Trichostatin A Up-Regulates Human Papillomavirus Type 11 Upstream Regulatory Region-E6 Promoter Activity in Undifferentiated Primary Human Keratinocytes

Abstract

Human papillomavirus (HPV) gene expression in squamous epithelia is differentiation dependent in benign patient lesions and in organotypic raft cultures of primary human keratinocytes (PHKs). Using the lacZ reporter in raft cultures, we previously showed that this transcriptional regulation of the HPV type 11 (HPV-11) enhancer-promoter located in the upstream regulatory region (URR) appears to have resulted from coordination between the transcription transactivators AP1, Oct1, and Sp1 in differentiated upper strata and the repressor C/EBP in proliferating basal cells. We report here that trichostatin A, a specific inhibitor of histone deacetylase, dramatically stimulated reporter gene activity from the wild-type HPV-11 URR or the C/EBP mutation in PHKs grown in undifferentiated submerged cultures. In epithelial raft cultures, up-regulation occurred predominantly in basal and parabasal strata; this effect was promoter specific, as expression of the lacZ reporter gene driven by the murine leukemia virus long terminal repeat (LTR), the keratin 14 promoter, or the involucrin promoter was not altered, nor was expression of endogenous keratin 10 and profilaggrin affected. However, the responses were not cell type or species specific, as identical results were observed for both HPV-11 URR-lacZ and LTR-lacZ in murine retrovirus producer cell lines of fibroblast origin.

Human papillomaviruses (HPVs) infect squamous epithelia at various body sites, causing warty lesions. The low-risk HPV types, such as HPV type 6 (HPV-6) and HPV-11, cause genital condylomata and recurrent respiratory papillomatoses that almost never progress to high-grade lesions. On the other hand, the high-risk HPV types HPV-16 and HPV-18 can cause neoplastic progression in a low percentage of patients. In the productive infection program, expression of HPV genes is differentiation dependent (12, 14, 27, 36, 37; reviewed in reference 11). The HPV E7 protein inactivates the retinoblastoma susceptibility protein (pRB), a tumor suppressor, thereby reactivating the host DNA replication machinery in postmitotic, differentiated keratinocytes to support vegetative viral DNA amplification (10). The E6 protein inactivates another tumor suppressor, p53, and is thought to delay apoptosis, but this has not been directly tested. It is clear that inappropriate expression of the E6 and E7 oncoproteins in stem cells plays a major role in initiating viral carcinogenesis (reviewed in references 11 and 20). Since no progeny viruses are produced in high-grade dysplasias and carcinomas (36), it is critical for the virus to maintain its differentiation-dependent transcription program.

We have developed an epithelial raft culture system and demonstrated the squamous differentiation dependence of the HPV-18 and HPV-11 enhancer-promoter located in the upstream regulatory region (URR). In this system, the URR-driven reporter is introduced into primary human keratinocytes (PHKs) via acute infection with recombinant retroviruses. When the E7 gene is used as a reporter, squamous differentiation of raft cultures is not affected, but host DNA replication genes are reactivated and unscheduled cellular DNA synthesis is induced in differentiated keratinocytes (10). When the reporter is the bacterial lacZ gene, β-galactosidase (β-Gal) activity is predominantly detected in differentiated cell strata, with little or no activity observed in basal cells. Interestingly, in proliferating PHKs in submerged cultures, a significant fraction of the cells are positive for reporter gene activity, indicating that active repression of the URR occurs in basal proliferating cells upon stratification in raft cultures (33, 47, 48). Site-directed mutagenesis of sequence elements in both URRs has demonstrated that the integrity of the transcription factor binding motifs Sp1, Oct1, and AP1 is critical for conferring high β-Gal activities in differentiated keratinocytes. Conversely, the presence of two C/EBP transcriptional factor binding sites in the HPV11 URR diminishes reporter activity in the lower strata, particularly in the basal layer, indicating that they mediate URR repression.

