|Home | About | Journals | Submit | Contact Us | Français|
Kaposi’s sarcoma-associated herpesvirus (KSHV) replication and transcription activator (RTA) encoded by ORF50 is a lytic switch protein for viral reactivation from latency. The expression of RTA activates the expression of downstream viral genes, and is necessary for triggering the full viral lytic program. Using chromatin immunoprecipitation assay coupled with a KSHV whole-genome tiling microarray (ChIP-on-chip) approach, we identified a set of 19 RTA binding sites in the KSHV genome in a KSHV-infected cell line BCBL-1. These binding sites are located in the regions of promoters, introns, or exons of KSHV genes including ORF8, ORFK4.1, ORFK5, PAN, ORF16, ORF29, ORF45, ORF50, ORFK8, ORFK10.1, ORF59, ORFK12, ORF71/72, ORFK14/ORF74, and ORFK15, the two origins of lytic replication OriLyt-L and OriLyt-R, and the microRNA cluster. We confirmed these RTA binding sites by ChIP and quantitative real-time PCR. We further mapped the RTA binding site in the first intron of the ORFK15 gene, and determined that it is RTA-responsive. The ORFK15 RTA binding sequence TTCCAGGAA TTCCTGGAA consists of a palindromic structure of two tandem repeats, of which each itself is also an imperfect inverted repeat. Reporter assay and electrophoretic mobility shift assay confirmed the binding of the RTA protein to this sequence in vitro. Sequence alignment with other RTA binding sites identified the RTA consensus binding motif as TTCCAGGAT(N)0–16TTCCTGGGA. Interestingly, most of the identified RTA binding sites contain only half or part of this RTA binding motif. These results suggest the complexity of RTA binding in vivo, and the involvements of other cellular or viral transcription factors during RTA transactivation of target genes.
Infection by Kaposi’s sarcoma-associated herpesvirus (KSHV), also known as human herpesvirus 8 (HHV-8), is associated with the development of Kaposi’s sarcoma (Chang et al., 1994), primary effusion lymphoma (PEL) (Cesarman et al., 1995) and a subset of multicentric Castleman’s diseases (Soulier et al., 1995). Like other herpesviruses, the life cycle of KHSV consists of latent and lytic phases that specify expression patterns of four classes of viral genes: latent, immediate-early (IE), early, and late genes (Greene et al., 2007). The IE genes are made immediately after primary infection or upon reactivation from latency, and do not require de novo protein synthesis (Deng et al., 2007). IE genes generally encode for regulatory proteins, and are critical for initiating viral transcription (Miller et al., 2007). KSHV replication and transcription activator (RTA) is the gene product of a major IE transcript transcribed from the ORF50 locus (Chen et al., 2000; Lukac et al., 1999; Sun et al., 1998). The switch from KSHV latent to lytic replication can be initiated by specific intracellular signals or extracellular stimuli including chemical inducers such as 12-O-tetradecanoyl phorbol-13 acetate (TPA) and sodium butyrate that activate the expression of this key lytic switch protein. The expression of RTA is necessary and sufficient to trigger the full lytic program resulting in the cascade expression of viral proteins, release of viral progeny, and host cell death (Gradoville et al., 2000; Lukac et al., 1999; Lukac et al., 1998; Sun et al., 1998; Xu et al., 2005).
RTA has been shown to activate the transcription of several downstream viral genes, including ORFK1 (Bowser et al., 2002; Bowser et al., 2006), ORFK2 (Deng et al., 2002; Song et al., 2003), ORFK3 (Rimessi et al., 2001), ORFK5 (Haque et al., 2000), ORFK6 (Chang et al., 2005b), PAN RNAs (Song et al., 2001; Song et al., 2003; Song et al., 2002), ORF35 (Masa et al., 2008), ORF49 (Gonzalez et al., 2006), ORF50 itself (Chen et al., 2001; Deng et al., 2000; Sakakibara et al., 2001), ORFK8 (Lukac et al., 2001; Seaman and Quinlivan, 2003; Wang et al., 2004c; Wang and Yuan, 2007), ORF57 (Byun et al., 2002; Song et al., 2003; Wang et al., 2003), ORFK9 (Chen et al., 2000; Ueda et al., 2002), ORF74 (Liang and Ganem, 2004), ORFK12 (Song et al., 2003), and ORFK15 (Wong and Damania, 2006). RTA activates its target genes through at least two mechanisms: direct binding to the RTA-responsive elements (REs), or indirect binding to other cellular factors (Chang et al., 2005b). For example, RTA directly binds to the promoters of PAN RNAs (Song et al., 2003; Song et al., 2002) and ORFK12 (Chang et al., 2002) but does not bind to the ORF57 and ORFK6 promoters (Chang et al., 2005b). The indirect binding of RTA to its REs may be involved with the interaction of RTA with RBP-Jkappa, a Notch signal pathway transcription factor (Chang et al., 2005a; Liang et al., 2002; Liang and Ganem, 2003; Liang and Ganem, 2004). Other cellular transcriptional factors that are involved in the activation of viral genes include AP-1 and OCT-1 for ORF50 itself (Pan et al., 2006; Sakakibara et al., 2001; Wang et al., 2004a; Xie et al., 2008) and ORFK8 (Carroll et al., 2007), or AP-1 for ORF35 (Masa et al., 2008), AP-1 or SP-1 for ORFK9 (Chen et al., 2000; Ueda et al., 2002), and AP-1 for ORF57 (Byun et al., 2002; Wang et al., 2004a). It has also been shown that RTA recruits CBP, the SWI/SNF chromatin remodeling complex, and the TRAP/mediator co-activator to its target promoters, and that such recruitment is essential for the expression of RTA-dependent viral genes (Gwack et al., 2003).
