Phosphorylated STAT5 directly facilitates parvovirus B19 DNA replication in human erythroid progenitors through interaction with the MCM complex

Productive infection of human parvovirus B19 (B19V) exhibits high tropism for burst forming unit erythroid (BFU-E) and colony forming unit erythroid (CFU-E) progenitor cells in human bone marrow and fetal liver. This exclusive restriction of the virus replication to human erythroid progenitor cells is partly due to the intracellular factors that are essential for viral DNA replication, including erythropoietin signaling. Efficient B19V replication also requires hypoxic conditions, which upregulate the signal transducer and activator of transcription 5 (STAT5) pathway, and phosphorylated STAT5 is essential for virus replication. In this study, our results revealed direct involvement of STAT5 in B19V DNA replication. Consensus STAT5-binding elements were identified adjacent to the NS1-binding element within the minimal origins of viral DNA replication in the B19V genome. Phosphorylated STAT5 specifically interacted with viral DNA replication origins both in vivo and in vitro, and was actively recruited within the viral DNA replication centers. Notably, STAT5 interacted with minichromosome maintenance (MCM) complex, suggesting that STAT5 directly facilitates viral DNA replication by recruiting the helicase complex of the cellular DNA replication machinery to viral DNA replication centers. The FDA-approved drug pimozide dephosphorylates STAT5, and it inhibited B19V replication in ex vivo expanded human erythroid progenitors. Our results demonstrated that pimozide could be a promising antiviral drug for treatment of B19V-related diseases.

The results of this study confirmed that phosphorylation of STAT5 is essential for B19V DNA replication. Mechanistically, the B19V RF DNA genome harbors STAT5-binding element (STAT5BE) within the minimal origins of DNA replication (Ori), located to the ITRs at each end of the viral genome. The binding site specifically binds phosphorylated STAT5 (pSTAT5). Moreover, our experiments revealed a novel interaction between STAT5 and minichromosome maintenance (MCM) complex; B19V exploits this interaction to recruit MCM complex to the viral replication centers for initiation of B19V DNA replication.

Inhibition of STAT5 phosphorylation completely inhibited B19V DNA replication
Our results previously demonstrated that pSTAT5A has a critical role in B19V infection of human EPCs cultured under hypoxic conditions [25], leading us to consider in this study whether specific inhibition of STAT5 phosphorylation affects B19V replication. This possibility was tested by treating cells with a specific inhibitor of STAT5 phosphorylation, pimozide [34]. At a final concentration of 15 μM, pimozide abolished >90% of the STAT5 phosphorylation in CD36 + EPCs, without altering the total expression of STAT5 (Fig 1A,  lane 4). CD36 + EPCs were incubated with pimozide 6 h prior to infection, and, at 48 h postinfection, numbers of B19V-infected (capsid-expressing) cells were reduced by 4.7-fold and 18.5-fold at 15 μM and 25 μM pimozide, respectively, compared with DMSO-treated cells (Fig 1B). Pimozide abolished viral DNA replication at both concentrations (Fig 1C). STAT5 dephosphorylation was confirmed in pimozide-applied infected cells (Fig 1D). These results suggested that inhibition of STAT5 phosphorylation abolishes viral DNA replication in B19V-infected CD36 + EPCs. Notably, treatment with pimozide at 15 μM did not significantly inhibit cell proliferation, as assessed by the BrdU incorporation assay (Fig 1E  & 1F).
Pimozide treatment, at a concentration as low as 10 μM, also abolished DNA replication of the B19V RF genome M20 in transfected UT7/Epo-S1 cells (S1A Fig, lane 3), and inhibited STAT5 phosphorylation (S1B Fig, lane 3). As controls, at 10 or 20 μM pimozide, cell proliferation was not significantly affected (S1C & S1D Fig). Taken together, our results suggested that phosphorylation of STAT5 is essential for viral DNA replication.
Phosphorylated STAT5 interacts with a consensus STAT5-binding element in the B19V minimal replication origin (Ori) The requirement of pSTAT5 for B19V DNA replication suggested that there might be a direct involvement of pSTAT5 in viral DNA replication. In silico analysis of the B19V genome demonstrated the presence of several consensus STAT5-binding elements (STAT5BEs) throughout the genome. STAT transcription factor binds a GAS or GAS-like motif with a consensus sequence of TTCN3GAA, TTCN3TAA, or TTAN3GAA [35]. TTCN3TAA binds STAT5 [36] and is one of the top ten STAT5BEs identified in a genome wide analysis by ChIP-seq [37]. A consensus STAT5BE is located within the previously identified 67-nt Ori in the B19V genome (Fig 2A) [38].
Binding of pSTAT5 from nuclear lysates of UT7/Epo-S1 cells to the STAT5BE in the Ori was confirmed by EMSA. A shifted band, indicating binding of protein to the probe, was observed in the presence of wild-type (wt) Ori-derived probe wt-Ori-39, but not the mut-Ori-39 that has the STAT5BE mutated (Fig 2B and 2C, lanes 2 vs 3). On incubation with an anti-pSTAT5 antibody, the level of shifted band was dramatically decreased (Fig 2D, lane 3). Because the EMSA was performed in the presence of excess amounts of non-specific competitor poly dI-dC, these results indicated specific binding of pSTAT5 to the B19V Ori.

