DNA Repair 25 (2015) 84–96

Contents lists available at ScienceDirect

DNA Repair journal homepage: www.elsevier.com/locate/dnarepair

Proteome-wide analysis of SUMO2 targets in response to pathological DNA replication stress in human cells Sara Bursomanno a,1 , Petra Beli b,c,1 , Asif M. Khan d,2 , Sheroy Minocherhomji a , Sebastian A. Wagner b,3 , Simon Bekker-Jensen e , Niels Mailand e , Chunaram Choudhary b , Ian D. Hickson a,d , Ying Liu a,d,∗ a

Center for Healthy Aging, Department of Cellular and Molecular Medicine, University of Copenhagen, Panum Institute, DK-2200 Copenhagen, Denmark Department of Proteomics, The Novo Nordisk Foundation Centre for Protein Research, University of Copenhagen, Panum Institute, DK-2200 Copenhagen, Denmark c Institute of Molecular Biology (IMB), Ackermannweg 4, 55128 Mainz, Germany d Molecular Oncology Unit, Weatherall Institute of Molecular Medicine, University of Oxford, Oxford OX3 9DS, UK e Department of Disease Biology, The Novo Nordisk Foundation Centre for Protein Research, University of Copenhagen, Panum Institute, DK-2200 Copenhagen, Denmark b

a r t i c l e

i n f o

Article history: Received 31 July 2014 Received in revised form 26 September 2014 Accepted 28 October 2014 Available online 25 November 2014 Keywords: Common fragile sites DNA replication stress POLD3 Mass spectrometry SUMOylation

a b s t r a c t SUMOylation is a form of post-translational modification involving covalent attachment of SUMO (Small Ubiquitin-like Modifier) polypeptides to specific lysine residues in the target protein. In human cells, there are four SUMO proteins, SUMO1–4, with SUMO2 and SUMO3 forming a closely related subfamily. SUMO2/3, in contrast to SUMO1, are predominantly involved in the cellular response to certain stresses, including heat shock. Substantial evidence from studies in yeast has shown that SUMOylation plays an important role in the regulation of DNA replication and repair. Here, we report a proteomic analysis of proteins modified by SUMO2 in response to DNA replication stress in S phase in human cells. We have identified a panel of 22 SUMO2 targets with increased SUMOylation during DNA replication stress, many of which play key functions within the DNA replication machinery and/or in the cellular response to DNA damage. Interestingly, POLD3 was found modified most significantly in response to a low dose aphidicolin treatment protocol that promotes common fragile site (CFS) breakage. POLD3 is the human ortholog of POL32 in budding yeast, and has been shown to act during break-induced recombinational repair. We have also shown that deficiency of POLD3 leads to an increase in RPA-bound ssDNA when cells are under replication stress, suggesting that POLD3 plays a role in the cellular response to DNA replication stress. Considering that DNA replication stress is a source of genome instability, and that excessive replication stress is a hallmark of pre-neoplastic and tumor cells, our characterization of SUMO2 targets during a perturbed S-phase should provide a valuable resource for future functional studies in the fields of DNA metabolism and cancer biology. © 2014 Elsevier B.V. All rights reserved.

Abbreviations: DTT, dithiothreitol; EGTA, ethylene glycol tetraacetic acid; FACS, fluorescence-activated cell sorting; HA, human influenza hemagglutinin; HCD, higherenergy collisional dissociation; HPLC, high performance liquid chromatography; PCNA, proliferating cellular nuclear antigen protein (Homo sapiens); PCR, polymerase chain reaction; RPA, replication protein A; SILAC, stable isotope labeling with amino acids in cell culture; Strep, streptavidin; STRING, search tool for the retrieval of interacting genes/proteins. ∗ Corresponding author at: Center for Healthy Aging, Department of Cellular and Molecular Medicine, University of Copenhagen, Panum Institute, DK-2200 Copenhagen, Denmark. Tel.: +45 353 27761. E-mail address: [email protected] (Y. Liu). 1 These authors contributed equally to the work. 2 Current address: Department of Neurobiology, Institute for Molecular Medicine, University of Southern Denmark, J.B. Winsløws Vej 25, 2, DK-5000 Odense, Denmark. 3 Current address: Department of Hematology/Oncology, Johann Wolfgang Goethe University, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany. http://dx.doi.org/10.1016/j.dnarep.2014.10.011 1568-7864/© 2014 Elsevier B.V. All rights reserved.

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

1. Introduction Small Ubiquitin-like Modifiers (SUMOs) are a conserved family of polypeptides that are covalently attached to and detached from proteins to modify their cellular function. SUMOs are conjugated to their substrates via an enzymatic cascade analogous to that involved in protein ubiquitylation (reviewed in [1]). In yeast (Schizosaccharomyces pombe or Saccharomyces cerevisiae), there is only one form of SUMO, while in human cells there are four distinct SUMO isoforms: SUMO1–4 [2,3]. SUMO1–3 are expressed in all cell types, and SUMO4 is mainly expressed in the kidney, lymph nodes and spleen. There is 95% amino acid sequence similarity between SUMO2 and SUMO3, whereas the level of similarity between SUMO1 and SUMO2/3 is approximately 50%. It is thought that SUMO2/3 form a subfamily with cellular roles distinct from those of SUMO1. Indeed, previous studies have indicated that the conjugation status of SUMO2/3 and SUMO1 are different in several ways. First, non-conjugated SUMO2/3 is present in molar excess over conjugated SUMO2/3 in normally growing cells, while SUMO1 is virtually all conjugated constitutively to target proteins [4]. Second, when cells are exposed to various toxic stresses, including heat shock and oxidizing agents, SUMO2/3 become conjugated to substrates, while the level of conjugated SUMO1 remains unchanged [4]. Third, previous proteomic studies have revealed that distinct groups of substrates are conjugated to SUMO1 or SUMO2/3 in unstressed growth conditions [5,6] or following heat shock [7]. Fourth, the sub-cellular localization of SUMO2/3 and SUMO1 is generally different. In interphase cells, SUMO2/3 are distributed throughout the nucleoplasm, while SUMO1 is specifically localized to the nuclear envelope and the nucleolus [8]. Finally, only SUMO2 and SUMO3 contain within their polypeptide sequence a SUMO consensus modification motif (␺KxE/D) and can, therefore, form polySUMO chains [9]. Genetic studies in yeast have indicated that polySUMO chains play a role in chromosome segregation, recovery from checkpoint arrest, the DNA damage response and in meiosis [10–12]. A major source of DNA damage occurs during the process of DNA replication. For example, replication forks can encounter lesions in the template or bound proteins that interfere with fork progression [13]. This perturbation of replication can cause deletions or gene rearrangements at several genomic loci; in particular, at common fragile sites (CFSs). These sites manifest as gaps or breaks visible in condensed metaphase chromosomes, and are considered part of normal chromosome structure and are present in nearly all individuals. They normally are not prone to breakage (called ‘expression’), but can be induced to break by treating cells with replication perturbing agents such as a low dose of the DNA polymerase inhibitor aphidicolin (APH) [14,15] that still permits cells to traverse S phase. Previous studies have demonstrated that CFSs are a primary target for oncogene-induced DNA damage in pre-neoplastic lesions, suggesting that CFS instability could be a key player during tumorigenesis [16]. Mechanisms that regulate the expression of CFSs have been studied extensively and documented [17,18]. In S. cerevisiae, it has been shown that SUMO is essential for the viability [19], while in S. pombe SUMO is important for the cellular response to replication perturbation [20]. It has also been demonstrated in S. cerevisiae that DNA double strand breaks can trigger simultaneous multisite SUMOylation of several proteins involved in DNA damage response pathways [21,22]. In human cells, only a handful of proteins have been identified as being SUMOylated during the processes of DNA damage/repair and replication, including BLM [23,24], 53BP1 [25], BRCA1 [26], and TDG [27], most of which are targeted by SUMO1. The proteins that are modified by SUMO2 during DNA replication or repair events remain largely unknown. Considering the strong evidence that SUMO2/3 are involved in the cellular response to various stresses, we hypothesized that

