Nascent Chromatin Capture Proteomics Determines Chromatin Dynamics During DNA Replication and Identiﬁes Unknown Fork Components

RESOURCE

Nascent chromatin capture proteomics determines chromatin dynamics during DNA replication and identiﬁes unknown fork components

Constance Alabert1, Jimi-Carlo Bukowski-Wills2, Sung-Bau Lee1, Georg Kustatscher2, Kyosuke Nakamura1, Flavia de Lima Alves2, Patrice Menard1,4, Jakob Mejlvang1,4, Juri Rappsilber2,3,5 and Anja Groth1,5

To maintain genome function and stability, DNA sequence and its organization into chromatin must be duplicated during cell division. Understanding how entire chromosomes are copied remains a major challenge. Here, we use nascent chromatin capture (NCC) to profile chromatin proteome dynamics during replication in human cells. NCC relies on biotin–dUTP labelling of replicating DNA, affinity purification and quantitative proteomics. Comparing nascent chromatin with mature post-replicative chromatin, we provide association dynamics for 3,995 proteins. The replication machinery and 485 chromatin factors such as CAF-1, DNMT1 and SUV39h1 are enriched in nascent chromatin, whereas 170 factors including histone H1, DNMT3, MBD1-3 and PRC1 show delayed association. This correlates with H4K5K12diAc removal and H3K9me1 accumulation, whereas H3K27me3 and H3K9me3 remain unchanged. Finally, we combine NCC enrichment with experimentally derived chromatin probabilities to predict a function in nascent chromatin for 93 uncharacterized proteins, and identify FAM111A as a replication factor required for PCNA loading. Together, this provides an extensive resource to understand genome and epigenome maintenance.

Mammalian genomes are replicated by the simultaneous progress- packaged into a structure with nucleosomal periodicity and nuclease ion of thousands of replication forks1. In this process, chromatin resistance similar to bulk chromatin5,6. This process is referred to as organization is disrupted ahead of the replication machinery and chromatin maturation6,7. Restoration of nucleosomal density is rapid restored behind, on the two daughter strands2,3. This genome-wide and relies on recycling of old histones combined with addition of disruption of chromatin raises fundamental questions about how DNA new ones8. Old parental histones are thought to maintain their marks, replication is integrated with chromatin dynamics and how specific which are thereby transmitted to the daughter strands. In contrast, new chromatin structures are transmitted through mitotic cell division. histones H3–H4 are acetylated, principally at Lys 5 and 12 of histone These questions underpin how cells maintain epigenetically defined H4 (K5K12diAc; ref. 9). These histone H4 acetylations are removed gene-expression patterns and thus their identity, which is central to shortly after replication, as nascent (immature) chromatin matures development and disease avoidance2–4. into a nuclease-resistant state6,10,11. Chromatin restoration on newly synthesized DNA is a multi- The sliding clamp proliferating cell nuclear antigen (PCNA) plays layered process, including nucleosome assembly and remodelling, a central role in coupling replication with chromatin restoration restoration of DNA methylation and histone marks, deposition of through its ability to recruit the nucleosome assembly factor histone variants and establishment of higher-order chromosomal chromatin assembly factor 1 (CAF-1) the maintenance DNA structures including sister-chromatid cohesion2,3. Correspondingly, a methyltransferase DNMT1, several chromatin remodellers and large number of proteins are involved, for many of which timing and lysine deacetylases and methyl transferases2,3. Still it remains largely action are yet to be elucidated or even their identity to be revealed. unknown when and how most chromatin constituents and modifiers Pioneering work using nucleases to probe chromatin assembly, argues are recruited. As a discovery approach, proteomics has been used for a time window of approximately 15–20 min for new DNA to be to define the composition of purified telomeres12 and of mitotic

1Biotech Research and Innovation Centre (BRIC) and Centre for Epigenetics, University of Copenhagen, Ole Maaløes Vej 5, 2200 Copenhagen, Denmark. 2Wellcome Trust Centre for Cell Biology, University of Edinburgh, Edinburgh EH9 3JR, UK. 3Department of Biotechnology, Technische Universität Berlin, 13353 Berlin, Germany. 4Present addresses: Novo Nordisk Foundation Center for Biosustainability, DK-2970 Hørsholm, Denmark (P.M.); Institute of Medical Biology, University of Tromsø, 9037 Tromsø, Norway (J.M.). 5Correspondence should be addressed to J.R. or A.G. (e-mail: [email protected] or [email protected])

Received 12 October 2012; accepted 15 January 2014; published online 23 February 2014; DOI: 10.1038/ncb2918

RESOURCE abBiotin–dUTP cRelease from of new DNA thymidine Nascent block Biotin–dUTP chromatin

Strong chromatin 3 h 20 min crosslinking Mature 2 h chromatin 5 µm5 µm Sonication into d 2-3 kb fragment

Purification on streptavidin beads Collecting mature chromatin 5 µm 5 µm Labelling and collecting nascent chromatin Proteins Histone marks DNA G1 Thymidine block DAPI, biotin–dUTP and PCNA G2 e g PCNA H4 H1 90 DLSHIGDAVVISCAK DNIQGITKPAIR ASGPPVSELITK 75 100 531.61 100 442.59 447.27 100 559.84 z = 3 z = 2 60 z = 3 z = 3 603.84 z = 2 45 528.29 z = 3 30 15 0 0 0 Relative abundance m/z Light Heavy m/z Light Heavy m/z Light Heavy 0 (mature) (nascent) (mature) (nascent) (mature) (nascent)

Biotin–dUTP-positive cells (%) Early Mid Late S phase SILAC ratio N/M (log2) = 2.51 SILAC ratio N/M (log2) = 0.00 SILAC ratio N/M (log2) = –0.38 f NCC pulldown h Clamp and Fork stability Okazaki fragment

) 3 CMG DNA polymerases clamp loader complex processing Nascent Mature Control 2 PCNA 2 H4K12ac 1 Histone H4 NF POLE3 NF Histone H1 Histone H3 0 Histone H2B SILAC ratio N/M (log TIPIN RPA1 RPA2 RPA3 FEN1 ELG1 –1 RFC1 RFC2 RFC3 RFC4 RFC5 AND1 DNA2 DCC1 PCNA GINS1 GINS2 GINS3 GINS4 PRIM1 PRIM2 MCM2 MCM3 MCM4 MCM5 MCM6 MCM7 POLE1 POLE1 POLE4 CHTF1 CHTF8 POLA1 POLA2 POLD1 POLD2 POLD3 POLD4 CDC45 MCM10 Nascent Mature CLASPIN TIMELESS DNA LIGASE I

Figure 1 Isolation of replication forks and nascent chromatin by NCC with 300 cells counted in total. (f) NCC pulldown analysed by western technology. (a) Outline of the NCC protocol. See Methods for details. (b) blotting. Control: no addition of biotin–dUTP. Uncropped blots are provided Biotin-labelled newly synthesized DNA co-localizes with PCNA in replication in Supplementary Fig. 7. (g) Mass spectra showing the SILAC intensities foci. (c) Experimental design for comparison of nascent and mature for peptides of PCNA, H4 and H1. (h) Nascent chromatin enrichment of chromatin. Cells released from a single thymidine block were labelled with core replication fork components. Enrichments are given as the median, biotin–dUTP in early–mid S phase and collected directly (nascent) or after 2 from three independent experiments, of log2 of the SILAC ratios of nascent h chase without biotin–dUTP (mature). (d) Cell cycle proﬁles representative (heavy) over mature (light) (N/M). Error bars represent standard error of of experimental design in c.(e) S-phase distribution of cells labelled the mean (s.e.m.), and indicate the precision of ratio measurements for according to the experimental design in c. Cells were scored according to each protein. POLE1 and histone H1 are included as a reference in all their PCNA pattern50. Error bars represent standard deviation (s.d.), n = 4 enrichment plots. chromosomes13. The latter study revealed a surprising complexity insight into the process of building chromatin and maintaining of whole chromosomes and a machine learning approach, multi- epigenetic information. Moreover, our proteomic analysis provides classifier combinatorial proteomics (MCCP) was used to predict a discovery approach to identify proteins recruited to replication functional significance of associated proteins13. Given the power forks and/or nascent chromatin. Applying a MCCP-derived of these tools in defining chromosomes, we applied them to study chromatin-probability list to our data, we predict a role for chromatin replication. 93 uncharacterized proteins at replication forks or in newly We profile DNA replication and chromatin maturation in replicated chromatin behind the fork, and identify FAM111A as human cells by NCC, a biochemical approach to isolate newly a PCNA interaction partner required for DNA replication and synthesized DNA for quantitative proteomic analysis. By determining S-phase entry. the dynamic association of 3,995 proteins with nascent newly synthesized chromatin and mature post-replicative chromatin, RESULTS we reveal three classes of factors: enriched in nascent chromatin; Isolation of replication forks and nascent chromatin by NCC present in both nascent and mature chromatin; and enriched Mammalian replication and chromatin assembly factors are in mature chromatin. This rich data set oﬀers unprecedented commonly recognized by co-localization with newly synthesized