The RTA protein is mainly encoded by ORF50 but obtains an additional 60 amino acid (aa) to its N-terminal through a splicing event that shifts its start codon across ORF49 (Chen et al., 2000; Sun et al., 1998). The addition of the 60 aa is critical for its function as a viral transactivator because a cloned protein from ORF50 itself neither locates to the nuclei nor has any transactivation capacity (Chen et al., 2000). The full-length functional RTA protein contains 691 aa, and is predicted to have a molecular mass of 73.7 kDa. However, cellular RTA migrated as a series of polypeptides with the major ones at around 119 and 101 kDa, suggesting that RTA is post-translationally modified (Chang and Miller, 2004; Lukac et al., 1999). RTA lacks significant homology with any cellular proteins but is a functional homolog of the RTA proteins from Epstein-Barr virus (EBV), herpesvirus saimiri (HVS) and murine gammaherpesvirus 68 (MHV-68) (Wu et al., 2000). The RTA protein consists of an N-terminal DNA binding domain, a central dimerization domain, a C-terminal acidic activation domain, and two nuclear localization signals (NLSs) (Chen et al., 2000; Lukac et al., 1999). The DNA binding domain of RTA is located at the N-terminus from aa 1 to 530. A deletion mutant of the activation domain sequences, containing only the DNA binding domain (aa 1–530), has been shown to be a dominant-negative mutant of RTA transactivation (Lukac et al., 1999). This mutant maintains its DNA binding activity for RTA REs but no longer activates the expression of downstream lytic genes (Lukac et al., 2001; Lukac et al., 1999; Song et al., 2001; Song et al., 2003). This construct binds to the RTA binding consensus sequence as a tetramer (Bu et al., 2007). Using a purified fusion protein of this mutant GST-50ΔSTAD (aa 1 to 530) expressed in bacteria in an in vitro DNA binding assay named systematic evolution of ligands by exponential enrichment assay (SELEX), Ziegelbauer and colleagues identified a total of 18 RTA direct binding sites in the KSHV genome including a binding site for the ORF8 promoter; however they did not recover the known high-affinity PAN promoter site, nor the sites for the promoters of ORFK2, ORFK8, ORF57, ORFK12 and ORF74 (Ziegelbauer et al., 2006). This may indicate that the bacterially expressed GST-RTA fusion protein and the in vitro binding assay might not be the ideal method for delineating the RTA binding sites in vivo. To address this problem, we applied an in vivo ChIP-on-chip approach coupled with a KSHV genome tiling array and a stable BCBL-1 cell line expressing a triple FLAG-tagged RTA DNA binding domain to identify the RTA direct binding sites in the KSHV genome. A full set of RTA direct binding sites in the KSHV genome including some that have not been previously described were identified.
Previous studies have shown that RTA activates its downstream genes by both direct and indirect bindings to the DNA sequences (Chang et al., 2005b). To identify the RTA direct binding sites in the KSHV genome, we generated a KSHV-infected stable BCBL-1 cell line, BCBL-1-dRTA cell line, expressing the RTA DNA binding domain tagged with 3FLAG epitope. Similar to the parental cell line, the majority of the cells in this new cell line were latently infected by KSHV but can be reactivated into lytic replication with chemical inducer such as TPA or sodium butyrate. The use of RTA DNA binding domain avoided the constitutive activation of viral lytic replication, and hence cell death by RTA (Gradoville et al., 2000; Lukac et al., 1998). This construct maintains the DNA-binding activity for the RTA REs albeit it is no longer capable of activating the expression of lytic genes (Lukac et al., 2001; Lukac et al., 1999; Song et al., 2001; Song et al., 2003). It also reduces the confounding effects of cellular factors that cause indirect RTA binding.