85

SUMO2/3 conjugation might play an important regulatory role in DNA replication and repair. Hence, we employed mass spectrometry (MS)-based, quantitative proteomics to identify proteins whose SUMOylation status is altered when cells are under DNA replication stress. We performed two biological replicate experiments and identified 976 putative SUMO2 target proteins under both normal growth conditions and during replication stress. Moreover, we found that the SUMOylation status of 22 proteins increased more than 2-fold when cells are exposed to DNA replication stress. Most interestingly, POLD3 was shown to have the greatest increase in SUMO2 conjugation in cells exposed to conditions that activate CFS breakage. Because POLD3 is the human ortholog of POL32 in budding yeast, and plays a role in break-induced recombinational repair in both yeast and human cells, we subsequently analyzed the involvement of POLD3 in the DNA replication stress response. We found that human cells deficient for POLD3 had a more prolonged DNA damage stress response when the cells were treated with a low dose of APH that is known to activate CFS breakage. Our data suggest that POLD3 is SUMOylated in response to replication stress and could play an important role in the regulation of cellular response to DNA replication stress. 2. Materials and methods 2.1. Cloning and stable cell line generation The complementary DNA (cDNA) of human SUMO-2 isoform a (NCBI reference sequence: NP 001005849.1) was amplified using the polymerase chain reaction (PCR) from an IMAGE clone. The PCR product was subcloned into a bacterial expression vector that contains Strep-HA tag (pcDNA4/TO-Strep-HA, made in-house). The Strep-HA-SUMO2 cDNA was then sub-cloned into the eukaryotic expression vector pcDNA3.1(+) (Life Technologies) at the KpnI and NotI sites. All constructs were verified by DNA sequencing. The pcDNA3.1-Strep-HA-SUMO2 vector was transfected into U2OS cells using FuGENE 6 transfection reagent (Promega). G418 resistant clones expressing Strep-HA-SUMO2 recombinant protein were selected. 2.2. Cells culture conditions and drug treatments U2OS and HeLa cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (Gibco), penicillin (100 U/ml) and streptomycin (100 ␮g/ml). Where required, G418 (400 ␮g/ml) was added to the medium to maintain the selection of U2OS cells with pcDNA3Strep-HA-SUMO2. To enrich for cells in a perturbed S phase, an asynchronously growing cell population was first synchronized at the G1/S boundary by incubation in medium containing 3 mM HU for 18 h, then rinsed twice with PBS and were released in medium without the drug for the length of time as indicated in Fig. 1A. To induce a perturbed S-phase, the HU-treated cells were released into medium containing 0.4 ␮M APH for 8 h. 2.3. Flow cytometry Following the treatment under the various conditions listed above, the cells were trypsinized, washed twice in PBS, and suspended in 1 ml of ice cold 70% ethanol (added drop-wise) and incubated at −20 ◦ C overnight. Fixed cells were rinsed twice in PBS, and were incubated 30 min at 37 ◦ C in 1 ml of PBS containing 40 ␮g/ml propidium iodide (Sigma–Aldrich) and 200 ␮g/ml RNase A. DNA content was analyzed by measuring propidium iodide staining of the DNA using a FACS Calibur (Becton Dickinson). Percentages

86

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

Fig. 1. Cell cycle regulation and DNA replication stress induction in human U2OS cells. (A) Schematic representation of the procedure for generating conditions of DNA replication stress in U2OS cells. Asynchronous U2OS cells were either untreated or treated with 3 mM HU for 18 h (HU). Cells treated with HU were then either released in drug-free medium for 7 h (HU-release), or released for 1 h and then exposed to 0.4 ␮M APH for 8 h (HU-APH). Cells were harvested for the following analysis at the end of the treatment. (B) Cell cycle profiles of cell populations generated as described in panel (A). Percentages of cells in different phases of the cell cycle are shown. (C) Representative immuno-fluorescence images showing RPA foci (red) in U2OS cells treated as described in panel (A). The locations of nuclei were defined using DAPI (blue). Scale bars represent 10 ␮m. (D) Global analysis of SUMO1 and SUMO2 protein conjugation in whole cell extracts of the cell populations described in panel (A). Tubulin was used as a loading control.

of cells in G1, S or G2/M phases were calculated using CellQuest software. 2.4. SILAC labeling and enrichment of SUMOylated proteins U2OS-Strep-HA-SUMO2 cells were cultured in medium containing either l-arginine and l-lysine, l-arginine-U-13 C6 and l-lysine-U-D4 , or l-arginine-U-13 C6 -15 N4 and l-lysine-U-13 C6 -15 N2 (Cambridge Isotope Laboratories) as described previously [28].

Cells were lysed in lysis buffer (modified RIPA buffer: 20 mM Tris–HCl, pH 7.5, 50 mM NaCl, 0.5% NP40, 0.5% sodium deoxycholate, 0.5% SDS, 1 mM EDTA) supplemented with protease and phosphatase inhibitors, and the lysates were cleared by centrifugation at 16,000 × g for 15 min at 4 ◦ C. Protein concentration was determined using a BCA protein assay kit (Thermo Scientific) and 10 mg of protein from each cell population were mixed in a 1:1:1 ratio. Strep-tagged SUMOylated proteins were enriched using Strep-Tactin resin (IBA) for 2 h at 4 ◦ C with rotation (10 rpm). The

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

87

precipitated proteins were washed four times in lysis buffer containing 1 M NaCl and once in lysis buffer. Proteins were eluted by boiling in sample buffer (4% SDS, 120 mM Tris–HCl, 1 ␮M DTT), resolved by SDS-PAGE and were stained using Colloidal Blue Staining Kit (Life Technologies). The gel lane was cut into 10 slices and proteins were digested in-gel with trypsin.

BLM (1:1000; Abcam ab476). An antibody specific for human SETX was kindly provided by Dr. Delia Domenico (Fondazione IRCCS Istituto Nazionale Tumori, Milano, Italy). Horseradish peroxidaseconjugated rabbit or mouse secondary antibodies (Sigma) were used at the dilution 1:2000. Proteins were visualized using an ECL Plus system (Thermo Scientific).

2.5. MS analysis

2.8. Immune-fluorescence (IF) analysis

Peptide fractions were analyzed on a quadrupole Orbitrap (QExactive, Thermo Scientific) mass spectrometer equipped with an EASY-nLC II nanoflow HPLC system (Thermo Scientific) as described previously [29]. Raw data files were analyzed using MaxQuant [30]. Parent ion and MS2 spectra were searched against protein sequences obtained from the UniProt knowledge base using the Andromeda search engine [31]. Spectra were searched with a mass tolerance of 6 ppm in MS mode, 20 ppm in HCD MS2 mode, strict trypsin specificity and allowing for up to 2 missed cleavage sites. Cysteine carbamidomethylation was included as a fixed modification and N-terminal protein acetylation and methionine oxidation were included as variable modifications. The dataset was filtered based on posterior error probability (PEP) to arrive at a false discovery rate of 1% for peptide spectrum matches and protein groups. Statistical analysis was performed using the R software environment. Gene Ontology term enrichment analysis was performed using the DAVID bioinformatics resource [32]. Network analysis of proteins was performed with the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database [33] and proteins were visualized using Cytoscape [34]. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository with the dataset identifier PXD000716 [35].

