A Protein Sequestering System Reveals Control of Cellular Programs by the Transcriptional Coactivator HCF-1* (cid:1)

The mammalian transcriptional coactivator HCF-1 is a critical component of the multiprotein herpes simplex virus immediate early gene enhancer core complex. The protein has also been implicated in basic cellular processes such as cell-cycle progression, transcriptional coactivation, and mRNA processing. Functions have been attributed to HCF-1 primarily from analyses of protein-protein interactions and from the cell-cycle-arrested phenotype of an HCF-1 tempera-ture-sensitive mutant. However, neither the mecha-nisms involved nor specific cellular transcriptional targets have been identified. As the protein is essential for cell viability and proliferation, a genetic system was developed to specifically sequester the nuclear factor in the cell cytoplasm in a regulated manner. This approach exhibits no significant cell toxicity yet clearly demonstrates the requirement of available nuclear HCF-1 for herpes simplex virus immediate early gene expression during productive infection. Addi-tionally, cellular transcriptional events were identified that contribute to understanding the functions ascribed to the protein and implicate the protein in events that impact the regulation of critical cellular processes. the HA tag and the actin coding sequences in pHA-actin. The coding regions from pHA-actin, pHA-TIF-Act, and pHA-TIFEH-Act ligated into pMEP4 (Invitrogen) to generate the respective metallothionein promoter controlled expression vectors. Cell Culture and Virus— Cell lines expressing the actin fusion pro- teins under the control of a metallothionein promoter were generated by transfection of HeLa cells with pMEP4 plasmids (Invitrogen) express- ing either the HA-TIF EHAY motif-actin fusion protein (T-Act), HA-TIF EHAY mutant motif-actin fusion protein (TEH-Act), or HA-actin fusion protein (Act). Hygromycin B-resistant colonies were pooled and used for all experiments. HSV-1 (F) viral stocks were produced and titered by infection of Vero cells according to standard procedures. Cells plated on glass coverslips and 4 (cid:4) M was for various defined time periods. The cells were fixed in 4% paraformaldehyde and permeabilized with 1% Triton X-100. The fusion proteins and HCF-1 were with anti-HA mono- clonal antibody (mHA.11, Covance) and anti-HCF-1 antibody (Ab2131 (3)), respectively, followed by the appropriate Alexafluor secondary antibody (Molecular Probes). Actin was using Texas Red-X phalloidin (Molecular Probes). Confocal were using a confocal laser scanning (Leica TCS-SP2) equipped with a (cid:2) PL Fluotar objective, a 488 horseradish peroxidase-conjugated sec- ondary Pico (Pierce) reagent. Protein quantitated GAPDH. cDNA h post-addi-tion .

The mammalian transcriptional coactivator HCF-1 is a critical component of the multiprotein herpes simplex virus immediate early gene enhancer core complex. The protein has also been implicated in basic cellular processes such as cell-cycle progression, transcriptional coactivation, and mRNA processing. Functions have been attributed to HCF-1 primarily from analyses of protein-protein interactions and from the cell-cycle-arrested phenotype of an HCF-1 temperature-sensitive mutant. However, neither the mechanisms involved nor specific cellular transcriptional targets have been identified. As the protein is essential for cell viability and proliferation, a genetic system was developed to specifically sequester the nuclear factor in the cell cytoplasm in a regulated manner. This approach exhibits no significant cell toxicity yet clearly demonstrates the requirement of available nuclear HCF-1 for herpes simplex virus immediate early gene expression during productive infection. Additionally, cellular transcriptional events were identified that contribute to understanding the functions ascribed to the protein and implicate the protein in events that impact the regulation of critical cellular processes.
HCF-1, 1 a cellular transcriptional coactivator, was originally identified by its requirement for the stable assembly of the herpes simplex virus (HSV) immediate early (IE) enhanceosome (1)(2)(3). In vitro, the multiprotein IE regulatory complex is nucleated by the recognition of the 5Ј sequences in the IE enhancer core element by Oct-1, a POU-homeo-domain protein (4 -7). The viral encoded IE transactivator protein, VP16 (␣TIF), recognizes the 3Ј sequences of the core and cooperatively interacts via specific recognition of the Oct-1 homeosubdomain (8 -11). HCF-1 interacts with VP16 with high affinity, resulting in a stable enhanceosome assembly.
