A local ATR-dependent checkpoint pathway is activated by a site-specific replication fork block in human cells

When replication forks encounter DNA lesions that cause polymerase stalling a checkpoint pathway is activated. The ATR-dependent intra-S checkpoint pathway mediates detection and processing of sites of replication fork stalling to maintain genomic integrity. Several factors involved in the global checkpoint pathway have been identified, but the response to a single replication fork barrier (RFB) is poorly understood. We utilized the E.coli-based Tus-Ter system in human MCF7 cells and showed that the Tus protein binding to TerB sequences creates an efficient site-specific RFB. The single fork RFB was sufficient to activate a local, but not global, ATR-dependent checkpoint response that leads to phosphorylation and accumulation of DNA damage sensor protein γH2AX, confined locally to within a kilobase of the site of stalling. These data support a model of local management of fork stalling, which allows global replication at sites other than the RFB to continue to progress without delay.


Introduction 4
In the E.coli genome, the DNA pausing sequences called terminator (Ter) are 68 recognized by a protein called Tus to cause a polar site-specific arrest of the RF at the 69 end of the bacterial chromosome replication (Hiasa and Marians, 1994;Hidaka et al., 70 1988;Mulcair et al., 2006;Roecklein et al., 1991). The Tus-Ter complex leads to a 71 temporarily locked complex on DNA that can be overcome by the fork arriving in the 72 opposite direction, displacing Tus to terminate replication. The artificial E.coli-based 73 Tus/Ter system has previously been employed in mouse embryonic stem cells to 74 measure homologous recombination repair products and investigate the DNA repair 75 pathway choice (Chandramouly et al., 2013;Willis et al., 2017Willis et al., , 2014. 76 In this study, we integrated a plasmid with 5 repeats of the TerB sequence in the 77 non-permissive orientation at a unique site within chromosome 12 of the breast cancer 78 cell line MCF7 (MCF7 5C-TerB clone). We utilized the integrated Tus-TerB system in 79 human MCF7 cells to artificially generate individual RFBs and investigated the 80 activation of the S-phase checkpoint signaling mechanism. We show that Tus/Ter 81 creates an efficient site-specific RFB in human cells and observed local activation of 82 ATR signaling which was responsible for the phosphorylation of DNA damage marker 83 γH2AX at the stall sites. When a replication fork pauses at the local Tus-TerB block, we 84 do not detect any alteration in global replication profiles. Our system allows us to study 85 the ATR-checkpoint activity as a local response to a single RFB. 86 87

