Spontaneous deamination of cytosine to uracil is biased to the non-transcribed DNA strand in yeast

Depósi o de in es igación de la Uni e sidad de Se illa h ps://idus.us.es/ “This is an Accep ed Manusc ip o an a icle published by Else ie in DNA Repai on June 2023, a ailable a : h ps://doi.o g/10.1016/j.dna ep.2023.103489 .” Spon aneous deamina ion o cy osine o u acil is biased o he non- ansc ibed DNA s and in yeas Jona han D Williamsa,1, Demi Zhua, Ma ía Ga cía-Rubiob, Saman ha Shal za, And és Aguile ab, Sue Jinks-Robe sona,* aDepa men o Molecula Gene ics and Mic obiology, 213 Resea ch D ., Duke Uni e si y Medical Cen e , Du ham, NC 27710 bCen o Andaluz de Biología Molecula y Medicina Regene a i a (CABIMER), Uni e sidad de Se illa-CSIC, Se ille, Spain Abs ac T ansc ip ion in Saccha omyces ce e isiae is associa ed wi h ele a ed mu a ion and his pa ially e lec s enhanced damage o he co esponding DNA. Spon aneous deamina ion o cy osine o u acil leads o CG>TA mu a ions ha p o ide a s and-speci ic ead-ou o damage in s ains ha lack he abili y o emo e u acil om DNA. Using he CAN1 o wa d mu a ion epo e , we ound ha C>T and G>A mu a ions, which e lec deamina ion o he non- ansc ibed and ansc ibed DNA s ands, espec i ely, occu ed a simila a es unde low- ansc ip ion condi ions. By con as , he a e o C>T mu a ions was 3- old highe han G>A mu a ions unde high- ansc ip ion condi ions, demons a ing biased deamina ion o he non- ansc ibed s and (NTS). The NTS is ansien ly single-s anded wi hin he ~15 bp ansc ip ion bubble, o a mo e ex ensi e egion o he NTS can be exposed as pa o an R-loop ha can o m behind RNA polyme ase. Nei he he dele ion o genes whose p oduc s es ain R-loop o ma ion no he o e -exp ession o RNase H1, which deg ades R-loops, educed he biased deamina ion o he NTS, and no ansc ip ion-associa ed R-loop o ma ion a CAN1 was de ec ed. These esul s sugges ha he NTS wi hin he ansc ip ion bubble is a a ge o spon aneous deamina ion and likely o he ypes o DNA damage. Keywo ds ansc ip ion; mu agenesis; RNA polyme ase; R-loops; ansc ip ion bubble; deamina ion 1. INTRODUCTION Mu a ions gene ally ha e nega i e consequences bu a e equi ed o adap a ion unde s ess condi ions and o e olu iona y change. Mos mu a ions a ise du ing genome duplica ion and e lec ei he a e e o s o he high- ideli y eplica i e DNA polyme ases o he bypass o DNA damage by e o -p one anslesion-syn hesis DNA polyme ases. In addi ion o *Communica ing au ho : Depa men o Molecula Gene ics and Mic obiology, 213 Resea ch D ., Duke Uni e si y Medical Cen e , Du ham, NC 27710, Phone: 919-681-7273, [email p o ec ed]. 1Cu en add ess: Labco p, Bu ling on, NC HHS Public Access Au ho manusc ip DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Published in inal edi ed o m as: DNA Repai (Ams ) . 2023 June ; 126: 103489. doi:10.1016/j.dna ep.2023.103489. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip eplica ion-associa ed changes in DNA sequence, mu a ions also a ise in he con ex o ansc ip ion. This may be a pa icula ly impo an sou ce o gene ic al e a ions in he pos -mi o ic cells o me azoans, such as hose in neu al issue [1]. In mic oo ganisms, ansc ip ion-associa ed mu agenesis (TAM) is ope a ionally de ined as an inc ease in he mu a ion a e o a a ge gene when i is ansc ibed a a high e sus low le el ( e iewed in [2]). Ea ly s udies in Saccha omyces ce e isiae demons a ed ha TAM inc eases when excision- epai o e o - ee damage-bypass pa hways a e inac i a ed, and dec eases when anslesion syn hesis is dis up ed [3]. The e also is syne gism be ween he s imula o y e ec s o exogenous mu agens and ansc ip ion on ecombina ion a es [4]. Collec i ely, hese gene