RESOURCE

SILAC ratio N/M (log2) purification of tagged chromatin (Supplementary Methods). Biotin– –10 1 2 dUTP can be introduced by a short hypotonic shift of HeLa S3 spinner POLE1 cultures without affecting S-phase progression or triggering DNA EXO1 MSH6 damage (Supplementary Fig. 1a,b). MSH2 RAD18 To follow DNA synthesis and its assembly into chromatin, MSH3 we compared nascent chromatin pulse-labelled for 20 min with RAD9A HUS1 post-replicative chromatin allowed to mature during a 2 h chase CHK2 UNG (Fig. 1c,d and Supplementary Fig. 1c). Given a fork speed of 1– BRCA1 2 kb min−1 (ref. 1), 20–40 kb of DNA could be labelled behind TP53BP1 RECQ5 each replication fork, in total corresponding to 5% of the genome. RAD50 NBS1 We thus isolate replication forks and nascent (immature) chromatin MRE11 BRCA2 behind, but shorter labelling times can be applied to focus on SMARCAL1 DNA replication (Supplementary Fig. 1d,e). Synchronized cells were RAD54B ATRIP labelled in early–mid S phase (Fig. 1d,e), including both euchromatin CHK1 MLH1 and heterochromatin regions in our analysis. Western blot of NCC MMS22L pulldowns confirmed enrichment of PCNA and new histone H4 TOPBP1 PMS2 acetylated at Lys 12 in nascent chromatin, whereas canonical histones TONSL RAD54A or L H3, H4 and H2B were present in both nascent and mature chromatin FANCJ (Fig. 1f). Histones and other factors were not detected in a negative APEX2 ERCC6 control without biotin–dUTP (Fig. 1f). MPG PMS1 To quantitatively compare the composition of nascent and mature RAD51 chromatin, we combined NCC and stable isotope labelling by amino GTF2H2 WHIP acids in cell culture16 (SILAC; Supplementary Fig. 2a–c). A low fold RAD17 DDB2 difference for most proteins between three independent biological CETN2 replicates illustrated high reproducibility and robustness of the BLM ATM technology (Supplementary Fig. 2d,e). We therefore present the full ERCC4 KU80 combined data set, normalized to histone H4 to adjust for variations SMC6 LIG4 in yield between independent experiments (Supplementary Table 1). KU70 We quantified in total 3,995 proteins providing a comprehensive RAD23B NEIL2 view of: factors enriched in nascent chromatin (for example, DNA-PK PRKDC PCNA); factors present in both nascent and mature chromatin (for MBD4 example, histone H4); and factors enriched in mature chromatin ERCC2 ERCC1 (for example, histone H1; Fig. 1g). Given that the SILAC ratio MUS81 ERCC5 generally reflects relative protein abundance, these groups typically ATR represent replication-associated factors (Fig. 1g, left), and early- RMI1 SMC5 (Fig. 1g, middle) or late-arriving chromatin components (Fig. 1g, XPC RECQ1 right). However, as crosslinking efficiency can be context dependent, NTHL1 extreme SILAC ratios may also indicate changes in the environment WRN MMS19 or conformation of a protein. Importantly, we identified 41 of the 44 GTF2H1 XRCC4 known core replication fork components enriched on newly replicated GTF2H3 DNA (Fig. 1h), arguing that we successfully isolated active human GTF2H4 Histone H1 replication forks. Mature Nascent Replication-associated repair Identification of accessory fork components Replication-independent repair We also identified a number of accessory factors dealing with DNA–RNA duplexes, protein degradation and DNA repair at the Figure 2 Genome maintenance factors at the fork. Replication-associated and fork (Fig. 2 and Supplementary Fig. 3a), whereas factors primarily replication-independent repair pathways annotated according to the literature are highlighted in red and blue, respectively. Enrichments are presented as involved in origin recognition and licensing were either not identified in Fig. 1h. or did not show strong enrichment (Supplementary Fig. 3b). Ranking replication-associated and replication-independent repair and checkpoint factors according to NCC–SILAC ratio illustrates that DNA labelled by nucleotide analogues. We have developed an the former group is highly over-represented in nascent chromatin analogous biochemical method, NCC, for isolation of protein (Fig. 2). This corroborates other reports that several genome caretakers complexes present on newly replicated DNA (Fig. 1a). This method serve functions and localize to sites of ongoing replication (for involves incorporation of biotin–dUTP (refs 14,15; Fig. 1b), example, MRN (ref. 17), SMARCAL1 (ref. 18) and TONSL–MMS22L strong chromatin crosslinking and sonication followed by affinity (ref. 19)).

RESOURCE

a Histonesb DNA methylationc Cohesion SILAC ratio N/M (log ) SILAC ratio N/M (log2) SILAC ratio N/M (log2) 2 –1 0 1 2 –1 0 1 2 –1 0 1 2 clamp loaders histones Alternative DNA methyl- POLE1

POLE1 transferase

H4 Core H3.1 / H3.3 CHTF8 H2B DNMT1 CHTF18 Deacetylation

H2A DNMT3A DCC1 Acetylation Cofactors variants H2A.x Histone DMAP1 ESCO2 HDAC8 MacroH2A1 UHRF1 UHRF2 H2A.Z/H2A.V SMC3 CenpA/CenpA-2 DNA methylation SMC1A binding factors and regulatory

MBD1 RAD21 Cohesin ring MBD2 histones H1.2 STAG1 factors H1.4 Linker MBD3 STAG2 H1.5 MBD4 PDS5A PDS5B H1x Kaiso ZBTB4 WAPAL H1.0 Sororin Mature Nascent Histone H1 Histone H1 Mature Nascent Mature Nascent

d Histone chaperones e H3K9 me3 regulators H3K27 me3 regulators

SILAC ratio N/M (log2) SILAC ratio N/M (log2) SILAC ratio N/M (log2) –5 –3 –1 0 1 2 –1012 –1012 POLE1 POLE1

POLE1 PRC2 H3.1–H4 G9a EZH2 CAF-1 P150 GLP Writer EED CAF-1 P60 SETDB1 SUZ12 Eraser CAF-1 P48 SUV39H1 KDM6B BMI1

H3–H4 SUV39H2 ASF1a Eraser RING1A ASF1b PHF8 LSD1 RING1B NASP CBX2 SPT6 JMJD1A

H3.3–H4 α CBX4

HP1 Reader ATRX HP1β CBX8 DAXX HP1γ PHC2 HIRA NF CDYL PHC3 NF MEL18

CDYL2 PRC1 FACT SSRP1 SCMH1 H2A–H2B Histone H1 FACT SPT16 NSPC1 NPM1 Mature Nascent CBX6 NPM3 Histone H1 NAP1 Nucleolin Mature Nascent NAP2 Histone H1 Mature Nascent

Figure 3 Chromatin assembly and maturation. (a) Nascent chromatin regulatory factors. (d) Nascent chromatin enrichment of histone chaperones. enrichment of canonical and replacement histones. (b) Nascent chromatin (e) Nascent chromatin enrichment of H3K9 (left) and H3K27 (right) enrichment of DNA methyl transferases, cofactors and methyl-binding methylation writers, erasers and readers. Only core components of PRC1 and proteins. (c) Nascent chromatin enrichment of the cohesin complex and PRC2 are shown. All enrichments are presented as in Fig. 1h.

Several accessory factors are recruited through PCNA (for example, replication. Although less sensitive, western blotting of selected factors RNAse H2 (ref. 20), EXO1 (ref. 21) and MSH2/3/6 (ref. 22)), and this generally mirrored SILAC enrichments (Supplementary Fig. 3d). prompted us to evaluate the coverage of known PCNA interactors in As expected, we found chaperones such as CAF-1 and FACT, our analysis23. Whereas DNA replication and chromatin restoration DNMT1 with cofactors UHRF1 and DMAP1, and the cohesin factors are comprehensively represented and enriched in nascent acetyltransferase ESCO2 enriched in nascent chromatin (Fig. 3b–d). chromatin, many factors involved in cell cycle control and survival The Asf1a and Asf1b histone chaperones also play important are not identified, and those we find are not enriched (Supplementary roles during replication24,25, yet they were not enriched in nascent Fig. 3c and Table 2). Furthermore, factors carrying a PIP-degron, chromatin (Fig. 3d). Likewise, previous studies could not identify these which undergo PCNA-dependent degradation in S phase, were not chaperones at replication sites by immunofluorescence microscopy25. found. The NCC data thus identify those factors from the large PCNA This probably reflects their dynamic behaviour and additional interactome that function at active replication forks. replication-independent functions. Unexpectedly, the centromeric H3 variant CENP-A was highly enriched in nascent chromatin (Fig. 3a). Chromatin assembly and maturation However, we did not identify the CENP-A histone chaperone HJURP. We quantified 806 known chromatin components and transcriptional Other interesting factors enriched in nascent chromatin include regulators in our analysis (Supplementary Table 1 and Fig. 3), the H3K9me3 reader and writer HP1 and SUV39h1/2 (Fig. 3e), providing insight into their dynamic behaviour during chromatin the NASP histone chaperone (Fig. 3d), the H3K4me3 demethylase

RESOURCE

a Nascent Mature Control b 1× 2× 1× 2× 1× 2× 15 min NCC pulldown H4K5ac biotin–dUTP Soluble Nascent –+CHX (2.5 min) – +CHX H3K9me1 Nascent H4K5ac H4K5ac S-phase β H3K9me3 H3K27me1 cells -actin Nascent CHX Histone H4 H3K9me3 H4K5ac H3K27me3 CHX (2.5 min) H3K27me3 Histone H4 Histone H3

Figure 4 Histone mark dynamics during chromatin maturation. (a) Histone in the presence of CHX to prevent new histone deposition. Left, experimental marks analysed by western blotting of NCC pulldowns of nascent and mature design. Middle, western blot of new histone H4 acetylated at Lys 5 in the soluble chromatin. 2x indicates double the amount of sample loaded in 1x. Antibody fraction before biotin–dUTP labelling. Right, western blot of histone marks after speciﬁcity was veriﬁed25. (b) Histone marks in nascent chromatin replicated NCC pulldown. Uncropped blots are provided in Supplementary Fig. 7.