We carried out ChIP with TPA-induced BCBL-1-dRTA cells as described in Materials and Methods. The RTA-binding enriched DNA was hybridized with a KSHV whole-genome tiling microarray. We identified a set of RTA direct binding sites in the KSHV genome. Fig. 1 shows the genomic locations of the RTA binding sites. Peaks indicated the relative enrichment of RTA-bound DNA (ratio of ChIP-enriched DNA over non-enriched DNA). The selected criteria for each RTA binding site were that it had at least five consecutive peaks, and each peak had a minimum two-fold ChIP/Input ratio of enrichment. Because the microarray has a 60-mer resolution, this represents a total of 300 11bp in length for the RTA binding sites, which is comparable to the original sheared immunoprecipitated DNA fragments. Based on this criteria, we identified a total of 19 RTA binding sites in the KSHV genome, which are located in the promoters, introns or exons of KSHV genes including ORF8, ORFK4.1, ORFK5, PAN RNAs, ORF16, ORF29, ORF45, ORF50, ORFK8, ORFK10.1, ORF59, ORFK12, ORF71/72, ORFK14/ORF74, and ORFK15, the two origins of lytic replication (OriLyts) OriLyt-L and OriLyt-R, and the microRNA cluster. The finding that RTA binds OriLyts confirms its crucial role in viral lytic replication. As a major IE gene, RTA binds not only to the promoters of other IE genes such as ORFK4.2, ORFK5, ORF29b, ORF45 and ORF48, as well as itself, but also to early genes and a late gene ORF8. Interestingly, RTA strongly binds to the latent gene cluster region (ORFK12 to ORF74), and thus may serve as a transactivator for the expression of these latent transcripts, including those of KSHV microRNAs. In fact, an inducible promoter (LTi) and a constitutive promoter (LTc) immediately upstream of ORF73 have been described, and RTA activates LTi but not LTc (Matsumura et al., 2005; Lan et al., 2005).
Fig. 2 shows the detailed locations of RTA direct binding sites in the KSHV genome. There were two RTA binding sites located in the ORFK12 promoter region. The location of RTA binding sites in the P type ORFK15 promoter (K15P) was based on the KSHV sequence Accession No: U85269 (Nicholas et al., 1997). Table 1 listed all the identified RTA direct binding sites with the possible corresponding target genes. Some of these genes such as PAN RNAs, ORFK8, ORFK12, ORF74 and ORF8 have been previously identified to be the RTA direct targets, and their RTA REs described (Lukac et al., 1999; Lukac et al., 1998; Song et al., 2001; Song et al., 2003; Wang et al., 2003; Ziegelbauer et al., 2006). We also identified several novel RTA direct binding sites in this study (Table 1).
To confirm the putative RTA binding sites identified in the ChIP-on-chip experiments, we performed ChIP-qPCR analyses. The primers used in the qPCR are listed in Table 2. As shown in Fig. 3, ChIP-qPCR confirmed the RTA binding sites of ORF8, ORFK4.1, OriLyt-L, ORFK5, PAN RNAs, ORF16, ORF29, ORF45, ORFK8, ORFK10.1, ORF59, ORFK12, OriLyt-R, miRNAs, ORF71, ORF74, and ORFK15. RTA was not shown to bind to the promoter regions of ORFK2 and ORF57, indicating that these two genes may not be the direct targets of RTA as suggested by previous studies (Byun et al., 2002; Chang et al., 2005b; Ziegelbauer et al., 2006). However, it remains possible that the failure to detect the RTA binding sites in the ORFK2 and ORF57 promoters was due to sub-optimal condition in our analyses. As negative controls, RTA did not bind to the promoter regions of β-actin and CDK6.
We identified six RTA direct binding sites in the latent locus (Fig. 4). This region encodes at least 11 KSHV transcripts and all the 12 KSHV microRNAs in addition to OriLyt-R. These results indicate that RTA might be heavily involved in regulating the expression of these latent transcripts. Previous studies have identified a RTA RE between the LTi and ORFK14 transcriptional start sites (Matsumura et al., 2005; Staudt and Dittmer, 2006). The RTA binding site 1 falls in this region (Fig. 4). Thus, this site might represent the reported putative RTA RE. Further investigation is needed to confirm the function of this site.
