We have determined distinct roles for different proteasome complexes in adenovirus (Ad) E1A‐dependent transcription. We show that the 19S ATPase, S8, as a component of 19S ATPase proteins independent of 20S (APIS), binds specifically to the E1A transactivation domain, conserved region 3 (CR3). Recruitment of APIS to CR3 enhances the ability of E1A to stimulate transcription from viral early gene promoters during Ad infection of human cells. The ability of CR3 to stimulate transcription in yeast is similarly dependent on the functional integrity of yeast APIS components, Sug1 and Sug2. The 20S proteasome is also recruited to CR3 independently of APIS and the 26S proteasome. Chromatin immunoprecipitation reveals that E1A, S8 and the 20S proteasome are recruited to both Ad early region gene promoters and early region gene sequences during Ad infection, suggesting their requirement in both transcriptional initiation and elongation. We also demonstrate that E1A CR3 transactivation and degradation sequences functionally overlap and that proteasome inhibitors repress E1A transcription. Taken together, these data demonstrate distinct roles for APIS and the 20S proteasome in E1A‐dependent transactivation.
Adenovirus (Ad) E1A protein expression is essential for both viral replication and cellular transformation (Gallimore and Turnell, 2001). Ad E1A proteins are expressed predominantly, from two splice variant transcripts, 12S and 13S, which, in Ad2/5, give rise to proteins of 243 and 289 amino acids respectively. Ad‐regulated expression of viral early genes, and some cellular genes, is dependent upon conserved region 3 (CR3), a region unique to 13S E1A (Jones, 1995). Studies investigating CR3 have not only been important in elucidating E1A function, but have also served as a model system for understanding the basis of transactivation per se.
Comparison of E1A sequences from different Ad subtypes suggests that CR3 extends from residue 144 to 191 in Ad2/5 (Avvakumov et al, 2002). CR3 contains a C4 zinc‐finger, a single Zn2+ ion coordinated by amino acids C154, C157, C171 and C174; mutation of any of the critical C residues results in loss of transactivation function (Culp et al, 1988; Webster et al, 1991). Residues H160 and S172 within the finger domain, residues V147, P150, Y175 and R177, which lie proximal to the finger domain, and a contiguous stretch of amino acids, encompassing residues V183–S188 within the C‐terminal region of CR3, are also critical for transactivation (Webster and Ricciardi, 1991). The transactivating properties of CR3 reside in its ability to recruit the necessary components of the cellular transcription machinery specifically to transcription factor‐bound promoters. Indeed, the zinc‐finger binds TBP, as part of a functional holo‐TFIID complex (Lee et al, 1991; Boyer and Berk, 1993; Geisberg et al, 1994). The C‐terminal region of CR3 (i.e. V183–S188) also recruits, independently of TBP, multiple TAF components of TFIID (Chiang and Roeder, 1995; Geisberg et al, 1995; Mazzarelli et al, 1997). The C‐terminal region of CR3 is also required for promoter targeting of competent E1A‐containing transactivation complexes; CR3 binds to promoter‐bound transcription factors through their DNA‐binding domains (DBD) (Hurst and Jones, 1987; Lee et al, 1987; Lillie and Green, 1989; Liu and Green, 1990, 1994). Phosphorylation of residues S185 and S188 within the C‐terminal domain modulates E1A‐dependent transactivation (Whalen et al, 1997). The zinc‐finger domain of CR3 also binds hSur2 (MED23), a component of Mediator (Boyer et al, 1999). Thus, CR3 is unable to function efficiently as a transactivator of Ad early genes in sur2−/− embryonic stem cells following viral infection or DNA transfection (Stevens et al, 2002). Indeed, the ability of mouse Ad1 to stimulate early gene transcription, and replicate, in sur2−/− mouse embryo fibroblasts is reduced significantly, relative to its capacity to perform these functions in sur2+/+ mouse embryo fibroblasts (Fang et al, 2004).
The 26S proteasome is the major non‐lysosomal proteolytic machinery in eukaryotes, serving to degrade polyubiquitylated proteins in an ATP‐dependent manner (Ciechanover, 1998). The 26S proteasome is assembled from two large macromolecular complexes, the 20S proteolytic particle and the 19S regulatory complex (RC). The 20S proteasome degrades non‐ubiquitylated proteins in an energy‐independent manner and possesses trypsin‐, chymotrypsin‐ and postglutamyl hydrolase‐like activities. The 19S RC can be functionally subdivided into base and lid components. The base complex consists of six homologous AAA ATPases (S4, S6, S6′, S7, S8 and S10b) and three non‐ATPase (S1, S2 and S5a) components, whereas the lid of the RC is made of an additional eight non‐ATPase subunits. The 19S ATPases regulate the assembly of the 26S proteasome, unfold substrate proteins targeted for degradation and gate protein translocation into the central chamber of the 20S proteasome (Ciechanover, 1998).
The realisation that the Saccharomyces cerevisiae protein responsible for rescue of a class of Gal4 activation domain mutants, Sug1, was a component of the 26S proteasome was crucial in establishing a potential role for the proteasome in transcription (Rubin et al, 1996). Significantly, Sug1 had previously been reported to recruit transcription factors such as Gal4 to TBP (Swaffield et al, 1995). The belief that Sug1 functions specifically in a number of diverse transcription programmes was strengthened by the observation that the mammalian orthologue of Sug1, S8, interacted with the thyroid receptor, the retinoic acid receptor (RAR) α, the oestrogen receptor (ER) and other nuclear receptors in a ligand‐ and AF‐2 activation domain‐dependent manner (Lee et al, 1995; vom Baur et al, 1996). 26S proteasome‐mediated degradation is essential for both RARγ‐ and ER‐dependent transactivation (Lonard et al, 2000; Gianni et al, 2002).
