Summary
The BRCA2 tumor suppressor protein is involved in the maintenance of genome integrity through its role in homologous recombination. In mitosis, BRCA2 is phosphorylated by Polo-like kinase 1 (PLK1). Here we describe how this phosphorylation contributes to the control of mitosis. We identified two highly conserved phosphorylation sites at S193 and T207 of BRCA2. Phosphorylated-T207 is a bona fide docking site for PLK1 as illustrated by the crystal structure of the BRCA2 peptide bound to PLK1 Polo-box domain. We find that BRCA2 is in complex with PLK1, phosphorylated-BUBR1 and the phosphatase PP2A. Precluding BRCA2 binding to PLK1, as observed in BRCA2 breast cancer variants, alters the tetrameric complex resulting in mitotic delay, misaligned chromosomes, faulty chromosome segregation and aneuploidy. We thus reveal a direct role of BRCA2 in the alignment of chromosomes, distinct from its DNA repair function, with important consequences on chromosome stability. These findings may explain in part the aneuploidy observed in BRCA2-mutated tumors.
Introduction
The BRCA2 tumor suppressor protein plays an important role in DNA repair by homologous recombination (HR) (Moynahan et al, 2001; Jensen et al, 2010), which takes place preferentially during S/G2 phases of the cell cycle (Saleh-Gohari & Helleday, 2004). BRCA2 has also emerging functions in mitosis. For example, at the kinetochore, BRCA2 binds to BUBR1 (Futamura et al, 2000), an essential factor for the faithful segregation of chromosomes. BUBR1 is required for kinetochore-microtubule attachment. It is also a component of the spindle assembly checkpoint (SAC), which ensures chromosome segregation by blocking anaphase entry until all chromosomes align to the metaphase plate (Lampson & Kapoor, 2005; Lara-Gonzalez et al, 2012). These two activities of BUBR1 involve different partners and are functionally distinct (Elowe et al, 2010; Zhang et al, 2016). BRCA2 has been reported to promote the acetylation of BUBR1, thus facilitating SAC activation (Choi et al, 2012; Park et al, 2017). At the end of mitosis, BRCA2 localizes to the midbody and assists cell division by serving as a scaffold protein for the central spindle components (Mondal et al, 2012; Daniels et al, 2004; Takaoka et al, 2014). BRCA2 is phosphorylated by PLK1 both in its N-terminal region (Lin et al, 2003) and in its central region (Lee et al, 2004), although the functional role of these phosphorylation events remains unclear.
PLK1 is a master regulator of the cell cycle, especially in mitosis (Zitouni et al, 2014; Barr et al, 2004). Among other functions, PLK1 phosphorylates BUBR1 at several residues including the tension-sensitive sites S676 (Elowe et al, 2007) and T680 (Suijkerbuijk et al, 2012) in prometaphase allowing the formation of stable kinetochore-microtubule attachments. This activity needs to be tightly regulated to ensure proper alignment of the chromosomes at the metaphase plate (Elowe et al, 2007; 2010; Zhang et al, 2016); in particular, Aurora B is an essential kinase that destabilizes erroneous kinetochore-microtubule interactions (Hauf et al, 2003). The phosphatase PP2A protects initial kinetochore-microtubule interactions from excessive destabilization by Aurora B (Foley et al, 2011) and this function is achieved through interaction with BUBR1 phosphorylated in its Kinetochore Attachment and Regulatory Domain (KARD) motif (including residues S676 and T680) (Suijkerbuijk et al, 2012). Thus, BUBR1, PLK1, Aurora B and PP2A are necessary for the formation of stable kinetochore-microtubule attachments.
PLK1 is recruited to specific targets via its Polo-box domain (PBD) (Elia et al, 2003a). PBD interacts with phosphosites characterized by the consensus motif S-[pS/pT]-P/X (Elia et al, 2003b). These phosphosites are provided by a priming phosphorylation event, usually mediated by CDK1 or other proline-directed kinases (Barr et al, 2004); however, there is also evidence that PLK1 itself might create its docking sites (“self-priming”) during anaphase (Neef et al, 2003; Kang et al, 2006).
Several BRCA2 sites corresponding to PLK1 consensus motifs have been identified as phosphorylated in mitosis, some of which belong to a cluster of predicted phosphosites located in BRCA2 N-terminus around residue S193 (Lin et al, 2003). We set out to investigate which of these sites are phosphorylated by PLK1, and to reveal whether these phosphorylation events play a role in the regulation of mitotic progression. Here, we identify S193 and T207 as targets for phosphorylation by PLK1 and reveal that BRCA2 phosphorylated at T207 is a bona fide docking site for PLK1. By investigating the phenotype of BRCA2 missense variants S206C and T207A that limit phosphorylation of BRCA2 T207 and PLK1 binding, we reveal an unexpected role of BRCA2 in the alignment of chromosomes at the metaphase plate; we find that BRCA2 facilitates the phosphorylation of BUBR1 by PLK1, thus contributing to the establishment of kinetochore-microtubule attachments, which is functionally distinct from BRCA2 reported role at the SAC (Choi et al., 2012). A defect in this function of BRCA2 leads to mitotic delay and chromosome instability manifested in chromosome misalignment, chromosome segregation errors and aneuploidy.
Results
BRCA2 variants identified in breast cancer reduce the PLK1-dependent phosphorylation of BRCA2 N-terminal region
Several missense variants of uncertain significance (VUS) identified in BRCA2 in breast cancer patients are located in the N-terminal region predicted to be phosphorylated by PLK1 (around S193) (Breast information core (BIC)(Szabo et al, 2000) and BRCAShare (Béroud et al, 2016)), summarized in Table EV1. To find out if any of these variants affected PLK1 phosphorylation in this region, we purified fragments comprising amino acids 1 to 250 of BRCA2 (hereafter BRCA21-250) from human embryonic kidney cells (HEK293T) and used an in vitro kinase assay to assess the phosphorylation by PLK1 of the fragments containing either the WT sequence, the different BRCA2 variants M192T, S196N, T200K, S206C and T207A, or the mutant S193A, previously reported to reduce the phosphorylation of BRCA2 by PLK1 (Lin et al, 2003). As expected, S193A reduced the phosphorylation of BRCA21-250 by PLK1 (Fig. 1A, 1B). Interestingly, variants T207A, S206C and T200K also led to a 2-fold decrease in PLK1 phosphorylation of BRCA21-250 (Fig. 1A, 1B). In contrast, M192T slightly increased the phosphorylation above WT levels whereas VUS S196N did not significantly modify the phosphorylation of BRCA21-250 by PLK1 (Fig. 1A, 1B). The phosphorylation observed in BRCA2 fragments is specific of the recombinant PLK1 kinase as replacing the WT PLK1 by a kinase-dead (PLK1-KD) version of the protein (K82R) (Golsteyn et al, 1995) purified using the same protocol, or adding a PLK1 inhibitor (BI2536) to the reaction, abolished the phosphorylation of BRCA21-250 (Fig. 1C, lanes 4 and 5 compared to lane 3; Fig. 1D). Moreover, running a kinase assay with BRCA21-250 alone did not result in any signal excluding a kinase activity coming from a possible contaminant kinase in the preparation of BRCA21-250 produced in human cells, (Fig. EV1A, 1B).
Together, these results show that VUS T207A, S206C and T200K identified in breast cancer patients impair phosphorylation of BRCA21-250 by PLK1 in vitro.
BRCA2T207 is a target of phosphorylation by PLK1
The reduction of BRCA2 phosphorylation in BRCA21-250 containing T207A and S206C variants suggested that T207 could be a target of PLK1 phosphorylation. We investigated this possibility by following the PLK1 phosphorylation kinetics of a truncated fragment of BRCA2 N-terminus comprising T207 (amino acids 190-283) (hereafter BRCA2190-283) by Nuclear Magnetic Resonance (NMR) spectroscopy. NMR analysis allows residue-specific quantification of a peptide modification, provided it is 13C or 15N-labelled. Fig. 2A shows superimposed 1H-15N HSQC spectra of BRCA2190-283 at different time points of the phosphorylation reaction with recombinant PLK1. Analysis of these experiments revealed phosphorylation of S193 and of eight other phosphosites, among which three threonines including T207 and five serines (Fig. 2A). Four residues (T226>T207>T219>S193, Fig. 2B, Fig. EV1C) were favoured phosphorylation sites, modified with markedly faster rates than the five other sites (S197, S217, S231, S239, S273). Interestingly, while T219 and T226 conservation is poor, T207 and S193 are conserved from mammals to fishes (Fig. EV1D and EV1E) suggesting that both are important for BRCA2 function.
T207 phosphorylation together with the presence of a serine residue at position 206 creates a predicted docking site for PLK1PBD (Elia et al, 2003b). Hence, we conclude that T207 is phosphorylated in vitro by PLK1, this phosphorylation is efficient in the context of BRCA2190-283, and it may be a primary event for further BRCA2 phosphorylation via the recruitment of PLK1 to BRCA2.
BRCA2 variants T207A and T200K alter the phosphorylation kinetics by PLK1
Having identified T207 as a target of phosphorylation of PLK1, we next compared the residue-specific phosphorylation kinetics in the polypeptide WT BRCA2190-283 containing the variants T207A or T200K that displayed reduced overall phosphorylation (Fig. 1A, 1B). (The production of a 15N recombinant fragment BRCA2190-283 comprising S206C from bacteria yielded an insoluble protein precluding NMR analysis). Time-resolved NMR experiments revealed that PLK1 phosphorylates significantly less BRCA2190-283 containing the variant T207A than the WT peptide (Fig. 2C, 2D). Initial phosphorylation rates were decreased by a factor 5 (S193), 8 (T226) and 13 (T219) (Fig. 2C, 2D, Fig. EV2A). Variant T200K reduced by half the phosphorylation rates of S193, T207, T219 and T226 (Fig. 2E, 2F and Fig. EV2B). In agreement with the in vitro kinase assay using the BRCA21-250 fragment purified from human cells (Fig. 1), these results show that in BRCA2190-283, variants T207A and T200K impair the phosphorylation of T207 and the cascade of associated phosphorylation events.
