Herpesviruses of all subgroups have been shown to establish latency in certain cells of a host organism. During the state of latency the viral genome persists as a circular, nonintegrating episome that is associated with cellular histones to form regular nucleosome-like structures (10
). Viral gene expression is tightly regulated via epigenetic mechanisms and allows only the transcription of a few genes supporting the latent state. The structural organization and modification of viral chromatin is thought to be essential for herpesvirus persistence. We have shown that distinct histone modification patterns are established during HVS latency in transformed human T cells (1
Here we investigated the binding of the cellular insulator protein CTCF on the HVS genome, with additional fine-mapping and characterization of CBS conserved in primate rhadinoviruses. We identified two distinct CBS that specifically bind CTCF at the right end of the L-DNA ( and ). The CBS are located within a homologous region to the major latency control region of the Kaposi's sarcoma-associated herpesvirus, a close relative of HVS, where such CBS have been previously reported (42
). The two CBS lie within the intergenic region of HVS orf73/orf74
and more specifically locate to the intron region of the orf73
/LANA promoter, which is spliced out in orf73
/LANA mRNA transcripts ( and A, B). The binding of CTCF in close proximity to the initiation site of major latency transcripts has also been observed for the HSV-1 latency-associated transcript (LAT) and EBV EBNA latency transcripts (3
). Furthermore, our bioinformatic analysis also detected CBS in the corresponding promoter region of the LANA homolog in the related primate viruses Rhesus rhadinovirus and herpesvirus ateles, and also the T-lymphotropic Alcelaphine herpesvirus 1 (13
), but not in murine gammaherpesvirus 68 (MHV68) (data not shown). The orf73
/LANA gene product confers viral persistence by attaching the HVS episomes to metaphase chromosomes during mitosis (5
) and controls the reactivation of the viral lytic cascade as it efficiently represses transcription of the R transactivator protein (37
). Therefore, the timely regulation of the orf73
/LANA is essential for the establishment and maintenance of HVS latency as well as for reactivation.
The CCCTC-binding factor, short CTCF, is a protein that is able to structurally organize chromatin loci. CTCF is a multivalent, 11 zinc finger phosphoprotein that has been recognized to function in gene regulation as a chromatin boundary element or an enhancer-blocking insulator (15
). The detection of CBS in the promoter region of orf73
/LANA prompted us to search for an influence of CTCF on the structural organization, promoter activity and expression of orf73
/LANA, in order to study how CTCF might contribute to the state of HVS latency in human T cells. We found that CTCF negatively regulated the orf73
/LANA promoter in gene reporter assays, while there were only minor effects on the promoter activity of orf74
(C). The increase of the orf73
/LANA promoter activity upon mutation of the first CTCF binding motif indicated that CTCF can potentially act as a repressor or enhancer blocker at this site (C). The role of CTCF as an enhancer blocking insulator protein has been shown at human loci like the H19/Igf2 imprinting control region, where binding of CTCF also correlated with conformational changes of the chromatin (25
). At the human c-myc
locus, the upstream enhancer activation of the proximal promoters can also be blocked by CTCF (31
). Moreover, involvement of CTCF in the regulation of HSV-1 latency has been reported, where CTCF binds downstream of the LAT promoter, thereby preventing the expression of ICP0, the HSV-1 lytic (re-)activator gene (3
We also sought to analyze the effect of such CBS in the contexts of viral lytic replication and latency. For this purpose, we generated different recombinant viruses with CBS mutations, as well as their respective revertant viruses to control for undetected second site mutations (). Interestingly, CTCF seems to have different binding arrangements at the two distinct CBS and it is tempting to speculate that complexes of CTCF with different partners and configurations can exert different functions (D). As we assumed that CTCF mainly functions as an epigenetic modifier of chromatin structures which are present particularly during herpesviral latency, we were not surprised by the finding that mutant viruses were highly similar in regard to their lytic replication (). This is in line with the hypothesis that the CTCF protein has little or no influence during the lytic replication cascade of HVS.