Recent investigations have revealed that histone acetyltransferases and deacetylases are integral constituents of eucaryotic transcription complexes and that they participate critically in activating and silencing promoter activities, respectively (for a review, see reference 38). When histones H3 and H4 are acetylated on lysine residues near their amino termini, reduced positive charges lead to a relatively open chromatin conformation conducive to transcription activation. Conversely, histone deacetylation leads to a more condensed chromatin structure which is relatively inactive in transcription. Specifically, transcription coactivators such as p300, CBP, the nuclear receptor coactivator ACTR, and TAF(II)250 possess histone acetyltransferase activities (3, 8, 29, 32, 39, 44), whereas certain transcription repressors function by recruiting histone deacetylases to target promoters. For example, by collaborating with different factors, corepressor mSin3 inhibits Myc-responsive promoters as well as promoters targeted by retinoic acid and thyroid hormone receptors (1, 2, 15, 18, 21, 30). The family of retinoblastoma proteins repress genes necessary for cell cycle progression by binding to and inhibiting E2F transcription factors and by recruiting histone deacetylases (6, 13, 23, 24).

Trichostatin A (TSA), a specific inhibitor of histone deacetylases, has been used to assess a transcriptional regulatory role for promoter- or locus-specific histone acetylation and deacetylation (45). For example, TSA activates a reporter gene expressed from a cytomegalovirus or human α-globin promoter which has been stably transduced into cell lines via an adenovirus-associated virus vector (9). It also strongly induces the human immunodeficiency virus type 1 promoter reconstituted into chromatin in an in vitro transcription system (35). Interestingly, inhibitors of histone deacetylases induce remission of promyelocytic leukemia by activating transcription of retinoic acid receptor α-targeted genes critical for maturation of haematopoietic cells. In these patients, these genes are repressed by a fusion between the promyelocytic leukemia zinc finger protein and retinoic acid receptor-α, generated by chromosomal translocation (references 17 and 22 and references therein).

There has been no previous study of whether histone deacetylases regulate papillomavirus promoters. Since mutations of C/EBP repressor binding sites up-regulated URR promoter activity only in some basal cells, C/EBP cannot entirely account for promoter repression in these cells. We therefore examined the effect of TSA treatment on expression of the HPV-11 URR-lacZ reporter in PHKs. We show that in submerged, proliferating PHK cultures, TSA up-regulates reporter expression from the wild type HPV-11 URR and a URR in which the one or both of the known C/EBP repressor binding sites were mutated. In stratified raft cultures, induction was observed predominantly in basal and parabasal cells. These observations demonstrate that histone deacetylases also contribute to the relative inactivity of the HPV URR-E6 promoter in cells in the lower strata. Examination of raft cultures for expression of the lacZ reporter from additional viral or host promoters or for endogenous host proteins shows that this up-regulation exerted by TSA is relatively promoter specific; however, it is neither cell type nor species specific.

MATERIALS AND METHODS

Recombinant retroviral vectors.

The cloning vector pLN-lacZ is derived from pLNSX (28), in which the neomycin resistance gene is controlled by the murine leukemia virus (MuLV) long terminal repeat (LTR) promoter; the simian virus 40 (SV40) promoter has been removed so that the lacZ reporter is no longer expressed, for lack of a dedicated promoter (33). In pLN-11URR-lacZ, the wild-type or mutated URRs (spanning nucleotides 7072 to 7933, contiguous with nucleotides 1 to 99) were placed in the correct orientation downstream of the neomycin resistance gene to drive the lacZ reporter, as previously described (47). The C/EBP(distal [d]) and C/EBP(proximal [p]) elements were site-directed substitution mutations (48). The C/EBPM(dp) double mutation was created by replacing the BstEII and NdeI fragment from the C/EBP(d) mutation with the same fragment containing a C/EBP(p) mutation. pLN-K14-lacZ was similarly constructed by placing a 2.4-kb human keratin 14 (K14) promoter (41) between the neomycin resistance gene and the reporter gene. In pLJ-lacZ, the reporter gene is controlled by the MuLV LTR while the neomycin resistance gene is directed by the SV40 promoter (33).

Production of retroviruses and acute infection of PHKs.

