sT deletion from the MCV genome results in a loss of MCV propagation in transfected cells

Although a co-expression study of MCV sT and LT proteins revealed that sT activates MCV DNA replication through LT stabilization, the breadth of function of native sT expressed from the viral genome remains unknown. Thus, we exploited the native circular MCV genome, which was prepared by re-circularization of the previously reported MCV-HF clone [16]. We generated an sT null mutant (MCVΔsT) by deleting the entire LT intron that contains the sT-specific exon (nucleotide position 430–860, GeneBank accession no. JF813003). We also used a previously reported replication-incompetent MCV (MCV_rep-) with an MCC-derived point mutation in the NCCR [13] (Fig 1A). We anticipated that only sT expression would be ablated in MCVΔsT, since the non-sT functional elements are not mapped within the deleted sT exon.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. MCVΔsT cannot maintain its genome in transfected 293 cells.

(A) MCVΔsT with a deletion of sT-specific exon (dotted line in wild type MCV), wild type MCV (MCV_WT), and MCV_rep- with an NCCR mutation that ablates LT binding to the origin. (B) 293 cells transfected with MCV_WT, MCVΔsT, and MCV_rep- were harvested at various time points up to 8 days p.t. Extracted genomic DNA was digested with EcoRI and DpnI and subjected to Southern blot with an MCV probe. Ethidium bromide (Et-Br) staining was used as a loading control (bottom image). M denotes DNA size marker.

https://doi.org/10.1371/journal.ppat.1011039.g001

First, we examined replication activity by transfecting re-circularized MCV genomes into 293 cells. To avoid the dilution effect of cell passage, the transfected 293 cells were harvested at various time points without cell passage after the initial transfection. We extracted the replicating viral genome and to confirm its size, we used DpnI to remove non-replicating MCV DNA and EcoRI to linearize the full-length genome. The digested genomic DNA was then subjected to a Southern blot using a MCV probe for hybridization (Fig 1B). Both MCVΔsT (~5.0 kbp) and wild type MCV (MCV_WT, ~5.4 kbp) initiated replication and increased the DNA genome by day 2, consistent with successful initiation of viral genome replication from the transfected re-circularized MCV. However, after day 2, MCVΔsT could not maintain its replication activity, and its viral genome copy number decreased over the 8-day experimental time course. By contrast, MCV_WT replicated further, increasing its genome copies in transfected 293 cells (Fig 1B). As expected, no replication activity was detected from MCV_rep-, in which LT cannot bind to the replication origin, rendering it unable to initiate DNA replication. We detected a comparable amount of DpnI-sensitive template DNA on day 1 and day 2 between the three transfected MCV DNAs, confirming equal transfection efficiency. These results suggest a critical role of sT in maintaining the viral genome.

sT deletion attenuates viral early and late mRNA expression, inhibiting capsid protein production

To examine if expressing the sT protein exogenously could rescue MCVΔsT replication, we established 293 TRE-sT cells that inducibly express codon-optimized sT by doxycycline treatment. Cells were seeded in seven dishes and transfected with the MCVΔsT and MCV_WT genomes. We harvested cells from one dish at day 1 post-transfection (p.t.) and added either doxycycline or autoclaved water to the remaining three dishes. To analyze viral growth kinetics relative to day 1, cells with or without sT expression were harvested at days 2, 4, and 6 p.t. for viral DNA, RNA, and protein expression analyses. In the absence of sT expression, MCVΔsT DNA increased moderately by day 2, yet decreased below the starting copy number by day 6 (Fig 2A, left panel), consistent with the Southern blot assay (Fig 1B). However, MCV_WT DNA steadily increased over the course of the 8-day experiment (Fig 2A, left panel). Interestingly, induction of sT expression by doxycycline significantly increased both MCVΔsT and MCV_WT DNA, indicating that sT expression rescued MCVΔsT replication and further activated MCV_WT viral replication (Fig 2A, right panel).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Exogenous sT expression rescues viral DNA replication and mRNA expression in MCVΔsT.

MCV_WT and MCVΔsT viral DNAs were transfected into TRE-sT cells seeded for a 6-day time course experiment. Cells were harvested 24 h p.t. and the remaining cells were treated with or without doxycycline (Dox) at day 1 (0.5 μg/mL) and paired samples were harvested at day 2, day 4, and day 6 p.t.. Time kinetics in DNA replication, gene expression, and protein expression in both MCV_WT (red lines) and MCVΔsT (blue lines) were analyzed. Unpaired T-test was used for statistical analysis. *, P<0.05; **, P<0.01; ***, P<0.001. (A) Extracted genomic DNAs were digested with DpnI and EcoRI and subjected to qPCR using primer sets detecting MCV VP2 and GAPDH genes. Relative MCV DNA copy number to the day 1 sample was determined by the 2^-ΔΔCt method using GAPDH for normalization. (B) RNA was extracted and used to synthesize cDNA for qRT-PCR to determine early and late gene mRNA expression levels using MCV PanT and VP2 (set 1) primer pairs, respectively. Dotted lines represent Dox- condition while solid lines show Dox+ condition. Primers used in each experiment are listed in S1 Table. Relative MCV DNA copy number and early and late gene levels to the day 1 sample were determined by the 2^-ΔΔCt method using 18S RNA for normalization. (C) Protein was extracted and subjected to an immunoblot showing LT, sT, and VP1 expression levels. Immunoblots were stained with LT (CM2B4), sT (2T2) and VP1 (CM9B2) antibodies with Hsp70 antibody as a loading control. Asterisk (*) indicates non-specific band.

https://doi.org/10.1371/journal.ppat.1011039.g002

Using RNA from the same samples, we also examined early and late gene expression using primers targeting the panT (sT and LT) and VP2 transcripts. Both early and late transcripts were significantly lower in MCVΔsT when compared to MCV_WT (Fig 2B, left panels). It is possible that the decrease in early gene expression is due in part to the deletion of the LT intronic sequence that encodes sT [19]. Exogenous sT expression fully activated both early and late gene expression in MCVΔsT and MCV_WT (Fig 2B, right panels). Between days 1 and 6, the magnitude of viral mRNA increase by sT (~1000-fold) was significantly higher than that of viral DNA increase (~10-fold). Immunoblots of these samples further revealed a lack of LT and VP1 protein expression in MCVΔsT (Fig 2C). While LT and VP1 protein expression was detectable after day 2 p.t. in MCV_WT in the absence of sT induction, no protein expression was detected from MCVΔsT. Doxycycline induction of sT rescued LT and VP1 expression in MCVΔsT and further increased protein expression in MCV_WT. These results confirm that sT deletion does not affect LT protein expression. The levels of LT and VP1 protein induced by sT at each time point are nearly identical between MCVΔsT and MCV_WT, coordinating with the corresponding viral early and late gene expression that are fully activated by sT induction (Fig 2B). In the absence of doxycycline, no endogenous sT was detected (Fig 2C). Although LT protein expression decreased at day 6, VP1 protein expression increased in sT-induced MCVΔsT, as well as sT-induced and sT-noninduced MCV_WT transfected cells, consistent with progression of the viral life cycle.

The LTSD domain of MCV sT activates viral gene expression

To determine if LT stabilization is the only mechanism that activates viral replication by sT, we examined if the exogenous expression of LT could rescue MCVΔsT replication. We transfected MCVΔsT in 293 TRE-LT cells that inducibly express codon-optimized MCV LT by doxycycline as well as 293 TRE-sT cells and 293 TRE-Empty (Emp) control cells. Doxycycline was added at day 1 p.t., and samples were collected 4 days p.t. to examine viral replication by Southern blot. Immunoblots confirmed exogenous sT and LT expression induced by doxycycline treatment (Fig 3A, bottom panel). In the Southern blot, sT expression activated MCVΔsT replication substantially, indicating a full rescue by sT (Fig 3A, top panel). LT did not activate MCVΔsT replication to the same extent. These results suggest that LT stabilization is not the sole mechanism by which sT activates viral DNA replication. MCV sT binds to various host factors such as PP2A, Fbw7, and the EP400/NuA4/Tip60 histone acetyltransferase (HAT) complex. The protein domains of sT that are responsible for host protein binding were mapped in previous studies [5,6,20]. We first examined if sT domain mutants could stabilize LT in 293 cells (S1A Fig). When codon-optimized sT and LT expression vectors were co-transfected in 293 cells, sT increased LT protein levels by stabilizing LT [11]. As shown in S1A Fig, the LT protein was stabilized by the codon-optimized sT mutants, which ablate PP2A (sT_R7A, sT_L142A), Hsp70/DnaJ (sT_D44N), and both PP2A/Hsp70 (sT_L142A/D44N) protein interactions. Using an original sT_LTSD mutant (sT_91-95A), we confirmed the importance of the LTSD domain (aa 91–95), which interacts with Fbw7 [11], for LT stabilization. We also developed two additional LTSD mutants that cannot stabilize LT surrounding aa 90–95: sT_90-95A and sT_90-94A (S1A Fig). sT_90-95A has been reported as an EP400/NuA4/Tip60 interacting mutant along with two additional mutants sT_83-88A and sT_4M that have alanine substitution mutations at amino acid residues 83-88A and serine substitution mutations at residues 86, 87, 92, and 93 (4M) [20]. We found that sT_83-88A and sT_4M stabilized LT, whereas sT_90-95A did not (S1A Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Mapping of sT domain required for viral genome maintenance.

