Interaction and assembly of the DNA replication core proteins of Kaposi’s sarcoma-associated herpesvirus

ABSTRACT Kaposi’s sarcoma-associated herpesvirus (KSHV) encodes six highly conserved core replication proteins essential for the viral lytic DNA synthesis, including ORF6 (single-stranded DNA-binding protein), ORF9 (DNA polymerase), ORF40/41 (primase-associated factor), ORF44 (helicase), ORF56 (primase), and ORF59 (polymerase processivity factor). Since the protein-protein interactions among KSHV core replication proteins are largely unknown, this study aimed to decipher their interrelationships. Herein, we propose a protein-protein interaction network of these six core replication proteins according to the results obtained from confocal fluorescence microscopy, coimmunoprecipitation, and mammalian two-hybrid (GAL4/VP16) reporter assays. In this interaction network, ORF40/41 plays a central role in the connection of different replication subcomplexes. In addition to the well-conserved helicase-primase subcomplex (consisting of ORF44, ORF56, and ORF40/41) and the replisome subcomplex (consisting of ORF9 and ORF59), our data also suggest that several discrete, stable subcomplexes exist in the cell nucleus. Among these replication subcomplexes in the nucleus, the tetrameric subcomplex composed of ORF44, ORF56, ORF40/41, and ORF6 exhibited the ability to trigger a DNA damage response. By using an established GAL4/VP16-based reporter system and confocal fluorescence microscopy, we additionally found that the heat shock protein 90 (Hsp90) inhibitor, radicicol, significantly inhibited the formation of both the helicase-primase subcomplex and the replisome subcomplex in a dose-dependent manner. Collectively, these data not only provide further insights into the interaction and assembly of KSHV-encoded core replication proteins but also suggest a critical role of Hsp90 in assisting the construction of the viral core replication machinery. IMPORTANCE Eukaryotic DNA replication is a highly regulated process that requires multiple replication enzymes assembled onto DNA replication origins. Due to the complexity of the cell’s DNA replication machinery, most of what we know about cellular DNA replication has come from the study of viral systems. Herein, we focus our study on the assembly of the Kaposi’s sarcoma-associated herpesvirus core replication complex and propose a pairwise protein-protein interaction network of six highly conserved viral core replication proteins. A detailed understanding of the interaction and assembly of the viral core replication proteins may provide opportunities to develop new strategies against viral propagation.

that control the expression and interaction of different core replication proteins and prevent the mislocalization and accumulation of immature replication subcomplexes in the nucleus of host cells.Furthermore, due to the fact that most of anti-herpesvirus drugs are known to target viral replication enzymes (3)(4)(5)(6), understanding the assembly of herpesviral core replication components in detail may be helpful to develop novel therapies against viral replication.
In this study, we aimed to investigate the protein-protein interactions of KSHVencoded core replication proteins and the assembly of their associated replication intermediates.Different methods including confocal fluorescence microscopy, coimmu noprecipitation, and mammalian two-hybrid (GAL4/VP16) reporter assays were utilized in the study.Based on the results obtained, a protein-protein interaction network of the six viral core replication proteins was proposed.Importantly, we found that radicicol, a heat shock protein 90 (Hsp90) inhibitor, could disrupt the formation of both the conserved helicase-primase subcomplex (consisting of ORF44, ORF56, and ORF40/41) and the replisome subcomplex (consisting of ORF9 and ORF59).Our results, therefore, suggest that Hsp90 may play a critical role in the assembly of KSHV core replication complex.

Subcellular localization of six KSHV-encoded core replication proteins
To characterize the expression and subcellular localization of KSHV core replication proteins, the viral DNA fragments corresponding to the coding regions of ORF6 (SSB), ORF9 (POL), ORF40/41 (PAF), ORF44 (HEL), ORF56 (PRI), and ORF59 (PPF) (Fig. 1A) were individually cloned into the expression vector pFLAG-CMV-2 (carrying a coding sequence of a FLAG tag) or pEGFP-C2 [carrying a coding sequence of a green fluorescent protein (GFP) tag].The resultant plasmids were transfected into 293T cells, and the expressions of these core replication proteins in cells were analyzed by Western blotting.Our results showed that the molecular masses of F-ORF6, F-ORF9, F-ORF40/41, F-ORF44, F-ORF56, and F-ORF59 were 120, 110, 70, 80, 85, and 55 kDa, respectively (Fig. 1B), whereas the molecular masses of GFP-ORF6, GFP-ORF9, GFP-ORF40/41, GFP-ORF44, GFP-ORF56, and GFP-ORF59 were 150, 140, 100, 110, 115, and 85 kDa, respectively (Fig. 1C).Although all viral proteins migrated to their predicted positions in gels, we noticed that ORF40/41 with a FLAG tag or GFP tag was prone to form high-molecular-mass aggregates in the experiments (Fig. 1B and C).The subcellular localization of these core replication proteins in 293T cells was examined by confocal fluorescence microscopy.Confocal microscopic analysis revealed that F-ORF6 and F-ORF59 were mainly localized in the nucleus, whereas F-ORF9, F-ORF44, and F-ORF56 were restricted to the cytoplasm (Fig. 1D).Only F-ORF40/41 was shown in both the nucleus and cytoplasm (Fig. 1D).Consistently, the viral core replication proteins with a GFP tag also exhibited the same subcellular localizations to their respective counterparts with a FLAG tag in 293T cells (Fig. 1E).Notably, three core replication proteins with enzymatic activities, including ORF9, ORF44, and ORF56, were confined to the cytoplasm, suggesting that the cytoplasmic retention of these core replication proteins may be an important regulatory point to prevent untimely DNA synthesis.