We further investigated the transcriptional regulation of the upstream and intron region of the ORFK15 gene by RTA, and determined the RTA direct binding sequences in this locus. KSHV ORFK15 type P and M forms are highly diverged with less than 30% amino acid homology to one another (Choi et al., 2000; Glenn et al., 1999; Poole et al., 1999). Our ChIP-on-chip data indicated that RTA binds to the upstream of the ORFK15P gene from the BCBL-1 cells (Fig. 1). We constructed ORFK15 upstream/intron-luciferase reporters from BCBL-1 (type P) and BC-3 (type M) cells (Fig. 5A). The ORFK15 gene is located next to the right TR of the KSHV genome, and is transcribed in a leftward orientation (Fig. 5A). Reporter assay indicated that RTA activates both constructs (Fig. 5B) and the transactivation of these two genes by RTA was dose-dependent (Fig. 5C and 5D).
To determine the location of RTA RE in the ORFK15 upstream/intron region, the K15P reporter construct was further used to generate a series of deletion mutants (Fig. 6A and 6B). Surprisingly, luciferase reporter assay showed that all the deletion mutants were responsive to RTA transactivation (Fig. 6C), indicating that the RTA RE must exist within the minimal sequence of D6, a 106-bp deletion mutant located in the first ORFK15 intron. Next, we generated five overlapping luciferase reporter clones covering this 106-bp sequence region by PCR-based cloning strategy (Fig. 7A). Each clone is 30 bp in length with 10 bp overlapping sequence. Reporter assay indicated that, similar to the full-length (600-bp) and D6 (106-bp) constructs, the RE1 construct was responsive to RTA transactivation (Fig. 7B). The sequence of RE1 is shown in Fig. 7C. Computational analysis of this sequence revealed a palindromic structure with two tandem-repeats (bold sequence in Fig. 7C).
To test whether RTA could bind to the RE1 sequence in vitro, we performed a competitive EMSA assay (Fig. 8), using AP-1 consensus sequence, RE1, and the putative RTA binding sequence (RE11) as probes. We found that all the labeled probes could form complexes with nuclear proteins from the 293T cells transiently transfected with the p3FLAG-RTA plasmid. However, the DNA-protein complex detected by the AP-1 probe could not be eliminated by addition of excess unlabeled RE1 (lane1) and RE2 (lane 2) probes, indicating that the AP-1-protein complex was not related to the RE1 and RE2 sequences. In contrast, the DNA-protein complex formed by RE1 (lanes 4–6) or RE11 (lane 7–9) probes could be eliminated by addition of excess unlabeled RE1 (lanes 4 and 7) but not RE2 (lanes 5 and 8). To examine whether this DNA-protein complex is due to the RTA binding, we performed a supershift assay with the RE1 probe and an antibody to FLAG. As shown in Fig. 8C, the DNA-protein complex formed by the RE1 probe could be eliminated by excess unlabeled RE1 (lane 4) and RE11 (lane 6) but not by AP1 (lane 3) and RE2 (lane 5). Additionally, this DNA-protein complex could be supershifted by specific antibodies against FLAG (lane 7) and RTA (lane 8), demonstrating that RTA specifically bound to the RE1 region in vitro (Fig. 8C).
To define the RTA RE sequence, we tested the RTA binding efficiency with several known RTA REs from previous reports (Fig. 9). Fig. 9B indicated that the complex formed by the RE1 probe could be eliminated by the excess amount of unlabeled RE11 (lane 3) and RE12 (lane 4), as well as by the known RTA REs previously identified in the promoters of ORFK2 (lane 8), ORFK8 (lane 9), PAN RNAs (lane 10), ORFK12 (lane 11), and partially ORF57 (lane 12), but not by the AP1 consensus sequence (lane 2) and RE2 (lane 5). Antibody against FLAG supershifted this complex (lane 7), though antibody against RTA only formed a weak supershift band (lane 6). The fact that RTA binds to ORFK2 and ORF57 promoter in vitro in EMSA (Fig. 9B) but no in vivo in ChIP-on-chip (Fig. 1 and and2)2) indicated that in vitro EMSA assay might not be ideal for examining the RTA direct binding sites.
Using Motif Discovery scan (MDscan) (Liu et al., 2002), we determined the RTA binding sequence in the ORFK15 region as a palindromic element with two tandem-repeats TTCCAGGAA TTCCTGGAA (Fig. 9C). This two tandem-repeats structure is similar to that of p53 consensus binding elements (el-Deiry et al., 1992), and may indicate the binding of RTA to DNA as a tetramer (Bu et al., 2007). Detailed analysis of other RTA binding sites identified the RTA consensus binding motif as TTCCAGGAT(N)0–16TTCCTGGGA, which contains two imperfect inverted tandem-repeats (Fig. 10A). The gap that separates the tandem repeats also varies with different RTA binding sites ranging from 0 to 16 bp. Interestingly, some of the tandem-repeats in the RTA REs contain only half of the tandem-repeat sequence (TTCC) (Fig. 10A). We further examined the sequences of other RTA binding sites that failed in the initial alignment analysis. The majority of these sites with the exception of RTA and ORFK5 identified in this study also contain at least half of the tandem-repeat sequence (Fig. 10B).
