SPRTN protease-cleaved MRE11 decreases DNA repair and radiosensitises cancer cells

Abstract

The human MRE11/RAD50/NBS1 (MRN) complex plays a crucial role in sensing and repairing DNA DSB. MRE11 possesses dual 3'-5' exonuclease and endonuclease activity and forms the core of the multifunctional MRN complex. We previously identified a C-terminally truncated form of MRE11 (TR-MRE11) associated with post-translational MRE11 degradation. Here we identified SPRTN as the essential protease for the formation of TR-MRE11 and characterised the role of this MRE11 form in its DNA damage response (DDR). Using tandem mass spectrometry and site-directed mutagenesis, the SPRTN-dependent cleavage site for MRE11 was identified between 559 and 580 amino acids. Despite the intact interaction of TR-MRE11 with its constitutive core complex proteins RAD50 and NBS1, both nuclease activities of truncated MRE11 were dramatically reduced due to its deficient binding to DNA. Furthermore, lack of the MRE11 C-terminal decreased HR repair efficiency, very likely due to abolished recruitment of TR-MRE11 to the sites of DNA damage, which consequently led to increased cellular radiosensitivity. The presence of this DNA repair-defective TR-MRE11 could explain our previous finding that the high MRE11 protein expression by immunohistochemistry correlates with improved survival following radical radiotherapy in bladder cancer patients.

Fig. 1. MRE11 is found in truncated form in bladder cancer cell lines and primary bladder tumours.

A Two out of nine bladder cancer cell lines show a truncated form of MRE11 (approximately 65 kDa). Calgem is gemcitabine-resistant cells created from Cal29 by exposure to increasing gemcitabine doses. B Eight out of nine nuclear cell-free extracts from human bladder tumours show a truncated form of MRE11 (approximately 65 kDa). Ta = papillary tumour, T1 = tumour invading lamina propria, T2 = tumour invading muscularis propria. C Schematic diagram of MRE11 and the epitope regions of MRE11 of two antibodies, spanning amino from acids 182 to 582 (ab214) and 650 to C-terminus (ab30725), respectively. D In RT112, ab214, spanning amino acids 182–582, detected both FL and TR MRE11, but not in VM-CUB1. In contrast, the C-terminus anti-MRE11 antibody (ab30725) did not detect the truncated MRE11 both in RT112 and VM-CUB1. E-i The eluted sample from super-paramagnetic beads with recombinant protein A-bound MRE11 was loaded onto a precast 4–20% polyacrylamide gel. E-ii, iii The sequence coverage of FL-MRE11 digested with trypsin (red) and elastase (blue) was 71% coverage for both cases. E-iv, v Protein sequence coverage of TR-MRE11 digested with trypsin (red) and elastase (blue) were 23% and 28% respectively. The green box at K568 indicates the last amino acid of TR-MRE11 model.

Fig. 2. The putative cleavage site in MRE11 lies within a 22 amino acid region, namely amino acids 559–580.

A A map of the lentiviral vector including either intact MRE11 or MRE11 with the initial putative cleavage site (PCS) at amino acids 539–600 removed. The functional domains of MRE11 are indicated^,. B Schematic representation of MRE11 with the various mutants, including complete deletion of the region (MRE11 PCS removed) and sequential deletions of 12 amino acids (MRE11 Δ539–550, Δ549–560, Δ559–570, Δ569–580, Δ579–590 and Δ589–600). C The MRE11 mutant lacking the initial putative cleavage site failed to generate truncated MRE11. D MRE11 mutants lacking either 559–570 or 569–580 aa did not generate the approximately 20 kDa size short form of cleaved MRE11 in RT112 cells, as seen using anti-myc-tag and anti-MRE11 ab 30725 antibodies. E pLenti-TR-MRE11 vector was designed by deleting 569–708 aa from FL-MRE11 (amino acids deleted shown in upper panel). Lower panel shows the truncation in diagrammatic form.

Fig. 3. TR-MRE11 lacks nuclease and DNA binding activities.

A Nuclease assays of FL-MRE11 (lane 1, 2, 3, 4) and TR-MRE11 (lane 5, 6, 7, 8) on 3′-overhang DNA substrates (n = 3). The FL-MRE11 shows clear dose-dependent nuclease activity which is absent in TR-MRE11. B Quantification of nuclease activity data in (A). Error bars represent SEM (P < 0.0001 (****), P < 0.05 (*)). C Measurement of fluorescence polarisation (n = 1). The apparent dissociation constant for FL-MRE11 was 34.56 nM, whilst TR-MRE11 showed significantly reduced DNA binding (242.15 nM). D TR-MRE11 was still able to efficiently interact with RAD50 and NBS1, and this has been shown through immunoprecipitated MRE11 from VM-CUB1 protein lysates.