Dynamic changes in the subcellular localization of ORF56 in the presence of other core replication proteins
ORF56 is a primase protein essential for initiating the viral DNA synthesis.As shown in Fig. 1, GFP-ORF56 was mainly localized in the cytoplasm when expressed individu ally.However, when GFP-ORF56 was coexpressed with the other five core replication proteins including F-ORF40/41, F-ORF44, F-ORF6, F-ORF9, and F-ORF59 in 293T cells, GFP-ORF56 displayed a punctate pattern in the nucleus (Fig. 2A).To examine the subcellular localization changes of GFP-ORF56 in more detail, different combinations of the viral core replication proteins were coexpressed with GPF-ORF56 in 293T cells.Confocal microscopic analysis found that coexpression of both F-ORF40/41 and F-ORF44 with GFP-ORF56 allowed the nuclear transport of most GFP-ORF56 (Fig. 2B; Fig. S1, panel a).However, coexpression of F-ORF9, F-ORF59, and/or F-ORF6 with GFP-ORF56 did not convey GFP-ORF56 from the cytoplasm to the nucleus (Fig. S1, panels b and c).Furthermore, omission of either F-ORF40/41 or F-ORF44 failed to mediate the nuclear transport of GFP-ORF56, even in the presence of other core replication proteins in transfected cells (Fig. S1, panels d-g).These results suggested that both ORF40/41 and ORF44 were absolutely required for the nuclear transport of ORF56, thereby forming a trimeric helicase-primase subcomplex.
Although coexpression of both F-ORF40/41 and F-ORF44 was essential for the nuclear transport of GFP-ORF56, there were no punctate structures detected under the condition (Fig. 2B).To further investigate the minimal requirement for the formation of GFP-ORF56 puncta in the nucleus, different core replication proteins were coexpressed with the helicase-primase subcomplex consisting of GFP-ORF56, F-ORF40/41, and F-ORF44 in 293T cells (Fig. S2).We found that the formation of GFP-ORF56 puncta in the nucleus required the participation of F-ORF6 but not F-ORF9 and/or F-ORF59 (Fig. 2B; Fig. S2).As compared to GFP-ORF56 puncta formed by coexpression of GFP-ORF56 with F-ORF40/41, F-ORF44, and F-ORF6, a full set of the viral core replication proteins coexpressed in cells substantially increased numbers of GFP-ORF56 puncta in the nucleus (Fig. 2B).In KSHV-positive HH-B2 cells transfected with the plasmid encoding GFP-ORF56, we also found that GFP-ORF56 was expressed exclusively in the cytoplasm under the latent condition (Fig. 2C).However, GFP-ORF56 entered the nucleus and formed a punctate pattern in HH-B2 cells after treatment with sodium butyrate (SB), a lytic-inducing agent (Fig. 2C).Overall, these data suggest that the subcellular localization of ORF56 is tightly controlled by other viral core replication proteins.

Confocal microscopic analysis of the pairwise protein-protein interactions among the viral core replication proteins
Although the viral proteins ORF56, ORF40/41, and ORF44 were believed to form a conserved helicase-primase subcomplex in the cell, the protein-protein interactions among these three proteins actually remain unclear.Moreover, little is known about the interaction of the helicase-primase subcomplex components with the other core replication proteins.To determine the pairwise protein-protein interactions among these viral replication proteins, we first chose GFP-ORF56 as a target protein to analyze its association with the other FLAG-tagged core replication proteins in 293T cells by confocal fluorescence microscopy (Fig. 4A and B).Confocal microscopic analysis showed that GFP-ORF56 had a high degree of colocalization with F-ORF9, F-ORF40/41, and F-ORF44 (Fig. 4A and B; Pearson's correlation coefficient r = 0.80 ± 0.15, r = 0.78 ± 0.11, and r = 0.68 ± 0.11, respectively).However, no significant colocalization was detected between GFP-ORF56 and F-ORF6 or F-ORF59 (Fig. 4A and B; r = -0.41± 0.12 and r = -0.37 ± 0.16, respectively).In these experiments, we additionally noticed that coexpression of GFP-ORF56 and F-ORF40/41 in cells caused retention of F-ORF40/41 in the cytoplasm.
When GFP-ORF40/41 served as a target protein, we observed that GFP-ORF40/41 seemed to colocalize with all five core replication proteins in 293T cells (r = 0.74 ± 0.11 for F-ORF6; r = 0.64 ± 0.15 for F-ORF9; r = 0.68 ± 0.10 for F-ORF56; r = 0.62 ± 0.11 for F-ORF44; r = 0.26 ± 0.12 for F-ORF59) (Fig. 5A and B).However, the value of Pearson's correlation coefficient for the relationship between GFP-ORF40/41 and F-ORF59 was relatively low (r < 0.5).Since the fluorescence signals of GFP-ORF40/41 were diffusely distributed in both the nucleus and the cytoplasm (Fig. 5B), the association between GFP-ORF40/41 and different core replication proteins, particularly F-ORF59, needed to be further determined by using other approaches or methods (see below).