Given the emerging role of the proteasome in transcriptional regulation, and the specific ability of 13S E1A to function as a transactivator, we undertook a comprehensive study to determine whether there was a role for the proteasome in CR3‐dependent transcriptional activation. Data presented here offer new insights into E1A‐dependent transactivation and clearly establish roles for the 19S ATPase, S8, as a component of 19S ATPase proteins independent of 20S (APIS), and the 20S proteasome in Ad E1A‐CR3‐regulated transcriptional initiation and elongation in vivo. We also show for the first time that E1A transactivation capacity is dependent upon its inherent stability. CR3 transactivation and degradation sequences functionally overlap, suggesting that the 20S proteasome also regulates the termination of E1A transcription programmes. Taken together, these data define key and novel roles for proteasomal complexes in transcription programmes regulated by E1A, and offer insights into proteasome function, per se, in transcriptional regulation.
The 19S ATPase S8 binds to the CR3 transactivation domain of E1A
We have previously shown that S8 binds to the N‐terminal region of E1A (Turnell et al, 2000). In order to determine whether S8 can also bind to E1A through CR3, we utilised an N‐terminal double point E1A mutant, L1920A, which in the context of 12S E1A does not bind S8 (Rasti et al, 2005). Using GST‐w.t.‐12S E1A and GST‐L1920A‐12S E1A fusion proteins we initially confirmed, using pull‐downs from A549 cell lysates, that the N‐terminal E1A binding proteins S8, TBP and Ran all bind w.t. 12S E1A but do not bind L1920A‐12S E1A (Figure 1A, compare lanes 2 and 3). Interestingly, however, S8 and TBP retained significant binding capacity for L1920A‐13S, whereas Ran did not (Figure 1A, compare lanes 4 and 5), suggesting that S8 and TBP bind CR3 but Ran does not. The ability of pRB to bind in equal measure to each of the GST‐E1A fusion proteins validates the structural integrity of the GST proteins used (Figure 1A, upper panel).
In order to establish whether S8 binds to CR3 in vivo, we generated a number of A549‐derived, clonal cell lines that constitutively express either w.t. 13S E1A or L1920A‐13S E1A. Significantly, immunoprecipitation studies revealed that relative to w.t. 13S E1A, L1920A‐13S E1A retained appreciable binding capacity for S8 (Figure 1B) and TBP (Figure 1C), indicating that these proteins bind CR3 specifically, in vivo. Using cell lysis conditions that dissociate the 19S proteasome from the 20S proteasome (Grand et al, 1999), we next determined, by GST pull‐down, the binding site within CR3 for S8. These experiments revealed that the S8 binding site extends from amino acid 169 to 188 (Figure 2A, upper panel). S8 retained appreciable binding capacity for the zinc‐finger mutant C157S (Figure 2A, upper panel), and had w.t. CR3 binding capacity for the H160Y and Y175F zinc‐finger mutants (Figure 2A, middle panel). S8 also bound to CR3 species from all six human Ad subgroups (Figure 2A, lower panel), suggesting that this is a conserved function in E1A. TBP had a similar propensity to bind these same CR3 mutant proteins (Supplementary Figure S1A).
As S8 and TBP physically interact (Swaffield et al, 1995), we next examined by selective siRNA treatment whether S8 and TBP bind independently to CR3. Significantly, TBP retained full binding capacity for CR3 following the knock‐down of S8, and S8 retained full binding capacity for CR3 following the knock‐down of TBP (Supplementary Figure S2A and B). As S8 is one of six, homologous ATPases that comprise the base of the 19S RC, we also examined the capacity of CR3 and TBP to bind S4, S6, S6′, S7, S8 and S10b in vitro. Interestingly, TBP bound to all the 19S ATPases (Supplementary Figure S2C), although it bound to S8 with the highest capacity. CR3, on the other hand, bound S8 strongly, but had little or no capacity to bind other 19S ATPases (Supplementary Figure S2C). CR3, thus binds to S8 independently of TBP, and selectively binds S8 in preference to other 19S ATPases.
The 20S proteasome binds to the CR3 transactivation domain independently of S8
Given the ability of CR3 to bind S8, we wished to determine whether 20S proteasomes similarly bound to CR3. Using the protocol outlined above, we assayed the relative abilities of the CR3‐E1A mutants to bind 20S proteasomes from A549 cell lysates. Under these conditions, the 20S proteasome bound to CR3 (Figure 2B, upper panel). 20S binding to CR3 differed appreciably from that of S8. The binding site within CR3 for 20S extends from amino acid 161 to 177 (Figure 2B, upper panel). In contrast to S8, the 20S proteasome had little or no affinity for the zinc‐finger mutants, H160Y and Y175F (Figure 2A and B, compare middle panels). The 20S proteasome did however bind to CR3 species from all Ad subgroups, suggesting that this is a conserved function of E1A (Figure 2A and B, compare lower panels). Taken together, these data suggest that S8 function and 20S proteasome function, in the context of E1A transcription, are separable. MED23 had a similar, but not identical, propensity as the 20S proteasome to bind these same CR3 mutants (Supplementary Figure S1B).
