Actin nucleators safeguard replication forks by limiting nascent strand degradation

Abstract Accurate genome replication is essential for all life and a key mechanism of disease prevention, underpinned by the ability of cells to respond to replicative stress (RS) and protect replication forks. These responses rely on the formation of Replication Protein A (RPA)-single stranded (ss) DNA complexes, yet this process remains largely uncharacterized. Here, we establish that actin nucleation-promoting factors (NPFs) associate with replication forks, promote efficient DNA replication and facilitate association of RPA with ssDNA at sites of RS. Accordingly, their loss leads to deprotection of ssDNA at perturbed forks, impaired ATR activation, global replication defects and fork collapse. Supplying an excess of RPA restores RPA foci formation and fork protection, suggesting a chaperoning role for actin nucleators (ANs) (i.e. Arp2/3, DIAPH1) and NPFs (i.e, WASp, N-WASp) in regulating RPA availability upon RS. We also discover that β-actin interacts with RPA directly in vitro, and in vivo a hyper-depolymerizing β-actin mutant displays a heightened association with RPA and the same dysfunctional replication phenotypes as loss of ANs/NPFs, which contrasts with the phenotype of a hyper-polymerizing β-actin mutant. Thus, we identify components of actin polymerization pathways that are essential for preventing ectopic nucleolytic degradation of perturbed forks by modulating RPA activity.


INTRODUCTION
Faithful genome duplication during cell division ensures accur ate tr ansmission of genetic information to daughter cells. This process relies on the replication of the entire genetic material during S-phase by thousands of replication forks (RFs) emanating from numerous origins of replication. Howe v er, replication is constantly challenged by endogenous and exogenous insults that may damage RFs (1)(2)(3). To respond to replicati v e stress or to prevent stalled forks from collapsing into DNA double-strand breaks (DSBs), or to ascertain that a mutagenic lesion is repair ed corr ectly, cells hav e e volv ed intricate genome surv eillance mechanisms underpinned by the ATR and ATM kinases ( 2 , 4-5 ). Mutations in genes involved in the DNA r eplication and r epair (DDR) pathway lead to human disorders that clinically manifest in de v elopmental abnormalities and cancer, underscoring their importance to human health ( 1 , 6-7 ).
A key step in the replication stress response (RSR) is the formation of RPA-coated single stranded DNA (RPA-ssDNA) complex at sites of RF damage. Replication protein A (RPA) is a heterotrimeric complex (RPA1, RPA2, RPA3) with essential roles in se v eral aspects of DNA metabolism, with conservation from yeast to humans. During DNA replication, RPA protects ssDN A transientl y formed at perturbed forks against nucleolytic degradation and serves as a platf orm f or r ecruitment and r egulation of se v eral DDR factors, including ATR ( 8 , 9 ). Interestingly, work in yeast Sacchar om y ces cer evisiae (Sc) and Xenopus showed tha t cells acti v ely control the ability of RPA to associate with ssDNA, suggesting an importance of this process for DNAassociated transactions and genome stability ( 10 , 11 ). Indeed, d ysregula tion of RPA expression has been implicated in tumour pro gression, chemothera py-resistance, or DN Adamage tolerance (12)(13)(14). Accordingly, we have recently identified a novel function for the Wiskott-Aldrich syndr ome pr otein (WASp), a regula tor of ARP2 / 3-media ted actin nucleation, in promoting the assembly of RPA-ssDNA complexes in cis via its direct interaction with RPA ( 15 ). Interestingly howe v er, se v eral other factors involved in actin dynamics, for example N-WASp (nucleationpromoting factor, NPF) or ARP2 / 3 and Formins (actin nucleators, ANs) have all been shown to participate in a range of responses essential for genome maintenance, including DNA r eplication, centromer e maintenance, or DSB r epair (16)(17)(18)(19)(20)(21)(22)(23).
Here we find that besides WASp, another member of the NPF family, N-WASp, as well as the ANs ARP2 / 3 and DIAPH1, collecti v ely r eferr ed to hence after as (NPF / ANs), ar e r ecruited to RFs and function to facilitate the formation / stability of RPA-ssDNA complexes and safeguard genome duplication. Consequently, their loss triggers defecti v e ATR acti va tion, global RF d ysfunction and genome instability. Significantly, we discover that monomeric actin (G-actin) interacts directly with RPA, and cells expressing a hyper-depolymerization ␤actin (G13R) mutant phenocopy defects associated with NPF / ANs loss i.e. impaired RPA f oci f ormation, impaired ATR activation and fork stability. In contrast, cells expressing a hyper-polymerization ␤-actin (S14C) mutant do not show replication defects, likel y impl ying a more direct role of actin polymerisation in RF protection . In conclusion, we propose that the association of NPF / ANs with perturbed RFs plays a critical role, either directly or indirectly through actin state changes, in facilitating formation of RPA-ssDNA complexes essential for fork protection.