KDM5D/C (Supplementary Fig. 3e), the deacetylases HDAC8 and first 2 h after replication (Fig. 4a). H3K9me1 levels were increased in SIR2L, components of the NuA4 (YEATS4, MORF4L2, and BRD8) mature chromatin, probably reflecting that a fraction of new histones and CREBBP acetyltransferases, RIF1 and several nuclear lamins are mono-methylated during this time window. SILAC analysis of new (lamin-A/C, lamin B1, B2; Supplementary Table 1). histones has shown that establishment of H3K27me3 and H3K9me3 The canonical histones showed a ratio close to 0 (Fig. 3a), consistent is slow, whereas mono-methylation is more rapid27,28. Together with with their rapid assembly onto newly synthesized DNA and thus a our NCC data, this predicts that H3K9me3 and H3K27me3 marks similar occupancy on nascent and mature chromatin. This was also detected in nascent and mature chromatin originate from recycled old true for H2A.X and MacroH2A1 (Fig. 3a), probably reflecting that histones. To test this directly, we modified our NCC set-up to label they are transferred during replication and additional incorporation DNA under conditions where new histone biosynthesis is repressed does not take place within 2 h of maturation. A large number of other by cycloheximide (CHX). Short pre-treatment with CHX depleted the components also showed similar occupancy in nascent and mature soluble pool of new histones marked by H4K5ac (Fig. 4b, left), thereby chromatin, including cohesin rings (Fig. 3c), CTCF, the chromatin reducing their incorporation into nascent chromatin (Fig. 4b, right). remodellers SMARCA1, 4 and 5, HDAC1 and 2, and the PRC2 However, H3K9me3 and H3K27me3 levels were not aﬀected by the complex (Fig. 3e and Supplementary Table 1). This probably reflects lack of new histone deposition (Fig. 4b, right), corroborating the long- that these components, like canonical histones, are only transiently standing model that histone marks are transferred with old histones released during replication. onto newly synthesized DNA (refs 2,3). Most linker histone H1 isoforms and the replacement variant H2A.Z were enriched in mature chromatin (Fig. 3a). This implies Predicting replication-coupled functions for uncharacterized that incorporation of these histones is separated from replication proteins by in silico puriﬁcation (replication-independent), but commences within 2 h after DNA Our NCC–SILAC experiments quantified 3,995 factors, manually synthesis. Several methyltransferases catalysing active methylation distributed into seven categories based on literature (Fig. 5a). marks such as H3K4me3 and H3K36me2 were also enriched Overall, more factors were enriched in nascent chromatin compared in mature chromatin (for example MMSET, ASH1, MLL1 and with mature, possibly owing to its open and highly acetylated MLL2; Supplementary Fig. 3e)26, correlating with RNA Pol II state. Importantly, DNA replication stands out as the most highly association (Supplementary Fig. 3f). Strikingly, the de novo enriched category in nascent chromatin, suggesting a replication- DNA methyltransferase DNMT3A and the PRC1 complex also coupled function of proteins that are similarly enriched (Fig. 5b and preferentially occupied mature chromatin (Fig. 3b,e). Most methyl- Supplementary Fig. 4). We found 878 functionally uncharacterized binding proteins behaved like DNMT3 (Fig. 3b), as expected if their proteins as well as many factors with no expected chromatin function recruitment awaits restoration of methylation by DNMT1–UHRF1. that for the most part are likely to be ‘hitchhikers’, considered Thus, late recruitment characterizes a broad group of factors involved biological/biochemical background13. Unexpectedly, some hitchhikers in both transcriptional activation (MMSET, MLL1, H2A.Z) as well as showed enrichment in nascent or mature chromatin (Supplementary repression (H1, PRC2, DNMT3). Fig. 4), limiting the predictive value for functions of uncharacterized proteins. To add a second dimension to the data, so that functionally H3K9me3 and H3K27me3 are transferred with old histones relevant proteins can be distinguished from background, we used To explore the relationship between histone marks and the dynamics a chromatin probability table for 7,600 proteins29. This is based on of the modifying enzymes, we followed H3K9me3 and H3K27me3 MCCP (ref. 13), integrating many chromatin proteomics experiments levels. In contrast to the rapid loss of H4K5K12diAc, neither unrelated to the present study but collectively indicating chromatin H3K27me3 nor H3K9me3 levels changed significantly within the function. Applying chromatin probability as an in silico purification

RESOURCE

a DNA replication 59

DNA repair 125

Transcription 461

Other chromatin 345

RNA processing 282

No expected chromatin function 1,845

Uncharacterized 878

Total 3,995

b 1× 1012 c 1.0

0.8 1× 1010

0.6

1× 108 0.4 Total protein intensity 1× 106 Chromatin probability 0.2

1× 104 0 –6–4 –2 0 2 4 6 –6–4 –2 0 2 4 6 Nascent chromatin enrichment Nascent chromatin enrichment

(log2 SILAC ratio N/M) (log2 SILAC ratio N/M)

d Nascent chromatin (log2 > 0.4) Filtered by Unfiltered Protein class chromatin probability

43 DNA replication 37

52 DNA repair 36

127 Transcription 74

96 Other chromatin 71

79 RNA processing 55

613 No expected chromatin function 60 286 Uncharacterized 93

1,296 Total 426

Figure 5 Identification of uncharacterized proteins with a predicted function probability against nascent chromatin enrichment. (d) Barrel plot illustrating in chromatin replication. (a) Pie diagram illustrating the distribution of all the distribution of proteins with a SILAC log2 ratio <0.4 before (left) and after proteins identified in NCC–SILAC into 7 functional classes (Supplementary (right) using chromatin probability to separate chromatin from non-chromatin Table 1). (b) Volcano plot showing the logarithmic ratio of protein intensities in proteins (Supplementary Table 3). Factors with a chromatin probability above the combined data set plotted against their median logarithmic ratio of heavy 0.4 were considered chromatin, as this included roughly 90% of the core versus light peptides (nascent chromatin enrichment). (c) Plot of chromatin replication factors (Fig. 1h). step to our data set allowed us to discriminate factors with a chromatin a function at replication forks or on newly replicated chromatin function from those with no expected chromatin function according behind the fork (Supplementary Table 3). to literature (Fig. 5c and Supplementary Table 1). Focusing on factors with nascent chromatin enrichment larger than the 40 most Identification of chromatin replication factors enriched core replication factors (SILAC log2 ratio < 0.4), in silico First, we tested the localization of seven uncharacterized proteins purification removed 90% of the factors with no expected chromatin enriched on nascent chromatin, three with high chromatin probability function, but keeping 90% of the replication factors. A large number and four with low probability29. All three factors with a highly likely of uncharacterized proteins (60%) were also removed, ultimately chromatin function, FAM111A, FAM178A and ATAD2B, showed identifying 93 uncharacterized factors with high probability of having nuclear localization, exhibited a focal staining pattern and were

RESOURCE

a Pre-XTag PCNA Merge DAPI b c

Nuclear Cytoplasmic Both 1.0 – 100 0.8 80

FAM111A + 0.6 60 0.4 40 – 0.2 Distribution (%) 20 (Pearson coefficient)

0 Co-localization with PCNA 0 FAM111A FAM178A ATAD2B FAM111A FAM178A ATAD2B FAM178A + NCC enrichment 2.27 1.46 0.88

– d e Nuclear Cytoplasmic Both 1.0 ATAD2B + 100 80 0.5 60 – 40 0 20 Distribution (%) PTCD3 + (Pearson coefficient) 0 –0.5 Co-localization with PCNA PTCD3 RPL29 KIAA0391 MPST NCC PTCD3 RPL29 KIAA0391 MPST enrichment 0.66 0.62 0.57 0.51 –

f g RPL29 + GST pulldown NCC pulldown ++– – GST Input – ++– GST–PCNA Nascent Mature + – +–+ – FAM111A – FAM111A – + – – + + PIPmt 35S PCNA KIAA0391 + GST–PCNA GST Histone H3

– 0.2

0.1 MPST FAM111A

+ bound/input 0

10 µm

Figure 6 Localization of uncharacterized proteins with a predicted function Horizontal lines represent the median. (f) Western blot of NCC pulldowns. in chromatin replication. (a) Analysis of GFP- or FLAG-tagged proteins in (g) In vitro analysis of PCNA binding using GST–PCNA to pull down in vitro- U-2-OS cells stably expressing RFP–PCNA. Cells were fixed directly or after translated 35S-labelled FAM111A wild type or PIP box mutant (PIPmt). The pre-extraction with Triton (Pre-X). (b–e) Localization pattern (b,d) and co- diagram shows the quantification of bound FAM111A relative to input with localization with PCNA measured by Pearson coefficient in individual nuclei error bars representing the s.d. (n = 3). Uncropped blots are provided in were scored after direct fixation (c,e) or Pre-X (Supplementary Fig. 5a–d). Supplementary Fig. 7. resistant to Triton extraction (Fig. 6a,b and Supplementary Fig. 5a), as FAM178A APIM motif did not affect its localization (Supplementary expected for chromatin-bound proteins. In contrast, the four proteins Fig. 6b), disruption of the FAM111A PIP box led to diffuse pan-nuclear with low chromatin probability were mainly cytoplasmic (Fig. 6a,d distribution (see below). We thus focused our attention on FAM111A and Supplementary Fig. 5c). FAM111A and FAM178A co-localized and first confirmed the enrichment in nascent chromatin by western almost perfectly with PCNA, whereas ATAD2B showed partial co- blotting (Fig. 6f). Then we addressed PCNA binding by GST pulldown localization consistent with the lower NCC enrichment (Fig. 6c and and found that FAM111A interacts directly with PCNA in a PIP-box- Supplementary Fig. 5b). We noted that both FAM111A and FAM178A dependent manner in vitro (Fig. 6g). contained putative PCNA interaction motifs, a PIP box and an APIM Green fluorescent protein (GFP)-tagged FAM111A co-localized motive, respectively (Supplementary Fig. 6a). Whereas mutation of the with PCNA at replication sites, but at the same time impaired EdU

