Differentiation-dependent changes in levels of C/EBPβ repressors and activators regulate human papillomavirus type 31 late gene expression

Abstract

The liver-enriched transcriptional activator protein (LAP) isoform of CCAAT/enhancer binding protein β (C/EBPβ) is shown to be a major activator of differentiation-dependent human papillomavirus (HPV) late gene expression, while the liver-enriched inhibitory protein (LIP) isoform negatively regulates late expression. In undifferentiated cells, LIPs act as dominant-negative repressors of late expression, and upon differentiation, LIP levels are significantly reduced, allowing LAP-mediated activation of the late promoter. Importantly, knockdown of C/EBPβ isoforms blocks activation of late gene expression from complete viral genomes upon differentiation.

Fig 1

The C/EBPβ isoform LAP activates HPV-31 late gene expression upon keratinocyte differentiation. (A) Linear representation of the HPV-31 promoter-driven luciferase reporter constructs (p31-luc) used in this study. p31-luc-7045–891 contains sequences from nt 7045 to 7891 that include the 5′ end of HPV-31 URR through the E6/E7 open reading frames into E1. The deletion constructs depicted below p31-luc-7045–891 have been described previously (28). (B) Relative luciferase activity in extracts from 293T, C33A, HaCat, and CIN-612 cells transfected with p31-luc-7045–891 construct in the presence or absence of an LAP expression plasmid. (C) LAP activates HPV-31 late gene expression in a dose-dependent manner. Relative luciferase activity levels in extracts from 293T cells transfected with p31-luc-7045 or p31-luc-611–891 and in the presence of an increasing amount of an LAP expression construct are shown. Western blotting of C/EBPβ expression in the cells is shown below the graph. The membrane was subsequently stained with Ponceau S and used as a loading control. (D) Relative luciferase activity in extracts from C33A cells transfected with the indicated p31-luc deletion constructs in the presence or absence of an LAP expression plasmid. For the luciferase assays, the firefly luciferase activity was normalized to that of linked Renilla luciferase and the normalized luciferase activity level in the absence of LAP was set to 1; for panel D, p31-luc-7045–891 in the absence of LAP was set to 1. Each histogram bar represents the mean and standard error of the mean of the results of three independent experiments.

Fig 2

Localization of the C/EBPβ-responsive sequences in HPV late promoter sequences with respect to nt 794. (A) Relative luciferase activity in extracts from 293T, HaCat, and CIN-612 cells transfected with wild-type p31-luc-611–891 or p31-luc-611–891 containing a mutated C/EBP binding site along with LAP expression vectors. Renilla luciferase was used as an internal control, and the mean normalized luciferase activity level in the absence of LAP was set to 1. (B) Late promoter reporter activity increases upon differentiation and is reduced when the C/EBPβ binding site at nt 794 is mutated. The histogram shows fold increases in luciferase activity upon keratinocyte differentiation of CIN-612 cells that were transfected with wild-type p31-luc-611–891 or p31-luc-611–891 containing a mutated C/EBP binding site at nt 794. At 24 h after transfection, half of the cells were induced to differentiate in methylcellulose for 24 h and the other half remained in a monolayer culture as previously described (28). Renilla luciferase was used as an internal control. Fold activation values represent the ratios of the normalized luciferase activity in extracts from differentiated CIN-612 cells to that of undifferentiated cells. (C) Chromatin immunoprecipitation (ChIP) assay representing the binding of C/EBP to the viral late gene promoter (nt 611 to 891) in both undifferentiated and differentiated states of CIN-612 cells. The IgG control ChIP fold level was set at 1 to facilitate comparisons. The binding of C/EBP to the viral late promoter was enriched about 5-fold upon differentiation and may have been due, in part, to increased template numbers due to amplification. The ChIP assay protocol was performed as described by Wong et al. (31), and quantitative PCR (qPCR) was done using a Lightcycler 480 Sybr green I Master kit (Roche Applied Science). The data are consistent with those of a previous study in which C/EBPβ was normalized to the amount of input DNA (32). Each histogram bar represents the mean and standard error of the mean of the results of three independent experiments.

Fig 3

LIP repressor levels are significantly reduced upon differentiation, while LAP/LAP* activator levels are only moderately reduced. C/EBP expression patterns in undifferentiated (Undiff) and differentiated (Diff) CIN-612 cells are shown. CIN-612 cells were cultured either as a monolayer or in methylcellulose for 48 h. The cells were lysed and analyzed by Western blotting as described previously (23) using a rabbit polyclonal antibody specific to C/EBP (Santa Cruz; sc-746). Upon differentiation, the protein level of the inhibitor isoform LIP was reduced nearly to background levels. The activator isoforms LAP* and LAP (38 kDa) were also reduced in level upon differentiation but were retained at significantly higher levels than LIP. The levels of the RNA binding protein CUGBP1, which has been shown to be important for LIP translation, were also reduced upon differentiation. GAPDH (glyceraldehyde-3-phosphate dehydrogenase) levels are shown as loading controls.

Fig 4

The C/EBPβ LIP repressor inhibits C/EBPβ-mediated transcriptional activation of the HPV-31 late promoter. (A) Relative luciferase activity in lysate from 293T cells transfected with p31-luc-611–891 along with a constant amount of an LAP expression plasmid and an increasing amount of an LIP expression plasmid. Renilla luciferase was used as an internal control, and the mean normalized luciferase activity level in the absence of LAP and LIP was set to 1. (B) Relative luciferase activity in lysate from CIN-612 cells transfected with wild-type p31-luc-611–891 or p31-luc-611–891 containing a mutated C/EBP binding site. Renilla luciferase was used as an internal control, and the mean normalized luciferase activity level of wild-type p31-luc-611–891 was set to 1. (C) Fold increase in luciferase activity upon keratinocyte differentiation in the presence and absence of LIP. CIN-612 cells were transfected with p31-luc-7045–891 or p31-luc-611–891, in the presence or absence of an LIP expression plasmid. At 24 h after transfection, half of the cells were induced to differentiate in methylcellulose for 24 h. Renilla luciferase was used as an internal control. Fold activation values represent the ratios of the normalized luciferase activity in extracts from differentiated CIN-612 cells to that of undifferentiated cells. Each histogram bar represents the mean and standard error of the mean of the results of three independent experiments.

Fig 5

Lentivirus-mediated shRNA knockdown of C/EBPβ resulted in reduced viral late gene expression upon differentiation of cells stably maintaining HPV-31 episomes. Lentivirus virions were produced from five different validated shRNAs targeting C/EBP (3 μg each) by transfection along with pTatRevGagPol (2 μg) and pVSVG (1 μg) into HEK 293T cells. At 48 h posttransfection, the virions were collected and were individually concentrated by the use of centrifugation filters (Millipore) per the manufacturer's instructions. The concentrated virions from all the shRNAs were pooled at equal volumes and were used to infect CIN-612 cells. (A) Western blot representing the knockdown of C/EBPβ (LAP*, LAP, and LIP) in CIN-612 cells by sh C/EBPβ compared to sh Scr (scrambled control) and mock-infected cells. GAPDH served as a loading control. (B) Northern blot analysis was performed for HPV transcripts as previously described by the use of CIN 612 cells in which C/EBPβ levels were reduced. In cells in which C/EBPβ levels had been reduced, the levels of the major late transcripts encoding E1̂E4,E5 were significantly reduced upon differentiation compared to shScr and mock-infected cells. Levels of early E6*,E7,E1̂E4,E5 transcripts were found to be increased in cells in which C/EBPβ levels had been reduced.