The replication checkpoint protects fork stability by releasing transcribed genes from nuclear pores

The Replica ion Checkpoin P o ec s Fo k S abili y by Releasing T ansc ibed Genes om Nuclea Po es Rod igo Be mejo, 1,8, *Thelma Cap a, 1,8 Rachel Jossen, 1,8 A ianna Colosio, 1 Camilla F a ini, 1 Wal e Ca o enu o, 1 And ea Coci o, 1 Ylli Doksani, 1,2 Hannah Klein, 3 Bele ´nGo ´mez-Gonza ´lez, 4 And e ´s Aguile a, 4 Yuki Ka ou, 5 Ka suhiko Shi ahige, 6 and Ma co Foiani 1,7, * 1 Fondazione Is i u o FIRC di Oncologia Molecola e (IFOM) a IFOM-IEO Campus, Via Adamello 16, 20139 Milan, I aly 2 Rocke elle Uni e si y, New Yo k, NY 10065, USA 3 New Yo k Uni e si y School o Medicine, New Yo k, NY 10016, USA 4 Cen o Andaluz de Biologı ´a Molecula y Medicina Regene a i a CABIMER, Uni e sidad de Se illa, 41004 Se ille, Spain 5 Tokyo Ins i u e o Technology, 226-8503 Yokohama Ci y, Japan 6 Resea ch Cen e o Epigene ic Disease, Ins i u e o Molecula and Cellula Biosciences, Tokyo Uni e si y, Bunkyo-ku, 108-8639 Tokyo, Japan 7 DSBB-Uni e si a `degli S udi di Milano, 20139 Milan, I aly 8 These au ho s con ibu ed equally o his wo k *Co espondence: od igo.be m[email p o ec ed] (R.B.), [email p o ec ed] (M.F.) DOI 10.1016/j.cell.2011.06.033 SUMMARY T ansc ip ion hinde s eplica ion o k p og ession and s abili y, and he Mec1/ATR checkpoin p o ec s o k in eg i y. Examining checkpoin -dependen mechanisms con olling o k s abili y, we ind ha o k e e sal and do man o igin i ing due o check- poin de ec s a e escued in checkpoin mu an s lacking THO, TREX-2, o inne -baske nucleopo ins. Gene ga ing e he s ansc ibed genes o he nuclea pe iphe y and is coun e ac ed by checkpoin kinases h ough phospho yla ion o nucleopo ins such as Mlp1. Checkpoin mu an s ail o de ach ansc ibed genes om nuclea po es, hus gene a ing opolog- ical impedimen s o incoming o ks. Releasing his opological complexi y by in oducing a double- s and b eak be ween a o k and a ansc ibed uni p e en s o k collapse. Mlp1 mu an s mimicking cons i u i e checkpoin -dependen phospho yla ion also alle ia e checkpoin de ec s. We p opose ha he checkpoin assis s o k p og ession and s abili y a ansc ibed genes by phospho yla ing key nucleo- po ins and coun e ac ing gene ga ing, hus neu al- izing he opological ension gene a ed a nuclea po e ga ed genes. INTRODUCTION ATR/Mec1 and ATM/Tel1 s abilize s alled (Lopes e al., 2001) and e minal eplica ion o ks (Doksani e al., 2009), p e en ing hei collapse (B anzei and Foiani, 2010) and accumula ion o ecombinogenic X-shaped e e sed o ks (Doksani e al., 2009; Sogo e al., 2002). Re e sed o ks a e p ocessed by he Exo1 nuclease (Co a-Ramusino e al., 2005; Lopes e al., 2001) in o gapped s uc u es (Sogo e al., 2002). Checkpoin mu an s expe iencing eplica ion s ess ail o esume o k p og ession (B anzei and Foiani, 2010; Segu ado and Di ley, 2008) and i e addi ional eplicons (San ocanale and Di ley, 1998; Shi ahige e al., 1998). Hyd oxyu ea (HU) causes dNTP deple ion, inhibi ing eplica- i e polyme ases. Consequen ly, eplica ion o ks s all and accu- mula e sho RPA-ssDNA s e ches (Sogo e al., 2002; Zou and Elledge, 2003) ha igge Mec1 ac i a ion. Mec1 hen ac i a es he Rad53 kinase, which p o ec s s alled o k s abili y h ough mechanisms ha a e s ill obscu e. DNA opology in luences o k e e sion. In i o, e e sed o ks accumula e a posi i ely supe coiled plasmids upon eplisome dissocia ion (Pos ow e al., 2001b). The he modynamic ene gy accumula ed as o sional s ess in on o eplica ion o ks causes o k e e sal cha ac e ized by pa en al s ands eanneal- ing and nascen s ands ex usion and pai ing (Pos ow e al., 2001a; Sch a zman and S asiak, 