild- ype
cells (Fig. 1B and ig. S2A). Nex , we examined PAR le els associa ed
wi h ongoing eplica ion o ks by “in si u p o ein in e ac ion wi h na-
scen DNA eplica ion o ks” also known as SIRF assay. Compa ed o
wild- ype MEFs, Fancd2−/− cells displayed heigh ened EdU- PAR oci,
which we e u he exace ba ed by 5hmdC ea men (Fig. 1C).
The inc ease in nuclea PARyla ion o EdU- PAR oci obse ed in
Fancd2−/− cells was la gely dependen on PARP1, as PARP1 knock-
down signi ican ly educed bo h PARyla ion and EdU- PAR oci (Fig.
1, B and C). Fancd2−/−
MEFs also displayed signi ican spon aneous
SSBs measu ed by alkaline come assay, which we e u he inc eased
by 5hmdC ea men , unlike in wild- ype cells (Fig. 1D). These da a
sugges ha 5hmdU misinco po a ion esul s in ssDNA gap accumu-
la ion in he absence o FANCD2, p obably accoun ing o heigh ened
PARyla ion a eplica ion o ks. Consis en wi h his, we also ound
ha 5hmdC exposu e caused inc eased eplica ion o k asymme y
by DNA ibe assay (Fig. 1E), indica ing ha 5hmdU- dependen ep-
lica ion o k impai men was no a consequence o a global check-
poin esponse. Mo eo e , 5hmdU- media ed DNA damage inc eased
sis e ch oma id exchanges (SCEs) in Fancd2−/−
cells (Fig. 1F), sug-
ges ing ha BRCs elici ed an HR- dependen epai mechanism.
Howe e , we did no obse e any e idence o checkpoin ac i a ion
measu ed as Se 345- CHEK1 o Se 33- RPA2 unde hese condi ions
( ig. S3). These da a sugges ha sho 5hmdC exposu e does no ac i-
a e he DNA damage checkpoin despi e inc eased SSBs and indica e
ha ssDNA gaps a ising du ing 5hmdU emo al igge eplica ion
o k impai men and sis e ch oma id ecombina ion in HRD cells.
ssDNA gaps a ising om he conce ed ac i i ies o SMUG1
and APEX1/2 on misinco po a ed 5hmdU al e eplica ion
o k p og ession
P e ious s udies ha e epo ed ssDNA gap accumula ion as a majo
con ibu o o BRCA1/2 umo cell dea h (18,31,41). As SMUG1 is
he main BER DNA glycosylase esponsible o 5hmdU elimina ion
om genomic DNA (42), we sough o examine i s con ibu ion o
ssDNA gap o ma ion in nascen DNA. In ag eemen wi h a p e ious
epo (27), SMUG1 deple ion in 5hmdC- ea ed Fancd2−/− cells
signi ican ly dec eased nuclea PARyla ion o un ea ed le els in bo h
o al and EdU+ cells compa ed o wild- ype cells (Fig. 2A and ig.
S2B). SMUG1 deple ion supp essed 5hmdC- dependen EdU- PAR
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
3 o 17
Fig. 1. Genomic ins abili y induced by 5hmdC in HRD cells. (A) (3-(4, 5-dime hyl hiazolyl-2)-2, 5-diphenyl e azolium b omide) assay (M ) cell p oli e a ion assay o
5hmdc- ea ed wild- ype o Fancd2−/− (Fd2−/−) cells knocked down o dc d ( op) o dcK (bo om) exposed o he indica ed dose o 3 days. (B) le : Rep esen a i e
PAR ( ed) immuno luo escence images o edU+ (g een) wild- ype, siPARP1, Fancd2−/−, and Fancd2−/− siPARP1 cells exposed o 5hmdc (10 μM) o 3 hou s. 4′,6-diamidino-
2- phenylindole (dAPi; blue) s ains nuclea dnA. Righ : Plo depic ing PAR mean in ensi y signal pe nucleus. A.U., a bi a y uni s. (C) le : Rep esen a i e images o edU+-
PAR oci ( ed) by SiRF analysis om wild- ype, siPARP1, Fancd2−/−, and Fancd2−/− siPARP1 cells exposed o 5hmdc (10 μM) o 3hou s. dAPi (blue) s ains nuclea dnA and
edU (g een) s ains S- phase cells. Righ : Plo depic ing edU- PAR oci pe nucleus (n=2). (D) le : Rep esen a i e images o alkaline come assay om wild- ype and
Fancd2−/− cells ollowing 5hmdc ea men (10 μM) o 3hou s. Righ : Plo depic ing come ail momen pe cell (n=4). (E) op le : Scheme o he dnA ibe o igin sym-
me y. Bo om le : Rep esen a i e images o dnA ibe s om wild- ype and Fancd2−/− cells upon 5hmdc ea men (40 μM) o 30 min. Righ : Box plo o he asymme y
index. (F) op le : Scheme o Sce assay. Bo om le : Rep esen a i e images o Sces om wild- ype and Fancd2−/− cells exposed o 5hmdc (10 μM) o 40 hou s. Righ : Ba
plo o Sce/ch omosome pe me aphase sp ead (n=50 o each o he wo biological eplica es; ba ep esen s means±Sd). B dU, 5- b omo- 2′- deoxyu idine.
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
4 o 17
Fig. 2. 5hmdC- media ed HRD genomic ins abili y is dependen on SMUG1. (A) le : Rep esen a i e PAR ( ed) immuno luo escence images o edU+ (g een) wild- ype,
siSMUG1, Fancd2−/−, and Fancd2−/− siSMUG1 cells exposed o 5hmdc (10 μM) o 3hou s. dAPi (blue) s ains nuclea dnA. Righ : Plo depic ing PAR mean in ensi y signal
pe nucleus. (B) le : Rep esen a i e images o edU- PAR oci ( ed) by SiRF assay om wild- ype, siSMUG1, Fancd2−/−, and Fancd2−/− siSMUG1 cells exposed o 5hmdc
(10 μM) o 3hou s. dAPi (blue) s ains nuclea dnA and edU (g een) s ains S- phase cells. Righ : Plo depic ing edU- PAR oci pe nucleus (n=2). (C) le : Rep esen a i e
images o alkaline come assay om wild- ype, siSMUG1, Fancd2−/−, and Fancd2−/− siSMUG1 cells ollowing 5hmdc ea men (10 μM) o 3hou s. Righ : Plo depic ing
come ail momen pe cell (n=2). (D) op le : Scheme o dnA ibe assay. Bo om le : Rep esen a i e images o dnA ibe s om Fancd2−/− and Fancd2−/− siSMUG1 cells
un ea ed (Un ) o exposed o 5hmdc (40 μM) o 30 min. Righ : Box plo ep esen ing he equency o 5′- iododeoxyu idine (idU)/cldU a io o wild- ype, siSMUG1,
Fancd2−/−, and Fancd2−/− siSMUG1 cells a e 5hmdc ea men (n=200). (E) M cell p oli e a ion assay o wild- ype, siSMUG1, Fancd2−/−, and Fancd2−/− siSMUG1 cells
exposed o he indica ed dose o 5hmdc o 3 days (n=4; means±Sd).