CR3 binds preferentially to ‘free’ 20S proteasomes in vivo
Given that the 20S core particle can exist in vivo as a distinct entity, we wished to determine which 20S species were targeted by CR3. To do this, we fractionated 26S proteasomes from 20S proteasomes derived from A549 and 13S‐L1920A A549 cell lysates by fast protein liquid chromatography (FPLC) (Figure 3). 26S proteasomes and 20S proteasomes from A549 and 13S‐L1920A A549 cells eluted with identical profiles (cf. Figure 3A and B), suggesting that E1A binding does not affect 26S proteasome assembly. Immunoprecipitation studies confirmed this observation (Supplementary Figure S3A). Significantly, immunoprecipitation of 20S proteasomes from fractionated 13S‐L1920A A549 cells revealed that E1A bound preferentially to ‘free’ 20S proteasomes, and that only a minor proportion of 20S bound E1A as a component of the 26S proteasome (Figure 3C).
CR3 binds preferentially to APIS‐associated S8 in vivo
The 19S RC exists predominantly in vivo associated with one or both ends of the 20S proteasome as functional 26S complexes. The base component of the 19S RC, of which S8 is a part, has also recently been shown to exist as a distinct functional species, APIS, that is not associated with 20S proteasomes (Gonzalez et al, 2002; Figure 4A). Using fractionated eluates from 13S‐L1920A A549 cells (Figure 4A), we endeavoured to resolve whether CR3 targets S8 associated with 26S proteasomes or S8 associated with APIS. Intriguingly, immunoprecipitation studies revealed that CR3 bound preferentially to S8 associated with APIS (Figure 4B, upper panel); only a minor proportion of S8 bound to CR3 was through the 26S proteasome (Figure 4B, upper panel, compare lane 2 with lanes 6 and 8). Significantly, a ‘minor fraction’ of 20S proteasomes co‐eluting with APIS were found not to be associated with E1A; CR3 targeted ‘free’ 20S proteasomes (Figure 4B, lower panel). To confirm that APIS components associate with E1A in vivo, we immunoprecipitated 19S base components S8, S2, S6, S6′ and S10b from 13S‐L1920A A549 cells and subsequently Western blotted these samples for E1A. Immunoprecipitation revealed that 19S base components are found associated with E1A in 13S‐L1920A A549 cells (Figure 4C, upper panel). Reciprocal immunoprecipitation similarly revealed the association of 19S base components S2 and S10b with E1A (Figure 4C, lower panel). Consistent with these findings, E1A expression did not affect association of S8 with other 19S base components (Supplementary Figure S3B), indicating that E1A targets the whole APIS complex. The 19S base non‐ATPase component S2 co‐fractionated with APIS (Figure 4A) and also bound E1A (Figure 4C), establishing S2 as an integral component of APIS. Taken in their entirety, these studies indicate that CR3 targets 20S proteasomes and APIS, independently in vivo, to perform distinct functions.
S8 functions to enhance CR3‐dependent transactivation in vitro
To determine whether CR3 utilises S8 for its transactivation function, we initially investigated the effects of S8 overexpression upon the ability of a Gal4DBD‐CR3 construct to transactivate a Gal4‐responsive luciferase reporter gene in HCT116 cells. Interestingly, low‐level expression of exogenous FLAG‐tagged human S8 potentiated CR3 transactivation (Figure 5A, compare lanes 1 and 3 with lanes 4 and 5), whereas high‐level expression of FLAG‐tagged human S8 repressed CR3 transactivation (Figure 5A, compare lanes 1 and 3 with lanes 6–8). As low levels of exogenous S8 enhanced CR3 transactivation, we next determined by RNA interference (RNAi) the requirement for endogenous S8 in CR3‐dependent transactivation. Treatment of HCT116 cells with non‐silencing siRNAs allowed efficient transactivation by exogenously expressed Gal4DBD‐CR3 (Figure 5B, compare lanes 1 and 2). Interestingly, however, knock‐down of S8 expression by RNAi dramatically reduced the ability of DBD‐CR3 to transactivate the Gal4‐responsive reporter gene (Figure 5B, compare lanes 2 and 6). Knock‐down of TBP expression by RNAi similarly reduced the ability of DBD‐CR3 to transactivate the Gal4‐responsive reporter gene (Figure 5B, compare lanes 2 and 4).
Given these findings, we next investigated whether knock‐down of S8 (or TBP) expression by RNAi affected 13S E1A‐dependent transactivation of the Ad5 E4 promoter. Significantly, knock‐down of either S8 or TBP expression by RNAi attenuated 13S E1A‐dependent transactivation of the E4 promoter (Figure 5C). Importantly, exogenous expression of FLAG‐tagged human S8, or the ablation of endogenous S8 expression by RNAi, did not affect the transactivation capacity of Gal4DBD‐CBP (Supplementary Figure S4A and B), suggesting that not all transactivators utilise S8 for transcription. Taken together, these experiments demonstrate that S8 function is integral to CR3 transactivation function in mammalian cells.