2004). Ou cu en knowledge o he in i o a chi ec u al o ganiza ion o eplica ing ch omo- somes is limi ed. I has been p edic ed ha hose si es ancho ing ch omosomes o memb anes beha e as opological ba ie s, p e en ing he o a ion o he DNA helix s ands a ound each o he and hus impeding he di usion o opological changes (Pos ow e al., 2004; Wang, 2002). In p inciple, opological ba ie s migh a ise a si es whe e ansc ibed genes associa e o ixed nuclea en elope s uc u es. In euka yo es, messenge RNA (mRNA) syn hesis is coupled wi h mRNA p ocessing, mRNP (messenge ibonucleopa icle) assembly, and expo o he cy oplasm (Aguile a, 2005; Ko ¨hle and Hu , 2007). These p ocesses a e coo dina ed by p o ein complexes e he ing an- sc ibed ch oma in o he nuclea po e complex (NPC). THO is Cell 146, 233–246, July 22, 2011 ª2011 Else ie Inc. 233 Open access unde CC BY-NC-ND license. a complex, composed o Tho2, Hp 1, M 1 and Thp2, ha a els wi h he RNA polyme ase II (Luna e al., 2008). THO also associ- a es wi h Y a1 and Sub2 ( o o m he TREX complex) and ac o s in ol ed in mRNA expo , including he TREX-2 (o THSC) com- plex. TREX-2 is cons i u ed by Sac3, Thp1, Sus1, and Cdc31 p o eins (Ko ¨hle and Hu , 2007). Sac3 in e ac s wi h Cdc31 and Sus1 (Jani e al., 2009) and media es he associa ion o he complex o he nuclea en elope by binding o he Nup1 nu- cleopo in a he NPC inne baske . Mu an s in THO and TREX-2 genes sha e ansc ip ional elonga ion de ec s, ansc ip ion- associa ed hype ecombina ion (TAR), and mRNA expo de ec s (Luna e al., 2008). THO and TREX-2 a e also equi ed o he associa ion o ansc ibed genes o he NPC (Cabal e al., 2006; D ubin e al., 2006; Rougemaille e al., 2008). This p ocess, known as gene ga ing, assis s apid gene exp ession coupling mRNA ansc ip ion and expo o he cy oplasm and has been implica ed in a myosin-like Mlp1-dependen ansc ip- ional memo y mechanism (Tan-Wong e al., 2009). The THO-TREX-2-media ed physical associa ion o he an- sc ibed DNA wi h he NPC migh hinde DNA s and o a ion es ablishing opological ba ie s. T ansc ibed genes ep esen he mos abundan si es o eplica ion o k pausing in he yeas genome, and eplisome pausing a ansc ibed genes is inde- penden o he pola i y be ween eplica ion and ansc ip ion (Az olinsky e al., 2009). The mechanisms causing eplica ion/ ansc ip ion in e e ence may no simply e lec he clash be ween o ks and ansc ip ional machine ies, as he R m3 heli- case, which acili a es o k p og ession by emo ing p o ein obs acles, does no ac a highly ansc ibed genes (Az olinsky e al., 2009). Mo eo e , DNA opoisome ase II (Top2) seems o modula e DNA opology a si es o eplica ion/ ansc ip ion in e - e ence (Be mejo e al., 2009). We ound ha mu a ions in THO and TREX-2 genes coun- e ac HU sensi i i y and o k e e sal in checkpoin mu an s, hus sus aining eplica ion o k p og ession. Rad53 eleases ansc ibed genes om he nuclea en elope in esponse o eplica ion s ess. Se e al nucleopo ins, including Mlp1, a e phospho yla ed by checkpoin kinases (Smolka e al., 2005, 2007). mlp1 mu an s mimicking cons i u i e phospho yla ion by Rad53 escue ad53 HU sensi i i y. Dis up ing he e he ing o ansc ibed genes o he NPC o in oducing a double-s and b eak (DSB) be ween a o k and a highly ansc ibed gene alle i- a es he HU sensi i i y and coun e ac s o k e e sal in check- poin -de ec i e cells. We p opose ha he Mec1/ATR and Rad53-dependen eplica ion checkpoin p omo es eplica ion o k s abili y by con olling gene ga ing, hus coun e ac ing opological s ess-d i en o k e e sal. RESULTS Iden i ica ion o ad53 Supp esso s ad53-K227A mu an s (Zheng e al., 1993), in 0.2 M HU, accumu- la e e e sed o ks and