Growth Transformation of B Cells by Epstein-Barr Virus Requires IMPDH2 Induction and Nucleolar Hypertrophy

ABSTRACT The in vitro growth transformation of primary B cells by Epstein-Barr virus (EBV) is the initial step in the development of posttransplant lymphoproliferative disorder (PTLD). We performed electron microscopic analysis and immunostaining of primary B cells infected with wild-type EBV. Interestingly, the nucleolar size was increased by two days after infection. A recent study found that nucleolar hypertrophy, which is caused by the induction of the IMPDH2 gene, is required for the efficient promotion of growth in cancers. In the present study, RNA-seq revealed that the IMPDH2 gene was significantly induced by EBV and that its level peaked at day 2. Even without EBV infection, the activation of primary B cells by the CD40 ligand and interleukin-4 increased IMPDH2 expression and nucleolar hypertrophy. Using EBNA2 or LMP1 knockout viruses, we found that EBNA2 and MYC, but not LMP1, induced the IMPDH2 gene during primary infections. IMPDH2 inhibition by mycophenolic acid (MPA) blocked the growth transformation of primary B cells by EBV, leading to smaller nucleoli, nuclei, and cells. Mycophenolate mofetil (MMF), which is a prodrug of MPA that is approved for use as an immunosuppressant, was tested in a mouse xenograft model. Oral MMF significantly improved the survival of mice and reduced splenomegaly. Taken together, these results indicate that EBV induces IMPDH2 expression through EBNA2-dependent and MYC-dependent mechanisms, leading to the hypertrophy of the nucleoli, nuclei, and cells as well as efficient cell proliferation. Our results provide basic evidence that IMPDH2 induction and nucleolar enlargement are crucial for B cell transformation by EBV. In addition, the use of MMF suppresses PTLD. IMPORTANCE EBV infections cause nucleolar enlargement via the induction of IMPDH2, which are essential for B cell growth transformation by EBV. Although the significance of IMPDH2 induction and nuclear hypertrophy in the tumorigenesis of glioblastoma has been reported, EBV infection brings about the change quickly by using its transcriptional cofactor, EBNA2, and MYC. Moreover, we present here, for the novel, basic evidence that an IMPDH2 inhibitor, namely, MPA or MMF, can be used for EBV-positive posttransplant lymphoproliferative disorder (PTLD).

dehydrogenase 2 (IMPDH2), which is the rate-limiting enzyme for de novo GTP synthesis, was induced by day 2. This induction was required for nucleolar growth and the increased sizes of the nuclei and cells. IMPDH2 induction depended on EBNA2 and MYC. Pharmaceutical IMPDH2 inhibition prevented EBV-driven growth transformation and resulted in prolonged survival in a mouse xenograft model. Notably, the stimulation of B cell proliferation by the CD40 ligand (CD40L) and interleukin-4 (IL-4) was associated with IMPDH2 induction and nucleolar hypertrophy. Therefore, IMPDH2 is a key molecule that is involved in nucleolar hypertrophy in B cell growth transformation by EBV and also in proliferation in the absence of EBV. These results suggest that IMPDH2 inhibitors may be used as immunosuppressants in transplant patients to prevent posttransplant lymphoproliferative disorder (PTLD).

RESULTS
Morphological changes in B cells after EBV infection. To analyze the morphological changes in EBV-infected B cells during growth transformation, primary B cells were infected with EBV at a multiplicity of infection (MOI) of 3. The cells were harvested at the indicated time points after infection for a TEM analysis. At day 0 postinfection (dpi), a large area of the peripheral nuclear region was enriched with dark heterochromatin, which is a typical feature of resting B cells, and the nucleolus was not evidently detectable (Fig. 1A). However, the heterochromatin structure receded after infection, likely representing hypomethylation (3,(16)(17)(18). The size of the nuclei and cells increased after EBV infection (Fig. 1A), as reported previously (11,14). The nuclear size increased by twofold to fourfold at 4 to 14 dpi and then gradually decreased, although it remained larger than that observed at 0 dpi (Fig. 1B). The nucleolar size remained increased from that observed at 2 dpi (Fig. 1A). The nucleolar area was at its maximum (about 2 mm 2 per nucleolus, on average) at 4 dpi, and this was followed by a slight decrease (Fig. 1C). However, the nucleolar areas remained enlarged. The proportion of nucleolar area in the nuclear area increased immediately after infection and peaked at 4 dpi (Fig. 1D).