Functional 19S ATPases are required for CR3 transactivation in yeast
Recent studies have utilised 19S conditional mutants of S. cerevisiae to demonstrate a functional requirement for Sug1 and Sug2, as components of APIS, in transcription initiation and transcription elongation (Gonzalez et al, 2002; Ezhkova and Tansey, 2004; Lee et al, 2005). Using these same yeast mutants, we investigated whether the ability of CR3 to stimulate transcription in S. cerevisiae (Shuen et al, 2002) was dependent on the function of these 19S base proteins. Following transformation of w.t. yeast and sug1‐25 and sug2‐1 strains with Gal4DBD‐CR3 and a Gal4‐responsive reporter possessing the GAL1 regulatory region upstream of the bacterial LacZ reporter gene, and their subsequent growth selection, we assayed CR3 transactivation capacity in these cells. In w.t. cells, CR3 efficiently stimulated transcription, whereas CR3 was unable to stimulate transcription in sug1‐25 and sug2‐1 mutant cells (Figure 6A), despite the Gal4‐CR3 protein being expressed to similar levels (Figure 6B). These data unequivocally establish a requirement for Sug1 and Sug2 in CR3‐dependent transcription in yeast, and are consistent with a role for orthologous proteins in CR3‐dependent transcription in human cells.
S8 function is required for efficient CR3 transactivation during Ad infection
As S8 interacts directly with CR3 in vitro and in vivo, and regulates CR3 transactivation function, we wished to determine whether S8 modulated the ability of CR3 to transactivate viral early promoters during Ad infection. To do this, we initially treated A549 cells with either non‐silencing siRNAs or with siRNAs directed against S8. At 24 h post‐transfection, we infected cells with w.t. Ad5 at a multiplicity of infection (m.o.i.) of one plaque forming unit (PFU)/cell. At the appropriate time post‐infection, we quantified Ad early gene expression. Significantly, the mRNA levels of 13S E1A, the 2.2 kb E1B transcript, E3A and E4 orf6/7 were all reduced significantly in Ad‐infected cells where S8 expression was reduced by RNAi, relative to their levels in Ad‐infected control cells, as measured by quantitative real‐time polymerase chain reaction (PCR) (Figure 7A). Consistent with these findings, Ad early protein expression was also reduced appreciably following S8 knock‐down, relative to cells expressing normal levels of S8 (Figure 7B, compare lanes 1–6 with lanes 7–12). These data provide substantive evidence to suggest that S8 functions in vivo during Ad infection to regulate CR3 transactivation function.
Recruitment of S8 and 20S proteasomes to Ad early promoters during Ad infection
We next investigated by chromatin immunoprecipitation (ChIP) whether E1A, S8 or the 20S proteasome could associate with Ad early promoters during viral infection. A549 cells were thus either mock‐infected or infected with w.t. Ad5 at an m.o.i. of 1 PFU/cell. At 24 h post‐infection, protein–DNA complexes were recovered by ChIP and the specific recruitment of E1A, S8 and the 20S proteasome to early region promoter elements was determined by PCR (see Supplementary Figure S5 for schematic representation of promoter regions amplified). Interestingly, these analyses showed that in addition to E1A, S8, α4 and α6 subunits of the proteasome were all found associated with Ad early promoters during viral infection (Figure 8, panels 1–5, respectively; compare lanes 1–5 with lanes 6–10). The apparent difference between α4 and α6 reflects the differential abilities of the respective antibodies to immunoprecipitate the 20S proteasome (data not shown). Consistent with earlier reports that E1A does not form specific complexes with E2l promoter, we found only modest amounts of E1A and S8 associated with the E2l promoter, compared to the association of E1A and S8 with the E2e promoter (Figure 8, panel 6, compare lanes 2–4 with lanes 7–9). These studies suggest that S8, in the context of APIS, and the 20S proteasome function, in concert with E1A, to initiate transcription.
S8 recruitment to early region promoters during infection is dependent upon CR3
As S8 associates with Ad early promoters during infection, we next determined whether S8 recruitment was dependent upon E1A. Thus, following infection of A549 cells with either w.t. Ad5, or Ad5 dl1113 (Δ169–177; Figure 9A, upper panel) or Ad5 dl1114 (Δ180–184); Figure 9A, lower panel) mutants that do not bind S8, we determined by ChIP that association of S8 with the E3 promoter (Figure 9A, upper and middle panels, compare lane 8 with lane 12) and other Ad early promoters (data not shown) was dependent upon the expression of a 13S E1A species that bound S8. Western blotting revealed that w.t. E1A and E1A mutant proteins were expressed to similar levels following infection (Figure 9A, lower panel). These data suggest that S8 recruitment to Ad early promoters by E1A is mediated exclusively through CR3 and is not through the alternative S8 binding site located in the N‐terminal region (Turnell et al, 2000).
As studies in yeast have suggested that S8 and 20S proteasomes can be found associated with actively transcribing gene units (Ferdous et al, 2001; Gonzalez et al, 2002), we investigated whether E1A and proteasomal subunits could also associate with actively transcribed Ad gene regions during infection. Significantly, E1A, S8 and 20S proteasomes could be found associated with E1B gene elements across the length of the E1B gene (Figure 9B, upper and lower panels), indicating that E1A, 19S and the 20S proteasome are also required for transcriptional elongation.