Analysis of the protein-protein interactions among the viral core replication proteins by coimmunoprecipitation
Based on our confocal microscopic analysis, a protein-protein interaction network of the six KSHV-encoded core replication proteins was proposed (Fig. 8A).To further confirm the interactions between these core replication proteins, coimmunoprecipitation (Co-IP) experiments were performed using different combinations of the viral core replica tion proteins (summarized in Fig. 8B: i-viii).By using GFP-ORF56 as a target protein, we demonstrated that the core replication proteins including F-ORF9, F-ORF44, and F-ORF40/41 could be coimmunoprecipitated with GFP-ORF56 (Fig. 8C-i and -ii).However, both F-ORF6 and F-ORF59 could not be coimmunoprecipitated with GFP-ORF56 (Fig. 8C-i and -ii).For studying the interaction of ORF40/41 with the other core replication proteins, Co-IP experiments were performed using coexpression of GFP-ORF40/41 (or F-ORF40/41) with the other core replication proteins in 293T cells (Fig. 8C-iii, -iv, and -v).Besides the interaction between ORF40/41 and ORF56 (Fig. 8C-ii), our results demonstrated that ORF40/41 could specifically interact with ORF44, ORF6, and ORF9 but not ORF59 (Fig. 8C-iii, -iv, and -v).These results suggested that ORF40/41 could interact with all core replication proteins except for ORF59 (Fig. 8A).
As mentioned above, ORF9 was able to interact with ORF56 and ORF40/41 in Co-IP assays (Fig. 8B-i and B-v).We next examined the interaction of ORF9 with the other three core replication proteins including ORF44, ORF6, and ORF59 by Co-IP assays.Our data showed that F-ORF44, but not F-ORF6, could be coimmunoprecipitated with GFP-ORF9 (Fig. 8B-vi).Furthermore, Co-IP analysis also demonstrated the interaction between F-ORF9 and GFP-ORF59 (Fig. 8B-vii).Collectively, these results concluded that ORF9 could interact with all core replication proteins except for ORF6 (Fig. 8A).

Evaluating the assembly of the subunits of the helicase-primase subcomplex by the GAL4/VP16-based reporter system
Due to some limitations in Co-IP assay (e.g., largely depending on the strength of the underlying protein-protein interaction) and in confocal microscopy (e.g., limited resolution for the detection of specific protein-protein interactions), we attempted to use mammalian two-hybrid (GAL4/VP16)-based reporter system to characterize the assembly of the multi-subunit protein subcomplex.In this assay system, the DNA-binding domain of a yeast transcription factor GAL4 and the activation domain of HSV-1 VP16 were separately fused to different core replication proteins.A physical contact between the resultant GAL4-and VP16-fused proteins in the nucleus might lead to the reconstitution of a functional transactivator able to activate the reporter construct containing five GAL4-binding sites [pGal4(5×)-Luc].For studying the assembly of the heterotrimeric helicase-primase subcomplex, the GAL4 DNA-binding domain was first fused to the N-terminus of ORF56 (designated as GAL4-ORF56), and the VP16 activation domain was fused to the N-terminus of ORF44 (designated as VP16-ORF44) (Fig. 9; Fig. S3).As expected, confocal microscopic analysis revealed that the GAL4-ORF56 and VP16-ORF44 fusion constructs, when expressed individually or co-expressed together, were localized in the cytoplasm (Fig. 9A and B).Only coexpression of GAL4-ORF56, VP16-ORF44, and GFP-ORF40/41 together in 293T cells allowed the nuclear transport of all these pro teins (Fig. 9C).When various combinations of the plasmids expressing GAL4-ORF56, VP16-ORF44, and F-ORF40/41 were cotransfected with pGal4(5×)-Luc in 293T cells, we found that only cells transfected with all three expression plasmids showed robust reporter gene activation up to 55-fold (Fig. 9D-i).These results reflected the assembly of the trimeric subcomplex containing GAL4-ORF56, VP16-ORF44, and F-ORF40/41 in the nucleus of transfected cells.
To verify the consistency of this assay system, another fusion construct VP16-ORF40/41 was combined with GAL4-ORF56 in the experiments (Fig. 9D-ii).As shown in Fig. 9D-ii, different sets of the plasmids encoding GALF4-ORF56, VP16-ORF40/41, and F-ORF44 were cotransfected with the reporter plasmid into 293T cells.Our results showed that only coexpression of GAL4-ORF56, VP16-ORF40/41, and F-ORF44 together in cells led to an abundant enhancement of luciferase activity (up to 280-fold).Similarly, the fusion constructs including GAL4-ORF40/41 and VP16-ORF44 were also included in the assay system (Fig. 9D-iii).Consistently, only cells expressing GAL4-ORF40/41, VP16-ORF44, and F-ORF56 were able to abundantly activate the reporter expression (Fig. 9D-iii, up to 60-fold).Collectively, these results strongly suggested that the GAL4/ VP16-based reporter system would be a useful tool to evaluate the formation of the multi-subunit subcomplex.