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
5 o 17
oci o un ea ed le els (Fig. 2B), sugges ing ha mos eplica ion
o k–associa ed PARyla ion upon 5hmdC exposu e depends on
SMUG1. Consis en wi h his, SMUG1 knockdown also abolished
5hmdC- induced SSBs in Fancd2−/− cells (Fig. 2C), sugges ing ha
SMUG1 is esponsible o 5hmdC- induced ssDNA gap o ma ion.
Mo eo e , SMUG1 deple ion la gely es o ed eplica ion o k p o-
g ession measu ed by DNA ibe assay and cell iabili y in 5hmdC-
ea ed Fancd2−/− cells (Fig. 2, D and E). These da a demons a e ha
SMUG1 loss supp esses ssDNA gap accumula ion, eplica ion o k
impai men , and le hali y obse ed in 5hmdC- ea ed HRD cells,
mos likely h oughou he abolishmen o AP si e o ma ion.
Mammalian cells con ain wo class II APEXs, APEX1 and APEX2
(43,44). While APEX1 has s ong APEX ac i i y, APEX2 exhibi s
s ong in i o 3′- phosphodies e ase and 3′- 5′ exonuclease ac i i ies
(44). We he e o e examined he con ibu ion o APEX1 o APEX2
o PARyla ion and ssDNA gap o ma ion du ing 5hmdU elimina-
ion. Simila o SMUG1 loss, APEX1 knockdown had ha dly any
e ec on nuclea PARyla ion in un ea ed cells (Fig. 3A). Howe e ,
5hmdC- dependen nuclea PARyla ion in o al o EdU+ Fancd2−/−
cells signi ican ly dec eased o nea –wild- ype le els (Fig. 3A and
ig. S2C). APEX1 loss signi ican ly supp essed spon aneous and
5hmdC- induced EdU- PAR oci in Fancd2−/− cells o wild- ype le -
els (Fig. 3B). Mo eo e , loss o APEX1 signi ican ly educed 5hmdC-
induced SSBs obse ed in Fancd2−/− cells o almos un ea ed le els
(Fig. 3C). These da a sugges ha APEX1 is he main endonuclease
esponsible o mos 5hmdC- induced SSBs in Fancd2−/− and, o a
lesse ex en , in wild- ype cells. APEX1 deple ion also subs an ially
supp essed 5hmdC- media ed Fancd2−/− eplica ion o k impai -
men and cell le hali y (Fig. 3, D and E), indica ing ha APEX1-
dependen SSBs accoun o he eplica ion o k de ec s and le hali y
obse ed in 5hmdC- ea ed Fancd2−/− cells. As APEX1 knockdown
does no ully es o e cell iabili y o o k dynamics, hese esul s
also sugges ha APEX1 may play addi ional oles du ing s abili y
o esump ion o eplica ion o ks. Simila o APEX1 deple ion,
APEX2 knockdown also educed 5hmdC- media ed nuclea PARy-
la ion in o al o S- phase (EdU+) Fancd2−/− cells popula ions ( ig.
S4, A and B), o a lesse ex en , han APEX1 deple ion. APEX1 and
APEX2 codeple ed S- phase Fancd2−/− cells showed 5hmdC- induced
PAR le els simila o hose obse ed in APEX1- deple ed 5hmdC-
ea ed Fancd2−/− cells, indica ing an epis a ic ela ionship be ween
APEX1 and APEX2. Mo eo e , APEX1 o APEX2 knockdown com-
ple ely supp essed 5hmdC- induced EdU- PAR oci and SSBs (Fig. 3F
and ig. S4C). These da a indica e ha APEX1, in combina ion wi h
APEX2, is esponsible o he inc eased PARyla ion and SSBs associ-
a ed wi h nascen s and du ing 5hmdU emo al. Consis en wi h
his no ion, APEX1 and APEX2 showed an epis a ic ela ionship on
PAR le els and le hali y obse ed in 5hmdC- ea ed Fancd2−/− cells
(Fig. 3E). These da a poin ou o a conce ed unc ion o APEX1
and APEX2 a p ocessing 5hmdU- de i ed AP si es, likely gene a -
ing ssDNA gaps.
5hmdU- de i ed ssDNA gap pe sis ence a eplica ion o ks
exace ba es genomic ins abili y and le hali y in HRD cells
In unpe u bed S- phase, ssDNA gap pe sis ence due o ine icien
Okazaki agmen liga ion is signaled by PARP1- dependen PARyla-
ion, ollowed by XRCC1- media ed PARP1 e ic ion, o p omo e
ai h ul gap epai (45). We easoned ha de ec i e 5hmdC-
dependen ssDNA gap p ocessing ei he by excessi e PARP1 e en-
ion o by PARP1 o XRCC1 loss would exace ba e ssDNA gap
pe sis ence and cell le hali y, hus mimicking pe sis en unliga ed
Okazaki agmen s. XRCC1 knockdown inc eased u he nuclea
PARyla ion in 5hmdC- ea ed o al o EdU+ Fancd2−/− cell popula-
ions while emaining una ec ed in wild- ype cells (Fig. 4A and ig.
S2D). Likewise, he SIRF assay e ealed a signi ican inc ease in
EdU- PAR oci in 5hmdC- ea ed Fancd2−/− cells upon XRCC1 de-
ple ion (Fig. 4B). PARP1 o XRCC1 deple ion u he exace ba ed
5hmdC- media ed Fancd2−/− cell le hali y (Fig. 4C) simila o PARP1
apping by olapa ib (25) as a esul o ine icien epai o ssDNA
gaps. Howe e , XRCC1 deple ion did no u he a ec 5hmdC-
media ed Fancd2−/− eplica ion o k impai men (Fig. 4D). These
da a indica e ha ssDNA gaps associa ed wi h eplica ion o ks in
he absence o FANCD2 pe sis o accumula e o a la ge ex en
upon XRCC1 loss.
Replica ion o ks s alled by bulky base adduc s s imula e pos ep-
lica i e SCEs in an ssDNA gap– and PRIMPOL- dependen ashion
(14). We he e o e examined SCEs induced by 5hmdC- media ed
ssDNA gaps in nascen DNA by exposing cells o 5hmdC o less han
one cell cycle leng h (≈12 hou s). We ound ha 5hmdC igge ed a
signi ican inc ease in SCEs in Fancd2−/− cells compa ed o un ea -
ed cells (Fig. 4E). Mo eo e , SMUG1 o APEX1 deple ion supp essed
5hmdC- induced SCEs, while PARP1 o XRCC1 knockdown exace -
ba ed hem (Fig. 4E). These da a indica e ha 5hmdC- media ed
ssDNA gaps in nascen s ands occu ing du ing he cu en cell cycle
a e a sou ce o SCEs in HRD cells. Ou indings sugges ha de ec-
i e epai o pe sis en eplica i e 5hmdC- induced ssDNA gaps
accoun s o inc eased PARyla ion, eplica ion o k impai men ,
heigh ened SCEs, and le hali y obse ed in Fancd2−/− cells.
