Control of human adenovirus type 5 gene expression by cellular Daxx/ATRX chromatin-associated complexes

Abstract

Death domain-associated protein (Daxx) cooperates with X-linked α-thalassaemia retardation syndrome protein (ATRX), a putative member of the sucrose non-fermentable 2 family of ATP-dependent chromatin-remodelling proteins, acting as the core ATPase subunit in this complex, whereas Daxx is the targeting factor, leading to histone deacetylase recruitment, H3.3 deposition and transcriptional repression of cellular promoters. Despite recent findings on the fundamental importance of chromatin modification in host-cell gene regulation, it remains unclear whether adenovirus type 5 (Ad5) transcription is regulated by cellular chromatin remodelling to allow efficient virus gene expression. Here, we focus on the repressive role of the Daxx/ATRX complex during Ad5 replication, which depends on intact protein-protein interaction, as negative regulation could be relieved with a Daxx mutant that is unable to interact with ATRX. To ensure efficient viral replication, Ad5 E1B-55K protein inhibits Daxx and targets ATRX for proteasomal degradation in cooperation with early region 4 open reading frame protein 6 and cellular components of a cullin-dependent E3-ubiquitin ligase. Our studies illustrate the importance and diversity of viral factors antagonizing Daxx/ATRX-mediated repression of viral gene expression and shed new light on the modulation of cellular chromatin remodelling factors by Ad5. We show for the first time that cellular Daxx/ATRX chromatin remodelling complexes play essential roles in Ad gene expression and illustrate the importance of early viral proteins to counteract cellular chromatin remodelling.

Figure 1.

Functional inhibition of Daxx/ATRX repressive complex efficiently stimulates Ad5 viral gene expression. Hepa RG cells were infected with wild-type H5pg4100 at moi of 50 FFU per cell. (A) Viral particles were harvested 24, 48 and 72 h. p. i., and virus yield was determined by quantitative E2A-72K immunofluorescence staining on HEK293 cells. Shown are averages from three independent experiments. Error bars indicate the standard error of the mean. (B) Total cell extracts were prepared and treated with proteinase K. PCR was performed, and identicle amounts of PCR product were separated on analytic agarose gels (1%) and quantified with Gene Snap Software (Syngene). The results shown represent the averages from two independent experiments.

Figure 2.

Daxx/ATRX complex represses Ad5 mRNA synthesis. HepaRG cells were infected with H5pg4100 Ad5 wild-type at moi of 50 FFU/cell. Then, 24 h. p. i., total RNA was extracted, reverse transcribed and quantified by RT-PCR using primers specific for E1A (E1A fwd: 5′-GTGCCCCATTAAACCAGTTG-3′; E1A rev: 5′-GGCGTTTACAGCTCAAGTCC-3′), E1B-55K (E1B fwd: 5′-GAGGGTAACTCCAGGG TGCG-3′; E1B rev: 5′-TTTCACTAGCATGAaGCAACCACA-3′, E2A/DBP and hexon (hexon rev: 5′-GAACGGTGTGCGCAGGTA-3′; hexon fwd 5′-CGCTGGACATGACTTTTG AG-3′. Data were normalized to 18S rRNA levels (18S rRNA fwd: 5′-cggctaccacatccaaggaa-3′; 18S rRNA rev: 5′-GCTGGAATTACCGCGGCT-3′). Values correspond to the mean of triplicates, and error bars indicate the standard error of the mean.

Figure 3.

Daxx/ATRX complex mediates Ad transcriptional repression. (A) Hepa RG cells were transfected with luciferase reporter plasmids under Ad promoter control (E1A, E1B, E2early, MLP). Then, 48 h after transfection, samples were lysed, absolute luciferase activity was measured. All samples were normalized for transfection efficiency by measuring Renilla luciferase activity. Promoter activity of E1A, E1B, E2early and MLP promoter in Hep parental cells was normalized to 1. Mean and STD are from three independent experiments. (B) HepaRG cells were infected with H5pg4100 Ad5 wild-ype at moi of 50 FFU/cell. Then, 24 h. p. i. cells were fixed with formaldehyde and analysed by ChIP assays (see ‘Material and Methods’ section). The average C_t-value was determined from triplicate reactions and normalized against non-specific IgG controls with standard curves for each primer pair (Table 2). The y-axis indicates the percentage of immunoprecipitated signal from the input (100%). The white dotted line highlights values >1% of input, commonly stated as significant chromatin/protein binding.