Determining the interactions among ORF9, ORF59, and ORF6 by the GAL4/ VP16-based reporter system
To apply the GAL4/VP16-based reporter system to analyze the interaction between ORF59 and ORF9, the fusion constructs GAL4-ORF59 and VP16-ORF9 were generated (Fig. S3).Confocal microscopic analysis showed that VP16-ORF9 alone was localized in the cytoplasm, whereas coexpression of GAL4-ORF59 allowed the nuclear transport of VP16-ORF9 (Fig. 10A).For the reporter assay, activation of the pGal4(5×)-Luc reporter gene was detected only in the presence of both VP16-ORF9 and GAL4-ORF59 in cells (Fig. 10A).In a similar way, the plasmids encoding GAL4-ORF9 and VP16-ORF59 were also included in the reporter assay (Fig. 10B).However, we noticed that cells expressing VP16-ORF59 alone showed a high background level of reporter expression (Fig. 10B).The high background reporter activation might be due to a high intrinsic DNA-bind ing activity of ORF59 (25,27).Thus, nonspecific DNA binding of VP16-ORF59 to the reporter plasmid DNA probably caused the high-level reporter activation.Despite having a high-level reporter activation mediated by VP16-ORF59, coexpression of GAL4-ORF9 with VP16-ORF59 could further increase the reporter activation of pGal4(5×)-Luc (Fig. 10B).
According to the results from confocal microscopic experiments and Co-IP experi ments, ORF9 or ORF59 alone could not interact with ORF6 (Fig. 8).To investigate whether the ORF9-ORF59 subcomplex could interact with ORF6 in the nucleus, the ORF6 fusion constructs including VP16-ORF6 and GAL4-ORF6 were individually used in the reporter assay (Fig. 10C and D).Our results showed that coexpression of VP16-ORF6 with GAL4-ORF59 and F-ORF9 (Fig. 10C) or coexpression of GAL4-ORF6 with VP16-ORF9 and F-ORF59 in 293T cells (Fig. 10D) could not activate the reporter gene expression.These findings implicated that there was no direct interaction between ORF6 and the ORF9-ORF59 subcomplex in the nucleus.

Analyzing the association between the helicase-primase subcompex and the replisome subcomplex
As described above, all the subunits of the helicase-primase subcomplex were able to individually interact with ORF9 (Fig. 5 and 8).Due to the fact that the helicase-primase subcomplex could enter the nucleus, we sought to determine whether the helicase-pri mase subcomplex could convey ORF9 to the nucleus.To test this possibility, we first examined the subcellular localization of GFP-ORF9 in the presence of F-ORF56, F-ORF44, and F-ORF40/41 (Fig. 12A-i).Confocal microscopy analysis showed that GFP-ORF9 could be partially transported into the nucleus in the presence of F-ORF56, F-ORF44, and nucleus (Fig. 12A-ii).These results, therefore, implied that ORF9 could be transported into the nucleus via the helicase-primase subcomplex (Fig. 12A-iii).

Effect of radicicol on the formation of both the helicase-primase and replicase subcomplexes
To study whether small molecules could specifically disrupt the assembly or nuclear transport of the viral helicase-primase subcomplex and the replisome subcomplex, a compound library was screened by using the GAL4/VP16-based reporter system.We here found that radicicol, an inhibitor of heat shock protein 90 (Hsp90), could significantly inhibit the reporter activation induced by coexpression of GAL4-ORF56, VP16-ORF44, and F-ORF40/41 (Fig. 13A) and by coexpression of GAL4-ORF59 and VP16-ORF9 (Fig. 13B) in a dose-dependent manner.It should be noted that radiciol at the concentrations used in the study (0.25-1 µM) displayed only weak cytotoxic activity on 293T cells (Fig. S4).To further verify the effect of radicicol on the assembly and nuclear transport of the ORF56-ORF44-ORF40/41 subcomplex and the ORF59-ORF9 subcomplex, the subcellu lar localization of individual core replication proteins in 293T cells was examined by confocal fluorescence microscopy.When GFP-ORF56, VP16-ORF44, and GAL4-ORF40/41 were coexpressed in 293T cells, we found that treatment with increasing amounts of radicicol (0.5 and 1 µM) greatly increased the cytoplasmic retention of these core replication proteins in cells (Fig. 13C).On the other hand, when GFP-ORF9 and F-ORF59 were coexpressed in 293T cells, treatment with radicicol specifically affected the nuclear localization of GFP-ORF9 but not F-ORF59 (Fig. 13D).These results implied that Hsp90 was required for the protein-protein interaction and/or the nuclear transport of both the helicase-primase and replisome subcomplexes.

DISCUSSION
Although the lytic DNA replication of herpesviruses has been studied for several decades, the protein-protein interactions of herpesviral DNA replication proteins are not yet fully understood.In this report, we characterize the protein-protein interactions of six KSHV-encoded core replication proteins by using different methods, and a complete protein-protein interaction network of these conserved core replication proteins is proposed.This study may be important not only to deepen our knowledge about the assembly of the viral DNA replication complex but also to provide opportunities to develop new strategies against viral propagation.