Relationship between 13S E1A stability and transactivation potential
As the transactivation potential of a number of transcription factors is controlled specifically by proteasome‐mediated degradation (Salghetti et al, 2000), we also addressed whether transactivation‐defective 13S mutant proteins had altered half‐lives relative to w.t. 13S E1A. To do this, we used A549 cells that stably expressed either w.t. 13S E1A or transactivation‐defective 13S E1A mutants H160Y and Y175F. We specifically chose to investigate these two transactivation‐defective mutants as they have little or no affinity for the 20S proteasome (Figure 2B). Significantly, 13S E1A mutants H160Y and Y175F had considerably longer half‐lives than w.t. 13S (Figure 10A); w.t. 13S E1A had a half‐life of approximately 1 h, whereas H160Y had a half‐life of approximately 2 h, and Y175F had a half‐life of approximately 3 h. We also determined that transactivation‐defective 13S E1A mutants (Δ161–168; Δ169–177), which similarly do not bind the 20S proteasome, also had extended half‐lives relative to w.t. 13S E1A (data not shown). Taken together, these experiments reveal that there is a close relationship between 13S E1A stability, 20S proteasome binding and transactivation potential. Consistent with these findings, w.t. 13S E1A protein levels were more sensitive to proteasomal inhibition than the transactivation‐defective 13S E1A mutants H160Y and Y175F (Figure 10B). Indeed, selective proteasomal inhibition repressed CR3‐dependent transactivation (Supplementary Figure S6), suggesting that proteasome function is important in the transactivation process.
Binding of 19S and 20S to the CR3 transactivation domain of E1A
We present considerable evidence to demonstrate that the 19S ATPase, S8, and the 20S proteasome bind to separate sites within CR3 (cf. Figure 2A and B). In Ad5, the binding site for S8 extends from residue 169 to 188, whereas the 20S proteasome binding site extends from residue 161 to 177. Indeed, specific point mutants H160Y and Y175F bind S8 as w.t. CR3, yet have little or no capacity to bind the 20S proteasome (cf. Figure 2A and B). Significantly, CR3 from all Ad subgroups bind both S8 and the 20S proteasome, suggesting a conserved Ad function. The ability of CR3 to target these proteins separately presumably reflects the capacity of 19S and 20S to perform distinct functions during E1A‐mediated transcription.
We also obtained evidence in vivo to indicate that CR3 targets S8 and the 20S proteasome separately. Following the resolution of 26S proteasomes, 20S proteasomes and APIS from 13S‐L1920A A549 cells by FPLC, we determined by immunoprecipitation that CR3 bound S8 as a component of APIS, independent of 26S proteasomes (Figure 4B), and similarly bound ‘free’ 20S proteasomes, independently of 26S proteasomes (Figures 3C and 4B). These studies also revealed that other 19S ATPase components of APIS were associated with CR3 of E1A in vivo, indicating that E1A CR3 targets the entire APIS complex through binding to S8 (Supplementary Figure S2C). Previous studies have suggested that APIS minimally comprises the six 19S base‐component ATPases (Gonzalez et al, 2002). Our immunoprecipitation studies reveal that the non‐ATPase 19S subunit, S2, is also a component of APIS (Figure 4C).
Our in vitro binding studies suggested that the binding sites for S8 and TBP were very similar (cf. Supplementary Figure S2A and B). However, although TBP and Sug1 (S8) can functionally interact (Swaffield et al, 1995), we determined by RNAi that TBP and S8 bind CR3 independently (Supplementary Figure S2), suggesting distinct functions for these proteins in the transactivation process. We also determined that the transactivation domain of E1A bound preferentially to S8 (Supplementary Figure S2C), whereas TBP was able to bind to all six 19S ATPases, although it bound most strongly to S8 (Supplementary Figure S2C). Interestingly, the Gal4 activation domain has previously been shown to bind both Sug1 and Sug2 preferentially in yeast (Gonzalez et al, 2002). Significantly, however, both CR3 (Figure 4C) and Gal4 (Gonzalez et al, 2002) associate with all 19S ATPases in vivo. It thus appears that different transactivators target different 19S ATPases for the specific recruitment of APIS. It is perhaps not surprising, therefore, that TBP binds to all six ATPases (Supplementary Figure S2C), facilitating the specific recruitment of APIS–transactivator complexes to the basal transcriptional machinery.
Role for APIS in 13S E1A‐mediated transcription initiation and elongation
Data presented within this report indicate that the proteasomal ATPase S8, as a component of APIS, is required for CR3 13S E1A transactivation function (Figures 4, 5, 6 and 7). In this regard, CR3 can utilise low levels of exogenous S8 for transactivation (Figure 5). Consistent with our findings that overexpression of exogenous S8 represses CR3 activity (Figure 5A), high levels of exogenous S8 have been postulated previously to interfere with endogenous S8 function in proteasomal‐mediated protein degradation, and hence activation of the nuclear receptor RARγ2 (Gianni et al, 2002). Using RNAi, we also established a role for endogenous S8 in CR3‐dependent transactivation in reporter assays (Figure 5B and C) and during w.t. Ad5 infection. Indeed, the ability of w.t. Ad5 to promote early gene mRNA expression, and consequently early gene protein expression, is reduced significantly in cells where S8 expression has been reduced by RNAi (Figure 7). The fact that expression of E1A itself is reduced in these circumstances is a significant finding and demonstrates an important feedback loop in the in vivo requirement for E1A and S8 in the transcriptional activation of the E1A promoter (Figure 7). The observation that E1A still retains transactivation potential following S8 knock‐down in these instances might reflect residual S8 activity, the ability of E1A to form partially competent transactivation complexes in the absence of S8, the ability of E1A to bind APIS independently of S8 or may merely reflect 13S E1A‐independent activation of these promoters. In addition to these findings, we also determined a role for 19S ATPases in CR3 transactivation function in yeast, such that CR3 could not function as a transactivator where Sug1 or Sug2, and hence APIS were functionally inactivated through mutation in their ATPase domains (Figure 6). These data are consistent with previous studies indicating a requirement for APIS in Gal4‐dependent transcription in yeast (Gonzalez et al, 2002; Ezhkova and Tansey, 2004).