(A) MCVΔsT DNA was transfected into 293 TRE-Emp, sT, and LT cells and harvested at day 6 p.t.. gDNA was digested with EcoRI and DpnI and run on a Southern blot showing the effect of exogenous pcDNA sT and LT on MCVΔsT DNA replication. Immunoblot shown below was performed to confirm expression of exogenous sT and LT levels. (B-D) 293 cells were co-transfected with MCV_WT (red) or MCVΔsT (blue) and various pcDNA6 sT mutants including pcDNA Empty (Emp). Cells were then harvested at day 4 p.t.. (B) qPCR was performed as described in Fig 2A legend. To determine the effect of sT, sT mutants, and LT on MCVΔsT and MCV_WT replication, MCV DNA levels, relative to the control sample (MCVΔsT co-transfected with pcDNA Empty), were analyzed by 2^-ΔΔCt method. Solid bar graphs represent viral DNA. Unpaired T-test was used for statistical analysis. Significance was determined in comparison to MCVΔsT co-transfected with pcDNA6 sT_WT. *, P<0.05; **, P<0.01; ***, P<0.001. (C) RNA was extracted and converted to cDNA for qRT-PCR analysis to determine early and late gene expression as described in Fig 2B legend. Relative early and late mRNA expression to the same control in Fig 3B was determined by the 2^-ΔΔCt method. Hatched bar graphs represent viral mRNA expression. (D) Corresponding protein levels shown on an immunoblot stained with MCV proteins (LT, sT, VP1) of each sample. Hsp70 was used as an internal control. A closed arrowhead indicates LT protein, and asterisk (*) designates nonspecific band. (E) 293 cells were transfected with MCT_WT, MCVΔsT, MCV_rep-, and two MCV mutants with different LTSD mutations (MCV.sT_LTSD and MCV.sT_90-94A) and harvested day 4 p.t. for Southern blot using an MCV probe. The same samples were analyzed by qPCR to quantitate MCV DNA levels as in Fig 3B. Relative MCV DNA abundance to MCV_WT is shown. (F) With the same samples used in Fig 3E, total RNA was extracted and converted to cDNA for qRT-PCR analysis of early and late gene expression as in Fig 3C. Relative early and late gene expression to MCV_WT is shown. All error bars indicate standard error of three independent experiments.

https://doi.org/10.1371/journal.ppat.1011039.g003

To determine the sT domain that controls native viral replication, we performed phenotype rescue experiments by co-transfecting MCVΔsT with the aforementioned codon-optimized sT mutants. As shown in Fig 3B, in the empty vector co-transfected cells, replication of MCV_WT was nearly 5.0-fold greater than that of MCVΔsT, as determined by qPCR (Fig 3B). When sT_WT was co-transfected, both MCVΔsT (50-fold) and MCV_WT (30-fold) DNA increased relative to their corresponding empty vector co-transfected cells. This indicates that exogenous sT expression in MCV_WT, regardless of endogenous sT expression, can further increase viral DNA replication. When we co-transfected MCVΔsT with various domain mutants, only three LTSD-related mutants (sT_LTSD, sT_90-95A, and sT_90-94A) failed to rescue MCVΔsT replication, whereas all DnaJ and PP2A binding mutants restored MCVΔsT replication to levels comparable to sT_WT (Fig 3B). These results support the importance of the LTSD for MCV genome replication. Although co-expression of LT increased DNA replication for both MCVΔsT and MCV_WT genomes (7.9- and 12-fold), the levels were once again not comparable to those produced by co-expression of sT (Fig 3B).

We also examined viral early and late gene expression by qRT-PCR (Fig 3C). In the absence of exogenous sT and LT, early and late gene expression from MCV_WT was 12- and 23- fold higher, respectively, than in MCVΔsT. However, the difference in MCV DNA levels between MCVΔsT and MCV_WT was only 5.0-fold. This suggests that the decrease in early and late gene expression induced by sT deletion is not simply a reflection of viral template DNA reduction. Interestingly, rescue of exogenous LT expression increased both early and late gene expression 5.9- and 6.9-fold, respectively, and a corresponding 7.8-fold increase in MCV DNA levels was also seen. This supports the notion that LT activates both early and late gene expression by increasing viral DNA replication. In contrast to LT, sT co-expression markedly increased early and late gene expression 342- and 523-fold, respectively, indicating that sT activates both early and late gene expression by a post-viral DNA replication mechanism. Of the sT mutants, LTSD mutants (sT_LTSD, sT_90-95A, sT_90-94A) failed to rescue early and late gene expression. All other DnaJ and PP2A mutants activated MCVΔsT mutants to a similar degree as sT_WT (Fig 3C). We also tested if SV40 sT (SV40 sT_WT) or a PP2A binding mutant of SV40 sT (SV40 sT_L133A), could rescue transcription in MCVΔsT. However, neither MCV early nor late gene expression could be rescued by SV40 sT (S1B Fig), confirming that MCV sT-mediated activation of viral transcription is independent of PP2A and specific to MCV sT. Therefore, host factors that promote viral replication downstream of sT are likely to be different between MCV and SV40. Immunoblots confirmed strong expression of exogenously expressed sT and LT proteins (Fig 3D). Endogenous expression of sT by MCV_WT, however, was too low to be detected in transfected 293 cells. LT expression by MCV_WT was detected, yet it was significantly lower than exogenous LT. LT expression from MCVΔsT itself was lower than that from MCV_WT and was not detected (Figs 2C and 3D). Co-expression of sT constructs that rescued MCVΔsT replication (sT_WT, sT_D44N, sT_R7A, sT_L142A, and sT_D44N/L142A) induced late VP1 protein expression significantly and increased LT expression minimally as compared to the pronounced LT stabilization seen in S1A Fig. By contrast, LTSD mutants failed to induce LT and VP1 expression (Fig 3D), consistent with attenuated early and late gene induction by the LTSD mutants (<3-fold for early gene and <5-fold for late gene) (Fig 3C). Although the LTSD appears to regulate viral gene expression and replication, expression of LTSD mutant protein levels were significantly lower than wild type sT and other mutants that rescued MCVΔsT transcription and replication (Fig 3D). Thus, we titrated the sT_WT and LTSD mutants’ expression vectors to determine the levels of sT that would rescue the MCVΔsT phenotype. MCVΔsT was co-transfected with different amounts of sT expression vectors to achieve comparable protein expression levels across sT_WT and different sT mutants in 293 cells (S1C Fig). Three LTSD mutants, sT_LTSD, sT_90-95A, and sT_90-94A required 2 ~ 20-fold sT_WT expression vectors for transfection in order to express proteins comparable to sT_WT (S1C Fig). Strikingly, a small amount of sT_WT protein (10 ng of vector co-transfection) was sufficient to rescue LT and VP1 protein expression, whereas higher sT_LTSD, sT_90-94A, and sT_90-95A protein expression (500 ng of expression vector co-transfection) did not rescue LT and VP1 protein expression. We also examined if sT_WT restoration of LT and VP1 protein expression accompanies increased viral gene expression. In line with the protein expression results, a small amount of sT expression fully activated early and late gene expression, while the LTSD mutants could not activate early and late gene expression, regardless of protein levels (S1B Fig). Although endogenous sT protein expression from MCV_WT was undetectable by immunoblot (Figs 2C and 3D), these results suggest that even a small amount of sT_WT can sufficiently activate MCV gene expression and replication, and that the LTSD is required to stimulate viral DNA replication, gene expression, and protein expression.

Since the LTSD may rescue MCVΔsT replication, we examined if the LTSD is the sole domain controlling sT’s genome maintenance function. To investigate this, we introduced two LTSD mutations into the MCV full genome to generate MCV.sT_LTSD and MCV.sT_90-94A, and compared their replication to that of MCVΔsT in transfected 293 cells (Fig 3E). In a Southern blot, replication of the two LTSD mutants, as compared to MCV_WT, was significantly reduced to levels comparable to MCVΔsT (Fig 3E, top panel). qPCR analysis over three replicates of samples also confirmed a similar reduction (70~80%) in viral DNA replication for LTSD mutants and MCVΔsT (Fig 3E, bottom panel). When we examined early and late gene expression with the same samples, early gene expression from the LTSD mutants and MCVΔsT was reduced by ~90% while late gene expression was reduced by ~95% (Fig 3F). A larger reduction in viral gene expression by sT deletion and LTSD mutations suggests that viral gene expression might be sT’s primary regulatory target for MCV genome maintenance.