Variants T207A, S206C and T200K reduce the interaction of BRCA2 and PLK1
The finding that T207 is efficiently phosphorylated by PLK1 in BRCA2190-283 polypeptide (Fig. 2A) together with the observation that T207A mutation causes a global decrease in the phosphorylation of this region (Fig. 2C) and the prediction that T207 is a docking site for PLK1PBD binding (Elia et al, 2003b) made us hypothesize that T207 might be a “self-priming” phosphorylation event required for the interaction of PLK1 with BRCA2 at this site. If so, the variants that reduce phosphorylation of T207 by PLK1 would be predicted to alter PLK1PBD binding. To test this hypothesis, we examined the interaction of PLK1 with the VUS-containing polypeptides. We overexpressed 2xMBP-BRCA21-250 constructs carrying these variants in U2OS cells to detect the endogenous PLK1 that co-purifies with 2xMBP-BRCA21-250 using amylose pull-down. As expected, overexpressed BRCA21-250 was able to interact with endogenous PLK1 from mitotic cells but not from asynchronous cells (predominantly in G1/S) where the levels of PLK1 are reduced (Fig. 3A, lane 2 compared to lane 1). As previously described (Lin et al, 2003), mutation S193A reduced the binding to PLK1 (Fig. 3A, lane 6 compared to lane 2 and Fig. 3B). Interestingly, the variants reducing PLK1 phosphorylation (T207A, S206C and T200K) showed a weaker interaction with PLK1 than the WT protein (Fig. 3A, pull-down lane 4, 8 compared to lane 2 and lane 20 compared to lane 18, Fig. 3B) despite the protein levels of PLK1 remaining unchanged (Fig. 3A, compare PLK1 input lanes 4 and 8 to lane 2 and lane 20 to lane 18). In contrast, the effect of M192T and S196N on the interaction was mild (Fig. 3A, compare pull-down lanes 12 and 14 to lane 10, Fig. 3B). These results are consistent with the idea of a self-priming phosphorylation by PLK1 on T207.
To provide further evidence that the PLK1-mediated phosphorylation of BRCA2 favors BRCA2 binding, we performed an in vitro kinase assay followed by an amylose pull-down with recombinant proteins, and eluted the proteins with maltose. PLK1 was found in the maltose elution with WT-BRCA21-250 demonstrating that PLK1-phosphorylated BRCA21-250 binds to PLK1 (Fig. 3C lane 4, Fig. 3D). In contrast, the fraction of PLK1 in the eluate of T207A-BRCA21-250 was substantially reduced (Fig. 3C, lane 6 compared to lane 4, Fig. 3D) indicating that the phosphorylation of T207 is required for the efficient binding to PLK1 and confirming our results with cell lysates (Fig. 3A, 3B).
T207 is a bona fide docking site for PLK1
To directly demonstrate the recognition of pT207 by PLK1, we measured the affinity of recombinant PLK1PBD for a synthetic 17 aa peptide comprising phosphorylated T207. Using isothermal titration calorimetry (ITC), we found that recombinant PLK1PBD bound to the T207 phosphorylated peptide with an affinity of Kd= 0.09 ± 0.01 µM (Fig. 3E), similar to the optimal affinity reported for an interaction between PLK1PBD and its phosphorylated target (Elia et al, 2003b). Consistently, PLK1PBD bound to the fragment BRCA2190-283 with nanomolar affinity upon phosphorylation by PLK1 (Kd= 0.14 ± 0.02 µM; Fig. EV3A) whereas it did not bind to the corresponding non-phosphorylated polypeptides (Fig. 3F, Fig. EV3B). Mutation T207A also abolished the interaction (Fig. 3G), in agreement with the pull-down experiments (Fig. 3A-3D). A peptide comprising pT207 and S206C mutation could not bind to PLK1PBD (Fig. 3H), as predicted from the consensus sequence requirement for PLK1PBD interaction (Elia et al, 2003b). Last, a peptide containing phosphorylated S197, which is also a predicted docking site for PLK1, bound with much less affinity to PLK1PBD than pT207 (Kd= 17 ± 2 µM; Fig. EV3C).
To further characterize this molecular interaction, we determined the crystal structure of PLK1PBD bound to the T207 phosphorylated peptide at 3.1Å resolution (Table EV2). Analysis of this 3D structure showed that, as expected, the 17 aa BRCA2 phosphopeptide binds in the cleft formed between the two Polo boxes (Fig. 3I). Twelve residues of the peptide (from A199 to I210) are well-structured upon binding, burying about 694 Å2 in the interface with PLK1PBD. All BRCA2 residues from P202 to L209 are at least 25% buried in the complex. The interface is stabilized by 12 hydrogen bonds: the backbone of residues T200 to L209 as well as the side chain of S206 are bonded to residues from Polo Box 1, whereas the side chain of phosphorylated T207 is bonded to residues from Polo Box 2 (see the zoom view in Fig. 3I). Specifically, the backbone oxygen of T200 is bonded to the side chain of Y417 of PLK1PBD. Residues L204, S205 and S206 form a β-sheet with residues W414, V415 and D416, characterized by hydrogen bonds between the backbone atoms of BRCA2 L204, S206 and PLK1 D416, W414, respectively. The side chain of S206 participates in 2 hydrogen-bonding interactions with the backbone of W414, which explains the strict requirement for this amino acid at this position (Elia et al, 2003b). Moreover, the phosphate group of pT207 participates in 3 hydrogen-bonding interactions with the side chains of residues H538, K540 and R557 in Polo Box 2 (see the zoom view in Fig. 3I). This explains the critical dependence on phosphorylation for binding observed by ITC (Fig. 3F-3H). The presence of a buried leucine at the pT-3 position (Yun et al, 2009), as well as the electrostatic interactions of the serine at the pT-1 position with PLK1 W414 (Elia et al, 2003b) and the phosphorylated threonine with PLK1 H538 and K540, have been described as essential for the high affinity and specificity of the interaction in other PLK1PBD- phosphopeptide complexes (Elia et al, 2003b; García-Alvarez et al, 2007).
Thus, our biochemical and structural analysis demonstrate that the BRCA2 T207 phosphopeptide interacts with PLK1PBD as an optimal and specific PLK1PBD ligand. It supports a mechanism in which phosphorylation of T207 by PLK1 promotes the interaction of PLK1 with BRCA2 through a bona fide docking site for PLK1 and favours a cascade of phosphorylation events. In variants T200K and T207A the decrease in T207 phosphorylation impairs PLK1 docking at T207 explaining the reduction of binding to PLK1 and the global loss of phosphorylation by PLK1. S206C eliminates the serine residue at −1 position required for PLK1PBD interaction resulting as well in a reduction of BRCA2 binding.
Impairing T207 phosphorylation prolongs mitosis
PLK1 is a master regulator of mitosis (Barr et al, 2004). To find out whether the interaction between BRCA2 and PLK1 is involved in the control of mitotic progression we examined the functional impact of two of the variants that reduce PLK1 phosphorylation at T207 (S206C and T207A) in the context of the full-length BRCA2 protein in cells. For this purpose, we generated stable cell lines expressing the BRCA2 cDNA coding for either the GFPMBP-BRCA2 WT or the variants to complement DLD1 BRCA2 deficient human cells (hereafter BRCA2-/-). In this cell line, both alleles of BRCA2 contain a deletion in exon 11 causing a premature stop codon after BRC5 and cytoplasmic localization of a truncated form of the protein (Hucl et al, 2008). We selected two stable clones of each variant that show similar protein levels as the BRCA2 WT complemented cells (clone C1, hereafter BRCA2 WT) by western blot (Fig. EV4). We then tested the interaction of full-length BRCA2 with PLK1 in these stable clones by GFP pull-down. As expected, PLK1 readily co-purified with full-length BRCA2 WT from mitotic cells. Importantly, in cells expressing the variants S206C and T207A the level of co-purified PLK1 was greatly reduced (Fig. 4A, 4B) confirming the results obtained with the overexpressed BRCA21-250 fragments (Fig. 3A) now in the context of cells stably expressing the full-length BRCA2 protein. Thus, BRCA2 interaction with endogenous PLK1 is impaired in cells bearing variants S206C and T207A.
To examine the impact of BRCA2 variants on mitosis, we monitored the time individual cells spent in mitosis, from mitotic entry (defined as nuclear envelope break down) to complete daughter cell separation using live cell imaging. Cells expressing the endogenous BRCA2 (hereafter BRCA2+/+) and the BRCA2 WT cells showed similar kinetics, the majority of the cells (80% for BRCA2+/+ and 82% for BRCA2 WT) having completed mitosis within 60 min (Fig. 4C, D). In contrast, cells expressing variants S206C and T207A augmented the time spent in mitosis as manifested by a significant decrease in the frequency of cells dividing within 60 min (∼49-51%), slightly higher than the BRCA2-/- cells (38%) (Fig. 4C, D). The percentage of cell death during mitosis also increased in the mutated population compared to the BRCA2 WT complemented cells although the difference was not significant (Fig. 4C, D). Representative videos of the still images shown in Fig. 4C are included in Expanded View (movies EV1-EV4).
To find out the stage of mitosis at which the mutated cells were arrested we monitored the time elapsed from chromosome condensation to anaphase onset in cells labelled with SiR-DNA that allows visualizing the DNA in living cells. Interestingly, cells expressing S206C and T207A variants required significantly more time to reach anaphase (between 131 and 139 min) compared to the BRCA2 WT complemented cells (80 min) or BRCA2+/+ cells (77 min) (Fig. 4E, F). The BRCA2-/- also required more time on average (175 min). Representative videos of the still images shown in Fig. 4E are included in Expanded View (Movies EV5-EV7).
Taken together, cells expressing variants S206C and T207A display a strong delay in mitotic progression, specifically, from chromosome condensation to anaphase onset, compared to BRCA2 WT expressing cells.
