L3MBTL1 polycomb protein, a candidate tumor suppressor in del(20q12) myeloid disorders, is essential for genome stability

Depletion of L3MBTL1 Inhibits Cell Proliferation and Causes G2/M Arrest.

To assess the role of L3MBTL1 in cell cycle regulation, we depleted L3MBTL1 mRNA and protein in U2OS cells using lentiviral vectors that expressed several shRNAs directed against its ORF. Down-regulation of L3MBTL1 mRNA and protein was achieved efficiently with three different shRNAs (≥90% knockdown; Fig. 1 A and B). The depleted cells were monitored for S phase entry by using BrdU incorporation, and we detected a marked decrease of S phase cells and the accumulation of cells in the G2/M phase (Fig. 1C and Fig. S1A). This effect was also observed in MRC5 normal diploid fibroblasts, Cal51, T98G, and K562 cells following L3MBTL1 knockdown (Fig. S1 B–D), demonstrating that it is not cell-type specific; it occurred in noncancerous cells as well as hematopoietic K562 cells. These findings indicate that these cell cycle effects are relevant to the myeloid compartment.

Depletion of L3MBTL1 Generates DNA Breaks.

Cell cycle alterations triggered by L3MBTL1 depletion suggest that cells may be experiencing replicative stress, leading to DNA damage. To verify this hypothesis, we examined two DNA damage markers, the production of the phosphorylated histone H2A.x (γH2A.x), which localizes to sites of double-stranded DNA breaks (DSBs) (13), and the distribution of 53BP1, which is recruited to DNA within minutes following DNA damage by binding to H4K20me2 and possibly other modified histones (14). We observed an increase in both γH2A.x and 53BP1 foci following L3MBTL1 depletion with all three shRNAs in U2OS (Fig. 2 A and B) and in MRC5 fibroblasts (Fig. S2A).

We also measured the formation of DNA strand breaks in L3MBTL1-depleted cells using the comet assay, which visualizes damaged DNA on a single-cell level (15). Comet tails were detected in depleted U2OS cells (Fig. 2C and Fig. S2B), MRC5 and Cal51 cells (Fig. 2D), indicating DNA strand breakage. We treated MRC5 fibroblasts with varying doses of gamma irradiation to establish a range of tail moments (Fig. 2D) and determined that depletion of L3MBTL1 caused extensive DNA damage, similar to that induced by 5 Gy of irradiation.

Loss of L3MBTL1 Activates the DDR and Affects H4K20 Methylation Status.

DNA damage triggers a checkpoint response that prevents cells from progressing into mitosis. To explore this pathway, we first examined the phosphorylation of the ATM kinase, one hallmark of the DDR pathway (16), and detected increased phospho-ATM foci that overlapped with 53BP1 foci in L3MBTL1-depleted cells (Fig. 3A). We also examined whether the downstream components of the ATR/ATM-triggered DDR pathway were activated following L3MBTL1 depletion using immunoblotting to monitor changes in phosphorylation of the key components of this system, including Chk1 and Chk2. As shown in Fig. 3B, increased levels of phospho-Ser317-Chk1, phospho-Ser345-Chk1, and phospho-Thr68-Chk2 were detected in cells at 24, 48, and 72 h postinfection, similar to that observed in cells irradiated with 10 Gy.

Depletion of L3MBTL1 altered the activities as well as the levels of a number of proteins associated with checkpoint regulation and the DDR pathway (Fig. 3 B–D). These include: (i) the formation of phospho-Tyr15-Cdc2, a modification that inhibits Cdc2 kinase activity, which is required for cell cycle progression into mitosis (17) (Fig. 3B and Fig. S3A); (ii) the up-regulation of p21 and p53, downstream effectors of the DDR pathway, as well as Rad51, which plays a key role in double-strand break repair by homologous recombination (Fig. 3C and Fig. S3B), and (iii) an increase in the level of γH2A.x (Fig. 3D and Fig. S3C), in keeping with the data shown in Fig. 2. We also observed an increase in the amount of H4K20me2 (40- to 90-fold), which plays important roles in DNA replication and DNA damage recognition (18), but detected little effect on the levels of H4K20me1. Conversely, overexpression of L3MBTL1 in U937 cells that have monoallelic loss of the commonly deleted region of 20q had the opposite effect, namely, a decrease in H4K20me2 and an increase in H4K20me1 (Fig. 3E). Collectively, these data show that the DDR pathway is strongly activated following depletion of L3MBTL1 and suggest that changes in the level of H4K20 dimethylation could contribute to these effects.