HMCES- DPCs a e esponsible o 5hmdC- media ed ssDNA
gaps in HRD cells
AP si es and o he in e media e DNA ends du ing BER hampe ep-
lica ion o k p og ession, causing BRCs (28,29,36,46). AP si es a e
p o ec ed by HMCES o a oid cy o oxic DSB o ma ion a he epli-
ca ion o k while p omo ing TLS bypass mechanisms (35,36,40).
We he e o e easoned ha HMCES could eac o 5hmdU- de i ed
AP si es a ising du ing 5hmdC ea men . Upon p eex ac ion, we
obse ed a signi ican ly highe p opo ion o nuclei s ained posi i e
o ch oma in- bound Flag- agged HMCES in Fancd2−/− han in
wild- ype cells unde un ea ed condi ions (Fig. 5A). Fancd2−/− cells
also displayed a b igh e ch oma in- bound HMCES signal com-
pa ed o i s wild- ype coun e pa unde spon aneous, me hyl me h-
anesul ona e (MMS) used as a posi i e con ol, o 5hmdC exposu e
(Fig. 5B). Mo eo e , ch oma in- bound HMCES le els di ec ly
co ela ed wi h nascen s and 5hmdC misinco po a ion in a dose-
dependen manne , eaching signi ican ly highe le els in Fancd2−/−
han in wild- ype cells (Fig. 5B). We also con i med hese indings
by cellula sub ac iona ion and by apid app oach o DNA adduc
eco e y (RADAR) assays, showing a 5hmdC dose- dependen
HMCES ch oma in e en ion ( ig. S5, A and B). These da a sugges ha
Fancd2−/− cells p esen highe le els o endogenous AP si e–bound
HMCES han wild- ype cells, which inc ease upon 5hmdC ea -
men in a dose- dependen manne . We nex examined he ch oma in
e en ion ac i i y o cell lines o e exp essing ei he Cys2→Ala (C2A)
o A g212→Glu (R212E), wo mu an e sions o HMCES showing
de ec i e DPC o ma ion and ssDNA binding ac i i ies, espec-
i ely. Consis en wi h hei epo ed in i o biochemical ac i i-
ies (35), HMCES- C2A showed impai ed ch oma in ec ui men
upon 5hmdC o MMS exposu e, while HMCES- R212E comple ely
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
6 o 17
Fig. 3. APEX1 and APEX2 con ibu e o 5hmdC- induced ssDNA gap gene a ion in HRD cells. (A) le : Rep esen a i e PAR ( ed) immuno luo escence images o siA1-
deple ed (siA1) wild- ype o Fancd2−/− cells exposed o 5hmdc (10 μM) o 3hou s. dAPi (blue) s ains nuclea dnA and edU (g een) s ains S- phase cells. Righ : Plo depic -
ing PAR mean in ensi y signal pe edU+ nucleus. (B) le : Rep esen a i e images o edU- PAR oci ( ed) by SiRF assay om wild- ype, siA1 Fancd2−/−, o Fancd2−/− siA1 cells
exposed o 5hmdc (10 μM) o 3hou s. dAPi (blue) s ains nuclea dnA and edU (g een) s ains S- phase cells. Righ : Plo depic ing edU- PAR oci pe nucleus. (C) le :
Rep esen a i e images o alkaline come assay o APeX1 knockdown wild- ype o Fancd2−/− cells exposed o 5hmdc (10 μM) o 3hou s. Righ : Plo depic ing come ail
momen pe cell. (D) op le : Scheme o dnA ibe assay. Bo om le : Rep esen a i e images o dnA ibe s om Fancd2−/− and Fancd2−/− siA1 cells un ea ed o exposed
o 5hmdc (40 μM) o 30 min. Righ : Box plo o he equency o idU/cldU a io o siA1- deple ed wild- ype o Fancd2−/− cells upon 5hmdc ea men (n=200 o each o
he wo biological eplica es). (E) op: M cell p oli e a ion assay o siA1- deple ed wild- ype o Fancd2−/− cells exposed o he indica ed dose o 5hmdc o 3 days (n=9;
means±Sd). Bo om: M cell p oli e a ion assay o APeX1- o siA2- deple ed wild- ype o Fancd2−/− cells exposed o he indica ed dose o 5hmdc o 3 days. (F) Plo
depic ing edU- PAR oci pe nucleus by SiRF assay om wild- ype, Fancd2−/−, siA1, siA2, Fancd2−/− siA1, and Fancd2−/− siA2 cells exposed o 5hmdc (10 μM) o 3hou s
(n=2).
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
7 o 17
Fig. 4. XRCC1 deple ion exace ba es 5hmdC- induced genomic ins abili y and le hali y o HRD cells. (A) le : Rep esen a i e PAR ( ed) immuno luo escence images
o XRcc1- deple ed wild- ype o Fancd2−/− cells exposed o 5hmdc (10 μM) o 3hou s. dAPi (blue) s ains nuclea dnA and edU (g een) s ains S- phase cells. Righ : Plo
depic ing PAR mean in ensi y signal pe nucleus. (B) le : Rep esen a i e images o edU- PAR oci ( ed) by SiRF assay o XRcc1- deple ed wild- ype o Fancd2−/− cells ex-
posed o 5hmdc (10 μM) o 3hou s. dAPi (blue) s ains nuclea dnA and edU (g een) s ains S- phase cells. Righ : Plo depic ing edU- PAR oci pe nucleus. (C) op: M cell
p oli e a ion assay o wild- ype o Fancd2−/− cells lacking XRcc1 exposed o he indica ed dose o 5hmdc o 3 days (n=9; means±Sd). Bo om: M cell p oli e a ion
assay o wild- ype, siPARP1, Fancd2−/−, and Fancd2−/− siPARP1 cells exposed o he indica ed dose o 5hmdc o 3 days (n=5; means±Sd). (D) op le : Scheme o dnA
ibe assay. Bo om le : Rep esen a i e images o dnA ibe s om Fancd2−/− and Fancd2−/− siXRcc1 cells un ea ed o exposed o 5hmdc (40 μM) o 30 min. Righ : Box
plo o he equency o idU/cldU a io o XRcc1- deple ed wild- ype o Fancd2−/− cells upon 5hmdc ea men (n=200 o each o he wo biological eplica es). (E) op
igh : Scheme o he Sce assay. le : Ba plo o Sce/ch omosome pe me aphase om SMUG1- , APeX1- , XRcc1- , o PARP1- deple ed Fancd2−/− cells exposed o 5hmdc
(10 μM) o 12hou s (n=50 o each o he wo biological eplica es; ba ep esen s means±Sd).
