Depletion of CoREST Does Not Improve the Replication of ICP0 Null Mutant Herpes Simplex Virus Type 1 ^▿

Natural infections of individuals by herpes simplex virus type 1 (HSV-1) are characterized by the establishment of latency in sensory neurons, interspersed by episodes of active infection at the periphery (for reviews, see reference 19). The balance between latent and lytic infection is regulated by both cellular and viral factors, for which the viral immediate-early (IE) protein ICP0 is of significant importance (2, 4, 6, 18). ICP0 is a complex protein for which a large number of functions have been reported (reviewed in references 4, 5, 18, and 22). It is a member of the RING finger class of E3 ubiquitin ligases, and this activity is known to be essential for its core biological functions (1, 3, 6, 10, 26). However, a number of other functions have been proposed for ICP0, of which the most topical is disruption of a cellular corepressor complex, which includes the proteins CoREST, REST, and histone deacetylases 1 and 2 (HDAC1/2) (14-17).

ICP0 contains a serine- and alanine-rich sequence between residues 550 and 600 that has similarity to a region in CoREST, and based on this information, it was hypothesized that ICP0 might disrupt the CoREST/REST complex (14). This was indeed observed at mid- to late times of HSV-1 infection (14) and further worked mapped sequences of ICP0 (which do not coincide with the original sequence of interest) that are required for an interaction with CoREST (16). Evidence that this interaction might contribute to the biological functions of ICP0 came from improved (but not fully complemented) yields in single-step growth curves of a recombinant virus expressing a fragment of CoREST (predicted to act in a dominant-negative manner) in place of ICP0 (16) and from a slight decrease in replication at intermediate times of a mutant virus with a lesion in the CoREST interaction region of ICP0 (17). On the basis of these observations, it was proposed that disruption of the CoREST/REST complex is a major function of ICP0 that underlies its core biological properties. However, a number of observations do not fit easily with this hypothesis. For example, ICP0 null mutant HSV-1 exhibits severely compromised gene expression from the earliest stages of infection (6), whereas the disruption of the CoREST/REST complex was observed only at mid- to late times and also occurred, albeit less efficiently, during ICP0 null mutant infection (14). Furthermore, the proposed CoREST interaction region of ICP0 is within a region where several other functions have been described, and whereas complete deletion of this region compromises ICP0 function, this mutant still functions about 2 orders of magnitude more efficiently than a RING finger mutant in assays of both lytic infection and reactivation from quiescence (10).

An RNA interference approach was adopted to test the proposal that CoREST is involved in regulating ICP0 null mutant HSV-1 infection. A clear prediction of this hypothesis is that depletion of CoREST should weaken cell-mediated repression of the mutant virus gene expression and replication. In consequence, the plaque-forming efficiency and growth yields of the mutant virus should be increased in CoREST-depleted cells. Lentiviruses expressing control (shNeg) and candidate anti-CoREST shRNAs (purchased from Sigma-Aldrich) were used to transduce HepaRG cells, in which ICP0 is required for efficient HSV-1 replication, using cell culture and transduction methods described previously (9, 10). Two short hairpin RNAs (shRNAs; sh486 and sh902) gave partial depletion, while another (sh958) resulted in a very substantial decrease in CoREST expression (Fig. 1A). The coding strand sequences of sh486, sh902, and sh558 are TTGGATGAATACATTGCCAT, GTCTTATTTGAGCAAGCCTTT, and GTACAAGCCATCAGGAAATAT, respectively, starting at nucleotides 435, 592, and 1174 of the CoREST cDNA (taking the A of the initiating ATG as nucleotide 1). The negative-control shRNA sequence (shNeg; coding strand sequence TTATCGCGCATATCACGCG) was designed so that it does not target any human or herpesvirus gene. The plaque-forming efficiencies of neither the wild type (wt) nor ICP0 null mutant HSV-1 strains were affected by depletion of CoREST (Fig. 1B and C). Single-step growth curves of ICP0 null mutant virus in shNeg and sh958 cells showed no significant effect of CoREST depletion (Fig. 1D). Since the ICP0 null mutant phenotype is most pronounced in low-multiplicity infections, the yields of ICP0 null mutant virus were investigated in shNeg and sh958 cells at 24 h after infection with decreasing virus inputs. Consistent with the other experiments, the yields of the mutant virus were not improved by depletion of CoREST (Fig. 1E). These data are incompatible with the hypothesis that CoREST plays an influential role in regulating HSV-1 infection, at least in this cell type.