Recruitment of APIS and 20S proteasomes to promoters and gene elements in vivo
We also demonstrate in this report that 13S E1A, S8, and 20S proteasomes are recruited to both Ad early region promoters and Ad early region gene units during the early phase of an Ad infection (Figures 8 and 9). Roles for E1A, the 20S proteasome and APIS in transcription initiation are well documented (Jones, 1995; Morris et al, 2003; Ezhkova and Tansey, 2004). This is the first report, however, demonstrating a potential role for 13S E1A in transcriptional elongation. Pertinent to these new findings, Sug1, as a component of APIS in yeast, has recently been reported to associate not only with the activated GAL1‐10 promoter but also with GAL1 gene units that are being actively transcribed (Gonzalez et al, 2002) and participate in transcriptional elongation (Ferdous et al, 2001). Indeed Sug1, through its recruitment to ubiquitylated histone H2B, can also regulate transcription initiation through the modification of histone H3 methylation (Ezhkova and Tansey, 2004). Given the specific requirement for Sug1 in RNA polymerase II‐dependent transcriptional activation and elongation in eukaryotes, the ability of 13S E1A to utilise S8, as a component of APIS, for initiation and elongation of Ad early region genes suggests an evolutionarily conserved mechanism for transcription of cellular and viral genes. Moreover, recent evidence suggests that Sug1 can specifically recruit the SAGA acetyltransferase complex to transcriptional activators and enhance its recruitment to promoter elements to affect global histone H3 acetylation status (Lee et al, 2005). Interestingly, we have previously shown that CR3 recruits the SAGA complex, through binding SAGA component gcn5 acetyltransferase (Shuen et al, 2002). CR3, in this circumstance, will also bind the human gcn5 orthologue, P/CAF (Shuen et al, 2002). It will be of considerable importance in the future to establish if P/CAF and APIS interact to mediate Ad transcription function in human cells.
Role for proteasome‐mediated degradation in CR3 transactivation function
The transactivation domains from a number of short‐lived transcription factors functionally overlap with sequences that ultimately coordinate their destruction by the proteasome (Salghetti et al, 2000). The activation domain potency of a transcription factor is thus inversely related to the half‐life of the protein; mutations that inhibit proteasome‐mediated degradation inhibit transactivation function (Molinari et al, 1999). In this report, we present evidence to demonstrate that 20S proteasomes bind 13S E1A (Figures 2B, 3 and 4) and are found associated with both enhancer/promoter (Figure 8) and actively transcribed Ad early region genes (Figure 9B). Consistent with the possibility that 20S proteasome‐mediated degradation might regulate 13S E1A transactivation capacity, previous reports have indicated that the 13S E1A protein is a short‐lived protein, and significantly more labile than the 12S E1A protein (Spindler and Berk, 1984); 13S E1A possesses a PEST motif within CR3 that is lacking in 12S E1A (Turnell et al, 2000). Given these findings, we investigated the possibility that the transactivation function of 13S E1A is regulated by proteasome‐mediated degradation. Significantly, proteasome inhibitors repressed CR3 transactivation function (Supplementary Figure S6), indicating that the proteasome directly controls 13S E1A transactivation potential. The ability of E1A to stimulate transcription is related directly to the level of E1A expression (Supplementary Figure S6). This reflects the ability of proteasome–E1A complexes to regulate the E1A promoter, and is consistent with the effects of S8 knock‐down upon the level of expression of E1A following w.t. Ad5 infection (Figure 7). In consideration of whether proteasome‐mediated degradation of 13S E1A controlled its inherent transactivation capacity, we established that transactivation‐defective mutants H160Y and Y175F had little or no capacity to bind the 20S proteasome (Figure 2B) and had significantly longer protein half‐lives than w.t. 13S E1A protein (Figure 10A). Thus, the transactivation potency of 13S E1A is related to its inherent stability. Moreover, we also determined that in contrast to w.t. 13S E1A, H160Y 13S and Y175F 13S protein levels are resistant to the use of selective proteasome inhibitors (Figure 10B), consistent with the low affinity of these proteins for 20S proteasomes.
Taken together, these data indicate that 13S E1A has evolved such that its transactivation domain functionally overlaps sequences that target the protein for degradation. Proteasome‐mediated degradation of 13S E1A is required for efficient 13S E1A‐mediated transcription function. We suggest that during viral infection or cellular transformation, 13S E1A utilises S8, as a component of APIS, and 20S proteasomes independently and at multiple stages in the transcription process to regulate Ad transcription programmes. In this regard, we hypothesise that E1A would recruit APIS to enhancer/promoter elements and gene elements independent of 20S proteasomes, to activate transcription in a non‐proteolytic manner. As such, APIS recruitment might facilitate the recruitment of coactivators such as P/CAF and methyltransferases to the promoter to modify promoter acetylation and methylation, and hence transcriptional initiation. E1A might also recruit APIS to gene elements to promote transcriptional elongation. On the other hand, E1A might recruit 20S proteasomes, independently of APIS, to regulate transcription through proteolysis. In this regard, E1A might recruit 20S proteasomes to promoter elements to facilitate degradation of chromatin‐bound transcription factors, and/or E1A, in order to promote multiple rounds of transcription initiation. E1A might similarly recruit 20S proteasomes to gene elements to establish limits for transcription through targeted degradation of regulatory proteins. The efficacy of 13S E1A‐dependent transactivation is thus, in turn, dependent upon the combined activities of APIS and 20S proteasomes. The ability of 13S E1A to recruit these complexes separately allows for a temporally and spatially coordinated programme of proteasome‐mediated activation and inactivation. During an Ad infection, the magnitude of the proteasome‐dependent 13S E1A transactivation signal is modulated, in part, by a feedback loop involving 13S E1A itself, which regulates its own promoter. Whether APIS and 20S proteasomes can functionally cooperate in the transcription process by transiently associating on chromatin awaits further investigation. We propose a model whereby CR3 of 13S E1A binds to the basal transcriptional machinery (e.g. TBP), TBP‐associated factors, transcription factors (e.g. ATF‐2), transcriptional Mediator (e.g. MED23), transcriptional coactivators (e.g. P/CAF), APIS, the 20S proteasome, and potentially other transcriptional regulators to temporally coordinate E1A‐stimulated gene expression.