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
8 o 17
Fig. 5. HMCES accumula es in nuclei and is esponsible o 5hmdC- media ed genomic ins abili y obse ed in HRD cells. (A) le : Rep esen a i e immuno luo es-
cence images o Flag- agged hMceS (g een) posi i e nuclei om wild- ype o Fancd2−/− cells. Righ : Pe cen age o hMceS- FlAG+ wild- ype and Fancd2−/− cells. (B) le :
Rep esen a i e immuno luo escence images o hMceS posi i e nuclei om wild- ype o Fancd2−/− cells upon combined 5hmdc (160 μM) and N- ca bobenzyloxy- l- leucyl-
l- leucyl- l- leucinal (MG132; 10 μM) ea men s o 3 hou s. Righ : hMceS- FlAG mean in ensi y signal pe nucleus om wild- ype and Fancd2−/− cells upon 5hmdc (3 hou s)
o MMS (0.5 mM, 30 min). ea men s we e combined wi h MG132 (10 μM). (C) le : images o FlAG- agged hMceS posi i e nuclei om Fancd2−/− cells, exp essing
MmhMceS- W - FlAG, MmhMceS- c2A- FlAG, o MmhMceS- R212e- FlAG upon 5hmdc exposu e (160 μM, 3hou s). Righ : Plo depic ing hMceS- FlAG mean in ensi y
signal pe nucleus om wild- ype o Fancd2−/− cells exp essing MmhMceS- W - FlAG, MmhMceS- c2A- FlAG, o MmhMceS- R212e- FlAG upon 5hmdc (indica ed dose o
3hou s) o MMS (0.5 mM). ea men s we e combined wi h MG132 (10 μM). ns, no signi ican . (D) le : Rep esen a i e images o edU- PAR oci by SiRF assay om hMceS-
deple ed wild- ype o Fancd2−/− cells exposed o 5hmdc (10 μM) o 3hou s. dAPi (blue) s ains nuclea dnA and edU (g een) s ains S- phase cells. Righ : Plo depic ing
edU- PAR oci pe nucleus. (E) le : Rep esen a i e images o alkaline come assay om hMceS- deple ed wild- ype o Fancd2−/− cells ollowing 5hmdc ea men (10 μM)
o 3hou s. Righ : Plo depic ing come ail momen pe cell.
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
9 o 17
abolished i (Fig. 5C and ig. S6). These da a sugges ha 5hmdU-
de i ed AP si es on nascen DNA s and a e ecognized by HMCES,
o ming HMCES- DPCs. We p edic ed ha HMCES loss would in-
c ease cy o oxic ssDNA gaps on nascen DNA s and by APEX1-
media ed incision o AP si es igh a o behind he o k, esul ing in
a syn he ic le hal pheno ype in HRD cells. To ou su p ise, HMCES
knockdown ma kedly dec eased 5hmdC- induced nuclea PARyla-
ion in o al o EdU+ Fancd2−/− cell popula ions ( ig. S7, A and B),
EdU- PAR oci in S- phase cells (Fig. 5D), and SSBs obse ed in
Fancd2−/− cells (Fig. 5E), sugges ing ha HMCES- DPCs on nascen
DNA s and a e esponsible o a la ge subse o 5hmdC- induced
ssDNA gaps obse ed in Fancd2−/− cells. Consis en wi h he p o-
ec i e ole o HMCES du ing eplica ion (35), HMCES deple ion
inc eased nuclea γ- H2AX, which was u he ele a ed in he ab-
sence o FANCD2 ( ig. S7C). Ne e heless, γ- H2AX le els emained
ela i ely unchanged du ing 3hou s o 5hmdC ea men , sugges -
ing ha ssDNA gaps in nascen DNA a ising om AP si es unp o-
ec ion due o HMCES loss do no esul in cumula i e DSBs ( ig.
S7C). HMCES deple ion also es o ed Fancd2−/− eplica ion o k
p og ession and symme y o wild- ype le els (Fig. 6, A and B).
5hmdC- induced heigh ened ch omosomal abe a ions we e ma k-
edly educed upon HMCES deple ion (Fig. 6C), speci ically hose o
adial ch omosomes, which a ose a he expense o ch omosomal
gaps ( ig. S8). As a consequence, 5hmdC- induced HRD cell iabili y
was also no ably imp o ed upon HMCES deple ion (Fig. 6D), indi-
ca ing ha HMCES- DPCs o med a BER in e media es a e espon-
sible o eplica ion o k impai men , ch omosomal ins abili y, and
cell le hali y in HRD cells. To ule ou ha he pheno ypic supp es-
sion by HMCES knockdown was a consequence o pooled small
in e e ing RNA (siRNA) o - a ge s, we alida ed ou esul s using
ou independen siRNAs agains HMCES ( om numbe 1 o
numbe 4) compa ed o pooled siRNAs, ob aining simila esul s o
pooled siRNAs ( ig. S9). We also alida ed ou esul s using human
BRCA2–de icien DLD- 1 cells, which we e p e iously epo ed o
be sensi i e o 5hmdU (25). HMCES knockdown also supp essed
5hmdC- media ed BRCA2−/− cell le hali y (Fig. 6E), hus b oaden-
ing hese indings o di e en human gene ic backg ounds. To di-
ec ly examine he e ec s o HMCES deple ion on DNA eplica ion
in e media es, we use DNA elec on mic oscopy (EM) o moni o
he equency o b oken eplica ion o k in e media es, possibly
esul ing om ssDNA gap b eakage (47). In ag eemen wi h
ou gene ic da a, compa ed o un ea ed condi ion, 5hmdC- ea ed
Fancd2−/− cells displayed signi ican inc ease in DNA eplica ion
in e media es wi h asymme ic b anches, indica ing b oken o ks,
possibly esul ing om he p ocessing o ssDNA gaps p esen in
Fancd2−/− cells (Fig. 6F and ig. S10) (47). HMCES deple ion la gely
supp essed 5hmdC- media ed o k ins abili y and ssDNA gap accu-
mula ion (Fig. 6F). We also a emp ed o alida e hese indings by
S1 nuclease DNA ibe assay bu ailed o ob ain ep oducible esul s
in MEFs. To o e come his, we knocked ou FANCD2, HMCES, o
bo h genes in human RPE- 1 cells by CRISPR- Cas9 app oaches ( ig.