ic da a indica e ha ansc ip ion inc eases he exposu e o he unde lying DNA empla e o endogenous and exogenous damage. Mu agenic damage can also be in lic ed by opoisome ase I (Top1), which esol es he o sional s ess ha accumula es when DNA s ands a e sepa a ed du ing ansc ip ion [5, 6]. Top1-dependen mu agenesis speci ically gene a es small dele ions and e lec s sequen ial clea age o one DNA s and, wi h he ini ial incision occu ing a an embedded ibonucleo ide [7, 8]. Bo h eplica ion and ansc ip ion impa an asymme y be ween he wo s ands o duplex DNA. Du ing yeas eplica ion, complemen a y s ands a e copied by dis inc DNA polyme ases and he e iciencies o eplica ion-e o emo al di e , which can lead o mu a ion asymme ies on he leading and lagging s ands ( e iewed in [9]). Du ing ansc ip ion, only one s and ypically se es as he empla e o RNA syn hesis. The empla e s and is e e ed o as he ansc ibed s and o TS, while he complemen a y s and is he non- ansc ibed s and o NTS. S and- ela ed biases a e eadily e iden when mu a ions a e epo ed wi h e e ence o only one DNA s and and bo h eplica ion- and ansc ip ion-associa ed mu a ion biases ha e been epo ed in cance genomes [10]. S and- ela ed biases associa ed wi h ansc ip ion can e lec mo e obus epai o he TS han NTS and/o mo e damage o he NTS han TS. A lesion on he TS ha blocks/ s alls RNA polyme ase, o example, igge s mo e e icien emo al by he nucleo ide excision epai pa hway han does he same lesion on he NTS ( ansc ip ion-coupled epai ; e iewed in [11]). In addi ion, single-s anded DNA is mo e suscep ible o chemical and enzyma ic damage han is double-s anded DNA ( e iewed in [12, 13]). Mo e damage o he NTS ela i e o he TS p esumably e lec s i s enhanced single-s anded cha ac e . Du ing ansc ip ion, 10–15 n o he NTS a e exposed wi hin he RNA polyme ase (RNAP) ansc ip ion bubble, whe eas he TS is p o ec ed h ough i s pai ing wi h he nascen ansc ip . A much la ge segmen o he NTS can be exposed when he RNA h eads back a e exi ing RNAP and pai s wi h he empla e s and [14]. The ex ended RNA-DNA hyb id, oge he wi h he single-s anded NTS, is e e ed o as an R-loop. R-loops a e impo an o ansc ip ion e mina ion, bu con lic s wi h he eplisome can gene a e DSBs ha a e po en ini ia o s o homologous ecombina ion and gene ic ins abili y ( e iewed in [15]). The con ibu ion o R-loops o he accumula ion o spon aneous DNA damage is less clea . While ansc ip ion-associa ed gene ic ins abili y is gene ally p oblema ic, i is physiologically impo an in he ma u a ion o he e eb a e immune sys em. An igen- s imula ed soma ic hype mu a ion (SHM) imp o es an ibody a ini y, while class-swi ch Williams e al. Page 2 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip ecombina ion (CSR) di e si ies an ibody e ec o unc ions ( e iewed in [16]). Bo h p ocesses a e limi ed o B cells and equi e enzyma ic deamina ion o cy osine o u acil by he single-s and speci ic enzyme AID (ac i a ion-induced cy osine deaminase). In i o , AID a els wi h phage RNAP [17] and p ima ily deamina es he NTS, leading o he sugges ion ha i also a ge s he ansc ip ion bubble. A mo e ecen s udy wi h euka yo ic RNAPII, howe e , demons a ed equal deamina ion o he wo DNA s ands by AID and i was sugges ed ha bo h become accessible in a back acked ansc ip ion bubble [18]. By con as , o he epo s ha e shown ha deamina ion du ing CSR in ol es R-loop o ma ion [19]. AID exp ession in yeas was epo ed o ha e an NTS bias in an RNAPII- ansc ibed gene, bu only in a mu an backg ound whe e R-loops we e expec ed o accumula e a a high le el [20]. The me azoan-speci ic APOBEC (apolipop o ein B mRNA edi ing enzyme, ca aly ic pep ide) amily o cy osine