Nucleophosmin interaction with APE1: Insights into DNA repair regulation

1 Nucleophosmin interaction with APE1: Insights into DNA repair regulation. David J. López, Ander de Blas, Mikel Hurtado1, Mikel García-Alija2, Jon Mentxaka, Igor de la Arada, María A. Urbaneja, Marián Alonso-Mariño, Sonia Bañuelos* Biofisika Institute (UPV/EHU, CSIC) and Department of Biochemistry and Molecular Biology, University of the Basque Country (UPV/EHU), Leioa, Spain. Present addresses: 1NEIKER-Basque Institute of Agricultural Research and Development, c/ Berreaga 1, E-48160 Derio, Spain. 2CIC bioGUNE , E-48160 Derio, Spain. Abbreviations: AML: acute myeloid leukemia; APE1: apurinic apyrimidinic endonuclease 1; BER: Base excision repair; CD: circular dichroism; DDR: DNA damage repair; DSB: double strand break; FT-IR: Fourier transform infrared spectroscopy; IR: ionizing radiation; ITC: isothermal titration calorimetry; NPM1: nucleophosmin; PA: polyacrylamide; PAGE: polyacrylamide gel electrophoresis; THF: tetrahydrofurane. *Corresponding author at: Biofisika Institute, University of the Basque Country, Barrio Sarriena s/n, E-48940 Leioa, Spain. E-mail address: [email protected] This is the accepted manuscript of the article that appeared in final form in DNA repair 88 : (2020) // Article ID 102809, which has been published in final form at https://doi.org/10.1016/j.dnarep.2020.102809. © 2020 Elsevier under CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/) 2 Abstract Nucleophosmin (NPM1), an abundant, nucleolar protein with multiple functions affecting cell homeostasis, has also been recently involved in DNA damage repair. The roles of NPM1 in different repair pathways remain however to be elucidated. NPM1 has been described to interact with APE1 (apurinic apyrimidinic endonuclease 1), a key enzyme of the base excision repair (BER) pathway, which could reflect a direct participation of NPM1 in this route. To gain insight into the possible role(s) of NPM1 in BER, we have explored the interplay between the subnuclear localization of both APE1 and NPM1, the in vitro interaction they establish, the effect of binding to abasic DNA on APE1 conformation, and the modulation by NPM1 of APE1 binding and catalysis on DNA. We have found that, upon oxidative damage, NPM1 is released from nucleoli and locates on patches throughout the chromatin, perhaps co-localizing with APE1, and that this traffic could be mediated by phosphorylation of NPM1 on T199. NPM1 and APE1 form a complex in vitro, involving, apart from the core domain, at least part of the linker region of NPM1, whereas the C-terminal domain is dispensable for binding, which explains that an AML leukemia-related NPM1 mutant with an unfolded Cterminal domain can bind APE1. APE1 interaction with abasic DNA stabilizes APE1 structure, as based on thermal unfolding. Moreover, our data suggest that NPM1, maybe by keeping APE1 in an “open” conformation, favours specific recognition of abasic sites on DNA, competing with off-target associations. Therefore, NPM1 might participate in BER favouring APE1 target selection as well as turnover from incised abasic DNA. Keywords: APE1, NPM1, nucleophosmin, BER, base excision repair, protein-protein interaction. 3 Introduction Nucleolus, beyond its canonical ribosome assembly function, is being increasingly recognized as a key hub in coordinating progression of the cell cycle and responses to cellular perturbations, such as damage to DNA (Boisvert et al., 2007; Emmott & Hiscox, 2009; Antoniali et al., 2014; Ogawa & Baserga, 2017). Upon sensing different types of stresses, the localization of several nucleolar components is modulated and their involvement in repair tasks orchestrated from this membrane-less organelle (Scott & Oeffinger 2016; Antoniali et al., 2014). Nucleophosmin (NPM1) is one of the nucleolar proteins that seem to be involved in DNA damage repair (DDR) mechanisms. NPM1 is homopentameric and consists of multiple domains: an N-terminal compact, - structured core domain, responsible for oligomerization (Lee et al., 2007), is connected by long, flexible linkers, to small, helical, C-terminal domains (Grummitt et al., 2008). Although normally enriched in nucleoli, NPM1 continuously shuttles between cytoplasm, nucleoplasm and nucleolus (Lindström, 2011). NPM1 shuttling is mediated by transport receptors of the karyopherin family (importin  exportin CRM1) (Lindström, 2011; Arregi et al., 2015), whereas nucleolar accumulation probably depends on its affinity for G-rich DNA or RNA sequences adopting special structures known as G-quadruplexes (Bañuelos et al., 2013; Chiarella et al, 2013). NPM1 is in charge of multiple functions related to cell homeostasis, including biogenesis and export of ribosomes (Maggi et al., 2008), control of centrosome duplication (Wang et al., 2005), regulation of the stability of tumour suppressors such as p53 (Colombo et al 2002) and Arf (Colombo et al., 2005), and cell response to stress. Recent experimental evidence points to NPM1 as a key factor directly participating in DNA damage repair (DDR) pathways, although its role in these mechanisms remains to be established (Scott & Oeffinger 2016; Box et al., 2016; Koike et al., 2010; Moore et al., 2013; Ziv et al., 2014; Guillonneau et al., 2016; Yang et al., 2016). Regarding NPM1 activity in the context of stress response, it was found associated with chromatin upon induction of double strand breaks (DSBs) with ionizing radiation (IR) or etoposide (Lee et al., 2005). The putative changes in subnuclear localization of NPM1 upon DSBs and its participation in the management