We, and others, have found L3MBTL1 complexed with Rb/E2F proteins repressed the activation of some E2F target genes, including c-myc and Cyclin E1, two established oncogenes (7, 8, 19, 20). We tested whether these genes were up-regulated in L3MBTL1 knockdown cells, but noted no changes in their levels at 48 h following lentiviral infection with shRNAs against L3MBTL1 (Fig. S3D). Thus, induction of the DDR following L3MBTL1 depletion cannot be explained by the enhanced expression of c-myc or Cyclin E1.

L3MBTL1 Depletion Slows the DNA Replication Forks Movement.

The presence of DNA damage and activation of the DDR suggest that depletion of L3MBTL1 may trigger defects in DNA replication. We examined DNA replication fork progression by DNA fiber analyses. For this purpose, control and L3MBTL1-depleted cells were sequentially pulsed for 1 h each with IdU and CldU, lysed, and DNA fibers spread on slides. The fibers were labeled with antibodies to IdU and CldU and fluorescence-labeled secondary antibodies. Fibers containing green IdU fluorescent label flanked on each side with the red CldU fluorescent label represent DNA molecules formed by bidirectional movement of replication forks. Fig. 4A shows DNA fibers isolated from control and L3MBTL1-depleted MRC5 fibroblasts. The length of the green IdU-labeled fibers, indicating DNA replicated from origins, was shorter in depleted cells than in fibers isolated from control cells. Furthermore, the total length of red CldU- and green IdU-labeled fibers was shorter in L3MBTL1-depleted cells than in control cells, indicating slower replication fork movement. The average length of DNA fibers from L3MBTL1-depleted cells and control cells was 16.8 μM and 31.7 μM, respectively (Fig. 4B); 78% of the DNA fibers from L3MBTL1-depleted cells were <20 μM, whereas 33% of fibers from control cells were this length (Fig. 4C). Similar data were obtained using U2OS cells, indicating that this effect is not cell-type specific (Fig. 4 D and E). The average rate of fork progression in L3MBTL1-depleted cells was ~40% slower than in control cells, indicating that L3MBTL1 is required for the normal progression of DNA replication forks.

L3MBTL1 Interacts with Components of DNA Replication Machinery.

Because L3MBTL1 appears to play a role in replication fork progression, we tested whether it interacted with components of the DNA replication machinery. We overexpressed HA-tagged L3MBTL1 in 293T cells, and examined its interaction with members of the MCM2-7 complex, a critical component of the putative replicative helicase. As shown in Fig. 5A, HA-tagged L3MBTL1 interacted with the MCM2-7 complex, and L3MBTL1 was immunoprecipitated by antibodies against MCM2 and MCM5 (Fig. 5B). Cdc45 and the GINS complex form a replicative helicase complex with MCM2-7 (CMG complex) (21) that travels with the replication fork (22). These findings prompted us to examine whether L3MBTL1 interacted with these components as well as other replication proteins. As shown in Fig. 5A, the interaction of L3MBTL1 with Cdc45 was detected, whereas an interaction with Sld5 (a subunit of the GINS complex) was not observed. Interactions with PCNA, the DNA sliding clamp required for processivity of the replicative DNA polymerases, were also noted, whereas interaction with Fen-1, an enzyme involved in DNA repair and lagging strand processing, was not observed (Fig. 5A). We also examined whether the levels of replication proteins were affected by depletion of L3MBTL1 in U2OS cells. As shown in Fig. S4, we did not find significant changes. Thus, L3MBTL1 interacts with a number of proteins that play roles in different stages of DNA replication, consistent with the notion that it affects replication. The mechanisms contributing to these effects remain to be investigated.