Tus is found enriched at TerB sites integrated in human cells 89
The 23 bp TerB sequence in interaction with the Tus protein have been successfully 90 used in yeast and mice to create artificial RFBs (Larsen et al., 2014a, 2014bWillis et 91 al., 2018Willis et 91 al., , 2017Willis et 91 al., , 2014Willis and Scully, 2016). To study the effect of site-specific 92 replication blocks at a genomic locus in human cells, a previously established MCF7 93 clone with a unique integrated copy of a plasmid carrying two sets of 5 repeats of the 94 TerB sequence in the non-permissive orientation was used (Fig.1A, SupFig1). Whole 95 genome sequencing of the clone was carried out to confirm the single copy integration 96 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted March 26, 2023. ;https://doi.org/10.1101/2023 doi: bioRxiv preprint 7 described (Collis et al., 2008;Nandi and Whitby, 2012;Panday et al., 2021;Xue et al., 158 2015). Together, these results show that Tus expression induces a RFB at the TerB 159 arrays impairing the replisome progression during S-phase. 160 γH2AX is enriched at TerB sites after Tus expression before fork collapse 161 One of the earliest responses to replication stress is the phosphorylation of the histone 162 variant H2AX on serine 139 (γH2AX) by members of the phosphoinositide 3-kinase 163 (PI3K)-like family (PIKK) (Ward and Chen, 2001). As we demonstrated that TerB arrays 164 can block incoming replication forks, we asked whether we could see a γH2AX signal 165 enrichment due to fork arrest. Using a co-immunoprecipitation assay with the chromatin 166 fraction of cross-linked-cells expressing either GFP tag or GFP-Tus, we found that  Tus and γH2AX were immunoprecipitated together using GFP or γH2AX antibodies 168 ( Fig.3A). 169 To gain insight into the γH2AX signal at the RFB induced by the Tus-TerB 170 interaction, cells were transfected with either VC or HA-Tus expression plasmids, and 171 PLA was performed using antibodies against HA and γH2AX (Fig.3B). We found that 172 16% of the HA-Tus transfected cells were harboring at least one PLA signal versus 3% 173 of the VC transfected cells (Fig.3C). 174 To better characterize the γH2AX enrichment around TerB sites, we performed 175 ChIP-qPCR assays using a γH2AX antibody 24h after Tus or VC expression. We 176 observed an enrichment of γH2AX, strictly co-localizing with the Tus-Ter interaction 177 sites (PP9 and PP47), suggesting a local role of γH2AX in response to stalled RFs 178 ( Fig.3D, E). This is in stark contrast to the observed spreading of the γH2AX signal after 179 a site-directed DSB (Berkovich et al., 2007;Chailleux et al., 2014;Clouaire et al., 2018;180 Savic et al., 2009). To corroborate this observation in our integrated cassette, we 181 generated a site-specific DSB by expressing Cas9 targeted to the integrated cassette 182 (SupFig.4A-C). We assessed γH2AX using the ChIP-qPCR assay 24h after transfection 183 and noticed enrichment at the distal primer pair (PP10) contrasting with the very tight 184 local signal observed at the Tus-TerB RFB (SupFig.4D). However, we did not find any 185 significant γH2AX signal between VC or Tus expression (SupFig 5A-B) supporting the 186 lack of global response. This suggests that the RFB generated by the interaction 187 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made We hypothesized that if the local γH2AX enrichment at TerB sites was ATR-204 dependent, a reduction of ATR activity would lead to a decreased γH2AX signal. To 205 validate the ATR inhibitor activity (ATRi), VE-822 (Fokas et al., 2012), we performed 206 immunoblotting using total protein extracts of cells treated with 2mM HU and compared 207 the level of phospho-ATR TH1989 with or without an ATR inhibitor. A 4h treatment with 208 ATRi reduced the level of phospho-ATR TH1989, reflecting a decrease in ATR activity 209 ( Fig.4B). ChIP-qPCR assays were performed using a γH2AX antibody 24h after VC or 210 Tus expression and cells were harvested after a 4h ATRi treatment. Consistent with our 211 hypothesis, the γH2AX enrichment in the vicinity of TerB sites after Tus expression was 212 remarkably decreased upon ATRi treatment (Fig.4C PP9 and PP47), implying that the 213 Tus-TerB RFB activates a localized stress response that stimulated the local 214 phosphorylation of H2AX. Together, these data suggest a local activation of the intra-S 215 checkpoint via the ATR kinase and rapid phosphorylation of H2AX, which could mediate 216 the recruitment of repair factors near the damage site. 217 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

