|Home | About | Journals | Submit | Contact Us | Français|
Human parvovirus B19 (B19V) infection is highly restricted to human erythroid progenitor cells, in which it induces a DNA damage response (DDR). The DDR signaling is mainly mediated by the ATR (ataxia telangiectasia-mutated and Rad3-related) pathway, which promotes replication of the viral genome; however, the exact mechanisms employed by B19V to take advantage of the DDR for virus replication remain unclear. In this study, we focused on the initiators of the DDR and the role of the DDR in cell cycle arrest during B19V infection. We examined the role of individual viral proteins, which were delivered by lentiviruses, in triggering a DDR in ex vivo-expanded primary human erythroid progenitor cells and the role of DNA replication of the B19V double-stranded DNA (dsDNA) genome in a human megakaryoblastoid cell line, UT7/Epo-S1 (S1). All the cells were cultured under hypoxic conditions. The results showed that none of the viral proteins induced phosphorylation of H2AX or replication protein A32 (RPA32), both hallmarks of a DDR. However, replication of the B19V dsDNA genome was capable of inducing the DDR. Moreover, the DDR per se did not arrest the cell cycle at the G2/M phase in cells with replicating B19V dsDNA genomes. Instead, the B19V nonstructural 1 (NS1) protein was the key factor in disrupting the cell cycle via a putative transactivation domain operating through a p53-independent pathway. Taken together, the results suggest that the replication of the B19V genome is largely responsible for triggering a DDR, which does not perturb cell cycle progression at G2/M significantly, during B19V infection.
Human parvovirus B19 (B19V) is a small nonenveloped virus with a single-stranded DNA (ssDNA) genome of 5.6 kb (18) and belongs to the genus Erythrovirus in the family Parvoviridae (68). B19V infection in healthy adults is self-limiting, but in immunocompromised individuals, those with inherited hemolytic anemia, and pregnant women, B19V infection can cause aplastic crisis and hydrops fetalis, which can be fatal (74). B19V infection is restricted to human erythroid progenitor cells (EPCs) (2, 44, 51, 63). During B19V infection, nine major mRNA transcripts are generated by alternative processing of a single precursor mRNA (50) and encode one large nonstructural 1 (NS1) protein, two small nonstructural proteins (11-kDa and 7.5-kDa proteins), and two capsid proteins (VP1 and VP2) (37, 64, 78). The NS1 protein is essential for B19V DNA replication (78) and is a transactivator for viral gene expression (22, 25, 55), as well as for the expression of several cellular genes (24, 41, 48). NS1 is also thought to induce cell cycle arrest (45, 70) and apoptosis (42) of infected EPCs. The 11-kDa protein plays a role in the viral DNA replication (78) and apoptosis induced during infection (13); however, the function of the 7.5-kDa protein remains unknown. Apart from serving as a minor structural protein (30), VP1 also contains a unique region that is essential for the intracellular trafficking of the virus into the nucleus (75). VP2 is the major structural protein involved in virion formation (30, 31).
Ex vivo-expanded EPCs are highly permissive for B19V infection (62, 72) and support an efficient productive B19V infection when EPCs are cultured under hypoxic conditions (10, 53). A few erythropoietin-dependent cell lines, e.g., the megakaryoblastoid cell line UT7/Epo-S1 (S1) (47), support B19V infection with limited efficiency; however, replication of the B19V genome is increased approximately 100-fold when S1 cells are cultured under hypoxic conditions (10). After B19V infection, host cells show remarkable responses to viral infection. B19V induces a DNA damage response (DDR) (39), cell cycle arrest (45–47, 70), and cell death (13, 42, 46, 73), among which the B19V-induced DDR is hijacked by the virus to promote viral DNA replication (39). Three phosphatidylinositol 3-kinase-like kinases (PI3Ks), ataxia telangiectasia-mutated (ATM), ataxia telangiectasia-mutated and Rad3-related (ATR), and the DNA-dependent protein kinase catalytic subunit (DNA-PKcs) are activated; however, the ATR is the major determinant for the DDR induced during B19V infection, and both the ATR and DNA-PKcs are essential for promoting viral genome replication (39). Which viral components are responsible for triggering the DDR and whether the induced DDR is detrimental to host cells during B19V infection are currently unknown.