Figure 4.

Daxx/ATRX functional complex modulates productive Ad replication. U2OS cells were transfected with either empty vector or pEGFP-ATRX plasmid expressing human ATRX 24 h before infection with wild-type H5pg4100 virus or E1B-55K minus virus mutant H5pm4149 at moi of 50 FFU per cell. (A) U2OS cells were harvested at 48 h. p. i., total cell extracts were prepared and proteins were separated by SDS–polyacrylamide gel electrophoresis (PAGE) and subjected to immunoblotting using mouse ATRX-specific mouse Mab or Daxx-specific rabbit Ab. Corresponding β-actin was included as a loading control. (B) Viral particles were harvested 24, 48 and 72 h. p. i., and virus yield was determined by quantitative E2A-72K immunofluorescence staining on HEK293 cells. Averages from three independent experiments are shown. Error bars indicate the standard error of the mean. (C) U2OS cells were harvested at 48 h. p. i., total cell extracts were prepared and treated with proteinase K. Quantitative real-time PCR was performed using hexon-specific primers. Ad5 H5pg4100 bacmid was used to obtain a standard curve. The results represent the averages from three independent experiments. Error bars indicate the standard error of the mean. (D) Total cell extracts were prepared and treated with proteinase K. PCR was performed and identicle amounts of PCR product were separated on an analytic agarose gels (1%) and quantified with Gene Snap Software (Syngene).

Figure 5.

Daxx/ATRX functional complex modulates productive Ad5 mRNA synthesis. U2OS cells were transfected with either empty vector or pEGFP-ATRX plasmid-expressing human ATRX 24 h before infection with wild-type H5pg4100 virus or E1B-55K minus virus mutant H5pm4149 at moi of 50 FFU per cell. Then, 24 h. p. i., total RNA was extracted, reverse transcribed and quantified by RT-PCR using primers specific for E1A, E1B-55K, E4orf6 (E4orf6 rev: 5′-CCCTCATAAACACGCTGGAC-3′; E4orf6 fwd: 5′-GCTGGTTTAGGATGGTGGTG-3′) and hexon. Data were normalized to 18S rRNA levels. Values correspond to the mean of triplicates, and error bars indicate the standard error of the mean.

Figure 6.

Proteasomal degradation of ATRX in infected and transfected cells. H1299 cells were infected with wild-type (H5pg4100) and mutant viruses (H5pm4149, H5pm4154) at moi of 50 FFU per cell. (A) Total cell extracts were prepared 48 h. p. i., and proteins were separated by SDS–PAGE and subjected to immunoblotting using mouse MAb 2A6 (E1B-55K), mouse MAb RSA3 (E4orf6), Daxx specific rabbit Ab and ATRX-specific mouse MAb clone 39F. β-actin was included as a loading control. Co-immunoprecipitation (IP) of E1B-55K was performed with ATRX-specific mouse MAb clone 39F followed detection of co-precipitated E1B-55K with mouse MAb 2A6. (B) Infected cells (as aforementioned) were treated for 6 h with proteasome inhibitor (+MG 132), before total cell extracts were prepared and specific proteins detected as described in A. (C) H1299 cells were transfected with pcDNA3-derived plasmids expressing wild-type E1B-55K, E4orf6 or a combination of both. Cells were harvested 48 h. p. i. Total cell extracts were prepared, and specific proteins were immunoprcipitated and detected as described in (A).

Figure 7.

