Inactivation of checkpoint kinase 1 (Chk1) during parvovirus minute virus of mice (MVM) infection inhibits cellular homologous recombination repair and facilitates viral genome replication

ABSTRACT

During infection, the autonomous parvovirus minute virus of mice (MVM) induces cellular DNA breaks and localizes to such sites, which presumably affords an environment beneficial for genome replication. MVM replication also benefits from the DNA damage response (DDR) mediated by the ataxia-telangiectasia mutated (ATM) kinase, while the ataxia telangiectasia and Rad-3 related (ATR) arm of the DDR is disabled, which prevents activation of its primary target, checkpoint kinase 1 (Chk1). We find here that Chk1 inactivation strongly correlates with dephosphorylation of one of its targets, RAD51, known to play a pivotal role in homologous recombination repair (HRR), thus leading to substantial inhibition of DNA repair in infected cells. We demonstrate colocalization of replicating MVM DNA with cellular double-strand breaks (DSBs) during infection, and show that an agent that exogenously induces cellular DSBs significantly increases viral DNA replication levels, establishing a role for cellular genome damage in facilitating virus DNA replication. Additionally, overexpression of active Chk1 during MVM infection was found to re-establish the activating phosphorylation of RAD51 Thr 309, significantly suppress infection-induced reduction of HRR efficiency with a concomitant increase in cellular genome DSBs, and reduce viral DNA replication levels. Thus, we conclude that during infection, MVM inhibition of Chk1 activation enhances viral replication, at least in part, by inhibiting cellular HRR.

IMPORTANCE

The autonomous parvovirus minute virus of mice (MVM) has a compact DNA genome encoding a minimum number of proteins. During infection, it induces cellular DNA damage and both utilizes and modifies the subsequent cellular DNA damage response (DDR) in various ways to facilitate its replication. One of MVM’s activities in this regard is to inhibit one of the primary arms of the DDR, the ataxia telangiectasia and Rad-3 related (ATR) pathway, which prevents activation of checkpoint kinase 1 (Chk1), a key protein involved in controlling the cellular DDR and preserving genome integrity. We show that prevention by MVM of Chk1 activation leads to inhibition of homologous recombination repair (HRR) of cellular DNA, which helps sustain viral replication. This work illuminates another way in which autonomous parvoviruses adjust the cellular environment for their replicative advantage.

INTRODUCTION

Minute virus of mice (MVM) is an autonomously replicating parvovirus. It has a non-enveloped icosahedral capsid containing an approximately 5 kilobase (kb), negative-sense, single-stranded (ss) DNA genome (1). The genome is flanked by non-identical inverted terminal palindromes, which facilitate replication by a rolling-hairpin mechanism involving synthesis of double-stranded (ds) monomer replicative forms (mRFs, ∼5 kb), dimer replicative forms (dRFs, ∼10 kb) containing two copies of the original genome, as well as additional DNA structures (2). The MVM genome is arranged into two overlapping transcription units, producing nonstructural (NS1 and NS2) and capsid proteins (reviewed in reference 3). The 83-kDa phosphoprotein NS1 plays a number of roles essential for viral replication, while the roles of the smaller NS2 isoforms are less clear and are dispensable in some hosts (3). MVM relies heavily on cellular DNA replication machinery expressed during S-phase of dividing cells for its own DNA replication (3). Although MVM cannot induce cells to enter S phase, the virus causes a cell cycle arrest prior to mitosis, establishing a favorable environment for virus replication (4).

Previous research has demonstrated that as MVM infection proceeds, viral DNA tends to associate with common fragile sites of the host genome, which are prone to high levels of DNA breaks as cells progress through S-phase (5). MVM replicating genomes are localized to these sites via the MVM genome-binding protein NS1 (6), possibly, because factors such as DNA polymerase δ and its partners, which have been demonstrated to support MVM DNA replication (7, 8), are found in proximity with DNA breaks (9). The progression of MVM infection is associated with the continuing accumulation of cellular DNA breaks, evidently due, at least in part, to the ability of autonomous parvoviruses to induce the production of reaction oxygen species (ROS) (10, 11). This is accompanied by the concurrent spread of viral DNA to newly formed cellular genome damage sites (5). Therefore, the induction of cellular DNA breaks has been suggested to facilitate virus replication (5).

Many DNA viruses have been shown to trigger and engage with host DNA damage response (DDR) machinery, sometimes inactivating these responses, and sometimes exploiting them to enhance viral replication (reviewed in reference 12). MVM infection causes, and benefits from, a robust ataxia-telangiectasia mutated (ATM)-driven DDR, typically activated as a response to DNA double-strand breaks (DSBs) (13). While many cellular DNA DSBs are repaired by the non-homologous end joining (NHEJ) pathway, during S phase, a significant portion of these are eliminated by homologous recombination repair (HRR) (14). HRR involves resection of repaired breaks with consequent formation of single-stranded replication protein A (RPA)-covered DNA (reviewed in reference 15). These structures normally attract ataxia telangiectasia and Rad-3 related (ATR) protein along with its partners, which activate an ATR-dependent DDR, resulting in activation by phosphorylation of ATR’s main target, the checkpoint kinase 1 (Chk1), attached to the genomic DNA in its inactive conformation (reviewed in reference 16). A principal function of Chk1 following activation by ATR is the induction of a G2/M cell cycle block following DNA damage (16); however, an additional important role of activated Chk1 is the phosphorylation of the DNA repair protein RAD51 (17), which performs crucial tasks in HRR (reviewed in reference 15). In contrast to ATM-driven DDR, in advanced stage of MVM infection, the ATR-dependent DDR pathway is inhibited and Chk1 surprisingly remains inactive (18).

In this work, we provide evidence that inactivation of Chk1 following MVM infection resulted in the loss of RAD51 phosphorylation, causing a significant decrease in levels of HRR. We confirmed the assumption that host DNA breaks facilitate MVM DNA replication, and that MVM-induced inhibition of HRR contributed to an accumulation of cellular DNA DSBs, facilitating an environment that supported increased viral replication.