Materials and methods
Cells and viruses
Human, small‐cell lung carcinoma A549 cells and colon carcinoma HCT116 cells were grown and maintained in HEPES‐buffered Dulbecco's modified Eagle's medium containing 2 mM glutamine and 8% fetal calf serum. A549‐derived clones that stably express similar levels of w.t. 13S Ad5 E1A, or 13S Ad5 E1A mutant proteins were isolated after transfection and selection with G418 (800 μg/ml). w.t. Ad5 and Ad5 mutants Ad5 dl1113 and Ad5 dl1114 (Jelsma et al, 1988) were grown in Ad5 293 cells and the virus titre determined by plaque assay using Ad5 911 cells.
Ad E1A mutants
Ad 13S E1A and Ad CR3 mutants were generated by PCR in accordance with the manufacturer's instructions (Stratagene). Mutants were validated by direct sequencing of both strands of AdE1A cDNA using an ABI Prism 3100 Genetic Analyzer.
Immunoprecipitation and GST pull‐downs
For immunoprecipitation, cells were lysed in buffer containing 50 mM Tris–HCl (pH 7.4), 0.825 M NaCl and 1% NP‐40, sonicated and cleared by centrifugation. For GST pull‐downs, 10 μg of the appropriate GST‐fusion protein was mixed with either A549 cellular lysate or with the relevant in vitro‐translated protein (Promega), labelled with l‐α‐[35S]methionine (Amersham). Immunoprecipitation and GST pull‐downs were performed as described (Rasti et al, 2005).
PAGE and Western blot analysis
Samples, solubilised in 9 M urea, 50 mM Tris–HCl (pH 7.4) and 0.15 M β‐mercaptoethanol, were sonicated and cleared by centrifugation. Protein concentrations were determined by Bradford assay (Bio‐Rad). Samples were separated either on 12% polyacrylamide gels, run in the presence of 0.1 M Tris, 0.1 M Bicine and 0.1% SDS, or in the presence of 7 M urea, 93 mM Tris, 15 mM glycine but in the absence of SDS. For Western blotting, proteins were electroblotted onto nitrocellulose filters (Gelman Sciences), hybridised with the appropriate antibodies and visualised by enhanced chemiluminescence (Amersham Pharmacia).
Fractionation of 26S proteasomes, 20S proteasomes and APIS by FPLC
Cells were washed in ice‐cold saline and solubilised by homogenisation in 20 mM Tris (pH 7.2), 5 mM ATP, 20% (v/v) glycerol and 0.2% (v/v) NP‐40. Lysates were clarified by centrifugation at 40 000 r.p.m. for 30 min. Protein lysate (2 mg) was fractionated on a Superose 6 HR 10/30 column connected to an Amersham Pharmacia FPLC system. The column was eluted with 20 mM Tris (pH 7.2), 5 mM ATP, 0.1 M NaCl and 10% (v/v) glycerol at 0.25 ml/min and 0.25 ml fractions collected.
DNA‐LipofectAMINE complexes were incubated with cells for 6 h according to the manufacturer's instructions (Invitrogen). At the appropriate time post‐transfection, cells were lysed and luciferase or CAT activities determined according to the manufacturer's instructions (Promega).
Purified, annealed double‐stranded 21‐mer RNA oligonucleotides were from Ambion. The targeted gene sequences were S8 (nucleotides 485–507): 5′‐aagaagtgatcgagctgcctgtt; TBP (nucleotides 606–628): 5′‐gaggataagagagccacgaactt. A non‐silencing siRNA 5′‐aattctccgaacgtgtcacgttt, with no known homology to any human gene, was used as a negative control. A549 cells were transfected with siRNAs by electroporation (960 μF and 220 V in 4 mm cuvettes).
After mock or Ad infection, cells were treated with 1% paraformaldehyde (v/v) for 10 min at 37°C to crosslink DNA–protein complexes. ChIP was then performed according to the manufacturer's instructions (Upstate Biotechnology). The oligonucleotides used for amplification were as follows: E1A promoter: forward tgacgtagtagtgtggcggaagtgt and reverse tataaatacactacacgtcagctga; E1B promoter: forward gtgtctagagaatgcaatagtagtacggat and reverse taaccaagattagcccacggcgcattatat; E2e promoter: forward tcatgtagtttattcgggttgagtagtctt and reverse atactgcgcgctgactcttaaggactagtt; E3 promoter: forward tggtgtaccaggaaagtcccgctcccacca and reverse caagcgaggagctcaccgactcgtcgttga; E4 promoter: forward gcagagcgagtatatataggactaaaaaat and reverse taccttattttggattgaagccaatatgat; E2l promoter: forward gactgtcccagtattcctcctcgtccgtgggt and reverse ccacattcttggccaattgcaagccatcaa; E1B gene 1750: forward tgcgtaacttgctggaacagagctctaaca and reverse tcttcgctccatttatcctttataaaactc; and E1B gene 2591: forward tgtaaatatcaggaattgttgctacatttc and reverse tttcttaattgaagccctgcttttggggcg. Oligonucleotides were from Alta Bioscience (The University of Birmingham).