RTA is a critical switch gene for viral reactivation from latency. Delineation of the genome-wide locations of RTA binding sites in the KSHV genome in vivo is important for the comprehensive understanding of viral reactivation and gene regulation. However, this has turned out to be a difficult task because of the weak binding of RTA to its targets in vivo and the activation of full lytic replication and cell death following overexpression of RTA. Using a novel strategy, we were able to determine the full set of RTA direct binding sites in the KSHV genome. We designed our ChIP-on-chip experiments based on several considerations. The most important factor is the quality of the antibody used for the immunoprecipitation. Transcription factors generally are expressed at low levels in the living cells, and have weaker affinities for DNA than histone proteins. Therefore, ChIP application of transcription factors is particularly demanding since the antibody must be capable of recognizing the native protein as part of a cross-linked DNA-protein complex. We have tested both monoclonal and polyclonal antibodies against RTA from different groups (Okuno et al., 2002; Yang and Wood, 2007) as well as our own laboratory. Although these antibodies work well for other applications, they failed to enrich RTA-DNA complex sufficiently, and consequently, the signal-to-noise ratios were too low to make any meaningful direct microarray analysis of the RTA binding sites (data not shown). Among all the systems that we have tested, triple FLAG epitope and the corresponding anti-FLAG M2 antibody is the most sensitive antigen-antibody detection system. Detection of fusion proteins containing the 3xFLAG is 20–200 times more sensitive than other tags such as c-myc, 6xHis, GST or HA, and therefore, is ideal for the ChIP assays of proteins expressed at low levels in mammalian cells (Wu and Chiang, 2002). In this report, a stable BCBL-1 cell line expressing the DNA binding domain of RTA tagged with a 3FLAG tag was used in the ChIP-on-chip assay.
The advantages for using the DNA binding domain alone to identify the RTA direct binding sites are as follows: 1) Unlike structure proteins, constitutive expression of RTA leads to full viral lytic expression and cell death (Gradoville et al., 2000; Lukac et al., 1998), therefore, prevents the harvest of sufficient RTA-DNA complexes for the microarray analysis. 2) Previous reports have indicated that the RTA construct containing only the DNA binding domain maintains the DNA binding activity for the RTA REs albeit it is no longer capable of activating the expression of lytic genes (Lukac et al., 2001; Lukac et al., 1999; Song et al., 2001; Song et al., 2003), and therefore, can be used to establish a stable RTA expressing cell line. In fact, this construct has been used for studying the RTA binding sites by other groups (Bu et al., 2007; Ziegelbauer et al., 2006). 3) The use of DNA binding domain alone can reduce the confounding effects of cellular factors that cause indirect RTA binding.
While the use of RTA DNA binding domain is necessary to circumvent some of the technical difficulties associated with the ChIP assay, it also introduces additional confounding factors. 1) The RTA DNA binding domain does not fully reflect the action of the full-length molecule, which could lead to under- or over-representation of the RTA binding sites. 2) Although less likely than the full-length molecule, the RTA DNA binding domain could still bind to the RTA REs by partnering with other viral and cellular proteins. 3) The protein expression level in the stable expression condition might not reflect all the in vivo situations. Indeed, we have observed a wide dynamic range of protein expression levels of the wild-type RTA following lytic induction over time (data not shown). Furthermore, at a given time point following lytic induction, the protein expression levels of the wild-type RTA also vary vastly at the individual cell level. Obviously, it is not possible to examine the RTA DNA binding targets at all expression levels in vivo even if the ChIP approach works with the endogenous protein. Nevertheless, the stable expression approach at least captures some in vivo conditions, most likely at high RTA expression levels. Thus, when analyzing results obtained by different approaches including the ChIP assay that we used, it is necessary to consider their limitations, which could account for some of the discrepancies between the RTA REs identified in this study with those identified by other approaches.