hemi eplica ed in e media es ha can be isualized by elec on mic oscopy and wo-dimensional (2D) gels (Lopes e al., 2001; Sogo e al., 2002). Re e sed o ks mig a e on 2D gels as a cone signal cha ac e ized by c uci o m in e media es spanning om ully duplica ed molecules o s uc- u es o lowe mass, esul ing om Exo1-dependen esec ion (Co a-Ramusino e al., 2005; Lopes e al., 2001). A lowe HU concen a ions, Exo1 does no signi ican ly con ibu e o e e sed o k p ocessing and ad53-K227A cells accumula e un- esec ed X spikes (da a no shown). Two nonmu ually exclusi e hypo heses p edic ha o k e e sal esul s om opological ansi ions such as accumula- ion o posi i e supe coiling (Pos ow e al., 2001b) and/o uno o hemica enane s uc u es likely ep esen ing p eca enane de i a i es (Be mejo e al., 2008; Co a-Ramusino e al., 2005). We designed gene ic sc eens aimed a iden i ying ac o s in luencing ad53-K227A iabili y a low HU doses. We used he yeas dele ion lib a y (Tong e al., 2001), sea ching o hose gene dele ions escuing he HU sensi i i y caused by o e ex- p ession o he dominan -nega i e ad53-D339A mu a ion (Fay e al., 1997). The supp esso s, alida ed in he W303 gene ic backg ound (Thomas and Ro hs ein, 1989), we e also able o escue he ad53-K227A mu a ion. Dele ions o SAC3,THO2 HPR1,THP2,MFT1,SUS1, and THP1 (Figu e S1 a ailable online) supp essed ad53 and mec1 HU sensi i i y (Figu es 1A– 1C and da a no shown). THO and TREX2 mu a ions pe se do no coun e ac he inhibi o y e ec s caused by HU because he single-dele ion mu an s a e sensi i e o high HU concen a- ions and a e syn he ic sick in combina ion wi h ad53 (da a no shown; Go ´mez-Gonza ´lez e al., 2009). T ansc ip ome anal- yses showed ha SAC3 abla ion did no signi ican ly a ec he exp ession o he Rad53-dependen damage-inducible genes in wild- ype o ad53 cells (da a no shown). TREX-2 and THO mu a ions do no accele a e S phase p og ession a low HU concen a ions, as would be he case o mu an s exhibi ing ele a ed dNTP le els (da a no shown). Al hough THO/TREX-2 mu an s exhibi mild ansc ip ional de ec s (Go ´mez-Gonza ´lez e al., 2011), we canno ule ou ha he ad53 supp ession mechanism is, in pa , due o educed ansc ip ional obs uc- ions o eplica ion. TREX-2 and THO Mu a ions Rescue ad53 Fo k De ec s TREX-2 and THO mu a ions cause ecombinogenic e en s asc ibed o R loops (Gonza ´lez-Aguile a e al., 2008; Hue as and Aguile a, 2003). In i o and in i o obse a ions indica e ha nascen mRNAs can p ime DNA syn hesis (Pome an z and O’Donnell, 2008, 2010). We in es iga ed whe he he abili y o TREX-2-THO mu a ions o escue ad53 de ec s was due o ep iming o DNA syn hesis downs eam o collapsed o ks by engaging apped mRNAs a R loops. This would imply ha he mechanism o supp ession would be in luenced by RNase H o e exp ession, as in he case o R loop accumula ion and an- sc ip ional and ecombina ion abno mali ies in THO mu an s (Hue as and Aguile a, 2003). RNH1 (RNase H1) o e exp ession o abla ion o endogenous RNase H1 (RNH1) o RNase H2 subuni s (RNH201,RNH202) did no in luence he HU sensi i i y o ad53 o sac3 ad53 cells (Figu es 2A and 2B). Hence, he supp ession o TREX-2-THO mu a ions does no depend on he abe an accumula ion o R loops. We hen analyzed he e ec o SAC3 abla ion on ad53 eplicon dynamics (Figu e 3A) by immunop ecipi a ion o B dU- subs i u ed DNA ollowed by high-densi y oligo a ay hyb idiza- ion (B dU-IP-Chip) (Fachine i e al., 2010). A low doses o HU, wild- ype o ks emana ing om ea ly eplica ion o igins 234 Cell 146, 233–246, July 22, 2011 ª2011 Else ie Inc. p og essed slowly (Lopes e al., 2001), gene a ing B dU acks spanning se e al kilobases ou wa d o he o