Cells and Viruses

Vero cells were maintained as suggested by ATCC. E5 cells express complementing levels of ICP4, and have been described previously [22]. The viruses, wild type strain KOS and n208 [23], have been previously described. TAP-KOS and TAP-n208 were generated for this study by marker transfer with the TAP-pK1-2 and TAP-pn7 plasmids (described below), and their genotypes were confirmed by Southern blot and sequencing.

Recombinant Plasmids

TAP-ICP4 plasmids were constructed such that the TAP sequence would be in frame and 5’ to the ICP4 translation start site. The 77 amino acid TAP tag consists of a tandem arrangement of the calmodulin binding protein and the streptavidin binding protein from the plasmid pNTAP-B (Statagene). wtICP4 and n208 plasmids, pk1-2 and pn7 were utilized to construct the TAP-pk1-2 and TAP-pn7, respectively, by the RecET recombination system, as described previously [24,25]. The primers used in the initial recombination step to insert the TAP tag contained homology to ICP4 -60 to -20 and -7 to +31, relative to the ICP4 translation start site.

Streptavidin Affinity Purification

2 x 10⁸ Vero cells were infected at a multiplicity of 10 pfu/cell with KOS, TAP-wtICP4, n208, or TAP-n208 for 3, 6, or 12 hours at 37°C prior to harvesting. Cells were rinsed twice with TBS containing 0.1 mM N-p-tosyl-L-lysine chloromethyl ketone (TLCK) and scraped into 50 mL TBS + 0.1 mM TLCK. Cell pellets were collected by centrifugation at 5 K for 5 min in the Sorvell RC5 centrifuge. Nuclei were collected and proteins were extracted in 0.4 M KCl as described previously [20]. Nuclear extracts were diluted in half in 2 X low salt streptavidin binding buffer (2 X LS SBB) (20mM Hepes KOH pH 7.9, 20 mM KCl, 4 mM EDTA, 0.2 % NP40), treated with 300 U benzonase nuclease (Novagen) and incubated with 500 uL streptavidin conjugated agarose beads (Pierce) overnight at 4°C with end over end rotation. Beads were washed 4 times in 10 mL of streptavidin binding buffer (SBB) (20 mM Hepes KOH pH 7.9, 200 mM KCl, 2 mM EDTA, 0.1% NP40). Samples were eluted from the beads using 500 uL streptavidin elution buffer (SBB + 2 mM Biotin). Two elutions, each 30 min, and one elution, overnight, were combined and either concentrated in 7 mL/9 kDa iCon concentrator (Pierce) and used for western blot analyses, or sent to MS Bioworks (Ann Arbor, MI) for TCA precipitation, in gel digestion, and LC-MS/MS analysis.

LC-MS/MS Analysis

Samples were sent to MS Bioworks (Ann Arbor, MI) for processing. Breifly, protein samples were precipitated with TCA and separated on a 10% Bis-Tris Novex mini-gel using a MES based system, stained with Coomassie stain, and the gel was divided into 10 equal sized slices. Gel fragments were processed using the ProGest (DigiLab) robot. The gel slices were washed with 25 mM ammonium bicarbonate followed by acetonitrile. The samples were reduced with 10 mM DTT at 60°C followed by alkylation with 50 mM iodoacetamide at room temperature. The protein samples were digested with trypsin (Promega) for 4 hours at 37°C which was quenched with formic acid. The supernatant was then loaded onto a Waters NanoAcquity HPLC system interfaced to an Orbitrap Velos Pro (Thermo-Fisher). Data was analyzed against the Rhesus macaque and HHV-1 proteomes using the MASCOT program and validated using Scaffold software.