In a previous report, Ziegelbauer and colleagues identified a total of 18 RTA binding sites using purified GST-50ΔSTAD fusion protein expressed in bacteria and an in vitro SELEX assay (Ziegelbauer et al., 2006). We performed our ChIP-on-chip experiments and data analysis without previous knowledge of this dataset in order to avoid any bias in analyzing our results. Of 18 RTA binding sites identified by Ziegelbauer and colleagues, 10 (56%) were detected in our ChIP-on-chip assay. Interestingly, both studies failed to confirm the binding of RTA to the promoter regions of ORFK2 and ORF57, suggesting that these two genes might not be the direct targets of RTA in vivo, albeit both promoters contains two halves of the tandem-repeats of the consensus RTA binding motif (Fig. 10A), and RTA bindings to these promoters were detectable in vitro (Fig. 9). In addition to the RTA binding sites identified by Ziegelbauer and colleagues, we identified several novel binding sites, including those in the promoter regions of ORFK4.1, ORF16, ORF45, the miRNA cluster, ORF74, and ORFK15, as well as the well-defined RTA binding sites in the PAN and ORFK12 promoters. The discrepancies of these two datasets of RTA direct binding sites might indicate the fundamental differences between the in vitro and in vivo binding sites of transcription factors. A recent study examined the RTA direct targets in a reporter assay through conditional induction of RTA nuclear localization in the presence of cycloheximide, and identified 8 downstream genes activated by RTA, including PAN RNAs, ORF57, ORF56, ORFK2, ORF37, ORFK14, ORFK9, and ORF52 (Bu, et al., 2008), of which only the sites of PAN RNAs and ORFK14 were identified in our study. However, it remains possible that RTA could activate its downstream genes by partnering with other preexisting viral and cellular nuclear proteins in this system, even in the presence of cycloheximide.
Previous studies have identified two OriLyts in the KSHV genome (AuCoin et al., 2004; AuCoin et al., 2002; Lin et al., 2003). Both RTA and ORFK8, along with other cellular factors are required for the origin-dependent DNA replication (Wang et al., 2004c; Wang et al., 2008; Wang et al., 2006). Our data clearly indicated that RTA intensively bound to both OriLyts (Fig. 1, ,2,2, and and3),3), thus supporting a pivotal role of RTA in viral DNA replication during viral reactivation. The identification of RTA binding sites on other IE genes such as ORFK4.1, ORFK5, ORF29b, ORF45, ORF48, and ORF50 itself further support the conclusion that RTA is a true lytic switch gene of KSHV.
KSHV ORFK15 is expressed at both viral latent and lytic phases as complex transcription patterns (Choi et al., 2000; Glenn et al., 1999; Poole et al., 1999). Our study identified the RTA binding site in the first intron of the ORFK15 gene from BCBL-1 cells (type P). Nevertheless, we could not rule out the possibility that there were additional RTA REs located in the ORFK15 promoter region as reported in another study (Wong and Damania, 2006). Through the binding to the first intron, RTA likely contributes to the transcription regulation of the ORFK15 gene. Further investigation is undergoing to determine the role of RTA in regulating the differential expression patterns of various ORFK15 transcripts.
RTA activates its downstream genes through two distinct mechanisms: direct binding through RTA REs and indirect binding through cellular cofactors. In this study, we identified a total of 19 RTA direct binding sites in the KSHV genome, and defined the RTA binding sequence in the first intron of ORFK15 gene as a palindromic element with two tandem repeats (TTCCAGGAA TTCCTGGAA). Sequence alignment with other RTA REs identified the RTA consensus binding motif as TTCCAGGAT(N)0–16TTCCTGGGA, which contains two imperfect inverted tandem repeats (Fig. 10A). Some RTA REs contain only one, two halves, or even just half of the tandem repeat sequence (TTCC) (Fig. 10B). The gap that separates the tandem repeats also has large variations, ranging from 0 to 16 bp. These results indicate that RTA might bind to the majority of the target genes in more complicated manners, most likely, through recruitments of other cellular or viral transcription factors.
Polyclonal antibody against RTA was a gift from Dr. Charles Wood (Yang and Wood, 2007). Mouse monoclonal antibody against RTA was a gift from Dr. Koichi Yamanishi (Okuno et al., 2002). An anti-FLAG M2 antibody and the corresponding agarose beads used in immunoprecipitation were purchased from Sigma Life Science (St. Louis, MO).
p3FLAG is a vector for stable mammalian expression of protein fused with a triple FLAG epitope on the N-terminal (Chen, 2006). The full-length RTA (691 aa) expression vector (pcDNA-50Full) was previously described (Chen et al., 2001). A full-length RTA with the 3FLAG tag (p3FLAG-RTA) was constructed in this study. A C-terminal-truncated RTA DNA-binding domain (aa 1–390) tagged with the 3FLAG epitope (p3FLAG-dRTA) was constructed for the ChIP-on-chip assay by PCR amplification using the pcDNA-50Full DNA as a template, and primers RTA-NF (5′GACAAGCTAGCGCAAGATGACAAGGGTAA3′) and RTA-XR (5′ TTGTTCTAGAGCAGGAGTGGACGCTGAC3′), which contain the NheI and XbaI sites (underlined).