Coordinate regulation of the assembly of KSHV core replication proteins
During the KSHV lytic cycle, the expression and the interaction of the viral core replication components must be tightly controlled to allow the correct assembly of a functional replication complex in the cell nucleus.Although the lytic DNA replication of KSHV takes place within the cell nucleus, we here showed that three viral replication components with enzymatic activity including ORF56, ORF44, and ORF9 when expressed individually, were restricted to the cytoplasm (Fig. 1).It is possible that the cytoplasmic (Continued on next page) retention of specific viral replication components may be a key control mechanism to avoid unlicensed viral DNA replication when other necessary core replication compo nents are not yet available during the viral lytic cycle or when leaky expression of these viral proteins occurs upon some stress conditions.In our research, we demonstrated that the nuclear translocation of OR56 and ORF44 requires ORF40/41 (forming a trimeric subcomplex).Additionally, consistent with previous studies (26), we confirmed that the nuclear translocation of ORF9 requires ORF59 (forming a dimeric subcomplex).
Like HSV-1 and EBV, the KSHV core replication complex could be generally divided into three modules including the ORF56-ORF44-ORF40/41 trimeric subcomplex, the ORF9-ORF59 dimeric subcomplex, and the ORF6 subunit (Fig. 8A).To monitor the intermediates in the assembly of the multi-subunit replication complex, we chose ORF56 as a target to analyze its subcellular localization along with other core replication proteins (Fig. 2).According to our results, ORF56 initially entered the nucleus in the presence of ORF44 and ORF40/41, leading to the formation of the helicase-primase subcomplex.The resultant ORF56-ORF44-ORF40/41 subcomplex in the nucleus may interact with the ORF6 subunit and/or the ORF9-ORF59 heterodimeric subcomplex.Interestingly, coexpression of the ORF56-ORF44-ORF40/41 subcomplex with ORF6 could cause the formation of ORF56 puncta in the nucleus.Particularly, these ORF56 nuclear puncta observed in transfected cells significantly colocalized with the signals of γH2AX, 53BP1, or RPA foci (Fig. 3).Since ORF6 in the nucleus may preferentially bind to single-stranded DNA intermediates derived from cellular DNA replication, repair, or recombination, we speculate that the recruitment of the helicase-primase subcomplex to these DNA sites through ORF6 may trigger DNA replication stress, ultimately leading to double-stranded DNA breaks and significant DNA damage.In agreement with previous observations, several studies have shown that lytic induction of KSHV could lead to activation of the DNA damage response (DDR) pathway including increased phosphory lation of H2AX (36)(37)(38).In future studies, the detailed relationship among the viral core replication subcomplex, cellular DNA repair enzymes, and the potential DNA damage sites needs to be further characterized.In contrast, we did not find the punctate pattern of ORF56 when the ORF56-ORF44-ORF40/41 subcomplex was coexpressed with the ORF9-ORF59 subcomplex.As compared to the coexpression of four core replication proteins (including ORF56, ORF44, ORF40/41, and ORF6), the coexpression of all six core replication proteins significantly increased the number of ORF56 puncta in the nucleus (Fig. 2B).

Protein-protein interaction networks of KSHV-encoded core replication components
Detailed characterization of the pairwise protein-protein interactions revealed that all three components within the helicase-primase subcomplex could interact with each other (Fig. 8).Furthermore, we propose that the association of the helicase-primase subcomplex with ORF6 in the nucleus is mediated through the interaction between ORF40/41 and ORF6 (Fig. 8A and 11).For the association between the helicase-primase subcomplex and the replisome subcomplex, our results showed that each component within the helicase-primase subcomplex could individually interact with ORF9 but not with ORF59 (Fig. 8A).It is possible that during the viral lytic DNA synthesis, ORF9 may dynamically interact with the components of the helicase-primase subcomplex to coordinate DNA synthesis of the leading and lagging strands.On the other hand, we did not detect any significant interaction between the replisome subcomplex and ORF6.Besides the well-conserved helicase-primase and replisome subcomplexes in the nucleus, we additionally found that several distinct, stable subcomplexes exist in the nucleus.These subcomplexes include (i) the dimeric subcomplex containing ORF40/41 and ORF6 (Fig. 11); (ii) the tetromeric subcomplex containing ORF56, ORF44, ORF40/41, and ORF6 (Fig. 11); and (iii) the tetromeric subcomplex containing ORF56, ORF44, ORF40/41, and ORF9 (Fig. 12).Noteworthily, the formation of the latter subcomplex ORF56-ORF44-ORF40/41-ORF9 could be of potential importance because the helicaseprimase subcomplex may have the ability to convey ORF9 to the nucleus when ORF59 is absent (Fig. 12).Although a protein-protein interaction network of six KSHV-encoded core replication proteins is proposed here, the interaction network still needs to be further confirmed in the context of KSHV infection.To achieve the goal, both the preparation of reliable antibodies specific to endogenous core replication proteins and the creation of a set of PEL cell lines with specific gene knockout in KSHV genome may be helpful for experi mental investigations.In addition to the six KSHV replication core proteins, the assembly of a complete functional replication complex is believed to require other viral proteins (such as ORF50 and K8) or cellular proteins as well as the cis-acting lytic origin (oriLyt).The interrelationship of these replication components remains to be further determined.Moreover, since dimerization or multimerization of specific core replication proteins may be necessary for the assembly of the replication complex, future experiments may need to determine whether certain replication proteins are able to form dimers or multimers.
In the study, different methods including confocal fluorescence microscopy, Co-IP, and the GAL4/VP16-mediated reporter system were used to determine the protein-pro tein interactions of KSHV core replication proteins.However, these methods individu ally have certain limitations.For instance, low-resolution images taken by confocal fluorescence microscopy might lead to false data interpretation.Although Co-IP assay is considered a gold standard method for studying the protein-protein interaction, the low-affinity protein-protein interactions might not be easily detected.Herein, we attempted to establish the GAL4/VP16-based reporter system for probing interactions among subunits of different subcomplexes.This approach is a fast and highly sensitive assay, which may serve as a screening tool for identifying small compounds that inhibit the protein-protein interactions of the viral core replication proteins.Although the GAL4/VP16-mediated reporter system is useful to determine physiologically relevant protein-protein interactions, there are two major limitations for this approach.First, this established method is not suitable for detecting the assembly intermediates that are localized in the cytoplasm (such as heterodimeric subcomplexes: ORF56-ORF44, ORF56-ORF9, and ORF44-ORF9).Second, due to the fact that the formation of multi-subunit subcomplexes might cause steric hindrance of GAL4-and/or VP-16-containing proteins, this method might have a problem to detect the assembly of certain subcomplexes (e.g., the ORF56-ORF44-ORF40/41-ORF9 subcomplex) and the full replication complex (Fig. 12).