HCF-1 is a ubiquitously expressed family of polypeptides derived from a single 230-kDa precursor protein by site-specific proteolytic processing (1,3,12). With a single notable exception, the protein is localized predominantly in the nucleus and has been described as a transcriptional coactivator for VP16 (13) and for cellular transcription factors such as GA-binding protein (GABP) (14), KROX20 (15), E2F4 (15), and LZIP (13,16). In addition, multiple roles have been proposed for HCF-1 in basic cellular processes such as cell-cycle progression (17)(18)(19), positive and negative transcriptional regulation (20,21), chromatin remodeling (21), and mRNA processing (22). These functions have been based primarily upon identified HCF-1protein interactions and upon the pleiotropic phenotypes of cells containing a temperature-sensitive mutation in the amino terminal kelch domain of the protein.
While various genetic systems could be utilized to address these issues, HCF-1 appears to be essential for cell viability, general transcription, and cell-cycle progression. Approaches such as the generation of dominant negative mutants, ts mutants, or RNA interference-mediated depletion have inherent problems for the analysis of essential proteins including the resulting cell toxicity, the time frame for phenotypic effects to become evident due to protein turnover rates, and the difficulties in determining the primary effects of the mutation or depletion. Therefore, a system was developed to inducibly alter the specific localization of HCF-1 by sequestering the protein in the cell cytoplasm. This approach is reflective of the specific cytoplasmic sequestering of HCF-1 in sensory neurons, the single exception to the nuclear localization pattern of the protein in most cell types (30). This system produces a regulated shift in the nuclear-cytoplasmic ratio of HCF-1, exhibits low toxicity, and allows for the delineation of events that are critically dependent upon the levels of nuclear HCF-1.
Using this approach, cytoplasmic sequestering of HCF-1 inhibited HSV IE gene expression, indicating the potential of the system for analysis of essential protein functions. In addition, transcriptional regulatory targets were identified that may account for the pleiotropic effects of the protein and also defines HCF-1 as a critical transcriptional component with impact on many basic regulatory processes.

EXPERIMENTAL PROCEDURES
Plasmids-pHA-actin was constructed by insertion of an HA epitope tag and start codon (ATG) into the NheI-BglII sites of pEGFP-actin (Clontech). pHA-TIF-Act and pHA-TIFEH were constructed by insertion of the coding sequences for the ␣TIF-HCF-1 interaction domain (AK-LDSYSSFTTSPSEAVMREHAYSRARTKNNYGSTIEGLLD) or mutant domain (AKLDSYSSFTTSPSEAVMRAAAYSRARTKNNYGSTIEGLLD) between the HA tag and the actin coding sequences in pHA-actin. The coding regions from pHA-actin, pHA-TIF-Act, and pHA-TIFEH-Act were subsequently ligated into pMEP4 (Invitrogen) to generate the respective metallothionein promoter controlled expression vectors.
Cell Culture and Virus-Cell lines expressing the actin fusion proteins under the control of a metallothionein promoter were generated by transfection of HeLa cells with pMEP4 plasmids (Invitrogen) expressing either the HA-TIF EHAY motif-actin fusion protein (T-Act), HA-TIF EHAY mutant motif-actin fusion protein (TEH-Act), or HA-actin fusion protein (Act). Hygromycin B-resistant colonies were pooled and used for all experiments. HSV-1 (F) viral stocks were produced and titered by infection of Vero cells according to standard procedures.
Immunofluorescence-Cells were plated on glass coverslips and 4 M CdSO 4 was added for various defined time periods. The cells were fixed in 4% paraformaldehyde and permeabilized with 1% Triton X-100. The expressed fusion proteins and HCF-1 were detected with anti-HA monoclonal antibody (mHA.11, Covance) and anti-HCF-1 antibody (Ab2131 (3)), respectively, followed by the appropriate Alexafluor secondary antibody (Molecular Probes). Actin was detected using Texas Red-X phalloidin (Molecular Probes). Confocal images were obtained using a confocal laser scanning microscope (Leica TCS-SP2) equipped with a 63ϫ PL Fluotar objective, a 488 nm argon-ion laser, and a 568 nm krypton-ion laser. The images from 488 and 568 nm channels were collected independently and merged using the accompanying software. For quantitation of the HCF-1 fluorescence, polygons were drawn around the entire cell and the nucleus using the accompanying Leica software and the fluorescence values of each were obtained. The HCF-1 fluorescence in the cell cytoplasm was determined by subtracting the nuclear fluorescence values from those of the entire cell.