Cell lines and drug treatments
HeLa cells were a generous gift from Dr F. Esashi. U2OS cells and U2OS over expressing RPA (Super RPA) cells were a kind gift of Dr L. I. Toledo. These cell lines were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and standard antibiotics. Primary dermal fibroblasts were maintained in Dulbecco's modified Eagle's medium (DMEM; Life Technologies) supplemented with 20% fetal calf serum (FCS), 5% L-glutamine, and 5% penicillin-streptomycin (Invitrogen) antibiotics. Primary fibroblasts were immortalized with 293FT (Invitrogen)-deri v ed supernatant containing a human telomerase re v erse transcriptase (TERT) lentivirus that was generated with the plasmids pLV-hTERT-IRES-hygr o (gift fr om Tobias Meyer; Addgene #85140) ( 24 ), psPax2 (gift from Didier Trono; Addgene #12260), and pMD2.G (gift from Didier Trono; Addgene #12259). Selection was performed with hygromycin (Invitrogen) at 70 g / ml. Fibroblast complementation was carried out using a lentiviral vector (pLVX-IRES-Neo; TakaraBio) that encoded 3xHA-tagged DIAPH1 in combination with the lentiviral packaging plasmids described above. Selection was performed with geneticin (Invitrogen) at 400 g / ml. Expression of HA-tagged DIAPH1 was validated by Western blotting. All cell lines were routinely tested for mycoplasma. Cells were treated with hydroxyurea (Sigma-Aldrich) with indicated doses. Where indicated cells wer e pr e-tr eated with 100 M ARP2 / 3i (CK-666, Sigma-Aldrich) for 1 hour prior to further drug treatment or harvesting.

Rapid, efficient and practical (REAP) cellular fractionation
To fractionate cytoplasmic and nuclear compartments of the cell we carried out REAP fractionation as described previously ( 26 ). HeLa cells grown to 70-90% confluence in a 10 cm tissue-culture dish were washed 2 times in ice cold 1x PBS. Then, 1 ml ice cold PBS was added to a 10 cm tissue culture dish and cells were scraped and collected in 1.5 ml microcentrifuge tubes. Microcentrifuge tubes were pulse spun for 10 s and the supernatant was discarded. Cell pellets were then pipetted up and down 5 times in 1 ml ice cold 0.1% NP-40 / 1x PBS and pulse spun for 10 s in a microcentrifuge. The supernatant was then transferred to a new microcentrifuge tube (Cytoplasmic fraction). The cell pellet was then resuspended by pipetting up and down 5 times in 1 ml ice cold 0.1% NP-40 / 1x PBS and pulse spun and the supernatant was discarded. The nuclear pellet was then resuspended in 50 l of PBS and sonicated for 30 s at 10% amplitude using a soniprep 150 (MSE) probe sonicator. The protein concentration of the cytoplasmic and nuclear fractions were then quantified by Bradford assay and processed by SDS-PAGE and western blot transfer in advance of imm unoblot anal ysis.

YFP-NLS-␤-actin mammalian expression construct transfection for immunofluorescence microscopy
For YFP-NLS-␤-actin mammalian expression construct transfection, 2 × 10 5 cells were seeded into a well of a 6-well tissue culture plate. Cells were incubated at 37 • C for 24 h to allow them to adhere. Per condition, 0.75 g of YFP-NLS-␤-Actin mammalian expression construct was transfected following the Lipofectamine 3000 manufacturer's protocol. Cells were then incuba ted a t 37 • C for 24 h and treated with hydroxyurea at the indicated doses before harvesting cells for down-stream assays.

Cell survival assay
Alamar Blue survival assays were performed in accordance with the manufactur er's r ecommendations (Bio Rad). Briefly, 500 cells per well in 96-well plates were plated and untr eated or tr eated with indicated doses of hydroxyurea, cis-platin or Mitomycin C and incubated for 7 days. Alamar blue reagent was added to each well and fluorometric measurements taken after 4 h incubation at 37 • C.

