Identification of a Heteromeric Complex That Promotes DNA Replication Origin Firing in Human Cells

See allHide authors and affiliations

Science  24 May 2013:
Vol. 340, Issue 6135, pp. 981-984
DOI: 10.1126/science.1237448

Keeping Coordinated

Before a cell divides, it must replicate its genome so that both daughter cells receive a copy of the parental DNA. Replication must be tightly regulated to ensure that the genome is replicated only once, because over- or underreplication could result in aberrations that cause genome instability. Thus, replication must be coordinated with other events in the cell, such as the cell cycle and DNA damage response systems. Boos et al. (p. 981) analyzed the function of the Treslin/TICRR protein, an essential DNA replication factor regulated by cyclin-dependent kinases and the DNA damage checkpoint. Treslin interacts with the Mdm Two Binding Protein (MTBP), implicated in oncogenesis. MTBP and Treslin appear to integrate signals from the cell cycle and the DNA damage–response pathway, thereby controlling the initiation of DNA replication.


Treslin/TICRR (TopBP1-interacting, replication stimulating protein/TopBP1-interacting, checkpoint, and replication regulator), the human ortholog of the yeast Sld3 protein, is an essential DNA replication factor that is regulated by cyclin-dependent kinases and the DNA damage checkpoint. We identified MDM two binding protein (MTBP) as a factor that interacts with Treslin/TICRR throughout the cell cycle. We show that MTBP depletion by means of small interfering RNA inhibits DNA replication by preventing assembly of the CMG (Cdc45–MCM–GINS) holohelicase during origin firing. Although MTBP has been implicated in the function of the p53 tumor suppressor, we found MTBP is required for DNA replication irrespective of a cell’s p53 status. We propose that MTBP acts with Treslin/TICRR to integrate signals from cell cycle and DNA damage response pathways to control the initiation of DNA replication in human cells.

Eukaryotic cells must ensure their entire genome is replicated exactly once in each cell cycle (13). The yeast Sld3 protein has emerged as a focal point of regulation of replication origin firing: It is an essential cyclin-dependent kinase (CDK) substrate (46), its recruitment to origins is regulated by Dbf4-dependent kinase (DDK) (7, 8), it is inhibited by the DNA damage checkpoint kinase Rad53 (9, 10), and it is one of a handful of limiting factors involved in executing the temporal program of origin firing (7, 11). Treslin/TICRR (TopBP1-interacting, replication stimulating protein/TopBP1-interacting, checkpoint, and replication regulator) is the ortholog of Sld3 in multicellular eukaryotes and is also regulated by CDK and the DNA damage checkpoint (1216). Treslin/TICRR is, however, much larger than Sld3 and has additional domains, suggesting that it may interact with additional factors and may be regulated by additional pathways.

We used mass spectrometry to identify interacting proteins in FLAG immunoprecipitates from a cell line stably expressing FLAG-tagged Treslin/TICRR. Among the proteins most highly enriched (table S1 and fig. S1) was the Mdm two binding protein (MTBP). Endogenous Treslin/TICRR and MTBP coimmunoprecipitated with antibodies to both MTBP and Treslin/TICRR (Fig. 1A). Depletion of MTBP with an antibody to MTBP resulted in removal of ~50% of the Treslin/TICRR from the extract (fig. S2i), and immunodepletion of Treslin/TICRR removed between 50 and 75% of the MTBP from the extract (fig. S2ii). MTBP and Treslin/TICRR coimmunoprecipitated from all extracts made from cells synchronized with either nocodazole or thymidine and released into the G1, S, and G2/M phases (Fig. 1B). In these experiments, MTBP and Treslin/TICRR were present throughout the cell cycle, but their levels increased later in the cell cycle. Therefore, Treslin and MTBP are present in a complex throughout most or all of the cell cycle.

Fig. 1 MTBP and Treslin/TICRR form a complex throughout the cell cycle.