MATERIALS AND METHODS

Cells, synchronization, drug treatments, and virus

The mouse A9 (ATCC #CCL-1.4) cell line was cultured as monolayers in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 5% Serum Plus Medium Supplement (Sigma-Aldrich) at 37°C in a 5% CO₂ incubator. To para-synchronize the cells in G0, they were seeded and grown in the supplemented DMEM for 24 h. Then, the medium was replaced with isoleucine-free DMEM (19), supplemented with 5% dialyzed, i.e., devoid of amino acids, fetal calf serum (FCS) and incubated for 48 h under the cultivation conditions. Neocarzinostatin (NCS) solution in 20 mM MES (Sigma-Aldrich #N9162) was used to treat cells with 300 ng/mL concentration for 30 min to induce DNA damage. All cell treatments were performed in DMEM at 37°C in a 5% CO₂ incubator. MVM prototypic strain stock was prepared and quantified as described (20) and was used for all infections.

Plasmid constructs

Mouse wild-type (WT) Chk1 expression vector (pChk1-WT) was constructed by PCR amplification of Mus musculus Chk1 coding sequence (cds) (RefSeq: NM_007691.4) from vector pMD-mCHEK1 (Sino Biological #MG50248-M) with primers acggcaagtgggcggccgcggcagtgccttttgtgg and tcctctggtcgactcatgtaacaggaaaccaaacc, and cloning the resultant product into NotI and SalI sites of the p3xFLAG-CMV-7.1 expression vector (Sigma-Aldrich #E7533). An Invitrogen pcDNA3.1-based vector expressing solely MVM nonstructural protein 1 (NS1) was described previously (21). Plasmids, described below, were modified by introduction of single amino acid substitutions with QuikChange II (Agilent Technologies) Site-Directed Mutagenesis (SDM) Kit. Constitutively active mouse Chk1 (pChk1-CA) expression vector was constructed by introduction of the protein L449R substitution into pChk1-WT with SDM primers cgactttctaagggtgatggaagagagttcaagagacacttc and gaagtgtctcttgaactctcttccatcacccttagaaagtcg. Kinase-deficient mouse Chk1 vector (pChk1-KD) was constructed by introduction of the protein D148A substitution into pChk1-WT with SDM primers taacctcaaaatctctgcctttggcttggcaacgg and ccgttgccaagccaaaggcagagattttgaggtta.

Transfections and infections

All DNA transfections with plasmid constructs listed above were performed with SignaGen LipoD293 (#SL100668) according to the manufacturer’s protocol. To increase the proportion of the transfected cells selected for analyses, harvested cells were sorted using a transfected marker-expressing vector. For this purpose, 1 µg of an expression vector encoding pDsRed (Takara Bio #632429) for the comet assays (Fig. 4), or pEGFP (Clontech #6085-1) for MVM DNA replication assays (Fig. 7), was added to the transfection solution containing 4 µg of expression vector encoding a protein of interest per one 10-cm cell culture dish. Similarly, for the fluorescence-activated cell sorting (FACS) experiment (Fig. 3), the sorting was possible due to immunostaining of the FLAG-tag, fused with Chk1 coding sequence, in the transfected vector, as described below.

MVM infections were performed as described previously (22). By the end of the procedure, the infection inocula were replaced with the supplemented DMEM containing 0.1 mg/mL of neuraminidase (Sigma-Aldrich) to prevent the reinfection, thus ensuring single-cycle infection.

Briefly, to conduct the experiments involving the para-synchronization and MVM infection, the cells were seeded, grown under standard conditions (see above) for 24 h, and para-synchronized by isoleucine starvation for 48 h. Then, MVM infection was performed, and the infected cells were incubated in supplemented DMEM under the growth conditions for the desired periods of time. In order to perform infection of the para-synchronized cells overexpressing Chk1, co-transfection of Chk1 and a fluorescent protein expressing vector was carried out during the para-synchronization. Subsequently, sorting cells expressing the fluorescent marker with the Thermo Fisher Bigfoot Cell Sorter allowed selection of the transfected cells prior to the comet (Fig. 4) and Southern (Fig. 7) assays. For FACS analysis, an additional DNA repair reporter vector was transfected after the infection, as described below (Fig. S1).

SDS-PAGE and immunoblots

Total protein extraction from A9 cultures was performed with RIPA buffer (Thermo Fisher Scientific #89900) and quantified by Bradford assay (Bio-Rad #5000001). Total protein (10 µg) from different samples was resolved by discontinuous 10% SDS-PAGE and electrophoretically transferred to nitrocellulose membranes (23) using a Bio-Rad Mini-PROTEAN system. Following the transfer, the membranes were blocked in 5% low-fat dry milk and probed with primary antibodies against examined proteins (see below), washed with 0.3% Tween20 in 1× phosphate-buffered saline (PBS), and the bound primary antibodies were reacted with secondary horseradish peroxidase (HRP)-conjugated antibody. The detection was achieved by reaction with enhanced chemiluminescent substrate (Thermo Fisher Scientific #A38555), followed by exposure to ECL film (Cytiva #28906835).

MVM NS1 was detected with a mouse monoclonal antibody (5). Commercially available primary antibodies were anti-FLAG (Sigma F1804), anti-Chk1 (Abcam ab32531), anti-phospho-S345-Chk1 (Cell Signaling 2348), anti-RAD51 (Abcam ab133534), anti-phospho-T309-RAD51 (LS Bio LS-C381696), anti-Cdc25A (Sigma SAB4300430), anti-phospho-S76-Cdc25A (Sigma SAB4300114), and anti-Actin (Thermo Fisher Scientific MA5-11869). Abcam goat anti-rabbit (ab205718) and anti-mouse (ab205719) HRP-conjugated IgG H&L were used for the detection of rabbit and mouse-raised primary antibodies, respectively.