RNA and cDNA Preparation-Cells were harvested 32 h post-addition of CdSO 4 . Total RNA was isolated using TRIzol reagent (Invitrogen) followed by purification with an RNeasy midi-column (Qiagen) as directed by the manufacturer. RNA integrity and concentrations were assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Massy, France) and a Ultrospec 3000 spectrophotometer (Amersham Biosciences), respectively. Fluorescently labeled first strand cDNAs were prepared by direct incorporation of fluorescent nucleotide analogs during a reverse transcription reaction consisting of either 20 (Cy3 labeling) or 40 g (Cy5 labeling) of total RNA, 2 g of oligo(dT) 20 primers, 0.25 mM each of dATP, dCTP, and dGTP, 0.025 mM dTTP, 10 mM dithiothreitol, 300 units of reverse transcriptase (Superscript II, Stratagene), and 2 nmol of either Cy3-or Cy5-dUTP (Amersham Biosciences). The cDNA probes were concentrated by filtering through Vivaspin 0.5-ml concentrators (Vivascience). For each microarray hybridization, the appropriate Cy3-and Cy5-labeled cDNA pools to be compared were mixed.
Microarray Hybridization and Data Analysis-Microarray slides representing 14,000 human genes (NIAID Microarray Facility) were incubated for 1 h at 42°C with prehybridization solution (1% bovine serum albumin, 0.1% SDS, 5ϫ SSC), washed two times in doubledistilled H 2 O, washed once in isopropyl alcohol, and dried by centrifugation at 600 rpm for 3 min. Each microarray slide received 66 l of hybridization solution containing the appropriate Cy3-and Cy5-labeled cDNA pools, 1 g of poly(d(A)) 40 -60 , 10 g of Cot-1 DNA, 4 g of yeast tRNA, and 33 l of hybridization solution (10ϫ SSC, 50% formamide, 0.2% SDS). The mixtures were applied by capillary action under a coverslip (LifterSlip, Erie Scientific Co., Portsmouth, NH) placed over the microarray, and the assembly was incubated in a humidified hybridization chamber for 16 h at 42°C. Microarray slides were washed twice in 1ϫ SSC, 0.05% SDS for 2 min and twice in 0.1ϫ SSC for 2 min and dried by centrifugation at 600 rpm for 3 min. The slides were immediately scanned and analyzed with a scanner/software package (Axon GenePix 4000B/GenePix Pro 4.0; Axon Instruments, Inc., Foster City, CA). For each microarray set, four replicate hybridizations were done including a dye-bias hybridization control.
Microarray Analysis-The average signal intensity and local background measurements for each spot on the array were analyzed using custom mAdb software (developed by Center for Information Technology/BioInformatics & Molecular Analysis Section in collaboration with NCI/Center for Cancer Research & NIAID and available at madb. niaid.nih.gov/index.shtml). The local background was subtracted from the mean intensity value of each spot on the array. Spots were considered negative and eliminated from further analysis if the values for both channels were less than a threshold value defined as one S.D. above the background. The two channels were normalized with respect to the median values for the remaining set of spots. The Cy5/Cy3 fluorescence ratios were calculated from the normalized values and reflect the differential value. Three sets of microarray analyses were performed: T-Act versus TEH-Act, T-Act versus Act, and Act versus MEP (control). To ensure reproducibility in the cDNA preparation and hybridization steps, all experimental data were collected from the hybridization of four independent cDNA preparations from each RNA preparation.
Luciferase Reporter Constructs and Assays-Luciferase reporter constructs contained the promoter domains for ELK1 (Ϫ500 to ϩ34, NT_011568), c/EBP␤ (Ϫ500 to ϩ66, NT_011362), Rous sarcoma virus long terminal repeat (Ϫ303 to ϩ23), and ubiquitin C (Ϫ333 to ϩ877) in pGL3-Basic (Promega). pLG3FOR2 (gift of C. J. Ciudad, University of Barcelona) contained the Sp1 promoter (Ϫ217 to ϩ44)-luciferase reporter (31). For transfections, duplicate plates seeded with 4 ϫ 10 4 tsBN67 cells per 1.9-cm 2 well (19) were incubated at 31 and 40°C for 24 h prior to transfection with DNA mixes containing various amounts of the promoter-luciferase reporter (25,50,100,200, and 400 ng). The dishes were incubated at the appropriate temperatures for 24 h prior to the measurement of luciferase activity (dual luciferase assay system, Promega). In all cases, the firefly luciferase activity was normalized to the activity of a cotransfected Renilla control (pRL-TK), and the results of duplicate transfections were averaged.

Development of a Cytoplasmic Sequestering
System-In contrast to the predominantly nuclear localization of HCF-1 in most cell types, the protein is uniquely sequestered in the cytoplasm of sensory neurons. This localization has been hypothesized to regulate the transcriptional functions of HCF-1 and to be a major controlling determinant in the establishment of and reactivation from the latent state for herpes simplex virus (30). The phenomenon provided a concept for the development of an inducible cytoplasmic sequestering system for the analysis of the cellular functions of HCF-1.