deaminases a ge i al genomes and a e an impo an componen o inna e immuni y [21]. APOBECs also obus ly deamina e single-s anded genomic DNA, howe e , and a e po en d i e s o umo igenesis [22]. In yeas , APOBEC3G p ima ily a ge s he NTS o RNA genes [23], which a e s ong si es o R-loop accumula ion [24]. The NTS bias in RNA genes was exace ba ed by loss o RNase H, which deg ades he RNA componen o RNA-DNA hyb ids, consis en wi h a p ima y a ge ing o single-s anded DNA in he con ex o R-loops. An APOBEC om lamp ey also a ge ed he NTS in a genome-wide analysis in yeas , bu wi h p e e en ial deamina ion o ansc ip ion s a si es ha con ain s able p e-ini ia ion complexes [25]. While he single-s anded DNA o R-loops can be clea ly a ge ed by cy osine deaminases, a ole o R-loops in spon aneous deamina ion has no been epo ed. A p ima y ole o he ansc ip ion bubble in NTS-biased cy osine deamina ion, howe e , was p e iously in e ed in bac e ial cells [26], whe e co- ansc ip ional ansla ion is assumed o limi R-loop o ma ion. The apid assembly o yeas ansc ip s in o ibonucleop o ein pa icles simila ly seques e s ansc ip s o limi R-loop o ma ion [27]. Al hough mos wo k o da e has sugges ed g ea e suscep ibili y o he NTS o a ious ypes o DNA damage, he e e se bias was epo ed a a bac e ial p omo e when he ansc ip ion and eplica ion o ks we e in a head-on o ien a ion [28]. I was sugges ed ha his e lec ed a p o ec i e in e ac ion o sigma ac o wi h he NTS in a s abilized RNAP open complex. In he cu en s udy, we examined he spon aneous deamina ion o cy osine o u acil in yeas as a unc ion o low- e sus high- ansc ip ion o he CAN1 o wa d-mu a ion epo e . Mu a ions a ibu able o cy osine deamina ion occu ed a simila a es on he NTS and TS s ands unde low- ansc ip ion condi ions bu we e s ongly biased o he NTS unde high- ansc ip ion condi ions. No co ela ion be ween he ansc ip ion-associa ed mu a ion bias and R-loop accumula ion was de ec ed, howe e , sugges ing ha he NTS exposed in he ansc ip ion bubble may be he p ima y a ge o DNA damage in some con ex s. Williams e al. Page 3 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip 2. MATERIALS AND METHODS 2.1. S ain cons uc ion All s ains used o mu a ional analyses (Table S1) we e de i ed om SJR282 ( MATα ade2–101OC his3 Δ 200 u a3 Δ Nco gal80 Δ ::HIS3 ) by ans o ma ion. As p e iously desc ibed [5], high ansc ip ion o CAN1 was achie ed by eplacing he endogenous p omo e wi h he GAL1 p omo e ( pGAL ) linked o he bac e ial kan ma ke [29]. FCY1, MSH6 , MFT1 , RNH1 , RNH201 , TOP1 o UNG1 was dele ed by one-s ep allele eplacemen using PCR- gene a ed casse es con aining an app op ia e selec able ma ke . Fo expe imen s in ol ing RNase H1 o e exp ession, he high- xn ung1 Δ s ain was ans o med wi h he URA3-CEN ec o pRS316 [30] o a de i a i e con aining a pGAL1-RNH1 usion [31]. 2.2. Mu a ion a es and spec a A minimum o 24 cul u es o each s ain was used o de e mining cana anine- esis ance (Can-R) a es. Each cul u e was inocula ed wi h a single colony and was g own o sa u a ion ( h ee days) in YEP-GE (1% yeas ex ac , 2% Bac o-pep one, 2% glyce ol and 2% e hanol) on a olle d um a 30°C. In a gal80 backg ound, pGAL is cons i u i ely ac i e in YEP-GE medium [32]. Cul u es we e washed wi h wa e and he numbe o Can-R mu an s in each was de e mined by pla ing an app op ia e dilu ion on syn he ic comple e medium (SC; 0.17% yeas ni ogen base, 0.5% ammonium sul a e, 2% aga and 0.13% Ha well’s comple e amino acid mix) lacking a ginine and supplemen ed wi h 60 μg/ml cana anine (Sigma). To al cell numbe was es ima ed by pla ing an app op ia e dilu ion on non-selec i e YPD medium (1% yeas ex ac , 2% Bac o-pep