Electrophoresis, Silver Stain, and Western Blot

SDS polyacrylamide gel electrophoresis was carried out essentially as previously described [26]. Silver stained gels were processed according to the manufacturers directions (Pierce). For western blotting, proteins were transferred to either polyvinylidine fluoride (PVDF; Amersham) for chemi-luminescent detection with ECL reagent (Amersham) or nitrocellulose membranes for infrared detection by fluorescence using the LI-COR Odyssey infrared imager according to the manufacturers instructions. The antibodies used to probe the membranes were; N15, polyclonal rabbit serum, for ICP4 (1:500), Trap220/Med 1 (sc-8998-X), Med6 (sc-9434), Med 4 (NBP1-84977; Novus Biologicals), Med23 (550429; BD Bio), CDK8 (sc-1521), Med12 (NB100-2357; Novus Biologicals), Med 13 ((NB100-60642; Novus Biologicals), Med 26 (sc-166614), TAF1 (sc735), TBP (233-R Covance), p62 (sc292-X), p80/XPB (sc-20696-x), CPSF2 (sc-165983), CPSF7 (NB100-61600; Novus Biologicals), Brahma (sc-6450), CHD3 (ABE86; Milipore), RUVBL1 (06-1299; Milipore), RUVBL2 (06-1300; Milipore) at the manufacturers recommended dilutions.

TFIID- and Mediator-ICP4 Complexes

All the components of TFIID and approximately 60% of the components of Mediator were identified in complex with TAP-wtICP4 by tandem LC-MS/MS. The number of spectral counts and the percent coverage for the detected components of TFIID and Mediator are listed in Table 1. While both of these parameters reflect the relative amounts of the proteins present, table 1 also lists the number of spectral counts divided by the molecular weight of the protein. This is to partially account for the fact that larger proteins will generate more spectral counts by virtue of their increased size. This parameter, the number of spectral counts/molecular weight, was used as a measure of the relative amounts of a given protein, and is plotted for the TFIID and Mediator components found in the samples from the wt virus infections in Figure 3. An important observation in this analysis was that there were no spectral accounts for TFIID or Mediator components recovered in any of the experiments in the parallel control lacking the TAP tag (see Figure 1B, KOS lanes, for example).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Normalized spectral counts for TFIID and Mediator components copurifying with ICP4 as a function of time post infection.

A. The normalized spectral counts/MW for each component of TFIID detected by LC- MS/MS are plotted as a function of time post infection. B. The normalized spectral counts/MW for each component of Mediator detected by LC- MS/MS are plotted as a function of time post infection.

https://doi.org/10.1371/journal.pone.0078242.g003

More spectral counts of TFIID components were isolated with affinity purified ICP4 than any other cellular proteins. In fact the number of spectral counts for some of the TFIID components exceeded 10% of the number of spectral counts of ICP4 when normalized for MW (Table 1). The components of TFIID were found in complex with ICP4 as early as 3 hpi and these interactions were retained throughout infection. However, the relative abundances of the TAFs decreased over the duration of infection (Table 1, Figure 3A). TAFs 4,5,6, 9B and 12 were present in the greatest amounts in TAP-wtICP4 samples, with their normalized spectral counts being nearly double any other component of TFIID (Figure 3A). Interestingly, these TAFs are thought to be the core components of TFIID [27]. Additionally, TAF4 and TAF9B where detected relative to TAF4B and TAF9, respectively. In addition to the core TAFs, TBP and TAFs 1, 2, 3, 7, 8, 10 and 11 were detected. To our knowledge no viral activator has been previously shown to interact with TFIID complexes containing such a comprehensive collection of TAFs.

Mediator is a coactivator composed of approximately 30 subunits, divided into 4 modules, the head, middle, tail and kinase. Eighteen of the Meditator subunits distributed across the four modules were isolated with TAP-wtICP4 (Table 1). Western blot analyses were conducted on independently generated TAP and control samples. MED 6 of the head, MEDs 1 and 4 of the middle, MED 23 of the tail, and MEDs 12, 13, and CDK 8 of the kinase modules were all present in the TAP samples and absent in the parallel untagged control sample confirming the LC-MS/MS results (Figure 4). Of note, Med 26, which serves as a docking site for elongation factors and is associated with a coactivator form of Mediator [28,29], was absent from all preparations (Figure 4, Table 1). Additionally, all four subunits of the kinase module, which is typically associated with a repressive function of Mediator [30,31], were identified by LC-MS/MS.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Components from all modules of Mediator are found in TAP-wtICP4 samples.

