Role of ND10 nuclear bodies in the chromatin repression of HSV-1

Unlike other DNA viruses that can pack their virion genomes into minichromosomes to avoid being recognized as foreign DNA [4, 5], HSV-1 does not contain histones or histone-like proteins in the capsid [6]. Instead, early studies showed that HSV-1 had polyamines in the virion to neutralize the negative charges on viral DNA [7]. HSV-1 DNA is tightly confined within the capsid and endures a pressure of about 20 atmospheres [8]. This tremendous pressure drives a quick ejection of viral DNA into the cell nucleus upon infection [9]. The sudden injection of mostly naked viral DNA inevitably triggers an immediate alarm of foreign invasion. Indubitably the infected cell mobilizes all defensive forces and attempts to silence the viral DNA right away. One major host cell defense against the incoming viral DNA is the mobilization of histones and histone associated repressors to force the viral DNA into chromatin repression. Although the exact mechanism of how cells mobilize the histone pool is not clear, it has been shown that histones are more mobile after the HSV-1 infection [10‐12]. At least partial or unstable nucleosomes are formed in the lytic infection, albeit unevenly across the viral genome [13, 14].

The inhibitory effects of chromatin formation on viral gene expression are reflected in several lines of evidence. First, HSV-1 DNA was found to associate with histone H3 as early as 1 h post infection [6]. Early in infection, more histone association was found at β (delayed early) and γ (late) gene promoters than that of α gene promoters [6, 15]. Viral proteins such as VP16 and ICP0 are responsible for the removal or remodeling of the histones, which leads to the activation of viral gene expression (see below). The second observation that chromatin formation represses HSV-1 expression is the fact that inhibitors targeting chromatin deactivating enzymes, such as histone deacetylases (HDACs) [16, 17], promoted viral gene expression and DNA replication for a recombinant HSV-1 containing a growth defect [18], indicating the significance of the reversal of histone deacetylation in lytic HSV-1 infection. The third evidence is the demonstration of functional interactions between HSV-1 proteins and chromatin repressors during the infection. For example, a nuclear repressor complex REST/CoREST/LSD1/HDAC was disrupted during the HSV-1 infection by ICP0, a viral gene transactivator that enhances downstream gene expression without any sequence specificity (for reviews, see [19, 20]), and then later in infection, CoREST and HDAC1 were translocated into the cytoplasm [21]. A dominant negative CoREST interfering the CoREST-HDAC1 interaction partially rescued viral replication in the absence of ICP0 [22], whereas the ICP0 mutant virus defective in CoREST binding showed a growth defect and failed to hyperacetylate the histone H3 and H4 bound to the quiescent DNA in a superinfection assay [23, 24]. ICP0 also interacts with class II HDACs and the interaction is responsible for the relief of HDAC5-mediated gene repression [25]. ICP0 has a comprehensive role in both histone removal and histone acetylation in lytic infection [26]. It is capable of promoting a two-step heterochromatin removal from the ICP8 promoter [27]. Interestingly, LSD1, the histone demethylase in the REST/CoREST/LSD1/HDAC complex, is required for early gene expression in both lytic and latent HSV-1 infection [28]. Since histone methylation status (mono-, di- or tri- methylation) plays different roles in gene activation or repression [29], how the inhibition of LSD1 changes histone methylation and how different methylation status regulates the initial HSV infection are not yet clear. Another viral protein, tegument protein VP16, is responsible for excluding histones from α gene promoters upon the viral DNA entry [15]. VP16 recruits host cell factor 1 (HCF-1) and Oct-1 to stimulate α promoter activity. This immediate counteraction against chromatin repression allows the expression of α genes, including ICP0 which further de-represses the HSV-1 chromatin on β and γ promoters [20, 26], and ensures a full blown infection. In consistent with these observations, newly synthesized viral DNA is not chromatinized and is well associated with RNA polymerase II and transcription factors [6, 30].