RESOURCE

a GFP PCNA Merge EdU 97 +/– 4% 64 +/– 15% b c d G1Early S Mid–late S G2 100 80 150 80 60 100 60 40 40 PIPmt FAM111A

100 +/– 0% 50 20 20 EdU intensity (a.u.) PCNA-positive cells (%) Cell cycle distribution (%) H2B 0 0 0 FAM111A PIPmt H2B FAM111A PIPmt H2B FAM111A PIPmt H2B 10 μm e GFP PCNA DAPI f g ∗∗∗∗ NS ∗∗∗∗ 250 15 ∗∗

H2B ∗∗∗∗ 200 ∗∗∗∗ TIG-3 cells 10 150 WT 100 5 PCNA intensity (a.u.) 50 Control siRNA FAM111A siRNA1 FAM111A siRNA2

PCNA intensity (a.u.) 2× 1× 2× 1× 2× 1× PIPmt ∗ 0 0 20 μm FAM111A PIPmt H2B MG132 –+–+–+–+ H2B PIPmt H2B PIPmt Total PCNA Chromatin bound PCNA h i ∗∗∗∗ j k ∗∗∗∗ ∗∗∗∗ 40 1,000 ∗∗∗∗ 100 Asynchronous 10 h nocodazole Control siRNA FAM111A siRNA1 800 30 FAM111A siRNA2 75

600 20 50 400 10 25 EdU intensity in PCNA intensity in

200 EdU-positive cells (%)

EdU-positive cells (a.u.) 0 PCNA-positive cells (a.u.) 0 0 0 1418202224 Control FAM111A FAM111A Control FAM111A FAM111A Control FAM111A FAM111A Time after release in serum (h) siRNA siRNA1 siRNA2 siRNA siRNA1 siRNA2 siRNA siRNA1 siRNA2

Figure 7 FAM111A facilitates PCNA loading and S-phase entry. (a) Analysis treated or not with MG132. One representative experiment is shown. of PCNA and EdU in Triton-extracted U-2-OS cells. Localization patterns were Horizontal lines represent the median; **** P > 0.0001 (unpaired t-test, quantified in transfected cells for 3 independent experiments. GFP–H2B was n > 77). (g–k) FAM111A depletion in TIG-3 cells transfected with two used as control. (b) Quantification of EdU intensities in transfected cells. The independent siRNAs for 48 h. (g) Western blot. 2x indicates double the mean of 3 independent experiments with s.d. is shown, <200 cells scored amount of sample loaded in 1x. * unspecific band. Uncropped blots are per experiment. (c) Quantification of PCNA-positive cells. Only transfected provided in Supplementary Fig. 7. (h,i) Quantification of EdU incorporation cells were scored. Error bars represent s.d. (n=3). The distribution of PCNA (h) and chromatin-bound PCNA (i). Dot plots show EdU intensities in PCNA- patterns is shown in Supplementary Fig. 6c. (d) Cell cycle distribution based positive cells and PCNA intensities in EdU-positive cells. Similar results were on combined analysis of MCM2 pattern and EdU incorporation in transfected obtained in total cell populations and in U-2-OS cells (Supplementary Fig. cells. Supplementary Fig. 6d shows representative patterns. Error bars 6g–i). One representative experiment is shown. Horizontal lines represent represent s.d. (n = 3), 515 cells were counted in total. (e) Analysis of total the median. **** P > 0.0001 (unpaired t-test, n > 108). (j) Cell cycle PCNA in transfected cells fixed without pre-extraction. Left, representative profiles of cells treated without (left) or with (right) nocodazole for 10 h. (k) images; arrowheads mark transfected cells. Right, quantification; the mean S-phase entry scored by EdU incorporation after release from quiescence of three independent experiments with s.d. is shown. ** P =0.022; NS, not by re-stimulation with serum. One representative experiment is shown, significant, P =0.13 (one-sample t-test); <200 cells scored per experiment. <100 cells scored per time point. Statistics source data are available in (f) Quantification of total and chromatin-bound PCNA in transfected cells Supplementary Table 4. incorporation and arrested cells in early S phase (Fig. 7a,b and (Fig. 7a,c). This suggested that the mutant arrested cells outside Supplementary Fig. 6c). This probably reflects that PIP box-containing S-phase. We thus determined the cell cycle distribution based on proteins on overexpression can interfere with DNA replication by MCM2 staining patterns and EdU labelling for 1 h to detect slowly competing with other PCNA binding partners (that is, DNA pol replicating cells33 (Supplementary Fig. 6d). Intriguingly, a substantial δ, DNA pol ε, FEN-1 and DNA ligase I)30–32. The FAM111A PIP proportion of cells expressing the PIP mutant were in early S phase and box mutant (PIPmt) showed a pan-nuclear distribution (Fig. 7a), incorporated low levels of EdU (Fig. 7d and Supplementary Fig. 6e), and, surprisingly, these cells were all negative for PCNA and EdU regardless of the fact that PCNA foci were hardly detectable (Fig. 7a,c

RESOURCE

H2A.Z Cohesin

New H3-H4

Replisome

Parental H3-H4 Histone H1

CAF-1, DNMT1, HP1, SUV39h, ESCO2, RIF1 Replication-coupled (426/93 unknown) Lamins, CBP, KDM5D/C

H3, H4, H2A, H2B, Cohesin, PRC2, CTCF, HDAC1/2 Early stable (914/236 unknown)

H1, H2A.Z, NAP1, DNMT3a PRC1, ASH1, MMSET Delayed (135/32 unknown)

Chromatin factors Association dynamics

Figure 8 Chromatin dynamics during DNA replication. Association dynamics probability (Supplementary Table 1). Selected factors are listed along with of chromatin proteins during the first two hours of chromatin assembly the total number of chromatin proteins identified and those functionally and maturation determined by NCC–SILAC enrichment and chromatin uncharacterized. and Supplementary Fig. 6e). This suggested that the PIP mutant chromatin maturation, and predict a replication-coupled function might affect PCNA loading and/or stability in a dominant-negative for 93 uncharacterized proteins. Our analysis reveals that replication manner. Quantitative analysis showed that the PIP mutant reduced markedly perturbs chromatin composition, with 561 chromatin the level of both chromatin-bound and total PCNA in the nucleus factors changing association within the first hours post-replication (Fig. 7e,f). Similar results were obtained in cells stably expressing (Fig. 8). This correlates with deacetylation of new histones, but red fluorescent protein (RFP)-tagged PCNA from an exogenous surprisingly repressive marks such as H3K9me3 and H3K27me3 promoter (Supplementary Fig. 6f), excluding indirect effects on remain largely unchanged. Overall, this large quantitative analysis transcription. We thus tested whether PCNA stability was affected provides a unique resource to understand epigenome maintenance in by inhibiting the proteasome with MG132. This largely restored total proliferating cells. PCNA levels, but only moderately increased loading of PCNA on Two other methods for isolation of proteins on new DNA, chromatin (Fig. 7f). iPOND and Dm-ChP, were recently described34,35. These rely on Next we used RNA-mediated interference (RNAi) depletion to EdU labelling of DNA and chemical ligation of biotin post-fixation address directly whether FAM111A function is required for DNA (Click-iT). Introduction of biotin–dUTP requires a short hypotonic replication. Depletion of FAM111A significantly impaired EdU shift, but circumvents chemical ligation of biotin. Furthermore, NCC incorporation (Fig. 7g,h and Supplementary Fig. 6g,h) and this is developed for large-scale proteomics and thus optimized with was accompanied by reduced loading of PCNA onto chromatin respect to: strong crosslinking to preserve larger chromatin-bound (Fig. 7i and Supplementary Fig. 6i) in both primary TIG-3 primary complexes; isolation of nuclei to reduce contaminants; and use of fibroblasts and U-2-OS osteosarcoma cells. The cell cycle was not HeLa S3 spinner cultures and SILAC to quantify chromatin dynamics. markedly perturbed after 48 h of FAM111A short interfering RNA Our analysis captures the first well-established steps in chromatin (siRNA) treatment (Fig. 7j). However, treatment with nocodazole replication: DNA unwinding (CMG helicase), DNA synthesis (Pol to trap cells in mitosis showed that progression from G1 through α, β and ε), Okazaki fragment processing (DNA ligase I, FEN1), S-phase to G2/M was markedly delayed in the absence of FAM111A nucleosome assembly (CAF-1), maintenance DNA methylation (Fig. 7j). To address G1/S transition directly, we released siRNA- (DNMT1–UHRF1) and establishment of sister chromatid cohesion treated TIG-3 cells from quiescence and followed S-phase entry by (Esco2). The comprehensive identification of known fork components EdU incorporation. This revealed that FAM111A-depleted cells enter by NCC probably reflects our chromatin preparation method along S phase and start DNA replication substantially later than control with SILAC-based quantification of chromatin composition at two cells (Fig. 7k). distinct maturation states. NCC–SILAC should thus be a perfectly controlled approach DISCUSSION to discover DNA replication and chromatin maturation factors. Here, we describe a strategy for profiling dynamics of the chromatin Surprisingly, many proteins enriched in nascent chromatin had well- proteome and histone marks during replication. We provide established functions elsewhere in the cell. Whereas some may association dynamics for 3,995 factors during DNA replication and function in replication, others probably constitute background. This