Ecotropic and amphotropic recombinant retroviruses were produced from the helper cell line ψ-cre and pG+envAM12 (25), as previously described (33, 47). The latter producer cells were grown in 10% bovine serum in Dulbecco’s modified Eagle medium and selected with 800 μg of G418 (Geneticin; GIBCO/BRL) per ml for 6 days after infection with the ecotropic recombinant viruses. The cells were used directly for the TSA induction study or were used to generate amphotropic recombinant retroviruses. Amphotropic producer cells of pBabe-inv-gal, in which the lacZ reporter driven by a 2.5-kb involucrin promoter was inserted between the MuLV LTR and the SV40 promoter-driven puromycin resistance gene, were a generous gift of Joseph Carroll and Lorne Taichman (16). First-passage PHK cells at 30% confluence were infected with these recombinant retroviruses and then selected for 2 days with 400 μg of G418 per ml or 1.5 μg of puromycin per ml in serum-free medium (GIBCO/BRL). All uninfected cells died after selection. More than 50 to 70% of the cells usually survived selection. These cells were used immediately for experiments without further passage in order to minimize the possibility of clonal selection of cells with transcriptional properties not characteristic of the bulk population. A fraction of the keratinocytes was used for TSA induction studies in submerged cultures; another fraction was used for developing raft cultures at the air-medium interface.

TSA treatment.

Initial tests with a range of TSA (Sigma) concentrations established the optimal concentration of 0.6 μM, at which there was little or no toxicity while promoter induction was evident. Toxicity was evaluated by cell morphology and cell viability after removal of TSA (data not shown). The pG+envAM12 cell lines transduced with wild-type HPV-11 URR-lacZ and LTR-lacZ recombinant retroviruses were treated with 0.6 μM TSA for 24 h. Retrovirus-infected and G418- or puromycin-selected early passages of PHK cells were cultured in serum-free medium (GIBCO/BRL) for 2–3 days in the absence of fibroblast feeders. The cells were then treated with 0.6 μM TSA for 24 h. The treated and untreated cells were stained with X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside), as described before (47). Cells in duplicate plates were trypsinized, lysed, and used for in vitro β-Gal enzymatic assays (34). These experiments were performed twice with comparable results.

Epithelial raft cultures were developed on a dermal equivalent consisting of a rat tail type I collagen matrix containing Swiss 3T3 J2 fibroblasts. After 9 days of culturing at the medium-air interface, cultures were harvested and embedded in paraffin, as described previously (47). Selected raft cultures were treated with 0.6 μM or higher concentrations of TSA on day 8 for 24 h. Five-micrometer sections of the raft cultures were cut for histologic analysis and detection of β-Gal-positive cells. The slides were photographed after a light counterstaining with hematoxylin and eosin to correlate β-Gal activity with tissue morphology. Raft cultures were prepared in duplicate, and each experiment was performed several times with different batches of PHKs. Each batch was derived from several neonatal foreskins.

Immunohistochemical assays.

Tissue sections were subjected to antigen retrieval by heating in an 800-W microwave oven in 1% ZnSO₄ solution for 2 min at 30% power. A monoclonal antibody for keratin 10 (Biogenex, San Ramon, Calif.) was used at a 1:50 dilution, and that for profilaggrin/filaggrin (Biomedical Technologies, Stoughton, Mass.) was used at a 1:100 dilution. Immunohistochemical staining by peroxidase was performed with a kit from Zymed (San Francisco, Calif.).

RESULTS

We have previously shown that pLN-lacZ without a dedicated promoter (Fig. 1C) did not exhibit any β-Gal activity either in submerged, proliferating cultures or in organotypic raft cultures of primary keratinocytes due to an inefficient translation reinitiation of transcripts initiated from the LTR after termination of the upstream neomycin resistance gene product (33). pLJ-lacZ, in which the reporter is driven by the MuLV LTR (Fig. 1A), is active in submerged cultures and equally active in basal proliferating and differentiated spinous cells, indicating that LTR promoter activity is not influenced by squamous cell differentiation. In contrast, for pLN-URR-lacZ, in which the reporter is preceded by the HPV-11 URR (Fig. 1D) or HPV-18 URR, the activities are observed in a fraction of submerged, proliferating PHKs as well as in differentiated strata but not in the basal proliferating layer of epithelial raft cultures. Based on these observations and extensive mutagenesis of the URR, we have concluded that, in the sequence context of pLN-URR-lacZ, β-Gal was translated from transcripts initiated from the differentiation-dependent URR but not from RNAs derived from the upstream LTR (33, 47, 48). Using these constructs, we tested the effects of TSA treatment. The activities of the different reporters, however, should not be compared to one another, because titers of the different recombinant viruses were not equalized in this study.

FIG. 1.