DDR involves biochemical pathways that arrest cell cycle progression in response to DNA damage or replication stress. Activation of ATR or ATM phosphorylates downstream factors to induce cell cycle arrest and facilitate DNA repair (7, 35). As B19V infection activates the ATR and ATM signaling pathways, they presumably contribute to G2/M arrest during B19V infection. However, pharmacological inhibitors of DDR do not rescue G2/M arrest in infected cells, although viral DNA replication is significantly decreased (39). In contrast, the B19V NS1 plays a key role in inducing G2/M arrest during B19V infection in EPCs through the interaction between its nuclear localization signal (NLS) domain and the cellular transcription factors E2F4/E2F5 (70). It is arguable that DDR signaling and the interaction between NS1 and E2F4/E2F5 are redundant for arresting B19V-infected cells at the G2/M phase; nevertheless, the role of the DDR in cell cycle perturbation during B19V infection needs to be confirmed.
We recently demonstrated that hypoxic conditions promote efficient B19V infection in both EPCs and S1 cells (10). B19V infection of EPCs under hypoxic conditions in vitro mimics B19V infection of EPCs in the human bone marrow and fetal liver, where such hypoxic conditions exist (16, 52, 58, 67). To understand the cause of the DDR and to differentiate the role of the DDR from that of NS1 in inducing G2/M arrest, in this study we cultured both S1 cells and EPCs under hypoxic conditions; EPCs were transduced with lentiviruses expressing individual viral proteins, and S1 cells were transfected with the double-stranded DNA (dsDNA) form of the B19V ssDNA genome.
Primary human CD34+ cells were isolated from granulocyte colony-stimulating factor (G-CSF)-mobilized peripheral blood stem cells from healthy donors according to a protocol (02-H-0160) approved by the National Heart, Lung, and Blood Institute institutional review board. EPCs were expanded ex vivo from the primary human CD34+ cells in Wong medium as previously described (9, 72). Briefly, cells frozen on day 4 of culture were thawed and cultured under normoxic conditions until day 7. The cells were then transferred to hypoxic conditions (1% O2 and 5% CO2) for 48 h before infection or transduction (10). The S1 cells were cultured as described previously (26) and kept under hypoxic conditions for 48 h before electroporation, B19V infection, or lentiviral transduction.
The B19V plasma sample (no. P158, ~1 × 1011 genome copies [gc]/ml) was supplied by the ViraCor-IBT Laboratories (Lee's Summit, MO). B19V infection was carried out at a multiplicity of infection (MOI) of 1,000 gc/cell by following the methods previously described (10).
The electroporation of S1 cells was performed as previously described (26). The B19V dsDNA genome (M20) and its derivative mutants were recovered from SalI-digested pIB19-M20 and its derivative mutants.
The DNA coding sequences for the B19V NS1, 11-kDa, 7.5-kDa, VP1, and VP2 proteins were optimized at GenScript USA Inc. (Piscataway, NJ) to enhance protein expression efficiency (79). The pLenti-CMV-IRES-GFP-WPRE vector (36) was used for inserting C-terminally Flag-tagged optimized (opt) NS1-, 11-kDa protein-, 7.5-kDa protein-, VP1-, and VP2-coding sequences at BamHI and BsrGI sites, resulting in the following constructs: pLenti-optNS1, pLenti-opt11-kDa, pLenti-opt7.5-kDa, pLenti-optVP1, and pLenti-optVP2.
The B19V P6 promoter-based NS1-expressing lentiviral vector, pLenti-½ITR-P6-NS1, was constructed as follows: a DNA fragment from nucleotides 187 to 2628 of the B19V genome, which spans the left half of the inverted terminal repeat (ITR), namely, ½ITR, the P6 promoter, and the NS1-encoding sequence, with a Flag epitope at the C terminus, was inserted into ClaI/SalI-digested pLenti-CMV-GFP-WPRE, which was made by removing the puromycin gene expression cassette from pLenti-GFP-Puro (Addgene Inc., Cambridge, MA). The amino acid sequence of B19V NS1 (GenBank accession no. AAQ91878.1) was aligned with that of adeno-associated virus 2 (AAV2) Rep78 (GenBank accession no. AAC03775.1), which locates the conserved endonuclease domain (69) and helicase Walker A box site (43, 69) in the B19V NS1 (see Fig. 2C). Putative transactivation domains (TADs) at the C terminus of the B19V NS1 were predicted by the 9aaTAD program (54) (see Fig. 3A). Then, mutations of the putative endonuclease and helicase-A domains and the predicted TAD2 and TAD3 were introduced into the NS1-encoding region of pLenti-½ITR-P6-NS1, constructing pLenti-½ITR-P6-NS1(endo-), pLenti-½ITR-P6-NS1(heli-), pLenti-½ITR-P6-NS1(mTAD2), and pLenti-½ITR-P6-NS1(mTAD3), respectively. pLenti-½ITR-P6-RFP was made by replacing the NS1-encoding sequence with a red fluorescent protein (RFP) open reading frame (ORF) (13).