RESOURCE is possible if the functional and structural differences between nascent follow new and old histones27,28. We took advantage of short CHX and mature chromatin also impact on the background, in which treatment as a crude, but instant, approach to block new histone case there is no perfect control. To pick the best candidates for biosynthesis8,48. Our data suggest that H3K27me3 and H3K9me3 follow-up studies, we thus took advantage of a chromatin-probability in nascent chromatin largely reflect transmission with old histones, table for 7,600 proteins established by proteomics of interphase confirming present models of epigenetic memory2,3. A recent study chromatin29. By this means, we shortlist 93 uncharacterized proteins in Drosophila embryos could not detect H3K4me3 and H3K27me3 that with high probability have a function at replication forks or on newly synthesized DNA and proposed that these marks were not in nascent chromatin. Corroborating this resource, three of the top inherited with parental histone49. A transmission of histone marks candidates exclusively (FAM111A, FAM178A) or in part (ATAD2B) may differ in Drosophila embryos and human cells, it is important to localize to replication sites. We thus anticipate that FAM111A and note that our results are entirely consistent with recent SILAC-based FAM178A primarily function in chromatin replication, whereas comparison of new and old histone in cycling human cells27,28. ATAD2B serve additional roles in chromatin. Indeed, FAM111A Together these complementary studies support a model in which interacts directly with PCNA and promotes S-phase entry and DNA H3K9me3 and H3K27me3 are transferred with old histones and synthesis. A direct function of FAM111A at replication forks might imposed on new histones with slow kinetics, while mono-methylation explain recent findings that SV40 large T antigen binds FAM111A is more rapid. to overcome host-range restriction36 and that variations in the NCC combined with quantitative mass spectrometry is a strong chromosomal loci harbouring FAM111A (11q12) confer susceptibility tool to address how the epigenetic framework is maintained in to prostate cancer37. We speculate that FAM111A could play a role dividing cells. To understand cellular memory it is evident that in replication initiation by affecting PCNA loading. Overexpression we need to unravel how chromatin is replicated. The resource and of the FAM111A PIP mutant compromises PCNA stability and technology we provide here should thus be relevant to questions chromatin binding. Yet, stabilization of PCNA did not fully restore about cell identity during development, proliferative exhaustion of chromatin binding (Fig. 7f), and overexpression of PCNA together ageing cells and epigenetic alterations in cancer cells that experience with the PIP mutant could not rescue replication (data not shown). uncontrolled proliferation. Given the wide use of chemotherapeutic The FAM111A PIP mutant might thus interfere with PCNA loading drugs that target replication, NCC could also provide an avenue to or maintenance on chromatin in a dominant-negative manner. drug design. Consistent with this, cells lacking FAM111A show low levels of ACKNOWLEDGEMENTS chromatin-bound PCNA and impaired replication. We would like to thank J. Déjardin for fruitful discussions, K. Helin for reagents Our study defines the dynamic changes in chromatin composition (Biotech Research and Innovation Centre and Centre for Epigenetics, University during replication (Fig. 8). The challenge is now to identify recruit of Copenhagen, Denmark), Ib J. Christensen for support on statistics, and C. Wu, W. C. Earnshaw and Z. Jasencakova for critical reading of this manuscript. The ment mechanisms and uncover functional dependencies. Our finding A.G. laboratory is supported by a European Research Council Starting Grant that histone H1 is recruited in a delayed manner supports the model (ERC2011StG, no. 281,765), the Lundbeck Foundation, the Danish Cancer Society, that deacetylation of new histones pave the way for linker histone the Danish National Research Foundation (DNRF82) and Medical Research 6,38 39 Council, the Novo Nordisk Foundation and FP7 Marie Curie Actions ITN binding and compaction . PRC1 also promotes compaction and Nucleosome4D. The Wellcome Trust generously supported this work through a together with PRC2 constitutes an important cellular memory system3. Senior Research Fellowship to J.R. (084229), two Wellcome Trust Centre Core An in vitro study of SV40 replication proposed that PRC1 remains Grants (077707, 092076) and an instrument grant (091020). C.A. was supported 40 by fellowships from HFSP and the Danish Medical Research Council. G.K. was bound to chromatin during DNA replication . Our work argues supported by a FEBS Long-Term fellowship. that a substantial part of PRC1 is recruited later during chromatin maturation. In contrast the PRC2 complex is present in both nascent AUTHOR CONTRIBUTIONS A.G., C.A. and J.R. conceived the project and designed the study. C.A., S.L., K.N. and mature chromatin, consistent with rapid recruitment to nascent and J.M. performed and analysed experiments. J.B., G.K., F.A. and J.R. performed chromatin by parental histones carrying H3K27me3 (refs 41,42). As mass spectrometry and analysed SILAC data. P.M. generated F.A.M. mutants. The H3K27me3 levels did not change, other features must contribute to manuscript was written by C.A. and A.G. and edited by J.R., J.B. and G.K. delayed recruitment of PRC1. COMPETING FINANCIAL INTERESTS The centromeric H3 variant CENP-A was enriched in nascent The authors declare no competing financial interests. chromatin (Fig. 3a), despite the fact that stable de novo CENP-A Published online at www.nature.com/doifinder/10.1038/ncb2918 incorporation mainly occurs in early G1 (refs 43,44). We suggest Reprints and permissions information is available online at www.nature.com/reprints that CENP-A either associates spuriously to accessible genomic regions such as nascent chromatin (this study) and sites of DNA 1. Mechali, M. Eukaryotic DNA replication origins: many choices for appropriate repair45, or that structural changes in CENP-A nucleosomes after answers. Nat. Rev. Mol. Cell Biol. 11, 728–738 (2010). 2. Alabert, C. & Groth, A. Chromatin replication and epigenome maintenance. Nat. Rev. 46 replication influence crosslinking efficiency. The histone variant Mol. Cell Biol. 13, 153–167 (2012). H2A.Z also decorates centric chromatin in addition to pericentric 3. Margueron, R. & Reinberg, D. Chromatin structure and the inheritance of epigenetic 43 information. Nat. Rev. Genet. 11, 285–296 (2010). regions, promoters and gene bodies . In contrast to CENP-A, H2A.Z 4. Baylin, S. B. & Jones, P. A. A decade of exploring the cancer epigenome - biological levels increase with chromatin maturation. This probably reflects and translational implications. Nat. Rev. Cancer 11, 726–734 (2011). 47 5. Worcel, A., Han, S. & Wong, M. L. Assembly of newly replicated chromatin. Cell 15, dilution of H2A.Z during replication , followed by replication- 969–977 (1978). independent incorporation of new H2A.Z (ref. 43). 6. Annunziato, A. T. & Seale, R. L. Histone deacetylation is required for the maturation of newly replicated chromatin. J. Biol. Chem. 258, 12675–12684 (1983). NCC provides a tool to monitor transmission of histone 7. DePamphilis, M. L. & Wassarman, P. M. Replication of eukaryotic chromosomes: a modifications that complements present SILAC-based approaches to close-up of the replication fork. Annu. Rev. Biochem. 49, 627–666 (1980).