(A) Endogenous MTBP or Treslin/TICRR were immunoprecipitated from HeLa cell extracts, and extracts (Inp) and immunoprecipitates (IP) were analyzed for MTBP and Treslin/TICRR with immunoblotting. Ctr, control-IPs. (B) U2OS cells were released from a mitotic block with nocodazole (Noc) or from a G1/S arrest with thymidine (Thy) for the indicated times, and cell lysates and IPs of endogenous MTBP (MTBP-IP; Ctr-IP, control-IP) were analyzed with immunoblotting using antibodies to MTBP or Treslin/TICRR or Ponceau staining (Pon.). Cell-cycle analysis was done with flow cytometry of BrdU- and propidium iodide (PI)–stained cells.

MTBP interacted with Treslin/TICRR lacking the conserved CIT or STD domains (Fig. 2, A and B). However, two nonoverlapping deletions in the M domain between the CIT and STD domains (∆M1 and ∆M2) showed reduced MTBP interaction, and one, ∆M2, was completely defective in MTBP interaction. Moreover, a fragment containing just the M-domain was sufficient for interaction with full-length MTBP (fig. S3). A small interfering RNA (siRNA)–resistant version of ∆M2 (Fig. 2C) localized to the nucleus (fig. S4) but could not restore DNA replication after depletion of endogenous Treslin/TICRR (Fig. 2D), indicating that Treslin/TICRR’s ability to interact with MTBP is required for DNA replication.

Fig. 2 Treslin, MTBP, and TopBP1 form a ternary complex.

(A) Schematic of Treslin/TICRR and Sld3 domains. The indicated deletions (Del.) and point mutants (Mut.) (the amino acids deleted are available in the supplementary materials, materials and methods) of the Treslin domains CIT (conserved in Treslins), M-domain (middle-domain), STD (Sld3/Treslin domain), TopBP1-interacting region [2PM, double phosphomutant (16)] and CT (C-terminus) were used in (B) to (D). (B) 293T cells were transiently transfected (16) with MTBP-Myc and AcGFP-Flag-Treslin-wild type (WT) or the indicated mutants described in (A). Myc-IPs and cell lysates (Input) were analyzed with immunoblotting with antibodies to Myc and GFP. 220/160 is the molecular weight in kilodaltons. (C) U2OS control cells or clones stably expressing siRNA-resistant AcGFP-Flag-Treslin-WT, 2PM, or ΔM2 (A) were analyzed by immunoblotting using antibodies to Treslin and Tubulin after control (siCtr) or Treslin (siTres) RNAi (16). Asterisks indicate endogenous (endog.) and tagged Treslin/TICRR (∆M2, full-length). (D) Cells described in (C) were analyzed for BrdU incorporation by using flow cytometry and displayed as histogram plots FITC fluorescence (AU, arbitrary units; log, logarithmic scale). 2.5x, 1,1x, 3.1x, and 2.3x indicate fold decrease of FITC fluorescence upon Treslin-RNAi compared with siCtr. (E) U2OS cell extracts were used for IPs with antibodies to TopBP1 or control (Ctr), and flow-throughs and IPs were analyzed by immunoblotting with antibody-to-Treslin, -MTBP, and -TopBP1 or with Ponceau staining (Pon.). (F) 293T cells were transiently transfected (16) with AcGFP-Flag-Treslin-WT or Treslin-2PM (16) and MTBP-Myc. Myc-IPs from native cell lysates (Input) were analyzed with immunoblotting using antibodies to GFP, Myc, and TopBP1.

MTBP interacted with the ∆CT mutant lacking the C-terminal tail and the 2PM mutant of Treslin/TICRR (Fig. 2B). The 2PM mutant lacks the two key CDK sites in the C terminus required for interaction of Treslin/TICRR with the multi-BRCT domain TopBP1 protein (15, 16). Therefore, the interaction of MTBP with Treslin/TICRR is not mediated by TopBP1. Endogenous MTBP coimmunoprecipitated with endogenous TopBP1 using antibodies to TopBP1 (Fig. 2E). However, TopBP1 was not coimmunoprecipitated with MTBP in cells transiently transfected with the 2PM mutant of Treslin/TICRR (Fig. 2F). Treslin/TICRR can thus interact simultaneously with TopBP1 and MTBP and bridges their interaction, which requires Treslin/TICRR phosphorylation by CDK.