Phosphorylated STAT5 is associated with replicating viral DNA in the viral DNA replication centers of B19V-infected EPCs
The association of STAT5 with B19V NS1 and the viral capsid was demonstrated by immunofluorescence assays (Fig 3A & 3B). STAT5 colocalized with NS1 and the viral capsid in the nucleus of B19V-infected CD36 + EPCs. The association of STAT5 with viral capsid was confirmed by the observation of fluorescent foci in B19V-infected cells in a proximity ligation B19V genomes of the single-stranded (ss) DNA form and full-length replicative form (RF) are depicted, along with the sequence of viral Ori that contains a consensus STAT5-binding element (STAT5BE), terminal resolution site (trs), two NS1-binding elements (NSBE1 and NSBE2), and two putative cellular factor-binding elements (CFBE) [38][39][40]. (B) Probes used in electrophoretic mobility shift assay (EMSA). Sequences of two 39-nt probes, wt-Ori-39 and mut-Ori-39, are shown with the consensus STAT5BE and the mutated STAT5BE (mSTAT5BE) highlighted. (C&D) EMSA. (C) 32 Plabeled Ori probes wt-Ori-39 (lane 2) and mut-Ori-39 (lane 3) were incubated with UT7/Epo-S1 nuclear lysate (NL) in the assay (Fig 3C), which produces an amplified signal when two labeled molecules are within 20 nm of one another [41].
Proximity ligation assay (Fig 3D) and confocal microscopy (Fig 3F) both demonstrated that STAT5 colocalized with replicating viral DNA that was pulse-labelled with BrdU in B19Vinfected CD36 + EPCs, which are parvovirus replication centers [42,43] as shown by proximity ligation assay using anti-BrdU and anti-capsid antibodies (Fig 3E). Interaction of pSTAT5 with the viral genome in cells was confirmed by ChIP assays in B19V-infected CD36 + EPCs and M20-transfected UT7/Epo-S1 cells. The pSTAT5-DNA complexes were pulled down with anti-pSTAT5(Y694) antibody, and bound viral Ori was detected by PCR. In the ChIP assay, cellular DNA was sheared to < 500 bp by sonication (Fig 4A). A specific PCR band was amplified in samples from B19V-infected or M20-transfected cells pulled down by anti-pSTAT5 (Y694) (Fig 4B, lane 4). Moreover, in UT7/Epo-S1 cells transfected with the B19V RF genome (M20), we observed that application of pimozide significantly decreased the amount of the Ori-containing fragments of the M20, as assessed by the quantitative ChIP assay targeting Ori (S4C Fig, Pimozide). Thus, these results confirmed the association of pSTAT5 with B19V Ori in B19V-infected CD36 + EPCs and M20-transfected UT7/Epo-S1 cells.