igin sequences (Fig- u e 3A). In ad53 cells, sho e B dU acks a ound ea ly o igins we e obse ed due o o k collapse and consequen ailu e o inco po a e B dU (Feng e al., 2006; Lopes e al., 2001; Sogo e al., 2002). Concomi an ly, addi ional B dU acks appea ed a la e/do man o igins (San ocanale and Di ley, 1998; Shi ahige e al., 1998). Though SAC3 abla ion pe se did no cause A B GAL ad53-D339A sac3 GAL ad53-D339A ho2 YP Galac ose 10 mM HU YP Glucose 10 mM HU C Figu e 1. TREX-2 and THO Abla ions Rescue ad53 HU Sensi i i y (A) GAL- ad53-D339A SGA YP galac ose o YP glucose pla es con aining 10 mM HU. Whi e boxes ma k supp esso s o ad53-D339A HU sensi i i y. (B) WT, sac3D, ad53-K227A, ad53-K227A sac3D, hp1D, ad53-K227A hp1D,sus1D, and ad53-K227A sus1Dcells pla ed wi hou (YPDA) o wi h 5 mM HU. (C) Summa y o THO/TREX, TREX-2, and NPC genes and hei gene ic in e ac ions wi h ad53-K227A. Human o hologs a e indica ed. Null, gene dele ion pheno ype; ad53 escue, supp ession o ad53-K227A le hali y a 5–10 mM HU; ss wi h ad53, syn he ic sickness in combina ion wi h ad53-K227A;HU sensi i i y, g ow h de ec s a 100–200 mM HU; phosphosi es, he esidues a ge ed by he checkpoin kinases (in pa en heses) a e shown. Fo schema ic summa y, see Figu e S1. Cell 146, 233–246, July 22, 2011 ª2011 Else ie Inc. 235 signi ican changes in eplicon dynamics, in ad53, i escued bo h he sho B dU acks a ea ly o igins and he B dU inco po- a ion a la e/do man o igins. Because SAC3 dele ion es o es eplicon p og ession a ea ly o igins wi hou exhibi ing addi ional o igin-un ela ed B dU peaks, we conclude ha Sac3 abla ion somewha s abilizes genome-wide ad53 o ks. This u he a gues agains eini ia ion e en s media ed by R loops. The inding ha , in ad53 sac3 cells, s abiliza ion o ea ly eplicons co ela es wi h he lack o i ing o la e/do man eplicons sugges s ha he wo phenomena a e mechanis ically linked, a leas in his gene ic backg ound. Because Sac3 plays a global and de imen al ole on he p og ession o HU-challenged o ks Figu e 2. RNase H O e exp ession Does No In luence ad53 o ad53 sac3 Cells’ HU Sensi i i y (A) WT, sac3D, ad53-K227A, and ad53-K227A sac3Dcells ca ying ei he he ec o (pYES) o pGW-RNH1, exp essing RNH1 unde he GAL1 p omo e (GAL RNH1), we e pla ed in glucose (GLU)- o galac ose (GAL)-con aining media wi hou () o wi h 5 mM HU. (B) WT, nh1D, ad53-K227A, ad53-K227A nh1D, nh201D, ad53-K227A nh201D, nh202D, and ad53-K227A nh202Dcells we e pla ed wi hou (YPDA) o wi h 2.5 o 5 mM HU. in ad53 cells, we es ed whe he Sac3 a ec ed he accumula ion o abe an eplica ion in e media es in ad53 mu an s. Replica ion in e media es om wild- ype, sac3, ad53, and ad53 sac3 s ains ea ed wi h 25 mM HU we e isu- alized by 2D gels (Figu e 3B) a he ea ly o igin ARS305 (Newlon e al., 1993; Po- loumienko e al., 2001). Wild- ype and sac3 cells exhibi ed compa able 2D p o iles and accumula ed bubbles, which esul om o igin i ing, and la ge Ys, which a ise due o asymme ic o k p o- g ession ou side o he es ic ion ag- men analyzed (B ewe and Fangman, 1987). A 60–90 min, ad53 cells accumu- la ed X-shaped in e media es co e- sponding o unp ocessed e e sed o ks (Co a-Ramusino e al., 2005; Lopes e al., 2001). Re e sed o ks de ec ion was educed in sac3 ad53 double mu an s (Figu e 3B). Hence, SAC3 abla ion escues eplicon dynamics and o k e e sal in ad53 mu an s, sugges ing ha he TREX-2 complex, and by ex ension THO, may ac in coo dina ion wi h eplica ion o ks o e en be an in eg al pa o he mo ing eplisome. Howe e , he Hp 1 ChIP-chip genomic clus e s did no colocalize wi h DNA polyme ases a ac i e eplica ion o igins in HU- ea ed cells bu o e lapped h oughou he