For IF analysis with an HA antibody (Santa Cruz, sc-7392), U2OSStrep-HA-SUMO2 cells were grown on cover slips in 6-well cell culture plates. Cells were washed in PBS, fixed in 4% paraformaldehyde for 15 min, and permeabilized in 0.2% Triton X-100. Cover slips with cells were incubated with primary antibody (1:300 dilution) for 1 h at room temperature and then with a secondary antibody (Invitrogen) conjugated to AlexaFluor 488 for 30 min at room temperature. Cover slips were mounted in DAPI-containing mounting medium (Vectashield), and images were acquired on an LSM780 confocal microscope (Zeiss) using a 40× objective. For IF analysis with RPA antibody, U2OS cells were grown and treated on cover slips in 6-well cell culture plates. After the treatment, cells were washed in PBS, fixed in PTEMF buffer (0.2% Triton, 1 M PIPES, pH 6.8, 1 M MgCl2 , 0.5 M EGTA, pH 8, 4% formaldehyde) for 20 min, and permeabilized in 0.2% Triton, 3% BSA for 20 min. Cover slips were incubated with the RPA antibody (1:100; Abcam, ab2175-500) for 2 h at room temperature and then with secondary antibodies (Invitrogen) conjugated to Alexa Fluor 568 for 1.5 h at room temperature. Cover slips were mounted in DAPI-containing mounting medium (Vectashield). Images were quantified using a Nikon Eclipse 80i microscope and software (Nikon, ACT-1, v2.62) or a LSM 700 confocal microscope (Zeiss) and Zen imaging software (v7, 2011). 2.9. Colony formation assay

2.6. RNA interference POLD3 SMARTpool were an equimolar mixture of four siRNAs (5 -UGUAUAGCAAGCUGAGUAA-3 , 5 -CAAUUAGUGGUUAGGGAAA-3 , 5 -GGCAUUAUGUCUAGGACUA-3 and 5 -ACGAAAACGCGUACUAAAA-3 ) (Dharmacon, Cat. No. L-026693-01-0010). The sequence of a scrambled siRNA was 5 -GACCAAGUUCCGUCACUAA3 (Sigma–Aldrich). Lipofectamine RNAiMax reagent (Life Technologies) was used in siRNA transfection experiments. 2.7. Western blotting For the analysis of the global enrichment of proteins conjugated to SUMO2 during replication stress, U2OS cells were treated with the three conditions shown in Fig. 1A. After the treatment, cells from each condition were lysed in Laemmli buffer and an amount of total cell extract (20 ␮g protein per sample) was loaded onto 4–12% gradient gels (Biorad), and western blotted using a semi-dry blotting system. For the validation of SILAC results, U2OS-StrepHA-SUMO2 cells were treated with three conditions shown in Fig. 1A. The cells were then harvested with lysis buffer, 30 ␮g total protein from each sample was saved as input, and 2.5 mg total protein from each sample was used for the pull-down assay. All of the input or pull-down samples were separated on gradient gels (Biorad) and the proteins were transferred to Hybond-P membrane (GE Healthcare) using a semi-dry blotting system. After blocking in 5% milk, membranes were incubated with the following commercially available primary antibodies at dilutions as follows: POLD3 (1:500; Abnova H00010714-M01), HA (1:2000; Santa Cruz sc-7392), SUMO2/3 (1:2000; Abcam ab81371), BRCA1 (1:500; Santa Cruz sc-6954), PCNA (1:2000; Santa Cruz sc-56), FANCD2 (1:500; Novus NB100-182), FANCI (1:500; Bethyl A301-254A),

U2OS cells were transfected with scrambled or POLD3 siRNAs, and were seeded onto 10 cm plates (1000 cells per plate) 24 h after the transfection. Cells were then subjected to HU-release or HUAPH treatment (Fig. 1A) 45 h from the siRNA transfection. At the end of the treatment, cells were released to fresh drug-free medium and grown for 10–13 days until visible colonies could be observed. The colonies were then fixed and stained with Coomassie solution (0.1% Brilliant blue R, 7% acetic acid, 50% methanol) (Sigma–Aldrich), and the number of colonies was scored. Three plates were analyzed for each condition in each experiment, and two replicate experiments were performed. 3. Results 3.1. SUMO2 conjugation to proteins during a perturbed S phase Prior to embarking on a large-scale proteomic analysis of cellular SUMO targets, we determined whether the global pattern of SUMO conjugation during a perturbed S phase is in any way different for SUMO1 and SUMO2. For this, we designed a protocol to enrich for cells undergoing replication stress by first arresting the cells in early S phase using HU (3 mM for 18 h). The cells were then released from this arrest into drug-free medium for one hour, before being exposed to low dose APH (0.4 ␮M) for 8 h, a condition known to induce CFS expression [36,37]. Untreated cells, the cells treated only with HU and harvested (HU-release), and the cells treated with HU and released into APH-containing medium (HU-APH) were then analyzed in parallel (Fig. 1A). The HU-release condition was designed to enrich for cells in S-phase of the cell cycle, while the HUAPH sample defined a population where late-replicating CFSs are activated. The majority of cells were shown to be in mid S phase 7 h

88

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

Fig. 2. Generation of U2OS clonal cell lines stably expressing Strep-HA-SUMO2. Western blotting analysis of 3 clonal cell lines (#3, #5 and #9) with antibodies against SUMO2 (A) or HA (B). In each case, parental U2OS cells were analyzed as control. Tubulin was used as loading control. The positions of the endogenous SUMO2 and Strep-HA-SUMO2 are indicated with arrows. Asterisks denote endogenous proteins that bind to HA antibody non-specifically. (C) IF analysis of clonal cell line #5 used in the study. Cells were stained with anti-HA antibodies and DAPI.

after release from HU treatment into drug-free medium. However, for cells released into APH-containing medium, it took approximately 9 h for the population to reach a comparable stage of S-phase (Fig. 1B). These data indicate that HU effectively arrests U2OS cells in early S-phase without apparently affecting their subsequent progression through S-phase following its removal. However, mild APH treatment slows S phase progression by about 2 h, indicative of the induction of replicative stress (Fig. 1B). To evaluate whether any of the cell treatments employed could induce a DNA damage response, we analyzed the induction of RPA foci, a marker of ssDNA, using immuno-fluorescence (IF). We found, as expected, that untreated cells very rarely displayed any RPA foci, but approximately 70–80% of the cells arrested by HU treatment contained readily detectable RPA foci (Fig. 1C). In the cells treated under HU-release condition, there was a marked reduction in the level of RPA foci. Strikingly, cells treated with APH retained elevated levels of RPA foci (Figs. 1C and S1). These data indicate that a DNA damage response persists in cells exposed to low dose APH, consistent with the slowing of S-phase under these conditions.