Disruption of the interaction between STAT5 and viral Ori inhibits viral DNA replication
A small molecule STAT5-SH2 inhibitor (STAT5-SH2i, CAS no. 285986-31-4) specifically targets the SH2 domain of STAT5 and inhibits STAT5 binding to DNA [44]. EMSA was performed to determine whether STAT5-SH2i disrupts the interaction between the STAT5 and B19V Ori. Incubation of either UT7/Epo-S1 nuclear lysates or purified pSTAT5 with increasing concentrations of the inhibitor showed that the STAT5-SH2i prevented formation of the STAT5-DNA formation in a dose-dependent manner (Fig 5A and Fig 2G). To examine the effect of the inhibitor on virus replication, CD36 + EPCs were pretreated with STAT5-SH2i 6 h prior to infection with B19V. The results showed that, at a final concentration of 500 μM, the inhibitor significantly decreased the virus-infected cell population by 10.7-fold (Fig 5B), and the level of viral RF DNA by~10-fold (Fig 5C), but not the expression level of pSTAT5 (Fig  5D), compared with the cells with DMSO treatment. Cell proliferation was not significantly affected by this level of inhibitor in mock-infected CD36 + EPCs (Fig 5E & 5F). The inhibition of viral DNA replication by STAT5-SH2i was also demonstrated in M20-transfected UT7/Epo-S1 cells (S2 Fig), and STAT5-SH2i significantly disrupted the interaction of pSTAT5 with the Ori of the B19V RF genome (M20) in vivo as shown by a ChIP assay (S4C Fig, STAT5-SH2i).
Derivatives of B19V replicative form genome with mutated STAT5-binding elements do not replicate in UT7/Epo-S1 cells The effect of mutation of the STAT5BE of the viral Ori on replication of the B19V RF genome was determined. The viral genome has an Ori sequence adjacent to each ITR, and the presence of non-specific competitor poly dI-dC. Products were subjected to non-denaturing 5% polyacrylamide gel electrophoresis (PAGE). Gels were dried and exposed to a phosphor screen. (D) Similarly, EMSA was performed with 32 Plabeled wt-Ori probes and 5 μg of NL in the presence of 5 μg of anti-pSTAT5(Y694) or IgG control antibody. (E) PAGE analysis of purified pSTAT5. 20 μl of pSTAT5 was analyzed by SDS-10% PAGE. Gels were either stained with Coomassie brilliant blue (left panel/CBB staining), or transferred to a PVDF membrane for Western blotting with an anti-pSTAT5(Y694) antibody (right panel/Western blot). (F&G) EMSA with purified pSTAT5. (F) 32 P-labeled wt-Ori-39 (lane 2) and mut-Ori-39 (lane 3) probes were incubated with purified pSTAT5 in the presence of poly dI-dC. Samples were run on 5% nondenaturing PAGE, dried, and exposed to a phosphor screen.  (A&B) STAT5 colocalizes with B19V NS1 and capsids. Mock-or B19V-infected CD36 + EPCs were co-stained and examined with rabbit anti-STAT5 and rat anti-B19V NS1 antibodies (A) or with rabbit anti-STAT5 and mouse anti-B19V capsid antibodies (B). (C-E) Proximity ligation assay. Infected cells were co-stained with rabbit anti-STAT5 and mouse anti-B19V capsid antibodies (C), or costained with rabbit anti-STAT5 and mouse anti-BrdU antibodies (D), or co-stained with mouse anti-B19V capsid and rabbit anti-BrdU antibodies (E), followed by a proximity ligation assay, which produces amplified signal for labeled molecules in close proximity. (F) STAT5 colocalizes with the replicating viral genome. Mock-or B19V-infected CD36 + EPCs were BrdU labeled to identify replicating viral ssDNA genomes. The treated cells were co-stained with rabbit anti-STAT5 and mouse anti-BrdU antibodies, followed by incubation with secondary antibodies. Images were taken with an Eclipse C1 Plus (Nikon) confocal microscope at 100 × magnification. Nuclei were stained with DAPI. https://doi.org/10.1371/journal.ppat.1006370.g003 Phosphorylated STAT5 facilitates B19 DNA replication by interaction with MCM STAT5BE was mutated in either the left ITR (N8 mOriL ) or right ITR (N8 mOriR ) or both ITRs (N8 mOri ) of the N8 replicating RF DNA that has half ITRs at both ends, as shown in Fig 6A. The replication capability of these mutated RF genomes was examined in UT7/Epo-S1 cells. Although the N8 RF DNA replicated well, much less replication occurred with N8 mOriL and N8 mOriR , and no replication was observed with N8 mOri (Fig 6B). The mutations in the STAT5BEs were then introduced into both ITRs of M20 RF genome, to make the M20 mOri mutant. No viral DNA replication was observed in M20 mOri -transfected cells (Fig 6C, lane 2). Although both M20 and M20 mOri RF genomes expressed NS1, viral capsid (a hallmark of B19V DNA replication [45]) was present only in M20-transfected cells (Fig 6D).
pSTAT5 interacts with the MCM complex of the pre-initiation complex of cellular DNA replication During initiation of cellular DNA replication, the origin recognition complex (ORC) binds to autonomously replicating sequence sites and recruits cell division control protein (CDC6) and DNA replication factor CDT1 to replication origins [46]. CDT1 recruits the MCM complex and primes replication initiation [47]. Although viral DNA replicates independently of ORC/ CDC6/CDT1, DNA viruses may require the MCM complex to initiate viral DNA replication [48]. In the parvovirus adeno-associated virus (AAV), MCM complex is required for in vitro reconstitution of viral DNA replication [49]. In the case of B19V, we previously found that MCM complex is associated with the viral DNA replication centers and has a role in B19V replication [50]. Chromatin immunoprecipitation (ChIP) assay. ChIP assay was performed using either infected CD36 + EPCs or transfected UT7/Epo-S1 cells, as indicated. (A) Crosslinked chromatin was sheared by sonication to sizes of~500 bp. (B) An anti-pSTAT5(Y694) antibody or negative control IgG was used to pull down DNA-protein complexes. Recovered DNA from UT7/Epo-S1 cells or CD36 + EPCs was examined for viral DNA by PCR with primer sets of F1/R1 and F1/R2, respectively, which span the Ori sequences of the B19V genome. pM20 plasmid was used as a template for positive controls of PCR. (C) A diagram of the Oritargeting PCR. The primers used for PCR are shown. Initially, to determine whether the viral NS1 protein has a role in recruitment of the MCM complex to the viral replication origin, we performed pull-down assays using lysates from NS1-expressing UT7/Epo-S1 cells. With pull-down of NS1, MCM and pSTAT5 were not detected (Fig 7A, lane 3), but the positive control transcription factor E2F5 (which interacts with B19V NS1 [27]) was detected, which suggested that NS1 has no role in recruitment of the MCM complex. By contrast, co-immunoprecipitation (Co-IP) with an anti-pSTAT5 antibody pulled down MCM5 protein of the MCM complex from lysates of UT7/Epo-S1 cells (Fig 7B). Similarly, Co-IP with an anti-MCM5 antibody pulled down pSTAT5, in addition to MCM2 (Fig 7C). The interaction between pSTAT5 and the MCM complex was DNA-independent, as DNase treatment of the lysate did not disrupt the interaction (Fig 7D, lane 4). Also, we show that MCM2, MCM3, MCM5 and MCM7 were associated with viral Ori in M20-transfected UT7/Epo-S1 cells, as confirmed by ChIP analyses (S4A Fig).
STAT5 and the MCM complex colocalized in CD36 + EPCs, irrespective of whether the cells were infected (Fig 7E). An association of the MCM complex with STAT5 was confirmed in both B19V-and mock-infected cells by the proximity ligation assay (Fig 7F). This association was blocked by treatment of pimozide in CD36 + EPCs (Fig 7G). Immunofluorescence detection of the viral capsid demonstrated that, following B19V infection, most of the cells were infected (Fig 7H).
pSTAT5 recruits MCM complex to B19V Ori to facilitate initiation of viral DNA replication Our demonstration that pSTAT5 interacts with viral Ori as well as the MCM complex suggested that B19V might exploit these interactions to initiate viral DNA replication. To test this hypothesis, we infected CD36 + EPCs with B19V, and at 36 h post-infection (when B19V DNA replication was at its peak), we treated the cells with STAT5-SH2i (Fig 8A). At 6 h post-treatment, the cells were collected for ChIP assay with anti-MCM2 antibody, which showed that MCM abundance on viral Ori decreased significantly in the presence of STAT5-SH2i, compared with untreated control cells (Fig 8B). Results with three-color confocal imaging demonstrated that MCM2 and STAT5 colocalized in mock-infected cells (in the absence of viral NS1) (Fig 8C, Mock). In infected cells, viral NS1 (which binds viral Ori) colocalized with both STAT5 and MCM, indicating that they were localized at viral DNA replication centers (Fig  8C, B19V). These results suggested that B19V utilizes viral Ori-STAT5 and STAT5-MCM interactions to recruit the MCM complex to viral DNA replication origins, to initiate viral DNA replication.