cell cycle wi h RNA Pol II clus e s (Be mejo e al., 2009; Go ´mez-Gon- za ´lez e al., 2011). This is consis en wi h a global ole o Hp 1 in ansc ip ion elonga ion and mRNP biogenesis. We no e ha he Hp 1-binding si es also co ela e wi h Top2 and Hmo1 clus e s ha ep esen nonpola pausing elemen s o incoming o ks (Az olinsky e al., 2009; Be mejo e al., 2009). Rad53 In luences Gene Ga ing I is easonable o hink ha he TREX-2-THO-media ed physical con inui y be ween ansc ibed ch oma in and he nuclea en e- lope would es ablish nonpola ba ie s p e en ing he di usion 236 Cell 146, 233–246, July 22, 2011 ª2011 Else ie Inc. A B Figu e 3. SAC3 Dele ion Supp esses ad53 Cells’ Fo k De ec s (A) B dU-IP-Chip analysis o eplicon dynamics o WT, sac3D, ad53-K227A, and ad53-K227A sac3Dcells 60 min a e elease om G1 in o S phase in 25 mM HU. O ange (B dU-IP) his og am ba s in he y axis show he a e age signal a io in log2 scale o loci along he epo ed egion on ch omosome III. The x axis shows ch omosomal coo dina es. Posi ions o ea ly and la e/do man ARS elemen s a e in ed and g ay, espec i ely. Blue and ed ho izon al ba s ma k he B dU inco po a ion acks co esponding o o ks emana ed om ea ly eplica ion o igins in WT, sac3D, and ad53-K227A sac3Dcells o ad53-K227A mu an s, espec i ely. Red iangles ma k addi ional B dU acks gene a ed by unscheduled do man o igin i ing in ad53-K227A cells. (B) 2D gel analysis o eplica ion in e media es in WT (SAC3 RAD53), sac3D(sac3 RAD53), ad53-K227A (SAC3 ad53), and ad53-K227A sac3D(sac3 ad53) cells a he indica ed imes a e elease om G1 in o S phase in 25 mM HU. A schema ic ep esen a ion o he 2D gel p o iles obse ed in RAD53 and ad53 cells is shown. His og am plo s o he a io be ween quan i ied ‘‘Spike’’ and ‘‘La ge Y’’ in e media es signals a e shown. Cell 146, 233–246, July 22, 2011 ª2011 Else ie Inc. 237 o opological s ess, as he associa ion o DNA o a ixed nuclea s uc u e would limi he o a ion o helix s ands a ound each o he (Pos ow e al., 2001a, 2004). The consequen accumula- ion o posi i e supe coiling a ansc ibed genes migh he e o e p o ide he ene gy sou ce o p omo e e e sal o incoming o ks, pa icula ly in a checkpoin -de ec i e gene ic con ex ha exhibi s an al e ed eplisome- o k associa ion (Cobb e al., 2003; Lucca e al., 2004; Pos ow e al., 2001b). This hypo hesis would ha e he ollowing expec a ions: (1) pa o he o k ins a- bili y p oblems in ad53 mu an s should depend on ansc ip ion; (2) he Mec1-Rad53 checkpoin migh con ol eplicon in eg i y by eleasing he opological ba ie s imposed by he coupling be ween mRNA syn hesis and nuclea expo ; and (3) gene a ing a physical discon inui y ei he a he le el o he nuclea po e- bound TREX-2/THO complexes o in he DNA s ands a he bo de be ween an incoming o k and a ansc ibed gene should elie e he opological s ess causing o k e e sal. We es ed he con ibu ion o ansc ip ion on o k s abili y in ad53 mu an s. We ans o med wild- ype, sac3, ad53, and ad53 sac3 cells wi h a plasmid ca ying he URA3-selec able ma ke and a galac ose-inducible LACZ gene ha is ansc ibed head on wi h he le o k a ising om ARS209 (Figu e 4A). Wild- ype and sac3 cells we e able o s ably e ain he plasmid and g ow in he absence o u acil wi h o wi hou galac ose. ad53 AB CD Figu e 4. The Replica ion Checkpoin Nega i ely Regula es Gene Ga ing (A) WT, sac3D, ad53-K227A, and ad53-K227A sac3Dcells ca ying pGAL-LACZ-IN (IN) we e g own in syn he ic comple e U a pla es con- aining glucose (SC –U a) o galac ose (SGal –U a). (B) Rep esen a i e images o Nup49-GFP, TETR- GFP, TETO::GAL1/10/7, Nop1-Che y glucose- g own cells and pe cen age o cells showing cen al o NPC- e he ed (pe iphe