The B19V infectious clone pIB19-M20 (80), NS1 knockout mutant [pIB19-M20(NS1−)], and VP2 knockout mutant [pIB19-M20(VP−)] (78) were gifts from Kevin Brown and Ning Zhi at the NIH. The M20 DNA is a dsDNA form of the B19V ssDNA genome, which presumably is an intermediate replicative form (RF) of the B19V genome during virus replication (27). Mutations of the putative endonuclease and helicase-A domains and the predicted TAD1, TAD2, and TAD3 of NS1 (see Fig. 2C and and3A)3A) were introduced into pIB19-M20, creating pIB19-M20(endo−), pIB19-M20(heli−), and pIB19-M20(mTAD1, mTAD2, and mTAD3).
The validated coding sequence for p53 short hairpin RNA (shRNA) was obtained from Sigma (St. Louis, MO) as follows: 5′-CCG GTC GGC GCA CAG AGG AAG AGA ATC TCG AGA TTC TCT TCC TCT GTG CGC CGT TTT TC-3′ (clone identifier, NM_000546.x-1095s1c1). pLKO-shRNA-p53 was made by inserting a DNA fragment of this sequence into the AgeI/EcoRI-digested pLKO-GFP-scramble-shRNA vector (Addgene Inc.) as described previously (9).
All the nucleotide numbers of the B19V genome refer to GenBank accession no. AY386330.
Rat anti-B19V NS1 (anti-NS1) and anti-11-kDa protein polyclonal antibodies were previously reported (13, 39). Mouse anti-7.5-kDa protein polyclonal antibody was made by immunizing mice with purified glutathione S-transferase-fused 7.5-kDa protein by following the method previously described (36). The following antibodies were obtained commercially: anti-VP1/VP2 antibody (MAB8292, Millipore, Billerica, MA), anti-phosphorylated H2AX (γ-H2AX) antibodies (05-636, Millipore; 2212-1, Epitomics, Burlingame, CA), anti-phosphorylated replication protein A32 (p-RPA32) (3237-1, Epitomics), anti-p53 antibody (GTX102965, GeneTex, Irvine, CA), anti-Flag antibody (200472, Agilent Technologies, Santa Clara, CA), and anti-β-actin antibody (A5441, Sigma-Aldrich, St. Louis, MO).
Lentiviruses were generated as previously described (36). EPCs were transduced at an MOI of approximately 10/cell as previously described (9), whereas S1 cells were transduced at an MOI of 1/cell. S1 cells were preincubated with p53-shRNA lentivirus by slow rotation at 4°C for 1.5 h before culture at 37°C.
Immunofluorescence (IF) staining was performed as previously described (39). Texas Red-conjugated donkey anti-rat antibody (catalog no. 712-076-153) was used for NS1 or 11-kDa protein detection, and fluorescein isothiocyanate (FITC)-conjugated donkey anti-mouse antibody (catalog no. 715-095-151) was used for γ-H2AX (anti-γ-H2AX, Millipore) or Alexa Fluor 488-conjugated donkey anti-rabbit antibody (catalog no. 711-546-152) was used for p-RPA32. Rhodamine red-X-conjugated donkey anti-mouse antibody (catalog no. 715-295-150) was used for 7.5-kDa protein or VP1/VP2 detection, and Alexa Fluor 488-conjugated donkey anti-rabbit antibody (catalog no. 711-546-152) was used for γ-H2AX (anti-γ-H2AX, Epitomics) or p-RPA32. All the secondary antibodies were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA).
Images were taken under a confocal microscope (Eclipse C1 Plus, Nikon) with Nikon EZ-C1 software.