Pimozide is a promising candidate for the treatment of B19V infection
To confirm the efficacy of pimozide as a drug, we treated primary CD36 + EPCs with pimozide at various concentrations, and infected them with B19V. The cells were collected 48 h postinfection for quantification of viral DNA replication (RF DNA) by Southern blot analysis, which demonstrated that the IC 50 of pimozide for inhibition of viral DNA replication (the concentration at which 50% of viral DNA replication was inhibited) was 2.7 ± 0.69 μM derivatives with mutations in the STAT5BE of either the left ITR (N8 mOriL ), right ITR (N8 mOriR ), or both (N8 mOri ), were transfected into UT7/Epo-S1 cells. (C) M20, and M20 mOri , a derivative of the M20 RF DNA with STAT5BEs of both ITRs mutated, were transfected into UT7/Epo-S1 cells. At 48 h post-transfection, cells were collected for Hirt DNA extraction. And Hirt DNA samples were analyzed by Southern blotting with an M20 DNA probe. RF DNA (RF), ssDNA (ss), and Dpn I-digested DNA (shown with a line) are indicated. Mitochondrial DNA (Mito DNA) was used as a loading control (lower panels). (D) Viral protein expression of B19V DNA mutants. M20 or M20 mOri transfected UT7/Epo-S1 cells were stained with anti-NS1 or anti-capsid antibodies. Confocal images were taken with an Eclipse C1 Plus (Nikon) microscope at 100 × magnification. https://doi.org/10.1371/journal.ppat.1006370.g006 Phosphorylated STAT5 facilitates B19 DNA replication by interaction with MCM  Fig 7. pSTAT5, but not NS1, interacts with the MCM complex. (A) Immunoprecipitation (IP) assay. Cell lysates of NS1 Flagexpressing UT7/Epo-S1 cells were prepared for pull-down assays with either anti-Flag-conjugated beads or control beads. Immunoprecipitated proteins were examined for the presence of MCM2 by Western blotting. Blots were reprobed with rabbit anti-pSTAT5(Y694), anti-E2F5, and anti-Flag antibodies. Detection of E2F5 was used as a positive control for NS1 IP. (B) Co-IP assay. UT7/Epo-S1 cells were collected, washed, and lysed with RIPA buffer. After centrifugation, the supernatant was incubated with either rabbit anti-pSTAT5(Y694) or control IgG antibody. Immunoprecipitated proteins were blotted for the presence of the MCM complex with an anti-MCM5 antibody and for pSTAT5 with rabbit anti-pSTAT5(Y694). (C) Reverse Co-IP (mean ± standard error) (Fig 9A). To examine the effect of pimozide on colony formation in the absence of virus infection, CD36 + EPCs were incubated with pimozide at increasing concentrations on Day 7 for 2 days, and then cultured in methyl cellulose-based medium for colony formation. After 10 days, numbers of colonies were counted (Fig 9B). Pimozide only assay. Reverse Co-IP was performed with an anti-MCM5 antibody. Immunoprecipitated proteins were examined for pSTAT5, MCM2, and MCM5, respectively. (D) Co-IP of lysates treated with DNase. UT7/Epo-S1 cell lysates, either treated or untreated with DNase (750 units of Benzonase) were incubated with anti-pSTAT5(Y694) or control IgG antibodies for Co-IP assay, and immunoprecipitated proteins were examined for MCM2 by Western blot analysis. (E-H) Immunofluorescence analysis. (E&F) Mock-or B19V-infected CD36 + EPCs were co-stained with rabbit anti-STAT5 and mouse anti-MCM2 antibodies, followed by (E) incubation with respective secondary antibodies, or by (F) proximal ligation assay, which produces amplified signal for labeled molecules in close proximity. (G) CD36 + EPCs were incubated with either DMSO or pimozide (at 30 μM) for 2 days. And then the cells were co-stained with rabbit anti-STAT5 and mouse anti-MCM2 antibodies for proximity ligation assay. (H) Infected EPCs were stained with an anti-capsid antibody. Confocal images were taken with an Eclipse C1 Plus (Nikon) microscope at 100 × magnification.