Retrovirus constructs. (A) pLJ-lacZ, in which the lacZ reporter gene is directly driven by the MuLV LTR. (B) pBabe-inv-gal, in which the lacZ gene is under the control of a 2.5-kb involucrin promoter (16). (C) pLN-lacZ, in which the reporter has no dedicated internal promoter. (D) pLN-H11 URR-lacZ clone, in which the lacZ reporter gene is driven by a wild-type or mutated HPV-11 URR. (E) pLN-K14-lacZ, in which the reporter is driven by a 2.4-kb human K14 promoter.

TSA up-regulates both the wild-type HPV-11 URR and the URR with mutations in C/EBP transcription repressor binding sites in submerged, proliferating PHKs.

We tested the effects of TSA treatment on reporter activity in submerged, proliferating PHKs acutely infected with the pLN-H11URR-lacZ retrovirus. After a 2-day selection with G418, all uninfected cells died. Forty to 60% of infected PHKs were positive for reporter activity in the absence of TSA. Upon TSA treatment, more than 90% of the cells became positive, and the signal intensity was also increased (Fig. 2A, panels a and e). Upon treatment, PHKs assumed an enlarged and flattened morphology. A morphological change has previously been noted for cell lines (19). However, the effect was reversible upon removal of TSA, and the cells resumed proliferation and could be passaged, indicative of an absence of permanent toxicity (data not shown). To ascertain that the effect of TSA was on the URR rather than on protein translation in some unforeseen way, we tested in parallel experiments pLN-lacZ, which lacks a dedicated promoter. It remained completely negative with respect to β-Gal activity before and after TSA treatment (Fig. 2A, panels d and h). These results demonstrate that the HPV-11 URR is down-regulated by histone deacetylases in submerged proliferating keratinocytes. To investigate whether histone deacetylases are recruited to the URR by the C/EBP family of proteins bound to the two previously characterized sites, we tested the TSA responsiveness of mutations in which the distal site (d) or both the distal and proximal sites (dp) have been mutated. TSA stimulated the reporter activities in both cases, as it did for wild-type HPV-11 URR-lacZ (Fig. 2A, panels b and f, and data not shown).

FIG. 2.

β-Gal activities upon TSA induction of retrovirus-transduced cells in submerged, proliferating cultures. (A) Retrovirus-transduced PHKs. (B) pG+envAM12 retroviral producer cells. Panels a to d show cultures not treated with TSA. Panels e to h show cultures induced with 0.6 μM TSA for 24 h immediately prior to β-Gal assays. a and e, pLN-H11URR-lacZ; b and f, pLN-11-URR-C/EBP(d)M-lacZ; c and g, pLJ-lacZ; d and h, pLN-lacZ.

To quantify the overall extent of up-regulation, in vitro β-Gal assays of cell lysates were performed and the average induction from two independent experiments was determined (Table 1). The results show that TSA stimulated wild-type URR-lacZ by 3.4-fold, the C/EBPM(d) mutation by 3.7-fold, and the C/EBPM(dp) double mutation by 4.1-fold. These levels of induction are consistent with the visual impression that both the number of positive cells and signal strengths were elevated in TSA-treated cells. In contrast, pLN-lacZ-transduced PHKs did not exhibit any β-Gal activity with or without TSA treatment, relative to uninfected cells. These results support the notion that histone deacetylases can repress the HPV-11 URR independent of the two C/EBP binding sites.

TABLE 1.

Promoter regulation by TSA

Reporter	Relative β-Gal activitya
	PHK	pG+envAM12
pLJ-lacZ	1.0	0.9
pBabe-inv-gal	1.9	ND
pLN-lacZ	0.0*	0.0*
pLN H11-URR-lacZ	3.4	7.2
pLN H11-URR-C/EBP(d)M-lacZ	3.7	ND
pLN H11-URR-C/EBP(dp)M-lacZ	4.1	ND
pLN-K14-lacZ	6.3	ND
Uninfected PHKs	0.0*	0.0*

^apG+envAM12 producer cells were obtained after a 6-day G418 selection postinfection with ecotropic recombinant retroviruses. PHKs were infected with amphotropic retroviruses produced from pG+envAM12 cells or pBabe-inv-gal producer cells (16) and selected for 2 days with G418 or puromycin. All uninfected cells died under selection conditions. TSA induction was conducted at 0.6 μM for 24 h. Cultures were then harvested and assayed for β-Gal activities. Values are fold induction in TSA-treated cultures relative to identical cultures left untreated. Each value is the average of duplicate experiments. ND, not done; *, no detectable activity with or without TSA treatment. 