Next, to assess the extent of protein conjugation to SUMO1 or SUMO2, we used western blotting analysis on extracts from untreated, HU-release and HU-APH cell populations. The level of SUMO1 conjugation appeared largely unchanged, while there was a substantial increase of SUMO2 conjugation in the HU-APH condition (Fig. 1D). A small, but consistent, increase in SUMO2 conjugation was also seen in the HU-release cell population compared to the untreated control (Fig. 1D). These data are consistent with the previous findings using other types of cellular stress [4], and indicate that SUMO2 very likely plays an important regulatory role during the cellular response to DNA damage and replication stress. 3.2. Establishment of a cell line stably expressing SUMO2 tagged with Strep-HA To facilitate the identification of proteins conjugated to SUMO2, we generated a human U2OS cell line stably expressing SUMO2 tagged with Strep-HA at its N-terminus. Henceforth, we will refer to this cell line as U2OS-Strep-HA-SUMO2. The Strep-tag has an

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

89

Fig. 3. Identification of proteins SUMOylated in response to replication stress by SILAC-based mass spectrometry. (A) U2OS-Strep-HA-SUMO2 cells were SILAC labeled with light, medium, or heavy isotopes, and were then subjected to the indicated treatments. Cells were lysed, and SUMOylated proteins were enriched using Step-Tactin resin. After stringent washing, proteins eluated from the resin were resolved by SDS-PAGE and visualized by colloidal Coomassie staining. Proteins were in-gel digested and analyzed by LC–MS/MS. (B) Cell cycle profiles of differentially treated cell populations. The percentages of cells in different cell cycle phases are shown. (C) Venn diagram shows the overlap of quantified proteins in the two experimental replicates. (D) Distribution of the logarithmized SILAC ratios H/L from both replicate experiments visualized as lineplot. The pink shaded area shows the 22 proteins with a log2 (SILAC ratio H/L) > 1 that were used for subsequent bioinformatic analysis. (E) Comparison of log2 SILAC ratios for proteins SUMOylated under H versus L conditions (X-axis) and H versus M conditions (Y-axis). Note that POLD3 is a significant outlier.

90

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

Table 1 Proteins with increased SUMOylation in response to DNA replication stress with a combined SILAC ratio of H/L (HU-APH/untreated) > 2 (log2 > 1) from two replicate experiments. Their combined SILAC ratios of M/L (HU-release/untreated) are also shown. Most of these proteins have been found SUMOylated by SUMO2 in various previous studies. The proteins highlighted in green are very likely to be SUMOylated by SUMO2 in natural S phase [42].

Protein Name

Gene Name

Unique Peptide Number

Log2 Ratio H/L Combined

Log2 Ratio M/L Combined

SUMOylated by SUMO2 in previous studies* B, D, F, G

DNA polymerase delta subunit 3

POLD3

6

2.18

0,59

BRCA1-associated RING domain protein 1

BARD1

10

1.91

1,92

Breast cancer type 1 susceptibility protein

BRCA1

10

1.84

1,74

AT-rich interactive domain-containing protein 3B

ARID3B

8

1.81

1,89

TP53

2

1.80

2,62

F

Cellular tumor antigen p53

G

Ribonucleoside-diphosphate reductase subunit M2

RRM2

8

1.62

1,34

B

Transcription factor MafF

MAFF

5

1.62

2,16

B

Fanconi anemia group I protein Fanconi anemia group D2 protein Actin-binding protein anillin Protein Mis18-alpha FLYWCH-type zinc finger-containing protein 1 WD repeat and HMG-box DNA-binding protein 1 Ski-like protein Replication factor C subunit 1

FANCI

35

1.59

1,35

B, E, D, G

FANCD2

16

1.57

1,19

B, G G

ANLN

10

1.54

1,51

MIS18A

2

1.52

1,54

FLYWCH1

8

1.49

1,32

G

WDHD1

9

1.22

0,65

B, D, F

SKIL

2

1.22

1,1

RFC1

10

1.22

0,85

B, D, F

Ankyrin repeat domain-containing protein 32

ANKRD32

7

1.19

0,84

G

Protein MMS22-like

MMS22L

10

1.18

1,13

NAB1

17

1.17

1,64

NGFI-A-binding protein 1 Mis18-binding protein 1

B, C, D, F, G

MIS18BP1

16

1.10

0,9

Chromosome-associated kinesin KIF4A

KIF4A

47

1.08

0,97

Proliferating cell nuclear antigen

PCNA

11

1.04

0,83

E, F, G

Probable helicase senataxin

SETX

35

1.01

1,37

A, B, D, G

intrinsically high affinity toward Strep-Tactin resin. Therefore the Strep-tagged proteins can be purified in a single step from crude cell lysate under denaturing conditions, which is an essential requirement for isolating SUMO-conjugates, as both the activity of SUMO proteases and the non-covalent binding of SUMO are inhibited. Moreover, the high affinity of the interaction allows more stringent washing conditions to be adopted, thus greatly minimizing the non-specific binding. The expression of tagged SUMO2 in candidate stable cell lines was confirmed first by western blotting analysis. In the selected lines, the exogenously expressed Strep-HA-SUMO2 migrated on SDS-PAGE with an expected molecular mass, and could be detected using antibodies to both SUMO2 (Fig. 2A) and HA (Fig. 2B). Although the Strep-HA-SUMO2 was expressed at a relatively higher level than the endogenous SUMO2, its conjugation pattern was very similar to that of the endogenous SUMO2 (Fig. S2). The cellular localization of Strep-HA-SUMO2 was also analyzed by immunofluorescence using an antibody against a HA tag (Fig. 2C). We found that Strep-HA-SUMO2 was predominantly localized to the nucleoplasm, but was excluded from the nucleolus, as expected from previous studies [8]. 3.3. Quantitative analysis of proteins conjugated to SUMO2 in human cells Stable isotope labeling with amino acids in cell culture (SILAC) [28] was employed to metabolically label the U2OS-Strep-HASUMO2 cells. Untreated, HU-release, and HU-APH treated cells

B, D

were labeled with light, medium, or heavy isotopes, respectively. The SILAC labeled cells were treated and harvested for quantitative proteomic analysis at the end of the treatment as described in Fig. 3A. The cell cycle distributions of the HU-release cells and those treated with HU-APH were comparable at the time when they were harvested for MS analysis (Fig. 3B). In total, we identified 1355 putative SUMOylated proteins from two biological replicate experiments. There was an excellent quantitative reproducibility (R = 0.9) between the replicate experiments (Table S1) and a majority of the proteins (976 out of 1355) were identified in both experiments (Fig. 3C). To identify the proteins that are significantly SUMOylated under DNA replication stress conditions, we extracted from these data the proteins with a SILAC H/L (HU-APH/untreated) ratio of over 2.0 (log 2 > 1). This identified 22 proteins, of which 16 were identified previously as SUMO2 targets under normal or heat shock conditions (Fig. 3D and Table 1) [38–44]. Most interestingly, SUMOylation of POLD3 was substantially increased in the HU-APH condition (labeled with heavy isotopes; H) in comparison with that in HU-release condition (labeled with medium isotopes; M), while most of the other proteins showed little difference in SUMOylation status between the H and M conditions (Table 1 and Fig. 3E). Gene Ontology term enrichment analysis using ‘biological process’ and ‘cellular compartment’ terms with proteins having more than a 2-fold increase in SUMOylation after replication stress (22 proteins) indicated a significant enrichment of proteins involved in DNA replication fork processing and DNA repair (Fig. 4A). Network analysis of these 22 proteins with the STRING database (Search Tool