MTBP depletion from HeLa cells with either of two siRNAs (siNo1 and siNo2) (Fig. 3A) reduced BrdU incorporation of cells in S phase without affecting Treslin/TICRR levels (Fig. 3A) and caused an accumulation of cells in late S phase (Fig. 3B, i, and fig. S6). These defects were suppressed by expression of an siRNA-resistant MTBP transgene fused at its C terminus to green fluorescent protein (GFP) (Fig. 3, C and D, and fig. S7), indicating that they are not off-target effects. An N-terminal GFP-MTBP fusion was unable to interact with Treslin/TICRR and did not suppress the replication phenotype of MTBP knockdown (fig. S8), providing additional evidence that the interaction between MTBP and Treslin/TICRR is critical for DNA replication.

Fig. 3 Depletion of MTBP results in slower replication and longer S phases.

(A) HeLa cells were treated with control (siCtr) or MTBP (siNo1/2) siRNA for 48 or 62 hours, and whole cells were analyzed by immunoblotting with antibodies to MTBP and Treslin or with Ponceau (Pon.) staining. (B) Cells of (A) were analyzed for BrdU incorporation and chromatin content by PI (1C or 2C: G1 or G2/M phase) by using flow cytometry. (i) Whole-population PI histogram plots (AU, arbitrary units). (ii) Histograms of S phase cells stained with a FITC-coupled antibody to BrdU (log, logarithmic scale). 2.4x and 2.9x indicate fold decrease of FITC fluorescence upon MTBP-RNAi for 48 hours. (C) HeLa cells and clones stably expressing siRNA-resistant MTBP-Flag-AcGFP were depleted of endogenous MTBP for 48 hours as in (A). Immunoblots of whole-cell lysates were analyzed with antibodies to MTBP (endog., endogenous; transg., transgenic) and Tubulin. (D) The experiment for (C) was repeated and analyzed with flow cytometry as in (B). (E) Live-cell imaging of HeLa cells stably expressing GFP-PCNA was used to quantify G1, S, and G2 phases of the cell cycle upon treatment with control (siCtr) or MTBP (siNo1/2) siRNAs. For technical reasons, S phase analysis is shown as the fraction longer than 10 hours, and G1 and G2 are in hours.

Using time-lapse microscopy of cells expressing GFP–proliferating cell nuclear antigen (PCNA), we found that the S phase (distinguished by the presence of multiple nuclear PCNA foci) was >10 hours in virtually all cells in which either MTBP or Treslin/TICRR had been depleted (Fig. 3E and fig. S9) but was almost never this long in control treated cells, in which the S phase averaged 8.3 hours. The length of the G1 phase was not affected in a consistent manner, but G2 was greatly shortened. G2 shortening may be a consequence of the greatly extended S phase or may reflect a second role for MTBP in G2/M control (17). We conclude that depletion of MTBP inhibits DNA replication during the S phase. The small amount of replication seen after MTBP siRNA may indicate that some replication is MTBP-independent but is more likely due to incomplete MTBP depletion.

MTBP was discovered as a factor that interacted in a two-hybrid assay with MDM2 (18, 19), the E3 ubiquitin ligase involved in p53 degradation. We found, however, that MTBP depletion inhibited replication in both p53-deficient HeLa cells (Fig. 3) and p53-positive U2OS cells (fig. S6), as well as p53-positive and -negative HCT116 cells (fig. S10A). In addition, inhibition of the MDM2 ubiquitin ligase with nutlin-3 did not alleviate the inhibition of DNA replication in p53-deficient cells upon MTBP RNA interference (RNAi) (fig. S10, B and C). These results indicate that MTBP is essential for DNA replication regardless of a cell’s p53 status and are consistent with previous work showing that MTBP is essential for early mouse development in both p53+/+ and p53−/− backgrounds (20). Although p53 was activated upon MTBP depletion (fig. S11) as previously reported (19), this happened at relatively late times and is likely to be a consequence of aberrant replication because cells accumulate in S and G2 phases rather than the G1 phase (Fig. 3B and figs. S5 and S6). Although p53 is not required for MTBP’s function in replication, it is intriguing to consider that MTBP might play a role in regulating p53, perhaps coupling some aspect of origin surveillance with G1/S progression.