FACS

Cells were para-synchronized, transfected with FLAG-tagged Chk1 expressing vectors, and infected with MVM, as described above. Sixteen or 24 h postinfection, the cells were co-transfected with HRR reporter vector pre-cut with HindIII (24), here designated as pHRR, and pDsRed. Following incubation for 24 h under growth conditions (see above), the cells were trypsinized, fixed with 4% paraformaldehyde (PFA), permeabilized with 0.1% Triton X-100, and stained with Invitrogen FLAG-Tag Monoclonal Antibody (L5), Alexa Fluor 647 (#MA1-142-A647). FACS analysis was performed with BD Accuri C6 Flow Cytometer. Initially, sequential gating was caried out to identify single A9 cell populations, based on their empirically determined size and granularity (FSC vs SSC), and Alexa Fluor 647-positivity, i.e., Chk1-overexpressing cells. Then, EGFP/DsRed-expressing cells, i.e., the repair/transfection frequency ratio, were determined. Data were analyzed with Flowing Software (Turku Bioscience).

Southern analysis

To generate the probe for detecting MVM DNA, the cloned virus genome (25) was PCR amplified with two pairs of primers: ccatggctggaaatgcttactc & ggctgttaccaaccatctgctc and gccatttgctctggtcaaactat & tggcgtccttggtttggtca, producing non-overlapping 482- and 962-bp products, respectively. The resulted PCR products were purified with QIAquick PCR Purification Kit and were labeled with 2 µCi/µL α-³²P-dATP using random priming mix (Sigma H0268) according to the manufacturer’s manual. Similarly, to produce mouse mitochondrial (mt) DNA probe, total DNA was extracted from A9 cells monolayer (26), and the mtDNA was PCR amplified with primers accttgcctagccacacccc and tgaggtagagcggggtttatcg. The resulting 450-bp product was labeled as mentioned previously. In addition, 1 kb DNA Ladder (NEB N3232) and λ DNA-HindIII marker (NEB N3012) were likewise labeled with α-³²P-dATP. Total DNA from MVM-infected cells was extracted, quantified, resolved with 0.8% agarose electrophoresis (500 ng per well), transferred to nylon membranes, and hybridized with the labeled probes, as described previously (27). Parallel runs for each sample were performed for separate probing with the mixed labeled PCR-amplified MVM DNA and 1 kb DNA Ladder, as well as mixed mtDNA and λ DNA-HindIII probes, allowing MVM DNA quantification and the loading control, respectively. Following the hybridization, the images were obtained with PharosFX Molecular Imager (Bio-Rad), and MVM replicative forms were quantified with Quantity One software (Bio-Rad). Mean values from three independent experiments were used to plot each data set.

DNA breakage detection–fluorescence in situ hybridization (DBD-FISH) for simultaneous detection of cellular genome breaks and replicating MVM DNA

To generate the probe for detecting the cellular genomic DNA, total DNA from A9 mouse fibroblasts monolayers was extracted (26), digested with restriction enzyme HaeIII, and labeled with DIG-11-dUTP by Nick translation (28). To detect the replicated MVM DNA plus strand, two non-overlapping fragments of cloned MVM genome (25) were PCR amplified with primer pairs ccatggctggaaatgcttactc & ggctgttaccaaccatctgctc and gccatttgctctggtcaaactat & tggcgtccttggtttggtca, producing non-overlapping 482 and 962 products, respectively. Following purification with QIAquick PCR Purification Kit, 1 µg of each product was incubated in primer elongation mix containing 1 mM of each dATP, dCTP, dCGP, dCTP, Biotin-11-dUTP (Thermo Fisher Scientific), 1 u of NEB Taq DNA polymerase with its standard buffer (NEB M0273) and primers designed for synthesis of only minus MVM DNA strand of both fragments, i.e., ggctgttaccaaccatctgctc (for the 482-bp product) and tggcgtccttggtttggtca (for the 962-bp product). Thus, the resultant elongation reaction produced exclusively a minus MVM DNA strand with incorporated Biotin-11-dUTP molecules.

To perform mild alkaline denaturation-based DBD-FISH (29), briefly 10⁵ A9 cells were seeded onto cover slips in 24-well plate and grown in supplemented DMEM for 24 h, then the para-synchronization and MVM infection were carried out as described above. At 16 and 24 hours postinfection (hpi) time points, the cells were fixed with 4% PFA, permeabilized with 0.1% Triton X-100. The alkaline denaturation was performed in 0.03 M NaOH, 1 M NaCl solution for 5 min at room temperature (RT), followed by neutralizing and protein lysing in 0.4 M Tris, 2 M NaCl, 1% SDS, pH 7.2 for 30 min at RT. The same conditions were empirically determined as sufficient for MVM DNA denaturation. The hybridization was achieved by incubation of the alkaline-denatured samples with temperature-denatured (100°C for 5 min) 1 µg cellular and 1 µg MVM DNA probes (equal molar amounts of both 482- and 962-bp derived probes) supplemented with 25 mg/mL sheared and denatured herring sperm DNA in 50% deionized formamide, 10% dextran sulfate, 2× saline-sodium citrate buffer (SSC) pH 8 at 37°C for 16 h. Following extensive washing with 2× SSC, the samples were stained with a mixture of 1 µg/mL anti-DIG-Rhodamine sheep Fab fragments (Roche) and 10 µg/mL Streptavidin-Alexa Fluor 488 (Thermo Fisher Scientific) in 1.5% bovine serum albumin (BSA), 1× PBS for 1 h at 37°C, washed in 1.5% BSA, 1× PBS, and mounted for microscopy.