of these lesions seem to be regulated by phopshorylation. T199-phosphorylated NPM1, which locates outside the nucleoli, is recruited to DSBs repair sites (foci) in response to IR (Koike et al., 2010). Moreover, NPM1 undergoes nucleoplasmic translocation upon different genotoxic 4 insults, such as oxidative stress, UV, cisplatin, etc. (Colombo et al., 2011; Scott & Oeffinger 2016; Yang et al., 2016) and, apart from DSBs repair, it has been implicated in yet further DNA repair pathways, namely translesion synthesis (TLS) (Ziv et al., 2014) and base excision repair (BER) (Vascotto et al., 2009; Vascotto et al., 2014). NPM1 has been described to interact with APE1 (apurinic apyrimidinic endonuclease 1) (Vascotto et al., 2009), a key enzyme of the BER pathway, aimed to repair DNA damage caused by oxidizing or alkylating agents, including chemotherapeutic drugs (Almeida & Sobol, 2007; Krokan & Bjoras, 2013; Parsons & Dianov, 2013). APE1 specifically recognizes abasic sites in DNA and hydrolyzes the phosphodiester bond 5´of the abasic sugar (Freudenthal et al., 2015), generated after removal of a modified / damaged base by a specific glycosidase, or by spontaneous depurination (Li and Wilson, 2014). Afterwards, a polymerase restores the lacking nucleotide using the complementary strand as template, and the phosphodiester link is re-formed by a ligase. APE1 performs additional functions in the context of DNA repair (being able to correct “non productive” ends of DNA strands in repair intermediates) and, probably, in RNA management (as an “RNA quality control”) (Antoniali et al., 2014). Besides its roles in DDR, APE1 acts as a redox coactivator of different transcription factors such as p53 and NF-B. APE1 catalytic cycle on DNA has been described in structural detail (Mol et al., 2000; Freudenthal et al., 2015); however, the mechanisms regulating APE1 functionality are largely unknown. NPM1 might have a role in BER repair, modulating the localization of APE1 (Vascotto et al., 2014). In the absence of DNA damage, APE1 locates through nucleoplasm and is enriched in nucleoli, which is probably mediated by its interaction with NPM1 through a flexible, N-terminal region of APE1. This region seems to also play a regulatory role in APE1 binding to DNA (Fantini et al., 2010; Vascotto et al., 2014). Furthermore, acetylation of lysine residues in this segment might modulate the interaction with NPM1 and thus APE1 nucleolar localization (Lirussi et al., 2012). The fact that, as mentioned, NPM1 itself may transit to the nucleoplasm depending on stress conditions and its own posttranslational modifications adds further complexity to the regulation of NPM1 and APE1 dynamics. A pool of APE1 seems to locate also in mitochondria, and could take part in the repair of oxidized DNA in this organelle (Li & Wilson, 2014). On the other hand, NPM1 participation in multiple, mechanistically different repair pathways might be also based on its chromatin remodelling (histone chaperoning) 5 capacities (Okuwaki et al., 2001). A further, as yet unexplored possibility is that NPM1 takes part directly in the BER enzymatic machinery. Indeed, NPM1 seems to slightly stimulate APE1 incision activity on abasic DNA in vitro (Vascotto et al., 2009). A direct involvement of NPM1 in repair tasks remains however to be confirmed, and the underlying mechanism deciphered. DNA repair machineries are valuable targets in cancer, as long as interfering with their functioning can enhance the efficiency of antineoplasic therapies (Pilié et al., 2018). In this sense, APE1 is considered a target for chemosensitization and radiosensitization (Wilson & Simeonov, 2010; Li & Wilson, 2014). On the other hand, NPM1 is regarded as an oncoprotein, being deregulated (overexpressed, mutated and/or aberrantly located) in several types of human cancer (Grisendi et al., 2006). In particular, it is mutated, misfolded, and partly accumulated in the cytoplasm of tumour cells in a subtype of acute myeloid leukemia (AML) (Falini et al., 2005; Federici & Falini, 2013). NPM1 binds many ligands, so that when aberrantly located as in AML, might induce subsequent mislocalization of some of these partners. Indeed, it has been described that APE1 (Vascotto et al., 2014), amongst other proteins, may be, as a result, dislocated to cytoplasm in AML, an alteration perhaps related with triggering of the disease. To better understand the role of NPM1 in DNA repair, as well as the pathological consequences of its mislocalization, we have explored the subnuclear dynamics of NPM1 and APE1 under stress conditions. In addition, we have characterized the interaction between both proteins and the effects of oncogenic mutations of NPM1 on the binding properties. Finally, we have evaluated the consequences of DNA binding on APE1 conformation, and the effect of NPM1 on APE1 binding to abasic DNA. Altogether, our results suggest a mechanism for regulation of APE1 activity by NPM1 and shed light on the putative role of APE1:NPM1 interaction in BER repair. 1. Material and methods. 