In latent HSV infection, all viral genes are turned off except the latency-associated transcript (LAT), which is actively transcribed throughout the latency [31]. HSV DNA exists as episomes in latently infected sensory neurons [32, 33]. The viral DNA itself is not extensively methylated [34, 35], but the typical nucleosome-protected DNA pattern is readily observed for latent DNA in micrococcal nuclease assays, suggesting that latent viral DNA is packed into a nucleosomal structure like the host chromatin [36]. The viral latent chromatin is also regulated in a mechanism similar to that of the host chromatin. For example, the histone H3K9 and H3K14 at LAT promoter are hyperacetylated whereas they are hypoacetylated at lytic promoters, consistent with the fact that LAT is the only transcript made in latency while all other viral expressions are repressed [35]. Furthermore, injection of HDAC inhibitor into latently infected mice induces reactivation [37, 38], whereas application of an inhibitor specifically blocking the demethylation of the repressive marker H3K27me3 reduces reactivation in cultured neurons [39]. These findings suggest that changes of histone modification status may control the switch between latency and reactivation.

Interestingly, part of the LAT transcript is complementary to the C-terminal region of ICP0, the powerful heterochromatin remover that stimulates lytic infection. The promoters for LAT and ICP0 are only about 5 kb away [1]. To set apart the euchromatin of LAT promoter/enhancer region from the heterochromatin of ICP0 promoter region in the latent infection, HSV evolves to contain chromatin insulator CTCCC repeats within the LAT intron, which recruits the CTCF protein and marks the boundary between the euchromatin and heterochromatin of latent HSV DNA [40].

Although in latent infection HSV genome DNA is clearly packed into chromatin and HSV genes are fully regulated through the host epigenetic machineries, the processes of how chromatinization is initiated to establish latency and how chromatin repression is released to reactivate from latency are largely unknown. The expression of LAT is very important for the HSV-1 latency, which is reflected in two lines of evidence: (i) Deletion of LAT expression resulted in a reduction of histone H3K9me2 and H3K27me3, markers for inactive heterochromatin, and an increase of histone H3K4me2, a marker for active euchromatin, at the lytic promoters, indicating the participation of LAT in regulating chromatinization at lytic promoters of HSV-1 [41, 42]; and (ii) several microRNAs derived from the LAT region inhibited the expression of ICP4 and ICP0, the two major gene transactivators for lytic infection, suggesting that LAT also regulates lytic expression on posttranscriptional level [43]. More interestingly, the lack of LAT expression did not eliminate the presence of latent viral DNA in mouse ganglia [38, 41, 44], but greatly reduced the rate of spontaneous reactivation in infected animals [38, 45]. These results demonstrate that LAT expression is not required in latency establishment but it is essential for latency reactivation. Although it is still unclear how LAT is involved in stimulating the reactivation, it is conceivable to postulate that LAT may monitor the basal level expression of lytic genes by modulating the chromatin status at lytic promoters and by controlling the leaky transcription via microRNAs. Consequently, LAT works to fine-tune the balance between latency and reactivation.

Several chromatin repressor complexes are shown to be important for either latency establishment or latency reactivation. One of them is the aforementioned REST/CoREST/LSD1/HDAC complex. Specific inhibition of LSD1 blocked HSV-1 reactivation from latency [28, 46]. Another component from this complex, REST, plays a critical role in latency establishment. Overexpression of wild type REST in infected neuron caused a reduction in reactivation from the explanted ganglia [47], whereas overexpression of a dominant-negative REST capable of binding to DNA but not to the other complex components led to a failure in latency establishment [48]. A second repressor complex has been implicated in latency regulation is the polycomb group proteins including polycomb repressor complexes (PRC) 1 and 2 [49, 50]. PRC1 component Bmi1 and PRC2 component Suz12 were each found on lytic promoters during latency by two research groups [49, 50], but the results were not reconciled with each other. A recent report showed that histone phosphorylation by JNK pathway in the presence of repressive methylation also contributed to the initiation of latency reactivation [51]. How these different pathways cooperate to control the reactivation switch is still largely unknown.