TSA stimulates wild-type and mutated HPV-11 URRs in PHK raft cultures.

To assess the effects of TSA on the HPV-11 URR during squamous differentiation, we prepared organotypic cultures of PHKs infected with recombinant retroviruses. Cultures were treated on the 8th day with 0.6 μM (Fig. 3) to 3 μM TSA (data not shown) and harvested after 24 h. As reported previously for untreated cultures, reporter gene activity was essentially restricted to suprabasal, differentiated cells. Upon TSA treatment, β-Gal reporter gene activity was dramatically induced in the lower strata, especially in undifferentiated basal cells. There was only a rather moderate effect in the differentiated upper layers (compare Fig. 3a and b). As described previously (48), the reporter activity from pLN-H11URR-C/EBPM(d) was up-regulated, mainly in the basal stratum (compare Fig. 3c and a). TSA significantly increased the number of strongly positive cells in the lower strata (compare Fig. 3c and d). Raft cultures infected with pLN-lacZ had no reporter activity in the presence or absence of TSA (Fig. 3g and h). Thus, we conclude that histone deacetylases contribute to the low-level activity of the HPV URR in cells in the lower strata.

FIG. 3.

β-Gal activities upon TSA induction of PHKs in raft cultures. TSA was added on the 8th day to a concentration of 0.6 μM to one of duplicate cultures, and rafts were harvested on the 9th day. Cultures were fixed and stained for β-Gal activity in situ. Panels a, c, e, g, and i show cultures not treated with TSA, whereas panels b, d, f, h, and j show cultures treated with TSA. a and b, HPV-11-URR-lacZ; c and d, HPV-11 C/EBP(d)M-lacZ; e and f, pLJ-lacZ; g and h, pLN-lacZ; i and j, pBabe-inv-gal.

The up-regulation of reporter activity by TSA is promoter specific.

To rule out that TSA might up-regulate viral or host genes nondiscriminatively, we examined the response of the lacZ reporter controlled by other promoters in PHKs grown in submerged and raft cultures. First, we examined PHKs transduced with pLJ-lacZ, in which the LTR controls reporter gene expression. β-Gal activity was observed in a population of submerged PHKs (Fig. 2A, panel c) and in some of the basal and differentiated keratinocytes in raft cultures in the absence of TSA (Fig. 3e), in agreement with our previous observation (33). Upon exposure to TSA over a range of concentrations from 0.6 to 3 μM, we detected no discernable induction in either submerged or raft cultures (Fig. 2A, panel g, and Fig. 3f; also data not shown).

We then examined PHKs similarly transduced with recombinant retroviruses in which the lacZ reporter was driven by the promoter of the basal-cell-specific human K14 (41) or that of the differentiated squamous-cell-specific involucrin (7), previously characterized in transgenic mice. In submerged PHKs transduced with pLN-K14-lacZ (Fig. 1E) or pBabe-inv-gal (Fig. 1B), 30 to 40% of cells were positive for β-Gal activity. TSA treatment doubled the percentage of positive cells. β-Gal assay of the respective cell lysates indicated a six- or twofold activity in these cultures relative to untreated cultures (Table 1). Interestingly, we rarely detected involucrin by immunofluorescence in untreated PHKs in submerged cultures (47). We do not know whether this discrepancy between reporter expression and endogenous gene product is due to a difference in sensitivity of the two assays or to posttranscriptional regulation of involucrin messages.

In raft cultures, β-Gal activity in pLN-K14-lacZ-transduced cells was confined to basal and lower spinous cells (data not shown). This observation is in agreement with the distribution of endogenous K14 mRNA in raft cultures, as the control of transcription shutoff or mRNA turnover upon differentiation in vitro is less stringent than in vivo (40). Conversely, in pBabe-inv-gal-transduced cultures, reporter activity was confined to the upper spinous cells (Fig. 3i). The presence of TSA at a 0.6 μM or higher concentration did not alter the distribution or intensity of reporter gene expression in either culture (Fig. 3j and data not shown). These results strongly suggest that promoter regulation by histone deacetylation has a certain degree of specificity.

