Adenovirus E1A oncogene induces rereplication of cellular DNA and alters DNA replication dynamics

Abstract

The oncogenic property of the adenovirus (Ad) transforming E1A protein is linked to its capacity to induce cellular DNA synthesis which occurs as a result of its interaction with several host proteins, including pRb and p300/CBP. While the proteins that contribute to the forced induction of cellular DNA synthesis have been intensively studied, the nature of the cellular DNA replication that is induced by E1A in quiescent cells is not well understood. Here we show that E1A expression in quiescent cells leads to massive cellular DNA rereplication in late S phase. Using a single-molecule DNA fiber assay, we studied the cellular DNA replication dynamics in E1A-expressing cells. Our studies show that the DNA replication pattern is dramatically altered in E1A-expressing cells, with increased replicon length, fork velocity, and interorigin distance. The interorigin distance increased by about 3-fold, suggesting that fewer DNA replication origins are used in E1A-expressing cells. These aberrant replication events led to replication stress, as evidenced by the activation of the DNA damage response. In earlier studies, we showed that E1A induces c-Myc as a result of E1A binding to p300. Using an antisense c-Myc to block c-Myc expression, our results indicate that induction of c-Myc in E1A-expressing cells contributes to the induction of host DNA replication. Together, our results suggest that the E1A oncogene-induced cellular DNA replication stress is due to dramatically altered cellular replication events and that E1A-induced c-Myc may contribute to these events.

Figures

Fig 1

Levels of mRNA of genes related to initiation of DNA replication (A) and levels of replication initiation proteins present in whole-cell extracts (B) and chromatin-bound fractions (C). (A) RNA levels as determined by qPCR. MCF10A cells (3 × 106/plate) were seeded overnight, serum starved for 32 h, and then infected with Adb-gal or AdE1A at a multiplicity of infection (MOI) of 50. One set of plates infected with Adb-gal was stimulated with serum. Another set of Adb-gal-infected plates (control) and E1A-infected plates was maintained in serum-free medium. Cells were harvested at 16 h, the total RNA was isolated, and the mRNAs for each gene were quantified by qPCR assays as described in Materials and Methods. GAPDH was used as an internal control to normalize the values. The fold increase relative to b-gal control sample values was calculated by the 2−ΔΔCt method. Average values (with error bars) determined for RNA samples prepared in duplicate are shown. (B) Western blot showing the levels of several replication initiation proteins expressed in serum-stimulated and E1A-expressing cells. Western blot analysis was carried out using whole-cell extracts. (C) Chromatin loading of the indicated replication initiation proteins. For chromatin-loading assays, chromatin-bound proteins were isolated as described in Materials and Methods. Virus infection conditions in the experiments represented in panels B and C were as described for panel A. In panel C, histone H3 levels indicate that equal amounts of chromatin proteins were present in the chromatin extracts. Membranes in Western blots were probed sequentially with several antibodies. The PCR primers used for the qPCR assays represented in panel A are shown in Table S3 in the supplemental material.

Fig 2

Flow cytometric analysis of PI-stained quiescent E1A-expressing cells showing rereplication of cellular DNA in the late S phase. (A) Cell cycle profiles of quiescent cells expressing E1A at various time points. Serum-starved MCF10A cells were infected with Ad viruses for various time periods as indicated and maintained in serum-free medium. Cells were harvested at indicated time points, stained with PI, and then analyzed by flow cytometry. (B) Quantification of the PI-stained cells present in various cell cycle fractions. Cell cycle profiles for 24-h samples in panel A are not shown. M, mock-infected quiescent MCF10A cells. The 0-h sample data refer to cells infected with virus for 60 min and then harvested.

Fig 3

Flow cytometric analysis of cells pulse-labeled with BrdU, showing active DNA replication in the late S phase. Serum-starved cells were infected with Adb-gal and then serum stimulated (left) or quiescent cells were infected with AdE1A (right) for various time periods as shown and pulse-labeled for 20 min with 100 μg BrdU/ml before harvesting. Cells were analyzed for BrdU incorporation after fixation in 70% ethanol (EtOH). The BrdU-labeled cells were then processed as detailed in Materials and Methods.