S11, A and B). Consis en wi h published da a, HMCES loss did no
a ec unpe u bed eplica ion o k p og ession (40) o he amoun
o ssDNA gaps, sugges ing ha unp o ec ion o spon aneous AP
si es upon HMCES loss does no a ec o e all le els o ssDNA gaps
( ig. S11C). Despi e S1 insensi i i y in un ea ed FANCD2−/− RPE- 1
cells, small ssDNA gaps we e isible in all Fancd2−/− MEFs samples
by EM (Fig. 6F and ig. S10), compa ibly wi h AP si e accumula ion
on empla e s ands. Mo eo e , exposu e o FANCD2−/− RPE- 1 cells
o 5hmdC (10 μM) signi ican ly a ec ed 5′- iododeoxyu idine (IdU)
ack leng h, sugges ing ha FANCD2 is equi ed o a oid ssDNA
gap o ma ion. No ably, HMCES deple ion supp essed 5hmdC-
induced ssDNA gap o ma ion in FANCD2−/− cells de ec ed by S1
nuclease ( ig. S11C), con i ming ha HMCES- DPCs a e esponsible
o 5hmdU- de i ed ssDNA gap accumula ion in HRD cells. We also
ound ha low exp ession le els o HMCES co ela ed wi h lowe
o e all su i al in wo independen coho s o pa ien s wi h b eas
adenoca cinoma, one om TCGA (1089 pa ien s) and ano he om
a se o cu a ed b eas cance coho s om GEO (2976 pa ien s)
(Fig. 6G and ig. S11D). These da a sugges ha HMCES loss migh
play a ole in b eas cance agg essi eness o pa ien ou come, al-
hough hese indings need u he alida ion.
HMCES- DPCs on empla e DNA s and showed an epis a ic
ela ionship o e APEX1 incision o a oid oxic DSB o ma ion
(34,35). We he e o e examined nuclea PARyla ion and iabili y
pheno ypes upon deple ion o HMCES, APEX1, APEX2, o hei
combina ions. Nuclea PARyla ion in HMCES APEX1 double-
knockdown EdU+ cells esembled he one seen in APEX1 knock-
down Fancd2−/− cells (Fig. 7A). Howe e , PARyla ion in HMCES
APEX2 double- knockdown Fancd2−/− cells esembled he one ob-
se ed in Fancd2−/− HMCES single- knockdown EdU+ cells (Fig.
7B). Mo eo e , knockdown o HMCES, APEX1, o APEX2 also
showed an epis a ic supp essi e pheno ype on Fancd2−/− cell su -
i al (Fig. 7, C and D). These da a sugges ha HMCES, APEX1, and
APEX2 unc ion in he same gene ic pa hway du ing he p ocessing
o 5hmdU om nascen DNA.
PRIMPOL media es a subse o 5hmdC- media ed ssDNA gaps
Replica ion o k s alling by bulky base adduc s gene a es pos eplica-
i e ssDNA gaps h oughou he p imase and polyme ase ac i i ies
o PRIMPOL (14,19). We he e o e examined he con ibu ion o
PRIMPOL o 5hmdC- induced ssDNA gap o ma ion in ou cell
lines. PRIMPOL deple ion had li le e ec on PARyla ion le els in
un ea ed EdU+ Fancd2−/− cells. Ne e heless, PRIMPOL knock-
down la gely supp essed 5hmdC- media ed PARyla ion, no eaching
wild- ype le els ( ig. S12). In addi ion, PRIMPOL deple ion pa ially
supp essed 5hmdC- induced EdU- PAR SIRF oci in Fancd2−/− cells
(Fig. 8A). Mo eo e , PRIMPOL deple ion signi ican ly educed
5hmdC- induced ssDNA gaps in Fancd2−/− cells, bu no eaching
le els seen in he un ea ed condi ion (Fig. 8B). These da a sugges
ha a p opo ion o 5hmdC- induced ssDNA gaps depends on
PRIMPOL. This si ua ion con as ed o APEX1- deple ed cells,
whe eby APEX1 deple ion comple ely abolished 5hmdC- induced
ssDNA gaps in bo h backg ounds. These da a indica e ha , in addi-
ion o PRIMPOL- media ed ssDNA gaps, 5hmdC induces o he
ypes o ssDNA gaps, p obably a ising om un epai ed APEX1-
media ed SSBs. Consis en wi h his, PRIMPOL pa ially accoun ed
o he 5hmdC- induced le hali y obse ed in Fancd2−/− cells (Fig.
8C). APEX1 PRIMPOL o HMCES PRIMPOL double- knockdown
cells displayed he same iabili y as single PRIMPOL–deple ed
Fancd2−/− cells (Fig. 8C). These da a sugges ha a subse o ssDNA
gaps a ising om cy o oxic HMCES- DPCs is PRIMPOL dependen ,
whe eas o he s a e independen ly o med. We also examined he
con ibu ion o HMCES- DPCs on nascen DNA o 5hmdC-
dependen SCEs. As p e iously epo ed o bulky adduc s on em-
pla e DNA s ands (14), PRIMPOL deple ion e icien ly supp essed
heigh ened 5hmdC- dependen SCEs obse ed in Fancd2−/− cells.
Mo eo e , HMCES loss also blun ed his esponse, and codeple ion
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
16 o 17
REFERENCES AND NOTES
1. A. ubbs, A. nussenzweig, endogenous dnA damage as a sou ce o genomic ins abili y
in cance . Cell 168, 644–656 (2017).
2. . lindahl, An N- glycosidase om Esche ichia coli ha eleases ee u acil om dnA
con aining deamina ed cy osine esidues. P oc. Na l. Acad. Sci. U.S.A. 71, 3649–3653 (1974).
3. e. Seebe g, l. eide, M. Bjo as, he base excision epai pa hway. T ends Biochem. Sci. 20,
391–397 (1995).
4. . lindahl, dnA epai enzymes. Annu. Re . Biochem. 51, 61–87 (1982).
5. d. e. Ba nes, . lindahl, B. Sedgwick, dnA epai . Cu . Opin. Cell Biol. 5, 424–433 (1993).
6. . lindahl, Recogni ion and p ocessing o damaged dnA. J. Cell Sci. Suppl. 19, 73–77
(1995).
7. J. c. F omme, G. l. e dine, Base excision epai . Ad . P o ein Chem. 69, 1–41 (2004).
8. K. W. caldeco , Mammalian dnA base excision epai : dancing in he moonligh . DNA
Repai 93, 102921 (2020).
9. M. Ga cia- diaz, K. Bebenek, J. M. K ahn, l. Blanco, . A. Kunkel, l. c. Pede sen, A s uc u al
solu ion o he dnA polyme ase λ- dependen epai o dnA gaps wi h minimal
homology. Mol. Cell 13, 561–572 (2004).
10. P. J. McKinnon, K. W. caldeco , dnA s and b eak epai and human gene ic disease.
Annu. Re . Genomics Hum. Gene . 8, 37–55 (2007).