LT binding to the NCCR is a prerequisite for sT-mediated activation of viral gene expression

To determine how LT function relates to sT function, viral gene expression in two replication-defective mutants, MCV_WTrep- and MCVΔsT_rep-, was compared by qRT-PCR to that in MCV_WT and MCVΔsT in transfected 293 cells. As seen in Fig 3F, sT deletion reduced early and late gene expression by ~90% and ~97%, respectively. Both rep- mutants showed an even greater decrease with no discernible difference between MCV_WTrep- and MCVΔsT_rep-, indicating that the effect of sT was cancelled (Fig 4A). We also examined if sT could rescue viral gene transcription in MCVΔsT_rep- in the absence of LT binding to the viral genome. qRT-PCR demonstrated that although sT co-expression activates early and late gene expression over 100-fold in MCVΔsT (Fig 3C) (~341-fold for early and ~522-fold for late gene expression), a single rep- mutation introduced into the MCVΔsT genome drastically reduced sT’s viral gene activation ability to only a few-fold (Fig 4B) (~2.5-fold for early and ~3.5-fold for late gene expression). While subtle activation was seen in early and late gene expression by sT_WT co-expression (Fig 4B), two LTSD mutants, an original sT_LTSD and sT_90-94A, also increased viral gene expression moderately. These results indicate that LT binding to the NCCR is essential for sT to activate viral gene expression.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. sT requires viral DNA replication for an efficient activation of viral gene expression.

(A) 293 cells transfected with MCV_WT, MCVΔsT, and their rep- counterparts were harvested at day 4. Total extracted RNA was subjected to qRT-PCR analysis by the 2^-ΔΔCt method with 18S ribosomal RNA for normalization. Relative mRNA expression to the MCV_WT-transfected sample is shown. Error bars indicate SD. (B) 293 cells co-transfected with MCVΔsT_rep- and either pcDNA empty (Emp), pcDNA sT_WT, or LTSD mutants (pcDNA sT_90-94A and pcDNA sT_LTSD) were harvested at day 4. Early and late gene expression was determined as in Fig 4A. Relative mRNA expression to the pcDNA6 empty co-transfected sample is shown. (C) 293 TRE-sT cells that can inducibly express sT by doxycycline were transfected with MCVΔsT in 13 dishes. At day 1 p.t., cells in one dish were harvested and doxycycline (0.5 μg/mL), DNA inhibitors (5 μM aphidicolin (APC) and 400 μM mimosine (Mimo)), and water (for mock) were added to the other 12 dishes. The remaining cells were harvested at day 2 and 4 p.t. To examine viral DNA replication, qPCR was performed, as described in the Fig 2A legend. (D) Using the same samples as in Fig 4C, RNA was extracted and converted to cDNA for qRT-PCR analysis, as described in the Fig 2B legend, to determine early and late gene expression.

https://doi.org/10.1371/journal.ppat.1011039.g004

We further confirmed this result by using a reporter MCV that expresses a fluorescent protein under a late promoter, since sT is essential for VP1 protein production. We generated both wild type and ΔsT MCV that encode the ZsGreen (ZsG) fluorescence reporter protein-FMDV 2A(F2A) sequence under the late gene promoter before the VP2 gene (S2A Fig), MCV_WT.L-ZsG and MCVΔsT.L-ZsG). Although ZsG-F2A insertion ablated VP1 protein expression, normal viral replication occurred from the MCV.L-ZsG genome [18]. When transfected into 293 cells, MCVΔsT.L-ZsG showed remarkably lower fluorescence than MCV_WT.L-ZsG, consistent with the significant reduction of late gene expression seen in the native MCV genome (S2A Fig). While ZsG positivity was comparable between MCV_WT.L-ZsG and MCVΔsT.L-ZsG, introduction of the rep- mutation in the two viral genomes (MCV_WT.L-ZsG_rep- and MCVΔsT.L-ZsG_rep-) decreased ZsG positivity, ablating the sT effect as seen in the native MCV genome (S2B Fig). By contrast, the mean fluorescence intensity of ZsG was decreased by both sT deletion and rep- mutation (S2C Fig). MCV.L-ZsG is a useful reporter virus to assess late gene promoter activity, and the results in native MCV and MCV.L-ZsG reporter viruses confirm that LT binding to the viral origin of the NCCR is a prerequisite for sT function.

Next, using general DNA replication inhibitors, we examined if ongoing viral DNA replication, which follows LT binding to the NCCR, is important for sT-induced viral gene expression. MCVΔsT-transfected 293 TRE-sT cells in one of 13 dishes were harvested at day 1 p.t., and the others were treated with or without DNA replication inhibitors aphidicolin and mimosine in the presence or absence of sT induction with doxycycline to be harvested at days 2 and 4 p.t.. Treatment with aphidicolin and mimosine prevented sT-induced viral DNA replication and decreased viral DNA on days 2 and 4 relative to day 1 (Fig 4C). In contrast to this decrease in viral template DNA, early and late mRNA expression continued to increase over time, indicating that, in the absence of sT, viral mRNA expression is independent of viral template DNA. However, sT-mediated increase of viral mRNA expression was significantly attenuated by aphidicolin and mimosine (Fig 4D). In the polyomavirus late gene expression, wraparound transcription that runs through the polyA site and circles the genome multiple times are reported [21–23]. Thus, both positive and negative strand transcripts are expressed from the early gene locus. Since qRT-PCR exploits random hexamers for cDNA synthesis, we used a positive strand-specific primer for cDNA synthesis to detect positive strand early mRNA. Semi-quantitative RT-PCR using the aphidicolin and mock-treated RNA harvested at day 4 p.t. demonstrated that sT induction by doxycycline increases early T antigen mRNA, while aphidicolin ablates the increase without decreasing basal mRNA expression. No signal was seen in the negative control samples that depleted RT in the reaction, confirming that the PCR signals originated from viral mRNA as opposed to transfected template DNA (S2C Fig). These results suggest that sT-mediated viral gene activation requires concurrent viral DNA replication.

MCVΔsT virions cannot maintain replication or induce VP1 protein expression in infected dermal fibroblast cells

Dermal fibroblasts have been shown to be target cells for MCV infection [24]. We examined if sT is also important for MCV genome maintenance and persistence in dermal fibroblast cells that are permissive to MCV infection. In order to produce infectious MCV, MCVΔsT and MCV_WT viral genomes were transfected in 293 TRE-sT cells and cultured for 10 days under doxycycline-induced sT expression [17]. In the final step of virion purification by iodixanol gradient ultracentrifugation, similar amounts of MCVΔsT and MCV_WT virions were sedimented in fractions #11–13 as demonstrated by MCV VP1 immunoblots (Fig 5A). These data suggest that, in the presence of sT expression, MCVΔsT virion production is as efficient as MCV_WT virion production. After the three MCV-positive fractions were combined, a viral genome titer (MCV genome copies/μL) was determined by qPCR for subsequent infection studies.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. sT regulates VP1 expression in infected permissive cells.

(A) Production of MCVΔsT and MCV_WT virions in transfected 293 TRE-sT cells with sT protein induction. Thirteen fractions from MCVΔsT- and MCV_WT-transfected 293 TRE-sT cell lysates sedimented by Opti-prep (1~13 from lighter to heavier) demonstrate similar virion packaging efficiency in fractions 11~13 between MCV_WT and MCVΔsT according to VP1 protein expression and genome copy numbers from the pooled fractions (fractions 11~13). VP1 protein was detected by CM9B2 by immunoblot, and absolute MCV genome copy number was determined by qPCR with a VP2 primer (set 1) pair. *VP1 dimer protein. (B) BJ.hTERT cells were infected with 10⁶ per cell genome copies of MCVΔsT and MCV_WT and harvested over a time period of 15 days after the addition of serum on day 3 post infection (top panel). qPCR was performed on genomic DNA using the 2^-ΔΔCt method with GAPDH for normalization. A relative MCV DNA abundance to the day 3 post-MCV-infected sample indicates the inability of MCVΔsT to maintain the viral genome. Significance was determined by unpaired T test in comparison between MCVΔsT and MCV_WT within the same date point *, P<0.05; **, P<0.01; ***, P<0.001. (C) Exogenous sT rescues VP1 protein expression in MCVΔsT infected cells. BJ.hTERT TRE-Emp and TRE-sT cells infected with MCVΔsT and MCV_WT were treated with doxycycline at day 5 p.t. to induce exogenous sT expression. Cells were harvested at day 8 for immunofluorescence with VP1 antibody. (D) Doxycycline (Dox) rescues VP1 expression in the MCVΔsT mutant to levels greater than that of MCV_WT. VP1 positive cells were counted to determine % positivity.