Fig. 4. TR-MRE11 co-localises less frequently than FL-MRE11with γH2AX after IR in VM-CUB1 cells.

A Western blot of established cell lines from VM-CUB1. Endogenous MRE11 was knocked down and replaced with either exogenous FL or TR-MRE11, and named MRE11 knockdown (KD), FL-MRE11, and TR-MRE11 cells, respectively. B Immunofluorescence microscopy of MRE11 and γH2AX foci showing co-localisation and some failure of co-localisation after IR (2 Gy, 24 hr post-IR). MRE11 foci were clearly knocked down in the shMRE11 sample and rescued in FL and TR. C TR-MRE11 colocalises with γH2AX less than FL-MRE11 does. Data are presented as means ± SEM (P < 0.001 (***), P < 0.01 (**)). D Correlation between MRE11 and γH2AX foci numbers from (B). MRE11 deficient and TR-MRE11 cells showed non-significantly lower slopes for γH2AX foci compared to FL-MRE11 and untreated cells. E Representative graphs for propidium iodide cell cycle. F Cell cycle distribution of RT112 UT, KD, FL-MRE11, and TR-MRE11 cells was quantitatively analysed using flow cytometry (****<0.0001, **<0.01, *P < 0.05; one-way ANOVA, n = 3, mean + SEM).

Fig. 5. Cells expressing TR-MRE11 are more radiosensitive and deficiency in HR repair than those expressing FL-MRE11.

A The protein expression levels of the MRN complex subunits were unchanged after IR (2 Gy). MRE11 KD and TR-MRE11 showed less increased level of γH2AX compared to WT after IR. B GFP based reporter assay showing the efficiency of HR repair pathway under FL-MRE11 or TR-MRE11 rescued condition as indicated. Graph represents the quantifications from three independent experiments (*P < 0.05; one-way ANOVA, n = 3, mean + SEM). C Clonogenic survival rates of TR-MRE11 and KD cells were lower than for FL-MRE11 and untreated cells. The response to ionising irradiation is represented by the linear-quadratic model. KD cells had significantly lower survival rate than UT cells (P < 0.05). TR-MRE11 cells had significantly lower survival rate than FL-MRE11 (P < 0.05) at 4 Gy. D The representative image of (C). E Flow cytometric analysis results of various RT112 cells after 8 Gy IR (48 h). F A statistical plot of annexin V-FITC/PI staining for (E) is shown for both early and late apoptosis after 4 and 8 Gy IR (48 h). The results are expressed as the means ± SD (****<0.0001, **<0.01, *P < 0.05).

Fig. 6. MRE11 is cleaved in cells and in vitro by the SPRTN metalloprotease and SPRTN overexpression increases radiosensitivity.

A siRNA-targeting SPRTN in RT112 cells regulates the levels of TR-MRE11. siRNA was used indicated concentrations 40, 60, 80 and 100 nM for 72 h, with scrambled siRNA as a positive control. Pictures are representative from 3 replicates with similar results. B Quantification of TR-MRE11 and SPRTN knockdown level from (A). C A significant correlation was found between TR-MRE11 and SPRTN protein expression levels (p = 0.011). D Clonogenic survival rates of SPRTN overexpressed cells were lower after IR than for WT cells in RT112. The response to ionising irradiation is represented by the linear-quadratic model. MRE11 KD+SPRTN cells had significantly lower survival rate than WT+SPRTN cells (P < 0.05). WT+SPRTN cells had significantly lower survival rate than WT (P < 0.05). E SPRTN overexpression in WT and MRE11 KD cells was confirmed by Western blotting. F The representative images of (D). G In vitro cleavage assay using recombinant proteins. TR-MRE11, indicated with arrow, was released only in the presence of active SPRTN. A blot for anti-MRE11 (C-terminus) antibody shows nonspecific binding to SPRTN protein, marked with asterisk. Cdc25A was used as a positive control for SPRTN activity (the size of the cleaved products is ~5 kDa smaller than the products in Fig. 2D, as this recombinant protein does not include tag part in C-terminus). H Immunoprecipitation of 293T cell lysate using SPRTN-WT and SPRTN-E112A. p97, the strong interacting partner for SPRTN, was used as a positive control for SPRTN interacting protein.