Immunofluor escence microscop y
Cells were grown and treated on circular 13 mm diameter coverslips, thickness 1.5 mm. For visualization of 53BP1, RAD51 foci and Cyclin A, cells were fixed with 4% paraformaldehyde in PBS for 10 min at room temperature, washed twice in PBS and permeabilised with 0.2% Triton X-100 in PBS for 10 min at room temperature. Coverslips were washed 3 × in PBS cells and were blocked in 10% FBS in PBS for 1 h a t room tempera tur e befor e incubation with primary antibodies in 0.1% FBS in PBS overnight at 4 • C. Coverslips were then washed 4 × 5 min in PBS followed by incubation with secondary antibodies for 1 h at room temperature. Slides were then washed 4 × 5 min in PBS and subsequently mounted with Vectashield mounting medium (Vector Laboratories) with DAPI. Micronuclei were quantified by assessing DAPI stained nuclei. For visualization of RPA2 foci and BrdU foci cells were preextracted on ice for 2 min in CSK buffer (10 mM PIPES pH6.8, 300 mM sucrose, 100 mM NaCl, 1.5 mM MgCl 2 , 0.5% Triton X-100), washed with PBS and fixed with 4% paraf ormaldehyde in PBS f or 10 min at room temperature. Coverslips were then blocked and processed as above. Primary antibodies employed for immunofluorescence were as follows: 53BP1 (MAB3802, Millipore, 1:1000), Cyclin A (ab19150, Abcam, 1:500), RPA2 (NA-18, Calbiochem, 1:200), RAD51 (ab133534, Abcam 1:500), BrdU (BU-1; RPN202 GE Healthcare Life Sciences 1:200). Secondary antibodies for immunofluor escence wer e as follows: Alexa Fluor 488 anti-rabbit (A21206, Invitrogen, 1:200) and Alexa Fluor 555 anti-mouse (A31570, Invitrogen, 1:400). Images wer e acquir ed using a Leica SP8 confocal microscope with a 63x oil objecti v e, using a Zeiss Axio Observer Z1 Marianas ™ Microscope attached with a CSU-W spinning disk unit using either a Hamamatsu Flash 4 CMOS camera or a Photometrics Prime 95b sCMOS camera built by Intelligent Imaging Innovations (3i) or a Zeiss Axio Observer Z1 Marianas ™ Microscope attached with a CSUX1 spinning disk unit and Hamamatsu Flash 4 CMOS camera built by Intelligent Imaging Innovations (3i). Quantification was carried out using FIJI (ImageJ) software and CellProfiler (Broad Institute).
iPOND iPOND was performed as decribed previously ( 28 ). Briefly, lo garithmicall y growing HeLa S3 cells (1 × 10 6 per ml) or HEK293TN cells were incubated with 10 M EdU for 10 or 15 min respecti v ely. Following EdU labeling, cells were fixed in 1% formaldehyde, quenched by adding glycine to a final concentration of 0.125 M and washed three times in PBS. Collected cell pellets were frozen at −80 • C and cells were permeabilized by resuspending 1.0-1.5 × 10 7 cells per ml in ice cold 0.25% Triton X-100 in PBS and incubating for 30 min. Before the Click reaction, samples were washed once in PBS containing 0.5% BSA and once in PBS. Cells were incubated for 2 h at room temperature in Click reaction buffer containing 10 M azide-PEG(3 + 3)-S -S -biotin conjugate (Click ChemistryTools, cat. no AZ112-25), 10 mM sodium ascorbate, and 1.6 mM copper (II) sulfate (CuSO 4 ) in PBS. The 'no Click' reaction contained DMSO instead of biotinazide. Following the Click r eaction, cells wer e washed once in PBS containing 0.5% BSA and once in PBS. Cells were resuspended in lysis buffer (50 mM Tris-HCl pH 8.0, 1% SDS) containing protease inhibitor cocktail (Sigma) and sonicated. Samples were centrifuged at 14 500 rcf. at 4 • C for 30 min and the supernatant was diluted 1:3 with TNT buffer (50 mM Tris pH 7.5, 200 mM NaCl and 0.3% Triton X-100) containing protease inhibitors. An aliquot was taken as an input sample. Streptavidin-agarose beads (Novagen) wer e washed thr ee times in TNT buffer containing protease inhibitor cocktail. Two hundred microliters of bead slurry was used per 1 × 108 cells. The streptavidin-agarose beads wer e r esuspended 1:1 in TNT buffer containing protease inhibitors and added to the samples, which were then incuba ted a t 4 • C for 16 h in the dark. Following binding, the beads were then washed two times with 1 ml TNT buffer, two times with TNT buffer containing 1 M NaCl, two times with TNT buffer and protein-DNA complexes were eluted by incubating with 5 mM DTT in TNT buffer. Cross-links wer e r e v ersed by incubating samples in SDS sample buffer at 95 • C for 20 min. Proteins wer e r esolved on SDS-PAGE and detected by immunoblotting using specific antibodies.
iPOND in ND1 cells was performed as described ( 29 ) with some modifications. Briefly, T cells ( ∼80 million cells per sample) were incubated with 20 M EdU in 30 ml culture medium at 37 • C for 20 min and then washed with fresh medium a t 37 • C . For HU or thymidine pluse, cells wer e r e-suspended either with 30 ml fresh medium at 37 • C as the control or 4 mM HU or 20 M thymidine in fresh medium. These cells were further cultured at 37 • C for another 2 h. After labeling / treatments, each cell samples were washed with PBS, and crossed-linked in 1% formaldehyde in PBS for 20 min at RT, quenched with glycine at the final concentration of 0.125 M for another 5 min, and washed 2 times in PBS. Cell pellets were permeabilized with 10 ml permeabilizing buffer (0.3% Triton-X / 0.5% BSA in PBS) 30 min at RT and washed with 0.5% BSA / PBS. Each cell pellet was resuspended in 10 ml PBS as the control, or 10 ml fresh pr epar ed Click buffer (10 mM Sodium ascorbate, 2 mM CuSO 4 , and 20 M Biotin-dPEG7-azide) and incubated for 1-2 h at room temperature (RT). Cells were washed with 0.5% BSA / PBS and then pellets either frozen at -80 • C or immediately used for lysis. Each cell pellet was resuspended in 0.8 ml lysis buffer (25 mM NaCl, 2 mM EDTA, 50 mM Tris pH 8, 1% Igepal CA630, 0.2% SDS, 0.5% sodium deoxycholate, and 1 × Halt protease and phosphatase inhibitor cocktail (Thermo Fisher)) and incubated for 10 min on ice. Samples were sonicated with a Branson 250 using the settings, 20-25 W, 20 s constant pulse, and 40 s pauses for a total of 4 min on ice. Cell lysates were centrifuged at 1.8 ×10 3 rcf for 10 min at RT. The supernatants were collected and diluted with the dilution buffer (the lysis buffer without SDS or sodium deoxycholate). Streptavidin-agarose beads (Millipore Sigma) (80 l / sample) were washed with the dilution buffer 3 times, and then incubated with the diluted samples overnight at 4 • C. The beads were again washed 3 times with RIPA buffer. Captured proteins were separated from beads by incubating beads in 50 l 2 × Laemmli Sample Buffer (Bio-Rad) at 95 • C for 25 min. The supernatant was collected, and proteins resolved on 4-15% SDS-PAGE and detected by immunoblotting.