Remarkably, CBS turned out to be of great importance in regard to the maintenance of HVS episomes in latently infected human T cells. Human cord blood lymphocytes that were infected with recombinant HVS viruses harboring mutations or deletions in the second CBS or both showed reduced proliferation capacities (A). As the viral genome copy number per cell was also significantly decreased in cells infected with these mutant viruses, we concluded a direct involvement of CTCF in episome maintenance. Due to a loss of HVS genomes in T cells infected with such mutant viruses, the transformed state, which is a consequence of the viral oncoproteins StpC and Tip, cannot be maintained (B). This finding strongly argues for a contribution of the interaction between CTCF and the second CBS to the persistence of viral episomes. Such an attribution is conceivable, since it has been reported that CTCF can form loops in cis
and can bridge sequences located on different chromosomes in trans
). Recent publications have suggested that loop formation promoted by CTCF is also involved the establishment of the different latency types of EBV. It was shown that CTCF generates distinct chromatin architectures of EBV episomes, which results in variable promoter selection typical for the different gene expression patterns of EBV latency types (43
). Moreover, our data are in agreement with publications for KSHV, where it has been shown that CBS mutations within the KSHV genome resulted in a decrease of transfected viral episomes over time at a comparable magnitude to HVS (22
). This underlines the importance of CTCF for the maintenance of stable rhadinovirus genome copy number in latently infected cells. However, the results for KSHV were obtained after transient transfection of KSHV bacmids into 293T cells, while our study investigated HVS transformed, latently infected T cells. Furthermore, the studies for KSHV did not analyze the CBS separate from each other; we were able to provide first indications that the different CBS may account for different functions and that especially the second CBS is of importance with regard to stable episomal maintenance of HVS genomes.
Cohesins have recently been found to be involved in CTCF function with regard to the establishment of intrachromosomal loop formation (17
). These molecules primarily function in maintaining sister-chromatid cohesion and are necessary not only for correct chromosome segregation but also for facilitating the repair of damaged DNA by homologous recombination (28
). It was discovered that cohesins share a highly similar consensus binding sequence with CTCF and colocalize with CTCF at many sites (32
). We suspect that cohesins might contribute to CTCF function by stabilizing the attachment of the viral episomes to metaphase chromosomes during cell division. This hypothesis is supported by an investigation in KSHV, where it has been shown that cohesion subunits colocalize to the CBS found at the major latency control region and that these molecules are involved in complex gene regulation and chromatin organization (42
). The loss of viral episomes might be, in part, the result of a reduction of orf73
/LANA mRNA transcript detected in these T cells (C). However, taking into account that there is a corresponding reduction of viral genome copy number of mutant viruses in such human T cells, one could also argue that, per genome, comparable levels of transcription of the orf73
/LANA gene occur at the remaining genomes; the loss of viral genomes in mutant virus-infected cells thus might not be a consequence of the deregulation of orf73
/LANA. A simple siRNA experiment to knock down LANA cannot address this, as the polycistronic nature of the LANA transcript would also influence the viral cyclin and FLIP homologs; further studies using recombinant viruses might be required to shed light on the role of LANA versus CTCF.
Taking our data together, it is possible that CTCF, when bound at the orf73/LANA promoter region, is able to exert two different functions during HVS latency. This assumption is assisted by the finding that CTCF establishes two distinct binding arrangements at this gene region (D). First, CTCF can act as a regulator of transcription of the orf73/LANA gene expression. By influencing this promoter region CTCF might be involved in the regulation of the viral lytic cascade initiated by the R transactivator protein of orf50, which is regulated by the orf73/LANA gene product.
More importantly, we speculate that CTCF might contribute to the function of the orf73/LANA gene product in viral genome segregation. CTCF (and cohesins) may be involved in tightly tethering the viral episomes to metaphase chromosomes, thereby guaranteeing proper separation of HVS genomes to daughter cells. In the future, studies of viral chromatin conformation and virus-host chromatin interaction using recombinant viruses will probably shed more light on the role of CTCF in rhadinovirus biology.