218
Replication stress occurs during the S-phase of the cell cycle, when the replication 219 machinery encounters DNA lesions that cause stalling of replicative polymerases and 220 can be a significant cause of genomic instability. If the stalled forks cannot be 221 processed, they can result in DNA breakage, mutations, and chromosomal 222 rearrangements leading to the development of many different human cancers (Gaillard 223 et al., 2015;Tubbs and Nussenzweig, 2017). In this study, we used the reconstituted 224 E.coli-derived protein-DNA barrier, Tus-TerB, to study the replication stress response at 225 a single RFB. We have shown that the Tus protein efficiently binds at the TerB 226 sequences using ChIP-qPCR and PLA in MCF7 cells (Fig.1). We confirmed that this 227 integrated system was able to cause site-specific RFB at TerB sequences when Tus 228 was introduced with the use of the SMARD technique and observed the accumulation of 229 replication fork protein MCM3 upstream of TerB sequences (Fig.2). Our results indicate 230 that the integrated Tus-TerB system acts as an efficient site-specific replication fork 231 barrier, providing valuable insights into the local processing of a stalled single fork in 232 human cells and its difference from global replication stress (Fig.5). 233 We show that the site-specific replication stress leads to the accumulation of 234 γH2AX at the site of fork block when Tus was expressed by co-immunoprecipitation of 235 Tus and γH2AX (Fig.3A). Additionally, we confirmed that γH2AX is localized at the site 236 of Tus-TerB using PLA and ChIP-qPCR assays (Fig.3B,C). We suggest that the γH2AX 237 signal is being constrained to a region of less than a kilobase from the TerB sites by 238 tortional stress generated when the replication fork encounters the barrier and is not 239 influenced by well-defined topological domains that are one measure of the functional 240 units of the genome (Dixon et al., 2012;Rao et al., 2014). Furthermore, there was no 241 observable difference in global replication profiles or the activation of global checkpoint 242 markers such as pRPA, pCHK1, and pATR, with or without Tus protein expression, 243 contrasting with the differences observed in the presence and absence of HU-induced 244 replication stress (Fig.4A, SupTable1 & SupFig.5,6). We conclude that Tus-TerB-245 induced replication fork stall does not activate global ATR-dependent S-phase 246 checkpoint signaling, suggesting that a local response is occurring at the stalled 247 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted March 26, 2023. ; https://doi.org/10.1101/2023.03.26.534293 doi: bioRxiv preprint replication fork. Interestingly, upon ATR inhibition, we observed that the enrichment of 248 phosphorylated H2AX at TerB sites was significantly reduced when Tus was expressed, 249 which indicates local ATR signaling is responsible for the phosphorylation of H2AX at 250 the stalled site (Fig.4C,D). This localized ATR-dependent γH2AX differs from the well-251 observed ATM-dependent spreading of γH2AX signal after a site-directed double-strand 252 break (Berkovich et al., 2007;Chailleux et al., 2014;Clouaire et al., 2018;Savic et al., 253 2009) and suggests a distinct local intra-S phase checkpoint is being activated in 254 response to the single RFB. 255 There are additional mechanisms to be understood in this newly described local 256 replication stress response, which includes whether the activation of ATR is dependent 257 on ATRIP. ATRIP requires binding to ssDNA, but in unpublished observations we have 258 not found significant ssDNA (using RPA-ChIP) in response to Tus-TerB replication 259 stalling. In the Tus-TerB system, there is no known dissociation of the replicative 260 helicase from the polymerase, which is responsible for the ssDNA signal. Therefore, just 261 like ATM can be activated independently of the MRE11 complex (Bakkenist and Kastan, 262 2003), we expect that ATR is activated independently of ATRIP, likely due to local 263 changes in chromatin that will be investigated in future studies. 264 During extensive DNA damage, the genome-wide checkpoint response activates 265 ATR/CHK1 globally, which leads to the slowing of all replication forks in the cell, 266 inhibition of cell cycle progression, and suppression of late origin firing; to provide 267 sufficient time for DNA repair. It has been proposed that a local checkpoint response 268 occurs when one or a few replication forks encounter a DNA lesion and activates 269 ATR/CHK1 signaling at local sites of fork stalling. The local ATR response has been 270 hypothesized to be transient, in which the fork moves slowly only at stalled sites to 271 promote fork stabilization, restart the stalled fork, and suppress recombination without 272 triggering the global checkpoint response (Iyer andRhind, 2017, 2013;Kaufmann et al., 273 1980;Merrick et al., 2004;Saxena and Zou, 2022;Willis and Rhind, 2009b;Zeman and 274 Cimprich, 2013). 275 We predict that the local ATR-dependent checkpoint signaling can result in a 276 rapid and controlled response to allow RFB resolution, through the recruitment of repair 277 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made and it is unlikely that shifting the position of the RFB would make it easier to resolve. 296 Therefore, fork cleavage is the most likely mechanism for resolving the RFB. 297 The upstream pathway of activation of the DNA damage response, as a result of 298 a single RFB, is not completely understood. Previous work has shown that FANCM acts 299 as a scaffolding protein for recruitment of different repair protein complexes involved in 300 the stalled replication fork rescue (Panday et al., 2021;Willis et al., 2017). We observed 301 that the FANCM protein is enriched at the site of the RFB induced by Tus-TerB 302 (SupFig.3), confirming a similar phenomenon observed in mouse cells. However, it is 303 yet to be determined if FANCM recruitment to the RFB precedes local H2AX 304 phosphorylation by ATR, or if FANCM is recruited as a consequence of the γH2AX 305 signal to help elicit the downstream DNA damage response (DDR). We also anticipate 306 that the protein complex, 9-1-1 (RAD9A, HUS1, RAD1), and the subsequently recruited 307 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted March 26, 2023. ; https://doi.org/10.1101/2023.03.26.534293 doi: bioRxiv preprint TOPBP1, may also play a role at this local ATR-dependent intra-S checkpoint (Helt et 308 al., 2005;Parrilla-Castellar et al., 2004). 309 While the Tus-TerB system is an important tool in deciphering signaling at 310 individual replication forks, it will be important to test if the local-S phase model remains 311 supported when exploring endogenous lesions that occur when replication forks 312 encounter secondary structures such as R-loops, G-quadruplexes and common fragile 313 sites characterized by trinucleotide repeat expansion [(CAG) n /(CTG) n ] (Brickner et al., 314 2022;Bryan, 2019;Kim et al., 2016). Here the mechanisms surrounding the 315 identification and resolution of the RFB may differ. Overall, our results indicate that Tus-316 TerB system acts as an efficient RFB and activates local ATR checkpoint signaling at 317 the stall site, leading to phosphorylation and accumulation of the DNA damage sensor 318 protein γH2AX, which is dependent on the ATR kinase. The local γH2AX accumulation 319 at the stalled region would lead to the recruitment of DNA repair factors for resolution of 320 the fork (Fig.5). Together, our findings reveal the signaling mechanism of the Tus/TerB 321 induced replication block and showed that the site-specific fork block is dependent on 322 the local ATR S-phase checkpoint signaling. For PLA, cells were seeded at 1X10 5 in a 12-well cell culture plate. For T7 assay, cells 331 were seeded at 2.5X10 5 in a 6-well cell culture plate. For ChIP, IP and cellular 332 fractionation, cells were seeded at 2.5X10 5 in a 10 cm cell culture dish. 333 MCF7 5C-TerB cells were seeded and transfected the day after with 10 μg of 334 pCMV3xnls Tus or pCMV3xnls using the Mirus Bio™ TransIT™-LT1 reagents for 24h 335 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