Western blot analysis of nuclear extracts (NE) or affinity purified (PD) samples using either wtICP4 (KOS) or TAP-wtICP4 (TAP) expressing viruses at 6 hpi. Results of the Western blot analysis for components of the head (Med6), middle (Med4 and Med1), tail (Med23), kinase (cdk8, Med12, and Med13), and the Med26 protein are shown.

https://doi.org/10.1371/journal.pone.0078242.g004

The number of spectral counts of mediator components that copurified with ICP4 was considerably less than for TFIID components (Table 1, Figure 3). Also unlike with TFIID, only 5 subunits of Mediator were identified by LC-MS/MS at 3 hpi. While the normalized spectral counts for TFIID components decreased from 3 to 6 hpi the normalized spectral counts for all the detected components of Mediator in the same sample increased during this time (Table 1, Figure 3). This indicates while ICP4 is found in complexes with TFIID at 3 hpi, the associations between ICP4 and Mediator become more prevalent later in infection (6 h). The situation is more complex at 12 hpi. Here several components remained relatively abundant (Meds 20, 31, 18, and 17), while the remainder of the detected Med components decreased in abundance relative to the 6h sample. Three of these proteins are components of the head module, suggesting the specific retention of this part of the mediator in ICP4-containing complexes late in infection. This is further discussed below.

To confirm the temporal differences observed in the isolation of TFIID and Mediator with ICP4, independent samples were subjected to Western blot analysis and the fluorescent intensities of the resulting bands were quantified using a Licor Odyssey infrared imager (Figure 5). Figure 5A is the image of the Western blot of TAP-wtICP4 samples collected at 3 and 6 hpi probed with antibodies to ICP4, TAF 1, TBP, and MED 1, while Figure 5B is a graphic representation of the protein quantification analyses, normalized to the amount of ICP4. There was a 2-3-fold reduction in levels of TAF 1 and TBP at 6 hpi compared to 3 hpi (Figure 5B). Conversely, there was a 3-fold increase in the amount of MED 1 present in samples collected at 6 hpi compared to 3 hpi (Figure 5B). Together these data are consistent with the LC-MS/MS analyses and indicate that TFIID associates with ICP4 before Mediator.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. TFIID and Mediator display different temporal associations with ICP4.

(A) Western blot analysis of nuclear extracts (NE) or affinity purified (PD) samples collected at either 3 hpi or 6 hpi using either wtICP4 (KOS) or TAP-wtICP4 (TAP) expressing viruses. Antibodies directed against ICP4, TAF1 and TBP of TFIID, and Med1 of Mediator were used. (B) Quantification of the western blots in (A). Images were quantified using the LiCOR Odyssey infrared imager and normalized to the amount of ICP4 present within the sample.

https://doi.org/10.1371/journal.pone.0078242.g005

To determine whether the C-terminus of ICP4 was required for interactions with transcription complexes, potentially accounting for the observed deficiencies in L gene expression, TAP-n208 was purified at 6 hpi and the co-purifying proteins were compared to those isolated with TAP-wtICP4 at 6 hpi. Notably, all 12 TAFs and TBP were isolated with TAP-n208, albeit some TAFs were less abundant when compared to those isolated with TAP-wtICP4 (Table 1). Of the core TAFs, TAFs 4 and 12 were present at similar levels in both the TAP-n208 and TAP-wtICP4 samples, while TAFs 5 and 6 were decreased and TAF 9 was increased in the TAP-n208 samples (Table 1). The decreased levels of some, but not all, core components of TFIID in TAP-n208 samples could be indicative of direct interactions between the N-terminus of ICP4 and certain core components of TFIID such as TAF 9B, however this remains to be experimentally determined.