Cells were stained and analyzed by flow cytometry as previously described (11, 36). For cell cycle analysis of the B19V-infected S1 cells and EPCs and of S1 cells transfected with M20 and its mutants, cells were fixed, permeabilized, and costained 48 h postinfection (p.i.) or posttransfection (p.tx.) in succession with the rat anti-NS1 antibody, a Cy5-conjugated goat anti-rat antibody (catalog no. 112-176-143) for detection of NS1, and 4′,6-diamidino-2-phenylindole (DAPI). For cell cycle analysis of the lentivirus-transduced S1 cells and EPCs, cells were stained 48 h posttransduction (p.td.) with the mouse anti-Flag antibody, followed by a Dylight 649-conjugated goat anti-mouse antibody (catalog no. 115-496-003) and DAPI.
To determine the level of a DDR, S1 cells were fixed, permeabilized, and costained 48 h p.i./p.tx. with the mouse anti-γ-H2AX (Millipore) and rat anti-NS1 antibodies, followed by FITC-conjugated donkey anti-mouse (catalog no. 715-095-151) and Alexa Fluor 647-conjugated donkey anti-rat (catalog no. 712-606-153) secondary antibodies for the detection of γ-H2AX and NS1, respectively. All of the secondary antibodies were purchased from Jackson ImmunoResearch Laboratories.
All of the samples were analyzed on a three-laser flow cytometer (LSR II, BD Biosciences) at the Flow Cytometry Core of the University of Kansas Medical Center. All flow cytometry data were analyzed using FACSDiva software (BD Biosciences).
Low-molecular-weight (Hirt) DNA was extracted from treated cells, followed by DpnI digestion and Southern blotting as previously described (26) with M20 DNA as a probe template. Images were developed and analyzed with a Typhoon 9700 phosphorimager (GE Healthcare).
We previously reported that B19V infection induces a DDR in infected EPCs cultured under normoxic conditions (39). Since hypoxic conditions promote the infection of EPCs by B19V (10), we cultured EPCs under hypoxic conditions and infected them with B19V. B19V infection of EPCs cultured under hypoxic conditions also induced a DDR, as evidenced by the phosphorylation of both H2AX and RPA32 (which are hallmarks of the DDR [5, 40, 65]), at levels approximately three times higher than those in infected cells cultured under normoxic conditions (Fig. 1A). Notably, p53 was phosphorylated in B19V-infected EPCs under normoxic conditions and phosphorylated further under hypoxic conditions.
Since B19V expresses the nonstructural proteins (NS1, 7.5-kDa protein, and 11-kDa protein) in addition to the structural proteins (VP1 and VP2) during infection (37, 64, 78), we examined whether the expression of the individual viral proteins induced a DDR in EPCs. We expressed the nonstructural proteins and structural proteins in EPCs via lentiviral transduction. As a positive control for the DDR, EPCs were treated with hydroxyurea (HU). At 48 h p.td., EPCs were analyzed for the expression of phosphorylated H2AX and RPA32 by IF staining. B19V-infected EPCs showed pan-distribution of γ-H2AX (Fig. 1B) and colocalized foci of NS1 with p-RPA32 (Fig. 1C) in the nuclei, which was consistent with that in infected EPCs cultured under normoxic conditions (39). However, we observed that NS1 was expressed in the cytoplasm to a greater extent than in the nucleus, likely due to the high infectivity of B19V in EPCs cultured under hypoxic conditions (10). In contrast to the formation of nuclear foci during B19V infection, NS1-transduced EPCs showed neither phosphorylation of H2AX nor colocalization of NS1 and p-RPA32 (Fig. 1B and andC).C). A proportion of the NS1 in transduced EPCs appeared to be in the cytoplasm, mimicking the NS1 expression pattern observed during viral infection. The difference in γ-H2AX and p-RPA32 expression between B19V-infected and NS1-transduced EPCs suggests that B19V NS1 alone is not sufficient to induce the DDR observed during B19V infection. Similarly, when the 7.5-kDa, 11-kDa, VP1, and VP2 proteins were expressed in EPCs by their respective lentiviruses, neither H2AX nor RPA32 was phosphorylated in transduced EPCs (Fig. 1B and andC),C), whereas HU treatment of control cells induced the phosphorylation of both H2AX and RPA32.