The sequestering system is dependent upon the inducible expression of an actin fusion protein containing a small, high affinity HCF-1 interaction motif derived from VP16 (7,25,26). As illustrated in Fig. 1, incorporation of the expressed actin fusion protein into cytoplasmic actin filaments would act as a high affinity anchor by binding HCF-1 and preventing the normal nuclear transport of the protein. In this manner, the nuclear to cytoplasmic ratio of the endogenous HCF-1 could be shifted in a regulated manner to allow modulation of HCF-1 localization without total ablation of the protein. Theoretically, this system would promote cell survival while affecting functions that are critically dependent upon the available nuclear pool of HCF-1.
In developing this system, the high affinity HCF-1 interaction motif from VP16 or a control containing mutations in the critical EHAY core of the motif were fused to actin under the control of an inducible metallothionein promoter ( Fig. 2A). Stable cell lines containing the transgenes were selected and analyzed by Western blot for the expression of the fusion proteins in the absence and presence of the inducer (CdSO 4 ). As shown in Fig. 2B, both the wild-type fusion protein (T-Act) and the fusion protein containing the EHAY core mutations (TEH-Act) were not detected in the absence of CdSO 4 and were expressed at equivalent levels in the presence of the inducer. Furthermore, both fusion proteins exhibited immunofluorescent patterns indicating that they had been incorporated into cellular actin filaments (Fig. 2C).
Cytoplasmic Sequestering of HCF-1-The effect of the expression of the actin fusion proteins upon the localization of endogenous HCF-1 was analyzed by quantitative immunofluorescent labeling (Fig. 3, top). T-Act, TEH-Act, and control cells containing the vector alone (MEP) were fixed at various times after induction with CdSO 4 . The cells were costained with anti-HA sera to detect the expressed actin fusion protein and anti-HCF-1 sera to determine the localization of HCF-1. As shown in Fig. 3 (top), cells expressing the T-Act fusion protein exhibited a clear visual increase in the level of cytoplasmic HCF-1 over time as compared with control MEP (Fig. 3, right panels) or TEH-Act cells (data not shown). Quantitative assessment of the cytoplasmic/nuclear ratio of HCF-1 in each cell line confirmed the visual interpretation and demonstrated that the cytoplasmic/nuclear ratio of HCF-1 increased in T-Act cells to nearly 1.4, while MEP and TEH-Act cells maintained a stable ratio ranging from 0.3 to 0.5 (Fig. 3, bottom right).
Sequestering of HCF-1 Inhibits the Expression of HSV IE Genes and Viral Replication-Induced T-Act cells exhibited a significant shift in the cytoplasmic to nuclear ratio of HCF-1. As in vitro data and studies using the ts-HCF-1 mutant have supported the model that HCF-1 is required for the regulated transcription of the HSV IE genes, T-Act, TEH-Act, and control cells expressing the HA-tagged actin (Act) were induced for 24 h and infected with HSV at either 0.01 or 0.1 pfu/cell. At 2 h post-infection, the expression of the viral IE protein ICP4 was determined by Western blot analysis of cell lysates. As shown in Fig. 4A (left panel), the level of ICP4 expression was signif-FIG. 1. Schematic representation of the sequestering system. The system utilizes the regulated expression of an actin fusion protein consisting of a high affinity domain that interacts with HCF-1. Upon induction, the expressed fusion protein is incorporated into the cytoplasmic actin filaments and sequesters the targeted HCF-1 in the cytoplasm.

FIG. 2. Actin fusion proteins are incorporated into cytoplasmic filaments.
A, the actin fusion proteins are schematically represented showing the sequences of the HCF-1 interaction motif derived from VP16 and the control point mutant. B, equivalent amounts of protein extracts derived from cells expressing the wild-type (T-Act) or mutant (TEH-Act) fusion protein were resolved in an 8 -16% Tris-glycine gel, transferred to membrane, and probed with anti-HA serum. U, uninduced; I, induced. C, cells expressing the wild-type or mutant fusion proteins were fixed and stained with anti-HA (green) to detect the fusion protein and Texas Red-X-phalloidin (red) to stain actin.
icantly reduced in induced T-Act cells relative to those in TEH-Act and Act cells (60% reduction, Fig. 4B). The levels of a second IE protein, ICP0, were similarly affected (data not shown). In contrast to this, infection of the T-Act cells at 0.1 pfu/cell resulted in expression of nearly wild-type levels of both IE proteins (Fig. 4A, right panel, ICP0 data not shown). The contrasting impacts of HCF-1 cytoplasmic sequestering on viral IE gene expression at 0.01 and 0.1 pfu/cell were likely due to the introduction of high levels of wild-type VP16 protein into the cells by the infecting viral particles (4) that would be expected to compete for the sequestered HCF-1 protein (see "Discussion").