We also examined the expression of two endogenous cellular genes, keratin 10 and profilaggrin, by immunohistochemistry in duplicate raft cultures that were transduced with pLN-H11URR-lacZ or the C/EBP(d) mutation but not stained for β-Gal activities. Keratin 10 is normally detected only in spinous cells and in granulocytes in cutaneous squamous epithelia, whereas profilaggrin is a marker for terminal squamous differentiation and is synonymous with keratohyalin granules in granulocytes in the more superficial layers. The pattern of expression of neither host gene was altered by TSA (Fig. 4, compare panels a to b and c to d; also data not shown). Notably, these two genes were not induced in the lower strata, where TSA significantly up-regulated HPV-11 URR-lacZ. While it is possible that the sensitivity of this assay may not be as high as that of the β-Gal assay, these observations agree with the notion that histone deacetylases do not regulate all promoters.

FIG. 4.

Expression of cellular protein in raft cultures as revealed by immunohistochemistry. A pLN-H11 URR-lacZ-transduced PHK raft culture treated with 0.6 μM TSA for 24 h (b and d) or an untreated control culture (a and c) (duplicate cultures of those shown in Fig. 3 that were not stained for β-Gal) was probed with a monoclonal antibody against keratin 10 (a and b) or profilaggrin (c and d).

Promoter responses to TSA induction are not cell type specific.

To examine whether the response of a given promoter to TSA is cell type specific, we examined reporter activities in the pG+envAM12 producer cells of four recombinant retroviruses: pLN-lacZ, pLJ-lacZ, and pLN-H11URR-lacZ and the C/EBP(d) mutation. The producer cells are derived from mouse 3T3 fibroblasts. As in PHKs, pLN-lacZ exhibited no β-Gal activity in the presence or absence of TSA (Fig. 2B, panels d and h). Not surprisingly, LTR-lacZ had relatively high-level activity, but there was no detectable induction by TSA at concentrations up to 3 μM (Fig. 2B, panels c and g). The reporter activity from the wild-type URR (Fig. 2B, panels a and e) and from the C/EBP(d) mutation (Fig. 2B, panels b and f) were both strongly stimulated when treated with TSA. In vitro β-Gal assays confirmed this qualitative visual evaluation. pLN-H11URR-lacZ was stimulated by 7.2-fold relative to untreated cultures, whereas neither pLN-lacZ nor pLJ-lacZ was affected (Table 1).

DISCUSSION

Multiple mechanisms control squamous cell differentiation-dependent up-regulation of the HPV-11 URR-E6 promoter.

Previous studies from our laboratory and others have revealed that expression of the HPV E6 and E7 genes is differentiation dependent in benign lesions from patients (11). We have recapitulated this differentiation-dependent URR activity in organotypic cultures of PHKs that were acutely transduced with recombinant retrovirus carrying an HPV-18 or HPV-11 URR-driven lacZ reporter (33, 47, 48). Site-directed mutagenesis of the URR in this system demonstrated that two mechanisms contribute to this promoter regulation. For both HPV-18 and HPV-11, binding to the enhancer-promoter elements of transcription activators such as Sp1, AP1, and Oct1 confers high-level activity in differentiated spinous cells. For HPV-11, the second form of regulation is transcription repression by a member or members of the C/EBP family in the lower strata, especially in the basal cells. The present study has shown for the first time that the state of histone deacetylation appears to comprise a third mechanism, as TSA, a specific inhibitor of histone deactylases, dramatically up-regulated both the wild-type URR and C/EBP mutations. Collectively, these three mechanisms lead to squamous cell differentiation-dependent promoter activity.

What might be the host transcriptional factor or factors that recruit histone deacetylases to the HPV-11 URR? Because mutations in one or both C/EBP binding sites in the HPV-11 URR maintain responsiveness to TSA treatment (Fig. 2 and 3; Table 1), C/EBP bound to these two sites would not appear to be responsible. However, the possibility of additional, yet-to-be-identified C/EBP sites cannot be ruled out. Alternatively, transcription factors bound to other sites may be responsible for recruiting histone deacetylases. One candidate factor is YY1, which interacts with mammalian histone deacetylases, and the ability of YY1 to repress transcription requires these interactions (42, 43). Although YY1 binding sites have not been reported in the HPV-11 URR, multiple YY1 sites in HPV-16 or HPV-18 URR modulate their respective E6 promoter activities in cell lines (4, 5, 26, 31). In HPV-18, YY1 interacts with C/EBPβ to enhance promoter activity. When the C/EBPβ binding site is deleted, YY1 becomes a repressor in a cell-type-specific manner. The detailed mechanism is not understood, nor has regulation by YY1 been reported in PHKs. Nevertheless, it is conceivable that YY1 also mediates histone deacetylase repression of the HPV-11 URR. A 5′ deletion mutation, 24-N, which retains nucleotides 7674 to 7933/1 to 99 but no longer contains the distal C/EBP site was also responsive to TSA stimulation (our unpublished results). This observation narrows the URR sequences to which these factors can bind and recruit histone deacetylases to a 350-bp segment upstream from the RNA initiation site at nucleotide position 99. During squamous differentiation, these host proteins either are no longer present, are not able to bind to the URR due to competition by other transcription factors, or are counteracted by other host proteins that recruit histone acetyltransferases.