Comet assay

A9 cells were seeded onto 10-cm dishes, para-synchronized, co-transfected with one of the FLAG-tagged Chk1 expression vectors pChk1-WT, pChk1-CA, pChk1-KD, or an empty FLAG-tagged vector and pDsRed, and infected with MVM at a multiplicity of infection (MOI) of 10 (see above). Following 16, 24, or 36 h of growth in supplemented DMEM, the cells were trypsinized and sorted for the presence of DsRed, as described previously. The sorted cells were embedded into low-temperature melted agarose and subjected to the assays with Trevigen’s CometAssay Kit (#4250-050-K), under the neutral assay conditions, according to the manufacturer’s manual. For application of FISH to the comet assays, a published protocol (30) was adapted. MVM DNA probe was produced by PCR amplification of cloned MVM genome (25) with primer pairs ccatggctggaaatgcttactc & ggctgttaccaaccatctgctc (producing 482-bp product) and gccatttgctctggtcaaactat & tggcgtccttggtttggtca (962-bp product) in the reaction mixture containing a pair of the named primers (0.5 µM each), 1 u of NEB Taq DNA polymerase with its standard buffer (NEB M0273), 200 µM of each dATP, dCTP, and dGTP, 100 µM dTTP, and 100 µM Biotin-11-dUTP. Total cellular (A9) genomic DNA was labeled by Nick translation as described above. The FISH procedure was performed following the regular comet assay, on the same Trevigen’s CometAssay Kit slide, according to the FISH protocol described above, except the DNA denaturation, which was performed by incubation of the slide in 95% deionized formamide at 65°C for 10 min.

RESULTS

MVM infection prevented the activating phosphorylation of RAD51

Prolonged MVM infection was previously found to disable the ATR pathway, either in the presence or in the absence of genotoxic stimuli, preventing phosphorylation of Chk1 at S345 (18). This modification is required for the Chk1 kinase to acquire its active conformation, detach from chromatin, and carry out its multiple functions ranging from G2/M checkpoint enforcement to DNA repair (reviewed in reference 16). In particular, activation of Chk1 is required for phosphorylation of RAD51 at T309, which is necessary for RAD51 to perform its central role in HRR (17). This was of particular interest because it suggested that MVM-infected cells might have diminished capacity for HRR.

Protein immunoblot analysis with phosphospecific antibodies against Chk1-P-S345 and RAD51-P-T309 demonstrated that by 24 hpi of murine A9 cells with MVM, both the active form of Chk1 and RAD51-P-T309 had fallen to undetectable levels (Fig. 1, compare lane 1 to lane 5, rows A and C, respectively; endogenous levels of unmodified Chk1 and RAD51 under the various experimental conditions, as well as the levels of NS1 expressed during infection, are shown in rows B, D, and E, respectively), allowing for the possibility, as examined further below, that by this stage of infection, HRR might be significantly impaired. In mock-infected cells and by 16 hpi, the small amounts of phosphorylated Chk1 detected in otherwise untreated cells (Fig. 1, lanes 1 and 3, row A) were consistent with previously reported observations showing that basal levels of the enzyme, required to prevent abnormally increased initiation of DNA replication, are constantly present during cellular interphase (31). As previously reported (18), genotoxic treatment (here with NCS) at early times postinfection (16 hpi) resulted in a substantial increase in Chk1-S345 phosphorylation (Fig. 1, compare lanes 3 and 4, row A). However, by times postinfection when activation of Chk1 and RAD51 had been suppressed (Fig. 1, lanes 5 and 7, rows A and C, respectively), NCS had no such stimulatory effect (Fig. 1, lanes 6 and 8, rows A and C).

Fig 1.

RAD51 is not phosphorylated in advanced stages of MVM infection. Murine A9 cells were synchronized by isoleucine starvation, infected with MVM at an MOI of 10, incubated in supplemented DMEM for indicated periods of time, harvested with RIPA buffer, and subjected to immunoblotting using antibodies specific for the indicated proteins. Mock-infected cells were harvested after 16 h of incubation in the supplemented medium. Cells of the indicated sample were treated with 300 ng/mL of NCS for 30 min prior to the harvesting.

Expression of exogenous active Chk1 resulted in efficient phosphorylation of its crucial targets

To confirm whether RAD51-T309 phosphorylation was dependent on activated Chk1, we monitored RAD51 phosphorylation status following overexpression of various forms of Chk1. To increase Chk1 kinase efficiency, we expressed a constitutively active, FLAG-tagged form of the enzyme (Chk1-CA), which had been engineered to carry an L449R single amino acid substitution, which disrupts the inactive conformation of Chk1, leading to its kinase activity regardless of DNA damage status (32). FLAG-tagged wild-type Chk1 (pChk1-WT) and a kinase-deficient enzyme (Chk1-KD, also FLAG-tagged) carrying the D148A substitution (33) were used as controls. As expected, even in the absence of the genotoxic stress inducer NCS, and the resultant low levels of activated, endogenous Chk1-P-S345, overexpression in A9 cells of the constitutively active Chk1, but not the kinase-dead version, resulted in significant increases of both RAD51-P-T309 and phosphorylated Cdc25A, a well-established (16) immediate downstream Chk1 target (Fig. 2, compare lanes 2, 6, and 7, rows F and D, respectively; endogenous levels of unmodified RAD51 and Cdc25A are shown in rows G and E, respectively; expressed protein levels from vectors are shown in row A). While expression of pChk1-WT led to a modest increase of both Cdc25A-P-S76 and RAD51-P-T309, notably, constitutively active Chk1 overexpression produced significantly higher levels of both, despite its lower expression levels (Fig. 2, compare lanes 5 and 6, rows D and F, respectively). The decreased levels of Chk1-CA observed (Fig. 2, row A) were likely due to the highly efficient SCF^Fbx6 and Cul4A^CDT2-dependent degradation of active Chk1, known to be required for efficient resumption of cell cycle after conclusion of a DNA damage response (34–36). Immunoblotting with anti-Chk1 and anti-Chk1-P-S345 antibodies revealed two distinct bands (Fig. 2, lanes 5–7, rows B and C, respectively), the larger of which is consistent with the additional mass conferred by the 3xFLAG tag on the exogenously expressed proteins. Our results, which demonstrate that RAD51-T309 phosphorylation is dependent on activated Chk1 is consistent with a model in which MVM inhibition of Chk1 activity during infection leads to loss of RAD51 activity and, potentially, inhibition of HRR.