We performed immunofluorescence assays using DAPI, anti-GAPDH, and anti-nucleolin or anti-fibrillarin antibodies to visualize the nucleus, cytoplasm, and nucleolus, respectively. The stained sections were observed using confocal microscopy and 3D reconstruction to quantify the nuclear/cellular volumes and nucleolar major axes ( Fig. 1E-K). The nuclear volume peaked at 4 dpi, and this was followed by a slight decrease (Fig. 1F), although there was some variation in volume. The cytoplasm volume was hardly detectable at 0 dpi, but it peaked at 4 dpi and slightly decreased thereafter (Fig. 1G). The cell volume (nucleus plus cytoplasm) was at its highest at 4 dpi and slightly decreased thereafter (Fig. 1H), whereas the volume ratio of the nucleus to the cell did not increase and, in fact, decreased slightly (Fig. 1I), likely because of the greater increase in cell volume than in the nuclear volume. A similar experiment was performed with nucleolar staining (Fig. 1J) to measure the major axis of the nucleoli. The results (Fig. 1K) were similar to those obtained from the TEM (Fig. 1C).
In addition, we counted the number of nucleoli in the nuclei by using 3D images (Fig. 1L). Noninfected primary B cells typically possessed either no nucleoli or one nucleolus (day 0), but EBV infection led to an increase in the number of nucleoli, which reached its maximum at 7 dpi and was followed by a slight decrease to 2 to 4 nucleoli per cell at 28 dpi.
Our findings are summarized in Fig. 1M. The nucleolar size increased as early as 2 dpi and slightly decreased after 4 dpi (Fig. 1K). EBV infection led to more nucleoli per cell, reaching its maximum at 7 dpi (Fig. 1L). The transformed cells (at 28 dpi) had more and larger nucleoli than did the noninfected cells (0 dpi). The heterochromatin domain receded at 2 dpi and persisted thereafter (Fig. 1A). The nuclear and cellular sizes were the largest at 4 or 14 dpi (Fig. 1F-H), indicating that nucleolar enlargement occurred earlier than nuclear and cellular enlargement.
Role of IMPDH2 in nucleolar hypertrophy. Cancerous cells typically have larger and more nucleoli, and enlarged nuclei and cells, compared to normal cells (19)(20)(21), although the underlying mechanisms are unclear. Recently, Kofuji et al. (22) found that  IMPDH2, which is a critical enzyme that is involved in GTP synthesis, is overexpressed in glioblastomas. This leads to nucleolar hypertrophy, the increased synthesis of pre-rRNA and pre-tRNA, and cell proliferation. Our RNA-seq data (12) have shown that IMPDH2 expression is increased by 2 dpi, and this is followed by a slight gradual decrease, although the level remains high ( Fig. 2A). Based on our RNA-seq data ( Fig. 2A) and the results of previous studies (22), we hypothesized that IMPDH2 may play a crucial role in the nucleolar enlargement and growth transformation of B cells via EBV infection. GTP and ATP levels were significantly increased by EBV infection ( Fig. 2B; Fig. S1A), and this is in line with previous results for glioblastoma (22). The synthesis of rRNA and tRNA was markedly upregulated at 2 dpi ( Fig. 2C and D). We found that the level of GNL3 (also known as nucleostemin), which is a nucleolar GTP-binding protein that is stabilized by GTP, was increased by EBV (Fig. 2E). Therefore, the EBV infection of primary B cells caused prompt IMPDH2 induction, and this led to the increased biosynthesis of GTP and other molecules that are required for cell proliferation within two days. Next, we analyzed the underlying mechanism of IMPDH2 induction by EBV infection. EBNA2 is the key molecule that is expressed immediately after infection ( Fig. 2F and H), and it regulates the transcription of viral and other cellular genes. The MYC gene has a low expression in primary B cells, whereas EBNA2 induces MYC gene expression by 2 dpi (Fig. 2G and I) (23,24). A previous study reported similar induction patterns for IMPDH2, EBNA2, and MYC after EBV infection in primary B cells ( Fig. S1B-D), although IMPDH2 induction was not evaluated in detail (10). Because MYC reportedly induces IMPDH2 in melanoma (25), we used a specific inhibitor of MYC (10058-F4). The inhibitor reduced the IMPDH2 level in a dose-dependent manner (Fig.  S2A). EBNA2 gene disruption in the EBV genome decreased the induction of IMPDH2 transcription (Fig. 2J) and MYC (Fig. 2K). Because the bromodomain (BRD) inhibitor JQ1, which typically inhibits super-enhancer activity, decreased the IMPDH2 induction caused by EBV infection (Fig. S2B), we assume that the IMPDH2 gene is induced in a manner that is dependent on a super-enhancer involving both EBNA2 and MYC. Therefore, we checked the accumulation of transcriptional activators/regulators, including EBNA2, EBNA-LP, EBNA3A/C, RelA, RelB, cRel, p50, p52, and MYC, and CTCF at the genomic region spanning the IMPDH2 locus in LCLs (Fig. S2C) (26)(27)(28). The ChIP-seq data show EBNA2, EBNA-LP, and MYC accumulation at the promoter region of the IMPDH2 (Fig. S2C, yellow) as well as active epigenetic marker H2K27 acetylation. Furthermore, a ChIA-PET analysis revealed that the IMPDH2 promoter region organizes 3D connections with several upstream and downstream enhancers, thereby suggesting that the transcriptional upregulation of the IMPDH2 gene by EBV is mediated by the super-enhancer.