Ad5 CR3 transactivation studies in yeast
Yeast strains, Sc507 (MATa, can1‐100, ade2‐1, his3‐11, ‐15, leu2‐3, ‐112, trp1‐1, ura3‐1) and congenic strains Sc660 (sug1‐25) and Sc671 (sug2‐1) were from Dr Tom Kodadek (UT‐Southwestern Medical Center, Dallas, TX). Ad5 CR3 was cloned into the yeast expression plasmid, pAS1, such that CR3 was expressed as a Gal4DBD fusion protein (Shuen et al, 2002). The Gal4‐responsive reporter construct used in these studies contains the GAL1 regulatory region upstream of the bacterial LacZ reporter gene, an ampicillin resistance gene for Escherichia coli selection, a LEU2 gene for yeast auxotrophic selection and a 2‐μm yeast origin of replication. For experimental purposes, yeast strains were transformed with the Gal4‐reporter construct and pAS1 or pAS1‐CR3 using a modified lithium acetate procedure and plated on synthetic complete media lacking appropriate nutrients. Yeast transformants were then grown in selective media at 30°C until they reached an OD600 of 1.0. Cellular lysates were subsequently prepared for either β‐galactosidase assays (Promega) or Western blotting analysis.
Quantitative real‐time PCR
RNA extraction and PCR were essentially as described previously (Zhang et al, 2004). Primers used for amplification were as follows: 13S E1A F, 5′‐atg ttt gtc tac agt cct gtg tct gaa; 13S E1A R, 5′‐gat agc agg cgc cat ttt agg; E1B F, 5′‐cgc gct gag ttt ggc tct ag; E1B R, 5′‐tca aac gag ttg gtg ctc atg; E3A F, 5′‐gcc gcc aca agt gct ttg; E3A R, 5′‐ctc gga gag gtt ctc tcg tag act; E4 orf6/7 F, 5′‐ctg ctg ccc gaa tgt aac act; E4 orf6/7 R, 5′‐tcc acc ttg cgg ttg ctt aa. FAM‐labelled Ad early probes were as follows: 13S E1A, 5′‐cca gaa ccg gag cct gca aga cct ac; E1B, 5′‐cgg cgg ctg ctc aat ctg tat ctt ca; E3A, 5′‐ccg cga ctc cgt ttc aac cca g; E4 orf6/7, 5′‐tga caa tgc aca agc gtg gac ttc tcc. All of the reactions were multiplexed with a preoptimised control VIC‐labelled probe for 18S ribosomal RNA (PE Biosystems). Primers and probes were from Eurogentec.
The E1A monoclonal antibodies (mAb) M2, M73 and M58, E1B‐55K mAb, 2A6, and the pRB mAb, IF8, were all obtained as a supernatant fluid from hybridoma cell lines. The E1A polyclonal rabbit antibody (pAb) was from Eurogentec, the E1B‐19K rat mAb was from Oncogene Science, the E2‐72K pAb was from Peter van der Vliet (University Medical Centre, Utrecht, The Netherlands), the E4orf6 mAb was from Thomas Dobner (University of Regensberg, Germany), the S8 pAb was from Wenlan Wang (AI duPont Hospital for Children, DE, USA) and the TBP pAb was from Nouria Hernandez (Cold Spring Harbor, NY, USA). The 20S mAb and α4 and α6 antibodies were from Biomol. The Ran mAb was from Transduction Laboratories. The N‐FLAG mAb was from Sigma. The Gal4 mAb was from Santa Cruz.
Ad5 12S E1A, 13S E1A and Ad E1A mutants were cloned into pcDNA3.1 (Invitrogen) for both in vitro transcription/translation and mammalian expression, and pGEX 4T‐1 (Amersham Pharmacia) for bacterial expression. Genomic w.t. and mutant Ad5 E1A clones, under the control of the E1A enhancer/promoter region, in expression vector pLE2, were also used for transient mammalian expression (Jelsma et al, 1988). Human proteasomal ATPase cDNAs in pBS were from Carlos Gorbea (University of Ohio at Salt Lake City, USA). S8 was also cloned into pcDNA3‐N‐FLAG. Gal4DBD, Gal4DBD‐Ad5 CR3 and the Gal4‐luciferase reporter were cloned into pcDNA3. The Ad5 E4 CAT reporter gene has been described (Dery et al, 1987). Sequences were verified by sequencing using an ABI Prism 3100 Genetic Analyzer.
Supplementary data are available at The EMBO Journal Online.
This work was supported by CR UK programme grant C2/A1113 and CIHR grant MOP‐74647. MR was sponsored by a studentship from the Iranian government. AFY is a scholar of the CIHR/UWO Strategic Training initiative in Cancer Research and Technology transfer.
- Copyright © 2006 European Molecular Biology Organization