11. K. W. caldeco , Single- s and b eak epai and gene ic disease. Na . Re . Gene . 9,
619–631 (2008).
12. S. Mou on, S. Rod iguez- Acebes, M. i. Ma inez- Jimenez, S. Ga cia- Gomez, S. choc on,
l. Blanco, J. Mendez, Rep iming o dnA syn hesis a s alled eplica ion o ks by human
P imPol. Na . S uc . Mol. Biol. 20, 1383–1389 (2013).
13. A. Quine , d. J. Ma ins, A. . essoni, d. Bia d, A. Sa asin, A. S a y, c. F. Menck, anslesion
syn hesis mechanisms depend on he na u e o dnA damage in U - i adia ed human
cells. Nucleic Acids Res. 44, 5717–5731 (2016).
14. A. l. Pibe ge , A. Bow y, R. d. W. Kelly, A. K. Walke , d. Gonzalez- Acos a, l. J. Bailey,
A. J. dohe y, J. Mendez, J. R. Mo is, h. e. B yan , e. Pe e mann, P imPol- dependen
single- s anded gap o ma ion media es homologous ecombina ion a bulky dnA
adduc s. Na . Commun. 11, 5863 (2020).
15. J. l. Quinones, B. demple, When dnA epai goes w ong: BeR- gene a ed dnA- p o ein
c osslinks o oxida i e lesions. DNA Repai 44, 103–109 (2016).
16. S. h. Wilson, . A. Kunkel, Passing he ba on in base excision epai . Na . S uc . Biol. 7,
176–178 (2000).
17. A. Quine , S. i man, e. cybulla, A. Me oni, A. indigni, o skip o no o skip: choosing
ep iming o ole a e dnA damage. Mol. Cell 81, 649–658 (2021).
18. K. cong, S. B. can o , exploi ing eplica ion gaps o cance he apy. Mol. Cell 82,
2363–2369 (2022).
19. A. agliala ela, G. leuzzi, . Sannino, R. cuella- Ma in, J. W. huang, F. Wu- Bae , R. Bae ,
. cos anzo, A. ciccia, Re 1- Polζ main ains he iabili y o homologous ecombina ion-
de icien cance cells h ough mu agenic epai o PRiMPOl- dependen ssdnA gaps. Mol.
Cell 81, 4008–4025.e7 (2021).
20. A. Mann, M. A. Rami ez- O e o, A. de An oni, Y. W. han hi, . Sannino, G. Baldi, l. Falbo,
A. Sch emp , S. Be na do, J. loizou, . cos anzo, POlθ p e en s MRe11- nBS1- c iP-
dependen o k b eakage in he absence o BRcA2/RAd51 by illing lagging- s and gaps.
Mol. Cell 82, 4218–4231.e8 (2022).
21. O. Belan, M. Sebald, M. Adamowicz, R. Anand, A. ance ska, J. ne es, . G inke ich,
G. hewi , S. Segu a- Bayona, R. Bellelli, h. M. R. Robinson, G. S. higgins, G. c. M. Smi h,
S. c. Wes , d. S. Rueda, S. J. Boul on, POlQ seals pos - eplica i e ssdnA gaps o main ain
genome s abili y in BRcA- de icien cance cells. Mol. Cell 82, 4664–4680.e9 (2022).
22. h. e. B yan , n. Schul z, h. d. homas, K. M. Pa ke , d. Flowe , e. lopez, S. Kyle, M. Meu h,
n. J. cu in, . helleday, Speci ic killing o BRcA2- de icien umou s wi h inhibi o s o
poly(AdP- ibose) polyme ase. Na u e 434, 913–917 (2005).
23. h. Fa me , n. Mccabe, c. J. lo d, A. n. u , d. A. Johnson, . B. Richa dson, M. San a osa,
K. J. dillon, i. hickson, c. Knigh s, n. M. Ma in, S. P. Jackson, G. c. Smi h, A. Ashwo h,
a ge ing he dnA epai de ec in BRcA mu an cells as a he apeu ic s a egy. Na u e
434, 917–921 (2005).
24. P. X. lim, M. Zaman, W. Feng, M. Jasin, BRcA2 p omo es genomic in eg i y and he apy
esis ance p ima ily h ough i s ole in homology- di ec ed epai . Mol. Cell 84, 447–462.
e10 (2024).
25. M. J. Pena- Gomez, P. Mo eno- Go dillo, M. na mon e, c. B. Ga cia- calde on, A. Ruksenai e,
S. Klimasauskas, i. . Rosado, FAncd2 main ains eplica ion o k s abili y du ing
misinco po a ion o he dnA deme hyla ion p oduc s 5- hyd oxyme hyl- 2′- deoxycy idine
and 5- hyd oxyme hyl- 2′- deoxyu idine. Cell Dea h Dis. 13, 503 (2022).
26. M. J. Pena- Gomez, M. Sua ez- Piza o, i. . Rosado, XRcc1 p e en s eplica ion
o k ins abili y du ing misinco po a ion o he dnA deme hyla ion bases
5- hyd oxyme hyl- 2′- deoxycy idine and 5- hyd oxyme hyl- 2′- deoxyu idine. In . J. Mol. Sci.
23, 893 (2022).
27. K. Fugge , i. Baj ami, M. Sil a dos San os, S. J. Young, S. Kunzelmann, G. Kelly, G. hewi ,
h. Pa el, R. Golds one, . ca ell, S. J. Boul on, J. MacRae, i. A. aylo , S. c. Wes , a ge ing
he nucleo ide sal age ac o dnPh1 sensi izes BRCA- de icien cells o PARP inhibi o s.
Science 372, 156–165 (2021).
28. A. Se ano- Beni ez, S. e. Wells, l. d ummond- cla ke, l. c. Russo, J. c. homas, G. A. leal,
M. Fa ow, J. M. edge on, S. Balasub amanian, M. Yang, c. F ezza, A. Gau am, J. B azina,
K. Bu do a, n. c. hoch, S. P. Jackson, K. W. caldeco , Un epai ed base excision epai
in e media es in empla e dnA s ands igge eplica ion o k collapse and PARP
inhibi o sensi i i y. EMBO J. 42, e113190 (2023).
29. S. Saxena, c. S. nabel, . W. Seay, P. S. Pa el, A. S. Kawale, c. R. c osby, h. ig o, e. Oh,
M. G. ande heiden, A. n. ha a, Z. Suo, l. Zou, Unp ocessed genomic u acil as a sou ce o
dnA eplica ion s ess in cance cells. Mol. Cell 84, 2036–2052.e7 (2024).
30. S. Ga cia- Gomez, A. Reyes, M. i. Ma inez- Jimenez, e. S. choc on, S. Mou on, G. e ados,
c. Powell, e. Salido, J. Mendez, i. J. hol , l. Blanco, P imPol, an a chaic p imase/polyme ase
ope a ing in human cells. Mol. Cell 52, 541–553 (2013).