To further explore the effects of EBNA2 and MYC on IMPDH2 induction, we constructed a correlation chart using the RNA-seq data. The levels of IMPDH2 and MYC were strongly correlated (Fig. 2L), suggesting that MYC is important for IMPDH2 induction ( Fig. 2L; Table 1). However, the IMPDH2 level was poorly correlated with EBNA2 ( Fig. S2D; Table 1). These results are explained by the high level of IMPDH2 gene expression in the restricted latency patterns of EBV-negative cancerous cell lines or EBV-positive cell lines, such as NK/T cell lines ( Fig. S2D; Table 1). In addition, the dpi, the nucleus and cytoplasm of the EBVinfected cells increased in size by almost twofold to fourfold, compared to noninfected cells (0 dpi). Nucleolar enlargement reached its maximum at 2 to 4 dpi. The number of nucleoli increased from 4 to 14 dpi and reached its maximum at 7 dpi. Two-tailed Student's t tests were used to indicate betweengroup differences. Ns, not significant; *, P , 0.05; **, P , 0.01.   Table 1. Correlations were examined via a Pearson correlation analysis. (M) Graphical summary of IMPDH2 induction and immortalization. EBNA2 induced IMPDH2 expression predominantly through MYC-dependent mechanisms. IMPDH2 increased the levels of GTP, rRNA/tRNA, and GNL3, leading to immortalization. The data are represented as the mean 6 standard deviation (SD). Two-tailed Student's t tests were used to indicate between-group differences. The P values are shown in each graph. presents our hypothesis regarding the mechanism underlying IMPDH2 induction, nucleolar hypertrophy, and immortalization of B cells by an EBV infection.
To clarify whether IMPDH2 induction is involved in the B cell proliferation in the absence of EBV, we used CD40L and IL-4 to transiently increase the proliferation of primary B cells (Fig. 3A). The cell number was successfully increased by 6.7-fold at 6 dpi ( Fig. 3B). The levels of IMPDH2 and GNL3 were significantly increased by CD40L and IL-4, but the MYC expression remained unchanged (  and numbers were also observed ( Fig. 3F and G), suggesting that IMPDH2 is the key molecule that is involved in the cell proliferation and nucleolar hypertrophy of primary B cells with or without an EBV infection, whereas MYC is not required to induce IMPDH2 in the case of activation by CD40L and IL-4. Therefore, IMPDH2 induction may be regulated by several pathways. An EBV infection stimulates the EBNA2-MYC axis to induce IMPDH2, whereas CD40L and IL-4 activate alternative pathways besides MYC.
Effect of IMPDH inhibition. To provide further evidence that IMPDH2 is involved in the EBV-driven nucleolar enlargement of primary B cells, the effects of a specific inhibitor of the enzyme mycophenolic acid (MPA) were examined (Fig. 4). A TEM analysis (Fig. 4A) and the quantification of the area (Fig. 4B-D) revealed that MPA inhibited the enlargement of the nuclei and nucleoli of primary B cells after infection in a dosedependent manner. Fig. 4E presents representative 3D images of an immunofluorescence assay of infected B cells that were treated with MPA at 4 dpi (data from 2 dpi are presented in Fig. S3A-F). The DAPI-positive volume (nuclear volume) was increased to 400 mm 3 , which corresponded to the previously analyzed nuclear size, by infection at 4 dpi (  (Fig. 4N), although tRNA synthesis was not affected by MPA due to unknown reasons (Fig. 4M). Western blotting showed that the protein levels of GNL3 and IMPDH2 were increased by EBV infection and decreased by MPA treatment (Fig. 4O).