Discussion
We have now demonstrated that STAT5 is directly involved in B19V DNA replication. Importantly, STAT5 specifically interacts with the MCM complex, the eukaryotic DNA helicase complex that is required for the formation and elongation of the cellular DNA replication fork [51]. We therefore propose a novel model of B19V DNA replication in human EPCs, in which STAT5 functions as a mediator protein that brings the MCM complex to the viral DNA replication origins. Our results also identify pSTAT5 as a target for inhibition of B19V infection. In addition, as the STAT5-MCM interaction is independent of infection, we envisage an important role of this interaction in the context of cellular replication and transcription in human EPCs, which warrants further investigation.

Phosphorylated STAT5A is directly involved in B19V replication
STAT5 is phosphorylated at a single conserved tyrosine residue (Tyr694 in STAT5A and Tyr699 in STAT5B), and these phosphotyrosine motifs, upon intermolecular interaction, enable formation of either homodimers or heterodimers of STAT5A/B [52,53]. These dimers accumulate in the nucleus and bind DNA, to transactivate target genes [52]. EPO-activated JAK2 phosphorylates STAT5 in human EPCs [54]. We examined the relative expression of STAT5A and STAT5B in UT7/Epo-S1 and CD36 + EPC lysates with a STAT5A/B pan-specific antibody, and found that STAT5A was predominantly expressed in both cell types (S3A Fig). This result agrees with the observations that JAK2 kinase predominantly phosphorylates STAT5A in cells of erythroid lineage [26], and a constitutively phosphorylated STAT5A (1 Ã 6) variant enhances virus replication, whereas knockdown of STAT5A inhibits virus replication in B19V-infected EPCs [25].
STAT5B promotes viral DNA replication, but, during replication of human papillomavirus 16 (HPV16), STAT5B enhances viral DNA replication indirectly via regulation of TopBP1 expression, leading to the activation of ATR kinase [55]. In a proof-of-concept experiment, fusion of STAT5BEs to the DNA replication origin of polyoma virus replicon DNA improved replication efficiency in transfected mouse lymphoid BA/F3 cells, corroborating the direct role of STAT5 in viral DNA replication [56]. CD36 + EPCs have to be cultured in the presence of EPO for proliferation and differentiation [24], which dominantly leads activation of STAT5A (S3A Fig) through the EPO-JAK2-STAT5 pathway [25]; however, a DDR or activation of ATR is not observed in normal (uninfected) CD36 + EPCs [29,57] (S5A Fig). Furthermore, in hydroxyurea-treated CD36 + EPCs, both ATR and ATM were activated; however, application of pimozide did not change the level of phosphorylated ATR or ATM (S5A Fig). As ATR activation enhances B19V replication [57], these lines of evidence suggest that pSTAT5 does not utilize the STAT5-ATR pathway to facilitate B19V replication in CD36 + EPCs. Moreover, B19V infection per se did not affect STAT5 phosphorylation (S5B Fig). Of note, the binding of pSTAT5 to the Ori, which locates in front of the B19V P6 promoter, did not obviously transactivate the P6 promoter (S6 Fig). Thus, our results provide the first evidence that an authentic virus, B19V, depends on direct binding of pSTAT5 to its replication origin (Ori) for viral DNA replication.