The overall effect of sequestering HCF-1 on viral replication was also assessed. T-Act, TEH-Act, and Act cells were induced and infected with HSV at 0.01 or 0.1 pfu/cell. The resulting progeny virus was harvested at 17 h post-infection. As shown in Fig. 4C, infection of T-Act cells at 0.01 pfu/cell resulted in a significant decrease in the resulting viral yield relative to the TEH-Act and Act control lines. Surprisingly, a decrease in the viral yield was also observed after infection of T-Act cells at 0.1 pfu/cell. The less significant but reproducible decrease in viral yield at 0.1 pfu/cell in T-Act cells suggests that HCF-1 may also play additional roles later in the HSV viral lytic cycle (see "Discussion").
Cellular Regulatory Targets of HCF-1-While the role of HCF-1 in HSV IE gene expression has been well studied, little is known concerning the cellular HCF-1 functions or targets. The effect of sequestering of HCF-1 on HSV IE gene expression and viral replication demonstrated that this system could provide an approach to the analysis of HCF-1-dependent cellular transcriptional functions including those events that might account for the various phenotypic functions ascribed to the protein. Therefore, to identify targets that might be critically dependent upon HCF-1 nuclear function, human oligonucleotide microarrays were hybridized with mRNA isolated from the HCF-1 sequestering and control cell lines in three independent microarray sets: T-Act versus Act, T-Act versus TEH-Act, and Act versus MEP.
The resulting microarray raw data were subjected to a series of successive criteria to ensure the validity of the final values and the selected gene list. As listed in Table I, the oligonucleotide microarrays represented 14,000 known and predicted human open reading frames. Those genes that were detected in an array set and that were represented in at least three of the four replicate array hybridizations were selected for further analysis. Genes identified in the T-Act versus Act array set and the control Act versus MEP set were then filtered for those with differential values Յ0.3 or Ն3.0 in at least one replicate hybridization. As the TEH-Act mutant control is not a null mutant and does have a limited sequestering phenotype (refer to Fig. 3), this array set was filtered less stringently for those genes with differential values Յ0.5 or Ն1.5 in at least two replicate hybridizations. From each of these array sets, genes represented in the filtered Act versus MEP array set were deleted to eliminate any impact of the overexpression of actin or fluctuations in the actin cytoskeleton. For those genes remaining in the T-Act versus Act and T-Act versus TEH-Act array sets, the values of the replicate hybridizations were averaged, and the gene list and values were compared between each independent array set. Only those genes that were represented in both the T-Act versus Act and T-Act versus TEH-Act arrays and whose values were consistent in both array sets were selected. Finally, the average of the replicate hybridizations of the T-Act versus Act and T-Act versus TEH-Act for each gene was filtered for a differential of Յ0.5 or Ն1.5. The final 188 selected genes are listed in Tables II, III, and IV.  Table II lists the 144 genes with known functions. These genes are organized into functional categories with descriptions based upon extensive investigation of published literature (PubMed: www.ncbi.nlm.nih.gov) and on information derived from the gene ontology data base (geneontology.org/). Gene products with multiple functions or effects upon basic processes have an additional category indicated. Within each major category, the genes are arranged in ascending order according to the average differential array value from those genes suppressed to those genes induced by the sequestering of HCF-1. In addition to genes with known functions, 33 genes represent unknown or novel hypothetical protein coding regions (Table III). Table IV lists those genes that have appropriate differential values but that may be represented due to overexpression of the actin fusion protein or to effects on the actin cytoskeleton (11 genes). While these genes may, in fact, be affected by HCF-1 sequestering, they have been listed separately due to the relationship of the gene products to actin polymerization or cytoskeletal structure.
As shown in Table V, the total number of selected genes (188) encode a wide range of proteins that are involved in critical processes such as cell cycle, apoptosis, development, DNA replication, RNA processing, transcription, and signal transduction. Interestingly, there is a predominance of genes whose products are involved in gene expression (transcription, 30 genes, 15.8%) and signal transduction (14.2%).