Docking of PLK1 at T207 of BRCA2 is required for the phosphorylation of BUBR1 at S676 and T680 by PLK1 and chromosome alignment
BRCA2 interacts directly with BUBR1 through its C-terminal region (Futamura et al, 2000; Choi et al, 2012). BUBR1 associates with and is phosphorylated by PLK1, which controls the stability of kinetochore-microtubule interactions and enables the alignment of chromosomes at the metaphase plate (Elowe et al, 2007; Suijkerbuijk et al, 2012). The strong delay in mitotic progression observed in the clones expressing S206C and T207A (Fig. 4C-F) led us to investigate BUBR1 levels in these cell lines. As previously described (Elowe et al, 2007; Huang et al, 2008), upon cell arrest in G2/M phase by Nocodazole treatment, BUBR1 displayed 2 bands, the non-modified BUBR1 and the up-shifted band corresponding to phosphorylated BUBR1 (Fig. 5A). Interestingly, the up-shifted band of BUBR1 strongly decreased in the cells expressing variants S206C or T207A compared to the cells expressing BRCA2 WT (Fig. 5A, lane 3-6 compared to lane 2) and the same trend was observed in BRCA2-/- cells (Fig. 5A, lane 8 compared to lanes 7 and 10). To find out the species of BUBR1 altered in the stable cells expressing these BRCA2 variants, we probed the same membrane with an antibody specific for pT680, a known BUBR1 phosphosite target of PLK1 (Suijkerbuijk et al, 2012). Interestingly, both clones of the cells expressing variants S206C and T207A displayed reduced pT680 levels of BUBR1 (Fig. 5A). As previously described (Suijkerbuijk et al, 2012), the phosphorylation of T680 is mediated by PLK1 since PLK1 inhibitors strongly reduced this signal (Fig. EV5A) and phosphatase treatment resulted in a decrease of the intensity of the band (Fig. EV5B) confirming the specificity of the anti-pT680-BUBR1 antibody in our cells. Furthermore, we observed a reduction in the PLK1-dependent phosphorylation of S676, another known target of PLK1 in BUBR1 (Elowe et al, 2007) (Fig. 5B), indicating that BUBR1 phosphorylation by PLK1 is impaired in cells expressing BRCA2 variants S206C and T207A, at least at these two sites.
BUBR1 facilitates kinetochore-microtubule attachments via its interaction with the phosphatase PP2A which requires the phosphorylation of BUBR1 at the KARD motif comprising S676 and T680 (Suijkerbuijk et al, 2012). So, we next tested whether BRCA2 and PLK1 formed a tetrameric complex with pT680-BUBR1 and PP2A. Interestingly, using a GFP pull-down to capture GFPMBP-BRCA2 from BRCA2 WT complemented mitotic cells BRCA2 pulled-down PLK1, pT680-BUBR1 and PP2A (detected with an antibody against the catalytic subunit of PP2A, PP2AC), confirming the formation of a tetrameric complex. Importantly, cells expressing the variants S206C or T207A showed a strong reduction in the interaction of BRCA2 with PLK1, pT680-BUBR1 and PP2A in the context of the tetrameric complex (Fig. 5C, D). To find out whether the binding of total BUBR1 to PP2A was also affected in these cells we immunoprecipitated BUBR1 from mitotic cells and detected the levels of PP2A (PP2AC antibody). Interestingly, although PP2A was readily copurified with BUBR1 in the BRCA2 WT cells, expressing BRCA2 variant T207A reduced the levels of PP2A to nearly half (Fig. 5E, F). These effects were not due to a reduced localization of PLK1 or PP2A to the kinetochores in these cells, as the levels of these proteins at the kinetochores remained unchanged (Fig. EV6A-D); nor to a reduced binding of BRCA2 to total BUBR1 as T207A did not alter the interaction of BRCA2 with total BUBR1 (Fig. EV6 E, F).
The phosphorylation of BUBR1 at S676 and T680 by PLK1 and its association with PP2A are required for the formation of stable kinetochore-microtubule attachments (Suijkerbuijk et al, 2012; Elowe et al, 2007), a defect in this interaction results in chromosome misalignments. Therefore, we next examined whether cells expressing the BRCA2 variants S206C and T207A altered chromosome alignment. Following thymidine synchronization, the cells were treated with Monastrol (Eg5 inhibitor) at a dose (100 μM) that for these cells was sufficient to enrich the population in prometaphase cells without precluding the assembly of bipolar spindles (as revealed by the microtubule marker α-tubulin). A subsequent treatment with the proteasome inhibitor MG132 for 1h was used to avoid exit from mitosis (Elowe et al, 2007). Chromosome alignment was then analysed by immunofluorescence. Strikingly, analysis of cells expressing S206C and T207A variants showed high frequency of faulty chromosome congression compared to the BRCA2 WT clone (52% in S206C A9, 42% in T207 B1 versus 16% in the BRCA2 WT clone) (Fig. 5G, H) as detected by a signal of the centromere marker (CREST) outside the metaphase plate (Fig. 5H). Interestingly, most of the misaligned chromosomes were located close to the spindle pole as revealed by the colocalization of CREST with the microtubule marker α-tubulin (Fig. 5G, H). Similar results were obtained when the cells were released from Monastrol treatment before adding the proteasome inhibitor MG132 (Fig. EV7).
Together, our results show that the level of phosphorylation of BUBR1 by PLK1 is reduced in cells expressing the variants S206C and T207A suggesting that PLK1 needs to be bound to BRCA2 to efficiently phosphorylate BUBR1. Moreover, BRCA2 forms a tetrameric complex with pT680-BUBR1, PLK1 and PP2A. This complex is strongly impaired in cells expressing variants S206C and T207A. The reduced interaction of pBUBR1 with PP2A, which is normally required for the establishment of proper kinetochore-microtubule attachment, causes chromosome misalignments.
The variants that reduce PLK1 phosphorylation of BRCA2 display strong defects in chromosome segregation and aneuploidy
Unresolved chromosome misalignment as observed in cells altering BRCA2 phosphorylation by PLK1 is expected to drive chromosome missegregation. To find out if this was the case in cells expressing BRCA2 variants S206C and T207A, we examined chromosome segregation by immunofluorescence in cells synchronized by double-thymidine block and released for 15h to enrich the cell population at anaphase/telophase stage. Importantly, 49% and 36% of cells expressing S206C and T207A, respectively, exhibited faulty chromosome segregation including lagging chromosomes and chromosome bridges as compared to 20% of the cells expressing BRCA2 WT (Fig. 6A, B).
Erroneous chromosome segregation generates aneuploid cells during cell division (Santaguida & Amon, 2015). Given the strong chromosome segregation defects observed in cells expressing S206C and T207A, we next analysed the number of chromosomes in these cells. Total chromosome counts carried out on metaphase spreads revealed that 37.1% of BRCA2 WT cells exhibited aneuploidy with chromosome losses or gains. In the case of S206C A7 and T207A B1, this number was elevated to 52.2% and 61.8% of the cells, respectively (Fig. 7A). An example of the images analyzed can be found in Fig. 7B. As the number of chromosomes was difficult to assess for cells with high content of chromosome gains we arbitrarily discarded cells that contained more than 65 chromosomes. Thus, tetraploid cells were not included in this measurement. Therefore, as a complementary test, we determined the frequency of tetraploid cells by assessing the incorporation of BrdU of asynchronous cell populations and measuring the frequency of S-phase cells with >4N DNA content (Fig. 7C). The quantification of these data show that similar to the BRCA2 WT complemented cells, the frequency of tetraploidy in cells bearing the variants is <1% of the total population (Fig. 7D) while the frequency of BrdU positive cells was equivalent between the BRCA2 WT and the VUS expressing cells (Fig. EV8A-D).
Together, these results indicate that, in addition to the severe chromosome misalignment phenotype, cells expressing S206C and T207A display high frequency of chromosome missegregation. As a consequence, the incidence of aneuploidy, but not tetraploidy, is greatly exacerbated in these cells.
The variants altering PLK1 phosphorylation of BRCA2 restore the hypersensitivity of BRCA2 deficient cells to DNA damage and PARP inhibition and are HR proficient
Since BRCA2 has a major role in DNA repair by HR, the prolonged mitosis observed in the VUS-expressing stable cell lines (Fig. 4) could result from checkpoint activation through unrepaired DNA. In addition, defects in chromosome segregation have been reported as a cause of DNA damage (Janssen et al, 2011). Thus, to test the DNA repair efficiency of cells expressing variants S206C and T207A, we performed a clonogenic survival assay in the stable clones after treatment with mitomycin C (MMC), an inter-strand crosslinking agent to which BRCA2 deficient cells are highly sensitive (Kraakman-van der Zwet et al, 2002). As expected, BRCA2 deficient cells (BRCA2-/-) showed hypersensitivity to MMC treatment whereas BRCA2 WT cells complemented this phenotype almost to the same survival levels as the cells expressing the endogenous BRCA2 (BRCA2+/+). The stable clones expressing variants S206C and T207A also complemented the hypersensitive phenotype of BRCA2-/- cells reaching similar levels as the BRCA2 WT cells (Fig. 8A) suggesting that the delay in mitosis is not a consequence of checkpoint activation via unrepaired DNA. Cells expressing VUS S206C and T207A showed a growth defect manifested in a reduced number of colonies observed in unchallenged conditions (Fig. EV9A), which is consistent with the mitotic phenotype observed (Fig. 4-6). To exclude a possible bias arising from the different ability to form colonies of these cells we also tested the sensitivity to MMC of the cell population by MTT assay. As shown in Fig. 8B, cells expressing S206C and T207A showed similar relative viability compared to BRCA2 WT complemented cells or the cells expressing endogenous BRCA2 (BRCA2+/+), confirming our results. To additionally address the DNA repair capacity of these cells, we tested their viability upon treatment with the poly (ADP-ribose) polymerase (PARP) inhibitor Olaparib. PARP1 is an enzyme required for the sensing of DNA single strand breaks (SSBs) and double strand breaks (DSBs) that becomes essential in the absence of a functional HR pathway (Bryant et al, 2005) and therefore is used as a surrogate of HR proficiency. In fact, PARP1 inhibitors, in particular Olaparib, are currently used in the clinic to treat breast and ovarian cancer patients carrying germline mutations in BRCA1/2. In our settings, the relative viability of BRCA2-/- cells was 45% upon 4-day treatment with the highest Olaparib concentration tested (5 µM); in contrast, 74% of BRCA2 WT complemented cells remained viable. Similarly, cells expressing S206C or T207A survived the treatment equally well as the cells expressing BRCA2 WT, the percentage of viable cells at 5 µM treatment ranging from 70 to 83 % depending on the clone (Fig. 8C) indicating that the mutated cells can rescue the phenotype of BRCA2-/- cells.
To determine directly the levels of spontaneous DNA damage in these cells and whether or not they can recruit RAD51 to sites of DNA damage we measured both the number of cells containing more than 10 nuclear foci of the DSB marker γH2AX (Fig. 8D) and the ratio of RAD51/γH2AX foci per nucleus (Fig. 8E), in cells unchallenged or 2h upon induction of DNA damage by ionizing radiation (6 Gy), (+IR). Our results show that the number of γH2AX foci, indicative of spontaneous DNA damage, is not increased in cells expressing S206C compared to the BRCA2 WT complemented cells and upon irradiation, the number of γH2AX foci in cells expressing S206C is increased to similar levels as in cells expressing BRCA2 WT (Fig. 8D). We noticed that the number of nuclear γH2AX foci in BRCA2-/- was remarkably similar to that of the BRCA2 WT cells; this is probably due to the adaptation of these cells to BRCA2 deficiency as opposed to a transient depletion.