al) GAL loci in WT o mlp1Dcells g own in he p esence o glucose (GLU) o galac ose (GAL). (C) WT and ad53-K227A cells we e g own o e - nigh in he p esence o galac ose (GAL) and ea ed wi h 0.2 M HU o he indica ed imes. The mean pe cen ages o cells showing pe iphe al GAL loci and s anda d de ia ions (his og am e o ba s) om h ee independen expe imen s a e shown. (D) Se ial dilu ions o WT, sac3DCID, ad53- K227A, and ad53-K227A sac3DCID cells pla- ed in he absence (YPDA) o p esence o 5 mM HU. cells in galac ose we e unable o o m colonies, sugges ing ha hey ailed o e ain he URA3 plasmid ollowing galac- ose-induced ansc ip ion o he LACZ gene. We obse ed he same phenom- enon using a plasmid ca ying he LACZ gene ansc ibed codi ec ionally wi h o ks a ising om ARS209 (da a no shown). We no e ha he GAL1 p omo e is leaky in he p esence o glucose, hus explaining he g ow h de ec s o glucose- ea ed ad53 cells. SAC3 abla ion was able o escue he g ow h de ec s in glucose and he cell le hali y in galac ose owing o he ad53 mu a ion. We conclude ha he TREX-2 complex is de imen al o ad53 iabili y when o ks a e des abilized by ansc ip ion, aising he possibili y ha , also in ad53 mu an s ea ed wi h low doses o HU (and he e- o e wi h an al eady comp omised o k p ocessi i y), TREX-2 igge s o k abno mali ies by e he ing ansc ip ion wi h nuclea po e-media ed mRNA expo . We hen es ed whe he Rad53 in luences he nuclea en e- lope associa ion o he ansc ibed GAL genes locus agged wi h he TET ope a o /TET -GFP sys em (Be ge e al., 2008). We analyzed he pe cen age o cells exhibi ing pe iphe al dis i- bu ion o he GFP loci in loga i hmically g owing cells in he p esence o glucose ( ep essing condi ions o GAL genes) o galac ose (Figu e 4B). In glucose, he GAL1/GAL10/GAL7 gene locus was localized o he nuclea pe iphe y in 39% o he cells, whe eas in galac ose, in 67%. Howe e , gene ga ing-de ec i e mlp1 mu an s exhibi ed 37% and 44% pe iphe al oci in glucose and galac ose, espec i ely, acco ding o p e ious esul s (B ickne and Wal e , 2004; Cabal e al., 2006; Tan-Wong e al., 2009). We hen analyzed he localiza ion o he GAL clus e in wild- ype and ad53 cells in he p esence o HU and galac ose 238 Cell 146, 233–246, July 22, 2011 ª2011 Else ie Inc. (Figu e 4C). Whe eas in wild- ype cells, he pe cen age o pe iph- e al GFP loci dec eased, in ad53, i emained a high le els. Con e sely, Rad53 did no a ec GFP dis ibu ion in HU- ea ed cells in he p esence o glucose (da a no shown). We hen in e - up ed in ad53 cells he physical con inui y wi hin he TREX-2 complex by in oducing he sac3DCID mu a ion ha p ecludes he associa ion wi h Sus1 and Cdc31 (Figu e S1 a ailable online) (Jani e al., 2009) and compa ed HU sensi i i y and o k e e sal in ad53 and ad53 sac3DCID cells. The sac3DCID mu a ion escued ad53 HU sensi i i y and o k e e sal (Figu e 4D) (da a no shown). Se e al NPC ac o s in he inne baske and some THO subuni s a e phospho yla ed in a checkpoin -dependen manne in yeas and mammals, and in pa icula , Mpl1 is a Rad53 a ge (Figu e 1C) (Chen e al., 2010; Ma suoka e al., 2007; Smolka e al., 2007). Abla ion o MLP1 in ad53 cells p e en ed he pe iph- e al localiza ion o he GAL locus ollowing HU ea men in galac- ose, esembling mlp1 mu an s (Figu e 5A). In p inciple, Rad53 migh coun e ac gene ga ing in S phase by nega i ely egula ing NPC ac o s. The HU sensi i i y o ad53 cells migh be asc ibed o he inabili y o elease he associa ion be ween ansc ibed genes and he nuclea en elope and, in u n, o coun e ac he opological ba ie s a ising when o ks app oach nuclea en e- lope-associa ed ansc ibed genes. We hen abla ed hose NPC p o eins exhibi