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

91

Fig. 4. Gene ontology (GO) and functional interaction network analysis of proteins SUMOylated after replication stress. (A) GO ‘cellular compartment’ (left) and ‘biological process’ (right) term enrichment analysis of proteins with increased SUMOylation after replication stress. Significance of the enrichment is indicated for each term on the right. (B) Functional interaction network analysis of proteins with increased SUMOylation after replication stress. All proteins with a SILAC ratio H/L > 2 are represented as nodes in the network and functional interactions were obtained using the STRING database. The node size depicts the SILAC ratio H/L.

for the Retrieval of Interacting Genes/Proteins) revealed that they are connected in functional interaction networks (Fig. 4B). 3.4. Validation of the SILAC data We next validated the SILAC results for selected hits from the MS-based screen using a Strep-tactin pull-down assay and western blotting analysis with the non-SILAC labeled U2OS-StrepHA-SUMO2 cell line. These cells were treated with the same conditions as those described in Fig. 3A. Of most interest, we validated that POLD3 is SUMOylated and that its SUMOylation status is highest in the cells exposed to low dose APH (HU-APH; Fig. 5A). We also performed similar analysis on proteins with an established role in DNA replication and repair, including BRCA1, FANCI, FANCD2, PCNA, SETX, and TP53. In all cases analyzed, we could detect a SUMOylated version of the protein that migrated more slowly on the gel compared to the unmodified protein (Fig. 5B). Importantly, and consistent with our quantitative MS data, the level of SUMOylation in each case was increased to a similar degree in the HU-release and HU-APH conditions compared to the untreated condition (Fig. S3).

3.5. POLD3 depletion leads to an increased persistence of RPA foci Based on the above findings, and the fact that the yeast ortholog POLD32 is known to play an important role in break-induced recombination repair, we investigated the involvement of POLD3 in the DNA replication stress response. To do this, we depleted POLD3 from U2OS cells using a validated POLD3 siRNA [45]. Western blotting analysis showed that POLD3 protein depletion was >90% after 72 h of exposure to POLD3 siRNA compared to the cells treated with the siRNA scrambled control (Fig. 6A). We then analyzed the cell cycle profile and survival of cells transfected with either POLD3 siRNA or the scrambled siRNA control. This showed that depletion of POLD3 did not affect the S phase fraction of the population (Fig. S4A), or the plating efficiency (Fig. S4B), which is consistent with previous findings [45]. In addition, there were no changes in the number of the RPA foci in cells treated with either siRNA scrambled control or POLD3 (Fig. S4C). We then compared the response of the cells transfected with scrambled or POLD3 siRNAs to replication stress. For this, we analyzed the siRNA-treated cells using the same protocol as for the MS analysis (Fig. 6B). We first investigated whether the POLD3 depleted cells showed evidence of

92

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

Fig. 5. Validation of the MS results by western blotting. U2OS-Strep-HA-SUMO2 cells were with either untreated, treated with HU alone (HU-release), or HU and APH (HU-APH) conditions. Strep-HA-SUMO2-conjugates were purified using Strep-Tactin resin. Proteins were detected with anti-POLD3 (panel A), or with anti-BRCA1, FANCD2, PCNA, SETX, FANCI, or TP53 antibodies (panel B). A sample of the input was used in each case as control of the un-modified form of the protein. Note the slower migration of the SUMO2-conjugated protein compared to un-modified protein. The total amount of pull-down was shown by the abundance of free Strep-HA-SUMO2, as revealed by immuno-blotting using anti-HA antibody. The asterisks denote trace amounts endogenous PCNA and TP53, which we found consistently to be bound by the Strep-Tactin beads.

an aberrant DNA damage stress response. We found that the POLD3 depleted cells showed both a higher number of RPA foci and a greater propensity to show pan-nuclear staining for RPA compared to the scrambled siRNA control cells, but only when cells were under replication stress (Fig. 6C and D). Moreover, the percentage of cells with persistent RPA foci was higher in the POLD3 deficient cells compared to the control cells even 4 h after release from the HU-APH treatment. Interestingly, the POLD3 deficient cells also showed a small, but consistent, delay in the timing of G/M phase entry after the release compared with the control cells (Fig. 6D).

4. Discussion Interest in the phenomenon of replication stress has grown dramatically in recent years with the realization that perturbation of DNA replication is a hallmark of pre-neoplastic cells and tumors. It has been proposed that replication stress might be both a key driver of tumorigenesis and required for the acquisition of oncogenic genome rearrangements by cancer cells, especially at CFSs [46,47]. Studies in yeast have illustrated that SUMOylation is also a key modification in DNA replication and repair processes

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

93

Fig. 6. POLD3 depletion induces the formation of more and persistent RPA foci following HU-APH treatment. (A) Western blot analysis of the total cell extracts (30 ␮g per sample) harvested from cells transfected with either scrambled or POLD3 siRNA 48 or 72 h after the transfection. Actin was used as loading control. Cell lysate from un-transfected cells (–) was used as background control. (B) Schematic illustration of the treatment applied to U2OS cells with siRNA transfection and HU-APH treatment. (C) Left: representative immuno-fluorescence images showing RPA foci in scrambled or POLD3 siRNAs transfected cells treated under HU-APH condition. Scale bars indicate 20 ␮m. Right: quantification of the number of RPA foci per cell. Data are means of two replicate experiments. An average of 300 cells was analyzed for each condition in all the experiments. Error bars show standard deviation (SD). (D) Left: cell cycle profiles of scrambled or POLD3 siRNAs transfected U2OS cells treated under either untreated or HU-APH conditions. Right: quantification of the number of cells positive with RPA foci with untreated, HU-APH condition, or 4 h after the release from HU-APH treatment. Data are means of three independent experiments. An average of 300 cells was analyzed in each condition in all the experiments. Error bars show SD (* indicates p < 0.01).