Importantly, cells expressing S206C showed equivalent ratio of RAD51/γH2AX foci per nucleus following irradiation as the BRCA2 WT cells indicating that, in these cells, BRCA2 can recruit RAD51 to the DNA damage sites. This is in contrast to the BRCA2-/- cells where the ratio of RAD51/γH2AX foci per nucleus in irradiated conditions is nearly zero (Fig. 8E and Fig. EV9B).
Finally, to directly assess the HR proficiency of these cells, we performed a cell-based HR assay by DSB-mediated gene targeting at a specific locus (AAVS1 site) within the PPP1R12C endogenous gene using a site-specific transcription-activator like effector nuclease (TALEN) and a promoter-less mCherry donor flanked by homology sequence to the targeted locus (Brunet et al, 2009; Hockemeyer et al, 2009). DSB-meditated gene targeting results in mCherry expression from the endogenous PPP1R12C promoter (Fig. 8F) which can be measured by flow cytometry (Fig. EV 10). Using this system, BRCA2+/+ and BRCA2 WT complemented cells show ∼ 7% of mCherry positive TALEN-transfected cells whereas the BRCA2 deficiency (BRCA2-/-) led to a reduction to ∼2% of mCherry expressing cells, as expected. Importantly, TALEN-transfected cells expressing BRCA2 variants S206C and T207A showed no significant difference with the BRCA2 WT complemented cells indicating an intact HR activity.
In summary, these results indicate that the role of BRCA2 in conjunction with PLK1 in mitosis is independent of the HR function of BRCA2 as the variants S206C and T207A affecting PLK1 phosphorylation of BRCA2 are not sensitive to DNA damage, are able to recruit RAD51 to DNA damage sites (as shown for S206C) and are efficient at DSB-mediated gene targeting.
Discussion
Our results demonstrate that residues S193 and T207 of BRCA2 can be phosphorylated by PLK1 (Fig. 2) and that pT207 constitutes a bona fide docking site for PLK1PBD (Fig. 3F-I). Accordingly, BRCA2 missense variants of unknown clinical significance reducing the phosphorylation status of T207 (T207A, S206C, T200K), result in a decreased interaction of BRCA2-PLK1 (Fig. 3A-3I, 4A). The phenotype of cells expressing two of these breast cancer variants (S206C, T207A) in a BRCA2 deficient background, allowed us to investigate in detail the possible role of BRCA2 phosphorylation by PLK1 in the control of mitosis. Unexpectedly, we found that the cells expressing S206C and T207A display defective chromosome congression to the metaphase plate (Fig. 5G, H) causing a strong delay in mitosis progression (Fig. 4C, D), particularly from chromosome condensation to anaphase onset (Fig. 4E, F).
Mechanistically, cells expressing T207A and S206C exhibit reduced PLK1- dependent BUBR1 phosphorylation at two known tension-sensitive phosphosites required for the establishment of kinetochore-microtubule attachments, S676 and T680 (Elowe et al, 2007; Suijkerbuijk et al, 2012) (Fig. 5A, 5B). Proper kinetochore-microtubule attachments also require the interaction of BUBR1 with the phosphatase PP2A-B56α to balance Aurora B kinase activity (Suijkerbuijk et al, 2012; Foley et al, 2011). This interaction is mediated through the PLK1-mediated phosphorylation of several residues (including S676 and T680) in the KARD motif of BUBR1. Accordingly, we found that BRCA2 forms a tetrameric complex with PLK1-pT680- BUBR1-PP2A and that BRCA2 variants S206C and T207A strongly reduce the interaction between BRCA2 and PLK1, pBUBR1 and PP2A in this complex (Figure 5C, 5D). The fact that the interaction of BRCA2 with total BUBR1 is not altered (Fig. EV8) and that the signal of PP2A at the kinetochores is unchanged in the mutated cells (Fig. EV6) suggests that it is the impaired interaction of pBUBR1 with PP2A rather than the recruitment of PP2A to the kinetochores per se that is responsible for the phenotype observed in these cells. Consistent with this hypothesis, the complex of endogenous BUBR1 with PP2A was strongly reduced in cells expressing T207A (Fig. 5E, F).
As a consequence of the chromosome misalignment, cells bearing T207A and S206C variants display chromosome segregation errors including lagging chromosomes and chromosome bridges (Fig. 6A, 6B). Importantly, these accumulated errors ultimately lead to a broad spectrum of chromosome gains and losses (aneuploidy) compared to the wild type counterpart (Fig. 7A), but not to tetraploid cells (Fig. 7C), suggesting that cytokinesis per se, in which BRCA2 is also involved (Mondal et al, 2012; Daniels et al, 2004; Takaoka et al, 2014), is not affected.
Finally, the function of BRCA2-PLK1 interaction in mitosis is independent of the HR function of BRCA2 as cells expressing these variants display normal sensitivity to DNA damage (MMC) and PARP inhibitors (Fig. 8A-C), normal recruitment of RAD51 to DNA damage sites and the HR activity, as measured by DSB-mediated gene targeting, is intact (Fig. 8E-F, EV9B, EV10).
Putting our results together we propose the following model (Fig. 8G): in cells expressing BRCA2 WT, PLK1 phosphorylates BRCA2 on T207 leading to the docking of PLK1 at this site. This step promotes the phosphorylation of BUBR1 at the tension-sensitive phosphosites (S676 and T680) by PLK1 in prometaphase allowing the interaction of BUBR1 with the phosphatase PP2A to balance Aurora B activity (Fig. 8G, panel 1). Once the kinetochore-microtubule attachments are established and proper tension is achieved, BUBR1 is no longer phosphorylated by PLK1 (Elowe et al, 2007). This leads to a full alignment of chromosomes at the metaphase plate, SAC inactivation (Fig. 8G, panel 2) and the subsequent faithful chromosome segregation in anaphase (Fig. 8G, panel 3).
In cells expressing the variants that impair T207 phosphorylation (S206C, T207A), PLK1 cannot be recruited to pT207-BRCA2, impairing the phosphorylation of BUBR1 by the same kinase (Fig. 8F, panel 1’), which in turn reduces its binding to PP2A for proper kinetochore-microtubule interactions. This leads to chromosome misalignment defects that prolong mitosis (Fig. 8F, panel 2’); as a consequence, these cells exhibit increased chromosome segregation errors (Fig. 8F, panel 3’) and aneuploidy.
A previous study has reported the involvement of BRCA2 in the SAC through direct interaction of the C-terminal region of BRCA2 and BUBR1 (Choi et al, 2012). In this work, the authors show that BRCA2 is required for the acetylation of BUBR1 by the acetyl-transferase PCAF, which also binds BRCA2, to facilitate the activation of the SAC. A weaken SAC is manifested in a faster progression in mitosis which is what the authors observed when depleting BRCA2 in nocodazole arrested cells. In stark contrast, the cell lines studied in the present work display a strong mitotic delay in unchallenged conditions (Fig. 4C-F), which is indicative of a functional SAC. Thus, the BRCA2 function we describe here does not involve the SAC but rather the attachment of kinetochore-microtubules and the alignment of chromosomes at the metaphase plate.
The mechanism by which PLK1 regulates the function of the BRCA2-BUBR1 complex is consistent with the “processive phosphorylation” model (as opposed to the “distributive phosphorylation” model) proposed before on how PLK1PBD regulates the kinase activity of PLK1 (Lowery et al, 2005), with some modifications. In this model, PLK1PBD would first bind the site phosphorylated by another mitotic kinase allowing the kinase domain of PLK1 to phosphorylate the same protein at another site. In this work, we show that the mitotic kinase would be PLK1 itself (“self-priming”(Kang et al, 2006)), and that the kinase domain of PLK1 would then phosphorylate not the same protein but BUBR1 bound to BRCA2. Interestingly, T207 was not predicted as a target for phosphorylation by PLK1. Indeed, the consensus sequence for PLK1 phosphorylation imposes an aspartic or glutamic acid at position - 2, and in the case of T207, it is a serine (S205) (Nakajima et al, 2003). Thus, this work extends the consensus sequence for PLK1 phosphorylation.
In all, we reveal an unexpected chromosome stability control mechanism that depends on the phosphorylation of BRCA2 by PLK1 at T207. We show that BRCA2 pT207 is a docking platform for PLK1 that ensures the efficient phosphorylation of BUBR1 required for PP2A phosphatase interaction through the formation of a tetrameric complex comprising BRCA2-PLK1-pBUBR1-PP2A. Why the reduction of BRCA2 phosphorylation alters pBUBR1 interaction with PP2A? One possibility is that BRCA2 directly binds PP2A. In support of this hypothesis, a previous report has shown that BRCA2 contains a putative binding motif specific for PP2A-B56α (Hertz et al, 2016). Regardless of whether or not BRCA2 binding to PP2A is direct, we show here that pBUBR1-PP2A interaction is regulated by PLK1 recruitment to BRCA2 T207. Through this mechanism, BRCA2 ensures a proper balance of kinases and phosphatases at the kinetochore required for the faithful segregation of DNA content to daughter cells; a function that is separate from its HR activity and from the previously reported function in the SAC (Choi et al, 2012).
BRCA2 harbours another docking site for PLK1 at T77, which is primed by CDK phosphorylation (Yata et al, 2014). BRCA2 is phosphorylated by PLK1 at several sites other than T207, such as S193 (Fig. 2), which is required for the midbody localization of BRCA2 (Takaoka et al, 2014). It is also phosphorylated by PLK1 in its central region (Lee et al, 2004). The chronology of the docking and phosphorylation events and how the activities they regulate are coordinated throughout the cell cycle deserve further study.
Recent findings implicating other DNA repair proteins in mitotic functions such as ATR (Kabeche et al, 2018) together with previous work implicating PLK1 in DNA repair checkpoint recovery (Smits et al, 2000) suggest that the processes of DNA repair and cell division may involve common players acting on different protein networks exquisitely regulated via post-translational modifications.