1.1. Cell culture, transfection, treatment, imaging and immunoblotting. The clone of human APE1 fused to EGFP was kindly provided by Dr. Izumi (Jackson et al, 2005). The gene was subcloned fused to C-terminal mCherry, between XhoI and SmaI sites of vector pCRY2PHR-mCherryN1, a gift from Chandra Tucker (Kennedy et 6 al., 2010). Human HeLa and HEK293T cells (DSMZ collection) were grown in Dulbecco’s modified Eagle’s medium, supplemented with 10% fetal bovine serum, 100 units/mL penicillin and 100 μg/mL streptomycin (all from Invitrogen). Transfections of cells, seeded onto sterile glass coverslips in 12-well trays the previous day, were carried out with XtremeGENE 9 (Roche) following the manufacturer’s protocol. Induction of oxidative damage and extraction of soluble proteins were done following a previously described protocol (Campalans et al., 2013). The next day after transfection, cells were treated with 0 – 300 mM KBrO3 in PBS for 30 min at 37ºC, and afterwards grown for 3 h longer in fresh medium, in order to “recover”. Then, prior to fixation with 3.7% formaldehyde in PBS for 30 min, cells were washed with 10mM PIPES pH 6.8, 100mM NaCl, 300mM glucose, 3mM MgCl2, 0.5% Triton X-100 and a cocktail of protease inhibitors (Roche) (CSK buffer, Campalans et al., 2013), to extract soluble proteins. Other treatments were done with 0.5 and 5 mM H2O2 (Sigma) for 30 min, 5 g/mL neocarzinostatin (Sigma) for 4 h or 0.5 g/mL nocodazole (Sigma) for 24 h. For fluorescence immunostaining, fixed cells were further permeabilized with 0.2% Triton X-100 in PBS for 10 min, and incubated for 1 h in blocking solution (3% BSA in PBS). Endogenous NPM1 was detected with mouse anti-B23 (Santa Cruz Biotechnology FC82291, dilution 1:800 in blocking solution), and NPM1 phosphorylated at Thr199 with rabbit anti-NPM1p199 (Abcam EP1857Y, dilution 1:150). The secondary antibodies were Alexa Fluor 594 anti-mouse (Invitrogen A11005, dilution 1:400) and Alexa Fluor 594 anti-rabbit (Invitrogen A11012, dilution 1:400). Washing between the different steps was carried out with PBS. Cells were finally mounted on to microscope slides using Vectashield containing DAPI (Vector Laboratories). Cell imaging was performed with the support of SGIker Service (University of the Basque Country UPV/EHU). Routine images were taken with Zeiss Axioscope or Axio Observer Apotome 2 microscope. High resolution images (Fig. 1) were acquired in a Zeiss LSM880 with a 63x lens (Plan Apochromat, 1.4 NA), using the Fast Airyscan superresolution mode Two days later, after the different treatments, they were gently washed with cold PBS, scraped on ice and resuspended in 50 L of lysis buffer (Pierce) supplemented with a cocktail of inhibitors (Roche). Following lysis for 30 min and centrifugation, samples were loaded in a 12.5% PAGE gel and then transferred to a nitrocellulose membrane (GE Healthcare) in a wet apparatus (Bio-Rad). Then, the membrane was blocked with 7 5% (w/v) non‐fat dry milk powder in TBS‐T and incubated with the same primary antibodies (anti-NPM1, diluted 1:800 in blocking solution, and antiNPM1p199, diluted 1:2500) as for immunostaining. As controls of loaded amount, antitubulin (Sigma Aldrich, T9026, dilution 1:5000) and anti-GAPDH (Santa Cruz Biotechnology, sc-25778, dilution 1:2000) were used. After washing steps with TBS-T, the membrane was incubated with the corresponding anti-mouse and anti-rabbit HRPconjugated antibodies and finally detection was made with the Luminata Forte Western HRP Substrate (Millipore). 1.2. Proteins overexpression and purification. Full length wild type NPM1, mutant A (NPM1MutA), and nucleophosmin spanning residues 1-188 (NPM1C106), all with an N-terminal His-ZZ tag, were overexpressed in E. coli BL21 (DE3), and purified as in Arregi et al. (2015). Purification involved NiNTA chromatography, tags removal with TEV protease, reverse Ni-NTA and a final size exclusion chromatography (SEC) step. Histagged NPM1 core domain was a clone obtained from Dr. Se Won Suh (Lee et al., 2007). For purification of this construct, cells were disrupted by sonication in 25 mM Tris-HCl pH 7.5, 100 mM NaCl, 10% glycerol, 1mM DTT, supplemented with lysozyme (20 mg/L culture) and a cocktail of protease inhibitors (Complete, EDTA-free, Roche). The clarified extract was loaded on a NiNTA affinity column (His Trap FF, GE Healthcare) and the protein eluted with an imidazole gradient. Then it was further purified by gel filtration chromatography with Superdex 200 (GE Healthcare) in 25 mM Tris-HCl pH 7.5, 100 mM NaCl, concentrated and frozen for storage. APE1 (full length and APE1N33) was subcloned in the same plasmid as NPM1 (pTGA20) and purified using the same chromatographic steps. SEC was done in buffer 25 mM Tris-HCl pH 7.5, 100 mM NaCl, 1 mM DTT, 10% glycerol. After purification, all proteins were concentrated, flash frozen with liquid nitrogen and stored at -80 ºC. They were quantified with NanoDrop 2000 (Thermo Scientific), based on 280 as calculated with ProtParam (Gasteiger et al., 2005). 1.3. DNA oligonucleotides. As model of abasic DNA, we used a dumbbell-shaped oligonucleotide of 46 residues, with a centrally located tetrahydrofurane (THF) (Freudenthal et al., 2015) 5´- CTGGAGCTTGCTCCAG CGCXCGGTCGATCGTA AGATCGACCGTGCG-3′, where X represents THF), synthesized by IDT (Leuven, Belgium). For anisotropy 8 binding assays, a fluorescein-labelled (in thymine 28) oligo with the same sequence was used. To prepare the “substrate” (“Dumbbell S” or “Dumbbell SF” when fluorescently labelled), the oligos were resuspended at 100 M in water, annealed by heating 5 min at 95 ºC and slowly cooling down in a thermoblock, and ligated with T4 ligase (New England Biolabs) (4000 U ligase per nmol of DNA). Ligation of near 100% of the DNA was confirmed by denaturing polyacrylamide-urea electrophoresis (see section 2.8). For activity and binding assays the DNA was purified from ligase and magnesium with the “Nucleotide Removal Kit” (QIAGEN) and re-annealed; this could not be done for CD and FT-IR, though, since due to the higher amount needed, the lost of yield was too high to be afforded. FT-IR spectra, most CD measurements