https://doi.org/10.1371/journal.ppat.1011039.g005

For the infection studies, we used human foreskin BJ fibroblasts immortalized by human telomerase (BJ.hTERT) to avoid the effect of replicative senescence on MCV replication. As reported by Liu et al, 10⁶ copies of MCV_WT and MCVΔsT per cell were infected into BJ.hTERT cells with serum-free DMEM/F12 medium containing the GSK3 inhibitor CHIR99021 and 1 mg/mL collagenase IV for 3 days [17]. Cells were harvested every 3 days up to 15 days post-infection to examine MCV replication. To study MCV replication kinetics, the relative copy number of MCV at day 3 was determined by qPCR. It should be noted that even though MCVΔsT and MCV_WT were infected with the same copy number of MCV, the MCVΔsT DNA copy number was 1.5-fold higher than that of MCV_WT at day 3. The copy number of MCV_WT significantly decreased by 50% between days 3 and 5, indicating that only part of the viral genome initiates replication in infected cells. Despite this initial drop in the genome copy number, the MCV_WT genome was sustained between days 6 and 15 post-infection (Fig 5B). By contrast, the MCVΔsT genome copy number decreased by 80% between days 3 and 6 and continued to decrease until day 15. This indicates that the MCVΔsT genome cannot persist in infected cells. We also examined early and late gene expression in the MCV infected BJ.hTERT cells. In MCV_WT-infected cells, both early and late gene expression was maintained until day 9 post-infection but decreased at day 12. In MCVΔsT cells, however, early and late gene expression was reduced by >95% at day 6 post-infection and continued to decrease over the course of the experiment (S3A Fig).

We also examined if exogenous sT induction could rescue VP1 capsid protein expression in MCVΔsT-infected fibroblast cells by establishing BJ.hTERT TRE-sT cells that inducibly express codon-optimized sT by doxycycline (S3B Fig). BJ.hTERT TRE-sT cells infected with MCVΔsT or MCV_WT were split into two wells and treated with or without doxycycline at day 5 post-infection. Cells were harvested at day 8 and examined for VP1 protein expression by immunofluorescence. In the absence of doxycycline, VP1 positivity was very low (<0.2%) in MCVΔsT-infected BJ.hTERT TRE-sT cells (Fig 5C and 5D). Doxycycline treated TRE-sT cells showed >10-fold VP1 positivity (2.2%), indicating that sT expression induces VP1 protein expression in infected BJ.hTERT cells. However, in the control MCVΔsT-infected BJ.hTERT TRE-Emp cells, VP1 positivity remained low irrespective of doxycycline treatment. Regardless of the cell lines infected, MCV_WT showed nearly 2% VP1 positivity in the presence of doxycycline (Fig 5C and 5D). These data suggest that, even in infected fibroblast cells, sT expression is required for VP1 capsid protein expression.

Proximity biotin-labeling with sT_WT and sT_LTSD

To identify the LTSD-dependent host protein interactions of MCV sT, we exploited a proximity-dependent biotin labeling assay (Bio-ID) [25] by fusing biotin-ligase BirA to the C-terminus of the sT protein. Prior to conducting the Bio-ID study, we confirmed that the C-terminal fusion of biotin-ligase BirA to sT did not alter the ability of sT to rescue MCVΔsT replication. First, we generated doxycycline-inducible 293 stable cells transduced with TRE-sT_WT-BirA, TRE-sT_LTSD-BirA, and TRE-Emp-BirA. MCVΔsT was transfected into the stable cell lines and replicated in the presence or absence of doxycycline. qPCR revealed increased MCV DNA by doxycycline induction of sT_WT-BirA, but not by sT_LTSD-BirA, indicating that BirA fused sT also activates viral replication through an LTSD-dependent mechanism (Fig 6A). BirA fusion, then, does not alter the function of the sT domain. Immunoblots also confirmed host protein biotinylation induced by sT_WT-BirA and sT_LTSD-BirA in doxycycline-treated 293 TRE stable cells in the presence of both doxycycline and biotin (Fig 6B). After purifying the biotinylated protein with streptavidin beads, we observed similar protein banding patterns between sT_WT-BirA and sT_LTSD-BirA, yet some bands were more intensive or specific to sT_WT-BirA (S4A Fig, asterisks). These data suggest that mutating the LTSD may alter the proximity-dependent host protein biotinylation sT-BirA.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Identification of LTSD-dependent MCV sT proxisome by Bio-ID.

(A) sT fusion with biotin-ligase BirA does not alter the function of sT. Prior to Bio-ID analysis, 293 TRE-sT_WT-BirA, 293 TRE-sT_LTSD-BirA, and 293 TRE-Emp-BirA cells were transfected with MCVΔsT, induced by doxycycline and harvested at day 4 p.t. qPCR with VP2 and GAPDH primers confirmed that MCV replication is still activated by sT_WT-BirA, but not by TRE sT_LTSD-BirA. Error bars indicate SD. (B) Host protein biotinylation is induced by sT-BirA fusion proteins only in the presence of biotin. Host cell biotinylation (top panel) and expression of sT-BirA fusion proteins (bottom panel). sT protein expression was confirmed by streptavidin and sT (2T2) immunoblots in sT_WT-BirA, sT_LTSD-BirA, and Emp-BirA cells after doxycycline and biotin treatment. (C) The flow chart of Bio-ID data analysis and Venn diagram summary of the distribution of related proteins identified by Bio-ID. 301 total proteins were identified as candidates for the sT_WT and sT_LTSD proxisome by the Bio-ID study. 57.1% were unique to sT_WT, 6.3% were unique to sT_LTSD, and 36.6% were common to both sT_WT and sT_LTSD. (D) Gene ontology analysis for the sT proxisome identified by Bio-ID. The 172 unique to wild type sT interactors and the 110 interactors common to both wild type and the LTSD mutant as identified by the Bio-ID approach were included in the analysis. Proteins were analyzed using the PANTHER gene ontology software. Interactors were separated into 19 protein classes. (E) Summary of sT interactors identified as chromatin binding and transcription regulator proteins by the PANTHER software to use for further confirmational studies. Bolded proteins indicate proteins unique to sT_WT. Underlined proteins were selected for further confirmational studies.

https://doi.org/10.1371/journal.ppat.1011039.g006

For the Bio-ID study, purified proteins were subjected to LC-MS/MS to identify potential host proteins that localize near sT_WT and sT_LTSD. Proteins identified from the Empty-BirA control were used as a background.

Based on the criteria outlined in Fig 6C, out of a total of 3,239 proteins, 301 proteins were identified as potential sT interactors. 172 (57.1%) proteins were deemed to be specific to sT_WT-BirA, while 19 (6.3%) were specific to sT_LTSD-BirA. Among them, 110 (36.6%) proteins were found in common between sT_WT-BirA and sT_LTSD-BirA. While the majority of sT_LTSD-BirA specific proteins overlapped with sT_WT-BirA (110/129, 85.3%), less than half of the sT_WT-BirA proteins (110/282, 39.0%) were shared with sT_LTSD-BirA, indicating that more proteins were uniquely labeled by wild type sT_WT-BirA in an LTSD-dependent mechanism. Ablating the LTSD domain showed a decrease in the number of protein interactors able to associate with sT. We also identified known MCV sT interactors including PP2A, EP400, Max, and PP4C. However, since PP2A and PP4C proteins were also identified from control samples, they were excluded from the analysis even though greater spectral counts were seen in sT_WT than in the control samples (S1 Dataset).

Next, to explore the cellular functions potentially affected by sT, we performed a Gene Ontology (GO) analysis using the PANTHER software to classify the 301 proteins that were found to have a greater than 5-fold difference in spectral counts between sT and Empty from the Bio-ID study. PANTHER revealed multiple protein classes that were potential interactors of sT (Fig 6D). The top 10 protein classes identified were: nucleic acid metabolism (19.3%), cytoskeleton (14.6%), metabolite interconversion (9.4%), membrane trafficking (8.2%), transcriptional regulation (7.0%), protein-binding activity modulation (6.4%), protein modifying enzyme (6.4%), chromatin binding (5.3%), scaffold/adaptor (5.3%), and translation (5.3%). Within major protein classes, we identified subclasses to determine if specific protein functional groups may be important for the sT phenotype (S4B Fig). Since the four largest protein classes identified by PANTHER comprised greater than 50% of the total proteins identified, we further studied the subclassifications for these groups. For these major protein classes, sub-classification group names are listed with corresponding percentages of total proteins per class to discuss potential cellular functions affected by sT (S4B Fig). Since our results suggest that sT activates early and late viral mRNA expression, sT may interact with gene expression regulators including gene specific transcription factors and chromatin binding proteins (Fig 6E). To verify that there were no false positives among the potential interactors, the spectral count of each was compared to the average spectral counts listed in the CRAPome 2.0 using the REPRINT (resource for evaluation of protein interaction networks) website (S4C Fig). All of the identified transcriptional regulators and chromatin-binding factors demonstrated greater spectral counts than previously reported CRAPome averages, except for MRTFA and MRTFB.