Plasmid construction 352
The doxycycline-inducible lentiviral plasmid used to express N-terminally myc-tagged, 353 C-terminally SNAP-tagged nuclear localized, human codon-optimized wild-type Tus 354 (Myc-NLS-TUS-SNAP) was generated as follows. The N-terminally myc epitope-tagged, 355  (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted March 26, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 PBS, and fixed and permeabilized with 4% paraformaldehyde containing 0.2% Triton X-424 100 for 20min on ice and blocked with PBS-BSA 3% overnight at 4. Coverslips were 425 incubated with primary antibodies (see Table 1 for dilution) for 1h at RT. Proximity 426 ligation was performed using the Duolink® In Situ Red Starter Kit Mouse/Rabbit  Aldrich) according to the manufacturer's protocol. The oligonucleotides and antibody-428 nucleic acid conjugates used were those provided in the Sigma-Aldrich PLA kit. Images 429 were quantified by counting the number of foci per nucleus using Nikon software. 430

Immunoprecipitation 431
For immunoprecipitation, MCF7 5C-TerB cells were pre-extracted using CSK100 buffer 432 (100 mM NaCl, 300mM sucrose, 3 mM MgCl2, 10 mM PIPES pH 6.8, 1 mM EGTA, 433 0.2% Triton X-100, anti-protease and anti-phosphatase) for 5 min on ice, fixed in 1% 434 formaldehyde for 10 min on ice and a 1% glycine solution was used to stop the reaction. 435 After scraping the cells in ice-cold PBS and centrifugation, pellets were lysed in SDS 436 lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.1% SDS, anti-proteases and anti-437 phosphatases) for 10 min on ice and sheared for 3 min. Samples were cleared by 438 centrifugation for 5 min at 4°C. Immunoprecipitations were performed on 10-fold diluted 439 lysates in dilution buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl,5 mM EDTA, 0.2% 440 Triton X-100, anti-proteases and anti-phosphatases). with antibodies against GFP, 441 γH2AX or IgG overnight at 4°C on a wheel. Beads were extensively washed in the 442 dilution buffer and denatured in 2X Laemmli buffer. 443 Proteins were separated on 4-12% acrylamide SDS-PAGE, transferred on 444 Nitrocellulose membrane and detected with the indicated antibodies described in the 445 table and ECL reagents. 446