one, 2% dex ose and 2% aga ). Mu a ion a es we e calcula ed using me hod o he median [33], and 95% con idence in e als (CIs) we e de e mined as p e iously desc ibed [34]. Independen Can-R mu an s o sequencing we e ob ained using a p ong-pla ing echnique ha gene a es ~100 mini-cul u es/pla e. B ie ly, cells we e dilu ed in wa e and a 100-coun , la - ipped cus om pinning de ice was used o ans e ~103 cells/pin on o non-selec i e YEP-GE pla es. Cells we e also spo ed on o SC-A g+Can pla es o ensu e he e we e no p e-exis ing Can-R mu an s. A e h ee days o g ow h, cells we e eplica-pla ed on o SC-A g+Can pla es and incuba ed o wo days o allow he appea ance o Can-R colonies. A single Can-R colony was picked om each spo and ollowing genomic DNA isola ion he CAN1 locus was ampli ied using MyTaq (Bioline) polyme ase. Fo PacBio analysis, 16-n ba code (a lis a ailable is a h ps://gi hub.com/Paci icBiosciences/Bioin o ma ics-T aining/blob/mas e /ba coding/ pacbio_384_ba codes. as a) was a ached o he 5′ end o each CAN1 -speci ic p ime and DNA o each mu an was ampli ied using a unique pai o ba codes. Ba codes o Can-R mu an s isola ed om he low- ansc ip ion, pCAN- CAN1 s ains o high- ansc ip ion pGAL-CAN1 s ains we e a ached o o wa d p ime s 5′-CAGTTTTTAATCTGTCGTCAATCGAAAG o 5′-GCCGAGCGGGTGACAG, espec i ely. Ba codes we e a ached o single CAN1 e e se p ime (5′- GGGAGCAAGATTGTTGTGGT). Following p oduc quan i ica ion using ImageJ so wa e (h ps://imagej.nih.go /ij/), ba coded amplicons we e pooled in equal amoun s o Williams e al. Page 4 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip subsequen SMRT lib a y p epa a ion. SMRT lib a ies we e cons uc ed and sequenced a Duke Cen e o Genomic and Compu a ional Biology using he PacBio RSII and Sequel sys ems. Ci cula consensus eads we e so ed and analyzed using an in-house pipeline [35]. The a es o C>T and G>A mu a ions we e calcula ed as he p oduc o he o al Can-R a e and he p opo ion o he mu a ion ype in he co esponding spec um. 95% CIs we e assigned o mu a ion p opo ions using he Vassa s a s web si e ( assa s a s.ne ). To assign uppe and lowe con idence in e als o a C>T o G>A mu a ion a e, he 95% CIs o bo h he co esponding Can-R a e and he mu a ion p opo ion we e combined using he oo o he squa e o he sums [36]. Fo expe imen s in ol ing he o e -exp ession o RNase H1, he high- ansc ip ion ung1 Δ mu an was ans o med wi h emp y ec o o wi h a plasmid in which RNH1 was egula ed by pGAL1 . Mu a ion a es we e calcula ed and Can-R mu an s we e isola ed as desc ibed abo e, excep ha all p e-g ow h was in SC medium con aining 2% glyce ol and 2% E OH as ca bon sou ces and missing u acil in o de o main ain plasmids. Following DNA isola ion and PCR ampli ica ion o CAN1 ( o wa d and e e se p ime s 5′- TTGGTATGATTGCCCTTGGT and 5′-ACAGCGTTGAAGATATGTGGC, espec i ely), nonsense mu a ions be ween posi ions 412 and 1056 we e cap u ed by Sange sequencing (p ime 5′-TATCCACACCTCTGACCAAC). This egion con ains six CAG/CAA codons and nine TGG codons o moni o C>T and G>A nonsense mu a ions, espec i ely. 2.3. DNA-RNA Immunop ecipi a ion (DRIP) Cul u es (100 ml) g own o mid-log phase in YEP con aining 2% galac ose we e washed wi h chilled H2O, esuspended in 1.4 ml sphe oplas ing bu e (1 M so bi ol, 10 mM EDTA pH 8, 0.1% β-me cap oe hanol, 2 mg/ml Zymolase 20T) and incuba ed a 30° o 30 min. Sphe oplas s we e pelle ed (5 min a 4000 pm), insed wi h H2O and esuspended in 1.65 ml o bu e G2 (800 mM guanidine HCl, 30 mM T is-Cl pH 8, 30 mM EDTA pH 8, 5% Tween-20, 0.5% T i on X-100). Samples we e ea ed wi h 10 μl o 10 mg/ml RNase A o 30 min a 37° and hen wi h 75 μl o 20 mg/ml p o einase K (Roche) o 1 h a 