EdU labeling of nascent DNA and proximity ligation assay
Analysis of the association of proteins to nascent DNA by EdU labelling and the Proximity Ligation Assay was carried out as previously described ( 28 , 30 ). Briefly, cells grown on coverslips were labeled with 10 mM EdU for 10 min followed by treatment with 1 mM or 4 mM hydroxyurea at various timepoints as indica ted. Fixa tion was carried out with 3% formaldehyde, 2% sucrose in PBS for 10 min at room temperatur e. Slides wer e then washed twice with PBS and incubated with blocking solution (3% BSA in PBS) for 30 min. Following this, slides were washed 2 × in PBS before EdU-Biotin Azide conjugation by click chemistry using the Click-iT reaction (Click-iT assay kit (Thermo Fisher, according to the manufacturer's instructions). Coverslips were washed 2 × with PBS before primary antibody incubation overnight at 4 • C in 1% BSA / 0.1% saponin in PBS. Following primary antibod y incuba tion coverslips were washed 2 × in PBS and then the proximity ligation assay was carried out (Duolink In Situ Red Starter kit (Sigma Aldrich) according to the manufacturer's instructions). Coverslips were mounted using Vectashield containing DAPI. Antibodies employed for the PLA assay were as follows: Biotin (Bethyl Laboratories, A150-109A, 1:3000), (Biotin (Jackson Immunor esear ch, 200-002-211, 1:1000), WASp (Santa Cruz, sc-5300, 1:500), N-WASp (Abcam, ab187527, 1:500), PCNA (Santa Cruz, sc-56, 1:500), PCNA (Abcam, ab18197, 1:500) and RAD51 (Sigma Aldrich, PC130) . Images were acquired using a Zeiss Axio Observer Z1 Marianas ™ Microscope attached with a CSU-W spinning disk unit using either a Hamamatsu Flash 4 CMOS camera or a Photometrics Prime 95b sCMOS with a 63 × objecti v e. Image analysis was carried out with FIJI (ImageJ) software. PLA foci positi v e cells were scored as cell nuclei with equal or greater than one PLA foci within them.

DNA fibre analysis
DNA fibre assay was performed as described previously with some modifications ( 28 , 31 , 32 ). In brief, exponentially growing cells were first incubated with 25 M iododeoxyuridine (IdU) and then with 125 M chlorodeoxyuridine (CldU) for the indicated times. Fibre spreads were prepared from 0.5 × 10 6 cells / ml. Slides were stained as described previously ( 31 , 32 ). Images were acquired with Leica SP8 or Carl Zeiss LSM 710 Meta confocal microscope using a 63x oil objecti v e. Analysis was performed using the Ima geJ software packa ge (National Institutes of Health). A minimum of 100 fibres from three, unless stated otherwise, independent experiments were scored. Mann-Whitney test was used to determine statistical significance.
qPCR RNA extraction and cDNA synthesis were performed using the RNeasy mini kit (Qiagen) and High Capacity cDNA Re v erse Transcription Kit (Applied Biosystems). QPCR analysis was performed with QuantStudio ™ 6 Flex Real-Time PCR System (Applied Biosystems) using SYBR Green PCR Master Mix (Life Technologies) and the following primers 5 -GTCCTACTTCATCCGCCTTTAC-3 and 5 -TCGTCTGCAAAGTTCAGCC-3 for WASP and 5 -GGCAT GGACTGTGGTCATGAG-3 and 5 -T GCACC A CCAA CTGCTTAGC-3 for GAPDH.