ing gene ga ing de ec s (Cabal e al., 2006; Tan-Wong e al., 2009) and es ed hei abili y o alle ia e ad53 HU sensi i i y. NUP1 o MLP1 dele ions ecapi ula ed he ad53 supp ession by TREX2 mu a ions (Figu e 5B). O he p o eins such as Nup2 and Nup60, which localize in he inne baske o he nuclea po e bu do no physically in e ac wi h TREX-2 and whose ole in gene ga ing is s ill con o e sial (B ickne e al., 2007; Cabal e al., 2006; Ligh e al., 2010), did no con ibu e o ad53 HU sensi i i y (Figu e S2). Al oge he , hese esul s sugges ha dis up ing he associa ion be ween he nuclea po e and TREX-2 p o eins migh alle ia e he inabili y o ad53 mu an s o de ach ansc ibed genes om he nuclea en elope. We hen mu a ed se ine 1710 in Mlp1 (Figu e 5C) o an aspa ic esidue o an alanine o mimic cons i u i e Mlp1 phospho yla ion o dephospho yla ion, espec i ely. mlp1-S1710D coun e ac ed nuclea pe iphe al localiza ion o ansc ibed GAL genes, hus esembling loss-o - unc ion mlp1 mu a ions (Figu e 5D). Con e sely, mlp1-S1710A cells beha ed like wild- ype. Mo e- o e , mlp1-S1710D, di e en ly om mlp1-S1710A, also escued ad53 HU sensi i i y (Figu e 5E). These obse a ions suppo he hypo hesis ha Rad53 inhibi s gene ga ing hough phospho yla- ion o Mlp1 and, possibly, o he NPC p o eins. Fo k Re e sal Is Coun e ac ed by Double-S and B eak Fo ma ion We hen explo ed he e ec o eleasing he opological ension wi hin a egion be ween a o k and a ansc ibed gene on o k e e sal. We induced a DSB by o e exp essing he HO endonu- clease (Lee e al., 2000) ha ecognizes a speci ic si e inse ed be ween ARS305 and he mos p oximal ansc ibed gene, PDI1, which causes o k pausing (Be mejo e al., 2009) and is bound by Top2, Hmo1, Hp 1, and Sac3 (Be mejo e al., 2009 and da a no shown). In heo y, he DSB-induced discon inui y in he helix s ands should pe mi o a ion o hei ee ends, hus leading o supe coiling elaxa ion and coun e ac ing o k e e sal. We induced DSB o ma ion and eleased he cells in he p esence o HU (Figu e 6). We analyzed he eplica ion in e me- dia es o ARS305 and ARS202, an o igin loca ed on a di e en ch omosome and no expe iencing DSB o ma ion. S ains bea ing a nonclea able HO sequence we e included as con ols (HO-inc). Wild- ype cells i ed bo h o igins, as indica ed by he p esence o la ge Y in e media es in ARS305 and bubbles in ARS202, ega dless o he p esence o absence o he DSB (Doksani e al., 2009). ARS202 i es la e han ARS305, and bubbles can be de ec ed a ARS305 a ea lie ime poin s (da a no shown). As expec ed, HU- ea ed HO-inc ad53 cells i ed bo h o igins and accumula ed X-shaped in e media es co e- sponding o e e sed eplica ion o ks. DSB induc ion educed e e sed o k accumula ion a ARS305, bu no a he ARS202 locus (Figu e 6). We no e ha 20% o he AR305-PDI1 egion emained uncu ollowing HO induc ion, hus accoun ing o he esidual Xs obse ed a ARS305 in ad53 cells. These obse - a ions sugges ha , in i o, posi i e supe coiling is a d i ing o ce o o k e e sal in checkpoin mu an s. Mo eo e , he ind- ings ha DSB o ma ion mimics he e ec o Sac3 abla ion in p e en ing o k e e sal wi hin he same locus sugges ha he in eg i y o he opological domain spanning he eplisome and he nuclea po e-associa ed ansc ibed gene (Casola i e al., 2005) in luences he a e o s alled o ks and ha Rad53, Top2, TREX-2, and THO complexes and NPC p o eins collabo a e in con olling he S phase a chi ec u e o ansc ibed loci o p e en abe an ansi ions a eplica ing ch omosomes. DISCUSSION T ansc ip ion is coo dina ed wi h eplica ion o main ain genome in eg i y. Because o k ad ance and he p og ession o he an- sc ip ion bubble gene a e posi i e supe coiling, a head-on collision