Phosphorylated STAT5 interacts with the MCM complex and recruits it to viral replication origins during DNA replication initiation
B19V infection induces late S-phase arrest in human EPCs, and S-phase factors are fully utilized by the virus to replicate its genome [50]. During cellular DNA replication, ORC-CDC6-CDT1 binding to the replication origin is a priming event that takes place in G1-phase [51]. Furthermore, CDT1 recruits the MCM complex and subsequently the whole replisome via formation of the MCM-CDC45 complex [51]. Notably, no such priming takes place during S-phase, so that chromosomes are not replicated multiple times [46]. However, viruses have evolved different mechanisms to initiate viral DNA replication. For examples, SV40 has the large T antigen that binds SV40 DNA replication origin and has helicase activity, and also recruits the replication machinery by interacting with DNA replication factors, such as replication factor A, DNA polymerase α and topoisomerase I [58]. Parvoviruses use the large non-structural protein NS1, which binds directly to the viral origin and has helicase and nickase activities that facilitate viral DNA replication [59]. In parvovirus AAV, the MCM complex is essential to AAV2 DNA replication in vitro [49], and is probably recruited by interaction with Rep78, the large viral nonstructural protein [60].
In the case of B19V, the MCM complex is localized to the viral DNA replication centers and is required for viral DNA replication [50]. However, we did not observe any interaction between the B19V NS1 protein and the MCM complex, suggesting that the complex is recruited to the viral DNA replication centers by an alternative mechanism. Here, our results provided evidence that STAT5 interacts with the MCM complex in human EPCs, without involvement of viral or cellular DNA. These cells express STAT5A more abundantly than STAT5B (S3A Fig), but both STAT5A and STAT5B proteins interact with the MCM complex (S3C Fig).
During B19V infection, STAT5 is recruited to the viral DNA replication origin by direct interaction with STAT5BEs in the Ori sequences of the viral genome, thereby bringing the MCM complex to the viral Ori. Outside of the Ori, there are additional 6 putative STAT5BEs, and we tested that two of them in the capsid proteins-coding region also bound pSTAT5 (S4D  Fig). Since there is no putative terminal resolution site (trs) and NS1-binding sites outside of the Ori, we speculate that these STAT5BEs outside of the Ori do not contribute to B19V DNA replication. We hypothesize that MCM complex recruited by pSTAT5 at Ori may contribute to virus replication through its helicase activity or the recruitment of other DNA replication factors to the viral origin [51]. Notably, PIF (parvovirus initiation factor), a member of the KDWK family of transcription factors, has been shown to bind two adjacent "ACGT" motifs in front of the NS1 binding site of left-hand replication origin (OriL TC ) of the Protoparvoviurs minute virus of mice (MVM) [61,62]. PIF stabilizes the binding of NS1 to the Ori, which is critical for the activation of NS1 nickase [63]. In B19V, at least in an in vitro nicking assay, B19V NS1 is sufficient to cleave the Ori [40]. However, whether the binding of STAT5 to B19V Ori or the recruited MCM complex also involves in NS1 nickase activity of the Ori at trs (Fig 2A) warrants further investigation.

Pimozide, an FDA-approved drug, shows promise for the treatment of B19V infection
To date, no specific treatment (either anti-viral or vaccine-based) exists for B19V infection.
We have now demonstrated that pimozide, an FDA-approved anti-psychotic drug that is used in the treatment of a wide range of diseases [64] and could be potentially used to treat chronic myeloid leukemia, in which it specifically targets cancer cells, without affecting CD34 + hematopoietic stem cells [34]. Pimozide specifically inhibits STAT5 phosphorylation without affecting JAK2 activation or JAK2-derived signaling pathways; however, the underlying mechanism is unknown yet [65]. The pSTAT5 is presumably required for recruitment of the MCM complex to the viral Ori, and facilitates B19V replication in human EPCs. Pimozide is a potent inhibitor of B19V replication, with an IC 50 of~2.7 μM. At 15 μM, pimozide does not have a significant effect on proliferation of human EPCs expanded ex vivo, and has only moderate effect (~15% reduction) on colony formation of EPCs. As STAT5A phosphorylation plays a key in B19V replication in human EPCs under hypoxic conditions [25], these lines of evidence suggest that the inhibition of B19V replication in CD36 + EPCs is not a side-effect of the pimozide. Antivirals such as cidofovir and ribavirin are used in the treatment of adenovirus infection, and have IC 50 values of 15 μM for cidofovir and 25 μM for ribavirin [66]. Importantly, when we applied both pimozide and STAT5-SH2i (at 15 and 250 μM, respectively), a significant synergistic inhibition of B19V infection was observed (S7 Fig). Therefore, we expect that a clinical trial should be conducted to examine pimozide as a treatment for B19V infection of patients with sickle-cell disease and immunocompromised patients and as anti-viral prophylaxis of transplant recipients.

Ethics statement
We purchased CD34 + hematopoietic stem cells, which were isolated from bone marrow of a healthy human donor, from AllCells LLC (Alameda, CA) without any identification information on the cells, and, therefore, an institutional review board (IRB) review was waived.

Primary cells and cell lines
Primary human CD36 + EPCs were expanded ex vivo from CD34 + hematopoietic stem cells as previously described [24,25,67]. Briefly, hematopoietic CD34 + stem cells, purchased from All-Cells, LLC (Alameda, CA), were grown in Wong medium under normoxia up to Day 4 and frozen in liquid nitrogen [25]. In each experiment, Day 4 cells were thawed and grown under normoxia in an atmosphere containing 5% CO 2 and 21% O 2 at 37˚C for 2-3 days, prior to incubation under hypoxia at 5% CO 2 and 1% O 2 .