We next used DNA fiber labeling (21, 22) to examine rates of DNA replication fork progression. Although MTBP knockdown reduced overall DNA synthesis (Fig. 3B), the lengths of remaining nascent DNA synthesis tracks were not shorter after MTBP knockdown (Fig. 4A), indicating that the defect in replication is likely due to reduced origin firing rather than reduced fork rate. Mcm2 bound to chromatin normally as cells progressed through mitosis and into G1 phase (3 hours after nocodazole release) after MTBP knockdown (Fig. 4B), indicating that MTBP is not required for mitotic progression or origin licensing. However, considerably less PCNA was associated with chromatin at later times (for example, 7 to 10 hours after release) when MTBP was depleted, similar to treatment with the CDK inhibitor roscovitine (lane 6). Levels of GINS (Fig. 4C) and Cdc45 (Fig. 4D), components of the CMG helicase, were also reduced on chromatin after MTBP knockdown. The decrease in PCNA, Cdc45, and GINS chromatin association is not due to a failure of progression into S phase because the E2F-induced cyclin A was expressed to the same level as in control cells (lanes 4 to 6). Therefore, MTBP is required after origin licensing for CMG assembly.

Fig. 4 MTBP is not required for replication elongation, licensing, and G1-S transition but is required for GINS, Cdc45, and PCNA recruitment to chromatin in S phase.

(A) MTBP was RNAi-depleted from HeLa cells as in Fig. 3A for 48 hours. Replication track lengths on chromatin fibers pulse-labeled with CldU were analyzed with fluorescence microscopy. Replication fork speeds are represented in kilobases per minute (kb/min) (error bars indicate SEM). (B) Control or MTBP-siRNA treated U2OS cells were released from a nocodazole block (Noc.-rel) for the indicated amounts of time. Chromatin-enriched fractions were analyzed for levels of Mcm2 and PCNA by means of immunoblotting and for histones by means of Coomassie (Coom.) staining. Rosc., roscovitine treatment from 3 hours after Noc-release. (C) Control cells or cells depleted of MTBP were synchronized in the G1 (3 hours nocodazole-release) or S phase (8 hours) as in (B), and chromatin or whole cells were analyzed by means of immunoblotting with antibodies against Mcm2, PCNA and Psf3, Cyclin A, MTBP, and Tubulin and by means of Ponceau staining. (D) Chromatin of cells treated as in (C) were analyzed by means of immunoblotting with a subset of antibodies used in (C) and with an antibody to Cdc45 and by means of Coomassie staining to detect histones (Hist).

MTBP may have a similar role to yeast Sld7 (23), given that both proteins interact with analogous regions of Treslin/TICRR and Sld3 and that a homolog of Sld7 has not been identified in humans. MTBP is amplified in >10% breast cancers and >20% ovarian cancers as part of a large amplicon on chromosome 8q24 that includes the MYC gene (24, 25). Mouse models have shown that MTBP haploinsufficiency significantly decreases myc-induced lymphomagenesis (26), suggesting that this step in DNA replication may be a useful drug target for some cancers.

Supplementary Materials

Materials and Methods

Figs. S1 to S12



References and Notes

  1. Acknowledgments: We are grateful to the LRI Protein Analysis and Proteomics facility for mass spectrometry analysis; M. Petronczki and colleagues for help with time lapse experiments; L. Sanchez-Pulido and C. P. Ponting for help with bioinformatics; J. Mendez, H.-P. Nasheuer, F. Grosse, and H. Pospiech for antibodies; and LRI Cell Services for assistance with cell lines. This work was supported by Cancer Research UK, European Research Council grant 249883-EUKDNAREP, Association for International Cancer Research grant 10-0270 (J.F.X.D.), and a European Molecular Biology Organization Long Term Fellowship (D.B).
View Abstract

Navigate This Article