To further evaluate the effect of IMPDH2 on nucleolar hypertrophy at the cell culture level, we transfected primary B cells with a control siRNA and a siRNA for IMPDH2, and this was followed by EBV infection. After two days, the sizes of nucleoli in the IMPDH2-abated cells appeared significantly smaller ( Fig. 5A; Fig. S4A, C, and D), and the number of nucleoli per cell was less ( Fig. 5B; Fig. S4B) than that observed in the control. The transduction of the CRISPR/Cas9 vector for the IMPDH2 gene also caused shorter major axes ( Fig. 5C; Fig. S4E) and smaller numbers of nucleoli ( Fig. 5D; Fig. S4F).
Next, we examined the effects of an IMPDH inhibitor on B cell growth transformation by EBV. Peripheral blood mononuclear cells (PBMCs) were infected with EBV at a MOI of 0.1, 0.01, or 0.001 in the presence or absence (DMSO) of MPA ( Fig. 6A; Fig. S5). All wells without MPA treatment showed cell clumping, suggesting growth transformation, by days 5 and 12, at a MOI of 0.1 or 0.01, respectively. Continuous MPA treatment, even at the lowest concentration (2.5 mM), blocked clump formation. Similar results were obtained at a MOI of 0.001 (Fig. S5). To test the early effects of MPA after infection, we treated cells with MPA for 5 days. Treatment with 2.5 mM MPA for 5 days that was followed by the removal of the inhibitor caused a slight increase in clump formation at a MOI of 0.1 (Fig. 6A). Therefore, MPA treatment efficiently inhibited in vitro EBV-driven immortalization.
We intraperitoneally injected EBV-infected human cord blood mononuclear cells (CBMCs) into NOD/Shi-scid-IL2Rg null (NOG) mice. The mice were administered daily with mycophenolate mofetil (MMF), a prodrug of MPA, or DMSO orally for four weeks (Fig. 6B). The longevity and body weights of the NOG mice were observed ( Fig. 6C; Fig.  S6A). At the time of sacrifice on day 29, measurements of spleen sizes suggested that MMF inhibited splenomegaly (Fig. 6D). Peripheral blood was collected at days 14 and 21 to analyze EBV DNA. MMF reduced the viral load from the peripheral blood at day 14 (Fig. 6E). The hematoxylin and eosin (H&E) staining of spleen sections (Fig. 6F, upper panels) showed that MMF treatment was associated with reduced mitosis (red arrowheads) and large, atypical lymphocytes with nuclear enlargement and prominent multiple nucleoli (blue arrowheads). Immunohistochemistry showed that the spleens of the DMSO-treated mice were strongly stained with LMP1, whereas MMF treatment suppressed the staining (Fig. 6F, lower panels). The administration of MMF for the first 14 days after injection also improved the survival of the mice ( Fig. S6B and C). MMF is a prodrug of MPA and is clinically used for the anti-rejection of posttransplant patients as well as the mitigation of autoimmune disorders, including systemic lupus erythematosus (SLE) (29). MMF treatment significantly improved mouse survival ( Fig. 6C; Fig.  S6C), indicating that IMPDH inhibitors decreased lymphomagenesis in vivo.

DISCUSSION
Because primary B cells are in the resting (G 0 ) state, EBV alters the gene expression machinery to increase the efficiency of growth transformation (30). This is accomplished partly through epigenetic changes, including CpG demethylation and histone acetylation, which result in the drastic reduction of nuclear heterochromatin (31)(32)(33). EBV infection triggers the formation of superenhancers that induce specific host genes, such as MYC, in which viral gene products, such as EBNA2 and EBNA3, are involved (13,27). Furthermore, previous studies have used gene editing to identify the factors that are essential for EBV growth transformation, such as IRF2, IRF4, BATF, SYK, CFLAR, RBPJ, CCND2, CDK4, CDK6, CD19, CD81, and RelA (34). In addition, studies have found that EBV subverts host mevalonate and fatty acid pathways (35) as well as induces cytidine metabolism through the CTPS1 and CTPS2 genes (36).