50°. Genomic DNA was ex ac ed gen ly wi h chlo o o m:isoamyl alcohol (24:1). DRIP was pe o med as p e iously desc ibed [37] wi h he ollowing modi ica ions. P ecipi a ed genomic DNA was spooled on o a glass od, washed wice wi h 70% E OH, esuspended gen ly in 250 μl TE (10 mM T is-HCl pH 8, 1 mM EDTA pH 8) con aining 2 mM spe midine and 2.5 μl o 10 mg/ml BSA, and diges ed o e nigh wi h 50U each o Hind III, EcoR I, Bs G I, Xba I and Ssp I. Hal o he DNA was ea ed wi h 3 μl RNase H1 (In i ogen Ambion; 1U/μl) o e nigh a 37° o an RNase H1- ea ed con ol. RNase H1- ea ed and un ea ed samples we e incuba ed wi h 10 μl o Dynabead-coupled S9.6 monoclonal an ibody (1 mg/ml) in 500 μl binding bu e (10 mM NaPO4 pH 7.0, 140 mM NaCl, 0.05% T i on X-100) o e nigh a 4°C and hen washed h ee imes wi h binding bu e . The S9.6 an ibody was coupled o Dynabeads P o ein A (In i ogen) o 2 h a 4°. DNA was elu ed in 100 μl elu ion bu e (50 mM T is pH 8.0, 10 mM EDTA, 0.5% SDS), ea ed 45 min wi h 7 μl p o einase K (20 mg/ml) a 55° and pu i ied wi h he Mache ey- Nagel DNA pu i ica ion ki . Real- ime quan i a i e PCR was pe o med using iTaq Uni e sal SYBR G een (Bio ad) and a 7500 Real-Time PCR machine (Applied Biosys ems). Williams e al. Page 5 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Fo wa d and e e se p ime s o PDC1 we e 5′-GAAGGTATGAGATGGGCTGGTAA and 5’ CCTTGATACGAGCGTAACCATCA, espec i ely. CAN1 -speci ic o wa d and e e se p ime s we e 5′ GGTTTTCTTGGGCAATCACTT and 5′ AATTGAATGACTTGGCCAACTACA, espec i ely. Th ee biological eplica es we e pe o med o each s ain; mean and s anda d de ia ions a e epo ed. As a posi i e con ol o R-loop de ec ion in DRIP expe imen s, genomic DNA ex ac ed om W303 and an sp 16–11 de i a i e was used. 2.4. Whole genome analysis o spon aneous deamina ion si es The da a o 162 C>T mu a ions de ec ed om whole-genome sequencing o an ung1 s ain passaged o 81 days was downloaded om Chen e al. [38]. The epo ed genome loca ion and lanking sequences we e used o map he exac loca ion o mu a ions in he Saccha omyces Genome Da abase ( e sion 6CA_00146045.2). The deamina ed s and (NTS o TS) was de e mined o mu a ions in RNAPII- ansc ibed genes using elease 91 o he Ensembl genome b owse . The ansc ip le el o each gene was downloaded [39, 40] and genes we e anked based on exp ession le el. 2.5. Addi ional s a is ical analyses The numbe s o C>T e sus G>A mu a ions obse ed in di e en s ain backg ounds we e compa ed using ei he a 2×2 con ingency chi-squa e es o a one- ailed Fishe exac es , as app op ia e ( assa s a s.ne ). These es s we e also used o compa e he ela ionship be ween ansc ip le el and NTS/TS deamina ion in he whole-genome analyses. Chi- squa e goodness o i was used o compa e obse ed o expec ed numbe s o e en s. Ra es we e signi ican ly di e en ly i he a e in each indi idual s ain did no o e lap he co esponding a e-associa ed CI o he o he s ain. 3. RESULTS Rela i e le els o spon aneous DNA damage associa ed wi h ansc ip ion o a a ge gene can be in e ed by examining he mu agenic consequences. An inhe en complica ion wi h mos base damages is ha hey o hei epai in e media es slow o s all subsequen DNA eplica ion and ansc ip ion. This can igge specialized bypass mechanisms du ing eplica ion and/o empla e-speci ic engagemen o nucleo ide excision epai du ing ansc ip ion, which complica es subsequen analyses and in e ences. The hyd oly ic deamina ion o cy osine o u acil is an excep ion, howe e , because he base-pai ing p ope ies o u acil a e iden ical o hose o hymine [41]. In wild- ype (WT) yeas cells, u acil is e icien ly emo ed by he Ung1 u acil-DNA N-glycosylase and is only a ely mu agenic. In he absence o Ung1 deamina ed cy osine pe sis s as pa o a U:G misma ch and he u acil is eplica ed as a hymine, leading o CG>TA ansi ion mu a ions [42]. 