94

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

[20–22]. Consistent with this, while characterizing a panel of factors enriched in the vicinity of newly replicated DNA, a recent study indicated that SUMOylation, but not ubiquitylation, plays an important role in the DNA replication process in human cells, although the relevant targets were not identified [48]. Based on the above background, we set out to identify the proteins SUMOylated when human cells are exposed to DNA replication stress. We could first observe that SUMO2, but not SUMO1, is conjugated to various proteins when cells are under these conditions (Fig. 1D), which is consistent with the previous findings that SUMO2 predominantly participates in the cellular response to various stresses. Using SILAC based MS technology, we subsequently identified a panel of 22 proteins that are conjugated to SUMO2 following exposure of cells to replication stress (the HU-release or the HU-APH conditions; Fig. 3). It has proven difficult to identify the targets of SUMO2 due to the fact that SUMO2 is mostly not conjugated in normal cell growth conditions [4], and the reversible nature of the SUMO conjugation process [1]. Nonetheless, considerable progress has been made toward the identification of SUMO2 targets using advanced MS technology. The most common strategy is to use His-tagged or tandem affinity protein (TAP)-tagged SUMO2 to facilitate the pull-down of SUMO2 target proteins [5,7,40–44,50–52]. A less common approach is to use a SUMO2 antibody to bind to SUMO2 and enrich for covalently bound SUMO2 conjugates in the immunoprecipitate [38]. With this approach, the targets detected are those conjugated to endogenously expressed SUMO2, although fewer targets are identified using this method in general as the affinity of the SUMO2 antibody is not as strong as those to the commonly used protein tags. Our study represents the first comprehensive analysis of SUMO2 targets under conditions of replication stress. Noteworthy, amongst the 22 proteins with significant increase of SUMOylation to SUMO2, 16 were found conjugated to SUMO2 in previous studies, while 2 (FANCI, and PCNA) were identified using the ‘SUMO2 antibody’ approach as well as the ‘tag’-based approach [38] (Table 1). This gives us the confidence that these 22 proteins are likely to be genuine SUMO2 targets. In addition, it should be noted that 4 of the 22 proteins were recently found to be SUMOylated by SUMO2 in a normal S phase [42] (Table 1). Most interestingly, POLD3 was the only protein that showed more significant SUMOylation in the APH-treated cells in both SILAC MS and western blotting analysis (Figs. 3E and 5A). POLD3 is a 66-kDa subunit of DNA polymerase delta. In budding yeast, its ortholog, Pol32, is known to play an essential role in break-induced recombination [53,54]. POLD3 was shown previously to bind to PCNA via its C-terminus [55]. Recent studies have demonstrated that POLD3 is recruited to sites of UV damage in human cells [56], and phosphorylation of POLD3 regulates its interaction with PCNA [57]. Very recently, it was shown that POLD3 was required for the continuation of DNA synthesis when human cells are under double strand break-inducing stress [45]. Our experiments with siRNA knockdown POLD3 in U2OS cells indicated that it might play an important role in regulating the DNA replication stress response. At this stage it is not known how POLD3 affects the cellular response to perturbation of DNA replication. One possibility that should be pursued in future studies is to analyze whether the interaction of POLD3 with PCNA is regulated by the conjugation of SUMO2. To address this and other questions, the first step would be to identify the SUMO2 conjugation sites in POLD3 under conditions of DNA replication stress, and then to carry out functional studies with mutant POLD3 derivatives lacking these target residues. In this context, a previous study mapped two sites in POLD3 (K258 and K433) that are conjugated to SUMO3, but not to SUMO2 [58]. More recently, these two sites were also identified as SUMO2 conjugation sites in cells treated with heat shock [44]. Considering that these two sites match the known SUMOylation consensus sequence

(KxD/E, where ␺ is a large hydrophobic residue, K is the modified lysine, x is any amino acid and D/E is an acidic residue), it would be worthwhile to validate whether these two sites are SUMOylated when cells are treated with APH. Because 50–70% of known SUMOylation sites conform to this tetrapeptide sequence (either in the forward or reverse orientation) [42,44], more detailed MS analysis should be carried out to identify other potential SUMO2 conjugation sites in POLD3 when cells encounter DNA replication stress. Although our study focused mainly on proteins whose SUMOylation is increased during a perturbed S phase, it is clear that a small set of proteins show decreased SUMOylation under these conditions. Using a threshold of a 0.5-fold change (log2 H/L: −1), we found 13 proteins in this category, of which only 4 were identified in both experiments (Table S1). Because KRT2 is a common contaminant in MS analysis, we propose that TCF12, TDG, and CAPN1 are the only proteins that we can say with confidence are de-SUMOylated in a perturbed S phase. Amongst these, TDG (thymine-DNA glycosylase) is of significant interest because it can act as a DNA repair enzyme, and is a known target of both SUMO1 and SUMO2/3 [27]. It was shown previously by in vitro studies that unmodified TDG remains bound to an unstable DNA repair intermediate (the abasic site) until a conformational change upon conjugation to SUMO1 induces its release [59,60]. Interestingly, a recent study demonstrated that TDG is part of transcription regulatory complexes and contributes to the epigenetic stability in differentiated cells by de novo methylation [61]. The substantial decrease in SUMOylation of TDG in a perturbed S phase revealed in our SILAC data (Fig. S5) is consistent with the notion that, in un-stressed or differentiated cells, TDG is SUMOylated by either SUMO1 or SUMO2 to restrict its access to chromatin. However, when cells are stressed with DNA replication inhibitors, such as HU and APH, TDG becomes deSUMOylated, which allows binding to promoter regions to protect them from aberrant CpG methylation [61]. 5. Conclusion We have identified a set of 22 SUMO2 targets that are modified at higher levels in human cells exposed to pathogenic DNA replication stress. In particular, we have shown that POLD3 is SUMOylated most significantly in stress conditions that cause CFS breakage, and that it could play a role in DNA replication stress responses. Our data provide a comprehensive description of SUMO2 targets when cells are under replication stress, which will serve as valuable foundation for future functional analysis of individual proteins required for DNA replication and repair. Conflict of interest The authors declare that they have no conflict of interest. Author contributions SB, PB, AMK, SM, SBJ, and YL carried out the experiments. SB, PB, SAW, and YL performed data analysis. CC, NM, IDH, and YL designed and supervised the study. SB, PB, IDH and YL wrote the manuscript. Access code The raw mass spectrometry proteomics data are deposited with the ProteomeXchange Consortium (http://www. proteomexchange.org) via the PRIDE partner repository with the dataset identifier PXD000716. Assess details: username: [email protected]; password: cJq8okp9.

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96

Acknowledgements We would like to thank members of the Choudhary, Hickson and Liu groups for helpful discussions, Drs. D. Huttner and H. Mankouri for comments on the manuscript, and Dr. D. Huttner for advice on the SILAC data validation experiments. Work in the authors’ laboratories is funded by the Nordea Foundation, the Danish Medical Research Council, and the Danish Natural Sciences Research Council. SB is funded by a PhD fellowship from the Faculty of Health and Medical Sciences, University of Copenhagen. PB is supported by the Emmy Noether Program of the German Research Foundation (DFG, BE 5342/1-1).

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.dnarep. 2014.10.011.