Lentiviral and Retroviral Transduction.

Lentiviral shRNA constructs were purchased from the TRC shRNA library at Thermo Scientific Open Biosystems and modified by subcloning the GFP gene instead of the puromycin gene into the pLKO.1 vector as a marker of viral integration. shRNA sequences against L3MBTL1 were gcctgcactttgatgggtatt (shRNA 1), gctggagtcatggctatgatt (shRNA 2), and gcagtcactcacaacaagaat (shRNA 3). shRNA with scrambled sequences in the same vector was used as a negative control. To produce lentiviruses, 293T cells were transfected with the shRNA constructs and helper plasmids (the envelope construct pMD2G and the packaging construct psPAX2) using the calcium phosphate method. To produce the retroviruses for the overexpression assays, Phoenix-A cells were transfected with either HA-tagged L3MBTL1 or empty MIGR1 vector construct. Lentiviruses or retroviruses were harvested by collecting and filtering supernatants 48 and 72 h after transfection, and concentrated by centrifugation at 4 °C for 2 h at 25,000 rpm in a Beckman L7-55 Ultracentrifuge. Target cells were infected with concentrated virus in the presence of 8 μg/mL polybrene by centrifugation at 450 × g for 45 min.

Cell Cycle Analysis by Flow Cytometry.

Cells were collected by trypsinization and fixed in 70% ethanol. They were washed with PBS, stained with 50 μg/mL propidium iodide and 0.5 μg/mL RNase A (both from Sigma-Aldrich), and analyzed by flow cytometry. For BrdU/propidium iodide staining, cells were incubated with 1 mM BrdU for 1 h and then stained with BrdU-APC antibody and propidium iodide, according to the instructions in the BrdU Flow Kit (BD Biosciences). Minimally, 10⁵ cells were analyzed on the BD FACSCalibur flow cytometer, and the data were collected using BD Cell Quest Pro software and analyzed using FlowJo software (Tree Star).

Immunoblotting and Coimmunoprecipitations.

U2OS cells were lysed in lysis buffer [20 mM Tris-HCl (pH 8), 0.5% Nonidet P40, 0.5% Tween 20, 1 mM EDTA, 1 mM DTT, 150 mM NaCl, 1 mM PMSF and protease inhibitor mixture (Calbiochem)], and then briefly sonicated and centrifuged. The protein concentration was measured using the Bradford assay (Bio-Rad), and equal amounts of protein were loaded on NuPage gels (Invitrogen). Protein transfer was conducted onto Immobilon-P PVDF membrane (Millipore), and the membranes immunoblotted with antibodies to the following proteins: phospho-Ser-317 Chk1, phospho-Ser354 Chk1, phospho-Thr68 Chk2, phospho-Tyr15 Cdc2, p53 (all from Cell Signaling), actin, Rad51, p21, geminin (all from Santa Cruz), α-tubulin (Sigma Aldrich), γH2A.x (Biolegend), H4K20me1, H4K20me3, H3 (all from Millipore). The anti-L3MBTL1 antibody was described previously (6). For coimmunoprecipitations, whole-cell extracts were prepared from 293T cells overexpressing HA-tagged L3MBTL1 using the same lysis buffer as above. The lysates for coimmunoprecipitations were treated with 1,000 units/mL of benzonase (Sigma Aldrich) overnight at 4 °C to degrade DNA and RNA. They were also treated with 50 μg/mL ethidium bromide to disrupt DNA-protein complexes and exclude DNA-mediated protein-protein interactions (42). The lysates were incubated overnight at 4 °C with HA-agarose beads (Roche) or normal rat IgG serum (Santa Cruz) followed by protein A agarose beads (Roche). The coimmunoprecipitates were washed four times with lysis buffer containing 250 mM NaCl before immunoblotting for the following proteins: MCM2-7 (all from Bethyl Laboratories), PCNA (Millipore), Cdc45, Fen-1, Sld5 (all from Santa Cruz). For reciprocal coimmunoprecipitation, Flag-tagged L3MBTL1 was overexpressed in 293T cells, and lysates treated as above were immunoprecipitated with antibodies against Mcm2, Mcm5, or rabbit IgG, followed by adsorption to protein A agarose beads.