A comparison of proteins isolated from TAP-wtICP4 and TAP-n208 samples revealed that some components of Mediator were comparably isolated. These included MEDs 4, 6, 18, 20, 23, and 27 (Table 1). Conversely, other components of Mediator were less abundant in TAP-n208 purified samples. These include MEDs 1, 12, 13, 15, 16, 17, 25, cyclin C, and cyclin dependent kinase (CDK) 8. The existence of Mediator components in TAP-n208 samples suggests that the C-terminus of ICP4 is not necessary for interaction with Mediator. Interestingly, many of the Mediator components that were reduced in abundance in the TAP-n208 sample (7 out of 9) belong to the tail and kinase region of Mediator, perhaps suggesting an interaction between the N-terminal transactivation domain of ICP4 and a component of the head or middle regions of Mediator. To validate these results, Western blot and protein quantification analyses were performed on independently isolated TAP-wtICP4 and TAP-n208 samples collected at 6 hpi using antibodies to ICP4, TAF 1, TBP, and MED 1 (Figure 6A). Protein quantification confirmed that TAF 1, TBP, and MED 1 were less abundant in TAP-n208 samples compared to TAP-wtICP4 samples (Figure 6B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. TAP-n208 associates with both TFIID and Mediator.

(A) Western blot analysis of nuclear extracts (NE) or affinity purified (PD) samples collected at 6 hpi using either wtICP4 (KOS), TAP-wtICP4 (TAP), the C-terminal truncation mutant (n208) or a TAP-tagged C-terminal truncation mutant (TAP-n208) expressing viruses. Antibodies directed against ICP4, TAF1 and TBP of TFIID, and Med1 of Mediator were used. (B) Quantification of the western blots in (A). Images were quantified using the LiCOR Odyssey infrared imager and normalized to the amount of ICP4 present within the sample.

https://doi.org/10.1371/journal.pone.0078242.g006

Novel ICP4-Containing Complexes

In addition to TFIID and Mediator, other transcription complexes were isolated with TAP-wtICP4. These include TFIIH, the cleavage and polyadenylation specificity factor (CPSF) complex, and chromatin remodeling complexes. These factors have roles in initiation, elongation, and mRNA processing, implying that ICP4 may function at stages of transcription beyond initiation.

TFIIH is a multi-subunit protein involved in promoter clearance [32-35]. It is composed of 5 subunits that form a core, and a dissociable catalytic activating kinase (CAK) module composed of MAT1, cyclin H, and CDK7. All 5 of the core components of TFIIH were identified by LC-MS/MS in TAP-wtICP4 samples at 6 hpi, while cyclin H and CDK7 of the CAK were not detected (Table 2). The kinase domain is partially responsible for phosphorylating the C-terminal domain of RNA PolII. Its absence in TAP-wtICP4 samples may reflect the observation that RNA PolII is alternatively phosphorylated in HSV infections [36]. Western blot analyses using antibodies against p80 (XPD) and p62 from the core of TFIIH confirmed their presence in the TAP-wtICP4 samples (Figure 6A). As Mediator is involved in recruiting TFIIH to preinitiation complexes [37], it is possible that TFIIH is present in the ICP4-containing complexes because it interacts with Mediator.

Components of the cleavage and polyadenylation specificity factor complex (CPSF) were also isolated in TAP-wtICP4 samples (Table 2). The CPSF complex is responsible for recognizing and cleaving the pre-mRNA such that the PolyA polymerase can polyadenylate the nascent transcript [38]. All but two of the CPSFs, 5 and 6, were identified by LC-MS/MS. Western blot analyses using antibodies against CPSF2 and CPSF7 confirmed their presence in TAP-wtICP4 samples (Figure 7B). In TAP-n208 samples, components of the CPSF complex were isolated, although again the association did not appear to be as strong as when the C-terminus of ICP4 was present (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 7. ICP4 containing complexes include TFIIH, cleavage and polyadenylation complex and chromatin remodeling complexes.