We showed here that BRCA2 missense variants identified in breast cancer patients affect BRCA2 phosphorylation by PLK1 with direct consequences on chromosome stability manifested by chromosome missegregation and aneuploidy. Although these individual variants are rare (Table EV1), the fact that BRCA2 deficient cells exhibit low levels of phosphorylated BUBR1 (Fig. 5A lane 8 vs lane 10) suggest that this function could be affected in pathogenic truncating variants, which are more frequent. Thus, the chromosome alignment function described here could be responsible, at least in part, for the numerical chromosomal aberrations observed in BRCA2-associated tumors. Moreover, mutations in BUBR1 cause mosaic variegated aneuploidy (MVA), which is characterized by mosaic aneuploidies and childhood tumors(Hanks et al, 2004). As the variants of BRCA2 presented in this study result in a dysfunctional BUBR1 and aneuploidy, it would be interesting to investigate whether these BRCA2 variants are found in MVA patients.
Finally, the lack of sensitivity to the PARP inhibitor Olaparib observed in our mutated cell lines (Fig. 8B) strongly suggests that breast cancer patients carrying these variants would not respond to PARP inhibitor treatment (unlike BRCA2-mutated tumors that are HR-deficient).
Materials and Methods
Cell lines, cell culture and synchronisations
The human cell lines HEK293T and U2OS cells (kind gift from Dr. Mounira Amor-Gueret) were cultured in DMEM (Eurobio Abcys, Courtaboeuf, France) media containing 25 mM sodium bicarbonate and 2 mM L-Glutamine supplemented with 10% heat inactive FCS (EuroBio Abcys). The BRCA2 deficient colorectal adenocarcinoma cell line DLD1 BRCA2-/- (Hucl, T. et al 2008) (HD 105-007) and the parental cell line DLD1 BRCA2+/+ (HD-PAR-008) was purchased from Horizon Discovery (Cambridge, England). The cells were cultured in RPMI media containing 25 mM sodium bicarbonate and 2 mM L-Glutamine (EuroBio Abcys) supplemented with 10% heat inactive FCS (EuroBio Abcys). The DLD1 BRCA2-/- cells were maintained in growth media containing 0.1 mg/ml hygromycin B (Thermo Fisher Scientific). The stable cell lines of DLD1-/- BRCA2 deficient cells expressing BRCA2 WT or variants of interest generated in this study were cultured in growth media containing 0.1 mg/ml hygromycin B and 1 mg/ml G418 (Sigma-Aldrich). All cells were cultured at 37°C with 5% CO2 in a humidified incubator and all cell lines used in this study have been regularly tested negatively for mycoplasma contamination.
For synchronization of cells in mitosis, nocodazole (100-300 ng/ml, Sigma-Aldrich) was added to the growth media and the cells were cultured for 14h before harvesting. For synchronisation by double thymidine block, the cells were treated with thymidine (2.5 mM, Sigma-Aldrich) for 17h, released for 8h followed by a second thymidine (2.5 mM) treatment for 15h.
Plasmids
2XMBP-, human 2XMBP-BRCA21-250 and EGFP-MBP-BRCA2 subcloning in phCMV1 expression vector were generated as described 1,2. In the case of 2XMBP and 2XMBP-BRCA21-250, a tandem of 2 nuclear localization signals from RAD51 sequence was added downstream the MBP-tag.
Point mutations (M192T, S193A, S196N, S206C, T200K and T207A) were introduced in the 2xMBP-BRCA21-250, EGFP-MBP-BRCA2 vector using QuikChange II and QuikChange XL site-directed mutagenesis kit (Agilent Technologies), respectively (see Table EV3-EV4 for primer sequences).
For expression of BRCA2190-283 in bacteria, the human BRCA2190-283 was amplified by PCR using full length BRCA2 as template (phCMV1-2xMBP-BRCA2, see Supplementary information, Table EV5 for primer sequences). The PCR product was purified and digested with BamH1 and SalI and cloned into in the pGEX-6P-1 vector (GE Healthcare) to generate GST-BRCA2190-283. The point mutations (M192T, T200K and T207A) were introduced in the same way as for 2xMBP-BRCA21-250 and the EGFP-MBP-BRCA2. The introduction of the point mutations was verified by sequencing (see Table EV3-EV4 for primer sequences).
The PLK1 cDNA (Addgene pTK24) was cloned into the pFast-Bac HT vector using Gibson assembly (NEB) (see Table EV6 for primer sequences). To produce PLK1- KD, the point mutation K82R was introduced in the pFast-Bac HT-PLK1 vector using QuikChange XL site-directed mutagenesis kit (Agilent Technologies), see Table EV7 for primer sequences.
The Polo-like binding domain (PBD) of PLK1 (amino acid 326 to amino acid 603) was amplified from the pTK24 plasmid (Addgene) and cloned into a pT7-His6-SUMO expression vector using NEB Gibson assembly (Gibson Assembly Master Mix, New England BioLabs, Cat. # E2611S) (see Table EV7 for primer sequences). A plasmid containing a smaller PLK1 PBD fragment (amino acid 365 to amino acid 603) with a N-terminal GST tag was a kind gift from Dr. Anne Houdusse (Institute Curie, Paris).
For the DSB-gene targeting assay, we replaced the GFP tag in the promoter-less AAVS1-2A-GFP-pA plasmid (kind gift from Dr. Carine Giovannangeli) with the mCherry tag from the pET28 mCherry plasmid using NEB Gibson Assembly (Gibson Assembly Master Mix, New England BioLabs, Cat. # E2611S). See Table EV10 for primer sequences.
Expression and purification of 2xMBP-BRCA21-250
The 2xMBP-BRCA21-250 was purified as previously described1. Briefly, ten 150 mm plates of HEK293T were transient transfected with the 2xMBP-BRCA21-250 using TurboFect (Thermo Fisher Scientific). The cells were harvested 30 h post transfection, lysed in lysis buffer H (50 mM HEPES (pH 7.5), 250 mM NaCl, 1% NP-40, 5 mM EDTA, 1 mM DTT, 1 mM PMSF and EDTA-free Protease Inhibitor Cocktail (Roche)) and incubated with amylose resin (NEB) for 3h at 4°C. The 2xMBP-BRCA21-250 was eluted with 10 mM maltose. The eluate was further purified with Bio-Rex 70 cation-exchange resin (Bio-Rad) by NaCl step elution. The size and purity of the final fractions were analysed by SDS-PAGE and western blotting using anti-MBP antibody. The 2xMBP-BRCA21-250 fragments containing the BRCA2 variants (M192T, S193A, S196N, T200K, S206C and T207A) were purified following the same protocol as for WT 2xMBP-BRCA21-250.
Expression and purification of BRCA2190-283 for NMR
Recombinant 15N-labelled (WT, T200K, T207A,) and 15N/13C-labelled (WT, T207A) BRCA2190-283 were produced by transforming Escherichia coli BL21 (DE3) Star cells (Protein Expression and Purification Core Facility, Institut Curie) with the pGEX-6P-1 vector containing human BRCA2190-283 (WT and the variants) following standard heat-shock transformation protocols. Cells were grown in a M9 medium containing 0.5 g/l 15NH4Cl and 2 g/l 13C-glucose when 13C labelling was needed. The bacterial culture was induced with 1 mM IPTG at an OD600 of 0.8, and it was further incubated for 3 h at 37°C. Harvested cells were resuspended in buffer A (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 2 mM DTT, 1 mM EDTA) with 5 % glycerol, 10% Triton X-100, 1 mM PMSF and protease inhibitors cocktail (Roche) and disrupted by sonication. Clarified cell lysate was loaded onto Glutathione (GSH) Sepharose beads (GE Healthcare) equilibrated with buffer A. After 2 h of incubation at room temperature, beads were washed with buffer A and eluted with buffer A containing 20 mM reduced glutathione. The tag was cleaved by the precision protease during an overnight dialysis at 4°C against buffer B (50 mM HEPES pH 7.0, 1 mM EDTA) with 2 mM DTT and 150 mM NaCl. The cleaved GST-tag was removed by heating the sample for 15 min at 95°C and spun it down for 10 min at 16,000 x g. Sample concentration was calculated using its estimated molecular extinction coefficient of 10,363 M-1 cm-1 at 280 nm. The protein sample was characterized for folding using NMR HSQC spectra, before and after the heating at 95°C. BRCA2190-283 was dialyzed overnight at 4°C against buffer B with 2 mM DTT.
Expression and purification of PLK1 and PLK1-kinase dead (PLK1-KD)
The recombinant 6xHis-PLK1 and 6xHis-PLK1-K82R mutant (PLK1-KD) were produced in sf9 insect cells by infection for 48h (28°C, 110 rpm shaking) with the recombinant baculovirus (PLK1-pFast-Bac HT vector). Infected cells were collected by centrifugation (1300 rpm, 10 min, 4°C), washed with 1xPBS, resuspended in lysis buffer (1xPBS, 350 mM NaCl, 1% Triton X-100, 10% glycerol, EDTA-free Protease Inhibitor Cocktail (Roche), 30 mM imidazole). After 1h rotation at 4°C the lysate was centrifuged (25000 rpm, 1h, 4°C) and the supernatant was collected, filtered (0.4 µm) and loaded immediately onto a Ni-NTA column (Macherey Nagel) equilibrated with Buffer A1 (1xPBS with 350 mM NaCl, 10% glycerol and 30 mM imidazole, the column was washed with buffer A2 (1xPBS with 10% glycerol) and the protein was eluted with Buffer B1 (1x PBS with 10% glycerol and 250 mM imidazole). The eluted protein was diluted to 50 mM NaCl with Buffer A before being loaded onto a cationic exchange Capto S column (GE Healthcare) equilibrated with Buffer A1cex (50 mM HEPES (pH 7.4), 50 mM NaCl and 10% glycerol), the column was washed with Buffer A1cex before elution with Buffer B1cex (50 mM HEPES (pH 7.4), 2M NaCl and 10% glycerol). The quality of the purified protein was analysed by SDS-PAGE and the proteins concentration was determined using Bradford protocol with BSA as standard. The purest fractions were pooled and dialyzed against storage buffer (50 mM Tris-HCl (pH7.5), 150 mM NaCl, 0.25 mM DTT, 0.1 mM EDTA, 0.1 mM EGTA, 0.1 mM PMSF and 25% Glycerol) and stored in −80°C. The purified proteins can be seen in Figure EV11A.