LTSD-dependent nuclear sT interaction with chromatin and transcriptional regulators in MCV replication

From these potential protein classes, we selected chromatin binding proteins and transcriptional regulators to validate the proximity-ligation assay results (Fig 6E). Of the factors unique to sT_WT (Fig 6E, bolded), we investigated transcriptional regulators Cux1 and c-Jun, and chromatin binding proteins BRD7, BRD9, CHAF1A, CBP and DEK (Fig 6E, underscored) by immunoprecipitation assays to study their physical binding. Despite CBP’s similarity to EP300, CBP-specific peptides were only detected in sT_WT-BirA by mass spectrometry. EP300-specific peptides were found in both sT_WT-BirA and the Empty-BirA background, and therefore excluded from our Bio-ID analysis according to our cut-off criteria.

We used doxycycline-inducible 293 stable cell lines that express StrepII-FLAG-tagged empty (SF-empty), C- and N-terminally SF-tagged codon-optimized sT_WT (CSF-sT_WT and NSF-sT_WT), and N-terminally tagged sT_LTSD (NSF-sT_LTSD) to perform immunoprecipitation (IP). CSF-sT_LTSD was not used because its protein expression was remarkably lower than that of NSF-sT_LTSD. We preliminarily tested the potential interactors using CSF- and NSF-sT_WT (S5A Fig), and if an interaction was detected, we used NSF-sT_WT and NSF-sT_LTSD to confirm LTSD-dependent interaction. The FLAG antibody, which immunoprecipitated similar amounts of NSF-sT_WT and NSF-sT_LTSD, as well as sT’s known interactor, PP2Ac [11] (Fig 7A, bottom panel), co-precipitated Cux1, c-Jun, CBP, and BRD9 proteins robustly with NSF-sT_WT, but faintly with NSF-sT_LTSD. These results suggest an LTSD-dependent interaction between sT and these four proteins. Cux1 is unique because it possesses multiple isoforms through proteolytic cleavage and differential splicing [26,27]. The Cux1 polyclonal antibody that recognizes the N-terminus of Cux1 protein detected full-length 200 kDa (p200-Cux1) and previously defined CASP (CDP/cut alternatively spliced product, 70kDa) isoform, an alternatively spliced product of Cux1 that localized at the Golgi apparatus [28]. Despite the background interaction with SF-empty, the p200-Cux1 repeatedly demonstrated a stronger interaction with NSF-sT_WT when compared to NSF-sT_LTSD, similar to other interactors. However, for CASP, a similar level of protein was co-precipitated with NSF-sT_WT and NSF-sT_LTSD. PP2Ac was used as a positive control as it has been shown to interact with MCV sT [6].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Validation of sT protein interaction with chromatin binding and gene-specific transcription factor proteins identified by Bio-ID.

(A) Confirmation of LTSD-dependent direct sT_WT interaction with Cux1, c-Jun, CBP, and BRD9 by immunoprecipitation (IP). N-terminally SF-tagged (NSF) NSF-sT_WT and NSF-sT_LTSD cell lysates were immunoprecipitated with FLAG Ab. 2.5% of IP lysate was used as input. SF-sT was pulled down to confirm equal precipitation of sT_WT and sT_LTSD. PP2Ac, which interacts with both sT_WT and sT_LTSD, was used as a positive control. (B) MCVΔsT transfected 293 TRE NSF-sT_WT and NSF-sT_LTSD cells were treated with Dox at day 1 p.t. and harvested day 4 p.t. for immunofluorescence. Cells were co-stained with antibodies against FLAG sT (red, Alexa 488) and either Cux1, c-Jun, or CBP (green, Alexa 568), and images were captured by confocal microscopy. Magnification 100x. Arrowheads designate sT-positive mitotic cells. White bars on DAPI images indicate 10 μm.

https://doi.org/10.1371/journal.ppat.1011039.g007

Since NSF-sT_LTSD was weakly co-precipitated with potential interactors, we sought to further confirm sT’s LTSD-dependent interaction with Cux1, c-Jun, and CBP by immunofluorescence (IF). BRD9 was excluded from this experiment since no IF antibodies are available. To study their subcellular localization, we co-stained MCV sT and MCV LT with Cux1, c-Jun, and CBP in MCVΔsT-transfected 293 TRE-NSF-sT cells harvested at day 4 p.t., and Z-stack scanning was performed by confocal microscopy. As shown in previous reports (5, 24), NSF-sT_WT predominantly localized in the nucleus (Fig 7Bd, 7Bl, and 7Bt). By contrast, NSF-sT_LTSD consistently showed more predominant cytosolic staining (Fig 7Bh, 7Bp, and 7Bx). These results suggest that the LTSD domain potentially regulates nuclear sT protein expression by affecting nuclear protein stability or nuclear import. All three transcriptional regulators predominantly localized in the nucleus and showed a partial nuclear co-localization with NSF-sT_WT (Fig 7Bd, 7BI, and 7Bt), showing strong Pearson’s correlation coefficient (r >0.5) (S5B Fig). Due to low nuclear expression, NSF-sT_LTSD protein showed no co-localization with the c-Jun or CBP, but Cux1 protein co-localization was seen in the cytosol of mitotic cells (Fig 7Bh). This phenotype was not observed with c-Jun or CBP (Fig 7Bp and 7Bx), consistent with the IP results in which CASP was the only interactor to both sT_WT and sT_LTSD (Fig 7A). CBP and c-Jun localized in the cytosol in mitotic cells, but no noticeable co-localization with NSF-sT_LTSD was observed in these areas (Fig 7Bm-p and 7Bu-x). Using LT staining, we also examined if these interactors co-localize with replication foci in the nucleus (25). Replication foci were detected in the MCVΔsT-transfected 293 NSF-sT_WT cells, but not in the MCVΔsT-transfected NSF-sT_LTSD cells except for a diffuse nuclear signal (S5C Fig). All three proteins co-localized with LT replication foci. Curiously, Cux1 and c-Jun protein showed distinctive accumulation at the LT replication foci in addition to diffused nuclear expression (S5C Fig), suggesting an accumulation of Cux1 and c-Jun at the site of viral DNA replication and transcription. Both the IP and IF analyses validated our Bio-ID study, showing that four out of the seven candidate proteins analyzed interact with sT in the nucleus, and that LTSD mutation suppresses these interactions by reducing sT protein expression in the nucleus.

Role of sT-interacting transcription regulators in MCV gene expression

To study the importance of the identified sT interactors in MCV transcription, we knocked down the four aforementioned transcriptional regulators and examined their effect on viral early and late gene expression. MCV_WT was co-transfected with two siRNAs, each with different targeting sequences, that knock down either Cux1, c-Jun, CBP, or BRD9 in 293 cells. Early and late gene expression was evaluated at day 4 p.t.. We achieved over 80% knockdown efficiency as quantitated by LI-COR immunoblot except for siCux1.1 (77%), sic-Jun.2 (63%), and siCBP.1(53%), which partially inhibited target protein expression (S6A Fig). Knockdown with siCux1.1, siCux1.2, and sic-Jun.1 increased early (4.3-, 4.9-, 8.4-fold, respectively) and late gene (4.8-, 6.1-, 8.7-fold, respectively) expression, suggesting that these transcriptional regulators may be negative regulators of viral transcription (Fig 8A). sic-Jun.2 showed a moderate increase in early (1.2-fold) and late gene (1.3-fold) expression, consistent with lower knockdown efficiency. On the other hand, knockdown with siCBP.2 and siBRD9.2 decreased early (0.6- and 0.4-fold) and late (0.7- and 0.5-fold) gene expression. We did not observe significant effects in siCBP.1 due to low knockdown efficiency. These results suggest that CBP1 and BRD9 may act as transcriptional activators (Fig 8A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Role of sT-interacting proteins in MCV transcription and viral replication.