The minimal TATA-promoter reporter plasmid pTAL-luc containing the firefly luciferase gene (ClonTech Laboratories, Inc., Mountain View, CA) was used for the construction of the ORFK15 upstream/intron or RTA RE reporter constructs. Serial deletion mutants of the ORFK15 upstream/intron reporter construct were generated using a deletion kit for kilo-sequences according to the instructions of the kit (Takara, Madison, MI). All constructs were sequenced and confirmed.
Human PEL cell line, BCBL-1, was grown in RPMI1640 supplemented with 10% fetal bovine serum (FBS). Human embryonic kidney 293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with 10% FBS. A stable BCBL-1 cell line expressing the RTA DNA binding domain tagged with 3FLAG epitope was established by transfection of BCBL-1 cells with the plasmid p3FLAG-dRTA followed by selection in 800 μg/ml G418 (Sigma) and single cell cloning. The resulting BCBL-1-dRTA cell line was shown to express the 3FLAG-RTA fusion protein by immunoblotting assay (data not shown). When necessary, cells were treated with 25 ng/ml of TPA (Sigma) to induce viral lytic replication.
Custom-designed KSHV whole-genome tiling microarray, which contains a total of 10,260 probes of 60-mer oligonucleotides, was designed mainly based on the type M BC-1 KSHV sequence (Accession No.: NC_003409) (Russo et al., 1996) and synthesized by Agilent Technologies (Santa Clara, CA). This DNA microarray containing 4 full sets of KSHV genome including the terminal repeat (TR) covers the sense and complimentary DNA strands with overlapping probes. In addition, this array contains partial KSHV type P sequence (Accession No: U85269) (Nicholas et al., 1997), representative ORFK1 sequences from various isolates, 332 human microRNA sequences, independent sets of KSHV microRNAs, consensus RTA binding sequences (PAN, ORF57, ORFK8, etc.), and sequences from cellular house-keeping genes serving as negative controls (Vandesompele et al., 2002).
BCBL-1-dRTA cells were induced with 25 ng/ml of TPA for 6 h to initiate KSHV lytic replication. A total of 5×108 cells were collected, washed three times with PBS, and cross-linked in 45 ml of PBS containing 1% formaldehyde for 10 min at room temperature with occasional inversion. The cross-linking reaction was quenched with 5 ml of 1.25 M glycine, and the cells were washed twice in ice-cold PBS. The cells were then suspended in 5 ml of ice-cold hypotonic buffer containing 10 mM Hepes at pH 7.9, 10 mM KCl, 1.5 mM MgCl2, 0.5 mM DTT. The cells were passed through a 21-gauge needle 10 times followed by incubation on ice for another 30 min to disrupt cell membranes. Nuclei were collected by spinning at 10,000 × g for 8 sec, washed once in the hypotonic buffer, and re-suspended in 5 ml of ChIP lysis buffer containing 50 mM Hepes at pH 7.9, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100 and 1xPIC (Protease inhibitor cocktails, Sigma). After incubation on ice for 30 min, the nuclear lysate was sonicated on ice for 10 min with a 10-sec burst/50-sec pause cycle to generate sheared DNA with an average size of 300 bp. The suspension was then centrifuged at 12,000 × g for 10 min at 4°C, and the supernatant volume was increased to 10 ml with lysis buffer. The FLAG-tagged protein-DNA complexes were immunoprecipitated by incubating with 400 μl of anti-FLAG M2 affinity agarose beads (Sigma) overnight at 4°C, and then centrifuging for 5 sec at 10,000 × g. The supernatant was kept as non-enriched control DNA. The agarose beads were washed three times sequentially in the ChIP lysis buffer, ChIP high-salt buffer (50 mM Hepes at pH 7.9, 500 mM NaCl, 1 mM EDTA, 1% Triton X-100), ChIP wash buffer (50 mM Hepes at pH 7.9, 250 mM LiCl, 1 mM EDTA, 1% Triton X-100), and then TE buffer (10 mM Tris at pH 8.0, 1 mM EDTA), all containing 1xPIC. The DNA-protein complexes were eluted from the beads by adding 400 μl of ChIP elution buffer containing 50 mM Tris at pH 8.0, 10 mM EDTA, 1% SDS into the beads followed by incubation at 65°C overnight with agitation. ChIP elution buffer at 400 μl was also added to 100 μl of the non-enriched DNA control. Both the enriched (immunoprecipitated) and non-enriched DNA samples were treated with 5 μl of RNase A at 10 mg/ml for 1 h at 37°C, and then with 20 μl of proteinase K at 20 mg/ml for 1 h at 50°C. The DNA samples were then extracted twice with phenol/chloroform, once with chloroform, precipitated with ethanol, and dissolved in 100 μl of TE.