Fig 4

DNA combing assay results, showing longer replication signals in E1A-expressing cells attributed to faster replication forks and increased interorigin distance. (A) Schematic images of the signals detected in a replicating DNA fiber stained with IdU (green) and CldU (red) and of the parameters analyzed in the single-molecule experiments. Fork velocity is measured as the average of the lengths of two consecutive signals divided by the time of the pulses. The total replicon length is the portion of DNA that replicates from one origin. The interorigin distance is the distance between two replication origins. (B) Representative images of DNA fibers (40× magnification). The top images show the density and the quality of the combed DNA fibers visualized after yoyo staining. The bottom images show the replication tracts as visualized by immunofluorescence after sequential IdU and CldU incorporation (see Materials and Methods). Serum-starved cells were infected with Adb-gal and then serum stimulated or infected with AdE1A and maintained in serum-free medium. The duration of infection was adjusted such that about 20% of each cell population was in the S phase. Cells were sequentially pulse-labeled with IdU and CldU for the last 60 min of the infection and analyzed as described in Materials and Methods. Replication tracts were visualized by immunofluorescence, and the images were translated to measurements as described in Materials and Methods. (C) Examples of images of two replication tracks with origins (Ori) from b-gal plus serum and E1A-expressing cells are shown (images are enlarged for clarity).

Fig 5

Summary of the DNA combing analysis, showing changes in fork velocity, replicon length, and interorigin distance of DNA replication patterns in E1A-expressing cells compared to serum-stimulated cells at both early (A) and late (B) time points. (A) Combing analysis of DNA at early S phase. Top: analysis of the fork velocity (Fv) distribution of the two samples calculated as (IdU Fv + CldU Fv)/2. Middle: analysis of the total replicon length. Bottom: analysis of the interorigin distance. For each sample, the mean value (Mean) and the number of observed events (N) are reported. The error bars indicate the 95% confidence intervals. Data sets were analyzed with the nonparametric test for unpaired data (Mann-Whitney). E1A-expressing cells and b-gal samples showed P < 0.001 for fork velocity and replicon length and P < 0.01 for interorigin distance. (B) Summary of DNA combing assay results determined for DNA samples isolated from cells harvested in late S phase. Details of the experiment were as described above with the exception that infection was allowed to proceed for 48 h. Top: analysis of the fork velocity distribution. Middle: analysis of the total replicon length. Bottom: analysis of the interorigin distance. Analysis of data sets and P values was performed as described for panel A.

Fig 6

Activation of the DNA damage response in cells infected with AdE1A. (A) Phosphorylation of Chk2 in E1A-expressing cells. Quiescent MCF10A cells were infected with the indicated viruses at 50 PFU/cell and harvested after 24 h. The indicated proteins were detected in Western immunoblots using equal amounts of protein. (B) Formation of γ-H2AX-containing foci in cells expressing E1A. Quiescent cells grown on coverslips were infected as described above and then analyzed for γ-H2AX-containing foci (green) by immunofluorescence as described in Materials and Methods. DNA was detected by staining with DAPI (blue). The percentages of γ-H2AX-positive cells were scored from three different fields, and the average value ± standard deviation (SD) is shown. At least 100 cells were counted in each field.

Fig 7

Myc and E1A protein levels in quiescent MCF10A cells infected with viruses as indicated and reversal of S-phase entry of E1A-positive cells by antisense Myc. (A) Schematic of the infection protocol. (B) Western immunoblot showing Myc protein levels in cells infected with Adb-gal, AdE1A, or AdE1A plus AdASMyc. Serum-starved cells were infected with Adb-gal (150 PFU/cell), Ad E1A (50 PFU/cell), or AdE1A after infection with AdASMyc (150 PFU/cell). Adb-gal virus was used for infection where appropriate to ensure that each group of cells received the same infection protocol. Cells were harvested 16 h after infection and then lysed, and Myc protein was detected in Western immunoblots. (C) E1A levels in serum-starved AdE1A- or AdE1A-plus-AdASMyc-infected cells at 16 h after infection. Infection conditions were as described above. (D) Myc activity levels determined using AdM4 promoter-reporter virus. Virus infections were as described above except that cells were also infected with AdM4 or AdΔM4 viruses (10 PFU/cell). AU, absorbance units. (E) Quantification of PI-stained cells in various cell cycle fractions at 24 h p.i. using flow cytometry.