Fig 2.

Expression of exogenous active Chk1 resulted in efficient phosphorylation of its crucial targets. Murine A9 cells were synchronized by isoleucine starvation, transfected with vectors expressing indicated Chk1 variants, incubated in supplemented DMEM for 24 h, harvested with RIPA buffer, and subjected to immunoblotting using antibodies specific for the indicated proteins. M, mock transfection; F, p3xFLAG; W, pChk1-WT; A, pChk1-CA; D, pChk1-KD.

Active Chk1 alleviated MVM inactivation of HRR

To directly assess the effect of MVM infection on HRR and to examine a potential role for Chk1 via inhibition of RAD51 in this process, we employed the fluorescent reporter system developed by the Gorbunova lab, which allows a relative quantitative measurement of HRR efficiency in cells under various conditions (24). The reporter vector (pHRR) contains an inactivated GFP coding sequence, which can be restored to wild type by HRR with a complementary GFP fragment encoded in cis on the same vector. The ratio of cells expressing GFP allows a quantifiable estimate of the HRR efficiency (24). For our experiments, pHRR cut with endonuclease HindIII inside the GFP sequence to induce the repair event was transfected into MVM-infected A9 cells (16/24 hpi) expressing either the constitutively active FLAG-tagged Chk1 (Chk1-CA), the FLAG-tagged kinase-dead (Chk1-KD) variant, or the empty vector (p3xFLAG). Following assessment of exogenous Chk1 expression, the percentage of GFP positive cells was determined with FACS 24 h after the transfection with pHRR (see Materials and Methods). First, to verify our experimental approach, we demonstrated that under these conditions, overexpression of Chk1-CA (Fig. 3A, right panel, lane 5, row D) (and, in separate experiments, Chk1-WT [Fig. S2, right panel, lane 6, row D]), but not Chk1-KD (Fig. 3A, right panel, lane 6, row D), led to significant levels of RAD51-T309 phosphorylation despite the MVM-induced absence of endogenous activated Chk1 (Chk1-P-S345) (Fig. 3A, right panel, lanes 4–6, row C. Expressed protein levels from vectors are shown in row A, levels of total Chk1 and endogenous total RAD51 are shown in rows B and E, respectively, and levels of NS1 are shown in row F). (The Western blot analysis shown in Fig. 3A was run on two separate gels to accommodate the necessary samples. However, the left panel, in addition to mock-infected samples [Fig. 3A, left panel, lanes 1–4], contained the same infected sample [Fig. 3A, left panel, lane 5] shown in lane 5 of the right panel [Fig. 3A, right panel, lane 5], to allow direct comparison between the gels).

Fig 3.

Exogenously expressed active Chk1 alleviated MVM inactivation of HRR. (A) Murine A9 cells were synchronized by isoleucine starvation, transfected with vectors expressing indicated Chk1 variants, infected with MVM at an MOI of 10, incubated in supplemented DMEM for indicated periods of time, transfected with HindIII-precut HRR reporter vector (24), further incubated for 24 h, harvested with RIPA buffer, and subjected to immunoblotting using antibodies specific for the indicated proteins (Materials and Methods). Cells of the indicated sample were treated with 300 ng/mL of NCS for 30 min prior to the harvesting. F, p3xFLAG; A, pChk1-CA; D, pChk1-KD. (B) The transfected and infected A9 cells were prepared in triplicate in parallel with samples for the panel A. After indicated periods of time, cells were harvested by trypsinization and were analyzed with a BD Accuri C6 Flow Cytometer. The percentage of GFP positive cells (HRR rate) was calculated with Flowing Software (Turku Bioscience). Error bars represent sample standard deviation. Unpaired Student’s t-tests were run to examine statistical significance of the differences between samples. ns (not significant), P value > 0.05; *, P < 0.05; **, P < 0.01. Dashed lines show comparisons between non-adjacent columns. M, mock transfection; F, p3xFLAG; D, pChk1-KD; A, pChk1-CA; W, pChk1-WT.

Figure 3B presents the level of HRR detected in our reporter assays under the various conditions. Results demonstrate that MVM in the presence of empty vector reduced the levels of HRR more than twofold by 16 hpi. Overexpression of Chk1-CA and to a lesser extent Chk1-WT (likely because, as mentioned previously and discussed more fully below, the Chk1-CA variant was substantially more efficient at phosphorylating its targets than the WT kinase), but not Chk1-KD, significantly alleviated MVM-induced inhibition of HRR at all time points, demonstrating that MVM inhibition of Chk1 activation during infection was important for the reduction of both RAD51 phosphorylation and subsequent inhibition of HRR.

Expression of active Chk1 ameliorates MVM-induced cellular DNA damage

We have previously used comet assays, a technique developed by others to examine the status of genome integrity of individual cells (37), to demonstrate that MVM infection, similar to other parvoviruses (11), caused significant DNA damage to host cell DNA (5). Because overexpression of active Chk1 suppressed the MVM-induced decline in HRR in our reporter assays, and HRR is a major contributor to the repair of DNA DSBs, we deployed comet assays to examine whether exogenously added active Chk1 could alleviate damage to the host genome caused by MVM infection. A9 cells were transfected with the FLAG-tagged expression vectors pChk1-WT, pChk1-CA, pChk1-KD, or an empty FLAG-tagged vector and were infected with MVM, sorted for expression of co-transfected pDsRed, and after the indicated time points, embedded in agarose, subjected to electrophoresis, and stained with a distinguishing fluorescent DNA-binding dye (see Materials and Methods). These experiments were performed under neutral conditions, allowing selective assessment of double-strand rather than single-strand DNA breaks (38). DSBs relax supercoiled cellular DNA, permitting migration of the fraction of the genomic DNA with free ends from comet “head” to its characteristic “tail.” The percent DNA in the tail multiplied by the tail length (defined as the “tail moment”) reflects the magnitude of DNA breaks (39). Because the distance of DNA migration in the comet assays is short (about 50 µM, Fig. 4A), small viral genomes and replicative forms are predicted to move out of the observable comet area during electrophoresis. Additionally, DNA fragments smaller than 50 kbp were reported not to be detectable by this assay (40). However, to confirm the validity of the assay, we applied fluorescent in situ hybridization using MVM probes to select comet assays (30), which confirmed that comets shown here did not include significant amounts of MVM DNA (data not shown).