and also some binding assays were performed with a model of product DNA (“Dumbbell P” and “Dumbbell PF”, respectively) (Freudenthal et al., 2015), with the same sequence as the substrate but starting with 5´ THF. 1.4. Native electrophoresis. Samples containing a fixed concentration of NPM1 and increasing amounts of APE1 in buffer 20 mM potassium phosphate, 50 mM NaCl, 5 mM MgCl2, 2 mM DTT, 0.01% (w/v) Tween-20, pH 7.0 (total volume of 20 L) were incubated for 30 min at room temperature and loaded in precast native 4-16% polyacrylamide (PA), 10 wells Bis-Tris gels (Invitrogen). Running buffer and sample buffer (without G-250) were also from Invitrogen. Gels were run for 130 min at 150 V at 4 ºC, stained with Coomassie solution, and photographed with Gel Doc EZ System (Bio-Rad). Densitometry analysis of the protein bands was performed with Quantity One (Bio-Rad), and the measured intensities were fitted to a quadratic function to obtain a binding curve. 3-5 experiments were averaged to estimate KD values. Ternary mixtures NPM1/APE1/DNA were also analyzed by native electrophoresis using the same conditions, but in this case gels were also stained with GelRed (Biotium) prior to Coomassie. 1.5. Isothermal titration calorimetry (ITC). NPM1/APE1 binding was measured with a Nano-ITC (TA instruments) (Low Volume, with a cell of 189 L). Both proteins were dyalized in buffer 20 mM potassium phosphate, 50 mM NaCl, 5 mM MgCl2, 2 mM TCEP, pH 7.0, centrifuged and degassed prior to the measurement. 200-500 M APE1, was titrated onto NPM1, either wild type 9 or mutant A (5-30 M pentamer), in 33 injections of 1.5 L, at 20 ºC. Dilution controls were performed diluting APE1 into buffer. For baseline correction, the average of the heat of the 3 last injections was substracted from the isotherm. Isotherms were analyzed with the Nano Analyze software (TA), using a model of independent binding sites. 1.6. Circular dichroism (CD). Circular dichroism measurements were performed in a Jasco 720 spectropolarimeter equipped with Peltier temperature control, in a cuvette of 0.2 cm pathlength, in buffer 20 mM potassium phosphate pH 7.0, 50 mM NaCl, and either 0.5 mM EDTA or 5 mM MgCl2. Concentrations were 4 mM of NPM1, APE1 or oligo. Temperature scans were done at 60 ºC /h. 1.7. Fourier transform infrared (FT-IR) spectroscopy. For infrared spectroscopy, samples contained 100 M (or 50 M, in the mixtures with DNA) APE1, either alone, or in the presence of equimolar amounts of NPM1 pentamer or DNA oligo (“product”). The corresponding control samples of NPM1 and DNA were also measured. For H/D exchange, samples were subjected to 6-7 cycles of concentration with Amicon Ultra 0.5 mL filter units (Millipore) of 10 K cuttoff (except for DNA alone, where 3K was used), and dilution in D2O buffer (20 mM potassium phosphate, 50 mM NaCl, 5 mM MgCl2, 2 mM DTT, 0.01% (p/v) Tween-20, pD 7.0). The final concentration was checked by UV absorption and readjusted as necessary. Spectra were recorded in a Thermo Nicolet 5700 spectrometer equipped with a MCT detector using a Peltier-based temperature controller with 25 m-pathlength calcium fluoride cells (BioTools). A 25-μl sample aliquot was deposited on a cell that was sealed with a second cell. Typically 370 scans were collected for each background and sample, and the spectra were obtained with a nominal resolution of 2 cm−1. Temperature was increased at a rate of 1 °C min−1 between 20 and 80 °C for all samples. 1.8. APE1 incision assays. Incision of 5 M abasic DNA model oligonucleotide (Dumbbell S) was assayed in buffer 20 mM potassium phosphate, 50 mM NaCl, 5 mM MgCl2, 2 mM DTT, pH 7.0, in the presence or absence of 5 M NPM1, and monitorized by denaturing 18% polyacrylamide (PA)-urea gels (Adachi et al., 2015). The assay was done either as a function of APE1 concentration (0-200 pM) in reactions of 15 min at 37 ºC, or as a 16 accounting for 10-15% of the peptide bonds, might be thought to correspond to rearrangements of the N-terminal flexible region, not visible in the crystal structures. We have also analyzed the CD and FT-IR spectra as well as thermal stability with and without DNA of the truncated form of APE1, lacking the N-terminal 33 residues (APE1N33), which is catalytically active (Mol et al., 2000; Lirussi et al., 2012). Binding of APE1N33 to the oligo Dumbbell P also results in similar alterations of the Amide I band (Fig. 6B) and thermal stability, as followed either by ellipticity at 222 nm (not shown) or width of the Amide I band (Fig. S64). Therefore, the changes induced by DNA binding cannot be assigned to the N-terminal 33 residues, but perhaps to the following ca. 30 residues region, also flexible, and/or protein loops. Overall, and in line with previous reports (Yu et al., 2010) our data indicate that APE1 (either full length or isolated globular domain) conformation or dynamics is sensitive to abasic DNA binding. We have also explored the conformational consequences of APE1/NPM1 complex formation, using wild type APE1 and the truncated mutant NPM1C106, which as mentioned, is sufficient for binding, but lacks thermal transitions in the 20-80ºC range. At concentrations well above the KD, we could not observe any change neither in the spectral properties of both proteins, nor in the thermal profile of APE1 (not shown). This observation supports the notion that NPM1 binds mainly to the N-terminal segment of APE1, as described (Vascotto et al., 2009), having little effect on the conformation of APE1 globular, major domain. Finally, in ternary mixtures APE1/DNA/ NPM1C106 we could not observe any effect of NPM1 on the DNAbound APE1 conformation or on the stabilization brought about by DNA (not shown). 