References [1] R.T. Hay, SUMO: a history of modification, Mol. Cell 18 (2005) 1–12. [2] D. Guo, M. Li, Y. Zhang, P. Yang, S. Eckenrode, D. Hopkins, W. Zheng, S. Purohit, R.H. Podolsky, A. Muir, J. Wang, Z. Dong, T. Brusko, M. Atkinson, P. Pozzilli, A. Zeidler, L.J. Raffel, C.O. Jacob, Y. Park, M. Serrano-Rios, M.T. Larrad, Z. Zhang, H.J. Garchon, J.F. Bach, J.I. Rotter, J.X. She, C.Y. Wang, A functional variant of SUMO4, a new I kappa B alpha modifier, is associated with type 1 diabetes, Nat. Genet. 36 (2004) 837–841. [3] F. Melchior, SUMO – nonclassical ubiquitin, Annu. Rev. Cell Dev. Biol. 16 (2000) 591–626. [4] H. Saitoh, J. Hinchey, Functional heterogeneity of small ubiquitin-related protein modifiers SUMO-1 versus SUMO-2/3, J. Biol. Chem. 275 (2000) 6252–6258. [5] A.C. Vertegaal, J.S. Andersen, S.C. Ogg, R.T. Hay, M. Mann, A.I. Lamond, Distinct and overlapping sets of SUMO-1 and SUMO-2 target proteins revealed by quantitative proteomics, Mol. Cell Proteomics 5 (2006) 2298–2310. [6] Y. Zhao, S.W. Kwon, A. Anselmo, K. Kaur, M.A. White, Broad spectrum identification of cellular small ubiquitin-related modifier (SUMO) substrate proteins, J. Biol. Chem. 279 (2004) 20999–21002. [7] G. Rosas-Acosta, W.K. Russell, A. Deyrieux, D.H. Russell, V.G. Wilson, A universal strategy for proteomic studies of SUMO and other ubiquitin-like modifiers, Mol. Cell Proteomics 4 (2005) 56–72. [8] F. Ayaydin, M. Dasso, Distinct in vivo dynamics of vertebrate SUMO paralogues, Mol. Biol. Cell 15 (2004) 5208–5218. [9] M.H. Tatham, E. Jaffray, O.A. Vaughan, J.M. Desterro, C.H. Botting, J.H. Naismith, R.T. Hay, Polymeric chains of SUMO-2 and SUMO-3 are conjugated to protein substrates by SAE1/SAE2 and Ubc9, J. Biol. Chem. 276 (2001) 35368–35374. [10] G.R. Bylebyl, I. Belichenko, E.S. Johnson, The SUMO isopeptidase Ulp2 prevents accumulation of SUMO chains in yeast, J. Biol. Chem. 278 (2003) 44113–44120. [11] C.H. Cheng, Y.H. Lo, S.S. Liang, S.C. Ti, F.M. Lin, C.H. Yeh, H.Y. Huang, T.F. Wang, SUMO modifications control assembly of synaptonemal complex and polycomplex in meiosis of Saccharomyces cerevisiae, Genes Dev. 20 (2006) 2067–2081. [12] D.C. Schwartz, R. Felberbaum, M. Hochstrasser, The Ulp2 SUMO protease is required for cell division following termination of the DNA damage checkpoint, Mol. Cell Biol. 27 (2007) 6948–6961. [13] A. Helmrich, M. Ballarino, L. Tora, Collisions between replication and transcription complexes cause common fragile site instability at the longest human genes, Mol. Cell 44 (2011) 966–977. [14] S.G. Durkin, T.W. Glover, Chromosome fragile sites, Annu. Rev. Genet. 41 (2007) 169–192. [15] T.W. Glover, C. Berger, J. Coyle, B. Echo, DNA polymerase alpha inhibition by aphidicolin induces gaps and breaks at common fragile sites in human chromosomes, Hum. Genet. 67 (1984) 136–142. [16] P.K. Tsantoulis, A. Kotsinas, P.P. Sfikakis, K. Evangelou, M. Sideridou, B. Levy, L. Mo, C. Kittas, X.R. Wu, A.G. Papavassiliou, V.G. Gorgoulis, Oncogene-induced replication stress preferentially targets common fragile sites in preneoplastic lesions. A genome-wide study, Oncogene 27 (2008) 3256–3264. [17] M. Debatisse, B. Le Tallec, A. Letessier, B. Dutrillaux, O. Brison, Common fragile sites: mechanisms of instability revisited, Trends Genet. 28 (2012) 22–32. [18] S. Minocherhomji, I.D. Hickson, Structure-specific endonucleases: guardians of fragile site stability, Trends Cell Biol. 24 (2014) 321–327. [19] K. Tanaka, J. Nishide, K. Okazaki, H. Kato, O. Niwa, T. Nakagawa, H. Matsuda, M. Kawamukai, Y. Murakami, Characterization of a fission yeast SUMO-1 homologue, pmt3p, required for multiple nuclear events, including the control of telomere length and chromosome segregation, Mol. Cell Biol. 19 (1999) 8660–8672.

95

[20] A. Skilton, J.C. Ho, B. Mercer, E. Outwin, F.Z. Watts, SUMO chain formation is required for response to replication arrest in S. pombe, PLoS ONE 4 (2009) e6750. [21] C.A. Cremona, P. Sarangi, Y. Yang, L.E. Hang, S. Rahman, X. Zhao, Extensive DNA damage-induced sumoylation contributes to replication and repair and acts in addition to the mec1 checkpoint, Mol. Cell 45 (2012) 422–432. [22] I. Psakhye, S. Jentsch, Protein group modification and synergy in the SUMO pathway as exemplified in DNA repair, Cell 151 (2012) 807–820. [23] S. Eladad, T.Z. Ye, P. Hu, M. Leversha, S. Beresten, M.J. Matunis, N.A. Ellis, Intranuclear trafficking of the BLM helicase to DNA damage-induced foci is regulated by SUMO modification, Hum. Mol. Genet. 14 (2005) 1351–1365. [24] K.J. Ouyang, L.L. Woo, J. Zhu, D. Huo, M.J. Matunis, N.A. Ellis, SUMO modification regulates BLM and RAD51 interaction at damaged replication forks, PLoS Biol. 7 (2009) e1000252. [25] Y. Galanty, R. Belotserkovskaya, J. Coates, S. Polo, K.M. Miller, S.P. Jackson, Mammalian SUMO E3-ligases PIAS1 and PIAS4 promote responses to DNA double-strand breaks, Nature 462 (2009) 935–939. [26] J.R. Morris, C. Boutell, M. Keppler, R. Densham, D. Weekes, A. Alamshah, L. Butler, Y. Galanty, L. Pangon, T. Kiuchi, T. Ng, E. Solomon, The SUMO modification pathway is involved in the BRCA1 response to genotoxic stress, Nature 462 (2009) 886–890. [27] U. Hardeland, R. Steinacher, J. Jiricny, P. Schar, Modification of the human thymine-DNA glycosylase by ubiquitin-like proteins facilitates enzymatic turnover, EMBO J. 21 (2002) 1456–1464. [28] S.E. Ong, B. Blagoev, I. Kratchmarova, D.B. Kristensen, H. Steen, A. Pandey, M. Mann, Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics, Mol. Cell Proteomics 1 (2002) 376–386. [29] A. Michalski, E. Damoc, J.P. Hauschild, O. Lange, A. Wieghaus, A. Makarov, N. Nagaraj, J. Cox, M. Mann, S. Horning, Mass spectrometry-based proteomics using Q Exactive, a high-performance benchtop quadrupole Orbitrap mass spectrometer, Mol. Cell Proteomics 10 (2011), M111.011015. [30] J. Cox, M. Mann, MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification, Nat. Biotechnol. 26 (2008) 1367–1372. [31] J. Cox, N. Neuhauser, A. Michalski, R.A. Scheltema, J.V. Olsen, M. Mann, Andromeda: a peptide search engine integrated into the MaxQuant environment, J. Proteome Res. 10 (2011) 1794–1805. [32] W. Huang da, B.T. Sherman, R.A. Lempicki, Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources, Nat. Protoc. 4 (2009) 44–57. [33] K.G. Engstrom, L. Ohlsson, Morphology and filterability of red blood cells in neonatal and adult rats, Pediatr. Res. 27 (1990) 220–226. [34] R. Saito, M.E. Smoot, K. Ono, J. Ruscheinski, P.L. Wang, S. Lotia, A.R. Pico, G.D. Bader, T. Ideker, A travel guide to Cytoscape plugins, Nat. Methods 9 (2012) 1069–1076. [35] J.A. Vizcaino, R.G. Cote, A. Csordas, J.A. Dianes, A. Fabregat, J.M. Foster, J. Griss, E. Alpi, M. Birim, J. Contell, G. O’Kelly, A. Schoenegger, D. Ovelleiro, Y. Perez-Riverol, F. Reisinger, D. Rios, R. Wang, H. Hermjakob, The PRoteomics IDEntifications (PRIDE) database and associated tools: status in 2013, Nucleic Acids Res. 41 (2013) D1063–D1069. [36] K.L. Chan, T. Palmai-Pallag, S. Ying, I.D. Hickson, Replication stress induces sister-chromatid bridging at fragile site loci in mitosis, Nat. Cell Biol. 11 (2009) 753–760. [37] S. Ying, S. Minocherhomji, K.L. Chan, T. Palmai-Pallag, W.K. Chu, T. Wass, H.W. Mankouri, Y. Liu, I.D. Hickson, MUS81 promotes common fragile site expression, Nat. Cell Biol. 15 (2013) 1001–1007. [38] J. Becker, S.V. Barysch, S. Karaca, C. Dittner, H.H. Hsiao, M. Berriel Diaz, S. Herzig, H. Urlaub, F. Melchior, Detecting endogenous SUMO targets in mammalian cells and tissues, Nat. Struct. Mol. Biol. 20 (2013) 525–531. [39] R. Bruderer, M.H. Tatham, A. Plechanovova, I. Matic, A.K. Garg, R.T. Hay, Purification and identification of endogenous polySUMO conjugates, EMBO Rep. 12 (2011) 142–148. [40] F. Golebiowski, I. Matic, M.H. Tatham, C. Cole, Y. Yin, A. Nakamura, J. Cox, G.J. Barton, M. Mann, R.T. Hay, System-wide changes to SUMO modifications in response to heat shock, Sci. Signal. 2 (2009) ra24. [41] I. Matic, J. Schimmel, I.A. Hendriks, M.A. van Santen, F. van de Rijke, H. van Dam, F. Gnad, M. Mann, A.C. Vertegaal, Site-specific identification of SUMO2 targets in cells reveals an inverted SUMOylation motif and a hydrophobic cluster SUMOylation motif, Mol. Cell 39 (2010) 641–652. [42] J. Schimmel, K. Eifler, J.O. Sigurethsson, S.A. Cuijpers, I.A. Hendriks, M. Verlaande Vries, C.D. Kelstrup, C. Francavilla, R.H. Medema, J.V. Olsen, A.C. Vertegaal, Uncovering SUMOylation dynamics during cell-cycle progression reveals FoxM1 as a key mitotic SUMO target protein, Mol. Cell 53 (2014) 1053–1066. [43] J. Schimmel, K.M. Larsen, I. Matic, M. van Hagen, J. Cox, M. Mann, J.S. Andersen, A.C. Vertegaal, The ubiquitin-proteasome system is a key component of the SUMO-2/3 cycle, Mol. Cell Proteomics 7 (2008) 2107–2122. [44] T. Tammsalu, I. Matic, E.G. Jaffray, A.F. Ibrahim, M.H. Tatham, R.T. Hay, Proteome-wide identification of SUMO2 modification sites, Sci. Signal. 7 (2014) rs2. [45] L. Costantino, S.K. Sotiriou, J.K. Rantala, S. Magin, E. Mladenov, T. Helleday, J.E. Haber, G. Iliakis, O.P. Kallioniemi, T.D. Halazonetis, Break-induced replication repair of damaged forks induces genomic duplications in human cells, Science 343 (2014) 88–91. [46] R.A. Burrell, S.E. McClelland, D. Endesfelder, P. Groth, M.C. Weller, N. Shaikh, E. Domingo, N. Kanu, S.M. Dewhurst, E. Gronroos, S.K. Chew, A.J. Rowan, A. Schenk, M. Sheffer, M. Howell, M. Kschischo, A. Behrens, T. Helleday, J. Bartek,