Virus and infection
Plasma samples containing B19V at~1 × 10 12 viral genomic copies per ml (vgc/ml) were obtained from ViraCor Eurofins Laboratories (Lee's Summit, MO). After 2 days of hypoxia, CD36 + EPCs were infected with B19V at a multiplicity of infection (MOI) of~1,000 vgc per cell. At 48 h post-infection, the infected cells were analyzed.

Proximity ligation assay
Duo link In-Situ Red Mouse/Rabbit kit (cat# DUO92101) was purchased from MilliporeSigma (St Louis, MO). Proximity ligation assay was performed following the manufacturer's instructions, as described previously [69].

Immunofluorescence assay and confocal imaging
Immunofluorescence assay was carried out as described previously [25,50]. Briefly, infected EPCs were deposited on slides by cytospinning, fixed with 3.7% paraformaldehyde for 30 min, and permeabilized with phosphate-buffered saline (PBS, pH7.2) containing 0.5% Triton X-100 (PBS-T) for 5 min at room temperature. Non-specific interactions were blocked with 3% bovine serum albumin (BSA) before subsequent incubation with primary and fluorescencelabelled secondary antibodies. The slides were visualized with a Nikon confocal microscope, and images were taken at 100 × magnification.
Plasmid construction pM20 contains the full-length B19V replicative from (RF) genome (nt 1-5596), and pN8 contains a half-ITR deleted B19 RF genome (nt 199-5410) [38,70]. They are diagramed in Fig  2A. pN8 mOriL and pN8 mOriR were constructed by mutating the STAT5BE of the Ori in the left and right half ITRs of the pN8, respectively. Both STAT5BE were mutated in pM20 and pN8 resulted in pM20 mOri and pN8 mOri , respectively, which are diagramed with mOri shown, and the sequence of mutated Ori in the half right ITR is depicted (Fig 6A).

Flow cytometry and cell-cycle analysis
B19V-infected CD36 + EPCs were examined for virus infection by flow cytometry analysis with an anti-B19V capsid antibody, as described previously [25,50]. For cell-cycle analysis, a bromodeoxyuridine (BrdU) incorporation assay was used, as described previously [50].

Southern blot analysis
Lower molecular DNA (Hirt DNA) was extracted from either B19V-infected CD36 + EPCs or transfected UT7/Epo-S1 cells by a Hirt extraction method, as described previously [45]. Hirt DNA extracted from UT7/Epo-S1 cells was further digested with Dpn I to remove non-replicated plasmid DNA input. Southern blot analysis was performed as reported previously [28,45]. B19V RF DNA M20 excised from pM20 with Sal I was used as a probe.

Chromatin immunoprecipitation (ChIP) assay
Chromatin immunoprecipitation (ChIP) assay was performed essentially as described previously [74,75] with modifications. Cells were fixed in 1% formaldehyde for 10 min at room temperature and then quenched in 125 mM glycine. Fixed cells were washed with PBS, and then lysed in 400 μl of Lysis Buffer (10 mM Tris-HCl, pH 8.0, 10 mM NaCl, 0.2% NP-40, 1 mM PMSF, and PIC) and incubated for 10 min on ice. After centrifugation at 2,500 rpm for 5 min at 4˚C, the nuclear pellet was resuspended in 100 μl of Nuclear Lysis Buffer (50 mM Tris-HCl, pH 8.1,10 mM EDTA, 1% SDS, and PIC) for 10 min on ice. One ml IP Dilution Buffer (20 mM Tris-HCl, pH 8.1, 2 mM EDTA, 150 mM NaCl, 1% Triton X-100, and 0.01% SDS) was added, and chromatin was sheared by sonication at 80% power for 10 cycles of 15 s pulse and 1 min rest. Sonicated samples were centrifuged to remove debris, and the supernatant was split aliquots. Antibody (2.5 μg) was added to each aliquot, and the mixtures were incubated overnight at 4˚C. For each sample, 10 μg of yeast tRNA was added to 40 μl of cold PBS-prewashed Protein A/G beads (Gold BioTechnology, Inc., St Louis, MO), and this mixture was added to the sample containing antibody and incubated with rocking for 6 h. Beads were collected by centrifugation and washed with IP Wash-1 (20 mM Tris, pH 8.1, 2 mM EDTA, 50 mM/ 500mM NaCl, 1% Triton X-100, 0.1% SDS) three times (first at low salt of 50 mM and then twice at 500 mM) for 10 min each at 4˚C, followed by one wash with IP Wash-2 (10 mM Tris, pH 8.1, 1 mM EDTA, 0.25 M LiCl, 1% NP-40, and 1% deoxycholic acid) for 10 min at 4˚C. The beads were then washed with cold TE, and protein-DNA complexes were eluted twice using 200 μl of Elution Buffer (100 mM sodium bicarbonate and 1% SDS) for 10 min at room temperature. Crosslinking was reversed by addition of 16 μl of 5 M NaCl and incubation at 65˚C. DNA was purified with a Qiagen PCR purification kit (Qiagen, Hilden, Germany), and ChIP product was recovered in 50 μl of H 2 O, and used for PCR or quantitative PCR (qPCR) analysis.