2.6. NPM1 favours the specific binding of APE1 to abasic substrate DNA. To get insight into the putative regulatory role of NPM1 on APE1 catalysis, we have analyzed the effect of NPM1 on the recognition and catalytic incision of an abasic DNA by APE1. First we have confirmed that our preparation of recombinant APE1 is highly active cleaving the model oligonucleotide of abasic DNA (Dumbbell S), as followed by denaturing polyacrylamide-urea DNA electrophosresis (Fig. 7A). In line with the reported stimulation of APE1 incision by NPM1 (Vascotto et al., 2009), but observing a greater effect than in their report, we have found a higher endonuclease activity in the presence of NPM1 (Fig. 7A). 17 In order to investigate the molecular basis of this activation, we have explored the effect of NPM1 on APE1 binding to the abasic DNA substrate. It had been previously described (Maher & Bloom, 2007; Freudenthal et al., 2015) that APE1 binds with high affinity abasic DNA and also binds, albeit with lower affinity, the product, nicked DNA, which has been related to a physiologically advantageous slow release of cytotoxic incised BER intermediates during DNA damage processing (Maher & Bloom, 2007). We have measured the binding by monitoring the increase in fluorescence anisotropy of fluorescein-labelled DNA upon addition of increasing amounts of APE1. The affinity we have determined for APE1 binding to the oligo Dumbbell S (KD 39 ± 8 nM (Fig. 7B,C) is similar to previously reported data with abasic-mimicking oligonucleotides (Hadi et al., 2000), but lower than that determined by Freudenthal et al. (2015) with the same DNA sequence that we have used. This could be explained because they use a buffer of lower ionic strength and/or we could not get reliable readings at a DNA concentration sufficiently below the KD. When the titration is performed in the presence of an excess of NPM1, no additional increase in the anisotropy is obtained, suggesting that NPM1 regulatory effect is not mediated by formation of a ternary complex DNA / APE1 / NPM1 (Fig. 7B). On the contrary, at high APE1 concentration, the maximum anisotropy value attained is lower in the presence of NPM1, suggesting that NPM1 can compete with the low-affinity, off-target binding of APE1 to DNA. However, APE1 affinity for the specific site on DNA (supposedly the abasic site-mimicking THF) is slightly enhanced in the presence of NPM1 (Fig. 7C), rendering a KD value of 25.3 ± 8 nM, as compared to 39 ± 8 nM in the absence of NPM1 (p value < 0.05). We have also explored the formation of complexes in ternary mixtures of APE1, DNA and NPM1 through native gel electrophoresis (Fig. 8). With equimolar concentrations of APE1 and the “product” abasic DNA (Dumbbell P), a defined band, probably corresponding to a 1:1 complex can be detected (Fig. 8A), while at higher APE1:DNA ratios (4:1) an heterogeneous species, putatively containing more molecules of APE1 bound to off-target sites is obtained (Fig. 8B). The addition of NPM1 partly competes with the heterogeneous, off-target binding, but not with the specific complex (Fig. 8A,B). On the contrary, when the different bands are analyzed by densitometry, it can be concluded that NPM1 favours the formation of the APE1:DNA (1:1) complex, either if its quantity is followed based on Coomassie staining (protein amount) or GelRed (DNA). Collectivelly, these results suggest that NPM1 may be able to favour the formation of a specific, discrete APE1:abasic DNA complex, opposing to alternative, 18 off-target APE1:DNA interactions. Apparently, this selection for the specific, productive binding results in a higher affinity for abasic DNA (Fig. 7C) and, ultimately, in the stimulation of APE1 activity (Fig. 7A). 3. Discussion Our results indicate that, under certain oxidative stress conditions, i.e. KBrO3 treatment, NPM1 is partly released from nucleoli and remains in chromatin regions (probably euchromatin), which might correspond to DNA damage repair platforms. Nucleoplasmic translocation of NPM1 upon a number of different stresses, including oxidative damage, has been previously reported (Colombo et al., 2011; Yang et al., 2016; Sekhar et al., 2014) and seems a common theme in the response to cellular perturbations of this and other nucleolar proteins (Antoniali et al., 2014; Scott & Oeffinger 2016). Moreover, NPM1 localization dynamics has been proposed to depend on posttranslational modifications (the protein is amenable to regulation by phosphorylation, ubiquitylation, sumoylation, acetylation, PARylation, etc). In this sense, we have observed that the oxidative treatment induces phosphorylation of NPM1 at least on T199. This modification perhaps reduces the affinity of the protein for Gquadruplex forming DNA sequences in the nucleolus, prompting nucleolar release upon DNA damage. Thus, phosphorylation of NPM1 would be related not only with DSBs repair (Koike et al., 2010), but also with repair of oxidative damage. Recently, glutathionylation of a Cys residue in the C-terminal domain of NPM1 has been described as a mechanism underlying nucleolar oxidation state sensing and release from nucleolus (Yang et al., 2016). The multiple posttranslational modifications affecting NPM1 might impact in different ways its localization dynamics and DDR-related activities. The exact role of NPM1 in BER is so far unknown. The participation of NPM1 in this and several other DDR pathways might be related to the histone chaperoning activity of the protein, since chromatin