Quantitative RT-PCR.

U2OS cells were harvested and RNA prepared by RNeasy Plus Mini Kit (Qiagen); cDNA was synthesized using SuperScript III RT-PCR kit (Invitrogen). Quantitative PCR was conducted on an ABI7500 Real-Time PCR system using SYBR Green PCR mix (both from Applied Biosystems) and primers corresponding to the genes of interest. Primer sequences are available upon request.

Comet Assay.

DNA breaks were measured using the Comet Assay Reagent Kit (Trevigen) with minor modifications. Briefly, cells embedded in low-melting point agarose on a slide were lysed, treated with alkali, and electrophoresed in 1× TBE buffer for 40 min at 1 V/cm, stained with propidium iodide, and imaged under the fluorescent microscope. To evaluate the extent of DNA damage present in each sample, the average tail moment (a product of percent DNA in tail and tail length) of 100 cells was calculated using CometScore software (Tritek).

Immunocytochemistry.

Cells were plated on coverslips, then fixed in methanol, permeabilized with 0.25% Triton-X/PBS, blocked in 2% FBS/PBS, and incubated with primary antibody for 1 h, followed by secondary antibody conjugated with a fluorescent probe for 30 min, both at room temperature. Cells were stained with DAPI, mounted onto coverslips in mounting medium (Vector Labs), and imaged on Leica Upright confocal microscope or Carl Zeiss fluorescent microscope using appropriate filters. The primary antibodies used were directed against γH2A.x (Millipore), 53BP1 (Novus Biologicals), and phospho-Ser1981 ATM (Cell Signaling), and secondary antibodies used were anti-rabbit Alexa488 and anti-mouse Alexa568 (both from Invitrogen). Quantitation of the DNA damage foci was determined by scoring foci of at least 100 cells per treatment used.

DNA Fiber Analyses.

Control and L3MBTL1 knockdown U2OS and MRC5 cells were sequentially labeled with 50-μM IdU and 250-μM CldU for 1 h each. DNA fiber spreads were prepared as described previously with some modifications (43). Briefly, three aliquots of cells were resuspended in PBS at 1 × 10⁶ cells/mL, spotted onto a microscope slide, and lysed with 15 μL of spreading buffer [0.5% SDS in 200 mM Tris-HCl (pH 7.4), 50 mM EDTA]. After 6 min, slides were tilted 15° to allow lysates to slowly move down the slide, and the resulting DNA spreads were air-dried, fixed in 3:1 methanol/acetic acid, and stored at 4 °C overnight. The slides were then treated with 2.5 M HCl for 30 min, incubated in blocking buffer (1% BSA/0.05% Tween 20/PBS) for 1 h followed by 1 h at room temperature with 1:500 rat anti-BrdU antibody (Abcam; to detect CldU) plus 1:100 mouse anti-BrdU (BD Biosciences; to detect IdU), diluted in blocking buffer, followed by incubation for 30 min in 1:350 Alexa488-conjugated goat anti-rat antibody and 1:350 Alexa595-conjugated goat anti-mouse antibody (Invitrogen). Slides were air-dried and mounted in Prolong Gold antifade reagent (Invitrogen). Microscopy was carried out using a Zeiss LSM 5 Live Confocal microscope, and the fiber lengths were measured using Volocity software (PerkinElmer).

Metaphase Spreads.

Cal51 cells were treated with Karyomax solution (Invitrogen) for 5 h and then lysed with 75 mM KCl. Cells were fixed with a 3:1 methanol/acetic acid solution, dropped onto slides, dried, and stained with DAPI. The metaphase chromosomes were imaged under the Carl Zeiss microscope and evaluated for abnormalities such as chromatid and chromosome breaks, gaps, and exchanges.