Colony formation assay
The colony formation assay was performed with methyl cellulose-based medium (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions, with modifications. Briefly, CD36 + EPCs were cultured in Wong expansion medium and were treated with pimozide at various concentrations on Day 7. After 48 hours, !3 × 10 4 cells from each well were cultured in semi-solid methyl cellulose-based medium for 10-12 days, at which time colony counts were assessed by someone who was blinded to the experimental conditions.

Immunoprecipitation assay and Western blotting
Co-immunoprecipitation (Co-IP) assay was performed as previously described [73,76]. Briefly, UT7/Epo-S1 cells were collected, washed with PBS, and lysed in radioimmunoprecipitation assay (RIPA) buffer. After centrifugation at 12,000 rpm for 20 min at 4˚C, supernatant was taken and split into aliquots. Each aliquot was incubated with 3 μg of an antibody of interest overnight at 4˚C, and then 40 μl of Protein A/G beads (washed with ice-cold PBS three times beforehand) was added, followed by incubation for 6 h. The beads were collected by centrifugation and washed three to five times with 1 × PBS, and then resuspended in 1 × Laemmli sample buffer. Samples were boiled for 10 min and run on 10% SDS-polyacrylamide gels for Western blot analysis, which was performed as described previously [25,50,77]. Pull-down assay was performed similarly to Co-IP, except that anti-Flag-conjugated beads or control beads were used.
Horseradish peroxidase (HRP)-conjugated anti-mouse and anti-rabbit secondary antibodies were purchased from Sigma, and fluorescein isothiocyanate (FITC)-, Texas Red-, and Dylight405-conjugated anti-mouse, anti-rat, and anti-rabbit secondary antibodies were all purchased from Jackson ImmunoResearch (West Grove, PA).

Statistics
Statistical analysis was performed using GraphPad Prism Version 7.0. Statistical significance was determined by using 1-way ANOVA analysis, followed by Tukey-Kramer post-test for comparison of three or more groups and unpaired (Student) t-test for comparison of two groups. Error bars show mean and standard deviation (Mean ± SD) unless otherwise specified. Differential expression of STAT5A and STAT5B. (A) Cell lysates of UT7/Epo-S1 and EPCs were subjected to Western blotting with STAT5A/B pan-specific, STAT5A-specific, or STAT5B-specifc antibodies. Asterisks indicate dimerized or degraded or non-specific protein bands. (B) Purified STAT5 of UT7/Epo-S1 cells was subjected to Western blotting with a STAT5A/B pan-specific antibody. (C) Both STAT5A and STAT5B interact with the MCM2 complex. UT7/Epo-S1 cells were collected and lysed with RIPA buffer. The lysates were incubated with an anti-MCM2 or control IgG antibody for co-immunoprecipitation (Co-IP). Immunoprecipitated proteins were blotted for the presence of STAT5A, STAT5B, and MCM2 with anti-STAT5A, anti-STAT5B, and anti-MCM2 antibodies, respectively. The precipitated IgG heavy chain is also shown. (TIF) S4 Fig. Analyses of MCM or STAT5 binding to B19V genome by ChIP assay. (A) UT7/Epo-S1 cells were transfected with M20 and allowed to replicate for 48 h under hypoxic conditions. Cells were collected for ChIP analysis. Anti-MCM2, anti-MCM3, anti-MCM5, anti-MCM7, and control IgG antibodies were used to pull down DNA-protein complex. Recovered DNA was analyzed by qPCR targeting the viral origin (Ori-qPCR). Error bars represent standard deviations taken from at least three experiments. P values were calculated using a Student's t test, compared to the IgG control. ÃÃ P<0.01; Ã P<0.05. (B) A diagram of the Ori-qPCR amplicon targeting the viral replication origin (Ori) at the left ITR (L-ITR). The starting nucleotide numbers of both forward and reverse (F and R) and the location of the probe are indicated. (C) UT7/Epo-S1 cells were treated with either DMSO, STAT5-SH2i inhibitor (at 500 μM) or pimozide (at 15μM), as indicated in the figure, at 6h prior to transfection. Then, the cells were transfected with M20 and cultured under hypoxic conditions for 48 h. Cells were collected for ChIP analysis using an anti-STAT5 and Ori-qPCR. Error bars represent standard deviation taken from at least three experiments. P values were calculated using one-way ANOVA followed by Tukey-Kramer post-test, compared with the M20 group. ÃÃÃÃ denotes P<0.0001. (D) Mock-or M20-transfected UT7/Epo-S1 cells, cultured under hypoxic conditions for 48 h, were collected for ChIP assay using an anti-STAT5 antibody, followed by PCR using primers: forward (F, nt 3135-3156), 5'-GGA CTG TAG CAG ATG AAG AGC T-3', and reverse (R, nt 3393-3373), 5'-GTG GCC CCC TCA CTC CAC AT-3', primes as indicated in the diagram. Rabbit IgG was used as a negative control of pull-down, and M20 DNA was used as PCR positive control.  [29]. (B) UT7/Epo-S1 (S1) cells or NS1-expressing UT7/Epo-S1 (NS1-S1) cells were transduced with Lenti-ATF/P6-GFP an