remodelling must be a requirement in all repair processes. On the other hand, the nucleolar enrichment in normal conditions of both NPM1 and APE1, and the fact that they are able to bind not only DNA but also RNA, and indeed display enzymatic activities on RNA, at least in vitro, has inspired the notion that they might collaborate in RNA quality control activities in addition to DDR (Antoniali et al., 2014). Moreover, our data point to a direct role of NPM1 in the BER machinery. The 19 similarity between the behaviour of NPM1 and APE1 when the cells are subjected to oxidation (dependence on KBrO3 concentration, localization within certain nuclear regions, and resistance to detergentsolubilization) strongly suggests that NPM1 could be recruited to the same structures on chromatin as APE1 (Campalans et al., 2013), i.e., platforms where active BER is taking place. If NPM1 is involved in BER regulation, the interaction between NPM1 and APE1 may be relevant as pharmacological target (Poletto et al., 2015), since interfering with DNA repair represents a cancer therapy strategy. An in vitro interaction between NPM1 and APE1 has been described implying the N-terminal, core domain of NPM1, and the flexible, N-terminal region of APE1 (Vascotto et al., 2009; Poletto et al., 2013). In our further characterization of the NPM1/APE1 association, we demonstrate a specific binding where part of the linker region of NPM1 (but not the region comprising the 106 last residues, including the C-terminal domain) is involved, explaining the ability of the AML leukaemia-related NPM1 mutant to also bind APE1 and apparently sequester it when aberrantly locating in the cytoplasm of AML cells (Vascotto et al., 2014, and our unpublished observations). In addition, we have found that more than one molecules of APE1 can bind to one NPM1 pentamer, and that the interaction is entropically driven. Although unfortunately we could not obtain evidences of an interaction between the two overexpressed proteins in cells, using the new “F2H” technology (from ChromoTek) (unpublished results), an in vivo association has been reported based on proximity ligation assay (PLA) (Poletto et al., 2013). It remains to be seen whether phosphorylation of NPM1 at Thr199, which according to our data correlates with nucleolar release, or further posttranslational modifications of either NPM1 or APE1 could modulate / enhance the association. NPM1 stimulates APE1 cleavage of abasic DNA (Fig. 7A, Vascotto et al., 2009), while interestingly, it has been described to inhibit APE1 endonuclease activity on RNA (Vascotto et al., 2009). The in vitro NPM1 regulatory effects, along with the similarity between the localization dynamics behaviour of both proteins (Fig. 1) suggest that NPM1 might participate in repair tasks on BER patches. Tell and co-workers have proposed that acetylation of Lys residues in APE1 would inhibit the association to NPM1 under stress conditions. Nevertheless, the same authors have reported “proximity ligation assay” (PLA) signal also outside the nucleoli; thus we think that an interaction between both proteins (even if the binding mode might differ) within the BER 20 machinery is feasible and could imply a modulation of APE1 functionality. The molecular basis of such regulation remains nevertheless elusive. We have found that NPM1 can compete with low-affinity, off-target APE1 binding to DNA, and could select for the specific, productive APE1/DNA complex. NPM1 interacts with the N-terminal flexible region of APE1 (Vascotto et al., 2009, and this study). This segment, ca. 40 residues long, has been related to regulation of DNA and /or RNA binding (Fantini et al., 2010; Poletto et al., 2013), and could be necessary for APE1 scanning of DNA to search target lesions, as disordered tails of other DNAbinding proteins (Vuzman et al., 2000). NPM1, by binding to APE1 N-terminal segment, could maintain APE1 in an open conformation, ready for productive association to abasic sites. This might be perhaps promoted under cellular conditions of heavy damage. NPM1 has been also proposed to regulate NF-B activity on DNA by keeping this protein in an open conformation (Lin et al., 2017). The modulation exerted by NPM1 resembles the role of other chaperones, in the sense that their activity is only evident in vitro at high chaperone: target protein ratios (Sharma et al., 2010), as it is the case in our experiments, where there is a significant molar excess of NPM1. We have shown that, upon binding to an abasic product-mimicking DNA, APE1 suffers conformational changes and becomes significantly stabilized, thus probably more compact. In line with this notion, it had been described, based on mass spectrometric protein footprinting, that binding to DNA after incision results in protection of some APE1 Lys residues (Yu et al., 2010). According to that study, APE1 conformation is still more “tensioned” when bound to the substrate, and then partly “relaxes” once the incision of the DNA has taken place, perhaps in order to facilitate discharge from DNA. Although in our experiments we could not demonstrate NPM1 competition with APE1 binding to the abasic product, NPM1, by grasping the N-terminal region of APE1 and acting on it as a lever, might somehow favour the release of APE1 from DNA once incised. This might be facilitated, in vivo, by subsequently intervening enzymes of the BER pathway, so that the cleaved intermediate is not released until it can be properly handled (Almeida and Sobol, 2007; Parsons and Dianov, 2013; Freudenthal et al., 2015). Then APE1 would be ready to dismount from DNA and scan chromatin for another abasic site. Other proteins participating in the BER route (XRCC1, Pol) have also been proposed to stimulate APE1 activity by releasing it from product inhibition (Vidal et al., 2001; Masuda et al., 1998). 