Verification of Microarray Results-A stringent series of criteria and multiple array sets were used to delineate those genes affected by the cytoplasmic sequestering of HCF-1. As an independent assessment of the microarray results, equivalent amounts of protein extracts from induced T-Act, TEH-Act, and Act cells were subjected to quantitative Western blot analysis with antiserum to selected proteins from Table II (Fig. 5). Protein controls that did not exhibit a differential value in the microarray sets were used as internal controls for normalization (GAPDH). Of the nine proteins assayed, four (MCM7, CDC34, Sp1, and SKP2) showed no difference in the T-Act cell protein levels when compared with the levels exhibited in the TEH-Act or Act cells (data not shown), while five proteins (ELK1, MCM2, ARF6, c/EBP␤, and CASP9) exhibited a T-Act to Act and T-Act to TEH-Act ratio that was comparable with the values obtained in the microarray analysis (Fig. 5). As the steady state protein levels depend upon a number of additional factors such as protein turnover rates, the 55.5% correlation of the Western blot and the microarray analyses is striking. This correlation indicates that the microarray results are an accurate, if conservative, representation of the genes whose expression is affected by the sequestering of HCF-1.
Expression of HCF-1 Regulatory Targets in ts-HCF-1 Cells-To demonstrate a direct effect of HCF-1 on the expression of target genes identified in the arrays, tsBN67 cells were transfected with luciferase reporter constructs containing the promoter domains from several genes whose expression was affected by sequestering HCF-1 (ELK1, c/EBP␤, and Sp1). tsBN67 cells contain a single missense mutation (P134S) in the amino-terminal HCF-1 kelch domain and exhibit a loss of HCF-1 chromatin association at the nonpermissive temperature (19,32). As shown in Fig. 6, the expression of each of the selected luciferase reporters was significantly reduced at the nonpermissive temperature (ELK1, 2.9 -5.6-fold; c/EBP␤, 2.6 -3.3-fold; Sp1, 2.7-4.4-fold) relative to the activity at the permissive temperature. In contrast, the activity of the control promoters (Rous sarcoma virus long terminal repeat and ubiquitin C) were nearly equivalent at both temperatures. The results strongly indicate that HCF-1 is an important regulatory component for the identified target genes. DISCUSSION HCF-1-HCF-1 has been implicated in the control of basic processes such as cell-cycle progression, transcriptional regulation, and mRNA processing. The protein interacts with transcription activators, repressors, coactivators, and chromatin remodeling components, suggesting that the multiple functions ascribed to the protein may be due primarily to its roles in controlling cellular gene expression. However, aside from the functions of the protein in the regulation of HSV IE gene expression, the targets of HCF-1 regulation are unknown as the protein has not been identified in specific cellular transcriptional regulatory complexes. T-Act, TEH-Act, and Act cells were induced for 24 h and infected with HSV at 0.01 or 0.1 pfu/cell. A, cells were harvested at 2 h post-infection, and equivalent amounts of protein extracts were resolved in 8 -16% Tris-glycine gels, transferred to membrane, and probed with anti-ICP4 and anti-GAPDH sera. B, the levels of ICP4 expression in induced cells infected at 0.01 pfu/cell were quantitated, normalized to endogenous GAPDH levels, and graphically represented relative to that of the uninduced cell levels. C, cells were harvested at 17 h post-infection. The resulting viral yields were determined by titration and are graphically represented relative to the control Act cells.  Genes where the average values of the T-Act vs. Act and T-Act vs. TEH-Act arrays were Յ0.5 or Ն1.5 are listed with the average values, gene name, unigene number, and functional description. The genes are arranged by average value in ascending order within general functional categories. Secondary categories or functions are indicated where applicable. ϩ or Ϫ indicates the anticipated stimulation or suppression via wild-type HCF-1, respectively. GR, glucocorticoid receptor; ER, endoplasmic reticulum; snRNP, small nuclear ribonucleoprotein; IP 3 , inositol 1,4,5-trisphosphate; MAP, mitogen-activated protein; MAPK, MAP kinase; HGF, human growth factor; PIP, phosphatidylinositol phosphate; VEGF, vascular endothelial growth factor; ERK, extracellular signal-regulated kinase; bHLH, basic helix-loop-helix. Cytoplasmic Sequestering of HCF-1-As an approach to identifying HCF-1 cellular regulatory events and contributing to the understanding of the roles of the protein in control of gene expression and cell-cycle progression, a system was developed to specifically alter the cytoplasmic to nuclear equilibrium of the protein. This approach allowed for the controlled alteration of HCF-1 localization while maintaining cell viability. As such, this system should be applicable to the analysis of other essen- tial nuclear proteins. It should be noted that the system involves the targeted sequestering of HCF-1 in the cytoplasm by an expressed actin fusion protein that likely targets the newly synthesized, nuclear exported, or free pool of protein. Therefore, the sequestering is likely to affect functions of HCF-1 that depend upon this protein pool. The induction of the T-Act protein expression specifically resulted in cytoplasmic accumulation of HCF-1 in contrast to the expression of either the HA-tagged actin or the TEH-Act control proteins. Furthermore, infection of cells exhibiting HCF-1 sequestering with HSV at low pfu/cell resulted in defects in viral IE gene expression and a significantly reduced viral yield. These results both validated the system and demonstrated the requirement for HCF-1 in the initial stage of infection. In contrast, HSV infection at a higher pfu/cell did not show any significant decrease in viral IE expression. This is likely to be due to competition by the viral encapsulated VP16 for the sequestered HCF-1 since each viral particle contains ϳ900 VP16 molecules that are released into the cell cytoplasm upon infection (4). As a representative HSV preparation might contain 100 viral particles per infectious unit, a 0.1 pfu/cell infection could result in 9000 molecules of VP16 being introduced into the cell.