In the present study, we found that EBV infections increased the numbers and sizes of nucleoli (Fig. 1C, K, and L). Nucleoli are nuclear bodies that are not surrounded by a membrane and are located at chromosomal regions called nucleolus organizer regions (NORs), which include rRNA gene repeats that encode rRNA sequences (37,38). They are the site of rRNA transcription, rRNA storage, and ribosome assembly. Increases in their sizes and numbers are pathological features of tumor cells, including EBV-associated B cell lymphomas. However, the underlying mechanisms have remained unclear for more than 100 years. In 2019, the rate-limiting enzyme for GTP synthesis, namely, IMPDH2, was identified, and it appears to be the bottleneck in nucleolar enlargement and increased biosynthesis (22,39,40). In a glioblastoma model, an increased GTP level due to IMPDH2 induction in cancer cells caused the transcriptional activation of ribosomal RNAs, thereby increasing ribosomal biogenesis, upregulating cellular genes, and promoting cellular proliferation.
A RNA-seq analysis showed that IMPDH2, which was expressed at low levels in primary B cells, was rapidly induced by EBV infection (Fig. 2A). Indeed, transcriptome data from previous studies have shown that IMPDH2 is induced by the EBV infection of primary B cells by 2 dpi (11,14). In addition, IMPDH2 induction by EBV infection is significantly alleviated by the deletion of the viral EBNA2 gene (Fig. 2J), indicating that EBNA2 is required for IMPDH2 expression immediately after EBV infection. These results are in line with those of a previous study that used EBNA2 that was C-terminally fused to an estrogen receptor; EBNA2 ablation by 4-hydrohytamoxifen depletion decreased IMPDH2 gene expression in LCLs (24). EBNA2 induces the MYC gene (41), which is a transcriptional activator of IMPDH2 (25,42). The expression levels of IMPDH2 in primary lymphocytes and lymphomas are strongly correlated with those of MYC but not EBNA2 ( Fig. 2L; Fig. S2D). In addition, EBNA2 overexpression did not increase the expression levels of IMPDH2 and MYC in P3HR1 cells (Fig. S2E-G). Treatment with the MYC inhibitor attenuated IMPDH2 induction in primary B cells (Fig. S2A), suggesting that MYC, but not EBNA2, is the major transcriptional activator of IMPDH2. However, interestingly, the transient mitogenic activation of primary B cells by CD40L and IL-4 was associated with the increased expression of IMPDH2 but not MYC (Fig. 3). Therefore, CD40L and IL-4 increase IMPDH2 expression independently of MYC. There may be multiple pathways for IMPDH2 gene induction because of its central role in cellular proliferation. Interestingly, although the MYC gene was not increased by CD40L or IL-4 treatment, the MYCL expression was significantly induced (data not shown). Therefore, MYCL may play a significant role in IMPDH2 expression.
The EBV-induced IMPDH2 gene was associated with an increased GTP level (Fig. 2B), which correlated with the biosynthesis of rRNA, tRNA, and the nucleolar protein GNL3 as well as nucleolar hypertrophy by 2 dpi. The sizes of the nuclei and cells were increased by EBV infection slightly after (4 dpi) the increase in nucleolar size (Fig. 1). Several studies have reported efficient cellular proliferation after EBV infection at 4 to 8 dpi (43,44). Therefore, we speculate that rapid cell division requires abundant rRNA, tRNA, nucleolar proteins, and ribosomes that support the efficient de novo synthesis of viral and cellular proteins. Furthermore, the administration of a specific IMPDH inhibitor, namely, MPA, blocked the EBV-induced biosynthesis of GTP, rRNA, tRNA, and GNL3 as well as the enlargement of the nucleolus, nucleus, and cell (Fig. 4). MPA also suppressed in vitro growth transformation by EBV infection (Fig. 6A), which is in line with previous data that show the MPA treatment of LCL causing cell growth inhibition and cell death (45)(46)(47). MMF treatment in a mouse xenograft model delayed tumor formation and prolonged survival ( Fig. 6B-D). This finding is important because MMF is approved for use as an immunosuppressant (29). Our results provide basic evidence that MMF use may prevent PTLD after transplants. In fact, previous studies have found that the use of MMF in transplant patients reduces the risk and incidence of PTLD (48)(49)(50)(51)(52)(53)(54)(55) or, at least, does not increase the risk for PTLD (56)(57)(58)(59). However, MMF increases the incidence of PTLD involving the central nerve system (CNS) (60,61). The underlying reasons for the increased risk for CNS PTLD with MMF use are unclear. However, the transport of MPA through the blood-brain barrier is inefficient because .95% of the MPA, which is the active metabolite of MMF, forms a complex with serum albumin (62,63).