Role of Hsp90 in the formation of both the helicase-primase and replicase subcomplexes of KSHV
Hsp90 is an evolutionally conserved molecular chaperone (39,40).Previous studies have shown that Hsp90 plays a critical role in the nuclear transport of the HSV-1 DNA polymerase UL30 and the EBV DNA polymerase BALF5, two homologs of the KSHV ORF9 (41,42).In the case of the HSV-1 UL30, it is normally localized in the nucleus when expressed alone (43); however, treatment of the Hsp90 inhibitor geldanamycin blocked its nuclear transport (41).Unlike the HSV-1 UL30, the EBV DNA polymerase BALF5 is localized in the cytoplasm when expressed alone.The nuclear transport of the EBV BALF5 requires the polymerase processivity factor BMRF1 (42).Pharmacological inhibition or genetic knockdown of Hsp90 markedly prevented the nuclear transport of the EBV BALF5 even in the presence of BMRF1 (42).It should be noted that inhibition of the Hsp90 activity did not affect the nuclear transport of BMRF1 (42).By using the GAL4/VP16-based reporter assay, we found that the Hsp90 inhibitor radicicol was able to reduce the reporter activation mediated by the combination of GAL4-ORF59 and VP16-ORF9 in a dose-dependent manner (Fig. 13B).Confocal microscopic analysis further demonstrated that radicicol significantly inhibited the nuclear transport of ORF9 but not the nuclear transport of ORF59.These results were similar to those obtained from the study of the EBV BALF5 and BMRF1 (42).Our findings, therefore, conclude that Hsp90 is involved in the interaction between ORF9 and ORF59.On the other hand, our studies also revealed that radicicol treatment could markedly inhibit the reporter activation induced by the combination of GAL4-ORF56, VP16-ORF44, and F-ORF40/41 in the GAL4/ VP16-based reporter assay (Fig. 13A).Confocal microscopic analysis demonstrated that radicicol could block the nuclear translocation of all the helicase-primase subcomplex components.These results suggest that Hsp90 may play critical roles in the protein-pro tein interactions of the helicase-primase subcomplex and/or the subsequent nuclear transport of this subcomplex.Although the detailed mechanisms underlying the role of Hsp90 in the assembly of both the KSHV helicase-primase and replisome subcomplexes still remain unclear, we think that Hsp90 could be a common factor essential for the assembly of the core replication complex in all of herpesviruses.In addition to Hsp90 inhibitors, our experiments have also identified several bioactive small molecules that could inhibit the reporter activation mediated by coexpression of GAL4-ORF56, VP16-ORF44, and F-ORF40/41 or by coexpression of GAL4-ORF59 and VP16-ORF9.Further research is needed to elucidate the mechanism of action of these identified bioactive compounds in affecting the assembly of the KSHV core replication complex.
In summary, we characterize the protein-protein interactions of six KSHV-encoded core replication proteins by using different methods and a complete protein-protein interaction network of these viral core replication proteins is proposed.In addition to the well-conserved helicase-primase and replicase subcomplexes, several stable assembly intermediates are also identified in the cell nucleus.Furthermore, we show that Hsp90 plays significant roles in the assembly of the KSHV core replication complex.

Coimmunoprecipitation
Coimmunoprecipitation assays were carried out as mentioned previously (49).Briefly, 293T cells were cotransfected with the plasmids expressing GFP fusion proteins and the plasmids expressing FLAG-tagged proteins.At 24 h after transfection, cells were harvested and lysed in the coimmunoprecipitation assay buffer (50 mM Tris-HCl pH 7.6, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, and 1 mM phenylmethylsulfonyl fluoride).Protein lysates were subsequently mixed with GFP-Trap MA beads (no.gtma-100; ChromoTek).After immunoprecipitation, the resultant immune complexes were subjected to Western blot analysis using specific antibodies.

Luciferase assays
293T cells were seeded on 24-well plates (7 × 10 5 cells/well) one day before trans fection.Each transfection used a fixed amount (0.8 µg) of plasmid DNA including the reporter plasmid pGal4(5×)-Luc and effector plasmids.At 24 h post-transfection, luciferase reporter activities of transfected cell samples were measured using the luciferase reporter assay system (no.E1501; Promega).Fold activation was calculated as the luciferase activity of the reporter construct in the presence of effectors divided by that in the presence of the empty control vector.All experiments were performed at least three times with duplicate samples.

Statistical analysis
All data were presented as means ± standard deviation (SD).The Mann-Whitney U test was used to evaluate differences between samples.All statistical analyses were conducted using SPSS (Statistical Package for the Social Sciences) software (version 18.0; IBM Corporation, Armonk, NY), and the P-values less than 0.05 were considered statistically significant.