Fig 4.

Chk1 overexpression ameliorates MVM-induced cellular DNA damage. (A) Murine A9 cells were synchronized by isoleucine starvation, transfected with vectors expressing indicated Chk1 variants, infected with MVM at an MOI of 10, incubated in supplemented DMEM for indicated periods of time, trypsinized, and assayed under neutral conditions with Trevigen’s CometAssay Kit according to the manufacturer’s protocol. The images were obtained with epifluorescence microscopy. F, p3xFLAG; D, pChk1-KD; A, pChk1-CA; W, pChk1-WT. (B) The comets’ images were examined with CaspLab Comet Assay Software to determine tail moments (comet tail length multiplied by percentage of genomic DNA-derived fluorescence in the comet tail) in the software-established procedure defined units (p.d.u.). Mean values for 50 comets were used to plot each data set. Error bars represent sample standard deviation. Unpaired Student’s t-tests were run to examine statistical significance of the differences between samples. ns (not significant), P value > 0.05; *, P < 0.05; **, P < 0.01. Dashed lines show comparisons between non-adjacent columns. F, p3xFLAG; D, pChk1-KD; A, pChk1-CA; W, pChk1-WT.

The comet tails were readily observed with epifluorescence microscopy starting from 24 hpi, indicating that MVM caused significant quantities of DSBs by this point during infection (Fig. 4A). Overexpression of constitutively active Chk1 (pChk1-CA) and to a lesser extent the wild-type kinase (Chk1-WT) (probably, as mentioned above, due to the greater efficacy of the constitutively active enzyme), but not carrier vector (p3xFLAG), or the kinase dead-expressing vector (pChk1-KD), led to significant reduction in DNA DS break damage levels at both 24 and 36 hpi (Fig. 4A; quantification of the tail moments shown in Fig. 4B). These results demonstrated that re-supplying Chk1 activity that had been depleted by infection could substantially improve cellular HRR capacity in those cells.

Host genome damage enhanced MVM DNA replication

An emergence of large quantities of DSBs during MVM infection was expected because autonomous parvovirus infection has been shown to cause elevated levels of reactive oxygen species (10, 11), known to cause various types of DNA damage, including DSBs (reviewed in reference 41). However, the ability of MVM to suppress repair of cellular DSBs suggested that sustaining this type of DNA damage might benefit MVM replication. We have previously shown cellular genome-wide association of MVM DNA with locations of cellular DNA damage (5). Moreover, Hashiguchi et al. demonstrated that DSBs attract DNA polymerase δ as well as its partners PCNA and RFC (9), which have been proposed to be factors required for MVM DNA replication (7), and are found in MVM replication foci (8). Thus, we attempted to directly examine the proximity of MVM DNA replication sites to cellular genome breaks. We utilized DBD-FISH, which allowed in situ detection of DNA breaks in the whole cellular genome (29), together with FISH analysis of MVM DNA in infected cells. Briefly, we labeled the entire mouse genome and specifically the MVM minus strand with DIG-11-dUTP and Biotin-11-dUTP, respectively, as our probes. Only the MVM minus strand was labeled to allow specific detection of the replication-derived viral plus strand (it is important to note that >99% of packaged MVM is a negative-sense, single-stranded DNA genome [42]). Denaturation of infected samples was performed with an alkaline unwinding solution calibrated to denature cellular DNA for hybridization at sites of DNA breaks, exposing these single strands, but not the DNA in intact genomic areas (29). The same conditions were empirically determined as sufficient for MVM DNA denaturation (data not shown). As shown in Fig. 5, fluorescence confocal microscopy demonstrated substantial colocalization of signals derived from hybridization to both cellular and MVM genomes at both initial and advanced stages of the infection (Fig. 5, merge, 16 and 24 hpi). These results indicated that MVM DNA replication mainly takes place in close proximity to cellular genome breaks during both initial and advanced stages of the infection.

Fig 5.

Colocalization of cellular genome breaks and MVM DNA replication foci. Murine A9 cells were synchronized by isoleucine starvation, infected with MVM at an MOI of 10, and incubated in supplemented DMEM for indicated periods of time. The DBD-FISH procedure (see Materials and Methods) was performed with combined probes of total mouse genomic DNA labeled with DIG-11-dUTP by nick translation and MVM DNA minus strand labeled with Biotin-11-dUTP by DNA synthesis elongation of primers corresponding to MVM DNA negative sequence (see Materials and Methods). The distinctive fluorescent signals were achieved by subsequent application of anti-DIG-Rhodamine and Streptavidin-Alexa Fluor 488 conjugates. Nuclei were stained with DAPI. Confocal images were taken with Leica SP8 microscope.