Increased IMPDH2 expression is present in various human cancers and may be related to a poor prognosis. For instance, four cohorts of glioma patients showed that higher IMPDH2 expression was positively correlated with the risk of malignancy, a poor prognosis, and even chemoresistance (22,64). Higher IMPDH2 expression in osteosarcoma is associated with a poor prognosis and even resistance to chemotherapy and radiotherapy (47,65,66). Similar results have been obtained for several other cancers (67)(68)(69)(70)(71)(72)(73)(74)(75). Clinical trials are already underway to examine the anti-cancer effects of MPA and MMF (76). For instance, multiple myeloma cells express higher levels of IMPDH2, and MMF use results in a positive clinical response (77). Therefore, IMPDH inhibitors may be used for several cancers in addition to PTLD.
Taken together, our results suggest that EBV increases IMPDH2 expression and thereby increases the GTP level, leading to nucleolar hypertrophy and the increased biogenesis of ribosomes. This exploitation of host biogenesis by EBV is required for B cell growth transformation. Our study provides basic, mechanistic evidence that the use of MMF in patients with transplants or SLE prevents PTLD. MMF may also be used for other cancers and viral infections, but further studies are required to confirm its usefulness. TEM. The primary B cells that were infected with wild-type EBV were harvested at the indicated time points. The cells were washed with phosphate-buffered saline (PBS) and fixed in half-strength Karnovky's fixative (2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M sodium phosphate [PB]) for 1 h at 4°C. The specimens were washed with PBS twice at room temperature (RT) for 5 min and were postfixed with 2% osmium tetroxide. Following postfixation, the specimens were washed with MQ twice at RT for 5 min, dehydrated in a graded ethanol series (50%, 70%, 80%, 90%, 100%, 100%, and 100% for 5 min each at RT), suspended in propylene oxide (QY-1) for 30 min at RT, and embedded in Epon 812 (Taab Laboratories, Aldermaston, UK) at 60°C. The embedded cells were sectioned using an ultracut microtome. The ultrathin sections were stained with 2% uranyl acetate solution and Sato lead solution. TEM was performed using a JEM-1400 Flash (JEOL, Tokyo, Japan) at an accelerating voltage of 80 kV.

MATERIALS AND METHODS
Immunofluorescence analysis. The primary B cells that were infected with wild-type EBV were harvested at the indicated time points. The cells were fixed with 70% ethanol overnight at 220°C, postfixed with 4% paraformaldehyde at RT for 15 min, washed with wash buffer (0.5% bovine serum albumin in PBS-T) three times, and blocked with 5% bovine serum albumin in PBS-T for 30 min. Then, the cells were stained with the primary antibody overnight at 4°C, and this was followed by staining with the secondary antibody for 1 h at 37°C. The primary and secondary antibodies were diluted with wash buffer. The stained cells were mounted on ProLong Diamond with DAPI (ThermoFisher Scientific, Waltham, MA, USA) for nuclear staining.
Confocal microscopy. Confocal microscopy was performed on a Zeiss LSM-710 system (Carl Zeiss, Jena, Germany). Images were photographed using a Plan-Apochromat 63Â/1.4 lens objective. For the 3D reconstruction, Z-stacks were performed at a Z-step of 0.35 mm.
Microscopic data analysis. To analyze TEM images, the areas of the nuclei and nucleoli were measured using ImageJ software. The nuclear and cellular volumes were measured using Imaris software. Serial sections were observed using a confocal laser scanning microscope (Zeiss LSM-710) and were reconstructed to a 3D isosurface at a fixed threshold. The long axis of the nucleoli was measured using serial images of the sections and the Zen software package.