RESOURCE

8. Annunziato, A. T. Assembling chromatin: The long and winding road. Biochim. 31. Li, R., Waga, S., Hannon, G. J., Beach, D. & Stillman, B. Differential effects by Biophys. Acta 1819, 196–210 (2012). the p21 CDK inhibitor on PCNA-dependent DNA replication and repair. Nature 371, 9. Sobel, R. E., Cook, R. G., Perry, C. A., Annunziato, A. T. & Allis, C. D. Conservation of 534–537 (1994). deposition-related acetylation sites in newly synthesized histones H3 and H4. Proc. 32. Warbrick, E., Lane, D. P., Glover, D. M. & Cox, L. S. A small peptide inhibitor Natl Acad. Sci. USA 92, 1237–1241 (1995). of DNA replication defines the site of interaction between the cyclin-dependent 10. Sirbu, B. M. et al. Analysis of protein dynamics at active, stalled, and collapsed kinase inhibitor p21WAF1 and proliferating cell nuclear antigen. Curr. Biol. 5, replication forks. Genes Dev. 25, 1320–1327 (2011). 275–282 (1995). 11. Taddei, A., Roche, D., Sibarita, J. B., Turner, B. M. & Almouzni, G. Duplication and 33. Masata, M., Juda, P., Raska, O., Cardoso, M. C. & Raska, I. A fraction of MCM 2 maintenance of heterochromatin domains. J. Cell Biol. 147, 1153–1166 (1999). proteins remain associated with replication foci during a major part of S phase. Folia 12. Dejardin, J. & Kingston, R. E. Purification of proteins associated with specific Biol. 57, 3–11 (2011). genomic Loci. Cell 136, 175–186 (2009). 34. Sirbu, B. M., Couch, F. B. & Cortez, D. Monitoring the spatiotemporal dynamics of 13. Ohta, S. et al. The protein composition of mitotic chromosomes determined using proteins at replication forks and in assembled chromatin using isolation of proteins multiclassifier combinatorial proteomics. Cell 142, 810–821 (2010). on nascent DNA. Nat. Protoc. 7, 594–605 (2012). 14. Maya-Mendoza, A., Olivares-Chauvet, P., Kohlmeier, F. & Jackson, D. A. Visualising 35. Kliszczak, A. E., Rainey, M. D., Harhen, B., Boisvert, F. M. & Santocanale, C. DNA chromosomal replication sites and replicons in mammalian cells. Methods 2, mediated chromatin pull-down for the study of chromatin replication. Sci. Rep. 1, 140–148 (2012). 95 (2011). 15. Koberna, K. et al. Nuclear organization studied with the help of a hypotonic shift: 36. Fine, D. A. et al. Identification of FAM111A as an SV40 host range restriction and its use permits hydrophilic molecules to enter into living cells. Chromosoma 108, adenovirus helper factor. PLoS Pathog. 8, e1002949 (2012). 325–335 (1999). 37. Akamatsu, S. et al. Common variants at 11q12, 10q26 and 3p11.2 are 16. Ong, S. E. et al. Stable isotope labeling by amino acids in cell culture, SILAC, associated with prostate cancer susceptibility in Japanese. Nat. Genet. 44, as a simple and accurate approach to expression proteomics. Mol. Cell Proteom. 1, 426–429 (2012). 376–386 (2002). 38. Perry, C. A. & Annunziato, A. T. Histone acetylation reduces H1-mediated 17. Maser, R. S. et al. Mre11 complex and DNA replication: linkage to E2F and sites of nucleosome interactions during chromatin assembly. Exp. Cell Res. 196, DNA synthesis. Mol. Cell Biol. 21, 6006–6016 (2001). 337–345 (1991). 18. Betous, R. et al. SMARCAL1 catalyzes fork regression and Holliday junction 39. Francis, N. J., Kingston, R. E. & Woodcock, C. L. Chromatin compaction by a migration to maintain genome stability during DNA replication. Genes Dev. 26, polycomb group protein complex. Science 306, 1574–1577 (2004). 151–162 (2012). 40. Francis, N. J., Follmer, N. E., Simon, M. D., Aghia, G. & Butler, J. D. Polycomb 19. Duro, E. et al. Identification of the MMS22L-TONSL complex that promotes proteins remain bound to chromatin and DNA during DNA replication in vitro. Cell homologous recombination. Mol. Cell 40, 632–644 (2010). 137, 110–122 (2009). 20. Bubeck, D. et al. PCNA directs type 2 RNase H activity on DNA replication and 41. Hansen, K. H. et al. A model for transmission of the H3K27me3 epigenetic mark. repair substrates. Nucl. Acids Res. 39, 3652–3666 (2011). Nat. Cell Biol. 10, 1291–1300 (2008). 21. Nielsen, F. C., Jager, A. C., Lutzen, A., Bundgaard, J. R. & Rasmussen, L. 42. Margueron, R. et al. Role of the polycomb protein EED in the propagation of J. Characterization of human exonuclease 1 in complex with mismatch repair repressive histone marks. Nature 461, 762–767 (2009). proteins, subcellular localization and association with PCNA. Oncogene 23, 43. Boyarchuk, E., Montes de Oca, R. & Almouzni, G. Cell cycle dynamics of histone 1457–1468 (2004). variants at the centromere, a model for chromosomal landmarks. Curr. Opin. Cell 22. Li, G. M. Mechanisms and functions of DNA mismatch repair. Cell Res. 18, 85–98 Biol. 23, 266–276 (2011). (2008). 44. Jansen, L. E., Black, B. E., Foltz, D. R. & Cleveland, D. W. Propagation of centromeric 23. Moldovan, G. L., Pfander, B. & Jentsch, S. PCNA, the maestro of the replication fork. chromatin requires exit from mitosis. J. Cell Biol. 176, 795–805 (2007). Cell 129, 665–679 (2007). 45. Zeitlin, S. G. et al. Double-strand DNA breaks recruit the centromeric histone CENP- 24. Groth, A. et al. Regulation of replication fork progression through histone supply and A. Proc. Natl Acad. Sci. USA 106, 15762–15767 (2009). demand. Science 318, 1928–1931 (2007). 46. Bui, M. et al. Cell-cycle-dependent structural transitions in the human CENP-A 25. Jasencakova, Z. et al. Replication stress interferes with histone recycling and nucleosome in vivo. Cell 150, 317–326 (2012). predeposition marking of new histones. Mol. Cell 37, 736–743 (2010). 47. Nekrasov, M. et al. Histone H2A.Z inheritance during the cell cycle and its 26. Albert, M. & Helin, K. Histone methyltransferases in cancer. Semin. Cell Dev. Biol. impact on promoter organization and dynamics. Nat. Struct. Mol. Biol. 19, 21, 209–220 (2010). 1076–1083 (2012). 27. Scharf, A. N., Barth, T. K. & Imhof, A. Establishment of histone modifications after 48. Sariban, E., Wu, R. S., Erickson, L. C. & Bonner, W. M. Interrelationships of chromatin assembly. Nucl. Acids Res. 37, 5032–5040 (2009). protein and DNA syntheses during replication of mammalian cells. Mol. Cell Biol. 5, 28. Xu, M., Wang, W., Chen, S. & Zhu, B. A model for mitotic inheritance of histone 1279–1286 (1985). lysine methylation. EMBO Rep. 13, 60–67 (2011). 49. Petruk, S. et al. TrxG and PcG proteins but not methylated histones remain associated 29. Kustatscher, G. et al. Proteomics of a fuzzy organelle: interphase chromatin. EMBO with DNA through replication. Cell 150, 922–933 (2012). J. http://dx.doi.org/10.1002/embj.201387614 (in the press) . 50. Dimitrova, D. S., Todorov, I. T., Melendy, T. & Gilbert, D. M. Mcm2, but not RPA, is 30. Fridman, Y. et al. Subtle alterations in PCNA-partner interactions severely impair a component of the mammalian early G1-phase prereplication complex. J. Cell Biol. DNA replication and repair. PLoS Biol. 8, e1000507 (2010). 146, 709–722 (1999).

METHODS DOI: 10.1038/ncb2918

METHODS MaxQuant to combine our data with the chromatin probability data. We apply Cell lines. For SILAC experiment, HeLa S3 cells in spinners were grown in SILAC normalization, and derived statistics from the MaxQuant evidence file using in- medium depleted for arginine and lysine (Thermo Scientific RPMI 1640 Medium) house Perl scripts. Normalization is based on the ratio of H4, which we assume to and supplemented with dialysed FBS (#26400-036 Invitrogen), MEM non-essential be stable on chromatin from nascent to 2 h mature. For each protein we represent amino acid mix (#11140 Invitrogen), Glutamax (#35050-038 Invitrogen), 66 mgl−1 the spread of data using the median absolute deviation, which we observe to be more of arginine and 100mgl−1 of lysine (SIGMA). Heavy medium was complemented robust than the standard deviation of the mean (s.e.m.) for those proteins that have a with heavy lysine and arginine (Cambridge Laboratory number CNLM-291-0.25, great many quantitative measurements, and it is more in keeping with reporting the number CLM-2265-0). Cells were maintained for 8 to 9 divisions in heavy or light median ratio observed for each protein. It should be noted that the median absolute medium before NCC analysis. deviation is estimated to be approximately 0.8 of the s.e.m.

Synchronization and drug treatment. Cells were synchronized at the G1/S border Quantification of histones by mass spectrometry. The large number of by single thymidine block (2 mM, 17 h) and released into fresh media containing histone genes with highly conserved but not completely identical sequences makes deoxycytidine (24 µM). TIG-3 cells were blocked in G2/M by nocodazole treatment it diﬃcult to unambiguously assign peptides to one particular histone gene. (100ngml−1, 10 h) 42 h after siRNA transfection. To follow S-phase entry, TIG-3 To simplify data processing of histone sequences, a simplified human histone cells were serum starved (0 % FBS) for 40 h and released by replating in 10% FBS. database was created. This database contains only one representative sequence Cells were treated with cycloheximide (CHX, 200 µgml−1) for 2.5 min before of each of the four replication-dependent core histones H2A, H2B, H3 and biotin–dUTP labelling and with MG132 (5 µM) for 6 h. H4. The protein sequence most closely approaching the consensus sequence was chosen as a representative sequence. In addition, this simplified histone database Immunofluorescence microscopy. Cells grown on coverslips or cytospun onto contains all non-canonical core histones (variants, isoforms of variants, but not glass slides were pre-extracted with CSK (10 mM PIPES at pH 7, 100 mM NaCl, pseudogenes) and all forms of the linker histone H1. Canonical, replication-