Expression and purification of PLK1PBD
The pT7-6His-Sumo-PLK1 PBD (326-603) plasmid was expressed in Tuner pLacI pRare cells (Protein Expression and Purification Core Facility, Institut Curie), 2L of TB medium with Kanamycin and Chloramphenicol antibiotics were inoculated with cells from the pre-culture. The cells were grown at 37°C until an OD600 of ∼ 0.85. The temperature was decreased to 20°C and the expression was induced by 1mM IPTG overnight. The cells were harvested by 15 min of centrifugation at 4690 x g, at 4 °C. The cell pellets were suspended in 80 ml of 1 x PBS, pH 7.4, 150 mM NaCl, 10% glycerol, EDTA-free Protease Inhibitor Cocktail (Roche), 5 mM β-mercapto-ethanol (β-ME). The suspension was treated with benzonase nuclease and MgCl2 at 1 mM final concentration for 20 min at 4°C. The suspension was lysed by disintegration at 2 kbar (Cell distruptor T75, Cell D) followed by centrifugation at 43000 x g, for 45 min, at 4 °C. The supernatant was loaded at 1 ml/min on a His-Trap FF-crude 5 mL column (GE Healthcare) equilibrated with PBS buffer, pH 7.4, 150 mM NaCl, 10% glycerol, 5 mM β-ME (A) and 20 mM imidazole. The proteins were eluted in a linear gradient from 0 to 100 % with the same buffer (A) containing 200 mM imidazole, over 10 column volumes (CV). The purest fractions were pooled and dialyzed (8 kDa cut-off) against 20 mM Tris-HCl buffer, pH 8.0, 100 mM NaCl, 0.5 mM EDTA, 5 mM β- ME, 10% glycerol at 4 °C. 6xHis-SUMO Protease (Protein Expression and Purification Core Facility, Institut Curie) was added at 1/100 (w/w) and incubated overnight at 4 °C to cleave the 6His-SUMO tag. The cleaved PBD-PLK1 was purified using Ni-NTA agarose resin (Macherey Nagel), washed with the following buffer: 20 mM Tris-HCl pH 8.0, 100 mM NaCl, 0.5 mM EDTA, 5mM β-ME and 10 % glycerol. The sample was incubated with the resin for 1h at 4 °C and the flow-through was collected. The sample was concentrated on an Amicon Ultra Centrifugal Filter Unit (10 kDa cut-off) and injected at 0.5 ml/min on a Hi-Load 16/60 Superdex column (GE healthcare), equilibrated with 20 mM Tris-HCl buffer, pH 8.0, 100 mM NaCl, 0.5 mM EDTA, 5 mM β-ME. The protein concentration was estimated by spectrophotometric measurement of absorbance at 280 nm. The purified protein is shown in Figure EV11A.
The GST-tagged PLK1PBD (365-603) was expressed in E. coli BL21 (DE3) STAR cells, induced with 0.5 mM IPTG at an OD600 of 0.6, and grown at 37 °C for 3h. The PBD (365-603) was purified by glutathione affinity chromatography. After GST cleavage (using a 6His-TEV protease), the tag and the protease were retained using GST- and NiNTA-agarose affinity chromatography, and the PBD collected in the flow-through was further purified by gel filtration chromatography. The protein was dialyzed against a buffer containing 50 mM Tris-HCl pH 8, NaCl 150 mM, and 5 mM β-ME.
In vitro PLK1 kinase assay
0.5 µg purified 2xMBP-BRCA21-250 or 25 ng RAD51 protein, was incubated with recombinant active PLK1 (0, 50 or 100 ng) or PLK1-kinase dead (100 ng) (purchased from Abcam or purified from sf9 insect cells as detailed above, see Figure EV11B for the comparison of the kinase activity of both PLK1 preparations) in kinase buffer (25 mM HEPES, pH 7.6, 25 mM ß-glycerophosphate, 10 mM MgCl2, 2 mM EDTA, 2mM EGTA, 1 mM DTT, 1 mM Na3VO4, 10 µM ATP and 1 µCi [γ32P] ATP (Perkin Elmer)) in a 25 µl total reaction volume. After 30 min incubation at 30°C the reaction was stopped by heating at 95°C for 5 min in SDS-PAGE sample loading buffer. The samples were resolved by 7.5 % SDS-PAGE and [γ32P] ATP labelled bands were analysed with PhosphorImager (Amersham Bioscience) using ImageQuantTM TL software (GE Healthcare Life Science). To control for the amount of substrate in the kinase reaction, before adding [γ32P] ATP, half of the reaction was loaded on a 7.5 % stain free SDS-PAGE gel (BioRad), the protein bands were visualized with ChemiDoc XRS+ System (BioRad) and quantified by Image LabTM 5.2.1 Software (BioRad). The relative phosphorylation of 2xMBP-BRCA21-250 was quantified as 32P- labelled 2xMBP-BRCA21-250 (ImageQuantTM TL software) divided by the intensity of the 2xMBP-BRCA21-250 band in the SDS-PAGE gel (Image LabTM 5.2.1 Software). In the control experiment where PLK1 inhibitor was used, 50 nM BI2536 (Selleck Chemicals) was added to the kinase buffer.
In vitro protein binding assay
To assess the interaction between recombinant PLK1 and BRCA21-250 after phosphorylation by PLK1, a kinase assay was performed with 0.2 µg recombinant PLK1 and 0.5 µg purified 2xMBP-BRCA21-250 (WT or the VUS T207) in kinase buffer supplemented with 250 µM ATP (no [γ32P] ATP) in a total reaction volume of 20 µl. After 30 minutes incubation at 30°C, 15 µl amylose beads was added to the reaction and incubated for 1h at 4°C. The beads were centrifuged at 2000 x g for 2 minutes at 4°C and the unbound fraction was collected before the beads were washed three time in kinase buffer (no ATP) containing 0.5% NP-40 and 0.1% Triton X-100. Bound proteins were eluted from the beads with 10 mM maltose, protein complexes were separated by SDS-PAGE and analysed by western blotting. To control for the amount of proteins in the reaction, 2 µl of the kinase reaction (before adding the amylose beads) was loaded as input.
NMR spectroscopy
NMR experiments were carried out at 283 K on 600 and 700 MHz Bruker spectrometers equipped with a triple resonance cryoprobe. For NMR signal assignments, standard 3D triple resonance NMR experiments were recorded on BRCA2190-283 WT and T207A. Analyses of these experiments provided backbone resonance assignment for the non-phosphorylated and phosphorylated forms of these BRCA2 fragments. To follow the PLK1 phosphorylation kinetics, 1H-15N SOFAST-HSQC experiments3 were recorded at each time point. These experiments were performed using 2048 × 256 time points, 64 scans and an interscan delay of 80 ms. Data processing and analysis were carried out using Topspin and CcpNmr Analysis 2.4.2 softwares.
Analysis of phosphorylation assays followed by NMR
In the HSQC spectra, the intensity of peaks of each phosphorylated residue (pT207, pT226, pT219, pS193) as well as the intensity of peaks corresponding to their non-phosphorylated form was retrieved at each time point of the kinetics. In order to estimate the fraction of phosphorylation for each residue at each point, the function Intensity(phospho) = f[Intensity(non-phospho)] was drawn for each residue, the trendline was extrapolated to determine the intensity corresponding to the 100% phosphorylated residue and then the percentage of phosphorylation could be calculated at each time point by dividing peak intensities corresponding to the phosphorylated residue by the calculated intensity at 100% phosphorylation. Peaks corresponding to residues closed to a phosphorylated residue (L209 and V211 for pT207; A227, K230, V229 and Y232 for pT226; F221, E218 and A216 for pT219; D191, S197 and S195 for pS193) and thus affected by this phosphorylation were also treated using the same protocol and they were used to obtain a final averaged curve of the evolution of the percentage of phosphorylation at positions 193, 207, 219 and 226 with time.
Isothermal Titration Calorimetry
ITC measurements were performed with the PLK1 PBD protein (amino acid 326 to amino acid 603) and BRCA2 peptides in 50 mM Tris-HCl buffer, pH 8.0 containing 150 mM NaCl and 5 mM β-ME, using a VP-ITC instrument (Malvern), at 293 K. We used automatic injections of 8 or 10 µl. The titration data were analyzed using the program Origin 7.0 (OriginLab) and fitted to a one-site binding model. To evaluate the heat of dilution, control experiments were done with peptide or protein solutions injected into the buffer. The peptides used for the ITC experiments were synthesized by GeneCust (Ellange, LU) or Genscript (Piscataway, NY). The peptides were acetylated and amidated at the N-terminal and C-terminal ends, respectively (see Table S8 for peptide sequences). Only peptide BRCA2190-283 was expressed in bacteria and purified as detailed above (see “Expression and purification of BRCA2190-283 for NMR” section).
Crystallization and Structure Determination
The purified PBD protein (amino acid 365 to amino acid 603) was concentrated to 6 mg/ml, and mixed to the 194WSSSLATPPTLSS{pT}VLI210 (pT207) BRCA2 peptide at a 3:1 molar ratio. The crystals were obtained by hanging drop vapor diffusion method at room temperature (293 K), by mixing 1µl of complex with 1µl of solution containing 10% PEG 3350, 100 mM BisTris pH 6.5, and 5 mM DTT. Diffraction data were collected at the Proxima 1 beamline (SOLEIL synchrotron, Gif-sur-Yvette, France). The dataset was indexed and integrated using XDS4 through the autoPROC package5. The software performs an anisotropic cut-off (Tickle et al., STARANISO (2018) Global Phasing Ltd.) of merged intensity data, a Bayesian estimation of the structure amplitudes, and applies an anisotropic correction to the data. The structure was solved by molecular replacement using PHENIX (Phaser) software6. Two molecules of PBD were consecutively positioned. Electron density for the peptide was clearly visible in the position previously reported in other PBD structures in complex with phosphorylated peptides (PDB 4O56 or 3P35). Refinement was performed using BUSTER7 and PHENIX8. The model was built with Coot9. A summary of crystallographic statistics is shown in Table S2. The figures were prepared using Pymol v.1.7.4.0 (Schrödinger, LLC).
Generation of stable DLD1 clones
For generation of DLD1 BRCA2-/- cell lines stably expressing human BRCA2 variants of interest, we transfected one 100 mm plate of DLD1 BRCA2-/- cells at 70% of confluence with 10 µg of a plasmid containing human EGFP-MBP-tagged BRCA2 cDNA (corresponding to accession number NM_000059) using TurboFect (Thermo Fisher Scientific), 48h post-transfection the cells were serial diluted and cultured in media containing 1 mg/ml G418 (Sigma-Aldrich) for selection. Single cells were isolated and expanded. To verify and select the clones, cells were resuspended in cold lysis buffer H (50 mM HEPES (pH 7.5), 250 mM NaCl, 1% NP-40, 5 mM EDTA, 1 mM DTT, 1 mM PMSF and EDTA-free Protease Inhibitor Cocktail (Roche)), incubated on ice for 30 min, sonicated and centrifuged at 10,000 x g for 15 min, 100 µg total protein lysate was run on a 4-15% SDS-PAGE followed by immunoblotting using BRCA2 and GFP antibodies to detect EGFP-MBP-BRCA2. Clones with similar expression levels were selected for functional studies.