FIG 2
FIG 2 Changes in the subcellular localization of GFP-ORF56 induced by other viral core replication proteins.(A) Confocal microscopic analysis of the subcellular localization of GFP-ORF56 in the presence of five FLAG-tagged core replication proteins in 293T cells.Scale bars, 20 µM.(B) A proposed model for changes in the GFP-ORF56 subcellular localization during the assembly of the viral core replication complex.When specific sets of core replication proteins were coexpressed, different localization patterns of GFP-ORF56 were observed in cells, including (i) exclusively localized in the cytoplasm, (ii) mainly localized in the nucleus, and (iii) mainly localized in the nucleus with puncta formation.Scale bars, 10 µM.(C) Subcellular localization of GFP-ORF56 in HH-B2 cells untreated or treated with SB.In the experiments, HH-B2 cells transfected with pCMV-GFP-ORF56 were left untreated or treated with SB for 24 h, and then the subcellular localization of GFP-ORF56 in transfected cells was analyzed by confocal microscopy.Scale bars, 10 µM.Notably, SB treatment induced the formation of GFP-ORF56 puncta in the cell nucleus (white arrow) and the expression of the viral lytic protein ORF50 (red).All experiments shown above were performed at least three times independently with similar results.

FIG 3
FIG 3 Colocalization of GFP-ORF56 puncta with the focal sites of γH2AX, 53BP1, and RPA in the cell nucleus.After 293T cells were transfected with the indicated expression plasmids for 24 h, the transfected cells were analyzed for the subcellular localization of GFP-ORF56 and the DNA damage markers including γH2AX (A), 53BP1 (B), and RPA (C) by confocal fluorescence microscopy.Scale bars, 10 µM.As noted, the foci formation ofγH2AX, 53BP1, or RPA was mainly detected in 293T cells expressing GFP-ORF56, F-ORF40/41, F-ORF44, and F-ORF6 together.White arrows indicate the colocalization of GFP-ORF56 puncta with γH2AX, 53BP1, or RPA foci in the nucleus of transfected cells.All experiments were performed at least twice independently with similar results.

FIG 4
FIG 4 Analysis of the colocalization of GFP-ORF56 or GFP-ORF44 with the other KSHV-encoded core replication proteins in 293T cells.(A) Summary of the colocalization of GFP-ORF56 with the other five core replication proteins in 293T cells."×, " no colocalization.(B) Quantitative analysis of the colocalization between GFP-ORF56 and the indicated FLAG-tagged core replication proteins in 293T cells.The degree of colocalization of two different fluorescent labels in cells was determined by the Pearson's correlation coefficient (r).Scale bars, 20 µM.Data are expressed as means ± SD (n = 15 cells, three independent experiments).(C) Summary of the colocalization of GFP-ORF44 with the other five core replication proteins in 293T cells."×, " no colocalization.(D) Quantitative analysis of the colocalization between GFP-ORF44 and the indicated FLAG-tagged core replication proteins in 293T cells.The degree of colocalization of two different fluorescent labels in cells was determined by the Pearson's correlation coefficient (r).Scale bars, 20 µM.Data are expressed as means ± SD (n = 15 cells, three independent experiments).

FIG 5
FIG 5 Analysis of the colocalization of GFP-ORF40/41 or GFP-ORF9 with the other KSHV-encoded core replication proteins in 293T cells.(A) Summary of the colocalization of GFP-ORF40/41 with the other five core replication proteins in 293T cells."×, " no colocalization; "#, " additional validation required.(B) Quantitative analysis of the colocalization between GFP-ORF40/41 and the indicated FLAG-tagged core replication proteins in 293T cells.The colocalization between GFP-ORF40/41 and the indicated FLAG-tagged core replication proteins in cells was assessed by the Pearson's correlation coefficient (r).Scale bars, 20 µM.Data are presented as means ± SD (n = 15 cells, three independent experiments).(C) Summary of the colocalization of GFP-ORF9 with the other five core replication proteins in 293T cells."×, " no colocalization.(D) Quantitative analysis of the colocalization between GFP-ORF44 and the indicated FLAG-tagged core replication proteins in 293T cells.The colocalization between GFP-ORF44 and the indicated FLAG-tagged core replication proteins in cells was assessed by the Pearson's correlation coefficient (r).Scale bars, 20 µM.Data are presented as means ± SD (n = 15 cells, three independent experiments).

FIG 6
FIG 6 Confocal microscopic analysis of the colocalization between GFP-ORF59 and the other five core replication proteins in 293T cells.(A) Summary of the colocalization of GFP-ORF59 with the other five core replication proteins."×, " no colocalization; "#, " additional validation required.(B) Quantitative analysis of the colocalization of GFP-ORF59 with the indicated FLAG-tagged core replication proteins in 293T cells.The degree of colocalization of GFP-ORF59 with the indicated FLAG-tagged core replication proteins in cells was determined by the Pearson's correlation coefficient (r).Scale bars, 20 µM.Data are presented as means ± SD (n = 15 cells, three independent experiments).(C) Confocal microscopic analysis of the wild-type and NLS-mutated GFP-ORF59 constructs expressed in 293T cells.Scale bars, 20 µM.(D) Quantitative analysis of the colocalization between the NLS-mutated GFP-ORF59 and the indicated proteins including F-ORF9, F-ORF6, and F-ORF40/41.Data are presented as means ± SD (n = 15 cells, three independent experiments).Scale bars, 20 µM.