3.1. T ansc ip ion inc eases cy osine deamina ion in CAN1 As a epo e o spon aneous cy osine deamina ion we used he yeas CAN1 gene, which encodes an a ginine pe mease ha anspo s he oxic analog cana anine and con e s cana anine sensi i i y. Any o wa d mu a ion ha inac i a es he encoded p o ein esul s in cana anine esis ance (Can-R) on d ug-con aining medium. To examine e ec s o Williams e al. Page 6 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip ansc ip ion on spon aneous deamina ion, s ains we e used in which he CAN1 gene was ei he unde con ol o i s na i e p omo e o used o he highly ac i e GAL1 p omo e (low- and high- ansc ip ion condi ions, espec i ely; [5]). Unde low- ansc ip ion (low- xn) condi ions, 31% (69/220) o Can-R mu an s isola ed in an UNG1 backg ound con ained a CG>TA base subs i u ion (Tables S2-S3). We no e ha his is highe han he CG>TA p opo ion among Can-R mu an s epo ed p e iously (60/277; p=0.018; [43]). This di e ence could e lec ei he s ain backg ound di e ences o he use o en imes he s anda d amoun o cana anine in he p e ious s udy. Dele ion o UNG1 was accompanied by a 1.6- old inc ease in he Can-R a e (Table S4) as well as a p opo ional inc ease in CG>TA mu a ions om 31% o 88% (217/246; p<0.0001). This ansla es in o a 4.6- old inc ease in he CG>TA a e in he absence o Ung1 (Figu e 1) and con i ms ha mos CG>TA mu a ions a ising in an ung1 Δ backg ound e lec cy osine deamina ion. Highly ele a ed ansc ip ion esul ed in an ~12- old inc ease in Can-R a e he WT backg ound and dele ion o UNG1 was associa ed wi h an addi ional 1.9- old a e inc ease. As unde low- xn condi ions, he e was a p opo ional inc ease in CG>TA mu a ions in he ung1 Δ backg ound unde high- xn condi ions: om 6% (41/663) o 45% (249/554; p<0.0001). The a e o CG>TA mu a ions was 7- old highe unde high- xn han unde low- xn condi ions in he ung1 Δ backg ound, demons a ing ha ansc ip ion inc eases deamina ion o he co esponding DNA (Figu e 1). The lowe p opo ion o CG>TA mu a ions unde high- han low- xn condi ions (45% and 88%, espec i ely; p<0.0001) e lec s he la ge con ibu ion o Top1-dependen 2–5 bp dele ions o he spec um unde high high- xn condi ions [5, 6]. 3.2. T ansc ip ion ele a es nonsense mu a ions mo e han missense mu a ions An inhe en p oblem wi h o wa d-mu a ion assays is ha missense mu a ions, which al e a single amino acid, a e o en o no unc ional consequence. Changes ha con e a sense o a nonsense/s op codon unca e he co esponding p o ein p oduc , howe e , and almos always disable unc ion. A he CAN1 locus i was p e iously es ima ed ha only ~5% o missense mu a ions con e cana anine esis ance [43]. The missense-de ec ion issue po en ially can be exace ba ed unde high- xn condi ions whe e he e may be enough o a highly dys unc ional p o ein o con e a cana anine-sensi i e pheno ype. Among Can-R mu an s analyzed, we indeed obse ed p opo ionally ewe missense mu a ions among CG>TA changes de ec ed unde high- han unde low- xn condi ions (83/249 and 114/217, espec i ely; p<0.0001; Tables S2 and S3). Because o he inabili y o eliably de ec missense mu a ions unde high- xn condi ions, only CG>TA mu a ions ha c ea ed a TAA, TAG o TGA s op codon we e conside ed in subsequen analyses. 