31. e. cybulla, A. indigni, le e aging he eplica ion s ess esponse o op imize cance
he apy. Na . Re . Cance 23, 6–24 (2023).
32. K. cong, M. Peng, A. n. Koushol , W. . c. lee, S. lee, S. nayak, J. K ais,
P. S. ande e e- ca ozza, K. S. Pawelczak, J. cal o, n. J. Panza ino, J. J. u chi, n. Johnson,
J. Jonke s, e. Ro henbe g, S. B. can o , Replica ion gaps a e a key de e minan o PARP
inhibi o syn he ic le hali y wi h BRcA de iciency. Mol. Cell 81, 3128–3144.e7 (2021).
33. A. Quine , S. i man, J. Jackson, S. S iko ic, d. lemacon, d. ca ajal- Maldonado,
d. Gonzalez- Acos a, A. . essoni, e. cybulla, M. Wood, S. a is, l. F. Z. Ba is a, J. Mendez,
J. e. Sale, A. indigni, PRiMPOl- media ed adap i e esponse supp esses eplica ion o k
e e sal in BRcA- de icien cells. Mol. Cell 77, 461–474.e9 (2020).
34. l. halabelian, M. Ra ichand an, Y. li, h. Zeng, A. Rao, l. A a ind, c. h. A owsmi h,
S uc u al basis o hMceS in e ac ions wi h abasic dnA and mul i alen subs a e
ecogni ion. Na . S uc . Mol. Biol. 26, 607–612 (2019).
35. K. n. Mohni, S. R. Wessel, R. Zhao, A. c. Wojciechowski, J. W. luzwick, h. layden,
B. F. eichman, P. S. hompson, K. P. M. Meh a, d. co ez, hMceS main ains genome
in eg i y by shielding abasic si es in single- s and dnA. Cell 176, 144–153.e13 (2019).
36. M. S i as a a, d. Su, h. Zhang, Z. chen, M. ang, l. nie, J. chen, hMceS sa egua ds
eplica ion om oxida i e s ess and ensu es e o - ee epai . EMBO Rep. 21, e49123
(2020).
37. M. donsbach, S. du aue , F. G une , K. . nguyen, R. nigam, d. Yane a, P. Weicke ,
R. Bezalel- Buch, d. R. Semlow, J. S ingele, A non- p o eoly ic elease mechanism o
hMceS- dnA- p o ein c osslinks. EMBO J. 42, e113360 (2023).
38. J. Rua-Fe nandez, c. A. lo ejoy, K. P. M. Meh a, K. A. Paulin, Y. . oudji, c. Giansan i,
B. F. eichman, d. co ez, Sel - e e sal acili a es he esolu ion o hMceS dnA-p o ein
c osslinks in cells. Cell Rep. 42, 113427 (2023).
39. d. Yane a, J. l. Spa ks, M. donsbach, S. Zhao, P. Weicke , R. Bezalel- Buch, J. S ingele,
J. c. Wal e , he FAncJ helicase un olds dnA- p o ein c osslinks o p omo e hei epai .
Mol. Cell 83, 43–56.e10 (2023).
40. K. P. M. Meh a, c. A. lo ejoy, R. Zhao, d. R. hein zman, d. co ez, hMceS main ains
eplica ion o k p og ession and p e en s double- s and b eaks in esponse o APOBec
deamina ion and abasic si e o ma ion. Cell Rep. 31, 107705 (2020).
41. A. agliala ela, S. Al a ez, G. leuzzi, . Sannino, l. Ranjha, J. W. huang, c. Maduba a,
R. Anand, B. le y, R. Rabadan, P. cejka, . cos anzo, A. ciccia, Res o a ion o eplica ion
o k s abili y in BRcA1- and BRcA2- de icien cells by inac i a ion o SnF2- amily o k
emodele s. Mol. Cell 68, 414–430.e8 (2017).
42. A. Masaoka, M. Ma suba a, R. hasegawa, . anaka, S. Ku isu, h. e a o, Y. Ohyama,
n. Ka ino, A. Ma suda, h. ide, Mammalian 5- o mylu acil- dnA glycosylase. 2. Role o
SMUG1 u acil- dnA glycosylase in epai o 5- o mylu acil and o he oxidized and
deamina ed base lesions. Biochemis y 42, 5003–5012 (2003).
43. . c. liu, c. . lin, K. c. chang, K. W. Guo, S. Wang, J. W. chu, Y. Y. hsiao, APe1 dis inguishes dnA
subs a es in exonucleoly ic clea age by induced space- illing. Na . Commun. 12, 601 (2021).
44. P. Bu ko ics, i. hajdu, . Szukacso , i. Unk, l. ha acska, Role o PcnA- dependen
s imula ion o 3′- phosphodies e ase and 3′- 5′ exonuclease ac i i ies o human Ape2 in
epai o oxida i e dnA damage. Nucleic Acids Res. 37, 4247–4255 (2009).
45. A. A. demin, K. hi o a, M. suda, M. Adamowicz, R. hails one, J. B azina, W. Gi ens,
i. Kalaso a, Z. Shao, S. Zha, h. Sasanuma, h. hanzliko a, S. akeda, K. W. caldeco , XRcc1
p e en s oxic PARP1 apping du ing dnA base excision epai . Mol. Cell 81, 3018–3030.e5
(2021).
46. J. Biayna, i. Ga cia- cao, M. M. Al a ez, M. Sal ado es, J. espinosa- ca asco, M. Mccullough,
F. Supek, . h. S acke , loss o he abasic si e senso hMceS is syn he ic le hal wi h he
ac i i y o he APOBec3A cy osine deaminase in cance cells. PLOS Biol. 19, e3001176 (2021).
47. Y. W. han hi, M. A. Rami ez- O e o, R. Appleby, A. de An oni, l. Joudeh, . Sannino,
S. Waked, A. A dizzoia, . Ba a, d. Fachine i, l. Pelleg ini, . cos anzo, RAd51 p o ec s
abasic si es o p e en eplica ion o k b eakage. Mol. Cell 84, 3026–3043.e11 (2024).
48. Y. lin, J. Raj, J. li, A. ha, M. A. hossain, c. Richa dson, P. Mukhe jee, S. Yan, APe1 senses dnA
single- s and b eaks o epai and signaling. Nucleic Acids Res. 48, 1925–1940 (2020).
49. M. A. hossain, Y. lin, G. d iscoll, J. li, A. McMahon, J. Ma os, h. Zhao, d. suchimo o,
Y. nakabeppu, J. Zhao, S. Yan, APe2 is a gene al egula o o he A R- chk1 dnA damage
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
Peña-Gómez e al., Sci. Ad . 11, eads3227 (2025) 26 Ma ch 2025
Science Ad AnceS | ReSeARch AR icle
17 o 17
esponse pa hway o main ain genome in eg i y in panc ea ic cance cells. F on . Cell De .