(A) Effect of siRNA knockdown of sT interactors on viral gene expression in MCV_WT. siRNA treatment of MCV_WT samples. In cells treated with siCux1.1, siCux1.2, sic-Jun.1, and sic-Jun.2, early and late gene expression increases relative to the siCtrl. In cells treated with siCBP.2 and siBRD9.2, early and late gene expression decreases. Total RNA was subjected to qRT-PCR analysis by the 2^-ΔΔCt method using VP2 (set 1) and PanT primer pairs with 18S ribosomal RNA for normalization. Error bars indicate SD. Unpaired T-test was used for statistical analysis to determine the significance relative to siCtrl. *, P<0.05; **, P<0.01; ***, P<0.001. (B) A-485 and I-BRD9 treatment of MCVΔsT transfected with empty or sT. A-485 treatment significantly decreases early and late gene expression at 1.0, 5.0, and 10.0 μM in MCVΔsT transfected with sT. I-BRD9 treatment at 0.5 and 2.0 μM also significantly decreases early and late gene expression. Total extracted RNA was subjected to qRT-PCR analysis as described in Fig 8A. Relative mRNA expression to the sT_WT-transfected sample is shown. Error bars indicate SD. Unpaired T-test was used for statistical analysis. *, P<0.05; **, P<0.01; ***, P<0.001.

https://doi.org/10.1371/journal.ppat.1011039.g008

Since siRNA knockdown was only partial and knockdown phenotype varied between the two siRNAs used for CBP and BRD9, we next tested if small molecule inhibitors that target CBP and BRD9 would affect sT-induced MCV transcription. For CBP inhibition, we used A-485, a potent and specific CBP/P300 histone acetyltransferase inhibitor shown to inhibit H3K18 acetylation [29,30]. To examine the effect of the inhibitor on sT-induced viral transcription, 293 cells co-transfected with MCVΔsT and sT_WT or an empty vector were treated with A-485 for 4 days p.t.. A-485 treatment significantly reduced sT-induced viral gene expression in a dose-dependent fashion while, in the absence of sT, no dose-dependent effect was seen despite less reduction (Fig 8B). At the highest dosage, 10 μM, early and late gene activation induced by sT (633.5-fold and 885-fold, respectively) from the MCVΔsT genome was attenuated by >97% (12.2-fold and 25.9-fold). For BRD9 inhibition, we used I-BRD9, a structure-based inhibitor that specifically binds BRD9. At concentrations of 0.5 and 2.0 μM, I-BRD9 led to decreased panT and VP2 expression by approximately 25% in MCVΔsT cells transfected with sT_WT, yet the magnitude of reduction was not as large as that seen with A-485 treatment (Fig 8B). We also tested the effects of A-485 and I-BRD9 on MCV DNA replication by MCV origin replicon assays (S6B and S6C Fig). sT-mediated replication activation by LT stabilization was not inhibited by either A-485 or I-BRD9, indicating that these inhibitors do not affect LT stabilization by sT and DNA replication from the viral origin. I-BRD9 demonstrated a dose-dependent increase in sT-induced replication activity up to 2.7-fold at 2.0 μM. Therefore, the significant reduction in early and late gene expression following inhibition of CBP and the modest reduction following inhibition of BRD9 strongly suggest that both CBP and BRD9 interact with sT to regulate MCV gene expression, but not DNA replication.

Since CBP inhibition substantially decreased MCV early and late gene expression, we next tested the effect of inhibiting CBP on early gene expression in four MCV-positive MCC cell lines (Fig 9A). After treatment with 1.0, 5.0, and 10.0 μM of A-485, panT expression significantly decreased in MS-1, MKL-1, and CVG-1 cell lines relative to their respective mock-treated samples. At 10 μM, panT expression decreased by approximately 98% in MS-1, 98% in MKL-1, and 95% in CVG-1. In the BroLi cell line, a significant decrease in early gene expression occurred beginning at 5 μM, as compared to 1 μM in the other three cell lines, suggesting that BroLi may be more resistant to A-485 treatment. Corresponding immunoblots further confirmed decreased LT protein expression with increasing A-485 drug concentration in all four cell lines, further suggesting dose-dependency (Fig 9B). Total H3 and H3k18ac were also measured as a control for A-485-mediated CBP inhibition since H3K18 is one of the major CBP/P300 acetylation sites [31]. H3K18ac protein expression decreased with increasing A-485 concentration, while total H3 expression remained constant, which confirms dose-dependent inhibition of CBP/P300.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. CBP regulates T antigen expression in MCV-positive MCC.

(A) CBP/p300 inhibitor, A-485, significantly decreases early gene expression. MCV-positive CVG-1, MKL-1, MS-1, and BroLi cell lines were treated with various concentrations of A-485 and harvested at day 4. Total extracted RNA was subjected to qRT-PCR analysis by the 2^-ΔΔCt method using the PanT primer with 18S ribosomal RNA for normalization. Relative mRNA expression to the mock-treated sample is shown for each respective cell line. Error bars indicate SD. Unpaired T-test was used for statistical analysis. *, P<0.05; **, P<0.01; ***, P<0.001. (B) Corresponding immunoblots for qRT-PCR shown in Fig 9A. MCV LT was detected by CM2B4. H3K18ac and total H3 were detected to confirm A-485 activity. HSP70 was used as a loading control.

https://doi.org/10.1371/journal.ppat.1011039.g009

EP400/NuA4/Tip60 histone acetyltransferase complex, a known interactor of sT, is another epigenetic transcriptional regulator that may contribute to viral transcription and replication. To test this possibility, we performed an MCV origin replication assay with EP400 and three LTSD mutants to test their MCV origin replication activities (S7A Fig). The sT_90-95A mutant that was originally identified as an EP400-binding mutant did not activate MCV origin replication, similar to sT_LTSD and sT_90-94A. However, two other EP400 mutants, 83-88A and 4M, stabilized LT and activated MCV origin replication. An MCVΔsT replication rescue assay with co-expression of the EP400 mutants demonstrated that, like sT_LTSD, the sT_90-95A mutant failed to restore MCVΔsT replication, consistent with continued LTSD control of MCV replication. On the other hand, the sT_83-88A and sT_4M mutants rescued replication (S7B Fig). As demonstrated by Southern blot (S7C Fig), when we transfected the MCV genome with EP400 binding mutations, the MCV_sT90-95A mutant showed reduced replication like that seen in MCVΔsT. Replication of MCV_sT4M was also moderately decreased. We also examined the importance of EP400/Tip60 in MCV mRNA expression by knocking down EP400 in MCV_WT transfected cells. Contrary to our prediction, shRNA knockdown of EP400 did not inhibit early and late gene expression, but rather increased expression (S6D and S6E Fig). These results are consistent with the fact that EP400/Tip60 targeting of sT inhibits MCV replication and gene expression despite its important role in sT’s transformation activity [20].

Cell lines, cell culture, and transfections

Cell lines that were used include: HEK293, BJ.hTERT, MKL-1, MS-1, CVG-1, and BroLi. HEK293 and BJ.hTERT cells were cultured in Dulbecco’s modified Eagle’s Medium (DMEM) while MKL-1, MS-1, CVG-1, and BroLi cells were cultured in RPMI media. All media was supplemented with 10% Fetal Bovine Serum (FBS) and 5 mL of Non-Essential Amino Acids (NEAA). Cells were kept at 37°C and 5% CO₂ in an incubator. Cells were tested for mycoplasma throughout the experiments to confirm negativity. For plasmid DNA transfections, 293 cells/293 TRE cell lines were transfected at 40–50% density using Lipofectamine 2000 (Invitrogen). To screen active siRNAs, Lipofectamine RNAiMAX transfection (Invitrogen) was used. All transfection reagents were used according to manufacturer’s instructions.

Plasmids

A full-length wild type MCV-HF genome (GeneBank accession no. JF813003) and MCV-HF with a rep- mutation (cytidine 44 to adenine in JF813003) were cloned into pSMART-LC-Amp cloning vector (Lucigen) to generate pSMART MCV-HF and pSMART MCV-HF_rep-. This was done using an EcoRI site that digests once in the MCV genome as described previously [16,18]. To construct sT mutants on pSMART MCV-HF (ΔsT, MCVsT_83-88A, MCVsT_90-95A, MCVsT_4M, MCVsT_LTSD, MCVsT_90-94A), we used overlapping PCR. First, PCR was performed using AvrII.MCV.NCR.F with a mutagenic reverse primer, and a corresponding mutagenic forward primer with BamHI.R [S2 Table]. The two PCR products containing overlapping sequences were purified by PCR/Gel purification columns (Thomas Scientific) and overlapping PCR was performed with AvrII.MCV.NCR.F and BamHI.R primers containing AvrII and BamHI restriction sites [S2 Table]. The PCR-ligated product was then cloned into pSMART MCV-HF using the unique AvrII and BamHI sites.

Using a similar strategy, MCV encoding ZsGreen (ZsG) under the late promoter was generated by inserting a ZsG-FMDV 2A sequence before the VP2 initiation codon (S2A Fig). A ZsG-FMDV 2A fragment synthesized ZsG cDNA (IDT) using VP2.NCR.ZsGreen.F and 2A.VP2.R. Two MCV fragments were also amplified from pSMART MCV-HF using AvrII.MCV.NCR.R and VP2.NCR.R, or 2A.VP2.F and VP2.PacI.R [S2 Table]. Three overlapping PCR fragments mixed with equal molar ratios were fused by PCR using AvrII.MCV.NCR.R and VP2.PacI.R. The PCR-ligated fragment was then cloned into pSMART MCV-HF, pSMART MCV-HF_rep-, pSMART MCVΔsT, and pSMART MCVΔsT_rep- with unique AvrII and PacI sites. pcDNA MCV LT, a codon-optimized MCV LT expression vector, and codon-optimized MCV sT expression vectors pcDNA6 MCV sT_WT, pcDNA6 MCV sT_LTSD, pcDNA6 MCV sT_D44N, pcDNA6 MCV sT_R7A, pcDNA6 MCV sT_L142A, pcDNA6 MCV sT_D44N, pcDNA6 SV40 sT_WT, and pcDNA6 SV40 sT_L133A were previously generated [5,6]. Additional mutants including pcDNA6 MCV sT_83-88A, pcDNA6 MCV sT_4M, pcDNA6 MCV sT_90-95A, and pcDNA6 MCV sT_90-94A were generated by the QuickChange site-directed mutagenesis kit (Agilent) using primers in S2 Table.