21 4. Conclusions Nucleophosmin (NPM1) interacts with the repair enzyme APE1 and is able to stimulate in vitro APE1 incision of abasic DNA, suggesting that it could take part in the BER pathway. Our finding that, upon oxidative damage to cells, part of nucleophosmin is released from nucleoli (being correlated with phosphorylation at Thr199) and recruited to insoluble patches on chromatin, resembling APE1 behavior, further supports this hypothesis. In this study, we have further characterized the interaction between NPM1 and APE1, showing that pentameric NPM1 can bind several molecules of APE1 with submicromolar affinity. Not only NPM1 core domain and APE1 N-terminal region, as previously proposed, but also the proximal part of NPM1 linker region and the globular domain of APE1 are required for proper binding. The fact that the NPM1 C-terminal 106 residues are dispensable for the recognition explains why the AML leukemiaassociated NPM1 mutant keeps the ability to bind APE1, which might lead to APE1 mislocalization and pathogenic deregulation. We have also found that NPM1 is able to compete with APE1 off-target binding to DNA while favouring the specific, productive binding to the abasic site. This stimulatory effect, as well as a putative induction of APE1 release from the incised DNA, would positively regulate APE1 repair activity. In summary, our results depict a new, operational role of NPM1 in the modulation of BER machinery. Funding This work was funded by the Spanish Ministry of Economy and FEDER (grant SAF2014-57743-R), the Basque Government (grant IT709-13) and the University of the Basque Country (grant GIU18/172) Author’s contribution SB designed the research. All of the authors performed the experiments. DJL, AdB, IdlA, JM and SB analyzed and interpreted the data. S.B. wrote the paper with the contribution of DJL and MAU. Conflict of interest 22 The authors declare that they have no competing interests. Acknowledgements The authors thank the staff from the High Resolution Microscopy Facility (SGIkerUPV/EHU) for technical support, Dr. José Martínez (University of Granada) for advice on interpretation of ITC data, and Dr. José .Antonio. Rodríguez (University of the Basque Country) for suggestions and advice. A.dB. and M.G. were recipients of an Ikasiker fellowship (Basque Government). 23 References Adachi H, Yu YT (2014) Purification of radiolabeled RNA products using denaturing gel electrophoresis. Curr Protoc Mol Biol 105, doi: 10.1002/0471142727.mb0420s105. Almeida KH, Sobol RW. (2007) A unified view of base excision repair: Lesiondependent protein complexes regulated by post-translational modification. DNA Repair 6, 695-711, doi:10.1016/j.dnarep.2007.01.009 Antoniali G, Lirussi L, Poletto M, Tell G. (2014) Emerging roles of the nucleolus in regulating the DNA damage response: the noncanonical DNA repair enzyme APE1/Ref1 as a paradigmatical example. Antioxid Redox Signal. 2014 20(4):621-639, doi: 10.1089/ars.2013.5491 Arregi I, Falces J, Olazabal-Herrero A, Alonso-Mariño M, Taneva SG, Rodríguez JA, Urbaneja MA, Bañuelos S. (2015) Leukemia-Associated Mutations in Nucleophosmin Alter Recognition by CRM1: Molecular Basis of Aberrant Transport. PLoS One.: 10(6), doi: 10.1371/journal.pone.0130610. Ballmaier D and Epe B (2006) DNA damage by bromate: mechanism and consequences. Toxicology 221, 166-171, doi: 10.1016/j.tox.2006.01.009. Bañuelos S, Lectez B, Taneva SG, Ormaza G, Alonso-Mariño M, Calle X, Urbaneja MA. (2013) Recognition of intermolecular G-quadruplexes by full length nucleophosmin. Effect of a leukaemia-associated mutation. FEBS Lett. 587, 2254-2259, doi: 10.1016/j.febslet.2013.05.055. Boisvert FM, van Koningsbruggen S, Navascués J, Lamond AI (2007) The multifunctional nucleolus. Nat Rev Mol Cell Biol 8(7), 574-585, doi: 10.1083/jcb.200702147. Box JK, Paquet N, Adams MN, Boucher D, Bolderson E, O'Byrne KJ, Richard DJ. (2016) Nucleophosmin: from structure and function to disease development. BMC Mol Biol 17(1):19, doi: 10.1186/s12867-016-0073-9. Campalans A, Kortulewski T, Amouroux R, Menoni H, Vermeulen W, Radicella JP. (2013) Distinct spatiotemporal patterns and PARP dependence of XRCC1 recruitment to single-strand break and base excision repair. Nucleic Acids Res 41, 3115–3129, doi: 24 10.1093/nar/gkt025. Chan PK, Chan FY. (1995) Nucleophosmin/B23 (NPM) oligomer is a major and stable entity in HeLacells. Biochim Biophys Acta 1262, 37-42. Chiarella S, De Cola A, Scaglione GL, Carletti E, Graziano V, Barcaroli D, Lo Sterzo C, Di Matteo A, Di Ilio C, Falini B, Arcovito A, De Laurenzi V, Federici L. (2013) Nucleophosmin mutations alter its nucleolar localization by impairing G-quadruplex binding at ribosomal DNA. Nucleic Acids Res. 41(5):3228-3239, doi: 10.1093/nar/gkt001. Colombo E, Marine JC, Danovi D, Falini B, Pelicci PG (2002) Nucleophosmin regulates the stability and transcriptional activity of p53. Nat Cell Biol 4, 529–533, doi: 10.1038/ncb814. Colombo E, Bonetti, P, Denchi EL, Martinelli P, Zamponi R et al. (2005) Nucleophosmin is required for DNA integrity and p19Arf protein stability. Mol Cell Biol 25, 8874–8886, doi: 10.1128/MCB.25.20.8874-8886.2005. Colombo, E., Alcalay, M. & Pelicci, P. G. (2011) Nucleophosmin and its complex network: a possible therapeutic target in hematological diseases. Oncogene 30, 25952609, doi: 10.1038/onc.2010.646. Emmott E and Hiscox