be ween eplica ion and ansc ip ion causes opolog- ical impedimen s and o k pausing (Liu and Albe s, 1995; Ola a ie a e al., 2002; Wang, 2002). Fo k es a can occu h ough he displacemen o he RNA polyme ase complex (Pome an z and O’Donnell, 2010). In i o s udies showed ha codi ec ional collision be ween eplica ion and ansc ip ion has li le e ec s on o k p og ession (Liu and Albe s, 1995; Pom- e an z and O’Donnell, 2008) unless he RNA polyme ase s alls (Elı ´as-A nanz and Salas, 1997). The e is e idence sugges ing ha he eplisome can use mRNA as a p ime a e colliding codi ec ionally wi h RNA polyme ase (Pome an z and O’Donnell, 2008; Kogoma, 1997). Hence, i is no su p ising ha , in p oka y- o es, which ini ia e DNA syn hesis om a single o igin, codi ec- ional collision be ween o ks and ansc ip ion bubbles is he p e e ed op ion (B ewe , 1988; Rocha, 2004). Howe e , in mul i- o igin species, such as euka yo es, ansc ip ion and eplica ion collide bo h in head-on o codi ec ional ways. Indeed, an- sc ibed genes always pause eplica ion o ks, ega dless o hei ela i e o ien a ion (Az olinsky e al., 2009; Be mejo e al., 2009), pe haps due o he Top2-media ed a chi ec u e o ansc ibed genes (Be mejo e al., 2009). We show ha he TREX-2-THO-dependen coupling o an- sc ip ion, gene ga ing, and mRNA biogenesis causes abe an Cell 146, 233–246, July 22, 2011 ª2011 Else ie Inc. 239 A C D E B Figu e 5. Nucleopo in Mu a ions Abolishing Gene Ga ing Supp ess ad53 Pheno ypes (A) ad53-K227A and ad53-K227A mlp1Dcells we e g own o e nigh wi h galac ose (GAL) and ea ed wi h 0.2 M HU. The mean pe cen ages o cells showing pe iphe al GAL loci and s anda d de ia ions (his og am e o ba s) om h ee independen expe imen s a e shown. (B) Se ial dilu ions o WT, nup1D, ad53-K227A, ad53-K227A nup1D,mlp1D, and ad53-K227A mlp1Dcells pla ed in he absence (YPDA) o p esence o 5 mM HU. (C) E olu iona y compa ison o a po ion o he Mlp1 globula domain con aining esidues phospho yla ed by checkpoin kinases. Conse ed esidues a e labeled in g een. (D) Pe cen age o cells showing cen al o NPC- e he ed (pe iphe al) GAL loci in WT, mlp1S1710A,mlp1S1710D, and mlp1Dcells g own in he p esence o glucose (GLU) o galac ose (GAL). (E) Se ial dilu ions o WT, mlp1S1710D, ad53-K227A, and ad53-K227A mlp1S1710D cells pla ed in he absence (YPDA) o p esence o 5 mM HU. Fo ela ed da a, see Figu e S2. 240 Cell 146, 233–246, July 22, 2011 ª2011 Else ie Inc. ansi ions a s alled o ks in checkpoin -de ec i e cells. These obse a ions, oge he wi h he indings ha he genome-wide dis ibu ion o Hp 1 o e laps wi h ha o RNA polyme ase II and Top2 a ansc ibed genes (Be mejo e al., 2009), he check- poin phospho yla es nuclea po e p o eins (Chen e al., 2010; Smolka e al., 2007), Rad53 s oichiome ically associa es wi h he Kap95 ka yophe in (Smolka e al., 2005), and he ela i e dis ance be ween nuclea pe iphe y and ansc ibed genes depends on he checkpoin , sugges ha he physical con inui y be ween ansc ibed genes and he nuclea en elope is modu- la ed by he checkpoin o assis o k p og ession. The TREX-2 and THO-dependen opological impedimen s could, in pa , Figu e 6. DSB Induc ion Coun e ac s Fo k Re e sal in ad53 Mu an s Replica ion in e media es in WT HO-inc, WT HO, ad53-K227A HO-inc, and ad53-K227A HO cells ollowing a ac o -induced G1 a es , DSB o ma ion by galac ose addi ion, and elease in 25 mM HU. Schema ic ep esen a ions o he 2D gel diges ion s a egies used a e shown. His og ams indica e he a io be ween e e sed o k signal in ensi ies a ARS305 and ARS202 o ad53 cells a each ime poin . Fo ela ed da a, see Figu e S4. Cell 146, 233–246, July 22, 2011 ª2011 Else ie Inc. 241