Lentivirus constructs that inducibly express codon-optimized sT (pLenti TRE-MCV sT) were described previously [54]. To generate an inducible lentivirus vector expressing MCV LT, a codon-optimized LT fragment digested from pcDNA6 MCV LT with KpnI and XhoI was cloned into pENTR1A. The resulting pENTR1A MCV LT was then LR-recombined into pTRE.DEST EF.Puro-rTet [55] to generate pLenti TRE-MCV LT by LR clonase (Invitrogen).

For the proximity ligation assay, we constructed a parental lentivirus vector that inducibly expresses a C-terminally BirA-tagged fusion protein. BirA (R118G) cDNA was amplified by primers containing KpnI.linker.BirA.F and BamHI.BirA.R sites [S2 Table], and the resulting fragment was cloned into the pLenti TRE-MCS EF.Puro-2A-rTet vector [54] with KpnI and BamHI restriction sites to generate pLenti TRE-MCS-BirA EF.Puro (Emp-BirA). Codon-optimized sT_WT and the sT_LTSD mutant (91-95A) fragments were amplified from pcDNA6 MCV sT_WT and pcDNA6 MCV sT_LTSD using AfeI.sTco.F and KpnI.sTco.R [S2 Table], and then cloned into pLenti TRE-MCS-BirA EF.Puro using AfeI and KpnI sites to produce pLenti TRE-sT_WT-BirA EF.Puro (sT_WT-BirA) and pLenti TRE-sT_LTSD-BirA EF.Puro (sT_LTSD-BirA), respectively.

To construct pcDNA3 vectors that express codon-optimized sT fused to the C- and N-terminus Strep-II-FLAG epitope (SF) tag, pcDNA3 CSF-empty and pcDNA3 NSF-empty vectors were used [56]. Codon-optimized sT cDNA fragments were amplified by two primer sets containing NheI and XhoI restriction sites for the NSF tag, and KpnI and NotI restriction sites for the CSF tag [S2 Table] from pcDNA6 MCV sT_WT and pcDNA6 MCV sT_LTSD. Next, the wild type and LTSD mutant sT cDNA were cloned into pcDNA3 CSF-empty and pcDNA3 NSF-empty vectors using corresponding unique sites (NheI and XhoI or KpnI and NotI). CSF-empty, CSF-sT and NSF-sT fragments were then transferred to the pENTR 1A vector using unique KpnI and XhoI sites. The resulting pENTR 1A vectors encoding CSF-empty, CSF-sT_WT, CSF-sT_LTSD, NSF-sT_WT, NSF-sT_LTSD were LR-recombined into pTRE.DEST EF.Puro-rTet to generate pLenti TRE-CSF-empty, pLenti TRE-CSF-sT_WT, pLenti TRE-CSF-sT_LTSD, pLenti TRE-NSF-sT_WT, pLenti TRE-NSF-sT_LTSD by LR cloning (Invitrogen). All sequences were confirmed by Sanger sequencing (MCLAB).

Antibodies

Primary antibodies include: Cux1 (11733-1-AP, Proteintech), BRD9 (24785-1-AP, Proteintech), c-Jun (24909-1-AP, Proteintech), CBP (7389S, Cell Signaling), DYKDDDDK Tag (FLAG) (A00187S, Genscript), Total H3 (14269, Cell Signaling), H3K18Ac (13998, Cell Signaling), Anti-SV40 T antigen PAb419 (sc-58665, Santa Cruz), and Hsp70 (sc-7298, Santa Cruz). Primary antibodies against MCV LT (CM2B4, VP1, CM9B2) were gifted from Dr. Patrick Moore and Dr. Yuan Chang’s laboratory. 2T2 was gifted from Dr. Christopher Buck’s laboratory. Secondary antibodies for immunoblots include: goat anti-rabbit IRDye 800CW (926–32211, LI-COR), goat anti-mouse IRDye 800CW (926–32210, LI-COR), IRDye 800CW Streptavidin (926–32230, LI-COR). Secondary antibodies for immunofluorescence include: Alexa488 anti-mouse (Invitrogen), Alexa568 anti-mouse (Invitrogen), Alexa488 anti-rabbit (A11034, Invitrogen), Alexa568 anti-rabbit (A11036, Life Technologies).

Genomic and episomal DNA preparation

Cell pellets were suspended overnight in lysis buffer (10 mM Tris, 100 mM NaCl, 0.5% SDS, 25 mM EDTA, pH 8.0) with 100 μg/mL of proteinase K at 37°C. 100 μg/mL RNase was then added for 1 h at 37°C. After lysis, genomic DNA was acquired by phenol chloroform and ethanol precipitation. To extract episomally replicating DNA, transfected cells were harvested and lysed in 600 μL of 0.6% SDS, 10 mM Tris, 1 mM EDTA, pH 8.0. 100 μL of 5M NaCl was added to 400 μL of the lysate and incubated overnight at 4°C. Samples were then centrifuged at 13,000 rpm for 30 min at 4°C to separate the supernatant from the insoluble fraction. The supernatant was subsequently removed and subjected to phenol chloroform and ethanol precipitation to acquire episomal DNA.

Re-circularization of MCV-HF clone

Wild type and mutant pSMART MCV-HF plasmids were digested with EcoRI overnight at 37°C followed by EcoRV for 5 h at 37°C. This reaction produces a full-length linear MCV-HF genome with self-ligatable EcoRI DNA ends and a pSMART-LC-Amp vector fragment with EcoRV ends. The digested sample was purified using plasmid purification columns (Thomas Scientific) to acquire linearized MCV DNA. Re-circularization of the MCV-HF genome was performed by the T4 DNA ligase system (NEB) overnight at 16°C. To prevent concatenation of the MCV genome, the ligation reaction was performed at a lower DNA concentration (2.5 ng/μl) overnight. The re-circularized MCV genome was purified using the same plasmid purification columns. DNA eluted with 10 mM Tris pH 8.0 was used directly for transfection experiments without removing the pSMART vector fragment.

Total RNA extraction, DNase treatment, and cDNA synthesis

Total RNA was extracted by the JET RNA Purification Kit (Thermo Scientific). Acquired RNA was then treated with the Invitrogen TURBO DNA-free Kit (Invitrogen). Following DNase treatment, the RNA was converted to cDNA by random hexamers using the iScript cDNA synthesis Kit (Bio-Rad). To detect the positive strand of the MCV early gene, cDNA was synthesized by a T antigen specific primer (Pan T2) using the Super Script IV cDNA synthesis kit (Invitrogen). To confirm the amount of input RNA, 18S rRNA expression was detected from the cDNA synthesized by random hexamers.

Quantitative PCR

Genomic DNA was digested with DpnI at 37°C overnight and purified using plasmid purification columns to remove non-replicating DNA (Thomas Scientific). 20 ng of DNA per sample was used for qPCR using the PowerUp SYBR Green Master Mix Kit (Applied Biosystems). For qRT-PCR, cDNA reverse transcribed from 25 ng equivalent of total RNA was used for each sample and cDNA reverse transcribed from 10 ng equivalent of total RNA was used for the 18S control. qPCR was performed using QuantStudio 3 (Applied Biosystems). Data was obtained using the ΔΔCT method. Primers used for qPCR and qRT-PCR are found in S1 Table.

Southern blot hybridization

gDNA was digested with EcoRI and DpnI overnight. 5–8 μg of DNA was loaded in a 0.8% agarose gel. The DNA ladder used was a 1kb Plus DNA ladder (NEB, N3200S). Equal loading and full digestion of DNA was checked with ethidium bromide staining. DNA was then transferred with 10X SCC onto a nitrocellulose membrane overnight and DNA was then cross-linked onto the membrane. To generate an MCV probe, MCV-HF genome that was excised from pSMART MCV-HF with BsrFI was biotinylated by using the BioPrime Array CGH kit (Invitrogen). The membrane was hybridized with the MCV probe at 65°C overnight in a hybridization buffer (5X SSPE, 2% SDS, 10% dextran sulfate, 1X Denhardt’s Solution, 10 μg/mL sonicated salmon sperm DNA). Hybridization signal was detected by Streptavidin-IRDye 800CW (LI-COR) and replication was detected by LI-COR imager.