JA (2009) Nucleolar targeting: the hub of the matter. EMBO Rep 10(3), 231-238, doi: 10.1038/embor.2009. Falini B, Mecucci C, Tiacci E, Alcalay M, Rosati R et al. (2005) Cytoplasmic nucleophosmin in acute myelogenous leukemia with a normal karyotype. N Engl J Med 352: 254-266, doi: 10.1056/NEJMoa041974. Fantini D, Vascotto C, Marasco D, D'Ambrosio C, Romanello M, Vitagliano L, Pedone C, Poletto M, Cesaratto L, Quadrifoglio F, Scaloni A, Radicella JP, Tell G. (2010) Critical lysine residues within the overlooked N-terminal domain of human APE1 regulate its biological functions. Nucleic Acids Res. 38(22):8239-56, doi: 10.1007/s00018-010-0486-4. Federici L, Falini B (2013) Nucleophosmin mutations in acute myeloid leukemia: a tale of protein misfolding and mislocalization. Protein Sci 22: 545-556, doi: 25 10.1002/pro.2240. Freudenthal BD, Beard WA, Cuneo MJ, Dyrkheeva NS, Wilson SH. (2015) Capturing snapshots of APE1 processing DNA damage. Nat Struct Mol Biol 22: 924-931, doi: 10.1038/nsmb.3105. Gasteiger E, Hoogland C, Gattiker A, Duvaud S, Wilkins M.R, Appel R.D, Bairoch A. (2005) Protein Identification and Analysis Tools on the ExPASy Server. In: John M. Walker (ed): The Proteomics Protocols Handbook, Humana Press, 571-607. Grisendi S, Mecucci C, Falini B, Pandolfi PP. (2006) Nucleophosmin and cancer. Nat. Rev. Cancer 6, 493-505, doi: 10.1038/nrc1885. Grummitt CG, Townsley FM, Johnson CM, Warren AJ, Bycroft M (2008) Structural consequences of nucleophosmin mutations in acute myeloid leukemia. J. Biol. Chem. 283, 23326-23332, doi: 10.1074/jbc.M801706200. Guillonneau M, Paris F, Dutoit S, Estephan H, Bénéteau E, Huot J, Corre I (2016) Oxidative stress disassembles the p38/NPM/PP2A complex, which leads to modulation of nucleophosmin-mediated signaling to DNA damage response. FASEB J. 30(8):28992914, doi: 10.1096/fj.201500194R. Hadi MZ, Coleman MA, Fidelis K, Mohrenweiser HW and Wilson 3rd DM (2000) Functional characterization of Ape1 variants identified in the human population. Nucleic Acids Res. 28, 3871-3879, doi: 10.1093/nar/28.20.3871. He H, Chen Q, Georgiadis MM. (2014) High-resolution crystal structures reveal plasticity in the metal binding site of apurinic/apurymidinic endonuclease I Biochemistry 53(41), 6520-6529, doi: 10.1021/bi500676p. Jackson EB, Theriot CA, Chattopadhyay R, Mitra S, Izumi T. (2005) Analysis of nuclear transport signals in the human apurinic/apyrimidinic endonuclease (APE1/Ref1) Nucleic Acids Res. 33(10), 3303-3312, doi: 10.1093/nar/gki641. Kennedy MJ, Hughes RM, Peteya LA, Schwartz JW, Ehlers MD, Tucker CL. (2010) Rapid blue-light-mediated induction of protein interactions in living cells. Nat Methods 7(12), 973-975, doi: 10.1038/nmeth.1524. Koike A, Nishikawa H, Wu W, Okada Y, Venkitaraman AR et al. (2010) Recruitment of 32 Figure 5. Comparison of APE1 binding by wild type NPM1, NPM1MutA, NPM1ΔC106 and NPM1 core. (A) Binding curves based on native electrophoresis assays with 0.5 M (NPM1core) or 0.1 M pentamer (for the rest of NPM1 variants). The corresponding fittings to a simple binding model are also shown, with indication of the estimated KD values. (B) Comparison of NPM1 binding by wild type APE1 and APE1N33. Figure 6. (A) Binding to magnesium and/or abasic DNA conformationally stabilizes APE1. Thermal scans based on ellipticity at 222 nm of 4 M apo APE1 (0.5 mM EDTA) (black), holo APE1 (5 mM MgCl2) (red), apo APE1 in the presence of equimolar amount of oligonuleotide “Dumbbell P” (dark blue) and holo APE1 plus DNA (green). Scans were corrected for the DNA contribution. The heating rate was 1ºC/min. (B) Secondary structure of APE1 and APE1ΔN33 is sensitive to DNA binding. Amide I band of the infrared spectra of APE1 (solid line) and APE1N33 (broken line) in the absence (blue) and presence (red) of equimolar amounts of the oligo “Dumbbell P”. The spectra were recorded in D2O buffer at 20ºC. The contribution of buffer was substracted and the baseline was corrected between 1700 and 1600 cm-1. Figure 7. NPM1 stimulates APE1 catalyzed incision of abasic DNA, and favors specific binding of APE1 to the substrate. (A) Incision of an abasicmimicking oligonucleotide (5 M) by increasing concentrations of APE1 as followed by DNA urea-polyacrylamide denaturing electrophoresis, in the absence (empty symbols) and in the presence (solid symbols) of 5 M NPM1. (B) Binding of 0 – 10 M APE1 to 10 nM Dumbbell S in the absence (empty) or presence (solid) of 5 M NPM1, based on fluorescence anisotropy. (C) The same binding data of panel B shown in the range 0-1 M APE1 and logarithmic scale and with the corresponding fitted curves (pvalue < 0.05). Data are averages of three experiments, bars indicating standard deviation. Figure 8. Effect of NPM1 on APE1/DNA complex formation as followed by native electrophoresis. (A) Native 4-16% PA gel of mixtures of 5 M oligo (O), 2 M APE1 (A) and 0-10 M NPM1 (N) (pentamer), along with the corresponding individual species and binary mixtures as controls. The lanes corresponding to the ternary mixtures are labelled according to NPM1 concentration. Images of the gel stained with Coomassie for detection of proteins (blue) and GelRed for DNA (white) are overlaid. (B) Native PAGE of a mixture of 5 M oligo and 20 M APE1 in the absence (OA) and 33 in the presence (OAN) of 1 M NPM1. The gel was stained as in (A). (C) Densitometry of the band corresponding to the APE1:DNA (1:1) complex stained with Coomassie (squares) or GelRed (triangles) as a function of NPM1 concentration. 34 Fig. 1 35 Fig. 2 36 Fig. 3 37 Fig. 4. 38 Fig. 5. 39 Fig. 6 40 Fig. 7 A 🞅 - NPM1 B 🞅 - NPM1 C 🞅 - NPM1 41 Fig. 8