RNA-seq. Briefly, RNA from primary B cells infected with EBV were collected at the indicated time points, and this was followed by polyA1 RNA isolation and library preparation using a NEBNext Poly(A) mRNA Magnetic Isolation Module and a NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (both from New England Biolab), according to the manufacturer's instructions. The sequencing was carried out by using a HiSeq X next-generation sequencer or a NextSeq 2000 using P2 Reagents (100 cycles) (Illumina).
Quantitative RT-PCR and Western blotting. For the qRT-PCR, total RNA was isolated from cells using TriPure Isolation Reagent (Roche, Rotkreuz, Switzerland). The qRT-PCRs were performed using a One-Step TB Green PrimeScript PLUS RT-PCR Kit (TaKaRa, Shiga, Japan) and a Step One Plus realtime PCR system (ThermoFisher Scientific). The primer sequences used in this study were as follows: The Western blotting was performed as described previously (12).
High-performance liquid chromatography (HPLC) for cellular ATP and GTP. The cell pellets were rinsed with ice-cold PBS, and the supernatant was removed via centrifugation at 16,000 Â g. Then, 1 mL 80% methanol was added, and the solution was mixed. The samples were centrifuged at 16,000 Â g for 10 min, and the supernatant was transferred to a new test tube. The samples were evaporated and resuspended with 100 mL mobile phase A buffer (20 mM phosphate buffer, pH 8.5) containing 10 mM tetrabutylammonium (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan). After centrifugation at 16,000 Â g for 10 min, the supernatant (70 mL) was transferred to a new test tube. Then, 10 mL of supernatant were injected into an HPLC system (Alliance e2695; Waters, Milford, MA, USA) with an XBridge BEH C18 column (3.5 mm; 4.6 Â 50 mm; Waters). The mobile phase B buffer consisted of CH 3 CN. The flow rate was set to 1.0 mL/min. The linear gradient program used was as follows: 5.5 to 6% for 0 to 15 min, 6 to 6.5% for 15 to 20 min, 10% for 20 to 25 min, and 5.5% for 25 to 45 min. The UV detection was set at a wavelength of 254 nm using a 2998 PDA Detector (Waters). Data were collected and processed using the Empower 3 software package for the HPLC system (Waters). The peaks were identified by comparing the retention times and the similarity between the chromatographic peak spectrum with the reference standards of ATP and GTP (Wako). To calculate the concentrations, the measured values were divided by the cell volume.
Ablation of IMPDH2. The siRNA for IMPDH2 or its control, purchased from Santa Cruz Biotechnology (Dallas, TX, USA), were transfected into 1 Â 10 6 primary B cells via electroporation (1,400 V, 20 ms, 2 N). The cells were infected with EBV at a MOI of 3 and were harvested at 2 dpi. The IMPDH2 CRISPR/Cas9 knockout (KO) plasmid and control CRISPR/Cas9 plasmid were purchased from Vector Builder, and 2.5 mg of each plasmid was transfected into Tet-Z/B95-8 cells utilizing lipofectamine 3000. After 5 days, the cells were harvested. Those cells were subjected to fixation and immunofluorescence assays, as mentioned above.
B cell growth transformation assay. An immortalization assay was performed as described previously (89). In brief, PBMCs from healthy donors (LONZA no. CC-2702) were seeded on 96-well plates with cyclosporine in the presence of DMSO or MPA. EBV that was obtained as previously mentioned was subjected to 10-fold serial dilutions and was mixed with PBMCs in the plates. Clump formation was examined under a microscope at the indicated time points.
Mouse xenograft model. The experiments on EBV-associated lymphoproliferative disorder, using a mouse model, were performed as described previously (90). Immunodeficient NOG mice (NOD/Shi-scid, IL-2RgKO) and human CBMCs were obtained from the Central Institute for Experimental Animals and RIKEN, respectively. Human CBMCs were inoculated with EBV in vitro and were intraperitoneally injected into NOG mice. DMSO or MMF (120 mg/kg) was administered orally. Viral DNA in the mouse blood was quantified, and a histological analysis was performed as previously described (91).
Data availability. The RNA-seq data presented in Fig. 2 were obtained from a previous study (12) (DDBJ accession no.: DRA011328). The RNA-seq data for Fig. 2L, Fig. S2D (DRA015073), and Fig. S3 (DRA015092) will be available from DDBJ when the manuscript is accepted for publication. The ChIP-seq and ChIA-PET data were obtained from the publicly available EBV regulome resource tracks (26)(27)(28).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, DOCX file, 13.8 MB.