RPA complex expression, purification and in vitro pull down
RPA comple x was e xpressed from a plasmid (a kind gift from David Cortez (Vanderbilt Uni v ersity, Nashville, Tennessee, US) ( 33 , 34 ) encoding 6xHis-tagged RPA1 (70 kDa), tagless RPA2 (32 kDa), and 6xHis-tagged RPA3 (14 kDa) in BL21 (D3) E. coli cells for 2 h at 37 • C in LB medium with 100 M / ml ampicillin after induction with 1 mM IPTG. Cells were harvested and lysed in buffer containing 50 mM Tris-HCl, pH 8.0, 400 mM NaCl, 1 mM PMSF, 10 M ZnCl 2 , 5% glycerol and protease inhibitor cocktail (Sigma, #S8830). Supernatant was applied to a HisTrap HP 5 ml column (DE Healthcare) calibrated in 5 volumes of calibration buffer containing 20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 1 mM PMSF, 10 M ZnCl 2 and 5% gly cerol. Ne xt, the column w as w ashed in 10 column volumes with washing buffer: 20 mM Tris-HCl, pH 7.6, 500 mM NaCl, 20 mM imidazole, 10 M ZnCl 2 and 5% glycerol. Then, the proteins were eluted from column with the linear imidazole gradient in final concertation as follow, 20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 300 mM imidazole, 10 M ZnCl 2 and 5% gly cerol. The RPA comple x of three proteins was bound to the column and was eluted from it by high imidazole concentration (about 250 mM). The eluate was diluted to bring the NaCl concentration down to 50 mM. It was then applied to a HiTrap Heparin HP 5 ml column (GE Healthcare) in 20 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM PMSF, 10 M ZnCl 2 and 5% glycerol, and eluted with 10 volumes of 50-1000 mM NaCl linear gradient. The fractions containing all thr ee proteins wer e joined together and purified RPA complex was r ebuffer ed into stora ge b uffer (50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM DTT, 5% glycerol), aliquoted and frozen in -80 • C. The complex composition was confirmed by western blot using antibodies recognizing His-tag peptide and RPA2 protein and by binding RPA complex to ssDNA and lack of biding it to dsDNA. Interaction of RPA complex with monomeric actin was analysed using G-Actin Sepharose beads (Hypermol). Reaction was performed in a buffer containing (10 mM Tris-HCl pH7.5, 150 mM NaCl, 1 mM DTT, 5% glycerol, 0.1% NP-40). Complex es wer e extensi v ely washed with IP buffer before elution in 2 × SDS sample buffer and subsequently boiled for 3 min followed by centrifugation. The resultant supernatant fraction was retained as the eluate.

String analysis
Gene accession codes from iPOND analysis were input into the multiple protein search function of STRING ( https://string-db.org/cgi/input? sessionId=b5U7THWxoyjd&input page acti v e form= multiple identifiers ) ( 35 ) with the detected organisms selected as Homo sapiens . The network was then exported from STRING and imported into Cytoscape (v3.8.2). Functional enrichment network analysis was then conducted using the STRING plug-in, with the network to be set as background set to genome. The STRING Enrichment tab was selected, and the generated network was filtered by GO Biological pr ocesses. GO Biological pr ocesses associated with all the actin related processes were selected and exported as a new network. The subsequently generated network was saved, and the file was uploaded to STRING ( https://string-db.org/cgi/input?sessionId= b5U7THWxoyjd&input page show search=on ).
Acti v e interaction sources settings were adjusted to display links associated with only experiments and databases. The minimum r equir ed interaction scor e was adjusted to high confidence (0.700) and the network was updated. GO terms GO:0034314 Arp2 / 3 complex-mediated actin nucleation and GO:0008154 Actin polymerization or depolymerization were selected in the analysis tab functional enrichment biological process list.

Statistics and reproducibility
Statistical analyses were done using GraphPad Prism 9 (GraphP ad Softwar e Inc.) or Microsoft Excel. Unpaired Students' t -test or Mann-Whitney test were used to determine statistical significance as indicated in the figure legends.

Contact for reagent and resource sharing
Further information and requests for r esour ces and reagents should be directed to the Lead Contact, Wojciech Niedzwiedz ( wojciech.niedzwiedz@icr.ac.uk ).

NPF / AN family proteins are enriched at perturbed replication forks
To address the possible role of NPF / ANs during normal and stressed DNA replication we performed a deep mining of our iPOND (isolation of proteins on nascent DNA) data set ( 28 , 36 , 37 ) combined with STRING analysis ( http://string-db.org ) ( 35 ) performed with or without hydroxyurea (HU)-mediated replicati v e stress (RS) in HeLa cells. This analysis identified a cluster of proteins involved in actin polymerisation, including the ARP2 / 3 complex, WASp family proteins, and members of the formin family of actin nucleators (DIAPH1, 2 and 3), but not Spire family proteins, the latter known to nucleate actin polymers directly without ARP2 / 3 (Figure 1 A). We validated the presence of these factors at replication forks (RFs) in two ways: a) iPOND / Western blot (WB) analysis and b) EdU-based proximity ligation assay (PLA), aka quantitati v e in situ analysis of protein interaction at RFs (SIRF) ( 15 , 38 , 39 ). Using iPOND / WB analysis we could readily detect W ASp, N-W ASp, and DIAPH1, as well as RPA2 (control) enriched at HU-perturbed RFs in both human T cells and HEK293TN cells (Supplementary Figure S1A-C). Notably, a thymidine chase decreased their association with nascent DNA as compared to HU treated samples, which indica tes tha t these factors ar e likely r ecruited specifically to RFs in a RSR-dependent manner ( 29, 40 ) (Supplementary Figure S1A-C). Accordingly, utilizing PLA we could also readily detect proximity of the NPFs (N-W ASp, W ASp) to newl y-synthesised DN A (monitored by EdU labelling of nascent DNA marking RFs) in HeLa cells. The nuclear PLA signal for N-WASp and WASp was significantly enriched upon HU-induced RS compared to untreated as well as control samples (Figure 1 B). Upon 4 mM / 3 h HU treatment, which is known to induce replication fork collapse ( 29,41 ), we observed a decrease in the PLA signal when compared to 1 mM / 1 h. This is consistent with the fact that fork collapse leads to dissociation of RF associated factors ( 29,41 ). Finally, thymidine addition resulted in a significant decrease in PLA signal between N-WASp or WASp and nascent DNA (Figure 1 C), which is in support of our iPOND data. To ascertain that the observed phenotype is not restricted to EdU labelling we repeated these analyses using an antibody against PCNA --an established marker of RFs. Again, we readily detected association between PCNA and N-WASp or WASp ( Supplementary Figure S1D and E) confirming their proximity to sites of DNA replication, which is in line with our recent analysis demonstrating WASp localisation to RFs in human T and B cells ( 15 ). Taken together, these data suggest that members of the NPF / AN family of proteins enrich at perturbed RFs in human immune and nonimmune cells, malignant or nonmalignant.