Biol. 9, 738502 (2021).
50. A. B amba i, O. Sacco, S. Po cella, J. heyza, M. Ka eh, J. c. Schmid , A. S ei , RhinO di ec s
MMeJ o epai dnA b eaks in mi osis. Science 381, 653–660 (2023).
51. Y. Sugimo o, Y. Masuda, S. iwai, Y. Miyake, R. Kanao, c. Masu ani, no el mechanisms o
he emo al o s ong eplica ion- blocking hMceS- and hiazolidine- dnA adduc s in
humans. Nucleic Acids Res. 51, 4959–4981 (2023).
52. R. Zellwege , d. dalche , K. Mu eja, M. Be i, J. A. Schmid, R. he ado , A. indigni,
M. lopes, Rad51- media ed eplica ion o k e e sal is a global esponse o geno oxic
ea men s in human cells. J. Cell Biol. 208, 563–579 (2015).
53. B. Gyo y, Su i al analysis ac oss he en i e ansc ip ome iden i ies bioma ke s wi h he
highes p ognos ic powe in b eas cance . Compu . S uc . Bio echnol. J. 19, 4101–4109
(2021).
54. A. hale, A. dhoonmoon, J. S aka, c. M. nicolae, G. l. Moldo an, Mul i- s ep p ocessing o
eplica ion s ess- de i ed nascen s and dnA gaps by MRe11 and eXO1 nucleases. Na .
Commun. 14, 6265 (2023).
55. n. Ga cia- Rod iguez, i. dominguez- Ga cia, M. d. c. dominguez- Pe ez, P. hue as, eXO1
and dnA2- media ed ssdnA gap expansion is essen ial o A R ac i a ion and o main ain
iabili y in BRcA1- de icien cells. Nucleic Acids Res. 52, 6376–6391 (2024).
56. S. Roy, K. Schlache , SiRF: A single- cell assay o in si u p o ein in e ac ion wi h nascen
dnA eplica ion o ks. Bio P o oc. 9, e3377 (2019).
57. G. ci iello, M. l. Ga za, A. h. Beck, M. d. Wilke son, S. K. Rhie, A. Pas o e, h. Zhang,
M. Mclellan, c. Yau, c. Kando h, R. Bowlby, h. Shen, S. haya , R. Fieldhouse, S. c. les e ,
G. M. se, R. e. Fac o , l. c. collins, K. h. Allison, Y. Y. chen, K. Jensen, n. B. Johnson,
S. Oes e eich, G. B. Mills, A. d. che niack, G. Robe son, c. Benz, c. Sande , P. W. lai d,
K. A. hoadley, . A. King, . R. ne wo k, c. M. Pe ou, comp ehensi e molecula po ai s o
in asi e lobula b eas cance . Cell 163, 506–519 (2015).
Acknowledgmen s: RPe- 1 TP53−/− o dld- 1 BRcA2−/− cell lines we e p o ided by P. hue as
(Uni e si y o Se ille/cABiMeR, Spain) and A. Aguile a (Uni e si y o Se ille/cABiMeR, Spain),
espec i ely. We hank F. Zhang (B oad ins i u e, USA) and h. leonha d (lMU, Munich) o
p o iding he pSpcas9(BB)- 2A- GFP (Addgene, #48138) and pcAG- BSR plasmids, espec i ely.
We also acknowledge n. S - denis (high- Fideli y Science communica ions l d.) o manusc ip
edi ion. We a e in deb o J. duxin (no o no disk Founda ion cen e o P o ein Resea ch,
U. copenhagen, dK), G. P. c ossan (As a Zeneca, UK), and l. B. Pon el (ins i u o Josep ca e as,
Spain) o c i ical eading o he manusc ip and M. Gianna asio o he iFOM dnA eM acili y
o assis ance. Funding: his publica ion is pa o he p ojec Pid2021- 128988OB- i00, unded
by MiciU/Aei/10.13039/501100011033 and by eRdF/eU. his esea ch wo k was also
suppo ed by cnS2022- 136055 inanciado po Mcin/Aei/10.13039/501100011033/
Unióneu opeanex Gene a ioneU/PR R. J.Y.M. is a ie i canada Resea ch chai in dnA Repai
and cance he apeu ics and was suppo ed by a canadian ins i u es o heal h Resea ch
Founda ion G an (Fdn- 388879). .c. ecei ed unding om AiRc unde iG 2023- id 28725
p ojec . M.J.P.- G. was suppo ed by conseje ía de ans o mación económica, indus ia,
conocimien o y Uni e sidades (PRedOc_00505). Y.R.- M. was suppo ed by Pid2021-
128988OB- i00/Mcin/Aei/10.13039/501100011033/FedeR, Ue. M.d.R.O. was suppo ed by
cnS2022- 136055/Mcin/Aei/10.13039/501100011033/Unión eu opea nex Gene a ioneU/
PR R. Y.W.h. was suppo ed by he 26596 AiRc ellowship o i aly. S.B. is suppo ed by an
FRQ- S pos doc o al ellowship. Au ho con ibu ions: M.J.P.- G.: in es iga ion and
me hodology, da a cu a ion, alida ion, o mal analysis, and isualiza ion. Y.R.- M.:
in es iga ion, alida ion, o mal analysis, and isualiza ion. M.d.R.O.: in es iga ion,
me hodology, alida ion, o mal analysis, and isualiza ion. Y.W.h.: in es iga ion, me hodology,
da a cu a ion, alida ion, o mal analysis, and isualiza ion. S.B.: in es iga ion and w i ing—
e iew and edi ing. R.F.: W i ing— e iew and edi ing and esou ces. J.Y.M.: W i ing— e iew
and edi ing, unding acquisi ion, and supe ision. J.c.R.: in es iga ion. .c.: concep ualiza ion,
in es iga ion, w i ing— e iew and edi ing, esou ces, unding acquisi ion, and supe ision.
i. .R.: W i ing—o iginal d a , concep ualiza ion, in es iga ion, w i ing— e iew and edi ing,
me hodology, esou ces, unding acquisi ion, da a cu a ion, alida ion, supe ision, o mal
analysis, p ojec adminis a ion, and isualiza ion. Compe ing in e es s: he au ho s decla e
ha hey ha e no compe ing in e es s. Da a and ma e ials a ailabili y: All da a needed o
e alua e he conclusions in he pape a e p esen in he pape and/o he Supplemen a y
Ma e ials. Raw da a a e a ailable a Zenodo (h ps://zenodo.o g/ eco ds/14738090).
Submi ed 7 Augus 2024
Accep ed 21 Feb ua y 2025
Published 26 Ma ch 2025
10.1126/sciad .ads3227
Downloaded om h ps://www.science.o g a Uni e sidad de Se illa on May 12, 2025