Interestingly, while the levels of the IE proteins were not affected by HCF-1 sequestering at the higher pfu/cell infection, the viral yield from these infections were consistently reduced. HCF-1 is relocalized to viral replication factories later in infection, and the protein may be required in later stages of the HSV lytic cycle for efficient replication. 2

Cellular Gene Expression
Affected by the Sequestering of HCF-1-The sequestering of HCF-1 was utilized to identify transcriptional targets including those that may account for the functions ascribed to the protein. Multiple microarray hybridizations were done with independent sets of mRNAs from T-Act, TEH-Act, Act, and MEP cell lines to ensure the accuracy of the resulting data. Furthermore, the data were subjected to a series of progressive stringent requirements. The final group of selected genes were represented in both the filtered T-Act versus Act and the T-Act versus TEH-Act array hybridizations with similar differential values. This process of selection would certainly eliminate some genes whose expressions were dependent upon HCF-1 by elimination of those that were not represented in both array sets at a minimum of 75% of the replicate hybridizations. More importantly, the TEH-Act control fusion protein is not a completely null mutant and does have some limited capacity to sequester HCF-1. Therefore, the requirement for a comparable differential in the T-Act versus Act and the T-Act versus TEH-Act would result in selection of only those genes that were critically dependent upon the normal cellular levels of nuclear HCF-1.
The resulting list of HCF-1-dependent genes include those with important functions in or control of basic cellular processes such as apoptosis, cell-cycle progression, DNA replication, DNA repair, RNA processing, signal transduction, and transcriptional regulation. Interestingly, there is a concentration of genes involved in signal transduction and gene expression suggesting that HCF-1 has very broad regulatory effects. The sequestering of HCF-1 results in both increases and decreases in gene expression, which is consistent with in vitro protein interaction studies showing the association of HCF-1 with both transcription activation and transcription repression components (5, 13-15, 20, 21, 29, 33).
In terms of the functions ascribed to HCF-1, the protein has an impact on global transcription and is required at multiple stages for cell-cycle progression. With respect to the regulation of components that affect general transcription, sequestering of HCF-1 impacted the expression of basal transcription factor components (i.e. TFIIH subunits p34 and p80, RNAPI general factor Rrn3), chromatin and chromatin remodeling components (i.e. histone 3, the KRAB-ZFP corepressor TRIM28/KAP-1, and a SWI/SNF complex binding regulator WDR9), and general transcription elongation and mRNA processing components (i.e. RNA cleavage and polyadenylation factor CPSF6, RNAPI transcription terminator TTF1, and the U2 small nuclear ribonucleoprotein subunit SF3A2/SF3a66). HCF-1 sequestering also impacted the expression of a number of important broad spectrum transcription factors (i.e. Sp1, NFI-C, c/EBP␤, CREB-H, JunB, TEAD1/TEF-1, ELK-1, and MLLT7). In general, sequestering HCF-1 resulted in altered expression of these components in a manner that would be expected to reduce general transcription by both RNAPII and RNAPI. Collectively, the data indicate that HCF-1 is essential for cell viability and the effects of depletion or alteration of HCF-1 would be expected to have global pleiotropic impacts on cellular gene expression.
Aside from the effects upon general factors, HCF-1 sequestering also alters the levels of a number of specific transcriptional regulators such as Tbx10, PBX1, MITF, Id3, Dlx6, Hes1, ZYX, ERF, and IRLB. A number of these factors are involved in cellular differentiation or developmental processes (i.e. Hes1, neuronal differentiation (34); Dlx6, jaw development (35)) suggesting that HCF-1 could be an important component in various developmental processes by controlling the levels of differentiation factors.