Deficiency in NPF / ANs prov ok es global replication dysfunction
To address the mechanism by which NPF / ANs manage RS, we employed the DNA fibre assay to analyse various aspects of str essed DNA r eplication ( 42 , 43 ). Strikingly, r eplicati v e stress induced by HU resulted in a significant reduction in the average fork velocity in cells transiently depleted of N-W ASp, DIAPH1 or W ASp by RNA interfer ence r elati v e to WT control (Figure 2 A, Supplementary Figure S2A-C). To validate the specificity of siRNA data we recapitula ted these observa tions using (i) multiple different siRNAs against N-WASp or WASp (Supplementary Figure S2D), (ii) patient-deri v ed DIAPH1-deficient cell-line and a complemented control cell line (Figure 2 B), as well as (iii) a specific ARP2 / 3 inhibitor (CK-666; Supplementary Figure  S2E). In addition, loss of W ASp, N-W ASp or DIAPH1 significantly impaired fork restart upon HU treatment (Figure  2 C). WASp family members function as positi v e regulators of ARP2 / 3-media ted genera tion of short branched F-actin filaments ( 44 ) whereas formins, of the DIAPH-family, nucleate long unbranched F-actin filaments ( 45 ). Thus, gi v en the striking similarity of the replication phenotypes we observed in the tested NPF / ANs, we next tested if these factors function in a single or parallel pathway(s) to manage RS. To this end, we tested the impact on DNA replication (measured as fork speed) of either WASp or DIAPH1 depletion alone or together with an ARP2 / 3 inhibitor under HU-induced stress. Targeting either WASp or ARP2 / 3 alone as well as inhibiting both together resulted in a similar replica tion defect, indica ting tha t WASp and ARP2 / 3 both function in the same pathway to manage RS (Supplementary Figure S3A). In contrast, inhibition of ARP2 / 3 in DIAPH1-deficient cells led to a significantly greater replication defect than the defect displayed by DIAPH1-deficient cells alone (Figure 2 D), indicating an additi v e effect between branched and unbranched actin polymerisation pathways in promoting DN A replication. Taken to gether, these data suggest that WASp / ARP2 / 3-and DIAPH1 (formin)mediated actin signalling is r equir ed to facilitate fork stability and global RF dynamics.

Interfering with actin polymerisation elicits replicationassociated DNA damage
Unstable RFs are prone to collapse into single-ended DSBs. Thus, to determine the fate of HU-stalled RFs when actin nucleation is impaired, we analysed the formation of 53BP1 foci (a canonical marker of DSB) in cells depleted for either N-W ASp, W ASp or DIAPH1, either in unperturbed conditions, after 1 mM / 1h HU treatment (mild RS associated with fork stalling) or 4 mM / 3h HU treatment (high RS, associated with elevated RF collapse) ( 46 ). We noticed a significant increase in the number of 53BP1 foci in cy clin A-positi v e (S / G2-phase) cells after depleting either of the three proteins as compared to their respecti v e controls, most likely due to e xtensi v e for k collapse (Figure 3 A).
Notably, we reproduced these findings in cells treated with an ARP2 / 3 inhibitor (CK-666) and in the patient-deri v ed DIAPH1-deficient cells (Supplementary Figure S3B). Fork collapse is known to induce genome instability resulting in the increased formation of micronuclei, which are particularly prominent in cancer cell models (e.g. HeLa) that display an intrinsically high le v els of replicati v e stress. Consistently, loss of WASp, N-WASp, or DIAPH1 as well as inhibition of ARP2 / 3 resulted in increased le v els of micronuclei compared to control cells e v en in the absence of any exogenous replication stressors (Figure 3 B and Supplementary Figure S3C). This observation further underscores the role of NPF / ANs in pre v enting genome-instability not only in response to exogenous genotoxic insults but also those arising from endogenous sources of RS, including but not limited to R loops (RN A-DN A hybrids), transcriptionreplication (T-R) conflicts, etc. ( 15 , 17 ). Collecti v ely, these data support the notion that NPF / ANs promote recovery of perturbed RFs and suppress replica tion-associa ted genome instability under normal growth conditions as well as in response to exogenous replicati v e stress.