3.3. T ansc ip ion-associa ed deamina ion is biased o he NTS By con en ion, mu a ions a e epo ed in ela ion o he coding/sense s and o a gene, which has he same sequence as he mRNA and co esponds o he non- ansc ibed s and (NTS). Al hough epo ing mu a ions in his way allows a s and bias o be easily disce ned, i o en p o ides no in o ma ion as o he sou ce o he bias. This can be in e ed, howe e , using DNA polyme ases wi h speci ic misinco po a ion pa e ns [44], disabling a epai pa hway ha is highly speci ic o damage o only one o he complemen a y bases [45] o c ea ing a Williams e al. Page 7 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip single-s anded DNA a ge [46]. Thus, deamina ion o cy osine on he NTS leads o a C>T change while deamina ion o a cy osine loca ed on he ansc ibed s and (TS) esul s in a G>A mu a ional ead-ou (Figu e 2A). C>T mu a ions in sense codons gene a e all h ee s op codons: CAA>TAA, CAG>TAG and CGA>TGA. By con as , G>A mu a ions c ea e only TGA and TAG s op codons, bo h o which a e de i ed om he TGG sense codon (TGG>TGA and TGG>TAG); a TAA s op codon a ises only ia a G>A change in a p eexis ing TAG o TGA s op codon. Wi hin he CAN1 ORF, he e a e 12 CAA/CAG/CGA codons plus 16 TGG codons (Figu e S1) and mos we e equen si es o C>T o G>A mu a ions, espec i ely. Excep ions we e he ou TGG codons nea es he 3′ end o he gene, mu a ion o which p esumably gene a es a unca ed bu unc ional p o ein, and a TGG loca ed 5′ o o he posi ions whe e nonsense mu a ions we e equen . Mu a ions a hese i e codons we e also absen om he spec a o Lang and Mu ay [43]. Excluding hese i e TGG codons, he e we e 12 and 22 posi ions (22 guanines wi hin he emaining 11 TGG codon a ge s) in CAN1 whe e C>T and G>A mu a ions, espec i ely, we e moni o ed. E en hough p opo ionally ewe C>T han G>A mu a ions ha c ea ed s op codons we e expec ed (12/34 and 22/34, espec i ely), simila numbe s o C>T and G>A mu a ions we e obse ed in he low- xn ung1 Δ backg ound: 55 and 48, espec i ely (Table S4; p=0.02). On a pe si e basis, he e we e 4.6 C>T changes a cy osines whe e deamina ion was moni o ed (55 mu a ions a 12 si es) and 2.2 G>A mu a ions a moni o ed guanines (48/22). The e is hus an app oxima ely 2- old C>T bias e en unde low- xn condi ions. The co esponding a es o C>T and G>A e en s, howe e , we e e y simila and a e shown in Figu e 2B. In con as o simila numbe s o C>T and G>A mu a ions unde low- xn condi ions, he e was a 3.4- old excess o C>T ela i e o G>A mu a ions in he high- xn ung1 Δ backg ound: 128 and 38, espec i ely (p<0.0001). On a pe si e basis, he e we e 10.7 C>T changes a cy osines whe e deamina ion was moni o ed (128 mu a ions a 12 si es) and 1.7 G>A mu a ions a moni o ed guanines (38/22). In e ms o a es, C>T mu a ions we e ele a ed 14.5- old unde high- xn condi ions while G>A mu a ions we e ele a ed only 4.8- old (Figu e 2B). 3.4. The ansc ip ion-associa ed C>T bias e lec s nei he enzyma ic deamina ion no biased U:G epai Cy osine in single-s anded DNA spon aneously deamina es mo e apidly han cy osine in duplex DNA [47], sugges ing ha he s and-speci ic deamina ion bias obse ed unde high- xn condi ions e lec s he enhanced single-s anded cha ac e o he NTS. An al e na i e possibili y, howe e , is s and-biased deamina ion o cy osine by he yeas Fcy1 p o ein, which has been epo ed o deamina e DNA as well as RNA [48]. I Fcy1 we e he p ima y sou ce o cy osine deamina ion, hen he a e o C>T mu a ions should be educed in i s absence. The e was no change in he a es (Figu e 2B; Table S4) o dis ibu ion (p=0.12; Tables S2-S3) o C>T and G>A mu a ions in he ung1 Δ cy1 Δ double mu an ela i e o he ung1 Δ single mu an , consis en wi h s and-biased hyd oly ic deamina ion o cy osine. The misma ch epai (MMR) sys em emo es e o s ha a ise du ing DNA eplica ion by ini ia ing excision and e-syn hesis o he newly syn hesized s and ( e iewed in Williams e al. Page 8 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip [56]. Wang D, Landick