With respect to the requirement for HCF-1 in cell-cycle progression, the protein has been implicated in at least two distinct stages: G 0 -G 1 and mitotic exit (18,19). The protein may play some role by direct interaction with cell-cycle regulatory components (36,37). However, it has also been proposed that the HCF-1 requirement may be in part a reflection of the transcriptional regulatory functions of the protein as the ts-HCF-1 is released from cellular chromatin at the nonpermissive temperature (32). Sequestering of HCF-1 does result in alteration of the expression of transcription factors and other regulatory components that are critical to the cell-cycle progression. Decreases in Skp2 and CCNT2 along with increased expression of JunB, ATF3, and TIP-1 may promote delay or arrest in the G 1 -S cycle phase (38 -44). For example, decreased Skp2 would result in accumulation of p27 and thus block cellcycle progression from G 1 to S (42). Similarly, up-regulation of ATF3 or JunB has been shown to slow the progression of cells from G 1 to S (38 -40, 43). For the G 0 -G 1 progression, increased expression of ERF, a Ets-2 repressor, could result in G 0 arrest (45). Finally, a decrease in the expression of ASPM, an essential regulator of mitotic spindle orientation, might result in late stage mitotic arrest (46).
In addition to those genes whose decreased or increased expression may result in cell-cycle arrest, there are also contrary effects of HCF-1 sequestering that would promote cell-cycle progression. It is important, however, to consider the balance of FIG. 5. Western blot verification of the microarray results. Left, cell lines were induced for 32 h, and equivalent amounts of protein extracts were resolved in 4 -20% Tris-glycine gels, transferred to membranes, and probed with antiserum to various genes (ELK1, MCM2, ARF6, c/EBP␤, and CASP9). Right, the levels of each protein were quantitated, normalized to GADPH or endogenous control proteins, and represented as the ratio of the protein levels in the T-Act to the TEH-Act or Act control lines. The ratios obtained for these genes in the microarray analyses are presented for comparison. effects and the regulation of components that are essential to proliferation. Regardless, the effects of HCF-1 sequestering clearly implicate the protein in multiple cell-cycle stages via transcriptional regulation of cell-cycle control components.
As indicated by the genes affected by HCF-1 sequestering, regulatory control via HCF-1 is both broad and complex. Important components of other critical cellular processes or programs such as apoptosis, DNA replication and repair, and stress response signal transduction are also represented. The shift in levels of these components may dictate the cell response to stress signals and determine cell viability. The potential regulation of these genes by HCF-1 can direct future studies into the regulatory processes governed by this essential coactivator.
Interestingly, the expression of factors that are affected by HCF-1 sequestering includes transcriptional activators with predicted binding sites in the HSV IE genes such as Sp1, ELK-1, ATF3, NFI-C, PBX1, c/EBP␤, and Hes1 (TESS, www. cbil.upenn.edu/tess/ and MatInspector, www.genomatix.de (47)). The impact of HCF-1 control of these components may not be readily apparent during an HSV infection of tissue culture cells but may be important in determining those components that would impact the expression of the IE genes in various cell types in vivo. Most notably, in sensory neurons, HCF-1 is normally sequestered in the cytoplasm and is translocated to the nucleus upon various stress stimuli that initiate the reactivation of HSV from the latent state in these cells. This reactivation model proposes that transport of HCF-1 is important for the direct induction of HSV IE genes by assembly of VP16independent transcriptional complexes during the initial reactivation process (30). However, HCF-1 may also be important in an indirect manner by stimulation of the expression of activators (i.e. ATF3, Sp1) that can play a role in the induction of the IE genes and/or in later stages of the HSV lytic cycle. Consistent with this hypothesis, ATF3 is transcriptionally induced by either UV irradiation or axonal injury, stimuli that also result in reactivation of HSV from the latent state in sensory neurons (48,49).
Finally, there is a striking requirement for HCF-1 for the expression of proteins that participate in UV-induced DNA damage repair (i.e. KARP-1, PMS2L8, POLE/DNA polymerase ⑀, and TFIIH subunits p34 and p80 (50 -55)) and in the DNA damage response (i.e. CASP9, GTSE-1, Skp2, ATF3, and Hus1 (39, 56 -59)). The data suggest that HCF-1 may be a critical, although as yet unrecognized, component of the cellular response to DNA damage by regulation of proteins involved in DNA repair. Studies to assess the cellular response to such exposure in the absence of available HCF-1 could be facilitated using the HCF-1 sequestering system.