Regulation by histone deacetylases is promoter specific but not cell type specific.

Our results with PHKs and amphotropic fibroblast producer cells demonstrate that promoter regulation by histone deacetylation is not cell type specific, as identical results were observed for the wild-type HPV-11 URR and for the LTR despite the different strengths of these two promoters in the two cell types (Fig. 2 and Table 1). Thus, histone deacetylase recruiting factors present in mouse fibroblasts are similar to those in PHKs. However, this regulation appears to be promoter specific in raft cultures. Of the four lacZ reporters tested, only the HPV-11 URR-E6 promoter was stimulated by TSA treatment, whereas the MuLV LTR promoter, which is active in all cell strata; the K14 promoter, which is restricted to less-differentiated PHKs; and the involucrin promoter, which is confined to more differentiated strata, were not altered (Fig. 3). Moreover, expression of the endogenous differentiation-dependent keratin 10 and profilaggrin proteins was not affected (Fig. 4). Collectively, these results rule out the possibility that TSA alters the transcriptional milieu of host cells in a dramatic, universal fashion, leading to a nonspecific promoter deregulation. We also infer that the squamous cell differentiation-dependent up-regulation of keratin 10, involucrin, and profilaggrin is mediated through mechanisms not identical to those responsible for that of the HPV-11 URR. Interestingly, in proliferating submerged cultures, all but the MuLV LTR were up-regulated by TSA. These results further highlight the differences in properties between proliferating cells in submerged cultures and those in the basal layer of a squamous epithelium. The absence of a response by pLN-lacZ without a dedicated promoter also shows that TSA does not have a discernable effect on mRNA translation. Lastly, the findings in this study further reinforce our previous interpretation that the URR, not the LTR, drives expression of the reporter in our HPV URR-based retroviral transduction system.

It is interesting to note that in submerged, proliferating cultures, only a fraction of cells were positive for β-Gal activity (Fig. 2). We do not believe that heterogeneity in transgene expression is related to cell cycle. First, TSA arrests cells in G₁ or G₂ phase (19, 46). Thus, it is unlikely that cells positive for β-Gal in the absence of TSA are in S phase when the chromatin could be relatively more open in newly replicated DNA prior to histone deacetylation. Second, heterogeneous expression of the lacZ reporter was observed with all four promoters regardless of whether they responded to TSA in submerged cultures (Table 1 and Fig. 2) or of their pattern of expression in raft cultures in the presence or absence of TSA (Fig. 3). These observations would be difficult to reconcile with the hypothesis that expression is cell cycle regulated. We suggest several more-probable explanations that are not mutually exclusive. First, since bulk cultures rather than clonal cell lines were examined, the sites of proviral integration undoubtedly varied in the infected cell population. Consequently, the transcriptional environment of the integrated proviruses could vary from site to site. Second, the copy number of proviruses in infected cells varies according to a Poisson distribution. These two factors can affect the levels of reporter transcripts. Third, because β-Gal is a tetrameric enzyme, the protein concentration may not reach the threshold to reveal enzymatic activity in some cells when the provirus copy number is low or when the transcriptional environment is less than optimal.

The observation that the HPVs contain multiple cis elements in the URR to down-regulate its enhancer-E6 promoter in basal and parabasal cells emphasizes that it is crucial for the virus to minimize E6 and E7 gene expression in these cells during productive infection. The adverse consequence of constitutive expression of the high-risk HPV E6 and E7 proteins in proliferating cells is evident; in vivo and in vitro, it dysregulates cell growth and differentiation, causing dysplasias and carcinomas, conditions not conducive to virus propagation.