To test the proposition that exogenous damage of the cellular genome facilitates replication of MVM DNA, we treated synchronized, MVM-infected A9 cells with NCS, known to cleave DNA and generate DSBs by abstracting hydrogen atoms from the deoxyribose sugar (43). At 0, 8, 16, and 20 hpi, cells were pulsed with NCS, or carrier (Mock), for 30 min. Cells in each group were subsequently collected at final time points of either 16 or 24 hpi (Fig. 6A, lanes 3–5 and 6–10, respectively) and were subjected to Southern analysis. Cells treated at 0 hpi and taken immediately (0 h) served as an additional control (Fig. 6A, lane 2). Although NCS pulsed at either 0 or 8 hpi did not cause any change in MVM replication by 16 hpi (Fig. 6A, lanes 4, 5; quantifications shown in Fig. 6B), treatments performed at 8 and 16 hpi (Fig. 6A, lanes 8, 9; quantifications shown in Fig. 6B), but not at 20 hpi (Fig. 6A, lane 10; quantification shown in Fig. 6B), resulted in a significant increase in MVM DNA replication levels when assessed at 24 hpi (Fig. 6A, lanes 8, 9; quantifications shown in Fig. 6B). This indicated that during the 24-h infection period, when sufficient time was allowed for MVM to benefit from NCS treatment, introduction of additional cellular DNA breaks resulted in augmented levels of the viral DNA replication.

Fig 6.

Induction of cellular DNA breaks enhances MVM DNA replication. (A) Murine A9 cells were synchronized by isoleucine starvation, infected with MVM at an MOI of 10, and incubated in supplemented DMEM. During the infection, the cells were treated with 300 ng/mL of NCS for 30 min at indicated time points. Total DNA was extracted at 0, 16, and 24 h postinfection. Extracted DNA (500 ng) and 10 ng of NEB 1 kb DNA Ladder (Thermo Scientific) were resolved by 0.8% agarose gel electrophoresis and subjected to Southern blot analysis using mixed 32P-labeled MVM and the Ladder DNA as a probe (upper panel). In parallel, as a loading control, the same amount of the total DNA and 10 ng of λ DNA-HindIII size marker (NEB) was resolved separately and subjected to the same Southern procedure, except the probing was carried out with mixed 32P-labeled murine mitochondrial and the marker DNA (lower panel). The images were obtained with PharosFX Molecular Imager (Bio-Rad). dRF, dimer replicative form; mRF, monomer replicative form; ss, single-stranded MVM genome; C. mtDNA, circular mitochondrial DNA; L. mtDNA, linear mitochondrial DNA; M, mock NCS treatment; K, DNA Ladder/Marker. (B) MVM mRFs identified on the obtained images were quantified with Quantity One software (Bio-Rad). Mean values from three independent experiments were used to plot each data set. The maximum mean value (NCS treatment at 16 h following MVM infection) was set as 100%, and, accordingly, normalized relative percentages of all values were plotted. Error bars represent sample standard deviation. Unpaired Student’s t-tests were run to examine statistical significance of the differences between samples. M, mock NCS treatment; ns (not significant), P value > 0.05; *, P < 0.05; **, P < 0.01. Dashed lines show comparisons between non-adjacent columns.

Exogenous expression of active Chk1 inhibited MVM DNA replication

Our results have indicated that MVM suppression of Chk1-S345 phosphorylation was directly involved in inhibiting host cell DNA repair, which resulted in an increase in DSBs and augmentation of MVM DNA replication. If this was true, we would expect that overexpression of constitutively active Chk1 would inhibit MVM replication. As can be seen in Fig. 7, we found that A9 cells infected with MVM generated significantly reduced levels of replicative viral DNA forms when transfected with constitutively active Chk1 (Chk1-CA) at both 24 and 36 hpi, compared to control samples expressing the inactivated Chk1 (Chk1-KD) or empty vector (p3xFLAG) (Fig. 7A, compare lanes 8 and 11 to lanes 7 and 9 and to 10 and 12, respectively; quantifications shown in Fig. 7B). Our results demonstrate that while MVM induces DNA damage upon infection, the virus’s ability to suppress DNA damage repair by interfering with Chk1 activation promotes its replication.

Fig 7.

Chk1 overexpression inhibits MVM DNA replication. (A) Murine A9 cells were synchronized by isoleucine starvation, transfected with vectors expressing indicated Chk1 variants, infected with MVM at an MOI of 10, and incubated in supplemented DMEM for indicated periods of time. Total DNA was extracted and subjected to Southern analysis as described in the legend for Fig. 6A. The image between the upper and the lower panels shows longer exposure of ssDNA from the upper panel. F, p3xFLAG; A, pChk1-CA; D, pChk1-KD; dRF, dimer replicative form; mRF, monomer replicative form; ss, single-stranded MVM genome; C. mtDNA, circular mitochondrial DNA; L. mtDNA, linear mitochondrial DNA; K, DNA Ladder/Marker. (B) MVM mRFs identified on the obtained images were quantified with Quantity One software (Bio-Rad). Mean values from three independent experiments were used to plot each data set. The maximum mean value (F [p3xFLAG] transfection, 36 hpi) was set as 100%, and, accordingly, normalized relative percentages of all values were plotted. Error bars represent sample standard deviation. Unpaired Student’s t-tests were run to examine statistical significance of the differences between samples. ns (not significant), P value > 0.05; **, P < 0.01. Dashed lines show comparisons between non-adjacent columns. F, p3xFLAG; D, pChk1-KD; A, pChk1-CA.

DISCUSSION

MVM infection in various cultured rodent and transformed human cell lines has previously been shown to cause cellular DNA breaks and a robust DNA damage response (DDR) (5, 13). Both these phenomena were suggested to enhance virus replication (5, 13). Early during infection, MVM DNA was found to be preferentially located in proximity to cellular DNA damage sites (5), and MVM NS1 was shown to actively traffic the viral genomes to these locations (6). At the same time, Chk2 and multiple additional DDR proteins belonging to the ATM arm of the DDR pathway exhibited DDR-associated phosphorylation modifications (13). Moreover, ATM activation was shown to play an important role in infection as its inhibition significantly decreased MVM DNA replication, although the ATM pathway is complex, and the exact mechanism remains to be determined (13). ATM is activated in response to DNA DSBs, which our comet assay showed to be abundant in MVM infected murine A9 cells by 24 hpi, presumably as a result of ROS generation characteristic for autonomous parvovirus infection (10, 11). Because autonomous parvovirus DNA replication takes place during S-phase, the infection-generated DSBs could hypothetically be repaired by either NHEJ or HRR. However, HRR has been reported to be a dominant DNA repair mechanism at its peak capacity during S-phase (14), and our data support its significant involvement.