Microarray probe labeling, hybridization, and scanning were performed according to Agilent’s protocol for mammalian ChIP-on-chip (Version 9.0) with some modifications. Briefly, ChIP-enriched or non-enriched DNA sample (5 μg each) was labeled directly with either Cy5-dCTP or Cy3-dCTP (GE Healthcare, Piscataway, NJ) without the PCR amplification step to reduce any potential amplification bias. Probe labeling was performed using the BioPrime Array CGH Genomic Labeling Kit (Invitrogen, Carlsbad, CA) in 75-μl reactions. The purified labeled samples were then mixed, and microarray hybridizations were performed at 65 °C for 2 days with gentle agitation in a SureHyb chamber and hybridization oven (Agilent). The hybridized arrays were then washed and scanned on a GenePix 4000A scanner (Agilent), and the images were analyzed with the Feature Extraction software version 9.1 (Agilent).
For data analysis, we assumed that all short chromatin intervals were subjected to a uniform probability of shearing during the sonication step, and that these short chromatin intervals were independent of each other. Since the array had a 60-mer resolution, the putative binding sites were identified by scanning regions that had a minimum of five consecutive peaks. Each peak was defined as a minimum of two-fold ChIP/Input ratio of enrichment. The length of five consecutive peaks is 300 bp, which is comparable to the original sheared immunoprecipitated DNA fragments.
BCBL-1-dRTA cells were treated with 25 ng/ml of TPA for 6 h before cross-linking and immunoprecipitated with the anti-FLAG M2 affinity agarose beads. ChIP-enriched DNA was dissolved in 30 μl of TE, and 2 μl each of ChIP-enriched and non-enriched DNA control were amplified by qPCR using the 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA). Primers used for the qPCR, some of which have been described in a previous study (Yoo et al., 2005), are listed in Table 2.
293T cells (107) transiently transfected with the p3FLAG-RTA plasmid were collected by low-speed centrifugation, washed three times with cold PBS, suspended in 1 ml of buffer A containing 10 mM Hepes at pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, and passed 10 times through a 27 1/2 gauge needle. Nuclei were collected by spinning at 10,000 × g for 8 sec, washed once with the buffer A, and suspended in 200 μl of cold buffer C containing 20 mM Hepes at pH 7.9, 25% glycerol, 420 mM KCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT, and 0.5 mM PMSF, and incubated on ice for 15 min. After addition of 200 μl of cold buffer D containing 20 mM Hepes at pH 7.9, 20% glycerol, 0.2 mM EDTA, 0.5 mM DTT, and 0.5 mM PMSF, the suspension was centrifuged at 12000 × g for 15 min at 4°C. The supernatant was transferred to a new tube, and the protein concentration was measured using the Bradford protein assay reagent (Bio-Rad, Hercules, CA) and stored at −70°C until use. EMSA reactions were performed by pre-incubating 5 μg of nuclear extract containing 2 μg of poly dI-dC, 2 μg of BSA and 100 pmol of unlabeled competitor oligonucleotide or 0.2 μg antibody where indicated, in the binding buffer containing 10 mM Hepes at pH 7.9, 100 mM KCl, 5 mM MgCl2 and 5% glycerol on ice for 30 min. Annealed oligonucleotide probes (2 pmol) were labeled with [γ-32P]ATP and added to the reactions. The reactions were incubated at room temperature for 1 h, and the complexes were resolved on 4% non-denaturing acrylamide gels.
293T cells were co-transfected with 1 μg of luciferase reporter constructs and the RTA expression vector or control vector using SuperFect transfection reagent (Qiagen, Valencia, CA). Total DNA in each transfection was equalized with the empty pcDNA3.1 vector. Transfection efficiency was normalized by cotransfection with 1 μg of the pCMV-β-gal internal control plasmid (Promega, Madison, WI). Cells were lysed and collected 24 h post-transfection using a luciferase assay lysis buffer (Promega). Luciferase activity was then measured with a Veritus microplate luminometer (Turner Biosystem, Sunnyvale, CA). All the reporter assays were carried out three times, each with triplicates. Results are the averages of data with standard deviations from one representative experiments.
This work is supported by an American Cancer Society Research Scholar Grant (#RSG-04-195), and grants from the National Institute of Health (CA096512, CA124332, CA119889 and DE017333) to S-J Gao. We thank Dr. Charles Wood at the University of Nebraska, Lincoln, Nebraska, and Dr. Koichi Yamanishi at Osaka University Medical School, Osaka, Japan for providing the RTA antibodies. We thank Drs. Kenneth Izumi and Anthony Griffiths for their valuable suggestions to this work, and appreciate the technical assistances from Daniel White and members of Dr. Gao’s Laboratory.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.