Taken together, these results confirm that B19V viral proteins, including NS1, the 7.5-kDa protein, the 11-kDa protein, VP1, and VP2, alone do not induce the DDR in EPCs that is observed during B19V infection.
Hypoxic conditions support the efficient replication of the B19V dsDNA genome (M20 DNA) in S1 cells at a level approximately 100 times greater than that observed under normoxic conditions (10). Since B19V replication in EPCs requires a DDR (39), we hypothesized that the replication of B19V M20 DNA in S1 cells induces a DDR as well. To this end, S1 cells were transfected with M20 and a mutant [M20(VP−)] in which the VP2 ORF was knocked out and neither VP1 nor VP2 was expressed (data not shown). Cells were infected with B19V as a control. We found that both wild-type M20-transfected and B19V-infected S1 cells generated DpnI digestion-resistant bands of ssDNA, monomer RF (mRF) DNA, and double RF (dRF) DNA on a Southern blot (Fig. 2A). This indicated that B19V M20 DNA replicated in the transfected cells. The negative control, M20(NS1−)-transfected cells, did not generate any DpnI digestion-resistant DNA bands, confirming that NS1 is essential for B19V DNA replication. Transfection of mutant M20(VP−) yielded clear mRF and dRF DNA bands but not ssDNA bands (Fig. 2A, lane 4). Production of ssDNA was significantly inhibited in the absence of capsid proteins during B19V DNA replication. When the putative endonuclease domain or helicase-A site was mutated, the M20(endo−) and M20(heli−) mutants did not replicate in transfected cells, as evidenced by the lack of DpnI digestion-resistant DNA bands (Fig. 2B, lanes 3 and 4).
The transfection of M20 induced phosphorylation of both H2AX and RPA32 and induced the formation of NS1 and p-RPA32 foci in the nuclei of transfected cells, similar to that observed in B19V-infected S1 cells (Fig. 2D and andE).E). Moreover, H2AX and RPA32 were phosphorylated and colocalized with NS1 in the nuclei of M20(VP−)-transfected S1 cells. Notably, M20(endo−)- and M20(heli−)-transfected NS1-positive (NS1+) cells did not express γ-H2AX (Fig. 2D) or p-RPA32 (Fig. 2E). Next, we carefully quantified the DDR induced in S1 cells infected with B19V or transfected with M20 and its mutants by flow cytometry. Replicative B19V M20 DNAs [M20 and M20(VP−)] and B19V infection induced significant expression of γ-H2AX in NS1-expressing cells (Fig. 2F and andG),G), which was significantly higher than those induced by nonreplicative M20 DNAs [M20(endo−) and M20(heli−)].
Taken together, these results show that replication of B19V M20 DNA per se, without ssDNA production, is sufficient to induce a DDR at a level similar to that induced during B19V infection.
B19V infection of EPCs triggers a DDR cascade with the activation of all three PI 3-kinases (ATM, ATR, and DNA-PKcs) (39). It also arrests infected cells at the G2/M phase (70). ATM- and ATR-mediated signaling is thought to be the major pathway leading to cell cycle arrest at the G2/M checkpoint (35, 61). However, in EPCs the expression of B19V NS1 alone can induce G2/M arrest (70). Hence, we examined the role of the B19V DNA replication-induced DDR in G2/M arrest.
We first tried to identify an M20 mutant that replicates but does not induce significant G2/M arrest in transfected cells. The DNA binding (endonuclease) and helicase domains of NS1 were critical for B19V DNA replication (Fig. 2B), and mutations in the NLS site abolished DNA replication (data not shown). We focused on the C terminus of NS1, particularly the predicted TADs, since the TAD of minute virus of mice (MVM) NS1 is dispensable for MVM DNA replication (33). Three TADs were predicted by analyzing the C terminus of NS1 using a TAD prediction program (54) (Fig. 3A). We mutated each of the predicted TADs in M20 and assessed their replication capability. Only the M20(mTAD2), which harbors mutations in TAD2, replicated in transfected S1 cells at a level comparable with that in the wild type (Fig. 3B, lane 3). Furthermore, we examined the ability of the M20-based TAD mutants to trigger a DDR in transfected cells. The results showed that in only the M20(mTAD2)-transfected (NS1+) cells were both H2AX and RPA32 clearly phosphorylated but not in cells transfected with M20(mTAD1) and M20(mTAD3) (Fig. 3C), supporting the notion that B19V DNA replication is critical for inducing a DDR.