DSB repair by HRR involves resection of DNA, leading to generation of ssDNA strands in the region of the DSB, which initially attracts replication protein A (RPA), followed by ATR with its auxiliary factors, as the major ATR’s downstream target kinase Chk1 tends to associate with chromatin in its inactive state (reviewed in references 15, 16). In addition, RPA-coated parvovirus ssDNA genomes and viral replication intermediates may also attract ATR with its co-factors (44), as well as Chk1, and consequently be a potent trigger of ATR-Chk1 pathway activation. Surprisingly, during MVM infection, despite the induction of DNA damage, the presence of both cellular and viral ssDNA, and the activation of ATM, both ATR and Chk1 were found to be inactivated. The objective of this study was to determine the significance of ATR pathway inactivation during MVM infection.

Chk1 is a key protein kinase involved in controlling the cellular DDR and, thus, preserving genome integrity. Its most well-defined role is activation of several factors such as Cdc25 phosphatases and the Wee1 kinase, following DNA damage events and its activation by ATR. The resultant activation of the G2/M checkpoint allows cells to repair the genomic impairments prior to M-phase (reviewed in reference 16). MVM actively blocks infected cells from entering mitosis, prompting a prolonged pre-mitotic cellular environment favorable for its replication (4). Thus, it seemed possible that rather than functioning to promote M-phase transition in the infected cells, inhibition by MVM of the ATR/Chk1 pathway served another purpose during infection. We considered the possibility that additional Chk1 functions, unrelated to the G2/M cell cycle block, might interfere with viral replication, which would be suppressed by MVM inactivation of the ATR-Chk1 pathway.

Sørensen et al. reported that a key HRR participant, RAD51, depends for its activity on Chk1 phosphorylation of amino acid Thr 309 (17). Additional reports corroborated that Chk1 and its interaction with RAD51 are required to maintain normal cellular HRR levels (45, 46). Our results demonstrated significant loss of RAD51 Thr 309 phosphorylation by 24 h post-MVM infection, which strongly correlated with a substantial decrease in HRR activity. A decrease in HRR activity may be particularly significant in the context of parvoviral infection, because during the MVM-induced pre-mitotic block, CDK1 and CDK2 typically activate and stimulate DSB end resection to ensure functional HRR (reviewed in reference 47). Disabling of HRR would be predicted to inhibit repair of DSBs and consequently increase levels of DSBs in MVM-infected cells. DSBs are known to attract factors such as DNA polymerase δ, PCNA, and replication factor C (9), required for MVM DNA replication (7), and, as mentioned, MVM genomes have been shown to localize and replicate in close proximity to DNA breaks present in infected cells (Fig. 5) (5). Thus, MVM-induced loss of active Chk1 and the resulting decrease in activated RAD51 levels led to inhibition of cellular HRR, which directly promoted virus replication by increasing the number of sites of DNA damage, which provided additional platforms for virus replication.

It should be pointed out that MVM has been shown to suppress host cellular genome replication (48), and so it remained possible that HRR levels could be limited upon infection solely due to paucity of newly synthesized sister chromatid DNA, considered to be a major template source for HRR (reviewed in reference 15). However, under the experimental conditions employed here, i.e., para-synchronization by isoleucine starvation prior to infection and release into complete medium, it has been reported that about 60% of the host cellular genome of murine A9 cells has been replicated by the peak amplification stage of infection (49), suggesting that there were adequate available DNA templates to sustain HRR in our experiments. Additionally, homologous chromosomes, like sister chromatids, can provide a general repair template for all chromosomal sequences, except an XY pair. Such allelic recombination repair was reported to reach 30% of all HRR events in mammalian cells upon the induction of DSBs (50), which is under the control of SWS1–SWSAP1-SPIDR complex (51).

Our DBD-FISH assay (Fig. 5) demonstrated that viral DNA replication foci localized in close proximity to cellular DNA breaks, supporting the possibility that these locations, replete with factors recruited for DNA repair, could provide a rich environment for MVM DNA replication. Consistent with this supposition, treatment of host cells with a DSB-inducing reagent substantially enhanced MVM DNA replication (Fig. 6A and B), which further demonstrated that the cellular DNA damage environment is favorable for MVM replication, which would be facilitated by the suppression of cellular DNA repair by MVM. As mentioned, in work previously reported, MVM genomes were found to localize to existing areas of cellular genome damage early in infection, and as the infection-amplified cellular DNA breaks accumulated, additional replicating MVM DNA localized to these areas (5). Thus, the ability of MVM to suppress the repair of cellular DNA breaks, leading to increased levels of DSBs, would be expected to facilitate increased replication of MVM. Overexpression of constitutively active Chk1 during MVM infection restored RAD51 Thr 309 phosphorylation and suppressed infection-induced reduction of HRR efficiency. This led to a consequent decrease in DSBs and reduced viral DNA replication levels. Together, this indicates that during infection, MVM inhibition of Chk1 activation enhances viral replication, at least in part, by inhibiting cellular HRR.

Interestingly, the enhanced DNA amplification and gene expression of autonomous parvoviruses seen in transformed cells (52) seemingly correlate well with the propensity of cancer cells to suffer an elevated extent of DNA damage caused by their high DNA replication rate, potent ROS levels, and other factors (reviewed in reference 53). Therefore, both the higher levels of DSBs in cancer cells and the viral suppression of cellular DNA repair may play important roles in parvovirus oncosuppression.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jvi.00889-24.

Supplemental figures. jvi.00889-24-s0001.pdf.

Fig. S1 and S2.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.