Histone acetylation/deacetylation plays a pivotal role in the regulation of HCMV gene expression 
. Analysis of proteins captured with pHDAC1gfp by mass spectrometry predicts that HDAC1 associates with a substantial number of cellular and viral proteins during infection ( and Table S1
). HDAC1 exists in different complexes, and the capture assay identified putative interacting proteins that are found within the NuRD, coREST, and Sin3 complexes. We performed reciprocal co-immunoprecipitations to confirm the interaction of pUL38 and pUL29/28 with HDAC1 and to verify that the viral proteins interact with the MTA2 subunit of the NuRD complex (). A relatively lower, but detectable, amount of pUL29/28 was coimmunoprecipitated with a component of the Sin3 complex (), consistent with the view that the viral protein exists in multiple HDAC1-containing complexes within the infected cell, as predicted by the mass spectrometry analysis.
In addition to these HCMV proteins, we detected peptides corresponding to the viral kinase pUL97 in the HDAC1 capture experiment (Table S1
). Although we did not validate this predicted interaction, other herpesvirus-coded kinases have been reported to hyperphosphorylate HDAC1, including the herpes simplex type 1 US3 kinase 
and the varicella zoster virus ORF66p kinase 
. Consequently, the possible HDAC1-pUL97 interaction could reflect a kinase-substrate interaction. We failed to detect IE1 or IE2 peptides in the HDAC1 capture experiment. These immediate-early proteins regulate HCMV gene expression, in part, by altering histone acetylation 
, and IE2 has been previously reported to interact with HDAC1 
in transfected cells. Our failure to observe IE2 may be in part due to the isolation conditions (e.g. lysis buffer) or poor ionization of corresponding peptides during MS analysis, although we have detected IE2 in other MS analyses that were focused on the immediate-early protein (data not shown). Alternatively, the proteins might interact at the start of infection but not at 24 hpi, when we performed the capture assay.
We have previously shown that pUL29/28 and pUL38 are needed for efficient HCMV replication 
. Our current results demonstrate that these proteins target the NuRD complex; but, rather than antagonize NuRD (), the viral proteins modulate its activity to support HCMV replication. We observed an ~10-fold reduction in both HCMV RNA and DNA accumulation in cells expressing shRNAs directed against two different components of the NuRD complex, CHD4 and RBBP4 (). A pUL29/28-deficient virus expressed ~2-fold-reduced levels of immediate-early RNAs, but accumulated ~10-fold less DNA when compared to wild-type virus 
. The more severe effect on viral RNA accumulation observed for a NuRD as compared to a pUL29/28 deficiency suggests that NuRD likely plays a more complex role than simply regulating the MIEP in HCMV replication. The relatively small drop in pUL29/28-induced MIEP reporter activity upon disruption of NuRD components supports this view ().
Our data demonstrate that components of the NuRD complex, with which pUL29/28 interacts, are needed for efficient replication. This is in contrast to the consequences of drug-mediated, global HDAC inhibition during infection, which enhances immediate-early gene expression 
, overcomes hDaxx-mediated repression on the MIEP 
and rescues an IE1-deficient virus 
. We observed similar increases in IE1 RNA expression in both pUL29-deficient and wild-type virus in the presence of TSA (), suggesting that the mechanism to overcome hDaxx-mediated repression on the MIEP is still in place in the absence of pUL29/28.
pUL38 interacts with NuRD components only in the presence of pUL29/28 (); and pUL29/28, but not pUL38, can activate the MIEP in a reporter assay (). The NuRD subunit-dependent () ability of pUL29/28 to modulate the MIEP in the absence of additional viral proteins argues that it plays a significant role in modulation of the NuRD complex during infection. Further, pUL29/28 failed to significantly activate reporters controlled by ISRE or NF-κB elements (), supporting the interpretation that that pUL29/28 modulates NuRD activity to generate promoter-specific outcomes.
How does pUL29/28 modulate the NuRD complex to favor expression of the MIEP? The HCMV genome undergoes chromatin remodeling and changes in acetylation during the course of infection through unknown mechanisms 
. The NuRD complex contains both HDAC and nucleosome remodeling activity 
and participates in both activating and repressing transcription 
. As discussed above, global inhibition of HDAC activity enhances viral replication, raising the possibility that NuRD HDAC activity is not important for HCMV. Perhaps pUL29/28 modulates the activity of the MIEP and accumulation of immediate-early RNAs by influencing nucleosome positioning through its interaction with NuRD. In this regard, we observed an interaction of the MTA2 component of NuRD and pUL29/28 with the MIEP (). There is precedent for a role of nucleosome remodeling in HCMV gene expression. The p150 subunit of the chromatin assembly factor 1 (CAF1) is localized to HCMV replication centers during viral infection 
, and it has been shown to influence MIEP activation 
. CAF1 consists of three subunits, p150, p60 and p48 (RBBP4), and functions in nucleosome assembly 
and p150 has been observed in remodeling complexes 
. Alternatively, a complex of pUL29/28 and NuRD may act globally to alter the nuclear environment. The NuRD complex deacetylates both cellular proteins and histones at cellular promoters 
. Consistent with a role at sites other than the HCMV genome, pUL29/28 and pUL38 are not restricted to viral replication centers during the course of infection () 
. The NuRD complex regulates expression of several cellular genes, including snail 
, wnt4 
, and brca1 
. Interestingly, expression of Snail
are rapidly down regulated during HCMV infection 
. Beyond histones, NuRD also has been demonstrated to directly deacetylate transcription factors, including p53 and HIF-1α 
. Casavant et al
have argued that p53 is necessary for efficient HCMV replication, influencing viral gene expression 
Previously, we have demonstrated that HCMV pUL38 also interacts with and antagonizes the TSC complex during infection 
. This association is clearly distinct from its association with the NuRD complex, because pUL29/28 is not associated with TSC2 (). Further, a lower molecular weight isoform of pUL38 binds pUL29/28 and MTA2 than interacts with TSC2 (data not shown). The different forms of pUL38 have been attributed to alternative usage of ATG start codons 
. In addition, the interaction of pUL38-pUL29/28 with NuRD subunits is evident by 6 hpi () within the nucleus (), while pUL38-TSC complex is detected after 24 hpi and predominantly within the cytoplasm 
. The consequences of the pUL38 interaction with NuRD subunits remains unclear. However, our studies suggest that the contribution of NuRD to replication is complex, and we speculate that pUL38 is important in regulating NuRD. Both TSC and NuRD complexes are targets of the human papillomavirus (HPV) E6 and E7 proteins 
. This conservation in cell targets among different viruses argues for their importance, and consequently for a central role of pUL29/28 and pUL38 in viral pathogenesis.