Stalled fork rescue via dormant replication origins in unchallenged S phase promotes proper chromosome segregation and tumor suppression

Abstract

Eukaryotic cells license far more origins than are actually used for DNA replication, thereby generating a large number of dormant origins. Accumulating evidence suggests that such origins play a role in chromosome stability and tumor suppression, though the underlying mechanism is largely unknown. Here, we show that a loss of dormant origins results in an increased number of stalled replication forks, even in unchallenged S phase in primary mouse fibroblasts derived from embryos homozygous for the Mcm4(Chaos3) allele. We found that this allele reduces the stability of the MCM2-7 complex, but confers normal helicase activity in vitro. Despite the activation of multiple fork recovery pathways, replication intermediates in these cells persist into M phase, increasing the number of abnormal anaphase cells with lagging chromosomes and/or acentric fragments. These findings suggest that dormant origins constitute a major pathway for stalled fork recovery, contributing to faithful chromosome segregation and tumor suppression.

Figure 1. Mcm4^{Chaos3/Chaos3} cells have reduced amounts of the MCM2-7 proteins on chromatin, resulting in a reduced number of dormant origins

(A) All components of the MCM2-7 complex are significantly reduced in Mcm4^{Chaos3/Chaos3} cells. Reduction levels of chromatin-bound MCM2/4/7 as well as other replication proteins in Mcm4^{Chaos3/Chaos3} (C3) cells were estimated by referencing wildtype (WT) proteins loaded in different amounts (left). In Mcm4^{Chaos3/Chaos3} cells, the MCM3/5/6 proteins were also reduced in both the chromatin fraction and whole cell extract (WCE) (right). Protein samples were obtained from cells cultured asynchronously. Actin and stained membranes were used as loading controls. (B) Schematic presentation of consecutive dual labeling in the DNA fiber assay. Replication forks were labeled with digoxigenin-dUTPs (dig-dUTPs, red) for 20 min followed by biotin-dUTPs (green) for 30 min. (C) There is no significant difference in the average density of active origins between wildtype and Mcm4^{Chaos3/Chaos3} cells in untreated conditions (UNT). However, APH treatment induced a significantly lower origin density in Mcm4^{Chaos3/Chaos3} cells (p<0.05, t-test). These values were determined by measuring the distances between adjacent origins as shown (left). Bars show standard error of the mean (SEM). (D) Mcm4^{Chaos3/Chaos3} cells show a significant increase in the frequency of asymmetric forks (see image on the left). The average frequencies are shown with SEMs and are compared by χ2-test (p<0.005, see asterisk).

Figure 2. Elevated levels of RPA, pRAD17 and γH2AX foci formation are observed in Mcm4^{Chaos3/Chaos3} cells

(A) An increased number of Mcm4^{Chaos3/Chaos3} cells are positive for RPA32 and pRAD17 (Ser645) foci. Shown are the average percentages of cells positive for each marker in the untreated (UNT) and APH (300nM for 24 hrs) treated conditions. Bars are SEMs for ten different fields obtained from two independently performed experiments. Representative images are shown on the right with a magnified view of the selected nuclei. Nuclei are stained with DAPI (blue). Scale bars are 40 μm. (B) Mcm4^{Chaos3/Chaos3} cells show a slight increase in the formation of γH2AX foci in S phase. Shown are the average percentages of cells positive for γH2AX foci and the distribution of the number of γH2AX foci per cell in the untreated condition (left). S phase cells were detected by incorporation of dUTPs. Bars are SEMs for three independent experiments. Representative images are shown on the right. Scale bars are 10 μm.

Figure 3. The mutant MCM2-7 complex is unstable but retains proper helicase activity in the CMG complex

(A) Silver stained 10% SDS-polyacrylamide gels (PAGE) show purified wildtype and mutant CMG complexes. (B) An autoradiograph of a helicase assay showing the radiolabeled products separated by PAGE (top). M13 circular DNA annealed with a radiolabeled oligonucleotide was used as a substrate. The migration of double stranded substrate and displaced oligo is shown with arrows. The amount of protein in femtomoles in each reaction is indicated. The first two lanes show the completely denatured substrate and the substrate with no protein. The reactions were performed in duplicate and quantified as a percentage of substrate processed as shown in the bottom graph. Error bars show standard deviations. (C) The salt elution profiles of wildtype and mutant MCM2-7 from Mono Q anion exchange chromatography show the relative instability of the mutant complex (top). Blue and red lines show the relative absorbance at 280 nm for the wildtype and mutant gradient runs, respectively (left Y axis); the grey axis indicates salt concentration (right Y axis). The peak protein fractions from the mutant gradient were separated by SDS-PAGE and stained with coomassie brilliant blue (bottom); the fraction numbers are shown above each lane. The starting stoichometric mutant MCM2-7 complex that was loaded onto this column is shown in the first lane and the pooled peak fractions (15 and 16) from the WT gradient are shown in the last lane.