ATRX is reduced via the E1B-55K/E4orf6 E3 ubiquitin ligase complex. (A) H1299 control and shCullin5 cells were infected with wild-type (H5pg4100) and mutant viruses (H5pm4149, H5pm4154) at moi of 50 FFU per cell. Total cell extracts were prepared 48 h. p. i., and proteins were separated by SDS–PAGE and subjected to immunoblotting using mouse MAb 2A6 (E1B-55K), mouse MAb RSA3 (E4orf6), Daxx- and Mre11-specific rabbit Ab and ATRX-specific mouse MAb clone 39F. β-actin was included as a loading control. (B) H1299 cells were infected with wild-type (H5pg4100) and mutant viruses (H5pm4149, H5pm4154, H5pm4139) at moi of 50 FFU per cell. (A) Total cell extracts were prepared 48 h. p. i., and proteins were separated by SDS–PAGE and subjected to immunoblotting using mouse MAb 2A6 (E1B-55K), mouse MAb RSA3 (E4orf6), Daxx- and Mre11-specific rabbit Ab and ATRX-specific mouse MAb clone 39F. β-actin was included as a loading control. (C) H1299 cells were infected with wild-type (H5pg4100) and E1B minus mutant virus (H5pm4149) at moi of 50 FFU per cell. Then, 48 h. p. i., total cell extracts were prepared after fractionation of soluble and chromatin fractions as described recently (60). Proteins were separated by SDS–PAGE and subjected to immunoblotting using Mre11-specific rabbit Ab, Daxx-specific rabbit Ab and ATRX-specific mouse MAb clone 39F, mouse MAb 2A6 (E1B-55K) and mouse MAb RSA3 (E4orf6).

Figure 8.

Inhibition of functional Daxx/ATRX complex modulates chromatin structure and condensation. Hepa RG cells were infected with H5pg4100 Ad5 wild-type at moi of 5 and 100 FFU/cell. 24 h p.i. (A) and 48 h p.i. (B), cells were fixed with formaldehyde and analysed by ChIP assays (see ‘Material and Methods’ section). Isolated chromatin was precipitated with histone variant H3.3-specific polyclonal rabbit Ab. The average C_t-value was determined from triplicate reactions and normalized with standard curves for each primer pair (Table 2). The y-axis indicates the percentage of immunoprecipitated signal from the input (=100%). Values above 1% of input (dotted line) indicate chromatin/protein binding. The term ‘% of input’ is commonly used as y-axis label for ChIP assays and reflects the percentage of signal normalized to the original sample prior to IP. Values between 0% and 0.05% reflect no binding to promoter sequences. (C) Hepa RG cells were infected with H5pg4100 Ad5 wild-type at moi of 5 and 100 FFU/cell. Then, 24 h p.i. cells were harvested and total cell extracts were prepared. Proteins were separated by SDS–PAGE and subjected to immunoblotting using Daxx rabbit polyclonal antibody 07–471, ATRX-specific mouse MAb clone 39F, mouse E1B-55K-specific mouse Mab 2A6 and ß-actin mouse MAb AC-15 as a loading control.

Figure 9.

MNase nuclease accesibility assay to monitor Daxx/ATRX-mediated modulation of chromatin structure. Hepa RG cells were either untreated or infected with H5pg4100 Ad5 wild-type at moi of 100 FFU/cell for 24 h. Chromatin sensitivity assays were performed using digestion with 20 U of MNase for the indicated periods followed by an RNase treatment. Digested chromatin was analysed on a 1.4% agarose gel using the G-Box system and Gene-Tools software (Syngene). Band intesities were quantified with ImageJ and analysed with GraphPad Prism software. Mono-, di-, tri and poly-nucleosomes are indicated on the right.

Figure 10.

Model for Ad5-mediated restriction of cellular Daxx/ATRX chromatin remodelling complexes. (A) A schematic representation of known cellular Daxx localizations in human cells. Nuclear Daxx is associated with either PML-NBs or ATRX at heterochromatin foci. Daxx ability to repress transcription is inhibited by its localization to the PML-NB (82). Cytoplasmic Daxx has been reported to be involved in cell death (56,83). (B) Daxx is reported to bind ATRX via two paired amphipathic helices. ATRX acts as the core ATPase subunit in this complex, whereas Daxx is the targeting factor, leading to deacetylation of histone tails (HDAC) and histone variant H3.3 depositioning resulting in transcriptional repression of target promoters. As the Daxx protein has no DNA binding region, ATRX is thought to be the molecular bridge connecting chromatin with Daxx bound to Daxx-interacting sequence-specific transcription factors (yellow sphere). (C) Daxx/ATRX repressive complexes assemble on viral genomes. Efficient transcription of Ad5 gene products necessitates inhibiting Daxx repressive complexes and/or preventing their assembly. Binding of E1B-55K triggers Daxx degradation via a proteasome, whereas ATRX restriction additionally requires the E4orf6 protein. These processes displace the repressive Daxx/ATRX complex from the viral genome, relieving negative regulation of Ad transcription.