FIG 7
FIG 7 Confocal microscopic analysis of the colocalization between GFP-ORF6 and the other five core replication proteins in 293T cells.(A) Summary of the colocalization of GFP-ORF6 with the other five core replication proteins."×, " no colocalization; "#, " additional validation required.(B) Quantitative analysis of the colocalization of GFP-ORF6 with the indicated FLAG-tagged core replication proteins in 293T cells.The degree of colocalization of GFP-ORF6 with the indicated FLAG-tagged core replication proteins in cells was determined by the Pearson's correlation coefficient (r).Scale bars, 20 µM.Data are presented as means ± SD (n = 15 cells, three independent experiments).(C) Confocal microscopic analysis of the wild-type and NLS-mutated GFP-ORF6 constructs expressed in 293T cells.Scale bars, 20 µM.(D) Quantitative analysis of the colocalization between the NLS-mutated GFP-ORF6 and the indicated proteins including F-ORF40/41 and F-ORF59.Data are presented as means ± SD (n = 15 cells, three independent experiments).Scale bars, 20 µM.

FIG 8
FIG 8 Analysis of the pairwise protein-protein interactions among the viral core replication proteins by coimmunoprecipitation.(A) A proposed protein-protein interaction network of the KSHV core replication complex.There are three major modules in this interaction network, including a heterotrimeric subcomplex, a heterodimric subcomplex, and the ORF6 subunit.Labels (i-vii) indicate the pairwise protein-protein interactions determined in coimmunoprecipitation (Co-IP) experiments shown in (C).(B) Schematic illustration of Co-IP experiments (i-viii) conducted in (C) using specific sets of viral core replication proteins.(C) Co-IPanalysis of the indicated protein-protein interactions.Typically, 293T cells were cotransfected with the indicated expression plasmids for 24 h, and then, the protein lysates from these transfected cells were immunoprecipitated using anti-GFP antibody.After immunoprecipitation, the precipitated proteins were analyzed by Western blotting using anti-GFP and anti-FLAG antibodies.All Co-IP experiments were repeated at least twice with similar results.

FIG 10
FIG 10 Analysis of the formation of the ORF9-ORF59 subcomplex and the possible interaction between the ORF9-ORF59 subcomplex and ORF6 using a two-hybrid (GAL4/VP16) reporter assay.(A) Confocal microscopic analysis and GAL4/VP16-mediated reporter analysis of the interaction between GAL4-ORF59 and VP16-ORF9.Scale bars in the confocal microscopic images represent 20 µM.The confocal microscopy experiments were performed at least three times independently with similar results.The results from reporter assays are expressed as means ± SD (n = 3).*P < 0.01, for results compared to those from the empty vector group.(B) Confocal microscopic analysis and GAL4/VP16-mediated reporter analysis of the interaction between GAL4-ORF9 and VP16-ORF59.Scale bars in the confocal microscopic images represent 20 µM.The confocal microscopy experiments were performed at least three times independently with similar results.

FIG 13
FIG13 Inhibition of the formation of the helicase-primase and replisome subcomplexes by radicicol.(A) Effects of increasing amounts of radicicol on the reporter activation mediated by coexpression of GAL4-ORF56, VP16-ORF44, and F-ORF40/41 in 293T cells.Data are presented as means ± SD (n = 3).*P < 0.01, for results compared to those from the empty vector group; #P < 0.01, for results compared to those from the untreated ORF56-ORF44-ORF40/41 group.(B) Effects of increasing amounts of redicicol on the reporter activation mediated by GAL4-ORF59 and VP16-ORF9 in 293T cells.Data are presented as means ± SD (n = 3).

FIG 13 (
FIG13 (Continued) *P < 0.01, for results compared to those from the empty vector group; #P < 0.01, for results compared to those from the untreated ORF59-ORF9 group.(C) Effects of radicicol on the subcellular localization of components of the helicase-primase subcomplex in 293T cells.In the experiments, 293T cells cotransfected with the plasmids encoding GFP-ORF56, VP16-ORF44, and GAL4-ORF40/41 were treated with DMSO (used as a solvent) or radicicol (0.5 and 1.0 µM) for 24 h, and then, the subcellular localization of each protein in transfected cells was analyzed by confocal microscopy.Scale bars, 20 µM.The subcellular localization patterns of GFP-ORF56 (acting as a representative) in different treatment groups were calculated and plotted in the right panel (n = 3).N, nucleus; N/C, both nucleus and cytoplasm; C, cytoplasm.(D) Effects of radicicol on the subcellular localization of components of the replisome subcomplex in 293T cells.In the experiments, 293T cells cotransfected with the plasmids encoding GFP-ORF9 and F-ORF59 were treated with DMSO (used as a solvent) or radicicol (0.5 and 1.0 µM) for 24 h, and the subcellular localization of each protein in transfected cells was analyzed by confocal microscopy.Scale bars, 20 µM.Changes in the subcellular localization of GFP-ORF9 in different treatment groups were shown in the right panel (n = 3).N, nucleus; N/C, both nucleus and cytoplasm; C, cytoplasm.As noted, radicicol treatment did not affect the nuclear transport of F-ORF59.