300 mM sucrose and 3 mM MgCl2) containing 0.5% Triton for 5 min on ice to dependent histones were distinguished from histone variants according to ref. 54 remove soluble proteins, and fixed with 4% formaldehyde for 7 min. For biotin and, where necessary, based on Uniprot annotations and sequence comparisons. For detection, we used streptavidin conjugated to Alexa Fluor 488 (Invitrogen #S32354). histone quantification, mass spectrometry raw data from the NCC experiments were For EdU (5-ethynyl-2’-deoxyuridine) labelling, cells were incubated with 40 µM searched against this simplified human histone database using MaxQuant 1.2.2.5 EdU (Invitrogen #A10044) for 10–15 min (Fig. 7a,h,k) or 1 h (Fig. 7b,d). EdU was software, omitting potential post-translational modifications from the search and detected using Click-IT Alexa Fluor 488 azide (Invitrogen #A10266) and Click- using only unique peptides for quantification. Each experiment was normalized to IT cell reaction buffer kit (Invitrogen #C10269). Images were acquired using an histone H4. Axiovert 200M confocal microscope equipped with an LSM510 laser module (Zeiss), or a DeltaVision system, analysed and quantified with SoftWorKx 5.0.0 software Chromatin probability. We used a chromatin probability table for 7,600 proteins29. (Applied Precision). For statistical analysis, data were processed in Prism.6 using This table is derived using MCCP (ref. 13) to integrate many chromatin proteomics a t-test when appropriate. For co-localization analysis, three-dimensional images of experiments unrelated to the present study but collectively indicating chromatin single cells were acquired and deconvolved before co-localization was measured by function as assessed by known chromatin proteins. the Pearson coefficient using SoftWorKx. The investigators were blinded to the group allocation during picture acquisitions and quantitative analyses. GST-pulldown experiments. 35S-labelled FAM111A wild-type and PIP mutant were produced using the TnT T7 Quick Coupled Transcription/Translation System Nascent chromatin capture. HeLa S3 cells growing in suspension were released (Promega) and incubated with recombinant GST–PCNA proteins in binding buffer from a single thymidine block for 3 h. For SILAC analysis, 5 × 108 cells were used (150 mM NaCl, 0.2% NP-40, 50 mM Tris, at pH 7.6, 2 mM EDTA, 5% glycerol as starting material for heavy and light cultures. For biotin–dUTP labelling, cells and phosphatase inhibitors) for 3 h at 4 ◦C. The reactions were washed 6 times were incubated for 5 min in a hypotonic buffer (50 mM KCl and 10 mM HEPES) in binding buffer containing 300 mM NaCl and analysed by autoradiography and containing biotin–dUTP and resuspended in fresh cell culture medium. Cells were western blotting. fixed in 2% formaldehyde after 20 min (nascent chromatin) or chased for 2 h in fresh medium before fixation (mature chromatin). Crosslinking was stopped after 15 min Transfection. siRNAs and plasmids were introduced by Oligofectamine by adding glycine to a final concentration of 1% and incubating for 5 min at room (Invitrogen) and Lipofectamine (Invitrogen), respectively, according to the temperature. Nuclei were mechanically isolated in a sucrose buffer (0.3 M sucrose, 10 manufacturer’s recommendations. mM HEPES–NaOH at pH 7.9, 1% Triton X-100 and 2 mM MgOAc), and chromatin was solubilized by sonication in a Diagenode Bioruptor at 4 ◦ C in sonication buffer Statistical methods. For Fig. 7e,h–j and Supplementary Figs 1d and 6h, the values (10 mM HEPES–NaOH at pH 7.9, 100 mM NaCl, 2 mM EDTA at pH 8, 1 mM for each independent experiment are provided in Supplementary Table 4. The EGTA at pH 8, 0.2% SDS, 0.1% sodium sarkosyl and 1 mM phenylmethylsulphonyl median is shown. An unpaired t-test with Welch’s correction has been performed. fluoride; Bioruptor setting: High, 28 cycles of 30 s sonication and 90 sec pause). The normality of the data has been tested by analysing the histogram distribution Biotinylated chromatin fragments were purified on streptavidin-coated magnetic in Prism.6, and by Q–Q plot in SAS (version 9.2 SAS Institute). The difference beads (MyC1 Streptavidin beads) by overnight end-over-end rotation at 4 ◦ C and of variance between two populations was measured in Prism.6. We used Welch’s 5 stringent washes (10 mM HEPES–NaOH at pH 7.9, 100 mM NaCl, 2 mM EDTA correction when the variances were not equal. P values are provided and defined in at pH 8, 1 mM EGTA at pH 8, 0.1% SDS, and 1 mM phenylmethylsulphonyl the legend of the figure. For Fig. 7f, each value on the graph is the median from one fluoride). For SILAC analysis, nascent and mature samples were mixed in the last experiment. Values for each independent experiment are provided in Supplementary wash. To release chromatin and reverse the crosslink, beads were boiled in LSB Table 4. The mean is shown with the s.d. We have used a one-sample t-test to for 40 min at 100 ◦ C, including a brief vortex and short spin every 10 min to compare the two groups to the control in SAS. P values are provided and defined prevent drying. in the legend of the figure.

Proteomics. Proteins from the NCC purification were digested with trypsin. The Primary antibodies, plasmids and siRNA. PCNA (Abcam ab29, clone PC10, resulting peptides were fractionated by offline SCX and analysed on an LTQ Orbitrap 1:1,000), FLAG (Sigma F7425, 1:2,000), H4K12ac (Upstate Millipore 07-595, Velos with online UPLC. Initial data processing was done using MaxQuant (ref. 51). 1:1,000), H4K5ac (Abcam ab51997, clone EP1000Y, 1:1,000), histone 3 (Abcam We performed in-gel digestion52 on NCC purified proteins and fractionated the ab10799, clone mAbcam 10799, 1:2,000), histone H4 (Upstate Millipore 05- peptides using a PolyLC SCX column on Dionex UltiMate 3000. The resulting 30 858, clone 62-141-13, 1:1,000), histone H2B (Abcam ab1790, 1:1,000), Smc3 fractions were desalted using StageTips53 and separated online on a Waters Acquity (Abcam ab9263, 1:1,000), Ctf18 (Bethyl A301-883A, 1:500), RPA/p70 (Abcam UPLC. Each fraction was run on a 2, 3 or 4 h gradient according to estimated material ab79398, clone EPR3472, 1:500), CAF-1 p150 (ref. 55; 1:500), CAF-1 p60 from the offline ultraviolet trace; fractions of very low content were combined. We (ref. 55; 1:500), HP1γ (Abcam ab56978, 1:500), Dnmt1 (Abcam ab16632, used a formic acid/acetonitrile buffer system, and gradients optimized for complex 1:5,000), H3K9me1 (Upstate Millipore 07-450, 1:1,000), H3K9me3 (Upstate samples. Analysis was performed with an LTQ Orbitrap Velos (Thermo Fisher 07-442, 1:1,000), H3K27me1 (Upstate Millipore 07-448, 1:500), H3K27me3 Scientific) using a nano-electrospray ion source. Precursor scans were acquired (CST 9756, 1:500), β-actin (Sigma A5441, clone AC-15, 1:25,000), FAM111A with lock mass in the Orbitrap, and the top 20 ions, with dynamic exclusion, were (Sigma HPA040176, 1:500), MCM2 (BD Transduction Laboratories 610701, selected for collision-induced dissociation fragmentation and measurement in the clone 46/BM28, 1:1,000). The FAM111A–tGFP (RG210012), FAM178A–FLAG LTQ. Initial data processing was done using MaxQuant v 1.2.2.5 (ref. 51), using (RC227146), PTCD3–FLAG (RC204119), RPL29–FLAG (RC203616), KIAA0391– the final release of IPI. We used the default parameters, except that we adjusted FLAG (RC206805) and MPST–FLAG (RC202466) plasmids were from OriGene. minimum unique peptides to 1 and minimum ratio count to 1. We also used The FAM111A PIP mutant was generated by site-directed mutagenesis of

DOI: 10.1038/ncb2918 METHODS

Y24A and F25A, and the FAM178A APIM mutant was generated by site- 53. Rappsilber, J., Ishihama, Y. & Mann, M. Stop and go extraction tips for matrix-assisted directed mutagenesis of Y142A. ATAD2B–FLAG is a kind gift from K. Helin. laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in siRNAs against FAM111A were 50-GGAAGAUAACCACAUAUUUGGCAGG-30 proteomics. Anal. Chem. 75, 663–670 (2003). (FAM111A siRNA1, OriGene) and 50-AGAGCUAAAUGCUUGAUUAGAAATG-30 54. Marzluff, W. F., Gongidi, P., Woods, K. R., Jin, J. & Maltais, L. J. The human (FAM111A siRNA2, OriGene). and mouse replication-dependent histone genes. Genomics 80, 487–498 (2002). 55. Green, C. M. & Almouzni, G. Local action of the chromatin assembly factor CAF-1 at sites of nucleotide excision repair in vivo. EMBO J. 22, 5163–5174 (2003). 51. Cox, J. & Mann, M. MaxQuant enables high peptide identification rates, individualized 56. Chon, H. et al. Contributions of the two accessory subunits, RNASEH2B and p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat. RNASEH2C, to the activity and properties of the human RNase H2 complex. Nucl. Biotechnol. 26, 1367–1372 (2008). Acids Res. 37, 96–110 (2009). 52. Shevchenko, A., Tomas, H., Havlis, J., Olsen, J. V. & Mann, M. In-gel digestion for mass spectrometric characterization of proteins and proteomes. Nat. Protoc. 1, 57. Gilljam, K. M. et al. Identification of a novel, widespread, and functionally important 2856–2860 (2006). PCNA-binding motif. J. Cell Biol. 186, 645–654 (2009).

SUPPLEMENTARY INFORMATION

DOI: 10.1038/ncb2918

Supplementary Figure 1 (a) Biotin-dUTP labelling does not affect S phase shown as a negative control. (d, e) NCC with shorter biotin-dUTP labelling progression. Cells synchronized in mid-S phase were labelled with biotin- times. (d) The percentage of biotin-dUTP positive cells (left) and the biotin- dUTP during a 5 min hypotonic shift (red) or incubated in PBS (blue), dUTP intensity per cell (right) after 5, 10, 20 and 40 minutes of labelling. and analysed by FACS at the indicated times. (b) Biotin-dUTP labelling Horizontal lines represent the median, ****p < 0.0001, *p = 0.0168, n.s. does not trigger DNA damage as measured by gH2AX staining. Cells were p = 0.2918 (unpaired t test, 108 < n). Statistics source data is available in treated with hydroxyurea (HU, 3 mM) for one hour as a positive control. (c) Supplementary Table 4. (e) Western blot analysis of NCC pull-downs after Co-localization between PCNA and biotin-dUTP is lost during chromatin 5 and 10 minutes pulse-labelling with biotin-dUTP. Labelling times of >5 maturation. U-2-OS cells stably expressing PCNA-RFP were pulse labelled minutes are preferable for efficient incorporation of biotin-dUTP. The amount with biotin-dUTP for 20 minutes and fixed directly (nascent) or left for of starting material should always be adjusted according to labelling time in 2 hours before fixation (mature). Cells not treated with biotin-dUTP are order to achieve sufficient material for a comprehensive proteomic analysis.