Actin nucleation protects r eplication f orks fr om uncontr olled degradation by promoting RPA localisation to ssDNA
Since unstable RFs undergo e xcessi v e nucleolytic degradation ( 47-50 ), we tested whether loss of NPF / ANs influences this process by employing a modified DNA fibre protocol ( 28 ). Interestingly, depletion of N-WASp, DIAPH1 or WASp in HeLa cells, resulted in a significant shortening of the CldU tracts (in the IdU > CIdU > HU labelling scheme) compared to controls (Figure 4 A). This da ta indica te tha t pa tholo gical nucleol ytic degradation of HU-perturbed forks is not restricted to WASp deficiency or to human T and B lymphocytes, but is a more general phenotype associated with depletion of multiple NPF / AN proteins ( 15 ). As pathological fork resection is attributed to the unrestrained activity of the MRE11 or DNA2 nucleases, we monitored fork degradation in cells treated with the MRE11 inhibitor mirin, or siRNA targeting DNA2. Treatment with mirin or depletion of DNA2 supressed e xcessi v e for k degradation observ ed in cells depleted of NPF / ANs (Supplementary Figure S3D). Moreover, depletion of SMARCAL1 (SWI / SNF rela ted, ma trix associa ted, actin dependent regulator of chromatin, subfamily A like 1) also rescued e xcessi v e for k degradation in cells depleted of N-W ASp, W ASp or DIAPH1 suggesting a role in fork protection downstream to RF re v ersal (Figure 4 B and Supplementary Figure S3E). Since pathological fork resection is attributed to fork deprotection ( 51 ) Figure S4C). In line with this, loading of RAD51 at perturbed / collapsed RFs, an e v ent dependent on RPA-ssDNA formation ( 8 ) was also defecti v e in WASp and DI-APH1 depleted cells, as measured by the PLA assay between nascent DNA (EdU labelling) and RAD51 ( Figure  5 A). Furthermore, the number of RAD51 foci detected by IF after HU treatment upon WASp depletion was also significantly reduced when compared to those treated with control siRNA (Supplementary Figure S4D), thus supporting our PLA-based analysis. Formation of RPA-ssDNA is r equir ed for ATR acti vation and consequently, we hav e shown recently that WASp is r equir ed for this r esponse in T and B cells ( 15 ). To ascertain whether other members of the NPF / ANs family are similarly r equir ed for this response we analysed the efficiency of ATR signalling in cells depleted for N-WASp or DIAPH1. Again, and in line with the defecti v e RPA loading onto ssDNA, ATR activity was significantly compromised as evidenced by impaired phosphorylation of CHK1 (pSer345) in N-W ASp, W ASp and DIAPH1 depleted HeLa cells, and incr eased r eplication origin firing in N-WASp or WASp depleted HeLa cells and in DIAPH1-deficient cells (Figure 5 B and C).
Since impaired formation of RPA foci and defecti v e DNA replication was not restricted to WASp deficiency alone, but was observed with loss of multiple other members of the NPF / ANs protein family, we considered that this phenotype could arise through at least three distinct mechanisms: (i) an overall decrease in the le v els of RPA expression in cells, (ii) defecti v e nucleolytic processing of HUperturbed forks failing to generate ssDNA, and / or (iii) defects in a downstream process that promotes RPA association with ssDNA. To distinguish between these possibilities, we first analysed total RPA2 protein le v els in cells depleted of NPF / ANs, which indicated that loss of these factors does   not affect RPA2 le v els (Figure 6 A), as we previously showed in WASp-deficient T cells ( 15 ). We also used sub-cellular fraction and WB to confirm that the cytoplasmic and nuclear pools of RPA1 and 2 did not alter upon depletion of NPF / ANs (Supplementary Figure S5A). Next, we analysed ef ficiency of ssDNA genera tion a t HU-perturbed RFs by BrdU staining under non-denaturing condition. Here too, we did not observe a decrease in ssDNA generation (endresection) at perturbed RFs upon depletion of these factors (Figure 6 B). Thus, we considered it likely that NPF / ANs may function as RPA 'chaperones' facilitating RPA loading and / or stability at RFs both, in cis as in the case of WASp ( 15 ) but also in trans , likely via modulating the actin state, i.e. G-actin (monomeric) versus F-actin (polymeric). A simple prediction from this model would be that a surplus of RPA may be enough to overcome limitation associated with suboptimal RPA chaperoning in NPF / ANdeficient cells. To test this hypothesis, we utilised the previously characterised SuperRPA U2OS cells generated in the Lukas lab, that display a modest two-fold excess of all three RPA subunits. Importantly, these cells retain a normal cellcycle profile and do not show spontaneous defects in either DNA replication or ATR activation ( 52 ). Strikingly, RPA ov er-e xpr ession largely r estor ed RPA loading / stability at HU-perturbed forks in DIAPH1-or WASp-deficient cells (Supplementary Figure S5B), as well as r estor ed global RF dynamics (Supplementary Figure S5C and D) with a concomitant rescue of ectopically-increased fork collapse -as measured by 53BP1 foci formation (Supplementary Figure  S6A and B). To examine the specificity of the rescue phenotype we tested if a surplus of RPA can rescue, in general, any phenotype associated with RF instability. To this end, we analysed DNA replication in SuperRPA cells depleted for BRCA2, since BRCA2-depletion confers a strong defect in fork stability ( 48,53 ). In support of a specific role for NPF / ANs in facilitating RPA association with ssDNA and RF stability, we did not observe a similar rescue of BRCA2associa ted replica tion phenotype in SuperRPA cells (Supplementary Figure S6C-E).