The presence of the point mutations in the genome of the clones was confirmed by extraction of genomic DNA using Quick-DNATM Universal Kit (ZYMO Research) followed by amplification of the N-terminal of BRCA2 (aa 1-267) by PCR using a forward primers that binds to the end of MBP and a reverse primer that binds to amino acid 267 in BRCA2, the presence of the point mutations was confirmed by sequencing of the PCR product (see Table EV9 and EV4 for primer sequences).
Cell extracts, immunoprecipitation and western blotting
For the interaction between BRCA21-250 and endogenous PLK1, U2OS cells were transfected with 2xMBP-BRCA21-250 construct (WT, M192T, S193A, S196N, T200K, S206C, and T207A) using TurboFect (Thermo Fisher Scientific), 30 h post-transfection cells were synchronized by nocodazole (300 ng/ml), harvested and lysed in extraction buffer A (20 mM HEPES (pH 7.5), 150 mM NaCl, 0.1% NP40, 2 mM EGTA, 1.5 mM MgCl2, 50 mM NaF, 10 % glycerol, 1 mM Na3VO4, 20 mM ß- glycerophosphate, 1 mM DTT and EDTA-free Protease Inhibitor Cocktail (Roche)). After centrifugation at 18,000 x g for 15 min, the supernatant was incubated with amylose resin (NEB) for 1.5h at 4°C. The beads were washed five times in extraction buffer before elution with 10 mM maltose. Bound proteins were separated by SDS-PAGE and analysed by western blotting. Where PLK1 inhibitor was used, the cells were synchronized in mitosis by nocodazole (14h) followed by 2h treatment with PLK1 inhibitor (50-100 nM BI2536 (Selleck Chemicals) or 50 nM BTO-1 (Sigma-Aldrich)) before being harvested. The cells were lysed in extraction buffer, pre-cleared by centrifugation and total protein lysate was separated by SDS-PAGE and analysed by western blotting.
For BUBR1 and pBUBR1 levels in mitosis, nocodazole (100 ng/ml) treated DLD1 BRCA2-/- clones were lysed in extraction buffer A, pre-cleared by centrifugation and total protein lysate was separated by SDS-PAGE and analysed by western blotting. For analysis of the interaction between BRCA2-PLK1, BRCA2-BUBR1 and for the tetrameric-protein complex BRCA2-pBUBR1-PP2A-B56α-PLK1 in mitosis, DLD1 BRCA2-/- stable clones expressing EGFPMBP-BRCA2 (WT or the VUS S206C or T207A) were synchronized with nocodazole, harvested and lysed in extraction buffer A. The lysate were pre-cleared by centrifugation before incubation with GFP-TRAP beads (Chromotek) for 2h at 4°C to pull-down EGFP-MBP-BRCA2. Around 3 mg total protein lysate was used per pull-down. The beads were washed 5 times in extraction buffer and 2 times in extraction buffer with 500 mM NaCl. Bound proteins were eluted by boiling the samples for 4 min in 3x SDS-PAGE sample loading buffer (SB), eluted proteins were separated by SDS-PAGE and analysed by western blotting using anti-mouse PLK1, anti-rabbit pT680-BUBR1, mouse anti-PP2A C and mouse anti-BRCA2 (OP95) antibodies.
For immunoprecipitation of endogenous BUBR1, nocodazole treated DLD1 BRCA2-/- stable clones expressing BRCA2 WT or the variant T207A were lysed in extraction buffer A. After centrifugation, 2000-3000 µg total protein lysate was pre-cleared by incubation with 20 µl Protein G PLUS-Agarose (Santa Cruz, sc-2002) for 30 min at 4°C. The pre-cleared lysate was incubated with 1.25 µg BUBR1 mouse antibody or control mouse IgG over night at 4°C before addition of 40 µl Protein G PLUS-Agarose, the lysate was incubated for additional 30 min before immunoprecipitates were collected by centrifugation. After four washes in extraction buffer and two washes in extraction buffer with 500 mM NaCl, the beads were re-suspended in SB, boiled and the immunocomplexes were analysed by western blotting using anti-rabbit BUBR1 and mouse anti-PP2A C antibodies.
Antibodies used for western blotting
mouse anti-MBP (1:5000, R29, Cat. #MA5-14122, Thermo Fisher Scientific), mouse anti-BRCA2 (1:1000, OP95, EMD Millipore), rabbit anti-GFP (1:5000, Protein Expression and Purification Core Facility, Institut Curie), mouse anti-PLK1 (1:5000, clone 35-206, Cat. #05-844, EMD Millipore), mouse anti-BUBR1 (1:1000, Cat. #612502, BD Transduction Laboratories), rabbit anti-BUBR1 (1:2000, Cat. #A300- 386A, Bethyl Laboratories), mouse anti-PP2A C subunit (1:1000, clone 1D6, Cat. #05-421, EMD Millipore), rabbit anti-pT680-BUBR1 (1:1000, EPR 19958, Cat. #ab200061, Abcam) and rabbit anti-pS676-BUBR1 (1:1000, R193, kind gift from Dr. Erich A Nigg). Horseradish peroxidase (HRP) conjugated 2nd antibodies used: mouse-IgGκ BP-HRP (IB: 1:10 000, Cat. #sc-516102, Santa Cruz), goat anti-rabbit IgG-HRP (IB: 1:5000, Cat. #sc-2054, Santa Cruz), goat anti-mouse IgG-HRP (1:10 000, Cat.# 115-035-003, Interchim), goat anti-rabbit IgG-HRP (1:10 000, Interchim, Cat.# 111-035-003).
Phosphatase treatment
DLD1 BRCA2-/- cells stably expressing EGFP-MBP-BRCA2 WT were synchronized in mitosis by nocodazole (14h), harvested, lysed in extraction buffer A without phosphatase inhibitors (NaF, Na3VO4 and ß-glycerophosphate), and pre-cleared by centrifugation. Increased amount (0-20U) of FastAP Thermosensitive Alkaline Phosphatase (Thermo Fisher Scientific Cat. #EF0654) was added to 15 µg of total protein lysate in FastAP Buffer in a total reaction volume of 60 µl. After 1h incubation at 37°C the reaction was stopped by heating at 95°C for 5 min in SDS-PAGE sample loading buffer, 30 µl of the reaction was loaded on a 4-15 % SDS-PAGE gel, the gel was transferred onto nitrocellulose membrane and the levels of pT680-BUBR1 were analysed by western blotting.
Cell survival and viability assays
For clonogenic survival assay, DLD1 BRCA2-/- cells stably expressing full-length GFPMBP-BRCA2 and the variants (S206C and T207A) were treated at 70% of confluence with Mitomycin C (Sigma-Aldrich) at concentrations: 0, 0.5, 1.0 and 2.5 µM. After 1 h drug treatment the cells were serial diluted in normal growth media containing penicillin/streptomycin (Eurobio) and seeded in triplicates into 6-well plates. The media was changed every third day, after 10-12 days in culture the plates were stained with crystal violet, colonies were counted and the surviving fraction was determined for each drug concentration.
Cell viability was assessed with 3-[4,5-Dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT, #M5655, Sigma Aldrich) after treatment with MMC and the PARP inhibitor Olaparib (AZD2281, Ku-0059436, #S1060, Selleck Chemicals). For MMC, the cells were plated in triplicates in 96-well microplates (3000-5000 cells/well) the day before treatment. The cells were washed once in PBS before addition of serum-free media containing MMC at the concentrations: 0, 1.0 and 2.5 µM. After 1h treatment the cells were washed once in PBS and incubated for 72h in normal growth media before the viability was measured by MTT assay. For PARP inhibition, the cells were seeded 4h before 4-days treatment in normal growth media with Olaparib at concentrations: 0, 2.5 and 5.0 µM.
Homologous recombination assays
We applied a DSB-mediated gene targeting strategy using site-specific TALEN nucleases to quantify HR in cells. DLD1 BRCA2-/- cells stably expressing full-length GFPMBP-BRCA2 and the variants (S206C and T207A) were transfected using AMAXA technology (Lonza) nucleofector kit V (Cat. # VCA-1003) with 3 µg of the promoter-less donor plasmid (AAVS1-2A-mCherry) with or without 1 µg of each AAVS1-TALEN encoding plasmids (TALEN-AAVS1-5’ and TALEN-AAVS1-3’, see Table EV11 for sequences, kind gift from Dr. Carine Giovannangeli). For each transfection, 1 x 106 cells were transfected using program L-024, the cells were seeded in 6-well plate in culture media without selection antibiotics. The day after transfection the media was changed to media with selection and 48h post-transfection the cells were trypsinized and reseeded on a 10-cm culture dish and cultured for additional five days. The percentage of mCherry positive cells was analysed on a BD FACSAria III (BD Bioscience) using FACSDiva software and data was analysed with FlowJo 10.4.2 software (Tree Star Inc.).
Analysis of tetraploid cells
For the analysis of S-phase tetraploid cells in the DLD1 BRCA2-/- stable clones, the cells were incubated with 10 µM BrdU for 20 minutes before they were harvest, fixed and stained for cell cycle analysis using a APC-BrdU flow kit (BD Bioscience, Cat. #552598) following the manufacturer’s instructions.
Labelled cells were analysed on a BD FACSCanto II (BD Bioscience) using FACSDiva software and data were analysed with FlowJo 10.4.2 software (Tree Star Inc.).
Immunofluorescence
DLD1 BRCA2-/- cells stably expressing full-length EGFP-MBP-BRCA2 and the variants (S206C and T207A) were seeded on coverslips in 6-well tissue culture plates and synchronized in mitosis. For analysis of chromosome alignment, the cells were synchronized by double thymidine block, released for 9h followed by treatment with monastrol (100 µM, Sigma-Aldrich) for 16h before the proteasome inhibitor MG- 132 (10 µM, Sigma-Aldrich) was added to the cells for an incubation of 1h. For the monastrol wash-out experiment (Fig. EV7A), the cells were washed twice in PBS to remove the monastrol before adding the MG-132.