R, Nuclease clea age o he ups eam hal o he non empla e s and DNA in an Esche ichia coli ansc ip ion elonga ion complex causes ups eam ansloca ion and ansc ip ional a es . J. Biol. Chem 272 (1997) 5989–5994. [PubMed: 9038220] [57]. Ba nes CO, Cale o M, Malik I, G aham BW, Spah H, Lin G, Cohen AE, B own IS, Zhang Q, Pulla a F, T akselis MA, Kaplan CD, Cale o G, C ys al s uc u e o a ansc ibing RNA polyme ase II complex e eals a comple e ansc ip ion bubble. Mol. Cell 59 (2015) 258–269. [PubMed: 26186291] [58]. Chu chman LS, Weissman JS, Nascen ansc ip sequencing isualizes ansc ip ion a nucleo ide esolu ion. Na u e 469 (2011) 368–373. [PubMed: 21248844] Williams e al. Page 15 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Figu e 1. Loss o Ung1 p e e en ially ele a es CG>TA mu a ions in CAN1 unde low- and high- xn condi ions. Whi e ba s co espond o he o al mu a ion a e and g ay ba s o he a e o CG>TA mu a ions. E o ba s a e he 95% con idence in e al (CI). Williams e al. Page 16 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Figu e 2. S and-speci ic mu a ion accumula ion. (A) Cy osine deamina ion on complemen a y DNA s ands yields C>T and G>A mu a ions. In an ung1 Δ backg ound, a u acil ha esul s om cy osine deamina ion pe sis s and empla es dAMP inco po a ion du ing subsequen DNA syn hesis. A e wo ounds o eplica ion, a deamina ed cy osine on he NTS (blue) o TS (yellow) gi es ise o a C>T o G>A mu a ion, espec i ely, when he NTS sequence is epo ed. Do ed lines ep esen he s a ing DNA s ands and a owheads co espond o 3′ ends. (B) In he absence o Ung1, high le els o ansc ip ion ele a e C>T mo e ha G>A mu a ions (blue and yellow ba s, espec i ely) demons a ing a bias o NTS deamina ion. This bias e lec s nei he Fcy1 no MMR ac i i y. E o ba s a e 95% CIs. Williams e al. Page 17 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Figu e 3. Biased deamina ion o he NTS unde high- xn condi ions is no exace ba ed in gene ic backg ounds whe e R-loops a e expec ed o (A) inc ease o (B) dec ease. Blue and yellow ba s e lec he a es o C>T and G>A mu a ions, espec i ely, in he indica ed gene ic backg ounds. E o ba s a e he 95% CIs. Williams e al. Page 18 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Figu e 4. Rela i e le els o RNA-DNA hyb ids a CAN1 and PDC1 . Samples we e o we e no (+ o −, espec i ely) ea ed wi h RNase H1, which deg ades RNA-DNA bu no RNA-RNA duplexes, p io o immunop ecipi a ion wi h he S9.6 an ibody. Values plo ed a e a bi a y uni s (A.U.) ha ep esen he mean o h ee biological eplica es a e immunop ecipi a ion (IP) wi h S9.6 ela i e o he IP inpu . E o ba s a e s anda d de ia ions. The sp 16–11 de i a i e o W303, which accumula es R-loops, was used as posi i e con ol o RNA-DNA de ec ion and was analyzed in pa allel wi h expe imen al s ains; duplica e samples we e used o WT bu only a single sample was analyzed om he sp 16–11 backg ound. Williams e al. Page 19 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Figu e 5. Vulne abili y o he NTS o DNA damage du ing ansc ip ion. A small po ion o he NTS is ende ed single-s anded as pa o each ansc ip ion bubble; e- o med duplex DNA and he nascen ansc ip lea e RNAP h ough sepa a e channels. I he ansc ip h eads back and pai s wi h he TS o o m an R-loop, a la ge po ion o he NTS is exposed. In he case o CAN1 , a con ibu ion o R-loops o he biased deamina ion o he NTS was de ec ed nei he gene ically o physically. Damage o he NTS ha occu s wi hin he ansc ip ion bubble is indica ed by a s a . Williams e al. Page 20 DNA Repai (Ams ) . Au ho manusc ip ; a ailable in PMC 2023 Sep embe 11. Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip Au ho Manusc ip