Western blot analysis of nuclear extracts (NE) or affinity purified (PD) samples collected at 6 hpi using either wtICP4 (KOS) or TAP-wtICP4 (TAP) expressing viruses. Antibodies directed against (A) p80 and p62 of TFIIH, (B) CPSF2 and CPSF7 of the cleavage and polyadenylation complex and (C) Brahma of Swi/Snf, CHD3 of the Nurd complex, and RuvBL1 and RuvBL2 of the Ino80 complex were used.

https://doi.org/10.1371/journal.pone.0078242.g007

In addition to the many components involved in transcription initiation and elongation, chromatin remodeling factors, including the SWI/SNF, Nurd, and Ino80 chromatin remodeling complexes were isolated in TAP-wtICP4 samples (Table 2). Chromatin remodeling complexes hydrolyze ATP to provide the energy to disrupt histone/DNA interactions, thus allowing for histone displacement during transcription, DNA replication, and DNA repair [39]. As such, a common characteristic of chromatin remodeling complexes is the presence of an ATPase domain. The ATPase containing proteins of the SWI/SNF, Nurd, and Ino80 complexes are Brahma related gene 1 (Brg1) & Brahma (Brm), CHD 3 and 4, and RuvBL 1 and 2, respectively [39]. With the exception of Brg1, each of these ATPases co-purified with TAP-wtICP4 (Table 2). The isolation of these ATPases was confirmed by Western blot analysis with antibodies directed against Brm, CHD 3, RuvBL 1 and RuvBL 2 (Figure 7C). In addition to the ATPase containing proteins, multiple adaptor proteins from each of these complexes were isolated, including SMARCA4, SMARCC2, and Bzb1 kinase from the SWI/SNF complex, HDAC2 from the Nurd complex, and Arp5 from the Ino80 complex (Table 2).

TAFs associated with ICP4

We have previously shown that ICP4 interacts directly with TBP and TAF1 in vitro [17,18]. Affinity purification of ICP4 from HSV infected cells resulted in the isolation of TBP and TAFs1, 2, 3, 4, 5, 6, 7, 8, 9B, 10, 11, and 12. The numbers of spectra counts of ICP4 associated TAFs were among the greatest of all the protein identified in this study, suggesting that TFIID is the major interacting partner of ICP4. Additionally, the abundances of the core TAFs, 4, 5, 6, 9B, and 12 were greater than those of the non-core TAFs in these samples. It has been proposed that the core TAFs are present twice within holo-TFIID [40,41]; therefore, the abundance of the core TAFs in TAP-wtICP4 samples may simply reflect the stoichiometry of the isolated TFIID. An alternative hypothesis is that the abundance of core TAFs is the consequence of a direct interaction between ICP4 and one or more of the core components of TFIID. Interestingly, only TAF4 and TAF9B were abundant in both TAP-wtICP4 and TAP-n208 samples. TAF9B is distinct from TAF9, but like TAF9 forms a histone like pair with TAF6 and is involved the transcription of a distinct but overlapping set of cellular genes relative to TAF9 [42]. Likewise, TAF4 and TAF4B are distinct, form a histone like pairs with TAF12 and are involved in TFIID binding to a distinct set of cellular genes [43]. The precise definition of the common characteristics of cellular genes that are differentially affected by TAFs4 and 9B are not known at present, but it is possible that they might share characteristics with viral or cellular gene that ICP4 regulates. The selection of TAF4 and TAF9B in ICP4 containing complexes also suggests a possible direct interaction between ICP4 and these TAFs. Experiments to test this hypothesis are ongoing.