We next tested the ability of wild-type M20 and its mutants to arrest the cell cycle at G2/M by transfecting them into S1 cells. Notably, the M20 mutants that did not replicate induced clear G2/M arrest at levels similar to or higher than those observed in the wild type, with the percentage of cells in G2/M phase ranging from 63.1% to 88.6% in NS1+ cells (Fig. 3D and andE).E). More importantly, the replicative TAD mutant, M20(mTAD2), significantly decreased the number of cells at the G2/M phase compared with that of the wild-type M20 control (32.8% vs. 57.7%, respectively) (Fig. 3E).
Collectively, these results show that replication of the B19V dsDNA genome is not required to induce G2/M arrest in transfected S1 cells. Since replication of the B19V dsDNA genome is essential for inducing a DDR, we conclude that the DDR is dispensable for G2/M arrest of the cells in which B19V DNA replicates. The results also suggest that B19V NS1 alone is sufficient to arrest the cell cycle at G2/M during infection.
To examine the sole role of NS1 in inducing G2/M arrest, we expressed only wild-type NS1 and the NS1 mutants, which originated from the M20 mutants, by lentiviral transduction in S1 cells and EPCs. Transduction of Lenti-½ITR-P6-NS1, in which expression of wild-type NS1 is driven by the native B19V P6 promoter (55), recapitulated G2/M arrest to the level induced by B19V infection in S1 cells, in which over 90% of NS1+ cells were arrested at the G2/M phase. This was also observed after the transduction of lentiviruses expressing the NS1 mutants except for NS1(mTAD2) (Fig. 4A). The expression of NS1(mTAD2) increased the number of NS1+ cells at G2/M only to 36.6%, a level similar to that induced by the expression of the RFP control (42.4%) (Fig. 4C). In EPCs, in agreement with the results observed in S1 cells, both B19V infection and transduction of lentiviruses expressing wild-type NS1 and the NS1 mutants, NS1(endo−), NS1(heli−), and NS1(mTAD3), induced over 90% of NS1+ cells at the G2/M phase (Fig. 4B). Notably, the expression of NS1(mTAD2) in EPCs resulted in 42.2% of NS1+ cells to arrest at G2/M, a level close to that induced by expression of the RFP control (35.0%) (Fig. 4D). We noticed that the expression of RFP in the context of Lenti-½ITR-P6-RFP induced a higher number of cells at G2/M (approximately 40%) than G2/M-phased cells (approximately 20%) in the mock group in both S1 cells and EPCs. It has been reported that a consensus Toll-like receptor (TLR) ligand sequence (5′-GTTTTGT-3′) exits in the B19V P6 promoter and inhibits proliferation of EPCs by arresting the cell cycle at the S and G2/M phases (28). We also observed that a lentiviral vector bearing a B19V sequence of nucleotides 187 to 615 (contains ½ITR and the core P6 promoter) arrested the cell cycle at G2/M (data not shown).
Taken together, these results confirm that the expression of B19V NS1 alone is sufficient to induce nearly complete G2/M arrest in the context of the B19V P6 promoter and that the predicted TAD2 of NS1 is critical for G2/M arrest induced by NS1 in both S1 cells and EPCs.
ATM and ATR activation during a DDR phosphorylates p53 at Ser15 (3, 60). The cyclin-dependent kinase Cdk1, which controls the G2/M transition, is inhibited by the product of the p53-targeting gene, p21 (6). Since B19V NS1 transactivates the expression of p21 (45) and p53 is phosphorylated at Ser15 during B19V infection of EPCs (70) (Fig. 1A), we decided to examine the role of p53 in B19V-induced G2/M arrest.
We first attempted to knock out p53 in EPCs using p53-specific shRNAs, which proved to be very difficult (data not shown). Fortunately, nearly perfect p53 knockout was obtained in S1 cells transduced with a p53-shRNA-expressing lentivirus (Fig. 5A). We found that the percentage of p53-deficient NS1+ cells (either infected with B19V or transfected with M20) at the G2/M phase did not decrease compared with that in scramble (scr)-shRNA-treated cells (Fig. 5B and andC),C), suggesting that the p53-mediated pathway is not used to induce G2/M arrest during infection or transfection. We also observed that the transfection of M20 caused fewer cells to arrest at G2/M than B19V infection (56.8% vs. 90.6%) (Fig. 5B); however, knockout of p53 did not ameliorate G2/M arrest significantly under either condition (Fig. 5C).