Immunoblotting

Harvested cells from each experiment were suspended in RIPA buffer with a proteinase inhibitor cocktail (Roche Complete Tablet) and sonicated until cell membranes were fully disrupted to release protein. The protein was then quantified using BSA (bovine serum albumin) standards on a Biotek Synergy 2 machine. After quantification, samples were mixed with NuPAGE LDS Sample Buffer 4X (Invitrogen) to make a 1X concentration with 5% beta-mercaptoethanol. Protein was loaded in an SDS-PAGE gel with percentages adjusted to detect the target protein of interest. Precision Plus Protein Kaleidoscope (Bio-Rad) was used as a ladder for all immunoblots. After running the gel, protein was transferred onto a nitrocellulose membrane overnight using a wet transfer buffer at 30 V. The membrane was blocked for 30 min in 5% skim milk in TBS at room temperature. Primary antibodies were diluted either in 5% BSA or 5% skim milk in TBS and incubated overnight at 4°C, or 5 h at room temperature. Membranes were washed 3 times in TBS-T. Secondary antibodies (anti-mouse or anti-rabbit) were diluted in 5% skim milk in TBS and membranes were incubated for 1 h at room temperature in the absence of light. Membranes were washed again in TBS-T, and lastly imaged on a LI-COR imager for analysis.

Proximity-based biotin labelling and purification of biotinylated proteins

293 TRE-Emp-BirA, sT_WT-BirA, and sT_LTSD-BirA cells were seeded in five 14.5 cm dishes at 5x10⁶ cells per dish. Five μg of MCVΔsT re-circularized DNA was transfected per dish 24 h after seeding by Lipofectamine 2000. Cells were supplemented with doxycycline 24 h p.t. at 0.5 μg/mL. At 72 h p.t. doxycycline was refreshed, and biotin was added at 50 μM. 24 h following biotin addition, cells were washed in PBS, resuspended in SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris HCl, 1 mM PMSF, pH 8.0) with protease inhibitor cocktail (Roche), and sonicated to acquire the soluble protein fraction. Total biotinylated proteins from sT_WT-BirA, sT_LTSD-BirA, and Empty-BirA were purified by Dynabeads MyOne Streptavidin C1 (Invitrogen) for mass spectrometry.

Mass spectrometry

Mass spectrometry protocol for Bio-ID experiments was previously described [57]. Briefly, samples were heated at 85°C for 5 min, separated ~1.5 cm on a 10% Bis-Tris Novex mini-gel (Invitrogen) with the MES buffer system, and stained with coomassie. This process separated lanes into 10 fragments of the same size, which were then reduced using dithiothreitol, alkylated with iodoacetamide, and digested with trypsin (Promega, Wisconsin). The digested samples were then subjected to Nano LC-MS/MS with a NanoAcquity HPLC system (Waters, Massachusetts) interfaced to a Fusion Lumos tandem mass spectrometer (ThermoFisher, California) using a 3 s cycle. Using a trapping column with Luna C18 resin (Phenomenex, California), peptides were eluted at a rate of 350 nL/min. A 30 min gradient was employed for each segment. Product ion data were searched against the combined forward and reverse Swissprot H. sapiens protein database using a locally stored copy of the Mascot search engine v2.6 (Matrix Science, London, U.K.) via Mascot Daemon v2.6. Proteome Discoverer v2.1 (ThermoFisher) was used to make peak lists. Search parameters were precursor mass tolerance 10 ppm, product ion mass tolerance 0.02 Da, 2 missed cleavages allowed, fully tryptic peptides only, fixed modification of carbamidomethyl cysteine, variable modifications of oxidized methionine, protein N-terminal acetylation and pyro-glutamic acid on N-terminal glutamine. Scaffold software v4.8 (Proteome Software) was used to read flat files (DAT) that were generated by Mascot search with a filter for 1% protein and peptide level false discovery rate and for detection of at least two unique peptides for each protein. Proteins with at least a 5-fold difference in spectral counts between the control Emp-BirA sample and either sT_WT-BirA or sT_LTSD-BirA were selected as proteins proximity biotinylated by either sT_WT or sT_LTSD, respectively. Proteins were classified as common to both sT_WT and sT_LTSD if both had a greater than 5-fold difference in spectral counts compared to Emp-BirA, but spectral counts between sT_WT and sT_LTSD were less than 5-fold different [58].

Immunoprecipitation

293 TRE-SF-Emp, CSF- and NSF-sT_WT, and NSF-sT_LTSD cells were seeded in 14.5 cm dishes and left to adhere for 24 h prior to doxycycline treatment to induce sT expression. After 72 h, cells were harvested from each plate and split into two pellets per dish. Protein was extracted using IP lysis buffer (25 mM Tris, 137 mM NaCl, 2 mM EDTA, 5% glycerol, 1% Triton X-100, pH 7.5) containing 0.5X protease inhibitor cocktail (Roche), 10mM NaF, and 10mM NaVO₃. The pellet was rotated in 1 mL of IP lysis buffer for 10 min at 4°C before being re-pelleted, and the supernatant was taken for immunoprecipitation. 800 μL of the supernatant was used for IP while the remaining 200 μL was taken for input. Anti-FLAG antibody was added at a 2 μg/mL concentration rotating overnight at 4°C. 24 h following antibody addition, Dynabeads G (Invitrogen) were added, and samples were again rotated at 4°C for 90 min. Following the antibody reaction, beads were washed 4 times with IP lysis buffer containing 0.5 mM PMSF. The third wash contained 150 mM NaCl, in place of 137 mM, to decrease the background of immunoblots. After washing, 20 μL of 2X LDS sample buffer containing 10% 2-ME was added to each sample and denatured at 70°C for 10 min prior to loading. We used 2.5% of the total lysate for input. Samples were subjected to immunoblots.

Immunofluorescence

BJ.hTERT TRE-Emp and TRE-sT cells infected with MCV_WT and MCV ΔsT were fixed with 10% buffered formalin for 8 min, permeabilized with 1X PBS containing 0.1% Triton X-100, blocked with 5% normal goat serum (NGS), and stained with anti-VP1 antibody (CM9B2, 1:2 dilution) for 3 h at room temperature. After washing with PBS, cells were treated with Fluor 488-conjugated anti-mouse IgG for 1 h (1:1000, Invitrogen). Antibodies were diluted in 5% NGS. For co-immunofluorescence staining in Figs 3B and S5D, 293 TRE-SF-Emp, NSF-sT_WT, and NSF-sT_LTSD cells transfected with MCVΔsT were treated with Dox day 1 p.t., seeded on 12 mm coverslips day 3 p.t. and harvested day 4 p.t.. Cells on coverslips were fixed, permeabilized, and blocked as described above. Co-immunostaining was performed sequentially by probing one antigen after another by using mouse (FLAG at 1:500 dilution and CM2B4 at 1:100 dilution) and rabbit (Cux1, c-Jun, and CBP at 1:50 dilution) primary antibodies, which were detected with anti-mouse and anti-rabbit IgG conjugated with either Alexa Fluor 488 or Alexa Fluor 568 (1:1000 dilution, Invitrogen). Samples were reacted with primary and secondary antibodies for 1 h at 37°C and washed with PBS for 15 min before and after secondary antibody reaction. All antibodies were diluted in 5% NGS in PBS. Slides were mounted onto coverslips using DAPI Fluoromount-G (SouthernBiotech, Cat #0100–20). All staining was performed on IX81-DSU Spinning Disk Confocal Microscope (Olympus). The slides were scanned in z-stack mode, using 9 levels with an interplane spacing of 0.22 μm. A stack that included the sectioned nucleus with highest sT expression was subjected to pixel matching colonization analyses to determine Pearson’s correlation efficient by CellSens dimension software (Olympus). For FLAG sT, we used 293 TRE-SF Empty cells as a negative control, and for LT we used mock transfected cells as a negative control (S5A).

Flow cytometry

MCV_WT.L-ZsG and MCV.L-ZsG_rep- as well as their corresponding sT deletion mutants (MCVΔsT.L-ZsG and MCVΔsT.L-ZsG_rep-) were transfected into 293 cells and harvested 4 days p.t.. Cells were pelleted and fixed in 1.5 mL of 4% paraformaldehyde for 8 min. Cells were suspended in 1% FBS and analyzed using a Cytoflex LX (Beckman Coulter). To determine precise GFP positivity, autofluorescent cells were gated out by using FL1 and FL3 scatter plots by using the FCS 7.0 software (De Novo Software).

siRNA and small molecule inhibitors

TriFECTa kits containing three Dicer-Substrate siRNAs for each target (Cux1, c-Jun, BRD9, and CBP) were purchased from IDT. Two of the best D-siRNAs were first screened by immunoblot. SignalSilence siRNA II targeting c-Jun was obtained from Cell Signaling. I-BRD9 and A-485 were purchased from Selleckchem Chemicals. DNA replication inhibitors aphidicolin and mimosine were purchased from Sigma, and used at 5 μM and 400 μM, respectively.

Statistical analysis

Unpaired T-test was used for determining statistical significance for qRT-PCR analyses. Statistical significance was determined as follows: *, P<0.05; **, P<0.01; ***, P<0.001. qRT-PCR analyses were each repeated at least 3 times.