Actin polymerisation mutants recapitulate the phenotypes associated with loss of NPF / ANs
Our above results suggest that actin nucleators, i.e. components of ARP2 / 3 or formin-dependent pathways, act early in the RSR pathway to promote the localisation of RPA to perturbed forks, stimulating efficient RPA-ssDNA complex formation both, directly through WASp-RPA interaction ( 15 ) but also indirectly through another putati v e mechanism. Indeed, a r ecent r eport by Pfitzer et al. ( 54 ) demonstra ted tha t RPA2 coimmunoprecipita tes (co-IPs) with nuclear actin, which proposes a model that G-actin / F-actin state itself may somehow influence the ability of RPA to form RPA-ssDNA complexes. To test this hypothesis, we first show that sepharose-coated monomeric G-actin directly binds purified RPA complex (Figure 7 A and Supplementary Figure S7A) by in-vitro pull-down assays, suggesting that the in vivo RPA:G-actin association is also likely to be a direct interaction. Furthermore, we show that actin co-IPs with RPA2 in human cells spontaneously (without HU-induced r eplication str ess), howe v er, upon HU treat-ment, this RPA:actin interaction is decreased (Supplementary Figure S7B). Notably, and in direct support of the role of NPF / ANs in 'chaperoning' RPA-ssDNA complex forma tion, we observe tha t a hyper-depolymerising G13R ␤-actin mutant (YFP-NLS-␤-actin G13R ) ( 55 , 56 ) pr efer entially interacts with RPA2 in co-IP analyses from human HEK293TN cell extracts, when compared to WT (YFP-NLS-␤-actin WT ) or the S14C hyper-polymerising mutant of ␤-actin (YFP-NLS-␤-actin S14C ) (Figure 7 B) ( 56 ). Accor dingly, cells ov er-e xpressing an actin polymerizationincompetent (YFP-NLS-␤-actin G13R ) mutant that preferentially associates with RPA phenocopy the abnormal phenotypes associated with the loss of NPF / ANs i.e. defecti v e RPA foci forma tion, ATR activa tion as well as global impairment in DNA replication (Figure 7 C, D, Supplementary Figure S7C). In contrast, cells expressing actin pol ymerization-hypercompetent S14C m utant (YFP-NLS-␤-actin S14C ) showed phenotypes similar to cells expressing the control YFP-NLS-␤-actin WT (Figure 7 C, D) Note, WT and S14C mutant have a similar RPA foci phenotype (Figure 7 C), suggesting a rate-limiting step within the actin pa thway tha t may be distinct from actin filament formation, especially since we also show that all 3 actins (WT, S14C, G13R) are expressed at a similar le v el (Figure 7 E). Thus, our data suggest a unique functional and cooperati v e interaction between RPA, actin and NPF / ANs, where a combination of cis and trans , actin-state dependent mechanisms calibrate RPA activity at the forks to safeguard DNA replication (Model, Figure 7 F).

DISCUSSION
The RPA-ssDNA-protein complex plays a key role in protecting ssDNA generated during DNA metabolism, including at RFs, as well as acting as a 'platform' for r ecruitment / signalling / r egulation of a plethora of DNA repair factors. Howe v er, increased RPA le v els can be toxic likely due to e xcessi v e formation of RPA-ssDNA complexes ( 13,52,(57)(58)(59). Our recent report ( 15 ) as well as work in yeast Sacchar om y ces cer evisiae (Sc) and Xenopus together point to acti v e mechanisms employed by cells to dynamically control the ability of RPA to associa te / dissocia te with ssDNA ( 10 , 11 ).
Here, we have shown that not only WASp, but also other members of the NPF / ANs protein family such as N-WASp, ARP2 / 3 or DIAPH1 as well as actin itself, promote efficient localisation of RPA to perturbed forks and by doing so promote ATR activation and supress ectopic fork resection. In line with this, we show that NPF / ANs involved in ARP2 / 3 or formin-dependent actin polymerisation are specifically enriched at perturbed forks and their loss preclude RPA localisation to ssDNA resulting in defecti v e acti vation of ATR, e xcessi v e for k degradation and ultimatel y RF colla pse. The e xtensi v e for k degrada tion tha t we observed is likely due to increased fork r egr ession since this phenotype can be rescued by SMARCAL1 depletion, howe v er the amount of ssDNA generated at perturbed RFs seems to be similar between WT and NPFs / ANs mutants. Ther efor e, we conclude that the kinetics of resection that is r equir ed to genera te ssDNA a t the fork is unaf fected, but fork r egr ession coupled with fork deprotection (lack of   Replicati v e stress leads to for k stalling and generation of ssDN A, subsequentl y the association of actin nucleators (ANs) and nucleation promoting factors (NPFs) with sites of ongoing replica tion facilita tes RPA deposition, either directly (via WASp) or indirectly (via actin polymerization), in order to protect ssDNA generated at perturbed forks. This allows for an efficient fork restart and promotes genome stability. Depletion of NPF / ANs and / or interference with actin polymerization impairs RPA loading / stability on ssDNA leading to e xtensi v e nascent strand degradation and ultimately genome instability.