In this study, we report that the replication of the B19V dsDNA genome, but not the individual viral proteins, induces a DDR. We identified an M20 mutant, M20(mTAD2), containing a mutated transactivation domain within NS1, which replicated well in S1 cells but did not arrest the cell cycle at the G2/M phase. In agreement with this result, the NS1 harboring the TAD2 mutation did not induce significant G2/M arrest in either EPCs or S1 cells by lentiviral transduction. Finally, we showed that the p53-mediated pathway is not involved in G2/M arrest during B19V infection. Thus, our results provide evidence that the DDR induced during B19V infection does not contribute to G2/M arrest and confirm that NS1 alone is sufficient to induce the G2/M arrest in the context of the B19V P6 promoter.
A DDR is a response by the cellular surveillance network that senses and repairs damaged DNA and is required to maintain genome integrity (14). During viral infection, a DDR plays a role in the intrinsic antiviral mechanism to eliminate the invasion of the viral genome; however, in some cases, the virus exploits the DDR signaling mechanism to promote its own replication (71). One of the major consequences of a DDR is cell cycle arrest, which ensures that there is time for DNA repair after DNA damage (57). B19V infection of both EPCs and S1 cells typically arrests the cell cycle at G2/M (39, 47, 70). Preventing replication of the B19V dsDNA genome by mutating the endonuclease domain or helicase-A motif within NS1, which prohibits the DDR, did not reduce the percentage of cells arrested at G2/M; however, mutating TAD2 could result in a replication-competent M20 mutant but with reduced G2/M arrest. NS1 alone is sufficient to cause G2/M arrest, which requires a predicted TAD2. Therefore, B19V utilizes the mechanism of a DDR only to promote viral DNA replication (39), and the NS1-induced G2/M arrest is dispensable for viral DNA replication. The role of the NS1-induced G2/M arrest during B19V infection warrants further investigation. The fact that the p53 pathway is not used by B19V NS1 to arrest the cell cycle indicates that the TAD2 of NS1 does not transactivate p53-regulating genes. NS1 expression in EPCs upregulates 44 genes and downregulates 28 genes that are involved in cell cycle regulation (70). Therefore, we hypothesize that TAD2 of NS1 may play an important role in regulating one or several of these genes. The NLS domain of NS1 is thought to be critical for inducing G2/M arrest, which is mediated by the interaction between NS1 and E2F4/E2F5 (70); however, mutation of the NLS, which prevents NS1 transport into the nucleus (70), may also prevent the function of the TAD. Clearly, our results highlight a novel mechanism underlying G2/M arrest induced by NS1 during B19V infection, which is independent of the DDR and p53 activation.
Viral proteins often play an important role in triggering a DDR during infection (8, 34, 71), particularly in the large T antigen of polyomaviruses (21, 49, 77) and the Vpr of HIV (76). In parvoviruses, the replication of the AAV2 genome during coinfection with adenovirus elicits an ATM-, ATR-, and DNA-PKcs-mediated DDR (17, 59). We reported that DNA replication of the minute virus of canines (MVC) genome is responsible for an MVC-induced DDR rather than the viral proteins (38). The MVM also induces a DDR during infection (1, 56), but the small nonstructural protein NS2 is dispensable (56). The AAV2 Rep78 induces a DDR by introducing nicks in the cellular chromatin (4), but it is only a minor contributor to the DDR induced during AAV2 infection (59). The present study shows that B19V NS1, the small nonstructural proteins (7.5-kDa and 11-kDa proteins), and the capsid proteins failed to induce an obvious DDR in EPCs when they were expressed individually. However, we have not ruled out the possibility of the combined expression of B19V proteins, which is the case during B19V infection, and transfection of the replication-competent M20 DNAs inducing a DDR. Knockout of the 11-kDa protein in M20 DNA showed a decreased level of DNA replication in S1 cells (data not shown). We speculate that a combined expression of B19V proteins supports B19V DNA replication to the highest extent, which induces a maximum DDR. This actually is true for the NS1 protein. NS1 is essential for B19V DNA replication (Fig. 2A), but NS1 alone could not induce a DDR (Fig. 1B and andC).C). Taken together, our results support the conclusion that the DDR induced during parvovirus infection is mediated mainly through replication of the viral genome, for which NS1 is essential.