The phenotype of ICP0 null mutant HSV-1 is heavily influenced by cell type, with the plaque-forming defect differing widely at about 1,000-fold in human fibroblasts (HFs), 5- to 100-fold in HeLa, Vero, and BHK cells, and absent in U2OS cells (3, 6, 23, 24, 27). HFs are the cell line of choice for many experiments because they are nontransformed, but they are more difficult to use for the production of transduced cell lines. HepaRG cells are much more amenable to this type of experimentation, and although they are a permanent cell line, they are minimally transformed (13). It is possible that CoREST has a regulatory role in HSV-1 infections in cell types other than HepaRG cells. However, it is unlikely that HepaRG cells differ substantially from HFs in this respect for the following reasons. The defect of ICP0 null mutant HSV-1 in HepaRG cells is of the order of 500-fold (9, 10) (only 2-fold down from that in HFs), and the burst sizes of the ICP0 null mutant in single-step growth curve experiments are similar in the two cell types (compare data shown in Fig. 1 with those shown in Fig. 6 and 7 of reference 12). Furthermore, mutations that affect ICP0 function do so to similar extents in HFs and HepaRG cells (10). Based on this evidence, it is a reasonable hypothesis that the same cellular factors are involved in restricting ICP0 null mutant HSV-1 in the two cell types. In addition, we have not detected any major differences between HepaRG cells and HFs in terms of plating efficiency or replication of wt HSV-1 or establishment of quiescent HSV-1 infections with defective HSV-1 mutants (10, 11) or in the use of many other assays of viral and cellular proteins in HSV-1-infected cells (8-12).

An event that occurs during the earliest stages of HSV-1 infection is the recruitment of the promyelocytic leukemia protein PML to sites that are closely associated with incoming HSV-1 genomes (7). Several other components of cellular nuclear substructures known as ND10 or PML nuclear bodies behave similarly (7). In a wt HSV-1 infection, this process is rapidly counteracted by ICP0 (7), and the evidence suggests that this function of ICP0 correlates well with its ability to stimulate viral gene expression and reactivation from quiescence (10). A plausible hypothesis is that the cellular response that leads to the recruitment of selected cellular proteins to sites associated with HSV-1 genomes is involved in a mechanism that represses transcription from foreign DNA and that by inhibiting this recruitment, ICP0 enables more efficient expression of viral genes. It has been proposed that CoREST, which is diffusely distributed in the nuclei of uninfected cells, is partially relocated to sites adjacent to PML in ICP0 null mutant HSV-1 infections of human embryo lung fibroblasts and is recruited into viral replication compartments as infection progresses (15). These data raised the possibility that CoREST might, like PML and other ND10 proteins, be recruited to sites associated with incoming HSV-1 genomes during the early stages of ICP0 null mutant infection. As observed previously (15), CoREST was diffusely distributed in the nuclei of HFs (Fig. 2, top row). However, compared to the control of PML recruitment (Fig. 2, second row), no evidence of recruitment of CoREST to viral genome foci in ICP0 null mutant infections was observed (Fig. 2, third row), even when recruitment of PML was evident (Fig. 2, bottom row). Essentially identical results were obtained in HepaRG cells (data not shown). Therefore, CoREST is not a representative of those proteins that accumulate markedly in viral genome-associated foci.

Previous work demonstrated that PML and a number of other ND10 component proteins are involved in the repression of HSV-1 gene expression that occurs in the absence of ICP0 (9, 11, 20, 21). These studies used lentivirus-mediated shRNA expression to make stable cells lines depleted of PML, Sp100, hDaxx, or ATRX. In all cases, a modest but reproducible increase in replication efficiency (up to 10-fold) of ICP0 null mutant HSV-1 was observed. Further increases in ICP0 null mutant plaque-forming efficiency were observed in cells depleted of both PML and Sp100, suggesting that the repressive effects of this group of proteins may be additive (9). Although these increases are insufficient to explain the full defect of ICP0 null mutant HSV-1, these data demonstrate that this technology is capable of revealing relevant cellular regulators of HSV-1 infection. The data presented here show that depletion of CoREST has no effect on the efficiency of plaque formation or replication of ICP0 null mutant HSV-1 infection in HepaRG cells, thereby putting into doubt the hypothesis that CoREST has a significant universal role in regulating HSV-1 infection.