Figure 4. An increase in RAD51 and BLM foci formation in Mcm4^{Chaos3/Chaos3} cells does not lead to a significant increase in homologous recombination events

(A) An increased number of Mcm4^{Chaos3/Chaos3} cells are positive for RAD51 foci. (B) BLM foci formation is drastically elevated in Mcm4^{Chaos3/Chaos3} cells. Shown in A and B are the average percentages of cells positive for ≥2 foci in the untreated and APH treated conditions (left). The average number of foci per cell is also shown. Bars are SEMs for ten different fields obtained from two independently performed experiments. Representative images are shown on the right. Scale bars are 20 μm. (C) No significant increase in HR events was detected at the FYDR locus in Mcm4^{Chaos3/Chaos3} cells compared to wildtype. Bars are SEMs for recombinant frequencies determined by analyzing at least 16 embryos per genotype. (D) No significant increase in HR events was detected at the FYDR locus in either wildtype or Mcm4^{Chaos3/Chaos3} cells after a low dose of APH treatment. CPT and a higher dose of APH were used as positive controls. Bars are SEMs for recombinant frequencies determined by analyzing at least 3 independent MEF lines.

Figure 5. Mcm4^{Chaos3/Chaos3} cells have a drastically increased number of FANCD2 sister foci at prophase, preceding abnormal anaphase and micronucleation

(A) An increased frequency of FANCD2 sister foci is found in Mcm4^{Chaos3/Chaos3} cells. The average percentages of cells positive for FANCD2 sister foci (top) and the average numbers of FANCD2 sister foci per cell (bottom) are shown with SEMs. Note that the number of FANCD2 sister foci per cell increases approximately 2-fold in Mcm4^{Chaos3/Chaos3} cells compared to wildtype cells, while nearly all cells become positive for such foci in the presence of APH. Representative images are shown on the right with a magnified view (indicated by squares). (B) An increased number of Mcm4^{Chaos3/Chaos3} cells undergo abnormal anaphase, forming micronuclei (MN). The average frequencies of abnormal anaphases containing lagging chromosomes and/or fragments are shown for wildtype and Mcm4^{Chaos3/Chaos3} cells with SEMs (left, top). B6 MEFs were used for anaphase analysis. The average MN frequencies are also shown for wildtype and Mcm4^{Chaos3/Chaos3} cells with SEMs (left, bottom). MN were detected using the cytokinesis-block micronucleus assay (Fenech, 2007). Representative images are shown on the right for a normal anaphase, an abnormal anaphase containing lagging chromosomes, a normal binucleated cell and one with a micronucleus. Scale bars are 5 μm. (C) Mcm4^{Chaos3/Chaos3} cells have an increased number of centromeric (CENP-A⁺) MN compared to wildtype cells. The average frequencies of CENP-A⁺ MN were determined from three independently performed experiments and are shown with SEMs. Representative images are shown on the right. (D) G-banding analysis of metaphase chromosomes shows no evidence for increased chromosome breaks but does reveal an increased occurrence of translocations in Mcm4^{Chaos3/Chaos3} cells. Representative karyotypes containing translocations are also shown bottom. (E) Mcm4^{Chaos3/Chaos3} cells have an increased number of γH2AX-foci-positive (γH2AX⁺) MN compared to wildtype cells. The average frequencies of γH2AX ⁺ MN were determined from three independently performed experiments and are shown with SEMs. Representative images are shown on the right for MN positive and negative for γH2AX (red) in binucleated cells. APH was used as a positive control (150 nM for 24 hrs). Nuclei are stained with DAPI (blue). Scale bars (A, C and E) are 10 μm.

Figure 6. Many different types of spontaneous tumors are observed in Mcm4^{Chaos3/Chaos3} mice

(A) Tumor-free survival curves for B6 and F1 Mcm4^{Chaos3/Chaos3} (C3) and wildtype (WT) females. (B) Representative images of histiocytic sarcomas. A hematoxylin and eosin (H&E) stain (top) shows a diffusely hypercellular liver area with round cells (black arrowheads) smaller than hepatocytes (unfilled arrowheads). Immunohistochemistry (IHC) with Mac-2, a macrophage marker, identified these small round cells (stained brown) as histiocytes (bottom). (C) Representative images of F1 Mcm4^Chaos3 gastrointestinal lymphomas, including an H&E image (top) and an IHC image with B220, a B-cell marker (bottom). (D) A model for chromosome instability driven by a loss of dormant origins. A lack of dormant origins increases the frequency of unresolved replication intermediates marked with FANCD2 sister foci (red diamonds) in M phase. Although such lesions can be rescued in a Fanconi pathway-dependent manner, a significant fraction persists into anaphase, interconnecting sister chromatids. As a result, the disjunction of sisters is disrupted and lagging chromosomes occur. This has three possible consequences: 1) Tetraploidy may occur due to cytokinesis failure, when the frequency of nondisjunction is high (Shi and King, 2005), 2) Aneuploidy could occur due to the nondisjunction of a few sisters, forming MN positive for CENP-A, or 3) Breaks may arise when unresolved replication intermediates are converted into DSBs, generating acentric fragments. In this case, MN positive for γH2AX foci would be formed. These aberrations lead to multiple types of chromosome instability, thereby contributing to tumorigenesis.