Because EPCs are difficult to transfect (13), we delivered the B19V dsDNA genome into S1 cells by electroporation. Replication of the B19V dsDNA genome, even in the absence of ssDNA production, was sufficient to trigger a DDR in transfected S1 cells at a level similar to that observed during B19V infection. These results indicate that replication of the B19V dsDNA genome during infection of EPCs initiates the DDR independently of the production of the ssDNA viral genome. However, the ssDNA genomes of B19V and other parvoviruses have hairpin termini at both ends (20). The gap between the termini displays a structure of a single-strand break (SSB) that is recognized by ATR (15). Since it was difficult to detect the ssDNA viral genome in the nucleus upon B19V infection (data not shown), we have not provided direct evidence that the B19V ssDNA genome elicits a DDR. However, when UV-inactivated B19V was used to infect EPCs at an MOI of 1,000 gc/cell, the same as that for the wild-type B19V in our study, neither γ-H2AX nor p-RPA32 was detected in infected cells (data not shown). It has been reported that infection of UV-inactivated AAV at a high MOI (5,000 gc/cell) elicited a DDR by mimicking the stalled replication fork (29). Further studies are warranted to determine the nature of the B19V ssDNA genome as an SSB substrate for ATR recognition.
B19V induces a DDR via replication of the RF genome, which involves replication intermediates (27) rather than via NS1 expression alone. The replication of parvovirus DNA follows the rolling-hairpin model of DNA replication, which involves a strand displacement step (19). Therefore, the formation of a replication origin complex at the replication origin, or other steps, such as nicking at the terminal resolution site by NS1 followed by unwinding of the RF DNA or hairpin by the helicase activity of NS1 (which resembles a stalled replication fork), is likely the key for triggering a DDR during virus infection. It is possible that ATR recognizes the nicked site within the terminal sequence of the genome and that the ssDNA binding protein RPA32 is recruited to the displaced ssDNA, which mimics a stalled replication fork. Thus, our observations have laid a solid foundation for further study of how replication of the ssDNA genome of autonomous parvovirus induces a DDR during virus infection.
Regulation of cell cycle progression, symbolized by arrest at the S or G2/M phase checkpoint, is a classic strategy for infecting viruses to exploit cellular components/mechanisms to aid viral genome replication (12, 23). MVM and MVC infections activate ATM-mediated signaling, which arrests infected cells at the S and G2/M phases; the DDR then directly leads to cell cycle arrest (1, 38). The Vpr protein of HIV is responsible for a DDR during HIV infection, which results in G2/M arrest through ATR-checkpoint kinase 1 (Chk1) signaling (66). It is intriguing that B19V DNA replication-induced DDR does not result in G2/M arrest, even though Chk1 and Chk2 are activated during infection (39). Therefore, our observations suggest multiple roles for virus infection-triggered DDR. In some cases, the DDR induces cell cycle arrest, which in turn facilitates virus replication or kills infected cells. In other cases, the DDR is directly involved in promoting DNA replication without arresting the cell cycle.
In summary, we have demonstrated that replication of the B19V dsDNA genome induces DDR signaling and that the DDR is dispensable for triggering cell cycle arrest during B19V DNA replication. However, the NS1 protein was instrumental in G2/M arrest, which did not require the putative activity of endonuclease and helicase but did require TAD2 function. Furthermore, this process was independent of the p53 pathway. Further studies are needed to examine how TAD2 functions to induce G2/M arrest and how this function coordinates with the function of the NLS to induce G2/M arrest. Thus, B19V may have evolved a unique strategy employing NS1 to overcome difficulties in arresting the cell cycle in rapidly proliferating EPCs. The reason why the induced DDR during B19V infection of EPCs, particularly the activation of ATR and ATM signaling, did not significantly result in G2/M arrest will be an interesting topic for future study.
This work was supported in full by PHS grant R01 AI070723 from the NIAID and grant R21 HL106299 from the NHLBI to J.Q.
We are indebted to Susan Wong at the Hematology Branch, NHLBI, NIH, for helping culture CD34+ human hematopoietic stem cells.
Published ahead of print 25 July 2012