For chromosome segregation analysis, the cells were synchronized by double thymidine block and released in normal growth media for 11h. The cells were fixed with 100% methanol for 15 min at −20°C, rinsed once in PBS before permeabilization with PBS containing 0.1% Triton-X for 15 min at room temperature. Nonspecific epitope binding was blocked with 4% BSA (Sigma-Aldrich) in PBS. The coverslips were rinsed in PBS, incubated with primary antibody diluted in PBS containing 0.1% Tween-20 (PBS-T) and 5% BSA for 1h at room temperature. After incubation, the coverslips were washed three times of 5 min in PBS-T before being incubated for 1h at room temperature with respective Alexa Fluor conjugated secondary antibody diluted in PBS-T with 5% BSA. The coverslips were washed two times of 5 min each in PBS-T followed by one rinse in PBS before being mounted on microscope slides.
For aneuploidy analysis the cells were treated with nocodazole for 14h (0.1 µg/ml) to enable chromosome spread; the cells were rinsed in PBS, incubated for 10 min with KCl (50 mM) at room temperature before they were spread on coverslips at 900 rpm for 5 min in a Cytospin 4 (Thermo Scientific). The cells were fixed with 3% paraformaldehyde (PFA) in PBS for 20 min followed by 15 min permeabilization in PBS containing 0.1% Triton X-100. The coverslips were rinsed three times in PBS, blocked with 5% BSA in PBS before incubation with primary antibodies diluted in PBS over night at 4°C. After incubation the coverslips were washed three times of 5 min in PBS before 1h incubation at room temperature with respective Alexa Fluor conjugated secondary antibody diluted in PBS. After three washes of 5 min in PBS the coverslips were mounted on microscope slides.
For staining of PLK1 and PP2A at the kinetochore, the cells were seeded on coverslips and treated with nocodazole (0.25 µg/ml) for 4h, fixed with 4% PFA in PBS containing 0.5% Triton X-100 for 20 minutes at room temperature. The coverslips were rinsed three times in PBS-T and blocked for 30 minutes with 4% BSA in PBS before incubation with primary antibodies diluted in PBS-T with 5% BSA over night at 4°C. After three washes of 5 minutes in PBS-T the coverslips were incubated for 2h incubation at room temperature with respective Alexa Fluor conjugated secondary antibody diluted in PBS-T with 5% BSA. After two washes of 5 min in PBS-T and one rinse in PBS the coverslips were mounted on microscope slides.
For analysis of γH2AX and RAD51 foci, the cells were seeded on coverslips the day before 6 Gy γ−irradiation (GSR D1, Cs-137 irradiator). Two hours after irradiation, the coverslips were washed twice in PBS followed by one wash in CSK Buffer (10 mM PIPES, pH 6.8, 0.1 M NaCl, 0.3 M sucrose, 3 mM MgCl2, EDTA-free Protease Inhibitor Cocktail (Roche)). The cells were permeabilized for 5 minutes at room temperature in CSK buffer containing 0.5% Triton X-100 (CSK-T) followed by one rinse in CSK buffer and one rinse in PBS before fixation for 20 minutes at room temperature with 2% PFA in PBS. After one rinse in PBS and one in PBS-T, the cells were blocked for 5 minutes at room temperature with 5% BSA in PBS-T before incubation for 2h at room temperature with primary antibodies diluted in PBS-T with 5% BSA. After primary antibody incubation, the coverslips were rinsed in PBS-T followed by two washes of 10 minutes in PBS-T and blocked for 5 minutes at room temperature with 5% BSA in PBS-T before incubation for 1h at room temperature with respective Alexa Fluor conjugated secondary antibody diluted in PBS-T with 5% BSA. After one rinse in PBS-T and two washes of 10 minutes in PBS-T the coverslips were rinsed in PBS before being mounted on microscope slides.
All coverslips were mounted on microscope slides with ProLong Diamond Antifade Mountant with DAPI (Cat. #P36966, Thermo Fisher Scientific).
For analysis of chromosome alignment and segregation, images were acquired in an upright Leica DM6000B wide-field microscope equipped with a Leica Plan Apo 63x NA 1.4 oil immersion objective. The camera used is a Hamamatsu Flash 4.0 sCMOS controlled with MetaMorph2.1 software (Molecular Devices). For Figures 6c and 7b, 7 to 20 Z-stacks were taken at 0.2 μm intervals to generate a maximal intensity projection image using ImageJ. For aneuploidy and kinetochore localization analysis, images were acquired in an inverted confocal Leica SP5 microscope with a plan Apo 63x NA 1.4 oil immersion objective. For Figure 7b, Z-stacks were taken at 0.13 μm intervals to generate a maximal intensity projection image using ImageJ. For the counting of chromosomes in the aneuploidy experiment, the quantification was performed in zoomed areas counting the CREST signal in separated stacks to ensure the counting of all chromosomes. We were able to count up to 65 chromosomes with certainty, thus >65 CREST signals were discarded and not included in the analysis. For the analysis of γH2AX and RAD51 foci, 20 Z-stacks were taken at 0.2 μm intervals to generate a maximal intensity projection using Image J. LH2AX and RAD51 foci per nucleus were counted by a customized macro using a semi-automated procedure; the nucleus was defined by an auto-threshold on DAPI, a mask was generated and applied onto the Z-projection to only count foci within the nucleus. For the definition of foci we applied the Plug In Find Maxima (ImageJ).
Antibodies used for immunofluorescence
Human anti-CREST (1:100, Cat. #15-234-0001, Antibodies Online), mouse anti-PP2A C subunit (1:500, Clone 1D6, Cat. #05-421, EMD Millipore), mouse anti-PLK1 (1:500, clone F-8, Santa Cruz Biotechnology, Cat. #sc-17783), anti-pT680-BUBR1 (1:100, clone EPR 19958, Abcam, Cat. #ab200061), mouse anti-pSer139- γH2AX (1:1000, clone JBW301, EMD-Millipore, Cat. #05-636), rabbit anti-RAD51 (1:100, clone H-92, Santa Cruz Biotechnology, Cat. #sc-8349) and mouse anti-α-tubulin (1:5000, GT114, Cat. #GTX628802, Euromedex). Alexa Fluor conjugated secondary antibodies used: goat anti-human Alexa-488 (1:1000, Cat. # A11013, Life Technologies), donkey anti-mouse Alexa-594 (1:1000, Cat. #A-21203, Thermo Fisher Scientific), donkey anti-rabbit Alexa-488 (1:1000, Cat. #A-21206, Thermo Fisher Scientific), goat anti-human Alexa-555 (1:1000, Cat. #A-21433, Thermo Fisher Scientific), donkey anti-mouse Alexa-488 (1:1000, Cat. #A-21202, Thermo Fisher Scientific).
Time-lapse video microscopy of mitotic cells
For phase-contrast video-microscopy DLD1 BRCA2-/- cells stably expressing full-length GFPMBP-BRCA2 and the variants (S206C and T207A) were seeded in 35 mm Ibidi µ-Dishes (Ibidi, Cat. #81156), synchronized by double thymidine block, released and cultured for 4h in normal growth media before the filming was started. The cells were imaged for 16h every 5 min, at oil-40X using an inverted video-microscope (Leica DMI6000) equipped with electron multiplying charge coupled device (EMCCD) camera controlled by Metamorph software (Molecular Devices). Images were mounted using Image J software (1.51s, NIH).
For the analysis of the time from chromosome condensation to anaphase onset, the cells were seeded in Ibidi 4 well µ-slide with glass bottom (Ibidi, Cat. #80427), synchronized by double thymidine block. At the time of release from the second block, the cells were washed twice in PBS, released and cultured for 7h in phenol red free culture media before the filming was started. To stain the DNA, 50 nM SiR-DNA probe (far red absorption) and 3 μM verapamil (SiR-DNA Kit, Cat. #SC007, Spirochrome) were added to the culture media at the time of release from the second block. The cells were imaged for 14h every 2 min, at 40x dry objective using an inverted spinning-disk microscope (Ti-E, Nikon and Yokogawa CSU-X1 spinning head) equipped with sCMOS Hamamatsu Orca Flash 4.0 (pixel size 6.5 µm) camera controlled by Metamorph software (Molecular Devices). Images were mounted using Image J software (1.51s, NIH).
Statistical analysis
In all graphs error bars represent the standard deviation (SD) from at least three independent experiments unless otherwise stated. Statistical significance of differences was calculated with unpaired two-tailed t-test, one/two-way ANOVA with Tukey’s multiple comparisons test or Mann-Whitney two-tailed test as indicated in the figure legends. All analyses were conducted using GraphPad Prism (version Mac OS X 7.0b).
Author Contributions
A.E. purified WT and mutated BRCA21-250, established the stable DLD1-/- cell lines, performed kinase assays, pull-down assays, western blots, time-lapse microscopy experiments, mitotic index measurements by FACS, clonogenic survival and MTT assays as well as the statistical analysis for these experiments. C.M. performed IF and image acquisition of metaphase plate alignment, chromosome segregation and karyotype analysis. M.J. performed the NMR experiments assisted by S. M., F.T. and S.Z.J. S.M. purified PLK1PBD, performed the ITC experiments and solved the X-ray structure assisted by V.R. V.B. assisted establishing stable clones and performing clonogenic survival assays. P.D. purified PLK1PBD. A.M. cloned and produced PLK1 and PLK1-KD from insect cells. A.C., A.E. and S.Z.J. designed the experiments. A.C. and S.Z.J. supervised the work. A.C. wrote the paper with important contributions from all authors.
The authors declare no conflict of interest.
Correspondence and requests for materials should be addressed to A.C.
Acknowledgments
We thank members of the AC lab for fruitful comments on the manuscript and Davide Panigada for the illustration of chromosomes in Figure 8. We thank Rene H. Medema for useful discussions on this work including the cell synchronization protocol used in Figure 7. We also thank Juan S. Martinez for construct BRCA21-250, Anne Houdusse for construct PLK1365-603, Carine Giovannangeli for TALEN plasmids and Eric Nigg for pS676-BUBR1 antibody. We acknowledge the Cell and Tissue Imaging Facility of the Institut Curie (PICT), a member of the France BioImaging National Infrastructure (ANR-10-INBS-04) and the French Infrastructure for Integrated Structural Biology (https://www.structuralbiology.eu/networks/frisbi, (ANR-10-INSB-05-01). We thank Charlene Lasgi from the Flow Cytometry platform of Institut Curie, Orsay.
This work was supported by the ATIP-AVENIR CNRS/INSERM Young Investigator grant 201201, EC-Marie Curie Career Integration grant CIG293444 to A.C. and Institut National du Cancer INCa-DGOS_8706 to A.C. and S.Z.J.; A.E. was supported by the Swedish Society for Medical Research.