96

[47]

[48]

[50]

[51]

[52] [53] [54]

S. Bursomanno et al. / DNA Repair 25 (2015) 84–96 I.P. Tomlinson, C. Swanton, Replication stress links structural and numerical cancer chromosomal instability, Nature 494 (2013) 492–496. J.A. Harrigan, R. Belotserkovskaya, J. Coates, D.S. Dimitrova, S.E. Polo, C.R. Bradshaw, P. Fraser, S.P. Jackson, Replication stress induces 53BP1-containing OPT domains in G1 cells, J. Cell Biol. 193 (2011) 97–108. A.J. Lopez-Contreras, I. Ruppen, M. Nieto-Soler, M. Murga, S. Rodriguez-Acebes, S. Remeseiro, S. Rodrigo-Perez, A.M. Rojas, J. Mendez, J. Munoz, O. FernandezCapetillo, A proteomic characterization of factors enriched at nascent DNA molecules, Cell Rep. 3 (2013) 1105–1116. A.C. Vertegaal, S.C. Ogg, E. Jaffray, M.S. Rodriguez, R.T. Hay, J.S. Andersen, M. Mann, A.I. Lamond, A proteomic study of SUMO-2 target proteins, J. Biol. Chem. 279 (2004) 33791–33798. H.A. Blomster, V. Hietakangas, J. Wu, P. Kouvonen, S. Hautaniemi, L. Sistonen, Novel proteomics strategy brings insight into the prevalence of SUMO-2 target sites, Mol. Cell Proteomics 8 (2009) 1382–1390. M.H. Tatham, I. Matic, M. Mann, R.T. Hay, Comparative proteomic analysis identifies a role for SUMO in protein quality control, Sci. Signal. 4 (2011) rs4. B. Llorente, C.E. Smith, L.S. Symington, Break-induced replication: what is it and what is it for? Cell Cycle 7 (2008) 859–864. J.R. Lydeard, S. Jain, M. Yamaguchi, J.E. Haber, Break-induced replication and telomerase-independent telomere maintenance require Pol32, Nature 448 (2007) 820–823.

[55] J.R. Pohler, M. Otterlei, E. Warbrick, An in vivo analysis of the localisation and interactions of human p66 DNA polymerase delta subunit, BMC Mol. Biol. 6 (2005) 17. [56] J. Chea, S. Zhang, H. Zhao, Z. Zhang, E.Y. Lee, Z. Darzynkiewicz, M.Y. Lee, Spatiotemporal recruitment of human DNA polymerase delta to sites of UV damage, Cell Cycle 11 (2012) 2885–2895. [57] L. Lemmens, S. Urbach, R. Prudent, C. Cochet, G. Baldacci, P. Hughes, Phosphorylation of the C subunit (p66) of human DNA polymerase delta, Biochem. Biophys. Res. Commun. 367 (2008) 264–270. [58] G. Liu, E. Warbrick, The p66 and p12 subunits of DNA polymerase delta are modified by ubiquitin and ubiquitin-like proteins, Biochem. Biophys. Res. Commun. 349 (2006) 360–366. [59] D. Baba, N. Maita, J.G. Jee, Y. Uchimura, H. Saitoh, K. Sugasawa, F. Hanaoka, H. Tochio, H. Hiroaki, M. Shirakawa, Crystal structure of thymine DNA glycosylase conjugated to SUMO-1, Nature 435 (2005) 979–982. [60] R. Steinacher, P. Schar, Functionality of human thymine DNA glycosylase requires SUMO-regulated changes in protein conformation, Curr. Biol. 15 (2005) 616–623. [61] D. Cortazar, C. Kunz, J. Selfridge, T. Lettieri, Y. Saito, E. MacDougall, A. Wirz, D. Schuermann, A.L. Jacobs, F. Siegrist, R. Steinacher, J. Jiricny, A. Bird, P. Schar, Embryonic lethal phenotype reveals a function of TDG in maintaining epigenetic stability, Nature 470 (2011) 419–423.