SUPPLEMENTARY INFORMATION

Alabert et al. Supplementary Figure 2

a HEAVY biotin-dUTP Lys8 Arg10 Nascent Release 3 hrs from chromatin thymidine block 20 min Mass spectrometry analysis Mix heavy and by Orbitab and quantitation light in 1:1 ratio by Maxquant LIGHT biotin-dUTP Lys0 Arg0 Mature Chase 2 hrs Release 3 hrs from chromatin thymidine block 20 min

b c Heavy - Light 80

60 % Max) of ( % un ts 40 l co

20 Ce l -dUTP (%) cells positive G1 G2 G1 G2 biotin 0 Time of biotin-dUTP Time of harvesting labelling Heavy Light

d e

log2 protein ratios (N/M) Fold Difference Histogram 8 2,000

1,500 4

Replicate A or C 2 Replicate B 1,000 Count -8 -6 -4 -2 64 -2 500 -4

-6 0 Rep A vs B: all proteins -5 -4 -3 -2 -1 0 1 2 3 4 5 -8 Rep A vs B: replication factors Rep C vs B: all proteins Median fold difference between Rep C vs B: replication factors replicates log2 protein ratios -10

Supplementary Figure 2 (a) Experimental design for NCC-SILAC. Two SDS-PAGE and mass spectrometry. (b) Biotin-dUTP incorporation scored by independent cultures grown in heavy and light amino acids were released immunofluorescence verified equal labelling efficiency in heavy and light into S phase from a single thymidine block. Cells were labelled with biotin- cultures. Approximately 250 cells were counted. (c) Cell cycle profile of dUTP for 20 minutes 3 hours after release into S phase. Light cultures heavy and light cultures verified matching synchronization at the time of the were left to progress in S phase for 2 hours after labelling, while heavy labelling (left) and that light cultures had progressed to late S phase at the cultures were harvested immediately. Synchronization and release of heavy time of harvest (right). (d) Scatter-plot comparing log2 SILAC ratios between cultures were shifted 2 hours with respect to light cultures, such that they replicates and (e) histogram of median of pairwise log2 fold difference of could be cross-linked and processed in parallel for NCC. After pull-down, SILAC ratios between three replicates, illustrating reproducibility of the NCC- the heavy and light samples were washed stringently and mixed prior to SILAC technology.

SUPPLEMENTARY INFORMATION

Supplementary Figure 3 (a) Nascent chromatin enrichment of factors proposed Moldovan et al., 2007. For details see Supplementary Table 2. (d) Analysis of 20, 56 to deal with DNA-RNA duplexes and protein degradation at the fork . (b) factors enriched on nascent chromatin by western blot. For comparison the log2 Nascent chromatin enrichment of origin recognition and licensing factors. nascent chromatin enrichment is indicated. (e) Nascent chromatin enrichment (c) Coverage of the PCNA interactome by NCC-SILAC. Bars illustrate factors of lysine and arginine methyltransferases (violet) and demethylases (grey). (f) identified in NCC relative to those reported in the literature, numbers are Nascent chromatin enrichment of RNA polymerases. Only unique subunits are indicated in brackets. Functional groups of PCNA interactors are according to shown. All Nascent chromatin enrichments are presented as in Figure 1h.

SUPPLEMENTARY INFORMATION

-8 -6 -4 -2 0 246

Supplementary Figure 4 Box plot of nascent chromatin enrichment of the individual classes defined in Figure 5a.

SUPPLEMENTARY INFORMATION

Alabert et al. Supplementary Figure 5

a b 1 Nuclear Cytoplasmic both 100 0.8 80 0.6 60 0.4 40 0.2 20 Distribution (%) (Pearson coefficient) 0

0 Colocalization with PCNA FAM111A FAM178A ATAD2B FAM111A FAM178A ATAD2B c d

1 Nuclear Cytoplasmic both 100 0.5 80

60 No signal No signal 40 0

20 Distribution (%) No signal (Pearson coefficient)

0 Colocalization with PCNA -0.5 PTCD3 RPL29 KIAA0391 MPST PTCD3 RPL29 KIAA0391 MPST

Supplementary Figure 5 Analysis of GFP- or FLAG-tagged proteins in were scored after pre-extraction. The Pearson coefficient is calculated for U-2-OS cells stably expressing RFP-PCNA. Localization pattern (a, c) individual nuclei and shown in a box plot (n≥ 9). Horizontal line represents and co-localization with PCNA measured by Pearson coefficient (b, d) the median.

SUPPLEMENTARY INFORMATION

Alabert et al. Supplementary Figure 6

a b c

No PCNA Early S Mid S Late S 100 WT APIM PIP Box 80

FLAG 60 11 1A FAM 40 DAPI PCNA patterns (%) APIM 20 20 μm 0 FAM111A PIPmt H2B FAM178A

d e f EdU MCM2 GFP PCNA EdU GFP PCNA-RFP DAPI G1 H2B H2B Early S WT WT Mid-late S PIPmt G2 PIPmt 20 μm 20 μm 20 μm

g h i **** **** U-2-OS **** 100 600 ****

80 400 60 siFAM111A-1 siFAM111A-2 siControl

1x 1x 1x 40 * 200 FAM111A 20 Ponceau S

EdU intensity in PCNA + cells (a.u.) 0 PCNA intensity in EdU + cells (a.u.) 0 siControl siFAM111A-1 siFAM111A-2 siControl siFAM111A-1 siFAM111A-2

Supplementary Figure 6 (a) Conservation of the FAM111A PIP box and co-stained with PCNA. Representative images are shown. (f) RFP-PCNA levels FAM178 APIM in mammals57. (b) Localization of FLAG-tagged FAM178A wild in a stable cell line transfected with FAM111A WT and PIPmt. Representative type (WT) and APIM mutant (APIMmt). (c) PCNA patterns in cells transfected images are shown with arrowheads marking transfected cells. (g-i) FAM111A with FAM111A WT or PIPmt. H2B was used as a control. The mean and of depletion in U-2-OS cells transfected with two independent siRNAs for 48 three independent experiments is shown with error bars representing s.d., 80 hours. Western blot (g), 2x indicates the double amount of sample loaded < n < 175. (d) Representative EdU and MCM2 patterns used to analyse cell in 1x. (*) unspecific band. Quantification of EdU incorporation (h) and cycle distribution in Figure 7d. The EdU pattern distinguishes S phase stage chromatin-bound PCNA (i). Dot plot of PCNA intensities in EdU positive (early, mid or late), while the MCM2 pattern distinguishes EdU negative cells cells are shown. Similar results were obtained in total cells population. One in G1 and G2. (e) Cells expressing the FAM111A PIPmt show a low level of representative experiment is shown. Horizontal lines represents the median, EdU incorporation and strongly reduced loading of PCNA on chromatin. Note ****p < 0,0001 (unpaired t test, 167 < n). Similar results were obtained that the GFP-FAM111A PIPmt does not localize to EdU foci, in contrast to the using a SMART pool (Dharmacon) directed against FAM111A (data not wild type protein. Transfected cells were labelled for one hour with EdU and shown). Statistics source data are available in Supplementary Table 4.

SUPPLEMENTARY INFORMATION

Alabert et al. Supplementary Figure 7 Figure 1f kD kD 38 17 28 14 Histone H3

17 28 14 17 6 14 14

Histone H2B 6 Histone H4 H4K12ac PCNA

Figure 4a kD kD 28 17 14 17

H4K5ac 14 6

6 17 14 17 14 6 H3k27me3 H3k9me3 H3K9me1

28 17 14 17 14 Histone H4 H3K27me1

Figure 4b kD kD 49 17 14 38 H4K5ac β - actin

17 14 17

H4K5ac 14 H4K5ac

17 17 14 14 H3K9me3 H3K27me3

17 14 17 14 Histone H4 Histone H3

Supplementary Figure 7 Uncropped Western blots.

SUPPLEMENTARY INFORMATION

Figure 6f Figure 6g kD 62 49

FAM111A 38 S

38 35 28

17 14 Histone H3 PCNA

Figure 6g S 35 Coomassie

Figure 7g

kD 98 62 FAM111A Ponceau

Supplementary Figure 7 continued Uncropped Western blots.

SUPPLEMENTARY INFORMATION

Supplementary Figure 1e Supplementary Figure 3d kD kD 38 38 28 28 PCNA PCNA

14 49 CAF1 p60

H4K5ac 6

17 188 14

CTF18 98

Histone H3 kD

Supplementary Figure 6g 188

kD DNMT1 98 98 62 62 FAM111A 49 HP1 γ 49

14 Histone H4 Ponceau

Supplementary Figure 7 continued Uncropped Western blots.

SUPPLEMENTARY INFORMATION

Supplementary Table Legends

Supplementary Table 1. List of the 3995 factors quantified by NCC-SILAC. Column S indicates the replicate experiment(s) (A, B and/or C) in which a protein was quantified.

Supplementary Table 2. List of known PCNA interacting factors and their NCC-SILAC enrichment. Categories are adapted from23.

Supplementary Table 3. List of chromatin proteins enriched in nascent chromatin (see Fig. 5d).

Supplementary Table 4. Statistics source data.