Cellular fractionation 447
For the cellular fractionation, MCF7 5C-TerB cells were scraped in PBS, divided in 2 448 different tubes (1/3 of the volume for the whole cell extract (WCE) and the remaining 2/3 449 for the fractionation) and centrifuged to keep the pellets. 450 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted March 26, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 For the WCE, cells were lysed in 1 volume of lysis buffer (50 mM Tris Ph 7.5, 20 Mm 451 NaCl, 1 Mm MgCl2, 0.1% SDS, anti-protease and anti-phosphatase) for 10 min at RT 452 on a wheel and denaturated in 2X Laemmli buffer. 453 For the fractionation, MCF7 5C-TerB cells were pre-extracted in 2 volumes of CSK100 454 (100 mM NaCl, 300mM sucrose, 3 mM MgCl2, 10 mM pipes pH 6.8, 1 mM EGTA, 0.2% 455 Triton X-100, anti-protease and anti-phosphatase) for 15 min on ice, and centrifuged. 456 The supernatant (SN) representing the soluble fraction was kept in a new tube (soluble 457 fraction). The pellet was washed with CSK50 (50 mM NaCl, 300mM sucrose, 3 mM 458 MgCl2, 10 mM pipes pH 6.8, 1 mM EGTA, 0.2% Triton X-100, anti-protease and anti-459 phosphatase and resuspended in 2 volumes of CSK50 containing benzonase for 10 min 460 on a rotating wheel, and the SN was kept after centrifugation (chromatin fraction). All 461 the fractions were denaturated in 2X Laemmli buffer. 462 Proteins were separated on 4-12% acrylamide SDS-PAGE, transferred on 463 Nitrocellulose membrane and detected with the indicated antibodies described in the 464 Detection of DSBs using the T7 endonuclease assay 476 MCF7 5C-TerB cells were seeded and transfected the day after with 120 pmol of Cas9 477 protein and 120 pmol of the sgRNA TerB1 using the Lipofectamine CRISPRMAX™ 478 Cas9 transfection reagents (previously described). 24 hours post-transfection, cells 479 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted March 26, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 were lysed using a lysis buffer (100 mM NaCl, 10 mM Tris-HCl pH 8, 25 mM EDTA pH 480 8, 0,5% SDS) and 50 ug of Proteinase K overnight at 50°C with shaking. After addition 481 of NaCl and mix for 1 min, the supernatant was put in a new tube and an ethanol 482 precipitation was performed. A PCR to amplify the 900 bp region surrounding the 483 sgRNA TerB1 site was conducted using PP9F and R. Heteroduplex were formed by 484 heating and cooling down the samples and a T7 endonuclease assay digestion was 485 performed prior electrophoresis for detection. 486 Cell Cycle progression 487 MCF7 5C-TerB cells were transfected as described before. Control cells were treated 488 with 2mM HU for 4 hours before harvesting. Cells were washed with PBS and fixed in 1 489 ml cold 70% ethanol for 30 minutes. Cells were pelleted and washed with PBS. 100 490 µg/mL RNase A was added 50 µg/mL PI solution was added directly to the pellet, mixed 491 well and incubated for 30 mins at room temperature in the dark. 50,000 cells per 492 condition were analyzed by flow cytometry. 493

Immunofluorescence (γH2AX) 494
Cells were fixed with 4% PFA/PBS for 20 min, permeabilized with 0.5% Triton-X 100 for 495 10 min, washed 3 times in 1X PBS, and blocked in 10% goat serum overnight at 4°C. 496 The primary antibody (1:500) was incubated for 2 hr at room temperature. Cells were 497 then washed 3 times in 1X PBS. The secondary antibody (1:1500) was incubated for 1 498 hr at room temperature in dark. Cells were then washed again 3 times in 1X PBS. 499 Stained cells were mounted with mounting medium containing DAPI. Slides were 500 imaged at 60X (immersion oil) using Nikon A1 spinning disk confocal microscope. 501

Image Analysis 502
For PLA and gH2AX experiments, slides were imaged at 60X (immersion oil) with Nikon 503 spinning disk confocal microscope. PLA foci per nucleus and gH2AX foci per nucleus 504 were calculated using Nikon Elements AR Analysis Explorer (version 5.21.03), where 505 DAPI was used as a mask for the nucleus. The number of PLA foci per nucleus were 506 quantified to get the percentage of cells with 1 or 2 foci indicating a positive signal at the 507 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted March 26, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 site-specific replication block. The number of gH2AX foci was counted for each DAPI to 508 obtain the average number of gH2AX foci in each condition. 509

NGS sequencing of the MCF7 5C-TerB clone 510
Genomic DNA was extracted from the cells and submitted to Novogene for sequencing. 511 The samples were processed and whole genome sequencing was performed using their 512 hWGS service pipeline. 513 information required to reanalyze the data reported in this paper is available from the 526 lead contact upon request. 527

528
We are indebted to Ino de Bruijn, Jorge S. Reis-Filho, and Yingjie Zhu for help with the 529 genomics data. We thank members of the Powell and Schildkraut lab for their 530 comments on the manuscript. This work was supported in part by NIH Grants R01-531 CA187069 and P50-CA247749 (to S.N.P.); 5R01-GM045751 and R01-CA085344 (to 532 C.L.S.); National Cancer Institute Cancer Center Support Grants, P30-CA008748 at 533 MSK, and P30-CA013330 for use of a core facility at Albert Einstein COM. S.T. was 534 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted March 26, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 30 715 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted March 26, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023